Eortljon; H. K. LEWIS AND CO, Ltd., 136, GOWER STREET, W.C. 1 EonSon: WILLIAM WESLEY AND SON, 28, ESSEX STREET, STRAND, W.C. 2nbsp;iHenj lorfe; G. P. PUTNAM�S SONSnbsp;Dombat, ffnlcutta antr fKaSraa: MACMILLAN AND CO., Ltd.

FOSSIL PLANTS

- A TEXT-BOOK FOR STUDENTS

OF BOTANY AND GEOLOGY

A. C. SEWARD

PKOFESSOE OF BOTANY IN THE �NIVEESITY ; MASTEll OP DOWNING COLLEGE AND HONOBABYnbsp;FELLOW OF EMMANUEL COLLEGE, CAMBRIDGE

GINKGOALES, CONIFERALES, GNETALES

CAMBRIDGE:

PREFACE

IN the Preface to Volunre III I expressed my thanks for help given to me by many friends in the course of the preparationnbsp;of the subject-matter of Volumes III and IV, but Dr Scott hasnbsp;again earned my gratitude by very willingly and to very goodnbsp;purpose continuing the tedious task of reading the proofs. It isnbsp;also a pleasure to acknowledge the help received from the Staffnbsp;of the University Press.

Since the publication of Volume III Palaeobotany has been deprived of the services of three senior investigators, Professornbsp;C. E. Bertrand of Lille, Monsieur Grand�Eury, and Mr Clementnbsp;Reid, men whose researches along different lines of inquiry havenbsp;played a prominent part in the progress of the science duringnbsp;the last few decades. By the death of Miss Ruth Holden, anbsp;graduate of Harvard University and a Fellow of Newnhamnbsp;College, Cambridge, Palaeobotany has lost an unusually giftednbsp;and promising worker; though a citizen of a country which wasnbsp;then neutral her strong sense of duty led her to lay aside,nbsp;temporarily as we hoped, botanical research for work with anbsp;British Medical Unit in Russia where she died in April of lastnbsp;year. Miss Holden�s last contribution to Palaeobotany (�On thenbsp;Anatomy of two Palaeozoic stems from India �; Annals of Botany,nbsp;vol. XXXI. p. 315, 1917) was published too late to be considerednbsp;in Volume III.

If it is possible to carry out my intention of supplementing the descriptive treatment of plants, which forms the basis ofnbsp;Volumes I-IV, by a general review of the Floras of the Past thenbsp;results will be published as an independent work more intelligible,nbsp;I hope, to the general reader than the text-book which, with anbsp;certain sense of relief, is now brought to a conclusion.

POSTSCEIPT

Through the death of Dr Newell Arber in June of last year at the comparatively early age of forty-seven Botany and Geologynbsp;have lost an able and indefatigable investigator. Since 1901,nbsp;the date of publication of his first paper, he laboured incessant!}quot;nbsp;and with success to advance palaeobotanical science. To hisnbsp;activity the Sedgwick Museum of Geology at Cambridge isnbsp;deeply indebted, and through his personal influence severalnbsp;younger men acquired something of their teacher�s enthusiasm.

The proofs of this Volume were passed for press in June, 1918, but owing to the exigencies of war conditions publicationnbsp;has been unexpectedly and inevitably delayed,

TABLE OE CONTENTS

G-OITES 10-33 ; Ginkgocladua 33 ; BAIERA 33-50; iii. Flowers and Seeds 50-58; STENORACHIS 54-58;nbsp;EEETMOPHYLLUM 58-60.

CHAPTER XLI

GEXERA BELIEVED TO BELONG TO THE GINKGOALES BUT WHICH ON THE AVAILABLE EVIDENCE CANNOTnbsp;BE REFERRED WITHOUT HESITATION TO THAT GROUP.

GINKGODI�M 61,62; CZEKANO WSKIA 62-68; FEILD-ENIA 68, nbsp;nbsp;nbsp;69; PHOENICOPSIS and DESMIO-

CHAPTER XLII

GLOTTOPHYLL�M76; GINKGOPSIS 77; NEPHROPSIS 77, 78; PSYGMOPHYLLUM 79-90; RHIPIDOPSISnbsp;90-92; SAPORTAEA 92, 93; DICRANOPHYLLUMnbsp;93-101; TRICHOPITYS 101-103 ; SEWARDIA 103-105.

CHAPTER XLIII

VII. PITYOXYLON 219-231; VIIL PEOTOPICE-OXYLOF 231-234; IX. WOODWORTHIA 234, 235;

X. AEAUCARIOPITYS 235, 236; XI. PROTO-CEDROXYLON 236-238; XII. XENOXYLON 238-242; XIII. ANOMALOXYLON 242,243; XIV. THYL-LOXYLON 243 ; XV. PLANOXYLON 244.

CHAPTER XLV

CONIFERALES (continued). Pp. 245�30].

DAMMARITES, PROTODAMMARA 245-252; Conites 252-256; ARAUCARITES 256-270; ELATIDES 270-274; PAGIOPHYLLUM 274-277.

Palaeozoic Conifers exhibiting certain features suggestive OP Araucarian affinity but which cannot be definitely assigned to that or to any other family

282-286; HAPALOXYLON 286, 287; GOMPHOSTRO-BUS 287, 288; VOLTZIA 289-295; SWEDENBORGIA 295, 296; Strobilites 296; ULLMANNIA 296-299;

CHAPTER XLVI

CONIFERALES (continued). Pp. 302�335.

CUPRESSINOOLADUS 303, 304, 307-310; THUITES 305; CUPRESSITES 305 307; ATHROTAXITESnbsp;311-315; BRACHYPHYLLUM 315-322, .324-328;nbsp;BRACHYOXYLON 322 324; TAXODITES 328-331;nbsp;CUNNINGHAMIOSTROBUS 331, 332; MORICONIAnbsp;332, .33.3; CRYPTOMERITES 334, 335; CRYPTO-MERIOPSIS 3.35.

CHAPTER XLVII

CONIFERALES (continued). Pp. 336�367.

SEQUOIITES 349-351, 352-357 ; Conites 351, .352 ; GEI-NITZIA 357-361; EUGEINITZIA 361, 362 ; PSEUDO-GEINITZIA 362; SPHENOLEPIDIUM 363 365.

CHAPTEE XLVIII

CONIFEEALES (continued). Pp. 368�4=04.

PITYITES 370-376; PITYOCLADUS 377-380; PITYO PHYLLUM 380, 381; PITYOSTROBUS 381-395nbsp;PITYANTHUS 395,396 ; PITYOSPERMUM 396-398nbsp;PITYOSPORITES 398, 399; ENTOMOLEPIS 399nbsp;CROSSOTOLEPIS 400; PREPINUS 400-404.

CHAPTEE XLIX

CONIFEEALES (continued). Pp. 405�423.

PODOCARPINBAE.........

PODOCARPITES406-409; STACHYOTAXUS 410-412; Strobilites 412, 413; Saxegothopsis 413.

........................

PHYLLOCLADITES 413, 414; Frotophyllodadm^X�, 416; Palae,ocladus 417 ; Phyllooladopsis 417.

Taxineae..........

TAXITES 417-419; TORREYITES 419-421; VESQUIA 422; CEPHALOTAXOPSIS 422, 423.

CHAPTEE L

CONIFEEALES INCEETAE SEDIS. Pp. 424�446.

TRIOOLEPIS 424; MASCULOSTROBUS 424-426; PA-LISSYA 426-428; ELATOCLADUS 429-432, 435, 436;

RETINOSPORITES 432, 433; ANDROVETTIA 436-438; DACTYOLEPIS 438; RARITANIA 438, 439; SCHIZOLEPIS 439-442; DREPANOLEPIS 442; SCHI-ZOLEPIDELLA 442, 443; OYPARISSIDIUM 443-445 ;nbsp;BENSTEDTIA 445, 446.

CHAPTEE LI

PODOZAMITES AND NAGEIOPSIS: GENEEA INCEETAE SEDIS., Pp. 447�458. .

CHAPTEE LII GNETALES. Pp. 459�472.

LIST OF ILLUSTRATIONS

635. nbsp;nbsp;nbsp;Oinkgoites adiantoides; 0 pluripaHita; O digitata; O. sibirica

652. nbsp;nbsp;nbsp;Baiera Phillipsi .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.

664. nbsp;nbsp;nbsp;Oinkgopsis Czekanowskii; Nephropsis integerrima. M. Zalessky

Xlll

PAGE

86 87nbsp;91

95, 97 102nbsp;104nbsp;106nbsp;107nbsp;109

109

110

111

112

113

114

116

119

120 121nbsp;126nbsp;130

132

133 137nbsp;141

144

145

146

147

148 150

152, 153

FIG.

667.

668.

669.

670.

671.

673.

674. 674*

675.

676.

677.

678.

679.

680. 681.nbsp;682.

683.

684.

685.

686.

687.

688.

689.

690.

691.

692.

693.

694. 69.5.

696.

697. . 698.

699.

quot;00.

701

702,

703.

Psygmophyllum Haydem. [Geolog. Survey, India. Psygmophyllum Orassertinbsp;Rliipidopsis ginkgoides. M. Zalesskynbsp;672. Dicranophyllum gallicumnbsp;Trichopitys heteromorpha.

. Sequoia sempervirens. A. Henry Phyllocladus trichommioides; P. liypophylla; P. alpinanbsp;Podocarpus latifolianbsp;Cryptomeria japonica

Pinus monophylla; Taxodium mucronatum; T. dislichum Sazegothaea conspicua. [New Phytologist.] ...nbsp;Torreya nucifera. [New Phytologist.] ....

Sequoia sempervirens; Abies sp. E. C. Jeffrey Araucaria Bidunllii ; A. Cookii; Agathis bornensis ; Larix amerinbsp;cana. R. B Thomson ......

Sequoia giganiea . nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;. -nbsp;nbsp;nbsp;nbsp;.

704. nbsp;nbsp;nbsp;Pinus excelsa .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;. T......

711. nbsp;nbsp;nbsp;Veronica Hectori; Callitris calcarata; Veronica cupressoides

729. nbsp;nbsp;nbsp;Protopiceoxylon Edwardsi. [Trustees of the British Museum.�

733. nbsp;nbsp;nbsp;Dammarites borealis. A. G. Nathorstnbsp;734-736. Gonites Juddi. [Royal Society of Scotland.]

Walchia piniformis Schizodendron speciosumnbsp;Oomphostrobus bifidusnbsp;Voitzia heterophylla; V. Liebeananbsp;Swedenborgia cryptomeroides .

Ullmannia frumentaria ; U. Bronni Cupressites taxiformisnbsp;Cupressinocladus salicornoidesnbsp;Aihrotaxites Ungeri; A. lycopodioidesnbsp;Brathyphyllum expansum

, 756. Brachyphyllum expansum. [Geolog. Survey, India.] Braehyphyllum obesumnbsp;Brachyoxylon notabile; Aramariopitys americana; Protodam-mara speciosa; Brachyphyllum crassum. E. C. Jeffreynbsp;Taxodites europaevsnbsp;Mamp;riconia cyclotoxonnbsp;Calliirites curia

Callitrites Brongniarti; C. helvelica; G. europaeus Frendopsis Hoheneggeri; F. ram.osissimanbsp;Sequoia magnifica. [CamVj. Univ. Press.]

Pityostrobus Benstedi. [Trustees of the British IMuseum.] Pityostrobus oblongus. [Trustees of the British Museum.]nbsp;Pityostrobus (Pinites) Andraeinbsp;784. Pityostrobus (Pinites) macrocephalusnbsp;Pinites Plutonis

798. nbsp;nbsp;nbsp;Protophyllocladus subintegrifolius; Brepanolepis rotundifolia

799. nbsp;nbsp;nbsp;Masculostrobns Zeilleri; Masculostrobus sp. [Royal Society of

806. nbsp;nbsp;nbsp;Androveltia statensis; Geinitzia Reichenbachii. E. C. Jeffrey

PAGE

401

406

408

409 411nbsp;413

416

425

428

430

431

433

434 437nbsp;437nbsp;439nbsp;441

443

444 448nbsp;450nbsp;453

456

457

458 463nbsp;466

CHAPTER XL,

In the account of the Coniferae contributed to Die Natiir-lichen Pjkmzenfamilien} the genus Ginkgo, in accordance with the prevailing custom, was included in the Taxeae with Taxus, Cephalo-faxus, and Torreya. Eichler had previously referred Ginkgo, ornbsp;Salishuria, to a separate family, the Salisburyeae^. Hirase�snbsp;discovery of motile antherozoids in the pollen-tube of Ginkgonbsp;hiloba in 1896, �the most remarkable event in plant morphologynbsp;during the last decade of the 19th century,� confirmed the suspicion that the association of this �unicum de la cr�ation actuelle�nbsp;with Taxus and other Conifers was inconsistent with a naturalnbsp;scheme of classification. At a later date Engler adopted thenbsp;family-name Ginkgoaceae, and in his survey of the Embryophytanbsp;Siphonogama the isolation of Ginkgo is emphasised by the reference of the Ginkgoaceae to a special class, the Ginkgoales�.

Ginkgo hiloba L. {Salisburia adiantifolia Smith) has a preeminent claim to be described in Darwin�s words as a �hving fossil.� It isnbsp;sometimes said to occur in China as a wild plant, but there appearsnbsp;to be no sufficient reason to believe that it would have escapednbsp;extinction had it not been carefully tended as a sacred tree innbsp;the gardens of temples^. Since its introduction into Europe innbsp;1730, Ginkgo has become familiar in cultivation in the northernnbsp;hemisphere and thus through man�s agency this monotypic genusnbsp;has been restored to regions where it survived as late as thenbsp;Tertiary epoch. In habit Ginkgo^ resembles many Conifers andnbsp;its long and short shoots recall those of Cedrus and Larix; thenbsp;short shoots may also be compared with the main trunk of a

^ Engler (97) pp. 19, 341. nbsp;nbsp;nbsp;quot; Elwes and Henry (06) Vol. I. p. 58.

Cycad. The large thin leaves with long and slender petioles are scattered on long shoots or crowded on slow-growing branchesnbsp;covered with leaf-scars. These short shoots are occasional!)^nbsp;branched^ and, as Tupper^ states, they may branch within thenbsp;wood of the axis out of which they grow, a feature exhibited bynbsp;the Triassic Conifer Woodworthia. The deciduous leaves arenbsp;usually more or less deeply bilobed (fig. 630, A, D, F) but those onnbsp;short shoots are often smaller, and the margin may be entire ornbsp;uneven (fig. 630, C). On young and vigorous shoots or on seedlings the lamina is often deeply divided into several cuneatenbsp;segments (fig. 630, E). In exceptional cases the lamina maynbsp;reach a breadth of 20 cm. (fig. 630, A) though as a rule it seldomnbsp;exceeds half that size. The leaf-scars show two small cicatrices.nbsp;The considerable range in size and form of the foliage-leaves isnbsp;an important consideration in connexion with the determinationnbsp;of fossil specimens. Two vascular bundles pass up the petiole:nbsp;at the summit each divides and the two outer branches follownbsp;the outer edge of the lamina, giving off a succession of forkednbsp;veins. Objection is taken by Prof. Johnson� to the statement thatnbsp;there are two marginal veins on the lower edge of the lamina;nbsp;he regards the �marginal� vein as the product of the successivenbsp;fusions of the forked veins of the lamina as they pass towards thenbsp;leaf-base. Whatever interpretation is adopted, the presence ofnbsp;two broadly diverging marginal veins is a noteworthy feature,nbsp;and the correct explanation is probably that each is derived fromnbsp;one of the two strands in the petiole and gives off a succession ofnbsp;dichotomously branched veins as it passes along the margin ofnbsp;the leaf-blade. The presence of short secretory tracts at intervalsnbsp;between the veins is a characteristic feature sometimes recognisable in fossil examples. Throughout the greater part of theirnbsp;course in the lamina the veins are accompanied by a small numbernbsp;of reticulate transfusion-tracheids (fig. 631, G, t): these increasenbsp;m amount near the distal end of each vein and the water-conducting elements may be eventually replaced by a group of short,nbsp;pitted^ tracheids^. A group of large cells with brown contentsnbsp;occurs above and below each collateral endarch bundle. The

^ Seward and Gowan (00) B. PI. ix. fig. 42. nbsp;nbsp;nbsp;* Xupper (11) p. 376.

Johnson (14) p, 171. nbsp;nbsp;nbsp;� Sprecher (07) pp. 68�71; Bertrand, C. E. (74).

stomata, irregularly scattered over the lower epidermis, consist of two guard-cells surrounded by 4�6 accessory cells which project towards the centre of the stoma as blunt cuticularised papillae^nbsp;(fig. 636, C). The epidermal cell-walls are slightly undulate^.nbsp;The distinctive form of Ginkgo leaves renders almost negligiblenbsp;the danger of confusion with those of other Gtymnosperms; butnbsp;impressions of certain Fern fronds, e.g. Ldndsaya reniformis Dry.,nbsp;Plerozonium {Gymnogramme) reniforme Mart., Trichomanes reni-forme Forst., and Scolopendrium nigripes Hook, might he mistakennbsp;for imperfectly preserved specimens of Ginkgo.

Ginkgo is dioecious. The male flowers occur in loose catkins (fig. 631, B) borne on short shoots in the axil of a scale-leaf: eachnbsp;microsporophyll consists of a short, slender, filament with a smallnbsp;terminal scale or knob bearing as a rule 2 but not infrequentlynbsp;3 or 4 elliptical microsporangia (fig. 631, A, A'). The microsporesnbsp;recall those of Cycads. Jeffrey� has recently called attention tonbsp;the occurrence of � wings � on the microspores of Ginkgo: these arenbsp;very slightly developed and hardly warrant the use of the termnbsp;wing; they present the appearance of very small shoulders givingnbsp;the spores a form similar to that of a brachiopod shell. The samenbsp;author expresses the view that Ginkgo presents striking resemblances to the Abietineae. It has recently been pointed out thatnbsp;the microsporangia have a hypodermal layer of cells with thickening bands comparable with the fibrous layer in the anthers ofnbsp;Angiosperms^. Jeffrey and Torrey� claim that certain anatomicalnbsp;features in the microsporangia of Ginkgo indicate a closer affinitynbsp;to the Abietineae than to any other section of the G-ymnosperms.nbsp;The vascular bundles of the microsporophylls end in transfusion-tissue which passes almost imperceptibly into the mechanicalnbsp;elements of the sporangial wall: a similar distribution of mechanicalnbsp;tissue occurs in Abietineous microsporangia and there is the samenbsp;intimate relationship as in Ginkgo between the tracheary andnbsp;mechanical tissues. The female flower consists of a comparatively long peduncle borne in the axil of a foliage-leaf, with two

Ovules at the summit, one on each side of the actual apex. Frequently one of these is larger than the other. The occasional occurrence of abnormal female flowers is interesting from thenbsp;point of view of palaeobotanical comparison. In extreme cases

the partially modified lamina of a foliage-leaf may bear marginal Ovules, the lamina being continuous with the collar (fig. 631, D, c)nbsp;at the base of the ovule. In other cases the peduncle may givenbsp;ofi several stalked ovules, as in the specimen described by Fujii^nbsp;^ Fujii (96); Seward and Gowan (00) B. PI. ix. figs. 1�.5; Sprecher (07) p. 138.

and reproduced in fig. 631^ D: the apex of the shoot is seen at a. Fertilisation is said to occur after the fall of the ovule^ but Hirasenbsp;states that some seeds contain an embryo while still attached tonbsp;the tree. The seeds are comparable in size with large cherries;nbsp;the broad integument consists externally of a thick sarcotestanbsp;rich in secretory tissue but without a vascular supply^ and annbsp;inner sclerotesta which is usually two-angled (platyspermic) butnbsp;occasionally three-angled and radiospermic (fig. 631, C). Annbsp;account has recently been published^ of some remarkable examplesnbsp;of Ginkgo seeds gathered from one tree: the stony coat showed 2,nbsp;3 and 4 ribs and many transitional forms. The sclerotesta is linednbsp;by a few layers of loose cells which form a papery membrane innbsp;ripe seeds. The absence of vascular tissue in the sarcotesta is anbsp;character in which the seeds differ from those of Cycads andnbsp;Taxads. At the base of the ovule is a single concentric strandnbsp;which splits into two branches and these pass through the shellnbsp;and divide into several bundles on the inner face of the integumentnbsp;forming a continuous mantle^ of short reticulate tracheids as innbsp;the Palaeozoic seed Stephanospermum^. The base of the ovulenbsp;is enclosed in a shallow cup or collar, a structure that is probablynbsp;homologous with the lamina of a foliage-leaf but which has received various interpretations. A suggestion has been made thatnbsp;the collar may be homologous with the cupule of Lagenostoma'^.nbsp;The nucellus is joined to the integument except at the apex wherenbsp;it forms a prominent cone in which a pollen-chamber is developed:nbsp;this chamber becomes roofed over by nucellar tissue and at anbsp;later stage a blunt outgrowth is produced from the summit of thenbsp;prothallus, serving as a �tent-pole� to support the roof of thenbsp;pollen-chamber. There are two or more archegonia at the apexnbsp;of the prothallus differing from those of Conifers and Cycads innbsp;the shorter and more spherical form of the egg-cell and similar tonbsp;those in some Palaeozoic seeds. Fujii� draws attention to thenbsp;remarkable capacity for pollination exhibited by Ginkgo andnbsp;speaks of the conveyance of microspores over a distance of 500nbsp;to 1000 metres. Another fact worthy of remark in view of thenbsp;wide distribution of the Ginkgoales in the Mesozoic era,, i.? tbs

� See VoL in. p. 326. nbsp;nbsp;nbsp;� Shaw, P. J. P. (08).nbsp;nbsp;nbsp;nbsp;^ pujjj (^q) p. 216.

germination of Ginkgo seeds after 45 days� immersion in sea-water^. The embryo has normally two hypogean cotyledons though three are not uncommon. Velenovsky^ mentions a peculiarity, another indication of the isolated position of the genus,nbsp;in which seedlings of Ginkgo differ from those of other Phanerogams ; the cotyledons are succeeded by two elongated scales withnbsp;a forked apex; the next higher leaves, in which a small bilobednbsp;lamina is a characteristic feature, show at the base two divergentnbsp;prongs representing the fork of the lower scales. The lamina ofnbsp;the foliage-leaf thus arises in the angle of the V-shaped distalnbsp;end of the earlier scale-leaf.

A microspore on germination developes 2�3 prothallus-cells and the generative cell forms two large (llOp. x 80/x) spirallynbsp;ciliated antherozoids. After fertilisation the egg-nucleus divides,nbsp;as in some Cycads, until 256 free nuclei are formed�, but in Ginkgonbsp;the subsequent production of walls results in a tissue, called bynbsp;Lyon^ the protocorm, which completely fills the egg; whereas innbsp;Cycas this tissue is massed at the base and in Zamia wall-formationnbsp;IS also restricted. In Conifers the number of nuclei is much lessnbsp;and the proembryo still further reduced. It is probably legitimate to deduce from these facts that Ginkgo is in respect of itsnbsp;embryogeny the most primitive of the Gymnosperms: in this andnbsp;other characters it is allied more closely to the Cycads than tonbsp;the Conifers. Saxton� who has described the later stages in thenbsp;embryogeny of Encephalartos draws attention to certain featuresnbsp;shared by that genus and Ginkge.

The leaf-traces arise from the stele as a pair of collateral bundles, 3�S in the Palaeozoic genus Mesoxylon, which pass up the petiole.nbsp;Annual rings are fairly distinct though, as Mcol� recognised, lessnbsp;obvious than in Conifers. The walls of the late-summer tracheidsnbsp;S'le hardly thicker than those of the spring-elements and the difference between the early and late wood is often slight^. Circularnbsp;oordered pits occur either in a single or double row on the radialnbsp;Walls of the tracheids and are fairl}^ common on the tangentialnbsp;Walls. The pits may be separate or in contact, occasionally

slightly flattened and alternate, but usually opposite (fig. 631, E, F). The pores of the pits are often crossed. Rims of Sanionbsp;form a well-marked feature on the tracheal walls, and Jeffrey^nbsp;points out that they are clearly shown in the mature wood butnbsp;not in close proximity to the primary xylem or in the wood of thenbsp;reproductive shoots and leaves. True bars of Sanio frequentlynbsp;occur on the tracheids^. The secondary phloem consists of discontinuous rows of fibres in addition to sieve-tubes and parenchyma.nbsp;Characteristic features are presented by the uniseriate medullarynbsp;rays: these are often 1�2 or 1�5 cells deep and do not appearnbsp;to exceed 11 cells in depth; they are comparatively large and innbsp;tangential sections of the wood present an inflated appearance.nbsp;There are 2�7 elliptical pits in the field of the ray-cells. Xylem-parenchyma though not abundant occurs here and there amongnbsp;the tracheids; the cells have thin walls and are larger than thenbsp;medullary-ray cells and characterised by the occurrence of stellatenbsp;calcium oxalate crystals�.

The characters of the wood of Ginkgo biloba summarised above are in general agreement with those of many Conifers,nbsp;and such anatomical features as have been described by authorsnbsp;as more or less distinctive of the genus do not afford very trustworthy guides to the identification of fossil wood. The comparatively large size and rounded contour of the medullary-ray cells,nbsp;as seen in tangential section, though Mmrthy of note as characteristic features, are hardly satisfactory criteria when applied to woodnbsp;that may have undergone partial decay and been exposed tonbsp;influences affecting the original form of the more delicate tissuesnbsp;before petrification. The untrustworthy evidence afforded by thenbsp;size of the medullary rays has been emphasised by Essner^ whonbsp;states that the ray-cells of Ginkgo are larger than those in anynbsp;genus of Conifers. It has been claimed bj^ Felix� that Ginkgo is

� For anatomical details, see also Kleeberg (85); Essner(86); Strasburger (91); Seward and Gowan (00) B.; Penhallow (07); Sprecher (07); Burgerstein (08);nbsp;Tupper(ll).nbsp;nbsp;nbsp;nbsp;* Essner (86).nbsp;nbsp;nbsp;nbsp;^ Felix (94).

an exception to the general truth of Essner�s conclusions and that the large dimensions and rounded form of the ray-cells are featuresnbsp;of diagnostic value, though in the Tertiary specimens comparednbsp;by him with the wood of Ginkgo the ray-cells do not appear tonbsp;differ appreciably in size or form from those of true Conifers.nbsp;Given well-preserved material, it is not improbable that in favourable cases the characters of fossil wood might furnish adequatenbsp;grounds for referring it to Ginkgo: the numerous obliquely elliptical pits in each �field,� the swollen medullary-ray cells, and thenbsp;frequent crossing of the pores of the tracheal pits are the featuresnbsp;nientioned by Gothan^ who considers that the wood of Ginkgo�nbsp;though difficult to define precisely in an analytical key�may benbsp;distinguished from that of Conifers.

Among the specimens of wood assigned to Ginkgo there are none, so far as I am aware, that can safely be accepted as entirelynbsp;above suspicion. In 1850 Goeppert^ proposed the generic namenbsp;Physematogntys^ for some Tertiary wood that he believed to possessnbsp;the anatomical characters of Ginkgo biloba. Kraus^ subsequentlynbsp;recognised resin-cells in the wood of Goeppert�s type-species,nbsp;Physematopitys salisburioides, and identified the specimens as thenbsp;root-wood of a Cupressinoxylon: he did not, therefore, includenbsp;Physematopitys in the list of woods contributed by him tonbsp;Schimper�s Trait� de Paleontologie, but mentioned it as a synonymnbsp;of Cupressinoxylon. Beust� and Barber� among other authorsnbsp;adopt the same course. It has more recently been stated bynbsp;KrauseB that Goeppert�s genus Physematopitys has the charactersnbsp;of Protopiceoxylon. Goeppert� afterwards described a secondnbsp;species, Physematopitys succinea, founded on a tangential sectionnbsp;of a piece of Oligocene wood from the Baltic amber, but thenbsp;data are clearly insufficient to justify its identification as Ginkgo:nbsp;Conwentz� includes the specimen in Pinus succinifera.

Schroeter^� described some wood from beds on the Mackenzie river in North Canada, referred to the Miocene period, as Ginkgo sp.

' Glothan (05) p. 103. nbsp;nbsp;nbsp;- Goeppert (60) p. 242, PI. xlix. figs. 1�5.

^ Kraus in Schimper (72) A. p. 370; Kr0.us(83); Schenk in Zittel (90) A. p. 871.

Beust (85). nbsp;nbsp;nbsp;0 Barber (98).nbsp;nbsp;nbsp;nbsp;� Kr�usel (13).

� Conwentz (90) A. p. 20. nbsp;nbsp;nbsp;Schroeter (80) p. 32, PI. ill. fig. 27�29.

on the ground of the large size of the medullary cells: no pits are described either on the tracheids or on the medullary-ray cellsnbsp;and the unusual size of the ray-cells may well be a pathologicalnbsp;or post-mortem phenomenon. The species PJiysematopitys ex-cellens described by Felix^ from beds, probably Eocene in age, innbsp;the Caucasus agrees with Cupressinoxylon in the presence of rowsnbsp;of resin-parenchyma in the wood, and the depth of the rays greatlynbsp;exceeds that in Ginkgo hiloba. Penhallow^ described some calcifiednbsp;wood from Upper Cretaceous beds in the Queen Charlotte Islandsnbsp;as Ginkgo pusilla, but the reasons for assigning it to that genus arenbsp;not convincing. A fuller description of another specimen regardednbsp;as the wood of a Ginkgo has been published by Dr Platen� undernbsp;the name Physematopitys Goepperti from material collected innbsp;Miocene beds in Milam County, Texas. The relatively large sizenbsp;of the medullary-ray cells is mentioned as the chief character onnbsp;which the determination was based.

It may be said that such fossil specimens as have been referred to Physematopitys or Ginkgo have very little value as records ofnbsp;the occurrence of the genus Ginkgo: in view of the abundance ofnbsp;leaves in Mesozoic and Tertiary strata that are hardly distinguishable from those of the surviving type it is remarkable�if thenbsp;anatomical characters of the genus afford in themselves a trustworthy basis of identity--that more satisfactory specimens havenbsp;not been found.

It has been customary to use the generic name Ginkgo both for the recent species and for fossil leaves from Mesozoic andnbsp;Tertiary strata, and in a few cases for Palaeozoic leaves. Innbsp;certain instances, for example such leaves as those from thenbsp;Island of Mull and other Tertiary localities referred to Ginkgonbsp;adiantoides (fig. 644) there can be no doubt as to generic identitynbsp;with the recent species and indeed, so far as concerns form andnbsp;venation, the Eocene leaves might well belong to Ginkgo biloba.nbsp;On the other hand even in the case of Ginkgo adiantoides we lack

the confirmatory evidence of flowers and seeds. From Wealden and Jurassic rocks numerous leaves have heen described that innbsp;some cases are practically identical with those of the living species,nbsp;but for the most part they are characterised by certain featuresnbsp;denoting at least a specific difference. For these and for othernbsp;Ginkgo-like leaves it would seem desirable to follow the usualnbsp;custom and adopt a designation that does not necessarily implynbsp;even generic identity. A few examples of seeds and male flowersnbsp;are known from Jurassic strata bearing a close resemblance tonbsp;those of Ginkgo hiloha, but such specimens are not common andnbsp;some of the few that have been found, though probably belongingnbsp;to the Ginkgoales, may not be correctly included in Ginkgo.

I therefore propose to employ the name Ginkgoites for leaves that it is believed belong either to plants genericallj'^ identical withnbsp;Ginkgo or to very closely allied types.

It is impossible in some cases to draw a sharp line between the genera Ginkgo and Baiera: typical examples of the latternbsp;genus are easily recognised by their narrow, relatively longer,nbsp;and more numerous segments, but it is obvious that charactersnbsp;based on the degree of division of a lamina and on the breadthnbsp;of the segments are at best unsatisfactory, and the inclusion ofnbsp;certain specimens in one or other genus is purely arbitrary.

A difficulty is presented by several types of Palaeozoic leaves assigned by many authors to the Ginkgoales and referred tonbsp;^inkgophyllum, Psygniophyllum, and other genera which, whilenbsp;hearing a general resemblance to the leaves of Ginkgo, cannot henbsp;regarded as evidence of the occurrence of the class that is nownbsp;represented by Ginkgo biloba. It has been suggested that Psygmo-pkyllum, Ginkgopthyllum, Bhipidopsis, and certain other generanbsp;should be included in a distinct group, the Palaeophyllales^, anbsp;group of which the affinities are unknown. Though the adoption

a distinctive group-name has the advantage of indicating the a-bsence of any trustworthy evidence of relationship to the Ginkgo-ales, it is open to question whether anything substantial is gainednbsp;hy the use of a term suggestive of relationship between differentnbsp;leaves that in themselves afford no clue as to the position of the

The name Ginkgoites as used in this chapter is restricted to leaves that are regarded as records of the Ginkgoales, while thenbsp;genera referred by Dr Arber to the Palaeophyllales are brieflynbsp;described as fossils that may or may not be closely related to onenbsp;another but which cannot as yet be assigned to any place in anbsp;natural system of classification.

The leaves discovered by Grand�Eury in Permian Uralian beds and described by Saporta as Salisburia primigenia^ should probably be referred to the genus Psygmophyllum: like many othernbsp;supposed Palaeozoic species assigned to the Ginkgoales or tonbsp;Ginkgo they afford no satisfactory evidence of affinity to thenbsp;surviving genus. Other examples of leavhs from Palaeozoic rocksnbsp;described as species of Ginkgo or Salisburia on inadequate groundsnbsp;are described in the latter part of this chapter. The Rhaeticnbsp;leaves described by Brauns as Gyclopteris crenata and afterwardsnbsp;referred by Nathorst, with some doubt, to Ginkgo are describednbsp;in the account of Psygmophyllum^.

Fig. 632 A shows the form of the specimen from the Rhaetic beds of Scania on which Nathorst� founded the species Ginkgo

Fig. 632. A. Ginkgoites obovata. B. Ginkgoites antarctica. (Nat. size; A, after Nathorst; B, drawn from a specimen in the University Museum. Brisbane.)

^ Saporta and Marion (85), p. 14.5, fig. 74. � Nathorst (86) p. 93, PI. xx. fig. 5.

ohovata: the obovate lamina is 5-6 cm. long with a maximum breadth of 2'5 cm.; the upper edge is partially torn and thenbsp;forked veins are about 3 mm. apart. Some dark patches betweennbsp;the veins are probably, as Nathorst suggests, secretory sacsnbsp;similar to those in the leaves of Ginkgo. In shape but not innbsp;venation this type resembles Ginkgodium NathorstiYok.^ (fig. 659,nbsp;p. 62); except in the absence of a deep median sinus it is, however, nearer to a species from the Jurassic of Dzungaria describednbsp;originally as Ginkgo Obrutschewi^ (fig. 642, p. 26).

The leaf from the Rhaetic beds of Scania on which this species was founded was originally referred by Nathorst� to Ginkgo butnbsp;subsequently transferred by him to Baiem; it consists of a slendernbsp;stalk and a sub-triangular lamina deeply divided into 4�6 linearnbsp;truncate segments with 2�4 veins dichotomously branched nearnbsp;the base (fig. 645, B, p. 38). An examination of the originalnbsp;specimen leads me to prefer the designation Ginkgoites to Baiera.nbsp;This species like many others from Rhaetic rocks is hard!/ distinguishable from some Jurassic types.

Under the name Salisburia antarctica Saporta^ described a single leaf from Australia believed to be of Lower Lias age, butnbsp;no precise information is given with regard to the locality. Shirley ^nbsp;fias also figured a specimen as Ginkgo antarctica from rocks thatnbsp;are probably of Rhaetic age at Denmark Hill, Ipsmch (Queensland). The lamina of Saporta�s specimen is broadly obcuneatenbsp;and 3-5 cm. broad, characterised by the presence of two marginalnbsp;veins like those in Ginkgo hiloba from which forked branches arenbsp;given oil. This leaf is practicalR identical with some of thenbsp;smaller, entire, examples on the short shoots of the recent type.nbsp;The rather larger specimen figured by Shirley does not present sonbsp;striking a similarity to those of the existing species. The lack ofnbsp;definite information as to the provenance of the type-specimen

* Saporta and Marion (85) p. 142, fig. 71, A.; Ratte (88) PI. in. fig. 1; Renault (85) A. PI. II. fig. 19,

is unfortunate, but whether or not Shirley�s fossil is identical with Saporta�s specimen there would seem to be no reasonablenbsp;doubt that it should be included in the genus Ginhgoites.

Fig. 632 B is drawn from a photograph of a specimen in the Brisbane Museum which I recently had an opportunity of examining ; it is from the Ipswich beds and is undoubtedly specificallynbsp;identical with Saporta�s type.

This widely spread Jurassic species founded on leaves from the Yorkshire coast, was first figured by Phillips in 1829^ asnbsp;Sphenopteris lalifolia, but under the same name Brongniart^ hadnbsp;a year previously recorded a Carboniferous Fern. In 1830 Brongniart� figured and described another Yorkshire specimen as Cyclo-pteris digitata: the generic name Cyclopteris was adopted by Dunkernbsp;and other authors until Heer* drew attention to the very closenbsp;agreement between the Jurassic leaves and those of the Maidenhairnbsp;tree, a similarity that led him to adopt the generic designationnbsp;Ginkgo. Leaves hardly distinguishable from the Jurassic impressions had previously been recorded from Tertiary rocks as speciesnbsp;of Salisburia or Ginkgo.

It is impossible to define precisely the several species of Ginkgoites founded on leaves: in the account of the recent species attention is called to the range in leaf-form and its bearing onnbsp;the determination of fossils. All that can be done is to adoptnbsp;certain specific names as a matter of convenience, recognisingnbsp;that the difierences on which the classification is based are notnbsp;either sufficiently sharply defined or morphologically importantnbsp;to be regarded as criteria of true specific distinctions. Manynbsp;authors have employed the specific name Huttoni, first used bynbsp;Sternberg�, for leaves identical in size and outline with G. digitatanbsp;but characterised by a deeply-lobed lamina; this difference is,nbsp;however, not greater than or even as great as differences metnbsp;with vdthin the species Ginkgo biloba. To facilitate descriptionnbsp;the designation Huttoni is retained as a form-designation for thenbsp;more deeply lobed examples included in the species G. digitata (e.g.

1 nbsp;nbsp;nbsp;Phillips (29) A. PI. VII. flg. 18. See Fontaine in Ward (05) B. p. 121.

2 nbsp;nbsp;nbsp;Brongniart (28) A. p. 51.nbsp;nbsp;nbsp;nbsp;� Brongniart (28^) A. p. 219, PI. Lxi bis, figs. 2, 3.

fig. 633). The number of Ginkgoites leaves from Jurassic strata is considerable and the student who attempts to classify specimens innbsp;a large collection under specific heads soon finds himself confrontednbsp;in an acute form with the constantly recurring difficulty of fixingnbsp;boundaries. As Knowlton^ says, �In dealing with such an abundance of specimens and multiplicity of forms one must needs makenbsp;either many � species� to accommodate this diversity, or only onenbsp;or two, and in view of the known variation exhibited by the singlenbsp;living species, the latter plan seems preferable.� In advocating thisnbsp;Use of specific names in a liberal sense I admit the probability or indeed the certainty that forms specifically distinct wll be grouped

Pig. 633. Ginkgoites digitata var. Huttoni. A leaf from the Upper Jurassic of Helmsdale, Scotland. (Stockholm Museum; nat. size.)

under one designation. It is, however, clearly impossible in the case of impressions of leaves of Ginkgoites to impose limitationsnbsp;based on the form of the lamina, the degree of dissection, and similarnbsp;variable features that cannot be accepted as trustworthy criteria ofnbsp;tTue specific distinctions. As material accumulates data may benbsp;furnished that will enable us to recognise characters of morphological significance: in carbonised impressions from Avhich cuticularnbsp;preparations can be made the form of the epidermal cells andnbsp;the structure of the stomata may supply a valuable aid to morenbsp;accurate diagnosis. The spacing of the veins is a feature worthynbsp;of attention in the description of well-preserved specimens.

There is also a further difficulty in regard to terminology: the employment of the two generic names Ginkgo or Ginkgoitesnbsp;and Baiera reveals a striking lack csf uniforihity among authors,nbsp;and the artificial nature of the characters determining the use ofnbsp;one or other generic name necessarily lead to diversity in practice.nbsp;As with the definition of species within the genus Ginkgoites, sonbsp;also the adoption of Ginkgoites or Baiera is to a large extent the

lift

Fig. 634. Ginkgoites digitata. Leaf from Kap Boheman, Spitzbergen. (Stockholm Museum; nat. size.)

result of individual preference and merely expresses an attempt to classify in an arbitrary fashion the numerous types of leavesnbsp;that in themselves afford no sure guide as to precise affinity. Thenbsp;South African, Ehaetic, specimen shown in fig. 635, L was originallynbsp;described as Baiera moltenensis^ but it might equally well be referrednbsp;to Ginkgoites.

The leaves of Ginkgo digitata have a long slendei petiole (fig. 634); the lamina is semiorbicuiar or obcmieate, entire, or more or lessnbsp;deeply divided into equal lobes, or irregularly divided into several

segments; the number and size of the segments and the form of their distal ends, truncate or obtuse, vary within wide limitsnbsp;(figs. 635, 637, 639, etc.). Numerous dichotomously branchednbsp;s. IVnbsp;nbsp;nbsp;nbsp;2

veins spread from the base of the lamina, the veins in the middle of the leaf being generally about 0-8�1 mm. apart.

The stomata are practically confined to the lower surface of the lamina. The epidermal cells are polygonal and the wallsnbsp;slightly sinuous as in the recent species, and over the veins thenbsp;cells are longer and narrower (fig. 636, A). Many of the largernbsp;epidermal cells have a cuticular ridge in the middle of the outernbsp;wall, represented in the figure by a black line. The stomata agreenbsp;closely with those of Ginlcgo biloba; the two guard-cells are surrounded by a group of subsidiary cells characterised by their

papillose heavily cuticularised walls overarching the stoma (fig. 636, B). The features shown in fig. 636 are not brought out innbsp;drawings from cuticles of the same specimen reproduced by Drnbsp;Stopes^ in her account of fossil plants from Brora: this may benbsp;due in part to a difference in the level at which the stomata werenbsp;drawn. The stoma of Ginkgo biloba represented in fig. 636, Cnbsp;illustrates the considerable difference produced by viewing a stomanbsp;in slightly different planes^.

No seeds have been found attached to stems bearing leaves of G. digitata, but seeds closely resembling those of the recent speciesnbsp;1 Stopes (07) p. 380.nbsp;nbsp;nbsp;nbsp;^ Seward (11) p. 47, PI. v. fig. 62.

occasionally occur in association with the foliage of this and othc? Jurassic species. Male flowers^ similar in habit to those of Ginkgonbsp;biloha are also found in beds containing impressions of Ginkgoites.

The abundance and wide geographical range of Ginkgoites digitata precludes anything more than a brief reference to somenbsp;representative types selected in illustration of the range in formnbsp;and the widespread occurrence of the species in Jurassic floras.

The leaf represented in fig. 637 is an unusually complete example from the Middle Jurassic beds of Scarborough; the laminanbsp;is 3-8 cm. deep and 6 cm. broad,nbsp;the venation agrees with that ofnbsp;Ginkgo biloha. A very similar typenbsp;of leaf is figured by Heer fromnbsp;Upper Jurassic (or Wealden) stratanbsp;of Spitzbergen as G. integriuscula^,nbsp;but with the proviso that it maynbsp;be merely a variety of G. digitata,nbsp;a view that Nathorst� has wiselynbsp;adopted. The latter author innbsp;speaking of the occurrence of G.nbsp;digitata in Spitzbergen states thatnbsp;sometimes the surface of the schists [shales] is as completelynbsp;covered with the leaves of Ginkgo as the soil beneath a livingnbsp;Ginkgo tree may be in autumn^.� In some specimens fromnbsp;the Yorkshire coast the lamina is practically entire as in a leafnbsp;from Scarborough in the York Museum figured in 1900�. Annbsp;exceptionally large form is shown in fig. 638; the lamina, 8 cm.nbsp;broad, is divided into several short and comparatively broadnbsp;obtuse or truncate lobes�. Uig- shows a leaf from thenbsp;8gt;tonesfield Slate, now in the Cirencester Museum; the lamina isnbsp;deeply divided into two broad cuneate lobes as in some forms ofnbsp;the recent species. The Stonesfield Slate specimens were origin-^by named by Buckman Noeggerathia {'() and later StrieJdandia

Xathorst (97) p. l.j ; for a discussion of the age of the Spitzbergen beds, see Xathorst (13^.

ncuminata^. A deeply divided obcuneate leaf, only 2�4 cm. broad, is figured by Dr Slopes^ from Jurassic strata (Lower Oolite)nbsp;on the Sutherland coast at Brora (fig. 635, J). The specimennbsp;shown in fig. 633, also from the Sutherland coast, a few milesnbsp;north of Brora and of Kimei�idgian age, illustrates the type ofnbsp;leaf that may be conveniently referred to Ginhgoites digit�tanbsp;var. Huttoni^. Heer�s Jurassic species. Ginkgo JaccarcW^, fromnbsp;Switzerland is almost certainly Ginhgoites digitata.

The leaf reproduced in fig. 634 from a drawing made for me in the Stockholm Museum through the kindness of Prof. Nathorstnbsp;Was originally figured by Heer� from Cape Boheman, Spitzbergen,nbsp;from beds assigned by Nathorst to the Upper Jurassic series,nbsp;laathorst includes in this series strata agreeing iir their fossil plantsnbsp;with the Wealden of England and North Germany; the �Ginkgo,nbsp;beds� may be Portlandian or even younger. The veins in thisnbsp;specimen are approximately 1 mm. apart.

The leaf represented in fig. 640 is one of several specimens collected bvDr Koettlitz in Franz Josefnbsp;Land from beds that are probably Ju-�assic. The lamina is 2-5 cm. in depthnbsp;ftnd the veins are about 0-8 mm. apart;nbsp;the lamina between the veins shows anbsp;fine transverse striation, not shown innbsp;the drawing, a feature occasional^ seennbsp;111 impressions of Ginkgoites and due innbsp;^11 probability to the shrinkage of thenbsp;mesophyll tissue into transverse bands�.

This form of leaf has been described as Gingico polaris Nath.�; it is smaller than most examples of Ginkgoites

� Newton and Teall (97) PI. xxxviii. figs. 1, 2 ; (98) PI. xxix. fig. 3; Nathorst lt;9fi) PI. 1. fig. 8 and Thomas (11) PI. rv. fig. 8.

digitata and the veins are more crowded. For this type I suggest the designation G. digitata var. polaris. Solms-Laubach^ figuresnbsp;an incomplete leaf from Franz Josef J.jand with a broader laminanbsp;divided into several broadly rounded segments as in many Britishnbsp;specimens but with rather closer venation. A similar leaf fromnbsp;the same locality is figured by Newton and TealP. Fig. 635, Knbsp;shows a lobed leaf similar to that represented in fig. 640, describednbsp;by Nathorst� from Franz Josef Land and compared by him withnbsp;G. sibirica Heer and G. flabellata Heer from the Jurassic of Bastnbsp;Siberia. A very small specimen similar in form to the largernbsp;example shown in fig. 635, K was figured by Nathorst^, also fromnbsp;Franz Josef Land, as Ginkgo polaris var. pygmaea. From thenbsp;west coast of Greenland Hartz described a leaf very similar tonbsp;some of the Yorkshire examples as Ginkgo (Baiera) Hermelini^:nbsp;Hartz regarded the beds as Liassic or Rhaetic, and the occurrencenbsp;of shells pointing to a Kellaways horizon immediately above thenbsp;plant-beds suggests that the latter may belong to the Middlenbsp;Jurassic series�. Ginkgo digitata is represented also in Jurassicnbsp;strata in the New Siberian Islands by a leaf figured by Nathorstnbsp;as Ginkgo sp. which agrees with the type G. digitata var. Huttoni��,nbsp;and Krystofovic* has described G. digitata from Jurassic beds innbsp;Ussuriland at the northern end of the Muravjev-Amurskyj peninsula. Some good specimens are recorded from Bornholm�, ofnbsp;Middle or Lower Jurassic age, as G. Huttoni which are identicalnbsp;in form and size with British specimens.

Several examples of leaves of the G. digitata type and some of the form Huttoni have been figured from Middle Jurassic rocksnbsp;in Oregon^� (fig. 635, H). Some particularly large examples arenbsp;named by Fontaine G. Huttoni var. magnifolia but these are not

� Bartholin (94) p. 96, PI. xii. figs. 1�3; (10) PI. in. figs. 9, 10 4� Fontaine in Ward (05) B. Pis. xxx.�xxxii., XLVi.

specifically distinct from G. digitata. G. digitata is represented in Upper Jurassic or Wealdeu beds in Alaska (Cape Lisburne)^.

Other Localities.

From Soutbern dlussia Thomas^ has recently described good specimens of Ginkgo digitata of Middle Jurassic age some of whichnbsp;agree closely with the large leaves figured by Fontaine from Oregonnbsp;as G. Huttoni var. magnifoliathe specimens previously figured bynbsp;Eichwald� as Cyclopteris incisa from the same district are examplesnbsp;of 6r, digitata. The species is recorded also from Jurassic rocksnbsp;m Turkestan^ (fig. 635, F), Chinese Dzungaria^ on the west bordernbsp;of Mongolia, from the region to the east of Lake Baikal�, and fromnbsp;Afghanistan�. The incomplete specimen figured by FeistmanteFnbsp;from the Jabalpur group of India as Ginkgo lobata, part of whichnbsp;is shown in fig. 643, A, agrees in the form of the lamina and innbsp;venation with G. digitata: the veins in the middle of the laminanbsp;are from -8 to 1 mm. apart. Feistmantel compares his speciesnbsp;with G. digitata, and an examination of one of his figured specimensnbsp;leads me to assign it to that type; it is indistinguishable from thenbsp;Afghan specimen already quoted. The piece of a leaf figured bynbsp;Feistmantel as Ginkgo sp. belongs to a similar leaf, but the venationnbsp;is finer and it may be identical with Ginkgo crassipes Feist.�

Leaves of the G. digitata type are recorded from Jurassic beds in Victoria^� (fig. 635, D) and from Japan^^ (fig. 635, C). Tuzsonnbsp;has figured a bilobed petiolate leaf from J urassic rooks in Hungarynbsp;as Ginkgo parvifolwd^; it is similar in form to most of the specimensnbsp;referred to G. digitata, but has relatively broader segments: it isnbsp;interesting as being the first recorded example of Ginkgoites fromnbsp;Hungary.

quot; Thomas, H. H. (11) p. 73, PI. iv. fig. 7; � Eichwald (65) PI. iv. fig. 6.

The specific name sibirica was given by -Heer^ to one of the most abundant forms in the rich plant-beds at Ust Balei nearnbsp;Irkutsk in Siberia; the specimens figured from Siberia as Ginkgonbsp;sibirica (fig. 635^ E), G. Schmidtiana, and G. lepida cannot benbsp;regarded as well-defined species; they agree in the deep divisionnbsp;of the lamina into several linear segments with obtuse or in somenbsp;cases more pointed apices. Heer draws attention to the resemblance between G. sibirica and G. pluripartita from Wealdennbsp;rocks but, as he says, the venation is rather coarser in the Siberiannbsp;leaves and the segments are generally narrower in G. sibirica.nbsp;Fontaine in his description of Jurassic leaves from Oregon apparently identical with Heer�s G. sibirica states that G. Schmidtiananbsp;is a smaller form of the same species�, but Ward� points out thatnbsp;as G. Schmidtiana is described on one page and G. sibirica is definednbsp;on the following page the former designation must be preserved.nbsp;It may be urged that as the name sibirica is the more widely usednbsp;and familiar term, considerations of convenience should overridenbsp;this meticulously strict interpretation of the rule of priority. Anbsp;revision of Heer�s Siberian material would, I have no doubt, resultnbsp;in the reduction of his specific terms; on comparing severalnbsp;specimens in the Museums of Copenhagen and Stockholm with thenbsp;illustrations in the Flora Fossilis Arctica I found that several ofnbsp;the published figures are far from accurate. For the present thenbsp;most convenient course would seem to be the retention of Gink-goites sibirica for leaves similar to some of the more deeply dividednbsp;forms of G. digitata and to G. pluripartita, but normally characterised by a lamina divided almost or quite to the base into oblong�nbsp;obtuse or more or less acute segments. Leaves of the G. sibiricanbsp;type, using the term in the wider sense and including Heer�snbsp;other species Ginkgo Schmidtiana and G. lepida, are fairly Commonnbsp;in Jurassic rocks and occur also in Cretaceous floras; they arenbsp;recorded from Kimeridgian beds in Scotland^ (fig. 641, A),nbsp;Jurassic strata in Siberia, China� (described by Yokoyama as

^ Heer (77) ii. p. 61. PI. vn. fig. 6; PI. IX. fig. 5 6; PI. XI.; Heer (82) ii A. p. 16 Pis. IV., V.

G.flahellata, also the similar leaves figured by Krasser as G. Schmidt-iana var. panifolia), Turkestan^, and Western America^; also from Upper Jurassic beds in Japan� and Franz Josef Land^ and from

Lower Cretaceous beds in Canada�. A similar form of Ginkgoites is recorded from Jurassic rocks in Victoria�.

It should be added that some Jurassic specimens described as species of Baiera, e.g. Baiera Ph�lipsi�� Nath, are very near tonbsp;U. sihirica and in such cases the choice of Baiera or Ginkgoitesnbsp;is not determined by any satisfactory standard.

\Ginkgo digitata (Brongn.), Schmalhausen (79) A. p. 33, PI. v. fig. 4 h. Gingho sihirica Heer?, Ibid. p. 34, PI. iv, fig. 2 6.]

The incomplete leaf-fragments from the Altai mountains referred by Schmalhausen to these species are too incomplete to be determined nuth any degree of certainty. The precise age of thenbsp;iisds is uncertain but, as Zeiller� has shown, they are probablynbsp;Permian. There is a certain resemblance between the specimennbsp;referred to Ginkgo digitata and some leaves from Permo-Carboniferous strata in Kashmir described as Psygniophylluni HollandP.

� Zeiller (96) A. nbsp;nbsp;nbsp;� Seward (07) p. 59, PI. xiii. figs. 3�6.

This name was proposed by Nathorst^- for a leaf from the Jurassic rocks of the Yorkshire coast in the British Museum similarnbsp;to some of the smaller forms referred to G. digitata hut characterisednbsp;by the deltoid form of the lamina, its deep dissection into sixnbsp;more or less pointed segments, three on each side of a broad mediannbsp;V-shaped sinus, and by the small size (1-5 cm. broad and 2-5 cm.nbsp;deep) of the lamina. It is hardly possible to decide whether thisnbsp;and similar small leaves should be regarded as varieties, e.g. G. digitata var. polaris, or assigned to a distinct species. Fontaine^ compares some leaves figured by him from Oregon as Ginkgo sp. withnbsp;Nathorst�s species, but they are probably nearer to the examplesnbsp;described by Nathorst and others as G. polaris. A small bilobednbsp;leaf figured by Raciborski� from Rhaetic beds near Cracow asnbsp;Ginkgo afi. whitbiensis is more likely to be a young leaf of thenbsp;Fern Hausrnannia.

This species, named after Prof. Obrutschew who made a collection of plants- from Jurassic rocks in Chinese Dzungaria^ on the western border of Mongolia (lat. 85� N. long. 45� E.), agreesnbsp;fairly closely with some forms of G. digitata. The lamina is deeply

' Nathorst (80) A. p. 74; Se-ward (00) B. p. 261, PI. ix. fig. 8 2 Fontaine in Ward (05) B. PI. xxxiv.

� Raciborski (91) PI. iv. fig. 13. nbsp;nbsp;nbsp;^ Seward (11) p. 46, Pis. iii.�vii.

bilobed and each half may be similarly divided: the segments ^re obtuse and not truncate (fig. 642, A); the veins are approximately 1 mm. apart. The walls of the epidermal cells are straightnbsp;and not sinuous; the stomata (fig. 642, B) are practically confinednbsp;to the lower surface and are less numerous than in Ginlcgo hilobci.nbsp;The guard-cells are surrounded by a group of broadly triangularnbsp;cells with papillose, thickly cuticularised, inner walls, but thenbsp;overarching papillae are rather less prominent than in G. digitata.

It is interesting to find short secretory tracts at intervals in the intercostal regions of this species agreeing vnth those in the leavesnbsp;of the recent species. A leaf described from Jurassic rocks innbsp;Amurland as Ginkgo sp., cf. G. OhrutschewP- agrees closely withnbsp;the type-specimens.

This Upper Gondwana species (fig. 643, B) was described from the Madras coast as Ginkgo crassipes'^. An examination of thenbsp;type-specimens enables me to confirm generally the accuracy ofnbsp;Teistmantel�s figures except in one point, namely the supposednbsp;presence of a median ridge extending from the petiole through thenbsp;lower third of the lamina, which gives the impression of a midrib;nbsp;this is merely a shallow groove that is clearly accidental. Thenbsp;leaves appear to be entire ; the lamina is obcuneate and passesnbsp;Into a fairly stout petiole; the veins are occasionally forked andnbsp;approximately 0-5 mm. apart. The piece of lamina described bynbsp;Teistmantel as Ginkgo sp.� may be specifically identical with thisnbsp;species.

This Wealden species was first described by Dunker^ as Gyclopteris digitata Brongn. and Ettingshausen� also regarded thenbsp;Herman specimens as identical with the English Jurassic type.nbsp;Schimper� proposed the name Baiera pluripartita because of thenbsp;deeply divided lamina and the comparatively narrow segments:

while substituting Ginhgoites for Baiera, it is advisable to retain Scliimper�s specific designation on the gronnd that the deep divisionnbsp;of the lamina appears to be the rule in the Wealden leaves whereasnbsp;in G. digitata the leaves are usually much less deeply dissected and

'A f' I/' I

Ginkgoites lohata (A) and Ginkgoites crassipes (B). (Indian Geological Survey, Calcutta.)

have broader segments. The largest specimen is one figured by Ettingshausen with a lamina 7-5 cm. broad and 4-5 cm. deep;nbsp;there is generally a deep median sinus and each half of the laminanbsp;IS subdivided into relatively narrow obtuse or truncate obcuneatenbsp;segments (fig. 635, B). Schenk^ describes the epidermal cells as

polygonal with straight walls; the stomata are surrounded by 6 accessory cells as in Ginkgo hiloba and they are confined tonbsp;the lower surface. Although there is practically no difference,nbsp;as regards form and venation, between this Wealden species andnbsp;some of the Jurassic leaves referred to G. digitata var. Huttoni thenbsp;distinctive specific name is retained for the reason already mentioned.

Leaves of a very similar form are figured by Heer as Ginkgo ^*iuUinervis^ (fig, 635, G), from Upper Cretaceous (Cenomanian)nbsp;beds of West Greenland, and Baiem arctica from the Rome beds

The specimen from the Atane (Cenomanian) beds of Greenland figured by Heer as Ginkgo primordialis^ appears to be an entire leafnbsp;With a long petiole 2-5 mm. broad; the original impression in thenbsp;Stockholm Museum, too incomplete to serve as the type of a species,nbsp;�bows a very imperfect lamina and a long axis that has probablynbsp;110 connexion mth the leaf.

The Tertiary leaves on which this species was founded were in ifie first instance described as Ginkgo biloba* and, as several writersnbsp;8'Ve pointed out, so far as regards form and venation there is nonbsp;good reason for drawing a distinction between the fossils and thenbsp;6aves of the recent species. In a note published in 1913 Depape^nbsp;^onmtely adopts the name Ginkgo hiloba for Tertiary leaves whichnbsp;� regards as specifically identical with Unger's species. In thenbsp;gt; ��iice of any satisfactory evidence as to the nature of the repro-uctive organs and in view of the considerable interval thatnbsp;separates the Tertiary and recent plants, it is clearly inadvisablenbsp;assume specific identity. In adopting the generic designationnbsp;inkgoites instead of Ginkgo I am following the custom generallynbsp;recognised of distinguishing fossil forms by a special termination.nbsp;Ugh there is no implication that all species so named arenbsp;generically distinct from the surviving type. In proposing the

lt;iPo nbsp;nbsp;nbsp;maps of these plant-beds on the coasts of the Noursoak peninsula,

name Salisburia adiantoides for the Miocene leaves from Sengallia in North Italy Unger^ indicated their probable identity withnbsp;Ginkgo biloba L., the generally adopted name for the existingnbsp;species which Smith in 1797 proposed should be called Salisburianbsp;adiantifolia^. The Miocene leaves from Sengallia figured bynbsp;Massalongo and Scarabelli� have an entire, irregularly crenulatenbsp;or a more or less deeply bilobed lamina very like that of Ginkgonbsp;biloba: the specimen named by them S. Procaccini'^ should alsonbsp;be included in G. adiantoides. The resemblance to the recentnbsp;leaves extends to the presence of short secretory tracts betweennbsp;the veins, but these were referred by Massalongo to a fungusnbsp;which he named Sclerotites Salisburiae^.

The two specimens reproduced in fig. 644 from the Eocene beds in the Isle of Mull illustrate the broad fan-like lamina that variesnbsp;from 5 to 10 cm. in breadth and may be entire, unevenly lobed ornbsp;symmetrically bilobed. The venation is identical with that of thenbsp;living species: the characteristic marginal veins on the lower edgenbsp;of the lamina are clearly seen in fig. 644, A. The preservationnbsp;of these British leaves described by Mr Starkie Gardner� isnbsp;exceptionally good; they occur as purple impressions in whitenbsp;clay interbedded with basaltic sheets in the cliffs of Ardtun Headnbsp;in the Island of Mull. Though perhaps on the average thesenbsp;Eocene leaves from Mull are larger than those of the Maidenhairnbsp;tree some examples of the latter exceed in size any of the fossils.

Leaves identical with or very similar to the Italian and Scottish specimens are recorded from both Tertiary and Upper Cretaceousnbsp;rocks in many parts of the world. Specimens collected by Dr Lyallnbsp;from Tertiary (Miocene or Eocene) beds on Disco Island off thenbsp;West coast of Greenland, lat. N. 70�, were described by Heer^ asnbsp;Salisburia borealis and he also speaks of them as Salisburia adiantoides var. borealis^. An examination of the original specimensnbsp;in the Kew Museum and an impression in the Dublin Museum

� Massalongo and Scarabelli (58) p. 163, PI. i. fig. 1; PI. vi. fig. 18; PI. VII. fig. 2; PI. XXXIX. fig. 12.

collected by Sir Leopold McClintock enables me to confirm Gardner�s view as to the identity of S. botealis and S. adiantoides. Examples of G. adiantoides were also obtained from Atanekerdluknbsp;On the Noursoak peninsula to the north of Disco Island. Heer

figures a very well preserved leaf with a long petiole from Atanekerdluk which he refers to S. adiantoides. In a later account of this flora he refers the species to Ginkgo. Heer regarded these

Greenland beds as Miocene, but it has been suggested that they may belong to the Eocene period^.

Heer^ also described specimens of Gihkgoites adiantoides from beds, assigned to the Miocene period, in Sachalin Island. Thisnbsp;species has been obtained from the Laramie series, a thick successionnbsp;of brackish water strata deposited on both sides of the Rockynbsp;Mountains � extending from Mexico far into British North Americannbsp;Territory� and including both Upper Cretaceous and Lowernbsp;Tertiary strata�. Leaves described by Lester Ward as Ginkgonbsp;laramiensis^ and regarded by him as intermediate between G. adiantoides and G. biloba are indistinguishable by any definite characternbsp;from G. adiantoides. Ward�s species is also recorded by Knowlton�nbsp;from the Montana formation, a series of beds formerly included innbsp;the Laramie group, in Wyoming. Ginkgoites adiantoides occursnbsp;in Upper Cretaceous beds in British Columbia and specimensnbsp;described as Salisburia pusilla^ by Dawson, which I believe tonbsp;belong to this species, were found in the Upper Cretaceous ofnbsp;Vancouver Island. From beds of the same age Penhallow^ hasnbsp;described some wood as Ginkgo pusiUa though it is not clear onnbsp;what grounds it is assigned to the genus Ginkgo.

It is interesting to find leaves of this type recorded from Lower Pliocene beds at Saint Marcel-d�Ard�che in France� and fromnbsp;Upper Pliocene beds in the Lower Main valley in Germany�.

It is therefore abundantly clear that trees, apparently indistinguishable as regards the form of the leaves from Ginkgo biloba, flourished as recently as the Pliocene period in western Europenbsp;and in the Eocene period grew as far north as latitude 70�.

Records of seeds referred to Ginkgo are very meagre and add nothing of importance to our knowledge of Tertiary species. Somenbsp;pyritised seeds were described by Ettingshausen and Gardner^�

� Dawson (93) p. 56, PI. vi. figs. 11�14. Heer�s species Ginkgo pusilla is founded on Jurassic leaves; Heer (77) ii. p. 61.

from the London clay of Sheppey as Salisburia eocenica and afterwards figured by Gardner as Ginhgol eocxmca'^. Thenbsp;specimens (11 x 9 mm.) are smaller than the seeds of the recentnbsp;species but in shape and in the keeled shell there is a fairly closenbsp;resemblance. The hard sclerotesta forms the surface of the fossils.

Similar seeds have been assigned to Ginkgoites adiantoides from ffre Upper Pliocene of the Prankfiirt district^, but neither thenbsp;flcrman nor English specimens possess any interest as records ofnbsp;Ginkgoites seeds.

An imperfect leaf-like impression described by Ettingshausen� from Eocene beds in New Zealand was made by him the type ofnbsp;a new genus Ginkgodadus and interpreted as a stalked phyllocladenbsp;similar to those of the recent Conifer Phyllodadus, but because ofnbsp;the presence of a slender stalk and the resemblance of the lateralnbsp;�^eins to the venation of Ginkgo Ettingshausen suggested an affinitynbsp;to that genus. The existence of a midrib is, however, an importantnbsp;difference. Neither the New Zealand species nor similar fragmentsnbsp;from Eocene strata in New South Wales {Ginkgodadus austm-Gensis*) are of value as botanical records.

This generic name was first used by Braun� for some Triassic and Jurassic leaves agreeing in shape with those of Ginkgoites butnbsp;distinguished by the greater number and less breadth of the linearnbsp;segments. Braun�s definition states that the primary veins arenbsp;dichotomously branched while between them secondary veinsnbsp;form irregular hexagonal meshes. Schenk� examined Braun snbsp;Rhaetic specimens and failed to discover any indication of thenbsp;presence of secondary veins. In 1877 Heer� emended the originalnbsp;definition of the genus; he refers to the presence of finer veinsnbsp;between the main vascular strands and this feature is shown in

� Ettingshausen (87) p. 39, PI. vii- fig- nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;C P W. (43) p. 20.

� Ihid. (88) p. 103, PI. vni. fig. 32. nbsp;nbsp;nbsp;Braun�s species di Baiera in

some of his figures: these interstitial � veins � probably mark the position of hypodermal strands of stronger cells, a feature that isnbsp;not represented in recent or fossil Ginkgo leaves and is by no meansnbsp;generally characteristic of Baiera. Both Braun and Heer describenbsp;male and female reproductive organs: Braun interpreted somenbsp;small specimens as sporocarps but these were recognised bynbsp;Schenk as young foliage leaves. The male organs are describednbsp;by Heer as �amenta staminifera pedunculata, nuda, filamentanbsp;filiformia, antherae loculis 5�12, verticillatis. Semen drupae-forme, basi cupula carnosa cinctum� and compared with Schenk�snbsp;Stachyopitys Preslii which that author afterwards regarded asnbsp;microstrobili of Baiera Muensleriana. Reference is made tonbsp;these and similar fossils in the account of examples of malenbsp;flowers^. Seeds have been referred to Baiera on evidencenbsp;iurnished by their occasional association with leaves and by theirnbsp;resemblance to those of Ginkgo. It has been suggested thatnbsp;specimens described from the Potomac group of Virginia andnbsp;Maryland as Carpolithus ternatus^ and other species may be seedbearing organs of Baiera, but there is no satisfactory evidence innbsp;support of this view. In all probability some of the associatednbsp;seeds belong to Baiera, also some of the microstrobili, e.g. Leuthardt�snbsp;Swiss specimens described on a later page, but in the presentnbsp;state of our knowledge it is preferable to regard these specimensnbsp;as reproductive organs that cannot be assigned with certainty tonbsp;any particular species of Baiera or Ginkgoites.

Leaves assigned to Baiera vary within wide limits as regards size, the number of linear segments and their angle of divergence.nbsp;In many cases the leaves are petiolate though in several instancesnbsp;the petiole is represented by a narrow basal region of the laminanbsp;as in Psygmophyllum. It is stated by some authors that the veinsnbsp;are undivided, but though dichotomy is less frequent in Baieranbsp;and may be absent in narrow parallel-sided segments it is by nonbsp;means rare. The difference in venation between such leaves asnbsp;Ginkgoites digitata and typical species of Baiera, e.g. Baiera gracilis,nbsp;is mainly the result of the different form and degree of dissectionnbsp;of the lamina. The choice between Baiera or Ginkgoites as the

�lore suitable name fox ceitain forms of leaf is not governed by 9'ny definite criterion; specimens described as Ginkgo sibirica,nbsp;lt;5*. lepida, G. concinna^, etc. are indistinguishable from leavesnbsp;referred by authors to Baiera. Similarly such a species asnbsp;Lindleyana differs very slightly from some forms usually innbsp;eluded in the genus Czekanoivskia. Although leaves of Ginkgmtesnbsp;8'Ud Baiera are abundantly represented in plant-bearing beds uenbsp;know very little of the habit of the foliage-shoots; in a few casesnbsp;Ikere is evidence of the occurrence of several leaves on a singlenbsp;short shoot (fig. 646); in Baiera paucipartila, for example^ thenbsp;habit is the same as that of Czekanowskia and Phoenicopsis, butnbsp;in view of the frequent preservation of Czekanoivskia leaves stdlnbsp;attached to an axis it is surprising that the leaves of Baiera almostnbsp;always occur as detached specimens. The explanation may benbsp;that in Ginkgoites and Baiera the foliage-leaves were borne onnbsp;long and dwarf-shoots as in the recent species Ginkgo hiloha,nbsp;whereas in Czekanowskia the leaves were confined to shoots ofnbsp;limited growth as in Pinus. Some specimens described by Salfeld�nbsp;Irom the Solenhofen beds of Bavaria as Baiera^. longifolia Heernbsp;interesting in this connexion; they consist of fairly stoutnbsp;branches bearing alternate leaf-like organs having the habit ofnbsp;Baiera longifalm but subtended by a short and thick recurvednbsp;spinous process. There is no means of deciding from the available material whether the resemblance of the leaves to those ofnbsp;Baiera is an expression of relationship or merely a case of parallelnbsp;development, nor have we any means of determining the morphology of the leaves and the subtending spines. The Solenhofennbsp;plant agrees in habit with Sewardia latifolia* from the Wealdennbsp;of England and is included under that genus. While relationshipnbsp;between Sewardia and Baiera is by no means excluded, it is clearnbsp;that the species of the former genus differ considerably from

of Baiera exceed in size the leaves of any example of Ginkgoiies, notably B. Simmondsi, B. spectahilis, and others.

Potomac group in Virginia as Baiera foliosa^, consisting of an axis bearing crowded leaves with a deeply and rather irregularly dividednbsp;lamina and a comparatively broad and flat basal region, are notnbsp;typical examples of the genus but agree more closely in habitnbsp;with the older genus Dicramfhyllmn: their precise positionnbsp;cannot be definitely determined.

The structure of the cuticles is known in a few species from accounts published by Schenk^, Nathorst� and Thomas^; thenbsp;epidermal cells are sometimes characterised by fairly prominentnbsp;papillae on the outer walls (fig. 647) and the stomata, morenbsp;abundant on the lower surface but present also on the uppernbsp;surface, closely resemble those of Ginkgo and Ginkgoites', thenbsp;guard-cells are slightly depressed and are surrounded by 5�6 subsidiary cells with strongly cuticularised and projecting wallsnbsp;(fig. 647). Nathorst� has drawn attention to the presence innbsp;B. sfectahilis of traces of some secreted substance in the mesophyiinbsp;recalling the secretory tracts in the leaves of Ginkgo.

Braun and other authors have included Baiera in the Filicales, and attention has been called to the danger of confusing truenbsp;Fern-fronds with leaves of Baiera. Berry�s discovery of sporangianbsp;on the linear segments of Baiera-Y^ke leaves from the Potomacnbsp;beds, originally referred by Fontaine to his genus Baieropsis andnbsp;regarded by him as Ginkgoaceous, illustrates the possibilities ofnbsp;error in determinations founded on leaves alone. The fertilenbsp;examples of Baieropsis have been made the type of a new genusnbsp;Schizaeopsis^, other species of Fontaine�s genus being transferrednbsp;to the genus Acrostichopteris; they differ from Baiera in theirnbsp;attachment to slender axes and are no doubt portions of compoundnbsp;Fern-fronds. It is impossible to define with confidence the precisenbsp;geological range of the genus; leaves from Permian and Uppernbsp;Carboniferous strata agreeing with Baiera in the deep dissectionnbsp;of the lamina have been assigned to the genus Ginkgophyllumnbsp;(Psygmophyllum) and compared with Saporta�s Ginkgophyllumnbsp;Grasserti {fig. 669, p. 87). In imperfect specimens it is not alwaysnbsp;possible to draw a sharp line between Baiera and species of Psygmo-

Phyllum. There are, however, certain Permian leaves that are legitimately included in Baiera; the genus appears to have beennbsp;quot;ividespread in Triassic floras, though more especially in those ofnbsp;tile Ehaetic and Jurassic age. Baiera shares with Ginkgoites annbsp;important position in the Jurassic vegetation of both hemispheres,nbsp;but in the Cretaceous period Baiera appears to have been a comparatively rare genus and in the Tertiary floras it was entirelynbsp;replaced by members of the Ginkgoales with leaves of the typenbsp;that still survives. Baiera is clearly an older form than Ginkgoites;nbsp;it is not recorded from India and it has not been found in thenbsp;Permo-Carboniferous rocks of Gondwana Land.

Baiera though unknown in a petrified condition may con fidently be included in the Ginkgoales; the habit of the leaves,nbsp;the structure of the epidermal cells and such evidence as there isnbsp;with regard to the fertile shoots favour this conclusion. It must,nbsp;however, be added that the position of the Palaeozoic examples isnbsp;less firmly established.

This species, from Permian beds in Virginiai, jg based on imperfect portions of laminae deeply divided into bifurcate segments with truncate apices and several parallel veins; it agrees in thenbsp;form of the lamina and in the linear divisions with the type-specimen of Brongniart�s Fucoides digitalus^ from Permian bedsnbsp;of Mansfeld, a species which Geinitz^ also recorded, but under thenbsp;generic name Zonarites. Potoni�^ and other authors, followingnbsp;ffeer, transferred the species to Baiera. Heers combination,nbsp;B. digitata, had, however, already been used by Schimper� for thenbsp;Jurassic species usually called Ginkgoites digitata.

Leaves of the form represented by B. virginiana may be closely allied to Saporta�s Permian species Ginkgophyllum (Psygmophyllum)nbsp;lt;^rasserti^. In the absence of the basal part of the lamina a complete diagnosis or accurate identification is impossible. Somenbsp;authors have referred fragments of similar leaves to the genus

Schizopteris, e.g. S. Guembeli fromUpper Carboniferous and Permian stratabut other Permian specimens assigned to ScMzopteris,nbsp;e.g. S. dichotoma and S. trichomanoides^ are generically distinctnbsp;and probably belong to Pteridosperms. B. virginiana and thenbsp;leaves referred to Baiera digitata�possibly specifically with Fontaine�s type�represent Permian forms that agree closely with thenbsp;larger species B. Simmond.si, Heer�s Keuper species B. furcata,nbsp;and with the larger examples of B. Muensteriana (Presl) fromnbsp;Rhaetic beds.

This French, Permian (Autunien), species� (fig. 645, A) differs but little from B. virginiana-, the narrow cuneate leaf is divided

portion of the lamina is 5 mm. broad and the ultimate segments With obtuse apices have a breadth of 2-5 mm. The venation isnbsp;imperfectly shown in the specimens figured by Renault and Zeilleri.nbsp;I^wera furcata Heer.

A type similar to B. multijida Font.^ but characterised by the more uniformly narrow segments (2�2-5 mm.), is described bynbsp;Heer 2 and Leuthardt� from Keuper beds of Switzerland. Leuthardt�snbsp;figures show a single vein in the segments, a feature which maynbsp;� a peculiarity of the species. It was in association with thisnbsp;species that Leuthardt found the male flowers referred to onnbsp;another page (p. 53).

The leaves of this Swedish, Rhaetic, species^ may reach a length lb cm.; the lamina is narrow and cuneate, deeply divided intonbsp;1 nrcate linear segments with obtuse a,pices. It differs from

B. longifolia Heer and other similar species in the smaller number of segments. There is no petiole but the leaves are attached bynbsp;a narrow basal portion to a short axis (fig. 646). The occurrencenbsp;of several leaves on a very short scale-covered dwarf-shoot is annbsp;interesting feature which affords evidence of relationship withnbsp;CzekanowsMa and Phoenicopsis. Baiera paucipartita is describednbsp;by Yokoyama^ from Rhaetic beds in Japan and an imperfectnbsp;specimen from Rhaetic strata in New Zealand has been assignednbsp;by Arber^ to this species. In the case of imperfect leaves it isnbsp;impossible to distinguish specifically between many �species� ofnbsp;Baiera characterised by the division of the lamina into bilobednbsp;linear segments.

The leaves of this Rhaetic species from the south of Sweden� reach a length of 25 cm.; the coriaceous lamina is obcuneate andnbsp;fan-like but narrower than the leaves of such a species as B. Sim-mondsi. There is no well-defined petiole; the lamina is deeplynbsp;divided into two symmetrical halves each of which is further subdivided into bifurcate linear segments, and the strongly contractednbsp;ultimate segments are a characteristic feature (fig. 647). Thenbsp;basal region forms a stalk-like portion varying in length and, innbsp;the specimens so far obtained, not exceeding 3 cm. The veinsnbsp;are parallel to the sides of the segments and on the average 1 mm.nbsp;apart; they are occasionally dichotomously branched. Stomatanbsp;occur on both sides of the lamina but are more numerous on onenbsp;surface, presumably the lower; the guard-cells are slightly depressed and surrounded by 5�6 subsidiary cells with thicklynbsp;cuticularised papillose walls (fig. 647, B) as in Ginkgo. The epidermal cells on the lower surface are also papillose like thosenbsp;described by Thomas in B. longifolia'^. Several dark spherical andnbsp;spindle-shaped patches found by Nathorst between the cuticularised layers are believed to be the remains of some resinous or othernbsp;substance comparable with that formed in the secretory tracts innbsp;the leaves of Ginkgo and some species of Ginkgoites.

resembles B. pulchella, a Jurassic species described by Heer^ from East Siberia and by Bartholin^ from Bornholm, but in typicalnbsp;examples of B. pulchella the lamina is divided into two segmentsnbsp;only. Comparison may be made also with B. longifolia a Jurassicnbsp;species distinguished by its narrower segments. A specimen fromnbsp;Bornholm referred by M�ller� to B. pulchella is probably, as Nathorstnbsp;suggests, a piece of a B. spectabilis leaf.

The leaves described by Shirley^ as Ginkgo Simmondsi from Denmark Hill near Ipswich in Queensland, from rocks that arenbsp;probably of Rhaetic age, are of the same type as the leaf on whichnbsp;Ratte� founded his species Jeanpaulia (?) palmata, which he afterwards transferred to Salishuria, from the Wianammata beds (Trias)nbsp;near Sydney. The precise age of the rich flora from Ipswich isnbsp;difficult to determine; a recent examination of several specimensnbsp;in the Brisbane collections led me to regard the plants as Rhaetic,nbsp;but further light on this question will be afforded by Mr Walkomnbsp;who is engaged in an investigation of the material. The Austrahannbsp;leaves agree closely vdth Fontaine�s Triassic species, Baiera multi-fida^, from Virginia; the plant-beds of the Richmond coalfieldnbsp;are correlated with the Lunz plant-beds in Austria^, the flora ofnbsp;which has never been adequately illustrated. The specimens fromnbsp;Virginia on which Fontaine founded his species do not afford anynbsp;evidence of a true petiole and the basal portion of the cuneatenbsp;lamina is narrower than in the Australian leaves: it is, therefore,nbsp;not improbable that B. multifida is a distinct though very similarnbsp;species. Ratte�s name B. palmata cannot be retained as Heernbsp;had jDreviously employed the same name for a Jurassic Siberiannbsp;form�; I have adopted Shirley�s designation in the belief thatnbsp;there are no differences of specific value between the Sydney andnbsp;Ipswich specimens.

The leaf reproduced in fig. 648 is Ratte�s type-specimen in the Australian Museum, Sydney; the whole leaf is nearly 30 cm.

- Bartholin (94) PI. xi. fig. 5. nbsp;nbsp;nbsp;^ M�ller (03) PI. iv. fig. 19.

long and 23 cm. in breadth; there are nearly 60 ultimate linear segments with obtuse apices and, in the smaller subdivisions,

^ 5 veins. There is a well-defined petiole and in outline the 'Whole leaf is identical with typical examples of Ginkgo hiloba.

Fontaine speaks of the lamina of his species as reaching a length of 25 cm.; both in the method of division and in the form of thenbsp;segments, B. multifida agrees closely with the specimen shown innbsp;fig. 648. A similar form of leaf is figured by Solms-Laubach^nbsp;from Rhaetic beds in Chile as Baiera ? Steinmanni, but the laminanbsp;only is preserved. Schenk�s B. tueniata^ from the Rhaetic floranbsp;of Franconia is another similar type.

The specimens described from the Stormberg series (Rhaetic) of South Africa� are portions of leaves that must have reached anbsp;length of 12 cm. or more and a breadth of 10 cm. The lamina isnbsp;deeply divided into broad linear segments which are furthernbsp;subdivided into narrower distal segments. In the lower part ofnbsp;the lamina the venation is comparatively coarse, but as the resultnbsp;of repeated dichotomy the veins are much more numerous in thenbsp;upper portion. This species may be merely a larger form ofnbsp;Feistmantel�s B. Schenki^ from the same beds, in which the lobesnbsp;are narrower as in B. longifolia Heer. B. stormbergensis resemblesnbsp;Nathorst�s B. spectabilis from the Rhaetic of Scania, but thenbsp;segments of the South African leaves have a coarser venation.

This Rhaetic species, originally figured by Presl as Sphaero-coccites Muensterianus and subsequently described by Braun as Baiera dichotoma, was named by Schenk Jeanpatilia Muensteriana. Schenk� examined Braun�s specimens from Franconianbsp;and identified the supposed sporocarps as partially expandednbsp;segments of foliage-leaves. The leaves are petiolate and the fan-, like lamina is deeply dissected into bifurcate linear segments;nbsp;the veins are numerous and dichotomously branched. The epidermal cells are elongate over the veins and elsewhere polygonal;nbsp;their walls are straight or slightly sinuous. The stomata are ofnbsp;the usual type, met with in Ginkgoaceous plants.

Baiera Muensteriana cannot be distinguished by any definite character from leaves that are referred to B. gracilis: in thenbsp;Solms-Laubach (99) PI. xiv. fig. 1.

type-specimen of the latter species the segments are fewer than in B. Muensteriana, hut in some Jurassic forms {e.g. fig. 651)nbsp;this difference no longer holds good. This is only one amongnbsp;several instances where Rhaetic and Jurassic �species� cannot benbsp;separated by any constant differentiating feature. B. Muenste-fiana is recorded also from Persia^ and several European localities,nbsp;hut it is impossible to determine its geographical range apart fromnbsp;that of Baiera gracilis.

The type-specimen of this species from the Middle Jurassic rocks of Yorkshire, as shown in fig. 649, is an impression of an

04:9. Baiera gracilis. Typ^' specimen oi Bunbury (Bun-Collection, Botanynbsp;School, Cambridge).

iBiperfect leaf with a fan-shaped lamina deeply divided into forked near segments^: a better example is reproduced in fig, 650. Bean

referred this type to Schizopteris but Bunbury, while adopting Bean�s ms. specific name, substituted the generic designationnbsp;Baiera. The leaves.are petiolate and the lamina is divided almostnbsp;or quite to the base into a varying number of linear segmentsnbsp;with obtuse apices. The veins, frequently indistinct, run parallelnbsp;to the edges of the lamina and there are several in each segment.

Leaves identical with or very similar to Baiera gracilis are very widely distributed among Jurassic floras in both hemispheres.nbsp;Some of the specimens described by authors as Ginkgo lepida arenbsp;hardly distinguishable from Bunbury�s species; G. concinna} Heernbsp;from the Siberian Jurassic flora is another very similar form;nbsp;also Baiera incurvata Heer^ from the Lower Cretaceous of Greenland, B. angustiloha Heer�, as figured from Siberia and China,nbsp;B. hidens (Ten.-Wood8,H from Queensland, B. australis McCoy andnbsp;B. delicatula Sew. from Jurassic rocks in Victoria�, also leavesnbsp;recently referred by Halle � to B. australis from the Lower Cretaceousnbsp;plant-beds of Patagonia. Baiera gracilis is recorded from Uppernbsp;Jurassic (or Lower Cretaceous) beds in Alaska�, but leaves of this

type are rare in the Jurassic strata of North America. The Khaetic species B. Muensteriana^ (Presl) described by Schenk fromnbsp;Franconia and by other authors is a closely allied type which

1 Heer (77) ii. PI. xili. figs. 6�8. nbsp;nbsp;nbsp;^ Ib d. (82) B. PI. xiii. fig. 6.

^ Ib.'d. (78) ii. PI. vn. fig. 2; Krasser (05) PI. ii. fig. 10; Schenk (83) A. PI. Liii. fig. 1.

� Fontaine in Ward (05) B. PI. xLiv. fig. 2. nbsp;nbsp;nbsp;� Schenk (67) A. PI. ix.

Cannot always be distinguished from B. gracilis. The example shown in fig. 651 from the Yorkshire coast has been named B. gracilis forma Mmnsieriana to denote its close resemblance to thenbsp;Rhaetic species^. B. Guilhaumali^ described by Zeiller from Rhaeticnbsp;^ccks in Tonkin is another similar form but the leaves are narrowernbsp;�^id the apices of the segments more obtuse. On the one handnbsp;gracilis approaches close to B. Lindleyana, a species characterised by still narrower segments, and on the other it shadesnbsp;leaves agreeing vnth Ginkgoiies sibirica.

RomeP described this Jurassic species as Dicropteris longifolia ^'id Heer substituted the generic name Baiem\ The leaves re-�6Kible those of B. Simmondsi in the division of the lamina intonbsp;narrow linear segments 2�9 mm. in breadth, but the leaf isnbsp;narrower and cuneate; the segments have obtuse apices. Heernbsp;scribes the veins as parallel and simple, 3�7 in each segment,nbsp;^ith this species Heer associates some male flowers similar tonbsp;nse shown in fig. 654, also some detached seeds, but in neithernbsp;is there any convincing evidence of connexion. The Siberiannbsp;species B.Czekanowskiana^, recorded also by Holler from Bornholm,nbsp;fjnbsp;nbsp;nbsp;nbsp;not a distinct type. Thomas� records B. longifolia

m the Middle Jurassic series of Yorkshire and gives new facts regard to the structure of the epidermal cells; one of hisnbsp;specimens of an incomplete lamina is 12 cm. long, the whole leafnbsp;at least 18 cm. in length. The epidermal cells have a verynbsp;^nticle; those on the lower surface are arranged in longi-thnbsp;nbsp;nbsp;nbsp;and most of them have a prominent papilla; on

e lower face the cells are more rounded or hexagonal and the ^ta are much more numerous; each pair of guard-cells isnbsp;^^rounded by a group of 5�6 srrbsidiary cells (fig. 647, C,nbsp;as m Ginkgo. Krasser�� records this species from Jurassic rocks

3 nbsp;nbsp;nbsp;(00) B. p. 264.nbsp;nbsp;nbsp;nbsp;2 Zeiller (03) B. PI. n figs. 10�19.

fiss R T n- P- 52, Pi. VII. figs. 2, 3; Pi. vm.; PI. ix. figs. 1�11; PI. x. PI- ^v. fig. 116.

(03\ PI *�� P 56, PI. X. figs, 1�5; PI. vii. fig. 1; (82) B. PI. iir. figs. 4�8; Moller ' V. fig,

in China but the photographic reproduction is unfortunately too obscure to afford any indication as to the nature of the specimen.

This Jurassic species^ (fig. 652) illustrates the absence of any definite dividing line between Baeira and GinJcgoites; it agreesnbsp;very closely with G. sibirica andnbsp;with leaves assigned to G. lepidanbsp;and other �species.� Fig. 652 isnbsp;drawn from Phillips� type-specimen^ which he named Sphenopterisnbsp;longifolia and afterwards transferred to Cyclopteris; his specificnbsp;name is not retained because Pomelnbsp;adopted it for a type subsequentlynbsp;called by Heer Baiera longifolia^.

Krasser records B. PhiUipsi from Jurassic strata in Sardinia*. Thisnbsp;species shades into B. gracilis andnbsp;the very similar B. australis'KcQoy,nbsp;especially resembling some leavesnbsp;included by Halle� in the latternbsp;species.

Leaves of this type were first figured by Lindley and Hutton� as Solenites ? furcata and transferred by Braun to Baiera.nbsp;Schimper'^ subsequently substituted Jeanpaulia and proposed thenbsp;specific name Lindleyana on the ground that Heer had employednbsp;the designation furcata for a Rhaetic species of Baiera. Saportanbsp;included this species in Trichopitys. Baiera Lindleyana isnbsp;characterised by the deep dissection of the lamina into verynbsp;narrow, filiform, segments and by the presence of a long andnbsp;slender petiole (fig. 653). Some forms of this type with rathernbsp;broader segments are hardly distinguishable from Baiera gracilis.

In a former account of this species^ I included the specimen repror lt;Iuced in fig. 661 (p. 66), also a similar specimen figured by Phillips^nbsp;a distinct type, Baiera micro-phylla. The examination of additional material collected from thenbsp;Yorkshire coast by Mr Hamshawnbsp;Thomas leads me to substitutenbsp;^^hinowsMa for Baiera as thenbsp;*^ore appropriate name for thenbsp;hunch of leaves represented in fig.nbsp;hhl which is in all probabilitynbsp;identical with B. microphylla asnbsp;hgured by Phillips. In the case ofnbsp;incomplete leaves it is by no meansnbsp;easy to distinguish B. Lindleyananbsp;from Czekanowskia microphylla; butnbsp;in the latter the branches of thenbsp;i^'inina are separated by a smallernbsp;^iigle and if cuticular preparationsnbsp;available the stomata afford anbsp;means of differentiation: in Baieranbsp;Ihe guard-cells are surrounded bynbsp;* circular group of cells, while innbsp;^^kanowskia the subsidiary cellsnbsp;nie longer and narrower, forming a more oblong group.

Baiera Lindleyana is recorded also from Middle Jurassic rocks n Chinese^ Dzungaria and from Upper Jurassic rocks in Scotland .nbsp;^ome specimens described by Fontaine^ from the Black Hillsnbsp;(Lower Cretaceous) as Czekanouishia nervosa Heer are, as Berry�nbsp;points out, probably leaves of a Baiera, and I am disposed to refernbsp;them to B. Lindleyana.

�assic rocks, agrees in the form and dissection of the lamina quot; Seward (00) B. p. 266, fig. 46 (p. 268).

quot; Phillips (75) A. p. 200, fig. 9. nbsp;nbsp;nbsp;� Seward (11) Ft iv. fig- 44.

� Seward (ll^) PI. v. fig. 105.

quot; Pontame in Ward (99) B. p. 685, PI. 169, figs. 1, 2.

'Berry (11) p. 374. nbsp;nbsp;nbsp;i Bunker (46) A. p. 11, PI. v. fig. 4.

S. IV

with B. gracilis but is distinguished by the smaller dimensions. The imperfect example shown in fig. 641, B from the Kimeridgenbsp;beds of Sutherland (Scotland) ^ illustrates the unsatisfactory characters on which specific distinctions are drawn in the case ofnbsp;Baiera leaves agreeing in habit with B. gracilis. Better examplesnbsp;are figured by Schenk^ from the Wealden of North Germany.

This species, one of the smallest representatives of the genus, is described by Nathorst� from Upper Jurassic rocks of Spitz-bergen; it is characterised by the very narrow but apparentlynbsp;cylindrical segments and, except in its smaller size, resemblesnbsp;B. Lindleyana.

Our knowledge of seeds assigned to Mesozoic and Tertiary representatives of Ginkgo or to Baiera is limited to casts andnbsp;impressions of detached examples: no reproductive organs havenbsp;been discovered either in a petrified state or in connexion with anbsp;foliar shoot. Eeference has already been made to some smallnbsp;GinkgoAdkB seeds from the Eocene beds of Sheppey described bynbsp;Gardner as Ginkgo ? eccenica. Many similar seeds are figured bynbsp;Heer from Jurassic strata in Siberia and elsewhere, in most casesnbsp;as detached seeds but in a few instances borne singly or in pairsnbsp;on an axis resembling the peduncle of Ginkgo biloba'^. Heer�snbsp;seeds are correlated with G. digitata, G. sibirica and other speciesnbsp;but only on evidence afforded by association with leaves; theynbsp;are preserved as oval nuts, sometimes enclosed in a carbonaceousnbsp;envelope possibly representing an outer flesh, and resemble Ginkgonbsp;seeds in shape though they differ from them in their smallernbsp;size (8�9 mm. long and 6�8 mm. in diameter): in some of thenbsp;Jurassic specimens the nuts have an apical beak. All that cannbsp;be said is that seeds similar except in their smaller size to thosenbsp;of the recent species are not infrequently found in associationnbsp;with different species of Ginkgoites.

It has been suggested that the seed-bearing shoots, which rruthers named Beania, from Jurassic beds on the Yorkshirenbsp;coast may have belonged to a member of the Ginkgoales, but it isnbsp;^t least equally probable that Beania is Cycadean and possiblynbsp;seed-bearing axis of Nilssonia. The genus is described innbsp;� xxxviiii. It is possible that specimens from Cretaceous andnbsp;assic rocks regarded by Heer as male flowers of Ginkgoitesnbsp;^ca and other species, also specimens described by him asnbsp;se ^nbsp;nbsp;nbsp;nbsp;Schmidtianus^, may be fertile shoots, which bore

s and not microsporangia, belonging to Ginkgoites or some how^nbsp;nbsp;nbsp;nbsp;of the Ginkgoales: the nature of these fossils is,

th^^^'^^��'�^^*ion is scanty and indecisive. Nathorst� first suggested some small carbonised bodies from Yorkshire Jurassic bedsnbsp;^e*ired by Phillips* as �unknown leaves� are probably fragmentsnbsp;quot;which^^ ^c^�crs of some species of Ginkgoites. The specimen ofnbsp;COnbsp;nbsp;nbsp;nbsp;figured a small portion is shown in fig. 654, B; it

later* ^ slender axis with several short and partially broken lon^^^ ^^^^ches bearing terminal groups of oblong bodies 4 mm.nbsp;'vith^^^ Inbsp;nbsp;nbsp;nbsp;wide, 2�4 in each group: these suggest comparison

with longitudinal dehiscence, and the habit of the ^ilobanbsp;nbsp;nbsp;nbsp;shoot agrees with that of a male flower of Ginkgo

�iiigle microsporophyll the resemblance between the fossil PossibT^^^nbsp;nbsp;nbsp;nbsp;fairly close�. Unfortunately it has not been

I,Qq- ^icrospores, and while the probability is that the oblong seeds ^^lerosporangia it is not impossible that they are smallnbsp;be a mnbsp;nbsp;nbsp;nbsp;�f identical bodies showing what appears to

the axis is 2-3 cm. long and some microsporangia are seen in their original position, while others are detached^. It is by no meansnbsp;unlikely that these specimens are portions of male flowers ofnbsp;Ginkgoites digitata or of some other species, but this cannot benbsp;definitely settled until better material is available. Some Rhaeticnbsp;fossils described by Nathorst^ as Antholithus Zeilleri present anbsp;certain resemblance to these supposed male flowers. One of

Nathorst�s specimens from Scania is reproduced in fig. 655 twice natural size; the photograph, for which I am indebted to Prof.nbsp;Nathorst who published it in 1908, shows a cuticular preparationnbsp;of the axis and microsporangia. The axis of Antholithus isnbsp;dichotomously branched and bears terminal clusters of microsporangia about 3-5 mm. long, usually eight in a cluster; several ofnbsp;them have dehisced longitudinally and the apices show a slightnbsp;separation of the two halves. In some of the sporangia Nathorstnbsp;found micro spores with an average length of 40�43 p, agreeingnbsp;closely with the spores of Ginkgo and recent Cycads. Nathorstnbsp;considered that Antholithus Zeilleri may be a male flower of somenbsp;Ginkgoaceous plant though a correlation with a Cycadean type isnbsp;by no means excluded. There is, however, a general resemblancenbsp;between the English Jurassic specimens shown in fig. 654 and thenbsp;Rhaetic species; the latter is distinguished by a greater tendency

at maturity into 10�12 spreading lobes each of which resembles in appearance a microsporangium of Antholiihus Zeilleri. Nonbsp;spores have been isolated and, as Nathorst points out, the agreement with the Scanian specimens is to a large extent superficial.

The Australian specimens, probably of Ehaetic age, described by Shirleyfrom Ipswich, Queensland, as Stachyopitys annularioidesnbsp;and S. Simmondsi require further investigation; they may benbsp;allied to Stachyojntys Preslii Sch., though neither their morphological nature nor systematic position can be settled without freshnbsp;data. Halle^ describes some examples of a similar kind from thenbsp;Jurassic beds of Graham Land as Stachyopitys, cf. annularioidesnbsp;Shir, and thinks it probable that they are portions of some Gymno-spermous male strobilus, but, as he points out, the absence of anynbsp;member of the Ginkgoales in these southern beds is noteworthy.nbsp;Specimens similar to those described by Schenk, Shirley, and Hallenbsp;are also figured from Ehaetic beds in South America as Sphefiolepisnbsp;rhaetica^ and from the Stormberg (Ehaetic) series of South Africanbsp;as Stachyopitys sp.^

The generic name Ginhgoanthus has been adopted by Nathorst� for a fragmentary specimen from the Upper Jurassic of Franznbsp;Josef Land which he considers may be a male flower of a Gink-goites', but the preservation is too imperfect to admit of satisfactory determination. As regards terminology, in the presentnbsp;state of our knowledge it is preferable to use the non-committalnbsp;designation Antholiihus^ for the English, Scanian, Swiss, andnbsp;Franz Josef Land fossils, leaving Schenk�s Stachyopitys Preslii asnbsp;a type apart. As regards the English and Swiss specimens, thenbsp;probability would seem to be that they are the microstrobili ofnbsp;some members of the Ginkgoales.

This generic name� was first used by Saporta� for Nathorst�s Zamiostrobus scanicus from Ehaetic and Liassic rocks in Scania�

* Shirley (98) p. 13, PI. xviii. nbsp;nbsp;nbsp;� Halle (13) p. 88, PI. vi. fig. 13.

� Geinitz (76) B. p. 12, PI. n. nbsp;nbsp;nbsp;* Seward (03) B. p. 66, PI. ix. fig. 2.

� Used by Nathorst in Linnaeus�s and not Brongniart�s sense; Nathorst (08 23.nbsp;nbsp;nbsp;nbsp;� crrer�s, narrow; pdx�S, the backbone.

for a Liassic species from Belgium, S. Ponceleti. Nath subsequently adopted Saporta�s genus. I have elsewhere suggesteanbsp;the application of Stenorachw to various species described by Meernbsp;from Jurassic and Cretaceous beds as male flowers of Gmkgoaceae.nbsp;Although there is no proof as to the morphological nature ol the

specimens included in this genus some of them, e.g. S. present the appearance of seed-bearing shoots though, as anbsp;is careful to point out, the seed-like bodies may not be ^rue seeds,

I am inclined to regard Heer�s supposed male flowers (fig. possibly fertile shoots of some members of the Gmkgoales w icnbsp;originally bore seeds, but this view is merely tentative, htenoracnbsp;is employed as a designation for specimens consisting o anbsp;axis, generally fairlv stout, bearing lateral appendages, w e ernbsp;axial or foliar cannot be definitely determined, either simple or .nbsp;�rked and in some cases with terminal seed-like bodies but usuallynbsp;�ith a small distal swelling or a few spreading lobes as mnbsp;(Heer)*. Some at least of the specimensnbsp;lu this genus probably belong to Ginkgoaceous plants, thougnbsp;regard to others, e.g. S. scanicus, it should be remembered thatnbsp;i^uthorst inclines towards a Cycadean affinity. The genus Beamanbsp;'^as founded on specimens similar in general habit to species ofnbsp;^tenorachis but in Beania the appendages have a comparative ynbsp;large terminal shield bearing on its adaxial side two seeds.

� nbsp;nbsp;nbsp;� scamcMsi and afterwards transferred to Stemrachis, is represented

� nbsp;nbsp;nbsp;several lateral appendages, spoken of by Nathorst asnbsp;sporophylls; these are split into two divergent arms each of which

, bears on the side away from the fork an oval, longitudina y striated, body described as thick and woody (fig. 656^ Thenbsp;uature of these bodies is uncertain and Nathorst is incline tonbsp;:nbsp;nbsp;nbsp;nbsp;think they are not seeds; he suggests as an alternative interpre

tation that they are laminar structures in which microsporangia jnbsp;nbsp;nbsp;nbsp;�^re embedded. The morphology of this Rhaetic and Liassic

^ Heer (77) �. PI. xi.; (78) ii. PI. vi. fig. 8; (82) it -J. PP. 18, 21, PH. vi,, ix.

species referred by Nathorst to the Cycadophyta^ is, therefore, uncertain. It is by no means certain that it has not an equalnbsp;claim to inclusion in the Ginkgoales; there are no substantial

grounds for such relationship, but the resemblance of this and other species of Stenorachis to abnormal seed-bearing shoots ofnbsp;Ginkgo hiloha may be significant (cf. fig. 631, D, p. 5).

A similar but rather smaller type was described by Heer^ from Upper Jurassic rocks in Spitzbergen as Carpolithes striolatus.nbsp;Nathorst^ examined Heer�s figured specimens and recognised onenbsp;of them as an example of Stenorachis, agreeing in the possessionnbsp;of forked appendages with S. scanicus and bearing seed-like bodies.

The fossils described by Shirley^ from Ehaetic(?) beds in Queensland as Beania geminata are similar in habit to Stenorachisnbsp;scanicus and differ from Beania gracilis Carr, in the absence ofnbsp;distally expanded sporophylls.

Another Rhaetic species is described by Nathorst� as Stenorachis Solmsi in which the �sporophylls� have a different form and arenbsp;characterised by a distal, erect, laminar expansion deeply dividednbsp;into two segments: no seeds or microsporangia have been found.

The species for which this name has been suggested was originally regarded by Heer as the male flower of the Jurassic species Ginkgo

pond P^��tice of some reproductive organs: some of the ap-ges are bilobed as in the Afghan example, at states that some of his specimens bear 2�3 pollen-sacsnbsp;j ^^P� of the appendages, but the published figures afford nonbsp;3nbsp;nbsp;nbsp;nbsp;ii PI. XI.nbsp;nbsp;nbsp;nbsp;2 Heer (82) A. p. 18, PI. VI. figs. 1�6. -

* � P- 21, PI. IX. nbsp;nbsp;nbsp;4 Seward (12) p. 23, PI. iv. fig. 52.

confirmation of this and an examination of some of Hner�s material lent to me through the kind offices of Dr Zalessky failed to revealnbsp;any indication of spores or sporangia. In Heer�s Antliolithmnbsp;Schmidtianus the lateral appendages are said to hear 3�4 pollen-sacs in a terminal whorl, but Heer also suggested the possibilitynbsp;that these bodies are the segments of a calyx-like envelope, a morenbsp;probable interpretation. It may be that the terminal bodies arenbsp;homologous with the slightly expanded distal ends of the appendages in S. lepida and possibly with the collar at the base of thenbsp;ovules of Ginkgo biloha, which in the case of S. Schmidtianus hasnbsp;the form of a lobed cupular organ which enclosed a seed. It isnbsp;noteworthy that Heer�s figures show a central scar surrounded bynbsp;the spreading lobes.

The incomplete Jurassic specimens from Victoria (Australia)^ described as possibly parts of a female shoot of a Ginkgoaceousnbsp;plant resemble Stenorachis lepida and should be referred to thenbsp;same genus.

A specimen like those represented in fig. 657 has been figured by Krystofovic^ from Jurassic rocks in Ussuriland as Ginkgo sp.nbsp;An imperfect fossil described from Jurassic beds in Australia asnbsp;possibly a seed-bearing shoot of a Ginkgoaceous plant^ should benbsp;included in Stenorachis, as also Raciborski�s Ixostrohus Siemirad-zkii* from Rhaetic beds of Poland.

In no case have we any decisive evidence with regard to the parentage or morphological nature of the specimens referred tonbsp;Stenorachis, but any material that may represent fertile shootsnbsp;belonging to Ginkgoales or Cycadophyta should be described innbsp;the hope that additional facts may be obtained.

A genus founded� on some well-preserved leaves from the Middle Estuarine (Middle Jurassic) beds of Gristhorpe Bay on thenbsp;Yorkshire coast, and named Eretmophyllum from the paddle-likenbsp;form of the lamina�. Leaves oblanceolate to linear reaching anbsp;length of 12 cm. and a breadth of 2 cm.; in the type-species.

E. puhescens Thom., the leaf is from 7 to 10 cm. long and 1 3 cm ^road; the apex is rounded or refuse (fig. 658, B), the base taperingnbsp;gradually towards the petiole. Veins 1�1-5 mm. apart, die o-tomously branched in the proximal part of the lamina and usual ynbsp;parallel and simple except where they converge at t e apexnbsp;epidermal cells (preserved in the Yorkshire species) are po ygona ,nbsp;^vith or without papillae (fig. 658, D); the stomata are characterisenbsp;an enclosing group of subsidiary cells as in Ginkgo.

Secretory tracts occur between the veins of the smooth lamina

Ginkgo. The polygonal cells with straight or slightly undulate wallsnbsp;^re characterised by papillae (fig.

^ewer epidermis to which the sto- I N Hiata are confined; the slightlynbsp;�depressed stomata are in regularnbsp;��ows and the guard-cells are sur-^oiinded by 4�7 subsidiary cells,nbsp;another species, E. wJiithien.^enbsp;Whitby, the surface of thenbsp;lamina,which maybe 7cm. longandnbsp;^�2 cm. broad and slightly falcate,nbsp;rough, and between the veinsnbsp;strands of elongated cells, pos-denoting the presence of hypo-�rnial stereome. Stomata occurnbsp;both surfaces and the papillaenbsp;confined to the subsidiary cells.

�^^^amata, which Schenk^ suggests may be a segment of should probably be included in Eretmophyllum.

This species (fig. 658, A), originally referred with some misgiving to Podozamites'^ and compared with Yokoyama�s Ginkgodium, isnbsp;from Jurassic beds in Afghanistan. There can be little or nonbsp;doubt as to its generic identity with the Yorkshire leaves. Thenbsp;broadly linear lamina tapers gradually to a slender petiole and thenbsp;veins, 1 mm. apart, are simple except at the proximal end.

Mr Thomas is certainly justified in his opinion that Eretmo-phyllum is a member of the Ginkgoales. In shape the leaves resemble Ginkgodium and differ but little from some Jurassicnbsp;specimens referred to Ginkgoites. They agree in venation, in thenbsp;presence of short secretory tracts, in the structure of the epidermalnbsp;cells and stomata with Ginkgo and species of Ginkgoites. Eretmophyllum is distinguished from Feildenia by its larger leaves, anbsp;coarser venation, and a more definite petiole. Some leavesnbsp;figured by Fontaine^ from Jurassic-Cretaceous rocks of Alaskanbsp;as Nageiopsis longifolia? Font, have little claim to be included innbsp;that genus� and may perhaps be allied to Eretmophyllum.

CHAPTER XLI.

believed to belong to the GmKGOALES BUT WHICH ON THE AVAILABLE EVIDENCE CANNOT BE REFERRED WITHOLFT HESITATION TO THAT GROUP.

defined the genus as follows: �Leaf coriaceous, th' 1 iobed, gradually narrowed towards the base which isnbsp;at its margin and gradually passes into a short petiole;nbsp;Henbsp;nbsp;nbsp;nbsp;simple, parallel; interstitial veins very fine.�

la � attention to the thickening of the lower margin of the quot;'�ein ^ feature reminiscent of Ginlcgo and to the course of thenbsp;� which run parallel to the median axis of the lamina insteadnbsp;�0nbsp;nbsp;nbsp;nbsp;from the base as in Ginhgoites and Baiera. Gink-

�Hucture,the

supposed affinity to Ginkgo rests therefore on leaf-form alone.

^^9odium Nathorsti Yokoyama.

Xti

j � type-species was founded on specimens from strata in thenbsp;nbsp;nbsp;nbsp;by Yokoyama^ to the Middle Jurassic series, but

�y8te �^?gests a somewhat higher horizon in the Jurassic � Some of the leaves are entire, obovate, and have anbsp;'vhic^^^nbsp;nbsp;nbsp;nbsp;end; others are cuneate and broader at the apex

th^ ^�sembles the Palaeozoic genus Whittleseya in the position Veins but the genera are unlikely to be confused; a com-jgg 6^ iriay also be made with the Jurassic genus Eretmophyllumnbsp;quot;^Hch has longer'and narrower leaves with a coarsernbsp;ion. We have no information with regard to the cuticular

deep median sinus into divergent obtuse segments (fig. 659). One leaf is described as 6-6 cm. long, 2-1 cm. broad with thirty veinsnbsp;and an interstitial � vein � betweennbsp;each pair: the interstitial � vein �nbsp;is due to the presence of an intercostal stereome strand. Thomas^nbsp;records this species from thenbsp;Bathonian series of Kamenkanbsp;in the south of Russia (fig. 659,

B). The specimens from Alaska named hy^ontaine^ Ginkgodium?nbsp;alaskense agree more closely withnbsp;Ginkgoites.

