ter verkrijging van den graad van doctor in
de wis- en natuurkunde aan de rijks-univer-
siteit te utrecht, op gezag van den rector
magnificus Dr. L. S. ornstein, hoogleeraar in de
faculteit der wis- en natuurkunde, volgens
besluit van den senaat der universiteit tegen
de bedenkingen van de faculteit der wis-
en natuurkunde te verdedigen op vrijdag
8 juli 1952, des namiddags te 4 uur

1; \'

Hoewel mijn dank uitgaat naar allen, die tot de vorming van
mijn geest hebben bijgedragen, wil ik mij op deze plaats beperken
tot woorden van erkentelijkheid aan U, Hooggeleerde Rutten,
ScHMUTZER en Nierstrasz.

Hooggeleerde Rutten, Hooggeachte Promotor. Ik meen in
Uw geest te handelen, wanneer ik er van afzie onder woorden te
brengen de zeer groote waardeering, welke ik voor U gevoel in

echter toegestaan er op te wijzen, dat de ononderbroken stroom van
vitaliteit, welke van U uitgaat en Uw uiterst snelle, kritische geest
van zeer veel beteekenis zijn geweest voor de totstandkoming en
de inhoud van dit werk.

Hooggeleerde Schmutzer. Assistent te zijn bij U is een voor-
recht. Zoowel door het nauwe contact tusschen U en mij, als wel
door de groote vrijheid, welke Gij mij hebt toegestaan, heb ik
ruimschoots gelegenheid gehad mijn kennis te verrijken. U kunt
er van overtuigd zijn, dat ik al mijn krachten zal aanwenden om
U bij te staan in Uw streven de mineralogische afdeeling van het
Instituut tot groote bloei te brengen.

Hooggeleerde Nierstrasz. Uw ongeëvenaarde wijze van do-
ceeren is één der voornaamste redenen, waardoor mijn studiën met
magnetische kracht in de richting der zoölogie getrokken werden.
Voor een groot deel dank ik het aan Uw aanbeveling, dat mij deze
kostbare collectie werd toevertrouwd. Moge de tijd leeren, dat ik
het door U in mij gestelde vertrouwen waard was.

Velen hebben op eenigerlei wijze bijgedragen tot het bereiken
van deze mijlpaal in mijn leven. In de allereerste plaats mag in dit
verband mijn vrouw genoemd worden. Haar daadwerkelijke en
moreele steun zijn voor mij van onschatbare waarde geweest.

Aan Uw groote ervaring op fotografisch gebied, waarde West-
broek, dank ik de voortreffelijke kwaliteit der lichtdrukken.

Waarde van Dijk, mijn hartelijke dank voor de vervaardiging
der kaartjes en tekstfiguren, alsmede voor het uitstekende prepareer-
werk.

Ik wil niet eindigen zonder met erkentelijkheid gewag te hebben
gemaakt van de zeer gewaardeerde medewerking, welke ik steeds
van de ambtenaren der diverse bibliotheken heb mogen onder-
vinden. Inzonderheid U, Zeergeleerde van Lummel, mijn dank
voor de correctie der bibliografie.

Bujfelus bubalus (L.) ?var. j-öWö/V/^j-(Schl, et Müll.)/öj-j-Z/Zj- 25
Bibos sondaicus (Schl. et Müll.) fossilis........

APPENDIX: \' \' ^
On the terms applied to the primary and secondary
elements of upper premolars and molars in general, and
those of rhinoceros in particular........... ^^

The material, the description and determination of which
forms the subject of the present paper, was part of an extensive
collection, brought together by the exertions of the Geological
Survey („Opsporingsdienstquot;) of the Dutch East Indies, and en-
trusted to the author for examination. The specimens of this collection
had been procured, partly from a new finding spot in the vicinity
of Bumiaju (Central Java), partly from Javan localities already
known for a long time. A renewed reconnoitring of the latter
took place namely, in order to prepare a scheme, which had in view
cooperation between the Survey mentioned and the American
Museum of Natural History. On that occasion a rather great number
of mammalian remains was collected, among which were some very
fine specimens. The plan above mentioned was not carried into
effect, so that the latter material was also sent to the writer.

The reader may form himself an idea of the extensiveness of the
whole collection, if is mentioned that this consisted of four sendings,
containing in total 31 cases, pardy of considerable size 1). The first,
third and fourth sending consisted exclusively of specimens, ob-
tained from the vicinity of Bumiaju. The second sending (18 cases)
contained the material, collected in the other localities.

The packing was so excellent that most of the specimens either
had experienced no harm whatever, or had hardly suffered any
injury on the Jong journey. Only one specimen viz., a cranium of
Bibos sondakus fossilisy arrived in a very damaged condition. This
was, however, not owing to the packing, but a result of the small
degree of fossilization of that specimen.

Though in Bandung already much time was spent on the pre-
paration, it appeared that by prolonged preparation a good deal of
the specimens described could be brought into a considerable
better condition. This work, wliich required both patience and skill
has aptly been done by Mr J. van Dijk, amanuensis of the Geolo-
gical-Mineralogical Institute of the University of Utrecht.

The collection sent to Utrecht consisted for the greater part of
mammalian remains. The cranial and dental remains of Buffelus
Bibos, Rhinoceros, Hippopotamus, quot;Elepbaf\', Stegodon, quot;Masfodonquot;

1) Beside these 31 cases one more case was rcccived, comprising molluscs from and
samples of rocks of the vertebrate bearing strata, near Bumiaju. The writer intends
to examine the rocks, possibly the molluscs too.

In accepting the examination the writer had undertaken to care
that the greatest possible number of specimens should be returned
before the beginning of the fourth Pacific Science Congress (Ban-
dung 1929). Consequently, the specimens, after having been care-
fully examined, described, measured and pictured, immediately
set out on the return journey, provided or not with a provisional
name. It will hardly be necessary to point out that this worldng
method involved difficulties. On the other hand, it yielded an ad-
vantage, which is not to be underestimated. The very fact that the
fossils were sent back before their determination was definitely
ended, gave rise to an absolutely objective description of each
specimen, and as a matter of necessity the description had to
remain objective, as later on its contents could not be changed
any more.

In what follows the reader will find an enumeration of all the
specimens described, and of the localities 1) from whence they
have been obtained 2).

Fr. upper jaw with inj. 1. and r. M^.
Fragment of cranium with 1. and r. M^.
Incisive tusk.
?Fr. pelvis.

In this list quot;fr.quot; means fragmentary and quot;inj.quot; injured. A sign of interrogation in front
ot the arcumscnption of a specimen indicates its possible belonging to the form under
which It is mentioned.

Tetralopbodon humiajuensis nov. spec.:
Fr. r. mandibular ramus with Mg.
? Two femora (n^s. z and 6).
? Three ulnae (n^s. i—^3),
? Two radii (n^®. i and 2).
? Tibia (n^. 2).
Hippopotamus sp.:

Fragment of cranium with 1. M^.
Archidiskodon planifrons (Falc. et cautl.)-
Detached 1. M^.

Buffelus huhalus (L.) .?var. sondaicus (Schl. et Müll.) fossilis-.
Fr. cranium and two detached horn-cores.

Buffelus huhalus (L.) ?var. sondaicus (Schl. et Müll.)/öj-j-///j-:
Three fr. crania.

Buffelus bubalus (L.) ?var. sondaicus (Schl. et Mull.) fossilis:
Two fr. crania.

Buffelus bubalus (L.) ?var. sondaicus (Schl. et Müll.) fossilis:
Two fr. crania.

= 25.000) relating to the matter in question, wc only
found Alas Wedilembut . Alas means forest,

According to Zwierzycki 2) the first finds were made by
Mr. N. de Zwaan at Limbangan, who discovered in 1922—\'23
some loose specimens in the Kali Glagah and Tji Saät It lasted
till March 1923 before Mr. Buning of Cheribon announced these
finds m the papers. The result was that Van der Vlerk, at
that time palaeontologist of the Geological Survey, by order of
that Survey made inquiries on the spot. A brief communication
ol his experiences will be found in the quot;Mijningenieurquot; of 1923 4).

The specimens found by Van der Vlerk, together with the
collection presented by De Zwaan to the Geological Survey
were sent for examination to Stehlin in Bale. The results of this
investigation were embodied in a paper entitled: quot;Fossile Säugetiere
aus der Gegend von Limbangan (Java)quot; %

All of Stehlin\'s specimens had been obtained from the Kali
Biuk and Tji Saät 6). They were considerably rounded, distinct-
ly indicating river transport. Only a few specimens allowed of
determimng the species. Noteworthy is that Stehlin recognized

According to the list of localities, which accompanied the second sending, and also
to the labels, Kedung Kendang is situated in the residency Madiun. From the staff-man
It appeared, however, that a campong of that name does not occur in the residency
w^r vi. p It does occur in the government Surakarta, and as it is this campong.
rWtoX r .nbsp;1) mentions as a locality of vertebrate remains, we may be sure

De Mijningenieur, Jrg. 7, 1926, p. 229.
See for the topographic names map nquot;. i.

Lower sandstone-conglomerate group.
Turritella zone.
Brcccia zone.
Limestone-marl zone.

of a molar described and figured by Martin gt;) under the name of
E hjsudnms Falc. et Cautl. He continued, however, quot;Die Ent-
scheidung der Frage, ob dieser javanische Elephant mit der fest-
ländischen Snvahkform specifisch übereinstimmt, ist cura posterior-
d e Wissenschaft verhert nichts dabei, wenn wir sie hinausschieben
bis von den javamschen Fundstellen ein etwas reichlicheres Beleg-
materiaj zusammen gebraciit istquot;»).

Stehlin\'s species III was only represented by a small fragment
of a grindmg tooth, only consisting of the third part of two ridge-
plates It was however, enough to reveal some primitive characters.
(It may already be mentioned now that the writer\'s collection
confined two pndmg teeth of elephant, the one found in the
Kdi Bmk, the other m the TjiPangglosoran, which are also character!
i^ed by the possession of a number of distinctly primitive characters)

.u ^ abweicht; wenigstens wüsste ich kein Argument namhaft
zu machen, das erlaubte, dieselbe mit Bestimmtheit für älter oder
fur junger als letztere zu erklärenquot;.

tkleT \'^Rn?\'f^\'\' Geological Survey, and en-
fntunbsp;Geological guide to the locality of

fossil vertebrates in the Kah Glagahquot;, and pubhshed on the occaLn
of the fourth Pacific Science Congress (1929).

In 1925 a new locality was discovered in the bed of the Kali
Glagah - If I understand well - by Mr. Boning, already men-
tioned As a result of his report a renewed investiga\'tion was made
by Oppenoorth at that time superintendant of\'the Java party
of Znbsp;J-^^ Van Es 3), mining engLe^

Leiden, IV. (1887), p. 57, pi. yi. figs. a-a..
I93I^p^■x6:nbsp;on^ithecanthropus. The Hague 1931. Dissertation Delft

would be possible to find the bones in situ, and if so, whether
excavations would take effect. This appeared indeed to be the
case. And it is the specimens, which have been excavated and
collected in the years 1925 and 1926, which were sent to the writer
for examination. At the same time some geological researches
were made in the neighbourhood by the mining engineer Ter Haar
and ZwiERZYCKi. All the excavation operations were put to a
temporary standstill during the negotiations about the cooperation
already mentioned of the Geological Survey and the American
Museum of Natural History. When it appeared that these plans
had to be abandoned, the operations were continued in the second
half of 1928, with the result that an interesting collection was
brought together, containing remains of hippopotamus, ruminants,
crocodile and tortoise.

The writer received a geological map of the region between
Limbangan and Bumiaju, together with two sections. Map and
sections were made by Zwierzycki. We were told that they were
not intended for publication. Nevertheless, we are able to give
a geological map of that region. The geological guide, already
mentioned, contains namely a geological map. Our map n°. 1 has
been made after a part of the latter. For the sake of clearness the
site of only those excavations has been added, wliich furnished
part of the specimens described in the following pages. One
of the two sections, which were sent to us, has been published
by \'t Hoen in the quot;Jaarboek van het Mijnwezen in Ned,-Indiequot;
(Verhandelingen), Jrg. 1929, 1930, p. 30, so that we should have
been entitled to reproduce it. From reasons which will be dealt
with presently, we refrained from it. Already now it may be mention-
ed that\'t Hoen\'s publication contains also a stratigraphic column
of the different strata of the region under consideration (after
Zwierzycki), and that according to this column the total thickness
of the vertebrate bearing layers is about 1200 m.

It is easily to see that the section, published by \'t Hoen is
more or less perpendicular to the strike. No mention was made
however, of the exact situation of the section. It will be desirable\'
therefore, to mention that, according to the unpublished map\'
received by us, the section is situated in the south-easterly part
of both synclines, more particularly going via that place between
both synclines, where the lower sandstone-conglomerate group
crops out between two strips of the Turritella zone, A superficial
comparison between map (Ter Haar) and section (Zwierzycki)
will reveal that in the latter respect no perfect accordance exists
between the two, the section showing a more simple structure. On
the other hand the section shows the presence of a vertical fault

m the soudiwestern syncline, wheieas on the map no fault- will
be found. Van Esgt;) remarked that the tectonicTtr^ctvT s lo
o simple as would appear from the section pub quot;b tha

SSyncline. At tte tte
^„Jf \' Pubhcation (19^9) the exact length of the fLm^?
(and probabty also of the latter) was not known fet. Both Uc Ls
are separated by an anticline, in the core of wWch the dayquot;mS

white sandstones.
. sandstone-conglomerate
series with argillaceous
sandstones, clay and
marl, locally with beds
of lignite, and mostly
bearing lime.

lt;u

tu
u
O
»—!

cS
u

Hippopotamus sp.
Several species of Me-
lania and Corbicula, 69 %
of which recent species.
Some marine molluscs
and foraminifera.

blueish-gray and greenish
argillaceous marls, subor-
dinate sandy marls,
mollusc bed.

H ^

quot;S §

3 N

(U
Ö
lt;u
u
O

o a .
s y o

cS S S

hard, usually coarse basic
andesitic breccia. Inter-
calations of tuff beds, and
-at 300 m. from the base-
a horizon of pumice-stone
breccia of a thickness of
zoo m; locally lava.

lt;u

a
o

u
u

fine and coarse greenish
andesitic sandstones and
lime bearing sandstones
with intercalations of marl
beds.

more towards the top the
sandstones become more
conglomeratic and quot;brec-
cia-likequot;,
a characteristic horizon is
formed by a series of
light coloured tuff-sand-
stones and grit with pieces
of pumice stone.

«J
Ö
O

■M
M

greenish, concretionary
Globigerina marls, in the
upper part bedded by thin
layers of sandstones; more
towards the base gray-
ish-green marls with li-
mestones, containing ma-
ny foraminifera; finally
unstratified concretionary
argillaceous marls.

^ y
H §

Trjhliolepidina rutteni
with several varieties.
Cycloclypeus neglectus.
Cristellaria sp.
Operculina sp.
Operculimlla sp.
Amphistegina sp.
Gypsina globulus.
Kotalidae, Algae, and
pricks of ecfinids.

Both the author of the report as well as Van Es agree as to
the pliocene age of the series composed by the Vertebrate zone,
Turritella zone and Conglomerate zone. I do not know, however,
on what grounds the lowest boundary of the Pliocene has been
drawn between the Conglomerate zone and the Breccia zone
which are both unfossiliferous. Furthermore, it may be pointed
out that in all probability in Ter Haar\'s map (and therefore
also in our map n°. i) both. Conglomerate zone and Turritella
zone have been indicated by the same vertical signature. I do not
know, however, with which zone the Tuff-sandstone zone has been

It will be remembered that in the foregoing part has been stated
that according to the original opinion of Zwierzycki the total
thickness of the vertebrate bearing layers is about 1200 m. ine
above list shows that according to more recent opinions, the thick-
ness is but 600 m. As Zwierzycki calculated the thickness from the
sections, it is highly probable that the section published by t Hoen
— and already mentioned in the foregoing pages — does not hold
good any more. That is, therefore, the reason why I have refrained

The Vertebrate zone lies at the K. Glagah conformably on the
Turritella zone an interjacent zone of some tens ot meters
consists of alternating Turritella layers and vertebrate layeis,
and begins with conglomeratic, andesitic sandstones, contaimng
lumps of hgnite. The latter horizon is better developed (ca 150 m.)
in the Bentarsari basin, and shows there a serks of layers of impure
lignite, containing 50 % of water, which high percentage should
indicate a pliocene age. The lowest vertebrate bearing layers are
argillaceous marls and tuffoid sandy marls of andesitic material
In the Turritella zone, thus called on account of the frequence of
TurMIa djadjariensis, 46 species of mol uses (gasteropods and
lamellibranchiates) were found, 22 of which, 1. e. 48 % are recent.
This percentage was, however, not used as an argument for a
pliocene age. According to Ter Haar this fauna has been
examined by Gerth, who established the age of the Turritella
zone as older Pliocene. In the report stress has been laid upon
the fact that the fauna shows resemblance to that of Tjidjurai
(Cheribon), examined by Martin, and held to be Pliocene by this
investigator.

As an argument for the pliocene age of the vertebrate bearing
sandstone-conglomerate series the oQc^lZlcnc^^^Mastod^pe^-

mensisy Stegodon airâwana and Hippopotamus sp. i) is mentioned.
Though in a final chapter we shall return to the age of these beds
on the basis of the determination of the mammalian remains, al-
ready now it may be pointed out that the cogency of the argument
mentioned, carmot be called sufficient. It must not be forgotten
that St. airâwana is an upper pliocene species according to Dubois 2),
a lower pleistocene species according to Martin 3), a middle pleisto-
cene species according to Osborn and an upper to youngest
pleistocene species according to Dietrich As to Hippopotamus^
it must be borne in mind that according to Pilgrim the youngest
horizon in which this genus occurs, is the Boulder Conglomerate
zone (uppermost Upper Siwaliks). Pilgrim®) regards the Upper
Siwaliks as of pliocene age. Matthew\'s \') recent investigations on
the Siwalik fauna led him, however, to the conclusion that there
are no valid reasons for referring the Upper Siwaliks fauna
to the Pliocene. Accordingly he reckons the Upper Siwaliks,
to the lower Pleistocene. Finally it may be mentioned that
according to Osborn®) Mastodon (Anancus) perimensis is an upper
miocene species. I must admit, however, I do not know on what
grounds.

Van Es (1. c.) determined in cooperation with Von Koenigs-
wald, one of the palaeontologists of the Survey, part of a collec-
tion of gastropods and lamellibranchiates from the Turritella beds,
collected by him and C. A. de Jongh. The following values were
found. Of 34 species determined, 19 or 56 % were still living. In
Van Es\'s opinion this percentage points decidedly to a pliocene
age, which would be confirmed by the fact that 24 species are
known to occur in the Miocene, against 3 3 in the Pliocene. If one
compares this percentage of 5 6 % with the value found by Van Es
and Von Koenigswald for the fauna (76 species) of the Turritella
beds of Sangiran (45 %), and with those established by Martin
for the rich marine fauna (150 species) of Sonde (53 %), and for
the marine fauna (64 species) of Tjidjurai (51 %), one would be

1) The mention of these forms was based on provisional determinations of the writer.
It may be repeated that in the writer\'s present opinion the mastodont of Bumiaju is
not identical with Mastodon (Anancus) perimensis, but represents a new species (Tetra-
lophodon bwrnajuensis).

Tijdschr. Kon. Ned. Aardr. Gen., scr. 2, XXV, 1908, p. 1257. Dubois reckons S/,
airdwana to St. ganesa, var. javanicus Dub.

Unscre palaeozoologische Kenntnis von Java. Leiden 1919, p. 144.
Proc. Amer. Philos. Soc., LXX, 1931, n°. 2, p. 189.
Sitz.ber. Ges. Naturf. Fr. Berlin, Jhrg. 1924, 1926, p. 139.

\') Bull. Amer. Mus. Nat. Hist., LVI, 1930, p. 445.
Proc. Amer. Philos. Soc., LXIV, 1925, p. 27.

inclined to believe that the Tunitelk beds »^f
greatest affinity to the Sonde beds. It must not benbsp;\'

Cever, that\'tesp. .. and z, Bumiaju species were found to
occur in Sangiran and Tjidjutai, agamst but to m Sond^ Van Es
therefoie, concluded: quot;A comparison with Sonde shows a peat
dhparity to exist, but the resemblance to Sangiran and T)id
fZ sMry obvious. As Sangiran (......Lower quot;f)

\'^d urai Peddle PUocene) are different in age the ataost
affinity to both faunas makes it rather difficult » ^ °de from t^^^
point of view to what horizon theTurritella bed
However considering the rather high percentage of livmg species

wkh theTtter quotation.and with the percentages mentioned above
7 hodd Ïe to point out that the difference between the percentages
o SaÎiran 4\'%) and of Tjidiurai (,1 is ƒ and toween
those of the latter and Bumiaju ;6 %) is 5 %■ Whilst Van ts
rS rffthe fauna of Sangiran to belong to the
coLiders the fauna of Bumiaju to be of the ^me ^ge G^ddle Plio
cene-1 as that of Tjidiurai, in spite of a difference of 5 /o \'\'^twe^quot;
Bu^ah and T idjurli. Herewith I do not intend to ma.ntam ha
age of both is to be excluded. It must not be forgotten that
Xe mage of 56 %, found for the fauna of Bmniaju, is based
on but M s^ whete\'as in Tjidjurai almost twice as many species

\'quot;v\'rEfgav; also the results of the dete^ination by Gerth
of a collectirof fresh-water molluscs (»Vnbsp;Of

Paludina) from the vertebrate bearing layers of Bumia)u_ Of
„ stdes determined, 9 or 69 % appeared to be still livmg. From
tWs vIn S drew the conclusion: quot;This percentage points to an
iSpeT plcLe age of the fresh-water bedsquot; 13 species, however,
Sm too small\\ number to justify so resolute a conclusion^
VanEs\'s statement is the more remarkable, because on p. .6 he
Umself points rightly out that it is .. necessary to exercée g^^^^^
caution L determining the age of the beds f\'om the ratio of living
species when too few Molluscs are present \' He even set some
eLmples, one of which is very noteworthy. In Sangiran Martin
found a percentage of 33 %nbsp;species on 21 species; Van bs

Gerth =) advanced another objection agamst Van Es s de erm
nations of the age. He remarked namely that Van Es, m placing

the young tertiary fauna of molluscs with 50—60 % living species
in the Middle Pliocene, started from a supposition, which lacks
sufficient ground, as long as we do not know that this percentage
is indeed a characteristic of the middle pliocene strata of Java.
We may add that Martin ■— without doubt the best connoisseur
of the Tertiary of Java — has hitherto refrained from subdividing
the Javan Pliocene

Later on we shall have the opportunity to return to Van Es\'s
very important paper. For the present we will pay attention
for a moment to another argument, mentioned by Van Es to
prove the tertiary age of the vertebrate bearing strata of Bumiaju,
viz., their strong folding. The report, already mentioned, of the
Geological Survey even speaks of very intense folding. And as
^— Van Es remarks — quaternary beds with a steep dip are un-
known till now, a tertiary age is most likely. That also this argument
has no absolute cogency, may be proved by a quotation, which we
borrow from Van Es liimself: quot;Arguments derived from the dip
of the beds are but of relative and local value and are insufficient
to serve as a proof for the age of the beds......quot;

We shall now proceed with the consideration of the other loca-
lities. We may begin to ask ourselves whether detailed stratigraphic
and tectonic data of these localities are available. In 1927 the state
of affairs was still such that Rutten ■— after the discussion of
the vertebrate bearing layers of Java ■— had to make the bitter
remark, that not only the tectonic structure and the stratigraphy
of the Trinil beds was very insufficiently known, but also the startl-
ing fact occurred that our knowledge of the geology of the surround-
ings of Trinil, famous by Dubois\'s discoveries of Pithecanthropus
erect us and of a very rich fauna of vertebrates, was but very small.
And had Van Es not published the results of his extensive investi-
gations, we should have been compelled to make exactly the same
remarks. Just because we had to criticize Van Es\'s paper in the
foregoing pages, we are the more eager to avail ourselves of the
opportunity afforded of throwing light upon the great merits of
this work. It is. entirely due to Van Es that at present we dispose
of a lot of stratigraphic and tectonic particulars of numerous local-
lities, all embodied in detailed geological maps and sections. More-
over, Van Es made extensive collections of molluscs from the

Sec his recent paper: quot;Warm löste sich das Gebiet des Indischen Archipels von der
Tethys?quot; Leidsche Geol. Meded., IV, i, 1931.

marine layers, which occur almost everywhere below the verte-
brate bearing series. And though Gerth\'s remark may be true,
namely, that only the relative age has been established by the deter-
minations of these fauna by Van Es (and Von Koenigswald),
it must not be forgotten that Van Es, in publishing so many new
data, considerably enriched our knowledge of these marine sedi-
ments.

I should like to draw attention to some more merits. On the
occasion of the 70th birthday of Professor K. Martin a jubilee
book was published 1), which is a sort of reasoned fossil catalogue
of the endre East and West Dutch Indies. This work has been
brought about by cooparation of a number of Dutch and some
foreign investigators. The Mammalia have been dealt with by the
present writer. The various localities, which I found mentioned
in the papers relaring to the subject, were united in a small sketch
map 2). If that map be compared with a similar map, occurring in
Van Es\'s paper, it will be seen that also in this respect Van Es
collected many new data.

Noteworthy furthermore is the way in which Van Es discussed
the problem of the age of the Trinil beds. As will be known, the
number of publications, dealing with this problem, is considerable,
and the number of opinions is hardly less large. Consequently
difficulties are met, if one tries to form a definite opinion from the
chaos of assertions and opinions. Van Es, however, had the original
idea to class the arguments, advanced by the various authors, with
eleven different headings viz., i. Orogenic movements, 2. Vul-
canism, 3. River terraces, 4. Culture remains, 5. The anatomical
features of Pithecanthropus erectus, 6. The process of fossilization,
7. Marine molluscs, 8. Fresh-water molluscs, 9. Plant remains,
10. Vertebrates, 11. Chmate. In this way an excellent synopsis
originated..

After this expariation, which appeared to us as wholly justified,
we return to our starting-point. We shall begin with the localities
Watualang, Pitu and Kedung Kendang, all situated near the Solo
river 3). It will be known that the famous localities Trinil and Sonde
are also situated on that river.

Van Es\'s publication contains a geological map of the Solo
river region between Gesi (N. of Sragen) and Ngawi. According

to the text this map is accompanied by one section. This is,
however, an erroneous statement, as the section is not present.
It will, however, be found in a geological guide of the vicinity of
Trinil, made by Van Es for the use of the participants in the fourth
Pacific Science Congress.

Van Es\'s map shows clearly that ■—roughly spoken ■—the Solo
river follows the boundary between the vertebrate bearing series
and the older beds. In consequence of numerous meanders the river
intersects both repeatedly. According to Van Es the following
stratigraphic succession occurs:

Conglomeratic sandstones, conglomerates, sandstones, tuffs, and black day. This
very variable succession represents the main Vertebrate zone. According to Van
Es\'s map, it is also this zone, which yields the vertebrate remains, found near Pitu,
Watualang, and Kedung Kendang.

b.nbsp;Volcanic boulder breccia.
In the region between Ngawi and Sonde. Generally underlying the main Vertebrate
zone, and corresponding to a similar horizon, E. of Ngawi, which is very persistent.

c.nbsp;Sand and conglomerate containing older material.
Locally directly covering the pliocene limestone. Containing fossil bones.

y k Argillaceous sands. Thickness but 50 m. Only locally exposed. In Sonde with rich
.0 ^ fauna of marine molluscs. On 150 different species 53 % recent forms (Martin 1919).
The apparent disappearance of Sonde beds E. of Trinil is ascribed by Van Es to an
unconformity existing between the Pliocene and overlying pleistocene beds.

i.nbsp;Alternating tuffs and marls, volcanic breccia and limestone.
The limestone contains species of Lepidocyclina and Miogypsina proving the beds
to correspond to Tertiary ƒ of Van der Vlerk and Umbgrove.

a
p

In chapter VIII of Van Es\'s paper we can read on what grounds
Van Es regards the Trinil beds to be of pleistocene age. His line
of reasoning is as follows. Field work resulted in the establishment
of a stratigraphic gap between Trinil beds and the in Trinil almost
directly underlying Sonde beds. In Bumiaju this hiatus does not
occur; at that place the vertebrate bearing series lies conformably
on the marine sediments. The latter correspond in age to the Sonde

-a

beds, each showing a percentage of recent species between 50 and
60 %. (As already mentioned this percentage surely proves the
beds to be of pliocene age; their supposed belonging by Van Es
to the Middle Pliocene, however, remaining to be solved). The
stratigraphic gap which occurs in Trinil, corresponds to the verte-
brate beds of Bumiaju. In Java, therefore, an upper pliocene and
a lower pleistocene fauna of vertebrates occur. The vertebrate
beds of Bumiaju represent, however, only one facies of the Upper
Pliocene. No less than four different facies were found, of which
the marine facies, discovered N. of Djombang, is of great impor-
tance of course. For sediments of that facies Von Koenigswald
found 66 % recent forms on 71 species determined (Sumberringin
layers 2 and 3), while Martin established a percentage of 70 %
on 50 species determined (fauna of Kedungwaru).

ever a marine fauna of the age of the Trinil beds is discovered in Java,
it will show to contain more than 70 % living species of Molluscsquot;
This high percentage, wliich is to be expected, would fill the gap
between the Upper Pliocene and the post pleistocene beds of
Grissee (90 % on 30 species) and Batavia (86 % on 22 species).

A totally different facies of the quot;Bumiaju bedsquot; is met with in
Surakarta (near Sangiran and Baringinan), where freshwater lake-
deposits occur. In these beds 16 different forms have been found,
9 of which have been determined hitherto. As only 3 species appeared
to be recent, the percentage of living species can never exceed
62^/2 %, whatever the result of the determination of the remaining
forms may be. Martin-Icke and Martin determined the fresh-
water molluscs, occurring in the Trinil beds: 83 % appeared to
be recent species. The Trinil beds therefore are decidedly younger
than the fresh-water deposits in Surakarta. Accordingly Van Es
classed the Trinil beds with the Pleistocene. As furthermore the
Trinil beds (i) underlie a mighty complex of volcanic sediments
(Notopuro beds) in the vicinity of Pandan and (2) occur in the basal
part of a complex of at least 500 m. thickness in the Kendeng Hills N.
of Djombang, he inferred that they occupy a rather low horizon of
the Pleistocene. Van Es regards the presence of Mastodon? sp. \\
Stegodon and Hippopotamus as another indication in that direction!

For the present the above will suffice. In our final chapter we
shall return to the problem of the age of the Trinil beds.

2)nbsp;See Van Es (I.e.), p. 117. Van Es borrows these particulars from a report by Martin
not yet published. It may be emphasized that Martin concludes to a pliocene age.

Before passing on to the consideration of the other localities,
it will be desirable to bring forward the following particulars con-
cerning the Trinil beds between Gesi and Ngawi.

According to Van Es the bones seldom occur as abundantly
as in Trinil. The bone beds of a. o. Watualang and Kedung Kendang
are mentioned as deposits, that might compete with those of Trinil.
Van Es protests against the opinion of some, that the origin of the
bone beds should be a result of the destruction of the existing fauna
by volcanic eruptions. He points out that in many cases the bones
were found in cross-stratified sandstones, containing rounded
pebbles indicating true river deposits. In his opinion the animals
died through natural causes; they were swept by flooded rivers to
some wirl-pool bend where they sank, or to sandy banks where
they finally decomposed. In several cases the bones were already
broken and weathered, before they were hurried in the sand. Cro-
codiles often caused accumulation of the bones.

quot;In other cases the bone-bearing bed consists of black clay,
containing fresh-water Molluscs and remains of fishes, crocodiles
and turtles i). This black clay has been formed in stagnant pools
and marshes or even in big lakes. Sometimes nearly complete skele-
tons of larger Vertebrates occur, owing to the fact that marshes
often form the dwelling place of big animals.quot;

That river accumulation, not volcanic activity, was predominant
is — according to Van Es ■— proved by the fact that the bone
beds contain detritus of miocene strata. It is these detrital products
which Van Es holds responsible for the sohdification of sandstones
and conglomerates.

Then Van Es discusses the opinion, that the bones should
have been washed off from the older beds. Admitting the prepon-
derant influencé of erosion, this possibihty may not be immediately
excluded. Many bones, however, do not show traces of wear.
Furthermore, he rightly points out that Stegodon tusks and crania
with the horn-cores attached, are too brittle in a fossil state to allow
of any transport. Moreover, bigger bones and skulls often occur in
medium-grained sand, containing only small pebbles, whilst big
boulders are absent. Transport of the bones in a fossil state is in
such cases highly improbable, as the specific weight increases by
the process of fossilization.

As to the tectonic structure of the Trinil beds he mentions that
the general dip is 6—io° S., showing the influence of tilting or
folding movements. Several transverse faults show, moreover,
that the beds did not remain undisturbed.

We think it entirely superfluous to occupy ourselves with the
geology of the surroundings of Redjuno. From that locality, namely,
we are able to mention only one specimen viz., a proboscidean
humerus, which did not even allow of determining the genus with
sufficient rehability. Those who are interested in the stratigraphy
and tectonic structure of the region, I may refer to Van Es\'s
publication.

As to the localities, which are situated in the districts Randu-
blatung and Tambakredjo, we are forced to be brief, because Van
Es did not make detailed investigations in these regions. On p. 19
Van Es mentions that he discovered in 1927 an occurrence S. of
Randublatung, which appeared to be very rich in vertebrates. As
Kuwung is situated SE. of Randublatung, it is highly possible that
our specimens from Randublatung have been obtained from that
locality. quot;The beds mainly consist of sandstone and gravel of
volcanic origin and overlie the Miocene hills. There is a very pro-
nounced unconformity between the slightly N.-dipping vertebrate
beds and the steeply folded Miocene marls.quot;

Still on p. 19 he cursorily deals with the localities in the neigh-
bourhood of Tinggang. In 1926—1927 he collected there a large
number of veriiebrate remains from gravel beds, mostly containing
pebbles of volcanic origin. quot;Owing to the bad exposures no data
were obtainable concerning the relations to the underlying Tertiary
beds. In the vertebrate beds in several spots a slight dip to the N.
not exceeding 5° was established.quot;

It may be mentioned that it is highly probable that the remains
of our collection derived from these localities have been collected
by Van Es, the label of numerous specimens of the second sending
bearing the mention: quot;Collection Van Esquot;,

We shall now drop the subject of the consideration of the loca-
lities, and pass on to a discussion of the few Mastodorf\' remains
which were hitherto found in the Dutch East Indies.

Martin described and figured under the name of Mastodon
sp. the posterior portion of a grinding tooth, and the distal portion
of an incisive tusk. Dubois 2) did not agree with the generic deter-
mination. According to this investigator both specimens might
rather belong to Stegodon. The fragment of the incisor shows a
distinct flattening as a result of wear. Dubois rightly pointed out
that this feature is not an exclusive character of the tusk of masto-
donts.

Samml. Gcol. Rcichs-Mus. Leiden, IV, 1888, p. 90, pi. XI, figs, i—2a.
Nat. Tijdschr. v. Ned.-Ind., LI, 1892, p. 95.

Lydekkeri) recorded from British Borneo a 1. M^ oi Mastodon
latidens Clift. 2) The determination is certainly correct. At present
tliis form is reckoned to the new genus Stegolophodon Schlesinger
(Prostegodon AIatsumoto), which Osborn reckons to the sub-
family of the Stegodontinae. But as far as I can see, this is merely a
matter of taste.

Van Es3) gave under the name of Mastodon\'^ sp. two fine
figures of a fragmentary grinding tooth, obtained from Sangiran,
and from beds, which he regards to be of lower pleistocene age.
In a note on p. 54 he mentions, furthermore, that Dubois showed
him part of a similar molar, collected from the Kendeng Hills.
In the latter\'s opinion these specimens might represent an atavistical
deviation of a Stegodon molar. The writer should not like to en-
dorse this statement. Be that as it may. Van Es\'s specimen does not
in the least resemble the specimens obtained from Bumiaju. As to
the latter, these were originally determined as belonging to Masto-
don longirostris by Stehlin 4). As already mentioned, originally
the writer identified the form of Bumiaju provisionally with
Mastodon perimensis.

Some forms of our collection appeared to be specifically identical
with still living species (kerabau,banting, Javan rhinoceros andlndian
elephant). In this connection I should like to make some remarks.

It is clear that everyone, who has to occupy himself with the
examination of a relatively young fauna, will make comparisons
with the recent fauna. In doing so, it is of the first importance that
not one or some specimens of the recent species be used, but the
greatest number possible. Stremme — in determining the greater
part of the mammalian remains of the Trinil collection of Mrs.
Selenka — neglected this requirement, and it played tricks on him.
As a matter of fact the German museums of Natural History will
not contain so great a number of Dutch East Indian specimens as
the Dutch museums do. I have not got the impression, however,
that Stremme troubled himself sufficiently. Moreover, he apparently
did not feel fully the seriousness of the requirement mentioned.
Otherwise he would not have ventured to draw such resolute in-
ferences by the help of so small a material for comparison.

As mentioned, the writer was in a far better position. And he
has made an eager use of the opportunity afforded. Not because
measuring skulls and composing tables of measurements is a plea-

sant occupation! On tlie contrary. Tiiis declares in my opinion
why in this respect a lot of work remains to be done. The necessity
to have the disposition of tables, which show in numbers the indivi-
dual variation of the cranium of recent species, may appear from the
fact that Franz Toula\'s tables of cranial measurements of
Dicerorhinus sumatrensis are repeatedly consulted, in spite of the fact
that they contain several miscalculations.

Incidently it may be pointed out that in studying the measure-
ments of the crania of recent forms of Hippopotamus we arrived
at an uaexpected result. Hippopotamus constrictus Miller, namely,
appeared to have no right to specific distinction, being identical
with H amphihius. One more proof of the correctness of the assertion
above mentioned.

As to the measuring of the skulls and the making of the tables
of measurements the greatest possible accuracy has been practised.
Repeatedly it has occurred that measurements, which distinguished
themselves either by a very high or low value, have again been
taken. The percent numbers in general have been obtained by the
help of an electric calculating machine. All the percent numbers
have twice been checked. Nevertheless I do not imagine my tables
to be without errors. This is practically an impossibility with a
material of several thousands of numbers. Also I fully realise that
the material measured, does not suffice to serve as a base for far-
reaching conclusions. It will be known that some anthropologists
are not wholly satisfied before they dispose of the measurements
of some 2000 crania.

Of course it has been tried to obtain from Prof. Dubois the per-
mission of comparing our specimens with those of his famous col-
lection. Prof. Dubois wrote us, however, that he regretted much that
he was forced to refuse our request, because his collection was not
in a state for demonstration in consequence of its repeated removals.

Most specimens described are strongly fossilized. This especially
concerns the specimens from Bumiaju. As to the degree of fossili-
sation of the remaining, some require special mention. The three
crania of Bihos sondaicus fossilis., and some specimens of Buffelus
huhalus ?var. sondaicus fossilis are but little fossilized. Accordingly
it was necessary to harden them. Contrary to the remaining speci-
mens, the specific weight of the specimens mentioned is low. Also
they do not show the gray colour of the specimens from Bumiaju,
but have a brown appearance. One specimen of our collection is
extremely little fossilized. It is the posterior portion of a cranium
of Hippopotamus sp., and shows a remarkable recent appearance.

It appeared that the photographs of the proboscidean grinding
teeth in general considerably gained in clearness, if the dentine
material of the teeth was blackened previously. It will be seen that
most specimens have been treated in that way.

Last not least the following remark. The present paper is the
work of a beginner. The writer has sufficiently realized that pa-
aeontolog^al science will profit more by very accurate descrip-
tions than by phylogenetical speculations of beginners. Consequently
very much attention is given to the former. Especially in describing
the proboscidean grinding teeth, the writer started from^the sup-
position that it is better to give too much than too little.

Mr. A C. de Jong, mining engineer. Intendant of the Geological
Survey (Opsporingsdienst) of the Dutch East Indies, to whose
mtermedmry I owe that this fine collection was sent to me for
examination.

Prof. Dr. L. M. R. Rutten, who had the supervision of this
work, and who, though not being a vertebrate palaeontologist
saved me by his clear criticism from faulty reasoning.nbsp;\'

Prof Dr. H. F. Nierstrasz, Director of the Zoological In-
stitute of Utrecht, to whose warm recommendation I owe for a
good deal that the Geological Survey entrusted me with the deter-
mmation, for his unremitting attention in my work and for permitting
me to measure some fine crania, contained in his Institute.

Prof Dr Max Weber for several important informations, and
tor his landly lending me many separates of his collection.

Prof. Dr. E. D. van Oort, director of the National Museum of
Natural History in Leiden for access to his museum, and the
permission to measure and photograph specimens under his charge.

V^Tefr\' ^he opportunity of measuring and photographing
the specimens contained in the various museums under his charge

Mr. P. Th. N. Moesveld for revision of the text,
suppirt.^^^ ^nbsp;^^^^ her real and moral

Fam. BO VI DAE.
Subfam. BOVINAE.
Buffelus bubalus (L.) ?var. sondaicus (Schl. et Müll.) fossilis.

Localities: Sentang Kedung Klampo, Bondol near Kuwung, Tegal Sambiduwur,
Tegaron, Pitu, Watualang, Kedung Kendang.

Buffalo remains are represented in our collection by eleven
fragmentary crania, and two detached horn-cores. They have been
obtained from seven different localities.

Unfortunately all the specimens of crania (n^s. i—n) are very
incomplete; some only consist of the fronto-parietal region with
a short portion of one or both horn-cores. Most are broken off,
either in front of the orbit or immediately behind it. In this respect
two specimens are somewhat better preserved, the one (n». 2)
showing the hinder half of the nasals, the other (n». 3) still having
the posterior portion of the left maxillary with some molar teeth.
One specimen (n^. 2) possesses a zygomatic arch. In only one
specimen (n^. 3) the orbit is tolerably well preserved. Two crania
(nos. 10 and 11) are in too bad a condition to allow of measuring.

All the specimens show a smaller or greater portion of one or
both horn-cores. Two crania, one of which is figured in figs. 3
and 4 of pi. I, are in this respect the best preserved, one having

the nearly complete right horn-core, the other being in the possession
of both horn-cores. But even of these a smaller or greater portion
of the tip IS wanting. Nevertheless, the specimens represented
fortunately enable us to state the curvature and course of the
horn-core of this fossil buffalo, to determine the relation of the
complete core to the fronto-occipital region of the cranium, and
to realize the enormous span of the horn-cores.

As far as can be gathered from these fragmentary crania they
al seem to have belonged to adult individuals of the same species.
I think therefore, it will be best to give a summarized descr^tion
ot all the specimens.

Before commencing with the descriptive part it will be desirable
to lay stress upon the fact that — in describing the crania in
question — these were supposed to be placed on the anterior
premolar teeth and the paroccipi\'tal processes.

The occipital surface is in a splendid state of preservation in
^ecimen no. i, two views of which are given in figs, i and 2 of pi I
The occipital surface is divided by the occipital crest into two
semi-distinct portions, a smooth supra-cristal portion, and a rough
infra-cristal portion (i. e. the true occiput) for muscular attachments
The occip^al crest is a wide arch, the summit of which is mostly
flattened. This crest is sometimes so prominent that it makes the
impression as if the upper half of the true occipital surface has
sunk into the strong frame of the crest. Above this crest the
temporal fossae, which terminate in blunt, round extremities cut
into the occipital surface. The degree of constriction of the occiput
will be dealt with below. In specimens i and 7 the course of the
posterior part of the temporal fossa is the same as in the cranium
of the kerabau. In specimens 2, 3, 5, 6 and 8, however, a peculiarity
occurs. In these specimens,namely, the posterior parts of the temporal
tossae hrst converge normally backwards; near the occiput how-
ever they change their direction, strongly diverging backwardly.
10 this pecuharity we shall have to return later on. The true
occipital surface is much more broad than high, the interval between
tiie upper margin of the for. magnum and occipital crest being
TrZ Inbsp;\'\'nbsp;contained in the greatest breadth of the

curve tJ supra-cristal portion joins the frontals by a gentle
wkr;h f \'\' \'\'\'\'\' \'^^Shtest trace of an intercornual ridge
Tf Lehe^dnbsp;^^nbsp;^^ ^^hich the transition

theless it is not difficult to state that the interval between occipital
crest and for. magnum is nearly equal to the interval between
occipital crest and the middle of the curve formed by the gradual
transition of forehead to supra-cristal portion of occiput. From
the middle of the occipital crest straight downwards runs the
vertical crest, affording attachment to the ligamentum nuchae
between the occipital muscles. In some specimens this crest is
ill-defined, in others very distinctly marked. The occipital condyles
are obliquely placed; the for. magnum is subcircular. Its upper
border, complete but in few specimens, is either convex or has
the form of a flat reversed V. The paroccipital processes are thick,
short, with blunt extremities. They are curved backwards and
inwards.

The basi-occipital has a relatively great width and its posterior
tuberosities are very well developed. The position of the post, nares
and of the various foramina on the inferior aspect of the cranium
does not differ from those of the living kerabau.

If we now continue to view the upper aspect of the cranium
we notice that the frontals are nearly flat or slightly convex across,
and also slightly convex from front backwards. Immediately in
front of the base of the horn-cores the forehead is contracted,
but not much, which is mainly caused by the fact that a strong
ridge runs from the middle of the posterior border of the orbit
to the antero-inferior angle of the horn-core. The upper border
of the orbit is — at least in the few specimens in which the orbit
is preserved — considerably oblique to the longitudinal axis of
the cranium. The orbit seems to be subcircular; in the only specimen,
however, in which the orbit is rather completely preserved, it is to
a great extent restored with plaster, so that its proper form cannot
be clearly seen. The orbits are closely approximated to the cores.

The supraorbital foramina are in general ill-preserved and often
covered and filled with matrix, which could not be removed. There
is, however, one specimen (n». 4), in which the left supraorbital
sulcus and foramen are finely preserved. As, moreover, the specimen
in question is in the possession of the posterior extremity of the
nasals, and of a part of the upper border of the orbit, we are able to
give from this specimen the following particulars.

The supraorbital foramen is large and situated at the height of
the posterior border of the orbit. The foramen itself is directed
obliquely backwards and slightly downwards, the posterior border
overhanging it. The supraorbital sulcus leads immediately into
the foramen. Its posterior portion is deep and distinctly marked.
More frontwards, however, the sulcus gradually dies out, so that
its total length cannot be given. In the specimen under consider-

From the imperfection of the anterior portion of the face it is im-
possible to deter^ne the relations of praemaxillaries and the anterior
half of nasals and maxi lanes mutually, and to the posterior portion
of the face. Only can be stated that the nasals are clearly Lvex
across between the supraorbital sulci, and nearly straight longi-
tudmally. In profile view of the cranium the nasals are Ituatedin
the elongation of the anterior half of the frontals. In front of the
orbits the contraction of the face is considerable

the core by a rather sudden decreasing of height and width.
If further on will be spoken of the quot;horn-corequot; this will always

in cross-section. The cross-
section of the base of the core
has a form represented in text
fig. I. The superior surface is
flat; in some instances even

1. Buffelus bubaius?w2.r. sondaicus fossilis. ^^^^^^^ ^s remarkably sharp;
Cross-section of left horn-core

mnn^.^ J 1 , 1 . . inferior surface much more
rounded, and blunted. Anterior and inferior border merge very
gradually into one another by a gentle curve.nbsp;^

The cores, which are of very large size and thickness, are set
obliquely to the longitudinal axis of the cranium, and are con-
siderably curved backwards in one regular curve.

