HONOEAE� MEMBEE OP THE EOTAL ACADEMY OF GOTTINGEN, AND PEOFESSOE OP BOTANY IN THE OWENS COLLEGE AND THE VICTOEIA UNITEBSITY, MANCHESTEE.

LONDON:

1887.

INTRODUCTION.

A PALAEONTOLOGIST, located for more tliaii half a century near the centre of an area throughout which the Carboniferous rocks predominate, I have necessarilynbsp;been brought into constant contact with various forms of 8tigmaria ficoides. Notwithstanding these advantages I long participated in the ignorance which prevailednbsp;so widely respecting them; but the discovery, especially in the districts roundnbsp;Oldham and Halifax, of those remarkable coal-seams overlying the Millstone Glrit,nbsp;from which such rich harvests have been reaped during the last twenty years, hasnbsp;materially altered the position of students of Stigmaria. During that period, Inbsp;have collected every fragment of the plant calculated to throw light upon itsnbsp;structure and affinities, and I think I may express my belief that my cabinets nownbsp;contain the largest collection of such illustrative specimens in existence.

Tinder the impression that I now possess the materials for drawing up a history of Stigmaria ficoides which would contain but few blank places, the time seemsnbsp;to have arrived for placing such a history in the hands of geologists, though therenbsp;are, even yet, some features of the organism respecting which we require more knowledge. The publication of a fair report of what we do know will probably facilitatenbsp;the acquisition of what is yet wanting. I am fully satisfied that Stigmaria,nbsp;viewed as an organ,^ is a root; I am equally so that it is the root of variousnbsp;species of Sigillaria and Lepidodendra. Hence it is far from improbable thatnbsp;specific differences may one day be found amongst the objects which we now knownbsp;by the name of Stigmaria ficoides; but at present we have wholly failed to discover any such differences.

Though our knowledge of the structure of the aerial stems of Sigillarias and Lepidodendra is steadily increasing, we yet require more than is at present withinnbsp;our reach; but this is already sufficient to correlate, with a fair measure of probablenbsp;accuracy, the relationship existing between the tissues of these aerial stems andnbsp;those of their Stigmarian roots. We have abundant evidence respecting thenbsp;external features of that relationship, but we have yet to discover the actual junction of the vascular, exogenous or centrifugally developed, cylinder of the root with

^ I employ this term as it is used by Prof. De Bary, i.e. as the instrument of physiological work. See � Comparative Anatomy of the Phanerogams and Perns,� English Translation, p. 1.

INTRODUCTION.

the corresponding cylinder in the aerial stems. The centrifugal cylinder of the latter must obviously have been continuous with the similar one of the root, but thenbsp;centripetal zone enclosed within the centrifugal zone obviously terminated at thenbsp;base of the stem, like the medullary sheath of an ordinary exogen. But how it doesnbsp;so yet awaits demonstration. Connected with this another problem arises. Judgingnbsp;from the specimens in our possession I should conclude that the diameter of thenbsp;vascular cylinder of a Stigmaria was about one fourth that of its entire diameter,nbsp;including its surrounding bark. The largest vascular cylinder I have yet seen isnbsp;recorded in my text as being ^ of an inch in diameter, which would give to thenbsp;root of which it was the centre, a diameter of about 4f inches, or a circumferencenbsp;of rather more than 14 inches. But in the text I have described roots, the proximal ends of which are more than three feet in circumference. Now, such roots,nbsp;according to the scale of proportions adopted above, would require a vascularnbsp;cylinder fully three inches in diameter. We have not yet found any that approachnbsp;this magnitude. W'e have yet to discover whether or not any roots of this sizenbsp;exist, the internal structure of which has been preserved. Unfortunately littlenbsp;more than the outermost bark remains either in stems or roots of these dimensions.nbsp;Then we want specimens showing the structure of the part of the cortex internal tonbsp;the prosenchymatous layer, which latter forms so conspicuous a feature alike in thenbsp;Stigmarian roots and in their aerial stems. We are familiar with the parenchymatous zone that occupies this position in the stems, and we have indications thatnbsp;the same parenchyma was continued into the root; but we want clearer evidencenbsp;than we yet possess whether or not such was the case.

It must further be borne in mind that all the numerous Carboniferous plants, great and small, had rootlets of some kind, and that some of them bear a superficialnbsp;resemblance to those of Stigmaria. This is especially the case with the rootletsnbsp;of the plant which I named Amyelon, yet none of these can well be mistaken fornbsp;those of Stigmarise, though I doubt not that some have been so mistaken.

Some readers may consider that I have dwelt needlessly upon minute details of structure which, in their opinion, have neither interest nor importance. This isnbsp;possible ; yet I have done so with a definite hope as well as with an object. Thenbsp;hope is that sooner or later we shall know more than we yet do of the morphologicalnbsp;and physiological links connecting the primmval forms of vegetation with their livingnbsp;representatives. No real facts, however small, can be absolutely worthless to thenbsp;botanist who strives to work out this subject. Anyhow, the faithful record ofnbsp;them can do no harm, and they may be found useful at some future time.

Two peculiarities may be observed in the form of this Monograph. One is the elaboration of the Index to the Plates. In stndying the works of some of ournbsp;most distinguished Botanists, I have found it difficult to discover in what part of the-text the detailed descriptions of the Plates were to be found, and the too brief refer-^

INTRODUCTION.

ences in the Index itself threw but little light upon the subject. I have endeavoured as far as possible to make this task easy to such students as may consult what Inbsp;have written. A second feature is equally important. My entire Cabinet ofnbsp;sections of these Carboniferous plants, of which a very elaborate descriptive catalogue is already prepared, is destined to find its ultimate home in the Botanicalnbsp;Museum of the Owens College, where it will be accessible to any palaeontologistnbsp;who may desire to consult it. In the catalogue referred to, the description of eachnbsp;specimen embodies a statement as to what, in my opinion, that specimen teaches ornbsp;proves. The result will be that, whether those who may follow me in thesenbsp;researches agree with my views or feel compelled to reject them, they can themselves examine the specimens upon which those views were based. In order tonbsp;make such references easy, so far as the present work is concerned, the Cabinetnbsp;number of each specimen figured is attached to the notice of each figure in thenbsp;Index to the Plates. The cases where no such numbers are recorded belongnbsp;to specimens not in my cabinet. Such examples, however, are extremely few,nbsp;nearly all the specimens figured being in my own possession.

Stigmari^, which are mere casts or impressions, no portion of their internal organisation being preserved, are widely, and often abundantly, diffused through thenbsp;entire series of the Upper Carboniferous Rocks of Lancashire and Yorkshire downnbsp;to the Millstone Glrit. But the beds, from which specimens, not only of Stigmaria,nbsp;but of numerous other plants having their internal structures exquisitely preservednbsp;have been mainly derived, are the thin, lowermost coal-seams of the Ganister series.nbsp;These plants occur in nodules, of various diameters from a foot downwards, whichnbsp;are embedded in the substance of the coal. In some localities these nodulesnbsp;are so numerous as to make the working of the coal-seam unprofitable; a factnbsp;unfortunate for the palaeontologist, since such commercially unprofitable collieriesnbsp;are liable to be closed. So far as the Lancashire seams in which these plant-bearingnbsp;nodules occur are concerned, an excellent summary of them will be found in Mr.nbsp;Binney�s � Memoir on Calamites and Oalamodendron.�^ For the following detailednbsp;section of the corresponding series of deposits in the Halifax district I amnbsp;indebted to my friend J. W. Davis, Esq., F.G.S., of Chevinedge, Halifax.*^

^ It is fortunate for geological science that similar deposits have now been discovered at three separate localities on the Continent. One of these is at Pith Vollmond, in Westphalia, where my friend,nbsp;Count de Solms, tells me he finds in great abundance Lepidodendron selaginoides, Lyginodendronnbsp;Oldhamiwm, and Raohiopteris aspera, as well as examples of Amyelon radicans and some other well-known Yorkshire and Lancashire forms. One of the remaining two is in the Banat, in South Hungary,nbsp;and the other in Moravia. In all these places the deposits appear to be of the same age as those ofnbsp;Yorkshire and Lancashire, each being also overlain by a bed containing the marine Aviculopecten, asnbsp;in our Ganister series; it is from these shells, along with other mollusks, that the lime, which hasnbsp;played so important a patt in the preservation of our plants, has been derived.

It now only remains to record my acknowledgments of the assistance I have received from many friends in accumulating the materials from which this monograph has been drawn. All such names as Mr. G. Wild, of the Bardsley Colliery atnbsp;Ashton-under-Lyne, Mr. J. Spencer, and Mr. Binns, of Halifax, Mr. Isaac Earnshawnbsp;and Mr. Neild, of Oldham, Mr. J. Butterworth, of Shaw, and Mr. Ward, F.O.S.,nbsp;of Langton, in Staffordshire, are already well known to those who are familiarnbsp;with my memoirs contained in the �Philosophical Transactions.� The cabinet of mynbsp;late friend, Mr. John Aitken, now in my possession, has furnished me with somenbsp;interesting facts. All the persevering collectors thus catalogued have poured theirnbsp;treasures into my lap with a generosity that is alike significant of their desire tonbsp;facilitate my researches and to aid in the advancement of science. I am alsonbsp;indebted to Professor Miall and Professor Green, of Leeds, to Professor Lebour, ofnbsp;Newcastle-upon-Tyne, and to Mr. Geddes, of Blackburn, for similar assistance.nbsp;The Natural History Society of Newcastle-upon-Tyne also has kindly granted menbsp;the loan of several interesting specimens from the Hutton Collection, now in itsnbsp;possession. But I must especially mention Mr. W. Cash, F.G.S., of Halifax, whosenbsp;unwearied diligence in accumulating new material for investigation has rendered menbsp;the greatest service. If I have been in any degree successful in removing somenbsp;of the clouds which have hitherto obscured the history of Stigmaria ficoides, nonbsp;small measure of that success is due to the friends who have generously assistednbsp;me to accumulate the rich store of materials to which my attention has so longnbsp;been given.

A MONOGRAPH

STIGMARIA FICOIDES.

Few of tlie objects studied by Palgeontologists have occasioned more speculation than those fragments of Carboniferous vegetation long known by the names of Variolaria, Phytolithus, and Ficoidites, but now recognised by that of Stigmarianbsp;ficoides assigned to them by Adolphe Brongniartd The opiniona,formed respectingnbsp;their botanical nature and relationship have varied even more than their names. Fornbsp;more than half a century nothing could be learnt respecting them beyond what theirnbsp;external contours could teach; and, since many misunderstood causes combined tonbsp;modify those contours, most of the opinions formed respecting them had no value.nbsp;But much light has been thrown upon these fossils during the last half century.nbsp;Nevertheless, even at the present time some eminent Palaeontologists continue tonbsp;give circulation to views respecting them which are, in my opinion, altogether inconsistent with what, to British geologists, are well-known facts. Some of these viewsnbsp;are reproductions of what we might have hoped �were exploded errors. Others arenbsp;new, but apparently no truer than the older ones.

To catalogue the vague guesses promulgated respecting Stigmarise in earlier days would be wasted labour. But the case is altered when we find such distinguished leaders of the French school of Palmo-botanists, as M. Renault, M. Grand-Eury, the Marquis of Saporta, and M. Marion giving currency to what I believe tonbsp;be serious errors, alike of observation and of interpretation, relating to the Stig-maria ficoides of Brongniart. It happens that the Coal-Measures of Great Britain,nbsp;Canada, and the United States of America are rich beyond most other countries innbsp;the supply they afford of specimens of this plant. This is true not only of thenbsp;structureless examples known to our forefathers, but of others in which thenbsp;internal structure is preserved with exquisite beauty and completeness. Hencenbsp;the rich stores contained in our cabinets enable us to speak with a decision that

2 nbsp;nbsp;nbsp;STIGMARIA PICOIDES.

would otherwise appear presumptuously dogmatic. My reasons for rejecting the views of my French fellow-workers will be given in the following pages. But innbsp;order to avoid needless interruptions to the continuity of my descriptions of whatnbsp;appear to me to be facts, and of the conclusions which I think legitimately dedu-cible from those facts, all controversial treatment of the subject will be limited tonbsp;the footnotes.

I have not thought it necessary to re-figure the common aspects of Stigmaria ficoides with its attached rootlets. Every geologist is familiar with these forms ;nbsp;such figures have been well supplied by Martin,^ under the name of Phytolitliusnbsp;verrucosus ; by Artis^ as Ficoidites furcatus, verrucosus, and major ; by Bindley andnbsp;Hutton� and by Corda^ as Stigmaria ficoides.

The first acquisition of some really scientific notions respecting Stigmaria date from the publication of a memoir by Mr. (now Sir John) Hawkshaw, the distinguished engineer. When constructing the railway between Manchester and Bolton,nbsp;in 1837, under Sir John Hawkshaw�s direction, some excavators, cutting through thenbsp;Carboniferous strata at Dixon Fold,near the present Clifton Station, discovered ahugenbsp;fossil tree with large out-spreading roots, standing vertically upon a seam of coal,nbsp;and soon afterwards they exposed several others in similar positions. Excellentnbsp;figures and descriptions of five of these trees were published by Sir Johnnbsp;Hawkshaw in 1839.� The largest of them was eleven feet high, seven and a halfnbsp;feet in circumference round its top, and fifteen feet round its base. A second lessnbsp;lofty tree exhibited four large roots radiating from its base; each of these rootsnbsp;soon divided, producing eight secondary roots which extended six feet from theirnbsp;parent stem. All these fossils were coated externally with a layer of coal from anbsp;quarter to three quarters of an inch in thickness. Within this coal each stem andnbsp;root was merely a structureless mass of clay or shale. The outer surface of thenbsp;coal, as well as the corresponding one of the subjacent clay, exhibited irregularnbsp;longitudinal flutings, but these surfaces afforded no definite evidence respectingnbsp;the character of the trees. This discovery established several very important facts,nbsp;first, that some of the largest stems of trees found in the Coal-Measures werenbsp;furnished with gigantic roots, which branched dichotomously; and second, thatnbsp;these roots must have extended downwards through a bed of undisturbed coal tennbsp;inches thick, by which the roots were abruptly cut off. It became obvious that thenbsp;trees must have grown where the fossils stood, and that the materials converted intonbsp;the bed of coal must have accumulated above their wide-spreading roots whilst thosenbsp;trees were growing, and that subsequent changes obliterated parts of the roots.

The second significant discovery was recorded by the late Mr. Binney.^ Some trees, similar in most respects to those of Dixon Fold, were laid bare at St. Helens,nbsp;in Lancashire, from which Mr. Binney learnt two important truths; first, that somenbsp;of these trees were true Sigillarige; and second, that their roots were the long-discussed Stigmarise. A second discovery of a similar kind was made at Ducken-field, in Lancashire, in the colliery of Messrs. Swire, Lees, and Co., where a shortnbsp;stem was exposed, from one side of which a large root issued. This root, whichnbsp;was three feet and one inch in circumference, ran horizontally for about sixteennbsp;inches, when it divided into two branches, each of which again divided. Thenbsp;result of these bifurcations was the production of four smaller roots, none of whichnbsp;underwent further subdivision, though one of them was prolonged to a distance ofnbsp;fifteen feet from the parent stem; in their proximal proportions the surfaces ofnbsp;these roots exhibited no Stigmarian features, but such were abundantly displayednbsp;beyond their second ramifications. Mr. Binney was again the recorder of thisnbsp;most important discovery.*

Our next evidence showing that Stigmaria was a root of a large tree, and not an independent plant, came from across the Atlantic. Mr. Richard Brown reported^nbsp;the discovery, in the Coal-field of Sydney, Cape Breton Island, of trees similar tonbsp;those observed in England; and a little later he pointed out, in addition, that whilstnbsp;some of these trees were undoubtedly Sigillari�, amongst them was what he believednbsp;to be a Lepidodendron,'*' whose roots were equally Stigmarian, a fact which hasnbsp;been confirmed by various later observers. Mr. Brown estimated that the spreadingnbsp;roots of one of his trees must have covered two hundred square feet of ground.nbsp;That the roots of Lepidodendra were Stigmarian was observed by Geinitz on thenbsp;Continent, and by other observers in the Coal-fields of South Wales. A few yearsnbsp;ago a remarkable Carboniferous forest was laid bare close to Oldham, in Lancashire,nbsp;in which some of the trees were unmistakeably Lepidodendra with Stigmariannbsp;roots.�

Nothing whatever was known of the internal structure of Stigmaria until 1838, when Mr. (now Professor) Prestwich obtained a specimen from the Coal-field ofnbsp;Colebrook Dale, of which a transverse and a tangential section were figured in the

1 nbsp;nbsp;nbsp;� London, Edinburgh, and Dublin Philosophical Magazine,� series 3, vol. xxiv, p. 105, 1844.

2 nbsp;nbsp;nbsp;� Quarterly Journal of the Geological Society of London,� vol. ii, 1846.

4 Whether or not Mr. Brown was right in concluding that the plant in question was a Lepido-dendron is unimportant, since we now possess abundant evidence that the root of Lepidodendron as well as of Sigillaria was a Stigmaria Jicoides.

^ In some of his memoirs Mr. Binney contended that the Halonise were the roots of Lepidodendron. It is now clear that these were the fruit-bearing branches of Lepidodendron�not its roots. M.nbsp;Eenault and M. Grand-Eury have contended that the Stigmarise are not always roots but rhizomes,nbsp;which have leaves, and send up aerial stems from their peripheral extremities. The absolute absencenbsp;of all foundation for these opinions will be shown on a later page.

4 nbsp;nbsp;nbsp;STIGMARIA PICOIDES.

* Fossil Flora of Great Britain.�^ The former of these sections (loc. cit., fig. 1) is, as far as it goes, a fair representative of such sections of the plant; but the other,nbsp;fig. 2, is wholly unintelligible to me. A much more accurate use of the samenbsp;specimen was made by Professor Prestwich himself who figured and describednbsp;additional sections of it in his classical memoir on the � Geology of Colebrook Dale.�**nbsp;He was indebted to the late Robert Brown for the accurate suggestion that certainnbsp;vascular bundles, springing from the vascular axis figured, terminated at thenbsp;depressed external tubercles so characteristic of all specimens of Stigmaria. Thenbsp;existence of these bundles had escaped the notice of the authors of the � Fossilnbsp;Flora,� though they certainly ought to have represented them in their fig. 2, asnbsp;they were illustrated in corresponding sections in Professor Prestwich�s memoir.

A memoir � On Some Peculiarities in the Structure of Stigmaria,� by Sir Joseph Hooker, appeared in 1848.^ The author, in this memoir, recorded all that was thennbsp;known about Stigmaria. But at that time illustrative materials were few, andnbsp;too often inconclusive. Some equally imperfect specimens previously describednbsp;by Professor G�ppert'* misled Dr. Hooker as to the origin of the vascular bundlesnbsp;noticed by Robert Brown, as they afterwards misled Mr. Binney. In 1858 thenbsp;latter author figured and described the central portion of a Stigmarian rootlet,�nbsp;and in a second memoir � he republished the same figure, and along with it henbsp;represented the fragment, from the interior of which his section of the rootlet hadnbsp;been obtained. This fragment was misinterpreted by Binney exactly as a similarnbsp;specimen had previously been by G�ppert.

Parts II and XI of my memoirs on the � Organisation of the Fossil Plants of the Coal-Measures � ^ contain some hitherto undescribed features in the structure ofnbsp;Stigmaria, as well as diagrammatic restorations both of that structure and of thenbsp;organic relations of the root to its Lepidodendroid and Sigillarian stems; thosenbsp;restorations need little, if any, alteration to adapt them to the present state of ournbsp;knowledge, though during the subsequent years a large amount of informationnbsp;has been obtained respecting the details of the organisation of Stigmaria.

On examining the trees discovered at Dixon Fold, it soon became obvious that the shaly materials of which they were composed would give way, however carefully they might be protected from the weather. Hence an Italian artist named

^ � Transactions of the Gieological Society of London,� 2nd series, vol. v, PI. 38.

� � Memoirs of the Gieological Survey of Glreat Britain and of the Museum of Practical Geology in Loudon,� vol. ii, part 2, 1848.

� � Observations on the Structure of Fossil Plants found in the Carboniferous Strata,� part iv, � Sigillaria and Stigmaria,� Palseontographical Society�s vol. for 1875.

Bally, well skilled In tke art of making moulds, was employed to obtain exact casts of the two most important of the above trees. These casts, which are now preserved in the Geological Museum of Owens College, are represented in Rigs. 1 andnbsp;2. That represented by Fig. 1 is 2 feet 6 inches high. The circumference of itsnbsp;upper extremity is 7 feet 9 inches, and 8 feet 5 inches at two feet from the floor.nbsp;From * to * measures 10 feet. Though the full length of the spreading roots whichnbsp;originally existed is not exhibited in either of these casts, both well show some ofnbsp;their primary dichotomous ramifications. The specimen, Fig. 2, is 65 feet high ; itsnbsp;circumference at its upper extremity is 7 feet 10 inches, and 10 feet 6 inches at twonbsp;feet above the floor. The longitudinal ridges and fux�rows seen on these casts donbsp;not represent the vertical flutings of a Sigillaria, neither do we find on the rootsnbsp;any traces of the rootlet-scars so conspicuous in Figs. 5 and 6. We learn fromnbsp;the specimen. Fig. 2, that in all probability no portion of the true aerial stem isnbsp;preserved in Fig. 1. It is a mere mass of coalesced roots. The upper part ofnbsp;Fig. 2 is doubtless cauline. The disappearance of all traces of leaf-scars from thenbsp;latter, and of rootlet-scars from both, is due to the same cause, viz. the internalnbsp;exogenous growth, which increased the diameter alike of the vascular and corticalnbsp;zones of both stem and root. This expansion necessarily caused the outer barknbsp;to become fissured and its superficial portions to be thrown ofiquot;; and, since thenbsp;cicatrices left by both leaves and rootlets were confined to the superficial cortex,nbsp;when this was cast oflf they inevitably disappeared along with it. This gradualnbsp;disappearance of both can readily be traced in numerous specimens.

As is now well known, the primary roots given off by all these trees are four in number. Fig. 1 only exhibits two ((X, a) of them. Three such (a, a, a) are seennbsp;in Fig. 2. In Fig. 1 both the primary roots, a, a, dichotomise. In Fig. 2 only onenbsp;such dichotomy is seen at A, but as the tree originally stood at Dixon Fold morenbsp;such were visible.

The thickness of each primary root (a) is often enormous in the older trees. The tetramorphic arrangement is best seen when the specimens are inverted so asnbsp;to show their inferior surfaces. Plate II, fig. 3, represents one such base now innbsp;the Museum of the Leeds Literary and Philosophical Society.^ Such specimensnbsp;are frequently marked by the defined crucial suture seen in Figs. 4 and 6, a factnbsp;which Was first pointed out by Mr. Binney in 1854.� These sutures indicate thenbsp;surfaces of contact of the four several roots. In Fig. 3 they are very feebly preserved. The circumference of the root a at e-h is 3 feet; of 5 at e-g is 3 feet 2nbsp;inches ; the breadth from y to /x is 21 inches, and from e to � 18 inches, whilst fromnbsp;a to d is 37| inches.

1 nbsp;nbsp;nbsp;T.his figure is lithographed from a photograph, for which I am indebted to Professor Miall, thenbsp;accomplished Curator of the Museum. Figs. 1 and 2 are also lithographed from photographs taken fornbsp;me by A. Brothers, Esq., F.R.A.8., of Manchester.

2 nbsp;nbsp;nbsp;� Quarterly Journal of the Gleological Society,� vi, p. 21, figs. 5 and 7.

STIGMARIA PICOIDES.

