(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"


m 


_ ^ mi 














. />P-^ 




^^^A^)*}^\ii^^ f^^H 


\ 


1 ^■Mfy^^^^^^^^^^^^^^^^^^^^H 


^^B 1 


f/ 1 


1'/^ ,^J.J\ 




■Hi 




(J iy-^l 


i 


I ■, 1 1 ■; ■ , ; . : . i ' - 




,,<""Uli„ -. 



\, ■ 



y . 



■-1/"^ 



I 



iiniii^nn 




World Distribution of Stegodonts, N 




WO -fathom , Plio - Plelstocerie 
Present coast line 

•■■\\V i "--r;V . v-t'- : ' i ;/.''''-'';--'---.;v'-"-r/-'.:--:'-:--:i'^>:^f-> 



30° 45° 60° 75° 90"E 105" 120° 135° 150° f 



Steg ol ophodon t/nae 
O sterol op ho don 

Removed to the Mastodontidae too 
late for plotting on Map, Fie;. 680, Vol. I. 



Stegodont/nae 

(D Stegodon 

AIAA1A4 O/VT/ nXe 

% Archidiskodon 



At A A 

® Me. 
© Pa? 
® Aia 



0TH5, LOXODONTS, AND ELEPHANTS 




165 



-^^ .W ..n°"^.n5° 90V 75° 60 45 



VT//\/AE 

idiskodon 

has 

nteus 



jr o^OOOA/T/NAi ^LEPHANT/NAE 

@ £ojcodonta © ZLephas 

e Paldeolopcodon ® Jiupselephas 

O Hesperolo^codon ® /^/afe/^^A g<^^ _^^^^ 



LIBRARY OF THE 
X^ FOR THE <^> 

^ PEOPLE ^ 

-ff FOR ^ 

^ EDVCATION O 

r^ FOR ^J' 

n SCIENCE ^ 



PROBOSCIDEA 




RESTORATION OF THE JEFFERSONIAN MAMMOTH {PARELEPHAS JEFFERSONII) 

After a painting by Charles R. Knight in 1909, under the (Urertion of Henry Fairfield Osborn 

The skeleton on which this painting was based is moulded in the American Museum of Natural Histuri/, Hull of 
the Age of Man. It was found near Jonesboro, Indiana, on the farm of Dora E. Gift, in 1903; purchased for the 
American Museum with the Jesup Fund in 1904; mounted in 1906; first described and figured by the present 
author in 1907. As found the skeleton was embedded in a muck deposit of late Pleistocene age, fifteen feet below 
the surface. This deposit is probably of post{?)-W i sconsi n age {according to the geologic time scale having been laid 
doivn about 15,000 years ago). Representatives of the Parelephas phylum appeared in Europe in the early Pleistocene 
and persisted into the Third Interglacial. This Third Inlerglacial period may mark the time of migration across 
Asia into North America. In fact, it is suggested by the present author that such migration might hare occurred in 
Second or eve^i in First Interglacial time. 

The most striking features of this individual are the complete incurvation and cro.ssing of the tu.'iks, indicating 
that it is an old bull, and the relatively small size of the head. It is here represented with a hairy covering, as Parele- 
phas is characteristic of the north temperate region , both of Europe a nd the ( 'n ited States. 



PROBOSCIDEA 

A MONOGRAPH OF THE DISCOVERY, EVOLUTION, MIGRATION 

AND EXTINCTION OF THE MASTODONTS AND 

ELEPHANTS OF THE WORLD 



BY 

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. Christiania, 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 Palaeontologist, 

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 II 

STEGODONTOIDEA 

ELEPHANTOIDEA 




PUBLISHED ON THE J. PIERPONT MORGAN FUND BY THE TRUSTEES 

OF THE AMERICAN MUSEUM OF NATURAL HISTORY 

THE AMERICAN MUSEUM PRESS 

NEW YORK, 1942 



PUBLICATION NOTE 

Volume I of this work was issued August 15, 1936. The present volume, 
containing the Stegodontoidea and Elephantoidea as well as tables, conclusions, 
and general index, has been prepared from the materials left by the late author 
(see Publication Note to Volume I). 



Copyright, 1942, by 
The American Museum of Natural History 






CONTENTS 

PAGE 

LIST OF ILLUSTRATIONS xv 

CHAPTER 

XIV. THE ROOF-TOOTHED STEGODONTS, SUPERFAMILY STEGODONTOIDEA 805 

Classification of the genera Stegolophodon and Stegodon 807 

History of discovery of the subfamily Stegodontinae. Principles of type revision of species 815 

The Stegodontinse and Mastodontinse of China 816 

Pliocene to Pleistocene Proboscidea of Japan 818 

Phylogenetic discussion of the thirty described species of Stegodon ts and Stegolophodonts 819 

Probable European — Asiatic origin and migration of the primitive Stegodonts 822 

Type revision of the species in order of original discovery and description 822 

First two Stegodonts discovered in Burma (1928) 825 

Mastodon latidens Clift, 1828 [ = Stegolophodon latidens] 827 

Mastodon elephantoides Clift, 1828 [ = Stegodon elephantoides] 828 

Discoveries in India and Burma 829 

Stegodonts of China, India, Java, the Philippine Islands, Austria, Japan, and Burma 831 

Characters of the subfamily Stegodontinae 837 

Stegolophodon Schlesinger, 1917, generic definition 839 

cautleyi Lydekker, 1886 840 

latidens Clift, 1828 842 

sublatidens Schlesinger, 1917 846 

stegodontoides Pilgrim, 1913 846 

nathotensis Osborn, 1929 847 

cautleyi progressus Osborn, 1929 848 

lydekkeri Osborn, 1936 851 

Stegodon Falconer and Cautley, 1847, 1857, generic definition 853 

sinensis Owen, 1870 860 

elephantoides Clift, 1828 861 

bombifrans Falconer and Cautley, 1846 863 

insignis Falconer and Cautley, 1845, 1846 866 

ganesa Falconer and Cautley, 1845, 1846 869 

insignis-ganesa 874 

insignis birmanicus Osborn, 1929 874 

orientalis grangeri Osborn, 1929 875 

pinjorensis Osborn, 1929 883 

orientalis Owen, 1870 884 

airdwana Martin, 1890 885 

ganesa var. javanicus Dubois, 1908 889 

trigonocephalus Martin, 1887 890 

mindanensis Naumann, 1890 892 

auroi-B Matsumoto, 1915, 1918 892 

orientalis shodoensis Matsumoto, 1924 893 

bondolensis van der Maarel, 1932 894 

trigonocephalus praecursor von Koenigswald, 1933 896 

(Parastegodon?) kwantoensis Tokunaga, 1934 897 

yiishensis Young, 1935 897 

officinalis Hopwood, 1935 898 

zdanskyi Hopwood, 1935 899 

(Parastegodon) sugiyarnai Tokunaga, 1935 899 

Matsumoto on the phylogeny and classification of the Japanese Mastodonts, Stegodonts, and Elephants 901 

Osborn's comments (1929) on Matsumoto's phylogeny and classification of 1924-1927 908 

XV. CLASSIFICATION OF THE ELEPHANTOIDEA BY THEIR DIVERGENT AND HARMONIC CRANIAL 

AND DENTAL CHARACTERS 911 

Elephantoidea Osborn, 1921, superfamily definition 912 

Elephantidae Gray, 1821, family definition 912 

vii 



i OSBORN: THE PROBOSCIDEA 

AFTER XV — Continued page 

Failure of previous dental classifications 914 

Classification by cranial and dental ciiaracters 914 

Cranial mechanics of Elephas (Weithofer, Osborn, Gregory) 915 

Comparative cranial sections of elephant skulls 918 

Ontogenetic cranial changes in Elephas indicus 919 

Dental and cranial adaptation to prevailing feeding habits the key to phylogenetic classification 927 

Ridge-plate formulae of primitive and progressive genera in adaptation to prevailing habits of feeding 927 

Food of the Indian and African elephants and of the mammoth 927 

Seasonal changes in food of the mammoth 929 

Summary of progression from browsing to grazing dentition 929 

Vertebral distinctions of Elephas, Loxodonta, Mammonteus, and Parelephas 930 

Vertebral formulae 930 

Synopsis of subfamily classification of the Elephantoidea 932 

XVI. THE GENUS ARCHIDISKODON (SUPERFAMILY ELEPHANTOIDEA), MOST PRIMITIVE MEMBER OF 

THE SUBFAMILY MAMMONTIN^ 935 

History of the subfamily Mammontinae 935 

Mammontinse Osborn, 1921, subfamily definition 937 

History of the genus Archidiskodon 939 

Older of discovery and description of twenty-two species of Archidiskodonts 942 

Archidiskodonts of Eurasia and America 943 

New Archidiskodonts and Loxodonts of Africa 944 

Approximate phylogenetic order of succession of species of Archidiskodon and Parelephas (1928) 946 

Archidiskodon Pohlig, 1885, 1888, generic definition 947 

planifrons Falconer and Cautley, 1846 [1845] 950 

Measurements of twenty-seven specimens collected by Barnura Brown in the Siwaliks 954 

Leith-Adamsia siwalikiensis 959 

planifrons rumanus Stefanescu, 1924 968 

meiidionalis Nesti, 1825 969 

lyrodon { = A. meridionalis female] 977 

Durfort skeleton 977 

meridionalis cromerensis Deperet and Mayet, 1923 980 

Archidiskodonts and Metarchidiskodonts of South Africa 983 

Archidiskodon (continued) 

proplanifrons Osborn, 1934 986 

subplanifrons Osborn, 1928 987 

broomi Osborn, 1928 989 

vanalpheni Dart, 1929 990 

milletti Dart, 1929 991 

loxodontoides Dart, 1929 991 

yorki Dart, 1929 992 

M etarchidiskodon Osborn, 1934, generic definition 994 

griqua Haughton, 1922 994 

Archidiskodonts of the United States and Mexico 996 

Archidiskodon (continued) 

imperalor Leidy, 1858 998 

imperator silveslris Freudenberg, 1922 1015 

imperalor falconeri Freudenberg, 1922 1016 

El. Coluinbi var. imperator Freudenberg, 1922 1017 

maibeni (skeletal characters) 1019 

hayi Barbour, 1915 1023 

imperator srolti Barbour, 1925 1025 

imperator maibeni Barbour, 1925 1027 

haroldcooki Hay, 1928 1029 

eiilis Stock and Furlong, 1928 1031 

sonoriensis Osborn, 1929 1033 

meridionalis nebrascensis Osborn, 1932 1033 



CONTENTS ix 

CHAPTER PAGE 

XVII. THE GENUS PARELEPHAS (SUPERFAMILY ELEPHANTOIDEA), OF THE SUBFAMILY MAMMON- 

TIN^, INTERMEDIATE BETWEEN ARCHIDISKODON AND MAMMONTEUS, DISTRIBUT- 
ED IN THE NORTH TEMPERATE ZONE OF EURASIA AND NORTH AMERICA 1039 

European north temperate origin. History of separation from other extinct proboscideans 1039 

Order of discovery and description of species of Parelephas 1047 

Parelephas Osborn, 1924, generic definition 1048 

trogontherioides Zuffardi, 1913 1055 

trogontherii Pohlig, 1885, 1888-1891 1056 

Irogontherii nestii Pohlig, 1891 1059 

armeniacus Falconer, 1857 1060 

intermedius Jourdan, 1861 1062 

wusti Pavlow, 1909 1065 

North and South American species of Parelephas 1067 

jacksoni Mather, 1838 1068 

{1)rmnfiissippiensis Foster, 1872 1070 

Columbian Mammoth (Parelephas columhi) 1070 

columhi Falconer, 1857, 1863, 1868 1071 

texlanus [ = coluvihi] Owen, 1859, Blake, 1861, 1862 1073 

Cohen Collection (Phosphate Beds of South Carolina) 1075 

Amherst skeleton 1079 

columhi felicis Freudenberg, 1922 1082 

columbi cayennensis Osborn, 1929 1083 

jeffersonii Osborn, 1922 1083 

Elephas roosevelti [ = Parelephas jeffersonii] Hay, 1922 1095 

progressus Osborn, 1924 1097 

washingtonii Osborn, 1923 HOI 

eellsi Hay, 1926 1104 

floridanus Osborn, 1929 1105 

XVIII. THE GENUS MAMMONTEUS (SUPERFAMILY ELEPHANTOIDEA), OF THE SUBFAMILY MAM- 

MONTINiE, THE TRUE NORTHERN WOOLLY MAMMOTH 1117 

Sceleto Elephantino Tonnse, 1695, confused with the mammoth, Blumenbach, 1799 1118 

Breyne's description (1735) of the Elephas primigenius of Siberia 1119 

Names successively apphed to the mammoth 1120 

Native Siberian origin of the word Mammut 1124 

Ides (1706) 1124 

Howorth (1882) 112^ 

Mammonteus Camper, 1788, Osborn, 1924, generic definition 1126 

External characters and feeding habits 1127 

Skeletal characters of Mammonteus primigenius 1129 

Historical order of naming of species of Mammonteus exclusive of species which are now known to belong 

to Parelephas trogontherii, etc ^^"^^ 

Aurignacian mammoth hunters of Moravia 1139 

Typical progressive Eurasiatic stages of Mammonteus 1140 

primigenius Blumenbach, 1799, 1803 1141 

Elephas odontotyrannus [ = M. primigenius] Eichwald, 1835 1146 

Primitive European stages of Mammonteus primigenius 1149 

primigenius leith-adamsi Pohlig, 1888 ^^^" 

primigenius hydruntinus Botti, 1891 -^ ^^ 

primigenius fraasi Dietrich, 1912 

primigenius astensis Dep6ret and Mayet, 1923 11^4 

Forest Bed, or Cromerian, fauna H^^ 

American stages of Mammonteus ^^ 

primigenius americaiius DeKay, 1842 



X OSBORN: THE PROBOSCIDEA 

CHAPTER XVIII — Continued page 

primigenius compressus Osborn, 1924 1157 

prirnigenius alaskensis sp. nov 1159 

Frozen mammoth of Siberia 1162 

Summary of the discovery and natural history of the woolly mammoth 1163 

XIX. THE GENERA LOXODONTA, PALiEOLOXODON, AND HESPEROLOXODON OF THE SUPERFAMILY 

ELEPHANTOIDEA, SUBFAMILY LOXODONTIN^ 1171 

Classification and histoiy of discovery of the Loxodontinse 1173 

Difficulties of generic nomenclature 1174 

Order of discovery and description of the fifty-three type species of the extinct Loxodontinae 1187 

Loxodontinae Osborn, 1918, subfamily definition 1191 

Loxodonta F. Cuvier, 1825, 1827, generic definition 1191 

Order of description of eighteen living African species and subspecies 1192 

Loxodo7ita (continued) 

africana Blumenbach, 1797 1 197 

africana {?)cottoni Eales, 1926-1929 1202 

comaliae Aradas, 1870 1204 

Palxoloxodon Matsumoto, 1924, generic definition 1207 

namadicus Falconer and Cautley, 1846, 1847 1211 

Hesperoloxodon Osborn, 1931, generic definition 1217 

aniiquus Falconer and Cautley, 1847, 1857 1217 

Upnor skeleton of H. antiquus 1222 

Coadaptation of the vertebral column with the superior incisive tusks of the elephants 1228 

Hesperoloxodon (continued) 

antiquus nanus Acconci, 1880 1230 

antiquus plaiyrhynchus Graells, 1897 1231 

antiquus ausonius Major, 1875, Verri, 1886, Dep6ret and Mayet, 1923 1232 

antiquus germanicus StefSnescu, 1924 1233 

Hesperoloxodon antiquus italicus of Southern Italy and H. antiquus germanicus of North Central Germany .... 1238 

antiquus italicus Osborn, 1931 1245 

aniiquus germanicus of Steinheim 1253 

Extinct dwarfed species of the Mediterranean Islands 1257 

Palseoloxodon (continued) 

melitensis Falconer, 1862 1262 

falconeri Busk, 1867 1263 

mnaidriensis Adams, 1870 1265 

lamarmorae Forsyth Major, 1883 1266 

Cypriotes Bate, 1903 1266 

creticus Bate, 1907 1267 

The dwarfed elephants of the Mediterranean Islands and the question of Pleistocene isthmuses (Vaufrey, 1929) . 1268 

Ancestral stages of Palasoloxodon in Africa 1273 

atlanticus Pomel, 1879 1274 

jolensis Pomel, 1895 1274 

recki Dietrich, 1916 1275 

Palxoloxodon and Loxodonta of South Africa 1277 

Palxoloxodon (continued) 

andrewsi Dart, 1929 1278 

hanekami Dart, 1929 1279 

yorki Dart, 1929 1280 

wilmani Dart, 1929 1280 

kuhni Dart, 1929 1281 

urchidiskodonloides Haughton, 1932 1282 

iransvaulensis Dart, 1927 1284 

sheppardi Dart, 1927 1285 



CONTENTS xi 

CHAPTER XIX — (Continued) page 

Loxodonta (continued) 

Zulu Scott, 1907 1286 

prima Dart, 1929 1287 

africana var. obliqua Dart, 1929 1287 

subantiqua Haughton, 1932 1288 

Loxodontines of Japan and Java 1289 

Summary of Matsumoto's final observations and theories of 1924 and 1929 1290 

Osborn's summary (1930) of the observations of Makiyama (1924) and of Matsumoto (1924-1929) 1292 

Two Japanese subspecies described by Makiyama (1924) 1293 

Pabeoloxodon namadicus naumanni Makiyama, 1924 1295 

Palseoloxodon namadicus namadi Makiyama, 1924 1296 

Five Japanese Loxodontines described by Matsumoto 1297 

Palseoloxodon protomammonieus Matsumoto, 1924, 1926 1297 

Palxoloxodon tokunagai Matsumoto, 1929 1298 

Palseoloxodon protomammonieus proximus Matsumoto, 1926 1298 

Palseoloxodon namadicus yabei Matsumoto, 1929 1299 

Palseoloxodon (Archidiskodonf) tokunagai mut. junior Matsumoto, 1929 1299 

Japanese species described by Saheki and Tokunaga (1931, 1934) 1300 

Parelephas protomammonieus malsumotoi Saheki, 1931 1300 

Palseoloxodon yokohamanus Tokunaga, 1934 1301 

Javanese species described by Dubois 1302 

Palseoloxodon hysudrindicus Dubois, 1908 1302 

Geographic distribution along the eastern coast of Asia 1304 

XX. THE SUBFAMILY ELEPHANTINE (SUPERFAMILY ELEPHANTOIDEA), OF EASTERN ASIA, IN- 
CLUDING THE RECENT ELEPHAS AND RELATED LOWER PLEISTOCENE SPECIES 1307 

Historical introduction and nomenclature (850? B.C. — 1936) 1308 

Falconer (1868) on the specific unity and vertebral formulae of the Asiatic elephants 1312 

Corse, de Blainville and Falconer on characters of the geographic varieties 1313 

Elephas indicus Isodactylus Hodgson 1313 

Elephas indicus Heterodactylus Hodgson 1313 

Elephas indicus ceylanicus de Blainville 1313 

Elephas indicus bengalensis de Blainville 1313 

Elephas indicus (Dauntela var.) 1314 

Elephas indicus (Mukna var.) 1314 

Living specific or subspecific forms, continental and insular 1315 

Fossil forms more or less closely related to Elephas indicus 1318 

Names of species and subspecies of the subfamily Elephantinse in order of description 1319 

Elephantinse Osborn, 1910, subfamily definition 1320 

Elephas Linnaeus, 1735-1758, generic definition 1322 

indicus Linnaeus, 1735-1754, collective species 1323 

indicus ceylanicus de Blainville, 1845 1327 

indicus bengalensis de Blainville, 1845 1327 

indicus sumalranus Temminck, 1847 1329 

indicus hirsutus Lydekker, 1914 1332 

indicus Buski [=? Palseoloxodon buski] Matsumoto, 1927 1333 

Distinctions and measurements of the Indian elephant 1334 

Characters of the Upper Pliocene and Lower Pleistocene species Hypselephas hysudricus and Platelephas plat- 

ycephalus 1339 

Hypselephas Osborn, 1936, generic definition 1340 

hysudricus Falconer and Cautley, 1845, 1846 1340 

Observations of Osborn on fourteen specimens collected by Barnum Brown in the Siwaliks 1345 

Cranial characters and affinities 1348 

Platelephas Osborn, 1936, generic definition 1358 

plaiycephalus Osborn, 1929 1359 

Appendix to Chapter XX 1362 



xii OSBORN: THE PROBOSCIDEA 

CHAPTER PAGE 

XXI. NOMENCLATURE OF THE PROBOSCIDEA 1363 

The American Mastodon {Mastodon americanus) 1363 

The Northern or Woolly Mammoth {Mammonteus primigenius) 1365 

List of superfamilies 1367 

List of famiUes 1368 

List of subfamilies 1369 

List of genera 1371 

List of species, subspecies, and varieties 1382 

XXII. THE GEOLOGIC SUCCESSION OF THE PROBOSCIDEA 1421 

Africa 1422 

Eocene and Oligocene of North Africa 1423 

Miocene of North Africa 1426 

Miocene of Central and East Africa 1428 

Pleistocene of North Africa 1429 

Pleistocene of Central and East Africa 1432 

Pleistocene of South Africa 1437 

The Orient 1439 

Miocene of Baluchistan and Sind 1439 

The Siwalik series (Miocene-Pleistocene) of North India 1442 

Pleistocene of Central India 1447 

Pleistocene of Ceylon 1450 

Pleistocene of Burma 1450 

Pleistocene of South China 1452 

Pleistocene of Indo-China 1453 

Pleistocene of the East Indies 1453 

Europe 1456 

Lower Miocene : Burdigahan 1457 

Middle Miocene : Helvetian and Tortonian — Vindobonian 1460 

Upper Miocene : Sarmatian 1464 

Lower Pliocene : Pontian 1466 

Middle Pliocene : Plaisancian 1469 

Upper Pliocene : Astian 1470 

Pleistocene 1472 

Asia 1477 

Miocene of MongoUa and Central Asia 1477 

Miocene of North China 1479 

Pliocene of Mongolia 1481 

Pliocene of North China 1481 

Pleistocene of North China 1483 

Miocene to Pleistocene of Japan 1490 

North America 1490 

Upper Miocene: Barstovian 1491 

Lower Pliocene : Clarendonian 1495 

Middle Pliocene : Hemphillian 1502 

Upper Pliocene : Blancan 1503 

Pliocene of Mexico 1506 

Proboscideans from undetermined levels in the Miocene and Pliocene of North America 1507 

North American Tertiaiy horizons containing fragmentary proboscidean remains 1508 

Pleistocene of North America 1510 

Pleistocene of Mexico 1515 

Central and South America 1516 

Pliocene of Central America 1516 

Pleistocene of Argentina 1516 

Pleistocene of the Andean valleys 1519 

Pleistocene of Brazil and French Guiana 1521 



CONTENTS xiii 

CHAPTER PAGE 

XXIII. AFFINITIES, MIGRATIONS, AND PHYLOGENY OF THE PROBOSCIDEA: A SUMMARY 1523 

Five superfamilies 1524 

Eight families 1525 

Twenty-one subfamilies 1526 

Forty-four genera 1526 

Valid species (352) 1527 

Osborn's final (1935) classification of the Mceritherioidea, Deinotherioidea, and Mastodontoidea 1529 

Osborn's final (1935) classification of the Stegodontoidea and Elephantoidea 1539 

Explanation of terms used throughout the text of the present Memoir 1545 

Characters, affinities, and migrations of the Proboscidea 1552 

Moeritheres 1552 

Deinotheres 1553 

Longirostrines (genera, Trilophodon, Megabelodon; subgenera, Genomastodon, Choerolophodon, Tatabelodon).. 1555 

Gnathabelodonts (Gnathabelodon) 1557 

Amebelodonts {Amebelodon, Phiomia) 1558 

Tetralophodonts (genera, Tetralophodon, Morrillia; subgenus, Lydekkeria) 1559 

Notorostrines {Cordillerion) 1560 

Rhynchorostrines {Rhynchotherium, Blickotherium, Aybelodon) 1561 

Brevirostrines {Anancus, Penlalophodon, Synconolophus) 1563 

Humboldtines (Cuvieronius, Eubelodon, Stegomastodon) 1566 

Serridentines (Serridentinus, Ocalientinus, Se.rbelodon, Trobelodon) 1568 

Platybelodonts {Platybelodon, Torynobelodon) 1570 

Notiomastodonts (Notiomastodon) 1572 

Palifiomastodonts {Palxomaslodon) 1572 

Mastodonts {Mastodon, Miomastodon, Pliomastodon) 1574 

Zygolophodonts {Zygolophodon, Turicius) 1575 

Stegolophodonts (Stegolophodon) 1578 

Stegodonts (Stegodon) 1579 

Mammontines {Archidiskodon, Metarchidiskodon, Parelephas, Mammonteus) 1582 

Loxodontines (genera, Loxodonta, Palseoloxodon, Hesperoloxodon; Sivalikia, Pilgrimia, synonyms of 

Palxoloxodon) 1590 

Elephantines {Elephas, Hypselephas, Platelephas) 1595 

Skeletal material 1600 

Heights of proboscideans, estimated and actual 1604 

APPENDIX TO VOLUME II. PROBOSCIDEAN DENTAL HISTOLOGY 1607 

LIST OF CATALOGUE NUMBERS OF SPECIMENS MENTIONED IN THE PROBOSCIDEA MEMOIR, 

VOLUME II 1609 

SUPPLEMENTARY BIBLIOGRAPHY 1623 

INDEX TO VOLUMES I AND II 1631 



LIST OF ILLUSTRATIONS 

PLATE PAGE 

Frontispiece. Restoration of the Jeffersonian Mammoth {Parelephas jeffersonii). After painting by Knight. 

XIII. Stegolophodontinae : Stegolophodon 850-851 

XIV. Migration and evolution of Mceritherium and Deinotherium 1552-1553 

XV. Migration and evolution of Trilophodon, Megabelodon, Gnaihabelodon, Phiomia, Amehelodon, Tetralophodon, and 

Morrillia 1556-1557 

XVI. Migration and evolution of Cordillerion, Rhynchotherium, Blickotherium, Aybelodon, Anancus, Pentalophodon, and 

Sxjnconolophus 1560-1561 

XVII. Migration and evolution of Eubelodon, Cuvieronius, and Stegomastodo7i 1566-1567 

XVIII. Migration and evolution of Trobelodon, Serbelodon, Serrideniinus, Ocalientinus, Plalybelodon, Torynobelodon, and 

Notiomastodon 1568-1569 

XIX. Migration and evolution of Palxomastodo7i, Miomaslodon, Pliomastodon, Mastodon, Turicius, Zygolophodon, and 

Stegolophodon 1572-1573 

XX. Migration and evolution of Stegodon 1578-1579 

XXI. Migration and evolution of Archidiskodon and Metarchidiskodon 1582-1583 

XXII. Migration and evolution of Parelephas, Mammonteus, and Elephas 1584-1585 

XXIII. Migration and evolution of Loxodonta, Palieoloxodon, and Hesperoloxodon 1590-1591 

XXIV. Osborn-Reeds Correlation Table of 1922 and 1929 1606-1607 

XXV. Northwestern India and adjacent territory, showing especially the Siwalik Hills where Falconer and Cautley made 

their classic collections and Dr. Barnum Brown recovered in 1922 the fine Proboscidea collections for the 
American Museum 1606-1607 

XXVI. Elephas indicus: (Fig. 1) Complete transverse section near tip of unerupted tusk. (Fig. 2) Area of thin section 

within circle on Fig. 1. (Fig. 3) Same as Fig. 2, photographed between crossed nicols 1630-1631 

XXVII. Elephas indicus: (Fig. 1) Area within circle on PI. xxvi, Fig. 2. (Fig. 2) Transverse section of about half of a small, 

mature tusk about 14 inches long. (Fig. 3) Part of section mthin circle on Fig. 2. (Fig. 4) Same as Fig. 3, with 
crossed nicols. (Fig. 5) A small area near outer edge of cement band of section shown in Figs. 2-4. (Fig. 6) 
Phiomia wintoni: Transverse thin section about one-half inch from tip of small, worn tusk 1630-1631 

XXVIII. Phiomia wintoni: (Fig. 1) Same as PI. xxvii, Fig. 6, with crossed nicols. (Fig. 2) Area within circle on PI. xxvii, 

Fig. 6 1630-1631 

XXIX. Trilophodon obscurus: (Fig. 1) Transverse section of part of tusk. (Fig. 2) Area within circle on Fig. 1. Crossed 

nicols. (Fig. 3) Same tusk as Figs. 1 and 2, thin section across enamel cut in a plane vertical to surface and 
parallel to longitudinal axis of tooth. Crossed nicols. (Fig. 4) Trilophodon (Megabelodon) sp. Transverse thin 
section of tusk near edge of enamel band. (Fig. 5) A different part of same thin section as Fig. 4. Crossed 

nicols 1630-1631 

XXX. Trilophodon (Megabelodon) sp.: (Fig. 1) Vertical thin section of broken molar tubercle. (Fig. 2) Part of section in 
circle on Fig. 1. Crossed nicols. (Fig. 3) Transverse thin section of a molar cusp, from same tooth as Figs. 1 
and 2 but a different cusp. (Fig. 4) Part of section in inner circle on Fig. 3. (Fig. 5) Part of section in outer 
circle on Fig. 3 1630-1631 

FIGURE 

681. Scene on the ancient Solo River, illustrating the Dietrich-Osborn theory of the Middle Pleistocene age of Pithecanthropus 

erectus. Restoration by Flinsch 804 

682. Family group of Stegodon orientalis grangeri. Restoration by Flinsch 804 

683. Stegodon elephantoides ( = cliftii), cotype molar. After Lydekker 808 

684. Stegodon orientalis grangeri, referred first right superior molar 808 

685. Structural evolution of the cones, conelets, and ridge-crests in the Stegolophodon phylum, in comparison with Paheomastodon . 810 

686. Structural evolution of the cones, conelets, and ridge-crests in the Stegodon phylum 811 

687. Gradual progressive hypsodontj' in superior grinders of Stegodon 813 

688. Gradual progressive hpysodonty in superior and inferior grinders of Stegodon 813 

689. Turicius compared with Trilophodon molars. After Mayet 819 

690. Turicius and Stegolophodon form of grinding teeth 821 

691. Map showing geographic distribution of types and referred specimens of Stegolophodon and Stegodon 823 

692. Fossil-bearing horizons along the Irrawaddy River, Burma 824 

693. Stegolophodon latidens, lectotype palate. After Clift 826 

694. Stegolophodon latidens, cotype third right inferior molar. After Clift 826 

XV 



xvi OSBORN: THE PROBOSCIDEA 

FIGtTRE PAGE 

695. Elephas cliftii Falconer and Cautley equals Stegodon elephantoides ( = cliflii), cotype first left superior molar. After Clift . . . 826 

696. Stegodon elephantaides, lectotype lower jaw. After Clift 826 

697. Stegodon insignis, lectotype and cotype molars. After Falconer and Cautley 829 

698. Stegodon ganesa, lectotype third superior molar. After Falconer and Cautley 829 

699. Stegodon bombifrons, cotype skull. After Falconer and Cautley 830 

700. Elephas cliftii, original type figure of first left superior molar, equals Stegodon elephantoides { = cliftii). After Clift 831 

701. Elephas cliftii, new type figure, equals Stegodon elephantoides ( = cliftii). After photograph of type caSt 831 

702. Stegodon sinensis, type third superior deciduous premolar. After Owen 831 

703. Stegodon orientalis, type. Portion of true molar and posterior end of milk molar. After Owen 832 

704. Stegolophodon cautleyi, lectotype third left superior molar. After Lydekker 832 

705. Stegodon trigonocephalus, type immature skull. After Martin 833 

706. Stegodon (Archidiskodon?) mindanensis, type molar. After Naumann 833 

707. Stegodon airmmna, type lower jaw. After Martin 834 

708. Stegolophodon stegodontoides, type right third superior molar. After Lydekker 834 

709. Stegodon aurorse, type right second superior molar. After Matsumoto 834 

710. Stegolophodon snblatidens, type molar. After photograph 835 

711. Stegodon pinjorensis, type palate. After photograph 835 

712. Stegolophodon cautleyi, lectotype left third superior molar. After Lydekker 841 

713. Stegolophodon cautleyi, lectotype left third superior molar. After Falconer and Cautley 841 

714. Stegolophodon cautleyi, cotype first superior molar. After Lydekker 842 

715. Stegolophodon cautleyi, referred left second superior molar. After Lydekker 842 

716. Stegolophodon latidens, lectotype (new figure) right second and third superior molars. Orthogonal projection after cast. . 843 

717. Stegolophodo7i latidens, cotype third inferior molar. After Clift 843 

718. Stegolophodon latidens, lectotype right second and third superior molars. Perspective. After Clift 844 

719. Stegolophodon latidens, lectotype molars. Section after Falconer and Cautley, in comparison with Mastodon americanus, 

third superior molar 844 

720. Stegolophodon latidens, lectotj^je third superior molar. After Gaudry 844 

721. Stegolophodon latidens, referred third right superior molar from Japan. Primitive stage. After Matsumoto 845 

722. Stegolophodon sublatidens, type. Posterior half of a third right superior molar. After photograph 846 

723. Stegolophodon stegodontoides, type third right superior molar. After Lydekker 847 

724. Stegolophodon nathotensis, type fragmentary molars. American Museum (Barnum Brown) collection 848 

725. Stegolophodon cautleyi pi-ogressus, type cranium. American Mu.seum (Barnum Brown) collection 849 

726. Stegolophodon cautleyi progressus, type right superior dentition. American Museum (Barnum Brown) collection 849 

727. Stegolophodon cautleyi progressus, type skull, left lateral, palatal, and right lateral aspects. After photograph 850 

728. Stegolophodon lydekkeri, type third left superior molar. After Lydekker 851 

729. Falconer's map of the geology of India 852 

730. Map of chief Miocene and Pliocene fo.ssil mammal depo.sits of A.sia. After Osborn 853 

731. Species of Stegodon from India, China, and Java. Restorations by Flinsch 855 

732. Stegodon ganesa, S. ifisignis, and S. bombifrons crania. After plates by Falconer and Cautley 856 

733. Stegodon ganesa, cranium with tusk extremities turned inward, in contrast to Falconer and Cautley's restoration (Fig. 732) 

with tusks turned outward and closely appre.ssed 857 

734. Stegodon bombifrons, lectotype. A generalized cranium. After Falconer and Cautley 858 

735. Stegodon insig7ns, a specialized cranium. After Falconer and Cautley 858 

736. Stegodon ganesa, a specialized male cranium. Restoration after Falconer and Cautley 858 

737. Stegodon sinensis, type third right superior deciduous premolar. After Owen 860 

738. Stegodon elephantoides, lectotype second and third left inferior molars. After Clift 861 

739. Stegodon elephantoides Clift { = cliftii Falconer). Cotype first left superior molar. After cast 862 

740. Stegodon elephantcndes C\iit { = cliftii Falconer). Cotype first left superior molar. After Falconer and Cautley 862 

741. Stegodon elephantoides { = cliftii), referred third left inferior molar. After Falconer 863 

742. Stegodon bombifro7is, lectotype, cotype, and referred crania. After Falconer and Cautley 864 

743. Stegodon bombifrons. Restoration by Flinsch 864 

744. Stegodon bombifrons, cotype .skull. After I'\alconer and Cautley 865 

745. Stegodon bombifrons, referred third right superior molar. After Falconer and Cautley 866 

746. Stegodon bombifrons, referred fragment of third right superior molar. American Museum (Barnum Brown) collection. . . 866 

747. Stegodon insignis, lectotype and cotype molars. After Falconer and Cautley 867 

748. Stegodon insignis, referred superior and inferior molars. American Museum (Barnum Brown) collection 868 



LIST OF ILLUSTRATIONS xvii 

FIGURE PAGfJ 

749. Slrgodun insignis, referred second right superior molar. American Museum (Barnum Brown) collection 869 

750. Sk-godon insignis, referred inferior mandil)lc. American Museum (Barinnn Brown) collection 869 

751. Stegodon insignis, referred juvenile and young adult lower jaws. American Museum (Barnum Brown) collection 870 

752. Stegodon insignis, referred crania. After Falconer and Cautley 870 

753. Stegodon insignis, referred third left inferior molar. American Museum (Barnum Brown) collection 871 

754. Stegodon insignis, oft-reproduced referred skull and tusks. After Falconer and Cautley 871 

755. Stegodon ganesa. Restoration by Flinsch 872 

756. Stegodon insignis. Restoration by Flinsch 872 

757. Sectioned molars (lectotypes) of Stegodon ganesa and Stegolophodon latidens. After Falconer and Cautley 874 

758. Stegodon insignis binnanicus, type ramus with third left inferior molar. American Museum (Barmim Brown) collection. 875 

759. Stegodon orientaUs grangeri, type and referred molars. American Museum (Walter Granger) collection 876 

760. Stegodon insignis-ganesa ref., <S. insignis hirmanicus type, and S. orientaUs grangeri type. Comparison of left third inferior 

molars 877 

761. Stegodon orientaUs grangeri, referred superior and inferior molars 878 

762. Stegodon orientaUs grangeri, type and referred superior and inferior molars. After photographs 879 

763. Stegodon orientaUs grangeri, infantile, juvenile, young adult, and mature adult crania 880 

764. Stegodon pinjorensis, S. insignis, S. airaivana, and S. orientaUs grangeri, sections of third left superior molars 881 

765. Stegodon pinjorensis, type skull. American Museum (Barnimi Brown) collection 882 

766. Stegodon insignis, referred lower jaw and right superior tusk (American Museum, l^arniun Brown, collection), compared 

with S. ganesa, referred skull with tusks, after Falconer and Cautley 882 

767. Stegodon pinjorensis, type skull, also same skull superimposed on referred skull of S. ganesa 883 

768. Stegodon pinjorensis. Restoration by Flinsch 883 

769. Stegodon orientaUs, type. Portion of true molar and posterior end of milk molar. After Owen 884 

770. Stegodon insignis{?) = orientaUs{f) ref. and Serridentinus lydekkeri type. Molars, after Schlo.sscr 885 

771. Stegodon airawana, type lower jaw. After Martin 886 

772. Map of Kendeng horizon, Trinil, Java. After Dubois 887 

773. Stegodon airawana, referred skull. After Janensch 888 

774. Stegodon airehvann. Restoration by Flinsch 889 

775. Stegodon trigonocephahis. Restoration by Flinsch 890 

776. Stegodon trigonocephahis, type skull, after Martin, and single ridge-crest of molar, doubtfully referred by Naumann to S. 

insignis or to <S'. ganesa 890 

777. Stegodont crania: Stegodon pinjorensis type, S. bornbifrons cotype, S. orientaUs grangeri ref., S. trigonocephahis type, S. 

airdwana ref 891 

778. Stegodon (Archidiskodonf) mindnnensis, type inferior molar (incomplete). After Naumann 892 

779. Stegodon airdwana, section of referred second molar, and section of ridge-crest of third superior molar. After Janensch. . 892 

780. Stegodon aiirorse, type second right superior molar. After Matsumoto 893 

781. Stegodon aiirora\ type second right superior molar (vertical section). Aft(>r i)hotograph 893 

782. Stegodon bondolensis, type mandible with third molars in siti(. After yiin d(>r Maarel 895 

783. Stegodon trigonocephahis praecwsor, type third left inferior molar. After von Koenigswald 896 

784. Parastegodon? kwantoensis, ty])e lower jaw with right second molar. After Tokunaga 897 

785. Stegodon yiishensis, type third left superioi- molar. After Young 897 

786. Stegodon offieinaUs, type molar fragment. After Hopwood 898 

787. Stegodon oflkinaU's, referred molar fragment. After Hojiwood 898 

788. Stegodon zdanskyi, type third right inferior molar. After Hopwood 899 

789. Parastegodon [Stegodon?] sugiyamai, type, probably a left second superior molar. After Tokunaga 900 

790. Thirteen fossil mammal-bearing formations of Japan. After Matsumoto 902 

791. Theoretic jjhylogeny of the Mastodontidse. After Matsumoto 903 

792. Theoretic phylogeny of the Stegodonts. After Matsumoto 904 

793. Theor(>tic phylogeny of the Elephantidae of Asia and Europe. After Matsiunoto 905 

794. Elephanti,d£e : Primitive, intermediate, and progressive mandibles and grinding teeth. After Falconer and Cautley 910 

795. General climatic distribution of the subfamilies of the Elephantoidea and Stegodontoidea, including theoretic migration 

lines (1938) 914 

796. Asiatic elephant, juvenile cranium. After Osborn and Gregory 916 

797. Asiatic elephant, juvenile cranium, also orbitosphenoidal region, left side. After Osborn and Gregory 917 

798. Asiatic elephant, infantile cranium, basis cranii 917 

799. Asiatic elephant, infantile cranium, occiput and jaws 918 



xviii OSBORN: THE PROBOSCIDEA 

FIGUKE ■ PAGE 

800. Asiatic cipphant {Ehphas indicus hengalen.sis), adult cranium, ])alatal view 918 

801. Five wire section lines. Key to sections. Young Elephas indicus bengalensis cranium 919 

802. Asiatic elephant, young. Interior view of skull, after Gregory 920 

803. Elephas indicus cranium. Fronto-occipital growth curves of vertex 921 

804. Elephas indicus cranium. Growth ciu'ves of vertex: midfrontal vertical section 921 

805. Preliminary study of fronto-occipital-basilar planes: Elephas indicus, Loxodonta africana (adult skulls), and Parelephas 

jeffersonii (young male skull) subsequently referred to Archidiskodon imperator 922 

806. Later study of fronto-occipital-niolar-3 planes. Para-occipitofrontal section of Loxodonta africana, Parelephas jeffersonii 

male and female (the latter subsequently made the type of Mammonteus primigenius compressus by Osborn), and 

Elephas iyidicus 922 

807. Elephas indicus, superior nasal growth stages (sections) 923 

808. Elephas indicus, transfrontal growth stages (sections) 923 

809. Elephas indicus, oceipitohorizontal growth stages (sections) 923 

810. Mid-occipitofrontal sections, vertical longitudinal : Crania of Loxodonta africana, Elephas indicus, and Parelephas jeffersonii 

[ = Archidiskodon imperator] 924 

811. Frontal .■sections: Crania o( Loxodonta africana, Elephas indicus, and Parelephas jeffersonii [ = Archidiskodon imperator] . . . 924 

812. Nasal contours (sections) : Loxodonta africana and Elephas indicus compared with Parelephas jeffersonii and Archidiskodon 

imperator 925 

813. Midfrontal or intertemporal forehead (Loxodonta africana, Elephas indicus, and Parelephas jeffersonii) 925 

814. Oceipitohorizontal sections through liack of occiput (Loxodonta africana, Elephas indicus, Parelephas jeffersonii) 925 

815. Map. Successive habitats and world migration routes of the archaic-toothed mammoth Archidiskodon 934 

816. Mid-cranial axes (I/jxodontines, Elephantines, Mammontines) 937 

817. Archidiskodon and Stegodon, comparative profiles of crania 938 

818. Hypsicephalic crania of the Mammontinse (Archidiskodon imperator, Mammonteus primigenius, Parelephas jeffersonii , and 

P. xcashingtonii) 939 

819. Comparative series of superior molars showing evolution of the ridges in the Elephantoidea and Stegodontoidea (Mam- 

monteus primigenius compressus, Archidiskodon planifrons, Stegodon aurorse, S. ganesa, S. insignis) 939 

820. Map of central region of the Siwalik Hills, 200 miles south and north of Simla 940 

821. Map of chief Lower and Upper Pleistocene localities of western Eurasia in which occur species of Archidiskodon, Parelephas, 

Mammonteus, Loxodonta, and Palseoloxodon. After Osborn 941 

822. Map showing geographic distribution of the principal species of Archidiskodonts 942 

823. Vaal River gravel terraces, South Africa 945 

824. Archidiskodon imperatw maibeni, family group along the Platte River, Nebraska. Restoration by Flinsch 947 

825. Archidiskodon planifrons, lectotype right second superior molar, and cotype left third inferior molar. After Falconer and 

Cautley 951 

826. Map of favorable exposures, southwest of Simla, of the Archidiskodon planifrons life zone, U))iier Siwaliks, India, eliiefiy 

Pinjor horizon 952 

827. Archidiskodon planifrons, new lectotype figure of right second sujx'rior molar (Miss Woodward's drawing) 952 

828. Archidiskodon planifrons, lectotype right second superior molar. After Lydekker 953 

829. Archidiskodon planifrons, referred right third superior molar. After Falconer and Cautley 953 

830. Archidiskodon planifrons, ])rofile of skull. Ke])roduced from (laudry after Falconer and C'autley 953 

831. Archidiskodon planifrons, referred first left superior molar. American Museum (Barnum Brown) collection 956 

832. Archidiskodon planifrons, referred first left and right sujierior molars. American Museum (Barnum Brown) collec-tion. 956 

833. Archidiskodon planifrons, referred second left sujjerior molar. American Museum (Barnum Brown) collection 956 

834. Archidiskodon planifrons, referred third left superior molar. American Museum (Barnum Brown) collection 956 

835. Archidiskodon planifrons, referred left third inferior molar. American Museum (Barnum Brown) collection 956 

836. Archidiskodon. planifrons, portion of refei-red fiist and second right inferior molars. American Museum (Barnum Brown) 

collection 957 

837. Archidiskodon planifrons, ref(>rred fourth inferior deciduous premolar. American Museum (Barnum lirown) collection. . 957 

838. Archidiskodon planifrons, fourth left inferior deciduous premolar and first left inferior molar. American Museum (liarnum 

Brown) collection 957 

839. Archidiskodon planifrons, portion of lower jaw with I'iglit third molar in situ. American Mu.seum (Barnum Blown) col- 

lection 957 

840. A rchidi.><kodon planifrons, port'iou of lower jaw with right third molar in situ. .Vmerican Museum (Barnum Brown) 

collection 957 

841. Archidiskodon planifrons, referred left third inferior molar. American Museum (Barnum Brown) collection 958 



LIST OF ILLUSTRATIONS xix 

FIGURE l'A(;i'; 

842. AirhifliskodoN plunifrons, roforrod tliird liglil inferior molar. American Mvisoum (liarniiin Urown) eolleetion 958 

843. Archuliskodoii phinifrons, referred third left inferior molar. American Miuscum (Barnum Brown) collection 958 

844. Archidiskodon planifrons, portion of referred lower jaw with left third molar in situ. American Museum (Bariumi Brown) 

collection g.^g 

845. Archidiskodon planifro?is, referred right third inferior molar. American Museum (Barnum Brown) collection 959 

846. Archidiskodon planifrons, referred left third inferior molar, also transverse vertical section. American Museum (Barnum 

Brown) collection 959 

847. Leith-Adamsia sixmUkiensis, type third right superior molars [ = synonym of Archidiskodon planifrons]. After Falconer 

and Cautley qqq 

848. Archidiskodon planifrons, referred cranium (supposed female). Redrawn after Falconer and Cautley 960 

849. Archidiskodon planifrons, three referred primitive mandibles from the Siwaliks, India, Chagny-Belleeroix and Seneze, 

France. After Mayet and Roman 962 

850. Archidiskodon planifrons of C'hagny, referred tusks, maxillae, condyle, and atlas. After Mayet and Roman 963 

851. Map of type locality of Archidiskodon proplanifrons, and type and referred localities of .4. planifrons; also theoretic range 

from supposed African center northward to France and Britain and eastward to India 964 

852. Archidiskodon playiifrons of Piltdown, England. Molar fragments, after Smith Woodward 965 

853. Archidiskodon planifrons, molar fragments from Piltdown, England. Redrawn by Miss G. M. Woodward 966 

854. Archidiskodon planifrons, lectotype sixth, seventh, and eighth superior ridge-plates. After dra\ving by Miss G. M. Wood- 

ward 9(56 

855. Archidiskodon planifrons of Piltdown, England (sectioned molar), in compari.son with lectotype and referred molar sections . 967 

856. Scene on the ancient river Ouse, illustrating the Osborn theory of the Upper Pliocene age of Eoanthropus dawsoni. Restora- 

tion l)y Flinsch 968 

857. Archidiskodon planifrons rumaniis, type third left inferior molar and portion of referred third right inferior molar. After 

Stefanescu 969 

858. Archidiskodon mei-idionalis, lectotype cranium (C). After Nesti 970 

859. Archidiskodon meridionalis, cotype cranium (A). After Nesti 971 

860. Archidiskodon meridionalis, restored cotype cranium (A). After Weithofer 971 

861. Archidiskodon meridionalis, lectotype cranium (C). After Weithofer 972 

862. Archidiskodon meridionalis, referred third inferior molars from the Val d'Arno and Norwich Crag. After Falconer and 

Cautley 973 

863. Archidiskodon meridionalis, referred third superior molar from Chagny, France. After Gaudry 974 

864. Elephas lyrodon [ = Archidiskodon meridionalis, female], type skull. After Weithofer 975 

865. Crania (19) of the Mammontinse {Mammonteus primigeniiis, Parelephas trogontherii , Archidiskodon meridionalis, A. impera- 

tor, and A. planifrons). After Pohlig, Falconer, Weithofer 976 

866. Archidiskodon meridionalis of Durfort, skeleton, largely restored. After photograph 978 

867. Archidiskodon meridionalis of Durfort. Restoration by Flinsch 979 

868. Archidiskodon meridionalis of Durfort. Skeleton redrawn after jjhotographic plate in Gaudry 979 

869. Archidiskodon meridionalis, referred third left (?) superior molar from Essex, England. After Lydekker 980 

870. Archidiskodon meridionalis crotnerensis, type third left superior molar. After Deperet and Mayet 980 

871. Archidiskodon, Parelephas, Mammonteus: Primiti^•e grinding teeth from the Forest Bed and Red Crag, England. After 

pencil sketches by the present author 981 

872. Map showing fossil Proboscidea route from southern equatorial Africa 983 

873. Archidiskodon proplanifrons, type right third superior molar. After original 987 

874. Archidiskodon subplanifrons, Osborn's original type figure of third right inferior molar 987 

875. Archidiskodon subplanifrons, type right third inferior molar (new figure) 988 

876. Archidiskodon planifrons, referred third right inferior molar 988 

877. Archidiskodon broomi, Osborn's original type figure of third right inferior molar 989 

878. Archidiskodon vanalpheni, type third left superior molar. After Dart 990 

879. Archidiskodon milletti, type third left superior molar. After Dart 991 

880. Archidiskodon lo.rodonloides, type third left superior molar. After Dart 992 

881. Archidiskodon yorki, type molar fragment. After Dart 993 

882. Metarchidiskodon griqua, type third left superior molar. ( )riginal figure of Haughton and new figure after Osborn 995 

883. Metarchidiskodon griqna, referred fragmentary molar, from Kaiso Bone-beds, Africa. After photograph 995 

884. Archidiskodon imperalor, tyi:)C fragment of a third right superior molar. Leidy's original typo figure 999 

885. Archidiskodon imperator, type third right superior molar. After Osborn 999 

886. A rchidiskodon imperator, Leidy's type molar and Osborn's neotype combined. After Osborn 999 



XX OSBORN: THE PROBOSCIDEA 

FIGUHE PAGE 

887. Archidiskoclon imperator, type molar compared with type molar of Parelephas rohinihi. After i)liotograpli 1000 

888. Archidiskodon imperator, ncotypc third right siip(>rior molar. After Osborn 1000 

889. Archidiskodon imperator, referred third superior and inferior molars of two individuals, showing mechanical reversal of the 

convex and concave surfaces. After O.sborn 1001 

890. Map showing distribution of Archidiskodon imperator west of the Mississippi River. After Hay 1003 

891. Archidiskodon imperator, referred cranium of young male from Texas. American Museum Cope Collection 1004 

892. Archidiskodon imperator, referred mandible, in comparison with mandibles of Parelephas jejfersonii 1006 

893. Mandibles of Elephas indicus, Loxodonta africana, Parelephas irashingtonii, and Archidiskodon hayi 1006 

894. Archidiskodon imperator, right superior tusk of record size, from Post, Texas, combined with a superb tusk formerly in the 

National Museum of Mexico, now destroyed. After photograph 1007 

895. Archidiskodon imperator, referred young male(?) skull from tar pools of Rancho La Brea, C'alifornia 1007 

896. Archidiskodon imperator, referred skull from Victoria, Texas 1008 

897. Archidiskodon imperator, referred skull from Victoria, Texas, as mounted in the American Museum 1009 

898. Archidiskodon imperator, referred jaw from Tule Canon, Texas 1010 

899. Archidiskodon imperator, crushed skull of aged male (Nebraska Museum) from Hay Springs, Nebraska 1011 

900. Archidiskodon imperator, crushed skull of adult male in the American Museum, from Hay Springs, Nebraska 1011 

901. Archidiskodon imperator, referred third right superior molar, from Zumpango, Mexico. After photograph 1013 

902. Archidiskodon imperator, referred mature male cranium, from Tepexpan, Mexico. After photograph 1014 

903. Archidiskodon hayi (?), referred mandible in Ceological Institute of City of Mexico. After jihotograph 1014 

904. Archidiskodon imperator silvestris, type third left superior molar. After Freudenbcrg 1015 

905. Archidiskodon imperator fakoneri , cotype jaw after Freudenberg. Originally figured by \'illada 1016 

906. Archidiskodon imperator, referred. Aged male. Reconstruction 1017 

907. Archidiskodon imperator, referred right forelimb, as mounted in the American Museum 1018 

908. Archidiskodon imperator (young adult) and Loxodonta africana oxyotis ("Jumbo"), referred limb bones 1018 

909. Archidiskodon imperator. After restoration by Osborn and Knight, 1908 1019 

910. Archidiskodon imperator maibeni, forclimbs of type skeleton, also cranium and tusks of Parelephas jeffersonii 1020 

911. Archidiskodon imperator maibeni, mounted type skeleton in Morrill Hall, University of Nebraska 1021 

912. Shoulder heights of living and extinct elephants (Elephas indicus, Loxodonta africana oxyotis, Archidiskodon imperator, A. 

imperator maibeni) 1022 

913. Archidiskodon hayi, tyi)e jaw. After Barbour 1024 

914. Archidiskodon planifrons, primitive mandibles from the Siwaliks, India, Chagny-Bellecroix and Seneze, France. After 

Mayet and Roman 1024 

915. Archidiskodon hayi, type mandible, compared with A. imperator ref 1025 

916. Archidiskodon imperator, referred juvenile jaw, subsequently made by Barbour the type of Elephas [Archidiskodon] scotli. 

