(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Proboscidea : a monograph of the discovery, evolution, migration and extinction of the mastodonts and elephants of the world"

. ,., 



f^ 



I 1 



R¥ FAIRFIELD OSBORN 






World Distribution of I. Moerithere 



^ 




I ® Aiderltherilnde 
n ® Deinotheriinde 



m 



( c 



( ( 



Ai,astodontlnde 
JZyffolopkodontlnde 
jCongrlros trends 
" # Amebe Lo dontinde 



I 



n. Deinotheres in. Mastodonts 



180° 165' 150° 135° IZCT 105' 9dW 75° 60' 45' 30' 15 



Osbi 




180° 165° 150° 135° 120° 105° 90W 75 



HI ® TetraiopkodontlrKR 
# Serridentinde 
® Plat^beLodontLnde 
® Rhynctiorostrindt 



C4 
(6 



in @ A^torostrLnde 
/iamboLdtinae 
3reuiro3 trlnde 






D.F.LevettbradUn 



PROBOSCIDEA 



This original edition {issued 6fc<^-*«-*-7/^ ^YJCs) 

is limited to six hundred and seventy-five copies, 
of which this is Number ^ 




RESTORATION OF THE WARREN MASTODON {MASTODON AMERICANUS) 
After a painting by Charles R. Knight in 1908, under the direction of Henry Fairfield Osborn 

This restoration is based on the rkissir skeleton known as the Warren Mastodon, which was unearthed in the year 16'4o just 
north (if the Highlands of the Hudson and presented by the late J. Pierpont Morgan to the Hall of the Age of Man of The American 
Museum of Natural History in the year 1.906. The animal took its name from John C. Warren, Professor of Anatomy at Harvard 
University, who described it in detail in his memoirs of lSo2 and 1855. 

The pointing represents the anim(d, a bull at least thirty years old, as he appeared in life before he sank into a shell-marl 
basin six miles northu-esi of .\ewhurgh-on-Hudson, New York; this event, so auspicious to palaeontologists of today, occurred in 
the Late Clacial Epoch between 30,000 and 40,000 ijears ago. In the right background is the north gap in the Hudson River High- 
lands, Storm King Mountain on the right and Break Neck on the left. In the foreground is one of the shallow sheets of water 
covering the .'shell-marl in which the remains of the animal were .so perfectly preserved. The flora of the forest corresponds icith the 
description by Asa dray, in a letter to Doctor Warren, of the contents of the mastodon's stomach: "An e.ramination with the micro- 
scope showed, from the structure of the woody fibre, that they were boughs of pine or spruce of .some sort, and that they minutely 
agreed with the wood of hemlock spruce: sit that this is very probably the species they belonged to, but there is no certainty of it." 



PROBOSCIDEA 

A MONOGRAPH OF THE DISCOVERY, EVOLUTION, MIGRATION 

AND EXTINCTION OF THE MAS'IODONTS AND 

ELEPHANTS OF THE WORLD 



HV 

HENRY FAIRFIELD OSBORN 

A.B. Princeton, 1877; D.Sc. Princeton, 1880; Honorary LL.D. Trinity, 1901; LL.D. Princeton, 1902; 
Sc.D. Cambridge, 1904; LL.D. Columbia, 1907; Ph.D. Christians, 1911; D.Sc. Yale, 1923; 
D.Sc. Oxford, 1926; D.Sc. New York, 1927; LL.D. Union, 1928; Doctor 
of the University of Paris, 1931 ; Doctor of Natural Science. 
JoHANN Wolfgang Goethe University, 1934 

Research Professor OF Zoology, Columbia University; Honorary Curator-in-Chief of Vertebrate 

Palaeontology, The American Museum of Natural History; Senior Pal.«ontologist, 

United States Geological Survey; Honorary President, The American 

Museum of Natural History; Honorary President, 

The New York Zoological Society 

EDITED BY MABEL RICE PERCY 



VOLUME I 
MCERITHERIOIDEA 
DEINOTHERIOIDEA 
MASTODONTOIDEA 




PUBLISHED ON fHK J. IMERPONT MOKCiAN FIND BV IHE IRl'STEES 

OF THE AMERICAN MUSEl'M OF NATURAL HISTORY 

THE AMERICAN MUSEUM PRESS 

NEW YORK. 1936 



PUBLICATION NOTE 

Volume I has been on the American Museum Press since 1924. In the course 
of the intervening eleven-year period to 1935 the field of research lias been greatly 
broadened and many discoveries have been made which alter the Classification of 
the Proboscidea; the consecjuent alterations appear in the Table of Contents and 
in the Phylogenetic Aj^pendix to the present volume. 

At the time of the author's lamented death on November 6, 1935, the final 
revision of the manuscript was in progress. This has been carried out along the 
lines laid down by him; since that time no changes have been made in his deter- 
minations and necessary corrections have been inserted as footnotes. 

Volume II, containing Elephantoidea, Appendices, and Index, will be issued 
as soon as possible, probably in 1937. 



Copyright, 1936, by 
The Amehican Museum of Natural History 



IN MEM OKI AM 
lOANNIS PIER PONT MORCJAN 

ARTn'M ET SCIENTIARX'M 

AonroRis ARDEXriS E'|- I,1BERAMS 

HOC \C)L\'MEN 

AMICITIAE MNEMOSYNVM 

D. D. D. 

SCRIPTOR 



During the Years 1907 to 1936 

The Explorations, Collections, Researches and Printing of this Monograph 

Have Been Promoted Chiefly by Funds Donated by 

JOHN PIERPONT MORGAN 
1837 — 1913 

Founder, Trustee and Benefactor 

OF 

The American Museum of Natural History 



PREFACE 

THE author began his researches for this Monograph by devoting the years 1889-1900 to zoogeographical 
research on the origin, migration and distribution of the twenty-two Orders of Mammals known at that 
time. The theoretic conclusion was reached that the Proboscidea originated in Africa. This conclu-sion was 
confirmed in the Faytim district of northern Africa by Hugh Beadnell (1901-4) and by Charles William Andrews 
(1901-3). In Andrews' Memoir of 190G iniblished by the British Mu.seum (see p. 36 of the present monograph) 
he wrote: 

The probability that Africa would be found to be the original home of these animals was pointed out by several writers, 
notably Osboni, Stehlin, and TullberK. The first of these sugge.sted that probably not only the Proboscidea but also the 
'Hyracoidea, certain Edentates, the Antelopes, theCiraffes, the Hippopotami, the most specialized Ruminants, and among the 
Rodents the Anomalures, the Dormice, and Jerboas, among Monkeys the Baboons,' and, as his map suggests, the Sirenia also, 
originated in this reginn. Osborn also jiut forward the theory that a .succe.ssion of migrations from Africa to Europe occurn'd, 
notably at tlie end of the Eocene, at the beginning of the Miocene, and again in the earliest Pliocene. It was in the early 
Miocene migration that the Proboscidea passed out of Africa for the first time so far as known. 

The American Museum expedition of 1907 amplified the epoch-making discoveries of Beadnell and Andrews 
and sub.sequent exploration has demonstrated beyond question that in Africa originated not only the Moeritheres 
and primitive Mastodonts but also the Deinotheres and the Elephants. Not a trace of these i^roboscideans has 
been found in any other continent earlier than Miocene time, when waves of Deinotheres and Mastodonts, already 
highly specialized, began to enter Eurasia and to migrate to all the continents excepting Austraha. 

The actual prejjaration of this monograph opened in the year 1907 with the author's expedition to the desert 
bordering the Fayum of northern Egypt, accomjiiinied by Walter Granger of the mammalian iiala^ontology staff 
of the American Musetun. In 1908 the author determined to give the Probo.scidea a monograjjiiic treatment 
similar to that he was then giving to the Titanotheres. Intensive work on the Proboscidea, first entered 
upon in 1920 during the concluding phases of work on the Titanotheres, has continued without interrui)tion for 
fifteen years. Thus the intensive study of tlie evolution and phylogeny of two dominant and widely different 
types of mammals, the Titanotheres and the Proboscideans, has occupied a total uninterrupted re.search period of 
thirty-five years, 1900-1935. 

The Titanothere Monograph involved methods of collateral research in geology, comparative anatomy, 
animal mechanics and biology that were largely new to the science of palaeontology, chiefly in the science of measur- 
ing the skull (craniometry) and in the science of animal ascent (phylogeny). P>om the outset the craniometric 
methods and results of the Titanothere researches partially i)repared the way for the solution of proboscidean 
problems. In their adaptations the Proboscidea present the widest possible contrasts to the Titanotheres. 
Moreover, there was little concejition of the new and rapidly oncoming difficulties in interpreting the classic 
proboscideans as a whole or the entirely unexi)ected adai)tati()iis in numerous new lines of ma.stodont and ele- 
phantoid ascent. Previously undreamt-of kinds of proboscideans have been discovered. Through the classic 
and the new kinds have been revealed a number of biological principles hitherto unknown. 



xii OSBORN: THE PROBOSCIDEA 

METHODS OF COMPARATIVE RESEARCH 

In the analysis and syntliesis of thousands of observations on the Proboscidea it soon became obvious that 
five main Unes of research were essential to a judicial, comprehensive and philosophical treatment: 

First, to survey, for the restatement of historic descriptions and figures, the ancient hterature of what 
may be called "the dawn period" of mammahan paleontology, including Robert Kerr's naming of the 
American mastodon, Elephas americanus, in 1792 and Blumenbach's naming of the European mammoth, E. 
primigenius, in 1799; second, to study all the subsequent scientific descriptions of fossil and living proboscideans, 
with repubUcation of essential parts of the original descriptions and reproduction of all the original figures; third, 
to directly or indirectly reexamine all the fossil proboscidean collections in the world and especially to verify 
and refigure all types of genera and species; fourth, to secure the cooperation of mammalian palaeontologists all 
over the world in the examination of distant collections inaccessible to the author; fifth and most important, to 
approach this great and difficult subject with an entirely open mind, unbiased by preconceptions, each problem to 
be considered in a fresh and free spirit of observation, induction and generalization. 

Thus approached, the Proboscidea have proved infinite in variety, full of surprises, creative of new ideas and 
demonstrative of new evolutionary principles. 

DISCOVERY OF MULTIPLE LINES OF PROBOSCIDEAN ASCENT 
A research period of unparalleled difficulty proved to be also a period of unparalleled opportunity — oppor- 
tunity such as has never before been afforded to any zoologist, palaeontologist or biologist. There has been vast 
expansion of our knowledge since the year 1859 in which appeared Darwin's "Origin of Species," since 1886 when 
Richard Lydekker catalogued the Proboscidea in the great collection of the British Museum (Natural History), 
and even since 1929 when D. M. S. Watson's article, "Proboscidea," appeared in Volume XVIII of the Fourteenth 
Edition of the Encyclopaedia Britannica. Even within the year 1935 most surprising discoveries have been made 
bearing on the origin and distribution, affinity and ascent of the Proboscidea, which as far as possible are em- 
bodied in the phylogenetic appendix at the close of the present volume. 

The outstanding result of the expansion of our knowledge of the Proboscidea is the demonstration, partly in 
fulfilment of the prophetic vision of Hugh Falconer, that the single genus Mastodon of Cuvier is the prototype of 
the suborder Mastodontoidea, which comprises all the mastodonts of the world, as fully set forth in the Appendix. 
This suborder embraces thirty-one distinct lines of generic ascent — in brief, of genera ; these genera are grouped 
into fifteen distinct subfamilies, which in turn are grouped into four famihes, out of one of which it is possible 
that the Elephantoidea, the second great suborder of the Proboscidea, evolved. 

In like manner, as fully set forth in Volume II, the single genus Elephas of Georges Cuvier is now known to 
have divided into eight to ten separate lines of generic ascent, which in turn are grouped into four subfamihes, 
actually very wide apart in structure, in geographic distribution, in feeding habits and in adaptation. None of 
these elephantoid lines is known prior to Middle to Upper Pliocene time. The line with the most noble geologic 
and geographic hi.story is that of the genus Archidiskodon, which apparently dated back to the Middle PHocene of 
South Africa and enjoyed a glorious migration of 15,000 miles to its Late Pleistocene cUmax in Nebraska, Texas 
and Mexico. 

Thus, taken altogether, the Mastodonts and the Elephants, the Deinotheres and the Moeritheres — the four 
great branches constituting the great order PROBOSCIDEA— are now known to have evolved along no fewer 
than forty-one Unes of generic ascent. 



PREFACE xiii 

LINN^AN CLASSIFICATION REPLACED BY OSBORNIAN 
The above revolution, by the discovery of multiple lines of ascent, is first manifested in the zoologic and 
classificatory treatment of the Proboscidea, in which we are impelled to substitute for the purely zoologic pre- 
evolutionary classification of Linnaeus an evolutionary or phylogenetic classification based on princijiles first 
set forth by the author in 1892 and fully exemplified in the Titanothere Monograph (1929). Inherent in phylo- 
genetic classification are entirely new conceptions of the real evolutionary meanings of the old Linnaean terms 
Species and Genus, Family and Order, as well as of the more recent terms Superfamily, Subfamily, Subspecies 
and Mutations. The classic Linnaean system based solely on zoologic observation and the creational concept 
is now replaced by a ))hylogenetic system in which all divisions from the Subspecies to the Order are ])laced 
vertically as .succeeding each other during millions of years of geologic time, rather than horizontally as observed 
in recent or existing time by Linnaeus and all zoologists. 

The Osbornian phylogenetic system that was .set forth in the Titanothere Monograph (Vol. I, Fig. 14, p. 1(3) 
is literally fulfilled in one proboscidean family known as the Serridentidae, embracing seven complex lines of ascent 
of the serridentine mastodonts, namely, Serridentinus, Ocalientinus, Serbelodon, Trobelodon, Platybelodon, Toryno- 
belodon, Notiomastodon; also in an elephantoid sul:)family known as Mammontinae, embracing three related but 
widely distinct kinds of mammoths — the woolly, the hairy, the southern — Mammonleus, Parelephas and Archj- 
diskodon. Thus for the first time we arrive at the real evolutionary significance of the cla.ssificatory terms of 
Linnaeus — Order, Suborder or Superfamily, Family, Subfamily, Genus, Species, Subspecies, Ascending Mutation. 
The facts upon which these phylogenetic results of mastodontoid and elephantoid a.scent are establi.shed 
may be summarized as follows: 

1) The Proboscidea originated exclusively in Africa, probably in Upper Cretaceous or Lower Eocene time. 

2) Not remote from the stem that s'l^t^ ri-'^*' tx) tiie manatees and duf^ongs (Sirenia), proboscideans were distinguished 
from the first by specialization of the second incisor teeth above and below, also by the potentiality of forming transverse 
ridge-crests which, especially in the third superior and inferior grinders, rise from primitive 2Jr-3 ridge-crests to progre.ssive 
30 ridge-crests. 

3) The subfamily radiation into all habitats, from the i)urely aquatic and amphibious habitat of the Mceritheres to 
swamp, river-border, shallow lake-border, savanna, forest, tundra and desert habitat, from humid to arid climate.s, from 
eciuatorial to boreal latitudes and from sea level to elevated mountains (for example, the Andean ma.stodon), in which thirty- 
one lines of mastodonts and eight to ten hues of elei)hants have become specialized in everj- organ. 

4) The conversion of a broad prehensile upper lip of the hippopotamus type in the Ma-ritheres into the broad preheiLsile 
upper lip of many of the mastodonts and, finally, into the true proboscis of certain mastodonts and of all Stegodonts and ele- 
phants. 

5) The genera are primarily distinguishable (a) by the adaptations of the grinding teeth which, when very carefully 
analyzed, as in the Appendix to the i)resent volume, are highly characteristic, and (b) by the respective adaptations of the 
superior and inferior incisive teeth to a great variety of i)roboscidean floral habitats and modes of feeding; (c) in the Deino- 
theres, in three kinds of shovel-tuskers, al.so in the Rhynchotheres, the lower incisive tusks are far more importaiU and diversi- 
fied than the upper, while in many mastodonts and in all the Stegodonts and elephantoids the lower tusks disappear entirely, 
the lower jaw being correspondingly abbreviated; (d) the .suiierior tusks function first as ofTen!>ive weapons of marked 
sexual variation and second as auxiliary feeding tools of great importance, excepting in one divi.sion of the elephantoids in 
which the gigantic superior tusks are withdrawn from feeding service and become only of .sexual distinction and value. 

6) There is consequently frequent reciurence of adaptive parallelism, namely, of similarly specialized organs, like the 
shortened jaw being independently produced in difTerent phyla, or the rounded and upturned enamel4ess sujjerior tu.sks 
servuig as weapons. 

7) With the various functions and forms of the superior and inferior grinders, the superior tusks, the upper lij) and the 
proboscis are coordinated not only the cranium and jaws but also the neck, backbone, and limbs. For example, the three 
tyjjes of manunoths— the tundra {Mnmmnnlfits), the north temperate (Parelephas) and the southern (Airhidiskodoii) — alike 
e.xhibit enormous incur\ed tusks of use only in combat or for sexual .selection purjMi.ses and not as auxiliary feeding organs; 
in these three mammoths the posterior half of the backbone is relatively weak because the tusks have relati\ely little work to 
do as auxiliary feeding organs, wher(>as in the .\friean and Indian elephants and their relatives the backbone is extremely 
strong, in correlation with active tusks. 



xiv OSBORN: THE PROBOSCIDEA 

PROBOSCIDEA, THE NEW STANDARD OF CHRONOLOGIC WORLD CORRELATION 

Not revolutionary, but in accord with the clocking of geologic time by similar stages of evolution (homotaxis, 
Huxley) long in use by invertebrate paleontologists, is the evidence yielded by intensive examination which 
the Proboscidea afford for Tertiary geologic correlation of stages in adaptive progression and retrogression in the 
widely distant distribution centers of mastodonts and elephants which migrated into all the continents except 
Austraha. The continuous evolution of far-separated adaptations of the grinding teeth, in northern and southern 
Africa, in Eurasia and in North and South America, can now be coordinated with unexpected precision, as in the 
outstanding metamorphosis of Archidiskodon propkmifrons of South Africa into the gigantic Archidiskodon 
maibeni of Nebraska (see Volume II). In this correlation the intensive study of the grinding teeth in the masto- 
dontoid and elephantoid divisions plays a leading part. 




PROBOSCIDEAN CHRONOLOGY 

Another result which will prove to be revolutionary in anthropology is the new means afforded of dating 
precisely the main periods of the prehistoric evolution of Man, by intensive measurement of the length, width 
and thickness of the enamel layers in the grinding teeth of the elephantoid division (Osborn 193L858) in successive 
stages from Upper Pliocene to closing Pleistocene time. Man was a mastodon and elephant hunter from early 
times. Remains of fossil elephants are occasionally found embedded in the same strata with remains of fossil 
man, and the total length of the enamel foldings in proboscidean grinding teeth enables us to date relatively the 
successive phases in the evolution of man. 

ARISTOGENESIS AND PROPORTIONAL MODIFICATION 

P>om the author's prolonged thirty-five-year research on Titanotheres and Proboscidea there issue not 

only the principles governing the classic modifying modes of evolution known to Lamarck and Darwin (variation, 

development, degeneration) but also the newly discovered and hitherto unrecognized principle and modes of the 

origin of new characters through aristogenesis or creative biomechanical rectigradation (Osborn 1894.92,1934.922). 



PREFACE XV 

Ascending mutations, s])ecies and genera are princiixilly defined by continuously progressive changes in the 
aristogenes and proportions of the grinding teeth and in tlie upi)er or lower incisive tusks. Although tlie rate of 
proportional adaptation varies enormously, for the first time we know api)roximately how long a period of geo- 
logic time it takes to produce a full-fledged and highly efficient adaptation, as in the metamorphosis of the lower 
incisive tusks of Oligocene Phiomia into the dominant shovel-tusks of Pliocene Amebelodon (see Plate V, pp. 
236-237), or in the metamorphosis of the posterior grinders of Trilophodon from Lower Miocene into Mio-Pliocene 
time (see Fig. 300, p. 341). 




PRINCIPLES OF ADAPTIVE PROBOSCIDEAN ASCENT 

Through the clear distinction between change of proportion (alloiometry) and the origin (aristogenesis) 
of new parts, also through the newly discovered multiple lines (forty-one or more) of ancestry and ascent technical- 
ly known as phyla, the Proboscidea afford a complete revolution in our biological philosophy and concepts of the 
nature and causes of evolution. The contrast between Darwin's knowledge of the Proboscidea when he wrote 
the Origin of Species in the year 1859 and the knowledge contained in this monograph is graphically shown by 
comparing the two accompanying charts (1933) witli Plate XI (1935). 

Never before has it been jiossible to follow many lines of phylogenetic ascent over extremely long i)eriods of 
geologic time, noting the i)rogressive ada))tive changes in each organ in each phylum to gain the perfection of 
certain mechanisms at the expense of other mechanisms. In general, the specialization of certain organs becomes 
more intense, while closely contiguous organs remain absolutely stationary. For examjile, among the shovel- 
tuskers, Phiomin osborni of the Oligocene of Egypt gives rise to tiu- incredilily specialized A mebelodon fricki of 
Nebraska, with a relatively similar and unchanging skeleton and limbs. Amebelodon is jiaralleled by the flat- 
tusker Platybeledon of the Gobi Desert and of Nebraska in which the whole jaw becomes an enormous shovel, 
the upper jaw and skull being sacrificed and thus greatly reduced in size. 

\Thit^ fteclioii of the Preface ira.s not completed by the author. — Editor.] 



xvi OSBORN: THE PROBOSCIDEA 

ACKNOWLEDGMENT OF COOPERATION 

Cooperation in the preparation of this monograph has been practically worldwide, with the exception of the 
naturalists of Australia, the only continent not invaded by proboscideans. A full Ust of the names of cooperating 
authors and investigators is given in Chapter I, pages 13-15, also in the Bibliography at the close of this volume, 
but I should like to notice here some of the outstanding collaborators and assistants. 

My lifelong and honored friend, Andrew F. West of Princeton University, has greatly added to the value of 
the work by his Latin rendering of the Dedication. 

In the United States I am especially indebted to my friend, Erwin H. Barbour of the University of Nebraska, 
who placed at my service all his materials, his writings and his illustrations. In recent years my research as- 
sociate, Childs Frick, has rendered most timely service in his explorations, researches and astounding discoveries. 
European collaboration that has been invaluable was that of Sir Henry Lyons in supporting the Fayum expedition 
of the year 1907 with all the resources of the Geological Survey of Egypt. In his assistance on the home stretch 
of research on this monograph Arthur T. Hopwood has been particularly helpful in the Calcutta and British 
museums. In both France and Germany I have had the aid of a host of friends. 

Craniometry. New methods of measuring the skulls of the proboscideans were initiated in the year 1903 
with the aid of Wilham King Gregory in continuation of the new systems of measurement introduced in the 
researches on the craniology of the Titanotheres. In every subsequent year this most helpful and talented col- 
league has responded instantly to every call for cooperation and counsel. 

Odontography. The grinding teeth are the remains most frequently found and constitute the chief bases 
of scientific description and classification. The preparation of measurements, calculations and difficult drawings 
of the evolution of enamel foldings in the stegodontoid and elephantoid branches of the Proboscidea has occupied 
two years' time in the American Museum and the British Museum (Natural History) ; this work includes the 
excellent aid of Edwin H. Colbert of the American Museum staff in the calculation of indices, of William Dixon 
Lang and William E. Swinton of the British Museum in the new ganometric measurement of the superior and 
inferior grinding teeth of the proboscideans, of Miss Barbara Hopkins of the British Museum in her admirable 
drawings of molar tooth sections and of Sir Henry Lyons of the Science Museum in the preparation of diagrams. 

Systematic Identification. During the period from 1922 to 1930 Charles Craig Mook assisted in the 
arrangement of the rapidly incoming collections of fossil mastodonts and elephants. In the year 1930 Doctor 
Mook was succeeded by Doctor Colbert, who has rendered substantial aid in the more recent identification and 
measurements of the material. The proboscidean collection now numbers 1235 specimens. 

Illustration. Especially outstanding is the superb work of illustration rendered by Lindsey Morris 
Sterling from the year 1902 to the time of her death in 1931. Beginning with pen drawings of crania in the 
elephantoid branches, executed in imitation of wood engraving, shown especially in Chapters VI, VIII, X, XIV, 
XX, Mrs. Sterling wdth rare intelligence, accuracy and constant enthusiasm prepared the greater number of pen 
drawings and diagrams for this monograph and is thus entitled to an enduring rank among the leading scientific 
draughtsmen of the pre.sent century. To our deepest regret Mrs. Sterling did not live to see the pubUcation of 
these volumes in which her work will be preserved for all time. 

Reminiscent of the splendid series of illustrations of the extinct reptiles and mammals of North America 
are the frontispiece and five plates reproduced from the masterly oil, water color and pencil sketches of Charles 
R. Knight. All these sketches were executed before the habits and characteristics of these proboscideans were so 
precisely known as at present. In design and artistic conception the Knight restorations are unsurpassed. In 



PREFACE xvii 

the year 1928 Margret Flinsch, guided by the most recent studies of the author, entered u|x)n the work of restor- 
ation of the fossil elephants and mastodonts, after prolonged graphic study of tlie living elephants. Her work 
combines a fine sense of aecuraey witii rare artistic appreciation of the motions and adaptations of these remarkable 
animals. Beginning witli the Ma-ritheres of the Fayum slie lias vigorously and freshly restored, under the author's 
direction, all the chief types of proboscideans as members of the forty or more genera and twenty-odd subfamilies 
into which the Order is now divided. A prolonged and very difficult undertaking of the la.st three years has been 
the outline restorations of most of the known fossil and living types for phylogenetic charts illustrating the adaptive 
radiation of the mastodonts and elephants, also the modeling in pla.ster of the outstanding mastodonts and 
elephants, aided by the figures of Muybridge. 

The work of illustration also calls for grateful mention of the following artists: Alastair Brown and Roger 
Bullard for their ])en and pencil sketches; D. V. Levett Bradley for her majijiing and lettering; Al)ram E. 
Anderson for his hne pliotograjjhs of skulls and skeletons in the American Museum collection and for his de- 
scription of the microscoi)ic characters of the enamel of the i)roboscideans in Volume II; Elwin R. Sanborn of 
the New York Zoological Society for his ])hotograi)hy. 

Editinu. Beginning in the autumn of the year 1920 and continuing up to the time of pul)lication, Mabel 
Rice Percy has been the editor-in-chief of these volumes. Rarely fitted by intense appreciation of the po.ssibilities 
and results of scientific research, as well as by unflagging energy and patience, exceptional memory and rare 
precision. Miss Percy has rendered imjierishable service to the science of jmlaeontology. The taste and beauty of 
the typographic arrangement, the adjustment of the illustrations, the great labor of examining and correcting 
succeeding galley and i^age proofs to make this monograph a consistent and liarnionious statement of fact and 
measurement, are all due to Miss Percy. 

In the closing years Ruth Tyler has assisted in tlie final proofreading and editorial work of the text as a 
whole and, esi)ecially, in the Tal)le of Contents, Bibliograiihy and tlie Appendix. The introductory pages and 
Preface have been prei)ared by Florence Milligan. 

Editorial aiipreciation is also due to Francesca LaMonte and Marcelle Roigneau in the work of translating 
the citations from the French, Spanish, Italian and Rus.sian languages; to the Staff of the American Mu.><eum 
Library and to Jannette May Lucas for library assistance ; to Miss Percy for the extremely difficult a.ssemblage 
of titles for the Bibliography and to Miss Lucas for assistance in this compilation; under Miss Percy's direction 
fine work has been done by the American Museum Department of Printing, originally headed by Stephen Klassen 
and including Miss Ceroid and Messrs. Warther, Caggana and Klassen, Jr., and on the home stretch headed by 
Edward A. Burns with Messrs. Caggana, Haddon and Lewis and Mrs. Price assisting. 

I am indebted also to Professor Barbour, Harold J. Cook, Morris Skinner and C. Bertrand Schultz for as- 
sistance in fieUl geology; to James L. Clark and Martin Johnson in comparative zoology; to Nels C. Nelson and 
J. Reid Moir in iirehistoric archaeology. 



^^yUr^OocfiU/^^^^ 



CONTENTS 

PAGK 

PREFACE xi 

LIST ( )F ILLUSTRATIONS xxvii 

CHAPTER 

I. DISCOVKKV OF THE LIVINC, AND EXTINCT i'ROHOSCIDEANS 1 

II. ORKJIN AND CLASSIFICATION OF THE PROBOSCIDEA BECilNNING WITH M(ERITHERIUM, PAL.EO- 

MASTODON, AND PIIIOMIA 17 

Linnseau zoological classification and Osborn's phylogenetie classification of the Proboscidea 19 

Osborii's theory (1900) as to the African origin or center of adaptive radiation of the Proboscidea 34 

Phylogenetie relations and classification of Mwritherium, PahioinaModon, and Phiomia 46 

Re\ision and synojjsis of genera, sijecies, and types of the Fayuni proboscideans in the order of their original description .51 

Falieoniastodon beadnelli Andrews, 1901 54 

Ma'n'lherium lyonsi Andrews, 1901 54 

I'hiomia serridens Andrews and Beadnell, 1902 55 

Ala-ritheriiim granle Andrews, 1902 56 

Maritherium tn'godon Andrews, 1904 57 

I'hiomia minor Andrews, 1904 58 

I'nhfomnxtodon parvus Andrews, 1905 59 

I'hiomia winloni Andrews, 1905 60 

Phiomia barroisi Pontier, 1907 61 

Mivrilherium andrewsi Schlosser, 1911 61 

PdlaomasUtdon intermedins Matsumoto, 1922 63 

Phiomia osborni Matsumoto, 1922 64 

Mwrilhrriiiiii uncealrale Petronievics, 1923 65 

111. THE MCERITHERES (MCERITHERIOIDEA) OF NORTH AFRICA IN EOCENE-OLIGOCENE TIME 67 

Mocritherioidea Osborn, 1921, superfamily definition 69 

Ma>ritheriida' Andrews, 1906, family definition 69 

Mteritlieriini Winge, 1906; Mceritheriina' Osborn, 1923, subfamily definition 71 

Moeritherium Andrews, 1901-1906, generic definition 71 

lyonsi Andrews, 1901 72 

gracile Andrews, 1902 ; 73 

andrewsi Schlosser, 1911 74 

Irigodon Andrew's, 1904 74 

anrestrale Petronievics, 1923 76 

Supposed Moeritherium of Baluchistan 78 

1\'. THE DEINOTIIERES (DEINOTHERIOIDEA) OF WESTERN EURASIA AND OF AFRICA 81 

Deinotlieres distinguished from other proboscideans 81 

Curtognathidae Kaup, 1883, family definition and synonymy 82 

Deinotheriinffi Bonaparte, 1850; Winge, 1906; Osborn, 1910 83 

Deinotherioidea Osborn, 1921, superfamily definition 83 

Deinolheriiim Kaup, 1829, generic definition and synonymy 84 

Twenty-six species in order of description, 1715 1935 84 

History of discovery, 1715-1822, summary by Weinsheimer 86 

Historical summary, 1883-1934, by Osborn 95 

Resume of the chief generic and specific characters 110 

Conclusions as to origin, phylogeny, and adaptive radiation 112 

Geologic and geographic distribution of the types of the superfamily Deinotherioidea 114 

Deinolherium hopwoodi up. nov 117 

xix 



XX OSBORN: THE PROBOSCIDEA 

CHAPTER PAGE 

V. HISTORY OF THE CLASSIFICATION OF THE MASTODONTOIDEA, FAMILIES AND SUBFAMILIES, 

1705-1927 119 

Cuvier's five classic 'species' of Mastodon 120 

Cuvier's original conception of his genus 'Mastodonte,' 1806: Mastodonte de I'Ohio, Mastodonle a dents Hroites, Petit 

mastodonte de Montabusard, Petit mastodonte de Saxe, Mastodonte des Cordiliires, Mastodonte humboldien 122 

The Zygolophodontinae : Separation from the Longirostrinae 124 

The Brevirostrinae : Separation from the Longirostrinae and Zygolophodontinae 125 

Final separation of the Zygolophodontinae from the Mastodontinae 126 

Separation of species of 'Mastodon' and 'Elephas' described in America (1792-1874) 127 

VI. HISTORY OF THE SUBFAMILY MASTODONTIN^, THE TRUE MASTODONTS 131 

Mastodontinae, r6suin6 of subfamily characters 132 

Chronologic list of genera and species 136 

Progressive characters and generic distinctions of the Mastodontinae 138 

Palxomastodon Andrews, 1901, generic definition 143 

parvus Andrews, 1905 146 

intermedins Matsumoto, 1922 146 

beadnelli Andrews, 1901 147 

Miomastodon Osborn, 1922, generic definition 151 

merriami Osborn, 1921 154 

tapiroides americanus Schlcsinger, 1921, 1922 156 

Fliomastodon Osborn, 1926, generic definition 157 

ma«/ieTO Osborn, 1921-1926 157 

americanus praetypica Schlesinger, 1919, 1922 159 

sellardsi Simpson, 1930 160 

vexillarius Matthew, 1930 161 

Mastodon Cuvier, 1806-1817, generic definition 170 

americanus Kerr, 1792 170 

americanus rugosidens Leidy, 1890 171 

progenius Hay, 1914 172 

oregoneyisis Hay, 1926 173 

americanus plicatus Osborn, 1926 173 

moodiei Barbour, 1931 174 

grangeri Barbour, 1934 175 

raki Frick, 1933 175 

americanus alaskensis Frick, 1933 176 

Warren and Whitfield Mastodons 179 

MI. HISTORY OF THE SUBFAMILY ZYGOLOPHODONTIN^, THE YOKE-TOOTHED MASTODONTS 191 

Chronologic list of genera and species 192 

Collateral relationship to the subfamily Stegodontinae 195 

Scjiaration of Turicius from Zygolophodon 198 

Resume of the generic and specific distinctions 198 

Zygolophodon Vacek, 1877, generic definition 203 

pyrenaicus Lartet MS. (in Falconer, 1857) ; Lartet, 1859 206 

pyrenaicus aurelianetisis Osborn, 1926 207 

borsoni Hays, 1834 207 

Mastodon vellavus Aymard, 1847 [ = Zygolophodon borsoni vellavus] 209 

Mastodon vialetii Aymard, 1847 [ = Zygolophodon borsoni nialetii] 209 

Mastodon Buffonis Pomel, 1848 [ = Zygolophodon borsoni buffonis] 209 

Mastodon Zaddachi Jentzsch, 1883 [ = Zygolophodon borsoni zaddachi] 209 

Turicius Osborn, 1926, generic definition 212 

tapiroides Cuvier, 1806, 1821-1824; Desmarest, 1820-1822 217 

turicensis Schinz, 1824 219 

turicensis simorrensis Osborn, 1926 219 

aHjcMs Wagner, 1857 220 

virgatidens von Meyer, 1867 221 

wahlheimensis Klahn, 1922 223 



CONTENTS xxi 

CHAPTER PACK 

VIII. HISTORY OF THE SUBFAMILIES LONOIROSTRIN.E AND AMERELODONTIN.t:. ORIC.IN AND GEO- 

(iRAPHIC DISTRIBUTION OF THE LONG-JAWED MASTODONT PHYLA 225 

Bunomastodontidae, family characters 226 

I^ngirostrinae and Totraloj)hodontinae, subfamily characters 231 

Phiumia Andrews and Boadnoll, 1902, generic definition 236 

minor Andrews, 1904 239 

serridens Andrews and Beadnell, 1902 239 

minion i (cf . serridens) 240 

wintoni Andrews, 1905 241 

oshorni Matsumoto, 1922 244 

■pygm^us Dep6ret, 1897 246 

Trilophodon Falconer, 1846, 1857, generic definition 249 

Chronologic list of genera and species of the Longirostrinse and Amebelodontinae 250 

Systematic description of genera and species 252 

Trilophodon (continued) 

angustidens Cuvicr, 1806, 1817 252 

angustidens gaillardi Osborn, 1929 259 

angustidens var. nustro-germanirus Wegncr, 1908, 1913 259 

angustidens var. libycus Fourtau, 1918 260 

Choerolophodon Schlesinger, 1917, subgeneric characters 261 

(Choerolophndon) pentelicus Gaudry and Lartet, 1856, 1859, 1862 263 

pandionis Falconer, 1857 267 

palseindicus Lydekker, 1884 268 

Telrabelodon {Heminmslodon) rrepusnili Pilgrim, 1908, 1912 [ = Sui!ia] 271 

rhinjiensis Pilgrim, 1913; Osborn, 1932 272 

macrognathus Pilgrim, 1913 274 

cooperi Osborn, 1932 275 

inopinatus Borissiak and BelJacva, 1928 278 

hasnotensis sp. nov 279 

sendaicus Matsumoto, 1924 280 

engelswiesensis Klahn, 1922 281 

esselbornensis Klahn, 1922 281 

steinheimensis Klahn, 1922 281 

Other mastodontoids of Klahn 

Turicixis wahlheimensis Klahn, 1922 282 

Tetralophodon gigantorostris Klahn, 1922 282 

Anancus miniitoarvernensis Klahn, 1922 283 

Anancus giganlarvernensis Klahn, 1922 283 

Trilophodon (continued) 

pontileviensis Mayet-Fourtau, 1918 283 

simplicidens Osborn, 1923 285 

obscurus Leidy, 1869 285 

Genomastodon Barbour, 1917, subgeneric definition 290 

Megabelodon Barbour, 1914, 1917, subgeneric definition 290 

dinotherioides Andrews, 1909 291 

{Genomastodon) willistoni Barbour, 1914 293 

(Megabelodon) lulli Barbour, 1914, 1917 294 

ligoniferus Cope and Matthew-, 1915 298 

{Genomastodon) osborni Barbour, 1916 298 

giganieus Osborn, 1921 304 

Amebelodon (Trilophodon) hicksi Cook, 1922 307 

Amebelodon (Trilophodon) paladenlatus Cook, 1922 309 

Trilophodon (continued) 

abe/y Barbour, 1925 311 

frirki Peterson, 1928 312 

phippsi Cook, 1928 315 



xxii OSBORN: THE PROBOSCIDEA 

CHAPTER VIII — Continued page 

Trilophodon (continued) 

-pojoaquensis Frick, 1926 320 

cruziensis Frick, 1933 323 

(Tatabelodon) riograndensis Frick, 1933 324 

(Tatabelodon) gregorii Frick, 1933 324 

joraki Frick, 1933 326 

Summary of the origin and phylogenetic succession of the Trilophodonts 329 

Amebelodontinse Barbour, 1929, subfamily definition 333 

Descent from Phiomia of the Oligocene of Nortii Africa 334 

Amebelodon Barbour, 1927, generic definition 335 

frtcki Barbour, 1927 335 

sindairi Barbour, 1930 337 

Torynobelodon Barbour, 1929, of the Platybelodontinae, generic definition 338 

loomisi Barbour, 1929 338 

Mastodon [ = Trilophodon] anguslidens, cast of Cuvier's original type 340 

Central conules distinctive of Trilophodon 341 

IX. HISTORY OF THE SUBFAMILY TETRALOPHODONTINtE. ORIGIN AND CJEOGRAPHIC DISTRIBUTION 

OF THE MEDIUM-JAWED MASTODONTS OF THE PHYLUM TETRALOPHODON 343 

Distinctions between the Trilophodon phylum and tlie Tetralophodon pliylum, including the typical genus Tetralophodon 

and the very progressive subgenus Morrillia 343 

Tetralophodon contrasted with Synconolophus 347 

Phylum Tetralophodon Falconer, including the doubtful Lydekkeria, the more progressive and typical Tetralophodon, 

and the highly specialized Morrillia 348 

Chronologic list of genera and species 349 

Lydekkeria Osborn, 1924, subgeneric characters 353 

Tetralophodon (Lydekkeria) falconeri Lydekkor, 1877 354 

Tetralophodon (Lydekkeria) sinensis Kokcn, 1885 355 

Tetralophodontinse van der Maarel, 1932 356 

Tetralophodon Falconer, 1847, 1857, generic characters 356 

longirostris Kaup, 1832 357 

grandincisivus Schlesinger, 1917 360 

gigantorostris Klahn, 1922 362 

punjabiensis Lydekker, 1886 362 

bumiajuerisis van der Maarel, 1932 365 

campesier Cope, 1878 369 

precampester Osborn, 1923 [ = Tetralophodon (Morrillia) barbouri] 372 

elegans Hay, 1917 372 

Trilophodon (1 Tetralophodon) brazosius Hay, 1923 374 

fricki sp. nov 375 

Morrillia Osborn, 1924, subgeneric definition 377 

Tetralophodon (Morrillia) barbouri Osborn, 1921 377 

Summary of evolution and migration 379 

X. HISTORY OF THE NEW FAMILY SERRIDENTID^, SERRATE-TOOTHED MASTODONTS. ORIGIN AND 
GEOGRAPHIC DISTRIBUTION OF THE MEDIUM-JAWED SERRIDENTINUS OF EUROPE, ASIA, 

AND NORTH AMERICA 381 

European forest origin and migration to North America 381 

Serridentine grinders distinguished from Turicius, Trilophodon 382 

C^hronologic order of discovery and description of thirty-four species of Serridentinse, four of Platybelodontinse, and 

two of Notiomastodontinse 384 

Serridentinffi Osborn, 1921, subfamily definition 389 

Serridentinus Osborn, 1923, generic definition 393 

Serridentinus subtapiroideus Schlesinger, 1917 394 

Serridentinus mongoliensis Osborn, 1924 396 

Ocalientinus (Serridentinus) florescens Osborn, 1929 397 

Serridentinus gobiensis Osborn and Granger, 1932 398 



CONTENTS xxiii 

CHAPTER X — Continued page 

Serridenlinus progressiis Osborn, 1923 401 

Serrideyitinus proavus ( -ope, 1873 403 

Scrridenlinua productufi Copo, 1875 404 

Ocalientinus {Serridenlinus) republicanus Osborn, 1926 414 

Ocalientinus {Serridenlinus) bifolialiis Osborn, 1929 415 

Omlientinus {SerridentinuK) floridaniis Loidy, 1886 416 

Ocalientinus {Serriderdinus) jloridanus leidii Frick, 1926 419 

Ocalientinus {Serridenlinus) obliquidens Osborn, 1926 419 

Serridenlinus serridetis ("'ope, 1884 423 

Serridenlinus unguirivalis Osborn, 1926 425 

Serridenlinus nebrascensis Osborn, 1924 429 

Serridcntinus serridens cimarronis Copn, 1893 429 

Serridenlinus breu'sterensis ( )sb()rn, 1926 430 

Dibelodon[ = Serbelodon{'!)] prsecursor Cope, 1892, 1893 431 

Serridenlin us gunlemalensis Osborn, 1926 432 

Ocalientinus Frick, 1933, generic definition 433 

Ocalientinus njoraliensis Frick, 1933 435 

Serbelodon Frick, 1933, generic definition 443 

Serbelodon barboureiisis Frick, 1933 443 

Serbelodon burnhatni Osborn, 1933 444 

Serbelodon (?) precursor Cope, 1892, 1893 445 

Trobelodon Frick, 1933, generic definition 446 

Trobelodon taoensis Frick, 1933 446 

Serridenlinus (continued) 

Serridenlinus barslonis Frick, 1933 447 

Proboscidean life zones and geologic horizons of India 448 

Serrideidinus of India, Japan, and China 

hasnotensis Osborn, 1929 452 

brmmi Osborn, 1926 452 

metachinjiensis ( )sborn, 1929 456 

chinjiensis Osborn, 1929 456 

prochinjiensis Osborn, 1929 457 

anneclens Matsumoto, 1924 457 

lydekkeri Schh)sser, 1903 (1906) 457 

wiinani Hopwood, 1935 458 

Platybelodontiiiae Rorissiak, 1928, subfamily definition 459 

History of discovery, separation from the Amebelodontinae 459 

Platybelodon Borissiak, 1928, 1929, generic definition 459 

danovi Borissiak, 1928 461 

grangeri Osborn, 1929 463 

Torynobelodon barnumbrowni Barbour, 1931 470 

Serridenlinus filholi Frick, 1926, 1933 473 

Serridenlinus {Ocalienlinusl) nebrascensis Osborn, 1924 473 

XI. HISTORY OF THE IHIVNCHOROSTRINES, SUBFAMILY RHYNCHOROSTRIN^. ORIGIN AND GEO- 

CiRAPHIC DISTRIBUTION OF THE BEAK-JAWED MASTODONTS 475 

Rhyncholherium distinguished from other bunomastodonts 475 

Rhyncliorostrina? Osborn, 1918, 1921, subfamily definition 477 

Chronologic order of discovery antl description of the lihynchorostrinse 479 

Rhyncholherium Falconer, 1856 (MS.), 1863, 1868, generic definition 480 

spenceri Fourtau, 1918 485 

brevidens C^ope, 1889 485 

shepardi Leidy, 1871 487 

rerlidi-ns Osborn, 1923 488 

[ = lilirkothcriuni] cuhypodon Cope, 1884 489 

anguirivale Osborn, 1926 491 

tlascalse Osborn, 1918, 1921 493 



xxiv OSBORN: THE PROBOSCIDEA 

CHAPTER XI — Continued page 

Hhynchotherium (continued) 

hrowni sp. nov 494 

falconeri Osborn, 1923 494 

shepardi edense Frick, 1921 496 

paredensis Osborn, 1929 [ = Rhynchotherium shepardi edense] 499 

Rhynchotherium (?) francisi Hay, 1926 501 

chinjiense Osborn, 1929 502 

New Honduras phylum : BUckotherium and Aybelodon 505 

BUckolheriumhlicki Frick, 1933 508 

Aybelodon hondurensis Frick, 1933 509 

Summary of the evolution of the Rhynchorostrines 511 

XII. THE SOUTH AMERICAN MASTODONTS, OF THE GENERA CORDILLERION, CUVIERONIUS, AND 

NOTIOMASTODON 515 

Historical introduction 515 

Chronologic order of discovery and description 537 

Separation of the genera Cordillerion and Cuvieronius 540 

Notorostrinse Osborn, 1921, subfamily definition 542 

Cordillerion Osborn, 1926, generic definition 543 

anditim Cuvier, 1806, 1824 549 

Mastodon [ = Noliomaslodon] argentinus Ameghino, 1888 550 

Cordillerion (continued) 

tarijensis Ameghino, 1902 550 

bolivianus PhiHppi, 1893 551 

Teleobu)ioniastodon{'?) Revilliod, 1931, bolivianus — Cordillerion (?) bolivianus ref 552 

tropicus Cope, 1884 553 

oligobunis Cope, 1893 554 

oligobunis antiquissimus Freudenberg, 1922 555 

Cordillerion (?) oligobunis felicis Freudenberg, 1922 556 

Cordillerion (?) oligobunis intermedius Freudenberg, 1922 557 

oligobunis progressus Freudenberg, 1922 558 

gratum Hay, 1917 559 

edensis Osborn, 1922 560 

orarius Hay, 1926 562 

defloccatus Hay, 1926 564 

bensonensis Gidley, 1926 565 

New knowledge of Cuvieronius afforded by recent discoveries in Bolivia (Tarija) and in Ecuador (near Punin and 

Alangasi) 567 

Type cranium and skeleton of ' Bunolophodon' [ = Cuvieronius] postremus of Alangasi, Ecuador, and type cranium of 'B.' 

[ = Cuvieronius] ayorse of Punin, Ecuador 571 

Humboldtidae fam. nov 575 

Humboldtinae Osborn, 1934, subfamily definition 575 

Cuvieronius Osborn, 1923, generic definition 575 

humboldtii Cuvier, 1806 (in Desmarest, 1818) 576 

brasiliensis Lund, 1841 (in Lesson, 1842) 578 

bonaerensis Moreno, 1888 579 

platensis Ameghino, 1888 579 

supei-bus Ameghino, 1888 580 

rectus Ameghino, 1889 580 

maderianus Ameghino, 1891 581 

chilensis Philippi, 1893 581 

pirayuiensis Gez, 1915 582 

mjorse S])illmann, 1928-1931 583 

postrem us Spillmann, 193 1 585 

Synopsis and critique of Dr. Angel Cabrera's revision of the Argentine Mastodonts, including the new genus and species 

Notio7naslodon ornntus 587 

N oliomastodon Cabrera, 1929, generic definition 590 

ornatus Cabrera, 1929 590 



CONTENTS XXV 

CHAPTKK XII — Continued page 

Notiotnastodon (continued) 

argentiniis Aiiiculiino, 1888 592 

Conipiirison of Cuvieronius plateiinis and Cumeronius superbus 593 

Eubelodon Barbour, 1914, generic definition 601 

morriUi Barbour, 1914 602 

XIII. THE SUBFAMILY liHEVIROSTRINiE ol laiiOl'IO AND ASIA, GENERA ANANCUS, PENTALOPHODON, 
AND SYNCONOLOPHUS. STEGOMASTODON OF THE SUBFAMILY HUMBOLDTIN^ IN NORTH 

AMERICA 617 

Historical introduction 617 

Synopsis of the nineteen species of Brevirostrinse r.nd eigiit of the Humboldtinae 625 

BunoniastodontidsB 627 

Brcvirostrina; ( ).sborn, 1918, subfamily definition 627 

Anancus Aymard, 1855, 1859, generic synonymy 630 

macropluK genotype 631 

aiTerni'iisi.f ( 'roizet and Jobert, 1828 632 

arverriensis brevirontrin Cii'Tviiia and de Serrcs, 1846 634 

falconerj Osborn, 1926 636 

minutoarvernennis and giganlarvernensis Klahn, 1922 637 

iniermedius Eichwald, 1831 639 

arverriensis progressnr Klionienko, 1912 639 

sinensis Hopwood, 1935 640 

perimensis Falconer and ( 'autley, 1847 643 

properimensis sp. nov 647 

Penlnlophodon Falconer, 1857, 1865, generic definition 647 

sivalensis Cautley, 1836 650 

falconeri sp. nov 653 

Synconolophus Osborn, 1929, generic definition 654 

plyrhodii.s Osborn, 1929 657 

corrugdtu.s Pilgrim, 1913 658 

hasnoli Pilgrim, 1913 659 

dhokpnthanensisOi^hoTU, 1929 661 

propdthanensis Osborn, 1929 665 

Humboldtidifi fam. nov 667 

HumboldtiniE Osborn, 1934 667 

Stegomastndon Polilig, 1912, generic definition and synonymy 667 

chapmaiii Hays, 1834, 1843 669 

mirijkus Leidy, 1858 669 

successor Cope, 1892 671 

texanus Osborn, 1924 673 

amona? Gidley, 1924, 1926 678 

aftonix Osborn, 1924 682 

priesllei/i Hay and Cook, 1930 684 

primiliriis sp. nov 684 

APPENDIX TO VOLUME 1 685 

Adapti\e radiation of the Mastodontoidea as discovered and analyzed up to and including the year 1935 685 

Palaeomastodontinae subfam. nov 691 

Miomastodon depereti sp. nov 693 

Mastodon pnvlowi sp. nov 694 

Mastodon grangeri Barbour, 1934 695 

Mastodon acutidens sjj. no\' 696 

Turicius turicensis from Up]ier Bavaria 697 

Tiiririiis sj). from ( 'hina 699 

Zi/golophodon borsoni ref. from ( 'liinu 699 

Stegolophodontinse subfam. no\- 700 

Stegolophodon lydekkeri sp. nov 700 



xxvi OSBORN: THE PROBOSCIDEA 

APPENDIX TO VOLUME I — Continued page 

Trilophodon connexiis Hopwood, 1935 702 

Trilophodon spectabilis Hopwood, 1935 702 

Tetralophodon exoletus Hopwood, 1935 "04 

The "Spoonbill" Mastodonts 707 

Gnathabelodontinse Barbour and Sternberg, 1935, subfamily definition 710 

Gnathabelodon Barbour and Sternberg, 1935, generic definition 711 

Gnathabelodon thorpei Barbour and Sternberg, 1935 713 

The Shovel-tuskers: Phiomia-Amebelodon 715 

Anancus sinensis Hopwood, 1935 721 

Humboldtidse, family definition 722 

Distinctions between Cuvieronius platensis and Cuvieronius superbus 722 

Stegomastodon primilivus sp. nov 726 

Serridentidae, family definition 729 

Notiomastodontinae subfam. nov '30 

Notiomasiodon ornatus Cabrera, 1929 731 

Serridentinus wimani Hopwood, 1935 732 

Oralienlinus ernmonsi Hay, 1930 733 

Osborn's final (1935) classification of the Mceritherioidea, Deinotherioidea, and Mastodontoidea 735 

LIST OF CATALOGUE NUMBERS OF SPECIMENS MENTIONED IN THE PROBOSCIDEA MEMOIR, VOLUME I 745 

BIBLIOGRAPHY OF THE PROBOSCIDEA, 1602-1935 761 



LIST OF ILLUSTRATIONS 

PLATE PAGE 

Fkontispiece. Restoration of t lie Warren Mastodon (Mastodon americanus). After painting by Knight. 

I. Mastodontinse : Pliomastodon, Miomaslodon, and Mastodon 134-135 

II. Zygoloijliodoiitina;: Zi/golophodon 134-135 

III. ZyKolopliodoiitinffi: Turicius 134-135 

IV. Stegolophodontinae; Slegolophodon 134-135 

V. .Anu'bclodontinie: Phiomin and Amehelndon 236-237 

VI. Anu'hciodontiiiie: Phiomin nxhorni i\ni\ Amebelodon fricki, primitive Lower Oligoeene shovpl-tuskers. Resttjrations by 

Flinsch 236-237 

VII. Miocene-Pliocene life zones of the long-jawed, shovel-tusked, and Huniboldtine mastodonts of Nebraska, Kansas, 

( 'olorado, and New Mexico 328-329 

VIII. I'leistoceiie of Iowa, Nebraska, and Kurope. Correlation chart 348-349 

IX. Type jaw of Notioiiuidudun urnatus ( 'abrera, also restoration by Flinsch 382-383 

X. Phylogeny of theMoeritherioidea, Deinotherioidea, and Mastodontoidea, 1934. Folding chart 684-685 

XI. Phylofieny of the Ma>ritlierioidea, Deinotherioidea, Mastodontoidea, and Klephantoidea, 1935. Folding chart. . . .684 685 
X 1 1 . TyiH' jaw and dentition of Nolionidslodun ornalus Cabrera. liy BuUard 730-731 

FIGURE 

1. Heads of twenty-nine types of jjroboscideans. Kcstorations by Flinsch 16 

2. Flood-plain scene of the ancient river Nile in Lower Oligocene time. Restoration by Flinsch 18 

3. Diagram showing Osborn's theory (.June, 1921) as to the adaptive radiation of the Proboscidea 20 

4. Fundamental arrangement of the cutting teeth, I--I2, in four of the suborders or superfamilies of the Proboscidea 23 

5. Chief head and dental forms of four of the suborders (I-V) of the Proboscidea 25 

6. Worldwide distribution of the Proboscidea in past and present time 29 

7. Osborn's diagram showing the known and unknown stages (April, 1925) in the adaptive radiation of the Proboscidea 32 

8. Osborn's diagram showing the adaptive radiation of the 37-41 generic phyla of the Proboscidea, as discovered up to 1934 33 

9. Division of the world into three realms (Huxley) and eight main geographical regions. After Osborn 35 

10. Orders of mammals as placed by Osborn in 1900 in their hypothetical chief centers of Tertiary adaptive radiation 35 

11. Section showing lowering levels of the ancient Lake Moeris, of the brackish Birket-el-Qunm, and of the sea. After Osborn. . . 37 

12. Map of the Faytim area, formerly covered by Lake Moeris, now reduced to the brackish Rirkct-el-Qurun 37 

13. Mariiheriiim, side view of head with eye and ear in position. Model by Christman 38 

14. Phiomin wintoni, side view of head. Model by Christman 38 

15. (Loft) Eosiren libyca, the Eocene sea cow; (right) Trichechiis americanus, the manatee. Restorations by Knight 39 

16. (Left) Moeritherium, Phiomia, and Parelephas, evolution of the head, proboscis, nostrils, and tusks. (Right) Maeritheriiitn 

nndrewsi, Phiomin oshnrni, and Pnrelephns jeffersonii. Restorations by Knight 40 

17. Phiomia osborni, primitive Lower Oligocene .sliovel-tusker of tiie Faytim, I']gypt. Restoration by Flinsch 45 

18. Mamtherium. lyonsi, first reconstruction of skull and mandible by Charles W. Andrews 46 

19. Mceritherium, front view of head. Model by Christman 48 

20. Phiomia wintoni, front view of head. Model by Christman 48 

21. Moeritherium, Palseomastodon, and Phiomia, superior molars 50 

22. Mceritherium, Palseomastodon, and Phiomia, inferior molars 50 

23. Section through Upper Eocene and Lower Oligocene formations north of Lake Qurun, I^'aydm, Egypt 51 

24. Map sliowing geographic distribution of species of the Fayuni Proboscidea 52 

25. Complete faunal list (1906) of Qa.sr-el-Sagha and Fluvio-nuirine Series. After Andrews 53 

26. Palseomastodon beadnelli Andrews, type figure 54 

27. Mwrithrrium lyonsi .\ndrews, type figure 55 

28. Phiomia serridens Andrew.s and Beadnell, type figure 56 

29. Mwritherium gracile Andrews, type figure 56 

30. Marithniitm triqndon .Andrews, type figure 57 

31. Palieomastodon [Phiomia] minor .\ndrews, type figure 95 

32. Pabeomastodon parvus Andrews, type figure 60 

33. PaLromaslodon [Phiomin] wintoni .Andrews, type figure 60 

34. Pnheomnslodon bendneUi, Phiomia wintoni, Palaomnstodon parvus, Phiomia minor, type second and third lower molars 60 

35. Pal-ifomastodon [Phiomin] barroisi Pontier, type figure 61 

36. Mceritherium andrewsi Schlosser, type figure. After .\ndrews 61 

37. Ma;ritherium amlrcwsi Schlosser, paratype and referred specimens. After Andrews 62 

xxvii 



xxviii OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

38. Palxomastodon irdermedius Matsumoto, type and paratype figures 63 

39. Palxomastodon intermedius Matsumoto, type jaw and paratype skull 64 

40. Phiomia osborni Matsumoto, type figxu-e 64 

41. Primitive Faylim proboscideans : Paheomaslodon, Phiomia, and Mceritherium. Restorations by Flinsch 66 

41a. PaIa'omastodo7i, Phiomia, and Mwritheriiim. Restorations by Flinsch - 66 

42. Osborn's reconstruction (1920) of cranium and jaws of Mceritherium andrewsi-trigodon 68 

42a. Ancient Mceritheres of northern Africa. Restorations by Flinsch 68 

43. Mceritherium lyonsi, axial and appendicular skeleton. After Andrews 70 

44. Mceritherium lyonsi Andrews, second type figm-e, also paratype figures "2 

45. Mceritherium gracile and M. lyonsi, referred mandibles. After Matsumoto 73 

46. Mceritherium Irigodon and M. andrewsi, American Museum Collection of 1907 75 

47. Mceritherium ancestrale Petronievics, type figiue 76 

48. Mceritherium., Palxomastodon, and Phiomia, comparison of superior and inferior molars 77 

49. Comparison of hexabunodont grinders of Palxomastodon with tetrabunodont grinders of Mceritherium 78 

50. Trilophodon (Moeritheriumf) pandionis (?) from the Bugti Hills, incomplete fourth superior permanent premolar 79 

51. Mceritherium andrewsi. Restoration by Flinsch 79 

52. Deinotherium giganteum type, as originally figured by Kaup (1829) with incisive tusks erroneously upturned 80 

53. Deinotherium giganteum type, as refigiu-ed by Kaup (1835?) with incisive tusks downturned 80 

54. Deinotherium giganteum, referred superior dentition. After Gaudry 80 

55. Fundamental arrangement of the cutting teeth in four of the suborders or superfamilies of the Proboscidea 83 

56. Map showing geographic distribution of types and referred specimens of the Deinotheriinse 84 

57. Deinotherium giganteum in the British Museimi, four views of reconstructed original Eppelsheim skull. After Andrews — 88 

58. Deinotherium giganteum, first Eppelsheim skull and jaws. After cast in Darmstadt Museum 89 

59. Deinotherium giganteum, restored Epplesheim skull and jaws. After Falconer 89 

60. Dinotherium intermedixim de Blainville, type figure 90 

61. Dinotherium gigantissimum, referred grinding teeth. After Stefanescu 96 

62. Dinotherium gigantissimum, tibia and pes. After Stef Snescu 97 

63. Deinotherium bavaricum ref., mounted skeleton from Franzensbad, Bohemia, in Vienna Museum 99 

64. Deinotherium bavaricum of Franzensbad, showing original and restored portions of mounted skeleton in Vienna Museum . . 100 

65. Deinotherium bavaricum in Vienna Museum. Model by Abel (1932) 101 

66. Deinotherium, restoration of head. Mter Gregory and Osborn (1910) 102 

67. Deinotherium, restoration. After Abel (1922) 102 

68. Dinotherium hobleyi Andrews, type figure 104 

68a. Deinotherium hopwoodi sp. nov., type molars 104 

69. Deinotherium, progressive increase in size of mandible 106 

70. Deinotherium bavaricum, D. giganteum, D. gigantissimum. Restorations (1932) by Flipsch 106 

71. Deinotherium, composite figure of deciduous and permanent dentition. After Lartet, 1859 108 

72. Proportionate restorations (1930) of Deinotherium (bavaricum, giganteum, gigantissimum) by Sterling 110 

73. Deinotherium gigantissimum Stefanescu of the Manza^l Valley, Rumania. Restoration by Flinsch 116 

74. Prodinotheriuni [Deinotherium] hungaricum Ehik, type figure 117 

75. Representatives of fourteen genera of the suborder Mastodontoidea up to the year 1935. Restorations by FUnsch 118 

76. Mastodonte de I'Ohio or Le Grand Mastodonte [Mastodon americanus], Cuvier's types 120 

77. Petit mastodonte [Turicius tapiroides] from Montabusard, Cuvier's original type figure 120 

78. Mastodonte d dents etroites [Trilophodon angustidens], Cuvier's original type figure 121 

79. Mastodonte des Cordilieres [Cordillerion andium], Cuvier's original type figure 121 

80. Mastodonte humboldien [Cuvieronius hiimboldtii], Cuvier's original type figiu-e 121 

81. 'Mastodon ohioiicus' [ = M. (?)pavlowi ref.] molar from Russia. After Pavlow 126 

82. Warren Mastodon as moimted in 1845-1846 130 

83. Warren Mastodon as remounted in 1849 130 

84. Two superior molars figured by Buffon in 1778, referable to Zygolophodon borsoni and Mastodon americanus 134 

85. Map of Big-Bone Lick, Boone County, Kentucky 136 

86. Map showing geographic distribution of the types of the Mastodontinae 137 

87. Mastodon americanus, jaw and inferior dentition. After Warren 140 

88. Pliomastodon matthewi and Palxomastodon intermedius, comparison of third inferior and superior molars 140 

89. Evolution of molar ridge-crests in the Mastodonts, Oligocene to Pleistocene 141 

90. Superior view of largest known jaws of Phiomia and Palxomastodon superposed for comparison 142 



LIST OF ILLUSTRATIONS xxix 

FIGURE PAGE 

91. Palxomxisiodon interrnedius Matsumoto, type and paratyiJc specimens 144 

92. Palxomastodon intermedins Matsumoto, type and paratype grinding teeth 144 

93. Key to hexabunodont ui)per and lower molar crown.s of Puln'onm>itodon intermedins 145 

94. Detailed studies of third superior and inferior molars of Palxomastodon intennedius and Marilherium lyonsi, M. trigodon, and 

M. andrewsi 145 

95. Palo'o mastodon beadnelli Andrews, second type figure 147 

96. Palxomastodon beadnelli, referred limbs and vertebrae. After Andrews 148 

97. Palxomastodon as a forest browser. Restoration (1932) by Fhnsch 149 

98. Miomastodon and Pliomastodon type molars compared with referred molar of the true Mastodon americanus. After Osborn . . 150 

99. Miomastodon merriami Osborn, type figure (superior molars and tusks) 154 

100. Miomastodon merriami refcrrcid mandible, inferior and sujx'rior molars and tusks, Pawnee C^reek, Colorado. After Frick .... 155 

101. Pliomastodon mattheivi Osborn, second type and paratype figures 158 

102. Pliomastodon mattheivi Osborn, third type and paratype figures 158 

103. Pliomastodon americanus praeiypica Schlesinger, cotype molars 159 

104. Pliomastodon sellardsi Simpson, type figure (lower jaw) 160 

105. Pliomastodon sellardsi Simpson, missing part of right lower jaw of type 161 

106. Pliomastodon vexillarius Matthew, type skull and jaws 162 

107. Pliomastodon vexillarius Matthew, type lower jaws 163 

108. Pliomastodon vexillarius Matthew, type femur 163 

109. Mastodon americanus male and female. Restoration (1933) by Flinsch 164 

110. Palxomastodon, Pliomastodon, Miomastodon, Mastodon, Zygolophodon, and Turicius. Restorations by Flinsch 164 

111. Superior grinding teeth of Zygolophodon and of Mastodon. After Buffon, 1778 166 

112. Le Grand Mastodonte, or Mastodonte de I'Ohio [Mastodon aiJiericanus], Cuvier's types 167 

113. Ohio-Incog nitum [Mastodon americamis], Blumenbach's original type figui'e 168 

114. Mastodon americanus from Catawba, Ohio, mounted skeleton in Ohio State University 169 

115. Mastodon americanus rugosidens Leidy, type figure (third left inferior molar) 171 

116. Ohio-Incognitum [Mastodon americanus], Blumenbach's original type figure 171 

1 17. Mastodon americanus, referred third inferior molar 1 ' 1 

118. Mammut progenium [Mastodon progenius] Hay, type figure (lower jaw) 172 

119. Mammut oregonense [Mastodon oregonensis] Hay, type figure (second left superior molar) 178 

120. Mastodon americanus plicatus Osborn, type superior true molar series 173 

121. Mastodon moodiei Barbour, type skull and mandible. After photographs 174 

122. Mastodon raki Frick, type; figure (incomplete mandible with dentition) 175 

123. Mastodon americanus alaskensis Frick. Type figure (incomplete mandible with dentition) 176 

123a. Eurasiatic ancestors of the typical Mastodon of America. Theoretic migration lines to North America 176 

123b. Geographic distribution of the typical mastodonts of the United States and Eastern Canada 177 

124. Warren Mastodon [Mastodon americanus], mounted skeleton in the American Museum 178 

125. Whitfield Mastodon [Mastodon americanus], mounted skeleton in the Senckenberg Museum, Frankfort, Germany 180 

126. Warren Mastodon [Mastodon americanus], superior, inferior, and lateral views of the tusks showing annular growth rings. 

After Osborn 182 

127. Elephas indicus (?) sumalranus, referred tusks showing annular growth rings. After photograph 183 

128. Mastodon americanus ref . from near Fulton, Indiana, palatal view of female skull 184 

129. Mastodon americanus ref. from Otisville, New York, palatal view 185 

130. Warren Mastodon [Mastodon americanus], superior view of skull 185 

131. Mastodon acutidens sp. nov., male skull and jaw from Hoclicstcr, Iiul., and skull of calf of M. americanus ref. from Hack- 

ettstown, N.J 186 

132. Mastodon americanus ref. male and female crania, from Ashley and Fulton, Indiana 187 

133. Mastodon americanus ref. from Buffalo, Kansius, crown \-i('w of male (?) palate and lower jaw of middle-aged individual . . . 188 

134. Mastodon americanus ref. from near Fulton, Indiana, middle-aged female palate and lower jaw 188 

135. Mastodon acutidens sp. nov. from Rochester, Indiana, showing sharp summits of partly developed molar crowns excavated 

from a young male skull 189 

136. American Mastodons along the ancient Missouri River of Kansas. After mural by Knight in the American Museum .... 190 

137. Map showing geographic distributon of types and referred specimens of Zygolophodon and Turicius 193 

138. Miomastodon depereti sp. nov., Trilophodon pontileviensis, and Turicius tapiroides molars of the Lower Miocene. After Mayet 196 

139. Turicius turicensis [ = tapiroides] and Trilophodon pontileviensis molars of the Lower Miocene. After Mayet 196 

140. Turicius turicensis referred molar, from Murinsel, Croatia (cf. T. virgatidens). After Vacek 196 



XXX OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

141. Turicius turicensis simorrensis Osborn, type molar. After Gaudry 197 

142. Cotypes of Mastodon [Stegolophodon] cautleyi Lydekker equal lectotypes of Osborn. After Falconer and Cautley 197 

143. Middle ( = £tage Helvetien) and Upper ( = £tage Tortonien) Miocene correlation. After Deperet (1905-1908) and Osborn 

(1910) 201 

143a. Ascending phylogenetic succession of three species of Turicius and one of Zygolophodon. Restorations by Flinsch 201 

144. Head of Turicius turicensis Scliinz, exhibiting supplementary superior incisive tusks. Reconstruction by Pontier and Anthony 202 

145. Craniimi of Loxodonta africana, exhibiting reduplication of superior incisive tusks. After Anthony and Prouteaux 202 

146. Chief Lower Miocene localities { = fitage Biudigalien, Sables de I'Orleanais stage) and correlation of France, Switzerland, and 

Austria. After Deperet (1905-1908) and Osborn (1910) 204 

147. Zygolophodon pyrenaicus Lartet, type molar. After Gaudry 206 

148. Turicius turicensis simorrensis Osborn and Zygolophodon pyrenaicus Lartet, type molars 207 

149. Mastodon americanus ref., crown view of inferior dentition and jaw. After Warren 208 

150. Zygolophodon borsoni, referred inferior dentition and jaw, from Vialette, France. After Lortet and Chantre 208 

151. Zygolophodon borsoni, referred second and third inferior grinders 208 

152. Zygolophodon borsoni zaddachi Jentzsch, type figure (molar) 208 

153. Zygolophodon borsoni, referred tliird superior molar from Le Petit-Rosey, near Lyons, France. After Lortet and Chantre. 208 

154. Zygolophodon borsoni Hays, type figiu-e (molar) 209 

155. Zygolophodon borsoni [ = PI ionmstodon praetypica reL], referred superior molar, from Neudorf, Austria. . After Vacek 210 

156. Zygolophodon borsoni, referred right inferior grinder, from Hidveg, Hungary. After Schlesinger 210 

157. Zygolophodon borsoni, referred anterior half of third inferior grinder (erroneously attributed to ?North America). After 

Schlesinger 210 

158. Zygolophodon borsoni and Mastodon americanus. Restorations by Flinsch 21 1 

159. Theoretic migration lines of the forest- and swamp-loving Zygolophodon and Turicius 215 

160. Petit mastodonte from Montabusard [Turicius tapiroides type]; molar described by Defay (1783), by Cuvier (1806) 217 

161. Petit mastodonte Cuvier, 1806 [Turicius tapiroides], new type figure. After Mayet 217 

162. Turicius turicensis, genotype and homcEotype molais fiom the Upper Miocene of Elgg, Switzerland 218 

163. Turicius turicensis simorrensis Osborn, type figure (molar) 220 

164. Turicius atticus Wagner, type figure (premolar and first molar) 220 

165. Turicius atticus, referred milk premolars, from Pikermi. Referred by Gaudry to Mastodon turicensis 221 

166. Turicius turicensis, referred molar from Murinsel, Croatia (cf. T. virgatidens). After Vacek 221 

167. Turicius virgatidens, referred molar from Terrasse vom Laaerberg, Austria, and Turicius{?) tapiroides, referred molar from 

Faluns de la Toiu'aine, France. After Schlesinger 221 

168. Turicius virgatidens von Meyer, type figure (molars) 222 

169. Long-jawed Trilophodon, the classic 'Mastodon' angustidens of Cuvier. After painting by Knight 224 

170. Phiomia and Serridentinus, comparison of crania 227 

171. Phiomia, Serridentinus, Trilophodon, and Tetralophodon, evolution of the second and third inferior molars 227 

172. Crania and incisive tasks in six bimomastodont subfamilies 229 

173. Dental succe.ssion: Bunomastodontidse and Elephantidae. After Frick, 1926 230 

174. Dental succession in Mastodon, Rhynchotherium, Serridentinus, and Trilophodon [ — Ocalientinus]. After Frick, 1926 232 

175. Dental succession in Rhynchotherium, Serridentinus, and Trilophodon [ — Ocalientinus]. After Frick, 1926 233 

176. Deciduous premolars in Serridentinus, Trilophodon [ = Ocalientinus], and Rhynchotherium. After Frick, 1926 234 

177. Deciduous and permanent premolars in the Serridentinse and Longirostrinae. After Frick, 1926 235 

178. Palates of the small Phiomia minor and of the large P. wintoni, illustrating differences of facial proportion 237 

179. Phiomia minor and P. wintoni, superior dentition 240 

180. Phiomia serridens Andrews and Beadnell, second type figure. After Andrews 240 

181. Phiomia wintoni (cf. serridens), juvenile dentition, compared with type lower jaw of P. serridens 241 

182. Phiomia minor ref., aged skull, and P. wintoni ref., juvenile jaw 242 

183. Phiomia minor, skull and task, and P. wintoni jaw, composition drawing 243 

184. Phiomia ivintoni erroneously combined with jaw of Palseomastodon beadnelli. Andrews' restoration of 1906 244 

185. Phiomia osborni Matsumoto, second type figure (mandible and dentition) 245 

185a. Five species of Phiomia, in ascending geologic order. Restorations by Flinsch 245 

186. Phiomia pygmseus Deperet, type figure (molar) 247 

187. Phiomia osborni type, P. pygmseus type, and Palxomastodon beadnelli ref., comparison of third inferior molars 247 

188. Lower jaws and comparative crania of the hyperlongirostral Trilophodonts, also of the Amebelodonts 248 



LIST OF ILLUSTRATIONS xxxi 

FIGURE PAGE 

189. Map showing geographic distribution of the types and referred specimens of Trilophodon and Amehelodon, also of the type 

of Phiomia pygviseus 250 

190. Mastodonte d dents etroites [ = Trilophodon angustidens], Cuvier's original type figure 252 

191. Petit mastodonte [ = Trilophodon angustidens minutus], Cuvier's original type figure 252 

192. Trilophodon, Tiiricius, and Deinotherium, deciduous incisive tusks. After Stehlin 253 

193. IMap of the Lower Miocene, Burdigalian, showing horizons of the Sables dc I'Orl^anais, etc. After Dep^ret (1905-1908) 

and Osborn (1910) 254 

194. Trilophodon angustidens from Simorre, referred second inferior molar. After Lydekker 255 

195. Trilophodon angustidens from Simorre, referred tiiird inferior molar. After (Jaudry 255 

196. Superior milk premolars of the left maxilla in the Trilophodonts and Tetralophodonts. After Gaudry 255 

197. Trilophodon angustidens ref. from Simorre, succession of milk and permanent teeth. After Gaudry 256 

198. Trilophodon angustidens gaillardi, type mandible. After pen sketch by Dr. Gaillard 257 

199. Trilophodon and {?)Serridenlinus, composition skeleton in Paris Museum; skeletal material discovered by Laurillard 258 

200. Elephas [ = Archidiskodon] meridionalis of Durfort, mounted skeleton, and Trilophodon angustidens, skull of mounted skeleton, 

in Paris Museimi. After pen sketches by Knight 259 

201. Trilophodon angustidens var. austro-germanicus Wegner, type figure (molars) 259 

202. Trilophodon angustidens var. libycus Fourtau, type figui-e (molars) and referred left lower jaw 260 

203. Trilophodon and Tetralophodon, milk dentition. After Sclilesinger 262 

204. Lower Pliocene (Etage Pontien) of western Eurasia; Middle Pliocene (fitage Plaisancien) in part, as far east as the 

Caspian Sea; Upper Pliocene (Etage Astien), a few important localities 263 

205. Trilophodon (Choerolophodon) pentelicus Gaudry and Lartet, type figure (juvenile skull and teeth) 264 

206. Trilophodon {Choerolophodon) pentelicus type (juvenile skull and teeth). After Gaudry 264 

207. Trilophodon (Choerolophodon) pentelicus, referr(>d superior milk dentition. Brown ( 'oUection 265 

208. Trilophodon (Choerolophodon) pentelicus, referred juvenile cranium and milk dentiton. After Schlesinger 265 

209. Trilophodon pandionis Falconer, type figiu-e (molar) 267 

210. Trilophodon pandionis, type molar. After Lydekker 267 

211. Trilophodon paheindicus Lydekker, type figure (third inferior molar) and asfsociated second superior molar 268 

212. Trilophodon pals-indicus, referred second inferior molar. After Lydekker 269 

213. Trilophodon cooperi, referred third inferior molars from the Bugti Hills, Baluchistan 270 

214. Trilophodon cooperi, referred molar fragment. After Lydekker 270 

215 Trilophodon cooperi, referred fourth right superior true premolar. After Lydekker 270 

216. Ilemimastodon crepusculi Pilgrim [ = Suina], type figure (molar) 271 

217. Trilophodon sp., fourth superior deciduous premolar, referred by Pilgrim to Hemimastodon crepusculi 272 

218. Trilophodon angustidens gaillardi, type, and Trilophodon chinjieiu^is, holotype (mandibles) 273 

219. Trilophodon macrognalhus Pilgrim, type molar 274 

220. Trilophodon paljeindirus, referred third right inferior molar 274 

221. JVLip of Bugti Hills, Lower Miocene-Upper Oligocene (?) horizons west of Dera Bugti, Baluchistan 275 

222. Evolution of the central conules in Trilophodon, showing T. cooperi type 276 

223. Trilophodon cooperi Osborn, paratype — 'older palate' 277 

224. Trilophodon inopinatus Borissiak and Beliaeva, type figures (cranium, mandible, tusks, and grinding teeth) 278 

225. Classic fossil mammahan exposures of northwestern India 279 

226. Trilophodon sendaicus Matsumoto, type figure (molar) 280 

227. Trilophodon esselbornensis Kliihn, type figure (molar) 281 

228. Trilophodon steinheimensis Kliihn, type figures (molar). After O. Fraas, 1870, and Klalm, 1922 282 

229. Turicius wahlheimensis Klahn, cotype figures (molars) 282 

230. Trilophodon pontileviensis Mayet-Fourtau, cotype figures (molars). Also type molar of Mioinasiodon depereti sp. nov. 

After Mayet 284 

231. Trilophodon simplicidens Osborn, type molar 285 

232. Trilophodon obscurus Leidy, type figure (molar) 286 

233. Trilophodon obscurus, new type figure 286 

234. Trilophodonts, primitive and progressive, of Eurasia and North America. Restorations by Flinscli 287 

235. Trilophodon (Genomastodon) osborni, conjectural restoration. After Barbour 290 

236. Trilophodon dinotherioides Andrews, type figure (mandible, third molar) 291 

237. Trilophodon (Genomastodon) willistoni Barbour, juvenile type skull and jaws 292 

238. Trilophodon (Genomastodon) willistoni Barbour, type lower jaw 292 

239. Trilophodon (Genomastodon) willistoni Barbour, type lower jaw. .\fter sketch by Barbour 292 



xxxii OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

240. Trilophodon (Genomastodon) ivilKstoni Barbour, type skull and jaws. After pencil sketch by Osborn 292 

241. Trilophodon {Genomastodon) willistoni, referred humerus and five tusks showing varying widths of enamel bands 293 

242. Lower to Middle Pliocene exposiues of Devil's Gulch on south side of Niobrara River, Brown Co., Nebraska 294 

243. Trilophodon (Megabelodon) lulli Barbour, type skull and jaws 295 

244. Trilophodon (Megabelodon) hilli Barbour, type figure (five views of mandible) 296 

245. Trilophodon (Megabelodon) lulli Barbour, type mandible in comparison with mandibles and rostra referable to either T. 

joraki or T. (Meg.) lulli 297 

246. Trilophodon ligoniferus Cope and Matthew, type figure (mandible) 298 

247. Trilophodon (Genomastodon) osborni Barbour, type mandible 299 

248. Trilophodon (Genomastodon) osborni Barbour, type skull and skeleton 300 

249. Trilophodon (Genomastodon) osborni Barboiu-, type superior and inferior grinding teeth, also mandible. After sketches by 

Osborn 301 

250. Trilophodon (Genomastodon) osborni Barbour, type skull and mandible (skull in plaster 'cinches'), also conjectural restoration 

of head. After Barbour 301 

251. Trilophodon (Genomastodon) osborni Barbour, type "fifth and sixth" superior molars 301 

252. Trilophodon (Genomastodon) osborni Barbour, type "fifth and sixth " inferior molars 301 

253. Trilophodon (Genomastodon) osborni, type skull, also referred skeletal parts compared with those of Elephas indicus. After 

Barbour 302 

254. Trilophodon (Genomastodon) osborni, left superior tusk found in pit near skull. After Barboiu 303 

255. Trilophodon (Genomastodon) osborni and Elephas indicus forelimbs compared. After Barbour 303 

256. Trilophodon (Genomastodon) osborni Barbour, type skull and mandible 304 

257. Trilophodon giganteus Osborn, type mounted skeleton 305 

258. Trilophodon giganteus Osborn, type figure (mandible) compared with mandibles of other Longirostrines 305 

259. Trilophodon giganteus Osborn, type skull and jaws, also inferior dentition 306 

260. Amebelodon (Trilophodon) hicksi Cook, type figure (molar) 307 

261. A7nebelodon (Trilophodon) hicksi Cook, type figure (mandible) 308 

262. Amebelodon (Trilophodon) /izcte', referred superior molar associated with type, presumably of same individual. After Cook. 309 

263. Amebelodon (Trilophodon) paladentatus Cook, type figure (mandible) 310 

264. Amebelodo7i (Trilophodon) paladentatus Cook, type figure (incomplete inferior molar) 310 

265. Parts of two superior tusks found with the types of Amebelodon (Trilophodon) hicksi and A. (T.) paladentatus. After Cook . . 310 

266. Trilophodon abeli Barbour, type figure (mandible) 311 

267. Trilophodon fricki Peterson, type cranium, mandible, and dentition. After original drawings 313 

268. Trilophodon fricki Peterson, restoration. After pencil drawing by Dr. A. Avinoff 314 

269. Trilophodon phippsi Cook, type mounted skeleton in Colorado Museum of Natural History. After photograph 315 

270. Trilophodon phippsi Cook, type skull and jaws. After photographs 316 

271. Trilophodon phippsi, referred cranium and jaws 316 

272. Mio-Pliocene beds of Brown County, Nebraska, generalized columnar section 317 

273. Key to intermingling of the Longirostrinae, Serridentinffi, Platybelodontinae, and Humboldtinae along the ancient Niobrara 

River 317 

274. Lower to Middle Pliocene fossil localities in north central Nebraska 318 

275. Trilophodon pojoaquensis Frick, type figure (skull) 320 

276. Trilophodon pojoaquensis Frick, type molar, and Ocalientinus ojocaliensis Frick, referred molar 321 

277. World migration lines of the typical Trilophodon (T. angustidens), of the 'shovel-tuskers' (Phiomia, Amebelodon), and of the 

'prod-tuskers' (Trilophodon), from Africa (theoretical), Eurasia, India, and North America 321 

278. Ocalientinus ojocaliensis, referred mandible and inferior molar, and Trilophodon pojoaquensis, largest cranium, also superior 

molar. After Frick (original drawings) 322 

279. Trilophodon cruziensis Frick, type cranium and mandible, also Ocalientinus ojocaliensis referred inferior molar. After Frick 

(original drawings) 323 

280. 'Prod-tuskers': Trilophodon (Tatabelodon) riograndensis Frick, genotype mandible; T. (Tatabelodon) gregorii Frick, type 

cranium and mandible. Also palate referred by Frick to Tatabelodon gregorii, by Osborn to Trilophodon osborni. After 

Frick (original drawings) 325 

281. Trilophodon, (7) Megabelodon, Ocalientinus, Serridentinus,a,ndTrobelodon. Prick's mandibular types. Original drawings ... 327 

282. 'Oblique-tuskers' (Trilophodon) and 'uproot-tuskcrs' (Serridentinus, Ocalientinus), comparison of mandibles 328 

283. Resemblances and contrasts between the ancestral Oligocene Phiomia, Mio-Pliocene ' uproot-tusker' Trilophodon, Plio- 

cene 'shovel-tusker' Amebelodon, and tuskless 'spoon-bill' Trilophodon (Megabelodon) lulli 330 

284. Amebelodon fricki, 'shovel- tusker ' of Nebraska. Restoration by Flinsch 332 

285. Platybelodon grangeri, 'flat-tusker' of Mongolia. Restoration by Flinsch 332 



LIST OF ILLUSTRATIONS xxxiii 

FIGUHE PAGE 

286. Contrasts between the 'flat-tasker' Platybeludon, the 'shovcl-luskfr' Amebeltxlon, the 'uproot-tiiskers' Trilophodon chin- 

jiemiii and Si-rrideiitinus gnbieiixis — of MioPUocene age, and tlie pr.iljable ancestral <)U(!;ocene I'luomiii oaboriii 334 

287. Ainebelodon fricki Harbour, tyjie figure (mandible) 335 

288. A mebelodon fricki Harb(iur, type molar 336 

289. Ainebelodon fricki, referred inferior mohir. After Harljour 336 

290. Ainebelodon fricki Barbour, second tyjie figure of mandible 336 

29L Amebelodon fricki, bntsli drawing of type mandibular tusk.s. After Barbour 336 

292. Amebelodon fricki, conjectural restoration, .\fter Harbour 337 

293. A mebelodon fricki, rcNi.sed restoration of head, .\fler Harbour 337 

294. Ainebelodon .sinclairi Harbour, lyjie figure (mandibular tusk) 337 

295. Amebelodon, imaginative sketch. After Barbour ;i38 

296. Turynobclodon loomisi Harbour, ty|)e figure (mandibular tusk) 339 

297. Torynobelodon looini.si, conjectural restoration of mandible. After Barbour 339 

298. Torynobelodon loomisi, conjectural restoration in supposed marshy habitat. Aitt-r Barbour 339 

299. 'Mastodon' angustidens Cuvier, type. After one of the original castjs in Mantell Collection, presented by the British Mu- 

seum to the .\merican Museum 340 

."iOO. 1 4,()()(),(ltl(l-y('ar cvolulion of the Trilojihodonts, from the Trilophodon cooperi of Baluchistan to the T. inacrognathus Mio-Plio- 

ecnc stage of ( 'iiinji, India 341 

301. Tetralophodon punjabieiisis and T. campesler. Restorations by Flinsch 342 

302. Telralophodonts of Kurjusia and North .\mcrica. Restorations by Flinsch 342 

303. Tetralophodon and Trilophodon, superior intermediate molars. After Falconer 344 

304. Tetralophodon, Choerolophodon, and Trilophodon, superior milk premolars of left maxilla 346 

305. Synconiilophiis ilhokpalhanensis Osborn, tyi)e craniiun. .\fter ])hotograph 347 

306. Tetralophodon piinjiihiensis, referred cranium, .\fter photograi)h 347 

307. Map showing geographic distribution of the types and referred specimens of the Tetralophodontinae 349 

308. .Anteroposterior evolution and hypsodonty of third superior and inferior molars in Tetralophodon, from base to summit of 

Pliocene 350 

309. Tetralophodon (Lydekkeria) falconeri Lydekker, type figures (jaw and molar) 353 

310. Tetralophodon (Lydekkeria) (?) falconeri, referred molars. After Falconer and Cautley 354 

311. Tetralophodon (Lydekkeria) sinensis Koken, type figure (molar) 355 

312. Tetralophodon longirostris Kau|), ty|)e figure (man<libl(') 357 

313. Tetralophodon longiroslris. Kaup's composition of referred skull and type jaws. After Owen, figured as 'Mastodon' 

angustidens 357 

314. Mastodon grandis Kaup antl SchoU [ = Tetralophodon longiroslris], type molar 358 

315. Tetralophodon longirostris, referrcti palate anil superior dentition. After Kaup 358 

316. Kaup's type jaw of Tetralophodon longiroslris compared with Phiomia, SerrideiUinus, and Trilophodon 359 

317. Tetralophodon grandinrisivus Sehlesinger, type fig\n'e (lower tiLsk) 360 

318. Tetralophodon grandincisirus Sehlesinger, superior molar (paratype, Osborn) 361 

319. Tetralophodon longiroslris, referred skull of young individual. After Sehlesinger 361 

320. Tetralophodon (Trilophodon?) grandincisii'us, reconstruction of referred skull and jaws. After Sehlesinger 362 

321. Tetralophodon grandincisivus (?), referred sui)erior and inferior grinders. /Vfter Sciilesinger 362 

322. Tetralophodon punjabiensis Lj'dekker, one of the cotypes (Osborn's lectotype). .\ftcr Lydekker 363 

323. Tetralophodon punjabieiisis Lydekker, new lectotype figure by Osborn 363 

324. Tetralophodon punjabieiisis, referred cranium in Brown Collection 364 

325. Tetralophodon campesler type and T. piinjabicnsis ref . (molars) . ( "ontrasts and resemblances 365 

326. Tetralophodon bumiajuensis van der Maarel, type figures (cranium, jaw, and grinders) 366 

327. Map of central Java, showing type horizon of Tetralophodon bumiajuensis 366 

328. Scrrideiitinus and Tetralophodon grinding tooth o|)|)osition 367 

329. Kxtreme Pliocene evolution of Tetralophodon molars 368 

330. Tetralophodon campesler Cope, type. Second figure, after Cope and Matthew 370 

331. Tetralophodon campester Co\h\ ty]ie skull and jaws, compared with restored skull and jaws of T. longiroslris Kaup 371 

332. Tetralophodon elegan.s Hay, type figure (molar), also refern-d (?) Trilophoilon molar from near Waco, Texas. Aft^-r Hay 373 

333. Trilophoilon ('!Tetralophodon) brazosiiis Hay, type figure (part of right ranuis of lower jaw) . 374 

334. Tetralophodon fricki sp. nov., type palate and dentiton. After Frick (original drawings) 375 

335. Tetralophodon fricki and PUitybelodon grangeri crania com])ared 376 



xxxiv OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

336. Tetralophodon campester Cope, type palate 378 

337. Tetralophodon {Morrillia) barbouri Osborn, type superior molar 378 

338. Tetralophodon punjabiensis and T. longirostris, type molars compared 378 

339. Tetralophodon precampester Osborn [ = Tetralophodon (Morrillia) barbouri ref .], type superior molar 378 

339a. World migration of Tetralophodon. Restorations by Flinsch 379 

340. Ocalientinus {Serridentinus) .floridanus Leidy. Restoration by Flinsch 380 

341 . Serridentinus productus Cope. Restoration by Flinsch 380 

342. Serridentinus serridens Cope, type figure (molar) 382 

343. Serridentinus progressus Osborn, type molar 382 

344. Map showing geographic distribution of the types and referred specimens of the Serridentinse and Platybelodontinae 384 

345. Proboscidean locahties of the eastern coast of the United States (map) 386 

346. Serridentines of the Florida Pliocene and contemporary Hipparion, Pliomastodon, rhinoceroses, and camels. After 

Matthew (map) 386 

347. Contrast between hyperlongirostral species of Trilophodon and medilongirostral species of Serridentinus and Ocalientinus. 

Restorations by Flinsch 388 

348. Serridentinus and Tetralophodon grinding teeth superposed 389 

349. Contrast of adult superior and inferior molars in Tetralophodon, Morrillia, and Serridentinus 390 

350. Serridentinus, Ocalientinus, and Trilophodon inferior grinders. Comparative series 392 

351. Serridentinus subtapiroideus Schlesinger, cotype molars, compared with Trilophodon angustidens referred. After Schlesinger 395 

352. Serridentinus subtapiroideus Schlesinger, cotype molars 395 

353. Serridentinus subtapiroideus Schlesinger, cotype molar 396 

354. Serridentinus mongoliensis Osborn, type figure (inferior dentition) 396 

355. Ocalientinus {Serridentinus) florescens Osborn, type figure (molar) 397 

356. Map of Tung Giu- tableland, Inner Mongolia 398 

357. Serridentinus gobiensis Osborn and Granger, type figure 398 

358. Pliocene and Lower Pleistocene Proboscidea, Perissodactyla, Artiodactyla, and Edentata of the Staked Plains of Texas 

and San Pedro Beds of Arizona (map) 399 

359. Mio-Pliocene Tertiaries of Florida (map) . After Simpson 400 

360. Serridentinus progressus Osborn, type figure (left ramus of lower jaw) 401 

361. Serridentinus progressus Osborn, new figure of type jaw, compared with referred skull and dentition of S. productus; also 

diagrammatic figure of type inferior third molar of both species 402 

362. Serridentinus proavus Cope, type figure (premolar) 403 

363. Serridentinus proavus Cope, new figure of type by Osborn 403 

364. Serridentinus productus and S. serridens, palatal view of crania 404 

365. Serridentinus productus, model by Flinsch 405 

366. Serridentinus serridens, referred skidl from Clarendon of Texas 405 

367. Serridentinus productus Cope, type figure (mandible) 406 

368. Serridentinus productus juvenUe and adult (referred and neotype) jaws compared with referred adult jaw of Ocalientinus 

ojocaliensis. After Frick 406 

369. Serridentinus productus, referred skull, compared with S. progressus, lower jaws and teeth; also enlarged inferior molar of S. 

productus ref 407 

370. Serridentinus productus, referred skull and jaws from Clarendon beds 409 

371. Serridentinus productus, adult skuU and restored skeleton in American Museum 410 

372. Trilophodon giganteus type and Serridentinus productus ref., restored skeletons in American Museum; also outline restora- 

tions by Flinsch 412 

372a. Primitive Serridentines of Eiu-asia and North America. Restorations by Flinsch 413 

373. World migration and adaptive radiation of the Serridentine Mastodonts (map) 414 

374. Ocalientinus {Serridentinus) republicanus Osborn, type and paratype figures (mandibles) 414 

375. Ocalientinus {Serridentinus) republicanus Osborn, type, paratype, and referred dentition. After Osborn 415 

376. Ocalientinus {Serridentinus) bifoliatus Osborn, type molar and paratype jaw 415 

377. Ocalientinus {Serride7itinus) floridanus Leidy, type and one of the metatype molars. After Leidy and Lucas 416 

378. Ocalientinus (Serridentinus) floridanus Leidy, type and one of the metatype molars 416 

379. Ocalientinus (Serridentinus) floridanus Leidy, type and metatype molars 417 

380. Ocalientinus (Serridentinus) floridanus Leidy, second type and metatype figures (molars) 419 

381. Ocalientinus (Serrideritinus) floridanus leidii Frick, new figure of type molars 419 

382. Ocalientinus (Serridentinus) obliquidens Osborn, type, paratype, and referred molars 420 



LIST OF ILLUSTRATIONS xxxv 

FIGURE PAGE 

383. Ocalienlinus (Serridenlinus) obliquidens Osborn, type and paratype molars compared with referred molar of Mastodon ameri- 

canus 421 

384. Ocalienlinus {Serridenlinus) obliquidens, referred molar, from Tarboro, North Carolina. After Leidy 421 

385. Serridenlinus and Ocalienlinus, comparison of inferior grinding teeth 421 

386. Map showing exposures of the Clarendon horizon of the 'Staked Plains' of Texas 422 

387. Serridenlinus serridens Cope, referred young adult .skull and jaws, Clarendon of Texas , 425 

388. Serridenlinus serridens Cope, type figure (molar) 423 

389. Serridenlinus serridens Cope, new type figure 423 

390. Serridenlinus serridens, referred dentition of young adult 424 

391. Serridenlinus anguirivalis Osborn, type figure (tyjx' and i)aratype molars and referred superior tusk fragments) 425 

392. Sheep Creek to Snake Creek exposures, Sioux County, Nebraska 426 

393. Sections through Sheep Creek and Snake Creek quarries, Sioux County, Nebrjiska 426 

394. Fauna (Mastodonts, Serridentines, and Rhynchorostrines), Snake Creek B horizon, western Nebraska 428 

395. Serridenlinus serridens cimarronis Cope, type figure (molar) 429 

396. Serridenlinus hrewslerensis Osborn, type figure (molar) 430 

397. Serbelodon{?) precursor Cope, type figure (molar) 431 

398. SerbelodonC!) precursor Cope, new type figiu-e 431 

399. Three teeth of Serbelodon{l) prxcursor, referred by Cope to Dibelodon tropicus 431 

400. Serridenlinus gualemalensis Osborn, type figure (molar) 432 

401. Serridenlinus gualemalensis Osborn, after photograph of type molar 432 

402. Ocalienlinus ojocaliensis Frick, type and referred specimens. After Frick (original drawings) 434 

403. Ocalienlinus, progressive evolution of the pentalophid 435 

404. Ocalienlinus ojocaliensis, referred third superior and inferior molar.s 437 

405. Ocalienlinus ojocaliensis, referred skull and jaws first described by Frick as paratjrptes of Trilophodon pojoaquensis 438 

406. Serbelodon and Trobelodon, blunt 'shovel-tuskers.' Restorations by Flinsch 441 

407. Adaptive radiation of three types of Serridentine crania, jaws, tusks, and grinding teeth 441 

408. Serbelodon barbourensis and Trobelodon taoensis, genotype and referred specimens 442 

409. Serbelodon and Platybelodon shovel-tusks compared 445 

410. Serbelodon, Platybelodon, Torynobelodon, and Amebelodon tusks superposed 445 

411. Trobelodon taoensis Frick, type figure (cranium, mandible, molars), also referred cranium. After Frick (original drawings). . . 446 

412. Serridenlinus barslonis Frick, type figure (premolars) 447 

413. India, proboscidean life zones and geologic horizons. Pilgrim (1910-1927), Pilgrim-Osborn (1927-1932), Osborn-Colbert 

(1935) 448 

414. Proboscidean life zones (Osborn, 1935) within the geologic horizons of India (Pilgrim, 1905-1932) 449 

415. Molar evolution of the Serritientina; in India, compared with Synconolophus and Tetralophodon molars 450 

416. Serridenlinus browni Osborn, type mandible, maxilla, and tusk 453 

417. Trilophodon hasnotensis sp. nov., type molar 454 

418. Serridenlinus broivni and /^I. metachinjiensis type molars compared 454 

419. Serridenlinus chinjiensis Osborn, type molar 456 

420. Serridenlin us prochinjien^is Osborn, type molar 456 

421. Serridenlinus annectens Matsumoto, type figure (molars) 458 

422. Serridenlinus lydekkeri Schlosser, type figure (molar) 458 

423. Anu'belodont sho\el-tusker.s and Platybelodont flat-tuskers of Mongolia, North (raucasus, Nebraska, and Colorado. Res- 

torations by Flinsch 460 

424. American (Andrews, 1922-1929) and Russian (Khomenko, 1912, Borissiak, 1929) explorations in eastern and western Asia 

(map) 461 

425. Platybelodon danovi Borissiak, type figure (cranium, mandible, and molar) 461 

426. Platybelodon danovi Borissiak, type mandible 462 

427. Platybelodon danovi Borissiak, type third inferior molar 462 

428. Platybelodon, four restorations of head, mouth-parts, etc. After Borissiak 463 

429. Platybelodon qrnngeri Osborn. type mandible. .\ft^r Osborn and Cranger 464 

430. Dentinal rod-cones of Platybelodon and hollow incisors of Phiomia. After Osborn 465 

431. Sections of incisors of Platybelodon and Phiomia 465 

432. Map of region around Iren Oabasu, Inner Mongolia, showing Tung Cur tableland 466 

433. Deciduous succe.ssion in Platybelodon, as displayed in mandible and dentition of Platybelodon grangeri ref. After Osborn .... 466 

434. Platybelodon grangeri ref., superior dentition. After Osborn 467 



xxxvi OSBORN: THE PROBOSCIDEA 

FIGURE ^A^J^ 

435. Plahjbelodon grangcri ref., juvenile fragmentary cranium and nearly complete mandible. After Osborn 467 

436. Platybelodon grangeri, referred cranium and jaws (female), also male superior tusk. After Osborn 467 

437. Platyhelodon grangeri, referred adult mandible and dentition. After Osborn 468 

438. Platybelodon grangeri ref., anterior portion of mandible, showing inferior aspect of lower incisors, placed beside coal shovel 468 

439. Platybelodon grangeri, referred third superior grinder 470 

440. Torynobelodon barnumbrowni Barbour, type figure (mandible) 470 

441. Platybelodon grangeri, referred second and third inferior grinders 471 

442. Torynobelodon barmimbrowni Barbour, type inferior molars. After pencil sketch by Barbour 471 

443. Torynobelodon barnumbrowni Barbour, type figure (mandible) 471 

444. Platybelodon grangeri, referred mandible 472 

445. Torynobelodon barnumbrowni Barbour, tyj^e mandible 472 

446. Serridentinus Jiebrascensis Osborn, type figure (molar) 473 

447. Rhynchotherium tlascalx, male and female. Restoration by Flinsch 474 

448. P'alconer's genotype cast of Rhynchotherium [tlancalu] and Osborn's type of R. browni sp. nov., both from Mexico (mandi- 

bles) 474 

449. Rhynchotherium bi-evidens Cope, type and referred third superior molars 476 

450. Rhynchorostrine types of lower jaw. After Osborn 476 

451. Map showing geographic distribution of type and referred specimens of the Rhynchorostrinse 480 

452. Rhynchotherium browni sp. nov., type mandible and dentition (formerly Osborn's neotype of R. tlascalx) 481 

453. Type jaws of the Rhynchorostrinse, compared with Trilophodon dinotherioides 484 

454. Rhynchotherium molars from Contra Costa County, California, and Tambla, Honduras 484 

455. Rhynchotherium spenceri Foiu-tau, type figvire (mandible, second molar) 485 

456. Rhynchotherium brevidens Cope, type figure (molar) • 486 

457. Rhynchotherium brevidens Cope, new type figure 486 

458. Rhynchotherium brevidens, referred molar from state of Washington 486 

459. Rhynchotherium shepardi Leidy, right and left superior type tusks reconstructed 487 

460. Rhynchotherium shepardi Leidy, original type figure (tusk fragment) 487 

461. Rhynchotherium rectidens Osborn, type right and left uj^per tusks 488 

462. Rhynchotherium [Blickotherium] euhypodon Cope, new type figure 490 

463. Rhynchotherium [Blickotherium] euhypodon Cope, type figure (mandible) 491 

464. Rhynchotherium [Blickotherium] euhypodon Cope, type figure (palate with superior molars and incisors) 491 

465. Rhynchotherium anguirivale Osborn, type figure (molar) 492 

466. Rhynchotherium anguirivale, type and referred grinding teeth 492 

467. Rhynchotherium browni sp. nov., type jaw and dentition 494 

468. Rhynchotherium falconeri Osborn, type jaw and dentition 495 

469. Rhynchotherium molar from Oak Springs, Contra Costa County, California, referred by Leidy to 'Mastodon' obscurus 495 

470. Rhynchotherium shepardi edense Frick, type figures (portion of skull, posterior maxillaries, and grinding teeth) 497 

471. Cordillerion edensis Osborn, type figure (premaxillse and tusks). After Frick 497 

472. Map showing position of San Timoteo and Eden deposits in relation to other Pliocene fossil mammal horizons. After Frick 498 

473. Rhynchotherium shepardi edense, referred superior and inferior grinders 499 

474. Rhynchotherium shepardi edense, referred immature jaw and maxilla, formerly type of R. paredensis Osborn 500 

475. Rhynchotherium francisi Hay, type figure (molar) 502 

476. Rhynchotherium chinjiense Osborn, type figure (left ramus, third molar) 502 

477. Rhynchorostrines, comparative restorations by Flinsch 503 

478. Rhynchorostrine mandibles and tusks. North Africa to Honduras 504 

479. Blickntherium blicki Frick, genotype mandible and third molar, and Rhynchotherium edense, neotype immature cranium and 

mandible. After Frick (original drawings) 506 

480. Aybelodon hondurensis Frick, genotype mandible and third inferior molar. After Frick (original drawings) 509 

481. Rhynchotherium types, as mounted in the American Museum 511 

482. Rhynchorostrinse, triphyletic evolution (mandibular series) 512 

483. Rhynchorostrinse, theoretic migration lines (maj)) 513 

484. Cordillerion andium, mastodont of the Cordilleras. Restoration by Flinsch 514 

485. Cuvieronius superbus, giant mastodont of the Argentine plains. Restoration by Flinsch 514 

486. Mastodonte des Cordilieres {'Mastodon' andium), Cuvier's original type figure 517 

487. Mastodonte humboldien {'Mastodon' humboldtii), Cuvier's original type figure 517 



LIST OF ILLUSTRATIONS xxxvii 

FIG U UK PAGE 

488. 'Maiitudon' humbokUii [ = Cuvieroniuii iDplateiinis] and Cordillerion andium, referred third superior molars. After Gervais. 517 

489. C uvieronius plaiensis Ameghino, tyi)e figure (tusk) 520 

490. Cuvieronius rectus Amoghino, type figure (tu.sk) 520 

491. Notiotnastodon argentinus Ameghino, type figure (tusk) 520 

492. Cuvieronius superbus Ameghino, type figure (tusk) 520 

493. Anipghino's scheme of the pliyletic relations of species of South .\merican proboscideans 521 

494. Trtrabdodon [Tn'lophodon\ (ingiistidena and Dibdodon [Cordilleridit] andium, crania. After Lull 526 

495. Midcranial sections of 'M.' andium, 'M.' humboldtii, and E. indicus. After Boule 528 

496. Cordillerion andium, single crests and trefoils and enamel foldings of molars. After Boule 528 

497. Cuvieronius humboldtii, cranium in Paris MiLsemn. After Boule 529 

498. Restorations of Notorostrines and Humboldtines, as known up to 1933, by FUnsch 531 

499. Phylogeny of the Mastodontidffi. After Boule 532 

500. Cordillerion edensis Frick. Restoration by Flinsch 535 

501. Map showing palaogeographic range of Proboscidea in western United States, Central and South America 536 

.502. Cordillerion andium in comparison with Cuvieronius superbus. Restoration by Flinsch 542 

503. Cordillerion andium, the Andean mastodont of the Cordilleras. Restoration by Flinsch 543 

504. Cordillerion and Cuvieronius grinding teeth. After Gervais 544 

505. Cordillerion andium, referred cranium, after Burmeister 545 

506. Cordillerion andium, referred cranium in Field Museum of Natural History. After pen sketch by Knight 546 

507. Cordillerion andium, Cuvieronius humboldtii, and Eubelodon morrilli crania 547 

508. Mastodonte des Cordilieres (Cordillerion andium), Cuvier's original type figure 549 

509. Cordillerion andium, referred jaw, after D'Orbigny 550 

510. Notiomastodon argentinus Ameghino, type figure (tusk), first described as ' Mastodon' argenlinus 550 

511. CordillerionC?) bolivianus, referred molar [genotype of Tcleobunomastodon Revilliod] 551 

512. Cordillerion bolivianus, referred mandible. After Pompcckj 552 

513. Cordillerion tropicus Cope, type figure (right ramus of jaw). After von Meyer's 'Mastodon humboldtV ref 553 

514. Cordillerion oligobunis Cojie, type figure (mandible). After Villada 555 

515. Cordillerion oligobunis antiguissimus P'reudenberg, type figure (molar). After original iihotdgrapji 556 

516. Cordillerion{?) oligobunis felicis Freudenberg, type figure (superior molar) 557 

517. Cordillerioni?) oligobunis felicis, referred inferior molar, after photograph 557 

518. Cordillerion{'!) oligobunis intermedius Freudenberg, type figure (molar), after photograph 558 

519. Cordillerion oligobunis progressus Freudenberg, type figure (molar), after photograph 558 

520. Cordillerion gratum Hay, type figure (maxilla, second and third molars) 559 

521. Map showing position of San Tinioteo and Eden deposits in relation to other Pliocene fossil mammal horizons. 

After Frick 560 

522. Cordillerion ederisis Osborn, type figure (premaxillae and tusks). After Frick 561 

523. Cordillerion edensis tyjie rostrum and sujierior tusks compared with those of C. andium. After Osborn 561 

524. Cordillerion orarius Hay, type figure (molars) 563 

525. Cordillerion defloccatus Hay, type figure (mandible) 564 

526. Cordillerion bensonensis Gidley, type figure (fragmentary cranium) 566 

527. Localities of mastodont discoveries in I'^cuador, Canyon of Chalang, near Punin, Province of Chimborazo 567 

528. Cordillerion andium, referred cranium from Tarija, in Field Museum of Natural History 568 

529. Mastodonte des Cordilieres {'Mastodon ' andium), Cuvier's original type figure 568 

530. Cuvieronius postremus Spillmanii ('Mastodon of .\langasi'), type cranium, after retouched photographs 568 

531. Cuvieronius grinding teeth and tusks, type and referred, .•\fter Carette 569 

532. Cordillerion and Cuvieronius crania and jaws. After Nordenskiold and Carette, also Osborn original 570 

533. Cuvieronius postremus Spillmann, skeleton 572 

534. Cuvieronius ayora', referred first left superior molar 573 

535. Cuvieronius ayora', referred tiiird right inferior molar, also mandible, Frick Collection 573 

536. Cuvieronius postremus, supposed flint spear holes in occiput. After Spillmann 574 

537. Mastodonte humbnldien (' Mnstndon' [Cuvieronius] humboldtii), Cuvier's original type figure 577 

538. Mastodon [Cuvieronius] humboldtii, Osborn 's paratype, after Cuvier 577 

539. Cuvieronius humboldtii, referred third superior molar. After Lydekker 577 

540. Cuvieronius and Cordillerion, third right su|)pri()r molars. After Gervais 578 

541. Cuvieronius superbus, mandible referred b\- Burmeister to Mastodon humboldtii - 578 



xxxviii OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

542. Cuvieronius bonaerensis (superbus) Moreno, type cranium. After Cabrera 579 

543. Cuvieronius platensis Ameghino, type figure (tusk) 579 

544. Cuvieronius superbus Ameghino, type figure (tusk) 580 

545. Cuvieronius rectus Ameghino, type figure (tusk) 580 

546. Cuvieronius rectus, primiti\'e referred molar. After Ameghino 581 

547. Cuvieronius chilensis Phihppi and Cordillerion bolivianus Phihppi, type figures (mandibles), in comparison with Cordillerion 

andium 582 

548. Cuvieronius ayorse Spillmann, type figure (cranium), after photograph 583 

549. Cuvieronius ayorse and Cordillerion andium, cranial sections 583 

550. Cuvieronius postremus Sijillmaiui, type mandible. After photograph 584 

551. Cuvieronius postremus Spillmann, type cranium. After photograph 585 

552. Cuvieronius postremus, referred molars. After photograph 585 

553. Cuvieroyiius and Stego7nastodon, parallelisms and divergences in inferior molars and mandibles 586 

554. Cuvieronius platensis, referred second and third inferior molars 587 

555. Cordillerion andium and Notiomastodon argentinus, second cervical vertebrae (axis). After Cabrera 587 

556. Cuvieronius superbus referred, tip of tusk artificially carved to convert it into a utensil. After Castellanos 589 

557. Notiomastodon ornatus Cabrera, type and referred tusks, type jaw, also type tusk of A'^. argentinus. After Cabrera 591 

558. Cuvieronius platensis Ameghino and C. rectus Ameghino, type and referred superior tusks. After Cabrera 592 

559. Cuvieronius humboldtii Cuvicr, paratype third inferior molar, after Cuvier, and C. platensis, referred third superior molar, 

after Cabrera 592 

560. Cuvieronius platensis, referred third superior and inferior molars. After Cabrera 594 

561. Cuvieronius superbus, referred third superior and inferior molars. After Cabrera 594 

562. Cuvieronius bonaerensis Moreno, type cranium (cf. C. superbus Ameghino), after Cabrera, also crania and mandibles referred 

by Cabrera to Stegomastodon [ = Cuvieronius] platensis 596 

563. Cuvieronius superbus type tusk and mandible; also tusks, cranium, and mandible referred by Cabrera to Stegomastodon 

[Cuv.] superbus, by Osborn (in part) to C. platensis 597 

564. Skeletal remains of Cuvieronius platensis ref. and of C. bonaerensis (superbus) Moreno, type. After Cabrera 598 

565. Cuvieronius superbus, referred skeletal parts. After Cabrera 599 

566. Cuvieronius superbus and C. platensis. Restorations by Flinsch 599 

567. Eubelodon morrilli Barbour, type skull and jaws (after reconstruction) also referred vertebral series. After Barbour 600 

568. Devil's Gulch section, Brown County, Nebraska, where types of Trilophodon willistoni and Eubelodon morrilli were found. . 601 

569. Eubelodon morrilli Barbour, type jaws and palate before reconstruction. After photograph 603 

570. Eubelodon morrilU Barbour, type jaw. After sketch by Barboiu* 603 

571. Eubelodon morrilli Barbour, crown pattern of third superior and inferior type molars. After pencil sketch by Osborn 603 

572. Eubelodon morrilli, type pelvis, also referred limb skeleton. After photographs 604 

573. Eubelodon morrilli skeleton as mounted in the Nebraska State Museum, 1931. Background by Elizabeth Dolan 605 

574. Eubelodon morrilli, referred cranium and superior dentition, found by Frick expedition. After Frick (original drawings) .... 606 

575. Eubelodon morrilli, type and referred molars and referred tusk 607 

576. Eubelodon morrilli and Cuvieronius bonaerensis (superbus) crania and jaws compared 608 

577. Eubelodon morrilli, the Humboldtine mastodont of western Nebraska. Restoration by Flinsch 610 

578. Humboldtinae : Proversion of inferior ridge-crests in Stegomastodon 612 

579. Humboldtinae: Proversion of inferior ridge-crests in Cuvieronius 612 

580. Humboldtinae: Retroversion and controversion of superior ridge-crests in Cuvieronius and Stegomastodon 613 

581. Brevirostrinse : Proversion of ridge-crests in Anancus and Synconolophus 613 

581a. Notiomastodon ornatus, type first and second inferior molars. New figure 614 

582. Anancus arvernensis, the straight-tusked mastodont of Auvergne, southern France, and of northern Italy. Restoration by 

Flinsch 616 

583. Synconolophus dhokpathanensis. Middle Pliocene mastodont of Dhok Pathan, northern India. Restoration by Flinsch 616 

584. Anancus arvernensis, mounted skeleton in Turin Museum. After Sismonda 618 

585. Anancus arvernensis, mounted skeleton in Bologna Museum. After photographic reproduction of figiu-e in Capellini's 

monograph of 1908 619 

586. WilUam Smith's ' Whitlingham Tooth' [ = Anancus falconeri Osborn] 620 

587. Anancus falconeri Osborn, fluted deciduous superior premolar from Norwich Crag. After Lydekker 620 

588. Primitive Anancus molars of Europe and of India. After Schlesinger 621 

589. Anancus perimensis from Perim Island, India, referred second and third inferior molars. After Falconer and Cautley 621 

590. Map showing geographic distribution of eighteen of the types of Brevirostrinse of Eurasia and of seven of the types of 

Himiboldtinae of North America, also referred specimens 624 



LIST OF ILLUSTRATIONS xxxix 

FIGURE PAGE 

591. Crania of Brcvirostiinai from Italy and India, and Slegomastodon (Hiunboldtinae) from the United States (Texas) 626 

592. Proversioii of inferior ridge-crests in riglit third inferior mohvr oi Anancus arvernensis brevirostris 627 

593. Strong proversion of ridge-crests in type third right inferior molar of Syiicoiwlophus corrugatus 627 

594. Synconolophus dhokpatlianensis, restoration of head by Flinscli (right) drawn directly from type craniiun and tusks (left) . 629 

595. Anancus arvernensis and Synconolophus dhokpathanensis, restorations by Flinsch 630 

596. Anancus UTvernensis Croizet and Jobert, cotype figures (mandible, deciduous molars) 633 

597. Anancus arvernensis brevirostris, two cotype inferior molars. After (lervais 634 

598. Anancus falconeri Osborn, type figure (molar), also referred molar of same species. After Falconer 635 

599. Anancus falconeri Osborn, type figure (third inferior molar). After Falconer 636 

600. Anancus arvernensis^?), referred mandible, after Sisnionda, described by him as 'Mastodon' angustidens 637 

601. Anancus arvernensis, referred grinding teeth, from Hungary. After Schlesinger 638 

602. Anancus arvernensis, referred right inferior molar, from Hungary. After Schlesinger 638 

603. Anancus intermedius Eichwald, paratype jaw. After Eichwald 638 

604. Anancus arvernensis progressor Khomcnko, type figure (mandible, molars) 639 

605. Anancus perimensis, paratype and referred skulls, after Falconer and Cautley 640 

606. Anancus perimensis, type and referred specimens from Perim Island. After Falconer and Cautley 641 

607. Map showing distribution of Anancus in India: Perim Island, the Punjab, and sub-Himalayan Siwaliks 642 

608. Map of Middle and Upper Siwaliks of the Salt Range of the Punjab, India 642 

609. Ana7icus properimensis sp. nov. and A. perimensis ref., showing evolution of left third inferior molar during deposition of 

7,500 feet of sedimentation 643 

610. Anancus perimensis, referred symphysis. After Lydekker 644 

611. Anancus perimensis, referred first superior molar. After Falconer 644 

612. Anancus perimensis, paratype skull. After Falconer and Cautley 645 

613. Anancus properimensis sp. nov., type molar 645 

614. Pentalophodon falconeri sp. nov., type cranium with superior grinding teeth. After Falconer and Cautley. Third left superior 

molar, new diagrammatic figure 646 

615. Pentalophodon sivalensis Cautley, type and referred, also Osborn's ideotype. Synconodont or compacted and proverted in- 

ferior molars 649 

616. Pentalophodon sivalensis ( "autley, type figure (molar) 650 

617. Anancus perimensis, referred third left superior molar compared with corresponding molar of A. falconeri. After Falconer 

and Cautley 651 

618. Pentalophodon sivalensis, Osborn's ideotype molar. After Falconer and Cautley 652 

619. Pentalophodon falconeri, referred cranium in profile. After Gaudry 652 

620. Pentalophodon sivalensis or P. falconeri, referred molar. After Lydekker 652 

621. Triple phyla of the Eurasiatic Brevirostrines. Restorations by Flinsch 653 

622. Synconolophus, comparison of type and referred specimens {Synconolophus hasnoti, S. corrugatus, types, and S. dhok- 

pathanensis ref.) 655 

623. Synconolophus ptychodus and S. dhokpathanensis, superior dentition 656 

624. Syriconolophus ptychodus Osborn, type molar 657 

625. Synconoloph us ptychodus, referred molar 657 

626. Synconolophus corrugatus Pilgrim, type figures (molar and symphysis). After Lydekker 659 

627. Synconolophus hasnoti Pilgrim, type figure (molar). After Lydekker 660 

628. Synconolophus hasnoti, referred juvenile craniiun in three aspects. Brown Collection 660 

629. Synconolophijs dhokpathanensis Osborn, tyjje cranium 661 

630. Synconolophus dhokpathanensis Osborn, tyjie cranium and sujM'rior dentition 662 

631. Synconolophus dhokpathanensis Osborn, paratype molar 663 

632. Synconolophus dhokpathanensis, referred superior and inferior molars 664 

633. Synconolophus propalhuncnsis, type mandible and dentition 665 

634. Stcgomastvdon arizonw Cidley, mounted skeleton of type in National Museum. Photographic reproduction 666 

635. Stegornastodon arizonx, the curved-tusked Stegomastodont of Arizona. Restoration by Flinsch 666 

636. Stegornastodon mandibular abbreviation. Ridge-crest internal proversion 668 

637. Stegornastodon chapmani Hays, type figure (molar) 669 

638. Stegornastodon mirificus Leidy, type figui'(> (l(>ft ramus of jaw) 670 

639. Stegornastodon lexanus Osborn, type skull 672 

640. Stegonmstod(ni arizonu' (Jidley, mounted skeleton of type. Restoration and sketch by Knight 672 

641. Stegornastodon successor Cope, type figure (mandible, molar) 673 



xl OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

642. Stegoniastodon texanus Osborn, type skull and dentition; also S. mirificus, S. arizonx, S. texanus, and S. a/towj» restorations 

by Flinsch 674 

643. Stegoniastodon texanus Osborn, paratype mandible and dentition compared with S. succsssor type 675 

644. Stegomastodon, ontogenetic and progressive jaw characters and inferior grinding teeth 676 

645. Stegomastodon texanus, type and referred superior gi'inding teeth 676 

646. Stegomastodon arizonx Gidley, type figure (mandible). After photograph 678 

647. Stegomastodon arizonx Gidley, type figure (cranium). After photographs 678 

648. Physiographic and faunal characters of San Pedro valley of southern Arizona in uppermost Pliocene time. After mural by 

Knight, in American Museum 679 

649. Stegomastodon arizonx, referred third right inferior molar : . .|^ 682 

650. Stegomastodon aftonix Osborn, type figure (superior molars). After original photographs by Samuel Calvin 683 

651. Stegomastodon priestleyi Hay and Cook, type figure (inferior molar) 684 

652. Molar diagrams showing typical crown pattern of third inferior molar of each of the four families of the Mastodontoidea 687 

653. Mioniastodon depereti sp. nov., type molar, compared with type molar of M. merriami 693 

654. Mastodon pavlowi sp. nov., type figure (superior molars) 694 

655. Mastodon granger! Barbour, type figure (skull and tusks) 695 

656. Mastodon acutidens sp. nov., type figure (juvenile skull and jaws) 696 

657. Turicius luricensis, referred inferior jaws with incisors, also superior second and third molars and superior tusks, from 

Upper Bavaria 697 

658. "Mastodon americanus" of Hopwood [ = Turicius sp.], from China (left ramus) 699 

659. Zygolophodon borsoni, referred third left inferior molar, from China 699 

660. Stegolophodon lydekkeri sp. nov., type figure (superior molar), after Lydekker 700 

661. Structural evolution of the cones, conelets, and ridge-crests in the Stegolophodon phylum, in comparison with Palxomastodon . 701 

662. Trilophodon conncxiis Hopwood, type figure (inferior molars) 702 

663. Trilophodon spectabilis Hopwood, type figure (inferior molar) 703 

664. Tetralophodon exoletus Hopwood, type figure (inferior molar) 704 

665. Megabelodon hiUi, referred mandible 708 

666. Megabelodon lulli, tips of referred rostra 708 

667. Megabelodon lulli, referred skeleton mounted in Nebraska State Museum 709 

668. Chart showing type localities of various members of the Mastodontoidea, including Gnathabelodon thorpei Of northwestern 

Kansas 711 

669. Gnathabelodon thorpei Barbour and Sternberg, type figure (mandible, dentition, superior tusk) 712 

670. Migration routes of the three phyla of the Brevirostrinae of the family Bunomastodontidse from a hypothetic African ances- 

tral origin to Europe and thence to India 720 

671. A nanciis sinensis {Pentalophodon sinensis of Hopwood), type figure (superior molars) 721 

672. C uvieronius platensis and C. superbus, referred jaws and teeth, in the Buenos Aires Museiun 724 

673. Correlation chart of the Pleistocene of Nebraska, Iowa, and Europe, including type horizon of Stegomastodon primitivus sp. 

nov 725 

674. Stego7)iastodon primitivus sp. nov., type palate and tusks 726 

675. Type palate and paratype third left inferior molar of Stegomastodon primitivus sp. nov., compared with type third left inferior 

molar of S. 7nirificus Leidy 726 

676. Diagram of Stegomastodon primitivus quarry, northeast of Ainsworth, Nebraska 728 

677. Serridentinus wimani {Trilophodon ivimani of Hopwood), type figure (superior molars) 732 

678. Ocalientinus emmonsi {Gomphotheriuml emmonsi of Hay), type figure (superior molar) 733 

679. Duplication of columnar section of the Mio-Pliocene Devil's Gulch and Valentine region of Nebraska, with important cor- 

rections (see Fig. 272, Chap. VIII) 733 

680. Geographic distribution of three of the great primary stocks of the Proboscidea (Mceritherioidea, Deinotherioidea, Masto- 

dontoidea) 734 



Chapter I 
DISCOVERY OF THE LIVING AND EXTINCT PROBOSCIDEANS 

1. Principles of revision of generic and specific names. 2. Revision and adoption of generic names of Proboscidea. 

Certain of the names applied to the mammoth and Examples of names of prior generic conception and 

mastodon. definition, with classic orthography. A-s adopted 

Linntean special creation system (1735-1766). by Oshorn up to the year 1923. 

Nomenclatural principles adopted in the present Indeterminate names, collective or section names. 

Memoir. Summary of total generic names adopted up to the 

Principle of soundly established usage and of common end of the j'ear 1933. 

Post-1836-selection of the first soundly established gen- ^- I^''yl"ge"^tic nomenclature and classification adopted in 

eric or specific name. the present Memoir. 

Determination of the characters, validity, and names of 4. Acknowledgments of cooperation and assistance. 

specific types. 

The proboscideans and the horses divide the honors of age-long association with man, not only in the historic 
period but far back into prehistoric times. Whereas the horses entered Eurasia late in Pliocene time and were 
hunted for food during the closing 100,0U0-year period, the association of man and the proboscideans appears to 
have been far more remote, certainly extending back into Upper Pliocene time where we discover more or less 
conclusive evidence that man hunted the primitive elephants of the period. During the severe environment of the 
late Palaeolithic age, man had become an extremely clever horse hunter, as shown at Solutre in southern France, 
as well as an extremely clever mammoth hunter, as shown in the "woolly mammoth pit" of Moravia. Whereas 
many horses were necessary to feed a tribe, a single elephant or mammoth would furnish an entire feast, aside 
from the priceless materials of bone and ivory to be fashioned for economic and artistic purposes. 

From Fable and Myth to Science 
From the dawn period of Palaeolithic art the mammoth has been looked upon with wonder and curiosity, 
while the horse has compelled admiration by its beauty and adaptation. This introduces us to a further contrast 
in tlie fabulous, mythical, and scientific history of both these remarkable mammals, namely, that the fossil 
proboscideans of Quaternary age were the first objects in all parts of the world to arouse scientific curiosity, specu- 
lation, and finally research. In the Eastern world there sprang up around the mammoth a distinct 'earth-burrow- 
ing' myth; in the W^estern world there was the natural Biblical interpretation that the fossil elephants were a 
product of the Flood. 

But the present volume is devoted entirely to the succession of discoveries in various parts of the world, in 
fact in all the continents except Australia. These discoveries through the dawn period of scientific research and 
interpretation have been steadily accumulating. They began in 1695 with the first independent ob.servations on 
the 'ancient elephant' by TentzeHus (published in 1698). They culminate in the summary (Chap. XXI) of the 
second volume, in which the full tide of experience and of knowledge brings us to the close of the most complete 
and remarkable evolutionary history of mammalian descent thus far revealed. 

Thus, with only occasional allusion to fabulous and mythologic interpretation which in itself would fill 
volumes teeming with interest, we set ourselves to the task of progressive scientific interpretation: 

First, a period of exploration, discovery, and description which virtually began in 1695 when Tentzelius 
victoriously argued with the doctors of Gotha that he had actually found a fossil or extinct elephant rather than 
the remains of a recent elephant as his colleagues hotly contended; continuing (1) with the special creational 

I 



2 OSBORN: THE PROBOSCIDEA 

influence of Linnsus, Buff on, Blumenbach, and Cuvier, and extending through (2) the intermediate period cul- 
minating in the speculations of Hugh Falconer, a contemporary of Darwin ; thence (3) into the post-Darwin 
evolution period, through the bypaths and mazes of erroneous speculation on proboscidean descent, and finally 
(4) into the illumination of recent time resulting from twenty-seven years of continuous research by the present 
author in the entire field of proboscidean exploration and discovery. 

In brief, the history of the discovery and interpretation of the fossil proboscideans — mastodonts as well as 
elephants — is coincident and coextensive with the dawn, rise, and development of the science of Vertebrate Palae- 
ontology, coupled with our slowly deepening and broadening penetration into the mysteries, principles, and laws 
of evolution. First, there had to be waged a century-long battle to prove that these more or less petrified speci- 
mens were actually fossils belonging to the still unknown world of extinct life; second, another century passed of 
more or less accurate scientific description and nomenclature, and, finally, a third century of evolutionary inter- 
pretation and adaptation of the nomenclature to evolutionary knowledge. 

In the various chapters of the present Memoir the history of every great proboscidean type is separately 
traced from the beginning to the very end of its scientific interpretation, with the fullest justice to the actual facts 
observed by the authors of each of the four great periods of proboscidean thought. So far as practicable the obser- 
vations are recorded Uterally in the language of each author, both as a matter of justice and as a practical matter 
of the dissemination of the widely scattered and at present inaccessible literature in many different languages. 
Also, so far as possible, errors of interpretation are omitted, otherwise this volume would embrace a "comedy of 
errors" and one would lose sight of the main lines of actual fact. 

Immense pains have been taken to reproduce every original description in the language of the author, to 
reproduce every original type figure in facsimile, to retain every original generic name to which the author appears 
to be entitled by originality of observation, as fully explained below in the section on nomenclature of the present 
chapter, and finally to give the fullest credit for priority and originality of thought to the great pioneers in prophecy 
and research, among which are the outstanding names of TentzeUus, Blumenbach, Cuvier, and Falconer. 

As to proboscidean bibhography the series of papers bearing on the subject of this Memoir is arranged alpha- 
betically by authors at the close of the present Volume. A clearer idea, however, of the historic development of 
discovery is shown in the following chronologic arrangement of the writers, greatly renowned, less renowned, or 
obscure, who have enUghtened us on this great subject, from Tentzehus (1698) to the present time. 



AUTHORS CHRONOLOGICALLY CLASSIFIED BY THEIR EARLIEST AND LATEST WRITINGS 







1. 1698-1836. Tentzelius, 


BuPFON, LiNN^us, Blumenbac 


H, Cu-\ 


flER 1 


1602 




Avalo y Figueroa, Diego de 


1767 


Baldassari, Giuseppe 


1792 




1692 




Witsen, Nicolaus 


1768 


Collinson, Peter 


1792 




1693 




Ray, John 


1769 


Hunter, William 


1795- 


1829 


1696 




Ludolf, Hiob 


1771-1793 


Pennant, Thomas 


1795- 


1836 


1698 




Tentzelius, Wilhelmus Ernestus 


1773 


Rozier, Frangois 






1706 




Ides, Eberhard Ysbrant 


1777 [1780] 


Pallas, Peter Simon 


1799 




1715 




R6aumur, Ren6 Antoine 
Ferchault de 


1778 


Buff on, Georges Louis 
Leclerc de 


1802 




1729 




Sloane, Hans 


1779-1810 


Blumenbach, Johann Friedrich 


1803 




1734 




Beyschlag, Joannes Fridericus 


1782-1786 


Merck, Johann Heinrich 


1807 




1734 




Seba, Albert 


1783 


De Fay 


1807 




1735- 


1768 


Linnaeus, Carolus 


1785 


Kennedy, Ildephons 


1808- 


-1825 


1741 




Breyne, John Philip 


1785 [1788] 


Joubert de, 


1808- 


-1837 


1751- 


1775 


Targioni-Tozzetti, Giovanni 


1786 


Fortis, Giovanni Battista detto 


1811 




1752[1756] 


Guettard, Jean Etienne 




Alberto 


1814 




1764 




Daubenton, Louis Jean Marie 


1788-1803 


Camper, Peter 







Gmelin, Johann Fridrich 

Kerr, Robert 

Geoffroy Saint-Hilaire, fitienne 

Cuvier, Georges Leopold Chre- 
tien Fr^d^ric Dagobert 

Corse, John 

Lacepede, Bernard Germain 
Etienne de la Ville 

Artaud, Soulange 

Adams, Michael 

Link, Heinrich Friedrich 

Nesti, Filippo 

Fischer de Waldheim, Gotthelf 

Illiger, Carolus 

Rafinesque-Schmaltz, Constan- 
tine Samuel 



DISCOVERY OF THE PROBOSCIDEANS 



1815 Tilesius von Tiliuiu, Wilhelm 1820-1823 

Gottlieb 1820-1824 

1816 Oken, Lorenz 1821-1823 
1816 Smith, William 1824-1829 
1818-1820 Desmarest, Anselme Gaetan 1824-1830 
1818(1821] Sommerring, Samuel 1824-1833 

Thomas von 1827 
1820 Breislak, Scipione 

1820-1823 Borson, Etienne 1827-1853 



Schlotheim, Ernst Friedrich von 1828 



Kriiger, Johann Friedrich 
Goldfuss, Georg August 
C'uvier, Frederic 
Godman, John D. 
Schinz, Heiiirich Rudolf 
Fitzinger, Leopold Joseph 

Franz Johann 
Eichwald, Eduard von 



1828 
1828 

1828-1835 
1829 



Clift, WiUiam 

Croizet, I'Abbe Jean Baptiste 
Jobert, Antoine C. G. 
Buckland, William 
Desnoyers, Jules Pierre 
Fran9ois Stanislas 



2. 1839-1865. Post-Cuvierian Period Terminating with De Blainville, Laurillard, and Falconer 



1821-1869 Gray, John Edward 1837-1889 

1824-1831 Cooper, William 1838 

1825-1843 Harlan, Richard 1838 

1829-1857 Kaup, Johann Jacob 1838-1850 

1830-1855 Lyell, Charles 

1831 Beechey, F. W. 1839-1858 

1831 Brayley, Edward William 1839-1872 

1831-1842 DeKay, James E. 1840-1842 

1831-1867 Meyer, Hermann von 1840-1882 

1832-1841 Hodgson, Bryan Houghton 1840 

1833-1869 Brandt, Johan Friedrich 

1834 Boue, Ami 1841 

1834 SchoU, Johann Baptist 1841 

1834-1846 Hays, Isaac 1842 

1834-1847 Cautley, Proby Thomas 1842 

1836-1848 Klipstein, August von 1842 

1836-1875 Lartet, Edouard 1842 

1837 Caneto, I'Abbe 1842-1843 

1837 Desor, Pierre Jean Edouard 1845 

1837 Dumeril, Andre Marie Constant 1845-1867 

1837 Geoffroy Saint-Hilaire, Isidore 1846 

1837 Jacquemin, Emile 1846 

1837 Morren, Charles 1846 

1837 Strauss, F. C. J. von 

1837-1864 Blainville, Henri Marie 1846-1855 
Ducrotay de 



Burmeister, Hermann 1846-1856 

Briggs, C. 1846 

Mather, William Williams 1846-1876 
Bonaparte, Charles Lucieri Jules 1847 

Laurent 1847 

Koch, Albert C. 1847-1859 

Foster, J. W. 1847(1850] 

Horner, William E. 1848 

Owen, Richard 1848 
Lichtenstein, Martin Heinrich 1848-1895 

Karl 1850 

Gloger, Constantin W. L. 1850-1853 

Lund, Peter Wilhelm 1850-1858 

D'Orbigny, Alcide 1850-1857 

Grant, (R.) E. 1850-1862 

Lesson, Rene Primev^re 1850-1864 

Nasmyth, Alexander 1851 

Couper, J. Hamilton 1852 

Jackson, J. B.S. 1852 

Falconer, Hugh 1852-1858 

Geinitz, Hans Bruno 1853 

Gray, Asa 1853(1854] 

Serres, Pierre Marcel 1853-1865 

Toussaint de 1854 

Giebel, Christoph Gottfried 1855 

Andreas 1855 



Laurillard, C'harles Leopold 

Hodgson, William B. 

Gervais, Frangois Louis Paul 

Gay, Claudio 

Temminck, Coenraad Jacob 

Aymard, Auguste 

Jiiger, Georg Friedrich 

Goppert, Heinrich Robert 

Poppelack, F. 

Pomel, Auguste 

Gibbes, Robert W. 

Quenstedt, Friedrich August von 

Hoemes, AIorLz 

Wagner, Andreas 

Agassiz, Jean Louis Rodolphe 

Costa, Oronzio Gabriele 

Sismonda, Eugenio 

Girard, Charles 

Glocker, Ernst Friedrich von 

Hingenau, Otto von 

Pictet, Frangois Jules 

Lockhart, Charles Frangois 

Foetterle, Franz von 

Richardson, John 

Dorlhac, J. 

Reuss, August Emanuel 



3. 1859-1894. Darwinian Period Terminating with Leidy, Gaudry, and Cope 



1852-1855 Warren, John C. 