PI. II, fig. 4, represents a fine specimen in Professor Green�s collection at tlie Leeds College of Science.^ This specimen measures 50 inches from a to h and 36nbsp;from c to d. Though marked as usual by longitudinal cracks and corrugations, itsnbsp;surface is smooth contrasted with that of Fig. 3. The lines of demarcation betweennbsp;the four roots are not exactly crucial, as is also the case in Plate II, fig. 6. Thenbsp;growth of two, a and h, has probably been a little in advance of that of thenbsp;other two, which seem to have been pushed asunder by the former. Anothernbsp;striking feature in this specimen is the length attained by each of the four roots,nbsp;especially by a, without dichotomising. This variability in the conditions of thesenbsp;roots in different individuals was probably dependent on local states of the soilnbsp;into which the roots had to penetrate. An instructive illustration of this probability is afforded by Mr. Binney in his description of the Duckinfield specimen,nbsp;now in the Owens College Museum ; one of the primary roots, 3 feet and 1 inchnbsp;in circumference, extended horizontally for sixteen inches, when it divided into twonbsp;secondary branches ; each of these again divided into two others, one of which rannbsp;horizontally to a distance of fifteen feet from the central stem, without any furthernbsp;subdivision. These four ultimate branches alone exhibited the rootlet-scarsnbsp;characteristic of Stigmaria. On the other hand, the St. Helen�s tree presented anbsp;very different appearance. Its primary roots penetrated the fireclay at angles,nbsp;varying from 50� to 60�, before assuming the horizontal direction. In the formernbsp;case the roots encountered a stiff clay, which they did not penetrate freely; thosenbsp;of the latter example had eight feet of silty clay beneath them, into which theynbsp;plunged more easily.

Plate III, figs. 5 and 6, represent an instructive specimen sent to me by Mr. Geddes, the Librarian of the Blackburn Free Library. Being the base of a muchnbsp;younger tree than any hitherto described, it illustrates one or two points ofnbsp;interest not seen in any other example that I am acquainted with. The entire surface of the specimen being covered with the characteristic rootlet-scars, it is obviousnbsp;that it consists wholly of four coalesced roots, no portion of the true aerial stemnbsp;remaining in connection with it. It further demonstrates that the absence of allnbsp;traces of the rootlet-scars from most of the large specimens is due, as I have alreadynbsp;suggested, to superficial decortication and not to their previous non-existence. Anbsp;comparison of the general outline of Fig. 5 with that of Fig. 1 suggests, as alreadynbsp;remarked, that the latter example also represents roots only. The maximum heightnbsp;of Fig. 5 above the table, when standing upon its four rootlets, is only 6 inches ; thenbsp;diameter of its upper extremity, a, is about 6 iuches. From c to d it measures 10nbsp;inches. The mean diameter of the broken end of the root d is slightly more than 2ynbsp;inches,.the corresponding part of c being 2^ inches. The diameter of the root d

^ This figure is lithographed from a drawing made for me by Miss Beatrice Boyle, of Leeds, who promises to become a valuable auxiliary to scientific investigators who are not themselves artists.

at its base, from e to � in Fig. 5, is slightly less than 4 inches, and that of c, from e to g, exactly 4 inches.

Fig. 6 represents the under surface of Fig. 5, in which the crucial ridges seen in Fig. 4 are replaced by deep grooves. The line e of Fig. 4 is now representednbsp;by a fossa, 2f inches below the level of a line drawn from c to d. Fig. 6, a,nbsp;represents six of the rootlet-scars, of their natural size, as seen at the base, a, ofnbsp;the root. Fig. 5, d; and 6, h, is a copy of a similar cluster from near Fig. 5, x, ofnbsp;the same root. These two figures show how the Stigmarian rootlets accommodatednbsp;themselves to the increased growth of the root, not by any additions to theirnbsp;numbers, but by a molecular growth of the bark which pushed the rootlets furthernbsp;apart, this separation being accompanied by a corresponding increase in thenbsp;diameter of each individual rootlet.

The Xylograph 1 represents the outlines of the underside of a fine Stigmarian base of Sigillaria reniformis obtained by Mr. George Wild from the roof of the Bardsley Colliery, near Ashton-under-Lyne. In this specimen, which is ofnbsp;large size, the four primary roots become separated so near to the base ofnbsp;the stem that their division into four can only be made out by carefulnbsp;examination. Xylograph 2 is an outline of another specimen from the roof of anbsp;coal-mine at Honeywell Lane, also in the neighbourhood of Mr. Wild s colliery. I

am indebted to Mr. J. Hampden Barker for this sketch and for the careful measurements which accompanied it. The greatest diameter of the specimen was rather

more than 9 feet from the tip of root 1 to that of root 7. The shorter roots on each side of this longer axis could not be traced, since they plunged deep into the sidenbsp;walls of the mine. Whilst this specimen appears to subdivide into twelve roots'nbsp;the division into four primaries is even less obvious than in Xylograph 1.nbsp;Though Mr. Binney�s crucial line frequently exists, it is not a constant feature ofnbsp;these matured roots.

The above facts show that considerable differences exist in the length attained by these primary roots before they subdivide into secondary ones; whilst thenbsp;Duckinfield tree demonstrates that their ultimate branches may extend to greatnbsp;lengths with but little variation in their diameter.^

1 nbsp;nbsp;nbsp;The division between the roots 10 and 11 is not very distinct, but the otherwise symmetricalnbsp;arrangement of the roots leaves little doubt that a division exists there.

2 nbsp;nbsp;nbsp;This fact is not unimportant since M. Eenault is disposed to recognise in such elongatednbsp;examples a distinct class of organs from those which dichotomise near the base of the stem. He says,nbsp;� Cette observation est difficilement explicable par I�existence des simples racines,� � Cours denbsp;Botanique fossile,� premi�re ann�e, p. 154. At p. 9 of his � �tude sur les Stigmaria, rhizomes etnbsp;racines de Sigillaires,� he says, � L�epaisseur considerable que conservent les racines principalesnbsp;jusqu�a leurs dernieres sous-divisions, la grande r�gularit� de la ramification, le mode de dispositionnbsp;et la desarticulation des organes appendiculaires donnent aux Stigmaria un caract�re tout particulier que ne se rencontre dans le systeme radiculaire d�aucun autre type v�g�tal, et Ton senbsp;demande si ces curieux fossiles ne repr�senteraient pas plutot des rhizomes que des racines. Dans cenbsp;cas, les organes pris pour de radieelles et auxquels Groeppert attribue le nom de fibrilles, ne seraientnbsp;autre chose que des feuilles souterraines repr�sentant les feuilles rudimentaires appel�es �cailles desnbsp;rhizomes ordinaires.�

I shall have to show on a later page that none of these statements are applicable to Stigmaria ficoides. Subterranean rhizomes amongst the living Lycopodiaceee differ very little in structure fromnbsp;the stems of which they are mere extensions. This is shown by the example of Psilotum triquetrum.nbsp;But Stigmaria differs wholly both from Lepidodendron and from Sigillaria in wanting the inner vascularnbsp;zone seen in both these genera, the �tui m�dullaive of Brongniart, from which alone all the vascular

MEDULLA. nbsp;nbsp;nbsp;9

Having tLus glanced at some of the general aspects of Stigmaria ficoides, we may now examine the morphology and histology of its several parts.

This was an exclusively parenchymatous tissue; and, since I found it to be hollow in every one of the innumerable specimens that passed under my eye, I longnbsp;ago arrived at the conclusion that Stigmaria possessed a fistular medulla. Owing

bundles going to the leaves originate. In the second place, what can be more regular than the dichotomy of the rootlets of the living Lycopods. I shall further show on a later page that thenbsp;rootlets of Stigmaria, the � feuilles rudimentaires � of M. Eenault, did not disarticulate like leaves, ornbsp;leave a true leaf-scar. They have a distinctive internal organisation common to the entire series,nbsp;respecting which M. van Tieghem, our highest authority on the structure of roots, says, � Par lanbsp;structure du cylindre central, et par la division dichotomique qu�on y observe en plnsieurs endroits,nbsp;VOS racines appartient bien certainement a un Lycopodiac�e de la familie des Selaginell�es � (� Organisation of the Posssil Plante of the Coal-Measures,� Memoir II, p. 294). That Stigmaria does exhibitnbsp;some characteristic peculiarities, is unquestionable, but they are very different from those enumeratednbsp;by M. Eenault.

Since writing the foregoing pages, I have ascertained, to my surprise, that even some of my Oerman friends hesitate to accept the testimony of Binney, Dawson, and others on these importantnbsp;fundamental facts, and call for additional evidence that they are facts. This demand is easily met.nbsp;A few definite points are unquestionably proven.

I. nbsp;nbsp;nbsp;The gigantic Sigillarian stems must have had large roots. The specimens figured on my Plate Inbsp;demonstrate not only that they had such roots, but that these roots branched dichotomously.

II. nbsp;nbsp;nbsp;It is a characteristic feature of these roots that we invariably find them separating at thenbsp;base of the aerial stem into four primary ones, as represented on Plate II.

III. nbsp;nbsp;nbsp;The specimen figured on Plate III demonstrates that, when obtained in a sufficiently youngnbsp;state, these four roots were Stigmarian, bearing the characteristic rootlets of Stigmaria ficoides up tonbsp;the base of the aerial stem which they sustained.

IV. nbsp;nbsp;nbsp;The well-known Duckinfield example, now preserved in the museum of the Owens College,nbsp;presents four such primary roots, which dichotomise as in my Pigs. 1 and 2, and though these displaynbsp;DO traces of Stigmarian structure in their thicker portions close to the central aerial stems, theirnbsp;prolonged branches are absolutely Stigmarian.

V. nbsp;nbsp;nbsp;The absence of Stigmarian rootlet-scars from the proximal portions of the Duckinfield specimennbsp;is manifestly due to growth. The enlargement of each root having led to the decortication of thenbsp;superficial cortical zone of which the exterior of the base of each rootlet was but an extension, the resultnbsp;was the reduction of the thick, proximal end of each root to the condition of those of Pigs. 1 and 2.

We have here a few fundamental facts that it is absolutely impossible to dispute. They establish the truth that the Stigmaria ficoides is a root of these large trees; scientific evidence is renderednbsp;worthless if plain observations like these, made by a number of experienced observers, is to benbsp;lightly rejected. But I may add, from my own recent personal observations, that the conclusionsnbsp;arrived at from the above five propositions were amply sustained by observations which I made in thenbsp;fine Carboniferous forest recently exposed at Oldham, where I found both Sigillarian and Lepidodendroidnbsp;stems furnished with Stigmarian roots.

STIGMAUIA PICOIDES.

to the delicate structure of its cells this tissue was frequently destroyed; the area which normally represents its position being often entirely empty. But I havenbsp;obtained numerous examples in which this medulla is preserved as a very thin layer,nbsp;lining the vascular cylinder, as represented at Plate IV, fig. 7, a. The medullarynbsp;character of this layer is demonstrated by the way in which extensions of it consti-

Another point respecting which some friends entertain doubt refers to the possibility of the secondary roots of large trees, whose trunks and primary roots stand upon the coal, ever being sent into andnbsp;through the coal so as to reach the subjacent fireclay. The fact that these large roots are so constantlynbsp;cut off by the coal is in itself significant. We know that such roots attained to a length of as much asnbsp;fifteen and twenty feet, yet those referred to are abruptly truncated by the coal-seam at not more thannbsp;three or four proximal feet of their length. What has become of the remainder of each such a rootnbsp;if it has not entered the vegetable soil now converted into coal ? The original description of the Dixonnbsp;Dold trees (Plate I) was written by JohnEddowes Bowman, E.G-.S. and P.L.S., whom I knew long andnbsp;well as an experienced and accomplished geologist, trustworthy in the highest degree. He says of thesenbsp;trees, � The material on which they stand is a thin bed of pure bright coal eight or nine inches thick.nbsp;It has already been said that the roots do not penetrate the coal, but are abruptly cut off at its surface,nbsp;and that the immersed portions have probably been acted upon by the chemical changes going on^ innbsp;the surrounding vegetable mass so as not to be distinguishable from it (� On the Fossil Trees latelynbsp;discovered on the line of the Bolton Bail way near Manchester,� �Transactions of the Manchesternbsp;Geological Society,� vol i). Having been personally familiar with these trees at the time of theirnbsp;discovery, I can vouch for the correctness of my old friend�s description. The Duckinfield tree, alsonbsp;referred to above, affords a still more striking example. The truncated stump of the aerial stem wasnbsp;found with the bed of coal, two feet six inches thick, resting upon its broken, upper extremity. HJot onenbsp;of the long roots, now in the Museum of the Owens College, went deeper into the fireclay than threenbsp;feet. Hence the stump, when a living tree, must have ascended high above the upper surface of thenbsp;vegetable soil now represented by the coal-seam. It would be easy to multiply examples of a similarnbsp;kind, found both here and in the North American coalfields. Not unfrequently such trees are metnbsp;with apart from coal-seams. The 8t. Helen�s tree was in this position, and many of the trees in thenbsp;Oldham Forest were similarly circumstanced. Some of these were very young ones, and may notnbsp;have lived long enough in the positions occupied by them to accumulate, over their roots, sufficient vegetable soil to form a bed of coal; or such as was accumulated may have been washed away again beforenbsp;sedimentary sand and mnd took its place. That such local denudations have occurred is well known.nbsp;In a Memoir read to the Manchester Geological Society, on the 2nd of February last, Mr. Wild,nbsp;speaking of a coal to which he has recently sunk, says, � In the seat of this mine, which is a brownnbsp;stone, Stigmarise are both abundant and good, and both roots and rootlets pass through 'fournbsp;different layers of rock and shale to the depth of six feet below the coal. At about ten yards belownbsp;the coal (New Mine) is found a coally-looking floor or parting, overlain by dark shale containing verynbsp;well-formed septaria, whilst under the parting is a well-developed coal-seat or warrant (certainly anbsp;misnomer in this case), four to five feet in thickness, the Stigmariae in which are exceedingly good.nbsp;Such cases of an under-clay crowded with once vigorous roots and rootlets capable of supportingnbsp;gigantic trunks with foliage and fruit, being almost entirely robbed of the vegetation it had succourednbsp;by that relentless disturber, denudation, are by no means rare.� Familiar with the pit to which Mr.nbsp;Wild refers, and having collected some of the magnificent Stigmarise of which he speaks, I cannbsp;confirm his statements respecting these seat-beds and their vegetable contents.

But I am further asked by one of my doubting friends, has anyone ever found a Stigmarian root or Sigillarian stem passing through the coal into the fireclay below ? To this I answer yes. I am

tute the primary medullary rays (fig. 7, V V). Vegetable fragments wMcb do not belong to the individual Stigmarian root, especially tbe ubiquitous rootlets of othernbsp;individual Stigmarim, frequently found their way into the fistular cavity. Severalnbsp;writers have seen the vascular bundles of these rootlets within the medullary cavitiesnbsp;and have mistaken them for elements belonging to the root within which theynbsp;observed them. In other words, they supposed that Stigmaria had not only a solidnbsp;medulla, but that vascular bundles ran longitudinally through the centre of that

again indebted to my experienced and accurate friend Mr. Wild for a good example represented in the accompanying diagram. (Xylograph 3.)

This example is from the Dandy bed of coal, nearly fifty yards above the Arley Mine, at Tulledge, near Burnley. The specimen was a good Stigmarian root which was traced from near thenbsp;roof of the coal (the latter being about two feet in thickness), through which it gradually descended,nbsp;and entered the fireclay seat below. About four feet of the root was in the coal and nearly nine feetnbsp;in the seat. The root in this case had been preserved from destruction by the agencies referrednbsp;to by Mr. Bowman, quoted on a previous page, viz. by the fact that its cast had been filled at annbsp;early period with sandstone derived from patches of similar sandy material found in the roof of the coal.

The next diagram represents a state of things met with in November last in Mr. Wild�s colliery nt Bardsley known as the Pomfret Mine.

In this instance a is a Stigmarian � stool � embedded in the fireclay. Immediately above it, in the roof of the mine, was a � pot-hole,� h, i.e. a hole from which part of an aerial stem was extracted andnbsp;which there can be no doubt was the stem of the roots a.

Xylograph 5 represents another instance just discovered at the Bardsley Colliery, where there are two seams of coal separated by a thin parting of clay. Mr. Wild found a large stem ascending

STIGMARIA FICOIDES.

medulla and were the sources whence the vascular bundles going to the rootlets were derived. This, however, is altogether a mistake.^

from the clay parting, through the upper coal, and protruding somewhat into the roof of the mine. On removing this stem, the impression left by its base on the subjacent fireclay, exhibited, verynbsp;clearly, the crucial marking of Plate II, fig. 4. The fireclay separating the two coals was thereforenbsp;the soil in which this stem had commenced its growth and into which its roots had penetrated.nbsp;In like manner, Mr. Bradbury has obtained in the Bent Mine, a stem which was bared eighteennbsp;inches below, and which not only ascended some distance into the coal but reappeared immediatelynbsp;above it. A record kept, for but a few months, of all the cases of stems penetrating, or continuous through, the coal would supply a sufficient number of them to make any doubt as to thenbsp;possibility of such occurrences most unreasonable. At the same time it is easy to understand whynbsp;such should not be the ordinary positions in which such stems and roots would be preserved.nbsp;Xylograph 5 certainly represents an instance of a tree which began to grow when the lower coalnbsp;was covered by a layer of fireclay, in which latter it took root, and which continued to grow sufiicientlynbsp;long to allow the base of its stem to be imbedded in the vegetable soil ultimately converted into thenbsp;upper coal. In the same way, if the stem amp; of Xylograph 4 really belonged to the base, a still morenbsp;remarkable survival of the tree-trunk must have taken place. But the length of the life of the stemnbsp;necessary, in all probability, for the accumulation of sufficient vegetable soil to form a thick bed ofnbsp;coal would be so great as materially to exceed the duration of any tree�living or dead. Hencenbsp;it is that we so frequently find these rooted stems resting upon the coal, into which, however, theirnbsp;Stigmarian roots freely plunged.

Having so many proofs that some of the examples of Stigmariee discovered in the fireclay or seat-bed are the downward extensions of Sigillarian and Lepidodendroid trees, it surely can nonbsp;longer be doubted that the fragments of this identical Btigmaria ficoides with which that clay is sonbsp;constantly filled must also be portions of similar roots. Such fragments, both of roots and rootlets,nbsp;are extremely abundant. Indeed, it is rare to find a fireclay in which such is not the case; but hownbsp;these roots have so often become disturbed and broken up is a question not easily answered.

1 In p. 214 of my Memoir II, � Phil. Trans.,� 1872, referring to a retrogressive tendency on the part of several writers on Stigmaria, I said �� the first movement in the wrong direction originatednbsp;with Professor Goeppert, who described a Stigmaria (� Genres des Plantes Fossiles,� tab. 13) withnbsp;vascular bundles passing longitudinally through the pith, and from which he believed the vascularnbsp;bundles going to the rootlets were supplied. In this he was followed by Sir Joseph Hookernbsp;(� Memoirs of the Geological Survey of Great Britain �), who clearly affirmed the existence ofnbsp;medullary rays and bundles, but adopted Goeppert�s idea as to their origin.� � Mr. Binney recognised

VASCULAR CYLINDER.

TLat the bundles which Goeppert and Binney found in the interior of these Stigmari^ were those of Stigmarian rootlets is undoubted, but those rootlets hadnbsp;no individual relationship to the plants in which the two authors found them.nbsp;They were such as had intruded themselves into the medullary cavity from without,nbsp;just as they have entered into nearly every fragment of a plant from the Oldhamnbsp;and Halifax deposits, whence most of our rich stores of specimens have been derived.nbsp;Permeating both the vegetable soil, now converted into coal, and the seat clay, innbsp;the most extraordinary profusion, the smallest opening in any part of a vegetablenbsp;fragment was penetrated by these rootlets, making the study of such specimensnbsp;extremely perplexing to palaeontologists whose eyes are not familiar with the aspectsnbsp;assumed by these erratic rootlets. Plate VII, fig. 14, y, presents an instance ofnbsp;a rootlet, with its central vascular bundle, � �, in the medullary cavity or a Stigmariannbsp;vascular cylinder. In Plate X, fig. 42, we have a large rootlet, y, into the interiornbsp;of which several younger, but otherwise similar, rootlets have penetrated. Mynbsp;cabinet contains another example in which a small rootlet has penetrated anbsp;somewhat larger one, and these two, in turn, have entered together into a thirdnbsp;of yet larger dimensions. Mr. Binney unfortunately adhered to his error even innbsp;his latest writings.

The transversely barred tubes composing this cylinder belong to the type designated by Brongniart Vaisseaux ray�es. They are either vessels or Tracheidsnbsp;(Plate VI, fig. 9, b; Plate VII, figs. 10, b, and 11, b), assuming the latter formnbsp;especially where lateral bundles are given ofi.

At the earliest appearance of this cylinder in a young root the vessels occupying the position, though not fulfilling the functions, of the protozylem of an exogenousnbsp;stem, constituted a thin ring of very small vascular bundles surrounding a medulla.nbsp;These bundles, the vessels of each of which retained their mutual parallelism, didnbsp;not themselves pursue a straight, longitudinal, but an undulating, course throughnbsp;the stem, as at Plate V, fig. 8, the undulating curves of one bundle being opposednbsp;to those of its neighbour on either side. The result of these wavy undulations wasnbsp;that contiguous bundles alternately touched and separated from one another,nbsp;enclosing, in the latter case, large, vertically elongated, lenticular spaces (fig. 8,b'),nbsp;occupied by extensions of the medullary parenchyma which thus reached the bark.nbsp;As the vascular cylinder grew exogenously each new, superadded vessel followed

tbe medullary rays, but, as already mentioned, again adopts Goeppert�s explanation of the origin of the vascular rootlet bundles.�

STIGMARIA FICOIDES.

exactly the undulating course of those upon which it rested. The result was that these lenticular spaces became widened, without interruption, into what I havenbsp;elsewhere designated primary medullary rays; and which thus continued to receivenbsp;peripheral additions to their length so long as the vascular cylinder continuednbsp;to increase in diameter (Plate VI, fig. 9, h').

When tangential sections are made of any portion of the vascular cylinder, these medullary rays are intersected transversely, and always present, in such sections, anbsp;vertically elongated lenticular outline (Plate V, figs. 8, h', and 16, 5). On makingnbsp;two such sections of the same ray, one near the cortex and another close to thenbsp;medulla, as in Plate IX, fig. 12, and Plate V, fig. 13, the former being thenbsp;medullary and the latter the cortical section, it will be seen that the size of thenbsp;latter greatly exceeds that of the former. In other words, these rays, whichnbsp;in my � Memoir,� Part II (� Phil. Trans.�), I have designated primary medullarynbsp;rays, increase in size as they proceed from within outwards.

The same result is seen in transverse sections of the vascular cylinder (Plate YII, fig. 14, b' V). These rays were normally filled with an outward extension ofnbsp;the delicate medullary parenchyma, but this tissue has often failed to be preserved.nbsp;In the section, Plate lY, fig. 7, V V, we find the cells of this tissue elongated innbsp;the direction of the ray, but in such sections as are made at right angles to the longnbsp;axis of each ray, as in Plate Y, fig. 16, 6, the tissue resembles a delicate small-cellednbsp;parenchyma. Vascular bundles derived from the xylem cylinder (Plate Y, fig. 16,/},nbsp;to be referred to again more fully, are deflected outwards through these rays onnbsp;their way to the rootlets.