In the posterior aspect of the cranium they slope slightly down-
wards from base to tip.

Both from the description and the figures it will have appeared
that the fossil form under consideration must belong to the genus
Bujfelus in its narrowest sense. Now there are only known the
following Asiatic forms of Buffelus s. str.

Bubalus platyceros, R. Lydekker, Ree. Geol. Survey India, X, 1877, p. 31; Mem. Geol.

Survey India, ser. X, I, 1878, p. 127, pi. XVIII; 1880, p. 173; L. Rütimeyer, Abb.

This species was founded by Lydekker in 1877 on the evidence
of one fragmentary cranium from the Siwaliks; in 1878 he published
an elaborate description of the type specimen, illustrated by drawings
of frontal and occipital aspect, and of a cross-section of the horn-
core^). Mention was, furthermore, made of a detached horn-core,
and of a cast of a cranium of the species, also from the Siwaliks,
the original of which Lydekker presumed to be in the British
Museum. This indeed was true, being the fragmentary specimen
on which Rütimeyer (1878) founded his Buffelus sivalensis. The
second part of Rütimeyer\'s memoir quot;Die Rinder der Tertiär
Epoche, etc.quot;, 1878, in which part the cranium of Buffelus sivalensis
Rütim. was described 3), was published after Lydekker\'s memoir

We may neglect the two fossil Chinese buffalo spccies which Koken (Palaeont. Abh. Ill,
H. 2, 1885, p. 67, Taf. II, figs. 14 and 20; p. 68, Taf.II.figs. 15 and 21) recognized among
the detached teeth of bovines in his collection, but which he left specifically undeter-
mined. Matsumoto (Sci. Rep. Tôhoku Imp. Univ., Sendai, Japan, 2nd ser., [Geology],
ill, I9I5-\'i8) referred the detached teeth and lower jaws of bovines of his collection
from Sze-chuan (China) also to two different species. The one, Buffelus sp. a, (p. 19, pi. IX,
figs. 4 and 5) was considered as possibly identical with Buffelus palaeindicus (Falconer);
the other Buffelus sp. b, (p. 20, pi. VIII, figs. 1-5) as perhaps identical with that species
of Koken, which the latter described on p. 67. Also of these forms the above mentioned
may suffice.

A partly restored view of the type specimen was given by 1,ydekker in 1898.
See figure in the first part (1877).

Buffelus bubalus (L.) var. pa 1 aeindicus (Falc).
p\'zio-, p.!. M«m. I, .868, p. 284 pi.nbsp;««quot;„i, \'iesch. des Rindes. Zwe,.ct

Lydekker 1878, p. 129. See also Rütimeyer 1878. P-
3) Concerning this figure Rütimeyer (1878,nbsp;__ ^eines-

ht,^bschon nach einem Entwurf von Falconer s ^and doch dem ^ ^^
gehalten. Der Schädel von Bubalus palaewd.cus ist in beiden Geschlecnt

In 1878 our knowledge of this form was considerably enriched
by publications both of Lydekker and Rütimeyer. Lydekker
fully described and figured a cranium (frontal and occipital aspect)
from the Narbada valley, while mention was made and dimensions
were given of some other specimens, two of which had been
found in the topmost clay beds of the Siwaliks near Bubhor (Punjab).
Besides this locality Lydekker mentioned the following locaHties,
from which remains of this fossil buffalo were obtained: greater
portion of the central Narbada Valley, Jamna Valley, Godavari
and Perim-Ganga Valleys, and Madras (?) i).

Rütimeyer (1878) discussed four crania in the British Museum,
all obtained from the Narbada Valley. Again he laid stress upon
the very close relationship between the fossil form and the living
ami, and doubted whether the fossil form had a sufficient right
to specific distinction. He maintained, however. Falconer\'s

zweitens eine weitere Verständigung über die fossile Form des
Arni nicht durch Auslöschen eines Namens zu erschwerenquot; 2).
In his review of Lydekker\'s work he is wholly convinced of the
specific identity of Lydekker\'s specimens with those in the British
Museum.

Contrary to Rütimeyer, Lydekker (1878) considered the
differences between the fossil Indian form and the living arni
important enough to distinguish between the two, so that Falconer\'s
name was maintained unchanged. In 1885 Lydekker had partly
revised his opinion stating that the fossil Indian form cannot be
but regarded as more than a large variety of the living buffalo
{Buffelus bubalus. L.). In his book quot;Wild oxen, sheep and goats of
all landsquot;, 1898, he stuck to his opinion, while the Punjab specimen
was regarded as probably belonging to the same race. In that
work we can also find the following summary of the characters:
quot;Very close to the typical race, but of larger dimensions, with a
more convex forehead, and the horns apparently always directed
to a great extent outwardly. In the horn-cores themselves the
transverse section is also somewhat different, tending to become
quadrangular, instead of being strictly triangular.quot;

We shall return to this species further on. Already now, however,
I may call attention to the fact that the specimen which Rütimeyer
mentioned in 1868 was said to be distinguished from the arni
a. o. by a quot;breitere und flachere Stirnquot;, whereas Lydekker (1898)

1) In 1898 Lydekker mentioned as localities, outside the Siwalik Hills: Narbada-, Godavari-
and Pern Ganga Valley.

mentions as one of the distinguishing characters the quot;more convex
foreheadquot; of the present variety. This proves — in my opinion —
that both forms of forehead may occur in the fossil race, which
in no wise surprises me as I could state exactly the same in about
thirty crania of the recent Javanese kerabau.

Bubalus palaeokerabau, E. Dubois, Tijdschr. Kon. Nederl. Aardr. Gen., ser. 2, XXV,
1908, p. 1265.

non: Buffelus palaeokerabau, H. Stremme, Pithecanthropus-Schichten, 1911, p. 124,
pi. XVIII, figs. 5—6; pi. XIX, fig. 8; pi. XX, figs. 7—9, 12—14.

This fossil Javanese form was originally considered by Dubois
(1892) as identical with Buffelus palaeindkus, which form he regarded
as the ancestor of the kerabau and specifically identical with the
latter. In 1908 he had changed his opinion as may appear from
the following quotation: quot;Den BufFel des Kendeng hielt ich früher
für Bubalus palaeindicus. Der fossilen javanischen Art kommt aber
weder die längliche Form des Schädels der Narbada-Art, noch
die fast gerade Streckung und quere Richtung und der nahezu
vierkantige Querschnitt von deren Hornzapfen zu. Die Schädelform
ist kurz wie diejenige der lebenden javanischen Art, und an den
Hornzapfen ist die frontale Fläche eben, mit scharfen Unten- und
scharfen Obenrand, während ihre orbitale und ihre temporale
Fläche abgerundet sind; und indem sie auch durch eine sehr stumpfe
Kante sich von einander abgrenzen, kann der Durchschnitt des
Hornzapfes von dreieckig fast halbkreisförmig werden. In der
Form des Hornzapfes nähert der KendengbufFel sich etwas dem
siwalischen B. platyceros und unterscheidet sich von der lebenden
Kerabauart, übrigens sind die Schädel der beiden javanischen
Arten einander sehr ähnlich; die lebende stammt wahrscheinlich
von der fossilen Art ab, wie der Name Bubalus palaeokerabau n. sp.
andeuten soll.quot; i)

In what precedes all has been quoted that Dubois hitherto
has written about this new species. Figures have not been given.

The collcction of the Trinil-expedition of Mrs. Selenka con-
tained buffalo remains, which have been described by Stremme.
Three crania have been found, one of which but relatively little
damaged. Especially of this cranium Stremme gave a detailed
description, together with measurements and figures. Stremme
too arrived at the conclusion that the relationship between the
fossil Javanese form of the German Trinil collection and the recent
kerabau is a very close one. He noticed, however, some differences

by which the fossil form seemed to be distinguished viz., . . durch
einen längeren Gesichtsteil im Verhältnis 2:um Hirnteil, durch
eine stärker ansteigende Stirn; durch deren Crista und Furchen,
.... durch schmalere Prämaxillen . . . .quot; Besides, he did not see
hitherto a specimen of kerabau of the same size. He did, however,
not exclude the possibility that more ample materials of the recent
kerabau, than he had at his disposal, should reveal that these
differences in reality did not exist.

Stremme united in a table the cranial measurements of his
fossil form, of the kerabau, ami, quot;Buffelus palaeindicus\'\' and Buffelus
platyceros \\ he calculated, moreover, a number of relations, and
drew some more inferences from them. At last he arrived at the
conclusion that his fossil form belonged without doubt to Dubois\'s
B. palaeokerabau.

As a matter of fact, in determining the fossil form of our
collection, we shall still have to pay constant attention to Stremme\'s
work. I should like, however, to observe already now that it will
have appeared from what precedes that Dubois mentioned only
one difference between Buffelus palaeokerabau and the living kerabau
viz., a different form of the horn-core in cross-section. Stremme
enumerated also a number of differences between the fossil form
and the kerabau. He did not notice, however, any difference in
form of the horn-core in cross-section. And nevertheless he con-
sidered his form as belonging without doubt to Buffelus palaeokerabau.
In my opinion the possibility that Dubois\'s and\'Stremme\'s form
of buffalo are identical is great, both authors emphasizing the
very close relationship to the kerabau. On the other hand, however,
Stremme\'s argumentation cannot be regarded as sufficient to make
the identity probable enough. To prove it, will be, of course, an
impossibility in consequence of the absence of any figure of Dubois\'s
specimens.

Bubalus sondaicus, L. Rütimeyer, Vers, einer natürl. Gesch. des Rindes, Zweiter Teil,

BubS fndicu/v^ar.^\'fontiaica, L. Rütimeyer, Vers, einer natürl. Gesch. des Rindes,

Bos bubalus var. sondaica, Schlegel and S. Müller, Verh. over de nat gesck der Neder .

For more extensive synonymy see: R. Lydekker, Wild oxen, sheep and goats, 1898, p. 118.

Of this wide spread recent species there are two races, which
deserve our interest viz., the wild buffalo of India, the ami, and
the buffalo of the Malayan Islands, the kerabau. Of the form first

mentioned an extraordinary fine cranium of an adult individual
is contained in the National Museum of Natural History at Leyden.
Its measurements will be found in table B of cranial measurements,
together with those of two more crania of the arni, borrowed
from Lydekker (1878) and Stremme (191 i). The Leyden Museum
is, furthermore, in the possession of a buffalo cranium, labelled quot;Bos
bubalus, Arnr, while in the collection of quot;Natura Artis Magistraquot;
at Amsterdam two skulls are preserved, which also are regarded
as specimens of the arni. These three crania, however, are deviating
in many points from the skull of the arni, so that I strongly doubt
the correctness of the determination. That is also the reason why
the measurements of these specimens have not been given. Above
we observed that both Dubois and Stremme arrived at the conclusion
that in Java a fossil form of buffalo occurred, closely related to
the kerabau. As this race, therefore, is of the uttermost importance,
it will be dealt with extensively in the sequel.

Often the kerabau will be found named as quot;Buffelus {Bubalus)
sondaicus Rütimeyerquot;. And indeed, Rütimeyer (i 868, p. 3 8) figured i)
an occipital aspect of the cranium of a male kerabau with the sub-
scription quot;Bubalus sondaicus\'\'. On p. 173, however, we can read
in the explanation of plates and text figures that fig. 4 in the text
represents quot;Bubalus indicus, var. sondaicaquot;. And it was Schlegel
and Müller (1839—P- 205), who for the first time distinguished
the present form as a variety (var. sondaicd) of Buffelus bubalus
(syn. Bos bubalus, Bubalus indicus). From Rütimeyer\'s text it appears
that he wholly agreed with Müller and Schlegel. Those who
are sufficiently acquainted with the studies of Rütimeyer on the
Bovidae, will have noticed that this author, who knew this family
as nobody else did, drew the limits of the quot;speciesquot; very widely.
In my opinion, therefore, the subscription of text fig. 4 must be
considered as an error, and the kerabau must be called Buffelus
bubalus (L.) var. sondaicus (Schlegel et Müller).

It will be known that the kerabau occurs in abundance as a
domesticated animal in the Malayan Islands. It is, furthermore,
a fact that here and there, e. g. in the uttermost SE. and SW.
of Java 2), the kerabau is met with in a wild state. Whether, however,
they reverted to this state from domesticity or whether they are
indigenous inhabitants of the island is a moot point. The first
supposition is generally accepted, but it has certainly not been
proved. Stremme (1911) stated as his opinion that the occurrence
of the fossil B. palaeokerabau in Java made it probable that the

kerabau — as Cuvier already believed — belonged to the original
fauna of the island. Merkens (1927) too, doubted — on histo-
rical grounds — the original domesticity of all wild specimens of
kerabau.

Stremme, in determining his fossil form as B. palaeokerabau,
had for direct comparison at his disposal but one cranium of kerabau.
In this respect the present writer was in a far better position, the
musea of quot;Natura Artis Magistraquot; containing not less than twenty
six crania, and the National Museum of Nat. Hist, six specimens.

The being in use of M^ was taken as a criterium of adultness.
That I was right in doing so, is proved by what follows. Merkens
(1927), on p. 106 mentions that the fourth pair of permanent
teeth, that is to say the lower canini, protrude, when the animal
is 51/2 to 6 years old. My friend Mr. P. van Rijn, veterinary surgeon,
was so kind as to make an investigation of the relationship between
the protruding of the C and the entering in use of the permanent
upper molars. In one of the musea of quot;Natura Artis Magistraquot;
he found a cranium with the undoubtedly corresponding lower
jaw. In this specimen the C was already in use, whereas the M»
was still untouched by wear. This single example would, of course,
not afford great evidence, were it not that the length of the horn cores
proved in a convincing manner the adultness of all the specimens,
of which measurements are given in table B. From this table appears,
first, that the shortest horn-core, which was measured, has a length
of 273 mm. and secondly that in those specimens in which the
horns were still present, and in which they could be removed, the
difference in length between horn and horn-core is very con-
siderable. From Merkens (1927) we derive, furthermore, the
following data. At the age of one year the length of the horn is
about 50 mm., in the second year 100 mm., in the third 150 mm.
As will be seen from table B, the length of the horn of the specimen
with the shortest horn-core is 460 mm. Even if the rapidity
of the growth of the horns should increase after the third year
of the animal — which appears to me as not very probable —
we may be sure that a length of the horn of 460 mm. will not be
reached before the seventh year. And — as we have seen above —
the kerabau reaches the adult state in his sixth year.

Of the thirty-two crania two specimens had to be excluded
as being not fully grown. Two specimens, furthermore, showed
such strong deviations from the normal type, that I esteemed it
advisable to exclude these specimens too. There remained, therefore,
the measurements of twenty-eight specimens, to which have been
added the measurements of three more crania viz., those of one
specimen given by Stremme, and those of two frontlets of enormous

size, contained in the Buitenzorg Museum Of course it would
have been better, if we had been able to compare our fossil form
with crania of the wild kerabau. But even in the Dutch East Indies
it will be very difficult to obtain a sufficient number of crania
of the wild kerabau. We may be sure, therefore, that most, if not
all the crania, which have been measured by the writer, belonged
to domesticated animals.

Fortunately for us, however, the kerabau does not interbreed
with the Indian or European cattle which excludes the possibility
of crania of hybrids being among the specimens measured. Castration
of the males, however, occurs on a rather large scale so that
we may be sure, that among the twenty-eight specimens — besides
males and females — a number of oxen will be represented. Three
specimens from Leyden are, according to the label, males. But
in the first place it is in general desirable to accept such data of
museum specimens with some degree of reserve, and in the second
place these specimens may represent bulls- as well as oxen.

Originally it was thought that our table should considerably
gain in value, if we should be able to exclude the oxen, and if we could
distinguish between the crania of cows and bulls. Merkens (1927)
tried to detect cranial differences between cows, bulls and oxen.
For that purpose he took twenty-one measurements of four adult
specimens of each; expressed each measurement in % of the total
length of the cranium, and calculated the average of each measure-
ment for cows, bulls and oxen. After comparison of these averages
he arrived at the conclusion that in some respects differences exist
between the crania of cows, bulls and oxen. Though everyone
will be inclined to distrust conclusions based on such few materials,
it cannot be denied — after examination of all the numbers given
by Merkens — that in general his conclusions seem to be not
devoid of foundation. One conclusion, however, must be rejected
as being totally wrong. The average length in mm. of the cranium
of cows, bulls and oxen appeared to be respectively 480, 479 and
489. Merkens considered a difference of i mm. too small to be of
any value regarding the small number of specimens measured ■
He did, however, attach value to a difference of 10 mm. Now it

Aly sinccrc thanks are due to Dr. K. W. Dammermann who was so kind as to
send me — at my request — the measurements of these specimens. The one
(Buit. Mus. n°. 760) has been obtained from Sumba, the other (Buit. Mus. n°. 761)
from Sumbawa. The measurements of the former have been published by Dammermann
m quot;Treubiaquot;, vol. X, 1928.

\') Even if Merkens had measured 400 specimens of cows, bulls and oxen cach, instead
ot 4, a difference of i mm. would be in my opinion of not the slightest importance.

The present writer took partly entirely different measurements
than Merkens, so that not all of Merkens\'s conclusions could be
tested by the help of our more ample materials. Those conclusions
we were able to test, are:

a.nbsp;Bulls have in relation to cows a longer forehead, and a shorter
nasal part. Oxen are intermediate in this respect.

b.nbsp;The breadth of the praemaxillaries is greatest in cows, smallest
in bulls. Also in this respect oxen are intermediate.

c.nbsp;The breadth of the base of the horn is greatest in oxen, smallest
in cows.

In order to be able to express the difference in length of the
forehead and the nasal part into % of the total length, Merkens
took(IV V)-(II III).

fig. 2). As none of these
four measurements had
been taken by me, I used
(1—7) — 7, which express-
es an almost identical
difference. The meaning

table B. As, furthermore,
by far in most of the crania of the kerabau measured, the horns
were wanting, I used instead of the breadth of the horn, that of
the horn-core\' against which no objections can be raised.

In table A will be found all necessary measurements (m %
of the total length). In the 4th, 5 th and 6th horizontal row the
smallest and greatest number have been heavy printed. _

I have tried to indicate, on the base of the characteristics, given
by Merkens, which of the crania are males, and which are females.
Table A shows clearly, that the number of contradictions at which
we arrive then, is so great, that we are certainly justified to draw
the following conclusions:

i The differences, which Merkens believed to exist between the
crania of oxen, cows and bulls, appear m reahty not to exist,
if tested on more extensive materials.
2. Consequently, it is impossible to distinguish in our own table
between oxen, bulls and cows.

3- This negative result proves that the (almost certain) presence
of measurements of oxen in our table, need not trouble us;
the differences — if indeed present — being still smaller than
the small differences which Merkens believed to exist.

Besides the cranial measurements of 31 crania of the kerabau
and of 3 crania of the arni, all already mentioned above, table B
contains the measurements of 9 specimens of our own fossil form,
of 2 specimens of quot;Bujfelus palaeokerabauquot;, borrowed from Stremme,
of 3 specimens of Buffelus bubalus, var. palaeindicus derived from
Lydekker, and — for the sake of completeness — of one specimen
of Buffelus platjceros, also borrowed from Lydekker.

As far as the condition of the specimens measured by the writer
allowed, thirty measurements have been taken. As to the choice
of the measurements I had to join Stremme i), who in his turn
for the greater part followed Lydekker. Two of Stremme\'s
measurements I did not use Besides Stremme\'s measurements
I added eight new ones (n^s. 14^ 24—30), in order to invest some
facts concerning the course of the posterior part of the temporal
fossae, and regarding length and position of horns and horn-cores.
Unfortunately, however, in most of the recent specimens measured
the horns were either wanting or — if present — could some-
times not be removed, so that the relation between horn-core
and horn could be stated in only some cases.

In order to make direct comparison possible, it is necessary
that the measurements of each specimen be expressed in % of a
certain measurement as unit. Usually as unit is taken the total
length of the cranium. We do not dispose, however, of this measure-
ment in our fossil crania, being all very fragmentary. For that
reason the smallest breadth of the forehead has been chosen as
unit in table C. In order to facilitate a review of the numerous
percent numbers of table C, the maximum and minimum values
of each measurement for the various forms have been united in
table D. Table E contains a series of relations between certain
measurements. In this respect my choice was wholly fixed, as I
had to join Stremme. Only the relations i : 20 and 19 : 7 have

\') In Stremme\'s table of measurements on p. 126 a lapsus calami occurs in the circum-
scription of measurement IV and Xll viz., quot;Orbitalkammquot; instead of quot;Occipital-
kammquot;.nbsp;\'

II. quot;Entfernung zwischen dem Hinterhauptskamm und dem Scheitelquot;. From the
description of our own fossil form it will appear that the transition of the frontals
into the supracristal portion of the occiput is often very gradual, so that the quot;Scheitelquot;
yytTtnbsp;defined. The same occurs in the cranium of the kerabau.

•AAin. tntfcrnung vom Occipitalkamm zum Vorderrandc der Orbitaquot;. As the
crista occipitalis is a wide arch, I could not know which point of this crest Stremme
meant.

been added. For the sake of an easy review the various minima and
maxima of table E have been collected in table F.

As is seen from table C, in all the specimens of the kerabau
the interval between the extremities of the temporal fossae (measure-
ment 13) is not only exceeded by the quot;greatestquot; width between
the temporal fossae (measurement 14), but also by the interval
between the extremities of the occipital condyles (measurement 11).
That m. 13 is exceeded by m. 14 is caused by the fact that the
convex posterior portion of the temporal fossa of either side
converge backwards. That, furthermore, m. 11 exceeds m. 13 proves
that the temporal fossae cut deeply into the supra-cristal portion
of the occipital surface, a fact which already Rutimeyer (1868, p. 33)
strongly emphasized. The percent numbers in table C show also
in a convincing manner that the individual variation is in this
respect extraordinarily great. In my own notices, made of the
recent specimens of kerabau which I saw in Amsterdam and Leyden,
I read: quot;The manner in which the temporal fossa terminates is
very variable. In some cases the temporal fossa and the plane of
the occiput are — roughly spoken — perpendicular to one another;
in others the posterior end of the fossa curves very gradually, and
the boundary can hardly be seen. Between these extremes are
a great number of transitionsquot;. This explains also why in table F
the relation 4 : 13 varies from 1.82—3.87, and the relation

As to the specimens of the ami, B. bubalus var. palaehidicus
and of quot;5. palaeokerabau\'\' m. 14 is only known from the arm cranium
of which measurements have been given by the author. In this
specimen m. 14 exceeds also m. 13, but only with 4 %. For the
rest the ami crania show — as in the kerabau — a considerable
variation in the difference between m. 13 and being in spec, a i,
in spec, b o and in spec, c 24. Of spec. A and B of the Siwalik
variety palaeindicus m. 11 and 13 are known. Apparently the temporal
fossae do not cut so deeply into the occipital surface m. 11 being —
in contradiction to the kerabau crania — smaller than m. 13. The
relation 4 : 13 is for spec. A of the variety in question but 1.66
which is distinctly smaller than the smallest value which was found
for the kerabau. Also in the relation 10 : 13 the fossil Siwahk
variety shows a tendency to smaller values. The only cranium of
quot;13. palaeokerabau- of which m. n and 13 are known, shows in this
respect no difference from the kerabau cranium.

Coming now to our own fossil specimens we notice that in
the five specimens, in which m. n and 13 can be compared, m 11
is W/./than m. 13, in two specimens even considerably smaller.
In this respect, therefore, they arc distinguished from the kerabau.

and in accordance with var. B. palaeindicus. This is also demonstrated
by the relation 4 : 13 of table F, varying in our fossil specimens
from 1.33—1.53 and being 1.66 in the only specimen of B. palae-
indicus, for which the relation 4:13 could be calculated. One would
perhaps be incHned to infer from the numbers quoted that the
degree of constriction of the occiput by the temporal fossae is
in our fossil form still smaller than in B. palaeindicus. It must not
be forgotten, however, that of only one skull of the fossil variety
the relation 4 : 13 is known. Moreover, the relation 10 : 13 differs
but slightly in both forms.

The percent numbers of m. 13 and 14 of table C show, further-
more, clearly the remarkable feature, already mentioned in the
description of our fossil specimens. In spec, i and 7 namely m. 14
exceeds m. 13, the course of the posterior portion of the temporal
fossae being the same as in the kerabau crania. In spec. 2, 3, 5, 6
and 8, however, m. 13 is larger than m. 14, which is caused by
the fact that near the occiput the temporal fossae change their
direction, diverging backwardly. Probably I should have been
strongly inclined to attach value to this difference, were it not
that in one of two fossil crania of Bibos in our collection — without
a shade of a doubt belonging to the living species — exactly the
same deviation occurred. I must admit, however, that I cannot
declare this phenomenon. Of course it cannot be connected with
a stronger development of the occipital muscles, as these are
attached to the true occipital surface, and not to the supra-cristal
portion of the occiput.

19 : 7. From table F appears that both differences in reahty do
not exist. Nevertheless there remain other differences. But in the
first place they are small, and secondly it is almost certain that they
would disappear if still more extensive materials of the kerabau
had been available. Stremme hinted already at this possibility.
On p. 125 Stremme mentions some more differences viz., a more
strongly ascending forehead, the presence of crests and grooves
on the forehead and (perhaps) a greater size. As to the difference
first mentioned I may add that in the kerabau the forehead appeared
to be very variable, in some cases being almost flat, in others very
distinctly convex antero-posteriorly. Concerning the second difference
I may mention that supraorbital sulci appeared to be either present
or absent in the crania of the kerabau. As to the difference in size
I may refer to spec. 30 of my table B, which shows a tip to tip
interval of the horns of about 3 m. !

Summari:(ing we can state that the fossil Japanese form, described by
Stremme under the name of Buffelus palaeokerabau Dubois, appears to
be indistinguishable from the recent kerabau. It has, therefore, no right
to specific distinctness, and must be called: Buffelus bubalus (L.) var. son-
daicus fossilis. 1)

Finally our own fossil form. As far as can be gathered from
our fragmentary specimens the resemblances to the cranium of
the kerabau is so close that I should not hesitate in classing it with
that variety, if not the degree of constriction of the occiput were
far too small. In this respect the fossil form of our collection is
distinctly separated from the kerabau, and shows resemblance with
B. bubalus (L.) var. palaeindicus (Falconer). In my opinion, however,
this single difference does not suffice to establish a new variety
so that I class it provisionally with Buffelus bubalus (L.), var. sondaicus
(ScHLEGEL et Müller) fossilis.

With this I do not intend to maintain that Bufftlupa acohrabau does not exist though -
^ r has been pointed out already above - identity between Stremme\'s fossil form
and that of Dubois is not improbable.

Of the genus Bibos I received three crania and one detached,
entirely complete horn-core, obtained from the following locaHties.

Though two crania had to be freed from considerable quantities
of adherent matrix and restored with glue and plaster, they are
in a rather fine state of preservation. The third cranium, however,
when unpacked, showed itself to be for the greater part fallen
into tiny pieces, so that it was absolutely impossible to restore
it. What remained was the left maxilla together with a portion
of the orbit, the greater portion of the right horn-core and a
smaller piece of the left core. The maxilla contained five grinding
teeth (P3—M^), all finely preserved.

The two crania of which views are given in figs, i—2, 4—5
of pl. II and figs, i—3 of pl. Ill agree so much in nearly every
feature that their belonging to the same species cannot be doubted.
The remains of the crushed cranium and also the detached horn-
core show no characters by which they are distinguished from
the better preserved specimens.

As will be seen from the figures, one cranium only possesses
one horn-core, the other both. The specimen with one horn-core
is in general better preserved. In the specimen with two horn-
cores the nasals are for the greater part wanting, the occipital
surface is injured; the pterygoid-sphenoid region very much damaged,
and what remained of it covered with plaster in order to give the
very desired support to this cranium which was broken into two
behind the orbits.

As Rütimeyer already gave such an excellent account of the
skull of the living banting, Bibos sondaicus, and of the great differences

which are connected to sex and age, it will be unnecessary to explain
why the two figured crania and the figured, detached horn-core
all belonged to male individuals. The two fragmentary cores be-
longing to the crushed cranium also indicate the skull of a male.
There are, however, differences in age. From the curvature of the
cores we may be sure that the cranium with two horn-cores is that
of a mature male individual, while the cranium with one core,
and the detached core belonged to old males. The molars prove this
assertion to be true. In both crania all the molars are in use, but
the degree of wear is greater in the cranium with one horn-core.
Though of the cores of the crushed cranium a considerable part
of the distal end is wanting, the curvature of the remaining part
proves the cranium to belong to a younger individual. And the
degree of wear of the molars supports this statement, for, though
the M^ is already in use, the remaining premolars are but little worn.

As the resemblance between the two figured crania on the one
side, and those of the living mature male Javan banting on the
other side, is very close, it will be best to commence by comparing
the two together. If, however, differences occur between the two
fossil crania, they will be mentioned separately.

For direct comparison I had at my disposal only one cranium,
but a very fine one, namely that of an old male, contained in the
Museum of the Zoological Institute of the University of Utrecht,
and kindly lent to me by its director. Prof. Nierstrasz. The fossil
specimens could, moreover, be compared with the descriptions
and figures of some fine crania of the banting i), preserved in the
National Museum of Natural History at Leyden, and with my
own sketches and notices of these crania and of two more specimens
contained in the museums of quot;Natura Artis Magistraquot; at Amsterdam.

Below the comparison between the fossil specimens and the
crania of mature male individuals of the Javan Bibos sondaicus
will be made in this way, that respectively upper-, hind-, side-,
and lower aspect of the cranium and the horn-cores will be dealt
with, each preceded by an enumeration of resemblances and dif-
ferences which the present writer could state in the recent specimens.
(N. B. The crania are supposed to rest upon paroccipital processes

This aspect shows but small individual variations. As has
been already pointed out by RÜTmEYER and clearly shown by the

figures, given by Schlegel, the whole skull becomes, from mature
to old, broader in reference to its length, with the orbits very
prominent, masseter crista and malar process very well developed.
Nasals broad and short. I should like to add that the supraorbital
sulci and foramina may show considerable variations concerning
length and convergence of sulci, position and number of the supra-
orbital foramina. I am, therefore, convinced, that no specific value
whatever may be attached to position and number of supraorbital
foramina, and to length and position of the supraorbital sulcus.

Concerning the prominence of the orbit, the shortness of the
nasals and their relation to the frontals, there is a very close resem-
blance between the fossil specimens and the recent crania. The
masseter crista and the malar process are, however, less developed
than in general is the case in the crania of the living species.

As will be seen from the figures the two fossil specimens show
slight, mutual variations concerning the length of the supraorbital
sulcus, and the number and position of the supraorbital foramina.
But these variations by no means exceed the boundaries of variation
found in the crania of the Javanese banting.

The crania of mature male individuals of Bibos sondaicus showed
that the transition of the plane of the frontals in that of the occiput
is in some specimens gradual, in others rather abrupt. The latter
is caused by the fact that the forehead — in profile view of the
cranium — instead of being flat, and situated in the elongation of
the nasals, is slightly concave and rising strongly to the vertex
cranii, producing a ridge, called by Rütimeyer quot;Frontalwulstquot;
and identical with Lydekker\'s quot;intercornual ridgequot;. The feature
above described appears to be connected with a rich development
of bone-tubercles on the posterior half of the forehead.

From the above it will be clear that the relation between height
of supra- and infracristal portion of the occiput has but little value.

As mentioned by Rütimeyer, the true occipital surface becomes
gradually broader, so that in an old male the lower portion of the

This is also the case in both fossil specimens. In the specimen
with one horn-core the interval between the posterior border of
the meatus auditivus externus osseus of either side exceeds with
some 7 cm. the interval between the bases of the horn-cores, and
in the specimen with two horn-cores the difference between both
measurements is about 8 cm.

There is a difference — though not essential — between the
fossil specimen with one horn-core, and that with both cores re-
garding the degree of constriction of the supracristal portion of
the occiput by the posterior extremities of the temporal fossae,
being in the quot;two hornedquot; specimen very considerable, and in
the quot;one hornedquot; cranium far less pronounced. It will be seen,
however, from table H of cranial measurements that in this respect
there occurs a considerable amount of individual variation in the
crania of the Javan Bibos sondaicus. This is caused by the fact that
in some specimens the curved walls of the temporal fossae gradually
converge backwards, whereas in other crania the extremities
of the temporal fossae change their direction near the occipital
surface, so that the utmost ends of the temporal fossae diverge
strongly backwards. It will be clear that in the latter crania m. 25
(of table G and H) will show relatively great values. But in none
of the six recent crania m. 25 exceeds m. 26 (greatest interval
between temporal fossae in the prae-occipital region). The quot;two
hornedquot; fossil cranium does not show divergence of the utmost
ends of the temporal fossae, so that in this specimen the constriction
of the occiput is very considerable and m. 25 distinctly smaller
than m. 26. In the quot;one hornedquot; specimen, however, this divergence
occurs in such a strong degree, that not. only the constriction of
the occiput is far less pronounced, but m. 26 is even slightly exceeded
by m. 25. But of course this small deviation does not afford any

In both fossil specimens the tips of the paroccipital processes
are wanting. In position, and curvature of the remammg portion,
however, they show no differences from those of the crania of

In the crania of Bibos sondaicus, and also — according to
Rütimeyer — in those of Bibos frontalis there runs, from the
indentations by the temporal fossae on either side, a crista obhquely
upwards meeting in the middle of the breadth of the supracristal
portion of the occiput and some cm. above the middle of the
occipital crest. In examining six complete crania of the living
Javan banting it was revealed that also in this respect some
variation may occur. In some crania these cristae originate there
where occipital crista and posterior extremities of the temporal
fossae meet; in other specimens they begm about 2 cm. higher,
namely at the upper angle of the indentations of the occiput by
the temporal foSae. In some specimens the point where both
cristae nieet is situated very near to the vertex of the cranium.

in others some cm. below it. In other words at times both cristae
converge rapidly, at other times they enclose a greater angle. In our
fossil specimens these cristae are but poorly developed, they ineet
about 3 cm. above the middle of the occipital crest and about 4 cm.
below the vertex cranii. Exactly as in the crania of the banting,
in or immediately above the point of meeting the supracristal
portion of the occiput bulges up, causing the quot;intercornual ridgequot;
to show m the upper aspect of the cranium a convex outline in
its middle third.

Viewed from the side all the crania of mature male specimens of
Bibos sondaicus which I examined were very much alike. The only
difference which may occur, and which has already been mentioned
in noticing the frontal aspect, is the sometimes abnormally strong
development of the posterior portion of the forehead causing the
plane of the frontals to slope stronger than that of the nasals.
For the rest, however, the relation between frontals, nasals, lacrymalsj
maxillae, and intermaxillae, the shape and position of\'the orbits\'
the development of masseter crista and malar process is in all the
specimens I saw, very much alike. Also in this respect both fossil
crania are perfectly in accordance with the skulls of the Javan
banting, with the only exception, already mentioned, that masseter
crista and malar process are not so strongly developed as seems
to be the rule in the crania of the banting.

Here also we may be short. Both fossil specimens show the
same relations of the bones of the bottom of the cranium as in
the banting. For dimensions, and comparisons with those of the
banting, I may refer to tables G and H.

In but one specimen of cranium of the Javan banting (E)
the horns could be removed from the cores. It revealed that the
core by no means fills the sheath of the horn to its top, but that
a distinct difference in length exists between horn and horn-core.
As, especially in old male individuals, the curvature of the horn
is strongest in its distal third, the core will always show a smaller
degree of curvature than the horn which belongs to it. That explains
why I ventured to refer the quot;one hornedquot; fossil specimen to an
Old individual, although the curvature of the core is not so extremely
pronounced. I shall not try to give an account of the various forms
or the horns of the crania which I saw. Their degree of curvature

is — as was clearly pointed out by Rütimeyer — a function of age.
I am not wholly certain, but I believe that the distance from tip
of horn to tip of nasals, which, as will be seen from tables G and H,
may vary considerably, is not only dependent upon age, but is
also liable to common individual variation. For the sake of complete-
ness I will try to describe the form and curvature of the horn-cores
of both fossil crania, and of the detached core.

Near its base the core is directed outwards, downwards and
somewhat backwards. This downward slope decreases gradually
so that the core in the middle third of its course is — roughly spoken
— horizontal. Then the core rises upwards in one regular curve,
its tip directed slightly inwards.

If the cranium is placed on paroccipital processes and cheek
teeth, and viewed in side aspect, the distal third of the horn-cores
slopes distinctly backwards. The lowest point of the inferior sur-
face of the core is at the level of the lower angle of the occipital
condyles. At the base the horn-core is in cross-section approxi-
mately an ellips; its longest axis forms the antero-posterior dia-
meter; inferiorly it is far more flattened than superiorly. The core
maintains the flattened lower border throughout its proximal
half; more distally it gradually disappears, and the core becomes
conical. The surface of the core is deeply grooved and chanelled.

The place to which the base of the former horn extended is
on lower and posterior surface of the core marked by a sudden
decrease of height and breadth.

This core resembles the one, described above, in nearly every
feature. It is only larger and thicker, and its tip is more directed
inwards. It apparently belonged to a still older male. Without
difficulty it can be identified as a left core.

The differences between the cores of this specimen and the one
described first are:

1.nbsp;In its proximal third it is directed somewhat more backwards
and less downwards.

2.nbsp;Its tip is directed straight upwards and not inwards (which
proves its belonging to a less old individual).

3.nbsp;Placed in the position above mentioned, and viewed in side
Lpect, the distal third of the horn-core stands nearly perpendicular.

4. The lowest point of the inferior surface is at the level of the
upper border of the for. magnum.

The cores belonging to the crushed cranium are already men-
tioned, and are in too imperfect a condition to make a detailed des-
cription possible.

The form of the cores in cross-section of our fossil specimens
s entirely similar to that of the Javan banting. Also concerning
he course of the cores there is in general a close resemblance be
tween the fossil form and Bibos sondaicus. In the quot;one hornedquot; and
two horned specimen, however, the maximum span of the horn-
cores, in proportion to the total length of the cranium, is distincdy

(See table H). That this is a point of minor importance is in
my opmion proved by the fact that the detached fossil horn-core

is considerably less
spread out, showing
almost exactly the
same degree of cur-
vature as the horns
of the type specimen
of the cranium of an
old male (Compare
fig. 4 pi. Ill, and text

between the molars and premolars of the fossil specimens and those
of the recent Javan banting, it will suffice to refer to fig. 3 of pi II
m which an upper view of the dentition has been given. In thé
1nbsp;specimen are preserved: leftnbsp;a fragment of the

fr.u^\'\'^^^ o^quot;quot; bave to compare our fossil form with the
following South Asiatic and Malayan species:

Bibos frontalis (Lamb.)
Bihos gaurus (H. Sm.)
Bihos palaeogaurus (Falc.)
Bihos geron Mats.
Bihos protocavifrons Dub.

The comparison between the fossil form of our collection with
the Javan Bibos sondaicus has already been made in what precedes,
but there are still some varieties of Bibos sondaicus which need dis-
cussion.

Bos (Bibos) gavaeus, L. Rütimeyer, N. Denkschr. Schweiz. Ges. XXII, 1868, p. 105.

Sec for more extensive synonymy of these two species: Lydekker, Wild oxen, sheep and

We may immediately exclude Bibos frontalis, the gayal, and
Bihos gaurus, the gaur from our comparison. These recent South
Asiatic species are readily distinguished from the recent Javan
banting — and therefore also from our fossil form whose striking
resemblances to the former has been repeatedly emphasized in the
foregoing part — by a number of features of which I shall mention
the following ones. In both species the cranium is less elongatedi).
In the Mus. of Nat. Hist, at Leyden one cranium of gayal is
preserved. It belongs, however, to an individual not fully grown
for which reason I thought it better not to give the measurements.
The relation between the smallest breadth of the forehead and
the condylo-basal length in this specimen is 221 : 421, whereas
from table G of cranial measurements may be seen that in all the
specimens of the Javan banting the same relation is decidedly
smaller. It must not be forgotten that the relation between the
smallest breadth of forehead and condylo-basal length is smaller
than may be expected in the cranium of an adult individual, as
the cranium becomes broader in proportion to its length. This
smaller width in reference to the total length coincides with a
triangular form, far more pronounced than in the banting head.

Another difference is that the horns are but very slighdy curved,
with no inward bending. Of the gaur I have not seen the cranium.
But this species is much better known. That its skull is less elongated
in comparison to that of the banting can be clearly seen from the

Bos (Bibos) palaeogaurus L. Rütimeyer, Abh. Schweiz, pal. Ges V 1878 n t..
Bos palaeogaurus, R. Lydekker, Cat. Foss. Mamm. Br. Mus. pan. II, ?r8\'5,^p. If

Rütimeyer mentions under the present name a fragment of
a cramum from the Narbada and preserved in the British Museum
Rütimeyer should have named the species without any reserve
Bibos gaurus, h^d he not thought it useful to give in general fossils
new names. quot;Dennochquot; - he continues - da ich Linen Unter-
schied von dem lebenden Bos Gaurus entdecken konnte, würde
ich auf einen Namen verzichtet haben, wenn derselbe nicht, freilich
ohne Nachweis, dass er diesem Fossil gewidmet war, in den Manus-
cripten Falconer s vorgefunden hättequot;. Lydekker stated as his

opinion, that the specimen is perhaps insufficient for specific deter-
mination.

however were already known from China long before ZtsZto
described the present species. Both Koken ^ an^ScHLSsER ^^
namely referred cheek teeth of their collections to the genus bL^

on such maJals. MLuMoxoTated aTL\'Ät ^hquot; s^Sequot;
under consideration m ght oossiblv be iA . tne species

W. Granger, palaeontologist of the Third Asiatic Expedition,
obtained also from Sze-chuan a fine collection in which the genus
Bibos was represented by quot;a series of skulls, skeletons, upper and
lower jaws, etc.quot; i), which were all referred to Matsumoto\'s new
species by Matthew and Granger. Of a complete skull, selected
as neo-type, figures of side-, upper- and lower aspect were given.
No detailed description, however. Nevertheless it is not difficult to
see that Bibos geron is specifically distinct from our fossil form.
The totally different outline of the forehead in side aspect of the
cranium, the different course of the horn-cores, especially in their
proximal portion, the greater breadth of the cranium in reference
to its length present positive indications in this respect. Among the
numerous remains from Chou-K\'ou-Tien, extensively described and
figured by Zdansky under the name of Bibos geron, the cranium was
only represented by a damaged horn-core and a fragment of ditto.

(1)nbsp;Bibos protocavifrons, Dubois, Tijdschr. Kon. Ned. Aardr. Gen., 2de ser. XXV,
1908, p. 1262.nbsp;, . , ^

(2)nbsp;Bibos palaeosondaicus, Dubois, Tijdschr. Kon. Ned. Aardr. Gen., 2de ser. XXV.
1908, p. 1262.

? Bibos palaeosondaicus, Stremme in: Die Pithecanthropus-Schichten auf Java,
1911, p. 136, pi. XVIII, figs. 7—8; pi. XIX, figs. 7, 9—11; pi. XX, figs. lo—11;
text fig. 10.

dem lebenden Banteng anschliessenquot; These forms he united into
a new species, Bibos palaeosondaicus. Other forms showed transitions
from the latter to the Javan fossil species of Leptobos, whereas
one form occurred which, in having a very high intercornual ridge
and concave forehead, strongly resembled the gaur. Whence the
name Bibos protocavifrons

Above has been repeated all that Dubois has mentioned
on these new species. It is clear that our own fossil form does
not belong to Bibos protocavifrons. Furthermore, it is not improbable
that our form is identical with Bibos palaeosondaicus. But as little
as we were able to prove the identity of the fossil Buffelus of our
collection with Buffelus palaeokerahau Dubois, as little will it be
possible to identify with sufficient certainty our form of Bibos with
Bibos palaeosondaicus Dubois.

Stremme, however, apparently esteemed the few sentences
which Dubois dedicated to Bibos paUeosondakus, sufficient enS

Mrs amp;elenka, as belonging without doubt to that species According
to Strome, Dubozs, in naming his form BiL pMeosold^fs
followed the same custom as Rutimeyer did, when fhe iX S

I do not know how Stremme came to this conception. AppafeX
others have not interpreted Dubois in this way, the absence rfanv
recent species among the mammalian remains o{ the Senfbeds

f.k.nV^\'Pquot;quot; these measurements with the corresponding ones
taken from 2 crama of the Javan banting and from^ specimen

will probably have compared it - though nlhe e k ifm^u^ned
with crania of adult male individuals of the aC^nrr
banting. This talcen for granted, it is extremelv Zk Ki ?
of the two crania, which^SxRBMME cLsiS LÄU\'!^
Javan banting, in reality belongs to the Balian nv quot;fnbsp;quot;

be an exception as\'t Hoen i) gave a photograph of a Uving specimen
of the adult male Balian ox, in which the horns are also but slighdy
curved.

As to the measurements taken by Stremme, the following remarks
may be made. What Stremme meant by m. i, 2 and 3 is clear; they
agree with m. 3, 25 and 24 of my own table G. The meaning of
m. 4 and 5, however, is not clear at all. We may even say that
Stremme did not mean by m. 4 the height of the supraoccipitale.
It will be known that in all Artiodactyla it is not the supraoccipitale
which bounds the for. magnum superiorly, but the median united
exoccipitalia 2). That
this is also the case
in the cranium of
Bibos sondaicus may
be clearly seen in
text fig. 4, which re-
presents the hinder
aspect of the type
cranium of the adult
Javan banting. The
figure of Stremme\'s
fossil specimen
shows, however, dis-nbsp;• • 1

tinctly the absence of the suture between supraoccipitale and
exoccipitalia, so that Stremme certainly cannot have measured the
height of the supraoccipitale. Originally, I supposed that Stremme
in reality might have meant by m. 4 the interval b^ween upper
border of the for. magnum and crista occipitalis. This interval,
however, appears - according to the figure, the exact scale of
which can be ascertained without any difficulty — not to be

The circumscription of m. 5 is very indistinct. What exactly
is meant by quot;Stirnmittequot; ? The highest point of the curve, formed
by the more or less abrupt transition of forehead into occiput
in the median One of the cranium? Probably not, for the distance
between this point and the upper border of the for. magnum is —
according to Stremme\'s figure — at most 176 mm.,mstead of 205 mm.
Wc go, therefore, not too far, if we conclude that Stremme\'s m. 4

In spite of all these objections it cannot be denied that Stremme\'s
fossil specimen possesses indeed a higher and narrower occiput

than m the adult male cranium of the Javan Bibos sondaicus appears
to exist. It must, however, strongly be doubted whether Stremme
had a right to consider his fossil form so closely related to the
recent Javan banting. For, beside the difference above mentioned,
Stremme\'s specimen is — in my opinion — very decidedly dis-
tinguished from adult crania of the Javan banting by totally different
form and course of the horn-cores. As Stremme himself observes,
the horn-cores taper rather rapidly, and are but slightly curved!
In all typical adult male crania of the recent Javan species, however
the horns are very strongly bent, and the horn-cores taper gradually i)\'
The above will suffice to show, I believe, that our fossil form is
by no means identical with Stremme\'s fossil specimen.

The question whether our form is identical with Bibos palaeo-
sondaicus Dub, of course remains unsolved.

Bos sondaicus, Schlegel and Müller, Verb. Nat. Gesch. der Ned. Overzeesche Bezittingen

Bibos banting. Gray, Knowsley Menagerie, 1850, p. 48.
Gavaeus sondaicus, Blyth, Journ. As. Soc. Bengal XXIX, i8lt;5o, p. 296.
Bos (Bibos) sondaicus, Rütimeyer, N. Denicschr. Schweiz. Ges. XXII, 1868 p 77-
Lydekker, Wild oxen, sheep and goats, 1898, p. 36.nbsp;\' ^

The two latter races, being founded on differences in coloration
of the skin, are of no interest here.