Heer� gave this name (after Czekanowski who discovered thenbsp;specimens) to fascicules of longnbsp;and narrow, filiform, leaves with a simple or occasionallynbsp;forked lamina borne on a short supporting axis covered withnbsp;broader and shorter scale-leaves. The deciduous fascicules ornbsp;dwarf-shoots are similar to those of Phoenicofsis. Bunchesnbsp;of Czekanowskia leaves with their short scale-covered supportingnbsp;axes resemble the dwarf-shoots of Pines'^. Heer assigned to thisnbsp;genus some seeds associated with the leaves, also what he believednbsp;to be a male flower�, an example of a reproductive shoot of thenbsp;type described on page 57 as Stenorachis. There is, however,nbsp;no conclusive evidence as to the nature of the reproductive organs.nbsp;The venation is seldom shown on the carbonised laminae; somenbsp;leaves are finely striated while on others there may be one or twonbsp;narrow ridges that represent veins, but as a rule the impressionsnbsp;afford no indication of the venation. Czekanowskia was placednbsp;by the author of the genus in the Ginkgoales, the short shootsnbsp;being compared with those of Ginkgo though, except in the largernbsp;number of the leaves, they closely resemble the foliar spurs of

� Heer (77) ii. p. 65. nbsp;nbsp;nbsp;�* Cf. Finns fiexilis; Bot. Hag. Tab. 8467.

fines. The dichotomous branching of the lamina in some forms another feature in which GzekanowsMa resembles Baiera andnbsp;Ginkgo, a resemblance which derives a certain significance fromnbsp;I'fe occurrence of stomata of the Ginkgoaceous type. Nathorst^nbsp;described the cuticular membranes of the superficial layers:

� epidermal cells have straight walls and the stomata, more iinmerous on the lower surface, are accompanied by four or fivenbsp;subsidiary cells: these do not form a circular group as in Baieranbsp;^nd Ginkgoites, but, as the result of elongation in the direction ofnbsp;e long axis of the leaf, the group is relatively long and narrow,nbsp;icular preparations can often be made from the well-preservednbsp;^ that occur in great abundance in the shales of Gristhorpenbsp;and elsewhere on the Yorkshire coast^, and some particularlynbsp;good examples were collected by Prof. Obrutschew from Jurassicnbsp;^ocks in the Djair Mountains in Chinese Dzungaria�; these occurnbsp;^ith carbonised remains of Ginkgoites leaves in papery massesnbsp;�oinlar in the manner of preservation to the Palaeozoic papernbsp;from Russia^. The epidermal cells of the Dzungaria Czek-det*^*^^^� possibly identical with G. rigida but too incomplete to benbsp;lon^^^*^^^ with certainty, have straight walls and are relativelynbsp;g) the stomata are scattered and appear as dark patches, theirnbsp;or colour being due to the thick cuticles of the two or threenbsp;. cells on the sides of the stoma; the epidermal features arenbsp;those described hy Nathorst�, but in the Rhaetic specimensnbsp;om Scania the heavilv cuticularised accessorv cells are generally

been suggested by Jeffrey� that Gzekanoivskia may be aucarian in its affinities, but this opinion rests on the slendernbsp;once of association of Gzekanowskia-Yike leaves in Middlenbsp;ceous rocks with a stem described as Araucariopitys sonbsp;fp i ^ because of the association of Araucarian and Abietineousnbsp;in f. Such evidence of affinity as we have would seem to benbsp;datnbsp;nbsp;nbsp;nbsp;relationship with Baiera and Ginkgo though decisive

4 Y^'^^rd (00) B. p. 278, fig. 48. nbsp;nbsp;nbsp;� Seward (11) p. 49, Pis. iv., v., vi., vii.

On the specimen shown in fig. 660, A Lindley and Hutton^ founded the species Solenites Murrayana which they compared

Murrayana, Lind, and Hutt., from the Middle Jurassic plant-beds of,Yorkshire. (British Museum, no. 3685, no. V. 3684.)

with Isoetes and Pilularia. The type-specimen is from the Middle Jurassic plant-bed at Gristhorpe Bay on the Yorkshire coast.nbsp;They describe the narrow leaves as converging to a common

but the actual axis is not preserved; the lamina is longi-I'Udinally striated but no veins are shown either on the carbonised ttnua or in the magnified cuticle figured in the original description.

the specimen reproduced in fig. 660, B the leaves are seen to � attached to a short and relatively broad axis covered withnbsp;�cale-leaves, one of which is shown bent over on one side of thenbsp;arf-shoot. The comparatively large size and the pendulousnbsp;position of the scales are characteristic features of the genus whichnbsp;Well seen in figures of CzekanowsTcia published by Nathorst^.nbsp;�s species was included by Saporta in Jeanpaulia and laternbsp;nsferred to Trichopitys, while Zignh^ and some other authorsnbsp;�8arded Solenites Murrayana as a species of Isoetes. The com-P^'rison with Isoetes suggested by the form of the leaves is notnbsp;^orue out by the structure of the epidermal cells. Phillips�nbsp;8ured a specimen in 1829 as Flabellaria ? viminea: this specificnbsp;Hot ^nbsp;nbsp;nbsp;nbsp;employed before the publication of Murrayana has

been adopted by authors. Some of the specimens included Heer in his account of the species C. rigida should be referrednbsp;j Murrayana, but in a previous description�s of the speciesnbsp;'''ent too far in uniting C. Murrayana and C. rigida. In Czeha-^Kia Murrayana the leaves, usually about 1 mm. broad butnbsp;arenbsp;nbsp;nbsp;nbsp;narrower, reach a length of more than 17 cm.; they

^ '^iibranched and in this respect and in their slightly greater differ from C., rigida.

^^nowskia microphylla (Phillips), as specimen figured by Phillips� from the Yorkshire coastnbsp;Hat microphylla is undoubtedly identical specifically withnbsp;p . ^�produced in fig. 661, and both were formerly included innbsp;The chief reasons for transferring them tonbsp;are the more acute angle of divergence of the filiformnbsp;�nts, the difference in the shape of the,leaves, the absencenbsp;hotnbsp;nbsp;nbsp;nbsp;hhd the occurrence of the leaves in a fascicle, a habit

�^quot;'llips (75) A. p. 200, fig. 9. nbsp;nbsp;nbsp;� Seward (00) B. p. 266.

unknown in the genus {e.g. B. paiicifartita). Some specimens which may be identical with this type were obtained 'several years agonbsp;by Dr Nathorst from Yorkshirenbsp;but never fully described: annbsp;examination of his unpublishednbsp;drawings and of specimens collected by Mr Hamshaw Thomasnbsp;convinced me that some formsnbsp;of Czekanowskia are more freelynbsp;branched and exhibit more variation in the breadth of the laminanbsp;than I had formerly supposed.

In specimens of the type first noticed by Nathorst some of thenbsp;segments are comparativelybroadnbsp;and fern-like, a feature that isnbsp;not seen in the leaves shown innbsp;fig. 661. This species affords anbsp;striking contrast to Czekanowskianbsp;Murmyana in which the longnbsp;leaves are unbranched, and asnbsp;Nathorst^ suggests the namenbsp;Snleniies might be revived for the unbranched type; but in thenbsp;absence of any difference in the epidermal characters, it wouldnbsp;seem undesirable to raise to generic rank a feature depending onnbsp;the simple or branched habit of the leaves of otherwise similarnbsp;leaf-fascicles.

This species, founded on specimens from Siberian Jurassic rocks^, is characterised by its branched filiform leaves borne 0�nbsp;short shoots enclosed by scale-leaves, triangular or lanceolate innbsp;form and in some specimens pendulous on slender stalks. Nathorst�nbsp;states that the epidermal structure of the scale-leaves is similarnbsp;to that of the scale-leaves on short shoots of Ginkgo. Thenbsp;characters of the stomata are mentioned in the account of the

genus. The leaves are generally slightly narrower than the ^nhranched needles of C. Murrayana, but the habit of the dwarf-�toots is the same. The leaves often show line striations; innbsp;^ost specimens there is no indication of clearly marked veinsnbsp;though two or three vascular strands are sometimes visible.nbsp;Ileer on very slender evidence refers to this species some seedsnbsp;�^d a �male flower.�

It is not always easy to distinguish between imperfect examples C. rigida and Baiera Lindleyana: the leaves of the latter typenbsp;^quot;^6 petiolate and the segments diverge at a wider angle. Twonbsp;^6aves with spreading bifurcate segments figured by Fontaine^nbsp;Lower Cretaceous rocks in the Black Hills as Czekanowskianbsp;'^^^vosa Heer afford no indication that they were borne in clustersnbsp;dwarf-shoots but resemble the petiolate leaves of Baieranbsp;^-''^^dleyana. Berry^ points out a similarity between Fontaine�snbsp;^�^ssils and Baiera foliosa Font. Heer�s type-specimens of C.

from the Wealden of Portugal� are more like Czekanoivskia eaves. The leaves described from Siberia as C. setacea HeerLnbsp;though narrower than some forms of C. rigida, are probably notnbsp;specifically distinct.

Czekanowskia rigida is characteristic of Jurassic strata, and 'I'^curs in Europe, including Greenland, also in Siberia, China, andnbsp;'^apan.

^ The branched leaves described under these names� from �Cetaceous rocks in Greenland and North America are in mostnbsp;^^8es not sufficiently complete to be assigned with certainty tonbsp;� genus Czekanowskia', the examples figured by Hollick andnbsp;ffieys from Middle Cretaceous beds as C. capillaris occur innbsp;^^osely packed groups, but no specimens have been discoverednbsp;owing any scale-covered supporting axis. Mobile admitting thenbsp;^ability that these species and C. nervosa�� from Wealden strata

may be allied to Czekanowshia rigida, such evidence as is available points to a maximum development of Czekanowskia in the Jurassicnbsp;period,

In 1870 Heer^ described some small linear leaves from Tertiary strata in Spitzbergen for which he proposed the generic namenbsp;Torellia, defining it as follows: � Folia rigida coriacea, basin versusnbsp;angustata, articulata^ tenuiter costata, costis interstitiisque sub-tilissime striatis.� On the discovery by Capt. Feilden of additionalnbsp;specimens in Miocene beds in Grinnell Land (81� 46' N.) Heernbsp;published a further account of the genus and substituted Feildenianbsp;for Torellia because of the previous use of the latter name bynbsp;Zoologists. Heer compared Feildenia with Podocarpus, Araucaria,nbsp;and other Conifers but, mainly because of the occurrence of anbsp;leaf with a lobed lamina, he provisionally included the genus innbsp;the Taxineae^. The leaves are usually found as detached specimens but in one case several are spirally disposed on a stout axisnbsp;and one imperfect example shows at the base what appears to benbsp;a scale-leaf, suggesting that leaves were also borne on short shootsnbsp;like those of Phoenicopsis and Czekanoivskia. � Heer lays stressnbsp;on the ribbing and striation of the surface of the lamina as distinguishing features between Feildenia and Phoenicopsis, butnbsp;Nathorst^, in his revision of the genus, expresses the opinion thatnbsp;it is only in the tendency to a sickle-like form and a feeble expansionnbsp;of the slightly curved base that Feildenia, at best an ill-definednbsp;genus, can be distinguished from Phoenicopsis.

This species, from Miocene beds of Spitzbergen^ and Grinnell Land�, is represented by linear leaves 6�8 cm. long and 5�8 mm-broad at the widest part, usually rather nearer the apex than thenbsp;base; the lamina is often slightly falcate and tapers gradually tonbsp;a narrow base. There are 8�-11 veins for the most part parallelnbsp;but occasionally feebly convergent at the bluntly rounded apex.

* Heer (71) i.i. p. 44, PI. vi. figs. 3�12; PI. xvi. fig. 1 h. ^ Heer (78) i. p. 20, Pis. i., ii., viii.

A species from Upper Jurassic rocks in Spitzbergen founded by * norst^ on leaves similar to those of F. rigida but smaller; thenbsp;^ ^ Ilia is generally 3�4 mm. broad and may reach a length ofnbsp;'i cm. There are usually six veins and as in other species finernbsp;hnes occur between the true veins. A few smallnbsp;^xes very similar to F. Nordenshdldi are described by Nathorstnbsp;^eildenia sp.^ from Franz Josef Land, probably of Wealden age.nbsp;Until further evidence is available it is impossible to fix pre-^^6 position of the genus. Though often distinguished bynbsp;sickle-shaped lamina and the broad apical region from leavesnbsp;tKnbsp;nbsp;nbsp;nbsp;it is not always possible to separate the leaves of

^j.^^'^^nicopsis was founded by Heer� on linear leaves from lo ^ '^iii�a-ssic strata in Siberia; the leaves, in extreme cases 20 cm.nbsp;six^ varying in breadth from 2 mm. to 2 cm., occur in clusters ofnbsp;* Or more and even as many as twenty on very short and relativelynbsp;w ��Xes covered with small scale-leaves. These dwarf-shootsnbsp;deciduous: the lamina is fairly uniform in breadth but tapersnbsp;towards the slender base and is usually obtusely roundednbsp;�apex; the veins are parallel and very rarely diehotomouslynbsp;^ auched. The features on which species are founded are oftennbsp;le Systematic value; they are the breadth of the lamina, thenbsp;c . ^anient of the veins, the presence or absence of interstitialnbsp;,nbsp;nbsp;nbsp;nbsp;It is very doubtful whether much confidence can be

on the occurrence of the so-called interstitial veins: in ^ '�pecies of Phoenicopsis the parallel veins show no trace ofnbsp;'vein� between them, but occasionally in a leaf of thenbsp;Species there are indications of interstitial �veins.� Thenbsp;l^lie lamina is also an uncertain guide; well preservednbsp;show that the leaves may reach a considerable lengthnbsp;118nbsp;nbsp;nbsp;nbsp;lamina gradually decreases in breadth towards the

in the case of detached leaves which it is often impossible to determine even generically. The characters usually employed fornbsp;the separation of different forms are conveniently shown in a tablenbsp;published by Krasser^. There is no information available as tonbsp;the epidermal structure of the various types of Phoenicofsis leaves,nbsp;nor have we any data with regard to the reproductive organs. Thenbsp;genus is generally included in the Ginkgoales: the dwarf-shootsnbsp;agree closely with those of Czekanowskia which, in the structurenbsp;of the epidermis and in the bifurcation of the leaves, resemblesnbsp;Ginkgo and Baiera. The precise position of Phoenicopsis cannotnbsp;be regarded as settled. The only evidence with regard to anatomical structure is that furnished by Solms-Laubach^ whonbsp;described petrified leaves from Jurassic rocks in Franz Josefnbsp;Land which are probably examples of Phoenicopsis; but, assumingnbsp;that they belong to this genus, the anatomical data are insufficientnbsp;to determine the position of the genus within the Gymnosperms.nbsp;Incomplete and detached leaves agreeing in their venation and innbsp;the form of the lamina with those of Phoenicopsis cannot as anbsp;rule be distinguished from leaves of Podozamites, Feildenia, ornbsp;even from narrow forms of Cordaites. The Jurassic specimensnbsp;from North Germany on which Saif eld� founded the genus Phyl-lotenia should probably be assigned to Phoenicopsis. Solms-Laubach refrained from assigning the imperfect Franz Josefnbsp;Land leaves to Phoenicopsis because no dwarf-shoots were found;nbsp;he employed the non-committal generic name Desmiophyllum, anbsp;designation that might with advantage be more frequently usednbsp;for specimens which cannot be proved to belong to Phoenicopsis,nbsp;Podozatniies or other genera with similar leaves.

Desmiophyllum. Lesquereux established the genus Desmio-phyllum* for some narrow sublinear leaves from the Coal Measures of Pennsylvania similar to those of Poacordaites and attached tonbsp;an imperfectly preserved axis either singly or in small groups.nbsp;The type-species B. gracile is probably a species of Cordaites: thenbsp;name Desmiophyllum never came into general use until its revivalnbsp;by Solms-Laubach in 1904 as a convenient term to apply to linear

due to the section having passed slightly below the level of the epidermal surface. The cells of the epidermis are short and havenbsp;straight walls.

Though we should not be justified in asserting that the leaves named Desmiophyllum Solmsi are examples of Phoenicopsis, thenbsp;probability is that they belong to that genus. Phoenicopsis is

Fig. 662. Desmiophyllum Solmsi. A, transverse sections of leaves; B, tangentia section of the lamina showing veins and mesophyll; C, stomata. (Fromnbsp;sections in the Museum of the Geological Survey. London.)

especially characteristic of Jurassic rocks and is best represented in the Middle Jurassic series of Siberia. The genus is recorded alsonbsp;from Spitzbergen^, Franz Josef Land, and Bornholm^; a species,,nbsp;P. GunnP, has been described from Upper Jurassic rocks in Scotland, the only example of the genus in Britain. FeistmanteUnbsp;1 Nathorst (97), p. 16.nbsp;nbsp;nbsp;nbsp;^ Moller (03) p. 30.

in the lamina and a faintly marked interstitial �vein/ probably the impression of a stereome strand^ between each pair of veinsgt;nbsp;In one specimen Heer found 21 leaves in a single cluster butnbsp;usually the number on a single dwarf-shoot is smaller. It wasnbsp;the superficial resemblance of a cluster of these leaves (fig. 663)nbsp;to the leaf of some Palms that suggested the name Phoenicopsis.nbsp;The leaves described by Heer as P. latioP are not distinguishednbsp;from P. speciosa by any very definite character. Examples ofnbsp;detached leaves from Lower Jurassic rocks in Bornholm describednbsp;by M�ller^ as cf. Phoenicopsis latior may equally well be referrednbsp;to Podozamites.

The leaves are 4 mm. broad or less and have 6�8 veins without interstitial veins�. This species is recorded from Russia^ Siberia,nbsp;China, and the Arctic regions, and leaves of similar type arenbsp;represented by Phoenicopsis media Krasser�, which is probablynbsp;merely a form of P. angustifolia^ P. taschkessiensis Krass, fromnbsp;China ; also Chinese specimens first described by Potoni�^ withoutnbsp;a specific name and afterwards named by Krasser� P. Potoniei.

The species Phoenicopsis Gunni from Scottish Kimeridge beds is a similar type with leaves 3�4 mm. broad and 12 cm. long withnbsp;eight veins and indications of interstitial �veins.�

Morris� founded this species on a linear leaf, now in the British Museum, from the Jerusalem Basin (Triassic), Tasmania, which henbsp;referred to Zeugophyllites, a genus founded by Brongniart on anbsp;specimen from the Lower Gondwana rocks of India but nevernbsp;figured. To the same species McCoyi� referred some broadernbsp;leaves from Permo-Carboniferous strata in New South Wales:nbsp;these were shown by Etheridge^^ and Arber^^ to be distinct fromnbsp;Morris�s type and the latter author identified them with Noeg-

* nbsp;nbsp;nbsp;Seward (07'^), PI. viii. fig. 69.nbsp;nbsp;nbsp;nbsp;^ Krasser (00) B. p. 147, PI. iii. fig. 4. '

geraihiofsis Goe'pperti (Schmal.) a nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;Hislopt (Bunb.)-

Similar leaves have been describe y - absence of specimens from Bhaetic rocks in South America,nbsp;nbsp;nbsp;nbsp;confidence as to

very close resemblance to some o , gg ip Australia, Sou Phoenicopsis. The occurrence of thesenbsp;nbsp;nbsp;nbsp;. ^ke leaves recorded

Africa, and South America and the p^ioewicops-is m

1 Feiatmantel (90) A. p-1 nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;^ xO.

CHAPTER XLII.

In this section are included several Palaeozoic genera most of which have heen assigned to the Ginkgoales on evidence which innbsp;most cases is wholly inadequate.

Probably the oldest specimens referred to the Ginkgoales are some imperfect leaves from Middle Devonian rocks in Bohemianbsp;described by Potoni� and Bernard as Barrandeina Dusliananbsp;(Krejci)h but there is no evidence of affinity to this or to any othernbsp;class of plants.

This designation has recently been proposed by Zalessky� for Schmalhausen�s Permian species Ginkgo cuneata which has nonbsp;substantial claim to be regarded as generically identical withnbsp;Ginkgo hiloha. Zalessky considers that it may belong to thenbsp;Ginkgoales though the available data hardly justify more than anbsp;suggestion of possible relationship.

A Permian species from the Altai mountains� represented by an obovate spathulate leaf 11 cm. long and 4 cm. broad with a comparatively long � stalk � consisting of a narrow portion of the lamina:nbsp;the lamina is traversed by slightly spreading veins about 1 mm.nbsp;apart. In form this leaf resembles the genus Eretmophyllum�^ (cf.nbsp;fig. 658) and may be a species of that genus^, but in the absencenbsp;of any information with regard to the epidermal features it isnbsp;inadvisable to adopt a name implying affinity to the Ginkgoales.

This generic term proposed by Zalessky^ for Schmalhausen�s wkgo CzeJcanowsMi, but not defined by him, may be adopted fornbsp;certain leaves resembling those of Ginkgo but which there is nonbsp;^deqnate reason for regarding as the leaves of a member of thenbsp;'^mkgoales.

e Permian of the Lower Tunguska river as Ginkgo Czekanowskii more or less orbicular and the lamina, approximately 1-5 cm.nbsp;is^'i?*^'nbsp;nbsp;nbsp;nbsp;into several bilobed segments (fig. 664, C). It

y no means certain that all the fragments included in this pecies are specifically identical; some bear a close resemblancenbsp;Ginkgo minuta Nath.� from the Rhaetic of Scania, a type whichnbsp;assigns with hesitation to Ginkgo and compares with the leavesnbsp;tonbsp;nbsp;nbsp;nbsp;peUatuni. It is impossible without additional data

etermine either the svstematic position of the specimens Ttinbsp;nbsp;nbsp;nbsp;these two species or their morphological nature.

ough the kindness of Prof. Zalessky I have been able to examine p tographs of some of the original material and, as shown innbsp;�� 664, C, the supposed leaves are borne on a rachis-like axis,nbsp;possibly of some Pteridosperni, a circumstance which though notnbsp;P oving that they are leaflets of a compound frond, favours that

go martensis on the ground that these Permian specimens do lot afford satisfactory evidence of close affinity to the genusnbsp;ffo. The two sets of leaves are probably specifically identical.nbsp;^ephropsis integerrima (Schmalhausen).

Permian of the Lower Tunguska on which Schmalhausen founded his species Ginkgo integerrima^the characteristic features are:nbsp;the transversely elongate form of the entire lamina^ the spreading,nbsp;dichotomously branched, veins and the short and relatively broadnbsp;stalk-like basal portion of the lamina. As Zeiller^ says, thisnbsp;species is probably identical with Ginkgo martensis^ from Permiannbsp;beds near Toulon-sur-Arroux (Saone-et-Loire) founded on a singlenbsp;leaf 3 cm. broad and 1-6 cm. deep. Zeiller reproduces Renault�snbsp;figure in his volume on the flora of Blanzy and Creusot* and statesnbsp;that the specimen could not be found: he retains the name Ginkgo

-Pio. 664. C, Ginkgopsis Czekanowskii. A, B, Nephropsis integerrima. (From the originals of Sohmalhausen�s figures, supplied by Prof. Zalessky; nat. size.)

but with the addition of a question-mark. The Permian leaves differ from those of Ginkgo in the absence of marginal veins�thoughnbsp;this is of secondary importance�in the absence of a true petiole,nbsp;in the form of the lamina and, from the majority of recent Ginkgonbsp;leaves, in the entire margin. The supposed leaves may be leafletsnbsp;of a compound frond or possibly bracts from some fertile shoot.

This generic name^ was instituted by Schimpei^ for large ^u�ieate leaves (figs. 665�667) from Upper Carboniferous andnbsp;firinian rocks that had previously been included in Sternberg�snbsp;genus NoRggerathia: the type-species is NoeggeratUa flabellafanbsp;Hutt.� Psygmophyllum is thus defined by Schimper:nbsp;n pinnatisecta, pinnis erecto-patentibus, e basi valde angustatanbsp;eiiiforniis, longitudinaliter flabellatim plicatis^ plus minusnbsp;prof unde pinna tisectis vel margine lobatis seu crenatis; nervisnbsp;1 ries dichotomis, ereeto-radiantibus.� Among other speciesnbsp;annbsp;nbsp;nbsp;nbsp;Schimper to Psygmophyllum are Noeggerathia expansa

iV. cuneifolia from Permian beds in the Ural mountains ^gured by Kutorga*^ Brongniart and later authors. In 1878nbsp;published a note giving the results of an examinationnbsp;rongniart�s specimens: he expressed the opinioir that thesenbsp;species are portions of compound fronds comparable with Ereyno-

ich he regarded as a shoot bearing simple leaves. He also Webnbsp;nbsp;nbsp;nbsp;Ural fossils the leaflets have a more or less

defined midrib in contrast to the regular flabelliform dicho-''^^uation in the leaves of the English species, P. fiahelli-^6 � the latter he assigns to the genus Ginkgophyllum previously fished for a Permian species, G. Grasserti^, consisting of annbsp;^18 bearing spirally disposed cuneate and deeply divided leavesnbsp;��^ch'nbsp;nbsp;nbsp;nbsp;similar to some of the older Baiera leaves; while

ifiiper�s genus Psygmophyllum, is applied to the two Russian tl �nbsp;nbsp;nbsp;nbsp;^ expansum and P. cuneifoUum. Saporta included a

species in Psygmophyllum, P. santagoulourensis. An ifiation of some specimens of Brongniart�s Noeggerathianbsp;ifi the British Museum leads me so far to agree withnbsp;porta in the opinion that some of the specimens referrednbsp;Sch^^^^ �P6cies are generically distinct from P. flabellatum.nbsp;uialhausen�s figures of the Ural species, which he refers to

- .also Arber (12) p. 401,

Psygmophylluni, show a considerable variation in the venation of the leaves or leaflets: those represented in his Plate iii. figs. 8�10^nbsp;agree with P. flabellatum while others difier in the presence of anbsp;midrib {e.g. Schmalhausen�s PI. iv. fig. 3). There has been considerable confusion in regard to the determination of these Russiannbsp;specimens; as Arber says, Kutorga�s Gydopteris gigantea'^ is.nbsp;probably a true Psygmophyllum though other specimens subsequently referred to P. expansum should not be included in thatnbsp;genus. Zalessky� in 1912 proposed a new generic name Palamo-phyllum for the Russian species but retained Psygmophyllum fornbsp;a Mongolian Permian specimen which he named P. mongolicum:nbsp;this fossil is clearly a portion of a compound froncf with leafletsnbsp;like those of some forms of Palaeopteris. In a later paper* thisnbsp;author assigns Psygmophyllum mongolicum to Palamophyllum, butnbsp;on his attention being called by Zeiller to Saporta�s note of 1878,nbsp;Zalessky� decided to abandon his proposed genus Palamophyllumnbsp;in favour of Psygmophyllum.

Confusion has also been caused by lack of uniformity in the use of the two generic names Psygmophyllum and Ginkgophyllum.nbsp;The type-species of these genera I believe to be generically identical;nbsp;they agree in the general form of the leaves, the lamina being muchnbsp;more deeply divided in the type-species of Ginkgophyllum, also innbsp;the decurrence of the narrow basal portion of the lamina, and bothnbsp;are probably shoots, though the morphological nature of these andnbsp;other types included in Psygmophyllum is by no means clear.

Arber retains both names; as a matter of convenience he restricts Psygmophyllum to leaves that are entire or only slightlynbsp;lobed, e.g. P. flabellatum (fig. 665), and the more deeply dissectednbsp;leaves such as those of Psygmophyllum Grasserti (fig. 669) he refersnbsp;to Ginkgophyllum. This distinction is, however, purely arbitrarynbsp;and on the analogy of the leaves of Ginkgo biloba it would seemnbsp;preferable to include both deeply divided and more or less entirenbsp;leaves in the same genus. Cambier� and Renier prefer the namenbsp;Psygynophyllum to Ginkgophyllum on the ground that in the leaves

^ This decision was communicated in a letter (October, 1913). A. C. 8. � Cambier and Renier (10).

*^S8igned by some authors to GinkgophyUum there are no marginal ^eins Jijjg those in the lamina of a Ginkgo leaf. This objection,

^ugh not in itself fundamental, is based on a sound principle, flanaely an objection to the assumption of affinity to Ginkgo impliednbsp;y GinkgophyUum, an assumption that rests on a superficialnbsp;^�8emblance unsupported by any evidence of real value.

The name Psygniophylliim is adopted both for entire and deeply ivided leaves of larger dimensions than the similar leaflets ofnbsp;fon s included in such genera as Palaeopteris and Adiantites..nbsp;peciniens usually occur as detached leaves, but when the leavesnbsp;6 attached to an axis the lamina is usually contracted into anbsp;decurrent, basal portion; there is no true petiole.nbsp;� t-nbsp;nbsp;nbsp;nbsp;spread from the base of the lamina and are repeatedly

� ) they may he very numerous and in some forms occasion-p ^ ^^^^stonaose, as in P. flabellatum., or much farther apart, as in W�jMs Arb. and P. Brownii (Daws.). The genus is at best anbsp;'7 artificial one; we know nothing as to the reproductivenbsp;�S or the anatomical structure, nor is it possible to determinenbsp;fr instances whether the specimens are portions of compoundnbsp;nn s or shoots bearing simple leaves.

haetic leaves named Psygniophylliim ? erenatum (Nath.) are P ed as evidence, the vertical distribution must be extended.

^ ergen, and North America; it is also a member of the ��nio-Carboniferous floras of South Africa and India.

iiame Noeggerathia was given by Lindley and Hutton^ to exa ^nbsp;nbsp;nbsp;nbsp;from the Newcastle Coal Measures; of the two

se P es figured one consists of a slender supporting axis bearing euneate leaves, which they speak of as part of a cona-by*!*nbsp;nbsp;nbsp;nbsp;other is a single leaf or leaflet. It is stated

Lebour and Dr Arber� that the original specimens cannot Mrnbsp;nbsp;nbsp;nbsp;Hutton collection in the Newcastle Museum. In

^ wse 8 Catalogue^ a specimen is described as possibly the giiia of the larger example, and some years ago I examined a

leaf in the Hutton collection which was believed to be the specimen reproduced on Plate 29 of the Fossil Flora. Fig. 665 representsnbsp;an impression in the British Museum^ from the Newcastle Coalnbsp;Measures; the cuneate and partially torn lamina^ 15 cm. long^ isnbsp;characterised by the very large number of forked veins, approximately three to a breadth of 1 mm. In this and another specimen^

it can be seen that the veins occasionally anastomose, but this feature is more clearly shown in some better specimens in thenbsp;Sunderland Museum�. One of these is represented half naturalnbsp;size in fig. 666. The leaves have no true petiole and are attachednbsp;to the axis by narrow decurrent bases. A careful examination ofnbsp;the specimen did not enable me to detect any satisfactory evidencenbsp;^ No. 40578.nbsp;nbsp;nbsp;nbsp;� No. 38927.nbsp;nbsp;nbsp;nbsp;� For other figures, see Arber (12).

the sheathing nature of the leaf-bases described by Arber, but Was on this example that I noticed the anastomosing of thenbsp;^eins (hg_nbsp;nbsp;nbsp;nbsp;^^ character not hitherto recorded in the

genusi. The lamina is torn and it is difficult to determine the 8ree of original lobing. Dr Aiber believes that such specimensnbsp;s that reproduced in fig. 666 are portions of a herbaceous plant

Pio. 666. Psygmophyllum fiahellatum. A�C, anastomosing veins (Prom a specimen in the Sunderland Museum; J nat. size.)

quite distinct from Ginkgo. While agreeing with the latter statement I do not regard the shoot-nature as definitely de-�aonstrated, though that is the interpretation usually acceptednbsp;*^iid not improbably correct. The habit appears to be identical

with that of the Permian species named by Saporta Ginlcgophyllum Grasserti (fig. 669) and his figure^ of that type suggests a shootnbsp;rather than part of a compound frond. Some species referred bynbsp;authors to Psygmophyllum are certainly pinnae, while others bearnbsp;a closer resemblance to shoots with simple leaves. Until betternbsp;material is available we cannot determine either the morphologicalnbsp;nature or the systematic position of the various examples assignednbsp;to this provisional genus.

Nathorst^ has recently founded this species, naming it after Dr Kolderup, on specimens from Devonian strata in West Norwaynbsp;� consisting of pieces of shoots, or possibly compound fronds,nbsp;bearing spirally disposed fan-like leaves or leaflets on long stalksnbsp;and reaching a breadth of 15 to 30 mm.; the veins are fine andnbsp;repeatedly forked. The habit appears to be similar to that ofnbsp;Psygmophyllum fiahellatum, but it is hardly possible to say whethernbsp;we are dealing with fragments of a frond or branches bearingnbsp;simple leaves.

This species, very similar to P. fiahellatum, is founded on specimens discovered by Mr Leslie in the Permo-Carboniferousnbsp;rocks at Vereeniging, South Africa^. The cuneate leaves reach anbsp;length of 13 cm. and in some cases are deeply divided into twonbsp;truncate lobes (fig. 667). The veins appear to be identical withnbsp;those of P. fiahellatum though no definite anastomosing has beennbsp;detected. A photograph recently received of a new specimennbsp;shows some indication of a few cross veins, but the occasionalnbsp;anastomosing of veins should not be regarded as a feature of greatnbsp;importance. The axis of this species is broader than any so farnbsp;found in the case of the English species, and the leaves are attachednbsp;by a similar decurrent base. Incomplete leaves similar to P-Kidstoni though probably not specifically identical are describednbsp;by Dun* from Permo-Carboniferous strata, at Sydney as Rhipi'

P^yg�mophyllum WilUamsoni Natborst.

A species^ founded on imperfect leaves from the Upper Devonian of Spitzbergen agreeing closely in shape and venation with P.nbsp;fflahellatum. This is the oldest European example of the genus.nbsp;In answer to an enquiry with regard to the venation Prof. Nathorstnbsp;kindly informed me that he was unable to detect any definitenbsp;traces of anastomosis and that the veins agree in their spacingnbsp;quot;quot;ith those of P. jlabellatum.

, ^ --- ---- ^

i^racterised by broad and shallow lobes: as in P. WilUamsoni axis occurs with the leaves. As Arber points out, this species

^ Nathorst (94) A. PI. ii. figs. 1, 2; Arber (12) PL XLii. fig. 4.

^ Arber (12) p. 392, Pis. xlu.�xliv.

The large flabellate leaves^, often more than 16 cm. long and 15 cm. broad, representing this Lower Carboniferous or Uppernbsp;Devonian species from Newfoundland, are distinguished fromnbsp;fldhellatum by their broader and less kite-like lamina and bynbsp;coarser venation. The distal margin is almost entire or

bears some resemblance to Psygmophyllum Brownii originally described by Dawson^ as Cyclopteris Brownii from Upper Devoniannbsp;strata in Maine.

In 190'5 some incomplete specimens were described from Permo-Carboniferous rocks in Kashmir as Psygmophyllum sp.^nbsp;for which^ on the discovery of better material, a specific name wasnbsp;proposed�. The leaves reach a length of 13 cm. and are characterised by the division of the lamina into six or more obcuneatenbsp;segments, the divisions sometimes extending to the base of thenbsp;broad part of the leaf (fig. 668). In the upper part of the laminanbsp;there are three to four veins per millimetre but lower in thenbsp;lamina the veins are 1 mm. apart. Dr Arber * suggests that thisnbsp;species should rather be referred to Ginkgophyllum or Rhipidopsis.

This less satisfactory species is represented by some imperfect leaves from Carboniferous rocks of Kashmir�. The lamina is

divided by a deep median sinus into two bilobed segments and 8'grees closely with some species of Baiera. By some authors thisnbsp;Species would be included in GinJcgophylluni but, as already statednbsp;in the account of the genus, the degree of dissection of the leavesnbsp;is too variable and unimportant a character to be made the basisnbsp;a generic differentiation.

This Permian species from Lod�ve^, France, agrees closely with fldhellatum in the size and outline of the leaves as also in their

is figured by Saporta as nbsp;nbsp;nbsp;fcamensfeinnum .nbsp;nbsp;nbsp;nbsp;rocks

in Tasmania which he compares with G. Grasserti Sap., but as he gives no illustration of the fossil no opinion can be formed of itsnbsp;true nature.

A species^ recently described from the Upper Devonian grits of Kiltorcan, Ireland, characterised by fan-shaped leaves 7 cm.nbsp;long and 5 cm. broad, deeply divided into two symmetrical halvesnbsp;each of which is again divided into two ribbon-like segments withnbsp;a Ginkgo-Y\kQ- venation. The leaves agree closely with those ofnbsp;P. Grasserti in their general form and in the lobing of the lamina.nbsp;Johnson believes this type to be an ancestral form of Ginkgonbsp;though there are no adequate grounds for such a view.

The close resemblance presented by the Rhaetic leaves from near Braunschweig, originally described by Brauns^ as Cychpterisnbsp;crenata and subsequently figured by Nathorst� as Ginkgo ? crenata,nbsp;to some Permo-Carboniferous leaves included in Psygmophyllumnbsp;suggests affinity with that genus. The obovate lamina, approximately 12 cm. long, is slightly lobed on the upper margin andnbsp;contracted below into a stalk-like base; the forked veins arenbsp;nearly 3 mm. apart. Nathorst compares the species with Psygmophyllum Jlabellatum Lind, and Hutt., but the Rhaetic specimensnbsp;differ from the English type in their much coarser venation: innbsp;the lobing of the lamina and in the coarse venation there is a muchnbsp;greater similarity to the broader leaves described by Arber^ asnbsp;Psygmophyllum majus. In view of the incomplete nature of thenbsp;material it is inadvisable to adopt the name Psygmophyllumnbsp;without reservation. An examination of Nathorst�s specimen innbsp;the Stockholm Museum led me to regard it as more probably annbsp;example of Psygmophyllum than of Ginkgo.

^ Johnson (14). Since thi,? account was written Prof. Johnson has described additional material including stems and foliage [Johnson (17)] demonstrating thenbsp;occurrence of repeatedly forked filamentous leaves [or leaflets] attached to slendernbsp;axes hearing also the broader form of lamina. The plant, which he now refers tonbsp;a new genus Kiltorkensia, may well be a Pteridosperm with compound fronds andnbsp;dimorphic pinnules.

Psygmophyllum frimigenium (�apoita). Some leaves discovered by Grand�Eury in Permian rocks of the Urals were described by Saporta^ as Salishuria priniigenia and regarded asnbsp;tile prototype of the surviving species. The original specimensnbsp;are unfortunately not available, but from the published figuresnbsp;it would seem that the species is of the same general type as

Reference has already been made to an American Devonian species referred by Dawson to Cyclopteris and recently transferrenbsp;to Psygmophylluni. A leaf or leaflet described by Dawson fromnbsp;Rilboa, New York, as Noeggerathia gilhoensis^ affords a goodnbsp;example of a specimen that may be a Psygmophyllum leaf or anbsp;leaflet of a frond of the Noeggerathia type. LesquerenxS considersnbsp;liawson�s specimen to be a pinnule of PaUeopteris. The samenbsp;remark applies to Lesquereux�s species Noeggerathia ohtusa^ fromnbsp;the Coal Measures of Pennsylvania included by Arber in Psygmo-

The Russian species P. expansum and P. cuneifoliuni are discussed in the account of the genus. The species Psygmophyllum Pelvali^ Camb. and Ren. from the Westphalian of Belgium is nownbsp;admitted to be a leaf of Cordaites. A species described by Sand-berger� from the Permian of the Black Forest as Ginkgophyllumnbsp;^irtus has been assigned by SterzeU to Dicranophylluin. A leafnbsp;%ured by Schmalhausen* from the Permian of East Russia asnbsp;Paiera gigas is no doubt a Psygmophyllum allied to P. Kidstoni.

fragment figured by Schenk from China as Ginkgophyllum sp. IS too imperfect to determine, and the specimens from the samenbsp;Reality described as Psygmophyllum angustilobum^ are, as Zeillernbsp;points out, pinnules of a frond of the Eremopteris type.

A sufflcient number of examples have been described to illustrate the range of the genus and the unsatisfactory nature of

the material from a botanical point of view. Failing reproductive organs or petrified specimens some useful evidence might benbsp;afforded by an examination of the cuticular structure of wellnbsp;preserved leaves.

Schmalhausen^ instituted this genus for large petiolate oval leaves from the Permian rocks of the Petschora district, characterised by the division of the lamina into several obcuneate ornbsp;obovate segments closely resembling in their form and venationnbsp;some forms of Psygmophyllum especially P. Haydeni^. We havenbsp;no definite information as to the systematic position of the parent-plant; the genus has usually been regarded as a representativenbsp;of the Ginkgoales on the ground of similarity in the leaves, butnbsp;while admitting that a relationship between Rhipidopsis andnbsp;Ginkgo is not improbable it is the safer course to regard Rhipidopsisnbsp;as a genus of Gymnosperms of uncertain affinity. Schmalhausennbsp;attributes to Rhipidopsis some Samaropsis seeds� found in association with the leaves, and Kurtz^ states that he has found leavesnbsp;and �fruits� in the Argentine. No proof of any connexion betweennbsp;leaves and seeds has so far been discovered. The genus is recordednbsp;from Russia, South America, and India from strata that arenbsp;Permian or approximately Permian in age.

Rhipidopsis (fig. 670) is distinguished from Psygmophyllum by the presence of a petiole and from most forms of that genus by thenbsp;deeper dissection of the lamina, as also by a more pronouncednbsp;difference in form and size betiveen the several segments of thenbsp;lamina. Zeiller� has drawn attention to a close resemblancenbsp;between Rhipidopsis and a specimen figured by Schmalhausennbsp;from the Artinsk Permian beds as Psygmophyllum expansum^.

The type-species (fig. 670) is characterised by the large size of the leaves which, according to Schmalhausen�, may reach a

length including the petiole of 30 cm. and a breadth of 11 cm^ quot;file segments, 6�10 in number, are often free to the summit onbsp;petiole; they vary considerably in shape and size, the mediannbsp;^^gments are obcuneate with a broad rounded truncate margin,

reaches a length of 10 cm. This species has been recorded from the Argentine but no figures have been published. It is notnbsp;improbable that the seeds of the Samaropsis type associated withnbsp;the leaves in the Russian and Argentine localities may belong tonbsp;this genus, but proof is lacking.

This Indian species from the Raniganj group of the Damuda series^ (Lower Gondwana) is founded on some leaf-impressionsnbsp;very similar in size and form to Rhipidopsis ginkgoides. Thenbsp;presence of a petiole is shown on one of the figured specimens: thenbsp;lamina is deeply divided into about six obcuneate segments thatnbsp;appear to be irregularly lobed on the truncate margin. Rhipidopsis densinervis is distinguished by its dense venation and bynbsp;a difference in size between the lateral and median segments lessnbsp;than in the leaves of R. ginkgoides. Kurtz^ states that somenbsp;specimens found by him in Permo-Carboniferous beds in thenbsp;Argentine may belong to this species.

The specimens for which this name is proposed were described by Feistmantel as Rhipidopsis ginkgoides^ from the Barakar groupnbsp;of the Damuda series. My examination of the type-specimensnbsp;confirms Feistmantel�s statement that they agree closely withnbsp;Schmalhausen�s Russian leaves except in their much smaller size:nbsp;the Indian leaves reach a length of 3 cm. while in Schmalhausen�snbsp;species the lamina may be 14 cm. in length. In view of thisnbsp;difference and the wide geographical separation of the two formsnbsp;it would seem preferable to adopt a distinctive name. Thenbsp;lamina is divided, almost to the base, into 6�10 segments; thenbsp;larger are cuneate and the smaller obovate and obtuse.

Fontaine and White^ instituted this generic term for some incomplete impressions of large leaves from Permian rocks in Virginia having a broadly cunea te or suborbicular lamina characterised by

a thickened lower margin extending horizontally a short distance Oil either, side of the petiole and presenting the appearance ofnbsp;feeing formed by the bifurcation of the summit of the leaf-stalknbsp;a t right-angles to its long axis. The lamina is irregularly dissected,nbsp;feut from the published figures it is difficult to distinguish betweennbsp;original lobing and divisions due to tearing. The dichotomouslynbsp;feranched veins spread through the lamina from the centre of thenbsp;fease and are given off at a wide angle from the thick lower edge ofnbsp;the lamina. In Saportaea grandifolia the petiole has a length ofnbsp;10 cm. and the rest of the incomplete leaf is 8 cm. long and 9-5 cm.nbsp;feroad; the second species S. salisburioides is represented by portions of similar but smaller leaves with a slender petiole. Whilenbsp;comparing these fossils with Ginkgo the authors of the genus callnbsp;attention to the peculiar features of the lower edge of the laminanbsp;and of the venation. The general resemblance in leaf-form betweennbsp;^^poTlaea and Ginkgo is hardly sufficient to warrant any definitenbsp;statement as to relationship and this Permian genus must for thenbsp;Piesent be relegated to the class of Plantae incertae sedis.

This genus was first described by Grand�Eury^ who, before the publication of the full description of the type-species, suggestednbsp;tfee substitution of Eotaxites for Dicranophgllum^ the name finallynbsp;adopted�. The genus is fairly abundant in the Upper Carboniferousnbsp;rocks of France and occurs also in Portugal, Belgium, and Germany,nbsp;it has recently been recorded from England and is represented innbsp;the Coal Measures of the United States and Canada. It occurs innbsp;Permian strata in Germany but with a few exceptions the genusnbsp;IS characteristic of Stephanian beds.

The systematic position of Dicranophyllum is far from settled; fey many authors it is considered to be a member of the Ginkgoalesnbsp;und is compared also with the Taxeae. In all probability thenbsp;genus is allied to the Cordaitales, though, as stated in the case ofnbsp;Trichopitys, it cannot be assigned to a definite position in thenbsp;Gymnosperms until we possess, fuller information with regard tonbsp;the reproductive organs or the anatomical structure.

In habit Dicranophyllum resembles Lepidodendron; it is an arborescent plant sparsely and irregularly though sometimesnbsp;dichotomously branched; the leaves are crowded and spirallynbsp;disposed, in some species persistent�in the sense in which thenbsp;leaves of Araucaria are persistent�while in others they probablynbsp;fell at an earlier stage. The leaves (fig. 671) exhibit a wide rangenbsp;in size, in the amount of lobing and the angle of divergence of thenbsp;segments; there is no differentiation into a lamina and petiolenbsp;nor are there any short foliage-shoots as in Ginkgo) the wholenbsp;leaf is represented by a narrow lamina, in some species almostnbsp;spinous, which consists in the basal portion of a simple linearnbsp;�stalk� reaching in extreme cases a breadth of about 7mm.,nbsp;attached by a decurrent base which persists as an elongated cushionnbsp;closely resembling the leaf-base on some Lycopodiums or thenbsp;projecting cushions of Picea {cf. fig.'140, Vol. ii. p. 94). Thenbsp;cushions are contiguous and cover the surface of a branch as innbsp;Lepidodendron, but they are distinguished by the occurrence ofnbsp;the leaf-scar at the apex of the cushion in contrast to its sub-apical position in Lepidodendron. The typical form of the leaf-base is shown in fig. 671., A, but in Dicranophyllum Beneckianuninbsp;Sterz. the transversely elongated leaf-scars are almost contiguousnbsp;as in some species of Sigillaria. At a distance from the basenbsp;varying in different species the lamina is divided into two, generallvnbsp;equal, branches that diverge at an acute or wide angle, and innbsp;most species each arm undergoes one or more bifurcations in anbsp;single plane. The whole leaf may reach a length of over 20 cm.nbsp;In the basal portion of the lamina there are two or more parallelnbsp;veins, but in branches in which the leaf-scars are well preservednbsp;there is only a single vascular-bundle scar indicating a singlenbsp;leaf-trace up to the base of the lamina. Each segment of thenbsp;leaf has usually two veins and the acutely pointed ultimatenbsp;segments have a median vein. The so-called secondary or interstitial veins are no doubt due to the presence of hypodermalnbsp;stereome strands. The narrower Dicranophyllum leaves are verynbsp;similar to the deeply divided pinnae of Macrozamia heteromeranbsp;(fig. 671; cf. fig. 396, F, Vol. iii. p. 26). The branching of the laminanbsp;is generally regular but in several instances the subdivision isnbsp;irregular (fig. 671, D). On young shoots the leaves may be

nearly vertical but in most species they become widely extended and on older branches may be reflexed as in some Lycopodsnbsp;(fig. 121, B, Vol. II. p. 35). There is some evidence that the pithnbsp;Was discoid as in Cordaites^. The microsporophylls are said tonbsp;fie borne in small ovoid strobili in the axils of foliage leaves, but

671. Dicranophyllum galUcum. A, piece of a stem showing leaf-cushions. B�I), foliage-shoot and leaves. (After Grand�Eury.)

the only evidence as to their structure so far adduced is furnished y an imperfectly preserved specimen described by Zeiller^nbsp;associated with a shoot of Dicranophyllum robustum but notnbsp;actually attached; this consists of a small axis expanded into

a radially segmented distal portion bearing some imperfectly preserved ovoid bodies on its lower face which are probablynbsp;microsporangia. No spores are recorded. Zeiller compares thisnbsp;sporangiophore with the microsporophyll of a Taxus. On somenbsp;stems small scale-covered buds occur immediately above thenbsp;attachment of a leaf; these are probably fertile shoots but wenbsp;have no definite information with regard to their structure. Somenbsp;specimens from Commentry^ demonstrate the occurrence of smallnbsp;oval ovules or seeds, 4x3 mm., along the length of ordinarynbsp;leaves (fig. 672), and seeds are sometimes found associated with thenbsp;basal portions of foliage leaves though not in organic connexionnbsp;with them, except in an example described by Eenault fromnbsp;Autun as Dicranophylluni galhcum var. Parchemineyi^. Somenbsp;leaves of D. striatum^ are described as enlarged at the base andnbsp;slightly concave as if to hold a seed, but if this supposition isnbsp;correct it involves the admission of two types of seed-bearingnbsp;organs within the genus. The more probable conclusion is thatnbsp;the seeds were borne along the length of the lamina of the sporo-phylls and on the expanded bases. �

Some specimens from Lower Cretaceous beds in Virginia described as Baiem foliosa* resemble Dicranophyllum, but in viewnbsp;of the vast chronological gap between these beds and those innbsp;which Dicranophyllum occurs it is unlikely that the similaritynbsp;has any significance.

This species, one of the two described by Grand�Eury in 1877�, is the commonest representative of the genus; it is characterisednbsp;by persistent leaves with a base that is unbranched for a distancenbsp;of 15�20 mm. and then bifurcates into two equal or approximatelynbsp;equal segments at an angle of about 30�; these reach a length ofnbsp;10�15 mm. and divide into two acute segments 8�10 mm. long.nbsp;There are three veins in the basal portion of the lamina, one ofnbsp;which branches below the dichotomy, and each of the divergentnbsp;arms has two veins. The leaf-cushions are 2�3 times as long as

- Renault (96) A. p. 375. nbsp;nbsp;nbsp;� Renault and Zeiller (88) A. p. 632.

^ Grand�Eury (77) A. p. 272, Pis. xiv., xxx. Zeiller (80) A. PI. lxxvi. figs. 1, 2.

of foliage leaves and bear numero ovules (fig. 672). This species is recordednbsp;from the coal-fields of the Loire, omnbsp;nientry. Gard, Brive^ and elsewhere; i_nbsp;occurs also in the Coal Measures onbsp;Portugal.nbsp;nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;,

Eury from Gard as D. tripaHituni^, w rc I had an opportunity of examining in t lenbsp;�cole des Mines, Paris, are not specifically distinguishable from D. galhcwn-

imperfect specimens described by W hr e nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;with

This species was first figured by Gomes� �^A^some fragments, the ground of the similarity of the lamina ^ r ^ndley andnbsp;probably of Lepidodendroid leaves, describe y 1 Pynbsp;Hutton� as Cyperites hicarinata and subsequennbsp;nbsp;nbsp;nbsp;peount of

Heer- in his geL Distrigophyllum. In a note to bis account Mesozoic plants from Portugal Heer* renames Plantnbsp;'P^llim lusitanicum and compares it with Dicranop ynbsp;nbsp;nbsp;nbsp;sneci-

Crand�Eury. He Lima� recognised the true nature o mens from the Stephanian of Portugal and pubisnbsp;^ Zeiller (92�) A. p. 96.nbsp;nbsp;nbsp;nbsp;^ Lima, de (88) Pis. i-, m-

2� 4 mm. broad at the base; the lamina is once forked and thenbsp;forks diverge at a very small angle as in D. longifolium Ken.nbsp;Exceptionally good specimens figured by Gomes and de Limanbsp;show numerous long leaves spreading radially in the matrix fromnbsp;a comparatively slender axis. In D. longifolium the leaves arenbsp;given off. at a much more acute angle.

This type! is similar to D. gallicum but the leaves are only preserved in their basal portions; the lamina is 5�6 mm. broadnbsp;and bifurcates at a distance of about 15 mm. from the base at annbsp;angle of 20��30�. Ovoid buds occur in the axils of some of thenbsp;leaves. It was in association with this specimen from the Gardnbsp;coal-field that Zeiller found the microsporophylls already described.nbsp;The surface shows particularly well-preserved large and slightlynbsp;depressed cushions 3�4 cm. long and 4�5 mm. broad.

In the form of the leaves this Permian species from Baden closely resembles D. gallicum] it is chiefly of interest because ofnbsp;the almost complete absence of leaf-cushion; the leaf-scars,nbsp;characterised by their acute lateral angles, are almost contiguousnbsp;as in some species of Sigillaria.

In this Belgian Westphalian species the leaves are dichotomously branched into two linear segments at an angle of about 60�; itnbsp;differs from D. gallicum in the single bifurcation of the lamina,nbsp;the wider angle of divergence, and in the feebler relief of thenbsp;leaf-cushions.

This, the only British species, has recently been described from the Westphalian beds of Staffordshire'*. The crowded leaves,

3- 50 cm. long, are dichotomously branched three or four times intonbsp;slightly spreading linear rigid segments with a maximum breadth

XLII] DICRANOPHYLLUMnbsp;99

1�25 mm. The undivided portion of the lamina is about 7 mm.

Kidston speaks of the rhomboidal outline of the leaf and the repeated dichotomy of the lamina as distinguishing features.

Some fragments of forked leaves are figured by Schenkfrom the Coal Measures of China as Dicranophyllum latum, but thenbsp;material is too meagre for accurate determination. It is noteworthy that the broader type of Dicranophyllum leaf may easilynbsp;0 confused with an impression of a branched Stigmariaii rootlet,nbsp;narrower specimens described by Schenk from China as

nis species like several others is founded on detached leaves�, a circumstance that has led some authors to draw a distinctionnbsp;lea^*^'*'^ species with persistent leaves and those with caducousnbsp;There is, however, no good reason for assuming that allnbsp;pecies were not of the evergreen type. The leaves of this speciesnbsp;t^e characterised by their great length which may be 24 cm.;

e amina is once or twice forked and is 5�6 mm. broad at the base, which contains 4-7 veins.

Ill this Commentry species* the leaves, which reach 14 cm. in of tgt;i ' characterised by the very small angle of the divergencenbsp;jj ^ segments, 3� as contrasted with a divergence of 30� innbsp;licum. The leaves are almost erect and twice bifurcate.

to Dicranophyllum Beneckianum Sterzel has de-^ second species, D. latifolium^, from the Lower Permian a en characterised by leaves similar to those of Dnbsp;cut generally longer,nbsp;a well-defined type.

of Pennsylvania, but neither is represented by very satisfactory specimens: Dicranophyllum dicliotomum Lesq. is founded on anbsp;dichotonaously branched shoot bearing long and narrow leaves innbsp;the apical region only and very similar in appearance to Lepido-dendron except in the branched lamina. The second speciesnbsp;B. dimorphum Lesq. is represented by leaves and branches, whichnbsp;however are not well preserved. The peculiar subdivision of thenbsp;apical portion of the laminae suggests a simple leaf with a frayednbsp;termination.

Under this name Dr Stopes^ has recently described a well-preserved leaf from the Westphalian series of New Brunswick. The specific name was first applied by Dawson^ to specimens whichnbsp;he referred doubtfully to Psilophyton. The type-specimen is 9 cm.nbsp;long and 3 mm. broad at the base and the lamina is repeatedlynbsp;branched. This specimen bears a close resemblance to the leafnbsp;from Autun described by Renault as Trichopitys Milleryensis^.

The imperfect specimens described by Dawson'^ from Devonian rocks in Queensland as Dicranophyllum australicum and subsequently figured by Jack and Etheridge� consist of a slendernbsp;axis, 3 mm. wide, with elongate leaf-bases bearing leaves 3 mm.nbsp;long with two widely divergent apical segments like those characteristic of the sporophylls of Gomphostrobus. The fragments havenbsp;no claim to be included in Dicranophyllum.

There has been confusion between Dicranophyllum and Gom-phostrobus^: as shown by drawings reproduced by Potoni� of specimens of Gomphostrobus from the Permian of Thuringia, therenbsp;is a close resemblance in habit between the two genera, but innbsp;Gomphostrobus the foliage-leaves are falcate and entire, while thenbsp;bifurcate sporophylls differ from the leaves of Dicranophyllum innbsp;their widely divergent and small apical fork.

� Schenk (90) A. erroneously includes Sigillariostrohus hifidus Geinitz, (73) Pl. III. figs. 5�7, in Dicranophyllum. See also Sterzel (93) A. p. 111.

A foliage-shoot described by Benaulti from Autun as Pvmtes perniiensis, though too imperfect to be identified, is worthy ofnbsp;notice as possibly an example of Dicranophyllum or Tnchopitys;nbsp;it consists of an axis 3 mm. in diameter bearing numerous spirallynbsp;disposed leaves 3 cm. long, barely 1 mm. broad and triangular innbsp;section, at an angle of 45�. The leaf-cushions are elongate andnbsp;slightly prominent. It is, however, impossible to decide whethernbsp;this fossil should be referred to the Lycopodiales or to the Gymno-sperms. There is no evidence that the leaves are attached tonbsp;short shoots and the use of the generic name Pinites cannot benbsp;Justified by any trustworthy test.

Saporta^ proposed the generic name Trichopitys in 1875 for some shoots from the Permian beds of Lod�ve bearing long, narrow,nbsp;and deeply divided leaves; he defined the genus as follows:nbsp;Polia verosimiliter rigida cartilagineaqne, dichotome partitanbsp;etiamque pedato-partita, petiolo plus minusve elongato, sursumnbsp;in lacinias 4�6, anguste lineares, uninerviasque dissecta^.� Manynbsp;palaeobotanists'have followed Saporta in regarding Trichopitysnbsp;as a member of the Ginkgoales, but the evidence in support of thisnbsp;view is by no means conclusive. The only species so far describednbsp;that affords any information with regard to the habit oi fertilenbsp;shoots of the plant is the type-species T. heteromorpha (fig. 673).nbsp;A fairly stout branched axis bears leaves varying considerably innbsp;size and form; they may be long and filiform, apparently rigid,nbsp;simple or deeply divided, or short and entire, and in some casesnbsp;resembling the leaves of certain smaller species of Baiera exceptnbsp;in the less regular forking of the lamina. In the axil of some foliagenbsp;leaves are short, simple or branched, axes bearing seed-like bodiesnbsp;u�iginally described as buds and afterwards regarded as seeds.nbsp;A specimen figured by Zeiller'* from Lod�ve (fig. 673) shows anbsp;branched axillary shoot bearing several small ovules comparable

Renault^ has figured a portion of a large leaf from Autun as Trichopitys Milleryensis which may belong to the closely alliednbsp;genus Dicranophyllum; it is 12 cm. long and 3 mm. broad at thenbsp;base; the narrow basal part of the lamina has three parallel veinsnbsp;and forks into two arms, each of which again branches intonbsp;divergent linear segments. The leaf is larger and broader thannbsp;the leaves of T. heteroniorpha and agrees very closely with those ofnbsp;some species referred to Dicranophyllum: the fact that the branching of the lamina is not absolutely regular cannot be accepted asnbsp;a constant difference between Dicranophyllum and Tricho-pitys:nbsp;the leaves shown in fig. 671, which were found attached to undoubted Dicranophyllum branches, are no more regular in the

forking of the segments than T. Milleryensis, and a leaf recently figured by Dr Stopes^ from New Brunswick as Dicranophyllumnbsp;glabrum (Daws.) bears a close resemblance to Renault�s figure. Innbsp;some cases the more regular dichotomy of the leaves is a characteristic of Dicranophyllum, but it is not a constant feature. Somenbsp;imperfectly preserved specimens figured by White� from the Coalnbsp;Measures of Missouri as Dicranophyllum sp. bear irregularlynbsp;branched leaves wRich are hardly distinguishable from some ofnbsp;those on Saporta�s type-specimen of Trichopitys heteromorpha.

In the present state of our knowledge it is impossible to give a satisfactory definition of the genus or to state precisely on whatnbsp;grounds it is separated from Dicranophyllum. In Tfichopitys,nbsp;as represented by T. heferomorjjha, the leaves are more variablenbsp;nr form than in Dicranophyllum and less regular in the subdivisionnbsp;the lamina; there ate no persistent leaf-bases like those ofnbsp;^ilt;^ranophyUum, but this is a character that could not be seen innbsp;imperfectly preserved or partially decorticated specimens. A morenbsp;important difference would seem to be that in TricJiojyitys thenbsp;seeds are borne on special axillary shoots, while in Dicranophyllumnbsp;I^ey occur on ordinary leaves. Such evidence as we have suggestsnbsp;that Trichopitys is a Gymnosperm possibly allied to the Cordaitaiesnbsp;mill Ginkgoales, but the facts hardly justify its inclusion in eithernbsp;group. Its affinity to Dicranophyllum cannot be definitely determined though in all probability the two genera are closely related

^1 tJie latter species he included the specimens doubtfully assigned y Lindley and Hutton to Solenites as Solenites ? furcata^. Thesenbsp;other Jurassic leaves that are referred by some authors tonbsp;^ickopitys are usually regarded as examples of Baiera^; there arenbsp;110 adequate grounds for believing them to be closely related to

Zeiller* records a fossil from Triassic beds in Madagascar that thinks may be an example of Trichopitys.

This generic name was proposed by Zeilier in place of Witharnia ich, in ignorance of its previous use, I employed for somenbsp;specimens from the Wealden rocks of Sussex. The inclusion ofnbsp;u second species, Sewardia lonyifolia, necessitates an extension ofnbsp;e definition of the genus to include spinous branches hearingnbsp;spirally disposed leaves or leaf-like organs, either orbicular andnbsp;entire or fan-shaped and deeply divided, in the axil of recurved

Sphenozam�es latifolius Schimper^, Trait�, Vol. ii. p. 163. CycadoracMs armata Saporta, Plant. Jurass. p. 196, PI. 117, fig. 1.nbsp;Withamia armata Seward, Wealden Flora, Vol. ii. p. 174, PI. ii.nbsp;figs. 1, 2; PI. V. fig. 1.

This species is represented by woody axes, about 1 cm. in breadth reaching a length of 50 cm., from the Wealden beds ofnbsp;Sussex, bearing more or less orbicular, entire, leaves or leaf-like

organs, 6 cm. or more long, in the axil of stout recurved spinous processes (fig. 674, A, B). The leaves are sessile and the venationnbsp;is of the Cyclopteris type. Spinous axes of the same form hadnbsp;previously been described by Saporta from Kimeridgian rocks innbsp;France as Cycadorachis armata, but these show no indication ofnbsp;leaves and were regarded as Cycadean. In a letter written to menbsp;in 1895, shortly before his death, the Marquis of Saporta^ suggested

the generic name Acanthozamite-s as a substitute for Cycadorachis in view of the new data afforded by the English material, but itnbsp;Seemed preferable to adopt some provisional name which did notnbsp;imply affinity to the Cycads. The leaflets described by Saportanbsp;ns Sphe^iozatnites latifolius are apparently identical with thosenbsp;found in the English beds, but none of the French specimens werenbsp;nttached to a supporting axis. The relation of spines and leavesnbsp;suggests that the latter may be phylloclades borne in the axil ofnbsp;modified spinous leaves, but their morphological nature cannotnbsp;he determined. In this connexion attention may be called tonbsp;I^ioncophyllum Tholloni Baill. a West African shrub which bearsnbsp;on the long shoots leaves 2�3| inches long each of which has anbsp;pair of strong revolute hooks at the apex: in the axils of thesenbsp;leaves are short shoots with larger leaves without hooks. It isnbsp;�^88ested^ that the apparent lamina of the hooked leaves is anbsp;^^mged petiole, the hooks representing lateral leaflets.

This species was described by Salfeld^ from the Solenhofen beds (Upper Jurassic) of Bavaria as Baiera ? longifolia Heer. itnbsp;Is founded on branches nearly 30 cm. long bearing large fanshaped deeply divided leaves, or leaf-like organs, in the axils ofnbsp;lecurved spines similar to those in S. latifolia. The leaves arenbsp;identical in habit with those of some species of Baiera, but wenbsp;have no information with regard to the structure of the epidermis.nbsp;In view of the uncertainty as to the morphological nature of thenbsp;leaves or their relationship to leaves of Baiera, it is inadvisablenbsp;fo adopt a generic title that implies affinity to the Ginkgoales.

CHAPTER XLIII.

The Coniferales, by far the largest section of the Gymnosperms, present considerable difficulty to the student of fossil plants.nbsp;There is great divergence of opinion with regard to the relativenbsp;antiquity of the several families, and their position in an evolutionary series. The Abietineae are by some botanists regardednbsp;as the most primitive; on the other hand, and this is the viewnbsp;that in my opinion receives most support from the availablenbsp;evidence, it is held that the Araucarineae are both the mostnbsp;primitive and the oldest representatives of the Coniferales. Untilnbsp;recent years the study of fossil Conifers has suffered neglect andnbsp;little help has been afforded by palaeobotanists to the solution ofnbsp;the morphology of the ovulate shoots of the different genera, a

problem that has long exercised the ingenuity of investigators. The view expressed by Jeffrey^ that recent work on fossil Conifersnbsp;corroborates the interpretation of the seed-bearing scales asnbsp;metamorphosed shoots is based on facts furnished by a study ofnbsp;vegetative organs, which in themselves do not afford any decisive

I�m. 675. A, B, Phyllocladus trichomanoides. C, P. hypophylla. D, E, Megastro-tUus and seed of Phyllocladus alpina; a, arillus. [A�C from specimens in the British Museum; H, E, after Miss Robertson (Mrs Arber.)]

evidence as to the morphology of reproductive shoots. In view of these considerations it is important that an attempt should benbsp;made, even at the risk of disproportionate treatment, to give anbsp;general account of recent genera which, though necessarily farnbsp;1 Jeffrey (10) p. 331.

from complete, may afford assistance to students prepared to undertake a critical study of the fragmentary records of the rocks.

Conifers are trees or shrubs exhibiting a fairly wide range in habit; the � great ones of the forest� such as the Sequoias (fig. 674*),nbsp;the sugar Pines {Pinus Lamhertiana) and Douglas Firs {Pseudotsuganbsp;Douglasii) of the Eocky Mountains, Taxodium mucronatum} ofnbsp;Mexico, remarkable for its enormous bulk, the tall and slendernbsp;Cypresses, the less formal Podocarps of the southern hemisphere,nbsp;the shrubby Junipers, the dwarf Dacrydium laxifolium^ of Newnbsp;Zealand afford examples of recent types. In most species thenbsp;leaves are small and crowded, not infrequently dimorphic, and innbsp;Phyllocladus (fig. 675) reduced to inconspicuous and caducousnbsp;scales subtending phylloclades. Agathis is exceptional in havingnbsp;narrow ovate leaves reaching a length of nearly 20 cm. (fig. 695)nbsp;and a similar but smaller leaf is characteristic of some species ofnbsp;Podocarpus (fig. 676). The presence of long and short shoots isnbsp;a striking feature of Pinus, Larix, Pseudolarix, Cedrus, andnbsp;Sciadopitys: the short shoot, as Goebel says � takes no part innbsp;the construction of the permanent skeleton of the tree�.� Thenbsp;whorled arrangement of leaves characteristic of several Cupres-sineae and the Callitrineae is not a constant feature and, as innbsp;Lycopodium, both whorled and spiral foliage may occur on thenbsp;same shoot.

Conifers are monoecious or dioecious; the microsporophylls and megasporophylls are borne spirally or in whorls on separate shoots,nbsp;and in some genera on separate trees, except in the case of abnormalnbsp;bisporangiate strobilP. Proliferous cones are not uncommon innbsp;some genera: the prolonged axis of the cone of Cryptomerianbsp;japonica shown in fig. 677 bears microstrobili in the axils of thenbsp;small leaves. The microstrobili are for the most part constructednbsp;on a uniform plan; they are usually short-lived, small shoots, andnbsp;each microsporophyll often consists of a slender axis bearing twonbsp;microsporangia on its lower surface and prolonged as a smallnbsp;upturned distal expansion. In Pinus the sporangia dehiscenbsp;longitudinally, while Abies (fig. 684, E) affords an example of

2 nbsp;nbsp;nbsp;Hooker, J. D. (52) PI. 815.nbsp;nbsp;nbsp;nbsp;� Goebel (05) p. 444.

* For examples see Sterzel (76); Eichler (82); Worsdell (04); Bartlett (13); Shaw, W. R. (96); Robertson (06); Renner (04); Bayer (08); etc.

transverse dehiscence. The microstrobili of Cedrus are similar but longer. In Torreya (fig. 684, D), Taxodiuni, Widdringtonia,

VIA

and some other genera there are 4�6 microsporangia on each sporophyll. In Araucaria and Agathis there may be as many as

10�20 sporangia, longer and relatively narrower than in other genera and attached by one end, in contrast to the more completenbsp;union of sporophyll and sporangiumnbsp;in Pinus. In Araucaria Muelleri andnbsp;A. Rulei (fig. 678) the microstrobilinbsp;reach a length of 25 cm.: in ^4. excelsanbsp;and A. Cookii they are much smallernbsp;(fig. 679, A, B). In Taxus 4�7 sporangia are radially disposed on thenbsp;inner face of a flat distal expansion.

Pseudolarix^, and Keteleeria are borne in umbels, while in Cryptomeria andnbsp;Taxodium^ they occur in spikes. Thenbsp;microspores may be winged or wingless : in the Abietineae there are as anbsp;rule two conspicuous wings or bladdersnbsp;(fig. 684, B), but the spores of Pseudo-tsuga are wingless and in Tsuga bothnbsp;types occur. In M.icrocachrys�*� thenbsp;wings vary from 2 to 6 (fig. 684, C)nbsp;and in Dacrydium^ and Podocarpus^

Torreya, Sciadopitys, the Cupressineae, and some other Conifers there are nonbsp;prothallus cells: the microspores of the Abietineae are characterised by the occurrence of 2, or occasionally 3 or 4�, evanescentnbsp;prothallus cells (fig. 684, B); in Dacrydium there are 4�6 prothallusnbsp;cells; in Microcachrys 3 or 4; in Podocarpus (fig. 684, A) as manynbsp;as 8, while in Araucaria 15 cells have been recorded and as manynbsp;as 30 nuclei. The two male gametes are non-motile.

The term Conifer though appropriate as regards the majority of the plants so styled is misleading in the case of several genera

� Jeffrey and Chrysler (07); see also Thibout (96); Burlingame (08); (13); (15); Sinnott (13).nbsp;nbsp;nbsp;nbsp;� Hutchinson (14).

which possess ovulate shoots differing widely from cones as the ferm is generally understood. The cones of Araucaria (figs. 680,nbsp;fiSl) and Agathis reach a considerable size; those of Araucarianbsp;SidwilUA, similar to some cones of Encephalartos, may be 28 cm.nbsp;m diameter and in some species of Agathis^ they exceed 11 cm.

� ps

10. 679. A, Araucaria Cookii var. luxuriant. B, A. Cookii. C, Araucaria Cookii, microstrobilus. D, E, Araucaria Muelleri, part of a microstrobilus and a singlenbsp;sporophyll; ps, microsporangia. F, Araucaria Montana, branch. (Afternbsp;Seward and Ford.)

^nd are 14 cm. long. The cone-scales of Agathis are flat, woody structures bearing a single ovule (fig. 682): in Araucaria thenbsp;single seed is embedded in the scale, and a more or less prominentnbsp;appendage, the ligule, forms a characteristic feature (fig. 683, 1).

The cone-scales of some species, e.g. A. Cookii, A. excelsa (fig. 683, A, D) are flat and laterallv winged, while in A. hrasiliensisnbsp;the thick distal ends closely resemble those of the seed-scales ofnbsp;some Pines: the cone-scale of A. imhricata is larger and deeper,nbsp;and that of A. Bidwillii broad and woody (fig. 683, B, C). Innbsp;Pinus, with cones reaching a length of 2 feet, the mature scalesnbsp;are apparently simple like those of the Araucarineae; the distalnbsp;end is broad and rounded (P. silvestris) with a central umbo or.

Fig. 682. A, Agathis Mooreiy cone-scales; s, scar of seed. B, Agathis lorantkifolia, section of ovule showing integument, nucellus, and megaspore with threenbsp;archegonia, also part of the cone-scale with a projection close to the base ofnbsp;the ovule. (After Seward and Ford.)

as in P. Coulteri, the umbo is prolonged as a strong recurved spine, while in P. excelsa (fig. 704) and P. Cembra the scales are flatternbsp;like those of Picea. In the young Pine cone each scale is clearlynbsp;a double structure consisting of a lower portion, the bract ornbsp;carpellary scale, and an upper portion, the ovuliferous scale,nbsp;bearing two ovules. In the course of development the seminiferousnbsp;scale alone increases in size, and the bract-scale is hardly visiblenbsp;in the ripe cone or is represented by a small remnant. In Abies,nbsp;Larix, Pseudotsuga the dual nature of the scales is obvious at

maturity, the bract-scale usually extending beyond the edge of the seminiferous scale (fig. 705), in Abies bracteata reaching anbsp;length of 5 cm. In Sequoia (fig. 702 B) the cone-scales show nonbsp;outward sign of a double structure, but each scale contains two

Fig. 683. Cone-soales of Araucaria Coolcii (A), A. imbricata (B), A. BidwiUii (C), and A. excelsa (D, E); I, ligule. (After Seward and Ford.)

sets of bundles, the lower being normally orientated as in a leaf and the upper vascular strands inversely orientated {cf. fig. 684, R).nbsp;The occurrence of these two sets of bundles is often quoted innbsp;support of the view that the double cone-scale of the Abietineae

represents the lowest term of a series, the upper end of which is represented by the scales of Agathis which have lost all externalnbsp;signs of their supposed dual nature and retain only the inverselynbsp;orientated bundles as evidence of their descent from an ancestralnbsp;type in which the ovuliferous and bract-scales were separatenbsp;organs^. The scales of such genera as Sciadopitys, Athrotaxis,nbsp;Cryptomeria (fig. 684, S, N, M), on this hypothesis, occupy annbsp;intermediate position. The seminiferous scale of the Abietineaenbsp;is considered by many botanists to be a leaf or leaf-like organnbsp;borne on an axillary shoot subtended by a bract, and it is believednbsp;that the simple scale of Agathis has been produced by thenbsp;gradual fusion of two originally distinct organs. The ligule ofnbsp;Araucaria is held to be the outward and visible sign of the seminiferous scale that has almost lost its individuality, and with thisnbsp;ligular relic are homologised the upper half of the scale of Sequoia,nbsp;the deeply toothed upper part of the scale of Cryptomeria (fig.nbsp;684, M), the rounded ridge on the abaxial side of the seeds innbsp;Athrotaxis (fig. 684, N), the membranous outgrowth on the scalesnbsp;of Cunninghaniia (fig. 684, K, m), and the seminiferous scale of thenbsp;Abietineae. It has been pointed out in support of this hypothesisnbsp;that two vascular bundles are given off from the axis of a Pine cone,nbsp;one of which forms the bract-scale bundles and the other thenbsp;vascular supply of the seminiferous scale^.

In a recently published paper on the vascular anatomy of the Hiegasporophylls of Conifers by Miss Aase� additional facts arenbsp;given with regard to the origin and behaviour of the vascularnbsp;bundles of the cone-scales. In the upper part of a cone of Pinusnbsp;niaritima the bract-supply arises as a single bundle at the base ofnbsp;a gap in the stele, and the bundles of the seminiferous scale arenbsp;given off from the sides of the gap above the point of origin of thenbsp;bract-bundle: in the lowest sporophylls, on the other hand, thenbsp;bract and scale-bundles have a common origin. A separate originnbsp;for bract and seminiferous scale-bundles is recorded in severalnbsp;Abietineae and in some other Conifers. The origin of the vascular

^ For references to literature on the morphology of cones, see Coulter and Chamberlain (10); Worsdell (04); Eendle(04); Lotsy(ll); also Celakovsky (82);nbsp;Kramer (85); Bayer (08); Aase (15).

supply of the double cone-scale is^ however^ not constant even in the same cone, and in Araucaria BidmilliA each cone-scale isnbsp;supplied by two separate strands from the vascular axis thoughnbsp;in other species a single bundle enters the cone-scale and dividesnbsp;later. It would appear, therefore, that the single or double originnbsp;of the lower normally orientated bundles and of the upper set ofnbsp;inversely orientated bundles is far from constant, and the datanbsp;derived from anatomical study do not afford a satisfactory meansnbsp;of determining the morphological nature of the cone-scales. It isnbsp;held by Jeffrey and his school that the Abietineae represent thenbsp;oldest members of the Coniferales and that the Araucarineae arenbsp;a more recent development, the apparently single cone-scale ofnbsp;Araucaria and Agathis being derived from the double cone-scalesnbsp;of the Abietineae. Some botanists, e.g. Vierhapper^, whilenbsp;believing that the seminiferous scale of the Abietineae is an organnbsp;belonging to an axillary shoot subtended by a bract-scale and thatnbsp;the cone-scales of other Conifers are also double structures, whethernbsp;or not externally divided, regard the Araucarineae as earlier innbsp;origin than the Abietineae. If the cone-scales of the Araucarineae,nbsp;to take the extreme type, are in origin double and homologousnbsp;with the obviously double cone-scales of the Abietineae it is morenbsp;logical to regard the Abietineae as the precursors of the Araucarineae. The evidence afforded by fossils in my opinion lends strongnbsp;support to the greater antiquity of the Araucarineae, and I venturenbsp;to believe that no adequate reasons have been given for regardingnbsp;the cone-scales of the Araucarineae as other than simple leavesnbsp;bearing ovules. If, as seems probable, the Coniferales are mono-phyletic in origin the cone-scales of the different families are innbsp;all probability variants of a common type and, in opposition tonbsp;the view which is most in favour, I regard the double cone-scalesnbsp;of the Abietineae and the corresponding organs of other Conifersnbsp;which afford evidence of a double structure as derivatives of anbsp;simple form of sporophyll strictly comparable with the sporophyllnbsp;of a Lycopodium, the placental outgrowth assuming an increasingnbsp;degree of individuality in the different lines of evolution illustratednbsp;by various types of strobilus. It is noteworthy that the transitionnbsp;from foliage leaves to megasporophylls in the Araucarineae isnbsp;1 Worsdell (99).nbsp;nbsp;nbsp;nbsp;^ Vierhapper (10).

often very gradual in contrast to the much more sharply defined break between leaves and scales in most of the Abietineae andnbsp;many other Conifers. The cone-scales of the Abietineae arenbsp;recognised as more complex and more recent developments, thenbsp;seminiferous scale being an excessively enlarged placental outgrowth from a megasporophyll, while in the intermediate typesnbsp;such as Seqmia, Cryptomeria, and others the separation betweennbsp;the two parts of the cone-scale is much less completed Thisnbsp;morphological question is too complex to discuss fully in anbsp;general summary: students should, however, be warned thatnbsp;several botanists do not agree with the opinion that is herenbsp;expressed. It is at least fair to add that the views expressednbsp;by Prof. Jeffrey and his pupils with regard to the relative positionsnbsp;of the Araucarineae and the Abietineae in an evolutionary seriesnbsp;are stated with an assurance which is misleading to those unfamiliarnbsp;with the nature of the evidence^.

In Saxegothaea (fig. 685), Dacrydium (fig. 684, P), and some other genera each ovule is surrounded by a cup-like integument (fig. 684,nbsp;T a), formerly called the arillus but recently styled the epimatium�;nbsp;this is by some authors considered to be the equivalent of thenbsp;seminiferous scale.

This inadequate account may serve to call attention to a complex morphological problem which has an important bearing on questionsnbsp;connected with the relative positions of the several genera. It wouldnbsp;be out of place to enter fully into this difficult subject, but it is onenbsp;that demands careful attention by students of extinct types.

The number of seeds borne on each scale is an important feature in the recognition of genera. In the Abietineae eachnbsp;scale usually bears two seeds though it is not uncommon to findnbsp;single-seeded seminiferous scales such as those of Finns monophyllanbsp;(fig. 686, A�C) in which there is a deep cavity showing that thenbsp;seed was partially embedded in the supporting organ. Such anbsp;scale might, as a fossil, be easily mistaken for an Araucarian cone-scale. In Cunninghamia there are three seeds to each scale (fig.nbsp;684, K); in Athrotaxis and Cryptomeria 3�6; in Sequoia 5; and

^ For a recent discussion on the origin and relationship of the Araucarineae see Burlingame�s paper (15^) which appeared after this chapter was written.

Eig. 685. Saxegotkaea conspicua. Section of ovule; n, nucellus; i, integument; * st, stigma-like apex of nucellus; o, epimatium; c, young megaspore. (Afternbsp;Stiles.)

Eig, 686. A�C, Pinus monophylla, cone-scales with two seeds (A), one seed (B), and in distal view (C). D, E, Taxodium mucronatum, cone in surface-view (D)nbsp;and section (E); s, seed. E, Taxodium distichum, scale. (From specimens innbsp;the British Museum.) M. S.

in Sciadopitys 7�9. Fig. 686, E represents half a cone of Taxodium in which the distally expanded woody ends of the scales are tightlynbsp;joined by their edges and form a hard case enclosing as in annbsp;ovary several angular seeds, the slender stalks being shrivellednbsp;and inconspicuous. The cones of the Cupressineae and Calli-trineae are characterised by a whorled arrangement and a comparatively small number of the scales. In Cupresstis the conesnbsp;are oval or spherical and each scale bears 6�20 seeds; in thenbsp;Callitrineae the cones are valvular (figs. 703, 762, B�D) and thenbsp;scales vary from 2 to 3 in Callitris and fromnbsp;7 to 8 in Widdringtonia. The small conesnbsp;of Saxegothaea consist of one-seeded mega-sporophylls which become fleshy and partially concrescent (fig. 687); in Juniperusnbsp;the strobilus has the appearance of a berry;nbsp;in Microcachrys'the leaves pass graduallynbsp;into the single-seeded verticillate megaspo-rophylls, each with two vascular strands^

(fig. 684, T) and an epimatium, a, on one side of the ovule; in the ripe cone the mega-sporophylls are fleshy but not connate as innbsp;Saxegothaea^. In Dacrydium the megaspo-rophylls differ but slightly from the foliagenbsp;leaves in some species, e.g. in D. Balansaenbsp;(fig. 684, P) a single leaf at the apex of anbsp;branch bears an ovule partially covered bynbsp;a hood-like epimatium. In Torreya^ a verynbsp;short shoot in the axil of a leaf bears twonbsp;bracts and each subtends an ovule and twonbsp;pairs of bracteoles. The seeds of T. cali-fornica, which may be 4 cm. long, are enclosed by a thick integumentnbsp;differentiated into a sarcotesta and �sclerotesta surrounding anbsp;ruminated endosperm (fig. 688): there is a ring of vascular bundlesnbsp;at the limit of the free part of the integument and this is regardednbsp;by Oliver* as homologous with the tracheal plate at the base of

the nucellus of Cycadean seeds. The portion of the Torreya seed (fig. 684, V) below the free part of the nucellus has, according tonbsp;Oliver, been produced by the intercalation of a new basal regionnbsp;that has pushed up the chalaza. Cephalotaxus^ has plum-like seedsnbsp;similar to those of Torreya. In Phyllocladus^ (fig. 675, E) an ovulenbsp;enclosed in a papery epimatium occurs in the axil of a succulentnbsp;bract, and in Taxus a terminal ovule is borne on a short shoot

Pro. 688. Torreya nueifera, transverse section of seed; o, arillus; v, vascular tissue; sc, outer zone of integument in which the hard shell will be formed;nbsp;i, inner part of integument; mw, wall of megaspore; n, nucellus; ps, pro-thallus. (After Oliver.)

without any megasporophylls. In Podocarpus^ the megasporo-phylls are few and a part or whole of the reproductive apparatus is fleshy; the strobilus of P. Totara (fig. 684, G) consists of 3�4nbsp;bracts two of which are usually fertile. Similarty the strobilusnbsp;of P. imbric.ata is formed of 2�4, bracts with swollen bases formingnbsp;the receptacle; in the section shown in fig. 684, I, one bract is

fertile and its apex overtops the ovule, while a second bract is sterile. Fig. 684, H shows the strobilus of P. neriifolia in whichnbsp;the lowest bracts are leaf-like and the swollen bases of the uppernbsp;bracts are fused with the axis and each other to form a receptaclenbsp;analogous to the flower-axis of Anacardium. P. spicata has anbsp;long loose strobilus bearing several ovules (fig. 684, F); and innbsp;P. Nagi a single seed occurs on an axillary branch bearing smallnbsp;decussate scales: the flesh of the seed is formed from the epi-matium, and the sclerotesta from the integument.

The seeds of Conifers vary greatly in size and shape: those of some species of Pinus and several other Abietineae have a largenbsp;wing while others are wingless. The absence of a wing in a fossilnbsp;seed does not necessarily denote an original feature. The seedsnbsp;of Cedrus, Abies, Picea, Pseudotsuga, Keteleeria and Tsuga arenbsp;winged. Gliick^ calls attention to differences in the relation ofnbsp;seed to wing in certain Abietineae: in Picea the base of the wingnbsp;covers the seed like a spoon; in Abies, Larix, and Cedrus the seednbsp;lies in a pocket formed by the enclosing wing, while in Pinus thenbsp;wing embraces the seed like a pair of pincers. In Agathis (fig.nbsp;682, A) there are two very unequal wings. The much smaller seedsnbsp;of many Conifers have 2�3 wings: in Cupressus they are morenbsp;or less equal (fig. 684, Q), in Libocedrus and Fokienia markedlynbsp;unequal; in Fitzroya and Cunninghamia there are 2�3 wings.nbsp;Our knowledge of the relative vitality of Conifer seeds is meagre^,nbsp;and from the point of view of the possibilities of dispersal furthernbsp;research is desirable. The viviparous habit of Rhizophora andnbsp;other Mangrove Dicotyledons, believed by Guppy� to be a primitivenbsp;feature, is recorded in Podocarpus MakoyP.

The relation between nucellus and integument is less uniform in Conifers than in Cycads. In some genera, e.g. Agathis (fig.nbsp;682, B), Dacrydium, Phyllocladus, Fitzroya, Callitris and a fewnbsp;others the nucellus is free from the integument to the base; innbsp;Podocarpus the relation is variable; in Pinus and other Abietineae,nbsp;in Torreya (fig. 684, V) and some other genera the nucellar*apexnbsp;alone is free. The free summit often has the form of a steep cone:nbsp;in Araucaria^ and to a greater degree in Saxegothaea (fig. 685, st)

this protrudes through the micropyle: in Fitzroya patagonica the prolongation of the integument as a micropylar tube with a stigmalike terminal expansion is particularly striking. There is nonbsp;regular pollen-chamber as in Cycads and Ginkgo, but in Pseudo-tsuga^ a two-storied chamber, analogous to the pollen-chamber, isnbsp;formed by a knee-like bend in the integument. A peculiar typenbsp;of pollination characterises Araucaria: the microspores germinatenbsp;on the ligule or on the megasporophyll and their tubes grow overnbsp;or into the scale-tissues on their way to the ovule^. The archegonianbsp;of Conifers usually occur at the apex of the prof hallus and are fewnbsp;in number, they are separated by a few layers of cells (Abietineae)nbsp;or form a compact group (Cupressineae). In some Podocarpsnbsp;there may be as many as 14, in Taxodium 34, in AgalJiis 60,nbsp;irregularly distributed on the sides of the prothallus. In Widdring-tonia^ 100 archegonia are recorded occupying a lateral position ; innbsp;Actinostrohus^ Saxton has discovered groups of laterally placednbsp;archegonia. In Sequoia^ the archegonia are also numerous andnbsp;not confined to the apex. It is an open question whether or notnbsp;the greater number and irregular disposition of the archegonianbsp;are primitive features. The occasional occurrence of lateralnbsp;archegonia in Pinus may be a revival of an older habit.

The result of recent research into the morphology and life-histories of genera demand certain changes in the generally adopted grouping. The following classification is an attempt tonbsp;give clearer expression to the inter-relationships of existingnbsp;gen era �. Arnoldi ^ proposed to withdraw Sciadopitys from Sequoia,nbsp;Taxodium, and other members of the Taxodineae as the type of anbsp;separate family; he also suggested the isolation of Sequoia. Thenbsp;more recent work of Coker� and Lawson� favours the removal ofnbsp;Taxodium and Cryptomeria to the Cupressineae. Miyake�s researches^� point to a similar affinity in the case of Cunninghamia.nbsp;The genera Athrotaxis, Fokienia, and Taiivania are placed tentatively in the Cupressineae. The family-name Callitrineae, first

^ Lawson (09). nbsp;nbsp;nbsp;^ Thomson (07); EaineB(13); Burlingame (13); (15).

* nbsp;nbsp;nbsp;Saxton (10).nbsp;nbsp;nbsp;nbsp;lt; Ibid. (13).nbsp;nbsp;nbsp;nbsp;* Shaw, W. R. (96); Arnokli (01).

� See also Saxton (13*). nbsp;nbsp;nbsp;� Arnoldi (01); Lawson (10); Radais (94).

* nbsp;nbsp;nbsp;Coker (03).nbsp;nbsp;nbsp;nbsp;� Lawson (04).nbsp;nbsp;nbsp;nbsp;� Miyake (10).

used by Masters, has been revived by Saxton^ to give expression to the distinctive characters of CalUtris, Widdringtonia, andnbsp;Actinostrobus. Saxton�s work on Tetraclinis leads him to assignnbsp;it to the Cupressineae. Pilger^ makes Phyllodadus the sole genusnbsp;of Phyllocladoideae, while Miss Eobertson� includes it in thenbsp;Podocarpeae though recognising leanings towards the Taxineae.nbsp;It should be stated that the changes in classification suggestednbsp;are based mainly on characters of the gametophyte thoughnbsp;anatomical and vegetative features have not been entirelynbsp;neglected

II. nbsp;nbsp;nbsp;CUPRESSINEAE. Gupressus, CJiamaecyparis, Liho-cedrus, Thuya, Juniperus, Fitzroya, Diselma, Thujopsis, Taxodium,nbsp;Glyptostrobus, Cryptomeria, Cunninghamia, Taiwania, Fohienia,nbsp;Athrotaxis, Tetraclinis.

VI. nbsp;nbsp;nbsp;ABIETINEAE. Finns, Cedrus, Larix, Pseudolarix,nbsp;Picea, Tsuga, Abies, Pseudotsuga, Keteleeria.

VII. nbsp;nbsp;nbsp;PODOCARPINEAE. Podocarpus, Dacrydium, Micro-cachrys, Acmopyle, Pherosphaera, Saxegothae'a.

The order of the families is not intended to indicate their natural sequence in an evolutionary series, though the Arau-carineae are considered to be the most primitive. As certainnbsp;authors have suggested, Saxegothaea is probably closely allied tonbsp;Araucaria, but this is not indicated in the order adopted.

The distribution of the Conifers�, though too wide a subject for more than a brief notice, is of great interest from a palaeontological point of view. The ABIETINEAE, comparable in theirnbsp;present dominant role with the Polypodiaceae among the Ferns,

Saxton (10^); (13�). nbsp;nbsp;nbsp;� Pilger (03).nbsp;nbsp;nbsp;nbsp;� Robertson (06).

are the most widely spread; for the most part restricted to the northern hemisphere, they are not unrepresented south of thenbsp;equator. Pinus reaches the tree-limit in the north and extendsnbsp;as far south as Formosa^, Siam, the Malay region, the Philippines^,nbsp;S. Africa, and the West Indies. Picea has a similar distributionnbsp;in the north and reaches to the temperate regions of the southernnbsp;hemisphere. Abies ranges from Europe and Algeria to Siberia,nbsp;the Himalayas, Japan, and Formosa. Larix flourishes in northernnbsp;Europe and Siberia, Canada and the northern United States, thenbsp;Himalayas, and Japan. Tsuga is more especially a North American

Fig. 689. Araucaria imhricata on the Andes, Argentina. (From a photo* graph by Dr Wieland.)

genus, but it occurs in the Himalayas and in Japan. Pseudotsuga is characteristic of N.W. America and is recorded from Formosa.nbsp;Pseudolarix is a native of N.E. China and Formosa. Cedrus^nbsp;occurs in Algeria, Morocco, Syria, Cyprus, and the western Himalayas. The distribution of the ARAUCARINEAE affords anbsp;striking example of the contrast between the present and pastnbsp;range of a family. Araucaria occurs in Brazil, Chile and Argentina, in Australia, New Caledonia, New Guinea, the Pacific islands.nbsp;Agathis is confined to the Australian and Malay region. New Zea-

land, New Caledonia, and the Queen Charlotte Islands. The two trees of Araucaria Bidwillii shown in the Frontispiece are survivorsnbsp;of a forest on the hills of Queensland. The photograph reproducednbsp;in fig. 689, for which I am indebted to Dr Wieland^, illustrates thenbsp;habit of Araucaria imbricata on the eastern slopes of the Andesnbsp;in South-West Argentina where the trunks reach a diameter ofnbsp;two metres. There are few existing trees comparable with thesenbsp;venerable types in the impression they produce of the lapse ofnbsp;ages and the vicissitudes of a dwindled race.

CUPRBSSINEAE. Cupressus occurs in North America including the Californian coast and Mexico, in S.E. Europe, temperate Asia, China, and Japan. Chamaecyparis extends to the Sitkanbsp;Sound and flourishes in China, Japan, and Formosa. Lihocedrus,nbsp;one of the few genera met with in both hemispheres, has a discontinuous distribution; it occurs in California, Chile, Japan,nbsp;Australia, New Zealand, New Guinea, and New Caledonia. Thuyanbsp;flourishes over a wide area in North America and occurs in thenbsp;Far East. Juniperus is characteristic of temperate regions innbsp;both the old and new world and is represented in the Canaries,nbsp;the Azores, Somaliland, and Mexico. Fifzroya is confined tonbsp;Patagonia and Chile; Diselma to Tasmania. Thujopsis is exclusively Japanese. Taxodium is a native of Texas and Mexico,nbsp;while Glyptostrobus is a closely allied genus in China. The mono-typic Cryptomeria lives in China and Japan; Cunninghamia innbsp;China and Formosa. Taiwania and Fokienia have recently beennbsp;described from Formosa and East China respectively. Athrotaxisnbsp;is confined to Tasmania and Tetraclinis to North Africa.

CALLITRINEAE. CaUitris occurs iir Australia and New Caledonia; Widdringtonia grows in equatorial and South Africanbsp;and in Madagascar; Actinostrobus is restricted to West Australia.

SEQUOIINEAE. Sequoia is confined to the Pacific coast of North California; S. sempervirens the species with �the strongernbsp;hold upon existence � extends into Oregon, while S. gigantea formsnbsp;groves in the valleys of the Sierra Nevada.

PODOCARPINEAE. Podocarpus, one of the more successful genera, is essentially a southern type; in Africa it extends fromnbsp;1 Wieland (16), p. 224.

Cape Colony through East Africa to Abyssinia; it occurs in S. America from Patagonia to Brazil and replaces Finns on thenbsp;mountains of Costa Rica^; in the West Indies, Malaya, in thenbsp;Himalayas, China, Japan, Formosa^, Tasmania, New Zealand, Newnbsp;Caledonia, the Fiji Islands and New Guinea. Dacrydium has alsonbsp;a fairly wide range in the southern hemisphere, but like the othernbsp;members of the family, except Podocarpus, it does not cross thenbsp;equator; it is abundant in the Malay Archipelago� and occurs innbsp;New Zealand, Tasmania, New Caledonia, New Guinea, and onenbsp;species grows in the Chilean swamps. Saxegothaea is a monotypicnbsp;genus in Chile and Acmopyle an imperfectly known New Caledoniannbsp;genus. Pherosphaera'^ occurs in New South Wales, Victoria, andnbsp;Tasmania; Microcachrys, like Athrotaxis, is Tasmanian. Phyllo-cladus has a wide range in Tasmania, New Zealand, Borneo,nbsp;New Guinea, and the Philippines.

Some questions of exceptional interest from the point of view of the geographical distribution of Conifers in the Pacific regionnbsp;are ably discussed by Mr Guppy in the second volume of hisnbsp;admirable book Observations of a Nalumlist in the Pacific betweennbsp;1896 and 1899. He deals especially with Agathis, Podocarpus,nbsp;and Dacrydium, and his remarks illustrate the importance ofnbsp;taking into account palaeobotanical data in any general discussionnbsp;of the problems suggested by the present and often discontinuousnbsp;range of existing genera. �If,� he says, �there is a real difficultynbsp;in applying our canons of plant-dispersal to the distribution ofnbsp;Dammara [Agathis], it is merely the same difficulty that has sonbsp;often perplexed the botanist with other Coniferous genera innbsp;continental regions, such as, for instance, the occurrence of Finnsnbsp;excelsa on the far-removed mountains of the Himalayas, and thenbsp;existence of the Cedar in its isolated homes on the Atlas, thenbsp;Lebanon mountains, and the Himalayas. Such difficulties largelynbsp;disappear if we regard the present distribution of the Coniferae asnbsp;the remnant of what it was in an ancient geological period�.�

TAXINEAE. Taxus is chiefly a northern hemisphere genus; it occurs also in North Africa, Persia, India, the Philippines andnbsp;the Far East, and extends from Newfoundland to Pacific North

America, Mexico, and Florida. Torreya has a more restricted and less continuous range in China, Japan, Florida, and California.nbsp;Torreya taxifolia is almost extinct; it is separated by over 3000nbsp;miles from the other American species T. californica and the Pacificnbsp;separates the latter from the two species in China and Japan^.nbsp;Cephalotaxus lives in central China, Japan, and India.

The anatomy of Conifers, more especially from the point of view of the identification of families and genera, has long occupiednbsp;the attention of botanists, and although much has been done innbsp;the direction of more intensive study, the limits within whichnbsp;anatomical features may be safely used are still but vaguelynbsp;defined. . Jeffrey^ goes so far as to claim for the anatomicalnbsp;characters of plants a taxonomic value equal to that assigned bynbsp;zoologists to the anatomical features of animals. Though oftennbsp;extremely useful, in many cases anatomical characters do notnbsp;reveal more than an affinity between a fossil specimen and a groupnbsp;of recent genera. Statements are often based on insufficient datanbsp;and many authors have not appreciated the range of variationnbsp;in the vegetative shoots of a single tree. Attention has beennbsp;drawn to the fact that anatomical features are especially variablenbsp;in branches, and several authors have shown that characters tonbsp;which importance has been attached are much less constant thannbsp;has usually been supposed; many features, frequently acceptednbsp;as trustworthy criteria from the point of view of identification,nbsp;occur sporadically in other genera than those with which they arenbsp;usually associated. In the following summary attention is directednbsp;to the comparative value of different characters, and prominence isnbsp;given to possible sources of error in inferences based on anatomicalnbsp;features.

The wood of a Conifer consists only of tracheids, with or without resin-canals, and xylem-parenchyma and is characterised bynbsp;narrow medullary rays usually one-cell broad. For conveniencenbsp;in description it is proposed to speak of the wood of the Conifernbsp;type as pycnoxylic� in distinction to the Cycadean type ofnbsp;wood which is styled manoxylic. The presence or absence of

well defined rings of growth should he noted and attention paid to the breadth of the late summer (�autumn�) wood; Goeppert^nbsp;considered the breadth of annual rings a character of importance,nbsp;but Kraus^ and others have shown that this is of little significance.nbsp;In the Cupressineae (in the more restricted sense) it is probablynbsp;true that the rings are generally though not invariably narrowernbsp;than in Abietineae: in roots the later wood is smaller in amountnbsp;and there is a more sudden transition to the spring-wood than innbsp;stems�. Though as a rule there is a considerable difference innbsp;the thickness between the walls of the spring and summer tracheids,nbsp;in Podocarfus Nagi* the difference is slight. In some speciesnbsp;of Araucaria the rings are absent or feebly marked, a fact noticednbsp;long ago by Nicol�, and in other Conifers, e.g., Widdringtonianbsp;juniperoides and Tetraclinis^ there may be no definite rings; innbsp;^ Libocedrus macrolepis'^ there is but little difference in the thicknessnbsp;of the spring and summer tracheids. It is, however, impossiblenbsp;to say to what extent this is an inherent tendency and how far itnbsp;reflects the influence of external conditions: it may be that thenbsp;frequent absence of rings in Araucarian wood is explicable on thenbsp;hypothesis that this family is the oldest and most closely relatednbsp;to Palaeozoic types, which are almost invariably characterised bynbsp;an absence of rings: the habit of forming well defined spring- andnbsp;late summer-wood may have been acquired at a later stage�. Thenbsp;interest of annual rings is rather biological than taxonomic and itnbsp;is chiefly in connexion with fossil plants as tests of climate thatnbsp;attention has been directed to this feature�.

The genus Taxus is peculiar in having no resin-ducts in the cortex or stele of stem and root or in the leaves. In some generanbsp;resin-canals are a constant feature in the secondary wood, e.g.,nbsp;Pinus, Picea, Larix, Pseudotsuga; while in other Abietineaenbsp;canals do not usually occur in the xylem. This distinction is,nbsp;however, by no means constant and, as Jeffreyr� has shown, the

� Fujioka (13) p. 213. nbsp;nbsp;nbsp;* Thomson (13) p. 33; Gothan (07) p. 25.

great majority of Conifers which are normally without resin-canals in the wood have the power of producing them in response tonbsp;traumatic stimuli. In Cedrus, Pseudolarix, and Tsuga resin-canals are usually confined to the primary xylem of the root butnbsp;wounding induces the development of canals in other parts of thenbsp;wood. In Cedrus, however, both horizontal and vertical traumatic canals may occur whereas in other Abietineae the traumaticnbsp;canals are only vertical^. Eesin-canals may occur in the first-yearnbsp;wood of some species of Abies (fig. 690, B) and in Sequoia giganteanbsp;they are present in the first-year wood of vigorous branches andnbsp;in the peduncles of cones, but do not normally occur in the laternbsp;wood. In S. sempervirens canals are as a rule absent and arenbsp;developed only after wounding (fig. 690, A), In the Araucarineae

Fig. 690. A, Sequoia sempervirens. B, Abies sp. showing traumatic canals in the wood. (After Jeffrey.)

resin-canals are absent nor are they produced in injured stems: this failure to produce canals in response to disturbances set upnbsp;by wounds is considered by Jeffrey to be an indication of thenbsp;relatively late evolution of the family. Pinus, with abundantnbsp;canals, is regarded as one of the more primitive types; Abies,nbsp;with very few canals in healthy specimens but readily producingnbsp;them on wounding (fig. 690, B), is regarded as a slightly laternbsp;product of evolution, while Sequoia sempervirens (fig. 690, A) innbsp;which traumatic canals alone occur is still further removed fromnbsp;the original stock, and the Araucarineae, which are considered tonbsp;1 Jeffrey (05) p. 25.

have lost the power of reversion retained by Sequoia and Abies, are placed higher in the evolutionary series. The vestigial significance of resin-canals is by no means generally admitted. Pen-hallow^ holds, and I believe rightly, that they are not primitivenbsp;features; their occurrence in the young shoots of certain speciesnbsp;and in the peduncles of cones but not in the older wood may, asnbsp;Crothan^ suggests, be correlated with a greater need of protection.nbsp;Kirsch� considers that the development of canals in young woodnbsp;and in peduncles may be connected wnth the relatively greaternbsp;abundance of food in those regions which, in his opinion, wouldnbsp;induce a greater production of parenchyma and secretory passages.nbsp;Moreover, if the occurrence of canals in the axis of a female conenbsp;of Sequoia gigantea is attributed to the retention of an ancestralnbsp;character, why do not canals also occur in the axis of the micro-strobilil The facts demonstrated by Jeffrey and his pupils arenbsp;of great interest, but considered by themselves they may equallynbsp;well be interpreted as favouring the greater specialisation andnbsp;more recent development of those genera in which the productionnbsp;of resin-canals is a normal character.

The structure of the epithelial cells is employed as a taxonomic character though, as Conwentz* suggests, it is not a very satisfactory criterion and in petrified tissues it is often difficult tonbsp;distinguish between true thick walls and walls thickened bynbsp;secondary deposits. In Pinus the walls of the cells lining thenbsp;canals are frequently thin�, but in some species thick; Lanx andnbsp;Picea have thick-walled epithelial cells. The occurrence of tyloses,�nbsp;the parenchymatous cells that invade the cavities of water-conducting elements,�-has generally been regarded as the monopoly ofnbsp;-�Jngiosperins; though unknowm in recent Ferns they occur in somenbsp;extinct types. Chrysler� has showm that tyloses are produced innbsp;the tracheids of Pinus, apparently as a consequence of wounding.nbsp;Tyloses have also been found in some fossil coniferous woods.

The arrangement of the bordered pits on the radial walls cf the tracheids is the character to wKich most attention hasnbsp;been given. In the Abietineae they form either single or double,

and occasionally three or even four, rows (fig. 691, H); they are circular and not contiguous and the pits of a double row are as anbsp;rule on the same level; they are opposite and not alternate. Thisnbsp;type of pitting occurs also in all the other families except thenbsp;Araucarineae though in Agathis opposite pits are not unknownnbsp;(fig. 691, D) and Conwentz^ states that he has seen separate andnbsp;circular pits in the tracheids of recent species. The occasionalnbsp;occurrence of pits in clusters and not in opposite pairs has beennbsp;described in Finns Merhensis^: this is a feature characteristic ofnbsp;the tracheids of some fossil types, e.g., Cedroxylon transiens Goth.nbsp;In Agathis and Araucaria there may be 1�3 rows and as manynbsp;as 5 rows on the tracheids of cone peduncles (fig. 691, A). Thenbsp;pits are contiguous and flattened, and those of adjacent rows arenbsp;alternate and hexagonal^. Thomson has called attention to thenbsp;occasional occurrence, especially in the region of the rays, ofnbsp;transversely elongated or scalariform pits in the tracheids ofnbsp;Araucaria. A single series of flattened pits andnbsp;even the occasional occurrence of alternate hexagonal pits are not infallible criteria of an Arau-carian affinity; in Dacrydium the pits of a doublenbsp;row may be alternate though rarely contiguous,nbsp;and this is the case in some other genera, while innbsp;Saxegothaea^ (fig. 692) the pits are as a rule uni-seriate and often flattened. Worsdell� describesnbsp;circular and separate pits in the cone-scales ofnbsp;Araucaria and Thomson records alternate bi-seriate pits in the cone-axis and early wood ofnbsp;the Abietineae. Flattened pits are described innbsp;Podocarpus polystachya'^ and I have seen similarnbsp;pits in the wood of Torreya californica and several �!�- 692. Tracheids

^ Stiles (08). nbsp;nbsp;nbsp;� Worsdell (99).nbsp;nbsp;nbsp;nbsp;' Gerry (10).nbsp;nbsp;nbsp;nbsp;� Kraus (83).

Araucarian pits as small (9�12-8/n) in contrast to the broader pits (up to 21 /x) of the Abietineae, those in Cupressineae being intermediate in size. There may, however, be considerable differencenbsp;in the size of the pits in a single type^. The occurrence of spiralnbsp;thickening bands in addition to bordered pits is characteristic ofnbsp;the Taxineae, but spiral bands occur sporadically in the secondarynbsp;tracheids of other Conifers, e.g., Phyllocladus, Larix leptolepis,nbsp;species of Abies and other Abietineae^, also in some species ofnbsp;Cupressus^. In Psevdotsuga spiral bands may occur in all the tracheids of an annual ring. Some authors assert that the arrangementnbsp;and grouping of the bands in a tracheid constitute a character ofnbsp;generic value, but there is not complete agreement on this point^.

The walls of tracheids frequently exhibit well marked spiral patterns�, due to an entirely different cause, which, especiallynbsp;in some petrified woods, closely simulate spiral bands. In thenbsp;process of decay enzyme-action may etch into prominence thenbsp;striation or spiral method of wall-construction; but the spiralsnbsp;are steeper than those of the true thickening bands. The presencenbsp;of xylem-parenchyma, though of diagnostic value, is too uncertainnbsp;and variable a character to be used with great confidence. Innbsp;young shoots of Sequoia xylem-parenchyma may be absent thoughnbsp;it is present in older branches�. Such parenchyma occasionallynbsp;occurs in Abietineous wood�, but it is generally considered a characteristic feature of the Cupressineae though in the wood of somenbsp;members of that family it is not always obvious. The presencenbsp;of drops of resin in the cells which form vertical series in differentnbsp;parts of the wood, or only in the late summer wood, may rendernbsp;the xylem-parenchyma conspicuous both in transverse and longitudinal sections. Kows of parenchyma occur in the wood ofnbsp;Abies pectinata^, also in Podocarpus and Dacrydium. In Taxodium^nbsp;the thick horizontal walls of the cells are a characteristic feature.nbsp;Wood-parenchyma is rare in the Araucarineae and, as Penhallow^�

^ Gothan (05) p. 54; Penhallow (07) p. 41; Burgerstein (08) p. 104; Kraus (83) p. 103; Nakamura (83).

pointed out, tracheids with horizontal patches of resin may be mistaken for resiniferous parenchyma; but true parenchyma,nbsp;occasionally occurs^.

Attention has been called to the diagnostic value of the horizontal thickening bands which on staining, and often in fossil wood, stand out as conspicuous features on the tracheids of thenbsp;great majority of Conifers (fig. 693, C). Many authors speak ofnbsp;these bands as bars of Sanio^, apparently overlooking the fact thatnbsp;this term (Sanio�s �Balken�) was used by Midler� for the horizontalnbsp;bars previously described by Winkler* on the tracheids of Araucarianbsp;hrasiliensis (fig. 693, I). Groom and Eushton� have also callednbsp;attention to the inaccurate use of the term Sanio�s bars and theynbsp;suggest the more appropriate expression Sanio�s rims for thenbsp;persistent margins of the primordial pit-areas which appear asnbsp;horizontal lines between the bordered pits. An American airthornbsp;goes so far as to claim that �by far the most reliable criterion fornbsp;diagnosing coniferous wood is the occurrence of the bars [rims]nbsp;of Sanio�.� But if, as Jeffrey and his pupils assert, Sanio�s rimsnbsp;are present on the tracheids of all Conifers except the Araucarineaenbsp;the diagnostic value of this feature is exceedingly small. Jeffrey^nbsp;has shown that in the first-year wood of Araucaria and in the conenbsp;of A. Bidwillii the pits are not always contiguous and rims ofnbsp;Sanio may then be present. Moreover, as Thomson� states, thenbsp;darkly stained lines between contiguous pits on some Araucariannbsp;tracheids (fig. 691, F) may be regarded as feebly marked rims ofnbsp;Sanio. It is not surprising that in the case of tracheids with 2�3nbsp;series of contiguous hexagonal pits, which leave no free surface�,nbsp;Sanio�s rims are not represented*�.

The most recent contribution to our knowledge of the rims of Sanio is by Mr Sifton** who describes them in petioles of Cycas

quot; Holden (13) p. 252; (13^). nbsp;nbsp;nbsp;� Jeffrey (12) PI. vi. fig. b.

In a recent paper entitled �Gliding growth and bars of Sanio� (Grossenbacher, dmer. Journ. Bot. vol. I. no. 10,1914) the expression � bars of Sanio � is employed innbsp;�rn unusual sense and not in accordance with the ordinary usage of the term bars,nbsp;or rims, of Sanio.

revoluta. Jeffrey regards the occurrence of Sanio�s rims in the, cone-axis of Araucaria as a vestigial phenomenon. He failednbsp;to find any rims of Sanio in the cone-axes of Gycads and thisnbsp;negative evidence was regarded as favourable to his view that thenbsp;rims in the Araucarineae are derived from the more fully developednbsp;rims in the Abietineae. Sifton shows that the rims on the tracheidsnbsp;of Cycas revoluta agree closely with those in the xylem of thenbsp;Araucarian cone-axes and with those in the cone-axis and root ofnbsp;certain Pines. On the assumption that roots and cone-axes arenbsp;likely to retain ancestral characters, the resemblance of their rimsnbsp;to those found in the Araucarineae supports the view that thenbsp;Abietineae are descended from ancestors which had rims of Sanionbsp;of the Araucarian or Cycadean type. The conclusion is that thenbsp;shorter rims in the Araucarineae and on the tracheids of the cone-axis and root of the Abietineae represent the primitive form, thenbsp;broader rims met with in the Abietineae and most other Conifersnbsp;being later developments.