A transverse section of an entire vascular cylinder (Plate YII, fig. 14, and Plate VIII, fig. 15, h) exhibits its component vessels grouped in wedge-shapednbsp;segments of unequal sizes. This segmentation is almost wholly due to the intervention of the primary medullary rays. The variable diameters of the wedgesnbsp;depend upon whether the section has crossed individual rays at their broader part,nbsp;or at their narrower superior or inferior extremities, where they contract to thenbsp;dimensions of the ordinary, inconspicuous medullary rays. As already stated, thenbsp;cellular tissue, extended from the medullary parenchyma to fill the primary rays,nbsp;appears in such sections as Plate Y, figs. 13 and 16, as if it also was parenchymatous, but in sections like Plate lY, fig. 7, V it assumes a prosenchymatous form.nbsp;In vertical radial sections it exhibits a more mural aspect.^

Besides these large �primary� medullary rays the vertical laminae of the vascular cylinder are separated by numerous smaller �secondary� rays (Plate lY, fig. 7�, bquot;, and fig. 20, bquot;, hquot;). In radial longitudinal sections (Plate YI, fig. 9, hquot;'nbsp;b'quot;) these secondary rays are arranged in unequal groups often composed of severalnbsp;superimposed rows of cells. In transverse sections of the cylinder they are un-1 See � Memoir,� Part II, Plate xxx, fig. 43, �� (� Phil. Trans.�).

t quot;V^H

equally conspicuous. In tangential sections th y nbsp;nbsp;nbsp;consisting of several sucR

fig. 11, r hquot;), sometimes limited to a nbsp;nbsp;nbsp;radially elongated paren-

cells. In tlie majority of cases these rays^ nbsp;nbsp;nbsp;Tracheids, as in hquot;, Plate IV, fig_

chyma, hut they sometimes contain s nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;^ e of this substitution beyon

what follows. nbsp;nbsp;nbsp;,nbsp;nbsp;nbsp;nbsp;�� fin- 91 and in Part XI, P- 294,

In �Memoirs,� Part II, p- 236, nbsp;nbsp;nbsp;V p showed that similar barred

of Lepidodendron, and that Corda found em i nbsp;nbsp;nbsp;traeheids in themedullary

is probable, therefore, that the presence � nbsp;nbsp;nbsp;^ gigillarise, whilst they are

equivalent to a cambium-zone, is no lonpr isp nbsp;nbsp;nbsp;Staffordshire coalfield,

demonstrate this fact. My largest cylin er nbsp;nbsp;nbsp;possessor of the well-known

V, fig. 16, A). A third, equally perfect, is je nbsp;nbsp;nbsp;^^onging to the section of the

hark and rootlets represented in Plate 1 , g- � nbsp;nbsp;nbsp;original mean diameter

of the entire root, which has been about l� i constituted the xylem cylinder, disarranged state, are the vascnlar wedges t ^nbsp;nbsp;nbsp;nbsp;.pnbsp;nbsp;nbsp;nbsp;i..eh m

and which in all probability did not excee , nbsp;nbsp;nbsp;Oldham Forest, attache to

' speci�... quot;f stig�.ri.y

fcraile,' premi�re um�o, p. 153, 1881)- I� �be lt;�� .AAoieri�gt;� P''quot;' nbsp;nbsp;nbsp;�� ^ . I.

ftagmeatar, er.nrpl. o� thi. W- Sleiutam .aj. ( J'T/- nbsp;nbsp;nbsp;JetMted wUetnbsp;nbsp;nbsp;nbsp;*�

��ABoeget .t. L .�Wter �1 speciw.� e�m.a.d. nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;,k�

terminate, cloeing from a thickness of three me es o nbsp;nbsp;nbsp;� g�P,e author tells us a

STIGMARIA FICOIDES.

The rarity of small branching specimens, along with other facts, seems to show that after the first two consecutive bifurcations occurred within a limitednbsp;distance from the central stem, no others took place in the true roots. The long,

had seen no examples of Stigmarian roots less than two inches in diameter. Like Mr. Steinhaur, I am convinced that such examples of Stigmarian roots, terminating like thick cucumbers, werenbsp;abnormal, although my friend Sir William Dawson tells me that in Canada he has found ordinarynbsp;roots terminating in this obtuse form more frequently than we do in Great Britain. In allnbsp;probability some accidental cause had arrested the further longitudinal growth of such roots,nbsp;though they continued to swell transversely. Unquestionably the specimens described above, alongnbsp;with others in my cabinet, demonstrate that these roots ultimately tapered away to extremelynbsp;small twigs. Proof of this was abundantly furnished by the trees in the Oldham Porest, alreadynbsp;referred to. It is possible that some of the obtuse specimens spoken of above are fragments ofnbsp;examples like one described and figured by Mr. Eichard Brown, of which figure the accompanyingnbsp;Xylograph 6 is a copy. This memoir is entitled �Description of Erect Sigillariae with Conical

Tap-roots, found in the Eoof of the Sydney Main Coal in the Island of Cape Breton � (� Quarterly Journal of the Geol. Soc. of London,� vol. v, p. 354). The specimen figured is the under surface of thenbsp;base of a stem, in which the four primary roots divided and subdivided so rapidly that thirty-twonbsp;roots were seen within a circle of eighteen inches in diameter. � Besides this,� the author says,nbsp;� there are four large tap-roots in each quarter of the stump, as shown in fig. 7, and about five inchesnbsp;beyond these a set of smaller tap-roots, striking perpendicularly downwards from the horizontalnbsp;roots, making forty-eight in all, viz. sixteen in the inner and thirty-two in the outer set.�nbsp;Page 358, � The inner set of tap-roots vary from two to two and half inches in length, the diameternbsp;at their junction with the base of the trunk being about two inches.� � The outer set are muchnbsp;smaller, being about one inch in diameter at their junction with the horizontal roots, and from one tonbsp;one and a half in length. A thick tuft of broad, flattened rootlets radiates from the terminations ofnbsp;the tap-roots and a few indistinct areol� are visible on their sides.�

Eemembering that in many recent Lycopods the roots branch alternately in vertical and horizontal planes, the above description suggests that some Stigmarian roots have attempted to do the samenbsp;thing, though more or less abortively, owing possibly to unfavorable conditions preventing them fromnbsp;penetrating the soil. At any rate, we learn from Mr. Brown�s description that such undeveloped rootsnbsp;were capable of being produced exceptionally, which fact renders probable my explanation of the obtusenbsp;forms to which this note refers. It is possible, however, that the above plant may have been a formnbsp;distinct from our common type. If the letters i, k, I, and m, represent the subdivisions between the

VASCULAR CYLINDER.

terminal undivided roots, thus formed, may be identical with what led M. Renault to doubt their being roots and to regard them as rhizomes.

As already observed, the existence of this growth, alike in the aerial stems of Sigillarim, of most Lepidodendra, and in their common Stigmarian roots, is nonbsp;longer disputed. The great differences in the diameters of their several vascularnbsp;cylinders can be accounted for in no other way. But independently of the argument derived from the necessities of the case, our specimens demonstrate itsnbsp;occurrence. Thus in the section Plate VII, fig. 14, we find at hquot; hquot; a thin peripheral zone of vasoular laminae, sharply distinguished from those which it encloses,nbsp;by the much smaller size and less regular distribution of its component vessels.nbsp;A portion of this peripheral zone, enlarged 39 diameters, is shown in Plate IV,nbsp;fig. 19. At h we have the outer margin of the matured wedges of the oldernbsp;portions of the cylinder, whilst at V is the layer of new growths. Though thenbsp;vessels of the latter are of smaller, though unequal, sizes, they are obviouslynbsp;prolongations of the older laminae. They correspond, in these respects, withnbsp;similar young growths occurring in various other Carboniferous plants in whichnbsp;this exogenous development occurs.

The existence of such a process of exogenous growth demauds the pre-existence of some meristemic equivalent of a cambium. Plate IV, fig. iiOj represents such anbsp;layer. Like fig. 19, this is part of a transverse section of a vascular cylinder,nbsp;enlarged 75 diameters, the outermost vessels of which are seen at h separated bynbsp;the secondary medullary rays, V. At c we have a thin investing zone of what, innbsp;this section, appears to be an extremely delicate but otherwise ordinary form ofnbsp;parenchyma, the cells of which tend more or less to arrange themselves in radialnbsp;rows with parallel tangential divisions. Plate VII, fig- 10, is a radial, longitudinalnbsp;section through the same specimen as fig. 20, in which h again represents thenbsp;outermost vessels of the xylem cylinder. But we now see that the layer c of fig.nbsp;20 consists of narrow, vertically elongated cells with square ends, and which maynbsp;fairly be regarded as cambiform products of a cambial layer, the meristemic activitynbsp;of which may have manifested itself irregularly rather than periodically. Thenbsp;elements composing this cambiform layer being so very different from any whichnbsp;enter into the composition of the true cortex, we can scarcely doubt that theirnbsp;function has been as specialised as their structure and position, and that theynbsp;represent the zone within which the exogenous growths, successively added tonbsp;the exterior of the vascular cylinder, originated.

four primary roots, each of these has subdivided much more frequently, apart from the so-called Taproots, than is the case with our ordinary examples of Stigmaria ficoides..

Hitherto I have failed to discover any specimen showing the exact relations of the exterior of this camhiform zone to the innermost one of the true cortex ; hencenbsp;I am unable to say whether or not the former adds in any way to the growth of thenbsp;latter, or, in other words, whether any proper phloem zone exists in a Stigmariannbsp;bark. The probability that something of the kind will be found is suggested bynbsp;the fact that we appear to have a true phloem element in the vascular bundles ofnbsp;the rootlets, as will be shown on a later page. In Plate YII, fig. 10, we discovernbsp;dd some groups of what appear to be irregular cells intermingled with thenbsp;cambiform cells, and which appear to be inward extensions of a more externalnbsp;parenchymatous layer, the exact nature and relations of which I do not yet fullynbsp;understand.

This structure, so far as I understand it, consists of three zones, which pass more or less gradually into one another. In the cortex of my youngest specimennbsp;(Plate IX, fig. 18) only two zones are seen, d and d', as shown in Plate X, fig. 21,nbsp;enlarged 18 diameters. The outermost of these, d, is parenchymatous, andnbsp;the cells are without any special arrangement; in the inner zone, d', the cellsnbsp;are arranged in radial parallel lines; the transition from the one to the othernbsp;being rather abrupt. The thickness of the combined layers is about 2^ ( = �0^) ofnbsp;an inch, that of the inner one, h, being about to of an inch.

Having no longitudinal section of the above specimen, we learn from it nothing of the relations of the two layers. But when we turn to roots of largernbsp;dimensions and older growth, such sections throw some light upon the matter,nbsp;though their structure is more complicated. The outermost cortical layer (Platenbsp;VIII, fig. 15, d, and Plate VI, fig. 9, d) is always, as in the above-mentioned fig.nbsp;21, a simple parenchyma. But, though the lengths and breadths of the cells arenbsp;about equal in all directions, the thickness of the layer as a whole varies in differentnbsp;specimens. The sizes of the cells vary in the same specimen, and still morenbsp;in different specimens, the latter differences being due to age.

In the very young example, Plate X, fig. 21, the cells average from 4^ (�0025) to -g-^� (�00125), the mean thickness of this layer of parenchyma being about ^ ('016)nbsp;of an inch. In the older specimen represented in Plate VI, fig. 9, the correspondingnbsp;layer has a maximum thickness of about (�1) of an inch, and its cells vary fromnbsp;('006) to 3^ ('003) in diameter. Hence, whilst the outermost bark in thenbsp;older of these two specimens, as contrasted with the younger one, has increased innbsp;breadth six times, its component cells have less than three times the diameter ofnbsp;the younger ones. Such being the case these cells must have more than doubled

their mimlber. We thus learn that cell-growth in these primeeval trees seems to have obeyed the same laws as regulate that growth in plants now living.^

The second layer of the outermost Stigmarian bark is one of some interest, both morphologically and physiologically. In my � Memoir,� Part IX, p. 353,1 callednbsp;attention to the occurrence, in the Oldham calcareous nodules, of fragments of anbsp;singular form of bark, which I referred to Sigillaria and Stigmaria. I have nownbsp;no doubt respecting the accuracy of this reference, and that the bark is identicalnbsp;with that now under our notice.

To understand its true nature we must examine this tissue alike in transverse, radial, and tangential sections. Plate X, fig. 21, represents, as already observed, anbsp;small portion of the outer bark of the very young specimen, Plate IX, fig. 18, enlargednbsp;18 diameters; in this figure the layer now under consideration (fig. 21, d') appearsnbsp;as a series of cells, differing little from those of the investing parenchyma, d, exceptnbsp;in their more uniform size, and their regular arrangement in parallel, radial rows.

In Plate VIII, fig. 22, we have a transverse section of the corresponding portion of the specimen Plate YIII, fig. 15, also enlarged 18 diameters. In thisnbsp;section the layer d retains its parenchymatous form, but d', whilst still consistingnbsp;of radial lines of cells separated by tangential septas, not only has those septEe ofnbsp;very unequal lengths, but each radial group is circumscribed by a strongly-markednbsp;boundary line, dquot;, separating it from the contiguous parallel groups. In Plate YIII,nbsp;fig. 23, we have a similar section to fig. 22, also enlarged 18 diameters, but takennbsp;from a yet older root. We now find the layer d' of the preceding section hasnbsp;become more complicated. The rows of tangentially divided cells have lengthenednbsp;radially, and though tangential divisions still predominate, even at the morenbsp;external part of the layer d', we now find some radial divisions introduced amongstnbsp;the tangential septa of many of the groups. In addition, many of the innermostnbsp;cells of the parenchyma, d, contiguous to the outer ends of these groups, exhibitnbsp;a peculiarly disturbed arrangement that does not appear in fig. 22.

In Plate YI, figs. 9, d', and 45, d', we again see the layer just described, but now in radial vertical sections. The groups of cells are again disposed radially,nbsp;only the parallel tangential septa are now elongated vertically, instead of tangentiallynbsp;as in transverse sections. The radial boundary lines dquot; of fig. 23 are again seennbsp;in dquot; of this and similar sections. In my � Memoir � (IX, pp. 354, 355) I describednbsp;the cells thus separated from one another by tangential septa as � tabular cells

1 �An increase in the average breadth of the individual cell no doubt takes place, judging from estimates. It appears to rise rapidly to an approximately constant value, and then to maintain thisnbsp;during succeeding divisions, so that cells of the same layer in a stem a foot thick are no broader thannbsp;in one as thick as one�s finger, though they are of course more numerous in a corresponding degree.nbsp;The final, constant average dimensions are relatively little in excess of those existing originally at thenbsp;beginning of the growth in thickness; they may amount to scarcely more than two or three times thenbsp;latter.� (De Bary, � Comparative Anatomy of Phanerogams and Perns,� English translation, p. 538.)

whose broad parallel sides are parallel to the surface of the bark, whilst their shorter axes are radial. In a word, the cells stand upon their thin bevelled edges, withnbsp;their two flat parallel surfaces severally directed towards the medulla and thenbsp;periphery of the bark.� So far this description remains applicable to the corresponding cells of Stigmaria.

We learn something further respecting these curious cells from tangential sections of them. We have such a section in Plate VIII, fig. 24. We now findnbsp;that the dark boundary lines, dquot;, of Plate VIII, figs. 22 and 23, reappear as thenbsp;strongly marked walls of mother-cells, 24, dquot;, whose vertical lengths rather exceednbsp;that of their transverse ones, which explains the greater length of the parallelnbsp;tangential septa in the vertical section, Plate VI, fig. 9, d\ as contrasted withnbsp;their shorter lengths in figs. 22 and 23. But the tangential section further showsnbsp;that nearly every one of these tabular cells is undergoing secondary subdivisionsnbsp;(fig. 24, d!quot;, see also PI. V, fig. 49). Most of these secondary septa are horizontalnbsp;and parallel with one another, though occasionally a few vertical septa may also benbsp;noted.

I have dwelt with what may appear to be unnecessary minuteness upon the structure of this cortical zone, but the results of these detailed studies are notnbsp;unimportant physiologically. We are dealing with anomalous primEeval morphologies, and it is desirable to learn, so far as we can, the physiological truths whichnbsp;these morphologies seem to reveal. The facts stated above demonstrate that innbsp;the layer d' we have a peculiar meristem tissue or bark cambium of remarkablenbsp;activity, producing vertical cell divisions, seen in transverse and radial sections, asnbsp;well as horizontal divisions seen in tangential sections. Buch meristemic action isnbsp;obviously designed to make additions to the cortical structures. The next questionnbsp;requiring an answer is what part of the bark benefits by this meristemic activity ?

One portion of this question is easily answered. Internal to the zone d we have in all the more matured specimens a prosenchymatous zone, e, which, thoughnbsp;of very limited thickness in the younger roots, becomes the chief constituent ofnbsp;the bark in older roots. In all the transverse sections the cells of this zone arenbsp;arranged in straight, radial, parallel lines, the individual cells becoming graduallynbsp;smaller and more prosenchymatous as they approach the interior of the root.^ Innbsp;Plate VIII, fig. 22, e, they have a diameter of about tw� ('00083) of an inch. Atnbsp;fig. 23, e, of the same plate, their diameter is about -g-^o ('00125) to xw�' ('00083).

Plate IV, fig. 25, represents a transverse section, enlarged 3 diameters, of a specimen of Stigmaria where this prosenchymatous zone, e, approaches near tonbsp;the wedges of the vascular cylinder, b ; and Plate VII, fig. 26, represents a portionnbsp;of the same zone further enlarged to 18 diameters. The thickness of the prosenchymatous layer in the original of these figures, now about '65, must once havenbsp;1 Tangential sections of these cells are seen in PI. VIII, fig. 24, A.

been much greater, since the specimen has evidently been much compressed; the dark disorganised bands, e' (fig. 26), having normally been composed of the samenbsp;tissues as the intermediate bands, e, in which the radial lines of cells are lessnbsp;disturbed.

The preceding facts make it evident that the prosenchymatous zone e has originated in the centrifugal meristemic action of the zone d', a conclusion whichnbsp;agrees with my previous determination as to the centrifugal development of anbsp;similar prosenchyma in the stems of Sigillaria and Lepidodendron}

But a second physiological question is less easily answered. We have already seen that growth is accompanied, not only by a general increase in size and by anbsp;dilatation of the individual cells of the outer zone, d, of the bark, but by a considerablenbsp;increase in their number. Whence have these new cells been derived ? I havenbsp;never observed any signs of meristemic action amongst the cells actually composingnbsp;this layer, yet they must have been multiplied somewhere by such action. I amnbsp;therefore disposed to conclude that the meristem zone, d', of Plate VIII, fig. 23, hasnbsp;added centripetally to the outer parenchyma, d. If so, the true position of the bark-cambium must have been somewhere about the centre of the zone d', where the tangential septa were most closely approximated, and where the secondary horizontalnbsp;fissions, seen in the tangential section (Plate YIII, fig. 24), were being most activelynbsp;produced. This view, if correct, explains the disturbed arrangement at the innernbsp;part of the zone d in fig. 23, where the parenchymatous cells were being liberatednbsp;from the more internal radial lines in which they had undergone their meristemicnbsp;development, but had not yet assumed the characteristic form seen in the layer dnbsp;of fig. 22.

We have thus strong reasons for concluding that, in addition to a true cambial layer which produced fresh zones of xylem, these Carboniferous Cryptogams alsonbsp;possessed a bark-cambium which acted both centrifugally and centripetally, likenbsp;the phellogen of recent Exogens, only instead of producing true phellem externallynbsp;and phelloderm internally, this active zone produced parenchyma externally andnbsp;prosenchyma internally. What has already been described, however, shows thatnbsp;the meristemic energy must chiefly have been expended on its inner side, since thenbsp;prosenchymatous layer evidently constituted by far the largest element of thenbsp;Stigmarian bark.

In no case does this latter layer seem to have become a periderm. It is always, in the specimens which I have obtained, enclosed within and protected by thenbsp;parenchymatous zone, d, of which the rootlets are partly an extension, and uponnbsp;which they are planted. Hence we may be assured that so long even as the merenbsp;bases of these rootlets remained intact, the parenchymatous periderm remainednbsp;equally so. That such must have continued to be the case so long as the increase

STIGMARIA. FICOIDES.

in the size of the roots continued to progress is obvious, since it is clear that if the superficial layers of the bark had been successively thrown off until thenbsp;parenchyma wholly disappeared, the meristem zone, d!, must have been the nextnbsp;to perish, which would apparently have arrested all further growth in the thicknessnbsp;of the bark ; a result the occurrence of which does not seem to have taken place sonbsp;long as the trees continued to live. That the more superficial portions of thenbsp;parenchyma were so cast off is suggested by the gradual disappearance of thenbsp;rootlet-scars so conspicuous in Plate III, figs. 5 and 6, but which are either entirelynbsp;wanting, as in Plate I, figs. 1, 2, and Plate II, fig. 3, or very feebly represented innbsp;the extremities of Plate II, fig. 4, and Xylograph 1.

The Lower Coal-Measures of Lancashire and Yorkshire have furnished nearly all the trustworthy knowledge we possess respecting these organs. That knowledge, however, is now so complete that but little remains to be added to it.nbsp;Some little new light was thrown upon the subject in my Memoir, Part II, Platesnbsp;XXIX and XXX (�Phil. Trans.�). But since that Memoir was published muchnbsp;additional information has been obtained.

The longitudinal section of the zylem cylinder, Plate VI, fig. 9, illustrates this subject. All the vessels of special narrow radial laminse of the cylinder, from theirnbsp;innermost to their outermost margins, are suddenly deflected outwards, as at/, to benbsp;prolonged through the cortical zone as a rootlet-bundle. When the exogenous zonenbsp;of a very young root is in process of development, as in Plate VI, fig. 9, its outermost and newest formed Tracheids follow the course pursued by the older vesselsnbsp;on which they rest, and reappear in corresponding relative positions in transversenbsp;sections of each bundle as will be shown at page 31. As this bundle proceedsnbsp;outwards through the vascular cylinder, it increases in size transversely as well asnbsp;vertically. The formation of the bundle begins in the deflection of a solitary vessel,nbsp;which is the innermost and youngest one of a single radiating vascular lamina,nbsp;hence it is consequently a vessel of small diameter. As newer additions of largernbsp;vessels are made exogenously to the vascular cylinder, the portion of each newnbsp;growth that is in the radial line of the lamina originating the bundle contributes

^ M. Eenault says that these bundles � prennant naissance yers I�extremite interne des coins de hois,� � Cours de Botanique fossile,� premi�re ann�e, p. 156. This is true so far as the commencement of their formation is concerned; but, as will be shown in the text, they continue to arise from thenbsp;entire radius of that vascular wedge to whatever diameter it may attain. The erroneous notions ofnbsp;Professor Goeppert and Mr. Binney on this question have been already dealt with at p. 12.

to tLe further development of that bundle. But as this outward extension of the vascular cylinder progresses, more laminae than the primary one from which thenbsp;bundle originated, take part in that extension; hence, as it grows, the bundlenbsp;increases in diameter as well as in depth. This synchronous development, both bynbsp;the lateral and superior apposition of new vessels, explains the wedge-shaped contour which each of these bundles presents, when the deflected vessels are intersected transversely, as in the tangential sections of the vascular cylinder shown innbsp;Plate V, fig. 8, Plate IX, fig. 12, and Plate Y, figs. 13, 16. In consequence of thisnbsp;mode of development the smaller and oldest vessels of the bundle, forming thenbsp;apex of the wedge-shaped transverse section, invariably point downwards, or arenbsp;acropetal when considered in their relations to the root to which the rootlets belong.

Transverse sections of the xylem cylinder, like Plate VII, fig. 14, show these bundles passing outwards as at � and �. These sections also demonstrate thenbsp;increase in the lateral diameter of the bundle as it passes outwards. The twonbsp;directions of increase are well illustrated by comparing two transverse sections ofnbsp;the same bundle drawn to the same scale; the one, Plate IX, fig. 12, made tangentially through a vascular cylinder near its medullary surface, and the other,nbsp;Plate V, fig. 13, made close to the periphery of the cylinder. The primarynbsp;medullary ray, h, of fig. 12, like the similar one, Plate V, fig. 16, h, is comparativelynbsp;narrow, and the rootlet bundle, ��, passing through it, is also small, both in lengthnbsp;and breadth. Plate Y, fig. 13, represents a parallel section of the same ray asnbsp;fig. 12, but made close to the periphery of the cylinder. Both these figures arenbsp;equally enlarged 30 diameters. The increase in the magnitude alike of the primarynbsp;ray, h, and of the rootlet bundle, �, in fig. 13, is conspicuous, and correspondsnbsp;with what we observe in longitudinal sections of the rays and bundles as seennbsp;in transverse sections of the vascular cylinder like Plate YII, fig. 14, ��.

Each bundle of vessels, thus separated from the xylem cylinder to constitute a rootlet bundle, contains secondary medullary rays between its component vascularnbsp;lamin*, both in its vertical portions (Plate Y, fig. 16, d), and for a short distancenbsp;after they become deflected horizontally (fig. 16, d'). In these positions, unlikenbsp;those of the cylinder itself (Plate YII, fig. 11), they often consist of two or morenbsp;vertical rows of cells.