After i)h()tograi)h 11*26 

917. ComiJarisonof 1yp<> mandihlcs oi Archidiskodon imperator maibeni and A. imperator scotli; also enlarged views of right .second 

inferior molar of the type of A. scotti 1*'27 

918. Archidiskodon imperator maibeni, type superior and inferior dentition. After photograph 1028 

919. Archidiskodon haroldcooki, type mandible with third right molar in situ. After Hay and Cook 1029 

920. Archidiskodon eiilis, type, compared with A. imperator ref. After photograph 1030 

921. Archidiskodon exilis, type. Facial portion of skull, with tusks and lower jaw. Restoration. After photograph 1031 

922. Map showing location of some occurrences of fossil ele])hants on Channel Islands. After Stock 1032 

923. Archidi.skodon .sonoriensis, anterior portion of type mandible and maxilla, showing third sup(>rior and inferior molars. . . . 1033 

924. Archidiskodon meridionalis nebrasce7wis, type mandible. After Osborn 1034 

925. Archidiskodon meridionalis of Durfort, referred sujierior and inferior molars, found associated with sk(>leton. After casts. . . 1035 

926. Archidiskodon meridionalis of Durfort, second and third su])erior and inferior molars of skeleton. After casts 1035 

927. Archidiskodon meridionalis nebra.scensis, type. Restoration by Flinseh 1036 

928. Archidiskodon meridionalis of Durfort and A. meridionalis nebrascensis of Nebraska. Restorations by Flinseh 1037 

929. Parelephas trogontherii of Mosbach. Restoration by Flinseh 1038 

930. Parelephas jeffersonii, type. Restoration by Knight 1040 

931. Parelephas jeffersonii, type skeleton. Second figure after Osborn 1041 

932. Map of chief Lower to Upper Pleistocene localities in which occur species of Archidiskodon, I'urclcphas, Mammonteus, 

Loxodonta, and Pala-oloxodon. After Osborn 1042 

933. Map .showing geographic distribution of si)ecies of Parelephas, types and referred specimens 1047 

934. Mammonteus primigenius and Parelephas trogontherii, cranial profiles. After Polilig and l''aleoner 1050 

935. Left lateral profiles of seven species of Parelephas with progressive ridge formulae 1051 



LIST OF ILLUSTRATIONS xxi 



FIGURE 



PAGE 



936. ParelepJms of Eiiropo and America in comparison with Elephas indicus bengnlensis. Restorations by Flinsch 1052 

937. Cranial ]3rofiles oi Airhidiskodon imperator, Mammonteus pnmigenius, Parelephasjeffersonii, and /■*. washinglonii 1053 

938. Parelephas trogontherioides, lectotype and cotype molars. After Zuffardi 1054 

939. Parelephas trogonlherii, type third superior and inferior molais. After Pohlig 1057 

940. Parelephas trogonlherii, referred molars from Siissenborn and Weimar. After Wiist 1058 

94 1 . Elephas antiquus Neslii Pohlig [ = Parelephas{?) trogontherii nestii], cotype or syntype left third superior and inferior molars. 

After photographs 1059 

942. Parelephas armeniacus, type left third superior molar. After Falconer 1061 

943. Parelephas intermedius, referred molars. After photographs 1063 

944. Parelephas intermedius, restored skeleton in Lyons Museum. After Lortct and Chantre 1064 

945. Parelephas wiisti, cotype molars, from Tiraspol, Russia. After Pavlow 1066 

946. Parelephas jacksoni, type juvenile jaw, and referred Elephas [Mammonteus] primigenius jaw. After Mather 1068 

947. Diagrammatic cross-.scction of type locality of Parelephas jacksoni. Aitcv Mather 1069 

948. Parelephas rohnnbi, type third right lower molar (middle portion) longitudinally and vertically bisected. After Falconer. . . 1071 

949. Parelephas columbi, restored type molar, redrawn for present Memoir 1072 

950. Elephas lexiaiius [ = Parelephas cohimhi ref.], type right third inferior molar. After Blake 1073 

951. Parelephas columbi, Falconer's type third right inferior molar and Osborn's neotyi)e third left inferior molar. After Usborn . 1074 

952. Parelephas columbi, key to superior and inferior grinding teeth selected from the Cohen Collection, phosphate beds, Charles- 

ton, S. C 1076 

953. Map showing distribution of Parelephas jejfersonii, Mammonteus primigenius, Parelephas columbi, and Archidiskodon 

imperator in the United States and Canada. After Hay 1078 

954. Parelephas columbi, superior and inferior molars found in incomplete skull of Amherst skeleton 1079 

955. Parelephas columbi. Amherst skeleton 1081 

956. Parelephas columbi felicis, type third right superior molar. After Freudenberg 1082 

957. Parelephas columbi cayennensis, type fragmentary third right superior molar. After cast 1083 

958. Parelephas jeffersonii, showing ridge-plate compression at three levels of third right inferior molar 1085 

959. Parelephas jeffersonii, aged type third superior and inferior molars; also same superposed on type molars of Elephas 

rooseveUi 1086 

960. Parelephas jeffersonii, type and paratype [ideotype] grinding teeth and jaws 1089 

961. Frontal views of crania of Parelephas jeffersonii and P. loashingtonii 1090 

962. Profile views of type and referred crania of Parelephas jeffersonii, P. washingtonii, and Mammonteus primigenius; also 

front view of M. primigenius 1091 

963. Parelephas jeffersonii (Franklin County Mammoth) and Elephas indicus bengalensis crania 1092 

964. Parelephas jeffersonii (Franklin County Mammoth) skull in Nebraska State Museum. After photograph 1093 

965. Parelephas jeffersonii (Franklin County Mammoth), aged. Diagram showing ridge-plates of second and third superior 

molars, and portion of crown of third inferior molar. After Barbour 1093 

966. Parelephas jeffersonii, type skeleton in American Mu.seum. First published type figure by Osborn 1094 

967. Parelephas jeffersonii, paratype [ideotype] jaw 1096 

968. Elephas rooseveUi (syn. Parelephasjeffersonii), type third superior and inferior molars 1096 

969. Parelephas progressus, type third superior and inferior molars from Zanes\-ille, Ohio, side \-iews, originally figured li.y 

Osborn as paratypes of Elephas jeffersonii. After Osborn 1098 

970. Parelephas progressus, type molars, crown views. After Osborn 1098 

971. Parelephas washingtonii, referred cranium 1100 

972. Parelephas washingtonii, type adult jaw 1 101 

973. Parelephas washingtonii, referred young adult male skull, right lateral aspect 1102 

974. Parelephas washingtonii, referred young adult male skull, left lateral aspect, combined with type jaw 1102 

975. Parelephas washingtonii, type jaw compared with that of P. jeffersonii 1103 

976. Parelephas xcashingtonii, referred cranium with second and third superior grinders in situ 1103 

977. Parelephas ivashingtonii, referred second left inferior molar. After Peterson 1104 

978. Parelephasil) eellsi, type skull fragment. After Hay 1104 

979. Parelephas jloridanus, type and paratype crania and jaws compared with type molar fragment of P. columbi cayennensis .... 1 105 

980. Parelephas floridanus, type and paratype palates, with molars in situ 1106 

981. Parelephas floridanus, type right and left superior and inferior molars. After Osborn 1108 

982. Parelephas floridanus, type third left superior molar (detailed photograph) and drawing showing method of measuring length 

of superior molar crown. After Osborn 1 109 

983. Parelephas floridanus, type mandible. After Osborn 1110 



xxii OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

984. Porelephas flon'danus, paiatype mandible. After O.sborii 1110 

985. Parelephas floridanus, type inferior molar.s. After Osboni 1110 

986. Parelephas flon'danus, reconstructed type cranium. After Osborn 1111 

987. Parelephas floridanus, referred inferior mandible and milk dentition 1111 

988. New .standard method (1930) of i)robo.scidean .skeletal mea.surcment, illustrated on type skeleton of Parelephas jeffersonii. 

After Osborn 1112 

989. Parelephas floridanus, referred right third inferior molar, showing method of ridge-plate mea.surement 1115 

990. Maminonleus primigenius (Woolly Mammoth). After painting by Knight 1116 

991. Messerschmidt cranium of the mammoth of Siberia. After Breyne 1119 

992. Comparison of crania of Matnmonteus primigenius, Elephas indicus, and Loxodonta nfricana. After Cuvier 1121 

993. Mannnonteus primigenius, lectotype molars. After casts 1122 

994. Migrating Woolly Mammoth (Mammontevs primigenius) as it appeared on the river Somme, northern France. Restoration 

by Knight 1126 

995. ( omjiarison of the tip of the triuik of the mammoth {Mammonteus primigenius) with that of Elephas indicus and Loxodonta 

afrirana. After Flerov 1128 

996. Manimonteus primigenius, skeleton, from Kolyma-Beresowka River, Siberia. After Salensky 1130 

997. Mummonkus primigenius Jraasi, skeleton, from Stcinheim on the Murr, Wurttemlx'rg. After Abel 1130 

998. Mammo7iteus primigenius, skeleton, from Borna, Germany. After Abel 1130 

999. Munimonieus primigenius, skeleton, fnmi Lierre, Belgium. After Dupont 1130 

1000. Restoration of the Woolly Mammoth sketched on the wall of the cavern of Les C'ombarelles aux Eyzies (Dordogne), 

France. After Capitan, Breuil, and Peyrony 1131 

Outlines of the W^jolly Mammoth from the Oiotto of Oombarelles 1131 

Charging mammoth incised on a tusk of Mammonteus primigenius discovered in the rock shelter of La Madeleine (Dor- 
dogne), France. After Lartet and Chiisty 1132 

Maji showing geographic distribution of Mammonteus and Parelephas in North America. After Hay 1133 

Map showing location and principal disco^■eI■ies of fossil manmialian fauna of Alaska-Yukon to the year 1929 1134 

Diagram .showing geographic range of Mammonteus and Parelephas. Now superseded by figiu-e 795 above 1135 

Map sliowing geographic distribution of principal species of Mammonteus, types and referred 1136 

Mammonteus primigenius, skeleton from Moravia. After photograph 1139 

Mannnonteus primigenius, referred molars of Alaska 1 142 

Mammonteus pri7nigenius and Parelephas jeffersonii and P. irashingtonii crania compared 1144 

Mammonteus primigenius, male cranium, from the Yukon. After photograph 1145 

Mammonteus primigenius fiom Alaska, showing growth stages in the jaws and teeth 1145 

Type of Elephas odontoti/raiuius [ = third superior molar of Mammonteus primigenius]. After Eichwald 1146 

.Mammonteus primigenius, Elephas indicus bengalensis, and Loxodonta africana oxyotis. Restorations by Flinsch 1147 

Mammonteus primigenius ("Adams skeleton") from the Lena River, Siberia. Subsequently made by Brandt the ty]x> of 

Elephas brachyramphus. After Tile.sius 1148 

Mammonteus primigenius leith-adamsi, type third left inferior molar. After Pohlig 1150 

Mammonteus hydruntinus, type first left superior molar. After Botti 1 151 

Mammonteus primigenius fraasi, type cranium. After Dietrich 1152 

.Mammonteus primigenius fraasi, type mounted skeleton. After photograph 1153 

Mammonteus primigenius asten»is, type and paratype molars. After Dep^ret and Mayet 1154 

Mammonteus primigenius ('^)astensis, reierred molars, compared with (f) Parelephas, Archidiskodon or Hesperolo.rodon 

molars. After direct scale tracings by the present author 1155 

1021. Mammonteus primigenius americanus, type. Portion of upper molar from near Rochester, New York, .\fter DeKay 1156 

1022. .Mamnio)iteus primigenius compressus, type second and tliird superior molars, .•\fter Osborn 1157 

1023. .Mammonteus primigenius compressus, type female skull. After Osborn 1158 

1024. Mammonteus primigenius compressus, paratype third right superior molar. XHcr ( )sh()rn 1159 

1025. Mammonteus primigenius alaskensis, cotype crania, superposed outlines 1159 

1026. .Mammonteus primigenius alaskensis, cotype crania 1160 

1027. Discoveiy .sites of frozen carcasses of tlie Woolly JVLammoth and Rhinoceros. After Tolmachoff 1162 

1028. Map .showing geographic distribution of mammoths in Upper Pliocene and Pleistocene times. After Osborn 1164 

1029. .lefi'ersonian Mammoth of Indiana 1 165 

1030. Imperial ]\Iammoth of Nebraska 1 165 

1031. Arched tu.sks of the African elephant 1166 

1032. Mammonteus primigenius, referred skull and half-grown tusks of male specimen found on the Yukon River, Alaska 1166 



LIST OF ILLUSTRATIONS xxiii 

FIGURE PAGE 

1033. Circular tusks of the Woolly Mammoth of Siberia 1 166 

1034. Woolly Mammoth, Somme Kiver, France 1166 

1035. Strata of the Pekarna Cave, Moravia. After photograph 1168 

1036. Mammoth pit of V&tonice, Moravia. After photograph 1168 

1037. Ciiant killing stone of the Moravian hunters. After photograph 1168 

1038. Ivory figurine of a woman's head, from Brassempouy. After Pilloy in Piette's "L'Art pendant I'Age du Renne" 1168 

1039. Equine ivory statuette from Lourdes. After Pilloy in Piette's "L'Art pendant I'Age du Renne" 1169 

1040. Loiodonta africana, young adult bull of the Lake Paradise region, east central Africa. Photograph by Mr. and Mrs. Martin 

Johnson, and shown in film "Simba." Courtesy of Mr. Daniel E. Pomeroy 1170 

1041. Crania of Loxodonta africana, Palxoloxodon namadicus, and Hesperoloxodon antiquus. After Falconer and Cautley, Pilgrim, 

Weithofer, and Pohlig 1 172 

1042. Cuvier's figures and definitions of Elephas primigenius (Messerschmidt's cranium), E. indicus, and E. africanus 1173 

1043. Loxodonta africana and Elephas indicits, third inferior molars. After Owen 1175 

1044. Loxodonta africana and Elephas indicus, second .superior molars. After Lydekker 1175 

1045. Loxodonta africana oxyotis ("Jumbo"), cranium and jaws 1 176 

1046. Palseoloxodon namadicus, tyjie cranium, from the Nerbudda. After Falconer and Cautley 1176 

1047. Pala'oloxodon and Hesperoloxodon as depicted by the cave men of North Africa and Spain. After Pomel and after Breuil . . . 1184 

1048. Map showing geographic distribution of the principal species of the Loxodontinse 1186 

1049. Habitat of the African elephant (Loxodonta), forest and savanna of the Uasin (Jishu Plateau, Kenya ( 'olony. ( 'alf of old fe- 

male charging elephant. After photograph by Carl E. Akeley 1189 

1050. Habitat of the African elephant (Loxodonta). Same region as previous figure. Females and young bulls in forest. After 

photograph by Kermit Roosevelt 1 189 

1051. Small herd of African elephants (Loxodonta). After film photograph by Martin Johnson 1189 

1052. Akeley group of African elephants in the American Museum 1190 

1053. "Khartum" (Loxodonta africana oxyotis), formerly living in the New York Zoological Park. Two growth stages. After 

photograph 1 194 

1054. Remarkable cave paintings of white rhinoceros and African elephant, discovered in South Africa. Courtesy of London 

Illustrated News 1 194 

1055. Map showing distribution of existing African elephant (Loxodonta) 1195 

1056. Loxodonta africana oxyotis and L. africana pumilio. Restorations by Flinsch 1196 

1057. Loxodonta africana, Blumenbach's original figure of type, a right second inferior molar, described as Elephas africanus 1 197 

1058. Elephas priscus [ = Loxodonta africana]. Type molar after Goldfuss 1197 

1059. Loxodonta africana peeli, referred skull and tusks of adult male, from Mt. Kenya. After photograph 1198 

1060. Loxodonta africana oxyotis ("Jumbo"), referred middle-aged skull (24 years) 1199 

1061. Loxodonta africana oxyotis ("Jumbo"), superior and palatal views of cranium. Age twenty-four years 1200 

1062. Comparison of tusks of Loxodonta africana oxyotis and Mammonteus primigenius. After photographs 1201 

1063. Loxodonta africana albertensis and L. africana peeli, full-grown heads (male and female) 1202 

1064. Loxodonta africana (?)cottoni, foetal cranium, jaw, and milk dentition. After Eales 1203 

1065. Tusks of African elephant, believed to be the heaviest in the world, shown in front of a typical Arab door at Zanzibar. One 

tu.sk now in British Mu.seum. After Kunz, courte.sy of executors of his estate 1204 

1066. Loxodonta cornaliae, type right superior molar. After Aradas 1205 

1067. Young Addobush elephant (Loxodonta) from Cape Colony. After photograph 1205 

1068. Hesperoloxodon of Europe, Palieoloxodon of India and of the Mediterranean Islands, compared with drawing of Hesper- 

oloxodon by cave men of northern Spain. Restorations by Flinsch 1206 

1069. Comparison of crania of Palxoloxodon namadicus, Hesperoloxodon antiquus italicus, H. antiquus ausonius, and H. antiquus 

platyrhynchus. After Falconer and Cautley, Pilgrim, Weithofer, Graells, and Pohlig 1208 

1070. Palieoloxodon namadicus, type female(?) cranium. After Falconer and Cautley 1211 

1071. Channel of the Godavari near Nandiir Madm^shwar, India. After Pilgrim 1213 

1072. Comparison of Hesperoloxodon and Palseoloxodon (syn. Sivalikia) superior and inferior molars 1214 

1073. Three progressive broadening stages in the Palseoloxodon (syn. Sivalikia) and Hesperoloxodon superior grinding teeth. . . . 1215 

1074. Hesperoloxodon antiquus (Upnor elephant). Restoration by Flinsch 1216 

1075. Hesperoloxodon a?iiiquus, lectotype left second inferior molar. After Falconer and Cautley 1218 

1076. Hesperoloxodon antiquus, referred left second inferior molar from firay's Thiurock. After Falconer and Cautley 1219 

1077. Hesperoloxodon antiquus, referred third right superior molar. After Falconer and Cautley 1220 

1078. Hesperoloxodon antiquus (Upnor elephant). Referred second superior and inferior molars. After photographs 1220 

1079. Hesperoloxodon antiquus of Upnor. Mounted skeleton in British Museum. After photographs 1223 



xxiv OSBORN: THE PROBOSCIDEA 

FIGURE PAGE 

1080. Hesperoloxodon antiquus of Upnor, skeleton in the British Musoum. Redrawn to show original and restored parts 1224 

1081. Scapulae of Hesperoloxndon a7itiquus of Upnor, Loxodonla africana ("Jumbo") of the Sudan, and Elephas indicus. After 

Andrews and Cooper 1225 

Dorsolumbar vertebrae of Hesperoloxodon antiquus of Upnor and Elephas indicus 1225 

Upnor elephant, partly restored skeleton and flesh outlines 1227 

\'ertebral columns of Hesperoloxodon antiquus, Loxodonta africana oxyotis, Elephas indicus, Mammonteus primigenius, and 

Parelephas jeffersonii 1229 

Hesperoloxodon antiquus nanus, type, possibly a left second superior molar. After Acconci 1231 

Hesperoloxodon antiquus plalyrhynchus, type premaxillaiy rostrum with tusk and portion of right superior maxillary with 

second molar in situ. After Graells 1231 

Hesperoloxodon antiquus ausonius, type third right and left inferior molars. After Dep^ret and Mayet 1232 

Hesperoloxodon antiquus, H. antiquus germanicus, H. antiquus italicus, progressive stages in evolution of grinding teeth 1234 

Hesperoloxodon antiquus germanicus, type. Portion of right second inferior molar. After Stefanescu 1235 

Hesperoloxodon antiquus germanicus, right tusk in Field Museum of Natural History, from Steinheim, male and female 

tusks in Ciotha Museum, from Tonna, right male tusk in Stuttgart Museum, from Steinheim 1236 

Hesperoloxodon antiquus germanicus, right tusk excavated at Steinheim. Diagrammatic sketch reproduced through courtesy 

of Dr. Henry Field 1237 

Hesperoloxodon antiquus italicus. Restoration by Flinsch 1238 

Loxodonta africana albertensis. Restoration by Flinsch 1239 

Map of valley of the Liri, Italy, showing the location where Hesperoloxodon antiquus italicus was found, and other exposures. 

After De Lorenzo and D'Erasmo 1 239 

Mammalian fossils associated with type cranium of Hesperoloxodon antiquus italicus. After Osborn 1240 

Hesperoloxodon antiquus italicus of Pignataro Interamna, type cranium before removal 1241 

:Map showing location of Pignataro Interamna, region of valley of the Liri. After Century Atlas, 1913 1241 

Hesperoloxodon antiquus italicus, type cranium. After De Lorenzo's original sketch and measurements 1242 

Hesperoloxodon antiquus italicus, type cranium. After De Lorenzo and D'Erasmo 1243 

Hesperoloxodon antiquus italicus, front view of cranium. After De Lorenzo and D'Erasmo 1243 

Hesperoloxodon atdiquus italicus, type .superior grinders. After photograph 1244 

Hesperoloxodon antiquus italicus, type right second and third superior and inferior grinders. After Osborn 1244 

Hesperoloxodon antiquus italicus, type mandible, right lateral view with second and third superior teeth superposed on cor- 
responding inferior teeth, and superior view. After Osborn 1245 

Hesperoloxodon antiquus italicus, type third right inferior molar. After Osborn 1245 

Hesperoloxodon antiquus italicus, type cranium, three aspects 1246 

Hesperoloxodon antiquus italicus, type cranium as reconstructed and mounted in the American Museum. After Osborn .... 1247 

Hesperoloxodon antiquus italicus type and Loxodonta africana peeli ref. crania. After Osborn 1248 

Comparative bathycephaly of the Loxodontinae: Loxodonta africana, Palxoloxodon namadicus, and Hesperoloxodon 

antiquus italicus. After Osborn 1249 

Comparison of scapulae of Hesperoloxodon antiquus italicus, Loxodonta africana, and Elephas indicus 1249 

Palxoloxodon namadicus, referred cranium, of the Godavari Alluvium at Nandiir Madm^shwar, India. After Pilgrim 1250 

Hesperoloxodon antiquus italicus, referred femur, with Dr. Pohlig standing beside it. After photograph 1251 

Intracranial brain casts of Hesperoloxodon antiquus italicus, Loxodonta africana, and Elephas indicus. After Osborn 1252 

W. Bauer Quarry at Steinheim on the Murr, showing the site of the 1928 discovery of the cranium of Hesperoloxodon 

antiquus germanicus, ref. After photograph 1253 

1114. Hesperoloxodon antiquus germanicus, referred crania of 1926 and 1928. After photographs 1254 

1115. Hesperoloxodcm antiquus germanicus, referred third right superior molar from Steinheim 1255 

1116. Elephas antiquitatis Kruger [ = Hesperoloxodon antiquus germanicus ref.], type molar. After Breislak 1256 

1117. Bathymetric map of the Mediterranean Islands. By permission of Longmans, Green and Company, from map edited by 

Chisholm and Leete ^^^' 

1118. Dwarfed elephants of the Mediterranean Islands. Diagrammatic representation 1258 

1119. Dwarfed elephants of the Mediterranean Islands. Restorations by Flinsch 1259 

1 120. Loxodonta africana pumilio, young "pygmy" elephant pa.s.sing beneath adult Elephas indicus, female. After photograph ... 1259 

1121. Pala-oloxodon mnaidriensis, referred, the "Elephas {antiquu.'^) Melita'" of Pohlig. Fully adult cranium from the (Jrotta di 

Pontale, Carini, Sicily. After Pohlig • 1260 

1122. Palxoloxodon mnaidriensis, referred juvenile cranium, the "Elephas (antiquus) Melitie" of Pohlig. After Pohlig 1200 

1123. Pala-oloxodon rnelitensis, type third left superior molar. After Falconer 1262 

1124. Palwoloxodon melitensis, referred mandible from the Grotta di Pontale, Sicily 1263 



LIST OF ILLUSTRATIONS xxv 

FIGURE PAGE 

1125. Palseoloxodon 7nnmdriensis, third right inferior molar. After Leith Adams 1264 

1126. Palseoloxodon mnaidriensts, type and para type molars. After Leith Adam.s 1264 

1127. Denizens of ancient Malta. Restoration by Leith Adams as dwarfed African elephants 1265 

1128. Palseoloxodon Cypriotes, cotype molars. After Bate 1266 

1129. Palseoloxodon creticus, cotype molars. After Bate 1267 

1130. Section of Grotto of Luparello, Palermo, Sicily. After \'aufrey 1268 

11.31. Tusks in the three dwarfed species of the Mediterranean Islands: Palseoloxodon fakoneri, P. melilensis, P. mnaidriensis. 

After Vaiifrey 1271 

1132. Two types of molars belonging to Elephas [ = Palseoloxodon] mnaidriensis according to Vaiifrey, namely, 'tjqpe endioganai' 

and 'type pachyganal,' from Shantiun and Puntali. After Vaufrey 1271 

1133. Molars referred by ^■aufrey to Elephas [ = Palieoloxodon] melitensis, from Luparello, Sicily, and Benghi.sa, Malta. After 

Vaufrey 1272 

1134. Ulnae of Palseoloxodon mnaidriensis, P. melitensis, and P. fakoneri. After Vaufrey 1272 

1135. Palseoloxodon allanticus, cotype right second inferior molar. After Pomel 1274 

1136. Palseoloxodon atlaniicus, referred third left superior molar. After Pomel 1274 

1137. Palseoloxodon jolensis, type left third inferior molar. After Pomel 1275 

1138. Palseoloxodon recki, lectotype left second inferior molar. After Dietrich 1276 

1139. Palseoloxodon{1) andrewsi (Dart's tjrpe of Archidiskodon andrewsi). Left third inferior molar, restored. After O.sborn .... 1278 

1140. Palseoloxodon hanekomi, type ?third right superior molar. After Dart 1279 

1141. Palseoloxodon yorki, type right ?third inferior molar. After Dart 1280 

1142. Palseoloxodon vnlmani, type ?third inferior molar. After Dart 1281 

1143. Palseoloxodon kuhni, type "?lower left molar." After Dart 1281 

1144. Palseoloxodon archidiskodontoides, type ?second superior molar and part of right humerus. After Haughton 1282 

1145. Palseoloxodon transvaalensis, type right third superior molar. Modified after Dart's photographs 1284 

1146. Palseoloxodon sheppardi, type left third superior molar. Modified after Dart's photographs 1284 

1147. Loxodonta zulu, type third left inferior molar. After Scott 1286 

1148. Loxodonta zulu, referred third left inferior molar, from Kaiso Bone-beds, near Lake Albert, Africa. After photograph 1287 

1149. Loxodonta prima, type third left inferior molar. After Dart 1287 

1150. Loxodonta africana var. obliqua, type third right inferior molar. After Dart 1288 

1151. Loxodonta subantiqua, type "possibly a right lower molar, probably the .second." After Haughton 1288 

1152. Palseoloxodon najnadicus nanmanni, type, of Makiyama, in comparison with Hesperoloxodon antiquus germanicus ref. of 

Pohlig. Diagrammatic outline sketch 1294 

1153. Palseoloxodon namadicus namadi, type right third superior molar. After Makiyama 1296 

1154. Palseoloxodon protomammonteus, type left third infeiior molar. After Matsumoto 1297 

1155. Palseoloxodon protomammonteus proximus, type fragment of loft third inferior molar. After Mat.sumoto 1298 

1156. Pals'oloxodon namadicus yabei, type right ramus of mandible, containing third molar. After Matsumoto 1299 

1157. Palseoloxodon (Archidiskodonf) tokunagai mut. junior, type right second inferior molar. After Mat.sumoto 1300 

1158. Parelephas protomammonteus matsumotoi, type. Portion of left mandibular ramus with third molar //( situ. After Saheki. 

Not determined by the present author 1300 

1159. Palieoloxodon yokohamanus, type second right superior molar. After Tokunaga 1301 

1160. Palseoloxodon hysudrindiciis, cotype molars. After photographs 1302 

1161. Map showing Japan as part of the Asiatic continent in Plio-Pleistocene time. After Longmans' New School Atlas 1304 

1162. Map showing Japan as part of the Asiatic continent in Plio-Pleistocenc time. After Yabe 1305 

1163. Referred Elephas indicus. Male and female Ceylon elephants. After photograph by Plate Ltd 1306 

1164. First definition of the genus Elephas, bracketed with ?Rhinoceros. After first edition of Linnaeus' "Sy.stema Naturs," 

page 10 1309 

1165. Facsimile of portion of page 11 of Linnaeus' Memoir of the Museum Adolphi Friderici Regis, Stockholm, 1754, in which first 

appears the species name Elephas indicus 1309 

1166. Facsimile of page 33 of Linnaeus' original tenth edition of the "Systema Naturae," 1758, in which Elephas maximus is sub- 

stituted for Elephas indicus 1310 

1167. Indian elephant group in the American Museum of Natural History. Specimens from the hills in the Province of Mysore, 

shot in 1923 by Mr. Arthur S. Vernay 1311 

1168. Elephas indicus sumatranus, pair of young elephants from Sumatra captive (1921) in the Zoological Park of Wa-shington. 

.\fter photograph 1314 

1169. The Sumatran elephant, apparently a female, living in the Am.stcrdam Zoological Gardens, August, 1913. After photo- 

graph 1314 



xxvi OSBORN: THE PROBOSCIDEA 

FIGURE TAGE 

1170. EhphoH indints rri/laniciis and E. indicus botgalrnsi.s, eraniti. Types. After dc IMaiiivillc 1816 

1171. Crania of Elephus indicun (Dauntcla var.) and E. indicus (iMukna var.). After Falconer and Cautley 1317 

1172. Elephas indicufs bengalensis, type, and E. indicus ceylanicus, referred, crania. After photographs 1317 

1173. Map showing geograjihic distribution of the principal .species and subspecies (Hving and extnct) of Elephas, Hijpselephas, 

and Plalelephas 1318 

Platelephas platyrephalus, Hypselephas hysudricus, and Elephas indicus. Restorations of heads by Flinsch 1320 

Elephas indicus hengalensis. A herd of wild elephants in a bamboo jungle of Mysore. After ])hotograph 1322 

Elephas indicus, referred third inferior molar (vertical section) of an unusually large specimen from Assam. After Falconer 

and Cautley 1324 

Elephas asinlicvs. Blumenbach's original type figure of a first right superior molar 1325 

Elephas indicus ref . Section of a partly worn third inferior molar. After Gaudry 1325 

Elephas indicus hengalensis, known as the "giant tusker of Udiapur," showing abnormal length of tusks. After photographs. 1326 

Comparison of crania of Elephas indicus ceylanicus and E. indicus hengalensis 1327 

Elephas indicus sumalranus formerly living in the Rotterdam Zoological Gardens. After Lydekker 1329 

Elephas sumalranus, cotype male and female crania from Palcmbang, in Leiden Museum. After photographs 1330 

Elephas sumalranus crania (adult and infantile), in Munich Museum. After photographs 1330 

Sumatra!! elephant from Batang Serangan. Mounted specimen in Munich Museum. After photograph 1331 

Sumatran elephant (infantile) in Munich Museum. Mounted specimen. After photograph 1331 

Mounted Burmese elephant. After photograjih 1332 

Elephas indicus hirsulus, type, formerly living in the Gardens of the Zoological Society, London. Mounted specimen in the 

British Museum (Natural History). After Lydekker 1333 

Elephas indicus buski Matsumoto [=1Palseoloxodon huski], type first superior molar of the left side, from Japan. After 

Matsumoto 1333 

Second right superior molar from between Kanagawa and Tokio (Yedo), Jajjan, referred to Palseoloxodon by the present 

author. After Lydekker 1334 

Superior view of heads of a young African elephant and of an adult Lidian elephant. After Geoffroy St.-Hilaire and 

Frederic Cuvier 1335 

Loxodonla africana and Elephas indicus, crown views of third right inferior molars. After Owen 1335 

Comparison of low-browed African cranium {Loxodonla africana) and high-browed Indian cranium {Elephas indicus), 

showing deeply embedded brain below cranial air cells 1335 

Unguligradism. Radiograph of right foot of a young Indian elephant 1336 

Estimated shoulder heights of Indian elephant, .skeletal and fle.sh 1337 

"Elephas planifrons" and "Elephas hysudricus" life zones. After Pilgrim 1338 

Map showing Upper Siwalik exposures of the Simla foothills, India 1339 

Hypselephas hysudricus, type and jjaratype molars. Sections after Falconer and Cautley 1341 

Hypselephas hysudricus, paratype third right inferior molar. After Falconer and Cautley 1342 

Hijpselephas hysudricus, referred third left inferior molar. American Museum (Barnum Brown) collection 1342 

Hypseliphas hysudricus, icferred second left inferior molar 1343 

Hypselephas hysudricus, referred third left superior and inferior molar .sections. American Museum (Barnum Mrown) 

collection 1344 

Hypselephas hysudricus, referred third right superior molar, inner view, also transverse section, and photograph of occlusal 

surface. American Mu.seum (Barnum Brown) collection 1345 

Hi/pselephas hysudricus, referred superior and inferior molars from India. American Mu.seum (Barnum Brown) colhiction . . 1346 
( 'ompari.soii of two crania of Hypselephas hysudricus with Mukna and Dauntcla varieties of Elephas indicus. After Falconer 

and Cautley 1349 

Hypselephas hysudricus, referred adult mal(> craniimi. After Falconer and Cautley 1350 

Hypselephas hysudricus, portion of young jaw with greatly elongated rostrum. American Museum (Barnum Brown) 

collection 1351 

Plalelephas platycephalus, type cranium (palatal view). American Museum (Barnum Brown) collection 1352 

Archidi.skodon planifrons, referred adult cranium of supposed female, with small tusks. After Falconer and Cautley 1352 

Hypselephas hysudricus, male craniimi (palalal view). After Falconer and Cautley 1352 

Hypselephas hysudricus, referred female cranium in Amherst Museum, collected near Kulhi, a district of the Punjab, by 

M. M. Carieton 1353 

1211. /////wefe/j/io.s/i.v.saf/c/rM.s, restored juvenile skull. .American Mu.seum (Barnum Brown) collection, .\fter jjliotograpli 1354 

1212. Hypselephas hysudricus, referred cranium. After Falconer and Cautley 1354 

1213. Hypselephas hysudricus, referred juvenile crania in British and American Museums 1355 



LIST OF ILLUSTRATIONS xxvii 

FIGURE PAGE 

1214. Ifypnelephas hysudricus, referred juvenile eniiiium and jaws. American IMufieuni (Barnuni Brown) collection 1356 

1215. Fiypselephas hysudricus, referred right second superior molar, originally selected by Osl)orn as the type of Elephas platyce- 

phalus angustidens. American Museum (Barnum Brown) collection 1357 

1216. Upper Pliocene and Lower Pleistocene strata near Siswan, India, showing site where type cranium of Platelephas platyre- 

phalus was found. After ])hotographs 1358 

1217. Comparison of the types of Platelephas plaiycephalus and Stegodon pinjorensis. Cranial sections 1360 

1218. Plalelephas platycephalus and Hypselephas hysudricus, right cranial profiles 1360 

1219. Platelephas platycephalus, four aspects of type cranium. American Museum (Barnum Brown) collection 1361 

1220. Geological relationships of African Proboscidea. Columnar section by Colbert 1422 

1221. Geological relationships of Oriental Proboscidea. Columnar section by Colbert 1440 

1222. Geological relationships of European Proboscidea. Columnar section by Colbert 1457 

1223. Geological relationships of Asiatic Proboscidea. Columnar section by Colbert 1477 

1224. Geological relationships of North American Proboscidea. Columnar section by Colbert 1491 

1225. Geological relationships of South American Proboscidea. Columnar section by Colbert 1516 

1226. Models of Recent and extinct Mammoths and Mastodons. After Knight 1522 

1227. Map showing geographic distribution of the Moeritherioidea, Deinotherioidea, and Mastodontoidea 1528 

1228. Map showing geographic distribution of the Stegodontoidea and Elephantoidea (including the Stegolophodontinae of the 

Mastodontoidea) 1538 

1229. Four-coned ancestral grinders of the Proboscidea (e.g., Mwritherium) compared with the six-coned Palitomastodon molars. . 1544 

1230. Molar diagrams showing typical crown pattern (median sulcus, median conules, double trefoils, .serrate .spurs) in each of the 

four families of the Mastodontoidea 1546 

1231. Accelerated evolution of ridge-plates from Archidiskodon planifrotis into A. imperator; also Stegodon grangeri molar, with 

enamel foldings, inserted to show \'-shaped valleys of the stegodontoid molar as compared with I'-shaped valleys of the 

elephantoid molar 1547 

1232. Brevirost rinse: Proversion of ridge-crests in Anancus 1548 

1233. Humboldtinae: Retroversion and centroversion of superior ridge-crests in Cuvieronius and Stegomastodon 1548 

1234. Crown view of third inferior molars of the right side of Loxodonta africana and Elephas indicus. After Owen 1549 

1235. Archidiskodon subplanifrons, type third right inferior molar. Section showing cement, dentine, and enamel. Drawing by 

D. F. Levett Bradley 1549 

1236. Chief head and dental forms of four of the superfamilies (I H) of the Proboscidea (McEritherioidea, Deinotherioidea, Masto- 

dontoidea, Elephantoidea) 1550 

1237. Divergent adaptive radiation of crania and incisive tusks in six bunomastodont .subfamilies 1551 

1238. Deinotherium giganteum, juvenile jaws showing replacement molars. After Lartet 1554 

1239. Cencplasmie evolution of the archaic-toothed mammoths during a three-million-year period so far as known to April, 1935. 

After Osborn 1581 

1240. Alloiometrons: Adaptive speed and weight proportions. After Osborn 1581 

1241. General climatic distribution of the subfamilies of the Ele])hantoidea and Stegodontoidea including theoretic migration 

lines (1938). After American CJeographical Society North Polar Projection 1589 

1242. Worldwide distribution of the Proboscidea in past and present time. Same as figure 6 of Volume I , wit h modifications 1594 

1243. Foot trail of Indian elephant "Gunda," formerly li\ing in the New York Zoological Park, taken in sand 1598 

1244. Elephas indicus ref., showing contrast in proportions between adult and young. After photograph by Underwood aiul 

Underwood 1599 



Chapter XIV 
THE ROOF-TOOTHED STEGODONTS, SUPERFAMILY STEGODONTOIDEA 

Possible ancestry in the miocene zygolophodonts of western europe. Primitive forest-browsing 

ELEPHANTS OF THE ORIENTAL REGION. SlOWLY PROGRESSIVE FROM MIDDLE MIOCENE TO MIDDLE [upPEr] 
pleistocene TIME. BrACHYODONT TO SUBHYPSODONT, BUT PROGRESSIVE RIDGE FORMULA. EXTINCTION IN 
MIDDLE [upper] PLEISTOCENE TIME. 



I. Introduction. 

1. History of classification. 

2. Habits and general characters. 

3. Approximate descending ridge formulae, after Falconer, 

Lydekker, Martin, and Osborn. 

4. Geologic and diphyletic order of the Stegodontinse. 

5. History of discovery of the subfamily Stegodontinse. 

Principles of type revision of the species. 

6. The Stegodontinse and Mastodontinae of China. 

7. Pliocene to Pleistocene Proboscidea of Japan. 

8. Phylogenetic discussion of the thirty described species 

of Stegodonts and Stegolophodonts. 
Probable African-European-Asiatic origin and migra- 
tion of the primitive Stegodonts. 

II. Type Revision of the Species in Order of Original 
Discovery and Description. 

1. The first two Stegodonts, discovered in Burma, 1828. 

Mastodon laiidens Clift. 
Mastodon elephantoides Clift. 

2. Discoveries in India and Burma (1845, 1846). 

Third species, Elephas insignis, India. 
Fourth species, Elephas ganesa, India. 
Fifth species, Elephas hombifrons, India. 
Sixth species, Elephas cliftii, Burma. 

3. The Stegodonts of China, India, Java, the Philippine 

Islands, Austria, Japan, and Burma. 
Seventh species, Stegodon sinensis, China. 
Eighth species, Stegodon orientalis, China. 
Ninth species, Mastodon cautleyi of Perim Island. 
Tenth species, Stegodon trigonocephalus of Java. 
Eleventh species, Stegodon mindanensis, Philippine 

Islands. 
Twelfth species, Stegodon airdwana of Java. 
Thirteenth species, Stegodon ganesa var. javaniciis of 

Trinil, Java. 
Fourteenth species. Mastodon stegodontoides of Lehri, 

India. 
Fifteenth species, Elephas (Prostegodon, Parastegodon) 

auroTBe of Japan. 
Sixteenth species. Mastodon (Bunolophodon) longi- 

rostre Kaup forma siiblatidens of Austria. 
Seventeenth species, Stegodon orientalis shodoensis of 

Japan. 
Eighteenth species, Stegolophodon nathotensis, India. 
Nineteenth species, Stegolophodon cautleyi progres- 

sus, India. 
Twentieth species, Stegodon orientalis grangeri, 

China. 
Twenty-first species, Stegodon insignis birmanicus, 

Burma. 
'J'wenty-second species, Stegodon pinjorensis, India. 



Twenty-third species, Stegodon bondolensis, Java. 
Twenty-fourth species, Stegodon trigonocephalus prae- 

cursor, Java. 
Twenty-fifth species, Parastegodon? hvantoensis, Japan. 
Twenty-sixth species, Stegodon yiishensis, China. 
Twenty-seventh species, Stegodon officinalis, China. 
Twenty-eighth species, Stegodon zdanskyi, China. 
Twenty-ninth species, Parastegodon [Stegodon?] sugi- 

yamai, Japan. 
Thirtieth species, Stegolophodon lydekkeri, Borneo. 

III. Systematic Arrangement of the Stegolophodonts and 

Stegodonts in Phyloge.netic Order. 

1. Characters of the subfamily Stegodontinse. 

2. History of the generic names assigned to the Stegolopho- 

donts and to the Stegodonts. 
Generic characters of Stegolophodon Schlesinger. 
Systematic description of species of Stegolophodon. 

Stegolophodon cautleyi of the Upper Miocene 
[Middle Pliocene], Perim Island. 

Stegolophodon latidens of the Lower Pliocene 
[?Lower Pleistocene] of Burma and [Middle 
Pliocene] of India. 

Stegolophodon sublatidens of the Middle(?) Pliocene 
of Austria. 

Stegolophodon stegodontoides of the Upper(?) Plio- 
cene of India. 

Stegolophodo7i nathotensis of India. 

Stegolophodon cautleyi progressus of India. 

Stegolophodon lydekkeri of Borneo. 

IV. Succession of Species of the Genus Stegodon. 

Genus Stegodon Falconer and Cautley, 1847, 1857. 

Skulls of Stegodonts in the British and Indian Museums. 
Characters of referred skulls of Indian Stegodonts, 

after Falconer, 1868. 
Stegodont crania of China and of the East Indies. 
Systematic description of species of Stegodon. 

Stegodon sinensis of the Yangtze River, China. 
Stegodon elephantoides of Burma. 
Stegodon cliftii of the Irrawaddy River, Burma, and 
of the Dhok Pathan horizon, India. 
Falconer's Notes of 1868 on Stegodon cliftii. 
Stegodon bombifrons of the Lower [Middle] Pliocene, 
Dhok Pathan horizon. 
Falconer's Notes of 1868 on Elephas [ = Stegodon] 

bombifrons. 
Lydekker's Notes of 1886 on Elephas [ = Stegodon] 
bombifrons. 
Stegodon insignis of the Upper Pliocene [to Ujjper 

Pleistocene] of India. 
Stegodon ganesa of the Upper Pliocene [to Upper 
Pleistocene] of India. 



805 



806 



OSBORN: THE PROBOSCIDEA 



Falconer's Notes of 1868 on Stegodon insignis 

and S. ganesa. 
Lydckker's (1886) comparison of Stegodon in- 
signis, S. ganesa, and S. bombifrons. 
Stegodon insignis-ganesa a collective sjjecies. 
Stegodon insignia birnianicus of Burma. 
Stegodon orientalis grangeri of China. 

Cranial characters of .S. orientalis grangeri. 
Stegodon pinjorensis of India. 

Stegodon orientalis of Szechuan, northwest China. 
Lj'dekker's Notes of 1886 on Stegodon orientalis. 
Stegodon airdirana of Trinil, Pithecanthropus ereclus 
zone, Java. 
Stegodon airdwana fauna, Kendeng-Schichten 

layer of Trinil, Java. 
Comparison of Stegodon airdirana of Ja\-a with 
Stegodon insignis-gane.sa of India. 
Stegodon Irigonocephaliis of the vicinity of Surakarta, 

Java. 
Stegodon ganesa vav.javaiiiriisof theTrinil horizon, 



Middle Pleistocene, Java (synonym of .S'. 
airdwana or S. irigonocephaliis). 

Stegodon (Archidiskodon?) mindariensis of Minda- 
nao, Philippine Islands. 

Stegodon aurone of Mt. Tomuro, Japan. 

Stegodon orientalis shodoensis of Japan. 

Recently described Stegodonts fkom Java, China, and 
Japan. 

Stegodon bondolensis of Java. 

Stegodon trigonoccphalus praecursor of Java. 

Parastegodon'^. kwantoensis of Japan. 

Stegodon yilshensis of ("hina. 

Stegodon ojficinalis of China. 

Stegodon zdanskyi of China. 

Parastegodon [Stegodon'!] sugiyawui of Japan. 

Ari'ENDix: Matsumoto on the phylogeny and classification 
of the Japanese Mastodonts, Stegodonts, and Elephants 
(1924-1927). 



I. INTRODUCTION 

1. HISTORY OF CLASSIFICATION 

[The name Stegodontoidea first appeared in a diagram by Professor Osborii in his article of June, 1935, 

entitled, "The Ancestral Tree of the Proboscidea. Discovery, Evolution, Migration and Extinction over a 

oO,0()0,()00 Year Period" (Osborn, 1935.937, p. 407, fig. 2), in which superfamily he included both Stegolophodon 

and Stegodon under the family Stegodontidse of Young-Hopwood, thus removing them from the Elephantoidea. 

In Volume I of the present Memoir (published Aug. 15, 1936) Professor Osborn confirmed his separation of 
the Stegodontoidea from the Elephantoidea (pp. 22, 25) but he withdrew the Stegolophodonts, placing them in 
the superfamily Mastodontoidea, family Mastodontidse, and creating a new subfamily, the Stegolophodontinse 
(see pp. 700, 737, and PI. n ), to embrace the various species, owing to the "intermediate position of [the] molars 
between the true Mastodontidse . . . and the true Stegodontoidea," and suggesting (p. 191) the "possible derivation 
of the grinding teeth of the Stegodontoidea from those of Stegolophodon." The true Stegodonts he retained in 
the superfamily Stegodontoidea, subfamily Stegodontinai. 

As early as 1857 Falconer observed (1857.1, p. 314) that "The Stegodons constitute the intermediate group 
of the Proboscidea from which the other species diverge through their dental characters, on the one side into the 
Mastodons, and on the other into the typical Elephants." Later Falconer (in Murchison, Pal. Mem., 1868, 
\'ol. II, footnote, p. 268) remarked that "The Indian fossil species, which have been ranged under th(> designation 
of Stegodon, establish, through their molar teetli, a manifest and nearly unbroken passage from tlie Mastodons 
into the tru(> Elephants." Also, as recently as 1932, van der Maarel (1932.1, p. 162) expressed the opinion that 
"all the species of Stegolophodon . . . being all v(>ry primitive forms . . . may as well be reckoned to the family of 
t he M asiodontidx .' ' 

Therefore, while various scientific observers have regarded certain of the Stegodonts as transitional Ix'tween 
the Mastodontida^ and the Elephantida^, it remained for Professor Osborn to assign the superfamily name Stego- 
tlontoidea to the true Stegotlonts and to remove the Stegolophodonts to the sui)erfamily Mastodontoidea, under 
the new subfamily name Stegolophodontinse, the members of which he designated (Vol. I, j). 690) as "pro-stego- 
donts." 

Tlie regrettable deatli of Professor Osborn in No\eniber of 1935 precludes the full treatment of lliese groups 
as contemplated by him (see Vol. I, p. 197) ; it is deemed best, therefore, to allow the present chapter to remain as 



THESTEGODONTIN^: HISTORY 807 

first written by him, making such changes (either in square brackets or in editorial notes) as are consistent with 
the known opinions of the author, namely, as deduced from pubUshed statements, accumulated notes, and 
interlined text. 

The classification of the genera Stegolophodon and Stegodon, now to be described, would appear, therefore, to 
be as follows: 

SuPERFAMiLY : STEGODONTOIDEA Osborn, 1935, 1936 
Separated by Osborn (Vol. I, 1936, pp. 22, 25, of the present Memoir) from the Elephantoidea Osborn, 1921, 
as a distinct stock, but without diagnosis. However, according to his observations on the sectioned molars, the 
valleys separating the adjacent ridges are closed or V-shaped at the bottom in the Stegodontoids (Fig. 764) 
and open or U-shaped in the Elephantoids (Fig. 1231). He also considered that the extremely short face of the 
Stegodontoids could not have given rise to the longer face of the Elephantoids. 

Family: STEGODONTIDiE Young-Hopwood, 1935 
This name appears in Young (1935.1, p. 5) but without definition. Later in 1935, Hopwood, in his Memoir 
on the "Fossil Proboscidea from China," defines the family as follows (p. 71): "The animals included in this 
family have skulls which resemble those of the true elephants but which are more primitive. They have very 
long sockets for the tusks, and the grinding teeth and palate are well below the plane of the occipital condyles. 
The grinding teeth remain brachyodont throughout the whole of their evolutionary history, but they parallel 
the teeth of the true elephants in showing a progressive increase in the number of ridges. This is especially true 
of the third molars. With this increase in the number of ridges, and its accompanying increase of length, there 
is an ever increasing curvature of the occlusal surface which reaches its maximum in certain specimens referred to 
S. airmmna and S. insignis. Owing to the short palate, it was impossible for the whole of the very long teeth to 
be accommodated in the upper jaw at one time. The curvature of the crown allowed the tooth to follow a more or 
less circular path which brought it from a position practically parallel to the plane of the occiput to the correct 
position for mastication. There is an ever increasing amount of cement, and the ridges show an increasing number 
of mammillae on their crests." 

"This sub-family comprises two groups''' of animals. One, with compressed, tectiform, ridges is given the 
generic name Stegodon Falconer & Cautley. The other, in which the ridges are blunter, and composed of round- 
ed conules, is known as Stegolophodon Schlesinger. Both genera occur in India, but, so far as is known, Stegodon 
is the only genus found in China." 

Subfamily: Stegodontin^ Osborn, 1918, 1921 

Original reference: Bull. Geol. Soc. Amer., 1918, Vol. XXIX, pp. 135, 136 (Osborn, 1918.468); Araer. Mus. Novitates, 1921, 
No. 1, pp. 12, 13 (Osborn, 1921.515). 

Subfamily Definition (Osborn, 1918.468, p. 136): "The Stegodontinse may be distinguished as 
a phylum confined to Asia, in which the grinding teeth remain brachyodont, short-crowned, although 
a very large number of cross crests evolve, especially on the posterior grinding teeth. From an early 
member of this subfamily, perhaps of Middle Miocene time, were given off one or more branches of the 
elephant and mammoth phyla." 

(Osborn, 1921.515, pp. 12 and 13): "We observe that the Stegodonts are persistent browsers, 
probably tropical, forest-living proboscideans. According to Pohlig, from the skeleton discovered in 



'[Apcording to Osborn, 1936, Vol. I, p. 700, by removal of the Stegolophodonts to the family Mastodontidse, the Stegodontidae embrace the true Stego- 
donts {Stegodon Falconer and Cautley) only. — Editor.] 



808 



OSBORN: THE PROBOSCIDEA 



Trinil, Java, they have short, massive bodies hke those of the Mastodontinse of the north temperate 
forests. The skull and tusks do not lead into either the Elephantina^ or the Mammontinae types . . . The 
distinctive featiu'e of the grinding teeth is the rapid multiplication of transverse crests which rise from 
the fornuila 4.5.6.6.7-8 in S. [Stegodon] ciiftii (Lower Pliocene) to 5.9.10.12.13 in S. insignis (Lower Pleis- 
tocene). Jaw rapidly abbreviated. Upper tusks straight, parallel, slightly upcurved (adapted to dense 
forests). Grinders brachyodont to subhypsodont, crests breaking up into small mammillae, valleys 
filling with cement." 
Li the first article, namely, "A Long-jawed Mastodon Skeleton from South Dakota and Phylogeny of the 

Proboscidea," 1918, Osborn placed the Stegodontinae under the Elephantidae (p. 135), and in an accompanying 

table mentioned the species Stegodon ganesa, S. ciiftii, S. bombifrons, and S. latidem. The last-mentioned species 

was made by Schlesinger the genotype of Stegolophodon. 

The above definitions were based on both Stegoloplwdon and Stegodon. After the removal of the former 

genus to the superfamily Alastodontoidea, family Mastodontidse, subfamily Stegolophodontinae, no revised 

definition of the subfamily Stegodontinae was given by the author. 

LATE Tl':inTARY (iEOLOGY OF INDIA 

The latest opinion regarding the geology of Lidia (the Siwalik Hills and Perim Island) will be found in Edwin 
H. Colbert's Memoir of 1935 on "Siwalik Mammals in The American Museum of Natural History," pages 6 to 55; 
in a summary on page 21 it will be noted that Doctor Matthew in 1929 assigned to the Siwalik beds a somewhat 
liigher position in the geologic time scale than Doctor Pilgrim in 1927. Compare also Chapter XXII of the present 
Memoir. "Geologic Succession of the Proboscidea," which has been written by Doctor Colbert. 

Throughout the present Volume, therefore, the later determinations will be inserted in square brackets or 
added in footnotes. It should be recalled that the chapters constituting this Volume were written about eight to 
ten years ago and were awaiting final revision by Professor Osborn. — Editor.] 

2. HABITS AND GENERAL CHARACTERS 

We observe that this subfamily [Stegodontinae^] includes forest-hving browsers, which probably developed in a 
Fig. 20. forested or semi-forested oriental region, ranging through 

India into Burma, China, Japan, and southward into 
Java, Borneo,^ and the Philippines. The ancestors of the 
subfamily may be found in Miocene deposits of western 
Europe\ A single tooth, named Mastodon {Bunolopho- 
don) longirustre Kaup forma nublalidem, has been de- 





Hfjihas c!>f/i.— The first (?) lolt upper true molar : from (lie .'^Iwaliksof liuniia. 
I. Tlie lowpr borflor of tlic fi;:;uro is t)ic inner border of tlie ppccimcn. 
(KromGaudrTS ' Kncliainements.) 

CoTYPE OK Stegodon ki.kpiiantoidks Glut (=cLii'rn Falconeu) 

Kig. 683. Type of Ekphas ciiftii Falponer and Cautloy, 1846, a first loft 
upper true molar, l.M', onc-lialf natural si:!e. The same six erestcd tooth 
api)ear,s in figures 686, 7()0, and 701. After Lydekker, 1S86.2, j). 81, fig. 20 
(taken from a woodcut m (laudry, 1878, p. 176, fij,'. 232). 

'[At the time this chapter was written the subfamily Stegodontina; was thought to embrace both the .Stegolophoduuts (now removed to the Mnstod.mlnidea 
subfamily Stegolophodontina;, p. 700) and the true Stegodonts. — Editor.] 



RbFERKED superior MOT.AII OF STE(i()DON OKIBNTAI.IS OKANOEKl 

I'^ig. 684. Referred first right .superior molar, r.M' (rev.) of Slnioilnn 
oricnlalis grangcri (Amer. Mus. 18530- wrongly numbered 18.530). Acluiil 
median length 181 mm. 



•'[Type locality of Slegolophodon lydckkeri Osborn, 1936 (see Vol. I, p. 700). 
'[Sec Vol. I, p. 197.— Editor.) 



-Editor.; 



THESTEGODONTIN.^: HISTORY 



809 



scribed by Schlesinger from near Teschen (Schlesien), Austria; this resembles Mastodon [ = Stegolophodon\ 
lalideiiH of Burma and is icferred in the present Memoir to Stegolophodon subldtidens (see Vol. I, p. 737; Vol. 11, 
p. 846). 

The skull is subelephantine in type, brachycephalic, brachyopic, the rostrum being elongated to support the 
tusks; the grinding teeth and palate are depressed far below the occipital condyles (bathycephalic). As in the 
elephants, the jaw is greatly abbreviated. The upper tusks are straight or slightly upcurved, elongating, without 
trace of enamel band, to a length of about 10 ft. Considering the large size of the tusks the skull is relatively 
small. The grinding teetli are brachyodont to subhypsodont, yet the ridge-crests in the posterior molars, M3, 
multiply from five plus [Stegolophodon] to fifteen plus [Stegodon], each crest breaking up into small nipples, mam- 
illae, or conelets. The lower incisors disappear very early. 