The pitting oir the walls of medullary-ray cells has in recent years received special attention: in some Conifers the horizontalnbsp;and tangential walls are strongly pitted (fig. 693, A, G), and thisnbsp;feature is clearly seen in both radial and tangential sections asnbsp;also, in the case of the horizontal walls, in transverse sectionsnbsp;(fig. 693, D, E, F). In most of the Abietineae the pits on thenbsp;horizontal and tangential walls are a prominent feature while onnbsp;the other hand in some Abietineae the pitting of these walls isnbsp;feebly developed: to this type of pitting Gothan�^ has given thenbsp;name Abietineous pitting. In the great majority of recentnbsp;genera other than members of the Abietineae the horizontal andnbsp;tangential walls are smooth (fig. 693, L, 0); but there are exceptions.nbsp;The ray cells in the cone-scales of Agathis are pitted and speciesnbsp;of Juniperus^, Libocedrus decurrens and Fitzroya also exhibit anbsp;form of Abietineous pitting. Gothan points out that in somenbsp;Junipers and a few other Cupressineae the pits in the tangentialnbsp;walls differ in detail from the typical Abietineous form and thatnbsp;the pits in the horizontal walls are much less distinct than in thenbsp;Abietineae: there is, however, no very clear distinction between

1/quot; �
o	o	lt;

s		sp
Pi		1/^

iii

'0	�
�	�
0	0
oo	oo
oo	� o
H	�8

o	1) .	0 0	w.	Q Q
	//			0	0
lt;7	0 o		0 a \		o
	0 0	Q	Q 1	0 ^

M' nbsp;nbsp;nbsp;O

Fig. 693. A, Abies Veitchii, medullary ray. B, Pinus silvestris, medullary-ray tracheid. C, Abies balsamea, pits and Sanio�s rims. D, E, Pits in medullary-raynbsp;cells in Abies homohpis. P, Pits in tangential wall of ray cell of Juniperusnbsp;virginiana. G, H, K, L, M, N, 0, Pits in medullary-ray cells (radial view) innbsp;Cedrusatlantica iG); Taxodiumdistichum (H); Podocarpus'andina (K); P.salici-folia (L); Glyptostrobus (M); Sciadopitys (N); sp, spring wood; s, summer wood;nbsp;Thuya gigantea (0). I, Bars, b, in traoheids of Araucaria brasiliensis. P. Spiralnbsp;bands in tracheids of Torreya nucifera. (A, N, after Nakamura; B. D�H,nbsp;K, L, P, after Gothan; C, after Gerry; I, after Winkler; M, after Kraus;nbsp;O, after Penhallow.)

his Juniperoid and Abietineous types. Each medullary-ray cell is longer in a radial direction than the breadth of a single tracheidnbsp;and statements as to the number of pits on the radial wall of anbsp;ray cell have reference to the area bounded laterally by thenbsp;vertical walls of a tracheid: this area may be designated the fieldnbsp;(�Kreuzungsfeld�; �aire mitoyenne�). It is, however, incorrect innbsp;many cases to speak of pits on the radial walls of medullary-raynbsp;cells, and if pits occur they are never bordered: in the Abietineaenbsp;the walls are pitted, but in most other Conifers the pits seen in thenbsp;field belong to the tracheids in contact with the rays. But in viewnbsp;of the general use of the expression medullar3r-ray pitting it wouldnbsp;be inconvenient to discontinue the current terminology. There isnbsp;a difference of opinion as to the value of medullary-ray pitting asnbsp;a criterion of affinity, and it is probable that Gothan overestimates the taxonomic significance of this character. Withinnbsp;certain limits the pitting on the walls of medullary-ray cells isnbsp;undoubtedly important, but a comparison of sections of the woodnbsp;of a collection of genera shakes one�s confidence in the conclusionsnbsp;based by some authors on the form and number of the pits in thenbsp;field. In the Abietineae a single large simple pit (�Eipore�)nbsp;occupies the field in some species of Pinus but the same typenbsp;occurs also in Sciadopitys^, species of Podocarpus'^ (fig. 693, K, N),nbsp;Microcachrys, Dacrydium, and Phyllocladus^. The Araucarineaenbsp;are usually described as having several (2�6) oblique pits in eachnbsp;field and as a rule no pits in the horizontal and tangential wallsnbsp;of the ray cells, but Thomson^ states that the bordered pits in thenbsp;field of Araucaria and Agathis are confined to the tracheids andnbsp;an examination of macerated tissue confirms the absence of pits onnbsp;the walls of the ray cells. A similar absence of pits characterisesnbsp;some other Conifers. Gothan has suggested the term Cupressoidnbsp;pitting for Conifers in which the field in the spring-wood containsnbsp;bordered pits with a fairly broad pore in a more or less horizontalnbsp;position, a type of pitting found in some Cupressineae as also innbsp;Sequoia and certain other genera. He applies the name Podo-carpoid pitting to woods in which the field shows bordered pits

with a narrower pore occupying an obliquely vertical position, a form of pit well shown in some species of Podocarpus. In thenbsp;Podocarpineae and in most of the Cupressineae the tangentialnbsp;and horizontal walls of the ray cells are unpitted. An examinationnbsp;of sections of species of Thuya and some species of Cupressusnbsp;reveals the presence of pits in the field with an almost verticalnbsp;pore, and the variation in the breadth of the border and in the position of the pore is too great to admit of more than a restricted andnbsp;cautious use of this anatomical feature as a means of distinguishingnbsp;genera or even families. It is by no means easy even in sectionsnbsp;of recent woods to observe with accuracy the structure of the raynbsp;pits: in many cases they are more or less bordered, but the greaternbsp;� distinctness of the pore often leads to the neglect of the fainternbsp;border. Moreover the small medullary-ray pits may be converted into large pits by the action of fungal hyphae. The largenbsp;pits of some Pines, Sciadopitys, etc., represent one extreme;nbsp;intermediate types are represented by Cedrus, Taxodium, andnbsp;Glyptostrobus, while in Juniperus and several other genera the pitsnbsp;are smaller and more numerous^.

The depth of the rays as seen in tangential section is a feature to which much attention has been paid, but this is a very variablenbsp;and comparatively unimportant character^. In a single speciesnbsp;of Abies the depth varies from 1 to 63 cells�. Many authors innbsp;describing fossil wood state the number of rays per square millimetre of a tangential section. Characters such as these maynbsp;undoubtedly be useful in certain cases if used in conjunction withnbsp;others, due allowance being made for the range of variation withinnbsp;the limits of a single stem. A more important feature is thenbsp;occurrence of broad rays containing horizontal resin-canals suchnbsp;as those of Pinus, Picea, Larix, and Pseudotsuga. Another usefulnbsp;criterion is afforded by the association of horizontal tracheidsnbsp;(fig. 693, B) with the parenchyma of a ray usually at the uppernbsp;and lower margin but sometimes, e.g., Pinus canariensis*, in thenbsp;middle. The occurrence of such tracheids was formerly regardednbsp;as a trustworthy distinguishing feature of the Abietineae with thenbsp;exception of Abies and Pseudolarix, but they are now known to

occur in Abies^ and several members of other families. Chrysler^ states that Cedrus differs from Abies in having ray tracheids mixednbsp;with marginal parenchyma, and at the limit of an annual ring thenbsp;marginal tracheids may be replaced by shadowy cells or ghostsnbsp;of cells. The tracheids may have smooth walls as in Cedrus,nbsp;Tsuga, Larix, Pinus Strobus, or, as in other species of Pmiisnbsp;(�Hard Pines�), their walls are characterised by irregular ingrowthsnbsp;or pegs. Wettstein� states that in Picea omorica horizontalnbsp;tracheids though common in the main stem do not occur in thenbsp;rays of branches. De Bary^ recognised tracheids in the rays ofnbsp;Sciadofitys; they have been recorded also in Juniferus, Cupressus,nbsp;Thuja, Sequoia^ and, as the result of wounding, in Cunninghamia^.

The occurrence of idioblasts in the form of irregular thick-walled elements is characteristic of the pith and cortex of Araucaria, but similar cells are found in the pith of Torreya nucifera, Podocarjyus neriifolia, Dacrydimn cupressinutrA and Cryptomeria.nbsp;In some cases, e.g., Abies 7nagnifica^, Picea omorica, horizontalnbsp;rows of thick-walled cells form diaphragms in the pith. Thenbsp;structure of the secondary phloem has received relatively littlenbsp;attention and owing to its comparatively rare preservation innbsp;fossils it is less important to the palaeobotanist. Irr the Cupres-sineae the regrrlar alternation of tangential rows of hard and softnbsp;bast is a characteristic feature, while in the Abietineae the phloemnbsp;consists of sieve-tubes and parenchyma with a few scatterednbsp;stone-cells. The absence of albuminous cells in the medullarynbsp;rays of the phloem region in the Araucarineae is noteworthy andnbsp;Thomson� states that he found none in Podocarpus.

The structure of the stomata in the leaves of Conifers is fairly uniform: their distribution should be noted though this in itselfnbsp;is not of much value as a distinguishirrg feature. They may benbsp;confined to regular bands {Abies, etc.) or grooves {Torreya, etc.,nbsp;fig. 694, B), or irregularly distributed. The position and numbernbsp;of resin-ducts is often a useful guide: to quote one example only,nbsp;in the leaves of Agathis and Araucaria (fig. 694, C) the ducts

^ Kubart (ID) nbsp;nbsp;nbsp;s Jeffrey (05).nbsp;nbsp;nbsp;nbsp;� Thomson (13) p. 31.

occur between the veins, but in the leaves of some species of Podocarfus (fig. 695, D), externally indistinguishable from thosenbsp;of Agathis, the ducts are below-the veins^. Caution must be exercised in using the number of resin-ducts as a diagnostic character.nbsp;Schroetei^ draws attention to the occasional absence of ducts innbsp;Picea excelsa leaves while in others 1 or 2 are present. In mostnbsp;leaves there is some mechanical tissue immediately below thenbsp;epidermis either as scattered fibres or a continuous layer, but in

Taxus and Torreya there is none. The occurrence of branched idioblasts is a striking feature in the mesophyll of Agathis, Araucaria, Sciadopitys, and Podocarpus. The infoldings of the wallsnbsp;of the chlorenchyma are especially characteristic of Pine leavesnbsp;(fig. 694, D) and they occur also in Cedrus (fig. 694, F), Pseudo-larix and some other genera. The structure, extent, and positionnbsp;of short isodiametric tracheids in association with the conductingnbsp;tissue is an important feature. This tissue, the elements of whichnbsp;^ Seward and Ford (06) B.nbsp;nbsp;nbsp;nbsp;^ Schroeter (97).

are usually termed transfusion-tracheids^ is regarded by some authors as homologous with the centripetal wood of Cycadeannbsp;leaves (Vol. iii. p. 31) and Bernard definitely adopts the term centripetal xylem. Jeffrey^ and, more recently, Takeda� do not attachnbsp;a similar morphological significance to the short tracheids, andnbsp;they are probably justified in their sceptical attitude. Thenbsp;transfusion tracheids are often reticulately pitted: in many leavesnbsp;they form conspicuous flanges on the sides of the vascular bundlesnbsp;(fig. 694, B, C, t) or they may more or less encircle the vein. Innbsp;some leaves, e.ff., Araucaria, it is noticeable that the amount ofnbsp;transfusion tissue (fig. 694, C, t) increases as the vein is tracednbsp;towards the leaf-apex until the long and narrow elements may benbsp;entirely replaced by a group of short transfusion-tracheids.nbsp;Another type of accessory tracheid is occasionally met with,nbsp;namely elongated tracheids traversing the mesophyll between thenbsp;veins and the edge of the lamina. This is seen in the long andnbsp;narrow leaves of some Podocarps^.

The anatomy of cone-scales is too wide a subject for adequate treatment in this sketch. The general rule is that in each scalenbsp;there are two sets of vascdlar strands, a lower set of normallynbsp;orientated bundles and an upper inversely orientated series. Innbsp;some cone-scales, e.g., Araucaria, concentric vascular strands arenbsp;a prominent feature. The taxonomic significance of the anatomical characters of cone-scales has been discussed by Radaisnbsp;and more recently by Miss Aase�, to whose accounts the studentnbsp;is referred.

It has been shown that characters which it has been customary to associate with a definite type of wood may occur sporadicallynbsp;in several other Conifers; but this does not invalidate conclusionsnbsp;based on the prevalent occurrence of such features in a givennbsp;specimen. It is untrue to say that contiguous and alternate pitsnbsp;are the monopoly of the Araucaiineae and it is incorrect to assertnbsp;that in Araucarian wood the pits are never separate. Similarly

^ For further details with regard to leaf-anatomy, see Thomas, F. (66); Ber trand, C. E. (74); Mahlert (8.5); Striibing (88); Daguillon (90).

the pits on the medullary-ray cells, whatever relative value we may assign to this character, are in many cases of considerablenbsp;assistance even though we are not prepared to follow Gothan tonbsp;the full extent of his trust in the taxonomic importance of medullary-ray pitting.

Dr Groom^ has described the northern evergreen Conifers as architectural xerophytes having xeromorphic leaves with a xero-phytic structure. He discusses in his remarks on the Ecology ofnbsp;Conifers the correlation of the characteristic wood-structure andnbsp;the xeromorphic leaves, the cause of the survival of the Coniferaenbsp;in competition with Dicotyledons, and other questions of interestnbsp;to the student of the evolution and past history of the group.

Short summary of the characteristics of recent Conifers'^.

ARAUCAE.INHAE. There is a close agreement in the structure of the. wood in the two members of this family. Attention has already been called tonbsp;the normal type of pitting of the tracheids and to certain exceptional forms.nbsp;The occasional tendency towards a scalarrform type of pitting (fig. 691, D) isnbsp;an interesting point. Pits are fairly abundant on the tangential walls of thenbsp;xylem-elements. In the wood of Agathis robusta^ vertical rows of parenchymanbsp;are said to be fairly abundant. Annual rings not infrequently absent ornbsp;feebly developed. Medullary rays one-cell broad, rarely double, usuallynbsp;V�15 cells deep; but in Araucaria the depth may reach 26 cells: the ray cellsnbsp;occasionally present a distended appearance in tangential sections of the woodnbsp;(cf. Oinkgo biloha). The persistence of the leaf-traces in the old wood ofnbsp;Araucaria is a striking feature¹ considered by Lignier to possess diagnosticnbsp;importance: Jeffrey� states that traces are much less persistent in somenbsp;young stems. In Araucaria each trace arises as a single strand, but in Agathisnbsp;it leaves the perimedullary region as a double bundle�.

Agathis (fig. 695). Leaves sessile or slightly petiolate, opposite, subopposite or, on the main axis, spiral; ovate, broadly lanceolate {A. loranthi-folia, 13 X 5 cm.; A. macrophylla 17x5 cm.). Resin-canals between the veins; transfusion-tracheids fairly abundant, but less prominent than in Araucaria.nbsp;The almost spherical megastrobili (fig. 096) are very characteristic; they

^ In addition to the text-books on Conifers by Beissner, Veitch, and the account m Die Nal�rlichen PflanzenfamilUn the student should consult the Report of thenbsp;Conifer Conference, Journal of the Royal Horticultural Society, Vol. xiv. 1892.

bear a close superficial resemblance to Cedar cones but the latter are relatively narrow and often more or less flattened at the apex^. Bommer^nbsp;calls attention to the resemblance of Agaihis cones to those of the Dicotyledonnbsp;Dammaropsis hirujiana (Moraceae).

Araucaria. (Frontispiece; figs. 678�681, 689, etc.) The falcate tetragonal leaves of A. excelsa illustrate one type of leaf that is seen in its smallest form in A. Balansae (4�5 x 2-5 mm.). In A. Bidwillii (fig. 697) the leavesnbsp;are subsessUe and the fiat lamina may reach a length of 7 cm.: in A. Housteinii

the ovate-lanceolate leaves may be 10 cm. by 1 cm. Dimorphism^ in the foliage of a single shoot is not uncommon. The striking difference in some speciesnbsp;between the juvenile and adult foliage is illustrated in fig. 698. In the broadleaved species (Colymhea section) resin-canals occur between the veins (fig. 694,nbsp;C) and in the Eutada section, e.g., A. excelsa, the canals are scattered. Strobilinbsp;and cone-scales of Araucaria are described in an earlier part of this chapternbsp;(see page 113).

CUFBiESSINEAE. The absence of resin-canals in the xyleni is a feature shared by other families; but in the occurrence of xylem-parenchyina innbsp;different regions of the wood the Cupressineae differ as a rule from the Abie-tineae, though this is not a constant distinguishing character. The pits innbsp;the field vary from 1 to 6 or 8 in some genera, e.g., Taxodiwn and Glypto-strohns^: Gothan^ applies the term Cupressoid to medullary-ray pits character-

1) A

Fig. 698. Araucaria excelsa. A, Mature form. B, Seedling, (f nat. size.) ised by an upper and lower border (fig. 693, H) but, as pointed out on a previousnbsp;page, the position of the pore is by no means constant and in Taxodium thenbsp;ray pits are fairly large and almost simple.

Cwpressus and Chama�cyparis. By Beissner^ and many other authors both names are used in a generic sense, though Masters regards Chumaecyparis asnbsp;a subsection of Gupresaus. One distinguishing feature is the presence of more

than 2 seeds on each cone-scale of G-apressua. The leaves are whorled, in 4 ranks, and appressed to the axis (fig. 699). The young foliage of Gupressua,nbsp;Thuya, and other genera (fig. 700), which in the adult state has the form ofnbsp;scale-leaves, consists of spreading linear leaves: individuals in which thisnbsp;juvenile fohage persists are often spoken of as species of the �genus � Retinospora.nbsp;Cones oblong or globular, composed of a few pairs of scales with peltate distalnbsp;ends either smooth or provided with a prominent umbo (fig. 699).

Thuya. Very similar to Cupreaaua in the habit of its bi- or tri-pinnate shoots. The cones of 8�10 decussate scales are distinguished from those ofnbsp;Gupressua by their elongated, oblong, form and by the upwardly directednbsp;Scales with thickened apices in contrast to the more horizontal and peltatenbsp;scales of Gupressua-. there are 2 seeds to each scale, winged (sect. Euthuja)nbsp;or wingless (sect. Biota).

Libocedrua. Fohage shoots often broader than in Thuya-, in L. decurrena the appressed, flat, leaves are characterised by a long decurrent portion and

in L. Doniana the whorled arrangement is less obvious. There is a large canal below the midrib of the leaf as in Thuya and Cupressus. The conesnbsp;(4�6 valves),are longer and relatively narrower than in Thuya; seeds un-

Fig. 700. A, Young plant of Libocedrus decurrens; c, cotyledons; pr, primordial leaves; tr^ transitional leaves. B, Branch with adult (ad) foliage. (Afternbsp;Rendle, from Veitch.)

equally winged. Juniperus. The polymorphism of the shoots is especially striking; the leaves, 2�3 in each whorl, being small and appressed, spinousnbsp;and spreading, or flat and linear. The very narrow leaves of J. Sabina

differ widely from the broader, flat and sharply pointed l^ves of J. drupacea. There is a resin-canal near the lower surface^ or, in some species, e.g., J. oxy-cedrua, there may be no canal. The more or less globular or elongated fleshynbsp;strobilus consists of 1�1 whorls of megasporophylls.

Fitzroya and Disdma. In some examples of F. patagonica the leaves are short and crowded and shghtly falcate, in others more spreading; while innbsp;Diselma Archeri^ the shoots resemble those of some Lycopodiums. Thenbsp;leaves are in alternate ternary whorls and there is a single canal below thenbsp;vein. The cones consist of 3 alternate trimerous whorls and the fertile scalesnbsp;bear a variable number of winged ovules. Thujopais. Shoots similar tonbsp;those of Thuya but the decussate leaves are rather larger^. Cones narrow,nbsp;consisting of 8 clavate scales with 5 small winged seeds on each scale.

Taxodium. The slender deciduous shoots, 8�10 cm. long, bear 2-ranked linear leaves 2 cm. long, acute and lanceolate, also small leaves 10�17 mm.nbsp;long and barely 1 mm. broad. Circular depressed branch-soars are a characteristic feature of the leafless shoots. Globular cones^ composed of a fewnbsp;peltate scales with an irregular crenulate upper margin (fig. 686, D�P) eachnbsp;scale bearing 2 slightly winged seeds. The almost leafless spikes of mioro-strobili are a noteworthy feature. The wood of Taxodium closely resemblesnbsp;that of Sequoia sempervirens: the thicker horizontal walls of the xylem-parenchyma are quoted by Gothan as a characteristic feature^ of Taxodium.nbsp;Lingelsheim� says that this distinction is not valid; he states that simple pitsnbsp;occur in the cross-walls of the xylem-parenchyma of Taxodium but not in thenbsp;cells of Sequoia. Pits are, however, present in the parenchyma of both these 'nbsp;genera. Olypiostrobua. Similar in habit to Taxodium, but the leaves are notnbsp;2-ranked�; the cones are more pyriform; the seeds are smooth and narrownbsp;with lateral wings. The comparatively large circular pits on the radial wallsnbsp;of the ray cells are characteristic�.

Cryptomeria. Pohage-shoots (fig. 677) as in Araucaria excelsa: a large resin-canal occurs below the midrib of the laterally compressed leaves (fig. 694,nbsp;E). The cones have 20�30 peltate scales characterised by the deeply cleftnbsp;comb-like upper portion (fig. 684, M). The xylem-parenchyma is said bynbsp;Eujioka to be confined to the region between the spring and summer wood:nbsp;the pits in the field are variable in position and the breadth of the pore may benbsp;vertical or obhque.

Cunniughamia. Leaves densely spiral and spreading, narrow, lanceolate, acuminate and serrate, 2-5�6 cm. long, with a narrow decurrent base; a largenbsp;canal below the midrib. Cones ovoid-globular, 4 cm. long, composed ofnbsp;broad thin scales with a serrate edge and a fimbriate membrane on the abaxial

Kirchner, Loew, and Schr�ter (06) pp. 293, 314. Hooker, J. D. (60) PI. xcviii.

side of the 3 wing^ seeds (fig. 684, K). Branch-soars occur on the sterns^. Taiwania. This genus^ has the habit of Crypfomeria and cones recallingnbsp;those of CunningJiamia and Tsuga: each scale bears 2 seeds. Folcienia. Thenbsp;single species¹ ² is, in certain respects, intermediate between Cupressus andnbsp;Lihocedrus; the cones are globose like those of Chamaeeyparis and eachnbsp;scale bears two unequally winged seeds; the foliage is nearly identical withnbsp;that of Lihocedrus

Athrotaxis. (Figs. 684, N, 0; 701.) Leaves short, loosely spreading and slightly imbricate {A. selaginoides^), similar to those of Araucaria excelsa,

* nbsp;nbsp;nbsp;Lotsy (11) p. 51.nbsp;nbsp;nbsp;nbsp;2 Gard. Chron. Feb. 4, 1911.

more appressed and smaller, decussate or in close spirals (A. cupressoides) or, in A. laxijolia^, more like the foliage of Sequoia gigantea. The apex of eachnbsp;cone-scale has a sharp point at the distal end. TetracUnis (fig. 703, B). Thisnbsp;genus, which occurs in Algeria and Morocco, and is usually placed next tonbsp;Callitris and Widdringtonia, has recently been transferred by Saxton^, as thenbsp;result of his work on the life-history of T. articulata, to the Cupressineae: henbsp;believes that the Callitrineae were derived from the Northern Cupressineaenbsp;through some type resembling TetracUnis. The flattened fohage-shootsnbsp;resemble SaiicorpUa and bear 4-ranked small leaves; the cones consist ofnbsp;1 nearly equal decussate scales, each scale having 2 seeds with unequal wings.

SCIADOPITINEAE. Sciadopitys. The single species is characterised by the long and rigid �double needle� reaching a length of 12 cm. with a broadnbsp;furrow on the upper surface and a deeper groove on the lower face. Mrnbsp;Boodle^ has recently described an example of concrescence in needles of Pinusnbsp;Laricio Poir. var. nigricans Pari, which points to the probability that a morphological similarity exists between the double needle of Sciadopitys and abnormal,nbsp;fused, leaves of Pinus. Reference is made by Boodle to other possible viewsnbsp;that have been advanced with regard to the morphological nature of Sciadopitysnbsp;needles. The oblong woody cones, 7x4 cm.^ are fairly easy to identify bynbsp;the roinided and reflexed upper margins of the scales (fig. 684, S); there arenbsp;o�15 seeds, with a narrow wing, on each cone-scale.

SEQUOIINEAE. Sequoia. Attention has already been called to some of the anatomical features. The oval pits on the field (2�6) have an uppernbsp;and a lower border though the pore is not infrequently obliquely vertical.nbsp;Anatomically the wood of 8. sempervirens is considered by Gothan� to agreenbsp;more closely with that of Taxodiiini than with S. gigantea. The leaves ofnbsp;gigantea (=Wellingtonia) are 3-angled and decurrent, -5 mm. long; thosenbsp;on the fertile shoots are broader, shorter and imbricate. 8. sempervirensnbsp;bears 2-ranked linear, sessile, but not decurrent leaves with an abruptlynbsp;spinous apex. There is a resin-canal below the midrib and transfusion-tracheids form conspicuous lateral groups. The cones of the two species are .nbsp;of the same type, but those of 8. gigantea are larger and occasionally reach anbsp;length of 9�10 cm. (fig. 702).

CALLITRINEAE. The foliage-shoots, similar to those of some Junipers, are characterised by short decurrent scale-like leaves in alternate ternarynbsp;or decussate whorls; the ovate or globular cones are composed of a fewnbsp;valvate scales. Actinostrohus'^ (fig. 703, A). The very small leaves in ternarynbsp;whorls have a free apex and the slender shoots closely simulate those of somenbsp;species of Veronica and Thamnea depressa (Bruniaceae). The 6 cone-scalesnbsp;Me oblong, acute, and the base of the cone is invested by 6 rows of crowdednbsp;scales which gradually pass into the foliage-leaves: each scale has 2�3 winged

seeds. Callitris. Similar in habit to TeiracZ-i?!.*� (fig. 703, B): in some forms (C. arborea) the small leaves are closely appressed to the axis; G. glauca^ showsnbsp;a considerable range in the form of the leaves, and in G. rhomhoidalis the

shoots are especially slender. Saxton^ points out that the traoheids have a single row of separate pits and draws attention to the occurrence of horizontal

^ Saxton (10^); (10�). The wood of the Callitrineae requires more thorough investigation.

layer like that in the leaves of Callitris. Spirally thickened tracheids occasionally occur in the wood. The cones consist of 4 decussate thick and warty �'quot;alves, each scale bearing 7�8 winged seeds.

ABIETINEA.E. Pinus. The needle-like leaves on short shoots are a striking feature. In P. silvestris and many other species each short shoot

A nbsp;nbsp;nbsp;B

FiCr. 703. A, Acthiostrobus pyramidalis. Teiraclinis articulata. (After Saxton.)

bears 2 needles plano-convex in section. In P. monophylla (fig. 694, D) each shoot has usually a single sharp-pointed cylindrical leaf 4 cm. long. Three-needled Pines occur in N. America and the Himalayas but not in Europe.nbsp;In P. Cemhra, P. pence (the only European examples), P. Strobus, P. excelsanbsp;(fig. 704), P. koraiensis, etc., the short shoots bear 5 needles and each is triangular. Thomson^ regards the dwarf-shoot of Piniis as a derivative of a longernbsp;shoot with spirally disposed needles: he has recently described examples ofnbsp;dwarf-shoots in recent Pines bearing an abnormally large number of leaves.nbsp;In P. silvestris shoots with 3 leaves are not uncommon; in P. excelsa woundingnbsp;caused the development of as many as 15 needles on a single dwarf-shoot.nbsp;Jeffrey, on the other hand, regards the spur-shoot as a pi-imitive attribute ofnbsp;Ihe coniferous stock. The length of the leaf varies considerably and maynbsp;reach 30 cm.; the margin is generally entire, but in P. Cemhra the apicalnbsp;portion is finely serrate. The structure of Pine leaves^ is well known, butnbsp;reference may be made to the twin bundles in hard Pines, a distinction from thenbsp;single strand in the soft Pines^, the presence of a well-defined endodermis (fig. 694,nbsp;D, e) and the infoldings of the chlorenchyma^. There are two types of cone,nbsp;that of the Pinaster group in which the distal end of the woody seminiferousnbsp;scale is more or less pyramidal and has a central umbo (cf. fig. 785), prolongednbsp;in some species, P. ponderosa^, P. Jejfreyi^, etc,, into a short recurved spinenbsp;or, in P. Coxdteri, into a large curved spinous process. Another type, illustrated by the Strobus group, is distinguished by the flatter imbricate scalesnbsp;with the umbo at the tip of the free rounded margin (fig. 704). The cones ofnbsp;the latter type resemble those of Picea. The variation in the size and form ofnbsp;cones from the same tree in Pinus excelsa is worthy of notice from a palaeo-botanical standpointquot;.

Cedrus. The more slender needles of Cedrus, most of which occur in clusters on short shoots, are approximately triangular in section (fig. 694, E)nbsp;^nd have a single bundle and two canals next the lower epidermis. Thenbsp;persistent leaf-bases resemble those of Tsuga and Picea (fig. 706, B, H). Canals,nbsp;though not normally present in the wood, are induced by wounding; xylem-Parenchyma may occur in the late summer-wood and ray-traoheids are present,nbsp;but less prominent than in Pines and some other genera. The deciduousnbsp;nature of the mature cone-scales would lead one to expect their preservationnbsp;as fossils. A characteristic feature of the cones is the presence of radiallynbsp;disposed lines normal to the edge of the sporophylls�. Larix. The narrownbsp;iinear decurrent leaves, keeled on the lower surface, have a resin-canal close

^ Por anatomical features of hard and soft Pines, see Holden, R. (13�); Jeffrey ^nd Chrysler (06).

to the epidermis at each angle. There is a single vascular bundle but no thick-walled hypoderm. Bailey^ notes the occurrence of wood-parenchymanbsp;on the outer face of the summer-wood. The persistent subglobose or morenbsp;elongated cones, reaching a length of 10 cm. (fig. 705, G), are in most speciesnbsp;characterised by the large size of the bract-scales like those of Abies andnbsp;Pseudotsuga. Bommer calls attention to the superficial resemblance of Larchnbsp;cones to the fruit of Petrophila diversifolia^ (Proteaceae). Pseudolarixnbsp;resembles Larix in habit (fig. 705, A, B); the male strobili are umbellate;nbsp;the leaves linear lanceolate and up to 7 cm. long. The cones are ovate andnbsp;bear loosely imbricate, pointed deciduous scales; the bract-scale is shorternbsp;than in Larix. Picea. The short and narrow leaves are tetragonal or morenbsp;or less flat in section: a canal occurs below the vein or there may be two lateralnbsp;resin-ducts. The persistent bases of the leaves form prominent pegs (fig. 706,nbsp;D). The cylindrical or oval cones, reaching 16 cm. in length, consist of leathery,nbsp;concave imbricate scales; the bract-scales are concealed�. The stem of anbsp;variety of P. excelsa (var. tuberculata)^, characterised by large conical tuberclesnbsp;of cork, bears a close resemblance to that of Xanthoxylum (Rutaceae)�. Tsuga.nbsp;The leaves are flat, decurrent (fig. 706, B), similar to those of Taxus, Abies, andnbsp;some species of Picea-, but there is a single resin-canal below the vein and thenbsp;lamina is petiolate and not sessile as in Sequoia sempervirens. In some casesnbsp;the lamina is finely serrate. The cones closely resemble those of Picea exceptnbsp;in their smaller size; the bract-scales are usually concealed, but in T. Pat-toniana�^ they are longer and reach a length of 7-5 cm.

Pseudotsuga. The leaves resemble those of Tsuga, but there are two lateral canals and the lamina is attached by a narrow base to a slightly prominent leaf-cushion (fig. 706, C). The wood is like that of Larix but thenbsp;tracheids often possess spiral bands, and xylem-parenchyma occasionallynbsp;occurs. Resin-ducts are present in both normal and injured stems. Thenbsp;cones, S�^10 cm. long, resemble those of Tsuga but differ in the 3-prongednbsp;conspicuous bract-scale; they are pendulous and the scales are persistent.nbsp;Keteleeria. Shoots similar in habit to those of Abies; leaves flat, a canal atnbsp;each angle. Cones similar to those of Pinus Cembra: microstrobili umbellate.nbsp;Abies. Leaves usually flatter than in Picea and often larger; the apex maynbsp;be notched though this is not a constant feature�. There is no persistentnbsp;leaf-base (fig. 706, A). Two lateral canals� are generally present in the leaf,nbsp;but in some species the canals are median�. Transfusion-tracheids may form anbsp;ring round the phloem {A. magnifica). Cones large (in A. nobilis 25 x 10 cm.)nbsp;and in most species the bract-scales are conspicuous (fig. 705, C, F). The cones

* nbsp;nbsp;nbsp;Bailey (09).nbsp;nbsp;nbsp;nbsp;� Bommer, loc. cit. PI. ix. figs. 1G7, 168.

� For figures illustrating the range in form of cones, see Kirchner, Loew, and Schroter (06) pp. 152, 153.

� Barber (92). nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;� Gard. Chron. June 3, 1893, p. 659.

* nbsp;nbsp;nbsp;Ibid. p. 90.nbsp;nbsp;nbsp;nbsp;� Elwes and Henry (06) p. 717.

PlG. 705. A, B, Short shoot and cone of Pseudolarix Kaempferi. C, Abies bracteata, showing the long bract-scales. I), E, Abies concolor, cone-scale; b, bract-scale; s, seed. F. A6ie� Ercwcri, cone-scale; 6, bract-scale. G, LarixGriffitM',nbsp;b, bract scale; s, seed. (C�F, from the Gardeners� Chronicle.) M. S.

�\

�iG. 706. Branches of Abietineae. A, Abies pectinata. B, Tswga canadensis. C, Pseudotsuga Douglas�. D, Picea excelsa. E, Cedrus Lihani. F, Larixnbsp;europaea. G. Pseudolarix Kaempferi. (After Rendle from Veitch and afternbsp;Engler and Prantl.)

of some species with concealed bract-scales, e.g., A. violacea^, A. Webbiana^, A. concolor (fig. 70S, D, E), are hardly distinguishable from those of Pieea andnbsp;the larger cones of Tsuga, while those of Abies amahilis resemble Cedar cones.

PODOCARPINEAE. The wood agrees with that of the Cupressineae in the presence of xylem-parenchyma, but Gothan� states that the pores ofnbsp;the bordered pits on the radial walls of the medullary rays in the region ofnbsp;the spring-wood are narrow, elliptical, and vertical (fig. 693, K), while in thenbsp;Cupressineae the pore is horizontal (cf. fig. 693, H), a feature that is notnbsp;sufficiently constant to afford a trustworthy criterion. Podocarpvs. Leaf-lamina linear and short (P. nivalis, etc.) as in Taxus, longer and broader (in P.nbsp;elatus reaching a length of 29 cm. and 2 cm. broad), as in Cephalotaxus (fig.nbsp;694, A) and the pinnae of Cycas small appressed and scale-like (P. cupressina)nbsp;or ovate and provided with several veins (P. Nagi, 6x2 cm., P. Wallichiana,nbsp;13�15 X 3-5 cm.). The leaves are opposite in the section Nageia (fig. 707),nbsp;scattered in other species. Tison** has shown that in the broad-leaved speciesnbsp;only one vascular bundle is given off from the stele and this branches in thenbsp;petiole, not in the cortex as in Agaihis. The leaf of P. macrophylla^nbsp;(5�6 cm. X 7 mm.) represents a fairly common type: short reticulate tracheidsnbsp;occur on each side of the vein and elongated tracheids extend from the midrib tonbsp;the leaf-edge. There are three canals near the lower surface, but in some speciesnbsp;only one is present. P. formosensis^ (fig. 707) has leaves like those of P. Nagi,nbsp;but more rigid, thicker, and smaller; the epidermal cell-walls are thick andnbsp;elongated and 4�5 cells surround each stoma (fig. 707, a, b). The presencenbsp;of hypodermal mechanical tissue distinguishes Podocarpus leaves from thosenbsp;of Torreya, and another characteristic feature is afforded by the two kinds ofnbsp;accessory tracheids in place of the ordinarj^ transfusion-tissue in the great majority of leaves. Reference has aheady been made to the � cones.� Dacrydium.nbsp;The dimorphism of the foliage-shoots is illustrated in fig. 708. The verynbsp;simple type of megastrobilus is a striking feature (fig. 684, P). Microcachrys.nbsp;Leaves small, appressed, in decussate pairs As in Dacrydium there isnbsp;a single canal below the vein. Gothan notices the frequent occurrence of anbsp;single large pit in the field of the rays. The fleshy mulberry-like cone isnbsp;a peculiar feature. Saxegothaea (fig. 687). Lindley�, the author of the genus,nbsp;describes this Conifer as having the male flowers of a Podocarp, the cones ofnbsp;an Agathis, the fruit of a Juniper, and the habit of Taxus. There is a canalnbsp;below the vein, also lateral transfusion-tracheids but less numerous than innbsp;Podocarpus. The occurrence of pits in the horizontal walls of the ray cellsnbsp;is a character in which Saxegothaea differs from the genera comprised undernbsp;the general term Cupressinoxylon^. The cones have already been described.

Acmopyle. This generic name was given to an imperfectly investigated New Caledonian Conifer formerly known as Dacnjdium Pancheri. The sessile,nbsp;decurrenfc, falcate leaves on the lateral branches (1�Them, x 2-5 mm.) arenbsp;a characteristic feature, those on the main axis being small and soale-like^.

A nbsp;nbsp;nbsp;B

Fig. 708. A, Dacrydium elalum. B, Dacrydium araucarioides. (From specimens in the British Museum.)

Pherosphaera. The vegetative shoots resemble those of Microcachrys and some Lycopods. The genus is peculiar among the Podooarpineao in the absencenbsp;of an epimatium. Sinnott^ regards the megastrobili, composed of 2�S one-seeded scales, as �the last step in reduction,' but the difficulty is to discriminatenbsp;between simplicity as a primitive and as a reduction-phenomenon. Phyllo-cladus (fig. 675). The cladodes have a central rib and several vascular

^ Pilger (03) p. 117. Mr Sahni (Emmanuel College, Cambridge) has in preparation an account of Acmopyle based on material collected in New Caledonia by Mr Compton.

2 Smnott(13)p. 73; see also Masters (93) p. 6; Hooker. J. D.,/coH. Plaiit, 1882, PI. 1383; Baker and Smith (10).

strands, and stomata occur on both surfaces^ The small megastrobili consist of spiral or decussate scales, each with one seed. Miss Young^ mentionsnbsp;the occurrence of tracheids with spiral bands: a single large pit occupies thenbsp;field of a ray-cell as in Dacrydium Franklini^. The seed in the early stage ofnbsp;development shows signs of a rumino-ted endosperm.

TAXINEAE. The occurrence of spiral bands in the secondary tracheids is a family-character but similar tracheids are not uncommon in other Conifers.nbsp;Taxii^. The difference in leaf-form and in the arrangement of the fohage isnbsp;well illustrated by Kirchner and Schr�ter^. The leaves have recurved margins,nbsp;a mucronate acute apex, and a small median vein. Torreya. In habit thisnbsp;genus resembles Taxus but the leaves are longer (6 cm. in T. californica);nbsp;there is a resin-canal below the vein and two conspicuous stomatal grooves onnbsp;the lower surface with papihose epidermal cells (fig. 694, B): the midrib is notnbsp;prominent. Cephalotaxus. Leaves linear, more or less falcate, reaching anbsp;length of 12 cm. in C. Henryi (fig. 709), with a prominent midrib and one canal

(fig. 694, A); there are no stomatal grooves. Seeds like those of Torreya in size and in the thick fleshy sarcotesta and inner shell; the endosperm is notnbsp;ruminated. Rothert^ has described an interesting departure from the usualnbsp;structure in the stem of C. koraina (according to Beissner = C. pedunculatanbsp;var. fastigiata): a resin-canal occupies the centre of the pith and several shortnbsp;tracheids occur internal to the edge of the xylem-cylinder. There do notnbsp;appear to be any anatomical characters apart from the spiral bands of thenbsp;secondary tracheids by which the Taxineae can be distinguished from somenbsp;Cupressineae and other Conifers; xylem-parenchyma is said to be presentnbsp;in both spring and summer wood of Cephalotaxus drupacea, while it is um-epre-sented in Taxus cuspidata and Torreya nucifera^; the pits in the field varynbsp;in number and may be simple or bordered and in the latter case the pore isnbsp;obliquely vertical.

The determination of fossil Conifers is one of the most difficult tasks of the palaeobotanist. It is comparatively seldom that well-preserved cones are found in organicnbsp;connexion with the twigs that borenbsp;them and the cones rarely exhibit thosenbsp;features which are the best guides tonbsp;affinity. Excessive trust in superficialnbsp;similarity has frequently led to the employment of generic names suggestingnbsp;relationships which are thoroughly misleading. In comparing fossil and recentnbsp;forms authors are apt to confine theirnbsp;attention to the better-known types,nbsp;forgetting that it is often with the lessnbsp;familiar and geographically restrictednbsp;genera that extinct plants are mostnbsp;closely allied. Even the data suppliednbsp;by petrified wood are often insufficientnbsp;to enable the student to do more thannbsp;refer a specimen to some comprehensive genus based on characters sharednbsp;by several recent genera. Though itnbsp;is as a rule easy to distinguish betweennbsp;the wood of a Conifer and that ofnbsp;Cycads and Dicotyledons, the agreement between the xylem elements ofnbsp;many Cycads and those of the Arau-carineae is sufficiently close to affordnbsp;opportunity for error. The homogeneous structure of the secondarynbsp;wood of some Magnoliaceous genera,nbsp;e.g., Trochodendron and Drimys^ (fig.

710), closely simulates that of a Conifer, but the medullar}^ rays are approximately equal in breadth to the tra-cheids and the cells are more elongatednbsp;* Groppler (94); Solereder (99) p. 34; (08) p. 5.

vertically than in the rays of a Conifer. Attention has already been called to the difficulty of distinguishing between the foliage-shootsnbsp;of some Conifers, Dicotyledons and Lycopodiaceous plants. The

Dg. 711. A, Veronica Heciori; B, Callitris calcaraia; C, Veronica cupressoides; D, Alhrotaxis cupressoides; E, Crassula lycopodioides. M. S.

twigs reproduced in fig. 711 are examples of misleading resemblances, ^nd similar instances are cited by Bommer. The Conifer Podocarfusnbsp;dacrydioides was described by Banks, and Solander as Lycopodium

arboreum; Dacrydium Bidwi�lii^, D. FranJclini and D. araucarioides (fig. 708, B) recall some species of Lycopodium, and D. cupressinum^,nbsp;a dimorphic species, may simulate Lycopodium tetragonum. Thenbsp;long, linear, distichous leaves of some species of Podocarpus andnbsp;Cephalotaxus might, as fossils, be confused with the pinnate leavesnbsp;of Cycas; further, as Bommer points out, the leaves of Podocarpusnbsp;Blumei resemble those of Agaihis and the seeds are similar to thosenbsp;of Dehaasia media (Lauraceae). The Cupressus type of shootnbsp;occurs in Baccharis scolopendra as in other Dicotyledonous plantsnbsp;(fig. 711). The presence of a midrib in a linear Taxus-]ike leafnbsp;though usually easy to recognise is not always obvious, e.g., innbsp;Torreya nucifera the midrib cannot always be distinguished onnbsp;the upper face of the leaves. The recurrence of a similar habitnbsp;in many Conifers renders difficult the identification of vegetativenbsp;shoots, particularly as in fossil specimens the precise method ofnbsp;attachment of the leaves, their texture, and other features arenbsp;frequently unrecognisable. The tendency to dimorphism innbsp;many genera is another difficulty: exaynples of dimorphic shootsnbsp;are afforded hj Dacrydium laxifolium, D.Kirkii, D.elatium (fig. 708,nbsp;A), Juniperus chinensis, Araucaria excelsa, CaUitris glauca, Podocarpus cupressirwP, P. imbricata, Thuya occidentalis, etc. Allusionnbsp;has been made to the considerable variation in the length of leavesnbsp;on a single branch of different Conifers; in such a form as Crypto-meria japonica var. spiralis Sieb. and in similar varieties of othernbsp;genera the spirally twisted leaves, reminding one on a small scalenbsp;of the �wind-blown� Acanthus leaves on a Byzantine capital,nbsp;constitute a feature which might be regarded as of taxonomicnbsp;importance. The investigation of the cuticular membranes ofnbsp;Conifer leaves, as yet but little attempted, may supply usefulnbsp;criteria as in the case of Cycadean fronds.

CHAPTER XLIV.

The task of deciphering the fragmentary remains of Conifers is particularly difficult and no branch of palaeobotanical researchnbsp;makes greater demands upon the patience and self-control of thenbsp;student. As Saporta says, �Aucune �tude n�ouvre des perspectives plus �tendues, mais aucune aussi n�exige plus de r�serve et denbsp;tatonnements^.� The determination of impressions of ill-preservednbsp;vegetative shoots is often impossible and it is regrettable that manynbsp;authors have been too ready to employ generic names denotingnbsp;identity or close relationship with recent types on wholly inadequate grounds. A recent writer thus sums up the situationnbsp;created by an excessive faith in superficial resemblances and anbsp;lack of familiarity with existing representatives of the group:��nbsp;�Where a knowledge of reproductive parts is lacking, chaos reignsnbsp;supreme.� It may be added that impressions or casts of cones innbsp;many cases do not afford any real assistance. A comparison ofnbsp;the various forms of foliage-shoots and strobili met with amongnbsp;recent Conifers demonstrates the danger- of placing confidencenbsp;in external resemblance as a guide to affinity. It is seldom thatnbsp;reproductive organs are well enough preserved to enable us tonbsp;recognise features of primary systematic value. Though littlenbsp;has so far been done to test the value of epidermal characters asnbsp;aids to identification, such results as have been obtained^ favournbsp;the conclusion that this line of investigation promises to be lessnbsp;fruitful for Conifers than for Cycadean plants. Petrified wood ofnbsp;the Conifer type is abundant in plant-bearing strata from the laternbsp;Palaeozoic rocks upwards, and considerable pains ha*^e been takennbsp;to utilise to the full this source of information. Within widenbsp;limits anatomical characters are undoubtedly valuable, but thenbsp;recent tendency to subdivide comprehensive genera, which are

recognised as embracing several recent genera, into genera implying a limitation of affinity within narrower bounds has, I venture tonbsp;think, been carried too far. The investigation of fossil coniferousnbsp;wood, in spite of the disappointing quality of the data from thenbsp;systematist�s point of view is well worth attention. An examination of fossil wood from different geological horizons brings tonbsp;light many striking instances of a mixture in single plants ofnbsp;features now characteristic of distinct genera. It is the generalisednbsp;forms that throw light on the nature of the changes produced innbsp;anatomical structure in the course of evolution. The older andnbsp;more generalised types are of special importance to the student ofnbsp;phylogeny. The very difficult question as to the stock from whichnbsp;the Conifers are derived is too wide to be adequately discussed innbsp;a general treatise. It is probable that the Coniferales are mono-phyletic, the Araucarineae being the oldest representatives of thenbsp;group while the Podocarpineae are a closely related series. Thenbsp;widely held view that the Araucarineae are descended from Cordai-talean ancestors is by no means definitely established; it restsnbsp;mainly on anatomical evidence and the arguments based on anbsp;comparison of the reproductive shoots are far from convincing.nbsp;On the other hand those who favour a Lycopodiaceous ancestrynbsp;for the Coniferales are confronted with difficulties which, thoughnbsp;I venture to think they are not insurmountable, have not beennbsp;adequately met^. The suggested linking up of the Cordaitales,nbsp;through types in which the cylinder of secondary xylem is supplemented by separate primary strands of vascular tissue, withnbsp;Lyginofteris and other Pteridosperms leads to the inclusion of thenbsp;Coniferales among the descendants of an ancient Filicinean stock,nbsp;but here too the chain of evidence is incomplete particularly asnbsp;regards the lack of data as to the nature of the reproductive organsnbsp;of several Palaeozoic genera founded on anatomical characters.

The problem is still unsolved: the discovery of additional types and a more thorough comparative study of such data as wenbsp;possess may �nable us to see more clearly the paths along whichnbsp;evolutionary tendencies have operated, but the absence of recordsnbsp;of the vegetation of pre-Devonian times deprives us of the meansnbsp;of following to their common source the different phyla of vascularnbsp;^ For a useful summary of arguments see Burlingame (15).

plants which in the Permo-Carboniferous era had already advanced far beyond the simple ancestral forrns which the botanist seeksnbsp;but rarely finds.

The various examples of fossil genera founded on the anatomical features of vegetative organs are dealt with in a separate sectionnbsp;and not included with impressions in the descriptions of the severalnbsp;families, partly on the ground that it is rarely possible to demonstrate a connexion between the two sets of records and in part with anbsp;view to give a more connected account of the results so far derivednbsp;from a study of petrified wood. Cross-references to anatomicalnbsp;structure are given in the descriptions of vegetative and reproductive organs when there appear to be sound reasons for assumingnbsp;a generic or family connexion. The classification of woods is atnbsp;best provisional and the generic characters are far from constant.nbsp;The main point is that the student cannot afford to neglect thisnbsp;line of enquiry if he desires to obtain a comprehensive view ofnbsp;the changing combinations of structural features preceding theirnbsp;distribution among existing genera.

A comparison of recent Conifers and Cycads with their Mesozoic representatives brings out very clearly the fact that while on thenbsp;one hand the modern Cycads differ widely from the Cycadeannbsp;type which played a prominent part in Mesozoic floras, recentnbsp;Conifers on the other hand agree closely in their main featuresnbsp;with their Mesozoic ancestors. The Cycads as we know them nownbsp;are a more recent product of evolution than the Conifers thoughnbsp;it by no means follows that the Conifers in the wide sense arenbsp;the more ancient group.

The earliest attempts to identify petrified wood are summarised by Goeppertb Knowlton^, and other authors. Luidius (Lhwyd)�nbsp;at the end of the seventeenth century employed the general designation Lithoxylon, and the termination -xylon is still used in genericnbsp;names applied to fossil wood in conjunction with some prefix implying agreement in the more important anatomical featuresnbsp;with some recent genus or family. For woods exhibiting anbsp;combination of characters unknown in existing genera a distinctive

* Goeppert (50). nbsp;nbsp;nbsp;^ Knowlton (89^; Gothan (05).nbsp;nbsp;nbsp;nbsp;� Luidius (1699) A.

prefix is employed, e.g., Xenoxylon, and some authors make use of a name, e.g., Woodworthia, which does not indicate that thenbsp;diagnosis is based on anatomical features. Conwentz^ adoptednbsp;the method of adding the prefix Rhizo- to generic terms for woodnbsp;believed to belong to roots, and Felix^ and Lignier� have employednbsp;the prefixes Cormo- and Clado- for stem- and branch-wood respectively. It is, however, seldom that such differentiation is possiblenbsp;and it is questionable whether it is wise to attempt refinements ofnbsp;this kind. Barber^, in his critical paper on a species of Cuf ressino-xylon, calls attention to Strasburger�s description of an old moribund stem of Larix with root-like characters and Gothan� speaksnbsp;of a branch of Pinus silvestris with root-attributes. The differencesnbsp;between branches and the main stem are not sufficiently knownnbsp;even in the more familiar types to justify the use of the prefixesnbsp;Cormo- and Clado- in descriptions of fossil specimens.

The scientific study of fossil wood began with Nicol� and Witham whose work was rendered possible by methods of sectioncutting first employed, according to Nicol, by a Mr Sanderson, anbsp;lapidary. Opinions expressed by Nicol on methods of investigating petrified wood are still pertinent after a lapse of 80 years;�nbsp;�To pronounce with certainty whether a fossil Conifer be essentially different from any known individual of the recent kind, itnbsp;would be requisite to have a thorough knowledge of the structurenbsp;at least of all the different tribes of recent Coniferae; and yetnbsp;several distinct fossil genera have been indicated by a person whonbsp;has examined, and that too very superficially, only three slices ofnbsp;three recent Pines, differing not essentially from one another.�nbsp;In recent years the tendency has been towards a more detailednbsp;study of anatomical characters such as the distribution and formnbsp;of the pits on medullary-ray cells. The facts recorded in thenbsp;Chapter on Eecent Conifers illustrate the difficulty of arriving atnbsp;a thoroughly satisfactory classification of anatomical features thatnbsp;may serve as criteria in the identification of recent genera: evennbsp;in the case of well-preserved fossil wood we have as a rule to restnbsp;content with a generic name denoting a combination of charactersnbsp;met with in more than one existing genus. Moreover, as already

shown, recent work has tended to reduce the taxonomic value of certain characters such as the occurrence of ray-tracheids whichnbsp;are more widely distributed than has generally been supposed.nbsp;In this connexion a word may be added with regard to somenbsp;common sources of error in anatomical investigation. There isnbsp;the obvious danger of confusion between features due to petrifyingnbsp;agents or to decay before petrification and those present in thenbsp;living tree: the thickening of cell-walls, e.g., those of medullarynbsp;rays, has been shown in some cases to be a pathological phenomenon^. The partial obliteration of bordered pits by decay maynbsp;cause them to appear separate though originally in contact (fig.nbsp;475, B, Vol. III. p. 257). The recognition of pits on the tangentialnbsp;and horizontal walls of medullary-ray cells is often very difficult, andnbsp;negative evidence may be misleading. It is by no means alwaysnbsp;a siuLple matter to distinguish between true canals and canal-likenbsp;spaces formed by the destruction of groups of tracheids (e.g.,nbsp;Pityoxylon eiggense; fig. 725). In one case it has been shownnbsp;that leaf-traces traversing broad medullary rays were mistakennbsp;for horizontal resin-canals^. The spiral lines frequently seen onnbsp;the walls of petrified tracheids caused by the directive influence onnbsp;the structure of the membrane of the course of enzyme-action maynbsp;simulate the spiral bands characteristic of Taxus, Torreya, andnbsp;Cephalotaxus. These are a few of the pitfalls in the path of thenbsp;palaeobotanist, but despite the difficulties and the frequency withnbsp;which imperfect preservation prohibits complete diagnosis, thenbsp;investigation of fossil wood is well worth the attention of studentsnbsp;equipped with an intimate knowledge of recent Conifers. Thenbsp;unpromising nature of the material may be a deterrent, thoughnbsp;lignitic and other specimens not thoroughly petrified are amenablenbsp;to special treatment�.

In the account of recent Conifers attention is called to the significance of rings of growth: the subject has recently been exhaustively treated by Antevs^ and students should consult his memoir in the ProgressuS rei hotanicae for references to the literature. The

For methods, see Jeffrey and Chrysler (06); Hollick and Jeffrey (09) B. Gothan (09); Sinnott (09).

subject is interesting and beset with difficulties but well worthy of more thorough treatment than it has so far received. Thoughnbsp;petrified Coniferous stems are usually represented by the secondarynbsp;wood only, the phloem and cortical tissues are sometimes preservednbsp;and afford useful information. Examples of petrified phloem andnbsp;other extra-xylem tissues are described by Lignier and othernbsp;authors. In his description of silicified plants from Eranz Josefnbsp;Land, of Lower Cretaceous or Upper Jurassic age, Solms-Laubach^nbsp;includes some pieces of Coniferous bark showing patches of periderm alternating with secondary phloem consisting of sieve-tubes,nbsp;phloem-parenchyma, and fibres, also some stone-cells. Some ofnbsp;the sieve-tubes are shown in fig. 718, B, with well-preserved sieve-plates, a feature very rarely preserved. There is not enough woodnbsp;associated with the phloem and periderm to serve as a means ofnbsp;identification, but Solms-Laubach speaks of the bordered pits onnbsp;the tracheids and the pits of the medullary rays as indicatingnbsp;Pityoxylon or Cedroxylon.

A new generic name Vectia has been instituted by Dr Marie Stopes^ for a mass of petrified phloem which she compares morenbsp;especially with the phloem of recent Conifers: while recognising thatnbsp;the specimen cannot be assigned with confidence to a particularnbsp;group of Gymnosperms I venture to think it is almost certainlynbsp;a portion of a Cycadean stem.

Deference has already been made in the section devoted to the anatomy of recent Conifers to the relative importance and constancy of different characters from a taxonomic point of view andnbsp;this question need not be further considered. The method ofnbsp;classifying coniferous wood in general use is based on a schemenbsp;proposed by Kraus�. A modified form of this scheme was published by Schenk^ and more recently Penhallow�, Jeffrey, Lignier,nbsp;Gothan, Dr Stopes, and other authors have considerably extendednbsp;our knowledge. Dr Gothan�, whose memoir on the anatomynbsp;of Conifers contains much valuable information, employs severalnbsp;generic names denoting identity with recent types, and whilenbsp;admitting the great advance made by him and other workers in

this field, it is difficult to avoid a suspicion of overstraining the significance of certain anatomical minutiae beyond the limitsnbsp;of safety.

The great abundance of petrified wood in strata ranging from the late Palaeozoic through the Mesozoic and Tertiary formations,

often in places which have yielded few other plant fossils, is in itself a strong incentive to research in a department of palaeobotanynbsp;that has suffered from superficial work and hasty conclusions.nbsp;Petrified tree-trunks are among the most impressive monuments

of former ages: the petrified forests in the desert east of Cairo the vast accumulation of Triassic stems, some reaching a length ofnbsp;200 ft. and 7�10 ft. in diameter, over an area of 10 square, milesnbsp;in Arizona County^; the succession of Tertiary forests in 2000 ft.nbsp;of volcanic sediment exposed on the sides of Amethyst Mountainnbsp;in the Yellowstone Park� (fig. 712); and on a smaller scale thenbsp;Jurassic trees in the Portland quarries are a few of many strikingnbsp;examples of the wealth of material.

Bordered pits on the radial walls of the tracheids, if uniseriate flattened above and below, when in two or more rows alternatenbsp;and polygonal; separate and circular pits, though rare, may occur.nbsp;Rims of Sanio usually absent though their occasional occurrencenbsp;on the secondary tracheids in the cone-axis of recent species ofnbsp;the Araucarineae shows that they are not entirely foreign to woodnbsp;of the Araucarian type.

Xylem-parenchyma absent or rare and may be represented by resiniferous tracheids. Medullary rays uniseriate, rarely double;nbsp;horizontal and tangential walls smooth; there may be 1�15nbsp;small pits in the field, though whether they actually belong to thenbsp;walls of the ray cells or to the adjacent tracheids has not beennbsp;definitely determined in fossil species. The pits are simple ornbsp;bordered, circular or elliptical.

B. nbsp;nbsp;nbsp;Pitting on the radial walls of the tracheids of the commonnbsp;Coniferous tyfe ; the fits are sefarate and circular and, if in twonbsp;or more rows, offosite. Contiguous and more or less flattened fitsnbsp;occur sfasmodically on the tracheids of the ivood of the genera includednbsp;in this section. Well-develofed rims of Sanio usually occur on thenbsp;tracheids.

^ A good example of an Arizona tree-trunk is exhibited in the Plant-Gallery of the British Museum. See Ward (00?).

Xylem-parenchyma scattered through the wood and not confined to any particular region, often containing resin; the transverse walls may be thick and pitted. Medullary rays uniseriate, horizontal and tangential walls smooth (unpitted); there arenbsp;generally several small pits in the field though in some speciesnbsp;referred to this genus there may be a single pit. In the regionnbsp;of the spring-wood the pore of the apparently bordered ray-pitsnbsp;is more or less horizontal; but the form and position of the porenbsp;are variable.

III. nbsp;nbsp;nbsp;TAXOXYLON.

The same anatomical features as in Cupressinoxylon except that the tracheids of the secondary xylem have spiral thickeningnbsp;bands.nbsp;nbsp;nbsp;nbsp;Tertiary to Recent.

Xylem-parenchyma usually present and scattered, but � it is not so characteristic a feature as in Cupressinoxylon.

Medullary rays usually uniseriate; the pitting is confined to the radial walls as in Cupressinoxylon but in the region of the spring-wood the pore is oblique or more or less vertical; in the summer-wood the pits in the field are indistinguishable from those ofnbsp;Cupressinoxylon. There are often several pits in the field but innbsp;some species there may be one or two large simple pits in the field.

V. nbsp;nbsp;nbsp;PARACEDROXYLON.

A genus of doubtful affinity. Bordered pits on the tracheids usually separate; no rims of Sanio.

Xylem-parenchyma confined to wounded regions. Medullary-ray cells pitted only on the radial walls except in the injured parts of the wood where the other walls may be pitted. There are 4�6nbsp;circular, apparently bordered, pits with an oblique pore in the

field. In the characters of the normal wood this genus agrees most closely with Cupressinoxylon and Mesembrioxylon.

VI. nbsp;nbsp;nbsp;CEDROXYLON.

In some species included in this genus contiguous and flattened bordered pits are fairly common on the radial walls of the tracheids.

Medullary rays uniseriate; all the walls are pitted; tracheids may be present in the rays though they are usually absent. Therenbsp;are 1�6 or rarely more pits in the field, either simple or apparentlynbsp;bordered. Resin-canals confined to wounded regions.

VII. nbsp;nbsp;nbsp;PITYOXYLON. (Including Piceoxylon and Pinuxylon.)nbsp;Though in the great majority of cases the tracheids of the

secondary xylem have no spiral bands, the presence of such bands in the recent genus Pseudotsuga shows that this feature may occurnbsp;in wood of the Pityoxylon type.

Resin-canals present in the normal wood. Medullary rays of two kinds, uniseriate rays consisting of parenchyma and, in manynbsp;cases, ray-tracheids, also fusiform rays with horizontal resin-canals.nbsp;The walls of the ray-tracheids are either smooth or irregularlynbsp;dentate. All the walls of the medullary-ray cells are pitted;nbsp;there may be one large simple pit in the field or several small,nbsp;apparently bordered, pits.

VITI. PROTOPICEOXYLON.

Similar to Pityoxylon except in the absence of horizontal resin-canals in the normal wood; such horizontal canals as occurnbsp;are regarded as traumatic.

C. Genera in which Araucarian features, especially as regards the tracheal pitting, occur in association with characters tnet with innbsp;recent Abietineae. The genera included in this section afford examplesnbsp;of generalised types and do not resemble recent forms so closely as donbsp;the other genera.

IX. nbsp;nbsp;nbsp;WOODWORTHIA.nbsp;nbsp;nbsp;nbsp;�

Araucarian tracheal pitting; annual rings feebly marked. Short shoots and a subtending leaf are present in the secondarynbsp;wood. Resin-canals absent.

X. nbsp;nbsp;nbsp;ARAUCARIOPITYS.

Similar to Woodworthia in the possession of short shoots and in the pitting of the tracheids though separate and circular pitsnbsp;also occur. Vertical resin-canals abundant in wounded regions.nbsp;Medullary rays uniseriate; all the walls pitted.

XI. nbsp;nbsp;nbsp;PROTOCEDROXYLON. (Including Metacedroxylo7i.)nbsp;Bordered pits on the radial walls of the tracheids in 1�3 rows,

usually of the Araucarian type but separate pits also occur. No rims of Sanio and no resin-canals.

Xylem-parenchyma usually absent. Medullary rays generally uniseriate; all the walls pitted; 1�3 circular, simple, pits in thenbsp;field.

XII. nbsp;nbsp;nbsp;XENOXYLON.

Tracheal pits large, generally flattened above and below though not always, often transversely elongated. Resin-canals absent.

Medullary rays uniseriate; pitting confined to the radial walls; usually one large simple pit in the field.

XIII. nbsp;nbsp;nbsp;ANOMALOXYLON.

When uniseriate the pits on the tracheids are usually contiguous and flattened; if biseriate the pits are opposite; separate pits alsonbsp;occur.

True resin-canals absent, but canal-like spaces lined with small cells occur in some large medullary rays and constitute a characteristic feature.

Medullary rays uniseriate; pits confined to the radial walls; 2�3 circular simple pits in the field.

XIV. nbsp;nbsp;nbsp;THYLLOXYLON.

Xylem-parenchyma at the end of a year�s growth. No true, resin-canals, but the central parenchyma of some of the broadernbsp;rays is replaced by a canal-like space often filled with tyloses.

XV. nbsp;nbsp;nbsp;PLANOXYLON.

Wood with well-marked annual rings, resin-canals usually absent; tracheids with 1�3 rows of alternate hexagonal borderednbsp;pits on the radial walls and in the late wood there may be a singlenbsp;row of separate pits. Xylem-parenchyma occurs only between thenbsp;spring elements and the last-formed tracheids of the previous year.nbsp;Medullary rays almost entirely uniseriate; all the walls pitted.

The anatomical characters implied by the expression �wood of the Araucarian type� are enumerated in the Chapter on Recentnbsp;Conifers, and in Chapter xxxiii. reference is made to the impossibility of drawing a clear line of division between the wood ofnbsp;Araucarian plants and that of certain members of the Cordaitales^.nbsp;The name Araucarites was used by Goeppert^ for fossil wood ofnbsp;the Araucarian type, but the previous use of this name by Preslnbsp;for impressions of foliage-shoots and cones renders unsuitable itsnbsp;application to wood apart from the fact that Araucarites at oncenbsp;commits an author to a determination implying an affinity which innbsp;many cases cannot be demonstrated. Endlicher�s non-committalnbsp;genus Dadoxylon^ has been widely used, especially for Palaeozoicnbsp;wood having the characters of Araucaria or Cordaites. Thisnbsp;designation leaves open the question of precise systematic position.nbsp;In 1882 Kraus ^ instituted the genus Araucarioxylon, a name whichnbsp;has been widely adopted for fossil wood both from Palaeozoic andnbsp;later formations. The practice of limiting Dadoxylon to Palaeozoicnbsp;species and reserving Araucarioxylon for Mesozoic and Tertiary

Wood has already been criticised ^: it is pointed out that the application of an age-test is scientifically unsound and cannot fail to he misleading. Although it is probably true that species ofnbsp;Dadoxylon from strata later than the Ehaetic series are in thenbsp;majority of cases Araucarian, we have no adequate grounds fornbsp;definitely naming such types Araucarioxylon in the sense ofnbsp;membership of the Araucarineae. Similarly some Palaeozoicnbsp;species of Dadoxylon may well be more closely allied to the Araucarineae than to Cordaites: Dadoxylon valdajolense (Moug.)^ andnbsp;D. Rhodeanum (Goepp.)^ have both been referred to Walchia: thenbsp;latter species is included by Tuzson in the genus Ullmanniies. Wenbsp;do not know the lower geological limit of the Araucarineae, nornbsp;do we know when the Cordaitales became extinct. Tuzson^nbsp;subdivides wood of the Dadoxylon type into several genera including Pycnophyllum, to which is referred Dadoxylon Brandlingiinbsp;(Lind, and Hutt.), Ullmannites applied to the type recentlynbsp;named Eristophyton Beinertianum by Zalessky and previously included by Scott in Calamopitys, also to Dadoxylon Rhodeanumnbsp;and D. saxonicum, Pagiophyllites, including P. Jceuperianusnbsp;(Goepp.), and for Tertiary species the term Araucarites is used.nbsp;This nomenclature, based partly on age and in part on a supposednbsp;Connexion between the wood and foliage-shoots, is opposed tonbsp;sound principles and in some cases is at variance with the truenbsp;character of the species. AVhen evidence is available in support ofnbsp;a reference either to the Araucarineae or to the Cordaitales thenbsp;qualifying terms Cordaioxylon or Araucarioxylon may be addednbsp;after Dadoxylon. Such a species as Dadoxylon permicum Merck.�nbsp;18 one of many examples of a Dadoxylon that cannot be more precisely identified. Dadoxylon australe Cri� from Triassic rocks innbsp;New Caledonia must be included in the same category. Thisnbsp;species is founded on wood showing well-marked rings of growth;nbsp;there are two or rnore rows of alternate polygonal pits on the radialnbsp;i^alls of the tracheids and the medullary rays are uniseriate andnbsp;3nbsp;nbsp;nbsp;nbsp;15 cells in depth. Cri�s species should not be confused with

* nbsp;nbsp;nbsp;See p. 249, Vol. ill. See also Gothan (05) p. 14; Potonie (02) p. 229.

Dadoxylon australe Arber^ based on Palaeozoic wood from Australia: the substitution of Dadoxylon for Araucarioxylon, thenbsp;name used by Cri� for his New Caledonian species, necessitatesnbsp;a new specific designation for Arber�s type, which it is proposednbsp;to rename Dadoxylon Arberi. From Liassic beds in Yorkshirenbsp;Miss Holden^ has described a species as Araucarioxylon sp. whichnbsp;she suggests may be one of the oldest representatives of the Arau-carineae or perhaps a Jurassic example of Cordaites: the latternbsp;identification is supported by a reference to the recorded occurrencenbsp;by Lignier of an Artisia^ in French Jurassic beds, a test of affinitynbsp;that cannot be accepted as satisfactory.

There remains for consideration the debated question as to the value to be attached to the occurrence of contiguous and flattenednbsp;pits as an index of Araucarian affinity when this feature is associatednbsp;with a type of medullary-ray pitting foreign to Dadoxylon. Innbsp;Gothan�s genus Xenoxylon* the tracheids have usually largenbsp;flattened pits, but the pits on the radial walls of the medullary-ray cells are very different from those characteristic of Arauca-rineous wood. It is, moreover, not uncommon to find instances ofnbsp;contiguous and alternate pits on the tracheids of a stem in whichnbsp;the more usual type is the Abietineous arrangement. Gothan*nbsp;la3^s greater stress on the nature of the pitting on the walls ofnbsp;medullary cells, but Jeffrey� has discovered typical Abietineousnbsp;ray cells in the cone-axis of Araucaria and Agathis. Miss Holdennbsp;goes so far as .to maintain that the only feature which holdsnbsp;absolutely is provided by the rims of Sanio: these are invariablynbsp;absent in Conifers with Araucarian affinities except on the firstnbsp;few tracheids of the cone-axis of Araucaria and Agathis. Thisnbsp;author records as Araucarioxylon sp.� a wood (described on a laternbsp;page) from New Jersey possessing opposite pits on some of thenbsp;tracheids, also rims of Sanio. We cannot lay down any definitenbsp;rules with regard to the sporadic variation in tracheal or medullarynbsp;pitting or as to the relative value to be assigned to one or othernbsp;character. The statement by Thomson� that the ray cells of such

2 Holden, R. (IS^) p. 540, PI. XL. fig. 28. nbsp;nbsp;nbsp;^ gge p. 248, Vol. in.

^ See page 248, Vol. in. nbsp;nbsp;nbsp;� Gothan (05).nbsp;nbsp;nbsp;nbsp;� Jeffrey (12).

Araucarias as he examined have no pits even on the radial walls leads one to suspect the accuracy of some of the many recordednbsp;instances of fossil Araucarian wood with medullary-ray pits; butnbsp;this does not affect the value to be attached to the presence ornbsp;absence of well-defined pits on the vertical or horizontal walls ofnbsp;medullary-ray cells. It is generally true that in the ray cells ofnbsp;Abietineae these walls are pitted while in the Araucarineae theynbsp;are unpitted, and it is equally true that the predominance ofnbsp;alternate and contiguous pits on the tracheids is evidence of Araucarian affinity. Though generally absent from Araucarian woodnbsp;xylem-parenchyma occasionally occurs in stems otherwise identical with the usual Araucarian type. The genus Araucariofsis^nbsp;Was instituted by Caspary for specimens distinguished from mostnbsp;examples of Dadoxylon by the presence of scattered xylem-parenchyma but, as Gothan^ points out, a distinctive name is superfluous; the type-species of Araucariofsis, A. macractis, should benbsp;transferred to Dadoxylon; in the possession of xylem-parenchymanbsp;it agrees with Dadoxylon septentrionale Goth, from Spitzbergen.nbsp;The importance attached by Jeffrey and other American authorsnbsp;to the presence of Sanio�s rims on the tracheal walls is, I venturenbsp;to think, greatly overestimated. The determination of fossil woodnbsp;is to a large extent a question of relative values. There is clearnbsp;evidence, and it would be surprising were it not so, that in severalnbsp;extinct types there are combinations of character pointing to lessnbsp;sharply defined boundaries than those which delimit existingnbsp;families and genera. It is in the conclusions drawn from generalised types that authors differ. An outstanding fact is the predominance in Palaeozoic stems of the Araucarian form of trachealnbsp;pitting which is unquestionably a much older type than thatnbsp;characteristic of the Abietineae. The following definition is basednbsp;on specimens agreeing in the sum of their characters with recentnbsp;Araucarineae, but there are various genera described by Jeffrey,nbsp;Miss Holden, and other authors and believed by them to be morenbsp;cr less closely allied to the Araucarineae which are not providednbsp;for in the definition. These genera are treated separately asnbsp;generalised types and the decision as to the nature of the evidence

ttey afford with regard to the phylogeny of the Araucarineae and the Abietineae must be left to the individual student.

Annual rings occasionally well marked but not infrequently absent or indistinct. Tracheids with uniseriate and more or lessnbsp;flattened bordered pits or with two or more rows of alternatenbsp;polygonal pits on the radial walls. The alternate disposition,nbsp;even if unaccompanied by flattening and the polygonal contournbsp;of the pits, is a Dadoxylon feature if it is the dominant arrangementnbsp;and not a sporadic occurrence. Bordered pits occasionally occurnbsp;on the tangential walls but they are smaller and comparativelynbsp;rare. Rims of Sanio absent. Xylem-parenchyma usually absentnbsp;or feebly developed; resiniferous tracheids occasionally present.nbsp;Medullary rays uniseriate, very rarely double, homogeneous, 1�30nbsp;or as many as 50 cells deep; walls comparatively thin and withoutnbsp;pits on the horizontal and vertical walls; the radial walls may shownbsp;1�15 small pits, the oblique pore being occasionally enclosed in anbsp;feebly developed border. In view of the entire absence of pits onnbsp;the ray cells of at least some recent Araucarias the structure of thenbsp;ray cells in fossil stems requires careful revision^.

In the following brief descriptions of species of Dadoxylon a few examples are chosen to illustrate the wide geological and geographical range of fossil wood of this type, but it must be remembered that in many cases no positive statement can be made withnbsp;regard to the nature of the parent-plant beyond the facts affordednbsp;by the anatomical characters of the stem. Evidence bearing onnbsp;the geological age of the Araucarineae is discussed in the coursenbsp;of the description of genera founded on vegetative shoots andnbsp;reproductive organs. Species of Dadoxylon from Carboniferousnbsp;and Permian strata have already been described in Chapter xxxiii.nbsp;as more probably referable to Cordaites or at least to the Cordai-tales, and it is not by any means impossible that some of thenbsp;Dadoxylons recorded from Triassic or even higher strata maynbsp;belong to Cordaitalean species rather than to the Araucarineae.nbsp;The evidence afforded by petrified wood in conjunction with thatnbsp;� derived from vegetative remains lends probability to the view thatnbsp;Araucarian plants existed at least as early as the later Palaeozoicnbsp;^ See, in addition to Kraus and other authorities, Lignier (07^).

age. The lack of satisfactory knowledge with regard to the morphology of the reproductive organs of such genera as Walchia, Voltzia, and other plants closely resembling living Araucarias innbsp;the habit of their foliage-shoots precludes any definite statementnbsp;as to the precise degree of relationship between these and othernbsp;types and existing Araucarineae, though it is certain that thenbsp;Araucarineae were at least foreshadowed before the close of thenbsp;Palaeozoic era.

This species^, from Franconia and W�rtemberg, is considered fiy Schimper^ on the ground of association to be the wood ofnbsp;a Voltzia, and Tuzson^ adopts the generic name Pagiophyllitesnbsp;iruplying relationship with Pagiofliyllum; he includes in D.nbsp;iceuperianum the species Araucarioxylon wiirtembergicum Kr. andnbsp;A. thuringicum Born^. The tracheids of D. Jceuperianitni havenbsp;one or more rows of contiguous and more or less flattened pits;nbsp;the medullary rays are uniseriate and 2�50 cells in depth;nbsp;Tuzson figures 2�4 circular simple pits in the field. Thoughnbsp;possibly belonging to Voltzia or Pagiopliyllum this wood is bestnbsp;retained in Dadoxylon. Other Triassic Dadoxylons are describednbsp;tgt;y Wherry� from Pennsylvania: he records Araucarioxijlon vir-ginianum, a species described by Knowlton� from Potomac beds,nbsp;and A. vanartsdaleni: in both forms the tracheal pits are compressed and alternate, the rings of growth indistinct, and thenbsp;rnedullary-ray cells are said to have no pits. Reference hasnbsp;already been made to a Triassic species from New Caledonia,nbsp;-�. australe (Cri�).

This species'^, founded on material believed by Gothan to be Triassic in age, has the following characters:�Annual rings oftennbsp;*iistinct macroscopically but microscopically showing little contrast between spring- and summer-wood; bordered pits in a singlenbsp;row and separate or polygonal and in two alternate rows; they arenbsp;�ften arranged in stellate clusters as in some species, of Cedroxylon.

^ Gothan (10) p. 8, PI. nbsp;nbsp;nbsp;I.nbsp;nbsp;nbsp;nbsp;figs.nbsp;nbsp;nbsp;nbsp;4�8, PI.nbsp;nbsp;nbsp;nbsp;ii. fig. 1.

Medullary rays uniseriate, reaching a depth of 30 cells, 2�4 elliptical oblique pits in the field; wood-parenchyma, often withnbsp;dark contents, is not uncommon. Gothan emphasises the abundance of xylem-parenchyma as a character in which this speciesnbsp;differs from typical Dadoxylons.

Fliche^ described this species, from Liassic strata in Madagascar, under the name Araucarioxylon: the radial walls of the tracheidsnbsp;have two rows of contiguous and alternate pits; the medullarynbsp;rays are uniseriate, usually 8�16 cells deep, and small circularnbsp;pits occur on the radial walls.

An Oxfordian species from Normandy characterised by leaf-traces larger than those of Araucaria imbricata and, as Lignier^ states, suggesting leaves comparable in size with those of Cordaites.nbsp;The tracheids have 1�4 rows of pits and the medullary rays arenbsp;8�11 cells in depth.

Dadoxylon argillicola (Eichwald), recorded from Moscow�, is another example of a similar type of wood, and many other instances might be quoted in illustration of the wide distributionnbsp;of vegetative organs in Jurassic beds agreeing anatomically withnbsp;the Araucarineae.

A Cretaceous species from Amuri Bluff, New Zealand�, founded on a piece of decorticated stem 8 cm. in diameter preserved partly innbsp;silica and in part in a calcareous medium. The small pith, not morenbsp;than 1 mm. in diameter, is imperfectly petrified: the tracheidsnbsp;of the secondary xylem, which shows well-marked annual rings,nbsp;have biseriate, alternate, hexagonal, pits; there is no xylem-parenchyma. An interesting feature is the occurrence of tracheids onnbsp;each side of the medullary rays with thicker walls and containingnbsp;discs of resin: these resiniferous tracheids, similar to those described by Thomson� and other authors, afford particularly good

2 Lignier (07�) p. 257, PI. xvii. figs. 10�13. nbsp;nbsp;nbsp;� Eichwald (68) PI. v. fig. 12.

* nbsp;nbsp;nbsp;Stopes (14^) PI. XX.nbsp;nbsp;nbsp;nbsp;� Thomson, R. B. (13).

examples of this type of element. The uniseriate medullary rays are usually 3�4 cells deep and are described as having 5�6nbsp;bordered pits with oblique pores in the field. In view of Thomson�snbsp;conclusion with regard to the absence of pits on the ray cells innbsp;recent Araucarineae and considering the form of the pits as shownnbsp;in Dr Stopes� drawing it may be doubted whether the pits actuallynbsp;belong to the walls of the ray cells.

Dr Stopes comments on the scarcity of Araucarian remains recorded from New Zealand and adds a brief account of somenbsp;Tertiary wood described by Ettingshausen*^ as Araucaria Haasti,nbsp;a species founded on foliage-shoots which, without any adequatenbsp;reason, it is surmised belong to the petrified wood. The latter isnbsp;poorly preserved: Dr Stopes examined the type-specimen andnbsp;found that the bordered pits on the tracheids are circular and notnbsp;compressed; she expresses some doubt as to its Araucariannbsp;affinity, but renames the species Araucarioxylon Ettingshausem.nbsp;A specimen of wood from Amuri described by Ettingshausen asnbsp;Dammara Oweni^ appears to be undoubtedly Araucarian.

An interesting type of stem-wood has been described by Miss Holden� as Araucarioxylon sp. from the Cretaceous lignites ofnbsp;Cliffwood, New Jersey, which shows Araucarian characters in combination with certain anatomical features not usually associatednbsp;'with Araucarioxylon. The tracheids for the most part havenbsp;alternate compressed pits; the medullary rays consist of thin-walled cells and there is no xylem-parenchyma. Near the innernbsp;edge of the wood the tracheids are characterised by bordered pitsnbsp;in opposite pairs and rims of Sanio occur between adjacent pairs.nbsp;The specimen is said to supply the only missing link in the chainnbsp;of evidence pointing to the derivation of the Araucarineae fromnbsp;the Abietineae. Opposite pits are figured by Miss Holden in tracheids from the cone-axis of Araucaria Bidwillii and, as stated elsewhere, Sanio�s rims are not unknown in Araucarian wood. I havenbsp;adopted the name Dadoxylon because the characters as a wholenbsp;are consistent with that designation though it might be contendednbsp;that a new name is desirable to indicate the occurrence of unusual

features. The occurrence of opposite pits and rims of Sanio in the younger portion of the xylem is regarded as evidence in supportnbsp;of the view that the stem, while in the main Araucarian, exhibitsnbsp;features indicative of the origin of the Araucarineae from thenbsp;Abietineae. The presence of opposite pits in wood in which thenbsp;normal arrangement is alternate is not surprising if it is admittednbsp;that the Coniferous pitting is derived from an earlier scalariformnbsp;type. Even in stems in which the alternate or opposite pittingnbsp;is well established it is not very uncommon to find occasionalnbsp;departures from the normal pattern. This Cretaceous stem isnbsp;one of many generalised types, and the arguments based on thenbsp;admixture of characters in favour of the greater antiquity of thenbsp;Abietineae do not present any insuperable difficulty to the oppositenbsp;view namely, that the Araucarineae preceded the other families ofnbsp;the Coniferales.

This species from the Cenomanian of Normandy^ affords one of the few examples of the preservation of phloem, cortex, and pith.nbsp;Lignier adopted the generic name Arauca-riocaulon: he describes the tracheids asnbsp;having 1�3 rows of pits, crowded and alternate but not flattened. The medullary raysnbsp;are very short, 1�3 and rarely 4 cells innbsp;depth; there are said to be 8�15 pits in thenbsp;field (fig. 713). Resiniferous parenchyma isnbsp;abundant as in the wood referred by Casparynbsp;to a special genus Araucariopsis^ and innbsp;Dadoxylon septentrionale Goth. The phloemnbsp;includes well-preserved sieve-tubes and thenbsp;cortex is characterised by numerous sclerites, an Araucariannbsp;feature. This species appears to be an aberrant type the positionnbsp;of which is by no means clear.

Among other Cretaceous species are Dadoxylon albianum^ (Fliche) horn L�Aube; Dadoxylon Dantzii Pot.^ from beds probably of Upper Cretaceous age in East Africa, without definite rings

of growth and with a single row of contiguous pits on the tracheid walls; Dadoxylon virginianmn (Knowlton)^ from the Potomacnbsp;series; Dadoxylon bammianum (Fliche)^ from the Lower Cretaceous of France; Dadoxylon noveboracense (Holl. and Jeff.)� fromnbsp;the Middle Cretaceous beds of Staten Island; Dadoxylon Zuffardiinbsp;Negri ^ from middle Cretaceous rocks in the Gulf of Tripoli; Dadoxylon tankoense (Stopes and Fujii)� from Upper Cretaceous bedsnbsp;in Japan; Dadoxylon madagascariense (Fliche)� from Madagascar.

In 1881 Goeppert�^ mentioned some wood received from Kerguelen Island from Baron von Schleinitz, probably of Tertiary age, under the name Araucarites Schleinitzi et HooJceri, but, as Gothan�nbsp;points out, it is not clear whether he refers to one or two species,nbsp;and as there are no figures or full description Goeppert�s designationnbsp;cannot stand.

The sections on which this Kerguelen species is founded are in the British Museum�; they show the following characters:��nbsp;Annual rings narrow, often 15�20 tracheids broad, the summernbsp;Wood being frequently represented by only two rows of elements.nbsp;There are 1 or 2 rows of bordered pits on the radial walls of the

tracheids, contiguous, alternate, and often slightly flattened (fig. 714, B). The medullary rays have usually 5�8 ellipticalnbsp;pits in the field (fig. 714, A).

A species from Tertiary or possibly Upper Cretaceous rocks in Seymour Island (S. lat. 64� 16')^ agreeing closely with the woodnbsp;of recent Araucarineae. The annual rings are distinct; there arenbsp;1�2 rows of pits on the tracheids 10�12p, in diameter-; thenbsp;medullary rays, 2�10 cells deep, are usually uniseriate and therenbsp;are several small oblique pits in the field. Cross-bars^ (Muller�snbsp;�querbalken�) like those described in Araucaria hrasiliensis occurnbsp;in some of the tracheids.

Dadoxylon Doeringii Conwentz� is a Patagonian species of Sub-Oligocene age characterised by distinct annual rings; rays up to 40 cells in depth, with 1�2 pits in the field. Among other Tertiary species are Dadoxylon aegyptiacum Unger^, recorded fromnbsp;several localities in the Libyan desert; Dadoxylon Robertianumnbsp;(Schenk)� of Tertiary or possibly Cretaceous age from the province of Nagpur, India; Araucarioxylon Jcoreanum (Felix)� fromnbsp;Korea, characterised by the occurrence of a single row of continuous pits on the tracheids, is referred by Gothan to the genusnbsp;Xenoxylon and regarded as identical with X. latiporosum'^.

The name Cupressinoxylon^ or, as written by Kraus, Cupresso-xylon^, is usually applied to fossil wood exhibiting the following features;�Annual rings well defined, often narrow; vertical rowsnbsp;of parenchyma, often containing resin and recognisable by theirnbsp;dark contents even in transverse section (fig. 715, A), scatterednbsp;through the spring- and summer-wood. Bordered pits on thenbsp;tracheids usually separate and circular and if in more than one rownbsp;opposite; medullary-ray cells generally characterised b}^ thenbsp;presence of several small pits in the field. Used in this sense

Gufressinoxylon denotes wood similar to Cedroxylon except in the greater abundance of xylem-parenchyma and its occurrence innbsp;the spring- as well as in the summer-wood. The medullary raysnbsp;afford another distinction which according to some authors is morenbsp;trustworthy than the presence or distribution of the xylem-parenchyma. Go than who uses Cupressinoxylon in a morenbsp;restricted sense, lays stress on the pitting of the medullary cellsnbsp;as a distinctive feature: the pitting is confined to the radial walls,nbsp;or in other words there is no Abietineous pitting in Cupressinoxylonnbsp;in the stricter sense. The medullary-ray pits have a broadlynbsp;elliptical pore which is more or less horizontal at least in thenbsp;spring-tracheids�the Cupressoid type in contrast to the Podo-carpoid type in which in the spring-wood the pore is narrower andnbsp;more vertical, though in some Podocarps the bordered pits arenbsp;replaced by large simple pits: in the summer-wood the differencenbsp;between the Cupressoid and Podocarpoid type disappears. Innbsp;Cupressinoxylon the medullary rays are uniseriate and not verynbsp;deep though the depth is a variable character. Lignier^ statesnbsp;that 60�150 rays occur in 1 square millimetre, another featurenbsp;of doubtful value. In some species included in Cupressinoxylonnbsp;the pitting of the tracheids is partly Araucarian as it also isnbsp;in certain types of Cedroxylon. The presence of Sanio�s rims,nbsp;though not mentioned by many authors, is regarded by Jeffreynbsp;and some other American botanists as an important characternbsp;to be expected in all Coniferous wood other than that of thenbsp;Araucarineae. It is clear that unusually good preservation isnbsp;essential for the recognition of such features of the medullary-raynbsp;cells as Gothan includes in his definition of the genus; unless thenbsp;tissues are well preserved the generic separation of Coniferousnbsp;types except within very wide limits is impossible. The namenbsp;Cupressinoxylon may conveniently be restricted to wood havingnbsp;the usual type of tracheal pitting though pits of the Araucariannbsp;type may occur locally, with medullary rays in which the pittingnbsp;is confined to the radial walls and generally with several fairlynbsp;small and apparently more or less definitely bordered pits in thenbsp;field, the pores in the spring-wood being elliptical and more ornbsp;less horizontal. Resin-canals absent except in wounded regionsnbsp;1 Gothan (05) p. 39.nbsp;nbsp;nbsp;nbsp;^ Lignier (07*) p. 245.

of the wood; xylem-parenchyma normally abundant and not restricted to' the end of the year�s growth. The photographs

reproduced in fig. 715 of sections of a piece of petrified wood from Lough Neagh 1 in Ireland, probably of Upper Tertiary age, illustrate

very clearly the appearance of xylem- or resin-parenchyma as seen in transverse, radial and tangential section. The presencenbsp;of dark contents renders these cells conspicuous in transversenbsp;section (A) and both contents and cross-walls are seen in the longitudinal sections (B, C). In fig. 715, B, the opposite and scatterednbsp;bordered pits are shown on the radial walls of the tracheids withnbsp;an occasional tendency to an alternate arrangement. Goeppertnbsp;founded his species Pinites PritchardtP on wood from Lough Neagh,nbsp;the generic name being altered by Kraus to Cufressino'xylon'^.

Kraus� enumerates 46 recent species of Conifers with wood of the Cupressinoxylon type and Beust¹ mentions 37 species. Asnbsp;used by most authors Cupressinoxylon includes members of thenbsp;Cupressineae (in the wide sense as used on page 124), Podocar-pineae, Abies Webbiana, and some other recent genera. Asnbsp;already stated Gothan has essayed the difficult task of defining thenbsp;genus in such a way as to restrict the wood so named to a smallernbsp;number of recent Conifers, recognising as distinct genera certainnbsp;species previously included in Cupressinoxylon, e.g., Taxodioxylonnbsp;(= Cupressinoxylon Sequoianum Merck.�), Podocarpoxylon {= G.nbsp;Hookeri Arb.�), Xenoxylon (= Cupressinoxylon Barberi Sew.'^):nbsp;these and other species formerly included in Cupressinoxylon arenbsp;distributed among genera not always well defined but institutednbsp;with a view to increase the value of fossil species considered innbsp;relationship to recent types.

Miss Holden� has proposed a new generic name, Paracupressinoxylon for wood which agrees with Cupressinoxylon in having xylem-parenchyma, scattered through the year�s growth butnbsp;differing in the Abietineous pitting of the medullary-ray cells andnbsp;in the absence of Sanio�s rims. In Paracupressinoxylon are included two species, P. cedroides and P. cupressoides, both Jurassicnbsp;and the latter also Cretaceous in age, which appear to differ toonbsp;nauch from one another to be referred to one genus. Both species

Goeppert (50) p. 220. nbsp;nbsp;nbsp;^ Kraus in Schimper and Schenk (72) A. p. 376.

are regarded by the author of the genus as undoubted representatives of the Araucarineae, a determination that is hardly consistent with the affinity implied by the generic name. The species P. cupressoides, from Yorkshire and the Cretaceous lignites ofnbsp;New Jersey, is characterised by the restriction of pits to the radialnbsp;wall of the medullary-ray cells, the absence of Sanio�s rims, thenbsp;presence of scattered xylem-parenchyma, and by the occurrencenbsp;both of scattered and crowded pits on the tracheal walls. Moreover in this species the phloem shows an alternation of hard andnbsp;soft elements. The affinity suggested by these features wouldnbsp;seem to be to Cupressinoxylon. On the other hand, Paracupres-sinoxylon cedroides, founded on material from the Yorkshire coast,nbsp;is characterised by the Abietineous pitting of the medullary-raynbsp;cells, an admixture of scattered and compressed bordered pits onnbsp;the radial walls of the tracheids, scattered xylem-parenchyma, nonbsp;alternation of hard and soft bast, and by the absence of anynbsp;sclerous cells in the pith. This species also illustrates the occurrence of resin-canals in wounded regions of the wood. The reference of both these species to the Araucarineae, chiefly because ofnbsp;the absence of Sanio�s rims, though consistent with the principlenbsp;that this character is all important, implies the neglect of othernbsp;characters, more especially the nature of the medullary-ray pitting,nbsp;which in the case of recent Conifers are unquestionably of taxonomic importance. The species P. cedroides should not, in mynbsp;opinion, be included with P. cupressoides in one genus; it is probably more closely allied to the Abietineae than to any othernbsp;family. It should, however, be remembered that pitting of thenbsp;horizontal and tangential walls of medullary-ray cells is a featurenbsp;that is not confined to the Abietineae; it occurs also in some recentnbsp;Junipers and the extinct genera Protocedroxylon and Thylloxylon.

This Liassic species^ from Orne, France, is founded on the wood of a pentarch and hexarch root: the bordered pits on the radialnbsp;walls of the tracheids are usually in one row, occasionally in twonbsp;opposite series; smaller pits occur on the tangential walls; medullary rays 1�5 cells deep with a few ovoid-oblong pits, often simple

and rarely bordered, in the field. Eesin-cells are abundant. Lignier suggests that if the genus Cwpressinoxylon is subdividednbsp;this species might be referred to Glyptostroboxylon, though thenbsp;medullary-ray pitting is not consistent with the characters of thatnbsp;genus.

Founded on both stem and root wood from the Lower Greensand of the Isle of Wight and described by Barber^ with a thorough-

ness of detail rarely met with in accounts of petrified wood. Dr Stopes^ has recently re-described this species, adding a figure of thenbsp;niedullary-ray pitting. Annual rings well marked, often illustrating the occurrence of compound rings of growth (fig. 716) whichnbsp;are discussed in detail in the original account; bordered pits in a

single row, rarely double in the roots, free and circular in branches, often contiguous and compressed in roots; tangential pits common.nbsp;Medullary rays usually uniseriate, 1�16 cells deep, pits confinednbsp;to the radial walls, usually 1 but sometimes 2�4 oval and obliquenbsp;pits in the field. Resin-canals absent; resin-parenchyma innbsp;vertical rows, abundant and scattered. The pith consists of pittednbsp;parenchymatous cells separated by intercellular spaces; in the rootsnbsp;the rows of tracheids pass directly into the cells of the pith ; innbsp;the branches they terminate in small groups of cells irregularlynbsp;arranged.

This is one of several species from the Potomac lignites included by Knowlton^ in Cupressinoxylon. The annual rings are wellnbsp;marked; the tracheids have 2�3 rows of opposite and circular pitsnbsp;on the radial walls and small bordered pits are abundant on thenbsp;tangential walls. The uniseriate medullary rays, 2�49 cells deep,nbsp;have 1�2 oval apparently simple pits in the field and resin-parenchyma is abundant. Gothan^ has described some wood ofnbsp;LoAver Cretaceous age� from King Charles Land as Cupressinoxylon,nbsp;cf. C. McGeei, agreeing with Knowlton�s type in the medullary-raynbsp;pitting; there are 2�4 simple pits in the field, elliptical and horizontal ; an indication of a border was seen in some of the pits innbsp;the region of the summer-wood, but the general absence of a border,nbsp;if an original feature, is a difference between this wood and thatnbsp;of recent genera included in Cupressinoxylon. There is no Abie-tineous pitting on the ray cells.

The species Cupressinoxylon luccombense described by Dr Stopes^ from the Lower Greensand of the Isle of Wight closelynbsp;resembles C. vectense, but it has stone-cells in the pith, the tracheidsnbsp;are larger and there are usually 3�4 pits in the field in place of 1nbsp;or sometimes 2�3 in C. vectense] moreover in the latter speciesnbsp;the pits of the ray cells are more uniform in size.

� Gothan speaks of the King Charles Land fossils as Jurassic, but the beds have since been shown to belong to the Cretaceous system. See Burckhardt (11).

Fig. 717. Cupressinoxylon Hortii. Radial section showing the pits of the medullary rays, mp; the pits on the tracheids, and the wood-parenchyma, rp. (After Slopes.)

with small circular pits, forms projecting wedges in the perimedul-lary region; there are no resin-canals and resiniferous parenchyma is abundant; the medullary rays, generally 2�3 cells in depth,nbsp;have two circular pits in the field, a featnre regarded by Kraus asnbsp;indicative of afSnity to Glyptostrobus^. The cortex contains large

resin-canals. Dr Slopes draws attention to certain features, particularly the pitting of the medullary-ray cells and the structurenbsp;of the cortex, in which this species resembles the genus Crypto-mena.

This Lower Greensand species ^ from Bedfordshire is distinguished from other types by the very numerairs medullary rays, often separated from one another by a single row of tracheids,nbsp;which are both uniseriate and multiseriate, the same ray exhibitingnbsp;both forms at different heights. The rays attain a depth of 80nbsp;cells and there is generally one large oval or circular pit in the fieldnbsp;(fig. 717) though two are occasionally present. This species isnbsp;hardly a typical Cufressinoxylon and Dr Slopes points out thatnbsp;the large single pit in the field is suggestive of Podocarpoxylon: itnbsp;affords another illustration of the impossibility of identifying thenbsp;majority of fossil woods within narrow limits.

A species of Eocene age from the London Clay of Faversham in the Cambridge Botany School collection characterised by well-defined annual rings and the presence in some but not all the bandsnbsp;of summer-wood of resin-canals (fig. 718, C) which vary in size,nbsp;some of the larger being formed by the confluence of smallernbsp;adjacent canals. The presence of rather thick-walled, pitted, cellsnbsp;lining the canals or in close association with them is a characteristicnbsp;feature. Tyloses occur in some of the canals. The large numbernbsp;of canals in each row suggests their development in response tonbsp;traumatic stimuli. The bordered pits occur in single, separate,nbsp;rows or in double and opposite rows with rims of Sanio occasionallynbsp;preserved. In a few places the pits of a single row are in contactnbsp;and slightly flattened. Resin-parenchyma occurs in vertical rows innbsp;both spring- and summer-wood. Medullary rays uniseriate, reaching occasionally 30 cells in depth; the tangential and horizontalnbsp;walls are unpitted and on the radial walls the pits are preservednbsp;only in a few places; there appear to be 2�4 fairly large simplenbsp;pits in the field. The crowded series of canals (fig. 718, C) arenbsp;identical with the traumatic ducts described in Sequoia sempervirens^

' Stopes (15) p. 194, PI. xviii. text-figs. 56�58. nbsp;nbsp;nbsp;^ Jeffrey (03).

and Abies (cf. fig. 690, B, p. 130). The absence of Abietineous pitting in the ray cells, the distribution of the canals, and thenbsp;presence of scattered rows of xylem-parenchyma are featuresnbsp;indicating affinity to Sequoia sempervirens. The preservation isnbsp;not sufficiently good to warrant any definite statement with regardnbsp;to the pits on the radial walls of the ray cells: the absence of anbsp;border is in contrast to the pits in Sequoia, but the apparent lacknbsp;of a border may be due to imperfect preservation or to decay.nbsp;Miss Holden, who carefully examined the sections, called my attention to the occasional occurrence of obscure and narrow cells ofnbsp;unequal breadth on the edge of some medullary rays bearing anbsp;resemblance to the ghost-like ray-tracheids described by Thompson^.

Silicified wood is by no means uncommon in the Franz Josef Archipelago and several specimens have been found in talus-heapsnbsp;and in basaltic lavas. The age may be Upper Jurassic, Cretaceous,nbsp;or possibly Tertiary. A radial longitudinal section of a piece ofnbsp;wood is figured, though not named, by Newton and TealF andnbsp;without a full description. The following account is based onnbsp;sections cut from the same block in the possession of the Geologicalnbsp;Survey, which was collected at Northbrook Island by members ofnbsp;the Jackson-Harmsworth Expedition. The species is named afternbsp;Dr Koettlitz, the geologist of the Expedition�. Annual ringsnbsp;narrow and distinct: there are no resin-canals and no clearlynbsp;preserved xylem-parenchyma, though in a few places there arenbsp;indications of what appear to be elongated cells containing a darknbsp;substance. It is noteworthy that in some recent Cupressineae resin- //nbsp;parenchyma is not invariably present. The bordered pits on theMnbsp;radial walls of the tracheids are variable in their arrangement;nbsp;they occur in single rows (fig. 718, E), contiguous and sometimesnbsp;slightly flattened or more or less widely scattered, also in doublenbsp;rows with an opposite or occasionally an alternate disposition.nbsp;There are a few pits on the tangential walls of the tracheids and rimsnbsp;of Sanio are seen in places on the radial faces. The medullarynbsp;rays are 1�25 cells deep, uniseriate and very rarely two-cells

^ Newton and Teall (97) p. 508, PI. xli. fig. 11. nbsp;nbsp;nbsp;� Koettlitz (98),

I�lG. 718. A, D, Cupressinoxylon polyommatum. B, Phloem showing sieve-plates (Franz Josef Land). C, Cupressinoxylon Holdenae. E, Cupressinoxylon Koe.tl-litzi. (A, D, Dublin Museum; B, E, Geological Survey Museum, Jermynnbsp;Street; C, Cambridge Botany School.)

Cufressinoxylon polyommatum Cramer.

The petrified wood on which this species^ was founded was obtained from Banksland, from Miocene rocks, N. lat. 74� 40',

long. 122� W., during the voyage of H.M.S. Investigator in 1851. Through the kindness of Prof. Johnson I was able to have sectionsnbsp;cut from Dr Cramer�s specimen in the Dublin Museum.

Annual rings clearly marked; summer-wood narrow. The most striking feature is the irregular distribution and unusuallynbsp;large number of small bordered pits on the radial walls of thenbsp;tracheids. The pits vary in size and have an average diameternbsp;of 13-77ju.; they are usually separate but occasionally those of anbsp;single row are in contact. There are frequently as many as 3�4nbsp;rows (fig. 718, A) and occasionally 5 opposite and separate pits,nbsp;a feature suggesting comparison with Pinus MerkusiP, but in thatnbsp;species the pits of the three opposite rows are in contact. Kimsnbsp;of Sanio are occasionally present. Rows of narrow parenchymanbsp;occur in diflerent regions of the wood. Medullary rays uniseriate,nbsp;1�16 cells deep, with 2�4 simple large oval pits in the fieldnbsp;(fig. 718, D); there are no pits on the other walls of the ray cells.nbsp;The pitting in the field is very similar to that in Taxodium andnbsp;Cufressinoxylon {Taxodioxylon) Taxodii (fig. 720, A, B).

Under this name Conwentz^ describes some wood, probably of Pliocene age, from California which he compares with Sequoianbsp;sempervirens. He speaks of one stem 22 metres long and with anbsp;maximum diameter of 3�4 m.; the bordered pits are in 1�2 rowsnbsp;on the radial walls and small pits occur on the tangential walls;nbsp;the medullary rays are usually two-cells broad and have 3�4nbsp;generally elliptical pits in the field apparently simple and arrangednbsp;in a horizontal row. The rays are usually 15�20 cells deep butnbsp;may reach a depth of 56 cells. Resih-parenchyma occurs innbsp;vertical rows but it is not stated whether it is confined to anynbsp;definite region of the wood.

Schmalhausen�s species Cupressinoxylon (Glyptostrobusl) neosi-biricum^, characterised by medullary rays 13�20 or even 40�48 cells deep, and 1�2 circular or oval pits in the field, though compared by him with Glyptostrobus, cannot safely be regarded as morenbsp;nearly allied to that genus than to some other members of thenbsp;Cupressineae.

1 Groom and Rushton (13) p. 484. nbsp;nbsp;nbsp;^ Conwentz (78) Pis. xiii., xiv.

Many of the species included by authors are not described in sufficient detail to satisfy modern requirements with regard to thenbsp;structure of the medullary rays and other characters. From thenbsp;point of view of geographical distribution reference may be madenbsp;to Cupressinoxylon antarcticum described by Beust^ from Kerguelennbsp;Land.

	O	�
	�	o
o o	oo	oo
5	oo	O o
o o	\|oo	N-

Fig. 719. Cupressinoxylon tener (Kraus). (After Kraus.)

This generic name was first employed by Conwentz^ for wood from sub-Oligocene beds in Argentina which he described asnbsp;Glyptostroboxyhn Goepperti: no figures accompany the descriptionnbsp;and it is hardly possible to determine with accuracy the precisenbsp;affinity of the specimen. The annual rings are said to be distinct,nbsp;the pits on the tracheids uniseriate and contiguous; resin-parenchyma occurs, and the medullary-ray cells have large circular pitsnbsp;on the radial walls. It is suggested by Gothan� that this speciesnbsp;should be transferred to Podocarpoxylon, though in the medullary-ray pitting it differs from typical representatives of the Podocar-pineae. Kraus^ described the pits in the radialnbsp;walls of the ray cells of Glyptostrobus as largenbsp;and circular and in a Tertiary specimen fromnbsp;Niederw�llstadt, named Glyptostrobus tener Kr.,nbsp;he figures the pits in the field as simple (fig.

719); these are said to be 1�8 in number and they are arranged in horizontal series, a featurenbsp;characteristic of Taxodium.

Sections of wood of Glyptostrobus hetero-phyllus which I have examined show 2�8 pits in the field but there is a distinct border and thenbsp;pore is narrow and oblique or in the spring-woodnbsp;the pore may be broader and almost horizontal.

Xylem-parenchyma is scattered through the wood and the thick transverse walls are pitted. Gothan believes that the medullary-raynbsp;pitting of Glyptostrobus is distinguishable from the Cupressoid typenbsp;by the increase in the breadth of the pore as the ray cells arenbsp;followed from the summer- to the spring-wood and by the fact

that the pits in the region of the spring-wood may he simple. The same author^ points out that this type of medullary-raynbsp;pitting occurs in Cunninghamia. The impression produced bynbsp;an examination of the recent Glyptostrobus is that this so-callednbsp;Glyptostroboid pitting is not a sufficiently well-defined type tonbsp;serve as a trustworthy diagnostic character. The medullary-raynbsp;pitting in the region of the summer-wood is similar to that in somenbsp;species of Podocarpoxylon (= Mesembrioxylon), while in the spring-wood the pits in the field are rather of the type associated withnbsp;Cupressinoxylon and the scattered xylem-parenchyma is anothernbsp;characteristic of the latter genus. There would seem to be nonbsp;adequate grounds for regarding the two fossil species referred tonbsp;Glyftostroboxylon as more nearly related to Glyptostrobus than tonbsp;certain other recent genera. The retention of the name Glyftostroboxylon is inadvisable in that it implies an affinity which is notnbsp;supported by satisfactory evidence.

The generic name Taxodioxylon was applied to a Tertiary species from Hungary, originally referred by Felix^ to Rhizotaxo-dioxylon, on the ground of a resemblance in structure to the woodnbsp;of the recent genus Taxodium. Schenk�, who examined the type-specimen, confirmed this comparison. Taxodioxylon has beennbsp;adopted for fossils agreeing with the wood of Taxodium andnbsp;Sequoia sempervirens: Sequoia gigantea, on the other hand, agreesnbsp;. more closely with typical species of Cupressinoxylon. Taxodioxylon, while similar in most respects to Cupressinoxylon, is said tonbsp;differ in the medullary-ray pitting, the pits in the field beingnbsp;almost simple and elliptical with their long axis horizontal in contrast to the more definitely bordered pits of the Cupressoid type.nbsp;This distinction is, however, not entirely satisfactory: in the woodnbsp;of Taxodium the pits in the field are rather large and, though oftennbsp;simple, they occasionally present the appearance of pits with anbsp;well-developed border and the pore may be almost vertical ornbsp;horizontal. In the recent species, as in some fossil examples,nbsp;the tendency of the pits to arrange themselves in one or two

horizontal rows is a characteristic feature. The characters of Taxodioxylon may be summarised as follows:�Annual ringsnbsp;distinct; bordered pits on the radial walls of the tracheids in 1�4nbsp;rows, circular and separate and, if in two or three rows, opposite;nbsp;rims of Sanio present. Medullary-ray cells pitted only on thenbsp;radial walls; pits in the field, 2�8 in number, often arranged innbsp;horizontal rows (fig. 720, A), sometimes fairly large, simple, ornbsp;bordered and horizontally elliptical. The thick walls of the xylerh-parenchyma (fig. 720, B) are characteristic of Taxodium andnbsp;Taxodioxylon and this character has been quoted as a trustworthy

��

Fig. 720. A,Ji,Cupressinoxylon (Taxodioxylon) Taxodii. 0,Cupressinoxylon (Taxodioxylon) Sequoianum. (A, B, after Gothan; C, after Mercklin.)

distinction between Taxodium and Sequoia sempervirens though, as already pointed out, this is not a safe tesf-. In Glyptostrohusnbsp;heterophyllus the transverse walls of the xylem-parenchyma arenbsp;also thick and pitted and this tissue in some species of Cupressusnbsp;exhibits precisely similar features. It is in certain inconstantnbsp;features that Taxodioxylon differs from Gupressinoxylon. Innbsp;several instances the occurrence of wood referred to Taxodiq^ylonnbsp;in beds containing impressions of foliage-shoots like those ofnbsp;Taxodium and Sequoia sempervirens strengthens the conclusionsnbsp;based on anatomical characters.

In his description of the Tertiary beds at Senftenberg Potoni�^ identified some of the wood as Taxodium and compared the deposits with those of a Taxodium-sMaxup. Gothan^ has given anbsp;fuller description of the wood, which is characterised by xylem-parenchyma with thick transverse walls (fig. 720, B) and fairlynbsp;large elliptical pits in the field; in the region of the spring-woodnbsp;the medullary-ray pits have a horizontal pore and in the late woodnbsp;the pore is said to be vertical. The bordered pits on the verticalnbsp;walls of the xylem-parenchyma have a vertical pore in the summer-wood but it is horizontal in the spring-wood.

Gothan� recognised this species, originally described by Mercklin^ from an unknown locality, in Tertiary beds in Germany associatednbsp;with foliage-shoots agreeing with Sequoia sempervirens. The wallsnbsp;of the xylem-parenchyma are thinner than in C. Taxodii �, therenbsp;are 2�7 apparently simple pits in the field (fig. 720, C) in one ornbsp;two horizontal rows. There may be three rows of opposite pits onnbsp;the radial walls of the tracheids. It may be that C. Sequoianum isnbsp;the wood of a Sequoia. A closely allied species is represented bynbsp;Cupressinoxylon uniradiatum Goepp.� from Briihl afterwardsnbsp;recorded by Conwentz� as Rhizocupressinoxylon from Tertiary bedsnbsp;in Silesia. Schmalhausen�^ has described some interesting specimens of wood from Tertiary beds in Russia as Cupressinoxylonnbsp;Sequoianum characterised by the possession of elliptical simplenbsp;pits in the ray cells (from 2 to 6 in the field) of the transverselynbsp;elongated form characteristic of Taxodioxylon. Xylem-parenchyma is fairly abundant and the tracheids are peculiar innbsp;having three or rarely four bordered pits on the walls as in somenbsp;recent Pines.