In Lydekker\'s pubhcation of 1898 mention was already made
and frontlets were figured of the Bornean banting, which seemed
to be distinguished in having the horns less spread out and directed
more upwardly. At that time Lydekker did not know, however
whether this difference should prove to be constant. In 1912 2)
he returned to the Bornean form, the British Museum having
obtained new specimens both of Javan and Bornean banting
^rom these specimens it appeared that the difference noticed above
is indeed constant and, moreover, that the Bornean banting is
characterized by the flatness of the forehead and the straight inter-

In the National Museum of Natural History, furthermore, a cranium
of the Bornean banting is preserved, which is according to the
label that of a female. From the course and curvature of the horns
we may be sure, however, that the latter specimen represents the

In a relatively recent publication\'t HoenI) mentions the occur-
rence on the island Mojo, N. of Sumbawa, of the Balian ox, which
has there reverted from domesticity. He regards the wild Bornean
form as probably identical with the form of Mojo. There appears,
therefore, to exist still a great amount of doubt as to the origin

It will be known that the wild Javan banting interbreeds freely
with the European and Indian cattle. Furthermore, it will be re-
membered that higher up mention was made of the occurrence
of the domesticated banting in several islands of the Malayan
Archipelago. Consequently it is highly possible that among

Rütimeyer, Vcrsuch einer natürlichen Geschichte des Rindes, n. Denkschr. Schweiz.
Ges. 1867—1868.

museum specimens, labelled as Bibos sondaicus, crania of the
domesticated form, and of hybrids are contained. In this connection
I think It desirable to lay stress upon the fact, that the five crania of
male individuals, of which the measurements are given in table G
without any doubt belonged to wild specimens of the Javan banting\'
As to the crama D and E no explanation is needed; concerning
specimens A, B and C it will suffice to mention that they show
every essential feature which the type crania exhibit. Originally
the present writer intended to give in table G also the measurements
ot a rather large number of frontlets, which he found in the various
museums, and which were regarded as belonging to the Javan
banting. After all, their measurements have not been added for
fear that in frontlets alone the presence of the domesticated form
and of hybrids could perhaps not always be detected. Besides the
measurements of the five crania above mentioned, table G contains
those of the three male crania of the Bornean banting already
noticed on p. 54 and 55. They will, however, not be used in the
below, the Bornean banting perhaps not being, as we saw higher
up, an originally wild form. Perhaps that at some time or other
the measurements given — together with those of a much greater
number of crania — will provide sufficient data to solve this problem.
Though I myself did not want measurements of the crania of
females, I added the measurements of the type cranium of the
adult female, which may be of use to others.

As far as the materials allowed, forty-one i) different measure-
ments have been taken. As to the choice of the measurements
I partly joined Rutimeyer who measured also the type specimens 2)
In order to avoid confusions it will be very desirable to explain
what IS meant in m. 2 and 19—21 with quot;vertex craniiquot;. Text fig 5
shows some ways in which the transition of forehead into occiput
m the niedian line of the cranium may take place; point ^ that is
to say the middle of the uppermost curve, marks the place which
nas been taken as vertex crania.

In table H the various measurements have been expressed in %
of the condylo-basal length of the cranium, while in table I the
relations between certain measurements are given.

If we have a close look at table H — in which the maximum
and minimum values for the crania of the Javan banting have
been heavy printed —
we notice that also
the various percentage
numbers reveal the great

There remain still dif-
ferences, but everyone will have expected them, bemg the
measurements of but five recent crania available for comparison.
From table H also distinctly appears the greater maximum span
of the horn-cores (m. 13), and the greater values of those measure-
ments, which bear upon m. 13. in the fossil crania. I should have
been inclined perhaps to accept specific distmction on these grounds,
were it not — as has already been stated — that the form of the
detached horn-core of our collection is in perfect accordance with
that of the type cranium of the old male of the Javan banting.

At last what does table I reveal? The notorious fact that nearly
all the values found for the fossil form are situated between the (heavy
printed) maximum and minimum values, found for but five recent
crania! That the relation 19 : 25 is so small in our specimen a
is caused by the fact that m. 25 (width of occiput between extremities
of temporal fossae) is in that specimen abnormally great.

For the rest however, the accordance is remarkably great, so
that in my opinion the fossil specimen may not be regarded as
a distinct form, but specifically identical with the recent Javan
banting, Bibos sondaicus (Schlegel et Muller).

Our collection contains two crania of rhinoceroses, both obtained
from Bondol near Kuwung (District Randublatung, Regency Blora
Residency Rembang). Notwithstanding the fact that one specimen
is ill-preserved, we may say that these new founds enrich our
knowledge of the fossil rhinoceroses of Java not inconsiderably
as will appear in the sequel.

The middle portion of right zygomatic arch is wanting; the
left arch is remarkably well preserved. The praemaxillaries have
been totally lost; the processes of the maxillaries with which they
are connected too. The absence of the praemaxillaries need not
surprise us, as in museum specimens of recent rhinoceroses they
are often wanting. The pterygoid processes are injured; the par-
occipital processes broken off near their base. The palate is con-
siderably damaged, both in front and in back. The petrosal of
either side has been lost, their former presence being now indicated
by a hole on either side of the basioccipitale. (In museum specimens
1 often found the petrosals detached in the brain case). The thin
wall, which separates the left orbit from nasal cavity has been
destroyed. The vomer is also totally absent. Of the cheek teeth
only the crown of the left pi is wanting. The teeth are on the
Whole more or less injured. The degree of injury will be mentioned
later on, when describing them in detail.

Brachycephalic, with strong, frontal depression and moderately
high occiput, which slopes backwards i). Postglenoid and post-
tympanic processes of the squamosal ankylosed below the meatus
auditorius externus. Frontals smooth, rugosities totally absent,
proving that the specimen in question was not provided with (a)
frontal horn(s). Nasals rather pointed, separated by a groove ex-
tending from tips to vertex of nasal arch. On the nasals a strong
rugosity occurs, extending - on the upper side — nearly from
tips to vertex; on the lateral sides even some 6 cm. more backwards.
Cristae fronto-parietales do not meet in the median line. They
are sharper than would appear from fig. 3, F-. Some 5 cm.
in front of the occipital crest a low median crista occurs, which
extends about 5 cm. frontwards, dying out gradually. If one compares
the specimen in question with a sufficient number of equa ly aged
crania of the living Rhinoceros sondaicus, it will be revealed that
the eye cannot perceive any essential structural difference. In my
opinion, therefore, it would be a waste of words to describe the
cranium under consideration more detailed. Of course we shall
have to return to it, in discussing the tables of measurements.

In order to be able to give the inner aspect of the cheekteeth
a photograph has been taken (fig. 4, pl- V) from the image, produced
by a mirror placed obliquely between both rows of teeth.

Right: Antero-external angle damaged. Subtriangular shape Ground
down to the base of the crown, presenting a disc of dentine
Palatinally and posteriorly of the middle a very small and
shallow pit occurs, surrounded by an edge of enamel. This
pit represents the remainder of the medisinus.

P2 (antepenultimate premolar).
Left: The enamel of the ectoloph has been lost. Quadriform.
Deuterocone almost isolated. No crochet. Postfossette oval-
shaped; in its antero-external angle a very incipient enamel
fold Though the enamel of the ectoloph has been lost,
stillk can be seen, that the protocone presents on the outer
surface of the tooth an ill-defined rib, which dies out towards

gt;) Here and in the following pages the direction of the slope is interpreted in the
manner of the geologists 1

the base of the crown. In general the outer surface is dearly
convex with a concavity between protocone and tritocone
A small protostyle probably was present, the indications
Aerefore are, however, not absolutely conclusive.
The buccal extremity of the medisinus is triangular and
its bottom IS situated considerably below the entrance to
the medisinus.

Right: Preservation and structure as in the left one, with the only
exception that Ae postfossette does not show an incipient
enamel fold in its antero-external angle.

Anterior cingulum is present. Also internal cingulum, which
extends from the middle of the deuterocone to the middle
of the tetartocone and consists of a single row of very

upwards. The entrance to the medisinus is on the same
level with the middle of the internal cingulum. Both are
situated some 12 mm. above the base of the internal side
of the crown. In the succeeding teeth this distance decreases
gradually, so that in the M^ the entrance to the medisinus
is almost at the base of the crown.

Right: The state of preservation is better than that of the left
one only the anterior half of the ectoloph being damaged.
The crochet is still more blunt; an enamel fold in the antero-
external angle of the postfossette is absent.

Right: The only tooth which is not damaged. The ectoloph is
perfectly preserved, fortunately allowing us to study the
outer surface of the tooth. The ectoloph shows a very
strong protostyle, separated by the parastyle fold from the
protocone, which manifests itself on the outer surface by
a distinct rib (protocone style). The latter style, which
dies out towards the base of the crown, is well-defined in
front, but posteriorly it passes gradually into the outer
surface of the ectoloph. Outer surface and posterior surface
of the tooth are almost perpendicular to one another, forming
a sharp edge, which therefore may be called the tritostyle.
The crochet is double. The buccal one is lobe-shaped, the
palatinal one sharp triangular. The secondary enamel fold
in the postfossette is as distinctly developed as in the left P^.
Internal cingulum as in the left P^.

Left: Ectoloph for the greater part broken off. The protoloph
shows a distinct swelling in its protocone portion. The
single crochet is strong, broad and blunt, and approaches
the protocone very closely, being separated by a space of
about i mm. breadth. The metaloph is very obhque with
regard to the ectoloph. Already in the P^ the metaloph is
oblique, but the degree of obliqueness increases m the
succeeding teeth, culminating in the M^. The postfossette,
without a secondary enamel fold, is clearly ovalshaped, its
longest axis almost parallel to the oblique metaloph. The
ectàoph, though strongly injured, shows stiU a rather
strong metastyle. Internal cingulum is wanting Only a very
small tubercle is situated near the entrance to the medisinus.

Right: Ectoloph totally broken off. The structure is exceedingly
like that of the M^

Left: Ectoloph considerably injured. The worn protoloph is in
its top everywhere of equal breadth. More towards the
base of the crown the protoloph shows the same swelling
of its protocone portion as in the foregoing M^ The crochet
is sharper than in the M^ and does not approach so close the
protoloph. The medisinus is in general wider The post-
fossette is triangular, posteriorly bounded by a hne y
crenulated, posterior cingulum. Internal cingulum is totally

Right: The structure resembles that of the left one down to the
smallest detail. The state of preservation is better in so
far as only the antero-external portion of the ectoloph has
been broken off. A distinct metastyle is developed. The
outer side of the metacone is clearly concave, especially
in its upper portion, which is, moreover, strongly inclined
outwards.

Left: Strongly damaged. The enamel of the outer surface of
inner- and front side of the protoloph has been lost, while
the crochet also is considerably injured. It will therefore
be better to study the structure of the M^ in the right molar.

Right: The top of the united ectoloph and metaloph, and of the
crochet is somewhat damaged. The base of the crown is
subtriangular. Parastyle is present, but moderately developed.
Behind the parastyle another rather ill-defined rib occurs,
namely the manifestation of the paracone on the outer
surface. The crochet is sharp and elongated. Internal cingulum
is absent. The postero-external angle of the outer surface
of the tooth ends in a sharp tubercle, closely attached to
the outer surface, and representing a vestige of the posterior
cingulum.

Brachyodont. P^ and P^ are completely molarized. The metamor-
phosis of the P2 into the molar pattern is not so progressed in so far
as the deuterocone is still distinctly isolated. The pi is far too much
worn to show its former structure.

Though the ectoloph of most teeth is to a smaller or greater
extent damaged, we \'may be sure, from the evidence of the un-
damaged right that the outer surface of the teeth is marked
by two vertical ribs viz., a strong parastyle (protostyle of premolars)
and a distinct paracone style (protocone style of premolars), in
front separated by the parastyle (protostyle) fold. Backwards the
paracone (protocone) style passes more or less gradually into the
outer surface of the tooth. The edge where outer- and posterior
surface meet is sharp and may perhaps be called a metastyle (in
premolars tritostyle). The M^ is certainly provided with a metastyle.
Irotoloph and metaloph are oblique in regard to the direction
ot the tooth row. The metaloph is always more oblique than the
protoloph, while the degree of obliqueness of the metaloph in-
creases gradually in the teeth more backwards.

Crista and antecrochet are totally absent. A crochet is distinctly
developed from the P^ up to and including the M^. In the right P^
the crochet is distinctly double.

In the antero-external angle of the postfossette of the left P^
and P3, and of the left and right P^ a secondary enamel fold occurs.
In both P4 this fold is most distinct. It is totally wanting in the

An outer cingulum never occurs. Inner cmgulum is present
in ps and P^ of either side, and consists of a curved row of small
tubercles. In the molars the inner cingulum is sometimes present
in the form of a single incipient tubercle in the entrance to the
medisinus. An anterior cingulum is always present (except perhaps
in Pi). It runs from the antero-internal angle of the protoloph
(at a level equal to the inner cingulum, if this is present) in the
direction of the outer surface of the tooth, thereby bending strongly
upwards, and forming a valley, which probably may be called
praesinus i). A posterior cingulum is distinctly visible in both M^,
whereas it is represented in the M^ in the form of a tubercle.

The posterior side of the protocone (deuterocone in premo ars
and the anterior side of the hypocone (tetartocone in premolars)
are both either straight, shghtly concave or slightly convex in
profile view of the molar. Both converge strongly downwards,
so that the palatinal portion of the medisinus becomes gradually
narrower towards the base of the crown. The buccal extremity
of the medisinus is wide or narrow depending upon the development
of the crochet and the degree of wear of the tooth.

From the slightly worn M^, we may infer that the specimen
described above, belonged to an adult, though not old mdividua
The second specimen, however, the description of which wi
follow now, ce^rtainly belonged to a very old individual, as will
be seen below.

The whole of the occipital and parietal region of the cranium
is very much damaged. The brain caseJs t^lly uncover^^andjms

been filled for a good deal with plaster. As both petrosals, and a
portion ot the basioccipitale have been lost, the bottom of the
brain case presents a shapeless hole. The upper surface of the nasals
is so much injured, that it cannot be seen whether or not a rugosity
for a nasal horn occurred. Both zygomatic arches are remarkably
well preserved; locally, however, restored with plaster. Prae-
maxillaries and the processes of the maxillaries, with which they
are attached, have been lost. Both postglenoid processes have

been broken off for the
greater part. This is
also the case with the
right paroccipital pro-
cess, whereas of the left
corresponding process
only the base can be
seen. The tips of both
pterygoid processes are
absent. The palate is
injured both in front
and in back. The nasal
cavity is for the greater
part filled with matrix,
which could not be
removed, without run-
ning the risk to demo-
lish the whole of the
cranium. Of the right
occipital condyle the
greater part of the sur-
face of the lower side
is preserved, so that the
total length of the cra-
nium could be estimated with a tolerable degree of certainty.

Brachycephalic. Strong depression of frontal region. Union of
posttympanic and postglenoid process below the meatus auditorius
externus. No rugosities can be detected on the frontals, indicating
the absence of frontal horn(s). Due to the injury of the nasals we
cannot detect whether a nasal horn was present.

Of the right tooth row remains of P»—AP arc present; of the
left tooth row still less has been preserved viz., only remains of

Ml—M3. These remains show that the degree of wear is very great;
AP is ground down to a level only about 8 mm. above the bottom
of the palatinal extremity of the medisinus. We may, therefore,
be sure, that the cranium under consideration belonged to a very
old individual.

The state of preservation of these premolars and molars being,
in general, very bad, it will be of no interest to describe them detailed.

The little, which these poor remains still exhibit of the structure
of the cheekteeth, seems to show that by no single character the
cheekteeth in question differ from those of the foregomg specimen.

Before commencing to explain how we came to the conclusion
that the well preserved rhinoceros cranium of our collection was
indistinguishable from that of the living RJjinoceros sondaicus
will make a comparison between the two specimens described.

As far as the bad state of preservation of the one allows of
comparison, there appears to be an absolute accordance regarding
essential characters. Both show:

2.nbsp;union of postglenoid and posttympanic processes below the
meatus auditorius externus;

Above it was seen, furthermore, that the remains of the cheek-
teeth of the ill-preserved specimen did not show any character
by which the teeth could be distinguished from those of the well

locality, it is very probable that both belong to the same species.
On the other hand however, it cannot be denied that four of the
ten measurements, which could be taken i) from the fragmentary
specimen, show - if expressed in % of a certain measurement
as unit - rather great differences with the corresponding maximum
or minimum values found for 16 adult crania oilUnnoceros sondaicus.
Though I certainly do not neglect the fact that more ample materials
of the recent species would reveal a still greater individua variation
I thought it better to consider the identity between both fossil
specimens as not totally proved. This may account for the addition
of a sign of interrogation to the name under which the fragmentary
cranium has been described.

As has already been emphasized on page 59, the resemblance
between the well preserved cranium and that of the recent Rhino-
ceros sondaicus is extremely close. Hereafter will, furthermore, be
pointed out that also concerning the various measurements a nearly
complete accordance appears to exist. Moreover it will be show
that also the permanent cheekteeth dentition resembles that of the
recent species down to the smallest detail. We shall, therefore,
be wholly justified if the fossil specimen in question be compared only
with the recent and fossil Malayan forms of the genus Rhinoceros s. s.,
and with the only Asiatic form which shows distinct affinities to
Kh. sondaicus namely, Rh. sivalensis Falc. et Cautl. In this manner
the following species remain:

Rhinoceros kendengindicus, Eug. Dubois, Tijdschr. Kon. Ned. Aardr. Gen., 2nd ser.,
XXV, 1908, p. 1259.

Our knowledge of this fossil Javan form must be called absolutely
insufficient. Dubois only mentioning — after shortly dealing with
Rh. sivasondaicus —: quot;Die zweite Form von Rhinoceros schliesst
sich dem nur auf dem Kontinente lebenden R. indicus an, ist von
diesem aber in untergeordneten Punkten verschieden. Das hintere
Joch der oberen Molaren ist relativ breiter, die Nasenknochen
sind schmaler und scharfrändiger. Ich unterscheide diese zweite
Art als Rhinoceros kendengindicus n. sp.quot; i). Though this is very
little, yet it suffices in the present case to exclude Rh. kendengindicus
from the comparison. There are certainly a number of resemblances
between the crania of Ri. unicornis L. (R/;. indicus Cuv.) and of
Ry^. sondaicus, and therefore also between those of Rh. kendengindicus
and our own fossil form. The cheekteeth of the latter, however,
are readily distinguished from those of RJj. unicornis (and therefore
also from Rh. kendengindicus) by the absence of crista and medi-
fossette, the latter formed by the union of crochet and crista

2)nbsp;See e. g. W. H. Flower, On some cranial and dental characters of the existing species
of rhinoceroses. Proc. Zool. Soc. of London, 1876.

Rhinoceros sivasondaicus, Eug. Dubois, Tijdschr. Kon. Ned. Aardr. Gen., 2nd ser.,

non Rhinoceros sivasondaicus, H. Stremme in: Die Pithecanthropus-Schichten auf
Java, 1911, p. 89, pl. XVII, fig. 8; pl. XVIII, figs. 1-2.

Originally Dubois (1892) mentioned the occurrence of Kb.
javanicus Cuv. {RJ}. sondaicus Desm.) in the fossil state in Java.
In 1908 he apparently redetermined this form as R^. sivasondaicus
n. sp. As this species is of extreme importance to us, we shall
•quote all that Dubois mentioned about it. RJjinoceros sivasondaicus

lebend in Java, aber auch auf dem Festlande vorkommt. Diese
lebende Art is sicher von R. sivaknsis wenig verschieden, und
Lydekker betrachtete deshalb letztere als ihre Stammform. Es ist
nun sehr merkwürdig, dass die Kendeng-Form den geringen
Abstand, welche noch jene beiden trennt, überbrückt. Gerade die
wenigen Unterschiede die Lydekker angibt sind in der Kendeng-
Form noch geringer geworden. Namenthch vermittelt die fossile
Art von Java in dem Verhältnisse zwischen Lange und Breite
der oberen Molaren den Uebergang jener beiden. Ich nenne sie
deshalb VJnnoceros sivasondaicus n. sp. Jedenfalls war diese eine
mit der heutigen doch nicht identische Art. Von R. swalensis
unterscheidet sie sich auch durch den Besitz eines zwar sehr
kleinen medialen Unterkieferincisiven, von R. karnuhensis durch
das Fehlen des Cingulum an der Innenfläche der Vorderjoche und
des Höckerchens in dem Quertal der oberen Molarenquot; (1. c.

The Trinil collection of Mrs. Selenka contained of rhinoceros
remains, besides some detached upper and lower cheekteeth and
limb bones, a fine, well preserved cranium. All the remains were
reckoned by Stremme to Kb. sivasondaicus Dub., mainly from the
fact, that he thought to notice some differences between the cheek-
teeth of his fossil form and those of VJj. sondaicus. Though Stremme
stated that his material did not enable him to detect a transition
between Kb. sondaicus and KJ). sivaknsis, yet he came to the conclusion
that his form was specifically identical with Kb. sivasondaicus Dubois.
And again we may repeat what we observed already concerning
Stremme\'s identification of his fossil forms of Buffelus and Bibos
with resp. Buffelus palaeokerahau Dubois and Bibos palaeosondaicus
Dubois, namely: the probability is great that Stremme\'s form is
identical with Dubois\'s form, but Stremme had certainly no right
to accept this identity as if it had been wholly proved.

Hereafter we shall have the opportunity to deal with Stremme\'s
determination more detailed.

Rhinoceros indicus fossilis. Baker and Durand Journ. Asiatic Soc. of Bengal, IV,

Rh^no Ji S;iSsisrR\'\'LydeS;r:W Geol. Surv. India, ser. X, II, part z. t88x.
p. 28, pl. V, figs. i, 3—6; pl- X, fig- 4-

The literature relating to this Siwalik species is more extensive
than would appear from the above, where only has been men-
tioned what will be needed in the below. Furthermore stress must
be laid upon the fact that of each of the papers quoted only the
figures of those specimens have been mentioned which appear
to belong with certainty to the present species.

The name RIj. sivalensis appears to have been first applied to
a number of specimens figured in the F.A.S. a.o. comprising a
damaged, but still rather fine, adult cranium, m which both sets

In 1836 Baker and Durand published a paper in which they
described and figured a.o. a well preserved cranium, an occiput and
several upper cheekteeth. Though all the specimens were assigned
to one species (R/;. indicus fossilis), the authors hinted at the pos-
sibility of.there being a second species. Lydekker twice referred
to some of Baker and Durand\'s specimens viz., m 1876 and 1881.
He redetermined the cranium and two M« as certainly belonging
to the present species. A M^, which he in 1876 also reckoned to
KJj. sivalensis, was in 1881 considered as possibly belongmg to
Rh. palaeindicus. Another of Baker and Durand\'s specimens,
namely a symphysial part of the mandible, was thought probably
to belong to RJj. platjrhinus. Furthermore, Lydekker pointed out
the resemblance between still another specimen, namely the occiput
already mentioned and the hinder portion of the cramum described
and figured by him undernbsp;sivalensis, vxt.gajensisYxYi. )

We see, therefore, that among Baker and Durand\'s material
of the fossil VJmoceros indicus not only a number of different spccics
are contained, but even two different genera.

I thought it desirable to mention these details to make what
follows comprehensible. Baker and Durand\'s paper contains a

») This form was again redetermined as Aceratherium gajense Pilg. by Pilgrim (Mem.
Geol. Surv. Ind., new ser. IV, 1912). Matsumoto (Science Rep. Tohoku imp, Univ.
Sendai, Japan. 2nd ser. (Geology), V, (1918—1921) rightly pointed out that according
to the law of priority the name of the spccics must be Acer, gajense (Lyd.).

lot of measurements among others of the cranium and occiput,
both already mentioned, and of about twenty specimens of upper
premolars and molars. From the above it will be clear that the
measurements of the cranium may safely be used. Those of the
occiput and cheekteeth i), however, cannot.

Apparently Stremme was not acquainted with these particulars,
his tables containing, besides the measurements of Baker and
Durand\'s cranium of Rh. sivalensis, also those of the occiput and

Except the cranial measurements of Baker and Durand\'s
specimen and a few dental measurements which will be dealt with
in the sequel, we do not dispose of more material of numbers.

A comparison between the cranium of our own fossil form
with that of RJj. sivalensis will be made together with the discussion
of the crania of the recent Rh. sondaicus. At this moment we shall
pay attention to the differences between upper M of Rh. sivalensis
and R}j. sondaicus. Lydekker stated: quot;Between the true molars
of these two species, taking into consideration the smal variation
which I have noticed in those of the fossil, I am totally unable
to discover more than one point in their plan of structure which
can be taken as affording any certain indication of distinction.
This point is a difference in the relative dimensions of the molars

the greatest length of the anterior surface, measurmg to the second
quot;costaquot; 4) of the quot;buttressquot; is exactly equal to the greatest length
of the external surface; whereas in R.javanicus the former measure-
ment is greater than the latter.quot;nbsp;.

Specimen however, was later on reckoned by Lydekker
to Rh. sivalensis, var. intermedins «). Though by doing so, one of

«) Pilgrim redetermined this tooth as belonging to Acer, gyense, var. mtermcdmm (See
Rec. Geol. Surv. Ind., XL, 1910, part 3, p. 200).

the two specimens of U. sivalensis, of which Lydekker gave the
„emLs, must be omitted yet we ^^ ^
difference mentioned, because Lydekker stated that m a l tne
JdSins he could procure, this telationshtp appeared to be

greatest length of the anterior surface will be equal to ui
«the greatest length of the external surfacequot;. It will,
amp;e be dear!that the distinguishing character,
wlh Sdekkek discovered, will be very difficult to handle

Now k ™ be remembered that the only character, which
DuBO^mLtTons of Rh. sivasondaicus, is that its molars are inter-
TL TnX respect. I may add, that if afterwards, BJj. stvason-

tvTtf admire Dubois for his not overlooking such a very subde
S As already stated, Stremme did not succeed m using th
character! As to\'the teeth of my own fossil form they are in this
respect indistinguishable from those of Kb. sondaicus.

Busk described and figured a left and right fossi^lized W of
rhinocems belonging to a species «not distinguishable by its dental
ctrXs frori. sondaiJquot; They were obtained from Sarawak
fBorner^ locality unknown. Lydekker provisiona ly referred
KtStwo or M^ of opposite «dquot; and t t c ^

Rhinoceros javanicus, F. Cuvier, Hist. nat. des Mammiferes, III, l.v. 45.nbsp;P-

Of this recent species the present writer was able to examine
and measure thirteen crania of adult and old individuals, being m

all thirteen specimens the M^ to a smaller or greater extent abraded
by wear. The National Museum of Natural History at Leyden
contained five specimens; the museums of Natura Artis Magistra
six, and the Zoological Institute of the University of Utrecht two.
In one specimen (no. lo of table K) the lower jaw could not be
removed without demolishing the specirnen That is the reason
why length and breadth of the cheekteeth of tks specirnen have
not been given in table N. In cranium n° 4 only left M^ and M3
were present. Cranium no. 6 wanted left P^-M^. Furthermore
from table N will be noticed that in four specimens F of either
side is absent. In nearly all cases, there are, however, clear
indications that this loss has taken place after the death of the

Before proceeding with a discussion of the individual variation
which exists both in cranium, and permanent upper cheekteeth
dentition, it will be desirable to call attention to the various tables

one cranium i) of Kb. sondaicus, borrowed from Franz Toula,
Das NaZnln Hundsheim, Abh. K. K. Geol. Reichsanstalt,

two crania of the same species, borrowed from Cuvier,
Recherches sur les ossements fossiles, 1822, p. 37»

As to the measurements themselves we followed for the greater
part Stremme who in his turn partly )oined Toula.

In table L all the values are expressed in % of measurenient 13.
The total length of the cranium could not be used as un.t because
of three crania the total length was unknown. Maxima and m.mma
have been heavy printed for Klj. sondaicus.

Table M coniains a number of relations of certam measurements.
As to the choice of the measurements we followed Stremme as

far as possible. For R/j. sondaicus the greatest and smallest number
of each horizontal row have been heavy printed.

In table N the length, breadth (both measured at the base of
the crown) and the relation length : breadth of each cheektooth
have been united. Why this table does not contain — in contradiction
to Stremme\'s table — measurements of premolars and molars of
R/?. sivalensis will be clear, I believe, after what has been stated
in the foregoing part. Originally we added to table N also the
measurements of two cheekteeth sets which will be found mentioned
on p. 39 of Cuvier\'s paper quoted. When it appeared, however,

considerably exceeded the highest value found in the corresponding
teeth of twelve foregoing crania it was thought better to exclude
them from our table.

As will be seen we have given all dental measurements in mm.
Stremme, however, in tenths of mm. In my opinion this is absolutely
superfluous. For, though the structure of a rhinoceros tooth certainly
permits exact measuring of the breadth, it surely does not allow
of measuring the exact length of the tooth. Moreover it must
not be forgotten that length as well as breadth of corresponding
teeth of opposite sides may differ distinctly.

In table O, at last, the breadth and length have been expressed
in % of resp. the breadth and length of the fourth premolar. Both
in table N and O maxima and minima, found for the teeth of
Rh. sondaicus have been heavy printed.

We shall now return to the measurements of the sixteen crania
of Rhinoceros sondaicus. Both from table L and M it will be seen
that even such a relatively small number of specimens already
may show a considerable individual variation. Especially concerning
some points there appears to exist noticeable differences between
the crania of the present species. In this respect we may bring
forward:

a. The great differences in degree of depression of the frontal
region; smallest in cranium n«. 5 (text fig. 7), greatest in cranium
no. 8 (text fig. 8). Both text figures show, moreover, clearly
that a feeble frontal depression coincides with slightly curved
nasals. Table L exhibits, furthermore, decidedly that whereas
the difference between the greatest and smallest value, found
for m. 6 in 16 crania of JL6. sondaicus is 18 %, the greatest
value is exceeded by no less than 22 % by the only cranium
of Rh. sivalensis.

b.nbsp;The considerable difference in breadth of the frontals, narrowest
in cranium n». 2 (text fig. 9), widest in cranium n«. i (text fig. 10).
The specimen of Rhinoceros sivalensis appears to have a still
broader forehead than the specimen of Ri*. sondaicus drawn in
text fig. 10.

c.nbsp;The surprising variability in development and distance between
the cristae fronto-parietales.

d.nbsp;The distinct differences in shape of the occiput; broad and
therefore relatively low in cranium 13 (text fig. 11), narrow
and comparatively high in cranium 3 (text fig. 12).

Notwithstanding the considerable, individual variation in these
16 crania of Rh. sondaicus, their measurements show, in general
decidedly that the specimen of Rh. sivalensis of table L must be
specifically distinct.

As to the individual variation of the dental measurements I
may refer to table N and O. Of the structure of the premolars and
molars of the recent species we may give the following summary.

pi rather small, but not deciduous. Deuterocone of P^ more
or less isolated. P^ in general entirely molarized; top of deuterocone
very seldom free. Deuterocone and tetartocone rather closely
approximated in all the P, especially near the base of the crown, so
that union takes place after prolonged wear. Postfossette distinctly
more shallow than medisinus; consequently very worn teeth only
present one pit, namely the buccal part of the medisinus. Crochet
generally well developed, occasionally double. No antecrochet. As
a rule no crista and medifossette. The crochet is well defined,
towards the base it becomes more blunt. Ectoloph with distinct
parastyle (protostyle), paracone (protocone) fold, and paracone
(protocone) style.

In M2 and M^ metastyle. Outer cingulum always absent; inner
cingulum either absent in the molars or sometimes represented
by a small tubercle at the entrance to the medisinus. Inner cingulum
may also be absent in the P; it is, however, mostly present in the
form of a very fine row of incipient tubercles. Seldom this row
surrounds the whole of the internal side. Sometimes a short row
of tubercles is situated in the vicinity of the entrance to the medisinus,
in other cases it is attached to the tetartocone, more often, however,
to the deuterocone. Anterior and posterior cingulum either
smooth or finely crenulated. In M^ posterior cingulum generally
represented by a distinct tubercle at the postero-external of the
crown. An incipient secondary enamel fold in the postfossette of
the premolars may occasionally occur.

Lastly I will bring to the attention the following noticeable
peculiarity. The right P^ of

be in the possession of a
well developed crista which
had regularly united with
the crochet, forming a medi-
fossette. It is remarkable
that none of the other
teeth of the cranium in
question show any trace of
a crista. By the kindness of
the director of the Nat. Mus.
of Nat. Hist., Prof. Dr.
E. D. van OoRT, I am

enabled to give in text fig. 14 an upper view of the specimen, made
after a photograph, taken for me in Leyden.

After having dealt with cranium and cheekteeth of Kh. sondaicus
so extensively, we shall try to answer the question:

Is the form, which Stremme described under the name of Rh.
sivasondaicus, in reality specifically distinct from Rh. sondaicus, or
will it be possible to identify Stremme\'s form with the recent
species with the help of the more ample materials of the latter
we had at our disposal?

After a detailed comparison of the fossil cranium of his collection
with that of an old d and a young ? of RJj. sondaicus, Stremme
concluded: quot;Weist so die allgemeine Schädelform nur Unterschiede
von der rezenten auf, die innerhalb der individuellen Variationsbreite
liegen können, so bestehen doch in der Bezahnung Abweichungen,
diedie Aufstellung einer neuen Art gerechtfertigt erscheinen lassenquot; i).
My own tables L and M show that the correctness of Stremme\'s
first supposition is entirely proved by the facts.

länglichere, schwach eingedrückte vordere Grube 2) und einen
schärferen, bei einzelnen Zähnen geteilten Spornquot;. 3)

2.nbsp;quot;... der erste Prämolar, der bei allen rezenten Java . . . Nas-
hörnern ... ein verkümmerter und schon bei nicht allzuhohem

Alter abgekauter Zahn war, ist hier bei der abgekauten Zahn-
reihe des fossilen Nashorns noch relativ stattlich und zeigt zwei
deutliche Grubenquot;.

3. When length and breadth of the cheekteeth of Stremme\'s form
and those of BJ. sondaicus were expressed in % of resp. length
and breadth of P^ it was shown that Stremme\'s form gave
on the whole greater values.

Concerning the presumed first and second difference, I am
convinced they will be invalidated much more rapidly with the help
of a comparison of the right toothrow of Stremme\'s specimen
with the corresponding set of cranium 5 of Kb. sondaicus and drawn
(after a photograph) in text fig. 13, than by means of a lot of
words.

As to the third difference I may refer to my own table O.
Though I immediately admit that still the breadth of P^, P^ and M^
of quot;Rh. sivasondaicusquot; show the greatest values, we may be absolutely
sure that also these differences would disappear, if but we had been
able to collect the measurements of some more crania of the recent
Rh. sondaicus.

As, moreover, the other rhinoceros remains, which Stremme
described, do not afford any reason for specific distinction I
conclude:

The fossil form described by Stremme under the name of Rh. siva-
sondaicusI^viB. is specifically indistinguishable from the recent Rh. sondaicus,
and must therefore be called Rhinoceros sondaicus Desm. fossilis.

At last our own fossil cranium (specimen a of the tables).
Both from the tables and from comparison of the description
of the specimen in question with our enumeration of cranial and
dental characters and peculiarities of Rh. sondaicus appears —without
leaving a shade of doubt — that also our form is specifically identical
with the recent Rh. sondaicus, and consequently also with Stremme\'s
specimen.

The terminology of primary and additional cusps of upper
premolars and molars, based upon the tritubercular theory of Cope—
OsBORN, was originally as follows:

Scott\'s i) investigations, however, lead him to the conclusion
that the cusps of the premolars were not homologous with the
corresponding ones of the molars. Accordingly he proposed a series
of new names, which are for the primary cusps of upper premolars
as follows:

Protocone analogous with paracone of molars
Deuterocone „nbsp;» protocone „

Scott was convinced that the conules of the premolars
were not homologous with those of the molars. (quot;In position
these conules correspond to the proto- and metaconules of the
molars, but are obviously not homologous with themquot;) 2). He

1) The Evolution of the Premolar Teeth in Mammals. Proc. Acad. Nat. Sci., Philadelphia,
1893, p. 405—444-
Loc. cit. p. 413.

committed, therefore, an incompleteness in not proposing new
names for the conules of the premolars. Osborn, too, failed to
do so.

I should not have been at all surprised if Osborn had not
accepted Scott\'s new names for the primary cusps of the upper
premolars. This assertion will be sufficiently explained by the
following quotation. Osborn in dealing with the subject of

based upon the actual homologies of the primary elements of the
trigon and trigonid, but in extending it to the other parts of the
crown and to the secondary cusps it was found that we must
apply similar terms to some of the later elements in the upper

otherwise the terms soon multiply, so as to become a burden rather than
a conveniencequot;

Perhaps the present writer should not have ventured to draw
attention to these points, were it not that Osborn had been in-
consistent in another point. At the same time that Osborn introduced
the terms proto-, para-, meta- and hypocone for the primary cusps
of the upper premolars and molars, he proposed namely some
new names for the peripheral pillars, which occur in the upper
cheek teeth of all ungulates. In view of the rhinoceros premolar
and molar there are two pillars which may interest us, namely
those which Osborn gave the names of parastyle and metastyle,
the prefixes (para-, meta-, etc.) being appUed . . according to

Osborn after acceptance of Scott\'s terms quot;proto-, deutero-,
trito- and tetartoconequot; for the primary cusps of the upper
premolars, should have named the antero-external and postero-
external pillar of the premolar resp. protostyle and tritostyle, in-
stead of parastyle and metastyle. He did not, however. See e. g.
figs. ii6 and 192 in Osborn\'s work quot;Evolution of Mammalian
Molar Teethquot;.

Some subsequent investigators — I may mention Zdansky )
and Coopéré) — apparently met with the same inconsistency,
for they use the term protostyle for the antero-external pillar
of the premolars. It is, however, an enigma to me, why Cooper —

1)nbsp;Evolution of Mammalian Molar Teeth. Biological Studies and Addresses, vol. I,

Otto Zdansky, Die Säugetiere der Quartärfauna von Chou-k\'ou-tien. Palacontologia

C. F. Cqoper, On the skull and dentition of the Paraceratherium bugtiense. Phil.

in whose specimens of premolars also the postero-external pillar
was developed — does speak of protostyle, and not of tritostyle
instead of metastyle

Turning now to the rhinoceros premolar and molar of the
upper jaw, we have only some remarks to add. After what preceded

co^e - f-^ ri^nbsp;we/a cor? e

rc/\'ocane)nbsp;f ^

nrai\'oc£gt;/?enbsp;inbsp;^/lypcconp

(c/ea/eroconejnbsp;/nbsp;(///ar/acoi^e)

Text fig. 15. Diagram of a hypothetic left upper check tooth of
rhinoceros showing the terms applied to the various components
of upper premolars and molars. (The terms which only refer to
premolars are placed in parenthesis.)

it will need no explanation why we used, in describing the premolars
oi Khinoceros sondaicus fossilis, the terms protostyle fold and protocone
style. As to the greater part of the remaining terms I may refer
to text fig. 15. It will only be desirable to pay attention to Osborn\'s
terms praesinus, medisinus and postsinus. As far as my knowledge
goes OsBORN gave twice viz., in 18982) and 19078), a table in
order to show the parallelisation between his terms and those of
former authors. Of course hereafter only the terms applied to
the rhinoceros tooth will interest us. His table of 1898 contains
the terms medisinus and postsinus, which he regarded as identical
with resp. anterior and posterior valley of the English authors

[Boyd Dawkins (1867), Busk (1877), Lydekker (1881) i), Foote
(1874)]. According to the same table the authors above mentioned
should not have used terms identical with Osborn\'s prae-, medi-
and postfossette. Without any further explanation Osborn speaks,
furthermore, dealing in the text 2) with the rhinoceros molar of
quot;the three inner valleys .. . presinus, medisinus and postsinusquot;.
In his table of 1907, he sticks to his opinion concerning prae-,
medi- and postfossette, but now only praesinus and postsinus are
mentioned and regarded as identical with anterior- and posterior
valley of the EngUsh authors. There appears to exist, therefore,
a contradiction. But besides this, Osborn\'s tables are incomplete,
which may appear from what follows. Lydekker gave on p. 8
of his paper of 1881 3) an enumeration and explanation of the terms
applied by him to the various components of the rhinoceros molar,
from which we borrow:

quot;Anterior valley = hollow in front of anterior collisquot; (ant.
collis = protoloph).

quot;Posterior valley = hollow behind posterior collisquot; (post, collis
= metaloph).

quot;Accessory fossette = separate pit on the worn crown cut off
from the outer extremity of the median valleyquot; (and some sentences
further) quot;caused by the union of the crochet and combing-platequot;
(combing-plate = crista).

Furthermore, it may be mentioned that Lydekker did not
distinguish rigorously between the term valley and fossette, which

would present two fossettes on the crown (formed by the outer
part of the median valley and by the posterior valley)quot; With
these data it is not difficult to introduce the following correc-
tions :

Osborn mentions 1882, but apparently meant 1881 as I do not know a publication
of Lydekker on rhinoceroses of 1882.
Osborn, 1898, p. 105.
5) Siwalik Rhinocerotidae, Mem. Geol. Surv. of India, ser. X, II, part i, 1881.
*) Loc. cit. p. 16.

Osborn i) meant by postsinus the vestige of the primitive
valley, which may occur in the third upper molar of some species
between the ectoloph and the posterior surface of the metaloph.
This pit, therefore, is apparently considered as not homologous
with the postfossette of the remaining premolars and molars.
I cannot, however, see the reason. What Osborn meant by the
praesinus of the rhinoceros molar I have not succeeded in detecting.
I believe, however, that it is identical with the anterior valley of
Lydekker and other English authors.

The genus Hippopotamus is represented in our collection by a
number of remains which all — with the exception of one specimen—
have been excavated NNW. of Bumiaju at the base of the vertebr-
ate beds. The exact locality of each specimen will be mentioned
in dealing with each remain separately. After the description of
all the specimens we shall draw a comparison between the remains
of our collection and the fossil specimens which hitherto have
been found in SE. Asia, and Java.

At the back broken off some cm. behind the posterior extremity
of Mg. All the I are lost; three alveoli in the anterior border of the
ramus mark distinctly the presence of three I in this half of the
mandible. C is partly preserved, being broken off about 4 cm. above
its alveolar margin. Pj is lost; a dark coloured spot in front of the
P2 indicates, however, its alveolus filled with matrix. For the rest
the whole dentition is present. M^ is slightly damaged on its postero-
internal angle. The posterior half of Mg is for the greater part
wanting.

The premolars, except P]^ which probably was one-fanged,
consist of a single cone, supported by two fangs. Pg, which is
hardly worn, points strongly forward, so that — whereas the
fangs of P2 and Pg are closely approximated — the interval between
the top of P2 and the middle of the worn summit of Pg is not less
than 4 cm. In P2 and Pg a relatively slightly developed cingulum
occurs on the anterior and internal side, but not on the external

and postetior side. In the succeeding P4 also on the posterior side
the cingulum is distinctly developed.

Mjl is very much worn. Internal and external cingulum are
totally wanting. The cingulum is only present on the anterior side.
On the posterior side it occurred probably too, but must have dis-
appeared by the process of trituration. The degree of wear of Mg
is also considerable. In front the dentine islets of protoconid and
metaconid i), and in back those of hypoconid and entaconid, have
wholly become confluent, while even a narrow connection already
occurs between the dentine of anterior and posterior lobe. M2
shows a well developed cingulum on its anterior and posterior
side; internal and external side are — just as in M3 — devoid of
a cingulum. M3, the posterior extremity of which is lost, shows
a thick anterior cingulum and an undamaged, worn, anterior lobe,
clearly consisting of two rather regular trefoils, which are already
united for the greater part. These trefoils are formed by the fact
that each cone is pinched in by vertical grooves on its anterior and
posterior face. Of the second lobe only the enamel of the front
side of hypoconid and entaconid has been preserved. These remnants
show, however, that hypoconid and entaconid also present trefoils
after wear.

The enamel of molars and premolars is smooth and does not
show striations or other sculptures.

The mandibular toothrow is approximately straight from Mg
up to and including P3. Its frontmost portion, however, formed by
P2 and Pl, is considerably curved outwards.

The horizontal ramus shows the typical form of the corres-
ponding part of a hippopotamus mandiljle and need no detailed
description. It will suffice to refer to the measurements and to
figs. I and 2 of pl. VI in which upper and front aspect are given,
and to fig. i of pl. XIX in which a side view is represented. Stress
may be laid, however, upon the occurrence of three I. The alveoli
of Ij and I2 are of subequal size, that of I3 is smaller than either of the
others. The lower alveolar margin of Ig is situated somewhat above
the alveolar border of I^ and distinctly above that of I3. C is too
much injured, to fix its form in cross-section. Its enamel shows
regular longitudinal striations.

\') Osborn in his well-known work „Evolution of Mammalian Molar Teethquot; nowhere
deals with the hippopotamus cheek teeth separately, so that there exists some uncer-
tainty in respect of the terminology of some of the cusps. I have applied the term
metaconid (and not paraconid) to the antero-internal cusp from the evidence of the
terms occurring in Osborn\'s text fig. ijz in which are represented some lower teeth
of a primitive artiodactyl.

The M3 of the specimen in question having been fully in use,
we may be sure that it belonged to an adult individual.

The left half of the present specimen is in general better preserved
than the right one. Of the left half the anterior and posterior surface
of the ramus are injured, the top of the coronoid process is broken
off, while — which is very unfortunate — the descending plate
of the ramus, so peculiar to the present genus, is for the greater
part lost. As will be known this descending plate apparently furnishes
specific characters i). Of the right half the anterior border of the
ramus is damaged and the greater part of the lower posterior moiety
of the ramus has disappeared. The coronoid process, however,
is remarkably well preserved.

The left C, apart from small injuries, is complete. The right
C is broken off about 3—5 cm. above its alveolar margin. As a result
of the considerable injuries of the front border of the mandible
a good deal of the alveolar portion of both canines have been laid
open on their anterior face. Left and right P^ are wanting. Just as
in the foregoing specimen the alveole of P^ is filled with matrix and
is indicated by a dark coloured, circular spot in front of P2. In
all probability Pj was therefore one-fanged. Of the left Pg only
the posterior fang is preserved while the alveolus of the anterior
fang is again marked by a dark coloured, circular spot in the bone
material. The right Pg is broken off close to the base of its crown;
the two fangs and the enamel of the posterior side of the crown
are present. The top of the left P3 is injured; the right Pg has dis-
appeared. The left P4 is undamaged; the right P^ injured on its
external border. Mj, M2 and Mg of either side are present, all injured
to some extent, but, in general, well preserved. I, in situ, do not
occur. To this mandible belongs, however, a detached I which is
represented in pl. XVIII, fig. 3.

The resemblance between the premolars of the present lower
jaw with those of the foregoing specimen is so close that a des-
cription would be a needless repetition.

All three molars are considerably abraded by wear; M^, of
course, most of all. The structure of the molars is very similar to
those of the specimen just described. They show the typical pattern
of a worn hippopotamus molar. Mj, M^ and Mg possess a strong
anterior and posterior cingulum; both Mg show the common
feature of a strong posterior talon, consisting of three to four
cusps. External and internal cingulum are either totally wanting
or very slightly indicated. If we have a close look at the grinding
surface of the left Mg, we can state that, whereas the worn metaconid,
protoconid and hypoconid present a more or less regular trefoil-
shaped dentine islet, the entaconid is not, or hardly pinched in by
vertical grooves on its posterior and anterior face. The same feature
will be seen in the right Mg and in the left and right Mg. (M^ is too

The enamel of the molars is smooth and without striations or
other sculptures. Each toothrow is in its posterior half approximately
straight; its anterior half is clearly curved outwards.