As in certain of the Mastodontidee and as in all the Elephantidse, the grinding teeth increase the number of 
their ridges by adding crests both in front and behind, as first observed by Falconer. Thus in M 1 of Stegodon 
orienlalis grangeri (Fig. 684) the ridge formula may be written: """g""" . The descending scale of the ridge formulae 
in the premolars and molars of the principal species of Stegodonts is appro.ximately as below. 



3. APPROXIMATE DESCENDING RIDGE FORMUL.^, AFTER FALCONER, LYDEKKER, MARTIN, 

AND OSBORN 

Observe that (1) the maximum upper ridge-crests, rising from 21e to 51e, are, so far as known, less numerous 
than the maximum lower ridge-crests which probably rise to more than 51 ; (2) this is compensated for by the fact 
that the upper molars are throughout broader than the lower molars; (3) the differences in width and in the 
number of ridge-crests of the upper molars in comparison with the lower molars are beautifully shown in figure 687, 
also in figures 759 and 762, the type of Stegodon orientalis grangeri. See also details of progressive ridge-crest 
formulae, talons and half ridge-crests, under each species. 

Maximum Estimated maximum 
conelets upper and lowek 

















PER RIDGE- 


RIDGE-CHESTS 
















CREST 






Dp 2 


Dp 3 


Dp 4 


M 1 


M2 


M3 




Dp 3-M 3 


Stegodon airdwana 


2^ 


5-6 


7-9 


7-9 


9-1 1 


1 2-1 4 
1 3-15V4 


13-20-1- 


U. L. 

51e 


Stegodon insignis \ 


i 


5-e 
6 


7-7W 
7Vi>-9 


:7V5-8 

7»-10 


71S-M-8 


lS-1 1-15 


11 


42 49 


Stegodon ganesa j 


7W-8-M.9-li 


12W-13 


Stegodon orientalis grangeri 


M-2->i 


5«-6 
+ 5W 


5'4-6->j 


lt-6-7-« 
S 


M-8-W 
(?) 91,4 


lS-1 l-« 
VS-13 + 


-11 


41 46 


Stegodon bombifrons 




i 


5-6 

T 


6 


7 
7K-8 


9 IS 
9V4 


11 + 


34e 


Stegodon elephantoides 








6M 


6W 


T¥ 


5-8 




Stegodon elephantoides {=c.liftii) 




£ 


6^ 


6M 


6H 


8U 


10-12 


28e 


Stegolophodon stegodonfoides 












6» 


5-6 




Stegolophodon latidens 




^ 


i 


Vt-i-Vt 


4«-5 
4VS-3 


M-5-W 
V5-6-W 


4-5 


23(?) 


Stegolophodon caidleyi 






W-3-'/j 


M-i-M 


A 


5-W + 


4-5 


21e 



Formula: Ridge-crests or Lophs (Dp 2-M 3). — The above ridge-formula table, a.ssembled from several 
sources, is approximate, first, because various observers differ in the method of counting the ridge-crests, second, 
because within each species the ascending mutations may lead up to the next higher progressive stage, as, for 
example, in the transition of Stegodon insignis-ganesa to S. airdwana of Trinil, Java. 



810 



OSBORN: THE PROBOSCIDEA 



The individual number of ridge-crests from Dp 2 to M 3 is mainly assembled from a very careful collation of 
the ridge-crest formulae given by Falconer (see below), supplemented by the observations of Lydekker, of Martin, 
and of Osborn. 

CoNELETS. — The conelets of Stegolophodon and of Stegodon arise chiefly by binary, rarely by ternary, fission 
of the primary cones. Starting with the original loph of the Palseomastodon stage, consisting of two cones, normal 
binary fission would produce: 

2 — 4-t- {S. cautleyi) — 8 (*S'. elej)hantoides) — 20-1- i>^- airdwana), etc. 

But the binary fission is not so regular as this. The newer anterior and posterior crests exhibit fewer conelets 
tluin the older mid-crests, so that the highest number of conelets usually will be found in the third and fourth 
crests. 




6 J crests 



STEQOLOPHODON 5TE60D0NT0IDE5 




4--^ coTtelets 



5TEQ0L0PH0D0N LATIDEN5 





STEQOLOPHODON CAUTLEYI 




PALAE0MA5T0D0N 

Fig. 68.5. Structural Evolution of the Cone.s, Conelets, and Ridqe-crbsts in the Steoo- 
LOPHODON Phylum, in comparison with Pal.«omastodon 

(1) PAL.EOMA8TODON, PRIMARY Type. Four jjiimary cones; ridgc-ficsts, jjioto-, meta-, ami 
rudimentary tritoloph. [See pp. 143 and (191 of Vol. I, for subfamily position of Valxomaslndon. — Editor. 1 

(2-4) Stegolophodon Phylum. Binary fission of primary cones into four to five coneUls, vestigial 
conules in S. cautleyi and .S. lalidens (protoconule = p.l., metaconule = m.l.); addition of trito-, tetarto-, 
pcnta-, and hexalophs. Gradual loss of median sulcus, presistent, however, in the first two anterior 
crests. 

Stegolophodon Uilidens (3, left) is a third lower molar, r.Ma, inserted for romparison, and has seven 
phis (7J2) ridge-erests, four plus conelets, and median sulcus. 





Lrni 

/5 crests 20 or r,,ore ConeZet 10 STEGODON AIRAWANA 

-fissures 




i-4 crests // conelets 



9 STEQODON IN51GNIS- QANESA 




/34 crests // or less conelets 8 STEGODON ORIENTALIS qPANQERI 





7 STEQODON BOMBIFRONS 




W crests 5 To 8 coneZets 



5TEGOD0N E LEPHANT0IDE5 




Iml 




a^ crests JO -^ coneZets 5 STEQODON ELEPHANT0IDE5 (--CLIFTII ) 

Fig. 686. Strdctdral Evolutiox of the Cones, Conelets, and Rxdge-crests in the Stegodon 

Phylum, in Ascending Order (5-10). 
(5-10) Stegodon Phylum. Binary or ternary fis.sion of the cones into conelets (5-20); addition of 
anterior and po.sterior ridge-crests (10-15 in M 3) ; addition of cement. 

(5) Stegodon ekphantmdes (=cliftii) with six and a ([uarter ridge-crests and ten plus conelets (M'). 

(6) Stegodon elephantoides with ten ridge-crests and five to eiglit conelets (M3). 

(7) Stegodon bombifrons witli nine and a half ridge-crests and eleven plus conelets (M^). 

(8) Stegodon orientalis gnmgeri with thirteen and a half ridge-crests and eleven or less conelets (M3). 

(9) Stegodon in&ignis-ganesa with fourteen ridge-crests and eleven conelets (M3). 
(10) Stegodon airhwana with fifteen ridge-crests and twenty plus conelets (M'). 



811 



gl2 OSBORN: THE PROBOSCIDEA 

Figures 685 and 686. — Particularly interesting and significant in the Stegodontinae is the transformation of 
the original cones by fission into conelets. Thus in the archetypal Palasomastodon molar there are two cones in 
the protoloph, in Stegolophodon latidens and »S'. cautleiji each cone splits into two, making four plus conelets in the 
metaloph; in *S. stegodontoides each of these four conelets tends to split into two, tending to form from five to 
eight conelets, but this splitting is not regular and no loph actually attains eight. In Stegodon elepkantoides the 
equal splitting gives rise to from^we to eight conelets; in S. elepkantoides ( = cliftii), S. bombifrons, and *S'. insignis- 
ganesa the fission gives rise to from eleven to twelve conelets, hence each ridge is finally surmounted by twelve cone- 
lets which, when slightly worn, present eight loops. In Stegodon airdwana, the most progressive species, the 
conelets range from thirteen to twenty plus. Thus the maximum number of cones and conelets in each crest runs 
as follows: 

Molar cones and conelets: Primitive (Palxomastodon) 2-4-6-8-12-20+ progressive {Stegodon airdwana). 

Ridge-crest Evolution. — The Stegodontinae also furnish a beautiful example of the evolution through 

w liich each ridge passes in turn, from the primitive submastodontoid type seen in Palseomastodon into the highly 

progressive subelephantoid type seen in Stegodon insignis-ganesa and S. airdwana. The superior (-loph) and 

inferior (-lophid) ridge-crests may receive a brief numerical terminology, namely: 

Pro-protoloph— id = One-half, anterior rudimentary ridge Pentalopli— id = Fifth ridge 

Post-metaloph— id = One-half, posterior rudimentary ridge Hexaloph— id = Sixth ridge 

Protoloph— id = First primary ridge = protocone and Heptaloph— id = Seventh ridge 

paracone of Ungulata Octaloph— id = Eighth ridge 

Metaloph — id = Second primary ridge = hypocone and Ennealoph— id = Ninth ridge 

metacone of Ungulata Decaloph— id = Tenth ridge 

Tritoloph— id = Third ridge Endecaloph— id = Eleventh ridge 

Tetartoloph— id = Fourth ridge Dodecaloph— id = Twelfth ridge 

Crest Addition. — In Stegolophodon and in the Mastodontidae the homology of the protoloph and of the 
metaloph is simple as compared with other ungulates, but since the increment of ridge-crests in the elephantine 
molar is by addition to both the anterior and posterior ridges, namely, the pro-protoloph and the post-metaloph, 
it soon becomes difficult to determine which ridge-crests correspond with the primary protoloph and metaloph 
of Palseomastodon and of other ungulates. 

Intermediate Molars Uniform. — A constant feature in the Proboscidea appears to be the uniformity of 
the three 'intermediate molars,' namely. Dp 4, Ml, M 2, which tend to have the same ridge-crest formula in 
each species, for example: 

Stegodon insignis-ganesa Dp 4 7K> Ml 7}^ M2 7K-8 = intermediate molars 

Stegolophodon latidens Dp 4 4 M 1 4^ M 2 4^-5 = intermediate molars 

Trilophodon angastidens Dp 43 M13 M23 = intermediate molars 

Consequently it may be difficult to distinguish these 'intermediate molars' from each other by the ridge formula 
alone; whereas they may be distinguished by the character of wear, by the width of the crowns, and by the 
condition of the fangs. 

Ridge-crest Elevation.— The progres.sive elevation of the ridge-crests in the Stegodontinae is illustrated in 
two diagrammatic figures (Figs. 687, 688), which demonstrate the constant progressive heightening of the ridge- 
crests as we ascend from the Lower(?) Phocene Stegodon sinensis and [Middle Pliocene] S. bombifrons into the 
Middle Pleistocene S. airdwana stage. The early phases of ridge-crest elevation (Fig. 687) may be compared with 
the later phases (Figs. 688, 781) as follows: 

Stegodon orienlalis Owen Stegodon aurorx Matsumoto 

Stegodon orienlalis grangeri Osborn Stegodon airdwana Martin 

Stegodon bombifrons Falconer and C'autley Stegodon insiqnis l-'alconer and Caul ley 
Stegodon sinensis Owen 



THE STEGODONTINiE: HISTORY 



813 



Descending Order of Species. — (1) At the summit of the known Stegodontinae is Stegodon airdwana, the 
most progressive both in ridge-crests and in conelets — a Lower to Middle Pleistocene stage. The following extract 
from a letter by Doctor Dietrich (March 10, 1924, and notes) confirms the Middle Pleistocene age of this species: 

For a long time I have been trying to prove that the Stegodon [airdwana] species from the Pithecanthropus strata [Trinil, 
Java] are the very youngest, that is, young Pleistocene. In detail: "Geologisches Alter. Aus morphologischen (Jriinden muss 
die javanische Art jiinger sein als die chinesische und alle bekannten kontinentalen Stegodon-Arten. Dass St. Airawana im 
Pleistocan lebte, wird \on fast alien Autoren angenonimen ; ihr pleistocdnes Alter diirfte gesichert sein. Erweist sich die chines- 
ische Stegodontenfauna als alt-pleistociin, dann ist die Trinilfauna jiinger als alt-pleistociin. Ich halte es aus geologischen und 
anderen Griinden sogar fiir wahrscheinlich, dass die Trinilfauna (und damit Pitheacanthropus) ju/i^-pleistociin ist; das wird sich 
mit Hilfe des St. Airawana bei besserer Kenntnis der kontinentalen Stegodonformen vielleicht erweisen lassen." 




5/ \y ^-'r.c/pl 



5TEQ0D0N OR1ENTALI5 
Type 



% 




STEQCDON ORIENTALIS QRANiqERl 
A.M./aiOS •%- 




B0MBIFR0N5 %-" 




% 



5TEQ0DON SINENSIS 
-Type 

Fig. 687. Gradual Progressive Hypso- 
DONTY IN Superior Grinders 

Stegodon orientalis Owen, type. Superior 
deciduous premolar, r.Dp'. 

Stegodon orientalis grangeri Osborn. Superior 
deciduous premolar, r.Dp'. 

Stegodon bombifrons Falconer and Cautley. 
Third superior molar, M'. 

Stegodon sinensis Owen, type. Superior 
deciduous premolar, r.Dp'. 

AH figures four-fifths natural size, e.\cepting 
Stegodon t>omhifrons which is four-fifteenths 
natural size. Observe that the ridge-crests (1-5) 
in transition from S. sinensis type to S. orientalis 
type are broad, progressively elevated, and 
approximated. 



znner yict^ All '/^ nat: si7ef 




STEQODON AIRAWANA rs/. 
Ca^t. A.M. 6 33S 




r. m.- 



AURORAE 



Type 




STEQCOON 

-Type 



IN5IQNIS 



Fig. 688. Gradual Progressive Hypsodonty in Superior 
AND Inferior Grinders 

Stegodon airawana Martin. Middle Pleistocene of Java. 
Elevated and approximated ridge-crests (9-l.j}'>), right third 
inferior molar, r.Ms. 

Stegodon aurorx Matsumoto, type. Middle Pleistocene 
[?Upper Pliocene] of Japan. Elevated and closely approximated 
ridge-crests, right second superior molar, r.M^. 

Stegodon insignis Falc. and Caut., type. Lower Pleistocene, 
Upper Siwaliks, Boulder Conglomerate, of India. Third superior 
molar, M', with eleven-)- ridge-crests. 

All figures to same scale, one-third natural size. Observe that 
in Stegodon airawana and S. aurorx the ridge-crests are much 
more elevated and approximated than in S. insignis, but that S. 
irisignis is less elevated than S. orientalis (Fig. 687). 



Table III 



FSTEGOLOPHODON PHYT.UM 



RTECODON PHYLrM 





India, Burma 


1 
Japan Borneo 


India, Burma 


China, Japan 


Philippine Islands, Java 










Stegodon in- 






Upper 








signis ref. 






Pleistocene 








Stegodon gan- 
esa ref. 
















Stegodon ori- 


Stegodon airdwana type 












entalis 


Stegodon ganesa var. 


Middle 










shodoensis 


javanicas type = .S. airdwana or S. 


Pleistocene 










type 


trigonmephalus 
Stegodon {A rchidiskodon?) 
mindanensis type 


Lower 








Stegodon pin- 


Stegodon ori- 


Stegodon Irigonocephalus type 


Pleistocene 








jorensis 
type 


enlalis type 
Stegodon ori- 
entalis 




a> 










grangeri 




g 








Stegodon in- 


type 












signis lecto- 






{1, o § 








type and ref. 






K^^ 








Stegodon gan- 














esa lecto- 












type and ref. 








Stegolophndon 


Slegolophodon 


Stegodon 


Stegodon 




] Upper 


slegodonloi- 


lydekkeri 


ins ignis 


aarorx 




Pliocene 


des type 




type 


birnianicus 
type' 
Stegodon bom- 
bifrons ref. 


type 






Slegolophodon 






Stegodon bom- 


Stegodon bom- 






cautleyi 






bifrons lec- 


bifrons ref. 






lectotype 




totype 






Middle 


Slegolophodon 




Stegodon bom- 






Pliocene 


lalidens ref. 




bifrons ref. 








(AUSTRIA 














Slegolophodon 














sublalidens 














type] 










i 




Slegolophodon 


Slegolophodon 


Stegodon ele- 


Stegodon 






lalidens 


latidens 


phnnloides 


sinensis 




Lower 


lectotype and 


ref. 


1 


lectotype' 


type 




Pliocene 


cotype' 
Slegolophodon 
lalidens 
ref.' 






Stegodon ele- 
phanloides 
{=^clijtii) 
cotype' 








Slegolophodon 












caulleyi 














pi-ogressus 












Mio-Pliocene 


type 
Slegolophodon 
nathotensis 
type 













'[See note on page 824 below, in which the geologic age of Slegolophodon. latidens is given as Lower Pleistocene. As iSf. elephantoides and S. elephantoides 
( = diftii) were found in the same locality and at the same geologic level, they should al.w be regarded as of Lower Tleistocene age. Note also that S. birniani- 
cus is placed in the Lower Pleistocene. Compare (>)lbort, Chap. XXII, pp. 14.50, 14.51.— Editor.] 

814 



THESTEGODONTIN^: HISTORY 815 

(2) Next in descending order are referred Stegodon insignis and S. ganesa, of Lower to Middle [Upper] Pleisto- 
cene age. (3) Third in descending order is Stegodon orientalis grangeri, slightly more primitive than S. insignis 
both in ridge-crest and conelet progression — probably of Lower Pleistocene age. (4) Next is the [?Upper Pliocene] 
Stegodon auroras of Mt. Tomuro, Japan. (5) Far more primitive both in ridge-crests and conelets are Stegodon 
bombifrons [of the Middle Pliocene] and *S'. elephantoides ( = cliftii), known to be of Lower Phocene (type) age.' 
It is probable that the referred *S. bombifrons of the Middle [Upper] Pliocene, Tatrot horizon, will prove to be some- 
what more progressive than the lectotype of S. bombifrons of the Middle Pliocene, Dhok Pathan horizon. 

(6) Representing the Stegolophodon phylum is the Upper(?) Pliocene Stegolophodon stegodontoides type, of 
approximately the same age as Stegodon ganesa lectotype and S. orientalis grangeri type. (7) Stegolophodon 
latidens ref. is recorded in the same Middle Pliocene (Dhok Pathan) geologic level as Stegodon bombifrons lecto- 
type. (8) The Upper Miocene [Middle Pliocene] yields Stegolophodon cautleyi lectotype of Perim Island (Dhok 
Pathan). (9) The Middle Miocene [Mio-Pliocene] yields Stegolophodon cautleyi progressus and S. nathotensis. 

4. GEOLOGIC AND DIPHYLETIC ORDER OF THE STEGODONTIN^ 
The geologic level of Falconer's types of Elephas [ = Stegodon] ganesa and E. [ = S.] insignis, although unre- 
corded, is probably Pinjor (see Vol. I, Fig. 413). Pilgrim writes (letter of January 26, 1927) : "I am now almost 
convinced that the bulk of the fossils were collected from that zone, i. e.. Upper Pliocene [Pinjor zone, Moginand], 
but that some came from the Boulder Conglomerate zone Falconer distinctly states." As observed by Barnum 
Brown, the exact geologic horizon of the type and referred specimens of S. insignis and S. ganesa is uncertain, 
because apparently in the Pinjor horizon are obtained jaws and skulls redeposited from the Boulder Conglomerate 
above. 

5. HISTORY OF DISCOVERY OF THE SUBFAMILY STEGODONTIN^. PRINCIPLES OF TYPE 

REVISION OF THE SPECIES 
Genera. — Up to 1924 sixteen species of [Stegolophodonts and] Stegodonts had been discovered and described 
between the years 1828 and 1917, in the order shown in the list in Section II below. In 1924 Matsumoto described 
a new subspecies from Japan, namely, Stegodon orientalis shodoensis. In 1929 Osborn described Stegodon orientalis 
grangeri from the pits near Wanhsien, Szechuan, China, Stegodon insignis birmanicus from the Pliocene of Burma, 
Irrawaddy River, and Stegodon pinjorensis from the Lower Pleistocene (?Pinjor) of India; also two Stegolopho- 
donts (Stegolophodon nathotensis and S. cautleyi progressus). All the early species were described either as 
Mastodon {e.g., Mastodon latidens, Mastodon elephantoides) or as Elephas (e.g., Elephas insignis, Elephas ganesa). 
Although the name Stegodon appears as early as 1847 ("Fauna Antiqua Sivalensis," Falconer and Cautley, PI. 
XLii), it was not until the year 1857 (pp. 314, 318, table opp. p. 319) that Falconer ventured to select Elephas 
cliftii, E. bombifrons, E. ganesa{?), and E. insignis as representing a new subgenus, Stegodon. 

It was in 1922 that Matsumoto in a letter to the present author announced his intention of making Mastodon 
latidens the type of a new genus, Prostegodon. The name, however, is preoccupied by the Stegolophodon of Schle- 
singer, 1917 (genotype Mastodon latidens Clift), in which genus Osborn united (1929) the four species M. latidens 
Clift, M. cautleyi Lydekker, M. stegodontoides Pilgrim, and M. (Bunolophodon) longirostre Kaup /orma sublatidens 
Schlesinger, also Stegolophodon nathotensis Osborn from the lower Middle Siwaliks and Stegolophodon cautleyi 
progressus Osborn from 2,000 feet above the base of the Lower Siwaliks, India. ^ 

'[See footnote on preceding page (p. 814) regarding the Lower Pleistocene age of Stegodon elephantoides i=diftii). — Editor.] 

^[To these six species should be added Stegolophodon lydekkeri Osborn, described in Vohime I of the present Memoir, page 700. — Editor.] 



816 OSBORN: THEPROBOSCIDEA 

[Recently six additional species of Stegodon have been described, namely, Stegodon boiidolensis van der Maarel 
1932, and *S'. trigonocephalus praecursor von Koenigswald, 1933, from Java, *S. officinalis, S. zdanskyi Hopwood, 
1935, and S. yunhensis Young, 1935, from China, also Paraslegodon [Stegudon?\ sugiyamai Tokunaga, 1935, from 
Japan. The generic determination of Tokvmaga's species Paraslegodon? kwantoensis, 1934, has not been given in 
the present Memoir, owing to Professor Osborn's views regarding the genus Paraslegodon (see next paragraph). — 
Editor.] 

The genus Paraslegodon of Matsumoto, 1924, belongs to the genus Archidiskodon or to a progressive 
Slegodon, as the genotypic species, Elephas {Paraslegodon) aurorx, is slightly more primitive than Archidiskodon 
planifrons. 

Principles of Type Revision. — The consideration of these matters of descent and phylogeny, however, 
must be preceded by a very rigid review of each species of Stegodont in the order of its original description, 
following the standard methods of type analysis established throughout this Memoir, namely: 

1) Determination of the actual type specimen or the specimen first mentioned among a series of cotypes. 

2) Fixation of the original lype figttre, the one first published by the author or selected by the author 

from other publications. 

3) Enumeration of the lype characters observed in the type specimen by the author or by subsequent 

observers. 

4) Elimination from the type list of characters foimded on referred specimens which do not actually 

belong to the same species as the type. 

5) Determination of the type locality and distinction of the topotypes. 

Original Descriptions. — The revision of these various species of Stegodonts, on the strict application of 
the five rules above, has been a long and very difficult task. In order to establish absolutely the original author's 
intention, the author's original description is cited in full. When too prolix, as in the case of many descriptions by 
Owen and by Lydekker, excerpts are made in the author's own language. 

This revision and establishment of type characters must be followed by a restudy and revision of referred speci- 
mens, which in most cases can only be done in future by monographic research directly upon the specimens them- 
selves, amphfied by knowledge afforded by fresh materials. Consequently the present Stegodont chapter lays no 
claim to completeness or finality ; it leaves many questions wholly undecided, for example, the sexual or specific 
dental or cranial distinctions between Slegodon insignis and S. ganesa. 

In the meantime our present establishment of the types and of the type specific characters based upon the 
type specimens themselves as well as of characters derived from properly referred specimens in the same geologic 
horizon, as in the case of Slegodon bombifrons, may lay a firm foundation for future monographic research. 

6. THE STEGODONTINJ^ AND MASTODONTIN/E OF CHINA 
Two species of Stegodonts described by Owen in 1870, namely, Slegodon orienlalis and S. si^ietms, were 
determined from (collections l)rought in by dealers for medicinal purposes and not procured in sitii by palieontolo- 

gists. 



THE STEGODONTIN^: HISTORY 817 

a) Koken's Notes of 1885 and Schlosser's Notes of 1903 
Koken, "Ueber fossile Siiugethiere aus China," 1885, pp. 31-44; Schlosser, "Die fossilen Saugethiere Chinas," 1903, pp. 43-49 
Koken's specimens were also probably collected from the caverns of Yunnan by dealers for von Richthofen; 
his list is as follows (Koken, 1885, p. 33) : 

Proboscidia. 

1. Mastodon perimensis var. sinensis Yiinnan. 

2. Mastodon aff . Pandionis Yiinnan. 

*3. Stegodon Clijtii Shanghai ; oberer Hoangho (West-Kansu). 

*4. Stegodon insignis Yunnan ; Szechuen. 

5. Stegodon aff. bombifrons Yunnan. 

Schon friiher beschriebene oder erwahnte Arten sind durch *ausgezeichnet. 

The type fragment.s all appear as though they had been brought in by native collectors, for medicine dealers, 
since they consist of broken teeth only; the same is true of the Schlosser collection. Consequently we regard the 
specific and generic determinations made from these imperfect materials as of provisional value. The species of 
Proboscidea contained in Schlosser's article of 1903, following Koken's review of 1885, are as follows: 

Reference in 
Present Memoir 
Fokien, Kansu Stegodon orientalis Owen, 1870 = Stegodon insignis Falconer (fide 

Koken, 1885, fide Schlosser) Stegodon orientalis 

?Shanghai Stegodon sinensis Owen, 1870 = Stegodon clifti Falconer (fide Koken, 

1885), = Elephas clifti (fide Lydekker, 1886) Stegodon sinensis 
Stegodon bombifrons Falconer, 1846 = Stegodon aff. bombifrons (fide 
Koken, 1885), = Elephas bombifrons (fide 

Lydekker, 1886) IStegodon bombifrons 

Rothe Thone Mastodon latidens ("lift, 1828 = Mastodon aff. latidens (fide Schlosser) IStegolophodon latidens 

= Shansi 
Rothhche Sande = Mastodon Lydekkeri Schlosser, 1903, related to Mastodon latidens 

Tientsin, Honan, etc. (fide Schlosser, 1903), to Serridentinus (fide 

Osborn) 1 Serridentinus lydekkeri 

Mastodon perimensis var. sinensis Koken, 1885= Tetralophodon (Ly- 

dekkeria) sinensis (fide Osborn) Tetralophodon (Lydekkeria) sinensis 

Rothe Thone Mastodon pandionis Falconer (fide Koken, 1885) = Incertse sedis (fide 

= Shansi Osborn) Incertx sedis 

b) American Museum Discovery of Stegodon orientalis grangeri near the 

Yangtze River, China 

The first scientific party to collect fossils in situ in China, with records of the actual locality and geologic level, 
was that of the American Museum, under Walter Granger, in 1920-1921. 

In a preliminary notice of a collection secured during the winter of 1920-1921 by Dr. Granger of the Central 
Asiatic Expedition, Matthew and Granger (1923) described the material as occurring in a series of pits or fissures 
at the village of Yenchingkou in the vicinity of Wanhsien, province of Szechuan, about one hundred and forty 
miles distant in an air line from Chungkingfoo, the type locality of Stegodon orientalis Owen. This Yenchingkou 
material includes a fairly complete adult skull, two young skulls, a series of palates and lower jaws, and many 
teeth, which Matthew and Granger figured and partly described (pp. 567-571, figs. 3-6). They remark (p. 567) : 
"Stegodon orientalis Owen. Schlosser regards this species as identical with *S. insignis of India, basing the reference 
upon the fragmentary teeth described by Owen. Matsumoto regards it as distinct, upon the evidence of the 
referred material which he describes and figures. The Yen-ching-kao material includes a fairly complete adult 
skull, two young skulls, a series of palates and lower jaws and many teeth. It should enable us to estimate the 
affinities of the species more exactly when it has been cleaned up and studied." They deferred further description 
of this fine material to Osborn (see below, pp. 875-881). 



818 OSBORN: THE PROBOSCIDEA 

Geologic Age of Stegodon orientalis and S. orientalis grangeri Types. — Matthew and Granger 
(1923, pp. 563, 565) observe: 

The Chinese fossil mammals described by Owen in 1870 [Footnote: '(iuar. Journ. Geol. Soc, London, XXVI, jjp. 417-4.36, 
Pis. xxvii-xxix.'] came from 'a cave near the city of ( 'hung-king-foo in the province of Sze-chuan.' Chung-king is on the 
Yang-tse-kiang above Wan-hsien, about one hundred and forty miles distant [southwest from the ( iranger locality of Ycnching- 
kouj in an air line. . . . Possibly the Chinese informants of Consul Swinhoe, who sent the fossils to Owen, misled him, uninten- 
tionally or deliberately, as to the locality. . . . Owen regarded the [Chungkingfoo] fauna as Pliocene and described the follow- 
ing species: 

[Chungkingfoo] [Yenchingkou] 

Stegodon orientalis [type]. Parts of molars. [S. orientalis grangeri type 

Rhinoceros sinensis [type]. Parts of 4 upper and 4 lower molars. R. sinensis ref. 

Tapiriis sinensis [type]. Parts of 3 upper and 4 lower teeth. Tapirus sinensis ref. 

Chalicotheriimi sinense [type]. Part of an upper molar. Chalicotherium sinense ref. 

Hyaena sijiensis [type]. Canine, 2 premolars. Hyxna sinensis ref.] 

Owen's descriptions and figures accord very well with some of the species in our collection [the Yenchingkou collection 

of the .American Museum], so that we have referred them to his species, whether or not later investigation proves them to be 
exact topotypes. 

7. PLIOCENE TO PLEISTOCENE PROBOSCIDEA OF JAPAN 
Discovery. — Matsumoto's discoveries and descriptions of the Japanese Proboscidea up to 1924 are recited in 
more detail at the close of this chapter (pp. 901 to 909). His Elephas {Prostegodon, Parastegodon) aurorse, 1915, 
1918, 1924, from a compari.son with Stegodon airdwana of Java, appears to belong to Stegodon aurorse rather than 
to represent a distinct genus, i. e., Parastegodon. Matsumoto (1918, pp. 51, 52) gives the history of discovery of 
Stegodon and Elephas in Japan, beginning with Leith Adams (1868), Naumann (1881), Brauns (1883), Lydekker 
(1886), Martin (1886 [1887]), Tokunaga (1906), Sato (1914), Kato (1914), Matsimioto (1915, 1918). The Japanese 
species of Stegodon referred by Naumann to S. cliftii and .S'. insignis have been transferred by Matsimioto to 
Owen's Chinese species Stegodon sinensis and S. orientalis. Matsumoto remarks as to sexual characters (op. cil., 
1918, p. 52): 

Thus, the present writer's opinion [Footnote: 'This Vol., p. 10.'] that, St. sinensis, as well a,s St. orientalis, is geologically 
younger than ,S7. diSlii, evidently holds true also in the .Japanese specimens. One evidence noticeable is that the Stegodont 
species are usually found in couples. For example, St. cliftii and bombifrons are found associated with each other from the Dhok 
Pathan to the Tatrot horizon, St. ganesa and insignis from the Boulder Conglomerate horizon and the Lower Pleistocene of 
Narbada, .S7. orientalis and sinensis from the Uppermost Pliocene to Lower Pleistocene of China and .Japan, and .S7. airdwana 
and trigonocephalus from the Lower [to Middle] Pleistocene of Java. One may imagine the po.ssibility, that each couple of 
species represent sexual dimorphism of one and the same species. 

Matsumoto (1924, 1926, 1927) continues the history of discovery of the Stegodonts and Elephants of 
Japan up to the year 1927, as set forth in detail in this chapter, and presents his recent views as to the phylugeny 
of these animals as shown in figures 791 , 792, and 793. To the fossil fauna of Japan he adds the following five re- 
ferred and new species {op. cit., 1926.1, p. 1): 

Cenus Stegodon Falconer k Cautlcy. 4. ,S'. clifti Falconer & Cautley, ibid. [.Journ. Ceol. Soc. Tokyo, Vol. XX.Xl], 1924 
|1!)24.:^|, 1). :327. .\kira-mura, Kage District, Province of Isc— ijossibly Plaisancian-.\stian. 5. .S. sinensis Owen, ibid., 1924, p. 
328. Island of Sliodo (Shodo-siiima or Shozu-siiima), Inland 8ea — Milazzian-Tyrrhenian. 6. .S. orientalis Owen, ibid., 1924, p. 
330. (Also as S. bombifrons, ibid., j). 329.) Nagahama, Minato Town, I'rovince of Kazusa; Togane Town, same province— 
Cakbrian. Riuge, Ikadachi-mura, Province of Omi~Calabrian or possibly Cromerian. Okimisome, I'be Coal-Field, Province 
of Suwo — Cromerian. 7. S. orientalis ■shddoensi.s, nov., ibid., ]). 333. Islands of Mitsugo (.\Iitsugo-shima) and Island of Siiodo, 
Inland Sea; off Nagasaki, Eastern Sea— Milazzian-Tyrrhenian. Kashiwazaki, Province of Echigo (?this form). 

Of tlie al)()V(' the genus Parastegodon Matsumoto, 1924, is regardeil in tiie i)re.sent Memoir as belonging to 
the genus Archidiskodon or to a progressive Stegodon. 



THE STEGODONTINiE: HISTORY 



819 



8. PHYLOGENETIC DISCUSSION OF THE THIRTY DESCRIBED SPECIES OF STEGODONTS 

AND STEGOLOPHODONTS 

Double or Multiple Phyla of the Stegodonts. Schlosser, 1903. — Schlosser (1903, p. 191), in his 
revision of the Stegodonts of China, doubtfully suggested that west European species, originally described as 
Mastodon turicensis [ = il/. tapiroides, M. pyrenaicus, etc.], separated as Zygolophodon by Vacek in 1877 and (1926) 
by Osborn as Turichis (M. turicensis, M. tapiroides) and Zygolophodon {M. pyrenaicvs), may have given rise to 
such Upper Miocene [Middle Pliocene] Stegodonts of India as Stegolophodon cautleyi. 

Schlosser was also the first to suggest {op. cit., p. 191) that Mastodon turicensis [=Turicius tapiroides] of 
the Lower Miocene of Europe may have given rise to the Mastodon [=Stegolophodon] latidens of the Lower Pliocene^ 
of Asia from which in turn sprang off the true Stegodonts, such as Stegodon insignis. 

More in detail, Schlosser, who was the first to discuss the double phylogeny of the Stegodonts {op. cit., p. 
206), separated them into two divisions as follows: 

Mastodon latidens Clift Slegodon ganesa Falconer and Cautley 

Mastodon cautleyi Lydekker Stegodon insignis Falconer and Cautley 

Stegodon cliftii Falconer and Cautley 
Stegodon bombifrons Falconer and Cautley 

Two Phyla Suggested by Pilgrim, 1913. — The second to discuss the phylogeny of the Stegodonts was 
Pilgrim. The foundations of a diphyletic arrangement of the Stegodonts were laid by his observation (1913, 



JjOWEK Miocene. Turicius (4,5) compared with 

Trilophodon (2) 

Fig. 689. Left third inferior molar, I.M3. Comparison of 

Turicius turicensis [= tapiroides] (upper 4, .5) with Trilophodon 

pontileviensis (lower 2), one-half natural .size. After Mayet, 

1908, PI. XI, figs. 2, 4, and 5. 

These molar teeth exhibit the profound difference between 
the zygolophodont molar (above), resembling that of a primitive 
Stegodont with uninterruiited, widely open valley.s, and the 
bunolophodont molar (below), in which the valleys are closed by 
the central collides. 

(Upper) Figs. 4, 5. — "Mastodon turicensis [ = T. tapiroides]. 
Pontlcvoy. Dcrniere molaire infcrieure. Paris Museum. 
Grandeur naturolle." Fig. o. — "Id. vue d'en liaut." See also 
figure 138C, CI. 

(Lower) l''ig. 2.- "Mastodon angustidens. l''aluii de Pont- 
levoy. Dernicre molaire supcrieure. Communiciuee par M. .Joan 
de Bodard, Pontlevoy. Grandeur naturelle." [ = Trilophodon 
pontileviensis.] 



pp. 293, 294) that Mastodon [ = Stegolophodon] latidens occurs in the same Lower [Middle] Pliocene beds with 
Stegodon bombifrons. Pilgrim's discussion of the phylogeny of the Stegodonts may be paraphrased as follows: 

(a) In the lower deposits of Perim Island there appears a species Mastodon cautleyi representing a line of 
evolution which in all its earlier stages is entirely unknown in Europe [see Turicius (?) nnd Zygolophodon (?)]. 
The lectotype of M. cautleyi is a last upper molar, figured by Lydekker in 1886 (1886.1). 

'[See note on page 824 below where Stegolophodon latidens is given as of Lower Pleistocene age. — Editor.] 




820 



OSBORN: THE PROBOSCIDEA 



(b) In the higher Perim levels, as well as in the Middle Siwaliks of Lower [Middle] Pliocene age, is found 
a further development of this type in the species Mastodon latidens, with an increased number of ridges. The 
larger form of M. latidens is to be regarded as a direct descendant of Mastodon cautleyi. The last species of masto- 
dont to be referred to this Une is represented by a tooth from Lehri, for which Pilgrim proposed the name Mastodon 
[Stegolophodon] stegodontoides. 

(c) The type tooth of M. stegodontoides, of which the horizon is uncertain but may possibly be Upper Siwalik, 
is distinguished from that of M. latidens by the almost entire absence of accessory columns [conules of Osborn] ; 
M. stegodontoides, like M. latidens, carries on each of its ridges four columns [ = conelets of Osborn], while the 
anterior ridges of the next higher stage, Stegodon elephantoides {=cliftii),^ carries nine or ten mamillse [=conelets of 
Osborn]. 

(d) So close is M. stegodontoides to S. elephantoides ( = cliftii) that it is hard to separate the two genera. 

(e) The true Stegodon type represented by S. bombifrons appears in the Lower [Middle] Pliocene, Dhok 
Patlian horizon, before the extinction of the Mastodon [ = Stegolophodon] cautleyi-latidens-stegodontoides phylum 
which is parallel with it. 

Pilgrim accordingly divides the Stegodon tinse into two generic phyla: to the first he applies the name Masto- 
don [ = Stegolophodon], to the second he applies the name Stegodon. This may be graphically represented as 
follows: 



Mastodon Series [ = Stegolophodon] 
Upper Pliocene Mastodon stegodontoides 
Lower Pliocene' Mastodon latidens 
Upper Miocene [Middle Pliocene] Mastodon cautleyi 



Stegodon Series [ = Stegodon] 
Stegodon bombifrons [Middle Pliocene] 
Stegodon cliftii [ = Stegodon elephantoides {- 
Lower Pliocene]' 



■cliftii), 



SCHLESINGER, 

latidens Clift. 



1917. — Schlesinger recognized the distinctness of Stegolophodon, basing his type on Mastodon 



Matsumoto, 1922. — The fact that Mastodon latidens appears in the same Lower [Middle] Pliocene geologic 
horizon as S. bombifrons is very significant; it tends to support the idea that the Stegodonts were diphyletic. 
This idea is perhaps carried a step further by Matsumoto who writes (letter, Nov. 20, 1922) : "In my report just 
in preparation on the Japanese 'Mastodonts,' I follow you to refer 'Mastodon' latidens to the genus Stegodon 
creating however a subgenus Prostegodon for it. Prostegodon is the primitive representative of the Stegodon- 
phylum, representing half bunomastodontine and half stegodontine dental characters. Schlosser's opinion, that 
Prostegodon might be ? a descendant of 'Mastodon' turicens, does not appear to be correct at all." 

Osborn, 1923. — Osborn (1923.601, ji. 2) erroneously adopted the generic name Prostegodon Matsumoto, 
based on the genotypic species Mastodon latideyis Clift. Prostegodon, however, is preoccupied by Stegolophodon 
Schlesinger, 1917. Thus the diphyletic arrangement of the Asiatic and European species and genera would appear 
as follows: 



Stegolophodon stegodontoides 
Stegolophodon cautleyi 
Stegolophodon latidens 
Turicius (?) sp. 
Zygolophodon (?) .sp. 



Stegodon ganesa 

Stegodon insignis 

Stegodon bombifrons 

Stegodon elephantoides ( = cliftii) 



Matsumoto, 1924, 1920. -A more recent stej) is lliat of Mntsutiiol-o in a dingrain enii)odied in a, letter 
dated Sendai, November 20, 1924, in which the polyphyletic Stegodontinae are tlivitletl theoretically into five 

'[See footnote on page 824 regarding the Lower Pleistocene age of Stegodon elephantoides (=cliftii). — Editor.] 



THESTEGODONTIN^: HISTORY 



821 



phyla (Fig. 792), and Trilophodon [ = Zygolophodon] pyrenaicus (Fig. 791) is placed as ancestral to Tetralophodon 
falconeri, Parastegodon [ = Stegolophodon] lalidens, P. [ = S.] stegodontoides, and Stegodon. 

In other words: (1) The type of Mastodon { = Zygolophodon\ pyrenaicus of the Middle Miocene or its an- 
cestors Zygolophodon pyrenaicus aurelianensis Osborn of the Lower Miocene of western Europe may be ancestral 
to the Stegolophodon and Stegodon species of southern Asia. 



Ji^: S 00. 



Rg:3. 





B 




Fig. 230. — l)i.Tniijre niobire inrunciiie du yiustodon luricenms (tujnroidei;) , 
aux 2/5 de grandeur. — Miocuiic moycii dc Siniorrc, C.crs. (D'aprus Lartet.) 

TuRicius (C) AND Stegolophodon (A, B) Form of Grinding Teeth 
Compare with Stegodon Molars (Fig. 686) 
Fig. 690. (C) Type third right inferior molar, r.Ma, of Turicius turicensis simorrensis Osborn, 1926, erroneously determined by Lartet (1859, PI. xv, fig. 3) 
as Mastodon lapiraides, two-fifths natural size. Upper Middle Miocene of Simorre. After Gaudry, 1878, p. 174, fig. 230. Reversed in drawing. See also 
Vol. I, pp. 207 and 220 of the present Memoir. 

Cotype and lectotype of Mastodon [=:Stegohphodon] cautleyi Lydekker, 1886. 

A (Cotype). First superior molar of the left side, l.M', one-third natural size. After Falconer and Cautlcy, 1846 [1847, PI. XL, fig-s. 3, 3a], as "Mastodon 
lalidens." Length 4 inches, width 2.3 inches. Brit. Mus. M.2817. Cast Amer. Mus. 26965. Perim Island. See Lydekker, 1886.1, p.xv, fig. 5. 

B (Lectotype). Third superior true molar of the left side, l.M', one-third natural size. After Falconer and Cautlcy, 1846 [1847, PI. xxxi, figs. 6, 6al as 
"Mastodon latidcns." Length 8.5 inches, width 4.5 inches. Brit. Mus. M.2705. Cast Amer. Mus. 26966. Perim Island. See also Lydekker, 1886.1, p. xv, fig. 
6, and 1886.2, p. 73, fig. 18. Same as figures 141 and 142 of Volume I of the present Memoir. 

(2) Matsumoto held (1924) that the Stegodonts are not merely diphyletic (e. g., Stegolophodon and Stegodon), 
but polyphyletic by subdivision of the species of Stegodon into five distinct hnes of descent, as clearly displayed in 
figure 792. 



Osborn, 1927. — The present author (1927) takes the more conservative view that there are certainly two 
distinct phyla, namely: (a) Stegolophodon cautlexji, S. lalidens, S. stegodontoides, and (b) Stegodon elephantoides 
{ = cliftii) to S. airdwana. The latter appear to present a progressive series in the increasing number of conelets 
and ridge-crests or lophs, but without very marked divergence, as shown in the geologic succession table (Table 
III, p. 814). 



822 OSBORN: THE PROBOSCIDEA 

Suggested European-Asiatic Origin and Migration of the Primitive Stegodonts 
The suggestion by Sehlosser (1903), rejected by Matsuinoto (1922), that the species of StegolopJwdon may 
1)0 derived from mastodonts related to the Turicius or Zyyolophodon of western Europe, should not be lightly 
dismissed. These west European animals have been grouped in the genus Zygolophodon by Vacck (1877), based 
on the three species Mastodon borsoni, M. turicensis, and M. tapiroides ( = M. pyrenaicus). The genus Zygolo- 
phodon embraces a type molar tooth fundamentally homologous and analogous, as shown in the accompanying 
comparison of the lectotype of Mastodon [ = StegolopJwdon] cautleyi from the Middle Pliocene of Perim Island, 
with a third inferior molar erroneously referred by Lartet (1859) to M. tapiroides [ = type of Turicius turicensis 
simorrensis Osborn, 1926— Fig. 690] from the Middle Miocene of Simorre. Molars of the Turicius turicensis 
[ = tapiroides] type occur in the Lower Miocene, Burdigalian, of the Falun de Pontlevoy, contemporaneous with 
tlie referred Trilophodon angustidens [ = T. pontileviensis — Fig. 689, 2]. 

Actual relationship to primitive Stegodonts [Stegolopiiodontina^] of Asia is rei)rescnted by the Mastodon 
{Bunolophodon) longirostre Kaup forma suhlatidens of Schlesinger from the Pliocene of Teschen (Schlesien), 
Austria (Fig. 722). A striking analogy to the Stegodon type is seen in the molar teeth referred by von Meyer to 
M. [Turicius] turicensis from the lignites of Elgg and Kapfnach, and finally in the strictly stegodont Mastodon [T.] 
virgatidens of von Meyer (Fig. 168). The only way to test this theory, however, is to place models or casts of the 
teeth of Zygolophodon, of Turicius, and of Stegolophodon side by side to see whether they compare in close detail, in 
which case the genus Zygolophodon Vacek would replace the genus Stegolophodon Schlesinger. 

Conclusion: Probable African-European-Asiatic Origin and Migration of the 

Primitive Stegodonts 
Osl:)orn, 1927' : If, as now appears probable, (1) the Trilophodon. pliylum first arrived in southern Europe and 
migrated eastward into India, (2) it is also probable that certain primitive species of the forest-living Zygolopho- 
don or Turicius phyla gave ri.se in Lower Miocene time to forest-living animals which spread into the forests of 
southern Asia and developed into the Stegodon series, as first adumbrated by Schlosser (1903). 

The European species actually resembling these animals is the Mastodon (Bunolophodon) longirostre Kaup 
forma sublatidens Schlesinger from Schlesien, Austria (Fig. 722), as described below. 



II. TYPE REVISION OF THE SPECIES IN ORDER OF ORIGINAL DISCOVERY 

AND DESCRIPTION 

In the previous Section I of the present Chapter XIV we have discussed the habits and general cluiracters, the 
lidge formulae, the geologic order, the history of discovery, the principles of type revision of the species, the 
Stegodonts of China, Japan, and the East Indies, the phylogcnetic succession, and, finally, the probable origin of 
the Stegodontinse [and Stegolophodontinse] in western Europe, and hence more remotely in Afri(^a. 

We now pa.ss in Section II to the very complicated subject of the type revision of the species on tlu> jjriiiciples 
enumerated above (p. 816). For this purpose we will review the thirty species described in I he y{>ars between 
1828 and 1936 in the order of their description, quoting extensively from the original type descriptions and re- 
producing every available type or lectotype figure dir(M;(ly after the oiiginal author. 

'[Cami>arc Vol. I, i>ii. 19-3, H)7, iilso Pis. ii lo iv.— Editor.) 



THESTEGODONTIN^: HISTORY 



823 




Fig. 691. Geograpliic distribution of the iwiiicipal species of Stcgolophodon and Slcgodon. Tlie white dots within tlie black areas represent the 
approximate localities where the types of these thirty species were discovered. Numbers 1, 9, 14, 16, 18, 19 and 30 are Stegolophodonts (see also Fig. 1228). 
The white crosses represent referred specimens. 

STEGODONTS [AND STEGOLOPHODONTS] IN ORDER OF THEIR DISCOVERY AND DESCRIPTION 

See Figure 691 



1. 


Burma 


1828 


2. 


Burma 


1828 


3. 


India 


18461 


4. 


India 


1846 


5. 


India 


1846 


6. 


Burma 


1846 


7. 


China 


1870 


8. 


China 


1870 


9. 


India 


1886 


10. 


Java 


1887 


11. 


Phihppines 


1890 


12. 


Java 


1890 


13. 


Java 


1908 


14. 


India 


1913 


15. 


Japan 


1915, 


16. 


Austria 


1917 


17. 


Japan 


1924 


18. 


ImUa 


1929 


19. 


India 


1929 


20. 


China 


1929 


21. 


Burma 


1929 


22. 


India 


1929 


23. 


Java 


1932 


24. 


Java 


1933 


25. 


Japan 


1934 


26. 


China 


1935 


27. 


China 


1935 


28. 


China 


1935 


29. 


Japan 


1935 


30. 


Borneo 


1936 



Original Name 
Mastodon laiidens Clift, Irrawaddy River, Burma 
Mastodon elephantoides Chft, Irrawaddy River, Burma 
1846 [1845] Elephas insignis Falconer and Cautley, Siwahk Hills, India 
1846 [1845] Elephas ganesa Falconer and Cautley, Siwalik Hills, India 
Elephas bombifrons Falconer and Cautley, Siwalik Hills, India 
Elephas cliflii P'alconer and Cautley, Irrawaddy River, Burma 
Stegodon sinensis Owen, vicinity of Shanghai, China 
Stegodon orientaJis Owen, Chungkingfoo, province of Szechuan, 

China 
Mastodon cautleyi Lydekker, Perim Island, India 
Stegodon trigotiocephalus Martin, ?vicinity of Surakarta, Java 
Stegodon mindanensis Naumann, Mindanao, Philippine Islands 

Stegodon Airdwana Martin, Alas-Tuwa, Ja\'a 
Stegodon Ganesa var. javanicus Dubois, Trinil, Java 



Mastodon stegodontoides Pilgrim, Lehri, Punjab, India 
1918, 1924 Elephas {Prostegodon, Parastegodon) aurorx ]\Iatsu- 

moto, Mt. Tomuro, Kaga, Japan 
Mastodon (Bunolophodon) longirostre Kaup forma sublatidens 

Schlesinger, Teschen (Schlesien), Austria 
Stegodon orientalis shodoensis Matsumoto, Island of Mitsugo 

(Mitsugo-shima) and Island of Shodo, Inland Sea; off 

Nagasaki, Eastern Sea, Japan 
Stegolophodon nathotensis Osborn, near Nathot, India 
Stcgolophodon cautleyi progressus Osborn, near Chinji Bungalow, 

India 
Stegodon orientalis grangcri Osborn, Yenchingkou, China 
Stegodon insignis birmanicus Osborn, Mingoon opposite Man- 

dalay, Burma 
Stegodon pinjorensis Osborn, near Siswan, India 
Stegodon bon'dolensis van der Maarel, Bondol, near Kuwung 
Stegodon trigonocephalus praecursor von Koenigswald, Bumiaju 
Parastegodoni /cwjantoerests Tokunaga, Kakio, Kanagawa Prefecture 
Stegodon yiishensis Young, YiJshe 
Stegodon oJficinaUs Hopwood, Szechuan(?) 
Stegodon z'danskyi Hopwood. Exact locality unknown 
Parastegodon sugiyamai Tokunaga, Iruhi in Saida Village, Shikoku 
Stegolophodon lydekkeri Osborn, near Bruui 



Specific Reference 

IN Present Memoir 
= Stegolophodon latidens 
= Stegodon elephantoides 
= Stegodon insignis-ganesa 
= Stegodon insignis-ganesa 
= Stegodon bombifrons 
= Stegodon elephantoides { = cliftii) 
= Stegodon sinensis 

= Stegodon orientalis 
= Stegolophodon ccmtleyi 
= Stegodon trigonocephalus 
= Stegodon {Archidiskodon?) 

mindanensis 
= Stegodon airawana 
= Stegodon airdwana [or S. 

trigonocephalus] 
= Stegolophodon stegodontoides 



= Stegodon aurorse 

= Stegolophodon sublatidens 



= Stegodon orientalis shodoensis 
= Stegolophodon nathotensis 

= Stegolophodon cautleyi progressus 
= Stegodon orientalis grangeri 

= Stegodon insignis birmanicus 

= Stegodon pinjorensis 

= Stegodon bondolensis 

[Not determined 

! by the 

[ present author 

= Stegodon officinalis 

= Stegodon zdanskyi 

Not determined by the present author 
= Stegolophodon lydekkeri 




\ 56 

——^ ^ i'alay 







Fio. 692. FossiL-BEAKiNO Horizons along the Irkawaddv River, Burma 

(l.<-ft) Map illustrating J. Crawfurd's journey to Ava and Martaban in the years 1826 and 1827. After William Clift, 1828, Pi. xi.iv. (UiKlil) M:i|> illus- 
trating the explorations of Barnum Brown for the American Museum of Natural History in the year 1923. Based on the ofh.ial Indimi Survey Cazetteer. 

3 -3. Ava, upper level.s of the Irravvaddy .Series (600 feet in thiekness) = Upper Pliocene, Upper .Siwaliks of India, containing SOijoilim binminicus Usl.orn 
and undetermmed Hos. (Lower Pleistocene (cf. Colbert, chap. XXII, pp. H.'.O, 14.51 of the present Memoir.-Kditor.] 

2. Pondaung Clays (.50 feet in thickness) = Upper Eocene, containing the suillinc Anthracotheriida-; also of the Order Peris.sodactyla: Fam. Titano- 
tl.enidic. Gen. Sivatilarwps, sp. ,S. cotlen, S. hirmanicum, S. rugosidrns, Gen. Eotitanotherium, sp. /?. lahirii; Fam. Amynodontiila-, Gen. Paramynodon, sp. 
/ . coitcn, P. birmaniciis; Fam. Tapiridic [ = I^phiodontidffl, Gen. Indolophus, sp. /. guplai, Gen. Chaamothmiim, sp. C. hirmanicum.. [For additional members 
of the Pondaung fauna, see Colbert, 1938.1, pp. 2.).)-.398. -Editor.) 

1. Yenangyaung (2.50 miles south of Ava), lowest levels of the Irrawaddy Series (estimated at 1.500 feet in thickruss), base of the Middle Siwaliks, eon- 
tanung Slrgohphodon Uitidem type, Skgodon deplmntoidcs type, S. cli/lii type [SIcgodon clephanloidcs (^cliflii)]. 

(Note by Edwin H. Colbert:— The total thi<knes.s of the Irrawaddy Scries is estimated at 5,000 foot, of which I lie upper levels onlv (about 000 fec-t), so 
far as known, are mammal bearing and are now reganled by recent geologists and paheonlologists as of T-ower I'leislocerLe age, Stamp (1022, pp. WT, IDS), 
for example, has shown that .Uaslndon [Slrgnhplwdo,,] Inlidnm and llippopolamvs irrmmlicu.'i are probably limited to the upper Irrawaddy l)cds (Lower Pleisto- 
cene). He considers that Pilgrim erred (1910, p. 190) in pl.icing the.se species in the lower Irrawad.ly fauna (cf, Colbert, Chap. XXII, pp. 1 1.50, 14.51, of the 
present Memoir). Colbert places all the probosci<Ieans discovered in Burma up to the present time, namely, SUgolophodon hl.idrns, Slajodoii dcplmnloidrs, S. 
cleplianloidrsi^cliflu). and N. birmmiicu.',. in the upper Irrawad.ly .Series, thus it-ssigning them to the I,ower Pleistocene. The lowest levels of the Series are 
considered of Middle to Ui.pcr Pliocene .age. Likewise the Pondaung clays are of a total thiekness of 6,.500 feet, the mammal-bearing portion apparently not 
exceeding oO feet.— Editor.] it j 

524 



THE STEGODONTIN^: HISTORY 825 

1. THE FIRST TWO STEGODONTS, DISCOVERED IN BURMA, 1828 

Geology, Irrawaddy River, Burma. — In the years 1826 and 1827, J. Crawfurd, F.R.S., while on an em- 
bassy to Ava, Burma, discovered an extensive deposit of organic remains in that unknown and distant region. 
On the Irrawaddy River, 250 miles below Ava, a gravel and sand deposit contained fossil bones, as mapped and 
described by Buckland (1828). This is the type region of Mastodon latidens Clift, 1828, and of Mastodon elephant- 
aides Clift, 1828, also of "Elephas diftti" Falconer and Cautley, 1846. According to Buckland (1828, p. 378) the 
exposure is an extensive one: 

These plants were found most abundantly in the same region with the fossil bones, but occur also along nearly the whole 
course of the Irawadi from Ava to Prome. They were principally collected from a tract of country (Footnote: 'See annexed 
map, Plate xliv.'] extending over a square of more than twenty miles on the east bank of the Irawadi, near the town of Wetma- 
sut, about half-way between Ava and Prome, between lat. 20° and 21° N. The occurrence of bones was most abundant in 
a small .space near the centre of this district, occupying about one third of the above-named area, the svuface of which is com- 
I)osed chietiy of barren sand hills mixed with gravel; beneath these are strata containing shells and lignite, thnnigh which they 
sink wells about two hundred feet to collect petroleum. 