1853-1896 I^idy, Joseph 

1855-1900 Blake, William Phipps 

1856 Hauser, Franz 

1856-1893 Gaudry, Albert 

1858 Rouault, Marie 
1858-1861 Jourdan, Claude 

1859 Abich, Otto Wilhelm 

Hermann von 

1860 Theobald, William 
1861(1858] Gastaldi, Bartolomeo 
1861-1862 Schlegel, Hermann 
1861-1863 Blake, Charles Carter 

1862 Sclater, Philip Lutley 

1863 Logan, William E. 
1863-1864 Suess, Eduard 
1865 Claudius, Matthias 
1867 Le Hon, Henri 
1867-1868 Busk, George 
1867-1868 Murchison, C^harles 
1867-1891 Gunn, John 
1867-1896 Marsh, Othnicl Charles 
1868-1881 Adams, Andrew Leith 
1868-1894 Cope, Edward Drinker 
1869-1878 Blanford, \\illiam T. 
1870 Aradas, Andrea 



1870 Fraas, Oscar Friedrich von 

1871 Hall, James 

1871 Peters, Karl F. 
1871-1892 Dall, William Healey 

1872 Hayden, Ferdinand V'andeveer 
1872-1877 Gill, Theodore 

1872-1879 Chantre, Ernest 

1872-1879 Lortet, Louis 

1873 Dupont, Edouard Frangois 

1873 Marschall, Augusto de 
1873-1876 Biedermann, W. G. A. 

1874 Davies, William 

1874 Yates, Lorenzo G. 

1875 Bachmaim, Isidor 
1875 Christy, Hem-y 

1875 Dana, James Dwight 

1876 Wallace, Alfred Russel 
1876-1883 Major, Charles Lnmanuel 

Forsvth 

1877 Coues, Elliott 
1877 Vacek, Michael 
1878-1899 Stefanescu, Gregoriu 
1878-1902 Amcghiiio, Florentino 

1879 Fuchs, Theodor 

1880 Acconci, Luigi 



1881 Baretti, Martino 

1881 Grewingk, Constantin Caspar 

Andreas 

1881 Hoemes, Rudolf 

1881-1883 Brauns, David August 

1882-1890 Naumann, Edmund 

1882-1890 Nicolucci, Giustiniano 

1883 Jentzsch, Carl Alfred 

1883 Reiss, Wilhehu 

1883 Weinsheimer, Otto 
1883-1887 Howorth, Henry H. 

1884 Fritsch, K. von 

1884 Thomas, PhiUppe 
1884-1890 Martin, Johami Karl Ludwig 

1885 Koken, Ernst 

1885-1908 Kittl, Ernst Anton Leopold von 

1886 \'erri, Antonio 

1887 Jentink, F. A. 

1888 Carles, Enrico de 
1888 Moreno, Francisco P. 

1888 Riitimej'er, Ludwig 
1888-1891 Weithofer, K. Anton 

1889 Baur, Georg 
1891 Botti, Ulderico 
1891 Cantamessa, Filippo 
1891 Filhol, Henri 



OSBORN: THE PROBOSCIDEA 



1891 Whitfield, Robert Parr 

1893-1911 Hutchinson, Henry Neville 
1893 Philippi, R. A. 



1893-1896 Portis, Alessandro 

1894-1895 Hayes, Seth 

1897 Graells, D. Mariano de la Paz 



1897-1899 Troueesart, E. L. 

1899 Heck, L. 

1899 Wagner, George 



4. 1900-1933. Phylogenetic Period Terminating with the Publication of the Present Memoir 

1876-1916 Lydekker, Richard 

1883 Branco [Branca], Wilhelm 

1885-1912 Pohlig, Hans 

1885-1923 Deperet, Charles 

1888-1908 Capellini, Giovanni 

1890-1909 Hofmann, Adolf 

1890-1922 Newton, Edwin Tulley 

1890-1933 Barbour, Erwin Hinckley 

1891 Flower, William Henry 

1891 Lenk, H. 
1891-1901 Zittel, Karl Alfred von 
1891-1912 FelLx, Johannes 

1892 Wortman, Jacob L. 
1892-1934 Osborn, Henry Fairfield 
1893-1911 Thomas, Michael Rogers 

Oldfield 

1894-1931 Pavlow, Marie 

1898-1912 Lambe, Lawrence Morris 

1899-1900 Tullberg, Tycho Fredrik Hugo 

1899-1920 Boule, Marcellin 

1899-1926 Stehlin, Hans Georg 

1 899-1930 Hay, Oliver Perry 

1899-1930 Matthew, William Diller 

1900 Matschie, Paul 

1900 Sclater, William Lutley 
1900-1908 Lucas, Frederick Augustus 

1901 PhUHps, P. Lee 

1901 Wiist, Ewald 
190 1-1928 Andrews, Charles William 

1902 Andrusov, Nikolai Ivanovich 
1902 Beadnell, Hugh John Llewellyn 
1902 Blake, John Frederick 
1902 Herz, Otto 

1902 Leney, Frank 
1902-1930 Sherborn, Charles Davies 

1903 Bourg de Bozas, Robert 
1903 Clarke, John Mason 
1903 Dawkins, William Boyd 
1903 Nordenskiold, Erland 
1903-1930 Sternberg, Charles Hazelius 
1903 Stromer, Ernst 
1903 Villada, Manuel Maria 

1903 Zalensky, Vladimir Vladimiro- 

vich 

1903-1907 Bate, Dorothea M. A. 

1903-1911 Schlosser, Max 

1903-1929 Gidley, James Wilhams 

1903-1934 Gregory, William King 

1904 Bartolotti, C. 
1904 Blanckenhorn, Max 
1904 Palmer, Theodore Sherman 

1904 Seguenza, Luigi 
1904-1908 Dubois, Eugen 
1904-1931 De Lorenzo, Giuseppe 

1905 Anderson, Netta C. 
1905 Maddren, Alfred Geddes 
1905 Pompeckj , Joseph FelLx 
1905 Udden, Johan August 
1905-1914 Schuchert, Charles 
1905-1926 Pfizenmayer, E. W. 
1905-1926 Pilgrim, Henry Guy Ellcock 



1905-1928 


Hornaday, William Temple 


1906 


Beck, Richard 


1906 


Noack, Th. 


1906 


Tokunaga [Yoshiwara], 




Shigeyasu 


1906 


Winge, Herluf 


1907 


Barbiani, A. 


1907 


Fraas, Elserhard 


1907 


Merrill, George Perkins 


1907 


Sanderson, George P. 


1907-1913 


Scott, William Berryman 


1907-1928 


Abel, Othenio 


1907-1929 


Pontier, G. 


1908 


Brown, Barnum 


1908 


Gilmore, Charles Whitney 


1908 


Wegner, Richard Nikolaus 


1908-1910 


Bach, Franz 


1908-1911 


Haug, fimUe 


1908-1915 


Athanasiu, Sava C. 


1908-1917 


Lull, Richard Swann 


1908-1923 


Mayet, Lucien 


1909 


Quackenbush, L. S. 


1909-1910 


Shimek, Bohumil 


1909-1911 


Calvin, Samuel 


1909-1919 


Wadia, D. N. 


1909-1927 


Freudenberg, Wilhelm 


1909-1930 


Cook, Harold James 


1910 


Evans, George Henry 


1911 


Janensch, Werner 


1911 


Sierra, Lorenzo 


1911 


Stremme, Hermann 


1911 


Terra, Paul de 


1911-1922 


Schlesinger, Giinther 


1911-1924 


Breuil, Henri 


1912 


Khomenko, J. 


1912-1924 


Dietrich, Wolfgang 0. 


1912-1929 


Soergel, Wolfgang 


1913 


Hilzheimer, Max 


1913 


Zuffardi, P. 


1913-1916 


Sellards, Elias Howard 


1913-1930 


Woodward, Arthur Smith 


1914 


Buwalda, John P. 


1914 


Heller, Edmund 


1914 


Kato, T. 


1914 


Lomnicki, M. 


1914 


Newton, Richard Bullen 


1914 


Niezabitowski, Edward Lubicz 


1914 


Oswald, FelLx 


1914 


Roosevelt, Theodore 


1914 


Sato, Denzo 


1914 


Schouteden, Henri 


1914-1932 


Reck, Hans 


1915 


Gez, Juan W. 


1915 


Sinclair, William J. 


1915-1928 


Cooper, Clive Forster 


1915-1929 


Matsumoto, Hikoshichiro 


1916 


Cotter, G. deP. 


1916 


Kunz, George Frederick 


1916 


Merriam, John Campbell 


1917-1922 


Airaghi, Carlo 



1918 


Aichel, Otto 


1918 


Fourtau, Rene 


1918-1927 


Stefanescu, Sabba 


1919 


Bolk, Louis Carles Enrico de 


1919 


Brives, Abel 


1919 


Carette, Eduardo 


1920 


Thevenin, Armand 


1920-1928 


Joleaud, Leonce 


1921 


Castellanos, Alfredo 


1921 


Neuville, Henri 


1921-1933 


Frick, Childs 


1922 


Reeds, Chester Albert 


1922-1931 


Klahn, Hans 


1922-1931 


Revilliod, Pierre 


1922-1932 


Haughton, Sidney Henry 


1923 


Granger, Walter 


1923 


Hartnagel, Chris Andrew 


1923 


Petronievics, Branislav 


1923 


Reyes, Alicia E. 


1923 


Roman, Frederic 


1924 


Capitan, Joseph Louis 


1924 


Caterini, Francesco 


1924 


Majer, Istvan 


1924 


Makiyama, Jiro 


1924 


Palmer, Rupert William 


1924 


Peyrony, D. 


1924 


Stauffer, Clinton R. 


1924 


Vaughan, Thomas Wayland 


1924 


Zukowsky, Ludwig 


1924-1925 


Frade, Fernando 


1925 


Lang, Herbert 


1925 


Plate, Ludwig 


1925 


Sandford, Kenneth Stuart 


1925-1928 


Furlong, Eustace L. 


1925-1929 


Bales, Nellie B. 


1926 


Loomis, Frederic Brewster 


1926 


Pohle, Hermann 


1926 


Schmidtgen, 0. 


1926 


Wayland, E. J. 


1926-1927 


Hopwood, Arthur Tindell 


1927 


Anderson, Robert Van Vleck 


1927 


Bather, Francis Arthur 


1927 


Culver, Harold E. 


1927-1929 


Cabrera, Angel 


1927-1929 


Dart, Raymond A. 


1927-1931 


D'Erasmo, Geremia 


1928 


Beliaeva, E. 


1928 


Broom, Robert 


1928 


Fabiani, Ramiro 


1928 


Peterson, Olof August 


1928 


Romer, Alfred Sherwood 


1928 


Stock, Chester 


1928 


Uhle, Max 


1928-1929 


Borissiak, Alexei Alexievich 


1928-1929 


Tolmachoff, Innokenty Pavlo- 




vich 


1928-1931 


Spillmann, Franz 


1929 


Absolon, Karel 


1929 


Anthony, Raoul Louis 




Ferdinand 


1929-1933 


Hayasaka, Ichiro 



DISCOVERY OF THE PROBOSCIDEANS 



1929 Heurn, F. C. van 


1930 


f.hik, Julius 


1929 Prouteaux, M. 


1930 


Moodie, Roj' Lee 


1929 Vaufrey, Raymond 


1930 


Simionescu, loan 


1929 Yabe, Hisakat.su 


1930 


Szalai, Tibor 


1929-1930 Berckhemer, Fritz 


1930-1933 


Simpson, George Gaylord 



See supplementary Bibliography for 1934 and 1935 at the close of \'oluinc I. 



1931 Colbert, Edwin Harris 

1931 Flerov, Constaiitiue C. 

1931 Teilhard de Chardin, Pierre 
1931-1932 Sunamoto, Etsujiro 

1932 Maarel, Frans Henddrik van der 

1933 Koenigswald, Ralph von 



1. PRINCIPLES OF REVISION OF GENERIC AND SPECIFIC NAMES 

We should always remember that nomenclature is the tool rather than the master of palaeontologic thought; 
also that no technical principles should override the work of the early discoverers and naturalists. 

In revising the seventeenth to early nineteenth century generic and specific names of Proboscidea, the author 
has found it impossible either to adopt any consistent rules of nomenclature or to adhere to the International 
Code of Zoological Nomenclature of the successive editions of 1901 to 1916, which was formulated by zoologists 
for zoologic practice in whicii the problems are entirely different from those in palffontologic nomenclature. An 
ex post facto code is difficult in zoology; it is impossible in palajontology until after the time of Cuvier. Zoologic 
types are complete individual specimens, whereas the early and some of the more recent palaeontologic tj'pcs are 
often partial, imperfectly known, and still more imperfectly understood. For example, see the succession of 
names appUed to the 'Mammoth,' which will be fully set forth in Chapter XVIII, Volume II, and to the 'Mas- 
todon,' fully set forth in Chapter VI, as well as the generic names Usted more fully in chronologic order below. 



Certain of the Names Applied to the Mammoth and Mastodon 

Insertions by the author are in square brackets 

1696 Mammontova Kost Ludloff [Ludolf] (Mammontova, probably derived from the Tatar word mamn signifying earth, and 
kost signifying ivory tusks). Original citation inaccessible; cited by Fischer de Waldheim ("Orvctographic," 
III, Fossiles du Gouvern. de Moscou, 1830-18.37, p. 111). 

"I. Elephant. 1. Le Mammont . . . Elephas mammonteus. m. [Fischer, 1830-1837] . . . Je conserve la 
denomination la plus anciennc, et je n'ecris point Marnmouth paree que cc noni n'a pris origine que par corruption 
ou une fau-sse lecture du mot Mammont [Footnote: ' MaMMOHiirb en russc; le Yer (i,) terminal a ete change par les 
anglais en h et le n a ete pris pour un u. D'aillcurs Ludloff est le premier qui en parle (1696) et appelle ces 
ossemens Mammontova Kost et justifie ainsi ma denomination systematique.']." 

Cited by Cuvier (Ann. Mus., VIII, 1806, p. 45): 'C'est sous le nom de comes de mammont, mammontova-kost, 
qu'ils designent les defenses. " 

1788 Mammonteum Camper (Nova Acta Acad. Sci. Imp. Petropol. Communicanda, II, 1787, p. 251). 

[Type: E. primigenius.] . . . "os humanum petrifactum, aut fossile, etiamsi Mammonteorum, Elephan- 
torum, . . . Adserere ex eodem principio audeo Mammonteum animal extinctum non modo esse, sed nullam 
omnino habuisse cum Elephanto similitudincm!" 

1806 Mastodonte [ = Mastodon] Cuvier (Ann. Mus., VIII, 1806, p. 270). 

[Type: M. americaitus.] " Sur le Grand Mastodonte, . . . Animal trds-voisin de I'^l^phant, mais ^ machelieres 
herissees de gros tubcrcules, dont on trouve les os en divers cndroits des deux continens, et surtout pres des bords 
de rOhio, dans I'Amerique Septcntrionale, impropremcnt nomm(5 Mammouth par les Anglais et par les habitans des 
Etats-Unis." 

1816 Mastodon Oken (Lehrb. Naturges., Dritter Thcil, p. 789). 

"Mastodon" defined, distinguished from Mammoth and Elephant. 
Saxony. 2. M. of Simorre [atigustidens]. 3. M. of Chile [humboldtit]. 4. 
Ohio \americanus]. 



Cited: 1. [?tapiroides\ Montabusard, 
M. of Cordilleras \andium]. 5. M. of 



The full and most confusing chronologic sequence of nomenclature of the Mammoth and the Mastodon 
is set forth in detail in Chapters XVIII, Volume II, and VI, of which the above are examples. 

In surveying the above names the question arises whether Mammonteum Camper, 1788, may be regarded as 
a genus with Elephas primigenius as the type; also, what is the proper orthography of the word Mastodonte G. 



6 OSBORN: THE PROBOSCIDEA 

Cuvier, 1806, as compared with the word Loxodonte F. Cuvier, 1825. If Loxodonte is changed to Loxodonta as by 
some authors, why not change Mastodonte to Mastodonta? Similar difficulties arise in connection with every 
generic name proposed for the Proboscidea, excepting only the original Elephas Linnaeus, and even with Elephas 
we are in difficulties, because the Ceylon species, E. indicus Linn., 1754, is not admitted by the Code as antedating 
E. maximus Linn., 1758. Cuvier, the founder of vertebrate palaeontology and the first to clearly conceive 
and define the real characters of these extinct mammals, was the last to adopt the Linnsean binomial terminology. 
Blumenbach first called the American mastodon Ohio-Incognitum, and then Mnmmut ohioticum, because he entirely 
lacked the comparative anatomical genius of Cuvier. 

LiNN^AN Special Creation System (1735-1766) 
The early editions (I to IX inclusive) of the "Systema Naturae" of Linnaeus are not accepted in the Inter- 
national Code of Zoological Nomenclature. It is important to note that Linnaeus used the term Elephas indicus 
in 1754 and called the same animal Elephas maximus in 1758 ("Systema Naturae," 1758, p. 33). The Editio 
Decima, Reformata, of the "Systema Naturae" by Caroli Linnsei, 1758, is accepted in the International Code as the 
starting point of modern systematic zoology. In Chapter XX (the Elephantinse) the specific nomenclature 
under Elephas will be fully discussed. 

(Osborn, "From the Greeks to Darwin," 1894, pp. 128-1301: Finally the turning-point to modern Zoology and Botany 
was marked by the great work of Linnseus, the Systeyna Naturse. The binary system of nomenclature therein proposed was a 
mere tool for the expression of his broad conceptions of the relation of animals and plants to each other. Species were in his 
mind the units of direct Creation: each species bore the impression of the thought of the Creator, not only in its external form 
but in its anatomical structure, its faculties, its functions; and the end of classification was to consider all these facts and to 
arrange animals in a natural system according to their greater or less likeness ... he adopted the aphorism of Leibnitz natura 
non facii saltum: to him every species was exactly intermediate between two others: 'We reckon as manv species as issued in 
pairs from the hands of the Creator.' These were his earlier views in all his writings between 1735 and 1751, in which the 
sentence 7iullx specix novx recurs, expressing his idea of the absolute fixity of species from the period of their creation as de- 
scribed in Genesis, the only change being that of the extension in numbers, not of variation in kind. Yet Linnseus was too 
close an observer to continue to hold this idea of absolute fixity, and in 1762 we find his views had somewhat altered, and 
this is of particular interest because of the hypothesis which he advanced to explain the origin of new species: 'All the species 
of one genus constituted at first (that is, at the Creation) one species, ah initio unam constituerint speciem; they were subse- 
quently multipUed by hybrid generation, that is, by intercrossing with other species.' He was thus inclined to admit a great 
increase of species, more or less recent in origin, arising by hybridity, and losing their perfection of type. He elsewhere sug- 
gested that degeneration was the result of the influences of climate or environment. In the last and thoroughly revised edition 
of the Systema Naturse, which appeared in 1766, we no longer find this fundamental proposition of his earlier works, nullx 
specix novx. 

NOMENCLATURAL PRINCIPLES ADOPTED IN THE PRESENT MeMOIR 

In the present revision the author has therefore adopted four principles as applicable up to the year 1836. 

1) Recognition of priority of conception and of definition rather than of technical priority of the name. 

2) Principle of soundly established usage and of common sense. 

3) Principle of selection of the first generic or specific name founded on a clearly established character or a clearly designated type. 

4) Well-known principle of ' elimination' of species which do not belong, and of ' restriction ' to species which do belong to the genus. 

The need for such common-sense principles may be illustrated by the case of the names applied to the 
Mammoth and the Mastodon. The early writers on fossil Proboscidea were without the guidance of modern 
rules of nomenclature now embodied in the International Code, either as regards: (1) The orthography or ety- 
mology of names, or (2) the designation of type specimens, or (3) the consideration of priority of publication; nor 
did they realize the value of (4) the Linnaean classic orthography, nor (5) the designation of the type locality, nor 
(6) the geologic level or formation in which types occur. 

Consequently to rob Cuvier of his clear conception of grinding tooth structure, which he termed Mastodonte, 
and to substitute the barbaric term Mammut, signifying 'earth-burrower,' would be gross injustice to the founder 
of vertebrate palaeontology. 



DISCOVERY OF THE PROBOSCIDEANS 7 

Principle of Soundly Established Usage and of Common Sense 

Osborn's second principle of nomenclature accords in general with the successive revisions by G. Cuvier 
(1799, 1806), by Oken (1816), by G. Cuvier (1817), by Lartet (1836), by Burmeister (1837), by De Blainville 
(1839-1864), by Gloger (1841), by D'Orbigny (1842), by Falconer (1846, 1857, 1865. 1868), by Laurillard (1846), 
by Pictet (1853), by Leidy (1869), by Cope (1889), and by Trouessart (1899). 

Barbaric and Classic Terms Applied to Elephas primigenius. — -The mammoth {Elephas primigenius 
of Eurasia) received a series of christenings beginning in 1696 when Ludolf termed the mammoth "Mammontova 
Kost," a term cited by Cuvier (1806.1, p. 45), also by Fischer deWaldheim (1830-1837) who insisted that Mammoni 
is the proper expression of the Ru.ssian vernacular and that Mammouth is a corruption. Camper's Mammonteum 
(1788) is cited by Leidy (1869, p. 392). This serves to show that in the nomenclature of the seventeenth and 
eighteenth centuries the terms Mammont and Mammouth were repeatedly applied both in vernacular and in 
scientific usage to the animal now known as Elephas primigenius. Camper's Latin term Mammonteum, by trans- 
position Mammonteus, is the first classic term apphed to the mammoth; a full list of the terms subsequently 
apphed will be cited in Volume IL 

Barbaric and Classic Terms Applied to Mastodon americanus. — The corrupt vernacular name 
Mammut Blumenhsich. (1799), although preoccupied in the vernacular for Elephas primigenius and a synonym 
of Ohio Blum. (1797.2), was revived as a genus for the American mastodon by Hay (1902); so far as we know 
Hay was the first author to recognize Blumenbach's vernacular term Mammut, a term not employed in the 
designation of the type figures by Blumenbach himself. Our greatest American authority on generic names. 
Palmer (1904), following Hay, has accepted Mammut {op. cit., 1904, p. 397); it has also been adopted by Lull 
(1908). It certainly seems in the interest of palaeontology to regard the term Mammut as technically pre- 
occupied by Ohio-Incognitum Blum, and in historic justice to Cuvier to retain firmly in proboscidean nomen- 
clature the appropriate historic name Mastodon Cuvier. 

De Blainville's citations (1839-1864, p. 245) of the barbaric and classic generic names apphed to Mastodon 
americanus are as follows : 

r Le Gband Mastodonte. 

0/iio incognitum , Blumenbach. /Ibildungen, n° ig, 1797- 
Llephas Americanus. Pennant. 

Mammouth Ohioticum , Blumenbach. Manuel d'Hisl. nat. 
E. Americanus, molaribus mullicuspidibus , lamellis post detri- 
tionem quadri-lohatis , G. Cuvier, sur les esp. d'Eleph. viv. et foss. 

'798- / 

Mastodon giganleum , G. Cuvier. Ann. du Mui. yi, p. 370, pi. 49-56i 
.80/ / 

'Mastodon ' soundly established for Mastodon americanus. — Towards the beginning of the eighteenth 
century the American mastodon was discovered. In 1771 Pennant named this animal the American Elephant. In 
1792 Kerr named this animal Elephas americanus; in 1797 Blumenbach (1797.2, Abbildungen, No. 19) first named 
and figured it as Ohio-I ncognitum; in 1799 he named it Mammut ohioticum, using the preoccupied word Mam- 
mut, but in 1810 (Abbildungen, No. 19) he again reproduced this figure with the name Ohio-Incognitum on the 
plate. According to Falconer (1846, p. 18) and De Blainville (1839-1864, p. 245) the name Mastodon giganteum 
Cuvier dates 1805; this date is not accepted by Leidy (1869, p. 393) or Palmer (1904, p. 401). In 1806 Cuvier 
described the animal as "Le Grand Mastodonte," after defining the term in earlier papers (see De Blainville, 
1839-1864, p. 245). Cuvier did not at the time adhere to the Linnaean classic orthography in naming his genera, 



8 OSBORN: THE PROBOSCIDEA 

e.g., Mastodonte { = Mastodon). His name was cited, however, by Oken in 1816 as Mastodon and by Cuvier 
himself as Mastodon in 1817. 

Cuvier's admirable selection of the diagnostic name Mastodon, which has since been reprinted millions of 
times in all the scientific and popular literature, should not be set aside for any form of the barbaric and pre- 
occupied term Mammut. 

Post-1836-Selection of the First Soundly Established Generic or Specific Name 

Osborn's third principle of palseontologic nomenclature is of an entirely different kind; it applies to the 
rules which should govern us in the selection of generic names founded on two or more species which are now 
known to belong to one or more different genera, e.g., the name Trilophodon Falconer, et seq. 

In the chronologic list of Generic Names and Genotypic Species given below, in each instance the following 
eight rules are observed: 

(1) The original orthography is replaced by the classic orthography. 

(2) The earliest known use of the name is cited in the author's orthography. 

(3) The words nomen nudum indicate that the author gave no definition by which the 'species' could be determined; or that the name 

is a synonym. 

(4) The earliest type figure is cited and where valid and available is reproduced. 

(5) The genotype species are invariably cited in the order given by the original author. 
{6) The geographic locality is cited where given ( = Hab.). 

(7) The geologic age or level is cited also ( = Type loc). 

(8) The name or specific determination adopted by the present author is given. 

Important Note. — Few if any of the eight principles of the selection of generic or specific names can with fair- 
ness be applied retroactively to the 1698-1836 Cuvierian period, as listed above. A notable illustration of the 
perils attached to such procedure is afforded by the first mentioned type specimen of the species Elephas primigenius 
Blum.; in this classic description Blumenbach first refers to the fossil skeleton described by TentzeHus as un- 
earthed at Burgtonna near Gotha ; this skeleton is now positively known to be a specimen of the straight-tusked 
elephant named Elephas antiquus by Falconer in 1847, 1857. The technical type of Elephas primigenius Blum. 
is the specific type of Elephas antiquus Falc. Technically, therefore, the species named primigenius antedates the 
species named antiquus. Imagine the dislocation of the entire literature of the Proboscidea if the name antiquus 
were substituted for the name primigenius on technical grounds covering the modern selection of the type, holo- 
type, cotype, etc. 

Determination of the Characters, Validity, and Names of Specific Types 
Beginning with the close of the Cuvierian period in the year 1836, we may with greater justice apply the 
modern principles of palseontologic nomenclature which came into being during the Darwinian period (1859- 
1894) and which have been gradually clarified and intensified in the phylogenetic period (1900-1933). During the 
phylogenetic period two important steps have been taken: First, in revising the work of Cope, Leidy, and Marsh, 
Osborn discovered that the type locality and geologic level were of extreme importance in the study of phyletic 
series of species; second, Schuchert and others sharpened and clarified the conception of the word type, which may 
have an entirely different significance in palaeontology from that conveyed in zoology. Accordingly the modern 
technical procedure may be summarized under the following eleven rules: 

(1) We must first determine the geographic locality and where possible the geologic level as the starting point of the 
determination of the characters of the type, because both these facts have a very important bearing on generic and specific defini- 
tion. Experience teaches that the characters of the original type specimen, the geographic locality, and the geologic level on which 
the specimen was found afford the permanent facts to which all questions of nomenclature must finally be referred in mono- 
graphic investigation. 

(2) The type, where not specifically designated by the original author, is chosen (lectotype) as the first specimen men- 
tioned, the first described, the first figured by the author, consequently great attention has been paid in this Iconography to 
the selection of the particular specimen or specimens which conform with the author's first description and definition. 



DISCOVERY OF THE PROBOSCIDEANS 9 

(3) Do not confuse the holvlype characters with other characters based on cotype, paratype, or neotype specimens which 
the author may have erroneously considered as belonning to the same species as the type. Much of the early systematic work 
on the Proboscidea was done without discrimination between the permanent facts to be derived from the type specimen and 
the opinions which the original or subsequent authors derived from erroneously associated specimens. Osborn's rule is that 
specific definition must be based on original types or holotypos only, unless there is absolutely no possibility of doubt that the 
associated specimens are of the same geologic level and belong to the same species. 

Osborn's usage of the terms type (holotype), cotype, paratype, lectotype, and neotype conforms in general with that of 
Oldfield Thomas (1893), of Cossmann (1904), of Schuchert (1905.1, 1905.2), and of Osborn (1918.473). 

(4) Type = Holotype. A particular individual specimen "deliberately selected by the author of a species, or it may be 
the only example of a species known at the time of original publication. A holotype therefore is always a single individual, but 
may embrace one or more parts, as the skin, .skeleton, or other portions" (Schuchert, 1905.2, p. 10). The holotype must 
usually be determined from the original description. 

(5) CoTVPEs = Coordinate or Equivalent Types. This term is appUcable when the author's type description refers 
to two or more individuals without selecting or designating one as the holotype, so that all appear to be identified equally, by the 
original author, with the specific name given. Osborn's custom in such cases is to designate as the type (lectotype) the 
specimen first mentioned, described, or figured. 

(6) Paratype = Subordinate or Supplementary Type. Besides the particular individual specimen deliberately 
selected by the author of a species as the type, other specimens mentioned or enumerated in the original description which may 
supplement the characters of the type may be chosen as paratypes. This procedure is especially valuable in paleontology, for 
the type specimen very seldom consists of a complete individual. 

(7) Lectotype = Chosen Type. "Where the original diagnosis is without illustrations or is accompanied by figures 
based on two or more specimens, the first subsequent author is at liberty to select from these cotypes a type for the old species, 
adhering, as far as can be ascertained, to the intention of the original author. Such a type specimen is to be designated as a 
Lectotype ( = a chosen type)." (Schuchert, 1905.2, p 12). The lectotype practice of Osborn in palaeontology is either (a) to 
select the first individual specimen named by the original author, because it sometimes happens that the second individual 
specimen belongs to a distinct species, or (b) to select the specimen to which the specific name or description obviously refers. 

(8) Neotype = New Type. Defined by Schuchert (1905.2, p. 14) as a new "supplementary type selected by an author, 
on which a species is to rest because of the loss of the original type or where the original material still extant is so poor 
or fragmentary that from it the characters of the species can not be determined with certainty." Great care must be taken 
that the neotype comes from the same geologic level as the type. 

(9) Metatype. Defined by Schuchert, after Thomas (Schuchert, 1905.2, p. 14) as " 'A specimen received from the original 
locality [in palaeontology, the exact stratum as well] after the description has been published, but determined as belonging to 
his own species by the original describer himself.' " 

(10) TopoTYPE (Schuchert, 1905.2, p. 14). "In palaeontology it is further demanded that the topotype should come not 
only from the exact locality but also from tlie identical stratum that furnished the species." 

(11) Ideotype (Schuchert, 1905.2, p. 15). "These; are the specunens from any place except the original locality, named 
by an author of a species after pubUcation [e.g., M. sivalensis]." 

2. REVISION AND ADOPTION OF GENERIC NAMES OF PROBOSCIDEA 
We now come to the practical test of the above principles of nomenclature in vertebrate palaeontology and 
find that step by step the application of these principles enables us to do full justice to the work of tlie great leaders 
and discoverers in proboscidean palxontology, such as Cuvier and Falconer, only by settiTig aside in certain cases the 
generic and specific names proposed by revisers and reviewers of less authority and of little or no original tfiought. 
The student will observe the following five points: 

(1) Great confusion exists in the use of the eighty or more generic names which have been proposed for the Proboscidea. 

(2) Much research has previously been devoted to the revision of this nomenclature by De Blainville, Falconer, Leidy, Cope, 
Trouessart, Ameghino, Palmer, Hay, Matthew, Lull, Schlesinger, and Freudenberg. In the preparation of the present Memoir 
the author has enjoyed the cordial cooperation of the late Joel A. Allen dean of American mammalogists and highest American 
authority on zoological nomenclature, of T. S, Palmer author of the invaluable "Index Generum Mammalium," of the late 
W. D. Matthew who devoted much time and thought to this subject; also of the British Museum authorities, Oldfield Thomas, 
Charles W. Andrews, and F. A. Bather, as well as C. Davies Sherborn author of the "Index Animalium." 

(3) The notes and correspondence relating to this difficult subject are bound and filed in the Osborn Librarj', including 
the author's inquiries and replies, under the title: Osborn's Correspondence, Nomenclature ofthe Proboscidea, 1921-1935. 
As in the case of the revision of species, reference is made to these notes and letters as, for example: (Allen, letter, 1921). 

(4) The reader will observe between 1758 and 1931 a progressive departure from the Linntean conception of a genus. The 
generic names first proposed are really collective or group names, e.g., Elephas equals all elephants, Mastodon equals all masto- 
donts, Trilophodnn equals all mastodonts with three crests m the intermediate molars. Even so recently as 1886 Lydekker em- 
braced all known elephants, recent or fossil, under the single term Elephas, and all known mastodonts under the single term 
Mastodon: Freudenberg (1922) has followed this practice. 

(5) Recently certain authors who have gone to the opposite extreme have proposed new generic names based on single 
characters or single stages of evolution. 



10 OSBORN: THE PROBOSCIDEA 

A. Examples of Names of Prior Generic Conception and Definition, with 
Classic Orthography. As Adopted by Osborn up to the Year 1923 
The generic names proposed or adopted since 1923 are listed below (p. 11) 

Elephas Linnaeus, 1735-1758, for all the elephants of Indian type. Type Elephas indicm = maximus. 

Mammonteus Camper, 1788. Type Elephas primigenius. This Latinized vernacular name {Mammonleum Camper) pre- 
occupies the long series of generic names subsequently appUed to the Mammoth, e.g., Dicyclotherium, Cheirolites, Syno- 
dontherium, and Polydiskodon. 

Mastodon G. Cuvler, 1806, 1817, for all the zygolophodont mastodonts which are actually related to the genotypic species 
M. americanus. 

Loxodonta F. Cuvier, 1825, 1827, for all the elephants of African type. Type Elephas africanus. 

Deinotherium Kaup, 1829, for all the Deinotheres. Type Deinotherium giganteum. 

Stegodon Falconer and Cautley, 1846 [1847], 1857, for all pro-elephants which resemble the genotypic species E. cliftii, E 
bombifrons, E. ganesa, E. insignis. 

Anancus Aymard, 1855, for all the bunolophodont, brevirostrine mastodonts which are closely related to Aymard's genotypic 
species Anancus macroplus = M. arvernensis. 

Trilophodon Falconer, 1846-1857, collective name, restricted to M. angusHdens. 

Tetralophodon Falconer, 1847-1857, collective name, restricted to M. longiroslris. 

Pentalophodon Falconer, 1857-1865, collective name, restricted to M. sivalensis. 

Rhynchotherium Falconer, 1856-1868, based upon a very distinctive kind of lower jaw from Mexico and including a large 
number of American forms of similar character. Type Rhynchotherium tlascalse. 

ZyGOLOPHODON Vacck, 1877, an excellent genus based on the species M. borsoni, M. turicensis, M. tapiroides, M. pyrenaicus, 
all of which are probably interrelated and are certainly distinct from the ancestry of M. americanus. Osborn (1926.706) 
removed M. turicensis from Zygolophodon making it the type of the genus Turicius. He also refers M. tapiroides to 
the genus Turicius. 

Archidiskodon Pohlig, 1885-1888, founded on the very distinctive specific types Elephas pkmifrons and E. meridionalis. 

PAL.EOMASTODON Andrcws, 1901, founded on the very distinctive type P beadnelli. 

McERiTHERiuM Andrews, 1901, for all the true Mceritheres. Type Maeritherium lyonsi. 

Phiomia Andrews and Beadnell, 1902, founded on the very distinctive type Phiomia serridens. 

B. Indeterminate Names, Collective or Section Names 
Unfortunately several of the remaining sixty or more generic names which have been proposed do not rest 
on such indisputable grounds of dear conception and definition as the above. The erroneous or invalid names fall 

into four categories: 

(1) Indeterminate names, that is, names defined without a clear conception of generic characters or without a clear designa- 
tion of the genotypic species, or with an erroneous designation of genotypic species, or other obvious misconceptions. Examples 
of such names are Ohio Blum., 1797, Mammut Blum., 1799, Gomphotherium Burm., 1837 = Gamphotherium Gloger, Cseno- 
basileus Cope, 1877, Hernimastodon Pilgrim, 1912, Promastodon Pohlig, 1912, also Choerolophodon Schlesinger, 1917. 

(2) Collective, Section or Group Naines. Names founded on parallel or convergent deiital characters which arise entirely 
independently within different genera and different phyla. The earliest outstanding examples of this kind are Falconer's collec- 
tive section or group names, Trilophodon, Tetralophodon, and Pentalophodon, e.g.: 

Trilophodon Falc, 1846-1857, first used as a Section to include Mastodon ohioticus, Deinotherium giganteum, M. tapiroides, 
M. angustidens, and M. andium, species with three-crested intermediate molars. 

Trilophodon Falc, 1857, subsequently defined as a subgenus, with the genotypic species M. angustidens. 
Tetralophodon Falc, 1847-1857, first used as a Section to include Mastodon perimensis, M. sivalensis, M. arvernensis, M. 
longiroslris, and M. latidens, species with four-crested intermediate molars. 

Tetralophodon Warren, 1852, with the genotypic species: M. latidens [ = Stegolophodon], M. arvernensis [ = A7ian- 
cus], and M. sivalensis [ = Pentalophodon]. These species obviously belong to several distinct genera or phyla. 

Tetralophodon Falc, 1857, subsequently defined as a subgenus, with M. longiroslris as the genotypic species. 
Pentalophodon Falc, 1857-1865, defined to include species with five-crested intermediate molars; genotypic species M. 
sivalensis. 

(3) Collective or group names founded upon an erroneous assemblage of species belonging to two or more different genera or 
phyla. Examples of names of this kind are in chronologic order as follows: 

Mastotherium Fischer, 1814. Genotypic species: M. megalodon [ = Mastodon], M. leptodon [ = Trilophodon], M. microdon 
[ = Turicius], M. hyodon { = Cordillerion]. 

Bunolophodon Yacek, 1877. Genotypic species: M . arvernensis [ = Anancus], M.pentelicus [ = Trilophodon (Choerolopho- 
don)], M. atticus [ = Turicius], M. longiroslris [ = Tetralophodon], M. angustidens [ = Trilophodon]. 

Dibelod on Cope, 1884. Genotypic species: D. shepardi [ = Rhynchotherium]. Referred: D. tropicus [ = Cordillerion], D. 
humboldtii { = Cuvieroniu.s]. 

Tetrabelndon, Cope, 1884. Genotypic species: T. angustidens [= Trilophodon]. Referred: T. andium [ = Cordillerion], T. 
productus = Serridentinus]. 

Polydiskodon Pohlig, 1885-1888. Genotypic species: Elephas primigenius [ = Mammonteus], E. indicus [ = Elephas], E. 
namadicus [ = Palxoloxodon]. 

(4) Obvious synonyms, as listed below, which are also frequently collective, are readily disposed of, e.g.: 

Euelephas Falc, 1857. Genotypic species: E. indicus [ = Elephas], E. armeniacus [ = Parelephas], IE. hysudricus 
[ = Elephas (Hypselephas)]. 



DISCOVERY OF THE PROBOSCIDEANS 11 

Several Disputable Names. — From the above analysis it appears that there remain several generic names 
improperly defined and as to the validity of which there may be decided differences of opinion, namely, Gampho- 
therium Gloger, 1841, Trilophodon Falc, 1846-1857, Tetraloplwdon Falc, 1847-1857, Pentalopliodon Falc, 1857- 
1865, Mastotherium Fisch., 1814, Bunolopliodon Vacek, 1877, Dibelodon Cope, 1884, Tetrabelodon Cope, 1884, and 
Polydiskodon PohHg, 1885-1888. 

It is in the usage of such names that the widest differences of opinion have been manifested and that the 
greatest confusion as to the true phylogenetic evolution of the Proboscidea has arisen, for the common reason that 
all these names were founded on misconceptions of true relationships. It is in such a group of names also that it 
proves impossible to adopt any uniformly consistent principle of revision. Consequently for reasons more fully 
stated below we have cut the Gordian knot by courtesy to Falconer and have adopted three collective generic 
names, as follows: 

Trilophodon Falc, 1846-1857, collective name, restricted to .1/. anguslidens. 
Tetralophodon Falc, 1847-1857, collective name, restricted to M . longirostris. 
Pentalophodon Falc, 1857-1865, collective name, restricted to M. sivalensis. 

We have eliminated five other collective names, because they are synonyms of collective names previously 
proposed, namely: 

Mastotherium Fischer, 1814, collective name, also a synonym of Mastodon, etc. 

Bunolophodon Vacek, 1877, collective name, also a synonym of Anancus, of Tetralophodon, of Trilophodon. 

Dibelodon Cope, 1884, collective name, also a synonj'ni (in part) of Mastotherium and of Tetnibclodon. 

Tetrabelodon Cope, 1884, collective name, also a synonym of Trilophodon and of Dibelodon. 

Polydiskodon Pohlig, 1885-1888, collective name, also a synonym of Mammonteus, of Elephas, and of Loxodonta. 

Summary of Total Generic Names Adopted up to the End of the Year 1933 
The sixteen generic names adopted up to the year 1923 as representing distinct and separate phyla, namely, 
Elephas, Mammonteus, Mastodon, Loxodonta, Deinotherium, Stegodon, Anancus, Trilophodon, Tetralophodon, 
Pentalophodon, Rhynchotherium, Zygolophodon, Archidiskodon, Palseomastodon, Maritherium, and Phiomia, are 
now amplified by the adoption of Stegomastodon, Eubelodon, Stegolophodon, Miomastodon, Cuvieronius, 
Serridentinus, Ocalientinus, Serbelodon, Parelephas, Pliomastodon, Turicius, Palseoloxodon, Cordillerion, Atnebel- 
odon, Platybelodon, Synconolophus, Notiomastodon, Hesperoloxodon, Platelephas, (?) Hypselephas, Trobelodon, 
Blickotherium, and Aybelodon. 

The addition of twenty-three generic names to the sixteen first listed, making thirty-nine' in all, does not 
signify the discovery of thirty-nine' distinct phyla or branches, because certain of these genera represent geologic 
sections in certain branches; for example, it is probable that Miomastodon, Pliomastodon, and Mastodon were 
actually successive genera; similarly that Phiomia branched into Amebelodon. 

EUminating such succession of genera, we have thus far discovered no less than twenty-six great phyla in 
which our knowledge includes one or more sections conveniently designated as Gener.\. 

Stegomastodon Pohlif;;, 1912, founded on the Ma.^todon mirifirus of Leidy 

Eubelodon Barbour, 1914. Type Eubelodon morrilli, a single individual from Devil's Gulch, Nebraska. 

Steoolophodon Schlcsincer, 1917, founded on the distinctive type of .Mastodon lalidens (lift. 

Miomastodon Osborn, 1922 Type .Mastodon merriami, a Middle IMiocene mastodont intermediate in character between Palxo- 

mastodon and Mastodon. 
CuviEKONUTs Osborn, 1923, based upon the species Mastodon /iumboW/it Cuvier, characterized by double trefoils and tusks 

without enamel band. 
Serridentinus Osborn, 1923. Genotypic species: Mastodon productus Cope, M. serridens Cope, M. floridanus Leidy, M. 

obscurus Leidy, and Serridentinus [= Trilophodon] simplieidens Osborn. 



'Forty-one by the additiuii of Tonjnobelodon mid Gnathabelodon. 



12 OSBORN: THE PROBOSCIDEA 

Parelephas Osborn, 1924. Upper Pliocene to Upper Pleistocene genus based on the species Elephas jeffersonii Osborn, of 

Indiana. 
Pliomastodon Osborn, 1926, a Pliocene stage based on the species first described by Osborn as Mastodon matthewi, subse- 
quently as Pliomastodon matthewi. 
Pal^oloxodon Matsumoto, 1924, founded on Elephas namadicus naumanni Makiyama. 

TuRicius Osborn, 1926, based on Mastodon turicensis Schinz, one of the genotypes of Zygolophodon Vacek, 1877. 
CoRDiLLERiON Osborn, 1926. Type Mastodon andiwn Cuvier, characterized by single trefoils and spirally twisted tusks with 

enamel band. 
Amebelodon Barbour, 1927. Type Amebelodon fricki, distinguished especially by its shovel-shaped tusks. 
Platybelodon Borissiak, 1928 Type Platybelodon danovi, distinguished especially by its flat tusks. 
Synconolophus Osborn, 1929. Type Synconolophus dhokpathanensis, a very progressive species of this phylum from the 

Middle Pliocene of India. 
Notiomastodon Cabrera, 1929, founded on the distinctive type N otiomastodon ornatus. 
Hesperoloxodon' Osborn, 1931, provisionally proposed, with Palxoloxodon antiquus italicus as the type, representing a 

phylum quite distinct from that of ' Elephas namadicus ' of the SiwaUks. 
Blickotherium Frick, 1933. Type BHckotherium blicki Frick. A Rhynchorostrine with peculiarly formed symphysis and 

laterally compressed, enamel-banded incisors. 
Aybelodon Frick, 1933. Type Aybelodon hondurensis Frick. A Rhynchorostrine with slender jaws, upcurved rostrum, and 

enamelless tusks. 
Trobelodon Frick, 1933. Type Trobelodon taoensis Frick. Of Serridentine relationship. 
OcALiENTiNiTs Frick, 1933. Type Ocalientinus ojocaliensis Frick. Serridentine grinders, abbreviate cranium, and elongate, 

troughlike rostrum. 
Serbelodon Frick, 1933. Type Serbelodonbarbiiirensis Frick. Short, broad symphysis and laterally flattened incisors. 
Platelephas gen. nov.' Type Elephas platycephalus Osborn, having a relatively elongate, dolichocephalic, and platycephalic 

cranium. 
Hypselephas gen. nov.' Genotypic species: Elephas hysudricus Falconer, 1845, 1846, and Elephas platycephalus angustidens 

Osborn, 1929. 



3. PHYLOGENETIC NOMENCLATURE AND CLASSIFICATION ADOPTED IN THE PRESENT 

MEMOIR 

Modern terminology and classification express the discovery of Lines of descent. As clearly shown in Chapter 
II of the present Memoir, the Osborn phylogenetic system (1892-1931) adapts the Linnsean 'special creation' 
nomenclature to the Darwinian 'divergent-evolution ' principle of descent. As set forth in a long series of writings 
on this subject, beginning with Osborn, 1892.67, an essential outcome of Darwinism is the substitution of a 
nomenclature conformable to the divergent branching or polyphyletic adaptations of the Proboscidea to diverse 
habitats and different feeding habits. 

Evolutionary Stage Theory. In the 14th edition (1929) of the Encyclopaedia Britannica the articles 

Proboscidea, Elephant, Mammoth, and Mastodon simimarize the general knowledge of these animals as 

prevaiUng in the decades previous to the year 1929. These divide the Proboscidea into the following nine stages: 

1st Stage. Mceritherium, Upper Eocene of Egypt. Very primitive. 

2d Stage. Palxomastodon, Lower Oligocene of Egypt. Great advance on Mceritherium. 

3d Stage. Tetrabelodon angustidens, Lower Middle Miocene of Europe. Advance on Palxomastodon. 

4th Stage. Tetrabelodon longiro.ttris, Pliocene of Europe. In external appearance entirely elephantine. 

5th Stage. Mastodon, Pleistocene. Terminal stage of the mastodonts. 

6th Stage. Deinotherium, Mio-Pliocene of Eurasia. Side branch. 

7th Stage. Stegodon, Pliocene of India. Transitional to the elephants. 

8th Stage. Elephas antiquus of Eurasia and other Plio-Pleistocene species. 

9th Stage. Elephas maximus, E. africanus, of India and Africa. 

The arrangement of the Proboscidea in these nine stages accords with the outworn belief among zoologists 

that one or more of these stages represents a successive or ascending order; for example, that Moeritherium gives 

rise to Palseomastodon, as, in turn, Mastodon passes into Stegodon and then into Elephas. This prevaiUng belief 

is entirely erroneous; it is a vestige of monophyletic theories of animal descent which in the past have been largely 

due to the breaks and gaps in our palseontologic collections. Now that these gaps are in a great measure filled 

by discovery and these animals intensively studied, the older monophyletic theories prove to be illusory; they 

'Description in forthcoming Volume II (The Elephantoidea). 



DISCOVERY OF THE PROBOSCIDEANS 13 

are wholly superseded and replaced by pohjphyletic systems in which each of the generic phyla, even of the Oligo- 
cene, passes back into fundamental divisions that occurred during Eocene times. It is proven that the divergence 
or branching of the Proboscidea, like that of all other ungulates, was well advanced in the Eocene. 

In brief, it becomes necessary to replace the old evolutionary stage theory by the new polyphyletic adaptive 
radiation theory of Osborn. 

Polyphyletic Adaptive Radiation Theory. — Each of the successive chapters of this Memoir (Chap. II, 
"Origin and Classification of the Proboscidea," to Chap. XX, "The Subfamily Elephantinae") contains a full 
exposition not only of the principles of adaptive radiation but of the necessary modification of the nomenclature 
by polyphyletic discovery. Finally in Chapter XXI ("Affinities, Migrations, and Phylogeny ") the whole subject 
of the adaptive evolution of the Proboscidea along polyphyletic lines is fully set forth. 

Each of the forty-one genera finally established in the present Memoir represents a clearly distinct stage 
of descent; for example, it has very recently been determined that the line of descent terminating in the 'Elephas 
namadicus' of India is quite distinct from the line of descent terminating in the 'Elephas antiquus italicus' of 
western Europe. It follows that, not only as a matter of convenience but as an expression of truth, different 
generic names must be applied to these two Unes of descent and to their African relatives, namely : 

Elephas antiquus Une of descent = Hesperoloxodon Osborn, 1931. 

Elephas namadicus line of descent = Palseoloxodon Matsumoto, 1924. 

Elephas africanus Une of descent = Loxodonta F. Cuvier, 1825, 1827. 

Thus our modern phylogenetic knowledge completely revolutionizes all the older taxonomic sj'stems and 
theories. 

4. ACKNOWLEDGMENTS OF COOPERATION AND ASSISTANCE 
As recited in the Preface and more fully in Chapter II, the preparation of this Memoir began in the year 1907 
with the author's expedition to the Fayum of northern Egypt, accompanied by Walter Granger as head of the 
mammalian palaeontology staff of the American Museum. This eventful journey verified and amplified the 
remarkable discoveries of Hugh J. L. Beadnell and Charles W. Andrews, and determined the author to give 
the Proboscidea monographic treatment similar to that of the Titanotheres. 

At the time there was little conception of the difficulties of the undertaking. First entered upon in the 
midst of the concluding phases of research on the Titanotheres — the Titanothere Monograph having been sent to 
press in 1920 — intensive work on the Proboscidea Memoir has continued with periods of interruption for fourteen 
years — altogether twenty-seven years. 

It soon became obvious that the cooperation of practically all the mammaUan palaeontologists of the world 
was essential to a comprehensive and thorough treatment of the subject. It would be impos.sible to express to 
each of these collaborators the full measure of the author's appreciation; it is, moreover, difficult to give a com- 
plete list of those who have helped in many ways. 

Othenio Abel, Professor of PaliEobiology, Universitiit, Vienna, Raoul Anthony, Professor of Comparative Anatomy, Museum 
Austria. National d'Histoire Naturelle, Paris, France. 

Karel Absolon, Curator of Zoology and Pala-ontology, Moravski? Erwin Hinckley Barboith, Professor of Geology, University of 
Zemsk(5 Museum, Bruiin, Czechoslovakia. N^l^'''^^u'^ ' ^"'"'^^o''' Nebraska State Museum, Lincoln, 

*Carl Ethan Akeley, Associate in Mammalogy, The American _ hi a ^d n en ■ , c, ■ .-^ i-^t , • r^ 

Museum of Natural History, New York Citv. Dorothea M A. Bate, Unofficia Scientific Worker, in Depart- 

*JoEL Asaph Allen, Curator of Mammalogy, The American Mu- '{"^Ln EngSd' " '"■^^' 

seum of Natural History, New York City. *Francis Arthur Bather^ formerly Keeper of the Department of 

♦Charles William Andrews, Assistant m Department of Geologj-, Geology, British Museum (Natural History), London, 

British Museum (Natural History), London, England. England. 

Harold Elmer Anthony, Curator of Mammalogy, The American Fritz Beuckhemer, Curator, Geologische Abteilung, Wiirttem- 

Museum of Natural Historj', New York City. bergische Naturaliensammlung, Stiittgart, Germany. 



14 



OSBORN: THE PROBOSCIDEA 



William Reid Blair, Director and General Curator, New York 
Zoological Park, New York City. 

Alexei Alexievich Borissiak, Geological Museum of the 
Academy of Sciences, Leningrad, Union of the Socialist 
Soviet Republics. 

Marcellin Boule, Professor of Palaeontology, Museum National 
d'Histoire Naturelle; Director, Institut de Paleontologie 
Huinaine, Paris, France. 

£douard Bourdelle, Museum National d'Histoire Naturelle, 
Paris, France. 

Isaiah Bowman, Director, American Geographical Society, New 
York City. 

Ferdinand Broili, Director, Institut fiir Paliiontologie und 
historische Geologic, Bayerische Ludwig-Maximilians- 
Universitiit, Munich, Germany. 

Robert Broom, formerly Profesisor of Geology and Zoology, Vic- 
toria College, Stellenbosch, South Africa. 

Barnum Brown, Curator of Fossil Reptiles, The American Mu- 
seum of Natural History, New York City. 

William Alanson Bryan, Director, Los Angeles Museum, Los 
Angt'lcs, Calif ornia. 

John P. Buwalda, University of California, Berkeley, California. 

Angel Cabrera, Profes.sorof Palaeontology, Institute del Museo y 
Escuela de Ciencias Naturales, Universidad Nacional de 
La Plata; Curator of Palaeontology, Museo de La Plata, 
La Plata, Argentina. 

'Giovanni Capellini, formerly Professor of Geology, University, 
of Bologna, Italy. 

Joel Ernest Carman, Professor of Geology, Ohio State Univer- 
sity, Columbus, Ohio. 

*JoHN Mason Clarke, formerly Director of the New York State 
Museum, Albany, New York. 

Edwin Harris Colbert, Assistant Curator of Vertebrate Paleon- 
tology, The American Museum of Natural History, New 
York City. 

Harold James Cook, formerly Curator of the Department of 
Palaeontology, Colorado Museum of Natural History, 
Denver, Colorado. 

Clive Forster Cooper, University Museum of Zoology, Cam- 
bridge, England. 

Rav.mond a. Dart, Professor of Anatomy and Dean of the Faculty 
of Medicine, University of the Witwatersrand, Johannes- 
burg, South Africa. 

Giuseppe De Lorenzo, Professor of Geology, Universita di Napoli, 
Naples, Italy. 

♦Charles Deperet, formerly Dean of the Faculty of Sciences, 
Universite de Lyon, France. 

Geremia D'Erasmo, Professor of Palaeontology, LTniversita di 
Napoli, Naples, Italy. 

Wolfgang 0. Dietrich, Geologisches Palaeontologisches Institut 
und Museum der Universitat, Berlin, Germany. 

Raymond Lee Ditmars, Curator of Mammals and Reptiles, New- 
York Zoological Park, New York City. 

EuGEN Dubois, Koninklijke Akademie van Wcteaschappon, 
Amsterdam, Netherlands. 

Nellie B. Eales, Lecturer, Department of Zoology, LTniver.sity of 
Reading, England. 

Jesse Dade Figgins, Director, Colorado Museum of Natural 
History, Denver, Colorado. 

WiLHELM Freudenberg, formerly of Georg August Universitat, 
Gottingen, Germany. 

Childs Frick, Honorary Curator of Late Tertiary and Quaternary 
Mammals, also Trustee, The American Museum of 
Natural History, New York City. 

Eustace L. Furlong, Museum of Paleontology, University of 
California, Berkeley, California. 

Claude Gaillard, Director, Musee des Sciences Naturelles, Lyon, 
France. 



♦James William Gidley, Assistant Curator of Fossil Mammals, 
United States National Museum, Washington, D. C. 

Charles Whitney Gilmore, Curator, Department of Vertebrate 
Palaeontology, United States National Museum, Wash- 
ington, D. C. 

Allen H. Godbey, Professor, Old Testament, Duke University, 
Durham, North Carolina. 

Walter Granger, Curator of Fossil Mammals, The American 
Museum of Natural History, New York City. 

Madison Grant, President of the New York Zoological Society, 
New York City. 

William King Gregory, Curator-in-Chief of Living and Extinct 
Fishes, Curator of Comparative and Human Anatomy, 
Research Associate in Palaeontology, Associate in Anthro- 
pology, The American Museum of Natural History, 
New York City. 

Herman Gunter, Sate Geologist, Tallahassee, Florida. 

C. R. Halter, formerly of The American Museum of Natural 
History, New York City, and of the Department of 
Biology, Southern College, Lakeland, Florida. 

Robert Torrens Hatt, Assistant Curator of Mammalogy, The 
American Museum of Natural History, New York City. 

Walter W. Holmes, Field Associate in Palaeontology, The Ameri- 
can Museum of Natural History, New York City. 

Arthur Tindell Hopwood, Assistant Keeper, Department of 
Geology, British Museum (Natural History), London, 
England. 

William Temple Hornadav, formerly Director of the New York 
Zoological Park, New York City. 

♦Sir Harry H. Johnston, Poling, Arundel, England. 

A. C-. DE JoNciH, Director, Geological Survey of Dutch East Indies, 
Bandoeng, Java. 

H. MuNNiKS DE JoNGH, The Hague, Netherlands. 

George Frederick Kay, Head, Department of Geology, Univer- 
sity of Iowa, also State Geologist of Iowa, Iowa City, Iowa. 

Hans Klahn, Professor of Geology, Universitat, Rostock, Ger- 
many. 

Ernst C. Krohn, First Vice-President of the Greater St. Louis 
Mu.seum of Natural History, St. Louis, Missouri. 

♦George Frederick Kunz, Research As.sociate in Gems, The 
American Mu.seum of Natural History, and Vice-Pres- 
ident, Tiffany and Company, New York City. 

Gustav KiJsTHARDT, Inspector, Zoologische Staatssammlung, 
Munich, Germany. 

Francois Antoine Alfred Lacroix, Secretaire perpetuel pour 
les sciences physiques de I' Academic des Sciences, Institut 
de France; Professor of Mineralogj-, Museum National 
d'Histoire Naturelle, Paris, France. 

Herbert Lang, The Transvaal Muiseum, Pretoria, South Africa. 

William Dixon Lang, Keeper, Department of Geology, British 
Museum (Natural History), London, England. 

Louis Laurent, Director-Curator, Museum d'Histoire Naturelle, 
Marseille, France. 

Frederic Brewster Loomis, Professor of Geology, Amherst 
College, Amherst, Massachusetts. 

EiNAR Lonnberg, Naturhistoriska Riksmuseum, Stockholm, 
Sweden. 

*Frederic Augustus Lucas, Honorary Director, formerly 
Director, The American Museum of Natural History 
New York City. 

Richard Swann Lull, Sterling Professor of Vertebrate Palae- 
ontology, Director of Peabody Museum, Yale University, 
New Haven, Connecticut. 

Frans Henddrik van der Maarel, De Bilt, Netherlands. 

♦Paul Matschie, Museum fiir Naturkunde, Berlin, Germany. 

HiKOSHiCHiRo Matsumoto, Professor of Palaeontology, Tohoku 
Imperial University, Sendai, Japan. 



ACKNOWLEDGMENTS 



15 



*WiLLiAM DiLLER Matthew, Professor of Palceontology and 
Director of the Museum of Palaeontology, University of 
California. Berkeley, California, formerly Curator of 
Geology and Paleontology, The American Museum of 
Natural History, New York City. 

LuciEN Mayet, Professor of Anthropology, University de Lyon, 
France. 

John Campbell Merriam, President, Carnegie Institution of 
Washington, Washington, D. C. 

Elmer Drew Merrill, Director-in-Chief, New York Botanical 
Garden, New York City. 

*George Perkins Merrill, Head (Curator of Geology, United 
States National Museum, Washington, D. C. 

Gerrit Smith Miller, Jr., Curator of Mammals, United States 
National Museum, Washington, D. C. 

*RoY Lee Moodie, Professor of Palseodontology, College of Den- 
tistry, University of Southern California, Los Angeles, 
California. 

Charles Craig Mook, Associate Curator of Geology and Palaeon- 
tology, The American Museum of Natural History, 
New York City. 

Alfred T. Nester, formerly Consul of the United States of Ameri- 
ca at Naples, Italy. 

W. F. F. Oppenoorth, Bandoeng, Java. 

Raymond C. Osbukn, Department of Geology, Oiiio State 
University, Columbus, Ohio. 

Theodore Sherman Palmer, United States Biological Survey, 
Washington, D. C. 

Marie Pavlow, Professor, University de Moscow, Institut 
Geologique, Leningrad, Union of the Socialist Soviet 
Republics. 

*Olof August Peterson, Curator, Mammalian Palaeontology, 
Carnegie Museum, Pittsburgh, Pennsylvania. 

Guy Ellcock Pilgrim, formerly Superintendent of the Geological 
Survey of India, Calcutta, India. 

Hans Pohlig, formerly of the Universitiit, Bonn, Germany. 

Chester Albert Reeds, Curator of Geology and Invertebrate 
Pateontology, The American Musum of Natural His- 
tory, New York City. 

Graham Renshaw, Sale, Manchester, England 

Jules Joseph Repelin, Curator, Museum d'Histoire Naturelle, 
Marseille, France. 

Pierre Revilliod, Director, Musee d'Histoire Naturelle de la 
Ville de Geneve, Switzerland. 

Alicia E. Reyes, Assistant Curator, Museo de Historia Natural, 
Mexico City, Mexico. 

Elmer Samuel Riggs, Associate Curator of Palaeontology, Field 
Museum of Natural History, Chicago, Illinois. 

GiJNTHER Schlesinger, Director, Niederosterreichische Landes- 
sammlungen, Vienna, Austria. 

*Max Schlosser, Bayerische Staatssammlung fiir Palaontologie, 
Munich, Germany. 

C. Bertrand Schultz, State Museum, University of Nebraska, 
Lincoln, Nebraska. 

Elias Howard Sellards, Associate Director, Bureau Economic 
Geology, Professor of Geology, University of Texas, 
Austin, Texas. 



Charles Davies Sherborn, Bibliographer, British Museum 

(Natural History), London, England. 
George Gaylord Si.mpson, Associate Curator of \'ertebrate Palae- 
ontology, The American Museum of Natural History, 

New York City. 
Wolfgang Soergel, Professor, Geology and Palaeontology, and 

Director, Geologisch-paliiont. Institut und Museum, 

Schlesische Friedrich-Wilhelms-Universitiit, Breslau, 

Germany. 
Franz Spill.mann, Director, Cabinete Zool6gico, Museo, Universi- 

dad Central del Ecuador, Quito, Ecuador. 
•Gregoriii Stefanescu, Professor of Palaeontology, University 

of Bucharest, Rumania. 
Sabba Stefanescu, University of Bucharest, Rumania. 
Hans Georg Stehlin, President, Naturhistorisches Museum, 

Basel, Switzerland. 
R. A. Stirton, Department of Geological Sciences, University of 

California, Berkeley, California. 
Chester Stock, California Institute of Technology, Pasadena, 

California. 
Witmer Stone, Director Emeritus, Museum of the Academy of 

Natural Sciences, Philadelphia, Pennsylvania. 
W. E. SwiNTON, British Museum (Natural History), London, 

England. 
Pierre Teilhakd de Chardin, National Geological Survey of 

China, Peiping, China. 
•Michael Rogers Oldfield Thomas, British Museum (Natural 

History), London, England. 
Innokenty Pavlovich Tolmachoff. Professor of Palaeontology, 

University of Pittsburgh, Pittsburgh, Pennsylvania. 
Edward Leffingwell Troxell, Professor of Geology, Trinity 

College, Hartford, Connecticut; formerly Research 

Associate in Palaeontology at Yale University, New Haven, 

Connecticut. 
G. M. Vevers, Superintendent of the Gardens, Zoological Society 

of London, England. 
David Meredith Seares Watson, Jodrell Professor of Zoologj' 

and Comparative Anatomy, University of London, 

England. 
Max Weber, formerly Director, Zoologisch Institut en Museum, 

Amsterdam, Netherlands. 
Coleman Shaler Williams, Department of Vertebrate Palae- 
ontology, The American Museum of Natural History, 

New York City. 
M. WiLMAN, Curator, McGregor Museum, Kiniberley, South 

Africa. 
Arthur Smith Woodward, formerly Keeper of Geology, British 

Museum (Natural History), London, England. 
E. D. Van Oort, Director, Rijks-Museum van Natuurlijke His- 
toric, Leiden, Netherlands. 
Hisakatsu Yabe, Professor of Geology and Palaeontology, Tohoku 

Imperial University, Sendai, Japan. 
Naohide Yatsu, Professor of Zoology, Tokyo Imperial University, 

Tokyo, Japan. 



The outstanding names of the collaborators in the Volume include several great and highly esteemed col- 
leagues who have fallen by the way, such as Charles Dep^ret of Lyons, William D. Matthew, Joel A. Allen, 
Frederic A. Lucas and Carl E. Akeley of the American Museum, James W. Gidley of the U. S. National Museum, 
Paul Matschie of the Natural History Museum of Berlin, John Mason Clarke of the New York State Museum, 
and finally, but not least, Charles W. Andrews of the British Museum. These names, indicated on the Honor Roll 
by a star, are included in the above list of those who have cooperated more or less extensively with the author 
in the present work. 




Fig. 1. Heads op twenty-nine Types or Proboscideans belonging to seventeen of the nineteen Subfamilies as determined by the 

PRESENT Author. Restor.wions by Margret Flin.<!ch (1932), under the direction of Henry Fairfield Osborn 

All figures one-hundredth natural size, excepting 1 a which is one thirty-fourth natural size. 



1,1a Mwritherium lyonsi Andrews. 
2 Phiomia osbomi Matsumoto. 

Trilophodon anguslidens Cuvier. 

Platybelodon grangeri Osborn. 

Amehelodon fricki Barbour. 

Serridentinus serridens Cope. 

Telralophodon punjabiensis Lydekker. 

Palseomaslodon beadnelli Andrews. 

Mastodon americanus Kerr. 

Rhynchotheriuni tlascalx Osborn. 

Deinotherium giganlissimum Stefinescu. 

Ananciis arvernensis Croizet and Joliert. 

Synconolophus dhokpathanensis Osborn. 

Siegomastodon arizon^ Gidley. 

Noiiomastodon argentinus Ameghino. 



16 Eubelodnn morrilli Barbour. 

17 Cordillerion andium Cuvier. 

18 Cxwieronius superbus Ameghino. 

19 Slegodon ganesa Falconer. 

20 fslegodon trigonoccphalus Martin. 

21 Archidiskodon imperator Leidy. 

22 Parelephas jejfersoriii Osborn. 

23 Mammonieus primigenius Blumenbach. 

24 Hesperoloxodon antiquus italicus Osborn. 

25 Pahsoloxodon falconeri Leith Adams. 

26 Palasoloxodon namadicus Falconer. 

27 Elephas indicus Linnseus. 

28 Loxodonla africana Blumenl)ach. 

29 Loxodonla africana pumilio Noack. 



16 



Chapter II 

ORIGIN AND CLASSIFICATION OF THE PROBOSCIDEA BEGINNING 
WITH MCERITHERIUM, PAL/EOMASTODON, AND PHIOMIA 



I. LiNN^AN Zoological Classification and Osborn's Phylo- 

GENETic Classification of the Proboscidea. 

1. Principles of the Osborn Phylogenetic Classification. 

2. Origin of the Proboscidea. 

3. The primarj^ stocks established by primary choice of 

food in different habitats. 

4. Ordinal, subordinal-superfamily, family, subfamilj'-phy- 

letic, generic-phyletic characters employed in the phylo- 
genetic classification of the Proboscidea. 

5. Harmony of Linnisan and phylogenetic classification ex- 

pressing our knowledge up to the close of the year 1933. 

n. Osborn's Theory (1900) as to the African Origin or 
Center of Adaptive Radiation of the Proboscidea. 

1. Andrews confirms the African origin of the Proboscidea, 

Sirenia, and Hyracoidea. 

2. American Museum Fayflm Expedition of 1907. Interpre- 

tation and discovery bearing on the ijhylogeny of the 
Fayftm Proboscidea. 

3. Relations of Mwritherium, Palseomastodon, and Phiomin 

to each other and to other mammals. 

4. Matsumoto's complete separation of Meeritherium, Palseo- 

mastodon, and Phiomia. 

5. Africa demonstrated as the homeland of the three families 

Moeritheriidffl, Mastodontidae, and Bunomastodontidae. 



III. Phylogenetic Relations and Classification of M(eri- 

therivm, Pal.eomastodon, and Phiomia. 

1. Principle of adaptive radiation expressed in the phylo- 

genetic classification of the Faylim Proboscidea. 

2. Amphibious habits of Mwritherium. 

3. Lowland habits of Phiomia, ancestor of the Amebelodont- 

ina;. 

4. Forest-loving habits of Palxomaatodon, not ancestral to 

Mastodov. 

IV. Revision and Synopsis of Genera, Species, k^d Types of 

the Fayum Proboscideans in the Order of Their 
Original Description. 

1. Palxomastodon beadnelli Andrews, 1901. 

2. Maeritherium lyonsi Andrews. 1901. 

3. Phiomia serridens Andrews and Beadnell, 1902 

4. Moeritherium gracile Andrews, 1902. 

5. Mwritherium trigodon Andrews, 1904. 

6. Phiomia minor Andrews, 1904. 

7. Palxomastodon parvus Andrews, 1905. 

8. Phiomia wintoni Andrews, 1905. 

9. Phiomia barroisi Pontier, 1907. 

10 Mwritherium andrewsi Schlosser, 1911. 

11. Palxomastodon intermedins Matsumoto, 1922 

12. Phiomia osborni Matsumoto, 1922. 

13. Moeritherium ancestrale Petronievics, 1923. 



INTRODUCTION 

Classification. — The classification of the Proboscidea progre.sse.s with our ever increasing knowledge of 
origins, adaptive radiations, and phyletic successions. Consequently classification is progressively modified by 
discovery and research. The classification in the present Memoir attempts to set forth our general knowledge of 
the evolution of the Proboscidea up to the end of the year 1933; the nomenclature, that is, hsts of genera, species, 
etc., chiefly terminates at the close of the year 1933, because it is impossible to include all the new specific and 
generic lists of 1934 up to the moment of going to press. Tlie classification of 1935 concludes the present Volume. 

Our rules of nomenclature will be a shock to many zoologists, especiallj^ to those who carry the principle of 
priority to the point of adopting misspelled, misapplied, and practically unpublished names. After the most 
honest and painstaking endeavor, we have found it impracticable to apply uniformly all the principles established 
in zoology to nomenclature in palaeontology. The fundamental reason is that the imperfect data of many original 

17 



18 



OSBORN: THE PROBOSCIDEA 



descriptions in palaeontology are in wide contrast to the usually perfect data of zoology. The art of nomenclature 
is entirely different in these two branches. Consequently the rules of nomenclature framed by zoologists will 
inevitably be different in certain respects from those which can be adopted by palaeontologists. Certain of the 
reasons for this statement will become clear in the following sections. 

Origin. — Our present knowledge of the origin of the Proboscidea begins with the wonderfully varied and 
interesting forms discovered in the Fayum region of North Africa in 1901 and successively examined by Andrews, 
Schlosser, Matsumoto, Petronievics, and Osborn between 1901 and 1923. In the systematic revision all of these 
Fayiim animals will have to be considered together; they are therefore treated in chronological order of descrip- 
tion in the present chapter, although it is now known that they belong to at least three distinct phyla. After being 
thus separated phyletically, they are again treated in succeeding chapters in their true phylogenetic order, namely: 

Palseomastodon, forest loving, the possible phyletic ancestor of the Mastodontidae. 
Phiomia, shore loving, certainly the collective ancestor of the shovel-tusker Amebelodon. 
Moeritherium, water loving, standing and ending in a phylum by itself. 




Flood-plain Scene of the Ancient River Nile, North .\KnicA, in Lower Oligocene Time 
Restoration by Margret Flinsch (1932), under the Direction of Henry Fairfield Osborn 
Fig. 2. Theoretic restoration of Palseomastodon beadnelli (upper left), Phiomia osborni (upper right), and Mcerilheriuin andrewst 
(upper center), all one-fiftieth natural size. Moeritherium lyonsi in foreground one-thirtieth natural size. 



I. LINN^AN ZOOLOGICAL CLASSIFICATION AND OSBORN'S PHYLOGENETIC 

CLASSIFICATION OF THE PROBOSCIDEA 

LiNN^AN Classification (Linnaeus, 1735-1766) aimed to define the apparent relationships and affinities of 
existing animals as they were created and spread contemporaneously on the present surface of the earth, and as 
revealed by the comparative anatomy and zoology of existing animals. The "Systema Naturae" of Linnaeus is 
zoological and creational. 

Phylogenetic Classification (Osborn, 1892-1933) aims to distinguish ties of blood and descent among 
ancestral as well as existing mammals; it also aims to adapt and modify the Linnsean classification to the dis- 
coveries of paltEontology, namely, to include animals belonging to phyla or ancestral lines and to connect them with 
existing animals. It expresses the nearness or the remoteness of phyla of all degrees of magnitude, namely, phyla 
of species, genera, subfamilies, families, orders. The term Phylum has a larger taxonomic significance, e.g., 
Phylum Chordata = animals with a notochord. 

The word phylogenetic signifies the origin or genesis of lines of descent of ancestral animals unknown to Lin- 
naeus. The word phylum is derived from the Greek (cf. Liddell and Scott, "A Greek-English Lexicon," 1883, 
pp. 1698, 1703): "<t>v\r], 17 (cpvw) properly, like <j)v\ov, a set of men naturally distinct; ... In usage ^uXi? corre- 
sponded to the Roman tribus, and signified I. a body of men united: 1. by supposed ties of blood and descent, 
a cian or sep<, such as tho.se among the Dorians, ... or 2. by local habitation," [p. IQ98). "<t>vo}, ... A. trans., 
. . . tobring forth, produce, put forth, (t>v\\a . . . . B. Pa.ss., . . . to grow, ivax, spring up or forth, arise, cnme into 
being," (p. 17U3). 

1. PRINCIPLES OF THE OSBORN PHYLOGENETIC CLASSIFICATION 

In 1891, Osborn began to apply a system of phylogenetic classification to the order Perissodactyla (1892.67) 
and developed it in subsequent years especially in treating the family Rhinocerotidae (1900.192) and finally 
the superfamily Titanotherioidea (1929). The purpose of this classification of the Proboscidea is to express 
our present knowledge of the more or less profound gaps between the greater and the lesser branches. Most of the 
older classifications were based on the idea that there were but two genera, Elephas and Mastodo7i. Cope (1889.2) 
divided the Proboscidea into many genera but did not propose a phylogenetic classification. Gaudry (1862-1878), 
Leith Adams (1877-1881), Pohhg (1888-1909), Weithofer (1890), Pilgrim (1905-1926), Soergel (1912-1921), 
Lull (1908), Matthew (1915), Schlesinger (1917-1922), Boule (1920), and Freudenberg (1922) have successively 
discussed and enriched our knowledge of the phylogeny of the Proboscidea but have not expressed this knowledge 
in classification. 

Osborn (1907) began his observations upon the Proboscidea and his survey of the hterature covering the 3'ears 
1735 to 1934; in 1918 (1918.468), he began the special series of studies on classification as expressive of the 
phylogeny of the Proboscidea which has culminated in the present Memoir. 

Osborn, 1934: The advantage of a phylogenetic classification of the Proboscidea is that the superfamilies or 
suborders expressed with the terminal -oidea indicate the four' larger branches into which the Proboscidea are now 
known to be divided ; the family subdivisions corresponding with the terminal -id^e express the five' known subdivi- 
sions, leaving space for the unknown families; the third system of branching, expressed with the subfamily terminal 
-iN^, indicates the seventeen' subfamilies, three of which are now positively known to be as ancient as Lower 



'(Osborn, 19.35) Five larger branches (-oidea), eight families (-ID«), and twenty subfamilies (iN«) are now known. See Appendix of the present Volume. 

19 



20 



OSBORN: THE PROBOSCIDEA 



Oligocene time; beneath these branches are the forty-one or more generic phyla, each of which embraces a very 
long line of specific and subspecific descent. This phylogeny of 1933-1934 is shown in figure 8. 

The chief difficulty with a phylogenetic classification of the Proboscidea, as of the Perissodactyla or other 
orders of mammals, is a technical one, namely, to harmonize it, on the one hand, with the Linnsean "special 
creation" system of nomenclature, and, on the other, with the prevailing usage of family, subfamily, and generic 
nomenclature in mammalian zoology, which systems are alike based on affinities displayed between the existing 
terminal twigs of the branches and subbranches rather than on the affinities of the phylogenetic lines which connect 
the existing twigs with their geologically remote and fundamental ancestral branches. By Linnaean and post- 
Linnaean systematists upwards of fifty' generic names have been applied to the mastodonts and elephants, but there 
is no uniformity in the application of these generic names among paleontologists or zoologists, nor has any other 
principle of revision or of arrangement been worked out by which one may at least begin an advance toward a 
permanent system of nomenclature of this highly important and interesting group, except the present phylogenetic 
system. 




Fig. 3. Diagram showing Osborn's theory so far as developed up to June, 1921, as to the Adaptive Radiation of the Proboscidea. 
After Osborn, 1921.532, p. 233, Fig. 1. Compare diagrams of 1925, 1933-1934, and 1935 below. 



Apart from its theoretic value there is an honest difference of opinion about the practicability of a phylo- 
genetic system of classification ; some of the leading palaeontologists, like W. D. Matthew (letters, February, 1918), 
do not think it is practicable; other colleagues, Hke W. K. Gregory (letters, February, 1918), do not consider that 
the subfamily division (-IN^) can be properly applied to vertical lines of succession. But to Osborn there does not 



'This estimate was made in 1923; at the close of the year 1933 upwards of eighty genera had been listed by the present author (see Chap. I, p. 9); also 
for complete list of genera and species, see Volume 11. 



ORIGIN AND CLASSIFICATION 21 

appear to be any alternative between the continuation of the hopeless confusion which at present prevails in the 
nomenclature and arrangement of the Proboscidea and the absolute illumination and clarity which are thrown 
upon the whole subject by the adoption of the pliylogenetic classification. 

In our present phylogenetic revision the only possible way of straightening out the genera is by first straighten- 
ing out the -phyla in which they belong. The same is true in regard to the species; the species can be understood only 
in their phyletic relations. Consequently the prelude to our phylogenetic classification has been the prolonged 
and very difficult research on the actual characters of the types of genera and of species, upon which the author 
has been closely engaged for the last sixteen years (1918-1934). During this period the author has learned a great 
deal about the phylogeny of the Proboscidea from other observers and from his own observations, but many 
phyla still remain obscure and imperfectly known ; much is to be added to our knowledge by the more intensive 
study of materials already in hand and of the very rich materials still to be discovered, especially in central and 
northern Asia and northern Africa. 

Thus a phylogenetic classification is never final; it is always progressive and adaptive to future discover}'. 
Osborn's classifications of 1921 to 1926 appear to be marked advances upon that of 1918; doubtless it will be 
progressively and perhaps largely modified through future discovery. This is as it should be, because nomenclature 
and classification are the tools of thought, not the masters of thought as some systematists would have it. Other 
phylogenies are those of 1925 (Fig. 7) and of 1933-1934 (Fig. 8) bringing us up to 1935, see the Appendix of 
this Volume. 

With a reasonable rather than a strained regard for priority, nomenclature and classification are the plastic 
vehicles of expression of the actual knowledge of relationships and phylogeny. A reasonable recognition of all the 
great ideas of the past combined with provision for the ideas and the discoveries of the present and of the future 
— these are the watchwords of Phylogenetic Classification. 



2. ORIGIN OF THE PROBOSCIDEA 

As remarked above, phylogenetic classification depends progressively upon the discovery of origins, of 
adaptive radiations, and of phyletic successions. 

The discovery of primary stocks by palfeontologists naturally comes last, long after the zoologists have dis- 
covered recent or existing species. Thus tlie existing elephants have been known by man for thousands of years, 
but the three primary stocks from which they were derived were not discovered until 1901. These primary stocks 
give us the roots and greater branches of our system ; the lesser branches have afforded the progressive knowledge 
of adaptive radiation and the still more elusive knowledge of direct phyletic succession. 

The distinguishing feature of this Memoir, which we trust will be of permanent value, is the separation of 
many direct phyletic lines of descent from each other and the recognition of many mastodonts and elephants 
which are very much aUke in certain characters, but which are still more unlike in other characters and cannot 
possibly be descended one from the other. 



3. THE PRIMARY STOCKS ESTABLISHED BY PRIMARY CHOICE OF FOOD IN 

DIFFERENT HABITATS 

If we watch ungulates while feeding, we observe that they are very fastidious in the choice of food. Conse- 
quently the beginning of adaptive radiation lies in the choice of food and of feeding habitat. There is a great deal 
of truth in Lamarck's idea that the wish precedes the organ. Coadaptation and compensation follow the primary 
choice of food. 

The origin and evolution of the three kinds of Fayum proboscideans as described below (p. 34 to p. 50) 
are only part of the evidence for a long continued antecedent adaptive radiation process beginning in the Upper 
Cretaceous and extending throughout Eocene time, by which the Proboscidea, Uke other hoofed animals, were 
broken up into several great primary stocks, namely: 

An amphibious stock, adapted to rivers and swamps, of limited powers of migration. Represented only 
by the imperfectly known Mceritherium of northern Africa and possibly (Pilgrim) of southern 
Asia in Oligocene time =M(ERITHERIOIDEA 

A southern forest stock adapted to forested lowlands. Represented by the Deinotherium of northern 
and central Africa and of southern Eurasia; known from more abundant remains in Miocene 
to Middle Pliocene time =DEINOTHERIOIDEA 

A northern stock, adapted to lowlands, to savannas, and to forests, with better developed limbs and 
powers of wide migration. Represented from the Lower Oligocene of northern Africa to the Mio- 
cene-Upper Pleistocene of Eurasia and North and South America =MASTODONTOIDEA 

A stegodont-elephant stock, adapted to plains, savannas, and steppes; of browsing (forest) or grazing 
(plains) habits, also with powers of wide migration. Represented from the Lower Pliocene of 
southern Asia to the Pleistocene of North America and the Pleistocene and Recent of southern 
Asia and of Africa' =ELEPHANTOIDEA= 

These four- great primary stocks of Eocene time, and probably other stocks still to be discovered, gave off 
from two to six branches each, so that the Proboscidea, as a whole, from Upper Eocene time onward are not two 
branched, or diphyletic, as formerly expressed in the adoption of only two genera, Elephas and Mastodon, but many 
branched or polyphyletic. Species of Moeritheres and Deinotheres remained in Africa and southern Eurasia; 
whereas the mastodonts and elephants, forest and savanna browsers and grazers of the plains, steppes, and for- 
ests, were the long distance travelers and either from an African or from an Asiatic homeland center of origin 
and adaptive radiation in Eocene time reached all the continents of the world, except AustraUa, in Miocene and 
Pliocene times. 

Always intelligent, resourceful, independent, adaptive, well defended from their enemies of all kinds, we know 
at present of seventeen' entirely distinct main branches of proboscideans several of which persisted into Upper 
Pliocene if not into Pleistocene time, when they were overtaken by the worldwide elimination of the larger 
Mammalia. Only two branches, namely, the Loxodontines and the Elephantines, survive at the present time. 

'(Osborn, 1934) Plio-Pleistocene ancestors of Archidiskodon and Palseoloxodon recently discovered in South Africa (see p. 25). 
'(Osborn, 1935) The STEGODONTOIDEA superfam. nov. separated from the Elephantoidea as a distinct stock. 
^(1935) See footnotes on pages 27 and 30. 

22 



4. ORDINAL, SUBORDINAL-SUPERFAMILY, FAMILY, SUBFAMILY-PHYLETIC, GENERIC- 
PHYLETIC CHARACTERS EMPLOYED IN THE PHYLOGENETIC CLASSIFICATION 

OF THE PROBOSCIDEA 

The Order PROBOSCIDEA is distinguished by the following characters common to all its members: (1) A 
pair of second upper and lower incisors greatly enlarged and originally opposed as feeding organs, a function 
which resulted in the loss of both the first and third pairs of incisors and in the hypertrophy of the second pair, 
P-L; (2) originally tetrabunodont to hexabunodont superior and inferior grinding teeth; (3) originally penta- 
dactyl feet with subequal distribution of weight on all five digits. 





e- 



FlG. 4. FONDAMENTAL AKHANCEMENT OF THE CUTTING TEETH, I'-Ij, NAMELY, THE 
SECOND SDPERIOR AND INFERIOR INCISORS, IN POUR SUBORDERS OR SUFERKAMILIES OF 
THE PROBOSCIDEA. AFTER OsBORN, 1921.515, FIG. 1, P. 3. 

I. Mwrilherium type (MCERITHERIOIDEA): Median incisors, I'-I,, persisting; 
second incisors, I^-Ij, opposed, as in the Rodentia, front encased in enamel. 

II. Deinnlherium type (DEINOTHERIOIDEA): Superior incisors entirely 
aborted; inferior incisors (probably I2) downturned and retroverted. 

III. Rhynchnthenum type (MASTODONTOIDEA): I^-Ij, tusks downturned, 
laterally compressed, partly covered with enamel band (e). 

IV. Elephas type (ELEPHANTOIDEA), includinR Elephants, Loxodonts, Mam- 
moths, etc., with inferior incisors, I2, reduced or aborted, superior inci.snrs, I^, gradually 
losing enamel, typically upturned, used lioth as fighting and as feeding organs. 

On this primary ordinal constitution the Proboscidea radiated adaptively into all habitats known to the her- 
bivorous quadrupeds, namely, in) fluviatile, lowland, swamj), and forest, (h) forest border, (c) plains and upland 
habitats, with their accompanying vegetation, softer and harder kinds of food, and reactions of the feet and limbs 
to the modes of locomotion corresponding to habitats a, b, c. 

SUBORDERS— SUPERFAMILIES AS PROPOSED BY OSBORN 

SuPERFAMiLY Rank. — As regards the subordinal Separation or rank of the four' primary divisions of the Probo- 
scidea which have been discovered up to the present time, these divisions are certainly far more profound than 
those of the four suborders of the Rodentia, namely: I. Sciuromorpha, II. Myomorpha, III. Hystricomorpha, 
IV. Lagomorpha, which are separated from each other not by modifications in the paired incisor, P-Lj, function, 
but by fairly profound muscular adaptations in the mechanism of the jaws. The four primary divisions of the 
Proboscidea are also far more profound than Osborn's main divisions of the Perissodactyla, namely: I. Titano- 
therioidea, II. HipPOiDEA, III. Tapiroidea, IV. Rhinocerotoidea, which are distinguished by profound differ- 
ences in the fundamental pattern of the grinding teeth and by adaptive divergence in the original tetradactyl 
manus and pes. 



'Sec footnotes on pages 19 and 22. 



23 



24 OSBORN: THE PROBOSCIDEA 

Therefore the I. MCERITHERIOIDEA, II. DEINOTHERIOIDEA, III. MASTODONTOIDEA, and IV. 
ELEPHANTOIDEA are separated from each other by adaptive divergences in the function of the upper and 
lower incisor teeth which exceed in their coadaptive results those in the subordinal divisions of any other order 
of ungulates. If the Mceritheres and Deinotheres had lived in the time of Linnaeus, it is doubtful whether he 
would have united them in the same order with the mastodonts and elephants which are of more obvious affinity 
to each other. 

I. MCERITHERIOIDEA. — Renewed study of Moeritherium by Osborn and Matsumoto entirely confirms 
Andrews' original opinion that Moeritherium belongs in the order Proboscidea, as well as Osborn's opinion that it 
stands very far apart from the other proboscideans and is not directly or indirectly ancestral to either of the other 
three groups. The enlargement of the second upper and lower incisor teeth into mutually abrading tusks, girdled 
with enamel, presents a firm ground of affinity with a still unknown primitive Lower Eocene proboscidean stem 
form. There the resemblance ends. These Mceritheres had no proboscis. The face, of brachyopic type, is 
markedly abbreviated. The cranium is elongated. Thus the facial and cranial proportions are analogous to 
those of the Sirenia. The upper grinding teeth are bilophodont, pointing to a tetrabunodont ancestry, and 
different from the bunomastodont grinders of Pnlseomastodon, which point to a hexabunodont ancestry. The 
lower grinders exhibit a rudimentary third crest. Andrews' opinion that the Mceritheres were amphibious is 
probably correct. 

II. DEINOTHERIOIDEA. — All agree that these animals were chiefly browsers and partly amphibious in 
habit, in this respect resembling the Mceritheres but differing in the entire loss of the superior incisor teeth. 
Early loss of the superior tusks released the inferior. In the downturning of the inferior tusks the Deinotheres are 
paralleled by the Rhynchorostrinse among the Mastodontoidea. In .skull form and in limb and foot structure 
the Deinotheres parallel the true proboscideans. They diverge very widely from proboscideans in the evolution 
of the upper and lower grinding teeth. The primitive Deinotheres present simple, bilophodont grinders, similar to 
those of Mceritherium, and are progressively trilophodont. The upper grinders attain a stage which parallels the 
molar pattern of the tapir (Tapirus) among the perissodactyls, but which shows a tendency to the trilophodont 
structure characteristic of the primitive mastodonts and elephants. 

III. MASTODONTOIDEA.— The fundamental character of the front teeth in this superfamily is seen in 
primitive members of the Rhynchorostrinse and Longirostrinse, namely: (1) Second superior incisors enlarged, 
downturned, divergent, with enamel band on outer side only; (2) second inferior incisors downturned (as 
in Deinotheres) with enamel band on outer side (Rhynchorostrin^) or procumbent with no enamel band 
(Longirostrinse). 

The important functional distinction of the Mastodontoidea is that for a very long period of time the upper 
tusks abraded the outer side of the lower tusks ; this probably explains the retention of the superior enamel band. 
In certain lines (Longirostrinse) the procumbent lower incisors persist and the upper incisors retain their primitive 
downcurved position as in Phiomia. In other lines (Mastodontinse, Brevirostrinse) the lower incisors practically 
cease to function; the upper incisors finally turn upward and inward, but may retain the enamel band for a 
long period (Mastodontinse, ^de Schlesinger, and Notorostrinse, vide Cordillerion andium). 

A distinctive character of the grinding teeth of the Mastodontoidea is evidence of the descent from a 
tetrabunodont ancestral grinder (i. e., without intermediate tubercles or 'conules'), as is the case with the tetra- 
bunodont ancestral type of Mceritheres and Deinotheres. The rudiments of ancestral conules gave rise to various 



ORIGIN AND CLASSIFICATION 



25 



trefoils or paired median outgrowths or crests, so characteristic of all the Bunomastodontidae whether beak jawed 
(Rhynchorostrinae), or long jawed (Longirostrinae), or short jawed (Brevirostrinff). In each of these subfamilies 
the grinders independently undergo a more or less closely parallel evolution, evolving single trefoils in Upper 
Oligocene and in Miocene time, and double trefoils in Pliocene time. 

Unhke the Mccritheres and Deinotheres, the three intermediate molars (i. e., fourth premolar and first and 
second true molars) invariably become trilophodont, while the third true molars become tetralophodont. At this 
point there is a divergence into (1) Mastodontida^, ])urely forest Uving, brachyodont, with simply crested teeth, 
in which the intermediate molars are persistently trilophodont, with arrested trefoils, and into (2) Bunomastodon- 
tidae, which pass into tetralophodonty and polylophodonty in some lines, with evolving trefoils. The grinder 
evolution is adapted to a leaf-browsing habit, in distinction to the prevaihng grazing habit developed among the 
elephantoids. The development of hypsodonty, and choerodonty (Schlesinger), among these (longirostrine and 
brevirostrine) browsers is analogous to that in the hippopotami and the hypsodont suilhnes. 

IV. ELEPHANTOIDEA. — One prime distinction in this superfaniily is the very early complete loss of the 
lower incisor teeth, accompanied by the early development of the upper incisors into horizontal or upturned tusks 
finally devoid of enamel except at the tips in the young stage. Vestigial enamel bands are recorded in early stages 
of the Stegodonts. A second distinctive character is the absence of conule development into trefoils, so character- 
istic of the mastodontoids, and the early tendency to form evenly transverse, more or less mamillate, crests which 
become in the highest degree hypsodont and polylophodont in adaptation to chiefly grazing habits. 

It has been assumed by practically all palaeontologists that the Elephants were descended from the Stegodonts. 
This assumption now proves to be erroneous, for neither the Stegodon grinding tooth with enamel valleys closed at 
the bottom, nor the Stegodon cranium with its extremely short face, can give rise to the elephantoid molar or the 
face of the elephantoid cranium. The earliest elephants known (1934) are the primitive species of Archidiskodon 
and of Palxoloxodon of the PHo-Pleistocene of South Africa. The Stegodonts were of independent origin and 
formed an independent parallel branch terminating in the highly specialized Elephas [ = Stegodon] aurorge Matsu- 
moto from the Upper(?) Phocene of Mt. Tomuro, Japan, now (1935) separated by Osborn as the STEGODONT- 
OIDEA superfam. nov. 




Chief Head and Dental Forms of four of the Suborders (I-V) ok the Proboscidea 
FiR. 5. (1) Moeritherioidea, one-twelfth natural size. (2) Deinotherioidea, (3) Mastodontoidea, and (4) Elephantoidea, all one-fiftieth natural size 



26 OSBORN: THE PROBOSCIDEA 

FAMILIES OF PROBOSCIDEA AS ORIGINALLY PROPOSED 

Only one family up to recent time' has been proposed in three of the four' suborders or superfamilies of the 
Proboscidea. In the Mastodontoidea there are two^ families, the Mastodontidae and the Bunomastodontidse. 

I. M(ERITHERIOIDEA Osborn, 1921.515, p. 2. Family: Moeritheriidse Andrews, 1906, p. 99. 

II. DINOTHERIOIDEA Osborn, 1921.515, p. 2 [ = Deinotherioidea of the present Memoir]. Family: 
Curtognati [ = Curtognathidae] Kaup, 1833.4, p. 516; Dinotheridse Bonaparte, 1845 (fide Palmer, 
1904, p. 738), Dinotheriidae Bonaparte, 1850; Dinotheridse Zittel, 1891-1893, p. 454; Dino- 
theriidae Osborn, 1918.468, p. 134 [ = Curtognathidae of the present Memoir]. 

Note. — The superfamily Deinotherioidea as well as the family Deinotheriidse are partly anticipated in the 
term Curtognati Kaup (Kaup, Bull. Soc. Geol. de France, V, 18.34, p. 444): "Curtognati. — M. Kaup a regu de 
nouveaux ossemens d'Eppelsheim, et a reconnu que le Dinotherium medium etait veritablement bien different du 
D. giganteum. M. Kaup a pu faire aussi quelques corrections et additions a sa description du Dinotherium. II 
pense que ses deux enormes defenses ne lui servaient pas seulement pour extraire des racines de la terra, mais 
encore, comme k la morse, pour I'aider a mouvoir son corps si lourd. D'apres la forme des os intermaxillaires, cet 
animal devait avoir ime trompe pour porter la nourriture a sa bouche. II est a placer entre les Mastodontes et les 
Bradypus, et il formerait une famille particuliere a laquelle M. Kaup donne le nom de Curtognati. Elle serait 
caracterisee par la machoire infdrieure courbee vers le bas, et les deux defenses dirigees vers le bas et en arriere 
(AT. Jahrh. /. Mineral. Geognosie, etc., 1833 [1833.4], cah. 5, p. 509 avec 1 pi.)." 

III. MASTODONTOIDEA O.sborn, 1921.515, p. 2. Family: Mastodonada? Gray, 1821, p. 306 (fide Palmer 
1904, p. 752); Mastodontidae Girard, 1852, pp. 326, 328; Mastodontidae Osborn, 1918.468, p. 
134, including the Palaeomastodontidae of Andrews, 1906, p. 130. 

Osborn has recently added the following family: Bunomastodontidae, 1921.515, p. 2, to 
include (1921-1922) all the bunomastodonts which develop the intermediate conules into, first, 
single and, then, double trefoils; also a very important branch with several side branches of mas- 
todonts in which the lophs are transformed into well-defined transverse crests paralleling the 
zygolophodont condition of the Stegodonts, and perhaps giving oflf a branch into the Stegodon- 
tinae. The Bunomastodontidse constitute a very distinct family, from which arose several sub- 
families also widely separated from each other. 

IV. ELEPHANTOIDEA Osborn, 1921.515, p. 2. Family: Elephantida; Gray, 1821, p. 305 (fide Palmer, 
1904), p. 740 ( = Elephasidese Lesson, 1842, p. 156; Elephantida; Bonaparte, 1838, p. 112, and 
1850; Girard, 1852; Zittel, 1891, p. 458; Osborn, 1910.346, p. 558). 

The above five families of the Proboscidea are distinguished from each other by profound modifications 
in the fundamental structure of the grinding teeth. These differences in fundamental structure are almost as 
profound as those which separate the eight families of the order Perissodactyla, or the seven superfamilies of the 
order Rodentia, e. g., Aplodontoidea, Scixjroidea, Castoroidea, Geomyoidea, Myoidea, Anomaluroidea, 
Bathyergoidea. 

It is noteworthy that the giant Proboscidea closely parallel the humble Rodentia in the profound modes of 
evolution of the grinding teeth. 



'See footnotes on pages 19 and 22. 

''(Osborn, 1934) To these two families have been added the Serridentidae fam. nov. and the Huinboldtidse fam. nov. 



ORIGIN AND CLASSIFICATION 



27 



SUBFAMILIES OF PROBOSCIDEA AS ORIGINALLY PROPOSED 

The following seventeen' subfamily divisions of the Proboscidea have successively been proposed, as ex- 
plained in full chronological detail in Chapters II-XX of the present Memoir: 

I. MCERITHERIOIDEA: 



II. DINOTHERIOIDEA: 



III. MASTODONTOIDEA: 



1. Moerithoriini — Herluf Winge, 1906, p. 172; Moeritheriina? Winge- 
Osborn, 1923.601, p. 1. 

2. Dinotherina — Bonaparte, 1841, p. 253, Dinotheriina^Bonaparte, 1850; 
Dinotheriini — Herluf Winge, 1906, p. 172; Dinotheriinae— Osborn, 
1910.346, p. 558 [ = Dcinotheriinae of the present Memoir]. 

3. Mastodontina— Brandt, 1869, p. 35; Mastodontinae— Osborn, 1910.346, 
p. 558. 

Bunomastodontinae — Osborn, 1918.468, p. 134. Replaced by Buno- 
ma.stodontidae — 0.sborn, 1921.515, p. 2. 

4. Longirostrinse — Osborn, 1918.468, p. 136. 

5. Rhynchorostrinae — Osborn, 1918.468, p. 136. 

6. Brevirostrinae — Osborn, 1918.468, p. 136. 

7. Serridentinae {T. serridens phylum, Osborn, 1921.515, p. 8) — Osborn, 
1921.526, p. 330. 

8. Notorostrinae — Osborn, 1921.526, p. 330, distinguished and defined. 

9. Zygolophodontinae — Osborn, 1923.601, p. 1, to include the species: 
Zygolophodun tapiroides, Z. turicensis, Z. virgatidens, Z. borsoni, 
and Z. pyrenaicus, as more fully defined below, which were 
mistakenly confused with the Serridentinae by Osborn (1921.515). 

10. Platybelodontinaj— Boris.siak, 1928, p. 119. 

11. Amebelodontinae— Barbour, 1929.2, p. 139. 

12. Tetralophodontinae— van der Maarel, 1932, p. 108. 
13 Humboldtinse — subfam. nov. 

14. Stegodontina? — Osborn, 1918.468, p. 135, to include the Stegodonts of 

southern Asia. 

15. Loxodontinae — Osborn, 1918.468, p. 135, to include the phylum of 

African elephants. 

16. Elephantina — Bonaparte, 1838, p. 112, 1850; Elephantini — Winge, 

1906, p. 172; Elephantinae— Osborn, 1910.346, p. 558, 1918.468, 
p. 135. 

Euelephantinae — Osborn, 1918.468, p. 136. Abandoned as invalid, 
because the genus Euelephas is invalid. 

17. Mammontinae — Osborn, 1921.515, pp. 1, 14. 

Great Antiquity.— The above seventeen^ subfamilies have been separated from each other for very long periods 
of geologic time, many of them since Lower and Middle Eocene times, others since Oligocene time. This very 
ancient geologic and adaptive habitudinal radiation is especially characteristic of the subfamilies included within 
the Mceritherioidea and the Deinothorioidea as well as of the eleven' subfamilies included within the Mastodon- 
toidea. As regards the Elepliantoidea, it would appear that the Stegodontin;?, the Lo.xodontinae, and the 
Elephantinjr had already widely radiated in Miocene time, but we do not as yet know when the Mammontinae 
were given off from their closest relatives, the Elephantinae. 

'(1935) See Appendix of tlic picsi'iit Volumo I for Pal.bomastodontin* subfam. nov., based on the genus PaUeomastodon Andrews, and Gnathabelo- 
DONT1N.E, based on GnalhaModon Barbour, also tin- removal of Stegolnphndon from the Steijodont idac, subfamily Stcgodontinte, to the new subfamily Stego 
LOPHODONTIN.E, making twenty subfamilies in all, fourteen of which are includi'd within the superfamily Mastodontoidca. 



IV. ELEPHANTOIDEA: 



28 OSBORN: THE PROBOSCIDEA 

Subfamily Divergence. — Whereas the families of Proboscidea are distinguished from each other by pro- 
found modifications in the fundamental structure of the grinding teeth, the subfamilies of Proboscidea are distin- 
guished mainly by secondary adaptations of the grinding teeth correlated with profound adaptations in the upper 
and lower incisor teeth. For example, compare the lower incisor teeth in the Mastodontinse, the Longirostrinse, 
the Rhynchorostrinse, the Brevirostrinse, and the Notorostrinse and observe how profoundly different each is from 
the other. This adaptation or loss of the lower incisor teeth is correlated with important adaptations of the upper 
teeth and finally with the structure of the grinding teeth. Here the principle of adaptive compensation is observed ; 
e. g., in the Rhynchorostrinse, with their well-developed lower tusks, the grinding teeth remain excessively simple, 
whereas in the Brevirostrinse, with the loss of the lower tusks, the grinding teeth become excessively complex. 

The phylogenetic basis of this seventeen-fold^ subfamily division is to separate clearly every direct line of 
descent of species and genera from every other direct line of descent; the subfamilies are observed as vertically 
ascending ancestral lines which may be subdivided into one or more generic stages. 

Contrast with Cope. — This vertical phylogenetic use of the subfamily by Osborn (1918.468, pp. 133, 134) 
presents the most direct contrast to the system of Cope: "Cope's family classifications were morphological and 
horizontal rather than phylogenetic and geological. Finding one or more single characters possessed in common 
at certain horizontal periods of geologic time by mammals in different lines of evolutionary descent, he seized on 
these common characters as convenient keys to classification. ... I have reached the opinion that Cope's method 
of morphological classification is untenable, that the only true and permanent classification is phylogenetic. . . . 
Sometimes the subfamilies proposed by zoologists conform to the phyletic lines discovered by paleontologists; 
sometimes they do not." 

GENERIC VALUES IN PROBOSCIDEA 
Inasmuch as the Order PROBOSCIDEA is distinguished by profound adaptations of the incisor teeth, the 
Suborders by profound modifications of the incisors, P-L, the Families by profound adaptations of the grinding 
teeth, the Subfamilies by secondary adaptations of the upper and lower incisor teeth and compensatory adapta- 
tions of the grinding teeth, it remains to distinguish the Genera by successive stages in the evolution of these second- 
ary and compensatory adaptations. 

For example, the four genera of the phylum Mastodontinse in descending order: 

Pleistocene Genus Mastodon: Lower incisor tusks small, variable; upper tusks large, without enamel. 

Pliocene Genvs Pliomastodon: Lower incisor tusks intermediate ; upper tusks rounded, without enamel. 

Miocene Genus Miomastodon: Lower incisor tusks intermediate; upper tusks upcurved with broad 
enamel band. 

OUgocene Genus Palaeomastodon: Lower and upper tusks unknown; upper tusks probably downturned 
with broad enamel band. 

Or, for example, the three genera of the phyla Tetralophodontinse, Longirostrinse, and Amebelodontinse in 
descending order: 

Miocene Genus Tetralophodon: Lower tusks procumbent; upper tusks downturned. Intermediate 
grinders tetralophodont. 

Miocene Genus Trilophodon: Upper tusks downturned with enamel band. Intermediate molars tri- 
lophodont. 

OUgocene Genus Phiomia: Lower incisors procumbent without enamel; upper incisors fighting weapons 
with enamel band. Intermediate molars trilophodont. 



'See footnotes on pages 19 and 27. 



ORIGIN AND CLASSIFICATION 



29 



Thus in the phylogenetic classification of the Proboscidea the Genus embraces and distinguishes the successive 
stages of adaptation and evolution of single or multiple characters, as observed in ascending geologic horizons. 

Among Genera the phenomena of paralleh.sm, of convergence, and of homoplasy are especially rife. For 
example, the superior incisors of Mastodon, of Ananctis, and of certain species of Curdillerion tend to lose their 
enamel and to become upturned and outwardly arched so as to mimic each other closely, and there occurs a 
correlated homomorphy or parallelism in skull structure. Most confusing, however, to the early students (such 
as Falconer) of the Proboscidea was the parallel evolution of the trilophodont intermediate molars, e. g.. 
Dp 3 t, M 1 ^, M 2 's, which arose separately in Phiomia, in Palaeomastodon (rudimentary), in Anancus, and in 
Stegodon. In fact, it is certain that in many distinct branches of proboscideans trilophodonty was independently 
developed. Similarly the tetralophodont condition of the intermediate molars arose independently in Tetralopho- 
don and in Anancus, also in the ancestral Stegodonts and ancestors of all the true Elephantoidea. The third 
form of generic parallehsm is the independent evolution in all the genera of the bunomastodonts of single trefoils 
followed by double trefoils; for example, as in the transition from Trilophodon angustidens (single trefoils) to Tetra- 
lophodon longirostris (double trefoils), or in the transition from Cordillerionandium (single trefoils) to Cuvieronius 
humboldtii (double trefoils). 



/OO Fathom.. Pllo -Pleistocene.. EquaL-area 146rldMgp- 1935. 




\m^f€* ^r ^^ \ m ^ i«» 



^^ Fossil Elephants ondMastodonts. 1934- 



Y\s.. 6. Worldwide Distkihition ok the Pkoboscidea in Past and Present Time 
African Elephant: After Blanc, 1897; Booage, 1890; Buckley, 1876; Chapman, 1868; Chubb, 1909-1919; Claridge, 1915; Cox, 1900; 

Heilpiin 1887- Hippolyte, 1907; Johnston, 1907; Mavd.m, 1932; Roosevelt and Heller, 1914; SchulUe, 1907; Sclater, 1900. 

Indian Ei.ei'iiant: .\fter Bartholomew, 1911; Hlanford, 1888-1891; Blyth, 1872; Bishop, 1921; Champion, 1928; Hornaday, 1885: 

Hunter, 1868; .lerdon, 1874; l.aufor, 102.J; l.yilekker, liXK) 1908; Sanderson, 1907; Sclater, 1899. 



5. HARMONY OF LINN^EAN AND PHYLOGENETIC CLASSIFICATION EXPRESSING 
OUR KNOWLEDGE UP TO THE CLOSE OF THE YEAR 1933 

Thus by applying the principle of adaptive radiation in geologic time we find it is practicable to develop 
consistently a phylogenetic classification out of the Linnsean system which included within the order Proboscidea 
only the single species Elephas indicus. This phylogenetic system provides a consistent terminology for the main 
stock, its branches and subbranches, parts of which may be Anglicized for convenience as follows: 

CLASSIFICATION FIRST ADOPTED (1933) BY OSBORN IN THE PRESENT MEMOIR 

Osborn, 1935: This classification of the year 1933 has been recently modified by successive discoveries up to July of the year 1935. 
Accordingly the reader is referred to page 685 et seq, for the classification of 1935 which introduces new forms and new concepts of 
relationship and lines of descent. 

Order PROBOSCIDEA (Proboscideans) — Illiger, 1811, p. 96. Upper Eocene to Recent time. 

Suborder I. MCERITHERIOIDEA (Moeritheres)— Osborn, 1921.515, p. 2. A primitive, amphibious, 
swamp-loving stock. 

Family MCERITHERIID^— Andrews, 1906, p. 99; Osborn, 1921.515, p. 1. 

Subfamily Mceritheriin^ Winge-Osborn — Osborn, 1923.601, p. 1 ( = Moeritheriini Winge, 
1906, p. 172). Upper Eocene, Lower Oligocene of Africa. 

Suborder II. DEINOTHERIOIDEA (Deinotheres)— Osborn, 1921.515, p. 2. Specialized, fluviatile 
or river border proboscideans. Mio-Pliocene. 

Family CURTOGNATHID^— Kaup, 1833.4, p. 516. 

Subfamily Deinotheriin^ — Osborn, 1910.346, p. 558 ( = Dinotherina Bonaparte, 1841, p. 
253, Dinotheriina Bonaparte, 1850; Dhiotheriini Winge, 1906, p. 172). Large, river- 
border, fluviatile forms of southern Eurasia; throughout the Miocene to the close of the 
Pliocene, into the Pleistocene. 

Suborder III. MASTODONTOIDEA' (Mastodonts and Bunomastodonts)— Osborn, 1921.515, p. 2. 
Chiefly forest and savanna browsers. 

Family MASTODONTIDyE ("True Mastodonts")— Girard, 1852, pp. 326, 328 ( = Mastodontida? 
Osborn, 1918.468, p. 134), including Palseomastodontidae of Andrews, 1906, p. 130. Chiefly 
browsers in the north temperate and boreal forests. 

Subfamily Mastodontin^ — Osborn, 1910.346, p. 558 ( = Mastodontina Brandt, 1869, p. 35). 
Springing from unknown African ancestors; Oligocene of North Africa; including Mio- 
mastodon and Pliomastodon (Miocene and Pliocene of Holarctica), and Mastodon americanus 
(Pleistocene forests of North America). Persistent forest browsers. Grinders tetrabun- 
odont, sublophodont, lacking trefoils. 

Subfamily Zygolophodontin^ ("Yoke-toothed Mastodonts") — Osborn, 1923.601, p. 1. 
Browsers of the temperate forests of Europe and Asia, including Turicius turicensis, T. 
tapiroides, T. virgatidens, Zygolophodon pyrenaicus, terminating with Z. horsoni. Miocene 
to Upper Pliocene. Grinding teeth polybunodont to purely lophodont, lacking trefoils. 

■(1935) To the Mastodnntoidca have been added two new families, the Serridentid^ and the Humboldtid.b, and two new subfamilies, the Pal.eomas- 
TODONTiN-E and tlie Steqolophodontin^e (see Appendix of the present Volume for final classification). 

30 



ORIGIN AND CLASSIFICATION 31 

Family BUNOMASTODONTID^ ("True -Bunomastodonts")— Osborn, 1921.515, p. 2. Arising 
from forms similar to Phiomia of North Africa, probably lowland and swamp browsers, 
radiating into nine' subfamilies. Oligocene to Lower Pleistocene. Grinders provided 
with single or double trefoils — hence "bunomastodont." 

Subfamily Longirostrin^ ("Long-jawed Bunomastodonts") — Osborn, 1918. 408, p. 136. 
Long-jawed typical bunomastodonts, arising in North Africa {Phiomia'), migrating all over 
North Africa, southern Europe, Asia, and North America north of Mexico; including 
Trilophudun. Grinding teeth originally bunodont, trilophodont. Central conules. 

Subfamily Amebelodontin.e (typical "Shovel-tuskers")— Barbour, 1929.2, p. 139. Includ- 
ing the shovel-tuskers of Nebraska. Typified by Amebelodon fricki Barbour. 

Tetralophodontin^ ("Four-ridge-crested Bunomastodonts") — van der Maarel, 1932, 
p. 108. Typified by 'Mastodon' longiroslriK Kaup and Tetralophodon bumiajuensis, van 
der Maarel. Second molars with four ridge-crests, i.e., tetralophodont. 

Subfamily Serridentin^ ("Medium-jawed Bunomastodonts") — Osborn, 1921.526, p. 330. 
Modifying the trefoils to the outer walls. Grinding teeth sublophodont; upper and lower 
incisor teeth as in the Longirostrinse. Upper Miocene to Upper Pliocene of North America 
(typified by Serridentiruis serridens and S. produclus) ; Upper Pliocene of Texas, of Florida, 
and of Guatemala; Miocene of Mongolia {Serridentinus mongoliensis) ; Upper Miocene 
of India {Serridentinus brorvni); PUocene of North China {Serridentinus lydekkeri). 

PLATYBELODONTiNyE ( "Flat-tuskei's ")— Borissiak, 1928, p. 119. Typified by Plaiybcludon 
danovi Borissiak, including P. grangeri of Mongolia and Torynobelodon barnumbrowni of 
Nebraska. Extremely broad shovel-tusks. 

Subfamily Rhynchorostrin^ ("Beak-jawed Bunomastodonts")— Osborn, 1918.468, p. 136. 
With powerful downturned upper and lower tusks. Known only in the southern United 
States and northern Mexico. Typified by Rhynclwtherium tlascalse of Mexico, R. euhypo- 
don, R. brevidens. 

Subfamily Notorostrin^ (Notorostrines)— Osborn, 1921.526, p. 330. First known in the 
Andean region of South America {Cordillerion andium) ; subsequently discovered in the 
southern United States and Mexico. Distinguished by the loss of the lower tusks, the 
abbreviation of the lower jaw, the spiral tusks (C andium). 

HuMBOLDTiN^ ( "Humboldtines")— new subfamily described in the present Memoir. Typified 
by Cuvieronius humboldtii with straiglit or upturned tusks without enamel; molar enamel 
ptychoid, highly fokled. Includes Siegomasludon mirificus. 

Subfamily Brevirostrin.e ("Short-jawed Bunomastodonts")— Osborn, 1918.468, p. 136. 
Springing from the same stock as the Longirostrinip; distinguished by the rapid abbrevia- 
tion of the lower jaw, by the early loss of the inferior tusks, by the hypsodonty and laby- 
rinthine multiple trefoil pattern of the grinding teeth. First known in Europe {Anancus 
arvernensis) , thence wandering into Asia (Peulalophodon sivalensis, Syncunolophus dhok- 
pathanensis) ; not known to enter North America. 

'(Osborn, 1935) Seven subfamilies by removal of the S<-rrici.-ntiniP, Platybelodontinsp, ami Ilumboldtinoc iiiui the addition of the Gnathabelodontina.-. 
-Y'/a'jwia now removed to the Amebelodontina; (see .\ppendix of the present Volume for (in:il classification). 



32 



OSBORN: THE PROBOSCIDEA 



Suborder IV. ELEPHANTOIDEA' (Elephants and Mammoths of Africa, Eurasia, and North 
America) — Osborn, 1921.515, p. 2. Browsers (Stegodontinse), chiefly tree browsers 
(Loxodontinse), general browsers and grazers (Elephantinse), browsers and grazers 
(Mammontinse). 

Family ELEPHANTID^— Gray, 1821, p. 305 (fide Palmer, 1904, p. 740); Osborn, 1910.346, p. 558. 

Subfamily Stegodontin^ (Stegodonts) — Osborn, 1918.468, p. 135. Original members closely 
similar to, and perhaps identical with, certain of the Zygolophodontinae of Europe. Dwellers 
in the very warm forest and savanna regions of southern Eurasia and western India. 
Grinders rapidly multiplying crests, but persistently short crowned (brachyodont) . Known 
from the Middle Pliocene (Stegolophodon cautleyi) to the Lower Pleistocene {Stegodon 
insignis, S. ganesa) of India. 

Subfamily Loxodontinse (Loxodonts) — Osborn, 1918.468, p. 135. Distinguished by long 
narrow grinding teeth with relatively few crests; embracing Loxodonta africana and all 
the species of African elephants, the cranial varieties of Loxodonta antiqua [ — Hespero- 
Zoxodon], Upper Pliocene to Upper Pleistocene, Loxodonta namadica [ = Palseoloxodon] 
of India which radiated widely over Africa and wandered all over southern Eurasia, 
attaining the tallest stature known among the elephants ; also the dwarfed insular forms 
Loxodonta (Pilgrimia) falconeri, Loxodonta {Pilgrimia) melitensis, Loxodonta (Pilgrimia) 
creticus, Loxodonta (Pilgrimia) Cypriotes, etc. 

Subfamily Mammontin^ (Mammoths) — Osborn, 1921.515, p. 1. Of close original affinity to 
the Elephantinae, including (a) the southern mammoths Archidiskodon planifrons and A. 
meridionalis of southern Eurasia, A. imperator of North America, all with broad-plated 
teeth and few crests; also (h) the northern mammoths which apparently include Par- 
elephas trogontherii of western Europe, P. columbi and P. jeffersonii of North America, and 
the widespread woolly mammoth (Mammonteus primigenius) of the northern steppes. 




ADAPTIVE RADIATION OF THE PROBOSCIDEA 1525 H.F.O. 

Fig. 7. Diagram showing the known and unknown stages in the Adaptive Radiation of the Proboscidea up to April, 1925. After 
Osborn, 1925.662, p 21, fig. 2. Compare previous diagram of 1921 (Fig. .3 of the present Chapter II), also final diagrams of 1933-1934 
(Fig. 8 opposite) and 1935 (Appendix). 



"See footnote on page 22. 



ORIGIN AND CLASSIFICATION 



33 



Subfamily Elephantine ("True Elephants")— Osborn, 1910.346, p. 558 ( = Elephantina 
Bonaparte, 1838, ]). 112, 1850; Elephantini Winge, 1906, p. 172). Typified by Elephas 
indicus Linn, of India, wliich apparently evolved during Tertiary and Lower Pleistocene 
times in northern Eurasia and did not appear in southern Asia (India) and Asia Minor 
until the Upper Pleistocene. 

To those monophyletists who cUng to Mastodon-Stegodon-Elephas descent, this seventeen-foW subfamily branch- 
ing of the proboscideans is very difficult to credit until we examine in detail the anatomy of the known genera and 
species of which these branches are composed and which we soon realize make up a very small proportion of the 
total number of genera and species entering into this titanic order from its origin in Lower Eocene time to its great 
climax in Upper Pliocene and Pleistocene times. The contrast between the kn(jwn and the unknown is presented 
graphically in the accompanying diagrams (Figs. 7 and 8). 

In the succeeding pages of this Memoir the author will demonstrate that this Adaptive Radiation and Phylo- 
genetic Classification are not imaginary but are based on the solid ground of our present knowledge. This strong 
statement needs several reservations, among them, the ancestry of the Elephantidff, now traced to the summit 
of the Pliocene only. 




Adaptive Radiation of the Probo.scidea (1933-1934) 
Fig. S. DiaKiam showing tlic known ailaptivi' liidiation of the 37-41 gonoiic pliyla i>f the I'roboscidea as discovered up to the year 1934. Compare 
previous diagrams (Figs. 3 and 7), also pliylogonetic diagrams of 193.') in the App<'ndix. 



'See footnote on page 27. 



II. OSBORN'S THEORY (1900) AS TO THE AFRICAN ORIGIN OR CENTER OF ADAPTIVE 

RADIATION OF THE PROBOSCIDEA 

In 1899-1900 Osborn, Tullberg, and Stehlin independently advanced the theory that Africa was the con- 
tinent on which the Proboscidea originated. The original homeland of the Proboscidea is no longer in doubt; 
the evidence (1933) favors the geologically ancient continent of Africa. 

Osborn (1900).— Until the end of the nineteenth century Asia was favored as the homeland of the Proboscidea 
and it is still possible, although not at present probable, that the order originated there. The reasons why Asia is 
not at present favored as the Proboscidea center are threefold, namely: (1) The total absence of Proboscidea in 
the known Upper Eocene and Lower Oligocene exposures; (2) the apparent derivation of the earliest Asiatic 
Proboscidea from ancestral African prototypes; (3) the Miocene appearance of African Proboscidea in Europe and 
North America. If the Proboscidea had originated in Asia in Eocene time they probably would have migrated into 
North America in Eocene or at least in Oligocene time, since the Proboscidea were great travelers. If not the 
original homeland or primary center, Asia certainly became a secondary homeland in which occurred the chief 
evolution and adaptive radiation of the family Elephantidse. Current exploration in central Asia will probably 
clear up this whole problem. 

Theory of the African Center. — On the contrary, Africa, which remained the "dark continent" of 
palaeontology until the beginning of the twentieth century, was seldom regarded as a very important center of 
evolution and adaptive radiation of the Mammalia, despite some excellent suggestions by Rutimeyer (1888) on 
this subject. The first palaeontologists and zoologists to make a bold claim for Africa were Osborn (1900), Tullberg 
(1899), and StehHn (1900). On this subject we may refer to Osborn's addresses of 1900 (April, 1900.182, "The 
Geological and Faunal Relations of Europe and America during the Tertiary Period and the Theory of the Suc- 
cessive Invasions of an African Fauna," and July, 1900.187, "Correlation between Tertiary Mammal Horizons of 
Europe and America"). The following citation is from Osborn's July article (1900.187, pp. 56-58) : 



III. OSBORN'S THEORY OF SUCCESSIVE INVASIONS OF AN AFRICAN FAUNA INTO EUROPE 

In Europe there are in the upper Eocene two classes of animals, first, those which have their ancestors in the older rocks; 
second, the class including certain highly specialized animals which have no ancestors in the older rocks — among these, perhaps, 
are the peculiar flying rodents or Anomaluridx, now confined to Africa, and secondly the highly specialized even-toed ruminant 
types — the anoplotheres, xiphodonts and others, the discovery of which in the Gypse near Paris Cuvier has made famous. It 
is tempting to imagine that these animals did not evolve in Europe but that they represent what may be called the first inva- 
sion of Europe by African types from the Ethiopian region. 

It is a curious fact that the African continent as a great theater of adaptive radiation of Mammalia has not been sufficiently 
considered. It is true that it is the dark continent of palaeontology for it has practically no fossil mammal history; but it by 
no means follows that the Mammalia did not enjoy there an extensive evolution. 

Although it is quite probable that this idea has been advanced before, most writers speak mainly or exclusively of the 
invasion of Africa by European types. Blanford and Allen it is true have especially dwelt upon the likeness of the Oriental and 
Ethiopian fauna but not in connection with its antecedent cause. This cause I beUeve to have been mainly an invasion from 
south to north correlated with the northern extension of Ethiopian cUmate and flora during the Middle Tertiary. It is in a less 
measure due to a migration from north to south. Let us therefore clearly set forth the hypothesis of the Ethiopian region or 
South Africa as a great center of independent evolution and as the source of successive northward migrations of animals, some of 
which ultimately reached even the extremity of South America — I refer to the Mastodons. This hypothesis is clearly implied 
if not stated by Blanford in 1876 in his paper upon the African element in the fauna of India. 

34 



ORIGIN AND CLASSIFICATION 



35 



The first of these migrations wo may suppose brought in certain highly specialized ruminants of the upper Eocene, the 
anomalurcs or pppuliar flying rfwlcnts of Africa; with this invasion may have come the pangolins and aard varks, and possibly 
certain armadillos, Ddnypudidw. if M. I'ilhol's idcntificatitin of Xerrodasypus is correct. A second invasion of great distinct- 
ness may be that which marks the beginning of the Miocene when the mastodons and dinotheres first appear in Europe, also the 
earliest of the antelopes. A third invasion may be represented in the ba.se of the Pliocene by the increasing number of antelopes, 
the great girtifiVs of the /Egean plateau, and in tiie upper Pliocene by the hippopotami. With these forms came the rhinoceroses 
with no incisor or cutting teeth, similar to the smaller African rhinoceros, li. hicornis. Another recently discovered African 
immigrant upon the Island of Samos in the .Egean plateau is Pliohyrax or Leptodon, a very large member of the Hyracoidea, 
probably aquatic in its habits, inilicating that this order enjoyed an extensive adaptive radiation in Tertiary times. 

It thus appears that the Proboscidia, Hyracoidea, certain cdentata, the antelopes, the giraffes, the hippopotami, the 
most specialized ruminants, and among the roilents, the anomalurcs, dormice, and jerboas, among monkeys the baboons, may all 
have enjoyed their original adaptive radiation in Africa; that they survived after the glacial period, only in the Oriental or 
Indo-Malayan region, and that this accounts for the marked connnunity of fauna between this region and the Ethiopian as 
observed by Blanford and Allen. 

In Ills April and July, 1900, articles, Osborn printed two world maps (Figs. 9, 10) illustrating his theories 
regarding Africa and Asia; the Africa theory of Osborn was confirmed in 1901; the Asia theory of Osborn was 
confirmed in 1922. 

ZoooEOOR.\PHic Re.\lms a.\d Reqions op the M.\mmali.\ 




Fig. 9. Division of the world into three Realms (Huxley) and 
eight main Geographical Regions. The continental platform is 
raised to the 200-meter line showing the main Tertiary land con- 
nections. After Osborn, 1900.1S7, Fig. I; compare Osborn, 1900.182, 
Chart II. 






Imeditofi. Cheinpten, Crtodopia. ' ^trnhfora 



Qdontoceli 



PhmatesTiMe^donla , _^ 

'Umb/ypoda, Cwidyhrthra^^i^rissodactifh, /iaqrlcpoda ^rtiodxtfla ^^ 

w^ L Archaeoceii ^— *- 

^ii y{'^ f^isixcceti 




Fig. 10. Orders of mammals as placed liy Osborn in 1900 in their 
hypothetical chief centers of adaptive radiation during the Tertiary Period. 
The Pkoboscidea are pla'ced very near the point of actual discovery in 1901 
of their Faydm ancestors, .\fter Osborn, 1900.187, Fig. Ill; compare 
Osborn, 1900.182, Chart IV. 



As regards the Proboseidca and Hyracoidea, llu- aliove geographic charts of 1900 were conlirnud by Charles W. .\ndrcws in 1901. As regards central 
Eurasia as the chief center of radiation of fourteen orders of mammals, the present Central Asiatic Expeditions of the .\merican Museum are strongly confirma- 
tory. .\s regards the Edentata, Lilopterna, Toxodonlia, and Typotheria, .South .America remains the chief center. As regards the Sircnia and Cetacea, it 
now appears proliable that they also originated on the .\frican continental shorelines. 



This African or Ethiopian center theory of 1899-1900 was destined to enjoy almost immediate confirmation 
through the discovery in 1901 of the jjreviously unknown Ujjiier Eocene-Oligocene land fauna of North Africa. 
This Fa.vtlm region had yielded between 1879-1885 a promising Eocene marine fauna and a Miocene land fauna 
containing one species of proboscidean (found to the north in Moghara, after 1898) related to Triloplwdon 
angustidens. 



1. ANDREWS CONFIRMS THE AFRICAN ORIGIN OF THE PROBOSCIDEA, SIRENIA, 

AND HYRACOIDEA 

In April, 1901, Charles William Andrews of the British Museum had an opportunity of accompanying Hugh 
Beadnell, the geologist, into the Fayum district. Here a marine (Zeuglodont) fauna had previously been found, 
and on this occasion a considerable number of vertebrate remains, including portions of skeletons of the probo- 
scidean Moeritherium and of the sirenian Eosiren, were collected from beds of Upper Eocene age ; the discovery 
of fluvio-marine beds of Lower Oligocene age resulted in the finding by Andrews of the first traces of the animal 
named Palseomastodon. Moeritherium was the first Eocene proboscidean to be discovered in any part of the world, 
and the type of the second genus, named Palseomastodon beadnelli, proves to be a possible ancestor' of the classic 
Mastodon of Cuvier. The expeditions of Beadnell (1901-1904), of Andrews (1901-1903), and collections now 
preserved in the British Museum and the Geological Museum of Cairo, form the subject of a series of papers by 
Andrews, culminating in his great Memoir of 1906, entitled, "A Descriptive Catalogue of the Tertiary Vertebrata 
of the Fayum, Egypt. Based on the Collection of the Egyptian Government in the Geological Museum, Cairo, 
and on the Collection in the British Museum (Natural History), London." 

After discussing the principal characters of the vertebrate fauna, Andrews (1906) presents the following con- 
clusions regarding the homeland of the Proboscidea and the early history of these hitherto unknown mammals : 

[Andrews, 1906, p. xv] Although Arsinoitherium is certainly the most extraordinary of the Ungulates found in these beds, 
nevertheless the remains of the primitive members of the Proboscidea are perhaps of greater scientific interest, because they 
help to fill, at least to a large extent, one of the most obvious gaps in oar knowledge of the extinct Mammalia. Previous to their 
discovery the earliest Proboscideans known were from the Lower Miocene (BurdigaUen) of Europe and Northern Africa, and 
although many earlier deposits rich in mammalian remains were known in various parts of the world, in none of them was any 
trace of Proboscidea found, so that their appearance in Europe at the beginning of the Miocene period must be the result of 
their immigration from other regions. The probability that Africa would be found to be the original home of these animals 
was pointed out by several writers, notably by Osborn, Stehlin, and TuUberg. The first of these [Osborn, 1900.182, also 
1900.187] suggested that probably not only the Proboscidea but also the 'Hyracoidea, certain Edentates, the Antelopes, the 
Giraffes, the Hippopotami, the most specialized Ruminants, and among the Rodents the Anomalures, the Dormice, and Jerboas, 
among Monkeys the Baboons,' and, as his map [Fig. 10] suggests, the Sirenia also, originated in this region. Osborn also 
put forward the theory that a succession of migrations from Africa to Europe occurred, notably at the end of the Eocene, at the 
beginning of the Miocene, and again in the earliest Pliocene. It was in the early Miocene migration that the Proboscidea 
passed out of Africa for the first time so far as known. Stehlin [1900], who also emphasized the importance of Africa as a 
probable centre of mammalian evolution, expressed much the same views. TuUberg [1899] Ukewise regarded Africa as a centre 
of mammalian radiation, and pointed to Hystrix (or the whole of the Hystricognathi), the Simiae (Anthropoidea), and the 
Proboscidea as having probably migrated thence in the early Miocene. It is therefore very satisfactory that the earliest traces 
of land-mammals from the Eocene of Africa include remains of primitive Proboscidea, as well as early forms of Hyracoidea, 
Sirenia, and perhaps some of the other groups. 

The earliest-known Proboscidean is Moeritherium, which occurs first in the Qasr-el-Sagha beds (Middle Eocene) and 
persisted till the Upper Eocene, its remains having been found in the Fluvio-marine series. This animal was about the size 
of a Tapir, which it must have greatly resembled in general appearance. ... [p. xvii] The Umbs are unfortunately not well 
known. The hiunerus differs considerably from that of the later Proboscidea, but some of the smaller species of Palseo- 
mastodon (see text-fig. 56) from the Upper Eocene seem to supply intermediate forms: probably the difference arises from 
the fact that Mceritherium was a more or less amphibious type, while the later Elephants became fitted for progression on 
firm ground. The femur approximates very nearly to the form found in the later Proboscideans. As already mentioned, 
Moeritherium was probably an amphibious, shore, or swamp hving animal, and it was no doubt owing to the continuation of the 
conditions favourable to its mode of life that it persisted into the Upper Eocene period. . . . 

The largest species of Palseomastodon {P. beadnelli) must have been about the size of a half -grown Indian Elephant: 
in its general appearance it was Elephant-Uke, but differed in having a longer neck and the symphysial portion of the mandible 
prolonged beyond the skull (see text-fig. 48, p. 131) and covered only with the fleshy snout. Probably it could reach the 
ground with its lower incisors, and the end of the snout may already have been prehensile. ... [p. xviiij In general appear- 



'(Osborn, 1935) This proves to be an error on the part of both Matsumoto and Osborn; Palseomastodon is not an ancestor of Mastodon, Zygolophodon, or 
Turicius (cf. p. 43, Matsuraoto's phylogeny). See Pal.eomastodontin« subfam. nov., page 6.5 and Appendix. 



ORIGIN AND CLASSIFICATION 



37 



ancc Palseomastodon must have resembled a small rather long-necked Elephant, the most notable difference being that the 
trunk, instead of being freely flexible, was supported by, and formed the upper covering of, the elongated mandibular symphy- 
sis; its extremity, however, may have been free and to some degree prehensile. 

The further history of this group can only be briefly summarised here (see Phil. Trans. 196b, 1903, p. 99). The next form, 
Telrahelodon angustidens from the European Lower Miocene, has the symphysis still more elongated and the narial opening shifted 
further back. At the same time the molars, or at least the posterior ones, are greatly increased in size and possess more trans- 
verse ridges. ... [p. xix] This animal attained the size of a moderately large Indian Elephant, and except for the inflexibility 
of the mandiljle-supported trunk must have been very similar in appearance. In the later Miocene the mandibular symphy.sis 
shortened, leaving the trunk mobile and unsupported, as we now know it; at the same time traces of its original elongated 
condition are retained in the occurrence of deciduous lower incisors in some species of Mastodon and in the peculiar sharp process 
of the symphysis in the Elephants. . . . The above is, of course, only a general summary of the succession of forms which lead 
up to the modern Elephants, and no doul)t there have been many checks and side-branches leading only to extinction in the 
course of the vast period that has elapsed since Marithcrium existed. ... [p. xxi) The question of the origin of the Sirenia is 
of great interest, and there seems to be a considerable amount of evidence in favour of the view first put forward by de Blain- 
ville, that they are intimately related to the Proboscidea. In the first place, the occurrence of the most primitive Sirenians 
with which we are acquainted in the same region as the most generali.sed Proboscidean M(eritherium is in favour of such a 
view, and this is further supported by the similarity of the brain-structure and, to some extent, of the pelvis in the earliest- 
known members of the two groups (see pp. 202 dfe 214). . . . [p. xxvii] Another consideration which adds to the importance 
of Africa as a centre of mammalian evolution has been pointed out by Stromer [1903], namely, that part of it at least has 
probably never been submerged since the Paleeozoic period, and formed a portion of a vast Permo-Triassic land-area inhal)- 
ited by a great variety of mammal-like Theriodont reptiles from which the Mammalia may have actually arisen. This being 
the case, it is not only the Tertiary, but also the Secondary, deposits of this region that may be expected to yield most im- 
portant data for the history of the Mammalia. 



Map of the Region Explored by thk Ameiucan Mdseum Expedition of 1907 
Route of the Expedition indicated by dotted lines and arrows 



U^ff Bo*t iMTtt Ifntit Otm/n 

foMannity 4rnnin'tJtrrium, 




UArHrrh Ltrtl 




fiicfou Bone Urtf M0i>£jrvAitm£ Di^sits 
Cotnaimn^ Sireiiia'ia.ZetjyIeifeiits , 
Artrttrat Cle/ihttntal M^rilfimiinii r, i^ n 



^ Lamm Stȣ Urttt fUirnrr Oi^itrs 

J fatiattutf Smumnx *«.* Zruf/firrtrt" i 




Fie. 11 

Fir. ! 1. Section showing the lowering level.s of the anoient Lake Mceris, of the 
brackisli Birket-el-Qiirun, and of the sea, with the (1) lower marine Zeughdon beds 
(Middle Eocene), (2) middle Qasr-el-Sagha Mrerilherium beds (Upper Eocene), 
(3) upper Fluvio-raarinc Palseomastodon beds (Lower Oligocene). .\fter Osltorn, 
HH)7.302, Century Magazine, p. 827. 

Fig. 12. The horizontal line.s indicate the fertile lands of the Nile and of the 
FayAm. The F"ay(5m area formerly covered by Lake Mocris is now reduced to the 
brackish Birket-el-Qurun. The oblique dots north of Hirket-el-Qurun indicate the 
area of the marine Qasr-el-Sagha lieds. On the bluffs to the northwest are the 
Fluvio-marine beds, on which was placed the American Mu.seum camp. .\fter 
Osborn, 1907.302, Century Magazine, p. 826. 




Fig. 12 



Andrews thus adopted the Osborn-Stehlin-Tullberg theory that Africa was an important center of ada]:)tive 
radiation of the MammaUa and called attention (p. xxv) to the fact that "the species known must be a mere frac- 
tion of the faunas inhabiting the Etliiopian region during the Middle and Upper Eocene periods"; he maintained 
that Osborn, Stehlin, and Tullberg were fully justified "in their assumption that the Ethiopian continent in early 
Tertiary (and perhaps pre-Tertiary) times was a very important centre of mammalian evolution." 