The degree of wear of the cheek teeth is somewhat less than
of the foregoing specimen. As Mg, however, has been fully in use,
the mandible in question certainly belonged to an adult individual.

The detached I, which is broken off 126 mm. from its top, is
compressed in cross-section, greatest and smallest diameter being
resp. 24 and 18 mm. The same shape in cross-section is maintained
throughout the posterior half; towards the top a very gradual
tapering occurs. The whole tooth is, furthermore, slightly bent.
Its top is obliquely truncated on the convex surface. The thin enamel
is smooth. As we do not know, whether in the present fossil form
the incisors were abraded on the external or internal surface, it is
impossible to decide whether the incisor under consideration must

As has been already mentioned the anterior surface of the man-
dible is considerably damaged, especially the upper third. It will,
furthermore, be remembered that the former place of P^ of either
side is marked by a dark coloured spot, representmg the alveole
filled with matrix. It will be clear, that if the anterior surface of the
mandible exhibits dark coloured spots with a regular form, these
represent either foramina or alveoli of incisors, filled with matrix.
The right half of the anterior surface showed indeed dark coloured
spots in a light coloured material. A microscope section revealed
that the light coloured material consisted of calcined bone, the
dark coloured material of a tuffoid substance. As the damaged,
anterior surface of the mandible was very irregular, it was very
difficult to decide whether the dark coloured spots were regular

in shape or not. The regular outline of these spots was, however
splendidly revealed by grinding a plane to the upper portion of
the right half of the mandible i). The result can be seen in PI. XIX,
fig. 2. If this figure be compared with fig. 2 of pi. VI it will
easily be seen that the three uppermost spots represent the alveoli
of I^, I2 and I3, while the lower spot marks the place of a foramen
mentale. In consequence of the grinding and the inward slope of
the grinded plane, the three uppermost spots cannot be taken as
representing the real oudine of the incisors at the alveolar margin.
They do show, however, the mutual position of the incisors, reveal-
ing that also in this respect there is a great resemblance between
the present lower jaw and the specimen described above.

C are pearshaped in cross-section; the top of the pear is directed
backwards and inwards. The left canine, which is nearly complete,
shows the usual strong curvature; it is obliquely truncated on its\'
posterior surface over a considerable distance. Due to the injury
of the ramus the quot;fangquot; of both C can be seen, the posterior surface
of the fang of the left canine exhibits a shallow broad groove.
The posterior surface of the right canine is in too damaged a condi-
tion to show this feature. As in the foregoing specimen the enamel
of the canines is provided with fine regular, longitudinal striations.

The lower jaw itself does not present pecuUarities which need
description. Its form may be easily judged from the figures. In
what follows we shall have an opportunity of reverting to the
relations between certain of its measurements.

Very damaged specimen, on either side broken off some cm
behind the posterior extremity of Mg. The left half still possesses
a portion of the descending plate, expecially a good deal of its
anterior border. The upper surface of the anterior half of both
horizontal rami and the whole front border of the mandible are

1) Fig. 2 of pl. VII shows the original state in which the specimen was received.
The specimen in question was contained in one of the cases (n°. 17) of the fourth
sending. Though it did not occur on the list of specification, we may be sure that it
has been secured in one of the excavations near Bumiaju, as the case referred to
contained only specimens of various excavations near that village. Furthermore it
is not improbable that the specimen under consideration, which showed the number i*
has been found in excavation 13.nbsp;quot;

very considerably injured. Due to this fact both C, all the I, P^—P^
of either side are wanting. Even no remains of the alveoli of the

Of the cheek teeth dentition only fragmentary left and right M^,
left and right M^, (right one injured) and right M3 are preserved.

The whole fragment, furthermore, has been subjected to a
crushing force, which fact is clearly shown by the unusual obhque
position of the left ramus with reference to the right one. M3 having
been in use the mandible in question belonged to an adult specimen.

If the remaining molars be compared with those of the foregoing
specimens it will be seen that there are no essential differences in
structure. Only the manner of wear is somewhat different. Both M^
are in too damaged a condition and too incomplete to show much
of their structure. Both Mg and the M3 show a strong anterior and
posterior cingulum while — as in the molars of the foregoing
specimens — external and internal cingulum are practically absent.
Just as in the molars of the mandible described just before, the
entaconid presents a simple structure, its anterior and posterior
face not being pinched in and, therefore, not presenting a trefoil-
shaped dentine islet.

The present specimen consists of the detached corresponding
parts of left and right maxillary, each containing Mi-M«.

In both fragments the structure and the degree of wear of the
cheek teeth arc exactly the same. The state of fossilization and the
colour are also strikingly ahke. The cheek teeth of the one fragment
being furthermore almost exactly the reflected image of the other,
we may be sure that both fragments belong to one another. It was,
indeed, possible to match both fragments with a tolerable degree

The reservation of the teeth is splendid. The degree of wear
is considerable, M^ being ground down near to the^ bottom of its
transverse valleys. The four cusps of M« are abraded by wear into
distinct trefoils. The trefoil-shaped dentine islets of protocone and
paracone in front, and those of hypocone i) and metacone in back,

\') Perhaps this cusp ought to be called: metaconule, for Osborn (Evolution of Mamnialian
ISkr Tee h ?oo7 p 173) mentions: „In all Artiodactyls the metaconule is very large,
oSn repSdi\'com^^^^^^^^ the cingulum-hypoconequot;. See also the note on p. 83.

have, however, not yet become confluent. The structure of M^,
M2 and M^ is very similar and relatively simple.

Each tooth consists of four cusps, the inner ones are wider
than the outer ones, the longitudinal valley being consequently
situated buccally of the longitudinal median line of the tooth. Each
cusp wears into a rather regular trefoil, anterior and posterior face
of each cusp being pinched in by vertical grooves. A cingulum
surrounds the whole of the crown. It is especially strongly developed
on the anterior and posterior surface, giving rise, in M^ and M^,
to something which may be called resp. front and back talon!
Inner and outer cingulum are on the whole crenulated; they obstruct
the entrances to the transverse valley. The enamel is in general
smooth and not sculptured.

spec. a. Locality: Bumiaju, Excavation 9, Pl. V, fig. 5.
spec, b—e. „nbsp;„nbsp;„ 8.

The pearshaped cross-section of spec, a and b, together with
the fact that they are truncated of the concave side, show that
they represent lower canines. Specimen a is figured, being the
one which is best preserved. Its left side shows but a very slight
curvature across, whereas the right side is strongly convex in the
same direction. The left C of the lower jaw described on page 84
has its outer side strongly curved proving that spec, a represents
a right C. All the specimens are strongly curved; the top is obliquely
truncated and the enamel provided with fine, regular, longitudinal
striations, which are parallel to the curvature of the tooth. Spec, b
is a left lower C. It differs from spec, a in having on its inner side
a shallow, broad, longitudinal groove which occupies the middle
portion of the inner face along the whole length of the tooth.
Spec, a shows a similar groove on its posterior surface. Whether
this groove is also present in spec, b cannot be ascertained as the
whole posterior surface has been restored with plaster.

Spec, c—e are longitudinal halves of lower C. They need no
detailed description, as their resemblance to the foregoing specimens
is very close.

This specimen represents that part of the cranium situated
behind a plane which contains the (broken) tips of both zygomatic
processes of the squamosal and which intersects the frontals at a
distance of 14 cm. from the occipital crest.

The preservation of the present fragment is exceedingly fine.
Only the left petrosal has fallen out and the right paroccipital
process has broken off.

The degree of fossilization is notoriously small; the holes between
the spongiosa of the bone are unfilled and the surface of the bone
has a very recent appearance. Of our whole collection in the specimen
under consideration the process of fossilization has least progressed.
Of course, this need not to be regarded as an indication of a young
age. FossiHzation does not depend upon mere time.

The occipital surface is comparatively broad and low. Regarded
in hind aspect, the occipital crest is strongly and regularly convex.
The sagittal crest is thin in its anterior extremity, that is to say,
at the point where the cristae fronto-parietales meet. From that
point the sagittal crest becomes gradually wider towards the occipital
crest, thus exhibiting a very gradual transition into the latter.
Both cristae fronto-parietales are thin, sharp ridges. The braincase
is narrow, the curvature of the parietals being from above downwards
but slight. As will be seen from PI. XVIII, fig. 2, the frontals extend
far backwards, so that only a narrow space remains between the
cristae fronto-parietales and the suture between frontals and
parietals. PI. XVIII, fig. i shows that if the occiput is placed
vertically, the upper surface of the braincase slopes distinctly
forwards.

The specimen in question probably belonged to an immature
individual, as the various bones are not yet ankylosed, but separated
from one another by clearly visible sutures.

Hippopotamus (Hexaprotodon) sivalensis, H. Falconer and P. T. Cautley Asiat
R^arches, XIX, 1839, p. 39 (reprinted m Pal. Mem. I, 1868, p. 130); Fauna Antiqua

Sivalensis, 1845, pl. LIX—LXI, pl. LXII, figs, i—10, pl. LXIII—LXVI (name
and figures); Pal. Mem. I, 1868, p. 142.

Hippopotamus sivalensis, R. Lydekker, Mem. Geol. Surv. India, ser. X, III, 1884,
p. 37; Cat. Foss. Mamm. Br. Mus., II, 1885, p. 297; G. E. Pilgrim, Rec. Geol. Surv.
India, XL, 1910, p. 202 and 203, Rec. Geol. Surv. India, XLIII, 1913, p. 324.

Hexaprotodon sivalensis, W. D. Matthew, Bull. Am. Mus. Nat. Hist., LVI, 1930, p. 557.

This species was founded by Falconer and Cautley on the
evidence of an almost complete cranium from the Siwaliks. It was
figured in the Fauna Antiqua Sivalensis (pl. LIX, figs, i—\\b),
together with numerous specimens of crania, lower jaws, vertebrae,
limbbones, etc., apparently all obtained from the Siwahk Hills. The
name Hexaprotodon was originally appHed in a sub-generic sense,
but was — according to Lydekker — adopted as a generic term
by Owen. Lydekker pointed out in 1884 that among the material
of H. sivalensis, preserved in the British Museum, probably two
different varieties were represented, which he provisionally dis-
tinguished as, var. latidens and var. augustidens, only on the ground
of the different breadth of the upper true molars. The type specimen
of H. sivalensis was reckoned by Lydekker to var. latidens. There
existed some more cranial differences between both varieties, but
two other crania, which showed transitional characters concerning
the latter differences, prevented Lydekker from considering var.
augustidens as a distinct species. From the catalogue of the Br. Mus.
in which — as will be known — Lydekker gave his last, definite
opinion it is not difficult to infer that he thought it better to cancel
both varieties.

Besides extensive references to specimens of the Br. Mus.
Lydekker shortly described and figured in his paper of 1884 a
fragmentary lower jaw from the Punjab, which he had originally
considered as possibly belonging to H. iravaticus Subsequent
examination of the original specimens on which the latter form
was founded, convinced him that the lower jaw under consideration
was too large and had too short a symphysis to belong to H. iravaticus.
Lydekker thought it, however, not improbable that it would be
identical with the form which he had provisionally termed K.
sivalensis, var. augustidens.

In a faunal list given by Pilgrim in 1910, the latter variety is
said to occur in the Middle and Upper Siwaliks of the Salt Range
(near Asnot and Dhok Pathan), and H. iravaticus in the Middle
Siwaliks of Yenangyaung (Burma) and Asnot. In a subsequent
paper Pilgrim (1913) recorded only the latter form from the Dhok
Pathan Zone (mid Middle Siwaliks), and not a single mention is
made of H. sivalensis, var. augustidens.

Matthew, who in the winter of 1926—1927 had the opportunity
of making a critical re-examination of the type collections of the
Siwalik fauna preserved in the Indian Museum in Calcutta and in
the British Museum in London, records only H. iravatkus from
the Middle Siwaliks i) and H. sivalensis from the Upper Siwaliks 2).
He apparently does not accept any variety of the latter species.

As to the specific characters of the present form we may quote
the summary which Lydekker gave in 1885 viz.,

quot;This species is somewhat smaller than the existing K amphibius,
and has three incisors in both jaws, those of the mandible being
of subequal size; the groove in the upper canine is much deeper
than in H. amphibius, and the mandibular symphysis relatively
longer.quot;

As will be known, Lydekker included in the genus Hippopotamus
both Hexaprotodon and Tetraprotodon Falc. et Cautl. His reasons
were that quot;Tetraprotodon\'\' palaeindicus, appeared in reahty to be in
its lower jaw a Hexaprotodon while, moreover, an instance of
unilateral hexaprotodontism mH amphibius^) was known. Matthew
remarked that the only character he could find to separate Hexa-
protodon is the presence of six subequal incisors. This being the
state of affair, it is, of course, merely a matter of taste whether or
not one will use the term Hexaprotodon.

Hippopotamus (Hexaprotodon) iravaticus, H. Falconer ^d P. T Cautlcy, Fauna
Antiqua Sivalensis. 1845. pl- LVII, figs. lo-u. pl- LXXXIII fig. 12 (name and
figures); Pal. Mem. I, 1868, p. 142; Pal. Mem. II, 1868 p. 406.

Hippopotamus iravaticus, R. Lydekker, Cat. Foss. Mamm Br. Mus., II, 1885. p. 309;
G. E. Pilgrim, Rcc. Geol. Surv. India, XL, 1910, p. 196 and 203; Rcc. Geol. Surv.

Hexaprotodon iravaticu^ H. Falconer, Journ. Ac. Nat Sci. Philad., ser. 2,1,1849. p. 237;
W. D. Matthew, Bull. Am. Mus. Nat. Hist., LVI, 1930. P- 557-

The present form was founded by Falconer and Cautley on
the evidence of two specimens of the symphysis, obtained from
Burma. They will be found figured in pl. LVII of the Fauna Antiqua
Sivalensis. In fig. 12 of pl. LXXXIII of that work the distal end
of a radius is figured under the name of Hippopotamus (Hexaprotodon)
iravaticus. According to Lydekker this specimen has also been
derived from Burma. Finally Falconer reckoned to the species

From the evidence of H.palamdkus, the lower I^ of which is about to disappear,
Lydekker concluded that also in H. ampUbius it is the second incisor which is wanting.
But here Lydekker was wrong as subsequent investigations showed that the very
young milk dentition still possesses the third upper and lower incisor and that it is
I3 whkh disappears (See Max Weber, Die Saugetiere, II, 1928, p. 546).

in question the distal end of a radius and femur which he supposed to
have come from Ava (Burma) i). As has already been stated under the
heading of K sivalensis, according to Pilgrim the present form occurs
in the Middle Siwaliks of Burma and of the vicinity of Asnot (Hasnot).
The occurrence in the Asnot beds, however, appears to be open to a
certain degree of doubt. Pilgrim (1913) stating on p. 300 ,. the ...
species seems to have been found in the Hasnot beds,.. .quot;2). Without
determining the species. Pilgrim (1913) mentioned, furthermore, the
abundant occurrence of hippopotamus remains in the Tatrot beds
(lowermost Upper Siwaliks) of the Salt Range. According to Mat-
thew 3) the latter material is probably comparable with K iravaticus.

With the exception of a note pubhshed in Pal. Mem. II (p. 406),
in which it is stated that quot;K iravaticus is a size larger than H. libe-
r/mj-zV,quot; neither Falconer nor Cautley left any further information
concerning the species in question.

In 1884 Lydekker pointed out the differences which he believed
to see between H. iravaticus and K sivalensis, and which he considered
important enough quot;... as to leave little doubt of the specific
distinctness of the small Irawadi hippopotamus ...quot;. In the Catalogue
of the Br. Mus. he gave the following summary:

quot;This species is slightly larger than the existing H. liheriensis,
and has three mandibular incisors; the cranium being unknown.
The outer lower incisor is laterally compressed, and the mandibular
symphysis apparently relatively longer than in any of the foregoing
species (ff. sivalensis, H. palaeindicus, and H. namadicus). Other
specimens are required to indicate its full affinities.quot;

It remains with us to mention that according to Lydekker
the type specimen (F.A.S., pl. LVII, fig. 10) belonged to a sub-adult
specimen, P^ being half protruded. The second specimen of the
symphysis is in my opinion entirely worthless, being too fragmentary
and figured on too small a scale.

Hippopotamus (Hexaprotodon) namadicus, H. Falconer and P. T. Cautley, Fauna
Antiqua Sivalensis, 1845, pl. LVll, fig. 12, pl. LVIII, figs. 1—3 (name and figures);
Pal. Mem. II, 1868, p. 406.

Hippopotamus namadicus, R. Lydekker, Mem. Geol. Surv. India, ser. X, III, 1884,
p. 45; Cat. Foss. Mamm. Br. Mus., II, 1885, p. 294.

Hexaprotodon namadicus, W. D. Matthew, Bull, Am. Mus. Nat, Hist., LVI, 1930,
P- 557.

The Asnot beds belong to the Dhok Pathan Zone (middlemost Middle Siwaliks).
1930, p. 557-
0 1885, p. 309.
0 1884, p. 42.

and Cautley. Falconer mentioned only that it was larger than
H. amphibius or K sivalensis, and that it had wellmarked specific
characters. In the F.A.S. are figured under the present specific
name four fragmentary specimens of the mandible which show
considerable variation in size. In 1884 Lydekker mentioned the
following specific characters.

quot;The incisors are three in number on either side; the second
being smaller than either of the others, and its alveolus placed
slightly above the level of those of the other twoj this character
is more marked in some specimens than in others ..The mandibles
figured in the F.A.S. quot;... are readily distinguished from the
mandible of K sivalensis by the more abrupt angle at the inferior
border of the anterior extremity, and by the relatively shorter
symphysis.quot; One would be inclined to infer from Lydekker\'s
words in the Catalogue of the Br. Mus. that a year afterwards he
did not attach much value neither to the mutual position of the
lower incisors nor to the relatively smaller Ig and to the abruptness
of the angle formed by inferior and posterior border of the horizontal
ramus, as may appear from the following quotation:

quot;This species is somewhat smaller than the last (K palaemdtcus,
which he considered as fully equal in size to K amphibius) quot;and
disdnguished by having three small and subequal-sized mandibular
incisors. The mandibular symphysis is slightly longer than in the
last species.quot; As appears from my own observations of twenty
three lower jaws of H. amphibius, the characters, the specific value
of which Lydekker apparently doubted, are indeed subject to
too strong an individual variation to be very reliable.

Remains which can certainly be referred to the present form
have been obtained only from the Narbada (Nerbudda) River beds.

? Hippopotamus (Tetraprotodon) palaeindicus, H. Falconer and T- Camley, Fauna
Antiqua Sivalensis, 1845. pl- LVII, figs. 6-9. pl- LVIII, figs. 4-10, pl. LXII,
figs. II—12. H. Falconer, Pal. Mem. I, 1868. p. i47- .

Tetraprotodon palaeindicus, H. Falconer, Journ. As. Nat. Sci. Ph.lad., ser. 2,1,1849.

Hiroopoiamus palaeindicus, R. Lydekker. Mem. Geol. Surv. India, ser. X III, 1884,
p 44; Cat. fIss. Mamm, Br. Mus., II, 1885, p. 293; W. D. Matthew, Bull. Amer.
Mus. Nat. Hist., LVI, 193°. P- 55^-

With the exception of the four specimens of the mandible,
which Falconer and Cautley reckoned to the foregoing species,

they considered all other hippopotamus remains of the Narbada
deposits as belonging to H. palaeindkus. Among the latter remains
there is only one symphysial part of the lower jaw which still possesses
the lower incisors. It is represented in pl. LVII, %. 5 of the F.A.S.
and must be taken — as Lydekker pointed out — as the type!
The remaining specimens comprise fragmentary crania (in which
the praemaxillaries are wanting), vertebrae and limbbones. I wholly
agree with Lydekker, who stated that there is apparently not the
slightest reason for assigning these specimens to K palaeindicus
as Falconer and Cautley did, and not to the other Narbada form
If. namadicus. The same holds good with regard to a damaged
cranium from the Narbada and noticed by Falconer in the Catalogue
of the Museum of the Asiatic Society of Bengal i).

Falconer and Cautley never pubhshed any detailed description
of the species. It was again Lydekker, who suppHed this want
This author pointed out % that the restoration of the type mandible
in the F.A.S. as simply tetraprotodont is most probably incorrect.
He figured namely a symphysis of a hippopotamus mandible
obtained from the Narbada, which resembled the type specimen
in every feature except that between the two large incisors, occurring
in each half of the type specimen, a very minute teeth was wedged
evidendy corresponding with Had this specimen been the only
instance of hexaprotodontism of quot;Telraprolodonquot; palaeindicus the
present writer should certainly not have been convinced of the
hexaprotodont character of the present form. For, as will be seen
in dealing with twenty three skulls of the recent H. amphibius,
the lower jaw of this form shows in some cases notorious abnormi-
ties 3). Lydekker mentioned, however, that the British Museum
after Falconer\'s death acquired two similar mandibles from the
Narbada, which exhibited a precisely similar arrangement of the
mcisors. As, moreover, the young milk dentition of H. amphibius
is characterized by the presence of upper and lower I3, there cannot
remain any doubt with regard to the invahdity of the term quot;Tetra-
protodonquot;.

In the Catalogue of the Br. Mus. Lydekker gives the following
specific characters of this Narbada form:

quot;This species is fully equal in size to the existing H. amphibius.
I^he number of upper incisors is unknown; but there are three in
the mandible, the middle one being very small, and perhaps dis-

appearing in the adult; the other two are large, and equal in size.
The mandibular symphysis is extremely short.quot;

Like H. namadicm, the present species is only certainly known
from the Narbada (Nerbudda) River beds.

Hexaprotodon sivajavanicus, Eug. Dubois, Tijdschr. Kon. Ned. Aardr. Gen., ser. 2,

Hip^potaTus £vajatlnicus,.W. Soergel, Palaeontographica, Suppl. iv, Abt. iii,
Lief. I, 1914, p. 2.

Dubois reckoned the fossil hippopotamus remains of his collection
to a new species of which he gave a very short notice, which will

quot;Es ist ... ein Hexaprotodon, wie die fossilen Flusspferde des
indischen Festlandes. Sie entfernt sich jedoch weit von den beiden
Narbada-Arten: Hexaprotodon namadkus und H palaeindicus, durch
die Länge der Unterkiefersymphyse und die relative Grösse der
Incisiven, nähert sich in diesen Punkten sowie in der Schädelform
sehr beträchdich H. sivalensis, von der sie sich, ausser durch ihre
etwas geringere Grösse, durch die im Verhältnis zur Länge grössere
Höhe zwischen den Gaumen- und Nasenbeinen, durch den relativ
breiteren Oberkiefer, und besonders auch durch die nicht geradlinige
Lage der Austretungsstellen der unteren Incisiven unterscheidet.
Sie steht der Siwahkform deudich sehr nahe, und während sich
die Narbadäformen in der überwiegenden Entwicklung des
lateralen Incisiven als Endzweige des Hexaprotodon-St^mmts zu
erkennen geben, stand Hexaprotodon sivajavanicus n. sp., sogar noch
H sivalensis etwas übertreffend in dem relativ stärkeren medialen
Incisivus, dem Stamme des afrikanischen Flusspferdes sicher noch

quot;quot;^^Sotrgel determined a collection of fossil mammalian remains
collected by J. Elbert in the Kendeng beds of Java. Though he
gave the results of all the determinations, only the stegodonts were
fully described and figured. Hippopotamus was represented in
Elbert\'s collection by two lower jaws and one cranium (all the
•three without incisors and camnes, for the greater part also without
molar teeth); furthermore by one specimen of Mg, fragmems
of C and I some limbbones and vertebrae. Apparently Soergel
determined\' all these remains as belongmg to Hippopotamus siva-
javanicus. Of course, Soergel\'s determinations are of not the
shghtest value to us; they have only been mentioned here for the
sake of completeness.

From Java. H. Stremme in: Die Pithecanthropus-Schichten a\\:if Java, loii, p. 104.
H. G. StehHn, Wet. Med. Dienst Mijnb. Ned. Ind., 1925, p. 3.
J. Cosijn, Verh. Geol. Mijnb. Gen. v. Ned. en Kol., Geol. ser. IX, 1931, p. 119.

Non M. Schlosser, Abh. Math. Phys. Kl. K. Bayer, Ak. Wis., XXII, 1903, p. 95,
(from China).

The collection obtained by the Trinil Expedition of Mrs.
Selenka contained four upper true molars, partly damaged, two
right astragali of hippopotamus and a detached tooth, considered
as possibly representing an incisor of the present genus. The speci-
mens were described and well figured by Stremme. They did,
however, not enable him to determine the species.

Stehlin who determined a collection of fossil mammalian
remains obtained from Kali Biuk and Kali Saat (vicinity of Bumiaju),
mentions some fragmentary limbbones and a portion of the
pelvis as belonging to Hippopotamus. Interesting is that the distal
end of a radio-cubitus indicated the presence of a dwarf individual,
whereas the other specimens pointed to individuals of the size of
small specimens of H. amphthius. Whether or not the dwarf individual
belonged to another species Stehlin could not determine.

Cosijn detected a new findspot of fossil vertebrate remains
N. of Djetis and Perning (Central Java). Umbgrove, to whom the
specimens were sent for determination, recognized among them
a mandible with teeth of Hippopotamus sp.

Schlosser described and figured a deformed, injured molar
tooth from Shansi, which he considered to belong to Hippopotamus.
Ringstrom^) redetermined this specimen, however, Sinotheriums^.

From what precedes it will have appeared conclusively, I believe,
that very small differences exist in general between the Siwalik
hippopotamus species mutually and between the latter and the Javan
fossil form H. sivajavanicus Dub. If ever it was necessary to determine
the degree of the individual variation of a recent species of the
genus, it was now. In doing so the writer met, however, with
rather great difficulties, as will be seen from what follows.

Besides the dwarf species Cboeropsis liheriensis (Morton) of
Liberia there have been recorded three recent hippopotamus species:

i. Hippopotamus amphibius Linnaeus, Syst. Nat., I, 1758, loth
ed., p. 74 (Nile).

Syn: Hippopotamus typus Duvernoy, ITnstitut, XIV, 1846,
p. 333 (Senegal and Abyssinia).

3.nbsp;Hippopotamus constrictus Miller, Smiths. Misc. Coll. LIV,
no. 7, 1910, p. I (Angola).

H. australis from the Cape is founded upon differences in
the bony cranium, but also upon a number of differences in the
skeleton; H. constrictus, however, is founded only upon differences

Hippopotamus appeared to be well represented in the Dutch
collections. The writer measured only skulls in which at least
the first true molar was in use. Nevertheless he found not less
than 23 skulls, viz., 2 in the Zoological Institute of the University
of Utrecht, 7 specimens in the museums of Natura Ards Magistra
in Amsterdam, and 14 specimens in the Nat. Mus. of Nat. Hist,
in Leyden. The two Utrecht specimens were labelled H. amphibius-,
the localities from where they are obtained were unknown. All
Amsterdam specimens were unlabelled. With the exception of one
unlabelled specimen, all Leyden specimens were labelled H. amphibius.
From table P it will be seen that of most of the latter the locality
is known. As localities are mentioned: Congo, Mozambique, Cape,
Nile, Zambezi and Zanzibar. With regard to -the last locality I
strongly doubt the correctness of the statement for it occurs as
highly improbable to me that in this relatively small island near
Africa\'s E. coast hippopotami should occur, these being surely
bound to large rivers. Furthermore it must be borne in mind that
in a museum, so old and venerable as the Leyden Nat. Mus. of Nat.
Hist., the data of the labels are perhaps not m all cases reliable.
For all these reasons I thought it advisable to start from the supposi-
tion that of all recent specimens measured the locahties were un-
known, so that I should be obliged to determine all. And this seemed
to be an easy task by the help of the data which the literature

Miller the latter specimens showed conclusively that individual
variation is not quot;unusually great ...quot; i) so that he was convinced
that the Zambezian and Angolan specimen represented distinct
species. As, furthermore, the skulls from British East Africa agreed
in all essential characters with the Zambezian skull, they were for
the present assumed to represent true Hippopotamus amphibius of
the upper Nile. The Angolan species was named H constrictus
because of the deep rostral constriction.

The type skull of H. constrictus is that of an immature male
(W not fully in place); no. 34787, U. S. Nat. Mus. Miller gave
the following diagnosis: quot;Skull in general like that oiHippopotamus
amphibius, but rostral constriction deeper, dorsal surface of cranium
more flattened, difference in level between anterior border of
maxillary and of premaxillary more pronounced, and mandibular
symphysis much shorter, its median length less than one-third
length of mandible; teeth smaller than in Hippopotamus amphibius,
the transverse diameter of m^.^) scarcely more than one-half that
of palate.quot; Miller mentioned more differences but apparently
did not esteem them important enough to add them to his diagnosis.
Furthermore, he gave the measurements of the Angolan specimen
of H constrictus and of the adult male Zambezian skull, already
mentioned, and calculated a number of relations between certain
measurements in order to show most of the differences mentioned
in numbers. His paper is illustrated by fine figures of the two
specimens of which he gave the measurements.

Without doubting the exactness of Miller\'s assertion concerning
the variation which he nodced in the eight skulls from British
East Africa, we may say that Miller\'s foundation of a new species
would have been far more convincing, if he had given the measure-
ments of all these skulls so that the reader might have formed a
notion of the range of variation. For it appears to me as not impro-
bable that Miller will have selected among the nine skulls of
H. amphibius which he had at his disposal, the one showing the
greatest differences from the Angolan specimens. Premising the
correctness of this supposition, Miller gave too favourable an
idea of the differences which in general exist between the Angolan
specimen and the skulls of H. amphibius. But apart from this. Miller

depth at level of anterior border of orbit contained about 4 times in
condylobasal length, instead of 31/2 dmes, as in H. amphibius-,
rostral constricdon very pronounced, its least breadth contained

Meant is M^, which appears from the text and also from pl. II where a palate view of

nearly 7 times instead of about 5 times in condylobasal length ...quot; i).
If one calculates, however, these relations it appears that for the
relation first mentioned the values are resp. 3,81 and 3,51 (instead
of 4 and 3 and for the second relation resp. 6,27 and 5,07 (instead
of 7 and 5). Furthermore, I cannot see how Miller came to put in
his diagnosis quot;... the transverse diameter of M^ scarcely more
than one-half that of palate ...quot; If one calculates namely the relation
between width of palate at anterior lobe of M^ and width of first
upper molar, one finds for the Angolan skull 1,65 and for the
Zambezian skull 1,67. Of course this difference is too small to be
of any signification.

As Miller compared the measurements of an immature specimen
(namely the Angolan skull in which M^ was not fully in place)
with those of a mature skull (the Zambezian one) I was justified in
using not only the measurements of the skulls in which M^ has been
in use, but also those with worn M^. In this way there remained 16
skulls. The measurements of these specimens, together with those
of Miller\'s specimen and those of 7 crania in which only the
first true molar has been in use, will be found in table P. Of each
skull the measurements mentioned by Miller have been taken.
I joined Miller also with regard to the circumscription of each
measurement. Only there where the possibility of a wrong inter-
pretation was not excluded, I added a short explanation between
quadrangular parenthesis. In order to be able to settle the question
whether H. constrictus was represented among the skulls above
mentioned, and if so, which skulls had to be assigned to that species,
we united in table Q the results of a number of calculations.

The relations 1:7, i : 3, 4 : 2, 13 : 12 and 15 : 12 are those
selected by Miller. I added:

I : II and i : 6 in order to express in numbers a difference which
exists between Miller\'s Angolan and Zambezian skull
and which Miller described as: quot;... occipital region and
dorsal surface of orbits much less elevated above level of
median portion of frontals than in H. ampJnhim ...quot;
9 : 10 in order to show the difference described by Miller as
quot;... the vertical diameter of orbit not conspicuously
greater than transverse diameter ...quot; (in Angolan specimen,
contrary to Zambezian skull where vertical diameter dis-
tinctly exceeds the horizontal diameter).

The measurements of these crania are given, as I expcct that they may be of use to

When making table Q we asked ourselves: Which specimens
show typical constrictus-characters and which typical quot;amphibiusquot;
features. We used therefore only those values, which lay outside the
interval restricted by the values of resp. Miller\'s Angolan and Zam-
bezmn skull. In spite of the fact that in doing so we accepted individual
variation in only one direction, we came to remarkable results

It will be seen that only specimen no. 13 (and perhaps no.quot; 7)
paramountly show constrictus characters, whereas the other ones
most y show typical constrictus as well as typical amphibius charac-
ters ). The occurrence of such a large number of intermediate
skulls, proves, however, that K constrictus Miller has no right to
specific distinction and that it is identical with H. amthibiul The
reader who has taken the trouble to study the tables, which demon-
strate the individual variation of kerabau, banting and Tavan
rhinoceros, will not be astonished by this result

Finally it may be remarked that quot;... the difference in level
between anterior border of maxillary and ofpremaxillary .. .quot;which
difference Miller mentioned in his diagnosis and which can
hardly be expressed m numbers, also appeared to be of no value
as a specific character.

We shall still have to occupy ourselves with the question whether
perhaps H austrahs of the Cape region is present among the material
measured by us. Therefore it will be necessary to consider the
original hterature, which deals with that form.

Desmoulins started from the opinion, that when the dentition of
two forms shows no structural differences quot;... la figure, la propor-
tion et la direction de leur usure peuvent servir à exprimer d\'une
mamere abreviative et générale l\'ensemble des différences de deux
ammaux comparés, et par conséquent, à déterminer leur espècequot; 2).

The difference in use of the canines of a specimen of the Cape
and of one of Senegal it was, which led him to make further investi-
gations. The result was, that he beHeved to see sufficient essential
differences (in cranium as well as in skeleton) to justify specific
distinction. Only the cranial differences may be of interest to us.

The Senegal skeleton was that of an adult individual, M3 being
worn, whereas the Cape skeleton belonged to an immature individual,
as it is stated that M3 was not yet in use. It will be permitted,
therefore, to make use of the same 16 recent crania in what follows.
Desmoulins mentions a.o. the following differences:

a- Worn surface of canines almost vertical in Senegal form,
about 25° more obhque in Cape form.

b. Point of C a quot;poucequot; i) below alveolar border of C in Senegal
form; in Cape form a quot;poucequot; above it.

Hind border of for. infraorbitalis above the middle of P^ in
Senegal form; Cape form above front border of P^

Duvernoy compared the skull of a skeleton obtained from
Abyssinia with those of two skeletons from Senegal and with that
of a skeleton from the Cape. It is highly probable that this material
contained the two skeletons, which Desmoulins had at his disposal.
Duvernoy concluded that the forms of Abyssinia and Senegal
were identical, but specifically distinct from the form of the Cape.
The one mentioned first was named K. tjpus, the second K australis.
The former ought, however, to be named of course H. amphibius.

d. The greater length of the teeth row (P^—AP at alveolus)
in Abyssinia form, compared with Cape form.

The smaller length of the cranium in proportion to its width
in K amphibius of Senegal and Abyssinia.

/• The circular or even ovalshaped form of the orbit with its
longer diameter vertical in the latter species instead of horizontal
as in H. australis.

With the differences above mentioned, we shall try to answer
the question formulated at the beginning.

We may commence by observing that the difference mentioned
under a. is of no value, as a great number of intermediate conditions

Desmoulins does not mention 1» and P^ but the jth and 4th molar. Apparently he
numbered the chcck teeth from back to front.

occur. Furthermore may be remarked that in composing table R,
we proceeded in the same way as in table Q.

Table R shows convincingly that also the remaining 5 differences
are of no value as species-characters. Only , one specimen (n°. 6)
has exclusively quot;australisquot; characters. But that this also is of no
signification may appear from comparison with table Q, which
reveals that spec. 5 also possesses typical constrictus- and amphibius
characters. Thus the impression is strongly made, that the skull
of H. australis in reality cannot be distinguished from that of
H. amphibius. With this remark I do not intend to say that H. australis
is not a distinct species. Therefore I should have been obhged to
implicate in my considerations also the differences of the skeleton,
which Desmoulins mentions. But this would have led me too far.

In order to avoid confusions, I will lay stress upon the fact
chat I do not imagine that I proved rigorously the entire accordance
of the crania of K. amphibius and H. australis. I only made it more
or less probable.

As may appear from what follows we took all this trouble in
order to get data, on the one side with regard to the relation between
length of the symphysis and width of lower jaw across canines,
on the other side in respect to the relation between length and
breadth of the occiput. For assurance of those, who esteem the
possibility not excluded of H. australis being represented among
the recent material measured, we may add that neither Desmoulins
nor Duvernoy did mention any difference, which concerned the
above relations.

In table S all the measurements are given, which Lydekker gave
of the mandible remains of H. sivalensis, H. iravaticus and H. palaeindi-
cus. Moreover to this table are added the corresponding measurements
of the mandible remains of our own collection. With the data of
this table, combined with the conclusions, which we can draw from
table Q, we shall try to determine whether or not the three mandibles
of our own collection belong to one species and if so, whether this
species is identical with one of the S. E. Asiatic fossil forms. Also
we shall have to ask the question whether these continental fossil
forms really belong to four different species.

Comparison of measurements and descriptions of the Javan
specimens show that there is not the slightest reason for accepting
the presence of more than one species.

As to the right of the continental forms to specific distinction
the following observation may be made. From the discussion of the
continental forms it will have appeared what large value was placed
on the relative length of the symphysis. In table Q we may notice
the degree of variation of the relation between the length of the

symphysis (m. 15) and the interval between outer surface of canines
(m. 14) in 18 skulls of (very probably) one recent species. If we had
had at our disposal more recent material for comparison, the minimum
and maximum values of relation 15 : 14 would undoubtedly differ
still more. It is, however not probable that the difference would
be very much greater, i)

Returning now to table S, we notice that on the basis of the
length of the symphysis both Narbada forms, H. namadicus and
H. palaeindicus, certainly cannot be distinguished 2).

I cannot see, therefore, why Lydekker stated that the mandi-
bular symphysis of H. namadicus is ... quot;slightly longer ...quot; than
that of H. palaeindicus. Nevertheless there is no doubt that both
forms are specifically distinct, as the incisives afford ample reasons
to separate the two.

The value found for the relation i : 2 for the type specimen of
K iravaticus (which is also the only specimen of which measurements
are known) shows that this species is certainly not identical with
one of the Narbada species. The specific distinction between U.
iravaticus and H. sivalensis appears less convincing from the values
of the relation i : 2. It does, however, appear from the comparison
of the absolute values in H. iravaticus, measurement i distinctly
exceeding m. 2, and the reverse being the case in K. sivalensis.
The decidedly smaller size of K iravaticus, together with its lateral
compressed outer incisor, give some more differences, which cer-
tainly justify specific distinction.

The values of the relation i : 2, together with the absolute
values of measurements i and 2, show that specific identity of
U. namadicus and H. sivalensis may be considered as improbable.

Thus we arrive at the conclusion that on the ground of the
material hitherto found, 4 different continental species of hippopo-
tamus have indeed to be distinguished.

And now the question: Is our own fossil form identical with
one of the continental species?

It will be clear — and moreover it will appear from table S —
that the Javan form is certainly not specifically identical, neither
with K palaeindicus nor with H. iravaticus.

1)nbsp;One must not forget, that these i8 specimens are certainly partly derived from very
separate localities. As a matter of fact we may expect, therefore, a larger range of
variation than should have been found if we had had at our disposal 18 skulls of the
same locality.

2)nbsp;N.B. Measurement 2 of table S is not identical with measurement 14 of table P.
In the first case I had to follow Lydekker, in the second case Miller.

Consideration of tlie values of the relation i : 2 and of the
absolute values of measurements i and 2 reveal, that the affinity
of the Javan form to the Narbada form H. namadicus is far more
close than to K. sivalensis. On the other hand, however, I immediately
admit that the values of m. 5—13 of the Javan form in general
fairly agree with the corresponding ones of the small variety of
K. sivalensis. With a view to the fact, however, that we do not
dispose of these values in H. namadicus, too much value cannot be
attached to this agreement.

From table S it will be seen that in all the specimens of H.namadicus
and H. sivalensis of which Lydekker gave the measurements the
diameter of I2 is smallest in comparison with those of and I3,
whereas in our own specimen II, the diameter of the alveolus
of I3 is smallest. However, this need not be an argument against
identity with one of the two forms mentioned. First it appeared
that the relation between the diameter of \\ and Ig in H. amphibius
may vary considerably secondly in fig. 4 of pl. LXII of the
F.A.S. a front aspect of the anterior portion of a lower jaw of
H. sivalensis is represented, showing I3 which is in cross-section
smaller than Ig and the latter slightly smaller than I3. Therefore
it exhibits a similar condition as our specimen 11. The above mention-
ed Siwalik specimen deserves also in another respect our full attention.
From fig. 2 of pl. VI, it will have been seen that the alveoli of the
incisors of specimen II are remarkably shallow. Remarkably because
of the fact that in all recent hippopotamus mandibles which the
writer saw and in which the incisors could be removed, the
alveoli appeared to extend almost throughout the whole length
of the symphysis. In the figure of the Siwalik specimen referred,
however, the alveolus of right Ig (which is the only incisor which
is wanting) appears to be, as in our specimen II, very shallow. As
to H. sivalensis, however, this appears to be not a constant character.
In fig. 4a of pl. LXII of the same work there is namely represented
the front aspect of a vertical section through the anterior portion
of lower jaw, and that section — which is very distinctly situated
behind the anterior margin of the mandible — shows distinctly
sections through the six incisors. As far as can be seen from those
figures in the F.A.S., which represent mandibles of H. namadicus.

For the sake of completeness may be mentioned here, that the writer did not see any
skull, in which the diameter of Ij was equal to or greater than that of Ii- At the same
time I may draw attention to two peculiarities which I noticed in studying the recent
crania. In the lower jaw of specimen 6 of table P only one incisor occurs in each half
of the jaw, while in the upper, quite normally, four incisors arc developed. In the
lower jaw of specimen j of\'table P there occur two alveoli between the outer incisors,
but considerably above the level of these teeth. The peculiarity is, that the diameter
of the right alveolus is 39 mm. and of the left one 11 mm. In this specimen the upper
jaw is also normal.

the alveoh of the incisors are much deeper than those of our own
specimen IL But for reasons which will appear from the above I dare
not attach any value to this difference. We arrive, therefore, at the
conclusion that the mandibles of our own collection in one important
detail, namely in the relation between measurements i and 2 corres-
pond with those of K namadicus, while on the other hand affinities
to H. sivalensis cannot be denied.

Before occupying ourselves with the question whether our
fossil form exhibits affinities to ff. sivajavanicus Dub., it will b^-.
desirable to direct our attention for a moment to the two remaining
specimens of our collection viz., the portion of the palate with
Ml—M^ of either side, and the posterior portion of the cranium.
Beforehand we may expect that these specimens will be of
little use in our effort to determine the species. First it will be
known that no sufficient distinction can be drawn between the
different species of hippopotamus from the structure of the
molars; secondly we saw that only of H. sivalensis the cranium is
known.

In the table below will be found (a) the measurements of
the molar teeth of the referred palate of our collection, (b)
those of the type form and of an aberrant form of H. sivalensis, (c)
those of Hippopotamus sp. described and figured by Stremme and
obtained from Java.

The molar teeth of our own specimen have been measured at
the base of the crown, and at the outer side of the teeth. Stremme
mentions that he took his measurements at the base of the crown.
Lydekker does not notice where he took the measurements of
H. sivalensis. Like the molars measured by Stremme, those of our
specimen are quadrangular at the base of the crown and seem to
differ in this respect from H. sivaknsis.

In the F.A.S. four hind aspects of hippopotamus crania are figured
namely three of H. sivalensis and one of quot;H. palaeindicusquot;. As already
noticed Lydekker pointed out that the cranium remains, which
Falconer assigned to the latter species, may partly belong to
H. namadicus. The values of the relation between height (from
lower border of for. magnum to occ. crest) and breadth in these
four specimens i) will be found below, together with the measure-
ments and the corresponding relation of our own specimen.

With a view to the fact that the variability which we noticed
in this respect in i8 skulls of K. amphibius (see table Q), we may
say that our own specimen shows more affinities to the skull of
H. ^\'\'namadicus-palaeindicus\'quot; than to that of ff. sivalensis.

Both upper molars as occiput should therefore point in the direc-
tion of a species, distinct from H. sivalensis.

At last the question: Can our specimen be assigned to H.
sivajavanicus Dub? Of the differences which Dubois mentioned the
following ones may interest us.

Concerning these points it shows affinities to H. sivalensis
from which species it is especially distinguished by:

a in having the alveoli of the lower incisors not situated in
a straight line, and furthermore by:

Concerning difference a it will be remembered that in this
respect our specimens are* in perfect accordance with JFf. namadicus.

As to the size of the incisors (difference b) a closer affinity to H.
sivalensis can hardly be detected. To difference a I do not attach
much value. Finally it may be remarked that Ig of specimen II of
our collection is indeed more strongly developed with regard to \\
than is often the case in H. sivalensis. (See table S). It will, however,
be remembered that an identical development occasionally appeared
to exist in the latter species.

We conclude that we are not justified in classing our fossil
form with H. sivajavanicus Dub.

If we summarize the results, we may say that the teeth of the
palate as well as the occiput — though of course totally insufficient
for specific determination — seem to point in the direction of a
form distinct from K sivalensis. As to the other remains, we noticed
the greatest affinities to K namadicus. In my opinion, however,
the specimens, which I had at my disposal, do not allow of answering
the question whether they belong to H. namadicus or not. Better
preserved Javan specimens are needed to show full affinities. I
therefore thought it better to leave our specimens provisionally
specifically undetermined, and to describe them under the heading
of Hippopotamus sp.

The horizontal ramus is broken off some 3 cm. in front of the
anterior extremity of the molar so that nothing of the mandibular
symphysis has been preserved. There is not the slightest indication
of the presence of lower incisors.

The upper portion of the coronoid process and the upper part
of the ascending ramus are lost. The lower border of the horizontal
ramus is slightly but clearly convex, and sweeps round by a bold
curve into the posterior border of the ascending ramus. Inner-
and outer surface of the horizontal ramus are slightly convex
antero-posteriorly, and converge frontwards. Below the anterior
extremity of the molar the greatest width of the horizontal ramus
is, however, still considerable. The outer side of the horizontal
ramus is strongly convex from above downwards. The inner side
is in the same direction but slightly convex in its upper portion;
towards the lower border the convexity increases. The outer side
presents a circular foramen mentale (diameter 13 mm.) situated
some 7 cm. below and about 2 cm. in advance of the anterior extre-
mity of the molar. The inner-upper side of the ascending ramus
shows the base of the orifice of the dental canal, still filled with
matrix.

The alveolar margin of the molar and the lower border of the
horizontal ramus converge but slightly posteriorly. Consequently
the difference in height between the horizontal ramus at the front
side of the molar and at the posterior extremity of the latter is
small.

Extraordinary fine specimen. Only the enamel of the outer cone
of the pentalophid is somewhat damaged.

The tooth carries five transverse ridge-crests, a fore- and aft
talon. The bases of proto-, meta-, and tritolophid are of approxi-
mate equal width; from the tritolophid, however, the transverse
diameter of the crown decreases gradually. The posterior extremity
of the tooth is rounded off. At the anterior extremity a large disc
of pressure occurs which can distinctly be seen in fig. 3 of pl. IX.
Posteriorly any such a feature is totally absent. The tooth under
consideration is certainly a third molar, as may be determined from
its size, form, number of ridge-crests and from the absence of any
indication of a pressure-effect behind.