This indicates that the cotype specimens came from lower and higher geologic levels, a fact not realized in 
Buckland's paper, nor in subsequent descriptions excepting those of Pilgrim.' Referring to Cliffs paper of 1828, 
Buckland mentions (p. 380) : 

. . . two new and strongly characterized species, one of which, from its approximation to the elephant in the structure of the 
teeth, Mr. Clift proposes to designate by the name of Mastodon elephantoides: to the other he has given the name of Mnstodon 
latidens. 

Lydekker (1886.2, p. 81) states that the type of E. cliftii, l.M\ "was obtained near Yenankhoung, on the left 
bank of the Irawadi in Upper Burma, by Crawfurd in 1826, and is preserved in the Museum of the Geological 
Society. "2 

D. N. Wadia (1919, p. 213) includes this Burmese [Yenangyaung] deposit in the "Irrawaddy system" 
[series], which combines marine and fluviatile strata. The upper part ( = 2,000 feet), composed of "sands and clays 
with abundance of fossil wood and mammals," is of fluviatile origin and corresponds to the Manchhars of Balu- 
chistan and to the Siwaliks of the sub-Himalayas. 

Barnum Brown visited this region in 1923 for the American Museum of Natural History and prepared a new 
map (Fig. 692, right) of this clas.sic collecting ground which should be compared with Crawfurd's original map 
(Fig. 692, left). The three geologic levels, discovered along the Irrawaddy River, Burma, are shown in these maps. 

HISTORY OF DISCOVERY OF BURMESE STEGODONTS 
Clift, 1828. — The first proboscidean described from Burma was Mastodon latidens Clift (1828, p. 371), 
the second species was Mastodon elephantoides Clift (1828, p. 372), both species founded on excellent cotypes 
with the excellent figures which are reproduced herewith (Figs. 693, 694, 695, 696). 

Mastodon latidens. — The cotypes of M. latidens (Figs. 693, 694) were found in 1826 on the Irrawaddy 
River, 250 miles below Ava, near Yenangyaung, Burma. Cliffs figured specimens include: Plate xxxvi, upper 
jaw (palate with anterior molar teeth, M^ much worn), which rightly (Osborn) belongs to Mastodon { = Stegodon] 
elephantoides; Plate xxxvii, upper molar teeth of the right side, M^, M^ (a younger animal), also Plate xxxviii, 
fig. 1, anterior part of the lower jaw; these certainly are the cotypes, because they conform with Clift's descrip- 
tion — all from the left bank of the Irrawaddy. 

'[See this Volume, Chapter XXII, p. 1450, by EJwin H. Colbert.— Editor.] 
2[Now in British Museum (Brit. Mus. M. 10520)]. 



FOUR CLASSIC STEGODONT AND STEGOLOPHODONT TYPES OF CLIFT AND FALCONER AND CAUTLEY 




Lectotype Palate of Stegolophodon latidens 
Fig. 693 (left). Lectotype palate (perspective of r.M') of 
Mastodon latidens Clift, 1828, PL xxxvii, fig. 1, with r.M^"' in 
situ, about one-third natural size. Inverted to show natural 
position of molars. From near Yenangyaung, Burma. 

Observe beginnings of binary fission of cones in tritoloph. 
.•\n accurate scale drawing of these teeth is shown in figure 71() 
from a cast (Amer. Mus. 21978). 



Fig. 093 



CoTTPE Third Inferior Mol.^r of Steoolophodon latidens ^ 

Fig. 694 (right). Cotype third right inferior molar, r.Ma, of Mastodnu 
latidens Clift, 1828, Pi. xxxviii, fig. 1, one-third natural size. Jaw omitted. 
From near Yenangyaung, Burma. 





Fig. 694 



CoTYPE OF Stegodon elephantoides Clift (=cliftii Falconer) 
Fig. 695 (left). Type first left superior molar, l.M', of Elephas cliftii Falconer and Cautley, 
1846, after photograph of cast Amer. Mus. Warren Coll. 10382, one-half natural size. Inverted 
to sliow the molar in natural position. From near Yenangyaung, Burma. See also figure 
683. Compare Chft, 1828, PI. xxxix, fig. 6. Original in Briti.sh Museum (Natural History) 
M. 10520. 



Fig. 695 



Lectotype Lower Jaw of Stegodon 
elephantoides 
Fig. 690 (right). Lectotype second 
and tliird left inferior molars, I.M3-2, of 
Mastodon lU pliantoides Clift, 1828. 
PI. xxxviii, fig. 2, one-third natural 
size. From M<';ir '^'enangyaurig, Hurma, 




n.cUftU esju^iA,} 



J^nidtj: fy iT^WirwwW 



Fig. 696 
826 



THE STEGODONTINiE: HISTORY 827 

Mastodon latidens (op. cit., i>. 370).— On comparing the teeth of our Mastodon latidens witli those of the Mastodon of 
the < »hi() (M. Riganteum), we shall hnd the elevate<l points or ridges in the tooth of the former more numerous, le.ss distant, and 
the interstices less deeji tlian in those of the latter; in short, we shall obser\e that the teeth begin to assume the apiiearance of 
those of the elephant. ( )n advancing to Mastodon elephantoides, we shall find all these features of similarity more strongly 
developed; — the points and ridges are still more numerous, and the structure, wei'e it not for the absence of crusta petrosa, 
becomes almost that of the tooth of the elephant. 

Mastodon elephantoides.— This species (Fig. 696) was found in the same locality as the cotypes of M. 
latidem. Cliffs figiu-ed specimens include: Plate xxxviii, fig. 2 (lower jaw with M,, M;i in situ); Plate xxxvi 
(palate figured by CUift as M. latidens), and Plate xxxix, fig. 6 (a first superior molar, l.M') afterward made the 
type of Elephas cliftn by Falconer and Cautley, referred in the present Memoir to Stegodon elephantoides {=cliftii). 

Mastodon elephantoides (op. cit., p. 372).— The tooth [Fig. 690], which is eleven inches long and three inches and a half 
broad, has no less than ten denticules [i.e., ridges], and each of the.se denticules is mammillated with small points; five being 
the smallest number, and eight the greatest on any one <lenticide. In front of this beautiful tooth we have a remnant of the 
preceding one, . . . 

Osborn, 1924: Pending final revision, which can only be made by a ree.xamination of the specimens and the 
localities from which they came, we may designate these Stegodonts from Burma as follows: 

Mastodon [ = Stegoloi)liu(hm] latidens ( 'lift, 1828, founded upon a palate from near Yenangyaung, Burma, with five and a half 

ridge-crests in the third superior molar, four and a half to five in the second superior 
molar, and with four to five mamillse (or conelets) on each crest. See ( 'lift, 1828, 
PI. XXXVII, fig. 1 (Fig. 693 of the present Memoir). 

Also a lower jaw (.see Clift, PI. xxx\iii, fig. 1), r.Ms with seven ridge-crests 
(Fig. 694 of the present Memoir). 

Madodon [=,Slegodo)i] dephanloides Clift, 1828, founded upon a third inferior molar of the left side, I.M3, from near 

Yenangyaung, Burma, with ten ridge-crests, five to eight conelets on each. See 
('lift, 1828, PI. XXXVIII, fig. 2 CFig. 696 of the present Memoir). 

Also a palate showing r.'SV- and l.M'-, with six and a quarter ridge-crests (see 
Clift. 1828, PI. xxxvi), erroneously marked "Upper Jaw of Mastodon latidens." 
(Not figured in the present Memoir.) 

Stegodon elephnutoides {=diftii) Also a left first superior molar, l.M', with six and a quarter ridge-crests, from near 

Yenangyaung, Burma. See Clift, 1828, PI. xxxix, fig. 6, marked "Upper molar of 
M. Elephantoides." Afterward made the type of Elephas cliftii by Falconer and 
Cautley. (Figs. 683, 695, of the present Memoir.) 

Original Cotypes of Mastodon latidens Clift (Figs. 693, 694) 
First Species. — Cliffs designation (1828) of Mastodon latidens as including a lower jaw (PI. xxxviii, fig. 
1 — Fig. 694 of the present Memoir) and a palate containing M^, M^ (PL xxxvii, fig. 1 — Fig. 693 of the present 
Memoir) establishes these specimens as the cotypes. The same superior teeth, r.M", r.M^, were sectioned and 
figured by Falconer and Cautley (1846, p. 48 [1845, PI. iii, fig. 8]) : 

The last tooth shows five principal ridges with a posterior talon ridge and a subordinate ridge in front. The ridges are 
transverse, and divided by a longitudinal cleft into two pairs of princijjal points without intermediate mammillae in the hollows. 
The enamel is very thick, and the cement is reduced to a thin layer which is only observable in the bottom of the hollows. . . . 
The anterior tooth [M'^] had been a long time in use, and the ridges are nearly all worn out. They were four in number, in this 
as well as in the two teeth which preceded it in the jaw. We believe this to be a small or dwarf variety of M. latidens. . . . 

(Clift, 1828, p. 371) : Dentition. — Each tooth of the lower jaw [Fig, 694] consists of seven denticules, which are elevated, 
rounded, and mammillated: the mammillae being from three to four in number. The dentition both in this s])ecies and in M. 
elephantoides, \'ery much resembles that of the ele])hant. We have the molar tooth gradually protruded forward, and rising 
as the fangs are added, according to the demand made by the abrasion of the exposed crown, and the consequent absorption 
of the anterior fang; the posterior part of the tooth not having cut the gum, while the anterior portion is completely worn away. 
The relics of the preceding tooth, the place of which the tooth in use was progressively supplying, are plainly to be seen [Foot- 
note: 'See Plate xxxvii. fig. 1. Plate xxxviii. fig. 2.']. 



828 OSBORN: THE PROBOSCIDEA 

According to the above, the ridge formula is as follows: 

Ridge formula of Mastodon [ = Stegolophodmi] latidens: Dp 4| M ly M 2t M 3-,^. 
Osborn, 1927: The above is the main ridge formula given by Falconer, omitting the anterior and posterior 
talon ridges (compare tables above, Section I, No. 3, and below Section IV of this chapter, also full description). 

Original Cotypes of Mastodon elephantoides Clift (Fig. 696) 

Second Species. — The second species is Mastodon elephantoides Clift. 

Clift, 1828, p. 372. — Mastodon elephantoides. — M. dentibus molaribus latis, denticulis numerosis, compressis. This 
species must have been smaller than the last; and though we have one fine example of the lower jaw, showing the tooth in the 
highest degree of perfection, that is the onlj^ portion of the animal from which we can safely draw any inference as to its struc- 
ture and habits. The tooth, which is eleven inches long and three inches and a half broad, has no less than ten denticules 
[i. e., ridge-crests], and each of these denticules is mammillated with small points [conelets]; five being the smallest number, 
and eight the greatest on any one denticule. . . . The denticules of the tooth are much more compressed than those in the 
species last described; they are closer together [Footnote: 'Eight denticules of M. elephantoides occupy the same space as five 
denticules of M. latidens.'J, and the enamel appears to be not so thick. They form a series of plates mucronated with small 
points. There is no apparent commissure, neither is there any central depression : on the contrary, the plates rather rise in the 
middle. 

Clift uses the word denticules in the sense of ridges, or crests, or lophs; subsequently Falconer uses the word 
denticles in the sense of "mammillae" or conelets. Osborn introduces the word conelets, because these small, 
roimded "mammillae" appear on the summits of the primary cones, as e.xplained above (p. 812). 

(Clift, 1828, "Explanation of Plates," PI. xxxix, fig. 6): "Upper molar tooth of Mastodon elephantoides." 
[No mention of this tooth is found in the text.] As reproduced herewith (Fig. 695) this molar corresponds closely 
in scale and structure with Cliffs type of M. elephantoides (Fig. 696). 

Osborn, 1927 : From these two teeth, figured and described together by Clift, 1828, not improbably represent- 
ing the same species, also from the palate with r.M^ and l.M^ in situ (Chft's figure, PL xxxvi), the following ridge 
formula may be written: 

Ridge formula of Mastodon [ = Stegodon] elephantoides: M 1^ M2^ M3ro. 

The above ridge formula is similar to that of Stegodon bombifrons and much more primitive than that of S. 
insignis-ganesa. Consequently Mastodon elephantoides Clift, 1828, of Lower Pliocene age,' based upon two figured 
specimens (Figs. 695, 696), appears to be well established as the second species of Stegodon described from Burma 
in 1828. The name elephantoides was dropped, however, by Falconer and Cautley in 1846 and a new name, Elephas 
cliftii, was apphed to the second specimen (Fig. 695 of the present Memoir) figured by Clift, PI. xxxix, fig. 6. 

Species M. elephantoides Dropped by Falconer and Lydekker. — In 1846, Falconer and Cautley 
erroneously alleged that Clift had confused the remains of the two species Mastodon latidens and M. elephantoides 
under the name M. elephantoides; they accordingly {op. cit., p. 47) dropped the name elephantoides and proposed 
a new name for the species of "transitional Mastodons," which had been partly confused with Mastodon latidens. 
Thus to the Stegodonts with six ridges on the intermediate molars they gave the name of Elephas cliftii (Figs. 
683, 695), and to the Stegodonts with a greater number of ridges the name of E. insignis (Fig. 697). In a sub- 
sequent paper (Falconer, 1857, p. 314), these and other species appeared under the subgeneric name of Stegodon 
Falconer. Lydekker (Palaeontologia Indica, 1880, pp. 256, 257) also set aside the prior specific name elephantoides 
Clift, 1828, and made it a synonym of Stegodon cliftii Falc. and Caut., 1846. In subsequent literature (e. g.. 
Pilgrim) the Falconer-Lydekker usage is followed, i.e., M. elephantoides is dropped. 

M. elephantoides Revived. — In the present Memoir the specific name elephantoides is revived by Osborn 
for reasons given in the .systematic revisions above and below. 

'[See note on page 824 above regarding the Lower Pleistocene age of Siegudon elephantoides. — Editor.] 



THE STEGODONTIN.E: HISTORY 



829 



2. DISCOVERIES IN INDIA AND BURMA (1845, 1846) 
Third Species. — The third species of Stegodont {Elephas insignis) was described (Falconer and Cautley, 
1846, p. 37) as follows (see Fig. 697) : 

. . . the four anterior ridges being affected by wear, and the six posterior ridges entire, . . . The white mass in the centre [dentine] 
represents the body of ivory, which is projected upwards in ten angular lobes terminating in a sharp edge. . . . The interspaces 
of the five posterior ridges of enamel are completely filled up by a mass of cement, or 'cortical,' much exceeding the enamel in 
thickness; and in cjuantity iu nearly as great an amount of development as the ivory core of the ridge. 



^v fi 



Kj <■•■■ 





Fig. 697. Lectotype and Cotype of Steqodon insignis 
(Left) Lectotype superior molar, l.M^ of Elephas insignis Falconer and Cautley, 1846 [1845, PI. ii, fig. ffa], one-third natural 
size. Probably Pinjor horizon, Upper Pliocene [Lower Pleistocene], Upper Siwaliks, India. Inverted to show natural position 
of molar. 

(Right) Elephas [ =Siegodon] insignis cotype. Anterior portion of a thii-d inferior molar, M3. After Falconer and Cautley, 
op. cit, PI. II, fig. 6h. 



Falconer distinguished this species as possessing a ridge formula of M 3 --, with valleys between acute ridges 
deeply filled with cement, as compared with Stegodon elephanioides which has ten less acute ridges without cement 
(Fig. 696). Lectotype ridge formula (Fig. 697) oi Elephas [ = Stegodon] insignis: M 3 — . 

Fourth Species.— The fourth species of Stegodont (described by Falconer and Cautley, 1846, p. 45) was 
named Elephas ganesa: 

Fig. 7a, pi. 3., represents a section of the last upper molar of an undescribed Indian fossil species, named E. Ganesa, in this 
work. The crown consists of ten principal ridge.s, with a subordinate 'talon' ridge in front and behind [i.e., M 3 ^^"^°"''' ]. The 
anterior seven ridges have their summits worn, the two in front being ground down to the common base of ivory, the tooth 
having been a considerable time in use. A small portion is broken off at the anterior end. The disposition and relative propor- 
tions of the ivory, enamel, and cement, bear the closest resemblance to 
those of the corresponding tooth of E. insignis (pi. 2, fig. 6a), and the 
nimiber of ridges agrees. The section presents the same chevron-formed 
character in the ridges, but the interspaces are narrower, the cement is 
in less quantity, and the layer of enamel is thicker. 



Lydekker observes (Lydekker, 1880.1, p. 268): . . . "these 
latter teeth, however, cannot be distinguished from those named 
Stegodon insignis, and as we shall see subsequently, it is only the 
adult skulls of these two very closely allied species that can be 
distinguished." This has led to the opinion that S. insignis may 
represent a female and S. ganesa a male of the same species. 




Lectotype of Stegodon ganesa 
Fig. 698. Lectotype M' of Elephas Ganesa Falconer and 
Cautley, 1846, [1845, Pl. in, fig. 7a], one-third natural size. 
Probably Pinjor horizon, Upper Pliocene [Lower Pleistocene]. 
Upper Siwaliks, India. 



830 



OSBORN: THE PROBOSCIDEA 



The ridge formula of the lectotyi)e third superior molar of Stegodun ganesa (Fig. 698) appears to he practi- 
cally tli(» same as that of the lectotyjie of ;S'. insignis (Fig. 697). Other more progressive ridges foruuihr are 
shown in the comparative ridge formulae tables above and below in this Memoir. 

Ridge formula (Fig. 698) of Elephas [ = Stegodon] ganesa: M 3 '^'=^^^. 

Fifth Species. — Falconer described (Falconer and (Jautley, 1846, p. 46) Elephas bombifrons, the fifth species 

of Stegodont, as follows: 

This species, of the distinctness of which we are assured, by possessing several crania containing perfect teeth, belongs to the 
same group [Stegodon] as the two species last described. The crown is divitled into similar transverse ridges, composed of numer- 
ous mammillse, which yield a corresponding chevron-shaped section, and the interspaces are occupied by a thick coat of cement; 
but they differ, in being broader and less elevated, with more open hollows. The principal ridges of the last molar [IVP] do not 
exceed eight in the upper jaw, and nine in the lower [M3]; while in E. insigniti they amount to ten in the former [AP], and 
reach as many as thirteen in the latter [M3]. The last tooth of the upper jaw measures eleven inches in length, by four and 
a half in width. 

Falconer and Cautley (1846 [1847, PL xxviii]) assign to M'* nine ridges and a heel. 

Lectotype ridge formula of Stegodon bombifrons: M 3 f . 

Sixth Species. — A sixth species of Stegodont (Figs. 683 and 695) was erroneously proposed by Falconer and 
Cautley in 1846 (1846, p. 47) under the name of 'Elephas clij'tii.' They selected the type of this species as 
follows : 




CoTYrE OK Stuqodon bomhikhons 
Fig. 699. Cotypc skull of Elephas bombifrons 
FiilconiT and Cautley, 1846 [1847, Pis. xxvii, 
xxviiil, ont'-sixtli natural size. Brit. Mus. 
M.2979; cast Amcr. Mus. Warren Coll. 101578. 
From the Siwalik Hills, India, probably the Dhok 
Pathan horizon, fpper Siwaliks, Middle Pliocene. 
As a view of the third molars was not given in the 
loctotyiJC of Lydekker (Falconer and Cautley, 
op. cit., PI. XXVI, Brit. Mus. M.2978), the cotypc 
skull is figured here. 



THESTEGODONTINiE: HISTORY 



831 



In our view, the tooth represented in pi. 39, fig. 6, of Mr. CUft's memoir in tlie Geological Transactions [Clift, 1828], 
under the name of Mastodon Elephantoides, and the palate specimen represented in pi. 36 of the same memoir, under the 
name of M. latidens, belong to this species. 

This was an unfortunate error. As explained above, the tooth in PI. 39, fig. 6, of CHft's Memoir, measuring 
155 mm. in length, 83 mm. in breadth, has a ridge formula of M 1 — ; the palate specimen (PI. 36), erroneously 
identified by Clift as "Mastodon latidens,'' also has a ridge formula of (?)M 2—. Consequently they belong to 
the same species, namely, Stegodon elephantoides, but to preserve the name cliftii, which runs all through the 
previous literature, they are designated in the present Memoir as Stegodon elephantoides { = cliftii). 





Stegodon elephantoides (= cliftii) 



Fig. 700. Original type figure of Elephas cliftii Falconer 
and Cautley, 1846, a first superior molar of tlie left side, l.M', 
one-half natural size. From near Yenangyaung, Burm;i. 
After Clift, 1828, PI. xxxix, fig. G, figured as an "Upper 
molar of Af. EUphntdoides." Original in the British Museum 
(Brit. Mus. M. 10520); cast Amcr. Mus. Warren Coll. 10382. 
Inverted to show natural position. 



Fig. 701. New figure of type of Elephas cliflii Falconer and 
Cautley, 1846, a first superior molar of the left side, l.M', de- 
scribed and figured as M. Elephantoides by Clift, 1828, PI. xxxix, 
fig. 6. Subsequently seli^cted by Falconer and Cautley as the ty]>e 
of Elephas cliftii. After i^hotograph of type cast (Amer. Mus. 
Warren Coll. 10382). One-half natural size. Observe that the 
molar is here placed in its natural position. Original in the 
British Museum (Brit. Mus. M. 10520). 



3. THE STEGODONTS OF CHINA, INDIA, JAVA, THE PHILIPPINE ISLANDS, AUSTRIA, 

JAPAN, AND BURMA 

Seventh Species. — The seventh species of Stegodont described (Fig. 702) was the Stegodon sinensis of 
Richard Owen (Owen, 1870, p. 417), alleged to be (p. 421) "from marly beds in the vicinity of Shanghai," China, 
which has been mistakenly regarded by some authors as close to, or as a synonym of, Stegodon cliftii (Owen, op. 





Fig. 702. Type r.Dp^ of Stegodon sinensis Owen, 1870, PI. xxvii, figs. 1, 2, natural size. From "marly beds 
in the vicinity of Shanghai," China, of Upper Miocene [Lower Pliocene] age. 



832 



OSBORN: THE PROBOSCIDEA 



cit., p. 418; Brauns, 1883, p. 44). This synonymy is not borne out by careful comparison with the type figures, 
because S. sinensis is nuich more primitive than ti. cliftii. 

Eighth Species. — In the same paper Owen {op. cit., 1870, p. 421) describes Stegodon orientalis, the eighth 
species of Stegodont (Fig. 703), "from a cave, near the city of Chung-king-foo, in the province of Sze-chuen," 
based upon molar fragments which he rightly observes "more resemble the teeth of Stegodon Cliftii, St. insignis, 
and St. ganesa of Falconer than does the St. sinensis; and in the apparent quantity of coronal cement . . . as 
well as in the evidence of a hinder talon . . . they are more like *S^. insignis than St. Cliftii." 






Fig. 703. Type of Stegodon orienlalis Owen, 1870, PI. xxviii, figs. 1-4. "Portion of true molar" (figs. 1, 2) .anil "hind enj 
of milk-molar" (figs. 3, 4). From near Chungkingfoo, China, probably of Lower Pleistocene age. 



Fie, 18. 



Ninth Species.— il/astodow cautleyi Lydekker, 1886 (1886.1), was founded on five teeth of th(> upper jaw, of 
whicli wo s(>]oct as the type (Fig. 704) the third left upper true molar, from Perim Island, India, a tooth first figured 

by Falconer and Cautley and referred to the species 
Mastodon latidens Clift (PI. xxxi, figs. 6, 6a — see 
Vol. I, fig. 142 of present Memoir, the caption of 
which is erroneous— corrected in present c^hapter, 
p. 821). The four cotypes, obviously smaller and 
inucli more primitive teeth than the type of Mas- 
todon latidens, are also from Perim Island and are 
recorded, like the lectotype, from an older geologic; 
horizon, namely. Upper Miocene [now (1935-1938) 
regarded as Middle Pliocene, Dhok Pathan]. Ac- 
cording to these specimens, the ridge formula of 
Mastodon [ = Stegolophodon] cautleyi is about the 
same as that of Mastodon [ = Stegolophodon] latidens. 




Mastodon cautleyi. — The third left upper true molar, io an unworn condition ; 
from the Siwaliks of Perim Island. J. The lower border of the figure is 
the inner border of the specimen. 

Lectotype of Stegolophodon cadtleyi 
Fig. 704. Third left superior molar of Mastodon cautleyi Lydckkor, 18S(i, 
selected as the type (sec Pilgrim, 1913, [). 294). Roprod\iwd after Lydekker, ISSO 
(1886.2, p. 73, fig. 18), one-half natural .size. 



Mastodon [ = Stegolophodon] cautleyi: M 2^ M 3^^^. 

Tenth Species. — In 1887 Martin described 
("Fossile Siiugethier-reste von Java und Jajjan") 



THE STEGODONTIN^: HISTORY 



833 



from Java a species to whicli he gave the name Stegodon trigonocephalus (Fig. 705), in reference to the triangular 
shape of the head. The geologic level is regarded by Matsumoto as equivalent to the Lower Pleistocene, Boulder 
Conglomerate beds of India. It may be of the same geologic age as the type of Pithecanthropus erectus. The 
juvenile type does not admit of giving the mature ridge formula; the ridges are closely compressed, buried in 
cement, and each ridge is surmounted with ten to twelve conelets, the exact number being indeterminable from 
the figures. 






Fig. 705. Type of Stegodon trigonocephalus Martin, 1887, immature skull, one-eighth natural size. I'Vom vicinity of Surakarta, 
Java, probably of Lower Pleistocene age. (Right figures) Tab. ii, figs. 1, la; (left figure) Tab. iii, fig. 1. 



Eleventh Species. — The eleventh species (Fig. 706) was described by Naumann in 1890 from Mindanao, 
Philippine Islands, as Stegodon mindanensis. It was first referred to Stegodon trigonocephalus by Naumann in 
1887, but subsequently was made the type of a distinct species. The compressed cement-covered ridges with 
multiple conelets indicate a Lower to Middle Pleistocene stage of evolution, similar to that of S. trigonocephalus. 

Twelfth Species. — A twelfth species, Stegodon airdxvana (Fig. 707), was described by Martin in 1890 from 
Alas-Tuwa, Java, based upon a type jaw containing the right and left third inferior molar teeth, M3. In its 
high ridge formula the author compared it to Stegodon insignis and S. ganesa. In a later paper by Janensch 
(1911, 1). 187), he distinguished S. airdwana from both S. insignis and S. ganesa and wrote the ridge formula 
as follows, according to the present writer's understanding of the Janensch system [cf. Table V] : 




Ridge formula of Stegodon airdwana: Dp 3 '-^ Dp 4 *'-^ M 1 '-^^^ 



M2 



li-0-!.i 



M3 



14-1 1-1 2-1,4 
15-1 3-(i • 



l. %. 

Type or Stegodon (Archidiskodon?) mindanensi.s 
Fig. 706. Type of Strgodon mindanensis 
Naumann, 1890, molar tooth. Mindanao, Philip- 
pine Islands, Lower to Middle Pleistocene age. 
.\rtor Naumann, 1887, Taf. i, figs. 1 and 2, under the 
name Stegodon trigonocephalus. 



This species is important because it is highly characteristic of the 
Trinil Pithecanthropus erectus beds. 

Thirteenth Species. — In 1908 Dubois described from the Trinil 
Kendeng-Schichten, Java, Stegodon ganesa javanicus, as a variety of S. 
ganesa [which proves to be a synonym of either S. airdwana or S. tri- 
gonocephalus (see p. 889 below)]. 



834 



OSBORN: THE PROBOSCIDEA 



Fourteenth Species.- — Very important is the Mastodon stegodontoides of Pilgrim, 1913, from Lehri, Punjab, 

India (Fig. 708), wliich Pilgrim described as follows (p. 294): 

The last species of Mastodon whicli can be referred to this line [the M. cautleyi-M. lalidens Hne] is the tooth from Ix-hri, 
of which the horizon is uncertain but ma,y be possibly Upper Siwahk, figured by Lj^dekker in Pal. Ind. ser. 10, Vol. 1, Plate 39, 
as .1/. lalidens but which is better recognized as a new species, owing to the ahnost entire absence of accessory columns, for 
which I propose the name Mastodon stegodontoides. So close is this to Stegodon difti that it is hard to separate the two genera. 
It will be seen that Mastodon stegodontoides carries on none of its ridges more than the usual four columns while anterior ridges 
of Stegodon difti carry nine or ten mammillae. 

Fifteenth Species. — Very progressive is the 

Elephas {Prostegodon, Parastegodon) mirorx of Mat- 

sumoto, 1915-1924, from Mt. Tomuro, Kaga, Japan 

(Fig. 709), which Matsimioto first regarded as a stage 







Type of Stegolophodon steqodontoides 
I'^if!. 70S. Type r.M' of Maslodon slegodontoidcs Pilgrim, 1913. .\ftcr Lydek- 
k(_T, ISSO, PL xxxix: "Maslodon (Tetralophodon) lalidens, Clift. The third right 
upper true molar: from Lehri, in the Punjab. The specimen is drawn of the 
natural size, and is viewed from the inner [outer] side." Ind. Mus. A.86. Reduced 
to one-half natural size. Provisionally placed in the Upper Pliocene, Pinjor 
formation (.see Fig. 413, also PI. xiii). 



^i^V^ffft 





l''ig. 707. Type of SIfyodon Airawiiiid .Martin, 1890, Tab. i, figs. 
Kcndcng-Schichtcn horizon (Pilhccanlhropus ereclus zone), Middle 
1, 2, lower jaw, one-fourth natural size. From .\la.s-Tuwa, Triiiil, .lava. 
Pleistocene. 




Type of Stegodon auror.*; 

Fig. 709. Type r.M^ of Elephas (Proslegodon, Parastcyodon) aurora- 
Matsumoto, 1918, PI. xx, figs. 1 and 3, one-half natural size. 



THESTEGODONTIN^: HISTORY 



835 



in Elephas close to the Upper Pliocene Elephas [ = Archidiskodon\ planifrons of India, but which he subsequently 
made the genotype of a new genus Paraslegodon. With our fuller knowledge of Stegodon airdwana, the present 
species may be placed in the true Stegodon phylum, distinguished by cranial characters from species of the Ele- 
Ijhantidse, as Stegodon aurorx. 

Sixteenth Species.— The subspecies Mastodon {Bunolophodon) longirostre Kaup forma sublatidem Schle- 
singer, 1917, from near Teschen (Schlesien), Austria (Fig. 710), appears to be very close to the Mastodon [ = Stego- 
lophodon] stegodontoides of Pilgrim ; consequently it is removed to this genus, namely, Stegolophodon snblatidens. 

Seventeenth Species. — The subspecies Stegodon orientalis sliodoensis 
of Matsumoto, 1924, named from the Island of Shodo, Inland Sea, is 
regarded by its author as a descendant of S. orientalis Owen, of China, 
contemporaneous with the referred S. insignis of the Middle [Upper] 
Pleistocene, Narbada of India. Not figured in present Memoir. 

Eighteenth Species. — The type of Stegolophodon nathotensis Osborn, 
1929, consists of fragmentary molars (Fig. 724) collected by Barnum 
Brown in 1922 in the Lower Chinji horizon, near Nathot, India. 




Type of Stegolophodon sublatidens 
Fig. 710. Tyjjc of Mastodon {Bunolophodon) 
longirostre Kaup forma sublatidens Schk'singcr, 
1917, one-half natural size. Compare full legend 
below of figure 722. 



Nineteenth Species.— Somewhat more advanced than Stegolophodon 
cautleyi of the Upper Miocene [Middle Phocene] is the Stegolophodon caut- 
leyi progressus Osborn, 1929 [of the Mio-Pliocene], the type of which (Amer. 
Mus. 19446) is a complete cranium (Figs. 725, 727) collected at the summit of the Lower Chinji horizon, 2,000 
feet above the base of the Lower Siwaliks, India. This is a young individual in which the molar ridge-crests 
are intermediate in formula and pattern between the 'Mastodon' cautleyi of Lydekker and the 'A/.' latidens of 
Clift. 

Twentieth Species. — Somewhat more primitive and an- 
cient than the Upper Pliocene [Lower Pleistocene] Stegodon in- 
signis Falconer type is the subspecific stage Stegodon orientalis 
grangeri Osborn, 1929 (Fig. 762) of the Yangtze River region 
near Wanhsien, Sezchuan, 140 miles northeast of the type 
locality of S. orientalis Owen. This subspecies is now very 
fully knowai and is amply illustrated and defined in the present 
Memoir. See type figures on pages 876, 877, 879, and 881 below. 

Twenty-first Species. — Stegodon insignis birmanicus Os- 
born, 1929 (Fig. 758), from the Up])cr Pliocene of Burma. See 
tjqje figure on page 875. 

Twenty-second Species. — Stegodon pinjurensis Osborn, 
1929, from the Lower Pleistocene of India, upper levels of the 
Pin j or horizon (Figs. 711, 765, 767). 

|The following species have been descril)ed since the author's Fig. 711. Photographic reproau.ti.m ..f type palate of 

intensive review of the true Stegodonts, consequently the de- stegodon idnjormsis Oshom, \<m (Amer. Mus. i'j772), .•olle<-te,l 

by Barnuni Brown three miles north of Siswan, India. Alx^iil 

terminations of Professor Osborn cannot be given and the spe- one-eighth natural size. See figure 765 below. 




o 

1 

o 






Stegolophodon lydek- 
keri type 










< 
>-> 




Stegodon airdwana 
type 

Stegodon ganesa var. 
javantcus type 
[ = jS. airdwana or 
-S. trigonocephalus] 


t 

o 

fa 

IS 








PHILIPPINE 
ISLANDS 




11 
Iff 










< 

Oh 
< 




(Si 


1 

s 
e 

t/3 


1 

s 


1 

O K 

IS 

55 




< 
1— ( 

o 






09 00 

S g g 

-S <u-2 g 




00 

g 

g 

ii 




< 






•2-aj 
fi a 
•2= >> 

g 2 




BURMA 

Stegolophodon lati- 
dens lectotype, 
cotype, and ref.' 

Stegodon elephan- 
toides lectotype' 

Stegodon elephan- 
toides ( = cliflii) 
cotype' 




Q 


■is 
!§> 

s s 
o o 

^"§ 

coco 




Stegodon pinjorensis 
type 

Stegodon insignis lec- 
totype and ref. 

Stegodon ganesa lec- 
totype and ref. 

Stegolophodon stego- 
dontoides type 

Stegodon bombifrons 
ref. 


Stegodon bombifrons 
lectotype and ref. 

Stegolophodon caut- 
leyi lectotype 

Stegolophodon lati- 
dens ref. 




«g 

--» CO r* 

o 2 S S 
-g p-g -*^ 

<s S, o g 


> 

.J 
ca 
< 


z 

o 


w 
o 

w o 

J H 
O to 

1 ^^ 


g 9U900^SI9IJ U 

a g jaMOT ^ j^ 

"2 0^ 9U90 M g 

o3 -oiidJadda &2 


w 

3 ^ 
o 2 




is 2 

O J 


w 
o 
o 
!j 
P- 
o 



42 

a 

03 



O -T 



hr 


-4J 




d 


T3 




& 


^ 




o 


l_ 




9 


W) 






c3 




a 






43 


£ 




T3 


0) 


M 


4) 


o 


oo 



a E 



£ 


& 


133 


o 




h-l 


OS 




(M 




05 




T— f 


OS 


<u 


-n 


o 


0) 


a 


T3 






T3 


a 


.s 


fe 


•c 


o 


o 




CQ 


QJ 


(U 


t^ 


■a 


rt 


a 




03 


o 


a 


CJ 


oS 


a 


1-5 








'V 


OJ 


a 


_c 


OS 


-4-9 




0) 


oj 


m 


c 


-C 


-fi 




C) 


i: 






03 

> 


J3 


o3 


<D 


"-5 


O 


s 


CtJ 


X 


-t* 


<s 






t)0 


^ 




a ' .' 


a 


a> u 


u. 


ao 


c 


m "ti 


o 


fl'T3 


<u 


y W 


o 


> 1 


a 


o 




^ 


« 


-4.J 




3 




93 





THE STEGODONTINiE: HISTORY 837 

cies are listed here according to the nomenclature of the various authors, without comment, with the exception 
of Parastegodoni kwantoensis and P. sugiyamai. 

Twenty-third Species. — The type of Stegodon bondolensis van der Maarel, 1932, is a fragment of a mandible 
with what is inferred to be the third molar of each side, found at Bondol, Java (Fig. 782 of the present Memoir). 

Twenty-fourth Species. — In 1933 von Koenigswald described from Bumiaju, Java, a lower jaw with 
third molar of both sides complete, which he named Stegodon trigonocephalus praecursor (Fig. 783 of the present 
Memoir) . 

Twenty-fifth Species. — The species Parastegodon? kwantoensis Tokunaga, 1934, PI. ix, a portion of a jaw 
with second right molar in situ, is indeterminate, owing to the fact that Professor Osborn regarded Parastegodon 
as either a progressive Stegodon or a primitive Archidiskodon. vSee type figure 784 on page 897 below. 

Twenty-sixth Species. — Slegodmi T/iishensis Young, 1935, from Yiishe, China. Type, a well preserved 
upper left third molar, PL v, fig. 1 (Fig. 785 of the present Memoir). 

Twenty-seventh Species.— The type of Stegodon officinalis Hopwood, 1935, is a fragment of an unworn 
lower molar, said to have come from Szechuan, China. This is figured in Hopwood, 1935, PI. vii, fig. 3 (Fig. 
786 of the present Memoir). 

Twenty-eighth Species. — The type of Stegodon zdanskiji Hopwood, 1935, is a fragment of a right third 
lower molar, consisting of the first four ridges, and figured in Hopwood, PI. vii, fig. 5 (Fig. 788 of the present 
Memoir). 

Twenty-ninth Species. — The type of Parastegodon sugiyamai Tokunaga, 1935 [Stegodon? sugiyamai] is 
either a first or a second molar, probably of the upper left side, found in Shikoku, Japan (Fig. 789 of the present 
Memoir) . — Editor.] 

Thirtieth Species. — In Volume I of the present Memoir (p. 700, fig. 660) will be found the type descrip- 
tion of Stegolophodon lydekkeri, dedicated by the present author to his friend Richard Lydekker. The type is 
a third left superior molar from Borneo, "much more progressive than the S. latidens type." 



III. SYSTEMATIC ARRANGEMENT OF THE STEGOLOPHODONTS AND STEGODONTS 

IN PHYLOGENETIC ORDER 

. 1. CHARACTERS OF THE SUBFAMILY STEGODONTIN^^ 

Subfamily Characters. — (1) Habits chiefly browsing; tropical forest living; crushing of coarse 
leafage, herbage, and wood fiber; not progressing to the grazing type. (2) Cranium relatively abbre- 
viated, mesocephalic, bathycephalic; grinding-tooth plane deeply depressed to occipital condyles; 
occipitofrontal plane neither elevated nor expanded, non-acrocephalic, non-hypsicephalic. (3) Tusks 
straight or slightly curved, horizontal or subhorizontal in direction, continuously serving in browsing 



habits. (4) Grinding teeth brachyodont to subhypsodont, ridge-plates of M 3 increasing from 
[Stegolophodon] to "I'^J;' [Stegodon] 



5W-6H 
'/4-6-W 



'(In Volume I (1936) of the present Memoir, the aiitlior si'i)aratecl the Stcgodimtoidea (true St(>gocloiits) from the Elepliaiitoidea (pp. 22, 25), also the 
genus Stegodon from Stegolophodon, placing all the Stcgolophodonts in the superfamily Mastodontoidea, family Mastodontidae, new subfamily Stcgolophodon- 
tinac (p. 700) and the true Stegodonts in the superfamily Stegodontoidea, family Stegodontidse, subfamily Stegodontinae (see pp. 806-808 above). — Editor.] 



838 



OSBORN: THE PROBOSCIDEA 



Tliese cliariictors are observed in the species and subspecies discovered since 1828, many of which were origi- 
nally referred to the genus Mastodon and the genus Elephas and are now referred to the primitive genus Stegolo- 
phodon and to the more progressive genus Stegodon. 

2. HISTORY OF THE GENERIC NAMES ASSIGNED TO THE STEGOLOPHODONTS AND TO THE 

STEGODONTS 

The history of the generic term Stegodon Falconer is fully given in Chapter XXI, also in the present Chapter 
XI\', which may be sunnnarized as follows: (1) The name Stegodon applies chiefly to the more progressive 
Stegodonts. (2) To the more primitive Stegodonts, e.g., the Upper Miocene [Middle Phocene] MastodoJi 
cautleyi, the Lower Pliocene' M. latidens, and the Pliocene M. stegodontoides, the name Stegolophodon Schle- 
singer is applicable. The name Stegolophodon also applies to the Mastodon {Bunolophodon) longirostre Kaup 
forma suhlaiidens Schlesinger of Austria. The name Stegolophodon Schlesinger, 1917, preoccupies the name 
Prostegodon Matsumoto, 1922-1924. (3) The name Parastegodon Matsumoto, 1924, based on the genotypic 
species Elephas aurorse, appears to be in part a synonym of Archidiskodon. These interpretations of the generic 
names ado])ted in the present Memoir are displayed in the following table. 



i'i(vi,( )( ;i;\v. s ri;(;oi.()PHODONTs and stegodonts arranged uy countries in approximate ascend- 
ing ()I{Di:i{ OI' EVOLUTION, THE MOST PRIMITIVE FORMS BELOW, THE MOST PROGRE-SSIVE FORMS ABOVE 

|Th(; species described by various authors since 1929 are omitted here owing to tiic fact that they had not been studied in 
detail by Profe.ssor Osborn. — Editor.] 

STEGODON 

Stegodon orienlalis shodoensis Matsumoto, Island of Shodo, In- 
land Sea, Japan 
-1924 Elephas {Prostegodon, Parastegodon) aiirorw Matsumoto, 
Mt. Tomuro, Kaga, .lajmn 

Stegodon 7nindanensis Naumann, Mindanao, Phihppine Islands 

Stegodon ganesa var. javanicus Dubois, Kendeng. Trinil, Java 

Stegodon trigonorephalns Martin, vicinity of ?Surakarta, .Ia\-a 

Stegodon Airdirana Martin, Alas-Tuwa, Kendeng, Trinil, Java 

Stegodon orieiitolis Owen, Chungkingfoo, Szechuan, China 

Stegodon orientalis gratigeri Osborn, Yenchingkou, Szechuan, 
China 

Stegodon pinjorensis Osborn, near Siswan, India 
(184.5] Stegodon (ranesa Falc. and Caut., Siwahk Hills, India 
] 184.5] Stegodon insignis Falc. and Caut., Siwalik Hills, India 

Stegodon insignis birmanicns Osborn, Mingoon opposite Man- 
dalay, I^vu-ma 

Elephas honibifrons I'alc. and Caut., Siwalik Hills, India 

Mdtitodou elephanloides Clift, near Yenangyaung, Irrawaddy 
River, Burma 

Elephas cliftii Falc. and Caut., near Yenangyaung, Irrawaddy 
River, Burma 

Stegodon sinensis Owen, near Shanghai, China 



Jajiaii 


1924 




1915^ 


Piiilippines 


1890 


Ja\a 


1908 




1887 




1890 


China 


1870 




1929 




1929 


India 


1846 




1S46 


r>in'iiia 


1929 


India 


IS46 


liurma 


IS2S 




1840 


( 'liiiia 


1870 


India 


19 1:5 


Austria 


1917 


Borneo 


19.% 


l?urma 


1828 


India 


1929 




ISSG 




1929 



= Stegodon orieidaUs shodoensis 

= Stegodon aurorx 

= Stegodon (Archidiskodon?) mindanensis 

= Stegodon airdwana orS. trigonocephalus] 

= Stegodon trigonocephalus 

= Stegodon airdwana 

= Stegodon orientalis 

= Stegodon orientalis grangeri 
= Stegodon pinjorensis 
= Stegodon insignis-gnncsa 
= Stegodon insignis-ganesa 

= Stegodon insignis birmanicus 
= Stegodon bombifrons 

= Stegodon elephantoides 

= Stegodon elephantoides { = cliftii) 
= Stegodon sinensis 



STEG()LOPHODON 
Mastodon stegodonloides Pilgrim, Lehri, Punjab, India 
Mastodon {Bunolophodon) longirostre Kaup forma sublatidens 

Schlesinger, Teschen (Schlesien), Austria 
Stegolophodon lydekl\~eri Osborn, near Bruni, Borneo 
Mastodon latidens Clift, near Yenangyaung, Irrawadd.y River, 

Burma 
Strg(ili}])hi}doii i-onltci/i progressus Osborn, near ( 'hinji I'ungalow, 

Iiiiiia 
Mastodon raidlei/i Lydekker, Perim Island, India 
Stegolophodon nuthotensis Osborn, near Nathot, India 



'[See note on page 824 above regarding the Lower Pleistocene age of Mastodon [=Siegolophudon] latidens. — Editor 



= Stegolophodon stcgudontoides 

■■ Stegolophodon sublatidens 
■■ Stegolophodon lydekkeri 

■■ Stegolophodon latidens 

■Stegolophodon ranltei/i progrcsmis 
■ Stegolophodon roiilUyi 
Stegolophodon nothotensis 



THESTEGOLOPHODONTINiE: STEGOLOPHODON . 839 

Superfamily: MASTODONTOIDEA Osborn, 1921 
Family: MASTODONTID^ Girard, 1852 
Subfamily: Stegolophodontin^e Osborn, 1936 

Genus: STEGOLOPHODON Schlesinger, 1917 

Oiiginal reference: Denk. Xaturhist. Hofmus., 1917, I, p. 115. 

Genotypic species: Mastodon latidens Clift, 1828. 

Compare Stego (lopho) don Pohlig, 1888, p. 252. 

Syn.: Prostegodon Matsumoto MS., in Osborn, 1923; Matsumoto, 1924, p. 325. 

Generic Characters. — (Schlesinger, 1917, p. 115, footnote): "Ich schlage fur M. latidens, 
das sich durch seine kurze Syraphyse von dem Subgenus Bimolophodon, durch seinen Molarenbau von 
Dibunodon entfernt, den Untergattungsnamen Stegolophodon vor. Der Name bringt einerseits die 
nahen Beziehungen zum Genus Stegodon, anderseits die Loslosung der Untergattung von Bimolophodon 
und ihre Sonderstelhmg gegeniiber Dibunodon zum Ausdruclc." 

Osborn, 1926: (1) Six^ species of iStegolophodon, namely. Mastodon latidens, the more primitive M. 
cautleyi, and the more progressive M . siegodontoides; also S. cautleyi progressus, S. nathotensis, and 
Mastodon {Bunolophodon) longirostre Kaup/orma snblatidens. As defined by the teeth, these species have 
in common the following generic characters: (2) Lophs as in Mastodon and Zygolophodon, tendency to 
form from four to six transversely arranged cones and conelets (conelets somewhat irregular) and to con- 
solidate into ridge-crests; (3) molar pattern transitional between the Zygolophodon type and the 
Stegodon type; (4) ridge-crest formula known in Stegolophodon latidens as follows. Dp 3- Dp 4^ M 1 ^i^i^ 
M 2 |||4 M 3 li^. [Anterior ridge-crests with persistent median sulcus (see Vol. I, p. 700).— Editor.] 

This genus resembles Trilophodon and Tetralophodon in the retention of a broad enamel band on the straight 
superior incisive tusks (Figs. 725 and 727) ; it differs widely from Trilophodon in the presence oifour ridge-crests 
on the intemediate molars (Fig. 726) ; it also differs from Tetralophodon in the absence of trefoils and in the 
progressive tendency of the cones and conelets to form regular transverse ridge-crests surmounted by regular 
conelets, as seen in the genotypic species Mastodon [ = Stegolophodon] latidens, also in Mastodon [ = Stegolophodon] 
stegodontoides. 

Four of the known species are provisionally distinguished as follows : 

^Middle Pliocene Lower Pliocene- Middle(?) Pliocene Upper Pliocene 
Stegolophodon cautleyi Stegolophodon latidens Stegolophodon sublatidens Stegolophodon stegodontoides 
M2A M3^^ M2f|f| M3'^, (?)M3^' M3^* 
4-5 conelets on each ridge- 4-5 conelets on each ridge- 4-5 conelets on fourth, 5-6 conelets on each ridge- 
crest, very irregular. crest, more regular. fifth, and sixth ridge-crests, crest, very regular. Posterior 

Posterior ridge-crests arched ridge-crests arched or convexo- 

or slightly convexo-concave, concave. 

Stego(lopho)don (see Pohlig, 1888, p. 252). By strict rules this term may be regarded as a nomen nudum and 
should not stand in the way of Schlesinger's excellent name Stegolophodon. 

Prostegodon. (1) For the first printed use of the name Prostegodon, see Osborn, 1923.601, p. 2: "Prostegodon, 
new genus, Matsumoto. In a letter from Dr. H. Matsumoto, dated November 20, 1922, from Sendai, Japan, he 
writes: Tn my report just in preparation on the Japanese 'Mastodonts,' I follow you to refer "Mastodon" latidens 
to the genus Stegodon, creating however a subgenus Prostegodon for it. Prostegodon is the primitive representative 
of the Stegodon-phylum, representing half bunomastodontine and half stegodontine dental characters. Schlosser's 
opinion, that Prostegodon might be ? a descendant of "Mastodon" turicens, does not appear to be correct at all.' 
Genotypic species Mastodon latidens Clift. Tliis genus should be credited to Doctor Matsumoto." 



'[Seven species including Stegolophodon lydekkeri described in Vol. I, p. 700, of the present Memoir. — Editor.] 

-[See page 824 above where it is stated that S. latidens is limited to the upper Irrawaddy beds (Lower Pleistocene).— Editor.] 



840 . OSBORN: THE PROBOSCIDEA 

(2) In 1924 Matsunioto's report on the Japanese Mastodonts, referred to above, appeared in the Journal of 
the Geological Society of Tokyo, Volume XXXI, in which lie defined the genus Proslegodon (p. 325). This, 
however, was published in the Japanese language. 

(3) In 1926 Matsumoto published his English text on this genus ("On Two New Mastodonts and an Arche- 
typal Stegodont of Japan," 1926, p. 9), from which the following is a direct quotation: 

Skull and mandible onl.y inipcrfeptly known, brevirostral. Lower incisor-tusks might be absent, or abortive if present at 
ail. Intermediate molars four- or five-ridged, last molars five- or six-ridged, (irinders essentially lophodont, though their first 
and second ridges may show a slight tendency of bunodonty and of trefoil pattern of cusps; mesial longitudinal cleft evident; 
inner and outer cusps opposite, instead of being alternate; valleys widely open, free of cement. 

(4) Thus Proslegodon Matsumoto-Osborn, 1923, 1924, 1926, becomes a synonym of Stegolophodon Schlesinger, 
1917. 

(.')) The genus Parudegodon Matsumoto, 1924 (1924.2), founded upon the genotypic species Elephas {Prostego- 
dun) aurorx Matsumoto (Fig. 709), also probably [in part] becomes a synonym of Slegodon, because as shown below 
the genotypic species E. (P.) aurorse. [ = Slegodon aurorx of the present Memoir] is somewhat more primitive than 
Slegodon airdicana Martin of the Middle(?) Pleistocene and quite distinct from E. [Archidiskodon] plmiifrons. 
The name Paraslegodon was originally published by Matsumoto in 1924, pp. 256, 257: Paraslegodon gen. nov. 
= Slegodon mmdanensis-Elephas aurorse group ; type Elephas aurorse; it was subsequently cited by Matsumoto 
(1926.1, p. 1). Elephas aurorse was originally figured in 1918, PI. xx, as reproduced in figure 709 of the present 
Memoir. From the isolated type second superior grinder of the right side, r.M", it was difficult to determine 
whether this was a progressive Slegodon like S. airdwana or whether it was transitional to a primitive elephant 
like Archidiskodon planifrons. We have finally referred it to the genus Slegodon. 

SYSTEMATIC DESCRIPTION OF SPECIES OF STEGOLOPHODON 

Stegolophodon cautleyi Lydekker, 1886 Geologic Horizon. — According to Pilgrim-Osboru (see Fig. 

Figures 142, 08.5, 090, 701, 712-715, PI. xiii 413 of Vol. I), the type of Stegolophodon cautleyi from the Perim 

Lcctotype: Middle Pliocene, Perim Island, India; referred, Salt Range, Inland formation is of Middle Pliocene Dhok Pathan age, equal 

Simla Hills, India. approximately to the Plaisancian (of France) and Levantin (of 

Falconer's very accurate figure of the lectotype (Fig. 713) Austria and Hungary). In the same formation occur the types of 

displays clearly (1) the characteristic median fissure [or sulcus], Anancus perimensis, Deinothcrium indicum, and D. angiistideris, 

a i)rimitive character (cf. Fig. 68.5, 2-4), (2) vestigial median also Hipparion pernncnge type. Referred specimens of ,S'. cautleyi 

conules, also observable in Stegolophodon latidens, and (3) internal occur again in the Salt Range, Simla Hills, in the same level with 

conelets blocking the valleys. If N. cautleyi is not really ancestral specimens of Anancus perimensis, of Hipparion thcohaldi, and of 

to S. lalidens, it is certainly a much more primitive and less perfect Aceratherimn perimense. In the Lower Pliocene' it is rci)laced by 

Stegolojjhodont, retaining certain resemblances to the less regular Stegolophodon lalidens. 

molars of primitive species of Zygolophodon. The species is repre- Mastodon cautleyi Lydekker, 1886. "Addenda to Synopsis of 

sented in the present Memoir by the lectotyi)e (Figs. 712 and 713), Siwalik & Narbada Mammalia." Mem. Ool. Surv. Ind., Pal. 

also by the cotype of Lydekker (Fig. 714), and by the referred Indica, 1886, Ser. X, \o\. Ill, pp. xiv xix. Lk( totype and 

upi)cr molar (Fig. 715). Compare the superior grinding teeth with cotypes. — {Op. rit., j). xiv) : "The specimens on which this provi- 

the somewhat more progres.sive species Stegolophodon caidlcyi sional species is founded are five in ntmiber, and are all cheek-teeth 

progressus (Figs. 72.5-727). of the ui)per jaw; four of them being in an unworn condition." 