38 



OSBORN: THE PROBOSCIDEA 



Finally, the Deinotheres, although first discovered in Europe and Asia, have been found in recent years in 
deposits of Lower Miocene age in central {Demotheriimi hobleyi) and northern (D. cuvieri) Africa, as fully described 
in Chapter IV. Consequently early members of the Moeritheres, of the Deinotheres, of the true mastodonts, 
and of the bunomastodonts have been discovered in deposits of Upper Eocene to Lower Miocene age in the 
comparatively unexplored continent of Africa. In the Pleistocene of Africa occurs Deinotherium hopwoodi sp.nov. 

2. AMERICAN MUSEUM FA YUM EXPEDITION OF 1907. INTERPRETATION AND DISCOVERY 
BEARING ON THE PHYLOGENY OF THE FAYUM PROBOSCIDEA 

The general inference drawn from Andrews' memoir and popular articles and restorations was that the 
genera Mf^ritherium and Palseomastodon constituted the first and second links in a chain of descent which led 
up to the classic Mastodon angustidens of Europe, thus: 

Mastodon {Tetrabelodon [= Trilophodon]) angustidens of the Miocene of France. 
Palseomastodon, including a variety of species. Upper Eocene to Oligocene of Egypt. 
Moeritherium, including at least two species, Middle and Upper Eocene of Egypt. 



Second Restoration of Mcerithbrium and Phiomia by Osborn and Christman, 1908 
Compare front views (Figs. 19 and 20) 




Fig. 13. Side view of Uhe head of Mmrilherium with the eye 
and ear in position, the form and position of the nostrils being 
somewhat conjectural. Observe the elongated cranial region and 
abbreviated facial region; also the direct opposition of the upper 
and lower incisor teeth. Modeled by, Erwin Christman, under the 
direction of Osborn in 1908. After Osborn, 1909.332, p. 140, 
fig. 2. 




Fig. 14. Side view of the head of Phiomia ivintoni, as modeled in 1908 by 
Erwin Christman, imder the direction of Osborn. Observe the abbreviated cranial 
region, the elongated facial region, the upper tusks passing on the outer side of 
the lower tusks. Originally described as Poteomostodore. After Osborn, 1909..332, 
p. 139, fig. 1. 



This monophyletic interpretation of Andrews' conclusions was due rather to his broad inferences than to an 
examination of his actual descriptions in which the divergent adaptations to habit and to habitat, which would 
militate against any monophyletic theory, were clearly pointed out. But despite Andrews' admirable work, the 
opinion widely prevailed up to the year 1909 that Mceritherium gave rise to Palseomastodon and that Palseomastodon 
gave rise to Mastodon angustidens. 



ORIGIN AND CLASSIFICATION 



39 



The first to combat this monophyletic interpretation was Osborn who led an expedition into the Fayum 
region early in the year 1907, immediately after the publication of Andrews' great memoir, assisted by Walter 
Granger and George Olsen. A preliminary study of the extensive collections of Moeritherium, Palseomastodon, and 
Pfdomia, followed by com|)arison and restoration for purposes of exliibition in the American Museum, was entirely 
against the theory that Moeritherium gave rise to Palseomastodon; in fact, it was in this process of restoration, as 
shown in the accompanying figures (Figs. 13, 14), that profound divergences between Palseomastodon { = Phio- 
miu) and Moeritherium were brought out. This led to the publication in "Nature," July 29, 1909, of Osborn's 
paper, entitled, "The Feeding Habits of Moeritherium and Palseomastodon." As will be fully set forth in Chapter 
XXI on "Affinities, Migrations, and Phylogeny of the Proboscidea," differences in feeding habits give us 
the key to divergence, divergence gives us the key to phylogenetic separation and succession, and phylogenetic 
separation gives us the key to classification. 

Recent studies by Gregory (1910-1920), Matsumoto (1923), and Kkstokation op the Eocene and Recent sirenians 

,,„„^> . 11- • TT yi 1 (EoslRKN AND ThICHECHUs) 

bimpson (1932) pomt towards the existence in Upper Cretaceous and 
Lower Eocene time in Africa of a common ancestral form of mammal 
which by adaptive radiation through ground-loving, shore-loving, and 
water-loving habits, may have given rise to the Sirenians (Fig. 15), 
the McERiTHERES, and the Proboscideans. Widely 
different and profoundly divergent as the two great 
orders of Sirenians and Proboscideans are today, they 
still exhibit certain common characters in their inter- 
nal anatomy, certain common characters in their 
cranial and labial structure, as well as one unique char- 
acter in their grinding teeth, namely, trilophodonty or 
the evolution of three crests on the superior and 
inferior molars. Recent studies by Simpson (1932) 
show that the Pre-Pleistocene forms of America are 
Dugongs rather than Manatees. 




Fig. 15 (L<"ft). The Eocene sea 
cow, Eosiren libyca, a contemporary 
of Mirrilhrrium lyonsi and M. gracile 
in the Qasr-el-Sagha l)e<ls, retaining 
ihc small hind limhs. Restoration 
and drawing liy Charles R. Knight, 
under the direction of the author. 
After Oslx)rn, 1907.301.', Century 
Magazine, p. S3I. 

(Right). The existing Manatee, 
Tricherhiia anirricntiuf, of the rivers 
of Florida, in which Ihc hind liml>.s 

have entirely disappeared. Restoration and drawing liy Charles R. Knight. 

under the direction of the author, .\fter Osborn, 1907.302, Century Magazine, 

p. 831. 



3. RELATIONS OF MCERITHERIUM, PALSEOMASTODON, AND PHIOMIA TO EACH OTHER AND 

TO OTHER MAMMALS 

It appears that Osborn in 1909 went too far in separating Moeritherium from direct relationship to the Probo- 
scidea; he erred in the opposite direction, as shown in the following quotation (Osborn, 1909.332, p. 139): 
"The conclusion is that Moeritherium was a confirmed and continual river-living animal, feeding mainly under 
water and on the banks, more specialised for a(}uatic life than the hippojiotami, as indicated by its feeble pelvis, 
but less specialised than the Sirenians. It would not be far from the truth to say, from our present knowledge 
of the animal, that Moeritherium is an offshoot of the Proboscideo-Sirenian stock, with slightly nearer kinship to 
the elephants than to the Sirenians." This conclusion should now certainly be greatly modified, because Moeri- 
therium proves to be of much nearer kinship to the Proboscidea than to the Sirenia; it is in fact one of the primary 
stocks of the Proboscidea. 



40 



OSBORN: THE PROBOSCIDEA 



To Osborn's extreme view Andrews replied in 1909, p. 305 ("The Systematic Position of Mceritherium," 
Nature, LXXXI, p. 305): "On tlie whole, it seems that the weight of evidence is in favour of regarding Mceri- 
therium as a proboscidean, though perhaps not on the direct line of ancestry of Palseomastodon, and retaining 
some characters of the original Proboscideo-Sirenian stock." Schlosser (1911) considered it doubtful whether 
Moeritherium stands in the direct Hne of ancestry of Palseomastodon. Gregory ("The Orders of Mammals," 
1910, p. 368) concluded: "The genus [Moeritherium] represents a very primitive offshoot from the Proboscideo- 



.jl^ijmj^^.iia^ 



The Osborn-Knight early Restorations, in the year 1907, of 
MtERiTHERiUM (Upper Eocene), Phiomia (Lower Oligocene), and 
Parei.ephas (Upper Pleistocene). Compare more recent (1932, 1933) 
Restorations in this Memoir. 




M^^ a 





Fig. 16. 



Fig. 16. (Left) Evolution of the head, proboscis, nostrils, and tusks in: 
(1) Mwrithcrium, (2) Phiomia u'intoni, (3) Parelephas jeffersonii, as restored 
by Charles R. Knight in 1907, under the direction of the author. After 
Osborn, 1907.302, Century Magazine, p. 833. 



(Right) Comparative restorations to scale of: (1) Mmrilherium andrewsi, 
(2) Phiomia osborni, (3) Parelephas jeffersonii. Drawn by Charles R. Kniglit 
in 1907, under the direction of the author. After Osborn, 1907.302, Century 
MagazinCj p. 834. Scale approximately one-fiftieth natural size. 




Fig. 16 



Sirenian stock. Its dentition and certain other characters indicate a nearer alliance with the Proboscidea than 
with the Sirenia, but it is far more primitive than any other known representative of either order." More recently, 
Gregory (1920, p. 180) in his exhaustive study on the lacrymal bone of the vertebrates declares: "The orbital 
region '[of Moeritherium]- is much more primitive than that of other Proboscidea, and suggests the sirenian type. 
. . . [p.' 245] The Sirenia, although very highly speciaUzed for aquatic hfe, show special resemblances with 
Moeritherium in the skull (including the orbital region) and dentition, and are generally regarded as a derivative 
of the proboscidean stem." 



4. MATSUMOTO'S COMPLETE SEPARATION OF xMCERITHERIUM, PAL.EOMASTODON, 

AND PHIOMIA 

Matsumoto (1922, 1923, 1924) was the first to make a profound examination of the rich collections of Maeri- 
therium, Palxomastodon, and Pkiomia in the American Museum, in comparison with the type and referred speci- 
mens in the British and Cairo Museums. Aided by the new evidence presented in this rich American Museum 
material, he deserves the chief credit: (1) Of clearly distinguishing the genus Palxomastodon from the genus 
Phiomia, (2) of clearly defining and distinguishing three species of Phiomia and three species of Palseomastodon , 
(3) of distinguishing species of Mwrithcriinn, (4) of closely comijaring these three genera with each other and 
with representatives of the great orders Sirenia and Hyracoidea which occur in the same deposits. 

In the present Memoir Osborn entirely confirms Matsumoto's principal systematic conclusions. Matsumoto's 
separation of the genera Palseomnstodon and Phiomia is a confirmation of one of Andrews ' observations, namely (An- 
drews, 1905, p. 562) : "The species of Palaeomnsiodon fall into two sections, in one of which the posterior end of the 
symphysis of the mandilile is situated considerably in front of the level of the anterior premolar, while in the other 
it is only very little in front of that point. The first group, moreover, is distinguished by the comparative simphcity 
of the molars, in which the accessory cusps are scarcely at all developed, and by the small size of the talon of the 
last lower molar; into this subdivision the original species, P. beadnelli, falls, together with a much smaller form 
for which the name P. parvus is now proposed. The type-specimen of this new species is the right ramus of the 
mandible, with the premolars and molars in situ, though somewhat crushed." 

The profound structural and adaptive differences between these three genera of proboscideans, Moeritherium, 
Palseomastodon (sensu strictu), and Phiomia, will be fully set forth in the systematic section below. A significant 
feature of this separation is that it tends to greatly strengthen the theory that Africa iras the original homeland 
of the order Probosciden in which they widely diverged from each other during Lower and Middle Eocene times, so 
that as we find them in Upper Eocene time they are distinctly polyphyletic. This gives a death blow to all the 
monophyletic inferences based on Andrews ' earlier writings. 

The relations of these three great phyla to each other are set forth in the section on phylogeny and classifica- 
tion (Vol. II) and in the present chapter. The relations of these three phyla of Proboscidea to the Sirenia 
and the Hyracoidea respectively are discussed in detail in Matsumoto's paper of 1923, entitled, "A Contribution 
to the Knowledge of Mwritherium," in which he takes up with great fullness the following topics: 

A. — Characters Possibly Indicating Aquatic Adaptation (p. 102). B. — Characters Indicating Terrestrial and 
Non-aquatic Adaptation (p. 104). C. — Characters in Common with the Hyracoids (p. 105). D. — Characters 
Distinctive from the Hyracoids (p. 109). E. — Characters in Common with the Sirenians (p. 111). F. — Char- 
acters Distinctive from the Sirenians (p. 112). G. — Proboscidean Characters (p. 114). H. — Peculiar and 
Pre-Palaeomastodont Characters (p. 117). I.— Natural Position of Moeritherium (p. 121). J.— Evolutionary 
Tendencies in the Earlier Proboscideans (p. 121). K. — Phylogenetic Relations Among the Hyracoids, Earlier 
Proboscideans, and Sirenians (p. 123). 

Phylogenetic Relations of Proboscideans to Sirenians and Hyracoids. — Matsumoto's conclusions may 
be paraphrased as follows: (1) That the proboscideans and sirenians were very closely related to each other in their 
earlier stages of evolution has been maintained by many eminent authors, such as De Blainville, Andrews, Osborn, 
Gregory, etc. ; (2) it is very obvious that in their earlier stages they resembled each other in certain respects, but 

41 



42 OSBORN: THE PROBOSCIDEA 

the phylogenetic relationship between the Proboscidea and Sirenia is somewhat less close than is considered by 
those authors (e. g., Osborn) who emphasize chiefly the resemblances between Maeritherium and the earher 
sirenians; (3) certain similarities between Mceritherium and the earlier sirenians are probably due to their 
original phylogenetic relationship, other resemblances are attributable to convergence, i. e., similarity of habits, 
still other resemblances are primitive characters common to all three groups of early sirenians, early proboscideans, 
and early hyracoids; (4) as to progressive development, it is obvious that the differences observed between the 
sirenian series of Prorastomus, Eotherium, Prosiren, Eosiren, Miosiren, Halitherium, Halicore, etc. and the progres- 
sive development of the proboscideans Maeritherium, Palseomastodon, Trilophodon, and Megahelodon are very 
great; (5) doubtless no known sirenian can be looked upon as an ancestral type of Mceritherium, nor can Maeri- 
therium be looked upon as an ancestral type of any known sirenian ; (6) consequently both the sirenians and the 
proboscideans may have descended from unknown ancestors that stand closer to each other than they do to the 
Hyracoidea stock. Detailed examination and comparison of thirty-four characters in sirenians, hyracoids, and 
early proboscideans led Matsumoto (1923, p. 114) to the following somewhat conflicting conclusion: "Among 
these thirty-four characters examined, thirteen, viz. (2), (9), (10), (12), (13), (15), (16), (17), (18), (19), (21), (25), and 
(26), are common to both the hyracoids and proboscideans in contrast to the [earlier or later] sirenians; ten, viz. 
(4), (8), (11), (14), (20), (22), (27), (28), (29), and (30), are characteristic of the proboscideans in contrast to both 
the hyracoids and earlier or later sirenians; three, viz. (5), (7) and (23), are characteristic of the proboscideans in 
contrast, at least, to the earlier or later sirenians; three, viz. (32), (33), and (34), are characters vaguely distinctive 
of this genus from the earlier or later sirenians; three, viz. (9), (13), and (23) [these characters are referred to in 
duplicate], are characters of some of the terrestrial mammals in contrast to the aquatic ungulates; and one, viz. 
(31), is characteristic of this genus in contrast to all the hyracoids, sirenians, and other proboscideans or earlier 
forms of these groups. It may easily be recognized that Maeritherium has many characters common to both the 
hyracoids and proboscideans in contrast to the earlier or later sirenians, and many characteristics of the probo- 
scideans in contrast to both the hyracoids and later sirenians or earUer sirenians." 

Osborn, 1924: In our opinion Mceritherium, a true proboscidean of primitive type, stands much closer to the 
stem forms of the Sirenia than to the stem forms of the Hyracoidea ; for the degree of this closeness we must wait 
for fuller evidence from Middle and Lower Eocene deposits in Africa. 

Proboscidean Characters of Mceritherium (Matsumoto, 1923). — Having dismissed the resemblances 
to the Sirenia and Hyracoidea as principally due to inheritance from primitive mammalian characters in general 
and also to convergence, Matsumoto applies a thirty-five character test and sets forth a very strong argument 
for the dominant relationship of Mceritherium to the Proboscidea and especially for its peculiar and pre-palaeo- 
mastodont characters. After comparing and contrasting Mceritherium with Palxomastodon, we may paraphrase his 
conclusion as follows (p. 120) : (1) In all the twenty-three characters examined, Mceritherium is structurally a pre- 
palseomastodont type so far as we admit the conception that Palseomastodon, Trilophodon, and Megahelodon form 
together a fair series of evolutionary stages; (2) it is, of course, beyond doubt that the structural gap between 
Mceritherium and Palseomastodon is fairly great, yet the fact should not be neglected that Mceritherium stands 
structurally before Palseomastodon in the majority of its characters, that is, as a more primitive form; (3) con- 
sequently a presumed ancestral type of Palseomastodon should resemble Mceritherium in many characters. 

Osborn, 1924: In our opinion Andrews and Matsumoto have positively estabhshed the proboscidean relation- 
ship of Mceritherium. Osborn differs from Andrews and Matsumoto in placing Mceritherium in an entirely separate 
line of descent, of superfamily rank known as Moeritherioidea, distinguished by direct opposition of the upper 
and lower incisor teeth, as in rodents, by many aquatic adaptations in the skull, by profound differences in skull 



ORIGIN AND CLASSIFICATION 



43 



proportion, and probalily by an amphibious mode of life approaching that of the hippopotami; in contrast to the 
purely terrestrial life of the Palxomastodon and Phiomia phyla. As to the amphibious habits of Moeritlwrium, 
Osborn and Matsumoto substantially agree. 

Palxomastodon Ancestral to the True Mastodon (Matsumoto, 1924) .^In Matsumoto's succeeding 
paper (1924.1), "A Revision of Pala;omastodon," etc., he concludes by completely .separating Mcerilherium and 
placing it in an independent line, as shown in the accompanying diagram {op. cit., p. 58) : 

In my opinion, tho pliylogenetic relation.ship of the KWiTa just referred to, can be diaKrammatically shown a-s follows: 



Mwrithtrium 



Mastodon 



Telrahphodon 
(longirostral) 



Megahelodon 



Zygolophodon 



Trilophodon 



Chverolophodon 
I [pentelicus] 



Palxomastodon 



I ,-- 



Phiomia 



In discussing the phylogeny of the earUer Proboscidea, Matsumoto (1924.1, pp. 55-57) makes a final comparison 
between the skull and dentition of Mcmiherium, Palxomastodon, and Phiomia; he riglitly places Maeritherium 
in aside-line of its own, but wrongly determines that the true Palxomastodon points towards Zijgolophodon and the 
true Mastodon of the Mastodon americanus type for the following reasons. (1) The skull of Palxomastodon is not 
clearly known ; judging from the shape of the palate it is probably shorter skulled than Phiomia which is distinctly 
long skulled; the skull of Zygolophodon of the European Miocene is not yet clearly known; the true Mastodon 
is distinctly short skulled. (2) The jxilate of the true Palxumustodon is very wide in proportion to the length of the 
cheek teeth, while that of Phiomia is rather narrow in the same proportion; the palate of the true Mastodon is 
very wide. (3) The mandibular symphysis of Palxomastodon appears to be rather short, while that of Phiomia is 
very long; the symphysis of the true Mastodon is short. (4) The largest and most conspicuous of the anterior mental 
foramina lies just below the first cheek tooth, Pg; in the true Mastodon also the largest and most conspicuous of 
these foramina lies just below the anterior cheek tooth ; in many short-jawed mastodonts and elephants the largest 
and most conspicuous of the anterior mental foramina lies just below the anterior cheek tooth also; the position 
of this foramen may be correlated with the ilevelojiment of the relatively short symi)hysoal region and the rela- 
tively small lower tusks, in contrast with the long synipiiysoal region and large lower tusks in the Phiomia-Trilopho- 
don series. (5) The cheek teeth of Palxomastodon are proportionately shorter and wider than those of Phiomia; 
the cheek teeth of Zygolophodon and Mastodon are also proportionately shorter and wider than those of Trilophodon. 



44 OSBORN: THE PROBOSCIDEA 

(6) The cheek teeth of Palxomastodon show a lower ridge formula than those of Phiomia; the potentiality of getting 
a higher ridge formula was lower in the Zygolophodon-Mastodon phylum than in the Phiomia-Trilophodon phylum. 

(7) The cheek teeth of Palxomastodon are bunolophodont attaining a typically lophodont character when mode- 
rately worn, while those of Phiomia are typically bunodont; the cheek teeth of Zygolophodon and Mastodon are 
lophodont, while those of Phiomia and Trilophodon are bunodont. (8) In the cheek teeth of Moeritherium, Palxo- 
mastodon, Zygolophodon, and Mastodon no trefoil pattern of cusps is developed, while in the Phiomia-Trilophodon- 
Tetralophodon phylum the trefoil pattern of cusps is well developed. (9) In the cheek teeth of Moeritherium, 
Palseomastodon, Zygolophodon, and Mastodon the crests are not very thick anteroposteriorly, the valleys are widely 
open and even the walls and bottoms of the valleys are worn, while in those of Phiomia, Trilophodon, and Tetra- 
lophodon the crests are very thick anteroposteriorly, the valleys are not so widely open, the worn surfaces are 
almost even. In the cheek teeth of Moeritherium, Palxomastodon, Zygolophodon, and Mastodon the surface of the 
enamel is rather smooth, while in those of Phiomia, Trilophodon, and Tetralophodon the same is very rough. 

For all these reasons, Matsumoto continues, Palxomastodon appears nearly to correspond to a theoretical 
ancestral type of the Zygolophodon-Mastodon phylum ; Palxomastodon appears to correspond to the beginning of a 
very great phylum, namely, the Zygolophodon-Mastodon phylum. 

Matsumoto thus announces a very important phylogenetic discovery, i. e., that Palxomastodon stands 
near the ancestry of the true Mastodontidae. This discovery was immediately approved by Osborn, before the 
publication by Matsumoto of his complete results, with the following comments (see Matsumoto, 1922, p. 6, 
"Note by Henry Fairfield Osborn, August, 1922"): "The significance of the above revision [Matsumoto's] is 
that the true Palxomastodon beadnelli has bilophodont intermediate molars and a relatively broad skull ; it is a 
rare animal both in the British Museum and American Museum collections; according to Andrews ([letter], 1922) 
the genotype {Palxomastodon beadnelli) was found at the very base of the Fluvio-marine Beds, Lower Oligocene, 
50 or 100 feet below the Phiomia level. . . . Associated with the type is a very large femur and humerus." Palxo- 
mastodon is a relatively short-jawed animal, with bilophodont intermediate molars, whereas Phiomia is a very long- 
jawed animal with trilophodont intermediate molars. 

Matsumoto's theory is now (1935) rendered improbable by Osborn's closer studies in the present Memoir 
of the limited material of the true Palxomastodon from the Fayiim (see footnote on p. 36 above). 

Phiomia Ancestral to the True Trilophodon.'— From the first Andrews recognized the kinship of Phiomia 
to Trilophodon angustidens; Matsumoto goes further and proves that Phiomia is directly ancestral to what he 
terms the Trilophodon-Tetralophodon-Megabelodon phylum, whereas it has nothing to do with the ancestry of the 
Zygolophodon-Mastodon phylum. The kinship of Phiomia to Trilophodon is clearly indicated in the following 
characters: (1) Phiomia, Trilophodon, and Megabelodon are distinctly long skulled; (2) the palate of Phiomia is 
rather narrow in proportion to the length of the cheek teeth, the palates of Trilophodon and Megabelodon are also 
distinctly narrow; (3) the mandibular symphysis of Phiomia is very long, while the mandibular symphysis of 
Trilophodon and of Megabelodon is extremely long; (4) the anterior mental foramen in Phiomia lies on either side 
of the symphyseal region far anterior to the front cheek tooth, P3, and to the posterior end of the symphysis, in 
both Trilophodon and Megabelodon the position of the anterior mental foramen being quite similar to that observed 
in Phiomia; (5) the cheek teeth of Phiomia show a trilophodont ridge formula almost similar to that of Trilophodon, 
the potentiality of a multiple ridge formula is thus high in the Phiomia-Trilophodon-Megabelodon phylum, whereas 
it is low in the Palxomastodon phylum and still lower in the Moeritherium phylum ; (6) the cheek teeth of Phiomia 
are typically bunodont, those of Trilophodon, Tetralophodon, and Megabelodon are also bunodont; (7) in Phiomia, 

'(Osborn, 1935) The three or four known species of Phiomia are now recognized as directly ancestral only to the shovel-tusker Amehelodon of Nebraska. 
See Plate v between pages 235 and 236. 



ORIGIN AND CLASSIFICATION 



45 



Trilophodon, Telralophodon, and Megabelodon the trefoil pattern of cusps is well developed ; (8) the transverse 
crests in Phiomia, Trilophodon, Telralophodon, and Megabelodon are very thick anteroposteriorly, the valleys are 
not so widely open, and the worn surfaces are almost even; (9) the surface of the enamel is relatively rough in 
Phiomia, Trilophodon, Telralophodon, and Megabelodon; (10) the basal cingula are very rough and strong in 
Phiomia, Trilophodon, Telralophodon, and Megabelodon, whereas they are rather feeble in Pal^omastodon and 
Maslodon and more or less strong in Moerilherium; (11) thus Phiomia appears to correspond to the beginning of a 
very great phylum, namely, the Trilophodon-Tetralophodon phylum of the Old World. 

Osborn, 1924: Quite independently of Matsumoto, Osborn had confirmed and extended Andrews' opinion 
as to the Phiomia-Trilophodon phylum, and as early as 1917 (Osborn, 1918.468) placed Phiomia { = Palxomastodon) 
in the true longirostral line which he termed the Longirostrinse; but Matsumoto deserves the priority not only for 
clearly distinguishing Palaeomuslodon from Phiomia, but for suggesting that Palaeomastodon was related to the 
subfamily which Osborn designates as Mastodontinae. 



5. AFRICA DEMONSTRATED AS THE HOMELAND OF THE THREE PRIMITIVE FAMILIES 
MCERITHERIID.E, MASTODONTID^E, AND BUNOMASTODONTID^ 

Thus Andrews, Osborn, and Matsumoto have successively prepared the way for a true understanding of the 
Ihree greal and widely distinct phyla of Proboscidea in northern Africa in Upper Eocene and Lower OUgocene times. 
As more fully pointed out in the phylogenetic classification of the forthcoming Volume II, the wide divergence 
of these phyla tends to confirm the theory that Africa was the homeland of the stem forms of all the Proboscidea, 
although it does not as yet finally demonstrate it. 

Osborn's present classification of the thirteen species of the primitive North African Proboscidea is shown on 
page 65. 




Fit;. 17- Phiomia osborni, Primitive Lower Oliqocene Shovel-Tusker of the FayCm, Egypt 
Restored (1932) by Margret Funsch, under the Direction of Henrv Fairfield Osborn 

All figures one thirty-sixth natural size 
Phiomia osborni frequented the flood-plain region of the Nile and the present restoration shows a group of these animals in their 
natural lialiitat in Lower Oligocene time. The three specimens are based on a single lower jaw, whieh Matsumoto in 1922 dedicated to the 
present author; their estimated shoulder height is 1345 mm. or 4 ft. 5 in. 



III. PHYLOGENETIC RELATIONS AND CLASSIFICATION OF MCERITHERIUM, 

PAL.€OMASTODON, AND PHIOMIA 

The systematic descriptions and the nomenclature of the Fayum Proboscidea will be clearly reviewed below. 
Let us first examine their structure more closely with regard to adaptations to certain habits and modes of life 
and to progressive tendencies towards increasing perfection and specialization in adaptation. The Proboscidea 
follow the principle which we have clearly observed in many other ungulates, especially the families Brontotherii- 
dae, Rhinocerotidse, and Equidae, which the author has been studying and comparing for the past thirty-four 
years, namely, a phylum having once started in a certain adaptive direction, if unchecked, will proceed to the extreme 
of adaptation. 

For example, elongation of the jaws, of the skull, and of the teeth begins in Phiomia and reaches an extreme 
in certain species of trilophodonts in which the progressive development is known as longirostral. 



Text-fig. 40. 





e-i-.oc. 



Fig. 18. First reconstruction by Charles W. Andrews of skull and mandible of M rerilherium lyomi. A, from above; B, from left side. 
After Andrews, 1906, text figure 40. About one-fourth natural size. 

The Osborn-Matsuraoto reconstruction (Fig. 42) of the skull of M. andrewsi-trigodon differs widely in the relations of the upper and lower incisor 
teeth, P-Io. These teeth are actually less tusklike and more gliriform than as restored by Andrews (Fig. 18). 



1. PRINCIPLE OF ADAPTIVE RADIATION EXPRESSED IN THE PHYLOGENETIC 
CLASSIFICATION OF THE FAYUM PROBOSCIDEA 
Classification should be expressive of phylogeny as brought about through adaptive radiation into diverse 
habits and habitats which influence, first, the entire dental and cranial structure, second, the entire foot, limb, and 
body structure, third, the adaptation of dental, cranial, limb, and body structure to defense and offense, and, fourth, 
coadaptation and compensation through gain and loss of other organs. 

The analysis of the feeding and locomotor habits is the basis of the phylogeny and therefore of the classifica- 
tion of the Proboscidea, from the first appearance of these animals in the Upper Eocene up to the surviving recent 
forms in Asia and Africa. 

46 



ORIGIN AND CLASSIFICATION 47 

Resume of Preceding Section.— Through the discoveries of Andrews (1901-1906) in the Eocene-Oligo- 
cene of the Fayum of Egypt of three very distinct types, namely, of Palaeomastodon (very rare), of Maritherium 
(numerous), of Phiomia (very numerous), these three animals have become famous as the earliest known stages in 
the ancestry of the elephants. Since the wish is always father to the thought, and nothing was more eagerly 
sought for prior to 1901 than the primitive progenitors of the Proboscidea, it was altogether natural to place 
Mceritherium in or near the direct line of ancestry of Elephas, and consequently, with this idea, to provide Moeri- 
iherium, Palaeomastodon, and Phiomia with a proboscis of less or greater length such as would befit more or less 
remote direct ancestors of the elephant. 

As such direct ancestors these three animals have gone forth into the general literature, and have therefore 
taken an entirely false and unnatural position in popular treatises on phylogeny. Andrews tried to correct this 
error in his great Memoir of 1906, and Osborn (1909.332) in his article "The Feeding Habits of McEritherium and 
Palaeomastodon," followed by Matsumoto (1923), pointed out that since the dental and cranial structure of 
these three animals was profoundly different, it followed that the feeding habits were profoundly different and 
that they fed in three different habitats, consequently that their limb and body structure must have been different, 
finally, that none was ancestral to the other or to the genus Elephas. In brief, we were all so eager to welcome 
these animals into the order Proboscidea, that we too hastily assumed their ancestral position. 

We shall see that Moeritherium, Phiomia, and Paheomastodon are widely separated botli hy habits and by 
structure from each other, as well as from the still to be discovered direct ancestors of the true elephants. 



2. AMPHIBIOUS HABITS OF MOERITHERIUM 

The first more cautious note against the true-Elephas ancestry theory was sounded by Andrews in his Memoir 
of 1906 (1906, p. xvii) from which Osborn (1909.332, p. 139) made the following citation and comments: 

'As already mentioned, Moeritherium was probably an amphibious, shore, or swamp hving animal, and it was no 
doubt owing to the continuation of the conditions favourable to it.s mode of life that it persisted into the Upper Eocene 
period. In the meantime, however, either from this or some clo.sely allied type, there had arisen another animal more 
adapted to terrestrial life and showing a great advance in the direction of the typical Proboscidea: to this creature the 
name Palteomastodon has been given.' Elsewhere (p. xxi) Dr. Andrews notes that Mceritherium favours the view, first put 
forward by de Blainville, of an original relationship between the Proboscidea and Sirenia. Later on in the same work (p. 119) 
the same author, in commenting on the similarity between the pelvis of Mceritherium and that of the Eocene sirenian 
Eotherium, observes: — 'Then it may fairly be suggested that Mceritherium and Eotherium, both occurring in the same region 
(one the most primitive Proboscidean, the other occu])ying the same position with regard to the Sirenia), are, in fact, closelj' 
related, and had a common ancestor in early Tertiary times, prol)ably in the Lower Eocene.' On page 105 we find a comment 
on the remarkable likeness between the brains of Moeritherium and the Sirenia. 

Cranial and Dental Adaptations. — After these suggestive comments by Andrews were pubUshed, other 
specimens were secured by the American Museum Expedition of 1907 to the Faytim, including two fairly well pre- 
served skulls of Ma^ritherium (Fig. 42), which gave witness afresh to the general analogy of the skull of Moeritherium 
to that of the Sirenia. This material proves that Mceritherium not only had no proboscis, but that it was a totally 
different animal from Phiomia both in its appearance and habits and only very remotely related to it; this material 
proves further that in caenotelic adaptation the Moeritherium skull is closer to that of the sirenians and of the 
hippopotami than to that of any of the proboscideans. The cutting teeth, lips, and mouth parts of Mceritherium 
oppose each other in an entirely different manner than do those of Phiomia, so that it may be truly said that there 



48 



OSBORN: THE PROBOSCIDEA 



was not the least resemblance between either the mouth parts or the feeding habits of these two animals. In 
Moeritherium the nasal bones do not greatly recede, there was consequently httle or no free retractile power of the 
upper hp, which is always the first step in the evolution of the proboscis as witnessed in living species of Tapirus. 
Comparison of the muzzle of Mceritherium with that of Hyrax, of Castor, of Hippopotamus, and of other mammals 
with an enlarged pair of front teeth, tends to show that the upper and lower Ups were heavy and fleshy, and some- 
what similar in form, in function, and in prehensile power; that the blunt tusks wore directly against each 
other and were entirely sheathed in enamel ; that the lips were capable of closing over the tusks when the mouth 
was shut, somewhat as in the hippopotamus. The tusks were adapted as feeding organs rather than as fighting 
weapons, probably because Mceritherium was protected from attack by its partly aquatic habitat. 

Second Restoration of Mceritherium and Phiomia bt Osborn and Christman, 1908 
Compare side views (Figs. 13 and 14) 




Fig. 19. Front view of the head of Maerilherium with the eye 
and ear in position; the form and position of the nostrils somewhat 
conjectural. The eyes are seen to be very far forward, well raised 
toward the top of the face; the ears al.so are raised high on the side 
of the head; both these peculiarities are adiiptations to aquatic 
life to bring these sense organs near to the surface of the water 
in swimming, so that they will emerge first and disappear last. As 
modeled liy Erwin Christman, under the direction of Osborn, 1908. 




Fig. 20. Front view of the head of Phiomia wintoni, as 
modeled by Erwin Christman , under the direction of Osborn, 1908. 



The above method of comparison of Phiomia and M a^ritherium is that of making life-size models of the skulls in which the sense organs, 
such as the eye and the ear, are placed exactly in position, while the contours of the lips, nostrils, mouth parts, and external ears are largely 
conjectural. In Osborn's opinion (1926) the ear of Phiomia is erroneously restored, for it is much too elephantine. After Osborn, 
1909.332, pp. 139, 140. 

The conclusions drawn from the front teeth, I-'-I,, from the simple bunodont grinding teeth, from the very 
short face (brachyopy), from the long cranium (doUchocrany) , and from the extremely small bony eye sockets, 
are that Mceritherium was a confirmed and continual river-living animal, feeding mainly under water and along 
the banks of rivers, more specialized for aquatic hfe than the hippopotamus, as proven by its feeble pelvic bones, 
but far less specialized for aquatic hfe in its limb structure than the Sirenia. 

This does not prove that Mceritherium is of the order Sirenia, as Osborn suggested in 1909; it is certainly an 
independent member of the Proboscidea, as Andrews originally maintained and as Matsumoto has stoutly 
contended. Its cranial analogies are with the sirenian skull; its cranial and dental homologies are with the Probo- 
scidea. Its limbs and skeleton relate it to the primitive Proboscidea. 



3. LOWLAND HABITS OF PHIOMIA, ANCESTOR OF THE AMEBELODOXTIN^ 

Phiomia without a Proboscis. — The restoration-model of the head of Phiomia ( = the Pabeomastodon of 
the previous literature) also is without a proboscis but is provided with a stout upper lip which was capable of 
pressing closely against the procumbent paired lower incisor teeth. As Osborn observed (1909.332), Phiomia in all 
probability had not developed a proboscis. A profound difference between Phiomia and Mfjeritherium is brought 
out in comparing the front (Figs. 19 and 20) and side (Figs. 13 and 1-i) views of the head, in which it is seen that 
whereas the eyes of Phiomia are in tlie typical mannnalian position above the first permanent grinder, those of 
Maeritherium are very far forward, well raised in the front part of the head, and of very diminutive size, as is 
shown by the shallowness of the eye sockets. 

The distinctive pecuharity of Phiomia is that its eyes are in the position typical among mammals, that is, 
above the first true grinders, M'-M,. The reason that the eyes appear to be so far back is that the lower jaws are 
extended unusually far forward. The upper jaws recede, terminating at the sides in the sockets of the very sharp, 
laterally compressed tusks which at this stage of evolution were chiefly developed as fighting or defensive weapons, 
while only of indirect value as feeding organs. When the upper and lower lips are restored in such a manner 
as to enable the animal to close its moutli, the upper tusks are so largely covered Ijy tlie lips tliat they are not 
especially prominent. In contrast with Mceritlterium the nasal bones and narial openings in Phiomia deeply 
recede; thus a very wide space is left to be filled by the large retractile, prehensile upper lip which could un- 
doubtedly be raised or lowered. 

Evidence against a Proboscis.— The question now arises (Osborn, 1909.332, p. 140): "How far had this 
lip begun to transform into a proboscis? Was there a free projecting proboscis as represented in several previous 
restorations? A negative answer appears to l)e furnished by the structure and mode of wear of the lower incisors. 
Together these form a broad, protrusive, spoon-shaped feeding organ, which is invariably greatly worn on the 
upper surface and somewhat less at the ends. This worn upper surface seems to prove that in tlie prehension of 
food the edge of the upper lip was constantly pressed downward against these teeth, thus, with tlie aid of fine 
particles of grit and sand, which were occasionally taken in, cau.sing wear. In brief, the food appears to have been 
seized between the upper lip and the spoon-shaped lower teeth. Palaeomastodon was a browser, and this lip could 
be turned up and retracted effectively to pull down smaller branches, but there is no reason to suppose that it had 
the free curling and independent prehensile power which characterises a true iiroboscis. If we critically consider 
the theory of the animal possessing a proboscis of considerable length, we find it rests upon the idea of kinship 
with the elephant rather than ujion careful study of the mouth parts themselves." 

If Phiomia possessed an independent prehensile proboscis extending beyond the line of the mouth for the 
seizing of food, we cannot assign any function to the large and much worn paired lower incisor teeth; the most 
probable theory of their function, therefore, seems to be that shown in the model (Fig. 14), namely, the prehension 
of food by pressing the upper lip against the lower incisors, rather than tlie prehension of food by grasping with 
a long and flexible proboscis; this function is progressively developed in the Amebelodont successors of Phiomia, 
also without a proboscis. 

The first rule in tlie restoration of adajitations is not to be too niucli influenced by kinship, but to adhere to 
the evidence afforded by the hard parts theni.selves. According to this rule Phiomia had no proboscis; it had a 
powerful upper lip which pressed against the tips of the lower incisor teeth. Osborn (1933) believes that this 
function was continued into the subfamily of the Amebelodontinae (Longirostrines) to which Phiomia directly 
gave rise. 

49 



4. FOREST-LOVING HABITS OF PAL^EOMASTODON. NOT ANCESTRAL TO MASTODON 

One proof of the forest habits of Palxomastodon is the rarity of its fossil remains; forest-living animals are 
relatively rare. 

The true Palseomastodon, typified by Palseomastodon beadnelli Andrews, 1901, and supplemented by the more 
recently described Palseomastodon intermedius Matsumoto, 1922, is only partly known, namely, the palate, the 
grinding teeth, and portions of the zygomatic arch; therefore we can only form a conjecture as to its habits from 
our knowledge of the animals which appear to be analogous to it, namely, the true mastodonts, known to be of 
forest-living habitat and of leaf -browsing habits. Consequently analysis of the habits of the Fayum proboscideans 



Phiomla wintonr 
Amer. MuS- 13450 Ref 





MoBnthenum 
Amor. Mus, 1343 1 Ref, Irev. 



Fig. 21. Superior molars, compare Matsumoto, 1923, fig. 10: 
" Mceritherium, Palseomastodon, and Phiomia. Left upper grinding 
teeth. A, Mceritherium, Amer. Mus. No. 13431; B, PalseonMslodon, 
Amer. Mus. No. 13449; C, Phiomia, Amer. Mus. No. 13450." All 
figures one-third natural size. 




B Palseomastodon intermedius 
Amer Uus 14547 



All 1/3 nat. size 



Fig. 22. Inferior molars, compare Matsumoto, 1923, fig. 11; 
" M (rrilherium , Palseomastodon, and Phiomia. Left lower grinding 
teeth. A, Maerilherium, Amer. Mus. No. 13437 (reversed), second 
and third lower molars; B, Palseomastodon, Amer. Mus. No. 14547, 
first, second and third lower molars; C, Phiomia, Amer. Mus. No. 
13468." All figures one-third natural size. 



is incomplete in respect to this genus. 
Phiomia and profoundly different from 
animals as follows: 

Mwritherium 
Rivers and swamps 

1. Face abbreviated ; cranium elongated. 

2. Upper and lower second incisor tusks 
of equal size, directly opposed, sheathed 
in enamel; feeding, not fighting, organs. 

3. Eye sockets extremely small, placed 
anteriorly; eyes and auditory openings 
elevated. 

4. Upper and lower lips opposing each 
other. 

5. Molars short and broad, tetra- 
bunodont. 

6. Amphibious or aquatic in habit, 
hence rarelj' fossilized. 



It is certain that the habits and habitat were different from those of 
those of Mceritherium. We may therefore contrast these three Fayum 



Phiomia 
Lowlands and savannas 

1. Face elongated ; cranium normal. 

2. Lower tusks horizontal, no enam- 
el; upper tusks sharply pointed, re- 
curved, enamel band on outer surface; 
fighting, not feeding, organs. 

3. Eye sockets and auditory open- 
ings in normal position. 

4. Upper Up opposing lower incisors. 

5. Molars narrow and elongate, tri- 
lophodont, large central conules 

6. Shore or lowland habitat, hence 
frequently fossilized. 



Palseomastodon 
Forests and savannas 

1. Facial and cranial proportions un- 
known. 

2. Tusks unknown. Upper tusks prob- 
ably circular in section, with lateral enamel 
band. 

3. Eye sockets and auditory openings 
in normal position. 

4. Upper lip probably forming a pro- 
boscis. 

5. Molars relatively broad, hexabuno- 
dont, subtrilophodont, central conules rudi- 
mentary. 

6. Probably of forest habitat, hence 
rarely fossilized. 



50 



IV. REVISION AND SYNOPSIS OF GENERA, SPECIES, AND TYPES OF THE FAYUM 
PROBOSCIDEANS IN THE ORDER OF THEIR ORIGINAL DESCRIPTION 

Revisions (1901-1923).— The Fayum proboscideans described by Andrews (1901-1906), and added to by 
Pontier (1907), Schlosser (1911), Matsumoto (1922-1924), and Petronievics (1923), are so fundamentally im- 
portant and so confused in the present literature that it is necessary to give them a revision by themselves prior 
to the phylogenetic consideration which they will receive in Chapters III, VI, and VIII of this Memoir. This 



LCVCL or LIITAN 0»t«T 



LIITAN Otil 



Afilrrvdon 

Anceiion 

Arnnoit^frmrn 

Metaphuitnya 

Apidium 
ARSINOITHERIUM ZONE 
Arstnoiltwrium 
Jtha^aUitnun fbtaeomajtodon 
HunCLtnum Megalohyraj 
Aruadon PuiternattL 

Sa^AtuAtfiam Phiomyi 
Pttrvdon Apterodon 




Fluvio-marine Beds 
Fhiomia osbomi Type 
Fhiomia serridens Type 
Phiomia wintoni Type 
PInumia minor Tyf)e 
I'aUtumastwUin htadnelli Type 
Palu-omaslodon intinmdius Type 
PaUfomastodon parvus Type 
Mcrritherium anJrewsi Type 
Mwritherium Irigodon Tyjje 



QAi5R-EI.-SAGHA BeDS 

Mwritherium lyonsi Type 
Mwritherium gracile Type 
[Mwritherium anceslraU Type] 



Fig. 23. Section through the I'pper Eocene and Lower OHgocene formations north of Lake Quran, FayOin, Egypt 
Arrows indicate levels richest in ri mains of fossil mammals, .\pproxiniate geologic distribution of proboscidean types shown 
in right-hand column. After Andrews, Beadnell, Granger, andOsborn. Reproduced from Osborn, 1910.346, p. 199, fig. S9 



revision is chiefly based on the authority of Dr. H. Matsumoto (1922, 1923, 1924), who reviewed the entire Fayum 
collection of the American Museum and the original type collection of the British Museum for the purpose, 
inasmuch as neither Osborn nor Andrews had been able to verify these observations in detail. In 1905 Andrews 
(1905, p. 562) observed that the species of the genus Palaeomastodon fall into two groups; in 1922 Matsumoto (1922, 
p. 1) divided these groups into two genera, viz.: Palseomastodon Andrews and Phiomia Andrews and Beadnell. 
In general Osborn accepts Matsumoto's main conclusions, which have a very important bearing on the 
phylogeny of the Proboscidea. 

The geologic levels of the type and genotype specimens are indicated in the above diagram. The Qasr-el- 
Sagha beds are now regarded as Upper Eocene; the Fluvio-marine beds as Lower Oligocene. 

51 



52 



OSBORN: THE PROBOSCIDEA 



SPECIES OF THE FAYUM PROBOSCIDEA IN ORDER OF THEIR ORIGINAL DESCRIPTION 

Original Name Reference in 

Present Memoir 






a 

CQ 
S 

p 



1. 

2. 
3. 
4. 
5. 
6. 
7. 
8. 



Fluvio 

Qasr-el 

Fluvio- 

Qasr-el 

Fluvio 

Fluvio-: 

Fluvio 

Fluvio 



marine 

-Sagha 
marine 
-Sagha 
marine 
marine 
marine 
marine 



beds 
beds 
beds 
beds 
beds 
beds 
beds 
beds 



9. Fluvio-marine beds 



W 10. 
11. 
12. 
13. 



Fluvio-marine beds 
Fluvio-marine beds 
Fluvio-marine beds 
Qasr-eI-Sagha(?) beds 



Palxomastodon Beadnelli Andrews, 1901 
Maeritherium lyonsi Andrews, 1901 
Phiomia serridens Andrews and Beadnell, 1902 
Mceritherium gracile Andrews, 1902 
Moeritherium trigodon Andrews, 1904 
Palxomastodon minor Andrews, 1904 
Palxomastodon parvus Andrews, 1905 
Palxomastodon mintoni Andrews, 1905 

Paleomasiodon Barroisi Pontier, 1907 

Mceritherium Andrewsi Schlosser, 1911 
Palxomastodon intermedins Matsumoto, 1922 
Phiomia osborni Matsumoto, 1922 
Maeritherium anceslrale Petronievics, 1923 



Genotype of Palxomastodon 
Genotype of Mceritherium 
Genotype of Phiomia 
= Maeritherium gracile 
= Maeritherium trigodon 
= Phiomia minor 
= Palxomastodon parvus 
= Phiomia wintoni 
\ Phiomia wintoni ( 9 ?) 
[Phiomia minor (cf ?) 
= Maeritherium andrewsi 
= Palxomastodon intermedins 
= Phiomia osborni 
= Mceritherium ancestrale 




Fig. 24. Geographic distribution of species of the Fayiim Proboscidea according to the heavy face numerals in preceding list. The white dots within the 
black areas represent the approximate localities where the types of these thirteen species were found; the figures in the circles represent the species. 
Mastodun [ = Phiiimia] pygmseus Deperet, 1S97, from Algeria, omitted in this map (see Fig. 189). 



Geologic Levels of Mceritherium Species. — The abo.ve chronologic Hst of species was corrected by 
Andrews himself as to geologic level ; thus the genotype (Moeritherium lyonsi) as well as the type of Moeritherium 
gracile are both from the Qasr-el-Sagha beds, whereas a referred specimen of Moeritherium lyonsi (?) as well as 
the type of Mceritherium trigodon are from the overlying Fluvio-marine series. It is this referred specimen of 
Mceritherium lyonsi (?) which Schlosser selected as the type of Moeritherium andrewsi. 

It is not known whether these smaller, intermediate, and larger species constitute two geologic ascending 
series; it, is probable that they do. 

Quarry Distribution Records. — The shifting river-channel sands and gravels in which the species of 
Moeritherium, Phiomia, and Palxomastodon were deposited are exposed in different quarries; the numerous speci- 
mens referable to these three genera and six species found by the American Museum Expedition of 1907 were 
recorded as occurring in Quarry A, Quarry B, and Quarry C, or at various distances from these chief quarries. 



SPECIES OF THE FAYUM 



53 



Perhaps future examination of these quarry (A, B, C) records will enable us to determine the relative geologic 
levels. So far as examined the American Museum quarry records do not indicate the separation of the species by 
the quarries, e. g., from Quarry B are recorded the following species: Palseomastodon part;us, a single specimen; P. 
intermedius, two specimens ; P. beadnelli, one specimen ; Pliiomin minor, seven or more specimens ; P. wintoni, thirty 
or more specimens. Thus the same (Quarry B of the American Mu.seum records contains three species of Palseo- 
mastodon and two species of Phiomia. This demonstrates that Quarry B represents a very long -period of deposition. 
The very progressive species Phiomia osborni was found eight miles west of Quarry A; it undoubtedly represents 
a higlier geologic level. See full quarry records in American Museum card catalogue, also as printed by Matsu- 
moto (1923, 1924). 

Geologic Levels of Pal^omastodon and Phiomia Species. — Andrews informs us that the type of Palseo- 
mastodon beadnelli came from the very base of the Fluvio-marine .series, 50 feet below the typical level of Phiomia; 
yet P. beadnelli is the largest and most progres.sive species of tlie true Palseomastodon. Unfortunately we have as 
yet no record of the geologic levels on which were found the type specimens into which six species of Palseo- 
mastodon and Phiomia are now divided, namely: 

Palseomastodon beadnelli 
Palseomastodon intermedius 
Palxomxistodon parvus 

The associated mammalian and reptilian fauna with these species of proboscideans is shown in the Andrews 
table (1906, p. ix) reproduced herewith. 



Large, progressive specific stages 
Intermediate specific stages 
Small, primitive specific stages 



Phiomia osborni 
Phiomia unnloni 
Phiomia minor 



i 
•a 




250 


Flu'iio-marine Series (4 in Section). — Variegated sands, sandstones, clays, and 
marls, with liniestoiie-grits and thin bands of limestone. In the lower beds of 
this Series are large numbers of silicified tress assoji-iated with vertebrate remains, 
including: — Arsinoitlurinm zitleh, A. andrewsi, Saghntherium antiquum, S. min'it, 
S. maf/niim, IS. iiHiJus, Meijalohyrax eoc<vnus, M. minor. PaliTornaslodon beadnelli, 
P. wintoni, P. parvus, P. minor, Mocritherium (?) lyonsi, M. trigonodon, Phiomia 
serridens, Ancodon (jorringei, A. jjarvus, A. sp., Rhayatherium (eyyptiacum, 
Oeniohyus mirus, G. fajumensis, G. major, BycFnodon sp., Plerodon ajVicanus, 
Apterodon macrof/7iathus,Sinopn ethiopica, Eremopezus eoc^nus, Procodilus articeps, 
C. megarhitius, Tomistoma gavialoides, Testudo amnion, T. beadnelli, T. isis, 
Stereoyenys lihyca, Podocnemis fajumensis, P. blanckenhorni \\\X\i var. ovata, Pelo- 
medusa progaleata, Aetobatis sp. 




i 

o 

i 

P 


155 


Qasr-el-Sayha Series (Carolia Beds) {5 in Section). — Alternating limestones, marls, 
clays, and sandstones. The vertebrate remains include : — Mceritherium lyonsi, 
M. gracile, Barythcrium grave, Eosiren libyca, Zeuglodon osiris, Crocoddus sp., 
Tomistoma africanum, Psephophorus eocttnus, Thalassochelys libyca, Podocnemis 
antiqua, P. stromeri and \ar. major, Stereogenys cromeri, S. podocnemioides, 
Gigantophis garstini, I'terosphenus achwcinfurlhi, Fajurnia schweinfurthi,SocnOj>a'a 
gra/idis, Prtstis fajumensis, P. ingens; Eopristis reinachi, Propristis schweinfurthi, 
Ji/yliobatis sp., Carcharodtin sp. 







Fig. 25. Complete faunal list 
(1906) of the Upper Eocene (155) 
Qasr-eKSagha Series, also of the 
Ivowcr Oligocene (250) Fluvio- 
marine Scries of .\ndrcws. .\ftcr 
.\ndrcws, 1906, p. ix. Compare 
figures 11, 12, and 23. 



Original Descriptions and Present Reference. — The remaining pages of the present Chapter II are 
devoted to the firm establishment of Charles W. Andrews' types, tijpe figures, type descriptions, and type geologic 
levels, in accord with the monographic system which prevails throughout this Memoir. It is a token of Andrews' 
signal ability as a palaeontologist that he interpreted so accurately the chief characters of this proboscidean fauna 
from materials invariably scattered and more or less fractured in the course of fluvio-marine deposition. In 
italics is given Andrews' original reference, e.g., Palseomastodon minor Andrews, 1904; in heavy face is given 
the present generic and specific reference, e.g., Phiomia minor Andrews, 1904. It is earnestly hoped that these 
references, representing years of labor on the part of Aiulrews, Matsumoto, and the present writer, may stand 
and be accepted as a basis of future monographic research. 



ORIGINAL DESCRIPTIONS IN CHRONOLOGIC ORDER (1901-1923) 
COMPARE WITH REVISED DESCRIPTIONS IN CHAPTERS III, VI, AND VIII 



Palseomastodon beadnelli Andrews, 1901 

Figures 2, 26, 34, 41, 41a, 90-95, 97; see also page 147 

Fluvio-marine formation of the Faydm, Egypt = Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

This was the first proboscidean to be described from the 
Fay6m and was properly given the name Palxomastodon, namely, 
the ancient Mastodon. 

(Andrews, letter, 1922) : The type was found at the very base 
of the Fluvio-marine beds, Lower Oligocene, 50 or 100 feet below 
the Phiomia level; this genotype specimen, with a very large 
femur and humerus, was the only true Palxomastodon material 
recovered. 

Palxomastodon Beadnelli Andrews, 1901. Zoologist, Vol. V 
(4), August 15, 1901, pp. 318, 319 (Andrews, 1901.1); name only. 
Tageblatt des V Intern. Zoologen-Congresses, Berlin, No. 6, 
August 16, 1901, p. 4 (Andrews, 1901.2), notice; published volume 
(Verhandlungen) dated 1902, see especially p. 528. Gcol. Mag., 
Dec. IV, N. S., Vol. VIII, pp. 400-109, September, 1901 (Andrews, 
1901.3); description and figure. Type.— (Op. cit., 1901.3, 

p. 401): . . . " nearly complete left ramus of the mandible of a 
Proboscidean." In Andrews' supplementary description "A 
Descriptive Catalogue of the Tertiary Vertebrata of the Fayijm, 




FiQ. 1. — Left ramus of mandible of PaUeomaatodcn Beadnelli. One- sixth natural 
size. (A) From above ; (B) outer surface. 

Fig. 26. Original type figure of Paia°o)«a.5(o(fon beor/Hcffi Andrews, 1901. 
After Andrews, 1901.3, p. 401, fig. 1, A, B. {Op. cit., p. 401): . . . "nearly 
complete left ramus of the mandible of a Proboscidean." Original (Geol. 
Mus., Cairo, C. 10014) destroyed; cast Brit. Mas. M. 8059; cast Amer. 
Mus 9984. One-sixth natural size. 



Egypt" (1906, p. 150, Pis. xii-xvi, text figs. 50, A, 51) the type 
specimen is given as "Left ramus of mandible with pm. 4 (broken) 
and m. 1-3 (PI. xv, figs. 1, 1a); Geological Museum, Cairo." 
Original (Geol. Mus., Cairo, C. 10014) destroyed in Custom House 
at Cairo; cast in British Museum (M.8059), also cast in American 
Museum (Amer. Mus. 9984). Horizon and Locality. — 



Fluvio-marine beds, Lower Oligocene, Fay6m, Egypt. Type 

Figure.— Andrews, 1901 3, p. 401, fig. I, A, B. 

Andrews (1901.3, p. 401) observes: "One of the most impor- 
tant specimens found in the higher beds (probably Lower Oligocene) 
is the nearly complete left ramus of the mandible of a Proboscidean, 
which is in many respects similar to that of Mastodon angustidens, 
but belonged to a much smaller and in several respects more gen- 
eralized form." Andrews (1904, p. 115) added the following meas- 
urements of the type and supposed female specimen of P. bead- 
nelli in comparison with P. minor: "This species may be called 
Palxomastodon jnirwr; its dimensions compared to those of P. 
beadnelli are shown in the following table . . . 

P. [ = Phiomia] minor P. beadnelli P. beadnelli 

(? female) (type) 

m. 3 47 mm. 65 mm. 78 mm. 

m. 2 45 " 55 " 65 " 

m. 1 32 " 41 " 48 " 

pm.4 28 " 39 " 48 " 

pm.3 28 " 30 " ? 

Matsumoto (1922, p. 3) confirms Andrews' description of 1901 
and adds Amer. Mus. 13481 as a referred specimen. In 1924 
(p. 4) he distinguishes Palxomastodon beadnelli from P. interme- 
dins and P. parvus as follows: 

"3. — Length of lower molar series measuring 194 mm.; 
that of lower premolar and molar series, 285 mm. (Andrews' 
type) beadnelli 

2. — Length of lower molar series measuring 159 mm. (type 
[Amer. Mus.] 14547); that of upper molar series, 150-152 
mm. (paratype [Amer. Mus.] 13449) ; that of upper premolar 
and molar series, 250 mm. (ditto) intermedins. 

1. — Length of lower molar series measuring 130 mm.; 
that of lower premolar and molar series, 197 mm. 
(Andrews' type) parvus." 

Moeritherium lyonsi Andrews, 1901 

Figure.*; 2, 27, 41a, 42, 42a, 43, 44, 49; see also page 72 
Qasr-el-Sagha formation of the Faylim, Egypt = Middle Eocene of 
Andrews = LTpper Eocene of the present Memoir. 

This was the second proboscidean to be described from the 
FajTim and was very appropriately given the name Mceritherinm, 
in reference to the ancient Lake Moeris of the Greeks. 

Moeritherium lyonsi Andrews, 1901. Tageblatt des V Intern. 
Zoologen-Congresses, Berhn, No. 6, August 16, 1901, p. 4 
(Andrews, 1901.2), notice; pubUshed volume (Verhandlungen) 
dated 1902, see especially p. 528. Geol. Mag., Dec. IV, N. S., 
Vol. VIII, pp. 400-409, September, 1901 (Andrews, 1901.3), 
description and figure; also "A Descriptive Catalogue of 
the Tertiary Vertebrata of the Fay^m, Egypt," 1906, 
p. 120, Pis. VIII, IX, X. Type.— (Andrews, 1906, p. 120): 



54 



SPECIES OF THE FAYUM 



55 



"Mandible associated with upper molar.s and a dorsal vertebra." 
Geol. Mus., Cairo, C. 10000; east Amer. Mus. 9977. Hori- 

zon AND Locality. — Qasr-el-Sagha beds, Upper Eocene, 
Fay(lm, Egypt. Type Figure. — Andrews, 1901.3, p. 404, 

fig. 2. 




Pio. 2. — Dentition of Maritherium Lyonsi. One-fourth natural size. (A) Upper 
teeth ; (B) front of snout, showing the tusk-like second incisors ; (C) left 
ramus of mandible from outer side. 

Fig. 27. OriKiiial type figure of M wrilhenum lyonsi Andrews, 1901. 
.'Vfter Andrew.s, 1901.3, p. 404, fig. 2. (Andrew.s, 1906, p. 120): "Mandible 
associated with upper molars and a dor.sal vertebra." Geol. Mus., Cairo 
(C. 10000); cast Amer. Mus. 9977. One-fourth natural size. (.Andrews, 
1901.3, p. 404, fig. 2): A, Upper teeth (C.lOOOl) = paratype; A, B, portions of 
large skull, B showing front of snout with tusklikc .second incisors (C. 10002, 
cast Amer. Mus. 9978) =paratype; C, left ramus of mandible from outer side 
(C. 10000) = type. (See Fig. 44.) 



The referred Mocrtthcrium lyonsi (?) recorded by Andrews in 
the overlying Fluvio-marine beds was chosen as the type of 
Mmrithcrium andrewsi by Schlo.sser. Matsumoto (1923, p. 124) 
distinguishes the two species of the Qasr-el-Sagha formation as 
follows : 

"(1.) Larger form of the Qasr-el-Sagha Formation. Lower 
premolars very short: P^'' [P2-4], ca. 69 mm. (Andrews); 
lower molars very long, M1-3, ca. 104 mm. (Andrews). All the 
lower cheek-teeth very wide. Pj, triangular, its widest part 
corresponding to the posterior lobe. P^'^, 67-78 mm.; 
M'"', 85 mm. (Andrews) M. lyonsi. 

(2.) Smaller form of the Qasr-el-Sagha Formation. Lower 
premolars not very short in comparison with the length of 
lower molars: Po-i, ca. 62 mm. (specimens in the American 
Mu.seum); lower molars very short: M1-3, 83 mm. (specimen 
in the American Museum). All the lower cheek-teeth are 
narrow. P2■^ 62 mm. (Andrews); M'"', 75— ca. 79 mm. 
(Andrews) M. gracile." 

Matsumoto {op. cit., p. 125) adds the following information 
from the American Museum collection: 

Amer. Mus. 13444 ; " two of the three fragments of mandibular 
rami of this specimen number appear to belong to this species. 



They are very peculiarly weathered, as a characteristic of the 
weathered specimens from the Qasr-el-Sagha Formation, with 
much-weathered and badly preserved molars in situ. Qasr-el- 
Sagha Formation of the Fayflm." 

"The dimensions of the teeth of these fragments, in compari- 
son with those of Andrews' specimens, are tabulated as follows 
(in mm.): 





Lower Teeth 


Upper Teeth 


[Amer. Mus.] 
13444 ditto 


(Andrews) 


P 2 1'^"^*^ 
1 width 


(Andrews) 








22 
16 


27 
23? 




p g jlength 
1 width 




23 
21 


26.5 
29.5 




p 4 [length 
l^width 




25 
23 


23 
27.5 




Mir"f.f 

1 width 




. 26.5 
24.5 


29 

27 




M2psf 
1 width 


29 28 35 
25 25 39 


26? . 
23.5 : 


30 

28 


M ^hf'^ 

(width 


40 39 42 
28 28 30 


.... ; 


52 . . 37±' 
28 . . 30±' 


Length of P 2-4 


.. .. 69±' 


68 


. 67 


Length of M 1-3 


. . . . 104±' 




. 85 



'"These dimensions are estimated from Andrews' figures." 

Phiomia serridens Andrews and Beadnell, 1902 
Figures 28, 180, 181; see also page 239 

Fluvio-marine formation of the Fayflm, Egypt = Upper Eocene of 
.\ndrews = Lower Oligocenc of the |)resent Memoir. 

This was the third proboscidean to be described from the 
Fayum and was appropriately given the name Phiomia, the Greek 
equivalent of the province now known as the Fayilm. 

Phiomia serridens Andrews and Beadnell, 1902. "A Pre- 
liminary Note on Some New Mammals from the Upper Eocene 
of Egypt," Survey Dept., Pub. Works Ministry, Cairo, 1902, pp. 
1-9 (original description); "Descriptive Catalogue of the Terti- 
ary Vertebrata of the Fayiim, Egypt," Andrews, 1906, pp. 169-171 
(supplementary description), PI. xviii, figs. 4, 4a. Type.— 

Anterior portion of a left mandibular ramus, bearing Di2 and Dp2-3 
in situ; Geological Museum, Cairo (C. 10007); cast Amer. Mus. 
9981. Horizon and Locality. — Beds of the Fluvio-marine forma- 
tion (Lower Oligoccne), Fayftm, Egypt, north of Birket-el-Qurun. 
Type Figure.— Andrews and Beadnell, 1902, figs. 1, 2. Matsu- 
moto (letter, 1921) remarks of the referred specimen shown in fig- 
ure 3 {op. cit., 1902) that it is not the type and not a proboscidean. 

Matsumoto (letter, 1921): Schlosscr first pointed out that 
this genus and species might be merely a juvenile form of a 
Pal^omaslodon [i. e., Phiomia] (Neues Jahrb. f. Min., Vol. I, Pt. 1, 
1905, p. 157, Refcrate); Andrews accepted Schlosser's view, after 
securing a beautifully preserved mandible of a very young Palseo- 



56 



OSBORN: THE PROBOSCIDEA 



mastodon [i.e., Phiomia] probably referable to Ph. imnloni (Andrews, 
Geol. Mag., Dec. V, N.S., Vol. IV, 1907, p. 97); though the type 
specimen is merely a very young individual, it shows certain indica- 
tions that it might belong to the wintoni-minor type of Palxomas- 
todon [i. e., Phiomia], as distinguished from the beadmlli-parvus 
type ; moreover, the specific name Phiomia serridens antedates any 
of the specific names of the wintoni-minor type. 

Matsumoto (1924.1, p. 14) remarks that the genotype of 
Phiomia serridens Andrews is a species probably identical with 
Palseomastodon [ = Phiomia] wintoni Andrews, 1905; or possibly 
with Palxomastodon { = Phiomia] minor Andrews, 1904. 

Osborn, 1924: The type (genotype) cast of Phiomia serridens 
is close in size to P. wintoni; this observation is validated by 
another but slightly larger milk tooth specimen, namely, Amer. 
Mus. 13458. 



/ 

■•■■ f~-\ 



y" 




PHIOMtA SERRIDENS. Anterior portion of left ramus 
of the mandible Dorsal aspect 




Geol. Mus. Cairo (0.10003); cast Amer. Mus. 9979. Hori- 

zon .4^ND Locality.— (Andrews, 1906, p. 127): "Qasr-el-Saghabeds 
(Middle [Upper] Eocene): north of Birket-el-Qurun," Fayum, 
Egypt. Type Figure.— Andrews, 1906, PI. xvii, figs. 1, 2. 

Matsumoto (1922, p. 5, 1923, p. 125) rightly characterizes 
Moeritherium gracile as the small species of the Qasr-el-Sagha beds. 
His definition of this species is as follows (op. cit., 1923, p. 125): 

"(2.) Smaller form of the Qasr-el-Sagha Formation. Lower 
premolars not very short in comparison with the length of 
lower molars: Pi-t, ca. 62 mm. (specimens in the American 
Museum); lower molars very short: M1-3, 83 mm. (specimen 
in the American Museum). All the lower cheek-teeth are 
narrow. P""*, 62 mm. (Andrews) ; M'"', 75 — ca. 79 mm. 
(Andrews) M. gracile." 

The additional information based on the American Museum 
collection from the Qasr-el-Sagha formation of the Fayum is 
quoted below (Matsumoto, 1923, p. 128) : 




PHIOMIA SCRRIOCNS . Anterior portion of left ramus 
of tfie mdndible . Lateral aspect 

Fig. 28. Original type figure of Phiomia serridens Andrews and Bead- 
nell, 1902. After Andrews and Beadnell, 1902, figs. 1 and 2. Anterior portion 
of a left mandibular ramus, bearing Di; and Dp2.3 in situ: Geol. Mus., Cairo 
(C. 10007), cast Amer. Mus. 9981. One-half natural size. 

Moeritherium gracile Andrews, 1902 

Figures 29, 41, 42a, and 45; sec also page 73 
Qasr-el-Sagha formation of the Faj-tlm, Egypt = Middle Eocene of 
Andrews = Upper Eocene of the present Memoir. 

Moeritherium gracile Andrews, 1902. "Preliminary Note on 
some Recently Discovered Extinct Vertebrates from Egypt," 
Geol. Mag., Dec. IV, N. S., Vol. IX, pp. 291-295 (original descrip- 
tion); "Descriptive Catalogue of the Tertiary Vertebrata of the 
Fayum, Egypt," 1906, pp. 127, 128 (supplementary description), 
PI. XVII, figs. 1, 2. Type.— (Andrews, 1906, p. 127) : "An im- 

perfect skull . . . including the palatal region, associated with 
cervical, dorsal, and lumbar vertebrae ; Geological Museum, Cairo." 



prru.3. \ 
pm.2 \ \/ JWttlI 



rw.2. 



mi. 3 



!_>•.■ {^yi\ ■ 



m'i^ 




Fig. 29. Original tyjio figure of Moeritherium gracile Andrews, 1902. 
After Andrews, 1906, PI. xvii, figs. 1, 2. {Op. cit., p. 127): "An imperfect 
skull . . . including the palatal region, associated with cervical, dorsal, and 
lumbar vertebra;;" Geol. Mus., Cairo (C. 10003), cast Amer. Mus. 9979. 
(Fig. 1) . . . "anterior portion of skull, palatal view: one-third nat. size. 
Qasr-el-Sagha beds (Middle Eocene)." (Fig. 2) . . . "occipital region of 
same skull, posterior view; one-third nat. size. Same horizon." 

Amer. Mus. 13443; "mandible, with P3-M:, of the left side 
and Mi-3 of the right side in situ. [Amer. Mus.] 13444; one of the 
three fragments of mandibular rami of this specimen number, with 
badly preserved molars t« sj<M. [Amer. Mus.] 1.3445; fragment of 
a right mandibular ramus of a young individual, with the teeth 
broken away. [Amer. Mus.] 13446; fragment of a left mandibular 
ramus, with the crowns of the teeth broken away. All from the 
Qasr-el-Sagha Formation of the Fayum." 

"The mandible of [Amer. Mus.] 13443 measures 305 mm. in 
length without incisors, 8.5 mm. in length of symphysis, 55 mm. 
and 50 mm. in the distance between the two first molars and the 
two last molars respectively, 225 mm. in the bicondylar width and 



SPECIES OF THE FAYUM 



57 



76 mm. in the height of the mandil)ular ramus at Mj without 
the teeth. In this specimen the symphysial depression already 
cited is observed to be present. In the fragmentary mandible of 
[Amer. Mus.] 13446, the same depression is clearly oh.scrvcd, also." 
"The dimensions of the cheek-teeth of the specimens at hand, 
in comparison with those of Andrews, are tabulated as follows (in 
mm.): 









Upper 




Lower Teeth 




Teeth 


[Amer. Mus.) 


[Amer. 






13443 


Mus.] 


(Andrews) 




right-left 


13444 




p 9 (length 
^ ^ \ width 






22 .... 






18 .... 


/length 
^ -^ 1 width 


22 




20 .... 


15 




23 .... 


/length 
^ ] width 


21 




20 .... 


18 




21? .... 


M'fSt 


23 23 


22 


23 25 


20 19 


19? 


23 21 


"^fcr 


28 28 


27 


24 27 


22.5 23 


22 


25 23 


M 3 /'"*'' 

\ Width 


34 33 




28 28 


24 24 




24 25 


Length of P 2-4 


63 ± 63 ± 
(alveoli) (tlitto) 




62 .... 


Length of M 1-3 


84 84 




75 79±' 



the three molars: of these m.3 is in perfect and unworn condition, 
while m.2 and m.l have lost portions of their outer sides." Brit. 
Mus. M.8499; cast Amer. Mus. 9980. Horizon and 

Locality. — (Andrews, 1906, p. 128): "Fluvio-marine beds 
(Upper Eocene [Lower Oligocenel); north of Hirket-el-Qurun," 
FayAm, Egypt. Type Figure. — Andrews, 1906, PI. ix, fig. 5. 



m. 1- 



'"This dimension is estimated from Andrews' figure." 

"Thealveoli of each Ii and Pof the mandible of [Amer. Mus.] 13443 
measure 10 mm. and 20 mm. in transverse diameter, respectively; 
the lateral extension of, and the minimum distance between, the 
two alveoli of lower tusks are 50 mm. and 9 mm. respectively. 
The two alveoli of first inci.sors are .situated just lielow and anterior 
to the part corresponding to minimum distance between the two 
alveoli of tusks. Judging from these alveoli, the lower first incisors 
might be located not strictly inside, but inside and below and 
anterior to, the pair of lower tusks, which might i)e rather closely 
set to each other. These lower tusks appear to be distinctly smaller 
than those of M. lijonsi and andrewsi." 

MoBritherium trigodon Andrews, 1904 
Figures 30, 42, 42a, 46, 49; see also page 74 
Huvio-marine formation of the FayOm, Egypt = Upper Eocene of 
.\ndrews = Lower Oligoeenc of the present Memoir. 

Maerilherium trigodon Andrews, 1904. "Further Notes on the 
Mammals of the Eocene of Egypt," Geol. Mag., 1904, Dec. V, 
N. S., Vol. I, pp. 109-115 (original description); "Descriptive 
Catalogue of the Tertiary Vertebrata of the Faytim, Egypt," 
1906, pp. 128, 129 (supplementary description, including change of 
name, \. c.,trigodoniotrigonodon). Type. — (Andrews, 1904, 

p. 112): . . . "portion of the right ramus of a mandible containing 



m. 2. 




Fig. 30. Original type figure of Mwrilherium trigodon Andrews, 1904. 
After Andrews, lOOt), PI. ix, fig. 5. (Andrews, 1904, p. 112): . . . "portion 
of the right ramus of a mandible containing the three molars: of these m.3 
is in ix>rfect and unworn condition, while ni.'2 and m.l have lost portions of 
their outer sides." Brit. Mus. M.S499; cast Amer. Mus. 9980. Natural size. 



Matsumoto, 1923, p. 134. — Mceritherium trigodon is a smaller 
form of the Fluvio-marine formation. In revi.sing the American 
Museum collection, Matsumoto refers the following specimens to 
this species (p. 134) : 

Amer. Mus. 13430; "greater part of a full-grown skull, bear- 
ing all the upper cheek-teeth in situ." 

Amer. Mus. 13431; "fragment of a skull, including a greater 
part of a right half of palate, bearing P'-M' in situ." 

Amer. Mus. 13433; "left M- attached to a fragment of upper 
jaw." 

Amer. Mus. 13435; "fragment of a right ramus of mandible, 
with all the cl eek-teeth in situ." 

Amer. Mus. 13436; "fragment of a left ramus of mandible, 
with all the cheek-teeth in situ." 

Amer. Mus. 13439; "right l\ and M,, with their roots 
broken away." 



5S 



OSBORN: THE PROBOSCIDEA 



On this material M. trigodon is redefined as follows (Matsu- 
moto, 1923, p. 125): 

"(4.) Smaller form of the Fluvio-marine Formation. 
Lower premolars not very short in comparison with the length 
of lower molars: P2-4, ca. 63 (specimen in the American Mu- 
seum) — 70 mm. (ditto, as well as Andrews) ; lower molars 
rather short : M1-3, 93 (specimens in the American Museum) — 
98 mm. (Andrews). All the lower cheek-teeth very narrow: 
P2, fusiform in upper view, its widest part corresponding to 
the middle part. P^"^ 60-63 mm. (specimen in the American 
Museum); M'-^, 83-85 mm. (ditto). Skull, small and lightly 
built. Sagittal and occipital crests very strong. Zygomatic 
width very large in comparison with the length of skull. 
Distance between the external auditory openings, as well as 

the width of occiput, very large M. trigodon." 

Matsumoto {op. cit., p. 134) gives a detailed description of the 
six specimens in the American Museum listed above as referred 
to this species, concluding (p. 137) with the detailed measure- 
ments of the grinding teeth as follows (in mm.) : 



9982. Horizon and Locality. — (Andrews, 1906, p. 168): 

"Fluvio-marine beds (Upper Eocene [Lower Oligocene]): north 
of Birket-el-Qurun," Faydm, Egypt. Type Figure. — 

Andrews, 1906, PI. xiv, figs. 1, 1a, and text fig. SOD, p. 143. 

Type Description. — (Andrews, 1904, p. 115): "A portion of 
the right ramus of a mandible shows that there existed in the Upper 
Eocene beds a species of Palseoma-siodon considerably smaller than 
P. beadnelli, even allowing for a very wide range of individual varia- 
tion in size in that species. . . . This species may be called Palxo- 
maslodon minor ; its dimensions compared to those of P. beadnelli 
are shown in the following table . . . 





P. 


= Phiomia] 


minor 


P. beadnelli 
(?female) 


P. beadnelli 

(type) 


m. 3 




47 mm. 




65 mm. 


78 mm. 


m. 2 




45 " 




55 " 


65 " 


m. 1 




32 " 




41 " 


48 " 


pm. 4 




28 " 




39 " 


48 " 


pm. 3 




28 " 




30 " 


? 





Lower Teeth 




Upper Teeth 




[A. 


ner. Mus.] 


(Andrews) 




[Amer. Mus.) 






13435 


13436 13439 


13430 


13431 


13433 








right 


left 




p 2 (length 
1 width 


19.5 


23? 


23 . . 


19 


22? 




10 


11 


11 


21 


18? 






P 3 /'^"^^'^ 
1 width 


22 


24? 


25 . . 


23 


23.5 24 






15 


14? 


17 


26 


24? 25? 






p ^ jlength 
1 width 


23 


23.5 25 


24 22 


20.5 


20 22 






18 


18.5 19 


20 


24 


23? 25.5 






Iwidth 


27 


24 27.5 26 . 


27 


25 


22 24 






21 


19 21 




21 


25 


26.5? 24 






«4s 


29 


31 


32 




29 


28 30 


29 


24 


24 






28 


27 25? 


24 


« n.S' 


37 


37 


40 




31 


32 30 




26.5 


25 


24 




26.5 


27 26 




Length of P 2-4 


63 


70 


7 


.. 


60 


63 




Length of M 1-3 


93 


93 


98 . 




85 


83 84 





Phiomia minor Andrews, 1904 
Figures 31, 34, 178, 179, 182, 183; see also page 239 
Fluvio-marine formation of the Fajilm, Egypt = Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

Palxotna.'itodon minor Andrews, 1904. "Further Notes on 
the Mammals of the Eocene of Egypt," Geol. Mag., 1904, Dec. V, 
N. S., Vol. I, p. 115 (original description); "Descriptive Catalogue 
of the Tertiary Vertebrata of the Fayflm, Egypt," 1906, pp. 168, 
169 (supplementary description), PI. xiv, figs. 1, 1a, and text fig. 
50 D, p. 143. Type.— (Andrews, 1904, p. 115): . . . "part 

of the ramus [right] and the coronoid process of an immature 
mandible, in which m.3 has not yet been cut, although it is 
completely developed." Brit. Mus. M.8479b; cast Amer. Mus. 



(Andrews, notes, 1922): " These measurements [of P. bead- 
nelli] are no doubt from a specimen of P. ivintoni." 

Matsumoto (1924.1, p. 16) reviews the type oi Palxomastodon 
minor Andrews and transfers it to the genus Phiomia. He then 
compares it very carefully with specimens referred by Andrews to 
Palieomastodon beadnelli and to P. [Phiomia] wintoni, also with 
the type of Pabeomaslodon barroisi Pontier, and concludes that 
Phiomia minor is a valid species which may be clearly distin- 
guished from both Phiomia wintoni and P. osborni. 

He refers to Phiomia minor (1924, pp. 18, 19) eighteen speci- 
mens in the American Museum collection, of which the principal 
numbers are the following (also Matsumoto, 1922, p. 3) : Ameri- 
can Museum referred specimens Nos. 13469, 13471, 13475, 13483, 



SPECIES OF THE FAYUM 



59 



13486, 13448, 13455, 13461, 13464, 13465, 13467. Of the total 
number of eighteen specimens it is important to note that Quarry B 
and vicinity yielded ten specimens; Quarry B also yielded Phiomia 
wintoni in abundance. (Quarry A yielded three specimens of P. 
minor, also /■'. wintoni; one important specimen, namely, Amer. 
Mus. 13469, was found eight miles west of Quarry A. 




Type of Phio.mi.\ minor 
Fig. 31. Original type figure of Palsenmaxlodnii minor ,\ndrews, 1904. 
.After .\titlrew.s, 1906, PI. xiv, figs. 1, 1.^, .and text fig. 50 D (see present Memoir 
Fig. 34D), p. 143. (.\ndrew.s, 1904, p. 115): . . . "part of the ramus [right] 
and the coronoid proeess of an immature mandible, in which m.3 has not yet 
been cut, although it is completely developed." Brit. Mus. M. 8479b, cast 
Amer. Mus. 9982. One-half natural size. 

The following (Matsumoto, 1924.1, p. 20) are the detailed 
measurements of the mandibles: 

"The mandibles of the specimens Nos. 13469, 13471 and 13475, 
in comparison with one described by Andrews, measure as follows 
(in mm.). 



Normal and Variant Characters of P. minor (Mat- 
sumoto, 1924.1, pp. 19-21).— (1) Amer. Mus. 13469 bears an extra 
pair of incisors, I3 or C, just behind the enlarged \, or C; these 
e.xtra teeth are either third incisors or canines. (2) Alandible of 
Amer. Mus. 13471 very large in its dimensions [referable to 
Phiomia wintoni]. (3) Skull (Amer. Mus. 13448) bearing all the 
ui)por cheek teeth, P^-M'*, measured and described by Matsumoto 
in detail. 

Palssomastodon parvus .\ndrews, 1905 

Figuns 32 and 34; .see also page 146 

Fluvin-rnarine formation of the Fajtlm, Egypt = Upper Eocene of 
.\nilrews = l./0\ver Oligocenc of the present Memoir. 

Palsrnmastodon parvus Andrews, 1905. "Note on the Species of 
I'alscomastodon," Geol. Mag., 1905, Dec. V, N. S., Vol. II, pp. 
562, 563 (original description); "Descriptive Catalogue of the 
Tertiary Vertebrata of the Fayflm, Egypt," 1906, pp. 162-168 
(supplementary description), text figs. 50C, p. 143, and 55, p. 163. 
Type. — (Andrews, 1905, p. 562): . . . "right ramus of the 
mandible, with the premolars and molars in silu, though somewhat 
crushed." Brit. Mus. M.8479a; cast Amer. Mus. 9976. Hori- 

zon AND Locality. — (Andrews, 1906, p. 163): " Fluvio-marine 
beds (Upper Eocene [Lower Oligocene]) : north of Birkct-el- 
Qurun," Fayum, Egypt. Type Figure. — Andrews, 1906, text figs. 
50 C, p. 143, and 55, p. 163. 

Type Description (Andrews, 1905, p. 562). — In his paper 
Andrews observes the two sections of PateowasWo/i, subsequently 
separated by Matsumoto into Pnlxomastodon and Phiomia: 

"The species of Palxomaslodon fall into two sections, in one of 
which the posterior end of the syinphj'sis of the mandible is situated 
considerably in front of the level of the anterior premolar, while in 
the other it is only very little in front of that point. The first 
group, moreover, is distinguished by the comparative simplicity 
of the molars, in which the accessory cusps are scarcely at all 
developed, and by the small size of the talon of the last lower molar; 
into this subdivision the original species, P. beadnelli, falls, 



Measurements of Phiomia minor 


[Amer. Mus.] 


[Amer. Mus.] 


[Amer. Mus.] 






13469 


13471 


13475 


Andrews 


Amer. Mus. 13471 = Phiomia wintoni 


Young; 




Young; 




Length from Tip of Symphysis to Posterior Side of 


prob. 9 


prob. cf 


prob. 9 


prob. cf 










Angle 


435 db 


590 ± 


.... 


600 


Length of Symphysis 


137 


215 




226 


Length from Tip of Symphysis to Posterior Side of M^ 




435 




475 


Minunum Antcro-posterior Width of Ascending Bar 


122 


155 






Maximum Width of Anterior Half of Symphysial 










Region 


65 






.... 


Minimum Width at the Constriction of Symphysial 










Region 


60 






55 


Height of Ramus at P^ 


61 


83 


70 




Ditto at My 


59 


90 


72 




Ditto at Anterior Lobe of M^ 




82 




.... 


Height of Ascending Bar at Condyle 


112=b 


190 







60 



OSBORN: THE PROBOSCIDEA 



together with a much smaller form for which the name P. parvus 
is now proposed. The type-specimen of this new species is the 
right ramus of the mandible, with the premolars and molars in 
situ, though somewhat crushed." 

Matsumoto (1922, p. 2) confirms Andrews' type description 
of 1905 and type figure of 1906 and adds as a single Ameri- 
can Museum specimen No. 13497 from the upper Fluvio-marine 
formation of the Faydm. (Matsumoto, 1924.1, p. 4): "Length 
of lower molar series measuring 130 mm.; that of lower pre- 
molar and molar series, 197 mm. (Andrews' type). . . . [P.\ 
parvus." {Op. cit., p. 4): "[Referred] Specimen. — [Amer. Mus.] 
No. 13497; a left lower third molar; Amer. Mus. E.xp. 1907, 
Quarry B, Fluvio-marine formation, Fayum, Egypt. This tooth 
measures 52 mm. in length and 32 mm. in width. It is longer 
than and as wide as the lower third molar of Andrews' type, 
which is .stated by Andrews to be 46 mm. long and about 32 
mm. wide." 

Osborn, 1924: Osborn confirms Andrews' and Matsumoto's 
opinion as to the specimens of P. parvus, namely, the unique type 
and the referred specimen, and gives a description also measure- 
ments in comparison with Palseoviastodun beadnelli and P. 
intermedius (Chap. VI, pp. 143, 146, 147). 

Text-fig. 55. 



Teil-fig. 53, 



pm 



[tmt 




Bigbt ramus of maodible of Pal<tomastodon parvus. Type specimen. 

Fig. 32. Original type of Palxomastodon parvus Andrews, 1905. After 
Andrews, 1906, text fig. 55, p. 163 (see also ovir Fig. 34C). (Andrews, 1905, 
p. 562): . . . "right ramus of the mandible, with the premolars and molars 
in situ, though somewhat crushed." Brit. Mus. M.S479a; cast Amer. Mus. 
9976. One-fourth natural size. 

Phiomia wintoni Andrews, 1905 
Figures 33, 34, 178, 179, 182, 183; see also page 24 1 
Fluvio-marine formation of the Fayllm, Egypt =Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

Palseomastodon unntoni Andrews, 1905. "Note on the Species 
of Palseomastodon," Geol. Mag., 1905, Dec. V, N. S., Vol. II, pp. 
562, 563 (original description); "Descriptive Catalogue of the 
Tertiary Vertebrata of the Fayum, Egypt," 1906, pp. 156-162 
(supplementary description), text fig. 53, p. 157. Type. — 

(Andrews, 1905, p. 563): ... "a mandible with the incisors and 
posterior molars TO S7'<?«." (Andrews, 1906, p. 156): "A mandible 
wanting the angular region and the anterior cheek-teeth. . . ; 
British Museum." Brit. Mus. M.8414. Cast (Amer. Mus. 14557) 
of two teeth from type mandible, r. M2-3. Horizon and 

Locality. — (Andrews, 1906, p. 157): "Fluvio-marine beds 
(Upper Eocene [Lower Oligocene]): north of Birket-el-Qurun," 
Fayflm, Egypt. Type Figufe. — Andrews, 1906, p. 157, 

text fig. 53. 




Mandible of Paltrtmaitodon mntoni, type Bpecimen, from abo\e. 
eond.. condyle ; cor., coronoid procegs ; d.e., dental canal ; i., incisor. The premolars and first molar 
on the left side have been restored fnm another specimen. 

Type of Phiomia wintoni 
Fig. 33. Original type figure of Palxomastodon unntoni Andrews, 1905. 
After Andrews, 1906, text fig. 53, p. 157. (Andrews, 1905, p. 563): ... "a 
mandible with the incisors and posterior molars z« situ." (Andrews, 1906, p. 
156): "A mandible wanting the angular region and the anterior cheek-teeth 
. . . ; British Museum." Brit. Mus. M.8414; cast Amer. Mus. 14557. One- 
eighth natural size. 

Type Description.— (Andrews, 1905, p. 563): . . . "the 
other is a much larger form, and is important as being 
probably the commonest of all the species; for this the 
name P. unntoni is suggested, in honour of Mr. W. E. de 
Winton, by whom the expenses of my third collecting trip to 
the Fayilm were defrayed. It is considerably larger than 
P. parvus and P. minor, but smaller than P. beadnelli. From this 
last species it is also distinguished by (1) the extension back of 
the symphysis so that its posterior end is only just in front of the 
anterior premolar; (2) the position of the mental foramen on 
the side of the symphysis instead of behind it; (3) the greater 
complication of the molars and their greater length in proportion 

Teit-fig. 50. 



lur^^^fiAN, 




^^j£iiiLit^2 





The second and third lower molars of : A, PdlcBOmastodon beadaelU (left side) ; B, P. wintoni 
(left side) ; C, P. parvus (right side) ; D, P. minor (right side). AU | nat. size. 

Fig. 34. After Andrews, 1906, text fig. 50, p. 143, all reduced from 
two-thirds to one-half natural size. Type second and third lower molars, as 
named in the present Memoir. 

A, Type: Palxomastodon beadnelli (left side), type specimen, Cairo 
Museum, M2 and M3. B, Type: Phiomia ^mntoni (left side), type specimen, 
British Museum, Mo and M3. C, Type: Palseomastodon parvus (right side), 
type specimen, British Museum, M2 and M3. D, Type: Phiomia minor 
(right side), type specimen, British Museum, M2 and M3. 



SPECIES OF THE FAYUM 



61 



to their width; the hist lower molar consists of three transverse 
crests and a distinct talon. The type-specimen is a mandible with 
the incisors and jjosterior molars in silii." 

P. wiNTONi Revision. — Matsumoto (1922, p. 3) points out 
that the immature type of Phiomia serridens [see also P. minor], 
referred specimens of Palxomastodon beadndli Andrew.s, 1901, and 
cotypes of Paleomaslodon barroisi Pontior, 1907, maj' all belong 
to the same animal, namely, Phiomia winUmi Andrews, 1905. He 
also adds hindy-fierrn referred specimens from the American Mu- 
seum collection of 1907. Matsumoto (1924.1, i)p. 24-40) gives a 
clear and very able review of the characters of this dominant 
species, referring to it fifty-five specimens in all. 

Phiomia barroisi Tonlier, 19U7 
Figure 35 
Fhivio-marine formation of tho FayOni, Egypt = Uiippr Eocene of 
Anilrcvvs = Lower Oligocene of the |)resent Memoir. 
Synonym of I'hiomia minor and of P. wiiiUmi. 

Paleomaslodon Barroisi Pontier, 1907. "Sur une esp6ce 
nouvelle de Paleomastodon {Paleomaslodon Barroisi)." Ann. Soc. 
gdol. du Nord, XXXVl, pp. 150-154. Type.— Last left 

superior and inferior molars. Horizon and Locality. — 

Bartonien of the Fayi!im, Egypt. Type Figures. — Pontier, 

1907, pp. 150, 151, text figs. 1, 2. 





Fig 1 

Paleomastodon Uarro>si Ponlier, Paleotnaslndon Barroisi Ponlier. 

Arriire molaire supcriciire gauche Arriere molaire inliHeure gauche. 

Echelle ■ 3/5. Eclielle : 3/5. 

Fig. 35. Type figures of Paleomastodon lidrrniai Pontier, 1907. 

After Pontier, 1907, pp. 150 and 151, text figs. 1 and 2. Last left 

superior and inferior molars. Three-fifths natural size. 

Matsumoto (1922, p. 3) considers that Paleomastodon barroisi 
is a synonym in part of Phiomia minor Andrews, 1904, and in part 
of Phiomia wintoni Andrews, 1905. He observes (1924.1, p. 16, foot- 
note) : "The last lower molar of Pontier's two cotypes falls within 
the limit of variation uf P. minor; it may belong to the sujiposed 
male types of the same species." {O-p. cit., p. 24, footnote) "The 
last upper molar of Pontier's two cotypes appears to me to belong 
to the presumed female type of the species [P. irinloni]." 

Osborn, 1924: Osborn confirms Matsumoto's reference and 
considers that the two cotypes of Palseomaslodon barroisi Pontier 
may come within the limits of male and female specimens of 
Phiomia minor and of P. wintoni. 

Moeritherium andrewsi Schlosser, 1911 
Figures 2, 3(). 37, 41 , 42, 42a, 40, 49, 51 ; see also page 74 
Fluvio-marine formation of the Fajilni, Egyi)t = Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

Mwrilherimn A ndrewsi Schlosser, 191 1. "Beitn'ige zur Kennt- 
nis der Oligoziinen Landsilugetiere aus dem Fayiim: Agypten." 




Fig. 36. Type figure of Mxrilherium Andrewsi Schlosser, 1911. -Vfler 
.\ndrews, 190t), Pis. vni and ix; specimens provisionally referred to Mari- 
Iherium lyon.si by Andrews. Geological Museum, Cairo (C. 7867); cast Amer. 
Mus. 9983. One-finirth natural size. 

Beitr. Pal. Geol. Osterr.-Ung., XXIV, p. 130. Type.— A 

young, nearly complete skull, Geological Mu.seum, Cairo (C.7867), 
cast Amer. Mus. 9983; paratypes. also based on specimens 
referred l)y Andrews to Maritheriiim li/onsi, namely, right and 
left series of lower premolars and molars in almost perfect preserva- 
tion (Brit. Mus. M.8501), right lower premolars of another individ- 
ual (Cairo Mus. C.8127), and right upper premolars of a large 
individual (Brit. Mus. M.8500). Horizon and Locality.— 

(Andrews, 190G, PI. ix): "Fluvio-marine beds (Upper Eocene 
[Lower 01igocene|)," Fayiim, Egypt. Type Ficure.— 

Andrews, "Descriptive Catalogue of the Tertiary Vcrtebrata of 
the Fayi'nn, Egypt," 1906, Pis. viii, ix (skull and teeth from the 
Fluvio-marine formation provisionally referred to M. lyonsi ))y 
Andrews). 



62 



OSBORN: THE PROBOSCIDEA 




rrL.3 




£?f--V';^vi^V^^"^''-' V-T^ .J ->J-- ■ , :t , "'■■ "\ 




Fig. 37. Paratype and referred specimens of Mceritherium 
Andrewsi Schlosser, 1911. All from "Upper Eocene" ( = Lower 
Oligocene) Fluvio-marine beds. After Andrews, 1906, PI. ix; 
specimens provisionally referred to Maritherium lyonsi by 
Andrews. 

Fig. 2. Paratype, Brit. Mu.s., M.8501. Figs. 3, 3.\. Para- 
type, Geol. Mus., Cairo (C.8127). Fig. 4. Paratype, Brit. Mus., 
M,8500. Natural size. (See also our Fig. 36.) 

Observe the primitive tetrabunodont form of the Maritlteri- 
um andrewsi molars 



Matsumoto (1922, p. 5) confirms Schlosser's definition of 
Mceritherium andrewsi based on the larger form from the Fkivio- 
marine (skull, Geol. Mus., Cairo, C.7867) which Andrews had 
referred to Moeritheriimi lyonsi, type figure Andrews, 1906 Pis. 
VIII and IX. In 1923 (p. 125) he describes M. andrewsi as follows: 

Specific Characters.—" (3.) Larger form of the Fluvio- 
marine Formation. Lower premolars not very short in comparison 
with the length of lower molars; P2-4, 70 (specimen in the American 
Museum)— 73 mm. (Andrews); lower molars rather long: M1.3, 99 
(Andrews)— 100 mm. (specimen in the American Museum). All the 
lower cheek-teeth are narrow. Po, fusiform in upper view, its widest 



part corresponding to the middle part. P^'"*, 70-75 mm. (Andrews) ; 
M'"^, yet unknown, but M'"", ca. 61 mm. (British Museum, cast in 
the American Museum). Skull large and heavily built. Sagittal 
crest rather weak. Zygomatic width rather small in comparison 
with the length of skull. Distance between the two external audi- 
tory openings, as well as the width of occiput, small, being smaller 

even than in the next form [M. trigodon] M. andrewsi." 

Referred SPECiMEisfs (Matsumoto, 1923, p. 129). — "[Amer. 
Mus.] No. 13432; right half of a full-grown skull, without teeth, 
well preserved in the limit of the parts represented. [Amer. Mus.] 
No. 13434; right upper tusk, viz. P, with well-worn crown. Extra 



SPECIES OF THE FAYUM 



63 



no.; left first upper incisor, with imperfectly preserved crown. 
[Amer. Mus.] No. 1.3437, greater part of a mandible, with the right 
series of cheek-teeth, except P< which might have heen acciden- 
tally lost in the life of the animal, and with left M2-.1 in situ. All 
from the Fluvio-marine Formation of the Fayum." 

Matsumoto (op. ciY., pp. 129, 132) observes that while Schlo.s.ser's 
type and the specific name Mwrilherium atidrewsi are valid, 
Schlosser erred in referring to this species other specimens which 
belonged to Mceritherium trigodon. He gives detailed descrip- 
tion and mea-sjrements of referred skull (Amer. Mus. 13432) in 
comparison with M. trigodon, as well as detailed comparative 
measurements with the Andrews-Schlosser type; also char- 
acters and mea.surements of the first upper incisors, I', of the second 
upper incisors, I-', which constitute the tusks, also of the third upper 
incisors, P, which are rarely found. A referred P has a triangular 
crown, it measures 14 mm. in length, S mm. in width, and 13 mm. 
in height of crown. All three upper inci.sors of Meerilherium 
a-ndrewsi appear to have had the crown and root well differentiated. 
In the lower tusks, I2, one measures 28 mm. in length of crown, 
as compared with 38 mm. in length of crown in I-. 

Osborn, 1924: Osborn confirms Matsumoto's reference of this 
large Fluvio-marine species, and with the cooperation of Doctor 
Matsumoto restores the skull (Fig. 42) of Maerilhcrium andrewsi- 
Irigodon partly with the aid of M. lyonsi materials. 



PalsBomastodon minus Andrews, 1905 = Phiomia minor 

Palxomaslodon minus Andrews, 1905. "Note on the Species 
of Palxomaslodon" Geol. Mag., 1905, Dec. V, N. S., Vol. II, p. 
562. 

(Andrews, letter, 1922): "This was only a slip, later in the 
same paper P. minor was used." 



PalsBomastodon intermedins Matsumoto, 1922 
Figures 38, 39, 49, 88, 91-94, 300; see also page 146 

Fluvio-marine formation of the Faydm, Egypt =Upper Eocene of 
.\nilrew.s = Lower Oligocene of the present Memoir. 

Palseomastodon intermedius Matsumoto, 1922. "Revision of 
Palxomaslodon and Maerilherium. Palxomaslodon intermedius, 
and Phiomia osborni. New Species," Amer. Mus. Novitates No. 51, 
November 21, 1922, pp 1-6. Type.— (Op. cit., p. 2): ... 

' ' fragment of left mandibular ramus bearing all three molars in situ, 
with parts of alveoli of penultimate and last premolars." Amer. 
Mus. 14547. Paratypes: Amer. Mus. 13480, "a fragment of left 
mandibular ramus bearing last molar and posterior root of penulti- 
mate molar in situ." Amer. Mus. 13449, palate. Amer. Mus. 
14548, maxilla, P'M-. Horizo.v A^fD Locality. — Fluvio- 

marine formation (Lower Oligocens), north of Lake Qurun, Faytim, 
Egypt. Paratypes: Amer. Mus. 13480, south we.st of Quarry B ; 
Amer. Mus. 13449, Quarry A; Amer. Mus. 14548, north of Lake 
Qurun. Type Figure. — Op. cil., p. 2, fig. 1. 

Type Description. — Matsumoto's original description is as 
follows (op. cil., 1922, p. 2): 

' ' Palxomaslodon intermedius, new species. Type : American 
Museum No. 14547; fragment of left mandibular ramus, bearing 



all three molars in situ, with parts of alveo'i of penultimate and 
last premolars." 

"Paratype: American Museum No. 13480; a fragment of left 
mandibular ramu.s bearing last molar and posterior root of penulti- 
mate molar in situ." 

"American Museum referred [paratype] specimens: Nos. 
13449, 14.548." 

"All from Fluvio-marine formation." 




B 



tkml'luiNol'JdSO 




Fig. 38. Palseomastodon intennediwi Matsumoto, 1922. Type and 
paratype specimens. After Matsumoto, 1922, p. 2, fig. 1, p. 3, fig. 2. .\, 
Type specimen, fragment of left mandibular ramus (.\mer. Mus. 14547). 
One-third natural size. External view. B, Paratype specimen, fragment 
of left mandibular ramus (Amer. Mus. 13480). One-third natural size. 
Superior view. 



"The palate of paratype specimen of Palxomaslodon inter- 
medins. No. 13499 [13449], measures as follows: " 

Length from the frontal plane tangential to the 
anterior limits of the crowns of the two P- 
to the tip of the posteriorly directed process 
at the posterior limit of the median suture 
between the two palatines 250 mm. 

Distance between the two P^ 53 mm. 

Distance between the two M' 77 mm. 

Distance between the two M' 75 mm." 

"All the upper molars are distinctly bilophodont, as a generic 
character, the rudiment of the third ridge being much feebler 
and much less conspicuous than that of the lower molars. The 
mode of wearing corresponds well to what is stated of the lower 
molars. Besides, all the generic characters of all the cheek teeth 
of this species are the same as those stated in the diagnosis of the 
genus." 

Characters of Pal.eomastodon i.ntermedius (Matsumoto, 
1924.1, p. 10). — In this Bulletin Palxomaslodon intermedius is 
described and measured in great detail from four specimens. (1) 



64 



OSBORN: THE PROBOSCIDEA 



laeomastodon intermedius 
Amer. Mus. 13449 Paratype 




first and second ridges; the posterior valley is distinctly narrower 
antero-posteriorly and shallower than the anterior valley ; so that 
the third ridge looks simply like a talon. As a generic character, 
the very bottoms as well as the walls of the valleys were worn, 
even in the very earlier stages of wearing." (3) "All the upper 
molars are distinctly bilophodont, as a generic character, the rudi- 
ment of the third ridge being much feebler and much less conspicu- 
ous than that of the lower molars. The mode of wearing 
corresponds well to what is stated of the lower molars. Besides, all 
the generic characters of all the cheek teeth of this species are the 
same as those stated in the diagnosis of the genus." 

Phiomia osborni Matsumoto, 1922 

Figures 2, 17, 40, 41, 41a, 90, 171, 18.5, 185a, 187, 222, 258, 283, 286, 300, 316, 
430 ; S(>e also page 244 
Fluvio-niarine formation, eight miles west of Quarry A, Alexandria 
Trail, Fayflm, Egypt = Upper Eocene of Andrews = Lower Oligocene of the 
present Memoir. 

Phiomia osborni Matsumoto, 1922. "Revision of Palxo- 
mastodon and Mceritherium. Palseomastodon intermedins, and 
Phiomia osborni, New Species," Amer. Mus. Novitates No. 51, 
November 21, 1922, pp. 1-6. Type.— (Op. oil., p. 3) : . . . "a 

nearly complete mandible, bearing all the teeth in situ." Amer. 
Mus. 13468. Type Locality. — Eight miles west of Quarry A, 
Alexandria Trail, Fayum, Egypt. Type Figure. — Op. cit., p. 

4, fig. 3. 




Fig. 39. Type jaw and paratype skull of Palseomastodon intermedius 
Matsumoto assembled and refigured in lateral view. One-fourth natural size. 

B, Bl, Type jaw of P. intermedius (.\mer. Mus. 14547). The cavity may 
represent a hollow space within the jaw or a continuation of the alveoli of the 
lower incisors. Bl, Crown view of same type jaw. 

A, Al, Paratype skull fragment of P. intermedius (Amer. Mus. 13449), 
lateral view. Al, palatal view of same specimen. 

Compare figure 93 of Chapter VI, enlarged diagram of hexabunodont third 
superior and inferior molars. 

The type (Amer. Mus. 14547); the paratypes (Amer. Mus. 13480), 
a fragment of left mandibular ramus from near Quarry B, also 
"Quarry B [A]. . . [Amer. Mus.] No. 13449; a large fragment of 
skull consisting chiefly of the palate, bearing anterior premolars 
(P''') and all molars of both sides in situ, and with alveoli of penulti- 
mate and last premolars of both sides; Amer. Mus. Exp. 1907. 
[Amer. Mus.] No. 14548; a fragment of skull and palate, bearing 
penultimate premolar to penultimate molar of left side in situ; 
purchase, 1909. All the specimens, Fluvio-marine formation of the 
Faytlm, Egypt." (2) "In all the lower molars [M1-3] the third 
ridge is rather poorly developed, being distinctly narrower than the 



Fig. 40. Original type figure of P/iionwo osborni' Matsumoto, 1922. After 
Matsumoto, 1922, fig. 3. Type specimen, mandible (Amer. Mus. 13468). 
One-seventh natural size. Lateral view, right side. 

Type Description. — Matsumoto's original description (1922, 
p. 5) is as follows: "This species, Phiomia osborni, appears to 
be more progressive than Phiomia minor and Phiomia wintoni in 
the better developed posterior ridge of the first and second lower 
molars and in the better developed posterior talon of the last lower 
molar; and to be more archetypal than the same in the more 
gradual increase in size posteriorly of the lower cheek teeth." 

Characters of Phiomia csborni (Matsumoto, 1924.1, pp. 
40-49).— The type (Amer. Mus. 13468) is regarded as a male and 
detailed measurements of mandible and teeth are given in compari- 
son with a supposed female of Phiomia u'intoni (Amer. Mus. 13476). 
(1) Judging from the large last molars, M3, as well as from the 
large symphyseal region and tusks, this mandible is regarded as 
a male with the following principal measurements: 

Length from tip of symphysis to posterior 

side of angle 640 mm. 

Length of symphysis 275 

Length of premolar-molar series 250-255 

Length of true molar series 177-180 



SPECIES OF THE FAYUM 



65 



(2) The cheek teeth occupy relatively more space than in the 
supposed female of P. unntoni (Amer. Mus. 13476); mandible 
shorter behind symphysis than in P. wintoni female (Amer. Mus. 
13476), see also Andrews, 1906, text figure 54 of P. wintoni. 

(3) Thus this supposed male type of P. osborni appears to have a 
shorter jaw behind the mandibular symphysis than the supposed 
female of P. wintoni; this indicates a post-symphyseal reduction 
of the jaw in P. osborni. (4) The left and right lower tusks protrude 
58 mm. and 83 mm. respectively from the jaw; each has a distinct 
notch at the tips which may be the result of wear in digging 
and uprooting plants. (5) Symphysis extends back as far as the 
middle part of P3. (6) Lower premolars comparatively large and 
elongate. (7) The last lower molar, M3, extremely elongate and 
comparatively narrow although not beyond the limit of variation 
of Ms in P. wintoni. (8) Posterior cheek teeth, Mi-,,, increasing in 
size more gradually than in P. wintoni; posterior lobe of P4 
distinctly wider than anterior lobe; differences of proportion are 
observed in the premolar teeth in Ixjth P. minor and P. wintoni. 
(9) Posterior talon of M3 constitutes an imperfect fourth lobe con- 
sisting of two prominent cusps, besides a few smaller cuspules, an 
important progressive character. (10) Basal cingula of all molars 
very strong. (11) Intermediate cusps (trefoils) well developed. 



Summary (Matsumoto). — This species appears to be more 
progressive than P. minor and P. wintoni in the more strongly 
developed posterior ridge of the first and second lower molars. Mi-., 
and in the better developed posterior talon of the last lower molar, 
M3; and to be more primitive than the same in the more gradual 
increase in size posteriorly of the lower cheek teeth, M1-3. 
Nothing is yet known about the skull and upper teeth; it 
is of course possible that some of the fragments of the skulls 
and some of the upper teeth referred to /■". wintoni may really 
belong to P. osborni, although it is impossible at present to 
be certain of this. 

Osborn, 1924: Osborn confirms Matsumoto's separation of 
this animal and adds drawings and diagrams showing its distinct- 
ness from PaUeornastudon beadnelU and also illustrating the progres- 
sively distinct trilophodont structure of Mi and Mj and the lelra- 
lophodont structure of Ms. 

Mceritherium ancestrale Petronievics, 1923 
Figure 47; see also page 76 

Qasr-cl-Sagha(?) formation of the Fayflm, Egypt = Middle Eocene of 
.Andrews = Upper Eocene of the present Memoir. 

A full account and figure of this recently described species of 
Mceritherium are given in Chapter III of the present Memoir. 



SYNOPSIS OF SPECIES IN ORDER OF PHYLOGENY AND THEIR PRESENT REFERENCE 



Lower Oligocene Fluvio-marine Beds 

Reference in Present Memoir 



MASTODONTIDiE 



BUNOMASTODONTID^ 



MOLRITHERIID^ 



[Palseomastodon beadnelU 
<j Palxoniastudun intermedius 
[Palxomastodon parvus 

{Phiomia barroisi 
I Phiomia osborni 

Phiomia serridens 

Phiomia wintoni 

Phiomia minor 

{Mceritherium trigodon 
\Maeritherium andrewsi 



Original Reference 
Palseomastodon beadnelU 
Palseomastodon intermedius 
Pabeomastodon parvus 

Paleomastodon barroisi 
Phiomia osborni 
Ph iom ia serridens 
Palseomastodon wintoni 
Palseomastodon minor 

Maeritherium trigodon 
Moeritherium andrewsi 



Upper Eocene Qasr-el-Sagha Beds 

{Mceritherium lyonsi 
MCERITHERIIDiE I Moeritherium gracile 

( M oeritherium ancestrale 



Mceritheri 
Meeritheri 



um lyonsi 
um gracile 



Moeritheriuyn ancestrale 



According to their present reference these FayUm species, treated in chronological order of description above, 
will be treated in subsequent chapters in the order of their relationships and phylogenetic succession, as follows: 

Phiomia: Fam. BUNOMASTODONTID.E, Sulifain. Amebelodontin.e (Chapter VIII). 

Palseomastodon: Fam. MASTODONTID/E, Subfain. nov. Pal/Eomastodontin^ (see Appendi.x). 
Moeritherium: Fam. MffiRITHERIID.E, Subfam. Mceritheriin.e (Chapter III). 

To the above members of the Bunomastodontidae should be added tlie Upper(?) Oligocene species Phiomia 
pygmseus originally described as Mastodon angustidens Cuv. mat. asc. pygmseus Deperet, 1897, and as occur- 
ring in the more recent formation of Kabylie, Algeria. 




1''te=_wv.&_^ ><)A» 




Restorations (1932) of the Primitive Fayum Proboscideans 
Uniform reduction to one thirty-sixth natural size 
Fig 41. When compared with the imaginary scene (Fig. 2, p. 18) on the 'Ur-Nile' or 'Ancestral- 
River-Nile' of the Faytlm region of North Africa, these reconstructions are designed to dispel the wholly 
erroneous notion that Mcerilherium, Phiomia, and Palseotnaslodun are in any way ancestral to each other. 
On the contrary, their cranial, dental, and skeletal characters show them to be profoundly different and 
divergent both in habit and habitat. The relative rarity of remains of Pala;omaslodon is consistent with 
its interpretation as a forest-loving form, in contrast to the more abundantly represented shore-loving 
Phiomia and the still more frequent water-loving Mierilheriuin. 

(Upper) Pakeomastodon beadnelli Andrews, after material described by Andrews, Matsumoto, and 
Osborn. Cranium, ears, proboscis, and tusks largely conjectural. 

(Left lower) Phiomia osborni Matsumoto, after cranial and skeletal parts described by Andrews, 
Matsumoto, and Osborn. 

(Right lower) Moeritheriuiii andrewsi Schlosser. After materials described by Schlosser, Andrews, 
and Matsumoto, and cranial reconstruction by Osborn (Fig. 42, p. 68). 

Of the three restorations, Palosomastodon is the most hypothetical ; as shown in the figures of Chapter 
VI, "The Subfamily Mastodontinae,', besides the skeleton, only the jaw and palate are known, while the 
upper and lower incisive tusks, cranium and mouth parts, and cranio-aural parts are still unknown. 
Phiomia, on the other hand (as shown in Chap. VIII, pp. 236-247), is relatively well known in its cranial 
structure. Mueritherium is also relatively well known (Chap. Ill, pp. 67-79) both in cranial and limb 
structure. The mouth parts and feeding habits of Phiomia are deductible from our recent knowledge of its 
shovel-tusker descendants (Amebelodon) , as conceived and restored by Osborn and Borissiak. 




PALAEOMASTODON BEADNELLI e328min-,7'7%' e 
M.AFRICA FAYUM 



PHIOMIA OSBORNI I345mni..4'5"e 
N AFRICA FAYUM 




MQERITHERIUM LYONSI &9i?m[I1..2'3/f"e 
N AFRICA FAYUM 



Fig. 41a. Restorations oNE-nnNDREDTii natural size by Maroret Flinsch, under the direction of Henry Fairfiei-d Osborn 

66 



Chapter 111 

THE MCERITHERES (MCERITHERIOIDEA) OF NORTH AFRICA 

IN EOCENE-OLIGOCENE TIME 



In the present chapter hi is given a detailed systematic revision of the specific types of MCERI- 
THERES. 

See CHAPTER II for (a) separation of the MCERITHERES FROM OTHER PROBOSCIDEANS OF NORTH AFRICA, 
(b) SYSTEMATIC CLASSIFICATION OF THE MCERITHERES, (c) COMPARISON OF MCERITHERIUM WITH PHIOMIA, (d) 
RELATIONS OF MCERITHERIUM TO PAL^OMASTODON AND PHIOMIA, (e) COMPLETE SEPARATION OF THESE THREE 
GENERA, (f) PHYLOGENETIC RELATIONS TO OTHER PROBOSCIDEANS, (g) PROBABLE AMPHIBIOUS AND FLUVIATILE 
HABITAT, (h) DETAILED COMPARISON OF THE GRINDING TEETH WITH THOSE OF PAL^OMASTODON AND PHIOMIA, 
AND (l) GEOLOGIC DISTRIBUTION OF MCERITHERIUM AS COMPARED WITH PHIOMIA AND PAL^OMASTODON. 



Subordinal, family, subfamily, and generic characters. 
Reconstructions of Mceritherium by Osborn. 

Revised descriptions and definitions of the species 
of Maeritherium in phj'logenetic order. 

Type and characters of Maeritherium lyonsi An- 
drews, 1901. 

Type and characters of Maeritherium gracile An- 
drews, 1902. 



Type and characters of Mcerilherium andrewsi 
Schlosser, 1911. 

Type and characters of Maeritherium trigodon 
Andrews, 1904. 

Type and characters of Maeritherium ancestrale 
Petronievics, 1923. 

Maeritherium molars compared in detail with those 
of Palxomastodon and Phiomia. 

Supposed Maeritherium of Baluchistan. 



1. SUBORDINAL, FAMILY, SUBFAMILY, AND GENERIC CHARACTERS 

Mceriiherium is very fully treated in the preceding Chapter II, in comparison with the contemporary probo- 
scideans, by a complete review of the opinions and theories of Charles W. Andrews, who first recognized and 
described the genus (1901-1906) and the majority of the species embraced within it. Also see the opinions and 
theories of Hikoshichiro Matsumoto and those of the present writer on the liabits, habitat, and relationships of 
this genus, which, while standing apart from and not directly ancestral to any of the other Probo.scidea, displays 
many primitive ancestral characters of the order Proboscidea that are lost in other phyla. 

In the present Cliapter III it remains to complete tlie systematic description of the species of this genus, 
to define clearly the suborder Moeritherioidea and the family characters, to establish the types and type geologic 
levels of the species, to quote Matsumoto and Petronievics in their detailed studies of the cranial structure of 
Mceritherium, and to make a very detailed comparison of the grinding teeth of Maeritlierium and of Palaeomastodon. 



87 




Osborn's Reconstruction (1920) of the Cranium and Jaws op Mieritherium andrewsi-trioodon to be compared with Andrews' 

Reconstruction (1906) of M. lyonsi (Fig. 18). 
Fig. 42. Second rccon.struction of the skull of Mceritherium (M. andrewsi, M. Irigodon), 1920, which is to he compared with the first 
reconstruction of M. lyonsi by Andrews (Fig. 18). 

This reconstruction by Henry Fairfield Osborn with the cooperation of Dr. Hikoshichiro Mat,sumoto assigns chiefly a gliriform 
feeding function to the superior and inferior incisor teeth. There is no evidence that the upper incisor teeth were chiefly weapons or 
tusks, as indicated in the British Museum reconstruction and the model of the skull by Andrews (see Fig. 18). 

A, Al, Top and .side views of skull chiefly based on Amer. Mus. 13432 (M. andrewsi), 13430 (A/. Irigodon). 

B, Bl, Top and side views of jaw chiefly based on Amer. Mus. 13437 (M. andrewsi), 13436, 13435 (M. irigodon). 

The American Museum specimens were used as follows: For size and general characters No. 13432; for the first incisor No. 13442; 
for the .second incisor No. 13434; the faint lines indicate position of the root of I--I2; for the structure of the grinding teeth Nos. 15898 
and 13430; for the structure of the pterygoids and occiput No. 13430. 



4it^ 



4j-i^ 



MOERITHERIUM ANCE5TRALE MOERITHERlUt^ GRACILE 

5oomm.,i'73/4"e 5oomm.,i'73/4"e 



>fe?^ 



4J^ 



MOERITHERIUM LYONSI MOERITHERIUM TRIOODON MOERITHERIUM ANDREWSI 

&99mm., 2'3i/2"e 625iDni.,2'i/a"e 65omm..a'i'/2"e 



Fig. 42a. Ancient Moeritheres of northern Africa. One-hundredth natural size. Restorations by Margret Flinsch, under the direction of 
Henry Fairfield Osborn. 



68 



THE MCERITHI'^RES 69 

Suborder or Superfamily: I. iMCEUITHERIOIDEA Osborn, 1921 

Original reference: Amer. Mus. Novitates, No. 1, 1921, p. 2 (Osborn, 1921.515). 

SuBORDiNAL CHARACTERS. — (1) Palustral, amphibious, and semi-aquatic quadrupeds, with ancestral 
genetic affinities to the Proboscidea in the dentition, and analogies to the Sirenia and Hyracoidea in the 
skull. (2) Vertebrae estimated: Cervicals 7, dorsals 19 2U, liunbars 4, .sacrals 4; vertebral structure indi- 
cating an ambulatory and amphibious habit (fide descriptions of Andrews, 1906, pp. 112-117). (3) 
Girdles, scapulse, and pelvis very j^rimitive with analogies to the sirenian type, not expanded as in the 
proboscidean graviportal type. (4) Proportions of humerus and forearm unknown; humerus dLstally of 
primitive locomotor type. Femur with ligamentum teres pit, greater "and lesser trochanter, absence of 
inner trochanter; straight, not unlike Proboscidea. (5) Probably pentadactyl. (6) Cranium primitive, 
greatly elongated, tubular brain cavity, small brain case of archaic type, zj^gomatic arches slender. (7) 
Face abbreviated, orbits small, nasals abbreviated. (S) Somewhat reduced eutherian dentition: Ii, Ch, 
Fi, Mf . Second incisors, P-I2, enlarged, curved, gliriform, partly surrounded by enamel. (?) Motion 
of jaw vertical-orthal. 

The subordinal characters are derived from the detailed observations, descriptions, and definitions of Andrews 
(1906, pp. 99-120), as confirmed by Osborn (1907-1923), and as extended by Matsumoto (1922-1923). The most 
recent review is that of Petronievics (1923) cited below under the description of Maeritherium ancestrale. 

The numerous archaic and primitive characters in the trunk and limbs of these animals and the 
feeble muscular areas of the scapulse and pelvic girdles appear to indicate that they had never passed through a 
long period of terrestrial locomotion either mediportal or graviportal, much less cursorial. Until the complete 
adaptation of the fore and hind limbs and the feet is known we cannot determine how far they were terrestrial or how 
far amphibious. The girdles and limbs are certainly more aquatic in adaptation than those of the hippopotami. 
On the other hand, the Mocrithere limbs and girdles are not to be compared for a moment in aquatic adaptation 
with those of the contemporary sirenians. (See Fig. 43.) 

Conclusion as to Habits. — In the absence of a more complete knowledge of the limbs, we may conclude 
that the Mceritheres were chiefly shore-hving or marsh-living animals, amphibious in shallow waters, adapted in 
dentition to a succulent vegetation which could be seized by the opposing upper and lower tusks and by the broad 
lips, which are sirenoid and hippopotamoid rather than elephantoid. 

Family: M(ERITHERIID^ Andrews, 1906 
Original Reference: "A Descriptive Catalogue of the Tertiary Vertebrata of the Fayflm, Egypt" (Andrews, 1906.1, p. 99). 

Definition. — The family Mopritheriid* Andrews may be redefined as follows: (1) Nares small 
terminal; (2) orbits opening upwards and outwards; (3) auditory meatus elevated, an acjuatic adap- 
tation; (4) superior premolars, P'', tritubercular ((luatlritubercular in Phiomin and Palxomastodon), pre- 
molar tetartocones rudimentary or absent; (5) molars tetrabunodont (quadritubercular), bilophodont, 
with incipient triloi)hodonty, quadrate or slightly elongat(> in jiroportion: (6) superior molars, M' 
strictly bilophodont, M- with rudimentary third crest, M' with enlargeil third crest; (7) inferior molars, 
M, trilophodont, rudimentary, M,, trilophodont, trilophodonty not pronounced; (8) all three pairs of 
premolars and of molars of l)oth jaws functioning at same time; (9) no trace of trefoil pattern or of in- 
termediate conules in any of the Mucritherium molars. 

Of the eight characters assigned by Andrews (1906, p. 99) to his family Moeritheriida^, /re are included above 
as subordinal characters. We do not agree with Andrews when he remarks: . . . "probably a small proboscis 
was present," because the structure of the muzzl(> proves that the animal had a square upper lip. 

The family McBritheriidie stands well apart in its dental adaptations from its contemporaries, either Palxo- 
mastodon or Phiomia. The food was seized by the gliriform incisors and there was an up-and-down or chopping 
motion (vertical-orthal) of the bilophodont, subtrilophodont grinders, as compared with the lateral grinding 



X 
ft 



> 

o 




•3 -^ i s 






§2 

eg 






5 ■<" '< 
1-^ 00 o> 

f^ 00 oi" 



CO 
O 

J 

I — 1 

w 

I— < 

9 



s 


i 






K 


■-H 


w 




•s-i 


hr 




tC 


Ce. 


^ 


O 


b< 


7. 


• 


o 


Oh 


H 




U 


<-p 




(-) 










!/J 








B3 


ft 




En 


P 


-S 


n 


c 


Q 


< 


7. 


L^ 


0. 




Oi 


« 


< 


c 


a 


o 


z 






3 



:5 ►J 



< '- -E 



o u — 



£ 


o 


13 


n 




(1) 






CO 


c 


S 


^, 


bib 


i 


c 


O 

a 

CO 




-J- 
















n 


< 


> 




-< 


•< 

CM 






-T 


CI 


cc 






THE MGiRITHERES 71 

motion of the molars in Palaeomastodon and in Phiomia. The deUcate zygomatic arches indicate feeble develop- 
ment of the masseter and temporal muscles. 

These family characters probably represent the climax in a long period of evolution. Close comparison of 
the grinding teeth of Mwritlieriutn (Fig. 49A, B, C) with those of Palseomastodon inlermedius (Fig. 49D, E) 
indicates that they are tetrabunodont and sublophodont, and it also shows that Moeritherium differs from 
Palaeomastodon in the total absence of any trace of the 'central conules' in the molars. 

Subfamily: Mceritheriini Winge, 1906; Mceritheriin^ Osborn, 1923 

Original reference: "Jordfundne og nulevcndc Hovdyr (Ungiilata) fra Lagoa Santa, Minas Goraes, Bra.silien." E. Musco Luiidii, 
1906, III, p. 172 (Winge, 1906.1); also Amcr. Mus. Novitates, No. 99, 1923, p. 1 (Osborn, 192.3.601). 

Subfamily Characters. — (1) First superior and inferior incisors, I'-I,, partly functional. (2) 
Second superior and inferior inci.sors, F-L, greatly enlarged. (3) Third superior incisors, l\ and su- 
perior canines, C, greatly reduced. (4) Third inferior incisors, I,, and inferior canines, C, entirely 
wanting. (5) Inferior molars, M1.3, elongated, with rudimentary third lobe (tritolophid) behind the 
protolophid and metalophid. (6) Incisors sheathed in enamel. 

Wingc's definition of tlic subfamily Mceritheriini, which he included within the family Elephantida;, corre- 
sponds with Osborn's redefinition of the family Moeritheriidtc above. Inasmuch as Mceritlierium in its five 
included species, M. lyonsi, M. gracile, M. irigodon, M. andrewsi, and M. ancestrale, are the only Mceritheres 
known, the definition of this subfamily is merely a formal matter. 

It is not unlikely that other subfamilies of Mceritheriidsc will be discovered, especially if the Ma?ritheres should 
be traced into southern Asia, as suggested by Pilgrim, and should be found in other parts of the African continent. 

Genus: MCERITHERIUM Andrews, 1901 1906 

Original reference: Tageblatt des V Intcrnat. Zool.-Cong., Berlin, No. 6, Aug. 16, p. 4 (Andrews, 1901.2). (Published Volume 
Vcrhaiidlungpn, 1902, p. 528.) 

Genotypic species: Moerithcrium lyonsi. 

Generic Characters. — (Andrews, 1906, p. 99): Dental formula: — i.i; c.v; pjnA; tn.i. The second 
incisors in both jaws greatly enlarged and tuskhke. The last premolar not bilophodont; molars hilo- 
phodont. This genus is represented by several species, of which Mwrithcrium lyonsi is the type. It is 
found in both the Middle and Upper Eocene deposits of the Fayum [ = Upper Eocene and Lower 
Oligocene of the present Memoir]. 

Matsumoto, 1923, p. 121. — After very detailed comparisons, not summed up in a generic definition, 
Matsumoto reaches the following conclusion: "I. — Natural Position of Ma?rit]ierium. It may be evident from 
the preceding statements that Marithcrium differs strongly from either the hyraeoids or the sirenians in many 
characters, on the one hand, and has many proboscidean and pre-palaeomastodont characters on the other hand. 
If much weight should be given to the sirenian resemblances of Moeritherium, then more weight should be given 
to its hyracoid resemblances. In my opinion Moeritherium is to be treated as a very archetypal member of 
the probo.scideans as correctly stated by Andrews at the first." Also: (1) Skull short in comparison with 
zygomatic width; (2) long nasofronto-parietal region in proportion to length of skull; (3) zygomatic arches 
relatively long; (4) sagittal and lanibdoidal crests well developed ; (5) orbits situated just above P-"^ in Moerithcriion, 
just above M'- in Palxomastodon; (6) mandible relatively short; (7) diastemata relatively short; (8) ascending 
bars of mandible incline slightly forwards; (9) lower border of mandible ascending; (10) symphysis of mandible 
rather short; (11) upper [and lower] tu.sksshort in comparison with tho.se of the other proboscideans; (12) premolars 
large in jiroportion to the molars, the series of cheek teeth increasing gradually in size backwards; (13) ridge 
formula simpler than in Palxomastodon, much simpler than in Phiomia; (14) Mcorithcriuni structurallj'^ a pre- 
palaeomastodont type, that is, Mceritherium stands structurally as more primitive than Palxomastodon in twenty- 
three characters which are peculiar to it and also which indicate its remote relationship to Palxomastodon. 



2. REVISED DESCRIPTIONS AND DEFINITIONS OF THE SPECIES OF MCERITHERIUM IN 

PHYLOGENETIC ORDER 

Compare original descriptions of the same species in Chapter II, wherein the chronological order of description is clearly set forth. 
Consequently the reader is referred to both Chapters II and III and to the more recent paper of Petronievics for our present knowl- 
edge of the species of the genus Moeritherium. 



Moeritherium lyonsi Andrews, 1901 

See Chap. II, p. 54 

Qasr-el-Sagha formation of the Fay<lm, Egypt = Middle Eocene of An- 
drews = Upper Eocene of the present Memoir. 

The establishment of this species by Andrews and its descrip- 
tion, based on forty-six specimens contained in the British and 
Cairo museums from the Qasr-el-Sagha beds and one specimen in 



short: P^"'' [P2-4I, ca. 69 mm. (Andrews); lower molars very long, 
M1.3, ca. 104 mm. (Andrews). All the lower cheek-teeth very 
wide. P2, triangular, its widest part corresponding to the posterior 

lobe. P--^ 67-78 mm. ; M'"', 85 mm. (Andrews) M. lyonsi." 

It is this genotypic species M. lyonsi of the Upper Eocene 
Qasr-el-Sagha formation which furnishes the chief characters of the 
genus Moeritherium as described and figured in great fullness by 



pm,2- 



l^'i>:W^ 



/!">•* nil 




Fig. 44. Second type figure of Mwritherium lyonsi .\ndrews, 1901. .\fter Andrews, 1906, Fl. x, figs. 1, U, 2, 3. Geol. Mus., 
Cairo (C. 10000); cast Amer. Mus. 9977. One-third natural size. 

Three individuals: (1, 1a) Type specimen (C. 10000), mandible associated with upper molars and a dorsal vertebra (op. cii., 
p. 120); (2) "Upper molars and premolars found clo.se to type" (p. 123), of the right side = paratype (C. 10001); (3) front portion 
of skull, side view (C. 10002, cast Amer. Mus. 9978) = paratype. Same as figure 27. 



the American Museum collection, are in part set forth in the 
systematic revision of this Memoir above, pages 54, 55. 

Specific Characters.— (Matsumoto, 1923, p. 124): "(1.) 
Larger form of the Qasr-el-Sagha Formation. Lower premolars very 



Andrews (1906, pp. 99-126, Pis. viii, ix, and x). The detailed 
measurements and indices of the superior and inferior teeth in com- 
parison with those of M. gracile, M. andrewsi, and M. trigodon are 
transcribed and estimated from Matsumoto as follows : 



72 



THE M(ERITHERES 



73 



Lower Teeth 




M. lyonsi 




M 


gracile 




M. andrewsi 


M 


. trigodon 






1. 


w. 


I. 


1. 


w. 


I. 


1. w. 




I. 


1. 


w. 


I. 


P2 


22 


16 


73 








24-26 11 


-14 


54 


19.5-23 


10 -11 


48 


P3 


23 


21 


90 


22 


15 


68 


27-29 17. 


5-19 


65 


22 -25 


14? -17 


68 


P. 


25 


23 


90 


21 


18 


86 


24-25 21 




87 


22 -25 


18 -20 


80 


M, 


26.5 


24.5 


92 


22-23 


19-20 


87 


28-30 21 


-22 


73 


24 -27.5 


19 -21 


76 


Mj 


28 


-35 25 -30 


86 


27-28 


22-23 


82 


32-34 26 


-27 


79 


29 -32 


24 


83 


M, 


39 


-42 28 -30 


71 


33-34 


24 


72 


37-40 26 


-28 


70 


37 -40 


24 -26.5 


66 


P2-4 


69+ 






63+ 






70-73 






63 ±-70 






M,.3 


104 + 






84 






99-100 






93 -98 






I2 width 










20+ 




28 








20 




Upper Teeth 


























p2 


27 


23? 


85 


22 


18 


82 


25-27 20 


-22 


81 


19 -22? 


18? -21 


95 


pj 


26.5 


29,5 


111 


20 


23 


115 


25-26 28 


-32 


123 


23 -24 


24? -26 


108 


P* 


23 


27.5 


119 


20 


21? 


105 


21-22 27 


-29 


132 


20 -22 


23? -25.5 


116 


M' 


29 


27 


93 


23-25 


21-23 


92 


29-32 25 


-27 


84 


22 -25 


24 -26.5 


106 


M2 


26? - 


-30 23.5-28 


93 


24-27 


23-25 


92 


32-33 28 


-29 


88 


28 -30 


24 -28 


93 


M» 


32 - 


-37 ± 28 -30 ± 


81 


28 


24-25 


86 








30 -32 


26 -27 


84 


P^" 


67 - 


-68 




62 






70-75 






60 -63 






Ml-' 


85 






75-79 ± 










83 -85 






P width 




28 










26 








21 





American Museum Collection. — Matsiimoto 
refers (1923, p. 125) to the species Mwrilherium lijonsi 
two of the three fragments of mandibular rami 
inchided in Amer. Mus. 13444; the otiicr frag- 
ment of a ramus he refers {op. cit., p. 128) to M. 
gracile. his descriptions being included above in 
Chapter II, pages 54 and 56 respectively. 

Moeritherium gracile Andrews, 1902 

Scf Cliap. II, p. 06 
Qasrn'l-SaKliii forniiition of the Fayflni, Egy])! =Midilli' 
Eocene of Andrews = Upi)er Eoeene of the present Memoir. 

This small form of the Qasr-ei-Sagha beds 
Upper Eocene, is rciativoly rare, four specimens 
being described by Anch'cws from the British and 
Cairo museum collections while four are described by 
Matsumoto and Osborii from the .•Xinerican Museum collection 

Specific Ciiar.\cters. — (Andrews, 1900, p. 127): "Tliis 
species is distinguished from M. lyonsi by its comparative lightness 
of structure, the narrowness of the palate, the smaller size of the 
upper molars and premohirs, jjarticularly of w. 3, the strong de- 
velopment of the cingulum in these teeth, and by the considerable 
inflation of tiie cranial region of the squamosal, which apparently 
contains extensive air-sinuses." 

(Matsumoto, 1923, p. 125): "(2.) Smaller form of the Qasr- 
el-Sagha Formation. Lower premolars not very short in compari- 
son with the length of lower molai-s: P2.4, ca. 62 mm. (specimens in 
the American Musevmi); lower molars very short: M1.3, 83 mm. 
(specimen in the American Museum). All the lower cheek-teeth 
are narrow. P^, 62 mm. (Andrews); M'"', 75 — ca. 79 mm. 
(Andrews) M. gracile." 




Am.tius No 13443 




Am. Mus.No, 15444 



Fig. 45. MiTrilhcrium gniciU and .1/. lyimsi from the Qasr-cl-Sagha for- 
mation, I'pix^r Eocene of the Fa\-<im. .\nierican Museum collection of 1907. 
.\ffer Matsumoto, 1923, figs. 3 and 4. 

.\mer. Mus. 13443, M. gracile; mandil)Ie with P.i-Mj of the left side 
and M1.3 of the right side 111 silii. 

.\raer. Mus. 13444, M. lyoitxi; one of the three fragments of the man- 
dibular rami, with badly preserved molars in situ. 



74 



OSBORN: THE PROBOSCIDEA 



Andrews {op. cil., 1906, pp. 127, 128) gives a detailed descrip- 
tion with measurements of the four specimens in the British Mu- 
seum; Matsumoto (op. cit., 1923, p. 128) also gives a detailed 
description with measurements of the specimens in the American 
Museum collection. In his comparison of the four species, Mat- 
sumoto's detailed dental measurements are as above. 

Phylogenetic Summary. — From the limited materials at 
hand neither Andrews, Matsumoto, nor Osborn has been able to 
establish the phylogenetic or ancestral relationship of the Qasr-el- 
Sagha species (M. bjonsi and M. gracile) to the Lower Oligoccne 
Fluvio-marine species (M. andrewsi and M. trigodon). 

We might suppose that the larger form M. lyonsi gave rise to 
the larger form M. andrewsi, but Matsumoto does not believe so. 
(1) Matsumoto observes that the larger Upper Eocene species M. 
lyonsi seems to be less closely related to its smaller contemporary 
M. gracile than the larger Lower Oligocene form M. andrewsi is to 
the smaller Lower Oligocene form M. trigodon; this inference is 
drawn from the absolute and relative dimensions of the grinding 
teeth. (2) M. andrewsi and M. trigodon appear to represent two 
entirely distinct species and not male and female representatives 
of the same species, as Schlosser has suggested. (3) So far as can be 
judged from the relative dimensions of the lower cheek teeth, M. 
gracile appears tq be more closely allied to the two geologically 
succeeding species M. andrewsi and M. trigodon than to its con- 
temporary M. lyonsi; in other words, it appears that M. gracile 
might be the ancestral type of both M. andrewsi and M. trigodon. 

Moeritherium andrewsi Schlosser, 1911 
Sco Chap. II, p. 61 
Fluvio-marine formation of the Fay dm, Egypt = Ui)pcr Eocene of .\n- 
drews = Lower OUgocene of the present Memoir. 

This palustral animal, based on specimens referred to M. 
lyonsi by Andrews, occurring in the Fluvio-marine beds with a 
rich terrestrial fauna, is naturally much less abundant than its 
predecessors in the Qasr-el-Sagha marine beds; it is represented by 
twelve specimens in the British Museum (all referred by Andrews 
to M. lyonsi) and by three specimens of M. andrewsi in the Ameri- 
can Museum, two of which (Amer. Mus. 13432 and 13437) have 
been largely used in the second reconstruction of the skull of 
Maritherium shown in figure 42. As set forth in the systematic 
revision above (p. 61), Schlo.sser and Matsumoto have clearly 
separated and defined this larger form as a distinct species, M. 
andrewsi, named in honor of Charles W. Andrews. 