The specimen is rather considerably abraded by wear; inner-
and outer cones of the protolophid, up to and including tetartolophid,
present dentine-islets. The dentine-islet of the outer cone of the

protolophid and of the metalophid is in direct connection with
that of the inner cone. The cones of pentalophid — though touched
by attrition — have their thick enamel not yet perforated, the hind
talon, which is much lower, is untouched.

Both grinding surface and base of the crown are distinctly con-
cave superiorly. The plane of wear, however, is not only concave
longitudinally, but also in the direction of the greatest diameter of
the transverse lophs. The grinding surface slopes obliquely from
the inside, which is higher, to the outside, which is lower, thereby
presenting the characteristic ^\'Mastodon positionquot; of the grinding
surface. The molar is slightly bent, as viewed in plan; the outer
side is concave, the inner side convex.

The ridge-crests are not perpendicular to the longitudinal axis
of the crown, the inner cones being in advance of the outer ones,
which causes a slightly alternative arrangement of inner- and outer
cones.

Protolophid and metalophid being too much ground down, it will
be necessary — in order to get an idea of the structure of the ridge-
crests — to begin with the consideration of a crest which is not so
much worn e.g. the tritolophid (see schematic drawing of this
crest in text fig. i8). The tritolophid consists normally of an outer-
and inner cone. The structure of the outer cone is easily understood,
especially when compared with the corresponding cone of the
tetartolophid. Both cones consist of the components a and b;
b is distinctly in advance of a. On the postero-internal angle of the

that the component e^, which is
closely attached to e, represents
another trefoil conelet. It will be
clear, that the outer cone of the
tritolophid, if more ground down,
would present a distinct trefoil-
shaped dentine-islet, similar to the
one shown already by the outer
cone of the metalophid.

The outer cone of the tetarto-
lophid resembles closely that of
the tritolophid, described above. The
only difference is that in the fourth crest the separation between
the components e and e^ is considerably less distinct.

The outer cone of the pentalophid — though injured — pre-
sents clearly the components a and b, while the component e
(? plus e^) occurs at the base of the postero-internal angle of

The structure of the inner cone of the tritolophid is not so clear.
Without doubt also the inner cone consists of two components
(c and d), comparable with the components a and b of the outer
cone (see the inner cone of the pentalophid). It seems, furthermore,
that a trefoil conule (f) occurs on its postero-internal angle. This
component f is certainly present in the inner cone of the metalophid,
while its presence is also indicated in the corresponding cone of
the tetartolophid. The inner cone of the pentalophid has lost com-
ponent f. On its antero-internal angle, however, a low, but distinct
trefoil conule occurs. The foregoing description will render it
perfectly clear to the reader that the inner cone possesses a tendency
to form an imperfectly trefoil-shaped dentine-islet, which, however,
will never be so regular as that of the outer cones (e.g. compare
inner- and outer dentine-islet of metalophid).

The cleft which separates inner- and outer cone of each crest
is not parallel to the longitudinal axis of the crown, but axis and
each cleft intersect anteriorly. This is caused by the fact that the
cleft runs from thé middle of component e of the preceding crest
to the middle of component d of the succeeding loph.

The greatest diameter of the inner cone is at right angles to
the longitudinal axis of the crown with the exception of the
tetartolophid, where component d of the inner cone is somewhat
in advance of component c^).

The transverse valleys are completely blocked, mainly by the
large component e. The bases of the — relatively wide — inner-
and outer extremities of the transverse valleys are each occupied
by an — on the whole — thick, blunt, accessory tubercle.

The accessory tubercles, occupying the outer mouths of the
valleys between proto-, meta-, and tritolophid are already abraded
or touched by attrition. Cement is present in very small quantities.
There are, however, some faint indications which seem to point
out that cement has been lost, so that we must take into consider-
ation the possibility that the real amount of cement has been
greater.

The enamel is thick and near the basis of the crown provided
with a wavy and finely grooved rugosity, more or less parallel to

The anterior talon is only slightly developed, due to the great
pressure, which the tooth under consideration exerted on the

1) This is perhaps an indication that component d is united with an internal trefoil conule
as would have occurred in the inner cone of the pentalophid after prolonged wear.

preceding r. Mg. The fore talon apparently consisted of two low
tubercles, the one situated in front of component a of the outer
cone, the other placed in front of component d of the inner cone
of the protolophid.

The hind talon consists of 3 conelets and is distinguished from
the pentalophid by its much lower elevation. Noteworthy, however,
is the arrangement of these three conelets. Two of them are arranged
in a transverse line parallel to the direction of the greatest diameter
of the ridge-crests; the third conule is situated behind the interval
of the two, first mentioned. The couple, which is arranged in a
transverse line, in this respect resembles closely a true crest, which
impression is strengthened by the fact that the inner conelet of that
couple presents a division into two. There is no doubt, that
Osborn would have spoken in the present case of a hexalophid plus
posterior talon (cingulum), which is of course merely a matter of
taste.

The molar under consideration being a third molar and carrying
five ridge-crests and a hind-talon, two conelets of which show the
apparent potency to form another crest, evidently belongs to a
tetralophodont mastodont. Its determination, however, will be
dealt with at the end of the descriptions of the mastodont remains.

Measured along median line of tooth.
In this case the total length of the two foremost conelets.
\') Lingual width of the total hind talon.

Fragment of the upper jaw, containing on either side a fragment-
ary tooth, which is ii^erred to be the third molar from its size,
form, and number of ridge-crests. The greater part of the 1. M^ is
lost; only fragments of the inner cone of proto- and metaloph are
preserved. The dentine-material of its fangs can easily be distinguish-
ed from that of the bone of the palate, enabling us to get an idea
of the position of this molar in the upper jaw and of its total length.
Left and right molar are approximated behind and divergent in
front, (width of the palate anteriorly 84 mm., posteriorly 73 mm.).
The surface of the palate is damaged.

The r. M^, though not inconsiderably injured, is much better
preserved. Therefore, it will be sufficient only to give of this spe-
cimen a detailed description.

This molar presents a regular form, as viewed in plan. Inner-
and outer side are approximately straight and converge posteriorly;
the tooth therefore becomes narrower towards its posterior extre-
mity. The hind talon and the outer cone of the protoloph are totally
wanting; the last crest and the fore talon are for the greater part
broken off. The inner cone of proto- and metaloph is injured.

1)nbsp;According to the first of the original lists of specificadon (dated Bandung, 27 December
1927, n°. 11567) case n°. i of the first consignment contained (literal translation):

I incisive tusk (length 83 cm.), i connection (bridge) from upper jaw to occiput.

Originally the writer believed that the meaning of the foregoing statement was
that all these remains belonged to the same skull. He had, however, to alter his
opinion after case n°. i had been unpacked. On that occasion it was revealed that
the quot;portion of the lower jawquot; in reality represented a portion of the upper jaw,
containing fragments of 1. and r. M®. As the quot;upper jawquot; (represented in fig. i of
pl. VllI and fig. 2 of pl. X) also contains r. and 1. M\', it was — for this reason alone
already — impossible that the two specimens belonged to the same individual. This
being so, there was in the writer\'s opinion no reason to believe that the remaining
two specimens belonged to one another and to one of the two specimens mentioned.

After the investigation of the two latter specimens was ended, they were immediately
sent back to Bandung. The other two (tusk and zygomatic arch), however, remained
provisionally in Utrecht. In a letter (dated Bandung, June 4th 1929, n°. H 498) was asked
for these specimens, because they were needed quot;for completion and mounting of
the skullquot;. From this appears that in Bandung one is convinced of the belonging of
tusk and zygomatic arch to the specimen of the upper jaw. That is the reason why
we ventured to describe the tusk in the following pages under the same name as the
other specimens mentioned.

If we are rightly informed, a reconstruction of the skull has been made in Bandung.
For that purpose undoubtedly the specimens will have been needed, which have
subsequently been found. (Sec p. 128.)

The degree of wear is moderate. The remnant of the inner
cone of the protoloph and the outer and inner cones of the metaloph
present dentme-islets. The outer cone of the tritoloph is flattened
and not yet perforated; the inner cone is just perforated by attrition
Inner- and outer cones of the tetartoloph are only slightly touched
by wear The grinding surface slopes obliquely from the inside
which is lower, to the outside, which is higher (the tooth is supposed
to be m the natural position).

A convexity of the plane of wear could not be ascertained In
my opmion the tooth carried five ridge-crests, a fore- and aft taion
As to the presence of five crests plus an anterior talon the reader
will immediately be convinced after examination of the fig. i of
pl. IX. But also the former presence of a posterior talon is abso-
lutely certain. The disc of dentine-material namely, which occurs
behind the fragments of enamel belonging to the front side of the
pentaloph, is undoubtedly too extensive for one crest.

The crests are rather tall and comparatively narrow antero-
posteriorly. They are cleariy inclined forward. The relatively wide
transverse valleys are in general obstructed at their extremities by
low, blunt, accessory tubercles. In the middle they are totally blocked
by trefod conules, as will be seen in what follows.

It will be desirable again to pay our attention in the first place
to the structure of one crest, e.g. the tritoloph. (Text fig. 19 presents
a schematic drawing of this crest, only the degree of wear is suppos-
ed to be somewhat greater).

It will be seen that this crest — like the other crests — is nor-
mally composed of an inner and an outer cone, each subdivided
into two components (resp. c and d; a and b). The components a
and b of the outer cone are arranged in a line which is nearly per-
pendicular to the longitudinal axis of the crown. The greatest

diameter of the inner cone is,
however, distinctiy oblique
to the same axis, the com-
ponent d being in advance
of component c. Both on
the postero-internal and the
antero-internal angle of the
inner cone a well-defined tre-
foil conule occurs (f and h);
f is slightly more developed
than h; f is attached to the
component c, while h is
fastened to the component d. The postero-internal angle of the
outer cone also presents a trefoil conule (e), while a very ill-defined

trefoil conelet (g) occurs on its antero-internal angle. But these
two trefoil conules are attached to the innermost component (b)
of the cone.

Having fixed the detailed structure of the tritoloph we shall
continue with the consideration of the resemblances and differences
between the tritoloph and the other crests.

As has been already mentioned, the pentaloph is present in the
form of fragments of the enamel of its front side. But even these
poor remnants show clearly that the inner cone was also composed
of the components c and d, the latter being distinctly in advance of
the former. The presence of a component h is not indicated.

The structure of the tetartoloph corresponds in general with
that of the tritoloph. The components e and h, however, do not
occur. Furthermore, it will be seen that the subdivision of inner-
and outer cone is less conspicious, when the cones are unworn or
. but little worn.

The difference between the second and third crest is small. In
the inner cone of the metaloph the component h is probably wanting,
while component f is strongly developed. The whole presents a
rather regular trefoil-shaped dentine-islet, one lobus being formed
by component d, the third by the trefoil conule f. In the outer cone
of the metaloph the component g is present in the form of a faint
rudiment of a trefoil conule; the component e also occurs, but its
development is still smaller (compared with that of component e
in the tritoloph).

So much being lost of protoloph and fore talon, their detailed
structure cannot be ascertained.

It will be clear that the inner cone of trito- and tetartoloph
should present, if ground down enough, a rather regular trefoil-
shaped dentine-islet, like that of the inner column of ridge 2.
The outer cones also show a tendency to form trefoil-shaped dentine-
islets, due to the presence of the components e and g. But as these
components are on the whole but very slightly developed a well-
defined trefoil-shaped dentine-islet will never occur in the outer
half of the tooth.

The clefts separating inner- and outer cones of trito- and tetar-
toloph run from the middle of the front side of component f in the
direction of the middle of component e of the foregoing crest;
consequently longitudinal axis of the crown and cleft are not parallel,
but converge posteriorly. The cleft, which separates inner- and
outer cones of the metaloph, is almost parallel to the longitudinal
axis of the crown.

The inner- and outer extremities of the transverse valleys are
occupied by low, thick, and blunt accessory tubercles. Those occupy-

Text figs. 20—21.
Locality: Bumiaju, Excavation 1-4 at the K. Glagah.
Description ofthe fragment ofthe cranium. (Pl. VIII, fi, , p, x, , ,,,, ^^^

fragment of the cranium containing the M^ of either
sents rudely theZ JZT I ^.quot;^^^\'^^^tion each sheath pre-

the bow of the circle. The angle which is enclosed between the
two radii is approximately 90°. The inner sides converge superiorly.

the edge formed by the sudden
transition of the lower side of the
incisive sheath to its inner side.
In the lateral aspect of the cranium
it will be seen that the lowernbsp;Text fig. 20.

slightly curved, while the above mentioned crista is clearly concave.
The acute angle enclosed between the lower border of the incisive
sheath and the lengthening of the alveolar margin of the molar
is about 22°.

Noteworthy, furthermore, is that the posterior extremity of the
incisive sheath is situated only some 4 cm. above the alveolar margin.

In upper view of the cranium the interalveolar fossa presents
approximately the form of an ovalshaped disk, which is considerably
concave in the transverse direction of the cranium, and less concave
in the longitudinal direction.

Interval between anterior margin of posterior nares and
middle of line connecting alveolar margins of both M^

Both teeth are very much ground down. The grinding surface
of both slopes obliquely from the inside, which is lower, to the

(andfoietaIon),and a fragment of the enamel of the inner siLrfnrot^
and trtoloph are lost. The left molar, being best preserved k wrn l;
desirable to begin with the detailed dese^ription^ rf hi

It carries five ndge-crests and a hind-talon, while a na«ow
edge of enamel, closely attached to the front side rf the n^ i T
mdicates, the presence of an anterior talon

The ridge-formuIa may therefore be reoresentprl « f„ir
X5 X. As will be seen from the figure theT^nnot h i f\'
that the tooth represents the thirl o^£Tru:roLr ^els Th^
degree of wear is so considerable, that not only the define Set Jof

but also the dentme masses of these crests mutually. The direcS
of the greatest diameter of each crest is almost perpendicular to
th longituAnal axis of the crown. It will be clear AafnoTe of ti e
firs three ridge-crests is suitable to allow of studying the struetu e

Outer and inner cone of the last mentioned crest are each com
posed of t^vo components (resp. a and b; c and d). It w[ll Te noted

that neither the direction of the
greatest diameter of the inner
p^tie, nor that of the outer cone
IS at right angles to the longitu-
dinal axis of the crown. In both
the innermost cone is in advance
pf the outermost one; in the
inner cone, however, in a far
stronger degree than in the outer
^one. The same can be seen in
he tetartoloph. On the postero-
internal angle of the outer cone

the inner cone of the\' pentaloph is somewhat irregular. But after
comparison with the corresponding cone of the M^ of the right
side 1), the structure of which is more simple, it is not difficult to
interpret its pecuharities. As has been mentioned above, component
d is very much in advance of component c. On its postero-internal
angle, and attached to component c, a thick trefoil conule (f) occurs.
It is not certain, whether or not a small trefoil conule is present
on its antero-internal angle. The crest is already too much ground
down to give clear indications on this point. The resemblance in
structure of the inner cone of the tetartoloph with that of the
pentaloph is close. The equivalent of the component f seems to
be less developed in the tetartoloph.

The hind talon consists of two thick and blunt conelets; the
outer one is situated immediately behind the component e of the
fifth ridge-crest, the inner one is closely attached to the hind side
of the component f. The outer tubercle is slightly stronger developed.
The posterior extremity of the narrow cleft, which separates both
tubercles of the hind talon, is occupied by a very faint rudiment of
another tubercle.

Cement is present in very small quantities. It must be borne
in mind, however, that the molar is very considerably ground
down so that the possibility of finding cement has become very
small, even if the amount of cement was originally greater.

Most crests being ground down below the level of the base of
the valleys it is rather difficult to form an idea of the form and
shape of the crests in lateral aspect of the molar. They seem to have
been rather tall. The cleft which separates outer- and inner cone
of the pentaloph forms a small angle with the longitudinal axis
of the crown. Cleft and axis converge posteriorly. The correspond-
ing cleft of the tetartoloph and the axis of the crown converge also
posteriorly, but the angle between both directions is considerably
greater. The extremities of the transverse valleys are in general
occupied by accessory tubercles, of which those of the outer extre-
mities are most strongly developed. The base of the sides of the crown
is provided with a wavy, finely grooved rugosity.

The differences between left and right M^ are small. The only
difference worth mentioning concerns the hind talon. The structure
of the posterior talon of the 1. M^ has already been dealt with.
That of the r. differs in having in the first place a sUghtly better
developed tubercle between its main components, and in the
second place by the presence of still another tubercle externally
of the left principal conelet (apply magnifying lens to fig. i of pl. VIII).

The reader will be able to make this comparison by applying the magnifying lens
to fig. I of pl. VIII.

as far as 14 cm. from the to^ quot; disTan« thquot;e H?
so that the tooth tapers very stadllfTh,
is provided with very shallow tnlTl- ,
width, separated bv ed- -nbsp;grooves of about i cm.

molar, both obtained Cleat lon iTaT^r,\' very damaged
undoubtedly belones tonbsp;\' ^\'\'Sah. Tiie molar

theless I thought it better to describe it under the heading quot;Mast-
don sp.quot;.

As to the fragment of the cranium, I may mention that, as far
as my knowledge goes, excavations i—4 have not yielded remains
of Stegodon and E/epbas. There exists, therefore, a great possi-
bility that the fragment under consideration belongs to Mastodon,
and in that case probably to the same species as the other remains.
I thought it advisable, however, to describe the fragment in question
under the heading quot;?Mastodon sp.quot;.

Very much injured and deformed molar, imbedded in a mass of
andesitic-tuf With some difficulty 5 ridge-crests may be recognized
so that it is probable that the specimen in question respresents a
third molar. Only two cones are well preserved. These are unworn
and show each an indistinct division into two at the top. The
remaining cones are present in the form of remnants of enamel
and dentine. In upper view the molar is considerably bent, but
this is certainly due to the same crushing force, which caused the
destruction of the various cones. Whether the specimen respresents
an upper or lower molar could not be determined. It may be ob-
served that the specimen under consideration does not present
a single character by which it may be distinguished from the
molars described in the foregoing paragraph.

Fragment of the cranium, comprising the surroundings of the
occipital foramen and of the posterior nares, both cavitates glenoi-
dales, left- and right proc. zygomaticus of the squamosal.

The occipital foramen is subcircular. Both occipital condyles
are perfectly preserved, they are egg-shaped, slightly accuminated
below. Some 5 cm. above the foramen magnum the lower part of
a median depression of the occiput occurs, still filled with matrix.

Left- and right cavitas glenoidalis are separated from the occi-
pital surface by a distinct furrow. The latter is accentuated by the
fact that the base of the occipital surface is situated below the level
of the cavitas glenoidalis.

sented in %. i of pj VIII and fil ^\'^ë^ntary cranium, repre-
individual.nbsp;^nbsp;^^^ \' pl- belong to the sLe

If the descriptions and figures of A ar^ R ko
e ev^ent that A and B h Je .hefai^l® ^^rcl^^^

mat ot a is in all probabihty also X5 X.
f. The extremities of. the vallevs nf a j rgt;
by accessory tubercles.nbsp;®

d.nbsp;In A and B the valleys are in the middle blocked by trefoil conules,
as well as by the distinctly oblique position of the inner cone in
respect to the direction of the transverse crests.

e.nbsp;The base of the crown of either is provided with a wavy and
finely grooved rugosity.

g.nbsp;In either trefoil conules may occur as a rule on front and back
side of inner and outer cones. In general the trefoil conelets
of the inner cone are more strongly developed than those
of the outer one.

h.nbsp;The clefts separating outer- and inner cone of each crest are
in both molars as a rule not parallel to the longitudinal axis
of the crown.

/. Though the amount of cement in A and B is at present small, we
have seen above, that the amount in B was certainly greater, while
we must take into account the possibility, that the quantity
of cement in A was also greater.

I have no doubt that the reader — after perusual of the foregoing
enumeration — will be convinced of the extremely close resemblance
between A and B. There will be therefore no hesitation in referring
the two specimens to the same species, so that we may continue
answering the second question.

It will perhaps be desirable first to lay stress upon the
fact, that in the dentition of the Bunomastodontidae constant differences
occur between the corresponding molars of the upper and lower jaw.
These differences are:

1.nbsp;Lower molars are always narrower and longer than the corres-
ponding upper ones. The Mg is often laterally bent (outer side
concave, inner side convex).

2.nbsp;In upper molars the transverse crests are as a rule at right angles
to the longitudinal axis of the crown, while in lower molars
the longest diameter of the crests is oblique to the axis, the
inner cone being in advance of the outer one.

3.nbsp;The valleys of the lower molars are relatively wider than those
of the upper ones.

Furthermore it will be known, that inner (resp. outer) cone
of an upper molar is comparable with the outer (resp. inner) conè
of the corresponding lower molar. .

These constant differences must be taken into consideration,
comparing the lower third molar (C) with A and B.

If the descriptions and figures of these molars be compared,
it will be observed, that the resemblance of C with A and B is also
very close. The resemblances are:

b.nbsp;The innermost component of the outer cone of C is distinctly
in advance of its outermost component. The same can be seen
in the inner column of A and B.

c.nbsp;The diameter of the inner cone of the first three crests of C
is perpendicular to the longitudinal axis of the crown. The
greatest diameter of the inner cone of the succeeding two crests
is slightly oblique to this direction. Precisely the same can be
seen in the outer cones of A and B.

d.nbsp;In all the three trefoil conules may occur on front and aft
side of both outer- and inner cones. The trefoil conules of the
outer cones of C and those of the inner cones of B are most
strongly developed.

The transverse valleys of A, B, and C are in the middle blocked
by the trefoil conules.
ƒ. The extremities of the valleys of A, B, and C are occupied by
accessory tubercles.

h. In all the three the cleft, separating outer- and inner cones
of each crest is not parallel to the longitudmal axis of the
crown.

The base of the crown of A, B, and C is provided with a wavy
finely grooved rugosity.
j. The enamel is in all the three of approximately equal thickness.

We see therefore — making allowance that C is a lower M3,
A and B upper M^ — that the accordance of the characters of C
with those of A and B is so extremely close that C certainly may
be reckoned to the same species.

If we summarize the essential structural characters of the third
molar of our species, we come to the following list:
i. Ridge-crest formula X5 X. There exists the apparent potency
of forming a sixth crest in the lower M3.

3. The clefts separating outer and inner cone of each ridge-crest
are on the whole not parallel to the longitudinal axis of the
crown.

4.nbsp;The innermost component of the outer cones of the Mg and of
the inner cones of the M^ is much in advance of the outermost
component.

5.nbsp;The greatest diameter of the inner cones of the Mg and of the
outer cones of the M^ is perpendicular (sometimes nearly
perpendicular) to the longitudinal axis of the crown.

6.nbsp;Trefoil conules may occur on the postero-internal angle of
both inner- and outer cones.

7.nbsp;The trefoil conules of the outer cones of the Mg and of the
inner cones of the M® are generally more strongly developed
than those of the cones of the other side.

8.nbsp;The transverse valleys are blocked in the middle. In the Mg
mainly by the large trefoil conules, which occurs in the postero-
internal angle of the outer cones; in the M^both by the trefoil
conule, which occurs on the postero-internal angle of the inner
cone, and by the oblique position of the inner cones in respect
to the direcdon of the transverse crests.

id. Thick enamel, near the base of the crown provided with a
wavy, finely grooved rugosity.

Below a list will be found containing the names of the twenty
seven species of \'\'Mastodonquot; of East and South Asia which arc
known to the writer. As far as possible we have tried to give
each species the generic name under which it has been mentioned
by OsBORN. For that purpose we had to peruse a rather great
number of Osborn\'s publications on the present subject, as at the
time this was written (March 1932) Osborn\'s announced memoir
on quot;The Evolution of the Proboscideaquot; had not yet been pubhshed.
Of only two forms viz., \'\'Mastodonquot; Ijdekkeri and quot;Mastodonquot;
pandionis the systematic position remains uncertain. That is the
reason why these forms have not been placed under the heading
of one of the subfamilies.

The list below-mentioned does not pretend to completeness as
far as the quoted hterature is concerned. As a rule only those papers
are mentioned, which either contain the descripdons of the type
specimens, or which supply the reader with sufficient data. The
mention of two papers behind the name of a species means that in
the first paper the species has only been curtly mentioned, but
that in the second it has been fully described and figured.

Having adopted Osborn\'s classification i) all the forms of Stegolo-
phodon ScHLESiNGER (syn. Vrostegodon Matsumoto) were reckoned
to the Stegodontinae (subfam. of the Elephantidae). At present in the
genus Stegolophodon the following forms are united: Stegolopbodon
stegodonfoides (Pilgr.), cautleji (Lyd.), cautleji progressus Osb.,
nathotensis Osb., and latidens (Clift). It will be unnecessary, I
believe, to prove that our species certainly does not belong to
Stegolopbodon.

Hemimastodon crepusculi Pilgrim 1908, Rec. Geol. Surv. Ind., XXXVII, p. 157. 1912,

Mem. Geol. Surv. Ind., New Ser., IV, 2, p. 17.
Hemimastodon annectens Matsumoto 1924, Journ. Geol. Soc. Tokyo, XXXI, p. 405.

1926, Sci. Rep. Tôhoku Imp. Univ., Sendai, Japan. Sec. Ser. X, i, p. 4.
Trilophodon macrognathus (Pilgrim) 1915, Rec. Geol. Surv. Ind., XLIII, p. 309. (Very
short description).

Trilophodon angustidens (Cuv.) var. palaeindicus (Lydekker) 1884, Mem. Geol. burv.

Trilophodon angustidens (Cuv.) var. chinjiensis (Pilgrim) 1913, Rec. Geol. burv. Ind.,

Trilophodon sendaicus Matsumoto 1924, Journ. Geol. Soc. Tokyo, XXXI, p. 408.
1926, Sci. Rep. Tôhoku Imp. Univ., Sendai, Japan. Sec. Ser., X, i, p. 4-

p. 83. 1880, Mem. Geol. Surv. Ind., Ser. X, I, 5, p. 202.
Tetralopbodon (Lydekkeria) sinensis (Koken) 1885, Palaeont. Abh. Ill, 2, p. 6. (See

Tetralopbodon punjabiensis (Lydekker) 1886, Cat. Foss. Mamm. Br. Mus., IV, p. 60.

Serridentinus mongoliensis Osborn 1924, Am. Mus. Nov., n°. 148. P- i-
Serridentinus browni Osborn 1926, Am. Mus. Nov. n . 23^, p. 4.
Serridentinus hasnotensls Osborn 1929, Am. Mus. Nov. n . 393, p. 2.
Serridentinus metachinjiensis Osborn 1929, Am. Mus. Nov., n . 393, p. 4.
Serridentinus chinjiensis Osborn 1929, Am. Mus Nov. n . 393, P- 5-
Serridentinus prochinjiensis Osborn 1929, Am. Mus. Nov., n . 393. p. 6.
Serridentinus florescens Osborn 1929, Am. Mus. Nov., n . 393, p. 6.

Rhynchotherium chinjiensis Osborn 1929. Am. Mus. Nov., n°. 393, p. 8.
Subfam. Brevirostrinae.

Synconolophus dhokpathanensis Osborn 1929, Am. Mus. Nov. 393, p. 10.
Synconolophus ptychodus Osborn 1929. Am. Mus Nov n°. 393, p. 12.
Synconolophus propathanensis Osborn 1929, An^ Mus. Nov., n°. 393, p. 12.
Synconolophus corrugatus (Pilgrim) 1913. Rec. Geol. Surv. Ind., XLIII, p. 293.
Synconolophus hasnoti (Pilgrim) 1913. R^c. Geol. Surv. Ind., XLllI, p. 293.
Anancus perimensis (Falconer et Cautley) 1845, Fauna Antiqua Sivalensis.
Pentalophodon sivalensis (Cautley) 1836, Journ. As. Soc. Bengal., V, p. 294.

Platybelodon grangeri Osborn 1929, Nat. Hist. XXIX, n°. i, p. 13. 1931, Am. Mus.
Nov. n°. 470, p. 6.

Mastodon pandionis Falconer 1868, Pal. Mem. I, p. 124. (Sec furthermore Lydekker,

Cat. Foss. Mamm. Br. Mus., IV, p. 37.
Mastodon lydekkeri Schlosscr 1903, Abh. math, physik. Kl. K. Bayr. Ak. Wis., XXII,
Abt. i, 1903, p. 46.

gt;) See the excellent summary given by Dietrich in N. Jhrb. f. Min. ctc. I, 1927, B, p. 311.

We shall begin with excluding those species from which our
Javan form can be easily distinguished.

As has been already mentioned, the Mg of the Javan form
shows a distinct tendency to hexalophodontism. Consequently
all true forms of Trilophodon may be excluded from comparison.

crests ascending on the outer cones of the lower molars and on the
inner cones of the upper molars; true trefoils 2), i.e. intermediate
conules in the center of the valleys, observed in all the species of
Trilophodon, are wanting.quot; (It will be noticed that the second
part of the quotation yields another character by which the Javan
form is distinguished from Trilophodon). In 1925 Osborn

There cannot be any doubt that the Javan form does not belong
to the genus Serridentinus. In the first place the grinding teeth of
our species show no quot;serrate crestsquot;, and secondly all the species of
the genus in question are apparently trilophodont. In none of the
forms, of which the third molar is known, these carry more than
four crests while in the remaining species the known Mg bears
three crests.

May also immediately be excluded. The molars of this genus
present a peculiar structure, which is easy to recognize, and which

distinguished by the multiplication of the cones and conules which
finally render the molar a bewildering complex, of a warped, choe-
rodont, ptychoid pattern.quot;

=) This in contradiction to crcsts with double trefoils. In that case the trefoils arc not
exactly situated in the centre of the valleys.
Proc. Am. Philos. Soc. Philad., LXIV, 1925, p. 23.
Only M\' of Serricientimu browni with rudiment of pentaloph.
Am. Mus. Nov. n°. 393, 1929, p. 10.

The M3 of the only S. E. Asiatic species of that genus carries
four crests, which excludes this form.

This remarkable genus with its short and broad shovel-tusks (I^
is represented by one S. E. Asiatic pliocene species from Central
Asia, namely, Platybelodon grangeri. Though several grindmg teeth
of that species have been found, no mention has been made of their
structure. Nevertheless, we may be sure that our form is not i^n-
tical with that species which may appear from what follows. Uur
own specimens described in the foregoing pages, were sent back
to Bandung some years ago. It appears that since that time ii^ew
specimens have been found, von Koenigswald i) mentiomng that,
besides bones and teeth, fragments of two crania and of three lower
laws have been obtained. Of course, von Koenigswald, palaeon-
Xgist of the Geological Survey of the Dutch East Indies will
not have failed to compare our specimens thoroughly with the
remains subsequently found. From his paper (1. c.) it is not difficult
to infer that in his opinion remains of only one species have been
procured from Bumiaju. The new material enabled him to sta^te
that the lower jaw of the species is devoid of lower incisors This
fact proves the distinctness from Vlatyhelodon grangeri (which was
a priori to be expected).

Of this species but very little is known. Schlosser described
and figured under the present name a fragment of a M-^ consisting
of last crest plus posterior talon. It suffices, however, to point out tha
Mastodon lydekkeri exceeds our species considerably in size aiul that
its molars apparently do not possess trefoil conules, thus showing
another, very essential difference.

The distinctions of the subgenus in question from the true
Tetralophodon stage are according to Osborn 2) as follows: (i)
second upper and lower molars subtetralophodont to tetral^opho-
donf (2) molars with single central trefoils instead of double
trefoils- (3) third upper and lower molars apparently with four and
a half to five crests instead of five and a half to six crests; (4) molars
extremely short-crowned or brachyodont....quot;. From this quo-

tation follows that our form certainly does not belong to the sub-
genus Lydekberia, so that two more species are excluded.

Of the twenty seven species there still remain four forms with
which we shall have to compare the Javan species.

Before the specimens of our collecdon were sent back to Bandung
most of them were provisionally named by the writer. So the
\'\'Mastodonquot; of our collection received the provisional name of
Mastodon perimensis. In the writer\'s opinion this identification
however, cannot hold good. There are certainly resemblances
between the molars of the two forms, but on the other hand there
are enough differences to justify distinction. Lydekker i) reckons
but a few specimens to the species in question; only four grinding
teeth have been figured viz., one 1. M^, two r. AP and one r. M .
If one compares the figures of these teeth one is inclined to beheve
that two different species have been united under the name of
M. perimensis. Two specimens of M^ appear to show a more com-
plex pattern of the crown than the remaining teeth. The former
are in this respect also distinguished from our species. As to the
specimens with a more simple structure, these differ in having more
distinctly alternate inner and outer cones, and only trefoil conules
on one cone of each crest. The figured M3 shows a considerably
greater amount of cement. To this difference, however, I do not
attach much value, first because the amount of cement may have
been greater in our own specimen, and secondly because the quan-
tity of cement may be very variable in the grinding teeth of one
and the same species.

Lydekker descnbed under the present name both specimens
from Sind and Dhok Pathan. Pilgrim 2) showed that the Dhok
Pathan specimens belonged to a disdnct species for which he pro-
posed the name Tetrahelodon corrugatus. Osborn 3) changed the
generic name in Synconolophus. The type of Mastodon pandionis is
a first upper molar, the locality of which is not precisely known,

Rcc. GcoI. Surv. Ind., XLIII, 1913, p. 253,
Am. Mus. Nov., n°. 393, 1929, p. 10.

but which Lydekker considers — according to Pilgrim quot;with
a fair show of probabilityquot; — to have come from Sind. This speci-
men will be found represented in Pal. Mem. I, pl. XXXIV, figs.
6—7; also (and far more distinct) in Mem. Geol. Surv. India, Ser.
X, III, i, 1884, pl. V, figs. i—ia. Comparison of this figure with
those of our specimens will reveal that the latter are distinguished
by a series of differences, of which I may bring forward the absence
of a strong outer cingulum in the lower grinding teeth.

Pilgrim (1. c. p. 293) separated under the name of Mastodon
hasnoti the Middle Siwahks specimens, which Lydekker had referred
to M. sivalensis, from the typical M. sivalensis of the Upper Siwaliks.^
Osborn (1. c. p. 10) changed the generic name of \'\'Mastodonquot; hasnoti
in Sjnconolophus. The intermediate molars of the typical M. sivalensis
have five crests, the third molar is in general hexalophodont. Lydek-
ker namely called attention to a M^ (represented in fig. i of pl.
XXXIV of the F. A. S.) which is pentalophodont. All known AI3,
however, carry six crests. As to the number of crests there is,
therefore a distinct affinity of our form to Pentalophodon sivalensis
of the upper Siwaliks. Nevertheless I do not believe that the two fornis
are identical. One of the distinguishing features of P. sivalensis is
the alternate arrangement of outer and inner cones, both in upper
and lower grinding teeth. And it will be immediately admitted that
alternate arrangement of the cones — though shghtly present in
our M3 as a result of the oblique position of the crests — is totally
absent in the M^.

It lasted until 1886 before Lydekker separated this form from
M. perimensis (and M. cautleji). Lydekker i) mentions four grinding
teeth, one of which is a 1. M^ and represented by the same author
in Mem. Geol. Surv. Ind., Ser. X, I, 5, 1880, pl. XLII. This specimen
is readily distinguished from our upper third molar by a number
of differences of which I mention (i) the considerably greater size (2)
the presence of a crenulated inner cingulum, and (3) a totally dif-
ferent posterior talon.

Concluding that the Javan form is not identical with one of
the twenty seven species of S. and E. Asia, we are wholly entitled
to consider the Javan form as a distinct species for which we pro-
pose the name of Tetralophodon bumiajuensis.

We may add that the structure of the molars of T. humiajuensis
resembles very closely that of the teeth oiTetralophodon longirostris\'^)\\
by the absence of lower incisors our form is, however, readily
distinguished from the latter species.

teeth of our specimens) were sent, determ ned the spec LTsTS.kL
Mastodon longirostris. (See: Dc Mijningenieur 1926. p Tjo) ^ ^

An imperfect cheek tooth of which the base of the crown is
convex antero-posteriorly, thereby showing that it belongs to the
upper jaw. The tooth is implanted in a fragment of the maxillary,
which enables us to state with absolute certainty that the tooth is
from the left side of the upper jaw. It is namely at the right side
of the tooth that portions of the palate have been preserved. The
tooth becomes gradually narrower from front to back. It is this
characteristic shape of the crown, together with the absence of any
disc of pressure on the posterior aspect, which show conclusively
that the tooth represents the last of the true molar series. It is,
therefore, a 1. M^.

The anterior part of the crown and of the fangs being broken
off, it is impossible to estimate the exact number of ridge-crests
which are lost. As will be seen below, I have no doubt in referring
this molar to Stegodon airawana Mart. The ridge-crest formula of
the M^ of this species is in most cases X12X. The tooth under
consideration carries a posterior talon plus 9 ridge-crests, so that
probably 3 ridge-crests plus an anterior talon have been lost. To
avoid confusions, I call the first ridge-crest of the molar in question
ridge-crest 3 i; the second ridge-crest 3 2, et. seq.

The greater part of ridge-crest 3 -f i is broken off; the inner
half of ridge-crest 3 2 and the outer cusp of ridge-crests 3 4,
3 5) 3 6, are somewhat damaged.

Ridge-crest 3 2 up to and including ridge-crest 3 -f 5 present
dentine-islets; the greater part of the cusps of ridge-crest 3 6

is perforated; the mammillae of ridge-crest 3 7 and the big
median cusp of ridge-crest 3 -f 8 are flattened. Ridge-crest 3 -f 9
and the very low hind talon are untouched.

Besides small, regular unduladons, the enamel of each ridge-
crest shows a certain number of constrictions (6—8) which
disappear near the base. The ridge-crests are somewhat bent, but
as the figure gives a far better idea of the form of the dentine-
islets than any descripdon may do, I think it unnecessary to enter
into details.

In order to show the proper form of the tall and rather slender
ridge-crests in side aspect, the cement occurring in the palatinal
extremides of the valleys between ridge-crests 3 4, 3 5j 3 6,
3 7 and 3-1-8 has been removed. A considerable amount of
cement is present. The last three valleys are completely filled; the
last two ridge-crests and the hind talon are for the greater part
covered by cement. The number of cusps, borne by the unworn
ridge-crests cannot be clearly ascertained. The penuldmate ridge-
crest, however, appears to carry 7 cusps; the median cusp is the
highest.

The enamel consists of two layers. quot;Stufenbildungquot; is moderately
developed. The enamel of the buccal and palatinal side of the ridge-
crest only consists of one layer, namely the inner one. Owing to
this peculiarity the hind and aft side of each ridge-crest is blackish,
while the outer and inner side possess the yellow whitish colour
of the inner layer. One would perhaps be inclined to believe that
the absence of the outer layer on buccal and palatinal side does
not represent the original state, but that it might be a result of
subsequent transport of the specimen. This is, however, highly
improbable, which may appear from the fact that the surface of
inner and outer side is provided with distinct, parallel wrinkled,
horizontal lines.

The grinding surface as well as the base of the crown are spir-
ally twisted. (See for the former fig. 2 of pl. XV). That is the
reason why the position of the gdnding surface could not be used
in the present case to determine to which side of the upper jaw the
molar belongs. In order to be able to show this feature in numbers,
we placed the median plane of the tooth vertically, and measured
the slope of the top of each worn ridge-crest, and also the slope
of the base of each ridge-crest. It will be clear, that the numbers,
obtained in this manner, are only approximately right, and have
only relative value, as we do not exactly know the position of the
present specimen in the upper jaw.

b = broken. — = slope from the inside outwards ) if the fangs are above and the
u - unworn. = „ ,, ^^ outside inwards ) crown below.

There cannot be a shade of a doubt that the specimen under
consideration belongs to Sf. airdivana. First, it shows all the essential
characters of the permanent upper molars of that species, and
secondly its measurements agree closely with those of the upper
third molars given by Janensch. The latter will be easily verified
by the help of table T.

The resemblance of this molar with the foregoing specimen is
so close, that there is no doubt, that the tooth in question is a third
upper molar of the same species.

This tooth also is implanted in a fragment of the maxillary. As
in the foregoing specimen a portion of the palate is present, enabling
us to classify the present specimen without any difficulty as a r. M^.
But suppose a fragment of the palate had been absent, we had sdll
been aljle to determine this specimen as a molar belonging to the
right half of the maxilla. For its spirally twisted grinding
surface is the reflected image of the plane of wear of the 1. M^,
just described.nbsp;\'

Both molars agree in almost every feature: form of the dendne-
islets, degree of quot;Stufenbildungquot;, thickness of the enamel, shape
of the crown, colour, and state of fossilization. Consequently it
seems almost sure, that the 1. M^ already described and the present
r. M3 belong to the same individual. The only diffquot;erence is the de-
gree of wear, the 1. M3 being somewhat more ground down than
the r. M3. But this cannot be used of course as an argument against
the assertion above stated, for it occurs frequently that the cheek

teeth of left and right side of the jaw present a different degree of
wear.nbsp;°

In the present specimen ten ridge-crests plus a posterior talon
are preserved so that probably two ridge-crests plus an anterior
talon have been broken off. Of ridge-crest 2 -f i only the enamel
of a portion of its posterior side is present; the palatinal portion
of ridge-crest 2 2 and the outer cusps of ridge-crest 2 4 are
damaged.

Ridge-crests 2 -f- 2, 2 3 and 2 4 present dentine-islets; the
cusps of ridge-crest 2 -f 5 are perforated; the mammillae of ridge-
crest 2 -f 6 flattened; the median cusps of ridge-crest 2 7 just
touched by wear. Ridge-crests 2 -f 8, 2 9, 2 10 and the very
low hind talon are untouched and covered by a thick coat of
cement; only the sides of these ridge-crests and the summits of
some of their cusps are partly free from, or project above the
cement.

Ridge-crest 2 -]- 7 carries 6 cusps. The succeeding ridge-crests,
being for the greater part concealed by cement, do not allow of
ascertaining the number of cusps, borne by each. Ridge-crest 2 2
requires more consideration. It presents namely a deep constriction,
somewhat palatinally of the median line. This median constriction
is flanked on either side by a less distinct lateral constriction. This
ridge-crest is therefore divided into four. The outer lateral con-
striction, as well as the median one, cannot be perceived in the
succeeding ridge-crests. The inner lateral constriction, however,
is more or less distinct from ridge-crest 2 2 up to and including
ridge-crest 2 7.

The buccal half of the anterior portion of the present specimen
shows a peculiar divergency. The transverse valley between ridge-
crests 2 -f I and 2 -f 2 is namely blocked by a rudiment of a ridge-
crest.

The degree of twisting of the grinding surface and of the base of
the crown has been expressed in the same manner as in the foregoing
specimen. The various values will be found in the table below.

It consists of that part of the cranium situated between a plane,
extending from the front side of the outer narial aperture to the
inner nares, and a plane, beginning at the front side of the M^ and
intersecting the incisive sheaths at a distance of about 15 cm. below
the outer narial aperture. It contains the W of the leftside, and the
fangs of the corresponding molar of the right side.

The postero-external part of the left premaxillary is lost. Only
the posterior part of the inter-alveolar fossa has been preserved
presenting a divergence of its side walls frontwards. The anterior
portion of the right canahs infraorbitahs is well preserved. Its front
is subcircular (diameter 4 cm.). Behind it is ovalshaped (7 x 4 cm.).

The palate is considerably injured. Its hindmost part, which is
undamaged, is concave in cross-section. Though of the r. W only
the fangs are present, it is easily stated that both molars are ap-
proximated in front and divergent behind (interval in front 6 cm. •
behind 10 cm.).nbsp;\'

Very fine specimen. The tooth carries ii ridge-crests, while it
is possible that a fore- and aft talon are present. The relatively great
width of the foremost dentine-islet namely, may be an indication
of the presence of an anterior talon, the dentine of which has now
been united with that of the first ridge-crest. This foremost dentine-
islet is only at its buccal- and posterior side bounded by an edge
of enamel.

The last four ridge-crests were absolutely concealcd by a thick
coat of cement. In order to be able to show the profile view of
these ridge-crests, the cement covering their palatinal sides has been
removed. At the same time it was shown that behind the last ridge-
crest a considerable amount of cement is present, which possibly
conceals a posterior talon.

The plane of wear as well as the base of the crown are clearly
convex. Ridge-crests i, 2, 3 and 4 present dentine-islets; the cusps

of ridge-crests 5 and 6 are perforated; the median cusps of ridge-
crest 7 flattened. The succeeding ridge-crests are concealed by
cement, only two median cusps of ridge-crest 8 project through
the cement. The worn summit of ridge-crests 2, 4 and 6 is concave
in the palatinal-buccal direction. The worn summit of ridge-crest 5
is flat, that of ridge-crest 3 convex. As a result the top of the buccal
extremity of the worn ridge-crest 3 is some 7 mm. lower than
that of the corresponding part of ridge-crests 2 and 4.

The paladnal portion of ridge-crests i and 2 is wanting. Two
inner cusps of ridge-crest 5 are damaged. Ridge-crest 2 presents,
besides small crenulations, three constrictions of the enamel viz.,
one constriction in the median line of the molar and two lateral
constrictions, resp. buccally and palatinally of the median one. Both
lateral constrictions can be followed up to ridge-crest 4. The enamel
of the posterior side of ridge-crest i shows only a median- and
a buccal inlet. We may be sure that the median constriction of
ridge-crests i and 2 indicates the presence of a longitudinal median
cleft, manifestating itself in ridge-crests 3 and 4 by a distinct division
of these ridge-crests into two parts, the buccal half being so much
in advance of the paladnal half that inner- and outer halfs of ridge-
crests 3 and 4 alternate. Ridge-crest 5 is straight. Its palatinal
extremity is not developed, so that its width is three quarters of
that of the preceding ridge-crest. Ridge-crest 6 extends nearly
straight across the crown; ridge-crest 7 and probably ridge-crcsts 8,
9, ID and II too. The number of cusps, borne by ridge-crests 6
and 7, is resp. ten and eleven. Ridge-crcsts 2, 3, 4 and 5 are clearly
chevron-shaped, the valleys between them are consequently V shaped.
The succeeding ridge-crests present more the form of a wall, which
fact is accompanied by an increase of the height.

The enamel consists of two layers. This could clearly be perceived
in the remnants of the enamel of the r. AP which is broken off
partly near, pardy below the base of the crown. The worn ridge-
crests present a moderately developed quot;Stufenbildungquot;.

The separadon between crown and fangs is disdnct. Inner- and
outer side of the tooth converge posteriorly, thereby showing the
characteristic shape of a third molar. As has been already mentioned,
the amount of cement is considerable. The coat of cement, which
covers the last four ridge-crests, does not present furrows, extending
in transverse direction straight across the crown, so that the position
of the posterior ridge-crests is entirely hidden by the cement.

The fragment itself does not present enough characters by which
it can be distinguished; the molar, however, does. The structure
of the anterior half of the crown is — as was mentioned above_

b.nbsp;the same relation between height and width of each ridge-crest;
the same thick coat of cement, covering the ridge-crests totally,\'

the same curvature of the base of the crown and of the upper
surface of the coat of enamel.

In brief, the resemblance is so extremely close, that I have
not the slightest doubt in referring the fragment described above
also to Stegodon airâwana.

This almost complete, finely preserved tooth having grinding
surface and base of the crown concave antero-posteriorly, shows
Itself to belong to the lower jaw. The right side being concave
proves, furthermore, the tooth to belong to the right ramus of
the mandible.