Another example of wood of the Taxodioxylon type is afforded by T. palustre described by Felix� from Tertiary rocks in Hungarynbsp;and recorded also from sub-Oligocene beds in Silesia. Gothan�

assigns to Taxodioxylon two specimens from Tertiary strata in Spitzbergen and it is probable that the wood may belong to plantsnbsp;which bore the twigs described by authors as Taxodium distichumnbsp;and Sequoia Langsdorfii.

The name Taxodioxylon is retained as a section of Cupressino-xylon and not as a separate genus on the ground that the characters on which it is based do not appear to be sufficiently distinctive ornbsp;constant to warrant its recognition as a well-defined generic type.

Unger^ gave this name to fossil wood characterised by the presence of spiral bands in the secondary tracheids, a featurenbsp;especially associated with the recent genera Taxus, Torr�ya, andnbsp;Cephalotaxus, but by no means unknown in other Conifers^. Thenbsp;name, in the form Taxoxylum, was substituted for Taxites employed by Goeppert� for some species of Tertiary wood. Apartnbsp;from the presence of spiral bands Taxoxylon agrees with Cupres-sinoxylon, though according to Lignier the medullary rays arenbsp;deeper in the latter genus.

This Tertiary species, originally described by Goeppert from Hungary as Taxites scalariformis, was renamed hjXJngei: Taxoxylumnbsp;Goepperti. According to Schenk* it is the only species amongnbsp;those recorded by Goeppert which should be retained in Taxoxylon,nbsp;the spiral pattern of the tracheids being due to the presence of truenbsp;bands and not, as in the other species, the result of enzyme actionnbsp;on the wood which produces a spiral striation closely simulatingnbsp;spiral bands. The bordered pits on the ^racheids are circular andnbsp;separate; the medullary rays are uniseriate and from 1 to 10 cellsnbsp;in depth ; there are no resin-canals.

Goeppert�s species Taxites Ayckii (after Herr Aycke) of Tertiary age was retained by Kraus� as an example of that genus but afterwards transferred to Cupressinoxylon on the ground of thenbsp;absence of true spiral bands in the tracheids. Lingelsheim� also

states that there are no true spiral hands in Geoppert�s supposed Taxoxylon. The Permian species Taxoxylonginkgoidesnbsp;nbsp;nbsp;nbsp;and

Grand�Eury�s Upper Carboniferous species T. stephanense'^ are probably founded on wood of the Araucarian type in which thenbsp;tracheids show spiral striation.

An Aptiari species Taxoxylon anglicum Stopes� is referred to that genus because of the occurrence of a spiral marking on thenbsp;tracheids which the author of the species believes to indicate thenbsp;presence of true spiral bands in the wood of recent Taxaceae, andnbsp;because of the groups of 3�4 bordered pits in the fields of thenbsp;medullary rays.

There are no resin-canals in the wood; the tracheids have a single row of circular pits on the radial walls and occasional rims of Sanionbsp;are preserved. There are 1�6 pits in the field and the presence ofnbsp;a border is regarded by Dr Stopes as an argument in favour of thenbsp;tracheal nature of some of the medidlary-ray elements, though thenbsp;appearance of the cells does not afford any substantial ground fornbsp;interpreting them as other than parenchymatous elements.

An examination of the type-specimens convinced me that the spiral markings on the tracheids are not true bands like those ofnbsp;recent Taxineous wood and the pitting of the medullary-ray cellsnbsp;is in itself by no means a trustworthy criterion. There are, Inbsp;venture to think, no good reasons for referring this wood to thenbsp;genus Taxoxylon.

This generic name* is proposed for fossil wood exhibiting certain features associated with several recent genera which havenbsp;a southern distribution. It is intended to replace Gothan�s twonbsp;genera Podocarpoxylon and Phyllocladoxylon, types differing fromnbsp;one another in features which, as Dr Stopes� points out, are toonbsp;inconstant to justify the retention of both designations. Moreovernbsp;the use of Gothan�s names implies affinities to recent genera whichnbsp;there are no adequate reasons for assuming. In this instance, asnbsp;in many others, the anatomical characters do not enable us tonbsp;^ Renault (85) p. 163.

assign fossil species to a position within the Coniferales sufficiently definite to be denoted by the use of a name implying close relationship to a particular genus as distinct from a group of alliednbsp;types.

Podocarpoxylon. This name^ has been applied to wood agreeing in structure with recent species of Podocarpus and Dacrydiumnbsp;more closely than with other Conifers. As generally understoodnbsp;the genus stands for wood without resin-canals, possessing xylem-parenchyma not necessarily confined to a particular region ofnbsp;the year�s growth. In Podocarpoxylon aparenchymatosum Goth,nbsp;xylem-parenchyma is absent. The bordered pits on the tracheidsnbsp;are in 1�2 lows and, if in two series, the pits are opposite or subopposite (Podocarpoxylon Schwendae Kub.); rims of Sanio arenbsp;present. There is no Abietineous pitting in the ray cells; thenbsp;pits in the field are typically Podocarpoid, that is there aie few innbsp;the field and these appear to be bordered and characterised bynbsp;an elliptical or linear pore which is oblique or more or less vertical.nbsp;The medullary-ray pitting next the summer-wood does not affordnbsp;a satisfactory means of separating Podocarpoxylon and Gupressino-xylon, but in the region of the spring-elements the Podocarpoidnbsp;type is a distinguishing feature of Podocarpoxylon, though asnbsp;stated in the account of recent Conifers the position of the porenbsp;is by no means a constant character. On the other hand, the pitsnbsp;in the field may be large and simple as in Sciadopitys, some speciesnbsp;of Podocarpus, in Phyllocladoxylon, Xenoxylon and some othernbsp;genera.

Phyllocladoxylon. This name was given by Gothan^ to wood similar to Podocarpoxylon but differing chiefly in the occurrence ofnbsp;large, simple pits in the field [�Eiporen�], a feature shared withnbsp;Sciadopitys and some species of Podocarpus, by Microcachrys,nbsp;Dacrydium, and Pherosphaera. The tracheids have 1�2 rows ofnbsp;bordered pits on the radial walls, scattered and circular, but notnbsp;infrequently contiguous and flattened, and if in two rows they maynbsp;be alternate. There are no clear indications of Sanio�s rims innbsp;the specimens figured by Gothan and Schenk. There are no resin-canals and no xylem-parenchyma. Phyllocladoxylon agrees closelynbsp;with Xenoxylon^, but in Phyllocladoxylon the tracheal pits are oftennbsp;^ Gothan (05) p. 48nbsp;nbsp;nbsp;nbsp;^ Ibid. p. 55; (10) p. 37.nbsp;nbsp;nbsp;nbsp;� See page 238,

separate and are smaller than in Xenoxylon. The pits on the medullary-ray cells are also smaller, though it is doubtful if thisnbsp;is a constant character. Miss Holden^ has instituted a new genericnbsp;name, Paraphyllocladoxylon, for two specimens of wood fromnbsp;Jurassic rocks on the Yorkshire coast which do not appear tonbsp;differ from Mesembrioxyhn in any respect calling for genericnbsp;recognition. In ParaphyUocladoxylon eboracense, from the Oolitenbsp;of Scarborough, the tracheids have usually scattered and circularnbsp;pits on the radial walls and pits are also abundant on the tangentialnbsp;walls: in ParaphyUocladoxylon araucarioides the pits on the radialnbsp;walls are always closely compressed and flattened. There is nonbsp;Abietineous pitting and there may be one or occasionally twonbsp;large simple pores in the field like those in Xenoxylon and Mesembrioxyhn {� Phyllocladoxylon of Gothan) but smaller than thosenbsp;of Xenoxylon. Xylem-parenchyma is absent, but some tracheidsnbsp;have apparent cross-walls that are believed to be resin-plates.nbsp;Miss Holden recognises the close resemblance of her species tonbsp;Gothan�s Phylhcladoxylon, but a new name is employed on thenbsp;ground that the absence of Sanio�s rims shows that the wood ofnbsp;the Yorkshire plants is Araucarian. The absence of Sanio�s rimsnbsp;cannot be confidently regarded as an original feature and, assumingnbsp;this negative character to be a real one, it does not differentiatenbsp;the specimens from those described by Gothan; Gothan�s figuresnbsp;afford no evidence of the presence of Sanio�s rims in his species ofnbsp;Phyllocladoxyhn. If the Yorkshire stems are Araucarian so too arenbsp;those from King Charles Land and Seymour Island^. In one ofnbsp;Miss Holden�s species the tracheal pitting is not of the Araucariannbsp;type, while in the other it is Araucarian; the pitting of the medullarynbsp;rays is opposed to an affinity to any recent Araucarian Conifer.nbsp;Both of the Yorkshire species are therefore transferred to Mesem-brioxylon: their anatomical characters indicate that they arenbsp;generalised types which cannot legitimately be included in anynbsp;family based solely on existing Conifers. While recognising that itnbsp;is not always easy to draw a definite distinction between Xenoxylonnbsp;and Mesembrioxyhn the two names may be conveniently retained,nbsp;the former being used in a much more restricted sense than thenbsp;latter. Mesembrioxyhn is applied to woods in which the general

' Holden, R. (l.S^) p. 536, PI. xxxix. figs. 7�10. nbsp;nbsp;nbsp;^ Gothan (07^); (08).

features are similar to those associated with Cufressinoxylon, but the xylera-parenchyma may not be always present and the medullary-ray cells have one or two large simple pits, or two or morenbsp;smaller bordered pits, in the field, the pore being rather verticalnbsp;than horizontal. Mesembrioxylon undoubtedly includes speciesnbsp;which if additional data were available would be assigned tonbsp;distinct genera. Apart from the probability that anatomicalnbsp;characters were even less restricted in their range through differentnbsp;types in former periods than they are in existing genera, thenbsp;impossibility of discriminating between certain closely alliednbsp;recent Conifers points to the advisability of employing designationsnbsp;for fossil woods in a wide sense and thus avoiding the danger ofnbsp;misleading students in search of material on which to base conclusions with regard to the relative antiquity of existing genera.

This wood from Bathonian rocks in Russian Poland^ affords an example of the difficulty of distinguishing clearly betweennbsp;Podocarfoxylon and Glyftostroboxylon: the bordered pits on thenbsp;radial walls of the tracheids are separate or contiguous andnbsp;slightly flattened; xylem-parenchyma occasionally occurs; therenbsp;are usually two fairly large circular simple pits in the field (fig. 722,nbsp;C). The systematic position of this wood cannot be regarded asnbsp;well established.

Miss Holden^ has recently described two specimens from the Jurassic beds on the Yorkshire coast as Podocarfoxylon sp. butnbsp;the evidence in support of affinity to the Podocarpineae is not bynbsp;any means conclusive. In one specimen there is no xylem-parenchyma and in the other parenchyma occurs at the end of thenbsp;year�s growth. The pits on the medullary rays are described asnbsp;piciform, 1�2 in the field. The anatomical features describednbsp;hardly afford adequate reasons for assigning the wood to Podocarfoxylon rather than to Cufressinoxylon.

A specimen of imperfectly preserved wood from Jurassic rocks in the Izium district, in South Russia, described by Thomas� as

^ Gothan {06^} p. 456, fig. 5. nbsp;nbsp;nbsp;^ Holden, R. (13^) p. 542, PI. xl. figs. 31, 32-

Phyllocladoxyhn sp. may perhaps be included in Mesemhrioxylon though the preservation is hardly sufficiently good to admit ofnbsp;accurate determination. The bordered pits on the tracheids arenbsp;circular and separate, about 15p in diameter; the medullary raysnbsp;have one or rarely two large simple pits in the field.

An Aptian (Lower Greensand) species from Bedfordshire founded on two blocks of secondary wood and referred to Podo-carpoxylon^. The tracheids have 1�2 rows of bordered pits, thenbsp;pits in two series being opposite; Sanio�s rims are present. Eesin-parenchyma is abundant all through the wood; the medullarynbsp;rays are for the most part 3 cells deep but vary from 1 to 25; therenbsp;is one large circular or oval pit, or sometimes two, in the field,nbsp;and a narrow border is occasionally preserved. This species isnbsp;near to M. Schwendae but there are fewer pits in the field in thenbsp;English type.

This Aptian species^ is especially characterised by the arrangement of the bordered pits on the radial walls of the tracheids; the pits are uniseriate and occur in chains of 3�10, the border beingnbsp;flattened above and below by contact (fig. 721, lt;): the narrowernbsp;parts of the xylem-elements are often without pits. Xylem-parenchyma is scattered through the wood and the medullary-raynbsp;cells have an oval or nearly circular large pit, sometimes with anbsp;border (fig. 721, m, p), in the field. The contiguous pits constitutenbsp;an Araucarian feature though similar pits occur in Cedroxylon andnbsp;in some other genera.

Dr Stopes regards this species�, from the Aptian of the Isle of Wight, as highly suggestive of the genus Phyllocladus. Thenbsp;medullary rays are generally 2�4 cells deep and there are 1�2nbsp;large oval simple pits in the field. Xylem-parenchyma is sparselynbsp;scattered through the wood, and stone-cells occur in the pith.

This species is recorded from Attersee in Upper Austria and though probably of Tertiary age it may be derived from Cretaceousnbsp;strata^. Xylem-parenchyma is present; the bordered pits on thenbsp;radial walls of the tracheids are in 1�2 rows, usually separate butnbsp;if contiguous not flattened; if in 2 rows opposite or sub-opposite;nbsp;the medullary rays reach a depth of 13 cells; there are generallynbsp;1�3 pits (fig. 722, A, B) but occasionally as many as 5 in the field;

1 fo o
	oo '
o'	oo
\ .	od
	�7d
DO	Oo
O	o

A, B, Mesembrioxylon Schwendae. C, Mesembrioxylon sp. (A, B, after Kubart; C, after Gothan.)

they are bordered and the pore is obliquely vertical, though this feature is inconstant and in some places the pore is circular or thenbsp;bordered pit may be replaced by a large simple pit. Sclerous cellsnbsp;occur in the pith.

In this Tertiary species, included by Gothan in Podocarpoxylon, from Seymour Island ^ there is no xylem-parenchyma; the medullary-ray cells have 1�2 elliptical-circular pits in the field and thenbsp;Biajor axis is oblique. The evidence as to affinity is far from conclusive.

* Kubart (IP) PI. in. text-figs. 1�10. � Gothan (08) p. 8, PI. i. figs. 9�11.

Gothan^ described some wood as Phyllocladoxylon sp. which he originally stated to have been derived from King Charles Land,nbsp;but Nathorst pointed out that it came from Scoresby Sound innbsp;East Greenland, N. lat. 70� 50'. The pits on the tracheids are innbsp;1�2 rows, about 16p. high, and smaller than in Xenoxylon; theynbsp;are scattered or contiguous and flattened, but more often separate;nbsp;the medullary rays are composed of a small number of cells andnbsp;there are 1�2 simple pores in the field.

In this Tertiary species from Seymour Island, included by Gothan in Phyllocladoxylon^, the bordered tracheal pits are generallynbsp;uniseriate and separate; the pitting of the medullary-ray cells isnbsp;like that in Mesemhrioxylon sp. (fig. 722, C). In the account ofnbsp;this wood Gothan points out that the similar pits on the walls ofnbsp;the medullary rays of Sciadopitys are smaller than in the fossilnbsp;type; in contrast to the tendency towards a vertical elongationnbsp;of the ray pits in some recent Podocarpineae those of Mesembrio-xylon are usually more horizontally stretched as in Phyllocladusnbsp;and some species of Pinus. Gothan compares Cupressinoxylonnbsp;Hooheri Arb.� with Mesemhrioxylon antarcticum, but the former isnbsp;distinguished by the smaller pits on the ray cells and by thenbsp;occurrence of xylem-parenchyma. It is noteworthy that, asnbsp;Gothan states^, the only recent South American Conifer possessingnbsp;large simple pits in the medullary-ray cells is Podocarpus andina,nbsp;while on the other hand there are several Australian generanbsp;agreeing closely with the Seymour Island species in the characternbsp;of the medullary-ray pitting; from this it is concluded that innbsp;Tertiary times there was a closer connexion between the Southnbsp;American and Australian regions than at the present day, annbsp;inference which, though not improbably correct, rests on slendernbsp;evidence in this particular case.

This species from Pliocene strata in New South Wales, was described by Schenk� as Phyllocladus'. the pits on the tracheids

� Ibid. (08) p. i, PI. I. figs. 4�8. nbsp;nbsp;nbsp;� Arber (04).nbsp;nbsp;nbsp;nbsp;'* Gothan (08) p. 25.

are represented by Schenk as widely separated and the single pits in the field are narrower and more oblique than in other examplesnbsp;of the genus. The impression made by these species, formerlynbsp;referred to Phyllocladoxylon and distinguished from one anothernbsp;by no very well-defined characters, is that they agree with certainnbsp;recent Podocarpineae and with Sciadofitys more closely than withnbsp;any other recent Conifers; but in the absence of any definitenbsp;evidence with regard to foliage or reproductive organs it is impossible to select any one existing genus as the modern representative of the Arctic and South American fossil species. If thenbsp;absence of Sanio�s rims is accepted as a criterion of affinity, somenbsp;species of Mesembrioxylon would be included in the list of typesnbsp;allied to the Araucarineae, but even assuming that the preservationnbsp;of the wood is such as to admit of their recognition, were theynbsp;present, their absence does not nullify the evidence afforded by thenbsp;tracheal and medullary-ray pitting.

A species^ founded on a splendid specimen of silicified wood nearly 9 ft. long and with a diameter of 3 ft. from Tasmanianbsp;exhibited in the Geological Department of the British Museum.nbsp;The stem was discovered early in the nineteenth century in Tertiary basaltic lava on the Macquarie plains. Dr Arber quotesnbsp;Sir Joseph Hooker who gives an interesting account of the methodnbsp;of preservation of the decorticated wood. The annual rings arenbsp;well marked and narrow; the tracheids have usually a single rownbsp;of circular and scattered bordered pits on their radial walls andnbsp;smaller pits are abundant on the tangential walls. Sanio�s rimsnbsp;are clearly shown on some of the tracheids. The medullary raysnbsp;are generally uniseriate and in exceptional cases reach a dej)th ofnbsp;over 20 cells. Arber speaks of the occurrence of a small simplenbsp;pit on the radial wall of the ray cells; an examination of the sections in the British Museum showed that for the most part thenbsp;pitting on the ray cells is not preserved but in some places a single,nbsp;fairly large, simple pit occurs in the field. Resin-parenchyma isnbsp;present in both spring- and summer-wood.

It has been suggested by Gothan^ that this species should be referred to Podocarpoxylon: the pitting of the medullary-ray cells,nbsp;though seldom preserved, seems to differ from the typical Cufress-inoxylon form. It is therefore referred though with some hesitationnbsp;to Mesembrioxylon.

Paracedroxylon scituatense Sinnott. The generic name Para-cedroxylon^ was instituted for this Cretaceous species from Scituate, Massachusetts, in order to indicate its resemblance to Cedroxylonnbsp;as defined by Kraus, more especially as regards the pitting of thenbsp;tracheids and the absence of resin-canals and xylem-parenchymanbsp;in the normal wood, and to show that in some features it differsnbsp;from that genus, namely in the absence of Sanio�s rims and innbsp;the smooth and thin unpitted horizontal walls of the medullary-raynbsp;cells. In typical species of Cedroxylon xylem-parenchyma isnbsp;present and the tracheal pitting alone is not a distinguishingnbsp;feature. The annual rings are broad and not well defined; resini-ferous parenchyma is restricted to wounded regions. Groups ofnbsp;thin-walled cells, which it is suggested may be abortive resin-canals, though there is no evidence that this is the case, and thick-walled parenchyma occur in the wounded tissue. The tracheidsnbsp;have a single row of bordered pits, generally circular. The medullary rays, 2�12 or more cells in depth, are pitted only on the radialnbsp;walls and there are 4�6 circular pits with an oblique slit-like porenbsp;in the field. The occurrence of bands of much thickened andnbsp;pitted parenchyma is regarded as evidence of wounding: canal-likenbsp;spaces occur in the traumatic tissue. Moreover in the affectednbsp;regions the medullary-ray cells often show pitting on their horizontal and tangential walls.

Sinnott regards the absence of Sanio�s rims as indicative of Araucarian affinity while the traumatic phenomena are interpretednbsp;as Abietineous characters. The genus rests on a slender basis:nbsp;except for the absence of xylem-parenchyma the normal woodnbsp;differs very slightly from Cupressinoxylon and it is not distinguishednbsp;by any well-marked features from Mesembrioxylon.

This generic name was instituted^ for fossil wood agreeing with Ctipressinoxylon in the arrangement of the pits on the tracheidsnbsp;and in the absence of resin-canals, but differing in the scarcity ornbsp;absence of xylem-parenchyma. As defined by Schenk^, Cedroxylonnbsp;stands for fossil wood agreeing generally with that of recent speciesnbsp;of Cedrus, Abies, and Tsuga with or without tracheids in thenbsp;medullary rays. Brongniart�s genus Eleoxylon^ is included bynbsp;Schenk as a synonym of Cedroxylon. The chief distinguishingnbsp;character of Cedroxylon as compared with Cufressinoxylon, asnbsp;used by some authors, is the more restricted occurrence of xylem-parenchyma ; in Cedroxylon it is confined to the end of each year snbsp;wood whereas in Cupressinoxylon the parenchyma is not so limitednbsp;in its distribution. A closer examination of different types ofnbsp;wood included in Cedroxylon shows that the xylem-parenchymanbsp;is an unsafe guide: Barber* states that he found more xylem-parenchyma per square millimetre in Cedrus wood than in Crypto-meria (a genus included in the general term Cupressinoxylon) andnbsp;Lignier� speaks of the absence of parenchyma in some species ofnbsp;Cedroxylon. Gothan�, who has discussed the distinctive featuresnbsp;of these and other genera in considerable detail, points out thatnbsp;in Abies Webbiana xylem-parenchyma is abundant as in Cupressinoxylon, while in some Cupressineae the parenchyma is so scarcenbsp;that it is often difficult to discover. It is clear that a test basednbsp;on the presence or distribution of xylem-parenchyma is unsatisfactory ; the application of such a test would lead to the inclusionnbsp;of both Abietineous and Cupressineous genera in one generic type.nbsp;In typical cases the distribution of xylem-parenchyma is nonenbsp;the less a useful character, but Conifers with parenchyma scatterednbsp;through the year�s growth are not confined to types usually included in the comprehensive genus Cupressinoxylon'. many Podo-carps and some other genera not members of the Cupressineaenbsp;possess abundant parenchyma in the wood. The structure of the

^ Kraus in Schimper (72) A. p. 370. nbsp;nbsp;nbsp;^ Schimper and Schenk (00) A. p. 862.

inedullary-ray cells affords an important distinguishing feature. In Cedroxylon they are characterised by the occurrence of pits onnbsp;all the walls as in Abies, Tsuga, Cedrus, Larix, Picea, Pseudolarix,nbsp;while in wood of the Cufressinoxylon type there is no Abietineousnbsp;pitting but only pits on the radial walls. Another distinguishingnbsp;feature, mentioned by Lignier,�whether important or not is opennbsp;to question�is based on the greater number of medullary raysnbsp;per square millimetre in Cufressinoxylon.

The characters of Cedroxylon may be briefly summarised as follows;�Annual rings well marked; bordered pits on the radialnbsp;walls of the tracheids usually circular and separate and if in morenbsp;than one row, opposite, but in some species the Araucarian type ofnbsp;pitting also occurs (fig. 723), the pits being contiguous and alternatenbsp;or .sometimes arranged in stellate clusters. Xylem-parenchymanbsp;typically confined to the end of an annual ring, but sometimesnbsp;absent; medullary rays. generally uniseriate and composed exclusively of parenchyma though horizontal tracheids may occur;nbsp;pits on all the walls of medullary-ray cells as in the Abietineae;nbsp;on the radial walls there may be 1�6 apparently simple circularnbsp;pits in the field. There are no resin-canals except as the resultnbsp;of inj ary.

This species from Upper Jurassic rocks in Spitzbergen^ and from Lower Cretaceous beds in King Charles Land^ illustrates thenbsp;admixture of opposite, separate, and contiguous, alternate, pitsnbsp;in the same wood (fig. 723, A, B); stellate groups also occurnbsp;(fig. 723, A), an arrangement occasionally seen in some recent Pines,nbsp;e.g., Finns Merkusii^. The Araucarioid type of pitting is characteristic of the spring-wood where it is associated with the more usualnbsp;opposite and separate arrangement. Smaller separate pits occur onnbsp;the tangential walls of some of the tracheids. Resin-canals arenbsp;absent, though a solitary example is recorded by Gothan. Thenbsp;medullary rays have several small pits in the field, also pits on thenbsp;tangential and horizontal walls; the xylem-parenchyma is confinednbsp;to the end of the year�s wood; the cross-walls are pitted (fig. 723, B).

Gothan makes no mention of Sanio�s rims. The Abietineous features predominate over the Araucarian, the latter being limitednbsp;to the local occurrence of polygonal and alternate bordered pits.

tnT B

Wood of similar type was described by Schroeter^ from King Charles Land as Pinus {Larix) Johnseni: resin-canals, possiblynbsp;due to wounding, occur in the summer-wood. The medullary raysnbsp;are 1�18 cells deep and there are 1�3 simple circular pits in thenbsp;field; all the walls of the ray cells are pitted. As in Cedroxylonnbsp;transiens the Araucarian type of pitting is represented on some ofnbsp;the tracheids.

An Upper Jurassic species from Helmsdale^, Sutherland, a locality from which Hugh Miller recorded numerous specimens ofnbsp;fossil wood which is still abundant on the beach immediately northnbsp;of Helmsdale. The annual rings are well defined: the borderednbsp;pits are usually in a single row on the radial walls of the tracheids,nbsp;occasionally in contact and flattened; double rows of oppositenbsp;pits are not uncommon. Xylem-parenchyma is confined to thenbsp;late wood. Medullary rays, 1�26 cells deep, generally 8�12,

and uniseriate; there are 2�4 simple or faintly bordered circular pits in the field and pits occur also on the tangential and horizontalnbsp;walls.

			1 c
o o		Q	o
	lt;5 o	O o2	O nbsp;nbsp;nbsp;O
o	Ilixr	o nbsp;nbsp;nbsp;o
A.			o o
A.	�� - O o		c
			o

Fig. 724. A, Brachyphyllum eatkiense; stoma. B�F, Cedroxylon Hornei. (After Seward and Bancroft.)

In this species from King Charles Land^, probably Lower Cretaceous, the pitting of the tracheids is in the main of the usualnbsp;Coniferous type and not Araucarioid, though in places the pits are

alternate and form stellate groups. The medullary rays, reaching a depth of 30 cells, show very clearly the Abietineous pitting andnbsp;there are 4�5 simple circular pits in the field. The occasionalnbsp;occurrence of single large pores in the field would seem to be due,nbsp;at least in part, to the destructive action of fungi. Xylem-parenchyma occurs in the summer-wood. The structure agreesnbsp;with that of the wood of Cedrus, Pseudolarix, and Tsuga; thenbsp;Abietineous features are relatively more conspicuous than innbsp;Cedroxylon transiens in which the tracheal pitting is more Arau-carian.

In this wood^, from the Lower Greensand of Kent, the rings of growth are well marked; the bordered pits on the radial wallsnbsp;of the tracheids are usually uniseriate and Sanio�s rims are distinctly preserved. Xylem-parenchyma is absent or very rare andnbsp;there are no resin-canals. The medullary rays are seldom deepernbsp;than 10 cells; there are 4�6, or occasionally more, oval ornbsp;circular pits in the field and some have a slit-like pore and arenbsp;bordered; pits are clearly shown on the tangential walls of thenbsp;ray cells.

This species^ of the same geological age, from Potton in Bedfordshire, differs from C. maidstonense in the comparative abundance of xylem-parenchyma: the medullary-ray cells showquot; very clearlynbsp;the Abietineous type of pitting.

In this species from the gault of Bl�ville (Seine-Inf�rieure)^ the tracheal pits are usually uniseriate and separate but if in twonbsp;rows they may be either opposite or alternate. The pits in the fieldnbsp;are small, numerous, and have an oblique pore (fig. 723, C). Thenbsp;tangential walls of the ray cells are sometimes pitted. There isnbsp;no resiniferous parenchyma. The characters afford another illustration of the impossibility of drawing any clearly defined linenbsp;between Cedroxylon and allied generic types.

1 Stopes (15) p. 149, PI. xii. text-figs. 41�43. nbsp;nbsp;nbsp;^ jhid, p. 154, text-fig. 44.

= Lignier (07q p. 267, PI. xviii. figs. 1.5�17; PI. xxi. fig. 66; Pi. xxii. fig. 72; PI. xxiTi. fig. 87.

A Triassic species described by Wherry^ as Brachyoxylon penn-sylvanicum may, as that author suggested, belong to Cedroxylon: the tracheids have 1�2 rows of pits, usually separate but sometimesnbsp;alternate and hexagonal as in Cedroxylon transiens and C. Hornei;nbsp;no description is given of the medullary-ray pitting or of anynbsp;xylem-parenchyma. There are no adequate grounds for referringnbsp;this Triassic wood to Cedroxylon. Several species of wood fromnbsp;Triassic and higher horizons have been assigned to Cedroxylon,nbsp;but in many cases the descriptions fall short of modern standardsnbsp;and accurate determination is impossible. Cri�^ describes anbsp;species, C. australe, from the Trias of New Caledonia thoughnbsp;his figures and descriptions do not afford satisfactory evidencenbsp;in support of this reference. Schenk^ mentions Cedroxylonnbsp;pertinax (Goepp.) as the oldest representative of the genus andnbsp;speaks of it as Rhaetic, while Gothan refers the species to a Jurassic horizon. A species founded by Goeppert and described bynbsp;Mercklin^ from Jurassic.rocks of Russia, Pinites jurassicus, maynbsp;be a Cedroxylon: the bordered pits are usually separate andnbsp;opposite but sometimes in contiguous groups. Mercklin statesnbsp;that small thick-walled cells, often with dark contents, occur atnbsp;the outer limit of each ring. A specimen described by Felix� asnbsp;Cormocedroxylon jurense from the Braun Jura of Galicia is compared by him with Pinites jurensis. Fliche� records, though without complete diagnoses, some French Lower Cretaceous species:nbsp;the tracheids of Cedroxylon reticulatum Sap., from the Albian ofnbsp;L�Aube, are characterised by pits which are usually separate butnbsp;may be contiguous and flattened. Cones closely resembling those ofnbsp;Cedrus occur in the same beds. This author gives partial descriptions of C. barremianum Fliche� from the Lower Cretaceous ofnbsp;Haute Marne and a Cenomanian species C. manekildense Fliche�,nbsp;but in neither case are the data adequate.

C. matsumurae Stopes and Fujii� is an Upper Cretaceous Japanese species with 1�2 rows of tracheid-pits, generally opposite

1 Wherry (12) PI. iv.

^ Schimper and Schenk (90) A. p. 871.

* Mercklin (55) p. 48, PI. vni. figs, C�10.

Fliche (97) p. 7.

8 Ibid. (96) PI. XV. fig. 3.

8 Slopes and Fujii (10) p. 42, PI. i. fig. 10; PI. iv. figs. 20�23,

but sometimes alternate though not contiguous. The medullary rays, 5�12 or rarely 20 cells deep, are imperfectly preserved.nbsp;Another species, C. Yendoi St.^ and Fuj. from the same locality isnbsp;also founded on material that is insufficient for accurate determination. Sporadically occurring resin-ducts are regarded asnbsp;traumatic.

Among Tertiary species reference may be made to Cedroxylon affine Kraus ^ from Sicily, without resin-parenchyma and characterised by usually two large simple pits in the field; C. Hoheneggerinbsp;Felix� from the Eocene of Moravia figured by Schenk as fromnbsp;Cretaceous strata; C. Hermanni Sch.^, an incompletely describednbsp;species from Assam, probably of Tertiary age.

Kraus� included in this genus some of the species previously referred by Goeppert to Pinites; others he assigned to Cedroxylon.nbsp;Pityoxylon is distinguished from Cupressinoxylon and Cedroxylonnbsp;by the normal occurrence of resin-canals in the wood and by thenbsp;presence of horizontal tracheids in some of the medullary rays.nbsp;Within the limits of the genus the following differences occur innbsp;the characters of the medullary rays and the resin-canals;�thenbsp;walls of the ray-tracheids are smooth or provided with dentatenbsp;ingrowths; the pits on the medullary-ray cells are large and simplenbsp;or smaller and apparently bordered, and there may be one ornbsp;several pits in the field; the parenchyma of the resin-canals hasnbsp;thin or thick walls. As generally employed Pityoxylon includesnbsp;species exhibiting anatomical features met with in Pinus, Picea,nbsp;Larix, Pseudotsuga, and some other Abietineae. Gothan� makesnbsp;use of two generic names, Piceoxylon and Pinuxylon, to denote thenbsp;possibility of more precise comparison with recent types than isnbsp;implied by Kraus�s more comprehensive term. Piceoxylon isnbsp;characterised by thick-walled epithelial cells lining the resin-canals,nbsp;by small pits in the ray cells, spiral tracheids in the summer-wood,nbsp;the absence of teeth in the ray-tracheids, clearly marked Abietineous

pitting in the ray cells, and by the occurrence of numerous pits in the tangential walls of the summer tracheids.

Pinuxylon is used by Gothan in preference to Pinoxylon, the name adopted by Knowlton^ for wood in which there are no resin^nbsp;canals in the medullary rays. In Pinuxylon the walls of thenbsp;epithelial cells are thin, rarely thick; the medullary rays havenbsp;large simple pits in the spring-wood; there are no spiral bands innbsp;the tracheids. The horizontal tracheids have smooth or dentatenbsp;walls and the Abietineous pitting is much reduced. The distinctionsnbsp;on which these two genera are based are thus not very clearlynbsp;defined and it is only in particularly well-preserved materialnbsp;that the two generic types can be recognised with certainty.nbsp;Dr Stopes^ follows Jeffrey and Chrysler� in regarding Gothan�snbsp;twofold division as unnecessary.

In the majority of species referred to Pityoxylon the published information is insufficient for a sub-division in Gothan�s sense andnbsp;as a rule the generic name stands for wood of an Abietineous typenbsp;which cannot be assigned with confidence to any one recent genus.nbsp;The question of the antiquity of the Abietineae has been confusednbsp;by the too liberal use of the term Pinites by Goeppert and somenbsp;other authors for stems which have no claim to be placed in thenbsp;genus Pityoxylon. Jeffrey and Chrysler^, who follow previousnbsp;authors in quoting Pinites Conwentzianus Goepp.�, described as anbsp;Carboniferous species from Waldenburgh, as evidence of a Palaeozoic Pinus-like wood, state that the species receives �full confirmation from the description of a similar type, Pityoxylon chasensenbsp;Pen.� from the Permian of Kansas.� Goeppert and Stenzel statenbsp;that Pinites Conwentzianus was found on a rubbish-heap (�Halde�),nbsp;but Goeppert apparently entertained no doubt as to its Carboniferous age. Through the courtesy of Prof. Freeh of Breslau I wasnbsp;able to examine the original sections and convinced myself thatnbsp;the wood is Abietineous: the rings of growth are well defined;nbsp;horizontal tracheids occur in some of the rays and the trachealnbsp;pits, 1--2 rows, are widely separated, though occasionally the pits

of a single row are in contact and slightly flattened. Feeling sceptical as to the Carboniferous age of the wood I wrote to Drnbsp;Conwentz who confirmed my doubts with regard to the value of thenbsp;evidence as to the geological horizon. Thomson and Allin havenbsp;shown that Penhallow�s Pityoxylon cannot be accepted as trustworthy evidence of the occurrence of a Palaeozoic Abietineousnbsp;type. Pityoxylon chasense is not an Abietineous species; it isnbsp;founded on Dadoxyhn wood devoid of annual rings and withoutnbsp;resin-canals traversing the medullary rays.

The fragments of wood from the Muschelkalk of Recoaro figured by Schleiden and Schenk^ as Pinites Goep-pertianus affordnbsp;no evidence of Abietineous affinity beyond the occurrence ofnbsp;separate bordered pits on the walls of the tracheids.

The petrified wood first named by Witham^ Pinites eiggensis and afterwards � referred by him to the genus Pence was originallynbsp;recorded by Macculoch in 1814 from below the massive and precipitous ridge of pitchstone which forms a striking feature abovenbsp;the basaltic lavas of the Sgurr of Eigg in the Inner Hebrides.nbsp;Lindlev and Hutton ^ and Nicol� also gave short descriptions of thenbsp;structure and Miller� in the Cruise of the Betsey alludes to a fossilnbsp;trunk as �an ancient tree of the Oolites.� The wood occurs withnbsp;fragmental sedimentary rocks below the pitchstone and notnbsp;actually in situ; Mr Harker�s thorough examination of the islandnbsp;led him to the conclusion that the wood and associated rock-fragments are derived from Jurassic (Oxfordian) strata and werenbsp;carried up by volcanic agency�. Mr Harker tells me that he hasnbsp;never seen the Pityoxylon with any undoubted matrix adherent;nbsp;it occurs with wood of a different type (Dadoxylon) which is embedded in a white sandstone agreeing exactly with the Greatnbsp;Estuarine Sandstone of Eigg in which similar wood has been foundnbsp;in place. It is, however, possible that Pityoxylon did not come

� Ibid. (33) A. Pis. xiv., xv. nbsp;nbsp;nbsp;* Lindley and Hutton (33) A.nbsp;nbsp;nbsp;nbsp;PI. xxx.

� Harker (06) p. 55; (08) p. 52. In these memoirs Mr Harker discusses the earlier conclusions of Sir Archibald Geikie as to the geological history of Eigg andnbsp;gives reference.s to previous notices of the fossil wood; Seward (IP) p. 652.

from the same source as the Dadoxylon wood. Though probably Jurassic, a Tertiary source is by no means ruled out.

Kra.us^ transferred Witham�s species to the genus Pityoxylon and that name is used by Schroeter ^ and Schenk �. The specimensnbsp;on which the following account is based are most of them in thenbsp;British Museum. Annual rings clearly defined, usually 1�T5 mm.

broad; the reduction in diameter of the summer-tracheids extends over several rows, the transition being much more gradual thannbsp;in some types of Coniferous wood. A characteristic feature is thenbsp;occurrence of more or less circular patches where the tracheidsnbsp;have been destroyed with the exception of a single tracheid or a

small group in the centre of a clear crystalline matrix. Some of these patches simulate resin-canals, a fact which led Schenk tonbsp;deny the existence of true canals. All stages of decay are shown,nbsp;from the partial obliteration of a circular group of tracheids tonbsp;the destruction of the group, one central element being left, or tonbsp;the formation of a canal-like cavity (fig. 725, A). It is oftennbsp;difficult to decide whether a clear space in the wood is a canal ornbsp;the result of 'post-mortem changes, but there is no doubt as to thenbsp;occurrence of some true secretorj^ canals in different regions of thenbsp;wood. There is very little parenchyma accompanying the canals.nbsp;The medullary rays are of two kinds, uniseriate, 1�13 cells innbsp;depth, though usually about 6 cells in depth, and lenticular raysnbsp;with a central canal (fig. 725, C, c) identical with those in a modernnbsp;Pine. Fig. 725, D shows part of a vertical canal with somenbsp;parenchymatous lining in continuity with a horizontal canal in anbsp;broad medullary ray. Several small pits occur on the tangentialnbsp;and horizontal walls of the ray cells, and the radial walls, whichnbsp;are less clearly preserved, occasionally show 1�3 elliptical pits.nbsp;In radial section the upper and lower cells of a medullary ray arenbsp;often distinguished by their less uniform breadth and resemblenbsp;in this respect ray tracheids. A careful examination of sectionsnbsp;revealed the existence of bordered pits in the tangential v/alls ofnbsp;these elements and confirmed their tracheal nature

The bordered pits are generally single and sparsely scattered on the radial walls of the tracheids; occasionally the pits are innbsp;contact and a few double and opposite rows occur. An interestingnbsp;feature is the occurrence in some tracheids of a biconcave patchnbsp;of some brown substance agreeing closely with resinous depositsnbsp;described by Penhallow^. If this species is from Jurassic stratanbsp;its close resemblance to recent types of Pinus is a fact of considerable interest.

This species^ is founded on a specimen of wood obtained by Mr Eufiord from Wealden beds at Ecclesbourne on the Sussex

�lam indebted to Miss Ruth Holden for calling my attention to this feature and for other assistance in the examination of the sections.

coast: the sections on which the description is based are in the British Museum. Annual rings well marked, varying in breadthnbsp;from 1 to 3 mm.; resin-canals are abundant both in the spring- and

Fig. 726. Pityoxylon Sewardi. Radial section showing tracheids, tr, in a medullary ray; c, parenchyma of the ray with simple pits. (After Stopes.)

autumn-wood and horizontal canals occur in the lenticular medullary rays; some of the canals are occupied by large rounded cells like tyloses. The bordered pits on the radial walls of the tracheidsnbsp;form single or double rows; in the latter case the pits are generally

opposite but stellate groups also occur as in Cedroxylon transiem Goth. (fig. 723, A) and several other species. The uniseriatenbsp;medullary rays reach a depth of 30 cells: there are usually 2�4nbsp;oval or circular pits in the field.

A similar type of stem is represented by PityoxyIon NathorstP (Conw.) from the Lower Cretaceous of Sweden.

This species^ is founded on a petrified branch, not less than 18 cm. in diameter, from the Lower Greensand of Kent. It exhibitsnbsp;the usual features characteristic of the genus; the wood containsnbsp;horizontal and vertical canals with thin-walled epithelial cells. Thenbsp;medullary rays are larger and more abundant than in most Coniferous woods and horizontal tracheids (fig. 726, tr) occur interspersed with the parenchymatous cells, c, as well as on the uppernbsp;and lower margins, an arrangement in which the fossil bears anbsp;striking resemblance to the recent species Pinus monticola^.nbsp;Pityoxylon Benstedi Stopes.

In this Lower Greensand species* from Kent the resin-canals often contain tyloses as in P. Nathorsti Conw. and the epithelialnbsp;cells have very thick walls, a feature suggesting comparison withnbsp;the genus Larix. The medullary rays show well-marked Abie-tineous pitting (fig. 727, a) and ray-tracheids (rt) occur. ' Rims ofnbsp;Sanio are shown in fig. 727 between the circular bordered pits,nbsp;tr. The difference between the tracheal and ordinary parenchymatous elements of the rays, as represented in fig. 727 from anbsp;drawing by Dr Marie Stopes, is not very clearly defined and in thenbsp;upper ray shown in the figure part of a tracheid is seen abuttingnbsp;laterally on parenchymatous ray cells, the only difference betweennbsp;them being in the form of the pits, a criterion which is largelynbsp;dependent for its value on the state of preservation. Dr Stopes isnbsp;inclined to regard this species as most nearly allied to Larix.

A species from the Middle Cretaceous of Staten Island� found association with the short shoots described by Jeffrey andnbsp;* Conwentz (92) p. 13, Pis. i.�in., vi., vii.

� Stopes (15) p. 95, Pis. iv., v. text-figs. 23, 24. nbsp;nbsp;nbsp;� Ibid, text-fig. 25, p. 103.

Hollick as Finns triphylla, etc. The anatomical features are as follows;�annual rings narrow, not clearly marked owing to thenbsp;walls of the summer-tracheids being thinner than in recent speciesnbsp;of Finns; xylem-parenchyma confined to the periphery of the

resin-canals which occur in any part of the wood and are often filled with tyloses. The tracheids have a single row of pits notnbsp;contiguous or. flattened; the pits on the tangential walls arenbsp;confined to the summer-tracheids; both linear and fusiform

medullary rays occur, the latter with horizontal resin-ducts. There are no ray-tracheids. There is usually one circular or ellipticalnbsp;pit in the field. A second species from the same locality, Pityoxylonnbsp;scituatense, differs only in some unimportant features from P. staten-ense. These fossils differ from recent Pines as also from Picea,nbsp;Pseudotsuga, and Larix in the absence of ray-tracheids. In thenbsp;restriction of bordered pits to the tangential walls of the tracheidsnbsp;of the summer-wood they agree with the soft Pines, but thoughnbsp;this character is generally lacking in hard Pines, Jeffrey and Chrysler point out that in some hard Pines without pits on the tangentialnbsp;walls of the tracheids of vegetative shoots the summer elementsnbsp;of the cones have tangential pits. The occurrence of borderednbsp;pits on the tangential walls of the late wood and the absence of ray-tracheids are regarded by the authors of the species as ancestralnbsp;features.

A Middle Cretaceous species^ from New Jersey, referred by Miss Holden to Pinus, showing the following features:�annualnbsp;rings well developed; linear and fusiform medullary rays, horizontal and vertical resin-canals, bordered pits uniseriate and scattered on the radial walls of the tracheids; none on the tangentialnbsp;walls. Eims of Sanio are present. There are 1�2 pits in thenbsp;field with a lenticular pore and circular border; the other wallsnbsp;of the ray cells are abundantly pitted. Ray-tracheids occur onnbsp;the margins of the medullary rays and rarely interspersed with thenbsp;parenchyma; their walls are denticulate as in recent hard Pines.

The presence of horizontal tracheids in the medullary rays is an important character: in Pityoxylon scituatensiformis (Bailey)^,nbsp;another Middle Cretaceous species, ray-tracheids are present butnbsp;they have smooth walls and are not met with in the first 10�15nbsp;rings of wood, whereas in P. protosclerofitys they occur even in thenbsp;wood of the first year. In this connexion the presence of ray-tracheids in Pityoxylon eiggense is noteworthy at least if that speciesnbsp;is from a Jurassic source. Pityoxylon protoscleropitys is considerednbsp;by Miss Holden to be �probably the earliest form with all thenbsp;characters of a modern hard Pine, yet retaining certain ancestral

features, as the association of primary and fascicular leaves, the latter borne on brachyblasts subtended by a foliar trace.�

Miss Holden^ has also described from New Jersey two species which she compares with Preminus, namely Pityoxylon foliosumnbsp;and P. anonialmn.

Under the name Pinus Nathorsti Conwentz^ described in considerable detail specimens of petrified wood from the Senoniannbsp;Holma sandstone of Sweden, also a cone and two detached needles.nbsp;The rings of growth agree with those of stems and older branchesnbsp;in the gradual increase in the thickness of the tracheid-walls innbsp;passing from the spring to the late summer elements. The bordered pits on the radial walls of the tracheids are uniseriate andnbsp;separate and none were found on the tangential walls. Bothnbsp;horizontal and vertical resin-canals occur, several of them withnbsp;well-preserved tyloses; the epithelial cells are thin-walled andnbsp;unpitted as in Pinus silvestris. It may be that the abundance ofnbsp;tyloses is connected with the presence of fungal mycelia as innbsp;wood of Hevea stems recently described by Mr Brooks�. Thenbsp;medullary are uniseriate, generally 5�7 cells deep, also fusiformnbsp;and with resin-canals: the preservation is not sufficiently good tonbsp;admit of any definite statement as to the occurrence of horizontalnbsp;tracheids.

This Upper Cretaceous Japanese species was described as Abiocaulis zezoensis*, but in view of the occurrence of featuresnbsp;suggesting comparison with Pinus as well as with Abies the morenbsp;comprehensive genetic name is preferable. In the presence ofnbsp;pits on the tangential and horizontal walls of the medullary-raynbsp;cells and in the pitting of the tracheids this wood conforms to thenbsp;Abietineous type. There are no ray-tracheids; the pitting onnbsp;the radial walls of the ray cells agrees in part with that in A bresnbsp;and in some of the cells there are large circular pits like those ofnbsp;Pinus. Normal resin-canals are present in the second ring only,nbsp;others being interpreted as traumatic. Xylem-parenchyma is

gt; Holden (13*). nbsp;nbsp;nbsp;� Conwentz (92) p. 18, Pis. i.�in., vi., vii.

sparsely distributed. There would seem to be little difference of importance between this species and wood referred to Cedroxylon.

The number of Cretaceous examples of Pityoxylon might be considerably extended: for an account of French species referencenbsp;should be made to Lignier and to Fliche.

Gothan described this species, from South Nevada and probably of Tertiary age, as Piceoxylon Pseudotsugae^: it is interesting as a type of Pityoxylon agreeing closely with the recent genusnbsp;Psevdotsuga in the presence of spiral bands in the tracheids. Therenbsp;are both vertical and horizontal resin-canals and the ray cells havenbsp;Abietineous pitting. Xylem-parenchyma occurs next the summer-wood and the epithelial cells have thick walls, features in agreement with Gothan�s genus Piceoxylon. Bailey^ points out that innbsp;the absence of spiral bands in the ray-tracheids the fossil speciesnbsp;resembles Pseudotsuga Douglasii, while in Psevdotsuga macrocarpanbsp;the tracheids of the rays have spiral bands.

Fritel and Viguier� have described a species from Eocene beds in the Paris Basin as Piceoxylon Gothani in which some of the xylem-tracheids have spiral bands.

A Tertiary species^ originally described from material collected by Sir Leopold MacClintock in Banksland as Cupressinoxylonnbsp;pulchrum. A piece of wood in the Dublin Museum labelled �fromnbsp;Ballast Bay, Baring Island, given by Sir L. MacClintock� agreesnbsp;very closely with Cramer�s type-specimen, and as the resemblancenbsp;extends to most of the anatomical characters, I believe it to benbsp;the material on which C. pulchrum was founded. The chiefnbsp;difference is that the Dublin wood has resin-canals as in thenbsp;specimen described by Cramer as Pinus MacClurii (?) Heer^;nbsp;in C. pulchrum no resin-ducts are recorded. On the other handnbsp;in the sum of its characters the Dublin specimen agrees muchnbsp;more closely with C. pulchrum. Annual rings well marked;

bordered pits large, approximately 25 p in diameter, in 1�2 opposite rows, sometimes in contact and slightly flattened.nbsp;Medullary rays uniseriate, !��14 cells deep, also fusiform raysnbsp;containing a horizontal canal; 2 or 3 large oval pits occur on thenbsp;radial walls of the ray cells and in a few cases pits on the tangentialnbsp;walls. Ray-tracheids with bordered pits occur on the edges of thenbsp;medullary rays.

Among other Tertiary species reference may be made to Pity-oxylon farryoides Goth.^ from the Braunkohle of Rheinland, so named from its resemblance to the North American Finns Parry a,nbsp;characterised by horizontal tracheids with smooth walls' and thin-walled epithelial cells; also Pilyoxylon pineoides Kraus^ a Siciliannbsp;Tertiary species without ray-tracheids.

This species from the Oligocene amber beds of the Baltic coast was first named Pinites succinifer^ and several years later fullynbsp;described and admirably illustrated as Finns succinifer'^. Itnbsp;affords a striking illustration of the possibilities of amber as anbsp;petrifying agent and shows several features of anatomical interest..nbsp;The roots are represented by pieces of wood in a pathologicalnbsp;state: the tracheids have 1�3 rows of pits on their radial wallsnbsp;and some of them contain tyloses; the walls of the ray-tracheidsnbsp;have dentate ingrowths. The stem and branch wood is morenbsp;complete. Sieve-tubes and sieve-plates are exceptionally wellnbsp;preserved and both cortex and pith tissues are represented. Thenbsp;tracheids have 1�2 rows of separate pits; a spiral sculpturingnbsp;on the walls of the tracheids was mistaken by Menge for thenbsp;spiral bands characteristic of the Taxineae and he named thenbsp;species Taxoxylum electrochyton. Conwentz describes tyloses innbsp;the tracheids, also a crescentic patch of parenchyma in the woodnbsp;passing into a mass of resin�, a feature occasionally seen in recentnbsp;wood. The medullary rays have 1�4 pits in the field ; both ray-tracheids and horizontal resin-canals occur and in some cases

rays are said to consist exclusively of tracheal tissue^. Pine needles and cones have been obtained from the amber beds.

The Fossil forests of the Yellowstone Park include examples of Pityoxylon trees some of which have been described by Knowl-ton^ and Felix�, but unfortunately the anatomical details are notnbsp;as a rule well preserved. The most striking exposure of thenbsp;Tertiary (probably Miocene) trees is on the slopes of Amethystnbsp;mountain (fig. 712), where a succession of forests is representednbsp;throughout the 2000 ft. of strata. Felix describes a species,nbsp;Pityoxylon fallax, chiefly interesting from the point of view of anbsp;comparison between the stem and root wood of the same tree:nbsp;the elements of the root are in general larger than those in thenbsp;stem. Knowlton gives an account of P. Aldersoni and P. ame-thystinum, species which may be identical: the pits on the tracheidsnbsp;and medullary rays are seldom preserved, but the occurrence ofnbsp;both vertical and horizontal resin-canals is clearly shown.

Protopiceoxylon exstinctum Gothan. The generic name Proto-piceoxylon was proposed for some Lower Cretaceous wood from King Charles Land^ possessing Abietineous characters, intermediatenbsp;between Cedroxylon and Pityoxylon in having only vertical resin-canals, at least in uninjured wood. The anatomical features ofnbsp;the type-specimen are complicated by-the occurrence of additionalnbsp;resin-canals in wounded portions of the stem. It is difficult tonbsp;determine the precise extent of the traumatic influences, butnbsp;the presence of callus-wood healing a wound leaves no doubt asnbsp;to the correctness of Gothan�s conclusion that certain featuresnbsp;are abnormal and due to the effects of wounding. In the speciesnbsp;P. exstinctum are also included specimens from Spitzbergen�nbsp;and some of the material on which Cramer� founded his speciesnbsp;Pinites cavernosus: the later specific name is not retained on thenbsp;ground that the original diagnosis is incorrect and it was onlynbsp;after examining sections of the type-specimen that Gothan recognised the true nature of Cramer�s species.

2 nbsp;nbsp;nbsp;Knowlton (99) p. �63, Pis. cvi.�cviii., cxii.�cxv., cxviii., oxix.

3 nbsp;nbsp;nbsp;Felix (96) p. 254.nbsp;nbsp;nbsp;nbsp;^ Gothan (07^) p. 32, figs. 16, 17; PI. i. figs. 2�6.

5 Gothan (10) p. 15, PI. ii. figs. 5�8; PI. ill. figs. 1�4, 6�8. nbsp;nbsp;nbsp;� Cramer (68).

Protoficeoxylon exstinctum shows the following characters:�� annual rings well marked; vertical resin-canals occur in the woodnbsp;but there are no canals in the medullary rays except a few ofnbsp;unusually large diameter in wounded areas; there is no xylem-parenchyma apart from the resin-canals. Tracheids with f�2nbsp;rows of bordered pits on the radial walls, separate and circular,nbsp;also contiguous and flattened, opposite, or sometimes alternate:nbsp;in the occurrence of the Araucarian type of pitting on somenbsp;tracheids this species agrees with several types of Mesozoic wood.nbsp;Medullary rays uniseriate, characterised by well-developed Abie-tineous pitting; on the radial walls there are 2�4 circular andnbsp;bordered pits in the field. No undoubted ray-tracheids werenbsp;noticed; numerous small pits occur on the horizontal walls of parenchymatous cells associated with the resin-canals. The pith consistsnbsp;of parenchyma with thin sclerenchymatous diaphragms.

The horizontal canals, presumably traumatic, in some of the medullary rays resemble in their large size those in Anomaloxylonnbsp;but in that genus there is no Abietineous pitting on the medullary-ray cells; similar canals are described by Jeffrey^ in woundednbsp;wood of Cedrus and other Conifers. In the occurrence of verticalnbsp;canals only in the normal wood Protoficeoxylon is intermediatenbsp;between Cedroxylon, which has no canals, and Piceoxylon andnbsp;Pinuxylon of Gothan (= Pityoxylon of Kraus), the fossil representatives of such recent genera as Larix, Picea, and Pinus, in whichnbsp;both vertical and horizontal ducts occur. Gothan holds, andnbsp;probably with good reason, that vertical canals preceded thosenbsp;in the medullary rays and regards the fossil species as a primitivenbsp;type.

A species from the Black Hills described by Knowlton^ as Pinoxylon dacotense agrees with Protopiceoxylon in having onlynbsp;vertical canals, but it is not clear whether they are normal ornbsp;traumatic: Piceoxylon would seem to be the more appropriatenbsp;designation for Knowlton�s species.

This species is founded on a specimen from Cape Flora, Franz Josef Land, probably Oxfordian in age. Annual rings are distinct

and, narrow; several oval or circular spaces are conspicuous in transverse section, some being true canals and others the resultnbsp;of decay. There are 2�3 opposite rows of bordered pits on thenbsp;radial walls of the tracheids. Partially destroyed rows of resin-parenchyma occur which probably belong to secretory canals.nbsp;The uniseriate and comparatively deep medullary rays, 20�30nbsp;cells, are characterised by rather thick and pitted horizontal and

Fig. 729. ProtopiceoxyhnEd-wardsi. Longitudinal view of the thick-walled, pitted,nbsp;epithelial cells of the resin-canals. (After Stopes.)

vertical walls (fig. 728); 4�5 small simple pits occur on a few of the cells and on the upper and lower edges of some of the rays arenbsp;empty elements of unequal breadth which in all probability arenbsp;ray-tracheids. The wood agrees in the presence of verticalnbsp;canals only and in the structure of the medullary rays with Proto-ficeoxylon exstinctum Goth. In Gothan�s species there are 2�4nbsp;bordered pits in the field, but the absence of a border innbsp;the Franz Josef Land wood may be a consequence of imperfectnbsp;preservation.

Founded on a branch from the Lower Greensand of Sussex, showing 17 annual rings, having the following characters^: a largenbsp;pith nearly 3 mm. in diameter composed of paj-enchyma withoutnbsp;stone-cells; tracheids with usually one row of circular borderednbsp;pits; vertical canals in the summer-wood and associated withnbsp;a small amount of resiniferous parenchyma; medullary raysnbsp;uniseriate, with Abietineous pitting and 2�4 more or less circularnbsp;pits in the field. The small size of the resin-canals is a characteristic feature, also their thick-walled pitted epithelial cells (fig. 729).nbsp;The species differs from Gothan�s P. exstinctum in the smallernbsp;diameter of the canals, the absence of traumatic horizontal canals,nbsp;and in the greater number of the vertical secretory passages.

Woodworthia arizonica Jeffrey. This genus ^ is founded on specimens from the Triassic petrified forest of Arizona characterised by the occurrence of short shoots in the secondary wood comparable with those in the stem of Araucariopitys. In the type-specimen the annual rings are not very clearly defined: the pittingnbsp;on the tracheids is definitely Araucarian. The medullary raysnbsp;are uniseriate, 2�9 cells deep: they appear to have pits only onnbsp;the lateral walls.

On the surface of the wood are several small scars and a few larger ones, the former representing short shoots subtended by anbsp;leaf-trace; the shoots are not infrequently branched as they passnbsp;through the secondary xylem, a feature recorded also in Ginkgo^.nbsp;Jeffrey describes the short shoots as having a limited existencenbsp;and disappearing in the wood at a comparatively short distancenbsp;from the pith; they have no rings of growth, a character associatednbsp;with short-lived leaf-spurs in recent species but a feature in whichnbsp;they differ from those of Ginkgo. The leaf-traces subtending thenbsp;short shoots, in contrast to those of Araucaria, are not persistentnbsp;throughout the secondary wood. Jeffrey regards this fact as annbsp;argument against the view that the persistence of the traces innbsp;Araucaria is a primitive character; but it is worthy of note that

the leaves accompanying the shoots of Woodtvorthia are not strictly comparable with those of the foliar organs of recentnbsp;Araucarias which have no short shoots in their axils. Jeffreynbsp;regards the short shoot as a primitive attribute of the coniferousnbsp;stock and its occurrence in the stems of Woodtvorthia and Arau-cariopitys is held to be evidence in support of the interpretationnbsp;of the seminiferous scales of Abietineous genera as metamorphosednbsp;short shoots, an interpretation which is open to question. Thenbsp;presence of short shoots is not a monopoly of the Abietineae andnbsp;their presence in a stem may be regarded as a point of contact withnbsp;Ginkgo as well as with Abietineous plants. Attention is callednbsp;elsewhere to the probability that foliar spurs like those of Pinusnbsp;are specialised forms of ordinary shoots. However we maynbsp;interpret the characters exhibited by Woodworthia, the genus isnbsp;an interesting example of an extinct type illustrating the combination with Araucarian characters of a morphological feature thatnbsp;is no longer represented in the Araucarineae.

A genus founded by Jeffrey^ on a stem from the Middle Cretaceous beds of Staten Island, New York, showing on its decorticated surface scars of short shoots and in the structure of the wood both Abietineous and Araucarian features.

Araucariopitys americana Jeffrey. The bordered pits on the radial walls of the tracheids are often contiguous and flattenednbsp;though in places separate and circular, usually arranged as a singlenbsp;row. All the walls of the ray cells are pitted as in the Abietineae.nbsp;The large number of vertical resin-canals (fig. 758, C, D, page 323)nbsp;in a single tangential row is regarded as evidence of traumaticnbsp;origin. This conclusion is based on the fact that when canals arenbsp;present in wood that is known to have been wounded they occurnbsp;in crowded tangentially arranged rows in contrast to their sparsernbsp;distribution in the normal wood. There are no canals in thenbsp;uninjured tissues of Araucariopitys. Diaphragms of sclerous cellsnbsp;occur in the pith as in some recent Abietineae. The short shootsnbsp;are shown in tangential section of the stem and stated to be accompanied by a subtending leaf.

It is suggested by Jeffrey that Araucariofitys may be the stem of Heer�s genus Czekanowskia^, but there is no proof of a connexionnbsp;and evidence afforded by Czekanowskia favours an alliance with thenbsp;Ginkgoales. Araucariofitys is regarded by Jeffrey as �unquestionably Araucarian� though �nearer in structure to the Abietineaenbsp;than any other known Araucarian genus living or extinct.� Thenbsp;chief Araucarian feature would appear to be the occurrence ofnbsp;flattened tracheal pits, but Jeffrey points out that not only arenbsp;the pits in a single row and sometimes separate, but even whennbsp;in double rows the pits may be opposite. The other characters,nbsp;e.g., the pitting of the medullary-ray cells and the presencenbsp;of resin-canals, are more Abietineous than Araucarian^. Innbsp;view of the occasional occurrence in Abietineous genera of contiguous and alternate pits on the tracheids it is open to doubtnbsp;whether there are adequate grounds for assuming a definitenbsp;Araucarian affinity. Araucariofitys is one of several generanbsp;described by Jeffrey and other American authors exhibitingnbsp;features shared by recent Araucarineae and Abietineae which arenbsp;claimed as evidence of the greater antiquity of the Abietineousnbsp;type. In this genus the balance of evidence would seem to be innbsp;favour of an Abietineous alliance, the tendency towards an Araucarian pitting of the tracheids being reminiscent of ancestralnbsp;types in which that character was more pronounced.

Protocedroxylon araucarioides. Gothan. The type-species of the genus founded on Upper Jurassic wood� from the Esmarksnbsp;Glacier, Spitzbergen, is one of the most striking examples of anbsp;group of generalised types from Upper Jurassic strata especiallynbsp;from the Arctic regions. The generic name emphasises the Abietineous characters while the specific term gives expression to thenbsp;presence of Araucarian features. The following account is basednbsp;chiefly on the description by Gothan, and a few additional factsnbsp;are taken from an account of some specimens from Liassic andnbsp;Oolitic rocks on the Yorkshire coast by Prof. Jeffrey^ and Missnbsp;Holden�.

� Gothan (10) p. 27, Pis. v., vi. nbsp;nbsp;nbsp;^ Jefirey (12) p. 533, PI. i. figs, a, b.

Annual rings well marked; bordered pits on the radial walls of the tracheids in 1�3 rows; in the Spitzbergen wood these arenbsp;from 20 to 24/i in height, dimensions larger than in recent Arau-carineae, and in the type-specimen the pits are always contiguous,nbsp;more or less flattened and alternate�that is Araucarian; in thenbsp;English specimens the pits when in a single row are often separatenbsp;and circular but equally often contiguous. Jeffrey points outnbsp;that the alternate pits when in more than one series are lessnbsp;crowded than in Araucarian wood. There are no Sanio�s rims.nbsp;The medullary rays are uniseriate and characterised by Abietin-eous pitting on the horizontal and tangential walls; on the radialnbsp;walls there are 1�3 circular, apparently unbordered, pits in thenbsp;field. Xylem-parenchyma is practically absent. An interestingnbsp;feature is the abundance both in the Spitzbergen and Yorkshirenbsp;material of tyloses in many of the tracheids, a feature occasionallynbsp;met with in recent Conifers^ as in some other fossil species. Thenbsp;pitting of the tracheids in the type-specimen may be described asnbsp;exclusively Araucarian, but in the English specimens separatenbsp;pits also occur though on the whole the Araucarian type is dominant. The pitting of the medullary rays is on the other handnbsp;definitely Abietineous. The American authors, particularly Missnbsp;Holden^, consider that the absence of Sanio�s rims suffices to tipnbsp;the balance on the Araucarian side. On most of the tracheidsnbsp;the crowding of the pits precludes the occurrence of Sanio�s rimsnbsp;and in other cases their absence is not necessarily an originalnbsp;feature. Abietineous pitting is recorded by Jeffrey in the cone-axis of an Agathis and it has also been found in Araucaria] butnbsp;in the Araucarineae it is very exceptional: its occurrence as anbsp;constant featur� in Protocedroxylon may be regarded as an indication of Abietineous relationship. No substantial assistance isnbsp;afforded by impressions in Spitzbergen rocks: the abundance ofnbsp;Elatides is consistent with the occurrence of Araucarian wood, butnbsp;impressions of Abietineous Conifers afford at least as strong annbsp;argument in favour of the occirrrence of Abietineous wood.

Protocedroxylon scoticum (Holden). This species, described by Miss Holden^ under the generic name Metacedroxylon from Coral-liarr beds on the Sutherland coast of Scotland, is founded on a piecenbsp;^ See page 178.nbsp;nbsp;nbsp;nbsp;^ Holden, R. (15)-

of stem showing 75 rings of growth. There are no resin-canals and no xylem-parenchyma; the tracheids have uniseriate borderednbsp;pits on the radial walls only and they are almost invariably com.-pressed by mutual contact; the presence of a torus is a featurenbsp;characteristic of the Abietineae and not of the Araucarineae.nbsp;Bars of Sanio are present but there are no rims of Sanio. Tylosesnbsp;are abundant in the tracheids. The medullary rays are 2�20nbsp;cells deep, generally uniseriate though occasionally biseriate; thenbsp;pitting is of the Abietineous type. In the abundance of tylosesnbsp;and in other characters the wood resembles Protocedroxylonnbsp;araucarioides Goth, a species transferred by Miss Holden tonbsp;Metacedroxylon, but it differs in the absence of tangential trachealnbsp;pits and in the occurrence of biseriate medullary rays.

Some fossil wood, which is not very well preserved, from Middle Cretaceous rocks near lefren in the Gulf of Tripoli is described by Negri as Protocedroxylon ParonaP. The tracheal pitsnbsp;where biseriate are often alternate and compressed; rims of Sanionbsp;are absent: the latter feature, deduced from negative evidence,nbsp;and considering the state of preservatioir, is surely of little value.

The presence of Araucarian pitting on the tracheids in several Jurassic species is far from surprising in view of the prevalencenbsp;of that type of pitting in Palaeozoic stems; moreover an admixture of characters is a natural result of progressive development. It is a matter of opinion with regard to the relative valuenbsp;of tracheal or medullary-ray pitting whether Protocedroxylonnbsp;should be placed nearer to the Araucarineae or to the Abietineae.nbsp;Miss Holden^ discards the name Protocedroxylon for Metacedroxylonnbsp;on the ground that the former implies Abietineous affinity, a finenbsp;shade of difference that hardly gives adequate expression to hernbsp;conclusion that �Metacedroxylon araucarioides cannot be othernbsp;than an Araucarian Conifer.�

Gothan� instituted this generic name for some Upper Jurassic wood, originally described by Cramer* from Green Harbour,

2 nbsp;nbsp;nbsp;Holden, R. (14) p. 538.nbsp;nbsp;nbsp;nbsp;,nbsp;nbsp;nbsp;nbsp;* Gothan (05) p. 38.

Spitzbergen, as Pinites latiporosus in order to give expression to the combination of distinctive features in both the tracheal andnbsp;medullary-ray pitting. The most striking characteristics are thenbsp;very large size of the bordered pits of the tracheids, the occasionalnbsp;(X. fhyllocladoides) or constant {X. latiforosum) occurrence ofnbsp;contiguous and vertically flattened and transversely elongatednbsp;pits on the radial walls, the absence of pits on the transverse andnbsp;tangential walls of the medullary-ray cells, and the presence ofnbsp;large simple pores on the lateral walls. There is no�definitenbsp;evidence as to the nature of the foliage, though Nathorst^ hasnbsp;suggested a possible connexion between Elatides and Xenoxylon.nbsp;Xenoxylon would seem to have been widely distributed in laternbsp;Jurassic floras.

Fliche^ refers to Xenoxylon a piece of wood from the Muschel-kalk of the Vosges characterised by (i) the occurrence of uniseriate bordered pits compressed above and below and occupying the wholenbsp;breadth of the tracheids and (ii) a single elliptical pit in the field.nbsp;The regularity of the pores in the medullary-ray cells and the factnbsp;that they are most clearly shown where the preservation is bestnbsp;favour the conclusion that they are an original feature. Thenbsp;medullary rays are usually from 5 to 10 cells in depth. There arenbsp;no resin-canals and no xylem-parenchyma; Fliche states that thenbsp;pits on the tracheids are rather less flattened than in X. lati-forosum and they occupy a greater breadth of tracheal wall. Thenbsp;photographs accompanying the description are unfortunately toonbsp;small to show the important characters. The annual rings arenbsp;faintly marked and the summer elements are confined to 4�5 rows.

Despite the resemblance between this Triassic species and those previously described it is by no means certain that Fliche�s speciesnbsp;is generically identical with the younger types. Large simplenbsp;pores occur in the medullary-ray cells of recent Conifers belongingnbsp;to different families, and it is not uncommon to find the borderednbsp;pits on the radial walls of tracheids in contact and slightly flattenednbsp;in wood normally characterised by circular and separate pits. In

this connexion it is noteworthy that Fliche states that the pits in his wood are occasionally circular.

Gothan�s examination of the specimens on which Cramer founded this species^ enabled him to confirm the main points ofnbsp;the original description: he regards Cramer�s species Pinitesnbsp;pauciforosus as identical with the type-species with which he alsonbsp;identifies Araucarioxylon koreanum Felix^. Xenoxylon latiporosumnbsp;is characterised by the large size of the pits on the radial walls of

CD O	quot;2�	, Q o
O	o	, Q o
/o	o	O
O	o	O
D	o	lo

A, Xenoxylon phyllocladoides. B, C, Xenoxylon latiporosum. (A, after Gothan; B, C, after Cramer.)

the tracheids, 20�40/x broad and 15�20 p, high, their vertically flattened form (fig. 730, B) and their occurrence in one or two contiguous rows, the pits of double rows being generally opposite. Thenbsp;medullary rays are uniseriate, reaching 17 cells in depth, characterised by the narrow form of the cells, the absence of pits on thenbsp;horizontal and vertical walls and by the presence of large simplenbsp;pores on the lateral walls, usually one pore in the field (fig. 730, C)nbsp;which it almost fills, or occasionally two; there is no xylem-

The nature of the pitting led Kraus^ to include Cramer�s species in Araucarioxylon and Miss Holden^, who records thisnbsp;species from the Yorkshire coast, regards the absence of Sanio�snbsp;rims as evidence of Araucarian affinity. The medullary-raynbsp;pitting is, however, very different from that in recent Araucarineaenbsp;and the absence of Sanio�s rims may well be a natural consequencenbsp;of the crowded arrangement of the tracheal pits.

This species, founded on material from the Bathonian of Russian Poland�, differs from X. latiporosum in the not infrequentnbsp;occurrence of separate and circular pits on the tracheids: in it arenbsp;included specimens from Liassic rocks at Gallberges near Salzgitternbsp;in Germany described by Conwentz^ as Araucarioxylon latiforosumnbsp;(Cram.) and, with some hesitation, Cwpressinoxylon Barberi Sew.�nbsp;from the Yorkshire coast. The tracheal pits are Uniseriats,nbsp;flattened or separate and circular (fig. 730, A), or in two rows,nbsp;generally though not invariably opposite; they vary in size fromnbsp;22 X 30/X to 24 X 36/x; the medullary rays are generally less thannbsp;10 cells deep and in pitting agree with those of the type-species.nbsp;This species is recorded from Poland, Spitzbergen, King Charlesnbsp;Land�, Yorkshire, and Germany.

Though similar to Araucaria and Agathis in the flattened contiguous pits, Xenoxylon differs in the elliptical form of the border and pore, also in the occurrence of separate and circular pits andnbsp;in the occurrence of opposite pairs. In the form of the pits on thenbsp;tracheids Xenoxylon resembles the Palaeozoic species Dadoxylonnbsp;protopityoides Fel.� and pits of similar form occur in the wood ofnbsp;the recent Magnoliaceous plant Drimys Winteri^. From thenbsp;Abietineae the genus is distinguished by the restriction of thenbsp;medullary-ray pitting to the radial walls, though the large pores

(fig. 730, C) in the field resemble those of some Pines and other Abietineae as also those of Sciadopitys (fig. 693, N): in Xenoxylonnbsp;there are no resin-canals and no xylem-parenchyma. Gothannbsp;considers that while differing in the sum of its characters from anynbsp;other type of Conifer, Xenoxylon shows most resemblance tonbsp;Gothan�s genus Phyllocladoxylon (= Mesembrioxylon Sew.)^, a genusnbsp;including fossil species which suggest affinity not only with Phyl-locladus but with other members of the Taxaceae. In his accountnbsp;of Xenoxylon phyllocladoides from King Charles Land Gothannbsp;describes instructive examples of the effect of the action of fungalnbsp;hyphae on the structure of tissues. The genus may be described asnbsp;a generalised type exhibiting features shared by the Araucarineaenbsp;and Taxaceae.

Anomaloxylon magnoradiatum Gothan. Gothan proposed this name for some Upper Jurassic wood from Spitz-bergen^ which cannot be definitely assigned to anbsp;family-position: its most striking feature is thenbsp;occurrence of large spindle-shaped medullarynbsp;rays containing a large �canal,� or spaces lined bynbsp;a single layer of cells (fig. 731). The rings ofnbsp;growth are well marked, the summer-woodnbsp;being composed of a very few rows of tracheidsnbsp;in abrupt juxtaposition to the larger springnbsp;elements, a character associated with roots.

There are no vertical resin-canals and no regular or typical horizontal canals. Xylem-parenchymanbsp;is rare or absent. The bordered pits on thenbsp;radial walls of the tracheids are in 1�2 rows,nbsp;separate and circular or, more frequently, contiguous and more or less flattened but, if in twonbsp;rows, not alternate. The medullary rays arenbsp;uniseriate, generally 5�7 cells deep, or in placesnbsp;forming broad and deep spindle-shaped areasnbsp;either empty or containing a large circular canallike passage. These peculiar rays, as seen in anbsp;tangential section of the wood, are a conspicuous feature and are

often lined by a layer of small cells (fig. 731). In the presence of these large medullary xamp;js Anornaloxylon �esemhles Thylloxylon, butnbsp;in the latter genus the rays are smaller and more uniform in size.nbsp;Gothan discusses the nature of these medullary rays and inclinesnbsp;to the view that they agree more closely with abnormal or traumaticnbsp;formations in certain Conifers than with any normal structures.nbsp;There are no pits on the horizontal or tangential walls of the raynbsp;cells and there are 2�3 simple circular pits in the field.

The general impression gained from an examination of Gothan�s photographs is that no true canals occur, and that the peculiarnbsp;medullary rays owe their form to partial decay of abnormalnbsp;patches of parenchyma possibly produced as the result of wounding.

Though on the whole nearer in structure to the Taxodineae^ than to any other family Anomaloxylon is a type which cannotnbsp;be assigned to a definite position.

Thylloxylon irregulare Gothan. The generic name Thylloxylon was given to a single species of Upper Jurassic age from Spitz-bergen^ on account of the occurrence of tiillen-like parenchymanbsp;in horizontal canal-like spaces in some of the larger medullarynbsp;rays. The wood is characterised by separate bordered pits'in thenbsp;summer tracheids and 1�2 rows of alternate contiguous, Arau-carioid, pits on the spring elements; xylem-parenchyma occursnbsp;only at the end of the year�s growth. The medullary cells havenbsp;Abietineous pitting and there are 2�3 small circular, apparentlynbsp;simple, pits in the field, or occasionally only one in the regionnbsp;of the late wood. The rays are uniseriate or 2�3 cells broad andnbsp;some medullary rays closely resemble those of certain Abietineaenbsp;possessing horizontal resin-canals; but in Thylloxylon there are nonbsp;true canals. The central parenchyma of some of the broad raysnbsp;is replaced by a canal-like space and these spaces are often fillednbsp;with spherical tiillen-like tissue, a feature shared with Anomaloxylon, but in that genus there is no Abietineous pitting on thenbsp;medullary-ray cells. There are no vertical resin-canals.

^ That is AthrotaxiSf Cryptomeria, Sequoia, Taxodium, and Sciadopitys, genera which are now (see page 126) assigned to different families.

Dr Marie Stopes^ has recently instituted this generic name for a piece of Coniferous wood from Middle (or Upper 1) Cretaceousnbsp;rocks in New Zealand and in it she also includes the Liassic speciesnbsp;Araticarioxylon Lindleii (Withamj. The genus is a striking example of a combination of Araucarian and Abietineous characters^,nbsp;and, as Dr Stopes points out, it resembles in this respect Cedroxylonnbsp;transiens Goth, and other generalised types.

Planoxylon Hectori Stopes. The type-specimen, from Amuri Bluff, New Zealand, is part of a stem 150 years old or more. Thenbsp;lings of growth are well marked; the tracheids have 1�3 rows ofnbsp;alternate and hexagonal bordered pits on the radial walls and therenbsp;may be a single row of separate pits on the elements at the end ofnbsp;an annual ring. The medullary rays are nearly always uniseriate,nbsp;1�24 cells deep but usually from 3 to 9 cells in depth; all thenbsp;walls of the ray cells are pitted and there are 1�2 vertical rows ofnbsp;three pits in the field in the neighbourhood of the spring tracheidsnbsp;and generally a single vertical pair in the region of the late wood.nbsp;Xylem-parenchyma appears to occur only between the springnbsp;tracheids and the latest formed wood of the previous year. Likenbsp;many other fossil stems this species indicates the existence ofnbsp;Conifers with typical Araucarian pitting on the tracheids andnbsp;equally well defined Abietineous pitting on the medullary-ray cells.nbsp;It is especially interesting as showing the presence in the southernnbsp;hemisphere of a type very similar to Cedroxylon transiens and othernbsp;species recorded from high northern latitudes.

This Liassic species from Whitby was originally referred by Witham� to the genus Pence-, subsequently included in Araucario-xylon*� it has recently been transferred by Dr Stopes to her newnbsp;genus Planoxylon^. The pitting of the tracheids is essentiallynbsp;Araucarian; there are 1�3 rows of alternate hexagonal pits onnbsp;the radial walls, but the pitting of the medullary-ray cells, asnbsp;Dr Stopes has shown, is typically Abietineous.

^ irMvaoixat, to wander; suggesting that �the forms comprising the genus were moving from one position to another in a systematic sense.�

CHAPTER XLV.

The majority of the vegetative and fertile shoots, cone-scales, seeds, etc., selected in illustration of the past history and geographical distribution of the Coniferales are described under the differentnbsp;families enumerated on page 124. Under each family are includednbsp;not only specimens which, with a fair amount of confidence, cannbsp;be assigned to a family-position but also genera of doubtfulnbsp;affinity which it has been contended afford evidence of greater ornbsp;less value in favour of an alliance with the family under whichnbsp;they are described. The inclusion of certain genera in a chapternbsp;or section devoted to a particular family does not necessarily meannbsp;that they show clear evidence of relationship to that family:nbsp;many of the genera might with equal propriety be relegated tonbsp;Chapter L, w'hich is devoted to Coniferales incertae sedis. On thenbsp;other hand some genera included in that category would by othernbsp;authors be given a place in the Araucarineae or some other family.nbsp;I have endeavoured to state the different views expressed bynbsp;authors with regard to the affinity of imperfectly known genera,nbsp;but in many instances the available data do not afford any trustworthy evidence of relationship to existing types.

The distinctive characters of the recent genus Agathis {Dammara) are briefly described in Chapter XLIII. Fossil records bearing onnbsp;the past history of Agathis are more meagre and more difficult ofnbsp;interpretation than those relating to Araucaria. The evidencenbsp;at present available points to the greater antiquity of Araucarianbsp;at least, as regards the type of cone characteristic of that genus.

On the other hand the type of foliage-shoot represented by existing species of Agathis�without taking into account the Palaeozoicnbsp;leaves assigned to Cordaites, some species of which bear a closenbsp;superficial resemblance to those of certain examples of the recentnbsp;genus�is widely represented in Ehaetic and Jurassic floras bynbsp;Podozamites^. There is, however, no proof that Podozamites wasnbsp;nearly related to AgatJiis, and, indeed, such information as we havenbsp;with regard to the reproductive organs of that genus does notnbsp;point to any very close Araucarian affinity. Fossil wood givesnbsp;no help towards a distinction between the two members of thenbsp;Araucarineae nor do impressions of vegetative shoots materiallynbsp;aid us.

Palaeobotanical literature contains a few records of leaves referred to Dammara or Dammariies but in no case is there anynbsp;conclusive evidence of generic identity of the fossils with thenbsp;recent genus. Leaves from Lower Cretaceous rocks in Bohemianbsp;described by Velenovsky and by Fric and Bayer ^ as Dammaro-fhyllum striatum and D. hohemicum exhibit a close agreement innbsp;shape and venation with those of some species of Agathis, thoughnbsp;they differ but slightly from some forms of Podozamites, e.g.,nbsp;P. Reinii Geyl. (fig. 814, p. 456). Other leaves that may belong tonbsp;plants similar to Agathis are represented by Dammarites caudatusnbsp;and D. emarginatus Lesq. from the Dakota series�: these, probably specifically identical, forms present, as Lesquereux says, a closenbsp;resemblance to Agathis robusta. It is impossible without additionalnbsp;data to determine the true position of these and similar leavesnbsp;though it is permissible to regard them as possible examples ofnbsp;the foliage of Conifers closely allied to Agathis. Similarly, somenbsp;detached leaves from Cretaceous and Tertiary strata referred tonbsp;Podozamites may well be more akin to Agathis especially in viewnbsp;of the fact that Podozamites is essentially a Rhaetic and Jurassicnbsp;genus. The leaves figured by Saporta^ from Lower Cretaceousnbsp;beds in Portugal as Podozamites ellipsoideus agree closely withnbsp;those of Agathis. In the case of separate linear leaves like thosenbsp;described by Hollick� from the Cretaceous of Long Island as

Podozamites lanceolatus the term Desniiophyllum would be a more appropriate generic designation, the name Dammarites beingnbsp;adopted for broader foims. This distinction is purely arbitrarynbsp;and it must be admitted that there is no substantial justificationnbsp;for the use of a generic name implying affinity with Agathis.nbsp;Unless there are adequate grounds for assuming generic identitynbsp;of detached Tertiary and Cretaceous leaves with Podozamites itnbsp;is inadvisable to make use of that designation. As Schenk^ pointsnbsp;out Velenovsky�s Tertiary species Podozamites miocenicus may benbsp;a leaf of Agathis or possibly a Podocarpus. ,

This name was given to some oval-lanceolate leaves from Upper Cretaceous beds in Bulgaria varying in length from 10 to 12 cm.nbsp;and from 15 to 30 mm. broad agreeing closely with Heer�s Podozamites marginatus from the Cenomanian of Greenland but wiselynbsp;excluded by Zeiller^ from that genus, though on grounds whichnbsp;are no longer cogent if the interpretation of Podozamites impressionsnbsp;as shoots and not pinnate leaves is accepted.

Ettingshausen� records two species of Dammarites from Tertiary rocks in New Zealand: Dammarites Oweni includes innbsp;addition to leaves a cone-scale, the impression of a cone, and somenbsp;petrified wood of the Araucarian type. There is no proof thatnbsp;these disjuncta membra belong to the same plant though it is notnbsp;improbable that they are parts of a Conifer closely allied tonbsp;Agathis. Ettingshausen�s second species D. univervis is foundednbsp;on a leaf and a supposed cone-scale of doubtful value.

The data furnished by leaves alone are of little value. In addition to the cone described from New Zealand by Ettingshausennbsp;other examples are recorded as species of Dammarites but withoutnbsp;any satisfactory evidence of affinity to the recent genus, e.g.,nbsp;Dammarites albens Presl.^ from the Quadersandstein of Bohemianbsp;and D. crassipes Goepp.� These two species are united by

� Sternberg (38) A. PI. Lii.; Corda in Keuss (46) B. PI. XLViii.; Goeppert (50) p. 237; Schimper and Schenk (90) A. p. 279, fig. 292 6

Velenovsky under a single type which he calls Krannera mirabilis'^ from a name suggested by Corda: additional examples superiornbsp;in preservation to those previously figured are illustrated innbsp;Velenovsky�s memoir on the Bohemian Cretaceous Gymnosperms.nbsp;Velenovsky regards the supposed cones as stems bearing crowdednbsp;woody scales which originally had long Cordaites-Y\kamp; leaves attached to a transverse ridge just internal to the thickened distalnbsp;ends: there appears to be no absolute proof in support of thisnbsp;connexion between scales and foliage-leaves, but one specimennbsp;figured shows portions of leaf-like organs attached to two of thenbsp;scales, though these may be petiolar and not pieces of laminae.nbsp;Reference is made elsewhere to the Krannera leaves. It is probablenbsp;that, as Velenovsky believes, the supposed cones are stems similarnbsp;to tuberous Cycadean species but it is doubtful if they were provided with leaves of the type included in Krannera mirahilis.nbsp;Schmalhausen^ figures a Tertiary cone from Russia as Dammaranbsp;Armaschewskii which in the form of the distal ends of the scalesnbsp;resembles Araucaria brasiliensis and species of Pinus, but somenbsp;detached scales agree closely in shape and in the possession of anbsp;single seed with those of Agathis. Small detached^ cone-scalesnbsp;of Tertiary age, described by Schmalhausen� as Dammara Tollinbsp;from the New Siberian Islands,

(fig. 732), may be allied to Agathis; they agree generally with those ofnbsp;D. borealis and other western types.

It is, however, from detached cone-scales obtained from Cretaceous strata in Greenland and somenbsp;European localities but especiallynbsp;from the Eastern United States that the most promising information has been gained. Hitchcock first recorded these scalesnbsp;from Martha�s Vineyard and spoke of them as �seed-vessels� ofnbsp;some Coniferous plants^, but it was Heer� who compared themnbsp;with the cone-scales of Agathis. The latter author describednbsp;several examples-from Cenomanian strata in West Greenland asnbsp;Dammara borealis, D. microlepis, etc.

Dammarites borealis (Heer). Though it is clearly impossible to define with any precision the limits of species based on detachednbsp;scales varying considerably in size and shape, several types havenbsp;been recorded, particularly from different localities on the Atlanticnbsp;Coastal plain of North AmericaThe larger forms may conveniently be included in Dammarites borealis Heer and smallernbsp;forms are illustrated by Protodammara speciosa Holl. and Jeff.^nbsp;It is probable that these two types are generically identical, butnbsp;the name Protodammara implies the presence of certain structuralnbsp;features while Heer�s species is founded onnbsp;casts or impressions. A specimen of thenbsp;latter species from Greenland is shown innbsp;fig. 733, the scale is 22 mm. broad andnbsp;is characterised by several parallel lines,nbsp;either vascular bundles or resin-canals, andnbsp;the white patches represent some exudednbsp;resinous material. Other Greenland examples are more elongated basally and arenbsp;identical in shape with the smaller scalesnbsp;from Staten Island seen in fig. 758, E, F,nbsp;page 323. Fossils of similar form were described by Heer fromnbsp;the same locality as Eucalyptus Geinitzii^. Krasser^ and somenbsp;other writers have retained the generic name Eucalyptus on thenbsp;ground of association with Eucalyptus-li]ce leaves. Hollick speaksnbsp;of scales like D. borealis as � among the most abundant and characteristic remains found in the Cretaceous deposits of America andnbsp;Europe��: he adds that the name Dammarites is chosen for thenbsp;sake of convenience rather than from a conviction that it representsnbsp;their true generic relationship. Newberry in describing this typenbsp;from the Amboy clays states that some of the scales have grooves,nbsp;corresponding to the dark lines in fig. 733, filled with amber�,nbsp;and anatomical evidence derived from Protodammara supports thenbsp;view that the cone-scales were rich in resinous substance. Both

1 Hollick (97) PI. XI. figs. 5�8; (06) p. 37; Newberry and Hollick (95) p. 46.

large and small cone-scales are recorded by White^, Berry^, Knowlton, and especially by Hollick from several places on thenbsp;Atlantic Coastal plain. Knowlton�s species Dammarites acicu-laris^ is probably identical with D. borealis. As examples ofnbsp;smaller forms reference may be made to D. northfortensis*�, D.nbsp;minor and the scales now included in Protodammara. Thesenbsp;numerous scales occur as detached specimens and without seeds,nbsp;but their resemblance to the cone-scales of Agathis and the anatomical features exhibited by the lignitic examples described bynbsp;Hollick and Jeffrey afford strong arguments in favour of annbsp;affinity to Agathis. We have no proof as to the nature of thenbsp;vegetative shoots of the parent-plants. Newberry states that innbsp;the Amboy clays the scales occur in association with shoots likenbsp;Heer�s Juniferus macilenta, in some cases, apparently, attached.nbsp;On the other hand Krasser considered the association of the specimens from Moravia, which he called Eucalyptus Geinitzii, withnbsp;dicotyledonous leaves as evidence of original connexion.

It is probable that these scales� are generically identical with the larger forms referred to D. borealis and other species, e.g.,nbsp;D. cliffwoodensis'^, but until anatomical evidence is obtained anbsp;distinction should be recognised. The type-specimens were foundnbsp;at Kreischerville, Staten Island, New Jersey, in Middle Cretaceousnbsp;beds, and the species is thus defined:�kite-shaped cone-scalesnbsp;from 4 to 6 mm. long by 4�6 mm. broad above, abruptly narrowednbsp;from about the middle to the base (fig. 758, E, F, page 323),nbsp;rounded, incurved, and apiculate above; resin-ducts five or more,nbsp;extending down the lower surface of the limb; seed-scars three innbsp;number, crescentically arranged above the middle and approximately in the broadest part of the scale, with the central one highernbsp;up than the laterals. Near the base of a scale there is a singlenbsp;vascular bundle with the xylem uppermost: at a higher level anbsp;single strand with reversed orientation is given off and the original

^ Hollick (04) p. 405, PI. nxx. figs. 1, 2. nbsp;nbsp;nbsp;� Ibid. (12) PI. ii. figs. 35�37.

bundle divides into three. In the lower portion of the scale there are seven resin-canals and above these is a band of transfusion-tracheids surrounding and connecting the vascular bundles. Innbsp;median longitudinal section a scale shows a terminal spinousnbsp;process similar to that in the scales of Conites Juddi^ (fig. 734);nbsp;on the adaxial side of this the scale is swollen and internal to thenbsp;swollen part is a small pit marking the position of the middle ofnbsp;the three seeds. The upper surface of the scales is covered withnbsp;periderm and stone-cells occur in the ground-tissue. In anatomicalnbsp;characters Protodammara resembles the scales of recent Arauca-rineae more closely than those of any other Conifers, and in thenbsp;absence of a definite ligule and in the relation of the seeds to thenbsp;scale the fossil scales are similar to those of Agathis. In thenbsp;description of a Scotch Upper Jurassic cone, Conites Juddi,nbsp;attention is called to a close resemblance in anatomical featuresnbsp;to Protodammara. The American scales occur in association withnbsp;shoots of the type represented by Brachyfhyllum macrocarfumnbsp;Newb.^ {�B. crassum) (fig. 758, G), an association noted also innbsp;other localities than Kreischerville. The structure of these shootsnbsp;is described under the genus Brachyphyllum^, but as regards thenbsp;scales the important point is that if this association means originalnbsp;connexion, the habit of the parent-plant was wholly differentnbsp;from that of any recent Agathis or Araucaria. Brachyfhyllumnbsp;macrocarfum is regarded by Hollick and Jeffrey as certainlynbsp;Araucarian. Wood of the Araucarian type is also found in association with the Protodammara scales and the Brachyfhyllum shoots.nbsp;Considering the cone-scales by themselves, their position wouldnbsp;seem to be next to Agathis though they differ in bearing threenbsp;seeds in place of the single seed in the recent genus: the numbernbsp;of seeds borne on the larger scales siich as B. borealis is not known.nbsp;Protodammara affords an interesting illustration of the co-existencenbsp;of characters now characteristic of the Araucarineae with othersnbsp;no longer exhibited by members of that family: assuming a connexion between Brachyfhyllum macrocarfum and the cone-scales,nbsp;the habit of the vegetative shoots furnishes a further illustrationnbsp;of a wider range in the morphological features of fossil Conifersnbsp;allied to existing Araucarineae.

^ See page 252. nbsp;nbsp;nbsp;^ Hollick and Jeffrey (09) B. p. 33.nbsp;nbsp;nbsp;nbsp;� See page 322.

This name was given to partially petrified cones of Upper Jurassic age collected by Hugh Miller on the North-east coast ofnbsp;Scotland^: the fossils though differing in size and to some extentnbsp;in form are included under one specific term but distinguished as

Fig. 734. Conites Juddi. A, forma 7; ab, space where a section was cut; s, seed. B, forma a; s, spaces, probably resin-canals, filled with crystalline materialsnbsp;and simulating seeds. C, forma ab, ridge.

forma a, y, and 8 (fig. 734). The type-specimens are in the Eoyal Scottish Museum, Edinburgh. Spirally disposed thicknbsp;scales are attached by a comparatively narrow base to a thicknbsp;axis and the individual scales agree closely in shape with thosenbsp;described as Dammarites borealis and with smaller forms referrednbsp;Seward and Bancroft (13) p. 873, PI. I. figs. 9�12; PI. ii. figs. 14�21.

by Hollick and Jeffrey to Protodammara. The parenchymatous tissue of the cone-scales contains several thick-walled idioblastsnbsp;and the resin-canals and spaces form a conspicuous feature. Anbsp;series of vascular bundles runs radially through the scale, but nonbsp;evidence has been obtained of the occurrence of a double set ofnbsp;vascular strands like those in Protodammara. The seeds�thenbsp;number of which, whether one or more, cannot be determined�nbsp;lie in a depression near the proximal end of the scales and there

vb.-

K^vb.

Fig. 735. Conites Juddi^ forma a; cone-scale in median section; Z, ligule; 5, 5, (?) resinous material; r6, vascular bundles. ( x 6.)

is a ligule on the abaxial side of the seed or seeds (fig. 736, B, 1). The uppermost scales on the two cones shown in fig. 734, B and C,nbsp;illustrate the striking similarity to such detached scales as thosenbsp;of Dammarites borealis: the raised patches, s, s, simulating seeds,nbsp;are formed by a crystalline substance filling cavities in the scalesnbsp;and probably corresponding to the resin-ducts which form a characteristic feature in the scales of Dammarites and Protodammara.nbsp;A cone-scale from the specimen represented in fig. 734, B shows

in longitudinal section (fig. 735) a large cavity in the lower part of the scale containing at each end a dark patch of some secretednbsp;substance s, s; above this is a vascular strand vb, extending intonbsp;the distal end of the scale near which is a ligular outgrowth I, andnbsp;below this is a depression on the upper face of the scale in which

a seed was originally situated. The cones shown in fig. 734, B, C, have lost their seeds and indicate a persistent habit in contrastnbsp;to the cones of Agathis and, presumably, the cones which possessednbsp;scales like Dammarites borealis. The cone seen m fig. 734, A, isnbsp;probably younger; the scales are more crowded and in one of

them is the cast of a seed, s. One of the scales, 1-7 cm. long, of this cone is represented in section in fig. 736, B: a vascular bundle, c,nbsp;runs through the length of the tissues towards the blunt spinousnbsp;distal end above which is a prominent hump and next to this anbsp;ligule, I, close to the depression in which the seed, s, was situated.nbsp;Below the vascular bundle, c, a band of periderm surrounds anbsp;central area of decayed tissue, a, a. Next the lower surface atnbsp;b are a few layers of palisade cells, a characteristic feature. Thenbsp;specimen shown in section in fig. 736, A, is described as forma S:nbsp;the secondary xylem, x, of the axis is not sufficiently well preserved to throw any light on the nature of the tracheal pitting.nbsp;The large (white) cavities at first sight suggesting seeds near thenbsp;axis are bounded by periderm and correspond to the partiallynbsp;destroyed tissue in fig. 736, B: a delicate structure, the nature ofnbsp;which could not be determined, occurs in the cavity d, fig. 736, A.nbsp;In the scale e several smaller cavities are seen near the upper facenbsp;above the vascular strands and below the latter is a larger cavity;nbsp;similar cavities are shown in the scale � (fig. 736, A).

Cones similar to Conites Juddi are described by Velenovsky^ as Fricia nobilis and Sequoia fastigiata, both from Lower Cretaceousnbsp;strata, but it is impossible to say whether the resemblance has anynbsp;significance. In several anatomical characters the scales ofnbsp;Conites Juddi resemble those of Protodammara described by Hollicknbsp;and Jeffrey from Kreischerville. The only indication of pits onnbsp;the xylem tracheids in the Scottish cones was seen in the scalenbsp;shown in fig. 736, B: the pits are for the most part uniseriate butnbsp;occasionally contiguous though generally not actually in contact.nbsp;Large idioblasts and resin-cavities occur in both the Scottish andnbsp;American cones, but in the former the occurrence of a ligule is anbsp;distinguishing feature in which they agree with cone-scales ofnbsp;recent Araucarias. In Araucaria the seeds are embedded in thenbsp;substance of the scales while in the fossil species they are situatednbsp;in a depression on the upper face, a feature in which Conites Juddinbsp;agrees more closely with the cones of Agathis. In the sporophyllsnbsp;of Conites Juddi, which anatomically are close to those of recentnbsp;Araucarineae, characters occur which are now shared betweennbsp;Araucaria and Agathis. The apparently small size of the seedsnbsp;1 Velenovsky (85) B. PI. in. fig. 6; PI. vin. fig. 13.

and their relation to the ligular outgrowth, as well as the occurrence of separate bordered pits on the tracheids suggest comparisonnbsp;with the recent genus Cunninghamia, though the structure of thenbsp;scales is more akin to that of Araucarian sporophylls. The combination of features which are now distributed among differentnbsp;genera is to be expected in extinct types belonging to evolutionarynbsp;stages anterior to the divergence of characters along independentnbsp;lines. The main conclusion is that the affinities are Araucariannbsp;though the morphological characters are such as to indicate a combination of features no longer found in a single genus.

This species (fig. 737), from Inferior Oolite rocks at Bruton, Somersetshire^, affords a good example of a large Araucarian conenbsp;13 cm. in diameter very similar in form to some recent speciesnbsp;{cf. fig. 680 and fig. 681). The rhomboidal scales, 2 cm. broad atnbsp;the distal end, are laterally winged as in Araucaria Cookii (fig. 638,nbsp;A) and bear a single seed embedded in the middle of the uppernbsp;surface: on the exposed distal ends is a transverse groove and onnbsp;some of the more complete examples a short rounded umbo isnbsp;seen below the groove; in some scales a transverse row of pitsnbsp;marks the position of vascular bundles just below the transversenbsp;depression.

This species was originally described by Carruthers as Kaida-carf um ooliticum^ from the Great Oolite of Northamptonshire and regarded as an inflorescence of some Pandanaceous plant. Zigno�nbsp;transferred it to Pandanocarpum. An examination of the type-specimen in the Northampton Museum led me to refer the cone tonbsp;Araucarites'^. The type-specimen (fig. 738) is a portion of a conenbsp;9 cm. long consisting of a stout central axis covered with spirallynbsp;disposed deep pits bounded by a crystalline reticulum; the pits

being cavities in the proximal portion of the scales in which the seeds were embedded. Numerous imbricate scales are attachednbsp;laterally to the central region and partially hidden in the matrix.nbsp;The scales are approximately 1-7 cm. broad and� slightly winged.nbsp;The single seed on each scale, the general form of the cone, the shapenbsp;of the individual scales, and the occurrence of sterile scales at thenbsp;base of the axis are features in which the fossil is practicallynbsp;identical with recent forms. Fig. 739 shows a piece of a smaller

cone (in the Northampton Museum), of the same type; this specimen shows the appearance of the scales in end-view (C), in surface-viewnbsp;(A), and as seen from the proximal end with the seed-cavity (B).nbsp;In fig. 739, A, the base of a seed is seen projecting from the middlenbsp;of the laterally expanded scale. An oblong-ovate cone describednbsp;by Carruthers^ from the Coralline Oolite at Malton, Yorkshire, asnbsp;Araucarites Hudlestoni is probably another example of this species:nbsp;in one of the specimens of A. Hudlestoni in the York Museum anbsp;broad central region is occupied by a mass of pisolite to whichnbsp;numerous cone-scales are attached. The scales are shown innbsp;section and in several of them there is a single seed lying in a cavitynbsp;occupying the proximal end of the scale precisely as in A. ooliticus.nbsp;It is possible that A. ooliticus is specifically identical with A. sfJiae-rocarfus; it is at least a closely allied type.

The specimen figured by Bindley and Hutton^ as Strobilites Bucklandi appears to be indistinguishable from A. ooliticus.nbsp;Similar cones are illustrated by Araucarites Cleminshawi Mansell-

PleydelF from the Inferior Oolite of Dorsetshire, A. sfhaericus (Carr.)^, originally referred to Cycadeostrobus, and other species.

Araucarites pifpinyfordensis (Ung.)� is a Wealden species first described by Fitton and named by Unger Zamiostrobus fipjiing-fordensis but recognised by Carruthers as a cone closely allied to A.nbsp;sfhaerocarfus. Detached scales which may belong to this speciesnbsp;have recently been figured from the Wealden beds on the Sussexnbsp;coast. The cone described as Araucarites (Conites) sp. from thenbsp;same locality is possibly identical with A. fiffingfordensis.

The specimens from the Stonesfield slate^ on which this species was founded are rather larger than the scales of A. Phillifsi: thenbsp;scale shown in fig. 740, 4, is 3 cm. long and 1-7 cm. -broad; a raisednbsp;edge just beyond the single seed no doubt corresponds to the so-called ligule on an Araucarian scale and the distal spinous processnbsp;is another feature shared with recent types.

A specimen described by Fliche� from Lower Cretaceous rocks in the Haute-Marne, France, is made the type of a new genusnbsp;Sarcostrobus on the ground that the seeds are not so completelynbsp;covered by the tissues of the scales as in recent Araucarias. Thenbsp;elliptical cone is 5-5 cm. long and 3-8 Cm. in diameter; in form,nbsp;in the stout axis, and in the shape of the single-seeded cone-scalesnbsp;it closely resembles the megastrobili of Araucaria excelsa andnbsp;Jurassic species such as Araucarites ooliticus (Carr.) (fig. 738);nbsp;the small seeds are sunk in a cavity at the proximal end of thenbsp;scale, but Fliche states that'they are not covered on their uppernbsp;side by the substance of the scale. He is no doubt correct innbsp;assigning the cone to the Araucarineae, but the slight differencesnbsp;between the relation of seeds to scales referred to by Fliche do notnbsp;appear to be sufficiently important to justify the creation of anbsp;distinctive generic name; moreover the preservation of the specimens renders accurate description of details very difficult.

* Carruthers (69�) p. 3, PI. V. figs. 1�6; Seward (04) B. p. 137, PI. in. fig. 5; PI. XII. fig, 2.nbsp;nbsp;nbsp;nbsp;� Fliche (00) p. 11, PI. i.

Araucarites hespera Wieland. This type is described by Wie-land^ from a specimen obtained from Upper Cretaceous rocks in South Dakota consisting of half an eroded cone bearing scalesnbsp;with small seeds. Wieland�s description, though brief and lackingnbsp;details, and the photograph support his conclusion as to thenbsp;Araucarian affinity of the specimen.

The generic name Doliostrobus was instituted by Marion^ for specimens of foliage-shoots from Oligocene beds in the South ofnbsp;France agreeing with Araucarites Sternhergii, on the ground thatnbsp;the reproductive organs exhibit features more like those'charac-teristic of Agathis than Araucaria. Laurent� refers a small piecenbsp;of a foliage-shoot from the Aquitanian beds in the Puy-de-D6me tonbsp;Doliostrobus Sternhergii, though there is no information with regardnbsp;to the cones. Gardner^ describes branches from the Bembridgenbsp;marls in the Isle of Wight as Doliostrobus Sternbergii (Goepp.) andnbsp;accepts Marion�s conclusion as to the intermediate character of thenbsp;genus; the foliage-shoots, though rather more slender than thosenbsp;from Bournemouth referred by him to Araucaria Goepperti, arenbsp;exactly similar in habit and cannot be distinguished by any featurenbsp;of importance. Gardner reproduces a drawing communicated bynbsp;Marion of a foliage-shoot bearing a terminal cone-axis from whichnbsp;the scales have fallen, also several detached cone-scales, agreeingnbsp;closely in size and shape with scales of Araucaria excelsa and othernbsp;recent species, and an impression described as a seed with anbsp;lateral wing. The supposed seed has, however, a terminal wingnbsp;and moreover it is as large as the detached scales: it is permissiblenbsp;to suggest that it may be an imperfectly preserved cone-scale,nbsp;but without examining the actual specimen any definite assertionnbsp;would be hazardous. Gardner states that a reason for comparingnbsp;the scales with those of an Araucaria is that in Agathis the scalesnbsp;are persistent, but as pointed out elsewhere^ cones of Agathisnbsp;very readily fall to pieces and the scales easily become detachednbsp;from the axis. Having regard to the nature of the sterile shoots,nbsp;the form of the cone, as shown in a drawing published by Gardnernbsp;of a specimen sent to him by Ettingshausen from Haring, and an

unconvincing specimen of a winged seed figured from Marion�s drawing, there would seem to be no valid reason for drawing anbsp;distinction between Doliostrobus and Araucarites or for regardingnbsp;Marion�s and Gardner�s fossils as intermediate between Araucarianbsp;and Agathis. Attention has been called on a previous page^ tonbsp;the danger of placing too much confidence in the resemblance ofnbsp;foliage-shoots of fossil specimens to those of recent types, but innbsp;this case the presence of cones and scales like those of Araumrianbsp;supplies confirmatory evidence.

The generic name Pseudo-Araucaria was given by Fliche^ to several cones from the Lower Cretaceous beds of the Argonnenbsp;which he described under three specific names. Pseudo-Araucarianbsp;Loppinetti, P. major, P. Lamberti. Externally they are similarnbsp;to those of some recent Araucarias and in shape agree with conesnbsp;of Cedrus: a stout axis bears deciduous scales with two seeds, thenbsp;seeds of each pair being separated from one another by a mediannbsp;ridge of the cone-scale which covers them laterally. The seedsnbsp;appear to bear a relation to the scale similar to that between thenbsp;single seed and the cone-scale of an Araucaria. The cone-scalesnbsp;are slightly expanded laterally as in the Eutacta section of thenbsp;recent genus. Fliche�s descriptions are unfortunately inadequatelynbsp;illustrated and it is difficult to obtain a very clear impression ofnbsp;the structural features. The most interesting peculiarity of thesenbsp;cones is the occurrence of two seeds in each cone-scale agreeingnbsp;in their position on the sporophyll with the single seed of Araucaria:nbsp;the author of the genus regards it as a type intermediate betweennbsp;the Abietineae and the Araucarineae.

The question of the lower geological limit of cones or cone-scales of the Araucarian type is one which cannot be settled with any certainty; there are many examples of vegetative organsnbsp;very similar in habit to Araucaria excelsa and allied species recordednbsp;from Triassic, Permian, and to a less extent from Upper Carboniferous strata, also others which agree in the broader form of the

leaves with Araucaria Bidwilli and A. inibricata; but the majority of these shoots are referred to such genera as Voltzia, Walchia,nbsp;Albertia, and Ullmannia. It is pointed out in the description ofnbsp;these genera that there are reasons for believing them to havenbsp;Araucarian affinities, though there is no definite evidence that anynbsp;of them bore cones exhibiting the same order of resemblance tonbsp;those of recent Araucarineae as is the case with Jurassic andnbsp;Cretaceous types.

One of the very few Palaeozoic species of seed-bearing scales that can reasonably be referred to the genus Araucarites isnbsp;A. Delafondi founded by Zeiller^ on some detached scales fromnbsp;Permian beds at Charmoy; the scales are broadly triangularnbsp;10�12 mm. long and 8�10 mm. broad, the base is cuneate andnbsp;truncate, the apical margin is rounded and has a small mediannbsp;depression instead of the usual spine. In the middle of the scalenbsp;is a shallow depression which contained a single seed 8�10 mm.nbsp;long and 2 mm. broad. As Zeiller says, there is no absolutenbsp;certainty as to the affinity of this species but the scales are unquestionably very similar to those of Mesozoic and recent speciesnbsp;of Araucarites and Araucaria. It is suggested that the vegetativenbsp;shoots of Ullmannia frumentaria (fig. 750) from the same beds maynbsp;belong to the plant which bore cones with scales of A. Delafondi.

The occurrence of widely distributed Jurassic cone-scales, bearing a single seed and agreeing very closely in their shape andnbsp;size, as also in the laterally expanded borders and in many casesnbsp;in the presence of a distal spinous process, with those of recentnbsp;species of Araucaria especially those belonging to the sectionnbsp;Eutacta, bears striking testimony to the former extended geographical distribution of Araucarian plants. It has been pointednbsp;out in a previous chapter that a single seed occasionally occurs onnbsp;the seminiferous scales of recent Pine cones (fig. 686, B), but in thenbsp;scales now under consideration the occurrence of a single seed isnbsp;a constant feature and moreover the form of the scales is identicalnbsp;with that of such species as Araucaria excelsa and A. CooTcii. Thenbsp;number of names given to the fossil scales is but a rough index of

the number of actual species: it is obviously impossible to decide with any assurance how much value should be attached to differences in size or to slight variations in form, but the main point isnbsp;that cones and cone-scales of the Araucarian type are among thenbsp;most familiar Jurassic fossils. The following selected examplesnbsp;are chosen in illustration of this statement and reference to othersnbsp;will be found in some of the sources quoted in the footnotes.

Carruthers^ described this species from the Middle Jurassic rocks on the Yorkshire coast: the type-specimen is in the Leckenbynbsp;collection in the Sedgwick Museum, Cambridge. The scales are

cuneate, nearly as long as broad (fig. 740, 9), and in shape similar to those of the cone already described as Araucarites ooUticusnbsp;(Carr.).

Examples of French Jurassic cone-scales are afforded by Araucarites Moreauana Sap.^, from Corallian beds near St Mihiel

^ Carruthers (69^) p. 6, PI. ii. figs. 7�9; Seward (00) B. p. 285, PI. x. fig. 4.

and other localities, similar to those of A. Brodiei (fig. 740, 4, 7) but reaching a length of 4-5 cm., kite-shaped and provided with anbsp;terminal spine; also Amucarites microphylla Sap. ^ represented bynbsp;foliage-shoots and cone-scales; the shoots bear linear-lanceolatenbsp;leaves similar to those of Araucaria Bidwilli but smaller, and thenbsp;scales are of the Eutacta type. Amucarites Falsani Sap.^ is foundednbsp;on twigs similar to those of Araucaria excelsa and scales characterised by stout terminal spines. Both Araucarites Falsani and A.nbsp;microphylla are from Kimeridgian strata in Ain.