Specific Characters. — Matsumoto's characterization (Mat- 
sumoto, 1923, p. 125) of this species is cited in full on pages 62, 63 
of this Memoir. These characters agree with the reconstruction of 
the skull by Osborn and Matsumoto shown in figure 42, in which it 
appears that there are considerable differences in size and propor- 
tions between the larger M. andrewsi (Fig. 36) and the smaller 
M. trigodon (Fig. 30). 

Moeritherium trigodon Andrews, 1904 

See Chap. II, p. 57 
Fhivio-marine formation of the FayAm, Egypt = Upper Eocene of ,\n- 
drews = Lower Oligocene of the present Memoir. 

American Museum Collection. — Mwrilherium trigodon is 
represented by the single type specimen (Brit. Mus. M.8499) 



described by Andrews in 1904, and by six specimens in the American 
Museum collection, described, figured, and measured by Matsumoto, 
namely, Amer. Mus. 13430, 13431, 13433, 13435, 134.36, 13439. 

The cranium is inferior in all its dimensions to the larger and 
more massive M. andreivsi and there appear to be minor differences 
in the characters of the median and lateral incisors, I"^, P. A well- 
preserved second right upper incisor, P, with its casing of enamel 
(Fig. 42), illustrates the profound difference between the gliriform 
second superior incisor of M. andrewsi and the superior tusk of 
Phiomia (Fig. 183); observe especially the posterior surface 
beveled by the action of the second inferior incisor in contrast to 
the tusk of Phiomia which is worn on the internal side. Observe 
also the profound difference between the lower jaw in superior 
aspect (Fig. 42) and the lower dentition of Phiomia and Palseo- 
mastodon (Fig. 48) as seen in superior view; also the uniformly 
worn grinding series, P2-M3, of Amer. Mus. 13435 (Fig. 46), as 
compared with the successional wear in Phiomia. 

Specific Characters. — (Andrews, 1906, pp. 128, 129 — see our 
Fig. 30): "Type Specimen. — Posterior portion of right ramus of 
mandible (PI. ix, fig. 5); British Museum. This species is dis- 
tinguished by the form of the posterior lower molar, which nar- 
rows posteriorly, the talon consisting almost entirely of one large 
tubercle; also by the rapidity with which the teeth decrease in 
size from behind forwards. . . . This difference in the talons 
appears to justify the separation of the present form as a distinct 
species at least, and not improbably further material will show 
that a new genus must be established. The enamel of the whole 
tooth is raised into irregular ridges and small tuberosities. The 
dimensions (in centimetres) of the teeth are: — 

Length. Width. 

rrt. 1 2.6 ? 

m.2 3.2 ? 

??i. 3 4 2.4 

The length of the molar series is 9.8 cm." 

Matsumoto's definition (1923, p. 125) is cited above (pp. 57, 
58) under the systematic revision. He considers that Andrews lays 
too much weight on the shape and structure of the talonid or 
tritolophid of M3 in distinguishing this species from M. andrewsi, 
remarking that the posterior talon of M3 appears to be one of the 
most variable features in the teeth of Moeritherium, and that its 
surface varies according to age and the degree of wearing action 
of the tooth. 

Cranial Characters. — Matsmnoto observes (1923, p. 134) 
in his description of the skull: "The skull of [Amer. Mus.] No. 
13430 measures 240 mm. in the length from the point at which the 
vertical plane tangential to the anterior borders of both P- meets the 
median longitudinal line on the palate, to the basion, 145 mm. in 
the length from the same point to the median point of the posterior 
border of palate, about 180 mm. in the length from the anterior 
limit of the temporal vacuity in palatal view to the posterior lower 
border of squamosal, 42 mm. in the minimum width of the mid- 
cranial region (this dimension might be somewhat less than it 
ought to be in primary condition, as this part of this specimen 
appears to be crushed secondarily from side to side), 280 mm. in the 
zygomatic width, about 190 mm. in the distance between the upper 
borders of external auditory openings (including reliably restored 



McKRITHERIUM TRIGODON (A-C, E, F) AND M. ANDREWSI (D) FROM TKE FlUVIO-MARINE FORMATION, LoWER OlIGOCENE OF THE FaTUM. SeE FiGURE 23 




Am Mus No 13430 





Am MusNq 13430 





li 



f 



^m Mus No. 13435 



:^^ft& 




Fig. 46. .\merican Muspum • ' 

Collection of 1907. .\fter iMatsii- 
moto, 1923, figs. 1, 2, 6, 7, 8, and 9. 

A, B, Amer. Mus. 13430, M. 
trigodon, a full-grown skull, lateral 

and superior views; partly restored ''i - 

occipital and sagittal crests, nasal AmMusNal343l 

and symphyscal region, and anterior 

teeth. This skull is described and measured in detail by Matsumoto (1923, p. 134) in his 
description of M. Irigndoii. About one-fourth natural size. (1) Observe the excessively 
slender, widely arching zygomata, the tubular cranial region, the slender sagittal and 
lambdoidal crest,s. (2) Zygomatic width 280 mm. 

C, Amer. Mus. 13435, 3/. trigodon, fragment of the right ramus of the mandible, all 
the cheek teeth, P0-M3 in nilii. One-fourth natural size. 

D, Amer. Mus. 13434, Ma-ritherixtm andreuKJ Schlosser, second upper right incisor 
loDlli. 1^; upper, internal view, lower, external view. One-half natural size. Through error 
this appeared in the original legend (Matsumoto, 1923, p. 126) as M. lyonsi. 

E, Amer. Mu.«. 1343t), MariHirrium trigmlnn. "ramus of mandible, with all the cheek 
teeth in xilu." After Matsumoto, 1923, p. 126, fig. 1, through error referred to M. lyonsi 
in the legend. 

F, Amer. Mus. 13431 , M. trigodon, fragment of skull including greater part of right half 
of palate bearing P'-M' in situ. One-fourth natural size. 

75 



76 



OSBORN: THE PROBOSCIDEA 



parts to a slight extent), 220 mm. in the width of occiput, 91 mm. 
in the lateral e.xtension of the two occipital condyles, 52 mm. in the 
width of palate between the two P^, 34 mm. in the same between 
the two M', and about 150 mm. in the ma.ximum height of the 
skull, including the upper cheek-teeth (including a restored part of 
sagittal crest to a slight extent). In this skull the sagittal and 
occipital crests are extraordinarily well developed; and the occiput 
is strongly concave, so that its upper half, including the lambdoid 
crest, inclines distinctly backwards." 

Moeritherium ancestrale Petronievics, 1923 

Qasr-el-Sagha(?) formation of the Faydm, Egypt = Middle Eocene of 
Andrews = Upper Eocene of the present Memoir. 

Moeritherium ancestrale is regarded by Petronievics as nearer 
to Palseomastodon than any previously described species of 
Moeritherium. 

Moeritherium ancestrale Petronievics, 1923. "Remarks upon 
the Skulls of Moeritherium and Palseomastodon," Ann. Mag. Nat. 
Hist., Vol. XII (9), pp. 55-61. Type.— Skull and mandible 

(Brit. Mus. M. 9225). Horizon and Locality. — Fayum 

district of Egypt, Qasr-el-Sagha (?) formation = Middle Eocene of 
Andrews = Upper Eocene of the present Memoir. Type Fig- 

ure. — Petronievics, 1923, p. 56, fig. 1. 

Pis. 1. 




Part of the left .side of the skull of Moeritherium ances/rnle, i nat. size, 
■as., nlisphenoid ; pt., pterygoid j sq., sijuamoeal. 

Type of Mceritherium ance.strale 
Fig. 47. Type of Meeritherium aneeslralc Petronievics, 1923, p. 
56, fig. 1. Reproduced same size as the author's original figure and 
including the author's legend. Brit. Mus. M.9225. 

Type Description. — (Op. cit., 1923, p. 57): "There are two 
main charactere of our specimen which justify the supposition that 
it represents a distinct species: — (1) the elongated palatine vacuity 
between anterior maxillary prolongations forming the inner border 
of the sockets for the tusks; (2) the backward direction of the 
occipital plate. As the first of these two characters brings our 
specimen nearer to Palseomastodon than any other, we will give to 
the new species the name Moeritherium ancestrale, sp. n." 

The following is a direct quotation, with .some omissions, 
from the description of Doctor Petronievics (op. cit., pp. 55-57): 

"[1] There is in the British Museum [M.9225] a skull and 
mandible of Mceritherium not yet described, which is in several 
respects important for deciding the question of the ancestral 



relationship of Moeritherium to the Proboscidia. This skull was 
found by Baron Nopcsa .several years ago (1905) in Middle Eocene 
beds in the Fayum district of Egypt [Footnote: 'Comp. C. W. 
Andrews, 'A Descriptive Catalogue of the Tertiary Vertebrata of 
the Fayum, Egypt': London, 1906, p. 121 seq. (specimen M. 9225 
in the British Museum).']. I propose to describe the features of 
this skull, which either give some supplementary data as to the 
morphology of the skull of the genus Mwritherium or are peculiar 
to the species. I propose also to give some additional evidence on 
the skull of Palseomastodon. Finally, I will add a list of the re- 
semblances and differences in the skull, mandible, and dentition 
of the two genera. [2] In the skull in question the basioccipital is 
separated from the basisphenoid by a suture situated somewhat 
obliquely on a ridge at the level of the glenoid surface. The suture 
between basioccipital and exoccipitals is not plain. The exoccipi- 
tals with the condyles are as in Moeritherium lyon.'ii. [3] The same 
is to be said about the supraoccipital bone; but the whole occipital 
plate in our specimen is directed backwards, while in MoeritherixiTn 
lyonsi it is directed forwards, and the lateral depressions accompany- 
ing the median ridge are deeper in our specimen than in the old 
one. [4] The squamosal and its relation to the parietal and supra- 
occipital are the same as in Moeritherium lyonsi. [5] The jugal 
seems to be more slender than in the latter. [6] The tympanic in 
this specimen is better preserved than in the skull of Moeritherium 
lyonsi. [7] The parietals are as in the latter. It is uncertain if 
they meet the alisphenoids — probably not. [8] The frontals are 
as in Mceritherium lyonsi, with no postorbital process. The exact 
relation of the frontals to alisphenoids and parietals is uncertain, 
but they reach the former. [9] The existence of a separate lachry- 
mal cannot be established. JIO] While in Moeritherium lyonsi the 
position of the alisphenoid on the side-wall of the skull could not 
be determined, in our specimen the limit of the alisphenoid in this 
respect is quite clear (text-fig. 1). Above the alisphenoid canal 
the alisphenoid forms a triangular bone. The backward side of 
this triangle is constituted by the suture separating the alisphenoid 
from the squamosal, while the upperside is marked by a suture 
separating the alisphenoid from the frontal (perhaps also from the 
parietal) . It is to be remarked that the position of this part of the 
alisphenoid bone in our specimen is, perhaps, exactly the same as in 
Palseomastodon. [11] As the prema.xillse are crushed in front, their 
relation to the nasals and maxillae cannot be made out. [12] The 
maxillae are very large and greatly elongated as in Moeritherium 
lyonsi; but, while in this latter there is only a small palatine vacuity 
between the slightly divergent anterior prolongations of the 
ma.xillse, this vacuity is larger in our specimen, and the anterior 
prolongations of the maxillae undoubtedly form part of the inner 
border of the sockets for the tusks, although they are somewhat 
crushed anteriorly. In this respect our specimen is nearer to 
Palseomastodon than any other species of Moeritherium. Mceri- 
therium gracile has these prolongations longer than Moeritherium 
lyonsi, but they do not reach the inner border of the tusks (comp. 
Andrews, I. c. pi. xvii. fig. 1). [13] The exact shape of the ptery- 
goid is not to be established, but it is certain that its vertical plate 
projects beyond the end of the descending plate of the alisphenoid 
(fig. 1). [14] The upper teeth are so worn that their exact structure 
is not to be made out; but they seem to be the same as in Mceri- 
therium lyonsi." 



THE MCERITHERES 



77 



The author follows with a very detailed comparison of the 
cranial characters of the type of Moeritherium ancestrale with those 
of Pala>omaslodon beadnelli and P. witdoni Andrews, and concludes 
(pp. 58, 59): "The ancestry of M cerilherium to Palseomastodon 
has become very probable from the new additional features of the 
skull in the above-described specimen of Mmrilherium. But, in 
order to make it possible to judge better on this question (Foot- 
note: 'On this point three main opinions have been advanced: — 
(1) Moeritherium lies in the direct line of ancestry to the later 
Proboscidca; (2) Moeritherium is a primitive offshoot of the Probo- 
scidean stock; and (3) Moeritherium is an offshoot of the Sirenian 
stock allied to Proboscidca. The first opinion has been maintained 
by Andrews in his first fundamental memoir on the question (comp. 
C. W. Andrews, 'The Evolution of the Proboscidca,' in Phil. 
Trans. 1903, p. 101), but later on (comp. his Catalogue, Introduc- 
tion, p. xvii) he was rather inclined to adopt the second view. The 
third opinion was expressed for the first time by H. F. Osborn (in 
his article 'The Feeding-habits of Moeritherium and Palseomastodon,' 
in Nature, July 29, 1909), but abandoned (comp. his work 'The 
Age of Mammals,' 1910, p. 203 seq.) after the vigorous refutation of 
Andrews (in his article ' The Systematic Po.sition of Moeritherium,' 
in Nature, Sept. 9, 1909). In a recent paper, however, the 
eminent American paleontologist seems to have returned to his 
former wrong opinion on the subject (comp. H. F. Osborn, 'Palseo- 
mastodon, the Ancestor of the Long-jawed Mastodons only,' in 
Proc. Nat. Acad. Sci. vol. v. 1919, pp. 265-266).'], we give here a 
list of the resemblances and differences in the skull, mandible, and 
dentition of the two genera (Footnote: 'A similar but much shorter 
list is to be found in W. K. Gregorj', The Orders of Mammals, 
1910, p. 367 seq.'\, which are mainly founded on the characters of 
Moeritherium ancestrale on the one hand, and of Palseomastodon 
wintuni on the other." 

Osborn, 1925: In Chapter II of the present Memoir many 
rea.sons are given by both Matsumoto and Osborn why we should 
exclude Mceritherium entirely from the ancestry of Pateomastodon. 

MCERITHERIUM MOLARS COMP.\RED WITH THOSE OF 
P.\LiEOMA.STODON AND PHIOMIA 

Figuivs 42, 4.5, 46, 48, 4!) 

The five new figures and reconstructions of the grinding teeth 
assembled in the present Memoir, namely, figure 48, a comparison 
of the grinding tooth of three genera, figure 45, a comparison of 
Moeritherium lyonsi and M. gracile, figure 46, a comparison of M. 
trigodon and M. andrewsi, figure 42, second reconstruction of the 
skull of Marilherium andrewsi and M. trigodon, and figure 49, 
life-size detailed studies of Moeritherium and Palieoma^todon 
grinding teeth, exhibit very clearly the fundamental resemblances 
and the progressive differences between the dental and cranial 
structure of these three Fayiim genera. Figure 49 beautifully 
displays the superior grinders of Moeritherium trigodon in contrast 
with those of Palseomastodon intermedins, demonstrating that 
Palseomastodon cannot in any way be consitleretl as a descendant of 
Moeritherium. 

(1) In both Moeritherium and Palseomastodon Ma is trilophodont 
(Fig. 49C, E); (2) in both Moeritherium and Palseomastodon the 
secontl and third superior molars, M^^, are in an incipient bilopho- 
dont condition; (3) in Palseomastodon Mi-j have a rudiment of a 




Mcenlhenum 
A/ner. Mus. 13431 Ref. (rev. 



All l/3naLsi2e 




B PaiaeofTiaslodon 

Amer. Mus 14&47 



AiivSnat sijrc 



Fig. 48. C;<)inparison of upper and lower molars to the same 
scale of: 

A, True Mxrilherium, (lower) Amer. Mus. 13437, (upper) 
Amer. Mus. 13431. 

B, True Palseomastodon, (lower) .\mer. Mus. 14547, (upi)er) 
Amer. Mus. 13449. 

C, True Phiomia, (lower) .\mer. Mus. 1346S, (upper) Amer. 
Mus. 13450. 



tritolophid not observed in Moeritherium: (4) Palseomastodon 
is hexabunodont or sexitubercular, conules persisting in M-"', while 
Moeritherium is tetrabunodont or quadritubercular, the conules 
having disappeared; (5) the inner and outer cones are united (Fig. 
49D, E) by a transverse crest in Palseomastodon, M-"', Mj-.i; 
they are separated by a longitudinal cleft in Marilherium. (6) 
As to adaptation, this ilemonstrates that Moeritherium had more 
of the omnivorous, chcerodont habit, while Palseoma.itodon had 
more of the browsing, lophodont habit. The evolution of the 
Palseomastodon molai-s is discussed in full below under Palseomas- 
todon (Chap. \l). 

Conclusion. — This detailed comparison serves to demonstrate 
that both Moeritherium and Palseomastodon may have descended 
from an original primary tetrabimodont stock; that the inter- 
mediate tubercles or 'central conules' never developed (Fig. 37) in 
Moeritherium, while they did develop in Palseomastodon inter- 
medius. M*"', Mj-j. These conules, moreover, are in process of 
development in Palseomastodon, the lophs being composed of six 
cones (Figs. 93, 94). 



78 



OSBORN: THE PROBOSCIDEA 



SUPPOSED MCERITHERIUM OF BALUCHISTAN (Fig. 50) 
In 1912 Pilgrim erroneously referred to Maritherium a frag- 
mentary tooth (Ind. Mus. A 452) contained within "The Verte- 
brate Fauna of the Gaj Series in the Bugti Hills and the Punjab," 
found near Khajuri, Bugti Hills, Baluchistan, which he described 
as follows (p. 15) : " Mceritherium (?), sp. Plate in, fig. 3. Amongst 
the Bugti specimens is ... a fragmentary tooth, possessing char- 
acters which lead one, with no uncertainty, to diagnose it as Probo- 



to render it possible to recognize its presence. The incompleteness 
of the transverse crests entirely precludes the possibility of its 
belonging to Dinotherium. It is clearly distinguished from premolars 
of Tetrabelodon, not only by the rectangular instead of rounded 
corners, but also by the perfectly symmetrical arrangement of the 
cusps, and the breadth and absolutely transverse situation of the 
valley between the crests. The surface of the enamel also possesses 
irregular striae, which are not seen in Tetrabelodon. In all these 




An 13431 




AJ^J3-. 





AJ^. 14547 



Fig. 49. M centherium (A, B, C) and PaUeomasloihm (D, E) grinders figurod and compared in close detail, denionstrating 
\\\a.i. t\\c liexahunndoni grinder.s o! Palxomaslodon contrast with the tetrahunodnnl grinders ai Mceritherium. .See .'ianic fi(;ure in 
Chapter VI (Fig. 94), see also key figure 93. Natural size. 

A, Mceritherium lyonsi ref. (Amer. Mus. 15898 cast). First and second right superior molars, M'-M^. 

B, Mceritherium trigodon ref. (Amer. Mus. 13431). Right superior grinding teeth, P--M'. 

C, Mceritherium andrewni ref. (Amer. Mus. 13437). Second and third inferior grinders, ^Tj-Ms. 

D, Palsetnnasindon intermedins paratype (.\mer. Mus, 13449). Third right superior molar, r.M^. 

E, Palnmniistodon intenncdiiis type (Amer. Mus. 14547). Third left inferior molar, I.M3 (reversed in drawing). 



scidean. Although only three quarters of it remain there is no 
doubt as to its being a rectangular, oblong tooth, possessing two 
transverse crests, separated by a ])road straight transverse valley. 
Each crest consists of an inner and outer pointed cusp, perfectly 
distinct from one another. The presumed postero-internal cusp is 
continued into a small hinder lobe. A beaded cingulum is well 
developed on that side of the tooth, who.se preservation is such, as 



particulars, it agrees exactly with the hinder upper molars of Maeri- 
therium lyonsi, Andrews. It is, however, rather larger than this spe- 
cies, and the relative height of the cusps differs. It is by no means 
unlikely that further material of the Bugti species will show that it 
belongs to a difTerent genus from Mceritherium. At present its 
interest lies in the proof it affords of the existence in Baluchistan in 
Gaj times of a primitive Proboscidean allied to Mceritherium. 



THE M(ERITHERES 



79 





TrILOPHODON (MtERITHERIUM?) PANDIONIS (?) 

Fig. 50. Incomplete fourth sup rior pprmanent premolar, P^ pos.sihly 
referable to Trilophodtm pandionis. Originally referred by Pilgrim to Mteri- 
Iherium. After Pilgrim, 1912, PL in, figs. 3, 3a: " Mxrilherium. (?) sp. Pilg. 
rp|)er nidlar, 3, surface view, 3a, side view. From the base of the Oaj near 
Kliajuri, Bugti Hills. (Ind. Mus., No. A 452.) Page 15." 

Provisionally, therefore, it may be regarded as an indeterminate 
species of the Egyptian genus. The measurements of the tooth 
arc approximately as follows: — 



Length 42 mm. 

Breadth 31 " 

Locality. — The specimen was found at Khajuri, near the very 
base of the series, and hence, is probably lowermost aquitanian." 

Osborn, 1926: The supposed Mcerilheriurn superior molar of 
Baluchistan proves to be a fourth superior premolar, P', probably 
referable to Triluphodon pnrvlionis, a characteristic species of 
Sind, India. The original comparison by Pilgrim of this tooth 
with Moeritherium lyonsi, as quoted above, was qualified by his 
remark: "It is by no means unlikely that further material of the 
Bugti s])ccies will show that it belongs to a dififerent genus from 
Man-itheriurn." Similar tetrabunodont superior premolar teeth 
occur as P''-P4 ijoth in Trilophodon and in SerriderUinus; such a 
tooth was selected by Cope as the type of his 'Mastodon' proavus 
{=Serndentinus proavus of the present Memoir), as illustrated 
in figures 362 and 363. 

Unfortunately Pilgrim's original reference of this premolar 
tooth to Moeiitheriumi?) has been widely quoted as demonstrating 
the presence of the genus Maeritherium in India. 




'^\-^ 



■^:\Uv^^,,. 



v'Ji: 



Fig. 51. Restoration (September, 1932) of Mterilhenum atidrewsi, after materials described and figured by .Andrews (1906), Schlosser 
(19II), and Matsumoto (1922-1923), and reproduced in figures 41 and 42. One-sixteenth natural size. 

This restoration embodies the Amlrews-Matsumoto-Osborn conception of Maritherium as of a swamp- and river-loving stock. As 
detailed above, the conclusion is, that it was a confirmed and continual river-Hving animal, feeding mainly under water and on the banks, 
as indicated by its feeble pelvis, more specialized for aquatic life than Hippopolaiiius, but less specialized than the Sirenians. The resem- 
blances to the Sirenians are largely parallelisms. Mwrilhcriiim is one of the primary stocks of the Proboscidea. The marked resemblances 
to the river-living Hippopotamus are also i)arallelisms, namely, the small, elevated orbits, with small eyes and probably small ears, the 
elevation of the eye and ear openings to the upper level of the face, the prominent anterior tusks, the relatively short, stout limbs, probably 
encased in fat. The habitat was on a river-bed meandering over a low-grade flood-plain, as indicated in the restoration. 



.---■•■■^ 



Deinot/irrium aiya/i2eum 




Kaufiai/ nat «V. 



Fig. 52. Type of Deinolherixim giganleain as originally figured by Kaup, with the incisive tusks erroneously upturned, 
1829, XXII, Heft IV, Taf. i, one-ninth natural size. 




Fig. 53. Type of Deinolherium giganleum as refigured by Kaup, 1.S35.1, Add. Tab. I, fig. 5, 
with incisive tusks downturned. One-fifth natural size. Original in Museum of Darmstadt. 
Cast Amer. Mus. 14286. 

The date of publication of this figure is somewhat doubtful, but it would appear (see 
Kaup, 1841.1, p. 18) that this was prepared to supplement Kaup's Cahier IV of the "Ossements 
Fossiles." 




Fig. 54. Superior dentition of Deinotherium giganleum ref., one- 
fourth natural size. Copied from Gaudry, 1878, p. 187, fig. 247. 



80 



Chapter IV 

THE DEINOTHERES (DEINOTHERIOIDEA) OF WESTERN EURASIA 

AND OF AFRICA 

I. Early separation of the deinotheres from the other proboscideans. Marked distinction 

BY dental, cranial, AND SKELETAL CHARACTERS. FamILY, SUBFAMILY, AND GENERIC DISTINCTIONS. OrIGIN, 
HABITAT, MIGRATIONS, PHYLOGENETIC SUCCESSION, AND EXTINCTION OF THE DEINOTHERES. 

II, Geologic distribution of deinotherium, trilophodon, anancus.rhynchotherium, zygolophodon, 
TURicius, miomastodon, serridentinus, and tetralophodon. Causes of extinction. 

1. Chief distinctions of the Deinotheres from other proboscideans. 4. Conclusions as to origin, phylogeny, and adaptive radiation. 

2. History of discovery of the Deinotheres between 1715 and 5. Geologic distribution oi Deinutherhun, Trilophodon, Anancus, 

1934. Rhynchotherium, Zygolophodon, Turicius, Mioviastodon, 

Historical .surninary l)y Wein.sheimer, 1715-1882. Serridentinus, and Tetralophodon. Causes of extinction. 

Historical summary by Osborn, 1883-1934. Appendix: Deinotherium hioigaricum ii^hik and Deinotherium hop- 

3. R^sum^ of the chief generic and specific characters. woodi sp. nov. 



1. CHIEF DISTINCTIONS OF THE DEINOTHERES FROM OTHER PROBOSCIDEANS 

It is not pcssible for the author to treat the Deinotheres in the same critical or exhaustive manner as the 
species of proboscideans belonging to other families are treated in the present Memoir ; nor is it possible to give 
the type descriptions or reproductions of the type figures as in other chapters. Such exhaustive revision and 
monographic treatment can only be made after a tour of the museums of Europe and of India where the actual 
type specimens are to be found and the precise records of the type localities are given. 

In the succeeding historical summaries by Weinsheimer and by Osborn there is recited the wide range of 
opinion as to the structure, habits, and ordinal relationships of these remarkable mammals between the date of 
the first descri]ition in 1715 and the present time; the author's personal opinions on these points are summarized 
at the end (pp. 112-115). 

Following the ruling of the International Commission on Zoological Nomenclature, the generic name must 
be spelled Deinotherium throughout, in accordance with Kaup 's original orthography of 1829 derived from the 
Greek Seifos and 9r]plov. The same principle applies to the family name Curtognati, proposed by Kaup in 1833, 
which may be given the modern family terminal and orthography CItrTOCNATHID.E, although as in all 
matters of proboscidean nomenclature the early authors naturally did not follow modern nomenclatural canons. 
The history of the term 'Curtognati,' equivalent in German to 'Krummkiefer,' in reference to the sharp downward 
curvature of the lower jaw, is given as follows in the Bulletin of the Geological Society of France for 1834 (Bou6, 
Bull. Soc. g6ol. de France, 1834, V, p. 444): 

81 



82 OSBORN: THE PROBOSCIDEA 

Curlognati.—M. Kaup a regu de nouveaux ossemens d'Eppelsheim, et a reconnu que le Dinotherium medium etait veri- 
tablement bien different du D. giganteum. M. Kaup a pu faiie aussi quelques corrections et additions a sa description du 
Diiwiherium. II pense que ses deux enormes defenses ne lui servaient pas seulement pour extraire des racines de la terre, 
mais encore, comme a la morse, pour I'aider a mouvoir son corps si lourd. D'apres la forme des os intermaxillaires, cet 
animal devait avoir une trompe pour porter la nourriture a sa bouche. II est a placer entre les Mastodontes et les Bradypus, 
et il formerait une famille particuliere a laquelle M. Kaup donne le nom de Curtognati. Elle serait caracterisee par la 
machoire inferieure courbee vers le bas, et les deux defenses dirigees vers le bas et en arriere (A^. Jahrb.f. Mineral, Geognosie, 
etc., 1833, cah. 5, p. 509 [516] avec 1 pi.)- 

The following is from Kaup's original description, 1833.4, p. 516: 

Nach meiner Ansicht steht es am besten zwischen Mastodon und Bradypus als eigene FamiUe, die ich Curtognati 
(Krummkiefer) nenne, und durch den nach Unten gekriimmten Unterkiefer mit den zwei Stoss-Zahnen ebenfalls nach Unten und 
Hinten gerichtet, charakterisire. Die Gattungs-Diagnose wurde seyn # Stoss-Zahne. Der ausfiihrlichere Charakter: Keine 
Stoss-Zahne im Oberkiefer, zwei nach Unten geki-iimmte Stoss-Zahne im Unterkiefer, f Backen-Zahne, wovon der dritte 
dreihiigelig, die ubrigen zweihiigelig, und nur der erste des Unterkiefers mit einem schneidenden Rand auf der vorderen 
Halfte versehen ist. In der Jugend ist vor dem ersten Backen-Zahn noch ein uberzahliger, und der Milchzahn des zweiten des 
Oberkiefers ist in die Lange gestreckt und dreihiigelig. Russel, Krallen zum Aufscharren, Gang auf den Hand-Randern, wie beim 
Faulthier. 

In 1868 Falconer erroneously assigned the name A ntoletheriiim to specimens from Attock belonging to the genus 
Deinotherium ; this error was corrected by Lydekker in 1885. 

Antoletherium Falconer, 1868. — (Falconer, 1868, I. p. 416): "Among the fossils discovered by Lieut. 
Garnett, near Attock, is a portion of the lower jaw of a tapiroid animal containing what appear to be the last 
premolar and the first and second true molars. Of this specimen I have received an excellent drawing executed by 
Col. Baker, who regarded the species as alHed to Tapir. (Plate xxxiv. figs. 1 and 2.) The teeth certainly differ 
generically from those of Dinotherium. in the massive connecting bridge between the two ridges, speedily running 
them into one confluent disc. The bend of the first ridge is very tapiroid. The specimen appears to me to be of an 
undescribed genus for which the provisional name Antoletherium (avToXi] the east, and dr\plov) would be 
appropriate." 

(Lydekker, 1885, p. 105): "A. 20. Fragment of the right ramus of the mandible, containing p,n 4^ „, 1 
and .iTi ; from Kushalgarh. Figured in 'Falconer's Palseontological Memoirs,' vol.i., pi. xxxiv., figs. 1 and 2 (as 
Antoletherium), and in the 'Pal. Ind.' op. cit. [Lydekker, 1880] pi. xxix., figs. 2, 3. Presented by Lieuts. Garnett 
and Trotter." No. A. 20 is listed by Lydekker (1885, p. 105) under specimens of Dinotherium pentapotamise in 
the Indian Museum, Calcutta. See also Lydekker, 1880, p. 184, in which he states: "As I have already observed 
in the above-quoted passage in the 'Records' [Geol. Surv. India, 1877, Vol. X, p. 33], the figure of this specimen 
given by Dr. Murchison in the 'Palseontological Memoirs' [Footnote: 'Vol. I, Plate xxxiv, figs. 1-2.'] under the 
name of Antoletherium is entirely incorrect. It appears probable that Dr. Falconer never saw the original specimen, 
but made his new genus Antoletherium solely on the evidence of the imperfect drawing from which the figure was 
taken; a copy (or the original) of this drawing was sent by Dr. Falconer to Professor Owen, whose note on the 
subject will be found on page 416 of the first volume of the 'Palseontological Memoirs.'" 



THE DEINOTHERES 



83 



Original and subsequent nomenclatural references and modifications for this Memoir are as follows: 



Original Name 
p. 401 
All subsequent authors adopted the abbreviated 



Deinotheriwn Kaup, 1829 
Dinotherium Kau]), 1832. 

spelling 
Curtogntiti Kaup, 1833 
Dinutheridae Boniiparte, 18)5, in his "Catalogo nietodico dci Mammiferi 

Europei," cited by Palmer, 1904, p. 738 
Uinotheriidte Bonaparte, 1850, cited by Osborn, 1918.468, p. 134 
Diiiotherina Bonaparte, 1841 (1837, fide Graells, 1897, p. 573); Dino- 

theriina Bonaparte, 1 850, cited as a subfamily 
Antolelherium Falconer, 18G8 
Dinotheriini Winge, 1906, p. 172 
Dinotheriinse Osborn, 1910.346, p. 558 
Dinotherioidea Osborn, 1921.515, p. 2, "typified by the Miocene and 

Pliocene Dinotheres of Eurasia" 



Reference in 
Present Memoir 

= Deinotherium 

= Deinotherium 

= CURT()C.NATniD.'E 

= CURTOGXATHID.E 
= CURTOO X ATH I HJP. 

= Deinotheriin^ 
= Deinulherium 

= DEINOTlIERIINiE 

= Deinotheriin^ 

= DEINOTHERIOIDEA 



The following is Osborn's definition of the suborder or superfamily DEINOTHERIOIDEA (Osborn, 
1921.515, p. 3): 

II. — Dinotherioidea. All agree that these animals were chiefly fluviatile and amphibious in habit, in this 
respect resembling Maritheres but differing in the entire loss of the superior incisor teeth. Early loss of upper tusks released 
the inferior. In the downturning of the inferior tusks the Dinotheres are paralleled by the Rhj'nchorostrinae among the Masto- 




Fig. 55. Fundamental arrangement of the cutting teeth in the four su|)er- 
faniilie.s of the Proboscidea. For lateral aspect of the lower cutting teeth, compare 
fig. 1. p. Hi. See also fi<i. 4, p. 23. 

A. — Mirrilhirium. B. — Deinotht-rium. C. — Rhynchotheriurn. Composition of two species. 
D. — Sttvodon. S. itaiiesa stage. Scale not uniform. 

dontoidea. In skull form and in limb and foot structure the Dinotheres parallel the true proboscideans. Thej' diverge very 
widely from proboscideans in the evolution of the upper and lower grinding teeth. The primitive Dinotheres present 
simple bilophodont grinders, similar to those of Ma'riihvrium, and are persistently bilophodont. The upper grinders attain a 
stage which parallels the molar pattern of the tapir {Tapiroides[Tapirus]) among the perissodactyls, but show no ten- 
dency to the trilophodont, tetralophodont, or polylophodont structure characteristic of the mastodons and elephants. 



Thus the technical range of classification and nomenclature for these animals stands as on the following page. 



84 



OSBORN: THE PROBOSCIDEA 



Superfamily: DEINOTHERIOIDEA Osborn, 1921 

Family: CURTOGNATHID.E Kaup, 1833 

Subfamily: Deinotheriin^ Bonaparte, 1850, Winge, 19U6, Osborn, 1910 

Genus: DEINOTHERIUM Kaup, 1829 

So far as known the Deinotheres are certainly monophyletic. There is a single phylum of successive ascending 
mutations and species extending from the base of the Miocene, where they suddenly appear in Europe {D. cuvieri), 
to the Lower or Middle Pliocene of Rumania, where the pre-final stage is attained in D. gigantissimum. The final 
stage is D. hopwoodi of the Pleistocene of Africa. 




Fig. 56. Geographic distribution (according to the numbers in the list below) of the types of Deinotherium. The approximate localities where these 
twenty-six' types were found are represented by the white dots within the black areas; these dots each carry a number in a circle representing the chronologic 
seqnence of type description. The white crosses represent referred specimens. 



ORDER OF DISCOVERY AND DESCRIPTION OF TWENTY-SIX' SPECIES OF DEINOTHERIUM 
DESCRIBED OR NAMED, 1715-1935. COMPARE TABLE I, P. Ill, AND TABLE II, PP. 114, 115 

See Figure 56 



Original Name 
1715 Reaumur publishes Antoine de Jussieu's drawing of a grinding 

tooth discovered near Lyons, France, in the previous 

century 
1773 Rozier describes and figures a tooth discovered near Vienne, 

Dauphiny, France 
1798-1822 Cuvier ascribes grinding teeth found in France to "Tapirs 

gigantesques" 

1. 1828-1829 Kaup describes a lower jaw from near Eppelsheim, 

Germany, now in the Museum of Darmstadt, assigning 
in 1829 the name Deinotherium giganteum 

2. 1831 Gmiind, Bavaria Deinotherium Bavaricum 

von Meyer 

3. 1832 Eppelsheim, Germany Dinotheriwn maximum Kaup 

(MS.), fide von Meyer, 
1832.2, p. 78 



Specific Reference 
IN Present Memoir 



= Deinotherium sp.(?) 
= Deinotherium sp.(?) 
= Deinotherium sp.(?) 

= Deinotherium giganteum 
= Deinotherium bavaricum 



Geologic Age 



Ix)wer Pliocene 
Upper Miocene 



'See No. 6a, p. 85, and footnote 1. 



= Deinotherium giganteum Lower Pliocene 



THE DEINOTHERES 



85 



4. 

5. 
6. 


1832 

1833 
1833 
1835 


Comminge, Carlat-le-Comte, 

and Chevilly, France 
Kaup proposes the family name 
Eppelsheim, Germany 
Podolia, Russia 


6a 
7. 
8. 


1835- 

1835 

1836 


1837 Podolia, Russia 
Ural Mountains 
Simorre, France 


9. 


1839 


Locality (?) 


10. 


1839 


Locality (?) 


11. 


1839 


Locality (?) 


12. 
13. 


1841 
1841 


Eppelsheim, Germany 
Mosskirch, Germany 




1843 


Darling Downs, Australia 


14. 

16. 
16. 


1845 

1845 
1845 


Compubay [Cambay, cf. p. 90], 

India 
France 
Perim Island, India 



17. 1861 Grive Saint-Alban, Isere, France 

18. 1868 Perim Island, India 

19. 1876 Attock, Indus Valley, India 

20. 1880 Sind, India, Salt Range 

21. 1892 Gaiceana, Tecuciu, Rumania 

1908 Bugti beds, Bugti Hills, 
Baluchistan 



22. 1911 Karungu, British East Africa 

23. 1912 Bugti beds, Bugti Hills, 

Baluchistan 

24. 1930 Kotyhaza (Dep. Nograd, 

Hungary) 

25. 1935 Olduvai, Tanganyika Territory, 

British East Africa 

26. Kavirondo Gulf, Africa 



Dinolherium Cuvieri Kaup 

CuRTOGNATi for the Deinotheres 
Dinotherium medium Kaup 
Dinolherium proavum, {Dinotherii 

proavi Eichwald) 
Mastodon podoKcum} Eichwald 
Dinotherium uralense Eichwald 
Dinotherium secundarium Kaup 

{teste Lartet, 1836, p. 218) 
Dinolherium [gig.\ var. majus de 

Blainville (fide Weinsheimer, 1883, 

p. 210) 
Dinotherium [gig.] var. medium de 

Blainville (/rfe Weinsheimer, 1883, 

p. 210) 
Dinotherium [gig.] var. minus de 

Blainville {fide Weinsheimer, 1883, 

p. 210) 
Dinolherium Konigii Kaup 
Dinotherium minutum von Meyer 

(in Bronn) 
Dinotherium Australe Owen 



Dinotherium anguslidens Koch 
Dinotherium intermedium de Blainville 
Dinolherium indicum Falconer 

Dinolherium levius Jourdan 
Dinotherium Perimense Falconer, cited 

in Murchison (name only) 
Dinotherium pentapotamix Lydekker 
Dinotherium sindiense Lydekker 
D. [Dinotherium] giganlissimum 

Stefanescu 
Dinotherium ndricum Pilgrim (name 
abandoned by author and sub- 
specific term gajense substituted — 
see below) 
Dinolherium hobleyi Andrews 
Dinotherium indicum Lyd. var. 

gajense Pilgrim 
Prodinolherium hungaricum 
Ehik 

Deinotherium hopwoodi sp. nov. 

Deinotherium^ sp.? Leakey 



= Deinotherium cuvieri 

= CURTOGNATHID/E 

= Deinotherium medium 

= Deinotherium proavum 
= Deinotherium podolicum 
= Deinotherium uralense 
= Deinotherium secunda- 
rium (?) 



Lower Miocene 



Lower Pliocene 

Upper Miocene(?) 
PlioceneC?) 
Lower Pliocene 

Middle Miocene 



= Deinotherium giganleum Lower Pliocene 



= Deinotherium giganleum Lower Pliocene 



= Deinotherium giganleum 



-Diprolodon auslralis (not 
a proboscidean) 

= Deinotherium sp.(?) 

= Deinotherium intermedium 

■- Deinotherium indicum {fide 

Lydekker) 
= Deinotherium levius 

■ Deinotherium indicum 
- Deinotherium indicum 
= Deinotherium sindiense (?) 
= Deinotherium giganlissi- 
mum 



Lower Pliocene 
Lower Pliocene 

Upper Miocene 



Middle Pliocene 
Upper Miocene(?) 

Middle Pliocene 
Upper Miocene 

Middle Pliocene 
Upper Miocene 
Middle Miocene 

Middle Pliocene 



= Deinotherium indicum ga- 
jense 

- Deinotherium hobleyi 

- Deinotherium indicum 

gajense 



Lower Miocene 
Lower Miocene 
= Deinotherium hungaricum Lower Miocene 



= Deinotherium hopwoodi 

sp. nov. Middle Pleistocene 

= Deinotherium hopwoodi (?) Pleistocene 



2. HISTORY OF DISCOVERY OF THE DEINOTHERES, 1715-1934 

The history of discovery is most picturesque and even romantic because of the wonder and astonishment 
caused by the discovery of these pre-Diluvian mammals and the many and often grotesque theories as to their 
habits and external appearance. 

For the early history of discovery up to the year 1883 we may depend chiefly on the extremely thorough and 
scholarly article published by Dr. O. Weinsheimer in the Palseontologische Abhandlungen, Band I, Heft 3, BerUn, 
1883, entitled, "Uber Dinotherium giganteum Kaup," pp. 208-216. This history we have translated with 

'Type mandibular symphysis. First named Mastodon podolicum by Eichwald in 1835 (p. 736, Tab. LVi, LVii, see also p. 734); subsequently (Neues 
Jahrb., 1837, p. 44) referred by Eichwald to Deinotherium. Falconer (1816, p. 4) states in a footnote that the specimen appears to be a fragment of the sym- 
physis of the lower jaw of Deinotherium giganteum. 

^An undescribed specimen, of Pleistocene age, associated with the type remains of Homo kanamensis (see Leakey, 1934, pp. 204, 206, 207, also p. 105 of 
the present Memoir). 



86 OSBORN: THE PROBOSCIDEA 

marginal inserts and a few comments of our own together with illustrations of some of the important specimens 
alluded to by Doctor Weinsheimer in his interesting narrative of the progress of discovery. For the bibliography of 
the Deinotheres, the reader is referred not only to Weinsheimer and to the authors mentioned by him in the 
following historical summary but to the Bibhography at the close of the present Volume I. The more recent 
history of discovery is compiled by Osborn (p. 95). 

HISTORICAL SUMMARY BY WEINSHEIMER, 1715-1882 

Translated by C. D. Matthew from the original of Wein.sheimer, with notes, modifications, and additions by Osborn, 192,5 

Reaumur, 1715.— In 1715 Reaumur (1715)' published an illustration from a drawing by Antoine de Jussieu of 
a tooth of Deinotherium found near Lyons in the beginning of the previous century, without, however, ascribing it to 
any particular animal. From that time no further remains of this remarkable animal of the Miocene Tertiary 
were found until Rozier (1773) described and figured another — a molar tooth discovered in Dauphiny in the 
neighborhood of Vienne by Gaillard — without knowing of the tooth previously figured by Reaumur. 

The third and most remarkable discovery of Deinotherium remains was of two lower jaw fragments, unearthed 
near Comminge, France, and described and figured by de Joubert (1785). In the same year Ildephons Kennedy 
(1785) published his "Abhandlung von einigen in Baiern gefundenen Beinen," in which he describes five teeth and 
an astragalus as the best preserved parts of a complete skeleton of Deinotherium found April 6, 1762, near Reichen- 
berg in Lower Bavaria. Kennedy says that these remains show striking resemblance to remains [of mastodonts] 
from the Ohio River in North America, from Peru in South America, from Asia, and from the small town of Flirth 
in Lower Bavaria (1785, t.2 [IV], f.6), and remarks at the close of his article that these remains belonged to an 
animal that had long been extinct, and that could not be distinguished from the Siberian mammoth, the 
"Ochsenvater der Louisianer." 

CuviER, 1798-1822. — These teeth, described by de Joubert and Kennedy, were more carefully studied by 
G. Cuvier (1798) who, on account of their similarity to the teeth of the tapir, ascribed them to some gigantic 
extinct individual of that genus, which was contemporary with Mastodon and Elephas. Some of the teeth wliich 
Kennedy described were better figured by Sommering (1818) who ascribed them to Cuvier's giant tapir. Other 
teeth and jaws of Deinotherium, or fragments of such, were found later at various sites in France, such as St. 
Lary, Arbeichan, Grenoble, Carlat-le-Comte, CheviUy, and others, and — together with the specimen already 
described by Cuvier (1822) — ascribed to a new genus; these Cuvier named "Tapirs gigantesques," for he already 
suspected, in spite of their similarity of form, that on account of their great variation in size they did not belong 
to one single species, and therefore he used the plural. 

Kaup, 1828-1829.— Somewhat later Kaup (1829) made a report to the Naturforscherversammlung of Berlin, 
through Berthold, describing the left half of a large lower jaw, found in 1828 near Eppelsheim in Hesse-Darmstadt 
and sent to the Museum of Darmstadt where it remains to the present day. Much as the molar teeth in this jaw 
resembled those of the tapir, the teeth in the anterior part of the jaw were so widely different from the correspond- 
ing ones in the tapir that Kaup felt j ustified in distinguishing a new genus and species, DeinotJierium giganteum, which 
he believed to be related to Hippopotamus. The tusks of the lower jaw he erroneously placed [Fig 52] pointing 
upward. 

Since that time — particularly during the 30's— a great many remains of Deinotherium have been found at 
various sites in Europe, that is, in France and Germany, as well as in Asia, so that the material and the literature 

'Dates in parentheses after authors' names refer to the biljliograjjhic list given on pp. 214-216 of Weinsheimer 's article. 



THE DEINOTHERES 87 

respecting this genus are now very extensive. Moreover, the various authors who have occupied themselves with 
the specimens have reached widely divergent conclusions in regard to the dentition, the various species, and the 
classification of Deinotherium. 

Theories AS TO Habits and Habitat. — Thus, according to Kaup (1832), /)etno//ienMm apparently had four 
to six tusks in the upper jaw and two in the lower jaw, which were used to stir up the earth when seeking the 
roots of which its food consisted. There were apparently six molars in the upper jaw, and to all appearances yet 
a seventh, anterior to these, which, however, as in Hippopotamus, was lost early. In the lower jaw, on the 
contrary, there were only five molars. He distinguishes two species — one large, Deinotherium giganteum Kaup, 
and one smaller, Dinotherium Cuvieri Kaup. 

According to H. von Meyer (1833 [1832.1]) Deinotherium is most nearly related to Rhinoceros incisivtis; he 
recognizes the same species as Kaup, but — on account of its frequent occurrence in Bavaria — he calls the Bavarian 
animal Deinotherium havaricum, which is one quarter smaller than Deinotherium giganteum. He dissents from 
Cuvier's view that Deinotherium was contemporary with Mastodon and Elephas because their remains have never 
been found associated, claiming that it was contemporary with 'Mastodon' angustidens Siud Rhinoceros incisivus, 
and became extinct earlier than the mastodon of Cuvier, and that it belongs to the latest Tertiary deposits. In 
the plates given by Kaup (1832) and H. von Meyer (1833 [1832]) the tusks of the lower jaw are pictured point- 
ing upward. 

Kaup, 1833. — In 1833 the Museum of Darmstadt (Kaup, 1833, p. 419) acquired an almost perfect left half 
of a lower jaw from Eppelsheim, from which it could be seen that the anterior part of the jaw was inclined down- 
ward, whereupon Kaup (1833, [1833.4, Taf. vii], p. 509) correctedhisearlier mistake (1829 — see our figure 52) . He 
now states that Deinotherium lived only on land, on which — like the sloth Brady pus — it dragged itself along. Its 
principal food consisted of roots and tubers which it dug out of the earth with the two tusks of its lower jaw, and 
stuffed into its mouth with its trunk (Kaup, 1833 [1833.2], p. 172); but the tusks also served, Uke those of the 
walrus, to help its progress on land. It forms a separate family, intermediate between Mastodon and Bradypus, 
which he named Curtognati (i.e., crooked jaw: Kaup, 1833 [1833.4], p. 509 [516]). 

Pallas (1777) — Eichwald (1835). — Eichwald (1835) describes two molars from Podolia found associated 
with teeth of Mastodon podolicum,^ and assigns them to a new species which he names Dinotherium proavum. In 
it he also includes a molar from the Ural described by Pallas (1777), subsequently (1835) named Dinotherium 
uralense. 

Lartet (1835 [1836]). — Lartet (1835) ascribed a few teeth found in the sands of the freshwater deposit 
of Sansan, Gers, to the species D. secundarium Kaup. 

BucKLAND, 1835. — From the e.xistence and downturned position of the tusks Buckland (1835) draws the 
following conclusions in regard to the life and habits of Deinotherium. For a land animal the monstrous lower jaw 
and the consequent length of body would be very cumbrous and wearisome. With an aquatic animal these would 
not be disadvantages. He considers, therefore, disagreeing with Kaup (1833 [1833.2], p. 172), that Deinotherium 
frequented freshwater lakes and rivers, and was most nearly related to the tapirs. The tusks were used for up- 
rooting and plucking roots and large aquatic plants. From their construction it is seen that for this purpose they 
uiiitcMl the nuvhanical strength of the ])ickaxe with that of a very heavy harrow. The weight of the head, pressing 
on tluulownwardly pointing teeth, would increase the effectiveness of such a use, just as a heavy weight is fastened 
to a harrow to increase its effect. Possibly Deinotherium also used its tusks to lay hold of the bank and so to 



'See footnote 1 on page 85. 



88 



OSBORN: THE PROBOSCIDEA 



anchor itself. Thus the nostrils would come above the level of the water, so that the creature might sleep without 
danger and could breathe, while its body, floating in the water, remained concealed with the muscles relaxed and 
resting; for the weight of head and body would also serve to keep the tusks fast hooked in the same place, Uke 
the weight of a sleeping bird whose claws hold it fast to the branches on which it sits. It is also most probable that 
Deinotherium, like the walrus, used its tusks to drag itself on the land, or in case of danger or of attack as a means 
of defense. From the construction of the shoulder-blade it is clear that the fore feet were used in conjunction with 
the tusks for uprooting and dragging out roots and large aquatic plants. Hind limbs it did not have. 




-/l.o/^ 



Fig. 57. Four views of the reconstructed original E])nelsheini skull of Dcinotherium giganleum in the British Museum (Natural History). After 
Andrews, 1921, pp. 525-534, figs. 1-4. 

This is the first skull found at Eppelsheim and referred to by Weinsheimer (188.3, p. 210) as having been exhibited in Paris, March 15 and 16, 1837, 
to the members of the Academy of Sciences and other savants for their consideration, resulting in widely divergent views, and afterward .sold to the 
British Museum in London. 

According to Andrews (1921, pp. 525, 526), this skull was discovered in 1835 by von Klipstein in the well-known ossiferous sands of Eppelsheim (Hesse- 
Darmstadt); the .specimen was described by von Klipstein and Kaup in 18.36 under the title " Bcschreibung und .\bbildungen von dem in Rheinhessen auf- 
gefundenen colossalen Schiidel des Dinotherii gigantei"; while in Paris the skull was examined by de Blainville, who has given one of the best descriptions 
as yet published; it was then intended to send the specimen to London, but whether this was done at that time is uncertain; some years later, in 1849, 
It was certainly in London, and was offered to the British Museum for purchase; it was then examined and reported on by Buckland and Owen; the pur- 
chase, however, was not comi.leted. and the skull seems to have been sent back to Darmstadt. About 1866, von Klipstein sold his collection to Dr. Old- 
ham, Director of the Geological Survey of India, and from him the British Museum acquired the specimen, which is in an excellent state of preserva- 
tion, desjiite the widespread idea that it was injured on its journey to London. 



Kaup AND VON Klipstein, 1836. — Shortly after, one of the most interesting palaeontological discoveries was 
made in the Lower Pliocene of Eppelsheim, Germany, namely, a complete skull of Deinotherium which was 
described by Kaup (von Klipstein and Kaup, 1 836) . Kaup holds to his original view, and thinks that the walrus-like 
structure of the posterior half of the skull— on account of which Buckland considered Deinotherium an aquatic 
animal— offers no difficulties, and that such a flattened occipital region would permit the insertion of more 
powerful muscles than an abruptly ending occiput. Beside, these muscles would have to be excessively powerful 
if Deinotherium u.sed its tusks for tearing and rooting up the earth. This Eppelsheim skull— later sold to the British 
Museum in London, uninjured in transport— was exhibited by von Klipstein and Kaup in Paris, March 15 and 16, 
1837, to the members of the Academy of Sciences and other savants for their consideration, and again there 
resulted widely divergent views. (See Fig. 57 of the present Memoir.] 



THE DEINOTHEllES 



89 



De Blainville, 1837.— Thus de Blainville (1837.1) gives a minute description of the Eppelsheim skull, and 
classes Deinotherium as a genus belonging to the family of aquatic pachyderms [i.e., Sirenia] together with Halicore 
and Manatus, while Elephas, Mastodon, and Tetracaulodon belong to the terrestrial pachyderms [i. e., Proboscidea]. 
In regard to the habits of Deinotherium he practically agrees with Buckland. Later (1839) de Blainville recognizes 
two species, Dinotherium Cuvieri Kaup (1832) and Deinothenum giganteum Kaup (1829), and of the latter he 
distinguishes three separate varieties — Dinotherium majus de Blain., Dinotherium medium de Blain., and Dino- 
tlierium minus de Blain. — based on the .size of the teeth. 

Dumeril (1837, p. 427) and Isidore Geoffroy Saint-Hilaire (1837, p. 429) agree with de Blainville, while Strauss 
(1837, p. 529) considers Deinotherium exclusively aquatic and not able, Uke Hippopotamus and Manatus, to come 
on land. From the large narial openings he concludes that it had a trunk — the existence of which de BlainviUe 




Fig. 58 




0. giganteum 
Fate. 1847, PI. XLII, Fig. 1 



D. giganteum 
Falc. 1847, PI. XLIV, Fig. 1 Irev.) 



Fig. 59 



Fig. 58. First Eppelsheim skull and jaws uf Deinotherium giganlcum, after (he cast and restoration of Kaup, about one- 
twelfth natural size. Copied from Gaudry, 1878, p. 188, fig. 248. Type jaw in Darmstadt Museum; skull in British 
Museum (sec Fig. 57). 

Fig. 59. Restored Eppelsheim skull and jaws of Deinotherium giganteum Kau|). .\fter Falconer, 1846 [1847], from 
original in Museum of Darmstadt, 

.\1, Op. cit., PI. xi,ii, fig. 1. A, Up. cit., PI. XLIV, fig. 1 (rev.). 

doubted — used like that of the elephant for seizing food. The tusks were not u.scd for uprooting plants, because 
they showed no signs of wearing down, but were used like those of the elephant for sclf-defen.se only. It formed a 
separate family, tran.sitional between pachyderms and cetaceans. Jacquemin (1837) placed it between Rhinoceros 
and Elasmotherium, and Owen (1840) between the tapir and the pachyderms with trunks. Owen [1840-1868] 
gives the dental formula of Deinotherium as follows: 

1. Deciduous teeth; Dir:}; DcS^S; DpM = 16. 

2. Permanent teeth; IM; CM; Pj^; Mf^ = 22. 

He claims that of the three milk molars in the young animal, the anterior has no corresponding permanent tooth, 
that the tusks of the male are twice as large as those of the female, and that in both cases they were used for 
uprooting aquatic i)lants, since, Hke Hippopotamus, it was an aquatic pachyderm. 

Kaup, 1841. — After the discovery of the first great [Epi)elsheim] skull [now in the liritish Museum (Natural 
History)], the Museum of Darmstadt received a great number of additional Deinotherium remains from Eppelsheim 



90 



OSBORN: THE PROBOSCIDEA 



and Westhofen, which Kaup studied and described in the "Akten der Urwelt" in 1841. By that time he already 
had extensive material at his disposal, including the complete [Eppelsheim] skull, three skull fragments, 13 lower 
jaws, and 115 molar teeth. He now changes his former views, and considers that the adult Deinotherium had only 
five molars in both upper and lower jaw, while in the young six were present, anterior to the first being a "super- 
fluous" one which, however, soon fell out. 

All the teeth of Deinotherium previously known belonged to only one species [i. e., D. giganteum Kaup] with a 
wide range of variation in size. Dinotherium proavum Eichwald [1835], DinotJierium medium Kaup [1833.3] (see 
below), Dinotherium Cuvieri Kaup [1832] sive bavaricum H. von Meyer [1831] are therefore not admitted by Kaup as 
species. These teeth do not justify the recognition even of large or small varieties. Nevertheless Kaup introduces 
(1841, p. 49) ill addition to his Deinotherium giganteum a new species which he calls Dinotherium Konigii Kaup. 

'Dinotherium giganteum varies greatly in size, so that individuals of this species may have attained a length 
of 18 Parisian feet, while others again may have been barely 1 1 feet long.' On the other hand, the smaller species, 
Dinotherium Konigii, is supposed not to have exceeded the Indian Rhinoceros in length. In regard to the life and 
habits of Deinotherium Kaup reiterates his former opinions. 



'y 




B.i'nrEKx.Euwm , 



Fig. 60. De BlainviUe's original type figure of Dinotherium intermediam, 1839-1864 [1845], PI. iii. 
Possitily from the Upper Miocene of Simorre (cf. p. 111). Right mandibular ramus. One-third natural size. 



Owen's Error, 1843. — Up to 1843 remains of Deinotherium had been found only in Europe, but now it was 
claimed that remains of the creature had also been found in Austraha. Certain of these were sent to London 
and studied by Owen (1843) who ascribed them to a separate species which he named Dinotherium australe 
Owen. Later, however, these remains were recognized as belonging to [the giant marsupial] Diprotodon. 

Koch, 1843. — In 1843 Koch (1845, p. 41) saw in the British Museum in London a part of a lower jaw, 
discovered in Compubay,' western India, and labeled "Mastodon angustidens." By a detailed examination he 
found that it did not belong to the genus Mastodon but to Deinotherium, and, if not to a very old specimen, at 
least to one fully adult, which — on account of the comparatively small molars — he named Dinotherium angusti- 
dens Koch. He completely agrees with Buckland (1835) in regard to its life and habits, and classes it with the 
Proboscidea as intermediate between Tetracaulodon Godman and Missourium Koch. Geinitz (1846) agrees with 
him. 

Falconer, 1845. — Shortly after this remains of Deinotherium were also found in western India. Falconer 
( 1846 [ 1 845] ) describes the posterior half of a lower left molar of Deinotherium from a chalky ferruginous conglomerate 

[ Im Jahre 1843 sah ich im britischen Museum zu London eine in Compubay in Ostindien aufgefundene, mit der Ueberschrift: 'Mastodon angustidens' 
versehene untere Kinnlade." (Koch, 1845, p. 41.) For "Compubay," a locality not on any existing map, probably read Cambay, western India.) 



THE DEINOTHERES 91 

from Perim, a small island at the mouth of the Narbada [Nerbudda River] in the Gulf of Cambay on the west coast 
of India, which differs from the corresponding parts of an Eppelsheim tooth in having the transverse crest and 
the enamel thicker. He therefore ascribes it to a new species which he names Dinotherium indicum Falconer. 
He considers Deinotherium very nearly related to Mastodon Ohioticus (Falconer, 1868, 1, p. 85) and belonging 
hkewise to the Proboscidea. 

Gervais, 1848. — Gervais (1848) classes Deinotherium with Mastodon in the order Proboscidea, and recog- 
nizes three species — Deinotherium giganteum, Dinotherium intermedium de Blainville (see fig. 60), and Dinotherium 
Cuvieri — while Wagner (1850), Girard (1851 [1852]), and Pictet (1853) continue to place it in the order Sirenia. 
Burmeister (1851) holds it for an herbivorous sea monster, 15-20 feet long, with short, thick neck, powerful, 
slender rump, and broad, effective flippers for creeping like the walrus, which like all tlie Sirenia frequented the 
mouths of large rivers, and used its tusks for assisting its movements on shore and for uprooting aquatic plants. 

Discoveries in Western Europe, 1850-1858. — Again in the 50's a great number of remains of Deinotherium 
were found in a variety of places, and announced chiefly in brief reports. Opinions in regard to species, dentition, 
and classification continued divergent. Jiiger (1850) reports the occurrence of Dinotherium Cumeri Kaup sive 
bavaricum H. von Meyer at Mosskirch [Upper Miocene], and of Deinotherium giganteum Kaup near Salmendingen 
in the pisoUtic iron ores of the Swabian Alb. Quenstedt (1851 [1850]) reports the occurrence of the latter in the 
pisolitic iron ores of Melchingen south from Tubingen, and later (1853) in those of Frohnstetten, and says that 
Deinotherium had not five, but six, molars in both jaws. Von Hingenau (1852) found molars near Nikolsburg and 
Keltschau; Glocker (1852) a tusk and .several molars in the blue marl near Abtsdorf ; Fotterle (1853) the right half 
of a lower jaw with tu.sk near P\infkirchen and a molar near Keltschau; and Merian (1854) a molar in the Delsberg 
valley in the Bernese Jura. Pomel (1854 [1853]) classes Deinotherium with Elephas and Mastodon among the 
Proboscidea; Bronn (Lethaea, III, p. 803) as intermediate between Sirenia and Pachydermata; and Agassiz 
(Bronn, Lethaea, III, p. 805) among the herbivorous Cetacea. 

In 1853, in consequence of the collapse of an embankment on the Prague-Briinn railway, a skeleton of Deino- 
therium giganteum was exposed in the blue marl near Abtsdorf, the bones of which lay together, covering a space 
about 24 feet long and 18 feet wide. Many of them fell to dust after being for a time exposed to the air; others, 
before word of the discovery was noised abroad, were shattered or cleared away. All that was preserved was the 
complete jaws and dentition, the first and second cervical vertebrae, a few dorsal and caudal vertebrae, a large part 
of the tarsus and metacarpals, and a few large fragments of the limb bones. These skeletal parts of a juvenile 
specimen, which are now in the Bohemian Museum of Prague, prove incontestably that Deinotlierium should be 
classed not with the Cetacea, but with the Pachydermata as nearly alUed to Mastodon (Reuss, 1855). Neverthe- 
less Giebel (1855) again classes it among the Sirenia as nearly allied to Halitherium. 

Deinotherium remains were found in France at this time by Lockhart (1853) in the Department of Loiret: 
by Bayle (1855) near La Chaux-de-Fonds in the Miocene Molasse which is identical with the deposit of Simorre 
and Sansan [Miocene] and the Faluns of Touraine; and by Rouault (1858) near Rennes, his discovery consisting 
of a large molar of Dinotherium Cuvieri. Hauser (1856) discovered remains of a Deinotlierium skull in the Leitha 
chalk near Breitenbrunn on the Neusiedler See, Hungary. Gaudry and Lartet (1858) found a tibia 95 cm. long, 
and Wagner (1857) a skull fragment with five permanent teeth and a forearm of a young individual at Pikermi [Lower 
PUocene]. Wagner l)elieves that all Deinotherium remains known up to this time belong to one and the same 
species. Aichhorn (1857) found teeth of Deinotherium giganteum near Wies in Steiermark [Miocene]. Fotterle 
(1857) found a fragment of the right side of the chin with a molar and several bones of the torso embedded in fine 



92 OSBORN: THE PROBOSCIDEA 

sand intermixed with rubble (Schotter) lying above the upper brackish blue marl. At other sites in Austria Poppe- 
lack (1858), Homes (1858), and von Hingenau (1858) found various more or less important remains of 
Deinotherium. 

Lartet, 1858 [1859].— Lartet (1858) holds the opinion that the first dentition in the upper jaw of the young 
Deinotherium consists of three milk teeth on either side, followed by the first molar of the second dentition, and 
then by the second molar. The first milk tooth falls out but is not replaced; the second and the last are replaced 
by premolars which are simpler than the corresponding milk teeth. Accordingly he gives the following dental 
formula: 

1. First dentition: DiM; Dp2-4M=14. 

2. Second dentition: IM; P3-4M; MM = 22. 

Lartet recognizes four species: 

1. Dinotherium giganteum Kaup, syn. Dinotherium proavum? Eichwald. 

2. Dinotherium sp.? Lartet. 

3. Dinotherium bavaricum? H. v. Meyer, syn. Dinotherium intermedium? de Blainville. 

4. Dinotherium Cuvieri Kaup. 

Gaudry, 1860-1878. — Gaudry (1860) in his excavations at Pikermi, Greece [Lower Pliocene], collected a 
great number of skeletal fragments of Deinotherium giganteum which he considers as an ambulatory terrestrial 
animal related to Mastodon and Elephas, which was also his opinion in 1861 and 1862 when (p. 162) he gives a 
detailed description of the Greek remains and feels obliged to consider the various species of Deinotlierium merely 
as varieties of one species (1. c. p. 171). Miller (1861) feels obliged to doubt whether Deinotherium was exclusively 
an aquatic animal, and Hensel (1862 [1863]), on account of its skeletal structure, classes it among the pachyderms 
with trunks. Suess (1863) reports a small Deinotherium {Deinotherium havaricum?) from the [Upper Miocene] coal 
seams of Parschlug, Eibiswald, Turnau, Wies, etc., in Steiermark, and a large one, similar to the unnamed species 
of Lartet, in purely marine deposits. Rachoy (1863) and Stur (1864) found remains oi Deinotherium havaricum in 
the overlying sandstone of the coal bearing Tertiary basin of Leoben in Steiermark. 

Cranial Characters, Claudius, 1865. — Claudius (1864 [1865]), from the structure of the auditory laby- 
rinth of Deinotherium, shows positively that this genus, Uke the elephant, belongs to the family of the Proboscidea 
and among the pachyderms. According to his researches both Elephas and Deinotherium have a depressed auditory 
cochlea with less than three convolutions, but notably broad at the base. The planes of the superior and posterior 
convolutions of the semicircular canals which intersect in the crus commune, form an angle less than 90°, while in 
Rhinoceros and Tapirus the cochlea is higher, the angle formed by the superior and posterior convolutions is at 
least 90°, and the cochlea has hardly more than two convolutions. From the size of the canalis facialis the exist- 
ence of a trunk may be inferred with certainty. In Deinotherium the facialis nerve is led from the inner ear through 
an elliptical foramen 5'" long and 2'" wide. The latter figure, wliich remains the same for the passage through the 
tympanic cavity, gives the diameter of the nerve which, from the inner ear outward, penetrates the thin bone in a 
diagonal direction and thus produces an eUiptical opening. The nervus facialis of Deinotherium was about the size 
of the medianus nerve in man, and served like it for the innervation of an enormous number of muscle fibers. In 
the elephant this canal has the same dimensions: in the tapir it is proportionally much narrower. The stapes is 
similar in structure to that of the elephant; the incus somewhat smaller (6'": 4'"); in its morphological rela- 
tions the labyrinth corresponds perfectly to that of the elephant; similarly the canaUs ganglionaris of the cochlea 



THE DEINOTHERES 93 

agrees perfectly with that of the elephant, and consequently shows marked differences from all other mammals in 
this respect. 

Gervais (1864), Falconer (1868). — In Herault, between Narbonne and St. Chinian, in the freshwater 
Molasseof Montredon,Gervais (1864) discovered teeth of Deinotherium giganteum in the Upper Miocene [Lower 
Pliocene], as well as fragments of tusks, several bones of manus and pes, a metacarpal, and a phalanx, also an in- 
ner ear. In Haute-Garonne, Salaro (1864)['] discovered a pelvis oi Deinotherium in which he believed he recognized 
marsupial epipubic bones, and therefore concluded that it was a monstrous pachyderm-Uke marsupial related to 
Megatherium, a view also supported by H. von Meyer (1865, p. 1 et seq.). Quenstedt (1867) has doubts whether 
Deinotherium should be classed with the pachyderms or the cetaceans, while Cams (1868) would place it between 
the Sirenia and the Proboscidea. According to Falconer (1868) Deinotherium had only two milk molars in both 
upper and lower jaws — the posterior with three, and the middle one with two transverse crests. The anterior or 
third remained undeveloped. The two milk teeth were replaced vertically by an equal number of premolars, 
both of which have two crests. Following these come the three true molars, of which the first has three transverse 
crests, and the others each two. He divides the Proboscidea into Mastodon and Elephas, and holds Deinotherium for 
a divergent form of the latter. 

Owen (1840) to Lydekker (1876). — Among the more recent works on Deinotherium is that of Brandt (1869). 
He agrees with Owen (1840) and Lartet (1858) that of the three milk molars [Dp 2-4] the anterior one [Dp 2] 
has no corresponding replacing tooth [P 2], while both the others when they fall out are replaced vertically by two 
premolars [P 3-4]. The reconstruction of Deinotherium which he figures is very similar in form to the elephant, 
and he also says that Deinotherium was a true member (genus) of the family of elephant-like animals, more nearly 
aUied to Mastodon than to the elephants, more nearly allied than either Mastodon or Elephas to many other 
pachyderms, and related — though slightly — to the Sirenia. Its life and habits resembled those of the elephant. 
In a separate chapter he lists the various species of Deinotherium, with the remark that he will not increase the 
confusion and divergence of opinions by offering new and unproven views. 

Peters (1871) describes a number of Deinotherium remains from the highest Miocene deposits of southern 
Steiermark, which include Deinotherium giganteum Kaup, Dinotherium medium Kaup, Dinotherium Cuvieri Kaup, 
Deinotherium bavaricum H. von Meyer, the latter being only a variety of Deinotherium giganteum Kaup. He considers 
Deinotherium as a creature of fluviatile habitat, spending more than half its time immersed in river water. Bach- 
mann (1875) describes and pictures a lower jaw from the Delsberg valley in the Bernese Jura, which he identifies 
with Deinotherium bavaricum H. v. M., and considers as belonging to the pachyderms and not nearly aUied to 
Manatus. Zittel (1875) ranks it with the Proboscidea, between Mastodon and Elephas. Hollander (1877) also 
classes it with the Proboscidea, but says that the structure of the skull is not unlike that of the Sirenia, and that 
it was of aquatic habit. In regard to the dental succession in Deinotherium, Vacek (1877) fully agrees with Lartet 
(1858). He considers Deinotherium as the direct ancestor of Mastodon in the family Proboscidea, and so also 
does KoUner (1882). 

Lydekker (1876) describes a molar from India which he identifies with 'Dinotherium pentapotamicum' 
Falconer and later (1880) recognizes a third Indian species, beside Dinotherium pentapotamicum and Dinotherium 
indicum Falconer, which he founds on a fragment of lower jaw from Sind and names Dinotherium sindiense Lyd. 
He classes Deinotherium with the Proboscidea. 

See Sanna-Solaid in Bibliography at close of present Volume I. 



94 



OSBORN: THE PROBOSCIDEA 



The above historical summary is on the authority of Dr. 0. Weinsheimer (1883), and the reader is referred to 
his monograph "Ueber Dinotherium giganteum Kaup," in which he also gives an invaluable Literatur-Verzeich- 
niss (pp. 214-216) of one hundred and fifteen titles, from Reaumur (1715) to Vacek (1882). Weinsheimer then 
reviews (1883, pp. 217-238) the whole subject of the dentition as observed chiefly in the collections of the mu- 
seums of Germany. He describes the newly found jaw in the Darmstadt Museum (pp. 238-244) also the cranial 
and skeletal material in the same Museum (pp. 244-247), and then considers the nomenclature (pp. 247-264) 
of the following sixteen species described or named up to the year 1880: 



Deinotherium giganteum Kaup type, Eppelsheim, 1829 
Dinotherium cuvieri Kaup type, Eppelslieim, 1832 
Dinotherium medium Kaup type, Eppelsheim, 1833 
Dinotherium maximum Kaup type, Eppelsheim, 1831, 1832 
Deinotherium bavaricum von Meyer type. Upper Miocene, 1831 
Weinsheimer gives a very full discussion with measure- 
ments relating to this synonymy 
Dinotherium proavum Eichwald type, 1835, Podolia, Russia 
Dinotherium secundarium Kaup (ref. Lartet), 1835 

Dinotherium kmigii Kaup type, 1841 

Dinotherium minutum von Meyer type (in Bronn, 1841), 

Upper Miocene, Mosskirch 
Dinotherium australe Owen type, 1843 
Dinotherium angustidens Koch type, Compubay, western India 

[Cambay (?), western India], 1845 
Dinotherium intermedium de Blainville type, 1845 

Dinotherium indicum Falconer type, 1845, compare Dinotherium 

perimense, Falconer, 1868, p. 415, Pliocene, Perim Island 
Dinotherium levius Jourdan type, 1861 

Dinotherium pentapotamix Falconer type (MS.), 1868 
Dinotherium sindiense Lydekker type. Middle Miocene, Sind, 1880 



Synonymy 
fide Weinsheimer 
= Deinotherium giganteum 
= Deinotherium cuvieri 
-Deinotherium giganteum 
-- Deinotherium giganteum 
-■ Deinotherium cuvieri 



= Deinotherium proavum 

= Deinotherium giganteum {fide 

Lartet- Weinsheimer) 
= Deinotherium cuvieri Kaup sive 

D. bavaricum von Meyer 

= Deinotherium bavaricum (?) 
= Diprotodon australe 

= Deinotherium angustidens 
= Deinotherium bavaricum {fide 
Lartet, 1858) 

= Deinotherium indicum 

= Deinotherium giganteum {fide 

Gaudry-Lartet) 
= Deinotherium pentapotamix 
= Deinotherium sindiense 



Monospecific 
fide Weinsheimer 
Deinotherium giganteum 
Deinotherium giganteum 
Deinotherium giganteum 
Deinotherium giganteum 
Deinotherium giganteum 



Deinotherium giganteum 
Deinotherium giganteum 
Deinotheriu^n giganteum 
Deinotherium giganteum 

Deinotherium giganteum 

Deinotherium giganteum 

Deinotherium giganteum 

Deinotherium giganteum 
Deinotherium giganteum 
Deinotherium giganteum 



Monospecific Error. — The above synonymy is partly based on the erroneous treatment by Weinsheimer 
of Deinotherium giganteum [Lower Phocene] as a collective species which embraces many of the Miocene stages 
also. He closes, accordingly, with the following "Schlussbemerkungen" (pp. 280, 281), that in view of the 
great variation in size of the teeth of Deinotherium, and despite the various geologic horizons or layers in which 
the remains of Deinotherium occur, loe have to do uilh one single European species, namely, Deinotherium 
giganteum: 



Hiernach sind also alle Species, welche nach Zahnen oder Kiefern aufgestellt sind, die in Grosse und Gestalt von den dem 
Dinotherium giganteum Kaup zugerechneten nicht mehr abweichen, als von den betreffenden Autoren angegeben wurde, zu einer 
einzigen Species zusammenzufassen, fur welche die von Kaup zuerst gebrauchte Bezeichnung Dinotherium giganteum beizube- 
halten ist. Der Grund der mehr oder weniger grossen Abanderungen in Grosse und Gestalt der Zahne und Kiefer ist, wie wir 
gesehen, theils auf individuelle, theils auf sexuelle und besonders auch auf Altersverschiedenheiten zuriickzufuhren. . . . Um 
dies zu erklaren, mussen wir annehmen, dass das Dinotherium verschiedener Zeitraume und Landschaften die bedeutenden 
physischen Veranderungen, welche allenfalls zwischen die Ablagerungszeiten der verschiedenen Schichten der jiingeren Tertiar- 
periode fallen, uberdauert hat, ohne in merklicher Weise von denselben beeinflusst worden zu sein und ohne wahrend der sehr 
langen Zeitdauer, welche zwischen diesen Veranderungen liegt, die geringste Schwankung seiner Artenmerkmale zu erfahren. 



THE DEINOTHERES 



95 



Weinsheimer's Memoir concludes with an invaluable review (pp. 269-280) of the geographic distribution of 
Deinotherium in the provinces of Germany, Switzerland, France, Austria-Hungary, Russia, Greece, and India. 

HISTORICAL SUiMMARY BY OSBORN, 1883-1934 
The historical summary below by the present author relates to some of the chief and more important papers 
which have appeared since Dr. O. Weinsheimer's review (1883), namely, between 1883 and 1934, but also includes 
citations from some of the earlier authors, as follows: 



Richard Owen, 1840-1868 
Giegoriu Stefanescu, 1892-1899 
E. Kittl, 1908 
Lucien Mayet, 1908 



Charles W. Andrews, 1911 
R. Fourtau, 1918 
A. Brives, 1919 
Charles W. Andrews, 1921 



C. Forster Cooper, 1922 
Gunther Schlesinger (letter), 1923 
Othenio \he\, 1924 
R. W. Palmer. 1924 



Richard Owen (1840-1868).— First we may revert to Richard Owen (1840-1868), w^ho records (1840, p. 609, 
1868, III, pp. 358, 359) the permanent dentition in Deinotherium as, I"^; C"-^; Po^; M3-| = 22, and adds three 
superior and inferior deciduous premolars and one deciduous incisor, namely, Dp 2-4§, Dir. 

As regards the tusks, Owen observes (1868, III, p. 359) : "The generic peculiarity of the Dinotherium is most 
strongly manifested in its tusks. These, fig. 288, i, are two in number, implanted in the prolonged and deflected 
symphysis of the lower jaw, in close contiguity with each other, and having their exserted crown directed down- 
ward and bent backward, gradually decreasing to the pointed extremity. In jaws with molar teeth of equal size, 
the symphysis and its tusks offer two sizes; the larger ones, which have been found four feet in length, with tusks 
of two feet, may be attributed to the male Dinothere; the smaller specimens, with tusks of half size, to the female. 
The ivory of these tusks presents the fine concentric structure of those of the Hippopotamus, not the decussating 
curviHnear character which characterises the ivory of the Elephant and Mastodon. No corresponding tusks, nor 
the germs of such, have yet been discovered in the upper jaw of the DinoUierium." 



1878-1899. THE GIANT DEINOTHERES OF RUMANIA 
Stefanescu, 1892. — Professor Gregoriu Stefanescu of the University of Bucharest discovered remains of 
Deinotherium far exceeding in size any previously described, to which he gave the name Dinotherium gigantissimum, 
in allusion to Kaup's term Deinotherium giganteum of 1829. 

There was first discovered in 1878, at Gaiceana, district of Tecuciu, a molar tooth, probably an M2, of 120 
mm. in length and 120 mm. in width, thus exceeding in size the corresponding tooth of Deinotherium giganteum. 
This molar becomes the type grinding tooth of Stefanescu's new species Dinotherium gigantissimum, 1892. It was 
embedded in yellowish-gray sand {op. cit., 1892, fig. 24) belonging to the Middle Miocene [Lower or Middle PHocene]. 
An abbreviated summary of this very important discovery taken from his paper published in 1892 is as follows: 

(Stefanescu, 1892, pp. 81, 82): "I received in 1878 a fossil molar tooth found at Gaiceana, in the judet 
(district) of Tecuciu. It was the last but one molar of a Dinotherium, but it was so large that it could not have 
belonged to the usual D. giganteum, as you may judge from the following dimensions: 

Antero- posterior diameter meters 0.12 

Transverse diameter " 0.12 

Height of the crown " 0.08 

Height of the root " 0.14 

Distance between the hills of the crown " 0.05 

Thickness of the hills at their basis " 0.05 



96 OSBORN: THE PROBOSCIDEA 

These uncommon dimensions should lead us to look at these remains as belonging to another species than the 
usual D. giganteum, which may be named D. gigantissimum." 

In the same paper Stefanescu describes a second discovery at Manzati of which the following is an 
abstract (p. 82) : 

Manzati Valley. — In 1890, in the valley of Manzati, district of Tutova, was discovered a head of Deinotherium, 
partly removed by visitors, also other remains, preserved as follows: 

1. Right branch of lower jaw with five molars, P3-M3, referred to D. gigantissimum [Stefanescu, 
1896 (1899), Tab. i]. 

2. Portion of left branch of lower jaw with two posterior molars, referred to D. giganteum 
[Stefanescu, 1895, p. 175]. 

3. Fragment of right branch of upper jaw with portion of palatal bone and three molars, 
referred to D. gigantissimum [Stefanescu, 1896 (1899), Tab. in]. 

4. Ten ribs, one of which measures 1200 mm., of doubtful reference, Stefanescu, 1892, p. 83. 

5. Scapula, measured in situ, transverse diameter 1150 mm., Stefanescu, 1892, p. 83. 





Fig. 61. Outline of rpfcrrod grinding tepfh nf Demn/Zicnjini gigat^liasitmim, after Stefanescu, 1895, Tab. iv .ind v. Red\iced to two-thirds 
natural size. 

A, Second left inferior molar, I.M2, from Manzati. 

B, Third left suiicrior molar, l.M', from Gaiceana. 

According to the above descriptions of Stefanescu of 1892, 1895, 1896 [1899], remains referred to Deinotherium 
giganteum and to D. gigantissimum occur in the same horizon in the Manzati Valley (see Stefanescu, 1892, fig. 24, 
section through Manzati Valley) . 

Historical Review.— In the Introduction to his paper of 1895 (pp. 173-177) Stefanescu gives a full historical 
review, "Historique du Dinotherium," of the successive discoveries of remains of Deinotherium together with 
discussions as to the anatomy and affinities of this animal, beginning with Cuvier (1822) and ending with 
Hollander (1877) and Bieber (1884), namely, from the Tapir gigantesque of Cuvier to the Franzensbad 
Deinotherium. 



THE DEINOTHERES 



Description of Dinotherium gigantissimum Stefanescu, 1878, 1892, 1895, 1896 [1899] 

Stefanescu's original description of the type molar of Dinotherium gigantissimum, "Decouverte d'une molaire 
de Dinotherium en Roumanie," 1878, is not accessible for the present Memoir. The accompanying outlines (Fig. 
61) copied from Stefanescu's paper of 1895, Tab. iv and v, exhibit a second inferior molar and a third superior 
molar referred by him to this species. 

In 1895, however, Stefanescu gave a full description of the discovery of Deinotherium in Rumania, under the 
subtitle "Le Dinotherium en Roumanie," in which he mentions the discovery in 1878 of remains at Fundul 
Boghitii, Gaiceana, district of Tecuciu, referring them in part to Deinotherium giganteum and in part to the 
largest species which he had named Dinotherium gigantissimum. 

After reviewing (op. cit., 1895, pp. 173-177) the 
chief materials found in Rumania referable to D. gigan- 
teum and to D. gigantissimum, Stefanescu describes and 
figures in detail (pp. 177-195, Tab. i-v) the skeleton 
and teeth from Manzati referred to the giant species D. 
gigantissimum, including tlie successive stages of dis- 
covery of: (a) Inferior tusk (1889), ih) left inferior 
maxilla and large part of right inferior maxilla, also 
part of left superior maxilla (1890), two ihae, two 
scapulae, vertebrae, four Hmbs, eighteen ribs, one tusk, 
fragments of cranium, debris of teeth and of different 
bones (1890). He concludes (p. 183): "Comma on le 
voit. j'ai eu le bonheur de trouver a Manzati le squelette 
presque complet de ce colossal Dinotherium; je n'ai 
trouv6 cependant que deux vertebres, mais il est probable 
qu'on trouvera aussi les autres." The bones were 
found within a space 10 m. in length and 3 m. in 
breadth, 30 sq. m. in all. Stefanescu indicates the fol- 
lowing conditions regarding depth (op. cit., p. 183): 
"Ces faits prouvent ou que le Dinotherium de Manzati 
est mort envase dans un endroit bourbeux, en cherchant 
k y boire ou a s'y rafraichir, ou que, apres sa mort, 
il a 6t6 presque aussitot recouvert d'une couche de 
limon, qui a empeche le corps de se decomposer a 
I'air, ou d'etre la proie des animaux feroces qui eussent disperse les os et y eussent laisse la trace de leurs dents 
aigucs." 

The following details are important : (1) In this excavation all the parts belong incontestably to one individual ; 
(2) nearly all the bones were found in normal connection ; (3) no doubt remains that Deinotherium had but three 
digits [D.II, III, IV — Fig. 62], with a vestigial D.I in the pes; the tarsal and metatarsal bones are more or less 
cylindrical. 




■*m. 



Fig. 62. Front and side views of tibia and pes of Dinotherium gigan- 
lissimum, after Stef&nescu, 1899, Tab. iv. This retouched drawing is one- 
twelfth natural size. 



OSBORN: THE PROBOSCIDEA 



From the detailed measurements of all these parts, we may extract the following: Found with the skeleton 
from Manzati, a second left inferior molar, 1. M,, diameters, ap. 1 11 mm., tr. at protolophid 90 mm., index 81. 
Found at Gaiceana, a left third superior molar, 1. M^ diameters, ap. 107 mm., tr. 105 mm., index 98. 



Comparative Diameters and Lengths of Grinders and Limbs' 



Anteroposterior diameter of Mo 
Anteroposterior diameter of M3 
Space occupied by five inferior 

molars, Ps-M;) 
Length of inferior tusks 
Length of tibia (cf. Fig. 62) 
Height of tarsus (cf. Fig. 62) 
Total length of median Mts. HI 



Tapir 


Deinolheriuvi 


Dinotherium 


Dinotherium 


Dinotherium 


gigantesque 


giganleum 


indicum Falc. 


gigantissimum 


gigantissimum 


Cuvier 


Kaup 


and Pilgrim 


Stefanescu 
Manzati 


Stefanescu 
Gaiceana 


85 


81 


99 


111 






91 




114 


107 




453 




540 

560 






920e 




930-976 
210 
152 





The author concludes (1895, p. 197): 



Deuxieme molaire de la mdchoire 


Tapirus gigan- 


Dinoth. gigan- 


Dinotherium 


Dinotherium 




inferieure [second inferior molar, 

Md 


teus de France 
Cuvier 


teumd'Eppels- 
heim Kaup 


Indicum de Perim 
Falconer 


gigantissim. 
de Manzati 
Stefanescu 




Diametre anteroposterieur 
Diametre transversal de la colline 


0,085 


0,081 


0,099 


0,111 




posterieure (a la base) 
Diametre transversal de la colline 


0,073 


0,072 


0,088 


0,097 




anterieure (a la base) 


0,075 


0,076 




0,099 




Derniere molaire de la machoire 




Dinotherium 




Dinotherium 


Dinotherium 


superieure [third superior 
molar, M^] 




giganteum 
d'Eppelsheim 




de Manzati 


de Gaiceana 


Diametre anteroposterieur 
Diametre transversal de la colline 




0,091 




0,114 


0,107 


posterieure (base) 
Diametre anteroposterieur de la 




0,084 




0,107 


0,104 


colline anterieure 




0,091 




0,117 


0,117 



The author observes in his paper of 1896 [1899], p. 127, as follows: "Si, d'une part, nous considerons les 
dimensions de beaucoup plus grandes des dents du Dinotherium Gigantissimum, en comparaison avec ceUes des 
autres Dinotheriums connus, et surtout avec celles du Dinotherium Giganteum, si, d'autre part, nous prenons en 
consideration I'existence d'un bourrelet crenele continu a la base de toutes les molaires de la machoire superieure, 
et surtout de la premiere molaire, ce qu'on ne trouve pas de la sorte chez les autres especes, nous pouvons, a bon 
droit, regarder le Dinotherium de Manzati comme appartenant a une autre espece, et consequemment, regarder 
comme etant assez justifiee, la creation de I'espece Gigantissimum." 



'Limb measurements of /; bamricum of Fraiizen.sl)ad inserted below (pp. 100, 102). 



THE DEINOTHERES 



99 



1908. FIRST RECONSTRUCTION OF THE SKELETON OF DEINOTHERIUM BAVARICUM 

UpjK'i- Miocene of Franzensbad, Bohemia 

E. KiTTL, 1908. — The first description of this skeleton is that of Kittl in his article "Das Dinotheriumskelett 
von Franzensbad im k. k. naturhistorischen Hofmuseum" (1908). In describing the skeleton, ffittl remarks that 
this almost completely restored skeleton of Deinotherium arouses wide attention in the Vienna Museum becau.se 
the only other skeleton which appears to be nearly as complete is in the University Museum of Bucharest. He 
speaks of the abundance of scattered skeletal materials in sand and gravel deposits of western Europe, chiefiy of 
fluviatile and flood-plain origin, in contrast to the rarity of entire skeletons, in fact, an entirely complete skeleton 
had never been discovered up to 1908. The two most complete discoveries of this animal are those made near 
Franzensbad in Bohemia and near Manzati in Rumania, the former exhibited in Vienna, the latter in Bucharest. 

These animals belong to different species and are widely dis- 
tinguished by their size. The Franzensbad Deinotherium attains 
the dimensions of a half-grown elephant, whereas the Rumanian 
skeleton [D. gigantissimum] equals in size the largest of the Uving 
elephants. Its specific distinctions, however, rest less upon size 
than upon the structure of the grinding teeth; the condition of 
wear of the teeth in both skeletons indicates prolonged use. 

Comparing the [Upper Miocene] Franzensbad Deinothere with 
other skeletons, it appears to belong to a small, perhaps to the 
smallest, species of the genus Deinotherium, whereas the Rumanian 
skeleton \D. gigantissimum] belongs to the largest variety of this 
genus. Between these two extremes we can range a series of inter- 
mediate size. It is the middle-sized Deinotherium which was first 
observed; as early as 1812 Cuvier designated them as large tapirs 
owing to the resemblance of the back teeth to those of the tapir; in 
1833 [1829] they received the generic name Deinotherium with 
reference to their very large size. The discovery of a complete 
skull of Deinotherium in 1836 in a sand pit near Eppelsheim south 
of Mainz enabled us to understand these animals more clearly. 
It was, however, through the skeletal structure that the Deinotheres 
were first compared with the trunk-bearing pachyderms, i. e., 
elephants and mastodons, to which they have the greatest number 
of resemblances. The most striking differences of the genus Deino- 
therium, as contrasted with other proboscideans, he in the structure 
of the skull and especially of the teeth ; above all in the structure 

of the lower jaw which is strongly elongated and sharply bent downwards, the great incisor tusks bending back- 
wards and not forwards. Their second distinction is in the structure of the grinding teeth in which two sharp 
transverse crests appear, except in the middle back tooth, M-'-M,, wliich is trilophodont. In the permanent 
dentition of the Deinotheres there are five grinding teeth in each jaw, namely, P -P3, P*-P4, M'-M,, M--M2, M^-Mg. 
in all twenty grinders. 




Fig. 63. Moimtod skeleton of the I'pper Miocene 
Deinotherium of Franzensbacl, Bohemia, in the Royal 
Museum of Natural History, Vienna, referred to Oeitm- 
Iherium bavaricum. .■Vfter photograph kindly furni.shed 
the present author hy Prof. Othenio .\l)el of Vienna. 



100 



OSBORN: THE PROBOSCIDEA 



The Franzensbad skeleton, attaining a length of 3.2 m. ( = 10 ft. 6 in.) and a height of 2.5 m. ( = 8 ft. 2]/^ in.), 
is referred to the smallest species Deinotherium havaricum. It was found in a cleft of calcareous rock near Franzens- 
bad in the year 1883. The bones belonging to it were partly complete, partly broken into numerous fragments, 
the assembling of wliich occupied a long time. The complete skull with a lower jaw, several of the large bones of 
the extremities, numerous vertebrae and foot bones, and a large number of ribs were modeled. For this purpose 
fossil bones and models were employed from the Vienna and other museums, also the skeleton of an elephant as an 
object of comparison. 

Referring to the stature of the species of Deinotherium: 

Deinotherium bavaricum (or D. cuvieri) includes animals of the smallest size. Middle Miocene, second Mediter- 
ranean layer. 

Deinotherium levius (or D. medium) includes animals of middle size, generally distributed; D. giganteum, 
animals of the larger size, Pliocene, Congerienschichten and Belvedereschotter. 

Comments on the Deinotherium bavaricum Skeleton of Franzensbad. Miocene 
ScHLESiNGER, 1923. — The original materials included in this mount are shown in figure 64B. The mount 
as it appears in the Vienna Museum of Natural History, with the iron-work removed, is shown in figure 64A. 




>0 Sair 




Fig. 64. Deinothkhium bavaricum of Franzensbad. Original and Restored Portions of Skeleton as mounted in the Vienna Museum 
From notes and observation.s kindly forwarded by Dr. Giinther Sehlesinger (letter, Vienna, December 17, 1923) a modified description and figure of the 
skeleton of Deinotherium bavaricum are presented herewith. 

A, Photograiih, side view of skeleton as mounted, with the iron uprights removed by retouching of the photographic plate. 

B, Sketch furnished by Doctor Sehlesinger showing the original parts of the skeleton preserved. 

(1) Original Parts.— Lower jaw without incisor teeth, atlas, axis, cervicals 2, 4, dorsals 5, 9, 10, lumbars 17, 18, 22, parts of sacrum, caudal vertebrse. 
Of left forelimb: Left scapula, left humerus without articular end, ulna, distal end of radius and upper proximal end, unciform, and metacarpals II, III. Of 
right foreUmb: .Scapula, ulna, radiale of the carpals, ulnare, trapezoid, magnum, metacarpals 1, III, IV. Of the hindlimb: Pelvis, femur of right and left 
side, left and right patellx-, right fil)ula, left calcaneum, right and left astragali, left navicular. (2) Parts Restored in Plaster.— Sacrum partly restored. 
Cervicals, dorsals, and himbars apparently too abbreviate. Skull and dorsal spines partly restored. (3) Errors in MouNTiNG.—Scapula and pelvis appar- 
ently too upright; possibly the restoration of the (extremities elongates them too much. 

Measurements by Phofes.sor Abel (letter, January 13, 1926).— Length of humerus 81.5 cm., of ulna SO cm., of femur 105.5 cm., and of mandible 
63 cm. 



THE DEINOTHERES 



101 



It was at the request of the present author that further information regarding the original materials and the 
restored portions of this skeleton was kindly sent by Dr. Giinther Schlesinger in a letter dated Vienna, December 
17, 1923, as foUows: 

Ihien Brief vom 15 November 1923 iiiit Dank bestatigend beeile ich mich, die verlangten Photographien zu iibermitteln. 
Durch die IJebon.s\viirdigl<cit unsoros Kollegen Prof. Schaffcr vom Naturliistorischon Museum ist es mir moglich, Ihnen zwei 
Aufnahmen in 3 Kopien zu ubermitteln [Fig. 64A, B]. . . . Mcine niilieren Angabeii iiljer die Verteilung von Originalstiicken 
und Erganzungen in Gips werden Sie allerdings iiber das Skelett sehr enttauschen. Wirkliche Fossilreste sind nur: Mandibei 
(ohne Incisoren), Atlas, Epistropheus, 2, 4. Halswirbel, 5, 9, 10-17, 18, & 22, Rumpfwirbei, das Sacrum und der Seliwanz, 
ferner von der linken Vorder-Extremitiit Scapula, Humerus (ohne Geienkskopf), Ulna, distalcs Ende und oberstes proximales 
Ende des Radius, Unciformc und Metacarpals 2 & 3. Von der rechten Vorder-Extremitat sind original: Scapula, Ulna, 
Radiale, Ulnare, Trapezoid, Magnum, und Metacarpal 1,3,&4. Von den Hinter-Extremitaten ist das Becken original ferner: 
Femur sin. et dext.. Patella sin. et dext., Fibula dext.. Calcaneus sin.. Astragalus sin. et dext., Naviculare sin. Alles andere ist 
Gips. 

Die Montierung des Skelettes halte ich mit Ihnen fiir zu hocli. Der Fehler durfte daher gekommen sein, dass das Sacrum 
nur zum Teil vorhanden ist, und die Ergiinzuiig der Wirl)elkorper anscheinend zu kurz au.sgefallcn ist. Auch scheinen mir 
Schulterblatt und Becken zu stcil zu sfehcn. Violleicht ist auch bci der (iipsergiinzung der Extremitatenknochen etwas zu 
viel an Lange hcrau-sgekommen. Allerdings diirfte dies kaum erhebliche Werte ausmachen. 

Abel, 1924. — At the request of the present author, Prof. Othenio Abel in a letter dated Vienna, January 8, 
1924, kindly supplemented in somewhat more detail the information conveyed by Doctor Schlesinger, accom- 
panying it with a new and more direct frontal photograph (see Fig. 63) : 

[1. Original and restored parts] Vor allem durfte es fiir Sie wichtig sein, zu erfahren, welche Knochen erhalten und 
welche ergjinzt sind. Ich erfuhr gestern, dass Dr. Schlesinger Ihnen bereits geschrieben hat, wovom er mich nicht verstandigt 
hat; so weiss ich nicht, ob er Ihnen diese Liste mitteilte. Ich habe sie heute Vormittag aufgenommen und zwar sind von diesem 
Skelette erhalten : 



Unterkiefer (ohne Incisivem) 


10 Brustwirbel (mit Neurapophyse) 


Atlas 


17 " ( nur Zentrum) 


Epistropheus 


18 


4 Halswirbel 


4 (vorletzter) Lendenwirbel (nur Zentrum) 


6 


Sacrum 


5 Brustwirbel (nur Zentrum) 
9 

Linke Korperseitc: 


1 bis 5 Caudahvirbel 


Rechte Korperseite: 


2 Rippc 


1 Rippe 


3 " 


3 " (nur Oberende) 


G " 


4 " 


7 " 


5 " 


8 " 


7 " (nur Oberende) 


17 " 


10 " 




11 " 




12 " 


Scapula 


Scapula 


Humerus 




Radius 


Raili\is (luir untere Epiphyse) 


Ulna 


Ulna 




Ulnare 




Radiale 


Unciforme 


Unciforme 




Magnum 


Metacarp. II, III 


Metacarp. I, 111, IV 


Becken (sehr fragmentarisch) 




Femur 


Femur 


Patella 


Patella 




Fibula 


Astragalus 


Astragalus 


Calcaneus 




Naviculare 






Fii{. t)5. Deinotherium bavaricum in the \'ienna 
-Museum. Model by Othoiiin .\bel (1932), about one- 
fiftieth natural size. 



102 OSBORN: THE PROBOSCIDEA 

[2] Dimensionen: Distanz vom Unterende des linken Unterkiefers (AngulusJ bis zum 

Vorderrande der Stosszahnalveole : 66 cm. [ = 2 ft. 2 in. approx.] 

Lange des linken Humerus: 75 cm. [ = 2 ft. 5'^ in. approx.] 

Lange des linken Femur: ca. 100 cm. [ = 3 ft. 3'^ in. approx.] 
[3] Was die Rekonstruktion des Skelettes anbetrifft, so scheint mir vor allem ein Fehler daiin zu liegen, dass die Richtung 
der Neurapophysenachsen facherformig divergiert, was unmoglich ist. Ferner steht das Sacrum viel zu hoch iiber der Stand- 
flache. Es muss tiefer herabgesenkt werden. Die Stellung ware etwa so wie auf meiner Rekonstruktion des Dinotherium 
giganteum, von der ich Ihnen eine Photographic beischliesse (vgl. ' Lebensbilder aus der Tierwelt derVorzeit,' Fig. 82, pag. 91). 
Sodann steht die linke Scapula ganz uni-ichtig, weil zu weit nach einwarts gedreht und ausserdem ist die Hohenlage zum 
Thorax uiu-ichtig angenommen, wie aus dem Vergleiche mit dem rezenten Elefanten hervorgeht. Wenn Sie meine Pause der 
Photographic mit dem Skelette eines Elefanten in den 'Recherches sur les Ossemens fossiles,' T. I, PL i, pag. 204, zur 
Deckung zu bringen versuchen, mit dem die Photographic des Dinotherium bavaricum in der Grosse gut iibereinstimmt, so 
werden Sie sclien, dass sich Dinotherium bavaricum in tiberraschender weise als in den allgcmeinen Proportionen durchaus wie 
ein rezenter Elefant verhalt und sich von Bunolophodon etc. unterscheidet. Daium hal^e ich auch in meiner Rekonstruktion 
des D. giganteum, das sich von D. bavaricum nur unwesentlich, abgesehen von der Grosse und einigen anderen unbedeutenderen 
Merkmalen unterscheidet, den lebenden Elefanten zum Vorbilde genommen. 

[4. Habits] Ich halte Dinotherium fiir einen Laubfresser, der seine Stosszahne zur Nahrungsaufnahme nur insoweit 
verwendete, als er mit ihrer Hilfe Baumiiste abbrach, wie ich dies in meiner Rekonstruktion andeutete. Ausserdem hat Dino- 
therium seine Stosszahne wohl auch als Waffen verwendet, hat aber damit kaum den Boden erreichen konnen, ebensowenig wie 
ein Elefant in der Stand stelhmg dies zu tun vermag (ich denke an den indischen Elefanten). Dinotherium ist selbst verstandlich 
etwas ganz anderes als Mastodon oder Bunolophodon hinsichtlich seines Korperbaues. Unter den fossilen Proboscidien der 
Tertiarzeit kommt es meines Erachtens dem Elefanten der Gegenwart am nachsten. 





Fig. 66 Fig. 67 

P"ig. 66. Restoration of the head of Deinolhcrium modified from Gregory and Oshorn (O.sborn, 
1910.346, p. 247), It will he observed that the probofeis is drawn much shorter than that in the 
restoration of Abel (Fig. 67); this short proboscis is probably erroneous. 

Fig. 67. Restoration of Deinolherium, after Othenio Abel, "Lebensbilder aus der Tierwelt der 
Vorzeit," 1922, p. 91, fig. 82, showing the animal with a hairy coating. It will he oh.served that the pro- 
boscis is drawn much longer than in the Osborn-Gregory restoration (Fig. 66) ; this is probably correct. 

1921-1924. RECONSTRUCTION OF SKULL OF DEINOTHERIUM GIGANTEUM 

British Museum. W. D. Matthew, September, 1920. — The skull of Deinotherium restored and figured by 
Kaup is complete and very little crushed. The occiput pitches forward almost horizontally, with a great nuchal 
pit. The upper tusks are certainly absent. The breadth and the overhang of the nasal region are mostly natural. 

Andrews, 1921.— The original skull purchased from Kaup by the British Museum was carefully described by 
Dr. Charles W. Andrews in the year 1921 (p. 532) as follows: 

The skull of Dinotherium is, in many respects, one of the most remarkable known. Although fundamentally its 
characters are clearly Proboscidean, nevertheless it differs widely from the skulls of the other members of the group and, 
indeed, in some respects from that of any other mammal. In the true Elephants and Mastodons the peculiar form of the skull 
is mainly due to the enormous development of cellular bone in the occipital region, increasing the area available for the 
attachment of the muscles necessary for the support of the heavy trunk and tusks. 



THE DEINOTHERES 103 

In Dinotherium, although the upper tusks are wanting, the trunk, judging from the large size of the nasal opening, must have 
been enormous, and the weight of the head was further increased by the great deflected mandibular symphysis with its large 
tusks. In this case, however, the area for the attachment of the supporting muscles was supplied by the widening out of the 
occipital surface, which was further increased laterally by the extension outwards of the squamosals. Little or no cellular bone 
seems to have been developed, the occipital surface above the post-tympanic flanges being nearly flat except for the depression 
for the nuchal ligament. This flattening of the occiput, combined with its forward inclination, must have made it possible 
for the animal to move its head up and down through a large arc, a movement perhaps connected with the use of the downwardly 
directed lower tusks. The great width of the proboscis, probably rendered possible by the absence of upper tusks, led to the 
widening out of the skull in the orbital region, producing the shelf-like projection of the maxillae above noticed. The great width 
of the glenoid surface for the mandible is a peculiarity for which it is difficult to account, unless it is correlated with the general 
widening of this region of the skull. 

The skeleton of Dinotherium is still very imperfectly known, but such bones as have been described show that the animal 
must have been quite Elephantine in structure and appearance except as to its head, the legs being pillar-like and the neck 
short. A femur probably associated with the skull above described measured 150 cm. in length. The numerous speculations 
as to the appearance and habits of Dinotherium have been summarized by de Blainville and StefSnescu in their works referred 
to above. Most writers seem to suppose that the animal was chiefly fluviatile and aquatic in its habits, but there appears to 
be no good reason for believing that it was more so than the Elephants. 

India. Palmer, 1924. — In the Memoirs of the Geological Survey of India there appears an account by 
R. W. Palmer of an incomplete skull of Deinotherium wMch was published after his death in October, 1922, and 
the paper of 1921 by Charles W. Andrews, cited above, and that of 1922 by C. Forster Cooper, cited below, were 
therefore not consulted (see Palmer, Nature, Oct. 25, 1924, p. 624) : 

The specimen described is an incomplete fragment of a skull showing the basal surface from the condyles to the pterygoid 
region. The anatomical features shown do not differ markedly from those described by the late Dr. Andrews for the celebrated 
skull of Dinotherixim giganteum in the British Museum, but the new specimen being in better condition, certain points are 
more clearly established. With regard to the question of the validity of the Indian species D. indicum and D. pentipotamiae, 
the author concludes that they cannot be upheld and that they are both to be referred to the European form D. giganteum. 
This agrees with the views already published by Forster Cooper which were the outcome of a study of material from Baluchistan. 

(Palmer, 1924.1) Found in the Lower Chinji Sandstone, Lower PUocene (fide Pilgrim, 1913). 



1911-1919. DISCOVERY OF DEINOTHERES IN AFRICA 

Deinotherium hobleyi is the first species of the genus Deinotherium described from the continent of Africa; 
no mention is made by Fourtau (1918) of the presence of Deinotherium in the Lower Miocene of Moghara, although 
the Moghara locality (Fourtau, op. cit., p. 91) is described as certainly the richest in northern Africa in the remains 
of mastodonts, containing the types of 'Mastodon' [=Rhynchotherium{?)]spenceri and oi 'Mastodon' [ = Trilophodon] 
angustidens var. libyca, as described by Fourtau. In fact, Fourtau remarks {op. cit., p. 91) : 

Mais il est un fait assez bizarre a constater, c'est 1 'absence totale de restes de Dinotherium, le compagnon habituel des 
mastodontes miocenes. II semblerait done que ces animaux ne sont pas d'origine africaine, et que les restes signal^s dans 
I'Ouganda et en Ethiopie [Footnote: 'E. Haug. — Traite de Geologic, tome II (2), p. 1727.'], a des horizons bien superieurs, 
proviendraient d'une migration posterieure a I'apparition des Dinotherium en Asie et en Europe. 

Sir Harry Johnston independently supports the African origin of the Deinotheres and in 1925 writes to Osborn 
as follows : 

Johnston (letter, March 16, 1925). — You assign to the Dinotheria a 'European and Asiatic' habitat, but unless I 
have been grossly misled I rather fancy that the discoveries of Hobley, Fraas and one or two other Germans investigating 
Upper Egypt have shown the Dinothere group of Proboscideans to have originated in Egypt or Equatorial East Africa. 
Hobley discovered remains of pygmy Dinotheria in the Miocene formations lying east of the Victoria Nyanza — some such 
year as 1908. When I was in Germany just prior to , the War I was shown specimens collected by German palaeontologists 



104 



OSBORN: THE PROBOSCIDEA 



from Stuttgart and Munich, which, though fragmentary, certainly seemed to point to a primitive Dinotherium. These, 
I understood, were discovered in Upper Egypt [i. e.,Tebessa]. They or I may have been wrong, however, and it is possible that 
Hobley 's East African Proboscideans may not have been Dinotheres. But I think their statements are worth your investigation. 

Andrews, 1911.— Seven years previous to Fourtau's paper and twenty-eight years after the pubHcation of 
Weinsheimer's history, Charles W. Andrews recorded (1911.1, p. 35, Abstract, Proc. Zool. Soc. London) the dis- 
covery near Karungu, British East Africa, of a species which he named Dinotherium hobleiji, after C. W. Hobley, 
Commissioner of Mines in British East Africa. 

(Andrc\v.s, op. cil., p. 35) : The specimens described were sent to the British Museum by Mr. C. W. Hobley, Commissioner of 
Mines for British East Africa. They included portions of the mandible with teeth, acalcaneum.and a patella of a small species 
of Dinotherium nearly allied to D. cuvieri, from the Lower and perhaps Middle Miocene beds of France. The new species, 
which he [Andrews] proposed to call Diyiotherium hobleyi, differed from D. cuvieri in several particulars— e. g., the inner anterior 
column of pm 3 was more distinctly developed, and the talon of Ms had a distinct tubercle on its inner side. 



LM3 








Amer. Mus. 27006 
LM3 




Fig. 68. Original type figure of Dinotherium hobleyi Andrews, 1911.2, 
PI. XLViii, figs. 1, la, one-third natural size. From near Karungu, British 
East Africa. 



Amer, Mus. 27005 
DEINOTHERIUM HOPWOODI. Type. 

Fig. 68a. Type third superior and inferior molars (casts Amer. Mus. 27006, 
27005) of Deinoiherium hopu'oodi sp. n()V.,from Olduvai, near the southeast shore 
of Lake Victoria, regarded as of Middle Pleistocene age. One-half natural size. 
Collected by Mr. Arthur T. Hopwood of the British Museum. Observe especially 
the subdivision of the summits of the crests into 14 to 16 conelets- Originals (Brit. 
Mus. M. 14118, M. 14119 respectively). See page 117 below for description. 



The same author pubhshed (1911.2, p. 943) a full description of Deinotherium hobleyi, from which the following 
has been extracted: 

. . . from the neighbourhood of Karungu on the east side of Lake Victoria Nyanza. Most of the specimens are indetermin- 
able fragments, probably picked up on the surface, but in addition to these there are some beautifiUly preserved teeth with a 
portion of the mandible (PI. xlviii, figs. 1, la) of a small species of Dinolherium; a small imperfect Proboscidean calcaneum 
(fig. 5), a patella, and some other fragments probably belong to the same animal. ... In size and, on the whole, in the pattern 



THE DEINOTHERES 105 

of the teeth this Dinotherium is very similar to D. cuvieri Kaup, which is from the lower and middle Miocene of France, being 
apparently especially characteristic of the Burdigalien horizon. Detailed comparison however shows some differences. . . . 
These differences, coupled with the remoteness of the localities in which the two forms are found, seem to justify the establish- 
ment of a new species for this East African animal and I propose that it shall be called Dinotherium hobleyi Andrews. 

Brives, 1919. — A. Drives notes in 1919 (p. 90) the discovery of a single tooth of Deinotherium in the sands 
of Djebel Kouif near T^bessa, which he refers to the Lower Miocene Deinotherium cuvieri and describes as follows: 

Le Djebel Kouif est un vaste plateau de 16 a 18 kilometres de tour d'une altitude moyenne de 1150 a 1200 "', qui est situe 
pres de la frontiere tunisienne a 27 kilometres au Nord-Est de Tebessa. . . . C'est dans ces sables qu'a et6 trouvee une belle 
dent de Dinotherium en parfait etat de conservation. C'est la premiere fois que cet animal monstrueux est signals dans I'Afrique 
du Nord et ce fait a une importance capitale, car il permet de fixer d'une maniere precise I'age des sables qui le renferme. . . . 
Les seuls debris organiques trouv^s dans les assises consistaient en fragments de troncs d'arbres silicifi6s depourvus d'^corce et de 
racines. . . . Au col de Beccaria, Ph. Thomas a recueilli et cite: Araucariorylon segyptiacum Krauss, Bambusiles Thomasi 
Fliche, Palmoxylon Cossoni Fliche, Ficoxylon cretaceum Schenk, Acacioxylon antiquum Schenk, Jordanta iunetana Fliche, Nicolia 
segyptiaca (?) Unger. . . . La presence du Dinotherium ne laisse plus aucun doute sur I'age mioc^ne de ces depots; il restei 
pr^ciser maintenant si ces assises sont du Tortonien ou du Pontien. ... La dent que je prfeente est identique comme dimen- 
sions a un exemplaire de Dinoth. Cumeri des faluns de Maine et Loire ... La dent du Kouif est une 2*^ molaire superieure; 
elle est presque carrf^e (5 centimetres Vi au collet), les cretes sont l^gerement us6es, les racines manquent. 

Leakey, 1932-1934. — To the north of Olduvai, about 300 km., is Karungu, the type locaUty of Deino^/ienum 
hobleyi (Fig. 56, 22), on the east shore of Lake Victoria. About 30 km. distant from this was found by Dr. L. S. B. 
Leakey, on the islands in the Kavirondo Gulf, a tooth of a Deinothere probably referable to the Pleistocene species 
D. hopwoodi, associated with remains of Homo kanamensis of Kanam, East Africa (see footnote, p. 85, of present 
Memoir). 

1922. REVISION OF THE DEINOTHERES OF INDIA 
FoRSTER Cooper, 1922. — In the Proceedings of the Zoological Society for the year 1922, C. Forster Cooper 
published the following observations on the Deinotheres discovered in the Lower Miocene Bugti beds of Baluchis- 
tan, also in higher geologic levels. 

Middle Pliocene, Perim Island, is the type geologic level of Falconer's Dinotherium indicum, 1845; its specific 
characters lie in the shape of the jaw, there being no difference from D. giganteum in the teeth excepting a greater 
thickness of the enamel. Falconer (1868, p. 415) apparently refers to the same species as D. Perimense, a name 
nowhere defined. 

Upper PHocene, Attock. Falconer mentioned (MS. labels) a smaller variety from Attock but did not 
definitely name it; this smaller variety was subsequently named Dinotherium pentapotamise by Lydekker in 1876, 
but in 1886 he united this species as a synonym with D. indicum, regarding the differences as individual variations. 

Middle Miocene, Salt Range, Sind, India, is the type geologic horizon of Dinotherium sindiense Lydekker, 
1880; regarded by Lydekker (1880, p. 196) as distinct from D. pentapotamise; the type (Lydekker, op. dt., PI. 
XXXI, fig. 4) regarded by Forster Cooper as too fragmentary to afford adequate evidence of the specific distinctions. 

Lower Miocene, Bugti beds, Bugti Hills, Baluchistan, is the type geologic level of Dinotherium ndricum 
Pilgrim, 1908. (Pilgrim, 1908, p. 157) : "It differs very markedly from the other known species." Subsequently, 
however (1912, p. 16), Pilgrim withdrew the name D. ndricum and substituted the name D. pentapotamise 
[indicum] var. gajense. 

Forster Cooper (1922, p. 621, 622) considers that the characters assigned by Falconer and Pilgrim, and the 
published figures, lie within the Umits of variation of D. giganteum of Darmstadt, e. g., Brit. Mus. collection No. 
M.3494. In this Cooper agrees with Weinsheimer (see p. 94 of the present Memoir). 



106 



OSBORN: THE PROBOSCIDEA 



BuGTi Beds (Forster Cooper, 1922, pp. 622-624). — The material of the present collection from the Bugti beds is, as 
is the common e.xperience, both scanty and fragmentary, and consists only of part of a lower jaw and some separate upper and 
lower teeth. The lower jaw (text-figs. 8 & 9) belongs to a small form, and in size and structure compares closely with D. hobleyi 
[Footnote: 'Andrews, P. Z. S. 1911.']. The third ridge of the last molar is not parallel to the front ones, but slopes at an angle 
similar to that found in D. hobleyi and D. Ixvius. Of the separate teeth, none shows any great difference from specimens of 
European forms with which it has been compared. An ujsper molar, probably a second, is here figured (text-fig. 10). There is 
nothing in it which can be construed as a 'longitudinal ridge,' nor is there anything in the nature of a 'cusp blocking the valley' 
which has not been found within the range of variation of European specimens. A third upper premolar (PI. iv, fig. 6) is also 
figured to show the separation of the inner cusps said to be characteristic of the Indian forms, but which, as has just been 
stated above, occurs also in D. giganleum. The remaining teeth in the collection are not distinguishable from the smaller forms 
from Europe, and call for no comment except for three specimens. 

Comparative measurements (pp. 625 and 626) of the lower and upper grinding teeth oi Deinotherium sp. and 
of Trilophodon angustidens from the Bugti beds of Baluchistan, from Chevilly (Lower Miocene), and from 
Sansan and Simorre (Middle Miocene) of France are given, and the following conclusions are reached. 

Summary. — The Proboscidea of the Bugti beds of Baluchistan include Trilophodon angustidens [ = cooperi 
Osb.] and a small Deinotherium. The Indian T. angustidens is a small animal, more primitive than the Middle 
Miocene forms of Sansan, for which the name T. angustidens palseindicus may be retained. 

The Deinotherium resembles the smaller European Deinotheres, 
also the D. hobleyi of Africa. It is doubtful in our present state of 
knowledge whether the Indian Deinotheres are separable from the 
European. 





DEINOTHERIUM GIGANTISSIMUM 3900mm.,12'Q>^"e 
HUNGARY 




DEINOTHERIUM GIGANTEUM 3000mm.,9'lO"e 
GERMANY 



All 1/12 nat. size 



Progressive Increase in Size of Mandible of Deinotherium 
Fig. 69. Outline to same scale, namely, ono-lwelfth natural .size, of: 
C, .\d\ilt lower jaw of Deinotherium (lignuti.ssimum as jiartly restored by Stefanescu 
after Stefanescu, 1899, Talj. i. 

B, Adult lower jaw of Deinotherium giganleum, after Kaup, 1835.1, Add Tali, i, fig. 5. 
A, Juvenile lower jaw of Deinotherium cuneri ref., after Lartet, 1859, PI. xiii', fig. 4. 




DEINOTHERIUM BAVARICUM 2500mm.,8'2y8" 

S. GERMANY 

Fig. 70. Restorations (1932) by Margret 
Flinsch, under the direction of Henry Fairfield 
Osborn. One-hundredth natural size. 

These are based on the known skeleton of 
Deinolherium bavaricum and the incomplete 
skeleton of D. giganlissimum. 



THE DEINOTHERES 
1908-1925. SPECIFIC REVISION OF THE DEINOTHERES OF EUROPE 



107 



Mayet, 19U8, pp. 199-206. — This valuable specific revision is the first to treat the 'mutations' and species of 
Deinotherium in ascending geologic order, beginning with the Lower Miocene Deinotherium cuvieri of the Sables de 
rOrleanais and continuing with the Middle Miocene 'D. bavaricum,' with the Upper Miocene D. levius, and with 
the Lower Pliocene D. giganteum, as follows: 



D. cuvieri 
Buidigalicn-Helv6(ien 

Dinothcrium do taillc rclativt!- 
mont petite, atteignaiit a 
peine la moifio de celle du 
D. giganteum. 



Superior Dentition 
(1) (joneavity of crests fac- 
ing posteriorly, M' three 
crested, M-, M^ two crested, 
crowns, with the exception of 
M', less broad than long. M', 
crests gently concave, trito- 
loph diminishing in breadth, 
total length 72 mm. M-, 
transverse crests incompletely 
reunited internally, tritoloph 
narrower than protoloph, 
length CO mm. M', length 
02 mm. 



Inferior Dentition 
(2) Concavity of crests 
turned forward, molars longer 
than broad. P3, protolophid 
thick and divided into two 
points, length 43 mm. P4, 
length 48 mm. Mi, atrophy 
and narrowing of tritolophid, 
length 60 mm. M2, thickening 
of cingulum representing the 
talon, length 59 mm. M3, 
talon thick, narrow, triangular, 
detached from the crown, re- 
flected backward, length 71 
mm. 



(3) Tusks somewhat short 
and obtuse. 



D. bavaricum 
(Burdigalien?) Helvetien 

Taillc scnsiblement plus 
grande que celle du D. 
Cuvieri, mais inoindre que 
celle du D. levius et surtout 
du D. giganteum. 

Superior Dentition 
(1) Same characters as in 
D. cuvieri. M', same char- 
acters as in D. cuvieri. M", 
same characters as in D. 
cuvieri, excepting length 72 
mm., breadth of protoloph 69 
mm., of metaloph 66 mm. 



Inferior Dentition 
(2) Same observations as in 
D. cuvieri. P3, protolophid 
thick and bifid. Mi, trito- 
lophid narrower than the 
anterior crests, length 73 mm. 
M2, talon greatly reduced, 
length 63 mm. Ms, talon 
triangular, contracted as in 
D. levius, well detached from 
the crown, and thrown back- 
ward, length 72 mm. 



D. levius 
Helvetien 

Taille un peu moindre que 
celle du D. giganteum. 



Superior Dentition 
(1) Same observations as in 
D. cuvieri. P^, narrow in- 
ternally and somewhat tri- 
angular, length 66 mm. P'', 
length 70 mm. M', marked 
abbreviation in breadth of 
tritoloph, other characters 
more pronounced, length 84 
mm. M-, metaloph narrower 
than protoloph, length 73 mm. 
M^, two lobes more unefiual 
than those of M-, metaloph 
smaller, length 77 mm. 



Inferior Dentition 
(2j Same observations as in 
D. cuvieri. P3, anterior lobe 
compressed into a single tri- 
angular point, etc., small pos- 
terior talon, length 55 mm. 
P4, proto- and metalophid 
gently concave anteriorly and 
somewhat united on outer side. 
Pre- and postcingula little de- 
vclojjed, length 70 mm. Mi, 
reduction of tritolophid in 
breadth, length 78 mm. M2, 
postcingulum smaller, more 
detached than in M2 of D. 
giganteum, length 73 mm. 
M;!, postcingulum narrow, tri- 
angular, detached, length 80 
nun. 

(3) Tusks shorter, thicker, 
less sharp than in D. gigan- 
teum. 

(4) Mental foramen placed 
more anteriorly below P3. 



D. giganteum 
Helv^tien-Pontien 

Tres grande taille. Dimen- 
sions des dents fort 
variables. 



Superior Dentition 

(1) Same observations as in 
D. cuvieri. P^, of subquadrate 
form, length 80 mm. P^ 
length 78 mm. M', three 
transverse crests, i. e., proto-, 
meta-, and tritoloph, gently 
concave posteriorly, tritoloph, 
less strong and less broad, 
length 95 mm. M-, two trans- 
verse crests, i. e., proto- and 
metaloph, and crenulated cin- 
gulum, length 91 mm. M^, 
proto- and metaloph approxi- 
mately equal, length 91 mm. 

Inferior Dentition 

(2) Same observations as in 
D. cuvieri. P3, anterior lobe 
simple, narrow, single pointed, 
length 63 mm. P4, two trans- 
verse crests, i. e., proto- and 
metalophid, length 70 mm. 
Ml, three transverse crests, 
subequal, tritoloijhid slightly 
more narrow than metalophid, 
length 104 mm. M2, two 
transverse crests, length 83 
mm. M3, talon nearly as 
broad as metalophid, crenu- 
lated, closely applied to base of 
metalophid, length 95 mm. 



(3) Inferior tusks long, point- 
ed at e.xtremity. 

(4) Mental foramen at level 
of P3 or at interval between 
P3 and P4. 




Fig. 3' 





Fi|,l. 





H' 




B' 



Lartet's Composite Figure of the Deciduous and Permanent Dentition of Deinotherium 
Fig. 71. Proportions of Deinotherium gignnteum, after Lartet, 1859, PI. xiii, figs. 1-4, reduced to one-third natural size. Lartet remarks (op. cil., pp. 
506, 507): "Lcs pieces en nature quiont .servi de modeic au dessinateur pour recomposcr, dans los figures 1 et 2 de cette planche, les series thcoriquesde pre- 
miere et de seconde dentition, sont de provenances diverges et probablement aussi d'especes distinctes; mais on s'est attacfie i rainener uniformdment toutes 
ces dents aux proportions du DinolhcHum giganteum, avec reduction k demi-grandeur de nature." Compare Palmer, 1924, PI. iii, Deinotherium sp. {?). 
Fig. 1, a, b, c, Superior milk premolars, Dp", Dp', Dp'' (upper row), 
a', 1/, c', Inferior milk premolars, Dpj, Dps, Dp4 (lower row). 
Fig. 2, A-E. Permanent dentition, A = P',B = F',C = M',D = M-,E = M^ (ujjper row) . 

A'-E', Permanent dentition, A' = P3, B' = P4, C' = Mi, D' = M2, E' = M3 (lower row). 
Fig. 3, Fragment of jaw, c = Dp4 being replaced by B = P4; C = M] [H.F.O.]. 
Fig. 4, Juvenile jaw showing below A and B = P3 and P4, D = M2, above a-c = Dp2, Dps, Dp4, C = Mi. 

Materials 
The above synthetic illustration contains materials from Lower Miocene to Lower Pliocene horizons, as follows: 
Fig. 1, a, h, = Deinotherium cuvieri, Lower Miocene of Pontlevoy (Loir-et-Cher). 

c, Middle Miocene of the sub-Pyr6n<Ses. 

a'. Upper Miocene of Simorre (Gers). 

b', Miocene of Gers. 

c', = Deinotherium bavaricum (?) ref.. Upper Miocene. 
Fig. 2, A = Deinotherium cuvieri (?), Miocene of Gers. 

E, M , = Deinotherium giganteum, Lower Pliocene of Eppclsheim. 
Fig. 3, Fragment of jaw from Ile-on-Dodon (Ilaute-Garonne), Middle Miocene. 
Fig. 4. Jaw, Ilc-en-Dodon (Haute-Garonne), Middle Miocene. 
For further details, see Lartet's minute description of Plate xni, 1859, pp. 506-510. 



108 



THE DEINOTHERES 
1908. Synonymy and Citations (fide Mayet, 1908) 



109 



LOWER MIOCENE 

Dinolherium Cuineri Kaup, 
1831 [1832] 
1812 Tapir gigantesque Cuvier 

1825 Tapir gigantesque Cuvier 

1831 [1832] Dinolherium Cuvieri Kaup 



MIDDLE TO UPPER MIOCENE 

Deinotherium bavaricum von Meyer, 
1831 

1831 Dinutherium medium et secon- 

darium Kaup 

1832 Dinolherium bavaricum von Meyer 
1850 Dinolherium intermedium de Blain- 

ville 
1859 Dinolherium bavaricum Lartet 



1850 Dinolherium Cuvieri do Blainville 

1858 [1859] Dinolherium Cuvieri Lartet 1875 Dinolherium bavaricum Bachmann 



UPPER MIOCENE TO LOWER PLIOCENE 

Deinotherium giganleum Kaup, 

1829 

ISGl Dinolherium levins Jourdan 



Deinotherium cuvieri (Observations of Mayet, 1908, pp. 199-206). — Remains very abundant in the 
Lower Miocene Orl6anais, especially at Chevilly, chiefly isolated bones and teeth, type mandible from Chevilly. 
Entire dentary series represented by remains from Chevilly, Bricy, Boulay, etc., species clearly distinct from the 
Middle Miocene D. bavaricum, the Lower Pliocene D. giganteum, or the Upper Miocene D. levins. 

Dinolherium cuvieri is the sole species found in the Sables de I'Orleanais; certain molars approach those of D. 
bavaricum, a form of much larger size. Characters of D. cuvieri notably different from those of D. giganleum or D. 
levius, the latter type coming from the Upper Miocene of Grive-St.-Alban, Rhone basin. Thus D. cuvieri appears 
suddenly at the beginning of the formation of the Sables de I'Orleanais, survives into the Faluns du Blesois, and 
then disappears, to be succeeded by D. bavaricum, an ascending mutation of larger size. 

Deinotherium bavaricum (Observations by Mayet, 1908, pp. 205, 206). — Von Meyer's type of Deino- 
therium bavaricum from Gmlind, Bavaria, is a Deinothere of medium size closely intermediate between the Lower 
Miocene D. cuvieri and the Lower PHocene D. giganteum. The dental characters of D. bavaricum resemble those 
of D. cuvieri more closely than those of D. giganteum. From the Sables de I'Orleanais, Chevilly, one molar only 
is mentioned [?] as representing D. bavaricum. Although D. bavaricum of the Upper Miocene approaches in 
size D. levius of the Lower Pliocene [Upper Miocene] the dental characters of D. levius show much stronger 
affinity to those of D. giganteum. Consequently the larger teeth of the Sables de I'Orleanais probably belong to 
large examples of D. cuvieri. 



Ascending Increase in Size [including D. gigantissimum and D. indicum] 



Anteroposterior Measurements 
Deinotherium giga ntissiin um 
" indicum 

" giganleum 

" levius 

" bavaricum 

" cuvieri 



80 
66 



78 
70 



M' M' M^ 



95 
84 

72 



91 

77 



91 
73 

72 

60 62 



63 
55 

43 



70 
70 

48 



Ml 

100 

104 

78 

73 

60 



M2 
111 
99 
83 
73 
63 
59 



M. 



95 
80 
72 
71 



M1.3 
348-357e 

282 
231 
208 
190 



Osborn, 1925: From the above observations of Mayet it becomes perfectly clear that Deinotherium, like all 
other large herbivorous quadrupeds, passed through a series of ascending mutations, witnessed (a) in every single 
character of the superior and inferior grinding teeth, also (6) in the steady increase in size. The net anteroposterior 
diameter of the three lower grinding teeth, obtained by adding the length in millimeters, and including a 
comparison with D. giganleum, is as above. 



3. RESUME OF THE CHIEF GENERIC AND SPECIFIC CHARACTERS 

1. Progressive Doubling in Size. — In the above table the measurements given by Mayet of Deinotherium 
cuvieri, D. havaricum, D. levius, and D. giganteum are amplified by the Falconer (1868, p. 407) and Lydekker 
(1880, p. 189) measurements of D. indicum and by the Stefanescu (1896 [1899], Tab. i, iii) measurements of D. 
gigantissimum; they demonstrate a progressive increase in size, from the Lower Miocene stage (D. cuvieri) to tlie 
PUocene stage {D. gigantissimum) . 

The second inferior molar, Ma, of D. gigantissimum ( = 111 mm.) is about 95% longer than the second inferior 
molar of D. cuvieri ( = 59 nmi.). The inferior molar series, M1.3, of D. gigantissimum ( = 348-357e mm.) is also 
about 95% greater than the inferior molar series of D. cuvieri ( = 190 mm.). 

Consequently, corresponding to the doubhng in size of the dental series, we may roughly estimate that the 
Mid-PUocene D. gigantissimum was an animal about double the size of the Lower Miocene D. cuvieri. 

These proportions shown in the lower jaws of Deinotherium cuvieri, D. giganteum, and D. gigantissimum 
(Fig. 69) are also expressed in the diagrammatic restorations (Fig. 72) in which D. gigantissimum is represented 
as nearly double the size of D. bavaricum, with D. giganteum as intermediate in its skeletal proportions. 

CM FEET 




BAVARICUM 



D qiGANTEUM 

Three Specific Stages in Deinotheriitm 



D, qiQANTISSIMUM 



Fig. 72. Proportionate restorations of the year 1930, in profile, of Middle to Upper Miocene and Lower to Middle Pliocene specific stages of Deinotherium, 
mth an estimated shoulder height of: 

Deinolherium bavaricum 8 ft. 2^2 in. 

Deinotherium giganteum 9 G 

Deinolherium gigantissimum 12 

2. Geologic Succession of Species. — The Eurasiatic distribution of the Deinotheres, as carefully 
worked out for the first time in the accompanying correlation table, displays the ascending geologic succession of the 
species in western and eastern Europe, in southern Asia, and in Africa. When we compare the small Deinotherium 
indicum gajense from the Lower Miocene of India with the D. pentapotamise [ = D. indicum] of the Upper 
Miocene of India, we observe that there is a closely corresponding progressive increase in size of the 
Deinotheres in all parts of Eurasia where they have been discovered. The approximate geologic parallel between 
the Deinotheres of western Europe and those of corresponding horizons in India is displayed in Table I (Geologic 
Correlation) and Table II (Specific Distribution) of the present chapter. 

3. Specific Characters in Succession. — As explained in the introductory paragraph of this chapter, 
"It is not possible for the author to treat the Deinotheres in the same critical or exhaustive manner as the species 
of proboscideans belonging to other famiUes are treated in the present Memoir; nor is it possible to give the type 
descriptions or reproductions of the type figures as in other chapters." The reader is referred to the original type 

110 



THE DEINOTHERES 



111 



descriptions and figures wherever published, as indicated in the "Order of Discovery and Description of Twenty- 
six Speciesof Deinotherium described or named, 1715-1935" (p. 84 above) and in the corresponding bibliographic 
references below to the works of Kaup, von Meyer, Eichwald, Koch, de Blainville, Falconer, Jourdan, Lydekker, 
Stefanescu, Pilgrim, Andrews, and Ehik, the only palaeontologists to our knowledge who have described type species 
and specimens of Deinotherium. See also Haug, Joleaud, and Leakey in Bil)liogra])hy. 

4. Dentition Relatively Non-progressive. — As compared with other proboscideans both mastodontoid 
and elephantoid the dentition of the Deinotheres is relatively non-progressive; the fundamental pattern of the 
grinding teeth was established extremely early in geologic time certainly during the unknown Oligocene stages, 
because in the Lower Miocene specific stages it is fully established, especially the number of ridges and the char- 
acter of the crests of the upper and lower true molars, M 1-M 3. 

Mayet (1908, pp. 199-206) in his valuable table of specific distinctions (translated in full above) pointed out 
that the chief progressive characters between Deinotherium cuvieri, D. bavaricum, D. levius, and D. giganteum 
consist in («) progressive increase in size, (6) detailed changes of proportion between the anterior and posterior 
crests, protoloph, metaloph, and tritoloph, (c) in the greater or lesser union of the crests externally and internally, 
(d) in the shape and size of the incisive tusks, and (e) in the relative position of the mental foramen of the jaw. 



Mid. Pliocene 
Low. Pliocene 

Up. Miocene 

Mid. Miocene 
Low. Miocene 


Africa 


France, 
Switzerland 


Germany 
Bavaria 


Rumania 

Russia 

Bohemia 


India 








D. gigantissimum 

Type 
D. giga?iteum ref. 


D. indicum Type 
D. perimense Falc. 
( = D. indicum) 

D. angustidens 
Type 




D. gig. majus 
D. gig. medium 
D. gig. minus 


D. giganteum Type 
{ = D. maximum T.,D. 
medium T., D. secun- 
darium T. (?), D. 
kmigii T.) 


D. uralense Type 






D. leHus Type 

D. irdermedium Type 


D. bavaricum Type 
D. minutum Type 


D. proavum Type 
D. bavaricum ref. 


D. pentapotamix 
Type 




D. giganteum{1) ref. 


D. bavaricum ref. 

D. secundariu?n (?) ref. 




D. sindiense Type 


D. cuvieri rcf. 
D. hobleyi 
Type 


D. cuvieri Tj-pe 






D. ndricum Type( = 
D. indicum gajense) 



Table I.— Approximate Geologic and Geographic Distribution of Twenty-three of the Type Species of Deinotheres, 
after the Observations Recorded in the Preceding P.\ges of the Pre.sent Chapter IV of this Memoir. 

Compare pages 84, 85, 114, and 115. 



112 



OSBORN: THE PROBOSCIDEA 



These specific changes, as pointed out by Mayet, are of minor grade as compared with the profound changes 
which occur in the grinding teeth of other proboscideans by the addition of crests, by progressive hypsodonty, etc. 

The conclusion is therefore justified that Deinotherium is relatively conservative in the structure of its dentition, 
while sharing with all other proboscideans progressive increase in size. 



4. CONCLUSIONS AS TO ORIGIN, PHYLOGENY, AND ADAPTIVE RADIATION 

Geologic Origin and Range. — Deinotherium appears simultaneously in the Lower Miocene of Africa in D. 
hobleyi type and D. cuvieri referred, which apparently are in a stage of evolution similar to that of D. cuvieri type 
of the Lower Miocene Sables de I'Orleanais of France, which, in turn, are similar in size to the D. indicum gajense 
type of the Lower Miocene Bugti horizon of Baluchistan. These three specific types, which agree approximately in 
size, require very close examination and comparison before it can be determined which is the most primitive. 

The evidence seems to point to an African origin of Deinotherium, since there is no indication of ancestral 
forms of this genus in any Eocene or Oligocene horizon of Eurasia at present known. 

Migration. — Well defended by its tusks, well nourished by its precociously lophodont grinding teeth, well 
transported by its long limbs, Deinotherium spread rapidly through northern Africa, southern France, and far 
eastward into the peninsula of Baluchistan and India. Its traveUng companions were the Trilophodon angustidens 
libycus of the Moghara desert, northern Africa, the Trilophodon angustidens of the Loire, Sables de I'Orleanais, and 
the Trilophodon cooperi Osb. of Dera Bugti, Baluchistan. 

While ranging in the same geographic region, Deinotherium was very divergent in habit and local habitat from 
Trilophodon, as indicated by the following adaptive and generic contrasts: 



Deinotherium 

Cranium and jaws abbreviate. Proportions 
brachy cephalic. A very elongate proboscis capable 
of reaching higher branches of trees also the ground. 

Superior tusks early aborted. Inferior tusks 
rounded, downcurved, recurved, aiding the probos- 
cis in the prehension of branches and leafage. 