At the front the tooth is damaged, but the course of the fangs
shows decidedly that only the anterior side of the first ridge-crest
(plus perhaps a fore talon) are wanting. The lingual side of ridge-
crest 3 and the greater part of the inner moiety of ridge-crest 4
are also broken off. Only the upper portion (1-3 cm.) of the fangs
IS present. The tooth is slightly narrower in front than behind
Degree of wear:

Ridge-crests i, 2 and 3 present dentine-islets; the conelets of
ridge-crest 4 are perforated; the cusps of ridge-crest 5 are only
flattened; the median cusps of ridge-crest 6 arc just touched by
wear; ridge-crest 7 and the much lower hind talon are untouched
Ihe ridge-crests are tall, slender, chevron-shaped. The last five
ridge-crests slope backwardly. Ridge-crests i and 2 are so much
worn down, that it is difficult to determine whether or not they
slope backwards. The unworn ridge-crest is the highest in the

middle. Ridge-crest 5 consists of a big cusp on lingual and buccal
side and between these two five smaller ones. The clefts, which
separate the two lateral cusps from the median portion of the
ridge-crest, are much more distinct than those separating the five
smaller cusps mutually. We see, therefore, that this ridge-crest is
divided into three. The same may be seen in all preceding and
following ridge-crests. Ridge-crests i and 2 not only present two-
constrictions, which mark a lingual and a buccal cleft, but show
also a constriction in the median line. The latter constriction is
almost as distinct as the two lateral ones, lingually and buccally
of it. Probably it indicates the presence of a median, longitudinal
cleft in the first two ridge-crests. While we can easily follow the
two lateral, longitudinal clefts in the ridge-crests situated more
backwards, this is not the case with the median cleft. In the third
ridge-crest there is perhaps still a slight indication for its presence,
but it is totally absent in the following ridge-crests.

The structure of ridge-crests 6 and 7 is on the whole similar
to that of ridge-crest 5. The buccal extremity of the transverse
valley between ridge-crests 6 and 7 is for the greater part blocked
by a tall (2 cm.), rather sharp, accessory tubercle, which is closely
attached to the outer cusp of the seventh ridge-crest.

The hind talon is much lower than the last ridge-crest and placed
more to the inner side of the tooth. It consists of a large median
tubercle, flanked by two smaller tubercles on the inner side and
four on the outer side.

The outer extremity of each transverse valley — with the
excepdon of the valley between penultimate and ultimate ridge-
crests, which is occupied by the large accessory tubercle, already
mentioned — possesses a small incipient tubercle. The inner mouths
of the valleys are free from tubercles.

Ridge-crests i and 2 are nearly straight; ridge-crests 3, 4, 5
and 6 are slightly bent, the portion inwards of the buccal cleft
being in advance of the portion right of it. Ridge-crest 7 is almost
straight.

The transverse valleys are narrow. The first and second valleys
have but little cement; in the succeeding valleys the amount of
cement increases gradually, so that the valleys between ridge-crests 5,
6 and 7 and the space between the last ridge-crest and the hind
talon are completely filled.

The enamel is moderately thick and composed of two layers.
quot;Stufenbildingquot; is rather well developed. The enamel of each
fidge-crest presents, besides the two or three constncdons, small
regular crenulations. The degree of crenulation of the enamel agrees
exactly with that of the enamel of the cheek teeth oi Stegodon airmvana.

I have no doubt to refer the tooth under consideration to
Stegodon airdwananbsp;this lower tooth resembles in almost

evety essential character a r. M^ of that species figured by Janensch
in Die Pithecanthropus-Schichten auf Javaquot;, pl. XXII fig 4
Both specimens show the same form, the same degree of crenulation
of the enamel and of quot;Stufenbildungquot;, the presence of small acces-
sory tubercles in the outer extremities of the valleys, the presence
of a bigger accessory tubercle «auszen neben dem letzten Jochquot; i) 2).

It now only remains to consider the serial position of the tooth
in question. It will be remembered that its ridge-crest formula is
either yx or XyX. Furthermore, it will be known that cheek
teeth of J/, atrdwana have been described by Martin Janensch
SoergelS), Dietrich«) and Stehlin^. It is especially by Janensch\'s
excellent investigations that our knowledge of the present form
has been considerably enriched. Janensch gave the following
ridge-crests formulae:nbsp;^

In 1917 Janensch and Dietrich extensively described the
occurrence of the second upper premolar (P^) in a specimen of
the upper jaw of St. airdwana. Schlesinger ») showed, however,
that this supposed premolar in reality represents the first upper
milk molar (m^). Dietrich «) agreed with Schlesinger\'s re-
determination. The referred tooth carries three ridge-crests. Among
Soergel\'s material a r. Mg was represented, which this author
described and figured under the name of St. cf. airdwana. I have
no doubt that this specimen really belongs to St. airdwana. Though
914 ridge-crests and a posterior talon had been preserved, Soergel
did not hesitate to estimate the ridge-crest formula at X 11 X,

bigger accessory tubercle in (or\'near) the outer mouth of the
C X.nbsp;ultimate and penultimate ridgc-crest seems to be rather constant in

the M. , oi Stegodon airawana. Besides the example above mentioned, this peculiarity

wll n ^quot;a ^ \' ^S^tf::\'quot;\'\'\'quot;\'\'\'nbsp;^^^^^^ Verh. Kon Akad. v.

same author in: Samml. Geol. Reichs-Muscums Leiden. IV. 1884-1889 pl VI
hg. i. The latter specimen was originally referred by Martin with some diffidencc-
to St. trigomcephalus-, later on, however (Verh. Kon. Akad. v. Wetensch. XXVIII
1890. P- 10) to St. airdmana.nbsp;\'

») In cyration with Janensch. Sitz.-Ber. Ges. Naturf. Fr. zu Berlin. Jhrg. 19,6, 1917,

because the course of the fangs showed that about one ridge-crest
plus an anterior talon were wanting. In all probability the number
of ridge-crests in the lower third molar of the present species may,
therefore, vary from ii to 13. These are, however, the only additions
to Janensch\'s ridge-crest formulae, afforded by the investigations
subsequent to those of Janensch.

It will be clear that the specimen of our collection either represents
mg or Mj. Of the last lower milk molar but very little is known.
Janensch (1. c. p. 8) considered a fragment, consisting of three
ridge-crests, as representing a portion quot;of mg of i\'A airawana. Both
determination of the species and serial position appear to me
to be right. Stehlin (loc. cit. p. 5) described and figured the greater
part of a lower cheek tooth which he determined as M^ of St. airmvana.
He gave the length of the last three ridge-crests viz., (from front
to back) 73, 72 and 62 mm. In my opinion this specimen does not
represent a Mj, as Stehlin believed, but certainly a Mg. The
correctness of this assertion may be seen from comparison of the
measurements above stated with those of table U. From the same
table it will be seen that the present specimen of our collection
cannot but represent a M^ It is, therefore, the first specimen of Mj
which has hitherto been described.

Very fine specimen, containing on either side the posterior
part of the totally worn Mg and the whole of the partly worn Mg.

Horizontal ramus of either side and mandibular symphysis are
perfectly preserved. The ascending ramus of the left side is broken
off near the neck of the condyle, while the top of the coronoid
process has been lost. The ascening ramus of the right side possesses
a greater portion of the coronoid process than that of the left side.
For the rest, however, the right ascending ramus is broken off
some 13 cm. lower than the left one. The mandible is broken into
two in the mandibular symphysis, but both parts match exactly.

The lower border of the horizontal ramus is nearly straight,
that of the ascending ramus strongly curved (portion of the cir-

cumference of a circle with a radius of about 25 cm.). Alveolar
border and lower border of the horizontal ramus converge strongly
backwards. As a result the difference between greatest and smallest
height is considerable, and the height of the ramus at the base
of coronoid process is small. Inner and outer surface of the horizontal
ramus converge strongly frontwards, the ramus becoming consid-
erably narrower towards the symphysis.

The mandibular symphysis presents a spout-Hke, rather narrow
termination. Diastemal ridge and upper border of the symphysis
pass into each other uninterrupted. The diastema slopes obUquely
frontwards at an acute angle of about 25° with the inferior border
of the ramus. Diastema and alveolar margin intersect at an obtuse
angle of about 105°.

The anterior border of the upper portion of the coronoid process
is perpendicular to the lower border of the ramus. The upper
portion of the internal surface of the left ascending ramus presents
the oval-shaped orifice of the dental canal.

Both plane I and II are perpendicular to outer and lower surface
of horizontal ramus.

It will be sufficient to give a detailed description of the 1. Mg
and of the remains of the 1. M2 only, the corresponding molars
of left and right side being very similar.

Only the last two ridge-crests, absolutely ground down, are
present. The whole consists of an oblong fragment of dentine,
in the middle with two edges of enamel: remnants of resp. posterior
side of the penultimate, and anterior side of the ultimate ridge-
crest. Both edges are separated by a very narrow cleft, filled with
cement. There is no clear indication of a hind talon.

Fine specimen. Only the inner side of the first three ridge-crests
and the outer side of the fore talon are somewhat damaged.
Ridge-crest formula X11 X. This cannot be seen from the figures
because the photographs were taken before in the left ramus quot;bonequot;
material had been removed in order to expose the whole tooth.
It was then revealed, that besides the ten ridge-crests plus anterior
talon which can be seen in the figures, another ridge-crest plus a
very small and low posterior talon are present.

The tooth is considerably laterally bent; inner side convex,
outer side concave. The plane of wear is very slightly concave
antero-posteriorly. A slope of the grinding surface inwards or

outwards could not be stated. That part of the surface of the crown
which is not touched by wear, slopes obliquely from the outside
inwards.

The anterior portion of the tooth is so much worn down that
the dendne-islets of the first ridge-crest and of the fore talon are

united, presenting an irregularly shaped disk of dentine. A narrow
cleft, filled with matrix, separates M3 and remaining portion of M2.
An inlet of the enamel near the buccal side of the ridge-crest marks
the presence of the fore talon. Ridge-crests 2 and 3 present dentine-
islets; the cusps of ridge-crests 4 and 5 are perforated, some cusps
of ridge-crest 6 are touched by wear, the cusps of ridge-crest 7
and of all succeeding ridge-crests are untouched.

The amount of cement is considerable. The last four ridge-
crests plus the hind talon are totally covered by a thick coat of
cement; furrows, extending straight across the crown indicate the
presence of these ridge-crests. Some cusps of ridge-crest 7 project
above the cement, but others are still hidden, so that the number
of cusps of this ridge-crest cannot be ascertained. Ridge-crest 6
carries seven cusps; ridge-crest 5 ditto; ridge-crest 4 eight. The
narrow, deep valleys between the worn ridge-crests are for the
greater part still filled with cement.

The ridge-crests extend straight across the crown. Only ridge-
crest 2 presents a slight curvature. The ridge-crests are more or
less perpendicular to the curved sides of the tooth, consequently
they converge outwards. They are tall and narrow, but still clearly
chevron-shaped, the valleys are consequendy V shaped. The enamel
consists of two layers and consequently quot;Stufenbildungquot; is present.
Compared with the specimens described in the foregoing pages the
quot;Stufenbildungquot; makes the impression of being less distinct. This
impression is made, however, by the fact that the molars of this
lower jaw are strongly weathered. They totally lack the fresh
appearance of the foregoing specimens.

The enamel of ridge-crests i and 3 presents near the median
hne a constriction, which is more distinct than the remaining
constrictions. Perhaps this median constriction indicates the presence
of a median, longitudinal cleft in the anterior portion of the crown.

Crown and fangs are separated, which fact could be stated
at the outer side of the anterior portion of the tooth; everywhere
else the transition crown-fang is concealed by the bone.

The main difference between the molars of left and right side
of the lower jaw in question is that the r. Mg presents, besides
the two edges of enamel, already mentioned in describing the
1. M2, portions of two other edges of enamel namely of the
posterior side of the antepenultimate ridge-crest and of the anterior
side of the penultimate ridge-crest.

As to the identificadon of the present specimen I do not believe
that the reader will be convinced at first glance of the correctness

of the determination. Neither to me the specimen under consideration
made originally the impression of belonging to Sf. airdwana. I was,
however, misled by its weathered appearance, which not only
affects the distinction of the quot;Stufenbildungquot;, but also troubles
the degree of crenulation of the enamel. Only by blackening the
dentine material of the worn ridge-crests, it was possible to produce
a figure which enables the reader to form a notion of the structure
and of the degree of crenulation of the enamel. The figure shows,
especially when the magnifying lens is apphed to it, that the degree
of crenulation of the enamel is not so small as it seems to be at
first sight. Moreover, it must be borne in mind that the plane of
wear and the upper, unworn surface of the tooth are placed very
obliquely to one another, consequently the two frontmost ridge-
crests are worn down almost to the bottom of the transverse valleys.
And as a matter of fact the degree of crenulation becomes smaller
towards the base of the ridge-crests.

Neither ridge-crest formula (x ii x) nor measurements i) afford
any reason to suppose that the specimen might belong to a species
distinct from St. airdwana. Furthermore, I may call attention to
the remarkable likeness of the present lower jaw with the specimens
of lower jaw of St. airdwana, described and figured by Janensch
(1. c.) and Soergel (1. c.). Our specimen exhibits decidedly those
characters which appeared to be constant, viz., . . . die nach vorn
sich kräftig und ziemlich geradlinig verschmälernde Form des
Umrisses in der Ansicht von oben, die langgezogene, sehr flach
bogenförmige untere Kontur der Mandibeläste und die sehr geringe
Höhe derselben in der Gegend des Vorderrandes des Kronfort-
satzes . . . .quot;

Finally I may mention the following peculiar coincidence.
Janensch (1. c.) represents in text figs. 4 and 5 upper and side
aspect of a lower jaw of St. airdwana, containing the posterior
portion of cither M^ and 1. and r. M3. In this respect it resembles
closely our own specimen, which only differs by a slightly stronger
degree of wear, indicating somewhat higher age. The third molar
of Janensch\'s specimen resembles very closely those of our specimen
as may a. o. appear from table V where specimens B represent
1- and r. M3 of the referred lower jaw of Janensch\'s collecdon.
The peculiarity is that Janensch observes of the molars of his
specimen: quot;Die Zähne... machen den Eindruck, als ob sie von
der Verwitterung oberflächlich angegriffen wären. Auch der Schmelz
ist an den vorderen Jochen auffallend rauh und in seinen Kontupen

ziemlich unklar und verwaschenquot;. There exists, therefore, also in
th^ respect a remarkable resemblance between both specimens
1 do not mtend of course to say that the latter resemblance affords
an argument for specific identity. I only mentioned it to prove
that by weathermg the original structure of the ridge-crests of
the teeth of S/. airâwana may be troubled considerably.

In short, in my opinion there remains not the sHghtest reason
to regard this lower jaw of our collection as belonging to a species
distinct from St. airâwana.

Case n^ 17 of the fourth consignment contained a number of
specimens obtained from different excavations. According to the
original list of specification the present case contained also remains
from excavation 6, viz., quot;i cranium of a fossil elephant together
with I detached tusk and detached cheekteeth in a small case quot;i)
From this quotation one would be inclined to believe that\' the
detached cheek teeth (nine fragments) belonged to the same individual
as the cramum; the more so, if be added, that the detached tusk
really appeared to belong to it. Originally the present writer
believed this indeed. He had, however, to alter his opinion.
From the measurements given on p. 148 it will namely be seen
that the cranium in question belonged to an immature individual 2)
On p. 165 It will be seen that among the detached grinding teeth
above mentioned, two fragments occur which are worn, and which
represent the hinder portion of (in all probability) the third molar
As would already appear from what precedes, we may be sure that
m the cranium m question by no means the third molar was in use.
ihe cramum itself proves this assertion to be true. For though
It lacks grinding teeth it was not difficult to infer, that on either
side a sma 1 molar of about 11.5—12 cm. was in use. From the tables
given by Janensch it appears that it was in all probability the first
We are, therefore, wholly justified in regarding the

Injured specimen. No deformation. As already mentioned
probably the AP of either side was in use, but both are completely
hammered down to the margin of the alveoli. The maxillaries are
considerably injured. Behind both molars, mentioned above, and
on either side of the posterior nares a hole occurs, which is now
filled for the greater part with plaster. These holes represent the
places of left and right succeeding molar, which was certainly
not yet in use, and the crown of which was surely still partly
concealed by the bone.

Though grinding teeth are wanting we were able to determine
the Stegodon nature of the present specimen with absolute certainty.
After a careful examination, namely, small remnants of the enamel
have been found belonging to the posterior part of the succeeding
molar, mentioned above. And the enamel appearing to consist
of two layers, we may be sure, that the cranium under consideration
does not belong either to Mastodon or to Elephas.

The left incisive sheath has been pardy lost; the right one is
better preserved. The detached tusk was broken into three; the
parts, however, matched exactly, and after being glued together,
revealed that its proximal end fitted close to the portion contained
in the right incisive sheath. The incisor is almost straight in lateral
view of the cranium, but distinctly bent outwards, if regarded in
the upper aspect. Near the top it is compressed, the cross-section
being an ellips, the longest diameter of which is parallel to the
median plane of the cranium. More backwards the tusk becomes
gradually circular in cross-secdon.

The forehead is convex across. Consequently, the vertex cranii
presents a distinct, convex curvature, as seen from the aspect of
the occiput. In upper view of the cranium the vertex cranii is
convex backwards with a distinct depression in the middle.

The preservation of the lower half of the occiput is bad. There
IS a shapeless hole, indicating the posidon of the foramen magnum
and the median depression above it.

The walls of the outer nasal aperture are damaged so that a
hole has been formed, the dimensions of which are certainly greater
than those of the original nasal orifice.

Both zygomatic arches arc wanting. The preservation of the
left side of the cranium is somewhat better than that of the right
side so that we shall describe the former. The temporal fossa is
rather deeply concave from front backwards. As a result the temporal
contraction of the forehead is distinct. The crista spheno-frontalis

can be followed from\'above over a distance of some 13 cm In
upper view of the cranium the arcus supraciliaris manifests itself
by a shght curve, concave outwards. If we draw on either side
a tangent to the upper margin of the orbita and to that of the
temporal fossa, both hnes are approximated behind and divergent
m front. The degree of divergence is, however, small.

The interalveolar fossa is deep and presents the form of an
elongated oval, accuminated at both ends. The inner nares are
oval-shaped, with longest diameter in the median plane of the
cranium. The palate is narrow.

Total length (from vertex cranii to anterior extremity of
premaxillaries)............

If fig. I of pl. XX, in which the upper aspect of the present
cranium is represented, be compared with the corresponding aspect
of the young cranium of Janensch\'s collection i) it will be stated
hat besides a number of resemblances also some differences occur
And the question is: Are these differences essential or arc they
the result of different age and (or) individual variadon? Of some

that of Janensch inwards. Janensch i) describes the course of
the tusk as follows: quot;Bemerkenswert ist, dasz der dem Schädel
ansitzende rechte Stoszzahn bei schwacher gegen die Aiediane
konkaver Krümmung mit seiner Spitze über die Mitte nach links
hinausgewachsen ist. Würde der .... linke Stoszzahn die gleiche
Stellung wie der rechte gehabt haben, so müszten sie sich gekreuzt
habenquot;. I am, however, convinced that Janensch\'s specimen re-
presents in this respect a deviation. For it will be known that
even in those forms in which the tusks are curved inwards in old
individuals (e. g. Mammonteus primigenius Blumenb.) the course of
the tusks of younger specimens is such that the animal can use
them (e. g. for uprooting food).

As to the different form of the interalveolar fossa, it may be
observed that this is strongly affected, of course, by the development
of the incisors.

As to the remaining differences, however, it was necessary to
study a sufficient number of crania of one spccies in order to get
an idea of the influence of age and the range of variation. As f^ar
as my knowledge goes there is not one species of Stegodon of
which a sufficient number of crania has been described. Consequently
I had to turn to a recent species viz., to Elepbas nmximus L., of
which the National Museum of Natural History in Leyden possesses
a fine collection, consisting of about twenty crania of most different
age. Unfortunately time failed to take a great number of measu-
rements of these specimens so that I am not able to illustrate
the range of variation in the same manner as I did for kerabau,
banting and Rhinoceros sondaicus 2). Notwithstanding this hasty
examination it was not difficult to notice a.o. two prominent facts.

The relation between height and smallest breadth of the forehead
may vary considerably viz., from about 0.87 to about 1.26.
In determining the minimum and maximum value of this
relation no use has been made of the measurements of very
young crania. Furthermore, in measuting the height of the
forehead the median projection of the nasals is not included.
It appeared that the great difference between maximum and
minimum values is not only caused by differences in the relative
height of the forehead, but also by a different degree of con-
traction of the forehead by the temporal fossae.

After the publication of the present paper, 1 hope, however, to have the opportunity
of making a detailed study of the cranium of E. maxinius. Then I intend to
point out why in my opinion St. airänma and St. trignnocephalus arc not identical,
in spite of the fact that this has been asserted from authoritative side, namely by
Dubois (Tijdschr. Kon. Ned. Aardr. Gen., sec. sen, XXV, 1908, p. 1257) and by
Dietrich (Sitz. ber. Ges. naturf. Fr., Berlin, 1917, p. 126, note 2).

b. The posterior outhne of the cranium, as seen in upper aspect,
appeared to be remarkably constant. The depression in the
middle may vary a htde, but in all specimens examined, it is
decidedly present.

Provided with this knowledge we return to the comparison
of our specimen with the young cranium of St. airawana of Janensch\'s
collecdon. One of the most striking peculiarides of the latter specimen
is Its very high forehead, which is relatively little contracted by
the temporal fossae. In these respects our specimen differs not
mconsiderably, even if one is mindful of (i) the strongly damaged
outer narial odfice of our specimen, causing its forehead to seem
less high than it is in reality, and (a) the injury of left and right
postorbitary process of Janensch\'s specimen by which the im-
pression is made of a smaller degree of contraction of the forehead
than really exists. Janensch describes the forehead of his specimen
as follows; quot;Die Sdrn ist flach und sehr hoch. Ihre Höhe, vom
Oberrand der Nasenöffnung an gerechnet, beträgt 20 cm.,\'eben-
soviel wie die geringste Breite, die etwa unter der halben Höhe
hegt. Dabei ist noch zu berücksichtigen, dasz der von oben in
die Nasenöffnung vorragende mediane Vorsprung nicht erhalten
und deshalb nicht in den Betrag der Sdrnhöhe einbcrechnet ist.
Die Stirn verbreitert sich von ihrer schmälsten Stelle an nach oben
zu langsam und gleichmäszigquot; i). If one calculates, however, the
relation between height and smallest breadth of the forehead by
the help of the excellent figures of Janensch\'s specimen it will

our specimen the same reladon will be about 0.61. If we accept
that the range of variation in St. airâwana is not considerably smaller
than in E. maximiis (and why should it?) the different values for
the reladon mentioned does not afford a reason for specificdistinction.
The more so as our specimen presents a conformity with the
cranium of St. airawana to which I am inclined to attach much
value, after what I have seen in E. maximis. I mean the form of
the posterior oudine of the cranium as seen in frontal aspect.
A still younger cranium of St. airmvana of Janensch\'s collection
in which the posterior portion of the forehead was excellently well
preserved, showed a disdnct depression of the oudine mendoncd.
Ihe same feature is clearly exhibited in our specimen. In this
respect both crania differ essentially from the cranium of J^/./r/W
cephalus Martin.

as far as it has been preserved — no distinguishing characters
from the cranium of St. airdivana, there remains in my opinion
not the slightest reason to distinguish between the two.

Injured and slightly deformed specimen. Both zygomatic arches
are wanting; the occiput is very much damaged; the upper part
of the right incisive sheath has been lost; the outer nasal aperture
is so considerably injured that it is now in open communication
with both incisive alveoli.

The right side of the cranium is considerably better Reserved
than the left one. At the left side all is broken off, which — in frontal
aspect of the cranium — is situated buccally of the line, connecting
the outer side of the left incisive sheath with the innermost point
of the upper part of the temporal fossa of the same side.

Only the right molar is present, which is inferred to be the M^
from the characteristic shape of the crown, and from the absence
of any other tooth behind.

As has been already mentioned, the occiput is very much dam-
aged: the foramen magnum has become a big, formless hole, condyh
are absent.

We shall describe the right lateral side, as being best preserved.
The temporal fossa is distinctly separated from the orbita by the
crista spheno-frontalis, running from the postorbitary process of
the frontal to near the zygomatic process of the squamosal, and
presenting — in lateral view of the cranium — a very slightly curved
line, convex frontwards.

There is an abrupt transition of the upper part of the temporal
fossa into the forehead. The upper border of the former is nearly
parallel to the alveolar margin of the molar; anterior and posterior
border are almost perpendicular to the upper border, the lower
border is formed by the crista spheno-frontalis. Roughly spoken
the temporal fossa presents, therefore, the form of a rectangular
trapezium; it is deeply concave transversely with a slight longitu-
dinal convexity in the middle.

The irfra-orbital canal, which is rather well preserved, is oval-
shaped. The posterior lower angle of the orbit presents a hole
which for the greater part is filled with matrix, and indicating the
position of for opticum plus for. rotundum. In anterior aspect of
the cramum nothmg can be seen of the orbit, the arcus supraciharis
not contracting the upper surface.

The forehead is low and considerably convex across. Forehead
and upper surface of the premaxillaries are very markedly oblique
to one another, as seen in side aspect. The forehead is contracted
by the temporal fossae. The cranium broadens decidedly towards
the occipital surface. Stress may be laid upon the fact that in frontal
aspect the posterior ouriine of the cranium presents an uninterrupted
strongly convex line without any depression in the middlenbsp;\'

The outer nasal aperture is — as already mentioned — verv
much damaged. Only parts of the right wall, of the right half of
the posterior wall, and of the right side of the median projection
of the nasal of that side have been preserved. As, moreover the
interalveolar fossa is complete, we are able to form an idea of the
shape and the dimensions of the outer nares.

left M the fangs are present, it can easily be seen, that the molars
were approximated in front and divergent behind. The interalveolar
fossa IS well preserved. Those parts of the incisive sheaths bordering
it, are perpendicular to its bottom. The middle portion of the right
wall of the fossa even slopes from the left to the right, probably a
result of some deformation. Near the nasal aperture tL interalveolar
fossa is shallow, more frontwards the depth increases quickly still
more frontwards the fossa becomes again more shallow. The side
walls diverge frontwards. Some 20 cm. from the tips of the pre-
maxillaries the alveolar fossa ends abruptly in a lobe on either side
ot the median hne. The sutute between both ptemaxillaries presents
a very slight curvature, and is situated somewhat left of the median
line. Illustrating some deformation.

Total length (from vertex cranii to anterior extremity of
premaxillaries......................^

Interval between tips of postorbital process of dther side
Breadth of forehead at occipital surface .......

105

64
93

Rather injured and damaged specimen i). The tooth, having
Its anterior ridge-crests very much worn down and damaged, it
IS difficult to ascertain the ridge-crest formula. In order to determine
whether or not a hind talon is present, some cement of the hindmost
part of the tooth has been removed. It was revealed, that a posterior
talon is absent. There is no reason to consider the last ridgc-crest to
be a hind talon, as it does not disdnguish itself from the preceding
ridge-crests. The presence of ten ridge-crests can easily be ascer-
tained. Ridge-crest lo from behind, however, does certainly not
represent the foremost part of the tooth, for it is still preceded by
a disk of tooth material. The transverse diameter (length) of this
disk is considerably smaller than that of the remnants of the fore-
going ridge-crest. The well-preserved lower part of the paladnal
side of the ridgc-crest, last mcndoned, passes namely uninterrupted
into an undamaged plane, forming the front side of the palatinal
portion of the ridgc-crest in question. In this manner the impression
IS made as if the ridgc-crest under consideration really represents

I regret that the figure of the molar is so defective, not at all giving a fair idea of its
structure. The cranium was, however, so heavy and the situation in the old building,
m which the Utrecht Institute of Geology and Mineralogy was settled at that time,
such, that it was impossible to find a place which allowed of a better exposure.

small pieces of the posterior side l,av bL„ n eservel
3 shows a small portion of the enamel of tranterior sll Tnd
he greater part of the enamel of the posterior side nne ajouter
ide of ndge-crest 4 have been lost; inner side of ridgtcrestYtoo

seem to be convex from behind. Ridge-crest 8 carries six cusps;
the palatinal one, which is much damaged, is the biggest. The
number of cusps, borne by the succeeding ridge-crests could not
be ascertained, because the ridge-crests are for the greater part
concealed by cement.

The inner extremities of the valleys between ridge-crests 5, 6,
7 and 8 present a low, blunt, accessory tubercle. A big accessory
tubercle is fastened to the front side of the inner cusps of the ultimate
ridge. It can clearly be seen in the figure.

The enamel, which is distinctly crenulated, is very thick and
consists of two layers. Inner- and outer layer are nearly equally
thick. The degree of quot;Stufenbildungquot; being in direct proportion to
the thickness of the outer layer, might lead us to expect a strong
development of quot;Stufenbildungquot;. That this is less conspicious,
is caused by the fact that the molar is considerably weathered.

It will already be understood from what precedes that the
amount of cement is considerable.

89c

22P

As to the specific determination we may make the following
remarks. The relation between the height of the forehead (median
projection of nasals not included) and smallest breadth of ditto is
24

Jn dealing with the specific determination of the foregoing specimen
It was seen that in the young type cranium of St airdivana the same
relation amounts to 0.81. Having found, furthermore, that this
relation varies already from 0.87 to 1.26 in sixteen crania of the
recent E, /;/axi/////s, there should be no reason to consider on
this ground our specimen specifically distinct from St. airawana.
As already pointed out on page 150 we may, however, in all
probability attach much value as a specific character to the form

of the posterior outline of the cranium in front aspect. Though
the vertex cranii of our specimen is slightly injured, it was not
difficult to see that a median depression of that outline is undoubt-
edly absent, which distinguishes in my opinion our specimen decid-
edly from St. airawana, and points in the direction of St. trigono-
cephalus Martin. And there are more indications in that direction.
In the young type cranium of St. trigonocephalus the relation between
height and smallest breadth of the forehead is — according to

respect there exists, therefore, a closer resemblance between our
specimen and the type specimen of St. trigonocephalus than with
the type cranium of St. airawana^), though it will be hardly necessary,
I believe, to point out, that this resemblance alone would be of
very little value.

Still another indication is afforded by the cranium. It will be
known that Martin named St. trigonocephalus after the triangular
shape of the cranium in front aspect. According to Dubois^)
this form is not original. I have seen both type crania of J\'A trigono-
cephalus. They are contained in the National Museum of Geology
and Mineralogy in Leyden. As far as the young cranium is concerned
there is in my opinion not the slightest reason to believe that its
triangular form should not represent the original state. In our
specimen, being a cranium of an adult animal, the incisive sheaths
are much more strongly developed than in Martin\'s young type
cranium. Consequently the triangular form has totally been lost in
the anterior half of the upper aspect. In the posterior half it is, ho-
wever, the more distinct. Unfortunately this feature is not clearly
shown in- figure i of pl. XI, which represents the cranium not only
as seen from above, but also from the right side. It is, however,
very distinctly shown by comparison of two measurements, namely
(i) interval between tips of postorbitary process of either side
64 cm., and (2) breadth of forehead at occipital surface 93 cm. In
this respect our specimen is even more quot;trigonocephalus-XiVo!\'
than the young type cranium of that species. In the latter specimen,
namely, the relation between the measurements is — according

1) Jaarb. Mijnw. Ned. O.-L, 1887. Wetensch. Ged., pl. ill, fig. i; also in Samml. Geo).
Reichs Mus., Leiden, IV, 1884—\'89, pl. Ill, fig. i.

If, furthermore, the side aspect of our specimen be compared
with that of the young cranium of St. trigonocephalus i), it will be
seen — after making allowance of course that there exists a great
difference in age, and that the greater part of the occiput of our
specimen has been lost — that also on the basis of this comparison
no sufficient distinction can be drawn.

We have still to consider the question whether the M^ of our
specimen is apt to strengthen or to weaken our opinion. Martin
believed to notice a number of structural differences between the
grinding teeth of St. airdwana and St. trigonocephalus. Our knowledge
of the teeth of the former species have since been considerably
enriched, especially by Janensch\'s thorough investigations. If one
tests the differences, mentioned by Martin, by these more ample
materials, and also by the experiences gained in the Siwalik stego-
donts, it appears that many of them do not hold good. Janensch 2)
added some more differences, but concluded at last: quot;Sicherlich
wird man Zahnfragmente beider Arten nicht immer aus einander
halten könnenquot;. In my opinion the differences by which the grinding
teeth of St. trigonocephalus may be distinguished from those of
St. airdwana are: (i) thicker enamel, especially of the outer layer,
(2) consequently a more distinct quot;Stufenbildungquot;, (3) generally
more strongly indented enamel, and (4) greater size.

In describing the r. M^ of the present cranium mention has
already been made of the thick outer layer of the enamel. At the
same time it was pointed out, that its quot;Stufenbildungquot; is strongly
troubled by the weathering of the crown. The latter fact also
affects the impression which the enamel makes to the observer,
regarding its degree of crenulation. Nevertheless, we may say
that in this respect our j\\P does not show any clear indication,
neither in the direction of St. trigonocephalus, nor in that of St. airdwana.

Finally the size. Comparison between the various measurements
of the r. AP of the present specimen with the corresponding ones
of upper third molars of St. airdwana will reveal that the former
distinctly exceeds in size almost all upper third molars of St. airdwana.

In my opinion, therefore, not a single essential difference stands
in the way of reckoning our specimen to St. trigonocephalus Mart.

Fine fragment of the mandible, comprising the whole of the
horizontal ramus and the lower portion of the ascending ramus
of either side; containing on either side a molar which is inferred
to be the Mg from the ab^sence of the indication of any other tooth
behind.

The mandible is broken into two in the symphysis, but both
parts match exactly. The alveolar border of both molars is damaged.
Nevertheless fig. 4 of pl. XIV, shows distinctly that the difference
in height of the horizontal ramus respectively at anterior and
posterior extremity of the molar is far less pronounced than in
the lower jaw of St. airâwana, described in the foregoing part.

The diastema descends obliquely forwards at an acute angle
of about 35° with the inferior border of the ramus. The mandibular
symphysis is produced into a relatively broad, spoutlike termination.
Its lower border is damaged. As seen in crown view, and compared
with the lower jaw of St. airmvana, above mentioned, the distinctly
curved outer border of left and right horizontal ramus enclose a
considerably greater angle. On the whole the present ramus makes

a more robust impression. The outer surface of the horizontal
ramus presents two foramina mentaha, the one situated some
8 cm. below the anterior extremity of the alveolar margin, the
other about 8 cm. in advance of the former and nearer to the inferior
border of the ramus.

Fine specimens. Only the anterior part is damaged. Besides
small differences both teeth are very similar so that it will be sufficient
to describe one of the two detailed, e. g. the left one.

As the front side of both molars is damaged, it seems — at
first sight — impossible to ascertain the true number of ridge-crests
carried by each of them. Fortunately, however, in the 1. Mg a small
portion of the front surface of the fangs has been preserved. It
is situated below the anterior side of the foremost ridge-crest,
and proves with absolute certainty, that this ridge-crest is really
the first one. That is to say: not a single ridge-crest has been lost.
The ridge formula is, therefore, 8 X.

The tooth is distinctly curved outwards. The base of ridge-
crests I—7 are more or less of the same length; from ridge-crest 8
inner and outer side of the molar converge suddenly backwards.

The base of the crown and the grinding surface are strongly
concave. The grinding surface slopes obliquely from the outside
inwards. The plane of wear and the base of the crown converge

anteriorly, enclosing a very small angle. The separation between
crown and fangs is very clear, especially on the mner side of the

All the ridge-crests and the hind talon are abraded by wear. Ridge-
crests 1—6 present dentine-islets; the cusps of ridge-crests 7 and 8
are perforated; the cusps of the hind talon are flattened

The hind talon consists of two thick cusps of which the lingual
one is the biggest. Three constrictions of the enamel of ridge-crest 8
mark the presence of four cusps. Ridge-crest 7 possessed five or
six ridge-crest 6 six mammillae. The preceding ridge-crests being
too much worn down, it is impossible to estimate the precise
number of cusps, borne by each. The first three ridge-crests show
clearly, that near the base of the crown the enamel is thrown into

The form of the dentine-islet of the second ridge-crest is ap-
proximately that of an elongated rectangle, only slightly curved,
Ld but little wider in the middle. The dentine-islet of ridge-crest 3
and of the succeeding worn ridge-crests differ from that of ridge-
crest 2 in having inner and outer extremity of the rectang e acu-
minated. This is - at least partly - caused by the fact, that he
grinding surface is not only concave antero-posteriorly, but also

converge buccally. The elongations of the first and of the last ridge
intersect at an angle of about 30° (in r. M3 ca 3^°). To be accurate,
it must be added, that this is not quite true. For, as the grinding
surface is spirally twisted, intersection will never take place. 1 he
twisting of the plane of wear is such, that the slope increases from

The ridge-crests are, with respect to those of the corresponding
molar of St. airdmna and of St. trigonocephalus, considerably lower
and much more clearly chevron-shaped; consequentiy the valleys
are in cross-section also more distinctly V shaped.

The enamel is moderately thick and consists of two ayers
(see the buccal cusps of ridge-crest 7). A distinct Stufenbildung
can, however, only be seen in the last three ridge-crests. In this
connection attention may be drawn to the fact that in the molars
of the present specimen enamel, cement and the greater part ot
the dentine had the same, irregularly spotted appearance. As a
result the crenulation of the enamel was very indistinct, and it was
only possible to make it distinctly visible in the photograph by
blackening the dentine of the tooth. Another result is, however, that
quot;Stufenbildungquot; though present, is in general not clearly exhibited.

There is no indication of the presence of a median longitudinal
cleft. Cement is present, though in small quantities. The valleys
between ridge-crests i, 2, 3, 4 and 5 possess but little cement.
More backwards the amount is somewhat less small, not only
the base of the valleys being filled to a greater extent, but also
the lingual and buccal sides of the ridge-crests, and the posterior
side of the hind talon being covered by a thin coat of cement.
Noteworthy is the stronger, local development of cement near
the buccal side of ridge-crest 7.

As has been already mentioned, the r. Mg resembles in almost
every feature the 1. Mg.

Differences are :
a. The r. Mg is somewhat less ground down.
h. The r. Mg does not present the strong, local development of

cement near the buccal side of ridge-crest 7.
c. The hind talon of the r. Mg is less developed, and still totally
covered by cement.

Total length of r. M3 190 mm. | „.casurcd along median line of crown.
Ditto „ 1. Mj 197 „ )

86c
18

86c
22

62c

6ic

84c

79c
20

32c

I2C
I2C

In adopting Osborn\'s classification of the Proboscidea, the
following genera and species of S. and E. Asia must be reckoned
to the subfamily of the Stegodontinae.

Stegolophodon latidens (Clift) 1828, Trans. Geol. Soc., Ser. 2, II, p. 371.
Stegolophodon stegodontoides (Pilgrim) 1913, Rec. Gcol. Surv. Ind., XLIII, p. 294.
Stegolophodon cautleyi (Lydekker) 1886, Mem. Geol. Surv. Ind., Ser. 10, III, p. xiv
(of introduction).

Stegolophodon cautleyi progressus Osborn 1929, Am. Mus. Nov., n°. 393, p. ij.
Stegolophodon nathotensis Osborn 1929, Am. Mus. Nov., n°. 393, p. 13.

Stegodon sinensis Owen 1870, Qu«t. Journ. Geol. Soc. London XXV^^^ p^ 4 7-
(See also: Matsumoto, Sci. Rep. Tohoku Imp. Univ. Sendai, Japan, Sec. Ser. (Geol.},

Stegodo\'n^aur^rJe^) (Matsumoto) 1918 Sci. Rep. Tohoku Imp Univ.. Sendai, Japan,

u Ber rSoL Anthr. Ethn. Mus\'oresden, 1887. n°. 6, p. 5 (as St. tngonocephdus .

SteVd\'on trigonocephalus Martin 1887, Samml. Geol. Reichs-Mus. Leiden, IV, p. 56.
(Also: Jaarb. Mijnw. N. L, 1887, p. 12).

The above list does not pretend to be complete as far as the
papers quoted are concerned. For the rest the same remarks hold
good which have already been made on p. 125 m connection with

It will be evident that it will not take us much trouble to exclude
from our comparison most of the species above mentioned. As
a matter of fact all the species of Stegokphodon may be left out ot
consideration, being all very primitive forms which in my opinion
may as well be reckoned to the family of the Mastodontidae. ^

On the other hand, most of the remaining species are consid-
erably more progressive than the form to which the lower jaw

With the exception of St. clifti and St. bombifrons the remaining
forms are readily distinguished by a higher ridge-crest formula
of the Mo or, if this molar has not yet been found, by higher ridge-
crests greater amount of cement, and more cusps a ridge-crest.

There remain, therefore, only the two species already mentioned
viz St. clifti and St. bombifrons. It is, however, not difhcuit to
dismiss also the former. According to Lydekker the lower
third molar of that species carries either seven or eight ridge-
crests. In this respect there exists therefore accordance between

gt;) At present St. ganesa is often considered to be the male form of St. insignis.

^ E/ephas. In 1929 this author again referred to it as genotypic spec.es of a new m
namely Parasliodon (Mats. 1929). Dietrich (Neues Jhrb. 9. quot;I. ^
pointed out that quot;ParastegoJonquot; aurorae is very closely alliednbsp;f\'^V^/

still slightly more progressive. The present writer strongly doubts, therefore, the
right of Parastegodon to generic disdnction.

our form and the species referred to. Nevertheless there cannot be
any doubt of the specific distinction of either, Lydekker (1. c.),
mentioning that . . the plane of wear is the same as in Mastodon\'quot;
i. e. sloping outwards, instead of inwards as in all remaining
Stegodon species.

The Mg of St. bomhifrons carries — according to Lydekker i) —
eight to nine ridge-crests. In general the valleys of the grinding
teeth of St. bomhifrons are completely filled by cement i), whereas
in the teeth of our specimen cement is present in but small quantities.
On the other hand, however, Lydekker reckons to that species
a fragmentary lower jaw, containing Mg of either side, which
possess quot;. . . hardly any cementquot; The different amount of cement
would, therefore, not give any argument for specific distinction.
As, furthermore, the structure of the grinding teeth of Javan and
continental form is in several points very close, the writer provi-
sionally named the specimen in question St. bomhifrons, before it
was sent back to Bandung. He overlooked, however, some diffe-
rences which are now esteemed important enough to separate both
forms.

Among the lower third molars, which Lydekker reckons to
St. bomhifrons, only two specimens are tolerably perfect. They will
be found represented in F. A. S., pl. XXV, figs. 3 and 3a, and
pl. XXIX, figs. 7 and 7a. According to the description of the plates
the dimensions of the former are 12.9 X 4.4 inch = 32.8 X 11.2 cm.,
and of the latter 13.4 X 4.2 inch = 34 X 10.2 cm. Comparison
between these data and the measurements of 1. and r. Mg of our
specimen shows that the corresponding molar of St. bomhifrons is
not only considerably larger, but also that the Mg of the Javan
form is much broader in reference to its length than appears to
be the case in the Mg of the continental form.

There seems to be one more difference. Janensch observes:
quot;Bei den Stegodonten der Sivaliks habe ich eine nennenswerte
Stufenbildung nicht gefundenquot;. Unfortunately I have not had the
opportunity to visit the British Museum, so that I myself could
not make comparisons in this respect. The latter remark is added
not because I doubt the correctness of Janensch\'s asserdon,
but because quot;nennenswertquot; is a rather elastic conception. As the
enamel of the teeth of our specimen, however, is distinctly composed
of two layers, and the last three ridge-crests show a distinct quot;Stufen-
bildungquot;, also in this respect both forms probably differ.

In my opinion we may, therefore, conclude, that the present
lavan form is distinct from all other species. I propose the specific
name of bondolensis, indicating the locahty from whence it has
been procured.

Remains to describe twelve fragments of molars, ten of which
are unworn, two abraded by wear. On page 146 mention has
.already been made of nine of them, namely the two worn and

The various fragments carry from one to four ridge-crests.
Not a single fragment allows of ascertaining the ridge-crest formula
of the tooth, to which it belonged. Owing to the broken condition
of the base of the crown of nearly all the fragments, it is only in
some cases possible to determine whether the specimen belongs

Two fragments have been longitudinally and vertically bisected;
another specimen and a (half) ridge-crest of one of the specimens
vertically bisected have been sawn up in a horizontal direction.

As the ten unworn fragments all carry equally-shaped ridge-
crests, and as these ridge-crests carry the same average number
of cusps, it is almost sure, that they belong to the same species.
A detailed description of all the ten is, therefore, unnecessary;
an enumeration of the essential characters will be sufficient.

The ridge-crests are tall, rather narrow and clearly chevron-
shaped They are placed closely together, so that the form of the
valleys between them is that of a sharp V. This is clearly shown
in those specimens, which have been vertically and longitudinally
bisected, and in those fragments which have lost the cement.
Most ridge-crests carry from ten to twelve cusps. Of two fragments
the cusps are somewhat thicker and blunter, while their average
number is eight. The impression is made as if they belong to
another species than the remaining unworn specimens. In order
to get an answer to the question, whether this impression was
right or not, both a specimen with eight cusps a ridgc-crest, and
a specimen with ten to twelve cusps a ridge-crest were sawn up
in a longitudinal-vertical direction. They are represented respectively
in figs. 3 and 5 of pl. XIII. These cross-sections reveal, that both
fragments are absolutely alike, as far as regards essential characters.
Since therefore, the only difference between both fragments in

question, is the number of cusps, there appears to me to be every
probabihty that all unworn specimens belong to the same species.

We will now continue with the enumeration of the characters.
The cusps of the ridge-crests are most elevated in the middle,
rendering them strongly convex across. Some ridge-crests extend
straight across the crown, but most show a slightly S-shaped
curvature.

The enamel consists of two layers. The average thickness of
the enamel is 6 mm. The ridge-crests, if worn, would have
shown therefore quot;Stufenbildungquot;. Those ridge-crests, which have
been horizontally bisected, show that the enamel is considerably

A large amount of cement is present. Some fragments do not
possess cement anymore, but the former presence of a large amount
of cement is still clearly indicated.

We have still to consider the two unworn fragments. Both
consist of the posterior part of the original tooth, in all probability
a third molar, which is inferred from the form and size of the
fragments. Both contain the hind talon and the last two ridges.
The resemblance between the two specimens is extremely close. The
ridge-crests and the hind-talon are disdnctly convex backwards.
The hind talon consists of six cusps. Ridge-crests and talon are
abraded by wear. The enamel consists of two layers. quot;Stufenbildungquot;
is present and rather well developed. The enamel — at any rate
the inner layer — is strongly crenulated.

The amount of cement was — without any doubt — considerable.
The ridge-crests arc chevron-shaped. Inner and outer side of the
fragment converge posteriorly, and pass gradually into the curved
posterior side.

I think it advisable to leave all the specimens specifically un-
determined, being in too fragmentary a condition.

Before starting with the description of those specimens of our
collection which belong to the genus E/ephas in its widest sense,
we shall have to pay attention to Osborn\'s classification of the
Elephantidae.

In 1925 this author pubhshed i) a paper entitled: quot;Final
conclusions on the evolution, phylogeny and classification of the
Proboscideaquot; 2).

In this important paper Osborn divides the Elephantidae in
six different quot;racesquot; (subfamilies) viz.,

Already in 1924 3) Osborn had proposed the subdivision of
the subfam. Eoxodontinae in:

Loxodonta F. Cuvier. quot;Progressive from unknown ancestors to
L. africana and distinguished by narrow grinding teeth with few
ridge-plates and prominent quot;loxodont sinusquot; quot;;

Sivalikia,.n. g. quot;typified by Loxodonta namadica Falconer, type
species, and distinguished by broad grinding teeth, numerous
ridge-plates, and absence of quot;loxodont sinusquot; quot;

Pilgrimia, n. g. quot;typified by Elephas falconeri Busk, type species,
.... distinguished by narrow grinding teeth, numerous ridge-plates,
quot;loxodont sinusquot; vestigial or absentquot;

It is Osborn\'s thirty-sixlh communication sincc 1900 on that subjccti (See: Osdorn,
Fifty-two years of research, observation and publication. New-York, 1930, p. 86). All
these and subsequent preliminary papers will culminate in Osborn\'s monograph to
be entitled: quot;The Proboscidea; Evolution, Phylogeny and Classification.quot; (L. c. p. 57).
Am. Mus. Nov. n°. 152, 1924, p. 2.