Specific Characters (Osborn).— Distinguished by median The third superior molar of the left side (Lydekker, 1886.1, p. xv, 

fissure [or sulcus], irregular formation of ridge-cre.sts .surmounted fig. 6) has been selected as the type (see Pilgrim, 1913, j). 294). 

by from four to five 'conelets' each, with smaH'conules' or 'acres- Hoiuzon and Locamtv.- -Perim Lsland, India; .Middle Plio- 

sory tubercles' in the valleys, as shown in the lcctotyi)c and cotype cenc. Lkctotvpe and Cotvpe Vigvhks.— (Op. cit., j). xv, 

figures (Figs. 714 = cotyi)e, 712 = lec1otype). M'' with .">+ ridge- figs. .5 and 6) : Fig. .5 (lirit. Mus. M. 2817, cast Amer. Mus. 2696.5), 

crests; ap. 210 mm., tr. 112 mm. figured in the "Fauna Antiqua Sivalensis," PI. xl, figs. 3, 3a, as 

'(See note on page 824 above regarding the Lower Pleistocene age of Stegolophodon lalidens. — Editor.l 



THE STEGOLOPHODONTIN.^: STEGOLOPHODON 



841 



M. kitidens (appears as M. cautleyi, Lydekker, 1S86.2, ]). 72, fig. 
17); V'lg. 6 (Brit. Mils. M.2705, cast Amcr. Miis. 26966), figured in 
the "Fauna Autiqua Sivalen.sis," PI. xxxi, figs. 6, 6a, as M. latidens 
(appears as M. cautleyi, Lydekker, 1886.2, p. 73, fig. 18) chosen as 
the type (by Pilgrim); Brit. Mus. M.2884 (no history), figured in 
the "Fauna Antiqua Sivalensis," PI. xl, figs. 2, 2a, as M. latidens 
(appears as M. cautleyi, Lydekker, 1886.2, p. 71); Ind. Mus. A.48 
(cast Brit. Mus. M.3428),' figured in Lydekker, 1880, PI. xl, as 
M. ])erimenf<is (appears as M. puujabiensis, Lydekker, 1886.2, p. 
60); Ind. Mus. A. 437 (cast Brit. AIus. 2887), figured in Lydekker, 
1884.3, PI. XVI, fig. 2, as M. perimensis (appears as AI. cautleyi, 
Lydekker, 1886.2, p. 72). 



distinction between typical molars of those two species, it has 
been thought, after considerable hesitation, advisable to pro- 
visionally apjiiy a distinct specific name to the aberrant form, 
which may be called M. cautleyi. . . . The specimens on which this 
provisional species is founded are fi\e in number, and are all 
cheek-teeth of the upper jaw; four of them being in an unworn 
condition. Three of these teeth, which are all from Perim Island, 
are in the British Museum, and are figured in the 'Fauna Antiqua 
Sivalensis,' under the name of M. latidens: the first (pi. xl. figs. 
2, 2a [Brit. Mus. M.2884]) is the right "'— ; the second (pi. xl. 
figs. 3, 3a [Brit. Mus. M.2817]) is the left"^, and is refigured of 
the natural size in woodcut fig. 5; while the third (]>!. xxxi. figs. 



Fiif. 18. 





MastQdo7i catUleyi. — Tlie third lei't upper true molup, iu an unworn condition; 
from the Siwaliks of Pcrim Island. J. The lower border of the figure ia 
the inner border of the specimen. 




Lectotype of Stegolophodon cautleyi 



Fig. 712. Lectotype (see Pilgrim, 101.3, p. 294) of Mastodon cautleyi Lydekker, 
1886, l.M', one-half natural .size. Middle Pliocene, Perim Island, India. Brit. Mus. 
M.2705. After Lydekker, 1886.2, p. 73, fig. 18; .same as Lydekker, 1886.1, p. xv, 
fig. 6. See also four cotype figures of Lydekker, namely, Lydekker, 1886.1, p. xv, fig. 
5; PI. XVI, fig. 2 (as M. perimensis); Lydekker, 1880.1, PI. xl (as M. perimensis), 
1886.2, p. 60 (as M. punjahiensis); Falconer and Cautley, 1846 [1847, PI. xxxi, figs. 
6, 6o (as M. latidens)]. 

Observe the irregular character of the ridge-crests and the presence of rudimen- 
tary intermediate conules in the first and second valleys (Lydekker, 1884.3, PI. xvi, 
fig. 2, a, a, b, a). 



Fig. 713. Lectotype of Mastodon cautleyi Lydek- 
ker (.same lei-totype tooth as that shown in figure 712). 
Third upper true molar of the left side, I.M', from Perim 
Island, one-third natural size. Originally figurotl by 
Falconer and Cautley, 1846 (1847, PI. xxxi, figs. 6, 6a| 
as Mastodon latidens, and described as follows (Falconer, 
1868, Vol. I, p. 463): "Figs. 6 and 6o.— .W. latidens. 
Upjjer true molar, very perfect. — B.M. Length, 8..) in. 
Width, 4.5 in." Brit. Mus. M.2705. 



Original Description (Lydekker, 1886.1, p. xiv). — 
"Mastodon cautleyi, n. sp. Lyd. — ... In recently describing 
a molar of Mastodon latidens from Borneo, the present writer 
[Footnote: 'Proc. Zool. Soc. 1885, pl. xlviii.'] referred to certain 
Siwalik specimens in the British and Indian Museums which 
appeared to indicate a more or less complete transition between 
typical molars of that species on the one hand and those of M. 
perimensis on the other. A subsequent examination of the speci- 
mens in question has however led to the conclusion that they 
cannot apjiarently be satisfactorily referred to either one of those 
species; and as there can be no question as to the strongly marked 



6, 6a [lectotype, Brit. Mus. M.2705]) is the left"- , and is refigured 
on a larger scale in woodcut fig. 6. The other two specimens are 
in the Indian Museum, and are figured in the present work under 
the name of M. perimensis; the first'" [Footnote: ' "Cat. Siwalik 
Vert. Ind. Mus." pt. I, p. 97. No. A.48. (1885). M. perimensis.'] 
(vol. I. pl. XL.) being the partially-worn left "— , with the as.sociated 
^^, the former [Footnote: 'When describing this specimen in 
Calcutta the writer could not identify it with the one figured in 
the "F.A.S." pl. XL. fig. 3, owing to the small size of the figures in 
that work, which renders them almost useless for comparison.'] 
of which apparently agrees precisely with the homologous British 



'[Lydekker subsequently (1886.2, p. 60) made this (cast Brit. Mus. M.3428) a cotype of Mastodon [ = Tetralophodon] punjabiensis (cf. Vol. I, p. 363, of the 
ent Mftmoir"^ — P^Hitnr 1 



present Memoir). — Editor.] 



842 



OSBORN: THE PROBOSCIDEA 



Museum specimen, and the other'" [Footnote: '"Cat. SiwaUk 
Vert. Ind. Mus." pt. I. p. 97. No. A.437 (1885). M. perimcn.v.s.'] 
(vol. ITT. pi. x\i. fig. 2) the imperfect right "", in an unworn 
condition." 

Specific Characters (Lydekker, 1886.1, p. xv). — "All 
these five specimens agree so exactly with one another that there 
can be little or no hesitation in referring them to one and the 
same species. Their essential characters are that the ridges are 
moderately tall, and inclined forwards; the valleys partially 
blocked by accessory tubercles, of which there are none on the 
outer side of the median longitudinal cleft. The first inner column 
always has accessory tubercles on both sides, and there are similar 
tubercles on the hinder side of both the second and third inner 
columns: the hind talon of the 'intermediate' molars (woodcut 
fig. 5) is relatively small; while in unworn e.xamples the hinder 



also from Perim Island. According to these specimens, the ridge 
formula of Mastodnn [ = Stegolophodon] caiitlci/! is as follows: 

Ridge formula: Dp i'-^=^ M l*-^"^^ M 3^^. 

According to P'alconer and Lydekker, (1) the same ridge 
formula is observed in specimens of Mastodon [ = Stegolophodon] 
cautleyi from Perim Island (type locality) and from the Dhok 
Pathan, namely, M 2- M 3—; (2) the open and irregular ridge 
structure of the molar teeth in S. cautleyi is more primitive than 
the close-set, regular ridge structure in the type of .S'. lalideitti, as is 
seen at once by comparison of the type figures of the two species; 
(3) Lydekker separated the species Mastodon cautleyi from speci- 
mens in the Perim Lsland beds which had been referred by l''alconer 
to Mastodon latidens; (4) Pilgrim (1913, p. 294) accepted this 
species, regarding it as a direct ancestor of one of the larger forms 
of M. latidens. 

Fig. 14. 



Fie;. 17. 





Maiiodon penmcnsis.— Uhe second left upper triie molar, in nn unworn con- 
Masiodon caullcT/i.— The Rrsl Mt upper true molar in an unworn contlition; dition ; from the Siwaliks of Perim Island, j. a. Eiternal accessory 

from the Siwaliks of Perim Island. }. The lower border of the figure is tubercles. The lower border of tlie figure is tlio inner border of the speci- 

the inner border of the specimen. men. (From the ' Palajontologia Indica.') 

CoTYPE First and Refebred Second Superior Molar.s op Stegolophodon cautleyi 

Fig. 714. Cotypc of Mastodon cautleyi Lydekker, 1886 (Brit. Mu.s. Fig. 71.5. .1/astoiore pmmens^'s (Inil. Mus. A.355, cast Brit. Mu.s. M.2851). 

M.2817), p. XV, fig. r,. Compare "Fiiuna .'\iiti(|iia Sivalensi-s," Fnle. and After Lydekker, 1886.2, pp. 57 and 58, fig. 14. 

Caut., 184611847, Pi. xi-, figs. 3, 3a (;l/. Inlulcns)]. After Lydekker, 1886.2, Reierrrd by Oshorn to Stegnlopkmlon caiillfi/i. Compare type of ,S. mi/^r'i/J 

p. 72, fig. 17. progressus (Fig. 726). 

Compare type of S. cautleyi progressus (Fig. 726). 



aspect of the outer column of the first ridge is deeply concave, and 
the arrangement of the tubercles on the inner column of the same 
ridge forms a V. The third true molar (woodcut fig. 6) is very 
wide, tapers but little posteriorly, and carries five ridges and a 
simple hind talon, the latter consisting of a narrow ridge with six 
small tubercles. All the teeth are relatively wide, with a well- 
marked median longitudinal cleft; they appear to have no a]3preci- 
able quantity of cement in the valleys, and when worn . . . present 
trefoils on their inner columns." The cotypc figure reproduced 
above (Fig. 714) is a first left superior true molar from Perim Islaiui ; 
the lectotype (Figs. 712 and 713) is a third left .superior true molar. 

'[Cast Brit. Mus. M.2887.1 



Stegolophodon latidens Clift, 1828 
Figures 685, 693, 694, 716-721, 757, PI. xiii 

Lectotype: Lower Pliocene, lowest levels of the Irraw.addy Series 
(fluviatile)- near Yenangyaung, Burma; referred, Middle Pliocene of India 
(Dhok Pathan zone), and Ixnver (7) Pliocene of China and Japan. 

This is the chai'acteristic Lower Pliocene- stage of the Irra- 
waddy Series, Burma (lectotyi)c and cotypcs) and of the Middle 
Pliocene stage of the Siwaliks and Perim Island, India (referred 
specimens). 

Specific Cmauacters (Osborn). — Distinguished by more 
regular formation of the ridge-crests than in Stegolophodon cautleyi, 



'[See note on page 824 above regarding the Lower Pleistocene age of Stegolophodon latidens. — Editor.) 



THE STEGOLOPHODONTIN^: STEGOLOPHODON 



S4:{ 



siinuounted by four to five conelcts each, with two vestigial 
'conules'; median fissure [or sulcus] in anterior ridge-crests; ridge- 
crests more stegodontoid. Hidge-crest formula: M 2 HtjM 3 ^'.01. 
This important stage, originally described as Mastodon by 
( lift and as Slegodon by Falconer, has been aptly chosen as the 



STEqOLOPHODON 



LATIDENS 




C(ist 


/Imer. Mu.z. 2/978 


r. m? 


7- TTi ^ 




'/2 naf. Sije. 



New Figure of Lectotype of Stegolophodon latidens Superior Molars 
Fig, 716. New fiKurc of lectotype of Mastodon lalideiis Clift, 1828, PI. xxxvii, fig. 1 
r.M'-', r.M"'. Strong orthogonal projection after cast (Anier. Mils. 21978), one-half natiira 
size. Compare figure 693. From near Yenangyaung, Burma. Original in British Museum 



genotyjjic species of Slcgotophodon Schlesinger, and subsequently 
of Prostegodon Mastumoto-Osborn. The ci'own of the superior 
molar teeth, r.M'-, rM', from i^urma, represents an advance upon 
the species Stegolophodon eaitlli'i/i towards the true Stegodon type of 
molar; the three anterior ridges are closely compacted, closing the 
bottom of the three valleys (sec Fig. 719). Each ridge consists of 
four plus compres.sed, rounded conelets, as shown in the protoloph 
and metaloph of M'; these conelets soon wear into uniform, trans- 
verse ridges (see tritoloph of M-). None of the specimens referred 
by Falconer, from Perim Island, agrees in these characters with 
the cotypes of ^. latidens; some of the specimens referred by 
Falconer and Lydekker, from the Middle Pliocene Dhok Pathan, 
and other beds, also do not agree with the cotypes of S. latidens. 
We must accordingly define the species from the lectotype, 
following closely the original definition of Clift, together with the 
characters displayed in the se\'eral views of the lectotype molar, 
M^, reproduced herewith. 

The measurements of these molars (Figs. 718, 716, 719, 720, 
717) in millimeters (with equivalents in inches) are as follows: 



foiuid on the loft Bank of the Irawadi." Trans, (ieol. 8oc. London, 
(2), II , P( . 1 II, 1828, pp. 309 375. Lectotype.— Palate con- 

taining righl M-, Ml C'OTYi'E. — Kight lower jaw with M.-i 

in siiu. Originals in British Museum (Natural History), London; 
ca.st of r.M-, r.M' (Amer. Mus. 21978). Horizon and Locality. — 
Left bank of the Irrawaddy River, near Yenangyaung, 
250 miles below Ava, Burma; Lower Plioeene, lowest 
levels of the Irrawaddy Series (fluviatile).' Lecto- 

type Figure. — Op. cit., PI. xxxvii, fig. 1. C'otype 

Figure. — PI. xxxviii, fig. 1. 

Characters (compare Falconer and Cautlev, 
1846, p. 48). — Fi\e and a half ridges in the third supe- 
rior molar, including a posterior talon ridge and a 
subordinate ridge in front; four ridges in the second 
superior molar; four plus mammilla> on each ridge; 
ridges trans\erse, di\ided by a longitudinal cleft into 
two pairs of principal points [conelets], without inter- 
mediate mammillse [conules] in the hollows. Enamel 
very thick; cement reduced to a thin layer, observed 
in the bottom of the hollows. 

(Clift, 1828, p. 371) : "Each tooth of the lower jaw 
[Fig. 717] consists of seven denticules, which are ele- 
vated, rounded, and mammillated; the mammillije being 
from three to four in number." 
Characters Shown in Three Figi res of the Cotypes. — 
The three figures of the cotype molar teeth reproduced herewith 
show that Clift's original figtu-e (Fig. 693) gives a somewhat 
wrong impression as to the rounded character of the conelets of 
which the ridges are composed. The more acuminate character 
of the.se conelets is displayed in Falconer's section reproduced 
herewith (Fig. 719, fig. 8) in which it appears that the three 
anterior valleys tend to close in at the bottom, unlike the open 




Second right superior molar, r.M^ 




Anteroposterior 


98 = 3JUnches 


Transverse 


6S = 2K 


Third right superior molar, r.M^ 




Anteroposterior 


140 = 5K 


Trans\'erse 


76 = 3 


Third right inferior molar, r.Ms 




Anteroposterior 


300=11?^ 


Transverse 


101=4 



CoTYPE Third Inferior Molar of Stegolophodon latidens 
Fig. 717. Cotype third right inferior molar, r.Ms, of Mastodon latidens 



Clift, 1828, PI. xxxviii, fig. 1, about one-third natural size. 
From near Yenangyaung, Burma. 



Jaw omitted. 



Ma.ilodon latidens Clift, 1828. "On the Fossil Remains of two 
New Species of Mastodon, and of othei vertebrated Animals, 



valleys in Mastodon ameficanii.'i (Fig. 719, fig. 9). Thus a more 
accurate figure than Cliffs is that of Gaudry (Fig. 720) probably 
made directly from a cast of Cliffs specimen. This figure shows 
traces of trefoil conules in the anterior and second valleys and alto- 
gether presents a closer resemblance to the Stegolophodon rautletji 
form of molar tooth, jvistifying Pilgrim's statement that a large 
form of .M. latidens is derivable directly from M. cautleiji. 



'[See note on page 824 above. — Editor.] 



844 



OSBORN: THE PROBOSCIDEA 



History of Discovery. — The history of discovery of this 
species (Stegolophodon latidens) in 1826 by CHft is very fully set 
forth in this Memoir (pp. 825, 827 above) and need not be repeat- 
ed here; also Cliffs original description and ridge formula as 
abbreviated above. Attention should be concentrated on the 
characters of the cotype specimens as shown in the four figures of 




Tetralophodon punjabiensis, Hipparion pimjabiense, Tetraconodon 
magnus, and Tauroiragus latidens. 

Referred specimens of Stcgolophodnn latidciif:, according to 
Pilgrim, also occur in the overlying Dhok Pathan beds (Salt Range) 
associated with the types of Tetralophodon punjabiensis, Tetra- 
belodon [ = Synconolophus] cornigatus, Mastodon [ = Si/nconolophus] 
hasnoti, and Stegodon bombifrons. It is probable that 
specimens referred by Falconer to Stegolophodon lati- 
dens from the Upper Miocene [Middle Pliocene] of 
Perim Island belong rather to *S'. cautletji.. 

r 




Lectotype Molars of Stegolophodon latidens 
Fig. 718. [Perspective.] Lectotype of 3/o.5/o(ton iahticHsClift , 1828, reduced from Cliffs 
PI. XXXVII, fig. 1, to one-half natural size. Found on the left bank of the Irrawaddy River, 
below Ava, Burma. (Clift, 1828, p. 371 and Explanation of Plate.s): "The i)alate, and 
molar teeth of the right side of Mastodon latidens (a younger animal than the last). The 
anterior tot)tli is very much worn. The anterior part of the posterior tooth appears to have 
been just brought into use." These teeth include a right M- and M^. The ridge-crests arc: 
M 2^^^ M Z^iiSiA^, Cast Amer. Mus. 21978. See orthogonal projection (Fig. 71(i)- 



231. Premifere arrifere-molaire superieurc du Mastodon latidens, 

a 1/2 grandeur, dicouverte par Crawfiird, dans I'Avu. 

Fig. 720. Lectotype M' of Mastodon [Stegolophodon] latidens 
Clift, 1828, reproduced from an engraving by Gaudry (1878, 
p. 17.5, fig. 231). The third superior molar discovered by Craw- 
furd below Ava. 

Geologic and Geographic Distribution. — The 
cotypes oi Stegolophodon hitidens are from the Irrawaddy 
River, 250 miles below Ava, near Yenangyaung, 





Fig. 719. (Left) Section of molars, M^ M^ (Jde Falconer, Pal. Mem., 1868, Vol. I, p. 424) of Ma.-ilodon Uilidens Clift, 1828 [Stego- 
lophodon latidens, lectotype], reduced to one-third natural size. After Falconer and Cautley, 1846 [184.'3, PI. iii, fig. 8). Compare witli 
tlie crown view of same molar reproduced above. 

(Right) Section of M^ of Mastodon americanus, from Missouri, op. cil., PI. in, fig. 9, inserted by Falconer for comparison. Rrit. 
Mus. 17420. 



the superior molars reproduced above(Figs. 718,719,720,716), and 
in the lower jaw (see Fig. 717 for r.Ma). 

Geologic Age. — The type of Stegolophodon latidens, according 
to Pilgrim (notes of April .3, 1923), probably occurs in the lowest 
levels of the Irrawaddy Series (fluviatile), near Yenangyaung, 
Burma, of Lower Pliocene age.' In beds of the same age in the 
Salt Range, Pabbi Hills, Jammu, and Kangra, are also found 

'[See note on page 824 above. ^Editor. 1 



Burma. (1) In the Calcutta Museum are listed by Lydekker 
(1885, pp. 94-97) thirty-three specimens referred to M. latidens 
('lift, as follows: (a) Burma— six specimens of grinding teeth; (b) 
Asnot, Punjab— five grinding teeth; (c) Sind — five grinding teeth; 
(d) the remainder are from the Punjab, Potwar district, Gadari, 
Lehri, Jabi, Niki. (2) In the British Museum collections are listed 
by Lydekker (1886.2, i)p. 74-78) eleven specimens from: (a) The 



THE STEGOLOPHODONTIN^: STEGOLOPHODON 



845 



Pliocene Siwaliks of the Punjab; (b) from near Yenangyaung, 
upper Burma; (c) Perim Island, Gulf of Cambay; (d) the Plio- 
cene of Bruni, northwest Borneo [made by Osborn (1936, Vol. 
I, p. 700, of this Memoir) the type of Stegolophodon lydekkeri]; 
(e) Lehri, Punjab. 

Lower Pliocene' Age. — Pilgrim assigns (Ghap. XXII below) 
Stegolophodon latidens cotypes to Burma, Irrawaddy Series, lowest 
levels (fluviatile), and referred specimens to the Jammu and 
Kangra beds of India, also to the Salt Range and Pabbi Hills. 
According to Pilgrim the species ranges upwards into the upper 
Irrawaddy levels (fluviatile), Burma, and into the Salt Range, 
Hasnot, Dhok Pathan zone of India, equivalent to the Middle 
Pliocene Plaisancian (of France) and Levantin (of Austria and 
Hungary). Stegolophodon latidens may be regarded therefore as 
a characteristic Lower Pliocene' stage of the Stegodonts; whereas 
the more primitive Stegolophodon cautleyi specific type belongs 
in the Middle Pliocene of Perim Island, also referred specimens 
of the Salt Range, Simla Hills. 

NOTES OF FALCONER, 1868, ON MASTODON [STEGOLOPHODONl 
LATIDENS, ALSO OF LYDEKKER, 1886 
Falconer, "Palaeontological Memoirs," VoL I, pp. 424, 463, 471, Pis. iii, 
XXX, XL of the "Fauna Antiqua Sivalensis"; Lydekker, "Catalogue of the 
Fo.s.sil Mammalia in the British Museum (Natural History)," Part IV, 1886, 
p. 74, and "Siwahk & Narbada Probosc-idia," 1880, p. 236. 

Stegolophodon latidens. Lectotypc. Plate iii, fig. 8, IVP 
ridges 4)2, M' ridges )^-5-)^, 4 conelets or denticles, enamel thick, 
cement thin. 

Upper Molars. — Plate xl, figs. 1, la, ?second upper milk 
molar, Dp', with 2 ridges; figs. 2, 2a, ?third upper milk molar. 
Dp'', with 4 ridges; figs. 3, 3a, M' ridges 4 and back and front heel 
()4-4-'4 ridges), length 4 in. = 101 mm., width 2.3 in. = 59 mm. 
Lower Molars. — Plate xxx, fig. 6, lower jaw, I.M3, 5 ridges and 
a double heel (5J2 ridges), no cement, length of molar 11.3 in. = 
288 mm., width 4.5 in. = 114 mm. 

From these observations of Falconer, and from those of 
Lydekker below, we may write the following collective formula, 
jM-imitive and progressive : 

Stegolophodon latidens: Dp 3^ Dp 4- M 1 ^^^^^ M 2^ 

^^^ "^ i4-5-^i■ 

Osborn, 1927: Cliffs figures (Figs. 717 and 718) as well as 
his description assign to M 3 the following ridge-crests -^ which 
should be interpreted as ^fff^f. This is approximately in accordance 
with Falconer and Lydekker's formula, although, as in all the Stego- 
donts, it is difficult to count the new or rudimentary ridge-crests. 

Characters (Falconer, 1868.1, Lydekker, 1886.2). — (1) 
The presence of a median furrow between the primitive inner and 
outer lobes, (2) of four conelets or denticles on each ridge, (3) of 
thick enamel and thin cement, and (4) of conules in the two anterior 
valleys — the.se are the chief characters assigned by Falconer, 1868. 
Lydekker (1886.2, p. 74) characterizes the dentition as follows: 
"The mandibular symphysis is not certainly known, but it was 
probably short and tuskless. The upper molars are \ery wide, with 
no cingulum and relatively low transverse ridges, the median longi- 
tudinal cleft being frequently indistinct, the accessory tubercles 
small and the valleys comparatively open, the trefoils on the worn 
ridges imperfect, and the worn dentine surfaces on the inner and 

'[See note on page 824 above. — Editor.) 



outer columns soon uniting. The hind talons are very large, while 
the inner border of the crown is concave and without a distinct 
cingulum ; a fifth ridge is sometimes present in "' - . . . ; and the 
hind talon of "^ is always large, and may sometimes be reckoned 
as a sixth ridge. The enamel is very thick, and in the hinder 
teeth quite smooth ; cement is usually absent, and premolars were 
probably developed." 

Summary of Characters (Lydekker, 1880.1, p. 236). — 
"Having now passed in review the molar series of M. latidens, we 
may sum up what is known regarding the species. The adult 
cranium is unfortunately quite unknown ; the palate is noticeable 
from the extent to which the molars converge anteriorly. The 
mandible is known by a specimen of the greater part of the right 
ramus, containing the two last molars, in the collection of the 
Indian Museum. This mandible is very long and slender, and 
sub-circular in cross-section, in the middle its vertical diameter at 
the penultimate molar being 5.8 inches and its transverse diameter 
5.6 inches. The lower border is nearly straight up to the symphy- 




Referred Stegolophodon latidens of Japan, a very Primitive Stage 
Fig. 721. Referred 4}l'+ ridge-crested Stegolophodon latidens {cautleyi^) 
from Sliiwogama, Miyagi District, Province of Rikuzen, Japan; "probably of 
typical Pontian age [Lower Pliocene]." Cast Amer. Mus. 22610. Inner and 
superior aspects of a third right upjier molar, r.M', with 4}^+ ridge-crests. 
(Upper) Internal aspect; (lower) crown aspect. After Matsumoto, 1926, 
PI. V, figs. 1 and 3: "Fig. 1. Proslegodon latidens (Clift); right D* or M'; 
inner view; natural size. . . Fig. 3 . . . [crown] view; natural size." Reduced 
in the present figure to one-half natural size. Length of r.M' along median 
line 115 mm. Compare Mastodon (Banolophodon) [ = Stegolophodon] sublatidens 
Schlesinger (Fig. 710). 

sis, with a slight convexity in the middle. From the extremely 
small size and circular section of the ramus where broken off at 
the commencement of the symphysis, the latter must have been 
short, and was probably unprovided with incisors." 

"The ridge formula of the milk-molars is constant, and con- 
forms to the normal tetralophodont order; in the true molars, 
however, there is not an unfrequent tendency to assume a pentalo- 
phodont type, the hind-talon of many of these teeth not un- 
frequently, partly or entirely, taking the form of a fifth ridge. In 
the following ridge-formula these varieties are indicated : — 

Milk-molars. True molars. 



2+3+4 
2+3+4 



(4-5)+(4-5)+(5-6) 
4+(4-5)+5 



846 



OSBORN: THE PROROSCIDEA 



This tendency to the prfKluction of an additional ridge in tlie 
true molars of M. Jnlideiix will be siil)se(|uently shown to he 
a character which it possesses in connnon with M. .s/ra/cz/.s/.s. 
This tendency to variation enables us easily to com])rehend how 
the passage from the tetralophodont Mastodons to the hexalopho- 
dont Stegodons (S. cliftii) was effected." 

STEGOLOPHODON LATIDENS IN CHINA AND JAPAN 
According to Schlosser (1903, p. 45), in tiie Munich collection. 
from an unknown locality, is a fragment of a lower molar tooth 
referable to -1/. hilidens. Schlosser's figure of this fr.agnient i.s re- 
produced in figure 770, 7. 

A primitive Japanese stage (Fig. 721) has been referred to this 
species by Matsumoto; it apparently possesses but four and 
a half ridge-crests and thus approaches the »S'. caiitlcyi stage (cf. 
Figs. 713 and 712). 

Stegolophodon sublatidens Schlesinger, 1917 

Figures 710, 722, PI. xiii 

MiddleC?) Pliocene age. From near Te.sehen (Sehlesien), Au.stria. 

This species (Fig. 722) is founded ujion the posterior half of 
a superior molar quite closely comparable to Pilgrim's type (Fig. 
723) of Mastodon [ = Stegolophodon] stegodontoides; it is more 
prinuti\'e in every way but the affinity is very striking; it thus 




Type of Stegolophodon sublatidens 
Fig. 722. T\\)C of Mastodon (Bunolophodon) lungirostre Kaup funiia sub- 
latidens Sclilesingcr, 1917, posterior half of a tliird right superior mohir, 
r.M'. From near Teschen (Schlesien), Austria, Middle(?) Pliocene. Collec- 
tion of the Royal Museum of Natural History, Vienna ("alter Sammlungs 
bestand"). Photograph loaned by Doctor Schlesinger. Reproduced one-half 
natural size. Compare Schlesinger, 1917, Taf. xv^i, fig. 2. 

The five conelets, of which two of the ridge-erests are i-omposcd, are 
arched transversely, a character not observed in Stegolopliodon liUidcns, in 
which the ridge-cre.sts are directly transverse, but seen in the type of Stego- 
lophodon stegodontoides. 

appears to represent the first true Slrgolnphodoti disco\-erc(l in 
Europe. 

Spkcific Characters. — Inferior in size to Stegoloitliddnii 
stegodontoides] valleys between ridge-crests more oi)en, four to 
five conelets on fourth, fifth, and sixth ridge-crests; each j)osterior 
ridge-crest arched or convexo-concave; width of tetartoloph 82 
mm., width of tetartoloph in .S'. stegodontoides 107 mm.; rudimen- 
tary cement in the valleys. 

Mastodon (Bunolophodon) longirostre Kauj) fon/io sulilnliilens 
n.f. .Schlesinger. "Die Mastodonten des K. K. Xaturhistorischen 
llofmuseums." Dcnk. Xatiuhist. Ilofmus., I, 1917, pp. 101, 



102. Typk. — A third light superior molar, v.'SP. ('ollection 

(if the Royal Museum fif Natur.al History, \'ierma (alter Samm- 
lungs bestand). IIohizo.n axd Focamiv. — From near 

Teschen (Schlesien), Austria; MiddleC?) Pliocene. Tvi'k Fk;- 

I'RE. — Op. cit., Taf. XVII, fig. 2. 

Type Description.— (Schlesinger, 1917, i)p. 101, 102): "Es 
ware unmoglich, diesen Zahn, den icli als Typus e'lner forma sublati- 
dens n.f. auffasse, innerhalb der Variationsbreite von M. longi- 
rostre anzufiihren, wenn nicht die vorbeschriebenen Molaren direkt 
zu ihm iibei'leiten wiirden. PrakLisch ist der Zahn— stegodont, d.h. 
cr weist jene Form von Jochzithnigkeit auf, die im (iegensatze 
zu den Verhaltnissen bei den tapiroiden Mastodonten durcli Um- 
bildung der Sjierrhocker zu Xel3en])feilern und jMnordnung in die 
Flucht gleichgrosser. Mammillen erreicht wird, wahrend .jene 
.Vrten Ilaupt- und Nebenhocker versehmelzen und die Spei-r- 
hoeker als Crista nach aussen drilngen und dann unterdriicken 
[Footnote: 'Ich kann nicht umhin, auf die liestiitigung des "(ie- 
setzes der Nichtumkehrbarkeit der Entwicklung" (L. Dollo) hin- 
zuweisen, die darin liegt, dass eine funktionelle Zygodontie von 
der Bunodontie aus nicht durch ofi'ene Riickkehr zum Urtypus 
(M. pygmaeum), sondern durch die Rildung der ganz anderen 
Stegodontie erreicht wird.']." 

History. — In defining the genus Stegolophodon, with the geno- 
typic species Mastodon latidens ('lift, of Burma, Schlesinger implies, 
but does not so state with definiteness, the inclusion of his species 
'Mastodon stihlatidens' within the genus Stegolophodon. His type 
of Mastodon sublatidens (Fig. 722) is an imperfect third superioi- 
molar of the right side, r.M', from Schlesien, Austria, of Pliocene 
age. The Jive conelets, of which two of the crests are comi)oscd, 
ai'e arched transversely; this arcuate character is exceptional and 
does not appear in the two species of Zygolophodon (Z. borsoni, 
Z. pyrenaieus) nor in the genotypic species of Stegolophodon, name- 
ly, Mastodon latidens of Burma, in which the ridge-crests ai'C 
directly transverse. The resemblance to Stegolophodon stegodon- 
toides is much closer, although S. sublatidens is a smaller and more 
jirimitive form, with more open valleys and less distinct division of 
the cones into five conelets, as admirably shown in the accompany- 
ing figures 722 and 723, the types of these two .species. 

Stegolophodon stegodontoides I'iliirim, 1913 
Figures 68.5, 708, 723, PI. xiii 
Lehri, in tlic Piuijal), India, possibly tlpi)er Siwaliks. ITpperC?) Pliocene. 

This is the characteristic Upi)er(?) Pliocene stage of the Pin- 
jor horizon. Upper Siwaliks. 

Specific Characters (Osbohn). — Larger and much more 
|)rogressive than Stegolopfiodon latidens. M' measures ap. 210 mm., 
tr. 110 mm., approximately same dimensions as the type of 
Mastodon [ = Stegolophodon] cautleyi. Ridge-crests 6)2 with five to 
six conelets. ^'alleys compressed or part ly closed, less ojjen t han in 
S. latidens. \estige of median fissure on three anterior ritlgc-ciests 
(Fig. 723). 

I'he tyi)e tooth of this species is transitional between the 
Stegolophodon and the Stegudon pattern, yet the species belongs to 
an independent line of descent contemporaneous in the rp])er 
Pliocene with an advanced stage of Stegodon such as Stegodon 
insignis. From Lehri, in the Punjab, India, Pilgrim (1913) selected 



THE STEGOLOPHODONTINJE: STECJOLOPHODON 



847 



a six and a half crested third right upper true molar, r.M', as the 
ty]>e of a new species Mastodon slegodoiitonles. lie remarked {op. 
ciL, p. 294): "The last species of Mastodon which can be referred 
to this [cauUeyi-latidens] line is the tooth from l.ehri, of which the 
horizon is uncertain but may be possibly Upper Siwahk." The 
type specimen was originally figured natural size by Lydekker in 
"Palseontologia Indica," ser. 10, Vol. I, 1880, PI. xxxix, as: 
"Mastodon (Tetralophodon) lalidens, Clift. The third right upper 
true molar: from Lehri, in the Punjab. The specimen is drawn 
of the natural size and is viewed from the inner [outer] side." 

Mastodon stegodonloides Pilgrim, 1913. "The Correlation of 
the Siwaliks with Mammal Horizons of Europe." Pec. fieol. 
Surv. India, Vol. XLIII, Pt. 4, p. 294. Type.^A third supe- 

rior molar of the right side, r.M^, Ind. Mus. A.86. Horizon 

.\ND Locality. — Possibly Upper Siwaliks, Pinjor formation, 
Upper (?) PHocene; Lehri, Punjab, India. Type Figure. — 




Type of Steoolophodon stegodontoides 
Fig. 723. Type r.M' of Mastodon stegodonloides Pilgrim, 1913. After 
Lydekker, 1880, PI. xxxix: "Mastodon {Tetralophodon) latidens, Clift. The 
third right upper true molar: from Lehri, in the Punjab. Tlie specimen 
is drawn of the natural size, and is viewed from the inner [outer] side." Ind. 
Mus. A.86. Reduced to one-half natural size. Provisionally placed in the 
Upper Pliocene, Pinjor formation (see Vol. I, fig. 413, also PI. xm). 

Lydekker, 1880.1, PI. xxxix, figured as Mastodon {Tetralophodon) 
latidens. 

Type Figure. — This fine half-size drawing of the type tooth 
measures: ap. 210 mm., tr. 110 mm. The ridge formula is M 3—. 
In the anterior three ridges of the tooth there is still a trace of 
subdivision into inner and outer lobes, which disappears in the 
posterior three ridges. The conelet formula is: protoloph — six, 
metaloph — five to six, tritoloph — four to five, tetartoloph — five, 
pentaloph — five, hexaloph — four; the conelets are more or less 
discrete, the valleys apparently open. Consequently this molar 
crown appears to be descended from that of Stegolophodon latidens 
and represents a marked advance towards the true Stegodoti type. 

Specific Characters (Pilgrim). — (1) Owing to the almost 
entire absence of accessory columns the name Mastodon stegodon- 
loides is proposed. (2) It will be seen that the type molar carries 
on none of its ridges more than the usual four columns [character- 
istic of Stegolophodon latidens], while the anterior ridges of Stegodon 
elephantoides { = cliftii) carry nine or ten mammillae [i.e., conelets]. 



Osborn, 1924: While we recognize the usual four columns in 
this molar tooth, we observe in Lydekker's figure that the summits 
of the two inner columns tend to subdi\ide ; thus in the tetartoloph 
and the pentaloph there are five distinct conelets, as compared with 
four plus observed in Stegolophodon latidens and in aS. cautleyi. 

Lydekker's Description (1880.1, pp. 235-237).— "77mf/ 
upper true molar. — The large tooth represented in Plate xxxix is 
a specimen of the last upper molar of the right side of M. latidens, 
collected by Mr. A. B. Wynne near the village of Lehri, in the 
Punjab. This specimen is implanted in a fragment of the maxilla, 
which also contains the two last ridges of the preceding or penulti- 
mate tooth. The figured tooth is entirely unworn, and was still 
covered by the gum at the death of the animal. The crown carries 
six transverse ridges, the hindmost of which is considerably smaller 
than the others, and probably rejiresents an ultra-developed talon, 
as v,e saw to be the case in the penultimate molar represented in 
fig. 1 of Plate xxxviii. The tooth consequently belongs to a 'pen- 
talophodont' type of dentition. The ridges are low and simple, 
and with the exception of the first, are slightly convex anteriorly, 
and as concave posteriorly. Each ridge is divided by a longitudinal 
cleft placed somewhat externally to the mesial antero-posterior 
axis of the tooth. The internal moiety of each ridge, with the ex- 
ception of the second and sixth, bears three mammillae or cusps, 
while the external moiety bears only two on each ridge. The val- 
leys are quite simple and uninterrupted. In this tooth, as also in 
all the previously described specimens, there is no trace of cement. 
The length of the specimen is 8.6 inches, and its greatest width 4.2 
inches." 

Lydekker's description and figure concur in giving five cone- 
lets as characteristic of each ridge and as showing the three poste- 
rior ridges more progressive than the three anterior. 

Also exhibiting this six+ ridged, five to six conelet condition, 
is a third left upper true molar from Borneo erroneously referred 
by Lydekker to Mastodon latidens.'^ 

Stegolophodon nathotensis Osborn, 1929 
Figure 724, PI. xiii 
Lower Chinji horizon, near Nathot, India; Mio-Pliocene. 

This type was discovered by Barnum Brown in 1922 in the 
Lower Chinji horizon, near Nathot, India, the exact level being 
unrecorded. As shown in figure 413 the species is of approximately 
the same geologic age as Trilophodon chinjiensis type, Deinotherium 
pentapotamise type, and Serridetdinus browni type. The Chinji 
horizon, of a total thickness of 2,400 feet, doubtless represents 
a \-ery long period of geologic time. 

Stegolophodon nathotensis Osborn, 1929. "New Eurasiatic and 
American Proboscideans," Amer. Mus. Novitates, No. 393, Dec. 
24, pp. 13-15 (Osborn, 1929.797). Type.— Amer. Mus. 

19455. Posterior half of a fragmentaiy right third suj^erior mo- 
lar, r.M^ (Fig. 724, C, CI); posterior half of a right third inferior 
molar, r.Ms (A, Al) ; anterior half of a left second superior molar, 
l.M^ (B, Bl). Horizon and Locality. — Found near Nathot, 

India, "Lower Chinji horizon, the exact level being unrecorded, 
lower Middle Siwaliks ; Middle to Upper Miocene [Mio-Pliocene]." 
Type Figure.— Op. cit., 1929.797, p. 14, fig. 14. 



■[Made by Professer Osborn in 1936 the type of Slegoloptiodon lydekkeri (see Vol. I, p. 700, fig. 660). — Editor.] 



848 



OSBORN: THE PROBOSCIDEA 



Type Characters. — Ridge-crests low, blunted, with four to 
five blunted conelets (Fig. 724 B, A) on each crest; ridge-crests 
transversely arcuate or arched, as in Stegolophodnn stegodonioides. 
Enamel thick with slightly indented border. Hugose externa! 
cingulum. Ridge-crest formula unknown, probably M 3 Ij^; 
characters of ridge-crests and conelets quite distinct from those 
of Stegolophodon latidens, ridge-crest formula probably lower than 
in S. latidens. Species at present known by the type specimen 
only, which belongs in a lower geologic level, namely, Mio-Pliocene, 
than the type of S. latidens, which belongs in the Lower Pliocene.' 
See measurements in legend (Fig. 724). 



No. 393, Dec. 24, pp. 13 15 (Osborn, 1929.797). Type.— 

Amer. Mus. 19446. A juvenile cranium (Figs. 725 and 727), 
containing right and left .superior tusks with broad enamel band, 
also in .situ third and fourth superior deciduous premolars and 
first molar of both sides (see Fig. 726, r.Dp'"'', r.M'). Horizon 

AND Locality. — "Twelve miles east of C'hinji Bungalow, India . . . 
Collected by Barnum Brown in 1922 at summit of ]>ower Chinji 
horizon, 2,000 feet above base of Lower Siwaliks. Middle Miocene 
[Mio-Pliocene]." Type Figvre.— Op. cit., 1929.797, ]x 14, 

fig. 15. See also figures 725 to 727 of present Memoir. 

Type Characters. — Superior tusks laterally compressed. 




STEqOLOPHODON NATHOTENSIS 



^j ouier vieiv^' i, , °-"'^ icppe 





Fig. 724. Type of Stegolophodon nalhotensis Osborn, 1929 (Amor. Mils. 19455), Lower Chinji horizon, from near Nathot, India. One-half natural size. 

A third superior molar of the right side, r.M'*, probably with four and a half ridge-crests. Meas.: ap. 180e mm., tr. 88e mm.; height of nietalo|)li .38e mm.; 
I.M^ (protoloph) tr. 73 ram., (metaloph) 36e mm.; r.Ms, tetartolophid .and jjentalophid (talon only). Ridge-crest of tetartoloph of r.M' with five arcuate 
conelets. Mjussive, brachyodont. Osborn, 1929.797, p. 14, fig. 14. 



Stegolophodon cautleyi progressus Osborn, 1929 

Figures 725-727, PI. xiii 
Summit of Lower Chinji horizon, 2,000 feet above base of Lower Siwaliks, 
twelve miles east of Chinji Bungalow, India; Mio-Pliocene. 

This type is of great importance and interest as yielding for 
the first time a knowledge of the cranial structure of Stegolophodon 
in its Miocene stage of evolution. Distinguished as Stegolophodon 
by its four-crested intermediate molars, its superior tusks, and its 
cranial profile, which remind us strongly of the primitive type of 
skull and tusks seen in Trilophodon and in Serridentijius; the four- 
crested intermediate molars parallel those observed in Telralopho- 
don, a genus from which the present Stegolophodon is shar])ly re- 
moved by the absence of trefoils, vestiges of which ai)i)car only in 
the valleys between the ridge-crests. Altogether Stegolophodon 
cautleyi progressus (probably of the summit of the Miocene or 
lower levels of the Pliocene) is a progressive ascending mutation 
of the typical Stegolophodon cautleyi of Perim Island, which we 
regard as of Middle Pliocene age. 

Stegolophodon cautleyi progressus Osborn, 1929. "New 
Eurasiatic and American Proboscideans," Amer. Mus. Novitates, 

'[See notes on pages 814 and 824 above. — Editor.! 



with broad external enamel band ; ridge-crest formula as compared 
with that of Stegolophodon cautleyi, as follows: 

Stegolophodon cautleyi progressus: r.Dp 3^ r.Dp4 — - r.Ml^^^^ 



)i-3-ii 



Stegoloj}hodon cautleyi (typical): r. Dp 3 — r. Dp 4 
r.M 1^^ 

Four ridge-crests in intermediate molars, r.Dl)^ r.M', with 
four irregular conelets on each crest; rudimentary anterior and 
posterior ridge-crests in r.M'; conelets less blunt and crowns less 
brachyodont than in Stegolophodon nalhotensis; median fissure in 
r.M' wanting, as shown in comparison of figure 726 with figure 714, 
cotype of Stegolophodon cautleyi, or decidedly less distinct than in 
S. cautleyi lectotype from Perim Island (Fig. 712) and the referred 
l.M- (Fig. 715); traces of irregular internal trefoil conelets on 
r.Dp< and r.M'. 

Of very great importance and interest is the i^re.sence of 
superior incisive tusks with lateral enamel band as well as other 
evidence in the structure of the superior grinding teeth of the 
affinities of this tetraloijliodont tyjie with the much more i)riniiti\e 
trilophodont mastodonts of the Lower Miocene of France (cf. 
also Figs. 725-727). 




STEQOLOPHODON CAUTLEYI PROqRESSUS 



Amer AfU5. I9^^b '^yP^ 



~ 7ia.t. sije 




Fig. 725. Type cranium of Stegolophodon cautkyi progressus Osborn, 1929 (Amer. Mus. 19446), same specimen as figure 727. 
Al, Superior aspect, exhibiting cranium with anterior narial opening and prcmaxillaries. 

A2, Palatal view, exhibiting lateral enamel band of tusks, much worn Dp' with remnants of 3+ ridge-crests, Dp^ with 4+ ridge-crests, M' with 4+ ridge- 
crests, also posterior narial opening and basicranial foramina. 
A3, Right lateral aspect, exhibiting enamel band on tusk. 
.\4, Occii)ital aspect, slightly compressed by crushing. 



Amer Mus /9^^6 T'ype 




i nat szje %^ 



Fig. 72(5. Type of SlignlophoJon caiilkiji progressus Osborn, 1929. Detailed study of right 
sui)erior lientition, Dp'-M', combination drawing of two sides. Coronal ridge-crests intermediate in 
formula and pattern between the typical Mastodon [=Slrgolopho<hm] cautleyi Lydekker and M. 
1 = SlegohpfiodoH] lalidins Clift; similar in the regular disposition of the cones and conelcts. 



849 




Fig. 727. Type skull (Amor. Mus. 19440) of Slcgohphodon catd- 
Iciji progreasus Osborn, 1929, collected by Banuini Brdwii twelve miles 
east of Chinji Bungalow, summit of the Lower Cliinji horizon, 2,000 
feet above tiie base of the Lower Siwaliks, India. This skull, containing 
the right and left superior tusks, third and fourth superior deciduous 
premolars, and the fii'st superior molars (partly or\ipted), shows the 
laterally compressed tusks with enamel band, 4+ ridgc-crests on 
Dp"*, 4+ ridge-crests on M'; it is somewhat more progressive than 
the type of Mastodon [ = Stcgolophodon] caulleyi Lydekker. 

Left lateral, palatal, and right lateral aspects, one-fifth natural size. 



850 



STEGOLOPHODON 





ExtLTnal 
S STECODONTOIDES TYPE 



S.LATIDENS TYPE 




Extfrmxt 
S. CAUTLEYI TYPE 



H 



S.LATIDENS after CUft 





5 CAUTLEYI PT^QOPIESSUS TYPE 




S.SUBLATIDENS TYPE 



S.NATH0TEN51S 



PLATE XIII 

Stegolophodontin.e, Stegolophodon: Primitive Longitudinal Sulcus, Shaded (A, B), Persisting in I-IV Anterior Ridge-crests (E, El), 

IN I-III Anterior Ridge-crests (F), Vestigial in Ridge-crests I-III (G, H, I). Conelets Rounded, Increasing from 4-5 (A, D), 

5-7 (C, H, I, G). Posterior Ridge-crests IV-VI Progressive, with Conelets 5-7 (G), Conelets not Exceeding 5 (H, I) 
Mio-Pliocene B, Stegolophodon cautleyi progressus type, r.M'. Amer. Mus. 19446, near Chinji Bungalow, India. Summit of Lower Cliiiiji 
horizon, 2,000 feet above base of Lower Siwaliks. 
C, Stegolophodon natholensis type, r.Ms. Conelets 5-6. Amor. Mas. 19455, near Nathot, Lower Middle Siwaliks, India. Lower 
Chinji horizon. 
Lower Pliocene' D, Stegolophodon latidens cotype, r.Ms. Ridge-erests I-VIII, conelets 4, sulcus on ridge-crests I-IV. After Clift, 1828, 
PI. XXXVIII, fig. 1, near Yenangyaung, Burma. 

E, Stegolophodon latidens cotype, r.Ml Ridge-crests I-VI, sulcus on ridge-crests I-IV. E 1 (.section), ridge-crests coalescent 

at base (I-III). After cast (Amer. Mus. 21978) of Cliffs type, 1828, PI. xxxvii, fig. 1, near Yenangyaung, Burma. 

F, Stegolophodon latidens ref., r.M'. Sulcus on ridge-crests I-III, conelets 4-6, ridge-crests 4)2- After Matsumoto, 1926.1, PI. 

V, figs. 1 and 3 (Prostegodon), Sliiwogama, Miyagi District, Province of Rikuzen, Japan. 
Pliocene (?) G, Stegolophodon hjdekkeri type, l.M'. Ridge-crests I-VI, conelets 4-7, sulcus on ridge-crests I and II only. After Lydekker, 

1886.2, fig. 19 (as M. latidens), Borneo. 
Middle(?) Pliocene A, Stegolophodon sublatidens typo, r.Ml Conelets 4-5. After Schlesinger, 1917, Taf. xvii, fig. 2, Teschen (Schlesien), Austria. 
Middle Pliocene H, Stegolophodon cautleyi loetotype, I.M^. Ridge-crests I-V, sulcus on ridge-crests I and II, conelets 5. After Lydekker, 

1886.1, p. XV, fig. 6, Perim Island, India 
Upper(?) Pliucone I, Stegolophodon stegodontoides Pilgrim, type, r.M'. Ridge-crests I-VI, sulcus on ridge-crests I-III, conelets 5. After Lydekker, 

1880, PI. XXXIX, Lehri, Punjab, India, possibly Upper Siwaliks. 
CoMPAR,\TivE Observations (1935) 
Molars of Stegolophodon are readily distinguished from those of the Mastodon, Zygolophodon, and Turicius phyla by the following characters: 
First, by the persistence of the median stdcns separating the inner and outer pairs of cones of all the crests (A, B, D), of the three to four anterior 
crests (E, F, I), of the two anterior crests (G, H). 

Second, by the rounded, bunoid conelets separated by median sulcus (AH). 

Third, by the closure of the enamel in the base of the transverse valleys, as seen in section (El), very characteristic of Stegodon. 
The second and third of the Stegolophodon characters enumerated above link this genus with the genus Stegodon. But we must remember that 
Stegolophodon cautleyi is of Middle Pliocene age (Perim Island), contemporary with the true Middle Pliocene Stegodon bombifrons (Dhok Pathan). 



'[See note on page 824 above regarding the Lower Pleistocene rather than Lower Pliocene age of Stegolophodon latidens. — Editor.] 



THE STEGOLOPHODONTIN^: STEGOLOPHODON 



851 



(n'ideuce in the structure of the superior grinding teeth of the 
affinities of this tetralophodont type with the much more primitive 
trilophodont mastodonts of the Lower Miocene of France (cf. also 
Figs. 725 and 727). 

Stegolophodon lydekkeri Usborn, 193G 
Figure 728, PI. xin 
riom vicinity of Bruni, northwest coa.st of Borneo. 
I In an intensive reexamination of \arious species of Stegolo- 
jjhodon, Professor Osborn discovered wide differences between tire 

Fig. 19. 




Mastodvii latidens. — The third left upper true molar of a small iiuliviclual in a 
partially-worn condition ; from the Pliocene (?) of Borneo. J. The lower 
border of the figure is the inner border of the specimen. (From the 
' Palasontologia Indica.') 

Fig. 728. Stegolophodon lydekkeri Usljorn, 1936, Vohime I, p. 700, fig. 
(itjO, of the present Memoir. Third left superior mohir exhibiting six ridge- 
i-rests and a talon. Figured by Lydekker, 1886.2, fig. 19, as Mastodon lalidens. 
Compare Lydekker, 188o.2, PI. xLviii, also PI. xni of the present Memoir. 
Two-thirds natiwal size. 



tyjies of Stegolophodon sublatide/is Schlesinger, 1917 (PI. xiii, A) 
and S. lalidens Glift, 1828 (PI. xiii, E) and a molar tooth from 
Borneo described and figured by Lydekker as 'Mastodon lalidens' 
Clift (1885.2, 1886.1, 1886.2). This molar Professor Osborn 
selected as the type of a new Pliocene species, namely, Stegolo- 
phodon lydekkeri (Vol. I, p. 700). 

Type. — Third superior molar of the left side, l.JVP, original 
in the Zoological Society of London; cast in the British Museum 
(M.2498). Horizon and Locality.— (Lydekker, 1885.2, p. 

777): "The specimen forming the subject of the present notice 
was forwarded from Borneo to the Secretary of this Society by 
Mr. A. H. Everett, C.M.Z.S., who stated that it was found during 
the early part of the present year [1885] by a Kadayan in the 
jungle in the vicinity of Bruni, on the north-west coast of Borneo." 
Pliocene (?). Type Figure.— Lydekker, 1885.2, PI. xlviii, 

figs. 1 and 2; .see also Lydekker, 1886.1, fig. 7, 1886.2, fig. 19, as 
well as PI. xiii and figures 660 and 728 of the present Memoir. 

Description. — (Lydekker, 1885.2, pp. 777-779) : "The speci- 
men is the crown of the last left upper true molar of a tetralopho- 
dont Mastodon, and agrees so closely with Indian teeth of the 
Siwalik Mastodon lalidens, Clift [Footnote: 'Trans. Geol. Soc. 
ser. 2, vol. ii, pt. 3, p. 371 (1828).'], that it may be safely referred 
to that species, although it indicates a very small individual. . . . 
It will be seen that the Borneo specimen agrees [with .1/. lalidens 
of the Punjab — made by Pilgrim in 1913 the type of Mastodon 
stegodont aides] in the number of ridges (although the hind talon 
is considerably smaller), but is of greatly inferior size, the dimen- 
sions of the two specimens being as follows, in inches: — 

Punjab Borneo. 

Extreme length 8.6 6.3 

Width of first ridge 4.2 2.95"] 




Fig. 729. Falconer's map of the geology of India (see "PahEontological Memoirs," 
1868, Vol. I, description of PI. ii). The red fossiliferous areas, here colored black, are de- 
scribed by Falconer as follows; 

"The red [black] stripe represents the Sewalik Hills, stretching from the Hydaspes to 
the Gundiick River, 800 miles. The small red [black] patch behind the Himalayahs repre- 
sents the ossiferous plain of Tibet about 16,000 feet above the sea. The other red [black] 
patches represent the Nerbudda [Pleistocene], and the [black dot within the circle the] Gulf 
of Cambay [Pcrim Island-Middle Pliocene] fos.sil tracts." 

The following is Falconer's interpretation of the geology of India: 

"The great mass of light shadin/] represents the sujjposed insular form of the contineni 
of India at an early period of the Tertiary epoch, the island forming a sort of triangle, of 
which the eastern and western Ghats formed the sides and the great Vindhya range the base, 
with an irregular patch of mountainous country stretching north forming the Aravalli range." 