Grinding teeth .sharply crested; Dp4 and Ml 
trilophodont, P4, M2, and M3 bilophodont; total 
number of complete transverse crests not exceed- 
ing seven; all grinders in use at same time; per- 
sistently brachyodont, never hypsodont; prob- 
ably resembling Tapirus in browsing habit on leaf- 
age. Superior grinders tending to form ectolophs, 
like those of primitive tapirs. 

Limbs elongate, increasingly elephantoid, raising 
body well off the ground. Feet reduced to three 
short functional digits in manus and pes. Dorso- 
lumbar vertebrse and trunk abbreviate. 



Trilophodon 

Cranium elongate; lower jaws excessively 
elongate. Limited development of the proboscis, 
probably a prehensile upper lip, compensated for by 
great elongation of the lower jaw. 

Superior tusks persistently progressive in size, 
downcurved, aiding the prehensile upper lip in the 
prehension of food. Inferior tusks horizontal, spatu- 
late, aiding the proboscis in the prehension of food, 
also employed in the digging and uprooting of plants. 

Grinders persistently bunomastodont in pattern ; 
transverse crests increasing to four or five, adapted 
to the crushing of food, tending to hypsodonty; 
finally only the posterior grinders, M'-Ma, in use in 
advanced age. Browsing and partly grazing habit 
on low leafage and uprooted plants. 

Limbs mastodontoid. Feet tetradactyl to pen- 
tadactyl; digits abbreviate. Trunk elongate; low 
bodied; pelvis broad; mastodontoid in proportions. 



The above contrasts of structure and habit prove that from the very first the Deinotherioidea were sharply 
divergent from the Mastodontoidea. While the geographic range is very similar, as far at least as southern Asia, 
Deinotherium has been found only once in the north (D. uralense), whereas Trilophodon was a bold northern 
migrant reaching the 40th parallel of the United States in Middle Miocene time. 



THE DEINOTHERES 113 

Consistent with tlieir divergence of structure is the fact that, while fossil remains of Deinotfierium are some- 
times found in the same horizon with fossil remains of Trilophodon, more frequently they are found separate. 

Probable Habits and Habitat. — There is little in the Umb, foot, trunk, or skeletal structure of Deino- 
therium, now that it is quite fully known from the wonderful Franzensbad specimen (Fig. 64), to support the 
original theory that this was a fluviatile or water-loving animal which frequented streams and used its lower tusks 
for the prehension of food from the banks. Yet during warm seasons it doubtless bathed (Fig. 73) in shallow 
waters, like the modern Indian elephant. 

Rather we incline to conclude that it was a forest-living animal, sub.sisting upon leafage and tree boughs, to 
which its sharply crested grinding teeth were adapted like those of the tapir and of the tree- or shrub-browsing types 
of rhinoceroses. The very powerful trunk, indicated in the Eppelsheim cranium (Figs. 57, 58, 59), was eminently 
adapted to the collecting of tree boughs and leafage, as in the case of the existing African and Indian elephants; 
the open supranarial space is enormous. The absence of large superior tusks, which are seen in all the Mastodonti- 
dae and Elephantidae, explains the relatively flattened form of the top of the cranium and the forwardly inchned 
occiput, in wide contrast to the elevated, hypsicephaHc and brachycephalic cranium, with vertically placed occiput, 
for the attachment of large neck muscles to counteract the strain of the greatly elongated tusks. 

Consequently Deinothermm was neither mastodontoid nor elephantoid in profile but relatively flattened and 
depressed. Tlie profile aspect of the skull shown in our restorations (Figs. 72, 73) is totally different from that of 
either Mastodon or Elephas, wliile the proboscis was quite as large as that of these animals, and the body height 
and length of limb approximated that of the largest Proboscidea, as also shown in the restoration (Fig. 70) in 
comparison with other proboscideans drawn to the same one-hundredth scale. 

5. GEOLOGIC DISTRIBUTION OF DEINOTHERIUM. CAUSES OF EXTINCTION 

Species of eight mastodont genera, namely, Trilophodon, Anancus, Rhynchotherium, Zygolophodon, Turicius 
Miomastodon, Serridentinus, and Tetralophodon, were contemporaneous with Deinotherium, as shown in Table 
II below (Geologic and Geograpliic Distribution of the Types of the Superfamily Deinotherioidea, as compared 
with the Types of the Contemporary Mastodontoidea and Stegodontoidea'). 

The student should turn immediately to Table II below, in which the geologic and geographic distribution 
of most of the types of these deinotherioid and mastodontoid species are carefully set forth, also to the geographic 
distribution charts sho\ving exactly where these types were found ; by this means the zoologic and geographic 
environment of the Deinotheres is clearly brought out. 

It will be observed that while the Deinotheres belong chiefly in competition with animals of the INIasto- 
dontoid period, they are now known to survive into the period of Elephantoid distribution. 

Aided by different local adaptive radiation, probably in the forested regions of southern Eurasia, the Deino- 
theres with their brachyodont, tapiroid grinding teeth were able to compete with the bunomastodont genera, 
Trilophodon, Tetralophodon, and Anancus. It is probable that they enjoyed an entirely different local geographic 
range from the purely zygolophodont genera, Zygolophodon and Turicius. 

It is not an unreasonable hypothesis of extinction that the brachyodont Deinotheres were unable to compete 
with the incoming Stegodonts with their very numerous transverse crests and tendency to hypsodonty. At all 
events, the climax of the series {Deinotherium gigantissimum) in southern Eurasia becomes extinct in Middle Plio- 
cene time, just prior to the appearance of numerous species of Stegodon. In Africa the Deinotheres survived. 

This hypothesis of the failure of the brachyodont, tapiroid types of Deinothere grinding teeth to compete 
with the subhyiwodont and hyp.sodont types of the rising superfamily Elephantoidea' will afford a parallel in the 
Proboscidea of what we observe in the ungulates all over the world, because in PUocene time hypsodont ungulates 
began to prevail over the brachyodont ungulates, as first observed in Kowalewsky's great Memoir of 1873. 

'Sec footnote 2 on page 22. 



- fa 

§ ? 

St 



10 



Z 

o 

Q 
O 

o 
a B 






o 

S 
s 



o 
S 

■s 



o 

« 

fa 
o 

a 
K 

H 
O 

g 

Q 



o 

Q 
O 



CO 



« g 2 



H 
Z 

o 



bo 

OS 

a .u 



<u 



(u e S 

e <^- 
s & 






z 



o a 

> 

u 

Z 03 -3 < 
W t- < b 



gg 





B 


O 


o 


O 


^ 




a 


>5 




II 


H 


,^-x 


z 


r^ 


w 


K 


rn 


< 


W 


'.■> 


« 


z 


Ph 



Oh 



O 



W o 



§1 

o e fl 



a 



o -5 

-a -c! 

tf.S. 

dj S V. 
3 ^ i2 

c» pa 



o 
-a 

Q 



4J 






s 



o o 
ei. OS 



Oi o 
C a; 



•S ^ SS 

fees 
-« .S e 

^ 03 -C 
I": I- 

t^ 1 



a 

ei3 

J3 

03 
P^ 

O 

P 

<p 
o. 
>> 

H 



o 
S 

— , CO 

o -c 



S3 

Pu 

o 
Q 



4) 



"i; 



03 



C o C 
c/2 



« 9 

-c .2 

t| 
s a 

O o3 
CO 



-T3 q 
^ a e: 

,^ -O oS a; 

|S a? 
^ .a -g g 

1) Ph -t: 
p^ -— ' a> ■ 

<u ft >^ ^ 
a >:>H II 

•§ :§ -5^ l2 

§1 « 






o 



se 






a. -c 



s 

e 

■^ CQ Q 



0> '(Si 



CI :S c 

c3 o o3 

'^ Si "^ 

' l-H '<>> t—i 



^« 

ii 

I J 

ll 
•It 

g ^ 
11 



.2J 

'3 
to 3 



o .S 



.2 

s 

o3 3 

■3 p:i 

'i ^ 

a -— 

cS oS 

o e 
■.3 2 
O ^ 



a 



73 

a 



P^ 



O 



O 



03 
c3 






■s 

s 
s 

I 

s 

-B 03 
S. 0) 

e 





>. 




i 


>) 






-o 




03 


TS 




!? 


-c 




;-< 


■a 




OS 




—t 


03 




•S 


S 




'"-^ 


& 




a 


03 




CU 


03 




l-H 


^ 




ft 


IH 




CO 


►— t 










i— ( 






,_^ 






03 


<D 




". "^ 


0) 




_g 


ft 




^ 


ft 






>) 




>» 




co^ 


H 




'^ 


H 




a 


<i5 




II 


'e? « 




CO 


s 




60 


a-8 

a S 




s 


J 




^ 


m s 


^ 


i 




g 

P3 


e 




OS 

a 


1 


s 







a 
PQ 


s 



1 


a. 




> 


s 






"SI 


> 


§» 


s 


.Si 


^ S> 


s 


00 


^ 




CO 


33 


^ 



0) <u 

ft ft 



a; 
ft 



03 <^j. 

ft" ' 



<^ 



1^ 



I '8 -« 

i 1 1 



03 « 

•^ 2 



1 
s 






ft 



Oi S 

s s 

S 3 



ft 







O o o o o 
S s S s S 

'S 'S 'S 'S ■« 
Q Q Q Q Q 

0) 

ft 
H 

CO 

s 

S 



s s 

•■Si ■<>» 



oj a, 

00 












o 



a. a. 



CO 

s 

s 

s 

-§1 
-I 
o .2 



13 

s 
c 

e 
o 



SI 



e 

o 



Ex &-( E^ &; 6-. 



^^ 



03 

a 
a 
K 



'a ^-- a 
£ .5 g 

Qj CO s 

Ph Ph 3 
._ 03 

^-a ^ 

t^ M a; 
a g S2 
°3 fe ^ 

a^l 

o 

« a i o 
vw a g E 

fa 'S <^ a 
£ ;« ^ -5 

S ft " 

U W ^f; r^ 

O. II 



a; II u 
.^^ ,-, rv 

O ^^ ft 

o a '^ bfl 

-Sat" 
^ ^ S ^ 

-s ^ ft <: 



.03 - 

l-H ^ K 

II -3 ^ -^ 

« g § 

'^^:-^ a 

O aj 

i3 3 O 

>>S a.2 
ei< So Q 



*cfi 
t» 

3 

g ^ 

iJ l-H 



03 

3 
a 
pa 



03 

OS 



+5 



a 



a. 



-« .9 
•fi O 












<f 






o 




^ 






>> 




H 


al 








CV-. 


i^"^ 




s 




rO 


H s 




,o 


s 


3 


c 






•»^ 


Si 


1 £ 




s 

Si 

e 


;3 


o 


1-1 






ft. 












S ft 












s 


ft 


5 


.3 S 




s 


s 


_P- 














q; 




-c 


"5 








•♦■a 


o 


'o "S 




"S 


O 


iS c 


^' 


::^ 


ft 


_tf 






•rO 


C^J 


ss 






Q 



a 

y ^ 
o £ 

« ^ 
o CO 

w > 

go 



'3 ^ 

o 

M 



O. 



■= M^ 



„r"tf^^ iw 



s 



o a 

C3 o 
:c!3 II 






03 g 
3 II 

.2. iT 



CO 

-a 
_ c 

M <^ 
b£ CO 

S s 
a 2 



1-5 c« 



^ s .. 

0^ ^- § '^ « - ••^ 
-.§ B 

ti S I' 



S o 



m O 

i2 hJ 






a 



H ^ o 

03 



O 



a 
IE 

O 



§2 



s 






-a 

oj 

c 

03 






a S 



:« "s 






a; 

n. 
a t-i 

>> to 

it 






CO 

S 

s 



o 
^ (a 



S s 

O O 

. 03 

s i 

H CO 



c 
-a 

O 






II T3 

'? ^ 

. 2 w 

CO o 

.• i aJ" 

vS ■§ j3 

" "" 2 

;-. o3 

X5 







1 




o; 








GO 




O. 




? 


.5 


Q 




^ 




03 


■5 


<A 








JS 


a 


C 




CO 




SP 


»-H 


03 








o 


•^ 


^ 




o 




i«s 


■a 


tri 




(>» 




S 


o 


03 




"« 




— 


'm 


hJ 




ft 




a 


*^-^ 




J 




^ 


0) 


a; 




e 




a 


D. 




a. 






,>> 


>~> 








J- 


H H 




ft 


_03 


CO 

S 


e 

1 


CO 

■fS> 

s 

o 


-d 




'5" 

3 


■<s> 


S 

o 


c 




!s 


s 


CC 


»— 1 


5 




1 


S 




S 


s 

03 


"S 




1 


m 




^ 


•« 


-c 


» 


r< 


tl 




o 


s. 


-♦-^ 


a. 


o3 


Si 


ft 


o 




o 


n 


0^ 


•s 




■*^ 


^ 




Q t. 




E-, 





o 

a 

CO 



a 
o 
-a 
o 



fc 



J3 03 

o -a 

c3 g 



ZJ QJ 



03 

a 

<p to 



03 

bC 

O 



1=3 CO 






a 




116 



OSBORN: THE PROBOSCIDEA 




Fig. 73. Restor.vtion of Deinotherium gigantissimdm SxEKANEscn of the ManzatI Valley, Rumania 
By Margret Funsch (1933), under the direction of Henry F.airfield Osbohn 

The gigantic Deinotheres represented in this drawing are based on the dimensions of the type lower jaw and five grinding teeth 
described in the year 1899 by GregoriQ Stefinescu of the University of Bucharest, discovered in the Manzafl Valley, central Rumania. 
As stated on pages 95 to 98, the skeletal material described and figured is not from the type locality; the tibia and pes are figured on 
page 97 (Fig. 62). The restoration scene is in the MdnzaJI Valley during the Middle Pliocene period when these animals reached their 
climax in size, to which the specific name gigantissimum appropriately refers. Warm midsummer sea.son. 

The limb proportions of the skeleton are certainly mastodontoid rather than elephantoid; they are based on proportions of the 
Deinotherium bavaricum of Franzensbad (Fig. 63, p. 99). The three different feeding poses and attitudes represent a theoretic interpreta- 
tion of the adaptations of the lower tusks as well as of the proboscis in these remarkable animals. The tusks were used along river banks for 
uprooting function; also in combat between the males and in warding off enemies, while the proboscis was used in the conventional probo- 
scidean way in reaching for and drawing down foliage. 



APPENDIX OF CHAPTER IV 
Deinotherium hungaricum Ehik, 1930 
Figure 74 



Type: Kotyhdza (Dep. N6grdd), Hungary. Lower Miocene (?), Burdi- 
galian age. Paratype: Kiriild (Dep. Borsod), Hungary. 

Prodinotherium hungaricum £hik, 1930. " Prodinotherium 
hungaricum N.G., N.Sp." with an Apijendix by T. Szalai "On the 
Geological Occurrence of Prodinotherium hungaricum Ehik," 
Geol. Himgarica, Ser. Palaeont., Fasc. 6, pp. 3-21. Type. — 

Lower mandible containing I.P4-M3, r.Vi, M2, Ms; also fragment 
of scapula and limb fi-agment.s (humerus, ulna, radius, carpals, and 
portions of metacarpals). Locality and Horizon. — 

Kotyhaza (Dep. Nograd), Hungary. Embedded in a blue clay 
belonging to the Aquitanian (?) period. Paratype. — 

Molar and tusk fragments, from Kirald (Dep. Borsod), 
Hungary. Type Figure. — Op. ciL, PI. i, figs. 1-3, Pis. 11, 

III, and IV. Paratype Figures. — Op. cil., Pi. i, figs. 4-7. 

Specific Characters. — (Cf. fihik, p. 14) Third premolar. 
. . . "in the case of D. Cuvieri and D. Hobleyi the hypocoiiid and 
the entoconid [of Pnis] are connected by a posterior cingulum, while 
with thespechnen of Kirald the entoconid of pm31ies inside of the 



cingulum and is thus in no connection with it. This peculiar forma- 
tion of the pm3 distinguishes the Hungarian specimen from all 
Dinotheria hitherto described, so as from the larger species too." 
Mental foramen. . . . "in the ca.se of the Hungarian animal it 
[posterior mental foramen] is to be found in the vertical separating 
the pm3 from the pm4. The place of the anterior mental foramen 
is unknown with D. Hobleyi: in the case of D. Cuvieri it falls below 
the middle of pm3 ; in the Hungarian specimen it is situated below 
the anterior edge of the pm3. In consideration of all these cir- 
cumstances I give a new name viz: D. hungaricum to the specimen 
found at Kotyhaza and Kirald. . . . From the systematic point 
of view the mesatipody is a much more primitive character, than 
the systematically highly progressed dolichopody. This very 
interesting and important difference in the structure of the foot 
induces me to class the D. hungaricum into a new genus, under the 
genetic name of Prodinotherium." 

"The Prodinotherium hungaricum must be by all means one 
of the eldest Dinotheria hitherto known from an exactly determined 
geological age." 

Geology. — (Szalai, op. cit., p. 19): "On both find places of 
Prodinotherium hungaricum fihik, namely at Kotyhaza, in the 



THE DEINOTHERES 



117 



neighborhood of Salgotarjan, anil at Kiiald not very far from the 
former locality, the bed layers of the coal-measures are generally 
formed by a tough clay. Its material is sometimes of a swelling 
kind, sometimes of a sandy one, and its colour is bluish gray if wet, 
and greenish gi-ay when dry. Embetlded vegetal remains are to 
be found rather abundantlj' in this clay, among which Calamus 
. . . and Cinnamomum . . . are the most frecjucnt, whilst the 
Vertebrate fauna is represented by Mastodon aiiguslidens Cuv., 

. . . Aceralherium letradaclylum Lart Prodinolherium htin- 

garicum Ehik, . . . Testudo Fejervdryi n. sp. . . . and some 
Trionyx . . . remains." 

Osborn, 1932: The presence of 'Mastodon' august ideas in- 
dicates that these coal measures are of Burdigalian (Lower Mio- 
cene) age rather than of Aquitanian (Upper Oligocene) age; in 
fact, Szalai remarks (op. cil., p. 21) "The Prodinotherium remains 
described by Ehik are especially important from the stratigraphical 
point of view: for they are the only Dinotherian fo.ssils which un- 
doubtedly belong to the transition beds connecting the Upi)er 
Oligocene with the Lower Miocene, precisely this set of strata 
representing the Aquitanian. The biohistorical importance of 
such statement lies, moreover, in the fact that no Dinotherian 
remains are known, up to now, from strata older than those be- 
longing to the Aquitanian." 




Type of Deinotherium hungarichm 
Fig. 74. Tyix' mandible of Prodinntherium hungariciim fihik, 1930, I'l. 
u, figs. 1 unci 2, in oblique (uppor) and riglit lateral (lower) aspects. About 
onc-ei(thth natural size. 



In the opinion of the present author, the characters of the 
p:iratype premolar of Kirild (PI. i, fig. 7) serve to distinguish a 
new specifir rather than a new generic stage. Consequently we 
are inclined to consider Prodinotherium as a synonj'in of Deino- 
therium until further evidence is adduced to the contrary. 

DEINOTHERES RECENTLY DISCOVERED IN AFRICA' 
Deinotherium hopwoodi t^p. nov. 

KiKUre fiSa 
From Olduvai, near the southeast shore of Lake Victoria, TanKanyika 
Territory, Africa. Middle Pleistorene. 

The most surprising discovery of recent years is the survival 
of Deinotherium into the Middle Pleistocene of central Africa. 
The beautifully preserved type molars (Fig. 68a, p. 104) were 
regarded by Hop\\ood (letter, Feb. 5, 1934) as referable to the Lower 
Pliocene D. giganteum stage, from which they are quite distinct. 
For reasons given below they are made the type of Deinotherium 
hopwoodi. 

These grinders (Fig. 68a), an I.M3 (Brit. Mus. M. 14119, cast 
Amer. Mus. 27005), ap. 97 mm., tr. 87 mm., and an l.M^ (Brit. 
Mus. M. 14118, cast Amer. Mus. 27006), ap. 91 mm., tr. 96 mm., 
compare in size with tliose of the Lower Pliocene Deinotherium 
giganteum Kaup (l.Mj, ap. 95 mm., tr. 83 mm.; l.M', ap. 91 mm., 
tr. 93 mm.), while inferior to those of the Middle Pliocene D. 
gigantissimum (I.M3, ap. 114 mm., tr.?; l.M', ap. 114 mm., tr. 
117 mm.). 

In the same pit with the above tyjx; grinders were found 
three other molars," namely, third inferior premolar, l.P.i, ap. 
81 nmi., tr. 63 mm., index 78, fourth superior premolar, r.P', 
ap. 89 mm., tr. 90 mm., index 101, and finally a second su- 
Iierior molar, l.M'-', ap. 101 mm., tr. 100 mm., index 99. 

The teeth foimd (Fig. 56-|-) in the Lower Pleistocene of 
Abyssinia (Haug, 1911; Joleaud, 1928), referred to D. gigan- 
teum, probably belong to this species. 

The type locality (Fig. 56, 25), Olduvai, Tanganyika 
Territory, Lat. 2° 58' S, Long. 3° 24' E, is a most inaccessible 
semi-desert place, a day's march from the nearest good water 
sui)j)ly. Hopwood (letter, Feb. 5, 1934) reports that the type 
teeth were found in the low'est of the four horizons referable 
to the Middle Pleistocene. They were mixed up with a partial 
skeleton of a very young Elephas antiguus recki, and on the 
other side of the dry water-course were found otluv remains 
of the same elephant at the same level; also that teeth belong- 
ing to Deinotherium have been found in the Lower Pleistocene 
of Abyssinia (Haug, 1911, Joleaud, 1928), as well as in the 
Pleistocene of Kanam, Kenya CoK)ny. 

CoNELETs. — Beautifully preserved and fitting closely in 
api)ositiuii, as if belonging to a single individual, the.se tj'pc 
grinders display the conelets of which the crests were formed, 
namely: protolophid of I.M3 (16 conelets), metalophid of 
1.M.1 (14e conelets); protolojih of l.M' (14 conelets), metaloph 
of l.M' (12 conelets). On this character, which to Osborn's 
knowledge has not been mentioned before, also on the superior 
size, the .'species Dfinolhcrium hopwoodi has been ba.sed by 
Osborn, in recognition of Doctor Hoi)wood's never-failing 
cooperation with the author in tlie preparation of this Memoir, 
in addition to his distinguished work in the African field. 



'See also Leakey's discovery of DeinoUuriuin sp.?, of Pleistocene age, in central .\frica, Kavirondo Gulf (pp. 85 and 105 of the present Memoir). 
•The originals of these referred molars are in the British Museum, namely, Brit. Mus. M. 14116, M. 14115, and M. 14117, respectively. 




TRILOPHODON MACROGNATHUS 
INDIA 



325onimjo'8"e 




AMEBELODON FRICKI 3047 mm., 10' e 

NEBRASKA 




TETRALOPHODON GRANDINCISIVUS 2700mni.,8':0;4"e 
PERSIA HUNGARY 




OCALIENTINUS FLORIDANUS 27l4nim.,8'IO%"e 
FLORIDA 




SERBELODON BURN HAM I 2050iniH.,6'8y4"e 
CALIFORNIA 




PLATYBELODON GRANGERI 2135nim.,7'e 
MONGOLIA 




AYBELODON H0NDUREN5IS ZZOOmMJ 2M e 

HONDURAS 




CORDILLERION ANDIUM 2200inm,.7'2/2" e 

ECUADOR 




CUVIERONIUS SUPERBUS 3022 nim.,9'1 1' C 

ARGENTINA 




EUBELODON MORRILLI 2520mni.,8'5/6" 

NEBRASKA 




STEGOMASTODON ARIZONAE 26-'2IIim.,6'8/4" 
ARIZONA 




ANANCUS ARVERNENSIS 2550mm.,8'41'4" 
ITALY 




RHYNCHCTHERIUM TLASCALAE 2650m,ni,8'8>'4" e 
MEXICO 




SYNCONOLOPHUS DHOKPATHANENSIS £64-2 mm., 6' 8" 9 
INDIA 



Fig. 75. Fourteen Genera of the Suborder M.\stodontoidea up to the vear 1935 
Restorations to a one-hcndredth scale by Marqret Flinsch, under the direction op Henry Fairfield Osborn 



118 



Chapter V 

HISTORY OF THE CLASSIFICATION OF THE MASTODONTOIDEA, FAMILIES 

AND SUBFAMILIES, 1705-1927 

Gradual separation of the 'species" composing the Cuvierian genus Mastodon into two distinct 

FAMILIES, MaSTODONTID^ AND BuNOMASTODONTID^, AND INTO SEVEN DISTINCT SUBFAMILY PHYLA, MaSTO- 
DONTINvE, ZyGOLOPHODONTIN^, LoNGIROSTRINvE, SeRRIDENTINvE, RhYNCHOROSTRIN^, NOTOROSTRINyE, AND 

Brevirostrin^. 

1. Cuvier's five classic 'species' of Madodon represent four 5. The ZyRnlnphndontinse { = Mastodon borxoni): Final 

[five] subfamilies. separation (Russia, 1894) from the Mastodontinse 

2. Cuvier's original conception of his jjeiuis ' Maslodonte,' ( = Mastodon amcricanus) . 

1806. Types and definitions of five species. 6. Separation of species of 'Mastodon' and ' Elephas' 

3. The ZvRolophodontinie: Separation by Schinz and Vacek described in .\merica (1792-1874). 

from the Longirostrinie. 7. Osborn's classification (1927) of the superfamily Masto- 

4. The Breviro.strina: Separation by Lortet and Chantre dontoidea. 

from the Longirostrinie and Zygolophodontinie. 8. Cuvier and the Revolutions of the Globe. 

(January, 1934) Many discoveries since 1927 have altered or modified the statements made in this chapter. During the seven years 
(1927-1934) four additional subfamilies were discovered or defined, namely, the Amebelodontinse, the Platybelodontinse, the Tetralo- 
phodontinee, and the Humboldtinae subfam. nov.; also the families Serridentidse fam. nov. and Humboldtidae fam. nov. (1935) 
Three new subfamilies are added, namely, Palaeomastodontinae, Gnathabelodontinae, and Stegolophodontinae. The final classification, 
therefore, will be found in the Appendix at the close of the present Volume. 

Cuvier, 1806.— Cuvier, the founder of vertebrate palseontology, fell heir to all the collections in the Mu.s6um 
d'Histoire Naturelle of Paris and was the first to distinguish specifically the true mastodont form of molar without 
trefoils from the bunomastodont form of molar with trefoils. Cuvier was more impres.sed with the re.semblances 
between these two forms of molars than with the differences; consequently he included both forms within his 
single genus 'Masiodonte' { = Mastodon), which he divided into five classic species that became the starting point 
of all future descriptions and which survive to the present day; these species also are subdivisible according to 
modern discovery, as below; he did not adopt the binomial system of Linnaeus until 1817. 

Osborn, 1927.— We now know that these five Cuvierian 'species' belong in two' distinct families, namely, the 
Mastodontidae and the Bunomastodontidse, and in four' distinct subfamilies, namely, the Mastodontinae (including 
Mastodon giganteum Cuv.), the Zygolophodontinae (including M. tapiroides Cuv.), the Longirostrinae (including 
M. angustidens Cuv.), and the Notorostrinse (including M. andium Cuv. and M. humboldtir Cuv.). 

It has taken more than a century of research, concluding with the present Memoir, to discover that^t^e entirely 
different lines of descent actually separated these five Cuvierian 'species' in geologic time and divided them up 
among five generic and four' subfamily phyla, as follows: 

Species of Cuvier Genera, Subfamilies, Families, and Superfamilies 

OF THE Present Memoir 

Mastodon giganteum Cuv. . .Mastodon americanus Subfam. Mastodontinae Fam. Mastodontidae Superfam. Mastodontoidea 

Mastodon angustidens Cuv . .Trilophodon angustidens... " Longirostrinae.... " Bunomastodontidse 

Mastodon tapiroides Cuv Turicius tapiroides " Zygolophodontinae " Mastodontidae 

Mastodon andium Cuv Cordillerion andium " Notorostrinae " Bunomastodontidse 

Mastodon humboldtii CvLV .. .Cuvieronius humboldtiir .. . " Notorostrinae " Bunomastodontidse 

'(Osborn, 1934) These Cuvierian 'species' now belong in three families and five subfamilies through the addition of the Humboldtidffi fam. nov. and the 
Humboldtina; subfam. nov. 

'(Osborn, 1934) now referred to th<' lluinhoUltinu' subfam. nov., of the Humboldtida; fam. nov. 

119 



1. CUVIER'S FIVE CLASSIC 'SPECIES' OF MASTODON REPRESENT FOUR' SUBFAMILIES 

It was characteristic of Cuvier that he was very reluctant to admit the existence of more than one genus 
of either mastodonts or elephants, namely, Mastodon aiidi Elephas; or to admit the existence of more kinds of 
fossil proboscideans than he liimself had seen or described, This reluctant spirit has been manifested by many 
of his followers down to the present time. 

Subfamily Mastodontin^ Osborn, 1910, of the Present Memoir 
See Chapter VI for detaOed descriptions 

Cuvier was not in the habit of clearly designating his type specimens; consequently we select for the types 
of the Mastodon the teeth from the Ohio River first discovered, described, and figured as the types of Masto- 
donte de I'Ohio [ = Mastodon americanus Kerr — Fig. 76]. 




^ M.ts rujj(t.\ tjl . PL I. 

Fig. 76. Cuvier's types of Mastodonte de I'Ohio or Le Grand Mastodonte, 1S06.2, 
PI. 49 [i]. [ = Mastodon americanus Kerr of the present Memoir.] See figure 112. 

Subfamily Zygolophodontin^ Osborn, 1923, of the Present Memoir 
See Chapter VII for detailed descriptions 

Petit Mastodonte { = Mastodon [Turicius] tapiroides Cuvier, 1806 — see Figs. 77, 160, and 161), from the 
Lower Miocene of the Calcaire de Montabusard, with lophoid crests, without trefoils, has 
become the prototype of the zygolophodont phylum of the Miocene, including M. turicensis 
Schinz, M. affinis Jourdan, M. borsoni Hays, as shown in the conspectus of Mayet (1909, 
p. 45, B). This phylum is treated by Schlesinger (1917, pp. 146-162) under the heading 
"Mastodon (Zygolophodon) tapiroides Cuvier." Mastodon tapiroides Cuv. is treated by 
Osborn in the present Memoir as Turicius tapiroides. The reasons for the separation by 
Osborn in 1926 (Osborn, 1926.706) of the genus Turicius from the genus Zxjgolophodon are 
given in Chapter VII, pp. 198-202, of the present Memoir. See also Pis. ii and in, between 
pp. 134-135. 




Fig. 77. Cuvier's 
type of Pelil masto- 
donte, 1806, from 
M on tab usard = Turici- 
us tapiroides. 



'(Osborn, 1934) Five subfamilies (see footnote on page 1 19). 



120 



CUVIER'S FIVE CLASSIC SPECIES 



121 



Subfamily Longirostrin.e Osborn, 1918, of the Present Memoir 
See Chapter VIII for detailed description (p. 252) and new figure (Fig. 299) of Cuvier's type 
Cuvier's type (Figs. 78, 190, 299) from Simorre {Mastodonte a dents etroites = Mastodon [Trilophodon] 
angustidens Cuv.), with narrow molars and single trefoils, appears in the modern 
literature (e. g., Schlesinger, 1917) as a collective species of all the Miocene Lon- 
girostrincs of Europe; it includes, as synonyms or trinomials, Mastodon minutus 
Cuv., 1824, of Saxony, M. simorrense Lartot, 1851, M. (janjaci Lartet, 1851, and 
M. cuvieri Pomel, 1854 [1848], as revised and listed by Mayet (Mam. Fos., 
1909, p. 44, A). This phylum is treated most fully by Schlesinger (Denk. Natur- 
hist. Hofmus., I, 1917, pp. 5-63), under the heading "Mastodon (Bunolophodon) 
angustidens Cuv." A new figure of the type has been prepared (see Chap. VIII, 
Fig. 299) from a cast presented to the American Museum in 1927 by the British 
Museum. This cast was one of two sent to Gideon Mantell by Cuvier, both 
of which were acquired by the British Museum with the purchase of the Mantell Fig. 78. Cuvier's type of Masmdnnic d 
Collection about the middle of the last century. ^f^ f ™'"^*' '^"f ■ = TriiopMon angusU- 

•' aervs of trie present Memoir]. See new figure 

from cast (Fig. 299) in Chapter VIII. 




Subfamily Notorostrin^ Osborn, 1921, [and HumboldtinyE subfam. nov.] of the Present Memoir 

See Chapter XII for detailed descriptions 

Regarding Cuvier's South American types: (1) It is important to note that the type (Fig. 79) of Mastodonte 
des Cordilieres [ = Cordillerion andium of the present Memoir], distinguished by single trefoils, is from near a 




NOTOROSTRIN* 

Fig. 79. Cuvier's type of .\fn.ilii(l(>iilc lirs 
Cordilikres, 1806 I = Cordillerion andium of the 
present Memoir). 




Hl'MBOLDTlN.E 

Fig. 80. Cuvier's type 
of Mastodonte humlxitilicn, 
1 806 [= Cuvieroniuahumboldtii 
of the present Memoir). 



volcano in the vicinity of Quito, Ecuador; (2) the type (Fig. 80) of M. humholdien, distinguished by double 
trefoils (now referred to a distinct genus, Cttvieronius, in honor of Cuvier) , is from near Concepcion, Chile. Further 
characterization of these South American species is given in the phylogenetic section below (Chap. XII). 



2. CUVIER'S ORIGINAL CONCEPTION OF HIS GENUS 'MASTODONTE,' 1806 

(January, 1935) Written by Osborn in the year 1927, Sections 4-7 are now somewhat out of date (see Appendix at close of the present Volume). 

Cuvier conceived his genus Mastodonte as embracing animals possessing the five different types of grinding 
teeth shown in his five type figures above; this is a 'collective' genus in the same sense that most of the original 
genera of Linnaeus were collective, that is, embracing several 'species' and 'genera' in the modern usage of these 
terms. 

According to all authors, Cuvier's first references to his genus Mastodonte are in the Annales du Museum 
d'Histoire Naturelle of 1806 (1806.2, p. 270). In the same year, in a subsequent paper, he observed (1806.3, p. 
413) : "Ainsi le genre se trouvera compose de cinq especes, toutes egalement inconnues aujourd'hui sur la terre." 

His original descriptions, references, and figures of these five species are as follows (Cuvier, 1806.2, pp. 270, 
293; 1806.3, p. 412): 



" Je nommerai done la grande espece [op. 
cit., pp. 412, 413], 
Mastodonte de VOhio; 



Celle de Simorre et d'ailleurs, 
Mastodonte a dents etroites; 



Celle des petites dents, 
Petit mastodonte'^ 
[de Montabusard] 

[de Saxe] 



La grande a dents carries, 

Mastodonte des Cordilieres 
[presdu volcand'Imbaburra, 
Quito, Ecuador] 

Et la plus petite, 

Mastodonte humholdien 

[de la Conception du Chili]" 



Original and Present Reference 



[Mastodon giganteum Cuv. = Mastodon 
americanus] 



[Mastodon angustidens Cuv. = Trilopho- 
don angustidens] 



[Mastodon tapiroides Cuv. 
tapiroides] 



■ Turicius 



[Mastodon minidus aut. minor = (?) Tri- 
lophodon angustidens minutus] 



[Mastodon andium Cuv., 1824,' 
= Mastodon cordillerarum Desm. 
1822, = Cordillerion andium] 



1820- 



[Mastodon humboldtii Cuvier in Des- 
marest, 1818, p. 447 =Cuvieronius 
humboldtii] 



Ann. Mus., VIII, p. 293, PL 49 [i]. Figs. 
1-5; also Pis. 50 [ii] to 56 [viii], fide 
de Blainville. (Tj^pe confirmed by de 
Blainville in his revision of 1839-64, 
p. 245.) 

Ann. Mus., VIII, p. 405, PI. 66 [i], 
Fig. 4; PI. 67 [ii], Figs. 2, 4, 6, 11, 
and 13 (fide de Blainville). (Type 
confirmed by de Blainville in his 
revision of 1839-64, p. 246.) 

Ann. Mus., VIII, p. 411, PI. 68 [in], 
Fig. 6. (Type confirmed by de 
Blainville in his revision of 1839-64, 
p. 251.) 

Ann. Mus., VIII, p. 411, PI. 67 [ii], 
Fig. 11. (Type confirmed by de 
Blainville in his revision of 1839-64, 
p. 250.) 

Ann. Mus., VIII, pp. 411, 413, PI. 67 
[ii]. Fig. 1. (Type confirmed by de 
Blainville in his revision of 1839-64, 
p. 249.) 

Ann. Mus., VIII, pp. 412, 413, PI. 67 
[ii]. Fig. 5. (Type confirmed by de 
Blainville in his revision of 1839-64, 



p. 249.) 
Le Grand Mastodonte Cuv. (Cuvier, 1806.2, Ann. IVIus., VIII, pp. 270, 293) or Mastodonte de VOhio Cuv. 
(Cuvier, 1806.3, p. 412) = iWastodon giganteum Cuv., 1817 ("Le Regne Animal," p. 233). Type.— Cuvier, 

1806.2, PI. 49 [i], figs. 1-5; confirmed by de Blainville (1839-1864, p. 245). Type Description.— Cuvier, 

1806.2, pp. 270, 273, 293; 1806.3, p. 412. 

= Mastodon americanus of the present Memoir. 

"Mastodonte a dents etroites" Cuv., 1806 (Ann. Mus., VIII, 1806, pp. 405, 4^12) = Mastodon angustidens Cuv., 
1817 ("Le Regne Animal," p. 233). Type.— Cuvier, 1806.3, PI. 66 [i], fig. 4; confirmed by de Blainville 

'In 1806 Cuvier regarded the little teeth of MontaV)Usard { = M. tapirmde Cuv., 1821) and of Saxony { = M. minuliis Cuv., 1824) as belonging to the 
same species, namely, Pelil mastodonte. See Cuvier, 1806.3, p. 411: "La dent de Montabusard, pi. iii, fig. 6, correspond si bien k celle de Saxe pour sa 
largeur, que je ne doute pas que ce ne soit un gernie de I'une des post(5rieurps de la meme esp&ce, cass(5 en avant." 

The specific names M. andium Cuv. and M. cordillerarum Desm. refer to the same type description (Cuvier, 1806.3) of the Mastodonte des Cor- 
dili^es. We may follow de Blainville and Falconer in choosing the specific name andium. 

122 



CUVIER'S FIVE CLASSIC SPECIES 123 

("Osteographie des Mammiferes," 1839-1864, p. 246). Type Description.— (Cuvier, 1806.3, p. 405): "Je 

commence par une dent de Simorre, pi. i, fig. 4." (Detailed description and measurements of this tooth and com- 
parison with other teeth from Simorre also of M. giganteum, pp. 401 to 420). 

= Trilophodon angustidens of the present Memoir. 

Petit mastodonte, celle des petites dents, Cuvier, 1806.3, p. il3 = Mastodonte tapiro'ide Cuv., 1821 (Cuvier, 
1821-1824, Vol. I, p. 268), subsequently written Mastodon tapiroides (see Desmarest, 1820-1822, p. 386, also 
Cuvier, 1821-1824, Vol. V, Pt. 2, p. 527). Type.— Cuvier, 1806.3, PI. 68 [in], fig. 6; confirmed by de Blain- 

ville (1839-1864, p. 251). Type Locality. — Calcaire de Montabusard. (See below Petit mastodonte 

= Mastodon mimdiis.) Type Description. — (Cuvier, 1806.3, p. 411): "La dent de Montabusard, pi. iii, 

fig. 6, correspond si bien a celle de Saxe pour sa largeur, que je ne doute pas que ce ne soit un germe de I'une 
des posterieures de la meme espece, casse en avant." 

(de Montabusard, L. Miocene) = Turicius tapiroides of the present Memoir, 

Petit mastodonte, celle des petites dents, Cuvier, 1806.3, p. 4:13 = Mastodon minutus Cuv., 1824 (Cuvier, 1821- 

1824, Vol. V, Pt. 2, p. 527). Type.— Cuvier, 1806.3, PI. 67 [ii], fig. 11; confirmed by de Blainville (1839- 

1864, p. 250). Type Locality. — ^Saxony. (See above Petit mastodonte = Mastodon tapiroides.) Type 

Description. — (Cuvier, 1806.3, p. 411): "Telle est la dent de Saxe, envoyee autrefois par le professeur de Got- 

tingue, Hugo, a Bernard de Jussieu, et que I'illustre neveu de celui-ci a bien voulu me communiquer, PI. in [n, 67], 

fig. 11, entierement semblable en figure et en proportions a celle de la fig. 4, pl. i. EUe est exactement d'un tiers 

moindre." 

(de Saxe, Miocene) = Trilophodon angustidens minutus of the present Memoir. 

Mastodonte des Cordilieres, la grande a dents carrees, Cuvier, 1806.3, p. 413 = Mastodon Andium Cuv., 
1824 (Cuvier, 1821-1824, Vol. V, Pt. 2, p. 527) , and M. cordillerarum Desmarest, 1820-1822, p. 385. Type.— 

Cuvier, 1806.3, PI. 67 [n], fig. 1; confirmed by de Blainville (1839-1864, p. 249). Type Locality.— 

Volcano near Quito, Ecuador. Type Description. — (Cuvier, 1806.3, p. 411): "Les plus grandes ont les 

memes dimensions que leurs correspondantes de I'Ohio. M. de Humhold en a rapports une qu'il a trouvee pres du 
volcan d'Imbaburra, au royaume de Quito, a 1200 toises de hauteur. EUe est assez decomposee et encore enduite 
de cendres volcaniques. Son email est teint en roussatre; elle est longue de0,12, et large de 0,085. Voyez pl. 
II [67], fig. 1." (De Blainville, 1839-1864, p. 249) : "Cette espece a et6 proposee, pour la premiere fois, par M. G. 
Cuvier, en 1806, dans son memoire sur differentes dents de Mastodontes. Ann. du Mus., tome VIII, p. 411, et 
dans les m^moires reunis tome I, p. 11 [tome II, 1812.1, p. 11], et enfin dans la seconde edition de ses Recherches 
sur les Ossements fossiles de Quadrupedes (tome I, p. 266, 1821 a 1825). Elle ne reposait et ne repose encore que 
sur un petit nombre de pieces, trois dents molaires. {Loc. cit., pl. 2, fig. 1 et fig. 12.)" 

= Cordillerion andium of the present Memoir. 

Mastodonte humboldien, la plus petite, Cuvier, 1806.3, p. 413 = Mastodon Humboldtii Cuv. (see Desmarest, 
1818, p. 447); M. humboldii Cuvier, 1821-1824, Vol. V, Pt. 2, p. 527). Type.— Cuvier, 1806.3, Pl. 

67 [n], fig. 5; confirmed by de Blainville (1839-1864, p. 249). Type Locality.— Near Concepcion, 

Chile. Type Description.— (Cuvier, 1806.3, p. 412) : "M. de Humbold est encore celui qui les a decouvertes. 

Je lui en dois une qu'il a rapportee de la Conception du Chili; elle est fort usee, mais bien conserv6e, teinte 
en noir, longue de 0,08, et large de 0,06. Voyez pl. ii [67], fig. 5." 

= Cuvieronius humboldtii of the present Memoir. 



3. THE ZYGOLOPHODONTIN.E: SEPARATION BY SCHINZ AND VACEK FROM THE LONGIROSTRINiE 

(1) One of the earliest revisions of the mastodonts, in the broad sense, following Cuvier, is that of Desmarest 
(1820-1822) in which all of Cuvier 's five species are reviewed and the name Mastodon cordiller arum appMed to 
Cuvier 's Mastodonte des Cordilieres. In this revision six species of 'Mastodon' are recognized, namely, 
Mastodon giganteum, M. angustidens, M. cordillerarum Desm. ( = M. andium Cuv.), M. Humboldtii, M. minus 
( = M. angustidens), and M. tapiroides (pp. 384-386). M. minus { = minutus) is a synonym of M. angustidens. 

(2) The next revision is that of Schinz (1824, p. 278), in which five of the same species are recognized and 
the zygolophodont species 'Mastodon turicense' from the lignites of Elgg, Canton Zurich, is added. 

(3) There followed successive revisions in which it was recognized that M. tapiroides Cuv. and M. turicensis 
Schinz, animals possessing lophodont grinders without trefoils ( = Zygolophodontinse) , should be separated off 
from animals possessing bunomastodont grinders with trefoils ( = Bunomastodontidae), namely, M. angustidens 
and M. longirostris. 

(4) Finally, in 1879, Lortet and Chantre distinguished two groups of species, separating in group II ( = Zygo- 
lophodontinse Osborn), "Mastodontes a dents presentant des coUines larges et tapiroides," the following (p. 304) : 

II Lortet and Chantre (1879) (Present Memoir, 1927) 

Ma.storfon Borsorji Hays, 1834 = Zygolophodon borsoni 

Mastodon Turicensis Meyer, 1839 = Turicius turicensis 

Mastodon Tapiroides de Blainville, 1839 = Turicius tapiroides 

Mastodon Buffonis Porael = Zygolophodon borsoni bujfonis 

Mastodon Vellavus Aymard, 1846 = Zygolophodon borsoni vellavus 

Mastodon Vialetti Aymsird, 1846 = Zygolophodon borsoni vialetii 

(5) Practically the same zygolophodont phylum is adopted by Mayet (1909, p. 45) who, with Dep6ret, 
enjoys the advantage of direct examination and comparison of the type and referred specimens included within 
Mayet's Section B "Mastodontes a Mamelons Disposes en Cretes Transversales (Type Lophodonte)." These 
species, namely, Mastodon borsoni, M. turicensis, M. tapiroides, [and M. pyrenaicus], are those chosen by the 
Austrian palaeontologist M. Vacek in 1877 as typical of the genus Zygolophodon. 

BUNOLOPHODONT AND ZYGOLOPHODONT MiOCENE AND PlIOCENE SpECIES OF EuROPE AFTER MaYET, 1909 

A. - MASTODONTES A MAMELONS ARRONDIS (TYPE BUNOOONTE) B. - MASTODONTES A MAMELONS DISPOSES EN CRETES 

TRANSVERSALES (TYPE LOPHODONTE) 



Mastodon angustidens Cuvier. 



Mastodon turicensis Schinz. 



1825. Mastodon angustidens. Ca\ier,OssenKnls fossiles, I. I, p. i5o. 

i8»5. ilfastodon m/nutus. Cuvier, W, p. 267, pi, II, fifr. II. o c m , j . ■ >.. ^ ^ , ,• . . c 

tR'it Mactnrinn Ci'm^.....>».... t . . \t . ■ i ,,- j r- 1025. Mastodon tapiroides. Cuvier, Osscments fossucs, I. i. p. 267, 

1001. masioaon bimorrense. Lartet, Notice sur la colline de Sansan, 1 m r ^ r-r -^ •■ j «</ • ^m ._ 

, * pi. Ill, fig. (>. Cf. aussi Gueltard. M^moires. I. VI, io*in., 

i85,. Mastodon Gau/aci. Larlel, H, p. 24. „ „ pi. VII, fig. 4. j,..-,jj o- 

.854. Mastodon Cuvieri. Pomel, Calalog„c mModique, p. ^6. "*''• ^^stodon turicensis. Sch.nr, Naturg. und Abbdd. d. Sauge- 

1859. Afastodon angruslidens. Lartel, Nole sur la denlilion des pro- „, „ t/iierc, 1827, p. 243. .^ i- 

boscidiens vivants et fossiles (Bullrlin dc la SociM gMo- '^<=- '«'^^°'i<''' «'^'««°"«- "■.™" "^^'er^ P'l'onlolog.ca zuf C«cA.- 

ninttf ^» c^-,«/.- .a- ,c I ■v-\ri e- . < w ^""'^ dcp trdc ufid ihrer (jescfiopfe, rranclorl. 

r 6) ' "• '■ "'• ^ • ^- "*• P'- ^^' -85°. Mastodon tapiroides. Blainville, Osl/ographir. G Elephant. 

.859. Mastodo°n angustidens. P. Gervais. Zoologle cl PaUonlolonie '^'S" ^^^^odonBorsonii. G. Gervais (p. part.), Zoologic et PaUonlo- 

franfaise,. ^' Mil. log.e .fran,;aises, ^' ed., p.6i. 

.870. itfa.lodon arvernensis. Fraas, Die Fauna von Sleinhcim, Slutl- 'I'l' '"^'^°^''° amnis.lomda.. Archive, duMme^.mde Lyon, t. I. 

cart 1870 1878. Masiodon tapiroides. Lortet et Chantre, Hccherches sur les 

.887. Mastodon angustidens. Dcpdret, Verlchrds miocenesde la vallee Mastodontes M'cA„« du A/us^um de Lyon, t. II, p. 285. 

du Rhone (/lrc/i«c5duJ\/us^umd/!is(oirena(urc/tcc/eZ,vo/l, 00 „ P , . ''. „,.,,,,.., ■ , „. 

^ jV I) '^ 1887. Mastodon turicensis. Dep^ret, Vert^br^s miocenes de la vallee 

i8u8 MastnHnn o n m. el Irion t. \1.>.,ai M.,™ T" ■ • j , i <^n Rhone (Archives du Musi*um de Lyon, t. IV p. i3i). 

loyo. fliasioaon angrustiaens. tMayei, Maminifercs miocenes des sables o »» . .. - ■ . . .* . ». -r- ^ . . T /, 

dp I Orlranni.: oi Hoi f,i,.„. ^. 1, T /. . 1908. JIfastodon tunconsis. L. Mavct, Mammiferes miocenes dcs sablcs 

ac JUileanais et aes laluns de la rourainc (Annates a tnf . j f i j i x ra i.. ,1. 

deVUniversilf de Lyon, fasc. 24) : _ des sables dc TOrlea- ^' ' ^rleanais et des faluns de la Touraine (A'>''ales de 

nais n ,/nnl VII f,„1 < KK- A t, J Lr/ ■ t Universili de Lyon.lasc. 2^; — des sab\es ds VOrUsaais, 

nais,p. 140, pi. VII, liR. J, 4, 5 6 .— des falunsdu B^sois, , , ,,,,," „ . ' j i- i j 01.1. „■. 

p. 297, pi. XI, fig. a et 3). P- '94. P'- ^IH, fig. 1 et 2 ; — des faluos du BI^sois 

p. 298, pi. XI, fig. 4 et 5. 

Species, including types and referred specimens, related to Type and referred species related to the Mastodon turi- 

the Mastodon angustidens phylum. After Mayet, 1909, p. 44. censis phylum. After Mayet, 1909, p. 45. 

[ = Longirostrina! and Brevirostrinaj of the present Memoir] ( = Zygolophodontinse of the present Memoir] 

124 



MASTODON AND MAMMOTH 



125 



(6) In Vacek's invaluable Memoir of 1877, "tjber Osterreichische Mastodonten," we not only find a clear 
separation of the Zygolophodontinse from the Longirostrinae {Trilophodon, Tetralophodon, etc.), but a clear separa- 
tion of the Brevirostrinae {Anancus, etc.) from the Longirostrinae. Vacek's classic memoir is fully treated in 
Chapter VII (the Zygolophodontinae) . 



4. THE BREVIROSTRIN.E: SEPARATION BY LORTET AND CHANTRE FROM THE 
LONGIROSTRINiE AND ZYGOLOPHODONTIN^ 

Following Cuvier, in a long period of exploration, discovery, and research, representatives were found of a 
new and important phylum, unknown to Cuvier, namely, the Brevirostrinae, including 'Mastodon' arvernensis of 
Auvergne and related forms, while many additional members of the phyla known to Cuvier were discovered and 
traced into genera by Falconer. 

The state of knowledge and opinion in 1878 is reflected in the splendid memoir "Recherches sur Les Masto- 
dontes" of Lortet and Chantre, 1879, pp. 285-311, Pis. i-xvi (bis). In this invaluable memoir the mastodonts 
are grouped as follows: 



Mastodontes a Dents Etroites 

MASTODONTES DissiMiLis, Jourdan 
[p. 297] 
Mastodon dissimilis Jourdan, 1840. 

" Arvernensis. . . .Croizet et Jobert, 1828. 

" Arvernensis de Blainville, 1839. 

'■ angustidens Cuvier, 1836 (pro parte). 

" angustidens de Blainville, 1839. . {pro parte). 

" angustidens. . . .Laurillard, 1846 {pro parte). 

" breviroslris P. Gervais, 18.59. 

Anancus macroplus Aymard, 1846. 

1 = Brevirostrinae and Longiro.strina; of the present Memoir] 



II 

Mastodontes a Dents Pr^sentant des Collines 

Larges et Tapiroides 

MASTODo.x BoRSONi, Hays 

[p. 304] 

Mastodon Borsoni Hays, 18.34. 

Turicensis H. v. Meyer, 1839. 

" Tapiroides de Blainville, 1839 {pro parte) 

Borsoni de Blainville, 1839. 

" Bujfonis Pomel. 

" Vellavus .\yniard, 1846. 

Vialetti Aymard, [1846]. 

" Borsoni Vacek, 1877. 

" Borsoni Lartet, 1859. 

[ = Zygolophodontinffi of the present Memoir] 



MASTODON LONGIROSTRIS, Kaup 

[p. 303] 

Mastodon luiigiro.stris Kauji, 1835. 

" angustidens de Blainville, 1S39 {pro parte). 

" angu.ttidens P. Gervais, 1859 {pro parte). 

" longiro.stris Owen, 1861. 

" longirustris H. v. Meyer, 1807. 

" longirostris Falconer, 1868. 

" longirostris Vacek, 1877. 

[ = LongirostriniE and Tetralophodontinae of the present Memoir] 

It will be observed that while Lortet and Chantre (1879) follow Cuvier in uniting the brevirostrine, longi- 
rostrine, and zygolophodontine mastodonts into the single genus 'Mastodon,' they practically separate the species 
into three groups wliich correspond respectively with the subfamilies Brevirostrinae, Longirostrinae, and Zygolo- 
phodontinae of the present Memoir. 

Consofiuently the grouping of species by Lortet and Chantre (1879) is nearer the truth than that of Mayet 
(1909), becau.se Mayet includes under his Group A, "Mastodontes a Mamelons Arrondis (Type Bunodonte)," 
species belonging in both the Longirostrine and Brevirostrine phyla, while he rightly includes in his Group B, 



126 



OSBORN: THE PROBOSCIDEA 



"Mastodontes a Mamelons Disposes en Cretes Transversales (Type Lophodonte)," the entire Zygolophodontine 
phylum. 

In the Miocene of Europe no representatives of the Brevirostrinse have been found, but this phylum is appar- 
ently represented in the PUocene of India, e.g., Anancus perimensis. The author has not yet had an opportunity 
of personally examining the classic collections of the Proboscidea of Miocene Europe, which must be very carefully 
reexamined before their true phyletic, generic, and specific characters can be determined. We have positive evi- 
dence of the existence of four entirely separate and distinct proboscidean lines of descent, and we see some evidence 
of the existence of a fifth Une in Miocene Europe, although the fifth phylum has not been positively demonstrated 
as yet. These theoretical lines' are as follows: 



I II 

MASTODONTID.E 

MaSTODONTINjB ZYGOLOPHODONTINyE 

flcscencled from perhaps a branch of 

Miomatstodon Zygolophodon 

Pliomastodon Turicius 



III IV 

BUNOMASTODONTID^ 

LoNGIRO.STRINiE SerRIDENTIN^ 

descended from perhaps a branch of 



Phioniia 

Trtlophodon 

Tetralophodon 



Phiomia 
Serridentinus (?) 



V(?) 

ELEPHANTID^ 

Stegodontin^ 
a branch of 
unknown ancestors 
[Stegolophodon] 



These four to five subfamiUes are represented by species of Miocene and Pliocene age as follows : 

1. Mastodon tinse by species of Miomastodon and Pliomastodon, e.g., Miomastodon tapiroides americanus and 

Pliomastodon americanus praetypica. 

2. Zygolophodontinse by the species Turicius tapiroides, T. turicensis, and Zygolophodon borsoni. 

3. Longirostrinse by the species Trilopliodon angustidens; all longirostral types. 

4. Serridentinse, medilongirostral Serridentines with medium length of jaw, specifically defined by Serri- 

dentinus filholi in Europe; represented in Asia by S. mongoliensis; in America by S. productus, 
S. serridens, and other species. 

5. StegodontinsB (i.e., ancestors of) possibly represented by a species, Zygolophodon [== Stegolophodon] sub- 

latidens, recently described by Schlesinger from eastern Europe. 



5. THE ZYGOLOPHODONTINiE ( = MASTODON BORSONI): FINAL SEPARATION (RUSSIA, 1894) 
FROM THE MASTODONTIN^ ( = MASTODON AMERICANUS) 

It appears, from Marie Pavlow's revision (1894) cited below, that all 
Russian students of the Proboscidea worked under the spell of Cuvier and 
conservatively adhered to his generic name Mastodon even though they 
clearly recognized the wide distinctions which exist between the genotype 
species Mastodon americanus and many other species which he included within 
this genus. As detailed above, the last phylum to be spUt off from the 
Mastodontinae of the true Mastodon americanus type is that embracing a large 
number of European and Asiatic species which may be grouped under the 
name Zygolophodontine. Beginning in 1824 with the discovery of Mastodon 
turicensis and continuing in 1834 with the recognition of M. borsoni, this 
important and entirely distinct phylum, the Zygolophodontine, was grad- 
ually recognized, until in 1877 Vacek proposed the generic name Zygolophodon. 




Fig. 81. Tooth referi'eJ to 'Mastodon 
ohiolicus' by Pavlow, from Pestchana, 
Podolia, Russia. (Pavlow, 1894, PI. i): 
"Fig. 4, une m'[M2] inferieure droite . . ." 
Univ. de Moscou. [ =M. (?)pavloun ref.] 

That this is a true species of the true 
genus Mastodon appears from the two 
principal 'mamelons' composing each crest. 



'(Osbom. 1934). These theoretical lines of generic descent are considerably modified in the Appendix of the present Volume. 



6. SEPARATION OF SPECIES OF '\L\STODOX' AND 'ELEPIL\S' DESCRIBED IN 

AMERICA (1792-1874) 

The American palaeontologists up to the year 1878 were no less conservative than their European confreres 
in retaining the two generic names Mastodon and Elephas for all kinds and forms of proboscideans. The new generic 
names Euelephas, Tetralophodon, and Trilophodon of Falconer were reluctantly admitted only by certain authors. 
Between 1884 and 1922, however, the tendency to spUt the mastodonts into a number of genera did not assert 
itself generally until Osborn began the present revision, as shown in the following chronological list of names as 
originally used and as revised in the present Memoir. 

Original Genf.ric .\xd Specific Names Applied Generic and Specific Names as Subfamily Phyla 

TO Species of Mastodonts Revised by Osborn in the to the Year 

Present Memoir 1927 

1792 Elephas americanus Kerr. Mastodon americanus Mastodontin.e 

1834 Mastodon borsoni Hays Zygolophodon borsoni Zygolophodontin.e 

1834 Mastodon chapmani Hays Stegomastodon chapmani (Hcmboldtin.e subfam. nov.] 

1838 Elephas jacksoni Mather Parelephas jacksoni Mammontix.e 

1842 Elephas americanus DejKay Mammonteus primigenius americanus Mammontix.e 

1857-1868 Elephas columbi Falconer Parelephas columbi Mammontin.e 

1858 Mastodon (Tetralophodon) mirificns Leidy Stegomastodon mirificus [Humboldtin.e subfam. nov.) 

1858 Elephas imperator Leiily A rchidiskodon imperator Mammontin.e 

1859-1861 Elephas texinnus Owen 1859, and Blake, 1861 Parelephas columbi M.^mmontin.e 

1868 Mastodon americanus Leidy Mastodon americanus Mastodontin.e 

1869 Mastodon obscurus Leidy Trilophodon obscurus Longirostrin.e 

1871 Mastodon shepardi Leidy Rhynchotherium shepardi Rhy'.nchorostrin.e 

1873 Mastodon proavus Cope Serridentinus proavus Serridentin.e 

1874 Mastodon productus Cope Serridentinus productus Serridentin.e 

A full Ust of the generic and specific synonjins of Mastodon americanus will be found in Chapter VI (Sub- 
family Mastodontinae) , under the heading of this species. 



7. OSBORN'S CLASSIFICATION' (1927) OF THE SUPERFAMILY MASTODONTOIDEA 

From the foregoing sections it appears that the collective genus 'Mastodonte' of Cuvier, 1806, has become, 
through discovery, analysis, and research, equivalent to the superfamily Mastodontoidea. This superfamily, 
as it comprises all the mastodonts in all parts of the world, is far more comprehensive than Girard imagined 
in his Mastodontidae, because it includes two families, Mastodontidae and Bunomastodontidae, and not less than 
seven subfamilies and seventeen genera, the number of which is constantly being enlarged.' 

Through discoveries by Schinz (1824), by Vacek (1877), by Lortet and Chantre (1879), by Pavlow (1894), 
by Mayet (1909), and by Osborn (1918-1925), the seven distinct subfamily phyla, Mastodontinae, Zygolopho- 
dontinae, Longirostrinae, Serridentinae, Rhynchorostrinae, Notorostrinae, and Brevirostrinae, have been separated 
off. 

Thus the superfamily Mastodontoidea requires a definition broad enough to embrace the characters common 
to these seven subfamihes. 

'This, like all other cla-ssifications of the 1927 period, is now replaced by the 1935 classification set forth in the .\ppendi.T at the close of the present Volume. 
See also page 11 and footnotes on pages 27, 30, and 31 aljove. .\t the present time (July, 1935) the genera included within the Mastodontoidea number 
twenty-nine. 

127 



128 



OSBORN: THE PROBOSCIDEA 



Suborder or Superfamily: MASTODONTOIDEA Osborn, 1921 
Families:^ Mastodoiitidae Girard, 1852, Osborn, 1918; Bunomastodontidse Osborn, 1921 
Original reference: Mastodontoidea Osborn, Amer. Mus. Novitates, No. 1, p. 1 (Osborn, 1921.515); Masto- 
dontida; Girard, Proc. Amer. Assoc. Adv. Sci., 1852, pp. 326, 328; Osborn. Bull. Geol. Soc. Amer., XXIX, 1918, 
p. 134 (Osborn, 1918.468). Bunomastodontidaj Osborn, Amer. Mus. Novitates, No. 1, 1921, p. 2 (Osborn, 
1921.515). 

Definition. — Chiefly forest and savanna living proboscideans; widely migrating in the north and 
south temperate regions. Browsers on tree and shrub leafage and on twigs ; uprooters in part. Superior 
and inferior incisive tusks adapted to varied food habits, branching and uprooting, and to varied modes 
of defense; retaining or losing enamel bands. Deciduous premolar series prevaiUng over permanent 
premolars. Intermediate molars, P4-M2, early becoming trilophodont ; third molars, M'-Ms progres- 
sive to 4-7 crests; cones, conules or conelets progressive to lophodont, bunolophodont, or ptychodont 
form. Body prevailingly elongate, broad, massive, short limbed. 



MASTODONTIDiE 



BUNOMASTODONTID^ 



1. Subfamily 
Mastodontin^ 
Nipple-toothed mastodonts 

Tusk enamel disappearing. 
Grinders sublophodont to 
subhypsodont. 



1. 



[Palxomastodon^ 
Mastodon 
Miomastodon 
PUomnatodon 
= Mastodon giganteum Cuvier 



2. Subfamily 
Zygolophodontin^ 
Yoke-toothed mastodonts 

Incisive tusks oval, (?) re- 
taining enamel band. Grind- 
ers becoming truly crested, 
with numerous conelets. 



2. 



Zygolophodon 
Tun'ciiis 
= Mastodon tapiroides Cuvier 



3. Subfamily 

longirostrin^ 

Long-jawed mastodonts 

Upper and lower tusks 
functional in feeding, superi- 
or tusks retaining enamel 
band, inferior tusk.s spatu- 
late, losing enamel band. 
Grinders with central con- 
ules or conelets in valleys. 



Phiomia 
Trilophodon 
Tetralophodon 
3. = Mastodon angudidens Cuvier 



4. Subfamily 

Serridentin^ 

Serrate-toothed mastodonts 

Superior and inferior inci- 
sive tusks functional in feed- 
ing, retaining enamel band. 
Rostrum moderately elon- 
gate. Grinders without 
central conules or conelets, 
with external and internal 
serrated spurs. 



Serridentinus 
Unknown to Cuvier 



5. Subfamily 

RhyNCHOROSTRINjE 
Beak-jawed mastodonts 

Superior and inferior tusks down- 
turned, functional in feeding, retain- 
ing enamel band. Grinders with inter- 
mediate conules or conelets. 



Rhynf.hotherium 
5. Unknown to Cuvier 

The above synopsis shows that in 
seven subfamilies. To these are now 
Appendix of the present Volume I). 



BUNOMASTODONTIDiE 
6. Subfamily 

NoTOROSTRINyE 
South American mastodonts 

Jaws gradually abbreviating. Su- 
perior incisive tusks purely offensive 
(without enamel) or used in feeding 
(with enamel). Grinders with inter- 
mediate conules or conelets. 



Cordillerion 
Cuvieronius 
Eubelodon 

= Mastodon andium and 
M. humholdtii Cuvier 



7. Subfamily 
Brevirostrin.«! 

Short-jawed mastodonts 

Superior tusks offensive weapons, 
losing enamel band ; inferior incisive 
tusks disappearing. Grinders with 
alternating external and internal 
lobes, i.e., cones, lacking interme- 
diate conules or conelets, developing 
double or quatlruple trefoils. 

Anancus 
Pentalophodon, [Synconolophus] 
7. Unknown to Cuvier 



1927 Cuvier's 'Mastodon' was known to embrace five classic species and 
(1935) added seven subfamilies, making fourteen subfamilies in all (see 



'(Osborn, 1934) The .suliordcr Mastodontoidea now includes the additional families Rcrridentidse fam. nov. and Humboldtida; fam. 
^(Oaborii, 1931) The subfamily rclationsliips of [Paluiotmtstudon] ai 



nov. in Appendix of the present Volume. 



I arc now uncertain. (193.5) See p.agcs 143-1-19 below, also Palaeomastodnntina> subfam. 



8. CUVIER AND THE REVOLUTIONS OF THE GLOBE 

This historic chapter may appropriately conclude with quotations from "Les Transformations du Monde 
animal," a work published in the year 1907 by Charles Deperet, who next to Cuvier and Gaudry ranks among the 
leaders of the palaeontology of France. Deperet ardently sets forth and defends Cuvier as the creator of com- 
parative anatomy and of palaeontology. (Deperet, 1907.2, pp. 7, 10, 13, 14, 19, and 22): 

L'admiration g6n6rale suscit^e dans le monde savant par la s6rie de m^moires que G. Cuvier publia k partir de 1798 et qui 
furent r^unis, en 181 2, sous le titre de : Recherches sur les ossementsfossiles, n'a pas diniinu^ [sic] de nos jours, inalgre un recul de plus 
d'un sifecle. Tout naturaliste qui veut se familiariser avec I'organisation des aniniaux .superieurs vivants ou fossiles doit encore 
aujourd'huicommencerses Etudes par la lecture decet ouvrage magistral, oil se trouvent expos&s, avec une clart^et une precision 
luniineuses, les notions fondamentales des deux sciences sceurs: V A natomie co7nparee et la Faleontologie des vert^br^s. . . . 

Ainsi, Cuvier a non seulement d6montr6 la presence dans les couches s6dimentaires d'une sirie de faunes terrestres super- 
poshes et distinctes, mais il a eu le premier, et tres nettement, I'id^e du perfedionnemenl organique graduel de ces faunes depuis 
les plus anciennes jusqu'aux plus modernes. C'est \h, une notion fondamentale, dont on oublie trop souvent d'attribuer le 
m6rite k Cuvier, dans les jugements d'une s6v6rit6 excessive et souvent injuste, que les transformistes ont port6 et portent encore 
sur les idees cuvieriennes en matiere de pal^ontologie philosophique. . . . 

L'illustre naturaliste expose ses id^es, avec sa clart6 habitueJle, dans I'admirable Discours sur les revolutions du globe, qui 
forme I'introduction de son grand ouvrage sur les ossements fossiles. Pour lui, les extinctions de faunes ont Hi k la fois 
completes et brusques, provoqufes par des 6v6nements geologiques violents ou revolutions du globe, d'un caractSre de g6n6ralit6 
assez grand, niais non cependant absolu. En faveur de cette hypothese, Cuvier invoque de nombreux faits, d'ordre g^ologique, 
qui, pris isolement, et en tenant compte des documents connus k cette epoque, sont d'une rigoureuse exactitude, et dont les 
relations r^ciproques seules deviennent discutables ou meme inexactes. . . . 

Quant au proc^dd de ce renouvellement, le reproche a 6t6 bien souvent £>,dress6 a Cuvier d'avoir admis a son tour une autre 
hypothese tout aussi peu d^montrable scientifiquement, celle des creations successives. Mais c'est encore 1^ une critique tout 
k fait injustifi6e. Nulle part, le mot de creation ne se rencontre dans I'oeuvre de Cuvier et il suffit de lire avec attention le 
Discours sur les revolutions du globe pour voir que dans I'esprit de l'illustre savant, il s'agit seulement d'invasions de nouvelles 
formes animales venues brusquement de contr^es lointaines inconnues. Ici, l'id6e est assez fondamentale pour qu'il vaille la 
peine de citer: 

'Au reste, lorsque je soutiens,' dit Cuvier, 'que les bancs pierreux contiennent les os de plusieurs genres et les couches 
meubles ceux de plusieurs especes qui n'e.xistent plus, je ne pretends pas qu'il ait fallu une creation nouvelle pour produire les 
especes existantes, je dis seulement qu'elles n'existaient pas dans les memes lieux et qu'elles ont du y venir d'ailleurs.' . . . 

Le lecteur voudra bien excuser sans doute la longueur de cette citation, dont I'interet est considerable; elle d^montre 
jusqu'^ I'^vidence qu'il faut reporter k Cuvier tout I'honneur d'avoir pose, avec une nettete et une exactitude admirables, 
I'hypoth^se si importante et si f^conde du renouvellement des faunes par voie de migrations. . . . 

Ainsi, deux points paraissent absolument certains pour d'Orbigny: d'une part, la creation en bloc de faunes entieres — 
opinion que Cuvier n'avait jamais formulae — de I'autre, la disparition brusque de chacune de ces faunes. Pour le premier de ces 
grands faits, le savant pal6ontologiste ne tente meme pas, on I'a vu plus haut, le moindre essai d'explication scientifique. . . . 
C'est, on le voit, la reproduction k peu pres integrate des revolutions du globe de Cuvier, avec une generalisation encore plus 
grande. . . . 

Si I'explication de ce grand fait par des creations successives ne saurait nous satisfaire au point de vue scientifique, nous 
aurons plus loin k en chercher une interpretation rationnelle par des ph6nomenes de migration de faunes ou de migration de 
milieux, analogues k celles que Cuvier avait d6j^ si bien mis en Evidence pour les aniniaux terrestres. 



129 




vhp american mastono!. , 

The Warren Mastodon as it was First Mounted (1845-1846), shortly after its Discovery 
Fig. 82. The figure is reproduced from tlie Americati Journal of Agriculture and Science, Volume II, Number 2, condueted by E. Emmons, 
Albany, and A. J. Prime, Newburgh. In their article, "The Great American Mastodon," Messrs. Emmons and Prime remark: "The skeleton has 
since been arranged and set up, and this has been done with great care and the strictest attention to the articulating surfaces of all the bones, which 
we believe has not been the case with others which have been put together." 



The Warren M.\stodon 
as remounted in 1849 
As photographed in the 
interior of the Warren Mu- 
.seum, 92 Chestnut Street, 
Boston. 

Fig. 83. For fifty-seven 
years, 1849-1906, the War- 
ren Mastodon, mounted as 
shown herewith, was ex- 
hibited in the Vi'arren Mu- 
seum in Bo.ston. In 1900 it 
was acquired, thanks to the 
generosity of the late J. Pier- 
pont Morgan, by the Ameri- 
can Museum. 

The skeleton, as here de- 
picted, is covered with heavy 
black varnish. The imitation 
tusks are made of pafiier- 
m^chfi and were so length- 
ened as to sweep the ground 
and curve outwards at the 
extremities. The chest and 
ba<'kbonc were raised two 
feet al)ove the top of the 
shoulder blade, or scapula, 
and as a result the natural 
height of the animal was 
increjised from nine feet to 
twelve feet. Beneath the 
Warren Mastodon are tusks 
and grinding teeth of other 
specimens. Around the base 
of the walls are many verte^ 
brie of the giant Zeuglodon, 
the archaic fossil whale of 
the st)Uthern United States. 

The Warhen Mastodon 
as reiiKHinted for Tlie Ameri- 
can Mu.'iciun of Natural 
Hi.story, under the direction 
of Curator Henrv I''airfield 
O.sborn, July, 19(J7, by chief 
preparator .Adam H(>rmann 
and assistants, is shown in 
figure 124, page 178. 




130 



Chapter VI 
HISTORY OF THE SUBFAMILY MASTODONTIN/E, THE TRUE MASTODONTS 
Discovery of mastodon americanus, of pliomastodon matthewi, of miomastodon merriami, and 

OF PALiEOMASTODON BEADNELLI. SUBFAMILY, GENERIC, AND SPECIFIC DISTINCTIONS. HaBITAT, MIGRATIONS, 
PHYLOGENETIC SUCCESSION, AND EXTINCTION OF THE MASTODONTIN^. 

1. History of the subfamily MastodontiniE, 1705-1934. 8. Characters of Pliomastodon and its included species. 

2. R^.sum^ of the subfainilj' characters. 9. Historical review of the discovery of the genus Mastodon 

3. Separation of niastodonts of the Mastodon americanus (1705-1934) and of the species Mastodon americanus 

and Zygoluphodon borsoni phyla in Russia. (1792-18G9). 

4. R^sum(5 of discovery and description of the Mastodon- 10. Characters of Mastodon and its included species and 

tinae. varieties. 

5. Progressive characters and generic distinctions of the 11. Skeletal characters of Warren and Whitfield mastodons. 

Mastodontinie. 12. Age and progressive cranial and dental characters of 

6. Characters of Palseomastodon and its included species. Mastodon americanus, revealed in the American 

7. Characters of Mioftnastodon and its included species. Museum collections. 

In Chapter II it is shown how the Mastodontoidea may be clearly distinguished from the Mceritherioidea, 
the Deinotherioidea, and the Elephantoidea. In Chapter V it is shown how Cuvier's original genus 'Mastodonte' 
{=Mastodon) came to embrace two' distinct famiUes, seven' distinct subfamily phyla, and seventeen' distinct genera. 
In the present Chapter VI Cuvier's genus Mastodon, based on the type Mastodon americanus, becomes the 
terminal member of the subfamily Mastodontinae, otherwise known as the 'true mastodonts,' with a geologically 
prolonged lineage of its own certainly traceable into Asia and Europe and possibly back to the Lower Ohgocene 
Palseomastodon of North Africa. Accordingly Palxomastodon is systematically treated in this chapter, and the 
reader is referred to Chapter II where Palseomastodon is described in its North African environment of Oligocene 
time in competition with Moeritherium and Phiomia. The discovery of a complete cranium and jaws may 
compel us to remove Palseomastodon from the subfamily Mastodontinae^, in which it now rests on the evidence 
at present available. It is certainly not a direct ancestor of Mastodon. 

(October, 1934) The very important addition of the Lower Miocene Miomastodon depereti sp. nov.,from the Sables de I'Orleanais 
of France, of the Mastodon pavlowi sp. nov., from Pestchana, Podolia, Russia, has recently been made to the subfamily Mastodon- 
tinae, as described in the phylogenetic appendix of the present Volume; see fig. 230D, p. 284, of Chap. VIII, and PI. I, p. 134; also of 
the Mastodon acutidens sp. nov. of the Pleistocene of North America (see Figs. 131 A1-A4, 135, and PI. I, also Appendix). 

1. HISTORY OF THE SUBFAMILY MASTODONTIN.E, 1705-1934 

Original reference: Mastodontina Brandt, 1869, Mem. Acad. Imp. Sci. St. Petersb., (VII), XIV, No. 1, p. 35. Mastodontinae 
Osborn, 1910, "The Age of Mammals," p. 558 (Osborn, 1910.346). 

Although the first fossil remains of a member of this great subfamily were discovered on the banks of the 
Hudson River as early as 1705, also near the banks of the Ohio River (Big-Bone Lick, Kentucky) by the Indians 
who gave them in 1739 to a French officer named Longueil, the origin, migrations, and early geologic history 
of the Mastodontinae are still comparatively obscure. 

This we believe is because the Mastodontinae have included forest-living animals from the very beginning of 
their history, for fossil remains of forest-living animals are always rare in Tertiary deposits. 

Only by such a forest-loving hypothesis can we explain the unparalleled abundance of the remains of 
Mastodon americanus in the ancient forest, bog, and swamp deposits of the eastern and middle United States, as 

'See page 128. 

■'See pages 143-149 below, also Palieomastodontiiia; in Appendix. 

131 



132 OSBORN: THE PROBOSCIDEA 

summarized in the nation-wide records collected by Dr. J. M. Clarke, Dr. 0. P. Hay, and others. And only by this 
hypothesis also can we explain the abundance and remarkable preservation of the more or less complete skeletons 
of our American Mastodon, which we believe outnumber those of all other extinct proboscideans put together. 

Mastodon. — In the northern portions of the American continent these remains become very rare, and as we 
pass westward into the plains region and northward into Eurasia not a single fossil record is found until we reach 
Podolia, Russia, from which locality Marie Pavlow in 1894 described a single tooth (Fig. 81) which she referred 
to 'Mastodon ohioticus' This central Eurasiatic region also yields teeth of the species belonging to the related 
subfamily Zygolophodontinse. When the grinders of Mastodon americanus are placed side by side with those of 
Zygolophodon borsoni we observe that they are really profoundly different; we verify this observation as we 
proceed westward into Austria, Germany, and France and discover that the bilobed grinding tooth of the true 
Mastodon may be traced back into its simple ancestral forms, which are quite distinct from the quadrilobed 
grinding tooth of the true Zygolophodon. 

MioMASTODON. — This new stage of ancestral Mastodon grinder, as it appears in the Miocene of Holarctica, is 
placed in the distinct genus Miomastodon Osborn, because with these simple grinders are associated upper incisive 
tusks with broad enamel band; no Lower Pleistocene Mastodon thus far discovered has the enamel band. Thus the 
Pleistocene Mastodon americanus certainly leads back into the Miocene Miomastodon of western Eurasia and 
of North America. An intermediate stage is the PHocene Pliomastodon lacking the enamel band. 

Pal^omastodon. — Somewhat doubtful at present is the theory advanced by Matsumoto, and more recently 
supported by Osborn, that the original ancestor' of the Mastodontinse is to be found in the true Palseomastodon 
(as distinguished from the longirostral Phiomia) which occurs with relative rarity in the Oligocene river sands of 
the Fayum of northern Africa. Fossil remains of Palxomastodon are relatively rare, the skull is little known, but 
so far as our present knowledge goes the Palaeomastodon grinding teeth are not ancestral in type to those of the 
true American Mastodon. 



2. RESUME OF THE SUBFAMILY CHARACTERS 

Subfamily Characters. — Skull brachycephaUc, brachyopic; cranium relatively broad and low. 
Mandibular rami rapidly abbreviating, with persistently rounded inferior tusks becoming variable. 
Superior tusks rounded, with enamel band in the Palaeomastodon^ and Miomastodon stages disappearing 
in the Upper Pliocene (FZzomas^odon) and Pleistocene (true Mastodon) stages; superior tusks upturned 
.... and out-turned, adapted to browsing and uprooting purposes, powerful offensive and defensive weapons. 
Grinding teeth hexabunodont in Palseomastodon, primitively tetrabunodont in Miomastodon and 
Mastodon by the absence of the intermediate conules, gradually acquiring a bunolophodont form; 
intermediate grinders, Dp^-Dp^, M'-Mi, M'-Mj, becoming trilophodont; third superior grinders, M^ 
progressive to tetralophodont stage orily; third inferior grinders progressive to tetralophodont + 
or four and a half ridge-crests only. 

Warning the reader of the present limitations of our knowledge, which chiefly arise from the very great rarity 
of fos.siI materials in the Eastern Hemisphere, we may give a theoretic summary of the ancestral history of the 
Mastodontina;. 

. Geographic and Geologic Range.— Springing from unknown ancestors of the Eocene-Oligocene of North 
Africa, the true Mastodontinae appear in the Lower Miocene Miomastodon of western Eurasia and soon find their 
way eastward into North America, appearing in the Middle Miocene species of Nevada, Miomastodon merriami, 

'See footnotes on pages 36 and 131. 



THE MASTODONTIN.E: HISTORY 133 

and in the Lower Pliocene species of western Nebraska, Pliomastodon matthewi, meanwhile leaving behind 
in Austria-Hungary their relatives, Miomastodon tapiroides americanus and Pliomastodon americanus praetypica, 
which give rise to the rare true Mastodon of southern Russia, referred to 'Mastodon ohioticus' by Pavlow. Rarely 
found in northern and western North America, they multiply rapidly in the favorable forests of the middle and 
eastern United States in the typical form Mastodon americanus. 

Subfamily Habital Adaptation. — These broad, massive, low-bodied, low-headed, well-defended Mastodon- 
tinae probably evolved chiefly in the north temperate forests of Eurasia and of North America avoiding the ex- 
tremes of warm cUmate in both hemispheres. The American Mastodon is known to have lived on the foliage of 
trees and shrubs aided by the vertical chopping motion of its jaws and the free interlocking action of its superior 
and inferior ridge-crests, the valleys not being filled or blocked with the trefoils, as in the valleys of the Longi- 
rostrinae molar. That these true mastodonts were temperate and cold-loving animals, enjoying coniferous herbage, 
is shown by their avoidance of southern Eurasia, where no trace of the subfamily is found, also by their early 
disappearance in northern Africa. 

3. SEPARATION OF MASTODONTS OF THE MASTODON AMERICANUS AND ZYGOLOPHODON 

BORSONI PHYLA IN RUSSIA 

The reader is referred to Chapter V for the history of the separation of the Zygolophodonts in Europe as 
distinguished from their separation in Russia. See also Pis. i, ii, and iii, between pp. 134-135 of the present 
Chapter. 

We owe to the learned Russian palaeontologist Marie Pavlow the full historic record, from the time of 
Buffon, of the occurrence of the true Mastodon (ohioticus) americanus in Russia and its clear distinction from the 
true Zygolophodon borsoni of Russia, a molar of which was presented to the great French naturalist Buffon in 1770 
and beautifully decsribed and figured by him in 1778 (Fig. 84A). 

Pavlow, 1894, pp. 1, 2, 36. — Avant d'aborder la description de tous ces restes fossiles, il me semble necessaire de faire une 
cotirte revue de la iittcrature sur ce qui est connu jusqu'a present comme restes de Mastodon en Russic ... [1] Buffon. en 
1775 ['], a 6t(5 le premier a dccrire [Footnote: 'Buffon. Epoques de la Nature. 1775. PI. i-iii.'] deux dents provenant de la 
Russie, et rapport6es plus tard par Lartet au Mastodon Borsoni. Une d'elles que le comte de Vergennes lui avait donn^ en 1770 
(1. c. PI. I et ii) a et^ indiqu6e eomnie trouvee dans la Pctite-Tartarie [Footnote: 'On dfeignait par ce nom a la fin du XVIII 
si^cle toute la Nouvclle Russie d'aujourd'hui.']. Une autre dent, apportee par I'abb^ Chappe de Siberie (PI. iii), a eveille un 
grand doute sur sa provenance. 

[2] Buffon, sans indiquer le genre auquel ont appartenu cos dents, les croit identiques a celles trouvees pres de VOhio et ne 
doute pas qu'il y avait un animal inconnu jusqu'alors et commun aux deux continents. Ces ^chantillons tj'pes se trouvent 
maintenant dans la Grande Galerie du Museum a Paris. . ^ 

[3] Presque en meme temps Pallas a d^crit, en 1770-77 [Footnote: 'Pallas. Acta Acad. Petropolitanae. 1777. PI. 9.'] une 
dent tres us^e et trouvee dans les sables ferrugineux, pres de Bclaja, affluent de la Kama. Pallas rapprochait cette dent de 
celles de I'animal de rOhio et plus tard Blainvillc ct Eichwald [Footnote: 'Blainville. Ostcographie. Eichwald. Palcontologie 
de la Russie.'] Font rapport6e au Mastodon tapiroides. Le dernier de ces savants indique que cette dent se trouve a I'lnstitut 
de Mines a St. P^tcrsbourg. 

[4] II me semble que, autant qu'on pcut on jugcr d'aprcs le dessin dc ce debris tres mal conserve, cette dent a du appartenir 
a un Mastodon du type 'Zigolophodon', c'est a dire poss6dant des crates absolument nettes, d^pourvues de mamelons inter- 
m^diaires. Ce type est le mieux reprdscntd par Mastodon Borsoni. 

[5] Soixante ans a peu pr^s se sont (^coules apres cette description de Pallas, sans que personne mentionnc des mastodontes 
trouvds en Russie, et ce n'est que dcpuis 1835, que des nouvcllos trouvailles et des descriptions ont ete faitcs. 



'In the bibliography Pavlow gives the reference as follows: "Buffon. Supplement 4 I'histoire naturelle. Epoques de la nature. Tome V. PI. 1-5, 1778." 



134 



OSBORN: THE PROBOSCIDEA 



[6] Ainsi Fischer de Waldheim en 1835, en determinant d'apres les dessins les fossiles tiouves pres de Riazan, indique dans 
une courte notice [Footnote: 'Fischer de Waldheim. Bull. Natur. Moscou p. 394, PI. x.'] una jeune dent d'un Mastodon qu'il 
considere comme la premiere trouvaille de ce genre faite en Russic. Le dessin n'est pas assez bien fait, pour pouvoir determiner 
I'espece de ce Mastodon. L'echantillon est indique comme se trouvant dans la collection du Lycee de Riazan — 

[7. Op. cit., p. 36] : Age geologique et repartition geografique du groupe Zygolophodon. L'age geologique de ce groupe des 
Mastodon est bien prolongc. On rencontre leurs differents representants depuis le mioccne, durant le pliocene, en Europe et le 
pleistocene en Amerique. Les formes les plus anciennes ont ete indiquees en Espagne (Mast, tapiroides) dans les lignites de 
Brihuega dans le miocene nioyen (ou inferieur — Lartet p. 475 1. cit.). Dans le miocene moyen elles abondent: la France en a 
deux representants: Mast, tapiroides Lartet et Mast, turicensis Schinz (af. Borsoni), dans les faluns de Touraine, graviers 
d'Orleanais, lignites de Soblay (Ain). 

[8] Mast, turicensis Vacek (af. Borsoni) a ete trouve en Silesie. Le miocene en Suisse (Elgg) et la molasse de Winterthur 
sent tres riches en Mast, turicensis Schinz (af. Borsoni), ainsi que le niio-pliocene d'Europe qui debute a Oeningen (Suisse). 

[9] Mast. Borsoni Lartet, et Borsoni Hays provient du Pliocene inferieur d'Asti, d'Auray. Celui de Bravard, Lortet et 
Chantre du Pliocene superieur du Puy-de-D6me, Auvergne. Celui de Mr. Forsyth Major du Plioce7ic superieur du Val d'Arno. 

[10] Le Mast, virgatidens de v. Meyer n'est designe que comme provenant des d6p6ts tertiaires de Foulda (Allemagne) 
sans que l'age soit precise. 

[11] Les diff^rentes dents de M. Borsoni Vacek proviennent de divers depots; ainsi, PI. vi, f. 3 indiquee comme provenant du 
miocene superieur de Neidorf eveille les doutes de I'auteur sur I'exactitude de cette indication, a cause de I'anciennete des depots. 
Pourtant cela ne nous parait pas impossible, prenant en consideration la trouvaille de M. turicensis Schinz dans ces depots et la 
parente, presque I'identite de ces 2 formes. 

[12. Op. cit., p. 37] Quant au rapport genetique de ces formes 11 nous semble possible d'exprimer les suppositions suivantes: 
a) que Mast, tapiroides Lartet (non Schinz) de Simorre est I'espece la plus ancienne dans ce groupe, et qui a donne naissance 
a b) Mast, turicensis et Borsoni de Touraine, de Soblay, de Zurich et d'Asti, laquelle a son tour a precede c) Mast. aff. Borsoni de 
I'Auvergne, de 1' Allemagne, de I'Autriche et de la Russie; et qu'une branche, qui a du se d6tacher de ce dernier a la fin du 
Miocene (aff. ohioticus), a donne les formes de d) Mast, ohioticus de la Russie, developpees dans le Pliocene. 





A' A A 




BuFFON, 1778. Superior Grinding Teeth of Zygolophodon (left) and of Mastodon (right). Originals said to be in the 

Paris Mttseum (fide Pavlow, 1894) 

Fig. 84. Grinding teeth of Zygolophodon borsoni referred (left) and of Mastodon ameiicamis referred (right), after Buffon, 1778, 
Pis. I and IV, one-fourth natural size. |Inverted by H. F. O.j 

A, Buffon's figure of the molar of an animal which we now know to be related to Mastodon [= Zygolophodon] borsoni, found in Russia 
("la petite Tartarie"), presented to Buffon in 1770 by M. le Comte de Vergennes. Observe three to four lobes, i.e., cones or conelets, in 
each tran.sver.se crest. 

B, Buffon's figure of the molar of an animal now known as Mastodon americanus, from Big-Bone Lick, Kentucky, near the Ohio 
River, sent to Buffon by M. Collinson. Observe two lobes, i.e., cones, in each transverse crest, the true Mastodon type. 

Both of these beautifully engraved molar teeth appear to be third superior grinders, M', although they lack the rudimentary fifth 
crest or pentaloph characteristic of the third superior molars of Mastodon americanus (Fig. 133). 



PLIOMASTODON, MIOMASTO D ON, 

AND MASTODON 




PLIOMASTODON PRAETYPICA ref 

D 





'MASTODON OHIOTICUS ref. 



PLIOMASTODON PRAETYPICA ref 

H 




PALAEOMASTODON BEADNELLI ref 



MASTODON AMERICANUS 



PLATE I 
Mastodontin.e: Persistent Longitudinal Sulcus, Two Cones and Four Conelets in Each Ridge-crest, Rudimentary Ecto- (Inferior) 

AND Entothkkoils (Superioh). Fioures less than one-third natural size 
Ixnvcr OliKoociic A, Pnluuinastndon beadiuili ref., r.M^. No sulcus. CompU'tc transverse crests and prominent central conulcs (C, C) prove that 

Falxomaslodon is not ancestral to Mastodon. Amer. Mus. 1348L FayOm of Egypt (cf. Fig. 187c). 
Pleistocene L, Mastodon acutidens sp. nov., l.M'. Longitudinal sulcus, progressively sharpened ridge-crests. See figures 131 Al -.\4 and 

135. Amer. Mus. 17727, Rochester, Indiana. 
K-.I, Mastodon amcricanus ref., r. and l.M"'', Mj^. Sulcus persistent, 4 conelets, superior entotrefoils (K), inferior ectotrefoils (J) 
{=losanges of Cuvier). See figures 133 and 134. .\mer. Mus. 12464, Buffalo, Kansas, and Amer. Mus. 14294, near 
Fulton, Indiana. 
1, G, Mastodon pavlotri sp. nov., type, l.M"'', l.Mn. Primitive. Longitudinal sulcus, 4 conelets, no trefoils. After Pavlow, 1894. 
PI. 1, figs. 2 and 3, Pestchana, Podolia, Russia. 
1", 'Mastodon oliiolicits' ref. | = .U. pavlotri n-l.], r.M-.. Rounded conelets 4-5, entotrefoils. After Pavlow, 1894, PI. i, fig. 4 
(.see Fig. 81 of pre.sent Memoir), Pe.steliana, Podolia, Russia. 

Pliocene C, I), K, 11, I'liomantodon praetypica cotypes and ref. Primitive. Rounded conelets 4, persistent longitudinal sulcus, rudimentary 
trefoils. 
C, D, Cotypes r.M^^ l.M.j, after Schlesingor, 1922, Taf. xv, figs. 2 and 4, Szabadka and Batta-Crd, and I.M3, Taf. xix, fig. 2, 

Batta-Erd, Hungary. 
E, H, Referred l.M' (rev.), after Vacek, 1877, Taf. vi, fig. 3, Xeudorf, Austria, and l.M\ Taf. vi, figs. 1, la, near Theresiopel, 
Hungary. 
Lower Miocene H, Mioniasloilon drprnti sp. nov., l.M''. [Erroneously determined as an l.M,-), Fig. 230, p. 284, of the present Memoir.) The 

most primitive stage thus far known. Deep longitudinal sulcus, 3 4 conelets, 4 ridge-crest* and rudimentary penta- 
loph. After Mayet, 1908, PI. vii, fig. 3, Chevilly, Sables de rOrl<5anais, France. 

Compar.ative Observ.\tions (1935) 
Molars of the true Mastodon phylum arc readily distinguished: (1) By the persistence of the median sulcus even into the progrc-ssivo stage (L) 
of Mastodon acutidins- (2) by the persistent four conelets: (3) by the rudimentary ectotrefoil and entotrefoil spurs. 



Zy GO LOP HO DON 






External 



J7 ZYGOLOPHODON BORSONI 



Ant, 



r-cCX Vy TYPE 

ZYGOLOPHODON BORSONI 






Internal 
ZYGOLOPHODON BORSONI 

External 



PALAEOMASTODON INTERMEDIUS TYPE 



ZYGOLOPHODON PYRENAICUS TYPE 




External 




ZYGOLOPHODON BORSONI 



PLATE II 

Zygolopiiodontin.e, ZYGOLOPHODON : Primitive Longitudinal Sulcus (B, F, (1), Gradually Disappearing 

(E, D), CoNELETs 4 (B) Progressive to 6 (D, E, F', G), no Ecto- or Extotrefoils. Kidgb-ckests 

Obliquely Transverse (D). Not drawn to a uniform scale 



LoWl'lUliKOCClK' 

Middle Miocein' 
Middh' Pliocene 
Upper Pliocene 

Middle Pliocene 

Upper Pliocene 



A, Palseomastodon intermedius type and paratype, M1.3, M'"' (sec Fig. 92). Longitudinal sulcus blocked by proto- and meta- 

conviles. Central conules (C, C) in inferior molars. Not ancestral to .l/n.s/orfoH or Zi/gnlophodon. 

B, Zijgoloitluiildn pyrcnnicus type, r.Mi (.sec Fig. 148). Primitive longitudinal sulcus, 4 eonclets on three posterior crests. 

C, ZiigolophiiiloH borsoni type, r.M^ (see Fig. 154). 4-6 eonelets on three anterior ridge-crests. 

D, Zygiilof)k()(lon borsoni ref., r.M-jj, l.Mo.) (see Fig. \oO). Progressive 45-6 eonelets on perfected inferior molar ridge-crests. 

Vestigial sulcus. 

E, Zygolophodon borsoni ref., l.M', 'intermediate molars' 4 .") eonelets. After Schlesingi^r, 1922, Taf. .xiv, fig. .'> {praelypica), 

Rilkoskereszti'ir, Hungary. M3, after Schlesinger, 1922, Taf. xv, fig. 5, B;lcs-Bodrog, Hungary. 

F, Zygolophodon borsoiii ref., l.M' (see Fig. 1.53). Primitive irregular longitudinal sulcus, 5-<j eonelets. 

d, Zygolophodon borsoni ref., l.M' (.see Fig. 111). Primitive sulcus, 5 eonelets on proto-, mcta-, tritolophs, 4 eonelets on tetarto- 
loph. After Bnffon, 1778, PI. 11. 



Compar.\tive Opserv.\tions (1935) 

I'ahcomnslodon inlcrmcdius (A) is distinguished from Zygolophodon as follows: (1) Median siile\is lacking; (2) central conules (C, C) present in 
anterior valleys; (3) entotrefoils sometimes present. 

Molars of the relatively rare representatives of the Zygolophodon phylum are readilj- distinguished by the following three outstanding characters: 

First, the median sulcus separating the four eonelets in Zygolophodon pyrenaicus (B) is not very distinct; it is reduced or vestigial in Z. borsoni 
(D-G). 

Second, by the subdivision of the /our transverse eonelets of Zygolophodon pyrenaicus (B) into five eonelets in Z. borsoni superior molars (D-G). 

Third, by the entire absence of the eeto- and entotrefoil spurs, which are more or les.s prominent on prolonged usage in specimens of Pliomaslodon 
and Mastodon (PI. i, C-K), 



TURICIUS 




r-TAPIROIDES ref 




Ant, 



E Median 



Am. 




-VIRCATIDENS TYPE 



Am 

-TAPIROIDES 
Am-.-\ _^ _^^ Ant, 

-TAPIROIDES TYPE 



Post. .-^ , .. ... ^ 

-TURICENSIS,^ -TURICENSIS 



■VIRGATIDENS 






Ant. ^-~-<_>L:>'Port. 
-ATT/CUS TYPE 
Am.^_ 




-WAHLHEiriENSIS 



PLATE 111 
Zygolophodontin^, Turicius: Longitudinal Sulcus (A), Vestigial or Absent (A-J), Ridge-crests Constant 4 (A-I), 5 (J), Rudimentary 
EcTo- AND Entotrefoil Spurs (A, F, H), Conelets 4-5-6 (A, B), Rapidly Multiplying, 7 (C)-IO (E)-ll-13 (K)-10-25 (H H3), 
Compre.ssed at Summit (J). All figures (excepting I and J) about one-third natural size 

Turicius tapiroides type, r.Mj. Primitive longitudinal sulous, 4 6 Conelets. .\fter Mayet, 1908, fig. 66, Calcaire de 

Montabusard, Fiance. 
Turicius ttipiroides ref., I.M3. After Mayet, 1908, PI. viii, fig. 2, Sables de I'Orleanais du Bldsois, Poiitlevoy, France. 
Turicius tapiroides ref., r.M,i. Conelets o 7, ridge-cre.sts 4. After Mayet, 1908, PI. xi, figs. 4 and o, Pontlevoy, France. 
Turicius turiccnsis type, l.M'-. After Schinz, 1833, Taf. i, fig. 1 , Elgg, Canton Zurich, Switzerland. 
Turicius turicensis lioma-otype, r.M;. Conelet.s .")-10. .\fti'r von Meyer, 1867, Taf. 11, fig. 1, Elgg, Canton Zurich, 

Switzerland. 
Turicius I uricensis ref., r.Mo. Conelets .J -13, trefoil rudiment. After von Meyer, 1867, Taf. 11, fig. 2. 
Turicius turicensis ref., r.Mo, r.M'-. Conelets 4-7. After casts of originals in the Munich Museum (1899 I.\ 1^ = 

.\mer. Mus. 22492, from Freising, Upper Bavaria, and 1S91 I o0 = .'\mer. Mus. 22491, from Moosburg, Bavaria, 

respectively). 
Turicius virgatidens types, r.Mo (H), r.M^ (H 1), 1.M-" (H 2, 3). Conelets 6-25. After von Meyer, 1867, Taf. iv, 

figs. 1-5, 'Gelben Lehmc' bei Fulda, northeast of Frankfort, Germany. 
Turicius atticus type, l.P'', M'. Conelets 4-5. After Wagner, 1857, Tab. vii, fig. 16, Pikermi, Greece. 
Turicius loahlhcimensis cotype, l.M'. Conelets compressed, 4-9, 5 ridge-crests. .After Kliihn, 1922, p. 77, fix- 17, 

Esselborn, Rheinhessen, Germany. 

Compak.\tive Observ.\tions (1935) 

Molars of Turicius are readily distinguished from those of the Mastodon and Zygnlophodon phyla by the following characters: 

First, by the accelerated sharpening and elevation of the transverse crests (.\-F). 

Second, by the accelerated multiplication of the conelets from six (A) to seven (C) to thirteen (F) to twenty-five (H), thus producing a lofty, 
sharpened crown parallel with that of Stegodon, with minute conelets more numerous than in Sttgohphodou or primitive species of Stegodon. 

Third, by the early reduction of the median sulcus, more apparent in Miocene species (A-F), vestigial in Pliocene species (H-H3), absent in 
specialized Turicius ua)ilheimensis (J). 



Lower Miocene 




A, 
B, 

c. 


Upper Miocene 




D, 
E, 

F, 
G, 


Milldll' (?) PlioCflK 


H 


H3, 


Lower Pliocene 




I, 


Pliocene 




J, 



STEGOLOPHODON 





ExUnuil 

S STEiCODONTOIDES 



TYP£ 



S LATIDENS TYPE 




S CAUTLEYI 



TYPE 



S. LATIDENS after CUft 








5 CAUTLEYI PR0CPES5US TYPE 




S.NATHOTENSIS 



S.SUBLATIDENS TYPE 

PLATE IV 

Stegolophodoxtdi.e, Stegolophodos: Primitive Losgitttdix.u. Sclcts. Shaded fA. B). Persistisg rs I-IV Anterior Ridge-crests (E, El), 

IX I-III Anterior Ridge-crests (F), Vestigial ix Ridge-crests I-III (G, H. I). Cont:lets Rocxoed, Increasing from 4-5 {.K, D), 

5-7 (C, H, I, G). Posterior Ridge-crests IV-VI PHoGREssn-E, with Coxelets 5-7 (G), Cont:lets not Exceeding 5 (H, I) 
Niio-Pliocene B, Stegolophodon cautUyi progressus tj-pe, r.M'-. Amer. Mus. 1&446, near Chinji Bungalow, India. Summit of Lower Chinji 
horizon, 2,000 feet above base of Lower Siwaliks. 
C, Stegolophodon nathoteruis t>-pe, r.Mj. Conelets 5-6. .\mer. Mus. 19455, near Xathot, Lower Middle Siwaliks, India. Lower 
Chinji horizon. 
Lower Pliocene D, Stegolophodon latidenM t>-pe, r.Mj. Ridge-crests I-VII, conelets 4, sulcus on ridge-crests I-IV. After Clift, 1828, PI. xixvin, 
6g. 1, near Yenangyaung, Burma. Lowest levels of Irrawaddy Series (Suviatile). 

E, Stegolophodon latidem t\-pe. r.M'. Ridge-crests I-\7, sulcus on ridge-crests I-IV. E 1 (section), ridge-crests coalescent at 

base (I-III). After cast (.\mer. Mus. 21978) of Clift "s t>-pe, 1828, PI. xxxvn, fig. 1, near Yenangyaung, Burma, lowest 
levels of Irrawaddy Series (flu%-iatile). 

F, Stegolophodon latident ref., r.M'. Sulcus on ridge-crests I-III, conelets 4-6, ridge-rrests 4^. After Matsumoto, 1936.1, PI. 

v, figs. 1 and 3 (Pro»tegodon), Shiwogama, Miyagi EHstrict, Province of Rikuzen, Japan. 

Pliocene (?) G, Stegolophodon lydekkerisp.nov..\y'i'- Ridge-crests I-VI, conelets 4-7, sulcus on ridge-crests I and II only. After Lydekker, 

1886.2, fig. 19 (as M. latidenji), Bome<j. 

Middle(?) Pliocene A, Stegolophodon tublatidens t>-pe. r.M'. Conelets 4-5. After Schlesinger, 1917, Taf. xvn, fig. 2, Teschen (Srhlesien), Austria. 

.Middle Pliocene H, Stegolophodon cautUyi t>-pe, r.M'. Ridge-crests I-V, sulcus on ridge-crests I and II, conelets 5. .\fter Lydekker, 1886.1, 
p. XV, fig. 6, Perim Island, India. 

rpper(?) Pliocene I, Stegolophodon itegodontoides Pilgrim, t>-pe, r.M'. Ridge-crests I-VI, sulcus on ridge-crests I-III, conelets 5. .\fter Lydek- 
ker, 1880, PI. XXXIX, Lehri, Punjab, India, possibly Upper Siwaliks. 

Comparattve Observatioxs (1935) 

Molan^ of Stegolophodon are readily distinguished from those of the Mastodon, Zygolopkodon, and Turieius phyla by the foUo«-ing characters: 

First, by the persistence of the median tulcut separating the inner and outer pairs of cones of all the crests (.\, B, D), of the three to four an- 
terior crests (E, F, I), of the two anterior crests (G, H). 

Second, by the rounded, bunoid conelets separated by median sulcus (.\-H). 

Third, by the closure of the enamel in the basse of the transverse valleys, as seen in section (El), very characteristic of Stegodon. 

The second and third of the Stegolophodon characters enumerated below link this genus with the genus Stegodon. But we must remember that 
Stegolophodon cautlet/i is of Middle Pliocene age (Perim Island), contemporary with the true Middle Pliocene Stegodon bombifrons (Dhok Pathan). 
Sec Volume II, Chapter XV, Stegodontoidea. 



THE MASTODONTIN^: HISTORY 135 

True Mastodon americanus Type in Eurasia. — It is interesting to note that in addition to the lophodont 
and zygolophodont forms related to Mastodon borsoni and M. turicensis, described by Pavlow and beautifully 
figured by her (1894), there also occurs in several locahties, e.g., Pestchana and Krasnoie, Podolia, Russia, 
the molar type which she rightly compares closely with that of M. ohioticus, i.e., M. americanus (see Y'lg. 81 of 
the present Memoir). Of these Russian specimens Marie Pavlow (1901, p. 12) speaks as follows: 

Jc vcnix profiter de roecasioii, puis(iu(! jo jiarle dp.s mastodontcs, pour I'oniplctor la description des restes fossiles du 
Mastodon Ohioticus, trouve a Pestchana en 1S92, et decrit j)ar moi en 1894 (/. cit.). lin parlant de la partie antdricurc de la 
mandihulo de cet animal, je me suis bornee h, dire (p. 10), qu'on ne trouve sur cc morceau de mandibule aucune trace d'alv6ole 
pour la defense. Cela me semblait suffisant pour fairc croire, que cet animal adultc n'avait pas de defenses inf(5rieures et qu'il 
pourrait etre rapporte par ce caractere au Mast. Ohioticus, les autres caractcres de dents venant appuyer cette determination. 
Mais ayant rencontr(5 dans la litterature qiiolqucs oljjcctions a ce sujet, jc me crois obligee de donncr ime description plus 
dctaillde de cette partie do mandibule, en I'accompagnant d'un dessin (PI. i, fig. 3). 

This description, with photographic figures (1894, PI. i, figs. 2 and 3), convinces us that we have from Pest- 
chana a form closely resembhng Mastodon americanus but which may not be specifically identical' ; the Pestchana 
specimens are certainly closer to M. americanus than to Zygolophodon borsoni. 

The fact that this recent review by Pavlow of the Eurasiatic Proboscidea yields examples of only four 
individual specimens of Mastodon americanus from two localities, Pestchana and Krasnoie, as compared with the 
relative abundance of Zygolophodon borsoni, is another proof of the great scarcity of members of the true Mastodon 
americanus phylum up to the period of its abundant revelation in the eastern forests of North America. As shown 
in PI. I, pi). 134-135, other specimens referred by Vacek to Zygolophodon prove to belong to Mastodon. 

4. RESUME OF DISCOVERY AND DESCRIPTION OF THE MASTODONTIN^ 

The earliest discoveries of the American Mastodon on the Hudson River in 1705, and those in Kentucky near 
the Ohio River in 1739, antedated by thirty years the dissemination of the various editions of the "Systema 
Naturae" (1735-1766) of Linnajus. From 1752 to 1778 Buffon described the "os d'^Iephans," as cited in full below, 
of this Kentucky animal without assigning to them a specific name ; subsequently Kerr 's description of Elephas 
americanus appeared (1792). Nevertheless these fossils attracted Uttle attention, in fact, even as late as 1797 
Blumenbach aflfixed the clumsy name Ohio-Incognitum to one of the fossil teeth found on the Ohio River, and as 
late as 1806 Cuvier named the same animal Le Grand Mastodonte and Mastodonte de I'Ohio. 

Consequently the history of proper nomenclature lags far behind the history of discovery, and the very last 
stage in the long progress of palaeontology is the ancestral history or phylogeny. 

Type Locality, Big-Bone Lick, Ky. — The type locahty of the first discoveries of Mastodon in America is 
certainly the Big-bone-swamp (fide Kerr, 1792, p. 116), subsequently known as the 'Big-Bone Lick,' a few miles 
southeast of the Ohio River, in Kentucky. Thus the animal took its second name Ohio-Incognitum Blumenbach, 
1797, from the Ohio River, and became known as Mammut ohioticum Blumenbach in 1799, Mastocbnte de I'Ohio 
Cuvier in 1806. 

Chronologic notes respecting early explorers of Big-Bone Lick were jjublished by William Cooper in 1831 
[Monthly American Journal of Geology and Natural Science, Vol. I, 1831, p. 159]. This list contains the state- 
ment that a French oflficer named Longueil secured fossil bones from a morass near the Ohio River in 1739, and 
that they were collected by some Indians and not by a Frenchman. The list goes on to say that Col. George 
Croghan visited Big-Bone Lick in 1765. 

'(1035) Made the type of Mastodon pavlowi (sec .\ppendix). 



136 



OSBORN: THE PROBOSCIDEA 



CuviER, 1806.2, p. 280, ix paht, and 1834, Vol. II, p. 266.— Le plus celebre de ces d6p6ts, celui qu'ont visite Longueil, Croghan 
et tant d'autres, celui qui a fait donner au mastodonte le nom d'animal de I'Ohio, porta lui-meme celui de Big-Bone-Strick ou 
Great-Bone-Lick. II est dans I'etat de Kentucky, a la gauche et au sud-est de I'Ohio, a quatre milles du fleuve, trente-six 
milles au-dessus de Tembouchure de la riviere de Kentucky [Footnote: ' Volney, Tableau du climat et du sol des fitats-Unis 
d'Amerique, I, page 100.'], presque vis-a-vis celle de la riviere dite la Grande Miamis. C'est un lieu enfonce entre des coUines, 
occupe par un marais qu'entretient un filet d'eau salee, et dont le fond est d 'une vase noire et puante. Les os se trouvent dans 
la vase et dans les bords du marais, au plus a quatre pieds de profondeur, suivant le rapport que nous en a fait feu le general 
Collaud, qui avait ete sur les lieux, et qui, en fouillant pendant trois jours seulement, avait recueilli vingt-quatrc morceaux. 
Leur abondance y est etonnante. Dej& Croghan croyait y avoir vu des restes de plus de trente individus; mais on en a recueilli 
depuis un bien plus grand nombre. 



Fig. 8.5. Big-Bone Lick, Boone County, Kentucky, 
from map in Anthony Finley's "A new American atlas," 
etc., folio, Pliiladelphia, 1826, No. 10 (Jde P. Lee 
Phillips, "A List of Maps of America in the Library of 
Congress," 1901, p. 838). 

This map is of great historic interest as showing one 
of the birthplaces of vertebrate palaeontology in America. 
Visited by the French ofBcer Longueil, mentioned and 
also fully described by Buffon, Blumenbach, and 
especially Cuvier, founder of vertebrate palaeontology. 
The Mastodon in the New World took the part played 
by the Mammoth in the Old World, in fact, it is re- 
markable that Elephas americanus Kerr (1792) antici- 
pates Elephas primigenius Blumenbach (1799). 



Genotype 

Locality : 

Big-Bone Lick, 

Kentucky 





tfS^»y V--^^ y \ V K N () L E TM N [ ' *< 


i 


1 


I- v"^ \y^x ,V- rw- N. , > if ^ ■■■■■': ^ 





In the following generic list only the chief specific names applied to the American Mastodon are included; a 
full Ust of this polynomial species is given below under the heading "Mastodon americanus" (p. 165). 



1705 
1714 



1739 



1778 



1. 


1792 


2. 


1797 


2. 


1799 


3. 


1803 


4. 


1806 


4. 


1806 


6. 


1808 


6a. 


1814 



CHRONOLOGIC ORDER OF DISCOVERY AND PRELIMINARY DESCRIPTION 

See Figure 86 



New York 



Big-Bone Lick, 
Ohio River 



(?) 



No. America 



Original Name Specific Reference in Present 

Memoir 
First mention of fossil remains near Albany, New York. 
First published account of two teeth and a thigh bone found 
thirty miles south of Albany, New York, on the Hudson 

River. 
First fossil remains found by Indians in the Big-Bone Lick, 

Ky., near the Ohio River, and given to a French officer 

by the name of Longueil; subsequently mentioned by 

BufTon, Blumenbach, and Cuvier. 
A single molar tooth from the Ohio River figured by Buffon 

(Fig. 84 B). 

Elephas americanus Kerr = Mastodon americanus 

Ohio-I ncognitum Blumenbach = " " 

Mammut ohioticum Blumenbach = " " 

Elephas macrocephalus A. Camper = " " 

Le Grand Mastodonte Cuvier = " " 

Mastodonte de VOhio Cuvier = " " 

Harpagmotherium canadense Fischer de Waldheim (cf. Sherborn, 

1924, p. 1022, " Harpagonotherium canadense, Anim. foss. 

Siberie.") = " " 

Mastodon Macrodon Rafine.sque = " " 



THE MASTODONTIN^: HISTORY 



137 



6. 


1S14 


4. 


1817 


4. 


1824 


7. 


1854 



[1835 



18G8 



8. 


1890 


9. 


1894 


10. 


1901 


11. 


1905 


12. 


1914 


13. 


1919 


14. 


1921 


15. 


1921 


16. 


1921 


17. 


1922 


18. 


1926 


19. 


1926 


20. 


1930 


21. 


1930 


22. 


1931 


23. 


1933 


24. 


1933 


24a. 


1934 


25. 


1935 


26. 


1935 


27. 


1935 



CHRONOLOGIC ORDER OF DISCOVERY AND PRELIMINARY DESCRIPTION (Continued) 

Original Name Specific Reference in Present 

Memoir 

Ohio River Maslolherium megalodon Fischer de Waldheim = Mastodon americanus 

" " Mastodon giganteum ( 'uvier = " " 

" " Madodon maximum (kivier = " " 

Canada Elephas Rujyertianua Richardson, Swan River, Lake Winni- 
peg basin, Canada = Mastodon americanus rupertianus 

Russia Mastodon podolicum Eichwald = Deinotherium podolicum (see Chap. IV, 

p. 85)] 
Oliio River Mastodon americanus Kerr [Lcidy]. First use of Kerr's 

specific name in America = Mastodon americanus 

S. Carolina Mastodon rugosidens Lcidy, Beaufort County = Mastodon americanus rugosidens 

Russia Mastodon ohioticus rcf., Pavlow, Podolia = Mastodon sp. (?). See No. 26 below 

Faydm Palieomastodon BcadneUi Andrews, Egypt = Palu'omastodon beadnellt^ 

" PaUvomastodon parvus Andrews, Egypt = Paln'omastodon parvus' 

Iowa Mammut progenium Hay, Harrison County = Mastodon progenius 

Hungary Mastodon (Mammut) americanus Pcnn. forma praetypica 

Sclilcsinger = Pliomastodon artiericanus praetypica 

Nebraska Mastodon niatthcwi Osborn, Sioux County. = Pliomastodon matthewi 

Nevada Mastodon merriami Osborn, Thousand Creek, Humboldt 

Co = Miomastodon merriami 

Hungary Mastodon tapiroides americanus (Schlesinger in Osborn) . . . = Miomastodon tapiroides americanus 

FayCim Palieomastodon intermedins Matsumoto, Egypt = Pahromastodon intermedius' 

Illinois Mastodon americanus plicatus Osborn, Walnut = Mastodon americanus plicatus 

Oregon Mammut (nrgonense H.ay, Baker ( 'oinity = Mastodon oregonensis 

Florida Pliomastodon sellardsi Simpson, Brewster = Pliomastodon scllardsi 

California Pliomastodon vexillariusMniXhevi, VresnoConnXy = Pliomastodon verillarius 

Nebraska Mastodon moodiei Barbour, Milford = Mastodon moodiei 

New Mexico Mastodon raki Frick, Hot Springs - Mastodon raki 

Alaska Mastodon americanus alaskensis Frick, near Fairbanks = Mastodon americanus alaskensis 

Nebraska Mastodon grangeri Barbour, near Pender, Thurston Co. . . = Mastodon grangeri 

France M iomastodon depereti sp. nov.,- Chevilly = Miomastodon depercti sp. nov. 

Ru.ssia Mastodon pavlowi sp. nov.,-' Pestchana = Mastodon padowi sp. nov. 

Indiana Mastodon acutidens sp. nov.,- Rochester = Mastodon acutidens sp. nov. 




Fie. S6. Gengraphic distribution (arcordinK to the numbers in the li.st above) of the principal s|)ecie8 of the Mastodontinae. The white dots within the 
black areas represent the approximate localities where the types were discovered. The white crosses represent referred s|>ecimens. Nos. '24a and 27 
omitted on this map. Nos. 25 and 2(1 are omitted here but appear on map (I'lR. 123a). See also geographic distribution map forming the inside cover, or 
front end-paper, of tlie present Volume. 



The above list of genera and species shows that there are many gaps in this great IMastodontinjp phylum to 
be filled by future discovery; also that while the Pleistocene Mastodon stage is well known, the Pliocene and Mio- 
cene stages of Pliomastodon and Miomastodon are known only from a few grinding teeth and tusks. The Oligo- 
cene stage, Palaeomastodon, is still very imperfectly known.' 

'See pages 143-149 below, also Palseomastodontinae in Appendix at close of the present Volume. 

'See Appendix at close of the present Volume for description of Miomastodon depereti sp. nov., of Mastodon pavlowi sp. nov., and of Mastodon acutidens sp. nov. 



5. PROGRESSIVE CHARACTERS AND GENERIC DISTINCTIONS OF THE MASTODONTIN^ 

Generic and Specific Succession. — As explained in the preceding section, the inclusion of the Lower 
Oligocene genus Palseomastodon} as a form related to the phylum Mastodontinae is provisional and depends 
upon our further knowledge of this relatively rare animal. While we are certain that the true Palseomastodon 
possesses a longer jaw but a broader and shorter skull than its Lower Oligocene contemporary Phiomia, while the 
superior grinding teeth are comparatively short and broad, and while in the crowns of the superior and inferior 
grinding teeth the proto- and metaconules forbid the ancestral relationship of Palasomastodon to Pliomastodon, 
Miomastodon, and Mastodon, we must await further knowledge of the cranium and of the cutting teeth, also of the 
incisors of Palseomastodon, before we can form a positive opinion on this very important and interesting question. 

Meanwhile we may contrast the characters of these four genera so far as they are known : 



Lower Oligocene of North 
Africa 

Pabsomaslodon Andrews, 1901 

Palate and grinding teeth relatively 
short and broad. 

Dental formula: IJ-f^ Dp? PfH 
M|-f. 

Ridge formula: P 2* P 3^71 P4* 
M 1^= M 2=-^' M 3'-p. 

Grinding teeth brachyodont, buno- 
lophodont, with proto- and meta- 
conules (see Fig. 93). No median 
sulcus. 

Rudimentary third crests in 'inter- 
mediate molars,' P2-M 1. 

Lower jaw and symphysis greatly 
elongated. 

Incisive tusks oval in section, with 
broad enamel band. 



Palseomastodon parvus. 
Palseomastodon intermedius. 
Palseomastodon beadnelli. 



Miocene and Pliocene (?) of 
Europe and of North 
America 

Miomastodon, Pliomastodon Osborn, 
1922, 1926 

Palate and grinding teeth somewhat 
elongated. 

Dental formula: I t^i^ Dp? P? 

Ridge formula: Probably P 4f M if 
M2fM3,-^n^. 

Grinding teeth with more elevated 
lophs. A median sulcus; no conules. 



Fully formed third crests in 'inter- 
mediate molars. ' 

Lower jaw and symphysis mod- 
erately elongated. 

Incisive tasks rounded, with enamel 
band (Miomastodon), without enamel 
band (Pliomastodon). 

Pliomastodon malthewi. 

Pliomastodon americanus praetypica. 

Miomastodon tapiroides americanus. 

Miomastodon merriami. 



Pleistocene of Eurasia and 
OF North America 

Mastodon Cuvier, 1817 

Palate and grinding teeth relatively 
elongated. 

Dental formula: I S^M Dp M 
P^^^ (vestigial) M fff • 

Ridge formula: Dp 2| Dp 3| Dp 4f 
M U M 24 M 3 



^ "^ " 4 3/4- 5* 

Grinding teeth subhypsodont. 
median sulcus ; no conules. 



Third crests fully formed in 'inter- 
mediate molars.' 

Lower jaw and symphysis ab- 
breviated. 

Incisive tusks rounded, without 
enamel band. 

Mastodon americanus. 
Mastodon americanus rugosidens. 
Mastodon americanus plicatus. 
Mastodon progenius. 



Dental Succession.— The relations of the deciduous premolars, Dp^(?), Dp3(?), of Palseomastodon in dis- 
tinction from the permanent premolars, P\ Fo{?), have not been observed. Nor do we know the relations of the 
deciduous premolars to the permanent premolars in Miomastodon or in Pliomastodon. 

According to the observations of Hays (1834) and of Warren (1852, 1855) cited below, the true Pleistocene 
Mastodon americanus agrees with Pleistocene and recent species of the Elephantidae in the suppression of the 
permanent premolar teeth; the fourth true superior and inferior premolars, P*-P4, form in the jaw but they do not 
erupt. In Elephas the corresponding permanent teeth, P'-Pi, are suppressed entirely. 

'See pages 143-149 below, also PalseomastodontinEe in Appendix. 

138 



THE MASTODONTINiE: GENERIC DISTINCTIONS 139 

PRINCIPLES OF GRINDING TEETH EVOLUTION, PLATES I, II, III, AND IV 

The succession and relationship of members of the true Mastodontinae phylum depend upon our very close 
examination and comparison of the structure of the grinding teeth, as displayed in the accompanying comparative 
figures. As first observed by Lartet, the grinding teeth of the true Mastodon americanus are fundamentally diflfer- 
ent from those of Mastodon borsoni which Vacek has rightly referred to Zygolophodon. 

Median Sulcus. — In the carefully drawn Plates i, ii, iii, and iv, illustrating the evolution of the grinding 
teeth in Mastodon, Zijgolophodon, Turicius, and Stcgolophodon, a median longitudinal sulcus separates the exter- 
nal and internal cones both in the superior and inferior molars. This demonstrates that the ancestral probo- 
scidean molar was tetrabunodont, as in Ma:ritherium, not hexabunodont, as in Palseomastodon. 

Ridge Formulae. — In the carefully drawn figures of the superior and inferior grinders of Palaeomastodon the 
construction of the ridge-crests of the grinding teeth should be examined with great care, because it is difficult to 
express the rudimentary condition of the third crest or tritoloph in a formula. It will be seen (Figs. 90B, 91, 92, 
93, 94, 95) that the tritoloph and tritolophid are in a formative stage; the tritoloph is rudimentary even in M^ (Fig. 
94D); it consists of a single cone in M', M^ (Fig. 91); the <n7o/op/«d is completely formed in Mi (Fig. 92B); it 
is less fully formed in Ma (Fig. 92B) . Consequently Palseomastodon is far more primitive than Mastodon in the for- 
mation of the tritoloph and tritolophid, yet it exhibits the ancestral stages in the development of this third crest so 
characteristic of all the Proboscidea. 

By close comparison of all the figures of the upper and lower grinding teeth of Palseomastodon, Miomastodon, 
Pliomastodon, and Mastodon, it is observed: 

(1) That the molar crowns in Palseomastodon are mainly tetrabunodont, i.e., each protoloph (superior) and 
each protolophid (inferior) is composed of a main external bunoid cone and a main internal bunoid cone ; in the 
superior molars (Fig. 94D) where the conules persist the main crown is hexabunodont. The presence of proto- 
and metaconules blocking the median sulcus forbids the ancestry of Palseomastodon to Mastodon. 

(2) The vestigial intermediate protoconules and metaconules are observed in the hexabunodont superior 
molars of Palseomastodon intermedins, thus the crested upper grinders are hexabunodont or six coned, whereas 
the lower grinders are subtrilophodont (Fig. 93, M^-Ms, Fig. 94). This primitive condition of the cones connects 
Palseomastodon with its undiscovered sexitubercular-quadritubercular ungulate ancestors; the conules observed in 
the third superior molar, M^, of P. intermedins (Figs. 93 and 89) are not seen in M^ of the same species (Fig. 92) ; 
the conules are vestigial or disappearing structures. 

(3) The protoconules and metaconules of Palseomastodon (Fig. 93) obstruct the sulcus of Mastodon. In ad- 
vanced stages of Palseomastodon and of Miomastodon the superior grinders, Uke the inferior, become trilophodont. 



(4) To the protoloph and -lophid also metaloph and -lophid the superior and inferior grinders, M'^^-M 



1-3 



add a rudimentary tritoloph and -lophid or third superior and inferior transverse ridge-crest (Figs. 92 and 94). 
The genesis of the tritoloph and -lophid is very clearly shown in Palseomastodon intermedins (Fig. 92) ; this tri- 
lophodont structure in all the superior and inferior true molars becomes a distinctive character of the grinding teeth 
in all the Proboscidea, except Deinotherium in which Dp 4 and M, are trilophodont (Figs. 54, 60, 71); conse- 
quently it is an ordinal character. 

(5) In Palseomastodon the fourth superior and inferior premolars exhibit the protoloph and -lophid 
and metaloph and -lophid only, i.e., no tritoloph and -lophid, like the fourth superior premolar of Phiomia 
(Fig. 179). 



140 



OSBORN: THE PROBOSCIDEA 



The fourth deciduous premolars of Phiomia, Dp*-Dp4, also acquire a tritoloph and -lophid. The fourth 
deciduous premolars of Miomastodon and Pliomastodon are not known. The fourth deciduous premolars of Mastodon 
americanus have a fully developed tritoloph and -lophid (see Warren, 1852, p. 213, PI. v). Consequently a 
trilophodont fourth deciduous premolar appears to be a progressive ordinal character of the Proboscidea. 

(6) Thus the third lophs, tritoloph and tritolophid, are successively established in M^-Ms, in M^-Ma, in M'-Mi, 
and in P^-P4 by the addition of a third crest. 



Fig. 87 



Amer, Mus. 18237 Type 





P. matthewl 



Amer. Mus. 19246 8 Ref. 

'■ Tia (rev.) 



outer views 
1/4 nat. size 




Amer, Mus. 13449 Paratype 



Palaeomastodon inlermedius Amer. Mus 14547 Type 

Fig. 88 



Fig. 87. Crown view of inferior dentition and jaw of Mastodon americanus, 
containing Dpj, Mi, M2, and M.5 anterior lophs. After Warren, 1852, PI. v, 
one-sixth natural size. Compare figures 150 and 151 of Zygolophodon borsoni. 

Fig. 88. Comparison of third inferior and superior grinding teeth of Plio- 
masiodon mallhciei (upper) and Palxomastodon intermedins (lower), one-fourth 
natural size. 



(7) The main internal and external bunoid cones in the Mastodontinse phylum, namely, in Miomastodon, 
Pliomastodon, and Mastodon, remain separate, they never unite transversely into a completed ridge-crest; thus all 
the Mastodontinse remain bunolophodont, they never become fully zygolophodont. 

(8) The passage from extreme brachyodonty to subhypsodonty is observed as follows: (a) In Palaeomastodon 
the paired cones or lophs are extremely brachyodont; (b) the cones or lophs in the three successive stages of 
elevation are seen in the ascending species of Miomastodon merriami, Pliomastodon matthewi, Miomastodon 
tapiroides americanus, and Pliomastodon americanus practypica, as described in the Miomastodon section; (c) 
the summits of the cones are converted into sharply acute ridge-crests, as observed in the unworn grinders of 
Mastodon acutidens (Fig. 135 and PL i, L). 

(9) Observe wide differences between Palasomastodon and Phiomia in coniiles and trefoils: (a) In Palseo- 
mastodon the central conule (Fig. 90) is arrested, proto- and metaconules close the sulcus (Fig. 93); (b) 
in Phiomia the central conules progress strongly (Fig. 90); in the inferior molars of (c) Miomastodon and 
Pliomastodon the ectoconelets expand into trefoil spurs (Fig. 89); in (d) Mastodon americanus rudimentary 
trefoil spurs are observed on ectoconelets below and entoconelets above, the mesoconelets being reduced. 

Summing up, the progressive 'Mastodontinse' characters are as follows: (1) Early loss of the conules; (2) 
very gradual development of the tritoloph; (3) relatively rapid development of the tritolophid; (4) persistent 
separation by a sulcus of the internal cones from the external cones, hence bilobed; (5) early development of the 
internal cingulum on the superior molars, retarded development of the external cingulum on the inferior molars. 



M. americanus 

Amer. Mus. 14293 Ref. 

I female) 




f.ms 



M. amencaniis 
Amer. Mus. 14294 Ref. 




I p. matthewl 

\ Amer. Mus. 18236, 16239 
I I Paratyoes 




i?^'7R>^ 



1 M. mernami 

« Colo. Mus. 92 Type 

Drawn from cast of Type Amer. Mus. 14471 







I. mi mj (rev.> 



Palaeomsstodon Intermedius 
Amer. Mus. 13449 Paratype 



\ Palaeomastodon intermedius 
\Amer. Mus. 14547 Type 





Fig. 89. Evolution of the Molar Riuge-crests in the Mastodonts.Oliqocene to Pleistocene 

Compare Pi.s. i-iv, pp. 134 135 



Inferior Series 
Mastodon americantis, r.Ma; r.Mn with 4/i ridge-crests. 
Plujniastodon americanus praelypica, r.Mj; r.Ms with 5 

ridge-crests. 
PUomaslodon mallhcm paratypes, M.i with i)h ridge-crest-s. 
Miomastodun merriami type, r.M.i with 4)5 ridge-crests. 
I'alaiomnsUxlon inCermeditis, I.M3 with 3 ridge-crests. 



Superior Series 
Mastodon americamis, tM-; r.\P with 4!'3 ridge-crests. 
Pliomastodon americanus praelypica, r.M-; r.M'^witli 4>i 

ridge-crests. 
Pliomastodon mallhewi type, r.M^ with 873 ridge-crests. 

PaLromastodon intermedius, r.M' with 2)2 ridge-crests. 

Observe: (1) Inferior molars more progressive than superior molars; (2) a) anterior ridge-crests (protoloph, -lopliid) rela- 
tively constant, b) second ridgc-cresls (mctaloph, -lophid) progressively broadening, c) third ridge-crests (tritoloph, -lophid) 
progre.s-sively broadening, il) fourth ridge-rrests (tctartoloph, -lophid) progressively broadening, e) fifth ridge-crests (pentaloph, 
-lophid) more progressive in M.i than in M''; (3) third inferior molar elongate, gently narrowing posteriorly, third superior molar 
less elongate, more rapidly narrowing posteriorly; (4) absence of median trefoil spurs in PaLtumaslodon, presence of external 
trefoil spurs in Miomaslodon and Pliomastodon- (5) rudimentary eclotrcfoil spurs in Mastodon inferior molars. Observe (6) 
that in this figure the grinding teeth of all tlie true Mastudontiiia- exhibit a median siikux and four primary conelets on each 
ridge-crest; also a rudimentary trefoil .•<pur extending from the entoconelets if the sui>erior molars and from the ecto- 
conelets of the inferior molars. (7) In PahKimasiodim intermedius the proto- and mctaconules dose the median sulcus. 



141 



142 



OSBORN-. THE PROBOSCIDEA 



Final Ridge Formulae. — The ridge formula of the deciduous and permanent teeth of Mastodon americanus 
is as follows (see Warren, 1852, pp. 64-73) : 



Lower Jaw 

Dp2 small, two ridge-crests, bilophodont. 

Dps larger, two ridge-crests and talon, bilophodont. 

Dp4 three ridge-crests, cinguliun, trilophodont, six fangs. 

Ml larger, three ridge-crests. 

Mg three ridge-crests. 

Ms four ridge-crests, with complex talon, tetralophodont. 



Upper Jaw 

Dp^ (?) two ridge-crests, bilophodont. 

Dp' two ridge-crests, bilophodont. 

Dp'' three ridge-crests, trilophodont. 

M' three ridge-crests, trilophodont. 

M- three ridge-crests, trilophodont. 

M' four ridge-crests, and small talon, tetralophodont. 



The ridge formula of Mastodon americanus appears to be as follows: 

Dp 2 f Dp 3 * Dp 4 f M 1 f M 2 I M 3 



4 + 



The dental formula of Mastodon americanus accords with the observations of Hays (1834) and of Warren 
(1852, p. 63). It would appear from Warren's description that the fourth permanent premolar, P*-P4, does not erupt 
in Mastodon americanus. It is known in Serridentinus productus to be a bilobate tooth. Thus the final true 
Mastodon stage in the grinding teeth of the Mastodon tinae appears to give us the following dental formula : 

Deciduous premolars = Dp^"*-Dp2.4. 

Permanent premolars = P4 in germ, P2, P3 suppressed. 

Permanent molars = M'"^-M,.3. 



Total number of grinding teeth which come into use in both jaws = 24 



Phiomia osborni 
Amer. Mus. 13468 Type 




Palaeomsstodon beadnelli 
Drawn from cast of type Amer Mus. 9984 



Mandibular Contrasts: Phiomia (upper), PaL/EOMastodon (lower) 
Fig. 90. Superior view of largest known jaws of (A) PAiomia and (B) Palseotnastodon superposed for comparison. One-fourth natural size. 

A, Typn jaw of /■'/iiomia osborni Matsumoto (Amer. Mus. 134G8). Observe; (1) The relatively long, narrow grinding teeth; (2) the ex- 
tremely elongate symphysis extending back between right Pa and left P.-;; (3) an ancestral s ho vfl-tusker (cf. Amebelodon); incisive alveoh elongate. 

B, Type jaw of PaUeomasUjdon beadnelli Andrew.s (Cairo Mus. C. 10014), drawn from cast of type (Amer. Mus. 9984). Observe: (1) The 
relatively broad inferior molars; (2) the abbreviated inferior grinding series; (3) the extreme anterior position of mandibular symphysis; (4) 
incisive alveoli rounded, abbreviate. 



6. CHARACTERS OF PAL^OMASTODON AND ITS INCLUDED SPECIES 

Superfamily: MASTODONTOIDEA Osborn, 1921 
Family: .MASTODONTID.E Ciirard, 1852 
Subfamily: Pal/eomastodontin^ siibfam. nov. 

Osborn, 1934: The genus Palseomastodon was originally placed by Osborn. following Matsumoto, in the 
family Mastodontida?, subfamily Mastodon tirup, on the basis of the broad similarity of its grinding teeth to those 
of Mastodon. Recent and more intensive observation compels Osborn to remove Palaeomastodon} from its supposed 
ancestral relationship to Mastodon or to any other genus of the Mastodontina;. We must await further knowledge 
of this relatively rare animal before we can determine its phyletic relationships. 

Genus: PAL/EOMASTODON Andrews, 1901 

Original reference: Zoologist, 1901, (4), V, Aug. 1."), pp. 318, 319, (Andrews, 1901.1); Tageblatt des V Internat. Zool.-Cong., 
Berlin, No. 6, Aug. 16, p. 4. (Published volume, Verhandlungen, 1902, p. 528). (Andrews, 1901.2). 
Genotypic species: Palxomastodon Beadnelli (Andrews, 1901.1, pp. 318, 319). 

Generic Characters.— (Matsumoto, 1924.1, p. 3): "A genus of Probo.scidea. Skull imperfectly 
known. Palate rather short and wide, as compared with that of the next genus [Plnomia] ; judging from 
the form of the palate this genus might be less long-skulled than the next one. Mandible elongated 
anteroposteriorly; mandibular synii)hysis rather short as compared with that of Phiomia; posterior end 
of .symphysis lying a con,siderable distance anterior to the first cheek tooth (P3) ; largest and most con- 
spicuous one of mental foramina Iving just below the first cheek tooth, and far behind the posterior end 
of the symphysis. Dental formula: It. CS. Pf. Mi Ridge formula : Dm ■ H- PrF- [M',",']-' Cheekteeth 
markedly brachyodont; last premolars and all molars short and wide, bunolophodont, wearing like 
typically lophodont teeth, attaining rather sharp ridges and very widely open valleys when moderately 
worn; a rudimentary intermediate cusp is pre.sent in the anterior valley of each lower molar; no trefoil 
pattern of cusps; surface of enamel rather smooth; basal cingula neither very strong nor very rough." 