German cone-scales hardly distinguishable from some of the British and French examples are described by Sir William Thisel-ton-Dyer� from Solenhofen as Araucarites Haherleinii (fig. 740, 5).nbsp;Salfeld figures some detached scales from the Malm of South-West Germany as Araucaria 1 which are undoubtedly Araucariannbsp;cone-scales agreeing closely with A. Milleri from Scotland (fig.nbsp;740, 1).

Araucarites Rogersi Seward. The scales of this species (fig. 740, 3)* from the Uitenhage (Wealden) series of Cape Colony reachnbsp;a length of 3 cm. and the straight distal margin bears a mediannbsp;spine; there is no indication of a ligule. A specimen in the Britishnbsp;Museum collected by Atherstone and referred to by Tate� showsnbsp;several scales still in their natural position. This type bears anbsp;striking resemblance to some of the Indian specimens describednbsp;by Feistmantel and is similar to the North American Neocomiannbsp;species A. wyomingensis Font.

Several examples of typical Araucarian cone-scales are figured by Feistmantel from Upper Gondwana rocks in India. Thenbsp;scales described as A. macropterus*^ are distinguished by theirnbsp;large size; specimens from the Kajmahal series reach a breadth ofnbsp;5 cm. The scales of A. cutchensis'^ recorded from the Cutch floranbsp;(fig. 740, 8) and elsewhere are smaller, but in some cases it is

^ Tate (67) p. 147. nbsp;nbsp;nbsp;� Feistmantel (77^) p. 180, PI, viii. figs. 9�12

impossible to draw any sharp line between the two species; they agree very closely with both British and French Jurassic types.

Araucarites cutchensis Feist, is recorded by Halle^ from the Upper Jurassic flora of Graham Land. The scales exhibit a considerable range in size and shape and more than one type may benbsp;represented. They are always more or less cuneate and have anbsp;narrow truncate base; some of them show broad lateral wing-likenbsp;extensions: the distal end is nearly truncate and bears a narrownbsp;linear appendage. As Halle says, the scales closely resemble thosenbsp;of A. Brodiei Carr.

Cone-scales from Jurassic rocks in Victoria^, Australia, described as Araucarites sp., A and B, demonstrate the occurrence of cones with scales almost identical with A. Phillifsi and othernbsp;European forms. There is a comparative scarcity of Araucariannbsp;cone-scales in Jurassic and Cretaceous strata in North America butnbsp;some examples are recorded. Araucarites wyomingensis Font.� fromnbsp;the Lower Cretaceous of the Black Hills is represented by broadlynbsp;cuneate scales 1-2 cm. long and with a maximum breadth of 9 mm.nbsp;and a broad beak at the apex bearing seeds 4�5 mm. long.nbsp;Larger cone-scales are described by Berry^ from Middle Cretaceousnbsp;rocks in North Carolina as Araucarites Jejfreyi (fig. 740, 2). Thesenbsp;scales are associated with the foliage-.shoots referred to Araucaritesnbsp;bladensis and the two may belong to one plant. The same authornbsp;also figures a specimen from the Upper Potomac series as Araucarites patafscoensis^ from Virginia.

There are numerous examples of foliage-shoots among Mesozoic, and to a less extent Palaeozoic, strata which bear a striking resemblance to branches of recent species of Araucaria, especiallynbsp;species of the Eutacta section, but in many cases confirmatorynbsp;evidence such as would be afforded by reproductive shoots isnbsp;lacking. The practice adopted by some authors of referring

^ Fontaine in Ward (99) B. p. 669, PI. CLXiii. figs. 1-* Berry (08) p. 258, PI. xvi.; (14) p. 20. s Ibid. (11) p. 399, PI. Lxxvii. fig. 5.

impressions of vegetative branches to the genus Araucarites solely on the ground of similarity in habit and leaf-form to the recentnbsp;genus is not in accordance with sound principles, though in somenbsp;instances the implied relationship may be a reality. Pendingnbsp;more satisfactory evidence many of the sterile Araucaria-\�ienbsp;shoots are referred to Pagiophyllum, while branches of similarnbsp;habit bearing oval cones are included in the genus Elatides.

An example of a Palaeozoic fossil which has been assigned to Araucarites on slender grounds is afforded by Araucarites Oldhaminbsp;Zeiller.

The specimen from the Lower Gondwana rocks in India to which this name is applied^ consists of an axis bearing spirallynbsp;disposed lanceolate-acuminate leaves reaching 4-5 cm. in length,nbsp;slightly contracted at the base and longitudinally striated:nbsp;portions of the axis show rhomboidal and feebly convex areasnbsp;separated by narrow scars where the laminae have been broken off.nbsp;As Zeiller says, the resemblance of the shoot to a branch of Araucaria irnbricata is very close, but considering the age of the bedsnbsp;and the absence of any Araucarian cone-scales from rocks at thisnbsp;horizon in India it is questionable whether it is wise to adoptnbsp;the name Araucarites. It is not unlikely that a small specimennbsp;figured by Feistmantel^ from the Karharbari coal-field as possiblynbsp;a Fern rhizome is a portion of a leafless axis of Zeiller�s species.

This species from Cretaceous strata in New Jersey� is founded on fragments of sterile branches bearing elliptical-ovate leavesnbsp;resembling the foliage of Araucaria irnbricata and the fossil speciesnbsp;Araucarites Nathorsti Dus. A very similar type is represented bynbsp;Araucarites bladenensis, described by Berry^ as Araucaria hladen-ensis, from the Upper Cretaceous rocks of Carolina and Alabama:nbsp;the leaves are decurrent, ovate-lanceolate, about 1-6 by 0-8 cm.nbsp;with a cuspidate apex and rounded base; there are 14�16 parallel

' Zeiller (02) B. p. 36, PI. vii. fig. 6. nbsp;nbsp;nbsp;� Feistmantel (79^) PI. xiii. fig. 6.

'� Berry (08) p. 255, Pis. xii.�xiv.; (14) pp. 19, 105, PI. iii. figs. 6, 7; PI. xix. figs. 1, 2.

veins and imperfectly preserved stomata occur in rows on the lower surface. Berry compares the species with Araucaria Tou-casi figured by Saporta^ from Turonian rocks in the South ofnbsp;France. Another species founded on a sterile shoot is Araucantesnbsp;Hatcheri described by Wieland^ from Upper Cretaceous rocks innbsp;Wyoming.

These and other examples that might be quoted, though referred to the Araucarineae on evidence that cannot be considered conclusive, are probably correctly determined; the comparisonnbsp;with Araucaria Bidwilli and A. imbricata suggested by the strikingnbsp;resemblance of the leaves is supported by the occurrence of Arau-carian cone-scales' in some of the localities.

This species was founded on sterile branches, from the rich Eocene flora of Haring in the TyroU, practically identical in habitnbsp;with foliage-shoots of Araucaria excelsa and other recent species.nbsp;From the same locality Goeppert^ figured an imperfectly preserved cone approximately 6 cm. long and 3 cm. in diameternbsp;characterised by imbricate, spirally disposed scales with reflexednbsp;apices which he compares to a male cone of Araucaria imbricatanbsp;incorrectly spoken of as A. excelsa: Goeppert suggests the possiblenbsp;specific identity of A. Sternhergii and A. Goepferti Sternb.: thenbsp;latter species was founded by Sternberg� on a Tertiary cone fromnbsp;Haring in the Tyrol. Ettingshausen� subsequently figurednbsp;several goo4 examples of vegetative shoots of this type fromnbsp;Haring and described a subglobose cone, figured by Gardner�,nbsp;which he refers to the same species: this author also recordsnbsp;A. Sternhergii from Bilin in Bohemia*, but under the genericnbsp;name Sequoia: in his account of the occurrence of the speciesnbsp;in Carinthia� he adopts the designation Araucarites. Ettings-hausen figures a single cone-scale from Eocene beds in Styria asnbsp;Araucarites .schoeneggensis^^ and conipares it to the scales of

A. Sternbergii. The latter species is recorded also by Massalongo^ from Eocene rocks in Italy and on imperfect evidence by Heer^nbsp;from Switzerland. Gardner� describes several good specimens ofnbsp;vegetative shoots from the Eocene flora of Bournemouth whichnbsp;he names Araucarites Goepperti Sternb. though the specific namenbsp;Sternbergii would be more appropriate ^nbsp;as that designation was first appliednbsp;to similar branches from Haring andnbsp;A. Goepperii was founded on a detached cone. Two small pieces ofnbsp;larger specimens in the British Museum from Bournem.outh are represented in fig. 741 in illustration of thenbsp;very close resemblance of the leavesnbsp;to those of recent species. Gardnernbsp;draws attention to the similarity ofnbsp;some of the fossil examples to deciduous shoots of Araucaria Cunning-hamii: with reference to the absencenbsp;of cones or cone-scales he quotes the Big. 741. Araucarites Sternbergii.nbsp;fact, communicated to him by an (British Museum, V, 523; nat.nbsp;observer in Madeira, that the foliage

of A. Cunninghaniii requires two or three days to sink while mature seeds do not begin to sink before the fifth or sixth day,nbsp;so that in moving water shoots and seeds would necessarily benbsp;deposited separately.

Some of the fragments of branches described by Gardner as Athrotaxis (?) subulata* may well belong to Araucarites. It mustnbsp;be admitted that in the case of the English specimens, as in manynbsp;others, the use of the generic name Araucarites is based on thenbsp;evidence of vegetative branches only, but Gardner correctly statesnbsp;that in the shoots of similar habit referred to Cryptomeria thenbsp;leaves are straighter, and moreover the presence on some of thenbsp;shoots of the latter of persistent cones like those of Cryptomerianbsp;japonica constitutes a clear distinction. Having regard to thenbsp;very striking resemblance of the widely spread Tertiary specimens

included in Araucarites Sternbergii or A. Goepperti to those of A. Cunninghaniii and other species the probability of genericnbsp;identity is such as to justify the retention of the designationnbsp;Araucarites.nbsp;nbsp;nbsp;nbsp;'

Ettingshausen^ described this species as Araucaria Haastii from beds at Shag Point, New Zealand, believed to be of Eocenenbsp;age; it is represented by sterile branches bearing crowded ovate-lanceolate, acuminate, leaves apparently of leathery texturenbsp;reaching a length of 5 cm. and 2 cm. or less in breadth. Asnbsp;Ettingshausen says, they agree very closely with the leaves ofnbsp;Araucaria imbricata but like those of A. Nathorsti Dus. they havenbsp;a less spinous apex than in the recent species. Some petrifiednbsp;wood from Malvern Hills in New Zealand is referred by Ettingshausen to the same species but without any evidence of connexion between the wood and the foliage-shoots. The samenbsp;author describes a branch similar in habit to Araticaria excelsa,nbsp;horn Shag Point, as Araucaria Danai^, but the specimen is toonbsp;imperfect to warrant the use of the designation Araucarites.

This species, described as Araucaria Nathorsti, is recorded by Dus�n� from Punta Arenas on the Magellan Straits: the age of thenbsp;beds is believed to be Oligocene though the precise horizon hasnbsp;not been determined. The material consists of fragments ofnbsp;foliage-shoots bearing short and relatively broad leaves of leatherynbsp;texture, varying from linear to ovate; they agree closely with thenbsp;leaves of Araucaria imbricata, differing chiefly, as Dus�n states,nbsp;in their blunter apices.

Nathorst^ first suggested a reference to Araucaria of the single leaf on which this species was founded�; it was collected in anbsp;marine volcanic tuff in Seymour Island and is probably of Lowernbsp;Tertiary age. The leaf is linear, 6 cm. long, and tapers graduallynbsp;towards an incomplete apex; it agrees in form and size with

^ Ettingshausen (87) p. 154, PI. il. nbsp;nbsp;nbsp;^ Ibid. p. 155, PI. i. fig. 18.

leaves of A. Bidwilli and A. hrasiliensis, but the single impression is hardly sufficient to demonstrate the existence of Araucaritesnbsp;in this southern flora (lat. 64� 16' S.)- On the other hand thenbsp;occurrence of wood of the Araucarian type^ in Seymour Islandnbsp;in beds that are either Lower Tertiary or Upper Cretaceousnbsp;supports the conclusion of Nathorst and Dus�n.

Heer^ proposed this name� for some Jurassic Coniferous remains from Siberia characterised by spirally disposed falcatenbsp;leaves (figs. 742, 743) and cones similar externally to those ofnbsp;Picea, Abies and other Abietineae. The genus is based primarilynbsp;on the form of the cones and cone-scales. In the new genusnbsp;were included three species, E. ovalis, represented by oval conesnbsp;2-7 cm. long and 6�7 mm. broad, E. Brandtiana characterised bynbsp;cylindrical cones, and E. falcata founded on vegetative branchesnbsp;very similar to those of E. Williamsonis. Nathorst^ has includednbsp;these three species in Elatides curvifolia (Dunk.) a Wealdennbsp;species abundantly represented in the plant-beds of Spitzbergennbsp;(fig. 743). In the absence of cones it is impossible to draw anynbsp;satisfactory distinction between foliage-shoots belonging to Elatidesnbsp;and those referred by authors to Sequoia, Pagiofhyllum, and othernbsp;genera. It is therefore only in cases where cones are presentnbsp;that the designation Elatides is admissible. The vegetativenbsp;characters of Elatides are those of Araucaria Gunninghamii, A.nbsp;excelsa, and allied species while the cones consist of flat imbricatenbsp;scales with narrower and more or less pointed or spinous distalnbsp;ends. There is some reason to believe that the cone-scales werenbsp;monospermic but the evidence is not conclusive and rests on anbsp;single species. The data are insufficient to fix definitely the position of the genus, though it is in all probability a member of thenbsp;Araucarineae. Elatides is characteristic of Rhaetic, Jurassic, andnbsp;Wealden floras.

A Rhaetic species originally described by Nilsson from Rhaetic rocks in the South of Sweden as Abies Sternbergii, subsequently

^ Gothan(08). nbsp;nbsp;nbsp;^ Heer (77) ii. pp, 77�79, PI. xiv.nbsp;nbsp;nbsp;nbsp;� �Xari], Pir.

included by Nathorst^ in Palissya Braunii but afterwards recognised by him as a distinct species^ and recently transferred to Elatides^. An examination of specimens in. the Stockholmnbsp;Museum leads me to agree with the substitution of the designationnbsp;Elatides. Nathorst has also pointed out that some of the conesnbsp;from the Ehaetic of Franconia referred by Schenk^ to Palissyanbsp;Braunii are of the Elatides type and distinct from cones of Palissya which are characterised by their more open habit and bynbsp;other more important morphological features. Elatides Sternbergii,nbsp;though similar in the habit of the vegetative shoots to E. Williarn-sonis from Jurassic strata, differs in the narrower and straighternbsp;leaves which may reach a length of 2�3 cm. and are either straightnbsp;or slightly curved in contrast to the stouter and strongly falcatenbsp;leaves of E. Williamsonis and E. curvifolia. A cone figured bynbsp;Nathorst� is practically identical in external form with one ofnbsp;E. Williamsonis illustrated in volume i. of the Jurassic Flora of thenbsp;Yorkshire Coast^. We have no knowledge of the structure of thenbsp;reproductive shoots and no evidence other than the habit of thenbsp;foliage-shoots with regard to systematic position: it is, however,nbsp;probable that this Rhaetic species is closely allied to the laternbsp;Jurassic and Wealden types.

This Jurassic species described by Brongniartas Lycopodites Williamsonis, was figured by Phillips� as L. uncifolius and bynbsp;Bindley and Hutton� under Brongniart�s name. The specimensnbsp;figured by the English authors are in the York and Manchesternbsp;Museums respectively. Schimper transferred the species to Pachy-phyllum and it has usually been assigned to that genus or tonbsp;Pagiofhyllum^^, the name substituted by Heer for Pomel�s Pachy-phyllum, a designation now reserved for sterile shoots and therefore inapplicable to the present species which possesses conesnbsp;of the Elatides type. The vegetative shoots are monopodially

^ Nathorst (86), PI. xxv. fig. 8. nbsp;nbsp;nbsp;� Seward (00) B. PI. x. fig. 3.

� Brongniart (28) A. p. 83. nbsp;nbsp;nbsp;� Phillips (29) A. PI. viii. fig. 3.

branched, the smaller branches being given off at an acute angle; the leaves are crowded, fleshy, tetragonal and falcate (fig. 742),nbsp;agreeing closely with the foliage of the Eutacta species of Araucarianbsp;and with Cryptomeria. The megastrobili are cylindrical, approximately 6 cm. long and 2 cm. in diameter, bearing imbricate, flat,nbsp;scales with narrow pointed distal ends resembling the free portion

Ac!

of the foliage-leaves. No specimens have been described showing seeds attached to the scales. The more slender microstrobili,nbsp;2 cm. long, bear sporophylls at right angles to the axis with triangular upturned distal ends characterised by a median keel. Innbsp;the vegetative shoots this species closely resembles the Liassicnbsp;Pagiophyllum peregrinum^ (fig. 744) (Lind, and Hutt.), but in the

absence of cones the latter species is retained in Pagiophyllum. From the Wealden species E. curvifolia, E. Williamsonis differs innbsp;its stouter and more crowded leaves though the differences arenbsp;slight both in the cones and vegetative shoots. In all probabilitynbsp;this species is represented in several Jurassic floras, but unlessnbsp;cones are present specimens should be referred to Pagiophyllum.

Dunker^ first described this Wealden type (fig. 743) from North Germany as Lycopodites and it was referred by Ettingshausen^ tonbsp;Araucarites, the generic name, though probably correctly express-

ing the position of the fossil Conifer, being used without adequate reasons. The identity of the cones discovered by Nathorst� innbsp;the Wealden or Upper Jurassic beds of Spitzbergen with those onnbsp;which Heer founded the genus Elatides led to the adoption of thatnbsp;generic term. Nathorst�s discovery of several fertile branchesnbsp;justifies his reference of Heer�s specimens from Spitzbergennbsp;described as Sequoia ReichenbachiP to Elatides curvifolia, as alsonbsp;the employment of Hunker�s specific term for Elatides ovalis andnbsp;E. Brandtiana Heer. The cones of E. curvifolia are cylindrical ornbsp;oval and it is suggested by Nathorst that these forms might benbsp;regarded as varieties, the oval form being spoken of as var. ovalis

after Heer�s specific name and the cylindrical cones being distinguished as var. Bmndtiana, but it is doubtful whether the retention of these varietal names is advisable. The cone-scales have pointednbsp;apices and agree closely with those of E. Williamsonis. On anbsp;specimen of a cone of this species from Kimeridge strata innbsp;Scotland^ one scale afforded evidence of the occurrence of a singlenbsp;seed as in Araucaria. Nathorst regards some smaller cones on anbsp;branch from Spitzbergen as microstrobili and suggests that longitudinal striae on the sporophylls may represent long microsporangia like those of Araucaria; but the preservation is too imperfectnbsp;to demonstrate the nature of the specimen. The vegetativenbsp;branches bear falcate leaves rather more slender and as a rule lessnbsp;crowded than in E. Williamsonis: on older branches from whichnbsp;the free part of the lamina has fallen there are leaf-bases or in somenbsp;cases an oval leaf-scar. This type is characteristic of Wealdennbsp;strata in Spitzbergen, North Germany, and other Europeannbsp;localities; it is no doubt represented by some of the impressionsnbsp;of branches assigned to Sphenolepidium Sternbergianuni^ �, it isnbsp;also recorded from Kimeridgian strata in the North of Scotland.

Heer� instituted this genus in place of Pachyphyllum, previously adopted by PomeH for a section of his genus Moreauia, on thenbsp;ground that the latter name had been applied to a member of thenbsp;Orchidaceae. Some of the species referred to Pagiophyllum havenbsp;also been included in Araucarites and Brachyphyllum. Tuzson�nbsp;instituted a new genus Pagiophyllites for petrified Mesozoic woodnbsp;having Araucarian features, the type-species being P. Iceuperianusnbsp;(Goepp.), but no evidence is furnished in support of a connexionnbsp;of this wood with foliage-shoots of Pagiophyllum. Schimper andnbsp;Saporta include in their diagnosis of the genus both vegetativenbsp;and reproductive shoots and consider Pagiophyllum to be alliednbsp;to Agathis, Cunninghamia, and Araucaria. Certain species havenbsp;in recent years been transferred to Elatides because of the occurrence of cones conforming to Heer�s genus.

1 Seward (IP) p. 684, fig. 10. nbsp;nbsp;nbsp;quot; Ibid. (95) A. p. 205; (IP) p. 685.

ently be reserved for vegetative branches of Conifers (fig. 744) possessing foliage like that of Araucaria excelsa and allied species,nbsp;which in the absence of cones cannot safely be referred to Elatidesnbsp;or other genera based, in part at least, on strobilar characters.nbsp;Pagiofhyllum is essentially an artificial genus; as Solms-Laubachnbsp;says, �it is only in accordance with old custom to distinguish thenbsp;Ullmanniae of the Zechstein from Pagiophylliim^,� and it is equallynbsp;difficult to draw any clearly defined line between this genus andnbsp;some forms included by authors in Brachyphyllum. A Triassicnbsp;species from Raibl originally referred to Voltzia heterophylla^nbsp;afterwards named V. Foetthri by Stur�, Pagiophyllum Sandberginbsp;by Schenk^ and figured by Sch�tze� as P. Foettleri, has the habitnbsp;of a Brachypliyllum. This is one of many examples of sterilenbsp;shoots illustrating the arbitrary use of generic names for coniferous remains which afford no definite evidence of their systematicnbsp;position. The Araucarian habit is in itself of little value asnbsp;evidence of affinity, but the abundance of petrified wood withnbsp;Araucarian features (Dadoxylon) in strata yielding Pagiofhyllumnbsp;shoots suggests an Araucarian alliance, and the fact that somenbsp;Pagiofhyllum shoots bear Elatides cones affording indications ofnbsp;Araucarian characters points in the same direction. It cannotnbsp;be assumed that all Pagiofhyllum shoots bore similar cones, andnbsp;it is mainly on this account that the employment of Pagiofhyllumnbsp;as a provisional designation is recommended.

Pagiofhyllum is widely distributed in Jurassic strata and extends into Cretaceous and Tertiary rocks; it occurs also innbsp;pre-Jurassic floras and has recently been described by Zeiller�nbsp;from the Permian of France. It should be recognised that thisnbsp;extended use of the name is not in accordance with general practice,nbsp;but it is adopted on the ground that, as in recent Conifers so in thenbsp;case of extinct types, similarity in the habit of vegetative branchesnbsp;does not necessarily imply close relationship as regards the morenbsp;important characters of the reproductive shoots.

This species, first named by Lindley Araucaria peregrina^, was founded on material from the Liasnbsp;of Lyme Regis in Dorsetshire^.

It is possible that the generic name chosen by Lindley correctlynbsp;expresses the position of thenbsp;species, but decisive evidence isnbsp;lacking. Vegetative shoots bearnbsp;crowded imbricate, spirally disposed, leaves tetragonal in section,nbsp;broadly triangular, sometimes falcate and more or less appressed tonbsp;the stem in the lower portion ofnbsp;the lamina (fig. 744), There is anbsp;distinct dorsal keel and occasionally rows of papillae are visiblenbsp;on the lamina; the apex is obtusenbsp;or acute. The leaves vary considerably in size and shape. Zeiller�nbsp;describes the cuticle of the dorsalnbsp;and ventral surfaces of somenbsp;leaves on Permian specimens fromnbsp;Blanzy: the stomata occur innbsp;longitudinal rows on the lowernbsp;face only, the guard-cells beingnbsp;usually at right-angles to the longnbsp;axis of the leaf.

In habit this species agrees closely with Elatides William-sonis, a Middle Jurassic type; itnbsp;occurs in Jurassic rocks of England, France, Germany, Italy, andnbsp;elsewhere, the oldest recorded examples being those described by

� Seward (04) B. p. 48, PJ. v.; Saporta (84) p. 383, Pis. 173-� Zeiller (06) B. p. 219.

Zeiller from the Permian of France which he refers to Pagio-�phyllum in preference to Ulhnannia, pointing out that the leaves are relatively longer and less appressed to the axis than in thenbsp;shoots known as �. Bronni Goepp. (fig. 750, D, E). Triassicnbsp;specimens from North Italy in the Bologna Museum namednbsp;Pagiofliyllum Rotzoanum appear to be indistinguishable from thenbsp;English species.

It is unnecessary to describe other examples of the genus as the PagiofJiyllum type is illustrated by many Mesozoic andnbsp;Tertiary species referred by authors to Sequoia, Geinitzia, Elatides,nbsp;Sphenolefidium, and other genera. The important point is thatnbsp;in place of generic names connoting definite forms of cone, thenbsp;designation Pagiophyllum should be adopted for all foliage-shootsnbsp;of a certain habit which afford no satisfactory evidence as to thenbsp;nature of the reproductive shoots.

Palaeozoic Conifers exhibiting certain features suggestive OF Araucarian affinity but which cannot be DEFINITELY ASSIGNED TO THAT OR TO ANY OTHER FAMILYnbsp;OF CONIPERALES ON THE EVIDENCE AT PRESENT AVAILABLE.

The name Walchia^ is applied to foliage-shoots, occasionally bearing terminal cones, from Permian and to a less extent Uppernbsp;Carboniferous rocks, which present a striking agreement in habitnbsp;with branches of Araucaria excelsa and other recent species of thenbsp;section Eutacta of Araucaria. Information with regard to reproductive shoots is very incomplete and we have little more thannbsp;circumstantial evidence as to the anatomical features of the stem.nbsp;In many cases the ultimate branches bear terminal cones similarnbsp;to the megastrobili of Elatides^, but it is only in a few specimensnbsp;that seeds are preserved on the cone-scales: in some species, e.g.,nbsp;IF. frondosa Ren. and IF. fertilis Ren. the fertile shoots appear tonbsp;be of a distinct type though the evidence is not wholly satisfactory.nbsp;It is probable, as several authors have suggested, that the speciesnbsp;included in IFafc^m, were our information fuller, would be referrednbsp;to more than one generic type. The resemblance of branches ofnbsp;Walchia to the foliage-shoots of Lefidodendron, especially in thenbsp;^ Sternberg (26) A. p. xxii.nbsp;nbsp;nbsp;nbsp;^ See page 272.

case of specimens too small to show the characteristic branching-habit, has led to confusion between the two genera. It is often very difficult to draw a definite line between Walcliia and UUmannia,nbsp;and in the absence of sporophylls the genus Gomphostrobus may benbsp;easily confused with species of Walchia.

Foliage-shoots characterised by a pinnate arrangement of the ultimate branches (fig. 745) attached at right-angles or obliquelynbsp;to an axis of higher order. Leaves spirally disposed, crowded andnbsp;imbricate, short and ovate or linear and spreading, usually tetragonal and more or less falcate and decurrent. The dimorphism ofnbsp;branches and differences due to age or position on the tree rendernbsp;a satisfactory delimitation of species almost impossible thoughnbsp;a few fairly well defined types can be recognised with reasonablenbsp;certainty. As Bergeron^ says, in the absence of strobili thenbsp;separation of species represented only by sterile shoots is hardlynbsp;possible. Further reference is made to the features exhibited bynbsp;reproductive shoots in the appended account of a few selectednbsp;types. Information with regard to the anatomical characters ofnbsp;Walchia is very scanty and is based on evidence afforded by thenbsp;association of foliage-shoots and petrified wood or on inferencesnbsp;drawn from unconvincing considerations. Among specimensnbsp;which may belong to this genus one of the more interesting is thatnbsp;on which Mougeot^ founded the species Araucarites valdajolensis.nbsp;The type-specimen, from the Permian of Val d�Ajol in the Vosges,nbsp;has been refigured and critically discussed by Fliche� though nonbsp;complete investigation of its structure has been made. Thenbsp;cylindrical piece of stem, 9 cm. in diameter, has a large pith and anbsp;broad zone of secondary wood composed of tracheids, with twonbsp;rows of alcernate bordered polygonal pits, and narrow medullarynbsp;rays. It'seems clear from Mougeot�s brief account and from thenbsp;description of other specimens by Fliche that the anatomicalnbsp;features are Araucarian though we have no information as tonbsp;the structure of the inner edge of the xylem, a region of specialnbsp;importance as regards comparison with other types possessingnbsp;a similar Araucarian pitting on the tracheids. The surface ofnbsp;Mougeot�s specimen is characterised by numerous spirally disposed, elliptical projections 5�7 mm. long and 3�4 mm. wide

1 Bergeron (84). nbsp;nbsp;nbsp;2 Mougeot (52) A. p. 27, PI. iv.nbsp;nbsp;nbsp;nbsp;� Fliche (03).

which Fliche regards as leaf-bases and compares with those on Araucarian stems. While admitting the possibility that the woodnbsp;belongs to Gomphostrobus or some Cordaitean species Fliche considers WalcMa the most likely genus. The comparative closenessnbsp;of the leaf-bases would seem to be a difficulty: in stems as largenbsp;as the type-specimen of W. valdajolensis one would expect to findnbsp;the leaf-bases more widely separated and tangentially stretched.nbsp;It is by no means unlikely that the supposed surface-features maynbsp;belong to a deeper zone of the cortex of a partially decorticatednbsp;stem; but in any case they do not suggest a stem of Cordaites ornbsp;Mesoxylon. It is impossible to assign the species with confidencenbsp;to Walchia though Fliche may be correct in his opinion as to thenbsp;likelihood of that being its true position. The generic namenbsp;Araucarites implies a degree of affinity which has not been established and the designation Dadoxylon would be more in keepingnbsp;with the facts.

Walchia is especially characteristic of Permian floras though it has been shown to occur in the Stephanian of several countries.nbsp;In Britain Walchia is recorded from a very few Permian^ andnbsp;Upper Coal Measures'^ localities. We cannot speak with confidencenbsp;as to the position of the genus: the striking resemblance in thenbsp;system of branching and in the foliage-shoots to certain species ofnbsp;Araucaria at once suggests a possible affinity to the Araucarineae,nbsp;and this slender basis of comparison receives support from thenbsp;occurrence in a few instances of single seeds on the upper face ofnbsp;sporophylls and from the Araucarian type of pitting in woodnbsp;associated with Walchia branches. It may fairly be said thatnbsp;although proof is lacking there is a strong presumption in favournbsp;of regarding this Permo-Carboniferous genus as more nearly alliednbsp;to the Araucarineae than to any other family of Gymnosperms.

This, the commonest species, was originally described by Schlotheim as Lycopodiolithes piniformis^: it occurs in bothnbsp;Permian and Stephanian strata^. The pinnately branched shoots

^ ^'.(jr. Stefani (01) p. Ill; Weis.s, C. E. (72) p. 179; Heer (76) A. p. 57; Goeppert (65) p. 2.36; Grand�Eury (77) A. p. 514.

(fig. 745) are characterised by the more or less oblique insertion of the slender branchlets and by the comparatively long, narrow,nbsp;falcate, decurrent leaves. The ovoid or cylindrical megastrobilinbsp;terminal on the ultimate shoots bear imbricate ovate-lanceolatenbsp;sporophylls, but their preservation is not such as to throw anynbsp;light on the structure of the seed-bearing organs. Zeiller^ mentionsnbsp;a cone from Lod�ve (Permian) 10 cm. long and 1�1-2 cm. innbsp;diameter, but the average length is less than this. A branchnbsp;figured by Potoni�^ from the Permian of Thuringia shows ellipticalnbsp;leaf-cushions very like those on the larger stem described bynbsp;Mougeot as Araucarites valdajolensis. Potoni� has drawm attention to the difficulty of distinguishing small specimens of thisnbsp;species from W. filiciformis and W. linearifolia, and some formsnbsp;described as W. imbricata and W. hypnoides are by no meansnbsp;clearl}^ distinguished from W. piniformis. In W. filiciformisnbsp;(Schloth.) the leaves are characterised by the downward curve ofnbsp;the lamina near the base though this in itself is hardly a decisivenbsp;criterion. In W. linearifolia Goepp. the leaves are rather morenbsp;delicate and less falcate, while in W. imbricata they are usuallynbsp;shorter, relatively broader, and more strongly imbricate andnbsp;incurved. W. hypnoides (Brongn.) is a smaller form though, asnbsp;Kidston� suggests, this may not be a specific character. Renanlt�snbsp;species W. fertilis* represents a similar form but with smallernbsp;leaves, and each branchlet ends in a long and narrow strobilusnbsp;'vhich affords no indication of the nature of the sporophylls.

This species, recorded from Permian and Stephanian beds, and differing but slightly from IF. piniformis is important as supplyingnbsp;more satisfactory evidence as to the nature� of the megastrobili.nbsp;Zeiller� has described a fertile specimen from the Permian ofnbsp;Rrive in which the ovate-lanceolate cone-scales (sporophylls) bearnbsp;on their upper concave face single ovoid seeds, 7�8 mm. long.nbsp;The strobili appear to be lax in the arrangement of the monospermicnbsp;sporophylls.

In habit this type^ closely resembles the foliage-shoots of Araucaria Rtdei on a smaller scale, the crowded leaves beingnbsp;strongly incurved and imbricate: it is readily distinguishablenbsp;from Ullmannia Bronni G-oepp. (fig. 750, D, E). A good examplenbsp;is figured by Zeiller^ in which the branches are unusually large,nbsp;8 mm.�1-2 cm. in diameter: the species occurs in Permian andnbsp;Stephanian strata and is recorded by Kidston from the Uppernbsp;Coal Measures of central England.

This species� from Charmoy is characterised by the long filiform leaves, -6 mm. broad and 1-2 cm. long, usually straight andnbsp;decurrent. The ultimate branches, some of which bear comparatively long and slender cones, are oblique and alternate as innbsp;W. piniformis. W. foliosa Eich.^ from the Permian of Russia is anbsp;similar form but with less delicate leaves.

A species from the Permian of Autun having slender leaves rather shorter and more falcate than those of W. Schneideri:nbsp;some of the branches bear a terminal globular bud superficiallynbsp;resembling the ovuliferous shoot of Taxus. It is, however, notnbsp;improbable that the buds are purely vegetative like those figurednbsp;by Bergeron on a specimen of W. piniformis.

Prof. C. E. Weiss� instituted the generic name Tylodendron for casts from Upper Carboniferous and Permian strata in Germanynbsp;which he described as branches of a Conifer with spirally arrangednbsp;rhomboidal raised areas or pulvini each of which has a mediannbsp;slit in its apical portion (fig. 746). The elongate, narrow, raisednbsp;areas {cf. the medullary cast of a Voltzia, which shows preciselynbsp;similar areas, represented in fig. 748) were regarded as casts of leaf-cushions and the slit was interpreted as a resin-canal.

1 Schimper (72) A. p. 239. nbsp;nbsp;nbsp;^ Zeiller (06) B. p. 211, PI. xlix. figs. 1, 2.

Schizodendron speciosum (Weiss). One of the casts figured by Weiss and assigned by him to this species has a length of 70 cm.nbsp;and at intervals of about 30 cm. shows periodic swellings where itnbsp;assumes a barrel-shaped form. Pieces of wood attached to somenbsp;of the casts were investigated bynbsp;Dippel who found that they agreednbsp;anatomically with Araucarian stems.

Weiss considered his specimens to be generically identical with casts figurednbsp;by Eichwald^ from Russia as speciesnbsp;of Schizodendron and Angiodendron.

Potoni�, while uncertain as to the close agreement with some of Eich-wald�s fossils, regarded Tylodendronnbsp;as identical with Eichwald�s Schizodendron, and Zeiller^, in view of thisnbsp;agreement, adopted the older namenbsp;Schizodendron. The latter author�nbsp;formerly beheved Tylodendron, asnbsp;described by Weiss, to be distinguished from Schizodendron by thenbsp;apical occurrence of the slit on thenbsp;so-called leaf-cushion in contrast tonbsp;the basal slit in Schizodendron, butnbsp;Potoni�^ proved that in both casesnbsp;the median groove extends up thenbsp;lower portion of each projecting Fig. 746. Schizodendron specio-area from its base and representsnbsp;nbsp;nbsp;nbsp;(After Potom�; f nat.

an out-going leaf-trace; he also *^**��^ demonstrated that Schizodendron is a pith-cast, the taperednbsp;areas being the inner ends of medullary rays. In some specimensnbsp;the casts afford some indication of a discoid pith. The relationnbsp;between the wood and the pith-casts is also very clearly shownnbsp;in a section of a petrified stem of Permian age from Princenbsp;Edward Island described bv Miss Holden�. The true nature of

tlie Tylodendron casts is also shown in specimens from the Lower Permian of Saxony in the Chemnitz Museum^. Casts similar tonbsp;those described as Tylodendron and Schizodendron were recordednbsp;by Schleiden in 1846 and referred to a new genus Bndolepis: henbsp;believed them to be casts of the pith-cavity of some Dicotyledonousnbsp;stem. Examples of Endolepis have been described by Schenk andnbsp;more recently by Eliche^ who discusses the history of the genusnbsp;and on the ground of priority adopts Schleiden�s name in preference to Schizodendron. It is, however, preferable to retainnbsp;Schizodendron for the larger oasts with periodic swellings. Thenbsp;smaller type represented by Endolepis has in several instancesnbsp;been found in connexion with the foliage of Voltzia^ (fig. 748) andnbsp;it is questionable if a special designation is needed. In the formnbsp;of the raised areas on the surface of the cast Schizodendron andnbsp;Endolepis appear to be identical: while suggesting the advisabilitynbsp;of retaining the former name I recognise that the Permian andnbsp;Triassic casts may belong to stems which are closely allied or evennbsp;generically identical.

The structure of the wood of Schizodendron speciosum is of the Araucarian type; the tracheids have 1�3 rows of contiguous andnbsp;alternate pits on the radial walls and the medullary rays arenbsp;usually uniseriate. Potoni� compares the pith-casts of the Palaeozoic stems with those of recent species of Araucaria and Agathis:nbsp;the pith of the recent species is much smaller but in both fossilnbsp;and recent medullary casts there are periodic swellings where thenbsp;presence of scars, sometimes in a whorl or pseudowhorD, marksnbsp;the position of branches. The pith-cast of a recent Cycad (fig. 398,nbsp;p. 29, Vol. III.) bears a general resemblance to Schizodendron : innbsp;Araucaria the medullary raj^s are narrower and so producenbsp;narrower raised areas on a pith-cast.

Prof. F. E. Weiss� has recently described an interesting example of Schizodendron which throws some fresh light on structuralnbsp;features. The specimen was found in Cheshire but not in situnbsp;and nothing is known as to its geological age; it consists of a

petrified barrel-shaped piece of pith with portions of the inner edge of the xylem-cylinder. The surface-features agree with thosenbsp;of S. sfeciosum, each rhombic area being divided for a third of itsnbsp;length by a median groove. The pith is composed of thin-walled .nbsp;parenchyma with several secretory canals in the outer region;nbsp;patches of xylem are preserved in the depressions between thenbsp;lozenge-shaped areas showing the same anatomical characters asnbsp;those described by Potoni�: internal to the secondary xylem are verynbsp;small groups of tracheids separated by 1�2 rows of parenchymanbsp;from the secondary elements, which pursue a sinuous longitudinalnbsp;course. These tracheal strands are, as Weiss points out, at leastnbsp;superficially comparable with the primary xylem of such a type asnbsp;Pitys antiqua. The innermost elements of the secondary xylemnbsp;are usually scalariform and these pass gradually into tracheidsnbsp;with two alternate rows of bordered pits often slightly polygonal.nbsp;The leaf-traces are formed of two endarch strands which coalescenbsp;as they pass downwards and eventually merge laterally with thenbsp;secondary xylem.

Medullary casts with the external features of Schizodendron might well belong to stems which are not identical in anatomicalnbsp;characters, and from casts alone all that can be inferred is thenbsp;presence in the vascular cylinder of medullary rays with fairlynbsp;broad inner faces separated by prominent wedges of tracheids,nbsp;also the spiral disposition of leaves each supplied with a singlenbsp;vascular bundle given off from the lower angle of the xylem-meshes. In the case of Schizodendron Cowardi the presence ofnbsp;small strands of primary xylem suggests comparison with such anbsp;genus as Pitys or Mesopitys, while in the other examples there isnbsp;no indication of any xylem internal to the main cylinder. Thenbsp;characters of the secondary xylem point to an Araucarian ornbsp;Cordaitean affinity and the pith agrees with that of Araucaria,nbsp;though in S. Cowardi the presence of secretory canals is a Cycadeannbsp;feature. Bain and Dawson \ though they did not correctly interpret the surface-characters of Schizodendron, referred to it asnbsp;representing decorticated branches of the Conifer Walchia. Sterzelnbsp;recorded the association of Schizodendron with Walchia foliage-shoots in Saxony, and Zeiller, who noticed a similar associationnbsp;^ Bain and Dawson (85).

in French Permian rocks, expressed the opinion that the casts belonged to Walchia stems. In this connexion it is noteworthynbsp;that shoots of Voltzia^ also possess medullary casts (fig. 748, A, B)nbsp;.with the superficial features of Schizodendron. Though we havenbsp;no proof of a connexion between casts and leaf-bearing branches,nbsp;it is probable that some forms of Schizodendron represent the pith-casts of Walchia: if this view is correct it affords another argumentnbsp;, in favour of connecting Walchia with the Araucarineae, but hownbsp;close the connexion is cannot be definitely settled without furthernbsp;evidence as to the reproductive shoots.

Schizodendron, though not confined to Permian rocks, is most abundant in beds of that age; it is recorded from several localitiesnbsp;in Germany^, from France, Russia�, and Canada while the Britishnbsp;specimen, though presumably from English rocks, was not foundnbsp;in situ.

Renault^ instituted this genus for a cylindrical stem 2 cm. in diameter from the Permian of Autun characterised especially bynbsp;the parenchymatous structure of the secondary xylem. Thenbsp;type-species, Hapaloxylon Rochei, resembles Araucarites valda-jolensis, a Permian species founded by Mougeot, in its spirallynbsp;disposed leaf-scars each with an elongated groove marking thenbsp;position of the leaf-trace. The solid parenchymatous pith isnbsp;surrounded by a narrow zone of 2�3 layers of tracheids with anbsp;single row of bordered pits which Renault speaks of as primarynbsp;xylem: this forms the inner edge of a broad cylinder of homogeneous parenchyma traversed by uniseriate medullary rays 1�3 cellsnbsp;deep. The secondary-xylem elements are rectangular 7�8 timesnbsp;as long as broad and without pits. Beyond the cambium is anbsp;broad zone of secondary phloem consisting of a regular alternationnbsp;of well preserved sieve-tubes with lateral sieve-plates and parenchyma. The cortex contains some secretory sacs and is boundednbsp;by periderm.

The inference drawn by Renault is that the leaves were small like those of Walchia and each had a single vein. There is,

1 Seward (90). nbsp;nbsp;nbsp;^ quot;vyeiss, C. E. (72); (74); Potonie (88); (93) A. etc.

however, no definite evidence as to the nature of the foliage; the stem structure represents a type previously unrecorded amongnbsp;Gymnosperms, but comparable with the structure of the stem ofnbsp;Aeschynomene a recent genus of the Leguminosae.

This generic name was instituted by Marion^ for Permian foliage-shoots from Lod�ve bearing vegetative leaves similar tonbsp;those of Wahhia and Araucaria, excelsa but distinguished by thenbsp;bifurcate form (fig. 747) of the relatively long sporophylls bornenbsp;in a crowded cluster on the apical region of the axis. Marionnbsp;referred Gomfhostrobus to the Coniferales. Geinitz� had previously described detached sporophylls from the Lower Permiannbsp;of Saxony, of the same type as those on which the genus Gom-phostro-bus was founded, as Sigillariostrobus bifidus. The Saxon specimensnbsp;are represented in the drawings published by Geinitz as distally-forked scales bearing a single seed at the base. Potoni�*, whonbsp;examined the original fossils in the Dresden Museum�and 1 amnbsp;able to confirm his view�states that there are no undoubtednbsp;seeds but only a faintly outlined area near the proximal end ofnbsp;each scale which no doubt marks the position of a seed or sporannbsp;gium.

Sigillariostrobus bifidus, Geinitz, Neues Jahrbuch Min. p. 700, PL v. figs. 5�7.

Dicranophyllum gallicuni, Schenk in Schimper and Schenk, p. 266. Oomphostrobus heterophyllus, Marion, Conipt. Rend. cx. p. 892.nbsp;Psilotiphyllum bifidnni, Potoni�, Ber. deutsch. Bot. Ges. Bd. ix.nbsp;p. 256.

Oomphostrobus bifidus, Zeiller, Bassin Houill. Perm. Brive, p. 101, PI. XV. fig. 12.

Potoni� in his account of the genus and type-species reproduces Marion�s original drawings showing a Walchia-Wke axis with short falcate leaves bearing crowded linear and distally forkednbsp;sporophylls reaching a length of 8 cm. in the apical region. Thenbsp;sporophylls, which are bent to one side, giving the impression ofnbsp;wind-blown foliage, consist of a simple lamina 8-5 to 25 mm. long,

1 See vol. II. p. 26. nbsp;nbsp;nbsp;2 Marion (90) A.nbsp;nbsp;nbsp;nbsp;* Geinitz (73).

with two divergent distal prongs varying considerably in the angle of divergence, a variation noticed also by Zeiller. On one examplenbsp;(fig. 747, B) Potoni� records the occurrencenbsp;of two scars; a lower scar, a, representingnbsp;the attachment of the lamina and a secondnbsp;scar, 6, which he attributes to a sporangium.

The species, represented usually by detached sporophylls only, is recorded from Lod�ve, Brive, and other French localities^nbsp;also from Permian localities in Germany.

It was referred by Schenk to Dicrano-phyllum, but in that genus it is the foliage-leaves that are forked and there is no reason to assume any close relationship betweennbsp;the two imperfectly known types. If the scarnbsp;at the base of the sporophylls marks the position of a seed a comparison with the Araucarineae is suggested, and in this connexion itnbsp;is noteworthy that SterzeP records the association of Gomphostrobusnbsp;with Dadoxylon wood. Potoni�, who at first overlooked Marion�snbsp;paper, proposed the name Psilotiphyllum to give expression to hisnbsp;opinion that the Permian plant is a Palaeozoic member of thenbsp;Psilotales, a conclusion based on insufficient evidence. We havenbsp;no definite information with regard to the nature of the organnbsp;borne on the sporophylls. The same author compares the sporophylls of Gomphostrobus with the leaves of Sphenophyllum thoughnbsp;the verticillate disposition of the leaves of the latter genus is a well-defined difference. It would seem, as Zeiller says, that Gomphostrobus is probably allied to Walchia though its position cannot benbsp;precisely determined without further, data.

A recent examination of some specimens from Lower Gondwana rocks in India described by Peistmantel� as Voltzia revealed thenbsp;occurrence of some small distally forked leaves very similar to thenbsp;sporophylls of the European Gomphostrobus.

1 nbsp;nbsp;nbsp;Zeiller (922) ^ p, iqI, PI. xv. fig. 12; (OG) B. p. 213, PI. L. figs. 6�8.

Brongniart^ instituted this genus for foliage-shoots from the Bunter sandstones of the Vosges, the name being chosen in commemoration of Voltz; he compared the branches with those ofnbsp;Araucaria excelsa but added that the cone-scales bore three ovules.nbsp;The leaves show considerable variation even on the same axis,nbsp;a feature shared with Walchia and Ullmannia: the megastrobilinbsp;are characterised by a lax disposition and the fan-like, lobed ornbsp;crenulate form of the megasporophylls, which in the best preservednbsp;type, F. Lieheana (fig. 748, C�F), bear three ovate seeds on thenbsp;upper surface. Many authors compare the Triassic genus withnbsp;members of the Taxodineae, e.g., Cryptomeria, and the Araucari-neae: wood of the Araucarian type has been referred to Voltzianbsp;though without proof of connexion with the vegetative shoots.nbsp;Gothan^, who favours a Taxodineous alliance, points out thatnbsp;wood associated udth Voltzia has Araucarian pitting on thenbsp;tracheids, though he adds that the occurrence of typical Araucariannbsp;pitting in stems possessing other characters foreign to the recentnbsp;Araucarineae justifies the conclusion that the presence of alternatenbsp;polygonal pits on the tracheids is not necessarily proof of Araucarian affinity. An examination of some carbonised fragmentsnbsp;attached to cone-scales of F. Lieheana in the British Museum fromnbsp;Gera revealed the occurrence of uniseriate pits both separate andnbsp;in contact with one another. It is probable that Voltzia is relatednbsp;to the Araucarineae though in what degree is uncertain. A recentnbsp;view� that Voltzia affords an illustration of a generalised typenbsp;combining Araucarian and Abietineous features is in part basednbsp;on an assumption that the cone-scales are double like those of thenbsp;Abietineae. That the genus is a generalised type is probable,nbsp;but the data are insufficient to warrant any definite statement asnbsp;to which Coniferae are the nearest allies. The range of the genusnbsp;IS difficult to define: if we include the species F. heuperiana, alsonbsp;Heer�s genus Leptostrobus, the geological range extends from thenbsp;Permian to Middle Jurassic floras. The typical species are characteristic of Permian and Lower Triassic rocks. The similarity innbsp;habit of Walchia, some species of Ullmannia, and Voltzia renders

exceedingly difficult the determination of sterile branches. Further, the fact that the specimens of this presumably arborescentnbsp;genus are usually small branch-fragments sets a limit to our Irnow-ledge of the external features of the individual plants.

The examination of numerous specimens from the Bunter sandstone of the Vosges led Schimper and Mougeot to include

Fig. 748. Voltzia. A, B, Voltzia heterophylla (B, medullary oast). C�F, Voltzia Liebeana. (A, B, after Seward; C�F after Geinitz.)

under this name Voltzia brevifoUa and F. rigida of Brongniart^. The Triassic species Voltzia heterophylla (fig. 748, A) is the bestnbsp;known representative of the genus: the vegetative shoots agreenbsp;very closely with those of Araucaria excelsa and differ but littlenbsp;from shoots of Walchia, some forms of Ullmannia, and Pagiophyl-

1 Brongniart (28) p. 446; Schimper and Mougeot (44) A. p. 21, Pis. I., vi.�ix.

lum. Sch�tze^ has given a long list of references to records of V. heterophylla, but in making use of such lists it should be remembered that in the absence of reproductive organs the specific ornbsp;even generic determination of specimens resembling in habitnbsp;Araucaria excelsa is a hopeless task. The heterophylly of Voltzianbsp;heterophylla is a striking feature: long linear obtuse leaves, 2�5nbsp;cm. long, occur in close association with falcate decurrent foliage.nbsp;Small oval strobili characterised by crowded imbricate appendagesnbsp;are figured by Schimper and Mougeot as male cones, but in no. casenbsp;have any sporangia been detected: similar strobili are also figurednbsp;by Leuthardt^ from the Keuper of the Basel district. The mega-strobili are longer and bear cuneate cone-scales, with 3�5 roundednbsp;lobes on the upper surface, arranged in a lax spiral. The marginalnbsp;lobes of the scales are less deeply separated from one anothernbsp;than in F. Liebeana (fig. 748, D�F). We have no satisfactorynbsp;information with regard to the nature or method of attachmentnbsp;of the seeds. Saporta� figures a cone from Soultz-les-Bainsnbsp;showing, as he asserts, the impressions of seeds, but the drawingnbsp;affords no definite evidence as to the relation of cone-scales andnbsp;seeds. Saporta regards the cone-scales as double, each consistingnbsp;of an ovuliferous scale and a bract-scale more or less completelynbsp;fused as in the recent genus Taxodium. The assumption that thenbsp;scales are double rests on a very slender basis, and even in thenbsp;much better preserved specimens of F. Liebeana �there is nothingnbsp;to indicate that the scale was double^.� In a recent paper Missnbsp;Holden� speaks of Voltzia cone-scales as double in terms suggestingnbsp;a well-established fact, though this is by no means the case.

Blanckenhorn� and other authors have described specimens of Voltzia heterophylla showing elongated leaf-cushions which theynbsp;compare with similar raised areas on the Permian casts on whichnbsp;Weiss founded the genus Tylodendron [Schizodendroni�). In thenbsp;latter genus the supposed leaf-cushions are casts of medullarynbsp;rays at the inner edge of the secondary xylem, and an examinationnbsp;of Voltzia specimens in the Strassburg Museum� convinced me that

the same explanation applies to Voltzia heterofhylla. Triassic Voltzia casts were referred by Schleiden to a distinct genus Endo-lefis: examples figured by Schenk^ afford a good illustration ofnbsp;their close resemblance to Schizodendron. The medullary castsnbsp;of the Triassic genus differ from those of Schizodendron in theirnbsp;smaller diameter and jn the absence of periodic swellings: narrownbsp;slits in the elongated areas mark the position of out-going leaf-tracesnbsp;(fig. 748, B). A similar though larger form of cast is figured bynbsp;Miss Holden^ from Coburg and New Brunswick and referred tonbsp;Voltzia coburgensis. Fliche� has described a form from thenbsp;Muschelkalk of France with more slender shoots than in mostnbsp;examples of V. heterofhylla. Specimens from Swiss Triassic bedsnbsp;figured by Heerquot;* and Leuthardt^ as examples of this species arenbsp;too incomplete to be identified with certainty. Feistmantel�snbsp;Indian specimens� referred to V. heterofhylla, which I have recently examined, from Lower Gondwana strata, show a variation innbsp;leaf-form suggestive of the European species, but the determinationnbsp;is open to question. A supposed cone-scale figured by Feist-mantel resembles in outline the lobed scales of V. Liebeana. Somenbsp;very incomplete branches regarded by Feistmantel as pieces ofnbsp;Albertia shoots are probably identical with the impressions assignednbsp;to Voltzia. On a few of the smaller Indian specimens I foundnbsp;leaves 5 mm. long divided into two slightly divergent prongs,nbsp;a feature unknown in Voltzia but suggesting Gomphostrobus ornbsp;small leaves of Dicranophyllum. Some small seeds figured bynbsp;Zeiller'^ from the Karharbari beds of India as probably belongingnbsp;to Voltzia cannot be determined with confidence.

Fliche� gives this name to vegetative shoots from the Bunter of the Vosges characterised by a Walchia-like habit of branching,nbsp;the pinnately disposed lateral branches being given off at aboutnbsp;40�; the leaves are elliptical, short and broad, more or less appressednbsp;and less spreading than in Walchia. It is, however, impossiblenbsp;without the confirmatory evidence of strobili to distinguish

clearly between certain forms of Walchia and Voltzia. A very similar type is figured by Scbiitze^ as Widdringtonites Jceuperianusnbsp;Heer^ from Stuttgart but with no justification for the use of anbsp;generic name implying relationship with Widdringtonia. Thenbsp;fragments of foliage-shoots on which Heer founded this speciesnbsp;are too small and of too common a type to be referred to a genusnbsp;implying any definite position in the Coniferales.

A Permian species�, characteristic of the Zechstein copperbearing beds of Gera and other localities, represented by foliage-shoots (fig. 748, C), well preserved cone-scales, and strobili. The vegetative branches closely resemble those of V. heterophylla andnbsp;Ullmannia selaginoides: there is the same inconstancy in leaf-formnbsp;as in the Bunter species. The strobili are also similar to those ofnbsp;V. heterophylla: the largest example figured by Geinitz is 2-5 cm.nbsp;in diameter and 7 cm. long. The cone-scales (fig. 748, D-�F)nbsp;have five lobes, deeper than in V. heterophylla, and the central lobenbsp;is longer than the others. The occurrence of three seeds is anbsp;characteristic feature; these are ovate, 5 x 3 mm., and have anbsp;narrow marginal wing. From their close association with strobilinbsp;Geinitz identified some spherical seeds, formerly described by himnbsp;as Cyclocarpon eiselianuni, as those of F. Liebeana. This authornbsp;figures an imperfectly preserved strobilus as a male catkin; itnbsp;resembles those of F. heterophylla, but no microspores have beennbsp;discovered. Heer describes some foliage-shoots and detachednbsp;lobed cone-scales from the Permian of Hungary as F. hungarica^:nbsp;the scales are similar to those of F. Liebeana but have slightlynbsp;narrower lobes.

This specific name� was given to fertile shoots characterised by long and lax strobili called by many authors F. coburgensisnbsp;Schauroth�, a designation first applied to a cast resembling Lygi-nodendron {cf. fig, 401, Vol. iii. p. 37) from the Keuper of Coburgnbsp;and having no proved connexion with Voltzia. The strobili and

^ Schiitze (01) PI. x. nbsp;nbsp;nbsp;^ Heer (65) A. fig. 31; (76) A. PI. xxx. figs. 4, 5.

megasporophylls constitute the distinctive features of V. keuper-iana: Schimper figures two strobili approximately 18 cm. long characterised by fan-shaped scales; the lamina has a fairly longnbsp;stalk gradually passing into a broad rounded distal portion withnbsp;a crenulate edge, the sinuses between the numerous crenulationsnbsp;being continued as grooves over the face of the expanded portionnbsp;of the scale. No information is available as to the seeds. Thenbsp;similarity in the general plan of the strobili, apart from the clearlynbsp;marked distinguishing feature of the megasporophylls, points tonbsp;a generic affinity between this species and V. Liebeana and V. hetero-phylla. Schimper states that the strobili of V. Iceuperiana occurnbsp;in groups in contrast to the solitary cones of other types, and innbsp;view of this distinction and the form of the cone-scales he employednbsp;the generic name Glyptolepis for which Heer substituted Glypto-lepidium'^ on the ground of the previous use of Glyptolepis for anbsp;fossil fish. Schimper refers to this species the wood named bynbsp;Goeppert Araucarites heuperianus (= Dadoxylon heuperianum) butnbsp;there is no proof of actual connexion. Schenk^ adopted thenbsp;generic name Voltzia and Potoni� proposed a new term Voltziopsis^nbsp;to be used in a provisional and wide sense for Voltzia keuperiana,nbsp;Cheirolepis EscJieri Heer, Heer�s Leptostrobus^ and Nathorst�snbsp;Swedenborgia^ (fig. 749), including species ranging from the Keupernbsp;to Middle Jurassic strata in contrast to the Lower Triassic andnbsp;Permian range of typical representatives of Voltzia. The speciesnbsp;Cheirolepis Escheri, included by Potoni� in his genus Voltziopsis,nbsp;was founded by Heer on an imperfectly preserved scale from thenbsp;Lower Lias of Switzerland resembling the lobed cone-scales ofnbsp;Cheirolepis Miinsteri Schenk�. The genus Cheirohpis was instituted by Schimper^ as a substitute for Brachyphyllurn for thenbsp;Rhaetic species B. Miinsteri Schenk, the new name being chosennbsp;because of the digitate margin of the cone-scales which are said tonbsp;bear single seeds. My former employment of Schimper�s genericnbsp;term for sterile branches originally named by Phillips Brachy-phyllum setosum^ was hardly justifiable in the absence of sporo-

1 Heer (77) ii. p. 72. nbsp;nbsp;nbsp;^ Schimper and Schenk (90) A. p. 290, fig. 199.

phylls. The species Voltzia recubariensis (Mass.) represented by several vegetative shoots, imperfect cones, cone-scales, and seedsnbsp;in the Muschelkalk beds of Recoaro^ illustrates the impossibilitynbsp;in the case of sterile specimens of drawing any satisfactory linenbsp;between Voltzia and Pagiophyllum. Heer instituted the genusnbsp;Leptostrobus for strobili .from Jurassic strata in Siberia agreeingnbsp;closely in habit and in the form of the megasporophylls with thosenbsp;of Voltzia. The strobili, referred by Heer to three species, do notnbsp;exhibit any well-marked specific differences; the longest example,nbsp;L. crassipes, is 7 cm. in length and 1-5 cm. broad: the scales,nbsp;7�8 mm. broad, are entire at the distal margin or more or lessnbsp;lobed and in some specimens the scales are hardly distinguishablenbsp;from those of V. heterophylla. Heer states that two-winged seedsnbsp;are borne on some of the scales though the evidence is not clear.nbsp;In a later account Heer^ includes in Leptostrobus clusters of longnbsp;linear leaves apparently borne on short shoots and resembling thenbsp;needles of Pityites Solmsi Sew., Schizolepis Braunii, and Jeffrey�snbsp;Prepinus. These leaves, described as L. rigida and L. angusti-folia, though in close association with strobili are not actuallynbsp;connected with them; they differ considerably from the short,nbsp;triangular, imbricate leaAms shown in one of Heer�s figures immediately below the fertile portionnbsp;of an axis of L. crassipes^. Itnbsp;would be difficult to draw anynbsp;generic distinction between Leptostrobus and Voltzia especially F.nbsp;heuperiana.

This genus * is founded on lax oval strobili bearing small cone-scales with long stalks and anbsp;single seed (fig. 749, B). In thenbsp;type-species, S. cryptomerides, thenbsp;cones reach 7 cm. in length and the scales are divided distallynbsp;into 4 or 5 acute digitate lobes. Nathorst compares the strobili

with those of Cryptomeria and Voltzia but mentions the presence of a single seed as an important distinguishing feature. Tonbsp;unite Swedenborgia with Voltzia would be misleading, and therenbsp;is no valid reason for replacing Nathorst�s term by Potonie�snbsp;genus Voltziofsis.

This name was applied to a lax strobilus, 30 cm. long and T3 cm. broad at the base, from Ehaetic beds on the Orange River, Southnbsp;Africa^, which may be allied to Voltzia, though in the absence ofnbsp;seeds its position cannot be determined. The sporophylls consistnbsp;of a lamina with a rounded distal edge and a radially folded surfacenbsp;attached by a short horizontal stalk resembling the seed-bearingnbsp;scales of Voltzia heterophylla, V. cohurgensis, and to some extentnbsp;Heer�s Jurassic Leptostrobiis.

Goeppert^ in his description of Ullmannia refers to the extensive literature on the fossils from the Permian copper mines of Frankenberg on which the genus was founded: the most complete ofnbsp;these earlier accounts is that of Ullmann. In habit similar tonbsp;Walchia, Ullmannia is represented by various forms of foliage-shoots and impressions of buds and cones, but the data are insufficient to settle its position in the Coniferales. Ullmannianbsp;Bronni (fig. 750, D), the type-species, is practically identical innbsp;leaf-form and habit with the Mesozoic genus Pagiophyllum, whilenbsp;the species U. frumentaria (fig. 750, A) agrees closely with suchnbsp;recent Conifers as Araucaria excelsa and A. Bidwilli. The branchesnbsp;bear spirally disposed crowded leaves with a median vein andnbsp;numerous longitudinal striations on the lamina. The associationnbsp;of impressions of foliage-shoots with wood having the Araucariannbsp;type of pitting� affords contributory evidence, though by nonbsp;means proof, of Araucarian affinity. In the absence of any definitenbsp;information as to the structure of the reproductive shoots Ullmannia must be left for the present as a Conifer which cannot benbsp;assigned with certainty to a systematic position. Tuzson* uses

1 Seward (08) B. p. 101, text-fig. 7; PI. v. fig. .3. nbsp;nbsp;nbsp;^ Goeppert (50) p. 185.

� Solms-Laubach (84) PI. iii. fig. 16; Schimper and Schenk (90) A. p. 275, figs. 190, 191.nbsp;nbsp;nbsp;nbsp;4 Tuzson (09) p. 23.

the generic name Ullmannites for wood having the Araucarian type of tracheal pitting which he believes to belong to Ullmannia, butnbsp;under the former genus are included types of wood that are notnbsp;generically identical and afford no evidence of connexion withnbsp;Ullmannia. Ullmannia is essentially a Permian genus especiallynbsp;characteristic of the copper-bearing rocks of Frankenberg innbsp;Hessen and Ilmenau in Thuringia; it is recorded also from Francenbsp;and by Eichwald and Schmalhausen from Kussia, the species

U. hiarmica Eich.^ being represented by both sterile and fertile branches. Imperfectly preserved impressions from the Permiannbsp;of Durham^ have been referred to Ullmannia but no satisfactorynbsp;specimens have been discovered in English beds. The fragmentnbsp;figured by Lindley and Hutton^ as Voltzia Phillifsi may belong tonbsp;an Ullmannia.

The fragments of foliage-shoots on which Goeppert^ founded this species were described by some earlier authors as the Frankenbergnbsp;ears of corn and by Schlotheim as Poacites phalaroides. Bronn, whonbsp;first identified the fossils as Coniferous, named them Cupressitesnbsp;Ullmanni. We know nothing of the structure of the reproductivenbsp;shoots, and the cones referred by Goeppert to this species havenbsp;since been assigned to the genus Strobilites. In habit and leaf-formnbsp;Ullmannia Bronni is indistinguishable from certain Triassic and Jurassic foliage-shoots referred by most authors to Pagiophyllum. Thenbsp;leaves are imbricate (fig. 750, D, E); the lamina is oval or broadlynbsp;linear, elliptical, and characterised by longitudinal striations. Somenbsp;imperfectly preserved leaves examined by Solms-Laubach showednbsp;clear indications of the presence of a midrib with lateral groups ofnbsp;transfusion-tracheids, as in �. selaginoides and U. frumentaria.nbsp;Ullmannia selaginoides (Brongniart).

This and the following species, U. frumentaria, are founded on vegetative shoots from the Permian of Ilmenau (Thuringia) andnbsp;both were described by Brongniart as examples of Fucoides^. Thenbsp;leaves of U. selaginoides are longer than those of U. Bronni, linearnbsp;and almost uniform in breadth, elliptical in section. There is anbsp;single vascular bundle accompanied by wings of reticulate transfusion-tracheids (cf. fig. 750, C) associated with parenchyma�.nbsp;As in the leaves of recent Conifers the transfusion-tissue persistsnbsp;in the apical region of the lamina. There are 1�2 rows of hypo-dermal fibres below the epidermis with sunken stomata and thenbsp;mesophyll consists largely of palisade-cells {cf. fig. 750, B). Thenbsp;stele of the shoot has a large pith with nests of dark cells enclosednbsp;by a cylinder of secondary xylem consisting of tracheids having anbsp;single row of separate circular pits on the radial wall and uniseriatenbsp;medullary rays 1�6 cells deep.

This species, originally named Carpolithes frumentarius^, is the commonest fossil in the Ilmenau mines. The leaves are lanceolate,

1 Goeppert (.50) p. 185, PI. xx. For synonymy, see Solms-Laubach (84); Geinitz (80).nbsp;nbsp;nbsp;nbsp;^ Brongniart (28) A. PI. ix.

acute, decurrent, and more or less falcate and like those of other species characterised by longitudinal striae on the dorsal facenbsp;(fig. 750, A). Well-preserved impressions of this species in thenbsp;Dresden Museum bear a close resemblance to shoots of Araiicarianbsp;Bidwilli and A. brasiliensis. Some specimens show laterallynbsp;attached oval cones, but it is not clear if these are reproductivenbsp;shoots or vegetative buds and nothing is known as to the naturenbsp;of the sporophylls. The seeds, Cardiocarpus triangularis'^, referrednbsp;by Geinitz to this species on the ground of association, cannot benbsp;safely assigned to JJllmannia. In leaf-structure (fig. 750, B, C)nbsp;V. frumentaria closely resembles U. selaginoides but the hypo-dermal fibres form oval strands instead of 1�2 layers. Thenbsp;species has also been recorded from the Permian of France^.

Under this provisional name Solms-Laubach� described the problematical fossils from the Frankenberg copper mines whichnbsp;Gfoeppert believed to be the megastrobili and cone-scales ofnbsp;Ullniannia Bronni. In the absence of any satisfactory evidencenbsp;of connexion with the vegetative shoots on which TJ. Bronni wasnbsp;founded i^ is better to follow Solms-Laubach in the adoption of thenbsp;non-committal name Strobilites. These star-stones (�Sterngrau-pen�) of the miners consist of more or less circular bodies bearingnbsp;some resemblance to the peltate cone-scales of Cupressus; theynbsp;occur either singly or in cone-like groups. No seeds have beennbsp;found attached to the scales nor is there any proof that they werenbsp;borne by a Conifer. The larger scales, 15�25 mm. in diameter,nbsp;are characterised by 8�12 radial ridges and a central depressionnbsp;(? umbo), and to the under surface is attached a cylindrical stalknbsp;Usually in the centre but occasionally excentric. The scales arenbsp;sometimes found in almost spherical clusters and another formnbsp;described by Solms-Laubach consists of a cylindrical cone-likenbsp;�'�gregate 4 cm. x 2 cm. of rather smaller scales without radial ribsnbsp;and characterised by a stout, rounded, peripheral border.

This generic name was given by Schimper to vegetative branches from the Bunter of the Vosges agreeing in habit withnbsp;shoots of some species of AgatJiis but differing in the broadernbsp;insertion of the lamina. For Albertia Endlicher^ substitutednbsp;Haidingera. Schimper and Mougeot^ figure reconstructions ofnbsp;both male and female cones and a single cone-scale bearing anbsp;median seed. Schenk�, who examined the original specimens,nbsp;states that the supposed male cone is a young megastrobilus ofnbsp;Voltzia, and Solms-Laubach^, who also examined the material innbsp;the Strassburg Museum, considers that Schimper�s statement thatnbsp;the seed-scale and cones belong to Albertia is � altogether arbitrarynbsp;and unsupported.� There would seem to be no reason for connecting the cones figured by Schimper and Mougeot with the shootsnbsp;referred by those authors to four species of Albertia^. Albertia lati-folia is founded on branches bearing fairly large (2-5 x 1 cm.),nbsp;obovate, slightly decurrent leaves with numerous longitudinal stria-tions. The branches described as A. elliptica are not distinguishablenbsp;by any clearly marked feature from A. latifolia. Albertia Brauniinbsp;has larger obovate leaves and A. speciosa has broadly linear leavesnbsp;reaching a length of 4-5 cm. There is a very close resemblancenbsp;between the shoots from the Bunter beds and those of some formsnbsp;of TJllmannia especially U.fnmentaria (Schlot.)�, and it is doubtful whether any useful purpose is served by the retention of thenbsp;designation Albertia: the descriptions of the reproductive shootsnbsp;are misleading and rest on no substantial basis and the sterilenbsp;branches exhibit no characters by which they can be genericallynbsp;separated from Ullmannia. The important point is that there arenbsp;no grounds for regarding the specimens usually referred to Albertianbsp;as Araucarian other than the uncertain and untrustworthy evidencenbsp;afforded by a similarity to Agathis. The fragmentary impressionnbsp;from the Karharbari beds of India assigned by FeistmanteF to

Albertia is too imperfect to be determined; it may be identical with those referred by the same author to Voltzia.

The abundance and wide distribution of wood with Araucarian features in Palaeozoic rocks though, for reasons already stated,nbsp;not admissible as proof of the occurrence of members of thenbsp;Araucarineae, at least shows the great antiquity and predominancenbsp;of the Araucarian type. There can be no reasonable doubt thatnbsp;nauch of the wood described in Chapter xxxiii. as Dadoxylonnbsp;belonged to Araucarian plants, more especially the examplesnbsp;furnished by Mesozoic and Tertiary strata. In considering thenbsp;past history of the family the evidence of the wood must be takennbsp;into account.

The conclusions drawn from a survey of the fossil records are; (i) the type represented by Araucaria is older than that now illustrated by Agathis. In other words Araucaria possesses features,nbsp;especially those associated with the megastrobili, which extendnbsp;farther back without departing far from the existing type than is thenbsp;case with Agathis. (ii) The Araucarineae foreshadowed in the laternbsp;Carboniferous and earlier Permian periods were in all probabilitynbsp;established as a family in Ehaetic times, and in the Jurassic andnbsp;earlier Cretaceous periods the Araucarineae were almost cosmopolitan and represented by numerous forms, (iii) Such evidencenbsp;as is afforded by Tertiary records, though meagre and often incomplete, points to the continued existence of the family in thenbsp;Northern Hemisphere at least in the older Tertiary floras.

CHAPTER XLVl.

The published records of fossil Conifers would seem to justify the conclusion that the Cupressineae were widely distributed andnbsp;represented by a wealth of genera during the latter part of thenbsp;Mesozoic era particularly in the later Jurassic floras, but on closernbsp;inspection of the material a student, having any familiarity withnbsp;the external features of recent genera, cannot fail to recognisenbsp;the wholly inadequate nature of the data on which the systematicnbsp;determinations are based. It is undoubtedly true that in thenbsp;later Jurassic and Lower Cretaceous floras Conifers agreeingnbsp;generally in habit and in the possession of appressed imbricatenbsp;leaves with such genera as Cufressus, Chamaecyfaris, and Thuyanbsp;were among the most characteristic types: some have the leavesnbsp;in decussate pairs with an occasional tendency to a spiral phyllo-taxis while others possess leaves of the same form but spirallynbsp;disposed. Almost all are sterile and when cones are present thenbsp;form and arrangement of the scales often suggest comparison withnbsp;recent types other than the Cupressineae. In the Chapter onnbsp;recent Conifers attention is called to the inconstancy of leaf-arrangement in certain species and to the close resemblancenbsp;between vegetative shoots of plants belonging to different families.nbsp;Fossil coniferous branches referred by authors to the Cupressineaenbsp;afford a striking illustration of the insufficiency of the evidencenbsp;on which sterile impressions have been named. This statement,nbsp;though primarily concerned with Mesozoic records, applies alsonbsp;to many Tertiary species. The records of the rocks clearly shownbsp;that European Tertiary floras contained a considerable number ofnbsp;Cupressineous types that are now confined to other regions, butnbsp;a critical examination of the older fossils leads to the conclusionnbsp;that in very many cases accurate determination of the affinitiesnbsp;of sterile branches, superficially resembling existing members ofnbsp;the Cupressineae, is impossible without additional information.

In the absence of well-preserved cones or anatomical data it is possible that a comparative examination of cuticular membranesnbsp;might furnish useful results. It is, however, only rarely that suchnbsp;information can be obtained and the only safe course to follow is tonbsp;use, with greater freedom than has generally been the practice,nbsp;provisional generic names which do not imply affinities to recentnbsp;genera. Provisional names that have reference only to vegetativenbsp;features should be superseded by designations denoting charactersnbsp;of greater taxonomic significance when the necessary informationnbsp;is available. As a preliminary to the description of a few selectednbsp;types it may be useful to consider the sense in which some genericnbsp;names have been employed and at the same time to statenbsp;whether the retention of certain names or their use in a modifiednbsp;sense is advisable.

Brongniart^ instituted this term for �Branches like those of Thuya; fruit unknown.� The name, in the form Thuyites, hadnbsp;been used a few years previously by Sternberg and it was adoptednbsp;1*7 Unger^ as Thuites for both sterile branches and cones similar tonbsp;those of the recent genus. Thuytes has been widely used fornbsp;Vegetative branches agreeing generally in habit with those ofnbsp;Thuya or Cupressus and some other Cupressineae, but with thenbsp;exception of some Tertiary species the designation has referencenbsp;in nearly all cases to the form and arrangement of the leaves.nbsp;The employment of Thuytes in this wide sense is open to criticismnbsp;on the ground that in accordance with the usual practice specimensnbsp;so named would be considered to be more nearly related to Thuyanbsp;than to any other genus. It is therefore proposed to adopt thenbsp;generic designation Thuites only for such specimens as affordnbsp;evidence of close affinity to the recent genus and to discontinuenbsp;its use for sterile shoots which suggest comparison not only withnbsp;Thuya but also with Cupressus, Libocedrus and other genera.

CUPRESSINOCLADUS. Gen. nov.

Goeppert�s term Cupressites^ is retained for fossils which there is reason for associating with Cupressus and should not be employed

in a wider sense. Bowerbank^ adopted the form Cupressinites for some fossil cones from the London Clay in order to avoid thenbsp;implication of affinity only to Cupressus which is suggested bynbsp;Cupressites. This generic name would be convenient for Cupres-sineous branches had it not been restricted in the first instance tonbsp;cones: to avoid the revival of a term and its employment in a newnbsp;sense it is proposed to adopt the name Cupressinocladus for vegetative shoots agreeing in the habit of branching and in the predominance of a decussate arrangement of appressed leaves with recentnbsp;Cupressineae such as Gupressus, Thuya, Libocedrus and similarnbsp;types. When cones are present which throw any light on genericnbsp;affinity some other term should be adopted. It will, however, benbsp;found in practice that the choice of the most appropriate name isnbsp;exceedingly difficult; and no sharp line can be drawn betweennbsp;certain specimens which conform in part to Cupressinocladus andnbsp;in part to the characters of Brachyphyllum.

The published illustrations of Conifers included by Saporta^ in this genus afford examples of the inconstancy of leaf-arrangementnbsp;in a single type and demonstrate the impossibility of drawing anynbsp;definite distinction between this genus and Thuytes as used by'nbsp;Saporta. With one exception all the specimens from Jurassicnbsp;rocks referred by Saporta to his genus are sterile and in habitnbsp;agree with several recent genera of the Cupressineae particularlynbsp;Gupressus, Chamaecyparis, and Thuya. In Palaeocyparis are included species previously referred to Echinostrobus, Thuytes, andnbsp;Athrotaxites. The branching is in one plane; the leaves are statednbsp;to be usually though not invariably decussate, more or less triangular, appressed, and imbricate, rarely free at the apex of thenbsp;lamina. The supposed cone described in the case of Palaeocyparisnbsp;elegans^, a species from Upper Jurassic beds in France, is only 7nbsp;by 9 mm. and it is not clear whether it is a true cone or a vegetativenbsp;bud or perhaps a male flower. The genus is practically foundednbsp;on vegetative characters only. An objection to the retention ofnbsp;Saporta�s term is that several of the specimens may legitimatelynbsp;be included in a previously established genus Brachyphyllum.

As stated on a previous page it is proposed to limit this name to fossils affording good evidence of close affinity to the recent genusnbsp;Thuya. Among the few examples that appear to fulfil this condition are those described as Thuya occidentalis L. succinea Goepp.nbsp;and Biota orientalis Endl. succinea Goepp. ^ from the Oligocenenbsp;amber deposits on the Baltic coast. SchlechtendaB recordsnbsp;specimens from Oligocene beds at Weimar, consisting of shoots,nbsp;with opposite pairs of leaves, on which a resin-canal is seen belownbsp;the apex, and the remains of male flowers, which he names Thuyanbsp;occidentalis var. thuringica.

This name was given ^ to a piece of shoot of GwpfessMS-like habit from the Eocene beds of Antrim. Baily s figure, whichnbsp;Gardner says is inaccurate, shows the small scale-leaves as spiral,nbsp;but the specimens subsequently described by Gardner� fromnbsp;the same locality leave no doubt as to the decussate arrangementnbsp;of the foliage. Gardner adopts the name Cupressus Pritchardi onnbsp;the assumption that the fossil wood from Lough Neagh describednbsp;by Goeppert� as Pinites Pritchardi belongs to the species whichnbsp;furnished the Antrim specimens; the wood, subsequently referrednbsp;to Cupressinoxylon'^, affords a typical example of that genus asnbsp;is clearly shown by the photographs reproduced in fig. 715 (p. 188),nbsp;but there is no proof of any connexion between it and the branchesnbsp;from the Antrim leaf-beds. The shoots are characterised bynbsp;their slender pinnately arranged branchlets with small decussatenbsp;leaves, and the cones, about 14 mm. long, bear 10 hexagonalnbsp;scales of the Cupressus type very similar to those of Cupressitesnbsp;taxiformis. No seeds were found attached to the cone-scales.nbsp;The abundance of specimens in the Irish beds indicates that thisnbsp;Cupressineous species was a common tree in the forests which

flourished on the western edge of Europe during the period of volcanic activity responsible for the widespread sheets of lava innbsp;the North-East of Ireland and the Western Isles of Scotland.

This species was founded by Unger ^ on sterile and fertile shoots from Eocene beds in the Tyrol. Many of the sterile branches arenbsp;similar in the form of the linear leaves to Taxus, but other leavesnbsp;are appressed to the axis and free only at the apex, resembling onnbsp;a small scale those of Sequoia gigantea. The cones consist ofnbsp;polygonal, peltate, scales probably verticillate and superficiallynbsp;similar to the strobili of Cupressus: they are borne on shoots withnbsp;scale-like leaves. The species is recorded by De la Harpe^ fromnbsp;the Isle of Wight and several specimens are figured by Gardner�nbsp;from the Middle Bagshot beds of Bournemouth. The materialnbsp;from the latter locality consists of sterile shoots with linear Taxus-like leaves from 5 to 15 mm. long and occasionally, on the samenbsp;axis, smaller decurrent leaves, though generally the two forms arenbsp;found on different twigs. The cones occur onlynbsp;in connexion with the shoots bearing small appressed leaves: the specimen shown in fig. 751nbsp;has been re-drawn from one of Gardner�s figurednbsp;specimens*: it is characterised by peltate scalesnbsp;with a central umbo and a wrinkled surface.

The evidence in favour of assigning all the sterile shoots to the same species is not convincing, though a similar combination is metnbsp;with in the recent species Glypttostrobus hetero-phyllus. If we confine our attention to the conesnbsp;they may reasonably be retained in the geiius Cupressites andnbsp;regarded as evidence of the existence in Western Europe in thenbsp;Eocene period of a type closely allied to the genus Cupressus.nbsp;The position of the sterile shoots cannot be determined withoutnbsp;further investigation.

� Gardner (86) p. 26, PI. i. figs. 1�13; PI. v. figs. 13, 14; PI. vii. fig. 8; PI. ix. figs. 22�26, 28�30.

Another Tertiary representative of the recent genus Cupressus is that described from the Oligocene amber beds of East Prussianbsp;as Cupressus sempervirens L. succinea Goepp. and Menge'^. Thisnbsp;species is founded on fragments of sterile shoots with a well preserved male flower showing very clearly the form of the sporophylls.

The following examples of Cupressineous shoots that do not afford satisfactory evidence of relationship to any particularnbsp;recent genus are given in illustration of the desirability of employing such a non-committal generic term as Cupressinocladus.

The Cretaceous specimens from the Atane beds of West Greenland described by Heer^ as Libocedrus cretacea are unaccompanied by any cones and may equally well be compared with species ofnbsp;Thuya: specimens from the Amboy clays, believed to be identicalnbsp;with Heer�s, are assigned to the latter genus by Newberry�.nbsp;Cupressinocladus salicormides (Unger).

A sterile piece of branch figured by Lindley^ from Provence as Thuya is probably identical with Unger�s species recorded bynbsp;Saporta� from the same locality. The type-specimens on whichnbsp;Unger founded the species Thuites salicornoides (fig. 752) are fromnbsp;Eocene beds in Croatia�; they do not bear any mature cones andnbsp;cannot be assigned with confidence to any recent genus. Thenbsp;flattened shoots bear appressed leaves in decussate pairs and thenbsp;decurrent lamina may reach a length of 1'5 cm. By later authorsnbsp;this species, recorded from Styria�, Bohemia�, Switzerland�, thenbsp;Oligocene beds of East Prussia^�, Italy^^, the Miocene of Francenbsp;and from other localities�^�, is spoken of as Libocedrus, but the fewnbsp;examples of cones that have been figured do not exhibit with

� Saporta (65^) p. 42, PI. i. fig. 4. nbsp;nbsp;nbsp;� Unger (47) p. 11, PI. n.

* nbsp;nbsp;nbsp;Ibid. (67h p. 109, PI. V. figs. 1�7, 14; Engelhard! (85) PI. rirr. figs. 27�30.

sufficient clearness morphological features that justify the conclusion that the specimens agree more closely with Libocedrus than with Thuya or Thujopsis. The present discontinuous distribution of Libocedrus (page 126) is favourable to the view thatnbsp;it was formerly much more widely spread, but despite the very

close resemblance between the sterile shoots of the Tertiary Conifer to those of some existing species of the genus it would benbsp;unwise to adopt the designation lAbocedrus or Libocedrites.

Heer^ described fragments from Miocene strata in Greenland as Libocedrus Sabiniana including a supposed cone-scale, whichnbsp;is too indistinct to be determined. To this species Beust refersnbsp;some petrified wood from the same locality. The sterile specimensnbsp;recorded by Gardner^ from the Woolwich beds of Kent as Libocedrus adpressa, though similar to L. salicornoides, are no doubtnbsp;specifically distinct: they are valueless as evidence of the existencenbsp;of Libocedrus. Laurent� also records the species from Aquitaniannbsp;beds in the Puy-de-D6me.

1 nbsp;nbsp;nbsp;Heer (83) p. 58, Pis. lxx., lxxxvi., lxxxvii. ; Beust (85) p. 40, Pis. iii., v.

2 nbsp;nbsp;nbsp;Gardner (86) p. 25, PI. n. figs. 17�20.nbsp;nbsp;nbsp;nbsp;� Laurent (12) p. 69.

Cupressinocladus massiliensis (Saporta). The small twigs described by Saporta from Provence as Thujopsis massiliensis^ and compared by him with Thuya occidentalis are very similar to thosenbsp;named by the same author T. europaea^: in neither case is therenbsp;any justification for the use of the generic name Thujopsis. Heer�nbsp;records T. europaea from Miocene beds in Greenland and in a laternbsp;account adopts the name Biota orientalis on the ground of thenbsp;occurrence of imperfect cones and elongated cone-scales comparablenbsp;with those of Biota {Thuya), but the figured specimens are toonbsp;indistinct to warrant the employment of the generic term Thuitesnbsp;in the more restricted sense advocated above. Goeppert andnbsp;Menge'* refer some fragments from the Baltic amber beds tonbsp;Thujopsis europaea, but no reproductive organs are figured.

Several Jurassic and Lower Cretaceous specimens formerly included in Thuites are now transferred to Brachyphyllum on thenbsp;ground that the choice of one or other of these names has fre-c|uently been determined by characters that are both inconstantnbsp;and of little morphological importance. There are, however,nbsp;several examples of Coniferous shoots from Mesozoic and Tertiarynbsp;strata that are clearly distinguished from such types as Thuitesnbsp;expansa (= Brachyphyllum expansum), in which the verticillatenbsp;arrangement of the leaves is not a well marked or constant feature,nbsp;by the very regular disposition of appressed leaves in decussatenbsp;pairs as in recent species of Libocedrus and some other Cupres-sineae: for this form of shoot the generic name Cupressinocladusnbsp;is now adopted.

A species described as Thuites valdensis^ from a single specimen from Wealden beds on the Sussex coast, characterised by decussatenbsp;appressed leaves with a comparatively long basal portion con-crescent with the axis of the branch and a free short triangular

apex, agreeing closely with some forms of Libocedrus and with the younger branches of Frenelofsis. Though accurate determinationnbsp;of the position of such specimens is impossible, they afford evidencenbsp;of the fairly widespread occurrence of Conifers in Mesozoic andnbsp;Tertiary strata exhibiting a striking resemblance in habit tonbsp;recent Cupressineous genera.

Vegetative branches from Miocene beds in Spitzbergen and Greenland referred by Heer^ to Thuites Ehrenswaerdi and T.nbsp;Meriana respectively afford examples of specimens which wouldnbsp;be more appropriately included in the genus Cupressinocladus.nbsp;Similarly the fragments described by Goeppert and Menge asnbsp;Thuya Mengeana Goepp.^ afford no convincing evidence of genericnbsp;identity with the recent genus.

The slender sterile branches from the Lower Cretaceous beds on the West of Greenland described as Juniperus hypnoides^ afford nonbsp;substantial evidence of relationship to Juniperus rather than tonbsp;some other member of the Cupressineae or Callitrineae. Specimensnbsp;from the same locality which may be specifically identical withnbsp;Juniperus hypnoides are described by Heer as J. macilenta*. Thenbsp;leaves are very small and occur on the slender axes in oppositenbsp;pairs. Newberry records J. macilenta^ from the Amboy claysnbsp;and states that cone-scales of Dammara are associated with thenbsp;foliage-shoots, though Hollick in a note to the description saysnbsp;that he was unable to find any such scales with the vegetativenbsp;branches. Hollick and Jeffrey� figure specimens from the lignitenbsp;beds at Kreischerville as J. hypnoides and believed them to benbsp;identical with those described by Newberry as J. macilenta; theynbsp;also speak of the association of Dammara [Agathis] scales.

The Tertiary Greenland species J. tertiarius and J. gracilis'^, founded by Heer on sterile branches, are equally unsatisfactory as

i Ibid. PI. XXXV. figs. 10, 11. nbsp;nbsp;nbsp;� Newberry and Hollick (95) PI. x. fig. 7.

records of Conifers closely allied to Juniperus, and the same remark applies to Juniperites eocenica described by Ettingsbausen^ fromnbsp;Haring in the Tyrol. A single male flower figured by Goeppertnbsp;and Menge^ from the Baltic amber as Juniperus Hartmannianusnbsp;may be correctly referred to that genus though other recent generanbsp;are not excluded.

Proposed in the first instance by Schimper� for linger s Athro-taxites lycopodioides, this term was adopted for several sterile shoots such as those named by Brongniart Thuytes expansus,nbsp;characterised by the possession of decussate leaves like those ofnbsp;Thuya and Gupressus with others agreeing more closely withnbsp;Brachyphyllum. As the name has reference to the spinous naturenbsp;of the cone-scales, and as it is now agreed that Unger�s earlier namenbsp;Athrotaxites may be appropriately employed, Echinostrobus is discarded.

This generic name was given by Saporta^ to an Upper Jurassic fertile shoot with whorled leaves of the Thmtes form bearing anbsp;single cone compared with those of Libocedrus. The impressionnbsp;conveyed by Saporta�s figures is that the preservation of the conenbsp;is too imperfect to warrant the institution of a new generic term.

This name applied to specimens from the Solenhofen slates� has reference to the elbow-like insertion of lateral branches: thenbsp;foliage is like that of Brachyphylluni, and the cones, which arenbsp;imperfectly preserved, are compared with those of Thuya. As innbsp;the case of Saporta�s Phyllostrohus the cones are too obscure tonbsp;admit of any satisfactory description.

Athrotaxites Unger.

This name was proposed by Unger� for a branched cone-bearing shoot from Solenhofen agreeing in vegetative characters with

specimens previously figured by Sternberg as Caulerfiies. Tbe specimen in the Munich Museum on which the type-species Athro-taxites lycopodioides was founded was examined and re-figured bynbsp;Schimper^ and by Saporta (fig. 753, C)^; the former authornbsp;substituted for Aihrotaxites a new genus Echinostrobus in order tonbsp;avoid the implication of relationship with Athrotaxis which henbsp;was not prepared to accept, and without adequate reason alterednbsp;Unger�s specific name lycopodioides to Sternhergii. Saporta, whonbsp;believed Unger�s type to be intermediate between Athrotaxisnbsp;and Cryptomeria as regards the features of the cones, retainednbsp;Schimper�s designation Echinostrobus Sternbergii. Unger� in subsequent accounts of Solenhofen plants extended the application ofnbsp;Athrotaxites to sterile shoots, and this course was also followed bynbsp;Thiselton-Dyer^ who expressed agreement with Unger as regardsnbsp;the resemblance of the fossil cones to those of the recent genus.nbsp;In accordance with the principle advocated on a previous page itnbsp;is suggested that Unger�s generic name should be retained onlynbsp;for specimens which afford evidence, other than mere resemblancenbsp;of foliage-shoots, of affinity to the recent genus Athrotaxis: thisnbsp;use of Unger�s term has recently been adopted by Halle�. Mostnbsp;of the sterile specimens referred to Athrotaxites by Unger and othernbsp;authors should be transferred to Brachyphyllurn.

Fontaine� instituted this name for some fertile Coniferous shoots from the Potomac formation similar in habit to species included by Saporta in his genus Palaeocyparis and to the genusnbsp;Thu.tes but bearing cones different from those of Athrotaxitesnbsp;lycopodioides. Berry ^ subsequently reduced the number of Fontaine�s species and pointed out that some of his specimens arenbsp;indistinguishable from shoots included in Sphenolepidium, a conclusion to which I had been led� by a comparison of Wealdennbsp;specimens of Sphenolepidium Kurrianum with Fontaine�s figures.nbsp;Some of the Potomac examples agree in vegetative characters withnbsp;the genus Brachyphyllum.. In the absence of any substantial

* Fontaine (89) B. p. 239. nbsp;nbsp;nbsp;� Berry (lU).nbsp;nbsp;nbsp;nbsp;� Seward (95) A. p. 202.

grounds for assuming any direct relationship between the Potomac Conifers and Aihrotaxis, and in view of the fact that the Americannbsp;specimens can be accommodated in previously instituted genera,nbsp;there are no good reasons for retaining Fontaine�s name.

The Tasmanian genus Aihrotaxis (page 150) is one of the existing Conifers of which our knowledge of morphological features is very

Pig. 753. A, B, D, Athrotaxites Ungeri. C, Athrotaxites lycopodioides. (A, B, D, after Halle; A, nat. size. C, after Saporta; nat. size.)

incomplete: its restricted range and the striking resemblance of the vegetative characters to those of many Jurassic species included in Brachyphyllum suggest antiquity, but palaeobotanicalnbsp;records have furnished very little evidence of value in support ofnbsp;this conclusion. It is however probable that like several othernbsp;plants now confined to the southern hemisphere Aihrotaxis is a

Reference has already been made to the vegetative characters of this Upper Jurassic species from Solenhofen^. In the crowdednbsp;imbricate leaves and in the blunt stiff branches it agrees verynbsp;closely with Athrotaxis cupressoides: the globular cones (fig. 753,nbsp;C), though incompletely preserved, also exhibit in the comparativelynbsp;small number of cone-scales and their thick spinous distal ends anbsp;distinct similarity to those of the recent genusnbsp;of the seeds.

This species^, founded on fertile specimens from the San Martin flora of Patagonia, probably of Upper Jurassic or Wealden age,nbsp;is practically identical with Unger�s type: the branches are morenbsp;slender and the globular cones, 10�15 mm. in diameter, appear tonbsp;be of the same type. The cone-scales have a cuneate base and anbsp;thick spathulate distal end prolonged into a short pointed apexnbsp;(fig. 753, A, B); the thickening of the scales close behind the apexnbsp;recalls the form characteristic of recent cones (cf. fig. 684, N, p. 116).nbsp;As Halle says, the vegetative features (fig. 753, D) of this and thenbsp;preceding type are those of Brachyphyllum.

In his memoir on British Eocene Gymnosperms Gardner� referred to Athrotaxis some of the foliage-shoots and cones fromnbsp;Bovey Tracey in Devonshire which had previously been includednbsp;in Sequoia Couttsiae Heer: the reasons for the change of genus arenbsp;by no means adequate. Mr and Mrs Clement Reid^ in their recentnbsp;investigation of the Bovey Tracey material, which they refer tonbsp;an Upper Oligocene age, made a careful examination of numerousnbsp;Sequoia fragments including a comparison of fossil cuticularnbsp;membranes with the epidermis of both Sequoia and Athrotaxisnbsp;leaves: they were unable to discover any evidence of the presencenbsp;of representatives of the latter genus. Gardner also assigns some

pieces of vegetative organs and in one case a cone to Athrotaxis, but the evidence on which the species Athrotaxis (?) subulata^ isnbsp;founded has little value.

Biongniart^ proposed this name for a Jurassic species, Brachy-'phyllum mamillare, founded on sterile branches characterised by pinnate branching in one plane and spirally disposed appressednbsp;leaves with a thick lamina of triangular, conical, or hexagonal form.nbsp;He afterwards� extended the term to other Jurassic species andnbsp;called attention to the striking resemblance of the fossil shoots tonbsp;those of Athrotaxis. The photograph of Athrotaxis cupressoidesnbsp;shown in fig. 701 (p. 150) affords a very good idea of the habit ofnbsp;Jdrachyphyllum. Specimens in which the pinnate ramification is anbsp;conspicuous feature are more like shoots of Thuya or Cupressus, andnbsp;On the smaller branches the leaves may assume a decussate arrangenbsp;�ient. On older branches the leaves are often hexagonal and morenbsp;Or less convex, while on the branchlets they are more triangular ornbsp;conical and are free at the apex, which in some forms is bent outwards from the axis (fig. 756). In nearly all cases specimensnbsp;referred to Brachyphyllum are sterile and, except in examplesnbsp;where the preservation of the cones is too imperfect to afford anynbsp;evidence of morphological characters, it is suggested that thenbsp;generic name should be reserved for sterile branches and regardednbsp;O'S purely provisional. As Saporta* points out in his account of thenbsp;genus, considerable confusion has been caused by attempts tonbsp;assign species to several positions in the Coniferales on whollynbsp;insufficient evidence. Unger� figured a fertile specimen from thenbsp;Rhaetic rocks of Franconia, which he referred to Brachyphyllumnbsp;speciosum Miinst., and this was re-figured by Schenk� as Palissyanbsp;aperta though as others have pointed out the cones are verynbsp;different from those of Palissya-. as Nathorsf^ says, they have thenbsp;characters of the genus Elatides, and the same is true of some conesnbsp;figured by Saporta� and assigned by him to Brachyphyllum. Thenbsp;name Elatides is reserved for specimens characterised by a certain

1 Gardner (86) p. 43, PI. xi. ^ Brongniart (28) A. p. 109. nbsp;nbsp;nbsp;� Ibid. (49) A. p. 69.

type of cone (fig. 742) and althougli typical examples of the genus bear leaves differing in their greater freedom from the axis fromnbsp;the more concrescent foliage of Brachyphyllum, it is not possiblenbsp;in all cases to draw a definite line between the two forms of shoot.nbsp;Until a few years ago nothing was known as to the anatomicalnbsp;features of Brachyphyllum but the researches of Hollick and Jeffrey^nbsp;have partially made good this deficiency: these authors investigatednbsp;the structure of Brachyphyllum macrocarpum Newb. {= B. crassum)nbsp;(fig. 758, G) from the Cretaceous beds of Kreischerville and demonstrated a close resemblance in some characters to recent Araucarias.nbsp;They fully recognise that it would be unsafe to assume the presencenbsp;of similar anatomical features in other species, though it is reasonable to expect the occurrence of such characters in many speciesnbsp;not yet found as petrifications. Other examples of shoots ofnbsp;the Brachyphyllum habit furnishing information with regard tonbsp;anatomy are B. eat.hiense Sew.^ and Banc, and a Japanese speciesnbsp;originally described as Yezonia vulgaris by Drs Stopes and Fujii^nbsp;and since transferred to Brdchyphyllum. As the great majority ofnbsp;specimens referred to Brachyphyllum give no anatomical information the generic name is usually applied to fossils exhibitingnbsp;only external features; it is a form-genus. The introduction ofnbsp;anatomical characters, based on the examination of a very smallnbsp;number of examples, into a general definition might seriouslynbsp;mislead students with regard to the affinities of species knownnbsp;only as impressions. On the other hand as some species of Brachyphyllum exhibit anatomical features of diagnostic value thenbsp;definition of the genus may be extended, in certain cases only, tonbsp;include the information furnished by such examples as those described on pages 322�328.

The inclusion of some species, e.g. the well-known type usually referred to Thmtes, T. expansus, in Brachyphyllum is a changenbsp;which may be regarded as retrograde, but an examination ofnbsp;specimens of that type shows the impossibility of recognising anynbsp;constant verticillate disposition of the leaves such as would justifynbsp;the adoption of Cupressinocladus or some other name implying

affinity to recent Cupressineae. By slightly extending the use of Br achyphyllum we avoid the danger of giving a false impression ofnbsp;affinity and lighten the task of dealing with material which is ofnbsp;secondary botanical importance.

In transferring this widely spread Jurassic species, founded by Sternberg on a specimen from the Stonesfield (fig. 754), Oxford-

�hire, as Thuites expansus^, from Thuites to Brachyphyllum the application of the latter name is extended to include Coniferousnbsp;shoots in which the decussate arrangement of the leaves is morenbsp;apparent than in the majority of species usually referred to

^ Sternberg (23) A. p. 38, PI. xxxvin. figs. 1, 2. For synonymy see. Seward (04) P. p. 142.

Brachyphyllum. Schimper^ included Sternberg�s species in Echino-strohus and Saporta^ adopted the designation Palaeocyparis. The small amount of evidence with regard to the structure of the conesnbsp;does not afford an adequate reason for retaining the generic namenbsp;Thuites.

Specimens from Jurassic rocks in India described by Feist-mantel� (fig. 755) as Echinostrobus expansus, superficially at least very similar to the European Thuites expansus, have recently beennbsp;examined by Miss Holden^. The epidermal cells of the smallnbsp;decussate leaves are irregular in shape; the stomata are scatterednbsp;but there is an astomatic area down the centre of the lamina.nbsp;The stomata are sunk and have four accessory cells. Miss Holdennbsp;points out that the epidermal features of this Indian type arenbsp;different from those of Brachyphyllum macrocarpum, B. Miinsterinbsp;and B. affine^ in which rows of stomata alternate with strands ofnbsp;sclerenchyma: this diflierence is legitimately used as an argumentnbsp;in favour of retaining the generic name Thuites rather than employing Brachyphyllum. It is, however, as a rule impossible tonbsp;obtain any information with regard to the cuticular features, andnbsp;from the external characters of impressions of foliage-shoots wenbsp;cannot draw any satisfactory line between specimens referrednbsp;to Brachyphyllum and Thuites. Miss Holden�s work affords annbsp;illustration of the possibility of employing epidermal features asnbsp;a means of separating shoots which in habit appear to belong tonbsp;one generic type. So far as I know we have no data with regardnbsp;to the epidermal structure of the European Thuites expansus andnbsp;we cannot therefore say whether the Indian species are identical ornbsp;not with those included in the same species from other regions.

In habit Brachyphyllum expansum agrees with B. mamillare and other types as also with recent species of Thuya and Cupressus: innbsp;some examples the branchlets are crowded and in others thenbsp;ramification is much more open; the small appressed leaves arenbsp;broadly triangular or longer and relatively narrower than in such

* Holden, R. (15^) p. 221, PI. xi. figs. 2, 5, 6. The specimens examined were kindly sent to the Cambridge Botany School by the Director of the Indian Geological Survey.

I'lG, 756, Brachyphyllumexpansum�i. (Theoriginal of Teistmantel s Pachyphyllutn heterophyllum. Calcutta Museum, Geol. Surv. India; nat. size.)

cones occur on the foliage-shoots characterised by spirally disposed scales: in a specimen from the Stonesfield slate described iii 1904

each cone-scale has a funnel-like cavity near one edge and the upper side of the cavity is radially ridged^. No seeds have beennbsp;found in connexion with the cones. The male flowers are longernbsp;and narrower and consist of numerous sporophylls attached atnbsp;right-angles and expanded distally into a peltate lamina. Thenbsp;specimen reproduced in fig. 7.56 is the original of FeistnianteTsnbsp;Pachyphyllum heterophyllum^ from Indian Jurassic beds: on thenbsp;stouter axis there are spirally disposed triangular leaf-bases whilenbsp;on the smaller branches the leaf-lamina is preserved and appearsnbsp;to be thick, sub-falcate, and tetragonal. This specimen is in mynbsp;opinion indistinguishable from that shown in fig. 755, whichnbsp;Feistmantel figures as Echinostrobus expansus^ and both agreenbsp;superficially at least with European examples of Brachyphyllumnbsp;expansum.

This specific name* has been applied to specimens from many Jurassic localities and it might well be extended to others regardednbsp;by authors as distinct species. An accurate specific determinationnbsp;of the numerous Brachyphyllum shoots is indeed hopeless withoutnbsp;other characters than those afforded by impressions and casts. Innbsp;habit the species resembles Athrotaxis cupressoides: the branchesnbsp;are given off at a fairly wide angle; the leaves are small, fleshy,nbsp;and more or less triangular with a median dorsal keel and usuallynbsp;spirally disposed. There has been some confusion between thisnbsp;species and Sternberg�s Thuites expansus: the specimen from thenbsp;Yorkshire coast figured by Bindley and Hutton� under the latternbsp;name, now in the Manchester Museum, is undoubtedly identicalnbsp;with Brongniart�s species. There is a considerable difference in thenbsp;degree of freedom of the upper part of the lamina from the axis;nbsp;in some specimens the leaf is almost entirely concrescent with thenbsp;axis and in others the leaves are more open and attached only bynbsp;the basal part of the lamina.

Feistmantel figures several specimens of Brachyphyllum, from . Indian Jurassic localities under different names, many of whichnbsp;appear to be indistinguishable superficially from B. mamillare.

Among these are some of the shoots referred by him i expansus^ and others described as E. rajmahalensis Feist, andnbsp;E. fhombicus^. An examination of some of the figured specimensnbsp;referred by Feistmantel to Pachyphyllum (= Payiophyllum) peri-gyinum (Lind, and Hutt.) leads me to include them at least provisionally in B. mamillare. The generic distinction between thenbsp;form-genera Brachyphyllum and Pagiophylluni is by no meansnbsp;always clearly marked.

Among many European examples of the Brachyphyllum marnil-lare form of Conifer, reference may be made to the illustrations by Saporta of the French Jurassic specimens referred to Brachy-phyllmn Moreauanum Brongn., B. nepos Sap. and a form with

Zeiller^ records specimens of foliage-shoots with cones superficially resembling those of Sequoia from Lower Jurassic beds in Madagascar which he assigns to Brachyphyllum and compares withnbsp;E. nepos.

Erachyphyllum spinosum Sevyard.

A Wealden species^ founded on several well preserved specimens from the coast of Sussex characterised by the possession of short,nbsp;thorn-like, lateral branches clothed with fleshy leaves with anbsp;longitudinally striated lamina of the usual Brachyphyllum type.nbsp;Two or three of these spinous shoots occur at the same level on thenbsp;parent-axis. The stouter branches are covered with spirally disposed polygonal leaf-bases, while on the more slender branches thenbsp;broad and short leaves assume a more or less regular decussatenbsp;disposition. In leaf-form and branching-habit this species agreesnbsp;closely with several other examples of the genus, but the spinousnbsp;shoots are a distinctive feature.

Erachyphyllum obesum Heer.

This species originally described from Lower Cretaceous strata in Portugal� is represented in the Potomac formation by specimensnbsp;referred by Fontaine i to Brachyphyllum crassicaule, and there are

many specimens recorded both from Jurassic and Cretaceous rocks which differ in no important features from Heer�snbsp;type (fig. 757). An examination of branches ofnbsp;the recent species Cupressus Lawsoniana showsnbsp;a considerable difference in the form of ramification depending on the development of numerous or few lateral shoots, and such differencesnbsp;afford an argument against the use of distinctivenbsp;names such as B. obesiforme and others adopted bynbsp;Saporta�^ for Portuguese specimens. Apart fromnbsp;the absence of thornlike branches this species is Fig. 7,57. Brachy-

This generic name was proposed for pieces of wood from the Middle Cretaceous beds in Staten Island originally regarded as thatnbsp;of the plant which bore the foliage-shoots described from the samenbsp;locality by Hollick and Jeffrey as Brachyophyllum macrocarpum,nbsp;but as the result of further study it was recognised that lack ofnbsp;proof of any connexion between wood and shoots necessitated anbsp;new genus^.

The tracheids of the xylem have separate pits usually in a single row, but they are occasionally flattened and very rarelynbsp;there are two alternate rows of polygonal pits (fig. 758, A).nbsp;Normally there are no resiniferous cells in the xylem though thesenbsp;occur in wounded specimens. The medullary rays are said to havenbsp;numerous pits on the radial walls. Jeffrey has described in detailnbsp;the wound-reactions of Brachyoxylon^fig. 758, B represents partnbsp;of a transverse section showing a mass of resiniferous parenchymanbsp;and a row of resin-canals stretching tangentially from the woundednbsp;area. Wood exhibiting the same normal and traumatic features isnbsp;mentioned by Jeffrey from Martha�s Vineyard and the Potomacnbsp;formation. It is pointed out that Brachyoxylon differs from typicalnbsp;Araucarian wood in the frequent occurrence of circular and separatenbsp;bordered pits and in the power of developing traumatic resin-canals.

Fig. 758. A, �, Bruchyoxylon noiabile; A, tracheida of the secondary xylem; B, traumatic resin-eanak. C, D, Araucariopiiys americana; C, traumaticnbsp;resin-canaLs; D, section of stem. E, F, Protodammara speciosa, cone-scalesnbsp;(x7), see page 250. G, Bmchyphyllum crassum. (After Jeffrey.)

In the combination of the Araucarian and the common type of tracheal pitting Bmchyoxylon agrees with some other genera ofnbsp;Mesozoic woods^ e.g. species of Cedroxylon, and in the formation ofnbsp;traumatic resin-canals it resembles Abies and other genera ofnbsp;Abietineae as also Sequoia. Jeffrey�s view is that Brachyoxylon isnbsp;undoubtedly Araucarian though in its wound-reactions it differsnbsp;from the present representatives of the Araucarineae: in thisnbsp;respect he considers the genus to hold the same relation to recentnbsp;Araucarineae as Sequoia holds in respect of its power of developingnbsp;resin-canals in response to injury to other allied genera in whichnbsp;no such reaction occurs. Admitting the Araucarian arrangementnbsp;of pits on some though by no means on all tracheids, the sum ofnbsp;characters hardly warrants the inclusion of Brachyoxylon in thenbsp;Araucarineae; as in several other Mesozoic genera there is in somenbsp;degree a mixture of characters indicative of a generalised type and,nbsp;while Jeffrey sees in this combination evidence of the derivation ofnbsp;Araucarian Conifers from an Abietineous ancestry, I venture tonbsp;regard the spasmodic recurrence of the Araucarian type of pittingnbsp;as a partial persistence of characters inherited from an ancientnbsp;Araucarian stock.

Miss Holden^ has described some wood from Cliffwood, IV^ew Jersey, which she refers to Brachyoxylon, differing from thatnbsp;described by Jeffrey in the presence of fibres in the secondarynbsp;phloem, a feature associated generally with Cupressineae, Taxo-dineae, and the Podocarpineae. In the Cliffwood material thenbsp;medullary rays are said to have smooth walls, a feature in whichnbsp;they differ from those of the Abietineae.

This name was given by Lesquereux^ to a large branched vegetative shoot from the Dakota group, and for specimens,nbsp;believed by Hollick to be identical with Lesquereux�s species, fromnbsp;the Amboy clays Newberry proposed the name macrocarpum^: thisnbsp;specific name was not published and in the Amboy clay monograph^ the designation Brachyphyllum crassum is adopted. The

* Holden, R. (14) p. 171. nbsp;nbsp;nbsp;- Lesquereux (91) p. 32, PI. ii. fig. 5.

impressions of Brachyphyl�uni leaves. In the younger branches the vascular tissue consists of separate bundles and a gap is formed onnbsp;the exit of the single leaf-trace; the trace divides in the outernbsp;cortex into a number of fine strands � which finally become lost innbsp;a continuous band of transfusion-tissue� beneath the palisadenbsp;parenchyma^. No details are given with regard to the pitting ofnbsp;the tracheids or the structure of the medullary rays, but thenbsp;authors state that the phloem showed no indication of the presencenbsp;of any thick-walled fibres. In a later paper Jeffrey^ states that innbsp;older stems of Brachyphyllum crassum the pits are flattened bynbsp;mutual contact though in younger branches this feature is oftennbsp;not distinguishable. A single specimen is described as probablynbsp;a cone of Brachyphyllum^'. this is, however, much smaller thannbsp;any cone previously recorded in connexion with Brachyphyllumnbsp;shoots and the anatomical data do not furnish any proof of itsnbsp;morphological nature.