*) ^c statement of the absencc of a loxodont sinus must be regarded cum grano salis,

UsDoiw himself mentioning on p. 31 of his referred paper of 1925 quot;.........absencc

Osborn had, however, to replace his generic name Sivalikia
by Valaeoloxodon Matsumoto 1924 i), a name a few weeks prior
to the former 2). Matsumoto created Valaeoloxodon as a subgenus;
Osborn uses it in a generic sense.

Though Osborn\'s before mentioned paper of 1925 has the title
of quot;Final conclusions____it appears from subsequent communi-
cations of this author, that later on several additions and changes
have been made. At this place only those changes may interest
us which concern the classification of the Elephantidae. The latter are
to the effect that all the genera of mammoths, namely, Mammonteus,
Parelephas and Archidiskodon are united in the subfamily of the
Mammontmae

As far as my knowledge therefore goes, Osborn\'s present
classification of the Elephantidae may be represented as follows:

h. Subfam. Loxodontinae:
Loxodonta F. Cuvier.
Palaeoloxodon Matsumoto.
Sivalikia Osborn.

Manniionteus Camper (Northern Mammoths).
Parelephas Osborn (Trogontherian Mammoths).
Archidiskodon Pohlig (Southern Mammoths).

After the description of the specimens of our collection, wc
shall have the opportunity to see to which genera the various fossil
forms of \'\'Elephas\'\' of South and East Asia arc reckoned.

Journ. GcoI. Soc. T6ky6. XXXI, 1924. p. ^57 amp; 260.
\') See OsnoRN, Am. Mus. Nov. nquot;. 460, 1951. P-

The full description of Pahrolosodonh given by ^IAT^quot;OTO m Sci. Rep.T6hoku
Imp. Univ.. Sendai, Japan, Sec. Ser. (Geol.), Xlll, i, 19^9. P- 7 (O-
lt;) See OsnoRN, Journ. Am. Mus. Nat. Hist. XXX, 3, i93°. P-nbsp;Am. Mus. Nov.

Fine specimen of the lower jaw, containing in each ramus the
finely preserved Mg. This specimen is broken in the symphysis,
but the two parts match exactly. As appears to be the rule in fossil
lower jaws of elephants, both condyli have been lost, and both
coronoid processes are for the greater part broken o\'ff. Of the
left process a greater pordon has been preserved than of the right one.
The hinder side of each ascending rami is furthermore consid-
erably injured.

If the figures of this specimen be compared with the adult
lower jaw of Elephas maximus, it will be seen that the resemblance

between the two is extremely
close. The posidon and num-
ber of the various mental
foramina may perhaps some-
what differ, but it will be
known, that the position and
number of the mental foramina
may vary considerably. And
just the same may be said of
the form and shape of the
symphysial gutter.

The lower border of the
horizontal ramus is in its front
portion almost straight; more
backwards, however, a slight
curvature occurs, passing im-
perceptibly into the strongly curved, lower border of the ascending
ramus. The alveolar border of the molar is clearly concavc; the
inner alveolar border is more concave than the outer one. Alveolar
border and lower border of the horizontal ramus converge poste-
riorly; the jaw becomes lower from front to back. At the base
of the coronoid process the height of the horizontal ramus is,

however, still considerable. The diastemal ridge descends obliquely
at an acute angle of about 64° with the lower border of the
ramus horizontalis.

As seen in upper view, the convexity of the outer side of the
horizontal ramus increases strongly towards the symphysial gutter.
The angle enclosed by the anterior portion of the outer side of
both rami is consequently a very obtuse one. The symphysial
gutter is narrow and comparatively short. The front border of
the coronoid process is perpendicular to the lower border of
the ramus.

As both M3 arc exceedingly alike, it will be sufficient to describe
one of the two in detail, c. g. the left one.

Very fine specimen. Though the tooth is somewhat injured
in front, not a single ridge-plate has been lost. The molar carries
twenty-three ridge-plates. If the fore- and hindmost ones are supposed

to be resp. a fore- and aft talon, the ridge-plate formula is
X 21 X Perhaps, however, the hindmost ridge-plate does not
represent the aft talon, because we must take into consideration
the possibility that a low posterior talon is still concealed by the
quot;bonequot; of the jaw. In that case the ridge-plate formula would

The relatively narrow crown tapers strongly from ridge-plate 9
towards the posterior extremity. This feature combined with the
high ridge-plate formula makes it sure, that this molar is the last
of the tme molar series. From ridge-plate 7 the crown also tapers

The inner side of the tooth is strongly convex, the outer side
concave; the curvature of the latter is considerably less than that
of the inner side. The anterior talon plus fifteen ridge-plates are
in use The worn anterior talon is almost totally confluent with
the disc 1) of the first ridge-plate. Only a small portion of the
separating enamel plates on the outer side of the tooth indicates

The grinding surface is clearly concave m profile view and
somewhat spirally twisted. The plane of wear^ slopes obhquely
from the outside, which is higher to the insi^de, which is
lower. The grinding surface of left and right M3 form therefore

^ ^OnTy one slightly worn cusp of ridge-plate 15 can be seen.
The other digitations are still concealed by cement. Ridge-platc 14
already shows four digitations, two of which are ground down
into distinct annular discs. Ridge-plate 13 presents five annular
discs the middlemost of which is smaller than those immediately
buccally and lingually of it. The structure of ridge-plates 12 11
and 10 resembles closcly that of ridge-plate 13. As a resuh of the
increase of the degree of wear towards the anterior extremity
of the molar, the three middlemost discs of ridge-plates 11 and 10
have become confluent into a compound oval-shaped disc, while
the buccal and lingual discs have also got a more oval shape.

Ridge-plates 12, n and 10 are therefore clearly divided into
three; the middle portion of each ridge-plate consisung of three
digitations. The longitudinal clefts, which cause the triple division,
have been lost in the foregoing ridge-plates, which arc too much
ground down. The buccal longitudinal cleft, however, is still

Ridge-plate 8 and all foregoing ridge-plates are worn down
into transverse discs. The discs of ridge-plates i and 2 arc rectangular

shaped. Those of ridge-plates 3—8 present the form of a very flat V,
the limbs of which converge posteriorly, while each limb is slightly
curved, convex frontwards. The lateral extremities generally point
frontwards. Anterior and posterior enamel plates of ridge-plates 3—8
each are in general mutually parallel, and no mesial dilatation
occurs, except in ridge-plate 7. This is caused by the fact, that
the anterior enamel plate of the latter is almost swaight in the middle.
It may be pointed out, however, that a dilatation, formed in such
a manner, often may be observed in the ridge-plates of the molars
of Elephas maximus.

The enamel is dictinctly quot;chanelledquot; i), and strongly quot;crim-
pedquot; or quot;plaitedquot; 2). This crimping especially occurs in the
posterior plates of the well worn ridge-plates. Those ridge-
plates, which are less worn (10 and 11) only present secondary
quot;flexuresquot; caused by the confluence of the discs of the separate
digitations.

The ridge-plates are close packed, the intervening valleys
therefore being comparatively narrow. The anterior five ridge-plates
are so much ground down, that the separating valleys have become
narrow clefts.

The envelopment by the cement of ridge-plates i—14 is such,
that the sides of this part of the molar are totally concealed, so
that in profile view nothing can be seen of their position. As seen
in the aspect of the grinding surface, however, the position of
ridge-plates 6—10 can be ascertained, as the cement in the intervening
valleys is worn down 5—10 mm. below the level of the rim of
the discs. It was shown that ridge-plates 6—10 are inclined forwards.
In the succeeding ridge-plates the slope gradually increases. As
portions of the buccal side of ridge-plates 15—21 project through
the coat of cement, it was possible to state that these ridge-
plates arc distinctly inclined backwards. The ridge-plates diverge,
therefore, towards their bases in a fan-shaped manner, which is
a common phenomenon in the M3 of the genus Elephas contrary
to the Mo, the ridge-plates of which depart but slightly from

As has been already mentioned 1. and r. M3 are very much alike,
so that a detailed description of the latter would be of no interest.
The only difference is, that instead of fifteen, only thirteen ridge-
platcs arc in use (front talon not included).

\') See: Falconer, P.1I. iMcm. h, p- 146 and 147-
») L. c., p. 147.
\') L. c. p. 147.

Greatest length of each ridge plate in the level of the grinding surface..........

Interval between the middle of two ridga-plates, measured in median line of the grinding surface

The resemblance between our specimen and some of the
mandibles of E. maximus, which I saw, is extremely close, both
regarding structure and shape of M3, and form of lower jaw.
I should therefore not have hesitated to identify it with that species,
were it not, that Dubois mentions that a close resemblance exists
between the grinding teeth of E. fmximus and E. hjsudrindicus,
the fossil Javan form of his collection. Before the cranium of
E. hjsudrindicus had been found Dubois accordingly believed that
E. maximus was represented among the Kendeng fauna. There is,
therefore, a possibility that our specimen belongs to Dubois\'s species.
On the other hand it must be borne in mind that the Mg of the
mandible of our collection carries at least twenty-one ridge-plates,
whereas the third molar of E. hjsudrindicus possesses — according
to Dubois — at best nineteen ridge-plates. As moreover Dubois
did not mention any more structural detail of the teeth of his
species, and as we are not able to distinguish our specimen from
E. fmximus, we believe to be wholly justified to class it with the
latter species. The addition of a sign of interrogation to the specific
name in the heading may indicate the possibility of its belonging
to E. hjsudrindicus Dubois.

\') Only in the left half of the mandible material has been removed in order to expose
the whole length of the molar.

47
ca 46

i ii

Superb specimen. Only the front talon is damaged. Ridge-plate
formula X 7 X. Shght tendency to x 8 x, as behind and closely
attached to the posterior talon two small, low tubercles occur.

The front talon, all the ridge-plates, and the back talon are in
use, though of the posterior talon only the two median cusps
are flattened by wear. The grinding surface is slighdy convex,
and there is a peculiar slew or twist in the wear, the inclination
not only diminishing towards the posterior extremity of the tooth,
but even becoming in ridge-plates 6 and 7 the reverse of that of
the preceding pordon of the grinding surface.

The plane of wear being convex proves the tooth to be an upper
grinding tooth, and sloping obliquely from left to right 2) shows
that the tooth belongs to the left side of the upper jaw. Its serial
position will be considered afterwards.

As seen from above the crown presents — roughly spoken —
the form of a rectangle, the posterior extremity of which is rounded
off regularly, and the greatest diameter of which is very slightly
curved, concave buccally.

Crown and fangs are very distinctly separated from one another.
The base of the crown is in profile view very slightly convex;
this convexity also proves the tooth to be an upper molar. In
profile view the tooth shows approximately the form of an isosceles
trapezium. The front talon, ridge-plate i, and the palatinal moiety
of ridge-plate 2 are supported by an anterior fang, which curves
considerably to the palatinal side. Another anterior fang, supporting
the buccal portion of ridge-plates 2 and 3, and which also curves
to the palatinal side, is situated buccally of and behind the fang already
mentioned. Both are separated by a third fang. The remaining
ridge-plates are supported by a coalescence of fangs. All the fangs
are closely attached to one another.

The front side of the tooth presents a disc of pressure which
is clearly portrayed in pl. XVII, fig. 2. The worn anterior talon,
which is for a good deal confluent with the disc of the first ridge-
plate is mainly situated in front of the palatinal moiety of ridge-plate i.
The posterior talon carries four digitations, the two median of
which are thickest and highest, and flattened by attrition. Ridge-
plate 7 also carries four cusps; ridge-plate 6 appears to consist
of 5 conelets. In the preceding ridge-plates, the number of digitations
borne by each, cannot be estimated, the ridge-plates being too
much worn down. The two median digitations of ridge-plate 7
are worn down into irregular annular discs, the palatinal cusp is
just perforated, the buccal one only flattened. The worn summit
of this ridge-plate slopes very slightly from the inside out-
wards. This fact already appears from the increase of the degree
of wear of this ridge-plate from buccal to palatinal, but could
also easily be stated by regarding the molar in question from
behind.

From fig. I of pl. XVII it will be seen that in ridge-plate 7
the two median digitations are thicker than the buccal and palatinal
ones. The structure of ridge-plate 6 resembles that of ridge-plate 7
very closely. There appears only to be one cusp more, while the
three palatinal discs have just become confluent. Those two discs
of ridge-plate 6, which are comparable with the two median discs
of ridge-plate 7, are marked by their relatively great antero-postcrior
thickness. Their position is such, that the valley between ridge-
plates 6 and 5 is partly obstructed. There is a distinct cleft in the
middle of ridge-platc 6. Though ridge-plates 5 and 4 are already
in a more advanced state of wear, their general structure can easily
be understood after comparison with that of the sixth ridge-platc.
The posterior enamel plate of ridge-plates 5 and 4 throws out
two curved projections, separated from one another by a sharp
inlet, indicating the presence of a similar cleft as in ridge-plate 6.

The anterior enamel plate of ridge-plate 5 shows palatinally
of the median line of the crown an angular projection, which almost
touches the posterior enamel plate of ridge-plate 4. A similar angular
projection of the anterior enamel plate is still more distinctly visible
in ridge-plate 4. But in the latter it is situated more to the median
line of the tooth. Ridge-plates 5 and 4 present therefore an im-
perfect lozenge shape.

Ridge-plates 3, 2 and i, though considerably worn down,
show in general the same structure as the succeeding ridge-plates.
The inlet of the posterior enamel plate near the median line of
the crown is still indicated. The angular projection of the anterior
enamel plate appears to disappear gradually after advanced wear
of the ridge-plates.

The enamel is thick; crimping (folding, plaiting, festooning)
is but slightly developed, and disappears near the base of the ridge-
plates. The enamel of buccal and palatinal sides of the ridge-plates
is rugous, and provided with irregular, anastomosing, vertical
grooves.

The wide valleys are filled with cement. The buccal and palatinal
sides of the ridge-plates are in general free from cement. Small
remnants of cement on these sides reveal, however, that they were
probably covered by a thin coat of cement.

In the aspect of the grinding surface, the \'quot;»l^^ho^^y
the form of an oval, with a contraction some ^ «n ™ the

have become confluent into a transverse band. R\'^ge-pl^« 7 F«^
a prominent, curved projection of its ameriornbsp;p\'quot;^nbsp;^

becomes convex. This triangular form of the dentine-islet, whereby
the top of the triangle is directed forwards, can easily be recognized
in ridge-plates 6, 5 and 4. It will furthermore be seen that towards
the base of the crown the projection of the anterior enamel plate
gradually disappears.

The relatively wide valleys are V shaped, becoming very narrow
clefts if the ridge-plates are ground down. They are in the
middle obstructed by the projection of the anterior enamel plate
or — if the ridge-plates are more ground down — by the fact,
that the two halves of the anterior enamel plate converge frontwards.

The enamel is thick and rather strongly plaited. The plication,
however, is coarse. The enamel of inner and outer sides of the
ridge-plates is rugous. The valleys are filled with cement. Outer
and inner sides are in general devoid of cement, but the buccal
side of the crown offers clear indicadons that cement has been
lost, and that the inner and outer sides of the ridge-plates have
been covered by a coat of cement.

Interval between the middle of two
ridge-plates, measured in the
median line of the grinding
surface...........

Before pointing out, how we came to the conclusion that the
two specimens of our collection belong to Archidiskodon planifrons,
wc shall have to consider how many different forms of \'\'Yllephas\'\'
have hitherto been found in South and East Asia, and to which
genera they are reckoned.

Falconer and Cautley disdnguished three forms viz., E, plani-
frons, E. hjsudricus, E. namadicus.

The first is reckoned by Osborn i) to Archidiskodon, who
followed in doing so Pohlig, the creator of that genus. Matsumoto
esteemed it nccessary to found on two fragmentary, worn grinding
teeth, figured under the name of E. planifrons by Falconer and

\') Proc. Am. Phi!. Soc., LXIV, i, 1925, P-
\') Jap. Journ. Gcol. a. Gcogr., V, 1926—\'27, P- 2i4-

Cautley in the F. A. S. (pl. XI, fig. 4, and pl. XIV, fig. 8) not only
a new species, but even a new genus, for which he proposed the
names of Leith-Adamsia siwalikiensis. I wholly agree with Dietrich i)
who criticized this redetermination with the words: quot;Wert Nullquot;.

The systematic position of E. hysudricus is still somewhat un-
certain. PoHLiG 2) has proposed to idendfy this species with
E. {Archidiskodon) meridionalis. His view has, however, not
been generally adopted. Osborn reckoned in 1921 the species in
question to the Mammontinae, but added: quot;The position of E. hysu-
dricus in this phylum is doubtfulquot;. The cranium referred to this
species by Falconer is not of the Mammontine typequot;. In 1924 s),
however, he reckons it — though provided with a sign of inter-
rogation — to the Elephantinae. There is indeed a great possibiUty
of its belonging to that subfamily, Lydekker already having
expressed as his opinion that E. hysudricus might be the direct
ancestor of E. maximus, and Dubois \') having mentioned a fossil
form from Java which is in his opinion intermediate between
E. hysudricus and E. maximus {indicus). The latter species Dubois
accordingly named E. hysudrindicus.

Stehlin who determined a collection of fossil vertebrate
remains, obtained from the vicinity of Bumiaju, believed to^ecognize
among his material three different species of quot;Elephasquot;. He left
them specifically undetermined, but indicated them as spec. I, II
and III. Stehlin\'s species III might be identical with the form
which we identified as A. planifrons.

E. namadicus became the subgenotype of Matsumoto\'s new
subgenus Valaeoloxodon As already mentioned Osborn uses
Valaeoloxodon in a generic sense. Some investigators, among who
Dietrich 1®),considernamadicusP. aniiquus identical.
Osborn 11), however, regards them as distinct species. This author 12)
is even inclined to found a new genus {Hesperoloxodon = loxodonts
of the West) on a form from Italy, described by him under the
name of P. antiquus italicus.

Sitz. her, nicdcrrhcin. Gcs., Feb. 4th, 1884,
Sec Osborn 1. c. 1925, p. 31.
Am. Mus. Nov., n°. i, 1921, p. 14.
Am. Mus. Nov., 11°. 152, 1924, p. i.

») Sec Sci. Rep. Tohoku Imp. Univ., Sendai, Japan, Sec. Ser. (Geol.), XIII, i, 1929, p. i.

Makiyama 1) distinguished two different subspecies {namadi and
naumanni) of the species under consideration. According to
Matsumoto, Makiyama laid in doing so . . . the corner-stone
of further progress in the study of Elephas namadicus Falconer
amp; Cautley in Japan . . . 2). Matsumoto added still another
new subspecies, jahei and also a new species under the specific
name of tokunagai Dietrich opposed these determinations
strongly. Without agreeing with this author in every respect, there
is also in my opinion no sufficient reason to separate these specimens
from P. namadicus.

Under the name of E. trongontherii Pohlig, Makiyama described
and figured two fragmentary grinding teeth from Japan. It is the
first record of the occurrence of that species in Japan. Osborn \')
reckons this species to his new genus Parelephas In 1926
Matsumoto founded on some detached grinding teeth and on
a left mandibular ramus, containing a finely preserved molar, two
new forms of Parelephas, namely Parelephas protomammonteus tjpicus
and Parelephas proto?mmt?ionteus proximus. According to Dietrich iquot;)
Matsumoto\'s forms belong partly to E. antiquus, partly to a new
primitive race of true mammoth for which he proposes the name
of E. primigenius matsumotoi n. var. In Dietrich\'s opinion, further-
more, not only the belonging of the specimens to the left or to
the right side is not always rightly determined, but also concerning
the serial position misdeterminations occur. As to the latter objections
I wholly agree with Dietrich. As to the redeterminations of this
author, however, I wonder, how he has been able to form so
definite an opinion by the help of figures alone, the preservation
of some specimens being such, that the figure — though exceedingly
fine — only gives an absolutely insufficient idea of the structure.
Furthermore it must not be forgotten that Mammonteus primigenius
is a form, which — as far as my knowledge hitherto went — did
not range so far South in E. Asia.

In Naumann\'s paper quot;Uebcr Japanische Elephantcn der Vorzeitquot;
a short description of a well-preserved grinding tooth o£E. primigenius

*) Sci. Rep. TOhoku Imp. Univ., Sendai, Japan, Sec. Scr. (Geol.), Xlll, i, 1929, p. i.

\') According to Osborn (I. c.) all English authors, namely Falconer, Lr.irn Adams,
and Lydekki\'.r, confused the teeth and cninia of the various spccics of Pardephas
with those of the true mammoth, Mammonteus primigenius.

*) Sci. Rep. TOhoku Imp. Univ., Sendai, Japan, Sec. Ser. (Geol.), X, 2, 1926, p. 43 (,).

will be found i). Naumann himself mentions, however, that the
specimen has been possibly obtained from N. Asia.

The occurrence of E. maximus {indicus) in the fossil state
in Japan was for the first time recorded by Leith-Adams 2).
Lydekker doubted strongly the correctness of the determination,
and reckoned the specimen to E. namadicus. In 1927 Matsumoto
stated as his opinion that Leith-Adam\'s determination was right.
On the basis of a few differences he distinguished the Japanese
form under the new subspecific name of buski. To this subspecies
Matsumoto reckons also a tooth described and figured by Naumann
(1. c. p. 31, pl. VII) under the name of E. namadicus

Fossil teeth of E. maximus have been furthermore recorded
by Naumann «) from Malakka, and by Martin (under the name
of E. sumatranus) from Sumatra \'), and from Banka «).

At last Osborn described two new primitive forms of EJepbas,
namely E. platjcephalus and E. platycephalus augustidens.

It will be evident, I believe, that in what precedes no effort
has been made to discuss all the papers dealing with the present
subject. Only those papers have been used which were necessary
for the composition of the following list. The latter contains all
fossil forms of Elephantidae (exclusive of Stegodontinae) of S. and
E. Asia, which are known to me, and which have in my opinion
right to specific or subspecific distinction.

Quart. Journ. Geol. Soc. London, XXIV, 1868, p. 496. (Busk, Add. Remarks, id.
p. 498).

*) Sei. Rep. Tohoku Imp. Univ., Sendai, Japan, Sec. Ser. (Geol.), X, 3. 1927, p. 57 (1).
\') Already Lydekker-(1. c. p. 169) was convinced of the specific identity of this tooth

with Leith-Adam\'s specimen.
«) Abh. u. Ber. K. Zool. etc. Mus. Dresden, 1887, n°. 6, p. 9.
\') Samml. Geol; Reichs-Mus. Leiden, 111, 1883—\'87, p. 5.
») Ibid., IV, 1884—\'89, p. 13.
•) Am. Mus. Nov., n°. 393, 1929, p. 21 amp; 22.

It will need no further explanation why the two specimens of
our collection, which were determined as belonging to A. planifrons,
belong indeed to a primitive form. Already for that reason alone
we may exclude the greater part of the species occurring in our list.
There remain: A. planifrons, E. hysudricus, E. platycephalus and E.
platycephalus augustidens. Of the dentition of the last two forms but very
little is known, Osborn making only a few remarks on the structure
of the IVP, and only figuring the only specimen of molar, namely the
unworn type IVP of the subspecies. From this figure appears that the
ridge-plates are higher, narrower and more closely approximated
than in our specimens. As far as can be gathered from Osborn\'s
unworn specimen there seems to be no mesial dilatation of the ridge-
plates. Besides these few remarks, however, we shall have to leave
these forms out of consideration, the denddon being too little known.

Proceeding with the dentition of A. planifrons and E. hysudricus
we borrow the following ridge-plate formulae (exclusive of talons)
from Lydekker i).

Being an upper tooth with ridge-plate formula X 7 X, it will
be seen that if it really belongs to one of the two remaining species,
it represents either m^ or m^ of E. hysudricus, or m^ or Mi of

There are but a few specimens known of m- and m-^ of E. hysu-
dricus. They are mentioned by Lydekker in his Cat. Foss. Mamm.
Br. Mus., IV, 1886; they are figured in the F. A. S., and their
dimensions will be found in the quot;Description of Platesquot; Borrowing
data from these works, we arrive at the following table.

7- (8—9) • (10—12)
7- (8—9) • (10-13)
(9—12). (10—12). (15—17)
(9—12).(12—13).(14—18)

\')nbsp;I YDrKK..R Kivcs as ridge plate f..rmula of the last upper m.lilt;molar (9-11). Me men-

^nbsp;tiX^owiler twlt;gt; sjecfmens (n-. JI14 «quot;d 16,94) wh.ch carry but 7 r.dge-plates.

The dimensions of our specimen are 125 x 53 mm. Supposing it
to belong to E. hjsudricus, it may only represent the ultimate milk-
molar. I am, however, convinced, that the molar in question does
not belong to that species. Not only that in the aspect of the grinding
surface our specimen differs widely from the m® of E. hjsudricus
represented in fig. 3 of pl. VI of the F. A. S., but also comparison
of the lateral aspect reveals important differences. As the figures,
mentioned in the foregoing table, represent milkmolars still fastened
in the upper jaw, the latter comparison must be made by the help
of the lateral aspect of a lower, detached m^ of the referred species
(F. A. S., pl. VII, fig. 8). It will be seen that the ridge-plates of
our specimen are lower, and that the separation between its crown
and fangs is much more distinct. So we may exclude E. hjsudricus.

In the following table the reader will find the dimensions of
m^ and M^ which Lydekker reckons to A. plamfrons.

As already mentioned the dimensions of our specimen are
125 x 53. All the specimens of the foregoing table are therefore wider
than our specimen, while the length of the latter is intermediate
between those of and M^ of the table. Furthermore it will be

more closely to the M^ than to the m^ of A. plamfrons. It will be remem-
bered (i) that according to Lydekker, the m^ of the latter species
carries six to seven, the M^ seven ridge-plates, and (2) that our own
specimen shows a slight tendency to a transmission of a tooth with
seven ridge-plates into one with eight. Supposing its belonging
to A. planifrons, it is therefore more probable that it represents
a M^ than a m^.-

It cannot be denied, however, that the difference in size between
the Ml of A. planifrons represented in figs. 4—5, pl. VI of the
F. A. S. and our specimen is considerable. On the other hand
Lydekker, in dealing with the grinding teeth of the referred species,
mentions: quot;...the relative width and the absolute size of the
teeth . . . vary very considerablyquot; i). Tliis fact is clearly illustrated in

the following table, in which I united the dimensions of the various
complete Mg reckoned by Lydekker to A.planifrons. The dimensions
are again borrowed from the F. A. S.

If therefore the various characters of our specimen should appear
to be in accordance with those of the molars of A. planifrons, its
aberrant size and shape may never be used as an argument against
specific identity. Now, if the figures of our specimen be compared
with those of the molars which Lydekker reckons to A. planifrons,
it will be seen that the former is absolutely indistinguishable. Con-
cerning (i) the relative height of the ridge-plates, (2) the thick
enamel, almost quite devoid of plication, (3) the expansion of
the ridge-plates in the middle, assuming an imperfect lozenge shape,
(4) the distinct separation between crown and fangs, the molar
in question resembles so closely those of A. planifrons, that in my
opinion it may in all probability be classed with that species.

We shall now return to our lower molar, represented in pl. XV,
figs. 5—6. Its ridge-plate formula is X 7 X - Supposmg it to belong
either to A. planifrons or to B. bjsndricus, it represents either a nig
or M. of the former, or a or mg of the latter.

The data of both following tables have again been derived
from Lydekker, and from Falconer and Cautley.

The dimensions of our specimen being 105 x 46 mm. it will be
clear that it may be either a mg of E. hjsudricus or of planifrons. Of
these teeth but very few specimens have been figured in the F. A. S.
As furthermore these figures are on far too small a scale, and as
the milkmolars of both species are in several characters very much
ahke, it is extraordinarily difficult to answer the question, whether
our specimen belongs to A. planifrons or to E. hjsudricus.

Our specimen presents, however, some characters, which seem
to point in the direction of A. planifrons. They are:

a.nbsp;The coarse crenulation of the thick enamel. (In the teeth of
E. hjsudricus the enamel is either very thin, and much plicated,
or thicker with slight plication).

b.nbsp;The relatively low ridge-plates. The estimated height of the last
ridge-plate does not exceed 42 mm.

c.nbsp;The dimensions correspond far better with those of the mg
of A. planifrons than with those of the mg of E. hjsudricus,
especially if we take into consideration the fact, that our specimen
has lost the greater part of the cement covering its inner and
outer side, so that the molar seems to be some mm. narrower
than it is in reality.

Wc conclude therefore: The specimen in question either belongs
to A. planifrons, or to E. hjsudricus, but its belonging to the former
spccies is most probable. In that case it represents a last lower
milkmolar.

\') According to Lydekker (1, c. p. 120, nquot;. 16621) a m» (mm4), but in all probability
an error, the number of ridgc-platcs being seven or eight, the dimensions being very
small, and Falconer (Descr. of Plates, p. 8) considering the specimen to be a sccond
milkmolar (m,).

Besides tlie cranial and dental remains already described,
our collection contains seven femora, three humeri, two tibiae,
three ulnae, two radii, and one left half of the pelvis. From the
following enumerations it will appear that the bulk of these spe-
cimens has been obtained from the vicinity of Bumiaju.

Excavations 1—4 and 8 are situated at the base of the verte-
brate bearing series. Of proboscidean cranial remains found in
excavation 8 our collection contains only one mandibular ramus,
namely, the one described in the foregoing pages under the name of
Tetralophodon htwiiajuensis n.sp. From excavations 1—4 the other
remains of that species of our collection were obtamed. Moreover
one badly preserved grinding tooth, and the posterior portion
of the cranium of Mastodon sp. Neither in these excavations
remains of Stegodon or quot;Elephasquot; were found. It is therefore very
probable that the proboscidean skeleton remains procured from
these excavations belong to Mastodon, more particularly to Tetra-

lophodon humiajuensis. The same cannot be said of the specimens
which have been found in the Tji Djedjawai and Tji Saat, both
streamlets, which break transversely through the vertebrate bearing
strata. For it must be borne in mind that from the vicinity of
Bumiaju also Stegodon and \'■quot;\'Elephas\'\'\' remains have been procured.

PoHLiG^) — who studied the proboscidean skeleton remains
collected by the Trinil Expedition of Mrs. Selenka — already
pointed out the great lack of a sufficient number of data for com-
parison, experienced by everyone who is occupied with the deter-
mination of skeleton remains. As to proboscidean bones Pohlig\'s
paper supplies this want for a deal. Accordingly we shall use it
repeatedly.

We shall begin with the femora, more particularly with their
state of preservation.

i. Text fig. 27. Upper half of left femur of gigantic
size. It is very unfortunate that this specimen is not complete, as
the remaining portion is perfectly preserved. It is characterized
by a low trochanter major.

Text fig. 28. Right femur (Nquot;. 5).
Hind aspect; 15 = Inner aspect; C = Front aspect.
About Vu nat. size.

N°. 2. Pl. XVII, fig. 4. Left femur. Restored with plaster.
Front side, distal and proximal extremity injured. Consequently

\') quot;Zur Osteologie von Sugodonquot; in quot;Die Pithecanthropus-Schichten auf Javaquot; Leipzig
1911.

the fossa intercondyloidea seems wider than it already really is.
Smallest width 35 mm.

N°. 3. Right femur. Middle pordon of corpus wanting.
Great resemblance to N°. 4.

4. Pl. XVII, figs. 5—7. Right femur. Rather well pre-
served ; considerably restored. Fossa trochanterica better developed
than would appear from fig. 7.

N°. 5. Text fig. 28. Right femur. Fine specimen; was broken
into two in the middle of the corpus. Both parts have been glued
together later on. External epicondylus feebly developed.

N°. 6. Right femur. Badly preserved; proximal and distal
extremities broken off.

N°. 7. Probably right femur of elephant. Very badly preser-
ved. Only pordon of corpus.

In table W the measurements of our specimens will be found
together with those mentioned by Pohlig i). It will be remembered
that femora n®®. 2 and 6 have been found in excavation 8 from
which excavation no other remains than of Mastodon have been
obtained. And the question is whether it is possible to identify
2 and 6 as belonging to Mastodon by the help of the data, given
by Pohlig. According to that author, the long limb bones of
Mastodon are (i) thick and short and (2) characterized by the almost
total absence of a torsion: quot;Die hintere Begrenzungsebene der
Condylen ist bei Mastodon der vorderen des proximalen Femurrandes
fast parallel.quot; 2) From n°. 6 only two measurements could be taken,
which are in almost perfect accordance with the corresponding
ones of n°. 2. But of course only the latter can be used for comparison.
If the relation between total height and smallest breadth of the
corpus be calculated, it will be seen that n°. 2 seems to be slightly
more slender than Pohlig\'s specimen of St. airdwana. If, however,
the relation between total height and smallest girth of corpus be
calculatcd, it will be revealed that in this respect our specimen is
intermediate between St. airdwana and M. giganteus, and distinctly
separated from E. mxims. In all probability Pohlig attached most
importance to the latter relation, as he did not give the value of the
smallest breadth of the corpus femoris in the Cambndgc specimen
of M. giganteus. As to the angle of torsion % however, n°. 2 has
exactly the same value, which Pohlig mentions of the femur of

St. airawana. As to 4 and 5 this angle is even considerably
greater; in this respect, therefore, decidedly more quot;Elephas-likequot;
than St. airawana. Peculiar is, however, that the relation between
total height and smallest girth of corpus in n«®. 4 and 5 points in
the direction of a form which is still more quot;Mastodon-likequot; than
M. giganteus. Conclusion: By the help of Pohlig\'s data it is not
possible to determine our specimens of femora. I repeat, however,
that in all probability n^s. 2 and 6 belong to Mastodon.

N°. I. Text fig. 29. On the whole very fine specimen. Was
above the middle broken into two, but both parts have been glued

N°. 2. Considerably smaller
than n°. i. Nevertheless belonging
to an adult individual as no separ-
ation of the epiphysis can be
perceived. Strongly damaged. Caput
totally wanting. No essential diffe-
rences with specimen n°. i. Right
specimen.

N°. 3. As to size intermediate
between n«». i and 2. Still more
damaged than no. 2. A. o. distal
extremity wanting. For the rest
no differences with n°. 2. Right
specimen.

Pohlig\'s specimens of humeri
are so damaged that comparison
with our specimen is hardly pos-
sible. From table X, in which the measurements of the humeri of
our collection and those of Pohlig\'s best preserved specimen
have been united, it will be seen that n°. i, though not inconsi-
derably greater, shows in general exactly the same relations as
Pohlig\'s specimen. N°. i may therefore belong to St. airawana.

As already mentioned all the ulnae and radii of our collection
have been found in excavation 8, and belong therefore in all
probability to Mastodon. The belonging of ulna n°. i to radius
n°. 2 and of ulna n°, 2 to radius n°. i could be ascertained with
absolute reliability.

Ulna n°. i plus radius n°. 2. Pl. XVI, figs. 4—7. Very fine
specimens of the right side. Only slightly restored. Proc. anconacus
of ulna broken off. Radius but reladvely shghtly curved.

Text fig. 29. Right humerus (N°. i).
A = Hind aspect; B = Front aspect.
About Vi5 nat. size.

Ulna 2 plus radius n°. i. Left specimens. Very close resem
blance to foregoing pair. Somewhat less better preserved. Proc
anconaeus wanting. From table Y and Z it will be seen that while
the two ulnae are nearly of equal size, radius n°. i is somewhat thicker
and longer than radius n°. 2.

I and 2. It wants the distal extremity and the greater portion
of the olecranon. The planes of fractures seem to indicate that
the epiphyses have been lost, so that the specimen may have belonged
to an immature individual.

wo dieLängsrichtung der carpalen Gelenkfläche nahezu perpendikulär
zu derjenigen der proximalen ist, fast parallel stehen beide bei E. indiens
bei E. africanus and E. meridionalis etwas mehr dem Verhältnis von Ste-
godonangtn\'i\\\\enquot;lt appeared,however,that this character is not easy
to handle. What must be understood by quot;Längsrichtung der carpalen
Gelenkflächequot;, the latter being approximately triangular in shape ? If
we consider the line through top and middle of base of the triangle as
representing the longitudinal direction,there occurs indeed but a slight
torsion. If, however, the external side of the triangle be taken as longi-
tudinal direction, there exists an angle of about 40° between the latter
and the longitudinal direction of the proximal plane of articulation.

As to the ulna Pohlig i) observes : quot;Wie alle anderen langen
Knochen, so ist auch die Ulna bei Stegodon und Elephas schlanker,
graciler als bei Mastodon .........quot; From table Z it will be seen im-
mediately that in this respect the two complete ulnae of our collection
point decidedly in the direction of Mastodon. From which we may
infer that our first supposition of what Pohlig meant with quot;Längs-
richtung der carpalen Gclenkflächequot; of the radius is right.

At last the tibiae and the pelvis. All the three are very fragment-
ary. Tibia n°. i consists of the proximal portion; n°. 2 is still
more fragmentary, proximal as well as distal extremity being
wanting. The latter has been found in excavation 8, and belongs
therefore probably to Mastodon. The fragment has a length of 44 cm.,
smallest breadth is 7,7 cm. and smallest girth 24 cm.

The left half of the pelvis will be found figured in pl. XVII,
fig. 3. The ilium is very much damaged; the ischium has been preser-
ved for the greater part; of the pubis only a small portion is present.
Furthermore the border of the acetabulum is damaged. Dimensions
of the latter 16 X 14 cm. This specimen, having been found in
excavations i—4, belongs in all probability to Mastodon.

The molars of this species appeared to show very great affinities
to Tetralopbodon longirostris. According to Von Koenigswald i),
however, the form of Bumiaju is easily distinguished from the
European species by the absence of lower incisive tusks.

The remains did not allow of determining the species. We
were able, however, to show that the Bumiaju form is specifically
distinct from the species of Dubois\'s collection viz., Hippopotamus
sivajavanicus Dub.

The question is: Are these determinations apt to strengthen
or to weaken the opinion of an upper pliocene age of the Bumiaju
beds? The prevalent opinion is that \'\'Mastodonquot; docs not occur
in the Pleistocene of Eurasia, whereas it does appear in that of
America. This supposition, however, remains to be proved.
Matthew 2), namely, in correlating European with Amencan
mammal faunas, came to the conclusion that it is either necessary
to revise the American succession downward, or the European
succession upward. In the latter case \'\'Mastodon\'\' should occur,
therefore, also in the Pleistocene of Europe. In a quite recent
publication Osborn does not cxclude this possibility cither

\') quot;The Geologic Age of Pitljecanthropusquot;. Rcp. Centenary Meeting Brit. Ass. for the
Advancement of Science, London 1931, p. 451. London 1932.

*) The correctness of this affirmation may appear from the following quotation:
quot;Eoantbropus dawsoni was the companion of Arcbidishodon planifrons and Anancus
arverntnsis, hence of upper Pliocene or lower Plcistoccne agequot; (1. c. p. 432).

It will be known that Pilgrim i) places the Upper Siwaliks
in the Pliocene. Matthew (1. c.), however, stated as his opinion
that he could not find any valid reason for referring the Upper
Siwalik fauna to the Pliocene. If, therefore, Matthew\'s assertion
should turn out to be correct, quot;Mastodonquot; [in the form of Igt;enta-
lophodon sivalensis (Cautl.) 2)] should occur in the Pleistocene of
Asia. In the writer\'s opinion quot;Mastodonquot; in all probability occurs
in the Javan Pleistocene. As already mentioned in the Introduction
in Sangiran a mastodont-Iike, fragmentary molar has been found
by Van Es in beds, which he regards contemporary with the Trinil
beds. Moreover, it was mentioned that a similar grinding tooth
had been obtained by Dubois from the Ttinil beds, but that it
might represent in the latter\'s opinion an atavistical deviation of
a Stegodon molar. In our opinion, however, this supposition is
highly improbable. In what follows it will be seen that some of
our determinations give a very strong support in favour of a
pleistocene age of the Trinil beds. Consequently quot;Mastodonquot; in
all probability occurs in the Pleistocene of Java

Turning now to Stegodon airawana, we may be brief. We shall
return to that form later on. For the present it will suffice to call
to mind that remains of St. airawana are abundant in the Trinil
beds, and that Van Es\'s researches have shown the Trinil beds
to be of younger age than the Bumiaju beds.

As to Archidiskodon planifrons, the following remarks may be
made. According to Pilgrim (1. c.) this species occurs in the Pinjor
horizon (middlemost Upper Siwaliks). In Matthew\'s opinion,
therefore, it occurs in the Pleistocene of the Siwalik Hills. It may
be added that it is said also to occur in the Narbada deposits, which
are generally considered as Lower Pleistocene. It must be borne
in mind, however, that our knowledge of the Narbada beds must
be called totally insufficient.

Remains to consider Hippopotam/s spec. According to Pilgrim
Hippopotamus sivalensis occurs in the Boulder Conglomerate Zone
(uppermost Upper Siwaliks). Matthew (1. c.) pointed out that the
British Museum specimens of the Upper Siwaliks show two diverse
types of fossilization, and that the specimens, which are well fossilized.

*) The name of this spccics is erroneously left out of the faunal list of the Upper Siwaliks,
given hij MAmiiiW (1. c. p. 443)-

*) As far as my knowledge goes Ptnialopljodon sivalensis is the only mastodont spccics
which has hitherto been found in the Upper Siwaliks. I cannot see, therefore, how
VoN Koenigswald (1. c.) came to the assertion that the Bumiaju form resembles
certain progressive forms of the Upper Siwaliks so closely, that it is either identical
with, or at least closely allied to one of them.

may belong to an older horizon, quot;perhaps Pliocenequot;. As among
the specimens of H. sivalensis both types of fossihzadon are re-
presented, there is, therefore, a possibility that this species occurs
both in the Pliocene and Pleistocene. Hippopotamus iravaticus is,
however, an undoubtedly tertiary species. According to Pilgrim i)
it occurs in the Dhok Pathan Zone (middlemost Middle Siwaliks).
Pilgrim places the Middle Siwahks in the upper Miocene \\
Matthew in the Pliocene. As already pointed out, however, the
Bumiaju form is certainly not idendcal with H. iravaticus.

In the Narbada beds Hippopotamus is represented by Hippopotamus
(^^Tetraprotodon\'\') palaeindicus, and Hippopotamus {quot;Hexaprotodonquot;)
namadicus. It will be remembered that we arrived at the conclusion
that the Bumiaju form, being hexaprotodont, is certainly disdnct
from H. palaeindicus, but that it shows affinities to H. namadicus.
It appears, therefore, that neither the Hippopotamus remains afford
conclusive evidence in favour of an upper pliocene age of the
Bumiaju beds. We will return to the problem of their age later on.

In this region are situated our localities Pitu, Watualang, and
Kedung Kendang. Furthermore the famous locality Trinil. Stratigra-
phy and tectonic structure are owing to Van Es well known.

It will be remembered that the mammalian remains of the rich
collection, brought together by the Trinil Expedition of Mrs.
Selenka, have been examined by Janensch {Proboscidea) and
Stremme (the rest). It was the latter who drew his conclusions
from the determinations, made by either. One of the first con-
clusions at which Stremme arrived, was: quot;So sehen wir hier eine
Fauna, die recht beträchtlich von der heutigen verschieden ist
und hine mit ihr gemeinsame Art enthält ; von den gut bestimm-
baren Arten ist nicht eine mit einer heutigen identischquot; It has

That is to say to the Pontian. By some investigators, among who Pilgrim, the
Pontian is referred to the Upper Miocene, by most however, to the Lower Plioccne.

been shown that this assertion is not correct. In our opinion there
are no valid reasons for distinguishing the forms, reckoned bv
Stremme resp. to Buffelus palaeokerahau Dub. and to Rhinoceros
sivasondaicus Dub., from the still living species Buffelus huhalus var
sondaicus, and Rhinoceros sondaicus.

Another conclusion of Stremme was that — though he could
not find conclusive evidence either in favour of an upper pliocene
age, or of a lower pleistocene age — it could not be denied after
comparison of the Trinil fauna with upper phocene European

pHocäne Alter der Kendeng-Fauna spricht.quot; i) In Iiis comparison
Stremme made use of a faunal list, given by Schlosser, and con-
taining all the genera of some classic faunas, a.o. those of Val d\'Arno.
The latter fauna, however, formerly regarded as Phocene, is con-
sidered at the present moment by some of the best modern Italian
and French authorities as early Pleistocene 2).

Be that as it may, we do not doubt that Stremme would have
come to a pleistocene age, had he — as we — arrived at the
conclusion that the Trinil fauna contains at least three still living
species viz., Bihos sondaicus fossilis, Buffelus huhalus var. sondaicus
fossilis, and Rhinoceros sondaicus fossilis.

Our knowledge of the. stratigraphy of the vertebrate bearing
layers is very insufficient. Regarding the mammalian remains,
there seem to be few reasons to accept another age than for the
Trinil beds. It may be emphasized that Dubois did not accept
a different age.

As to the first three determinations comment will be needless.
As already pointed out on p 172, the lower jaw, which we deter-

mined as possibly belonging to the recent E. maximus, might
perhaps belong to E. hjsudrindtcus Dub., a form, which is said
to be intermediate between E. indiens {maximus) and the Siwalik
E. hjsudricus (Boulder Conglomerate Zone, uppermost Upper
SiwaHks). Even if Pilgrim\'s opinion of the age of the Upper
Siwaliks is adopted, we may E. hjsudrindicus expect only to occur
in pleistocene deposits

Concerning Stegodon trigonocephalus we may mention that in
the opinion of some this form is specifically indistinct from
St. airdwana. Though we do not agree with that opinion, we
immediately admit that the grinding teeth of both Javan forms
show exactly the same height of development.

Since 1914 three different investigators have paid attention to
the degree of specialization of the grinding teeth of St. airdwana.
The first was Soergel 2). He pointed out that in the anterior portion
of the teeth of that species the top of the ridge-crests is divided
into three by two longitudinal, shallow clefts. As the same pheno-
menon frequently occurs in Elephas teeth, Soergel concluded:
quot;In der Dreipfeilerbildung an Proximalende von Stegodonten-
zähnen haben wir ein fortgeschrittenes, zum speziellen Zahnbau
der Gattung Elephas überleitendes Merkmal zu erblickenquot;
Soergel pointed out, furthermore, that in the Siwalik stegodonts
this feature seems to be absent; consequently they would be older.
This difference, together with the greater number of ridge-crests
and the thinner enamel, is one of the chief reasons, why Soergel
regards the Kendeng fauna to be of undoubtedly pleistocene age.
The question is, however, not so simple as would appear from
the above. From our descriptions and figures of the teeth of
St. airdwana, it will namely be seen that sometimes the ridge-crests
are divided into four, because of the presence of a median, an
outer, and an inner constriction. The median constriction indicates
the presence of a median, longitudinal cleft. As such a condition
is a characteristic of Mastodon teeth, its occurrence in Stegodon
teeth (e. g. those of St. clifti) is regarded as a primitive feature.
Taken together, wc may say that the grinding teeth of St. airdwana
show in most cases an quot;quot;Elephas characterquot;, but in some cases
beside it a quot;Mastodon characterquot;. Nevertheless it cannot be denied
that St. airdwana is a progressive species, more specialized than
St. ganesa-insignis, the species of the Upper Siwaliks, and which

The more so, if is added that Lydekker already suggested the possibility of a lowermost
pleistocene age of the topmost beds of the Siwaliks, and that Pilgrim (1. c. p. 325)
refrained from quot;in the least disputing Lydekker\'s suggestionquot;.
Palaeontographica, Suppl. IV, Abt. Ill, Lief. I, 1914.
=■) L. c. p. ij.

is said also to occur in the Narbada deposits. If therefore — Soergel
observes —■ the Narbada stegodont should turn out to be really
specifically identical with the form of the Upper Siwahks, and if
indeed should appear that the Narbada deposits are of lower pleisto-
cene age, the Kendeng fauna has to be assigned to the Middle
Pleistocene.