As pointed out clearly in Chapter II, Sections II-I V, the genus Palaeomastodon, based on the genotype P. beadnelli 
and now amplified by our present knowledge of the species P. parvus Andrews and P. intermedius Matsumoto, 
represents a group of species entirely distinct from the Phiomia group. Inasmuch as Andrews' final definitions 
(1906, p. 130) both of the genus Palaeomastodon and of the family Palaeomastodontidae based on this genus are 
entirely founded on the generic characters of Phiomia, we are compelled with regret to abandon Andrews' defini- 
tion both of the genus Palseomastodon and of the family Palo'omastodontida^ and to substitute, with a few omis- 
sions and slight modifications, the definition of Matsumoto which is founded on a correct diagnosis of the characters 
of Pal3eomastodon so far as known. In Palxomastudon the lower incisors are apparently short and rounded; in 
Phiomia they are long nndjlattened. 

The above generic characters are confirmed and amplified in the three ascending species included within this 
genus and distinguished by measurements as follows: 

Palseomastodon parvus Palxomastodon intermedius Palaeomastodon beadnelli 

Andrews, lOOri Matsumoto, 1922 Andrews, 1901 

Lcufilli of lower molar series, M1-3, Len}.^tli of lower molar series, M1-3, Length of lower molar series, M,^, 

130-i;il nun.; of lower premolar-molar 159-101 mm.: of lower premolar-inolar 191-194 mm.; of lower premolar-molar 

series, Pj-Mj, 1G9 mm. series, P3-M3, 19()c mm. series, P3-M3, 279 mm. 

Rarity of Palxomastodon Remains. — The relative rarity of specimens referable to the three species of 
Palseomastodon is very significant indeed ; in the collections of the British, Cairo, and American museums com- 
bined the total number of specimens of teeth and jaws referable to these three species may be approximately 
enumerated as follows: 

'(103.')) Now pl.iocd in tlie now subfamily PAL.toMA.'iTonoNTiN.E (spp .\ppon(lix). 

■Till' Miitsunioto riilRc formula (1924.1, p. .3) of Pnhinnmstoihu is apparently » lapsus calami or a misprint. (The ridge formula in Palseomastodon 
beadmlli and /' intermedius is: M 1 '~r^ M 2 Ht" M '-^ '-^"-1 

143 



144 



OSBORN: THE PROBOSCIDEA 



Palseomastodon parvus: 4 specimens Brit. Mus., 1 specimen Amer. Mus., total 5 

Palseomastodon intermedius: 4 specimens Amer. Mus 4 

Palxomastudon beadnelli: 2 specimens Cairo Mus., 2 specimens Amer. Mus., total _4 

Grand total skull, tooth, and jaw specimens certainly referable 13 

In the American Museum collection there are seven specimens of Palseomastodon and seventy-nine specimens of 
Phiomia, the relative frequency of Palaeomastodon to Phiomia being as one to eleven. 

Osborn, 1934: We observe (Figs. 91-93) that the progressive third superior and inferior molars of Palseomasto- 
don intermedius exhibit the presence of proto- and metaconules, hence hexabunodont; these conules block the median 
sulcus which distinguishes the genera Mastodon, Zygolophodon, Turicius, and Stegolophodon (as shown in Pis. 
i-iv, pp. 134-135). These characters remove Palaeomastodon from the ancestry of the Mastodontinae. 



Paraty^e /J4^9 





Fig. 91. Type and paratype specimens of PaUeomastodmi. intermedius Matsunioto, 1922, one-fourtb natural 
size. Drawings prepared under the direction of Dr. H. Matsunioto. 

B, Type of Palsomastodon intermedius Matsunioto, 1922, Amer. Mus. 14547 (reversed in drawing). 151, 
Crown view of the same type specimen (reversed in drawing). 

A, Paratype palate of Palseomastodon intermedius (Amer. Mus. 13449); P', P' drawn in from Amer. Mus. 
14548 (reversed). Al, The same paratype specimen, palatal view. 

outer side 




Fig. 92. Type and paratype 
lower and upper grinding teetli 
of PaUeomastodon intermedius 
Matsumoto, 1922. Both figures 
two-thirds natural size. 

B, Crown view, left lower mo- 
lars of type (Amer. Mus. 14S47). 

A, Crown view, left upper den- 
tition of paratype (Amer. Mus. 
13449); P^ P' drawn in from 
paratype specimen (Amer. Mus. 
14.148). 



Palaeomastodon mtermedius 
Amer Mus 13449 Paratype 



B 




Palaeomastodon intermedius 
Amer. Mus. 14547 Type 



2/3 nat. size 



Palaeomastodon intermedlus 
Amer. Mus. 13449 Paratype 



Palaeomastodon intermedlus 
Amer. Mus. 14547 Type 



Di 



mefdcone 
metaconule •/ f 



upper m.3 
extBrnal 



r.m.3 




meiaconic/ 



fiypocone 



'^proiocone 
interna/ cingu/um Natural size 



profoconid 



hypoconid 



CBntral conule 



Fig. 93. Key to the he.xabunodont upper and Imver molar crown.s of PalsomasUxton intermedins: Ma drawn from the type (Amer. 
Mus. 14547), M' from the paratype (Amer. Mus. 13449). See figure 92A, B, for same type and paratype. Natural size. 

Observe; (1) Relativehj broad proportions; (2) hexabunodont crown.s of M^, namely, protocone, protoconule, paracone, hypo- 
cone, metacone, metaconule; (3) fundamental arrangement in two transverse crests, i.e., protoloi)h, mctaloph; (4) two intermediate 
cusps, i.e., protoconule, metaconule; (.5) lower molar M3 h.'us rudimentary trilopliodont crown; (0) both upper and lower molars wholly 
distinct in proportions from those of any species of Phiomia; (7) aidnd conule rudiment in M.i; (S) absence of median sulcus. 



AM.I5898(:Qst) 





'^2/ 



An. 13437 



AM. 14547 




Fig. 94. Detailed studies of: Tnie PnUfomasiodon itUermediiis third superior and inferior hexabunodont (D, E) teeth, M'-Mj, 
compared with the e.s.sentially titrabunodont (.\, B, C) Afoeritheriuin teeth, superior and inferior. From originals and casts in the 
American Museum. All figures natural size. 

.\, Mwritherium lijonsi ref., right M'"" drawn from cast Amer. Mus. 15898. 

B, Masrithcriiim trigodon ref., right superior grinders drawn from .\mer. Mus. 13431. 

C, M n rilhcrium andrewsi ref., left Mj-.i drawn from Amer. Mus. 13437. 

D, Palaomastodon inlermedius paratyiie, third right su|)erior molar, r.M', Amer. Mus. 13449. 

E, PaUromasiodon inlermedius type, third left inferior molar, I.M3, Amer. Mus. 14547 (reversed in drawing). 

145 



146 



OSBORN: THE PROBOSCIDEA 



Confusion of Pal^omastodon beadnelli with Phiomia wintoni. — As Matsumoto remarks, the greater 
portion of Andrews' descriptions and all his text figures of Palxomastodon beadnelli, except those of the type 
jaw and teeth, belong not to this species but to Phiomia wintoni. All the principal specimens of skulls and upper 
jaws referred by Andrews to Palseomastodon beadnelli also appear to belong to Phiomia wintoni, except the 
specimen (Cairo Mus. C. 10014a) illustrated in his PI. xv, fig. 2, representing a fragment of a maxilla with the 
second and third molars, M--M^ in situ which appears to truly belong to P. beadnelli; these teeth are clearly 
shown to be bilophodont, a generic character of the Palseomastodon superior molars. 



Palseomastodon parvus Andrews, 1905 

For original description and type figure, see pp. 59, 60, of the present 
Memoir. 

Fluvio-marine formation of tlie Faj-flm, Egypt = Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

This is the smallest and most primitive Palseomastodon known; 
it is probably of much greater geologic age than P. beadnelli, the 
first species discovered. The full characterization of this species 
in the systematic revision of Chapter II above and the compara- 
tive measurements as given by Matsumoto and corrected by Osborn 
leave little to be added. It is a relatively rare animal because it is 
known at present from only five specimens, namely, the type 
(Brit. Mus. 8479a) and three other specimens in the British Mu- 
seum, also referred Amer. Mus. 13497. 

Specific Charactkr.-s. — The specific characters cited in 
Chapter II from Andrews (1905, p. 562) and Matsumoto (1922, 
p. 2, 1924.1, p. 4), together with corrected measurements by Osborn, 
may be summed up in the following sentence : P. parvus is distin- 
gui.shed by (1) M1-3 131 mm. as compared with 161 mm. in P. 
inlermedius and 194 mm. in P. beadnelli, (2) P3-M3 169 mm. as 
compared with 190e mm. in P. intermedius and 279 mm. in P. 
beadnelli, (3) molar proportions of M1-3 substantially similar to 
those in P. intermedius and P. beadnelli. 

Palseomastodon intermedius Matsumoto, 1922 

For original description and type figure, see pp. 63, 64. 
Fluvio-marine formation of the Fayum, Egypt = Fpper Eocene of 
.\ndrews = Lower Oligocene of the present Memoir. 

.\s the name intermedius indicates this animal is intermediate 
in size between Palxomastodon parvus and P. beadnelli; it is more 
progressive than the former species in its molar teeth, the molars 
being relativelj' narrower. It is known from four specimens only. 

CiEOLOGic Level of Four Specimens. — The quarry records 



Specific Characters. — Osborn confirms Matsumoto's separa- 
tion and definition of this species and adds the following characters 
which are clearlj- displayed in figures 92 and 94: (1) The hexa- 
bunodont structure of M^, less strongly indicated in M^ which is 
tetrabunodont; (2) the sublophodont arrangement of the cones and 
eonules both in the superior and inferior molars; (3) the arcuate 
arrangement of the superior and inferior molars of opposite sides. 
In addition to the specific characters quoted in full from Matsu- 
moto in Chapter II (Matsumoto, 1922, p. 2, 1924.1, p. 10, and e.\- 
teuded bj' Osborn), it may also be pointed out that the grinding 
teeth of P. intermedius, besides being of smaller dimensions 
throughout, are relatively narrower than those in the t3'pe of P. 
beadnelli, namely : 

M, M.. Ms 

Palxomastodon beadnelli, type 

indices 77 

Palxomastodon intermedius. 

type indices 64 



68 



66 



62 



It may be premature, however, to establish these indices as 
specific characters until we secure more material. 

Phylogenetic Relationship. — It would also be premature to 
say whether Palseomastodon parvus, P. intermedius, and P. bead- 
nelli constitute an ascending series of species increasing in size, 
losing certain primitive characters and acquiring certain progres- 
sive characteis. We must await further evidence which will be 
afforded, first, by the discovery of the complete dental and cranial 
characters of these primitive mastodonts, and, second, by the 
determination of the geologic levels in which they occur. Compar- 
ing Palseomastodon with Phiomia the same question will arise in 
respect to the latter genus, namely, whether the smaller, inter- 
mediate, and larger species constitute an ascending series. Tenta- 
tively the two lines may be arranged as follows: 



Larger, more progressive 
Intermediate in character 
Smaller, more primitive 
Smaller, more primitive 



Palseomastodon beadnelli 
Palxomastodon intermedius 
Palseomastodon parvus 



Phiomia osborni 
Phiomia wintoni 
Phiomia serridens 
Phiomia minor 



show that the level of the type (Amer. Mus. 14547) is not recorded; 
one paratype (Amer. Mus. 13480) is from Quarry B, another 
paratype (Amer. Mus. 13449) is from Quarry .4, while a third, a 
ma.xilla (Amer. Mus. 14548), is not recorded. As shown in Chapter 
II, systematic revision of Palseomastodon intermedius above, there 
is no means of ascertaining the exact geologic levels of the different 
species of Palseomastodon: all that we can say is that the species 
P. intermediuji is well represented by the four specimens in the 
American Mu.seum collection, clearly illustrated in figures 39, 
91-94. 



Cranial Proportions. — The cranium of Palseomastodon 
intermedium (Amer. Mus. 13449) constitutes the only evidence we 
have at present as to the proportions of the skull in Palseomastodon 
(see Figs. 91, 39). These proportions are more brachycephalic 
than those of Phiomia, as evidenced (1) by the relatively widely 
arching palatal series, (2) by the harmonically broader and shorter 
grinding series, and (3) by the more widely arching zygomata. 
We may expect to find Palseomastodon with a broader and a shorter 
skull than that of Phiomia, with a large pair of upper and lower 
incisive tusks, probably rounder than those of Phiomia, with a broad 



PAL^OMASTODON 



147 



enamel band on the concave outer surface of the upper tusks, as in (Matsunioto, 1924.1, p. 9), as well as tlie following new compara- 
the Miocene species of Miomastodo7>. tive measurements by Osborn: 



Dental Measurements and Indices in the Three Species ok Pal^omastodon 






M 1-3 




Ml 




M2 


M3 


P3-M3 


Inferior Molars 




ap. 


tr. 


I. 


ap. 


tr. 


I. 


ap. 


tr. 


I. 




Palseomastodon beadnelH type, Cairo Mus. C;. 10014 


M,.M3=194 


48 


37 


77 


65 


51 


78 


78 


53 


68 


P3-M3 = 279 


Amer. Mus. 13481 
















81 


45 


56 




" intenmdixis type, Amcr. Mus. 14547 


M, -Ma =161 


42 


27 


64 


53 


35 


66 


65 


40 


62 


P3-M3 = 190e 


" " paratype, Anier. Mus. 13480 
















60 


37 


62 




" parvus type, Brit. Mus. 8479a (cast Amer. 
























Mus. 9976) 


M,-M3=131 


41 


28 


68 


44 


32 


73 


43 


30 


70 


P3-M3 = 169 


" Amer. Mus. 13497 
















52 


33 


63 




Superior Molars 
























Palseamastodon inlermedius paratype, Amer. Mus. 13449 


1. M'-M»=150 


42 


34 


81 


50 


42 


84 


57 


46 


81 


P'-M» = 200e 




r. M'-M'=149 


43 


35 


81 


50 


43 


86 


57 


45 


79 


P»-M' = 200e 



Dental Characters. — The dental characters in P. inler- 
medius are beautifully illustrated in figures 93, 94, and especially 
in figure 92, showing the detail of the upper and lower grinding 
teeth. We observe that both upper and lower grinders are progress- 
ing towards trilophodonty; at the same time the three lobe? are 
showing a tendency to unite into three transverse crests, so that 
the term trilophodont exactly describes them. This dental struc- 
ture does not sustain Matsumoto's generalization that the genus 
Palxomastodon includes the ancestral forms of Mastodon. 



Palsomastodon beadnelli Andrews, 1901 
For original description and type figure, see p. 54 

Fluviu-niarine formation of the KayUm, Egypt = Upper Eocene of 
Andrews = Lower Oligoccnc of tlu^ present Memoir. 

This is important as the largest and probably the most pro- 
gressive, as well as the first discovered and genotypic species of 
Palxo7nastodon. 

GEOLOciic Level. — Andrews (letter, 1922) informs the writer 
that the large type jaw (see original type figure, I'^ig. 26) was 
found at the very base of the Fluvio-marine beds, 50 or 100 feet 
below the typical Phiomia level; this genotype jaw (Cairo Mus. 
C.10014),with a very large referred femur (Cairo Mus. C110017), 
tibia (C.10015), scapula (C.10016), humerus (C.10013), and axis 
vertebra (C. 10061), was the only true P. beadnelli material found 
at this locality. This referred limb material (Fig. 96) is very 
important because of (1) its primitive Mastodon-Xike characters 
and (2) its use in estimating the size and limb proportions of 
Palxomastodon, as shown in Andrews' restorations. 

Specific Characters of P. beadnelli. — Th(; .specific char- 
acters are shown in the second type figure by Andrews (Fig. 95), 
in the type jaw as refigured by Osborn in comparison with Phiomia 
osborni (Fig. 90), al.so in Amer. Mus. 13481 (Fig. 187, PI. i, p. 134), 
a right lower third molar, M3, attached to a small fragment of a 
mandibular ramus from Quarry B, measuring 81 mm. in length 
and 45 mm. in width. See Matsumoto's definition in the systematic 
revision above; also his detailed comparative measurements 



Significant Molar Indices.— The dental projjoitions of 
Palieomastodon are shown in the molar indices M'-Mi, M--M2, 
M'-Ms as above. The molar indices (I.) in the three species of 
Palxomastodon aie printed above as showing the ratio of transverse 
(tr.) to anteroposterior (ap.) measurement; although greatly 
modified by crushing of the teeth (as for example in comparison of 
Amer. Mus. 13481 with Cairo Mus. C. 10014) they nevertheless 
serve to distinguish Palitomastodon very clearlj' from Phiomia, as 
shown in the following contrast: 

Ratios (I.) of Breadth (tr.) to Lencjth (ap.) in Palieo- 
mastodon and Phiomia 

Mi M., M, 
Palseomastodon beadnelli type indices 77 78 68 
Phiomia wintoni average 08 62 56 




Fig. 95. Second type figure of Palxomastodon beadnelli .Vndrcw.s, 1901. 
After Andrew.s' "De.scriptive Catalogue of the Tertiary Vertebrata of the 
Fayvim, Egypt," PI. xv, figs. 1, Ia, 1906. Original tyiw in Geological Museum, 
Cairo (C. 10014). One-sixth natural size. 

From this comparison we see that whereas in Sf)ecies of 
Pahomastodon the l)readth of Mj is from 62*^ to 70'^^ of the 
length, in species of Phiomia the breadth of M3 is from 54% to 



^ 


S 


m 









r3 


■a 




r> 








S 


_D 


J 


i! 


a 








> 






•3 


p 




< 




PAL^OMASTODON 



149 



63% of the length. Thus the indices in r'nhrnmastodon of 62% to 
70% are uniformly broader than those in I'hiomia of 54% to 63%. 
These molar indices therefore are very profound and significant 
characters, because the molar indices of Falsfomnstndon point 
towards those of Mastodon; the molar indices of Phiomia point 
towards those of Trilophodon. For example, in Serridentinus 
produclus (Amer. Mus. 10.582) the third molar indices are 
47-60, the proportions of M3 being about the same as in Phiomia 
wintoni. 

In Palseomnstodon the lower molars are relatively shorter and 
broader; in Phiomia they are relatively longer and narrower. 
There is a less wide contrast in f-he upper molars. 

Characters of the Skeleton. — Andrews (1906, pp. 145- 
147) observes of the only bones found in the same locality as the 
type of P. beadnelli as follows : 

"Fore limb. — An incomplete left scapula collected in the same 
locality as the type mandible is the only evidence of that bone yet 
discovered. It is very closely similar to the scapula of Elephas, 
differing only in the rather slighter prominence of the coracoid 
process and in widening out less rapidly posteriorly above the 
glenoid cavity, the whole blade being probably rather less expanded 
above than in Elephas. The glenoid cavity is an elongated oval, 
the prescapular fossa is extremely narrow, and the process project- 
ing backwards from the middle of the spine in Elephas here seems 
to be situated a little nearer the glenoid end. In the case of the 
humerus (PL xvi. figs. 2, 2a) also the only specimen that can be 
definitely determined as belonging to Palxomaslodon beadnelli, 



was collected near the type mandible. In its general character 
this bone is like that of the later Proboscideans, but is somewhat 
stouter and broader in proportion to its length, and approaches 
more nearly some specimens of the humerus of Ma.'itodon with 
which it has been compared. ... A femur (PI. xvi. figs. 3, 3a) 
from the same locality is the only specimen of that bone in the 
collections of Cairo and London that is definitely Proboscidean in 
character and can be referred to Palxoma.itodon — a circumstance 
which still further emphasises the extraordinary rarity of bones of 
the skeleton of that animal. ... A left ((6m (PI. xvi. figs. 4, 4a) 
from the same locality is, on the whole, like that of the Elephant. 
It is, however, rather more slender in the shaft, and the distal 
articulation iliffers in several important points. . . . No bones of 
the foot were found with the limb-bones just described, but a 
single calcaneum (text-fig. 51), from some distance off, so nearly 
resembles in essential characters the calcaneum of the later Pro- 
boscideans, more particularly of the Lower Miocene Tetrabclodon 
angustidens, that it may be safely referred to Palseomastodon , and 
from its large size probably belongs to P. beadnelli." 

Andrews combined these materials, probably chiefly refer- 
able to Palxomaslodon beadnelli, with other bones probably refer- 
able to Phiomia wintoni to produce his reconstructions of the 
animal. It is probable that the skeletal parts which he referred 
to Palawuastodon parvus belong to the species Phiomia minor. 

Osborn's restorations of Palxomaslodon as a forest browser 
(Figs. 2, 97) arc entirely different both as to form and function from 
Osborn's restorations of Phiomia (Figs. 2, 17) !is an ancestral 
shovel-tusker. 




Restoration (1932) ov P.\l.B()M.\stoD().n .\.s ,\ Fokest Bkowsek 

By MarRR't Flin.scli, uiulor the direction of Henry Faiifield Osborn 

Skeletal parts after Andrews, lOOti, Pis. XV, xvi (FiR. 96 of the present Memoir) 

Fig. 97. The rarity of Paki nmiiMmhm hcadrwlli in the Fluvio-marine beds, as compared with the great abundance of Phiomia, 

indicates that this was a forest-living animal wliicli rarely entered the Phiomia habitat. Forest-liviiiR animals throiiKhout the entire 

Tertiary are not often found fossil in association either with plains-living, palustral, or amphibious forms. For limb proportions of this 

restoration, see figure 96 opi)osite. Osborn, 1934: The restoration of the primitive proboscis and of the single pair of lower incisive 

tusks may prove to be incorrect ; we await furtlier knowledge of the anterior portions of the upper and lower jaws and of the tusks. 





M. mcmami 

Colo. Mus. 92 Type 

Drawn from cast of Type Amer. Mus. 1447 




JiPJIoaJes. Schles, 
amcicanus 



All r4 nat. size 



Miocene and Pliocene True Mastodons, Miomastodon and Pliomastodon of Europe and America. Compare Pl. i, pp. 134-135 

Fig. 98. Comparison of the type molars of Miomastodon and of Pliottmstodon of the Miocene and Pliocene and of the true Mastodon americanus of the 
Pleistocene of America and Europe. After Osborn, 1921.522, fig. 1, except Al, which is after Frick, 1933, fig. 28. All figures one-fourth natural size. 

A, Type, r.M^ of Miomastodon [ = Plvotnastodon] malthe.wi Osborn, Snake Creek B, Pliocene, Nebraska. Amer. Mus. 18237. 
A2, A3, Paratypes of M. \=P.] matthemi. Same locality as type. Amer. Mus. 18238, 18239. 

Al, L.Ms of Miomastodon merriami rpf.. Pawnee Creek formation, Middle Miocene, Colorado (F:A.M. 23345). After Frick, 1933, fig. 28. 

B, Type, I.M3 and r.Ms, of Miomastodon merriami. Virgin Valley formation. Middle Miocene, Nevada, Colo. Mus. 92; cast Amer. Mus. 14471. See also 
figure 99 for superior molars and tusks of same type. 

C, Referred molar, I.M3, of Mastodon americanus, from the phosphate beds of South Carolina, in the American Museum (Amer. Mus. 9845). 

D, Dl, Ori<;inals after photographs by Schlcsinger. Types, I.M2-I.M3, l.M', Mastodon tapiroides americanus Schlesinger, Tasmtd, Usztat6 Kom., Hungary. 
Lower Pliocene. Compare Schle-^inger, 1922, Taf. xiv, fig. 1 [ = Miomastodon tapiVoidfs amencanu.s ot the present Memoir]. 

150 



7. CHARACTERS OF MIOMASTODON AND ITS INCLUDED SPECIES 

Superfamily; MASTODONTOIDEA Osborn, 1921 

Family: MASTODONTID^ Girard, 1852 

Subfamily: Mastodontin^ Brandt, 1869 — Osborn, 1910. 



Genus: MIOMASTODON O.sborn, 1922 

Original reference: Amer. Mus. Novitates, No. 49, p. 4 (Osborn, 1922.564). 
Genotypic species: Mastodon merriami Osborn, 1921. 

Generic Characters of Miomastodon (compare Osborn 1922.564, p. 4). — A genus of the 
Mastodontina^ apparently directly ancestral to Pliomaslodon and Mantodon. Skull and .'skeleton 
unknown. Dental formula incompletely known: I^^ Dp, P} Mjfg. Inci.sive tusks with broad enamel 
band, broadly suboval in section, downturned (M. merriami). Grinding teeth somewhat more elevated 
than in Palseomastodon; much less elevated (subhypsodont) than in Mastodon. Ridge formula prob- 
ably: P4^MllM2iM3 TT^jTi. Crests on the 'intermediate molars' probably fully formed, i. e., 
trilophodont. Internal and external lobes (i.e., cones) of each loph separated by median or longitudinal 
sulcus, as in Mastodon; summit of each lobe (i.e., cone) double or bifid (Hg. 98B) as in Mastodon, unlike 
Palaeomastodon in which there is no longitudinal sulcus. Each lobe transversely oval in horizontal section 
but less compressed anteroposteriorly than in Mastodon. Lower tusks rodlike, horizontal, vertically 
oval. 

The above generic characters are confirmed, amplified, and sUghtly modified in the four ascending species 
included within the genera Miomastodon and Pliomastodon and distinguished as follows : 



Miomastodon^ 



Miomastodon merriami Osborn, 
1921. Genotypic species 
(Colo. Mus. 92). 



Tusks concave on outer side, 
with broad enamel band, down- 
and out-turned. Lower tusks 
rodlikp, horizontal, vertically 
oval. 

M3 inea.suros 179 mm. antero- 
postoriorly, transverse 81} mm.: 
Mj-3 measure 10)2 in. ( = 269 
mm.). 



Middle Miocene of Nevada 
and Colorado. 



Miomastodon tapiroides amrri- 
ranus (Sehlosingcr in Osborn, 
1921). Type. (Fig. 98D, 
Ul.) 

Incisive tusks with broad 
enamel band on convex e.xternal 
surface. 



Hidge-crcsts brachyodont or 
subhypsodont, as in M. mcr- 
rinmi; less erect than in P. 
matthetvi; Mj-3 measure 284 mm. 
anteroposteriorly. 

Lower Pliocene of Hungary. 



Pliomaslodon 

^///wVfj Schjpsingor, 1919 1922 
Cotypes. (Fig. 103.) 



Inc-isivp tusks not (iotcriiilMpd. 



Pliomastodon 

amrriranus praf- Pliomaslodon matlhewi O.sborn, 

1921. Type (Amer. Mus. 

I82;i7); paratypes (Amer. 

Mus. 18238, 18239). (Fig. 

98 A). 

Superior(?) incisive tusks 
rounded, upturned, enamel 
band vestigial or absent. 

Lobes of grinders slightly 
more elevated or subhypsodont. 
Second inferior molar, Mj, meas- 
ures 85e mm., anteroposteriorly : 
third superior molar, r.M', 
measures 14.5e mm. antero- 
posteriorly. 



Estimated length of third 
superior molar, M', 144 mm., 
of third inferior molar, M3, 190 
mm. M*"'' measure 246 mm., 
Mj-3 mea-siu'e .304 mm. In 
size nearly efpial to molars of 
Mastodon americanus: M-"^ 
measure 31 1 mm., Mj-j measure 
330 mm. 

Middle Pliocene of Hungary. 



Ix)wer Pliocene of Nebraska. 



The genotypic species of Miomastodon is Mastodon merriami Osborn (Figs. 98B, 99) from the Middle Miocene, 
Virgin Valley formation, Nevada; it is distinguished by low-crowned grinding teeth, especially by the persistence 
of a broad enamel band on the outer side of the superior inci.sive tusks. The geologically succeeding species 
Mastodon matthem becomes genotypic of the new genus Pliomastodon, which entireh' lacks the enamel on the 
incisive tusks, as also probable in Mastodon americanus practypica Schles. ; moreover, the ridge-crest-s in Plio- 
mastodon are somewhat more elevated (compare Figs. 89, 98, 102, 394). 

The genotypic specimens (Fig. 98B, 99) from the Middle Miocene of Nevada are now reinforced by the 
more complete referred specimen (Fig. 100) recently discovered by Frick in the Pawnee Creek, Colorado, which 
fortunately reveals the lower jaw and tusks. 

'The nevs- and very primitive species Miomaslodon depereti of Uie I>uwer Mioci-ne of France is described in the .\piH'ndix at the close of the present 
Volume. See al.so comparative PI. i, pp. 134 13.5, also figs. I38A, and 230D. 

161 



152 OSBORN: THE PROBOSCIDEA 

Discovery of Miomastodon and Pliomastodon in Europe and America. — We owe to Schlesinger the dis- 
covery of several very distinctive specimens related to the true Mastodon from the Lower and Middle Pliocene of 
Hungary, to which he gave the name M. tapiroides americanus; original photographs of these specimens were 
kindly forwarded to the author of the present Memoir, who commented (Osborn, 1921.522, p. 2) as follows: 
"There cannot be the least doubt, however, as to the affinity [to the true Mastodon] of the grinding teeth found 
in the Lower Pliocene of Hungary, to which Schlesinger applies the name M. tapiroides americanus. These teeth 
are reproduced herewith (Fig. 1, D, Dl) from unpublished photographs, kindly forwarded by the author, to the 
same scale with corresponding grinders (A, Al, A2, A3) from the Lower Pliocene, Snake Creek formation, of 
western Nebraska, also with lower teeth (B) from the Middle Pliocene [Middle Miocene, Virgin Valley formation], 
near Thousand Creek, Nevada, and with (C) the posterior lower molar of M. americanus from the American 
Pleistocene." The figure alluded to in this citation (Fig. 1) is reproduced as figure 98 of the present Memoir, 
the lettering of the plate being unchanged, but with the substitution of M. merriami ref. (Al) for the erroneous 
paratype of P. matthewi. 

OsBORN, 1921. — In the same article Osborn (Osborn, 1921.522, p. 2) published under the name of Schlesinger 
the species "Mastodon tapiroides americanus Schlesinger" based, as described in detail below, upon "upper and 
lower grinders from the Lower Pliocene, Tasnad, Usztato Kom., Hungary," embracing "a third left superior 
molar (Fig. 1, Dl, see PI. xiii, fig. 5 [error, H. F. O.j, Schlesinger), also two left inferior molars, ma-mg (Fig. 1, D, 
see PI. XTV, fig. 1, Schlesinger). The linear measurement of the crowns agrees closely with that of the Pleistocene 
M. americanus, but the vertical measurement is apparently less, i.e., less hypsodont. This indicates that already 
in the Lower Pliocene the mastodonts had attained the massive proportions of their Pleistocene descendants. 
The lophs are similarly composed and show no trace of a trefoil ridge. There is nothing to debar these Lower 
Pliocene mastodonts of Hungary from the true ancestral line of our Pleistocene Mastodon." 

The type incisive tusks of Mastodon tapiroides americanus Schlesinger from the Lower Pliocene of Hungary 
agree with the type incisive tusks of Mastodon merriami from the Middle Miocene, Virgin Valley formation, near 
Thousand Creek, Humboldt County, Nevada, in the presence of a broad enamel band; as Osborn observed (op. 
cit., p. 4), "A very important character is the presence of broad enamel bands on the upper tusks (Fig. 2, C, D, E), 
which are perhaps similar to the enamel bands observed by Schlesinger in the true Miocene and Plio- 
cene mastodons of Hungary." 

In the same communication (pp. 4, 6, Figs. IB and 2) the species Mastodon merriami Osborn was founded 
upon two left lower molars and portions of the upper molars and of the two upper tusks (Colo. Mus. 92) discovered 
in April, 1909, in the Virgin Valley formation. Middle Miocene age, near Thousand Creek, Humboldt County, 
Nevada. These teeth, constituting the first evidence of the arrival of the true Mastodon in America, were discovered, 
as described by the geologist Mr. Richard C. Hills, in a formation consisting of more or less stratified volcanic ash 
containing much opalized wood. Mr. Hills presented his priceless specimens to the Colorado Museum of Natural 
History (Colo. Mus. 92) and they were loaned by Director Figgins to the present author and figured as the types 
of Mastodon merriami (op. cit., figs. 1,2). 

In Science, 1921 (Osborn, 1921.523), these discoveries were summed up as follows: 

If the Ma.'stodon merriami of Nevada proves to be of Middle Miocene age, it will demonstrate that the.se true 
mastodons came to this country much earlier than we have been led to suppose. 

I am greatly surprised and interested by the Middle Miocene appearance of the true mastodons in America, if the above 
report by Dr. Merriam is correct, as I have no doubt it is. Middle Miocene age is, in fact, quite consistent with the structure of 
the superior . . . tusks, which bear a broad enamel band on a concave outer side, a fact that puzzled me greatly because Dr. 
Schlesinger describes the Lower Pliocene mastodons of Hungary as bearing an enamel band on a convex outer surface. We 
should expect the earlier mastodons to show just the difference in the curvature of their tusks which these two observations 
would indicate. 



THE MASTODONTIN^: MIOMASTODON 153 

Middle Miocene of America and Eurasia. — A year later (1922) Osborn realized that the presence of a 
broad enamel band on the incisive tusks in these three species necessitated their separation from the true genus 
Mastodon, and he accordingly proposed (Osborn, 1922.564, p. 4) for the reception of these species the new genus 
Miomastodon which he defined as below. The genotypic species Mastodon merriami from the Middle Miocene, 
Virgin Valley formation, of Nevada, is the only specimen in which the incisive tusks were found in association 
with the grinding teeth, thus constituting an excellent type. 

The rarity of the true mastodonts in Eurasia and America is attributable to their forest-living habits; they 
occur rarely in Austria, as recently described by Schlesinger of Vienna. The Mastodon merriami of Nevada, as it 
proves to be of Middle Miocene age, demonstrates that these true mastodonts came to America much earlier than 
we have hitherto been led to suppo.se. Middle Miocene age is quite consistent with the structure of the superior 
incisive tusks, which in M. merriami bear a broad enamel band on a concave external surface, whereas Schlesinger 
describes the Lower Pliocene mastodont M. tapiroides americanus as bearing an enamel band on a convex outer 
surface. We should expect the early true mastodonts to show just the progres.sive differences in the curvature 
and enamel of the tusks which these two observations would indicate. The species Mastodon merriami from the 
Middle Miocene of Nevada is apparently much more primitive than the species Mastodon matthewi from 
the Pliocene of the Snake Creek formation of Nebraska. 

Osborn, 1922.— Osborn's formal description of the genus Miomastodon was as follows (Osborn, 1922.564, p. 4) : 

Miomastodon, new genus 

Genotypic Species. — Manlodon merriami Osborn, 1921 . . . from the Virgin Valley formation, Middle Miocene of Nevada. 

CIeneric Characters. — A member of the true Mastodontina? phylum leading into the Mastodon americanus type, distin- 
guished from the true Pabeomaslodon bcadnelii of the Lower Oligocene of the Fayuin, F^gypt, by rounded, greatly enlarged, up- 
curved superior tusks ; form of inferior tusks not certainly known, probably rounded and more or less encased in enamel ; distin- 
guished from the true Pleistocene Mastodon americanus by the presence of a broad enamel band extending from the base to the 
summit of the tusk. Ridge formula: M 2 f , M 3 ^-^, as compared with the Pala-omastodon beadnelli ridge formula: M 2 ~^, 
M 3 "-j^l fis compared with the Mastodon americanus ridge formula: M 2 |, M 3 j-Jt-,. 

To this genus may at present be referred four species, namely : 

Genotypic species, Mastodon merriami, Middle Miocene, Nevada = Miomas<orfo7! merriami. 

Mastodon proavus Cope, 1873, late Middle Miocene, Pawnee Creek, Co\oTa.do = Miomastodon proavus [subsequently 
referred by the present author to Serridcntinus]. 

Mastodon matthewi Osborn, 1921, Middle Pliocene, western Nebraska = Mtomasiorfon matthewi [now Pliomdstodbn 
matthewi]. 

Also probably Mastodon tapiroides americanus Schlesinger, Lower Pliocene of Hungary = Miomastodon tapiroides 
americanus. 

The distinctive grinding tooth characters in all these MastodontiniE are: (1) that each loph (protoloph, metaloph, et seq.) is 
composed of a main internal and external bunoid cone; (2) the intermediate conule region does not develop; (3) the earliest 
grinder is tctrabunodont ; (4) as the third loph is added it becomes hexabunodont; (5) as the fourth loph is added it becomes 
octabunodont ; (6) whereas these four, six, and eight cones heighten (hypsodonty), they never unite transversely into a cre^t; 
thus none of the Mastodontina^ becomes [truly] zygolophodont. 

The broad enamel band of the tusks is apparentlj' placed on the concave surface of the tusk in Miomastodon merriami, on 
the convex surface of the tusk in Miomastodon tapiroides americanus. 

Middle Pliocene of Hungary. — In his Memoir of 1922 Schlesinger figures (Taf. xiv-xix) and describes (pp. 
115, 116, 227-230) a number of beautifully preserved upper and lower grinders under the name M. (Mammut) 
americanus Pennant forma praetxjpica from the Palaeontological Collections of Budapest. Unlike the previously 



154 



OSBORN: THE PROBOSCIDEA 



mentioned Mastodon tapiroides americanus of Lower Pliocene age, these teeth are regarded as of "Levantin" 
or Middle Pliocene age preceding that of Mastodon [Anancus] arvernensis. Consequently the species M. prae- 
typica is geologically the most recent of all the specific stages heretofore described. 

Through our present knowledge, therefore, there are five species known in Europe and in North America, in 
descending geologic order as follows: 



1919-1922 



1921 



1921 



1921 



1933 
1935 



Mastodon americanus praetypica Schlesinger, Hungary, of 

"Levantin" or Middle Pliocene age {fide Schlesinger) = Ph'omaslodon americanus praetypica 

Mastodon tapiroides americanus (Schlesinger in Osborn), 

Hungary, of Lower Pliocene age (fide Schlesinger) = Mimnaslodon tapiroides americanus 

Mastodon matthewi Osborn, Lower Pliocene, Snake Creek 

formation, western Nebraska (fide Matthew) = Pliomastodon inatthnri 

Mastodon merriami Osborn, Middle Miocene, Hiunboldt 

County, Nevada, Virgin Valley formation (fide Merriam) = Miomastodon merriami 

Miomastodon merriami rei., Middle Miocene, Pawnee Creek, 

Colorado = Miomastodon merriami ref . 

Miomastodon depereti sp. nov.. Lower Miocene, Sables de 

rOrlfenais, France = Miomastodon depereti sp. nov. 

Miomastodon merriami Osborn, 1921 
Figures 89, 98, B, 99 
Virgin Valley formation, Middle Miocene, Thousand Creek, Humboldt 
County, Nevada; also Pawnee Creek, Colorado. 

Mastodon merriami Osborn, 1921. "First Appearance of the 
True Mastodon in America," Amer. Mus. Novitates, No. 10, June 
15, 1921, pp. 4-6 (Osborn, 1921..522). Type.— (O.sborn, op. cit., 

pp. 4-6). "The type specimens include several bone fragments, 
portions of the two upper tusks, and parts of the upper molars, in 
addition to the well-preserved two lower molars here figured as 
the type. A very important character is the presence of broad 
^ ,, ^^1 II enamel band.? on the upper tusks (Fig. 2, C, D, E), which are 

perhaps similar to the enamel bands observed by Schlesinger in the 
true Miocene and Pliocene mastodons of Hungary. . . . Type: 
Colo. Mus. 92, found in 1909 in the Thousand Creek formation 
f ^^^^gl^^^^^^m [Virgin Valley formation], Htmiboldt County, Nevada, includes 

two left [now identified as left and right] inferior molars (Fig. 2, A, 
B [Fig. 1,B]; cast Amer. Mus. 14471, also portions of two upper 
tusks [and two superior molars]." Type Figure. — Osborn, 

1921.522, fig. 1 ,B, p. 3, and fig. 2, p. 5 (Figs. 89, 98, and 99 of the 
present Memoir). 

Type Description (Osborn, 1921.522, p. 6). — "The contours 
of these grinding teeth, as seen from above (Fig. 1, B), are convexo- 
(inner side) concave (outer side); the first crest is relatively 
narrow; the second, third, and fourth crests are relatively broad; 
the rudimentary fifth crest is little if anj' more advanced than in 
M. matthewi; crests two to four exhibit rudimentary intermediate 
cones and the spurs of a trefoil. The presence of an enamel band 
itT^^^^^^^^^^^m^ on the tusks and the somewhat more brachyodont character of the 

grinding teeth separate this stage from the M. americanus (Fig. 1, 




Type of Miomastodon merriami 

Fig. 99. Type superior molars and tusks of Miomastodon merriami 
Osborn, 1921 (Colo. Mus. 92), Virgin Valley formation, Middle Miocene, 
Nevada. See also figure 98,B. .After Osborn, 1921.522, p. 5, fig. 2. About one- 
third natural size. Cast Amer. Mus. 14471. 

.4, Two superior grinding teeth, M^, M', of the right maxilla, internal 
view. B, Same teeth, external view, with 11-inch scale indicated. C, Superior 
tusk of the left side, lateral view. E, External view of the same tusk. D, 
Superior tusk of the right side, external view. The presence of a broad 
enamel band on the concave outer surfaces of the right and left superior 
tusks is clearly indicated in D and E. 



THE MASTODONTIN-^: MIOMASTODON 



155 



C) of the Pleistocene. This species is dedicated to Professor John 
C. Merriam, in recognition of his pioneer work in describing the 
fauna of Thousand Creek." 

Miocene Type Locality. — The assignment of the type 
specimen to the Middle Pliocene of the Thousand Creek formation, 
Humboldt County, Nevada, was erroneous; this error was re- 
peated by the writer in his paper of September 1, 1921 (Osborn 
1921.526, p. 332j, but was corrected in the following footnote in 
the same publication: "Immediately after the publication of this 
paper the author learned [from Prof. John C. Merriam] that the 
type of Mastodon merriami occurred in beds of Middle Miocene 
age, which makes it geologically older than Mastodon matthevri." 
(Letter, Merriam, June 24, 1921): ... "I appreciate your 
courtesy in naming the Virgin Valley species in my honor. ... I 
note that Mastodon merriami is referred to the Thousand Creek 
formation. The [type] locality described bj' Mr. Hills, namely, 
that at which G. D. Mathewson secured his material, is, however, 



summer of 1909 I went to Thousand Creek to examine the occur- 
rence near there of precious opal. ... I continued the journey to 
Denio, Oregon. The following morning I left for a ranch on 
Thousand Creek. At the ranch I found specimens of the opal and 
fragments of a Mastodon tooth which my guide, George D. 
Mathewson, said he had found in digging one of the excavations 
along the opal outcrop. This outcrop is situated on a precipitous 
hill about 500 feet above the level of [the stream known as] Thou- 
sand Creek and between the main forks of the creek. . . . Mr. 
Mathewson had already dug out most of the bone fragments and 
thrown them out on the side of the excavation. I dug out a few 
remaining fragments myself and went very carefully over all the 
stuff that had been excavated and thrown out. Sifting was out of 
the question at that time as the ash was so wet as to form a stiff 
mud. I placed all the fragments, including the tusks, into a sack 
which was carried down to the ranch and later packed into a box 
for transportation on the stage 130 miles to the railroad and thence 




, 2— f A M.i334^5 

Fig. 100. Referred Miomastodon merriami Osborn, Middle Miocene of Pawnee Creek, Colorado 

Modified after Frick, 1933, Figs. 28, 24 
Jaw with tusk (F:A.M. 23345), one-eiglith natural size. Molars (F:A.M. 23345, 23337), one-third natural size. 
A-A3, Mandible with second and third molars, also inferior and superior tusks and sections (F:A.M. 23345). 
This appears to be a typical Miomastodon. 

B-B2, Superior molar and tusk, also .section, of another individual (F:A.M. 23337). This third superior 
molar, M', differs from the typical Miomastodon molar with trefoils. 



in the Virgin Valley formation, which is of approximately Middle 
Miocene age, not far from the zone of the Mascall of the John 
Day region. . . . this evidently brings the appearance of these 
Mastodons back to near Middle Miocene." 

Osborn corrected this geologic error specifically in Science 
(Osborn, 1921.523, p. 108). 

The Colorado Museum record of this specimen is (letter, 
Figgins, P'ebruary 25, 1921): "No. 92: from the John Day forma- 
tion (Miocene), Thousand Isle Creek, Humboldt county, Nevada, 
found April, 1909, by Geo. D. Mathewson, Denio, Ore." Mr. 
Richard C. Hills writes (letter. March 9, 1921): "Early in the 



by express to Denver. The parts must all belong to one individual. 
I spent considerable time in assembling the fragments. In addi- 
tion to the two lower molars, there are parts of the upper molars 
on one side embedded in a portion of the upper jaw. The tusks are 
not entire. You can, however, see the black enamel bands in the 
photograph.s. We can, if you so desire, furnish other photographs 
showing the upper molars." 

Osborn, 1922: The writer is greatly indebted to Director 
Figgins, also to the geologist, Richard C. Hills for the privilege of 
examining and describing this most important specimen which 
demonstrates the arrival of a member of the true Mastodontinae 



156 



OSBORN: THE PROBOSCIDEA 



New generic and specific characters revealed (Fig. 100) in the 
Pawnee Creek specimens are: (1) Mesial expansion of rudimentary 
trefoils in en to- and ectoconelets, upper and lower; (2) mandibular 
rostrum laterally compressed, abbreviate as compared with Tri- 
lophodon or Serridentinus; (3) rodlike inferior incisors, vertical 
oval section; (4) downturned superior incisors of moderate length, 
broad enamel band, broadly oval vertical section. 

Miomastodon tapiroides americanus Schlesinger, 
1921, 1922 

Figure 98 
Lower Pliocene, Tasndd, Usztat6 Komitat, Hungary. 

Mastodon tapiroides americanus Schlesinger (in Osborn). 
"First Appearance of the True Mastodon in America," Amer. 
Mus. Novitates, No. 10, June 15, 1921, p. 2. Schlesinger's descrip- 
tion: "Die Mastodonten der Budapester Sammlungen," 1922, Geol. 
Hungarica, II, Fasc. 1, pp. 224-227, Taf. xiii, figs. 6, 7, and xiv, 
figs. 1-4. Type. — A third left superior molar, l.M', also two 

left inferior molars, I.M2-3. Horizon and Locality. — Lower 

Pliocene, Tasniid, Usztato Kom., Hungary. Type Fig- 

ure.— Op. cit., Osborn, 1921.522, p. 3, fig. 1, D, Dl (see Fig. 98D, 
Dl of the present Memoir). 

Type Description.— (Op. cit., O.sborn, 1921.522, p. 2): 
"There cannot be the least doubt, however, as to the affinity [to 
the true Mastodon] of the grinding teeth found in the Lower 
Pliocene of Hungary, to which Schlesinger applies the name M. 
tapiroides americanus. These [type] teeth are reproduced [redrawn] 
herewith (Fig. 1, D, Dl) from unpublished photographs, kindly 
forwarded by the author, to the same scale with corresponding 
grinders (A, ... A2, A3) from the Lower Pliocene, Snake Creek 
formation, of western Nebraska, also with lower teeth (B) from 
the Middle Pliocene [Middle Miocene, Virgin Valley formation], 
[near] Thousand Creek, Nevada, and with (C) the posterior lower 
molar of M. americanus from the American Pleistocene." 

"The upper and lower grinders from the Lower Pliocene, 
Tasnad, Usztato Kom., Hungary, embrace a third left superior 
molar (Fig, 1, Dl, see PI. xiii, fig. 5 [error H. F. 0.], Schlesinger), 
also two left inferior molars, mo-ma (Fig. 1, D, see PI. xiv, fig. 1, 
Schlesinger). The linear measurement of the crowns agrees 
closely with that of the Pleistocene M. americanus, but the vertical 
measurement is apparently less, i.e., less hypsodont. This indi- 
cates that already in the Lower Pliocene the mastodonts had 
attained the massive proportions of their Pleistocene descendants. 
The lophs are similarly composed and show no trace of a trefoil 
ridge. There is nothing to debar these Lower Pliocene mastodonts 
of Hungary from the true ancestral line of our Pleistocene Mastorfon." 

Osborn, 1925: Since the above type description was written 
for Doctor Schlesinger the species has been transferred to the genus 
Miomastodon. The incisive tusks discovered in the same locality 
and probably attributable to the same species appear in the photo- 
graphs furnished by Doctor Schlesinger to bear a broad enamel 
band on the convex external surface; if this determination is 
correct, it would seem that in these Pliocene mastodonts the tusks 
are outwardly convex, as in Mastodon americanus, a progressive 
character; as observed in Miomastodon merriami, the tusks are 
outwardly concave. 

'Unworn molars of the Mastodontinae clearly demonstrate that the rudimentary trefoil spur springs from the ectoconelets in the inferior molars, from the 
enioconelets in the superior molars. The worn molar gives a false impression that the trefoil spurs arise from the mesoconelets. 



phylum in the Nevada region of America in Middle Miocene time. 
The presence of a broad enamel band on the upper tusks is a very 
important character ; these enamel bands are perhaps similar to 
the enamel bands observed by Schlesinger in some of the true 
Lower Pliocene mastodons of Hungary (as recorded in his published 
Memoir) to which he has assigned the name Mastodon tapiroides 
americanus. A significant difference is that in Miomastodon mer- 
riami the enamel is on the outer concave surface of the tusk, where- 
as in the photographs of M. tapiroides americanus furnished by 
Doctor Schlesinger the enamel is on the outer convex surface. 

MiOMASTODOX MERRIAMI OF COLORADO (FiG. 100). — TheSC 

most welcome specimens (F:A.M. 23345, 23337), discovered by the 
Frick Expedition of 1932 in the classic Pawnee Creek horizon of 
Colorado, were referred by Frick (1933, p. 611, fig. 28) to 'Mio- 
mastodon pronvus Cope.' They constitute a most timely and im- 
portant addition to our knowledge of the Middle Miocene genus 
Miomastodon and greatly strengthen Osborn's theory of the year 
1921 that Miomastodon merriami represents the arrival of ancestors 
of the true Mastodontinae in America in Middle Miocene time. 

The specific reference (Frick, 1933) of these specimens to the 
Pawnee Creek type of ' Mastodon ' pi-oavus Cope was an error, be- 
cause the two fragmentary grinding teeth of M. proavus (see Chap. 
X, fig. 363) clearly exhibit the characteristic serrate spur on both 
sides of the ectoconelet, both in the premolar and the fragmentary 
molar, also the reduction of the contiguous mesoconelet; these 
two features are highly characteristic of the Serridentina;, as pointed 
out in Chapter X, pages 403, 404, where a detailed description with 
figures of Serridentinus proavus is given. In the lower grinders of 
Miomastodon merriami, on the contrary (see especially Fig. 9SB, 
type lower grinders of Amer. Mus. 14471), there is no serration on 
the sides of the ectoconelet, but there is a distinct trefoil expansion 
of the contiguous mesoconelet. 

The expanded trefoils cin the ectoconelets' of the Mastodon- 
tinaj become more conspicuous as the teeth are worn down, as 
shown in the much worn r.Ms of F:A.M. 23345 of the Pawnee 
Creek (Fig. 100). This mesial expansion of the ectoconelet is also 
observed in Miomastodon tapiroides Schles. (Fig. 98 Dl), in Plio- 
m,astod(/n matthewi (Figs. 98 A2 and 102), as well as in certain speci- 
mens of Mastodon americanus (Fig. 89 — Amer. Mus. 14293, r.M^ 
Amer. Mus. 14294, r.Ms); this mesial e.tpansion is therefore char- 
acteristic of the phylum Mastodontinae. 

The comparative measurements are as follows: 



Type (Colo. Mus. 92) : 


ap. 


tr. 


I. 


R.M,, 


179 


80 


45 


Referred (F: A.M. 23345): 








R.Ma 


164 


77 


47 


Incisive tusk (upper) 








Length 






450 mm 


Vertical diameter 






67 


Transverse diameter 






56 


Width of enamel band 






49 


Incisive tusk (lower) 








Length 






250+ 


Vertical diameter 






40 


Transverse diameter 






35 



8. CHARACTERS OF PLIOMASTODON AND ITS INCLUDED SPECIES 

Subfamily: Mastodontin^ Brandt, 1869 — Osborn, 1910 

Genus: PLIOMASTODON Osborn, 1926 
Original reference: Amer. Mus. Novitates, No. 238, Nov. 30, 1926, p. 1 (Osborn, 1926.706) 

Generic Chakacters. — Incisive tiLsks witli enamel band ve.stigial or wanting, suboval to rounded 
in section, upturned {P. vexillarius) ; ridge-crests with expanded ectotrefoils {P. matthevn) ; intermediate 
in hypsodonty between Miomastodon and Mastodon. 

Mastodon (Miomastodon) matthewi, the genotypic species of Pliomastodon, is of Lower Pliocene age, whereas 
Mastodon merriami, the genotypic species of Miomastodon, is of Middle Miocene age. In the superior incisive 
tusks of Pliomastodon (genotype) the lateral enamel band is absent; in the superior incisive tusks of Miomastodon 
the lateral enamel band is present. Both of these generic stages are imperfectly known at present. 

This Pliocene stage is based on the species first described by Osborn as Mastodon maltheivi (Osborn, 
1921.522, pp. 2, 4), then as Miomastodon mattheivi (Osborn, 1922.564, p. 4), and finally made the genotype 
of the genus Pliomastodon (Osborn, 1926.706). This generic stage (Fig. 98) is typified by a right third 
superior molar, r.M' (Amer. Mus. 18237), and by paratypes r.M^ (Amer. Mus. 18238) and r.M, (Amer. 
Mus. 18239), also by a referred r.Mg (Amer. Mus. 19248«) and a referred tusk (Amer. Mus. 192486)— all 
from the Snake Creek B formation of western Nebraska, quarries 1 and 5, as shown in the table (Chap. X, p. 427). 
Associated with these specimens of Pliomastodon were the type and referred specimens of Serridentinus anguirivalis, 
S. nebrascensis, and Rhynchotherium anguirivalis. 

American Pliocene Serridentinus and Mastodon. — During the summer of 1908 an American Museum 
party mider W. D. Matthew and Harold J. Cook first collected in the "Lower Pliocene" deposits of western Nebraska, 
subseciuently known as the Snake Creek beds (Matthew and Cook, 1909, p. 363). The first proboscidean molars 
discovered belonged to the serrated molar group and are now referred to Serridentinus anguirivalis. 

Subsequently, in the seasons of 1916 and 1918, several bunolophodont grinders (Amer. Mus. 18237, 18238, 
18239, and 17217) were also discovered in the same Snake Creek formation; three years later, under the title 
"First Appearance of the True Mastodon in America" (Osborn, 1921.522, pp. 2,4), Osborn selected these specimens 
(Fig. 98 A-A3) as the type and paratypes of a new species which he named Mastodon matthewi, in honor of Dr. W. D. 
Matthew the author who first described tlie new and interesting fauna of the Snake Creek formation. Amer. Mus. 
17217, however, was erroneously figured as a paratype of Mastodon matthewi and has since been made the type of 
Serridentinus anguirivalis Osborn, 1926. The reader is referred to the new figures of the type, paratypes, and 
referred specimens of Pliomastodon matthein from the Snake Creek B level (Fig. 102), a description of which species 
follows. 

Pliomastodon matthewi O.sborn, 1921 1926 cene, New Genus." Amer. Mus. Novitates, No. 49, Oct. 23, 1922, 

Figures 89, 98, 101, 102 P- 4 (Osborn, 1922. 564). 

Snake Crock f.iniK.tion, LowiT Pliocni.-, li-vcl H, Sioux County, western Pliomastodon malthewi Osborn, 1926. "Additional New 

Nel.ni.ska. Collecii'd by the Ainericim -Museum expeditions of 1910 and I'JIS. Genera and Spi'cies of the Mastodontoid Proboscidea." Amer. 

Mastodon maiihem Osbom, 1921. "First Appearance of the Mus. Novitates, No. 238, Nov. 30, 1926, p. 1 (Osborn, 1926.706). 

True Mastodon in America," Amer. Mus. Novitates, No. 10, .June Typk Dkscription .\nd Type Locality.— (Osborn, 1921.522, 

15, 1921, pp. 2-6 (Osborn, 1921.522). p. 2): . . . "Subsequently, in I91S, several distinctive specimens 

Miomastodon mattheivi Osborn, 1922. " Dibetodon edensis were found in the same beds which may now be named as the tj-pe 

(Frick) of Southern California, Miomastodon of the Middle Mio- and paratypes of a new species of Mastodon (Mastodon mattluwi), 

1.57 



158 



OSBORX: THE PROBOSCIDEA 



A M. 1623 7 /- nT^(rev> 



A M. /92^8 b 




All '/^ natural sije 



PLIOMASTODON MATTHEWI 



AM. 1 92^3 a r. m 



A M /6238 r 772J 




Fig. 101. Second type and paratype 
figures of Plioinastodon matthewi Osborn, 
one-third natural size. Compare figure 
98A, A2, A3, also figures 102 and ! 



in honor of Dr. W. D. Matthew, the author who first described this 
interesting fauna. ... [p. 4] Type : the right third superior 

molar, Amer. Mus. 18237. Paratypes: the right second in- 

ferior [superior] molar (unworn), Amer. Mus. 17217' [this tooth has 
now been made the type of Serrideniinus anguirivalis] ; the posterior 
portion of a right third inferior molar (more worn), Amer. Mus. 
18238, also of a right second [first] inferior molar, Amer. Mus. 18239. 
The type and paratypes probably belong to four dififerent individu- 
als. . . . The type and Nos. 18238 and 18239 are from the Snake 
Creek B level {Procamelxis-Hi-pparion Zone) of Sioux County, 
Nebraska; the level of No. 17217 is not recorded. The type (Fig. 
1, A) is distinguished by the rapid narrowing of the posterior half 
of the crown of the third upper molar, including the third and 
fourth crests; the fourth crest is extremely narrow and bilobed; 
the rudimentary fifth crest consists of a single cusp. In these 
features M. matthewi is more primitive than the corresponding 
tooth of M. tapiroides americanus (Fig. 1, Dl) of the Lower Plio- 
cene of Hungary. The association of the lower molars, Amer. 
Mus. 17217 [error H. F. O.], 18238, 18239, as paratypes is provi- 
sional, because the Snake Creek deposition extended over a long 
period of time and may represent more than two life zones. Of 
these teeth, [m^] presents three unworn pointed crests with the rudi- 
ments of a trefoil (Fig. 1, Al'); in a second [first] molar ( . . . A2) 
the trefoil is less apparent : in the third lower molar ( . . . A3) it is 
not apparent at all. In the latter tooth, which is probably the 
posterior half of a third lower molar of the right side, the third 
and fourth crests are partly preserved ; crest five is represented by a 
broad tuberculate talon." Compare figure 98 of the present 
Memoir. 

Specific Characters (Fig. 102). — As compared with the 
type molars of Miomastodon merriami (Fig. 98, B), the type and 
paratype molars of Pliomastodon matthem (Fig. 98, A, A2, A3, 

'See Chap. X, fig. 391, p. 425. 



Amer. Mils. 18256, 18259 Paratypes 



r. 777 




r. mj 



/2 riat. 52 J€ 




Amer. Mu.5. /a237 Type 
r.TTZ? 




P. MATTHEWI 

Third Type and Paratype Figures of Pliomastodon matthewi 
Fig. 102. Type and paratypes of Pliomastodon matthewi, Snake Creek 

B horizon (upper), western Nebraska, one-half natural size. Type (Amer. 

Mus. 18237), a right third superior molar, r.M''. Paratypes (Amer. Mus. 

18238), a right third inferior molar, r.Mj, partly restored; (Amer. Mus. 

18239), a right first inferior molar, r.Mi. For comparison with other 

mastodonts of Snake Creek, see figure 394. 



THE MASTODONTIN^: PLIOMASTODON 



159 



and Fig. 102) show a greater elevation (subhypsodonty) and a 
greater anteroposterior compression (lophodont}'). The summits 
of the paired lobes, i.e., cones, seem to be somewhat more approxi- 
mated. Another distinctive feature is the posterior narrowing of 
the crown of the third superior molar (Fig. 98, A and Fig. 102). 
Since the original description and figures were prepared (Fig. 
98), another niolar tooth (Amer. Mus. 19248a), collected in 1918, 
has been identified as a third inferior molar of the right side, r.Ms. 
In the same quarry (Quarry 5) was found a section of a superior 
incisive tusk (Amer. Mus. 192486) somewhat abraded by river 
erosion, which shows no sign of the superior enamel band. These 
two specimens have been mentioned as indicating a more recent 
geologic age and with correspondingly progressive characters 
which signalize a new generic stage for which the name Pliomastodon 
(see Osborn, 1926.706, p. 1) seems appropriate (Fig. 101). 

Miomastodon Pliomastodon 
Comparative Measurements merriami matthewi 

Type Type and Ref. 

Third superior molar, M', 

anteroposterior 154e 160 

Height of internal cone of 

metaloph 60e 57 (type) 

Third inferior molar, Mj, 

anteroposterior 179 194 

transverse 80 

breadth-length index 45 

Height of internal cone of 

metalophid 54e 61e 

Superior incisive tusk, P 

Vertical diameter \ ._ 90 

Transverse diameter. . • . / 81 

The progressive development of the molar crowns in these 
respective Miocene and Pliocene species may also be carefully 
examined and compared from the following figures: 

Pliomastodon matthewi. Figure 98, A, A2, A3, original type and 
paratype figures; figure 89, type and paratypes in comparison 
with Miomastodon merriami, Pliomastodon americanus praetypica, 
Mastodon americanus, and Palieomastodon intermedius; figure 394, 
new type and paratype figures in comparison with other Snake 
Creek B mastodonts; figures 102 and 101, new type and paratype 
figures. 

Miomastodon merriami. Figure 99, type superior molars and 
tusks; figure 98, type molars in comparison with Pliomastodon 
matthewi, Mastodon tapiroides, and M. americanus; figure 89, type 
molars in comparison with Pliomastodon matthewi, P. americanus 
praetypica, M. americanus, and Palxomastodon intermedius. 

The above comparative measurements and figures reveal that 
in both Miomastodon merriami and Pliomastodon matthewi the 
crowns of the superior and inferior third molars, .\P and Ms, 
narrow posteriorly, the tetartoloph and -lophid have not attained 
the width seen in Mastodon americanus, the pentaloph and -lophid 
are very rudimentary, whereas in Pliomastodon americanus 
praetypica the pentalophid is almost as strong as in M. americanus. 
The distinction between Miomastodon merriami and Pliomastodon 
matthewi is partly seen in the height of the crowns, tlie lophs in M. 
merriami (e.g., Fig. 98, B) being somewhat less elevated than in 



P. matthewi (e.g., Fig. 98, A). The progressive difference is not so 
well marked in the grinding teeth as it is in the tusks, as observed 
by comparison of the type of M. merriami (Fig. 99) and new type 
figure of P. matthewi (Fig. 102). 

Pliomastodon americanus praetypica 

Schlesinger, 1919, 1922 

Figures 89, 103, and Pi. i, pp. 134-135 

Upper(?) Pontian anil Levantin age, Middle Pliocene. Batta-£rd, 
R;ikoskeresztiir, Szentlorinoz, and Ajniicsko, Hungary. 

Schlesinger originally described this true species of Mastodon 
from Hungary in his paper of 1919, as cited below, figuring his 
excellent types in Taf. vi, figs. 2-4. His final and more complete 
description appears in his Memoir of 1922. 

Mastodon (Mammutj americanus Penn. forma praetypica 
Schlesinger, 1919. "Die stratigraphische Bedeutung der euro- 
paischen Mastodonten," Mitt. Geol. Ges. Wien, XI, p. 142. 




COTTPF.S OF PlICMASTODON AMERICA.NUS PRAETYPICA 

Fig. 103. Colypes of Mastodon (Mammul) americanus Venn, forma prae- 
tyjrica Schlesinger selected in the present Memoir, namely, second and third 
right superior molars, r.M'-''', and second and third left inferior molars, l.Mjj, 
one-third natural .size. Compare Schlesinger, 1922, Taf. xv, figs. 2 and 4. 
Originals in Hungarian National Museum, Budapest. 

Right superior molars, r.M-"^: Szabadka ( = Maria-Thereaiopel), Komitat 
Pest, Hungary. 

Left inferior molars, l.Moj: Batta-fird, Komitat Pest, Hungary. 

OKKiiNAL Description.— Schlesinger, 1919, p. 142): "a) 
Die intermediiiren Molaren (vgl. Tafel vi, Figur2-1) dieser, friiher 
mit M. Borsoni zusammengeworfenen Spezies, sehliessen sich im 
Bau engstens an die vorbesprochene Art an, nur sind die Kronen 
breiter, die Tiiler enger. b) Fur die letzten Molaren (vgl. Tafel 
VI, Figur 1, 3, 4) gilt, abgesehen von der hoheren Jochformel 
(4 X im Ober-, 4 X —5 im Unterkiefer) das gleiche. Hier tritt meist 
noch das Merkma! der geringcren und etwa.s anderen Ausbildung 
der Sperrleisten helfend dazu. Trotzdem ist die Bestimmung 
schwierig und ohne breitere Vergleichsbasis die Trennung von der 
mioziinen Form oft uiunoglich." 



160 



OSBORN: THE PROBOSCIDEA 



M. [Mastodon] (Mammut) americanus Pennant forma praetypica 
n.f., 1922. "Die Mastodonten der Budapester Sainmlungen." 
Geol. Hungarica, Ed. Sep., Tome II, Fasc. 1, pp. 115, IIG, 
227-230. 

CoTYPES. — The cotypes selected for illustration in this Memoir 
are the .second and third superior and inferior molars from two local- 
ities, as follows (Schlesinger, 1922, Taf. .xv, figs. 2 and 4— same as 
Figs. 3a and 4 of Schlesinger, 1919, Taf. vi, on larger scale) : " Figur 
2: M^^dexl. von der Kauflache [Szabadka ( = Maria-Theresiopel), 
(Kom. Pest). Levantin]. . . . Figur 4: Linke Mandibel mit M2+3. 
Fundort: Batta-Erd, Komitat Pest. Horizont: Oberstes Pontikum 
(Unterplioziin). Wiedergabe aller vier Figuren: Yi natiirl. Gr." 
Originals of figures 1 and 3 in the Hungarian National Museum 
of Budapest ; of figures 2 and 4 in the Hungarian Reichsanstalt of 
Budapest. Cotype Localitie.s. — Aside from Batta-fird, Schle- 

singer mentions three other cotype localities, namely, Rakoskeresz- 
tur, Szentlorincz, and Ajnacsko, Hungary. Upper(?) Pontian or 
Levantine age, (?)Middle Pliocene. Cotype Figures. — Schle- 

singer, 1922, Taf. xiv-xix. These enlarged cotype figures include 
the original figures of Schlesinger, 1919, Taf. vi, figs. 2-A. 

Comparison with Mastodon americanus. — Schlesinger, 
1922, "Die Mastodonten der Budapester Sammlungcn," illustrates 
a large number of grinding teeth of the true Mastodon americanus 
in the collections of Budapest and of Vienna (op. cit., 1922, Taf. 
xv-xxii) from different localities in the United States, and com- 
pares in great detail the grinding tooth characteristics of Mastodon 
americanus tijpica with the grinding tooth characteristics of M. 
americanus praetypica and of M. [Zygolophodon] borsoni. He also 
summarizes in several parts of his invaluable Memoir of 1922 the 
resemblances and differences between the Mastodon americanus 
and the M. \Z\ borsoni phyla. 

Geographic Distribution of Pliomastodon americanus 
PRAETYPICA. — The cotypes selected in this Memoir (Fig. 103) are 
from Batta-Erd and Szabadka ( = Maria Theresiopel), Komitat 
Pest, Hungary; Schlesinger mentions three other cotype localities, 
namely, Rakoskeresztur, Szentlorincz, and Ajnacsko. (Schle- 
singer, 1922, p. 229): "Die Zahl der gut horizontierten Fund- 
punkte des M. americanus f. praetypica ist zwar nicht reichlich, 
doch sind es qualitativ hochwertige Stellen. An dreien von ihnen 
(Rakos, Szentlorincz und Batta-Erd) ist eine konkordante Auf- 
einanderfolge oberpontischer und Icvantiner Bildungen erwiesen; 
in alien Fallen kanien die Reste von M. americanus f. praetypica 
aus dem mitteiplioziinen Niveau." (Op. cit., p. 230): "Solche 
Belege konnen bei der Beurteilung einer Frage natiirlich nicht 
iibergangen werden. Ich bin vielmehr der festen Uberzeugung, 
dass M. americanus f. praetypica init dem Ende des Levantins aus 
unserem Gebiete endgiitig abgewandert ist." 

Osborn, 1926: In his volume of 1922, Taf. xiv-xix, Schle- 
singer figures numerous beautifully preserved upper and lower teeth 
from the Pliocene of Hungary which he refers to Mastodon (Mam- 
mut) americanus forma praetypica. As to the geologic age of these 
specimens, Schlesinger considers (op. cit., p. 227) that undoubtedly 
they are of "Levantin age" similar to that of M. arvemensis, 
namely. Upper Pontian [?Middle Pliocene). As to geographic 
di.stribution and migration, he remarks as quoted above. 

We may consider as the cotypes the eight grinding teeth 
figured in Taf. xiv-xix of Doctor Schlesinger's Memoir. 



Pliomastodon sellardsi Simpson, 1930 
Figures 104, 105 

Brewster, Florida; Lower Pliocene, Bone Valley formation. 

The type of Pliomastodon sellardsi was first described (1916) 
by Dr. E. H. Sellards, at that time State Geologist of Florida, as 
referable to Mammut progenium? (see Sellards, 1916, p. 95, PI. x). 
The author pointed out, however, its close resemblance to Mastodon 
americanus but emphasized the longer symphysis. The specimen 
was discovered in the mine of the American Cyanamid Company at 
Brewster and was sent to Doctor Sellards by the late Anton 
Schneider, then superintendent of the company. The right half 
of the mandible was returned to Mr. Schneider at his request and 
has since been mislaid, but the left half of the mandible is available 
at the Florida Geological Survey. Doctor Simpson in connection 
with his description gives three views of the specimen, the missing 
right mandible, internal and external aspects, also superior view of 
both sides, which are reproduced as figures 104 and 105 of the 
present Memoir. 

Pliomastodon sellardsi Simpson, 1930. "Tertiary Land 
Mammals of Florida," Bull. Amer. Mus. Nat. Hist., LIX, Art. 
Ill, pp. 203-206. Type.— "Left lower jaw with M2.3. 




F.S.V-3822 



Fig. 104. Pliomastodon sellardsi Simpson, type (Fla. Geol. Surv. V3822), 
from Brewster, Florida. Lower jaw with right and left M2-.1, crown view. 
One-fifth natural size. After Simpson, 1930, fig. 30. Compare figure 87, M. 
anwricxi?ius. 



THE MASTODONTIN^: PLIO.MASTODON 



161 



llight lower jaw of same individual known "^ 
from photographs, present location un- 
known. Presented by Anton Schneider." 
Fla. Geol. Surv. V3822 (old number 
6160). Horizon and Locality. — 

"Lower Pliocene, Bone Valley Fomiation, 
pit of American Cyanamid Company, 
Brewster, P'lorida." Type Figure.— 

Op. cii., fig. 30, p. 204, fig. 31, p. 205. 

Type Description. — (Simpson, 1930, 
p. 203): "Symphysis elongate, stout lower 
tusks, alveoli about 50 mm. in diameter near 
posterior end. Ms fully trilophodont. Fifth 
crest of M.I less developed than in Mastodon 
americanus, ridge-like, more distinct internal 
cusp and four closely appressed cuspules 
external to it. Cusps more elevated than in 
Miomasiodon merriarni, less than [in] Masto- 
don americanus, about as in Pliomastodon 
nmtthewi. Inner side of M3 strongly convex, 
outer border straight. First loph distinctly 
narrower than second, and last two lophs 
also relatively narrower than in Mastodon" 



Measurements 

Mo Length 110 mm. 

Width 80 

M, Length 164 

Width (max.), second loph 95 




Fig. 105. Pliomastodon seltardsi Simpson. Missing part of type. Right lower jaw with M2-3, internal 
and external views. One-fifth natural size. After Simpson, 1930, fig. 31. 



Height internal cone of first crest — actual 55 mm., estimated 
57 mm. 

Height internal cone of second crest — actual 56 mm., esti- 
mated 59 mm. 

Osborn, 1933 : Pliomastodon sellardsi, according to the above 
descriptions of Sellards (1916) and of Simpson (1930), is a most 
welcome addition to the American ancestry of Mastodon ameri- 
canus; it is distinguished by its longer rostrum, with ridge-crests 
intermediate between Miomasiodon merriarni and Mastodon 
americanus. 

To the same species may be referred a right mandibular ramus 
(Fla. Geol. Surv., V-5376, cast Amer. Mus. 26979), with M2, M3; 
measurements of M3 181X98 mm., index 54, found at a locality 
not very far from that of the type. 



Pliomastodon vexillarius Matthew, 1930 
Figures 106-108 

From southeast of Coalinga, Fresno County, California. Upper Etche- 
goin 1 = liite Pliocene], Plesippus proversus life zone. 

Faunal Level. (Stirton, August 8, 1932): "The type of PUomaslodon 
vexillarius was found in the upper portion of the Etchegoin section in the 
North Coalinga region. This section has been worked by Ralph Stewart 
(U. S. Geol. Surv.) who collected isolated teeth of Plesippus proversus and 
Castor californicus as his inde.x fossils. Late Pliocene." 

The Stantlard Oil Company of California, in excavating on its 
properties at the north end of the Kettleman Hills anticline, south- 
east of Coalinga, California, discovered the specimen described 



by Dr. W. D. Matthew as PUomaslodon vexillarius, the specific name 
signifying "standard bearer," chosen "in recognition of the very 
important contributions to paleontology made by the Standard 
Oil Company of California in the course of its operations and re- 
searches, and the care which the officers of the company have 
taken to preserve and turn over to scientific museums the collec- 
tions made by their parties and to facilitate in many ways field 
studies and collecting work of this [University of California] and 
other institutions." Doctor Matthew states (p. 338) that "What- 
ever be the final determination as to nomenclature and taxonomic 
status, the skull represents a type of American mastodon hitherto 
known only from teeth and jaw fragments. It is distinguished 
from the long-jawed species by the short jaws and upward curving 
enamelless tusks, from the mirifii-us group by simple molars with 
little trefoil development, from the 'American Mastodon' by the 
shortened skull base with overhanging occiput and the marked 
approach to anteroposterior succession of teeth." 

Pliomastodon vexillarius Matthew, 1930. "A Pliocene IMasto- 
don Skull from California Pliomastodon vexillarius, N. Sp.," 
Bull. Dept. Geol. Univ. Calif., XIX, No. 16, pp. 335-348, Pis. 
XLi-XLiv, text. figs. 1, 2. Type.— Back of skull and 

palatal region approximately complete, zygomata broken but 
parts of them preserved, frontonasal and orbital regions partly 
preserved. Dentigerous portions of both sides of lower jaws com- 
plete, but of the symphysis, angles, and coronoid, little is left. 
Atlas complete, one femur, an astragalus, patella, and parts of 
other limb bones. Crowns of teeth largely missing; most of M^ 



162 



OSBORN: THE PROBOSCIDEA 



right and left, and small parts of M', posterior halves of ^U, 
right and left, preserved, but only the bases of the crowns of the 
remainder of the second and third molars. First molars had been 
lost during life and their alveoli partly closed ; second molars well 
worn, third molars unworn in the posterior half, probably lightly 
worn in front. Original in the Museum of Paleontology, Uni- 
versity of California (No. 28301). Horizon and Locality. — 
North end of the Kettleman Hills anticline, southeast of Coalinga, 
Fresno County, California. Upper Etchegoin [ = late Pliocenel. 
Type Figure. — Op. cit., Pis. xli-xliv, text figs. 1 and 2. 

Geologic Age. — (Matthew, op. cit., p. 338): "The type and 
only known specimen of the genus is a last upper molar from the 
Snake Creek beds (referred by Matthew to Zygolophodon in 1918 
[made by Osborn in 1921 the type of Mastodon (Pliomastodon) 




of Vacek's genus or as a distinct genus, according to the relative 
weight given to one or another of the progressive characters." 

Distinctive Generic Characters. — (Cf. Matthew, op. cit., 
p. 336). Upper tusks large, cylindrical, close together and sub- 
parallel at base, curving gently upward and outward, a narrow 
strip of thin enamel showing at the base on the outer side but no 
clear evidence of enamel beyond the alveolus. Lower tusks prob- 
ably absent. Molar teeth wide, moderately crested, less so than in 
Mastodon, trefoils single, not prominent, valleys open, M 2 tri- 
lophodont, M 3 with 4% crests. The basicranial region of the skull 
is little elevated above palatal plane, much shortened as compared 
with Mastodon, occiput very low and wide, heavily overhanging 
backward instead of nearly vertical as in Mastodon and other 
genera. Posterior nares crowded forward so as to be opposite the 
back of M^; in Mastodon they are considerably behind it. The 
entire basicranial region is thus crowded forward along with the 
condyles, which are scarcely over half as far behind the teeth as in 
Mastodon. This forward crowding is paralleled in some of the 
other mastodontines and in the elephants, but mostly with a much 




Fig. 106. Type skull and jaws uf Hliomaslodon vexillariiis Matthew (Mus. tJniv. Calif. 28301). After drawings and photographs kindly furnished the 
present author l)y the University of California. Compare Matthew, 1930. 

(Left) Palatal and side view of type skull, one-fifteenth natural .size. Compare Matthew, 1930, text figs. 1 and 2. 
(Right) Same type skull. Compare Matthew, 1930, PI. xlii. One-fifteenth natural size. 



matihewi]) and owing to the crowns of the upper teeth being broken 
off in our .skull, exact comparison is impo.ssible. It agrees, however, 
in geological age, and the fauna of the Lower Etchegoin is largely 
compo.scd of the same genera as the Snake Creek fauna, so that 
provisionally I refer the akuU to Pliomastodon. As at present 
understood, it represents an advanced stage of the Zygolophodon 
phylum and might be regarded as a progressive species or subgenus 



higher elevation, especially in the latter group, of the basicranial 
above the palatal plane. The author states that the skull ap- 
parently cannot be placed in any of the better known Mastodon- 
tine genera, and gives its distinguishing characters from Gompho- 
therium, " Serridentinus," Mastodon, Cuvieronius, Anancus, Cordil- 
lerion, Tetralophodon, Pentalophodon, Rhynehotherhim, and Choe- 
rolophodon . 



THE MASTODONTIN^: PLIOMASTODON 



163 



Specific Charactehs. — (Op. cU., Matthew, p. 339): "Last 
upper molar with four full crests and a considerable heel behind 
the fourth; in /■". rnatthewt type there are three full crests with the 
fourth half -formed and a small rudiment of the fifth. Comparative 
measurements of this tooth are : 

[ Fliomastodmi ] [ Cordillerion ' 1 



Other measurements of teeth and skull — 

P. vexillarius \C.\ bensonenms 





vexillarius 


matlhewi 


bensonensis 


Antero-posterior length 


158 


144 


191 


Width at anterior end 


80 


80 


99 


Width at fourth crest 


71 


48 







107X71 



125X88.5 



165 
155 

88 



130 
120 
76" 



M-, dimensions 

Diameter of tusk (about a foot \x- 

yond alveolus) 
Width of palate between M' alveoli = 
Width of palate between M' alveoli = 

"The skull compares in bulk with the largest specimens of the 
American Mastodon. . . . From 'Tetralophodon' edensis this 
species is separated by upcurved tusks lacking enamel band. . . . 
The skeletal parts preserved do not show any remarkable features." 

Osborii, 1933: This interesting communication by Matthew, 
published posthumously, positively establishes Pliomastodvn as an 
ancestral stage of Mastodon clearly distinguished by a far more 
primitive cranium, bj' a "narrow strip of thin enamel" within the 
aheolar ba.se of tlie tusk, by the probable absence of inferior tusks, 
by the approximation of the condyles to the superior molars (M'), 
by the e.xtreme backward recession of the anterior nares and the 
shortening of the lower jaw, probably also with an unusual width 
of the occipital crest, suggesting an exceptionally large develop- 
ment of the proboscis. 

As to bodily proportions, the type femur {op. cil., PI. xLiv — 
Fig. 108) is much more massive than that of the Alaskan mammoth 
(Mammonleus primigenius) but indicating a massi\e body about 
2800 mm., or 9 ft. 2K in., in height at the shoulder. The cranium 
apparently debars this species from direct ancestry of Mastodon 
americanus, for it embraces a number of peculiar characters. It is 
nevertheless most welcome as exhibiting a very distinctive, low- 
browed, flat-crested, abbreviated postcranial region, a very broad- 
crested occii)ital region, and broadly divergent premaxillary region 
with sharply upturned tusks. These features are clearly displayed 
in the outline and photographs (Fig. 106) after Matthew, 1930. 
An outline restoration of tliis animal, to the same scale as the 
Warren specimen of M. americanus, is shown in figure 110. 





I'iK- 107. Pliomastodon vexillarius, type. Lower jawg, uiit^ide and 
occlusal views. One-fifth natural size. After Matthew, 1930, PI. xlhi. 



FiK- 108. Type femur of Pliomastotloit vexiUarius Matthew, front ami 
side views. After photograph kindly forwarded to the present author by the 
University of California (cf. Matthew, 1930, PI. XLiv). About one-sixteenth 
natural size. 



'Gidley's typ*- of 'Anancu.i' bensonensis is now referred to Cnntillerion bensonensis (see p. 565 below)- 




Fig. 109. Restor.^tion (1933) of Male and Female Mastodo.n amehicanus by Margret Flinsch, 

UNDER THE DIRECTION OF HeNRY FaIRFIELD OsBORN 

One-fiftieth natural size 
This restoration is based largely upon a record specimen (Fig. 114) of the American Mastodon in the Geological Museum of the Ohio State University, 
namely, with an estimated shoulder height in the flesh of 10 ft. 2 in.; it also embodies certain characters of the Warren Mastodon. The background is 
taken directly from a snow scene in the Hudson Highlands, near Newburgh, New York, where the Warren Mastodon was found. 





MASTODON AMERICANUS 3l05(lim.,IO'2/4" 
N AMERICA 



PLIOMASTODON VEXILLARIUS 28Q0mni,,9'£V4'e 
CALIFORNIA 




MIOMASTODON MERRIAMI 

NEVADA 



225omm.,7'4'e 





TURICIUS TURICENSIS 2b00ri(n.,8'6i/4"e 

SWITZERLAND 



ZYOOLOPHODON BORSONl 300omm.,<3'io" e 
ITALY IM. EURASIA 



PALAEOMASTODON BEADNELLI 23£8nim,7'7V->" 8 
N AFRICA FAYUM 



Fig. 110. Comparative Restorations of Pal.eomastodon, Pliom.astodqn, Miom.i.stodon, Mastodon, Zygolophodon, and Turicius 

One-hundredth natural size 
Mastodon americanus, Mus. Ohio State Univ., adult bull of maximum size, Turicius tuHcensis, restored after type specimen of Schinz, in the Zurich 

restored after skeleton (Fig. 114). Collection (Fig. 162). 

Zygolophodon borsoni, restored after grinding teeth from Asti (Fig. 154) 
Pliomaslodon vexillarius, Mus. Univ. Calif. 28301, restored after type and superior and inferior tusks after Schlesinger (p. 211 below). 
cranium and tusks, also femur, as described and measured by Matthew (pp. Palxomastodon beadnelli, as restored by Osborn (Fig. 97) from jaw, palate. 

161-163 above). and skeleton described and figured by Andrews, 1906 (Fig. 96). 

(Osborn, 1934) Palseomaslodon is now removed from its supposed ancestral 
Miomailodon merriami, restored after type specimen from Nevada and relationship to Mastodon. As here restored, also in fig. 97, p. 149, the mouth 
referred specimen from Colorado (pp. 154-15C above). parts and proboscis may be incorrectly drawn. 

164 



9. HISTORICAL REVIEW OF THE DISCOVERY OF THE GENUS MASTODON (1705-1935) 
AND OF THE SPECIES MASTODON AMERICANUS (1792-1869) 

Nomenclature. — The full history of this classic name is related in Chapter I, "Discovery of the Living and 
Extinct Proboscideans," in Chapter V, "History of the Classification of the Mastodontoidea, Families and 
Subfamilies (1705-1927)," and again in the introduction of the present chapter (Chapter VI), "History of the 
Subfamily Mastodontinae, the True Mastodonts." 

The first full summary of the vast generic and specific synonymy was that of Joseph Loidy in his great Memoir 
of 1869, p. 392. In our more recent and searching revision of the nomenclature of the Proboscidea we have found 
only one additional quaint and abortive name, Mastotherium megalodon Fischer de Waldheim, but doubtle.ss 
further bibliographic research will in the future reveal still others. 

Elephas americanus Kerr, 1792, Big-Bono Lick, Kentucky, Tetracaulodon Collinsii Hays, 1834. 

near the Ohio River. Mastodon Ohiolicum 1832, Neues Jahrb. f. Min., p. 355, 
Ohio-Incugnitum Blumenhach, 1797, same (?) locality. Gervais, 1848-1852. 

Mammul ohioticum Blunicnbach, 1799, same locality. Tetracaulodon Godmani Hays, 1834. 

Elephas macrocephahis A. Camper, 1802, same (?) locality. Missourium kochii Koch, 1840, Jefferson Co., Missouri. 

Le Grand Mastodonte Cuvier, 1806, same locality. Leviathan Missourii Koch, 1841, Mi.ssouri. 

Mastodonte dc I'Ohio, ("uvior, 180G, same locality. Tetracaulodon Tapi/rvides Koch, 1841, Missouri. 

Harpagmotherhim canadense Fischer de Waldheim, 1808, re^racaMZodon Osa*/;';' Koch, 1841, Missouri. 

Ohio (?) River (cf. Sherborn, 1924, p. 1022, "Harpagono- Tetracaulodon kochii , Koch, 1842, Missouri. 

therium canadense . . . Anim. foss. Siberie."). Tptracaidodon Haijsii CMvant, 1842. 

Mastothniutn megalodon Fischer dc Waldheim, 1814, Ohio. Tetracaulodon Bucklandi Grant, 1842. 

Mastodon giganteum CuvicT, 1817, Big-Bone Lick, Kentucky, Missourium Theristocaidodon Koch, 1843. 

Ohio River. Mastodon rufjatum Koch, 1845. 

Mastodon maximua Cuvier, 1824, Big-Bone Lick, Kentuckj', Elephas Ohioticus de Blainville, 1839-1864. 

Ohio River. Mastodon Ohioticus Falc. and Caut., 1845. 

Telracaidodon Mastodontoideum Godman, 1830, near New- Elephas Ruperlianus Richardson, 1854, Swan River, Lake 

burgh, Orange Co., New York. Winnipeg basin, Canada. 

Maxtodon Cuvieri Hays, 1834. Trilophodon ohioticus Falconer, 1868. 

Mastodon Jeffersoni Hays, 1834. Mastodon americanus Leidy, 1868. 

To our present knowledge the above names are chiefly synonyms of Elephas [Mastodon] americanus Kerr, but 
it is probable that some of these names were applied to specific or subspecific stages distinct from the type speci- 
men Kerr had in mind, which came from the Big-Bone Lick of Kentucky. We may at present consider only the 
names of Kerr and of Blumenbach, and of the species defined subsequent to Leidy 's Memoir of 1869, which are as 
follows : 

Elephas americanus Kerr, 1792. Mastodon moodiei Barbour, 1931. 

Ohio-Inaignilum IMumenhach, 1797. Mastodon raki Frick, 1933. 

Mammut ohioticum Blumenbach, 1799. Mastodon americanus alaskensis Frick, 1933. 

Mastodon rugosidefis Leidy, 1S90. Mastodon grangeri Barbour, 1934. 

Mammut progcnium Hay, 1914. Mastodon pavlowi sp. nov. 

Mastodon americanus plicatus Osborn, 1926. Mastodon acutidens sp. nov. 

Discovery and Description of the Ohio Mastodon from Guettard, 1752, to Buffon, 

1778, and to Cuvier, 1806-1836' 
Cuvier (1821, p. 250) states that the first engraving of "une grande molaire de VOhio" is that of Guettard 
(Hist. Acad. Roy. Sci. avcc Memoires, Paris, 1752 [1756], p. 360, PI. in, fig. 1 [Pis. 11, 12 of Memoirs]), but that 
the teeth of this animal did not attain real celebrity in Europe until between 1760 and 1770 through the 
Memoirs of Collinson (Phil. Trans., London, LVII, Pt. II, 1768, p. 468) and of Hunter (Phil. Trans., London 
LVIII. 1769, p. 34). 

'Continued from Chapter V, pp. 1 18 129 of the present Memoir. 

165 



166 



OSBORN: THE PROBOSCIDEA 



The name Ohio, moreover, is historic, since as early as 1765, as we learn from George Croghan's Journal and 
from Buffon (Biiffon, 1778, p. 505): 

. . . dans les contrees voisines de la riviere d'Ohio, environ a 4 milles sud-est de cette riviere, eloignee de 640 milles du fort 
de Quesne, (que nous appelons maintenant PifebMrgf/i) . . . a vu, aux environs d'un grand marais sale, oii les animaux sauvages 
s'assemblent en certains temps del'annee, de grands os & de grosses dents, & qu'ayant examine cette place avec soin, il a decou- 
vert, sur un bane eleve du cote du marais, un nombre prodigieux d'os de tres-grands animaux, & que par la longueur & la forme 
de ces os & de ces defenses, on doit conclure que ce sont des os d'elephans. 

Extracts from Croghan's Journal, communicated to Benjamin Franldin, appear in Buffon 's Supplement {op. 
cit., p. 507) accompanied by six admirable life-size steel engravings of mastodont grinding teeth, also citations 
from two little Memoirs by Collinson read before the Royal Society of London, in which ColUnson remarks 
(Buffon, op. cit., p. 509): 

Le marais sale ou Ton a trouve les os d'elephans, n'est qu'a quatre milles de distance des bords de la riviere d'Ohio, . . Les 
ossemens d'elephans se trouvcnt sous une espece de levee ou plutot sous la rive qui entoure & surmonte le marais a cinq ou six 
pieds de hauteur; on y voit un tres-grand nombre d'os & de dents qui ont appartenu a quelques animaux d'une grosseur 
prodigieuse ; 




Buffon, 1778. Superior GRiNmNG Teeth (A, Al) of Zygolophodon and (B) of Mastodon. Originals said to be in the Paris Museum 

(fide Pavlow, 1894) 

Fig. 111. Grinding teeth of Zygolophodon borsoni referred (A, Al) and of Mastodon amcricanus referred (B), after Buffon, 1778, Pis. i, ii, and iv, one- 
fourth natural .size. (Inverted by H. F. 0.] 

A, Al, Buffon 's figures of the molar of an animal which we now know to be related to Zygolophodon dorsom, found in Rassia ("la petite Tartarie"), pre- 
sented to Buffon in 1770 by M. le Comte de Vergennes. Observe three to four lobes, i.e., cones or conelets, in each transverse crest. (1778, Pis. i and ii.) 

B, Buffon's figure of the molar of an animal now known as Mastodon americanus, from Big-Bone Lick, Kentucky, near the Ohio River, sent to Buffon 
by M. Collin.son. Ob.serve two lobes, i.e., cones, in each transverse crest, the true Mastodon type. (1778, PI. iv.) 

Both of these beautifully engraved molar teeth appear to be third superior grinders, M\ although they lack the rudimentary fifth crest or pentaloph 
characteristic of the third .superior molars of Mastodon americanus (Fig. 133). 



To Collinson's characterization, Buffon {op. cit., p. 510) adds: 

C'e (|ue (lit ici M. Collinson, est tres-viai; ces grosses dents molaires different absolument des dents machelieres de 
I'el^phant, & en les comparant k celles de I'hippopotame, auxquelles ces grosses dents ressemblent par leur forme quarree, on 
verra qu'elles en different aussi par leur grosseur, etant deux, trois & quatre fois plus volumineuses que les plus grosses dents 
des anciens hippopotames trouvees de mcme en liberie & au Canada, quoique ces dents soient elles-memes trois ou quatre 
fois plus grosses que celles des hippopotames actuellement existans. 



THE MASTODONTIN.E: MASTODON 167 

Buffon {op. cit., p. oil J rightly compared the mastodont and hippopotamus teeth and rightly concluded: 

. . . comme on pout 1p voir pncomparaiit los figiirpsdos/;^;?)c/(f-.s i, ni& iv,avec cclk'sdcla^/^nWff v. Ilparoit done certain 
que ces grosses dents n'ont jamais appartonu a Tolophant ni a rhippopotamo; . . . jc crois done poiivoir prononeer avec fondc- 
inent que cette tres-grande espece d'animal est perdue. 

In brief, Buffon fully characterized this animal as a mastodont and designated it as an extinct species belong- 
ing to the epoch of the elephants but did not name it. This is sufficient evidence that the name Ohio was very 
prominent in the minds of the naturalists of the period, because the writings of Buffon dominated Europe at the 
time. 




6/-.//»y M.fS'l 'OUil. \TK . I'l.I, 

Fig. 112. Cuvier's types of he (hand Maslodonte, or Mastodonte de I'Ohio, 180(1. .\ftpr Cuvier, 1806.2, PI. 49|i|, one-half 
natural size. 

(Cuvicr, 1830, Atlas, PI. 19): Tig. 2. Molaire a dix pointo.s encore intactcs, donnoc par le due dp Plaisance. F'ig. 1. Dent 
a (lix pointes en partie us(''e, <lu cabinet de .loubert, vue de profil. Fig. 3. Le meme, vue par sa couronne. Fig. 4. Dent a dix 
pointes, rc'-trtoc en arricre, communiquee par feu M. Tonnellicr. Fig. .i. Molaire k six pointes a de'mi uscje, d'apre.'; uii dp.s,sin de 
M. Blumenbach. 



CuviER, 1806-1836. — Cuvier practically defined the genus 'Mastodonte' in his admirably engraved Planches 
and in the descriptive legends of these plates. In the descriptions he acknowledges his indebtedness for one of the 
drawings to Blumenbach and for one or two of the cotj'pe specimens to Fabri, to Daubenton, to Joubert, and to 
Tonnellier. Referring to Buffon's idea that the mastodont grinders were those of the "hippopotames gigan- 
tesques," he pointed out that they were distinguished by the worn 'losanges' which differed greatlj' from the worn 
"trifles de I'hippopotame." 

(Cuvier, 1800.2, p. 293): "Notre pi. i roprospnte quatre de ces dents de mnslndonic a nioitie grandeur. Fig. n en est une h 
six pointes ;\deini-u.sees:elle est copi<5e d'apres un dessin qu'a bien vouluni'envoyer M. Blumenbach. Nous en avonsau Mu.s(?um 
trois pareilles, anciennement rapportdes par Fabri. Ce sent elles que Daubenton (Hist. nat. XII, n." 1 106, 1 107, 1 108), et Buffon 
{Epoque.s (if la milure, pi. v) out pr'mospourdosdvnisd'hippopotnme.sgigante.tqucs. Elles .sont ais(^es a distinguer par ces lo.savges, 
dont notre figure donne une idee fort juste, et ([ui different heaucoup des trefles de \'hippopolnm< . D'ailleiirs Vhippopotanw n'a 
jamais que quatre trifles et non pas six." Compare Daubenton, 1764, p. 77, No. MCXI "Dent pltrifiie qui a quflque rapport a 
celle.^ de ritippopoinme." 



168 



OSBORN: THE PROBOSCIDEA 



Original Description of Elephas americanus by Kerr, 1792. No Figure 

Succeeding Buffon's description (1778) of the American Mastodon but anticipating Cuvier's description of 
Le Grand Mastodonte, 1806, the Scottish naturalist Robert Kerr proposed the specific name Elephas americanus 
for tusks and grinders found in the 'Big-bone-swamp' on the banks of the Ohio. He pointed out that these teeth 
were specifically different from those of the elephant and furnished with a double row of high conic processes 
reseinbUng those of carnivorous animals. This contribution appears in a translation of portions of at that time 
a recent edition of the "Systema Naturae" of Linnaeus, with improvements by Professor GmeUn, entitled, "The 
Animal Kingdom or Zoological System, of the Celebrated Sir Charles Linnaeus," pubhshed in 1792, as follows: 

(Kerr, 1792, p. 116): 2. American Elephant. — Elephas americanus. In America, on the banks of the Ohio, are found, 
several feet below the surface, in a marshy place called Big-bone-swamp, great numbers of tusks and grinders, supposed by many 
to belong to the Elephant: But the grinders are totally different, being covered uniformly with enamel, and furnished with a 
double row of high conic processes, like those of carnivorous animals; whereas those of the Elephant are composed of alternate 
perpendicular layers of bone and enamel, and are ribbed transverselj' on their upper surfaces, like those of graminivorous quad- 
rupeds : Hence the species must be entirely different ; and Mr. Pennant has chosen to suppose that they have belonged to an 
unknown species of this genus, which he names the American Elephant. Hist, of Quad. n. 71. 



Original Description of the Ohio-Incognitum by Blumenbach, 1797 

The next name applied to the animal was that of Johann Friedrich Blumenbach, pioneer vertebrate palae- 
ontologist of Germany, who received from the British Museum his type specimen (Fig. 113), a third superior 
molar tooth of the left side. Blumenbach subsequently corresponded and exchanged drawings and notes with 
Cuvier, as mentioned above. 




Fig. 113. A. Type left third superior molar 
of Ohw-Incognilum Blumenbach, 1797. Repro- 
duced in facsimile after Blumenbach. 1797.2, 
No. 19. Same figure in edition of 1810 desig- 
nated on the Plate as Ohio-1 ncngniium. Techni- 
cally a lype figure. ( = Mastodon americanus.] 



(Blumenbach, 1797.2, No. 19) : Bey jencm fossilen Ungeheuer der praadamitischen 
Vorwelt sind die zackichten Kronen ganz mit einer starken Lage vom so genannten 
Schmelz (substantia denlium vitrea) iiberzogen: da hingegen die Backenzahne beider 
Gattungen des Elephantengeschlechts aus vertical liegenden abwechselnden Schichten 
von Schmelz und Knochensubstanz bestehen. . . . Der vom fossilen Incognitum 
hingegen nach einem ungeheuern am Ohio ausgegrabenen Exemplar, das mir aus 
dem Britischen Museum fiir meine Sammlung iiberlassen worden. 

Osborn, 1922: Ohio-Incognitum is the name and figure cited by de 
Blainville (1839-1864, p. 245). We observe that Blumenbach's figure of 
Ohio-Incognitum (1797. 2, No. 19) is of a third superior molar of the left side, 
with very pronounced internal cingulum. There are four and a third lophs; 
there are two mamillate lobes on each crest in the figure. The above 
description by Blumenbach {op. cit., No. 19) shows that he secured this speci- 
men from the British Museum as part of its Ohio River collection. 



The Name Mastodon americanus as established by Leidy, 1868, 1869 

Specific Name. — Leidy concludes his full revision of the American species of Mastodon, in his great Memoir 
of 1869, with the .sentence (Leidy, 1869, p. 240): "One of the species, the well known Mastodon ohioticus, or M. 
americanus as T shall hereafter call it, appears to have roamed throughout the continent during the quaternary 
period." The dates of these species are Elephas americanus Kerr, 1792, and Mammut ohioticum Blumen- 
bach, 1799; to the best of our knowledge Blumenbach (1799) was the first to use the name Mammut 



THE MASTODONTIN^: MASTODON 



169 



ohiolicum; in the Neues Jahrbuch, 1832, p. 355 (see Eichwald), the term Mastodon ohioticum is used, and Leidy 
states in his Memoir of 1869, p. 394, that Gervais used this term 'ohioticum' in his "Zoologie et Paleontologie 
Francaises," 1848-1852, I, p. 187. The name Mastodon americanus was first used by Leidy in 1868, in Volume 
XX of the Proceedings of the Academy of Natural Sciences of Philadelphia, p. 175, to embrace the Mastodon 
ohioticus or M. giganteus of authors. 

Generic Name. — The generic name Mastodon was universally accepted by palaeontologists in all parts of 
the world until 1902, when the vernacular generic term Mammut of Blumenbach (1799) was revived by Hay, 
although sub-sequent to the vernacular term Ohio-Incognitum of Blumenbach. Unfortunately this vernacular 
synonym 'Mammut' entered the American literature between 1902 and 1925 and occurs in the writings of Palmer, 
Lull, Barbour, Stock, Daggett, and others. In Chapter I of the present Memoir, "Discovery of the Living and 
Extinct Probo.scideans," and in the foregoing pages of the present chapter (Chap. VI), "Historj^ of the 
Subfamily Mastodon tinae, the True Mastodonts," ample reasons are given for the banishment from all future 
proboscidean hterature of this barbaric term 'Mammut.' 

Subsequent to Leidy's Memoir of 1869 nine species have been proposed, namely. Mastodon rugosidens Leidy, 
1890, 'Mammut' progenium Hay, 1914, Mastodon americanus plicatus Osborn, 1926, Mastodon moodiei Barbour, 
1931, Mastodon raki Frick, 1933, and Mastodon americanus alaskensis Frick, 1933. The validity of Leidy's M. 
rugosidens, 1890, and of Osborn's M. americanus plicatus, 1926, is a question of further research. Mastodon 
pavlovn, also Mastodon acutidens, a very progressive stage, are proposed by Osborn in the present Memoir (see 
Appendix). In 1934 Barbour described the straight-tusked species Mastodon grangeri from Nebraska (see 
Appendix at close of Volume I for description and figure). 




MASTODON AMERICANUS 

Ohio State Z/T7it^. 



The Largest Fully .\ddlt Male Skeleton of Record; Estimated Shoulder Height 10 Feet 2 Inches in the Flesh. (See Fig. 109) 
Fig. 114. Mounted skeleton of Maslodon americanus from Catawba, Oliio, in the Ohio State University, Geological Museum, aft«r photograph kindly 
sent the present author by Prof. J. Ernest Carman, Chairman of the Department of Geology. Reproduced about one twenty-fourth natural size. 



10. CHARACTERS OF MASTODON AND ITS INCLUDED SPECIES AND VARIETIES 

Superfamily: MASTODONTOIDEA Osborn, 1921 

Family: MASTODONTID^ Girard. 1852 

Subfamily: Mastodontin^ Brandt, 1869 — Osborn, 1910 

Genus: MASTODON Cuvier, 1806-1817 

Original reference: Mastodonte, "Sur Le Grand Mastodonte," Cuvier, 1806.2, pp. 270, 272, 293; Mastodon, "Le Regne Animal," 
Cuvier, 1817, pp. 2.32, 233. 

Genotypic species: Mastodon giganteum. 

Synonyms: Mastotherium Fischer de Waldheim, 1814, pp. 337-.341. 

Generic Characters. — Incisive tusks without enamel; superior tusks large and rounded with 
slightly indicated annular ring growths. Inferior tusks straight, cylindrical, variable in old age. 
Marked sexual disparity in female tusks. Permanent premolars suppressed,' except the vestigial P^-F^. 

Dental formula: 1%^4^% Dpf^l Pi (vestigial) M+^f 

Intermediate molars, Dp 4-M 2, trilophodont, i.e., with three ridge-crests. Ridge-crests progressively 
elevated, lophodont to subhypsodont. Cingulum of grinding teeth slightly stronger on the inner side 
than on the outer side (in Serridentinus much stronger on the inner side). Grinders bilobate with 
strong median sulcus between the inner and outer lobes. Very rudimentary trefoil spurs on the superior 
inner lobes and on the inferior outer lobes (compare Fig. 133). Summits of lobes simple or crowned with 
small conelets. Progressive elevation, subhypsodonty, of the lobes (cf. Figs. 112, 135). Progressive pli- 
cation of the surface enamel of the grinding teeth (Fig. 120). Progressive development of the third crest 
(tritoloph and -lophid) in intermediate molars; progressive development of the fifth crest rudiment 
(pentalophid) in third inferior molars. Progre.ssive reduction of the rostrum of the lower jaw (see 
M. progenius and M. americanus) . 

The above list of generic characters enables us to distinguish clearly the Pleistocene species of the true 
American Mastodon from their Pliocene and Miocene ancestors Pliomastodon and Miomastodon. Whereas M^- 
M3 show a rudiment of the fifth crest in Mastodon americanus (Figs. 116, 117), no such rudiment is observed in 
Miomastodon merriami of the Middle Miocene. These progressive characters of the grinding teeth in the Masto- 
dontinse phylum are well displayed in figure 98, also in PI. i, pp. 134—135. 

Mastodon americanus Kerr, 1792 locality both of the genus 'Mastodonte' { = Mastodon) Cuvier and 

Figures 76, 112, 114, 116, 117, 124-126, 128-130, 132-134, 136, PI. i, p. 134 of the species Elephas[ = Mastodon] americanus Kerr. This famous 

Big-Bone Lick, Boone County, Kentucky, late Pleistocene, IV Glacial = salt swamp or salt lick was probably visited by the ungu- 

Wiscon.sin time. lates for a long period of geologic time, as shown by the latest 

The history of this classic species is fully recited above in the analj'sis of its contained fauna (see Hay, 1923.1). In the list of 

present chapter of this Memoir. Cuvier's types of the genus species by Hay (p. 403) cited below references are made in his 

'Mastodonte' { — Mastodon) are reproduced above in figure 112. Memoir on "The Pleistocene of North America and Its Verte- 

Unfortunately Kerr's type of the species Elephas [ = Mastodonl brated Animals" to pages whereon further information is given 

americanus was not figured. regarding the respective species. 

Type Description. — Kerr's type description is cited in full Hay believes (op. cit., 1923, p. 403) that the Big-Bone Lick 

above (p. 168) and need not be repeated here. The type locality, fauna followed Glaciation III ( = Ilhnoian) and belongs chiefly in 

however, deserves very careful description. Mastodon ameriramis 3rd Interglacial time ( = Sangamon interval); he accordingly 

from Big-Bone Lick, Kentucky, as figured by Cuvier (see Fig. 112 assumes that the animals were buried during the Sangamon stage 

of the present Memoir) probably belongs to IV Glacial = Wiscon- in the interval between the Illinoian (III) and the Wisconsin (IV) 

sin time; the molars are much more primitive than those of M. glaciations; the presence of Mammonteus primigenius in the Big- 

acutidens sp. nov. Bone Lick deposits rests (op. cit., p. 146) on the determination of a 

Type Locality of the specie.s Mastodon americanus. — single superior molar tooth with 23-24 plates, which Osborn con- 
Big-Bone Lick, Boone County, Kentucky, is undoubtedly the type siders may be referable to Parelephas jeffersonii. 

'See page 138 for observations of Hays and of Warren, namely, that P4 forms in the jaw but does not erupt; it is therefore vestigial. 

170 



THE iMASTODONTIN^: MASTODON 



171 



Mcfialonyx jeffcr.suiiii (p. 44 j. 
Mylodon harlani (p. 44). 
Ec|uus coniplicatus (p. 202). 
?Tapirus haysii (p. 209j. 
Odocoileus virginianus (p. 234). 
("orvus caiiadcn.sis (p. 243). 
('prvalfcs sootti. 
Alces aniericanuis. 
Ranp;ifor caribou (p. 247). 



iiootheiiuiu hoiiibilruii.s (p. 255). 
Symbos cavifrons (p. 255). 
Bison anti(nuis (p. 265). 
Bison bison (p. 270 j. 
.Maiiiinut iuiioricanuni (p. 128). 
lOlcphas priniigr'nius fp. 14(5). 
Klcpbas coiuMibi (p. 100). 
Ursus anirricamis. 



[ = Mastodon americanus type 
= MdmmonUua primigenius ref.(?) 
= Parclepliax jefferfsunii ref.(?jl 



Mastodon americanus rugosidens I-cidy, 1890 

P^ixvire ll.'i 

SantPc Ueds of Beaufort County, South Carolina; Pleistocene. 

Mastodon rur/osirfens Leidy, 1890. "Mastodon and Capybara 

of South ( aroiina." Proc. Acad. Nat. Sci. Pliila., Vol. XLIl, 1890, 

p. 184. Type. — . . . "conipletecrownof a last upper llowcr] 

molar, strikiuKly different from that tooth in the connnon American 

Mastodon . . . looking as if carved out of ebony, as is the case with 

many of the fossils from the same and similar localities." Typk 

bocwLiTY. —"It was found in the Santee Beds of Beaufort Co., S. 

M rugosidens 

Drawn from 14445 Cast of type 



TvpK OK Mastodon 

AMi:UIC.\NUH RUaOSlDKNS 

Fir. 1 1.'). bast inferior 
molarof the left side, I.Ms. 
Tyix" of Maxtndon rtignsi- 
(lcii.1 Leidy (Philadelphia 
Academy). Drawn from 
east of type (.\mer. Mus. 
14445), one-third natural 
size. 

A, Crown view; Al, 
external view; A2, inter- 
nal view. 




C." and "was presented to the Academy by Mr. .lames R. 
McKee." Type Figure. — Drawn from ca.st of type (Amer. 

Mus. 14445 — Fig. 115 of the present Memoir). 

Type Description. — (Leidy, 1890.2, p. 184); "It is worn only 
on the summits of the anterior pair of lobes, which display the 
usual e.xpo.sed tlentinal areas. Notwithstanding the many species 
of Mastodon which have been recorded in Nortii and South 
America, the present tooth seems to indicate a different one. It 
more nearly resembles the corresponding tooth of the M Jloiiddniis, 
recently described, or that of tlie M. aitgitstidcrits of Europe, than 
of the M. americanus. In comparison with the molars of these and 
other known species the tooth is remarkable for the greater propor- 



tionate length of the constituent lobes of the crown and their 
conspicuously wrinkled condition. The wrinkling is longitudinal 
and regular and apparently not the result of an al>normal state. 
Similar wrinkling is olxserved in some s|)ecimens of the same teeth 
in M. americanus, but mainly confined to the intermediate vallies 
of the crown, while it is well produced laterally in the present 
fossil. From the comi)aratively more prolonged condition of the 
lobes, the summits of the inner ones appear more tapering or 
narrowly pointed than in M. florida/ius; while with the fore and 
aft extensions of the same lobes the summits form acute and not 




./ 



Fig. 1 Hi. Type left third superior molar, l.M', 
of Ohio-incogtiilum filumenl>ach, 1797. Repro- 
(luecd in facsimile after Bhunenbach, 1797.2, 
No. 19. Same figure in edition of 1810 desig- 
nated on the Plate as Uhio-I ncognitum. [ = Mas- 
todon americanus.] 




Fig. 117. Inferior molar of the left side, l.Ma, of 
Mastodon americanus. One-fourth natural size. The 
record of the specimen, from which this beautiful 
drawing was made by Mr. H. Weber, has been lost. 

obtuse angles as in the latter. Regarding the specimen as indicat- 
ing a previously unknown species, this may be distinguished as the 

M.\.STOlJ(>.\ lUCOSIDKNS. 

Comparative nu^asuri'mcnts of the tooth are as follows; 

^f. rugo.iiden.t M. fioridanus 
Length of crown fore and aft 19(1 mm. 190 mm. 

Bictidth of ba.se at fore part 92 nun. 96 mm. 

Length of second inner lobe 90 mm. 80 mm. 

Length of second outer lobe 80 mm. 62 mm." 



172 



OSBORN: THE PROBOSCIDEA 



Osborn, 1925: The type cast (Amer. Mus. 14445), from the 
original type in the Academy of Natural Sciences, Philadelphia, 
agrees exactly in measurement with Leidy's description (Leidy, 
1890.2, p. 184) cited above; thus there is no mistaking the type. 
Leidy erred, however, in describing it as "a last upper molar"; 
it is certainly a last lower molar of the left side, I.M3, as shown by 
careful comparison of the type with third inferior grinding teeth 
and with the specimen beautifully displayed in figure 117. 

It is doubtful also whether the wrinkling of the enamel, to which 
the specific name rit/josidens applies, is actually of specific value. 




Type of Mastodo.n pkogenius 
Fig. 118. Type jaw of 'Mammut' progenium Hay, 1914, PI. .\i.iv, figs. 1 
and 2. From Cox gravel pit, Missouri Valley, Harrison County, Iowa. 
Described by the present author as Mastodon progcnius. 

Mastodon progenius Hay, 1914 
Figure 118 
Cox gravel pit, Missouri Valley, Harrison County, Iowa, early Pleisto- 
cene, probably 1st Intergtacial, Aftonian age, fide Hay. 

This species is of great interest as representing an early Pleisto- 
cene stage in the development of the true Mastodon. Of 1st Inler- 
glacial or Aftonian age it is much more primitive than the type of 
Mastodon nmericanus of Srdif) Interglacial or Sangamon age. 

Mammut progenium Hay, 1914. "The Pleistocene Mammals of 
Iowa." Iowa Geol. Surv., 1914, XXIII (Ann. Rept., 1912), pp. 
368-373. Type. — Lower jaw of a veiy old animal, with last 

molar in each side of jaw, worn down to the roots. Univ. of Iowa 
No. 292. Horizon and Locality. — Apparently found in 

1910, in the Cox gravel pit, at MisfBouri Valley, Harrison County, 



Iowa; of Aftonian age. Type Figure. — Hay, op. cit., 1914, 

PI. XLiv, figs. 1 and 2. "It was described and figured by Calvin 
in 1911 (Bull. Geol. Soc. Amer., Vol. XXII, p. 213, Pis. xx and xxi) 
under the name Mastodon americanus. Figures are here presented 
which are made from the same photographs as those used by Calvin 
. . . but reduced in size" (p. 368). 

Type Description. — (Hay, 1914, p. 369) . . . "symphysis 
of lower jaw longer than in M. americanum; chin less constricted 
at the symphysis, as viewed from above, and not truncated in 
front. . . . Thefollowingarethedimensionsof this jaw as furnished 
by the measurements of Calvin and the writer [Hayj. 

Length from fiont of symphysis to hinder border just 

above the angle 830 mm. 

Length from fiont of symphysis to rear of condyles, in 

straight hne 920 mm. 

Height of coronoid process above lower border of the jaw 425 mm. 

Height of condyles above lower border of jaw 410 mm. 

Length of the symphysis 212 mm. 

Height of the jaw at the front of m.3 192 mm. 

Thickness of the jaw at the front of m.3 120 mm. 

Width of the ascending ramus, from rear of the condyle . 290 mm " 

The symphysis of M. progenium is relatively much longer 
than in M. americanum; its length equals eighty per cent, of the 
width of the ramus. The rami cease to diverge opposite M3 and 
there is even some constriction. "Occasionally in the lower jaw of 
M. americanum there is a single tusk, rarely two of them ... in 
the lower jaw of the extremely old animal under description here, 
both tusks were present. Their presence is indicated by the widely 
open sockets. The vertical diameter of the socket measured, the 
right, is 73 mm. ; the transverse diameter, 50 mm. These measure- 
ments show that the tusks were far larger than those now and then 
found in M. americanum. The depth of the socket is 160 mm. 
The measurements show likewise that the tusks were confsiderably 
compressed, the horizontal diameter being about two-thirds of the 
vertical. What the form and the length of these tasks were, 
beyond the sockets, we can only surmise. ... In the type jaw the 
lingual gutter is somewhat peculiar in being overhung on each side 
by the upper borders of the jaw. These approach until they are 
only 28 mm. apart. On the outer face of this part of the jaw the 
surface is concave as it rises to the dental border. In this specimen 
the penultimate molar, m.2, had been pushed out on the right side 
of the jaw before the death of the ammal; that of the left side 
seems to have been lost after death, for there remains a part of one 
root. As stated, the crown of m.3 is worn down to its base. The 
grinding surface of the left molar forms a concavity which is sur- 
rounded by a ring of enamel ; but on the right side a part even of 
this is missing. This attrition of the tooth had so weakened it that, 
before the death of the animal, the teeth had each split into two 
parts. The inner wall of each had broken at the middle of the 
second crest and the cleft had run backward and outward to near 
the hinder end of the tooth. That this had occurred before death 
is evident, as Calvin remarked, from the fact that the edges of the 
fracture had been rounded off. The right tooth has another cleft, 
which crosses its front; but Calvin concluded that this was a post- 
mortem break. It is not improbable, however, that it happened 
under the strain of chewing just before death of the animal. It is 
evident that the hinder molar had four transverse crests and a 
heel which was essentially a crest." 



THE MASTODONTIN^: MASTODON 



173 



Referred Specimen. — Left ramus and symphysis of lower 
jaw, figured by ("aivin under the name Mdiiunut timcricanum 
(Calvin, 1909, p. 352, PI. xxv, fig. 2), referred provi.sionally by Hay 
to Mammut progenium because of the resemblanco.s of the jaw to 
the type. The referred specimen is figured by Hay on PI. XLViii, 
fig. 1; PI. ui, fig. 2. 



Type of Mastodon 
orkoonensis 

Fig. 119. Type figiiics 
of Mammut [= Ma!sloilon\ 
oregonenae Hay, 192G, figs. 
1, 2, about one-lialf natural 
sizfi, from Rye Valley, on 
Dixie Creek, Baker County, 
Oregon (Nat. Mus. -igil). A 
second suiXTior molar of the 
leftside, I. Ml The crenulate 
crown.s resemble those of 
Mastodon americanus /ilicalus 
Osborn (Fig. 120), a synonym. 




Mastodon oregonensis Hay, 1926 

Figure 1 19 
Rye Valley, Dixie Creek, Baker County, Oregon. Probably Pleistocene. 
Compare Mastodon americanus plicatus Osborn, 1926. 

Mammut oregonense Hay, 1926. "Two New Pleistocene 
.Mastodons." Journ. Washington Acad. Sci., Vol. XVI, No. 2, 
Jan. 19, 1926, pp. 39 to 41. Type.— A second superior 

molar of tlie left side, \.M-, Nat. Mus. 4911. Horizon and 

Locality. — This molar was "found by the Cartwright Brothers, 
placer miners, at Rye Valley, on Dixie Creek, in townsliip 13 
south, range 43 east [Oregon]." Probably Pleistocene. Type 

KiouRE. — Op. cit., p. 36, figs. 1 and 2. 

Type Description. — (Hay, op. cit., 1926, pp. 39 and 40): 
"In the U. S. National Museum is a mastodon tooth (Cat. no. 4911) 
which was sent tliere in Novemlier, 1900, by Dr. Waldemar Lind- 
gren, from Baker City, Baker County, Oregon. It had been found 
by the Cartwright Brothers, placer miners, at Rye Valley, on Dixie 
Creek, in township 13 south, range 43 east. Dr. Lindgren rcjiorted 
that the tooth liad been found in a fluviatile clay bed which had 
formed a part of a bench of auriferous gravels, overlying the 
I'ayette beds. He regartled the fluviatile clay as of Pliocene age. 
It appears more probable that the bed belongetl to tlie Pleistocene, 
for in it was discovered a tooth of Elephas columbi. . . . The type 
tooth here described and figured is the upper left second molar. 
It has been regarded as belonging to M. anwricanum, but it is so 
different that the writer ventures to give it a distinct name. The 
tooth had apparently not yet begun to suffer wear; or, if at all, 
only slightly on the first cross-crest. The length is 111 mm.: the 
width III' ihr fnint end, 74 niiii.; of the rear end, SO mni. The 



crown presents 3 cross-crests and, in the rear, a talon. The crests 
are high, and the valleys narrow. The ends of the cross-crests 
slope steeply and nearly equally. The summits of the two princi- 
pal cones of each crest are well separated, as follows: First crest, 
38 mm., second, 40 mm., third, 40 mm. ... AH around the tooth 
is a heavy cingulum composed, at the pretrite ends of the valleys, 
of 5 or 6 tooth-like conules. On the posttrite side the conules are 
smaller and more numerous." 

Mastodon americanus plicatus Osborn, 1926 
Figure 120 
Geologic age nneertuin, pos.sibly of Po.stglacial (IV) or post-Wisconsin age. 
Kr<iin Walnut, Bureau County, Illinois. 

The type of this species is a superior true molar series of the 
right side, r.M'"', from Walnut, Illinois, in the American Museum 
collection (Amer. Mus. 10666). 

Mantodon americanus plicatus Osborn, 1926. "Additional 
New Genera and Species of the Mastodontoid Proboseidea." 
Amer. Mus. Novitates, No. 238, Novemlx-r 30, 1926, p. 1. 
Type. — Superior true molar series of the right side, r.M'*'. Amer. 
Mus. 10666. Horizon and Locality. — Walnut, Illinois, 

(leologic age uncertain, possibly of Postglacial (IV) or post- 
Wisconsin age. Type Figure. — Op. cit., 1920.706, p. 2, fig. 1. 

Specific Ciiaracter.s.— (Op. cit., 1926.706, p. 1): "This 
progressive subspecies is distinguished by numerous minor fold- 
ings, valleys, and plications which break up all the surfaces of 
the lophs or lobes and which to a certain degree indent even 
tlie summits of the ridge-crests." 



M. americdnus 
Amer. Mus. 10666 




Fig. 



Type of M.^stodon .\mericanus plicatcs 
120. Tyj>e .superior true molar series of Mastodon americnnus 



plicatus Osborn f.\mer. Mus. 1(H)G6), from Walnut, Bureau County, 
Illinois. One-fourth natural size. After Osborn, 1926.706. p. 2, fig. 1. 
Compare Mastodon oregonensis Hay (Fig. 119). 

"As in the case of the type of Mastodon rugosidens Leidy, it is 
somewhat doubtful whether these foldings and plications represent 
constant specific characters. Consequently the teeth are given 
merely subspecific rank, pending a fuller study of the dental char- 
acters in the Mastodon americanus molar teeth in America." 



174 



OSBORN: THE PROBOSCIDEA 



Mastodon moodiei Barbour, 1931 

Figure 121 
West Blue River, about nine miles southwest of Milford, Seward County, 
Nebraska. Pleistocene, II Glacial, Kansan. See Pleistocene Correlation Chart 
in Appendix of the present Volume I. 

The type cranium of this sjjecies, discovered during the 
winter of 1931, was unearthed, together with the atlas, axis, 
two thoracic vertebrje, and ribs, between March and June of the 
same year. It was named by Doctor Barbour Mastodon moodiei, 
after Dr. Roy L. Moodie of tlie University of Southern Cahfornia. 

Mastodon moodiei Barbour, 1931. "The Milford Mastodon, 
Mastodon moodiei, Sp. Nov. A Preliminary Report." Neb. 
State Mus., Bull. 24, Vol. I, December, 1931, pp. 203-210 (Barbour, 
1931.3). Type. — Skull, together with atlas, axis, two thora- 

cic vertebrae, ribs, and mandible with teeth. Neb. Mus. 21-3-31. 



RebMus. 21-3-31 



protruded 6 or 8 feet or more. . . . By the position of the frag- 
ments the tusks were judged to have been 6 to 8 feet in length. 

Measurements 
Transverse diameter of condyle, 5 inches (127 mm.) 
Depth in front of the molar, lYs inches (180 mm.) 
Depth just back of the molar, 6% inches (172 mm.) 
Greatest thickness, 7 inches (182 mm.) 
Length of molar, Q% inches (175 mm.) 
Width of molar, 3% inches (96 mm.) 
Length of mandibular tusks, 9^4 inches (235 mm.) 
Diameter of tusk, 2 inches (51 mm.) 

The beds which yielded the relies of Mastodon moodiei are 
plainly Pleistocene in age, probably Etonian." 




Tifpe 

IJlastodon moodiei Barbour 



Fig. 121. Type of M.\.stodon moodiei Barbour, The Milford Mastodon of Nebraska, Morrill Pal.«ontological 

Collections (Neb. Mus. 21-3-31) 
(Left) Adult mandible with Mo, Mn, crown view. About (Right) Type skull and mandible. About one-twelfth 

one-sixth natural size. natural size. 

Rostrum relatively long, laterally compressed; short, blunt The mandible exhibits a downcurved rostrum with two 

incisive tusks worn off squarely at tips. Compare figure 87. blunt incisive tusks sharply worn off at the extremities. 



HoHizoN AND Locality. — Dam No. 7, across the West Blue River, 
about nine miles southwest of Milford, Seward County, Nebraska. 
Pleistocene, II Glacial, Kansan. Type Figure. — Barbour, 

op. cit., figs. 130, 131, 1326. See also new figures of tyi^e in the 
present Memoir (Fig. 121). 

Specific Charactkrs.— (Barbour, 1931.3, pp. 206, 207, 210): 
"Both mandibular tusks were .secured uninjured. They are dense 
and strong and slightly curved. They measure 9)4 inches (235 
mm.) in length, and the pulp cavity is shallow, measuring but X' 
inch (13 mm.) in depth. Each mandibular tusk ends squarely, 
for an indefinite amount has been worn off in some manner. To 
have thus ground them off it seems as though the creature must 
have designedly rubbed his tusks on rocks. Yet the interference 
offered by the great protruding upper tusks must have rendered 
such a procedure impossiljle. The incisive .sheaths .show that the 
upper tusks had a diameter of 4 inches, and they must have 



Mandibular Tusks as Recently Observed by Barbour 
1931).— The mandibular tusks of Mastodon americanus (cf. Barbour, 
1931.1, pp. 163, 164) were waning inheritances from the past, 
which plainly required ages for reduction and suppression, and 
complete elimination was never a reality. In the majority of 
jaws, the inferior tusks are wholly wanting; in others the tusks 
have not been erupted, but lie embedded in the bone; sometimes 
one or two small tusks are erupted; in still other cases large ones 
appear. 

The Warren mastodon in The American Museum of Natural 
History (Fig. 124) had a solitary dwarfed tusk in the mandible, 
and henffe was named Tetraraulodon, and the term might still be 
used conveniently to designate mastodons possessed of mandibular 
tusks. The tetracaulodont type of mastodon is not particularly 
rare, for there are two good examples in Amherst. Hay reported 
several, and Warren, studying mastodon jaws, found nine of this 



THE MASTODONTINiE: MASTODON 



175 



type. There is one example in tlie Nebra-ska State Miiwuni. In 
the museums of the eountrj' many others can doubtless l)e found 
The presence of mandibular tusks in mastodon jaws is partly a 
sexual character, those with tusks being males, those without, 
females. 

In the Amherst Museum is a jaw from Nine Mile Hottom, 
Southt'arolina, displaying inferior tusks 14 in. in length (Harbour, 
op. cjY., fig. 103), whereas the Warren Mastodon in the American 
Museum shows only one small vestigial mandibular tusk (op. cit., 
fig. 104). In the 'Monroe mastodon' of Orange County, New 
York, in the State Museum, Albany, there are two dwarfed and 
irregular mandibular tusLs. In the 'Seward County Mastodon' 
of Nebra-ska (Neb. Mus. 11-7-2.5), the two mandibular tusks 
were probably 12-18 in. in length {op. cit., fig. 106). 

From all these geologic, phylogenetic, and sexual variations 
of the mandibular tusks, we draw the induction that at the begin- 
ning of Pleistocene time not only was the rostrum of the mandible 
much longer than at the close of Pleistocene time, as illustrated in 
Mastodon progenium Hay, but the pair of lower mandibular tusks 
were much more fully develojied, as illustrated in the case of the 
'Seward County Mastodon' (Barbour, op. cit., fig. 106), which 
was found 75 feet below the surface in interglacial gravels. 

Mastodon granger! Barboiu', 19S4 

From near Pender, Thureton County, Nebra-ska. 

For full description, with figures, of Mastodon gmngeri, sec 
Appendix of the present Vohmie I. 

Mastodon raki Frick, 1933 
Figure 122 
From Hut .Sprinj;.?, New Me.xico. Pleistoeene. 

Mastodon ralxi Frick, 1933. "New Remains of Trilophodont- 
Tetrabelodont Ma^^todons," Bull. Amer. Mus. Nat. Hist., Vol. 
LIX, .\rt. IX, pp. 506, 510, 6.30. Type.— (Op. cit., p. 630): 

"Right ramus and symphysis and part of left ramus with nij-nii. 
F:A.M. 23335." From Hot Springs, New Mexico. Found in 1927. 
Pleistocene. Type Figure. — Op. cit., figs. 25A and 29A. 

Tvi'E Description. — (Frick, 1933, p. 630): "A right and 
broken left ramus with mi alveolus, mz-in^ and pai'tial symphysis, 
from the Pleistocene of New Mexico, is of the general proportions 
typical of Mustodon americanu.i-Mkc forms. On the character of 



the heel and the iili crown, which is narrower and taller than usual, 
the specimen is referred to a new species." 

Osborn, 1933; Osborn observes (Fig. 122) that the type third 
right inferior molar exhibits the following: (I) Serrated spurs from 
the ectoconelets, with reduced contiguous mesoconelet; (2) three 
to four conelets on worn inner lobes; (3) traces of cement in two 
anterior median valleys; (4) third inferior molar with well- 
developed pentalophid, crown relatively elongate and narrow (ap. 




lllastodon raki Frick 



FiK- 122. (Upper) Type third right inferior molar, r.Mj, distinguished 
by tlie unusual prominence of the trefoil crests extending down the slopes of 
the ectoconelets (see 5 below). After Frick, 1933, fig. 2oA. About one-half 
natural size. 

(Lower) "Right ramu.s and symphysis and part of left ramus with ms-ma 
lrev.|." After Frick, 1933, op. cit., fig. 29.4. .^bout one-sixth natural size. 



184 mm., tr. 80 mm., index 43), as compared with corresponding 
r.M:i of Mastodon americanus (Amer. Mus. 21920, ap. 181 nun., tr. 
93 mm., index 51) and with the large Shawaugunk cranium (.\mer. 
Mus. 2595, ap. 155 mm., tr. 84 mm., index 54). (5) This molar 



Table or Comparative Measurements of the Mastodon kaki Type and Three Eastern 

Specimens of Mastodon americanus 



mi 

m- 

nij 

Width at 2d crest. 
Mandible, length. . 
Diameter of t usk. . 



Mastodon raki Tyjje 
F:A.M. 2.3335 
New Mexico 



114 imn. 
184 

80 (crown tall) 
800e 
45 



A.M. 21920 
New York 



Mastodon americanus 
A.M. 17771 
Indiana 



88 mm. 
113 

181 

93 (crown low) 



114 

180 
96 



A.M. 2595 

Shawangunk' 

New York 



77.5-1- mm. 
103 
155 

84 



176 



OSBORN: THE PROBOSCIDEA 



with serrated eetoconelets and small mesoconelet appears to be 
distinct from the "losange" form prevailing in Mastodon americanus. 
Geologic Age. — Frick {op. cit., p. 630) observes: "The 
mandible was collected by Joseph Rak in the fall of 1927 from beds 
bearing teeth of Pleistocene Equus." 

Mastodon americanus alaskensis Frick, 1933 
Figure 123 
From the vicinity of Fairbanks, Alaska. Pleistocene. 
The specimen is especially interesting and important because 
of the rarity of the Mastodon in Alaska. Frick observes (p. 631): 




"Mastodon itself so far has been exampled b the Fairbanks area 
alone by a few detached molars. As associated remains of any of 
the extinct forms are almost unknown in the particular ai-ea, the 
present season's find of a mastodon mandible associated with m', 
superior tusk and a representation of the limb elements, is re- 
markable." 

Mastodon americanus alaskensis Frick, 1933. "New Remains 
of Trilophodont-Tetrabelodont Mastodons," Bull. Amer. Mas. 
Nat. Hist., Vol. LIX, Art. IX, pp. 506, 510, 631, 632. Type.— 

(Op. cit., p. 631) : "Partial mandible with left m-i, both mss, partial 

Comparative Measurements 



lUastodon amencanus alciskmsis Frick 



Fig. 123. Type of Mastodon americanus alaskensis Frick (F:A.M. 27009), 
about one-sixth natural size. 

Left ramus and symphysis, also left third .superior molar, l.M^. .\fter 
Frick, 1933, fig. 29A. 





M. a. alaskensis 


M. amencanus 


M. americanus 




A.C.-F:A.M. 


A.M. 14345 


A.M. 9951 




27009 






m2 


109-|-mm 


123 




ms. 


187 [ap.l 
[101 tr.) 


196 (left) 
(176, right) 








Himierus 


905 




910 


Radius 


680 




670 


Ulna 


675 
107 






Metacarpal 1 . . 


110 


2.. 


148 




150 


3.. 


175 




160 


4.. 


152 




150 


5.. 


123 




100 




Alnstodontindi type O O, referred •. Zygoiophodontindt type n, referred m. Ancient mastodonts type ®, referred +. 
American mastodon discoi^eries l°%h Famous Warren mastodon of JVewburyh ®. Osborn, 1934. 



Fig. 123a. Theoretic Miguation Lines of the Typical Mastodon of America 
"Yoke-crested" mastodonts (subfam. Zygolophodoiitinae, gen Zygolophodon, Turicins) indicated by squares: D types, ■ referred. Compare figure 137. 
True mastodonts (subfam. Mastodontinie, gen. Miomaslodon, Pliomastodon , Mastodon) indicated by circles: O types, • referred. Compare figure 86. 
Ancient mastodonts (order Mastodontoidea, gen. Pnln-omastodon) indicated by crosses: © types, + referred. 
Miomaslodon tapiroides americanus, Tasnid, Hungary. 
iMiomaslodon depereli sp. nov., Chevilly, France. 

Pliomastodon americanus praetypica, Maria-Theresio|jcl and Batta-Krd, Hungary. 
'Mastodon ohioticus,' P(«tchana, Podolia, Russia. 
Mastodon pavlowi sp. nov., Pestchana, Podolia, Russia. 

Zygolophodon. chiefly distributed in Italy, North Africa, France, and Hungary. Zygohphodon (?) (Matsumoto), Japan. 
.Mastodon tapiroides (Pallas, 1770-1777), Belaja, Ru.ssia [= Zygolophodon borsoni]. 
Mastodon abundant in f(;rmer forested regions of North America. See figure 1231). 
Miomaslodon and Pliomastodon rare in western United States. See figure 123b. 



THE MASTODONTIN^: MASTODOM 



177 



alveolus of mi and incisive alveoli; left m', superior tusk, humerus, 
both radii, ulnie. manus, and distal end of femur. Alaska College 
— F:A.M. 27009. Figured this paper, Fig. 29A." Hoiuzon 

AND Locality. — Vicinity of Fairbanks, Alaska. Type 

Figure. — Op. cil., fig. 29A. 

Type Dk.schiptio.n.— (Frick, 1933, pp. 631, 632): "The 
mandil)l(' exhibits a heavier vertical ramus and symphysis and 
narrower symi)hysial trough, the alveoli suggest much larger 
incisors, and the molar tooth-crowns are lower than in the American 
Mastodon, as seen in A. M. 14345, from Illinois. The limbs ap- 
proximate those of the Warren mount [Amer. Mus. 9951]. The 



tusk measures 7 ft. 1 in. on curve, and 19.5 in. circumference at 
base. 

Osborn, 1933: The ty(x> mandible of MaModon americanus 
alaskensis reveals (Fig. 123) relatively broad third inferior 
grinders, namely, l.M.i, ap. 187 mm., tr. 101 mm., index 54; (2) 
the mandible is correspondingly ma.s.sive, as well as the rostrum. 
Another feature, also observed in Mastodon raki Frick, is the 
.serration of the ectoconelets and relative reduction of the con- 
tiguous mesoconelet. Frick observes (op. cit., p. 631) that "The 
individual, while of as large size as the Warren Mammoth [Masto- 
don], was immature, the epiphyses being unconsolidated." 




Mastodon americanus in North America. affer O.P. Hay. 1925-1950 



I'iK- 1231). Geographic Distkihitiox of the TYPir.\i. Mastodont.< ok the United States and Canada 
(1) Typo locality of .Maslodon amrrica7>us Korr-Cuvirr. (7)-(24), ,•<('(■ full list of .'spcoies on 1)11^0 137. 

O Goographip distribution of subsppcios and mutations of Mastmlon americanus in tlip I'nitpd States and Canada, after detailed reports and re- 
searches of O. P. Hay on the mastodonts of the Ea.<itprn, Middle, and Wostorn I'nitrd States, and Eastprn Canada. 




01 J? 






2 

« CO 

a a 

a g 



— bo 

^ a **-■ 

S '^ 9 

a - § jj 



v 



S X 



o 5i =s 3 
., o I 






u 



a S 

< o 

^ S 

o 



t^ t- G -*^ 

rt « X g 

^- s ° 

— T3 r3 "«- 
3 OJ 2 
-5 t: -S "O 

g - il 

O ^ g Q 



•3 -c -fi S 



2 a 

T a 
a a 



t, a 

I 2 






.2 1) - 

^ 2 






•^ S g <« t. 

i lili 

=" <^ *^ -£ 
o t -2 



a 



a 
a 

o »■ 

CO 






o 
» 

u 

w 



O 03 
"^ CO Ml , - 

^ .- (- I— t 

= -^ -a ;§» 

3 ^ 'g c 

^ >;-S <N 

O to -iJ OS 



C _ 00 

<; ^ — 



+- J2 '-• ^ G 

^ ~ '^ "^ _r 



c tf 3 
3 c a 
o o 1^ 



'a 

si) QJ 






S Q -^ 



?? o -S 



11. SKELETAL CHARACTERS OF THE WARREN AND WHITFIELD MASTODONS 
The 'Warren Mastodon ' (Mastodon americanus) Skeleton 

See figures 82, 83, 124, 126, and 130 
Hudson River near Newburgh, New York, Pleistocene, of Postglacial (IV), or post- Wisconsin age 

Every bone of tliis superb skeleton has been separately figured and described with great accuracy in the two 
editions of Warren's great Memoir (1852, 1855). The reader is referred to Warren's Memoir for all details, except 
those relating to the tusks which are fully described below in the present Memoir. 