The species is recorded from the Magothy formation'*, Cliffwood, and from the Karitan formation; Berry also describes a largenbsp;example from the Woodbine formation, Texas, as B. macrocarpumnbsp;var. formosum^. It is practically impossible to distinguish thenbsp;present species so far as external features are concerned from suchnbsp;species as B. ohesum Heer, B. crassicaule Font, and others�.

The type-specimen of this species was originally figured by Miller as an 'imbricated stem� from Upper Jurassic rocks in thenbsp;North of Scotland^: it consists of a branched shoot bearing innbsp;places some broadly triangular imbricate leaves with longitudinalnbsp;ridges on the surface of the lamina. The pith includes somenbsp;scattered thick-walled elements: no information of importance wasnbsp;obtained as to the structure of the vascular tissue of the stele. Thenbsp;short fleshy leaves have a well-protected epidermis succeeded bynbsp;palisade-tissue and groups of hypodermal fibres while the rest ofnbsp;the mesophyll consists of parenchyma with secretory sacs and

1 For figures, see Hollick and Jeffrey (09) B. nbsp;nbsp;nbsp;^ Jeffrey (10^) p. 770.

� Hollick and Jeffrey (09) B. p. 37, PI. ix. figs. 3, 6; PI. xi. fig. 3; PI. xiv. fig. 3.

portions of leaf-traces. The most striking feature is the occurrence of reticulately pitted, isodiametric tracheids, closely resemblingnbsp;those in recent Araucarian leaves and in Brachyphyllum crassum.nbsp;Stomata were found on some of the leaves agreeing in the possessionnbsp;of four accessory cells (fig. 724, A, page 216) with those describednbsp;by Jeffrey! and Thompson^ in B. crassum. Though comparable withnbsp;Araucarian leaves in the structure and distribution of the transfusion-tissue and in the branching leaf-traces, there is hardlynbsp;sufficient evidence to warrant any positive statement with regardnbsp;to the relationship to recent genera of the American and Scottishnbsp;species.

In their account of Upper Cretaceous plants from Hokkaido, Japan, Drs Stopes and Fujii� instituted a new genus Yezonia^ fornbsp;some petrified shoots which they suggested should be placed in anbsp;special family of Gymnosperms. The specimens of foliage-shootsnbsp;they described as Yezoma vulgaris and for a cone, which they consider may belong to the vegetative branches, the generic namenbsp;Yezosfrobus was proposed. The slender foliage-shoots bear appressednbsp;leaves, apparently spirally disposed, agreeing closely with those ofnbsp;recent Cupressineae in their form and relation to the axis, but innbsp;the absence of impressions their surface-features cannot be clearlynbsp;determined. Anatomically the shoots agree very closely withnbsp;l^gt;'achyphyllum crassum: the pith contains groups of sclerous cells,nbsp;the leaf-traces branch repeatedly in the base of the leaf, and transfusion-tissue is abundant in the mesophyll. The secondary xyleinnbsp;shows uniseriatc separate pits on the tracheids, and the medullarynbsp;rays are 1�2 cells in depth. Jeffrey^ drew attention to the strikingnbsp;resemblance between Yezonia and Brachyphyllutn crassum andnbsp;fully justified his substitution of Brachyphyllum for the new genus.nbsp;Dr Stopes�, while agreeing with this conclusion, points out thatnbsp;evidence furnished by fructifications can alone settle the questionnbsp;of generic identity; she states that the supposed cone attributednbsp;Hollick and Jeffrey to Brachyphyllum differs widely fromnbsp;Yezosfrobus which may be the cone of the Japanese species. The

American cone may, however, be a vegetative bud, and nothing is known as to its seed-bearing appendages. Yezostrohus has not beennbsp;proved to have any connexion with the foliage-shoots of Brachy-phyllum. Disregarding the two cones, there can be no doubt as to thenbsp;very close similarity between the American and Japanese shoots.

This generic name, adopted by Endlicher^ from Unger, is employed for fossil species believed to be nearly related to thenbsp;recent genera Taxodium and Glyptostrobus. Reference is made innbsp;the chapter on Coniferous woods to supposed examples of thesenbsp;genera. The separation of the two recent Conifers is based onnbsp;features which cannot be applied to fossil impressions and even innbsp;the case of the existing types Beissner^, following Bentham andnbsp;Hooker, does not accept the Far Eastern species referred bynbsp;Endlicher to Glyptostrobus as representatives of a distinct genus,nbsp;but izrcludes them in Taxodium. Heef describes some fragmentsnbsp;of shoots from the Lower Cretaceous of Greenland as Glyptostrobusnbsp;groenlandicus^, but these are of little or no value as trustworthynbsp;records. Similarly his species G. intermedius from the Patoot beds^nbsp;founded on dimorphic, sterile, shoots affords no substantial evidence of affinity to Taxodium. or Glyptostrobus.

This species was first described by Brongniart� from Tertiary beds in Greece. The branches bear leaves varying in size and form,nbsp;some being appressed and triangular while others are more elongatenbsp;and freer from the axis: the oval or globular cones, with a maximumnbsp;diameter of 15 mm., consist of 18�20 scales agreeing in their roundednbsp;crenulate edges and radially grooved surface with those of Glyptostrobus. Brongniart� states that this species occurs also in Germany,nbsp;Bohemia, and at Oeningen. Heer� figures good examples of vegetative shoots and cones as Glyptostrobus europaeus from Oeningen;nbsp;the scale-leaves are decurrent and the oval cones have semicircular scales with 6�8 grooves (fig. 759). This is described asnbsp;one of the commonest fossils in the Swiss deposits and, as Heer

says, it bears a striking resemblance to the exis mg Glyptostrobus heterophyllus. Unger^ describes wenbsp;preserved specimens from Greece and the speciesnbsp;is recorded, on the evidence of cones as well asnbsp;sterile shoots, from Leoben^ and other localities innbsp;Styria^, also from Miocene beds in Bohemia .

Laurent� figures examples from Aquitanian be s in the Puy-de-D6me and Saporta and Marionnbsp;refer to Glyptostrohus europaeus fragments of sterilenbsp;branches and an imperfect cone from Pliocene bedsnbsp;in the Province of Ain and mention the occurrencenbsp;of the same type in Pliocene strata in the vallejnbsp;of the Arno. Vegetative shoots are recorded fromnbsp;Tertiary beds in Bosnia� and Nathorst� found-thenbsp;species in Arctic Ellesmere Land. The sterile fragments figured by Goeppert and Menge� from the ^

Oligocene beds on the Baltic coast, though possi j c � determined, afford no proof of affinity to thenbsp;strobus. Some very good specimens from Eocene e s anbsp;are described by Gardner�� as examples of thisnbsp;nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;.

cones are immature and do not furnish convincing ev close relationship to the recent genus. The samenbsp;to specimens figured by this author from Bournemou �nbsp;nbsp;nbsp;nbsp;,

from the latter locality, characterised by their long slender br ^ with spirally disposed leaves having long decurrent ases .nbsp;iecting apices, are referred to a distinct species Taxodmm eocaemcumnbsp;and compared with the Eloridan Conifer Taxodium ishc mmnbsp;imbricataria Mett. Gardner points out wdth reason thatnbsp;described by Heer� from Miocene beds in Greenlan an asnbsp;� as Glyptostrobus europaeus and from rocks of the same g

n Heer (68) i. Pis. iii., xlv.; (71) iii. Pis. i., m.; (77) i. Pk. xi., xii., xx^ ., ( Pis. IX., xin.; (83) Pis. lxx., nxxv.

Spitzbergen as G. Ungeri may be fragments of Conifers more closely allied to Sequoia.

The fossils originally referred by Lesquereux^ to Glyptostrobus gracillimus from the Dakota group were afterwards transferred bynbsp;him to Frenelites Reichii, at a later date removed by Newberry tonbsp;Sequoia and finally described by Jeffrey^ as Geinitzia gracillima-.

Neglecting records based on sterile specimens only it is clear that Conifers closely allied to Taxodium and especially to Glyptostrobus heterophyllus of China were abundant in the Tertiary florasnbsp;of Europe.

Numerous specimens of branches and some male and female flowers have been described by Heer� from Miocene beds in Greenland^ Spitzbergen, Grinnell Land, Northern Siberia, and Sachalinnbsp;Island as Taxodium distichum miocenicum. Though in certain casesnbsp;the material is too imperfect to determine with accuracy, some ofnbsp;the fossils bear a striking resemblance to Taxodium distichum bothnbsp;as regards vegetative features and cones. In this species Heernbsp;includes specimens originally named by Sternberg Phyllites dubiusnbsp;and afterwards transferred to Taxodium. The shoots bear distichous, linear leaves, reaching a length of 2 cm. and 2�3 mm.nbsp;broad. The absence of a decurrent base is spoken of as a characternbsp;distinguishing Taxodium from Sequoia Langsdorfii. Specimensnbsp;from Grinnell Land now in the Dublin Museum described by Heernbsp;as Taxodium distichum miocenicum bear leaves contracted at thenbsp;base but not decurrent. A specimen from Grinnell Land said to benbsp;a male inflorescence^ consists of an axis bearing a few oval budsnbsp;3 mm. long which may possibly be groups of microsporophylls, butnbsp;the figures are far from convincing. Nathorst� speaks of thenbsp;occurrence of foliage-shoots, flowers, and seeds in the Tertiarynbsp;Taxodium shales of Spitzbergen. Ettingshausen� figures shootsnbsp;and flowers from Bilin as Taxodium dubium which are in allnbsp;probability closely allied to the recent species. Goeppert and

� Heer (68) i. Pis. ii., xii., XLV.; (71) Pis. in., iv., etc.; (77) Pis. xm., xxv. (78) Pis. VIII., IX.; (83) Pis. Lxx., etc.

which he refers to Taxodtum and at t��nus. Engel-masses of pollen some of which he nbsp;nbsp;nbsp;^ faxodium type as

Heer�s species is also recorded from several floras* and, despite the fragmentary nature o .nbsp;nbsp;nbsp;nbsp;of the

widespread occurrence of a Conifer in Tertiarj u related to the Swamp Cypress of ^orthnbsp;nbsp;nbsp;nbsp;J'a from

side of the continent^; it is represented by deciduous seeds, stumps, and knees�: its occurrence pom s onbsp;of Cypress swamps over a wide area, also to the migration of

Cunninghamiostrobits yuhariensis Stupes and ^^3 amia rocks is founded on a single detached cone from Upper C-tac^^nbsp;at Hokkaido in Japan^ which in size and form

presumably close to the attachment of the ovules, or seeds were found and the open habit of the cone in ica esnbsp;the seeds had been shed at the time of fossihsation; three p

tuberances on a scale at tlie apex of the cone naay represent aborted ovules though the nature of these is problematical. Near the basenbsp;of a cone-scale there is a single transversely elongated vascularnbsp;bundle which subdivides higher in the scale into a series of normallynbsp;orientated vascular strands, and in one scale a much smaller bundle,nbsp;probably an ovular trace, was found immediately above the mainnbsp;strand. The parenchymatous ground-tissue contains a few sclerousnbsp;elements and several resin-canals, the larger ducts forming a seriesnbsp;across the scale, and near the base a single large canal occurs belownbsp;the broad vascular bundle as in Cunninghamia. Groups of trans-fusion-tracheids occur between the vascular strands.

The preservation of the tissues of the cone-axis is not good enough to throw any light on the question of affinity and it is fromnbsp;the morphology of the seedless scales that any conclusions mustnbsp;be drawn. The cone-scales show no indication of a division intonbsp;the two organs characteristic of the Abietineae nor is there anynbsp;evidence of a ligular outgrowth like that of an Araucarian scale.nbsp;The resemblances in both form and anatomical characters to thenbsp;sporophylls of Cunninghamia exhibited, by the fossil cone appearnbsp;to be such as to justify the employment of a generic name implyingnbsp;close relationship.

The employment of this name by many authors for sterile branches {e.g. fig. 805) superficially resembling foliage-shoots ofnbsp;Cunninghamia sinensis suggests an affinity which is not supportednbsp;by any substantial evidence and while in some cases the fossils maynbsp;belong to plants closely allied to the recent genus, there is nonbsp;definite justification for assuming such alliance. The Lowernbsp;Cretaceous species of Cunninghamites and similar forms are therefore relegated to Halle�s genus Elatocladus.

This generic name was applied by Debey and Ettingshausen^ to some obscure impressions from the Cretaceous beds of Aix-la-Chapelle which they described as portions of a plant �incertaenbsp;sedis�; the specimens superficially resemble the pinnae of a fernnbsp;with broadly linear pinnules, but the occurrence of curved linesnbsp;1 Debey and Ettingshausen (55) B. p. 239, PI. vii. figs. 23�27.

right-angles to the long axis of the pinnules (fig. 760) suggested that some at least of the vein-like markings might be the boundaries of small scale-like leaves similar to those of Libocedrus andnbsp;other Conifers. Saporta^, in his reference to the genus in an accountnbsp;of the Sezanne flora, assigns Moriconia to the Cupressineae, anbsp;determination in accordance with the habit of the foliage-shoots,nbsp;though in the absence of reproductive organs it is impossible tonbsp;hx its position more precisely. The characteristic features are thenbsp;pinnate branching, the flattened form of the branches, and thenbsp;geometrically regular decussate short and broad leaves. The genusnbsp;IS recorded only from Lower and Middle Cretaceous rocks.

This, the type-species, is recorded from Cretaceous rocks at ix-la-Chapelle, from the West coast of Greenland and the Atlanticnbsp;coastal plain. Heer^ figured an imperfectly preserved specimennbsp;rom Disco as Pecopteris kudlistensis in whichnbsp;an indication is given of the occurrence of thenbsp;actual leaves, but some years later� he described well preserved examples as Moriconianbsp;(cyclotoxon, and, as the result of an inspectionnbsp;d drawings supplied by Debey, identified themnbsp;''iith the type-species. The same type is recorded from the Amboy clays (fig. 760)^, Statennbsp;sland� and Block Island. As Hollick points

a large impression included by Heer in Fig.760. Moriconia cy-Moriconia should rather be referred to Brachy- ctotoxon. (AfterNew-IPHlum: in the arrangement of the leaves and in e form of the short and blunt lateral foliages' Moriconia agrees closely with some examples of Brachy-�gt; yllum crassum^ in which the leaves appear to be regularly decussate. The leaves of Moriconia are wholly appressed and thenbsp;upper edge of the lamina is rounded and almost truncate; a mediannbsp;lue, possibly due to the presence of a dorsal keel, runs down thenbsp;uuddle of the exposed broad surface of the shoots.

The species Cryptomerites divaricatus, for which Bunbury^ proposed this generic name^ is more probably Araucarian than a type allied to Cryptomeria: the choice of the term was suggested solelynbsp;by vegetative characters and Bunbury recognised that these agreednbsp;with species of Araucaria as well as with Cryptomeria. The designation Cryptomerites should be restricted to fossils which there is goodnbsp;reason for believing to be allied to the recent genus.

Baily^ figured a sterile piece of foliage-shoot from Eocene leaf-beds in County Antrim as Sequoia du Noyeri which Gardner� subsequently stated to be identical with specimens obtained from the same locality bearing cones similar to those of Cryptomeria.nbsp;Gardner described the Irish specimens and others from Mull asnbsp;Cryptomeria Sternbergii (Goepp.), the specific name being adoptednbsp;because he considered some examples figured by Ettingshausen^ asnbsp;Araucarites Sternbergii Goepp. to be identical with the Irish fossils,nbsp;though most of the specimens described by authors, including thenbsp;author of the species, as Araucarites Sternbergii are believed to benbsp;identical with Araucarites Goepperti Sternb. In these circumstancesnbsp;it seems desirable to employ the specific name du Noyeri used bynbsp;Baily. Gardner�s material consists of foliage-shoots agreeing iunbsp;their spirally disposed leaves, 4�7 mm. in length and falcate innbsp;form, with branches of some species of Araucaria, Dacrydium, andnbsp;Cryptomeria japonica. The occurrence of associated cones, iu somenbsp;cases attached to the vegetative shoots, affords fairly good evidencenbsp;in support of comparison with Cryptomeria. The sub-globose cones,nbsp;15�20 mm. in diameter, consist of a comparatively small numbernbsp;of scales attached by a narrow base and gradually widening towardsnbsp;the distal edge which is deeply fringed. The general appearance ofnbsp;the cones, especially those from Glenarm in Antrim, is similar tonbsp;those of Cryptomeria (cf. fig. 684, M) and taking into account thenbsp;characters of the sterile branches the assumption of affinity to thatnbsp;genus appears to be well founded, though actual proof of closenbsp;relationship is lacking. Gardner includes in Cryptomeria Sternbergii'

2 Baily (69) PI. xv. fig. 4. nbsp;nbsp;nbsp;* Gardner (86) p. 85, Pis. x., xx., xxi.

If the Eooene plant i. correctly regarded ns �'�f , f�2n dVj/pfonierio it supplies another striking 1 us ra ion ^nbsp;in the geographicai distribution of Conifers since the early part

ineriopsis^ was proposed for some petri e g �.,.T�etural CretaLus beds in Japan resembling m habrtnbsp;nbsp;nbsp;nbsp;structmal

the presence of the latter tissue is recorded y uzu i Japanese species C. mesozoica^- The medullarj rays a

one-cell deep in the type-species and there are a e ' n ^nbsp;nbsp;nbsp;nbsp;� t 1 V ;ro gt;iA field In C. antiqua the phloem is

canal nbsp;nbsp;nbsp;belownbsp;nbsp;nbsp;nbsp;thenbsp;nbsp;nbsp;nbsp;vascularnbsp;nbsp;nbsp;nbsp;bundlenbsp;nbsp;nbsp;nbsp;andnbsp;nbsp;nbsp;nbsp;twonbsp;nbsp;nbsp;nbsp;lateralnbsp;nbsp;nbsp;nbsp;dncts,

in the Araucarineae. Dr Stopes^ adheres to the view Japanese fossils are closely allied to Cryptomerm^nbsp;nbsp;nbsp;nbsp;�Uows

evidence of affinity to Araucaria: the structure of t e xy no Araucarian features in the pitting of the trac ei s an ,nbsp;accurate determination of'systematic position must depend up

favours comparison with Cryptonieria rather than u it ran ^ Heer (75) jii. PI. ii. figs. 1�i.

I nbsp;nbsp;nbsp;^ Stopes and Pujii (10) p. 52, i. fig.nbsp;nbsp;nbsp;nbsp;� b Slopes (IP)

I nbsp;nbsp;nbsp;Suzuki (10) p. 185.nbsp;nbsp;nbsp;nbsp;� Jeffrey (10 ) p. Pi-

CHAPTER XLVII.

It has already been pointed out that there is good reason for treating the three existing genera Callitris, Widdringtonia, andnbsp;Aciinostrobus as members of a distinct family. The genus Tetra-clinis, as Saxton^ has shown, while agreeing with the Callitrineaenbsp;in certain features, exhibits a closer resemblance in its gametophytenbsp;to the Cupressineae and is regarded as a type connecting the twonbsp;families Cupressineae and Callitrineae. So far as external charactersnbsp;are concerned, and these are the features from which the palaeo-botanist is compelled to draw such conclusions as^he can, Tetraclinisnbsp;falls into line with the Callitrineae. The discontinuous distributionnbsp;of the recent species of these four genera suggests antiquity and anbsp;former more extended range. Palaeobotanical literature containsnbsp;numerous records of Widdringtonia, Callitris, Frenela or Frenelitesnbsp;based in many cases on sterile shoots and sometimes on cones andnbsp;seeds more or less closely resembling those of recent forms. Thenbsp;generic name Frenela has now been discarded in favour of Callitris:nbsp;it was proposed by Miquel in 1826 to avoid confusion betweennbsp;Callitris and Calythrix, the latter being the name of a Myrtaceousnbsp;genus. An inspection of the published figures of supposed fossilnbsp;representatives of the Callitrineae shows that the name Widdringtonia or Widdringtonites has sometimes been applied to fertilenbsp;shoots with cones differing in the number of the valves from thosenbsp;of recent species and more closely resembling the cones of Callitris,nbsp;Tetraclinis, or Aciinostrobus. Moreover the number of valves innbsp;recent cones, though usually constant, is not invariably the samenbsp;and in imperfectly preserved specimens it is often difficult tonbsp;differentiate satisfactorily between the four genera. In the case ofnbsp;many sterile shoots preserved as impressions it is practically impossible to distinguish clearly between those of the Callitrineae and

cones were found. Similarly Widdringtoniies gracilis Sap. and W. creyensis Sap. from the Corallian and Kimeridgian of France^nbsp;respectively are founded solely on sterile shoots. The specimennbsp;figured by Eichwald^ from Jurassic rocks on the southern bordernbsp;of the Caspian sea as Widdringtoniies denticulatus has the habit ofnbsp;an Araucaria and the supposed cone, which may be some foreignnbsp;body not actually attached, affords no evidence of afifinity to thenbsp;Callitrineae. Zeiller� describes a small fragment from Liassic bedsnbsp;in the Commune of Cherveux bearing small rhomboidal decussatenbsp;leaves similar to Widdringtoniies liassinus (Kurr) as figured bynbsp;Salfeld and to W. lceuperia,nus, but the material affords no definitenbsp;indication of relationship to the Callitrineae.

This species, recorded from several Cretaceous localities in the Eastern United States and elsewhere, is in many cases representednbsp;only by slender sterile shoots and its position among the Coniferaenbsp;is by no means clearly established. It was founded by Ettingshausen ^ as Frenelites Reichii on some branched shoots from Cretaceous rocks in Saxony and afterwards described by Heer� fromnbsp;the Patoot beds of West Greenland under the generic namenbsp;Widdringtoniies though without satisfactory evidence in supportnbsp;of relationship to Widdringtonia. This species is one of the commonest Conifers in the Amboy clays of New Jersey, but no conesnbsp;are figured by Newberry� in his monograph except two smallnbsp;examples which it is suggested may be immature microstrobili.nbsp;Velenovsky^ figures sterile branches from the Perucer beds ofnbsp;Bohemia and an ovate cone, 13 mm. long, with four valves, whichnbsp;resembles a small cone of Aciinostrobus and those described bynbsp;Berry as Widdringtonites suhtilis. Some of the twigs bear terminalnbsp;elliptical bodies regarded as male flowers. The leaves of this speciesnbsp;are usually spiral and, with the exception of the apex, closelynbsp;appressed. Callitrites Reichii is also recorded by Krasser� from the

1 Saporta (84) Pis. 201, 202. nbsp;nbsp;nbsp;^ Eichwald (68) p. 43, PL iv. fig. 9.

� Zeiller (11) PI. ii. fig. 6. nbsp;nbsp;nbsp;* Ettingshausen (67) p. 246, PI. i. fig. 10.

Bowerbank referred several pyritised' cones from the London Clay of the Island of Sheppey to Cupres-sites and some of them he compared withnbsp;species of CaUitris. Gardner^, as the resultnbsp;of an examination of Bowerbank�s type-specimens, reduced the number of speciesnbsp;and adopted the name CaUitris. The specimens, in the British Museum, assigned tonbsp;Callitrites curta are conical cones composednbsp;of 4, 5, or rarely 6 thick and woody valvesnbsp;which are sometimes unequal in size: thenbsp;largest has a diameter of 2 cm. Fig. 761,

A shows a cone of five valves and B, C, are two views of a section of a cone consistingnbsp;of four valves. A similar type describednbsp;by Gardner as CaUitris Ettingshauseni^,nbsp;also from Sheppey, is represented by globular cones 12�15 mm. in diameter andnbsp;composed of 6�8 scales. These two species are probably correctlynbsp;referred to the Callitrineae though the pyritised cones are the onlynbsp;portions of the plant preserved in the Sheppey clay. Gardner statesnbsp;that Ettingshausen� is incorrect in recording Callitrites curta fromnbsp;the Isle of Wight.

This species, first described by Brongniart^ as Equisetum hrachyodon from the Paris Basin, is recorded from many Europeannbsp;localities, in some cases represented only by sterile shoots butnbsp;frequently also by cones and small winged seeds. Some well preserved specimens are figured by Unger� from the Tyrol (fig. 762,nbsp;A, A') under the name Thuyites callitrina characterised by regularlynbsp;whorled leaves, apparently four at each irode, with a relativelynbsp;long and narrow appressed lamina and a small free apex and bynbsp;valvate cones. Unger� subsequently described good specimens as

* Gardner (86) p. 21, PI. ix. figs. 7, 21. nbsp;nbsp;nbsp;^ Ibid. PI. ix. figs. 1�6.

^ Unger (47) p. 22, Pis. vi., vii. nbsp;nbsp;nbsp;� Ibid. (67) p. 42, PI. i. figs. 1, 2.

Gallitris BrongniaHi from Miocene beds in Euboea, but Saporta^ considers these impressions to be more closely allied to Widdring-torda and renames them Widdringtonia humensis. Good examplesnbsp;of quadrivalvate cones (fig. 762, B) are figured by Saporta^ fromnbsp;tbe Eocene beds of Aix and Armissan in Provence, showing in somenbsp;cases two outer broader valves and two internal laterally compressed valves. Ettingshausen� states that the species is verynbsp;abundant at Haring in the Tyrol; that author describes some

fio. 762. A, A', B, Gallifrites Brongniarti. 0, CalUirim helvetka. D, CMitritea europaea. (A, A', after Unger; B, after Saporta; C, after Heer; U, atter

Engelhardt and Kinkelin.)

Sterile shoots from Eocene beds in New South Wales as Callitns prisca^ which he compares with C. BrongniarU. Well preservednbsp;shoots are described by Watelet^ from the Paris Basin. Engelhardt� records the species from Oligocene beds in Bohemia but onnbsp;the inadequate evidence of a winged seed; it is recorded also bynbsp;Engelhardt and Kinkelin�' from the Pliocene beds of the Frankfurt

^ Saporta (68) p. 316.

^ Ibid. (62) p. 209, PI. ii. fig. 6; PI. m.

^ Ettingafiausen (55) p. 34, PI. v. figs. 7�35.

� Ibid. (86) p. 95, PI. vm. figs. 3, 4.

* Engelhardt (85) p. 314, PI. VHi. fig. 32.

� Engelhardt and Kinkelin (08) PI. xxni. fig- 5.

district though on slender grounds. It may be, as Masters suggested, that some at least of the Tertiary specimens included in Callitritesnbsp;Brongniarti are more closely allied to the existing genus Tetraclinisnbsp;than to Callitris.

Among other species that may be included in Callitrites are C. hrachyphylla (Sap.) and C. antiqua (Sap.) from Provence^, represented by shoots with spiral, sub-opposite or opposite appressednbsp;scale-leayes and by globular cones with four valves. As Solms-Laubach says^, the cones agree closely with those of Widdringionia,nbsp;though it would be difficult to decide between that genus andnbsp;Tetraclinis.

Some good specimens are figured by Heer� from the Oeningen beds as Widdringionia helvetica, now transferred to Callitrites, consisting of branched filiform foliage-shoots with small appressednbsp;leaves and cones with four valves (fig. 762, C).

Fragments of branches with small appressed leaves in opposite pairs from the Oligocene amber beds of the Baltic coast are described by Groeppert and Menge'* as three species of Widdringtonites,nbsp;and in one case, W. legitimus, the species is founded on a conenbsp;6 mm. long and 2-5 mm. broad which is not above suspicion as anbsp;record of a Callitrineous strobilus. Frenela europaea and F. Ewal-dana described by Ludwig � from Tertiary beds near Frankfurt arenbsp;founded on unconvincing specimens. Engelhardt and Einkelin�nbsp;describe pyramidal cones with 5�6 valves 1�1-5 cm. long (fig.nbsp;762, D), which they refer to Frenelites europaeus, from the Uppernbsp;Pliocene beds of the Lower Main valley. Many other similarnbsp;instances might be quoted, but on the other hand there is amplenbsp;evidence of the presence in the earlier Tertiary floras in Europe ofnbsp;Conifers agreeing both in vegetative and reproductive shoots withnbsp;existing species now confined to Africa and Australia.

Schenk^ instituted this generic name for specimens originally described by Ettingshausen� from Wealden beds in Silesia as

1 nbsp;nbsp;nbsp;Saporta (62) PI. Ii. fig. 7; (62^) PI. m. fig. 3; (652) pp j_ gg 4. (73) pg jj gg p

2 nbsp;nbsp;nbsp;Solms-Laubach (91) A. p. 60.nbsp;nbsp;nbsp;nbsp;2 Heer (55) A. p. 48, PI. xvi. figs. 2�18.

�* Goeppert and Menge (83) A. p. 39. nbsp;nbsp;nbsp;* Ludwig (59) A. pp. 69, 136.

� Engelhardt and Kinkelin (08) PI. xxm. fig. 5. nbsp;nbsp;nbsp;� Schenk (71) p. 13, PI. i.

^^uites Iloheneggeri on the ground that the external features of the vegetative shoots indicate an affinity to the recent genusnbsp;�^^'ewela (= Callilris) rather than to Thuya or Cupressus. Thenbsp;^sseinblance to Callilris was recognised by Ettingshausen. Thenbsp;^^st striking features of Frenelopsts are the comparatively longnbsp;lEternodes of the jointed stems and branches (fig. 763, A), thenbsp;Occurrence of appressed leaves in opposite pairs or four in a verticil,nbsp;ooncrescent with the whole internodal surface and projectingnbsp;�%htly above each nodal fine as small broadly triangular scales,nbsp;the presence of longitudinal lines of small dots on the internodalnbsp;legions due to rows of stomata characterised by 4�5 accessorynbsp;Ochs surrounding the depressed guard-cells^�(fig. 763, D, E). Thenbsp;��ialler branches closely resemble those of species of Cupressineaenbsp;763, C) or Callitrineae in leaf-form and branching, hut oldernbsp;ranches from which the leaves have partially or wholly disappearednbsp;often differ considerably from the younger foliage-shoots and bynbsp;CQiselves afford little or no indication of their true nature,nbsp;rther details are given in the description of representative species.nbsp;^fenelopsis is characteristic of Wealden or higher horizons innbsp;e Lower Cretaceous series; it occurs in Silesia, Bohemia, Portugalnbsp;^od the South of France and in some North American localities,nbsp;ps-rticularly in the Potomac formation. Heer^ records the speciesnbsp;ciL Lower Cretaceous rocks in West Greenland but some of thenbsp;^I�lginal specimens which I had an opportunity of examining in thenbsp;ockholm Museum afforded no satisfactory evidence of theirnbsp;^^feiaatio position. Though assigned by Heer to the Gnetales,nbsp;^neiopsis is usually regarded as a Conifer agreeing with Callilrisnbsp;^ore closely than with any other existing genus. In their descrip-^ of some fossil shoots referred by Newberry� to Frenelovsisnbsp;cihs Hollick and Jeffrey^, who institute a new genus Raritanianbsp;this species, state that they have reason to believe that somenbsp;ofnbsp;nbsp;nbsp;nbsp;specimens correctly assigned to Fremlopsis are examples

such information as we have with regard to the habit of the ^oStative shoots and the structure of the stomata would seem to

^ Heer (75) Li. p. 73, pj, xviii. figs. 5�8; (82) i. p. 7,.PI. 11. figs. 1�3.

be in favour of including this Lower Cretaceous genus among the Coniferales and regarding it as probably allied to the Callitrineae.nbsp;But the data are insufficient to form the basis of any definite statement as to the position of the genus.

Though the specimens referred to these two species may be correctly separated their close agreement in habit points to anbsp;single type so far at least as concerns the characters as a whole.nbsp;The specimen represented in fig. 763, A, B, originally described bynbsp;Zeiller from the province of Gard, illustrates the method of branching and the form of the leaves borne in whorls of four. Schenk, innbsp;his account of Lower Cretaceous material of F. Hoheneggeri fromnbsp;Wernsdorf in the Carpathians, states that the leaves are in decussatenbsp;pairs or sometimes in verticils of four. The epidermal cells havenbsp;straight walls and a thick cuticle; the stomata form longitudinalnbsp;rows on the internodes and are characterised by the presence ofnbsp;4�5 accessory cells overarching the stomatal depression^.

Specimens described from Bohemia by Velenovsky� as F. hohemica resemble F. Hoheneggeri both in habit and in the structurenbsp;of the stomata.

This species is represented by numerous well preserved specimens in the Potomac formation^: some of the stems have a diameter of 5 cm. and lateral branches are given off in whorls of 3�5; therenbsp;are three leaves at each node with broadly triangular apices andnbsp;concrescent decurrent bases as in F. Hoheneggeri (fig. 763, C). Thenbsp;stomata are arranged in longitudinal rows and agree in the possession of a rosette of accessory cells (fig. 763, D) with F. Hoheneggeri:nbsp;several of the epidermal cells are provided with short spinous processes^. This species is represented by specimens showing clearlynbsp;the cupressoid habit of the smaller foliage-shoots (fig. 763, C).

1 nbsp;nbsp;nbsp;Heer (81) p. 21, PI. xn. figs. 3�7; Saporta (94) B. pp. 139, 199, 214; Pis.nbsp;XXXVI., xxxvm.

2 nbsp;nbsp;nbsp;Thompson (12�).nbsp;nbsp;nbsp;nbsp;� Velenov.sky (88) figs. 1�3, 10.

are hardly sufficiently well exhibited to throw much light on the nature of the shoots. It is possible that some Wealden branchesnbsp;Irorn English strata described as BecUesia anoniala Sew.� maynbsp;belong to some species of Frenelopsis, but the absence of any leavesnbsp;nr nodal marks precludes their inclusion in this genus.

Further light will no doubt be thrown on the nature of Frene-lopsis when the results of the investigations of Hollick and Jeffrey are published: it may be that Heer was correct in his attributionnbsp;of the Portuguese specimens to the Gtnetales though in the habitnbsp;of the branchings especially in F. ramosissima, and in the structurenbsp;of the stomata there is a closer resemblance to recent Callitrineaenbsp;than to any other plants. The genus ranges from Wealden tonbsp;Cenomanian rocks.

In view of the restricted range of the two surviving species of Sequoia and the peculiarities of the genus, to which expression isnbsp;given by the institution of the family-name Sequoiineae^, thenbsp;question of geological antiquity and past distribution assumes anbsp;special interest. Reference has already been made to fossil woodnbsp;presenting features now found in Sequoia, but it is very doubtfulnbsp;� if the anatomical characters of the recent species are sufficientlynbsp;well defined to enable us to discriminate between the wood ofnbsp;^Sequoia and certain other Conifers. Many of the impressions ofnbsp;vegetative shoots and cones described as Sequoia from Jurassicnbsp;and especially Lower Cretaceous strata do not bear a close scrutiny.nbsp;The widely spread species often referred to as Sequoia Reichen-hachii affords no real evidence of affinity to the recent genus andnbsp;the same remark applies to specimens included in Heer�s genusnbsp;Sphenolepidimn and compared by authors with Sequoia. Some ofnbsp;the imperfectly preserved Jurassic cones agreeing superficially withnbsp;those of Sequoia may well belong to species of Sequoiineae.

Though in the majority of instances Jurassic and Cretaceous records do not prove the former presence of Sequoia or a closelynbsp;allied type, some of them afford justification for the belief that thenbsp;American trees are survivals from at least the later floras of thenbsp;Mesozoic era. On the other hand Tertiary strata in many parts ofnbsp;the world supply clear evidence of the wide distribution of Sequoianbsp;or some nearly related Conifers in Europe and elsewhere^. Thenbsp;inference suggested is that the recent species survive in Californianbsp;because of the greater possibilities of migration towards the more

2 nbsp;nbsp;nbsp;Mr E. W. Berry (16) has recently published a sketch-map illustrating the worldwide distribution of fossils referred to Sequoia.

Penhallowr records some petrified wood from Cretaceous strata iQ Alberta which he names Sequoia albertensis and regards as verynbsp;similar to the wood of Sequoia sempervirem, but the evidence innbsp;^vour of a reference to the existing genus is inconclusive. Resin-s are scattered through the wood; the medullary rays have 1�2nbsp;ordered pits in the field, the broadly elliptical pore being generally

Tertiary wood from different localities in North America is ^^quoia on evidence that is far from conclusive. Prof,nbsp;th '^^scribed a particularly well preserved piece of stem fromnbsp;e Miocene auriferous gravels of the Sierra Nevada, near thenbsp;j of Sequoia gigantea, as Sequoia Penhallowi; though I amnbsp;Innbsp;nbsp;nbsp;nbsp;inclined to refer the wood to the Abietineae.

^ e end-walls of the medullary-ray cells, the scarcity of xyleni-^ Rcliyma, and the presence of vertical and horizontal resin-na s, believed to be traumatic, which are certainly suggestive of ^jetineous affinity. Prof. Penhallow� described two species fromnbsp;8-nd^^^nbsp;nbsp;nbsp;nbsp;North-West Territory as Sequoia- Langsdorfii

'5. Burgessii, both of which were previously described by ^ awson but assigned by him to different positions. In the woodnbsp;renbsp;nbsp;nbsp;nbsp;belong to the plant which bore the well-known twigs

^ood containing resin-canals. No resin-canals occur in the . ttention has been called (p, 171, fig. 712) to the abundancenbsp;stone P stems in the Lower Tertiary deposits in the YelloA?-! tnaanbsp;nbsp;nbsp;nbsp;these are named by Mr Knowlton'* Sequoia

and a height of 30 ft. (fig. 764). The details are imperfectly preserved: a few of the tracheids show traces of single and double rows of small bordered pits, but no pits are shown on the walls ofnbsp;the medullary-rav ,cells. Resin-parenchyma is abundant and

Fig. 764. Petrified tree in the Yellowstone National Park (Sequoia magnifica Knowiton). (From a photograph kindly supplied by Prof. Knowlton.)

scattered, as in Cupressinoxylon: it is doubtful whether the wood of Sequoia can be distinguished from that of some other generanbsp;included in the genus Cupressinoxylon. Specimens of wood fromnbsp;the Tertiary coal-field of Aichi-Gifu in the middle region of Hondo,

the main Island of Japan, recently described by Yasui^ a,s. Sequoia ^ondoensis has the following characters; narrow annual rings,nbsp;tracheal pits usually uniseriate though often biseriate and oppositenbsp;on the broader tracheids, rims of Sanio present, medullary-raynbsp;cells with oval bordered pits on the lateral walls but unpitted elsewhere, resin-cells scattered through the spring- and summer-wood,nbsp;resin-canals present which are believed to be traumatic. Thenbsp;occurrence of this wood according to the author of the speciesnbsp;completes in an interesting way the evidence for the existencenbsp;of Sequoia �in Cenozoic times throughout temperate regions ofnbsp;the whole northern hemisphere.� A\hile it is probable that thenbsp;^equoiineae were very widely spread in the Tertiary period it isnbsp;Open to question if the anatomical evidence is sufficiently clear tonbsp;instify the reference of the Japanese wood to Sequoia. The chiefnbsp;reason for the adoption of that generic name is the occurrence ofnbsp;resin-canals similar to the traumatic ducts in the recent species.

The following descriptions include fossils which cannot be referred to Seq^uoiites and others which may reasonably be so named.

This species, originally described as Sequoia prohlematica'^, is founded on a small elliptical cone from Upper Jurassic rocks in thenbsp;Boulogne district; in the form of the scales, which show a ridgenbsp;extending from the edges of the distal surface to a central denbsp;pression in the middle of the cone-scales, the fossil suggests affinitynbsp;f0 the recent genus. ZeillerS also records a cone from Jurassic stratanbsp;fu Madagascar associated with branches of the Brackyphyllum typenbsp;which he says presents all the characters of Sequoia. It must, how-O�^er, be admitted that in both these cases close relationship to

Under the name Sequoia minor Velenovsky* describes specimens from the Lower Cretaceous strata of Bohemia consisting of foliage-shoots with small imbricate linear-lanceolate leaves and a smallnbsp;terminal, spherical, cone the sporophylls of which have rhomboidalnbsp;distal ends and a central umbo; but as in most fossils referred tonbsp;Sequoia the evidence of generic affinity is inadequate.

Yasui (17). nbsp;nbsp;nbsp;2 Fiiche and Zeiller (04).nbsp;nbsp;nbsp;nbsp;� Zeiller (00).

This species, under the name Sequoia giganteoides, has recently been founded by Dr Stopes^ on a small petrified fragment of a very

slender foliage-shoot from the Lower Greensand of Luccomb Chine in the Isle of Wight. The pith contains stone-cells, and a single,nbsp;i Stopes (15) p. 70, PI. II. text-fig. 16.

Undivided leaf-trace enters each decurrent leaf-base; there is a central large resin-canal in the leaves and a considerable development of transfusion-tracheids on each side of the secretory passage.nbsp;Palisade-cells are a conspicuous feature and one or two layers ofnbsp;hypoderm fibres occur next the lower epidermis. The author ofnbsp;the species points out the close resemblance between the fossil andnbsp;the leaves and shoot-axis of Sequoia gigantea.

Heer^ described several specimens of foliage-shoots and cones from the Patoot beds in West Greenland as Sequoia concinna, thenbsp;commonest Conifer in these rich Tower Cretaceous strata. Thenbsp;form of the sparsely branched shoots with their long and slendernbsp;branchlets and straight or slightly curved, decurrent, acuminate,nbsp;leaves (fig. 765) agree closely wdth those knowm as Sphenolepidiumnbsp;^ternbergianum from English and other Wealden rocks, as also withnbsp;the shoots of Sequoiites Couttsiae. The oval cones, 23 x 20 mm..nbsp;Consist of a few scales with 5�6 angled thick distal ends on whichnbsp;there is a median transverse line and a central scar.

This species, represented by sterile shoots and cones, has recently been recorded by Berry^ from Upper Cretaceous beds in Pikenbsp;County, Arkansas.

Carruthers� described a cone and a piece of vegetative shoot from the Gault of Folkestone as Sequoiites Gardneri but neithernbsp;specimen affords any satisfactory evidence of relationship withnbsp;^�^equoia. The shoot is of the Pagiophyllum type, and the cone,nbsp;^'5 X 1-5 cm., consists of spirally disposed scales with four-sidednbsp;rhomboidal distal ends. There is no information with regard tonbsp;the seeds: the data being wholly insufficient to serve as a criterionnbsp;of affinity, the generic name Comtes is substituted for Sequoiites.nbsp;A specimen figured by Lange ^ from the Aachen Sands as Carpolithesnbsp;^mlocinus Schloth. and compared by him to a Sequoia cone agreesnbsp;closely with the English species.

The type-specimen from the Gault of Folkestone^ is an oval cone 6 cm. long and about 2-5 cm. innbsp;diameter; the scales are cuneate and thenbsp;exposed ends transversely elongated andnbsp;hexagonal (fig. 766). It bears a closenbsp;resemblance to Geinitzia gracillima, but innbsp;the absence of any details with regard tonbsp;anatomical features or seeds the non-committal name Conites is employed.

This species^, from the Holma sandstone (Senonian) of Sweden^ is founded on fragments of foliage-shoots covered withnbsp;spirally disposed, appressed, broadly triangular leaves. The specimens are not wellnbsp;enough preserved to show in detail the anatomical features, but Conwentz considersnbsp;such characters as he was able to recognisenbsp;favourable to Nathorst�s adoption of thenbsp;generic name Sequoiites. The species is,

however, not above suspicion as a record of a Conifer closely allied to Sequoia.

Brongniart� instituted this Tertiary species under the generic name Taxites, and Heer^ in his description of foliage-shoots fromnbsp;Miocene beds in Switzerland adopted the designation Sequoia. Innbsp;habit S. Langsdorfii is practically identical with Sequoia sempervirensnbsp;and by many authors it is spoken of as the direct ancestor of thenbsp;recent species. Under this species Schimper� includes a fairly longnbsp;list of synonyms�species referred to Taxites, Taxodium, Cupres-sites, and other genera�which serves to emphasise the fact thatnbsp;impressions of sterile branches with distichous, linear, leaves cannot

2 nbsp;nbsp;nbsp;Conwentz (92) p. 28, Pis. in., iv., viii.nbsp;^ Brongniart (28) A. p. 108.

111 many cases be identified with Sequoia with absolute certainty, ^differences are pointed out by authors in their description of speciesnbsp;between the leaves of such recent forms as Taxus baccata andnbsp;Sequoia sempervirens, but an examination of actual specimens re-I'eals the inadequacy of such fine distinctions as are sometimesnbsp;Quoted. Our knowledge of the cones is confined to externalnbsp;characters and these afford a more substantial basis than thenbsp;foliage-shoots on which to form an opinion with regard to thenbsp;striking similarity between the Tertiary and existing species.nbsp;Sequoiites Langsdo)fii is recorded by Gardnepi from the Eocene bedsnbsp;1^ Mull, but the identification rests on sterile branches bearingnbsp;linear-lanceolate decurrent leaves 6�9 mm. long; the main axisnbsp;if a branched specimen bears scale-like leaves appressed to thenbsp;stem except at the distal end of the lamina and, as in the recentnbsp;species, scale-leaves occur at the base of each lateral shoot. Thenbsp;species is recorded also from Styria^, from Miocene beds in Greece�nbsp;where it is represented by foliage-shoots and cones, from Italy*,nbsp;crmany, and other European localities. A very similar form,nbsp;i^iginally named by Brongniart� Taxites Tournali, is described bynbsp;^a^rdner� from the Middle Bagshot beds of Bournemouth and withnbsp;^ be unites S. Hardti Heer founded on material from Bovey Tracey.

' p. 408) and smaller appressed leaves like those of Sequoia S^gantea, an association also met with in S. sempervirens. S. Tour-^'�li is recorded by Saporta^ who figures branches and cones fromnbsp;icene strata in Provence. Laurent� figures fragmentary speci-which he refers to S. Langsdorjii, from the Aquitanian seriesnbsp;the Puy-de-Dome.

Sequoiites Langsdorjii is very abundant in Arctic Miocene rocks: horst � speaks of numerous branches in Tertiary clays in Elles-Land in a remarkable state of preservation so that they couldnbsp;^ Washed out and isolated like dried specimens in a herbarium,nbsp;peculiarity of the Ellesmere specimens is the occurrence of very

g Gardner (86) p. 41, PI. x. fig. 1. nbsp;nbsp;nbsp;- Ettingshausen (57) PI. i. fig. 3.

Unger (47) PI. n. figs, 17�03. gaporta (68) refers 8. Lanysdorfii as figured by �ger to S. Tournali.

Sciuinabol (92) p. 26, Pis. xv., xvi. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Brongniart (28) A. p. 108.

Gardner (86) p. 40. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Saporta (652) pp �nbsp;nbsp;nbsp;nbsp;1,

Laurent (12) p. 65. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;laathorst (IP) p. 225.

fine teeth on the edge of the lamina^. Similar teeth are stated by Nathorst to have been seen in one or two examples of Sequoianbsp;senipervirens, and it suggested that the papillae, which are a normalnbsp;feature of the recent species, were more strongly developed in thenbsp;Tertiary type. Heer^ records the species from Miocene beds innbsp;Greenland and states that it is one of the commonest Conifers innbsp;Disco Island, from the Mackenzie Eiver, Alaska, Spitzbergen, andnbsp;Sachalin Island. The fragmentsnbsp;reproduced in fig. 767 were collected in Disco Island and are nownbsp;in the Dublin Museum with othernbsp;fossils described by Heer; the long |nbsp;linear leaves. A, are decurrent andnbsp;in some cases the lamina showsnbsp;fine transverse striations; thenbsp;smaller leaves shown in fig. 767, B,nbsp;are referred by Heer to a distinctnbsp;species S. brevifolia, but there is nonbsp;important difference between thenbsp;two forms. Palibin� figures sterilenbsp;shoots from the Sichota-Alin mountains. Penhallow* records the species from British Columbia and

specimens in the .Dublin Museum other localities, and to the same described by Heer.)

from the Queen Charlotte Islands though without any real evidence of connexion. Twigs and cones are described by Schmalhausennbsp;from Tertiary strata in the New Siberian Islands�, and the speciesnbsp;is said to be one of the most abundant and widely distributednbsp;types in the Yellowstone National Park�. Kemains of more thannbsp;one species of Sequoia are recorded from Plorissant, Colorado,nbsp;which has recently been described as a Miocene Pompeii; the sedi-

2 nbsp;nbsp;nbsp;Heer (68) Pis. il., xx., LV., etc.; (71) Pis. XL., XLiii., etc.; (75) ill. PI. il.;nbsp;(77) i. Pis. XII., XIII., XXV.; (78) v. Pi. i.; (82) i. PI. Liii.

ments of an ancient lake mixed with volcanic ash contain many plant and insect remains and Prof. Cockerell�s careful investigationsnbsp;have led to the discovery of several new types^. Staub^, whonbsp;records the species from Aquitanian beds in Hungary, gives a listnbsp;of references to other authors.

Making allowances for doubtful identifications based on sterile branches there remain a sufficient number of authentic records tonbsp;demonstrate the wide range of this species and allied forms innbsp;Europe and the Arctic regions during the Eocene and Miocenenbsp;periods. S. Langsdorfii is said to occur in beds of Lower Pliocenenbsp;3'ge in France� and a minute cone, only 2 by 1'9 mm. has recentlynbsp;been described by Mr and Mrs Reid from Pliocene deposits innbsp;Holland 1. The Dutch specimen is referred to Sequoia with somenbsp;hesitation and it is suggested it may be an immature cone of annbsp;Undescribed species, which possibly marks the last appearance ofnbsp;the genus in Europe.

This species was founded by Heer� as Sequoia Couttsiae on uiaterial from Oligocene beds which form a basin-shaped depression in the granitic rocks of Dartmoor in Devonshire. The materialnbsp;consists of foliage-shoots (fig. 768, A, B), similar in habit to those ofnbsp;i'he recent species Sequoia gigantea, and globose or sub-globosenbsp;cones with peltate scales and winged compressed seeds like those ofnbsp;^'^oquoia sempervirens. Several seeds are said to occur on each cone-scale, Beust� examined wood from Bovey Tracey in which henbsp;lound tracheids with separate bordered pits and resiniferous xylem-Parenchyma as in the recent species. Mr and Mrs Clement Reid^nbsp;have recently investigated the Bovey Tracey material and theirnbsp;conclusion is that �Sequoia Couttsiae is a true Sequoia and close tonbsp;^hc living Sequoia sempervirens and S. gigantea.� They give thenbsp;following description of the cones: �Broadly oval and abruptlynbsp;Harrowed into the stalk, or somewhat cordate; at the base are anbsp;few small recurved wedge-shaped barren scales, the lower onesnbsp;having their stalks strongly reflexed, the middle ones with stalks

^ Cockerell (06). (08), (08^), (08*). nbsp;nbsp;nbsp;* Staub (87) B. PI. xix. p. 249.

J Heer (62) p. 1051, Pis. lix.�LXi. nbsp;nbsp;nbsp;* Beust (85) PI. iii. figs. 1�8.

at light-angles to the axis; at the apex is a rosette of a few almost sessile barren scales; the arrangement of the scales is distinctlynbsp;spiral. It is not easy to count the number of the scales^ as none ofnbsp;the cones we found are perfect. There would seem to be 20�24nbsp;fully developed scales, besides a few undeveloped round the- apexnbsp;and base.� The scales vary in shape (fig. 768, C, D) and are cuneatenbsp;or umbrella-shaped, the rugose distal ends have lines radiatingnbsp;from a central umbo; the winged seeds are pendant beneath thenbsp;thick involuted margin and on the upper surface of one scale fivenbsp;seeds were found. Preparations of the cuticular membrane of thenbsp;leaves showed irregularly scattered stomata, each surrounded by

Fig. 768. Seqiioiites Couttsiae. Twigs A, B, and cone-scales C, D, from Bovey Tracey. {Photographs by Mr and Mrs Clement Reid; x 3.)

a ring of four or occasionally five cells. Gardner^ has also described specimens from Bovey Tracey and Hampshire characterised bynbsp;imbricate keeled decurrent leaves with a free, divergent or falcate,nbsp;apex and in older branches by more obtuse appressed leaves. Thenbsp;cones in size and form.resemble those of Sequoia sem-pervirens whilenbsp;the vegetative branches agree with S. gigantea. There are 3�5 seedsnbsp;on each scale. Several examples of Sequoiites Couttsiae are figurednbsp;by Heer^ from Miocene beds in West Greenland and he speaks of thenbsp;species as the commonest Conifer in Disco Island. Gardner pointsnbsp;out that the northern form has larger cones and stouter foliage-shoots than the British type and proposes for it a new specific namenbsp;1 Gardner (86) p. 36, PI. -vi.

^ Heer (68) p. 94, Pis. ni., viii., XLV.; (71) Pis. XL., XLiii., etc.; (83) PI. lxviH-

'S. Whymperi. This more robust form occurs also in Spitzbergen, ^�i the Mackenzie River, and elsewhere. Knowlton^ recordsnbsp;'S. Couttsiae along with other species from beds probably of Miocenenbsp;^ge in the Yellowstone Park; he assigns to this type specimensnbsp;described by Lesquereux from Colorado as Glyptostrobus Ungerinbsp;s-nd others from the Fort Union Group referred by Newberry tonbsp;^lyptostrobus europaeus. The cone-bearing branches figured bynbsp;Lesquereux^ from the Western Territories as Sequoia affinis bearnbsp;^ close resemblance to S. Couttsiae. The latter species is recordednbsp;dy Penhallow� from the Eocene beds on the Deer River in Canadanbsp;(lat. 51� and 54� N.). The same or a closely allied type is recordednbsp;from Miocene beds in Alsace*, and Saporta� describes very goodnbsp;examples of S. Couttsiae from the Eocene beds at Armissan innbsp;Provence. According to Gardner the material referred by Saportanbsp;fo Heer�s species includes at least two other species. Specimensnbsp;described by Schmalhausen� from Eocene beds in South-Westnbsp;Russia as S. Couttsiae, though possibly correctly named, are notnbsp;convincing. Palibin^ records this species from Oligocene beds atnbsp;Molotytchi in the Fatej district, Russia, and discusses the geologicalnbsp;uge of the strata from which Schmalhausen�s plants were obtained.

Ettingshausen�s specimens from Bilin in Bohemia assigned by Pirn to Taxodium duhiutn may, as Gardner says�, be examples of

The name Geinitzia was given by Endlicher� to a piece of sterile �Poot from Lower Cretaceous strata in Saxony previously figurednbsp;Py Geinitz*� as Araucarites Reichenbachii (fig. 769), and in the newnbsp;genus was also included Cryptomeria primaeva Corda**. Both thesenbsp;species were referred by Endlicher to Geinitzia crelacea. Corda�snbsp;�Pecies.was founded on several foliage-shoots from Lower Cre-faceous rocks in Bohemia with the habit of Araucaria excelsa andnbsp;iu one or two instances bearing what appear to be terminal budsnbsp;described by Corda as small cones. In the first instance Geinitzia

^ Knowlton (99) B. p. 681. nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;Lesquereux (78) B. PL lxv.

^ Penhallow (02) p. 50. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Bleicher and Fliche (92) p. 382.

� Saporta (65^) PL ii. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Schmalhausen (83^) Pis. xxxii., xxxvi.

�� Palibin (01) p. 499. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Gardner (86) p. 39.

� Endlicher (47) p. 280. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;Geinitz (42) PL xxiv. fig. 4.

Corda in Reuss (46) B. nbsp;nbsp;nbsp;PL XLlii. figs.nbsp;nbsp;nbsp;nbsp;1�11.

was applied to branches without any recognisable cones. In 1868 Heer^ figured specimens from the Kome beds in Greenland whichnbsp;he believed to be identical with Araucarites Reichenhachii Gein.nbsp;though the foliage-shoots bear shorter leaves than those on thenbsp;type-specimen of Geinitz: Heer states that he was able to examinenbsp;the type-specimen and assured himself of the specific identity of thenbsp;German and Greenland specimens; he substituted the generic name

Sequoia for Araucarites on the ground that some cones in the T�bingen Museum from Lower Cretaceous beds in Moravia,nbsp;attached to branches apparently identical with Araucarites Reichenhachii, presented a very close resemblance to those of recentnbsp;Sequoias. The Moravian specimens, which he afterwards figured^,nbsp;are oval and the cone-scales have distally expanded distal endsnbsp;(fig. 770) like those of Sequoia, but no evidence was obtained as tonbsp;the number of seeds. Additional examples of vegetative shootsnbsp;and cones were described by Heer� from Greenland as Sequoia

^sichenbachii and this species is recorded by authors from many Lower Cretaceous localities, but in no case is any conclusive evidence brought forward in support of the assumed generic identitynbsp;with Sequoia. Specimens of Sequoia Reichenbachii with foliage andnbsp;cones are figured by Velenovsky^ from Bohemia showing clearlynbsp;the characteristic peltate cone-scales, and similar examples thoughnbsp;with rather larger cones are described from Lower Cretaceous stratanbsp;in North America^. On the other hand the name Sequoia Reichen-bachii is applied in some cases to fragments of sterile branches^ un-8'Ccompanied with cones�: in one instance^ evidence was obtainednbsp;of the occurrence of separate circular bordered pits on the tracheidsnbsp;cf some vegetative branches from the Cretaceous beds of Aix-lanbsp;Lhapelle. It is impossible to say whether such shoots bore conesnbsp;like those of Geinitzia or Elatides; some at least belong to Elatidesnbsp;lt;^urvifolia.

Two conclusions are suggested by an examination of the records so far quoted: the use of the generic name Sequoia is not based onnbsp;solid foundation and, secondly, it is unsafe to assume thatnbsp;fragments of sterile branches bearing falcate leaves similar to thosenbsp;on fertile shoots referred to S. Reichenbachii belong to that species.nbsp;Lhe common occurrence of Mesozoic specimens agreeing more ornbsp;less closely with Araucaria excelsa, while demonstrating thenbsp;abundance of that form of vegetative shoot, by no means provesnbsp;Ihe equally wide occurrence of one specific type. It has, fornbsp;example, been shown by Nathorst� that the branches from Lowernbsp;Lretaceous or Upper Jurassic rocks in Spitzbergen figured bynbsp;H�er8 as Sequoia Reichenbachii are examples of Elatides curvifoUanbsp;(Bunk,). The genus Elatides�^ is characterised by cones differing innbsp;their flatter scales and more elongated form from those usuallynbsp;assigned to Sequoia though the foliage-shoots are of the same type.nbsp;It is therefore advisable to adopt some provisional generic term fornbsp;sterile shoots resembling in habit those of Araucaria excelsa andnbsp;which in the absence of cones cannot be safely assigned to a genusnbsp;tounded on the cone-characters. The name Pagiophyllum^ servesnbsp;^ Velenovsky (85) B. Pis. viii., ix.

quot; Ward (99) B. Pis. 165, 166; Hollick (06) PI. n. fig. 40; PI. ra. figs. 4, 5.

Heer (75) ii. Pis. xxxYi., xxxvii. � See page 270. nbsp;nbsp;nbsp;* See page 274.

this purpose and it should be applied to sterile branches of the Araucaria type which cannot reasonably be referred to Elatides,nbsp;Geinitzia, or other genera connoting certain types of fertile shoot.nbsp;It has, however, been pointed out that in the first instance Geinitzianbsp;was applied to sterile shoots, but later this designation came to benbsp;associated with cones of elongate-oval form bearing peltate scales.nbsp;In 1852 Unger^ applied Geinitzia to a specimen from Neustadtnbsp;consisting of a slender piece of foliage-shoot and an imperfectlynbsp;preserved cone similar to the cones of Heer�s Sequoia Reichenbachiinbsp;but longer in form. Subsequently Heer^ described under the namenbsp;Geinitzia formosa shoots and cones from Lower Cretaceous stratanbsp;at Quedlinburg: the cones are similar in form to that figured bynbsp;Unger and bear cone-scales with polygonal distal ends having anbsp;central umbo and radially disposed lines on the exposed surface.nbsp;Schenk� also gives good dra wings of Geinitzia formosa. A well preserved cone very like Heer�s G. formosa was described by Newberry^nbsp;from the Amboy clays as Sequoia gracillima, the specific namenbsp;having been previously used by Lesqiiereux for sterile branchesnbsp;from Dakota in conjunction with the generic name Glyptostrobus.nbsp;Newberry adopted Lesquereux�s specific term because he found innbsp;the Dakota beds cones like that from the Amboy clays associatednbsp;with the branches described by Lesquereux. Newberry�s cone isnbsp;practically identical with that of Heer�s Geinitzia formosa, but it isnbsp;noteworthy that the former is borne on a slender branch havingnbsp;small appressed leaves in place of the more spreading falcate leavesnbsp;of Heer�s species. This difference in the foliage is of secondarynbsp;importance in comparison with the close resemblance between thenbsp;cones. Subsequently Jeffrey� obtained good cones from thenbsp;Matawan formation apparently identical with Sequoia gracillimanbsp;(Lesq.) as figured by Newberry and he was able to investigate thenbsp;anatomical features. The pith of the cone-axis contains groups ofnbsp;sclerous cells; the phloem differs from that of Sequoia in the absencenbsp;of fibres, while the secondary wood has no resin-cells�anothernbsp;difference from Sequoia: the tracheal pits are circular and in nonbsp;case contiguous and there are no rims of Sanio. The latter feature

Heer (71-) p. 6, Pis. i., ii. Newberry and Hollick (95) jr. 50, PI. ix. figs. I�3.nbsp;nbsp;nbsp;nbsp;^ Jeffrey (11).

vascular bundles were found to be arranged round the margin of the peltate portion of the scales and completely surrounded bynbsp;transfusion tissue, �a feature of marked contrast to the scale-bundles in Sequoia and at the same time one which indicates anbsp;strong affinity with the Araucarineae.� The mature scales affordednbsp;no indication of the number or place of attachment of the seeds,nbsp;but an immature cone lent support to the view that each scale borenbsp;four ovules on the peduncle near the cone-axis. Hollick andnbsp;Jeffrey regard the scales as Araucarian and think it probable thatnbsp;they were connected with the twigs named by them Geinitzianbsp;Reichenbachii.

A special generic name^ is given to some four-sided scales on the ground that they not only differ in their tetragonal form from thenbsp;hexagonal scales of Eugeinitzia but probably belonged to a smallernbsp;cone. As in the former type the vascular bundles are enclosed bynbsp;transfusion-tracheids. The investigations of the American botanistsnbsp;show that the sterile branches, G. Reichenbachii, exhibit certainnbsp;Araucarian tendencies and that the cone, Geinitzia gracillinia, asnbsp;also the detached cone-scales, Eugeinitzia and Pseudogeinitzia,nbsp;cannot be included in Sequoia.

Until more is known of the morphological nature of the cones described by Heer and other authors as Sequoia Reichenbachii,nbsp;S. ambigua, etc., their relationship to existing Conifers cannot benbsp;settled, but meanwhile it would seem convenient to include bothnbsp;the smaller oval cones and the longer forms represented by G.nbsp;gracillima in the same genus Geinitzia, applying the name to conesnbsp;having spirally arranged scales with peltate distal ends superficiallynbsp;resembling those of Sequoia. The name Sequoia, much too freely usednbsp;by palaeobotanists, has in some cases^ been applied to cone-bearingnbsp;branches that are almost certainly identical with Sfhenolefidiumnbsp;Kurrianum (Dunk.). On the other hand for sterile foliage-shootsnbsp;unconnected with cones the non-committal name Pagiophyllum isnbsp;suggested on the ground that foliage-shoots alone cannot be morenbsp;precisely determined.

Heer^ instituted S'phenole'pidium in place of Sphenolepis, proposed by Schenk^ for Wealden Coniferous branches, because of the previous use of the latter name by Agassiz for a genus of fishes.

errys bas recently reverted to the original form Sphenolepis on the ground that its employment by zoologists is not a seriousnbsp;^^jection. Schenk�s definition of his genus includes both vegetativenbsp;Organs and cones, but it is desirable that the name Sphenolepidiumnbsp;should be restricted to fertile specimens or at least to specimensnbsp;^hich can with reasonable certainty be connected with cone-^earing examples. The habit of the foliage-shoots of the two best-own Wealden and Lower Cretaceous species, S. Sternbergianumnbsp;S. Kurriamim, is of the type which leads authors to employnbsp;�uch generic terms as Sequoia, Athrotaxites or Afhrotaxopsis,

the absence of cones it is impossible to feel confidence in any attempts to distribute such sterile specimens among genera whichnbsp;^re characterised not only by a certain form of foliage-shoot but

The generic name Sphetiolepidium should be retained only for �P6cimens with small, more or less globose, cones possessing spirallynbsp;^�posed cone-scales, cuneate, relatively broad and fairly thick.

6 cones are much smaller and have relatively broader and dignbsp;nbsp;nbsp;nbsp;than those of Elatides though there is no essential

and^described throwing any light on the affinity of the genus ofnbsp;nbsp;nbsp;nbsp;others it must be left for the present in the category

orn/erac inceriae sedis. The leaves are spirally disposed on the slender branches and are either ovate, triangular,nbsp;S'Udnbsp;nbsp;nbsp;nbsp;acuminate region, or longer and more spreading

^ falcate; the latter type agrees with Pagiophyllum while some Sphenolepidium cones are rather of the Brachyphyl-^ ^TP�- Many of the specimens recorded as Sphenolepidiumnbsp;^ no evidence as to the nature of the cones and should benbsp;^signed to Pagiophyllum or Brackyphyllum. The genus is charac-jQnbsp;nbsp;nbsp;nbsp;Wealden or Lower Cretaceous strata and is represented

This species was originally described by Dunker^ from North Germany as Muscites Sternbergianus and by later authors placednbsp;in Araucarites, Widdringtonites, and other genera^. It is impossiblenbsp;to determine the specific limits of this species� and S. Kurriamimnbsp;(fig. 771); the cones exhibit no well-defined distinguishing characters and the chief distinctionnbsp;is the more spreading foliage of the Ar'aucariannbsp;or Pagiofhyllmn type of S. Sternbergiamm. Asnbsp;Berry suggests, this species�described from thenbsp;Potomac formation and elsewhere in Northnbsp;America and from several European localities�nbsp;is probably represented in the Lower Cretaceousnbsp;flora of Greenland under such names as Glyfto-strobus groenlandicus Heer and Sequoia /asfigiata.

Some of the English and German fossils attributed to S. Sternbergianum are almost certainly examples of Elatides curvifolia.

Dunker ^ originally adopted the generic name Thuites; later authors pieteiied Brachyphyllum,

Eontaine�, who records this species from the Potomac formation, includes in his genus Athro-taxopsis specimens which cannot be distinguishednbsp;by any features of morphological importancenbsp;from Sphenolepidium Kurrianum. The leaves arenbsp;ovate, more or less appressed, agreeing withnbsp;Brachyphyllum or in some examples intermediatenbsp;between the type of foliage assigned to Pagio-phyllum and Brachyphylhim. The Wealden specimen reproduced in fig. 771 is placed in Sphenolepidium because of its association with branches,nbsp;identical in habit, bearing cones; if found

dossil it would be referred to Brachyphylluni. The figured specimen �bo\vs the variable form and size of the leaves and there isnbsp;�ood reason to believe that the plants represented by the frag-^^ents included in one or other species of Sphenolepidium werenbsp;characterised by a considerable range in the habit of the foliage-�ho.ots, a fact which renders of little importance the separationnbsp;S. Kurrianum and S. Sternbergianum based on the form of thenbsp;caves in detached branches. The small cones borne terminally onnbsp;�lcnder branches resemble superficially the cones of Athrotaxis, butnbsp;'^0 facts are available as to the structure of the cone-scales andnbsp;fhere is no evidence on which to found an opinion as to the positionnbsp;cf the genus.

clear what morphological feature is represented by the �

^ Schmalhausen (79) A. p. 39. nbsp;nbsp;nbsp;� Zeiller (96) A. p. 477.

^ J^athorst (97) p. 19. nbsp;nbsp;nbsp;* Schimper and Schenk (90) A. p. -93

' Goeppert and Menge (83) A. p. 36, PI. xm. figs. 117�123.

Though several fossil plants have been compared with the existing species Sciadopitys verticillata, in no case is there any con-wsive evidence of the occurrence of this type of Conifer. Schmal-iisen founded the genus Cyclopitys'^ for impressions of shootsnbsp;oni Russia bearing whorled linear leaves which he believed to benbsp;�sely allied to or generically identical with Sciadopitys. Zeiller^nbsp;''ho brought forward strong arguments for assigning the stratanbsp;^�cgarded by Schmalhausen as Jurassic to the Permian period,nbsp;considers Cyclopitys to be an Equisetaceous plant. Detached linearnbsp;^9,ves similar to those of Cyclopitys are abundant in many Jurassicnbsp;I'as .and, as ISTathorst� says, they may be compared with severalnbsp;*^ccent genera including Sciadopitys, but without anatomical datanbsp;^ccurate determination is impossible. It is stated by Schenk* thatnbsp;�.fquot;*�cfaceous leaves described by Heer as Pinus Crameri agree innbsp;cir epidermal features with the foliage of Sciadopitys, but in thisnbsp;in other cases generic identity or even close relationship has notnbsp;j cn demonstrated. Goeppert and Menge� describe some singlenbsp;^a-ves preserved in Baltic amber as Sciadopitytes linearis andnbsp;d^iucescens; they speak of the leaves as having a single vein onnbsp;upper face and two veins on the lower surface though it is not

Schenk states that these Oligocene leaves are Dicotyledonous and not the leaves of a Conifer.

Specimens of fossil wood have been described exhibiting certain features, especially the pitting of the medullary-ray cells, similar tonbsp;those of Sciadofitys^ but the occurrence of such features in othernbsp;recent genera precludes a definite reference to any one type.

This name has recently been revived by Halle^ for two species of Cretaceous leaves from Greenland, one of which, Pinus Cramerinbsp;Heer, though compared by Schenk with the leaves of Sciadopitys,nbsp;was not actually included in Sciadopitytes, while the other is a newnbsp;species, Sciadopitytes Nathorsti. Halle describes these leaves asnbsp;�Conifer-like in habit, with a dorsal groove which is protected bynbsp;elongated papillae and whose epidermal tissue differs from that ofnbsp;the rest of the leaf through a non-seriate arrangement of the cellsnbsp;and the occurrence of stomata.�

The outstanding feature of the leaves of Sciadopitys. is the double nature of the lamina and the morphological peculiaritiesnbsp;which have led to its recognition as a phylloclade; but, as Hallenbsp;admits, there is no evidence that the fossils are other than ordinarynbsp;simple leaves. The interesting characters described by Halle amplynbsp;justify the use of a generic name separating the leaves from thosenbsp;known only as impressions, without any structural features preserved, and referred to Pityophyllum. It is, however, open to question whether the name Sciadopitytes does not imply more than thenbsp;facts support. The leaves named by Heer� Pinus Crameri arenbsp;about 12 mm. long and 2-5 mm. broad: the apex is bluntly roundednbsp;and the base is slightly widened. Halle points out that there isnbsp;evidence that the leaves were cylindrical. The carbonised leavesnbsp;of this species form thick masses in the shale and excellent preparations of the cuticle can be obtained. Halle considerablynbsp;extends Schenk�s account of the epidermal characters. Thenbsp;apparent midrib is a groove and there is no indication of a truenbsp;median vein. Rather large stomata are crowded in the groove and

� Heer (68) j. PI. XLiv. figs. 7�18; (75) ii. PI. xxill. pp. 9�15; Halle (15) p. 509, PI. XIII. figs. 1�13.

sloping sides of the groove bear cylindrical papillae. On the whole l-lie structure recalls that of the recent Sciadofitys though as Hallenbsp;shows there are certain difierences.

The second species, Sciadopitytes Nathorsti^, was discovered by hiathorst in the Middle beds of Atanekerdluk in West Greenland.nbsp;The leaves are at least 40 mm. long and about 1 mm. broad. thenbsp;�tomata are confined to the groove as in S. Crameri and numerousnbsp;papillae are borne on the borders of the median depression. Thenbsp;stomata are not so crowded as in S. Crameri and differ less in sizenbsp;li'oin the other epidermal cells. These two species, though exhibitnbsp;liig some similarity to Sciadofitys, can hardly be assumed to belongnbsp;plants more closely allied to the recent Japanese Conifer thannbsp;to other existing forms. The occurrence of the characters describednbsp;hy Halle may be recognised by adding the name Sciadopitytes after

CHAPTER XLVIIl.

The relative antiquity of the different families of the Coniferales is a question which every student of the geological history of thenbsp;group desires to answer. Reference has already been made to thenbsp;different views that are held with regard to the phylogeneticnbsp;relations of the Araucarineae and the Abietineae: conclusions onnbsp;this subject are based partly on the morphological characters exhibited by recent types and in part on palaeobotanical data. Thenbsp;evidence afforded by petrified wood is briefly dealt with in Ch. XLiv:nbsp;this shows that the features associated with modern Abietineae donbsp;not stretch as far back into the past as is the case with the typenbsp;represented by the wood of the Araucarineae. The evidence derivednbsp;from a study of impressions of foliage-shoots and cones as well asnbsp;the meagre data supplied by petrified cones is less easy to interpretnbsp;because of the greater imperfection of the records. The southernnbsp;distribution of the Araucarineae predisposes the student in favournbsp;of a southern origin, while the essentially northern range of thenbsp;Abietineae suggests that this family had its birth north of thenbsp;equator. But conclusions based on such considerations requirenbsp;confirmation from other kinds of evidence. In the Jurassic-Wealdennbsp;period the Araucarineae were well represented in the northernnbsp;hemisphere and the impression gained from a survey of Jurassicnbsp;records is that the Araucarineae shared with other types an almostnbsp;world-wide distribution. It is much easier for a palaeobotanist tonbsp;form an opinion as to the period of maximum development andnbsp;vigour of a given set of plants than to discover a substantialnbsp;foundation on which to rest a view as to the first appearance ornbsp;the original home of the earliest representatives of the family-type.nbsp;It is, perhaps, significant that the Araucarineae are represented innbsp;the Jurassic floras of Graham Land on the edge of Antarctica,nbsp;Australia, and India. The Abietineae, on the other hand, do not

^ulk largely in Mesozoic floras before the closing stages of the Jurassic period and more especially in the earlier days of thenbsp;Cretaceous era. The abundance of Abietineous cones in Lowernbsp;Cretaceous strata, a period later than that in which the Arau-carineae are abundantly preserved in plant-bearing deposits, atnbsp;least points to a later maximum development of the Abietineae,nbsp;and such data as we have seem to favour a northern rather thannbsp;a southern origin. Winged seeds, hardly distinguishable from thosenbsp;nf modern Pines (fig. 788, p. 396), from Ehaetic beds in the South ofnbsp;Sweden, foliage-shoots from beds of the same age exhibiting featuresnbsp;now associated with the Abietineae, demand serious considerationnbsp;In connexion with the antiquity of the family, though it can hardlynbsp;maintained that they, furnish proof of the existence in Ehaeticnbsp;*^nd Liassic floras of true Abietineae. The occurrence of a wingednbsp;pollen-grain (fig. 491, G; Yol. iii. p. 298) in the partially decayednbsp;^mod of Antarcticoxylon might be urged as a plea for a southernnbsp;Origin of the family, but an extended bladder-like exine is not anbsp;monopoly of the microspores of the Abietineae.

The following types selected in illustration of the fossil records of the Abietineae show how difficult it is in many cases to deter-�line the precise position within the family to which cones ornbsp;foliage-shoots should be assigned. Palaeobotanical literature con-fmns many species referred to Abies or Ahietites, Cedriis, and othernbsp;8�nera, but it is usually impossible from the available data to carrynbsp;identification so far. A few examples may be quoted: certain Lowernbsp;Cretaceous cones bear a very close resemblance to those of Cedi us ,nbsp;f�rit an examination of some of the less familiar cones of existingnbsp;species of Ahies and Picea shows that the reasons for connecting thenbsp;fossils with Cedrus are not entirely satisfactory. The fossil woodnbsp;described under Cedroxylon does not denote that the parent-plantsnbsp;quot;'ere more closely allied to Cedrus than to some other genera of thenbsp;s^'Dre family. Boulay^ has described some seeds from Miocene bedsnbsp;Prance ae Cedrus vivariensis which he unhesitatingly regards asnbsp;^enerically identical with those of recent Cedars, and there is nonbsp;j^oason to doubt the correctness of this conclusion. Cone-scalesnbsp;Caring two seeds from Miocene beds in Spitzbergen described by

Heer^ as Pinus (Cedrus) Lopalini may belong to a true cedar cone, but the evidence is hardly convincing. It is not too much to saynbsp;that even Tertiary records of Conifers seldom enable us to discriminate between individual genera. In the absence of anatomicalnbsp;data the needle-like leaves scattered through Mesozoic and Tertiarynbsp;strata cannot be identified with reasonable certainty. From Uppernbsp;Pliocene beds in Germany Geyler and Kinkelin^ described a cone asnbsp;Abies Loehri, and this has more recently been identified as Keteleerianbsp;(fig. 786, C, p. 394) by Engelhardt and Kinkelin� on the strengthnbsp;of its external resemblance to K. Davidiana. The reference to Abies *nbsp;of some leaves enclosed in the Baltic amber affords an example ofnbsp;the assistance afforded by characters recognisable in well preservednbsp;material, and it is probable that a fuller knowledge of the epidermalnbsp;characters of recent Conifer leaves may supply a useful aid to morenbsp;precise identification.

Endlicher� employed the name Pinites for leaves, male flowers, and cones considered to be closely allied either to recent species ofnbsp;Pinus or to some other genus of the Abietineae, such as Abies,nbsp;Larix, or Picea. Many authors have adopted the generic namenbsp;Pinus in cases where the evidence appears to them sufficientlynbsp;strong to indicate identity with the existing genus, but it is onlynbsp;cones and foliage-shoots from Tertiary and Pleistocene beds thatnbsp;can as a rule be definitely assigned to such a position. It may,nbsp;perhaps, be carrying consistency too far to restrict Endlicher�snbsp;designation to such specimens as there is good reason for connectingnbsp;with the recent genus Pinus; but the more restricted use of Pinitesnbsp;has the merit of being less likely to mislead the student and,nbsp;chiefly on that account, I propose to adopt the genus Pityites fornbsp;Abietineous fossils which cannot with confidence be referred to anbsp;more precise position. In practice this designation will not oftennbsp;be employed as in most cases cones and vegetative organs occurnbsp;as separate fossils and are most conveniently described under the

� Engelhardt and Kinkelin (08) p. 216, PI. xxvi. fig. 7. � Goeppert and Menge (83) A. PI. xiii. figs. 107�110.

terras suggested by Nathorst and mentioned below. Pityites is, however, appropriate for such specimens as those represented innbsp;%s. 772, 773 which show a direct connexion between cones andnbsp;toliage-shoots.

Gloeppert adopted Pinites for fossil wood in a wide sense, but It has long been the custom to describe petrified wood agreeingnbsp;structurally with recent Pines and other members of the Abietineaonbsp;Wilder Kraus�s term Pityoxylon. Nathorst^, with a view to greaternbsp;�Convenience, proposed certain subgeneric names as qualifyingnbsp;epithets indicating the nature of the fossils but not implying anbsp;�direct connexion with Pinus: he adopted the names Pityanthusnbsp;^quot;Cr male flowers suggesting alliance with those of some Abietineousnbsp;genus, Pityostrohus for cones, Pityohfis for cone-scales, Pityosfer-^U�n for seeds, Pityocladus for vegetative shoots, and Pityofhyllumnbsp;^er detached leaves. To these the name Pityosporites^ has recentlynbsp;^�en added.

The generic or rather subgeneric term Pityophyllum is apt to jcfislead the student if used in conjunction with Pinites: thenbsp;^-Ves so named, as Nathorst admits, are in many instances almostnbsp;�Certainly derived from plants which do not belong to the Abie-^iReae. Under Pityophyllum are included both needle-like leavesnbsp;J'^hich are probably Abietineous with others having a broadernbsp;ttuna (fig. 770) and much more likely to be connected with suchnbsp;fe^aera as Cephalotaxus, Torreya, or Podocarpus.

The term Pityosporites^ is proposed for microspores provided ^ings similar to those of Pinus and other members of thenbsp;letineae, though in this case also relationship with anothernbsp;hy, namely the Podocarpineae, is not excluded. These termsnbsp;^her used as subgeneric titles or as generic designations serve anbsp;�6ful purpose for disjuncta membra, while the name Pityites isnbsp;^mployed for specimens of a more complete kind. The namenbsp;lias often been used for vegetative shoots and cones^nbsp;'*^h ther^ is no adequate reason for assigning to a positionnbsp;^�arer to Abies than to other genera of the same family; it isnbsp;��irable to restrict the term to fossils which afford evidence of

4 nbsp;nbsp;nbsp;(97) p. 62; (99) p. 16.nbsp;nbsp;nbsp;nbsp;quot; Seward (14) p. 23.nbsp;nbsp;nbsp;nbsp;a

affinity to the recent genus. Similarly, names such as Laricites Cedrites and others implying a more precise determination than isnbsp;suggested by Pityites'maj conveniently be used either as subgeneric or generic terms.

In the account of recent Conifers allusion is made to the views held by students of fossil plants with regard to the relative positionnbsp;of the Abietineae and the Araucarineae in a chronological sequence.nbsp;The types selected for description are intended to serve as guides tonbsp;those who wish to draw conclusions from the geological records,nbsp;but so long as we have to trust chiefly to impressions without the.nbsp;more certain guidance of anatomical data the inferences drawnnbsp;cannot be regarded as other than provisional. The evidence ofnbsp;fossil seeds is difficult to interpret, as its value depends on thenbsp;amount of importance to be attached to the occurrence of specimens closely resembling in the form of the wing the seeds of recentnbsp;Pines and other Conifers. The winged seeds of Agaihis differ in thenbsp;shape of the membranous appendage from those of Abietineousnbsp;species, and the oldest winged seeds attributed to the Abietineae,nbsp;from Khaetic rocks, exhibit a closer agreement with the Abietineousnbsp;type. On the other hand it is questionable whether the form of anbsp;wing constitutes a safe criterion of affinity. A similar difficulty isnbsp;presented by �winged� pollen-grains: a bladder-like extension ofnbsp;the exine though usually associated with the Abietineae is anbsp;character which is not confined to that family. Foliage-shoots likenbsp;those of recent Abietineae are recorded from Rhaetic rocks andnbsp;later Mesozoic strata, but we have no means of determining in thenbsp;case of the oldest examples whether their superficial resemblancenbsp;to branches of Cedrus and other genera has a phylogenetic significance. The generic name Pinites is applied by Renault^ to a slendernbsp;branch from Permian rocks in France bearing spirally disposednbsp;filiform leaves 3 cm. long apparently borne singly and directly onnbsp;the main axis, not on short shoots. It is elsewhere ^ suggested thatnbsp;this specimen, Pinites penniensis, may belong to a pl^nt allied tonbsp;Dicranophyllmn: there is certainly no adequate reason for thenbsp;employment of the generic term Pinites. Similarly an impressionnbsp;figured by Stur� from the culm of Altendorf as Pinites antecedens,

^'hich I was able to examine in the Vienna collection, is too frag-�'^entary to be determined. The occurrence of linear leaves in fascicles is in itself no real evidence of Abietineous affinity: thenbsp;flustered leaves of Czekanowskia and Phoenicofsis, especially thenbsp;�rmer, though essentially similar in habit to the foliage-shoots ofnbsp;Some Abietineae are generally believed to belong to plants ofnbsp;s-nother class. The evidence furnished by petrified wood hasnbsp;already been considered: the important point is that there is nonbsp;s^'tisfactory case of the occurrence of fossil wood of Palaeozoic age^

^ aving typical Abietineous features, a fact of importance in relation 0 the widely spread Palaeozoic woods agreeing in essentials withnbsp;� Araucarian type.

This name was proposed for some cones attached to foliage-^^oots as well as detached cones and vegetative branches from maiden rocks on the coast of Sussex^: the type-specimens formnbsp;P^rt of the rich EufEord collection in the British Museum. Thenbsp;^J�anches are covered with the elongated persistent bases of scale-a-ves and in the axils of these are borne numerous long needlesnbsp;g- 772). Tbe cones are oblong and bear broad, rounded, scalesnbsp;� those of Pinus Strobus, P. excelsa (fig. 773; cf. fig. 704), Piceanbsp;� Abies; they agree closely with Pityostrobus Carruthersi (Gard.)nbsp;also with P. Andraei (CoeiU.)� from Lower Cretaceous rocks innbsp;gium and with the smaller cones from the Potomac formationnbsp;cubed by Pontaine^ as Abietites ellipticus. The preservation isnbsp;sufficiently good to show the number of leaves in each foliage-needles may have been borne in dense clusters as innbsp;In general habit the species resembles Cednis and Lavixnbsp;^ ^g the greater length of the needles is more in accordance withnbsp;j ^ species of Pinus. Shoots similar to those of this species arenbsp;^f^nbsp;nbsp;nbsp;nbsp;by Prepinus statensis Jeff.� from the Cretaceous beds

j �^cischerville. Dr Stopes�, following the example of Berry, crs this species to Abietites.

See page 220, also Thomson and Allin (12). Seward (95) A. p. 196, Pis. xvin., xix.

In his account of petrified material from Franz Josef Land Solms-Laubach^ describes sections of a Pinus-like leaf.from Bell

Island (Eira harbour) probably of the same geological age as the plant-beds of Cape Stephen, which is believed to be Upper Jurassicnbsp;or Lower Cretaceous. Through the kindness of the Director of the

^ Solms-Laubach (04) p. 12, PL i. fig. 14; PI. Ti. fig. 3. For evidence as to geological age, see Newton and Teal] (97), (98); Nathorst (99).

Geological Survey I have been able to examine the sections in the Jernayn Street Museum. Graf Solms-Laubach describes the leavesnbsp;oval in section, the upper face strongly convex and the lowernbsp;almost flat as in two-needled Pines, but as shown in fig. 774, A thenbsp;leaves may be approximately cylindrical (1 mm. in diameter), likenbsp;l^hose of Pinus monofJiylla or the leaves of Cedrus. There is a singlenbsp;''^ein accompanied by some radially disposed transfusion-tracheids,nbsp;the whole being enclosed in a single layer of rather thick-walled

but^ quot;Inhere is no distinct division of the bundle into two halves la'rnbsp;nbsp;nbsp;nbsp;indications of the presence of a broad median medul-

y ray. The mesophyll-cells have prominent infoldings precisely ^�cent Pines, Cedrus, and some other Abietineae (fig. 774, B;nbsp;^nbsp;nbsp;nbsp;nbsp;694): the epidermis has a thick cuticle and below it are

istinguish any undoubted canals in the leaf shown in fig. 774. occasional absence of canals in Abietineous leaves normally

possessing them is mentioned in Chapter xliii. The leaf for which the specific name eirensis (from Eira harbour) is'proposed affords

an interesting example of an Abietineous type, in all probability of Upper Jurassic age, exhibiting a remarkable resemblance to certainnbsp;recent species especially Pimis monofhylla.

Under this name, used as a subgenus of Pinites, Nathorst includes branches bearing short shoots similar in habit to those ofnbsp;'Gedrus and Larix. Branches of this type, bearing leaves and cones,nbsp;nre illustrated by Pityites Solmsi (figs. 772, 773), but in some casesnbsp;such vegetative shoots occur as detached fossils and it is to themnbsp;that Nathorst�s term may conveniently be applied. The strikingnbsp;resemblance of the fossil specimens to shoots of Cedrus and Larixnbsp;nnd the frequent association or attachment of needle-like leavesnbsp;nfiord strong grounds for assigning the branch-fragments to thenbsp;Abietineae.

In his description of Schizolefis Follini Nath.^ from Ehaetic rocks in Scania, Nathorst includes not only cones mth lobed scalesnbsp;characteristic of Schizolepis but leaves and branches. Solms-Laubach� expressed the opinion that we know nothing of thenbsp;foliage of Schizolepis �for there is nothing to make it even probablenbsp;that the numerous needles which lie one above another in the beds

Palsj�, any more than the branches beset vuth needle-bearing shoots which Schenk has referred to this genus, have any connexionnbsp;quot;with Schizolepis.� In a later account of Schizolepis, Nathorstnbsp;�^ggests the advisability of separating the leaves and branchesnbsp;from the Schizolepis cones, though as he says the association of thenbsp;f'vo sets of organs in more than one locality may be significant.

is, therefore, preferable to assign the vegetative organs to ^dyocladus, at the same time keeping in mind the possibility of annbsp;Original connexion with the cones described under the generic namenbsp;^ohizolepis. lu order to avoid confusion I have removed thenbsp;I�ranches and associated leaves from Schizolepis Follini to a distinct species Pityocladus Nathorsti. The specimens figured bynbsp;i^athorst consist of (i) a fairly stout axis bearing a smaller lateralnbsp;�hoot like that on which the leaf-clusters of Cedrus and Lari* atenbsp;home; (�) separate short shoots characterised by zones of smallnbsp;�ears alternating with smooth areas; (iii) numerous crow ded linearnbsp;leaves.

Pityocladus longifolius (Nathorst). [And Pityophyllum longifolium Nathorst.]

This species was first described by Nathorst^ from Ehaetic beds in Scania as Taxites longifolius and afterwards as ? Cycadites.

Fig. 775. Pityocladus longifolius. Leaves attached to a short shoot covered with scales. (Nat. size;nbsp;a piece of lamina enlarged to show the fine transverse striations. From a specimen from Scanianbsp;in the Stockholm Museum.)

M�ller^ adopted Nathorst�s generic name Pityofhyllum for detached leaves and that designation should be used for specimens 1 Nathorst (78�) B. p. 50.nbsp;nbsp;nbsp;nbsp;^ Roller (03) p. 40, PI. vi.

'vhich afford no evidence as to the nature of the axis. The specimen from the Stockholm Collection represented in fig. 775, from thenbsp;Ifhaetic rocks of Scania, is especially interesting as affording onenbsp;of the few examples of leaves of the type known as PityofhyUumnbsp;^ongifolimn attached to an axis covered with short scales. Thenbsp;lamina varies from 1 to 5 mm. in breadth and may be broader. asnbsp;seen in the enlarged portion the lamina is transversely wrinkled, annbsp;Appearance characteristic of most forms of the broader Pityo-Vhyllum leaves and probably produced by contraction on drying.nbsp;There is little difference between this species and the leaves figurednbsp;I�y Nathorst from Scania as Schizolepis Follini. An accuratenbsp;specific delimitation of Pityophyllum leaves is hopeless. Detachednbsp;leaves (fig. 776) similar to those shown in fig. 775 are recorded bynbsp;Violier from Lower Jurassic beds in Bornholm, by other authorsnbsp;fiom Jurassic strata in Turkestan^, South Russia^, Oregon�, Spitz-Tergen^^ and elsewhere.

Schenk� also refers to Schizolepis several specimens of branches And leaves as well as cones which he includes in Schizoleyts Brauniinbsp;Schenk. The larger branches bear leafmushions and short lateralnbsp;shoots with scale-leaves at the base, and in some examples tufts ofnbsp;i^eedles occur on the short shoots. A specimen described by Braunnbsp;As Isoetes puniilus is identified by Schenk in his monograph of thenbsp;^�haetic Flora� as a leaf-bearing shoot like that of a recent Pine.nbsp;The Swedish specimen represented in fig. 775, also of Rhaetic age,

This species originally described from the Jurassic beds on the Rubuk River in Chinese Dzungaria as Pinites�� is founded onnbsp;ranches bearing short shoots almost identical with P. Schenkt.nbsp;^hort shoots not more than 1 cm. long are borne spirally on anbsp;thicker axis and covered with small leaf-scars (fig. 777) exactlynbsp;A� in the corresponding shoots of Cednis or Larix. With the

� Fontaine in Ward B. (05) PI. xxxv. nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;gg 2.

branches are associated numerous needles, 1 mm. broad and at least 5 cm. long, sometimes covering thenbsp;whole surface of the rock. The specimensnbsp;agree closely with the foliage-shoots ofnbsp;Pityites Sohnsi: similar examples arenbsp;described by Ettingshausen^ from Li-assic strata as Halochloris baruthinanbsp;Ett. A branch with short shoots fromnbsp;Jurassic rocks in Amurland describednbsp;as Pinites sp. cf. P. kobukensis^ may benbsp;specifically identical with the Dzungarianbsp;fossils: an example of the same typenbsp;lent to me by Dr Krystofovic fromnbsp;Jurassic beds of Amurland shows anbsp;forked lateral foliage-shoot. This authornbsp;has recently described a specimen fromnbsp;Jurassic rocks in Transbaikalia as Pinites (Pityophyllum) cf. P. kobukensis^.

�1 nbsp;nbsp;nbsp;11 ivT ,1nbsp;nbsp;nbsp;nbsp;,1 T. � �nbsp;nbsp;nbsp;nbsp;.T...nbsp;nbsp;nbsp;nbsp;Seward; nat. size.)

cladus) spp. a and b from Upper Jurassic beds in Spitzbergen, and compared by him with Pityites Solmsi.

This name� is applied to detached leaves of needle-like form like those of recent Pines or to long linear leaves broader and flatternbsp;than the needles of Pinus. Some of the specimens referred tonbsp;this genus are very similar to the leaves of Keteleeria. In a fewnbsp;cases (fig. 775) the leaves are still attached to a short shoot butnbsp;usually they occur as detached specimens (fig. 776). The genus isnbsp;met with in Rhaetic strata but is specially abundant in Jurassicnbsp;floras and persists through Cretaceous and Tertiary rocks. Thenbsp;leaves generally described under this generic term are broader andnbsp;flatter than such leaves as those of Pityites Solmsi^ and recent

� Ibid. (97) p. 62. nbsp;nbsp;nbsp;� See page 374; also Nathorst (97) PI. v. figs. 1-

and the presence of a fine transverse wrinkling on the lamina a characteristic feature. Pityophyllum, if employed for both thenbsp;i^arrower and broader forms, includes specimens which in all pro-^a-bility belong to Conifers of more than oa� family: some arenbsp;^^rtainly Abietineous, but the flatter and broader forms bear anbsp;closer resemblance to leaves of some species oi.Podocarpus, Cepha-^^axus or Toneya. Nathorst, who instituted the generic namenbsp;^^yophyllum, recognises that many of the specimens so namednbsp;ave no real botanical value. Detached leaves of the type includednbsp;this comprehensive genus are of little interest; but it is note-J^orthy that such species as P. Nordenski�ldi (Heer) and similarnbsp;cnns are characteristic fossils in Jurassic and Cretaceous strata.

Under this name Nathorst^ includes leaves described by Heer Cretaceous strata in Greenland as Pinus Quenstedti andnbsp;Peterseni, also specimens from Upper Jurassic beds in Spitz-cigen. The leaves reach a length of at least 8 cm. and are 1�2 mm.

^d; the lamina tapers gradually towards the base and is more rib narrowed in the apical region; there is a prominent mid-I one side and sometimes indications of two finer marginalnbsp;^�^ds, also other longitudinal striations which may mark thenbsp;position of rows of stomata. These leaves are broader than thosenbsp;Pityites Solmsi and narrower than very similar specimensnbsp;^ cribed by Heer, Nathorst, and other authors as Pinus ornbsp;^^y^hyllum Staratschini^ from Cretaceous and Jurassic rocks.nbsp;quot;^^y^piiyllum Nordenshi�ldi^ (Heer) from rocks of the same age isnbsp;Mother similar form having a tendency to a slightly sickle-shapednbsp;o transversely wrinkled lamina {cf. fig. 776). Pityofhyllum isnbsp;represented in Jurassic Floras^; the specimens are,nbsp;tonbsp;nbsp;nbsp;nbsp;''^ery little interest to the botanist as it is impossible

'rame is used in preference to Feistmaiitel�s genus Pino-� recently resuscitated by Dr Marie Stopes, on the ground

4 J, � PP- 41, 68, Pis. V., ri. 3 Heer (78) ii. PI. n.; Nathorst (97) p. 18. H. irr Tnbsp;nbsp;nbsp;nbsp;see Moller (03) p. 39; Seward (11) p. �3; Krystofovio (10)

that Nathorst�s term is more appropriate for specimens which do not aSord evidence of closer affinity to Finns than to other generanbsp;of the Abietineae. In cases where the specimens may reasonablynbsp;be regarded as mor� nearly allied to Finns than to any other genusnbsp;the designation Finites may be added.

There can be no question of the abundance of Abietineous Conifers in Tertiary floras and it is equally true that cones of thenbsp;Fityostrohus type are widely spread in Lower Cretaceous stratanbsp;especially in Europe. The evidence furnished by cones clearlynbsp;points to the existence in Upper Jurassic floras of Conifers closelynbsp;resembling in the general form of their strobili recent membersnbsp;of the Abietineae. The wide distribution of cone-scales and conesnbsp;of the Araucarian type in Middle Jurassic floras is in striking contrast to the scarcity of cones of the Abietineous form in rocks oldernbsp;than the uppermost Jurassic and Lower Cretaceous series.

Carruthers*^ speaks of this Kimeridge cone from Dorsetshire as the oldest example of a Pine-cone. It is represented by a singlenbsp;imperfectly preserved specimen, 2x2 cm., of globular form withnbsp;partially destroyed broad and thin cone-scales: though it superficially resembles some recent Abietineous cones there is scarcelynbsp;enough evidence to warrant its inclusion in the Abietineae. Thenbsp;cone was first described as Finites depressus but owing to thenbsp;previous use of that specific name by Coemans it was re-namednbsp;F. dejectus'^.

A species, described as Finites strohiformis^, from Portlandian rocks near Boulogne founded on a single incomplete cone similarnbsp;in form and in the possession of apparently flat, imbricate, scalesnbsp;to Finns excelsa (cf. fig. 704, p. 154). The surface-features are notnbsp;shown on the weathered specimen and there is no definite information with regard to the number or position of the seeds, but as thenbsp;authors of the species state the narrow elongate and slightly curvednbsp;form of the cone, which was probably about 17 cm. long, affords anbsp;valid reason for comparison with recent Pines.

The type-specimen of this species^, from the Portlandian of Boulogne, is an ovoid cone 4-5 cm. long characterised by distallynbsp;expanded scales and resembling the small cones of Pinus Laricio.nbsp;In the absence of further data precise identification is not possiblenbsp;though the fossil is probably correctly regarded as an Abietineousnbsp;cone of the Pinus Pinaster type. Having regard to the fact thatnbsp;this and the preceding species the determination is based solelj^nbsp;cn external form no very definite statement is admissible as tonbsp;systematic position, but such evidence as there is favours the viewnbsp;that in these two cones we have Jurassic representatives of twonbsp;sections of the genus Pinus.

Several detached cones bearing imbricate scales, broad and flat 0 those of Picea, some species of Abies in which the ovuliferousnbsp;scales are longer than the bract-scales, and certain species of Pinusnbsp;s-racterised by flat scales instead of the woody scales of the Pinusnbsp;^Vestris and P. Pinaster type have been described from Britishnbsp;ealden strata as also from other countries. Gardner^ institutednbsp;� following species; Pinites Carruthersi, P. valdensis, P. cylin--P. fottoniensis^, but an examination of the type-specimensnbsp;^ ows that the distinctive features are not sufficiently well markednbsp;Warrant so many specific names. The Lower Greensand specimennbsp;OBi Potton, P. cylindraides, is water-worn and the shape of thenbsp;l�iperfect scales is not the original form; it may possibly be identicalnbsp;P. valdensis, P. Carruthersi, and P. pottoniensis, and there arenbsp;important features in which these forms differ from the longernbsp;of^W �^nbsp;nbsp;nbsp;nbsp;Dunkeri. The cones from Brook in the Isle

^ ight named by Carruthers* Pinites Dunkeri were originally ascribed by Mantell as Abietites Dunkeri^-, they reach a length ofnbsp;relaf ^nbsp;nbsp;nbsp;nbsp;have a breadth of 3 cm., they are elongate-oval and

a ively narrow and the long scales are attached to a slender axis ing. 778). The seeds, apparently two on each scale, are oval andnbsp;pressed. Cones of similar form and length are described by

^slenovskyi from Lower Cretaceous rocks in Bohemia as Pinus ^issima, a species recently recorded by Dr Stopes^ from thenbsp;ower Greensand of England.

Though in the absence of foliage-shoots cones of this type cannot assigned with certainty to any one recent genus, their great lengthnbsp;^'^ggests comparison with those of Pinus Lambertiana and P. excelsanbsp;*�ather than with cones of recent species of Picea.

in species was first described by Carruthers� from a specimen ^ tbe Leckenby Collection, Cambridge, from the Lower Greensandnbsp;Isle of Wight. It is 10 cm. long and 5 cm. in diameter; the

PiCi. 779, Pifyostrobus Leckenbyi. From the Lower Greensand of the Isle of Wight. (After Carruthers; � nat. size.)

scales agree in external form with those of Cedrus and Dr Slopes recently proposed the generic name Cedrostrohus^ m order tonbsp;^�^Pbasise this resemblance. Prof. Fliche� described a cone iromnbsp;Argonne as Cedrus ohlonga which he believed to be identicalnbsp;specifically with Abies ohlonga of Lindley and Hutton,

Though superficially very like a cone of Cedrus (fig. 779), Pityo-strobus Leckenhyi also strongly resembles some species of Picea and Abies in which the bract-scales do not project beyond the seminiferous scales. I have adopted the non-committal term Pityostrobusnbsp;as it is by no means certain that Carruthers� type is more closelynbsp;allied to Cedrus than to Abies.

Dr Stopes includes in Cedrostrobus a second species, Cedrostrobus MantellP, from the Lower Greensand of Kent which Carruthersnbsp;originally named Pinites: she compares with it a Potomac conenbsp;described by Berry^ as Cedrus Leei. But these species do not affordnbsp;any proof of close relationship to the recent genus Cedrus. It isnbsp;probable that some of the numerous cones found in Lower Cretaceous rocks belong to trees having the characters of Cedrus,nbsp;though in the absence of more decisive evidence than has so farnbsp;been furnished it would seem preferable to retain the wider desig'nbsp;nation Pityostrobus.

The small oval cone on which MantelD founded the species .4bies Benstedi is from the Lower Greensand of Kent. It was subsequently

described by Carruthers'* as Pinites and regarded by him as prob' ably more nearly allied to Cedrus than to Pinus. Dr Stopes� ha�nbsp;recently made a further examination of the structure of this typ�

and believes it to be more closely related to Abies. The cone-scales, as seen in fig. 780, show their double nature, and on the seminiferous scale are two ovules provided wth wings, w. The ovulesnbsp;are immature and there is a small space in the middle of the endosperm (fig. 780, e).

This Aptian (Lower Greensand) species, originally named by Mantell Zamia sussexiensis^ and afterwards referred by Carrut ers^nbsp;fo Pinites, has recently been more fully described by Dr fetopesnbsp;rinder Pinostrobus. The cone, 14 cm. long and nearly 5 cm. innbsp;diameter, bears overlapping scales with a thickened, curve , istanbsp;�largin 2 cm. broad and, on the exposed surface of the specimennbsp;cm. deep. The seeds, two on each scale, have a corrugatednbsp;stone-layer in the testa and bear massive, broad wings. A sectionnbsp;through the middle of a scale shows an irregularly scattered doiAlenbsp;set of variously orientated vascular bundles and resin-canals. 1 enbsp;species closely resembles Pinus excelsa and P. Strobus and is considered by Dr Stopes to occupy a position between these two types.

The inference to be drawn from this and several other cones from Lower Cretaceous strata is that Abietineous cones havingnbsp;raore or less flat scales as seen on the surface were more ^^)rn antnbsp;ill Europe in the early Cretaceous forests than those in w ic t enbsp;distal ends of the scales are rhomboidal as in Pirns Pinaster.

Hoeppert assigns the species determine the positionbsp;among the Abietineaenbsp;^ Mantell (43) p. 34.

** Stopes (15) p. 123, Pis.

^ tViUiamson (86).

� nbsp;nbsp;nbsp;----A

�umahly from Lower Greensand rocks, was described by m ey and Hutton as AUes oblonga*: it was assigned by some authors tonbsp;P^nites. Williamsons gave an account of a cone from Sidmouth innbsp;^voiishire, which he referred to Pinites oblongns, though Dr Stopesnbsp;^xpresses a doubt as to the identity of his specimen with thatnbsp;described by Lindlev. Schimper� employs the name Gedrus andnbsp;quot;oeppert assigns the species to Abietites^ It is impossible tonbsp;the position of the specimen represented m fig. 781

2 Carruthers (66*�) p. 541, PI- sx. figs. S, 6. 4 Lindley and Hutton (35) A. PI. 137.

The type-specimen bears a resemblance to P. Leckenbyi (fig. 779); the scales are broad and thin at the distal end and the axis isnbsp;relatively slender. The French specimens from Lower Cretaceousnbsp;rocks referred by Fliche to this species as Cedrus oblonga are considered by Dr Stopes to be specifically identical with Pityostrobusnbsp;Leckenbyi.

A large cone 15 cm. long and 4 cm. in diameter composed of stout woody scales with hexagonal apophyses was described bynbsp;Carruthers as Pinites hexagonus from the Gault of the South ofnbsp;England^; it agrees externally with recent cones of the Pinasternbsp;type but the distal ends of the scales are almost flat and nothingnbsp;is known of the internal structure. The species may be comparednbsp;with the Lower Cretaceous species P. Quenstedti Heer^.

The cones of this species (fig. 782), the commonest type in the Lower Cretaceous rocks of Hainault�, are 10�14 cm. long and

2�25 cm. in diameter. The cone-scales are compared with those of Pinus excelsa, but the distal ends are stouternbsp;than in the recent species and more like thosenbsp;P. Pinaster. Heer^ compares P. Andraeinbsp;''^ith his Pimis Quenstedti from Moravia innbsp;which the scales have thick apophyses with anbsp;central umbo. The needles of the Moraviannbsp;species are 20 cm. long and appear to be eithernbsp;3 or 5 in a fascicle.

It is impossible within the limits of a general text-book to discuss the bearings of the numerous Tertiary records of Abietineous cones,

^any of them undoubtedly borne by species r*! Pinus. A few examples only are mentionednbsp;Prirnarily in order to draw the attention ofnbsp;students to the importance of making a criticalnbsp;examination of Tertiary and Pleistocene Conifers. The neglect of Tertiary plants is largelynbsp;lt;iue to the unscientific treatment by authors ofnbsp;detached leaves of Angiosperms which in manynbsp;instances are referred to recent genera on whollynbsp;inadequate grounds, but the more trustworthynbsp;nature of the material on which species of Fig. 782. Pityostrobusnbsp;-Abietineous cones are founded deserves careful (Pmites) Andraei.nbsp;Consideration and would probably yield resultsnbsp;nf considerable iniportance.

This species, founded on a cone 12 cm. long and 6 cm. in diameter, was in the first instance described by Lindley andnbsp;flutton^ from an account furnished by Prof. Henslow and namednbsp;macrocefhala', it was found near Dover and believed to benbsp;^erived from the � Greensand formation.� A second specimen fromnbsp;^versham in Kent was described by the same authors as Zaniianbsp;^vata^. Endlicher^ assigned the cones to Zamiostrobus and MiqueD

^ Heer (69) p. 13^ pi, u. fig. u, nbsp;nbsp;nbsp;2 Lindley and Hutton (35) A. PI. cxxv.

proposed the name Z. Henslowi: their Abietineous nature was first recognised by Corda^, and Carruthers^ subsequently gave somenbsp;account of the internal structure and employed the generic namenbsp;Pinites. The discovery of additional specimens in situ enablednbsp;Carruthers to assign P. macrocefhalus to Eocene beds at thenbsp;junction of the Woolwich and Thanet beds with the London Clay.nbsp;Both Carruthers and Gardner� retain both specific names, but annbsp;examination of the specimens convinces me that there are nonbsp;differences worthy of specific recognition. The following briefnbsp;account is based on an examination of sections in the Britishnbsp;Museum and in part on notes supplied by Mr Dutt of Queens�nbsp;College, Cambridge, who is preparing a fuller account of thenbsp;material^. The cones are ovoid-cylindrical and obtuse; thenbsp;weathered surface (fig. 783) shows slightly convex polygonal areasnbsp;without any trace of a central umbo. The axis is slender in comparison with that of most recent species of Finns ] the stele includesnbsp;a fairly large pith of thick-walled cells surrounded by a vascularnbsp;cylinder in which foliar gaps are formed by the exit of the doublenbsp;sporophyll-traces. It is noteworthy that no resin-canals occur innbsp;the xylem. A ring of large resin-canals lined with thin-wallednbsp;epithelial cells occurs outside the phloem. The cone-scales arenbsp;given off almost at right-angles and then bend sharply upwardsnbsp;and become slightly broader near the surface of the cone (fig.nbsp;784, B). In one section a portion of a subtending bract-scale wasnbsp;recognised. The seminiferous scales are composed of thick-wallednbsp;cells and contain idioblasts like those in Araucarian leaves, alsonbsp;resin-canals; two ovules occur in a depression near the base of thenbsp;scales. The sporophyII-trace divides in the scale into several bundles,nbsp;and in places there are indications of a second series of inverselynbsp;orientated strands. The comparatively large ovules, nearly 1 cm.nbsp;long, are attached by a short stalk, and in places the remains of anbsp;wing can be seen. Although the integument is thick and lignifiednbsp;and the micropyle closed there are no embryos and no indicationnbsp;of archegonia in the partially preserved nucellar tissue. In the

784 p � nbsp;nbsp;nbsp;�

^ ' ^^yo^lrobus [Pmiies]'inacrocephalits. A, section of an ovule; nucellus.

gt; nbsp;nbsp;nbsp;section of part of the cone. (From sections in the British

^seu�n- B, slightly reduced.)

tent-pole of Ginkgo ovules: at its blunt apex are two �e pollen-grains. Prothallus-tissue is also represented.

�^�utheis compared tbe species with Finns Pinaster but the

surface-features are more like those of the cones of Finns excelsa {cf. fig. 704^ p. 154) and similar types.

This species was founded by Bailyi on part of a cone from the plant-beds in the basalts of Antrim and described, in greater detail

and from better material from the same locality, by Gardner^. The cones are 7�9 cm. long and 2�3 cm. in diameter, characterised bynbsp;woody scales with sub-hexagonal apophyses with a central umbonbsp;and a rounded upper margin; there are two seeds with long andnbsp;narrow wings on each scale. In one case (fig. 785) three cones arenbsp;attached in an erect position to a branch covered with persistentnbsp;leaf-bases. The foliage-leaves were borne in pairs and reached anbsp;length of 10�15 cm. A similar type with shorter needles from thenbsp;^ Baily (69) PI. xv. fig. 1.nbsp;nbsp;nbsp;nbsp;^ Gardner (86) p. 69, Pis. xv.�xviii.

��nie locality is named by Gardner Finns BailyF. These Irish specimens agree in the cones and foliage-spurs with such recentnbsp;Species as Finns halepensis and P. Finaster, but the apparentlynbsp;�rect position of the cones of the fossil type is a distinctive feature.

^ Tertiary cones similar to Pityostrobus Plutonis are illustrated y Finns rohustifolia Sap.^ from Provence, P. Kotschyeana (Ung.),nbsp;^iginally described by Unger and recorded by Tuzson� fromnbsp;ungary, P. transsylvanica Pax^, a North American form, whichnbsp;author of the species compares \Yith Finns Balfouriana, andnbsp;� V^ae-montana described by Mogan� from Lower Austria.nbsp;^'^tyostrobus (Pinites) palaeostrobiis (Ettingshausen).

This type originally described by Ettingshausen� from Haring ^0 the Tyrol is recorded from many Tertiary localities. The conesnbsp;ovate sub-cylindrical with scales of the Finns Strobns form andnbsp;Needles are borne in fascicles of five. Heer� refers to this speciesnbsp;thin and long needles from the Miocene of Greenland, but it isnbsp;olear that the needles are in fives. The species is recorded fromnbsp;Germany, France, and elsewhere. A cone of siivularnbsp;^ IS described by Unger� from the Oligocene of Kumi as Finnsnbsp;^^bdlopis and it is associated with quinary fascicles.