Transverse sections made of these rootlet bundles after their escape from the periphery of the vascular cylinder assume the various forms represented in Platesnbsp;lY, IX, and XI, figs. 28 to 36. In this portion of their course the bundles, asnbsp;seen in such sections, exhibit an irregularly triangular or wedge-shaped outline,nbsp;the broad base of each wedge consisting of the latest additions to the bundle;nbsp;and its narrow apex, �, corresponding to the points indicated by the same symbolnbsp;in figs. 12, 13, and 16. The figs. 28�34 are all taken from the periphery of onenbsp;vascular cylinder, and are enlarged 80 diameters. Figs. 35 and 36, similarly

STIGMARIA PICOIDES.

enlarged, are taken from Mr. Ward�s large Staffordshire vascular cylinder, on the periphery of which similar wedge-shaped bundles are numerous.^

As we have already seen, each of these rootlet bundles originates, when the root is in its youngest state, as a single very small vessel, the dimensions of whichnbsp;bear a ratio to those of the young cylinders whence it emanates. In like manner,nbsp;as the vessels of the newer additions to that cylinder increase in size, such portionsnbsp;of those additions as enter into the formation of the pre-existing rootlet bundlesnbsp;do the same. Yet after many such additions have been made, and the bundle hasnbsp;attained to relatively large dimensions, we not unfrequently find, added to its broadnbsp;base, and especially to its two basal corners, a few vessels, the diameters of whiehnbsp;are little more than those (�) of what may be called the protoxylem of the bundles.nbsp;Such additions are seen at/, Plate lY, fig. 31. As already pointed out, thesenbsp;later additions are derived from new and half-developed zones that are being addednbsp;to the exterior of the vascular cylinder, as represented in Plate lY, fig. 19, V.nbsp;We shall find that similar conditions, due to similar causes, reappear after thesenbsp;bundles have escaped from the cortex and constitute the central structures of thenbsp;actual rootlets.

As to the regularity of their radial arrangement, to which M. Renault attaches so much importance, the vessels composing these bundles, whilst they are passingnbsp;through the bark, exhibit a considerable variability. Plate lY, fig. 28, on the onenbsp;hand, Plate IX, fig. 35, and Plate XI, fig. 36, on the other, illustrate this statement. Yet that all these figures, including their innumerable intermediate modifications, are but vegetative repetitions of the same organ, is certain.

Plate XII, fig. 37, is a fragment of a Stigmaria, split vertically, which has revealed the exterior of the vascular cylinder at 5, whilst the rootlet bundles, �, escapingnbsp;successively from that cylinder, bend downwards and outwards. This is the onlynbsp;specimen I have obtained which demonstrates the regularity of this arrangement.

Plate XII, fig. 38, is one end of a portion of the vascular cylinder of a specimen exactly resembling fig. 40, from Mr. Bayles� brickyard near Leeds, and for whichnbsp;I am indebted to Professor Green, of the Yorkshire College of Science. Primarilynbsp;they and others were derived from below the Crow Coal. The medullary cavitynbsp;of fig. 38 is occupied excentrically by a large intruded Stigmarian rootlet with itsnbsp;conspicuous vascular bundle. Fig. 39 is a lateral view of the same specimen. Bothnbsp;these figures are of the natural size. On the exterior of fig. 39 we see numerousnbsp;rootlet bundles,/, escaping from the periphery of the cylinder in a regular order,nbsp;disturbed only by the circumstance that, the inner cortical zones having disappeared,

^ On various occasions M. Renault has affirmed that there are two distinct classes of these bundles, which he illustrates by Plate xx, figs. 1, 2, 3 of his � Cours de Botanique fossile,� premi�renbsp;ann�e. He says that whilst his wedge-shaped bundle, fig. 1, has supplied a leaf, figs. 2 and 3 are thenbsp;bundles of rootlets. The absence of all foundation for this distinction will be discussed on a laternbsp;page.

a ferruginous clay has occupied the vacant space and pressed the free, broken ends of the bundles into close contact with the surface of the cylinder, instead of retainingnbsp;them in their normal positions. Beyond what we learn from fig. 37, none of mynbsp;specimens give me any further information respecting these bundles until we againnbsp;meet with them in Plate XII, fig. 41. This figure represents the inner surface ofnbsp;a thin portion of the outermost bark of a Stigmaria, the opposite surface of whichnbsp;displays the usual rootlet-scars. The free ends of several bundles have been leftnbsp;exposed by the decay of the tissues through which they had passed, and intrusivenbsp;clay has acted upon them as upon those of fig. 39, though in the opposite direction.nbsp;I am indebted for tkese interesting fragments to Mr. B. Holgate, of Leeds, fromnbsp;whom Professor Gireen obtained the specimen (Plate XII, figs. 38 and 39).

We have now to examine the free rootlet as an external appendage to the bark. These rootlets are long, cylindrical bodies, of nearly uniform diameter throughoutnbsp;their length, which radiate with vegetative regularity from the entire circumferencenbsp;of the root. Their length and diameter alike vary with their ages. The longestnbsp;examples which I have measured have been twelve inches in length, but othernbsp;observers record examples that have been fifteen. The greatest diameter attainednbsp;by any of my uncompressed specimens is �4 of an inch.

Plate X, fig. 42, g, exhibits the usual appearance of their transverse sections. It is one of six small rootlets which have found their way, through some accidentalnbsp;opening, into the interior of the larger rootlet, g. The peripheral zone, gquot;, consistsnbsp;of parenchymatous cells, which in this instance have nearly uniform diameters,nbsp;though very frequently they increase in size from without inwards. Within thisnbsp;cortical zone we have, with rare exceptions, the vacant space, g ; in the specimensnbsp;from our Lancashire and Yorkshire nodules this space is always occupied by thenbsp;white infiltrated mineral substance, which permeates all the vegetable fragmentsnbsp;found in the nodules, and which has contributed so materially to their exquisitenbsp;preservation. The probability is that these rootlets were fistular, as is the casenbsp;with the rootlets of the living Isoetes lacustris. I have seen no trace of tissuenbsp;occupying it even in the youngest and smallest rootlets met with, except at theirnbsp;extreme bases where each rootlet is embedded in the exterior parenchyma of thenbsp;Stigmarian bark. Within the zone gquot; is the vascular bundle/', enclosed in anbsp;cellular cylinder which technical accuracy can scarcely allow us to call a bundlenbsp;sheath, though it seems to act as one. A similar example of a transverse sectionnbsp;of a rootlet intruded within the medullary cavity of the Stigmarian root is seennbsp;in Plate YII, fig. 14 (�' and g).

But whilst the structure of sections like the two just referred to is very simple and easily understood, a more complicated organisation exists in the basal partsnbsp;of these rootlets ; and longitudinal sections of these basal portions enable us tonbsp;understand much of their relations to the roots upon which they are planted.

26 nbsp;nbsp;nbsp;STIGMARIA FICOIDES.

Plate X, fig. 43, Plate VI, fig. 45, and Plate VIII, fig. 15, alike show that the parenchymatous layer gquot; and g o� Plate X, fig. 42, is a cylindrical extension o� thenbsp;outermost parenchyma, d, of the cortex of the root. The exact sources whencenbsp;some of the more internal layers of the rootlets were derived is less easilynbsp;determined, owing to our not having yet succeeded in tracing any bundles, innbsp;unbroken continuity, from the point of emergence from the vascular cylinder tonbsp;the periphery of the bark.

No one of the numerous sections of rootlet bundles, made, like figs. 28 to 33, Plate IV, close to the vascular cylinder whence those bundles have just issued,nbsp;shows any definite indications of being enclosed in a special bundle cylinder, suchnbsp;as invests each one when it reaches the base of a rootlet. Yet the rounded contournbsp;of the similar bundles of fig. 39, Plate XII, suggests that they must have possessednbsp;such an investment. We recover the interrupted continuity of the bundles atnbsp;Plate X, fig. 43,/, and Plate VI, fig. 45, �, at which point each bundle penetratesnbsp;a specialized tissue upon which every rootlet is planted. This tissue forms a short,nbsp;broad, cellular cylinder, h, which I would designate the rootlet cushion. It isnbsp;composed of innumerable parallel rows of very small cells, arranged radially,nbsp;which pass outward through the prosenchymatous zone, d', of the bark, and extendnbsp;into the interior of each rootlet, within which they terminate in a conical projection,nbsp;as in Plate VIII, fig. 15, h, Plate X, fig. 44, h', and Plate VI, fig. 45, h'. In somenbsp;instances, as in Plate X, fig. 43, h, its radial lines of cells describe complex curves.nbsp;In the latter figure, as also in fig. 45, we find the vascular bundle entering thenbsp;base of the cushion at �. In Plate VI, fig. 45, the bundle is unquestionablynbsp;encased in a small cylinder of delicate cells, fquot;, which are elongated parallel to thenbsp;bundle. In the specimen Plate X, fig. 44, the outer portion of the bundle, �, thenbsp;basal part of which is similarly, though less conspicuously, invested, has enterednbsp;the interior of the rootlet, as is also the case with the corresponding bundle of thenbsp;central rootlet of Plate VIII, fig. 15.

A tangential section of the bark, crossing a rootlet cushion transversely, as in Plate V, figs. 46 and 47, reveals a mass of very small parenchymatous cells, h, withnbsp;the vessels of the rootlet bundle in their centre at �. In some such sectionsnbsp;these cells are distinguished with difficulty from the vessels. The figures of thenbsp;above sections are enlarged twenty diameters. In Plate V, fig. 48, in which thenbsp;centre of fig. 46 is further enlarged to 200 diameters, the vessels are seen at �, whilstnbsp;a small vacant space at �' looks as if it had been occupied by the phloem portion ofnbsp;the vascular bundle. Surrounding the cells, figs. 46 and 47, h, we have a zone, h',nbsp;composed of rather larger cells. In fig. 47 the vessels are undistinguishable fromnbsp;the surrounding cells. Externally to the zone h' we have in fig. 47 the fusiformnbsp;cells of the prosenchymatous layer of the bark, e, through which the rootlet cushionnbsp;passes radially. In all my sections an extension of the outer layer, h', of the

cushion seems to fill a vertically elongated lenticular cavity in the prosenchyma-tous zone, which cavity reminds us of the forms of the lenticular sections of the primary medullary rays as seen in tangential sections of the vascular cylinder.

Sections made parallel to figs. 46 and 47, but intersecting and successively encroaching upon the basal part of a rootlet, still exhibit the tissue, h, thoughnbsp;in circles of decreasing diameter, until we reach the conical apex of the cushionnbsp;(Plate X, fig. 44, h,quot;), when it disappears. Longitudinal sections of the base of thenbsp;rootlet reveal the existence of a thin layer of a very peculiar tissue springing fromnbsp;the entire conical surface of the cushion. It is composed of elongated branchingnbsp;tubular cells, having a diameter of from -g-^ (�0012) to x6�oquot;� (�0006) of an inch. Anbsp;few detached examples of these cells are represented in Plate X, fig. 50. Thenbsp;cells being frequently much disorganised, I was long ignorant of their truenbsp;arrangement, but since Plate X, fig. 43, was drawn I have discovered them at thenbsp;angles, i i, of the rootlet cavity, radiating upwards and outwards from the surface ofnbsp;the cushion in parallel lines; the lowest of these lines reach the cortical layer, g,nbsp;of the rootlet; the more central ones, pursuing a parallel upward and outwardnbsp;course, have been merged in the cells of the zone (Plate IX, fig. 51, g') of this partnbsp;of each rootlet yet to be described.^

Plate IX, fig. 51, represents a transverse section of the basal portion of a rootlet enlarged'50 diameters. The conical part of the rootlet cushion is intersected transversely, forming the dark-coloured central zone, hquot;, enclosing thenbsp;vascular bundle, �. Externally, we have some of the parenchyma of the root-barknbsp;at d, within which is the outer or cortical zone of the true rootlet, g. The closenbsp;continuity of the cells of these two zones shows that this section has been madenbsp;just below the line * *, of Plate VI, fig. 45, where the parenchymatous layer of thenbsp;cortex, d, has been bent back upon itself to form the outermost layer, g, of thenbsp;rootlet. That this outermost rootlet layer is merely an extension of the outermostnbsp;layer of the bark is demonstrated by all the sections in which both are preservednbsp;together. Between the two zones, Plate IX, fig. 51, g and h!', we have the very distinct parenchymatous zone, g', which is not preserved in one section in a thousand,**nbsp;even in this basal portion of a rootlet, its place being almost invariably occupiednbsp;by white infiltrated mineral matter. It is, however, well preserved in thisnbsp;section, as is also the case in a young longitudinal section of the corresponding part of a rootlet. (Cabinet, Xo. 746.) The cells of this tissue chieflynbsp;range betweennbsp;nbsp;nbsp;nbsp;('00437) and 4^ (= �0025) of an inch in diameter. At

2 nbsp;nbsp;nbsp;It is a noteworthy fact that tissue occupying this position is always absent from the rootlets ofnbsp;Isoetes lacustris; but in this plant the rootlet bundle, enclosed in its investing cylinder, becomes finallynbsp;united by a few cells to the inner side of the cortical wall of the rootlet. This union always takes placenbsp;on the same side of the Isoetes rootlet, viewed relatively to the position of the deep fissure intersectingnbsp;the base of the stem of this plant.

STIGMARIA FICOIDES.

the inner and outer borders of the zone we find a few cells only about ('00125) of an inch in diameter. The size of these cells, contrasted with the extremenbsp;thinness of their walls, probably explains their almost invariable destruction.

Plate IX, fig. 52, represents the central part, �, of fig. 51 enlarged 440 diameters. The xylem of the vascular bundle is seen at/; whilst at / we observenbsp;a vacant cavity which, interpreting this section by others made further away fromnbsp;the base of the rootlet, we may regard as having been occupied by the phloemnbsp;elements of the bundle, whilst the cells, /quot;, correspond to those indicated by thenbsp;same symbol and immediately surrounding the bundle in Plate V, fig. 48.

In most of the transverse sections of these rootlet bundles made like Plate V, figs. 46 and 47, where the latter are passing through the inner and broadernbsp;portion of the rootlet cushion, the elements constituting the vascular bundlenbsp;appear to have their morphological arrangements modified by their contact withnbsp;the tissue surrounding them. But we find that a change takes place as soon as, ifnbsp;not before, the bundle escapes from the apex (Plate X, fig. 44, hquot;) of the cone of thenbsp;cushion; sections of the bundle at and beyond this point assume the features thatnbsp;characterise them throughout the entire length of the free part of the rootlet. Thesenbsp;features are seen in all the bundles represented on Plate XI; but before dealingnbsp;with them some other points demand attention. The moment the bundle escapesnbsp;from the apex of the cushion cone, within the rootlet, we find it encased within anbsp;small cylinder (Plate X, figs. 43 and 44, /') composed of linear rows of small parenchymatous cells. These cells are usually as broad as long; but sometimes, as innbsp;Plate IV, fig. 53, the innermost of them, /, are more elongated. This figure represents a longitudinal section of a portion of a small bundle, /, enlarged 100 diameters.nbsp;Owing to the absence of all tissue from the rootlet cavity, g', the bundles andnbsp;their sheaths rarely occupy their normal position in the centre of these rootlets,nbsp;but are usually more or less excentric.

We have already seen, from such sections as Plate VI, fig. 9, that the deflection of its vessels to form rootlet bundles took place simultaneously with the first appearance of a rudimentary vascular cylinder; and that as the cylinder increasednbsp;in diameter by the addition of centrifugal exogenous growths, a correspondingnbsp;increase took place both in the diameter of the bundle and in the number and size ofnbsp;its component vessels. Transverse sections of free rootlets demonstrate the order ofnbsp;that development. After collecting and carefully weighing all available evidence,nbsp;I conclude that the number of the rootlets given off from a Stigmarian root wasnbsp;finally determined during the youngest stage of the development of that root; nonbsp;addition to that number being made during its subsequent growth. It is at first

difficult to believe that a large root several inches in diameter does not possess more rootlets than a small one whose diameter does not exceed a minute fraction ofnbsp;an inch. Yet series of tangential sections, made from successive planes of a vascularnbsp;cylinder, display exactly the same number of such bundles, whether such sections arenbsp;made close to the medulla, or in immediate contact with the cortex. The primarynbsp;medullary rays through which these rootlet bundles emerge differ in this respectnbsp;from the secondary ones. The former do not increase in number with age. Thenbsp;latter do so indefinitely. The two modes of growth, viz. individual dilatation andnbsp;meristemic multiplication of their elementary tissues, have caused each individualnbsp;rootlet to increase in size; the same actions, taking place simultaneously in thenbsp;bark upon which the rootlets were planted, pushed those rootlets further andnbsp;further apart from one another, as is demonstrated by Plate III, figs. 6, a, and 6, 6.

Thus, some rootlets in my cabinet have only a diameter of �025, others reach �4, the latter being sixteen times larger than the former, without any change whatever being made in their characteristic organisation. We shall see, what was firstnbsp;demonstrated in my �Memoir/ Part XI, p. 291�93 (� Phil. Trans.�), that this increasenbsp;in the diameter of each rootlet is accompanied by an approximately correspondingnbsp;increase of the number and diameters of the vessels forming the rootlet bundle.

Plate XI, figs. 54 to 61,/, represent transverse sections of vascular bundles from within the interiors of rootlets of different ages, each with more or less of itsnbsp;investing cellular cylinder,/, and all equally enlarged 100 diameters. The differences between the magnitudes, both of the bundles and of their investing sheathsnbsp;are seen when we contrast figs. 54�6 with 58. In fig. 54,/, the formation of thenbsp;bundle begins with three extremely minute vessels or Tracheids (�quot;) closely combined at one point of the wall of the investing cellular sheath, �'; to which vessels anbsp;fourth larger one,/, has been added centripetally. In fig. 55 a similar arrangement exists, only we have here a fifth and yet larger vessel, �', added to the innernbsp;side of / of fig. 54. In fig. 56 we have five Tracheids, between which number andnbsp;what we see in fig. 57 any number of examples could have been figured. Thusnbsp;far the absolutely monarch character of the vascular bundle is clear.^ In fig. 57nbsp;the equally monarch bundle, /, is greatly increased in size contrasted with figs.nbsp;54�6. It now consists of at least eleven vessels, the smallest of which,/', retainnbsp;their monarch character as definitely as their representatives, /quot;, in fig. 54. Thenbsp;large size of the newer vessels,/, is conspicuous, and a similar enlargement is seennbsp;in the cells of the sheath, /quot;. In fig. 59 the number of the Tracheids hasnbsp;increased to fifteen, and in fig. 58 to sixteen. All the last three figures show thenbsp;bundle to be as independent of the surrounding cellular sheath, except at the point

� This point is interesting viewed relatively to an opinion entertained by M. van Tieghem respecting the origin of similar bundles amongst recent Lycopods, to which attention will again benbsp;directed.

STIGMARIA FICOIDES.

of primary orientation, �', as in the three preceding smaller figures. In fig. 59 we see at fquot; two vessels somewhat detached from the next one, fquot;, and more deeplynbsp;embedded amongst the cells of the investing cylinder than is usual; nevertheless,nbsp;these three linearly arranged Tracheids, along with those at f' constitute thenbsp;monarch point, to which the remaining vessels of the bundle were added centri-petally. In figs. 60 and 61 a further increase in the same^ direction is seen.nbsp;There are about sixty-four vessels in fig. 60 and seventy-five in fig. 61; thenbsp;apparent adhesion of the left-hand side of the bundle to the bundle sheath innbsp;fig. 60 is merely accidental, not organic. The only material changes to be notednbsp;are that the section of the latter bundle has become increasingly wedge shaped,nbsp;and the newer Tracheids, �quot;, are more obviously arranged in radiating seriesnbsp;than are the older ones, �quot;, changes which are yet more conspicuous in fig. 60.nbsp;It is important to note that every intermediate condition exists between thenbsp;pyriform bundles of figs. 57 and 58 and the wedge-shaped bundles of figs. 60 and 61,nbsp;since M. Renault has attempted to show that these extreme modifications of anbsp;graduated series represent important morphological and functional distinctions;nbsp;a subject to be discussed immediately. Equally important is the fact that eachnbsp;bundle commences its growth from a single point, �', not at three points, as isnbsp;afiirmed by M. Renault; nor yet at two semi-contiguous points, which become onenbsp;by coalescence, as M. van Tieghem believes to be the case with the representativenbsp;bundles in the roots of living Lycopods.^

We now know that on its primary appearance in any young root, the vascular xylem almost invariably takes the shape of two or more groups of minute vessels,nbsp;symmetrically arranged round the periphery of a central cylindrical strand ofnbsp;small meristematic cells. Sachs has designated this axial strand a procambium,nbsp;Nageli a cambium strand, Russow a desmogen, and De Bary an axial meristematic or initial strand of a vascular bundle. The number of initial vascular pointsnbsp;varies much. But the number is conveniently designated by the terms monarch,nbsp;diarch, triarch, amp;c., whatever that number may be. The peculiarity of thesenbsp;initial vascular points is that additions are made to them centripetally, until, innbsp;many cases, these additional growths meet in the centre of the young root,nbsp;occupying, more or less completely, the medullary area. The name of xylemnbsp;plates is given to these vascular growths. Midway between these several initialnbsp;points, or points of orientation, we find small patches of phloem.

This question only concerns us now so far as it affects our study of the Stigmarian rootlets, which it does very materially, because of distinct peculiarities

1 Of ]ate years the primary structure and ultimate development of roots has attracted a large amount of attention amongst the most distinguished European botanists. An important memoir bynbsp;M. van Tieghem (� Recherches sur la sym�trie de structure dans les plantes vasculaires. � Annalesnbsp;des Sciences naturelles,� 5m� s�rie, tome xiii) has done much to stimulate further inquiry into thenbsp;subject.

displayed by them, in common with the roots of living Lycopodiaeeous plants and of Opbioglossum. In most of the former and all the latter we only find one xylemnbsp;plate opposed to, or more or less surrounded by, one phloem element. Hence thesenbsp;living roots are unmistakeably monarch. M. van Tieghem, finding that where anbsp;rootlet dichotomises its bundle divides into two, one half going to each of thenbsp;secondary branches, came to the conclusion that the monarch appearance of suchnbsp;a bundle was due to the fact that it had divided in order to supply two branches,nbsp;one of which had become abortive. M. van Tieghem�s conception seems to be thatnbsp;two xylem and two phloem plates initiated so near to each other that the twonbsp;xylems blended to form one, and that the two phloem strands did the same. I havenbsp;never been able to accept this explanation, because of the contradiction which thenbsp;rootlet bundle of Stigmaria gives to it. De Barry affirms that there is no basis ofnbsp;fact for it.^

The history of the development of the xylem plate in Stigmaria makes it clear that it is absolutely monarch. Every stage of that development, whether wenbsp;study its orientation in the vascular cylinder of the root, its appearance within thenbsp;root-cortex, or its final structure within the rootlet itself, leads us to the samenbsp;inevitable conclusion. And the establishment of this conclusion respecting whatnbsp;was indisputably a primseval Lycopodiaeeous rootlet, may react upon our interpretation of the same organisation in its living representatives.^

The rootlets of Stigmaria, springing from the axial root, always incline more or less, as they grow, towards the growing end of that axis, enclosing a more ornbsp;less acute angle as they do so. The first formed Tracheids of the young rootletnbsp;bundle (�quot; of PI. XI, figs. 55, 57, and 58) always originate on the side of thenbsp;rootlet nearest to the growing tip of the root. This relationship is absolutelynbsp;constant. We have already seen from PI. VI, fig. 9, �, that some of the vesselsnbsp;and Tracheids successively added exogenously to the entire exterior of the xylemnbsp;cylinder, are prolonged into each rootlet bundle. Such additions to a bundle arenbsp;always made on the side of it that is turned from the growing tip of the root. Hencenbsp;in all sections of these rootlet bundles, like PI. V, fig. 16, the vessels � representnbsp;those first formed, whilst � indicates the newest additions. We have here a secondnbsp;absolutely constant relation. The growth in thickness in each rootlet bundle wasnbsp;steadily upwards and outwards from an inner and lower monarch starting-point.nbsp;The small Tracheids seen at �� of Plate IV, fig. 31, instead of being two additionalnbsp;points of orientation of a triarch bundle, are really amongst the latest additions to

� � Comparative Anatomy of the Phanerogams and Ferns,� English translation, p. 561.

2 M. Benault having observed examples like my figs. 57 and 59, Plate XI, in which a few of the Tracheids, last added centripetally to the rootlet bundle of Stigmaria, were very small,nbsp;arrived at the conclusion that their relations to the surrounding bundle cylinder were as primary asnbsp;those of the Tracheids marked/quot; in my figures 57 and 58 ; in other words, he believes that thesenbsp;bundles are triarch. Their entire history completely contradicts this interpretation.

STIGMAUIA nCOIDES.

that bundle, and their smaller size is due, as has already been explained on p. 22, to the fact that they have been derived from a younger layer of half-developednbsp;Tracheids like that seen at h' in Plate IV, fig. 19.