A full history of the discovery in the summer of 1845 of this famous skeleton and of the various phases in its 
mounting and exhibition is given below (pp. 181-183). In 1906 the skeleton was presented to the American 
Museum by Mr. J. Pierpont Morgan, to whom this Memoir is dedicated. The entire year (1906-1907) was 
devoted to the work of cleansing, renewing, and remounting the skeleton; all the bones were given an alcohol 
immersion to remove the dark varnish with which they had been covered, thus the beautiful original hght brown 
color at the time of discovery was completely regained. The tusks had been completely shattered and were 
restored with great difficulty; twenty-three inches of each tusk are inserted in the sockets, the projecting part 
measures 6 feet 8 inches, as shown in figures 124 and 126. 



Me.\surements of the Warren Mastodon (Mastodon americanus) compared with those 

(LOXODONTA AFRICANA OXYOTIS) 

Mastodon americanus 
Total length of skeleton, base of incisive tusks to 

perpendicular drop of tail 4 . 546 14 ft. 11 in. 4 

Height to summit of dorsal spines opposite summit 

ofscapulif 2.78 9 IK 3 

Summit of scapulae from the ground 2 . 635 8 8 3 

Summit of highest part of pelvis or ossa innominata 

from the ground 2.668 8 9 - 2 

Incisive tusks: total length of right tusk on outside 

curve 2.615 8 7 

Incisive tusks: length of right tusk as exposed. .. . 2.033 6 8 

Thigh bones: length of right tibia 71 2 4 

length of right femur 1 . 06 3 A% 1 

length of right pes extended 53 1 9 (on ground) 

pelvis, width of ossa innominata. ... 1 . 826 6 1 

Shoulder girdle and forelimb : height of scapula. . .916 3 ){ 

right humerus, 

length of 950 3 IK 1 

right ulna-radius, 

length of 658 2 2 (ulna without ole- 1 

cranon) 
right extended 
manus, length of 

Seven cervical vertebrte, length of 51 1 8 

Dorsolumbar vertebrae, total length of 2 . 250 7 4>j 2 

Five sacral vertebrje, total length of 48 1 6K 

Caudal vertebrae, estimated total number (28), 

total length of 2.085 6 10 1 



OF the African Elephant "Jumbo"' 

Loxodonta africana oxyotis 
.090 13ft. Sin. 

,150 10 4 
045 10 

785 9 2 



788 2 

258 4 
217 

340 4 

900 2 

091 3 

037 3 



285 
550 
035 
268 



7 
IK 

8K 



4K 



WYi (without carpus) 
10 
8 

lOK (length 4 sacral 
vertebrae) 



455 4 9 (est. 21 caudal ver- 
tebrae) 



'Amer. Mus. Dept. Mam. 3283. Prepared from specimen presented by P. T. Baiuum in the year 1889. 

179 








.n 
■a 

T3 



6 



M 
C 

'Eb 

c 
o 



o 

a 



c 
o 



tac - 



S 

rn ^ CO , 

^ if -s 
S :5 ■- "2 



a 






"o -Si 

or. qj 

e > 

3 a, 
o ■« 



Si •£ 
JS Si 



O 03 



-6 a^ 



— !U 



a 



Ml as 

■3 t 



oS 



K 5 



S' .so^ 



a a.% 



_ .- dj •— 
c fe 5 ^ -a 



O si 



:« 






S 03 



OJ ' 






; ^ -S 



M 



5 --^ 



o 



CO 



£ S 



»2 — ( 



« ;g -S 
- - -^ J 

.2-2 5; i^~i; 

« g :.s -c ,s - 
c ^ ^ i. .fcj »*- 



a 












3 £ 

(U o 
to ' 



to 

a 



oj -a is 









to 

'.S 






Q 
O 

<U 4 



a 
o 

T3 

o 

*^ 

CS 



■-- S -a _g 
& OJ c -S 

- a o .2 o 9 

I -^ .5 ft- " g 

•-' '^ ^ 
— c- 

a d 

- ^ .=£ "^ '-' > 
•St 95 6-n 

o to 



=3 



c 

3 



;:l^ 



t- t, 03 « 5:- ^ 



OQ C 



3 (b (_; 



E 03 

■« " = 



2 I '^ - 
c3 — ^ ._ 



,rf 9 



bb 



CO G « 



S « pd 3 K 



.2 a 



T3 
S 



^ to I 

C c3 O 



THE MASTODONTIN^: MASTODON 181 

The contrast between the comparatively broad, long-backed, low-bodied proportions of Mastodon ameri- 
canus and the elevated, short-backed, narrow-bodied proportions of the African elephant is similar to that which 
generally obtains as between the Mastodontoidea and the higher Elephantoidea. 

MASTODONTOIDEA: Relatively broad, low-bodied, heavy-limbed, long-bucked proboscideans. 
ELEPHANTOIDEA: Relatively narrow, high-bodied, long-limbed, short-backed proboscideans. 

The 'Whitfield Mastodon' (Mastodon americanus) Skeleton 

Hudson River, near Newburgh, New York, Pleistocene, probably of PostRlacial (IV) or post-Wisconsin age 

The Whitfield Mastodon (Fig. 125), named in honor of the late Curator Robert Parr Whitfield of the Ameri- 
can Museum, is the assembled skeleton of a male mastodon in which the tusks are of a large size, but the jjropor- 
tions and mounting are much less accurate. This skeleton, mounted for the American Museum in 1879, was 
purchased by the Senckenberg Museum of Frankfort, Germany, where it is now on exhibition. 

History. — (1) This skeleton was discovered in 1879 near the town of Little Britain, about nine miles south 
of Newburgh, N. Y. It was imbedded in peaty material on the edge of what had been less than fifty years before 
an open pond ; the pond had been drained and was under cultivation at the time of discovery. In course of further 
drainage work, following the wet season of 1879, at a depth of fourteen inches, an apparent log was reached, 
which proved to be the leg bone of an animal. This induced the farmer to search for other bones, and in a few 
days more than one-third of the skeleton had been exhumed, including the head and lower jaws. (2) Three 
weeks later Professor R. P. Whitfield, first Curator of Geology in The American Museum of Natural History, 
visited the excavation at the request of Major T. B. Brooks of Newburgh. He found that the swamp on the side 
nearest the skeleton was bordered by a low hill of hard blue "bouldei- clay," mixed with shale and gravel. This 
hard layer sloped down and passed under the peat or "muck" of the swamp to form the original bottom of the 
pond. From this peat the skeleton derived its dark color. (3) The skeleton lay with the head farthest from the 
original shore of the pond and deeply imbedded. The right Hmbs were near the surface, the right humerus being 
buried only fourteen inches, which proves apparently that the animal had been mired and had fallen on its left 
side. At the point where the head was found, twenty feet from the place at which the "boulder clay" rose to the 
surface, the peaty material had accumulated over the head to a depth of ten feet. The humerus and other parts 
nearer the surface were consequently less perfectly preserved. 



12. AGE AND PROGRESSIVE CRANIAL AND DENTAL CHARACTERS OF MASTODON 
AMERICANUS, REVEALED IN THE AMERICAN MUSEUM COLLECTIONS 

Superior and inferior grinders, incisive tusks, age, sex, growth, cranial and mandibular development 

Age and Sex Characters of the Incisive Tusks 

In the assemblage and preparation of the tusks of the Warren Mastodon skeleton (Amer. Mus. 9951), Osborn 
(1910.347 and 1923.574) made the following observations which should be permanently recorded in this Memoir: 

A Means of Estimating the Age of the Mastodon. — (Osborn, 1910.347, pp. 1-3): In repairing the tusks, the outer 
sheathing of the dentine was found in large part absent. The inner sheathing e.xposed a series of concentric constrictions and 
px|)aiisi()ns which were observed to be approximately symmetrical on the twosides, as indicated by theseries of + signs in Fig. 3, 
.\ and H [Fig. 126, right, of present Memoir]. Secondly, it was noted tiiat the intervals between these constrictions are broader 
in the middle and fore part of the tusks (corresponding with the youthful stage of growth of the tusk) and become narrower 



182 



OSBORN: THE PROBOSCIDEA 



toward the base of the tusk (corresponding with the mature or adult stages of growth). Eighteen of these rings are preserved 
on one side and thirteen on the other. They are faintly indicated also in the waving surface of the dentine of the tusk (Fig. 3). 
On the hypothesis that these are actual annular increments of growth, the right tusk (Fig. 3, A) consisted of about twenty-eight 
segments, which allowing for the milk dentition and for the part worn off at the tip would assign to the Warren Mastodon an age 
of perhaps thirty years. Similar annular constrictions are observed in the tusks of the mammoth from Alaska; and are also 
indicated in the tusks of the African elephant. Since the age of the Indian elephant and the rate of tusk growth is definitely 
known the identification of similar concentric annular growths would be the means of testing the value of this hypothesis. 




M. amencanus 
"Warren Mastodon" 
Amer Mus. 9951 Hof. 




Fig. 126. Superior, inferior, and lateral view.s of Warren Mastodon (Amer. 
Mus. 9951) tusks, an adult male from Newburgh-on-Hud.son, New York. Both 
figures after Osborn (1910.347) and (1923.574). 

Left. Superior aspect. Observe the deep insertion of the right superior tusk 
and the annular growth rings feebly indicated. One twenty-fourth natural size. 

Right. Enlarged inferior aspect of the tusks showing the annular growth 
rings. One-twelfth natural size. 

A, Right tusk drawn from the outer inferior side showing growth rings 1 to 
19, the center of each ring being indicated by a plus sign (-I-). Of great interest 
is the contact point (c) just above the 19th growth ring, which positively 
enabled us to determine the actual length (8 ft. 7 in.) of the right tusk on the 
outside curve. 

B, I>eft tusk, outer view, sho\ving 13 growth rings. 

The Warren Mastodon thus represents a male animal at least thirty years of age at the time it sank into 
a bed of marl near Newburgh, New York. The story of the restoration of these tusks is as follows (cf. Osborn, 
1923.574, p. 18): 

Method of Restoring the Warren Mastodon Tusks.— There still remained the problem of the tusks, which are in- 
variably the most vital part of buried skeletons of the great proboscideans of the past. It appears that the original tusks could 
nf)t be piesorved entire by the methods known at the time of exhumation. The discoverers were unable to prevent the tusks 
from splitting, warping, and falling to pieces, especially at the butt. In order to preserve what could be saved intact, the 
butts of the tusks, already split and warped, were sawed off under Doctor Warren's direction, and only the tips, about three 
feet in length, were treated and preserved. The butts, fallen into fragments, but still lying undisturbed in two of the original 
boxes used for transporting the skeleton, were found in the Warren Museum when the skeleton was repacked to be sent to 
the American Museum. The tips, treated with preservatives, were still intact in another box; but neither had been used 
apparently for measurements in making the papier-mache restorations fitted to the skull in the Warren Museum. This documen- 
tary evidence certainly was not used by Doctor Warren, because in his three restorations he unfortunately accepted the 



THE MASTODONTIN^: MASTODON 



183 



erroneous original reports that the tusks as found were more than eleven feet in length; they were so described and illus- 
trated by him in the entirely impossible position shown in the photograph on p. 15 [Fig. 83 of the present Memoir]. 

When the Warren collection reached the American Museum, it was very carefully looked over in a search for remnants of 
the original tusks, and finally the fragmentary fossil ivory tips and butts were found, but inasmuch as most of the original 
records had been lost and no use of these materials had been made by Doctor Warren, it remained to be proved that the 
fragmentary butts of the tusks really belonged with the skull. The piecing together of these butts required several months of 
most ingenious and patient work on the part of one of our preparators, Mr. Charles Christman. The ends of each tusk were 
perfectly preserved, but there was no connection between these tips and the reconstruct od butts of either tusk. Fortunately, 
when the butts of the tusks were sawed off, a single splinter of bone broke off, and finally this splinter was found to fit e.xactly 
to a fragment of the butt. There was great rejoicing in the laboratory when the relationship of these two fragments was 
discovered, because it enabled us to determine positively the length of the tusks as 8 feet, 7 inches. 

The rebuilding of the tusks, which required several months of most patient work, had two very important results: in the 
first place, it enabled us to place them properly in the sockets of the skull and to prove for the first time the exact relations of the 
mastodon ivories; secondly, a very painstaking examination of these tusks led to an important and most interesting dis- 
covery, namely, that it was possil)le to determine very closely the age of the Warren Mastodon. The ivory exhibits a series of 
growth rings which, counted from tip to base, seems to prove that the Warren Mastodon was perhaps thirty years of age 
at the time it sank into the bed of marl near Newburgh. The right tusk included at least twenty-eight of these segments. 
The growth rings are shortest near the tip of the tusk when the animal is young, and increase in length from the tip toward 
the middle of the tusk, but not in a regular ratio. These growth rings do not correspond exactly in the opposite tusk, but 
in both tusks they are longest in the middle region. Nine smaller rings are in the lower part. The writer's theoiy regarding 
these growth rings is that during the summer season, when all the conditions of life were favorable, and perhaps during the 
rutting period, when tusk growth was hastened by internal secretions from the reproductive glands, the growth of ivory was 
very rapid, the maximum growth in the 17-18 ring being 108 centimeters, or 4}^ inches, perhaps the maximum growth of a 
favorable season at the most vigorous reproductive period of life. The Warren Mastodon is an adult but not an aged specimen ; 
the skeleton is apparently that of a younger animal than the one represented by the Shawangunk head. Some estimate the 
maximum age of the American mastodon at between thirty and forty years,— less than half the life span of the elephant, 
which attains more than one hundred years. 

Male and Female Tusks.— In all the known Proboscidea there is a marked disparity between the male 
and female incisive tusks both in length and in diameter. The adtdt female 
tusks never fully attain the length of the adult male tusks, but a still more 
striking difference is their slenderness of proportion and diameter. 

The female tusks are finely displayed in the female skull of Mastodon 
americanus (Amer. Mus. 14293) found near Fulton, Indiana, and represented 
one-tenth natural size in figure 128. We observe that the divergence of the 
tu.sks in their sockets produces the strikingly V-shaped arrangement of the 
grinding teeth and the tusks, so that a straight line drawn from the apex of 
the tusks traverses the grinders and passes through the occipital condyle 
of the opposite side; this explains the marked divergence of the grinding 
teeth anteriorly, and the convergence of the grinding teeth posteriorly. 
Undoubtedly the incisive tusks were extremely important in females for the 
defense of the young, and this strongly divergent tusk mechanism brought 
about a strongly divergent grinding tooth mechanism. This divergence, 
however, is much less marked in the male skull (Yale Mus. 12600), beauti- 
fully represented in figure 129, than in the Warren Mastodon skull (Fig. 130). 




Male and Female Crania. — The palatal aspect (Fig. 129) of the 
superbly preserved adult skull of the Mastodon americanus from Oti.sville, 
New York, in the Peabody Museum of Yale University (Yale Mus. 12600), 
is shown one-eighth natural size. The drawing is one of the masterpieces of 



Fig. 127. Tusks of Elephas indicia (/) auma- 
tranus showing 17-18 annular growth rings. 
After photograph kindly furnished by Mr. H. 
.Munniks de Jongh of the Hague, who purchased 
tliem some years ago without a record. It is 
probable that they are from the Dutch East 
Indies. One-tenth natural size. (Cf. Fig. 126.) 



184 



OSBORN: THE PROBOSCIDEA 



Mrs. L. M. Sterling, the artist of most of the original pen illustrations in this Memoir. All the constituent bones 
and the foramina are clearly indicated by abbreviations, as lettered under the direction of Dr. WiUiam K. 

Gregory. Viewed both from above (Fig. 130) and 
below (Fig. 129) the facial and cranial portions of 
the M. americanus cranium are subequal in length. 
There is a very marked development of the pre- 
maxillaries (P. mx.) and of the maxillaries {Mx.), an 
abbreviation of the palatines, a compression and an 
abbreviation of all the basicranial elements, which, 
however, are much less extreme than in the basi- 
cranial aspect of the Indian elephant {Elephas in- 
dicus), see Volume II, Chapter XX. 

The male cranium (v^xner. Mus. 14535) repre- 
sented in figure 132 differs strikingly from the 
Peabody Museum cranium (Yale Mus. 12600, Fig. 
129) in the fact that the facial portion, from the 
premaxillaries to the back of the grinders inclusive, 
is decidedly longer than the cranial portion; it 
differs also in the marked divergence of the grind- 
ing teeth of the opposite sides, whereas in the 
Peabody Museum skull (Yale Mus. 12600) the 
grinding teeth are more nearly parallel. 

The female cranium (Amer. Mus. 14292) repre- 
sented in figure 132 shows a very marked sexual 
disparity in size; this is probably the normal dis- 
parity between the male and female crania of 
Mastodon americanus, because this female cranium 
does not differ markedly from the female cranium 
(Amer. Mus. 14293) shown in figure 128. The more 
feeble and slender female tusks (Fig. 128) demand 
less prominent premaxillo-maxillary 'development, 
consequently the sockets for the incisive tusks are 
relatively short and small, and elevated above the 
grinding tooth border, as seen in both the female 
skull (Amer. Mus. 14293, Fig. 128) and in the fe- 
male skull (Amer. Mus. 14292, Fig. 132). Thus 
there is a marked disparity in size between the 
females and males in (1) the grinding teeth, (2) the 

Fig. 128. Female skull. Palatal v.ew of the finely preserved female skull i^^i^^^^ ^^S^^' ^^^ ^^e SOCketS of the incisive tUSks, 

(Amer. Mus. 14293) of Mastodon americanus from near Fulton, Indiana. One- and (4) the size of the cranium aS a whole. 
tenth natural size. 

Observe the strikingly V-shaped arrangement of the grinding teeth and tusks, JuVENILE MaLE CrANIA.— The fuUy adult age 

tne apex of the V hemg at the posterior nares. The tusks are relatively long, very 

slender, gently convex or outeurved, and slightly incurved at the extremities. characters pointed OUt in the above descriptions 




THE MASTODONTINiE: MASTODON 



185 



and figures of the male and female skulls should be compared with the juvenile and infantile characters of the 
immature male crania shown in figure 131. The immature male skull' (Amer. Mas. 17727, Fig. 131 Al), from 
Rochester, Indiana, represents a stage in which the three 'intermediate molars' only are in use, namely, Dp 4, 
M 1, M 2. All these teeth show the three transverse crests characteristic of this genus and species. The tetra- 
lophodont third superior and inferior molars, M'^-Mj, have been extracted from the skull and jaws and are repre- 



M. americanus 
Yale Mus. 12600 R«<. 



/Ji»~, 




,_ M. amencanus 
' Warren Mastodon' 
Amer. Mus. 9951 Ref. 




Fig. 129. Mastodon americanus palate of finely preserved 
adult skull (Yale Mus. 12600), from Otisville, New York. One- 
eighth natural size. Complete skeleton collected May 7, 1875. 



Fig. 130. Superior view of skull (Amer. Mus. UOol) of the 
"Warren Mastodon, " an adult male with the tusks removed. 
One-eighth natural size. 



sented in figure 135. In this young male cranium' (Amer. Mus. 17727) we observe the following points: (1) The 
facial and cranial portions are subequal in length; (2) the sides of the cranium are nearly parallel; (3) the juvenile 
tusks are imbedded in sUghtly expanded premaxillary sockets; (4) all the crests of the grinding teeth (Fig. 131 
Al, also Fig. 135) are extremely sharp, representing a high degree of subhypsodonty. 

The calf skull of Mastodon americanus (Fig. 131 B) found near Hackettstown, Warren County, New Jersey, 
about twenty miles from Newark, represents a much more immature stage (Amer. Mus. 10459). 

Immaturk and Mature Grinding Teeth. — The immature male skull' from Rochester, Indiana (Amer. Mus. 
17727, Fig. 131 Al) yields a .series of immature dental crests (represented in I'ig. 135) of the third left superior 
molar, l.M^ and of the third left inferior molar, I.M3. These ridge-crests are a revelation of the very sharp, 



'This specimen is now made the type of Mastodon acutidena (see Appendix at the close of the present Volume, also Figs. 131, A1-A4, 135, and PI. i, L, 
of this chapter). 





w' 
.0^: 



■ S^^Pfl!^? /.*; 



Al 



Fki. 131. (Al-Al) Type ok Mastodon acutidens sp. NOV.; compare Figure 135. (B) Immature male skull of M. americanus Ref. 
One-eiohth natural size. American Museum Collection, Nos. 17727,' 10459. 
Al 1 eft lateral V> superior A3, palatal, A4, occipital views of young Mastodon skull (Amer. Mus. 17727),' from Rochester Indiana, showing Dp , M , 
and m'' ar terior Srttn onlv in use M^-M,, extracted from the skull and jaws, shown in figure 135. Type of Ma.ioion -cuHdens sp. mv 

B. cTZilTlLrdnu. skull (Amer. Mus. 10459), superior view. Found near Hackettstown, Warren County, New Jer.sey, about twenty m.les from 
Newark. 



'See footnote on page 185. 



18fi 








(I-eft). Male 



Fig. 132. Male a.nd Female Crama of Mastodon .^mericands 

.45.35), from ..sh.e;Tndi™ a"' Ortt;!;^:!^'' '™'" "'"'^ "'" '"" ''"' ^"^' "■" '''""^ "''" ''''^ '^ " '"'"'' ^^"'* "'«'•"» ^A-- '^'-• 

rmnh.^^A'' ''L"'"'^r?^Ii!"'" °! '^^''^'<"'<"' ""•«■'•«'""«, of somewhat greater age than the male cranium, exhibiting similar aspects of the skull This 
cran.um (Amer. Mus. 14292) was discovered two and a half miles southeast of Fulton. Indiana. One-twelfth natural size. 



187 




•^ 


03 


CIS 


P 


e 


3 


o 




t3 


S 


o 


;j 


c3 




'C 


c 


t-> 


■D 


^ 




o 


P 


s 


j= 


'C 


03 




=3 
O 




1 

3 




O 










o 


:§ 




'm 


"3 


**5 


CJ 


J- 


to 


O 


a 




a> 


1 


C3 

'-3 


> 




(d 


c 




I-< 




o 


1 

O 


o 


3 


o 




a. 


"•J 



":s 



f-P 



as 




O 


t' 


^i 


§ 


^;- 


£ 


o . 


S 


^ 


'3[) 


a ST 


^ 


'o 

a 


c 


< OS 




c 


^ OJ 


u. 




hH 


^;2 

bjD 


S 


t-i 

o 




ci -r 




OJ 




:§§ 


■c 
a 

03 


!§ 


'3 


TJ CO 






a 









fe ^ 



■a ' 
a 




5 s 


inner 

s (see 
ir and 


•2.S 


01 - 
_0 CD 


.. 3 .2 


-a oj 


o o 

s § °- 






Cf, 


|1S 
1 a-c 


C3 g 




^ s ° 


OJ 


a =1^ o=" 


^ E 




^-^ ■< 


C3 « 


a o a 


s s 




t.-. c 


G "1^ ?r 


O o 


• — O bC 


^ -(* 




s fe 


1 1 




li 


2 -S "S fe 


u 1 


^ -O S T3 


^! 


3 "£ ^ _j 


-S"s 


j= 'c ^ 5 


C3 — 


t. ^ (^ t. 


2 ■> 


3 u 3 <U 


e ^ d O 


•||°S'^^ 


^ -a 

CD 


m .. ° 


CO f 


J " "§ .-F 


-H i, 




.Ej| 


^a 


i"P 



<1 ci 



THE MASTODONTIN^: MASTODON 



189 



almost knife-like summits of the molar ridge-crests, which give a truly subliypsodont character to these crowns. 
These sharply crested summits illustrate the very marked progression of Mastodun acutidens sp. nov. upon the low, 
compressed crowns in the two species Miomastodon merriami and Pliomastodon matthewi, and upon the rounded 
lobes, i.e., cones, in Palseomastodon. 

Hypsodont Grinders. — It is probable that when all the grinding 
teeth now referred to Mastodon americanus are carefully measured and 
compared, considerable progressive evolution toward hypsodonty will be 
observed through the long period of Pleistocene time. 

Much more sharply elevated and acute transverse crests than those of 
the female specimens found near Fulton, Indiana, and represented in 
figure 134, are the grinders of the young male (Amer. Mus. 17727, Fig. 135), 
the type of Mastodon acutidens sjd. nov. 

Typical Grinders. — The typical female grinders (Amer. Mus. 14293 
and 14294, Fig. 134) exhibit relatively obtuse crests both in the upper and 
lower jaws. It may be observed also: (1) That the individual lobes, i.e., 
cones, are not distinctly bilobed; (2) that the worn enamel presents 
simple, compressed loops in M--M2; (3) that the trefoil spurs rise from 
the internal lobes of M", from the external lobes of M2; (4) that a sharp 
median sulcus divides the internal lobes from the external in M^-Mg. It 
is not surprising that these rudimentary trefoil spurs should be compared 
with the homologous trefoils of Trilophodon and of Serridentinus; in 
position they resemble Serridentinus rather than Trilophodon. (5) The 
third inferior molar, M3, is distinctly longer and relatively narrower than 
the third superior molar, M\ These beautifully drawn grinding teeth 
very clearly illustrate the characters of the typical Mastodon americanus 
and agree closely with Cuvier's type grinders (Figs. 76, 112) from the Big- 
Bone Lick, near the Ohio River, Kentucky. 

Plicated Grinders. — A somewhat more folded or plicated condition 
of the superior grinders is shown in the palate and lower jaw (Amer. 
Mus. 12464, Fig. 133) of a middle-aged male (?) individual of Mastodon 
americanus found near Buffalo, Kansas. These plicated or folded grinders 
approach those of the type of Mastodon americanus plicatus Osborn (Fig. 
120), although the plication is somewhat less extreme. We observe in M'' 
(Fig. 133 A) the marked development of the cingulum, internal, anterior, 
and external, and the marked irregularity of the enamel surfaces of the 
lobes; the tetartoloph in M'^ is subdivided into four conelets; in M3 there 
is a strong anterior cingulum, but it exhibits no external or internal cingula. 

We conclude that the male and female grinding teeth of specimens 
from different parts of the United States, mostly of Upper Pleistocene and 
Postglacial time, exhibit marked |)rogressive ascending mutations and subspecific characters which may be found 
to distinguish more or less constantly the descendants of certain subspecific, specific, and phyletic branches. 




Akk£ 

Type of Ma.stouon- acitidens sp. nov. 

Fig. 13.5. Sharp summits of piirt!y developed 
molar crowns excavated from a young male skull 
(.\mer. Mus. 17727), from Rochester, Indiana. 
Compare figure 131, .\l-.^4. 

Al, Five crested left sui^erior molar, l.M', external 
view. 

.\2, Cromi view of the same five crests of l.M'. 
Observe the division of the transverse crests into 
inner and outer lobes, completely separate in the 
young condition. 

Bl, Crests of undeveloped left lower molar, I.M3, 
somewhat more advanced than the u[)i>er. Observe 
the bilol)ed condition of the fourth crest; fifth crest 
not developed. 

li'2, External view of same crest.s of I.M3. Observe 
the high anteroposterior compression of these crests 
adapted to the sharp cutting of the food. 



190 



OSBORN: THE PROBOSCIDEA 



Summary of Sexual and Progressive Characters Observed in the American Museum Collections 

In the Amorican Museum materials referred to Mastodon americanus, it is clear from the above brief descrip- 
tions and from the close examination and comparison of the carefully prepared figures that there is a verij marked 
progressive evolution in the grinding teeth, and probably in the tusks and portions of the skeleton as well, indicating 
varied ascending mutations and subspedfic if not specific stages. These characters are not mere individual varia- 
tions; they certainly mark progressive evolution extending over a very long period of time, namely, over the 
entire 1,000,000-year period, the minimum figure now assigned to the American Quaternary or Pleistocene time. 

(1) The primitive species Mastodon progenius Hay is believed to be of Aftonian or 1st Interglacial age. 

(2) The Mastodon americanus of Big-Bone Lick, Kentucky, is of much more recent geologic age, probably TV 
Glacial time, indicated by Ovibos and Rangifer. These are probably the typical deposits of the species Elephas 

[Mastodon] americanus Kerr, which came from the Big-Bone Lick. 

(3) The 'mastodon' deposits of Newburgh-on-Hudson, New York, including the 'Warren Mastodon,' are 
probably of Postglacial (IV) age, or the period of the retreat of the great Wisconsin glacier. 

If these geologic correlations are correct, the evolution of Mastodon americanus covered a period of at least 
1,000,000 years, from the more primitive stages, such as M. progenius, to the most progre.ssive stages, such as M. 
acutidens sp. nov. These stages, however, can only be clarified by very close and prolonged monographic research . 







t. 




■'•-^•f' 



Group of American Mastodons along the Ancient Missouri River of Kansas 
Left portion of the great late-Pleistocene mural in tlie Hall of the Age of Man in the American Museum. The faunal period of Bison lalifrons 

{Bison regius] and Equus complicatus {Equus scotti) 
Kig. 136. This mural, painted under the direction of Henry Fairfield Osborn, was completed by Charles R. Knight in the year 1920 and is one of the 
finest examples of his art. In the group there are one bull, two cows, and a ealf. 



Chapter VII 

HISTORY OF THE SUBFAMILY ZYGOLOPHODONTIN/E, THE YOKE- 
TOOTHED MASTODONTS 

Gradual separation of the 'mastodon' tapiroides and 'mastodon' turicensis from the 'mastodon' 

BORSONI and other PHYLA OF THE MASTODONTS. SuBFAMILY, GENERIC, AND SPECIFIC DISTINCTIONS OF THE 
ZYGOLOPHODONT GENERA, NAMELY, ZYGOLOPHODON AND TURICIUS. HaBITAT, MIGRATIONS, PHYLOGENETIC 
SUCCESSION, AND EXTINCTION OF THE ZYGOLOPHODONTIN^. COMPARISON WITH STEGOLOPHODON. 

1. History of the subfamily Zygolophodontinffi, 1770 1927- 6. (Characters of Zijgolophodon and its included species. 

1935. 7. Characters of Turicius and its included species and muta- 

2. Collateral relationship to the subfamily Stegodontinte. tjo^s, as compared with Zygolophodon and Serri- 

3. Zygolophodon and Turicius independent phyla (Osborn, denlinus 

1926) of the Zygolophodontinffi Osborn, 1923. ^ ,.'",..,. , . . , . 

4. Resume of the generic and specific distinctions. 8. Geographic distribution and extinction of species of 

5. Progressive characters of the grinding teeth. Turicius. 

(January, 1935) Since 1928, when this chapter was written, the fundamental patterns of the grinding teeth of Paleeomastodon, 
Mastodon, Zygolophodon, Turicius, and Stegolophodon have been intensively examined, compared, and figured, as shown in 
Pis. I-IV, between pp. 134 and 135. These new comparative figures positively demonstrate: ( 1 ) The complete phyletic separa- 
tion of these five genera, and (2) the possible derivation of the grinding teeth of the Stegodontoidea from those of Stegolophodon.' 
The species of Stegolophodon and of Slegodon are described in detail in Chapter XV of Volume II of this Memoir. 

We may open this chapter by calhng attention to the marked sexual and progressive characters which may be 
observed in the mastodonts in the American Museum collections. These serve clearly to distinguish the Mas- 
todontinse or true mastodonts from their most nearly related phylum the Zygolophodontins, because it is clear 
throughout that (1) the true Mastodon molar is profoundly different in character from the true Zygolopfu>don 
molar, and (2) that the final stage of evolution in the Mastodon acutidens molar (Fig. 135) converges towards 
the final evolution stage in the zygolophodont molar (Fig. 168, Turicius virgaiidens). 

This chapter covers the genera Zygolophodon Vacek and Turicius Osborn. The species included within 
Turicius show several resemblances to those included within the genus Serridentinus (Chap. X), as observed by 
the present author; in many other characters Turicius is clearly separable from Serridentinus and more closely 
related to Zygolophodon. The median longitudinal sulcus (Fig. 89) of Mastodon americanus disappears in Zygolo- 
phodon and transversely perfected crests evolve. Zygolophodon has four to six main conelets, Turicius has five 
to eight small conelets (T. turicensis) to eight to nine (T. virgatidens) . 

As pointed out clearly in the preceding and succeeding sections, the genus Zijgolophodon, based by Vacek 
(1877) on the three genotypic species Mastodon borsoni, M. turicensis, and M. tapiroides { = M. pyrenaicus), now 
proves to include two distinct generic phyla which may be traced chiefly in western Eurasia from the base of the 
Miocene to the summit of the Pliocene, namely: 

I. Zygolophodon borsoni, which may spring from Z. pyrenaicus, and from Z. pyrenaicus aurelianensis. Lower 
Miocene of France. This generic phylum parallels the true Mastodon americanus phylum in its evolution, but it is 
nevertheless quite distinct. Like the true mastodonts, it is little known until Middle and Upper Pliocene times, 
when the remains of Z. borsoni are very abundant. 

n. Turicius, typified by Turicius turicensis of the Upper Miocene, is a very distinct generic phylum, extend- 
ing from the Lower Miocene 'Mastodon' tapiroides of Cuvier to the Middle Pliocene. This phylum is abundantly 
represented especially in lignitic and fluviatile deposits. It is quite distinct (see Figs. 138 C, Cl; 139, 4, 5, of 
the present chapter) from the phylum Zygolophodon borsoni and presents certain parallels with the Serridentinus 
phylum, which is widely distributed in the Miocene and PUocene of America. 

'(1935) Stegolophodon rcmovpd from the Stcgodontinic and placed in the new subfamily Steqolophodontin* (see Appendix of the present Volume I). 

191 



1. HISTORY OF THE SUBFAMILY ZYGOLOPHODONTIN^, 1770-1927-1935 

Original reference: Zygolophodontinffi Osborn, 1923, "New Subfamily, Generic, and Specific Stages in the Evolution of the Probo- 
seidea," Amer. Mus. Novitates, No. 99, p. 1 (Osborn, 1923.601). 

BuFFON, 1770. — The history of discovery goes back to the year 1770, when Buffon received from the Comte 
de Vergennes an upper grinding tooth found in Russia ("la petite Tartaric ") and after examination and descrip- 
tion figured the tooth in 1778, as shown in our facsimile reproduction (Fig. 111). Marie Pavlow (1894) re- 
fers to Buffon's description of this tooth in the "Epoques de la Nature," 1775 [1778]. Almost at the same time 
(1777) Pallas described a tooth found near the Belaja, an affluent of the Kama River, Russia, a tooth which sub- 
sequently Eichwald and de Blainville referred to Mastodon tapiroides [now referable to Zygolophodon borsoni]. 
Pavlow's full description of these Russian and Siberian discoveries is cited above (Chapter YI, pp. 133-135). 

The chief succession in the discovery and description of the species of Zygolophodonts in Eurasia appears to 

be as foUows: 

CHRONOLOGIC ORDER OF DISCOVERY AND DESCRIPTION 

See Figure 137 

Original Name 



Specific Reference in 
Present Memoir 



1770-1778 Russia 



1777 



True zygolophodont molar, de- 
scribed and figured by Buffon 
(see Fig. Ill) 
Near the Belaja, Russia Zygolophodont molar described 

by PalLis 
Petit mastodonte Cuvier 
Mastotheriuni microdon Fischer 

de Waldheim 
Mastodon microdon Rafinesque 
Mastodonte tapiraide Cuvier 
Mastodon turicense Schinz 
Mastodon Borsoni Hays 
M. tapiroides-minus Lartet 

Mastodon Vialetii Aymard 
Mastodon vellavus Aymard 
Mastodon Buffonis Pomal 
Mastodon atticus G. & L. 

(in Wagner) 
M. (Triloph.) Pyrenaicus 

Lartet 
Mastodon affinis Jourdan (MS., 

labels in Lyons Museum) ; in 

Lortet and Chantre, 1879 
Mastodon inrgatidens von 

Meyer 
Mastodon Zaddachi Jentzsch 
Mastodon wahlheimensis Klahn 



Zygolophodon pyrenaicus aure- 

lianensis Osborn 
Turicius turicensis simorrensis 

Osborn 

192 



1. 


1806 


Montabusard, France 


2. 


1814 


Montabusard, France 


2a. 


1814 


Europe 


1. 


1821 


Montabusard, France 


3. 


1824 


Elgg, Switzerland 


4. 


1834 


Piedmont, Italy 


6. 


1835 


Simorre, France 


6. 


1847 


Vialette (environs du 
Puy), France 


7. 


1847 


Velay, France 


8. 


1848 


Auvergne, France 


9. 


1857 


Pikermi, Greece 


10. 


1857, 18.59 


Haute-Garonne, 
France 


11. 


1859, 1879 


Locality (?) 


12. 


1867 


Fulda, Germany 


13. 


1883 


Thorn, West Prussia 


14. 


1922 


Wahlheim and Essel- 
born, Germany 


15. 


1926 


Sables de I'Orl^anais, 
Chevilly(?),Casse- 
grain, Avaray, 
France 


16. 


1926 


Simorre, France 



= Zygolophodon borsoni 

= Zygolophodon borsoni 
= Turicius tapiroides 

= Turicius tapiroides 

= {1)Turicius tapiroides 

— Turicius tapiroides 

= Turicius turicensis 

= Zygolophodon borsoni 

= Turicius tapiroides-minus 

= Zygolophodon borsoni vialetii 
= Zygolophodon borsoni vellainis 
= Zygolophodon borsoni buffonis 

= Turicius atticus 

= Zygolophodon pyrenaicus 



Geologic Age 



Pleistocene (?) 

Pleistocene (?) 
Lower Miocene 

Lower Miocene 

Lower Miocene 
Upper Miocene 
Middle Pliocene 



Upper Pliocene 
Upper Pliocene 
Upper Pliocene 

Lower Pliocene 

Middle Miocene 



= Zygolophodon borsoni affinis Middle Miocene 



= Turicius inrgatidens 

= Zygolophodon borsoni zaddachi 

— Turicius wahlheimensis 

[see Chap. VIII, p. 282, for 
description and figure] 



= Zygolophodon pyrenaicus aure- 
lianensis 

= Turicius turicensis simorrensis 



Middle(?) Pliocene 
Upper Pliocene 



Upper Pliocene 

Lower Miocene 
Middle Miocene 



THE ZYGOLOPHODONTIN.f;: HISTORY 



193 



CuviER, 1806.— In Chapter V, "History of the Classification of the Mastodontoidea, Families and Sub- 
families, 1705-1927," we have recited Cuvier's original description (1806) of "Petit mastodonte" { = Turicius 
tapiroides), a tooth from Montabusard, compared with one from Saxony; this specific description was confirmed 
by de Blainville in 1839-1864. Schinz (1824, p. 278) recognized Cuvier's five cla-ssic species and added the 
zygoiophodont species 'Mastodon turicense' from the Miocene hgnites of Eigg, Canton Zurich, Switzerland. Sub- 
sequently tlie zygoiophodont relationship of 'Mastodon' tapiroides and of 'M.' turicensis was recognized. 

Hays, 1834, Aymard, 1847, Pomel, 1848.— The next step was that of Hays, clearly distinguishing the Middle 
to Upper Pliocene zygoiophodont species as Mastodon borsoni. Then followed the descriptions of Mastodon 
inaletii and of M. vellavus by Aymard, of Mastodon Buffonis by Pomel, and of Mastodon zaddachi by Jentzsch— all 
of which species we now believe to be referable to Zygolophodon borsoni. See Hays' type figure and description 
on page 207 below. 

Wagner, 1857.— In 1856 Gaudry and Lartet described Mastodon pentelicus. In 1857 Wagner cited (p. 140) 
the same animal as: "11. Mastodon atticus 0. & L.," thus apparently attributing the species pentelicus to Gaudry 
and Lartet, but under another name. In PI. vii, fig. 16, he referred to this species a maxillary fragment of another 
individual, on page 142 making reference to this specimen as follows: "In meinen friiheren Erorterungen 




I'^ig 137. Geographic distribution (according to the numbers in the list oppo-site) of species of Zygolophodon and of Turicius. The white dots within the 
hliick arciis represent the appmxiniale loralities where the types of these sixteen species w<Te discovered. The white crosses represent referred specimens. 
Tlie .lapanesc record is from tlie Tatsunokuchi, Lower Pliocene (Matsumoto, letter, 1924). 



(lieser Ueberreste habe ich es nicht gewagt, sie einer bestimmten Spezies von Mastodon zuzuweisen. Gaudry and 
Lartet haben sie jetzt als Mastodon atticus bezeichnet; ihre ausfUhrliche Abhandlung, die noch nicht erschienen 
ist, wird jedenfalls die Griinde angeben, warum sie dieselbe von M. angustidens getrennt haben." 

O.sborn (1927) in the present Memoir points out that Wagner's maxillary fragment (PI. vii, fig. 16) belongs 
to the genus Turidus and by courtesy adopts Wagner's specific name atticus. 

Gaudry, 1862.— Gaudry (1862, p. 153) pointed out that Cuvier established the name tapiroides on a single 
fractured milk molar ("molaire de lait"), a specimen which had been mentioned as early as 1783 by Defay; the 
third tooth cited by Defay was described by Cuvier as M. tapiroides. 



194 



OSBORN: THE PROBOSCIDEA 



Gaudry also pointed out, referring to Schinz, von Meyer, de Blainville, Lartet, and Kaup, that the teeth 
described throughout France as 'M.' tapiroides chiefly belong to the 'M.' turicensis Schinz. Gaudry then 
described (op. cit., p. 155) and figured (Pi. xxiv, figs. 1 and 2) the jaws of a young specimen of 'M.' turicensis from 
Pikermi, including three superior and inferior molars figured together in PI. xxiv, fig. 1 (Fig. 165 of present 
Memoir) ; the superior premolars enlarged (PI. xxiv, fig. 3) as compared with Cuvier's Lower Miocene type, 
the inferior premolars also enlarged (PI. xxiv, fig. 4 — cf. Fig. 165 of present Memoir for Dp4), and superior view of the 
jaw (PI. XXIV, fig. 2 — Fig. 165 of present Memoir). These Lower Phocene grinders of Pikermi age referred to 'M. ' 
turicensis appear to be somewhat more progressive than the Upper Miocene stage of 'M. ' turicensis of Schinz. 
The several characters cited by Gaudry are enumerated below in our specific description of 'M. ' turicensis in the 
present Memoir. See PL iii, pp. 134-135, diagrams of Turidus molar evolution. 

LoRTET AND Chantre, 1879. — The next great step was that of Lortet and Chantre (1879), when they brought 
together as Group II the following six species : 



Mastodon tapiroides 
" turicensis 



Mastodon borsoni 
" buffonis 



Mastodon vellavus 
" vialetii 



This Group II corresponds in large part with Vacek's Zygolophodon pubUshed two years earher. 
in 1909, practically the same zygolophodont group of species was adopted by Mayet. 



Finally, 



M. Borsoni Hays. 

M. Turicensis v. Pikermi 

Fonn V. Baltavir. 



M. aitictts Wagner 

M. hngirostris Kaup. 

Former aus dem Flinz der Tsar, 
d. Sandsteine von Veltheim, 
d. Leythakalke. ? Steinheim. 



M. tapiroides Cuv. M. pyrenaicus Lart. 1/. angustidens Cuv. 

Halbjoche in gleicher Linie. 

Zygolophodon 



M. arvernensis Cr. et Job. 
— M. Pentelici Gaud. 



Halbjoche alternirend. 



Bunolophodon 



Mastodon. 

Definition and phylogeny of Zygolophodon (subgenus Bunolophodon), in which the characters of the grinding teeth of 
Zygolophodon and of Bunolophodon {Mastodon angustidens, M. arvernensis) are contrasted. .'Vfter Vacek, 1877, p. 45. 



Vacek, 1877. — Meanwhile Michael Vacek, the Austrian palaeontologist, in his Memoir "Uber Osterreich- 
ische Mastodonten," Vienna, July 1, 1877, gave a complete revision of the mastodonts of Europe from the primi- 
tive period of the contributions of Baldassari (1767), of Pallas (1777), of Buffon (1784), of Ildephons Kennedy 
(1785), to the contributions of Biedermann (1876). He especially mentions the masterly memoirs of Etienne 
Borson (Turin, 1820, 1823), the palaeontologist to whom Hays dedicated his species 'Mastodon' borsoni in 1834. 

The most valuable parts of Vacek's contribution are his illuminating reviews of Mastodon tapiroides Cuv. 
(including M. turicensis, pp. 4-6), of M. borsoni Hays (pp. 6-11) in comparison with M. angustidens Cuv. (pp. 
12-25), with M. hngirostris Kaup (pp. 25-33), and with M. arvernensis Croizet and Jobert (pp. 33-39). In the 
case of each species valuable observations are given upon the known geographic distribution, as well as the 
resemblances and differences of the grinding teeth in relation to other species. Vacek sums up (pp. 39-45) the 
clear separation of his 'Group' Zygolophodon from his 'Group ' Bunolophodon, as presented in his diagram herewith. 



THE ZYGOLOPHODONTIN^: HISTORY 195 

We may by courtesy take Vacek's group name as a valid definition of the genus Zygolophodon which he based 
on the species Mastodon borsoni, M. turicensis, and M. tapiroides { = M. pyrenaicus). 

Vacek's diagnosis and diagram above embody a clear recognition of three distinct phyla corresponding 
with the three phyla of this Memoir, namely, the Zygolophodontinae, the Longirostrinae ( = Bunolophodon: "Halb- 
joche in gleicher Linie"), and the Brevirostrinae ( = Bunolophodon: "Halbjoche alternirend"). Vacek concludes 
as follows (p. 45) : 

^"^ ' B. — MASTODONTES A MAMELONS DISPOSES EN CRETES 

TBANSVERSALES (TYPE LOPHODONTE) 

Das Vorkommen von Uebergangstypen in einer Zeit, wo die beiden 

Foniieiignippen noch nicht so scharf gcschicdcii waren, wic wir dies in Mastodon turicensU Schmi. 

den jungsten Ablagerungen sehen, liat deninacli niclits Auffalleniies an .sjs. Afasiodon upiroides. Cu»icr, cement, fouiUt, i. .. p. ,0;, 

sich, ja dasselbe muss sogar von vornherein erwartet warden, und die i'| '!,'; "b- ^ ^'- '""' «"«»="■<), hUmoire,. i. vi, .o«m., 

, ^ , . . pi- Ml, fig. 4. 

Trennungdcr Forinon nach Massgabo ihror phyllogcnetischen [sic] Ent- i«j7. Mastodon turiceatis. Schim, .\3iurg. und Abbud. d. Siuge- 

wickelung odd' dii- Trcnnung in Zvgolophodontc und Hunolophodoiite „, iLf„„„i'"'7' '■*''' ''^ Vi' >• n. ,; ■ 1 /- i 

" j^ ^ c- 1 I 'Hi. Mastodon turiceatis. W. son Meyer, PtUronlologica :ur Getchi- 

ist sonach eine natiirliehe. WoUtou wir uns zum Schlusse nach den in <-hie derHnhund ihrer Ceichopfe.Fr^ncturi. 

d\ e t I r^ • I i li T'l ui- 1 •■ I I- i85o. Mastodon tapiroi'des- Blainvilk*. Oitvographie, O. Elepliaot. 

leseni Aufsatze gegebenen Gesiclitspunkten cu.en loberblick ubcr (ho ,85^,. „^stodon Borsonii. c. Gerva.s (p. pJ.),ZooiogicetPMUonio. 

bishcr bekaniiten euiopiiischen Foiinon der Clattung Mastodon schaffen, iugie .frani:ai>es, 2' ii]., p. 68. 

t .. . •11'/-, • 1 n -CI » 4 J 1 1 1878. Mastodon atfinis. Jourdan. Archiv'rs du Museum de Lyon. I. II. 

SO konnte man sich die Gruppirung derselben m folgender Art denken. ,«,«. Mastodon tapiroides. Loriet ei Ch.nire, Rccher<*es sur les 

Maslodonles (Archives du Museum Je Lyon, t. II, p. 285. 

Mayet, 1908. — Whereas Vacek (1877) and Schlesinger (1917, iss,. Maswdon^ur'icinsjs. ocp^ret, ven^br^s mioc^nes de u viii^e 

1 rvnrfc\ 11 j.xl tur 1 1 1 • • 1 > A^ i iji du hUone (Archii-es du Musf^um de Lyon, t. W, p. i3i). 

1922) adhere to the name Mastodon tapiroides throughout the igoS. AfastodontuWceosis. L.Mayet,Mamniireresmiocenesde.s.bl« 
_ _. _ , /if\r\t-^ -I rkrt\ 1 1 1 1 • r *^^ I'Orleanais et des faluns de la Touraine (.\nnztea de 

Miocene, Mayet (1908, p. 196) abandons the above name in lavor frn.i..rs.«'</<-L!/on,fasc.24); — dessabiesderori^anais, 

.... . /--\ • P- '9^' P'' ^'^'l» ^S- ' el a; ^ des faluns du Bl^sois 

of Mastodon turicensis owing to the uncertainty as to Cuvier s p. ags, pi. xi, ng. 4 ci 5. 

lost type specimen. Mayet, on the other hand, sharply distin- Type and refem-d speeios related to the Mastodon 

., , . irji ■ T e uTi'-j tapiroides, M. turiceruds, .\I . borsoni groups. After Mayet, 

guishes the species Mastodon pyrenaicus Lartet from M. tapiroides 1909, ,,. 4.5-aii referable to the Zygoiophodontins. 

( = M. turicensis) and points out that M. pyrenaicus occurs in the 

Lower Miocene, Burdigahan stage, as well as in the Middle Miocene horizon of the tj'^pe specimen of Lartet. 

Compare also Mayet, 1909, p. 45, Section B, "Mastodontes a Mamelons Disposes en Cretes Transversales 

(Type Lophodonte) " herewith. 

Klahn, 1922, 1931.— Klahn, in describing (1922) the mastodonts of Baden and Rheinhessen, cites the 
species Mastodon wahlheimensis, which the present author regards as belonging to the Zygolophodontinae. 
This is included, therefore, in the list of species on page 192 as Turicius wahlheimensis; it is not described, how- 
ever, in the present chapter, but is briefly treated together with other of Klahn's species in Chapter Mil, Sec. 
11, "Miocene and Pliocene Mastodonts of Baden and Rheinhessen," on pages 281-283. 

2. COLLATERAL RELATIONSHIP TO THE SUBFAMILY STEGODONTIN^ 
Compare Vol. II, Chap. XV, Stegodontin.e (see Stegolophodon, PI. iv, pp. 134-135) 

Schlesinger, 1917. — Schlesinger (1917) in describing his type specimen of Mastodoti {Bunolophodon) longi- 
TOsire Kaup forma sublatidens n.f., discovered near Teschen (Schlesien), Austria, proposes the new subgeneric 
name Stegolophodon for 'Mastodon' latidens Clift (see Schlesinger, p. 115, footnote): 

Ich schlage fiir M. latidens [Clift], dassich durch seine kurze Symphyse von dem Subgenus Bunolophodon, durch seinen 
Molarenbau von Dihunodon entferrit, den Untergattungsnamen Stegolophodon vor. Der Name bringt einerseits die nahen 
Bezioliungon zuin (icniis Stegodon. anderseits die Loslosung der Untergattung von Bunolophodon und ihre Sonderstellung 
gongeniiber Dihunodon zum Ausdruck. 

Schlesinger's type specimen of 'Mastodon' sublatidens from Teschen (Schlesien) certainly suggests the zy- 
golophodont form of molar, but it more strongly resembles Clift's type of M. latidens from Burma, and stimulates 



CONTRASTS BETWEEN LOWER MIOCENE MIOMASTODON (FIG. 138A), TRILOPHODON (FIG. 138B, AND FIG. 139, 2), AND TURICIUS 

(FIG. 138 C, CI, FIG. 139, 4, 5, AND FIG. 140) 




B 



0^%:m^ 




Lower Miocene Mium.^stodon deperbti 
sp. NOV. (A),Trilophodon pontileviensis 
(B), andTuriciustapiroidesRef. (C, Cl) 

Fig. 138. A, R, From the Sables dc 
rOrl6anais, Lower Miocene," C, from Pontle- 
voy, Lower Miocene, .\fter Mayet, 1908, PI. 
VII, figs. 3 and 5, PI. xi, figs. 4 and .5, reduced 
to one-third natural size. 

A, {Op. cit., PI. VII, fig. 3): "Z.—Maslodon 
angustiileus. Chevilly (.sahliere Cassegrain, 
au Glorieux, aotlt 1877). Derniere molaire 
superieure gauche. Paris, Museum. Grandeur 
naturelle." [ = Miumaslodon depereli sp. nov.| 

B, "a. — Mastodon angustidens. Chevilly. 
Derniere molaire superieure [H. F. O. inferi- 
eure]. Paris, Museum. Grandeur naturelle." 
[ = THlophodoii, pontileviensis.] 

We observe that the Trilophodon grinders 
are much narrower than those of Turiiiu.t. 

C,C1 (Op. eU.,V\. XI, figs. 4 and 5): "4.— 
Mastodon turicensis [ = T. tapiroides H. F. 0.|. 
Pontlevoy. Derniere molaire inffirieure. 
Paris, Musd-uni. Grandeur naturelle 5. — 
Id. vue d'en haul." (See Fig. 139 for en- 
larged view of these teeth.) 




Lower Miocene. Turiciu.s (4, 5) and Trilophodon (2) 
Fig. 139. Turicius turicensis [ = tapiroides] (upper 4, 5) 
compared with Trilophodon pontileviensis (lower 2), one-half 
natural size. After Mayet, 1908, PI. xi, figs. 2, 4, and 5. 

(Upper) Fig. 4. — "Mastodon turicensis [ = T. tapiroides]. 
Pontlevoy. Derniere molaire inf^rieure. Paris, Museum. 
Grandeur naturelle." Fig. .5. — "/d. vue d'en haut." (See also 
Fig. 138C, Cl.) 

These molar teeth exhibit the profound difference Ijetween 
the zygolophodont molar (above), resembling that of a primi- 
tiveStegodont with uninterrupted, widely open valleys, and the 
bunolophodont molar (below), in which the valleys are closed 
by the central conules. 

(Lower) Fig. 2. — "Mastodon angustidens. Falun de Pont- 
levoy. Derniere molaire superieure. Communiqu^e par ^L 
Jeande Bodard, Pontlevoy. Grandeur naturelle." [ = Trilopho- 
don pontileviensis.] 




Referred TuRicins turicensis Molar 

Fig. 140. A lower molar, l.Mj, from Murinsel, Croatia, 

referred by Vacek to Mastodon [ = Turiciwi] tapiroides, but 

which appears to be closer to the stage Turicius virgatidens. 

.\fter Vacek, 1877, Taf. vii, figs. 4, 4a, one-third natural size. 



Observe that in Triloplwdon pontileitiensis central eonules bIfKsk the two anterior valleys; there are /our conelets of unequal size in each loph. Observe in 
Turicius that all four of the valleys are open and free of central conules: each loph consists of five to six conelets, foreshadowing the ecto- and entoserrate 
lophs of the genotypic species Turicius turicensis (Fig. 102). In the Simorre stage, Turicius simorrensis (Fig. 141 ) displays serrate spurs from the ectoconelets. 

The Lower Miocene Trilophodon angustidens of France (cf. Trilophodon pontileviensis, Chap. VIII, fig. 230) is paralleled by T. cooperi in India. 

The Lower Miocene .Vliomaslodon depereli sp. nov. of France is fully described in the Appendix of the present Volume 1. 

196 



THE ZYGOLOPHODONTIN.E: HISTORY 



197 



our comparison of the type molar of Mastodon sublatidens Schlesinger with the type molar of Mastodon [ = Stego- 
lophodon] latidens Clift. Schlesinger does not state whether he also applies his new subgeneric name Stegolophodon 
to his species sublatidens, but it is obvious that M. sublatidens and M. latidens belong to the same genus, i.e., 
Stegolophodon; consequently Osborn has referred Mastodon sublatidens to the subfamily Stegolophodontinae' 
(Chap. XV), where it is treated in detail as Stegolophodon sublatidens. See PI. iv, pp. 134-135. 

ScHLOSSER, 1903. — Long prior to Schlesinger's note, Schlosser suggested that primitive Asiatic species of 
Stegodon may have been derived from mastodonts related to the 'Mastodon' turicensis of western Europe. A remote 
resemblance may be observed between specimens from the upper Middle Miocene of Simorre, France (Fig. 141), 
referred to M. tapiroides by Lartet (made the type of Turicius turicensis simorrensis by Osborn in 1926), and 
specimens from the Middle Pliocene of Perim Island, India, Lydekker's cotypes of M. cautleyi (see Fig. 142 A, B, 
Osborn's lectotypes of il/a.stodo/i [ = Stegolophodon] cautleyi). See Pis. ii-iv, pp. 134-135 {Zygolophodon, Turicius, 
and Stegolophodon). 





B 




FiC. 230. — Oernicic moUiiie infuricurc dii Maslodim turicfnsis [lujiirnides], 
aux 2/5 de grandeur. — Miocene mojcn do Siniorrc, Cicrs. (D'apres Larlet.) 

Turicius (C) and Stegolophodon (A, B) Form of- Grinding Teeth 
Fig. 141 . (C) Tv|if third right inferior molar, r.M.3. of Turiciti.^ titricetixU simorrensis, orronoously dctorminiHi !>>• Land 1 1859, PI xv. fig. 3 a.-- Mastodon 
tapiroides, two-fifths natural size. Ifjiper Middle Mioeenc of .'!imorro. .\fter ("laudry. 1S7S, p. 174, fig. 230. Reversed in drawing. S<h? also pages 207 and 220, 
Fig. 142, Cotypes of Mastodon [= Stegolophodon] cautleyi Lydekker, 1884. I-ectotypes (Osborn), compare Volume II. 

A, First superior molar of the left side, l.M', one-third natural size. After Falconer and Cautley, 1846 (1847, PI. Lx, figs. 3, 3a), as "Mastodon latidens." 
Length 4 inches, width 2.3 inches, Brit, Mus. M,2817, Cast .\mcr, Mus, 26965. Perim Island. 

B, Third superior true molar of the left side, l.M', one-third natural size, .\fter Falconer and Cautley, 1846 [1847, PI. xxxi, figs. 6, 6a] as "Mastodon 
lalidem." Length 8,.') inches, width 4.5 inches. Brit. Mus. M.2705. Cast Amer. Mus. 26966. Perim Island. See also Lydekker, 1886.1, p. 73, fig. 18. 

A further striking resemblance to the Stegodon type is seen in the molar teeth referred by von Meyer to M. 
turicensis from the Upper Miocene lignites of Elgg and Kiipfnach (Fig. 162). Finally the M. virgatidens type of von 
Meyer (Fig. 168) exhibits a progressive evolution somewhat closely parallel with progressive species of Stegodon. 

It is not impossible that the forest-living Zygolophodon i)hvlum of .southern Europe may have given rise in 
Lower Miocene time to forest-living animals which spread eastward into tlie forests of southern Asia and developed 
into the Stegodont phylum. This whole phylogenetic problem is so important that it is treated more fully in 
Vol. II, Cliap. XV, the Stegodontin£e, of the Stegodontoidea superfam. nov. 

Osborn, 1935. — Osborn is inclined to favor Schlosser's sugge.stion of 1903 and Schlesinger's suggestion of 
1917 that i)rimitive European species {Stegolophodon sid)latidens) may have given rise to primitive species of 
Stegodonts of A.sia to which Sdilesinger has given the new generic name Stegolophodon . See Pis. ii-iv, pp. 134— 
135, also full treatment in Chapter X\' (the Stegodont inie). 

'Sec Appendix of the present Volume I, pages 685 et seq. 



3. ZYGOLOPHODON AND TURICIUS INDEPENDENT PHYLA (OSBORN, 192G) 
OF THE ZYGOLOPHODONTIN^E OSBORN, 1923. PLATES II AND III 

As detailed above (pp. 192-197), beginning with Buffon in 1770 the Zygolophodontinse (Osborn, 1923.601) 
were clearly separated (0.sborn, 1926.706) from the Mastodontinse chiefly by (1) multiplying conelets, (2) increas- 
ing sharpness of the ridge-crests, (3) open central valleys, and (4) absence of central conules and reduction of sulcus. 

ScHLESiNGER, 1917-1922. — In his detailed researches on the morphology, phylogeny, environment, and 
stratigraphy of the European mastodonts, Schlesinger gives repeated illustrations and summaries of the 
profound differences which exist between the species described in Europe as Mastodon borsoni and the species 
described as Mastodon tapiroides, the latter including the M. turicensis Schinz. We may therefore credit to 
Schlesinger (1922) the observation of clear and sharp distinctions not only between M. americanus and M. borsoni, 
sufficient to place them in different phyla, but of still more importance the sharp distinctions between specimens 
referred to M. borsoni and those referred to M. tapiroides- M . turicensis. 

Osborn, 1927 : Finally in studying the detailed structure and geologic stages of evolution of these Miocene and 
Phocene Zygolophodonts, Osborn independently (1926.706, p. 3) concluded that they represent two genera, the 
included species of wliich may be given here as follows : 

Zygolophodon Vacek, 1877 Turicius Osborn, 1926 

Zygolophodon borsoni type, of Asti, Italy Turicius virgatidens type, of Fulda, Germany 

Turicius atticus type, of Pikermi, Greece 
Turicius turicensis type, of Elgg, Switzerland 
Zygolophodon pyrenaicus, Ile-en-Dodon, France Turicius turicensis simorrensis type, of Simorre, France 

Turicius turicensis ref., of Pontlevoy, France 
Zygolophodon pyrenaicus aurelianensis, Sables de Turicius tapiroides type, Calcaire de Montabusard, 

rOrleanais, France' France 

Whereas Vacek included Mastodon turicensis in his group Zygolophodon, Osborn points out that the type and 
referred specimens of Turicius, extending from the Lower Miocene of Pontlevoy to the Middle Pliocene of 
Germany, are clearly separable from the type of the true Zygolophodon and constitute a distinct ascending series 
of species with certain resemblances in the grinders to the species grouped under the genus Serridentinus Osborn. 
They thus constitute the types of a distinct genus Turicius typified by the typical Mastodon turicensis of Schinz 
but embracing a whole series of stages extending from the Lower Miocene to the Middle Pliocene. 

4. RESUME OF THE GENERIC AND SPECIFIC DISTINCTIONS 

It appears that the Zygolophodon of Vacek divides into two generic phyla. Before we can clearly summa- 
rize the generic and specific characters of Zygolophodon and Turicius and the species included within them, it is 
necessary to concentrate our attention on the type geologic locality and level and the characters of the type grinding 
teeth. Previous failure in this respect has led to great confusion which exists even to the present time. 

Geologically the oldest type is the Turicius tapiroides Cuv. from the Lower Miocene lacustrine Calcaire de 
Montabusard, France, but there is no certainty as to the precise characters of the type specimen. Geologically 
the most recent type is the Zygolophodon borsoni from the Middle to the Upper Phocene of Asti, Italy; al- 
though there is some reason to beheve that this species may have survived in eastern Europe into Lower Pleisto- 
cene time ; it certainly disappeared in western Europe at the close of Phocene time. 

198 



THE ZYGOLOPHODONTIN.E: HISTORY 199 

Type Localities. — The summary of the chief type locaUties, widely scattered through northern Italy 

Switzerland, Germany, and France, is as follows: 

Type Localities Original Name Specific ReferexNce in 

Present Memoir 

Middle Pliocene, Villanova, Asti, Italy Type of Mastodon borsoniVLsiys, 18M =Zygolophoflon borsoni 

Middle Pliocene, Fulda, Frankfort, Ger- Type of Madodon virgatidens von 

many Meyer, 18(37 = Turicius virgatidens 

Lower Pliocene, Pikermi, Greece Type of Mastodon atticus G. & L. (in 

Wagner, 1857) = Turicius atticus 

Upper Miocene, Elgg, ("anton Zurich Tj'pe of Mastodon luricense Schinz 

Switzerland 1824 = Turicius turicensis 

Middle Miocene, Ile-en-Dodon (Haute- Type of Mastodon pyrenaicus Lartet, 

Garonne), France 1859 = Zygolophodon pyrenaicus 

Lower Miocene, Calcaire de Montabusard, Type of Mastodon tapiroides Cuvier- 

France Desmarest, 1806-1822 = Turicius tapiroides 

Specific Characters. — Having thus distinguished the type geologic levels and localities, let us clearly distin- 
guish the characters of the type specimens in descending geologic order; these characters are based solely on the 
type grinding teeth, since our knowledge of the incisive tusks and of the characters of the jaws is very limited. 
Doubtless the tusks exhibit other distinct characters. See Pis. ii and iii, pp. 134-135. 

Type M', Mastodon borsoni (see Figs. 154 and 11L\, after Hays and Buffon) = Zygolophodon 

Third superior molars tetralophodont (4^)2 crests), each loph composed of/o;/?-' distinct and subequal conelets('mamelons'). 

Type Ms, Mastodon virgatidewi (see Fig. 168, after von Meyer) = Turicius 

Intermediate molars trilophodont, posterior molars subtetralophodont, lophs acute with_^oe to nine connate conelets (' Wulst- 
kanten,' 'mamelons'); a serrate spur on superior internal conelets. 

Type M-, Mastodon turicensis (see Fig. 162, after Schinz) = Turicius 

Second superior and inferior molars trilophodont, each loph with fire to six more or le.ss distinct conelets, no median 
sulcus; ridge-crests slightlj' arched or directly transverse; strong serrate spur on superior internal conelets. 

Type Ml, Mastodon pyrenaicus (see Figs. 147, 148, after Lartet) == Zygolophodon 

Third inferior molar tetralophodont plus, i.e., 4}, crested, each loph composed of three to four' more or less distinct conelets; 
no median sulcus; no serrate spur on external conelets; vestigial conules in anterior valleys. 

Type r.M;, of Mastodon tapiroides (see Fig. 160, after Cuvier, Fig. 161, after Mayet, also PI. iii A, pp. 134-135). = Turicius 
Third inferior grinders tetralophodont, each loph composed of five to six conelets; no median longitudinal sulcus; vestiges 
of conules in the middle of the valleys; trefoil spur apparently on metalophid. 

Generic Characters. — From the above principal characters of the type grinding teeth we observe that the 
universal generic character both of Zygolophodon and of Turicius is the absence or reduction of the longitudinal 
sulcus, fissure, or commissure, which Hays pointed out as the chief character distinguishing liis type (Fig. 154) 
of M. borsoni from M. americanus. The absence or reduction of this sulcus distinguishes all the ascending species of 
Zygolophodon and of Turicius, from early Miocene to late PUocene or early Pleistocene time, from all the ascending 
or ancestral stages of Mastodon in the corresponding period. We also observe that the fission of the cones into 
'conelets' is rapidly progressive and distinct in Zygolophodon and in Turicius, whereas it is decidedly retarded in 
Mastodon. In many other features (compare Schlesinger, 1922) the progressive dental and skeletal evolution of 
Zygolophodon parallels that of Mastodon, as described in the previous chapter, namely, (1) the abbreviation of 
the jaws, (2) the suppression of the premolar dental succession, (3) the reduction of the lower incisive tusks, (4) 
the progressive abbreviation of the cranium. (5) From the fact that the Zygolophodon molars are relatively 
broader and shorter than those of Mastodon, it is probable that the cranium of Ztjgolophodon was more brachy- 
cephalic than that of Mastodon. 

We may therefore give the following provisional generic distinctions of Zygolophodon and of Turicius from 
Mastodon, which will be ampUfied as we gain a more thorough knowledge of the skeleton. See also the full 
distinctions cited below (p. 211) from Schlesinger, 1922; also compare Mayet, 1908 (p. 204 of the present Memoir). 

'Four conelets in types of Zygolophodon pyrenaicus and Z. borsoni; four to six conelets in progressive stages (PI. 11 D-G). 



200 



OSBORN: THE PROBOSCIDEA 



Median valleys open, central conules rudlmentary or absent (5). See Pls. i-iii, pp. 134-135 



Mastodon 
Genotypif species: Elephas [= Mas- 
todon] americanus Kerr. 

1. Longitudinal sulcus persistent 
between internal and external cones. 

2. Main cones feebly subdividing 
at the summits into two conelets, pro- 
gressive to four (Fig. 89, Miomaslodon) 



3. Retarded 'trefoil spurs' on supe- 
rior internal cones and on inferior ex- 
ternal cones. 

4. Grinders, Dp-'-M', relatively 
longer and broader, breadth - length 
index of M3 not exceeding 57 (highest 
recorded index in Amer. Mus. speci- 
mens, Amer. Mus. 14345) — 60. 

5. Upper incisive tusks rounded, 
enamel band persisting into Miocene 
time. 

6. Lower incisive tusks straight, 
cylindrical, variable. 



Zygolophodon 
Genotypic species: Mastodon borsoni 
Hays. 

1. Longitudinal sulcus disappear- 
ing, vestigial. 

2. Main cones strongly subdivid- 
ing into four to five subequal conelets, 
progressive to six (Fig. 151 and PI. 11). 



3. 'Trefoil spurs' vestigial or absent. 



4. Grinders relatively shorter and 
broader, index of Ms = 43-45. 



5. Rudimentary conules or tubercles 
present in median valleys in early 
stages, i.e., Zygolophodon pyrenaicus ref. 

6. Upper incisive tusks rounded, 
enamel band disappearing early, before 
Lower Miocene time. 

7. Lower incisive tusks reducing 
early. Symphysis more abbreviated. 



TUEICIUS 

Genotypic species: Mastodon turi- 
censis Schinz. 

1 . Longitudinal sulcus disappearing, 
vestigial. 

2. Main cones subdividing into five 
to nine conelets, i.e., summits of lophs 
with quintuple, to sextuple, to octuple, 
to nonuple conelets (compare Serri- 
dentinus serridens, Figs. 388 and 389). 

3. Strong 'trefoil spurs' on superior 
internal and inferior external cones, 
hence 'serridentine.' 

4. Grinders relatively longer, in- 
creasing in breadth. 



5. Central conules vestigial in the 
median valleys in early stages (T. 
tapiroides) . 

(5. Upper incisive tusks oval; en- 
amel band sharp, persisting late, absent 
in T. virgalidens. 

7. Lower incisive tusks suboval, 
vestigial in Pliocene stage. Symphysis 
less abbreviated. 



Generic Distinctions into two Groups. — The Mastodontinse so far as we know embrace one phylum of 
species, the Zygolophodontinae, however, appear to embrace two phyla or groups (I and II) of species as follows: 

I. In Zygolophodon prjrenaicus, type and referred, and in Z. borsoni, each loph of the grinding teeth sub- 
divides into not more than four, to four and a half, to five to six conelets; the lophs are directly transverse not 
arched; trefoil spurs vestigial or absent. See Plate 11. 

II. In Turicius turicensis and T. virgalidens, each loph subdivides into from^^e to nine to Iwenlij-five conelets; 
strong trefoil spurs arise on the superior internal conelets and on the inferior external conelets. See Plate in. 

Consequently we appear to observe two ascending contemporaneous generic and specific phyla which do not 
intergrade. Geologically these phyla show two independent lines of ascent. 



5. PROGRESSIVE CHARACTERS OF THE GRINDING TEETH 



Upper Pliocene 
Middle Pliocene 
Lower Pliocene 

Upper Miocene 
Middle Miocene 

Lower Miocene 
Lower Miocene 



L Zygolophodon 
Four to six persistent conelets on each 

grinding crest, 'trefoil spurs' vestigial 

or absent 
Zygolophodon borsoni ref. 

Zygolophodon borsoni type, Asti, Italy 



Stage undescribed 

Zygolophodon pyrenaicus, Ile-pn-Dodon, 
France 



Zygolophodon pyrenaicus awelianensis, 
Sables de I'Orl^anais, France 

Parallel evolution with Mastodon 
americanus phylum 



II. Turicius 
Five to nine conelets on each grinding 
crest, progressive to twenty-five, 
'trefoil spurs' more or less prominent 

Turicius virgalidens of Fulda, Germany 

Turicius atticus of Pikermi, Baltavar, 

etc. 
Turicius turicensis of Elgg and Kapf- 

nach, Switzerland 
Turicius turicensis of Sansan and 

Simorre, France = 7'. simorrensis 
Turicius turicensis, Faluns de Pont- 

levoy, France 
Turicius (?) tapiroides, Calcaire de 

Montabusard, France 

Parallel evolution with Stego- 
lophodon and Stegodon phyla 



THE ZYGOLOPHODONTIN.E: HISTORY 



201 



Turicius. Undoubtedly constituting a phylum (Group II) distinct from above species of Zygolophodon 
(Group I) are the numerous grinding teeth referred by all recent European authors, especially Mayet, 1908, to 
'Mastodon' turicensis, first observed in the Lower Miocene Faluns de Pontlevoy and Sables de I'Orleanais, again 

Kig. 143. Miildle (=fitage Helvdtien) and Upper (=fitage Tortonien) Miocene 
correlation. .\fter Depc'Tet (190.5-1908) and Osborn (1910.346, p. 2.58, fig. 129). 
Horizons 1-69 arranged in ascending geologic order. 




MIDDLE AND UPPEB MIOCENE 

Horizon of Sansan. FRANCE. — Calcairo 
de 1 Snnaan (Gers), lacustrine. 2 Jcgun (Gers). Faluns de 3 Pontlevoy, 4 Saintc-M aure, 
6 Manthelan, in the basin of the Loire, marine. MoUasse de 6 pont dc V Herhassr, 7 Bren, 
8 C/^ricuj, near Romans, marine. GERMANY. — Mecresmolasse (in Swabia) von 9 Baltrin- 
Qen, Hnmmingcn, Hcgghach, Ursendorf, Hochgcland, 10 Hansen^ 11 Niederstozingcn, 12 Siisscn, 
SOsswasaerkalk von 13 Georgcnsgmund, 14 Engelswies (Bavaria). AUSTRIA-HUNGARY. — 
Braunkohte (in .Styria) von 15 Eibiswatd, Wiea, 16 Gdriach, Gamlitz, 17 Voitsherg, Parschhig, 
Seufd, Leoben, Leiding, FeisterniU. Marine sands of Grund at 17a Guntcrsdorf (Lower 
Austria). Horizon of Simorre. FRANCE. — Z Bonnefond, St. Cristan iGns). 2a. Ville/ranche 
d'Astnrac, calcaire de Simorre (Gers). lacustrine. CinfritesdelS Tournon, in south central 
France. 19 I'lsle-en-Dudon (Gers). 20 Saverdun (Arifege). Mollasse de 21 Mirabeau (Vau- 
cluse). marine. Marine deposits of 22 Sorgues (Vaucluse). Marine deposits of 23 Romans 
(Drome). Sid6rolithiques dc 24 La Grivc-Saint-Atban (in large part) (Is^re), 26 Mont Ceindrc 
(Rhone), 26 Pretty near Tournus (Saone-ct-Loire) , 27 Gray (Haute-Saone). GERMANY.— 
In Swabia, Bohnerz von 28 Willmardingen, Jungnau (in large part), Susswasscrkalk von 
Steinheim, Rics, I'rlau, 29 N6rdlingen, 30 Althausen. Bohnerz von 31 Mdsskirch, 32 Genkin- 
gen, 33 Heuhcrg. 34 Melchingen. Gipa von 35 HohenhUven (Baden) (10 meters). AU.STRIA- 
HUNGARY. — In the intra-Alpine basin of Vienna, 36 Dornbach, Vordersdorf, Loretio, Marga- 
rethen,Zl Fiinjkirchen (Hungary'), calcaire dc Zi Briu-k-a.-Leitha ; 39 Breitcnbrunn, Ahstdorf- 
Framtnshad, 40 Manmrsdorf, 41 Xcudorf. 42 Trauenzinen (Silesia). 43 Krivadia (Transyl- 
vania). 44 Gyulu-Mendru (Transylvania). 45 Pesth (Hungar>). 46 Ssoxkut (Hungary). 
RUSSIA. — 47 Wos.ikrcssensk. Horizon of Saint-Gaudens. FRANCE. — 48 Valentine, Saint- 
Gnudrn.'! (HautL-Garonne). i9 Monlrfjau (Haute-Garonnc). SWITZERLAND. — Sands of 
60 Dehberg, near Basel. 61 Le Locle, in the northwest. 82 La Chaux-de Fonds, in the north- 
west. .63 Kirmfs, near Basel. Mergelkalke von 64 Q?nins«n, near Zurich, fresh water. Lic- 
nitcs of 65 Elgg, 56 A'apjfnac/i, near Zurich. 57. VV/Mcim, near Zurich, GERM.\NY. — In 
Bavaria : 11 Giinshurg, Rrisensburg : Sande von 68 Hdder, Dinketscherbcn, Rciehenau, 69 Piis- 
ten. 60 Dosing, Fraising, Tulzing, SIfttzling, Reicherlsho/en, 61 Fronlenhausen, Ftinz of Munich. 
62 .Sankt Georgtn (Baden). AUSTRIA. — 36 Heiligcnsladt, Tegel von Hernnls. in Vienna 
ba,Hin. ITALY — Lignites of 63 Ml. Bamholi (Tuscany). SPAIN. — 64 .Snn hidrn, near 
Madrid. PORTUGAL. — 65 Aveiras de Baizo, near Oporto. RUSSIA— 66 Kriuoi-Rog 
(Kherson). 67 Xieotnieff, at the mouth of the Dnieper. 68 Sebaslopol (Crimea). 69 Tira-ipol. 
near the mouth of the Dniester, Correlation of Depirct. 




TURICIUS V1PGATIDE.NS 

GERMANY 



2400nira, 7'io*4'e 




TURICIUS TURICENSIS 2bOomni..e'6to'' e 

SWITZERIJ^ND 



TURICIUS TAP 




FRANCE 



2ioonini..6'io*»'e 




ZYGOLOPHODON PYRENAICUS aQrELIANENSIS E725mm,.8'llliCe 
FRANCE 

Fig. 143a. Ascending phvlogenetic succes-sion of three 
species of Turicius and one of Zygolophodon. Compare Appendix. 

4. Turicius virgatidenx of Fulda, Germany. Middle Plio- 
cene. 

3. Turicius luricensis of Elgg, Canton Zurich, Switzerland 
(see Fig. 143, 66). Upper Miocene. 

2. Turicius tapiroides, Calcaire do Montabusard, France. 
Lower Miocene. (See Fig. 146, 6). 

1. Zygolophodon pyrenaicu.i nureliattensis. Sables de I'Or- 
li?anais, France. Ix)wer Miocene. 

These outlines, restored to a one-hundredth scale, are chiefly 
after the grinding teeth and tusks, the skeleton and skull being 
relatively unknown or undescribed at the present time. 



in the Middle Miocene of Sansan and of Simorre, again in the Upper Miocene of Elgg and of Kapfnach, and 
finally in the Lower Pliocene of Pikermi and of Baltavar. These animals show the strong upper internal and 
lower external serrated 'trefoil spur' ('Wulstkante' of von Meyer) also characteristic of Serridentinus; in fact, 
the grinding teeth show certain resemblances to the species included mthin the genus Serridentinus, but the 
lower incisive tusks of Turicius are entirely different from those of Serridentinus. Von Meyer observes that the 



202 



OSBORN: THE PROBOSCIDEA 



lower incisive tusks of 'M. ' turicensis are reduced and that the upper incisive tusks are oval in section and exhibit 
an enamel band in the specimens referred to M. turicensis from Kapfnach. In brief, there are many strong 
resemblances between the grinders of Turicius turicensis and the grinders of Serridentinus; consequently we are 
disposed to think that there may be an affinity of the genus Turicius to Serridentinus which may be demon- 
strable by fuller knowledge. 

These animals certainly constitute a distinct genus, to which the name Turicius is applicable. Whereas 
Vacek defined Zygolophodon (1877) on three species, namely, Mastodon borsoni, M. turicensis, and M. tapi- 
roides { = M. pyrenaicus) , the genus Zygolophodon would now rest solely on the following species, Mastodon 
borsoni and M. pyrenaicus. 

Osborn, 1935: In the Appendix of the present Voliune are described two specimens of Turicius turicensis 
in the Munich Museum, which afford a fuller knowledge of the dentition, mandibular rami, and fore- and hind- 
limbs (see also PI. in, pp. 134-135). 



Anomalous Double Supekior Incisors 
Pontier and Anthony (Dec. 2, 1929) describe the right and left premaxillaries of a cranium referred to 
'Tetrabelodon' turicensis Schinz, of the following geologic age: "Villefranche d'Astarac, dans le Gers-Calcaire de 
Simorre-Helvetien-Miocene moyen." This interesting specimen, reconstructed in their figure 2 {op. cit., p. 1006 — 




Fig. 144. Reconstruction by Pontier and Anthony, 1929, 
fig. 2, of the head of Turicius turicensis Schinz, of Villefranche 
d'Astarac, exhibiting the supplementary superior incisive tusks; 
inspired by Abel's restoration of Trilophodon anguslidens. 




Fig. 145. Cranium of Loxodonla africana, from Fort 
Sibut, Equatorial Africa, exhibiting reduplication of the 
superior incisive tusks. After Anthony and Prouteaux, 
1929, PI. I, fig. 1. 



reproduced herewith as Fig. 144), contains on both sides a supplementary superior incisor, which emerges im- 
mediately above the normal superior incisor (P). This was regarded by Pontier (1926, p. 162) as representing a 
vestige of, or reversion to, the first superior incisor (P) of the ancestral proboscideans. As to this question of 
homology, it is doubtful whether this anomalous tooth represents such a reversion, or a dupUcation of the crown 
of the normal incisor (P). This tooth has a diameter of 36 mm. as compared with 112 mm. in a normal tooth; it 
curves upwards rather than downwards ; it is certainly an anomaly and not a specific character. 

Comparison may be made with an analogous Loxodonta africana in the dupUcation of the superior incisors, as 
described by Anthony and Prouteaux (1929), of a wild elephant found in the Oubangui Chari region in French 
Equatorial Africa. In this instance the paired superior tusks, beautifully illustrated by the authors (PI. i, figs, i-iv), 
emerge above the normal tusks and are directly horizontal in position, while the length nearly equals the normal 
(Fig. 145). According to the authors, such anomalies are not extremely rare among African elephants but they 
are decidedly exceptional ; they are regarded by the natives as princes of the elephant tribe. We accept the inter- 
pretation of the authors [op. cit., pp. 30, 31) that these anomalous teeth correspond with the second incisors (P) 
of Maritherium. 



6. CHARACTERS OF ZYGOLOPHODON AND ITS INCLUDED SPECIES 

Family: MASTODONTID^ Girard, 1852 
Subfamily: Zygolophodontin^ Osborn, 1923 

Genus: ZYGOLOPHODON Vacek, 1877 
Original reference: Vacek, Abh. geol. Ileichsanst. , VII, Heft IV, p. 45. Wien. 

Now resting;: on the genotypic species Mastodon borsoni and its apparently ancestral form Mastodon pyrenaicus. Compare Mastodon 
\ = Stegolophodon\ lalidens of Burma, and Mastodon [ = Stegolophodon\ sublatidens Schlesinger of the Middle(?) Pliocene of Teschen (Schle- 
sien), Austria. The genu.s Stegolophodon Schlesinger, 1917, proves to be distinct from the genus Zygolophodon Vacek. (See Plates 
ii-iv, between pp. 134-135.) 

Generic Characters of Zygolophodon. — (1) Each loph divided into four to six distinct subequal 
conelets. (2) No trace of median longitudinal sulcus. (3) Conules or tubercles in median valleys disap- 
pearing in early stages. (4) Crests, i.e., lophs, directly transverse, not arched. (5) 'Trefoil spurs' vestigial 
or absent. (6) Fifth inferior crest, i.e., pentalophid, slowly progressive. (7) Superior incisive tusks 
rounded without enamel band {Z. borsoni). (8) Grinders permanently blunt, brachyodont (Z. borsoni), 
not becoming subhypsodont. (9) Progressive adaptations similar to those of Mastodon americanus 
rather than to those of Turicius. Compare generic characters listed by Osborn above (p. 200) and by 
Schlesinger below (pp. 210, 211, 217). (10) No inferior enamel band {Z. pyrenaicus). 

The history of Vacek's generic name is related above in the present chapter. Vacek founded the genus 
Zygolophodon on three species, Mastodon borsoni, M. turicensis, M. tapiroides {=M. pyrenaicus). Osborn (1926.706) 
removed the species M. turicensis and made it the type of a new genus, Turicius; he also now removes the 
species M. tapiroides^ Cuvier to the genus Turicius. 

(1) The chief genotypic species is Mastodon borsoni^ Hays, Middle Pliocene of Asti, Italy. This stage is very 
distinctly characterized by the absence of a longitudinal sulcus between the internal and external cones 
and by the presence of four distinct conelets in each loph, as clearly displayed in Hays' type figure reproduced in 
facsimile (Fig. 154) below; many teeth from different parts of Europe have been referred to M. borsoni by various 
authors which lack the typical characters of the M. borsoni grinders; certain of the grinders figured by Lartet exhibit 
four and a half to five conelets in one or more of the crests. 

(2) Fortunately Lartet 's type of Mastodon pyrenaicus (Figs. 147, 148) from the Middle Miocene of Ile-en- 
Dodon clearly exhibits /owr distinct conelets also tubercles in the first median valley. Mayet (1908, p. 197) also 
records it in the Lower Miocene, Sables de I'Orl^anais, to which stage Osborn (1926.706) as.signed the new sub- 
specific name Zygolophodon pyrenaicus aurelianensis. 

(3) Uncertainty as to M. tapiroides Type Specimen. — The type molar of Mastodon tapiroides Cuvier from 
the Lower Miocene Calcaire de Montabusard, France, has been lost and there is a difference of opinion as to 
whether it is a true molar, or a deciduous premolar (Dp\ Dp^, von Meyer, Gaudry) : Cuvier's figure reproduced in 
facsimile below (Fig. 160) certainly shows five main conelets in each transverse crest and intermediate tubercles 
or conules in the first and second valleys. The anterior portion of the crown is broken off. Mayet's new figure 
(Fig. 161) of the type shows five conelets and a 'trefoil spur,' which relate this specimen to Turicius; he regards it 
Jis ancestral to Turicius (see p. 205). Osborn (1935) determines this molar as an r.Ms (see PI. in A). 

^Mastodon tapiroides was selected by Matthew as the type of Zygolophodon Vacek, 1877 (see Mitthew, 1918.1, p. 200, footnote). 

203 



204 



OSBORN: THE PROBOSCIDEA 



As von Meyer and Gaudry observed, the type appears small for a true molar. It is possible that when the 
type of M. tapiroides and other molar teeth from the Calcaire de Montabusard are studied, they may be found to 
exhibit characters other than those represented in Cuvier's type figure, on which our present description of this 
species rests. As independently observed by von Meyer and Gaudry, all the early French writers, including 
Lartet (1859), erred in referring similar specimens to Cuvier's species M. tapiroides. Von Meyer on the other 
hand states that there is no resemblance between the typical M. turicensis and the typical M. tapiroides; he con- 
siders (1867, p. 50) M. turicensis {tapiroides) as identical with M. borsoni. 

(4) Mayet (1908, pp. 194-197) considers that: (1) Cuvier's type is not certainly derived from the Calcaire de 
Montabusard but more probably from the Sables de Chevilly or d'Ingre (see also Gervais, 1859, p. 68); (2) this 
small type may well be regarded as a milk tooth ("dent de lait"). 

Fig. 146. Chief Lower Miocene localities ( =Etagc B\irdigalien, Sables de I'Orlc- 
anais Stage) and correlation of France, Switzerland, and .\ustria. After Deperet (190.5- 
1908) and Osborn (1910.346, p. 2.50, fig. 126). 




LOWER MIOCENE Burdigalian. FRANCE. — Sables de 1 N eutille-aux-Boig 
(Loiret), fluviatile. sables de Rebrtchien. 2 Marigny, 3 Fay-aux-Loges, 4 Beaugency, S Tavers, 
6 Les Barres, Artenay, marnea da Blesois, calcaire de Montabmard (Loiret), lacustrine. Sables 
de 7 Chilteurs, Chevilly (Loiret), 8 Neuvilly. near Belgian frontier, fluviatile. Sables de 9 Ruan, 
north central France. Calcaires et marnes de 10 Suivres (Loire-et-Cher). Faluns et calcaires 
de 11 Pontlevoy, faluns et sables de Thenay, in la Touraine. Sables de 12 Chilenay (Loire-et- 
Cher), fluviatile. Faluns de 13 Manthelan, in la Touraine, marine. 14 Saint-Nazaire-en- 
Royans (Drome). MoUasse blanche de 15 Angles (Card). LOWER-AUSTRIA. — MoUasse 
of 16 Eggenhurg, marine. Marine deposits of 17 Lim. SWITZERLAND. — Muschelsand- 
stein von 18 Brutlelen, 19 Macconens, La Moliire, 20 Bucheggberg, western Switzerland, 
marine PORTUGAL. — Mollassc of 21 Horia de Tripas. near Lisbon. GERMANY. — 
Spalte der Schieferbrilche von 22 Solevhofen (Bavaria). Correlation of Depfret. 



Osborn, 1925: Cuvier's type is refigured by Mayet (1908, p. 195, fig. 66), as reproduced in figure 161 of the 
present Memoir; Mayet's figure (of natural size) exhibits five to six conelets in each loph, also a 'trefoil spur,' 
and certainly resembles the M. turicensis type, i.e., Turicius, more closely than the M. pyrenaicus type with its 
four conelets in each loph. Mayet concludes that, owing to the uncertainty as to Cuvier's type, the specific name 
M. tapiroides Cuv. cannot be conserved and should be replaced by the name M. turicensis {op. cit., p. 196), "je 
crois preferable de suivre I'exemple de la majorite des auteurs en adoptant le nom de M. turicensis pour designer 
I'espece de I'Orleanais a dents tapiroides." Mayet continues (op. cit., p. 197) that M. pyrenaicus also occurs in 
the Sables de I'Orleanais: "Le M. pyrenaicus que Ton trouve represente dans les sables de I'Orleanais, se separe 



THE ZYGOLOPHODONTIN^.: ZYGOLOPHODON 205 

assez netteiuent du M. turicensis par ses molaires mi-partie tapiroides, ini-partie bunodontes pour que je n'hfeite 
pas a maintenir distincte cette espece creee par Lartet, d'apres un certain nombre de pieces provenant de la region 
l)yr6neenne (Saint-Frajou, Castelnau-Magnoac, Moncamp, etc.)." The species M. turicensis, i.e., tapiroides, 
extends throughout the entire Miocene period (Mayet, op. cit., p. 197) : "L'age geologique du M. turicensis com- 
prend tout I'ensemble du Miocene. On le voit apparaitre, comrne je viens de I'indiquer, dans le Burdigalien de 
rOrl^anais. On le retrouve pendant tout lo \'indobonien et il so rnontre a.s.sez abondant dans la faune des Faluns 
de la Touraine." 

Conclusions, O.sborn, 1925: Our conclusion from Mayet's ob.servations cited above is as follows: Both 
Mastodon pyrenaicus and M. tapiroides occur in the Lower Miocene Sables de I'Orleanais or BurdigaUan ; it follows 
that Group I {=Zygolophodon) and Group II { = Turicius) were separated from each other in Lower Miocene time. 

Deperet and Mayet, 1925. — The recent views of Deperet and Mayet (letter, November 9, 1925), differ- 
ing as regards Turicius, are as follows; 

J'ai dd differer de quelqucs jours ma reponse u votre lettrc du 21 octobrc dernier en raison de ee que je tenais a la souniettre 
a Monsieur Ch. Deperet. C'est done son opinion bien plus encore que la mienne, que je vous exprinie. La figure que j 'ai donn^e 
du aM. tapiroides du ealeaire de Montabusard, est la photographic agrandie de la figure type de Cuvier ramen^e a la grandeur 
naturclle. Le M. pyrenaicus est a considerer comme un representant d'un raineau different, independant, du M. turicensis. II 
ne .senible pas possible de le rattacher au phylum Turicius, non plus qua tout autre. M. tapiroides et M. turicensis se presentent 
comme deux mutations d'un nieme raineau phyletique: I'une, plus ancienne, du Burdigalien; I'autre, plus recenfe, du Vindo- 
bonien. M. Borsoni et M. iimericanus sont l)ien certainement deux representants d'un meme groupe, sinon d'un meme phylum. 
Toutefois M. Deperet se refuse absolument a admettre un passage quelconque ou une convergence de passage au Stegodon. M . 
Bursoni n'est connu imlle part en Eurojie avant le Pliocene. Vous connaissez certainement les deux beaux volumes, largement 
illustrds, de Sciilesinger sur les Mastodontes d'Autriche et de Hongrie ainsi que le travail du meme paleontologist e de Vienne 
paru en 1919: 'Die stratigrapliische Bedeutung der europaisehen Mastodonten,' Mitteilungeii der (ieolugischen (lesellschaft in 
Wien, Band XI, 1918. 11 indique bien que c'est seulenient dans le ' Levantin' que M. Borsoni apparait dans I'Europe orientale 
tandis qu'en P>ance, c'est dans le Pliocene inf^rieur (ancien). 

The above opinions of Deperet and of Mayet concur in many points with those independently reached by 
O.sborn (1925) in this Memoir, except that these distinguished colleagues do not place 'M.' pyrenaicus in the 
ancestral phylum leading to ^1/.' borsoni, as Osborn at present is inclined to do. See PI. ii, pp. 134-135. 

Rarity. — The extreme rarity of remains of specific stages of Zygolophodon is probably attributable to the 
strictly forest-living habits of these animals, similar to the forest-Living habits of the Mastodontinae, true masto- 
donts, which are rarely found fossil in the Tertiary of Europe or America but suddenly appear in great abundance 
in the Pleistocene forests of North America. Similarly Zygolophodon borsoni is relatively abundant in the Upper 
Pliocene deposits of all parts of Europe and of western Asia, but no traces of its ancestors have hitherto been 
described from the Lower Pliocene or Upper Miocene deposits of Europe. Consequently the succession of specific 
stages leaves many gaps, as shown in the following conspectus of the species placed in ascending geologic order. 

4. Middle to Upper Pliocene. Zygolophodon borsoni Hays, 1834. 

Typer.M^, ap. 155+ inm.,tr. 87+ mm.; four conelets ('inamelons') subequal insize, six conelcts in certain crests. 
Incisive tusks without enamel. 

3. Upper Miocene. Stage undescribed. 

2. Middle Miocene. Zygolophodon pyrenaicus Lartet, 1859. 

Type r.M^, ap. 0}^ inches, or 1.5.5 mm., tr. 2^4 inches, or 70 nun.; intermediate conules in first and secoiui valleys, 

third valley without tubercles (Lartet). Pentalophid consisting of two small cones; external and internal conelets 

much larger than median conelets. 

I. Lower Miocene. Referred Zygolophodon pyrenaicus Lartet (fide Mayet, 1908, p. 198), sibspecies airklianen- 

sis Osborn, 192(i. 

Sables de I'Orldanais, ("hevilly (?), Cassegrain, .\varay. Four conelets in each loph, iiitermeiliate tubercles in the 

valleys. This stage is now known as Zi/golophodon pyrenaicus aurelianensis, since the abundant remains listed by 

.Mayet (1908, p. 198) below are undoubtedly distinct specifically or subspecifically from the typical Middle Miocene 

Zygolophodon pyrenaicus of Lartet. 



206 



OSBORN: THE PROBOSCIDEA 



Referred Zygolophodon pyrenaicus of the Lower Miocene 

Mayet (1908, p. 198) confirms the clear specific distinction of M. pyrenaicus and adds the very important 
fact that several specimens referable to M. pyrenaicus occur in the Lower Miocene, Burdigahan, fauna: 

J'ai retrouve les principaux caracteres indiques par Lartet comme specifiques du M. -pyrenaicus sur quelques pieces decou- 
vertes dans les sables de la region d'Orleans: Un fragment de defense superieure (Chevilly, sabliere Cassegrain, collection 
Nouel) est de coupe sensiblement ovalaire, sans la moindre trace de bande longitudinale d'email; . . . Une Ms superieure 
gauche (Avaray, musee d'Orleans) determinee par M. Lartet lui-nieme, est a couronne tres large, supportant quatre collines 
dont les deux premieres ont leurs vallons partiellement interceptes par un tubercule accessoire ; les deux collines posterieures, 
plus comprimees, presque tapiroides, sent separees par des vallons libres; le talon posterieur est en forme de crete crenelee ; . . . 
Une M2 superieure droite du musee d'Orleans, . . . provenant de Beaugency et determinee par M. Lartet; . . . Une Mj 
inferieure (musee d'Orleans, sans localite indiquee) est a trois collines, avec, en arriere, un assez fort talon tubercule. . . . 
Le M. pyrenaicus avec sa dentition a earactere mixtes— buno-lophodonte— se rapproche a la fois du M. angustidens et du M. 
turicensis. II marque peut-etre le debut de I'apparition des Mastodontes a dents tapiroides au commencement du Miocene et ne 
parait pas depasser les horizons superieurs du Miocene moyen. 

Since Mayet observes both M. pyrenaicus ref. and M. turicensis ref. from the Sables de I'Orleanais, the 
separation of Z^jgoloplwdon ( = M. pyrenaicus ref.) from Turicius ( = M. turicensis ref.) in Lower Miocene, Burdi- 
gahan, time is completely established. 



Zygolophodon pyrenaicus Lartet MS. (in Falconer, 1857), 

Lartet, 1859 

Figures 147, 148, PI. ii B, pp. 134-135 

Type: Middle Miocene, Ile-en-Dodon (Haute-Garonne), France. 
Referred: Lower Miocene, Sables del'Grleanais, France. 

Lartet's type of M. pyrenaicus was examined by Falconer in 
1857 and, owing to the presence of intermediate tubercles or conules 
in the anterior valley, was confused with his Trilophodon. Lartet's 
description, however (1859, p. 513), cited below, clearly distin- 
guishes this excellent type molar from that of Trilophodon angusti- 
dens and compares it with that of Mastodon tapiroides and with 
that of M. turicensis, which Lartet considered as identical species; 
by others M. pijrenaicus has been identified with M. turicensis, 
but Deperet remarks (1887, p. 131) that it is clearly distinguished. 
The distinction between M. turicensis and M. pyrenaicus is well 
displayed in figure 148. Also see Plates n and iir, pp. 134-135. 

M. {Triloph.) Pyrenaicus Lartet MS. (in Falconer, 1857). 
Falconer, "On the Species of Mastodon and Elephant occurring 
in the foss^il state in Great Britain." Quart. Journ. Geol. See. 
London, XIII, Pt. I, table opposite page 319. 

M. [Mastodon] pyrenaicus Lartet, 1859. "Sur la dentition des 
proboscidiens fossiles (Dinotherium, Mastodontes et Elephants) 
et sur la distribution geographique et stratigraphique de leurs 
debris en Europe," Bull. Soc. geol. France, (2), XVI, p. 513. 
Type. — A right third inferior molar, r.Mj. Horizon and 

Locality. — From near Ue-en-Dodon (Haute-Garonne), France. 
Middle Miocene. Type Figure. — Lartet, 1859, PI. xv, fig. 4. 

Type Description. — (Lartet in Falconer, 1857, opp. p. 319): 
"5. M. (Triloph.) Pyrenaicus (Lart.MSS.) . . . Upper Miocene. 
. . . France." (Lartet, 1859, p. 513): "Dernieremolaire inferieure 
droite de M. pyrenaicus des environs de I'llc-en-Dodon (Haute- 
Garonne) . Cette espece forme le passage du type a dents tapiroides 
au type k dents mamelonnees. La moitie ant<5rieure de la dent a ses 
vallons en partie interceptes par des tubercules intermediaires. 
Les vallons qui s6parent les deux collines subtapiroi'des de la partie 



posterieure restent entierement libres. Longueur de la couronne 
de cette dent qui provient d'un petit individu, 0,155; largeur antra 
les deuxieme et troisieme coUines, 0,070." 

Deperet, 1887, p. 131. — It was recognized as a valid species 
by Deperet (1887, p. 131) as follows: "Le Mastodon Pyrenaicus, 
Lart., se separe assez nettement du M. Turicensis, au moins a 
I'etat adulte, par la forme des mamelons de ses molaires, qui ont 
conserve sur la moitie anterieure de chaque dent la forme arrondie 




- FiC. i!'i9. — Dcrniurc arrieie-molaire inferieure de Mastodon pijrenaicus , 
a 1/2 grandeur. — Miocene moyen de I'lle-en-Dodon. (D'aprcs Lartet.) 

Type of Zygolophodo.v pvRENAicns. Middle Miocene 
Fig. 147. Type r.Ms of M. {Triloph.) pyrenaicus (Lartet MS., fide 
Falconer, 1S57, table opposite page 319); Lartet, 1S59, PI. xv, fig. 4. 
Middle Miocene. After Gaudry, 1878, p. 174, fig. 229. See figure 148, 
lower, of the present Memoir. 

des Mastodontes omnivores, tels que le M. angustidens. M. le 
Prof. Gaudry {Mammif. tertiaires, p. 175) est dispose a admettre 
que les dents de M. Pyrenaicus seraient 'des molaires de M. 
angustidens qui tendent a devenir des dents de M. Turicensis.'" 

Osborn, 1922: The type of M. pyrenaicus apparently belongs 
to an animal of larger size than Cuvier's type of M. tapiroides, 
inasmuch as Mj attains a length of 6K inches. The stage of evolu- 
tion is closely similar. In evolution, the three posterior lophs, 
with /our 'mamelons' on each loph, agree quite closely with the 



THE ZYGOLOPHODONTINiE: ZYGOLOPHODON 



207 



corresponding three posterior lophs in the Cuvier type of M. 
tapiroides, of Lower Miocene age. The type from Ile-en-Dodon 
(Haute-Garonne) was treated by Gaudry and Lartet as of Middle 
Miocene age. Deperet regards Ile-en-Dodon as of the same age 
asSimorre (of. Osborn, 1910.340, p. 2.58^our Fig. 143) and a Httle 
more recent than Sansan, Middle Miocene. The original of the 
wood engraving after Gaudry (Fig. 147) is reproduced below (Fig. 
148) directly after Lartet. 




(Upper) Tcricius turicensis simorrbnsis Osborn Type and 
(I.,ower) Zvqolophodon pyren.^icus Lartet Type. Compare FionRE 141. 

Fig. 148. Lartet's type, r.Mj, of Mastodon pyrcnaiais (Lartet "SiS., fiiU: 
Falconer, 1857, table opposite page 319; Lartet, 1859, PI. xv, fig. 4) com- 
pared with an r.Ms (PI. xv, fig. 3) erroneously determined by Lartet as 
M. lapiroidea, now the ty[je of Turiciiis turicensis aimorrensia Osborn, 1926. 
After Lartet 1859, 

Upper figure. {Op. cii., p. 513) : " Fig. 3. Derai&re molaire inf^rieure droite 
du M. tapiroides des environs de Simorre (Gersl. La hauteur des collinos 
d(5passe I'^paisseur de leur base. Les vallon.s qui leg separent sont en partie 
intercept<5s par une arete riScurrenle tul)crcul(^e dans le fond. La base interne 
de la couronne est entourijo d'un collet saillant. Longueur de la couronne, 
0,205; largeur entrelesdeuxiemeet troi.sieme collinos, 0,083." 

This molar was chosen by Osborn, 1926.706, p. 4, fig, 2, as the type of 
the subspecies Turiciwi turicenais simorrensis. See pa»es 219 and 220. 

Lower figure. (Lartet, 1859, p. 513): "Fig. 4. Derniere molaire inferieure 
droite do .1/. pyrenaicus des environs de I'lle-en-Dodon (Haute Garonne). 
Cette cspi'^ci' forme le passage du type k dents tapiroides au type k dents mame- 
lonnies. La raoiti^ antdrieure de la dent a ses vallons en partie intercept^'-s 
par des tubercules interraddiaires. Les vallons qui separent les deux collincs 
aiibtapiroides de la partie post^rieurerestententierementlibres. Longueur de la 
couronne de cette dent qui provienf d'vm petil individu 0,155; largeur entre 
les deuxii'mc et troisii^mo cnllines, 0,070," 

Zygolophodon pyrenaicus aurelianensis Osborn, 1926 
Lower Miocene, Sables de l'0rl(5anais, Chevilly (?), Ctissegrain, .\varay, 
France, No type figure available to the author. 

In 192() Osborn (1926.700, p. 2) proposed the subspecific 
name Zygolophodon pyrenaicus aurelianensis for this Lower 
Miocene stage, to include the abundant remains listed by Mayet 
• 1908, pp. 198, 199), as given below,' whicii in the present writer's 



opinion are undoubtedly specifically or subspecifically distinct 
from the Middle Miocene Zygolophodon pyrenaicus of Lartet. 

Zygolophodon pyrenaicus aurelianensis Osborn, 1920. "Addi- 
tional New Genera and Species of the Mastodontoid Proboscidea." 
Amer. Mus. Novitates, No. 238, p. 2. Cotypes. — (Op. 

cil., p. 2); "Superior tusks and superior and inferior molars 
designated by Mayet as follows (Mayet, 1908, pp. 198, 199): 
' Un fragment de defense superieure (Chevilly, sabliere Cassegrain, 
collection Nouel)est de coupe sensiblement ovalaire, sans la moindre 
trace de bande longitudinale d'email; . . . Une Mj superieure 
gauche (Avaray, mus6e d'Orl^ans) d^terminee par M. Lartet 
lui-meme, est k couronne tres large, supportant quatre collines 
dont les deux premieres ont leurs vallons partiellement intercept's 
par un tubercule accessoire; les deux collines posterieures, plus 
comprim(5es, presque tapiroides, sont s6par6es par des vallons 
libres; le talon post6rieur est en forme decrStecr'nel'e; . . . Une 
Mj superieure droite du masee d'Orl'ans. . . . provenant de 
Beaugency et determin6e par M. Lartet; . . . Une Mj inferieure 
(musee d'Orl'ans, sans localite indiqu6e) est a trois collines, avec, 
en arriere, un assez fort talon tubercule.'" Horizon and 

Locality. — Sables de I'Orleanais, Chevilly (?), Cassegrain, Avaray, 
France. Lower Miocene. 

Specific Characters.— (Osborn, 1926.706, p. 2): "Four 
conelets in each loph, intermediate tubercles in the valleys. The 
superior incisors are of oval section and lack all traces of the longi- 
tudinal enamel band," 

Zygolophodon borsoni Hays, 1834 

Figures 84, HI, 154, and PI, li, pp, 134-135 
Type: Middle Pliocene, from near Villanova. Province of Asti, Pied- 
mont, Italy. 

Referred: Upper Phocene of western Eurasia, 

Syn,: Mastodon veUavns .\yinard, 1847; Mastodon Vialetii Aymard, 
1847; Mastodon Buffonis Pomel, 1848; Mastodon affinis Jourdan, 1859. 
According to Lortet and Chantre, also Dep6ret and Mayet, the above species 
are chiefly synonyraoas with Mastodon borsoni Hays, The obscure type descrip- 
tion of M. horsoni by Hays was probably not accessible to these authors. 

Hays' definition of this verj' important Middle to Upper Plio- 
cene species, genotypic of the genus Zygolophodon, is consistent 
with his type figure (Fig. 154) also with Biiffon's figure (1778) 
reproduced in figure 111 of the present Memoir. Etienne Borson, 
to whom this species was dedicated, described the same tooth as 
Mastodon giganteum, but called attention to its similarity to the 
tooth figured by Buffon. Hays observes (1834, p. 334): "we feel 
persuaded that it does not appertain to that species \M. giganteum], 
the denticles having no longitudinal commissure," thereby allud- 
ing to the absence of the median sulcus or longitudinal fissure which 
separates the internal and external cones in Mastodon americanus. 

BoRsox, 1820-1823.— In 1820 ("Mcmoire sur des Michoires 
et des Dents du Mastodonte dit Mammouth, Trouvees Fossiles en 
Pidmont." Mem. Accad. Sci. Torino, XXIY) and again in 1823 
("Note sur des Dents du Grand Mastodonte Trouvees en Pi'mont 
et sur des Machoires et Dents Fossiles Prises dans la Mine de Houille 
de Cadibona Proche Savone." Mem. Accad. Sci. Torino, XXVII, 
p. 31) Professor fitienne Borson described and figured (Tav. ii) 
from near Villanova, Province of Asti, Piedmont, the portion of 
a jaw with grinders of a species "a dents etroite^" to which 
Hays (Trans. Amer. Phil. Soc, 1834, p. 334) refers as follows : 



'Mayet's description (1908, pp. 198, 199) of this Lower Miocene 'Mast.' pyrenaicus is without figure. The specific distinctions are: (1) Oval superior 
incisors with no trace of enamel band; (2) third superior molar bunolophodont, four ridge-crests, crown relatively broad, posterior talon crest crenulate. 



CONTRASTS IN MASTODON AND ZYGOLOPHODON MOLAR CROWNS 




^ 



Fig. 149. Crown view of inferior dentition and jaw of Mastodon americanus, 
containing Dpi, Mi, \U, and M3 anterior lophs. .\fter Warren, 1852, PI. v, one- 
sixth natural size. Reproduced for comparison with Zygolophodon borsoni. 




Fig. 151. Second and third inferior grinders, 1.M.2, M3, of Zygolophodon 
borsoni; enlargement of same teeth as shown in figure 150. One-third natural 
size. Oliserve 4-5-6 conelets in proto-, meta-, and tritolophids respectively. 




Fig. 152. Type of Mastodon Zaddachi Jentzsch, a second right inferior 
molar, r.Ms, lenjjfh 112 mm., breadth 78 mm., index 69.6. From Thorn, 
west Prussia. Regarded as a subspecies or geographic variety of Zygolopho- 
don borsoni, namely, Zygolophodon borsoni zaddachi. One-third natural size. 
After .Jentzsch, 1883, Taf. v, fig. 6a. 

Observe in Mastodon a median fissure, separating single inner and outer 
lobes. In Zygolophodon no median fissure; multiple conelets, sharp ridge- 
crests. 






Fig. 150. Inferior dcntiliuu and jaw u{ ZygulopliiMlou Itursoui from Vialette (Haule- 
Loire), France. After Lortct and Chantre, 1879, PI. xvi, figs. 1, la,onc-.sixth natural size. 
Reproduced for comparison with Mastodon americanus. 



Fig. 153. Third superior molar, l.M^ of 
Zygolophodon borsoni from "Le Petit-Rosey, 
vallon de Rochecardon, pres Lyon," France. 
After Lortet and Chantre, 1879, PI. xvi (bis), 
one-third natural size. Observe four main 
conelets and crenulated or plicated slopes of 
the lophs, suggesting those in Mastodon ameri- 
canus plicatus Osborn (Fig. 120), as a parallel 
character. 



208 



THE ZYGOLOPHODONTIN^: ZYCOLOPHODON 



209 



Mastodon' hohsom Hays, 1834. " In thu third volume of the 
Memorie della Heale Aocademia delle Scienze di Torino, professor 
Borson has described a tooth, found at ViUanova d'Astiea, in 
Piedmont, and whicli he ascribes to the [sMastodon] gignnUuin. 
After a careful examination of the description and drawing, in the 
work referred to, and of a cast of tlie tooth, in the Cabinet of tlic 
.Academy of Natural Sciences of this city, we feel persuaded that it 
does not appertain to that species, the denticules havinsr no lonjii- 




ImK- 154. 'I'vpi' r..\I-'' uf Manliiiliin Imrsoiii Hays, 1S:{4. one-tliiril 
iiiitural size. Orisiiially lijjiircil two-thirds natvinil .size Ijy Borson 
(1823, Vol. XXVII, Tav. ii) and referred by him to .1/. gigaiikum. 
Trom ViUuiiova d'A.stica, Piedmont, Italy. Cast No. i:{:{.52, Phila- 
delphia .\(adeiiiy of Natural Sciences; east .\mer. Mus. 19238, gift 
of the Geological Institute of Turin, 1926. Reversed to show natural 
position of molar. 

Lorlet and Chantre (I.S79, p. 305) observe that Mastodon borsoni 
was described for the first time in 1823 [1820] by I'abbe Borson, 
11820, Mem. Accad. Sci. de Turin, III, p. Hit);, and that Hays, 
who had in his pos.scssion a tooth (cast) found near .\sti, Italy, ai>- 
plicd to it the name of homnrti in honor of the abbe. 

tudinal commissure. It probably belongs to a new species, and we 
would suggest the propriety of dedicating it to professor Borson, 
under the name of M. Borsoni." 

Osborn, 1925: This is the sixth reference to the occurrence of 
a member of the Zygolophodontina", namely, of M. borsoni, in 
Europe. As to the ti/pc, Professor Horson's designation {op. cit., 
1823, p. 33) is as follow.*: " La dent qui a quelque rapport avec la 
notre est celle que M. de Buffon a figurde dans le tome V du supple- 
ment Jl I'histoire naturelle, pi. \,pag. 512, qu'il avait re5ue de M. 



de V'ergennes; avec cette difference ccpendant que dans la notre il 
n'y a que des vallees trans versales; les pointes 6tant unies en- 
semble dans la largeur, ne laissent aucun lieu a des separations et 
cons(3quemment aux vallees longitudinales." The above observation 
by Professor Borson is quite correct. Hays' type of M. Borsoni 
is obviously similar to the grindinsj tooth from "petite Tartarie" 
(Russia) described and figurctl by Buffon in 1778 and reproduced 
above in facsimile (Figs. 84, 111). 

(Jsborn, 1926: Aglanceat thetypeof Jl/. borsoxt (Fig. 154), also 
at Butfon's figures of 1778, iniiicates that Zygolophodon borsoni 
belongs to a phylum distinct from M. ainericanius : (1) The molars 
are brachyodont, bluntly crested or lophoid; (2) each loph consists 
of four principal 'mamelons' or mamillse compressed fore and 
aft; (3) as observed by Haj's, there are no connecting crests, 
conules, or spurs between the lophs; thus when well worn, each of 
the anterior lophs shows four loops, as if this animal were a 
descendant of the Z. pyrenaicus molar type, with its four 
'mamelons'; whereas the true M. antericanus phylum always 
exhibits two loops from the wear of the two 'mamelons,' i.e., 
cones. From these distinctive characters, we infer that Zygolo- 
phodon borsoni was a short-limbed, forest-living, browsing type, 
with long upcurved tusks, distantly resembling M. ainericanus. 
Confirmation that it spread all over Europe and northwestern 
.\.sia, as originally observed by ButToii, is fully .set forth by Pavlow 
(Chap. VI, pp. 133-135 above). Hays' description of this species 
was api^arently unknown to Aymard, to Pomel, and to Jentzsch, 
who jjroposcd the four following species: 

'Mastodon' vellavus Aymard, 1847, i'. 414. — "M. 
.Aymard signale ensuite un Mastodonte surpassant de % le 
Mastodon maxinius de I'Ohio, d'apres un quatrieme metatarsien, 
long de 0,255, large en haul de 0,110, en bas de 0,145, et ayant 
0,095 dans son moindre diametre. II projjose pour cette espece le 
nom de Mastodon rellavus." 

'Mastodon' vialetii Aymaku, IS47,p. 415. — " Un troisieme 
Mastodonte, plus petit de Jj que le |)rec(jdent, caracterise par un 
metatarsien long de 0,095, large de 0,060 en haut, et 0,070 en bas, 
au milieu de 0,049, et par des molaires larges de 0,074, a coUines 
comme dans celui de I'Ohio; il devra se nommer Mastodon \'ialdii." 

'Mastodon' Buffonis Pomel, 1848, p. 258. — "Un mastodon 
a dents tapiroTdes accompagne presque toujours le M. Cwieri; 
c'est le type des .1/. tapiroides Cuv. : mais il en est un autre qui se 
trouve au contraire en Auvergne et en Italie avec Vangustidens, 
dont les dents sont beaucoup plus epai.^ses et plus courtes, plus 
grosses en tot alite, et qui nous parait d;'voir constituer une espece 
distincte. Les dents de Sib^rie, attribuees au grand mastodonte, 
devraient lui etre rapport ees. On poinrait hii donner le nom de M . 
liuJTonis." 

'Mastodon' Zadd.\chi .Jentzsch, 1883, p. 202. — 'Der Zahn 
hat drei scharf ausgepriigte Joche und gehiirt somit in die Gruppe 
des M. tapiroides, M. Borsoni und M. Ohioticus, welche Vacek 
trelTend als Untergattuag Zygolophodon zusammengefasst hat." 
See figure 152. 

Zygolophodon V.vcek, 1877. pp. 6 11, Taf. vl — Vacek gave 
a thorough review of the successive descriptions and characteriza- 
tions of this species, including those of Hays, de Btainville, Lartet, 
Kaup, Gaudry, von Meyer, and Jourdan; throughout he uses the 
term 'Mastodon' tapiroides as equivalent to 'M.' turicensis. 



210 



OSBORN: THE PROBOSCIDEA 



He records as characters of M. horsoni specifically distinct 
ivom Mastodon americanus the. following (after Lartet, 1859): (1) 
The greater proportional width of the grinders, the lesser height of 
the crests, the lesser prominence of the ' trefoil spurs ' (' aretes 
recurrentes'); (2) the clear distinctions of M. horsoni from 
'Mastodon' tapiroides { = M. turicensis), despite von Meyer's 
declaration (1867) that M. horsoni is inseparable from M. turicen- 
sis; (3) the absence in M. horsoni of 'trefoil spurs' ('aretes recur- 
rentes') which distinguish all species of M. turicensis (Lartet, 1859, 
Gastaldi, 1861) ; (4) the broad, short, and brachyodont crowns of 
M. horsoni as compared with the more prominent crests of M. 
tapiroides; (5) the ahsence of alveoli for lower incisors in the sym- 
physis of the lower jaw, as observed by Jourdan and corroborated 
by Brandt (1860); (6) the exceptionally short symphysis of M. 
horsoni and the elevation of the coronoid process, as observed in 
the collections in the Munich and Lyons Museums, presenting a 
resemblance to the lower jaw of Elephas. 

GEOGRAPHIC DISTRIBUTION' 

Austria (Vacek, 1877, pp. 9-11). — Beginning with the type of 
Mastodon horsoni from the [Middle) Pliocene of Asti, Vacek 
traces the distribution of M. horsoni and of M. tapiroides ( = turi- 
censis) throughout Austria. He observes that the remains in the 
Upper Phocene marine sands of Neudorf ' agree in the closest detail 
with the type teeth of M. horsoni of Asti; that the Lower Pliocene 
locality of Baltavar lacks M. horsoni but contains M. turicensis 
{i.e., tapiroides) and exhibits the closest relationship to the fauna 
of Pikermi, although the specimen from Baltavar appears to 
approach the Middle Pliocene type of M. horsoni more closely 
than the 'M.' turicensis described by Gaudry from Pikermi. 
From the illustrations in Vacek's Memoir, e.g., Taf. vi, figs. 3, 3a, 
the molar from the marine sands of Neudorf does not correspond 
with Hays' definition and type of the species M. horsoni: also 
the posterior molar given in the same plate (Taf. vi, figs. 1, la) 
presents a prominent median longitudinal sulcus or fissure, which 
by Hays' definition is absent in the type of M. horsoni. This 
molar, Ms, from the sands of Theresiopel presents resemblances to 
the Mastodon americanus f. praetypica [ = Pliomastodon americanus 
praetypica of the present Memoir]. See Plate i, pp. 134r-135. 

Hungary, Levantine, Schlesinger, 1922, p. 231. — 
Schlesinger observes (1922, p. 231) that 'Mastodon' horsoni occurs 
in Rakos, a locality near Budapest, which yields also a large number 
of 'M.'arnernensis and of 'M.' americanus f. praetypica (op. cit., 
p. 232) : " Die Fundstelle ist der ofters erwahnte und nach seinen 
geologisch-stratigraphischen Verhaltnissen auf S. 213 eingehend 
besprochene Schotter (Mastodonschotter). Sein levan tines Alter 
steht vollkommen fest. . . . Obwohl M. Borsoni nur von wenigen 
stratigraphisch sicheren Fundstellen bekannt geworden ist, weisen 
diese doch so untriiglich immer wieder auf den einen Horizont, auf 
den wir schon zwei Mastodonarten beschriinkt gesehen haben, das 
ich kein Bedenken trage, die Art als ein fur das osterreichisch- 
ungarische Gebiet rein levantines Tier zu erklaren. . . . AUer- 
dings ist die Bereicherung, die unsere Beweisfiihrung fiir das aus- 
schliesslich levantine Alter der Spozies in Osterreich-Ungarn und 
den unmittelbar angrenzenden Landern insbesondere vonseiten 
der rumanischen Funde erfahren hat, sehr erheblich. Wir konnen 



nunmehr mit grosserer Sicherheit diese stratigraphische Fest- 
stellung wiederholen und sehen M. Borsoni als ein in jeder Hinsicht 
sehr brauchbares Leitfossil des Levantins nur bestatigt." 

Clear separation of 'Mastodon' bor.soni from 'M.' 
americanus praetypica. — (Schlesinger, 1922, p. 227) : " Wohl der 
entscheidendste und auch fiir den StJratigraphen wichtigste Schritt 
in der Erforschung der Zusammenhange der jiingeren Zygodonten 
ist die Loslosung der beiden bisher stets in bunter Mischung als 
M. Borsoni beschriebenen Formen voneinander und die Klarung 




\ 



'^■^.A 



Fig. 157 



Fig. 155 




Fig. 156 

Fig. 155. Referred superior molar, I.M', of Zygolophodon bor- 
soni, from the Pliocene marine sands of Neudorf, Austria. After 
Vacek, 1877, Taf. vi, figs. 3, 3a, one-third natural size. [ = Plio- 
mastodon praetypica ref. See PI. i E, pp. 134-135.] 

Fig. 156. Referred right inferior grinder, r.Ms, of Zygolophodon 
borsoni from Hidv^g, Komitat Haromszek, Hungary. After Schle- 
singer, 1922, Taf. XVII, fig. 2, one-third natural size. Observe 
simple four conelets on each loph; four and a half lophs in all. 

Fig. 157. Referred anterior half of third inferior molar, r.Ma, 
of Zygolophodon borsoni (erroneously attributed to ?North America), 
one-third natural size. After Schlesinger, 1922, Taf. xx, fig. 3. 
Observe four chief conelets on each crest. 

(Osborn, 1934) Both Schlesinger and Vacek refer to Zygolophodon 
molars which prove to belong to Pliomastodon (see PI. i, pp. 134-135). 

der phylogenetischen Stellung beider zueinander, zur Stammart 
und zum echten M. americanus. Es ist sehr erfreulich, dass uns 
von den beiden fraglichen Arten etliche Stiicke aus sehr zuver- 
lassigen Fundorten unseres Untersuchungsgebietes bekannt gewor- 
den sind. Ajn&csko (Kom. Gomor) : Aus den Sanden und Schottern 
dieses Fundpunktes stammen die prachtvoUen Reste, die ich auf 
S. 101— 105, Taf. XVI, Fig. 1 u. 2, Taf. xvii, Fig. 1 und Taf. xviii, 
Fig. 5 dieser Arbeit bekanntgemacht habe. Es sind die sichersten 
Belege fiir die nahen Beziehungen dieses Tieres zum M. 
americanus." 



'The Upper Pliocene marine sands of Neudorf, also Theresiopel, Nikolsdorf, and Baltavdr (Vacek, 1877, Taf. vi, figs. l-5a,), yield a superb series of 
upper and lower grinding teeth which probably belong to Pliomastodon, although described and figured by Vacek as M. [ = Zygolophodon] borsoni. See 
alao specific revision in the phylogenctic appendix of the present Volume I. 



THE ZYGOLOPHODONTIN^: ZVGOLOPHODON 



211 



Distinctions of Zygoi-ophodon borsoni from Turicius 
Characters OF 'M.' horsoni. — Sclilesiiiger (1922, pp. 79-81, 
also pp. 116-123) describes arid figures numerous examples of the 
true Zygolophodun borsoni from Austria and Hungary, attributing 
this species to Upper Pliocene age. He distinguishes ' Mastodon ' 
borsoni from 'M.' tapiroides (i.e., Turicius turicensis) by the 
following six characters {op. cit., p. 79) : 

"Fur M. Borsoni wurden nachfolgende Merkmale geltend 
gemacht: 1. Vollkommen geradegestreckte, etwas nach aussen 
gerichtete schmelzbandlose Stosszalme im Oberkiefer. 2. Vor- 
handensein einer mehr oder weniger funktionellen Mandibel- 
schaufcl mit zwei kleinen unteren Stosszahnen. 3. Im Zusam- 
menhange damit Vorhandensein einer gegeniiber M. tapiroides 
wenig reduzierten Symphyse. 4. Fehlen von Pramolaren [i.e., 
P2-41. 5. Jochformeln: m|^|: [Dp]^lDp4]f: M|^: [M 1]?, 
(M 2] 37^: [M 3] ~^. 6. Charaktere der'echten Moiaren: 



o) Breite Kronen mit mehr oder weniger hohen Jochen, engeren 
Tiilern als M. tapiroides und breiten Jochgraten: 6) mediane Tren- 
nungslinien der Jochhiilften in schwankender Ausbildung, bis- 
weilen sehr kraftig und tiefeingeschnitten, bisweilen schwach; 
c) aufifallend starke Entwicklung der Nebenhiigel an den pratriten 
Halbjochen; b) schwankende Ausbildung der Sperrleisten an den 
pratriten Teilen." This valuable summary is as follows: 

1) Superior incisive tusks straight, without enamel band. 

2) Two small inferior incisive tusks, subfunctional. 

3) Symphysis more reduced than in Turicius. 

4-5) Permanent premolars lacking, intermediate molars 
(Dp 4-M 2) trilophodont. Third molare tetra- 
lophodont. 
6) Molars as described by Hays and Vacek; feeble 
'trefoil spurs.' 



Grinding Tebth Evolution 

Plate I B-L 

Miomaslodon , Flioniastodon , 

Mastodon 
No median conules. 
Conek'ts not exceeding 4. 

Persistent longitudinal sulcus. 

Ridge-crests incomplete. 

Trefoil spurs late in develop- 
ment {Plioniastodon, Mastodon). 

Pleistocene 
Mastodon acutidens (Indiana) 
.Mastodon americanus 
(Kansas) 

Plio-Pleistocene 
Mastodon pavlowi (Russia) 

Pliocene 
Pliomastodon praetypica 
(Hungary) 

Lower Miocene 

Miomastodon depereti 
(C'heviUy) 

.Miomaslodon americanus tap- 
iroides (see Fig. 98 D, Dl 
of Memoir), Hungary 



in the Mastodon, Zygolophodon, Tuhicius, and Stegolophodon 

Plate II B-G Plate III A-J 

Zygolophodon Turicius 



No median conules. 
C'onelets 4 progressive to 6 on 
each ridge-crest. 

Longitudinal sulcus irregular. 

Ridge-crests separated by 
sulcus. 

No trefoil spurs. 

Pliocene 
Zygolophodon bor.wni 

(France) 
Zygolophodon borsoni 

(Siberia) 
Zygolophodon borsoni type 

(Piedmont, Italy) 



Middle Miocene 

Zygolophodon pyrenaicus 
(France) 



No median conules. 
Conelets 4-7 9-25. 

Longitudinal sulcus vestigial 
or absent. 

Ridge-crests sharply trans- 
verse and elevated. 

Trefoil spurs present or ab- 
sent. 

Pliocene 
Turicius virgatidens (Fulda) 
Turicius atticus (Pikermi) 
Turicius wahlheiynensis 
(Rheinhcssen) 



Phyla (Pls. i-iv, pp. 134-135) 

Plate IV A-I 
Stegolophodon 

No media,!! conules. 
Conelets bunoid, 4-6. 



Longitudinal sulcus dis- 
appearing. 

Ridge-crests 4 -6. Progres- 
sively arched, post-concave. 

No trefoil spurs. 

Pliocene 
S. stegodontoides type, India 
S. latidens type, Burma 
S. cauileyi (Perim Island) 
S. sublatidens, Austria 



Miocene 

Turicius turicensis (Bavaria) ; 

D, ditto (Elgg, type); E, 

ditto (Elgg, ref.) 
Turicius tapiroides, France 



Middle Miocene 

S. cautleyi progressus, India 

S. nathotensis, India 

Collateral ancestor of Stego- 
don, fully described in Chapter 
XV 




ZYGOLOPHODON BORSONI 3000mm.,9'l0" e MASTODON AMERICANUS 3105(nni.J0'2/*" 

ITALY N. EURASIA N AMERICA 

Fig. 158. Comparison (1932) of Zygolophodon borsoni (I.EhT) a.nd of Mastodon americanus (Right), one si.ktv-second natural size 
(I-eft) Schlcsingor (1922, pp. 79-81, 116 123) .irscribcd llu- Zygolophodon borsoni oi Austria-IIungarv as of Upper Pliocene age. Cranial and bodv out- 
Imes largely conjectural, probably similar to MasUxion amencanus but less woolly. Estimated height 300() mm., 9 ft. 10 in. 
(Right) Chiefly restored from the Warren Mastodon of Newburgh (see Fig. 124). 



7. CHARACTERS OF TURICIUS AND ITS INCLUDED SPECIES AND MUTATIONS, 
AS COMPARED WITH ZYGOLOPHODON AND SERRIDENTINUS 

Family: .MASTODONTID.E Girard, 1852 

Subfamily: Zygolophodontin^ Osborn, 1923; compare Serridentin^ Osborn, 1921, also Stegolophodon- 
TiNiE subfam. nov. 

The species of animals described below were named Mastodon tapiroides by all early authors and more recently 
by Vacek and by Schlesinger. The same species are named M. turicensis by Schinz, von Meyer, Gaudry, Mayet, 
Schlosser, Cope, and other recent authors, except Schlesinger. Originally described from near Zurich, Switzer- 
land, species of this genus were widely distributed in Europe and are even mentioned as occurring in Spain 
(Pavlow, 1894, p. 36) ; they are not recorded from Asia, although Zijgolophodon is frequently recorded from 
northern Asia. 

Genus: TURICIUS Osborn, 1926 

Original reference: Amer. Mus. Novitates, No. 238, p. 3 (Osborn, 1926.706). 
Genotypic species: Mastodon turicensis Schinz, 1824. 

Generic Characters.— (Osborn, 1926.706, p. 3): "Transverse crests, lophs, with five conelets pro- 
gre.ssing to eight or nine conelets.'" Tubercles or conules retrogressive in the valleys between the crests. 
Median longitudinal sulcus vestigial in the early stages, completely disappearing in the progressive stages. 
Progressively strong 'trefoil spurs,' ['Wulstkanten,' 'aretes'] on the superior internal cones and on 
the inferior external cones. Conelets increasingly connate at the summit, rising into sharp, subhypso- 
dont transverse crests. Superior incisors oval in section, with sharply defined enamel band except in 
Mastodon [ = Turicius] virgatidens. Inferior incisors without enamel, straight, suboval in section, under- 
going progressive reduction. Symphysis of jaw progressively pointed and reduced in length, horizontal, 
unlike Serridentinus, in which the jaw remains elongate. Postero-inferior molars with four well- 
developed crests, the tetartolophid slowly progressive, the pentalophid rudimentary, but progressive 
in the higher stages. Gradual repression of the premolar dental succession, as in Mastodon americanus 
and in Zygolophodon borsoni." 

The generic name Turicius and the specific name turicensis are derived from Turicum, the Latin name of 
Zurich, near which historic city the type specimen was discovered. 

The grounds on which this genus is separated from Zygolophodon are recited above. Turicius in contrast 
with Zygolophodon is relatively well represented in all horizons from the Lower Miocene Faluns de Pontlevoy 
formation of France to the Lower Pliocene of Pikermi, Greece, and Baltavar, Hungary. Every museum in 
Europe contains a number of upper and lower grinding teeth, referable to various stages in the progressive 
evolution of Turicius, which are uniformly catalogued either as Mastodon tapiroides or as M. turicensis. 

In a closer study of the Turicius tapiroides of the Lower Miocene we observe several primitive characters, 
and in ascending geologic levels we further observe a steady progression not only in size but in the stronger 
development of the characters which distinguish these species throughout from species of the true Zygolophodon. 
Vacek (1877) and Schlesinger distinguish these species clearly from Mastodon [ = Zygolophodon] borsoni. Con- 
sequently there is no real resemblance between the final Pliocene stage in this generic phylum {Turicius 
virgatidens) and the final Pliocene stage of the Ztjgolophodon phylum {Zygolophodon borsoni). 

'To twenty-fivf coiiclits in certain .■ipocimoiis (sec PI. m, pp. 134 13.')). 

212 



THE ZYGOLOPHODONTIN.E: TURICIUS 213 

Whereas Zygolophodon is convergent or parallel with the true Mastodon americanus phylum, and in certain 
characters with the Stegodonts, the progressive molar stages of Turicius are similar in one character to those 
of Serridentinus, in other characters to Stegodon or to Stegolophodon. The upper tusks of Turicius are apparently 
distinguished by the persistence of the enamel band, while the lower tusks may readily be distinguished from those 
of Serridentinus by their suboval section and rodhke form, also by their gradual degeneration. All the specific 
stages referable to Turicius, however, require very careful reexamination before these generic distinctions can 
be positively determined. Meanwhile the generic characters now known to us are probably as above (pp. 200, 
212) and below (p. 217). 



Conspectus and Measurements of the Species of Turicius in Ascending Geologic Order 

See Plate hi 

The distinctive characters shown in the following series of geologic ascending species will doubtless prove of 
full specific value, but awaiting further examination, the trinomial nomenclature may be retained so as to connect 
these forms with the previous literature. In the succeeding pages of this chapter (VII) more detailed observations 
of von Meyer, of Gaudry, of Mayet, of Schlcsinger, and of others are quoted or abstracted. 

5. Middle (?) Pliocene. Turicius virgatidexs von Meyer, 1867. 

'Oclhon Loliine' bei Fulda, rjcrniany. Mucli nioro progressive than Turicius ntticHK: lophs sharply pointed, broader 
aiul shorter superior incisive tusks, devoid of enamel hand. 
4. Lower Pliocene. Turiciu.s atticus Wagner, 1857. 

Pikernii type (see Wagner, 18.^7, Tali. vii. fifi. Ifi). Left P', ap. 30 mm., tr. 26 mm. Left M' Irilophodont, ap. 
54 mm., tr. 4") mm., proportionate in size to the l)p4 of Ciaudrv's ' MoMndon' lurirr?h'<iK. 
3. Upper Miocene. Turicius turicensis Schinz, LS24. Type Stage. 

ElgR, Kapfnach, Schinz. 1824, von Meyer. 1867. Type 1.M-, ap. 94e mm., referred r.-Mj, ap. 92 mm., six ' nebenhiigeln ' 
or conelets in each loph ; median sulcus vestigial; strong 'trefoil spur' on second and third lophs. Superior incisors 
broadly oval to flattened, with sharply defined enamel band covering one-third of circumference. Small inferioi- incisors, 
without (?) enamel. 
2. Upper Middle Miocene. Turicius turicensis simorrensis Osborn, 1926. 

Simorre (Gers), France (see Lartef, 18.59 — Figs. 148, 163 of present Memoir). M3, up. 20.5 mm., tr. 83 nun. Strong ex- 
ternal trefoil crest; strong internal eingulum; tetartoloiihid well developed with four conelets; pentalophid rudimentary 
with two conelets. A premolar succession (Lartet). 
1. Lower Miocene. Turicius tapiroides. Referred. 

Faluns de Pontlevoy (see figure by Mayet, 1908 — Fig. 139 of present Memoir). M.i, ap. 147 mm., t r. 66 nun. Renmants 
of median sulcus; five to eight irregular conelets in each loph : no tubercles or conules in intermediate valleys. 

The i)hylogeny, conelet and ridge-crest evolution of Turicius, from T. tapiroides to T. virgatidens, is clearly 
shown in PI. iii, pp. 134-135. 