The next investigator, who tried to ascertain the age of the
Kendeng fauna by the help of the molars of St. airawana, was
Dietrich i). His train of thought was as follows. The general
progress of the phylogenetic evolution of the dentition of Ste-
godon is : reduction iii the region of premolars and (or) milkmolars,
addition, i. e. increase of length and of number of ridge-crests, in
the region of the molars. Therefore it would be possible to fix
in numbers the degree of specialization, if we but disposed of

From these values it appears that from St. clifti to St. airmvana
the number of ridge-crests an unit of length increases. But that
of course is not new. Dietrich, however, deduces more from them.
He pointed out that according to Pilgrim St. clifti and St. homhifrons
occur in the Dhok Pathan horizon (middlemost Middle Siwaliks)
as well as in the Tatrot horizon (lowermost Upper Siwaliks), while
St. ganesa-insignis occurs in the Boulder Conglomerate Zone (upper-
most Upper Siwaliks) and is said to occur in the Narbada deposits.
If, therefore, Pilgrim\'s determination of the age of the various
Siwalik horizons is adopted the difference between the values 33
and 27 (resp. 29 and 25) symbolises a lapse of time, comprising
the whole Pliocene. And as the values, wliich Dietrich calculated

Sitz. bcr. Gcs. Naturf. Fr. Berlin, 1924, p. 134.
») I cannot see why Dietrich did not prefer the formula

In our opinion it had been better not to use the quotient, but the product.
\') And if the Pontian is reckoned to the Lower Pliocene.

for St. ganesa-insignis, are situated almost in the middle between
the values found for St. airawana ( trigonocephalus) and St. clifti
( bomhifrons), he arrived at the conclusion: quot;Selbst wenn das
Entwickelungstempo rascher geworden ist, kommen wir für
Airawana zu einem sehr viel jüngeren Alter, nämlich zu Jung-
bis Jüngstpleistocänquot; i).

z. If Matthew\'s views of the age of the various Siwalik horizons
are adopted, the difference between Dietrich\'s values 3 3 and 27
(resp. 29 and 25) symbolizes a lapse of time, which reaches
at least to the base of the Middle Pleistocene. In that case
St. airawana would therefore be still younger than quot;Jung- bis
Jüngstpleistocänquot;.

3. The occurrence of a complex of layers of about 500 m. thicloiess
above deposits, contemporary to the Trinil beds argues
strongly against an upper- to uppermost pleistocene age.

The third author, who has occupied himself —■ be it indirectly —•
with the height of development of St. airmvana, is Osborn (in
cooperation with Edwin H. Colbert). Osborn maintains that the
grinding teeth of Elephas and Stegodon can be used as quot;priceless
enamel chronometersquot; Therefore it is only necessary to measure
very accurately the total enamel length. His method is called
ganometric (ganos = enamel). Concerning Stegodon grinders the
following — roughly estimated — values are mentioned:

For the present these values do not mean much. It was already
long known to us that the enamel of St. bomhifrons is less plicated
than of St. airawana. The difference of 100 mm. has no signification,
so long we do not dispose of the value found for the corresponding
molar of a species as St. ganesa-insignis. And even if we did know,
Osborn will have to prove the correctness of his assertion that
the total enamel length can be used as a chronometer.

\') Proc. Amer. PhiI6s. Soc., LXX, 1931, p. 187. Sec also Osborn\'s two papers in the

Rep. of the Cent. Meeting Br. Ass. London 1931.
*) Rep. of the Cent. Meeting Br. Ass. London 1931, 1932, p. 452.

Unfortunately enough in Osborn\'s provisional publications on
the present subject the technique of the ganometric method is
not dealt with. Has use been made of grinders with a certain
degree of wear, and has the total length of all the phcations
been measured at the grinding-surface ? Or has in one way
or another the height of the various ridge-crests been included?
Apparently it has. Colbert, namely, pointed out: quot;The measu-
rements thus far assembled are all rough and preliminary;
out of the thirty-six teeth studied, only two were unworn;
consequently there are large estimated factors in most of the
measurementsquot; i).

It remains with us to consider St. bondolensis nov. spec. On p. 15 8
it will be seen that this form has been obtained from Bondol near
Kuwung, a locality from which also remains of Buffelus bubalus ?var.
sondaicus fossilis and Rhinoceros sondaicus fossilis have been procured.
This is surprising, as St. bondolensis is a rather primitive species,
the height of development of wliich is equal to that of St. bombifrons.
It will be remembered that the youngest Siwalik horizon in which
the latter species occurs, is the Tatrot horizon (lowermost Upper
Siwaliks, Lower Pliocene according to Pilgrim, Lower Pleistocene
according to Matthew). If Matthew\'s views should appear to
be correct, the occurrence of St. bondolensis together with still living
species, would be comprehensible. On the other hand it must
not be forgotten that the stratigraphy of the locality mentioned
is very badly known. There remains, therefore, a possibility that
St. bondolensis has been obtained from older strata. Nevertheless
the problem remains that the grinders of all the Stegodon species,
hitherto found in Java, are characterized by enamel, clearly consisting
of two layers. Consequently quot;Stufenbildungquot; occurs. Why the latter
feature is distinctly exhibited only in the teeth of the Javan species
is an enigma to me.

To return now to our starting point, wc may say that
there are several indications, which point in the direction of
an older age of the Bumiaju fauna compared with the Trinil
fauna 2) viz.,

2.nbsp;In the Bumiaju beds a very primitive species of Elephas (A. plani-
frons) has been found; in the Trinil beds it is absent.

Proc. Amcr. Philos. Soc., LXX, 1931, p. 191.
\') With Trinil fauna I mean the fauna of the Trinil bcUs between Gesi and Ngawi.

3- The Bumiaju specimens are always well fossilized; in the Trinil
specimens the state of fossilization is in some cases far less
progressed.

Stress may be laid upon the fact that each argument alone has
but litde value. But all together they afford in our opinion strong
evidence in favour of an older age of the Bumiaju beds.

It will be noticed that we arrive at the same conclusion to
which Van Es came on geological-stratigraphical grounds i). The
question is: Do we agree with Van Es as to the lower pleistocene
age of the Trinil beds, and the upper pHocene age of the Bumiaju
beds ? As already pointed out in the Introduction, Van Es started
from a supposition which lacks sufficient grounds. Nevertheless
his assertion may be true. The occurrence of recent species among
the Trinil fauna proves in our opinion its pleistocene age. It may
therefore be Lower Pleistocene. And in that case the Bumiaju
beds have to be placed into the Upper Pliocene. As far as our
knowledge goes for the present, the mammalian contents of the
latter do not prove it, but neither do they afford sufficient evidence
against such a supposition.

We cannot conclude without drawing attention to the following
possibility. In a quite recent paper Martin 2) has pointed out that
the descriptions of tertiary molluscs from Burma and N. W. India
by Vredenburg and De Cotter have shown that in these regions
tertiary strata occur, which contain both Javan and European
species. The discovery is of course of great interest for the cor-
relation of tertiary beds in Europe and Asia. A. o. it enabled
Martin to emphasize the pliocenc age of the mannc Sonde beds.
He did not let on, however, about the question to which part of
the Pliocene the Sonde beds belong. Apparendy there is a possibility
that the Sonde beds represent (or include) the Upper Pliocenc.
In that case the Tnnil beds — being separated by a stratigraphic
gap from the underlying Sonde beds —• would belong to the
Middle Plcistoccne, and the Bumiaju beds — which according to
Van Es are contemporary to the hiatus in Tnnil — should be of
lower pleistocene age. In my opinion neither the mammalian
fauna of the Trinil beds, nor of the Bumiaju layers give sufficient
arguments against this supposition.

2.nbsp;The Trinil fauna is certainly of pleistocene age, more particularly
either Lower of Middle Pleistocene, but not Upper Pleistocene.

3.nbsp;Accordingly the Bumiaju fauna is of upper pliocene or lower
pleistocene age.

1930.nbsp;Anonymus. Java-kaarteering. Jaarb. Mijnw. in Ned.-Ind., Jrg. 1929, Algem. gcd.

1836. Baker, W. E., and H. M. Durand. Sub-Himâlayan fossil remains of the Dâdupur
collection. Journ. As. Soc. Bengal, V, 1836, p. 486—504; plates.

1908. Branca, W. Vorläufiger Bericht über die Ergebnisse der Trinil-Expedition der akade-
mischen Jubiläums-Stiftung der Stadt Berlin. Sitz. ber. d. kön. preuss. Akad,
d. Wiss., Jhrg. 1908, Halbbd. I, 1908, p. 261—273; sketch map, section.

1883. Braiws, D. Ueber japanische diluviale Säugethiere. Zeitschr. deutsch, geol. Ges
XXXV, 1883, p. 1—50; plate.

1869. Busk, G. Notice of the discovery at Sarawak in Borneo of the fossilized teeth of Rhino-
ceros and of a cervine ruminant. Proc. 200I. Soc. London, 1869, p. 409—416;
text figs.

1829. Clift, W. On the fossil remains of two new species of Mastodon, and of other verte-
brated animals, found on the left bank of the Irawadi. Trans, geol. Soc, London,
ser. 2, II, 1829, p. 369—375; plates.

1924- Cooper, C. F. On the skull and dentition of the Paraceratherium bugtiensc. Philos.
Trans, roy. Soc. London, ser. B, CCXII, 1924, p. 369—394; text figs.

1931.nbsp;CosrjN, J. Voorloopige mcdedecling omtrent het voorkomen van fossicle beenderen

in het heuvelterrcin ten Noorden van Djetis en Perning. Verh. gcol.-mijnbouwk.
Gen. V. Ned. en Kol., geol. scr. IX, 1931, p. 113—119; sections, sketch map.

1932.nbsp;- Tweede mededecling over het voorkomen van fossiele beenderen in het heuvel-

land ten Noorden van Djetis en Perning (Java). Ibid., IX, 1932, p. 135—148;
sections, sketch maps, plate.

1822. CirviER, G. Recherches sur les ossements fossiles. Nouv. ddit., t. 2, pt. i, Paris 1822.

1928. Dammerman, K. W. On the mammals of Sumba. Treubia, X, 2—3, 1928. Sec p. 312.

1868. Dawkins, W. B. On the dentition of Rhinocoros etruscus. Quart. Journ. geol. Soc.
London, XXIV, 1868, p. 207—218; plates.

1910. Deninger, K. Ueber einen Affcnkicfcr aus den Kcndcngschichtcn auf Java. Centr.
bl. f. Min. etc., Jhrg. 1910,1910, p. 1—3 ; text fig.

1825. Desmoulins, A. Détermination de deux espèces vivantes d\'hippopotame. Joum. de
Physiol, cxp. et Pathol., V, 1825, p. 354—366.

1863. Diest, P. H. van. Over fossiele zoogdierrcsten, gevonden in de mijn Banhin, n°. 8,
Socngei Liât, Bangka. Nat. Tijdschr. v. Ncd.-Ind., XXV, 1863, p. 470—471.

»923- Dietrich, \\V. O. Discussion a.o. of Schlcsinger\'s paper in Geologica Hungarica, 1922.
Neues Jahrb. f. Min. etc., Jhrg. 1923, II, 1925, p. 273—284.

1926.nbsp;- Zur Altersbestimmung der Pithecanthropus-Schichten. Sitz.ber. d. Ges. naturf.

1927.nbsp;-- Discussion of some of Osborn\'s papers viz., 1923, Am. Mus. Nov. nquot;. 99;

1924, Ibid., n°. 152; 1924, Ibid., n°. 154; 1925. Proc. Am. philos. Soc., XLIV.
Neues Jahrb. f. Min. etc., Ref., Abt. B, Jhrg. 1927, I, 1927, p. 311—313.

1928.nbsp;-- Discussion of Matsumoto\'s paper quot;On Leith-Adamsia siwalikiensisquot; in Jap.

Journ. Geol. and Geogr., V, 1928. p. 213. Neues Jahrb. f. Min. etc., Ref.,
Jhrg. 1928, III, 1928, p. 338.

1929.nbsp;-- Discussion of Matsumoto\'s papers in Sci. Rcp. Tohoku imp. Univ. Sendai,

Japan, XIIJ, 1929. Neues Jahrb. f. Min. etc., Ref., Jhrg. 1929, III, 1929, p. 466—467.

\'909. Dozy, C. M. Opgravingen bij Trinil in 1908. Tijdschr. kon. Ned. aardr. Gen.,
ser. 2, XXVI, 1909, p. 604—611; sketch map, section.

1889. Dubois, E. Over de wenschelijkheid van een onderzoek naar de diluviale fauna
van Ned.-Indië, in het bijzonder van Sumatra. Nat. Tijdschr. v. Ned.-Ind.
XL Vin, 1889, p. 148—165.

1890—\'95. - Palaeontologische onderzoekingen op Java. Versl. v. h. Mijnw., zde kwartaal

1892. - Voorloopig bericht omtrent het onderzoek naar de pleistocene en tertiaire

vertebraten-fauna van Sumatra en Java gedurende het jaar 1890. Nat. Tijdschr.
V. Ned.-Ind., LI, 1892, p. 93—100.

\'1907. - Eenige van Nederlandschen kant verkregen uitkomsten met betrekking tot

de kennis der Kendeng-fauna (Fauna van Trinil). Tijdschr. kon. Ned. aardr.
Gen., ser. 2, XXIV, 1907, p. 449—458. .

1926. -- Manis\' palaejavanica, het reuzenschubdier der Kendeng-fauna. Versl. kon.

Akad. v. Wetensch. Amsterdam, afd. Natuurk., XXXV, 1926, p. 947—958;
plate. Also in English: Manis palaejavanica, the giant pangolin of the Kendeng
fauna. Proc., XXIX, 1926, p. 1233—1243; plate, text figs.

1846. Duvernoy, M. Communication on the cranium of recent Hippopotamidae. L\'ln-
; .nbsp;stitut, sect, i, XIV, 1846, p. 333—334. (in French)

1908. Elbert, J. De nieuwste onderzoekingen over het Pithecanthropus-vraagstuk. Nat.
Tijdschr. v. Ned.-Ind., LXVII, 1908, p. 125—142.

1908.nbsp;- Ueber das Alter der Kcndeng-Schichten mit Pithecanthropus erectus Dubois.

Neues Jahrb. f. Min. etc., Beilage-Bd. XXV, 1908, p. 648—662; sketch map,
correlation table.

1909.nbsp;- Dubois\' Altersbestimmung der Kendengschichtcn. Centr.bl. f. Min. etc., Jhrg.

1911.--Die Selenka\'sche Trinil-Expcdition und ihr Werk. Centr.bl. f. Min., Jhrg.

1929. Es, L. J. C. van. Trinil. Guide of excursion E 5 of the Fourth Pacific Science Congress,
Java 1929. 14 pp.; map.

1931. - The age of Pithecanthropus. 142 pp.; correlation table, plates, maps, sections.

\'1849. Falconer, H. Communication on the Hippopotamidae, fossil and recent. In;

S. G. Morton. Additional observations on a new living spccics of Hippopo-
tamus. Journ. Acad. nat. Sei. Philad., ser. 2, I, pt. 3, 1849. See p. 235—237.

J 868. - Palaeontological memoirs and notes. (Compiled and edited by Ch. Murchison).

II.nbsp;Mastodon, elephant, rhinoceros, ossiferous caves, primeval man and his
cotempories. 675 pp.; plates. London 1868.

1845.nbsp;- and P. T. Cautley. Fauna Antiqua Sivalensis, being the fossil zoology

of the Sewalik Hills. London 1845; Plates only. (See for the explanation of the
plates: H. Falconer, Pal. Mem. II, 1868.)

1846.nbsp;- and - Fauna Antiqua Sivalensis being the fossil zoology of the Sewalik

Hills in the North of India. Letter-press, pt. I, Proboscidea, p. 1—64,
London 1846.

1876. Flower, \\V. H. On some cranial and dental characters of the existing species of rhino-
ceroses. Proc. zool. Soc. London, 1876, p. 443—457; text figs.

1874- Foote, R. B. Rhinoceros deccanensis. Mem. gcol. Surv. India, ser. X, I, pt. 1, 1874,
p. I—17; plates.

1904. Frech, F. Lethaea geognostica, Th. Ill, Bd. 2, Abth. i, 1904. See p. 30—32.

1932- Gerth, H. Discussion (in Dutch) of Van Es\'s paper: The age of Pithecan-
thropus, 1931. Tijdschr. kon. Ned. aardr. Gen., reeks 2, XLIX, 1932, p.

1929. Haar, C. ter. Boemi-Ajoe District. Guide of excursion E4 of the Fourth Pacific
science Congress, Java 1929. 15 pp.; sketch map.

1857. Hageman, J. Overlevering omtrent de reuzen van den berg Patiajam. (Vertaald
naar een maleisch verhaal geschreven te Pati in 1850). Aanteekening van een
verhaal door den nu overleden oud regent van Pati, Radhen Ario Pameeet
Sari, in 1814 gedaan. Tijdschr. v. Ind. taal-, land- en volkenkunde VI Ynieuwt;
ser. III), 1857, p. 275—278.

1921. Hilder, V. Alter der Pithecanthropus-Schichten. Centr.bl. f. Min. etc. Jhrg iqzi

1930- Hoen, C. W. A. P. \'t. Toelichting bij blad XVI (Midden Java) geologische over-
zichtskaart van den Nederlandsch-Indischen Archipel, Schaal i ; i.000.000
Jaarb. Mijnw., Verb., Jrg. 1929 atlas, 1930. See p. 28—35.

1921. Hoen, H. \'t. Buffel en rund. Onze koloniale dierentcelt I. 86 pp.; illustrations.
Haarlem 1921.

1857. Hogendorp, D. C. A. van. Over zoogdierresten van den Patiajam (Djapara) met
brief van C. S. A. Thurkow en aanbieding van die beenderen. Nat. Tijdschr.
v. Ncd.-Ind., XIII, 1857, p. 237—238 and 252—253.

1911. Janensch, W. Die Proboscidicr-Schädcl der Txinil-Expeditions-Sammlung. In: Die
Pithecanthropus-Schichten auf Java, 1911, p. 151—195; plates, text figs.

1917. Janensch, W. and W. Dietrich. Nachweis des ersten Prämolarcn an einen jugend-
lichen Oberkiefergebisz von Stegodon Airawana Mart. Sitz.ber. d. Ges. naturf.
Fr. Berlin, Jhrg. 1916, 1917, p. 126—136; plate.

1920. Joleaud, L. Contribution à l\'étude des hippopotames fossiles. Bull. Soc. géol.
de France, sér. 4, XX, 1920, p. 13—26; plate.

1857. Junghuhn, F. Over de fossiele zoogdierbeendcrcn te Patihajam, in de residentie
Djapara, eiland Java. Nat. Tijdschr. v. Ned.-Ind., XIV, 1857, p. 215—219.

1859—\'60. - Over fossielc(?) zoogdicrtanden van Bonthain (Sent by J. L. de Jeegcr,

see Nat. Tijdschr. v. Ned.-Ind., XX, 1859—\'60, p. 264). Nat. Tijdschr. v. Ned.-
Ind., XX, 1859—p- 385—386.

1832- Kaup, J. J. Description d\'ossements fossiles de mammifères inconnus jusqu\'à présent.
Cahier i. Darmstadt 1832.

1902. Keller, C. Die Abstammung der ältesten Haustiere. See p. 144—148. Zürich 1902.

193Koenigswald, R. von. Fossielen uit Chincesche apotheken in West-Java. De Mijn-
ingcnicur, n®. 11, 1931 p. 189—193; figs.

1931. - Sin-mthropus, Pithecanthropus en de ouderdom van dc Trinillagcn.. ibid.,

1885. Koken, E. Uebcr fossile Säugethiere aus China. Palaeont. Abh, (Dames u. Kayser)»
III, 2, 1885, 85 pp.; plates, text figs.

1856—\'57. Krajenbrink, J. A. Over rcuzcnbeendercn en schedels van den berg Patiajam,
res. Djapara. Nat. Tijdschr. v. Ncd.-Ind., XII, 1856—\'57, p. 489.

1868. Leith Adams, M. B. Has the Asiatic clcph.int been found in a fossil state? With some
additional remarks by G. Busk, Quart. Journ, gcol. Soc. London, XXIV,

1931- Leupold, W. and I. M. van der Vlerk. The Tertiary, In: Feestbundel (= Jubilee

book) K. Martin. Leid. gcol. Mcdcd., V, 1931. See p. 635—639.
187ö. Lydekker, R. Description of a cranium of Stegodon gancsa, with notes on the
sub-genus and allied forms, Rcc, gcol. Surv. India, IX, pt. 2, 1876, p, 42—49.

1876,nbsp;- Molar teeth and other remains of Mammalia. (With appendix). Mem. geol.

1877,nbsp;---Notices of new and other Vertcbrata from Indian tertiary and secondary rocks.

1878,nbsp;--- Crania of ruminants from the indi.-in Tcrtiaries. Mem. geol. Surv. India, scr.

1878.nbsp;--- Notices of Siwalik mammals. Rcc. gcol. Surv. India, XI, pt. i, 1878, p. 64—104.

1879.nbsp;- Further notices of Siwalik Mammalia, Ibid., XII, pt. i, 1879, p. 33—52;

i88o. Lydekker, R, Siwalik and Narbada Proboscidea. (With appendix). Ibid., I, pt. j,
1880, p. I (181)—113 (294); plates.

1880.nbsp;- Supplement to crania of ruminants. Ibid., I, pt. 4, i88o, p. i (171)—10 (181);

1881. - Supplement to Siwalik and Narbada Proboscidea. Ibid., II, pt. 2, 1881, p. i (63)

1884, - Additional Siwalik Perissodactyla amp; Proboscidea. Ibid., III, pt. i, 1884, p.

1884. - Siwalik and Narbada bunodont Suina. Ibid., III, pt. 2,1884, p. 55 (i)—104 (70);

1884.nbsp;- Mastodon teeth from Perim Island. Ibid., III, pt. 5, 1884, p. 149 (i)—154 (6);

1885.nbsp;- Description of a tooth of Mastodon latidens Clift, from Borneo. Proc. zool*

i88j. - Cat. foss. Mamm. in the British Museum, London, pt. 11, 1885, 324 pp.; text

1886.nbsp;- Ibid., pt. III, 1886, 186 pp.; text figs. (A.o. containing the Rhinocerotidae;

1886. - Cat. pleistocene and pre-historic Vcrtebrata of the Indian Museum, Calcutta.

1886. - Introductory observations to Mem. geol. Surv. India, ser. X, III, 1884—\'86,

1886.nbsp;- The fauna of the Karnul caves. Ibid., IV, pt. 2, 1886, p. 23—58; plates, text figs.

1887.nbsp;- The fossil Vertebrata of India. Rec. geol. Surv. India, XX, pt. 2,1887, p. 51—79.

1898. - On the geographical races of the banting. Proc. zool. Soc. London, 1898,

1898. - Wild oxen, sheep and goats of all lands; living and extinct. 318 pp.; plates, text

1909. - Exhibition of photographs of a spotted bull Tsaine or Bantin from Siam. Proc.

1912. - The Bornean bantin. Proc. zool. Soc. London., 1912, pt. IV, p. 902—906;

1931. Maarel, F. H. van der. quot;Mammaliaquot; in Fecstbundcl (= Jubilee book) K. Martin.
Leid. geol. Mcded., V, 1931, p. 471—484; sketch map.

1924. Makiyama, J. The occurrence of Elephas trogontherii in Japan. Jap. Journ. Geol.
and Geogr., Ill, 1924, p. 55—57; plate.

1883.nbsp;Martin, K. Palaeontologische Ergebnisse von Ticfbohrungcn auf Java. Samml.

geol. Reichs-Mus. Leiden, III, 1883—\'87; plates. Under the same title reprinted
in: Jaarb. Mijnw. in Ncd.-Ind., 1883, Wet. Gcd., p. 371—412; plates.

1884.nbsp;- Ucberreste vorweltlicher Proboscidier von Java und Bangka. Samml. geol.

Reichs-Mus. Leiden, IV, 1884—89, p. 1—24; plate. Under the same title
reprinted in: Jaarb. Mijnw. in Ncd.-Ind., 1884, p. 285—308; plate.

1887.nbsp;- Fossile Säugethierrestc von Java und Japan. Samml. geol. Reichs-Mus. Leiden,

IV, 1884—\'89, p. 25—69; plates. Under the same title reprinted in: Jaarb.
Mijnw, in Ned.-Ind,, 1887, Wet. Gcd,, p. 1—45; plates,

1888.nbsp;- Neue Wirbelthierreste vom Pati-Ajam auf Java. Samml. geol. Reichs-Mus.

Leiden, IV, 1884—\'89, p. 87—116; plates. Under the same title reprinted in:
Jaarb. Mijnw. in Ncd,-Ind,, 1888, Wet, Gcd,,p, 20—48; plates,

1890,nbsp;Ucbcr neue Stegodon-Reste aus Java. Verh. kon. Akad. v. Wetensch, Amsterdam,

1899—1902, - Die Eintheilung der versteinerungsführende Sedimente von Java. Samml.
geol, Reichs-Mus. Leiden, VI, 1899—1902, p. 135—245.

1908. Martin, K. Das Alter der Schichten von Sonde und Trinil auf Java Versl Won
Akad. V. Wetensch., afd. Wis- en Natuurk., XVII, 1908, p. 7—16. \'

1919. - Unsere palaeozoologische Kenntnis von Java. ij8 pp.; plates. Leiden 1919.

I93I- - Wann löste sich das Gebiet des Indischen Archipels von der Tethys? Leid

1918. Matsumoto, H. On some fossil mammals from Sze-chuan, China. Sei. Rep. Tóhoku
imp. Univ., Sendai, Japan, ser. 2 (Geol.),III,(1915—18), 1918, p. 1—28; plates
text figs.nbsp;\'

1918. -- On some fossil mammals from Ho-nan, China. Ibid., III, p. 29—38; plates.

1918. - One some fossil mammals from Tsukinoki, Ugo. Ibid., III, p. 39—49; plates.

1918. - On a new archetypal fossil elephant from Mt. Tomuro, Kaga. Ibid., Ill, p. 51—

1921. - Descriptions of some new fossil mammals from Kani District, prov. of Mino,

with revisions of some Asiatic fossil rhinocerotids. Ibid., V, (1918—1921), 1921!
p. 85—92; plates.

1926. -- On two new mastodonts and an archetypal stegodont of Japan. Ibid., X, n°. i,

1926.nbsp;- On the archetypal mammoths from the province of Kazusa. Ibid., X, n°. 2, 1926,

1926—\'27. - On Leith—Adamsia siwalikiensis, a new generic and specific name of

1927.nbsp;- On a new fossil race of the Asiatic elephant in Japan. Sei. Rep. Tohoku imp.

1929.nbsp;-- On Parastegodon Matsumoto and its bearing on the desccnt of earlier elephants.

1930.nbsp;MAmiEW, W. D. Critical observations upon Siwalik mammals. Ibid., LVI, (1926—

1923. - and W. Granger. New fossil mammals from the Pliocene of Sze-Chuan, China.

1927. Merkens, J. Bijdrage tot de kennis van den karbouw en de karbouwenteelt in Neder-
landsch Oost-lndie. 191 pp.; plates. Utrecht 1927. Thesis Utrecht 1927.

1910. Miller Jr., G. S. Description of a new species of Hippopotamus. Smiths, misc. Coll.,

LIV, 1910, n°. 7, p. I—3; plates.
1882. Naumann, E. Ueber japanische Elephanten der Vorzeit. Palaeontographica, XXVIII,
1882, p. I—39; plates.

1890. - Stegodon mindancnsis, cine neue Art von Uebergahgs-Mastodonten. Zeitschr.

1898. Osborn, H. F. ITie cxtinct rhinoceroses. Mem. Amer. Mus. nat. Hist. I, pt. 3, 1898,
p. 65—164; plates, text figs.

1900. - Phylogeny of the rhinoceroses of Europe. Bull. Amcr. Mus. nat. Hist., XIII,

1907. - Evolution of mammalian molar teeth. Biological studies and addresses, I, 250 pp.;

1917. Osborn, H. F. A long-jawed Mastodon skeleton from South Dakota and phylogeny
of the Proboscidea. Bull. geol. Soc. Am., XXIX, 1917. p. 133—137.

^921- - Adaptive radiation and classification of the Proboscidea. Proc nat Acad Sci

1921. - The evolution, phylogeny, and classification of the Proboscidea. Am. Mus.

1923.nbsp;— New subfamily, generic, and specific stages in the evolution of the Proboscidea.

1924.nbsp;— Serridentinus and Baluchitherium, Loh Formation, Mongolia. Ibid., n°. 148

I924\' .——, Parelephas in relation to phyla and genera of the family Elephantidae. Ibid.
n°. IJ2, 1924, 7 pp.; text figs.

1924.nbsp;—— Additional generic and specific stages in the evolution of the Proboscidea. Ibid.,

1925.nbsp;- Final conclusions on the evolution, phylogeny, and classification of the Probos-

1926.nbsp;- Additional new genera and species of the mastodontoid Proboscidea. Amer.

1929.nbsp;- New Eurasiatic and American proboscideans. Ibid., n°. 393, 1929, 22 pp.; text

1930.nbsp;—— Fifty-two years of research, observation and publication. 1877—1020. 160 dd •

.... stone age man by the elephant-enamel-method. Rep. centen. Meet. Brit. Ass.
Advanc. of Sci. London 193.1, p. 372—373.

1931.nbsp;The geologic age of Pithecanthropus, Eoanthropus, and other fossil men deter-
mined by the cnamel-ridge-plate-grinding-tooth-measuremcnt of the Probos-
cidea with which they were geologically contemporaneous. Ibid., p. 451—453

\'931- - Palaeoloxodon antiquus italicus sp. nov., final stage in the quot;Elephas antiquusquot;

1931, - and E. H. Colbert. The elephant enamel method of measuring pleistocene time.

Also stages in the succession of fossil man and stone age industries. Proc. Amer.
philos. Soc. Philad., LXX, 1931, p. 187—191.

1870. Owen, F. R. 8. On the] fossil remains of mammals found in China. Quart. Journ.
gcol. Soc. London, XXVI, 1870, p. 417—434; plates.

1908. Pilgrim, G. E. The tertiary and post-tertiary fresh-water deposits of Baluchistan
and Sind, with notices of new vertebrates. Rcc. gcol. Surv. India XXXVII
pt. 2, 1908, p. 139—166; plates.

1910. - Notices on new mammalian genera and spccies from the Tcrtiaries of India.

1910. - Preliminary note on a revised classification of the tertiary freshwater deposits

^912- - The vertebrate fauna of the Gaj series in the Bugti Hills and the Punjab. Mem.

1913. - The correlation of the Siwaliks with mammal horizons of Europe. Rcc. geol

,1917. -, . Preliminary note on some recent mammal collections from the basal beds of
the Siwaliks. Ibid., XLVIII, pt. 2, 1917, p. 98—101.

1911.. PonLiG II. Zur Ostcologie von Stegodon. In: Die Pithccanthropus-Schichtcn auf
Java, 1911, p.\' 156—213; plates, text figs.

1839- Rigg, J Sketch of the geology of Jasinga. Verh. Bataviaasch Gen. van Kunst. en
Wetensch., XVII, 1839. See p. 129.

1927. - Ueber quartäre und jungtertiäre Rhinocerotiden aus China/und der Mongolei

^877—\'78. - Die Rinder der Tertiär-Epoche nebst Vorstudien zu einer natürlichen

Theil i: Abh. Schweiz, pal.\'Ges., IV, 1877, p. 1—72; plates.
Theil 2: Ibid., V, 1878, p. 73—208; plates, text figs.

1878. - quot;Nachschriftquot; to his paper: Ueber Prof. M. Wilkens\' Brachycephalusrasse des

1927. Rutten, L. M. R. Voordrachten over de geologie van Nederlandsch Oost-Indië.
839 pp; illustrations. Groningen, Den Haag 1927.

1867. Saleh, Rhaden. Over ......... fossiele beenderen van den Pandan. Nat. Tijdschr

1867. —,— Over fossiele zoogdierbeenderen in het regentschap Scntolo, Jogjakarta, ibid.,
XXIX p. 423, 428 and 434—435.

1874. Sciieltema Beduin, p. Aanbieding van fossiele beenderen, uitgegraven in de vallei
Ajer Lajang op Banka. Ibid., XXXIV, 1874, p. 336.

1861. Schlegel, H. Bijdrage tot de geschiedenis van de olifanten, voornamelijk van
Elephas sumatranus. Verh. en Mcded. kon. Akad. v. Wetensch., afd. Natuurk.,
Xll, 1861, p. ici—112.

1839—\'44. - and S. Müller. Over de ossen. Verhandelingen over de natuurlijke

geschiedenis der Ncderlandsche Overzeesche bezittingen. Zoologie, 1839—\'44,
p. 195—208; plates.

1922. Schlesinger, G. Die Mastodonten der Budapester Sammlungen. Geologica Hun-
garica, 11, i, 1922, 284 pp,; plates, text figs,

1903. Schlosser, M. Die fossilen Säugethiere Chinas nebst einer Odontographic der
recente Antilopen, Abh, kön, bayr, Akad. Wiss., math,-physik. Kl,, XXII,
Abt, i, 1903, p, i—221; plates, text figs.

1924, - Tertiary vertebrates from Mongolia. Palaeontologia Sinica, scr. C, I, fasc. i,

1864, Schmülling, P, E, C, Fossicle lieenderen van zoogdieren bij Sanggiran (Java) ge-
vonden. Nat. Tijdschr, v. Ned,-Ind,, XXVII, 1864, p, 399—400,

\'1868, schneither, Aanbieding van een fossicle rib, afkomstig van Borneo\'s Zuidkust,
Ibid,, XXX, 1868, p. 394.

1910- Schuster, J. Ein Beitrag zur Pithccanthropus-Fragc, Sit2,ber, d. kon. bayr. Akad.
d. Wiss,, math-physik, Kl,, Jhrg, 1909, 1910, p, 1—30; plate, section,

1910. - De l\'âge géologique du Pithécanthrope ct de la période pluviale i\\ Java, Compt.

1893- Scott, W. B. The evolution of the premolar teeth in the mammals, Proc, Acad,
mt, Sci. Philad. (1892), 1893, p, 40J—444.

1909\' Selenka, M, Die fossiele Zähne von Trinil. (And answer by Eug, Dubois). Tijdschr.
kon. Ned. aardr. Gen., ser. 2, XXVI, 1909, p. 398—401.

19\'Selunka, M. L. .and M. Blanckenhorn. Die Pithecanthropus-Schichten auf Java,
Leipzig 19H. (This work consists of various papers. Those which were of
special interest to us will be found mentioned under the names of Stremme,
Janensch and Pohlig).

1858—\'59. Six)et van Oldruytenborgh, A. Fossicle beenderen, gevonden op de helling
van den Goenoeng Pandan bij Ketoegeocn, regentschap Ngawi, residentie
Madioen. Nat. Tijdschr. v. Ned.-Ind., XVI, i8j8—1859, p. 70—71.

heimensis nov. form. Mit Ausführingen über die vShStSe Ton .Ifl S ,
von Rhinoceros (Ceratorhinus) sumftrensis. Abh. SÏÏS g3. feic^^^^^^^
XIX. Heft X. X902. p. t-92; plates, text figs, table of meas^rlentf \'

■ 9a7--a8. W«e«, M. Die Slugeticrc, I cn II, ate Aufl. Jena ,9a7-\'a8
.899. WOODWARO. H. Noles on Elephas (Stegodon) ganesa Falc ct Cau,

Fig. 2. Right profile view of ditto, after removal of the right (detached)
horn-core. 0.30 nat.size. p. 25.

Fig. I. Palatal view of quot;one-hornedquot; dquot; cranium. 0.14 nat. size. p. 42.
Fig. 2. Front view of ditto. 0.16 nat. size. p. 42.

Fig. 3. Right P2—M3 of ditto, viewed from the gnnding surface. 046
nat. size. p. 42.

Fig. I.nbsp;Front view of quot;two-hornedquot; dquot; cranium. 0.15 nat. size. p. 42.

Fig. 4.nbsp;Hind view of left detached horn-core of old dquot;. 0.24 nat. size. p. 42.

■j.zi nat. size. p. 58. (Occipital
view in fig. i, pl. V).

Fig- I. Front view of cranium no. a. * ]
Fjg- Left profile view of ditto.
Fjg- 3- Palatal view of ditto.

Pig- I- Occipital view of cranium no. a. * 0.23 nat. size. p. 58. (The
same specimen as of figs. 1—3, pl. IV).

3-nbsp;Right pi—M^ of ditto, viewed from the grinding surface. 0.60
nat. size. p. 59.

4-nbsp;Reflected image of the inner view of the same tooth row. 0.52
nat. size. p. 59.

5-nbsp;Outer view of dght detached lower canine (specimen a). 0.50 nat.
size. p. 88.

I. Upper view of right horizontal mandibular ramus. 0.51 nat. size.
P- 82. (Right profile view in fig. i, pl. XIX).

P^g- 3- Palatal view of fragmentary right upper jaw with M^—M^.
nat. size. p. 87.

Pig- 2. Upper view of ditto. 0.31 nat. size. p. 82. (Anterior view in fig. 2,
pl. XIX).nbsp;^nbsp;^

1.nbsp;Palatal view of fragment of cranium with 1. and r. M^ in situ.
^ 0.22 nat. size. p. ii6. (Left profile view in fig. 2, pl. X).

2.nbsp;Inner view of fragmentary right mandibular ramus with M
0.29 nat. size. p. 108.

1.nbsp;Fragmentary r. M^, viewed from the grinding surface, o 8t
size. p. 113.nbsp;quot; \' ■

3- Right Mg (of ramus portrayed in fig. 2 of pl. VIII), viewed from
the grinding surface. 0.53 nat. size. p. 108.

Fig. I. Damaged mature cranium with r. M^ in situ, viewed from th
upper and partly from the outer side. 0.14 nat. size. p. 151.

1.nbsp;Left Mg and posterior extremity of Mg (of lower jaw portrayed
in figs. 2—3, pl. XIV), viewed from the grinding surface. 0.55
nat. size. p. 141.

2.nbsp;Right M^ (of cranium portrayed in figs, i—2, pl, XI), viewed
from the grinding surface. 0.63 nat. size. p. 151.

Fig- I. Left M3, viewed from the grinding surface. 0.64 nat. size
^ig- 2. Inner view of ditto. 0.51 nat. size. p. 136.

Pigs. 3 and 5. Fragments of (probably) upper grinding teeth vertir^,!!
and longitudinally bisected. 0.56 nat. size. p. 164. \'

Fig. i. Left Mg (of lower jaw portrayed in figs. 4—5, pl. XIV), viewed
from the grinding surface. 0.63 nat. size. p. 159.

0.15 nat. size. p. 158.
Fig. 5. Right profile view of ditto. 0.15 nat. size. p. 164.

o.io nat. size. p. 141.
Fig. 3. Left profile view of ditto, o.io nat. size. p. 141.

size. p. 132.
Fig z. Inner view of ditto. 0.52 nat. size. p.
Fi, , Outer view of right detached M. 0.68 nat. size. p. . 3 B-
Fig. 4. The same speamen. viewed from the grinding surface. 0.70 „at.
size. p. 138.

Fig I. Left M3 (of lower jaw portrayed in figs. 2—3, pl- XVI), viewed
from the grinding surface. 0.48 nat. size. p. 169.

Fig. 2. Upper view of fragmentary lower jaw with 1. and r. M3 in situ.
0.15 nat. size. p. 168.

Fig. 4. Proximal surface of articulation of ulna no. i * and radius no. 2
0.14 nat. size. p. 188.

Fig. 5. Distal surface of articulation of the same specimens. 0.13 nat.
size. p. 188.

Fig. i. Detached left U\\ viewed from the grinding surface. 0.77 nat.
size. p. 173.

Fig. 5. Right femur (no. 4¹), inner view. 0.12 nat. size. p. 187.
Fig. 6. Front view of ditto. 0.12 nat. size. p. 187-
Fig. 7. Hind view of ditto, o.ii nat. size. p. 187.

Fig. i. Right profile view of posterior portion of cranium. 0.48 nat.
size. p. 89.

Fie 2 Upper view of ditto. 0.45 «at. size. p. 89. (Hind view in fig. 3,
pl. XIX).

Fie 3 Detached lower incisor, probably belonging to the mandible,
portrayed in figs, i—2,,pl. VII. 0.69 nat. size. p. 85.

Fig I Right profile view of fragmentary lower mandibular ramus.
0.41 nat. size. p. 82. bame specimen as of figs. 1—2, pl. VI.

Fig 2 Anterior view of right half of symphysial extremity of the spe-
cimen, portrayed in hgs. 1—2, pl. VH. 0.50 nat. size. p. 84.

PLATE XX.
Stegodon airawana Martin.
Fig. i. Front view of damaged immature cranium. 0.13 nat. size. p. 146.

Hippopotamus spec.
Fig. 2. Upper view of fragmentary lower jaw. 0.42 nat. size. p. 86.

Hippopotamus constrictus Miller 1910 is stellig identiek met
Hippopotamus amphibius L. Afgezien van Choeropsis liberiensis Mort.
zijn er dus niet drie recente nijlpaardensoorten bekend — zooals
algemeen wordt aangenomen — maar hoogstens twee.

De terminologie der componenten van molaren en premolaren,
opgesteld door Osborn en Scott op grond van de trituberculair
theorie, bevat onvolledigheden en inconsequenties.

De karbouwen, welke thans in het wild in het uiterste ZW. en
ZO. van Java worden aangetroffen, behoeven niet uitsluitend af
te stammen van verwilderde exemplaren.

Bij dc beoordeeling van de vraag of een fossiele vorm al of niet
identiek is met een recente, rust op elk onderzoeker de verplichting
zich moeite te geven zijn fossiele exemplaren te vergelijken met een
zoo groot mogelijk aantal recente van ongeveer gelijke ouderdom.

Aan de hand van Stegoccphalen en Amphibiën ontbreekt niet
— zooals Abel meent — dc eerste vinger, maar dc vijfde.

Jaekel is er niet in geslaagd een aannemelijke verklaring te
geven voor het achterwaarts gerichte pubis van het vogelbekken.

Soergel\'s reconstructie van Chirotherium is het resultaat van
groote scherpzinnigheid en streng logische redeneering.

Wat betreft de beteekenis der begrippen „diabaasquot; en „dolerietquot;
bestaat er groote verwarring.

Het albitisatie-proces doet ons een verklaring aan de hand voor
de genese van sommige alkali-gesteenten.

Ook in woestijngebieden moet aan chemische verweering een rol
van beteekenis worden toegekend.

Door von Koenigswald\'s mededeelingen in „De Mijn-
ingenieurquot; van November 1931 komt de vraag over het al of niet
voorkomen van Devoon op Celebes in een nieuw licht te staan.

Wegener\'s hypothese der continent-verschuivingen ondervindt
een belangrijke steun door het feit, dat een ter zake kundig geoloog
(Du Toit) met klem gewezen heeft op de groote geologische over-
eenstemming, welke bestaat tusschen de O.kust van Z.-Amerika
en de W.kust van Afrika.

VoLZ\' argumentatie van de pleistocene ouderdom der Trinillagen
is ten eenen male onvoldoende.

		H. silvalensis	0 0
	Our own	F.A.S. pl. LX	.3 \'i
	specimen				rt 0 -
				E T3
		fig- 2b	fig. 3c	fig. 4C	- s K a 0.
I Interval between lower border of for. magnum and occipital crest...........	142
2 Breadth of occiput.....	236	—	—	—	—
2:1 ...........	1.66	1.30	1-33	1.28	I.61

Fore talon I	2	3	4	1 5	Hind talon
69 ca 68	ca 69 69	66 65	65 62	53	29 30	1. M3 r. M»
35 35	32 35	32 32	28 32	27 25	12 10	1. M3 r. AP
	—	—		—	24 25	1. M\' r. M3

	3 I	3 2	3 3	3 4	3 5	3 6	3 7	3 8	3 9	Hind talon
Slope of the base . . .	b	- 9°	- 7quot;	-5°	-4°	-2°	-2°	- 1°	-0°
Slope of the top . . .	b	-i6°	-13°	-5°	-1°	2°	4°	u	u	—

	2 1	2 2	2 J	2 4	2 5	2 6	2 7	2 8	2 9	2 10	Hind talon
Slope of the base . . .	b	b	-r	-6°	-4°	-3°	-2°	0°	2°	2°
Slope of the top . . .	b	-12°	-8°	-5°	-2°	0°	u	u	u	U	--

Length X width of m\'	Length X width of M^
F.A.S.,pl. VI, figs. 4-5	4 X 2.4 inch = 102 X 61 mm.	5.5 X 2.7 inch = 140 X 69 mm.
„ pl. XII, fig. 2	3.8 X 2.3 „ = 97 X 58 „
pl. XII, figs. 4-4d		— X 2.8 „ = — X 71 „

					H. sivalensis Lydekker 1884, p. 39	Hippopotamus sp.
	Our own	specimen	type form (quot;var. latidensquot;)	aberrant form (quot;var. augustidensquot;)	Stremme 1911, p. 104.
	left	right	A	b	a	b	left	right
	1	b	1	b	1	b	1	b	1	! ^	1	b	1	b	1	b
Ml	35	35	33	35		—	34	44	46	38	.45	41	—	—	38	37
M2	40	41	38	41	47	54	42	53	53	44	54	47	39	38	48	48
M^	41	41	40	42	46	52	48	50	53

	Fore talon 1	2	3	4	5	Hind talon
Length at the base of the crown.	66	66	66	63	31	35==)
Interval between lingual extremi- ties of transverse valleys.....	35	37	36	38	28	21 3)
Height of the unworn or slightly worn ridge-crest.......	—	—	—	—	—	25

Fore talon	I	2	3	4	5	6	7	Hind talon
—	44	50	52	52	52	54	54	45
—	18\')	15	17	17	17	\'5	15	9
							45»)	45

F.A.S., pl. VIII, fig. 2	8.8 x 3.8	inch = 224 x 97 mm.
„ , pl. XI, fig. 2	ii.8 x 3	„ = 300 x 76 „
„ , pl. XI, fig. 8 1)	12.1 x 3.8	„ = 307 x 97 „
„ , pl. XI. fig. 7	10.2 x 4	„ = 259 x 102 „
„ , pl. XII, figs. 12—12a	10 x 3.j	„ = 254 x 89 „

- -------------- -	Length x width nij
F.A.S., pl. XII, figs. 8—8a		4.4 x 2,4	inch = 112 x 61 mm.
„ , pl. XII, figs. 10—lotf		6.7 x 2.6	„ = 170 x 66 „
.. . pl. XI. fig. 6		— x 2.8	,, = — x 71 ,.

		- 184 —		J
	E. hjsudricus.
Length X width m^
F.A.S., pl. VII, figs. 7-7^71)	3.1 X 1.5	inch = 79 X 38 mm.
Length X width mg
F.A.S., pl. XIIIA, fig. 7		5.4 X 2.3	inch = 137 X 58 mm.
„ , pl. VII, fig. 8		5.5 X 2.2	„ = 140 X 56 „

	MM\'\'quot;\'ij,

0M