"The dark shading rejjresents the ))lains of India, forming the valley systems of the 
Ganges and Indus drainage, which were formerly narrow ocean straits. These straits were 
the n'cipients of the silt and alluviiun washed out of the Himalayahs, and were at length 
elevated above the ,sea, .so as to form th(^ existing continent. The Sewalik Fauna (hen spread 
over the continent, from the mouth of the Irrawaddi to the Gulf of Cambay 2,000 miles, 
and north to the .Iheluin 1,.')00 miles, .\fter the long establishim'nt of the Sewalik I'auna, 
a great upheavemcnt took place along the line of the Himalayahs, elevating a narrow bell of 
the plains into the Sewalik Hills, and adding many thousand feet to the height of the 
Himalayahs." 



852 



IV. SUCCESSION OF SPECIES OF THE GENUS STEGODON 



Fig. 730. After Osboni, 1910.346, i). 323, fig. 154. Coniparc figuics 72i) ;iihI 820 
of present Memoir. 



Superfamily: STEGODONTOIDEA Osborn, 1935, 1936 
Family: STE(iODONTID^ Young-Hop wood, 1935 
Subfamily: Stegodontin^ Osborn, 1918, 1921 

Genus: STEGODON Falconer and Cautley, 1847, 1857 

Original reference: Falconer and Cautley, "Fauna Antiqua Sivalensis," 1846 [1847, PL xlii]; also Falconer, Quart. Journ. 
Geol. Soc. London, 1857, Vol. XIII, pp. 314, 318, and Synoptical Table opposite page 319. 
Genotypic species: Elephas Cliflii, E. bombifrons, E. IGanesa, E. insignis} 
Syn.: i^j/iTMenodott Cope, 1889; Parasfejodon JNIatsumoto, 1924 (in part). 

Generic Definition. — (Falconer, 1857, p. 318): "Dentiuni molarium 3 utrinque intermedioruni 
eoronis complicata colliculis hypisonieris {e.g. 7 + 7 + 8), mammillatis, tectiformibus. Praemolares non- 
dum observati." 

Generic Characters. — Ridge-crests intermediate between Stegolophodon and Archidiskodon 
planifrons types, progressively multiplying from six to eleven in the intermediate molars, from nine to 
fifteen and a half in the posterior molars. Cones rapidly subdividing by binary or ternary fission into 
multiple conelets. Ridge-crests elevating from brachyodont {Stegodon sinensis) to subhypsodont {S. 
airatrdna and S. aurorx stages). Cement developing in the valleys. Crania of mastodontoid {S. 
bombifrons) to extremely abbreviated, female? {S. insignis), more elongated, male? (S. ganesa), more 
triangular (*S. trigonocephalus) form. Tusks attaining" great dimensions {S. ganesa). Phylum parallel 
to that of the true Archidiskodon and Elephas, not directly ancestral, readily distinguished by cranial 
and dental characters. 

The generic name Stegodon was first prnited 
in tlie "Fauna Antiqua Sivalensis" of Falconer 
and Cautley, 1846 [1847, PI. xlii]. It was ten 
years later that it appeared as a subgenus (Fal- 
coner, 1857, pp. 314, 318, table opp. p. 319) to 
include the following species, Elephas cliftii, E. 
bombifrons, E. ganesa{?), and E. insignis, and 
has subsequently been used as a genus by all the 
principal authors, except Lydekker who clung 
to the Cuvierian division of the mastodonts and 
elephants into the two genera, Mastodon and 
Elephas. Pohlig in 1888, p. 252, wrote the genus 
in this way: Stego(lopho)don. Schlesinger, 1917, 
p. 115, separated the species M. latidens as the 
type of a new genus Stegolophodon. 

Stegodon (omitting the six^ species described 

above under Stegolophodon) is readily definable _ chief Miocene and PUocene fossU mammal depoaits of Asia. 1. Maragha, 

ind rpfldilv di^tintriii^liihlp from «inv o-pnii'^ nf ^<'^^'^- 2. Perim Island. 3. Manchhar Beds of Sind. 4. Siwaliks of the Punjab. 5. Sub- 

auu ILdUay UlSliagUlSnaUie liom any genus Ul himalayan Siwallks (River Brahmaputra to River Jhelum). 6. Valley of the Lower Irawadi, 

thp Mn**tnflnntirlfP on tliP nnp linnrl or nf thp Burma. 7. Miocene and Pliocene deposits of China (Provinces of Shan-si, Shen-si, Sze-chuan, 

ine iViaSlOUOmiUcB, on me one nana, or or tne Kwang-Tung, Ho-nan, Hu-nan, Hu-peh). 8. Miocene and PUocene deposits of Japan. 

Elephantidae, on the other (see Vol. I, p. 25), by 

the brachyodonty to subhypsodonty of its ridge-crests, in contrast to the hypsodonty in all the species of the 

Elephas and Loxodonta phyla. The relative height attained is shown in figures 687 and 688. While the ridge- 

HHopwood in his Memoir of 1935 on the "Fossil Proboscidea from China," p. 72, remarks: "According to the modern school of priority-purists this 
necessitates the selection of E. cliflii as genotyjie as a matter of course. Fortunately, we can do so and adhere to the author's original intention, for, in his 
later writings Falconer expressly says that Stegodon corresponds 'with the forms collectively designated Mastodon elcphantoides by Cliff, (1857, p. 314)." 
—Editor.] 

-[To these six species should be added the Stegolophodon lydekkeri Osborn, described in Volume I of the present Memoir (p. 700). — Editor.] 

853 




854 



OSBORN: THE PROBOSCIDEA 



crests are composed exactly as we know that we shall find them hi the direct ancestors of Elephas, they 
remain short crowned, even in the progressive species Stegodon insignis and ;S. ganesa, and in the still more pro- 
gressive S. orientaUs and .S. airdwana. This is undoubtedly an adaptation to browsing on leaves and softer 
kinds of food, which leads us to believe that the Stegodonts were persistent browsers rather than grazers, as in 
all the i^liyla of Elcplnts and of Loxodonta. 



Table V. Si>i;cn;s ix Approximate Ascending Order of Collective Maximum and Minimim l\iiKii:-ci(EsT« 
l'"alconer (1868), Lydekker (1886), .Martin (1890), ^Nlatsumoto (1918), Osborn (1929) 





Dp 2 


Dp 3 


Dp 4 


M 1 


M2 


M3 


STEriOLdPHonox 

caNlleyi 






W-3-« 


',«-4-« 


4 


5-» + 


cautleyi progressus 




2. 


4-JS 


M-4-VS 






latidens 




S. 


i. 


W-4-V'. 


4M-5 
4M-5 


M-5-M 
M-6-1S 


lydekkeri 












6 + 


sublatidens 












?6)4 


slegodontoides 












e-K 


STEGODON 
sinensis 




tS-4 










elephanlaides {^cliflii, fide Falc.) 




i. 


5-6 


6-6 W 


6» 


8-M 


elephantoides Clift 








6W 


HV, 


nj 


homhifrons 




i 


5-6 


6 

7-714 


6-7 


7»-8-8W-9-9M 
B-9-9» 


)«-5-VS-6VS-7 


7-7W-8 


irigonocephalus 




4- 


4- 








., 1 

mindanensis {Archidiskodon?) Fragments, farmc 


)re progressiv 


e in the directi 


on oi Archidisl 


1 

•Mdon than either insignis or ganesa. 


insignis r 


5-C 



7li 

7M-9 


6W-7-7K-8 
4-7-7W-1 


7-VS-7-M-W-8 
7-8-M-0-V4-?12 


0-1 0-1 1-1 HHS-ll-W 


9-1 0-W-l 1-iS-l 2-1 214-1 3 


ganesa 


^ 


5, 


7 
Hi 


6V4-7 
5>4-7Vi 


7-7 M 
7 W- 8 


fll5-l 0-W-l 0-14 
714+-8-W-8-1S 


pinjorensis 












1 414-1 3 


insignis birmanicus 












12-M 


orienlalis grangeri 


M-2-H 
2 


5-M 
+ 5M 


54-6-W 
7-W 


VS-6-VS 
8 


M-8-V* 
(7)-9-» 


14-1 1-14 
14-13 


aurorx 










10« 




airdwana 


S. 


r,-e 


7-0 


7-0 


0-11 


12-1 i 
1 3-15-1'j 



THE STEGODONTIN^: STEGODON 



855 



The order of ridge-crest addition and development (Table V) corresponds approximately with the phylogenetic 
and geologic ascending order as shown in Table IV above of the present chapter, in which the species are grouped 
partly by geographic distribution. The ridge-crest formulae of certain species, e.g., Stegodon sinensis, S. ganesa 
javanicus, S. orientalis, and S. orientalis shodoensis, are too imperfectly known to determine precisely their phylo- 
genetic position. In general, S. sinensis appears to be the most primitive, while S. airdwana appears to be the 
most progressive and geologically recent. 




STEGODON GANE5A 3050min..lO'e 

INDIA SIWALIKS 



STEGODON PINJORENSIS 2745mm.. 9 e 

INDIA SIWALIKS 





STEGODOrJ INSlCrjlS I925mni.,0'4''8"6 

INDIA SIWALIKS 



STEGODON ORANGERI 

CHINA 



225&i;in;.,7'5"e 





STEGODON TR1G0N0CEPHA,LUS I428mm„4'8>i"e 

JAVA 



^m 



STEGODON BOMBIFRONS 2148 mm.. 7'>t"e 
INDIA SIWALIKS 



STEGODON AIRAWANA 
JAVA 




76min..3'lO/3"e 



Fig. 731. Species of Stegodon from India, Chin.\, and Java 
Restorations by Maboret Flinsch, under the direction of Henry Fairfield Osborn, One oNE-HU>fDRBDTH Natur.^l Size 



Skulls of Stegodon bombifrons, S. insignis, and S. ganesa in the British and Indian Museums 
Before considering in detail the succeeding species of Stegodon, it is necessary to examine and compare the 
known crania of various species with each other, as assembled in figure 732 from Falconer and Cautley's beautiful 
plates, with the crania of other Stegodonts from the East Indies, e.g., Stegodon airdwana (Fig. 773) and .S. trigono- 




ELtPHAS eOMBIFRONS B*», 
PI. XLIII, XIII 



ELEPHAS GANESA R»f. 
Pi, XXir, F.Q. 2 



I'lK- 732. The tlircr sprrios of Strgodonts of wliirh (ho cninia wcrt' known to Falconer iiro illiislratcd in tliis comparative plate in 
such a manner tliat their distinctive eliaraeters may be readily contrasted. TlirouKliout the Stegmlon crania are of relatively small size, 
laekiiiK the cancellate structure characteristic of EUphas. All one-twentieth natural size. Compare figure 777. 

Elii>lms [ = Slrgodon] bombifrons, a Middle [to Ujiper] Flioceni- stage of evolution, is illustrated in the lower line. [The cranium of E. 
Immbijnms at extreme right should read: PI. xr,v, xui (rev.).| 

Elrphns [=Strgo<ltm] in.iignis, a Ixnvpr to Middle |Up|)er] Pleistocene stage of evolution, is illustrated in the middle of the diagram. 

Elephas [ =Stcgodon\ gancsa, a Lower to Middle [Upper] Pleistocene stage of evolution, is illustrated in the top lines. It is extraordi- 
nary that tusks of such enormous length and width should be supported by a cranium of such small size. The extremities of the 
tusks should probably be turned inward, instead of outward as figured by Falconer and Cautley; the left tusk is complete; a portion 
of the right tusk is omitted in this drawing. [Compare figure 733 for revised ri'storation of the tusks.) 

856 



THE STEGODONTIN^: STEGODON 



85 



cephalus (Fig. 776) of Java, also with the newly found *S'. orientaUs grangeri (Fig. 763) of China. These Stegodont 
crania should be compared with the large craiiiuni of AirliidiHkddoii pUutiJron^ (Fig. 830), one of the most primitive 
of the true elephants. 

In general the cranium of Stegodon bombifrons (Fig. 732, bottom row, also Fig. 742 and Fig. 744) is more 
generalized and subelephantine in character, whereas the crania of Stegodon ganesa and S. insignis are very highly 
specialized, of relatively small size, and bear little resemblance to the crania of the true Elephantidae. Quite 
different are the small, triangular crania of Stegodon trigonocephalus and S. airdunna of the East Indies. 

Sexual Divergence very Marked. — The referred giant skull of Stegodon ganesa obviously belongs to 
a full-grown male Stegodont (as shown in fom- as]>ects in Fig. 732). On the contrary, the crania referred to S. 




Fig. 733. Cianium of Elephas [Stegodon] ganesa, after skull referred to this species by Falconer. Drawings made from original plates in Falconer and 
Cautley's "Fauna Antiqua Sivalensis" reduced to one-sixteenth natural size. See figure 732 opposite for frontal, lateral, and palatal views, reproduced to a one- 
twentieth scale. The frontal profile is seen to differ profoundly from that of Stegodon insignis and less profoundly from that of S. bombifrons. 

In an attempted restoration of Stegodon ganesa from the tusks as originally represented by Falconer and Cautley, it was found impossible to lower the 
proboscis between the closely appressed tnsks; it was also observed that the extremities of the tusks turned outward, unlike all other proboscideans. Accordingly 
on receipt of the gift of the cast from the British Museum (Jan. 23, 1931) the tusks were readjustctl in a position with the extremities turned inward, allowing 
a sufficient space for the descent of the proboscis between the butts. The present figure rcjirescnts the specimen according to this conception of the position 
of the tusks, in contrast (Fig. 732) with Falconer and Cautley's restoration. 

[The most painstaking examination of the original sijecimen by Professor Osborn, Doctor Gregory, and Doctor Colbert failed to reveal any evidence 
cither tliat the tusks have been transposed or tliat they are twisted in the alveolus by ijost-mortem changes. Nevertheless all felt that the space between them 
is insufficient for the trunk as restored by Falconer and Cautley and that the whole arrangement looks abnormal. — Editor.] 



858 



OSBORN: THE PROBOSCIDEA 



insignis (Fig. 732) obviously belong to small-tusked females, because the alveolar processes for the insertion of the 
tusks are extremely small ami narrow. 

If, as seems possible, all the crania referred by Falconer and Cautley to Elephas [ = Stegodon] insignis are 
females, and the great cranium referred to Elephas [ = Stegodon] ganesa is a male, there is in Stegodon a far greater 
sexual disparity and difference than prevails between the female and male crania of either Elephas indicus or 
Loxodonta africana, as figured below in the present Memoir. If this be true, the sexual disparity in cranial charac- 
ters constitutes an important specific distinction of Stegodon insignis and ;S. ganesa. 

In the present Memoir we treat the two species separately but agree with the theory suggested by more than 
one author, especially Lydekker and Matsumoto, that the crania of S. insignis represent the females of the same 
collective species as the referred male cranium of S. ganesa. 

British Museum, W. D. Matthew, September, 1920. — The skull of Stegodon bombifrom is essentially 
elephantine, the shortness of the enamel plates being the chief distinction; the skull is apparently shorter than in 
Loxodonta; the jaw is deeper; the symphyseal process may be a little heavier. In contrast, the skull of S. 
insignis has a supranarial region with a great thickening of the cellular tissue which appears to round back into 
the occiput, the occipital crest being very little developed; a very curiously shaped head and very small tusks. 

Characters of Referred Skulls of Indian Stegodonts, x 

after Falconer, 1868 -■ - 

(See Figs. 732-736, 754, 777) 

Skull of Elephas [ = Stegodon] bombifrons (Falconer, 1868, Vol. I, p. 
458, PI. xxvii): "Very fine and perfect skull, anterior view. Four other 
views of same skull are given in Plate xxviii. This head is very marked; 
it is convex from occiput to front and also across, and is very narrow 
at the temporal contraction. The bounding ridges sweep round by a 






a Generalized Cranium 
Fig. 734. Elephas [= Stegodon] bombi- 
frons. Front view of cranium, oiic-twclfth 
natural size. Brit. Mus. M.2979; cast 
.■Vmer. Mus. 1037S. .\fter Falconer and 
Cautley, 1840 (1847, PI. xxvii]. 



A Specialized Cranium 
Fig. 735. Elephas [=Slegott(m\ insig- 
nis, one-twelfth natural size. After 
Falconer and Cautley, 184611847, PI. xv). 



A Specialized Male Cranium 
Fig. 736. Re.storation: Elephas 

[= Stegodon] gnnesn, one-twelfth natural 
.size. Brit. Mus. M.3008. After Falconer 
and Cautley, 1846 [1847, PI. xxiri]. 



THE STEGODONTIN^: STEGODON 859 

l)old curve into the post-orbitary processes, as in E. [ = Archidiskodon] meridionalis. There is a deep furrow 
between the tusks. The nasal opening for the trunk is above the Hne (or nearly so) of the post-orbitary processes 
of the frontal bone. Above the infra-orbitary foramen on the right side there is another smaller opening." 

Skulls of Elephas [ = Stegodon\ insignis (Falconer, 1868, Vol. I, p. 448, PI. xv) : "This is the most remarkable 
of all the Indian fossil Elephants. The cranium is as singular and grotesque in construction as that of the Dino- 
fherium giganteum. The cranium is seen to differ remarkably from that of E. Ganesa (Plates xxi. and xxii.) 
notwithstanding that the molars of the two species agree so closely. That of E. insignis is flattened at the top, 
elongated from side to side and singularly modified, so as to bear an analogy to the cranium of Dinotherium 
giganteum, while that of E. Ganesa does not differ much from the ordinary type of the Elephants." 

(Op. cit., p. 449, PI. XVI, fig. 1) : "This head is very cubical in form, is old, very concave in front and vertically; 
teeth broken. Interval between incisive sheaths deep. No tusks." 

{Op. cit., p. 449, PI. XVII, figs. 3 and 4): "Anterior and lateral view of another cranium. Both zygomatic 
arches are missing, and the left side of the cranium is deficient. Shows the great length of the incisive sheaths." 

Skulls oi Elephas [ = Stegodon] ganesa (Falconer, 1868, Vol. I, p. 453, PL xxi): "Large skull, with fragment 
of left incisive in situ, and corresponding fragment of right incisive detached. The incisive alveoli are remark- 
ably elongated, as in E. primigenius. The plane of the incisives is continuous with that of the frontal, but with a 
tendency to obliquity forwards. The skull is very imperfect on right side." 

{Op. cit., p. 454, PL xxii) : "Fig. 1.— Elephas Ganesa. Lateral view of large skull figured in PL xxi. — B.M. 
Fig. 2. —E. Ganesa. Palate view of same skull. The right incisive is seen in section. The posterior true molar 
is seen on either side of palate. It has ten plates and a heel behind, and a small talon in front; the hind heel has 
few denticles; the foiu' front ridges are worn. The alveoli are parallel as in the Mammoth. — B.M. Fig. 3. — 
E. Ganesa. Sketch showing restoration of skull, with tusks, of iJ.Gawesa, profile view, one-thirteenth of natural size." 

{Op. cit., p. 454, PL xxiii): Restoration. "Sketch showing restoration of skull, with tusks, of E. Ganesa, 
oblique antero-lateral view, one-thirteenth of natural size." 

{Op. cit., p. 455, PL XXIV. A. ) : "Figs. 1 and la. — Elephas Ganesa. Fragment of skull with palate and back 
molars on both sides. This is a most remarkable specimen. I have called it E. Ganesa (H. F.), and it much 
resembles the molar of the big Ganesa specimen (Plate xxii. fig. 2) in form and in the compression of the ridges, 
but the ridges are few. . . . B.M." 

Stegodont Crania of China and of the East Indies 
(See Figs. 763, 776, 773, 777) 

A most fortunate discovery from the pits along the Yangtze River near Wanhsien, Province of Szechuan, 
China, is the extensive series of Stegodont crania in all stages of development, infantile, juvenile, young adult, 
and mature adult. This priceless collection made by Walter Granger of the Central Asiatic Expedition of the 
American Museum illustrates in a perfect manner (Figs. 759, 761-763, 686) the complete ontogeny, dental 
succession, and metamorphosis of the cranial form. The young adult and mature adult crania, together with 
the inferior mandible (Fig. 763), apparently resemble the cranium of S. bombifrons more closely than the crania 
of S. insignis and S. ganesa. This material was described by the present author as belonging to a new subspecies, 
namely, Stegodon orientalis grangeri. 

The cranium of the East Indian species Stegodon airdwana (Fig. 773) of Java belongs to a very advanced 
mid-Pleistocene stage, resembling *S'. bombifrons and *S. orientalis grangeri in profile but differing in the frontal 
aspect, which is flattened. The cranium of S. trigonocephalus (Fig. 776), as its specific name indicates, has a 



860 



OSBORN: THE PROBOSCIDEA 



triangular rather than a rounded superior profile, and, althougli of the same geologic age as S. insignis and *S. 
ganesa, it is entirely different from the Siwalik species in its profile and proportions. Consequently it appears 
that both in cranial and dental characters the East Indian species represent a distinct and somewhat dwarfed 
side-brancli of tlie northerly continental species. 

SYSTEAIATIC DESCRIPTION OF SPECIES OF STl'^GODON 



Stegodon sinensis Owen, ISTO 

Figures 087, 702, 737 

Alleged to be "from marly beds in the vieinity of Shangiiai," China. 
Tliis animal (Figs. 687, 702) is probably of Upper Miocene [Lower Pliocene) 
age, for it is more primitive than Slegodun bombi/rons, and possibly ancestral 
to the S. bomhijrons stage. 

Specific Characters. — This little-known true Stegodont, 
Stegodon sinensis (Dp 3 ^^), is somewhat more primitive than 
S. homhifroiis (Dp 3-), since the ridge-crests are less elevated or 
hypsodont as compared (Figs. 687, 688) with S. bombifrons, 
S. orienialis graiigeri, and S. insignis. Estimated number of cone- 
lets on fourth ridge thirteen to fifteen. Falconer, Lydekker, and 
Martin ascribe (see Table \ above) four ridge-crests to Dp' of S. 
elephantoides { = cliftii), S. bombifrons, and S. trigonocephalus. 

HiSTOHY. — (1) After comjiaring the type deciduous molar 
with all the Siwalik specimens in the British Museum, Owen 
concluded (1870, p. 420) that the above Chinese tooth was most 




parison is strengthened by the fact that Koken (1885) referred 
the Stegodon sinensis of Owen to the species Stegodon rliflii of 
Falconer. 

(3) Lydekker (1886.2, p. 79) also referred S. sinensis Owen to 
Elephas cliftii Falc. {Op. cit., p. 80) : Brit. Mus. 41925. "A third 
right upper milk-molar, provisionally referred to this species; from 
the Pliocene near Shanghai, China. This specimen is the type of 
Stegodon sineiisis, and is described and figured under that name by 
Owen in the Quart. Journ. (!eol. Soc. vol. xxvi. p. 417, pi. xxvii.; 
it is also figured and provisionally referred to the present species 
[Elephas cliftii] by the writer, in the 'Palseontologia Indica,' ser. 
10, vol. i. p. 257, pi. xlv. fig. 2. There are four complete ridges, and 
a large anterior talon, which is reckoned by Owen as a fifth ridge. 
The median longitudinal cleft is very indistinct. Presented by 
Prof. Sir R. Ou'en, K.C.B., 1870." 

(4) Osborn (1924) prefers to retain the name Stegodon sinensis 
until further local material can be secured for comparison. 




Fig. 737. Type of Stegodon sinensis Owen, 1870, PI. xxvil, figs. 1 and 2, natural size. Alleged to be from "marly 
beds" near Shanghai, China. (Op. cit., p. 417) : "The tooth in question is the second upper molar ((/3 of the type series) 
from the right side. Its crown, in a length of three inches, is divided into five transverse ridges." 

Falconer, Lydekker, and Martin (Taljle V above) ascribe four ridge-crests to Dp' of Stegodon elephantoides 
(= cliftii), S. bombifrons, and S. trigonocephalus. 



closely related to undetermined Siwalik specimens which he com- 
pared with the M. elephantoides of Clift. These two undeter- 
mined specimens are described by Falconer (1868, Vol. I, p. 460) 
as follows: Plate xxix. a, "Fig. 5. — E. bombifron.sl Fragment of 
molar, from lower jaw, right side, with four ridges. — B.M. Length, 
5.8 in. Width, 4.5 in. Fig. 6. — E. bojtibifronsl Fragment of molar 
with three ridges and a heel. 'Doubtful what figs. 5 and 6 are.' — 
H. F. Length, 4.4. in. Width, 4.5 in." This shows that Falconer 
regarded these Siwalik teeth as doubtfully related to E. bombifrons 
and that Owen also indirectly comjjared f hem with E. bombifrons. 
Stegodon bombifrons lectotype is of Middle Pliocene age. 

(2) One should also compare Owen's type tooth with figuie 700 
above of Mastodon [Stegodon] elephantoides (=cliftii}. Tiiis coni- 



Stegodon sinensis Ow-en, 1870. "On Fossil Remains of 
Mammals found in China." Quart. .lourn. Geol. Soc. London, 
Vol. XXVI, p. 417. Type.— (Op. cit., p. 417): ". . . Second 

upper molar {d 3 of the type series) from the right side [Dp']." 
Brit. Mus. 41925. Horizon and Locality. — Alleged to be 

"from marly beds in the vicinity of Shanghai," China. Probably 
tapper Miocene [Lower? Pliocene]. Type Figure. — Op. cit., 

PI. xxvii, figs. 1-3. 

Type Description. — (Owen, op. cit., pp. 417, 418) : "Stegodon 
sinensis, Ow. The tooth in question is the second upper molar 
(d 3 of the type series) from the right side. Its crown, in a length of 
three inches, is divided into five transver.se ridges, the jiroportions 
of which, as to height and basal breadth, with the ridged and 



THE STEGODONTINiE: STEGODON 



861 



wrinkled character of the enamel, suffice for its reference to a 
species of the group of Proboscidians discovered by C'rawfurd in 
the Irrawadi Tcrtiaries of Ava, and described by ('lift in the 
second volume of the second series of the Transactions of the 
Geological Society (p. 369, pis. 36-39, 1828). ... In the present 
tooth the first or foremost ridge (PI. xxvii. figs. 1 & 2, 1) is defined 
by a cleft on the outer side of the tooth, but not on the inner side, 
fig. 3; here the abraded surfaces or ridges 1 and 2 are blended by 
wear into a common hollow field of smooth dentine (fig. 1, a). 
There is a slight constriction near the part where the worn surface 
of the first ridge blends with that of the second ; and this constric- 
f ion, which may be detected in the succeeding ridges, I take to be 
a tiace of that stronger one which more completely divides the 



Stegodon elephantoides ('lift, 1828 
Figures 683, 686, 695, 696, 700, 701, 738, 739 741 

Lower PlioceiU', lowest levels of the Irrawaddy Series (fluviatile),' near 
Yenangyaung, 250 miles below Ava, Burma. 

Syn.: Elephas cliftii Falconer, 1846 [ = StegodoH elephantoides (=cliftii) 
of the present Memoir.] 

Specific Characters (Clift, 1828, Osborn, 1929). — Third 
inferior molar, M3, with nine complete ridge-crests and a well- 
developed half ridge, equaling ten; five to eight conelets on each 
ridge-crest; length 11 inches = 280 mm., est. breadth Sji inches = 
90 mm. ; no apparent cement. Superior molar, l.M^ ( = Falconer's 
type of Elephus cliftii) with six and a fourth ridge-crests, length 
155 mm., breadth 83 mm., ten to twelve conelets on each; cement 
in the bottom of the valleys. Palate with l.M^, r.M^, with six and 
a fourth ridge-crests; length of l.M^, 186 mm., breadth 102 mm.; 
riilge-crests worn; traces of cement. From the three specimens 




iM a.ij,^ Uiktj 



Jktial fy CS^Un-M-Jti 



Mus 



T^totJ. by /Cnyrfijirei ZsaJ SSi/ Mdt-; p^aw AwiJ. 

Lectotype of Stegodon elephantoides 
Fig. 738. Lectotype left Ms, M3, of Mastodon elephantoides Clift, 1828, PI. xxxvni, fig. 2, one-third natural size- 
{Op. cit.,p. 372 and Explanation of Plates): "Left side of the lower jaw of Mastodon elephantoides. The remains of 
the anterior molar tooth [l.Mj] are seen, and behind it, the posterior tooth which was advancing, and which, in 
consequence of the jaw-bone being broken away, is seen through its whole length. This tooth is eleven inches long and 
three and a half broad." 



transverse coronal ridge in the molars of better Mastodons into 
an inner and an outer part. A well marked tubercle (figs. 1 and 
2,/) projects at the outer side of the base of the first ridge, 1, 
near the interspace between that and the second ridge. . . . Never- 
theless in the number of ridges in a given tract of the grinding- 
surface, in their height and breadth of base, and in the absence of 
intervening cement, the conformity of the Chinese molar with the 
grinders of the Mastodon elephantoides is close. The enamel also 
shows the same vertical linear impressions and ridges, by which 
we may reckon that the summit (say, of the fourth ridge in the 
tooth here described), if it were unworn, might be cleft into from 
thirteen to fifteen small mamillse." Owen did not compare this 
specimen with Cliff's original type lower molar of Mastodon 
elephantoides, but with specimens referred by Clift to that species. 



enumerated below the lectotype and cotype ridge formula is 
compiled as follows: M 1^ M 2^ M 3rs- 

Osborn, 1927: This Stegodont is siinilar to S. bonibifrons in 
ridge-crest formula. 

History. — As fully explained above (pp. 855, 856), Stegodon 
elephantoides Clift is the second species of Stegodont based by Clift 
on a lower jaw (PI. xxxviii, fig. 2) and on an upper molar (PI. 
XXXIX, fig. 6). Unfortunately Falconer was led to abandon 
'Mastodon elephantoides' and to substitute 'Elephas cliftii'; conse- 
quently all the literature subsequent to Cliff's original description 
and most of the reproductions of his illustrations appear under the 
specific name Elephas cliftii, which is actually a synonym of 
Stegodon elephantoides. Falconer's mistake partly arose through 
Cliffs error in entitling the palate of Mastodon [ = Stegodon] 



'[See note on page 824 above regarding the Lower Pleistocene age of Stegodon elephanlnides. — Editor.) 



862 



OSBORX: THE PROBOSCIDEA 



elephantoides (figured in PI. xxxvi) as 'Maatodon latidens.' 
Consequently ('lift's species M. [ = Slpgodo)i] rlephnntnidcs rests 
upon three specimens: 

Lectotype: "Left lower jaw, Mastodon Pllephanloides" (C'lift, 
1828, PI. xxxviii, fig. 2 = figures 696 and 738 of the 
present Memoir). 

CotjTie: First left superior molar, l.M\ "Upper molar of M. 
Elephanloide.^" (C'lift, 1828, PI. xxxix, fig. 6 = figures 
700, 739, and 740 of the present Memoir), .sub- 
sequently made the type of Elephas rliflii by 
Falconer and Cautley. 

Referred: Palate, with l.M- and r.lSP, "Upper jaw of Mastodon 
latidens" (C'lift, 1828, PI. xxxvi), not figured in the 
present Memoir. 

Mastodon elephantoides Clift, 1828. "On the Fossil Remains 
of two New Species of Mastodon, and of other vertebrated Animals, 




cene, lowest levels of the Irrawaddy Series (fluviatile).'- LiccTO- 
TYPE Figure.— Clift, 1828, PI. xxxviii, fig. 2. Cotype 

Figure.— Clift, 1828, PI. xxxix, fig. 6 [ = !<t4'gudoH elephantoides 
{ = dijtii).\ Referred Palate.— Clift, 1828, PI. xxxvi. 

Clift's Original Description of 'Mastodon' elephan- 
toides. — We refer to Clift's clear and consistent description 
quoted in full above (p. 827) together with the beautiful lectotype 
figure reproduced in our figure 738 herewith. Clift's specimen 
comes from Burma and his figureagrees exactly with his descrip- 
tion, namely, M 3to, with five to eight conelets on each ridge. 

Fig. 2a 




Cotype of Stegodon elephantoides Clift (=cliftii Falconer) 



Fig. 740. Cotype l.M' of Stegodon elephantoides 
( = dijlii), after Falconer and Cautley, 184(1 [IS47, 
PI. XXX, figs. 2, 2a], loft M', said to he the same 
molar (see Pal. Mem., 1868, Vol. I, j). 461) as that 
figured in Clift, 1828, PI. xxxix, fig. 6, although 
the drawing does not agree well with Clift's original 
illustration nor with the reproduction from the 
cast (Fig. 739 opposite). 

Observe rudimentary anterior ridge (pro-protoloph) and rudimentary seventh ridge (heptaloph), also the six ridge- 
crests (proto-, meta-, trito-, tetarto-, penta-, and hexalophs), the ridge-crest summits each crowned with from ten to 
twelve conelets. 



Fig. 739. New figure of type (l.M') of Elephas cliflii 
Falconer and Cautley, 1846, first figured by Clift, 1828, 
PL XXXIX, fig. 6, as an "t'pper molar of M. Elephantoides." 
Reproduced herewith from cast (Amer. Mus. Warren 
Coll. 10382) one-half natural size. Original in British 
Museum (Brit. Mus. M. 10520). From near Yenangj'aung, 
Burma. 



found on the left Bank of the Irawadi." Trans, (ieol. Soc, 
London, (2), II, Pt. Ill, 1828, pp. 372, 373. Lectotype.— (1) 

Left side of lower jaw with Mj, M3; original in Museum of 
Geological Society of London; cast Brit. Mus. 7393 [referred by 
Falconer and Cautley in the "Fauna Antiqua Sivalensis," I'l. 
XX, figs. 9, 9a, PI. XX. A, fig. 6, to E. insignis].' ("otype. — 

(2) First upper molar of the left side, l.M' (original in the British 
Museum (Brit. Mus. M. 10520), cast Amer. Mus. Warren Coll. 
10382) ; same tooth was selected by Falconer and Cautley as the 
type of Elephas cliftii. Referred (Osborn). — Palate with 

l.M- and r.M- (figured as .Mastodon hUidens by Clift). llou- 

izoN and Locality. — Left bank of the Irrawaddy River, 
near Yanangj^iung, 250 miles below Ava, Burma. Lower Plio- 



COTYPE CH.\R.\CTERS OF STEGODON ELEPH.\NTOIDES 

( = CLIFTII) FALCONER AND CAUTLEY, 1846, PAGE 47, 

BASED ON CLIFT'S SPECIMEN AND FIGURE 

Falconer's designation of this type ('^left M') and of this 
species is as follows: "The same group comprises a fourth e.xtinct 
Indian species, named in this work, E. Cliflii, which furnishes the 
next link in the chain of forms presented by the molars of the 
Elci)liantidie. ... In our view, the tooth represented in pi. 39, 
fig. 6, of Mr. Clift's memoir in the Geological Transactions [Clift, 
1828(, under the name of Mastodon Elephantoides, and the palate 
.specimen represented in pi. 36 of the same memoir, under the name 
of M. latidens, belong to this species. . . . The penultimate and 
antepenultimate molars in the upper jaw have onlj' six transverse 



'[See Chakravaiti, D. K., Quart. .Tourn. Geol., Mining, Metalluig. .Soc. India, 1937.1, p. 34, who referred it to .S'. f;»7)/ia»i(oirfos.— Editor.] 
-[See note on page 824 above regarding the Lower Pleistocene age of Stegodon elephantoides. — Editor.) 



THE STEGODONTINiE: STEGODON 



863 



ridges, continuous, and chevron shaped, with numerous mammil- 
lae, as in K. insignis and E. Gnm-sn; hut the cement does not fill 
up the interspaces of the ridges, being reduced to a comparatively 
inconsiderable quantity in the bottom of the hollows. E. Clijiii, 
in the reduced number of the coronal ridges, and in the other 
characters of the teeth, appears to constitute the dental link which 
forms the immediate passage from Elephas into Mastodon." 

ADDITIONAL CHARACTERS OF THE COTYPE SPECIMEN OF 
STEGODON ELEPHANTOIDES ( = CLIFTII) 
(Falconer, 1868, Vol. II, p. 84): "But the detached tooth 
[Fig. 740] on the upper jaw is seen entire, and beautifully preserv- 
ed, in the specimen fig. 2 of the same plate [i.e., PI. xxx], presenting 
six ridges and a small hind talon. The same tooth is represented by 
fig. 6 of PI. XXXIX. of Mr. Cliffs Memoir (Geol. Trans., vol. ii. 
2nd series). It is there described as an upper molar tooth of 
Mastodon Elephanto'ides, under which title Mr. Clift included 
specimens that are referred in our arrangement to two distinct 
forms. . . . The Elephantine affinities of this tooth are indicated 
by the absence of a longitudinal line of division along the crown, 
and by the great number of points (about eleven in each) that enter 
into the composition of the ridges." Of the same cotype molar, 
Lydekker (1886, p. 81) observes: "7388. Cast of the first (?) left 
upper true molar in an early stage of wear (woodcut, fig. 20). 
The original was obtained near Yenankhoung, on the left bank of 
the Irawadi in Upper Burma, by Crawfurd in 1826, and is pre- 
served in the Museum of the Geological Society;'" . . . There is 
scarcely any trace of the median cleft, the cement is slight, and 
there are numerous cusps. Mantell Collection. Purchased, 1836." 

FALCONER'S NOTES OF 1868 ON STEGODON ELEPHANTOIDES 

( = CLIFTII) 
Falconer, "Palaeontological Memoirs," Vol. I, 1868, pp. 461, 462, Plate 
xxx of the "Fauna Antiqua Sivalensis." 

It is very important to observe that the four specimens 
described by Falconer, including the cotype of .S. elephantoides 
from Burma, agree with each other both in the number and char- 
acter of the ridges and in the number of conelets on each ridge. 

Upper Jaws. — Burma, left bank of the Irrawaddy River, 250 
miles below Ava. Plate xxx, figs. 1, la, E. cliftii F. & C, palate. 
Dp*, ? ridges; figs. 2, 2a, type,'" Burma, 250 mUes below Ava, 
l.^U, ridges &%, beautifully preserved, "six ridges and a small 
hind talon ... as many as eleven to twelve denticles. ... Its 
elephantine affinities are indicated by the . . . great number of 
points [denticles]"; fig. 3, superb palate, M^, ridges 6K, little 
cement (referred by Chft, 1828, PL xxxvi, to M. latidens); figs. 
4, 4a, 46, fragment of M-, right side, 5 ridges, cement moderate in 
quantity, from near Yenangyaung, upper Burma. Lower 

Jaws.— Burma. Plate xxx, figs. 5, 5a, I.M3, "eight ridges and 
a talon," little cement. 

Cotype Characters. — From the above it appears that the 
four specimens described by Falconer from near Yenangyaung, 
250 miles below Ava, Burma, exhibit the following characters: (1) 
Denticles very numerous, eleven to twelve on each ridge ; (2) ridges 
moderately high with little cement; (3) ridge formula as follows: 

Stegodon elephantoides { = cliftii): M l^^^M 2^' M i^^. 



PHYLETIC CONCLUSIONS 
Osborn, 1927: Falconer considered Cliffs Lower Pliocene- 
type of Mastodon [ = Stegodon] elephantoides as close to his Upper 
Pliocene S. insignia, but S. elephantoides proves rather to be close 
to the Middle Pliocene S. bombifrons, as shown in the comparative 
ridge formulae table above, the two formulae abbreviated being as 
follows : 

S. bombifrons: Dp 3i Dp 4 ^^ M 1 ^^^ M 2 i M 3 It 
S. elephantoides: Dp 3| Dp 4J M 1 ^* M 2 -p M 3 " 



Elephas cliftii [ = Stegodon elephantoides ( = 



and Caut. : Dp 3 } Dp 4 1 M 1 -^ M 2 ^- M 3 

Hg.l. 



»«• 



1 ^" " 10- 

cliftii)], fide Falc. 




Referred Stegodon elephantoides (= cliftii) 
Fig. 741. Referred I.M3 of Elephas cliftii, after 
Falconer, 1868, Vol. II, PI. v, figs. 1, 2, "Views in plan and 
profile of last true molar, lower jaw, left side, rather less 
than one-fourth (two-ninths) of the natural size." See also 
the "Fauna Antiqua Sivalensis," Falconer and Cautley, 
1846 [1847, PI. xxx, figs. 5, 5o]. From Burma, presented to 
the British Museum by Colonel Burney (Brit. Mus. 14759). 

The ridge formula of I.M3, namely, i^, assigned by Falconer 
and Lydekker, actually occurs in an aged specimen from Burma, 
Falconer's figure of which (Falconer and Cautley, 1846 [1847, PI. 
xxx, figs. 5, 5a] is reproduced herewith in our figure 741; this 
grinder is so old that the two anterior ridge-crests may have worn 
off, consequently the ridge formula is uncertain; the dimensions 
(length 12.7 in. =323 mm., breadth 4.5 in. = 115 mm.) considerably 
exceed those of the type of Stegodon elephantoides given above. 
There is no substantial basis, therefore, for the assignment to the 
synonymous Elephas [ = Stegodon] cliftii of a ridge formula inferior 
to that of Mastodon [ = Stegodon] elephantoides Clift. 

Stegodon bombifrons Falconer and Cautley, 1846 
Figures 686, 687, 699, 731, 732, 734, 742-746, 777, PI. xx 

Lectotyi)e; Siwalik Hills (Jide Falconer), probably from the Dhok 
Pathan horizon, Lower [Middle] Pliocene, India, in which this species is very 
abundant and characteristic (Pilgrim-Brown). 



'Now in the British Museum (Natural History), M. 10520= Stegodon elephantoides (=diftii). 
-[See note on page 824 above regarding the Lower Pleistocene age of Stegodon elephantoides. — Editor. 



864 



OSBORN: THE PROBOSCIDEA 



This is an curly Lower [Middle] Pliocene stage of the true 
Stegaddn based on types from the Dhok Pathan zone which also 
contains Slegoloiihodoii. latidena. From this species some of the 
earlier authors believed that the genus Elephas originated. Re- 
ferred Stegodon bomhifronn occurs in the Tatrot horizon. Middle 
[Upper] Phocene. A superior molar from the DIkjU Pathan 
horizon (Fig. 746) exhibits \ery clearly the cone and conelet 
structure of the crown. 

E. [Elephuii] hombifrons Falconer and Cautley, 1846. "Fauna 
Antiqua Sivalensis" . . . letterpress, 1846, pp. 46, 47. Llxto- 

TYPK AND CoTVPKS. — {Op. cit., p. 46): ". . . scvoral crania con- 
taining perfect teeth," of which the lectotype is figured in PI. 
XXVI. Horizon and LocALiTY.^Siwalik Hills, India, prob- 

ably from the Dhok Pathan horizon. Lower [Middle] Pliocene. 
Lectotype and Cotype Figures.— Op. cit., 1846 [1847, Pis. xxvi 
(lectotype), xxvii, xxviii], figures 742 and 744 of the present 
Memoir. 

Falconer and Cautley, 1846, p. 46 (see above p. 858) based 
this species on "several crania containing perfect teeth," described 
as "from the Sewalik Hills." Of these Lydekker (1886.2, p. 83) 
designated as the lectotype P.rit. Mus. M.2978. The exact locality 
is not given. According to Pilgrim (Vol. I, fig. 413, also Chap. 
XX n below) this species occurs abundantly in the Dhok Pathan 
horizon of Lower [Middle] Pliocene age, a horizon which also 
contains Teiralophodon punjabiensis, Synconolophus corriigalun, 
and S. hasnoli. Falconer assigned to this species a low ridge 
formula, namely : M 3 ^^■ 

Characters (Falconer and Cautley, 1846, p. 46); see 
OsBORN ABOVE, PP. 855-859.— (1) Type based on several crania 
containing perfect teeth. (2) Crown divided into transverse ridges, 
composed of numerous mammillae [conelets] of chevron-shaped 
section. (3) Intersjiaces occupied by a thick coat of cement. (4) 
Principal ridges of third upper and lower molars, M 3¥, in contrast 
to [S.] insignis. (5) Third upper molar measures 11 inches (279 
mm.) in length by 4K inches (114 mm.) in width. (6) The lower 
third molar of the left side, with nine [9)2] ridge-crests, measures 
13.4 inches in length ( = 340 mm.) by 4.2 inches in breadth (= 105 
mm.), coasiderably exceeding the dimensions of the lectotype of 



.S. elephantoides, namely, length 280 mm., breadth 90 nun. (7) 
The collective (Falconei-, T^ydekker, Osborn) ridge-crest formula 
(not including the uncxi)os('d ridge-crests) is as follows: 

Stegodon howbifrnns: Dp 3 t Dp 4 j^r^^^lj^ M 1 ^^YVi 



M 2 



M3 



7»t-8-8t>-9-9^t 
8-9-9VS [1 1+1 • 



FALCONER'S NOTES OF 1868 ON ELEPH.\S [=STEG0D0N1 
BOMBIFRONS 
Fiilconer, 'Tiilu'diitoloKic^al Memoirs," Vol. I, ISOS, pp. l.'iC), l.'jS^dl, 
Pliitos XXV -XXIX. H of tlic "Fuunii .Antiiiua Sivalensis." 

[Errors of lietenniiuitioM iiulii'ated hy Oshorii in scpiaR^ l)nu-kots.l 

Elephit.s boiNb/froiia. I'alconer and Cautley, 1846 [1847, Plates 
xxvi, XXVII, xxviii]. Lectotype and Cotypes. — Cranium 

represented in Plate xxvi (Lydekker's lectotype); Plate xxvii, 
very fine skull, M', ridges 9%, length of molar, 10.2 in. = 257 mm., 
width 3.7 in. = 93 mm., and Plate xxviii, fig. 2, same .skull. 




Fig. 743. Stegodon bombifrons. Much more ))rimilive 
than S. insignis. Observe transverse ridges of trunk to 
elevated nasals, extreme broadening of sumnut of oeeijiut. 
Ears conjeetural. Limbs given the same proportions as 
the Stegodonts throughout, without knowledge of skeletal 
material. Female to right, direetly after lectotype skull; 
no inferior tusks. Male to left. Restoration by Margret 
Flinseli, 1930. One-fiftieth natural size. 

Upper Jaws. — Plate xxix, fig. 1, broken cranium, M', ridges 
8K, length of molar 10 in. = 253 mm., width 4 in. = 101 mm.; figs. 





ELEPHAS BOMBIFRONS Ret. 
Pi. XLIII, XIII 



ELEPHAS BOMBIFRONS Lectotype Lyd 
PI. XXVI 



ELEPHAS BOMBIFRONS Cotype 
PI. XXVIII, Fig. 2 



ELEPHAS BOMBIFRONS Ret. 
PI. XLV, XIII (rov.) 



Steoodon Crania, after Falconer and Cautley's Illustrations in the "Fauna Antiqua Sivalensis" 
Fig. 742. Crania of Stegodon (Elcphas) bombifrons, lectotype, cotype, and referred. Outlines assembled from original plate drawings 
if this .sjieeies in Falconer and Cautley's "Fauna Antitiua Sivalen.sis," frontal, palatal, and lateral .aspects. All one-sixteenth natural size. 



THE STEGODONTIN.E: STEGODON 



865 



2, 2a, cranium, ]\P, 8 ridges; figs. 4, 4tt, r.lVP, 8+ ridges; figs. 5, 
5a, upper jaw, r.M', 7K ridges; figs. 6, 6a, palate with r.M', 9K 
ridges, length 10.9 in. = 276 mm., width 3.8 in. =97 mm. (at ends), 
4.3 in. = 112 mm. (in middle). 

Lower Jaws. — Plate xxv, figs. 3, 3a, lower jaw, r.Mj, 9% 
ridges, enamel very thick, scanty cement. Plate xxix.a, figs. 1, 
la, lower jaw with Dp4, ridges 6K, "probably the third [fourth] 
milk molar"; figs. 2, 2a, lower jaw, r.Mi, ridges 7'4, "certainly 
the first true molar"; figs. 3, 3a, lower jaw, I.M2, ridges Iji; figs. 
4, 4o, lower jaw, M2, ridges 7}i; figs. 7, 7a, lower jaw, Mi, ridges 7. 
Plate XXIX. B, figs. 5, 5a, lower jaw, r.Dp, "with 5 ridges and an 
anterior and posterior talon"; figs. 6, 6a, lower jaw, r.Mi, ridges 
7}i; figs. 7, 7a, lower jaw, I.M3, ridges 9>2 [11 + ]. 

In the two plates (xxix.a, xxix.b) of the "Fauna Antiqua 
Sivalensis" about sixteen specimens of inferior molar teeth are 
beautifully figured; they exhibit six to ten conelets on the unworn 
crown and a maximum of eleven conelets on the worn crown; 
five of the conelets double by dichotomy. Thus the number of 



conelets is approximately the same as in Stcyodon elephanloides 
{ = cliftn). 

From the above observations we deduce the ridge formula of 
Stegodon bombifrons practically as above under "Characters." 



LYDEKKER'S NOTES OF 1886 ON ELEPHAS (=STEGODONl 
BOMBIFRONS 

"Catalogue of the Fossil Mammalia in the British Museum (Natural History)," 
Part IV, 1886, pp. 82-88 

Lydekker (1886.2, p. 83) designates as his lectotype Brit. 
Mus. M.2978: "The cranium, showing the third true molars of 
both sides in an early stage of wear. This specimen is the type, 
and is figured by Falconer and Cautley in the 'Fauna Antiqua 
Sivalensis,' pi. xxvi"; reproduced in figure 742 of the present 
Memoir. 

Lydekker's notes are based on forty-four specimens in the 
British Museum referred to this species, chiefly from the Cautley 




CoTYPE OF Stegodon Bombifrons 

I''ig. 744. Cotype of Elcphas bombifrons I'alconer and Cautley, 1846. After Falconer and Cautley, 1846 [1847, Pis. xxvii, xxvin], one-sixth natural .size. 
Brit . Mus. M.2979; cast Amer. Mus. 10378. From the Siwalik Hills, India. 

Falconer, Pal. Mem., 1868, Vol. I, p. 458: (PI. xxvii) "Very fine and perfect skull, anterior view." (PI. xxviii) Fig. 1. "Lateral view of same skull, 
as figured in Plate xxvii. — B. M. Fig. 2. . . Palate view of same skull, showing sections of tusks, and last ? true molar on either side, with 9 ridges and a heel; 
the 8 front ridges worn. The interval between the molars in front is very narrow; behind they are extremely divergent. — -B.M." Fig. 5. Occipital view of 
another skull. 



866 



OSBORN: THE PROBOSCIDEA 



Collection (1842); the specific references therefore are based on 
Falconer's determinations. 

Specific Chakacters.— (Lydekker, 1886.2, p. 82): "The 
ridges are rather taller, some\\hat wider apart, and more numerous 
than in E. clifti [ = Stegodo7i clephantoides { = clifiii)], and the 
valleys are generally completely filled with cement; it is, however, 
sometimes very difficult to distinguish between the hinder teeth of 
the two species, while in the opposite direction it is often difficult 
to distinguish between those of E. bombifrons and E. insignis. 
The teeth figured by Falconer and Cautley under the name of 
E. ganesa cannot be distinguished from those of the present species, 
and are therefore provisionally classed under the same head. 
The teeth are frequently very large, and the ridges are often 



ridges, enable us to write the standard maximum formula as 
follows : 

^Maximum ridge formula of Ekphas [ = Stegodon] bombifrons: 
Dp 2j Dp 3i Dp 4 'i M 1 f M 2,\^ M 3 i 

Stegodon insignis Falconer and Cautley, 1845, 1846 
Figures 686, 688, 697, 731, 732, 735, 747-753, 756, 760, 764, 766, 776, 819, PI. xx 

Siwalik HiUs, India, Upper Pliocene, Pinjor horizon, to Lower Pleistocene, 
Boulder Conglomerate {fide Falconer and Pilgrim), to Upper Pleistocene. In 
the present Memoir (see Fig. 413) the upper levels of the Pinjor horizon are 
of Lower Pleistocene age {fide Barnum Brown). 

Falconer's types of this species agree in the ridge-crest foi mula 
with the type which he subsequently described as Stegodon ganesa 




Kekekked Right Third SupEnion Molar of Steqodon bombifrons 
Fig. 745. A partly worn right superior molar, r.M', referred by Falconer and Cautley 
to Elephaa bombifrons and corresponding very closely in all details to Falconer's description 
of the cotypcs of this species, namely : M 3 with nine and a half ridges; conelets — six divid- 
ing into eleven. After Falconer and Cautley, 1846 (1847, PI. xxix, fig. 6], one-third natural 
size. 



Fig. 746. Referred fragment of superior molar of 
Stegodon bombifrons collected by Barnum Brown in 
1922 in the Dhok Pathan horizon. Lower [Middle] 
Pliocene, two and a half miles northeast of Ha-snot, 
India. This specimen is wrongly numbered; it 
should read Araer. Mus. 20044. 

Three anterior ridge-crests of a right third superior 
molar, r.M' (compare Fig. 759 of Stegodon orien- 
talis grangeri type). 



curved; a trace of the median longitudinal cleft can often be 
observed in the first two or three ridges, and the inner columns of 
these ridges occasionally show accessory tubercles near the longi- 
tudinal cleft, where they assume a Maslodon-hke shape. The 
plane of wear of the teeth of this and the following species [E. 
ganesa] is similar to that of the true F^lephants. The mandibular 
symphysis is produced into a spout-like termination, as in E. 
indicus. The cranium has the fronto-parietal region very convex, 
the constriction of the frontals by the temporal fossff being more 
marked than in the other species. Ilab. India (Punjab to Siwalik 
Hills) and (?)China [Footnote: 'Koken, Pal. Abhand. vol. in. 
pt. 2, p. 12 (1885).']. The species may perhaps also occur in 
Java." 

The ridge formula of Lydekker (op. cit., 1886.2, p. 82), namely, 

MnA 1 MC.i-e) TV r 8.(fl.7).(8-9) . , . . ^ r • ^ 

m. I Dpi V ,(,,.7), M 7.(7.8).(8-9)' 's less precise m not refernng to 

tlic half-ridges but does not otherwise differ excepting in the supe- 
rior ridges (^) of Dp '.i fiom tiie type ridge fornuila gathered 
from Falconer's obs('r\ations abov(^. Lydekker's obser\ations fully 
substantiate Falconer's fornmla of 1868, and, omitting the half- 



but they exhibit profoundly different characters in the cranium, as 
shown in a comparison of figures 735 (S. insignis) and 736 (<S. 
ganesa), or in a comparison of figures 752 (S. insignis) and 732 
(S. ganesa). This profound difference, as explained above and 
below, is attributed to the fact that all the crania referred by 
F'alconer and Cautley to Stegodon insignis represent small-tusked 
and probably female individuals, while crania referred to S. ganesa 
represent large-tusked and probably male individuals. 

]''alconer originally described Elephus insignis in 1846, [). 37, 
and in the same communication (]>. 45) he named a fourth species 
Elephas Ganesa, describing a third superior molar (which had 
been figured in 1845, PI. in, fig. 7a) and remarking that the tooth 
bears the closest resemblance to the corresponding tooth in E. 
insignis. This doubt always remained in Falconer's mind, for 
in his notes of 1S67, ]). 4, and of 1S68, Vol. I, yi. 424, he remarked: 
"In f.'u't, there are no good characters by which the teeth of these 
two sj)ecies can be satisfactorily distinguished, although tiie crania 
are so remarkably different." fn I'^alconer's mind, flicrcfore, 
Elephus [ = Stegodon] insignis po.ssessed a cranium of the type he 



THESTEGODONTIN^: STEGODON 



867 



figured in PI. xv of the "Fauna Antiqua Sivalensis," reproduced 
herewith in our figure 735, in contrast to Elephas [ = Stegodon] 
gaiwsd which possessed a cranium of the type figured in PI. 
XXIII, reproduced in our figure 736. 