An invariable vegetative repetition of so complex an organisation as I have now described appears to be absolutely incompatible with the possibility of somenbsp;of the organs so constructed being phyllomes and others caulomes or emergencesnbsp;from caulomes. That members with such different functions as leaves and rootsnbsp;should possess so absolutely identical a structure, form, and direction of growthnbsp;seems to me too absurd to be conceived.

In several instances I have met with clusters of rootlets, a transverse section of one of which is represented in Plate XIII, fig. 79, where each rootlet has a thinnbsp;longitudinal cellular lamina, fig. 79, fquot;, looking like a centripetal extension of thenbsp;outer cortex, g, of the rootlet, and connecting that cortex with the bundle cylinder,nbsp;�. In many such examples the lamina appears to join the bundle cylinder exactlynbsp;opposite the point of departure of the bundle from its cylinder. I am not yet,nbsp;however, quite sure that this is a constant relationship between the position of thenbsp;lamina and the acropetal side of the rootlet.^

Plate XI, fig. 62, is a transverse section of a young rootlet with only three or four xylem Tracheids,/, but we find at/quot;the phloem cells of the bundle occupyingnbsp;the position which I have assigned to the phloem in Plate V, fig. 48, /', andnbsp;Plate IX, fig. 52,/.

Plate XIII, fig. 27, and Plate XI, fig. 63, illustrate another feature occasionally seen in these rootlets. Artis long ago represented some which were dichotomousnbsp;at their free extremities,^ and Oorda figured a similar example.� Plate XIII,nbsp;fig. 27, represents a similar dichotomous form, such specimens being occasionallynbsp;met with in our Lancashire deposits. Besides this my cabinet contains severalnbsp;transverse sections of what have been rootlets either preparing for or actuallynbsp;undergoing similar dichotomy. Plate XI, fig. 63, g, represents the inner surface ofnbsp;the external cortical zone of the rootlet, within which is the usual fistular cavity, g.nbsp;But at // we have two bundles that have originated from the subdivision of anbsp;primary one, each being enclosed in its separate bundle cylinder,/. The corticalnbsp;zone, g, has not yet shared this dichotomy, but in another of my sections it hasnbsp;done so. In it a broad belt of the cortical tissue g has extended completely across

^ M. Eenault has figured a rootlet bundle with its bundle sheath, from the exterior of which a similar cellular band radiates ; but he thinks he sees in the specimen evidence that a lateral branchnbsp;{radicelle) springs from the rootlet as well as that the rootlet bundle is double, half of whichnbsp;is developed centripetally, and the other half centrifugally. Nothing of this kind exists in any one ofnbsp;the innumerable rootlets in my cabinet. That a few of them dichotomise is shown in PI. XIII, fig. 27,nbsp;but such dichotomy is invariably truly terminal, not lateral or monopodial.

tlie section, dividing it into two areas, each of which has a bundle and bundle sheath, like those of fig. 63, in its centre. On the other hand, a section of a third rootletnbsp;displays the bundle divided into two, but even the bundle sheath has not yet begunnbsp;to divide. All these arrangements correspond very closely to what we find in thenbsp;branching rootlets of recent Lycopods. In the specimen figured by Artis, as wellnbsp;as in my fig. 27, the two branches appear to be joined to the primary one bynbsp;oblique articulations; but I find no trace of these in my sections. They werenbsp;probably mere constrictions of the cortical layer. In one section of a rootletnbsp;in my cabinet the bundle is enlarged laterally in a fan-shaped manner, as ifnbsp;preparing to divide. A union of the two bundles of fig. 63 would produce a verynbsp;similar contour to that seen in the above specimen. It appears as if this slightnbsp;tendency to dichotomous branching, manifested by the Stigmarian rootlets, wasnbsp;the forerunner of what became a normal condition amongst recent Lycopods.

Plate X, fig. 42, represents a transverse section of a large rootlet, g, into the interior of which six smaller ones have forced their way, in doing which they havenbsp;squeezed the true vascular bundle and its cylinder, �, of the invaded rootlet into anbsp;corner. This example affords a good illustration of the extraordinaiy way innbsp;which these rootlets penetrated openings, large or small, in any vegetable fragmentnbsp;within their reach.^

1 M. Eenault tas figured a similar specimen in his � Etude sur les Stigmaria, Ehizomes et Racines des Sigillaires,� � Annales des 8c. Gr�ol.,� xii, 1, PI. ii, fig. 1. Describing this specimen, he designatesnbsp;the invaded rootlet as a leaf, whilst the invaders are admitted to be true rootlets. But he gives nonbsp;adequate reason for thus applying different names to things that do not differ. Comparing his figuresnbsp;with similar specimens in my cabinet, I can only conclude that, misled by a foregone conclusion, henbsp;has allowed himself, in his pages 24�30, to be drawn into a confused maze of errors. He does notnbsp;deny that his leaves and his rootlets have the same external forms and internal organisation; the supposednbsp;difference to which he trusts in distinguishing leaves from rootlets being in the form of sections ofnbsp;their vascular bundles. It would be needless further to discuss a question with which I have alreadynbsp;alluded on p. 22, were it not for the important conclusions which M. Eenault draws from his supposednbsp;facts. I have already shown that the two types to which our author attaches so much importancenbsp;pass by imperceptible gradations into each other, and, I may add, that similar wedge-shaped and nonwedge-shaped bundles exist amongst the rootlets of living Selaginellm.

But the question assumes importance because it is made the basis of conclusions which set at defiance some of the most fundamental laws of botanical morphology, relating to the positions ofnbsp;members upon a common axis. The pages 24�30 contain a succession of statements which I cannotnbsp;accept. Describing a section like my fig. 14, Plate VII, he says that the vascular bundles which I havenbsp;indicated by �, �', � ne peuvent �tre pris pour des faisceaux de racines, dont ils n�ont aucun des caract�res�nbsp;(loc. cit., p. 21). I reply that, without one solitary exception, all those bundles go to those characteristic appendages of the Stigmarian axis which are now so widely recognised as rootlets. The illustrationsnbsp;given in the preceding pages surely demonstrate that point. M. Eenault further says: � Nous regardonsnbsp;ces faisceaux comme un portion des elements vasculaires destines a des appendices foliaires� (loc. cit.,nbsp;p. 22). We are thus carried back to the days of Artis, Corda, and the � Fossil Flora of Great Britain.�nbsp;The labours of Brongniart and Hooker, of Binney and Dawson, along with those of a host of othernbsp;observers are all to be cast aside as worthless. Before thus returning to the dark ages of Carboniferous

STIGMARIA PICOIDES.

There still remain for consideration some fragments of Stigmarise which, though not capable of microscopic examination, could not have been interpreted withoutnbsp;the knowledge which we owe to the microscope.

palaeo-phytology, I ask for some proof of the necessity for taking so extraordinary a backward step and I am referred to some insignificant differences in the forms of the transverse sections of somenbsp;small, variable, vascular bundles. After other equally inaccurate statements respecting some morphological details, M. Eenault says: � Le deuxi�me ordre de faisceaux ofire sur un section transversale lanbsp;forme de triangle �quilat�re ou scalene. Fig. 2 et 6.�nbsp;nbsp;nbsp;nbsp;� On distingue facilement sur deux ou trois

points, a, de la p�riph�rie du faisceau, des trach�ids de petit diam�tre, celles du centre �tant beaucoup plus larges, la section de ces derni�res est trois k quatre fois plus grande que celle des elements quinbsp;composent 1�ensemble des faisceaux du premier ordre d�crit plus haut, et que nous regardons commenbsp;appartenant a des organes foliaires; de plus,les trach�ides ne sont pas dispos�es en s�ries rayonnantesnbsp;a partir de 1�un des angles du triangle form� par la section, il est done evident que nous avons affaire anbsp;une autre sort d�organe et que ces cordons vasculaires sont des faisceaux de racines � (loc. eit.,nbsp;pp. 22, 23).

Tn the above passages M. Eenault enumerates what he regards as three distinctive characters, by means of which he recognises rootlet bundles. First, the unequal diameters of the Tracheids; secondly,nbsp;the triangular form of the section of the bundle, and lastly, the absence of a radiating arrangement ofnbsp;the vessels composing the bundle. It must be remembered that the bundles which he thus characterisesnbsp;are, according to him, something distinct from those seen at fig. 14 at �, �, and consequently also distinctnbsp;from those seen in tangential sections like fig. 8, �, of Plate V. That these latter sections merelynbsp;represent two aspects of the same organ is beyond all question; and since these are the only bundlesnbsp;discoverable within, or arising from the vascular cylinder, we may ask, whence and where do thesenbsp;apocryphal additional bundles arise ? Leaving this question, to which we get no answer, we maynbsp;inquire what value can be attached to the three other points ?

1. nbsp;nbsp;nbsp;Non-uniform size of the vessels. That this is a most variable feature I have already pointednbsp;out in my previous descriptions. The same bundle varies even in different parts of its course. Thus,nbsp;in sections like fig. 9 nothing is more common than to find bundles, the Tracheids of which are derivednbsp;from the larger vertical tissues of the vascular cylinder, reduced to an extremely small size whennbsp;deflected, as at Plate VI, fig. 9, �. I afflrm unhesitatingly that variations in the sizes of the vesselsnbsp;composing a bundle cannot be depended upon as a differentiating character.

2. nbsp;nbsp;nbsp;Triangular form of the bundle section. Where the bundles emerge from the cortical surface ofnbsp;the xylem cylinder and plunge into the bark, they almost always present more or less of the triangularnbsp;or wedge-shaped section; but their arrangement becomes entirely changed as they leave the outer cortexnbsp;to enter the rootlets, where their transverse sections become more or less pyriform. But besides thisnbsp;general fact, the forms of these sections vary considerably. We find many in which, as in Plate XI,nbsp;fig. 36, all the Tracheids are arranged radially; we have others, like Plate XI, fig. 61, in which onlynbsp;the outermost ones thus radiate, and others again, like Plate XI, figs. 57 and 58, in which there is nonbsp;radial arrangement whatever.

Definitions based upon such absolutely inconstant features are always worthless ; how much more so when they are depended upon to distinguish organs so widely different, both morphologically andnbsp;physiologically, as roots and leaves ? They would be worthless, even did other facts suggest a probablenbsp;existence of a combination of rootlets and leaves on the axis of a Stigmaria; but when, as is the case here,nbsp;all the known facts afford demonstrative evidence in the opposite direction, the employment of suchnbsp;variable features, for the purpose of overthrowing the conclusions of two generations of experiencednbsp;palffiontologists, can scarcely be regarded as wise.

In the first place, we have the most absolute vegetative uniformity in the orientation of all the

Plate XIII, fig. 64, is an inorganic cast of the medullary cavity of a Stigmaria, from the Hutton Collection, now in the Museum of the Natural History Society ofnbsp;Newcastle-upon-Tyne. The figure is of the natural size. The diagonal rows ofnbsp;oblong ridges covering the surface of the specimen are casts of the medullary endsnbsp;of the primary medullary rays of the vascular cylinder.

vascular bundles given off from the axial vascular cylinder of Stigmaria. Whether we examine transverse sections of that cylinder (Plate VII, fig. 14, �), longitudinal radial sections (Plate VI, fig. 9, �), tangential sections (Plate V, fig. 8), we arrive at the same conclusion. We find a number of vascularnbsp;bundles, springing from the vascular cylinder of the axis, in a uniformly characteristic manner. Allnbsp;these bundles pass outwards, through the primary medullary rays specially provided for their transmission, in a geometric order, which is not more disturbed by slight irregularities of growth than is thenbsp;case with the equally geometric phyllotaxis of recent leaves. When these vascular bundles emerge fromnbsp;the cylinder, to pass through the bark, they all bend downwards (Plate XII, figs. 37 and 39) in true rootlike fashion and which is the reverse of the course pursued by all the leaf bundles of Lepidodendroidnbsp;branches. On emerging from the outer bark, with the exception of an occasional derangement, resultingnbsp;probably from the arrested development of some rootlet when in a very young state, the quincuncialnbsp;arrangement of these rootlets in diagonal lines again becomes geometric. Turning to the structure ofnbsp;the rootlets, whose origin and course alike illustrate vegetative repetition in one of its most mechanicalnbsp;forms, we find that they display no material variation in the structure that is so characteristic of them.nbsp;The rootlet bundle is invariably monarch in its orientation ; its point of orientation is always acropetalnbsp;in relation to the growing root upon which each rootlet is planted.

In the face of facts like these, still to insist that structures, which are so obviously vegetative repetitions of one another, develop into a confused medley of rootlets and leaves, has no claim to ranknbsp;as a scientific inference. It becomes a mere preconceived idea adhered to in the face of an overwhelming array of opposing facts.

Unfortunately the mischief does not end here. What I have given my reasons for regarding as errors of observation have led M. Eenault to other conclusions equally unjustified by any known facts.

In his � Cours de Botanique � he has a paragraph headed �Mode de Croissance des Sigillaires� (loc. cit., pp. 162, 163,164). I regard this paragraph as full of unsupported hypothetical statements.nbsp;All our experiences in Great Britain, Canada, and the United States, in each of which countriesnbsp;Sigillarian and Lepidodendroid stems with Stigmarian roots are abundant, give to these hypothesesnbsp;an unqualified contradiction. No solitary instance can be shown in any of these countries in whichnbsp;� l�extr�mit� d�une branche de Stigmaria se relevait en bourgeon a�rien.� Even Brongniart tells us ofnbsp;�Tabsence de toute apparence d�un bourgeon terminale� (� Tableau des Glenres de V�g�taux fossiles,�nbsp;p. 56). It is unfortunate for science that M. Eenault should hold such views unless he could supportnbsp;them by more conclusive proofs than he has hitherto recorded. When we find these views receiving thenbsp;degree of countenance and circulation which MM. Saporta and Marion have given to them in theirnbsp;recent work, �L�Evolution du Kegne V�g�tale,� the matter becomes still more serious. It is thenbsp;support thus given to these retrograde opinions that in 1883 led Sir William Dawson to say, halfnbsp;despairingly, in his address to the American Association for the Advancement of Science, � Some onenbsp;will have to rescue from total ruin the results of our labours.�

I claim no monopoly of the knowledge of what is true; but I am entitled to ask that when a considerable number of practised observers, after many years of careful investigation, arrive at certainnbsp;definite conclusions, those conclusions should not be lightly disturbed. To justify such a course, thenbsp;disturber should be prepared with such strong evidence as very definite facts alone can furnish. Suchnbsp;facts, I contend, M. Eenault has not yet laid before us. In their place we have only got opinions !

STIGMARIA PICOIDES.

Plate XIII, fig. 65, is part of a fragment of a Stigmaria from a sandstone bed in the Mountain Limestone district of Weardale. When broken longitudinallynbsp;the specimen, fig. 64, was found loose in the cavity, 65, h. Both these specimensnbsp;were figured in Bindley and Hutton�s � Fossil Flora of Great Britain,� vol. i, pi. 35.nbsp;The concave surface, 65, 6, was supposed by the above authors to be merely a castnbsp;of the exterior of fig. 64, which latter was regarded as � a woody core communicating by means of woody elongations with the tubercles on the outside; this corenbsp;has evidently contracted, since the plant was embedded and now lies loose in thenbsp;cavity of the stem � (loc. cit., p. 106). This sentence affords a fair example of thenbsp;errors to which observers are liable when interpreting specimens of the histology ofnbsp;which they are ignorant. As we have just seen the supposed � woody core � is but annbsp;inorganic cast of the hollow interior of the true woody cylinder, the external surfacenbsp;of which cylinder is represented by the concavity, b. This cast of the latter surfacenbsp;displays the peripheral ends of the primary medullary rays where the rootlet bundlesnbsp;escaped from the cylinder to enter the bark. All the tissues between that cylindernbsp;and the outermost surface of the cortex have disappeared, being replaced by thenbsp;inorganic material, d, which has moulded itself upon the cylinder. The exteriornbsp;of this specimen shows the characteristic rootlet-scars.

Fig. 66 is a second specimen resembling fig 65, also from the Hutton Collection but which shows much more distinctly than the last does,nbsp;the casts of the oblique lines of large external orifices of the primary medullary rays. As in fig. 65, all the cortical tissues have disappeared, and werenbsp;replaced by soft sedimentary mud before the tissues of the vascular cylindernbsp;were decayed. This latter member also disappeared ultimately, both in figs. 65nbsp;and 66.

Plate XIII, fig. 67, is a transverse section of a Stigmarian root from which all the original organic elements have disappeared, the woody wedges of the vascularnbsp;cylinder, b, having been the last to do so. The inorganic sediment has herenbsp;occupied not only the whole of the cortical area but also the central medullarynbsp;cavity and the primary medullary rays, b', radiating from that cavity. The woodynbsp;wedges are now only represented by the dark, vacant spaces, b�b. Specimensnbsp;like this and the three just described are instructive. They demonstrate hownbsp;superimposed layers of tissue may have disappeared, not simultaneously, but innbsp;succession, and their places have been occupied by inorganic materials in a similarlynbsp;successive manner. Some of these materials have been introduced in a plastic state,nbsp;like those filling the areas a and d of fig. 67; but, had cavities like those left by thenbsp;decayed wedges, b, of fig. 67, instead of being left empty, been filled subsequentlynbsp;to the replacement of the other structures, a and d, by plastic sediment, this couldnbsp;only have been done by mineral matter in solution and capable of filtering throughnbsp;the clay. It has been such a deposition from infiltrated solutions that has occupied

ROOTS AND ROOTLETS.

not only tLe interior of most of the Stigmarian rootlets figured in this Memoir, but also of almost every cell and vessel found in the deposits from which a largenbsp;majority of my specimens have been obtained.

Plate XIV, fig. 68, is a specimen of a root, split vertically, drawn four-fifths the size of the original, for which I am indebted to Professor Green of Leeds. Itsnbsp;unseen exterior surface is furnished with the usual rootlet-scars. At a is the externalnbsp;surface of the inorganic cast of the medullary cavity, exhibiting, though rather morenbsp;closely aggregated, longitudinal ridges like those seen in Plate XIII, fig. 64. Thenbsp;well-defined longitudinal section of the vascular cylinder, h h, is transversely subdivided into small square areas by the primary medullary rays, V, which radiatenbsp;through the cylinder at right angles to its axis. The medullary cavity and all thenbsp;cortical zones are alike replaced by inorganic sandstone.

Plate XIV, fig. 69, is a specimen from the Burntisland deposit, represented of its natural size. This specimen would perplex an observer unfamiliar with thenbsp;internal structure of Stigmaria. It is the well-preserved external surface of anbsp;large vascular cylinder, exhibiting very definitely the lenticular external terminations of the primary medullary rays, h b, but amongst these are a few rootlet-scars,nbsp;as seen in the next specimen, fig. 70.

Plate XII, fig. 70, is also a fragment of the exterior of a compressed vascular cylinder, wholly composed of barred vessels or Tracheids; but in it the primary medullary rays are indistinctly shown. Forced rather deeply intonbsp;its substance are several rootlet-scars, g, arranged in their normal diagonal lines.nbsp;Such a specimen, seen apart from others, would inevitably indicate the directnbsp;orientation of the rootlets from the vascular zone. What has occurred is obvious.nbsp;The whole of the cortical tissues have disappeared, but with so little disturbancenbsp;that, on both sides of the specimen, the bases of the several rootlets have becomenbsp;impressed upon the exterior of the flattened vascular cylinder, without anynbsp;derangement of their normal relative positions. Such specimens teach cautionnbsp;ere we conclude that, because two tissues are found in the closest possiblenbsp;contact, they must once have been organically united.

Plate XIII, fig. 71. An impression on shale from the Hutton Collection. It is part of a dichotomising root, the surface of which exhibits, besides its rootlet-scars, parallel longitudinal ridges, which either represent fissures in the originalnbsp;bark, or elevations due to shrinkage ; between these ridges are fine undulating lines, also running longitudinally, which also appear to have been causednbsp;by a shrivelled state of the cortical surface. Another similar specimen, also fromnbsp;the Hutton Collection, exhibits these latter lines; but, in place of the coarsenbsp;longitudinal ridges of fig. 71, it has numerous strongly marked undulating ridgesnbsp;and furrows running transversely across the fragment. Varied modifications ofnbsp;the surface, especially in specimens of the larger roots from which all rootlet-scars

STIGMARIA FICOIDES.

have disappeared or are disappearing, are innumerable, and have no specific signification. The above figure is half the size of the original.

Plate XII, fig. 72. A cast of the outer surface of a small fragment of Stig-marian bark in which the spaces between the rootlet-scars are occupied by still more strongly marked undulating longitudinal lines. This seems to be the condition tonbsp;which Goeppert gave the name of Stigmaria undulata. The original is in thenbsp;Museum of the Owens College. I have seen similar specimens with correspondingnbsp;undulation, but so faintly preserved as to show that they merely represent anothernbsp;of the variable conditions of preservation just referred to.

Plate XIV, fig. 73, is a Stigmarian fragment displaying the more normal conditions of the rootlet-scars; they are depressions, disposed in diagonal lines,nbsp;producing what is known as the quincuncial arrangement. The large centralnbsp;cavity, h, is, like that of Plate XIII, fig. 65, the result of the disappearance of thenbsp;vascular cylinder from its medullary interior.

Plate XII, fig. 74, is a small fragment of Stigmarian bark from near Oldham, represented of the natural size. Since the tissues of the outer bark as well as ofnbsp;the bases of the rootlets occupying the bottoms of the sunken scars are preservednbsp;in this specimen, it becomes instructive, teaching the true histology of thosenbsp;depressed scars, and showing how the rootlets themselves have disappeared. Atnbsp;the outer margin, g, of each of these depressions we see the remains of the outermost or cortical zone of the base of each rootlet. Within this outer circle we havenbsp;the still deeper depression, h, produced by the disappearance of the second delicatenbsp;parenchyma, g', of Plate IX, fig. 51. At the bottom of the latter depression we seenbsp;traces of the apex of the rootlet cushion, h'. Specimens like this clearly shownbsp;that there has been no kind of � articulation� where the rootlet was planted uponnbsp;the bark. There was no definite plane corresponding to the cicatrix left by anbsp;fallen leaf at which the rootlet separated from the bark. The separation was thenbsp;result either of external force or of decay, producing a variable contour in whatnbsp;remained of the torn tissues of the rootlet. The rootlets of the living Isoetes lacustrisnbsp;illustrate this tendency to detachment by rupture, as contrasted with disarticulation.

Plate XIV, fig. 75, is an unusually fine specimen, from the neighbourhood of Newcastle-upon-Tyne, for which I am indebted to the kindness of Professornbsp;Lebour. Similar examples have been figured both by Sir Joseph Hooker^ and bynbsp;Mr. Binney,^ but both these authors have fallen into the same error in their interpretations of their specimens. Bach of them supposed that the surface which henbsp;figured was the true exterior of the bark, whereas it was exactly the reverse; it

1 Memoirs, � G-eol. Survey of Grreat Britain,� vol. ii, part ii, pi. ii, figs. 1, 2, and 3, 1848.

* � Carboniferous Plora,� part iv, Sigillaria and Stigmaria, pi. xxir, fig. 1, Palseontographical Society, 1875.

ROOTS AND ROOTLETS.

was merely tLe cast or impression of that exterior. They regarded the little projecting circles corresponding to those of my fig. 75, as identical with those of my Plate XII, fig. 74, believing the former to be the portions of the bark upon whichnbsp;the proximal ends of the rootlets were planted, whereas they are the actualnbsp;proximal ends of those rootlets. Specimens in my cabinet demonstrate that suchnbsp;is the case, since in them the remaining distal part of eaeh rootlet is seen passingnbsp;backwards through the stone to its opposite surface.

A glance at the diagram of a longitudinal section of a rootlet with a portion of the bark upon which it is placed (Plate XII, fig. 76) will probably make the historynbsp;of the specimen, fig. 75, intelligible. A fragment of such a bark has beennbsp;reduced to the condition seen in Plate XII, fig. 74. That this occurrence tooknbsp;place is shown by the fact that casts, which I have made of the surface of thatnbsp;specimen, correspond almost exactly with what we see in Plate XIV, fig. 75. Thenbsp;outer surface of the supposed bark was undisturbed, as at fig. 76, dquot;. Bach longnbsp;rootlet, 76, g, had either decayed from its tip backward or had been abruptly brokennbsp;off near the surface of the bark, where a little of the outer cylindrical wall of eachnbsp;rootlet stood in relief, as at fig. 76, g� gquot;, forming a funnel-shaped cup, as in Platenbsp;XII, fig. 74, g. At the bottom of this cup there would project the cone of thenbsp;rootlet cushion, as in the section Plate X, fig. 44, h'; embedded in mud, thenbsp;future matrix of the specimen fig. 75, that mud would fill the cavity, fig. 76, g,nbsp;and also surround its external wall and invest the outer surface of the bark, dquot;.nbsp;The unshaded paper to the right of the bark surface, dquot; dquot;, thus represents thenbsp;outlines of a section which would be identical with a similar section made throughnbsp;one of the rootlet bases of Plate XIV, fig. 75.