This species described by Heer^� as Finns (Abies) MacClurii Jom Miocene beds in Banks Land, lat. 74� 27' N. is represented bynbsp;^^Tociinen in the Dublin Museum. The narrow oval cone, 6x1-5nbsp;� ��nbsp;nbsp;nbsp;nbsp;of imbricate scales with the upper margin rounded or

S'^iarly truncate: some of the scales show indications of a pair Though similar to cones of Picea, the fossil cannot benbsp;assigned to any recent genus. Similarly, Miocene speci-�teepi^^nbsp;nbsp;nbsp;nbsp;scales, and leaves from Spitzbergen referred by

their generic position.

� Gardner (86) p. 73.

^ Tuzson (09^) p. 240, Pis. xiv., xv.

Mogan (03) figs. 1�3.

� Heer (83) PI. lxx. fig. 8; PI. Lxxxvii. figs. 5,6. ^ Unger (67) PI. xvi.

� Heer (68) i. p. 134, PI. xx. figs. 16�16-Ibid. (71) iii. p. 4.x^ pi. y. figs. 35�49.

While most of the Tertiary species of Pityostrobus agree closely with recent types some exhibit more or less striking peculiarities.nbsp;A species described by Engelhardt and Kinkelin as Pinus TimlerPnbsp;from Pliocene beds near Frankfurt is founded on pieces of largenbsp;cones characterised by cone-scales with a conical distal end havingnbsp;3 to 5 flat surfaces (fig. 786, A). The authors compare it withnbsp;Pinus Gerardiana from Afghanistan.

The material obtained from Pleistocene beds is often well enough preserved to afford trustworthy data with regard to the later geological history of different genera. Clement Reid^ identified a smallnbsp;cone from the Cromer Forest bed as Pinus silvestris, and from thisnbsp;horizon in Sweden Nathorst� records the same species; it is recordednbsp;also from Pleistocene deposits associated with Elefhas frimigeniusnbsp;in France and similar evidence has been obtained from Switzerland, Germany, Denmark, and other regions. At a later date thenbsp;former range of Pinus silvestris and other types is illustrated by the

evidence of submerged forests and, as we ascend the scale, tbe records become more legible and tbe prehistoricnbsp;�eerges into the historic era.

A cone apparently identical with the Spruce Fir (Picea excelsa) found in the Preglacial beds on the Norfolk^ coast (fig. 7S7)

Is a relic of the flora which existed in England when the Rhine after receivingnbsp;�many large tributaries�now separatenbsp;rivers-�seems to have flowed across thenbsp;present bed of the North Sea. The samenbsp;species is recorded from Pliocene beds onnbsp;l^be Dutch-Prussian frontier, also from thenbsp;Valleys of the Main and Neckar, the speci-mens from the latter locality being referrednbsp;by Gl�ck^ to Picea excelsa var. alfestris.

^exnander� has discussed the past history *^1 Picea in Scandinavia and quotes recordsnbsp;the occurrence of the genus in othernbsp;parts of Europe. Similar instances of thenbsp;^ider range of Abietineous genera are given

described Pleistocene plants in nbsp;nbsp;nbsp;d;;iey7 Norfolk. (After

merica. Prom the facts at present aval nbsp;nbsp;nbsp;nat. size.)

^ le it Would seem that Pinus and allied

��rrera were more abundantly represented in the ertiary an

Tertiary floras in Europe than in American strata of the same ag .

I*ITYaNTHUS. Nathorst. nbsp;nbsp;nbsp;^

Pityanthus granulatus (Heer). This species, descrited by

the Patoot (Cretaceous) beds in Greenland as 0^^^oglosslm S'ranulafuw and afterwards described by Newberry roni enbsp;^�iboy clays, has recently been identified by Dr Stopes as a longnbsp;�iicrostrobilus of some Abietineous Conifer, probably a wns

C and E. M. (08) H. xv. fig. 147 nbsp;nbsp;nbsp;^(14) for excellent

Sernander (93). See also Andersson (10) and W. B. g

TTi��^rieg of Pleistocene history. nbsp;nbsp;nbsp;/oot-pinbsp;nbsp;nbsp;nbsp;S 9

Bcrrv^07)� /ia3\ td v. ii nbsp;nbsp;nbsp;^ Heer(83)PI. laii. ^s.nbsp;nbsp;nbsp;nbsp;.

iU7); (10� ; Penhallow (04). nbsp;nbsp;nbsp;4 , , ^ v 9

�^ewberry and Hollick (95) PI. xx. figs. 11-13. nbsp;nbsp;nbsp;' Stopes (11 ) text-figs. 1,

This author examined the American specimen, which she regards as a fertile shoot of a three-needled Pine: the strobilus is 35 mm-long and from it winged pollen-grains were isolated. In its unusualnbsp;length the strobilus resembles the male flowers of Finns australisnbsp;from Florida.

The few specimens chosen for description afford examples of some of the oldest records of fossils, agreeing in the form of the wingnbsp;with recent Abietineous seeds and, as far as I know, none have beennbsp;discovered in strata below the Rhaetic. From Tertiary rocksnbsp;numerous winged seeds are recorded, but these are of no specialnbsp;interest and they are usually accompanied with foliage-shoots,nbsp;cones, or other fossils which afford more trustworthy data as tonbsp;relationship.

Nathorst described several winged seeds from the Rhaetic beds of Scania as Finns LundgrenF] they are 9�11 mm. long and 4 mni.nbsp;broad, the actual seeds being 3�4 mm. innbsp;length. Two examples from Stabbarp in thenbsp;Stockholm Museum are represented in fig.

788. To the same species Nathorst referred some imperfect cylindrical cones bearingnbsp;thin imbricate scales and reaching a lengthnbsp;of 3�5 cm. and a diameter of 1'2�2 cm.;nbsp;he also suggested the possibility that somenbsp;short shoots and long needle-like leaves described as Schizolepis Follini Nath, maynbsp;belong to the plant which bore the cones and seeds. In a laternbsp;account of Schizolepis^ he expressed the opinion that in thenbsp;absence of any proof of actual connexion the leaves and shortnbsp;shoots should be separated from Schizolepis and included iirnbsp;Pinites. These leaf-fascicles are described under the name Pityo'nbsp;phyllum and the seeds, which occur as separate fossils, are alonenbsp;included in Pityospermum Lundgreni. The striking resemblancenbsp;of the seeds to those from Franz Josef Land (fig. 789) and recent

1 nbsp;nbsp;nbsp;Nathorst (78) B, p. 31, PI. xiv. figs. 9 a, 1.3�17; PI. xv. figs. I�2.

^feietineous seeds is a valid reason for suggesting the inclusion the Rhaetic specimens in the Abietineae, though it would benbsp;too far to conclude that the seeds were borne on conesnbsp;8�nerically identical with or even closely related to those of anynbsp;fisting representative of the family. A Pliocene seed figured bynbsp;�gelhardt and Kinkelin^ as Finns Timleri (fig. 786, B) bears anbsp;ose resemblance in the form of the wing to some of the Rhaetic

. species, also from the Rhaetic flora of Scania^, is character-by the much longer wdng (2-7 cm.) which in size and form nets much more widely than Pityospermum Lundgreni, P. Nan-and other Jurassic types from the wings of any recent seeds.

i�lie seed shown m u^. nbsp;nbsp;nbsp;,

specimen in the Museum of the Survey (Jermyn Street) collected bynbsp;Kioettlitz in Rranz Josef Land annbsp;Lpper Jurassic or Wealden age. t isnbsp;other seeds are figured by New on

TealP. Tbe name Pityospermum Nansem

applied by Nathorst^ to similar speci �^ens obtained by Dr Nansen from t e samnbsp;��egion. Other winged seeds fromnbsp;'fosef Land closely resemble Heer s s]^cinbsp;Finns Maahiana^ from Jurassic roc snbsp;Siberia. A seed, 1-2 cm. long, from Meal-^nbsp;ben beds in the South of England is repronbsp;buced in fig. 789, B�; this is possibly anbsp;bistinct species, but the specific determina-^^

Fig. 789. A, Pityospennum Nanseni. B, Pityosper-mum sp. (A, drawn fromnbsp;a specimen in the Museumnbsp;of the Geological Surveynbsp;figured by Newton andnbsp;Teall; B, from a speci-- men, V. 232.3, in thenbsp;British Museum fromnbsp;Wealden rocks.)

� Newton and Teall (97) PI- xxxvni-Nathorst (99) p. 18, PI- n- nbsp;nbsp;nbsp;� 5 gg^vard (95) A, p- 19�-

and seeds of recent Pines and other Abietineae. It is not possible to determine the precise generic affinity of seeds of this type, butnbsp;their practical identity with recent Abietineous seeds warrantsnbsp;their reference to that family.

This generic name has been adopted^ for spores, provided with bladder-like extensions of the exine, agreeing in size and form withnbsp;those of recent Abietineous genera. Winged pollen occur also innbsp;the Podocarpineae, but the fossil examples so far recorded are muchnbsp;more like the microspores of Abietineous genera than those ofnbsp;Podocarpus, Dacrydium and Microcachrys.

In the course of examining sections of wood collected by Mr Priestley on the Priestley Glacier (approximately lat. 74� S.) Inbsp;noticed two small microspores in the siliceous matrix of the partiallynbsp;decayed stem^: one is shown in fig. 491, G (Vol. in. p. 298); thenbsp;longest axis is 80 p and the central part bears two bladders characterised by a fine surface-reticulation similar to that on recent.spores-A microspore of Pinus silvestris has a length of 75p. It is verynbsp;unlikely that the spores have any connexion with the stem in whichnbsp;they are preserved; they bear a much closer resemblance to thenbsp;microspores of Abietineous genera than to the spores of the Podocarpineae: the probability is that the Antarctic specimens belongnbsp;to some Abietineous Conifer though this cannot be definitely stated.nbsp;It is probable that the upper part of the Beacon Sandstone, fromnbsp;which the boulder containing the fossil is believed to have beennbsp;derived, is not older than Lower Mesozoic, e.g. Rhaetic.nbsp;Pityosforites sp.

Among the spores found by Nathorst� in Liassic clay from H�r in Scania were several winged microspores, one of which is reproduced in fig. 790, C from a photograph kindly supplied by Prof-Nathorst. The length of the spore is about 100p. and in thenbsp;shape of the bladders it agrees closely with the microspores ofnbsp;Picea excelsa^.

In his account of petrified plant-remains from Franz Josef I'and, probably of Wealden or approximately Wealden age, Grafnbsp;^olnas-Laubach^ mentions the occurrence of well preserved pollennbsp;^ith bladders and figures a piece of a cone with flat scales similar

�^90- A, B, Pilyosporites sp. from Franz Josef Land. C, Pityosporites sp. from Scania. (A, B, from specimens in the Museum of the Geological Survey;nbsp;G after Nathorst.)

that of a Picea. The drawings reproduced in fig. 790, A, B ^are made from specimens found in sections of the materialnbsp;^^�Dained by Solms-Laubach in the Geological Survey collection.

e longest diameter is 70�90p; in form and size the microspores y^�enible those of recent Pines. The apparently broad wall shownnbsp;fig. 790, A is the result of the obliquity of the section.

^fAomolepis cynarocephala Saporta. Saporta^ instituted this �6neric name for some cones from the Oligocene plant-beds ofnbsp;rovence 8�10 cm. long, ovate-elliptical, and composed of spirallynbsp;nged coriaceous scales not thickened at the apex but prolongednbsp;and�'^'^ Ifi� imbricate broad portion into a long recurved, acuminatenbsp;thenbsp;nbsp;nbsp;nbsp;spinous process. No seeds have been found and

fhe^^ evidence as to internal structure, Saporta considers cones to belong to some extinct type and, as Zeiller� says, theynbsp;� probably Abietineous.

^ Solms-Laubach (04) p. IL ^ �pTonos, cut up; XeTTts, scale.nbsp;� Zeiller (00) B. p. 278.

Crossotolepis Perroti Fliche. This generic nanie^ was proposed for an imperfectly preserved cone^ from Oligocene beds nearnbsp;Embrun in the French Alps, of elongate-cylindrical form, 13-3 cm.nbsp;long and 3-5 cm. in diameter, characterised by the fimbriate edgenbsp;of the imbricate, highly inclined, scales, which bear two seeds.nbsp;The cone agrees closely with several recent Abietineous typesnbsp;especially with Picea Menziesii and other species of Picea, but isnbsp;distinguished by the deeply fimbriate upper margin of the thinnbsp;cone-scales; it is difficult to determine how far this feature is thenbsp;result of secondary causes: Fliche believes it to be an originalnbsp;character comparable with that which led Saporta to found thenbsp;genus Entornolepis for an Oligocene cone from Armissan in Provence.nbsp;It is not certain whether the seeds are winged. Fliche is no doubtnbsp;correct in his conclusion that Crossotolepis is an Abietineous conenbsp;closely allied to Picea and probably related to Entornolepis. Henbsp;refers the two Oligocene genera to the Abietineae; they differnbsp;from any recent forms in the greater dissection of the distal edgesnbsp;of the seed-bearing scales, which in this respect are comparablenbsp;with the more feebly lobed scales of the cones of Picea Engelmanninbsp;and other species. Our knowledge of both genera is, however,nbsp;meagre and all that can be said is that the type-specimens affordnbsp;some evidence of the former occurrence of some Tertiary Abietineousnbsp;Conifers distinguished by the distally dissected scales from anynbsp;recent types.

This genus was instituted^ for lignitic specimens of short shoots and leaves from Middle Cretaceous beds on Staten Island, N.Y.nbsp;characterised by the large and indefinite number of leaves bornenbsp;on a single short shoot, the presence of a basal sheath of scale-leaves, and by certain anatomical features, particularly the mesarcbnbsp;structure of the single leaf-bundle, the occurrence of a complexnbsp;system of transfusion tissue, and other features.

The short shoots (fig. 791, B), rather less than 1 cm. long, consist of a relatively broad axis bearing on the upper part nume-rous spirally disposed truncate portions of leaves, in some cases

by nbsp;nbsp;nbsp;on a single shoot. Scars of scale-leaves are represented

tbe ��iDic areas near the base. There is no evidence as to branches on which the foliage-spurs were pro-arenbsp;nbsp;nbsp;nbsp;leaves are polygonal in section (fig. 791, A, C); there

a.s iv. '^^^ginal resin-canals and a single median vascular bundle

sclerous tissue in the ground-tissue of the leaf and a complete absence of mesophyll with infolded walls like that of recent andnbsp;some fossil Pines {cf. fig. 774). The phloem is represented by anbsp;crescentic space in the leaf shown in fig. 791, A. The dark zonenbsp;surrounding the bundle consists of thick-walled and relativelynbsp;long transfusion-tracheids and external to these is a broader sheathnbsp;of short transfusion-tracheids, but there is no endodermal layernbsp;aiid no admixture of parenchyma with the tracheids. The xyleninbsp;is composed partly of centripetal and in part of centrifugal elements:nbsp;the centrifugal xylem forms an uninterrupted arc next the phloem,nbsp;and between this and the transfusion-tissue on the lower side ofnbsp;the bundle seen in fig. 791, A the centripetal xylem is representednbsp;by radial rows of tracheids separated by spaces. Spiral proto-xylem elements occur between the two groups of metaxylem. IInbsp;is interesting to find similar transfusion-tissue in some leaves ofnbsp;true Pines described by Jeffrey from the same beds, but theirnbsp;bundles are double and composed of centrifugal xylem only as mnbsp;modern species.

The pith of the axis of a Prepinus shoot contains nests of sclerous cells: the leaf-traces pass through the cortex as singl�nbsp;bundles, and the single ring of wood contains a row of resin-canalsnbsp;blocked by tyloses. The tracheids have uniseriate bordered pA�nbsp;which are often contiguous and separated by Sanio�s rims.

Sew.^ (fig. 772) from English Wealden beds and with shoots described by Fontaine^ from the Potomac series under Heer snbsp;generic name Leptostrohus.

This species� was founded on specimens of short shoots from the Lower Cretaceous clays of Gay Head, Martha�s Vineyard,nbsp;Massachusetts, which are considered to be closely allied to Pityii^^nbsp;Solmsi. The wood of the axis, representing a single year�s growth,nbsp;contains two series of resin-canals and, as in the type-species, th�nbsp;canals of the leaves are in continuity with those in the cortex onbsp;the shoot-axis, whereas in recent Pines the leaf-canals end blindly-

2 nbsp;nbsp;nbsp;Fontaine (89) B. p. 227, Pis. oi.�oiv. etc.nbsp;nbsp;nbsp;nbsp;� Jeffrey (10).

As the choice of the name Prepinus implies, Jeffrey regards l'he type of foliage-shoot represented by these species as the directnbsp;Ancestor of the leaf-spurs of recent Pines. The short shoots ofnbsp;^fepinus are smaller than those of existing species of Pinus, butnbsp;in the numerous and spirally arranged leaves they resemble thosenbsp;nf Oedrus and Larix. Anatomically the fossil leaves differ widelynbsp;^lorn any Abietineous types, and were it not for the occurrence ofnbsp;tine Pine needles in association with Prepinus, which to somenbsp;extent bridge the gap between Prepinus and Pinus, one might benbsp;^ceptical with regard to the close affinity of Prepinus to recentnbsp;ines. Jeffrey compares the structure of the leaf of P. statensisnbsp;^ith that of some leaves of Cordaites, but the agreement is probablynbsp;�nt So close as Dr Stopes� descriptioni, quoted by Jeffrey, suggests.nbsp;enrey2 regards the short shoots of Pinus ?ind other Abietineaenbsp;a primitive attribute of the Coniferous stock and as one of severalnbsp;Reasons for believing the Abietineae to be the oldest tribe ofnbsp;inters. Prof. Thomson� has recently discussed the value of thenbsp;^i^idence based on the short shoots of Pinus and Prepinus andnbsp;loonies to the conclusion that the foliage-spurs of Pinus arenbsp;^Pecialiged shoots and do not belong to the category of primitivenbsp;^ lifts. In Cedrus, Larix, and Pseudolarix the leaves are spirallynbsp;^posed on the short shoots, while in Pinus they are fewer andnbsp;^yclic. q'}jg frequent occurrence of more than the normal numbernbsp;innbsp;nbsp;nbsp;nbsp;foliage-spurs of Pinus has already been mentioned:

plants supernumerary foliage-leaves are not uncommon aft increase in the number of needles is also induced bynbsp;jnbsp;nbsp;nbsp;nbsp;. The spirally arranged scale-leaves below the whorled

lea\ ^ short shoot of Pinus are homologous with the scale-6s on ordinary branches, and on seedling Pines they are re-fke primordial leaves: transitional forms occur between P three forms of leaf. The persistent short shoots of Cedrus,nbsp;Piirn't'^^*^nbsp;nbsp;nbsp;nbsp;as also of Ginkgo, are regarded as the more

^ycli ^ftftdition as compared with the deciduous nature of the of n- ^*^^^�g�'shoots of Pinus. Thomson notes that short shootsnbsp;Con 1nbsp;nbsp;nbsp;nbsp;proliferate like those of Cedrus and Larix. He

this condition.� The fossil shoots from the Potomac series described as Leptostrobus longifolius differ in the larger number of the needles from modern Pines and resemble abnormal short shootsnbsp;of Finns excelsa^ produced by wounding. The short foliage-shootnbsp;of Prepinus furnishes a more completely known example of anbsp;branch bearing spirally disposed leaves. In view of the palaeontological evidence and of the facts obtained from a study of recentnbsp;Pines it would seem that, as Thomson holds, the present form ofnbsp;the Pine spur is the result of specialisation and not a primitivenbsp;feature.

CHAPTER XLTX.

The data on which to base any conclusions as to the antiquity former distribution of the genus Podocarpus or of Conifersnbsp;believed to be closely allied to recent Podocarps are unfortunatelynbsp;derived from records which in the majority of cases are far fromnbsp;�8,tisfactory and consist mainly of detached leaves. Velenovsky'^nbsp;defers some linear leaves from the Perucer series of Bohemia tonbsp;Podocarpus {P. cretacea), but they exhibit no distinctive characters.nbsp;Some of the numerous leaves described as species of Podocarpusnbsp;in all probability correctly regarded as Tertiary representativesnbsp;of the recent genus, but it is often impossible to state with anynbsp;^^onfidence that detached leaves should be referred to Podocarpus,nbsp;preferably to Podocarpites, rather than to a more comprehensivenbsp;�enus such as Taxites or Elatocladus. In spite of the fragmentarynbsp;^^'ture of the evidence and the fact that no undoubted example ofnbsp;Podocarpus fertile shoot has been discovered, an examination ofnbsp;file published records leads to the conclusion that in Tertiary floras,nbsp;Particularly in those of Eocene age, species closely allied to existingnbsp;odocarps were abundant in Europe, a conclusion that is especiallynbsp;roteresting from the point of view of the present geographicalnbsp;istribution of the Podocarpineae. A brief account of some of thenbsp;otter known examples of Tertiary species of Podocarpus, which arenbsp;transferred to the genus Podocarpites in accordance with the practicenbsp;^�Oally adopted in the case of fossil species, may serve to illustratenbsp;0 nature of the material and the wide range of the specimens.

. Guppy 2 in his very suggestive remarks on the present dis-ntion and means of dispersal of Podocarpus writes: Tf we ^��rgn a home in the high latitudes of the northern hemisphere tonbsp;^ g�nus that was well represented in Europe in the Tertiary period,nbsp;,.^�vement of migration southward would explain most of thenbsp;unities in the present distribution. The great vertical rangenbsp;�^'elenov.sky (85) B. PI. xn. figs. 5�11.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;^ Guppy (06) p. 302.

of some of the species leads us to attribute a corresponding power of adaptation to the genus in respect of widely different climates....nbsp;With such a capacity for adaptation, migration of the genus wouldnbsp;be rendered easy over the globe.� The geological history of thenbsp;Podocarpineae is unfortunately very fragmentary but such datanbsp;as are available lend support to the view that there was �a centrenbsp;of diffusion in the extreme north,� the present distribution of the

family being as in the case of the Araucarineae the result of migration from other parts of the world where the plants are no'f^nbsp;represented only in the floras of a bygone age.

This species, from Eocene beds in South Styria, was founded by Unger^ on detached sub-falcate leaves 3�12 cm. long (fig'nbsp;792, A) with a short petiole or a more or less sessile lamina-1 Unger (51) p. 158, PI. xxiii. figs. 11�16.

^^rdner^ has described examples of this species from the Middle �9,g8hot beds of Bournemouth, the Lower Bagshot of Alum Baynbsp;the Isle of Wight, and from other British localities. A goodnbsp;^Ripression from Bournemouth is reproduced in fig. 792, B: thenbsp;is 7 cm. long and 3 mm. broad, the apex is sharply pointednbsp;there is no petiole. Fig. 792, C shows a leaf of slightly differentnbsp;which may belong to the species. No reproductive organsnbsp;^^ve been found. Ettingshausen^ records specimens of this typenbsp;from Eocene beds at Haring in the Austrian Tyrol where thenbsp;^ecies is said to be abundant, from Leoben� in Styria, Bilin^ innbsp;ohemia, Sagor in Carinthia^^ and elsewhere. Heer� states thatnbsp;species is common in Sw^iss Eocene deposits and Engelhardt^nbsp;^scribes examples from Oligocene rocks in Bohemia. A similarnbsp;perhaps specifically identical type was figured by Bindley� fromnbsp;ocene plant-beds at Aix in Provence as Podocarpus macrophylla,nbsp;^ Saporta�, who figured additional specimens from the samenbsp;�oality^ proposed the name Podocarpus Lindleyana on the groundnbsp;^at Lindley�s designation implies identity with a recent species,nbsp;leaf referred to P. eocaenica is figured by Massalongo^� fromnbsp;ortiary beds in Italy.

^ Some vegetative shoots originally described by De la Harpe^^ Alum Bay as Cupressites elegans are referred by Gardner^� tonbsp;� quot;^carpus and this determination derives support from the varietynbsp;the foliage illustrated by his specimens; in some branches thenbsp;leaves are two-ranked while in others the leaves are spirallynbsp;posed and three-sided, a diversity met with in recent species.nbsp;� imperfectly preserved fragment reproduced in fig. 793 from anbsp;^^teful drawing by Miss Woodward of the actual specimen isnbsp;Siired by Gardner as a fertile branch bearing a single seed with

a fleshy base as in certain existing species, but the details are too indistinct to aflord any proof of affinity to Podo-carpus. Gardner also describes a globose wrinklednbsp;seed, 16 mm. in diameter, as ? Podocarpus argillae-londinensis'^ from the London Clay which bearsnbsp;a close resemblance to the seeds of Podocarpusnbsp;data.' The specimens from Eocene beds in thenbsp;Island of Mull described by Gardner^ as Podocarpus borealis, consisting of small falcate leavesnbsp;and seed-like bodies, are too imperfect to benbsp;determined with accuracy. Fig. 794, C, C' represents a type from Bournemouth described asnbsp;Podocarpus wcerto which differs from other speciesnbsp;in the absence of a definite midrib; the linear-lanceolate coriaceousnbsp;leaves, reaching a length of 3 cm., are decurrent and appear to havenbsp;several parallel veins, a feature characteristic of the section Nagetu -

Fig. 794. A, B, Sequoiites Tournali (see p. 353). C, C', Podocarpites incerta. (From specimens in the British Museum described by J. Starkie Gardner; A, B, V. 524�nbsp;C, V. 522; nat. size.)

it is, however, doubtful whether this species should be included lO a genus implying affinity to Podocarpus. Some leaves figured bynbsp;Schmalhausen� from Oligocene strata in Russia as species oinbsp;Podocarpus are too imperfect to afford any trustworthy evidencenbsp;as to the occurrence of the genus Podocarpites. Ettingshausen

1 Gardner (86) p. 52, PI. ix. figs. 35, 36. nbsp;nbsp;nbsp;^ Ibid. (87) A. PI. xiii. figs. 3�R'

^ Schmalhausen (83*) PI. xxxii. nbsp;nbsp;nbsp;^ Ettingshausen (86) PI. vin. figs. 25�

lt;iescribes as Podocarpus prae-cupressina foliage-shoots and a very imperfectly preserved seed from Eocene rocks in New South Wales:nbsp;the same author records Podocarpus ParkerP, a doubtful species,nbsp;from New Zealand.

Shoots bearing leaves similar to those of P. eocaenica are ifescribed by Gardner from the Eocene plant-beds in the basalts

Mull as Podocarpus Camphelli^ (fig. 795). The linear acuminate straight or slightly curved leaves with a contracted decurrent basenbsp;about 7 cm. long and possess a well-defined midrib. Gardnernbsp;^t^tes that the late Prof. Oliver regarded the fossils as the branchesnbsp;Podocarp; they are compared with Podocarpus falcata ofnbsp;^ � Cape and Tropical Africa and with the South African speciesnbsp;'^hunbergii.

The records of the rocks afford very little information with regard to the past history of Conifers allied to the recent genusnbsp;Dacrydium. The marked dimorphism of the foliage-shoots (fig. 708,nbsp;p. 160), their close resemblance to branches of some other Conifers,nbsp;as also to Lycopodiaceous plants and some of the larger Mosses,nbsp;are serious difficulties in the way of recognising representatives ofnbsp;this genus among impressions of vegetative branches. It is interesting to find that the most promising piece of evidence of thenbsp;occurrence of a fossil type [Stachyotaxus) allied to Dacrydium isnbsp;furnished by a Rhaetic flora, a fact pointing to a high antiquitynbsp;of the plan of reproductive shoot characteristic of existing species.

Schenk^ compares with Dacrydium some obscure and small fragments from the Coal Measures of China which he described asnbsp;Conchophyllum Richthofeni, but there are no substantial groundsnbsp;for such comparison. The specimens consist of pieces of slendernbsp;axes bearing spirally disposed bracts or small leaves showing at thenbsp;base of the ovate-oblong lamina a slight depression from which anbsp;seed may have fallen. The Lower Cretaceous foliage-shoots fromnbsp;Bohemia described by Velenovsky^ as Dacrydium, densifolium havenbsp;no claim to be accepted as branches of a Podocarpineous Conifer.nbsp;Ettingshausen� figures from Eocene beds in Australia and NeWnbsp;Zealand sterile twigs assigned respectively to Dacrydium cupressi-noides and D. prae-cupressmum: in neither case is there any evidence as to the nature of the reproductive organs, and the formnbsp;of the foliage-shoots might with equal probability be interpretednbsp;as evidence of other Conifers or of some Lycopodiaceous plant.

The genus Stachyotaxus*^ was instituted for some Rhaetic specimens from Scania originally named by Agardh Sargassurt^�nbsp;septentrionale and Caulerpa septentrionalis] the former was renamed by Nathorst Carpolithes septentrionalis and the latter

^ypctfissidium septentrionale. The discovery of additional material Nathorst to transfer some of the specimens to a new genusnbsp;^^achyotaxus. The foliage-shoots of Stachyotaxus elegans'^ arenbsp;�dimorphic; some of the leaves are appressed and imbricate as in

Stachyotaxus elegans. A, B, f nat. size; C, slightly enlarged, D, X 3i. (After Nathorst.)

sessile and decurrent, with a lamina quot;quot;alls ^ ^^�^gth of about 1 cm. The epidermal cells have,straightnbsp;Thenbsp;nbsp;nbsp;nbsp;stomata occur in two rows on the lower surface.

^�gastrobili (fig. 796, B) have the form of spikes about ^ Nathorst (08^) p. 11, Pis. ii., iii.

5�6 cm. long; the fairly stout axis bears sporophylls, approximately at right-angles, consisting of a short and relatively thick stalk expanded into a triangular scale bearing two ovate seeds,nbsp;3�3-5 mm. long, each being enclosed basally in a cupule (fig. 796,nbsp;C, D); the distal end. of each sporophyll forms an upturned acuminate apex. The cuticle of the seed-coat shows that the latternbsp;consisted of thick-walled cells, and within the testa Nathorst foundnbsp;the remains of a much more delicate membrane, possibly representing the nucellus. Nathorst compares the sporophylls withnbsp;those of recent Dacrydiums, though in Stachyotaxus there arenbsp;normally two seeds on each sporophyll and not one as in the recentnbsp;genus, a difference possibly of no great importance. In habitnbsp;the fertile shoots of the fossil type are comparable with those ofnbsp;Podocarpus spicata. Miss Gibbs ^ in her apcount of recent Podo-carpineae expresses agreement with Nathorst�s view that Stachyotaxus is probably a member of that family. Nathorst describes anbsp;second Swedish species but from a slightly lower horizon in thenbsp;Rhaetic series. This species, Stachyotaxus elegans, is characterisednbsp;by longer and stouter megastrobili reaching at least a length ofnbsp;12 cm. and by longer linear leaves 10 cm. long.

Hartz^ refers to Stachyotaxus septentrionalis some sterile shoots from Lower Jurassic, or Rhaetic, beds in East Greenland, andnbsp;Halle� draws attention to the superficial resemblance to thonbsp;Swedish type of some vegetative twigs from Graham Land whichnbsp;he refers to the genus Elatocladus.

The specimen on which this species is founded^ was obtained by Hugh Miller from Upper Jurassic beds on the North-East coastnbsp;of Scotland and inaccurately figured in the Testimony of the Rochs^-Fig. 797 is from a careful drawing by Mr T. A. Brock of the originalnbsp;specimen in the Edinburgh Museum. A slender axis bears nuniC'nbsp;rous spirally disposed oval bodies (6x5 mm.) which are no doubtnbsp;seeds: each shows a differentiation into an inner portion surrounded

lation is tbat the oval bodies are seeds in l�itiniate association with fertile bracts.

The strobilus bears a close resemblance ^0 Stachyotaxus elegans Nath.'- from thenbsp;^haetic of Sweden compared by the authornbsp;the species with an ovuliferous shootnbsp;Podocarpusnbsp;nbsp;nbsp;nbsp;spicatanbsp;nbsp;nbsp;nbsp;and Dacrydiuni

^�^anklini and believed to be allied to the decent genus Dacrydiuni, a view upheldnbsp;V Miss Gibbs^ in her account of recentnbsp;Todocarps.nbsp;nbsp;nbsp;nbsp;Itnbsp;nbsp;nbsp;nbsp;is notnbsp;nbsp;nbsp;nbsp;improbablenbsp;nbsp;nbsp;nbsp;that

^^rohilites Milleri is more closely allied to the Podocarpineae than to any other familynbsp;Conifers.

In his nbsp;nbsp;nbsp;account ofnbsp;nbsp;nbsp;nbsp;a Tertiarynbsp;nbsp;nbsp;nbsp;flora,

Fig. 797. Strohilites Milleri. conspicua. The lamina, rather lessnbsp;nbsp;nbsp;nbsp;(After Seward; Edin-

^ nbsp;nbsp;nbsp;2 cm. long, is linear-lanceolate with anbsp;nbsp;nbsp;nbsp;burgh Museum; nat.

�pex and a short stalk; no veins j � shown in the drawing. There is no substantial reason fornbsp;6'rcting this solitary fossil as a fragment of a Conifer alliednbsp;Dus�n admits the lack of satisfactory evidencenbsp;generic identity, but the specimen hardly merits thenbsp;lion of being made the type of a new genus.

possibly Oligocene, from localities on the quot;Is-gellan .straits Dus�n^ describes a singlenbsp;as Saxegoihopsis fuegiantis on thenbsp;ground that it resembles the leaves of Saxe-

This generic name was given by Heer* to a fossil, .subsequently transferred to a new genus Drepamlepis^ (fig. 798, C), which affordsnbsp;* Nathorst (08^) PI. ii. figs. 1�27.nbsp;nbsp;nbsp;nbsp;quot; tlibbs (12) p. 539.

quot; Heer (75) ii. p. 124, PI. xxxv. figs. 17�21. nbsp;nbsp;nbsp;quot; Nathorst (97) p. 43.

no real evidence of a relationship to the recent genus Phyllocladus, but Heer�s term may be retained for a species described bynbsp;Ettingshausen as Phyllocladus asplenioides^.

This Tertiary species from New South Wales presents a close resemblance to Phyllocladus (fig. 675, p. 107) and is probably annbsp;Eocene representative of the genus. The specimens consist of fairlynbsp;stout axes bearing cuneate and irregularly lobed leaf-like organs,nbsp;in some cases apparently subtended by small scales�a circumstance which justifies Ettingshausen�s suggestion that the lateralnbsp;members are phylloclades. In one case a seed occurs at the basenbsp;of a phylloclade. The phylloclades are practically identical withnbsp;some forms of Thinnfeldia, particularly with American examplesnbsp;referred by Berry to his genus Protophyllocladus; the laminae of thenbsp;Australian species agree both in form and venation with those ofnbsp;the Cretaceous American impressions, but in the latter there isnbsp;no good reason for interpreting the leaf-like organs as flattenednbsp;branches.

Ettingshausen�s species is the only fossil that has come under my notice that has any substantia] claim to be considered a satisfactory record of the recent genus Phyllocladus. In the accountnbsp;of Thinnfeldia in Volume ii^ reference is made to the resemblancenbsp;of some impressions included in that genus to the phylloclades ofnbsp;Phyllocladus, a resemblance which led Ettingshausen to assignnbsp;the type-species of Thinnfeldia to the Coniferae. Berry� considersnbsp;that Ettingshausen�s comparison with Phyllocladus, though notnbsp;applicable to Jurassic and other of the older species of Thinnfeldia,nbsp;is valid in respect of certain Middle and Upper Cretaceous formsnbsp;for which he instituted the genus Protophyllocladus. Attention hasnbsp;previously been called^ to the inadequacy of the evidence in supporfnbsp;of the conclusion implied by the adoption of, the name Protophyllocladus. The specimens for which this name was institutednbsp;consist of comparatively large coriaceous leaf-like impressions, linearnbsp;or ovate-lanceolate with an entire, undulate, or crenulate margin^nbsp;provided with a short petiole prolonged as a stout midrib fromnbsp;which numerous simple veins are given of! at an acute angle.

2 nbsp;nbsp;nbsp;Page 543.nbsp;nbsp;nbsp;nbsp;� Berry (03) B.nbsp;nbsp;nbsp;nbsp;^ Seward (04) B. p. 3l-

'^�nation and to a large extent in shape the fossils conform to the characters of Thinnfeldia.

This species was originally described by Lesquereux^ from I*aljota beds in Nebraska as Phyllocladus subintegrifolius. Heer^nbsp;'^6scribed similar or possibly identical specimens fiom^he Atanenbsp;beds of Greenland as Thinnfeldia Lesquereuxiana and includednbsp;Lesquereux�s name as a synonym: these Greenland specimensnbsp;reach a length of 8 cm. and a breadth of 2 cm. Heer classes thenbsp;species among plants of uncertain position and compares thenbsp;r*�pressions with the ph^/lloclades of PhyUanthus. The species isnbsp;recorded by Hollick� from Martha�s Vineyard and other localities,nbsp;y Berry 4 from the Raritan flora, and as Thinnfeldia Lesque-'^^'u.xiana by Newberry and Hollick� from the Amboy clays andnbsp;^fher Cretaceous floras�: most of the examples are detached leavesnbsp;''rr 1 phylloclades), linear, spathulate, or ovate with an entire ornbsp;toothed margin while a few are branched (fig. 798, A, B). In nonbsp;�^rigle case is there any evidence in favour of regarding the speci-as phylloclades rather than leaves. Until additional factsnbsp;^^6 obtained it would seem preferable either to retain the genericnbsp;�ia'Bae Thinnfeldia used by several authors or to adopt some titlenbsp;does not suggest a relationship to any recent genus. Zeiller�nbsp;described a specimen from the Great Oolite of Marquisenbsp;� Trance) as Protophyllocladus sp.; this is the first Europeannbsp;^�^cord for Berry�s genus.

V similar species is represented by Protophyllocladm polytnorpkus first described by Lesquereux� from Vancouver Island asnbsp;souria polymorpha and afterwards transferred by Knowlton�nbsp;l^^^'^'^'hfeldia. Another closely allied type is Protophyllocladusnbsp;g Berryi� from Upper Cretaceous rocks in South Carolina,nbsp;of these supposed phylloclades closely resemble flattened

-- nbsp;nbsp;nbsp;*�nbsp;nbsp;nbsp;nbsp;i 1.nbsp;nbsp;nbsp;nbsp;�

Hollick (06) p. 36, PL V. figs. 1�6.

Newberry and Hollick (95) p. 59, PL ' besquereux (91) PL ii. figs. 1�3.

' Lesquereux (78) B. p. 84, PL lx. figs. ^0, 41.

� Knowlton. (93) p. 47, PL v. figs. 1�4; Berry (03) nbsp;nbsp;nbsp;. p.

quot; Berry (03) B.; (14) p. 17, PL II. figs. 9�13. .

leaf-like branckes from Kreischerville for which Hollick and Jeffrey instituted the genus Androvettia^ (fig. 806, A�C): some examplesnbsp;of that genus are clearly distinct as there are small leaves borne

on the flat phylloclade, but other specimens are hardly distinguish' able from Protophyllocladus. Though it would be foolish to den/nbsp;that the marked resemblance as regards form and venation betweennbsp;certain species of Protophyllocladus or Thinnfeldia and Phyllocladusnbsp;may be significant, it is true that no satisfactory evidence has beennbsp;produced in support of actual affinity to the recent genus.

This name was applied to a single species, Palaeocladus cunei-gt;nts, from New South Wales^ founded on a specimen described fg ^ cuneiform phylloclade with a median rib from which anbsp;simple veins are given off at an acute angle, each passing upnbsp;� middle of a lateral tooth; the teeth or serrations are regardednbsp;lateral phylloclades concrescent with one another and with thenbsp;flattened axis. A few small scales occur at the base of thenbsp;phylloclade. Ettingshausen�s ingenious interpretation,nbsp;�ugh possibly correct, rests on too slender a basis to justify thenbsp;^^urnptioir of any affinity to Phyllocladus.

ms generic terra was proposed for some branched foliage-Potomac formation^ characterised by small broadly 0^.^ leaves with spreading veins. The type-species Phylloclad-a clonbsp;nbsp;nbsp;nbsp;(fig-nbsp;nbsp;nbsp;nbsp;as Fontaine and Berry state, hears

htt] ^��mhlance to some forms of Nayeiopsis and there can be no ^nbsp;nbsp;nbsp;nbsp;as to the foliar nature of the appendages, which afford

^hidus. The species must be left for the present as a plant P��dion; it would seem more appropriate to adoptnbsp;name Nayeiopsis than to make use of a designationnbsp;�amp;esting a, relationship which has not been established.

fig- resembling in habit those of Taxus, �nnbsp;nbsp;nbsp;nbsp;Virens, Cephalotaxus and some other recent Conifers,

is Jjjj sence of anatomical characters or reproductive organs it determine the precise position of shoots of thisnbsp;form and the designation Taxites, as generally employed,

cannot therefore be interpreted as indicative of affinity to Taxus. Taxites has also been applied to fossil wood believed to belong tonbsp;Conifers allied to Taxus or other members of the Taxeae, but thisnbsp;name has been superseded by Taxoxylon. Taxites is misleadingnbsp;in its implication of relationship as the specimens so named affordnbsp;no proof of their systematic position within the Coniferales. Halle^nbsp;has proposed to employ the generic name Elatocladus for sterilenbsp;shoots of the Taxites type, the latter name being restricted tonbsp;fossils which there are good grounds for assigning to the Taxeae.nbsp;Among other genera to which authors have referred vegetativenbsp;shoots superficially similar to Taxus is Palissya^, a genus foundednbsp;in part on a definite type of reproductive shoot very different fromnbsp;that of Taxus. The investigation of the epidermal structure ofnbsp;some specimens of Jurassic age placed by Feistmantel in PalissrjO'nbsp;has led Miss Holden to institute a new genus Retinosporites^ basednbsp;on characters recognisable only in the case of fossils which retaiDnbsp;their cuticular features. Another example of foliage-shoots of thenbsp;Taxites habit, which are shown by their fertile shoots to have nonbsp;connexion with the Taxeae, is Stachyotaxus^. Taxites should benbsp;retained for fertile branches or reproductive organs which there i�nbsp;some good reason for believing to be more closely related to Taxu^nbsp;than to any other genus. So far as I am aware, none of thenbsp;specimens described as species of Taxites supply sufficient justifica'nbsp;tion for being so designated. Fragments of sterile shoots fromnbsp;the Miocene beds of Western Greenland described by Heer� a�nbsp;Taxites Olriki and now in the Dublin Museum afford no proof ofnbsp;close affinity to Taxus: the leaves are rather larger than those ofnbsp;Sequoiites Langsdorfii and not decurrent.

The generic names Cephalotaxites and Cephalotaxopsis appb�*^ by Heer� and Fontaine� respectively to foliage-shoots resemblmgnbsp;those of recent species of Cephalotaxus suggest an affinity that i�nbsp;not supported by data based on reproductive shoots. While sucbnbsp;a designation as Cephalotaxopsis may be used without a defind�nbsp;implication of relationship to the recent genus, CephalotaxiR^�nbsp;like Torreyites and Taxites, is more appropriately restricted

specimens which there is good reason for connecting with Torreya Taxus respectively.

Among the numerous specimens of wood from Mesozoic and 6rtiary beds referred by authors to Taxoxylon^ there are very fewnbsp;^Iiich show true spiral bands on the secondary-xylem elements:

are no undoubted examples of the Taxineous type of stem Tom pre-Tertiary strata.

The foliage-shoots described as species of Taxites are as a rule '^^.lueless as records of Taxus, Cephalotaxus, or Torreya. Somenbsp;^Diall seeds very like those of Taxus baccafa are described bynbsp;dwig from the Oligocene lignites of West Germany as Taxusnbsp;'^'''ynritifera^ and similar examples were described by Heer� asnbsp;^^^Polithes nitens from the lignites of Bovey Tracey: the Boveynbsp;��ds have recently been named by Mr and Mrs Reid Taxus (?)

latter authors refer some globose mucronate seeds the preglacial deposits of Norfolk to Taxus baccata^, andnbsp;�^ent Reid� records the same species from preglacial beds innbsp;oik, from interglacial beds in Suffolk and from the peat belownbsp;�vel nr the Thames valley. The genus Taxus ranges throughnbsp;Call ^nbsp;nbsp;nbsp;nbsp;Asia, North America, Algeria, and occurs sporadi-

y on the mountains of Sumatra, Celebes and the Philippines; .^�^�^''orthy that there is little difference between the severalnbsp;is �''^hich are probably mere geographical forms�. The Yewnbsp;loc rnbsp;nbsp;nbsp;nbsp;parts of Sussex, Hampshire, and MTltshire, in a few

sur �^�^�cene records that Taxus was formerly much more widely Conwentz� has shown that many places in Germanynbsp;British Isles derive their names from the Yew, and thenbsp;Byi ..^'^dhor found that several prehistoric wooden articles in thenbsp;Museum are made from Yew wood.

� evidence on which several fossil Coniferous branches from eeous and Tertiary rocks have been referred to Torreya is in

most cases unconvincing and with one exception no facts as to epidermal characters are available. Leaves of recent species ofnbsp;Torreya (fig. 694, B, p. 141) are characterised by two well markednbsp;stomatal grooves on the lower surface, and another feature is thenbsp;absence of a prominent midrib: the leaves of Cephalotaxus, similarnbsp;in form and size to those of Torreya, differ in the flat ungroovednbsp;lower surface and the prominence of the midrib on the upper surfacenbsp;of the lamina. We have no information with regard to any fossdnbsp;seeds of the Torreya type, a type to which reference is made iDnbsp;the account of fossil Palaeozoic seeds. The present distribution ofnbsp;Torreya suggests that it was formerly more widely spread, butnbsp;the data at present available do not admit of any very satisfactorynbsp;statement of its past history.

Berry described this species as Tumion carolianum}, using tb� unfamiliar generic name which has been substituted by purists mnbsp;nomenclature for Torreya. The material from Middle CretaceoU^nbsp;rocks in North Carolina consists of twigs with spirally disposelt;inbsp;flat linear-lanceolate leaves 2-5�3 cm. long and with a maximuurnbsp;breadth of 3 mm., gradually tapering towards the slender apo^nbsp;and slightly contracted at the decurrent base. There is no distinctnbsp;midrib, but in the proximal part of the lamina a more opaq*!�nbsp;band indicates the position of the vascular tissue: on either sid�nbsp;of the middle line is a band in which the stomata are scattered gt;nbsp;the long axis of the guard-cells tends to be at right-angles tonbsp;length of the leaf as in recent species and the fossil stomata genet'nbsp;ally resemble those of existing types. Some less satisfactorynbsp;specimens from Upper Cretaceous beds in Georgia are doubtfultnbsp;referred, by Berry^ to this species. In view of the characters ^nbsp;the vegetative fragments from Carolina it seems reasonable f�nbsp;adopt the generic name Torreyites. Berry has published a lU^Pnbsp;showing the distribution of Cretaceous representatives of TorreH^�nbsp;but it is questionable whether the nature of the records constitut�^nbsp;a solid foundation.nbsp;nbsp;nbsp;nbsp;^

parvifoUa, from the Lower Cretaceous beds of Greenland: the ^�aves of the former exhibit a close agreement mth those of thenbsp;^�cent species though the evidence in support of generic identitynbsp;far from decisive. The second species differs in the smallernbsp;^�aves: an examination of one of the figured specimens in thenbsp;Stockholm Museum led me to the conclusion that it may be identicalnbsp;^^ith an Upper Jurassic form from Scotland described as Taxitesnbsp;; there are no adequate grounds for the use of the name �nbsp;'^orreya. Fontaine�s Potomac species Torreya virginica^ is foundednbsp;S� piece of shoot bearing long linear leaves with no obviousnbsp;^^idrib but with two strong lines between the middle and the edgesnbsp;the lamina which suggest stomatal grooves. This specimen isnbsp;less value than the type-specimen of Torreyites carolinianusnbsp;I'ecause of the lack of information with regard to the stomata.

^ second species of very little botanical value is described from the

same formation as Torreya falcata^. Yokoyama�s Upper lassie or Wealden species Torreya venusta*' from Japan has nonbsp;lift to be included among records of Torreya.

The Miocene species from Greenland, Torreya borealis, described - tteer� is founded on sterile twigs with broad linear leaves whichnbsp;g no definite indication of relationship to the recent genus,nbsp;^^^gi�ients from Pliocene beds near Lyon described by Saportanbsp;l^arion� as Torreya nucifera var. brevifolia show two stomatalnbsp;�^ooves and may be correctly identified. These authors regardnbsp;asnbsp;nbsp;nbsp;nbsp;described by Ettingshausen from Bilin in Bohemia

Sequoia Langsdorfii as fragments of a Torreya and re-name the hecies p, bilinica.

^letached leaves and seeds from Upper Pliocene beds in Main Valley (Frankfurt) are referred by Engelhardt andnbsp;IP 6 to Torreya nucifera fossilis: the seeds bear a closenbsp;�oodnbsp;nbsp;nbsp;nbsp;those of the existing species. While there is fairly

oc nbsp;nbsp;nbsp;from Tertiary localities of the comparatively recent

^ ence of Torreya in Europe the records cannot be regarded

�onclnsive.

133quot;' P- 6^^- nbsp;nbsp;nbsp;^ Fontaine (89) B. p. 234, PI. cix. fig. 8,

Vesquia tournaisii Bertrand. The name Vesquia, after the French botanist Julien Vesque, was given by Bertrand^ to seedsnbsp;from Lower Cretaceous strata at Tournai which he described asnbsp;intermediate in certain respects between Taxus and Torreya; theynbsp;are two or three times as large as the seeds of Taxus and- aboutnbsp;one-third the size of those of Torreya. The ligneous shell is ribbednbsp;and on each side at the base is a large orifice marking the positionnbsp;of a vascular bundle which is continued through the length of thenbsp;marginal ribs. The seeds are elliptical in transverse section andnbsp;prolonged apically into a micropylar beak agreeing structurallynbsp;with the micropyle of Taxus and Torreya. Bertrand also foundnbsp;anatomical features in the shell similar to those of the recentnbsp;genera. In the absence of illustrations it is difficult to folio�quot;'nbsp;the description in detail, but the facts appear to favour Bertrand�snbsp;conclusions with regard to the affinities of the fossil species.

Fontaine^ gave this name to specimens of vegetative shoots, abundant in the Patuxent formation in the Potomac group, closelynbsp;resembling in habit recent species of Cephalotaxus: the character�nbsp;of the genus have been revised by Berry� who adds some particular�nbsp;as to the structure of the epidermis. Fontaine�s four species arenbsp;reduced by Berry to two. No seeds have been found attachednbsp;to the branches, but the American authors consider that sonr�nbsp;associated seeds may belong to the genus.

With this species Berry* includes C. ramosa Font. The branche� are fairly robust and in some cases bear lateral shoots in whorls ornbsp;pseudo-whofls; the leaves are distichous, linear-lanceolate, rathernbsp;abruptly rounded at the base and tapering gradually to a mucronat�nbsp;apex, with an average length of 4�5 cm. and a breadth of 3�4nbsp;Groups of bud^scale scars occasionally occur at the base ofnbsp;ultimate shoot. The thick lamina may be transversely wrinklednbsp;as in Pityophyllum. There is a distinct midrib and a short distanc�

^ Bertrand (83). nbsp;nbsp;nbsp;^ Fontaine (89) B. p. 235.nbsp;nbsp;nbsp;nbsp;� Berry (11) p-

either side of it is a stomatal groove in which stomata are '^regularly scattered; the orientation of the guard-cells though

constant tends to be parallel to the long axis of the leaf. The epidermal cells are thick-walled and quadrangular or hexagonalnbsp;^-iid arranged in regular rows. Berry states that the stomatalnbsp;fetooves are a prominent feature and that there is some evidencenbsp;ef the occurrence of woolly hairs, characters suggestive of Torreyanbsp;father than Cephalotaxus. Branches from the Potomac bedsnbsp;quot;^ith shorter leaves are referred to Cephalotaxopsis hrevifoUa innbsp;^bich is included C. microphylla Font. Berry points out that thenbsp;photograph of C. hrevifoUa which he gives serves to �emphasizenbsp;� idealisation and inaccuracy of the former figures of this plant.�

Such evidence as is available favours the comparison of these ^ecies with recent Taxineae, but the structure of the leaves ofnbsp;� ^(ignifoUa, so far as it is indicated in the epidermal preparationsnbsp;^ cribed by Berry, would seem to be in favour of a closer affinitynbsp;� Torreya than to Cephalotaxus. The presence of depressednbsp;^OKiatal regions is a characteristic feature of Torreya and not of

X 13 mm.) in the form of a mould. Without examining the �e-specimen it would be rash to accept the determination asnbsp;correct. Berry^ has described some seeds from Mid-Cretaceousnbsp;�ocks in Carolina as Cephalotaxospermuni carolinianum, approxi-mately lo by 8 mm. in size, ovoid acuminate, and resembling thenbsp;seeds of Cephalotaxus, but no anatomical features are preserved,nbsp;�aporta�s species Cephalotaxus europaea^ founded on sterile shootsnbsp;from the Aquitanian of Manosque in the South of France affordsnbsp;�10 convincing evidence of generic identity with Cephalotaxus.

Seeds from Upper Pliocene beds in the Main valley are referred oy Engelhardt and Kinkelin^ to three species of Cephalotaxus, thenbsp;type-species being C.francofurtana. They bear a close resemblance

lo the recent seeds but this is hardly sufficient to prove their generic identity.

� Heer (83) p. 10, PI. LIII. fig. 12. nbsp;nbsp;nbsp;- Berry (10�).

Saporta (93) p. 42, PI. V. fig. 4.

* Engelhardt and Kinkelin (08) p. 194, PI- xxin. fig. H-

Heeri described a specimen from the Lower Cretaceous beds Qf ^^��ii^^iid as Cephalotaxites insignis consisting of a small piecenbsp;^^^^�^i^g6-8hoot with, apparently attached to it, an oval seed

CHAPTER L.

Trioole'pis Leclerei Zeiller. This generic name was proposed by Zeiller for a cone from the Rhaetic flora of Tonkin^ which he atnbsp;first placed in the comprehensive genus Conites. In generalnbsp;appearance the specimen resembles a cone of a Picea; it is elongate-oval and incomplete, more than 10 cm. long and about 3 cm. innbsp;diameter. The impression shows numerous spirally disposed,nbsp;imbricate, scales apparently thin, oval-linear in form, 12�15 mm.nbsp;long and 6�7 mm. broad, suddenly contracted to an obtuse apex',nbsp;the surface is marked by more or less distinct longitudinal foldsnbsp;and close to the base are slight depressions indicating the formernbsp;presence of seeds 5 mm. long and T5 mm. broad. On some ofnbsp;the scales there is a faint curved trilobed line in the upper thirdnbsp;of the ventral face which, it is suggested, may possibly mark thenbsp;limit of an ovuliferous scale fused to a subtending bract-scale, butnbsp;there is no substantial ground for any conclusion as to the morphological nature of the cone-scales. Zeiller states that there is nonbsp;decisive evidence with regard to the systematic position of thisnbsp;fossil: the presence of three seeds recalls Cunninghamia thoughnbsp;this in itself is probably of comparatively small importance.

This designation was proposed for fossils which are in all probability male strobili of Gymnosperms^.

This, the type-species, was discovered by the late Dr Gunn in Kimeridgian strata on the North-East of Scotland: it consists of anbsp;slender axis 13 cm. long bearing numerous small branches withnbsp;1 Zeiller (03) B. p. 208, PI. L. fig. 15; PI. P, fig. 2.nbsp;nbsp;nbsp;nbsp;^ geward (IP) p. 686.

spirally disposed sporophylls. A few oval spores (fig. 799, c), 20�30 u in diameter, were obtained from some of the sporophylls.nbsp;The specimen is probably an inflorescence of male flowers of anbsp;Conifer; it bears some resemblance to Zeiller�s Pseudoasterophyllitesnbsp;VidalP from the Kimeridgian of Spain, but in that type the leavesnbsp;are in whorls and the shoot appears to be sterile.

The smaller example shown in fig. 799, d, e, may be part of a microstrobilus of Elatides curvifolia, but in the absence of anynbsp;connexion with that species it should be retained in Masculostrobus.

M�ller and Halle^ in their account of a flora from South-East Scania, probably of Wealden age, describe some fragments ofnbsp;microstrobili which they assign to Masculostrobus. The Cretaceousnbsp;specimens from Kreischerville described by Hollick and Jeffrey asnbsp;Strobilites microsjoorophorus^ may be transferred to Masculostrobus.nbsp;This species, represented by portions of small strobili a few millimetres long, is characterised by sporophylls in which the leaf-traces are enclosed by transfusion-tracheids, a feature suggestingnbsp;comparison with modern Araucarineae, and the spores have twonbsp;bladder-like wings. It may well be, as the authors of this typenbsp;suggest, that the extension of the exine of the microspores, nownbsp;restricted to the Abietineae and Podocarpineae, may be an ancientnbsp;character and formerly more widely spread among the Coniferales-

Endlicher^ instituted this genus for a type of Conifer previously described by Braun� from the Ehaetic flora of Franconia aSnbsp;Cunninghaniites sphenolepis, characterised by distichous, linear,nbsp;leaves and a megastrobilus composed of loosely imbricate scales.nbsp;The name Palissya was selected to commemorate Bernard Palissy-The type-species, P. Braunii, was first figured by Goeppert� andnbsp;described in more detail by Schenk� who, as Nathorst points out,nbsp;included two distinct types of reproductive shoot under the nam�nbsp;Palissya'. one of them is retained in that genus while the othernbsp;agrees with cones referred to Elatides^. Endlicher�s designation

^ M�ller and Halle (13) p. 36, PI. vi figs. 9� ^ Hollick and Jeffrey (09) B. p. 66.

tas been employed by several authors for Mesozoic shoots with linear leaves imaccompanied by reproductive organs and havingnbsp;therefore no claim to be assigned to a genus characterised by anbsp;Well defined type of strobilus. Further reference to the misapplication of Palissya is made in the account of some Indiannbsp;specimens recently assigned to Retinosporites. In the absence ofnbsp;^ny evidence of the occurrence of strobili, shoots similar in habitnbsp;to Palissya should be described under Halle�s generic namenbsp;^Itttocladus. An impression of a shoot from Triassic beds innbsp;Ilucks County, Pennsylvania, recently described by Wherry^ asnbsp;I^alissya longifolia, would be more appropriately referred to Elato-^^adus as it affords no indication of the nature of the fertile branches.

Nathorst^ in his recent and able account of this species reverts Iquot;� the older specific name on the ground that Endlicher�s namenbsp;Praunii is not in accordance with the laws of priority. It isnbsp;Possible though not certain that some specimens described undernbsp;*iifierent names by Presl prior to the publication of Braun�s accountnbsp;P'^nninghamites sphenolepis may belong to this species. Schenknbsp;Considerably extended Braun�s description and was the first tonbsp;Publish figures of ripe cones showing certain morphological featuresnbsp;^ the seed-bearing scales which authors have differently interpreted.nbsp;^l^ssya sphenolepis is a Ehaetic species recorded from Franconianbsp;Scania, possibly more closely allied to the genus Gunninghamianbsp;to any other existing Conifer; but, as Nathorst suggests, itnbsp;belong to an extinct section of Gymnosperms. The foliage-^^oots bear spirally disposed, two-ranked, leaves of the Taxitesnbsp;the lamina is narrow and linear with a median vein, decurrentnbsp;^ tbe axis as a persistent leaf-cushion. The epidermal cells havenbsp;aight Walls and the stomata, confined to the lower surface, occurnbsp;Wo rows. Palissya cannot be identified with any degree ofnbsp;^mty in absence of well-preserved strobili. The mega-difnbsp;nbsp;nbsp;nbsp;cylindrical and relatively narrow; in an immature con-

for^�^ closely resemble those of Elatides, the surface being scal^^*^ ^be lanceolate, imbricate, distal ends of crowded cone-It is the older strobili with elongated internodes that

constitute the most striking feature of the genus: the cone-scales (sporophylls) are entire, elongate-lanceolate (fig. 800) with annbsp;acuminate apex, and each bears 5�6 pairsnbsp;of seeds characterised by a cup-like basalnbsp;investment or cupide (fig. 800, C). Thenbsp;sporophylls have a strongly developed keelnbsp;on the lower surface and a less distinctnbsp;median rib between the two rows of seedsnbsp;on the upper face (fig. 800, A, B). Therenbsp;is no evidence to support the view that thenbsp;cone-scales are double^. Schenk^ describednbsp;the cone-scales as bearing 10�12 seeds onnbsp;the edge and Saporta� believed the seedsnbsp;to be lobes of a seminiferous scale, eachnbsp;lobe supporting one seed. Nathorst�s investigation of Scanian material has thrownnbsp;a welcome light on the nature of the mega-strobili as interpreted by previous authors.

The bodies described by Schenk as seeds are projecting spherical casts of cup-likenbsp;organs which originally embraced the lowernbsp;portions of the seeds. The morphologicalnbsp;nature of the cupule cannot be determined,nbsp;but as Nathorst suggests it may correspondnbsp;to the epimatium^ which partially enclosesnbsp;the seeds of Dacrydium and other recentnbsp;Conifers. Nathorst compares the cone-scales of Palissya with thosenbsp;of the genus Cunninghamia (fig. 684, K, p. 116) in which each sporo-phyll bears three seeds on the adaxial side of a membranous outgrowth stretched across the scale. The resemblance would seemnbsp;to be closer than Nathorst suspects as the membrane in Cunning-hamia does not arise as a continuous strip of tissue but as threenbsp;separate ligule-like pieces, one on the abaxial side of each seed.

This genus was founded with a view to reduce the �present state of intolerable confusion in the classification of the Conifers 1.�nbsp;Halle expresses the opinion, with which most botanists cannot butnbsp;agree, that sterile shoots of Conifers should not be described undernbsp;generic names which imply the possession of a certain type ofnbsp;cone. The occurrence of foliage-shoots of similar or even identicalnbsp;habit in certain recent genera possessing distinct types of reproductive shoots serves to emphasise the unscientific character ofnbsp;the too common practice of assigning fossils to genera distinguishednbsp;hy a particular form of cone even though the specimens in questionnbsp;afford no evidence of the nature of the fertile branches. Thenbsp;generic name Taxites has been widely employed for dorsiventralnbsp;branch-fragments bearing linear leaves with a pseudo-distichousnbsp;arrangement as in the recent Taxus, but notwithstanding thenbsp;employment of Taxites by authors who do not intend to conveynbsp;the impression of relationship to Taxus, it is preferable to reservenbsp;Taxites for specimens which there are reasonable grounds fornbsp;believing to be related to the Taxeae. Halle insists that it isnbsp;undesirable to use one name for dorsiventral shoots and anothernbsp;tor shoots with spirally disposed leaves, as fossil forms are known,nbsp;^�9- Stachyotaxus elegans, Nathorst, in which both types occur onnbsp;the same plant. Similar cases of dimorphism are well illustratednbsp;by Dacrydium and other recent Conifers (c/. fig. 708, p. 160).nbsp;Tntocladus is proposed for sterile coniferous branches of the radialnbsp;or dorsiventral type, �which do not show any characters thatnbsp;permit them to be included in one of the genera instituted fornbsp;uiore peculiar forms.�

The name serves a useful purpose for sterile shoots which it has been the custom to include in Taxites and for types such asnbsp;Thtocladus heterophylla Halle, which bear both distichous, linearnbsp;leaves and crowded scale-like leaves similar to those of Brachy-pTyllum and some forms referred to Pagiophyllum. It is, however,nbsp;*lesirable to retain Brachyphyllum and Pagiophyllum for sterilenbsp;shoots exhibiting no well marked dimorphism and bearing fleshynbsp;oppressed leaves and four-sided falcate leaves respectively. Used

in this narrower sense Elatocladus is more likely to serve the object which the author had in view. It is noteworthy that in somenbsp;specimens of Elatocladus {Taxites spp.) the leaves have a transverselynbsp;wrinkled lamina, a feature usually associated with the detachednbsp;linear leaves assigned to the genus Pityofhyllum.

The shoots of this species from the Jurassic flora of Graham Land^ are freely branched and the ultimate branches show a

tendency towards a distichous arrangement. The radially symmetrical branches bear short and thick, acute, more or less appressed leaves or the lamina may be slightly falcate, while the dorsiventralnbsp;shoots have narrow linear acute leaves in two ranks (fig. 801).nbsp;A faintly marked midrib is present in both forms of leaf. Hallenbsp;states that the shorter leaves differ from those of Brachyphyllumnbsp;in being rather longer and less closely appressed to the axis, thoughnbsp;this is a difference of secondary importance. The main interestnbsp;of the species is its dimorphism.

In Elatocladus is also included the Indian species, recorded by 1 Halle (13^) p. 84, PI. vni. text-fig. 18.

Hallei from Graham Land, originally described by Oldham and Morris as Cunninghamites confertus and subsequently removed bynbsp;Peistmantel to Palissya; the leaves are distichous with a sessile andnbsp;decurrent lamina attached at a wide angle. The apex of the leavesnbsp;is obtuse and a midrib is present. In Elatocladus conferta Halle includes the Australian form Palissya australis as figured by Stirling^.

This type, from the Middle Jurassic of \ orkshire, was described fcy Leckenby� as Cycadites zamioides and subsequently transferred

i-c Taxites^. The type-specimen, in the Sedgwick Museum, Cam-dridge, consists of a slender axis bearing two-ranked spirally

^ Halle (132) p. 86, PI. viii. nbsp;nbsp;nbsp;= Stirling (00) PL III. figs. 8, 9.

attached linear leaves slightly more than 2 cm. long and 1-8 mm, broad, basally contracted, with an acute apex. It closely resemblesnbsp;shoots of recent Taxeae and Sequoia seinpervirens, also some Potomacnbsp;species included by Fontaine^ in Cephalotaxopsis though there isnbsp;no evidence of relationship to the recent Cephalotaxm. Shoots ofnbsp;similar habit are figured by FeistmanteP from Indian Jurassic bedsnbsp;as Taxites planus but the lamina is not contracted at the base.nbsp;Two of Feistmantel�s figured specimens are reproduced in fig. 802,nbsp;These afford good examples of fossil branches which it has beennbsp;the custom to refer to Taxites, but without information with regardnbsp;to the epidermal characters it is impossible to determine theirnbsp;affinities. The form of the leaf-bases agrees with that shown innbsp;fig. 803 and it is probable that Taxites planus may be anothernbsp;example of Miss Holden�s genus Retinosporites, though in thenbsp;absence of anatomical data Elatocladus is the more appropriatenbsp;designation.

Feistmantel used the name Palissya for some Indian Jurassic vegetative coniferous shoots which afford no evidence of affinitynbsp;to that genus as represented by P. Braunii. Some of his fossilsnbsp;may be identical with the British species Taxites zamioides, noWnbsp;assigned to Elatocladus, while the examples described by Feistmantel a's Palissya sp. and Palissya indica have been transferrednbsp;to a new genus Retinosporites. The Indian impressions afford nonbsp;evidence of a midrib; the upper epidermis consists of cells withnbsp;straight walls and there are no stomata, while on the lower facenbsp;of the lamina stomata are irregularly scattered, the long axis ofnbsp;the guard-cells being more or less parallel to the margin of thenbsp;leaf. The absence of a midrib, at least so far as regards impressionsnbsp;and cuticular preparations, led Miss Holden� to separate the Indiannbsp;specimens from Palissya and Taxites as vegetative shoots includednbsp;by authors in genera having leaves with a distinct median veinnbsp;and in which the stomata are in rows on the lower surface. Thenbsp;generic name Retinosporites, spelt by Miss Holden Retinosporitis,nbsp;is proposed on the ground that the only flat-leaved Conifers among

those examined showing similar epidermal characters were certain Seedlings of the Retinospora type. Miss Holden recognises thatnbsp;^stinosporites suggests relationship to such recent Cupressineousnbsp;Species as are included under the genus Retinospora, but she statesnbsp;that no such implication is intended. The new designation, thoughnbsp;Hot very happily chosen, may be retained for shoots Avith linearnbsp;leaves (fig. 803) without a midrib and having the stomatal featuresnbsp;described in Miss Holden�s account of R. indica.

An Indian species originally described by Oldham and Morris Taxites indicus and transferrednbsp;Teistmantel to Palissya. Thenbsp;^�HVes are linear and decurrentnbsp;803, A) and without a midrib,nbsp;epidermal cells have straightnbsp;''^^lls and the stomata, though occasionally present on the uppernbsp;^^rface, are scattered on the lowernbsp;epidermis as in the Retinosporanbsp;oliage of Thuya or Juniperus,

quot;without any indication of a median nbsp;nbsp;nbsp;minosporites Mica. A, piece

^�tomatic region such as one would of shoot; B, stoma. (After R. Holden.) expect in leaves possessing a mid-

� There are generally six accessory cells and the guard-cells are below the level of the epidermis (fig. 803, B).

PresB gave the name Cunninghamites to some sterile shoots ^��ii Rhaetic and Lower Cretaceous strata on the ground of theirnbsp;Semblance to branches of Cunninghamia and both his specimensnbsp;the large number, particularly from Cretaceous beds, referrednbsp;^y authors to Presl�s genus afford no real evidence of affinity tonbsp;recent Conifer. The type-species Cunninghamites oxycedrus,nbsp;Q ^ Lower Cretaceous rocks in Saxony, is probably identical Avithnbsp;ce ^ species Cunninghamia elegans (fig. 804) from Lower Creta-beds in Bohemia. Presl also included in Cunninghamites

branches from the Keuper of Germany which he named C. dubius: this species is identified by Saporta^ with Palissya Bmunii, butnbsp;the latter name is now restricted to shoots bearing a particularnbsp;form of strobilus. Nathorst^, on the other hand, suggests thatnbsp;C. dubius may belong to Elatides, and the same author institutednbsp;a new generic name Camptophyllum^ for fragmentary foliage-shoots

from the Rhaetic of Scania which he named C. Schimperi (fig-804, B): these bear a close resemblance to C. elegans; the linear-lanceolate leaves are 15�20 mm. long and the lamina is recurved, probably as the result of drying. It is impossible to determin�nbsp;the position of this type and it may conveniently be transferrednbsp;to Elatodadus. Some of Nathorst�s specimens are also figured b/nbsp;Schenk^. The name Cunninghamites is given by Oldham and

^iorris^ to Jurassic Indian shoots which FeistmanteP afterwards ^��ansferred to Palissya: these have recently been included bynbsp;Kalle� in Elatocladus as E. conferta and recorded by him from thenbsp;^ich Antarctic flora of Graham Land.

In no case have we any definite information with regard to the Cones borne by the Cunninghamites type of shoot. Velenovsky*nbsp;figures some Lower Cretaceous cone-scales from Bohemia, which henbsp;Refers to Cunninghamia, resembling in shape the scales named bynbsp;Rollick and Jeffrey Protodammara, but the resemblance may benbsp;superficial.

^ The vegetative branches assigned by authors to Cunninghamites linear-lanceolate leaves usually shoving a distinct midrib andnbsp;citen other parallel lines on the lamina which are probably duenbsp;� fiypodermal fibres. The leaves may reach a length of 6 cm.nbsp;��d are 1�4 mm. broad; the edge is entire and finely serrate as innbsp;^'^ninghamia sinensis. A characteristic feature is the occurrencenbsp;persistent decurreiit leaf-bases on the branches which in somenbsp;^Pecirnens that have lost the free portion of the lamina present anbsp;^c^e similarity to Brachyphyllum. Some of the examples ofnbsp;j^^'gt;^inghamites may well be shoots of a plant allied to Araucarianbsp;^ * ^illii or A. brasiliensis. There is no doubt that under Cunning-are included branches of many different Conifers.

.P^iginally described by Corda� from Lower Cretaceous rocks ohemia as Cunninghamia elegans, this species is recorded fromnbsp;Cretaceous localities; from the Patoot beds in Greenland�,nbsp;�^avia�^ Westphalia�, Upper Cretaceous beds in Bulgaria�, thenbsp;clays^o, Cliffwood, Martha�s Vineyard^^, Georgia, Carolina^^,nbsp;fittnbsp;nbsp;nbsp;nbsp;places in North America 1�. The specimen shown in

am and Morris (63) PI. xxxii. fig. 10.

3 '-quot;�tmantel (762) p, 55.

_ Halle (132) p, ge, PI. viii. nbsp;nbsp;nbsp;� Velenovsky (87).

8 in Reuss (46) B. PI. xLix.

^ Heer (83) pi. lih. fig. 1. nbsp;nbsp;nbsp;� Egt;id. (69) PI. i-

, Hosius and von der Marok (80) B. PI. xxxvn. nbsp;nbsp;nbsp;^nbsp;nbsp;nbsp;nbsp;�

^ '^eiller (OS�) pi. yji 14. nbsp;nbsp;nbsp;1� Newberry and Hollick (95) PI. v.

^ fJollick (06) PI. ni. fig. 1. nbsp;nbsp;nbsp;Berry (10*) PI. xx.; (14) p. 106.

Hollick (06) for other references.

28�2

Bohemia specimen represented in fig. 804 the midrib is more obvious and the leaf-bases have a more regular form. The branching is sparse and not pinnate. Velenovsky^ assigns some branchesnbsp;to C. stemphylla but these may be younger forms of C. elegans.nbsp;Similarly C. squamosus^, as figured by Heer and other authors,nbsp;affords no satisfactory evidence of specific difference from C. elegans-Impressions from the Atane beds of Greenland described by Heer�nbsp;as C. horealis have been compared by Schenk with Sequoia and alsonbsp;referred by him to Torreya: there is no possibility of deciding thenbsp;precise systematic position of these and similar specimens. Ettings-hausen* has described as Cunninghamites miocenica fragments ofnbsp;shoots from Sagor in Carinthia bearing linear leaves with a finelynbsp;serrate edge.

This genus was instituted� for Cretaceous fossils from Stated Island superficially resembling Fern leaves with a pinnate venationnbsp;and an irregularly lobed or incised margin. The leaf-like fragmentsnbsp;are, however, stem-structures bearing minute scale-like leavesnbsp;attached to the edges and surface. In habit these phylloclade�nbsp;agree with Phyllocladus, but on anatomical grounds the authorsnbsp;of the genus regard it as Araucarian though the evidence is f**'nbsp;from convincing.

Some of the specimens show no indication of their phylloclads' nature and, as impressions, would be identified as Fern pinnuF�nbsp;or referred to Thinnfeldia. Others, after bleaching in chlorin�nbsp;water, showed a fairly stout vascular axis giving off simple ornbsp;forked branches at an acute angle; small decurrent leaves fr��nbsp;only at the apex occur on the margins of the shoots (fig. 806, A,

In a few cases the phylloclades bear short axillary branches immature cones, possibly microstrobili.

There are three vascular cylinders in the section reproduced fig. 806, C, and in the narrow Avings of the �lamina� there are th�

2 nbsp;nbsp;nbsp;Heer (71�) PI. 1. figs. 5�7; Schimper and Schenk (90) A. p. 282; Berry (nbsp;p. 64.

� Heer (82) B. PI. xxix. fig. 12. nbsp;nbsp;nbsp;^ Ettingshausen (72) PI. i. fig- ^

traces of two or three leaves. Several stomata occur on the surface of the phylloclade, each surrounded by 4�5 accessory cells. Sclerotic cells are present in the pith. The secondary xylem is of thenbsp;r^oniferous type and the uniseriate bordered pits on the tracheidsnbsp;^ay be either separate and circular or flattened by contact. Nonbsp;resin-cells, such as occur in the wood of Phyllocladus, were recog-nised. The medullary rays are not described. The data are

y sufficient to justify the inclusion of this type in the ; the occasional flattening of the tracheal pits andnbsp;Thquot;nbsp;nbsp;nbsp;nbsp;of resin-cells are not fatal to an alliance to Phyllocladus.

Sent �P�cies, from the Upper Cretaceous of Georgia, is repre-hy dorsiventral fern-like vegetative shoots which, as Berry^ ^ Berry (14) p. 103, PI. xviii.

points out, bear a close resemblance to species of Moriconia (cf-fig. 760). The branches are distichously arranged and the opposite lateral leaves are stout, falcate, and decurrent while those on thenbsp;upper and lower faces are represented by scales on the middlenbsp;line of the phylloclades.

The same author describes specimens from beds in North Carolina^ referred to the lower half of the Upper Cretaceous asnbsp;Androvettia caroUnensis.

Dactyolefis cryptomerioides Hollick and Jeffrey. The generic name^ was instituted for some detached, cuneate, cone-scales fromnbsp;the Cretaceous beds at Kreischerville in Staten Island, approximately 4 mm. long, composed of an upper and a lower segment.nbsp;The upper portion is divided distally into as many as sevennbsp;irregular short finger-like processes and the lower part is entire.nbsp;Each of the processes possesses a single vascular bundle � completelynbsp;surrounded by a cordon of transfusion-tissue, thus betraying itsnbsp;Araucarineous relationship.� The scales which are without seedsnbsp;are compared with those of Voltzia. There is, however, no proofnbsp;that Voltzia had double scales. The view that Bactyolepis isnbsp;Araucarian may fairly be said to rest on an insufficient basis.

The name Raritania^, after the Raritan formation, was given to some Cretaceous fossils from Kreischerville identical with Nen'nbsp;Jersey specimens described by Newberry as Frenelopsis gracilis^nbsp;on the ground that they belong to a type distinct from Frenelopsisnbsp;as generally understood.

Raritania gracilis (Newberry). The specimens so named consist of slender, dichotomously branched, axes bearing minute leaves resembling Psilotum triquetrum and in the form of thenbsp;branching the leaves of Baiera Ldndleyana (Schimp.). The distinguishing feature is the occurrence of the prickle-like leavesnbsp;(fig. 807, B) invisible to the unaided eye (fig. 807, A). A small

imperfectly preserved cone was found on a peduncle having leaves similar to those on the vegetative twigs. Some fragmentarynbsp;iignitic branches (fig. 807, C) associated rvith the impressionsnbsp;showed the anatomical characters of a Conifer; but Hollick andnbsp;Jeffrey, though believing that the fragments �almost certainly�

belong to Raritania, admit that there is no proof of their identity 'i^ith the dichotomously branched impressions. The secondarynbsp;^ylem of the cylindrical stele of the twigs, one of which is shownnbsp;bi fig. 807, C, agrees with that of Brachyfhyllum 'tnacrocarpuninbsp;^ewb. The uniseriate bordered pits are occasionally flattened;nbsp;^he cortex, confluent with the decurrent leaves, contains scleroticnbsp;Cells and each leaf has a resin-canal subtending the leaf-trace.nbsp;The genus is referred to the Araucarineae on the evidence of thenbsp;Occurrence of flattened pits at the ends of some of the tracheids,nbsp;^ conclusion difficult to accept without considerable reservation.

Braun^ instituted this genus for a strobilus from Rhaetic I�ocks in Germany which he called Schizolepis liaso-lceuperinus,nbsp;' Braun (47) p. 86.

characterised by its deeply split cone-scales. Schenk^ subsequently substituted the name S. Braunii and included under that desisna-tion Brann�s Isoetites pumilus, a species founded on a foliage-shoot,nbsp;also some other similar vegetative branches believed to belongnbsp;to the plant which bore the cones. As here used, the termnbsp;Schizolepis is restricted to cones and cone-scales since there is nonbsp;definite evidence as to the nature of the foliage-shoots connectednbsp;with the strobili. Schizolepis cannot be referred on any satisfactory grounds to a definite position among the Coniferales: it isnbsp;possibly an extinct type allied to recent Abietineae, but untilnbsp;more is known with regard to the morphology of the cone-scalesnbsp;the systematic position must be left an open question. The genusnbsp;is represented by strobili from Ehaetic beds in Franconia, Scania,nbsp;and Poland; detached scales from Middle Jurassic floras are alsonbsp;included in Schizolepis (fig. 808), and Nathorst has describednbsp;incomplete strobili from Upper Jurassic or Wealden strata innbsp;Spitzbergen. Attention has been called to a resemblance betweennbsp;Schizolepis scales and the fertile leaves of Tmesipteris^, but therenbsp;is no reason for regarding this as indicative of relationship. Morenbsp;than one author has compared the bilobed cone-scales of Schizolepisnbsp;with the 3�5-lobed scales of Voltzia and Cheirolepis though thisnbsp;comparison rests on a feature which in itself is no proof of affinity.nbsp;A comparison may also be suggested with the reflexed cone-scalesnbsp;of Picea Breweriana.

It has already been pointed out that under this name Schenk^ included both cones and vegetative shoots though he recognisednbsp;the lack of any decisive evidence of common parentage. Whilenbsp;agreeing with Nathorst that the association with Schizolepisnbsp;strobili of similar vegetative shoots both in Germany and Swedennbsp;may be more than accidental, in the present state of our knowledgenbsp;it is preferable to refer the leaves and branches to Pityophyllum ornbsp;Pityocladus. In the younger strobili the bilobed scales are morenbsp;or less pressed against the axis and in older examples they arenbsp;more spreading: each scale has two lanceolate lobes and is attached

by a narrow stalk-like basal portion (fig. 808, B). that there are two anatropous seeds tonbsp;each scale, but it is not clear if the actualnbsp;seeds are present.

In this species from the Rhaetic flora of Scania^ the bilobed scales are sessilenbsp;and broader than in S. Braunii. In somenbsp;specimens there appear to be two seedsnbsp;Hear the base of a scale. Saporta^ regardsnbsp;the cone-scales of the type-species and S.

Follini as double structures, the transverse line shown in Schenk�s figures below the forking of the scale being the limit ofnbsp;the bract-scale, while the thinner distalnbsp;^obed part represents the seminiferousnbsp;�cale; an interpretation which rests onnbsp;Very slender evidence. To this speciesnbsp;Jtaciborski� refers a specimen from rocksnbsp;possibly of Rhaetic age in Poland.

Under this name^ some detached bi-lobed scales without seeds were described ^I'ona Jurassic rocks in Turkestan (fig. 808, A) and the South ofnbsp;Itussia�, and Krystofovic has recently discovered a cone of thenbsp;�3'me species in Jurassic strata in Transbaikalia�.

Bounded on a long and narrow, incomplete, strobilus from ^Pper Jurassic rocks in Spitzbergen� bearing deeply lobed scales.nbsp;^ second species, S. retroflexa^ (fig. 808, C), assigned with some

doubt to Schizolepis, is characterised by pendulous stalked scales : in neither of these Spitzbergen forms are there any seeds onnbsp;the strobili. Nathorst quotes a species from Portugal describednbsp;by Saporta^ as Palaeolepis bicornuta as being possibly allied tonbsp;Schizolepis, but there are not sufficient grounds for assuming anynbsp;close affinity. He also draws attention to the resemblance of thenbsp;bilobed leaves or scales of Schizolepis {?) retroflexa to the fertilenbsp;leaves of Tmesipteris; the relationship of the Spitzbergen fossilsnbsp;must be left uncertain pending more evidence.

Heer described a specimen of Jurassic age from Spitzbergen as Phyllocladites rotundifolia^ which he considered to be closelynbsp;allied to Phyllocladus: an examination of the type-specimen lednbsp;Nathorst� to institute a new generic name Drepanolepis. Asnbsp;Nathorst�s revised description and more, accurate drawing show,nbsp;there are no substantial grounds for assuming any relationshipnbsp;between the fossil and Phyllocladus. Drepanolepis rotundifolitf'nbsp;consists of a fairly stout axis bearing spirally disposed thick,nbsp;falcate scales each of which bore a seed, or possibly a sporangium,nbsp;near the base (fig. 798, C). A similar type is described by Nathorstnbsp;as Drepanolepis angustior'^ characterised by the narrower form ofnbsp;the scales and a broader axis. Both species may be described asnbsp;strobili of open habit with single-seeded sporophylls: it is impossiblenbsp;to determine the systematic position of the genus, though asnbsp;Nathorst says it is probably a type of fertile Gymnospermousnbsp;shoot. There is no reason for comparing the specimens withnbsp;Phyllocladus.

Schizolepidella gracilis Halle. The specimens on which this genus is founded are from the Hope Bay flora in Graham Land�,nbsp;probably of Middle Jurassic age; they consist of slender steril�nbsp;shoots reaching a maximum length of 12 cm. and 2 mm. broad,nbsp;rarely branched and bearing small leaves, 2 X 1-5 mm., apparently

� Ibid. p. 71, PI. in. figs. 33�37. nbsp;nbsp;nbsp;= Halle (13q p. 90, PL ix. figs. 18-

spirally disposed (fig. 809). The lamina is rounded, obovate, and always bilobed at the broad apex:nbsp;no veins were detected. As Halle says, it is impossible to determine the affinities of the fragments but he thinks they may belong to pendulous branches of a Conifer. Attention is callednbsp;to a resemblance to some Hepaticae, and a possible relationship to Lycopodium or the Psilotalesnbsp;may also be suggested. The choice of the genericnbsp;name is not intended to imply anything morenbsp;than a superficial similarity between the leavesnbsp;and the bilobed strobilar appendages of Schizo-lepis^.

This name was given^ to foliage-shoots and B, nat. size.) cones, from the Urgonian rocks of Greenland, originally described asnbsp;^iddringtonites gracilis^. The smaller sterile branches are indistinguishable from specimens referred by authors to Widdringtonitesnbsp;while the larger examples might be included in Brachypltyllum. Thenbsp;leaves are small, appressed, and imbricate, similar to those of somenbsp;recent Cupressineae and Callitrineae but spirally disposed and notnbsp;''^erticillate (fig. 810); the shoots agree also with Microcachrys andnbsp;other recent Conifers. The cones are composed of a small numbernbsp;of flat scales (fig. 810, B) too imperfectly preserved to afford anynbsp;definite evidence as to the affinities of the genus. Heer statesnbsp;l-hat a detached cone-scale shows the impression of a single seed,nbsp;l*rit the material is insufficient to form the basis of a comparisonnbsp;^ith the Araucarineae; he points out a resemblance to Cunning-and mentions the striated surface of the fossil cone-scalesnbsp;a distinctive feature, though that may be due, in part at least,nbsp;lo the state of preservation. The fiat form of the cone-scales is anbsp;character in which Cyparissidium differs from genera such asnbsp;^^guoiites, and from the Callitrineae the cones are distinguished bynbsp;I'l'e spiral arrangement of the scales.

^ Heer (75) ii. PI. xvn. fig. 5 b, c; Pk. xix., xx., xxi.; (82) pp. 16, 50, Pis. i., vii.,

Cyparissidium is characteristic of Lower Cretaceous strata though Nathorst^ has recorded a species, C. Nilssonianum, fromnbsp;Rhaetic rocks in Scania with cones having scales more pointednbsp;and lanceolate than those of Heer�s species. A second Rhaeticnbsp;species, C. septentrionale, has been transferred by Nathorst to thenbsp;genus Stachyotaxus^.

The type-species (fig.- 810, A, B) is one of the most abundant plants in the Rome beds at Pattorfik in West Greenland and itnbsp;occurs also in the Atane beds. The cones are 2G mm. long andnbsp;11�20 mm. broad with scales having a breadth of 12 mm., ^

rounded distal border, a mucronate apex, and a striated dorsal face. Velenovsky^ records this species from Lower Cretaceousnbsp;rocks in Bohemia but without the essential evidence of cones: thenbsp;same author describes a smaller type from Bohemia as C. minimmn^nbsp;(fig. 810, C). Schenk� figures some fragments from Lower Cretaceous rocks in the Tyrol as Cyparissidium cretaceum, but in thenbsp;absence of cones there is no sufficient reason for their inclusion innbsp;Cyparissidium. Hollick�s record of C. gracile from the Cretaceousnbsp;of Block Island* is based on insufficient evidence.

In 1862 Mackie� figured a fossil stem from Lower Greensand rocks in Kent which K�nig afterwards named Dracaena Benstedtii.nbsp;This name was adopted by Morris and Mantell for the supposednbsp;�Dragon tree.� In 1868 Carruthers� expressed the opinion thatnbsp;the fossils are more likely to be Pandanaceous stems, while Gardner�nbsp;spoke of a possible Cycadean affinity. An examination of Mackie�snbsp;specimens and others in the British Museum led me to suggest anbsp;Comparison with stems of recent species of Cycads, particularlynbsp;stems of Zamia which do not retain the armour of leaf-basesnbsp;(fig. 381 B, vbl. III. p. 5) characteristic of most Cycadales, and tonbsp;institute a new generic na,me Benstedtia^ in preference to a designation implying an improbable relationship. The genus is definednbsp;as follows: stems characterised by irregular and interruptednbsp;grooves and broader ridges running transversely, with occasionalnbsp;small elliptical protuberances irregularly disposed on the surface.nbsp;There are no distinct leaf-scars but branch-scars occasionallynbsp;occur; the upward convergence of the transverse wrinklings indicates bifurcation in some specimens.

The English examples reach a length of over 40 cm. and a diameter of 15 cm. Smaller specimens with similar surface-characters are described by Fliche* from Lower Cretaceous beds innbsp;Trance as Coniferocaulon colymbeaeforme and compared by himnbsp;to stems of Araucaria imhricata: other examples closely resembling

* nbsp;nbsp;nbsp;VelenoFsky (85) B. p. 17, PI. vm.nbsp;nbsp;nbsp;nbsp;^ Ibid. Pis. ix., x.

� Gardner (86^) p, 201. nbsp;nbsp;nbsp;� Seward (96�) p. 216.nbsp;nbsp;nbsp;nbsp;� Fliche (00).

the French stems are recorded from the Uitenhage (Wealden) series of South Africa and the Kimeridge rocks of Sutherland^,nbsp;Scotland, in the former case as Benstedtia sp., and in the latter asnbsp;Coniferocaulon colymbeaeforme because of the presence of a smallnbsp;pith more suggestive of a Conifer than of a Cycadean branch. Anbsp;large specimen from Jurassic rocks in India is described by Missnbsp;Bancroft^ as Coniferocaulon sp.; this agrees very closely in surface-features with the casts from Kent, but an examination of transversenbsp;and longitudinal sections demonstrated that the apparent surfacenbsp;is not the actual surface, and such anatomical data as it wasnbsp;possible to obtain clearly indicated the Coniferous nature of thenbsp;wood, conclusions in agreement with those reached by Dr Stopesnbsp;in the case of some English specimens. The reference of thesenbsp;stems to Cycads or Conifers was based entirely on surface-charactersnbsp;and it was recognised that no definite conclusion was possiblenbsp;without anatomical confirmation. Dr Marie Stopes� succeeded innbsp;obtaining preparations of tracheids from a Lower Greensand castnbsp;showing uniseriate and separate bordered pits of the Abietineousnbsp;type thus disproving a Cycadean affinity. This discovery led tonbsp;the substitution of Coniferocaulon for Benstedtia, at least as regardsnbsp;the specimens which afforded anatomical evidence; in the Frenchnbsp;and South African examples no internal structure is preserved.nbsp;Dr Knowlton* pointed out that the generic name Benstedtia shouldnbsp;be preferred to Coniferocaulon on the ground of priority and henbsp;named the English specimens Benstedtia Benstedtii. Dr Stopes�nbsp;replied to this criticism by asking why the Dragon tree, which isnbsp;merely a partially decorticated piece of badly preserved Coniferousnbsp;wood, should have a name. Specimens exhibiting distinctivenbsp;surface-features, whether complete or decorticated and even ifnbsp;they are in some cases at least portions of Coniferous stems, arenbsp;none the less entitled to some recognition as a matter of convenience.nbsp;Some excellent illustrations of Benstedtia casts are given by Drnbsp;Stopes in her recently published Catalogue of Lower Greensandnbsp;Plants�.

CHAPTER LI.

The name Podozamites^ was instituted for certain species previously included in Zamites characterised by the possession of distant alternate pinnae with a contracted base and veins slightly spreadingnbsp;Hi the proximal part of the lamina but for the most part approxi-Hiately parallel. As defined by Braun Podozamites differs in nonbsp;Very important respect from Zamites, and the latter name is retainednbsp;Schenk for Z. distans Presl in preference to Podozamites appliednbsp;to that species by Braun. By most authors Podozamites has beennbsp;regarded as Cycadean, but Sehenk�s discovery of a specimen ofnbsp;Podozamites'^ in the Rhaetic beds of Franconia showing a clusternbsp;of small scale-leaves at the base of the axis led him to suggest anbsp;possible affinity to Agathis as an alternative to the generallynbsp;S'Ccepted view that the appendages are leaflets or pinnae homologous ivith those of a pinnate Cycadean frond. In a later papernbsp;Schenk included in Podozamites some undoubted pinnate frondsnbsp;which Schimper founded the genus Glossozamites^. Schenknbsp;quot;�as, however, influenced in his preference for a Cycadean alliancenbsp;the structure of the epidermal cells (fig. 812, E) which havenbsp;straight walls, and the same consideration weighed with Zeiller^nbsp;quot;�ho was strengthened in his view by the characters of the seed-bearing sporophylls described by Nathorst� and provisionallynbsp;Connected by him with Podozamites. Nathorst� described a specimen from the Rhaetic strata of Scania agreeing in the presence ofnbsp;^sal scale-leaves with that figured by Schenk, and more recently

vi:-:-4r: .. nbsp;nbsp;nbsp;�

Schuster^ has published photographs of some examples of Podo-zamites distans in which the base of the axis is invested by small imbricate scales and in connexion with it are two other clustersnbsp;of similar scales, probably unexpanded buds. In 1900 I expressednbsp;the opinion that Podozamites is probably a Conifer^, the supposednbsp;pinnate fronds (fig. 811) being foliage-shoots like those of recentnbsp;species of Agathis. The most important recent contribution to ournbsp;knowledge of Podozamites is due to Nathorst: in 1911 he publishednbsp;additional facts� with regard to some seed-bearing organs from thenbsp;Hhaetic of Scania for which he proposed the generic name Cycado-(^arpidium in 1886 and in 1902^ more fully described the type-speciesnbsp;C- Erdmanni. Until the publication of Nathorst�s more recentnbsp;paper Cycadocarpidium was known only as detached sporophyllsnbsp;found in beds containing Podozamites leaves. The following description is abridged from Nathorst�s account.

Cycadocarpidium Erdmanni is represented by ovate sporophylls Consisting of a sterile portion 9 mm. long and at most 6 mm. broadnbsp;'vith 4�5 simple veins, tapering to a short and slender pedicel onnbsp;each side of which is an oval seed (fig. 812, A�D) with an obliquelynbsp;placed triangular lamina compared by Nathorst with a cupule andnbsp;interpreted by Schuster as a leaflet. A specimen figured by Nathorstnbsp;shows several sporophvlls attached to a common axis, and supportsnbsp;his view that the seed-bearing organs were borne as imbricatenbsp;carpellary scales. Fig. 812, A is drawn from Nathorst�s restorationnbsp;cf a cone-like cluster of sporophylls. Another type of sporophyll,nbsp;Cycadocarpidium Swabii, is distinguished by the larger dimensionsnbsp;cf the lamina, 4-1 cm. long and 16 mm. broad, with 10 veins; innbsp;^his type the two small seeds are apparently without any appen-'^^ges (fig. 812, C, D). A third species, C. redivioum, is foundednbsp;small detached leaves and bud-like clusters previously assigned to Podozamites distans. These are now recognised as smallnbsp;Cycadocarpidium sporophylls. There is a strong probability thatnbsp;Cycadocarpidium was borne on a Podozamites shoot; in form andnbsp;'^cnation the sterile lamina of the sporophylls agrees with thenbsp;leaves of Podozamites and the two organs are constantly associatednbsp;the Scanian beds. Zeiller records C. Erdmanni in Rhaetic

rocks in Tonkin^ where Podozamites also occurs. Heer^ in his account of some impressions of Podozamites from Spitzhergennbsp;figured a seed in close association with what he believed to be anbsp;carpellary leaf like that of a Cycad and suggested a connexionnbsp;between the seed and Podozamites'. this supposed connexion has,nbsp;however, little to support it.

Nathorst�s view of the sporophylls is that each is a single fertil� leaf bearing two ovules comparable morphologically with thenbsp;megasporophylls of Dioon, but, as he points out, the termin*^nbsp;portion of the sporophyll of Cycadocarpidium is much more lea-f'nbsp;like than the sterile distal end of the megasporophylls of Dioo^�nbsp;Encephalartos Zamia. He considers that the -evidence afforded

by these seed-bearing organs favours a Cycadean alliance: on the other hand he has satisfied himself that some Podozamites specimensnbsp;are shoots with spiral linear leaves like those of Agathis. It is innbsp;many cases very difficult to say whether the axis of a Podozamitesnbsp;bears the leaves in two ranks or spirally. Nathorst speaks ofnbsp;some examples in which the leaves are not spiral and suggests thenbsp;existence of two kinds of branch some with spiral and some withnbsp;two-ranked leaves as in certain Conifers. But if this is the casenbsp;One can hardly imagine that the two-ranked arrangement is notnbsp;due to the twisting of the leaves of shoots with spirally disposednbsp;foliage. In a recent contribution to the systematic position ofnbsp;Podozamites Schuster^ speaks of Cycadocarpidium Erd-tnanni asnbsp;differing from C. Swabii in the presence of two �rudimentarynbsp;leaflets� (th� triangular lamina shown in fig. 812, B) which in thenbsp;latter species are represented only by two small swellings at thenbsp;Upper ends of the seeds: he regards C. Erdmanni as the morenbsp;primitive type. The lamina is homologised with the cover-scalenbsp;or bract of the double cone-scale of the Abietineae; he comparesnbsp;the two leaflets of C. Erdmanni and the swellings in C. Swabii withnbsp;abnormal seminiferous scales of an Abietineous cone. Schuster�snbsp;^lew is that Podozamites distans is a primitive Conifer evolvednbsp;from the base of a Cycadofilicinean line which gave rise to thenbsp;Irinkgoales, a supposition based on a very slender foundation.nbsp;Nathorst regards Podozamites as an intermediate type related bothnbsp;fo Cycads and Conifers; he does not, however, overlook the factnbsp;that the sporophylls of Cycadocarpidium may be'compared withnbsp;those of some Conifers even though their resemblance to Cycadeannbsp;�porophylls would seem to be closer. Additional data are needednbsp;before we can settle the position of Podozamites, but such information as we have may be said to point to the conclusion that it isnbsp;Hearer to the Conifers or the Ginkgoales than to any other group ofnbsp;Nymnosperms. Nathorst calls attention to a similarity betweennbsp;tokoyama�s Ginkgodium Nathorsti^ and separate leaves of Podo-^^niites; a similar comparison may be made between the latter genusnbsp;Hnd Eretmophyllum, a genus instituted by Thomas� and referrednbsp;the Ginkgoales (cf. figs. 658, 659, pp. 59,62). There is indeed somenbsp;resemblance bet\\*een Cycadocarpidium sporophylls and abnormalnbsp;Schuster (114).nbsp;nbsp;nbsp;nbsp;^ Yokoyama (89) B. Pis. viii., ix., � Thomas (13).

seed-bearing leaves of Ginkgo. We have as yet but little to guide us in our attempts to trace the ancestry of that remarkable survivalnbsp;Ginkgo biloba, and it is highly probable that, if more satisfactorynbsp;records of older members of the Ginkgoales were available, wenbsp;should be able considerably to extend the range in morphologicalnbsp;characters which in the present representative of the group isnbsp;comparatively restricted. The numerous leaf-bearing axes, manynbsp;of them branched, referred by Fontaine^ to his genus Nageio-psis,nbsp;should not be overlooked from the point of view of their possiblenbsp;relationship to Podozamites. The branching habit of these Potomacnbsp;specimens is no bar to an affinity to Podozamites if examples ofnbsp;the genus are no longer to be interpreted in terms of a Cycadeaiinbsp;frond. Berryin a recent revision of Fontaine�s genus, refersnbsp;some species to Podozamites which he still regards as Cycadean-It is interesting to find on a specimen of Nageiopsis figured bynbsp;Fontaine a zone of crowded scars^ (fig. 816, s) such as may be seennbsp;on an Agathis shoot.

Many of the leaves described as Podozamites are of little value as evidence of the occurrence of the genus. In the case of imperfectnbsp;specimens of detached leaves it is often impossible to distinguishnbsp;between Podozamites, Phoenicopisis, and the leaves of Araucariannbsp;plants, or pinnae of some species of Zamites. It is therefore notnbsp;possible to state with confidence the geological range of the genus-Undoubted examples of Podozamites are essentially Rhaetic andnbsp;Jurassic fossils, and there can be no doubt as to the abundancenbsp;and wide geographical range of the genus in both these periods-Such leaves as those recently figured by Hollick^ from Cretaceousnbsp;beds of Long Island as Podozamites lanceolatus certainly agreenbsp;closely in form with that species, but they are all detached specimens : the fragmentary leaves from the Middle Cretaceous beds ofnbsp;the Amboy clays described as P. angustifolius (Eich.) and P. niaiquot;nbsp;ginatus Heer� afford no proof of the presence of Podozamites-similarly Velenovsky�s species P. miocenica from Bohemia� migh^nbsp;equally well be referred to the genus Dammarites. Well pr�'nbsp;served specimens have been described by Zeiller� from fhe

Rhaetic flora of Tonkin^. There are few satisfactory records of the genus from the southern hemisphere, and we have no actualnbsp;proof of its existence in India, though FeistmanteP refers tonbsp;Podozamites detached leaves, which, as an examination of thenbsp;original specimens shows, may have been borne on Podozamitesnbsp;shoots, but they may also be examples of Phoenicojms. One ofnbsp;the leaves figured by Feistmantel from thenbsp;Jabalpur group as Podozamites lanceolatusnbsp;is reproduced in fig. 813 from a drawingnbsp;recently made from the actual fossil; thenbsp;lamina shows several fine parallel stria-tions between the more clearly markednbsp;veins. Miss Holden, who examined thenbsp;carbonised cuticles of some of the Indiannbsp;leaves lent to me by the Director of thenbsp;Indian Geological Survey, found that thenbsp;epidermal cells have straight walls andnbsp;the stomata, usually with six accessorynbsp;cells, occur in the intercostal regions onnbsp;both surfaces; the characters of the epidermis are favourable to a relationshipnbsp;with the Coniferales and they are not

inconsistent with the inclusion of the fossils in the genus Phoe-nicopsis. Halle� has recently described some imperfect leaves from Patagonia as probably Podozamites, but as he pertinentlynbsp;says the evidence is not enough to establish the correctness of thenbsp;determination. Some of the leaves from the Potomac beds included by Fontaine in Podozamites are of little value as authenticnbsp;records of the genus, but there is still considerable doubt as tonbsp;the relationship between this genus and Nageiopsis which was verynbsp;abundant in the Potomac flora. The leaves figured by Fontaine^nbsp;from the Jurassic of Oregon and from Alaska are also not abovenbsp;suspicion as records of Podozamites, though there is no doubt thatnbsp;the genus was represented in some of the North American floras.

This Rhaetic species^ differs very slightly from the Jurassic type Podozamites lanceolatus and there has been much confusionnbsp;on the part of authors between the two forms^ which, indeed,nbsp;cannot always be clearly distinguished. P. distans is often represented only by detached leaves but in some specimens the shootnbsp;reaches a length of 20 cm. The slender axis bears distant,nbsp;lanceolate or ovate-lanceolate leaves, sometimes slightly falcatenbsp;with a rounded or obtusely pointed apex and gradually contractednbsp;at the base which, as Zeiller says, may assume the form of a verynbsp;short pedicel. The leaves may be 4�7 cm. long and 5�14 mm.nbsp;broad; the veins, 0-4�0-7 mm. apart, are dichotomously branchednbsp;in the proximal portion of the lamina but elsewhere parallel andnbsp;simple, except that they slightly converge at the apex. Thenbsp;epidermal cells have straight walls and the stomata, which occurnbsp;on the lower surface, either in rows or scattered, are surrounded bynbsp;small subsidiary cells (fig. 812, E). The leaves are usually rathernbsp;broader in proportion to their length than those of P. lanceolatusnbsp;and the apex is less pointed. Braun^ instituted two varieties,nbsp;longifolius and latifolius, and to these Schenk* added others. Thenbsp;species is recorded from the Rhaetic of Scania�, where it is abundant,nbsp;from Persia�, Tonkin, and many other regions: it occurs also innbsp;Jurassic strata^, but on the whole P. distans is a characteristicnbsp;member of Rhaetic floras.

The Rhaetic species Podozamites Schenki Heer� founded on Jurassic specimens from Siberia and described by Zeiller� andnbsp;Nathorst from Tonkin, Persia, and Sweden is distinguished fromnbsp;P. distans by the smaller shoots and the more acuminate leaves.

The type-specimen of Zamia lanceolata Lind, and Hutt.*� in the Manchester Museum from the Middle Jurassic beds of Yorkshirenbsp;consists of a slender axis bearing scattered and distant linear-

^ See Zeiller (03) B. p. 159 for examples of P. distans referred to P. lanceolatus-

� Zeiller (05) p. 193. nbsp;nbsp;nbsp;�� For references, see Zeiller (03) B. p. 159.

lanceolate leaves up to 7 cm. long and 7 mm. broad; the lamina has a tapered acuminate apex and a less gradually though notnbsp;abruptly contracted base. The leaves differ from the leaflets ofnbsp;Zamia which they superficially resemble in their less abruptlynbsp;contracted proximal end. In habit a shoot of P. lanceolatusnbsp;(fig. 811) very closely resembles the fronds of Zamia media. Somenbsp;of the leaves in the type-specimen appear to be laterally attached,nbsp;while others appear to be given off from the upper surface. Thenbsp;leaves of this as of other species are frequently found detached.nbsp;The variability in the form of the leaves has led to the employmentnbsp;of several varietal names, and if not used too freely the additionnbsp;of some descriptive term to the specific name may often serve anbsp;Useful purpose. Fig. 811 represents a good example of the speciesnbsp;from the Yorkshire coast. The method of attachment of thenbsp;leaves is not always clear, but their irregular distribution and thenbsp;slender axis are features more in accordance with a foliage-shootnbsp;than a pinnate frond. Podozamites lanceolatus is a widely distributed Jurassic species^ recorded from many European localitiesnbsp;extending to North Siberia and Spitzbergen as well as from Northnbsp;America, Turkestan, Afghanistan, Japan, China, and elsewhere.nbsp;The specimens figured by Feistmantel from Upper Gondwananbsp;rocks in India as P. lanceolatus (fig. 813) should, I am inclined tonbsp;think, be assigned to Phoenicopsis.

These species from Jurassic strata serve as examples of a Ijroader fype of the genus represented in the last two species by detachednbsp;leaves only. In these as in many other cases one cannot feelnbsp;ubsolnte confidence as to the correctness of the determination. Innbsp;�ome of the Japanese examples of P. Reinii'^ (fig. 814) the broadlynbsp;uval leaves are attached to a slender axis. P. stonesfieldensis^nbsp;from the Great Oolite of Stonesfield is probably identical with thenbsp;leaves originally described by Buckman as Naiadea ovata and Lilianbsp;^(��Hceolata: the leaves are oblong-ovate, approximately 8 x 3-5 cm.;nbsp;^Ire lamina is rather abruptly contracted at the base and more

Fontaine recognised the similarity between Nageiopsis and shoots of Podocarpus belonging to the section Nageia: this suggested thenbsp;choice of the generic name. Berry^ in his revision of Nageiopsisnbsp;transfers some of Fontaine�s species to Podozamites] he also reducesnbsp;the number of the species retained in Nageiopsis on the ground thatnbsp;Fontaine attached too much importance to variations in the sizenbsp;and form of the leaves. I have elsewhere suggested^ that some ofnbsp;the shoots referred to Nageiopsis may be Arauoarian, as in habitnbsp;they closelv resemble Araucaria Bidwilli and Agathis. Untilnbsp;reproductive organs are discovered it is impossible to speak withnbsp;confidence with regard to the position of the genus. It may benbsp;closely allied to Podozamites or, as Fontaine believed, it may benbsp;related to Podocarpus. It should be noted that some of the speci-Daens included by Fontaine in Nageiopsis are hardly distinguishablenbsp;from Zamites Buchianus^.

Trees or shrubs characterised by irregularl}^ branched foliage-shoots bearing leaves usually in two ranks but spirally attached; the leaves exhibit a wide range in size andnbsp;shape, long and linear or lanceolate, acute ornbsp;subacute, more or less abruptly contracted atnbsp;the proximal end and attached by a very shortnbsp;stalk; there are several parallel veins dichoto-l�iously branched near the base of the lamina.

This species^ founded on the small specimen represented in fig. 815 has distichous leavesnbsp;1--T5 cm. long with several parallel veins.

The English Jurassic type agrees generally quot;''ith Nageiopsis microphylla Font, and N. des-^fescens Font.; a similar form is recorded fromnbsp;the Wealden beds of Sussex�. Though satisfactory evidence ofnbsp;S'ffinity is lacking it is permissible to suggest an Araucarian affinity.

The linear-lanceolate leaves reach a length of 8�20 cm. and ^�im. to 1-3 cm. in breadth; there are 9�12 veins unbranched

except at the base (fig. 816); the lamina is abruptly narrowed and attached by a short and slightly twisted stalk. Though apparentlynbsp;inserted laterally the leaves are in all probability spirally disposed.nbsp;In one of Fontaine�s figures there is a group of small scars, fig. 816, s,nbsp;presumably of bud-scales, at one place on the axis. This species

is very abundant in the Potomac beds of Virginia and Maryland^ and is recorded by Fontaine from several other localities thoughnbsp;for the most part on slender evidence.

In this species^ the leaves are ovate-lanceolate and shorter than in N. longifolia; they have a maximum breadth of 2 cm. andnbsp;reach a length of 7 cm. The example described from Englishnbsp;Wealden beds as Nageiopsis sp. cf. N. Iteterophylla agrees closelynbsp;with Fontaine�s Potomac species.

CHAPTER LH.

In this group of Gymnosperms are included three genera, Ephedra, Gnetum, and Welwitschia. They differ widely from onenbsp;another in vegetative features, and Ephedra, the most primitive,nbsp;is distinguished by certain important peculiarities of the reproductive organs.

Having regard to our exceedingly meagre knowledge of fossil representatives, it is unnecessary to deal fully with the recentnbsp;typesb but the members of this aberrant section of seed-plantsnbsp;exhibit morphological characters of interest from the point ofnbsp;view of comparison with the Bennettitales and the Angiosperms.nbsp;Though in external appearance the three genera are poles asunder,nbsp;they have in common certain features both in the vegetative andnbsp;reproductive organs which differentiate them from all othernbsp;Gymnosperms and connect them more closely than the Cycads ornbsp;Conifers with the Angiosperms. The leaves are opposite; thenbsp;Secondary xylem contains vessels in addition to tracheids; thenbsp;^ale and female flowers are characterised by the possession of onenbsp;or two envelopes in addition to the usual single integument; thenbsp;inflorescences, occasionally though not as a rule bisporangiate^, arenbsp;distinguished by a dichasial system of branching, a characternbsp;foreign to Gymnosperms as a whole though exhibited by the stemnbsp;of Wielandiella, a member of the Bennettitales. There are good

' For a general account of the group, with illustrations, the student is referred Wettstein (11) and Lotsy (11), or to Coulter and Chamberlain (10).

^ � Pearson (06^) p. 274; (09) p. 343; Berridge and Sanday (07) p. 172; Karsten w2); Porsch (10).

Welwitschia, in some species of Gnetum, and occasionally in Ephedra. The seeds are albuminous and the embryos have twonbsp;cotyledons. Archegonia are produced in the female prothallus ofnbsp;Ephedra while in Gnetum and Welwitschia these organs are represented by single cells as in the Angiosperms or by nuclei.