Sexual Disparity. — As explained above in the discussion of 
the crania. Falconer apparently selected the female crania as refer- 
able to Elephas [ = Stegodo7)] insignis and the wio/e crania as refer- 
able to E. [S.] ganesa. A detailed comparison of all the referred 
specimens has failed to establish any true specific distinction be- 
tween these two species. Consequently we may regard the lecto- 
type and referred specimens from the Pinjor, Boulder Conglomer- 
ate, and Godavari, Narbada Alluvium, as 'collective species' 
including a number of ascending mutations or subspecies which 
will be recognizable by profound monographic research; it is not 



Horizon and Locality. — Siwalik Hills, India, probably Pinjor 
horizon. Upper Pliocene or Lower Pleistocene. Lectotype 

Figure.— Op. cit., 1846 [1845, PI. ii, fig. 6a]. Cotype.— 0/). 

cit., 1846 [1845, PI. ii, fig. 66]. Description. — Falconer's 

original description of 1846, p. 37, quoted in part above in this 
Memoir (p. 829) was restated in 1868, Vol. I, p. 423, fig. 6a of 
PI. II, as follows: "Elephas iiisignis, from the Sewalik Hills [Fig. 
747 of the present Memoir]. Vertical section of last upper molar. 
The four anterior ridges are affected by wear; the six posterior 
ridges are entire, the fangs are fully developed, and their mode of 
implantation in the jaw is distinctly shown. The white mass in 
the centre represents the body of ivory, which is projected upwards 
in ten angular lobes, terminating in a sharp edge. The height of 
these lobes does not much exceed the width of their base, and closely 



7u/ eh 



Ft J t'li 





Lectotype axd Cotvpe of Stegodon insignis 

Fig. 747. Lectotype left superior molar, I.M', of Elephas inmgnis Falconer and Cautley, 1846, from the Siwalik Hills, India, after Falconer and 
Cautley, 1846 [1845, PI. ii, fig. 6a], also cotype third inferior molar, M3, after Falconer and Cautley, 1846 [1845, PI. 11, fig. 66j; both figures one-third 
natural size. 

(Falconer and Cautley, 1846, p. 37): "Fig. 6a, pi. 2 [lectotype l.M' — Brit. Mus. M.30151, represents a vertical and longitudinal section of the 
last upper molar of an Indian fossil species, which we have named Elephas insignis in this work." Length of this tooth 10.3 inches. Inverted to show 
natural position. (Falconer, "Palseontological Memoirs," 1868, I, p. 423): "The four anterior ridges are affected by wear; the six posterior ridges are 
entire, the fangs are full}' dcvelojied, and their mode of implantation in the jaw is distinctly shown." Fig. 66 (cotype M3 — Brit. Mus. M.3039) is a 
vertical section of a third lower molar of E. insignis, one third natural size (cf. Falconer, 1867.1, p. 4; 1868.1, I, p. 424; also caption to figure 097 above). 



probnble that a single true specific stage passed from the Pinjor 
into the Narbada Alluvium horizon. 

In the present Memoir we shall first treat Falconer's descrip- 
tions of the types and referred specimens of these two species 
separately and then unite them under the collective species name 
Stegodon insignis-ganesa. 

Cranium. — The lectotype cranium of the third species of 
Stegodont, described as Elephas insignis by Falconer and Cautley, 
1846, is from the Cautley Collection but is without record as to 
its exact geologic level. The specimen is now in the British Mu- 
seum. Lydekker (1886.2, p. 91) designates it as ".M. 3015. An 
imperfect cranium, showing the third true molar of either side." 

Elephas insignis Falconer and Cautley, 1846. "Fauna Anti- 
qua Si\alensis," letterpress, 1846, p. 37. Leciotype. — A 
third superior molar of the left side, l.M', in an imperfect cranium 
(Brit. Mus. M.3015) containing the third true molar of either side. 



applied over them is a thick layer of enamel reflected up and down 
in a continuous zig-zag plate. The interspaces of the five posterior 
ridges of enamel are completely filled up by a mass of cement much 
exceeding the enamel in thickness (vide Plate vi. fig. 7). This is 
the best illustration of the intermediate type of a proboscidean 
molar tooth, from which those of the other species diverge in 
opposite directions. It belongs to the Mastodon Elephanio'ides of 
CUft. The dark granulated shade below the portion of the ivory 
nucleus sustaining the five posterior ridges indicates the hollow of 
their common fang, which in the fossil is occupied by a core of 
sandstone. — B.M. (Reproduced in PL iv. fig. 1.) Length of tooth, 
10.3 in." 

Characters of Elephas [= Stegodon] i\sifiNis. — (Lydokkor, 
1886.2, p. 89): "The apparent impossibility of distinguisliing (Ik; 
dentition of this species from that of E. ganesa has been already 
mentioned. . . . The ridges of the cheek-teeth are usually rather 



J not. 31^ e 




I 777? 

-foresJiortenecl 



7i.Mj 




Fig. 748. Stegodon insignis ref. (Amcr. Mvis. 19869) suijcrior and inferior molars proVjably belonging 
to the same individual as the mandible shown in figure 7.50. Pinjor horizon, Upper Siwaliks, near Siswan, 
India. One-third natural size. 

A, Al, A2, Imperfect second superior molar, 1.M-, with +5^ ridge-crests much worn; a complete 
third superior molar, LM', with }j-8-}'j ridge-crests, of which the two anterior are slightly worn. 

B, Right third inferior molar, r.Mj, displaying K-9+ ridge-crests, of which the three anterior are 
slightly worn (eon\parc Fig. 7.")0). The l.M.i, now completely expo.sed, displays a total of 12 ridge-crests. 



868 



THE STEGODONTIN^: STEGODON 



869 



taller and narrower than in E. bombifrons, their average number 
greater, and the cement still more abundant. It is, however, not 
always easy to distinguish between the two. The third molar is 
usually narrower posteriorly, and the enamel frequently thinner. 
The taller and more numerous ridges indicate that the present 
species is intermediate in respect of dental characters between E. 
bombifrons and E. plmiifrons. The adult cranium is remarkable 
for the great depression of the fronto-parietal region, although this 
feature is less marked in some specimens than in others [Footnote: 
'Compare "Fauna Antiqua Sivalensis," pi. xliii. figs. 15, 156.']; 
but in the young cranium the contour is indistinguishable from 
that of the adult E. ganesa [Footnote: 'It is of course self-evident 
that these young crania (like detached teeth) might equally well 
have belonged to E. ganesa.']." 

Referkkd Materials in the American Museum. — A 
typical individual (Amer. Mus. 19869— Figs. 748, 750), In the 
primitive S. insigni.^-gancsa stage, was foimd by Barmun Brown 
near Siswan (Pinjor horizon), India, in which the upper and 
lower jaws are fortunately associated; this specimen exhibits the 
following ridge-crest formula: 



M2 



±5W=7) 



M3 



Also recorded from Siswan are two other si)ccimens, namely, 
a second sujierior molar, r.M" (Amer. Mus. 19804 — Fig. 749), in 




Fig. 749. Rpferrcd Stegodon insignis (Amer. Mus. 19804), 
a seeond superior mohir, r.M-, taken from palate found below 
the conglomerates, three miles northeast of Si.swan, Up))er 
Siwaliks, India. Ridge-erests, namely, }2~7-y>, agree elosely 
with S. insignis, somewhat too progressive for S. bombifrons. 



which the ridge-crests ( ^"^''' ) agree closely with the tyjjical 
formula of Stegodon iiisignis, and a third left inferior molar, l.Mj 
(Amer. Mus. 19859 — Fig. 753) with ^r^^i ridge-crests. A juvenile 
right lower jaw (Amer. Mus. 19858 — Fig. 751) exhibits Dp4 with 
7 ridge-crests; a young adult lower jaw (Amer. Mus. 19964 — Fig. 
752B) shows I.M2 in full u.se, I.M3 coming into jjlace; the ridge- 
crest formula of I.M2 is 7. Still another specimen from the same 
locality of Siswan is a lower jaw, also a right upi)er tusk complete 
(Amer. Mus. 19773— Fig. 766). 

The grinding teeth of these three individuals in the American 
Museum collection (see Figs. 748, 749, 7.50, and 753), all recorded 
from the Pinjor horizon. Upper Pliocene [or Lower Pleistocene], 
exhibit i)rogressive variation in the ridge count, as indicated by 
numerals in these figures; according to Doctor Brown, it is not 
certain that they belong in the Pinjor horizon proper, but may 
have come in by erosion from the conglomerates above. 



/t. M /3369 




Fig. 7.50. Referred inferior mandible, assoeiatod with two upper teeth, 
of Stegodon insignis (Amer. Mus. 19869), collected by Barn\nn Brown in the 
Pinjor horizon, Upper Siwaliks, near Siswan, India; it contains both second 
inferior molars, M2, well worn, also right and left third inferior molars, M3, 
little worn, with )^9-|- ridge-crests; the I.M3, after removal of the bone, 
displays 12 ridge-crcsts (compare Fig. 748, probably of the same individual). 

Stegodon ganesa Falconer and Cautley, 1845, 1846 
Figures 686, 698, 731-733, 736, 7.54, 755, 757, 760, 766, 819, PI. xx 

Siwalik Hills, India, Upper Pliocene Pinjor horizon, to Lower Pleistocene, 
Boulder Conglomerate (fide Falconer and Pilgrim), to Upper Pleistocene. 
In the present Memoir (Fig. 413) the upper levels of the Pinjor horizon are of 
Lower Pleistocene age (fide. Barnum Brown). 

This species (fide Falconer, Lydekker) exhibits the same deni al 
characters (Fig. 757) as Stegodon insignis but is distingiushed by 



dp^ 



STEqODON INSIGNIS i?e/r 



crown view j/^ not. 5ixe .rev. 




mer Mus. 19858 



j/aiv, rev. 

Fig. 7ol. Referred lower jaw, left aspect, of Slegodon insignis (Ainer. 
Mus. 19964), mature adult, colle<'ted by Barnum Brown in 1922 in the upper 
clays below the conglomerates, three miles north of Sisw-an, Pinjor horizon. 
Upper Pliocene, India. The juvenile right lower jaw of 5. insignis (Amer. 
Mus. 19858), collected by Dr. Brown also in the upper clays but two miles 
north of Siswan, summit of the Pliocene, is inserted for comparison. 

The ramus (Amer. Mus. 19964) exliibits a fully adult form, with abbrevi- 
ated symphysis, I.M2 in full use, I.M3 coming into place. The ridge-crest 
formula is M 2-J, M 3 |. One-sixth natural size. 

Observe Dp4 of Amer. Mus. 19858 which exhibits seven ridge-crests as 
compared with the typical Dp4 of S. insignis with seven to nine ridge-crests 
(of. Table V, p. 854 above). One-fourth natural size. 




/•^ncit. szje 
younq adult JCl^ 




ELEPHAS INSIGNIS Ret. 
PI. XVI. Fig. 3 




ELEPHAS INSIGNIS Bef. 
PI, XVI, Fig, 2 



ELEPHAS INSIGNIS Hof. 
PI, XLV, XV, B'fev.) 



ELEPHAS INSIGNIS Ref. 
PI. XV, 



I'ig. 752. Referred crania of Elephas [=Stegodon] insignis. Specimens 
reproduced in outline from original plates in Falconer and Cautley's "Fauna 
.Viitiqua Sivalensis," as indicated in the printed legends. Frontal a.spect of 
four crania, lateral aspect of one cranium, and posterior aspect of one cranium. 
.Ml one-twentieth natural siz(^ Compare figures 732 and 777. 



870 



THE STEGODONTIN^: STEGODON 



871 



profoundly different cranial contours, profile and facial aspects 
(compare Figs. 732, 752). Falconer, immediately after describing 
Eleplnis insignis (1846, p. 37), described Ehphas ganesa (1846, p. 
45) ; ajjparently he was very much puzzled by the exact similarity in 
the structure of the teeth as contrasted with the profound difference 




molar, M': "Plate iii, Fig. 7a. — Elephas Ganem, a fossil Indian 
species. Vertical section of last upper molar. The crown consists 
of ten principal ridges, with a subordinate talon ridge in front and 
behind. The anterior seven ridges have their summits worn. 
A small portion is broken off at the anterior end. The disposition 



Fig. 75.3. Rcferrod Slcgodon in- 
signis (.\mcr. Mus. 198.59), collected by 
Barnum Brown in 1922 two and a half 
miles south of Cliarnian, near Siswan, 
India, below the conglomerates, probably 
Pinjor horizon. Upper Pliocene. 

A third left inferior molar, I.M3; 
length 291 mm., width 93 mm.; ridge- 
crests of M 3 4-1 2 j«. One-third natural 
size. 






^rf* 



in the character of the skull, for in the "Palseontological Memoirs" 
of 1868 (Vol. II, p. 84) appears the following statement: 

"Regarding the specific distinctness of E. (Sicg.) Ganesa I am 
by no means so well assured; this species is chiefly founded on 
a huge cranium in the British Museum with long tusks, presented 
by Colonel Baker. I ha\e not been able to reconcile the form of 
this cranium with either that of E. (Steg.) insignis or E. {Steg.) 
bombifrons; but at the same time I must confess that I have 
failed in tracing its dentition satisfactorily as a distinct form 
through different ages." 

E. [Elephas] Ganesa Falconer and Cautley. "Fauna Antiqua 
Sivalensis," letterpress, 1846, p. 45. Lectotype. — A last upper 
molar, M'. Brit. Mus. 18489. Horizon and Locality. — 

Siwahk Hills, India, Pinjor horizon. Upper Phocene or Lower 
Pleistocene. Lectotype Figi he. — Op. cit., 1846 [1845. PI. in, 
fig. 7a — see figures 698 and 757 of the present Memoir]. 

Lectotype Description. — {Op. cit., 1846, p. 45) : "The crown 
consists of ten principal ridges, with a subordinate 'talon' ridge in 
front and behind. The anterior seven ridges have their summits 
worn, the two in front being ground down to the common base of 
ivory, the tooth having been a considerable time in use." 

Characters. — Falconer (1868, Vol. I, p. 424) remarks of this 



and relative proportions of the ivory, enamel, and cement bear the 
closest resemblance to those of the corresponding tooth of E. insig- 
nis, and the number of ridges agrees. In fact, there are no good 
characters by which the teeth of these two species can be satis- 
factorily distinguished, although the crania are so remarkably 

different.— B. M Length of tooth, 9.25 in." 

Lydekker (1886.2, p. 89) erroneously .selects Brit. Mus. M.3008 
as the type of E. ganesa: "The imperfect cranium, showing the 
l^artially-woin third true molar of either side, the base of the left 
and the greater portion of the right incisor; from the Pliocene of 
the Siwalik Hills. This specimen (the missing portions of which 
have been restored in wood) is the type, and is figured by Falconer 
and Cautley in the 'Fauna Antiqua Sivalensis.' pis. xxi., xxii., 
xxiii., and xliii. fig. 14. It is remarkable for the enormous size of the 
incisors. Presented bij Gen. Sir W. E. Baker, K.C.B., 1848." 



Fig. 754. This is the 
famous and oft-repro- 
duced referred skull 
and tusks of Elephas 
I = Slegodon] ganesa 
Falconer and Cautley, 
1845- 1S47, reproduced 
from Plate xxii, fig. 3, 
of the "Fauna An- 
tiqua Sivalensis." 
Same skull figures 732, 
733, and 736. Original 
in the British Museum 
(M.3008). 



^ 



X.' 







Sca/e o/'Iixt 



ELEPHAS GANESA 




Fig. 755. Restoration (1930 1933) of Slcgodon ganesa, to a onc-thirty-fifth scale, by Margret I'Miiisuh, under the direction of Henry Fairfield Osborn. 
Based on Falconer's figures of the skull and tusks [.see also Fig. 733.) 




Fig. 7.')G. Restoration (1934) of Slcgodon insigni«, by Margret Flinsch, under the dhection of Henry Fairfield Osborn. All figures about one thirty- 
sixth natural size. Restoration based on crania figured in tlic "Fauna .\nti(|ua Sivalensis" by Falconer and Cautley. Aged individual at left (PI. xvi, 
fig. 1), center (PI. XLiii, fig. xv..\, cf. PI. xv), right (PI. XLiii, fig. xv.B, of. PI. xvii, figs. 1 and 2). 

872 



THE STEGODONTIN^: STEGODON 873 

FALCONER'S NOTES OF 1868 ON STEGODON INSIGNIS AND S. GANESA 
FalcDner, "Paltoontological Memoirs," Vol. I, 1868, pp. 423, 424, 452, Plates ii -xxv of the "Fauna Antiqua Sivalensis" 

Slegodon i.nsigiiis. Lectotype of Slegodon insignis from the Siwalik Hills. Plate ii, fig. 6a, M^ with 10 ridges, the height of 
the.se ridges docs not much exceed the width at the base, closely covered with a thick layer of enamel, interspaces of five posterior 
ridges of enamel completely filled with cement. (Falconer and Cautley, 1846.1, p. 37) : "This tooth belongs to one of the forms 
which have been included under the name of Mast. Elephantoides, by Mr. Clift; and which Profe.ssor Owen names 'Transitional 
Ma.stodons.' " 

Stegodon ganesa. Lectotype. Plate in, fig. 7a, M^ length 9.25 in. = 235 mm. ; ". . . ten principal ridges, with a subordinate 
talon ridge in front and behind. . . . The disposition and relative proportions of the ivory, enamel, and cement bear the closest 
resemblance to those of the corresponding tooth of E. insignis, and the number of ridges agrees. In fact, there are no good 
characters by which the teeth of these two species can be satisfactorily distinguished, although the crania are so remarkably 
different.— B.M." 

Upper Jaws and Grinding Teeth 

Stegodon insignis. — Plate xix, figs. 1, In, upper jaw with Dp- and Dp', Dp' with 6 ridges; figs. 4, in, skull, M' with 7% 
ridges; figs. 6, 6a, skull with M-', J\F with ll}j ridges, abundant cement, length of M' 11 in. = 280 mm., width 3.8 in. =97 mm. 
Plate xix.A, figs. 2, 2a, palate with M^, M'; M 2 ridges 7, M 3 ridges K-ll-}-- Plate xxiv, figs. 6, 6a, iVP ridges 7)2, M- ridges 
}r-8, ridges very high and compressed, ten points or denticles on 5th ridge. Plate xxiv.a, fig. 2, skull fragment, M' ridges 
Qji, M- ridges }2-7-};', little cement. 

Stegodon ganesa. — Plate xxii, fig. 2, fine large skull, AP ridges 10}^, small talon in front. Plate xxiv, fig. 1, upper jaw, 
M' ridges 6)2; fig. 2, upper jaw, (?) M- ridges 7; fig. 3, upper jaw, M^ ridges 7)j; fig. 4, upper jaw, M' ridges &}i, last ridge 
with 7 denticles; fig. 5, imperfect upper molar — 6 ridges with 9 denticles on fourth ridge. 

Lower Jaws and Grinding Teeth 

Stegodon insignis. — Plate xviii.a, figs. 3, 3a, lower jaw, M3 with 12)2 ridges, 9 denticles on the 7th ridge; figs. 4, 4a, 
lower jaw with Mi, M2, Mi with 4 ridges, M2 with 7 ridges, plates very deej); figs. 5, 5a, lower jaw, I.M3 with 11}2 ridges. 
Plate XX, fig. 6, 6a, lower jaw, AIj with 9 ridges and front and back heel; figs. 7, 7a, lower jaw, M3 with 12 or 13 ridges; 
figs. 9, 9a, lower jaw, M2 with 9,''2 ridges. Plate xx.a, fig. 6, lower jaw, M2 7 to 8 ridges. Plate xxiv.a, fig. 3, M2 [M3] ridges 12)), 
length 11.5 in. =291 mm., width 4 in. = 101 mm. Plate xxv, fig. 4, lower jaw, Mi ridges 7)2. 

Stegodon ganesa. — Plate in, fig. 7b, M3 [?M2l "appears to have consisted of eight principal ridges, with a talon ridge 
behind, and a .subordinate ridge in front. ... It bears a close resemblance to the corresponding inferior tooth of E. insignis 
in the form of the ridges, thickness of enamel, and proportion of cement. — B.M." Plate xx.a, figs. 1, la, lower jaw, Mj with ' 
5)2 ridges, M2 with 8 ridges; figs. 2, 2a, Mj with 7)2 ridges. Plate xxv, fig. 1, M3 ridges +8. Plate xxv. a, fig. 1, lower jaw, 
M3 ridges 7)2+. Plate xxix.b, figs. 2, 2a, lower jaw, Dp^ ridges 7)2. 

Skulls of S. insignis and S. ganesa 
Slegodon insignis. Plate xv. "The cranium is seen to differ remarkably from that of E. Ganesa (Plates xxi. and xxii.) 
notwithstanding that the molars of the two species agree so closely. That of E. insignis is flattened at the top, elongated from 
side to side and singularly modified, so as to bear an analogy to the cranium of Dinotherium giganteum, while that of E. Ganesa 
does not differ much from the ordinary type of the Elephants." Plate xvi, figs. 1, 2, 3. "Fig. 1. — Elephas insignis. Broken 
cranium, oblique antero-Iateral view. Left orbit, &c., gone. This head is very cubical in form, is old, very concave in front and 
vertically; teeth broken. Interval between inci.sive sheaths deep. No tusks. A black .specimen in Cautley's collection. — 
B.M." Fig. 4, skull with M', number of ridges +10)2, length of M^ 9.4 in. =239 mm. Plate xvii, figs. 1, 2, same skull as Plate 
XVI, fig. 3. Extreme length from occipital bulge to plane of molars 23 in. = 583 mm., extreme width of occiput 25.5 in. = 647 mm., 
width of naso-maxillary opening 11.3 in. =288 mm., occipital condyles to anterior end of palate 22 in. =560 mm. See also 
Plate XVII, figs. 3, 4, and Plate xviii, figs. 1, 2, 3, 4. 

Summary of Ridge Formula 
The above observations on ridge formulae include teeth in different stages of development and of attrition, hence the many discrep- 
ancies. The adult ridge formulae are indistinguishable in the two species. Consequently we may combine the ridge numbers observed by 
Falconer (in Murchison) in the following collective ridge formula of Stegodon ijisignis-ganesa. 

Stegodon insignis-ganesa: Dp 3^ Dp 47^ M Ull M 2 '^f^ M 3 -'fUrrT^- 

This collective ridge formula includes the maxima numbers only; whereas that given below by Lydekker (1886.2, p. 89, under E. 
insignis) includes both maxima and minima. In a collective ridge formula we include ascending mutations, i.e., successive geologic stag(!s 
in the evolution of the ridges, some more primitive, some more progressive, all constituting a collective species. 

Lvdekker's (1886) Comparison of Stegodon bombifuons, S. insignis, and S. Ganesa 
"Catalogue of the Fossil Mammalia in the British Museum (Natural History)," Part IV, 1886, pp. 88, 89 
Characters of E. [ = Stegodon] ganesa. — (Lydekker, 1886.2, p. 88): "The third true molars in the type cranium of this species 
contain ten ridges, and thereby agree with the corresponding teeth of E. insignis rather than of E. bombifrons, a conclusion confirmed by 
a second cranium, in which there appear to be either ten or eleven ridges in the same tooth [Footnote: 'See Rec. Geol. Surv. Ind. vol. ix. 
p. 48 (1876).']. This close resemblance between the last molar of this form and of E. insignis renders it apparently impossible to draw any 
distinction between the earlier teeth of the two forms [Footnote: 'The majority of the teeth figured in the "Fauna Antiqua Sivalensis" 
under the name of E. ganesa have the low ridge-formula of E. bombifrons {q. v.).'], and all such teeth are therefore referred to the latter. 
Falconer [Footnote: 'See "Palseontological Memoirs," vol. ii. p. 84.'] had considerable doubts as to the specific distinctness of the present 
form, and as the resemblance between the type cranium and the young cranium of E. insignis [Footnote: 'See "Fauna Antiqua Sivalen- 



874 



OSBORN: THE PROBOSCIDEA 



sis," ])!. xliii. figs. 14, 15.'] indicates that (lie two are closely relatetl, it is jiossiblo that E. gaiiem may be the male form of E. insigiiis. 
athilt cranium iloes not differ \ery witlely from the tyi)e of E. indiciis, although the frontal constriction is less marked." 



Phe 



(7.8).(7-8).C9-l 1) 



Lydekker's ridge formula of Elephas [ = Slegodon] innlgniti-ganesa, namely, Mm. [Dp] fr-7:l^-i}y M (7rfoMi^2Mn:i3)> oinit** 
the half-ridges; it also gives a higher ridge formula to M 1 and M 2. Consequently the following maximum ridge formula according 
to l.ydekker is higher than that of Falconer; this ridge formula, how^ever, is not sustained by Jjydekker's own observations on fifty-nine 
specimens of the ('autley Collection in the British Museum, many of which are described and figured by Falconer. 

Lydekker's combined ridge formula of Elephas [ = Stegodon] insignis-ganesa may be written as follows: 



Dp 2 t Dp 3 r Dp 4 i M 1 t% M 2 



U 3 H. 



Stegodon insignis-ganesa 

Falc. ner and Cautley (1846)-Osborn (1928) 
ria\ing now reviewed in detail the observations of Falconer 
and of Lydekker, we appear to be forced to the conclusion that 
Stegodon insignis and tS. gaiiesa are respectively female and male 
representatives of a single 'collective species' which first appears in 
the Lower Pleistocene, Pinjor horizon, extending into the Boulder 
Conglomerate, while specimens at present referred to the same 
species occur in the Middle [LTpper] Pleistocene Godavari, Narbada 
Allu\ium,contem]ioraneous with Elephas (Hypselephas) hysudrictis 
and I'aUeoloxodoii iianiadicus of the Codavari. 

Doubtless we are dealing with a series of ascending mutations 
which in time may be distinguished by valid constant specific or 



presence of Slegodon airdwana in the same Kendcng-Schichten or 
Trinil horizon [Middle Pleistocene] of Java with Palxoloxodon 
hi/siidrindieus, which, according to Stremme (1911) and Janensch 
(1911) is comparable to Elephas anliquus (i.e., P. naniadkus). 
See note on the fauna of the Kendeng-Schichten below. 

In summation, the phylum Stegodon insignis-ganesa of India 
and the phylum Stegodon airdwana {=javanicasy of Java repre- 
sent the last surviving Lower and Middle Pleistocene [to 
l^pper Pleistocene] members of the southern forest-browsing 
Stegodonts. These two Stegodont phyla were geologically con- 
temporaneous with the forest- and plains-browsing Loxodonts, 
and with the grazing ancestral true elephants, such as Archi- 
diskodon . 




Sectioned Molahs of Stegodon ganesa and Stegolophodon latidens 
Fig. 757. (Left) Vertical section of loctotyiie left third sui>ei'ior molar, 1 . M'', of Elephas [ = Stegodon] ganesa Falconer and Cautley, 
1846 [1845, PI. HI, fig. 7a], less than one-half natural size. Brit. Mus. M. 18489 (erroneou.sly referred by Lydekker, 188G.2, p. 84, to E. 
bombifrons). From Siwalik Hills, India. (Falconer and Cautley, 1846.1, p. 45): "The crown consists of ten principal ri<lgcs, with a 
subordinate 'talon' ridge in front and behind. The anterior seven ridges have their summits worn, the two in front being ground 
down to the common base of ivory, the tooth having been a considerable time in use." Molar inverted to show natural position. 

(Right) Vertical section of two last upper molars of the right side, r.M^"', of Mastodon {= Stegolophodon] laiidens Clift, after 
Falconer and Cautley, 1846 [1845, PI. ill, fig. 8], le.ss than one-half natural size. Same molars as those figured by Clift in 1828, PI. 
XXXVII, fig. 1. From Iriawaddy River, Burma. Compare figure 719 above. 



subspecific characters; meanwhile we regard these ascending 
mutations as a collective sjiecies Stegodon insignis-ganesa. The 
collective ridge formula is probably as follows, indicating as 
iiiiniiiKi the ridge-crests characteristic of the more primitive stages 
and as maxima the ridge-crests characteri.stic of the more progres- 
sive stages so far as known: 

S. insignis-ganesa : Dp 2 i Dp 3 
M2^^tir-M3 «■"■** 



Dp 4 ^^„ M 1 



7H- S 
7 W- 1 



M-9.V', ^" " w-1 i-H-iz-iBM-is- 
The survival in India of referred Slegodon insignis-ganesa in 
the same Middle [Upper] Pleistocene le\cls with the true elejihants 
and Loxodonts, such as Elephas (Hypselephas) hysudrieiis and 
Palseoloxodon namadieus, is j)aralleled in the East Indies by the 



Stegodon insignis birmanicus Osborn, 1929 
Figures 758, 760 

Mingoon, opposite Mandalaj', Burma; upper levels of the Irrawaddy 
Series, Upper Pliocene." 

This left ramus (Amer. Mus. 20002), collected by Barnum 
Brown ill 1922, agrcM's with Stegodon insignis in the number of the 
ridge-crcsis of the third molar hut greatly exceeds this typical 
Siwalik species in size; the jaw and inferior grinder appear to be 
one of the largest of the Stegodonts heretofore described (Figs. 758 
760) ; the tooth indicates a higher degree of siiecialization than the 
tyjiical S. insignis-ganesa and a specialization different from that of 
>S. orientaiis grangeri; f he S. insign is birnian ints molar is of gigant ic 



'[Synonym of Stegodon airawaiia or S. trigonncephalus (see footnote on page 889 below). — Editor.] 

'[See note on page 824 above regarding the Lower Pleistocene age of the upper levels of the Irrawaddy Scries. — Editor.) 



THE STEGODONTIN^: STEGODON 



size, there are many ridge-crests, far apait, with a mciUuni iiumtjer 
of conelets, and considerable cement; the ,S. nrieiitalifi (jranqeri 
molar is of medium to large size, there are many ridgo-crests, 
ap])roximated, manj' conelets, and modei'ate de\elopment of 
cement. 

Stegodon insignis birmanicus Osborn, 1929. "New Eurasiatic 



second pair, are partly worn, ridge-crests slightly open; this is 
followed Ijy ridge-crests pentalophid to octalophid (5th to 8th), 
jiartly worn; conelets still separate; the ninth to twelfth ridge- 
crests are entirely unworn. The conelets are few and stout; they 
\ary in number from four to twehe on each ridge-crest. Cement 
is present all the way back. 




Fig. 758. Type of Slegodon insignis birmanicus Osborn, 1929 (Amcr. Miis. 20002), a twelve aiul a lialf erested third left 
inferior molar, l.Ms, in the same ridge-crest stage as Slegodon insigrtis-ganesa. One-tliird natural size, t^rom Mingoon, opposite 
Mandalay, Burma. 



and American Proboscideans." Amer. Mus. Novitates, No. 393, 
Dec. 24, 1929, pp. 15, 16. Horizon and Locality. — Upper 

Pliocene [now regarded as of Lower Pleistocene age]. Mingoon, 
opposite Mandalay, Burma. Type Figure. — Op. cil., 1929.797, 
p. 16, fig. 16. 

Type. — Amer. Mus. 20002, a very large and massive left 
inferior jaw containing the left third inferior molar, I.M3. The 
ridge-crests of I.M3, namely, 7^775. are the same in number as in 
Slegodon insigtds, but the elongation of this inferior molar and the 
open character of the ridge-crests are quite distinctive from S. 
insignis; the jaw is more massive and the inferior grinding teeth 
surpass in length measurement those of any other Stegodont type 
known; the grinders are larger and exhibit fewer conelets. The 
principal measurements (in millimeters) are as follows: 



Total length of third inferior molar, I.M3 

At the crown 

At the base 
Total maximum breadth of I.M3 
Breadth-length index of I.M3 
Length of mandibular ramus from base of 1 

to tip of symphysis 
Horizontal thickness of mandibidar ramus 
Height of ramus at symphysis 



M3 



333 
362 
97 
29^27 

560 
204 

208 



In detail, the entire jaw and posterior grinder are very mas- 
sive. Of the total of 12-)2 ridge-crests, the anterior pair are deeply 
worn and widely open; the tritolophid and tetartolophid, or 



Stegodon orientalis grangeri ()sl)oni, 1929 
Figures 682, 684, 686, 687, 731, 759-764, 777, 1231, PI. xx 

Upper Pliocene [Lower Pleistocene] of Yenchingkou, neai- Wanhsien, 
Province of Szechuan, China. 

The subspecies Stegodon orientalis gmngeri, as shown in fig- 
ure 687 is more primitive than the type of S. orientalis, also 
from a cave in Szechuan; the ridge-crests are less elevated and 
wider apart at the base and seem to be even more primitive than 
those of the iS'. insignis type (Figs. 688, 747) ; the cranium is much 
smaller and simpler than that of S. ins/gnis-ganesn (Figs. 735, 736) 
and resembles in its contour rather that of .S'. bomhifrons (Figs. 
732, 734, 777). 

Stegodon orientalis grungeri (_)sborn, 1929. "New Eurasiatic 
and American Proboscideans." Amer. Mus'. Novitates, No. 393, 
Dec. 24, 1929, pp. 16 and 17. Type.— Amer. Mus. 18714, 

a left third superior molar, l.M', and right and left third inferior 
molars, r.Ms, I.M3, of the same individual. Horizon and 

Locality'. — Upper Pliocene [now regarded as of Lower Pleistocene 
age]. Yenchingkou, near Wanhsien, Province of Szechuan, 
China. Type Figure. — Op. cit., Osborn, 1929.797, p. 16, fig. 16. 

Type and Referred Figures. — Aside from the type (Figs. 
759, 760, 762) there is very abundant referred material from the 
same locality, including the complete superior and inferior den- 
tition, Dp 2-Dp 4, M 1-M 3, in various stages of attrition and 
dental succession, as illustrated in figures 687, 759, 763, 761, 762 of 
the present Memoir, constituting the most complete material of 
dental succession known in the fossil Proboscidea. 



STEQODON ORIENTALIS QRANQERI 
PERMANENT MOLARS 



y^ /VZ /S636 





■ A./^. /a7/4- Type 




STEqODON 0RIENTALI5 QRANQERI 
MILK DENTITION -^ naT. si'je 



A.M. /8705 




i. dp 2 I. c^pj 




Fig. 7o9. Type (Amcr. Mus. 18714) and referred specimens (Amer. Mus. 18705, 18711, 18630a, 18636, 18642) of Slegodoii nrifnlnlis grangcri, part of the 
collection from Yenchingkou, Province of Szechuan, China, made by Walter Granger during the winter of 1920-1921. All figures one-fourtli luitunil size. (Cf. 
Figs. 761 and 702 for crown view.s, giving ridge-crests and conclets.) The specimens hearing numher 18642 have hccn sent to Peking, China, in cxcliiiMgc. 



Left lateral aspect of: 

R.Dp2, r.Dp' (rev.)— I.Dp2, l.Dp,, (Amer. Mus. 18705) 

L.Dp', l.Dp" (Amer. Mus. 18711), and l.Dp4 (Amer. Mus. 18630n) 

L.M', r.Mi (rev.) (Amer. Mus. 18636) 

L.M^, also r.Ma (rev.) (Amer. Mus. 18642) 

L.M^ I.Ms (Amer. Mus. 18714)— type. 



Crown view of: 

Type third left inferior molar, I.M.i, strongly concavo-convex, out- 
wardly arched. Amer. Mus. 18714. 
Type third left superior niolai', l.M'', strongly convex, parallel sides. 
Amor. Mils. 18714. 



876 



THE STEGODONTIN^: STEGODON 



877 



Materials of Stegodon orientalis grangeri, Type and 
Referred. — Collected by Walter (iranger in 1920-1921, all from 
the same pit; estimated specimens: 

Separate crania, more or less complete 6 
Separate crania, more or less fragmentary 7 
Right and left mandibular rami and complete in- 
ferior jaws 28 
Separate superior and inferior grinding teeth in 

various stages of succession 49 

Skeletal bones, very few, not exceeding 3-4 

Characters. — An ascending mutation or a subspecific stage 
regarded as somewhat more primitive than the type of Stegodon 



orientalis Owen (Fig. 769) which is recorded from a cave in the 
Province of Szechuan, northwest China, and much more progres- 
sive than Owen's type of Stegodon sinensis which appears to be 
comparable to S. bombifrons. The ridge-crest formula, beautifully 
shown in figure 759, is as follows: 



Dp2^^Dp3^, Dp4 



5i-6-V» 



Ml 



M-6-^s 



M2 



V4-8-W 



11-W 



(7)-9» J^l 3 jj.f 3 



Comparison with other Chinese Stegodonts. — On closely 
comparing Owen's types of Stegodon sinensis (Figs. 687, 702) and 
S. orientalis (Figs. 687, 769) with the Stegodon orientalis grangeri 
teeth (Figs. 687, 759, 761) collected by Doctor Granger at Yen- 
chingkou, it is certain that we have to do with three distinct spe- 
cific and subspecific stages broadly distingui.shed as follows: 



Upper Miocene [Pliocene] stage. 
Owen's type of 
Stegodon sinensis 
Brachyodont. 

Lophs compressed, far apart. 
Compare Stegodon bombifrons. 



Lower Pleistocene stage. 

Osborn's type of 
Stegodon orientalis grangeri 
Subhypsodont (first stage). 
Lophs more compressed, more elevated. 
Compare the typical Stegodon orientalis. 



(?) Lower Pleistocene stage. 
Owen's type of 
Stegodon orientalis 
Subhypsodont (second stage). 
Lophs still more compressed, still more 
elevated. 
Compare the referred Stegodon orien- 
talis and <S. airdwana. 




Amer. Mils /a7IA- 7y/>e 
STEqODON ORIENTALIS QRAN5ERI 



Tier y%e^ 
I. mi 



/3 ■naZ. stje 




Amer Mus 20002' Type' 
STEqODON INSIQNIS BIRMANICUS 




Arner. f^us. /SS69 J^e/r^^^^ ^ .^^; ; J-" ITtner VLen/ 

STECODON INSIQNIS - GANESA ^~--<=s^^i,«^__»£...— '-""'^ ^. „<.^ 

COMPARLSON OF StEGODON INSIGNIS-GANESA, S. INSIQNIS BIRMANICUS, AND S. ORIENTALIS GRANGERI 

Fig. 7t)0. Left third inferior molars, to a one-third scale, from the internal aspect, showing the eruption line of the grinder and exposing the six 
anterior ridge-crests in Stegodon orientalis grangeri (ujiper) and S. insignis ganesa (lower). After Osborn, 1929.797, p. 16, fig. 16. 

Stegodon orientalis grangeri, type (Amer. Mus. 18714). Large size with 13-|- ridge-crests of which 6+ anterior were erupted while 7-13 were 
buried in the jaw. 

Stegodon insignis birmanicus, type (Amer. Mus. 20002), with 5-6 ridge-crests erui)ted, 6-12 buried in the jaw and encased in cement (dotted). 

Stegodon insignis-ganesa ref. (Amer. Mus. 19869), with -|-l-6 ridge-crests erupted, 7-11 buried in the jaw, mostly lacking cement. 

This comparative figure illustrates the gigantic size attained; the largest {S. insignis birmanicus) slightly exceeds in length the correspond- 
ing grinder of »S. orientalis grangeri. 









<5 



mm 



Fig. 761. Superior and inferior second, third, and fourtli premolars and first molar; (Dp 2— M 1 ) of Slrgmhn nriinlnlis gmiigeri (Amcr. Mus. 
18636, 18711, 18630a, 1870.5), side and crown views, one-half natural size (ef. Fig. 7.59). 

(Upper pair). First superior and inferior molars (Amer. Mus. 18636), l.M> with +6+ ridgc-erests, conelets 10 niaxinmni, r.M, (rev.) 
with 8 ridge-crests, conelets 9. 

(Middle pair). Fourth superior and inferior deciduous premolars, l.Dp^ {Xim-T. Mus. 18711) with }.,&-}■: ridge-crests, conelets 12 13, 1.1)))4 
(.\mer. Mus. 18630a) with 7-)3 ridge-crests, conelets 12 maximum. ^ _ 

(Lower pair). .Second and third superior and inferior deciduous i)remolars (Amer. Mus. 1S70,')), r.Dp- (rev.) with +3+ ridge-crests, r. Dp 
(rev.) with 6+ ridge-crests, conelets 15 maximum, l.Dp2 with 2 ridge-crests, l.Dps with 6 1- ridge-crests, conelets 16 maximum. 

878 




Fig. 762. Type and referred .superior and inferior molars of Stegodon onentalis grangeri. Crown views (cf. Fig. 7.59). After 
retouched pliotographs, one-half natural size. 

(Upper pair) Type (Amer. Mus. 18714), left third superior molar, l.M^ with -I-IIJ2 ridge-crests. Left third inferior molar, I.M3, 
with Yr-M ridge-crests; inner side strongly convex, outer side strongly concave, in contrast to the parallel sides of the upper molar. 

(Lower pair) Amer. Mus. 18642. Left second superior molar, l.M^, with -f-8-f- ridge-crests. Right second inferior molar, r.M2 (rev.) 
with 9-1- ridge-crests. These molars have been sent to Peking, China, in exchange. 

879 



880 



OSBORN: THE PROBOSCIDEA 



CRANIAL CHARACTERS OF STEGODON ORIENTALIS GRANGERI 
(Figs. 7.')9, 701, 703, 777) 

The iiifiiiililc craniuin of Slcgndoii oricidalis graiigcri (Amer. 
Mvis. 18638, 18702), as represented in the composition drawing 
(Fig. 763), exhibits the small deciduous superior incisors and the 
deciduous premolars, Dp 2-4 (for crown views of these deciduous 



to lower grinders, M1.2, with ridge-crests 8 and 9 respectively (cf. 
Figs. 759, 762 lower); these ridge-crests, partly bathed in cement, 
of moderate hypsodonty, exhibit the characteristic convex coronal 
curvatiue of the upper ridge-crests, the concave coronal curvature 
of the lower ridge-crests, a feature much more strongly marked in 
the third molars (Fig. 759). In the inalure adult cranium (Amer. 



STEqOOON ORIENTALIS qPANCqERI 




/)// -g nat.sije 



Fig. 703. Infantile, juvenile, yoinig adult, mature adult crania showing eruption of Dp'— M', Dps — M3. Yenehingkou, Provinee of Szechuan, China. 
Infantile cranium (Amer. Mus. 18038) showing succession of Dp" — Dp'', also Di', together with j.aw (Amer. Mus. 18640) belonging to tliis sUull, showing 
Dp2— Dpi. 

Juvenile cranium (.\mer. Mus. 18702), jaw (Amer. Mus. 18711), showing eruption of Dp3, Dp4. 

Young adult cranium (Amer. Mus. 18630) showing M' and M^. Jaw (Amer. Mus. 18630) with Mj in xilu, .sent to British Museum. 

Mature adult cranium (Amer. Mus. 18708), jaw (Amer. Mus. 18029 — .sent abroad in exchange), showing M^, M', and M3 in silu. 



premolars, compare Fig. 761); the rounded jirofile, of subtriangu- 
lar form, should be compared with the infantile cranium of Mas- 
todon acutidens (Fig. 131) and that oi Elephas indicus (Fig. 799). 

The juvenile cranium (.\mer. Mus. 18702, sktdl, 18711, jaw — 
a composition drawing) has lost the deciduous incisors, also Dj) 2, 
but retains the functional Dp 3-Dii 4 (cf. Figs. 759 and 761). 

The young adult cranium (Amer. Mus. 1S630, skidl, 18636, 
jaw) exhibits the rounded and greatly elevated dome; of the l)rain 
case; all the deciduous teeth, Di-, Dp 2-4, have disappeared and I lie 
grinding function is now assumed by the first and second true 
molars, M 1 -M 2 (cf. Figs. 761, 763). It would a|)pear that 
the grinders M''-, with ridge-crests 6 and 8 respectively, are opposed 



Mus. 18708, .skull, 18629 jaw) the second superior molar, AP, is 
disappearing, while the third superior and inferior molars, M', 
M3, are fully functional, their reciprocal convexo-concave relations 
being beautifully shown in figure 763 (right). The upper jiortion 
of this mature adult cranium is fractured, leaving tlie exact 
profile contoui- in doubt. 

The third superior and inferior molars of this subspecies 
grandly r('|)resented in the li/pc si)ocimen (Amer. Mus. 18714) seen 
in both lateral and crown \iews in figures 759 and 762, add a great 
deal to our knowledge of the Stegodont dentition and show 
\ery marked mechanical (-ontrasts between the supeiior and in- 
ferioi' grinders as follows: 



THE STEGODONTINJ!;: STEGODON 



881 



Third SurEiuon Molaus 
], literal profile strongly convex. 

Posterior ridge-plates elongated, partly <jr wholly bathed 
in cement. 

Sides of crown parallel. 

Third Ixferior Molars 

l.ateral profile horizontal. 

Anterior and posterior ridge-plates of uniform height; 
cement chiefly in median plates. 

Sides of crown strongly arched, convex inner side, concaK 
outer side. 



A Acry careful study of these four ontogenetic stages of the 
cranium, namely, infantile, juvenile, young adult, and mature 
adult, reveals resemblances and contrasts with the Mastodon and 
Elephas crania which will probably prove of phyletic or generic 
value. Similarly, the remarkable ridge-crest succession, the 
mechanical warping of the superior and inferior grinders, the 
gradual increase in height (hypsodonty) of the superior ridge-crests 
from Dp 2-M 3, indicate that the highest and most progressive 
ridge-crests are reserved for the extremely adult stages in which the 
posterior ridge-crests of the eleven and a half crested third superior 
molar grind against the less hypsodont ridge-crests of the thirteen 
crested third inferior molar. 



STEGODON PINJORENSIS 







STEGODON mSIQNlS 
Br/t Mus. M.30/5 Type 




STEGOOON AIRAWANA 





cal 



Fig. 704. Cuni|)a.r!itivc sections of tliird left superior molars of .Stcgodonts of 
India, Java, and China, all to a one-third scale. 

A, Stegodnti pinjorcnsis Osborn, type. A 14-|- ridge-crested I.M^, from near 
Siswan, India. Amer. Mus. 19772. 

B, Sicgoclon insifjnis Falconer and Cautley, lectotype. All ridge-erest,etl 
l.M^ from the Siwalik Hills, India. Alter Falconer and Cautley, 1840 [184."), 
PI. 11, fig. Oal. Brit. Mus. M.3015. 

C, Stcgodon airawana Martin, rcf. Portion of an l.M' of the Middle Pleisto- 
cene of Trinil, Java, exiiibiting 7 of the 12-14 ridge-crests; much more recent in 
structure than the molars of Stcgodon. insignis of the Lower Pleisto<^ene. This 
tooth was presented to the Amei'ican Museum by the Geological-Palaeontologi- 

Tnstitutc and Museum of the University of Berlin, Gi^nnany, through the courtesy of Geheimrat Pompeckj and Doctor Dietrich. Amer. Mus. 22G30. 
D, Skgodon oriculalis grangcri Osborn, typ(^ An l.M^ with +11 }■> ridge-(Te.sts, from near Wanhsien, Province of S?;echuaii, China. Amer. Mus. 1S7I 1. 




STEGODON QANESA Ref. r Rev. > 
A-fter Thlc.oner 



Fig. 766. Referred lower jaw and right superior tusk (A) of Slcgodon ganesa (Amur. Mu.s. 19773), collected by Barnuni Brown in 1923 in the Upper 
Siwaliks, below the conglomerates, three miles north of Siswan, India, compared with skull and tusks (B), after Falconer, 1846 [1817, PI. .xxii, fig. 3]. 
Original in British Museum (Brit. Mus. M.3008). Both figures one-sixteenth natural size. 

882 



THE STEGODONTIN^: STEGODON 



883 



Stegodon pinjorensis Osborn, 1929 
Figures 711, 731, 704, 765, 707, 70S, 777, 817, 1217, PI. xx 

"Three miles nortli of Siswan, India." Upper levels of the Pinjor horizon, 
Lower Pleistocene. 

Slegodon pin jure nsis Osborn, 1929. "New Eurasiatic and 
American Proboscideans." Amer. Mus. Novitates, No. 393, p. 18. 
Type. — "Amer. Mus. 19772. A male cranium, rostrum wanting; 
portions of right inferior tusk preserved." Horizon and 

Locality. — "Recorded by Barnum Brown as follows: 'Skull. 
.Just below Conglomerate beds, Upper Siwaliks, three miles north 
of Siswan, India.' This progi-essive cranium does not belong in the 
Pinjor (I'pper Pliocene) horizon, as the specific name pinjorensis 
suggests, but was probably deposited from the overlying Boulder 
Conglomerate beds of Lower Pleistocene age." Type 

Figure. — Osborn, 1929.797, p. 17, fig. 17. 

Type Description. — "Superior grinding teeth distinguished 
from those of Slegodon insignis-ganesa bj' their superior size, 
nuich more numerous ridge-crests, progressive hypsodonty; the 
comparative ridge formula of M 3 are as follows: 

Slegodon pinjorensis: M 3 ^^^"^^ 
Slegodon insignis birmanicus: M 3 YzTi 
Slegodon insignis-ganesa: M 3 lilsfj't 

This type male cranium resembles that of the male type of 
Slegodon ganesa Falc, namely, with small rounded parieto- 



occipital crest, lofty and greatly abbreviated frontonasal surface, 
anterior nares correspondingly elevated, grinding surface of the 
large molars very strongly arched, but the cranium is relatively 
more depressed or bathycephalic than in S. ganesa." 



^5 A/atural 5tj& 




A. 



S. PINJORENSIS Type A. At /9772 
S. PANES* Tfef /l^ter rbZconer 



Fig. 707. Tyiic skull (Amer. Mus. 19772) of Slegodon pinjorensis (A), 
also skull of same individual (A) superimposed (dotted lines) on referred 
skull of <S. ganesa, after Falconer (B). Onc-sixtecntli natural size. 

Observe the extreme bathyroplialy in the skull of Slegodon ganesa in 
comparison with the more typical vStegodont form, as exemplified by the 
skull of S. pinjorensis. 








Fig. 768. Front view of Slegodon pinjorensis before present arrangement 
of the tusks, which now are regarded as possibly turning inward, as in new 
restoration of Slegodon ganesa (see Fig. 733, p. 857 above); otherwise head 
true to present proportions. Forelimbs entirely conjectural, drawn in pro- 
portion to size of head. Ungues 5 and 5. Restoration by Margret Flinscli, 
May, 1930. One-fiftieth natural size. 



884 



OSBORN: THE PROBOSCIDEA 



Stegodon orientalis Owen, 1870 
Figure 087, 703. 7ti9, 770 
Szechuaii, iiortliwcst China, near the city of Clmngkiiigfoo. Probiiljly 
Lower Pleistocene. Swinhoe Collection of 1870. 

The imperfectly known Stegodon orientalis, as shown in com- 
parison of figures 687, 688, 697, and 769, is apparently more 
progressive, with more elevated ridge-crests, than the type of >S'. 
insignis (Fig. 697); the ridge-crests (Fig. 769) are more acute and 
more compressed at the base, and the summits are more closely 
approximated; with numerous conelets and strong layers of 
cement in the valleys. It is, however, impossible clearly to define 
and separate this species until further material is found in the 
type locality. It is somewhat more progressive than S. orientalis 
gr anger i. 

This species described by Owen from an imperfect type figure 
(Fig. 769) is regarded by Lydekker (1886.2, p. 97) as a synonym of 




city of Chung-king-foo, in the province of Sze-chuen.' " Type 

Figure. — Op. cit., PI. xxviii, figs. 1-4. 

Type Desckiption. — {Op. cit., p. 421): "The dentine retains 
its original white colour, . . . the enamel also has its recent pearly 
tint; a thick mass of cement appears to have been retained in the 
intervals of the coronal ridges. One of these ridges, with the con- 
tiguous halves of two others, form a molar two inches nine lines in 
breadth (PI. xxviii. figs. 1 & 2); a portion of a posterior ridge 
with a low basal heel, from the same, or the same-sized tooth, and 
the last two ridges, with a terminal half ridge or talon, of a milk- 
molar, one inch and a half in breadth (ib. figs. 3 & 4), represent the 
present species. . . . The condition of the fragments agrees with 
the statement, viz. that they were from a cavern. . . . The ridge 
(ib. fig. 1) a a runs straight, or nearly so, across the tooth; the 
entire ridge is cleft at the summit into about a dozen mamillce 
by as many vertical grooves ; the dentine rises into the base of each 

3 





I'"ig. 769. Type oi Stegodon orientalis Owen, 1870, from a cavern in Szechuan, northwest China, .^fter Owen, 1870, PI. xxviii, figs. 14. 
Original in the British Museum (41926-7). (Op. cit., p. 433, fig. 1): "Portion of true molar, grinding-surfacc." Fig. 2. Same, "side vi(-w." 
Fig. 3. "Hind end of milk-molar, d 3, grinding-surface." Fig. 4. Same, "side view." 

Locality. — (Op. cit., p. 421): "These fragments form part of the series of teeth obtained by Mr. Swinhoe, and said to be 'from a cave, near 
the city of Chimg-king-foo, in the (jrovincc of Szc-chucn.' The condition of the fragments agrees witli tlie statement, viz. tliat they were from 
a cavern. . . . [)). 434] Mr. H. Woodward stated tliat Mr. Swinhoe liad himself obtained a .series of these fossils from a cave many miles in- 
land — he believed, on the course of the Yang-tse-kiang." 



Stegodon insignis (.sec citation below). Koken (1885) also regarded 
5. orientalis as a synonym of S. insignis, Schlosser (1903) agreeing 
with Koken. In Osborn's opinion, the imperfect character of the 
type molar (Fig. 769) and associated milk molar renders it tlifficult 
to determine whether these sjiecimens are in the Stegodon insignis 
stage; in both specimens the lophs are elevated. From another 
cave locality, 140 miles distant, near Yenchingkou, comes the 
superb material collected by (iranger, which proves to be somewiiat 
more priniiti\e than Owen's type of Stegodon orientalis ant! is 
described as Stegodon oricidalis grangeri (see p. 875 above). 
Owen's lyi)e description is in jjart as follows: 
Stegodon orientalis Owen, 1870. "On Fossil i{einains of Mam- 
mals found in China." (Juart. Journ. Cleol. Soc. ]>ondon, \'ol. 
XXVI, Pt. 1, p. 421. Type.— Molar fragments (Brit. Mus. 

41926-7). IIoiuzo.N' and L()c,\ijtv. — (Op. cil., ]i. 421): ". . . 

obtained by Mr. Swinhoe, and said to be 'from a ca\c, near the 



mamilla. The enamel (e) averages two lines in thickness." 

"From the above-defined characters it is plain that we have 
here, also, parts of a 'transitional Mastodon,' in other words, 
a species of Stegodon, Fr. In the straight, or nearly straight, 
direction of the coronal ridges, and the absence of any trace of 
mid cleft, these molar fragments more resemble Ihe (oeth of 
Stegodon Cliftii, St. insignis, and St. ganesa of Falconer than does 
the St. sinensis; and in the apparent quantity of coronal cement 
(ib. fig. 2r) as well as in the evidence of a hinder talon (ib. fig. 3 1), 
they are more like St. insignis than St. Cliflii. Yet the two hinder 
ridges, with the terminal talon of the tooth (ib. figs. 3 & 4), 
which, in breadth, corresjionds with the second vi])pcr deciduous 
molar of .S7. insignis and .S7. sinensis, clearly differ from both. 
The last two ridges run straighter across, are of the same extent, 
and are divider! by more ntmierous xerticnl grooves into smaller 
and correspondingly luuncrous apical maiiiilhe. The second of 



THE STEGODONTIN^: STEGODON 



885 



these ridges is cleft in tho middle. From the alleged conditions of 
discovery, and the little-altered condition of the above-described 
portions of proboscidian molars, one would be led to deem them 
to be of as comparatively recent geological age as our ordinary 
British Cave-fossils." 