Comparative Measurements 
(Notes, W. D. Matthew, September, 1920; References Modified by Osborn, 1925) 

Molars referred to Turicius turicensis are represented in the Paris Museum by four specimens from Simorre 
(typical geologic level of Trilophodon angustidens) , one from Sansan (Middle Miocene), and one from Pontlevoy 
(Lower Miocene), from which the extreme and average indices may be taken. The following measurements and 
indices were taken by Doctor Matthew, the specific references corresponding with labels in the Museum d'Hi.stoire 
Naturelle, Paris, 1920. 



M' 






M= 






M' 






Ms 




ap. tr. 


I. 


ap. 


tr. 


I. 


ap. 
143.5 


tr. 
80.2 


I. 

56 


ap. 
203 


tr. 
86 


I. 
42 


80.7 51 


63 


113.8 


67 


59 


130 


73 


56 


155 


66 


43 


81 56.1 


69 























M' 






M; 


ap. 


tr. 


I. 


ap. 


tr. 


144e 


91 


63e 







214 OSBORN: THE PROBOSCIDEA 

Lower, Middle, and Upper Miocene 

1872, Lartet, Simorre, 'Mastodon' turicen- 

sis ref. 
1784, Simorre, ' Mastodon ' turicensis ref. 
1838, Simorre, ' Mastodon ' turicensis ref. 
1862, Simorre, 'Mastodon ' turicensis ref. 
1791, Sansan, ' Mastodon' turicensis ret. 

1782, Miocene, 'Mastodon' turicensis rei. 121 70 58 

l897,Pont\evoy,' Mastodon' turicensis rei. 147 66.7 45 

Measurement.s of Type Molars 

M' M2 

ap. tr. I. ap. tr. I. 

1867, Type of 'Mastodon' virgatidens von 

Meyer, 'Gelben Lehme'beiFulda 112 89 79 

[1857, Type of Turicius atticus Wagner, 

Pikermi] (Dp^ ap. 30 mm., tr. 26 mm., M', ap. 54 mm., tr. 45 mm.) 

1824, Type oi Turicius turicensis, Schinz, (M-)94e 

Elgg (M2)92 

1926, Type of Turicius turicensis simor- 

rensis Osborn, Simorre 205 83 40 

1824, Type of Turicius tapiroides Cuvier, 

Calcaire de Montabusard 85e 

The above tables of type and referred grinding teeth partly demonstrate a progressive increase in size of the 
grinders when viewed in ascending geologic order. This is doubtless accompanied by retrogressive abbreviation 
of the mandibular symphysis correlated with reduction of the inferior incisors. 

Observations on Specimens from various Geologic Levels, Lower Miocene to Lower Pliocene, 

REFERRED TO 'MaSTODON' TURICENSIS OR TO 'MaSTODON' TAPIROIDES 

According to von Meyer, de Blainville, and Suess, as cited by von Meyer (1867, pp. 48-61), the M. turicensis 
lophodont type of mastodont is widely distributed over Europe — France, Austria, Greece (Pikermi). Its abun- 
dance in the lignites, swamp formations, of Elgg, Kapfnach, and Eibiswald in Styria, points to its being an in- 
habitant of dense and partly swampy forests, although it occasionally occurs under more open conditions of Ufe 
associated with Trilophodon angustidens and Trilophodon (Choerolophodon) pentelicus. 

Osborn, 1922: Following the name Mastodon turicense given by Schinz (1824, p. 278), and the type figure of 
the Elgg tooth with description by Schinz (1833, pp. 58, 59), there came the full description (partly cited above) 
of this animal by von Meyer in 1867 (1867, pp. 48-61) with admirable figures of referred specimens (Taf. ii and v, 
figs. 1-7), teeth associated with an Upper Miocene fauna in the lignites of Elgg and Kapfnach, beUeved to be of 
the same age as Saint Gaudens and somewhat more recent than that of Simorre which contains the type of Tri- 
lophodon angustidens. 

All the intermediate molars of Turicius figured by Gaudry (1862, PL xxiv) and by von Meyer (1867, Pis. 
II and V, figs. 1-7) are trilophodont with a tendency to form elevated transverse crests composed of numerous 
confluent mamillse or conelets grouped into outer and inner lobes. 

The jaw of the Pikermi stage figured by Gaudry (op. cit., PI. xxiv, fig. 2 — Fig. 165 of the present Memoir) 
shows a reduced and pointed symphysis with two vestigial incisors ; it is totally unlike the symphysis of Mastodon 
americanus. 



thl: ZYGOLOPHODONTIN^: turicius 




©-© Types oF sixteen species o'T Z(/golopfiodon and TurLcitLS. 
@ Turicius tapiroidLes referred. Preferred. H.F.0.1954. 



Fig. 159. Type localities (1-16) and theoretic migration lines of the forest- and swamp-loving Zygolophodon and Turicius. The crosses represent 
referred specimens, note especially Zygolophodon borsoni of Siberia and ZygolophodonC!) of Japan (fide Matsumoto); (o) represents Turicius tapiroides ref. of 
North Africa. 



The third inferior molar of 'M. ' turicensis is described by von Meyer (1867, p. 50) as tetralophodont: .... 
"sondern um einen letzten unteren Backenzahn von M. Turicensis handelt. Es scheinen vier Querreihen iiber- 
liefert, von der vierten freilich nur wenig. Die Querkiimme sind erst schwach abgenutzt. Der Zahn erinnert 
auffallend an die typischen Zahne von Elgg." The lower incisors {op. cit., Taf. v, fig. 6) are short, suboval in 
diameter, not dissimilar in form to the lower incisors of Mastodon americanus, totally dissimilar to those of 
Trilophodon angustidens. 

In the Pikermi type of 'M.' turicensis described by Gaudry (1862, p. 155), the lower incisors are still more 
reduced. The superior incisive tusks of the Elgg stage (von Meyer, 1867, Taf. ii, figs. 4, 7, 8, and Taf. v, figs. 6, 7) 
exhibit a sharply defined enamel band covering about one-third the circumference of the tooth, which is broadly 
oval to flattened, not circular Uke that of M. americanus, also of relatively small size. (Von Meyer, pp. 54, 55): 
"Nach diesen von drei bis vier Individuen herriihrenden Resten besass Mastodon Turicensis einen im Vergleich 
zu anderen Species auffallend kleinen Stosszahn. . . . Vergleicht man das, was ich flir die obern Stosszahne von 
Elgg hervorgehoben habe, mit meinen Angaben liber die oberen Stosszahne von Mastodon angustidens ... so 
wird man finden, dass in letzterer Species die oberen Zahne, den unteren entsprechend, gegen die in M. Turicensis 
hauptsachlich durch uberwiegende Grosse auffallen." 



216 OSBORN: THE PROBOSCIDEA 

Typical Uppeu Miocene Turicius turicensis Schinz of Switzerland 

Switzerland (Matthew, Notes, September, 1920. Turicius Osborn, 1926). — From the Upper Miocene 
lignites of Kiipfnach (near Zurich) in the Zurich Museum are a jaw and lower tusks resembUng Turicius turicensis 
rather than Mastodon { = Trilophodon] angtistidens to which they are referred in the label. From the Upper Miocene 
of Kapfnach also are specimens which exhibit only a moderate elongation of the jaw, resembhng that of Serriden- 
tinus prodticius, but with no decurvation of the symphysis. 

.\midst the interesting fauna of Elgg, Canton Zurich, in lignites of Miocene age, also occur specimens refer- 
able to Turicius turicensis. Among the Sansan teeth in the Paris Museum is one marked " M. turicensis"; a 
somewhat worn and broken M2 which may possibly be related to Turicius tapiroides ; a small last upper molar, 
M^, with only four crests, may also belong to this species. 

Munich Museum (Matthew, Notes, September, 1920). — From the Upper Miocene of Flinz in the Munich 
Museum is a trilophodont specimen consisting of lower jaws, upper tusks and teeth, with parts of skeleton asso- 
ciated: this animal is larger than the typical Trilophodon angustidens and the teeth are relatively broader; the 
tusks are arranged as in Serridentinus productus ref. and »S. serridens type from the Lower PUocene of Texas, 
although relatively larger, more massive, the symphysis heavier but not decurved; there is no twist in the upper 
tusks which are larger than those in the American species S. productus and S. serridens; these tusks curve down- 
wards, the enamel is external, the cross-section is oval; good fibulae and a femur are associated, also larger part 
of the manus and pes. This is much too large and broad-toothed an animal to be properly referred to Trilophodon 
angtistidens. Of the grinders, M^-Mo are trilophodont; M^-Mj tetralophodont, the tetartoloph being somewhat 
narrower; 'trefoil spurs' weak; grinding teeth relatively broad. 

Osborn, 1925: The above observations of Matthew in the Zurich and Munich museums unquestionably 
relate these specimens to the genus Turicius Osborn and serve to distinguish species of this genus from .species of 
the genus Serridentinus Osborn by the symphyseal characters of the lower jaw, namely: In the European Turicius 
the symphysis of the lower jaw is straight, not decurved ; in the American Serridentinus the symphysis of the lower 
jaw is decurved and contains two large tusks of flattened oval section. 

Turicius in Austria and Hungary 

Von Meyer (1867, pp. 48-61) describes in great detail the widely distributed 'Mastodon' turicensis of Austria, 
also the Upper Miocene Ugnitic forms from the typical locahties of Elgg, of Winterthur, and of Kapfnach. He 
describes and figures two small inferior tusks, laterally compressed, 110 nmi. in length, vertically oval, diameter 
ap. 27 mm., tr. 19 mm. (op. cit., Taf. v, figs. 6 and 7); these teeth are apparently larger than those of the 
Lower PUocene stage of 'M. ' turicensis described by Gaudry. He further describes referred superior incisive teeth 
(Taf. II, figs. 2 and 3; Taf. v, figs. 1-5) as relatively large, vertically oval, diameter ap. 37 mm., tr. 28 mm., with 
fluted enamel band on the external surface (Taf. 11, figs. 7 and 8) from the typical locality of Elgg. 

Throughout Schlesinger's invaluable Memoir of 1922 all the specimens referable to the genus Turicius are 
described by him as Mastodon tapiroides. He observes, however, that in six characteristics these animals appear 
to be clearly distinguishable from M. angustidens on the one hand and from M. borsoni on the other (op. cit., 
pp. 77, 78). C onsequently the characters by which Schlesinger distinguishes lus M. tapiroides are cited below as 
distinctive of the genus Turicius Osborn. 



THP: ZYGOLOPHODONTIN/E: turicius 



217 



(Generic and Specific Characters of Turicius. Osborn after Schlesinger, 1922, pp. 77, 78). — Den jungeren Zygo- 
donton gegeniibpi- ist M. tnpiroiden als Spozins weitaus arti siclicrstpii unterschicden und durcli folgendo Merkmale gekenn- 
zcic'hnot : I. Die ohpren Inzisoren sind nacli aufwiirts gckruiiimt, iiii (^iiorschnitt oval und tragcii oin wohlentwirkeltes 
Scliniolzhaiid an dc>r konvcxen Sritc. 2. Die untcrcn Inzisoren sind bis zii einom gpwissen Grade dencn des -U. fingu.y:tidens 
alinlicli, docli klcinor und hilden voroint koinrn Spaten, sondern einen an der Spitze schief von oben vorne nacli unten hinten 
al)gcslutztcn Stociiiippai:i( zuni Wtililcii in dor lOrde. 3. Die Unterkieferschaufel war den Inzisoren entsprechend und nacli 
Ubergangsfornien zu sclilicssen— spatclfcuniig ausgezogen, docli kiirzer als die des M. angu.stklens. 4. Sowohl im Ober-, wie ini 
Unterkiefer trug die .\rt jo einon Pratnolaren (pm^) (vgl. L.-V. 27, S. 149 und l.iC, Taf. xxi, Abb. 1 und 6). 5. An den droi 
Milch- und den drei echton Molaron koliren stets nachfolgonde Jocliformeln wiodcr: m [Dp] -J-^f : |Dp 2) f; [Dp 3] fS.'f^-'P'^l I- 
A/|^f : [M 1] §; [M 2| ^l\ [M 3] I't}.. li. Die Molarcn sind in nachgenannten Punkten von denen der jungeren Fornien scharf 
uiitorschiedoii: a)Schnialo Krone init verhiiltnismassig niodrigon .lochon, woiton Talorii untl ongon .Ifx'ligraten f.Jochkainnien) : 
/)) auffalloiid starko, ticfoingoschnitlono niodiane Tronnungsliiii(! dor Jochhalfton ; o schwacho I'',iit wicklung dor Xebonliiigol 
an don pratriten .lochloilon; d) stets bedeutende Betonung der pratriten Sperrleislen, die bisweilen ganz cnorin entfaltet sein 
kiinnon. 

Schlesinger concludes {op. cil., p. 78) that the enamel band of the superior incisors and the presence of 
permanent as well as milk premolars clearlj' distinguish these animals from the more recent Zygolophodonts 
("von den jungeren Zygodonten"). 



SYSTEMATIC DESCRIPTION OF 

Turicius tapiroides Cuvier, 1806, 1821-1824; 
Desniarest, 1820-1822 
I'iKiircs 77, HiO, Kil, iilso PI. in A, B, C, pp. 134-13.'> 
I.oHcr Miocene. Ciilcuiir de .Montahusiiril, Franee. Defay (1783) \va.s 
the first to record ' .U. ' UipirnuUs from this horizon (cf. Gaudry, 1862, p. l.")3). 

This Lower Miocene stage is unfortunately imperfectly known 
because of the uncertainty regarding the characters of the type 
specimen. It is a very important stage owing to its e.xtreme Lower 
Miocene age, since the Calcaire de Montabusard is regarded by 
Mayet as uiiderli/ing the typical Sables do I'Orloanais; at the time 
iMiropo had a warm and dry climate and broadly depressed laiul 
areas. In the Sables de I'Orloanais occur grinding teeth referred 
to M(i>:ti)don ptirennirux Lartet, with /<)«r conolets on each loph, 
wliidi may correspond in duiracter with ("uvior's lost type of .1/. 
tapiroides (see Zygolophodon pyrenaicus aurelianensis, p. 207.) 

Petit maslodonic Cuvior, 1806. "Sur DiffOrentes Dents du 
Genre des Ma.stodontes, mais d'especes moindres que celles de 
I'Ohio, trouvces en plusieurs lioux des deux continens." Ann. Mus. 
d'Hist. Nat., VIII, p. 411. Maslodonic Inpinnde Cuvier, 1821. 
"Reciierches sur les Ossemens Fossiles," Nouvelle Edition (2d 
edition), I, p. 268. Maxtodon tapiroides (in Desmarest, "Mam- 
malogie ou Description des Especes de Mammiferes," 1820-1822, 
p. 386); Cuvier, op. cit., 1821 1824, V, Pt. 2, p. 527. 

Type. — Considered by von Meyer (1867, p. 48) as a premolnr, 
i.e., milk tooth, Dp 4 ('Milciibackenzahn'); also by Gaudry 
(1862, p. 153) as a " molaire de lait." Probably an imperfect third 
right inferior molar, r.Ms, of diminutive siz<\ 

Horizon and Locality. — Calcaire de Montabusard, France; 
Lower Miocene. (Osborn, 1910.346, p. 250, and our Fig. 146): 
"The Miocene is said [by Mayet (1908, p. 314)] to be introduced 
... in the [lacustrine] mlaiiie de Monlnhmiiid, which records the 
return of the horses {Anckilherium). The overlying sables de 
I'Orlmmiis ]containing the remains of Trilophodon angnslidcn.i] 
are fluviatilo sands, laid down in successive stages which contain 
the entire rich Lower Miocono fauna, parallel with which are widely 
scattered deposits in France, (iorniany, Austria, Switzerland, and 
far to the southwest in I'ortugal near Lisbon. In this Lower 
Runiigalian stage the flora of .\uvorgiio (Myrica, Cinnamomum , 
Li'g!(i'da»i6ar) indicates a warm and drj- climate. . . . In the strata 



SPECIES OF TURICIUS. PLATE III 

of the basin of Mayence the Oligocene flora, Sabal, Sequoia, Lauras, 
Cinnamomum, also persists. . . . The marine cetacean of the 
period is Squalodon harieiisis." 

Type Figure.— Cuvier, 1806.3, PI. 68 (in), fig. 6. New type 
figure published by Mayet reproduced in figure 161 of the present 
Memoir. 



MuLAK DE.SCR1BEU 
Bv Def.\y (1783), BY 
CoviER (1806) 




FiK. 160. 'I'ype r.Mj. 
/'(•((( mnstoJonte Ciiv., 
1806, PI. 68 (ill), fig. 6. 
from the Calcaire de 
Montabusard, = Masto- 
rlonle Ininroiilr Cuv.. 
1821, subsequently 
written Mnslodon tapi- 
roides (in De.smarest. 
1820-1822, p. 386, in 
Cuvier, 1.S21-1824. V, 
Pt. 2, p. .527), =Tiiri^ 
ciits tapiroides in the 
pre.sent Memoir. This 
figure is one-half nat- 
ural size. 

Compare PI. in, pp. 134 
•son with T. Inpiroules ref. (C) 




Fig. 161. New type figure after .\layel, 
1908, p. 19.5, fig. 66, natural size. 

(Mayet, op. cil., p. 195): "O-lte dent 
est dvidemment tr6s petite et pout parfaite- 
ment ^tre regardfe comme une dent de hiit 
-Mais la figure de Cuvier, pour imparfaite 
qu'elle soit, montre suffisamment les rar- 
:iet«res es-sentiels du .U. tapiroides, mo- 
laires A eoUinc^ ('leviSes, plut6l er^nelc^s 
que divi.S(?es en manielons, s(^par(^s par de 
larges valleys, .sans manielons inter- 
mi^diaires, p<iur qu'il fdt possible A la 
rigueur de consorver le nom sp^cifique dr 
M. tapiroides donnd par Cuvier." 

135, type of Turicius tapiroides (A) in eompari- 



GUETTARD, (FIDE CuVIER, 1821, I, P. 267, AND 1834, II, 

P. 371).— The figures of Guettard, referred to by Cuvier, and the 
new type figure of Mayet (Fig. 161 of the present Memoir) relate 
this grinder more closely to Turicius luricensis than to Zygolo- 
phodon pyrennirwi. 



218 



OSBORN: THE PROBOSCIDEA 



Type Description.— (Cuvier, 1806.3, p. 411): "La dent de 
Monlabusard, pi. in, fig. 6, correspond si bien a celle de Saxe pour 
sa largeur, que je ne doute pas que ce ne soit un germe de I'une 
des post^rieures de la m^me espece, casse en avant." 

(Cuvier, 1821, 1, p. 268).— " Ainsi, independamment du grand 
mastodonte de I'Ohio, et du mastodonte a dents etroites, especes 
aujourd'hui bien connues et parfaitement determinees, nous 
trouvons des indices de quatre mastodontes, qui paroissent former 
d'autres especes. Les deux qui viennent d'Am^rique pourront 
s'appeler, lorsque leurs caracteres seront entierement confirmes, 
mastodonte des cordilieres et mastodonte humboldien. Je donnerai 
au premier de ceux d'Europe le nom de petit mastodonte, et au 
second, dont les collines ne sont pas completement divisees en 
mamelons, celiii de mastodonte tapiroide." 

(Von Meyer, 1867, p. 48) : "Cuvier's Mastodon tapiroides 
(Oss. foss., 4e. ed. II. p. 371. t. 28. f. 6) beruht einzig und allein auf 
einem noch dazu unvoUstandigen Milchbackenzahn, der mit noch 
zwei anderen Zahnen in dem tertiaren Lacuster-Gebilde von 
Montabuzard bei Orleans in Frankreich gefunden wurde." Cuvier 
in the Fourth Edition of his "Ossemens Fossiles" (1834-1836), Vol. 
II, p. 371, states: "J'en donne la figure r^duite a moiti6, pi. 28, 
fig. 6. C'est la meme qui a ete gravfe dans les Memoires de Guet- 
tard, tome vi, x'' Mem., pi. vii, fig. 4." [An error, cf. Gaudry, 1862, 
p. 153, footnote 4.) 

Osborn, 193.5 : It is very important to note that the type of M. 
tapiroides from the Calcaire de Montabusard, Lower Miocene, is 
possibly an imperfect tetralophodont third inferior molar with the 







anterior loph broken off; each loph in Cuvier's figure is composed 
of four to six transversely placed 'mamelons' (cf. PI. in A) in 
contrast with the type of M. americanus which has two ' mamelons' 
in each transverse loph; thus a complete M. tapiroides tooth 
would be comparable to a complete tooth of the M. pyrenaicus 
Lartet type which has four 'mamelons' on lophs 2, 3, and 4. 

Typical Lower Miocene Turicius tapiroides 
Gaudry's (1862, pp. 157, 158) Mastodon tapiroides 
('turicensis') of Pontlevoy. — (1) Superior milk molar (Dp^) 
less enlarged than in the Pikermi stage, distinguished by an anterior 
conelet; (2) third superior milk molar (Dp') with conelet on each 
side of the median valley, not observed in the Pikermi form; (3) 
fourth superior milk molar (Dp*) one-third smaller than the Pikermi 
form. (4) Third superior molar with feeble third crest (tritoloph), 
whereas the Pikermi molar exhibits an almost distinct third crest. 
The above specific characters of the Lower Miocene, or Burdi- 
galian, stage were regarded by Gaudry possibly as individual differ- 
ences; with our present knowledge of the established fact, the 
Lower Miocene grinders of ' Mastodon ' tapiroides of Pontlevoy are 
much more primitive than those of the Lower Pliocene Pikermi 
stage {Mastodon [ = Turicius] atticus Wagner) described by Wag- 
ner and by Gaudry. 

Mayet, 1908. — Mayet (1908, p. 194) observes: "Maiscomme 
le remarque P. Gervais (Zool. et Pal. franc, 1859, p. 68) il ne parait 
pas bien certain que ce morceau de dent vienne bien r^ellement du 
calcaire d'eau douce de Montabuzard; il provient beaucoup phis 




Fig. 162. Genotype and Homceotype of Turicius turicensis 
See PI. Ill D, E, pp. 134-135 



(Right). Upper Miocene of Elgg. Referred Mastodon turicensis, 
second inferior molar, r.M2. After von Meyer, 1867, Taf. ii, fig. 1. 
Sub-natural size (op. cil., p. 52): "Die Krone ergiebt 0,104 Lange, in der 
hinteren oder dritten Querreihe 0,070 grosste Breite, in der vorderen nur 
0,058." Winterthur Collection. 

The Schinz type l.M- (left) of the species " Afas<. Juricense " becomes 
the genotype of the genus Turicius Osborn, 1926. 



(Left). Type l.M^ of Mastodon turicense 
Schinz, 1824. After Schinz, 1833, Taf. i, fig. 1. 
Compare von Meyer, 1832.2, p. 72: "Mastodon 
Turicense. Schinz. Schinz, Mnspt." Subsequently 
written Mastodon turicensis (see von Meyer, 1839, 
p. 2). Original figure natural size; reproduced 
herewith one-half natural size. Zurich Collection. 
Elgg, Switzerland. 

Observe in l.M' six and a half conelets in protoloph, si.x and a half conelets in metaloph, and nine conelets in tritoloph. 
Observe in r.Mj eight to nine conelets in protolophid and strong postero-external trefoil spur, six conelets in metalophid and 
strong postero-external trefoil spur, six subequal conelets in tritolophid, no postero-external trefoil spur, strong posterior 
cingulum. Von Meyer's referred specimen (right) is certainly from the type locality of Elgg and represents the Upper 
Miocene Elgg stage of progression in the cones and conelets. Altogether there are twenty-two more or less distinct conelets 
in the three lophs of the second inferior molar. Compare these very distinct conelets with the closely connate, sharply 
pointed, and multiple conelets in the corresponding tooth, r.Mz, of the type of 'M. ' virgatidens (Fig. 168). 



THE ZYGOLOPHODONTIN.^: TUIIICIUS 



219 



vraisemblablement des sables de Chevilly ou d'Ingr6. D'un autre 
c6t6, plusieurs auteurs pensent que Cuvier aurait crd6 le type M. 
tapiroides sur des molaires de lait de M. angustidens [Footnote: 
'Cf. Lortet et Chantre, Rccherches sur les Mastodontes (Archives 
du Museum d'histoire naturelle de Lyon, t. II, 1879, p. 308).']. Je 
n'ai pu retrouver soit h Paris, soit k Orleans, les pieces (5tudiecs par 
Cuvier et toute la discussion ne peut porter que sur celle qu'il a 
figur6e, r^duite de moitid [Footnote: 'Ossements fossiles, t.I, 
pi. in, fig. 6. ') et que je rcproduis ramende a sa grandour naturelle. 
Cette dent est 6videmment tr&s petite et peut parfaitement etre 
regardde comme une dent de lait. Mais la figure de Cuvier, pour 
imparfaite qu'clle soit, inontre suffisamment les caractferes es- 
sentiels du M. tapiroides, molaires a collincs 61ev6e8, plutot cr6ne- 
16e8 que divisdes en mamelons, s4pardes par de larges valines, 
sans mamelons intermddiaires, pour qu'il filt possible h la rigueur 
de conservcr le noin spdcifique de M. tapiroides donnd par Cuvier." 

Turicius turicensis Schinz, 1824 
Figure 1G2, also PI. in D-G 
rp[)er Miocene, Tortonian .slage. Lignites of Elgg, Canton Zurich, 
Switzerland, also lignites of Kiipfnach; age of Saint Gaudcns. 

The type description by Schinz of this Upper Miocene species 
is very meagre, but our knowledge of this stage is greatly amplified 
by the figures and characterizations of von Meyer. As the geno- 
typic species of Turicius it is very important to observe closely the 
characters of the grinding teeth. The grinding teeth occur in rela- 
tive abundance in the lignitic formations of Switzerland and of 
northern Italy, in some instances accompanied by the jaws, portions 
of the cranium, and the superior and inferior incisive tusks. 

Mastodon turicense Schinz, 1824. " Naturgesehichte und 
Abbildungen der Saugethiere," 1824, p. 278. Type. — Second 

superior molar of the left side, l.M- (see Fig. 162, left). 

Geologic Horizon and Locality. — Elgg, Canton Zurich, 
Switzerland; Upper Miocene, Tortonian stage. (See Osborn, 
1910.346, p. 263): "This concluding phase of the . . . Miocene is, 
according to Depdret [Footnote: 'Depdret, L'dvolutiondesMam- 
mif^res tertiaircs (Miocene). C. R. Acad. Sci. Paris, Vol. CXLIII, 
Dec. 24, 1906, p. 1122.'], typified [in France] by the fauna of St. 
Gaudens [Fig. 143, 48] on the head waters of the Garonne, which 
are paralleled (48-69) by the freshwater dcpo.sits of (Eningen (54), 
near Zurich, the lignites of Elgg (55) and Kiipffnach (56); in the 
same region, the mammals of Giinsburg (11) and Statzling (60) in 
Bavaria, the lignitic fauna of Monte Bamboli (63) in Tuscany, and 
the deposits of San Isidro (64), near Madrid." 

Type Figure.— Schinz, 1833, Taf. i, fig. l,also PI. iii D. 

Type Description. — Schinz (1824, p. 278) gives a long de- 
tailed account of the various species of Mastodon of North and 
South America concluding with two lines on the Mastodon of 
Zurich, as follows: " Zurcherisches, Mast, turicense, Ziihne 
diescr wahrsrheinlich neuen Art wurden den Elgg im Canton 
Zurich in Rraunkohiengruhon gefundon." 

Osborn, 1925: The following description and the accompany- 
ing figure (Fig. 162) reproduced from von Meyer (1867, Taf. ii, 
fig. 1) arc very important, because they constitute tlie first full 
definition and characterization of this species. Observe that each 
crest or loph, as in M. americanus (ohioticus), consists of two main 
lobes, which in M. americanus are simple or slightly bilobed at the 



summit, but in 'M.' turicensis are subdivided into three or more 
'Nebenhiigeln,' 'mamelons,' or 'conelets,' six to nine in all. Von 
Meyer regards Cuvier's type of 'M.' tapiroides as a milk tooth, 
a statement which requires verification by examination of the type. 
Observe also the spur (' Wulstkante') on the anterior and posterior 
slopes of the main internal lol)e. Von Meyer rejects the direct 
relationship of 'M.' turicensis either to M. americanus or to 'M.' 
tapiroides. 

(Von Meyer, 1867, p. 48) : "III. Mastodon Turicensis Schinz. 
Mastodon Turicensis gehort zu den Trilophodonten. Er unter- 
scheidet sich von dem dersell^en Gruppe angehorigen Mantodon 
angustidens hauptsachlich dadurch, dass die Krone seiner Backen- 
zahne einfacher und dass die Stossziihne kleiner sind. Die Quer- 
reihen der Backenzahne stellen Querkamme, durch tief ein- 
geschnittene Querthaler getrennt, dar; die Querthaler werden 
nicht durch Ncbenhiigel oder Hiibel versperrt. Die Querkamme 
sind nicht einfach, . . . Die Querkamme zerfallen in eine 
deutlich zu unterscheidende aussere und in eine innere Halfte, 
deren jede aus zwei, drei, selten aus mehr Hiigeln oder Hiibeln, 
die in der Richtung der Querreihe mehr oder weniger verschmolzen 
sind, zusammengesetzt ist. Der aussere und der innere Haupt- 
hiigel eines Querkammes sind die starksten, und von ihnen zieht 
vorn und hinten cine Wulstkante herunter. . . . Mastodon 
Turicensis gleicht noch am meisten M. Ohioticus, dessen Quer- 
reihen aber deutlicher in zwei Halften getrennt erscheinen und 
deren Abnutzungsfliichen mehr zum Rhombischen hinneigen. 
Durch die Aohnlichkeit der Zaliiic des Miustodon Turicensis mit 
dem Nordamerikanischen M. Ohioticus hat man sich verleiten 
lassen, das Vorkommen letztercr Species auch fiir Europa anzu- 
nehmen. Man glaubte ferner, dass die Zahne von M. Turicensis 
zu denen gehorten, welche Cuvier unter M. tapiroides begrifT. 
Beides ist jedoch falsch. Cuvier's Mastodon tapiroides (Oss. 
foss., 4e. 6d. II. p. 371. t. 28. f. 6) beruht einzig und allcin auf einem 
noch dazu unvoUstiindigcn Milchbackenzahn, der mit noch zwei 
anderen Ziihnon in dem tertiaren Lacuster-Gebilde von Monta- 
buzard bei Orleans in Frankreich gefunden wurde." 

Turicius turicensis simorrensis Osborn, 1926 
Figures 141, 148, 163 

Tapper Middle Miocene. From near Siniorre (Gersl, France. 

This subspecific stage was based by Oslwrn (1926.706, p. 3) 
on a third right inferior molar, r.Mj, which Lartet in 1859, PI. xv, 
fig. 3, erroncouslj' determined as belonging to Mastodon tapiroides. 
The name Turicius turicensis simorrensis has been assigned to it, 
in reference to the localitj' in which the specimen was found. The 
distinctive characters of this molar tooth are readily seen by an 
examination of figure 148, where it is shown in comparison with the 
type right third inferior molar of Mastodon [ = Zygolophodon] 
pyrenaicus Lartet (see also Fig. 163 below). 

Turicin.<! turicensis simorrensis Osborn, 1926. ".\dditional 
New Genera and Species of the Mastodontoid Proboscidea." 
Amer. Mus. Novitates, No. 238, p. 3. Type.— (Op. cit., 

1926.706, p. 3) : "A third right inferior molar, r.M,." Hori- 

zon AND IvOCALiTY. — Upper Middle Mioccne. From near Simorre 
(Gers), France. Type Fiqdre.— Op. cit., 1926.706, p. 4, 

fig. 2. 

Specific Characters.— (Osborn, 1926.706, p. 3): "Strong 



220 



OSBORN: THE PROBOSCIDEA 



external trefoil crest; strong internal cingulum; tetartolophid well 
developed with four conelets; pentalophid rudimentary with two 
conelets. A premolar succession (Lartet). Anteroposterior meas- 
urement 205 mm., transverse measurement 83 mm." 




Type of Turicios turicensis simobrensis of the 
Upper Middle Miocene 

Fig. 163. Type third inferior molar of the right side, r. Ms, of Turiciua 
luricensis sirrwrrensis Osborn, 1926, after Osborn, 1926.706, p. 4, fig 2, 
one-half natural size. FromSimorre (Gers), France. Erroneously deter- 
mined by Lartet as Mastodon lapiroides. Compare Lartet, 1859, PI. XV, 
fig. 3 i"M. tapiroides"); also p. .513: "Derniere molaire infSrieure 
droite du M. tapiroides des environs de Simorre (Gers). La hauteur de," 
collines depasse I'^paisseur de levir base. Les vallons qui les apparent sont 
en partie intercept's par une arete recurrente tubercul^e dans le fond. 
La base interne de la couronne est entouree d'un collet saillant. Long- 
ueur de la couronne, 0,205; largeur entre les deuxieme et troisiemo 
collines, 0,083." 

It is noteworthy that this specimen from the upper Middle Mio- 
cene of Simorre contains five to six 'mamelons' on the three anterior 
lophs and is more progressively lophndont than the type of Zygolophodon 
pyrenaicus attributed to the Middle Miocene of Ile-en-Dodon. The 
wood engraving of this specimen after Gaudry is reproduced above in 
figure 141 of the present Memoir. 

Turicius atticus Wagner, 1857 
Figure 164, also PI. in, I 

Typical Lower Pliocene of Pikermi, Greece 

Compare 'Mastodon' turicensis described by Gaudry, 1862 (pp. 218, 220, 
and 221 of present Memoir). 

This species rests upon the teeth described and figured by 
Wagner (1857) as Mastodon atticus, a specific name which has 
erroneously been regarded as a synonym of Mastodon pentelicus 
Gaudry, 1856. Restudy of Wagner's figure and description makes 
it apparent that this animal is entirely distinct from the buno- 
ma.stodont Trilophndon (Choerolophodon) pentelicus and that it 
belongs to the zygolophodont phylum of Turicius. Consequently 
Osborn (1926) revives Wagner's specific name as appropriate for 
the [type] specimen of the Lower Pliocene stage of Greece, namely, 
Turicius atticus. 

Mastodon atticus G. & L., Wagner, 1857. "Neue Beitrage zur 
Kenntniss der fossilen Saugthier-Ueberreste von Pikermi." Abh. 
bayer. Akad. Wiss., Band VIII, CI. II, Abth. I, p. 140. [Type].— 
Left fourth superior premolar and first molar, l.P^, M'. Hori- 
zon AND Locality. — Lower Pliocene, Pikermi, Greece. [Type 
Figure].— Wagner, 1857, Tab. vii, fig. 10. 

[Type] Description. — From Wagner's description of this 
specimen we may extract the following (Wagner, 1857, pp. 140 - 
142): "Es ist nur ein kleines Bruchstiick vom Vorderende eines 



linken Oberkiefers (Tab. 5 [Tab. vii] Fig. 16), was sich zugleich mit 
seinen beiden ersten Zahnen erhalten hat. Der erste Backemahn 
hat eine ovale, nach vorn stark verschmalerte, etwas stumpf 
dreieckige Form, deren breitester Theil hinterwarts liegt; durch die 
Abreibung hat er noch nicht sonderlich gelitten. . . . Der zweite 
Backemahn ist weit grosser als der erste, viel langer als breit, und 
ninmit nach hinten an Breite zu. Er ist durch zwei tiefe Quer- 
furchen in drei Querhiigel abgetheilt, die durch eine seichte Kerbe 
in der Mitte abermals gesondert sind. ... In meinen friiheren 




Type of Turicius atticus 
Fig. 164. Left fourth superior premolar and first molar, l.P*, M', 
described by Wagner. 1857. pp. 140-142, as Mastodon atticus G. & L. 
Reproduced after Wagner's original figure. Tab. vii, fig. 16, reduced 
to two-thirds natural size. (Wagner, op. cii., p. 1.58): "Vorderende 
voni linken Oberkiefer des Mastodon atticus mit den lieiden ersten 
Backenziihnen." 

Left P*, measuring ap. 30 mm., tr. 26 mm., is functional as in 
Serridentiniis. The three crested or trilophodont M', measuring ap 
,54 mm., tr. 45 mm., is proportionate in size to the Dp4of Gaudry 's 
' Majitodon' turicensis (Fig. 165). 

Erorterungen dieser Ueberreste habe ich es nicht gewagt, sie einer 
bestimmten Spezies von Mastodon zuzuweisen. Gaudry und 
Lartet haben sie jetzt als Mastodon atticus bezeichnet; ihre aus- 
fiihrUche Abhandlung, die noch nicht erschienen ist, wird jeden- 
falls die Griinde angeben, warum sie dieselbe von M. angustidens 
getrennt haben." 

'Mastodon' turicensis of Pikermi. — The milk dentition 
probably belonging to the same species Turicius atticus was fully 
described by Gaudry, 1862, under the caption Mastodon turi- 
censis. Gaudry's Pikermi specimen (1862, pp. 152-159, PI. 
XXIV, figs. 1—4), fully described and figured as Mastodon turicensis, 
exhibits the following ten characters; (1) Ridge-crests: Dp 2i 
Dp 31 Dp 4f; consequentl}' the intermediate molars, Dp 4- 
M 2, are trilophodont; (2) the conelets are closely connate, re- 
sembling those in 'Mastodon' virgalidens von Meyer rather more 
closely than the clearly separated conelets of the type of 'M.' turi- 
censis Schinz; (3) the superior tusks are flattened oval, a layer 
of enamel covering the superior and inferior surfaces; Lartet 
describes an enamel band on the convex surface of the 'M'. 
tapiroides of Sansan; (4) the first superior milk tooth (?Dp^) bears 
four irregular cones and the rudiment of a talon; (5) the second 
superior milk molar (Dp') bears two subacute crests with partly di- 
vided cones and two cones on the talon; (6) the third superior milk 
molar (Dp^) is composed of three distinct crests surmounted with 
numerous conelets, with a trace of a trefoil spur ("une trace 
d'arete") on the inner side. 



IHi: ZYGOLOPHODONTIN.t:: TL'RICIUS 



221 





Fig. 165. Siipcricir :iiid infciiui ii.ilk piciiiolais. Dp'"*, I)|)2^, of ' Mastodon' turicensis ref., from 
Pikcrmi. After Ciaudiv. 1SC2, PI. .\.\i\,figs. I, 2. and 4, one-third natural size. The same specimen 
i.s now referred by Osborn to the species Turiciiis atiicus ( = Mastodon alticii.s of Wagner, 1857). 

(Left lower figure) Inferior milk premolar, Dp4, enlarged from lig. I (.side view) and Fig. '2 
(erown view), two-third.-: natural .size. 



LowKR J.wv. — (7) Symphysis abbreviated, reealling that of 
M. ainericaiius, its upper faee deeply holluw, the rami broad ami 
shallow; (8) small inferior milk incisois (Di-^) similar to those of 
.V. americaiiuf;, probably deciduous; (9) first inferior molar (Dp.) 
primitive, elongateti, laterally compressed, composed of a promi- 
nent cone and two posterior conelets; second inferior molar (Dpa) 
with double crests, third inferior molar (Dpi) with three very 
distinct crests and external trefoil spur {arete) on each crest : 
(10) no trace of vertical succession by permanent molars, again 
resembling M . americanun. The above progressive and retrogres- 
sive diaracters of the milk teeth serve to separate Turicius alticus 
very positively from T. turicensis (lypicus) and geologically preceti- 
ing stages. 

Turicius virgatidens von Meyer, 1867 

Figure. Hi7, Uis. al>o PI. ill H, HI H3 
Ty|»e: Middle(?) Pliocene, 'Gelben Lehme' bei Fulda, northeast of Frank- 
fort, (iennany Referred: .Middle Pliocene. Murin.sel, Croatia, and Laaerberg, 
Uistria 

Mastodon vtrg(ilkle)i.s von Meyer, 1807. "Studien iiber das 
Genus Mastodon." Palaeontographica, Band XVII, 1867 1S70, 
pp. 61-63. Type (fide von Meyer). — A right third lower 

molar, r.Ms, a second right lower molar, r.Mj, and a second left 
upper molar, l.M'. Horizon .\ni) Locality.— 'Gellx'ti 

Lehme' bei Fulda, northeast of Frankfort, Germany; Middle(?) 
Pliocene. Type Figure.— Op. cit.. Taf. iv, fies. 1-5. 




■S^S:i 



tig. Kit). Coiu\>j.K Turicius virgatidens (Fig. 168j. .\ lower 
molar, I.M2, from Murin.sel, Croatia, referred by Vaeek to .Va»- 
lodon iapiroides [ = Tnricius I ut ictna ij>], but which appears to 
be closer to the .stage Turicius rirgatidens. .\fter \'acek, IS77, 
Taf. VII, figs. 4,4a, one-third natural size. Same as figure 1 40 




Fig. 167. Anterior aspect of protoloph of Turicius. after Schlesinger, 1913, 
fig. 2, p. 718, one-half natural size. 

(Left) Turicius virgatidtns ref.. Middle Pliocene, Terrasse voni Laaer- 
berg, .\ustria. 

(Riglit) Referred Turicius (?) tapiroides. Middle Miocene, Faluns de 
la Touraine, France, Helv6tien. 



222 



OSBORN: THE PROBOSCIDEA 




Type of Turicius vikgatiue.ns 

Fig. 168. Type molars of Mastodon virgalidcns von Meyer, 1867, Taf. iv, figs. 1-5, one-half natural size. From 'Gelben 
Lehme' bei Fulda, northeast of Frankfort, Germany. 

Taf. IV, fig. 1 (op. cit., p. 62): "Letzter Backenzahn der rechten Unterkieferhalfte." Taf. iv, figs. 2, 3: "Vorletzter 
Backenzahn der rechten Unterkieferhalfte." Taf. iv, figs. 4, 5 (op cit., p. 63): "Vorletzter Backenzahn aus der linken 
Oberkieferhiilfte. ' ' 

Fig. 1. Third inferior molar of the right side, r.Mj; figs. 2 and 3, second inferior molar of the right side, r.Mz; 
figs. 4 and 5, second superior molar of the left side, I.M-. 



This final stage of Turicius seems to be rare; besides the type 
from Fulda, Germany, we observe one tooth from Murinsel, 
Croatia (Fig. 166), which may be referable to the same species. 
Schlesinger (1913, p. 716) also described from the locality of Laaer- 
berg, " MittelpUocan, Terrasse vom Laaerberg," a large tooth which 
he regards as transitional between Zygulophodon tapiroides Cuv. 
and Z. horsoni Hays and which (p. 720) he likens to Tetrabelodon 
virgatidens (of. Abb. 1, p. 716, Abb. 2, p. 718, and Abb. 3, p. 720); 
this Laaerberg animal, which Osborn is inclined to regard as re- 
sembling Turicius virgatidens more closely than the typical Zijgo- 
lophodon horsoni, appears to confirm the Middle Pliocene age of 
T. virgatidens. 

Geologic Age. — The uppermost Miocene age attributed by 
Sandberger and von Meyer probably corresponds with the Middle(?) 
Pliocene age according to more recent correlation; certainly Turi- 
cius virgatidens is much more progressive than the Turicius atticus 
of the Pikermi stage. 

The geologic age of this very important species is not determi- 
nable from von Meyer's description. The characters of the grind- 
ing teeth appear to represent a stage considerably more progressive 
than that from the Lower Pliocene of Pikermi, namely, Turicius 
aUicus. Vacek (1877, p. 7) observes: " Die Reste desselben stam- 
men aus einem gelben Lehme bei Fulda, . . . den aber in neuerer 
Zeit Prof. Sandberger in seinem schonen Werke iiber die Siiss- 
wasser-Conchilien (synchr. Taf.) als oberstes Miocan anfuhrt." 



Vacek (op. cit., pp. 7 and 8) cites von Meyer's characteriza- 
tion of this species and compares it closely with Mastodon horsoni 
and M. tapiroides, i.e., ' M' . turicensis, but fails to perceive the funda- 
mental distinctions in the character of the ridge-crests and in the 
multiplication of the conelets, especially in M^-M2 (Fig. 168, 2, 4), 
which clearly distinguish the species 'Mastodon' virgatidens from 
any less progressive species in the Turicius phylum. 

Type Description. — Von Meyer characterizes these teeth as 
follows (von Meyer, op. cit., 1867, p. 61): "Die Backenzahne sind 
von ausgezeichneter Schonheit; die beiden vorletzten sind drei- 
reihig, einen Trilophodonten verrathend, dessen Zahne sich nur 
denen in Mastodon Turicensis vergleichen lassen. Sie sind aber 
womoglich von noch einfacherer Bildung und etwas grosser als 
die von Elgg, namentlich verhaltnissmassig breiter; die Quer- 
reihen bestehen nicht aus Hiibeln oder zitzenformigen Theilen, sie 
gleichen mehr Querkammen und sind eher noch scharfer und 
flacher, die Querthaler den Kammen entsprechend tiefer und 
scharfer eingeschnitten und noch freier, und die Halften der Quer- 
kamme deutlicher getrennt als in M. Turicensis. . . . Die Wulst- 
kanten der Haupthiigel sind zwar deutlich aber eher schwacher, 
und zwischen denen der Aussen- und Innenseite besteht kein so 
deutlicher Unterschied in Starke als in M. Turicensis. Die durch 
das Ineinandergreifen der gegenstandigen Kronen bedingte Ab- 
nutzung ist daher mehr mit Zuscharfung der Kamme verbunden, 
als dass kleeblattformige oder ovale Kauflachen entstanden. Bei 



THE ZYGOLOPHODONTIN.E: TURICIUS 223 

den oberen Backenzahnen verlaufen die Haupthijgel der Innen- oval section, entirely devoid of enamel coating or enamel band, 

seite, bei den unteren Backenzahnen die Haupthtigol der Aussen- greatly exceeding in size the tusk referred to M. turicensis—aW of 

seite wulststreifig gegen ihre Basis hin, was wohl bei langerem which leads von Meyer {op. cil., p. 63) to conclude: "Ich habe 

Gebrauche der Zahne an Deutlichkeit verloren haben wiirde, aber daher diese Ziihne unter dem Namen Mastodon virgatidens be- 

doch entschieden dazu beitragt, diesen Ziihnen ein eigenthiimliches griffen." 
Gepriige zu geben. 

Osborn, 1925: Von Meyer rightly describes these beautifully Turicius wahlheimensis Klahn, 1922 

preserved types as more sharply lophodont than the M. turicensis Figure 229, also Pi. m .1 

type, the 'wulstkanten' ( = 'mamelon') being sharply compressed; From Walilheim and Esselborn, Rheinhessen, Germany. Pliocene, 

the teeth are broader and shorter than those of M. turicensis; For brief description and figure, see page 282, fig. 229, of 

the superior tusk ('stosszahn') is described as 5M feet long, of the present Memoir. 



8. GEOGRAPHIC DISTRIBUTION AND EXTINCTION OF SPECIES OF TURICIUS 

Compare Pl. hi A-J, pp. 134-135 

As fully described in the previous pages, Turicius occurs successively in the Lower Miocene of France, age of 
tlie Sables de I'Orl^anais, Calcaire de Montabusard, and Faluns de Pontlevoy, clearly enumerated by Mayet 
(1908). It occurs in the Middle Miocene of Sansan and Simorre, as described by Lartet, and of Styria, as described 
by Schlesinger, also in the Upper Miocene hgnites of Switzerland, as described by Schinz, von Meyer, and other 
authors. In all these Miocene stages the cones and conelets are more or less separate, rounded, and distinct at the 
summit. In the Lower PHocene stage {Turicius atticus), fully described and figured by Wagner as 'Mastodon' 
atticus and by Gaudry as 'Mastodon ' turicensis, the conelets become connate at the summit. This sharply con- 
nate character is still more accented in the Turicius virgatidens of von Meyer; the type tooth is sharply zygodont, 
which reminds one of some of the Pliocene stages of Stegodon, although the conelets (Fig. 168, 2) are much smaller 
and also less numerous than in the progressive Stegodon of India. Turicius apparently survived until the Middle 
Pliocene (Phocene ancien), but no trace of species of this Turicius phylum has been found in Upper Pliocene beds 
containing remains of Z. borsoni. 

Distinction from Serridentinus. — These zygolophodont progressive tendencies in Turicius and the devel- 
opment of a pure transverse crest, to which the name Zygolophodon so clearly appUes, is quite distinct from the 
development of the corresponding conelets and crests in the American Serridentinus in which the cones and conelets 
do not become connate and remain quite distinct at the summit. The second distinction of Serridentinus is the 
progressive strengthening of the 'trefoil spurs,' which become more and more prominent and conspicuous in 
progressive stages, surmounted with small conelets, whereas in progressive stages of Turicius, like T. atticus and 
T. virgatidens, the 'trefoil spurs' are reduced to fine ridges accented with numerous small conelets. Thus while 
there is a strong resemblance in the Miocene stages between species of Serrideyitinus and species of Turicius, 
there is very little resemblance between final species in these two phyla which are very widely divergent. This 
becomes clear in the following treatment of Serridentinus (Chap. X). 



Chapter VIII 

HISTORY OF THE SUBFAMILIES LONGIROSTRIN^E AND AMEBELODONTIN/E. 
ORIGIN AND GEOGRAPHIC DISTRIBUTION OF THE LONG-JAWED 

MASTODONT PHYLA 

Separation of the 'true bunomastodonts' of the family bunomastodontid^. Characters of the 
five' bunomastodont subfamilies, the longirostrines. serridentines, rhynchorostrines, notorostrines, 

AND BREVIROSTRINES. PhYLETIC HISTORY AND MIGRATIONS OF THE EXTREMELY LONG-JAWED PHIOMIA-AME- 
BELODON, ALSO OF TRILOPHODON. DeNTAL SUCCESSION IN THE LONGIROSTRINES AND OTHER MASTODONTS. 

1. Characters of the family Bunomastodont idae and separa- 10. Trilophodonts or Serridentines of Japan. 

tion of this family into nine' distinct subfamilies. j i. Miocene and Pliocene mastodonts of Baden and Rhein- 

2. Characters of the subfamilies Longirostrinse and Tetra- hcsseii, Klahn, 1922 (p. 281). 

lophodontinae. 12. The Ix)wer Miocene Trilophodont of France (Trilophodon 

3. Dental succession in the Longirostrines and other pontilerieiisis) compared with the primitive Tri- 

ma-stodonts (see also p. 230). lophodon cooperi of Baluchistan. 

4. Generic characters ofPhiomia and its division into three ,3 Migration of Eurasiatic Trilophodon at the close of the 

to four progressive species, r. minor, r. serndens Miocene 

{ = winloni) , P. lointnni, P. osborni. 

5. Transition from Phiomia osborni to Phmnia pygvuem. ^^- ^''''Kh AmericT'*''''"^ ^^"'^^ "^ *^^ '"^^ ''°^'* ""^ 

6. History and characters of the genus Trilophodon Fal- 

coner; chronological order of discovery and descrip- l'^- 'Uproot-tuskers' of Kansas, Nebraska, South Dakota, 

tion of the forty actual or related specie.-^ of Tri- Colorado, and New Mexico (p. 288). 

lophodon. 16. Review of the Frick Collection of Trilophodonts, 1924- 

7. Typical Trilophodon anguslidens of Cuvier including 1933. 

European and African stages. Separation from 17. Summary of the origin and phylogenetic succession of 
Serridentinus. the Trilophodonts. 

8. Trilophodon (Choerolophodon) of .southeastern Europe 18. Discovery and description of the 'shovel-tuskers' 

(Pikermi, Samos), and of Asia (Maragha, Persia). {Amebelodon, 1927), successors to Phiomia osborni. 

9. Asiatic species of Trilophodon, chiefly of Middle and 

Lower Miocene age. Evolutioii of the central .\ppkndix. Cuvier's tyjie of Mastodon anguMidens. 

conules (pp. 276-278). Central conules distinctive of Trilophodon. 

(December, 1933). The epoch making discovery of the 'shovel-tuskers' of eastern Eurasia and of North America, subsequent 
to the completion of the main portion of the present Chapter VIII , is fully described in Section 18, including the chief discoveries 
between the years 1927 and 1933. (October, 1934) Ancestral to the 'shovel-tuskers' Amebelodontinae was the Phiomia of northern 
Africa (see Pis. V and VI between pp. 235 and 236). 

This chapter is purposely comprehen.sive in order firmly to establish the di.stiiietions of the Bunomastodonti- 
diP from the Mastodontidae, namely, in the blocking of the intermediate valleys by a central cunule or a 
'trefoil spur' which is wanting in the superior grinders of two primitive species, Phiomia minor and P. inntoni 
(of. Fig. 179, B, A), but makes its appearance as a neomorph or rectigradation- in the inferior grinders of P. 
wintoni (Fig. 182 Bl) and in the inferior grinders of P. osborni (Fig. 185, A 1), probably also in P. pygniceus (Fig. 187 A). 
This very important conule reappears directly in the center of the valleys in the type of 'Mastodon' minutus Cuvier 
(Fig. 191) ; it attaches itself to the external cones in the inferior molars, forming the outer trefoil, and to the internal 
cones of the superior molars, forming the inner t refoil, as shown in figure 190, Cuvier's type of Mastodon' angustidens. 
This central conule of Phiomia and of Trilophodon (Figs. 185 and 195) and po.ssibly of Tetralophodon is readily seen 
in the unworn condition of the grinding teeth and clearly distinguishes the grinding teeth of the Longirostrinae 
from those of the Serridentinae, in which there is no central conule but lateral spurs that gradually arise on the in- 
ternal and external cones of the superior and inferior molars respectively. In tlie worn condition of the grinding 
teeth the molars of Trilophodon and of Serridentinus apjiear very much alike. 

This important distinction is explained below in describing the subfamilies of the Bunomastodontidae, also in 
the early part of Chapters IX and X. The progre.ssive development of the grinding tooth pattern in the Buno- 
mastodontidae is illustrated in a comparison of the species arranged in ascending order of specific evolution under 
the four most highly characteristic genera, as follows: 

'(193.">) Soo Appendix of prcsoiit Volumr for final classificiition of the Ma.stodontoidea. .\t date of Koii>K *<> press, the Buriomastodontida' embrace seven 
subfamilies, namely, the !x)ngiro.striiiff', .\melM'lodontina', Tetralophodontina', Rhynchorostrina', Notorostrina>, Brevirostrina", and (inatlialielodontinae. The 
Serridentina' and Platyhelodontina' have been removed to the Serridentida> fam. nov., the Humboldtinse to the Hnmboldtida> fam. nov. 

- 'Rectigradations' Osborn, 1908, are termeil aristOKcnes' Osborn, 1932-1933. 

'225 



226 



OSBORN: THE PROBOSCIDEA 



Intermediate molars with three 
or four crests; shovel-shaped incisors 
of Phiomia oxborni descent 



Amebelodan 
Trefoils progressive 

Amebelodon fricki, etc. 

'Shovel-tuskers' of Eurasia 
and America 



Rudimentary or progressive conules 
in the median valleys. Third inferior 
grinders with three and a third crests. 
No trefoils. Crowns primitively 
brachyodont. Central conules progres- 
sive. 

Phiomia Andrews 
Phiomia pygmxus 
Phiomia osborni 
Phiomia wintoni 
Phiomia serridens 
Phiomia ininor 

Primitive 'shovel-tuskers' of the 
African Oligocene 



Typically developed conules in the 
median valleys. Third inferior molars 
with four and a half to five and a half 
crests. Single external and internal 
trefoils pereistent; bare rudiments of 
double trefoils. Crowns persistently 
brachyodont. 

Trilophodon Falconer 
Single trefoils 
Trilophodon (Megabelodon) lulli 
Trilophodon giganteus 
Amebelodon {Trilophodon) hicksi 
Amebelodon (Trilophodon) paladentatus 
Trilophodon abeli 
Trilophodon pojoaquensis 
Trilophodon {Genomastodon) osborni 
Trilophodon [Genomastodon) willistoni 
Progressive species of the 
American Mio-Pliocene 
Trilophodon angustidens cuvieri 
Trilophodon angustidens var. austro- 

germanicus 
Trilophodon pandionis 
Trilophodon palieindicus 
Trilophodon macrognathus 
Trilophodon sendaicus 
Trilophodon angustidens 

Primitive and progressive 
species of Eurasia 
Trilophodon angustidens libycus 
Primitive species of 
Moghara, Egypt 



Third inferior molars with five to 
eight and a third crests; intermediate 
molars with four crests. Double tre- 
foils arising in inferior and superior 
grinders. Crowns brachyodont to 
subhypsodont, to hvpsodont. 



Tetralophodon Falconer 
Double trefoils 
Tetralophodon [Morrillia) barbouri 
Tetralophodon campester 
Tetralophodon elegans 
Tetralophodon pnnjabiensis 
Tetralophodon grandincisivus 
Tetralophodon longirostris 
Progressive species of the Pliocene 
of America and Eurasia 



Tetralophodon [Lydekkeria) sinensis 

(single trefoils) 
Tetralophodon (Lydekkeria) falconeri 
Primitive species of China 
and India 



1. CHARACTERS OF THE FAMILY BUNOMASTODONTID^ AND SEPARATION OF THIS 

FAMILY INTO NINE' DISTINCT SUBFAMILIES 

Family: BUNOMASTODONTID^ Osborn, 1921 
Original reference: Amer. Mus. Novitates, 1921, p. 2 (Osborn, 1921.515). 

(1) As previously described in Chapter II of this Memoir, the family Bunomastodontidse comprises the 'true 
bunomastodonts,' in which, as the two Greek words Powos and Mao-ros indicate, the grinding teeth retain or 
progressively develop 'conules' and single, double, or quadruple 'trefoils' in the valleys between the transverse 
crests; whereas in the 'true mastodonts' the valleys between the transverse crests are uninterrupted by either 
'conules' or 'trefoils.' Family characters arise independently in nine' subfamilies (see table, p. 228). 

(2) This type of grinding tooth does not appear in primitive bunomastodonts Uke Phiomia minor (Fig. 179 
B) or certain Phiomia wintoni (Fig. 179 A), but is clearly shown in progressive species Hke Phiomia osborni (Fig. 
185) in which a large conule appears between the crests of P4-M3, and is shown to connect itself on prolonged 
wear with the external cones of the inferior molars, thus constituting a true 'external trefoil'; this conule reappears 
in the grinding tooth of Phiomia pygmaeus (Fig. 187 A), in which the summits of the cones subdivide into two 
'conelets' as in all higher mastodonts. In certain lower grinders referred to Phiomia wintoni (Fig. 182 Bl) these 
conules also appear, but they are absent in upper grinders referred to Phiomia wintoni (Fig. 179 A). 

'See footnote on page 225. 



PROGRESSIVE LOPHODONTY AND HYPSODONTY 



S- productus 
Amor. Mus. 10682 R«f' 








T. giganteus 



S. p'ogressus 
Amer. Mus. 8529 Typo 



S. producius 
Amer. Mus. 10562 Ref. 



Phiomia osborni 
Amer. Mus. 13468 Type 



All 1/4 nat. size 









^rvyrf} 



Phiomia (A) and Serridentinus (B, C) Cranial Comparison 
Scale not uniform 

Fig. 170. Craiiiul profiles and lateral enamel bands in the African 
Phiomia and tlic American Serridenlinus. Scale not uniform. 

American Lower Pliocene. C, SerriderUinus prodtictus ref., Clarendon 
formation, Texas. 

American Ixjwer Pliocene. B, Serridenlinus serridens ref., of Claren- 
don age, Texas. 

N. Africa, Ixiwcr Oligocene. A, Phiomia minor (skull), wintoni (jaw). 
Fluvio-marine beds of the FayOra, Egypt. 

The Phiomiii minor skull is about half the size of the two Serridenlinus 
producius and S. serridens skulls; the proportions of the facial and cranial 
regions are very similar; the elongation of the jaws is practically similar and 
much less pronounced than in Trilophodon artgustidens. These animals, 
therefore, are of medilongirostral type. 



Phiomia (A), SERRiDEXTiNrs (B-D), Trilophodon (E), 
Tetrai.ophodon (F). Scale inifohm 

Fig. 171. Evolution of the left second and third inferior molars in the 
Tetralophodontinff, Longirostrinse, and Serridentina". Uniform reduction 
scale of one-fourth. 

F, Telralo/ihodon campester ref.. Neb. Mus.; M3 with 6'^ crests, Mi 
with 4 crests. 

E, Trilophodon giqanleiis type, .\mer. Mus.; M3 with 4^ to 5 crests, 
M2 \vith 3,^3 crests. 

D, Serridenlinus progressus type, Amer. Mus.; M3 with 4)^ crests, Mj 
with 3 crests. 

O, Serridenlinus producius ref., .\mer. Mus.; Mjwith i% crests, Mj 
with 3 crests. 

B, Serridenlinus serridens ref.. Amer Mus.; M3 with 4^3 crests, Ms 
with 3 cre.sts. 

A, Phiomia osborni type, .Amer. Mus.; M3 with 3% crests, Mj with 
3 crests. 



227 



228 



OSBORN: THE PROBOSCIDEA 



s 



s^ = 

^ r "3 

„ ° 2 c 

f3 o 0/ .5 






o ?; o 



C3 -5 

c — 






2 c c 
•3 .o 



^A 



o 0^ 3 h-— ? 



^ "3 



■a 1- a; rj o o 



o 



n >, 

•rz-C ■- 
■q. s^ 2 



::* «J o r ::^ 









£ o = 






Si. a. c 

o c S ■ 

o o £ 

e "S ~ S 



c 



IN CO ■* 



CO 

o ^ S 

"S a s s 
2 o o o 

e o o o 

S -K^ -l.^ -^A 

CC 60 fie 

oca 
o o o 

!» Ol Ci 
_^ ^ ^ 



9 

a 



3 j; 

■Ceo 

VI ^ 

C '{■ 



-^ "^ "^ 



D. 

a 

3 t 



■ see 
;s 3 Oj 



X 






5 g.l So 



= <); .5 



o c 
o — 



o 
2 



"5 g..2i g 
. ^ ^ -^ 1 



S 

_H 
'43 
•o 

o 



5 "S 
c a. 



-2 a> 3 o 

~ ?- -a aj 

~ S ~ =- 

^ ~T^ ^^ 3 ^ "^ 

-^ - ° o o 



-r 'S '2^ 



^'^< 



a. 3 

IS 72 *^ 

03 -^ 






'^ 2-5 §oo 



^'"3. 



cS ? 



is S 



s-s 



"S *^ 15 .2 

O O 'S 

■5? -S t:_^ 
2 C g £ 

3 >» •— tfl 






1^1 

I I o- 



- m V. 



O S 
C 5- 



~ a. 



s. 



ft o 



: ■^ 
> "S 






3 & 



5-r- 

03 



C'S'^ 



~ o o 

^ ^ " " "e ■ 

a ^ =: a « -ii ^ 



J^"? 



-5.S, S — 5J 



ro Tf 



10 CO 






03 =S .iri 
« 5 § 

-^ "S 5 



C 

2^ 



S 03 



-^ Q 

-£ c -a 
o o 



in 



c 

o 

O 



« ^-^ "^ 

C o z c 

J, 03 - rt 
CO ;. O — 

c- ^ , 

^'" c; 3 



'c ^ Cj 



c 

c 
o 



I £ £ 






X o 



s s s 



- c c 
.300 



si 



CSS 
^ 'ij "is ^ 



^ 'r^ ZO VU Jii 



e «u e 

r« S £ 
Coo 

O JS ^ 
3 - O 






"§ 






•s o 
A o 



; -H (M CO ■* in CO r-H 



3 -e — 









o 
-J 



o „ 

a - 

lili 



£ — a> 



S 3 
7; (V o 
■5 x "3 X 

■*- X ?^ 
Oj t- '' 
.S f. OJ 3 

0.-33 o 

— *f .s " 



2 fl 



!§: 

._ T3 oj 

^ — O) 
J- o o 

I -^ t' 
•° r.- -3 - 

» S of s 
<; = a- 



-2 i.i 

_i-' I — X 

35 a. 
o — >,■; 

-c 



•S _ 

C -r 
O 5= 

o ~ 

^ 5- 



c 

X 

c 



3 X 

S 



* "o.°i -S 

S .0 o c- s 



0- — 



■ -3 O o 

■ 21 

-CXI .S 






o ^^-c 



'See footnotp on page 225. 

^NotorostrinsB, Humboldlinse, and Serridentinie (Kolianiaslodmi). 



(3) Another dental characteristic of all primitive buno- 
mastodonts is the presence of a lateral enamel band on the 
superior tusks in apposition to the elongate and flattened 
inferior tusks which lack the enamel band. This betokens 
the primitive functional use of the upper and lower tusks in 
apposition, correlated with a grinding and triturating molar 
action rather than the chopping motion characteristic of the 
'true mastodonts' (P'ig. 172). 

(4) This combined functional action of the tusks and 
grinding teeth arises in Phiumia, persists in the Longirostrinse 
or 'long -jawed bunomastodonts,' in the Serridentinse or 
'medium-jawed mastodonts," and in the Rhynchorostrinae 
or 'beak-jawed bunomastodonts.' The latter undergo a 
more or less parallel evolution, the original functions of the 
tusks and of the grinding teeth being retained, but with 
the progressive addition of double trefoils in the genus 
Tetralophodon of the Tetralophodontingp (Fig. 172). 

(5) In three of the subfamilies, namely, the Notorostrinae, 
the Humboldtina;,' and the Brevirostrinse or 'Brevirostrines,' 
this original combined function is lost through the abbrevia- 
tion of the jaw and the disappearance of the lower tusks. 
Once released from their cutting function with the lower tusks 
the upper tusks begin to rise, first retaining the enamel (Cordil- 
lerion andium), then losing the enamel and becoming upturned 
{Cuirieronius humboldtii^). In parallel evolution the 'short- 
jawed bunomastodonts' (Brevirostrinse) lose their lower 
tusks at an early stage and develop long, projecting horizontal 
or upturned superior tusks without enamel (Fig. 172). 

(6) By these mechanical methods the Bunomastodontidae 
rose from small proboscideans similar to Phiomia of North 
Africa, probably frequenting the lowlands and swamps, 
adopting a browsing habit with their bunomastodont grinding 
teeth, and gradually differentiated into nine' subfamilies, 
members of which migrated throughout the entire Northern 
Hemisphere and gave ofif three branches," which pene- 
trated to the southern extremity of South America. 
We treat this subject of adaptive radiation and geographic 
distribution of the bunomastodonts more fully in Chapter 
XXI on "Affinities, Migrations, and Phylogeny of the Probo- 
scidea." Meanwhile these nine' subfamilies may be clearly 
distinguished, as in the accompanying table. 



Trilophodon 
iLilli 




Phyn ch otherium 
tlascalaJe 



All '/2if. natural size 



DlVEKGBNT AUAI'TIVE UaDIATION OK CrANIA AND INCISIVE TlISKS IN SlX BUNOMASTODONT SCBFAMILIFS. ScAI.E UNIKOHM 

Fig. 172. Comparative o\itliii(' illvistration.s, sitle view, to same scale, of the completely restored skulls, jaws, sinil tusks, in desoendin^ order, of: 



Trilophodon {\t egabelotton) lulli' 

Trilophniloii {(Irnoma.iln^lntt) iiilli-iltmi, juvenile 

Pliiomin minor 

Serriihfitinns protlurlns 

'I'lirolophoihn gromlincisirii^i 

Trtriilophnilon cotiifjester 



Stegomastodon arizona' 
Ananciis arvernensis 
Cuvieronius hiimboUtii, juvenile 
CnrdUlerion andvim 



RIn/ncholherium Uaxcals, restored 

<)l).serve persistent enamel band on inferior 
tusks, also ap|)<)sition of inferior and .superior 
tusks. In contnust with suix-rior enamel band 
only in Trilophodon. Phiomin. Srrridrnlinus, and 
Tetrolophodon. and nhxencr of enamel band on 
inferior tusks. 
(Osborn, liCil) Si-rridtntiniis \)vU\i\ii> in the new faniiix .SMiidentidie; .SVcgojnas/odoM and '^'i/niro/iii/.v in the new family lIu!Mlnildtida> 



'The type mandible of Trilophodon lulli, orieinally restored by Barbour in 1914 with a pair of slender inferior incisive tusks (KiR. 244 of present 
Memoir), is now found to be tuskless. The rostrum is expand.d at the tips. See Barbour, 1934.2. 



229 



**P3 "««• 



4pj <lPs -^P* 



>^^%^ 



A /^ p» p. 





RHYNCHOTHERlUn 



R. SHEPARDI nor Le^dtf 
AM. No. as 5 1 



ELEPHAS PLANIFRONS 
(HYPOTHETICAL) 




LOXODON AFRICANUS 



Dental Succession: Bunomastodontid-e and Elephantid.e. After Frick, 1926. Scale not uniform 
Fig. 173. Dental succe.ssion in the Amebelodontinae, Longirostrinae, and Rhynchorostrins compared with that in the Elephantidte. 
After Frick, 1926, p. 126, fig. IC. 



Phiojnia wintoni 

A, Juvenile jaw containing complete functional milk series, 
Di2, Dp2, Dp3, Dp4, also Mj; permanent premolars, P3, P4; 
M2 partly buried in jaw. 

B, Adult stage with adult functional dentition, I2, P3, P4, 
M1-M3. 

C, Incisors in section. 
Trilophodon [=SerTidentinus] produdus 

A, Juvenile jaw, partly hypothetical (compare juvenile 
Phiomia urintoni, also Amer. Mus. 21113 and 21112). 

C, Incisors in section. 

B, Adult mandible (Amer. Mus. 21111, neotype) containing 
M2, M3 only. 

C, Incisors in section. 

Elephas [=Archidiskodon] planifrons 

A, Juvenile jaw, hypothetical. 

B, Semi-adult jaw, hypothetical, P4, Mi, M2. 

Both figures based on Falconer and Cautley, 1846 [1845], 
PI. XII, figs. 8-11, and PI. xiv, fig. 10. 



Mastodon ainericanus 

A, Juvenile jaw containing Dp2-Dp4. After Kaup, 1835, 
PI. XX, fig. 1 (cast Amer. Mus. 10460). 

B, Semi-adult jaw containing Dp4, Mi, M2; M3 in jaw. 

C, Incisors in section. 

Rhynchotherium 

A, (?) /?. (Dibelodon) edensis Frick (Amer. Mus. 18216B), 
Eden Pliocene, Cahf. [= Rhynchotherium shepardi edensis 
of the present Memoir). Juvenile jaw, Di2, Dp2-Dp4 in situ, 
Ml in lower part of jaw. 

B, R. shepardi [ = Rhynchotherium falconeri Osborn, tj-pe] 
(Amer. Mus. Cope Coll. 8532), Blanco, Texas. Adult jaw, Mj, 
M3 in function. 

C, Section of I2. 

Loxodonin africana 

A, Semi-adult jaw containing functional M2; M3 still in jaw. 

B, M2 shedding, M3 coming into use 



230 



2. CHARACTERS OF THE SUBFAMILIES LONGIROSTRIX^ AND TETRALOPHODONTIN.E 

Original reference: Bull. Geol. Soc. Amer., 1918, \'ol. XXIX, p. 136 (Osborn, 1918.468), and "Contribution to the Knowledge of 
the Fossil Mammalian Fauna of Java," van der Maarel, 1932, p. 108. 

Subfamily Characters. — (1) Cranium uniformly abbreviated. (2) Ro.strum, facial region, 
moderately elongated (medilongiro.stral) or rostrum elongated (longirostral). (3) Lower jaw.s extremely 
elongated (longiro.stral) , moderately elongated (medilongirostral). (4) Superior inci.sive tu.sk.s down- 
curved, outcurved, persistently broad enamel band on concave outer suface. (5) Inferior incisive 
tusks without enamel, progressively spatulate or flattened, persistent (Trilophodon) , undergoing 
reduction (Tetrnlophodon); abrading dentine on inner side of superior enamelled tusks. (6) Grinding 
teeth brachyodont, with single trefoils, intermediate grinders trilophodont {Phiomia,Trilophodon) ; inter- 
mediate grinders tetralophodont {Tetralophodon) , with double trefoils, subhypsodont to hypsodont. 

(7) In final phyletic stages the grinding tooth action concentrated on third superior and inferior molars, 
M', Mg; in primitive forms [Trilophodon angustidens), two or three grinders persisting in action. 

(8) Progressive suppression or retrogressive suppression of true premolar teeth, P^"\P 2-4, in upper and 
lower jaws, in contrast to the Serridentinse. (9) Double trefoils slowly evolve in Trilophodon. 

This subfamily of long-jawed bunomastodonts was originally regarded as monophyletic, that is, it was 
generally believed that the Miocene genus and species first discovered {Trilophodon angustidens) passed into the 
Pliocene Tetralophodon longiroslris. Following studies by Andrews and Matsumoto, Osborn demonstrates that 
the species Phiomia osborni is directly ancestral to Amebelodon. The typical Trilophodon angustidens of the 
Miocene does not give rise to Tetralophodon longirostris but passes into a long line of direct descendants in 
which the Phiomia and Trilophodon characters are progressively developed in the elongation of the jaws and in 
the persistent use of the lower tusks as an uprooting 'shovel-tusk' (Greek om»/, shovel, ^i\o% dart, and 'ohov%, tooth), 
hence the final stage Amebelodon fricki of Barbour. Consequently it appears (1935) that the Longirostrinae are 
not monophyletic but diphyletic and that 'M.' longirostris and related species belong to a distinct subfamily 
named by Maarel (1932) Tetralophodontinae. 



Longirostral Extreme = Longirostrin^ Osborn, 1918 

Excessive elongation of the jaw, i.e., hjperlongirostral. 
Single mesotrefoils in the superior and inferior grinders. 
Persistent brachyodontj' to subhypsodonty. 
Persistent uprooting function of lower incisive tusks. 

Trilophodon, Miocene and Pliocene of Eurasia and 
North America. 

AMKBELODONTiN.e Barbour, 1929 
Phiomia, Oligocene of North Africa. 
Amebelodon fricki of Nebra-:ka, a direct descendant 
of Phiomia. 



Medilongirostral =TetralophodontinvE Maarel, 1932 

Moderate elongation of the lower jaw and of the rostrum of 
the cranium, i.e., medilongirostral. 

Early development of double trefoils on superior and in- 
ferior grinders. Progressive addition of crests and trefoils in 
third superior and inferior grinders. Addition of fourth 
crest in certain of the intermediate grinders, e.g.. Dp 4- 
M2. 

Subhypsodonty to hypsodonty, coronal cement on grinders. 

Tetralophodon (Syn. Tetrabelodon) , Pliocene of Asia and 
North America. 

Lydekkeria, Miocene of southern A.sia. 

Morrillia, Middle Pleistocene of Nebraska. 



With this explanation of a certainly diphyletic arrangement, we are able to define the subfamily Longi- 
rostrinae by enumerating the characters common to all species of Trilophodon; the subfamily Tetralophodontinae, 
as first distinguished by Maarel, by enumerating the characters of Tetralophodon (see also Chap. IX). 

Geologic Range. — The long-jawed mastodonts compose numerous radiating phyla both geologically and 
anatomically, beginning ^\-ith the numerous skulls of Phiomia, as well as the numerous skulls and partial skeletons 
of Trilophodon angustidens of the Miocene of France, and jaws and skulls of the Miocene and Pliocene of 
North America, continuing with the skulls and teeth of Tetralophodon of Europe, southern Asia, and North 
America; these long-jawed mastodonts appear to persist in savannas and forests until Middle or late Pliocene time. 

231 



232 



OSBORN: THE PROBOSCIDEA 



Geographic Range.— Gervais, Gaudry, and Pomel (see Poiiiel, 1895, pp. 8 and 9) record from Barbary (com- 
prising Morocco, Algeria, Tunisia, and Tripoli) Mastodon [ = Tetralophodon] longirostris in the lacustrine beds of 
Smendou north of Constantino; Mastodon [= Trilophodon] angustidens at Djebel-Chirichira, west of Kerouan, 
Tunis; Mastodon [ = Turicius] tapiroides from the region of I' Aures— geographic determinations of importance 
which may bear on the probable African origin of the Longirostrines. 

The Amebelodontines, as described below in the case of Phiomia, were probably savanna and stream-border 
dwellers, browsers rather than grazers. Tubers and leaves of trees were seized by pressing the elongated 
upper lip against the inferior teeth in Phiomia. The branches were drawn down by the sharpened superior 
tusks, which were also used as offensive and defensive weapons. The flattened lower tusks in Phiomia and 
Trilophodon were used in uprooting plants and smaller shrubs. The remarkable and rapid traveUng powers 



FIQ 3 .Jv'-vi^c-,* 



FIG 2. 




DKNTAr, Succession in Mastodon, RnvNCHOTHERinM, Serridentinus, and Trilophodon [ =Ocalientinus|. After Frick, 1926 

Fig. 174. Emergence and replacement of premolars and molars. After Fnck, 1926, p. 128, with the substitution of specific 
determinations corresponding to tho.se of the present Memoir. 

(Fig.s. 2, 2A, 3) Amer. Miis. 18216 |B|. Referred IRhijTichotherium {[Hbelodon) edensis Frick [=Rhyncholherium shepardi edcitsis 
of the present Memoirl. .Juvenile ramus, Pliocene of Eden, California. 

(Figs. 4.\, 4B) .\mer. Mus. 1434.5 IMnslodnn nmeriamus) and .\mer. Mus. 8532 (type of Rkynchotheriiim falconeri) introduced 
for comparison. 

(Fig. 5) Amor. Mus. 21113. Referred juvenile Serridentinus productiis, Santa F6 marls, New Mexico. 
(Fig. 6) Amer. Mus. 21112. Referred juvenile left ramus of Serridentinus productus showing unerupted Pi and M2. 
(Fig. 7A) Amer. Mus. 21111. Adult Serridentinvs jn-nductus, ncotype. Santa F^ marls, New Mexico. 

(Fig. 7B) .^iner. Mus. 21123, -\dult paratype jaw of Trilophodon pojoaquensi^ (1933, referred to Ocalientinus ojocaliensis) 
clearly showing the postsymphyseal (p.s.) region. 



/xi "^i, 



FIC 9 

'>i 'ft i,. 



FIG 8 



fiG II 




Dental Rucckssion in Rhynchothkhu'm, Serhidentinus, and Trilophodon [ =0calihlntini's1. After Frkk, 1926 
Kig. 175. .Su(K>rior aspprts of the mandibulnr rami ami dentition of Khijiu-holhenum, StrruUtUiniu<, and TriUn>hoilon [=()cnlienHnHs\. Aflcr Frii-k, 1926, 
pp. 132. 13.'i, fijj.s. 8-13. Kigs. 8 to 12 one-!=i\th natunil size; fijr. 13 one-eighth natural size. 

(KiRs. 8, 9i AiiH T Mu.-i. 18210 [HI 'HIitincholhiHum (Dibelodon) aleniiix Friok, Mt. Eden Plioeenc, California l=Rhyuchotherium shrpardi ederusix of the 
Memoir!. Deeiduous tusks, Di-j, and piemolars. 

(Fig. 10) .\mer. Miis. 21113. Serritlciitinus productus, Santa Fe marl.-*. Now Mexico. Referred juvenile dentition of ri^lit side, reversed. 
Serridt-ntiniis productus, Santa Fd marls. New Mexico. Referred juvenile left ramus and symphysis. 
SirridciUinus irroductus. Santa Ft^ New Me.\ieo. Xeotype adult left rainu.s. 
Sern-denlinui produclux. Referred left rainu."!. Santa F(^ marls. New Mexico. 
Tiihiphodon pojonqueusis ( = (1933) Omliriilinus njocaliensis\, .*!anta F6 marls. New .Mexico. External and internal views of 



(Fig. 11) Amer. Mus. 21112. 

(Fig. 12) Amer. Mus. 21111. 

(Fig. 13) Amer. Mus. 21119. 

(Kig. 13) Amer. .Mus, 21124. 
referred superior and inferior grinders, P-1-M2. 

(Fig. 13) .\iner. .Mus. 21 123. Trilophodon pojoaquensUi \ -■ 
lacking anterior rostrum. 



(1933) OcaUeiUmus ojocalicnsis\, .Santa F<5 marls. New .Mexico. Incomplete paratype mandible, 



233 



234 



OSBORN: THE PROBOSCIDEA 



FIG. 18 A 




N.M.No.4179 A.nNo.2.m4C - 
FIG. 2.0 C 



A.M.No.2ll24D 
FIG. 2.0 A 




N.n.No.Z062 





FIG. 20 B 



Fig. 



176. Deciduods Premolars in Sehridentinhs, Tri- 

LOPHODON [ =0CALIBNTINUS], AND RhYNCHOTHERIOM. AfTER 

Frick, 1926 

Deciduous premolars, Dp^-Dp* o( Serridentimis, Trilophodon, 
and Rhyncholherium. After Frick, 1926, p. 140, figs. 18A-20C. 

(Fig. 18A) Amer. Miis. 18218. ''.Rhyncholherium (Dibelodon) 
edensis Frick, referred I.Dp'-Dp*. Eden Pliocene, California. (Fig. 
18B) Crown views of same teeth. [ = Rhyncholherium shepardi 
edensis of the present Memoir.) 

(Fig. 19A) Nat. Mus. 4179. Serridentinus productus, re- 
ferred r.P'. Santa F^ marls. New Mexico. 

(Fig. 19B, C) Amer. Mus. 21124C, referred r.P' of Tri- 
lophodon jiojonquensis. Amer. Mus. 2ir24D, referred r.P' of 
same species. Santa F^ marls, New Mexico. [ = Ocalientinus 
ojocaliensis of the present Memoir.] 

(Fig. 20A-C) Nat. Mus. 3064, type of ITrilophodon (Serri- 
dentinus) le.idii Frick; crown and lateral views of r.P'"* (germ). 
Nat. Mus. 3062, referred left P^ of same species. Florida. 
=Ocalienliniit (Ser.) fioridanua leidii of the present Memoir.] 



these Longirostrines and Amebelodonts from Africa, through 
Europe, Asia, and the west central United States as far 
south as Colorado, prove that they were well fed and well 
defended. There is no evidence for the Lull-Schlesinger 
theory that some of them were transformed by the abbreviation 
of the jaws into Brevirostrines. 



3. DENTAL SUCCESSION IN THE LONGIROSTRINES 
AND OTHER MASTODONTS 

Referring to the dental succession in Mastodon americanus 
(Chapter VI) and to the dental succession in Archidiskodon 
planifrons (Vol. II, Chap. XVI), the recent researches of Frick 
(1926) have completely clarified the previously obscure modes 
of dental succession in the Longirostrinae as compared with 
other proboscideans. The six underlying adaptive principles 
are as follows : 

(1) The precocious or accelerated development of the 
milk teeth, Di 2, Dp 2-Dp 4, including the molariform pattern 
of Dp 4, to meet the tender browsing and feeding habits of the 
juvenile stage, as in many other Herbivora. 

(2) The retardation and final suppression of the replacing 
true premolars, P 2-P 4. 

(3) The early eruption of M 1, early taking its functional 
place behind Dp 4. 

(4) The retarded eruption of M 2 coincident with the shed- 
ding of Dp 2, Dp 3. 

(5) The uniform cresting of the 'intermediate molars,' 
Dp 4-M 2, e.g., three ridge-crests in Trilophodon and Serriden- 
tinus, four ridge-crests in Tetralophodon. 

(6) The greatly retarded eruption of M 3, with its multiple 
ridge-crests (4-6-8)0, after all the premolars have been lost. 

These six principles are admirably displayed in the ac- 
companying diagrams by Frick {op. cit., p. 126— Figs. 173-177 
of the present Memoir) . 



The above diagram, which clearly sets forth the principles of premolar sequence, is followed in Frick's 
researches by numerous figures (Figs. 174, 175, 177 of the present Memoir) which illustrate in greater detail the 
actual structure of the tooth crowns and actual sequence of Dpa and Mi, M2, also the symphyses and more or 
less complete jaws of Serridentinus productus and of Rhynchotherium shepardi edensis. 



FIG. I3A 



FIG. 15 B 



A.MNo.16115 




A.n.No. 2.111^ 



A.M.No.Zmi 
FIG. ISA 





A.M.No.dlll4-B 

ric. 16 c 

dp4 




n.No.zim 



FIG. I6A 
N.n.Noi.5057-8 p. 



FIG. 11 A 




N.n.No.306l 





FIG. 16 D 



A.M.No.lUIZ 

FIG. 16 B 
N.M. Nos.30S7-d ''' 




N.M.No.5061 



Fig. 177. Deciduous and Permanent Premolars in the Serridentin^ and Longirostrin^. After Frick, 1926 



Deciduous and permanent [)rcinolars, 2-4, of Serridentinus productnn, 
S. floridanus leiitii, Trilophodon anguslidens, all one-half natural size. 
After Frick, 1926, pp. 136, 137, figs. 13A to 17A. 

(Fig. 13A) Amer. Mus. 18113. Dps-Dpi of the right ramus of re- 
ferred Serridentinus produclu,<i. 

(Fig. 14A) Amer. Mus. 21114. Referred Serridentinus productus. Dpi, 
Ml of the right side. 

(Fig. 14C) .^mer. Mus. 21121. Referred Serridentimts -produclus. Dps 
of the left side (reversed). 

(Fig. 15C, A) Amer. Mus. 21124B and 21112. Pvnnanent inferior pre- 
molars, 1.P3.4, of Serridcnti7iiti< prodticttt.^. 

(Fig. 16C) Nat. Mus. 3061. Referred Dp4 of Trilophodon {Serridentinus) 
leidii [ =Ocnlirntinus (Ser.) floridanus leidii]. 

(Fig. 16A) Nat. Mus. 3057-305S. P3 of the right side referred to same 
species. 

(Fig. 17A) Referred Trilophodon ang^islidens, r Dp; 4. After cast Amer 
Mus. Warren Coll., Sansan, Miocene. 



External and internal views of deciduous and permanent mandibular 
premolars, all one-half natural size. After Frick, 1926, pp. 136, 137, 
figs. 13B to 17B. 

(Fig. 13B) Amer. Mus. 18113. Inferior deciduous premolars, r.Dpi- 
Dpi (compare Fig. 13A). 

(Fig. 14B). .-Vmcr. Mus. 21114. External view of r.Dp4, r.Mj (com- 
pare Fig. 14A). 

(Fig. 15B) .Amer. Mus. 21112. External \'iew of permanent I.P3- 
Vi feompare Fig. 1.5.\). 

(Fig. 16D) Nat. Mus. 3061. External view of referred Dp4 of Tri- 
lophodon (Serridentinus) leidii Frick (compare Fig. 16C). l=Ocal. 
{Ser.) floridanus leidii.] 

(Fig. 16B) Nat. Mus. 3057-3058. Externa! view of r.Pj (compare 
Fig. 16A). 

(Fig. 17B) Referred Trilophodon anguslidens, r.Dpz-Dp^. Sansan, 
Miocene (compare Fig. 17A). 



236 



4. GENERIC CHARACTERS OF PHIOMIA AND ITS DIVISION INTO THREE TO FOUR 

PROGRESSIVE SPECIES 

Superfamily: MASTODONTOIDEA Osborii, 1921 
Family: BUNOMASTODONTID.E Osborn, 1921 
Subfamily: Amebelodontin^ Barbour, 1929 



Genus: PHIOMIA Andrews and Beadnell, 1902 

Original reference: Surv. Dept., Pub. Works Ministry, Cairo, 1902, pp. 1-5. 
Genotypic species: Phiomia serridens Andrews and Beadnell, 1902. 

Generic Characters. — (Matsumoto, 1924.1, p. 3): "B. — Palate long and narrow. Symphysis 
long, its posterior end lying only a little anterior to, or posterior to the anterior end of the first cheek 
tooth (Ps) ; the most conspicuous one of the anterior mental foramina lying far anterior to the first 



cheek tooth, as well as to the posterior end of the symphysis. Ridge formula: Dmj 



2 3 
2 3' 



Ml 



Last premolars and all molars typically bunodont ; trefoil pattern of cusps well developed [P. wintoni, 
P. osborni] Phiomia." 

This is a continuation of Chapter II, in which the original discovery and descriptions of the several species of 
Phiomia, the separation of this genus from Palseomastodon and MoBritherium, the habits and ancestral and phylo- 
genetic relationships, the geologic succession, and the original type specimens and figures are fully recorded. 

As pointed out clearly in Chapter II, Section in, the genus Phiomia, based on the genotype Phiomia serridens 
and since ampUfied by our present knowledge of the species Phiomia minor, P. wintoni Andrews and P. osborni 
Matsumoto, represents a group of species entirely distinct from species of the Palseomastodon group. Remains of 
the several species of Phiomia are found abundantly in the Fluvio-marine beds, Lower Oligocene, of the Fayum, 
and are eleven times as numerous as specimens referable to Palseomastodon. Inasmuch as Andrews was misled by 
the paucity and the fragmentary nature of much of his material, his original (1902) and final (1906, pp. 169-171) 
definitions of the genus Phiomia and the genotype species Phiomia serridens became the cause of years of confusion 
both in Andrews' own splendid memoir (1906) and in all subsequent publications by Andrews, Osborn, Schlosser, 
and others, until the matter was supposedly set right b.y Matsumoto (1922.1, 1924.1), as fully set forth in 
Chapter II above, on the "Origin and Classification of the Proboscidea. " 

Osborn, 1934: The four species of Phiomia described above (pp. 55, 58, 60, 64) and below (pp. 239-246) 
are all 'shovel-tuskers' and now prove to be directly ancestral to the true 'shovel-tuskers' of the subfamily 
Amebelodontinae of North America (see Pis. v and vi, pp. 235-236). 

The above generic and phylogenetic characters are practically confirmed and amplified in the four species 
included within this genus and distinguished by measurement as follows : 



Phiomia minor 

Andrews, 1904, cf. 'Phiomia' 

wintoni in part 

Length of lower molar .scries, 
Mi.3 145 mm.; of lower pre- 
molar-molar series, P3-M3, 201 
mm. 



Phiomia serridens 
Andrews and Beadnell, 
1902, cf. Phiomia {?) 
wintoni 
Known from type antl para- 
type inferior and superior milk 
dentition: Dia, Dp"-Dp2, Dp'*- 
Dp3, Dp\ M'. 



Phiomia wintoni 
Andrews, 1905 = Syn. 
Paleomastodon hairoisi 

Length of lower molar series, 
M1.3, 166 mm.; of lower pre- 
molar-molar .series, Ps-M;;, 
236e mm. M3 breadth-length 
index 53-59, average 55.5; M' 
breadth-length index 69-80, 
average 74. 



Phiomia osborni 
Matsumoto, 1922 



Length of lower molar series, 
M,.3, 172 mm.; of lower pre- 
molar-molar series, P3-M3, 246 
mm. Central conules in Mo, 
M3. 



236 




t'-Vs 



PLATE V 

The Ancestral 'Shovel-tusker' Phiomia of the Olioocene op Egypt Evolves into the Typical 'Shovel- 

TrsKER' Amebelodon fricki of the Middle Pliocene of Nebraska. Compare Plate VI 

See also Appendix for Comparative Observations 

Mandibles A-E, one-sixth natural size; Kl, one-eighteciitli natural size; F, one-ninth natural size. Sections A-C, 

F2, F3, one-third natural size 

A, Juvenile mandible of Phiomia imntoni, Amer. Mus. 13470. Length of mandible .■)20 mm., length of tusks 140 mm. 

B, Adult mandible of Phiomia mnloni, Amer. Mus. 13471. Length of mandibli' 040 mm., length of tusks 180 mm. 

C, Type mandible of Phiomia osborni, -Amer. Mus. 13468. Ix-ngth of mandible 680 mm., length of tusks 2.50 mm., 

weight of tusks 12.'» grams or 4.4 ounces. Section C, 7 lamina; of dentine (1-7), also central cavity as in P. n-intoni 
(A) and in A. fricki (F2, F3). 

D, Type mandible of Phiomia u-i„loni, Brit. Mvis. M.8414. Length of mandible 670 mm., length of tusks 240e mm. 

E, Referred mandible of Phiomia mnloni, Brit. Mus. NL8476 (after Andrews, 1906, Pl. xiv, fig. 3). Length of tusks 

260e mm. 
F 1, Typ<> mandibl.' of .1 mibeUxlon fricki, Neb. Mus. 4-4-27. Length of mandible 1475 mm., length of tusks 1144 mm. 
Weight of tusks lOtitX) grams or 23 pounds, 6 ounces. 

F, Ty|H- left lo«cr tusk of A nuModon fricki. Neb. Mus. 4-4-27, after cast in the American Museum, with outline of 

same tusk, reversed. The mass of the Amehflmion fricki tusk is more than seventy times as great as the mass 
of the /'/ii'omin osftor/ii tusk (C). 
F2. F3. Tvoe tusks of Amebelodon fricki, cross-section sliowing nine (1-9) lamina' of dentine, also central ca\aty as 





PLATE VI 

(Upper) Restoration of Phiomia osborni of the OliKorciic of ERvpt. Onp thirty-sixth natural sizp. By Miirgret Flinsch, 1932. 
Compare figure 17, page 45. 

(Lower) Restoration of Antebehdon fricki, 'shovel-tusker' from the Middle Pliocene of Nebraska. One thirty-sixth natural size. By 
Margret I'liiisch, 1932. Compare figure 284, page 332. 



THE AMEBELODONTIN^: PHIOMIA 



237 



PlIYLOGENKTIC CHARACTERS OF PhIOMIA. — 

Osborn, 1923: In Osborn's phylogenetic system each 
genus includes a long series of geologically ascending 
species and ascending mutations in which certain generic 
characters are distinctively progressive. In a compari- 
son of the abov'^e four species, namely, Phiomia minor, 
P. serridens, P. wintoni, and P. oshorni, we observe 
three such progressive generic characters: (1) Progres- 
sive doUchodonty, the grinding teeth becoming longer 
and narrower in proportion; (2) progressive bunomas- 
todonty, the central conules, destined to form trefoils, 
becoming more prominent; and (3) progressive doli- 
chopy, the facial region elongating more rapidly than 
the cranial, the latter character being indicated in a 
comparison of the jaws. The successive stages in generic 
evolution are all tending towards the typical 'shovel- 
tusker' Amebelodon fricki of the American Pliocene. 

Pr()(;ressive Dolkhodontv. — The detailed com- 
parative measurements by the author, assisted by Dr. 
Charles C. Mook, of twenty-four specimens referable 
to the three species of Phiomia in the British, Cairo, 
and American Museums more than confirm the state- 
ment of Matsumoto that the grinding teeth of Phiomia 
are relatively long and narrow as compared with those 
of Palaeomastodon. These relative proportions reach 
the narrowest extreme in the largest and most progres- 
sive species Phiomia oshorni, index 54, as shown in the accompanying table : 




Fig. 178. Palates of the small (A) Phiomia minor and of the large (B) 
Phiomia wiidmii reduced to the same dental scale, to illustrate differences 
of facial [)roportion in these two species: (I) The rmiii^il proportions are the 
same in the two .species; i'i) the facial region is relatively longer (liolichopic) 
in (B) Phiomia icintoni than in (.\) Phiomia minor. 

A, .\dult Phiomia minor ref. {.\mer. Mus. 13448), Five-thirtieths natuVal 
sije. .\88ociated tusk (Amer. Mus. I34()7). 

B, Adult Phiomia tvintoni ref. (.Viner. Mus. i:i4.50i. Four-thirtieths 
natural aize. 



Typical and Average Mea.surements (in 






















Millimeters) and Indices in Phiomia 


M 1-M 3 




M 1 




M2 




M3 




P3-M3 


Lower Molars 




ap. 


tr. 


I. 


ap. 


tr. 


I. 


ap. 


tr. 


I. 




: Phiomia osborni type (Amer. Mus. 13468) 


M,-M,,= 172 


41 


2() 


t>3 


63 


35 


56 


71 


38 


54 


P3-M, = 246 


Phiomiii wintoni averages 


Mi-M3=166 


41 


28 


68 


56 


35 


62 


70 


39 


56 


P,rM3 = 236e 


Phimnin minor averages 


M,-M,= 14o 


34 


23 


68 


47 


30 


64 


55 


35 


63 


P.rM3 = 201 


Upper Molars 
























Phiomia oshorni (not known i 
























Phininin wintoni averages 


M'-M'=162 


41 


33 


SO 


57 


41 


72 


61 


44 


72 


P'-M' = 214 


Phiomia minor (Amer. Mus. 13455) 


M'-M»=129e 


33.5 


27 


81 


46 


37 


80 


50 


37 


74 


P^-M' = 185 



Progressive narrowing of M3, prophetic of Amebelodon fricki (PI. v), is seen in a comparison of the indices 
of the lower molars, as given in the table above, al.so as follows: 



Phiomia osborni, ty]H' indk-cs (Amer. Mus. 13468), Mi: 63 M2: .")6 
Phiomia wintoni, rof. " (Amer. Mus. 13476), Mi: 67 Ms: 57 
Phiomia minor type " (Brit. Mus. 8479b), Mi: 66 M^: 64 



M:i: 54 (relatively narrow) 

Ms: 56 

Mj: 60 (relatively brosid) 



238 OSBORN: THE PROBOSCIDEA 

According to these extremes the progressive narrowing affects the anterior molars, Mi.2, as well as the pos- 
terior molar, M3. The narrow indices of all Phiomia teeth, however, are in wide contrast to the relatively broad 
indices in the corresponding teeth of Palseomastodon, as follows: 

P^wwia wintorw, average indices, Mi: 68 M2: 62 M3: 56 = narrower form ; average 55.5. M' average 74. 

Palspomastodon beadnelli , type indices, Mi: 77 M2: 78 M3: 68 = broader form ; average 69. M^ average 76. 

Progressive Bunomastodonty. — A bunoid conule arising (Fig. 185 Al) in the center of the valleys between 
the lophoid crests marks the beginning of the bunomastodont character so distinctive of all the longirostrines. 
It would appear that these central conules, destined to close up and fill the valleys entirely, arise progressively in 
the three species; at least they are much more clearly marked in the more progressive P. osborni than in P. winioni 
or P. minor. Consequently progressive bunomastodonty is a very distinctive generic as well as phyletic character 
of Phiomia; it tends to place these animals in the direct line of ancestry of Amebelodon fricki. 

Abundance of Phiomia Remains. — In the Fluvio-marine formation. Lower Ohgocene, Phiomia is as abun- 
dant as Palaeomastodon is rare, the ratio being about eleven to one, since the greater portion of the specimens re- 
ferred by Andrews to various species of Palseomastodon really belong to various species of Phiomia. The relative 
abundance of the three species of Phiomia is indicated by the American Museum collections in which the numbers 
of important specimens run as follows : 

Most progressive, Phiomia osborni, 1 specimen, 8 miles west of Quarry A 
Intermediate in size, Phiomia wintoni, 27 specimens, Quarries A, B, and at large 
Intermediate in size, Phiomia serridens, 1 specimen [=P. {?) wintoni] 
Small form, Phiomia minor, 19 specimens, Quarries A. B, and at large 

It would be difficult to give the exact figures from the British and Cairo Museums, but from Andrews' refer- 
ences to the species P. minor and P. unntoni, we deduce the following: 

Phiomia osborni: not recognized by Andrews 

Phiomia wintoni: verj' abundant, the prevailing species 

Phiomia minor : rare; 5 specimens in British and Cairo Museums 

Similar Savanna Habitat of Phiomia and Trilophodon. — The rarity of Palseomastodon is attributable to 
its belonging to the forest-living mastodonts ; the abundance of Phiomia is attributable to its having a flood- 
plain habitat similar to that of Trilophodon and of other longirostrines which are by far the most abundant 
proboscideans throughout Ohgocene, Miocene, and Lower Phocene times. Our inference is that Phiomia avoided 
the close forests and hved in a flood-plain and low savanna country, partly open, partly forested, subject to flood- 
ing along the river borders. This theory is fully treated in Chapter II on the "Origin and Classification of the 
Proboscidea." 

DISTINCTIONS BETWEEN THE THREE SPECIES 
Matsumoto's Definition.— (Matsumoto, 1924.1, p. 14) With sUght verbal changes (Osborn's terminology) 
Matsumoto's definition of the three species in ascending order is as follows: 

(1) Posterior end of mandibular symphysis situated a httle anterior to P3; in M,.., tritolophid distinctly feebler 
than protolophid and metalophid; M3 not very long and narrow, trilophodont, the posterior talon being not very 
prominent; length of lower molar series measuring 126 (Andrews' type) — 147 (Amer. Mus. 13471) mm.; that of 
lower premolar and molar series, 180 (Andrews' type) — 202 (Amer. Mus. 13471) mm. ; that of upper molar series, 
125-130 mm. (Amer. Mus. 13455, 13448) ; that of upper premolar and molar series, 205-215 mm. (Amer. Mus. 
13455, 13448) [no median conules; lower grinders relatively broad] Phiomia minor. 

(2) Mandibular symphysis and lower molars similar to those of Phiomia minor except in size; length of lower 
molar series, 160 (Amer. Mus. 13474)— 173 (Andrews) mm.; that of lower premolar and molar series, 225- 



THE AMEBELODONTINiE: PHIOMIA 



239 



250 mm. (Amer. Mus. 13474, 13477) ; that of upper molar series, 145-170 mm. (Andrews) ; that of upper premolar 
and molar series, 255-284 mm. (Amer. Mus. 13474, 13477) [inferior median conules) Phiomia winloni. 

(3) Posterior end of mandibular symphysis situated a little posterior to the anterior end of P3; in Mi.2 the 
third ridge almost as well developed as the first and second, the widest part of the tooth corresponding to the third 
ridge; M3 exceedingly long and narrow, nearly tetralophodont, the tetartolophid corresponding to the prominently 
developed posterior talon; length of lower molar series, 177-180 mm.; that of lower premolar and molar series, 
250-255 mm. (type, Amer. Mus. 13468) [j^rominent median conules; lower grinders relatively narrow]. 

Phiomia osborni. 



Phiomia minor .\iulrew.s, 1904 
For original description and type figure, see pp. 58, 59. -Mso PI. v 
Fluvio-niarine formation of the Fayilm, Egypt = Upper Eocene of .\ndrew8 
= Lower Oligocene of the present Memoir. 

This small and less abundant animal, inferior in size both to 
Phiomia serridens and P. irintoni and greatly inferior in size to 
P. osborni, was clearly defined by Andrews in his two descriptions 
as Palspomaslodon minor, as shown in the systematic revision, 
Chapter II. al)ove. Mafsunioto confirmed this species by referring 
to it nineteen specimens in the American Museum collection, the 



The detailed measurements of three mandibles are cited from 
Matsumoto in Chapter II, p. 59. 

Osborn, 1923: (1) Osborn reviews all the .\merican Museum 
1907 collection and con6rms all the references by Matsumoto, with 
the exception of the jaw (Amer. Mus. 13471) which he refers to 
Phiomia winloni. (2j In the description in the present Memoir of 
the Fayuin collection, Osborn adds original drawings of Phiomia 
made under his direction also his own observations and comparative 
measurements of Phiomia minor, P. winloni, and P. osborni. (3) 
(Jsborn differs from the final opinion of Andrews that Phiomia 



Measurements (in Millimeters) and Indices 


















OF Phiomia minor in the British 


















AND American Museums P 4 


Ml 






M2 






MS 




Phiomia minor 






















Lower Molars 




ap. 


tr. 


I. 


ap. 


tr. 


I. 


ap. 


tr. 


I. 


P. minor type, B.M. M.8479b 




32 


21 


66 


45 


29 


64 


48 


29 


60 


/'.W/ito^/ref., A.M. 13471, r.Mi-M3= 147 




33 


26 


79 


47 


32 


69 


64 


38 


59 


P. winloni ref., A.M. 13471, l.P3-M3 = 202 




33 


24 


73 


49 


31 


74 


65 


38 


58 


P. minor ref., A.M. 13469, l.Mi-M2= 79 




33 


21 


64 


46 


30 


65 








P. minor ref., A.M. 13469, r.Mi-M2= 83 




35 


25 


71 


48 


31 


65 








P. minor ref., A.M. 13475, r.M,-M2= ? 




38 


24 


63 


48 


28 


58 








P. minor ref., A.M. 13483, r.M2-M3= ? 










46 


31 


67 


55 


32 


58 


Average meas. and indices 






















G of Ml, 8 of Mj, 4 of M3 




34 1 


23.3 


68 


47.3 


30.2 


63+ 


54.7 


34.7 


63+ 


Upper Molars 






















P. minor ref., A.M. 13455, I.M'-M'=129e 




33.5 


27 


' 81 


43 


35 5 


83 


46 


38 


83 


P. minor ref., A.M. 13448, l.M'-M==128 


27 32 111 


29 


33 


114 


48 


37 


77 


53 


40 


75 


/-•. minor ref., A.M. 13448, r.M'-M'=127 


28 27 97 


28 


37 


1.32 


47 


38 


81 


53 


41 


77 


Average meas. and indices 






















3 of M', 3 of M2, 5 of M' 




30.3 


32.3 


107 


45.6 


36.8 


80 


50 


36.8 


73+ 



dental measurements of seven of which are shown in the above 
table. 

Comparing the above measurements of the upper and lower 
molars with those on page 237, we observe that the average indices 
of the first two inferior molars. Mi ■>, of /''. minor are closely similar 
to the average indices of the same teeth in P. wintoni; in other 
words, the proportions of Mi and M2 in these two species are the 
same. In M3 the case is different. This tooth is much broader 
in P. minor (index 60) than it is in P. winloni (index 56), or in P. 
osborni (index 54). Consefiuently we are justified in the inference 
that the third lower molar is progressively narrowing (dolicho- 
donty) as we ascend from P. minor to P. mintoni, to P. osborni. 



serridens probably represents a juvenile individual of the same 
species as P. minor, and considers that P. serridens more probably 
belongs to P. winloni. 

Phiomia serridens .\ndrews and Beadneli, 1902 
For original description and typ<> figure, see pp. 55, 56 

Fluvio-marine formation of the FajTini, Egypt = Upper Eocene of 
.\ndrews = Lower Oligocene of the present Memoir. 

Osborn (1924) observes that the juvenile genotype jaw of 
Phiomia serridens Andrews and Beadneli belongs to a sjjecies prob- 
ably, but not certainly, identical with Palieomaslodon { = Phiomia) 



240 



OSBORN: THE PROBOSCIDEA 



Phiomia wintoni 
Amer. Mus. 13450 Ref 




Phiomia minor 
Amer. Mus. 13455 Ref. 



2/3 nat. size 




Fig. 179. Superior dentition of the small (B) Phiomia minor and of the large (A) Phiomia irinloni, t\vo-third.< natural .size. 

A, Phiomia winloni (.\mer. Mus. 13450), referred by Matsumoto, 1922, p. S, to Phiomin irinloni. 

B, Phiomio minor (.\mer. Mus. 13455), referred by Matsumoto, 1922, p. 3, to Phiomia minor. 



(Referate) p. 1.57.'] has sugfrested that the specimen in question is 
the anterior portion of thenianciible of a very young, Palseoinastodon, 
a supposition which not impossibly may be correct, though since 
this is by no means certain it seems preferable to keep the genus 
Phiomia for the present and place it in its present position." 
Matsumoto concludes that while the specific name P. serridens 
antedates both the specific names P. ivintoni and P. minor, he 
Hnally (1924.1, pp. 14, 33) inclines to relate it to P. wintoni. 

Osborn, 1924: Osborn concludes that the type of Phiomia 
serridens probably belongs to the animal described by all authors 
as P. wintoni, but does not think it wise at present to regard P. 
wintoni as a synonym of P. serridens. 

JUVENILE MAXILL.A A TOPOTYPE OF PHIOMIA SERRIDENS 

Specific Resembl.^nce to P. wi.xtoni. — The accompanying 
figure of the type of Phiomia serridens combined with the juvenile 
ma.xilla (Amer. Mus. 13458) affords for the first time a more com- 
plete knowledge of the juvenile characters of P. serridens and proves 
that it may probabh' be regarded as belonging to the same animal 
as certain of the specimens referred to Phiomia uyintoiii in previous 
descriptions of Andrews and of Matsumoto. From measurements 
now assembled bj- Osborn this maxilla may for the present be 
named Phiomia wintoni (cf. serridens), at least until the synonymy 
of these two species is absolutely confirmed (see Fig. 181). 

Milk Dentition of Phiomi.\ wintoni (cf. serridens). — 
The right maxilla (Amer. Mus. 13458) figured below may be 
associated with the type jaw of Phiomia serridens for two reasons: 
(1) Because Dp' agrees exactly in size and relative measurements 
with the inferior tooth, Dp3, in the P. serridens type; (2) the 
characters of the milk teeth and first molar, Dp--M', are exactly 
what we should anticipate in an ancestor of Amehelodon; (3) the 
first superior molar, M', still buried in the jaw, measures ap. 43 
mm., tr. 30 mm., index 70; (4) these proportions of M', namely, 
ap. 43 mm., tr. 30 mm., index 70, are extremely close to those of 
the average P. wintoni, namely, ap. 41 mm., tr. 33mm., index 80, and 




Fig. 180. Second type figure of Phiomia scrrideiis Andrews and Bcadnell, 
1902. After Andrews, 1906, PI. xviii, fig. 4: "Phiomia serridens .\ndrew.s 
& Beadnell; anterior portion of left ramus of mandible, .side and (4a) upper 
views: three-fourths nat. size. Same horizon [Fluvio-marine beds (Upper 
Eocene) = Lower Oligocene of present Memoir). [Type specimen, C. 10007.] 
Page 170." Geological Museum, Cairo. Compare similar juvenile incisive 
dentition of Plattjbelndon grangeri (Fig. 433). 

winloni Andrews, 1905, rather than to the smaller form Palie- 
nmaslodon ( = Phiomia) minor Andrews, 1904. 

Hlstoric Notes.— (Andrews, 1906, p. 169): "Known only 
from an imperfect mandibular ranms, in which the single spatulate 
incisor is serrated on its outer margin. It was originally .suggested 
that this mandible might represent a very aberiant type of Creo- 
dont, but this determination is certainly wrong. Schlosser [Foot- 
note: 'Neucs Jahrb. f. Min. Geol. u. Palsont. 1905, vol. i. pt. i. 



THE AMKBELODONTINiE: PHIO.MIA 



241 



greater by far than tlie largest specimen of /■". minor (Ainer. Mus. 
13455) in size and index, namely, ap. 33.5 mm., tr. 27 mm., 
inde.x 81; (5) M' is relatively narrower in ten specimens of P. 
winloni and relatively broader in eleven specimens of /-*. minor. 
Consequently we conclude that the juvenile maxilla (Araer. Mus. 
13458) certainly belongs to the same animal as that hitherto 
described as I'hiomia wintoni; it probably i)elongs to the same 
specific stage as the type specimen of Phiomia serridens. 




2/3 ml ■>it 



JUVKNII.E DbNTITION OK PlllOMIA WINTONI U'f SKRHIDENS) 

Fig. 181. Fragineiil of ina.villa rofcirfd to t'hioinia utintoni (cf. 
aerrideiis) ;Vndre«s, Drawing (reversed) fioiii .\iupr. Mils. 13458. Two- 
thirds natural size. I'rom Quarry B, FayOiu, Egypt. Museum Expedition 
of lfi07. 

A, Al, Maxilla, crown and lateral views of Dp", Dp', Dp\ M'. 

B, Type lower jaw of Phiomia serridens Andrews and Beadnell. Orig- 
inal drawing after cast from Cairo Mu.s. C. 10007, containing Di", Dp-j, 
Dp.i; tile Dpa fits e.vactly into the Dp' of the ma.xilla. 

Juvenile maxilla (Fig. 181 A, Al): Dp- partly fracliired; 
Dp' tetrabunodont with postero-internal cone; Dp* distinctly 
trilophodont. the protoloph and metaloph broad, the tritoloph 
narrow; M' measurements given above. l)p^ agrees clo.-^ely in size 
and fits precisely into Dpa in the Phiomia serridens type, as shown in 
the accompanying illustration. 



The literature, however, would l)ecome so badly confused if 
we should make P. wintoni a .synonym of P. serrideng, that it seems 
best to await the final determination of this question through the 
discovery of more complete material. 

Milk Dentition of PniOMi.\SKRRiDKNsTypE.—(Cf. Andrews, 

19(}('), p. 170) Tyi)e jaw: Di-.. with straight inner border, convex 
outer border marked by about six crenulations, upper face concave, 
ventral face convex, longitudinal ridges, enamel confined to the end 
of the crown (enamel entirely wanting in I-j of P. wintoni) ; diastema 
5 cm. in length; D|)2 small, laterally compressed, higli, slightly 
bifid anterior cusp, small posterior cingular cusp; Dpa subtrilophid, 
convex anterior lobe tricuspid, posterior lobe bicuspid, small 
po.sterior tubercle. 



Phiomia wintoni .\ndrews. 1905 

For original description and type figure, .«€e pp. (10, 61. .\Imi PI. v, pp 
2.3.J 23t) 

Fluvio-marine formation of the Faj-um, Egjpt = Upper Eocene of Andrews 
= Lower Oligocene of the present Memoir. 

Syn.: Phiomia wintoni (.\ndrews, 1905) may prove to be the same animal 
as Phiomia serridems (Andrews and Beadnell, 1902). 

This animal, Phiomia wintoni, is directly intermediate in size 
between P. minor and P. osborni. The detailed measurements of 
the teeth taken by Matsuinoto and by Osboni demonstrate that 
these three species do not overlap each other; they probably rep- 
resent three stages of progressive evolution, but this is not yet 
demonstrated by the geologic records, because P. minor and /■". 
wintoni occur frequently in the same American Mu.seuin (juarries 
(AandB). This may be explained by the theory that the quarry 
depositions represent long periods of time. This is fully discussed 
below. 

Specific Evolution. — The relative frequency of the three 
species in quarries A and B may prove to have some bearing on 
the question whether Phimnia minor, P. wintoni, and P. osborni 
are (1 1 three contemporaneous species in dififerent degrees of evolu- 
tion; or (2) extremes or variations of sexual difference: or (3) 
wliether each of these quarries represents a very long period of 
geologic time during which these three specific stages evolved from 
each other. On the whole, the third supposition is the most 
probable one. 

Chakactehs of P111U.MIA WINTONI (Matsumoto, 1924.1, pp. 
24-40).— (1) Matsumoto and Andrews consider that the type speci- 
men of Phiomia serridens belongs to a very juvenile individual 
probably of /'. ici/i^od/, orpos-sibly of P. minor. (2) Detailed mea.-*- 
urements are given of eight jaws of P. wintoni, four descril)ed by 
.\iulrews and four referred l)y Matsmiioto, distinguishing the prob- 
able males from the probable females by a slight superiority in 
size. (3) Detailed measurements are given by Matsumoto of the 
teeth of eleven specimens, also distinguishing prol)able males from 
probable females by superior size. (4) Inferior tusks, e.g.. Amer. 
Mus. 134(52, measure 250 mm. in length, 44 mm. anteroposteriorly, 
21 mm. transversely; distinction is therefore made lietween the 
larger male anil smaller female tusks. (5) Tlie principal cranial 
measurements from four British Museum and four .\merican Mu- 
seum specimens are given as follows: 



242 



OSBORN: THE PROBOSCIDEA 



Basioccipitallength, condyles to premaxillaries: min. 605 mm., forty-six specimens referred to P. wintoni, in which it is again 



max. 700 mm. 

Length of palate: min. 350 mm., max. 430 mm. 
Width across zygomata: min. 377 mm., max. 447 mm. 
Height of skull, basioccipital to sagittal crest: 220 mm. 



attempted to distinguish males and females. 

Dental Proportions in Phiomia wintoni. — The detailed 
dental measurements and proportions prove to be of great value, as 
shown in the probable identification of Phiomia wintoni with P. 



Tusk, 

Amer. Mu5. 13467 Rot 



Tusk, 
Amer. Mus. 13467 Ref: 




Fig. 182. Referred aged skull of Phiomia minor and juvenile jaw of Phiomia wintoni. One-fourth natural size. 

.^1, Palatal view of adult skull of Phiomia minor, Amer. Mus. 13448 (referred by Matsumoto, 1922, p. 3, to P. minor). Tusk inserted (Amer. Mus. 13467, ref.). 

A2, Superior view of same skull, Amer. Mus. 13448. 

Bl, Referred jaw of P. wintoni, Amer. Mus. 13470 (referred by Matsumoto, 1922, p. 3, to P. wintoni), M3 not erupted. Observe central conules. 



(6) Superior tusks (Amer. Mus. 13466, 13463): Total length serridens above. The absolute measurements and dental indices 

220-250 mm.; anteroposterior diameter 50-44 mm.; transverse given in the table below, however, indicate that certain speci- 

diameter 31-28.5 mm. (7) Measurements of one to six of the mens properly referable to PAiomiaosftomi, e.g., Cairo Mus. C.8457, 

superior cheek teeth, longitudinally and transversely, are given in may be included in the Phiomia wintoni column. 



THE AMEBP:L()D0NTIN.E: PHIOMIA 



243 



Upper and Lower P. wintoni Molars 


M 1-M3 




Ml 






M2 






M3 




Phiomia wintoni 






















Lower Molars 




ap. 


tr. 


I. 


ap. 


tr. 


I. 


ap. 


tr. 


I. 


P. unntoni type, B.M. M.8414 


M2-M3 = 132 








55 


35 


04 


74 


44 


59 


P. wintoni ref., Cairo Mus. C.8456 




35 


28 


80 


57 


35 


til 


70 


41 


58 


; P. wintoni rcf., Cairo Mus. C.8457 ( =P. 






















osborni '!) 




38 


25 


66 


54 


3L 


57 


00 


35 


53 


P. wintoni ref., B.M. M.8384 




41 


28 


68 


56 


33 


59 


69 


}|) 


58 


P. wintoni rcf., B.M. M. 88.53 




41 


28 


68 


55 


35 


64 


73 


41 


56 


P. wintoni ref.. Amor. Mus. 13470 


M,-M3=166 


40 


27 


67 


56 


32 


57 


68 


38 


56 


Average meas. and indices 






















lOof ^L, 14 of Ms, 12of M3 




41 « 


28 


67 


50 2 


34 9 


02 


69 7 


39 


56 


Upper Molars 






















P. wintoni ref., B.M. M.9450 


M'-M'=158 


42 


30 


71 


57 


39 


OS 


01 


41 


07 


P. wintoni ref., Amer. Mus. 13451 


M'-M'=1G7 


47 


35 


75 


63 


45 


71 


t)2 


48 


77 


P. wintoni ref., Amer. Mus. 134."i7 


LM'-M'=15(i 


37 + 


34 


92+ 


56 


43 


77 


59 


45 


76 


Average nieas. and indices 






















8 of M', 19 of M^, 22 of M^ 5 of M'-M' 


M'-M» = 161.6 


41 


32.7 


80 


56.7 


40.7 


72 


60.7 


44.4 


73 



Phiomia minor 
Skull, Amer. Mus. 13448 Rot 




Amor. Mus 13467 Rel 



Phiomit wintoni 
Amor Mus. 13470 Rot. 



All 1/4 nti. sue 



Fig 183. Hefcrrod aged skull 
of Fhiumia minor and juvi'iiilo 
jaw of Phiomia wintoni. One- 
fourth natural size. 

Conipo.sition drawing: A, 
Phiomia minor skull (Amer. 
Mu.s. 1:J44.S), tusk (.\nier. Mus. 
13407): ]i, Phiomia iciriforu jaw 
(Amer. Mus. I.'}470). 



Average Measurements ano Indices of Piiio.mia wintoni 

ap. tr. I. 

Superior molars, M', 41 32.7 80 M=, 

Inferior molars, Mi, 41.0 28 07 Ms, 



ap. 
50.7 


tr. 

40.7 


I. 

72 


M', 


ap. 

00.7 


tr. 
44 4 


I. 

73 


56.2 


34.9 


62 


Ma, 


09.7 


39 


56 



The average measurements and indices of the upi^er and lower twenty-seven specimens of Phiomia wintoni in the American 

teeth of Phiomia wintoni as contnisted above demonstrate that Museum collection iletermineil l)y Matsumoto, it is interesting to 

the inferior molars (Mi-3=166 mm.) occupy longer and narrower examine the quarry records, 

spaces than the superior molare (M'"'= 161.6 nun.). Qiiarky Records. — From (Quarry B [which contains also ten 

Amehkan Museum M.\terials. — lu connection with the species of P/iio/H/rt w(/ior] are the following speciinen.s of P. H'/H^oni 



244 



OSBORN: THE PROBOSCIDEA 



ref.: Amer. Mus. Nos. 13474, 13476, 13477, 13484, 13494, 13494a, 
c, d, e, g, 13453, 13454, 13458 (P. wintoni cf. serridens), 13459, 
13479, 13488, 13489, 13491, 13527, 13527b, c, f, g, i, k, m; 
also superior and inferior tusks, Amer. Mus. Nos. 13462, 13463, 
13466c. Quarry A. [also containing remains of three specimens of 
Phiomia minor] yields the following numbers: Amer. Mus. 13494b, 
g, 13451, 13456, 13457, 13492, 13527a, h, j, n; also the tusks, Amer. 
Mus. 13466a. 




A.MDREWS' (1906) ERRONEOUS COMBINATION OF THE SkULL OF PhIOMIA 
WINTONI WITH THE MaNDIBLE OF PaL-EOMASTODON BEADNELLI 

Fig. 184. Andrews' restoration of Phiomia urintoni combined with jaw of 
Palscomnslodon hcadmili. After .\n(lrews, 1906. This restoration conforms 
closely to what we now know to be the characters of the skull of Phiomia 
uintoni, although the restoration of the lower jaw (founded on Palseomastodmi 
beadnelli) is relatively too long. 

First restoration (190G) of the "skull and mandible of Palsemnaslodon" 
published by .\ndrews, 190(1, p. 131, text fig. 4S, .\, B. .\ndrews did not 
specifically designate the specimen on which his restoration was based, but 
it would appear to be a combination of four specimens, namely: (a) The 
ty])e jaw of Pal^otinistodon beadnelli (op. oil., PI. x\', fig. 1), (h) the symphysis 
and incisors of Phiomia wiiitoni (o;j. cil., PI. .xiv, fig. 3), (c) the skull of Phiomia 
mnloni {op. cil., PI. xii) erroneously referred by .Andrews to Palseomastodon 
beadnelli, (d) the anterior portion of a young skull (PI. xiii). 

Osborn, 1923: Andrews' restoration of 190() is historically interesting as 
furnishing the basis of oiu- misconceptions of Pnhiomastodon which, as the col- 
lections of the .American Museum expeditions ami the studies of Matsumoto 
and Osborn have shown, are confusi'<l with Phiomin winlord, a typical 
Amcbelodont (PI. v, pp. 23.") 2.36). 

Thus Quarry B yielded twonty-ninc specimens of Phiomia 
wintoni and ten specimens of P. minor; Quarry A yielded eleven 
specimens of P. wintoni and three specimens of P. minor. 



CRANIAL CHARACTERS OF PHIOMIA WINTONI 
(Andrews, letter, 1922): Referred skull of Phiomia wintoiii. 
This skull is beautifully figured by Andrews (1906, Pis. xii, xiii, 
and XIV, fig. 3) and described on the plates under an erro^(«OM.S' refer- 
ence to Palseomastodon beadnelli if). This reference was subsequently 
corrected by Andrews in his own handwriting: . . . "no doubt 
pi. XII, pi. XIII, pi. XIV, fig. 3, are all P. wintoni." This skull is 
referred to Phiomia wintoni by Matsumoto (1924.1) and by Osborn 
(present Memoir). 

Osborn, 1923: The skull of this species is also important as 
belonging to the commonest of all the species in the Fluvio-marine 
beds ( = Lower Oligocene). The adult cranium (PI. xii), together 
with the anterior portion of a young skull (PI. xiii) and the type 
jaw, is obviously the basis of Andrews' semidiagrammatic figure of 
the skull {op. cit., p. 131, text fig. 48) as here reproduced (Fig. 184). 
Specific Characters. — Osborn, 1923: (1) Osborn leviews 
and confirms Matsumoto's observations on the fifty-five specimens 
in the American Museum 1907 collection referred to P. wintoni, 
adding one jaw (Amer. Mus. 13471) which Matsumoto referred to 
P. minor; (2) he confirms Matsumoto's observations, compares, 
describes, and figures the best specimens, clearly distinguishing the 
P. wintoni stage of evolution from that of P. minor and of P. 
o.'iborni, (3) adds new drawings and tables of comparative measure- 
ments made under his direction with the new definition of Phiomia 
wintoni, and (4) gives additional evidence from the detailed evolu- 
tion of the P. wintoni grinding teeth that this species is directly 
ancestral to Phiomia osborni [and (1935) to Amebelodon fricki]. 

Phiomia osborni Matsumoto, 1922 
For original description and type figure, see pp. 64, 65. Also Pis. v and vi 

Fluvio-marine formation of the Fayiim, Egypt = Upper Eocene of 
Andrews = Lower Oligocene of the present Memoir. 

This animal, represented by a single specimen dedicated to 
Osborn by Matsumoto, is larger and much more progressive than 
Phiomia wintoni in the tritolophid of the first and second inferior 
molars, M1-2, and in the tetartolophid of the third inferior molar, 
M3. It is also more progressive in the development of the central 
conules (bunomastodonty) arising directly from the floor of the 
valleys and wearing into characteristic trefoil loops, as shown in the 
enlarged P4, Mi (Fig. 185 Al) taken from the type. 

Specific Character.s.— (Matsumoto, 1922.1, 1924.1). The 
characters 1 to 11 specified by Matsumoto (1924, pp. 40-49) and 
cited in full in the systematic revision. Chapter II, pp. 64, 65, are 
all confirmed by Osborn. The reader is also referred to the con- 
densed definition of this species by Matsumoto in the present 
chapter, page 236. Osborn (1924) has little to add to the specific 
characters excepting to refer to the above measurements and 
dental indices printed in full in the present chapter, page 237, 
and to give in the table (p. 246) the comparative measurements and 
indices of Phiomia osborni, Serridenlmus serridens, S. productus, 
and Trilophodon angustidens. 

Progressive Dolichodonty. — The average breadth indcix of 
M3 in Phiomia o.sborni ( = 54) is reduced to 42.7 in Trilophodon 
anijuMidens: Serridentinus productus is intermediate. It will be 
observed that the index of M2 in S. productus ( = 60e) exceeds that 
of P. osborni ( = 56), whereas in T. angustidens it is relatively 
broader, average 65.7. 



abrasion areas 




2'3 nat. size 



Fig. 185. Second type figure of Phiomia osborni Matsiiraoto, Amer. Mus. 13468 (see Matsumoto, 192'2, p. 4, fig. 3, for original type 
figure— Fig. 40 of the present Memoir). Observe central conules in P^-M,•^; for detailed structure of r.M.i, sec figure I87B, p. 247. below. 

A, (Matsumoto, 0/;, at., p. 3): . . . "a nearly complete mandible, be.iring all the teeth i>i «((«." One-fourth niitunil size. Observe especially 
the superior view of the paired incisors showing abrasion area." apparently caused by prehen.silc pressure of the upper lip against these teeth. 
This specimen, first confused with PaLvomastodon itintoni, was selected by RIatsumoto ll922) as the type of the more progressive sptKiies 
Phiomia oshomi. Al, Dentition, P3-M3, two-thirds natural size, of P. osborni. Same type figure as A. 



Gf^ <£^ 






PHIOMIA MINOR 

95omm,, 3'p/2"e 

FAyOm N.AFRICA 



PHIOMIA SERRIDENS 

i03onini„3'4i'2"e 

FAyOm N AFRICA 



PHIOMIA WINTONI 

isoomm., 3'iii/4."e 

FAyOm N.AFRICA 



PHIOMIA OSBORNI 

I345mm..4'5'e 

FAyOm N.AFRICA 



PHiOMIA PYGMAEUS 

i70omm.,5'7'e 

ALGERIA 



Fig. lH5a. Restorations of the five species of Phiomia in iLscending geologic c)rder, one-hundredth natural size, by Margret Flinsch. under 
the direction of Henrv Fairfield Osborn. 



245 



246 



OSBORN: THE PROBOSCIDEA 



Phiomia, Serridentinus, and Trilophodon 


















Comparative Measurements (in Millimeters) 


















AND Indices of Lower Molars 


P4 


M, 






Mj 






Ms 






ap. tr. I. 


ap. tr. 


I. 


ap. 


tr. 


I. 


ap. 


tr. I. 


P. osborni type, A.M. 13468 


l.Mi-M3=172 
1. P3-M,, = 246 


42 27 64 


41 26 


63 


63 


35 


56 


71 


38 54 


5. serndens ref., A. M. 10673 


P4-M2 = 290 


57 43 75 


92 62 


67 


139 


69 


50 






S. prodiictus ref., A. M. 10582 










124e 


75 


60e 


163 


76 47 


T. angustidens ref., Simorre 










84 


55 


66 


131. 


5 59 45 


T. angustidens ref., Simorre 










89 


58 


65 


145 


59 41 


T. angustidens ref., Simorre 










93 


63.3 


68 


160 


68 43 


T. angustidens ref., Tournan 










94 


60 


64 


160. 


4 68 42 


T. angustidens ref., Sansan 
















152 


70 46 


T. angustidens ref., Sansan 
















181 


78 43 


Mounted skeleton 










105 






176 


68 39 



The above comparative measurements and indices of the 
third inferior molar in Phiomia osborni (Lower Oligocene), 
in Serridentrnus produdus (Lower Pliocene of Texas), in Tri- 
lophodon angustidens from the type locality of Simorre and 
from the somewhat older Sansan beds, demonstrate the propor- 
tional relations of P. osborni to the extremely long-jawed or hyper- 
longirostral T. angustidens; also progressive narrowing of M3. 

Condensing the dental indices of these three species in ascend- 
ing geologic order, we observe the following: 

Trilophodon angustidens, proportions of Ma 

Serridenlinus produdus M2 

Phiomia osborni M2 

Dental Adaptation. — The two new figures of the type jaw 
of Phiomia osborni prepared for this Memoir (Figs. 90, 185) 
illustrate very clearly the adaptation of the pair of closely applied 
procumbent incisors and of the five parallel grinding teeth, also of 



the greatly elongated symphysis in contrast to the shorter 
structure of the jaw in Palseom,astodon beadnelli. First we observe 
the abrasion areas on the upper surface of the incisors which demon- 
strate that the upper lip pressed against these teeth in the prehen- 
sion of food, wearing off the outer surface of the dentine, because 
the enamel does not persist on these teeth. At the extremity of the 
right incisor, r. I2, is a worn groove which demonstrates that these 
teeth were used like a trowel in uprooting plants. To accommodate 
these greatly enlarged spatulate incisors, which functioned like a 

ap. 

160.4 

163 
71 

broad shovel,' the symphysis is increasingly elongated backwards 
until it reaches the space between r. P3 and 1. P3, whereas in P. 
beadnelli, an animal of the same size, the symphysis does not 
extend far backwards. 



ap. 


tr. 


L 




94 


60 


64 


M3 


124e 


75 


60e 


M3 


63 


35 


56 


M3 



tr. 


L 


68 


42 


76 


47 


38 


54 



5. TRANSITION FROM PHIOMIA OSBORNI TO PHIOMIA PYGM^US (COMPARE PAGES 276, 277) 



Phiomia pygmseus Dep6ret, 1897 

Figures 186, 187 
Cartennien de Kabylie, Algeria, near Isserville. Upper (?) Oligoeeue. 

Mastodon angustidens Cuv. mut. asc. pygmxus Deperet, 1897. 
"Decouverte du Mastodon angustidens dans I'fitage Cartennien 
de Kabylie," Bull. Soc. gM. de France, (3), XXV, pp. 518-521. 
Type. — Third inferior molar, M3 (cast Amer. Mus. 17728). 
Horizon and Locality. — Near Isserville (Kabyhe), Algeria, 
fitage Cartennien. Type Figure. — Op. cit., PI. xix, 

figs. 1,2,3. 

Type Description. — "La molaire en question (pi. xix) n'est 
pas entiere; elle est bris^e en avant, et de ce cote il manque une 
rangfe de mamelons; de sorte que la couronne compte seulement 
trois collines, plus un talon bien prononce. L'etroitesse g^n^rale 
de la couronne et la forme du talon etroit, triangulaire et pourvu 
seulement de deux tubercules, indiquent qu'il s'agit d'une molaire 
infdrieure et fort probablement d'une derniere ou sixieme molaire. 



en raison de la grandeur du talon. Chacune des trois collines se 
compose de deux gros mamelons principaux, arrondis, coniques, 
assez 6troit«ment accoles I'un a I'autre, mais qui ont une tendance 
a se dedoubler en dedans sous forme d'un mamelon accessoire moins 
^leve que le mamelon principal. Les vallees qui separent les 
diverses collines sont remplies d'une epaisse masse de cement du 
sein duquel on voit ^merger a peine les mamelons d'^mail, qui, en 
raison de la jeunesse du sujet, n'etaient pas encore entam^s par la 
detrition [op. cit., p. 519]." 

COMPARISON WITH PAL^OMASTODON AND PHIOMIA 
Osborn, greatly profiting by the rich Fayum collection in the 
American Museum, makes a very close comparison of Dep^ret's 
type of 'Mastodon' angustidens pygmxus with the corresponding 
third lower molar of Phiomia osborni type (Amer. Mus. 13468) 
and a characteristic third lower grinder referred to Palxomastodon 
beadnelli (Amer. Mus. 13481), as very carefully illustrated in figure 
187 A, B, C. 



'This passage was written by Osborn two years before the discovery by Barbour of the Nebraska 'shovel-tusker' Amebelodon fricki, which proves to be a 
direct successor to Phioynia osborni. See Plates v and vi, also Appendix at the close of the present Volume for the revised Amebelodontinse. 



THE AMEBELODONTIN^v: PHIOMIA 



247 



The type of Phiomia pygtrueus (original type figure 186, new- 
type figure 187 A-A4) appears to resemble (juite elosely the type of 
Phiomia osborni and to differ widely from the type and referred spec- 
imens of Palseomasiodon beadnelli. Consequently Osborn removes 
this species from the genus MaModon and from the subfamily 
Mastodontina' and places it in an intermediate position lietween 
Phiomia osborni and Trilophodon angustidens.' It may be observed 
that in P. pi/fimaus, as in P. osborni: (1) The summits of the cones 
are closely compressed towards the central apices of the respective 
lophs; (2) the summits of the cones are sharply divided into 'cone- 
lets,' a pair on either cone, (3) in contrast with Palsenmastodon 
beadnelli in which the summits of the cones are widely separated; 







Tyi'E OF Phiomia pygm.bus 
FiR. 186. Type of Mastodon angustidens Ciiv. miit. asc. pygm^us 
Dcpdrpt, 1897, two-thirds natural size. (Dep<^ret, 1897, PI. xix, figs. 
1-3, p. 521): "Demiere molaire inf(5rieure (bris^e en avant) de Mastodon 
angustidens Cuv. mut. asc. pygmanis Dep. dii Cart«nnien de Kabylie; 
vue par des,sus (1) et par les cM^s interne (2) et externe (3). — Grandeur 
naturclle." Compare figure 187A (Phiomia pygmasiis). Cast Amer. Mus. 
17728. 

(4) the type of P. pygmppua also has the elongate, narrow propor- 
tions of P. osborni and of Trilophodon angustidens, rather than the 
broader proportions of the grinder in Palxoviastodon and Mastodon; 

(5) the beginning of the tetartolophid is indicated in two cones 
in P. pygm^us and P. osborni, demonstrating a progression towards 
the well-developed tetartolophid of T. angmtidens. 

Osborn, 1925: Conclusion.^ From the above observations and 
figures, we conclude that Phiomia pygnurus is an ascending muta- 
tion beyond the stage of P. osborni in the direction of the much more 
highly developed stage of Trilophodon angustidens, in which the 
third inferior molars possess not only a complete tetartolophid but 
a rudiment of a pentalophid. 




Phiomia pygmaeus 
Amer. Mus. 1 7728 Type 




Phiom<9 osborn, 
Amer. Mus. I346B Type 




Pal«omastodon beadnslli 
Amer Mus 13481 rvl. 



Fig. 187. Comparison of (B) Phiomia osborni, (A-A4) tjTJe of Phiomia 
pygniaiis, and (C) Pnlsomaslodon headneUi third inferior molars. r.Mj, all 
to the same scale, two-thirds natural size; compare PI. i .\, pp. 134-13.>. 

A, .\1-A4, original new drawings of Dep(5ret's type of Mastodon angusti- 
denx Cuv. n)ut. asc. pygmseus, 1897 \= Phiomia pygmstus of the present 
Memoir], after cast of type (Amer. Mus. 17728). (A) crown, (Al) cxtemaJ, 
(A2) internal, (A3) anterior, and (A4) posterior views. In all four aspects 
this type is seen to clo.sely resemble (B), the type of Phiomia osborni. 
Compare figure 186, original type figure of M. angustide