Plate XII, fig. 40, is a Stigmaria, the interpretation of which was not easy on a merely casual glance. The fragment is but a portion of a much largernbsp;specimen, almost identical with, and from the same locality as, that from thenbsp;interior of which fig. 39 was extracted. We see from fig. 38, which representsnbsp;the upper end of fig. 39, that all the vascular tissues of that cylinder are wellnbsp;preserved; but both vascular and cellular tissues have disappeared from the area,nbsp;�� of fig. 40, with the exception of a confused mass of vascular bundles; these arenbsp;evidently the remains of such portions of rootlet bundles as passed through thenbsp;bark, like those seen on the exterior of fig. 39, but which were left exposed on thenbsp;decay of the cortical tissues. At fig. 40, g g, the rootlets of one side of thenbsp;specimen pass outwards to the right hand, each in a flattened condition. Thenbsp;surface is n cast of the outside of the bark opposite to that which supplied thenbsp;rootlets, g g. The rootlets of d have not collapsed and become flattened afternbsp;they were invested by the then plastic matrix, though they subsequently disappeared, leaving empty cylindrical cavities which pass downwards and outwardsnbsp;through the stone. In several of these cavities the vascular bundle of eachnbsp;rootlet can still be detected, and two or three of the rootlet cavities present similar

STIGMARIA HCOIDES.

conditions to those seen in Plate XIY, fig. 75, with the morphology of which this part of Plate XII, fig. 40 is almost identical.

Plate XIY, fig. 77, is another of the many variable conditions in which we find the external surface of a Stigmaria, and which were largely due to the shrivellingnbsp;of the specimen before its immersion in its muddy matrix, followed by externalnbsp;pressure. Its configuration is virtually that of Plate XII, fig. 74.

Plate XIII, fig. 78, represents a fine fragment in the Museum of the Natural History Society of Newcastle. It is undoubtedly a portion of the plant to whichnbsp;Goeppert gave the name of Stigmaria stellata,^ though the rootlets given off fromnbsp;the large symmetrically arranged tubercles are much less perfectly preserved thannbsp;in Goeppert�s specimen. That this object has been a root with rootlets verynbsp;similar to those of Stigmaria ficoides appears probable. Whether or not it can benbsp;generically united with Stigmaria is doubtful. We have other plants in the Coal-Measures furnished with succulent rootlets besides Stigmaria, e.g. my genusnbsp;Amyelon. It appears to me that no plant should be regarded as a Stigmaria, thenbsp;internal organisation of which is not at least typically identical with that ofnbsp;S. ficoides, and which consequently may be regarded legitimately as the probablenbsp;root of some Lepidodendroid or Sigillarian stem. We have no proof that eithernbsp;the one or the other of these affinities exists in the object in question; hence, whilstnbsp;recognising its unquestionable specific distinctness from Stigmaria ficoides, I shouldnbsp;prefer for the present to refer to it as Stigmaria {?) stellata. The Newcastle specimennbsp;was apparently derived from one of the Gannister beds. I have more recentlynbsp;received from Mr. Kidston, of Stirling, a cast of another similar specimen, foundnbsp;loose in a Boulder Clay at Town-Head, Riccarton, in Ayrshire, by Mr. P. Wright,nbsp;of Galston.

Having now described all the more important morphological and histological features of the Stigmarian root which I have thus far observed, some questionsnbsp;arise connected with its relations with other plants, extinct and living.

The fact that large quantities of Stigmarian fragments have been found in several localities unassociated with any Lepidodendroid or Sigillarian stems has lednbsp;some geologists^ to � consider Stigmaria as originally representing floating stemsnbsp;becoming roots under peculiar circumstances.�

We find nothing in Great Britain which supports this or any similar conclusion. Hence British geologists are unanimous in regarding Stigmarim as the roots alike

* jamp;. ff. Lesquereux, � Coal Flora of Pennsylvania,� vols. 1 and 2, p. 509.

CONCLUSIONS.

ui Lepidodendra and of Sigillari^; and they are equally unanimous in believing that these primmval Lycopodiacem, found in the Devonian, Carboniferous, andnbsp;Permian strata, are the remote ancestors of the modern Lycopodiacem. Thenbsp;question arises how far does our present knowledge respecting the morphology andnbsp;histology of these ancient arborescent forms enable us to detect connecting linksnbsp;between them and their degraded living descendants.

That the dichotomous ramifications of the branches, the structure and arrangement of the leaves, and the entire morphology of their reproductive organs, furnish such links is indisputable. But we are now familiar with other morphologicalnbsp;features presented by these ancient types, the relations of which to those of livingnbsp;ones are not quite so clear.

That the axial vascular strand of the living Lycopodium and Selaginella is the homologue of the non-exogenous inner vascular zone of Lepidodendron can scarcelynbsp;be doubted. In both cases these tissues constitute the only axial vascular elementsnbsp;possessed by the youngest branches on which all the leaves are developed, and tonbsp;which leaves they supply the only vascular bundles that those leaves ever possess.^nbsp;These ancient and modern vascular axes also agree in their mode of growth whichnbsp;is in both cases centripetal.

There can be equally little doubt that the rootlets of Stigmaria correspond to those of living forms, both in their structure and their acropetal order of development. The vascular bundle in the centre of the Stigmarian rootlet, as well as thenbsp;cellular zone which invests it, is almost identical with that of Selaginella, andnbsp;approximates still more closely to that of Isoetes. In two respects the affinitiesnbsp;of the Stigmarian rootlets with those of Isoetes are remarkable. In both organisms these rootlets are given off from the lower part of a downward prolongationnbsp;of a caulome, which prolongation never develops leaves; the rootlets, therefore, arenbsp;produced upon an axis which grew in the opposite direction to that in which thenbsp;leaf-bearing part of the stem grew. In addition to this, the rootlets of Isoetes andnbsp;of Stigmaria agree in the circumstance that, in both, they are converted, during life,nbsp;into fistular cylinders, owing to the disappearance or non-development of thenbsp;delicate parenchyma, which ought to occupy the space between the outer corticalnbsp;layer and the investing sheath of the central vascular bundle. But importantnbsp;differences have been produced by the introduction, into both the stems and roots

1 M. Renault�s idea that in the many extinct forms which possess a diploxyloid vascular axis, i.e. an inner centripetal and an outer exogenous axis, each of these two cylinders contributed to thenbsp;formation of the leaf-bundles, cannot be accepted. In all cases the leaves and their leaf-bundlesnbsp;were developed before any exogenous zone made its appearance; and in several known Lepidodendroidnbsp;plants the branches attained to a large size before any such zone began to grow. To suppose that, in suchnbsp;cases, the leaves had to wait for their complete vascular structures until, having done their chief work,nbsp;they were ready to be cast off, is impossible. Hence I must reject this assertion that the foliar bundlesnbsp;had a double origin, as it is alike contrary to probability and to observed facts.

STIGMARIA FICOIDES.

of the extinct forms, of a true, exogenously developed, vascular cylinder. This zone with its radial laminae of vessels, its true medullary rays, and its meri-stemic cambium zone, encloses the vascular, axial strand which supplied thenbsp;vascular bundles to the leaves. These leaf-bundles were all fully developed, andnbsp;extended continuously through the bark, from the axial vascular strand from whichnbsp;they originated to the leaves on the surface of each twig, before (in many casesnbsp;very long before) the exogenous cylinder made its appearance. Hence when thenbsp;first-formed vessels of that cylinder arranged themselves longitudinally roundnbsp;the central axis from which the leaf-bundles emerged, they had to bend roundnbsp;each of the many foliar-bundles that stood in their way, coming together again whennbsp;they had passed the obstruction. This was also done by each successive exogenousnbsp;growth without any contribution being made by the latter to the foliar-bundle. Eachnbsp;additional exogenous layer pursued the course followed by that upon which it rested.nbsp;Hence each foliar-bundle passed outwards through the exogenous cylinder, alongnbsp;horizontal lenticular passages, the vessels enclosing the peripheral portions of whichnbsp;passages were successively superimposed upon the pre-existent bundles. Thesenbsp;relations of consecutive, not coeval, origin explain what observation demonstratesnbsp;to be a fact, viz. that the exogenous cylinder was a vascular network, through thenbsp;meshes of which the foliar-bundles continued to reach the bark, but withoutnbsp;receiving any additional vascular contributions.

This exogenous zone made its first appearance in the various Lepidodendroid trees at very different stages of their growth. In the L. selaginoides of Halifaxnbsp;we find it existing even in very young twigs. In the Burntisland Lepidodendronnbsp;it appears, not in the twigs, but in young branches. In the Arran Lepidodendronnbsp;from Laggan Bay, no traces of it are seen until the branches have attainednbsp;to a large size. It evidently began to be developed at the junction of the stemnbsp;with the root of each plant; attaining to a greater relative importance in thenbsp;latter than in the former organs, since we find it, as shown by Plate IX, fig. 18,nbsp;reaching the extremities of the true roots, which it never does in the twigs.nbsp;Since every rootlet derived its vascular strand from this layer, and sincenbsp;rootlets were obviously furnished to the plant in its youngest state, this earlynbsp;development of the exogenous zone in the roots was a matter of absolutenbsp;necessity. Such was not the case with the aerial parts of the plant, the leaves ofnbsp;which, as we have seen, obtained their vascular strands independently of the exogenous zone. But the latter evidently crept up the stem from below as a successionnbsp;of investing cones, each newer investment reaching a higher point than thosenbsp;which preceded it. Xow if these views are correct, which I believe them to be,nbsp;we can understand the functions of the vascular cylinder of Stigmaria. Thenbsp;mineral and nitrogenous food-material absorbed by the rootlets was conveyed tonbsp;the stem through the exogenous zone, whence it was transmitted laterally to the

CONCLUSIONS.

more central, non-exogenous cylinder, througli tlie brandies of whicb it passed to the leaves. Assuming that my morphological data are accurate, about which Inbsp;have no doubt, the above appears to me to be the only physiological explanationnbsp;that can be given of the primary functions of the exogenous zone in these giantnbsp;Cryptogams. In the living Lycopodiace^, with the exception of Isoetes, nothingnbsp;like this arrangement exists. The rootlet bundles are all derived directly from thenbsp;same central source as the leaf-bundles, the two sets of bundles differing only innbsp;the upward direction followed by the former and the downward one pursued by thenbsp;latter; hence no lateral transference from one vascular system to another is necessary. Another structure introduced into the extinct plants is the bark cambiumnbsp;(Plate yill, figs. 22 and 23), which has no definite existence in any recent formnbsp;except Isoetes. In this plant the cortex is more dependent upon a cambial layer andnbsp;is developed by that cambium in a more conspicuous manner than in any othernbsp;known plant; but in this instance the same cambium adds centrifugally to the loosenbsp;axial mass of Tracheids on its inner side, and, centripetally, to the cortex on thenbsp;outer one. We thus see that the ancient Lycopods had two cambium zones, thenbsp;functions of the outermost being limited to bark-growth, and the inner one apparently to extension of the exogenous layers, the latter alone being concerned in thenbsp;orientation of the rootlet bundles. Isoetes, on the other hand, has but one cambiumnbsp;zone which is equally concerned in the orientation of the leaf-bundles, the rootletnbsp;bundles, and the development of the bark. Degradation from a higher to a lowernbsp;type of organisation has been followed by a generalisation of function, instead ofnbsp;the opposite process of multiplication of organs and specialisation of functionsnbsp;which attends progressive evolution. The supremacy of the Carboniferous Lyco-podiace* over their modern representatives therefore is not limited to their greaternbsp;magnitude, but includes a more complex organisation.

There appears no doubt whatever that the Stigmaria is found in the Devonian, Carboniferous, and Permian Eocks. In the two former it is in some cases associated with both Lepidodendron and Sigillaria. But in the Arran Deposit atnbsp;Laggan Bay, where it was not rare, though we have Lepidodendroid branches innbsp;profusion, we find no trace of Sigillaria. This association is yet more remarkablenbsp;at the plant-bearing deposit at Burntisland. The rock at that locality is largelynbsp;composed of Lepidodendroid twigs, but has not yet furnished, so far as I know, thenbsp;smallest trace of a Sigillaria. The late Professor Heer tells us that the Spitzbergennbsp;deposits supplied Lepidodendron Veltheimianum, along with fine and large Stigmariae,nbsp;but no Sigillaria. The same author also obtained Stigmaria along with severalnbsp;species of Lepidodendron at Bear Island, but again no Sigillaria; and M. Lesquereuxnbsp;cites Schimper�s authority for the fact that a deposit in the Yosges is � filled withnbsp;a prodigious quantity of fragments of Stigmaria without trace of any Sigillaria,�nbsp;but adds significantly, � that these strata contain abundant remains or trunks of

Knorria and Lepidodendron.� It is needless to add that all these cases only point to the fact that the Lepidodendra had Stigmarian roots as well as the Sigillarise.

This abundance of Stigmari* apart from Sigillariae has been referred to by more than one writer as a proof that Stigmaria was sometimes an independentnbsp;plant sui generis. The reply is obvious. We have abundant proof that, wherevernbsp;we obtain a Stigmaria connected with any aerial appendage, that Stigmaria is alwaysnbsp;in the position of a root; and seeing that all the other fragments referred to havenbsp;precisely the same characteristics of form and structure as those roots, wenbsp;have the strongest a ^priori ground for believing that they too were roots; anyhow, until the opposite view is demonstrated to be probable by more conclusivenbsp;evidence than has yet been discovered, deductive reasoning from what we do knownbsp;to be facts compels us to infer that true Stigmarise were always roots.

APPENDIX.

Since the preceding pages were put into type, probably the finest example of a fully developed Stigmaria ficoides yet seen has been discovered and has come intonbsp;my possession. Mr. Murgatroyd, the intelligent proprietor of a large quarry ofnbsp;Carboniferous Sandstone at Clayton near Bradford, in Yorkshire, removing somenbsp;stone by means of dynamite, found amongst the exploded materials some fragments of a Stigmaria. Seeing that the root extended into the undisturbed rock,nbsp;with a thoughtfulness highly creditable to him, he ceased to employ explosivenbsp;materials, and had the overlying stratum removed with great care. The resultnbsp;was the revealing of the magnificent specimen represented in Plate XV. Thisnbsp;representation is a copy of a beautiful photograph taken by Mr. E,. C. Clifford, anbsp;skilful young professional photographer, residing at Westgate, Bradford, to whomnbsp;I am indebted for permission to use the photograph in illustration of this memoir.nbsp;The photograph has been reproduced by the Automatic Engraving Company,nbsp;of Willesden Grreen, near London. The tree stands upon a fiat stratified surface,nbsp;composed of an arenaceous shaly bed, which is abundantly permeated by thenbsp;remains of its disorganised rootlets, and upon which its magnificent roots arenbsp;spread out with undisturbed regularity. The overlying stratum is a hard sandstone,nbsp;identical with the inorganic material of which the roots themselves consist. It isnbsp;obvious that the entire base of the tree became encased in a plastic material, whichnbsp;was firmly moulded upon these roots whilst the latter retained their organisationnbsp;sufficiently unaltered to enable them to resist all superincumbent pressure. Thisnbsp;external mould then hardened firmly, and as the organic materials decayed theynbsp;were floated out by water which entered the branching cavity; at a still laternbsp;period the same water was instrumental in replacing the carbonaceous elements bynbsp;the sand of which the entire structure now consists. It is obvious that we havenbsp;not got the ultimate divisions of the roots in their entire length. Their extremities have failed to be preserved, from a reason given at p. 29. Still the roots, asnbsp;seen in the plate, extend 29 feet 6 inches from right to left, and 28 feet in thenbsp;opposite direction.

STIGMARIA FICOIDES.

The above xylograph represents accurately the arrangement of these roots as seen from above; it is reduced from a large plan carefully prepared by Chas.nbsp;Brownridge, F.G.S., for Mr. Adamson, F.G.S., the Secretary of the Geologicalnbsp;Society of Leeds, who has kindly allowed the above reduced figure to benbsp;copied from it by photography. As the preceding pages have shown to be almostnbsp;invariably the case, four large primary roots, A, B, 0, and D, radiate from a large

centra] mass wHcli projects about 4 feet above the plane upon wbich the roots are outspread, and the top of which has a transverse mean diameter of 4 feet 4 inches,nbsp;equal to a circumference of 13 feet. The specimen confirms the statement madenbsp;in a previous part of this memoir (see page 16) that these Stigmarian rootsnbsp;normally dichotomised but twice. The following xylograph is a diagrammaticnbsp;representation in which I have placed the four roots, A, B, C, and D, with theirnbsp;respective branches, side by side, for the purpose of showing accurately theirnbsp;relative lengths and the varying positions of their several dichotomies. Theirnbsp;several diameters are not represented.

The following table gives the length of each division of the several roots between these dichotomies.�-

Length of each of the four undivided roots from the edge of the transverse section of the stem to the first dichotomisation.

B. nbsp;nbsp;nbsp;1 � 4nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;D. 1 � 5nbsp;nbsp;nbsp;nbsp;�

1 The Roman numerals attached to each of the XVI terminal branches of these roots are also attached to the same branches on Plate XV.

We do not know the full length of Y, for it disappeared in a hard face of the quarry, which, being under a roadway, could not easily be followed.

Interpreting this specimen as I have explained Fig. 2 (Plate I), aided by the light thrown upon both by the Figures 5 and 6 (Plate III), I conclude we have here no partnbsp;of the true aerial stem, which seems to have died down to its own base and disappeared. The central mass consists solely of the coalesced proximal ends of the fournbsp;primary roots. In this specimen the absence of every trace, either of rhizomatousnbsp;features, or of additional aerial stems ascending from these roots is also conspicuous.

In each of the four roots the rootlet-scars are distinctly seen covering the eight secondary branches a, h, amp;c. The ultimate divisions show them in the normalnbsp;form so characteristic of Stigmaria. How little the ramifications of this Stigmarianbsp;have in common with the diagram published by M. Renault^ need not be dweltnbsp;upon.

Still more recently a second and somewhat larger example has been discovered in the adjoining Fall-Top Quarry, belonging to Messrs. Briggs and Shepherd, not much above a hundred yards from, and resting upon the samenbsp;shaly bed, as that just described. The two specimens correspond in everynbsp;essential detail, only that several of the large roots in the Fall-Top fossil have notnbsp;yet been relieved of the thick mass of sandstone by which they are overlain. Thenbsp;following measurements are taken in the same way as those on the preceding page :

^ Since the xylograph 8 was prepared I have succeeded in laying this branch (xvi) bare, and tracing its length to a distance of seven feet, when it became completely flattened, so that its upper and lowernbsp;surfaces were almost in contact. The additional length thus revealed is correctly represented innbsp;xylograph 7.

APPENDIX.

B. nbsp;nbsp;nbsp;Nil. TEe cleft separating tEis root into its two primary divisions reacEing

WEilstj as tEis table sEows, tEe prolongations of ten of tEe ultimate root-brancEes were Eidden by tEe rock into wEicE tEey plunge. No. Ill sEows wEat may be expected of them whenever they are uncovered. TEis root ran in anbsp;direction more favorable for exploration. I followed it therefore until 22 feet werenbsp;uncovered�which, with the addition of the other two segments, made a totalnbsp;length of 31 feet 3 incEes from the exterior of the base of the central stem, andnbsp;even this did not give us its entire length. Where we ceased to follow it thenbsp;root was so completely flattened, that whilst its breadth was 4 incEes its maximumnbsp;thickness was only fths of an inch. Hence, in its uncompressed state, this rootnbsp;can scarcely have exceeded 2^ incEes in diameter. How much farther it extendednbsp;in length we have not yet ascertained.^

* This root has since been traced to its termination, its entire length being 37 feet 4 inches. It continued to be completely flattened to its apex, which also narrowed to a point. This specimennbsp;throws light upon what has taken place in at least many of these root-terminations. After thenbsp;vegetable elements had floated out, as described at p. 25, some obstruction prevented the inorganic sand, by which the cavity left by the disappearance of the organic material was fllled, fromnbsp;reaching the extremities of these narrowing tunnels. Thus deprived of all internal support, insteadnbsp;of retaining their cylindrical form these cavities were crushed down by the weight of the superimposednbsp;mud and sand, roof and floor being thus brought into close contact. In many cases all traces of thesenbsp;terminations have ultimately disappeared.

STIGMARIA PICOIDES.

Careful measurements of the circumference of each of the four primary roots of this specimen gave the following results :

A. nbsp;nbsp;nbsp;7nbsp;nbsp;nbsp;nbsp;feet 6nbsp;nbsp;nbsp;nbsp;inchesnbsp;nbsp;nbsp;nbsp;C.nbsp;nbsp;nbsp;nbsp;8 feet 8 inches

B. nbsp;nbsp;nbsp;8nbsp;nbsp;nbsp;nbsp;� 6nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;D.nbsp;nbsp;nbsp;nbsp;7 � 6nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�

We thus see that we have traced the gradual diminution of these roots from a mean diameter of about 32 inches to one of 2^ inches, which, as we have alreadynbsp;seen, was not the limit of the reduction.^ Such a diminution, combined with thenbsp;fact that no trace of aerial shoots presented themselves in either of these superbnbsp;specimens, is absolutely conclusive against the hypothesis that these roots arenbsp;rhizomes.

In both these examples a thin film of carbonaceous matter invested them throughout allnbsp;nbsp;nbsp;nbsp;theirnbsp;nbsp;nbsp;nbsp;ramifications.nbsp;nbsp;nbsp;nbsp;They hadnbsp;nbsp;nbsp;nbsp;evidently been hugenbsp;nbsp;nbsp;nbsp;andnbsp;nbsp;nbsp;nbsp;isolated

trees nbsp;nbsp;nbsp;growingnbsp;nbsp;nbsp;nbsp;uponnbsp;nbsp;nbsp;nbsp;the samenbsp;nbsp;nbsp;nbsp;horizontalnbsp;nbsp;nbsp;nbsp;plain,nbsp;nbsp;nbsp;nbsp;and not portionsnbsp;nbsp;nbsp;nbsp;ofnbsp;nbsp;nbsp;nbsp;anbsp;nbsp;nbsp;nbsp;dense

forest. To this circumstance, probably, was due the fact that they were not planted upon a bed of coal; any little vegetable soil that accumulated under theirnbsp;localised shade would attain to no thickness, and would readily be removed bynbsp;denudation.

I am indebted to my friend J. W. Davis, Esq., E.G.S., of Halifax, for the following sections, showing the geological horizon to which these trees belong.nbsp;At p. iv of my � Introduction � I have given, also on the authority of Mr. Davis,nbsp;a section of the beds between the Eiland Flagstones and the � Rough Rock �nbsp;or Millstone Grit. The present section shows the upward continuation of the samenbsp;series of beds, with a few observations in reference to them.

80	0	Shales (with 80 yards Band Coal)	. 180	0
70	0	Hard-bed Band Coal	. 1	2
7	6	Shales (with 36 yards Band Coal)	. 137	0
36	0	Hard-bed Coal (Ganister) nbsp;nbsp;nbsp;'nbsp;nbsp;nbsp;nbsp;.	. 2	2
2	6	Shale, amp;c.	. 29	6
130	0	Middle Band or Clay Coal	. 0	6
1	6	Stone and Shale .	. 62	0
54	0	Soft-bed Coal	. 1	6
30	0	Shales . nbsp;nbsp;nbsp;...	. 102	0
35	0	Coal and Seat Earth	. nbsp;nbsp;nbsp;5	6
180	0	Bough Bock

Clifton or Oakenshaw Eock Shales with Stone and Coalnbsp;Crow Coal with partingsnbsp;Shales with Ironstone (Low Moor)

� The Eiland Flag rock is of great thickness, and forms a bold range of hills from Sheffield and Wadsley, northwards to Penistone, west of Huddersfield, south andnbsp;east of Halifax to north of Leeds. In Lancashire it is the Rochdale and Uphollandnbsp;Flags ; probably when deposited it covered an area of 1500 square miles, and it isnbsp;the thickest and most persistent of the Coal Measures sandstones.