Ephedra^ has a wide distribution in the warm temperate regions of the northern hemisphere: in America it occurs on both sidesnbsp;of the equator and from the Mediterranean region it reaches tonbsp;Brittany in the west and North Africa in the south. Gnetumnbsp;extends both east and west in the tropics; Gnetum scandens is anbsp;widely spread Asiatic species, and the genus occurs in Angola andnbsp;in some other parts of Africa. Welwitschia is confined to a littoralnbsp;strip of desert in extra-tropical South Africa from 14� S. to 23� S.nbsp;and has not been found more than 50 miles from the coast.

Shrubs, in some species with climbing branches, characterised by an Equisetum-like habit of the younger shoots which form longnbsp;jointed and slightly fluted branches bearing whorls of two ornbsp;sometimes three, scaly, concrescent leaves. In rare cases, e.g-Ephedra altissima, the leaves may reach 3 cm. in length and anbsp;breadth of 1�T5 mm. Monoecious or dioecious; flowers unisexual; bisexual inflorescences are recorded in E. campylopoda^-The female flowers occur in strobili on a dichasially branchednbsp;inflorescence; each strobilus consists of three pairs of bracts, innbsp;some species the bracts are more numerous. There is generally anbsp;single ovule in E. altissima, but in most species there are two or asnbsp;many as six ovules in a single strobilus. The ovules are enclosednbsp;by two envelopes regarded by some authors as a perianth and annbsp;integument and by others as two integuments. In E. distachya, asnbsp;described by Mrs Thoday and Miss Berridge�, two vascular bundlesnbsp;supply the outer envelope (outer integument) one running up eachnbsp;angle of the flattened side of the flower. The thin inner integumentnbsp;becomes free from the nucellus at a distance of two-thirds it�nbsp;length and projects beyond the outer envelope as a long style-likenbsp;micropylar tube. A ring of bundles runs a short distance up thenbsp;inner integument but ends low down in a mass of transfusion-

1 Stapf (89). nbsp;nbsp;nbsp;2 Wettstein (11) p. 417.nbsp;nbsp;nbsp;nbsp;� Thoday (Sykes) and Berridge (12)-

tracheids. Attention is dra\An to the resemblance of the outer integument to the integument of Bennettites, and the single ovulenbsp;of Ephedra is considered to be the representative of the wholenbsp;ovulate strobilus of Bennettites of which it is a much reducednbsp;derivative. There is a deep pollen-chamber at the apex of thenbsp;nucellus^ and there are 2�8 long-necked archegonia at the summitnbsp;of the prothallus (endosperm). In its female prothallus andnbsp;sexual apparatus Ephedra differs considerably from Gnetum andnbsp;Welwitschia. The second envelope forms the hard shell of thenbsp;seed which is enclosed by bracts either in the form of membranousnbsp;wings (sect. Alatae) or as a red or yellow flesh (sect. Pseudobaccatae).

Ewart^ found that the seeds of Ephedra distachya germinated after 93 days� immersion in sea-water.

The male flowers� occur also in strobili on dichasial inflorescences, a single flower occurs in the axil of each of the fertile bracts. Anbsp;flower consists of a short axis bearing a pair of membranousnbsp;appendages and the flower-axis is prolonged as a simple or bifidnbsp;stalk bearing bilocular synangia, 2�6 according to the species.nbsp;In some cases the central stalk or antherophore of the flower isnbsp;flattened and laminar^ instead of the usual cylindrical form: itnbsp;Has been interpreted both as an axial and a foliar structure, butnbsp;the latter interpretation is probably correct. Arber and Parkin�nbsp;regard the antherophore as having been formed from two fusednbsp;members, and this view is adopted by Mrs Thoday and Missnbsp;Berridge. On the basis of this interpretation the microsporophyllsnbsp;of Ephedra, represented by the antherophore, are considered to benbsp;homologous with the disc of sporophylls of a Benettitean flowernbsp;and with the stamens in a male flower of Welwitschia. Anatomically� Ephedra exhibits a closer agreement with the Conifers andnbsp;m some respects with the Dicotyledons than with recent Cycadsnbsp;or the Bennettitales. The presence of vessels in the secondarynbsp;Xylem is an Angiospermous feature though in structure they differnbsp;from the Angiospermous type; the pitting of the tracheids is innbsp;the main Abietineous but the occurrence of compressed pits

Land (04); for other references, see Lignier and Tison (12); also Sigrianski (13) Ewart (08).

furnishes a point of contact with the Araucarineae; rims of Sanio occur and xylem-parenchyma is abundant; the medullary raysnbsp;are rnultiseriate as in Dicotyledons. The bast on the other handnbsp;is essentially gymnospermous. The occasional occurrence of spiralnbsp;bands in the tracheids and the presence of lignified trabeculae innbsp;the xylem-elements are other Coniferous traits. The leaf-trace isnbsp;double, a feature met with in Agathis as well as in recent Cycadsnbsp;but not in the Bennettitales. The anatomy of seedlings affordsnbsp;further indications of resemblance to Araucaria and the Podocarps^.nbsp;It would seem, then, that the case for a relationship between thenbsp;Gnetales and the Bennettitales founded on the facts of floralnbsp;morphology does not derive support from the anatomical featuresnbsp;of the most primitive genus of the group.

Small trees or climbers with long and slender stems; the internodes, sometimes reaching a length of 15 cm., bear pairs of ovate-oblong or lanceolate-acuminate leaves 11�18 cm. long by 4�7 cm-broad. The leaves^ agree in form and venation with those of many Dicotyledons and could not be distinguished from them innbsp;a fossil state. The epidermal cells have undulate walls. Thenbsp;flowers are in spikes; at each node two fused bracts form a cupularnbsp;structure in the axil of which the male or female flowers are bornenbsp;on an annular swelling. The male flowers� are in 3�5 whorls:nbsp;each consists of an envelope of two coherent leaves enclosing anbsp;central column, as in Ephedra, which bears at the apex one, two,nbsp;or rarely four unilocular sporangia or reduced synangia. Thenbsp;antherophore eventually elongates and pushes the anthers throughnbsp;an aperture at the summit of the floral envelope^. In appearancenbsp;the antherophore of Gnetum approaches most closely to the stamennbsp;of an Angiosperm. The female flowers� occur in a single series,nbsp;5�8 in a whorl; each consists of an ovule surrounded by threenbsp;envelopes; the outermost is coloured and succulent, the middlenbsp;envelope or outer integument is differentiated after fertilisationnbsp;into an inner sclerotesta and an outer sarcotesta, while the innermost covering is prolonged as a micropylar tube. There are no

archegonia: in some species the megaspore contains numerous free nuclei all of which are potentially sexual; after fertilisation anbsp;sterile nutritive tissue, or endosperm, is formed in the lower partnbsp;of the spore. In Gneturn Gnemon the endosperm is often formednbsp;before fertilisation. In the great reduction of the female apparatusnbsp;and in the nature of the endosperm Gneturn^ agrees much morenbsp;closely with Welwitschia than with Ephedra. Attention has beennbsp;called to certain resemblances between the seed of Gneturn and

that of Bennettites^. For convenience of comparison a diagram-static section of a seed of Gneturn africanum is reproduced in %� 817. The outermost envelope, a, forms a green succulentnbsp;Covering free to the apex; the middle envelope, h (outer integument),nbsp;ts free from the other envelopes except at the apex where it isnbsp;locked to the inner integument by the downward growth of anbsp;flange,/, from the apical region of the inner integument; the inner

integument, c, is united to the nucellus for about two-thirds of the length of the seed-body. The nucellus consists of a few layers of.nbsp;cells and at the apex forms a nucellar cap, n, the cells of which arenbsp;lignified; this cap is supported by a short tent-pole produced fromnbsp;the summit of the endosperm. A ring of vascular bundles enters-the base of the seed and forms three series, the two outer pass upnbsp;to the tip of both the two outer coverings, v, v, and the innernbsp;series extends up the inner integument as far as the level wherenbsp;nucellus and integument part company. In the seed shown innbsp;fig. 817 the micropyle is closed and the tissue in the closed regionnbsp;of the canal is lignified and dark. Among other features in whichnbsp;this seed agrees with that of Bennettites Morierei is the inner zonenbsp;of the outer integument, composed of a palisade and a fibrousnbsp;layer; the fibrous layer becomes five-angled in the upper part ofnbsp;the seed^ and in transverse section presents a striking similarity tonbsp;sections through the same region of a Bennettites seed (figs. 524, 527,nbsp;Vol. in. pp. 397, 402). The oval fleshy seeds are able to germinatenbsp;after lying some months in sea-water. There is a fairly closenbsp;resemblance between Gnetum and Ephedra as regards anatomicalnbsp;characters, but some species of Gnetum (sect. Thoa) are characterised by the formation of successive cambial cylinders as in Cycas.nbsp;Strasburger ^ pointed out that in the vascular bundles of the leavesnbsp;the parenchyma of the medullary rays forms continuous plates, anbsp;gymnospernious character.

This remarkable genus, discovered by Welwitsch in 1860 and described by Sir Joseph Hooker�, presents striking peculiarities innbsp;the habit of the vegetative body. A Welwitschia plant has beennbsp;aptly termed an'adult seedling^; the large and squat tuberousnbsp;stem, morphologically the swollen hypocotyl, may be as muchnbsp;as 4-5 m. in girth. The seedling has two cotyledons and an exceptionally long radicle: at an early stage a pair of isobilateralnbsp;leaves is produced at right-angles to the cotyledons and these

2 nbsp;nbsp;nbsp;Strasburger (91) p. 148. For an account of the anatomy of Gnetum, seenbsp;Duthie (12); La Riviere (16); Thomson, M. R. H. (16).

� Hooker, J. D. (63). For figures of Welwitschia, see also Gard. Chron. Jan. 22, p. 49, 1910.nbsp;nbsp;nbsp;nbsp;� Sykes (lO^) p. 333.

persist as the only leaves throughout the long life of the plant, attaining a length of 5 m. The tough lamina is torn into stripsnbsp;by the wind and �the extraordinary appearance of the shapelessnbsp;Hiass of coiled and twisted leaf-ribands standing out in bold reliefnbsp;from the sharp glistening dead landscape passes description^.�nbsp;The venation is parallel and there are numerous cross-connexions,nbsp;some ending blindly in the mesophylP. ' Welwitschia is dioeciousnbsp;and the flowers are borne in inflorescences with a dichasial branch-system produced from pits on the crown of the stem; the femalenbsp;Inflorescences, which are larger than the male, reach a length ofnbsp;30 cm. and bear cones about 7 cm. long. The female flowers occurnbsp;singly in the axils of bracts which form four orthostichies givingnbsp;a four-angled form to the cones. Each flower may produce twonbsp;small leaf-rudiments�, but the flower proper consists of an ovulenbsp;quot;'ith two envelopes; the outer, called by Hooker the perianth, isnbsp;considerably extended tangentially and in the ripe seed forms anbsp;quot;'ing-like appendage producing an appearance almost identicalnbsp;'quot;'ith that of some Samaropsis seeds. The inner � integument isnbsp;prolonged upwards like a long and slender hollow bristle for anbsp;'fistance of 4�5 mm. beyond the upper edge of the subtendingnbsp;bract. The inner envelope has no vascular supply. The secretionnbsp;cf sugar in the micropylar tube attracts the pollinating insectnbsp;^dontopus sexpunctulatus. The staminate cones are smaller andnbsp;fbe subtending bracts connate. The outer envelope of the flowernbsp;'S formed of two membranous segments without vascular bundlesnbsp;''�hich may be styled lateral prophylls of the axillary shoot; internalnbsp;these are two fused members forming a sac-like investmentnbsp;'^ith free rounded lobes also without a vascular supply. Withinnbsp;^bese perianth-segments is the staminal tube bearing six freenbsp;stamens each supplied with a vascular bundle and bearing anbsp;l'C�minal trilocular synangium (fig. 818). The centre of the flowernbsp;occupied by a pyriform ovule surrounded by a thin integumentnbsp;'^�ntinued as a slightly kinked stylar tube terminating in a flatnbsp;�^igmatic disc 1 mm. in diameter. There is no embryo-sac butnbsp;^be nucellus acts as a nectary, the drop-mechanism of the functional

ovule in the female flower being retained in the sterile ovule of the male^.

In contrast to the indefinite, spirally disposed, bracts or

Fig. 818. Welwitschia mirabilis. Staminate flower (sectional elevation), sub-tending bract and the two missing stamens indicated by dotted lines. (From a drawing kindly supplied by Dr A. H. Church.)

perianth of Cycadeoidea flowers Welwitschia has only two connate segments, and the staminal disc of Welwitschia is considerably

1 This account is based on the excellent description with illustrations by Dr Church (14).

reduced; the gynoecium consists of only one functional ovule instead of an indefinite number as in Cycadeoidea. Church regards the resemblance between the flowers of these two genera as an instancenbsp;of parallel development, which does not imply relationship. Henbsp;thinks there is � no indication whatever of any relation to the car-pellary flowers of the Angiosperms.�

Hooker�s account of the ovule has recently been considerably extended by the important researches of Pearson^. The megasporenbsp;consists of two regions, an upper fertile and a lower sterile portion;nbsp;each is composed of � cells � with more than one nucleus; some of thenbsp;� cells � of the fertile region grow upwards as tubes into the nucellarnbsp;cone where pollen-tubes are encountered and fertilisation ensues.nbsp;The nuclei in each �cell� of the sterile region fuse and uninucleatenbsp;cells are produced; this tissue now grows considerably in size andnbsp;cell-divisions occur resulting in the formation of an endosperm.nbsp;Pearson regards the free nuclei that are in the embryo-sac at thenbsp;time of septation into the multinucleate � cells � as all alike, and allnbsp;potential gametes. It follows, therefore, that the endospermnbsp;formed in the lower portion of the sac is the product of fusion ofnbsp;sexual nuclei; it is not a gametophyte or a sporophyte and Pearsonnbsp;proposes for it the new term trophophyte, � a bye-product resultingnbsp;from the fusion of potentially sexual nuclei and functioning in thenbsp;same manner as the prothallus of the lower seed-plants.� Morenbsp;recent work by this author confirms his opinion that the endospermnbsp;of Gnetum is also a trophophyte.

For an account of the anatomy of Welwitschia and Gnetum the student is referred to original sources. Miss Sykes^ called attention to certain interesting characters,�the occurrence of reticulatelynbsp;pitted protoxylem elements in the stem, the arrangement ofnbsp;separate and not contiguous bordered pits in 1�2 rows on thenbsp;tracheids, and to the presence of concentric steles and inverselynbsp;orientated bands of vascular tissue in the stem and inflorescencesnbsp;closely resembling Medullosean features.

2 nbsp;nbsp;nbsp;Sykes (10); (102); Worsdell (OP); Pearson (12); Bower (81); (82); Hill and denbsp;Praine (10); Boodle and Worsdell (94); Mary R. H. Thomson (16); Henriettenbsp;C. C. La Riviere (16).

the Cycads to the Angiosperms^. It has, for example, been suggested by Hallier^ that they are reduced Dicotyledons comparable with the Loranthaceae and Myxodendraceae; while Lignier andnbsp;Tison^ regard them as a group of Angiosperms nearest to thenbsp;Amentales. The question of relationship between the Gnetalesnbsp;and the Angiosperms, especially the difficult problems connectednbsp;with the endosperm, was fully considered by Pearson^ in a papernbsp;on the reproductive organs of Gnetum Gnemon published in 1915,nbsp;and in a later contribution�, published after his death, the morphological problems are reviewed in the light of more recent work.nbsp;The same subject is dealt with by Prof. Thompson� in a recentnbsp;paper in which he calls attention to the form of the inflorescence,nbsp;the arrangement of the parts of the flowers, the presence of an ovarynbsp;with a style, the germination of the microspores at some distancenbsp;from the nucellus, as evidence of affinity to the Angiosperms, andnbsp;concludes that the ancestors of the Angiosperms were �not farnbsp;removed from the genus Gnetum.�� On the other hand some botanistsnbsp;prefer to regard the Gnetales as a blindly-ending branch of Gymno-sperms with no direct relationship to the Flowering plants. Difficultnbsp;as it is to believe that plants so different, when the sum of charactersnbsp;is considered, as the Gnetales and the Bennettitales are off-shoots ofnbsp;a common stock, it would be rash to assume that such resemblancesnbsp;as have been emphasised by Miss Sykes and other authors have nonbsp;phylogenetic value.

At the time of his death (November, 1916) Professor Pearson was engaged upon a volume on the Gnetales: in April, 1916, henbsp;wrote, �A large part of the book on the Gnetales is written, though

alliance, there must be one, I think, but at present I cannot bring myself to believe that it is direct^.� Had Pearson been able tonbsp;complete his work it is certain that a statement of his most recentnbsp;conclusions would have enabled botanists to obtain a clearer view

1 Arber and Parkin (07); (08); references to other authors will be found in these papers. See also Lignier and Tison (12); Lignier (03); Lignier and Tison (11).

* nbsp;nbsp;nbsp;Pearson (15^): additional references to literature are given at the end of thisnbsp;paper. See also Caporn (16).

* nbsp;nbsp;nbsp;Pearson (17).nbsp;nbsp;nbsp;nbsp;� Thompson, W. P. (16).nbsp;nbsp;nbsp;nbsp;� Seward (17) p. ix.

of the true position of this puzzling group which, despite the lack of palaeobotanical evidence, is probably a survival from anbsp;remote past.

Arber and Parkin^^ and other authors have called attention to the lack of any trustworthy records of Gnetalean plants in thenbsp;sedimentary strata of different periods. Several specimens havenbsp;been described either as generically identical with Ephedra or asnbsp;probable representatives of the two other members of the group,nbsp;but while some are incorrectly determined others are too imperfect to be accepted as evidence. In view of the morphologicalnbsp;features characteristic of the present members of the Gnetales andnbsp;the geographical distribution of the species of Ephedra, Gnetuni,nbsp;and Welwitschia, it would seem safe to conclude that the absencenbsp;of fossil forms is not explicable on the hypothesis of a recent originnbsp;of the group, but is rather the result of the imperfection of thenbsp;geological record and of the difficulty of distinguishing betweennbsp;fragmentary remains of Gnetalean genera and vegetative or reproductive organs of similar external form belonging to other plants.nbsp;Reference has already been made^ to certain characters sharednbsp;by the seeds of Gnetuni and Bennettites and in spite of the greatnbsp;and obvious differences separating the Gnetales and Cycadales itnbsp;would seem probable that the striking similarity between thenbsp;seeds of Gnetum and those of the Bennettitales has some phylogenetic significance. But even granting a phylogenetic significancenbsp;to the evidence brought forward by Mrs Thoday and other authors,nbsp;we have still to admit that an indication of some former connexionnbsp;between the Gnetales and the Bennettitalean line is rather thenbsp;shadow of evidence with regard to the geological history of thenbsp;Gnetales and not a substantial contribution to our knowledge ofnbsp;the antiquity of this section of the Gymnosperms.

The specimens described by Unger� from Eocene beds in Styria as Ephedrites sotzkianus, though very similar to those of Ephedranbsp;fragilis with which they are compared, are too fragmentary to be

accepted as trustworthy records. The pieces of vegetative branches and the paired nuts described by Heer^- from Jurassic strata innbsp;Siberia as Ephedrites antiquus are of no botanical value. Portionsnbsp;of inflorescences preserved in amber from the Baltic coast andnbsp;named by G-oeppert and Berendt^ Ephedrites Johnianus and similarnbsp;specimens referred by Goeppert from the same Oligocene beds tonbsp;Ephedra Mengeana have been identified by Conwentz� as fragmentsnbsp;of flowering shoots of a Loranthaceous genus, Patzea. Engelhardtquot;*nbsp;refers some slender branches from Tertiar}^ beds in Chile to Ephedranbsp;but they, like most of the specimens recorded as fossil representatives of the genus, are too incomplete to be accepted as evidence.nbsp;In the absence of anatomical data or of well preserved flowers itnbsp;would be exceedingly difficult to recognise impressions of vegetativenbsp;shoots of Ephedra and to distinguish them from Dicotyledonousnbsp;twigs of similar habit. Similarly the torn lamina of a Welwitschianbsp;leaf bears too close a resemblance to other linear parallel-veinednbsp;leaves to be recognisable unless the preservation is such as to shownbsp;traces of the characteristic venation mentioned in the account ofnbsp;the recent genus. Comparisons between some fossil seeds andnbsp;the winged seeds of Welwitschia^^ though in some cases possiblynbsp;justified by actual relationship, cannot be considered to have anynbsp;importance unless supported by additional evidence. The seedsnbsp;named by Renault Gnetopsis and subsequently investigated bynbsp;Oliver and Salisbury� are now recognised as types closely alliednbsp;to Lagenostorm and other Pteridosperm seeds from Carboniferousnbsp;rocks.

In their monograph of the Pliocene Floras of the Dutch-Prussian Border Mr and Mrs Clement Reid figure under the name Gnetum scandens var. robustum'^ a piece of axis 8 mm. long andnbsp;4 mm. broad showing eight nodes bearing crowded scars of somenbsp;deciduous appendages. The authors speak of the specimen asnbsp;�a portion of a male inflorescence of a Gnetum...9,0 close to thatnbsp;of the living G. scandens that we cannot separate it.� If their

^ Goeppert and Berendt (45) A. Pis. iv., v.; Goeppert and Menge (83) A. PI. xvl-

identification is correct�and though the evidence is hardly conclusive the resemblance between the fragment from Renver and an inflorescence axis of Gnetum is undoubtedly striking�it pointsnbsp;to the occurrence in a Pliocene European flora of a genus that isnbsp;now mainly tropical and which had not so far been recognised withnbsp;any certainty in a fossil state.

The striking resemblance of Gnetum leaves to those of some Dicotyledons is an obvious difficulty in the way of the identificationnbsp;of impressions.

It is among the oldest examples of supposed Dicotyledons that search should be made for possible representatives of the genusnbsp;Gnetum. Among the earliest records of Angiosperms are thosenbsp;described by Fontaine^ from the Patuxent series of the Potomacnbsp;formation which rests on Palaeozoic crystalline rocks and containsnbsp;the remains of a flora that is clearly Jurassic or Wealden in itsnbsp;general facies; but with Jurassic Gymnosperms and Ferns arenbsp;associated some Dicotyledon-like leaves of ovate and linear formnbsp;for some of which Fontaine instituted the genera Rogersia, Fico-2)hyllum, Proteaephyllum and referred others to Ficus, Sapindopsisnbsp;etc. A revision of the Patuxent fossils by Berry^ has led to anbsp;considerable simplification in nomenclature and to the conclusionnbsp;that some at least of these Lower Potomac leaves are Gnetalean.nbsp;A comparison of some of Fontaine�s figures of Ficus virginiensis,nbsp;species of Ficophyllum, Proteaephyllum, and Rogersia with a leafnbsp;of Gnetum Gnemon reveals a very close agreement, as regards formnbsp;and venation, consistent with Berry�s suggestion. It is by no ,nbsp;means unlikely that these forerunners of the Dicotyledonous typenbsp;that occur as foreign elements in a typical Jurassic flora, withoutnbsp;an admixture of undoubted Angiosperms like those which occupynbsp;an important position in the upper beds of the Potomac formation,nbsp;may belong to plants more closely allied to Gnetum than to anynbsp;Angiosperm. Attention is especially called to the following speciesnbsp;as revised by Berry and illustrated in Fontaine�s monograph:nbsp;Ficophyllum ohlongifolium (Font.), Rogersia longifolia Font., Pro-teaephylluni ovatum Font.� It is possible that a careful study of

the venation-characters of these and other fossil leaves may lead to the discovery of criteria which may enable us to separate thenbsp;leaves of Gnetum from similar Dicotyledonous foliage.

It is with a keen sense of the incompleteness of my task that Volume IV. is concluded without any attempt to deal with thenbsp;abundant if, in very many cases, undecipherable records of Angio-sperms. The omission of this branch of Palaeobotany in whatnbsp;purports to be a general text-book calls for a word of explanation.nbsp;A mere summary of conclusions so far published with regard tonbsp;the geological history of Flowering plants would not yield resultsnbsp;commensurate with the labour involved. What is needed is anbsp;critical examination, as far as possible, of the actual specimensnbsp;and a careful scrutiny of the evidence on which determinationsnbsp;are based. It is undoubtedly the fact that a large number ofnbsp;leaf-impressions are practically valueless as trustworthy data, andnbsp;I venture to think that it is only with the cooperation of trainednbsp;systematists that any satisfactory estimate can be formed as tonbsp;the value of the fragmentary documents preserved in Cretaceousnbsp;and Tertiary strata. It is preferable to omit, at least for the present, this part of the subject than for the sake of completeness��nbsp;in a treatise that is very far from complete in its treatment of thenbsp;groups that have been considered�to essay a task for which thenbsp;author recognises that he is very inadequately equipped.

LIST OF WORKS REFERRED TO IN THE TEXT. (Volumes III. and IV.)

\_With a Jew exceptions this list does not include hooks and papers given in the Bibliographies in Volumes I. and //.]

The following are some of the Bibliographies which students will find useful for additional references:�Geological Literature added to thenbsp;Geological Society�s Library, published from time to time by the Society;nbsp;Prof. Zeiller�s lists in the �Revue G�n�rale de Botanique� (Paris); listsnbsp;given by Arber in the �Progressus Rei Botanioae� (Leiden), vol. r. Heft i.nbsp;p. 218,1907; Jongman�s �DiePalaeobotanische Literatur� (Jena), 1910-13;nbsp;also the International Catalogue of Scientific Literature (Botany andnbsp;Geology). For the Literature dealing with Cretaceous plants the studentnbsp;should refer to Dr Marie Stopes� �Cretaceous Flora,� r. and ii. (Britishnbsp;Museum Catalogues, 1913, 1915).

The dates of books published in parts given in the footnotes to this volume are as a rule those of the concluding part. For the datesnbsp;of separate parts of books relating to Palaeozoic floras the student isnbsp;referred to Prof. Zeiller�s valuable list at the end of the �Flore Fossile dunbsp;Bassin Houiller de Valenciennes.� Useful bibliographies of the writings ofnbsp;Saporta, Heer, and Ettingshausen have been compiled by Zeiller (96),nbsp;Malloizel and Zeiller (N.D.), and Krasser (97).

Aase, Hannah C. (15) Vascular anatomy of the megasporophylls of Conifers. Botanical Gazette, vol. LX. p. 277.

Affourtit, M. F. A. and H. C. C. La Rivi�re. (15) On the ribbing of the seeds of Qinkgo. Annals oj Botany, vol. xxix. p. 591.

Andersson, J. G. (10) Die Verknderungen des Klimas seit dein Maximum der letzten Eiszeit. (Collection of papers published by the Int.nbsp;Geol. Congress; edited by J. G. Andersson.) Stockholm.nbsp;Andrews, E. B. (75) Descriptions of Fossil Plants from the Coal Measuresnbsp;of Ohio. Geol. Surv. Ohio.

Antevs, E. (14) Lepidopteris Ottonis (Gopp.) Schimp, and Antholitlms Zeilleri Nath. K. Svensk. Vetenskapsakad. Hand. Bd. li. No. 7.

Mesozoischen H�lzern und das klimatische Zeugnis dieser Erschei-nungen. Geolog. Foren. Stockholm Forhand. Bd. xxxviii.

Arber, Agnes. (See also Robertson, A.) (10) On the structure of the Palaeozoic seed Mitrospermum compressum (Will.). Ann. Bof. voLnbsp;XXIV. p. 491.

Arber, E. A. Newell. (02) Notes on the Binney collection of Coal-Measure Plants. Pt. iii. The type-specimens of Lyginodendron oldhamiumnbsp;(Binney). Proc. Camh. Phil. 8oc. vol. xi. pt. iv. p. 281.

types of fronds belonging to the Cycadophyta. Trans, Linn. Soc. .vol. VII. pt. vii. p. 109.

- (12) On Psygmophyllum majus sp. nov. from the Lower Carboniferous rocks of Newfoundland, together with a Revision of the genus and Remarks on its affinities. Trans. Linn. Soc. vol. vii.nbsp;p. 391.

beds, New Zealand, collected by Mr D. G. Lillie, Biologist to Capt. Scott�s Antarctic Expedition in the �Terra Nova.� Proc. R. Soc.nbsp;vol. Lxxxvi. p. 344.

Arber, E. A. Newell and J. Parkin. (07) On the origin of Angiosperms. Journ. Linn. Soc. vol. xxxviii. p. 29.

Arnoldi, W. (01) Beitriige zur Morphologie einiger Gymnospermen. Bull. Nat. Moscow, No. 4, 1900.

Bailey, I. W. (09) The structure of the wood in the Pineae. Bot. Oaz. vol. xLvm. p. 47.

Baily, 'W. H. (69) Xotice of Plant-remains from Beds interstratified with the Basalts in the county of Antrim. Quart. Journ. Geol. Soc.nbsp;vol. XXV. p. 357.

� nbsp;nbsp;nbsp;Flora of Pi�inoe Edward Island. Canadian Bee. 8ci. vol. i. (1884-85)nbsp;p. 154.

Baker, R. T. and H. C. Smith. (10) Research on the Pines of Australia. Dpt. Public Instruction, Tech. Educ. 8er. Xo. 16. Sydney.

Bancroft, Nellie. (13) On some Indian Jurassic Gymnospernis and Bhexoxylon africanum, a new Medullosean stem. Trans. Linn. Soc.nbsp;vol. VIII. pt. ii. p. 69.

Barber, C. A. (92) On the nature and development of the corky excrescences on stems of Zanthoxylum. Ann. Bot. vol. vi. p. 155. .

Bartholin, C. T. (94) Xogle i den bornholmske Juraformation fore-kommende Planteforsteninger. Bot. Tidskrift (Copenhagen), Bd. XIX. p. 87.

Bartlett, A. W. (13) Xote on the occurrence of an abnormal bispo-rangiate strobilus of Larix europaea DC. Ann. Bot. vol. xxvii. p. 575.

Bassler, H. (16) A Cycadophyte from the Xorth American Coal Measures. Amer. Journ. Sci. vol. XLii. p. 21.

Bayer, A. (08) Zur Deutung der weihliohen Bliiten der Cupressineen nebst Bemerkungen �ber Cryptomeria. Beiheft Bot. Cent. Bd. xxiii.nbsp;Abt. I. p. 27.

Benson, Margaret. (08) On the contents of the pollen-chamber of a specimen of Lagenostoma ovoides. Bot. Gaz. vol. lv. p. 409.

Williamson), a Lower Carboniferous ovule from Pettyour, Fife-shire, Scotland. Trans. B. Soc. Edinburgh, vol. l. pt. i. No. i. p. 1.

Benson, M. andE. J. Welsford. (09) The morphology of the ovule and female flower of Juglans regia and of a few allied genera. Ann. Bot.nbsp;vol. XXIII. p. 623.

Berger, R. (48) De fructibus et seminibus ex formatione lithanthracum. Diss. Inaug. Vratislaviae.

Berridge, E. M. (11) On some points of resemblance between Gnetalean and Bennettitean seeds. New Phyt. vol. x. p. 140.

Berridge, Emily M. and Elizabeth Sanday. (07) Oogenesis and embryo-geny in Ephedra distachya. New Phyt. vol. vi. p. 127.

Berry, E. W. (03) The Flora of the Matawan Formation. Bull. New. York Bot. Gard. vol. iii. No. 9, p. 45.

- nbsp;nbsp;nbsp;(09^)nbsp;nbsp;nbsp;nbsp;Pleistocene Swamp deposits in Virginia. Amer. Nat.

- nbsp;nbsp;nbsp;(10-)nbsp;nbsp;nbsp;nbsp;The epidermal characters of .FreMelo/),si5nbsp;nbsp;nbsp;nbsp;Bot.

- nbsp;nbsp;nbsp;(10^)nbsp;nbsp;nbsp;nbsp;Additions to the Pleistocene Flora of New Jersey. Torreya,

- nbsp;nbsp;nbsp;(lOquot;*)nbsp;nbsp;nbsp;nbsp;Contributions to the Mesozoic Flora of the Atlantic Coastal

p. 181.

�� nbsp;nbsp;nbsp;(11*)nbsp;nbsp;nbsp;nbsp;A Revision of several genera of Gym nospermous plants from

the Potomac group in Maryland and Virginia. Proc. U. 8. Nal, Mus. vol. XL. p. 289.

- (12) The age of the plant-bearing shales of the Richmond coalfield. A.mer. Journ. 8ci. vol. xxxiv. p. 224.

Plain. XII. Arkansas. Bull. Torr. Bot. Club, vol. xliv, p. 167. Bertrand, 0, E. (74) Anatomie compar�e des tiges et des feuilles oheznbsp;les Gn�tac�es et les Contf�res. Ann. 8ci. nat. [v], vol. xx. p. 5.

- nbsp;nbsp;nbsp;(98) Remarques sur la structure des grains de pollen de Cordate.

- nbsp;nbsp;nbsp;(07) Les caract�ristiques du genre Rhabdocarpus d�apr�s les

pr�paraiions de la collection B. Renault. Buil. soc. bot. [4J, Tome vii, p. 654.

---- nbsp;nbsp;nbsp;(07^) Les caract�ristiques du genre Diplotesta de Brongniart.

- nbsp;nbsp;nbsp;(07^) Les caract�ristiques du genre Leptocaryon de Brongniart.

- nbsp;nbsp;nbsp;(08^) Les caract�ristiques du genre Cardiocarpus d�apr�s les graines

Bertrand, P. (11) Structure des stipes d�AsterocMaena laxa Sterz.el. M�m. soc. g�ol. Nord, Tome vii. i.

et � Steloxylon.� Compt. Bend, de I�assoc, frang. pour 1'Arancement des Sciences (Havre, 1914), p. 446.

Beust, F. (85) Untersuchung fiber fossile Holzer aus Gr�nland. Ne-ue Denksch. allgem. Schweiz. Oes. gesammt. Naturiuiss. Bd. xxix.

Binney, E. W. (66) On Possil wood in calcareous nodules found in the upper foot coal near Oldham. Proc. Lit. Phil. Soc. Manchester,nbsp;vol. V. p. 113.

Bleicher and Fliche. (92) Contribution a 1��tude des Terrains Tertiaires d'Alsace. Bull. soc. g�ol. France [3], Tome xx. p. 375.

Bodenbender, W. (96) Beobaohtungen fiber Devon- und Gondwana Schichten in der .Argentinischen Republik. Zeitsch. Deutsch. geol.nbsp;Oes. Bd. XLViii. p. 743.

Boodle, L. A. (15) Concrescent and solitary Poliage leaves in Pinus. New Phyt, vol. xiv. p. 19.

Boodle, L. A. and W. C. Worsdell. (94) On the comparative anatomy of the Casuarineae, with special reference to the Gnetaceae andnbsp;Cupuliferae. Ann. Bol. vol. vili. p. 231.

Boulay. (79) Recherches de paleontologie v�g�tale dans le terrain houiller du Nord de la Prance. Ann. soc. scient. Bruxelles, ann. iv.nbsp;pt, 2, 1880.

Bower, F. 0. (81) On the germination and histology of the seedlings of Welwitschia mirahilis. Qmrt. Journ. Mier. .soc. vol. xxi. pp. 15, 571.

and its relation to the Cyatheoideae and other Ferns. Ann, Bot. vol. XXVI. p. 269.

Bower bank, J. S. (40) History of the Fossil fruits and seeds of the London Clay. London.

Braun, A. (75) Die Prage nach der Gymnospermie der Cycadeen. Monatsber. K. Preuss. Akad. Wiss. Berlin, p. 289.

Sraun, C. F. W. (43) Beitrage zur Petrefactenkunde Bayreuth (Graf zu M�nster), Heft vi. Bayrevih.

Lias und Keuper des neu aufgefundenen Pflanzenlagers in dem Steinbr�che von Veitlahm bei Culmbach. Flora, p. 81.

eine neue Gattung fossiler Rhizantheen. Progr. Hi. Jahresber. K. Kreis-Landwirlhsch. und Gewerhschule zu Bayreuth.

Brauns, D. (66) Her Sandstein bei Seinstedt unweit des Fallsteins und die in ihm vorkommenden Pflanzenreste. Paleont. Bd. ix.nbsp;p. 47.

Brenchley, Winifred E. (13) On Branching specimens of Lyginodendron Oldkamium Will. Journ. Linn. Soc. vol. xni. p. 349.

Bristow, H. W. (62) The Geology of the Isle of Wight. Mem. Geol. Surv. Great Britain.

Brongniart, A. (25) Observations sur les v�g�taux fossiles renform�s dans les gr�s de Hoer en Scanie. Ann. Sci. nat. vol. iv. p. 200.

- nbsp;nbsp;nbsp;(74) �tudes sur les graines fossiles trouv�es a l��tat sihcifi� dans

Bronn, H. G. (58) Beitrage zur triassischen Schiefer von Raibl. Neues �Jahrb. Min. p. 129.

Brooks, F. T. and A. Sharpies. (14) Pink disease. Bull. No. 21, Depart. Agric. Fed. Malay States.

Brooks, F. T. and W. Stiles. (10) The stracture ol Podocarpusspinulosus (Smith) R. Br. Ann. Bot. vol. xxiv. p. 305.

Buchman, J. (45) Outline of the Geology of the neighbourhood of Cheltenham (in collaboration with R. I. Murchison and H. E.nbsp;Strickland). London.

Buckland, W. (28) On the Cyoadeoideae, a Family of FossilPlants found in the Oohte quarries of the Isle of Portland. Trans. Geol. Soc. [2],nbsp;vol. II. p. 395.

Burckhardt, C. (11) Bemerkungen zu einigen Arbeiten von W. Gothan und A. G. Nathorst. Cent. Min. Geol.; Paleont. p. 442.

Burgestein, A. (06) Zur Holzanatomie der Tanne, Fichte und Larche. Ber. deutsch. Bot. Ges. Bd. xxiv. Heft vi. p. 295.

Burlingame, L. (08) The staminate cone and male gametophyte of Podocarpus. Bot. Oaz. vol. xlvi. p. 161.

Burlingame, L. (13) The Morphology of Araucaria brasiliensis. Bot. Qaz. vol. LV. p. 97.

Butterworth, J. (97) Some further investigations of Fossil seeds of the genus Lagenostoma (Williamson) from the Lower Coal MeasureSrnbsp;Oldham. Mem. Proc. Manchester Lit. and Phil. Soc. vol. xli. ix. p. 1.

Caldwell, 0. W. and C. F. Baker. (07) The identity of Microcycas calocoma. Ibid. vol. xliii. p. 130.

Cambier, B. et A. Renier. (10) Psygmofhyllum Delvali n. sp. du Terrain houiller de Charleroi. Ann. soc. g�ol. Belg. Tome ii. p. 23. (M�m.nbsp;in Alo.)

Capellini, G. and Conte E. Solms-Laubach. (92) I Tronchi di Bennettitee dei Musei Italian!. Mem. Reale Acad. Sci. Istit. Bologna [5],nbsp;tom. II. p. 161.

Caporn, A. St. C. (16) A note on the male inflorescence of a species of Onelum from Singapore. Ann. Bolus Herb. vol. ii. pt. i.

Carano, E (04) Contribuzione alia conoseenza della Morfologia e dello sviluppo del fasoio vasoolare delle foglie delle Cicadacee. Ann. dinbsp;Bot. vol. I. p. 109 (Rome).

Carpentier, A. (11) Sur quelques fructifications et inflorescences du Westphalien du Nord de la Fi-ance. Rev. O�n. Bof. tome xxiii. p. 1.

Carruthers, W. (66) On Arauoarian cones from the Secondary beds of Britain. Geol. Mag. vol. iii. p. 249.

- nbsp;nbsp;nbsp;(69) On Beareia, a new genus of Cycadean Fruit from the Yorkshire

Carruthers, W. (93) On Gycas Taiwaniana, sp. nov. and C. Seemanni R. Br. Journ. Bot. vol. xxxi. p. 1.

Carter, M. Geraldine. (11) A Reconsideration of the origin of Trans-fusion-tissue. Ann. Bot. vol. xxv. p. 975.

K. Breuss. Geol. Landesanstalt, Bd. ix. Heft iii. p. 113. Celakovsky, L. (82) Zur Kritik der Ansichten von den Pruchtsohuppenbsp;der Abietineen. Ahh. K. h�hm. Ges. Wiss. Brag [vi], Bd ll.nbsp;Chamberlain, C. J. (06) The ovule and female gametophyte of Dioon.nbsp;Bot. Gaz. vol. XLii. p. 321.

Church, A. H. (14) On the Floral Morphology of Welwitschia mirabilis (Hooker). Bhil. Trans. R. Soc. vol. 205, p. 115.

Cockerell, T. D. A. (06) The Fossil Flora and Fauna of the Florissant (Colorado) shales. Univ. Colorado Series, vol. in. No. 3.

Coemans, E. (66) Description de la flore fossile du premier �tage du terrain Cr�tac� du Hainaut. M�m. Acad. R. Belg., tomenbsp;xxxvr.

Coker, W. C. (03) On the gametophyte and embryo of Taxodium. Bot. Gaz. vol. XXXVI. p. 1.

� (09) Vitality of Pine seeds and the delayed opening of cones. Amer. Nat. vol. XLm. p. 677.

s. IV nbsp;nbsp;nbsp;31

Conwentz, H. (78) Ueber ein tertiilres Vorkommen Cypressenartiger H�lzer bei Calistoga in Californien. Nenes Jahrb. Min. p. 800.

anstalt zu Berlin. Jahrb. K. preuss. geol. Land. BergaJcad. Berlin fur das Jahr 1881, p. 144.

- nbsp;nbsp;nbsp;(85) Sobre algunos arboles fosiles del Bio Negro. Bol. Acad. Nac.

Coulter, J. M. and C. J. Chamberlain. (03) The Embryogeny of Zamia. Bot. Oaz. vol. XXXV. p. 184.

Cramer, C. (68) Fossile H�lzer der Arctisohen Zone. Heer�s Foss. Flor. Arct. vol. I. p. 167.

Cri�, L. (89) Beitrage zur Konntniss der fossilen Flora einiger Inseln des Siidpacinsohen und Indisohen Oceans. Pal. Abhand. (Dames andnbsp;Kayser) (N.F.), Bd. i. Heft ii.

Daguillon, A. (90) Recherches morphologiques sur les feuilles des Conif�res. Rev. G�n. Bot. tome 11, p. 154.

Dawson, J. W. (46) Notices of some Fossils found in the Coal Formation of Nova Scotia. Quart. Journ. Geol. Soc. vol. ii. p. 132.

characters and affinities of Palaeozoic Gymnosperms. Canadian Rec. Sci. vol. iv. p. 1.

Dawson, Sir J. W. and D. P. Penhallow. (91) Note on the specimens of Fossil wood from the Brian (Devonian) of New York and Kentucky-The Canadian Rec. Sci. vol. iv. p. 242.

Depape, G. (13) Sur la pr�sence du Ginkgo biloba L. dans le Plioc�ne 'nf�iieur de Saint-Maroel-d�Ard�che. Compt. Rend. vol. 157, p. 9�7-

Depape, G. and A. Carpentier. (13) Presence des genus Gnetopsis B. Ren. and R. Zeill. el Urnafopteris Kidst. dans le Westphalien dunbsp;Nord de la France. Ann. soc. geol. du Nord, tome xlii. p. 294.

Dorety, Helen A. (08) The embryo of Oeratozamia, a physiological study. Bot. Gaz. vol. XLV. p. 412.

- (09) The extrafascicular cambium of Ceralozamfa. 76id. vol. XLVir.

Diimmer, R. (12) Podocarpus formosensis. Gard. Chron. Oct. 19, p. 296. Dun, W. S. (10) Notes on some Fossil plants from the roof of the coalnbsp;seam in the Sydney Harbour Colliery. Journ. Proc. R. Soc. Newnbsp;South Wales, vol. xliv. p. 615.

Duns, J. (72) On Cardiocarpon. Proc. R. Soc. Edinburgh, p. 692. Dusdnek, F. (13) Spalt�ffnungen der Cycadaceen. (Abstract in thenbsp;Bot. Cent. Bd. exxv. p. 340.)

Dus�n, P. (99) �ber die Tertiare Flora der Magellans-Lander. IFiss. Ergeb. Schwed. Exped. nach den Magella-n.iland.ern unter Leitung vonnbsp;O. Nordenshj�ld, Bd. l. No. iv. p. 87.

Duthie, Augusta V. (12) Anatomy of Gnetu-m africanmn. Ann. Bot. vol. xxvr. p. 593.

Dutt, C. P. (16) Pityostrobus macrocephalus,Jj. andH. A Tertiary Cone showing Ovular structures. Ann. Bot. vol. xxx. p. 529.

Eames, A. J. (13) The Morphology of Agathis australis. Ann, Bot. vol. XXVII. p. 1.

Eichler, A. W. (81) �ber die weiblichen Bliithen der Coniferen. Monats-ber. K. Akad. W'iss. Berlin, p. 1020.

Elkins, Marion G. and G. R. Wieland. (14) Cordaitean wood from the Indiana Black Shale. Arner. Journ. Sci. vol. xxxviii. p. 65.

Elwes, H. J. and A. Henry. (06) The trees of Great Britain and Ireland, vol. I. Edinburgh.

Endlicher, S. (40) Genera plantarum secundum ordines naturales disposita. Vindohonae, 1836-40.

EngelhaCrdt, H. (85) Die Tertiarflora des Jesuitengrabens beiKundratitz in Nordbohmen. Nov. Act. K. Leop.-Gar. Deutsch. Akad. Nalur-Jorsch. Bd. xlviii. No. 3, p. 299.

Engelhardt, H. and F. Kinkelin. (08) Oberphooene Flora und Fauna des Unter-Maintales. Abh. Senck. Naturforsch. Ges. Bd. xxix.nbsp;Heft iii. p. 151.

Eugler, A. (89) Engler and Prantl; Die Natiirlichen Pflanzenfamilien, Teil II. Abt. i.

Essner, B. (86) Ueber den diagnostisohen Werth der Anzahl und H�he der Markstrahlen bei den Coniferen. Abh. naturforsch. Ges. Halle,nbsp;Bd. XVI. p. 1.

Etheridge, K. (93) On the occurrence of a plant allied to Bchizoneura in the Hawkesbury Sandstone. Rec. Oeol. Surv. New South Wales,nbsp;vol. III. pt. iii, p. 74.

Ettingshausen, C. von. (51) Ueber einige neue und interessante Taeraio-pteris Arten au.s den Sammlungen des Kais. Hof Mineralien Cabinotes und der K.k. geol. Beiohsanstalt. Naturwiss. Abh. W.nbsp;Haidinger, vol. iv. p. 95.

��^ (67) Die Kreideflora von Niedersohoena in Sachsen. nbsp;nbsp;nbsp;Bd. lv.

Ettingshausen, C. von. (70) Beitrag zur Kenntniss der Tertiarflora Steiermarks. Sitz. K. Alcad. Wien, Bd. lx. Abth. i. p. 17.

Ewart, A. J. (08) On the longevity of seeds. Proc. R. Soc. Victoria, vol. XXI. f-^'.S.], pt. i. p. 1.

Feistmantel, 0. nbsp;nbsp;nbsp;(72) Beitrag zur Kenntniss der Ausdehnung des sogen-

nanten Wyraner Gasschiefers und seiner Flora. Jahrb. K.k. geol Reichs. Wien, Bd. xxii. p. 289.

- nbsp;nbsp;nbsp;(77^) Jurassic (Liassic) Flora of the Rajmahal group from Golapili,

Feistmantel, 0. nbsp;nbsp;nbsp;(80^) Note on the fossil generanbsp;nbsp;nbsp;nbsp;Sternberg,

Noeggerathiopsis Feist, and Rhiptozamites Sohmal. Rec. Oeol. Surv. India, voL xiii. pt. i. p. 61.

Verhaltnisse Siid-Afrikas. Th. i. Ahh. K. b�lim. Ges. W�iss. fvii], Bd. III.

Felix, J. (82) Beitrage zur Kenntniss fossiler Coniferen-H�lzer. EngleFs Jahrb. Bd. iii. p. 260.

Fliche, P. (96) �tude sur la flore fossile de I�Argonne (Albien-Ceno-manien). Bull. soc. sci. Nancy.

les-vaudes (Aube) dans les gr�s verts infraor�tac�s. Alem. soc. Acad, de I Aube, tome lx.

Fliche, P. and Bleicher. (82) �tude sur la flore de I�oolithe inf�rieure aux environs de Nancy. Bull. soc. sci. Nancy.

Fliche, P. and R. Zeiller. (04) Note sur une florale Portlandienne des envdi'ons de Boulogne-sur-mer. Bull. soc. geol. France [4], tome iv.nbsp;p. 787.

Fontaine, W. M. (93) Notes on some Fossil plants from the Trinity Division of the Comanche series of Texas. Proc. U. 8. Nat. Mm.nbsp;vol. XVI. p. 261.

Foxworthy, F. W. (11) Philippine Gymnospenns. Philipp. .Journ. 8ci. (C) Botany, vol. vr. No. 3, p. 149.

Fraine, E. de. (12) On the structure and affinities of Sutcliffia, in the light of a newly discovered specimen. .Ann. Bot. vol. xxvi.nbsp;p. 1031.

Fritel, P. H. and R. Viguier. (11) �tude anatomique de deux hois �oc�nes. Ann. sci. nat. [9], tome xiv. p. 63.

Fujii, K. (96) On the different views hitherto proposed regarding the morphology of the dowel�s of Ginlgo hiloha. Bot. Mag. Tokyo,nbsp;vol. X. No. 109, p. 13.

Fujioka, M. (13) Studi�n �ber den anatomischen Ban des H�lzes der japanischen Nadelbaume. Journ. Coll. Agric. Imp. Univ. Tokyo,nbsp;vol. IV. No. 4, p. 201.

Gardner, J. S. (86) A monograph of the British Eocene Flora, vol. ii. Gymnospermae. Palaeont. Soc. London.

of reporting on the ros.sil Plants of the Tertiary and Secondary beds of the United Kingdom. Rep. of the 56th Meeting Brit. Assoc.

Geinitz, H. B. (42) Charakteristik der Schichten und Petrefacten des Sachsischen Kreidegebirges. Heft ill. Dresden und Leipzig.

- nbsp;nbsp;nbsp;(63) �berzweineueDyadisohePflanzen. Aeiies Jalirfe. AH�, p. 525.

- nbsp;nbsp;nbsp;(75) �ber neue Aufschliisse iin Brandsohiefer der unteren Dyas

Gerry, E. (10) The distribution of the Bars of Sanio in the Coniferales. Ann. Bot. vol. xxiv. p. 119.

Geyler, T. and F. Kinkelin. (90) Oberpliocan Flora aus den Baugraben des Klarbeckens bei Niederrad etc. Abh. Senck. Naturforsch. Ges-Bd. XV. p. 1.

Gibbs, L. S. (12) On the Development of the female strobilus in Podo-carpus. Ann. Bot. vol. xxvi. p. 515.

Gl�ck, H. (02) Eine fossile Fichte aus dem Neokertahl. Mitt. Orossh. Bad. geolog. Landesanst. Bd. iv. Heft iv. p. 399.

Goc, M. J. le. (14) Observations on the centripetal and centrifugal xylems in the petioles of Cycads. Ann. Bot. vol. xxvill. p. 183.

Goeppert, H. R. (40) Uber die neulichst im Basalttuff des hohen Seelbach-kopfes bei Siegen entdeckten bituminosen und versteinerten Holzer, so wie ttber die der Braunkohlenformation �berhaupt. Arch, fiirnbsp;Min., Oeog., etc. (Karsten und von Declcen), Bd. xiv. p. 182.

die in Schlesien vorkommenden Arten. Uebersicht Arbeit, und Verand. ScUesisch. Ges. vat. Kult. 1843, p. 114. Breslau.

Tchihatcheff en Sib�rie. Voyage scientifique dans I�Altai oriental, p. 379. Paris.

Oberschlesien. JJebers. Arbeit, und Verand. Schles. Ges. vat. Kultur im Jahre 1845, Breslau, 1846, p. 139.

- (50) Monographie der fossilen Coniferen. Naturwerkundige Verhand. Holland. Maatschap. Wettenschappen Haarlem,. Leiden.

Schlesien, so wie �ber fossile Cycadeen. Denhsch. Schles. Ges. fiir Vaterland. Kultur.

Coniferen, insbesondere der Araucariten, und �ber die Descendenz-lehre. Bot. Cent. Bd. v., vi. p. 378.

Goeppert, H. R. and G. Stenzel. (81) Die Medulloseae. Palaeontol. [N.F.], Bd. VIII. p. 113.

Gomes, B. A. (65) Vegetaes Fosseis. Flora fossil do Terrens Carbonifero das visinlian9as dio Porto, serra do Bussaio, e moinho d�ordemnbsp;proximo a alcacer do Sal. Comm. geol. Portugal. Lisbon.

Gordon, Marjorie. (12) Ray-tracheids in Sequoia aempervi-rens. New Phyt. voi. XI. p. 1.

Gordon, W. T. (10) On a new species of Physostoma from the Lower Carboniferous rocks of Pettycur (Fife). Proc. Camb. Phil. Soc.nbsp;vol. XV. pt. V. p. 395.

the Calciferous sandstone series. Trans. R. Soc. Edinb. vol. XLvm. pt. iv. p. 813.

Gothan, W. (05) Zur Anatomie lebender und fossiler Gymnospermer-H�lzer. Abh. K. Prenss. geol. Landes. [N.F.], Heft XLiv. p. 1.

men im Laufe der geologischen Epochen und ihre physiologische Bedeutung. Sitz. Ges. Naturforsch. Freunde, No. 2, p. 13.

ll'iss. Ergeb. Schwedis. Siidpolar-Exped. 1901-03, Bd. iii. Lief. viii. Stockholm.

Kreideformation im Lichte palaobotanischen Tatsachen. Jahrb. K. Preuss, Geol. Landes, fiir 1908, Bd. xxix. Th. ii. Heft 2, p. 220.

- nbsp;nbsp;nbsp;(09) �ber Braunkohlenh�lzer des rheinischen Tertiars. Jahrb.

- (10) Die Fossile Holzreste von Spitzbergen. K. Svensk. Vetenskapsakad. Hand. Bd. xlv. No. viii.

Gourlie, W. (44) Notice of the Fossil Plants in the Glasgow Museum. Proc. Phil. Soc. Glasgow, vol. r. 1844, p. 105.

Grand�Eury, C. (00) Sur les tiges debout, les souches et racines de Cordaites. Compt. Rend, tome cxxx. (April 30).

donospermum et sur l�extr�me vari�t� des � graines de foug�res.� Ibid. p. 812.

la v�g�tation et la flore houill�res, en deux parties et dix livraisons. Pt. I. Livr. ii., Paris et Li�ge.

Graner, F. (94) Die geographische Verbreitung der Holzarten. I. Die Coniferen. Forstwissenschaft. Centralhlatt, Berlin (August).nbsp;Griffith, W. (59) Remarks on Gnetum. Trans. Linn. Soc. vol. xxii.nbsp;pt. iv. p. 299.

Groom, P. (10) Remarks on the Oecology of Coniferae. Ann. Bof. vol. XXIV. p. 241,

Groom, P. and W. Rushton. (13) The structure of the wood of East Indian species of Pinus. Journ. Linn. Soc. vol. xli. p. 457.nbsp;Groppler, R. (94) Vergleichende Anatomie des H�lzes der Magnoliaceen.

Grossenhacher, J. G. (15) Medullary spots and their cause. Bull. Torr. � Bot. Club, vol. XLii. p. 227.

Guppy, H. B. (06) Observations of a Naturahst in the Pacific between 1896 and 1899. London.

Gutbier, A. von. (49) Die Versteinerungen des Zechste�igebirges und Rothliegenden oder des Permisohen Systems in Sachsen. Dresdennbsp;and Leipzig.

Halle, T. G. (10) A Gymnosperm with Cordaitean-like leaves from the Rhaetic beds of Scania. Arkiv for Bot. Upsala, Bd. ix.nbsp;No. 14.

Tierra del Fuego and their Floras. K. Svensh. Vetenskapsakad. Hand. Bd. li. No. 3.

Hallier, H. (05) Provisional schemequot; of the Natural (Phylogenetic) system of Flowering Plants. New Phyt. vol. iv. p. isi.

Harker, A. (06) The Geological structure of the Sgurr of Eigg. Quart. Journ. Geol. Soc. vol. lxii. p. 40.

Harshberger, J. W. (98) Water-storage and conduction in Senecio -praecox DC. from Mexico. Contrib. Bot. Labt. {Univ. Pennsylvania), vol. II. No. 1.

Hayata, A. (06) On Taiwania, a new genus of Coniferae from the Island of Formosa. Journ. Linn. Soc. vol. xxxvii. p. 330.

Heer, 0. (62) On the Fossil Flora of Bovey Tracey. Phil. Trans. B. Soc. vol. CLii. p. 1039.

land, an Mackenzie, im Island und in Spitzhergen entdeckten fossilen Pfianzen. Flor. Foss. Arct. vol. i. Zurich.

- nbsp;nbsp;nbsp;(75) ii. Die Kreide-Flora der arctischen Zone. Flor. Foss. Arct.

V. Beitrage zur Miocenen Flora von Sachalin. Ibid. Contributions a la Flore du Portugal. Sext. Trav. Oeol. Port.nbsp;(Lisbon).

(81^) Zur Geschichte der Ginkgo-artigen Baume. Bugler's Bot. Jahrb. Bd. i. p. 1.

Herzfeld, S. (10) Die Entwicklungsgeschichte der weiblichen Bliite von Cryptomeria japonica Don. Ein Beitrag zur Deutung der Erucht-schuppe der Coniferen. Sitzber. Akad. Wiss. Wien, Bd. cxix.nbsp;Abt. i. p. 807.

Hick, T. (95) On KaloxylonHookeri Will, azad Lyginodendron oldhamium Will. Mem. Proc. Manchester Lit. Phil. Soc. [4], vol. ix. p. 109.nbsp;Hilderbrand, F. (61) Die Verbreitung der Coniferen. Rhein, undnbsp;Westphal. Verhand. Bd. xviii. p. 199.

Hill, T. G. and E. de Fraine. (10) On the seedling structure of Gymno-sperms. IV. Ann. Bot. vol. xxiv. p. 319.

Hirase, S. (98) �tudes sur la f�cundation et I�embryogenie du Ginkgo biloba. Journ. Coll. Sci. Imp. TJniv. Tokyo, vol. xir. p. 103.nbsp;H�hlke, F. (02) Ueber die Harzbehalter und die Harzbildung bei dennbsp;Polypodiaceen und einigen Phanerogamen. Beiheft Bot. Cent.nbsp;Bd. XI. p. 8.

Holden, Ruth.. (13) Some fossil plants from Eastern Canada. Ann. Bot. vol. xxvir. p. 243.

�� (15�) On the cuticles of some Indian Conifers. Bot. Oaz. vol. lx. p. 215.

Hollick, A. (97) The Cretaceous clay marl exposed at Cliffwood, N. J. Trans. N. Y. Acad. Sci. vol. xvi. p. 124.

on I.ong Lsland. Bull. N. Y. Bot. Oard. vol. viii. No. 28, p. 154. Hollick, A. and E. C. Jeflfrey. (06) Affinities of certain Cretaceous plant-remains commonly referred to the genera Dammara and Brachy-phyllum. Amer. Nat. vol. XL. p. 189.

Holmes, W. H. (78) Fossil Forests of the Volcanic Tertiary formations of the Yellowstone National Park. Ann. Sep. Oeol. Oeogr. Surv.nbsp;U.S.A. pt. n. p. 47.

Hooker, J. D. (52) Dacrydium laxifolium. Icone.s Plant, vol. v. pi. 815. - (60) Flora Tasmanica. London.

Hooker, J. D. and E. W. Binney. (55) On the structure of certain limestone nodules enclosed in seams of bituminous coal, with anbsp;description of some Trigonocarpons contained in them. Phil. Trans.nbsp;R. SoC. vol. CXLIX.

H�rich, 0. (06) Potoni�s Abbildungen und Beschreibungen fossilen Pllanzen-Reste. Lief. iv. 69, 70.

Howse, R. (88) A catalogue of Fossil Plants from the Hutton collection. Nat. Hist. Trans. Northumberland, Durham, and Newcastle-upon-Tyne, vol. X.

Hutchinson, A. H. (14) The male gametophyte of Abies. Hot. Oaz. vol. i.vii. p. 148.

Jeffrey, E. C. (03) The comparative anatomy and phylogeny of the Conifers. I. The genus Sequoia. Mem. Boston Soe. Nat. Hist.nbsp;vol. V. No. 10, p. 441.

Jeffrey, E. C. (14) Spore-conditions in hybrids and the mutation hypothesis of de Vries. Bot. Gaz. vol. LViii. p. 322.

Jeffrey, E. C. and M. A. Chrysler. (06) On Cretaceous nbsp;nbsp;nbsp;Bot.

Jeffrey, E. C. and Ruth D. Cole. (16) Experimental Investigations on the genus Drimys. Ann. Bot. vol. xxx. p. 359.

Jeffrey, E. C. and R. E. Torrey. (16) Ginkgo and the mierosporangial mechanisms of the seed plants. Bot. Gaz. vol. lxh. p. 281.

Johnson, T. (11) A seed-bearing Irish Pteridosperm, Crossotheca Honinghausi Kidst. Sci. Proc. R. Dublin Soc. vol. xiil. p. 1.

sperms from the Upper Devonian beds at Kiltoscan, Co. Kilkenny. Ibid. vol. XV. p. 245.

Johnston, R. H. (86) Fresh contributions to our knowledge of the Plants of Mesozoic age in Tasmania. Papers and Proc. R. Soc.nbsp;Tasmania for 1886, p. 160.

Johnstrup, M. F. (83) Recherches sur les fossiles appartenant aux formations Cr�tao�e et Mioc�ne, sur la c�te occidentale du Grpnland.nbsp;Medd. om Qnfnland, vol. v.

Jones, W. S. (12) The structure of the Timbers of some common genera of Coniferous trees. Quart. Journ. Forestry, April.

Karsten, G. (92) Beitrag zur Entwickelungsgeschichte einiger Gnetum Arten. Bot. Zeit. p. 205.

Kershaw, E. M. (09) The structure and development of the ovule of Myrica Gale. Ann. Bot. vol. xxiii. p. 353.

Kidston, R. (84) On a new species of Schiitzia from the Caloiferous sandstone of Scotland. Ann. Mag. Nat. Hist. vol. xiii. p. 77.

Airdrie, from the Lanarkshire coal-field. Trans. Geol. Soc. Glasgow, vol. VIII. p. 47.

Kidston, R. (90) The Yorkshire Carboniferous Flora. Trans. Yorks. Nat. Union, pt. xiv.

^� (04^) Some Fossil Plants collected by Mr A. Sinclair from the Ayrshire coalfield. Kilmarnock Olenfield Ramhlers Soc. Annalsnbsp;(1901-04), No. iv. Kilmarnock.

organic connection with the Foliage of Lyginodendron. Proc. B. Soc. vol. Lxxvi. p. 358.

The Fossil Flora of the Westphalian series of the S. Staffs, coalfield. Trans. R. Soc. Edinb. vol. l. pt. i. p. 73.

Kidston, R. and D. T. Gwynne-Vaughan. (12) On the Carboniferous Flora of Berwickshire. Pt. i. Stenomyelon tuedianum Kidst.nbsp;Trans. R. Soc. Edinb. vol. XLViii. pt. ii. p. 263.

Kidston, R. and W. J. Jongmans. (11) Sur la Fructification de Aotj-o-pteris obliqua Brongn. Arch. Neerl. sci. exact, nat. [in. B], tome r. p. 25. Kirby, J. W. (64) On some remains of Fishes and Plants from thenbsp;�Upper limestone� of the Permian series of Durham. Quart. Jmirn.nbsp;Oeol. Soc. vol. XX. p. 349.

Kirchner, 0., E. T. Loe-w, and 0. Schr�ter. (06) Die Coniferen und Gnetaceen Mitteleuropas. Stuttgart.

Kirsch, S. (11) The Origin and Development of resin-canals in the Coniferae, with special reference to the Development of Thylosesnbsp;and their correlation with the Thylosed strands of the Pteridophytes.nbsp;Trans. R. Soc. Canada, sect. iv. p. 43.

Kisch, Mabel H. (13) The Physiological Anatomy of the periderm of fossil Lycopodiales. Ann. Bot. vol. xxvii. p. 281.

Kleeberg, A. (85) Die Markstrahlen der Coniferen. {Inaug. Biss.) Bot. Zeit. Bd. xnm.

Klein, L. (81) Bau und Verzweigung einiger dorsiventral gebaute Polypodiaceen. Nov. Act. K. Leop. Gar. Beutseh. Ahad. Natur-forsch. Bd. xi.ii. No. 7, p. 335.

Knowlton, F. H. (89) Description of the fossil woods and lignites from Arkansas. Ann. Bep. Oeol. Surv. Arkansas, vol. n. p. 249.

compiled descriptions and partial synonymy of the species. Proc. U. S. 3Ius. vol. xn. p. 601.

Knowlton, F. H. (93) The Laramie and the overlying Livingstone formation in Montana. Report on the Flora. Bull. �. 8. Oeol.nbsp;Surv. No. 105.

Kny, L. (10) Uber die Verteilung des Holzparenchyma bei Ahies pectinata DC. Ann. Jard. Bot. Buit. [2], Suppl. iii. p. 645.

Koettlitz, R. (98) Observations on the geology of Franz Josef Land-Quart. Journ. Oeol. Soc. vol. liv. p. 620.

Kosmovsky, C. (92) Quelques mots sur les couches a v�g�taux fossiles dans la Russie ori�ntale et en Sib�rie. Bull. soc. Imp. Nat. Moscounbsp;[N.S.], tome V. p. 170.

Kramer, A. (85) Beitrage zur Kenntniss der Entwiokelungsgeschichte und des anatomischen Baues der Fruchtblatter der Cupressineen undnbsp;der Plaoenten der Abietineen. Flora, xliii. p. 519.

Krasser, F. (91) �ber diefossile Flora derrhatisohen Schichten Persiens-Sitzber. K. Ahad. Wiss. Wien, Bd. o. Abth. i. p. 413.

Flora von Ouricanga in Brasilien. Sitzber. K. Ahad. Wiss. Wien,. Bd. oxn. Abt. i. p. 852.

Kraus, G. (64) Mikroskopische Untersuchungen �ber der Bau lebenden Nadelh�lzer. W�rzb. Naturwiss. Zeitsch, Bd. v. p. 144.

KrS,usel, R. (13) Beitrage zur Kenntniss der H�lzer aus der Schlesischen Braunkohle. Teil I. (Inaug. Diss. Breslau.) Bot. Gent.Bd. csxni.nbsp;p. 123.

Krystofovic, A. (10) Jurassic Plants from Ussuriland. Mem. Ccm. Geol. [X.S.], Livr. 56.

Kubart, B. (08) Pflanzenversteinerungen enthaltende Knollen aus dem Ostrau-Karwiner Kohlenbecken. Sitzber. K. Akad. Wiss. Wien,nbsp;Bd. cxvii. Abt. i. p. 573.

- nbsp;nbsp;nbsp;(11) Corda�s sphaerosiderite aus dem Steinkohlenbeoken Radnitz-

Bfaz in B�hmen nebst Bemerkungen iiber Ckorionopteris gleicheni-oides Corda. Ibid. Bd. cxx. Abt. i. p. 1035.

Kurtz, F. (03) Remarks upon Mr E. A. Arber�s communication on the Clarke collection of Fossil Plants from New South Wales. Quart.nbsp;Journ. Geol. Soc. vol. lix. p. 25.

Kutorga, S. (42) Beitrag zur Palaeontologie Russlands. Verhand. Buss.-Kais. Mineral. Ges. St Petersburg.

Land, W. J. G. (04) Spermatogenesis and Oogenesis in FpAedro in/Mrca. Bot. Gaz. vol. xxxvm. p. 1.

Lang, W. H. (97) Studies in the Development and Morphology of Cycadean sporangia. I. The microsporangia of Stangeria para-doxa. Ann. Bot. vol. xi. p. 421.

p. 281.

Lawson, A. A. (04) The gametophytes, fertilization, and embryo of Gryptomeria japonica. Ann. Bot. vol. xviii. p. 417.

Lebour, G. A. (77) Illustrations of Fossil Plants; being an autotype reproduction of selected drawings. London.

Lesquereux, L. (74) Contributions to the Fossil Flora of the Western Territories. Pt. I. The Cretaceous Flora. Bep. U. S. Geol. Surv.nbsp;Territ. vol. vi.

Lesquereux, L. (78) On the Cordaites and their related generic divisions, in the Carboniferous formation of the United States. Proc. Amer.nbsp;Phil. Soc. (Philadelphia), vol. xvii. p. 31-5.

Lignier, 0. nbsp;nbsp;nbsp;(92) La nervation taeniopt�rid�e de folioles de Cycas et Ie

des Gymnospermes et celle des Angiospermes ? Bull. soc. Linn. Normand. [6], vol. vii. p. 55.

- nbsp;nbsp;nbsp;(04)nbsp;nbsp;nbsp;nbsp;Notes compl�mentaires sur la structure du Bennettites

- nbsp;nbsp;nbsp;(11)nbsp;nbsp;nbsp;nbsp;Le Bennettites Morierei (S. and M.) Lign. se reproduisait

probablement par parth�nog�n�se. Bull. soc. bot. France [4], tome xi. p. 224.

radicelles de Cordaitales. Assoc. Fran^. Avanc. Sci. XL. (Dijon), p. 509. [jSee also Lignier (06).]

- nbsp;nbsp;nbsp;(12)nbsp;nbsp;nbsp;nbsp;Stomates des �cailles inters�minales chez le Bennettites

.--(13) nbsp;nbsp;nbsp;V�g�taux fossiles de Normandie. VII. Contributions a la

Lignier, 0. nbsp;nbsp;nbsp;{J3^) Differenciation des tissus dans le Bourgeon v�g�tatif

Lignier, 0. et A. Tison. (11) Les Gn�tales sont des Angiospermes ap�tales. Gompt. Rend. Jan. 23.

Lima, W. de. (88) Flora Fossil de Portugal. Monographia do Genero Dicranophyllum. Comm, dos Trab. geol. Portugal.

Lingelsheim, A. (08) �ber die Braunkohlenholzer von Saarau. Jahres. Ber. Schles. Oes. Vaterland. CuUur. Bd. lxxxv.

Lomax, J. (02) On some iesttmesmi�lation to Lyginodendronoldhamium, Ann. Bot. vol. xvi. p. 601.

Lotsy, J. (99) Contributions to the life-history of the genus Gnetum. Ann. Jard. Bot. Buit. [2], vol. i. p. 46.

Ludwig, R. (61) Fossile Pflanzen aus der altesten Abtheilung der Rheinisch-Wetterauer Tertiar-Formation. Palaeontograph. Bd. vm.nbsp;p. 39.

Umgegend von Dillenburg, Bidenhopf und Friedberg und aus den SaaKeldischen. Ibid. Bd. xvir. p. 105.

Mackie, S. J. (62) The Dragon Tree of the Kentish Rag. Geologist, vol. V. p. 401.

McLean, R. C. (12) Two Fossil prothalli from the Lower Coal Measures. New Pliyt. vol. xi. p. 305.

McNah, W. R. (70) On the structure of a Lignite from the Old Red Sandstone. Trans. Bot. Soc. Edinburgh, vol. x. p. 312.

Mahlert, A. (85) Beitrage zur Kenntniss der Anatomie der Laubblatter der Coniferen mit besonderer Beriohsichtigung des Spalt�ffnungs-Apparates. Bot. Cent. Bd. xxiv. p. 54.

Malloizel, G. and R. Zeiller. (N.D.) Bibliographie et tables icono-graphiques (0. Heer). Stockholm.

Mansell-Pleydell, J. C. (85) Notes on a cone from the Inferior Oolite Beds of Sherborne. Proc. Dorset Nat. Hist. Antiq. Field Club,nbsp;vol. V. p. 141.

of the south-east of England. Quart. Journ. Oeol. Soc. vol. li. p. 51. Marion, A. F. (84) Sur les caraot�res d�une Conif�re tertiaire, voisine desnbsp;Dammar�es {Doliostrobus Sternbergii). Compt. Rend. vol. xoiv. p. 821.nbsp;Marsh, A. S. (14) Notes on the Anatomy of iSiawjrena parorfom. Ne^vnbsp;Phyt. vol. xiii. p. 18.

Marty, P. (08) Sur la Flore fossile de Lugarde (Cantal). Gompit. Rend. vol. CXLVII. p. 395.

Massalongo, A. (59) Studii sulla Flora Fossile egeologiastratigrafioa del Senigalliese. Verona.

Massalongo, A. and G. Scarabelli. (58) Studii sulla Flora Fossile e geologia stratigrafioa del Senigalliese. Verona.

Masters, M. T. (91) Review of some points in the comparative morphology, anatomy, and life-history of the Coniferae. Journ. Linn. Soc. vol. xxvm. p. 236.

Matte, H. (04) Recherches sur I�appareil lib�ro-ligneux des Cycad�es. Caen.

Mercklin, C. E. von. (56) PalaeodendrologiconRossicum. St Petersburg. Miller, H. (58) The Cruise of the Betsey. Edinburgh.

Mirande, M. (05) Recherches sur Ie d�veloppement et l�anatomie de Cassythac�es. Ann. Sci. nat. [ix], vol. I. p. 181.

Miyake, K. (06) �ber die Spermatozoiden von Gycas revoluta. Ber. Deutsch. bot. Oes. Bd. xxiv. p. 78.

Mogan, L. (03) Untersuchungen �ber eine fossile Konifere. Sitzber. K. Akad. Wiss. Wien, Bd. oxii. Abt. i. p. 829.

MoM, H. von. (62) Einige anatomische und physiologisohe Bemerkungen �ber das Holz der Baumwurzeln. Bot. Zeit. p. 225.

M�ller, H. (03) Bidrag till Bornholms Fossila Flora (Rhat ooh Lias). Gymnospermer. K. Svensk. Vetenskapsakad. Hand. Bd. xxxvi.nbsp;Xo. vi.

M�ller, H. J. and T. G. Halle. (13) The Fossil Flora of the Coal-bearing deposits of south-eastern Scania. Arkiv Bot. (Stockholm), Bd. xiil.nbsp;Xo. 7.

Mori�re, J. (69) Note sur deux v�g�taux trouv�s dans le d�partement du Calvados. M�m. soc. Linn. Normand. vol. xv.

Morris, J. (40) Memoir to illustrate a Geological map of Cutch (Grant, C. W.). Trans. Oeol. Soc. [2], vol. v. pt. ii. p. 289.

Murchison, Sir R. and R. Harkness. (64) On the Permian rocks of the Xorth-west of England, and their extension into Scotland. Quart.nbsp;Journ. Geol. Soc. vol. xx. p. 144.

Murchison, Sir R. and C. Lyell. (29) On the Tertiary Freshwater formations of Aix, in Provence, including the coal-field of Fuveau.nbsp;Description of plants by J. Bindley. Edinburgh New Phil. Journ.nbsp;p. 287.

Nakamura, Y. (83) Ueber den anatomisohen Bau des Holzes der wichtigsten Japanischen Coniferen. Unters. aus dem Forstbot.nbsp;Instit. zu M�nchen, m. Berlin.

Nathorst, A. G. (75) Om en Cycad�kotte fran denratiska formationens lager vid Tinkarp i Sk^ne. Ofver. K. Vetenskapsakad. Fork. No. 10.

- nbsp;nbsp;nbsp;(89) Sur la pr�sence du Genre D�ciyoxaTO�es, Old. dans les Couches

Nathorst, A. G. (99) The hTorwegian North Polar Expedition 1893-9�. Scientific Results, edit. P. Nansen. III. Fossil Plants from Franznbsp;Josef Land. London and Christiania.

Blatter von Dictyozamites Johnstrupi Nath. K. Svensk. Veten-skapsakad. Hand. Bd. xlii. No. 6.

teile. Palaobot. Mitt. 4-6. K. Svensk. Velenskapsakad. Hand. Bd. xLin. No. 6, p. 3.

siohtung Schwedisehen Arten. K. Svensk. Vetenskapsakad. Hand. Bd. XLIII. No. 12.

Nath, nebst einigen Bemerkungen fiber Podozamites. K. Svensk. Vetenskapsakad. Hand. Bd. XLVi. No. 8.

- (12) Die Mikrosporophylle von Wielandiella. Arkiv Hot. Stockholm, Bd. XII. No. 6, p. 1.

used? A question of nomenclature. Geol. Foren. Stockholm F�rh. Bd. XXXV. H. vi. p. 361.

schichten des Jura und der Kreide Spitzbergens. Geol. Foren. Stockholm Fork. Bd. xxxv. H. iv. p. 273.

Nathorst, A. G. (14) Zur Fossilen Flora der Polarlander. Teil I. Lief. iv. Nachtrage zur Palaozoischen Fl�ra Spitzbergens. Stockholm.

Norwegian Arct. Exped. in the �Fram � 1898-02, No. 35. Kristiania. Negri, G. (14) Sopra alouni legni fossili del Gebel Tripohtano. Boll. Soc.nbsp;geol. Ital. vol. xxxni. p. 321.

Nestler, A. (95) Ein Beitrag zur Anatomie der Cycadeenfiedern. Prings-Jieim'a Jahrh. Bd. xxvn. p. 341.

Newberry, J. S. (54) New Fossil Plants from Ohio. Annals of Science; including the Transactions of the American Association for the Advancement of Science, vol. I. Cleveland, 1863-54, p. 116.

Newton, E. T. and J. J. H. Teall. (97) Notes on a collection of rocks and fossils from Franz Josef Land, made by the Jaokson-Harmsworthnbsp;Exped. during 1894-96. Quart. Journ: Oeol. Soc. vol. lui. p. 477.

- nbsp;nbsp;nbsp;(98) Additional notes on rocks and fossils from Franz Josef Jjand,

Noack, F. (87) Der Einfluss des Klinias auf die Cuticularisation und Verholzung der Nadeln einiger Coniferen. Pringsheim�s Jahrh.nbsp;Bd. xvm. p. 519.

Noelle, W. (10) Studi�n zur vergloiehenden Anatomie und Morphologie der Koniferen Wurzeln mit R�oksicht auf die Systematik. Bot.nbsp;Zeit. p. 169.

Nor�n, C. 0, nbsp;nbsp;nbsp;(08) Zur Kenntniss der Entwickelung von Saxigothaea

Oliver, F. W. (02) On some points of apparent resemblance in certain Fossil and Recent Gymnosperm seeds. New Phyt. vol. l. p. 145.

�- (06) nbsp;nbsp;nbsp;The seed, a chapter in Evolution. Pep. 16th Meeting Brit.

Oliver, F. W. and E. J. Salisbury. (11) On the structure and affinities of the Conostoma group of Palaeozoic seeds. Ann. Bot. vol. xxv.

Oliver, F. W. and D. H. Scott. (03) On Lagenostoma Lomaxi, the seed of Lyginodendron. Proc. R. Soc. vol. lxxi. p. 477.

with a statement of the evidence upon which it is referred to Lygino-dendron. Phil. Trans. B. Soc. vol. cxcvii. p. 193.

Osborne, T. G. B. (09) The lateral roots of Amyelon radicans Will, and their Mycorhiza. Ann. Bot. vol. xxin. p. 603.

Palibin, J. (01) Quelques donn�es relatives aux d�bris v�g�taux oontenus dans les sables blancs et les gr�s quartzeux de la Russie m�ridionale.nbsp;Bull. Com. Oeol. tome xx. p. 447.

Pax, F. (07) Beitrage zur fossilen Flora der Karpathen. EnglePs Bot. Jahrb. Bd. xxxviii. p. 272.

Pearson, H. H. W. (06) Notes on South African Cycads. I. Trans. 8. Afr. Phil. Soc. vol. xvi. p. 341.

Reproduction in the genus Gnetum. Trans. Linn. Soc. vol. viii. pt. viii. p. 311.

with certain exotic species from Japan and Australia. Amer. Nat. vol. XXXVIII. pp. 243, 523.

of Alberta. The Ottawa Naturalist, vol. xxii. No. iv. p. 82. Phillips, J. (71) The Geology of Oxford and the valley of the Thames.nbsp;Oxford.

Pilger, R. (03) Taxaceae. Das Pflanzenreich (A. Engler), Heft xviii (iv. 5). Leipzig.

Platen, P. (08) Untersuchungen fossiler H�lzer aus dem Westen der Vereinigten Staaten von Nordamerika. Leipzig.

Pomel, A. (49) Mat�riaux pour servir a la flore fossile des terrains jurassiques de la France. Amt. Ber. Versam. Oes. deutsch. Natur-forsch. und Arzte. Aachen.

Potoni�, H. (88) Die fossile Pflanzen-Gattung Tylodendron. Jahrb. K. Preuss. geol. Landes, p. 311.

Die Reisen des Bergassessors Dr Dantz in Deutsch-�stafrika in den Jahren 1898-00. Mitt, aus den deutschen Schutzgebieten,nbsp;Bd, XV. Heft iv. p. 227.

Potoni�, H. and C. Bernard. (04) Flore Devonienne de l��tage H. de Barrande. Leipzig.

Prankerd, T. L. (12) On the structure of the Palaeozoic seed Lagenostoma ovoides Will. Journ. Linn. Soc. vol. XL. p. 461.

Prestwich, J. (54) On the structure of the strata between the London Clay and the Chalk in the London and Hampshire Tertiary systems.nbsp;Quart. Journ. Geol. Soc. vol. x. p. 75.

Raciborski, M. (91) Flora Retycka P�lnoonego Stoku G�r Swiqtokf-zyskich. Rozprawy Wydzialu Akad. Umiej. Krakowie, tom. xxni. - (92) Przyczynek do Flory Retyckiej Polski. Ibid. tom. xxil.

Radais, M. (94) L�anatomie compar�e du fruit des Conif�res. Ann. 8ci. nat. [7], vol. xix. p. 165.

Ratte, F. (87) Note on two new fossil plants from the Wianametta shales. Proc. Linn. Soc. N.S.W. [2], vol. l. p. 1078.

Rattray, G. (13) Notes on the pollination of some South African Cycads. Trans. B. Soc. 8. Africa, vol. iii. p. 259.

Reid, Clement and Eleanor M. Reid. (08) On the Preglacial Flora of Britain. Journ. Linn. Soc. vol. xxxvm. p. 206.

Renault, B. (79) Sur un nouveau groupe de tiges fossiles silicifi�es de l��poque houilli�re. Compt. Rend, tome lxxxviii. p. 35.

- (80) nbsp;nbsp;nbsp;Sur unenbsp;nbsp;nbsp;nbsp;nouvelle esp�cenbsp;nbsp;nbsp;nbsp;denbsp;nbsp;nbsp;nbsp;Poroxylon.nbsp;nbsp;nbsp;nbsp;Compt.nbsp;nbsp;nbsp;nbsp;Rend.

- (96) nbsp;nbsp;nbsp;Note surnbsp;nbsp;nbsp;nbsp;Ie genre M�tacordaite.nbsp;nbsp;nbsp;nbsp;Soc.nbsp;nbsp;nbsp;nbsp;dihist.nbsp;nbsp;nbsp;nbsp;nat.nbsp;nbsp;nbsp;nbsp;d'Aulun.

Renault, B. and R. Zeiller. (85) Sur un nouveau type de Cordait�e. Compt. Rend. vol. C. p. 867.

Rendie, A. B. (96) Gymnospermae. The Plants of Milanji, Nyasa-Land, collected by Mr A. Whyte. Trans. Linn. Soc. vol. iv. [2], p. 60.

Renier, A. (07) Trois esp�ces nouvelles, Sphenopteris Dumonti, 8. Corneti et Dicranophyllum Richiri du Houiller sans Houille de Baudour,nbsp;Hainaut. Ann. soc. geol. Belg. vol. xxxiv. M�m. p. 181.

Renner, 0. (04) Uber zwitter Bliithen bei Juniperus communis. Flora, Bd. xcii. p. 92.

Richards, J. T. (84) On Scottish Fossil Cycadaoeous leaves contained in the Hugh Miller Collection. Proc. R. Phys. Soc. Edinburgh.nbsp;Rivi�re, H. C. C. La. (16) Sur I�anatomie et l'�paississement des tigesnbsp;du Onetum moluccense Karst. Ann. Card. Bot. Buitenzorg, vol. xxx.nbsp;[2], p. 23.

Robertson, Agnes. (02) Notes on the anatomy of Macrozamia heteromera Moore. Proc. Camb. Phil. Soc. vol. xii. pt. i. p. 1.

Rosen, F. (11) Die biologische SteUung der abessinischen Baumlobelie {Lobelia Ehynchopetalum). Beit. Biol. Pflanzen. Cohn und Rosen,nbsp;Bd. X. Heft ii. p. 265.

Rothert, W. (99) Ueber parenchymatisohe Tracheiden und Harzgange im Mark von Cephalotaxus Arten. Ber. deutsch. hot. Ges. Bd. xvii.nbsp;p. 275.

Rushton, W. (15) Structure of the wood of Himalayan Junipers. Journ. Linn. Soc. vol. xi.ni. p. 1.

- nbsp;nbsp;nbsp;(16) The Development of � Sanio�s Bars � in PiwMS/�ops. Ann,

Russow, E. (72) VergleiohendeUntersuchungender I.eitbiindel-Krypto-gftmen. Mem, VAcad. Imp. Sci. St P�tershourg [vii], tome xix. p. 1.

Salisbury, E. J. (13) Methods of Palaeobotanical reconstruction. Ann. Bot, vol. XXVII. p. 273.

Sandberger, F. von. (64) Die Flora der oberen Steinkohlenformation im badischer Schwarzwalde. Verh. Nat. Ver. Carlsruhe, vol. i. p. 30.

Sanio, K. (74) Anatomie der Gemeinen Kiefer {Pinus silvestris L.). Pringsheim�s Jahrh. Bd. ix. p. 78.

Saporta, G. de. (62) �tudes sur la v�g�tation du sud-est de la France a l��poque Tertiaire. Ann. Sci. Nat. tome xvi. [4], p. 309.

groupe des Noeggerathia; g�n�ralit�s et type du Noeggerathia foliosa Sternb. Ibid. vol. lxxxve. p. 746.

Saporta, G. de. (78^) Sur le nouveau groups Pal�ozoique des Dolero-phyll�es. Ibid. vol. txxxvii. p. 393.

Saporta, G. de and A. F. Marion. (76) Recherches sur les v�g�taux fossiles de Meximieux. Arch. Mus. d�hist. nat. de Lyon. Lyon.

Saxton, W. T. (10) Contributions to the life-history of Widdringtonia cupressoides. Bot'. Oaz. vol. n. p. 31.

Masters, with some notes on the Phylogeny of the Cupressoideae and the Callitroideae. Ann. Bot. vol. xxvii. p. 577.

Schauroth, C. von. (52) Herr von Sohauroth an Herrn Beyrioh. Zeit. Deutsch. geol. Oes. Bd. iv. p. 538.

Schenk, A. (67) Ueber die Flora der schwarzen Schiefer von Raibl. Wiirzburg. Naturwiss. Zeitsch. vol. VT. p. 10.

Schenk, A. (89) Ueber Medullosa Cotta. Abh. K. Sachs. Ges. Wiss. Bd. XV. p. 523.

Schmalhausen, J. (83) Die Pflanzenreste der Steinkohlenformation am �stlichen Abhange des Ural Gebirges. M�m. Acad. S. Petersbourg,nbsp;vol. XXXI. No. 13.

�� (87) Die Pflanzenreste der Artinskischen und Permischen Ablagerungen im Osten des Enropaisohen Russlands. Mem.nbsp;Com. g�ol. St Petersbourg, vol. ii. No. iv. p. 1.

zur Erforschung des Janalandes und der Neusibirischen Inseln in den Jahren 1885, 1886 ausgesandten Expedition. Abt. il. Tertiarenbsp;Pflanzen der Insel Neusibirien. M�m. I�acad. Imp. Sci. St P�ters-bourg [vii], vol. xxxvn. No. 5.

Schneider, W. (13) Vergleichend-morphologische Untersuchungen �ber die Knrztriebe einiger Arten von Pinus. Flora [N.F.], Bd. v. p. 385.nbsp;Schroeter, C. (80) Untersuohung fiber fossile H�lzer aus der arotischennbsp;Zone. (Heer�s Flor. Foss. Arct. vol. vr. 1882.)

Vierteljahr. Naturforsch. Oes. Zurich. Jahrg. xlit. p. 125. Schuster, J. (11) Ueber Goeppert�s Raufneria im Zwinger zu Dresden.nbsp;Sitzber. K. Bayer. Akad. Wiss. p. 489.

Schiitze, E. (01) Beitrage zur Kenntniss der Triassisehen Koniferen-Gattungen Pagiophyllum, Voltzia, Widdringtonites. Jahresheft Ver. Vat. Naturkunde, Stuttgart, p. 256.

Scott, D. H. (97) The anatomical characters presented by the peduncle of Cycadaceae. Ann. Bot. vol. xi. p. 399.

Palaeozoic rocks. III. On Medullosa anglica, a new representative of the Cycadofilices. Phil. Trans. R. Soc. vol. oxoi. p. 81.

Scott, D. H. (09) The Palaeontological Record. II. Plants. Darwin and Modem Science, Art. xii. Cambridge.

Scott, D. H. and E. C. Jeffrey. (14) On Fossil Plants showing structure, from the base of the Waverley shale of Kentucky. Phil. Trans.nbsp;B. Soc. vol. 205, p. 315.

Scott, D. H. and A. J. Maslen. (07) The structure of the Palaeozoic seeds Trigonocarpus Parkinsoni Brongn. and T. Oliveri sp. nov.nbsp;Ann. Bot. vol. xxi. p. 89.

Sellards, E. H. (03) OorfoMot^eca, a new type of spore-bearing organ from the Coal Measures. Amer. Journ. Soi. vol. xvi. p. 87.

- (07) Kotes on the spore-bearing organ 6'orfoMotAec� and its relationship with the Cycadofilices. New Phyt. vol. vi. p. 175.

Sernander, R. (03) Die Einwanderung der Fichte in Skandinavien. EngleAs Bot. Jahrh. Bd. xv. p. 1.

Se�ward, A. C. (90) Tylodendron Weiss and Voltzia heterophyUa Brong. Geol. Mag. vol. vii. [3]*, p. 218.

on European and Eastern Mesozoic Floras. Quart. Journ. Oeol. Soc. vol. Lix. p. 217.

Seward, A. 0. (07�) Jurassic plants from Caucasia and Turkestan. M�m. com. geol. St Petersbmirg [N.S.], Livr. 38.

special reference to a collection of plants from Sussex. Quart. Journ. Geol. Soc. vol. lxix. p. 85.

- (14) nbsp;nbsp;nbsp;Antarcticnbsp;nbsp;nbsp;nbsp;Fossil plants. Britishnbsp;nbsp;nbsp;nbsp;Antarctic {Terra Nova)

- (11�) nbsp;nbsp;nbsp;Wealdennbsp;nbsp;nbsp;nbsp;Floras. Hastings andnbsp;nbsp;nbsp;nbsp;East Sussexnbsp;nbsp;nbsp;nbsp;Naturalist,

- (17) nbsp;nbsp;nbsp;H. H. W.nbsp;nbsp;nbsp;nbsp;Pearson, P.R.S., Sc.D.nbsp;nbsp;nbsp;nbsp;(Cambridge).nbsp;nbsp;nbsp;nbsp;Ann. Bot.

Seward, A. C. and N. Bancroft. (13) Jurassic Plants from Cromarty and Sutherland, Scotland. Trans. R. Soc. Edinb. vol. XLvm. pt. iv. p. 867.nbsp;Shaw, F. J. F. (08) A contribution to the anatomy of Ginkgo biloba.nbsp;New Phyt. vol. vii. p. 85.

- (09) nbsp;nbsp;nbsp;The seedling structure of Araucaria Bidwillii.nbsp;nbsp;nbsp;nbsp;Ann. Bot.

Shaw, W. E. (96) Contribution to the life-history of Sequoia sempervirens. Bot. Gaz. vol. XXI. p. 332.

Shirley, J. (98) Additions to the Fossil Flora of Queensland, Geol. Surv. Bull. No. 7. Brisbane.

Sifbon, H. B. (15) On the occurrence and significance of �bars� or �rims� of Sanio in the Cycads. Bot. Gaz. vol. lx. p. 400.nbsp;Sigrianski, A. (13) Quelques observations sur VEphedra helvetica Mey.

Univ. Gen�ve. Facult� des sciences. Prof. Chodat [8], Faso. x. Sinnott, E. W. (09) Paraoedroxylon, a new type of Arauoarian wood.nbsp;Rhodora, vol. II. No. 129, p. 165.

Smith, F. G. (07) Morphology of the trunk and development of the microsporangium of Cycads. Bot. Gaz. vol. XLiii. p. 187.

Smith, J. E. (1797) Characters of a new genus of plants named Sails-buria. Trans. Linn. Soc. voL iii. p. 330.

Solereder, H. (99) Systematische Ana,tomie der Dicotyledonen. Stuttgart. - (08) Ibid. Erganzungsband.

Solms-Laubach, Graf zu. (84) Die Coniferenformen des Deutschen Kupferschiefers und Zechsteins. Pal. A bhand. (Dames and Kayser)^nbsp;Bd. II. Heft ii. p. 81.

�� (91) On the Fructification of Bennettites Gibsonianus Carr. Ann. Bot. vol. V. p. 419.

Falkenberg in Sohlesien enthaltenen Structur bietenden Pflanzen-reste. Bot. Zeit. Jahrg. li. p. 197.

berg in Schlesien erhaltenen structurbietenden Pfianzenreste. IV. Volkelia refracta, Steloxylon lAidwigii. Zeitsch. Bot. Jahrg.nbsp;IT. Heft viii. p. 629.

South, F. W. and R. H. Compton. (08) Notes on the anatomy of Dioon edule Lind. New Phyt. vol. vii. p. 222.

Spiess, Karl von. (02) Ginkgo, Cephalotaxus, und die Taxaceen. Eine phylogenetische Studie. Osier. Bot. Zeitsch. Jahrg. lit. p. 432.nbsp;- (03) Ibid. Jahrg. Lin. p. 1.

Sprague, T. A. (16) Dionoophyllum. Kew Bulletin, No. 4, p. 89. Sprecher, A. (07) Le Ginkgo biloba L. Geneve.

Squinabol, S. (92) Contribuzioni alia flora fossile dei terreni terziarii della Liguria. Q�nes, 1889-92.

Stapf, 0. (89) Die Arten der Gattung Ephedra. Denksch. K. Akad. Wiss. Wien, Bd. lvi.

Staub, M. (96) Die Fossilen Ctenis Arten und Ctenis Jmngarica n. sp. Foldtani K�zl�ny, vol. xxvi.

B. Inst. Stud. sup. pratici e di perfezionamento in Firenze. Stenzel, G. (76) Beobachtungen an durchwachsenden Fichtenzapfen.

Sterzel, J. T. (83) Ueber Dicksonia Pluckeneti (Schloth.). Bot. Cent. Bd. xin. p. 282.

- nbsp;nbsp;nbsp;(12) Der �verkieselte Wald� im Garten desKonig Albert Museums

und des Orth-Denkmal in Chemnitz-Hilbersdorf. Bericht Natur-wiss. Ges. Chemnitz, Bd. xvill. p. 51.

Stiles, W. (08) The anatomy of Saxegothaea conspicua Lind. New Phyt. vol. VII, p. 209.

Stirling, J. (00) Notes on the Fossil Flora of South Gippsland Jurassic beds. Bep. on the Viet, coal-fields. No. 7. (Dpt. Mines, Victoria.)nbsp;Stokes and Webb. (24) Descriptions of some fossil vegetables of thenbsp;Tilgate Forest in Sussex. Trans. Geol, Soc. [2], vol. ii. p. 421.nbsp;Stopes, Marie C. (03) On the leaf of Cordate. New Phyt. vol. u.p. 91.

Slopes, Marie C. (14^) A new Araucarioxylon from New Zealand. Ann. Bot. vol. xxviii. p. 341.

�� (15) Catalogue of the Mesozoic Plants in the British Museum (Nat. Hist.). The Cretaceeus Flora. Pt. ii. Lower Greensandnbsp;(Aptian) plants of Britain. London.

Slopes, M. C. and K. Fujii. (10) Studies on the structure and affinities of Cretaceous Plants. Phil. Trans. R. Soc. vol. CCI. p. 1.

Slopes, M. C. andE. M. Kershaw. (10) The anatomy of Cretaceous Pine leaves. Ann. Bot. vol. xxiv. p. 395.

Slopes, M. C. and D. M. S. Walson. (08) The present distribution and origin of the calcareous concretions in coal-seams known asnbsp;�coal-balls.� Phil. Trans. R. Soc. vol. cc. p. 167.

Slrasburger, E. (66) Ein Beitrag zur Entwickelungsgeschichte der Spalt�finungen. Pringsheim�s Jahrh. Bd. v. p. 297.

Slriibing, 0. nbsp;nbsp;nbsp;(88) Die Vertheilung der Spalt�finungen bei den Coni

Slur, D. (68) Beitrage zur Kenntniss der geologischen Verhaltnisse der Umgegend von Raibl und Kaltwasser. Jahrb. K.h. geol. Reichs.nbsp;Wien, Bd. xviii. p. 71.

Suzuki, Y. (10) On the structure and affinities of two new Conifers and a new fungus from the Upper Cretaceous of Hokkaido (Yezo). Bot.nbsp;Mag. Tokyo, vol. xxiv. No. 284, p. 183.

Sykes, M. G. (10) � (See also Thoday, M. G.) The anatomy of Wel-witschia mirabilis Hook. f. in the seedling and adult stages. Trans. Linn. Soc. vol. vii. pt. ii. p. 327.

- nbsp;nbsp;nbsp;(10^) The anatomy and morphology of the leaves and inflorescences

Tate, E. (67). On the secondary fossils from S. Africa. Quart. Journ. Geol. Soc. vol. XXIII. p. 130.

Tenison Woods, J. E. (83) On a species of Brachyphyllum from Mesozoic coal beds, Ipswich, Queensland. Proc. Linn. Soc. N.S. W-vol. VII. p. 659.

Thiselton-Dyer, W. T. (72) On some Coniferous remains from the Lithographic stone of Solenhofen. Geol. Mag. vol. ix. p. 1.

Thiselton-Dyer, W. T. (01) The carpophyll of Encephalartos. Ann. Bot. vol. XV. p: 548.

Formosa, etc. Cycadaceae. Journ. Linn. Soc. vol. xxvr. p. 559. Thoday, Mary Gr. (Sykes, M. G-.). (11) The female inflorescence andnbsp;ovules of Gnetum africanum, with notes on O. secundum. Ann.nbsp;Bot. vol. XXV. p. 1101.

Thoday, M. G. and Emily M. Berridge. (12) The anatomy and morphology of the Inflorescences and flowers of Ephedra. Ann. Bot. vol. XXVI. p. 953.

Thomas, F. (66) Zur vergleichenden Anatomie der Coniferen-Laub-blatter. Pringsheimis Jahrh. Bd. iv. p. 23.

Thomas, H. Hamshaw. (11) The Jurassic Flora of Kamenka in the district of Isium. M�m. com. geol. St. Petersbourg [N.S.], Livr. 71.

Eretmophyllum, a new type of Ginkgoalean leaf. Proc. Gamb. Phil. Soc. vol. xvn. pt. iii. p. 256.

male flower of Williamsonia gigas. Proc. Gamb. Phil. Soc. vol. xvtii. pt. iii. p. 105.

Thomas, H. H. and Nellie Bancroft. (13) On the cuticles of some recent and fossil Cycadean fronds. Trans. Linn. Soc. vol. vm. pt. v.nbsp;p. 155.

Thompson, W. P. (10) The origin of ray-traoheids in the Coniferae. Bot. Gaz. vol. L. p. 101.

Thomson, R. B. (07) The Araucarineae, a Protosiphonogamic method of Fertilisation. Science [N.S.], vol. xxv. p. 272.

Thomson, R. B. and A. E. Allin. (12) Do the Abietineae extend to the Carboniferous? Bot. Gaz. vol. Lm. p. 339.

Thomson, Mary R. H. (16) A note on the wood of Gnetum Onemon. Ann. Bolus Herb. vol. ll. pt. ii. p. 81.

Tison, A. (09) Sur le Saxegothaea eonspicua Lind. M�m. soc. Linn. Normand. vol. xxin. p. 139.

' � (12) Sur la persistance de la nervation dichotomique chez les Conif�res. Bull. soc. Linn. Normandie [vi], vol. iv. p. 30.

Tapper, W. W. (11) Notes on Ginkgo biloha. Bot. Gaz. vol. lt. p. 3-74. Tuzson, J. (00) Monographic der foss�en Pflanzenreste der Balatonsee-gegend. Result, der Wiss. Erforsch. des Balatonsees, Bd. i. Teil i.nbsp;Budapest.

- nbsp;nbsp;nbsp;(52^) Iconographia plantarum fossilium. Abbildungenund Beschrei-

Velenovsk;^, J. (81) Die Flora aus den Ausgebrannten Tertiaren Letten von Vrsovic bei Laun. Abh. b�hm. Ges. [vi], Bd. ii.

Vernon, R. D. (12) On the geology and palaeontology of the Warwickshire coalfield. Quart. Journ. Oeol. Soc. vol. Lxvni. p. 587.

Vetters, K.L. (84) Die Blattsteile der Cycadeen. (Inaug. Diss.) Leipzig. Vierhapper, F. (10) Entwurf eines neuen Systems der Coniferen.nbsp;Jena.

Walkom, A. B. (17) Mesozoic Floras of Queensland. Queensland Oeol. Surv. Publications, No. 259.

trunks thus far determined from the Lower Cretaceous rim of the Black HiUs. Proc. U. S. Nat. Mus. vol. xxi. p. 21.

Cycadean trunks from the Jurassic of Wyoming. Proe. Washington Acad. Sci. vol. i. p. 253.

Warming, E. (77) Recherches et remarques sur les Cycad�es. Oversigt K. D. VidensJc. Selsk. Fork. (Copenhagen), p. 16.

Weiss, C. E. (72) Fossile Flora der j�ngsten Steinkohlen-formation und des Rothliegenden im Saar-Rhein-Gebiet. Bonn, 1869-72.

p. 1.

Weiss, F. E. (12) Report of the 80th meeting (Portsmouth) of the British Assoc, for the Advancement of Science, p. 550.

Weiss, F. E. (13) The root-apex and young root of Lyginodendron. Mem. Proc. Manchester Lit. Phil. Soc. vol. LVii. pt. iii.

Wernham, H. F. (11) Floral Evolution; with particular reference to the sympetalous Dicotyledons. New Phyt. vol. x. p. 73.nbsp;Wettstein, R. V. (90) Die Omorika-Fichte, Pieea omorica (Pane.).nbsp;Sitzher. K. Ahad. IFfss. Wien, Bd. xevin. Abt. i. p. 503.

Wherry, E. T. (12) Silicified wood from the Triassic of Pennsylvania. Proc. Acad. Sci. Philadelphia, vol. lxiv. pt. ii. p. 366.

White, D. (90) On Cretaceous Plants from Martha�s Vineyard. Amer. �Journ. Sci. vol. xxxix. p. 93.

White, D. and C. Schuchert. (98) Cretaceous series of the West coast of Greenland. Bidl. Geol. Soc. America, vol. ix. p. 343.

Wieland, G. R. (99) A study of some American Fossil Cyoads. Pt. ii. The Leaf structure of Oycadeoidea. Amer. Journ. Sci. vol. vii.nbsp;p. 305.

- (13) The Liassic Flora of the Mixteca Alta of Mexico, its composition, age, and source. Ibid. vol. xxxvi. p. 251.

Wiesner, J. (03) Die Rohstoffe des Pfianzenreiches. Bd. ii. Leipzig. Wild, G. (00) On new and interesting features in Trigonocarpus olivae-formis. Trans. Manchester Geol. Soc. vol. xxvi. pt. xv. p. 434.nbsp;Williamson, W. C. (40) On the Distribution of Fossil remains on thenbsp;Yorkshire coast from the Lower Lias to the Bath Oolite inclusive.nbsp;Trans. Geol. Soc. [2], vol. v. p. 223.

Williamson, W. C. (51) On the structure and affinities of the plants known as Sternbergiae. Manchester Lit. Phil, Soc. [2], vol. ix.nbsp;p. 340.

Lancashire. Rep. i�th meeting {Bristol) of the Brit. Assoc, p. 150.. - (86) On the morphology of Pinites ohlongus. Mem. Proc. Manchester Lit. Phil. Soc. vol. x. [3], p. 189.

Wiliamson, W. 0. and D. H. Scott. (94) The root of Lyginodendron oldhamium Will. Proc. R. Soc. vol. lvi. p. 128.

of the Coal Measures. III. Lyginodendron and Heterangium-. Phil. Trans. R. Soc. vol. clxxxvi. p. 703.

Wills, Lucy. (14) Plant cuticles from the Coal Measures. Oeol. Mag. [6], vol. I. p. 385.

Wills, L. T. (10) The fossiliferous Lower Keuper rocks of Worcestershire. Proc. Geol. Assoc, vol. xxi. p. 249.

Winkler, 0. nbsp;nbsp;nbsp;(72) Zur Anatomie von Araucaria brasiliensis. Bot. Zeit.

Witham, H. (30) On the vegetable fossils from Lennel Braes near Coldstream, upon the banks of the Tweed in Berwickshire. Phil.nbsp;Mag. vol. viii. p. 16.

tions of their internal structure as seen through the microscope. Edinb. Journ. Sci. vol. v. p. 183.

Worsdell, W. C. (96) The anatomy of the stem of Alacrozamia com. pared with that of other genera of Cycadeae. Ann. Bot. vol. x. p. 601

�� (98^) The comparative anatomy of certain genera of the Cycadaceae. Journ. Linn. Soc. vol. xxxiil. p. 437.

Worsdell, W. C. (00) The comparative anatomy of certain species of Encephalartos. Trans. Linn. 8oc. vol. v. pt. xiv. p. 445.

Yabe, H. (13) Mesozoische Pflanzen von Omoto. 8ci. Rep. of the Tokoku Imp. Univ. (ser. ii. GeoL), Bd. i. Heft iv. p. 57.

Yasui, K. (17) A Fossil Wood of 8equoia from the Tertiary of Japan. Ann. Bot. vol. xxxi. p. 101.

Yates, J. (55) Notice of Zamia gigas. Proc. Yorks. Phil. Soc. vol. i. p. 37. Yokoyama, M. (94) Mesozoic plants from Kozuke, Kii, Awa, and Tosa.nbsp;�Journ. Coll. Sci. Imp. Univ. .Japan, vol. vii. pt. iii. p. 201.

Young, Mary S. (07) The male gametophyte of Dacrydium. Bot. Gaz. vol. XLiv. p. 189.

Zalessky, M. (05) �ber Fr�ohte aus den Unter Carbon-Ablagerungen des Mstabeckens in Nord Russland. Bull. acad. Imp. sci. St. Peters-bourg, tome xxii. No. iii. p. 113.

von Jantai in der s�dlichen Mandshurei. Verhand. K. Russ. Min. Oes. St. P�tershourg [2], Bd. xlii. p. 385.

provenant du D�vonien sup�rieur du bassin du Donetz. B^dl. acad. Imp. sci. 8t. P�tersbourg.

balls in one of the seams of the Carboniferous strata of the Donetz basin. Ibid. p. 477.

identit� avec la Noeggerathiopsis Hislopi Bunb. sp. de la Plore du Gondwana. M�m. com. g�ol. 8t. P�tersb. [N.S.], Livr. 86.

Zalessky, M. (12^) On the impressions of plants from the coal-bearing deposits of Sudzenka, Siberia. Bull. soc. Natural. Orel. pt. iv.

Zang, W. (04) Die Anatomie der Kiefernadel und ihre Verwendung znr systematischen Gliederung der Gattung Pinus. (Diss. Inaug.)nbsp;Oiessen.

Zeiller, E. (78) Sur une nouvelle esp�oe de Bicranophyllum. Bull. soc. g�ol. France [3], vol. vi. p. 611.

- (12) Sur quelques v�g�taux fossiles de la Grande Oolite de Marquise. Bull. soc. acad. Boulogne-sur-nter, tome ix. p. 5.

Zeiller, E. and P. Fliche. (04) D�oouverte de strobiles de Sequoia et de Pin dans le Portlandien des environs de Boulogne-sur-mer. Compt.nbsp;Rend, tome cxxxvn. p. 1020.

Zeiller, E. and L. M. Vidal. (02) Sobre algunas impresiones vegetales del Kimeridgense de Santa Maria de Meya. Mem. Real Acad.nbsp;Cienc. y Aries Barcelona, vol. iv. No. 26.

Zigno, A. de. (53) D�oouverte d�une flore Jurassique analogue a celle de Scarborough dans les couches oohtiques des Alpes V�nitiennes.nbsp;Bull. soc. g�ol. France [2], vol. x. p. 268.

Zopf, W. (92) Ueber einige niedere AlgenpUze (Phycomyceten) und erne neue Methode ihre Keime aus dem Wasser zu isoliren. Ahh. Natur-Jorsch. Ocs. Halle, Bd. xvn. p. 79.

h;S'-

**#

:amp; #

IIT'’! ' ■

CAMBRIDGE UNIVERSITY PRESS C. F. CLAY, Managesnbsp;aotl�on ; FETTER LANE, E.C. 4

FOSSIL PLANTS

- A TEXT-BOOK FOR STUDENTS

OF BOTANY AND GEOLOGY

A. C. SEWARD

GINKGOALES, CONIFERALES, GNETALES

CAMBRIDGE:

PREFACE

POSTSCEIPT

TABLE OE CONTENTS

CHAPTER XLI

CHAPTER XLII

CHAPTER XLIII

CHAPTER XLV

CONIFERALES (continued). Pp. 245�30].

CHAPTER XLVI

CONIFERALES (continued). Pp. 302�335.

CHAPTER XLVII

CONIFERALES (continued). Pp. 336�367.

CHAPTEE XLVIII

CONIFEEALES (continued). Pp. 368�4=04.

CHAPTEE XLIX

CONIFEEALES (continued). Pp. 405�423.

........................

CHAPTEE L

CONIFEEALES INCEETAE SEDIS. Pp. 424�446.

CHAPTEE LI

PODOZAMITES AND NAGEIOPSIS: GENEEA INCEETAE SEDIS., Pp. 447�458. .

CHAPTEE LII GNETALES. Pp. 459�472.

LIST OF ILLUSTRATIONS

CHAPTER XL,

lift

20

Other Localities.

'A f' I/' I

Jl

54

shot

�bghtlv

Secretory tracts occur between the veins of the smooth lamina

CHAPTER XLI.

.J

72

Africa, and South America and the p^ioewicops-is m

CHAPTER XLII.

- .also Arber (12) p. 401,

^ ergen, and North America; it is also a member of the ��nio-Carboniferous floras of South Africa and India.

P^yg�mophyllum WilUamsoni Natborst.

CHAPTER XLIII.

VIA

� ps

v,.

yl

116

119

Tf

,ooo

lt;9(7 �

fA'

f�-.

M' nbsp;nbsp;nbsp;O

Short summary of the characteristics of recent Conifers'^.

145

146

1) A

147

148

149

150

152

154

A nbsp;nbsp;nbsp;B

156

157

�\

158