1 The note on the previous page shows that the root terminated in an absolute point.

APPENDIX.

� In the neighbourhood of Halifax, at South Owram, it is 150 feet thick ; thence it tops the hills at North Owram, and at Queensbury and Clayton is still thicker.nbsp;It is divided by a bed of shale at these places ; the best sandstone and flags are atnbsp;the bottom, 160 to 180 feet thick, then 35 feet of shale, and above the shale aboutnbsp;30 feet of flags, sometimes poor and raggy.

� The trees at Clayton are from the shale and rag above the lower flagstone, which in this quarry is worked to a depth of 80 to 100 feet.�

I may only add in conclusion that the magnificent fossil described in the above pages is now at the Owens College at Manchester, in the Museum of which institution it will shortly be mounted in a manner worthy of its perfection.^

^ In addition to the above specimens no less than seven similar examples, though of smaller size, have been discovered in excavating for the foundations of some buildings in Parley Street, Bradford.nbsp;Though of smaller size, these new specimens, so far as they have been uncovered, lead to the samenbsp;conclusions as those enunciated in the preceding pages.

GLOSSAEY OP BOTANICAL TEEMS EMPLOYED IN THE

MONOGEAPH.i

Aoeopetal.�Where a growing shoot develops other lateral structures or organs in succession, behind the growing tip of the shoot.

Baeeed Tissues.�Consist of tracheids or vessels, the walls of which are thickened internally by transverse bars of woody material (lignine) which alternate withnbsp;parallel thin spaces; modified scalariform vessels.

Cambieoem Tissue.�Cellular tissue of the phlo�m produced from the cambium, and which has assumed a permanent condition; phloem tissues the elementsnbsp;of which are similar in character to that of the cambium in their elongatednbsp;form and thinness of walls. See De Bary, � Phanerogams and Ferns,� Eng.nbsp;Trans., p. 327.

CxiMBiuM.�A thin zone of meristematic cells intermediate between the exterior of the exogenous vascular zone and the innermost surface of the bark, andnbsp;capable of developing new structures from either or both of its surfaces. Itnbsp;also appears between the xylem and the phloem elements of isolated vascularnbsp;bundles, e. g. foliar or rootlet bundles.

Cexteieugal oe Exogenous Yasculae Cylindee.�A zone of vessels or tracheids surrounding the centripetal cylinder where the latter is present, and increasing in thickness by additions made to its outer border through the action ofnbsp;a circle of meristem cells known as cambium; its growth in thickness therefore proceeds from within outwards.

Centeipetal Yasculae Cylindee.�A zone common to many Carboniferous Cryptogams, which, commencing as a medullary sheath, increases in size and thickness by a conversion of the medullary cells which it surrounds intonbsp;vessels or tracheids. Its component elements are never arranged in radiatingnbsp;lines or lamina, and it is the source whence the fibro-vascular bundles goingnbsp;to the leaves are derived. Its growth, therefore, is centripetal.

1 Introduced at the recommendation of the Editor for the use of geologists who may not be familiar

STIGMARIA FICOIDES.

Dichotomy.�An unvarying division of single organs into two more or less equal structures. A growing bud has an apical cell, or cluster of cells, which dividenbsp;equally, giving rise to two dichotomous branches.

Distal.�The end of an organ most remote from the organism to which it is attached.

Emergences.�Hair-like structures, which instead of being developed, like hairs, entirely from a single cell of the epidermis, also originate partly from somenbsp;of the structures that underlie the epidermis.

Fibres.�Prosenchymatous cells, whose walls are thickened by deposits of woody substance on their inner surfaces.

Fibro-vascular Bdndles.�Strands compounded, when perfect, of vessels, tracheids, fibres and cells constituting the xylem part, and of vascular (sieve-tubes),nbsp;cellular and fibrous elements, the phloem or hast portion of each bundle.

Foliar Bundles.�Fibro-vascular bundles primarily prolongations from the medullary sheath, or, in the case of the Carboniferous Cryptogams, from the centripetal vascular cylinder, and passing outwards through the cortex to the leaves.

Medulla or Pith.�The central cellular axis of a stem or branch enclosed within a circle of vascular bundles.

Medullary Cavity or Canal.�A hollow space in the centre of the cellular medulla, formed either by the absorption of its central cells or by their rearrangementnbsp;round the interior of the vascular cylinder, owing to the growth of the latternbsp;proceeding more rapidly than that of the medullary parenchyma does.

Medullary Rays.�Radial vertical plates of cells intervening between the lamina of the exogenous vascular cylinder, and connecting the medulla with the bark.

Primary Medullary Rays.�Those first seen between the vascular bundles when a vascular zone or cylinder commences its formation, and in which the connections between the medulla and the cortex are direct and undisturbed.

Secondary Medullary Rays.�The vessels of the first formed vascular bundles which intervene between the primary medullary rays are usually few innbsp;number. As the cylinder which they originate grows in thickness, thesenbsp;bundles enlarge laterally as well as radially. The additions assume the formnbsp;of lamime, which are elongated vertically, and arranged in radiating groups,nbsp;which latter constitute the vascular wedges of transverse sections; the laminaenbsp;of each of these wedges are in close contact at their inner extremities, butnbsp;spread out at their outer ends in a fan-shaped manner. At the same time newnbsp;vertical cellular lamina are intercalated, though but few of these reach thenbsp;medulla at their inner ends. These latter cellular laminae are the secondarynbsp;medullary rays.

Medullary Sheath.��tui meduUaire of Brongniart. The first formed ring of vascular bundles, separating the medulla internally from the outer zone of

GLOSSARY.

cortex. This sheath is chiefly concerned in supplying the leaves with their first vascular bundles, and is the centripetal vascular cylinder of manynbsp;Carboniferous stems.

Meeistbm.�Cells capable of multiplication by the formation of septa or walls crossing their inner cavities.

Monaeoh, Diaeoh, amp;c.�In a transverse section of a young root, the first-formed vessels appear at a variable number of points near the circumference of anbsp;central cellular cylinder. The vascular bundles of roots are called Monarchnbsp;when there is but one such initial point. Diarch if two. Triarch if three, amp;c.

Monopodial oe false Dichotomy.�Where the apex of a growing shoot advances whilst any lateral branches, however near to the apex, arise from a placenbsp;lower down than the true apex.

Oeientation.�The uprising of branches from some other organ, as of foliar bundles from the centripetal cylinder of a Lepidodendron or of rootlet bundlesnbsp;from the exogenous zone of Stigmaria.^

Peeideem.�An external protective layer of the bark replacing the detached epiderm of young plants. It is usually a Phellem, i. e. cork.

Phellem.�A peripheral cork-layer of the bark developed centripetally from the exterior of the Phellogen.

Phellodbem.�A cellular layer of the bark developed centrifugally from the inner side of the Phellogen.

Phloem.�The bast portion of a fibro-vascular bundle characterised by the presence in it of sieve-tubes or their representatives, as the xylem or woodnbsp;portion is characterised by the presence of vessels or Tracheids.

Phyllomb.�A modified part of a plant primarily capable of developing into a leaf, e. g. each of the parts of a flower.

Proximal.�The end of an organ nearest to the centre of the organism to which it is attached.

Quincuncial.�A term applied to the arrangement of the rootlet-scars of Stig-maria, each one of which occupies the centre of a surrounding group of four similar ones.

Rhizome.�A creeping subterranean stem, capable of giving oflF leaves or aerial shoots from its upper surface and from the ends of its branches, and rootsnbsp;from its lower surface.

^ This definition of � Orientation � represents the meaning of the word as used in the Monograph,

STIGMARIA FICOIDES.

Rootlets.�Secondary organs of Stigmaria performing root-functions, but having a different structure from the roots.

Rootlet bundle.�A fibro-vascular bundle derived partly from the exogenous or centripetal vascular cylinder of the root, and partly from the cambiform ornbsp;phlo�m zone which immediately invests the exogenous cylinder.

Rootlet-bundle Cyltndee.�A cylinder consisting of several layers of cells surrounding both the xylem and phlo�m elements of the bundle in the rootletsnbsp;of Stigmaria.

Rootlet-coetex.�The outer wall of each cylindrical rootlet of Stigmaria, an extension from the most external or parenchymatous layer of the root-bark.

Rootlet-cushion.�A solid cylinder of cellular tissue embedded in the meristemic zone of the root-bark, and transmitting the rootlet bundle through its centre.

Rootlet-scaes.�Hollow depressions left on the outer surface of a Stigmarian root by the disappearance of the rootlets. The latter have been mechanicallynbsp;broken off, not being deciduous like a falling leaf.

ScLEEOUS TISSUE OE SoLBEENCHYMA.�Cells or tubes of any kind of which the walls are uniformly thickened by woody deposits on their inner surfaces.

Teacheids.�Single elongated prosenchymatous cells, the ends of which have not been absorbed, which, when living, contained no protoplasm, and whose wallsnbsp;also are usually furnished with bordered pits.

Vegetative Repetition.�Multiplications of any organ, each of which multiplied examples has typically the same structure and functions as those which itnbsp;resembles.

Vessels.�Prolonged tubes, usually formed by the coalescence of linear series of cells, the divisions (septa) between the ends in mutual contact having beennbsp;partially or entirely absorbed.

DESCEIPTIYE INDEX TO THE PLATES.'

3. nbsp;nbsp;nbsp;II. Under surface of a Stigmarian root in the Museum of the Leeds Philosophical

4. nbsp;nbsp;nbsp;II. Under surface of a Stigmarian root in the Museum of the College of Science,

7. nbsp;nbsp;nbsp;IV. Medullary angles of three vascular wedges from a transverse section of the vascular

cylinder of a Stigmaria. a. Medullary cells. 5. Vascular wedges. 5'. Elongated cells of primary medullary rays. 5quot;. Secondary medullary rays. x 90.nbsp;Cabinet, No. 744 .nbsp;nbsp;nbsp;nbsp;,nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;10, 14

V. Tangential section from a vascular cylinder, showing transverse sections of the primary medullary rays, 5'. Vascular rootlet bundles, f. x 5. Cabinet,

No. 771 . nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;13, 14

VI. Longitudinal section through a young root of a Stigmaria. Medullary cavity, a.

Vascular cylinder, 5, 5. Vessels of the cylinder deflected to form a rootlet bundle, �, passing out through a primary medullary ray, V. Secondarynbsp;medullary rays, 5quot;'. Cortex parenchyma, A. Cortex meristem. A'. Base ofnbsp;rootlet, y. Eootlet cushion, 5. x 10. Cabinet, No. 776nbsp;nbsp;nbsp;nbsp;14,18,19,22,32

10. nbsp;nbsp;nbsp;VII. A radial section through the outermost vessels, 5, of the vascular cylinder. Cambi-

X 90. Cabinet, No. 779 nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.14,17, 18

11. nbsp;nbsp;nbsp;VII. A tangential section through a portion of a vascular cylinder. Xylem lamin�, 5.

12. nbsp;nbsp;nbsp;IX. Tangential section of part of a xylem cylinder, made close to the medulla, giving a

transverse section of a primary medullary ray, 5, and the vascular rootlet bundle, �, �'. Enlarged secondary medullary rays at the orientation of thenbsp;younger part of the rootlet bundle, A. X 30. Cabinet, No. 851nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;14, 23

^ The list of pages appended to each figure shows either where each specimen figured is described, or where some important reference is made to it.

Transverse section of the primary medullary ray, Tig. 12, but made close to the cortical surface of the vascular cylinder, amp;. Vascular rootlet bundle, �. Firstnbsp;formed portion of the rootlet bundle, y�. Secondary medullary rays, �f. X 30.nbsp;Cabinet, No. 856 .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;14,23

Transverse section of a medium sized vascular cylinder. Medullary cavity, a, containing an intruded Stigmarian rootlet, y. Vascular wedges, i. Primary medullary rays, b'. Obliquely longitudinal sections of rootlet bundles, normallynbsp;co-extensive with the entire thickness of the cylinder,/. Section of a rootletnbsp;bundle escaping from the cylinder,/, x 5nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;14, 17, 23, 24, 25

Transverse section of a vascular cylinder with its more external cortical layer and the bases of three rootlets. Medullary cavity, a. Wedges of the vascularnbsp;cylinder, h. Outermost cortical parenchyma, d. Meristem zone of bark, d'.nbsp;Cortical prosenchyma, e. Pootlet bundle,/. Eootlet bundle cylinder,/quot;. Cortical zone of rootlet, Eootlet cushion, h. X 6^. Cabinet, No. 759nbsp;nbsp;nbsp;nbsp;14, 18, 26

Transverse section of a young primary medullary ray. Vessels of the xylem cylinder,

V. Cells of the medullary ray, J. Primary vessels of the rootlet bundle,/'. Later formed vessels,/. Secondary medullary rays, lt;7. x 14. Cabinet, No. 824 14,23,31nbsp;Transverse section of the vascular cylinder of a very young Stigmaria. Cabinet,

No. 773. X 5. nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;,nbsp;nbsp;nbsp;nbsp;...

Part of a transverse section of a vascular cylinder, including a longitudinal section of a secondary medullary ray containing barred Tracheids. Two radial lines ofnbsp;vessels of the cylinder, J, amp;. Barred Tracheids, 5quot;. x 80. Cabinet, No. 767nbsp;Transverse section of an extremely young root of Stigmaria. Vascular wedges ofnbsp;the disturbed vascular cylinder, h. Superficial parenchyma of the cortex, d.nbsp;Cortical prosenchyma, d'. Eootlets belonging to the root, g, g. x 11.nbsp;Cabinet, No. 775 ......nbsp;nbsp;nbsp;nbsp;15

Peripheral portion of the transverse section of the vascular cylinder. Fig. 14, PI. VII, in which a new zone is being added exogenously. Older vessels of the cylinder,nbsp;h. Imperfectly developed vessels, amp;c., of the newer zone, V. x 39nbsp;Periphery of a transverse section of a vascular cylinder with an investing cellularnbsp;layer, which appears to be a cambium. Peripheral vessels of the vascularnbsp;cylinder, h. Procambial cells, c. X 75. Cabinet, No. 744nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;14, 17

Transverse section of the bark of Fig. 18, PL IX. Outer parenchyma, d. Pro-senchymatous zone, d'. x 18. Cabinet, No. 775 nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;18, 19

Outer cortex of a transverse section from the same specimen as Fig. 15, PI. VIII. Outermost parenchyma, d. Meristem or bark cambium zone, d'. Eadial walls,nbsp;apparently of the mother-cells of the procambial cells, d'quot;. Prosenchymatousnbsp;zone, e. x 18. Cabinet, No. 757nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;19, 43

Transverse sections of the cortex of a much older Stigmarian root than Fig. 22. Outermost cortical parenchyma, d. Meristem zone, d'. Cell walls of mothernbsp;cells ? dquot;. Prosenchymatous zone, e. X 18 .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;19, 21, 43

Tangential section through the meristem tissue, d', of Fig. 23. X 9 nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.20

Tangential section of the tissue, e, of Fig. 23. X 9 nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;20, 21

Transverse section of a fragment of Stigmaria with a considerable development of the prosenchymatous zone, e, of Figs. 22 and 23. X 3. Cabinet, No. 795nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;20

A portion of Fig. 25 further enlarged. External pressure has disturbed the

transversely intersected prosenchymatous cells along the lines d, e'. x 18 nbsp;nbsp;nbsp;.20

A dichotomous rootlet of Stigmaria nbsp;nbsp;nbsp;....nbsp;nbsp;nbsp;nbsp;32, 33

I Transverse sections of six rootlet bundles after they have escaped from the peripheral ends of the primary medullary rays, but before they have penetrated the cortex. In each of these, �' indicates the protoxylem of the wedge-shapednbsp;bundle and the broad opposite base of the wedge, including the small vessels,nbsp;f, its latest additions, x 80. Cabinet, No. 744nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;23, 24

Another section like Kgs. 28�33, but with all its newest small vessels derived from a half developed exogenous zone like Kg. 19, b', PI. IV. X 80nbsp;nbsp;nbsp;nbsp;.23

Two wedge-shaped rootlet bundles from the same position as the preceding ones, but from the exterior of the large vascular cylinder from Staffordshire. Protoxylem vessels,/'. X 80. Cabinet, No. 760nbsp;nbsp;nbsp;nbsp;...nbsp;nbsp;nbsp;nbsp;23, 24

Part of a fragment of a Stigmarian root split vertically. Exterior of the vascular cylinder, V. Eootlet bundles crossing the cortical region,/. Exterior surfacenbsp;of the cortex, d, d. Half the natural sizenbsp;nbsp;nbsp;nbsp;....

One extremity of Fig. 39. Vascular wedges, b, of the cylinder separated by primary medullary rays, b'. The central medullary cavity is occupied by an intrudednbsp;Stigmarian rootlet, in the interior of which the rootlet bundle is preserved.

Nat. size. Cabinet, No. 867 nbsp;nbsp;nbsp;.....

Side view of Eig. 38, with numerous rootlet bundles,/, enclosed within their several bundle sheaths, issuing from the primary medullary rajs of the cylinder. Nat.nbsp;size. Cabinet, No. 867nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;24, 26

Fragment of a large nodule enclosing a disorganised Stigmarian root, almost identical with one of which the specimen. Pigs. 38, 39, is the vascular cylinder, d represents a cast of the outer surface of one side of the root, with casts of the basalnbsp;ends of the rootlets penetrating the stone. Eootlets from the opposite side ofnbsp;the root are seen at y, y, with their free ends pointing to the right. / a mass ofnbsp;rootlet bundles resembling these given off from the exterior of Fig. 39nbsp;Inner surface of a fragment of the outermost bark, with rootlet bundles penetrating it to reach the rootlet-scars existing on the reverse side of the specimennbsp;Transverse section of a large Stigmarian rootlet, y, into the interior of whichnbsp;several other younger rootlets have penetrated. Vascular bundle of the largenbsp;rootlet,/. Cortex of one of the smaller rootlets, yquot;. Its inner vacant cavity,nbsp;y'. Its vascular bundle and sheath,/', x 11. Cabinet, No. 848 .nbsp;nbsp;nbsp;nbsp;25, 33

Longitudinal section of the proximal end of a rootlet. Parenchymatous cortex of the root, d, d. Prosenchymatous cortex of the root, e. Vascular rootlet bundle,

/ � Eootlet bundle-sheath, f. Branching cells investing the extremity of the rootlet cushion, i, i. Outer cortex of the rootlet, y, q. Fistular interiornbsp;of the rootlet, y'. X 18nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;26,27,28

Longitudinal section of the proximal end of another rootlet. Prosenchymatous zone of the root-bark, d!. Central portion of the rootlet cushion, h, h. Conicalnbsp;extremity of the rootlet cushion inside the rootlet, h'. Eootlet bundle, �nbsp;Eootlet bundle-sheath, /'. Outer cortex of the rootlet, y. Fistular interiornbsp;of the rootlet, y'. X 18. Cabinet, No. 819nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.26, 27, 28

Longitudinal section of the proximal end of another rootlet. Parenchymatous zone of the root-bark, d, d. Eootlet bundle, /, enclosed by its rootlet bundle-sheath, /quot;, and passing outwards through the rootlet cushion, h, h. Apex of thenbsp;rootlet cushion, h!. Outer cortical zone of the rootlet, y. X 18nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;19, 26

DESCRIPTIVE INDEX TO THE PLATES.

Transverse section of a rootlet cusHon below tbe base of a rootlet. Vascular rootlet bundle,/. Small cells of the rootlet cushion, A. Zone of large cellsnbsp;investing the rootlet cushion, A'. X 20nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.26

Similar section to 46, only partly invested by some of the tissues of the root-cortex. Small cells of the rootlet cushion, h. Investing zone of larger cells, hi. Pro-sencbymatous cells of the root-bark seen in tangential section, e. X 20.nbsp;Cabinet, No. 807nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.26

Central portion of Fig. 47, further enlarged. Rootlet bundle,/. Phloem portion of the bundle,/'. Cellular investment of the bundle,/quot;', x 200 .nbsp;nbsp;nbsp;nbsp;. 26

Some of the prosenchymatous cells of Fig. 47, e, undergoing fission by horizontal septa. X 40nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.20

Minute branching tubular cells investing the surface of the conical extremity, h', of the rootlet cushion, seen in the longitudinal rootlet sections, e.g. PI. X, fig. 43 i 27nbsp;Transverse section of a rootlet made about half way up the conical portion, hi, ofnbsp;the rootlet cushion. The outermost parenchyma of the root-bark, d. Thenbsp;cortical layer of the rootlet, g. The second cellular layer of the rootlet verynbsp;rarely preserved, g'. Section of the small cells of the rootlet cushion, h. Rootlet bundle and its sheath,/ x 50. Cabinet, No. 821nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.27

Vascular bundle,/ and part of the bundle cylinder,/quot;, of a larger and older rootlet.

Rootlet bundle,/ with part of its monarch initial vessels at/', but with two others

A second rootlet bundle, like Fig. 60, with its initial vessels at /', and a small portion of its cylinder at /quot;. Here again the vessels of the older part of thenbsp;bundle are non-radial; the younger additions are disposed radially. X 100,nbsp;Cabinet, No. 651 .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.30

Transverse section of a very young bundle, /, with its phloem elements preserved

at/'quot;. Bundle cylinder,/quot; . nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;.32

Transverse section of a rootlet about to dichotomise. Inner border of the cortical zone of the rootlet, g. Fistular cavity of the rootlet, g'. Two bundles,// eachnbsp;surrounded by its bundle sheath or cylinder,/',/', resulting from the divisionnbsp;of one single bundle, x 75.nbsp;nbsp;nbsp;nbsp;Cabinet,nbsp;nbsp;nbsp;nbsp;No.nbsp;nbsp;nbsp;nbsp;862 .nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.32

A natural cast of the medullary cavity of a Stigmarian root, the superficial, longitudinal ridges of which represent the prolongations of the medulla into the

inner extremities of the primary medullary rays. Nat. size. Hutton Collection. Newcastle Museumnbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.,

Natural cast of half of the exterior of the vascular cylinder, h, of the same specimen as Fig. 64, showing the peripheral terminations of the primary medullary rays ;nbsp;mineral matter replacing the entire bark, d. Hutton Collectionnbsp;A second natural cast, like Fig. 65, showing much more strikingly the peripheralnbsp;termination of the primary medullary rays. Nat. size. Hutton Collectionnbsp;Transverse section of a Stigmarian root, nat. size, from which all organic elementsnbsp;have been removed. The inorganic matrix occupies the medullary cavity, a, thenbsp;primary medullary rays, V, and replaces the entire bark, d. Medullary rays, b'.nbsp;Areas left vacant by the disappearance of the vascular wedges of the vascularnbsp;cylinder, b. Hutton Collectionnbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;�

68. nbsp;nbsp;nbsp;XIV. Fragment of a Stigmarian root split longitudinally.nbsp;nbsp;nbsp;nbsp;Exterior of medulla, a. Long

section of the vascular cylinder, b, b. Primary medullary rays passing horizontally through the vascular cylinder, 6'. Cortex replaced by inorganic matrix, d, d. Author�s Cabinet ....��

69. nbsp;nbsp;nbsp;XIV. Exterior of the vascular cylinder of a Stigmaria, from Burntisland. Nat. size.,

showing the large peripheral orifices of the primary medullary rays, intermingled with a few rootlet-scars. Author�s Cabinetnbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�

70. nbsp;nbsp;nbsp;XII. Fragment of the vascular cylinder of a Stigmaria, upon both surfaces of which the

bases of rootlet-scars have been impressed on the disappearance of all the other cortical tissues. Nat. size. Author�s Cabinetnbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;�

71. nbsp;nbsp;nbsp;XIII. Cast or impression of the surface of part of a dichotomising Stigmarian root,

exhibiting the casts of what were longitudinal fissures in the exterior of the bark, along with more minute undulating ridges and furrows. Half nat. size.nbsp;Hutton Collection.......

72. nbsp;nbsp;nbsp;XII. Cast of the outer surface of a fragment of Stigmarian bark with undulating corru

gations. Nat. size. Museum of the Owens College

73. nbsp;nbsp;nbsp;XIV. A normal portion of a Stigmarian root deprived of its rootlets. Cavity left by the

disappearance of the vascular cylinder, b. Nat. size. Author�s Collection

74. nbsp;nbsp;nbsp;XII. Fragment of Stigmarian bark in which the bases of the rootlets are preserved as

funnel-shaped depressions, within which the outer cortex of the rootlet forms a prominent ridge at g, and the conical apex of the rootlet cushion is seen at thenbsp;bottom of the funnel, h. Nat. size. Author�s collection .

75. nbsp;nbsp;nbsp;XIV. Cast of the outer surface of a Stigmarian bark which has been in the condition of

� the specimen, fig. 74. Nat. size. Author�s Collection

76. nbsp;nbsp;nbsp;XII. Diagrammatic section of a Stigmarian rootlet attached to its bark, d', dquot;, illustrating

Figs. 74 and 75. Inner surface of the parenchymatous layer, d'. Outer surface of the same, dquot;. Small remnant of the outer cortex of the rootlet,nbsp;gquot;, g,quot; being the point near which the rootlet, g, had been broken ofi' from its basalnbsp;end. Cylindrical interior of the rootlet, g', corresponding to the funnel-shaped depressions, h, of Fig. 74, and to the central prominences of Fig. 75.nbsp;Apex of the cone of the rootlet cushion, h, forming the apical depression ofnbsp;each central prominence of Fig. 75.nbsp;nbsp;nbsp;nbsp;.

77. nbsp;nbsp;nbsp;XIV, External surface of a Stigmarian bark modified by shrinkage or compression.

Nat. size. Author�s Collection nbsp;nbsp;nbsp;.....

Fig. 1.�Cast of one of the Stigmarian trees found at Dixon Fold. (Page 5.) Fig. 2.�Oast of a second of the same group of trees. (Page 5.)

PLATE TL

Fig. 3.�Under surface of a Stigmarian root in the Museum of the Leeds Philosophical Society. (Page 5.)

Fig. 4.�Under surface of a Stigmarian root in the Museum of the College of Science, Leeds. (Page 6.)

Fig. 7.�Medullary angles of three vascular wedges from a transverse section of the vascular cylinder of a Stigmaria. a. Medullary cells, b. Vascular wedges.nbsp;V. Elongated cells of primary medullary rays. hquot;. Secondary medullary rays.nbsp;X 90. Cabinet, No. 744. (Pages 10, 14.)

Fig. 17.�Part of a transverse section of a vascular cylinder, including a longitudinal section of a secondary medullary ray containing barred Tracheids.nbsp;Two radial lines of vessels of the cylinder, h, b. Barred Tracheids, bquot;. X 80,nbsp;Cabinet, No. 767. (Page 15.)

Fig. 19.�Peripheral portion of the transverse section of the vascular cylinder, Fig. 14, PI. VII, in which a new zone is being added exogenously. Older vesselsnbsp;of the cylinder, b. Imperfectly developed vessels, amp;c., of the newer zone, b'. X 39.nbsp;(Pages 17, 24, 32.)

Fig. 20.�Periphery of a transverse section of a vascular cylinder with an investing cellular layer, which appears to be a cambium. Peripheral vessels ofnbsp;the vascular cylinder, b. Procambial cells, c. X 75. Cabinet, No. 744. (Pagenbsp;17.)

Fig. 25.�Transverse section of a fragment of Stigmaria with a considerable development of the prosenchymatous zone, e, of figs. 22 and 23, Plate VIII. X 3.nbsp;Cabinet, No. 795. (Page 20.)

Figs. 28, 29, 30, 31, 32, 33.�Transverse sections of six rootlet bundles after they have escaped from the peripheral ends of the primary medullary rays, butnbsp;before they have penetrated the cortex. In each of these, �' indicates the proto-xylem of the wedge-shaped bundle and the broad opposite base of the wedge,nbsp;including the small vessels, �, its latest additions, X 80. Cabinet, No. 744.nbsp;(Pages 23, 24, 26, 31.)

Fig. 53.�Small portion of the longitudinal section of a young rootlet bundle. Vascular bundle, �. Cells of the bundle sheath, �,�quot;. X 100. (Page 28.)

Fig. 8.�Tangential section from a vascular cylinder, showing transverse sections of the primary medullary rays, h'. Vascular rootlet bundles, �'. X 5* Cabinet, No. 771. (Pages 13, 14, 34, 35.)

Fig. 13.�Transverse section of the primary medullary ray, Plate IX, fig. 12, but made close to the cortical surface of the vascular cylinder, b. Vascular rootletnbsp;bundle,/. First formed portion of the rootlet bundle, �'. Secondary medullarynbsp;rays, d. X 30. Cabinet, No. 856. (Pages 14, 23.)

Fig. 16.�Transverse section of a young primary medullary ray. Vessels of the xylem cylinder, b'. Cells of the medullary ray, b. Primary vessels of thenbsp;rootlet bundle, �'. Later formed vessels, �. Secondary medullary rays, d. X 14.nbsp;Cabinet, No. 824. (Pages 14, 23, 31.)

Fig. 16a.�Transverse section of the vascular cylinder of a very young Stig-maria. Cabinet, No. 773. X 5. (Page 15.)

Fig. 46.�Transverse section of a rootlet cushion below the base of a rootlet. Vascular rootlet bundle,/. Small cells of the rootlet cushion, h. Zone of largenbsp;cells investing the rootlet cushion, h'. X 20. (Pages 26, 28.)

Fig. 47.�Similar section to 46, only partly invested by some of the tissues of the root-cortex. Small cells of the rootlet cushion, h. Investing zone of largernbsp;cells, h'. Prosenchymatous cells of the root-bark seen in tangential section, e.nbsp;X 20. Cabinet, No. 807. (Pages 26, 28.)

Fig. 48.�Central portion of Fig. 47, further enlarged. Rootlet bundle, /. Phloem portion of the bundle, �. Cellular investment of the bundle, X 200.nbsp;(Pages 26, 28, 32.)

Pig. 49.�Some of the prosenchymatous cells of Fig. 47, e, undergoing fission by horizontal septa. X 40. (Page 20.)

Pig, 9,�Longitudinal section through a young root of a Stigmaria. Medullary cavity, cb. Vascular cylinder, b, b. Vessels of the cylinder deflected to form anbsp;rootlet bundle, �, passing out through a primary medullary ray, b'. Secondarynbsp;medullary rays, bquot;'. Cortex parenchyma, d. Cortex meristem, d'. Base of rootlet,nbsp;g. Rootlet cushion, fe. X 10. Cabinet, No. 776. (Pages 14, 18, 19, 20, 22, 31,nbsp;34, 35.)

Fig. 45.�Longitudinal section of the proximal end of another rootlet. Parenchymatous zone of the root-bark, d. Prosenchymatous zone of the root-bark, d'. Rootlet bundle, �, enclosed within its rootlet bundle-sheath, � , and passing outwards through the rootlet cushion, h, h'. Apex of the rootlet cushion, h . Outernbsp;cortical zone of the rootlet, g. X 18. (Pages 19, 26, 27.)

Lig. 10.�A radial section through the outermost vessels, b, of the vascular cylinder. Cambiform zone, c. Cells belonging to a more external corticalnbsp;parenchyma, d, d. X 90. Cabinet, No. 779. (Pages 13, 17, 18.)

Fig. 11.�A tangential section through a portion of a vascular cylinder. Xylem laminm, b. Secondary medullary rays, b'. Rootlet bundles, f. X 5. Cabinet,nbsp;No. 771. (Pages 11, 13, 15.)

Fig. 14.�Transverse section of a medium-sized vascular cylinder. Medullary cavity, a, containing an intruded Stigmarian rootlet, g. Vascular wedges, b.nbsp;Primary medullary rays, V. Obliquely longitudinal sections of rootlet bundles,nbsp;normally co-extensive with the entire thickness of the cylinder, �. Section of a rootlet bundle escaping from the cylinder, �'. X 5. (Pages 13,14,17, 23, 25, 33,35.)

Fig. 26.�A portion of Fig. 25, further enlarged. External pressure has disturbed the transversely intersected prosenchymatous cells along the lines, e', e.nbsp;X 18. (Page 20.)

Fig. 15.�Transverse section of a vascular cylinder with its more external cortical layer and the bases of three rootlets. Medullary cavity, a. Wedges ofnbsp;the vascular cylinder, h. Outermost cortical parenchyma, d. Meristem zone ofnbsp;bark, d'. Cortical prosenchyma, e. Rootlet bundle, �. Rootlet bundle cylinder,nbsp;�quot;. Cortical zone of rootlet, g. Rootlet cushion, h. X 6^. Cabinet, No. 759.nbsp;(Pages 14, 18, 26.)

Fig. 22.�Outer cortex of a transverse section from the same specimen as Fig. 15. Outermost parenchyma, c?. Meristem or bark cambium zone, d'. Pro-senchymatous zone, e. X 18. Cabinet, No. 757. (Pages 19, 20, 21, 43.)

Fig. 23.�Transverse sections of the cortex of a much older Stigmarian root than Fig. 22. Outermost cortical parenchyma, d. Meristem zone, d'. Cell wallsnbsp;of mother cells ? dquot;. Prosenchymatous zone, e. X 18. (Pages 19, 20, 21, 43.)

Fig. 24.�Tangential section through the meristem tissue, d', of Fig. 23. Boundaries of mother-cells, identical with those of d� of Fig. 23. X 9.nbsp;(Page 20.)

20, 21.)

Fig. 12.�Tangential section of part of a xylem cylinder, made close to the medulla, giving a transverse section of a primary medullary ray, h, and thenbsp;vascular rootlet bundle, �, f. Enlarged secondary medullary rays at the orientation of the younger part of the rootlet bundle, d. X 30. Cabinet, No. 851.nbsp;(Pages 14, 23.)

Fig. 18.�Transverse section of an extremely young root of Stigmaria. Vascular wedges of the disturbed vascular cylinder, h. Superficial parenchyma of the cortex,nbsp;d. Cortical prosenchyma, d'. Rootlets belonging to the root, g, g. X 11.nbsp;Cabinet, No. 775. (Pages 15, 18, 42.)

Fig. 34.�Another section like figs. 28�33, Plate IV, but with all its newest small vessels derived from a half-developed exogenous zone like Fig. 19, h', PI. IV.nbsp;X 80. (Page 23.)

Fig. 35.�A wedge-shaped rootlet bundle from the same position as the preceding ones, but from the exterior of the large vascular cylinder from Staffordshire, Protoxylem vessels,/'. X 80. Cabinet, No. 750. (Pages 23, 24.)

Fig. 51.�Transverse section of a rootlet made about half way up the conical portion, h', of the rootlet cushion. The outermost parenchyma of the root-bark,nbsp;d. The cortical layer of the rootlet, g. The second cellular layer of the rootletnbsp;very rarely preserved, g'. Section of the small cells of the rootlet cushion, hquot;.nbsp;Rootlet bundle and its sheath, �. x 50. Cabinet, No. 821. (Page 27.)

Fig. 52.�Centre of Fig. 51. Xylem portion of the rootlet bundle, �. Phloem area of the bundle, �'. Bundle sheath, �quot;. X 440. (Pages 28, 32.)

Fig. 21.�Transverse section of the bark of Fig. 18, PI. IX. Outer parenchyma, d. Prosenchymatous zone, d'. X 18. Cabinet, No. 775. (Pages 18,19.)

Fig. 42.�Transverse section of a large Stigmarian rootlet, g, into the interior of which several other younger rootlets have penetrated. Vascular bundle of thenbsp;large rootlet,/. Cortex of one of the smaller rootlets,/'. Its inner vacant cavity,nbsp;g'. Its vascular bundle and sheath, �'. X 11. Cabinet, Xo. 848. (Pagesnbsp;13, 25, 26, 33.)

Fig. 43.�Longitudinal section of the proximal end of a rootlet. Parenchymatous cortex of the root, d, d. Prosenchymatous cortex of the root, e. Vascular rootlet bundle, �, �. Rootlet bundle-sheath, �'. Branching cells investing thenbsp;extremity of the rootlet cushion, i, i. Outer cortex of the rootlet, g, g. Fistularnbsp;interior of the rootlet, g'. X 18. (Pages 26, 27, 28.)

Fig. 44.�Longitudinal section of the proximal end of another rootlet. Prosenchymatous zone of the root-bark, d'. Central portion of the rootlet cushion, h, h. Conical extremity of the rootlet cushion inside the rootlet, h'. Rootletnbsp;bundle, �. Rootlet bundle-sheath, �'. Outer cortex of the rootlet, g. Fistularnbsp;interior of the rootlet, /. X 18. Cabinet, Xo. 819. (Pages 26, 27, 28, 39.)

Fig. 50.�Minute branching tubular cells investing the surface of the conical extremity, Ji, of the rootlet cushion, seen in the longitudinal rootlet sections,nbsp;e. g. fig. 43, i, PI. X. (Page 27.)

PLATE XI.

Pig. 36.�A second wedge-shaped rootlet bundle, like fig. 35, PI. IX, from the exterior of the large vascular cylinder from Stafibrdshire. Protoxylem vessels,/',nbsp;X 80. Cabinet, No. 750. (Pages 23, 24, 34.)

Pig. 54.�Transverse section of the rootlet bundle with its bundle cylinder, of a young rootlet. Very small Tracheids at the point of departure of the monarchnbsp;rootlet bundle, �, �quot;. Bundle cylinder, �'. x 100. Cabinet, Xo. 831. (Page 29.)

Pig. 55.�A second rootlet bundle, like Pig. 54. X 100. Cabinet, No. 829. (Pages 29, 31.)

Pig. 56.�A third, more advanced, rootlet bundle, like Pig. 54. X 100. Cabinet, Xo. 775. (Page 29.)

Pig. 57.�Vascular bundle,/, and part of the bundle cylinder,/quot;, of a larger and older rootlet. Point of departure of the monarch bundle, /'. X 100.nbsp;Cabinet, Xo. 836. (Pages 29, 31.)

Pig. 58.�Rootlet bundle, /, and its cylinder, /quot;, of a yet older rootlet. X 100. (Pages 29, 31.)

Pig. 59.�Rootlet bundle, /, with part of its monarch initial vessels at/', but with two others yet deeper amongst the cells of the bundle cylinder at /'quot;.nbsp;(Page 29.)

Pig. 60.�Rootlet bundle, with part of its bundle cylinder, /quot;. The more recently added of the vessels of the bundle, /, radiate from the monarch point ofnbsp;departure at x 100. (Pages 29, 30.)

Pig. 61.�A second rootlet bundle, like Pig. 60, with its initial vessels at /', and a small portion of its cylinder at/quot;. Here again the vessels of the older part ofnbsp;the bundle are non-radial; the younger additions are disposed radially. X 100.nbsp;Cabinet, Xo. 651. (Pages 29, 30, 34.)

Pig. 62.�Transverse section of a very young bundle, /, with its phloem elements preserved at/'quot;. Bundle cylinder,/quot;. (Page 32.)

Pig. 63. �Transverse section of a rootlet about to dichotomise. Inner border of the cortical zone of the rootlet, y. Pistular cavity of the rootlet, y'. Twonbsp;bundles, /, /, each surrounded by its bundle-sheath or cylinder, /', /', resulting fromnbsp;the division of one single bundle. X 75. Cabinet, Xo. 862. (Page 32.)

Fig. 37.�Part of a fragment of a Stigmarian root split vertically. Exterior of the vascular cylinder, b. Rootlet bundles crossing the cortical region, �. Exterior surface of the cortex, d, d. Half the natural size. (Page 24.)

Fig. 38.�One extremity of Fig. 39, h. Vascular wedges, h, of the cylinder separated by primary medullary rays, V. The central medullary cavity is occupiednbsp;by an intruded Stigmarian rootlet, in the interior of which the rootlet bundle isnbsp;preserved. Nat. size. Cabinet, No. 867. (Page 24.)

Fig. 39.�Side view of Fig. 38, with numerous rootlet bundles, �, enclosed within their several bundle sheaths, issuing from the primary medullary rays ofnbsp;the cylinder. Nat. size. Cabinet, No. 867. (Pages 24, 26.)

Fig. 40.�Fragment of a large nodule enclosing a disorganised Stigmarian root, almost identical with one of which the specimen. Figs. 38, 39, is the vascularnbsp;cylinder, d represents a cast of the outer surface of one side of the root, withnbsp;casts of the basal ends of the rootlets penetrating the stone. Rootlets from thenbsp;opposite side of the root are seen at g, g, with their free ends pointing to the right.nbsp;�, a mass of rootlet bnndles resembling those given off from the exterior of Fig. 39.nbsp;(Pages 24, 39.)

Fig. 41.�Inner surface of a fragment of the outermost bark, with rootlet bundles penetrating it to reach the rootlet-scars existing on the reverse side of thenbsp;specimen. (Page 25.)

Fig. 70.�Fragment of the vascular cylinder of a Stigmaria, upon both surfaces of which the bases of rootlet-scars have been impressed on the disappearance ofnbsp;all the other cortical tissues. Nat. size. Author�s Cabinet. (Page 37.)

Fig. 72.�Cast of the outer surface of a fragment of Stigmarian bark with undulating coiTugations. Nat. size. Museum of the Owens College. (Page 38.)

Fig. 74.�Fragment of Stigmarian bark in which the bases of the rootlets are preserved as funnel-shaped depressions, within which the outer cortex of the rootletnbsp;forms a prominent ridge at g, and the conical apex of the rootlet cushion is seen atnbsp;the bottom of the funnel, h. Nat. size. Author�s Collection. (Pages 38, 40.)

Fig. 76.�Diagrammatic section of a Stigmarian rootlet attached to its bark, d', dquot;, illustrating Figs. 74 and 75. Inner surface of the parenchymatous layer of thenbsp;bark, d'. Outer surface of the same, dquot;. Small remnant of the outer cortex of thenbsp;rootlet, gquot;, gquot;, being the point near which the rootlet, g, had been broken offnbsp;from its basal end. Cylindrical interior of the base of the rootlet, g', correspondingnbsp;to the funnel-shaped depressions, h, of Fig. 74, and to the central prominences ofnbsp;Fig. 75. Apex of the cone of the rootlet cushion, li, forming the apical depressionnbsp;of each central prominence of Fig. 75. (Page 39.)

Fig. 64.�A natural cast of the medullary cavity of a Stigmarian root, the superficial longitudinal ridges of which represent the prolongations of the medullanbsp;into the inner extremities of the primary medullary rays. Nat. size. Huttonnbsp;Collection. Newcastle Museum. (Page 35.)

Fig. 65.�Natural cast of half of the exterior of the vascular cylinder, h, of the same specimen as Fig. 64, showing the peripheral terminations of the primarynbsp;medullary rays; mineral matter replacing the entire bark, d. Hutton Collection.nbsp;(Page 36.)

Fig. 66.�A second natural cast, like Fig. 65, showing much more strikingly the peripheral termination of the primary medullary rays. Nat. size. Huttonnbsp;Collection. (Page 36.)

Fig. 67.�Transverse section of a Stigmarian root, nat. size, from which all organic elements have been removed. The inorganic matrix occupies the medullary cavity, a, the primary medullary rays, h', and replaces the entire bark, d.nbsp;Medullary rays, V. Areas left vacant by the disappearance of the vascular wedgesnbsp;of the vascular cylinder, h. Hutton Collection. (Page 36.)

Fig. 71.�Cast or impression of the surface of part of a dichotomising Stigmarian root, exhibiting the casts of what were longitudinal fissures in thenbsp;exterior of the bark, along with more minute undulating ridges and furrows.nbsp;Half nat. size. Hutton Collection. (Page 37.)

Fig. 78.�A specimen from the Newcastle Museum of the Stigmaria stellata of Goeppert. (Page 40.)

Fig. 79.�Transverse section of a Stigmarian rootlet in which the rootlet bundle and its rootlet cylinder, �, is united to the inner side of the rootlet-cortex, g, by anbsp;cellular lamina, �quot;, running longitudinally through the rootlet. (Page 32.)

Fig. 68.�Fragment of a Stigmarian root split longitudinally. Exterior of medulla, a. Long section of tlie vascular cylinder, h, h. Primary medullary raysnbsp;passing horizontally through the vascular cylinder, V. Cortex replaced by inorganic matrix, d, d. Author�s Cabinet. (Page 37.)

Fig. 69.�Exterior of the vascular cylinder of a Stigmaria, from Burntisland. Nat. size. Showing the large peripheral orifices of the primary medullary rays, h, h,nbsp;intermingled with a few rootlet-scars. Author�s Cabinet. (Page 37.)

Fig. 73.�A normal portion of a Stigmarian root deprived of its rootlets. Cavity left by the disappearance of the vascular cylinder, b, Nat. size. Author�snbsp;Collection. (Page 38.)

Fig. 75.�Cast of the outer surface of a Stigmarian bark which has been in the condition of the specimen. Fig. 74, and is illustrated by the diagram,nbsp;fig. 76, Plate XII. Nat. size. Author�s Collection. (Page 38.)

Fig. 77.�External surface of a Stigmarian bark, like Fig. 74, modified by shrinkage or compression. Nat. size. Author�s Collection. (Page 40.)

Fig. 80.�View, taken obliquely from above, of a magnificent Stigmaria ficoides, recently discovered at Clayton, near Bradford, and now in the museumnbsp;of the Owens College, Manchester. (Page 45.)

�O J��.

A MONOGRAPH

MORPHOLOGY AND HISTOLOGY

8TIGMARIA FICOIDES.

WILLIAM CRAWFOED WILLIAMSON, LL.D., F.E.S.,

LONDON:

1887.

INTRODUCTION.

INTRODUCTION.

INTRODUCTION.

A MONOGRAPH

STIGMARIA FICOIDES.

2 nbsp;nbsp;nbsp;STIGMARIA PICOIDES.

4 nbsp;nbsp;nbsp;STIGMARIA PICOIDES.

STIGMARIA PICOIDES.

MEDULLA. nbsp;nbsp;nbsp;9

Having tLus glanced at some of the general aspects of Stigmaria ficoides, we may now examine the morphology and histology of its several parts.

This was an exclusively parenchymatous tissue; and, since I found it to be hollow in every one of the innumerable specimens that passed under my eye, I longnbsp;ago arrived at the conclusion that Stigmaria possessed a fistular medulla. Owing

STIGMAUIA PICOIDES.

STIGMARIA FICOIDES.

medulla and were the sources whence the vascular bundles going to the rootlets were derived. This, however, is altogether a mistake.^

VASCULAR CYLINDER.

STIGMARIA FICOIDES.

t quot;V^H

' speci�... quot;f stig�.ri.y

STIGMARIA FICOIDES.

The rarity of small branching specimens, along with other facts, seems to show that after the first two consecutive bifurcations occurred within a limitednbsp;distance from the central stem, no others took place in the true roots. The long,

.1

VASCULAR CYLINDER.

20

STIGMARIA FICOIDES.

STIGMARIA. FICOIDES.

STIGMARIA PICOIDES.

26 nbsp;nbsp;nbsp;STIGMARIA FICOIDES.

STIGMARIA FICOIDES.

STIGMARIA FICOIDES.

STIGMAUIA nCOIDES.

STIGMARIA PICOIDES.

There still remain for consideration some fragments of Stigmarise which, though not capable of microscopic examination, could not have been interpreted withoutnbsp;the knowledge which we owe to the microscope.

STIGMARIA PICOIDES.

ROOTS AND ROOTLETS.

STIGMARIA FICOIDES.

ROOTS AND ROOTLETS.

STIGMARIA HCOIDES.

CONCLUSIONS.

STIGMARIA FICOIDES.

CONCLUSIONS.

APPENDIX.

STIGMARIA FICOIDES.

ll

APPENDIX.

STIGMARIA PICOIDES.

APPENDIX.

GLOSSAEY OP BOTANICAL TEEMS EMPLOYED IN THE

MONOGEAPH.i

STIGMARIA FICOIDES.

GLOSSARY.

STIGMARIA FICOIDES.

DESCEIPTIYE INDEX TO THE PLATES.'

58

DESCRIPTIVE INDEX TO THE PLATES.

16a. V. 17. IV.

DESCRIPTIVE INDEX TO THE PLATES.

DESCRIPTIVE INDEX TO THE PLATES.

DESCRIPTIVE INDEX TO THE PLATES.

61

35

36

36

36

37

37

37

37

38

38

38

38

39

DESCRIPTIVE INDEX TO THE PLATES.