LYDEKKER'S NOTES OF 1886 ON STEGODON ORIENTALIS 
"Catalogue of the Fossil Mammalia in the British Museum (Natural History)," 

Ft. IV, 1886, p. 97 

Lydekker (1886.2, p. 97) treated this species as follows: 
"[Brit. Mus.] 41926-7. The last two ridges and talon of an un- 
worn fourth lower milk-molar and portions of two other cheek- 



Ste^odon airawana Martin, 1890 
Figures 686, 688, 707, 731, 764, 771, 773, 774, 777, 779, PL xx 

Kendeng-Schichten, Pilhecanthropxis erectus zone, Middle(?) Pleistocene, 
Alas-Tuwa, Trinil, Java. 

This is the most progressive Stegodont known, surpassing 
Stegodon orientalis and greatly surpassing S. insignis-gancsa in the 
elevation (hypsodonty) and number of the ridge-crests. As shown 
in figures 687 and 688, the fourteen to fifteen ridge-crests of the 
third right inferior molar, r.Mj, are almost columnar in side \iew 
and much more elevated than in »S. orientalis (Figs. 687, 779, and 
769). In front view the cranium (Figs. 77.3, 777) resembles that of 




10 +i 



Fig. 770. Referred Stegodon insigxis(?) = orientalis(?) and Type of Serri- 

DENTINUS LTDEKKERI. AFTER ScHLOSSER, 1903, TaF. XIV, FlGS. 7-10 

Fig. 7. Tooth referred by Schlosser to Mastodon a£f. latidens Clift, from the 
rothe Thone = Schansi; fig. 10, a left inferior molar, VMs, from ?Fokien, referred by 
Schlosser to Stegodon insignis. 

Fig. 8. Type of Mastodon [ = Scrridentinus] hjdekkeri Schlosser, from the rothliche 
Sande = Tientsin, Honan, north China, presumably a left M' (cast Amer. Mus. 10374); 
fig. 9, supposed inferior premolar referred by Schlosser to Mastodon lydetcheri [ =Serri- 
dentinus lydekkeri of the present Memoir). 



teeth, provisionally referred to the present species. These 
specimens, which were obtained from a cavern in Sechuen, north- 
west China, are the types of Owen's Elephas (Stegodon) orientalis, 
and are described and figured by him under that name in the 
(^uart. .lourn. fieol. Soc. vol. xxvi. pi. x.xviii. figs. 1-4. They show, 
however, no characters by which they can be distinguished from 
the teeth of the present species [i.e., Stegodon insignis], as the 
writer has already observed in the 'Pateontologia Indica,' ser. 10, 
vu\. i. p. 269. P urchased from R. Swinhoe, Esq., 1870." 



S. trigonocephalus (Fig. 776) and differs widely from that of S. 
insignis-ganesa or >§. bombifrons. This is a dwarfed insular form, 
very progressive in molar structure. 

History. — This species was named by Martin in 1890 and 
regarded by Janensch (compare Janensch, 1911, p. 171, fig. 12, 
also Taf. xxiii, fig. 3) as differing in skull structure both from 
Stegodon insignis and S. ganesa while similar in jaw structure. 
The assigned ridge formula is: Dp 3* Dp 4- M l-M2-^M3^y^. 
Dubois in his Trinil-I-'auna (1908, p. 1256) observes that this 



886 



OSBORN: THE PROBOSCIDEA 



species of Stegodon is highly characteristic of the Trinil Pithecan- 
thropus beds, also that the third lower molar rises to fourteen 
ridges, i.e., M 3 tt, and is more progressive than that of insignis. 
This would tend to place Pithecanthropus erectus as of Middle 
Pleistocene age. Dietrich (letter, March 10, 1924, and notes), 
as cited above in this ^Memoir (p. 813), on morphological grounds, 
regards this Javanese species as more recent than any of the 
known continental sjiecies of Stegodonts. 

Stegodon Airdwnna Martin, 1890. "Ueber Neue Stegodon- 
Reste Aus Java," Verh. Kon. Akad. Wetensch. Afdeel Natuurk., 
.Amsterdam, Deel XXVIII, p. 4, Type. — Incomplete man- 

dible with thirtl molar of either side in place. IIouizon and 

Locality. — Kendeng-Schichten, Pithecanthropus erectus zone, 
:Middle(?) Pleistocene, Alas-Tuwa, Trinil, Java. Type 

FiGiHE. — Martin, 1890, Tab. i, figs. 1 and 2 (mandible), also 
Tab. 11, figs. 3 and 4 (type Ms). 

Type Description. — {Op. cit., p. 4) : "Nur von Java in einer 
unvollstiindigen Mandibel und darin steckenden ]\Iolaren bekannt. 
Letztere mit 9 Jochen und 2 Talons. Durch den niastodonartigen 
C'harakter der Kronenspalte und die geringen Cementmengen 
schliesst sich die Art an .S7. Cliftii und St. bombifrons, durch die 
hoheren Joche an St. in.signis und St. ganesa an; sie nimmt somit, 
gleich dem ebenfalls auf Java beschriinkten St. trigonocephalus 
[Footnote: 'Vg\. Sammlgn. Ser. I, Bd. 4, pag. 102.'], eine Mittel- 
stellung zwischen beiden Gruppen ein." 

Janensch, 1911 ("Die Proboscidier-Schiidel der Trinil-Expedi- 
tions-Sammlung") concludes his detailed description of the skull, 
dentition, and skeleton of Stegodon airdwana with the statement 
(p. 192) that while Stegodon insignis and .S'. ganesa are closest in 
their dentition to S. airdwana, yet in certain details they are some- 
what more primitive; also in skull structure S. airdwana differs 
both from S. insignis and S. ganesa, while in jaw structure the 
three species are similar. 

Stremme ("Die Saugetiere mit Ausnahme der Proboscidier," 
1911, J). 143) observes: "Stegodon. Airdwana Mart, ist iiach 
Janensch mit Stegodon ganesa und Stegodon insignis am niichsten 
vcrwandt, die im indischen PliocJin und Plcistociin vorkonunen." 
Janensch (1911, p. 192) determines the Trinil Stegodon as Stegodon 
airdwana Martin, and observes regarding the ridge formula (op. 
cit.. 1). 187): "Die Jochforniel [Footnote: 'In dieser Formel be- 
deutct X den Talon.'] lautet, sowcit bis jctzt bekannt: 



Mm^ 



M 



X 7XX1IXX1 1-1 ax 

XI 3X 



l3ubois gibt iieucrdings (Trinil-Fauna S. 125t]) an, class die 
Zahl der Joche dei- letzten untcrcn Molaren bei dem Stegodon von 
Trinil bis zu wenig.stens 14 gehe, doch ist aus seiner Angabe nicht 
zu ersehen, ob er die Talons etwa mitzahlt." He also observes that 
this si)ecies of Stegodon is highly characteristic of the Trinil 
Pithecanthropus beds. The other Javan species, Stegodon trigono- 
cephalus Martin, does not occur at Trinil. 

Osborn, 1922: Our knowledge of this very progressive Javan 
type has been greatly extended by the researches of Janensch 
(1911) on the rich collection discovered by the Sclcnka-l^lancken- 
li(ini I'Apeditioii in the Trinil beds in which tiie type of Pithecan- 
thropus erectus occuis. The fauna of these i)eds, as desciibed by 
Strenune (1911, |ip. 82-150) includes (pj). 141, 142) the following. 



STEGODON AIRAWANA FAUNA OF THE KENDENG-SCHICHTEN 
LAYER, JAVA, AS LISTED BY STREMME, 1911 

This fauna was first regarded as of Lower Pleistocene age 
( = Boulder Conglomerate zone of Pilgrim), or transitional to 
Middle [Upper] Pleistocene ( = Godavari Alluvium, Nerbudda of 
Pilgrim) which contains Stegodon insignis ref., S. ganesa ref., and 
Palxoloxodon namadicus. Stremme remarks (1911, p. 144): "Ein 
wichtiges Leitfossil wiire eventuell Elephas, dessen Zahnbruchstiick 
Janensch dem Elcphas antiquus am niichsten stellt. Das Stiick 




Fifj;. 77L Typo of Stegodon Airawann Martin, ISiX), 'Pah. i, fifjs. 1 ami 
2, oiic-fuurth natural size. The type rif;lit inferior molar, r.Mj, is al.so figured 
in Tab. ii, figs. 3 and 4, two-thirds natural size. (Martin, op. eil., p. 4): 
"Nur von Java in einer unvollstiindigen Mandibel und darin steckenden 
Molaren bekaiuil. Letztere mit .loclien und 2 Talons." 



THE STEGODONTIN^: STEGODON 



887 



stammt nicht von Trinil; auch Dubois hat keine Elephas-Heste 
von Trinil in seiner grossen Saiiimlung." 



Primates : 



Ungulata : 



Artiodactyla : 



Rodentia : 
Edentata : 
Carnivora : 



Pilhccanthropus crcdus Dubois. 

Alacacus ncincstrinua mradana, related to exist- 
ing Zati of Sumatra and Ncmcutriuus of 
Borneo. 

Stcgodon gancsa javanicus Dubois = [>S7c(;ti(/(»/( 
Airdwana or .S'. trigonocephalus Martin]. 

Elephas hysiidrindicus Dubois = Elephas sp. an- 
iiquus [namadicus] Falconer (Jide Stremme, 
1911, and Janensch, 1911). 

[It is important to note that E. hysudriridicus 
does not occur at the locality of Trinil.) 

Rhinoceros sivasondaicus Dubois, intermediate 

between R. sivalensis and R. sondaicus}^^ 
Rhinoceros kendengindicus Dubois. 
Tapirus pandanicus Dubois. 

Sus hrachygnathus Dubois, related to recent 
(Sms verrucosus of Java. 

.S((.s macrognalhus Dubois. 

Hippopotamus (Hexaprotodon) sivajavanicus 
Dubois, related to Hexaprotodon sivalensis. 

Cervuhis kendengensis Stremme, related to 
existing muntjac. 

Cervus (Axis) Uriocerus Duhuis = Ccrvus {Axis) 
Lydekkeri INlartin. 

Cervus {Rusa) kendengensis Dubois. 

Cervus (Rusa) palseomendjangan Dubois. 

Duboisia (Tetraceros) Kroescnii Dubois, re- 
lated to existing Boselaphus and Tetraceros. 

Leptobos Grocnereldtii Dubois. 

Leptobos depend irornis Dubois. 

Bibos palxosondaicus Dubois, related to the 
existing Bihos sondaicus.^^^ 

Bibos protocavifrons Dubois. 

Buffelus {Bubalus) palseokerabau Dubois, re- 
lated to existing Javan Buffelus."' 

Hysirix. 

Manis palseojavanicu Dubois. 

Mccecyon trinilensis, related to recent Mececyon 

javanicus. 
Felis oxygnatha Dubois. 
Felis trinilensis Dubois. 
Felis rnicrogale Dubois = Feliopsis palxojavdnica 

Stremme. 
Hyxna bathygnatha Dubois. 
Lul7-a palseoleptonyx Dubois. 



The last word by Dietrich (1926.1, p. 1.39) makes it still more 
recent, namely: "Selbst wenn das Entwicklungstempo rascher ge- 
worden ist, kommen wir fiir Airawana zu einem sehr viel jiingeren 
Alter als bishcr fiir ihn und damit fiir die Trinilschichten fest- 
gesetzt wurdc, niimlich zu .Jung- bis Jiingstpleistocan. Diese Auf- 
fassung lies geologischen Alters der Pithecanthropusschichten 
bahnt der \'on Dubois 192.3 gewonnenen Erkenntnis, dass Pithe- 
canthropus ein tilled der Hominiden ist, den Weg zu dem weiteren 
Schritt, das P. bereits zur Gattung Homo gehort." 

Osborn, 1928: It now seems probable that Stegodon airdwana 
is of lower Middle Pleistocene age, somewhat more ancient than 
the Godavari, Narbada Alluvium, and with a more early fauna, 

Tafcl XXX IX. 




}?nii;;^ Jlrwchcn.tuJtrcrzr{c-Turrc 



VuJJccui.rn.arLtcC 



a. Kartcnsl;"uzc der Umgebung vun Trinil. M.isstal) I : I 500000. 

Fig. 772. Sketch of the Kendeng horizon (vertical shading), Trinil, Java, 
containing the type of Pithecayilhropm ereclus Dubois and the type of Stegodon 
gancxa var. javanicus {=S. airawana or S. trigonocephatus), also referred specimens 
of Stegodon airawana. After Dubois, 1908, Taf. xxxix. Middle Pleistocene. 

including the rhinoceroses related to the R. sivasondaicus and 
lacking the 7?. unicornis and Equus namadicus forms as well as the 
Palseoloxudon namadicus of the Godavari, Narbada Alluvium. 

Comparison of the beautiful figures of Janensch (1911, figs. 
1-16, Taf. xxi-xxv) together with casts (Amer. Mus. 6835) indi- 
cates that Stegodon airdwana is a much more progressive species 
than the S. insignis type of the Upper Pliocene [to Lower Pleisto- 
cene], and referred of the Lower and Middle [Upper] Pleistocene; 
it is nearly as progressive as the Stegodon aurorse type from Mt. 
Tomuro, Japan, as shown in the comparative figure (Fig. 688). 
Consequently we are justified in accepting Dietrich's recent con- 
clusion that Stegodon airdwana and Pithecanthropus erectus are 

'((Maarel, 1932, p. 193) ". . .we do not doubt that Stremme would have come to a |)lcistoceiie age, had he -as we— arrived at the conclusion that the 
Trinil fauna contains at least three still living species viz., Bibos sundaicus fossilis, Buffelus Ijubalus var. sondaicus fossilis, and Rliinoceros sondatcus 
f OSS His. "\ 



More recently (letter, 1924) Dietrich determined that the 
species Stegodon airdwana, which the Kendeng horizon contains, 
is somewhat more progressi\-e and consequently more recent geo- 
logically than S. insigni.i-ganesa. 



888 



OSBORN: THE PROBOSCIDEA 



of Middle rather than of Lower Pleistocene age. While the vertical 
heightening or hy])sodonty of the molar crown approaches that of 
the Archidiskodon planifrons type, the entire conformation of the 
S. airawana cranium (Fig. 773) is totally different and demon- 
strates afresh that the most progressive Stcgodontinw arc parallel 
with, rather than ancestral to, the Elephantidse. 

Sfteiika-Triiiit'Erii^iliitH. 




Fig 1-2. S'jorfon Airawana Makt. 
Vctbg von Wilhelm Engclmann id I.eipiig 

I'"if;. 773. Frdiit, and side views of referred Stegodon airawana Martin, as 
fi|i;uied by Jancnsch in his memoir of the Selenka-BIanckenhorn Trinil- 
Expcdition, 1911, Taf. xxi, figs. 1 and 2. One-sixth natural size. 

COMPARISON OF STEGODON AIRAWANA MARTIN WITH 

STEGODON INSIGNIS-GANESA 

Translated from the German of Dr. W. O. Dietrich with interpolations l)y 

H. F. Osborn and oorrected by Doctor Dietrich, April 18, 1021 

Stegodon airdwana Martin from the Pleistocene of middle 
Java (Pithecanthropus beds of Trinil) manifests evidences of insular 
dwarfing; it is in all its parts, both of teeth and skeleton, smaller 
tlian the continental west Asiatic and Indian races and s])C('ies. 
Si)ecial features of the .S. airawana teeth include the following char- 
acteristics: (1) Shallow longitudinal cleavage of the crowns; (2) 
relatively slender and narrow ridge-crests; (3) strong division of 
the ridge-crests into numerous mamillae [conelets], relatively 
limited cement covering; (4) small, thin, and strongly folded 
enamel on the abrasion sui'faces. (5) While Stegodon airdivana is 
a dwarfed lroi)ical form as compared with the giant S. insignis- 
ganesa, it exhibits more numerous, more elevated, and more com- 
pressed ridge-crests. 



It is in these respects the most specialized Stegodont of all 
known si)ccies, a terminal form based upon a Stegodont foundation. 
The dental formula of S. airdwana runs as follows: 

Stegodon airdicana : Dp 2 ? Dp 3^ Dp 4 ^ M 1 -^ M 2 ^ 

The greater or lesser development of the ridge formula de- 
pends upon the anterior and posterior half-ridges or talons, as in 
the reckoning of Falconer and Lydekker. In contrast [Osborn] the 
maximal ridge formula of the best-known continental .species, 
Stegodon insignis, is as follows : 



[Stegodoyi insignis-ganesa : Dp 3 f Dp 4 



AI2 



8-8H 



7» 
714-9 



M3 



"*^ •■■ 7M-10 



M-9-W -^'^ " 12W-X3J- 

Length of Molaes. — In the following table the characteristic 
maximal and minimal molar tooth length is expressed in millimeters : 



LTpper Molars Dp- 


Dp' 


Dp' 


M' 


M= 


M' 


Stegodon airdwana 18 


54 


90 


126 


189 


221 


Stegodon insignis 18 


68.6 


121.9 


246 


280 


310 


Stegodon elephan- 












toides ( = cliftii) 


70.7 


124.4 


155 


208 


236 


Stegodon bombifrons 






150 


188 


250 



Consequently Stegodon airdwana is two-thirds the size of S. 
insignis-ganesa and is also inferior in size to S. bombifrons. 

The construction of a specialized stegodont molar arises 
through the simple outgrowth of the anterior and posterior half 
ridge-crests. The lengths of the molar teeth in the above table 
are taken from Falconer and Lydekker and from original measure- 
ments by Dietrich. The maximal length measurements of M3 
of the lower jaw compared with the width give us the indices of 
the third lower molar which is the longest in the series: 



Stegodon airdwana ap. 304 
tr. ^^ 

ap. 310 



mm. = 22 length-breadth ridge index 



Stegodon insignis 
Stegodon cliftii 



tr. 



-yy- mm. = 28 length-breadth ridge index 



— 1^ mm. =36 length-broadtii ridge index 
^y^mm. = 33 Icngtli-broadlh ritlgc index 



tr. 
ap. 



Stegodon bombi- 
frons tr. 

By this means we may calculate the length-breadth ridge 
fiuotient. Such indices may be calculated also for the third upper 
molar, \P, when we secure similar figures which will enable lis to 




Fig. 771. SIcyoiUin airawana. JuveniU' craniniu with ,-<niall tn.sk.s. 
Elevation of narial openings as in <S'. trigonoccplialiis. Olisc r\i' I lie peculiar 
si raiglit front of the forehead and great, breadth across the (op of incipilal.s. 
Skelelon conjectural a»s in other Stcgodonts. Restorations by Margret Flinsch 
(lOIiO), miller the direction of Henry Fairfield Osborn. One-lift ietli natural 
size. 



THE STEGODONTIN.E: STEGODON 



889 



(leteiinine positively the progressive development within the Stego- 
dont series. In order to measure the specialization of the molar 
teeth, including anterior addition and reduction and posterior 
addition and reduction, a relative proportion may be established 
between the development of the two anterior teeth and the two 
posterior teeth. It is noteworthy that the proportion between the 
first and second molars in the Chine.se and Javanese .species is the 
same. From careful calculations it follows that M- in iS'. airdwana 
is more strongly reduced than in the corresponding tooth of the 
S. orientah's of Owen. 

On all these structural grounds Doctor Dietrich concludes that 
S. airdwana is more recent than either of the Asiatic species of the 
continent and that its newer Pleistocene age is rendered certain. 

STEGODON GANES.\ VAR. JAVANICUS DUBOIS, 1908 
Middle Pleistocene, Triiiil, Kendeng-Schichten, Java 

This subspecies, Stegodon ganesa javanicus, belonging in the 
same Kendeng beds as .S. airdwana Martin, 1890, is to be regarded, 
as remarked by Stremme, as a synonym of .S'. airdwana.'- 

In his paper of 1908, entitled "Das Geologische Alter der 
Kendeng-Oder Trinil-Fauna," Dr. Eugen Dubois, the discoverer 
of Pithecanthropus erectus, assigns to Pithecanthropus and the 
accompanying Kendeng-Trinil Fauna at the most a Lower Pleisto- 
cene age ("alt-diluvialen Alters")- The remains occur in the 
fossil-bearing tufa known as the "Kendeng-Schichten," in the 
center of which lies Trinil, as shown in Plate xxxix (Fig. 772 of the 
present Memoir). Recently Dietrich has assigned a Middle 
Pleistocene age to the Kendeng Trinil fauna. 

Dubois (1908, j). 1257) erroneously remarks that the two 
skulls described by Martin as S. trigonocephalus really belong 
specifically to Stegodon ganesa in their cranial and dental characters, 
but in order to distinguish this smaller Kendeng species, as an 
insular variety, from the giant continental form of ganesa, he 
proposes the subspecific name Stegodon ganesa var. javanicus. 
He considers it probable (1908) that this Javan subspecies be- 
longed to the Upper Pliocene fauna. 

Stremme, "Die Siiugetiere mit Ausnahme der Proboscidier," 
in "Die Pithecanthropus-Schichten auf Java," 1911, p. 142, regards 
this subspecies as closely related to Stegodon airdwana Martin, 
1890. 

According to Dubois (1908) the two skulls described by 
Martin as Stegodon trigonocephalus belong to this species." It is 
important to Ciote that the true .S. trigonocephalus of Martin does 
not occur in the same level as Pithecanthropus erectus. 

Stegodon Ganesa var. javanicus Dubois, 1908. "Das Geolog- 
ische Alter der Kendeng-Oder Trinil-Fauna." Tijdsch. Konink. 
Neder. Aardrijks. Genoots. Amsterdam, Tweede Serie, Deel xxvb, 



No. 6, p. 1257. Tvi'K. — Material from Keudciiu;. TTom- 

ZON AND Locality. — Trinil, Kendeng Schichten, Middle Pleis- 
tocene, Java. 

Description {Op. cit., 1908, pp. 1256, 1257). -" Die Untersuch- 
ung eines sehr reichlichen Stegodonten-materiales aus dem ganzen 
Kendeng, unter welchem auch mehrere Schadel von Jungen und 
alten Tieren, hat mieh nun zu dem Ergebnisse gefuhrt, dass alle 
diese l^eberreste einer einzigen Art angehiJren. Diese unterscheidet 
sich \on Stegodon ganesa kaum anders als durch ihre viel ger- 
ingere Grosse. Der Schadel besitzt dieselbe weite Temporal- 
grube, ist ebenso charakteristiscli stark brachycephal und auch 
iibrigens sind beide Formen sehr iihnlich. Nur iiltere Schadel 
weichen in einigen Beziehungen etwas von der typischen Ganesa- 
Form ab, indem niimlich die Frontal- und Occipitalteile sich 
gegeneinander abflachen und mehr oder weniger scharf von 
einander getrennt sind. Das braucht uns abcr nicht davon 
abzuhalten diese Formen derselben Art zuzuschreiben, denn es 
kann ja bei den Elephanten die Schadelform einer und derselben 
Art, innerhalb gewissen Grenzen, betrachtliche Verschiedenheiten 
zeigen. Der Ganesa-Typus bleibt bei unserer Form doch immer 
erkenbar." 

"Auch durch ihre Molaren ist sie von S. ganesa spezifisch 
nicht zu trennen, nur geht die Lamellenzahl im Unterkiefer bis 
zu wenigstens 14 anstatt 13; die Zahl der Lamellen hat aber, 
wie wir durch Pohlig wi.ssen, fiir die Unterscheidung der Elephan- 
tenarten nicht die grosse Bedeutung welche man ihr friiher zu- 
schrieb." 

"Zu dieser Art gehoren auch die zwei von M.\rtin als 
Stegodon trigonocephalus . . . beschriebenen Schadel. Waren 
diese gut erhalten gewesen, so hatten sie erkennen lassen, dass 
die dreieckige Form keine urspriingliche und der Art eigen- 
tiimliche ist. Der darauf Bezug nehmende Name ist also zu 
kassieren. Ich schlage nun vor, diese kleinere Kendeng-Form 
nur als VarietJit von der riesigen Festlandsform, der sie deutlich 
sehr nahe steht, zu trennen, ftihre sie also als Stegodon ganesa var. 
javanicus ein." 

Characters {Op. cit., 1908, p. 1257). — "Wenn man nun auch 
nicht zugeben kann, dass Stegodon ganesa und Stegodon insignis 
einer Art angehoren, so sind beide jedenfalls einander sehr nahe 
verwandt (die Molaren sind nicht oder kaum zu unterscheiden), 
Stegodon insignis aber kann man, mit M. Schlosser . . . ), 
geradezu als das Leitfossil der jlingeren Pliociinfauna Ostasiens 
ansehen. Es ist schon hierdurch wahrsclieinlich, dass unsere 
javanische Stegodon ganesa der gleichen jungpliocanen Fauna ange- 
hort. Unzweifelhafte Ueberreste von Stegodon ganesa sind nun 
aber noch nicht in jlingeren als Pliociincn Schichten angetroffen 
worden, denn der in den Narbada-Schichten gefuudene Stoss- 
zahn kann man dieser Art nicht mit Sicherheit zuschreiben ..." 



'[The following note was prepared by Dr. George Gaylord Simpson (October, 1937): "Stegodon trigonocephalus Martin, 1887, pp. 27, 36, 41, was 
founded i)iincipally on two skuHs, Professor Osborn selecting tlic younger as type. These were probably from the vicinity of Surakarta (Solo), Java (see 
Martin, 1S87.1, i)p. 2.5, 27). Duljois briefly cliaracterized abundant material from Kendeng, referring it to Stegodon ganesa but a.s a new variety javanicus 
(Dubois, 1908, pp. 1256, 12.57). He also referred S. trigonocephahis to S. ganesa. He probably considered trigonoceplialus as the same as his vari(>ty javanicus 
but he did not explicitly say so and the clear implication is tliat they luivc different types. Von Koenigswald considers S. trigonocephalus as distinct from 
S. ganesa but the same as S. airawana (1933, pp. 103-105). He implies that S. ganesa var. javanicus is the same as S. trigonocephalus and S. airawana and 
applies to the species the oldest of the three names, S. trigonocephalus. Maarel, however, considers .S'. trigonocephalus to be distinct from S. airawana. and this 
was Professor Osborn's opinion. It appears, in any case, that S. ganesa var. javanicus Dubois, 190S, is synonymous with S. trigonoceplialus, S. airawana, 
or both, and as both names antedate it, Dubois' name is invalid." — Editor.] 



890 



OSHOKN: THE PROHOSCIDEA 



Stegodon trigonocephalus Martin, IS.Si 

figures 70."), 731, 77."), 77(1, 777, I'l. xx 

Probably vicinity of Surakarta, Java. Dubois notes that Stegodon 
trigonocephalus docs not occur in tlie Kcndeng-Schichten of Trinil; Matsu- 
moto regards the geologic level of »S. trigonocephalus as equivalent to the 
Lower Pleistocene, Boulder Conglomerate bods of India. 

Stegodon trigonocephalus is a dwarfed insular Stcgndont, based 
chiefly upon an immature cranium (Fig. 776) which has sub- 
stantially the same characteristic insular form as the immature 
cranium of S. airawana (Fig. 773), but which is very different from 



4j?l.^r-^^^ 




Fig. 775. Stegodon trigonocepliolus. Rather young cranial profiles 
drawn directly from type. Frontal line much longer than in S. pinjorensis. 
Cranium extremely abbreviated. Skeleton conjectural, as in other Stegodonts. 
Restoration by Margret Flinsch (1930), under the direction of Henry Fairfield 
Osborn. One-fiftieth natural size. 

the crania of »S'. bombifrons or »S. insignis-ganesa. This ju^•enile 
type represents the immature condition of the Stegodont cranium 
and consequently is very interesting and important. It is now re- 
garded as in a Pleistocene stage of evolution, but the extremely 
immature condition of the teeth renders its specific determination 
very difficult. F'rom the following description and figures of 
Martin, it is difficult to distinguish this immature type from Stego- 
don airawana; it is apparently a somewhat more primitive Lower 
Pleistocene form. 



Stegodon trigonocephalus Martin, 1887. "Fossile Siiugethier- 
reste von Java imd ,l!i]);m," Samrnlung. (leolog. l{eichs-Museimis, 
Ijeiden, Beitriige z. (ieolog. Ost-Asiens und Australicns, f^er. I, 
Bd. IV, Heft 2, pp. 27, 36 (1887); also Jaarb. Mijnw. Neder. 
Oost-Indie, 1887, Wetensch. C!ed. 16, Palaeontologie van Neder- 
laudsch-Indie. Verhandeling No. 21, pp. 3, 12(1887). Type.— 
Young skull with third and foiu'th deciduous premolars in the 
CJoological Museum of Leiden. Hokizon and Locality. — 

Probably from vicinity of Surakarta, Java; said to be associated 
with remains of proboscideans referred to Stegodon bombifrons, 
Eiielephas namadicus, and Eiielephas hysudricus. Ty'pe 

Figure. — (Op. cit., Martin, 1887, Sammlung. Geolog. Reichs- 
Mu!5eums Leiden, Tab. ii, figs. 1, la, and Tab. iii, fig. 1. 

Description. — The original description by Martin (1887, pp. 
36-41, with figures) gives the characters of the immature type 
skull and teeth in great detail, and on page 4.5 offers the follow- 
ing distinctions from Stegodon cliftii and .S. bombifrons: "Zuniichst 
ist es von Bedeutung, dass die Ziihne von denjenigen des Stegodon 
Cliftii so voUig verschieden gebaut sind, dass eine Vereinigung der 
javanischen Art mit der genannten, von der ein Schiidel noch nicht 
bekannt ist, von vornherein ausgeschlossen wird. Ebenso be- 
stimmt unterscheiden sich aber (soweit unsere Kenntniss bis jetzt 
reicht) audi die Zt'ihne von St. trigonocephalus durch die Jochzahl 
der Praemolaren von alien anderen, bis jetzt bekannten Art.en von 
Stegodon, . . . Bezeichnend fiir den 2**^" Praemolaren von St. 
trigonocephalus ist ferner seine ovale Form, und ebenso fiir den 
gten pi<^pniolaren die kaum mcrkliche Convergenz seiner Seiten- 
fliichen nach vorne zu, Merkmale, die von alien bekannten, ent- 
sprechenden Ziihnen der iibrigen Stegodontenarten abweichen. 
Durch die hohcn, schmalen Joche und vor allem durch die feine 
Fiiltelung diirfte audi bei Bruchstiicken der ^Nlolaren von Si. 
trigonocephalus eine LTnterscheidung von St. bombifrons bisweilen 






..'^"■. '\. 




^iiUsihi'^^ 




Fig. 776. Type of Stegodon trigonocephalus Martin, 1887, Tab. n, figs. 1, la, and Tab. in, fin. 1. The side and front views are reproduced one-eighth 
natural size; the crown view of the tooth is reproduced one-third natural size. 

Taf. I, figs. 3 and 4 (Naumann, 1887). A single crest from .lava, distinct from ,S'. trigonocephalus and d()ul)lfully referred by Naumann to.S'. insignis Falc. 
ami Caut. (op. cit., p. 9): "Das Zahnbruchsfiick kann nur zu Si. in.'iignis oder zu St. Gane.ia gehoren." 



THE STEGODONTIN^: STEGODON 



891 




Fig. 777. Front and side view outlines to a one-sixteenth scale of Stcgodont crania of: 

A, Al, Stegodon pinjorensis type, mature adult (Amer. Mus. 19772), from near Siswan, India. See also figure 765. 

B, Bl, Stegodon bombifrons cotype, mature adult (Brit. Mus. M. 2979, cast Aracr. Mus. 10378). See figure 734. 

C, Stegodon orientalis grangeri ref., young adult (.\mer. Mus. 18(530) collected by Walter Ciranger, Wanhsien, Province of Szp<'huan, China. See figure 763. 

D, Stegodon orientalis grangeri ref., infantile (Anicr. Mus. 18032) collected by Walter Granger, Wanhsien, Pnivince of Szechuan, China. 

E, El, Stegodon trigonocephalus tyjie, juvenile, from vicinity of ?Surakarta, Java. Sammlung. Geolog. Reichs-Museums, Leiden. Sec figure 776. 

F, Fl, Stegodon airhwana ref., juvenile, Java. See figure 773. 



ermoglicht werden; dagcgen ist die Trennung solcher Reste \on 
St. insignis und ganesa mit sehr grossen Schwierigkeiten verbun- 
den." 

Ridge formula of S. trigonocephalus: Dp 3- Dp 4^^ as compared 
by the author with S. diftii: Dp 3^ Dp 4-; with .S. bombifrons: 
Dp3,^^Dp4^^^''; with.S. ^a/ifsa: Dp 2^ Dp 3- Dp 4^. The skull, 
however, is quite different in form from that of S. insignis. Martin 
inclines to compare S. trigonocephalus with S. insignis rather than 
with the Lower [Middle] Pliocene stage of S. bombifrons. 

It is important to note that these characters relate both to the 



immature type skull and to a second older skull which is a referred 
specimen, namely: {op. cil., p. 41) "6. Alter Schddel. (Tab. iv u. 
V Fig. 1)." 

A specimen from northwest Mindanao, Philippine Islands, 
referred to this species is described in detail by Dr. Edmund Nau- 
mann of IMunich University in his "Fossile Elephantenreste \'on 
Mindanao, Sumatra und Malakka," Abhand. u. Berichte des 
Konigl. Zoolog. Anthrop.-Ethnog. Museums zu Dresden, 1887, 
pp. 5-8. He subsequently (1890) made it the type of Stegodon 
mindanensis. 



StEGODONTS OF THE PHILIPPINE ISLANDS AND JaPAN 

We owe to Naumaiin (1890) and to Matsumoto (1915, 1918) the discovery and description of three very 
progressive species, originally referred to Stegodon, which in part or wholly seem transitional in structure to 
ArcMdiskodon or to Elephas. The true affinity of these types rests upon (a) the still unknown structure of the 
cranium and (b) the ridge formula and height and form of the ridge-crests as now revealed in the case of the species 
Stegodon aurorse (Fig. 781). In brief, until the cranium is known, we cannot be certain whether these animals are 
progressive Stegodonts or primitive Archidiskodonts. The history of discovery and description is as follows. 



892 



OSBORN: THE PROBOSCIDEA 



Stegodon (Archidiskodon?) mindanensis Naiiiu.imi, 1S9() 

■•'iguics 700, 77S 
Mindanao, Philippine Ishiiuls, Lowci(?) to Middle Pleistocene. 

Stegodon Mindanensis Naumann, 1S90. "Stegodon Minda- 
nensin, eine neue Art von Uebergangs-Mastodonten." Zeitschr. 
deutsch. geol. (ies., Bd. XLII, Heft 1, pp. 166 169. Type.— 
Fragment of a molar tooth. Horizon and Locality. — 

Mindanao, Philippine Islands, (?)Lower to Middle Pleisto- 
cene. Type FiGriu;. — Naumann, 1887, Taf. i, figs. 1 and 
2. Type Description. — (Naumann, 1890, p. 167): ". . . so 
bleibt (loch das friiher erzielte Rcsultat, nach wclchem (lurch die 




TvpE OF Stegodon (archidlskodon?) mindanensis 
Fig. 778. Type incomplete inferior molar of Stegodon 7Hi}idanettsis 

Naumann, 1890, from Mindanao, Philippine Islands (originally figured by 

Naumann in 1887, Taf. i, figs. 1 and 2, as Stegodon tiigonocephalus). 

.•\s compared with Stegodon airawana (Fig. 77S), the ridge-crests in 

S. mindanensis type are more vertically placed, as in the Stegodon aiirorx 

typo (Fig. 780). 



Untersuchung der beiden Zahnbruchstiicke von Mindanao 'die 
Verbreitung der Siwalikfauna iiber das Ciebiet der Philippinen 
bewiesen und die enge \'erkniipfung einer wahrscheinlich jungter- 
tijiren Saugethierfauna auf Java und den f^hilippinen durch eine 
in der Entwicklungsreihe der Stegodonten und EIe])hanten 
hochwichtige Art' constatirt sein .sollte, zu liecht bestehen. . . 
^lerkwiudig ist ferner ein medianer Einsclmitt der Krone, der 
jederseits \'on einem secundiiren Einschnitt begleitet wird. Durch 
diese Spaltungen werden die Mamillenreihen in (Iruppcn zerlegt." 
The type of this species was originally figured by Naumann in 
1887 as Stegodon trigonorephalus; it was regarded by him as an 
outlyer of the Siwalik proboscidean fauna. A single ridge-crest 
fiom Java (see I'^ig. 776, lower), distinct fiom S. tiigonocephalus, 



was doubifuiiy referred by Naumann to S. iiisignis or .S. ijniicsa 
Falconer and C'autley. (Naumann, 1887, p. 9): "Das Zahn- 
bruchstuck kann nur zu St. insignis oder zu St. Ganesa gehoren." 

If a true Stegodont, Stegodon mindanensis (Fig. 778) is even 
more progressive than S. airawana, because the valleys between 
the ridge-crests are entirely closed up; it compares somewhat more 
closely with Stegodon aurorx (Figs. 780, 781), also an imperfectlj' 
known sjjecies possibly referable to Archidiskodon. 

Osborn, 1924: It is very difficult to determine the characters 
of this type (Fig. 778) without examination of the original speci- 
men. The type lower molar is far more progressive in the direction 
of Archidiskodon or Elephas than the types of either Stegodon 
insignis or S. ganesa; it is, in fact, a true-crested tooth in which 
the ridges are closely compressed and the valleys closed. Conse- 
quently this tooth should be compared with that of a primitive 
species of elephant or of Archidiskodon. 



Stegodon auroras Matsimioto, 1915, 1918 
Figures OSS, 709, 7S0, 781, S19 
Uppci(?) Pliocene, Mt. Toniuro, Kaga, Japan. 

Like Stegodon mindanensis, Elephas (Prostegodon, Parastego- 
don) aurors- is either a highly progressive Stegodon or a primitive 
Archidiskodon, a point to be determined positively l)y the discovery 
of a cranium. 

Elephas (Prostegodon, Parastegodox) auuor.e Matsu- 
MOTO. — In describing the type specimen, determined as a second 
right superior molar, r.M^, Matsumoto remarked (1918, p. 52) : 
". . . this specimen represents a species new to science, being ... a 
transitional form from Stegodon to Elephas." Originally described 
as Elephas aurorsc Mats., 1915, also 1918, j). 52, Pl.xx, it was after- 
ward referred by Matsumoto (in Osborn, 1923.601) to Prostegodon, 
the genotype of which is Ma.<itodon latidens, and finally (1924.2) 
was made the genotype of Parastegodon Matsumoto, to include 
the Stegodon mindanensis-E. atirorx group. 

Elephas aurora: Matsumoto, 1915, 1918. Preliminary reports 
of this species were published in the Scientifie (Jazette, Tokyo, 
Vol. Ill, No. 5, 1915, pp. 308 315, and in the Journal of the 
Geological Society, Tokyo, Vol. XXIII, No. 275, 1916, i). 294 
(both in Japanese). Also (in English) "(^n a New Archetypal 





Fig. 779. Kcferrcd by Jancnsch to S. airaivnnn. Nine ridge-crested second superior molar, M'-', after .laiicnscli, 191 1, p. 171, fig. 12, and ridge-crest section 
of a third superior molar, M^, after Janensch, op. cit., p. 174, fig. 13. One-half natural size. Both figures inverted to show the molars in natural jmsition. 

(Left) Lateral view. (Right) Ridge-crest section. 

It will be observed that the convexity of the ridge-crests of the .second molar crown, resulting in the wide divergence of the cre.sts, together with the more 
numerous ridge-crests, tends to relate this animal to Stegodon rather than to Archidiskodon, in which the ridges are more vertical and more closely compressed. 



THE STEGODONTINi^^.: STECIODON 



893 





Fossil Elephant from Mt. Toimiio, Kuga." Sci. Uei)f. Tulioku 
Imp. Univ., 1918, (2), 111, Xo. 2, p. 52. Typk.— A right AP 

with ten and a half ridges, thus exceeding the ridge formula of 
JM- in Archidiskodoit planifroits, namely, M 2^^. Original in 
Geological Institute, Tokyo. Horizon and Locality. — Mt. 

Tomuio, Kaga, Japan, (?)Upper Pliocene. Type FiorRE. — 

Op. cit., 1918, PI. XX, figs. 1-3. SCPPLEMENTARY DESCRIPTION 

(Matsumoto, 1924.2, pp. 256, 257, 262).— Made the genotype of 
the genus Parasiegodon (see p. 903 of the present Memoir).' 

Type Description (1918, pp. 52-55). — "The unique type 
specimen corresponds to an ujiper, probably penultimate, molar 
of the right side. It measures 180 mm. in length, 75 mm. in 
maximum width and 48 mm. in the maximum height of the crown." 
The reasons why the autiior referred this new species to Elephas 
and not to Strrindon arc as follows: The ridge formula of M- 
corresponds to 10,'i>, while that of the known species of Slegodon 
corresponds to 6 to 9. The author gi\es six additional reasons 
for separating the animal from Stegodon, and six additional reasons 
for separating it specitically from E. planifro/is, to which he regards 
it as allied generically. The author concludes (p. 55): "The 
present species appears to the present writer to be more archety- 
pal and more Stegodon-\ike than E. planifruns in the second distinc- 
tive characteristic, but just the opposite in the third, fifth and 
sixth distinctive characteristics. At any rate, the discovery of the 
present species is worthy [to] be considered as an additional 
datum to pro\e the very intimate alliance of Stegodon and Elephas." 

Stegodon orientalis shodoensis ]\Iatsumoto, 1924 
Middle Pleistocene. I.sland of Mitsugo (Mitsugo-shima) and Island of 
Shodo, Inland Sea, Japan. 

Slegodon orieidalis tshodocnsis Matsumoto, 1924. "Preliminary 
Notes on the Species of Stegodon in Japan." Journ. Geol. Soc. 
Tokyo, XXXI, Xo. 373, pp. 333-335. Type.— A fragment of 

skull — left and right upper jaw bone with third molar attached. 
Uyeno Museum of Tokyo, No. 2194, presented by Mr. Tomekichi 
Ozaki. Horizon and Locality. — ]Mitsugo-shima, Yoshima- 

mura, Xakatadotsu-gun, Province of Sanuki, Japan. Milazzian- 
Tyrrhenian. ]\Iiddle Pleistocene. Type Figure. — Type 

figure not seen by the jiresent author. - 

Description in Japanese by Malsiiinolo; Iranslation by Messrs. 
Shoichi Ichikawa and Ushinosuke Narahara. — (Matsumoto, 1924.3, 
pp. 333-335). "Materials: A fiagment of skull: left and right 
upper jaw bone \vith third molar attached (found at Mitsuko- 
shima, Yoshima-mura, Nakatadotsu-gun, Province of Sanuki. 
Specimen No. 2194, Uyeno Museum of Tokyo, presented by Mr. 
Tomekichi Ozaki). A small fragment of third molar (found at 
Shodo-shima and the property of the Tokyo Imperial LTniversity). 
Third upper left molar (found at Shodo-shima, the cast is kept at 
Kyoto Imperial University). A fragment of an upper palate with 
third molar of left and right attached (found at Shodo-shima, the 
property of Kyushu Imperial L'ni\'ersity and the cast kept at 
Kyoto Imperial University)." 

HSee also Matsumoto's article (in English) "On Parastcgodon Mastumoto and its bearing on the Descent of Earlier Elephants," Sci. Kept. Tohokn Imp. 
Univ., (2), XIII, No. 1 (Matsumoto, 1929.3).— Editor.! 

-[Dr. Jiro Makiyama has recently reviewed the Japanese Proboscidea (see Makiyama, "Japonic Proboscidea," 1938, Mem. Coll. Sci. Kyoto Imp. Univ., 
(B), XIV, No. 1, pp. 1-59), in which he figures (p. 18, fig. 7) the holotype palate of Slegodon shodoctisis, with right an<l left third molars in situ. On page 19 he 
states that "The material of Matsumoto was not enough to give full details of the species, but it was named before Parasiegodon akashieiiMS Takai, 1936, which 
seems to me to be in the nearest relation, as it may have a new name form S. shodoensis akashien.fis." This subspecies he describes on pages 21 to 27, including 
three figures (Figs. 10-12).— Editor.) 



Type of Stegodon auroRvE 
I"ig. 7S0. Tyi)C of Elephas {Piostcgodon, Parasiegodon) aurorx 
Matsumoto, after Matsumoto, 1918, PI. xx, figs. 1 and 3. One-half 
natural size. Determined by Matsumoto (1918) as a ten and a half 
crested second right superior molar, r.M". Compare figures 699 of S. 
bombifrons and figure 762 of 5. orientalis grangeri showing convex 
external side aTid flat internal side; also figures 6S6, 7, and 6SS. 




Fig. 781. Vertical section of type molar, r.M-, of Stegodon aurone 
(compaie Fig. 780, crown and side views of tyiie of Elephas (Prostegodon, 
Parasiegodon) aurorse, also Fig. 82o, lectotype of E. [ = .Archidiskodon] 
planifrons). One-half natural size. After ])hotograph sent by Doctor 
Matsumoto. 



894 



OSBORN: THE PROBOSCIDEA 



"Ridge formula of molar: left, a+9X l = }i-Q-}Q; right, 
a+8X [ = ]4-8-}i\. Tlie exact measurement of ridges cannot lie 
determined. The length as it is, left 226 mm., right 215 mm. The 
greatest width, left 97 mm. at 2nd ridge, right 96 mm. at 3rd 
ridge to 6th ridge. The height of crown, left 52 mm. at 6th and 
7th ridges, right 49 mm. at 7th ridge. The 6th ridge of left, and 
7th ridge of right shows the beginning of wear. . . . There are four 
ridges within 100 mm. The length of each ridge is 25 mm. and 
this is rather narrow as a width for third molar of Stegodon. The 
valley is narrow, and the entrance of inside of valley has supple- 
mental cusps. The fold of enamel is very regular and fine. The 
cement is well develoi)ed. The interlocking of opposing surfaces 
of the molar is of Mastodon type." 

"Second example: Ridge formula may probably be XllX. 
Greatest width, 93K mm. at 6th ridge, height, 55 mm. at 7th ridge. 
There are four ridges within 100 mm. Judging by the measurement, 
it looks like a second molar, but by the .shape of its being attached 
to the palate is a third molar. Ridge formula, left a+9X ; right 



a + 8X. The exact measurements of ridges cannot be determined. 
Length, left 205 mm., right 20U mm. tireatest width, 85 mm. at 
6th ridge from back (left), 84 mm. at 5th ridge from back (right). 
Height, 45 mm. at 3rd ridge (left), 50 mm. at 2nd ridge (right). 
There are four ridges within 100 mm. The interlocking occlusion 
of the molar is elephant type." 

"A small fragment of third molar, which shows three ridges. 
I cannot determine whether it belongs to the upper or lower jaw. 
The length from inside to outside of each ridge is 90 to 96 mm. 
The width of front to back is 24 to 25 mm." 

"Among these specimens, the first has comparatively low 
crown and the second and third have very high crowns. The 
fourth specimen is rather low as absolute value but as a ratio it is 
high. The molars are different in sizes, and this differentiation 
(or varieties) is the one which occurs in the last period of the 
existence of the species." 

"These specimens resemble the Stegodon insignis which is 
found in the Narbada Valley of India." 



V. RECENTLY DESCRIBED STEGODONTS FROM JAVA, CHINA, AND JAPAN 

[The following species were described resi)ectively by Dr. Franciscus Hendricus van der Maarel ("Contribu- 
tion to the Knowledge of tlie Fossil Mammalian Faima of Java," 1932), by Dr. G. H. Ralph von Koenigswald 
("Beitrag zur Kenntnis der Fossilen Wirbeltiere Javas," 1933), by Dr. Shigeyasu Tokunaga ("Fossil Elephant 
teeth foimd at Yokohama and Kakio, Kanagawa Prefecture," Journ. Geog. (Tokyo), XLVI, No. 546, 1934, and 
"A New Fossil Elephant found in Shikoku, Japan," Proc. Imp. Acad. Tokyo, XI, 1935), by Dr. Chung-Chien 
Young("Miscellaneous Manmialian Fossils from Shansi and Honan," 1935), and by Dr. Arthur Tindell Hopwood 
("Fossil Proboscidea front China," 1935). While the species of Stegodonts in these pul^lications (other than 
Parastegodon? kwantoensis and Parastegodon sugiyamai of Tokunaga) were noted by Professor Osborn, they 
were not studied intensively by him, consequently they are hsted here together with excerpts from the original 
descriptions, including type figures, but without comment or determinations by the author of the present 
Memoir. 

Doctor Hopwood states on page 103 of his Memoir that "Insufficient is known of the Stegodontidie of China 
to make a comparison with those of India very profitable. At present it seems as though they occur earUer in 
India than in China, where they are a late invasion from the South. Palaeoloxodon, on the other hand, may be 
more recent in India than in China." — Editor.] 



Stegodon bondolensis van dor Maarel, 1932 
Figure 782 

IJondoI nciir KiiwiiiiK, district RiuMiiiblatiiiif;, Regency niora, Residency 
Remhaiig, Java. 

Stegodon bondolensis van der Maarel, 1932. "Contribution to 
the Knowledge of the I'ossil Mammalian Fauna of .Ia\a," ])p. 
158-164. TvPK. — "Fine fragment of the mandible, com- 

prising the whole of the horizontal ramus and the lower ])ortion of 
the ascending ramus of either side; containing on either side 
a molar which is inferred to be the M3 from the absence of the 
indication of any other tooth behind." Honizox and 

Locality. — Bondol near Kuwung, district Kandublatimg, Regen- 
cy Blora, Residency Rcmbaiig, .la\ a. T'i i'k Fkukk. — Op. 
n't., PI. XIV, figs. 1, 4, 5, also text figiu'os 24 and 25. 

Dkscriptiox.— (Van der Maarel, 1932.1, pp. 158-164): "The 
mandible is broken into two in the symphysis, but botii parts 
match exactly. 'I'he alveolar border of both molars is dninaged." 



"The diastema descends obliquely forwards at an acute angle 
of about 35° with the inferior border of the ramus. The mandibu- 
lar symi)hysis is produced into a relatively broad, spoutlike 
termination. Its lower border is damaged. . . . The outer surface 
of the horizontal ramus presents two foramina mentalia, the one 
situated some 8 cm. below the anterior extremity of the alveolar 
margin, the other about 8 cm. in advance of the former and nearer 
to the inferior bolder of the ramus." 

"Only the lowci- portion of the coronoid jirocess has been 
jtreserved." 

"Measurements of lower jaw in cm. 
See text fig. 24. 
greatest height = 16 



In plane I 
In i)lane II 



greatest width = 7.8 

greatest height = 12 

greatest width = 17 

Length of RC =47.5" 



THE STEGODONTIN^: STEGODON 



895 



"As the front .side of both moliii's is duinugeil, it seems — ut 
first sight — impossible to ascertain the true number of ridge-crests 
carried by each of them. Fortunately, however, in the I.M3 a small 
portion of the front surface of the fangs has been preserved. It is 
situated below the anterior side of the foremost ridge-crest, and 
proves with absolute certainty, that this ridge-crest is really the 
first one. That is to say: not a single ridge-crest has been lost. 
The ridge formula is, therefore, 8X." 



can, however, only be .seen in the last three ridge-crests. In this 
connection attention may be drawn to the fact that in the mohirs 
of the present sjiecimen enamel, cement ami the greater part of the 
dentine had the same irregularly spotted appearance. As a result 
the crenulation of the enamel was very indistinct, and it was only 
possible to make it distinctly visible in the photograph by black- 
ening the dentine of the tooth. Another result is, however, that 
'Stufenbildung' though present, is in general not clearly exhibited." 




Fig. 782. Type mandiblo and enlargement of third left inferior molar of Slegndon bondolensis. After 
van der Maarel, 1932, PI. xiv, figs. 1, 4, and 5. Molar about three-fifths natural size; mandible about 
one-seventh natural size. From Bondol, Java. 

(Upper) Crown view of third left inferior molar, I.M3, with 8 4- ridge-crests; same molar as in ac- 
companying lower jaw. 

(Lower, right) Superior view of fragmentary lower jaw with right and left third molars in situ. 

(Lower, left) Right lateral view of same jaw. 



"The tooth is distinctly curved outwards. The base of ridge- 
crests 1-7 are more or less of the same length ; from ridge-crest 
8 inner and outer side of the molar converge suddenly back- 
wards." 

"The base of the crown and the grinding surface are strongly 
concave. The grinding surface slopes obliquely from the outside 
inwards." 

"The enamel is moderately thick and consists of two layers 
(see the buccal cusps of ridge-crest 7). A distinct 'Stufenbildung' 



"There is no indication of the presence of a median longitudi- 
nal cleft. Cement is present, though in small quantities. The 
valleys between ridge-crests 1, 2, 3, 4 and 5 possess but little 
cement. More backwards the amount is somewhat less small, not 
only the base of the valleys being filled to a greater extent, but 
also the lingiuil and buccal sides of the ridge-crests, and the 
posterior side of the hind talon being covered by a thin coat of 
cement. Noteworthy is the stronger, local de\'elopment of cement 
near the buccal side of ridge-crest 7." 



896 



OSBORN: THE PROBOSCIDEA 



"MeASUUEMKNTS of U. and L.M3 IN MM. 

Total length of r. M3 190 mm.lnicasured along lucilian line 
Ditto " 1. Ms 197 " J of crown." 

"As a matter of fact all the species of Stegolophodoti may be 
left out of consideration, being all very primitive forms which in 
my opinion may as well be reckoned to the family of the Maslo- 
donlidx." 

"On the other hand, most of the remaining species [of Stegodon] 
are considerably moi-e progressive than the form to which the 
lower jaw in question belongs." 

"In my opinion we may, therefore, conclude, that the present 
Javan form is distinct from all other species. I propose the 
specific name of bondolensis, indicating the locality from whence it 
has been procured." 



schcinend weniger an Ihihe zunalun, als os bei der (ypischen Form 
der Fall ist. Die Ictzteii drei .lochc und der Talon liegen noch 
dick unter Zeraent." 

"Diesen letztaufgefuhrten Zahn bilde ich Taf. xxvii ab, 
neben dem entsprechenden Zahn von St. t. trigonocephalus von 
Pitoe, nahe Trinil. Dieser letztgenannte Zahn ist ca. 270 mm lang 
imd gehorte einem mittelgrossen Tier. Zwei letzte Molaren 
von Trinil sind (nach .lanonsch) 304 und 303 mm lang, auch solche 
von Watoealang und Ngandong erreichen 300 mm, (da sie z. T. 
mit der Hinterhiilfte noch im Kiefer stecken, kann ihre Grosse nur 
geschiitzt werden.) Dem gegenliber miissen die Ziihne \'on 
Boemiajoe mit ca. 240 mm klein genannt werden. Dabei haben 
sie dicselbe Breite und dieselbe Biegung, wie die anderen Ziihne. 
Die Verkiirzung kommt daher, dass die Ziihne von Boemiajoe nur 
X 11 X Joche besitzen gegeniiber X 13 X bei der jiingeren Form. 




Fig. 783. Tj'i)e left third inferior molar, I.M3, of Stegodon trigonocephalus praecursor, after von Koenigsvvald, 1933, Taf. xxvii, 
fig. 2, two-tliirds natural size. From Bumiaju, Java. 



Stegodon trigonocephalus praecursor 

\()ii Koonigswald, 1933 
Figure 783 
Untere Schiehton, Kali Glagah, Bumiaju, Java. 

Stegodon trigonocephalus praecursor \on Koenigswald, 1933. 
"Beitrag zur Kenntnis der Fossilen Wirbeltiere Javas." Wetens- 
chappelijke Mededeelingen, Dienst Mijnbouvv Nederl. -Indie, I 
Teil, No. 23, pp. 104, 105. Type.— Lower jaw with third 

molar of both sides complete. Horizon and Locality. — 

Untere Schichten, Kali (ilagah, Bumiaju, Java. Type 

Figure. — Op. cit., Taf. xxvii, fig. 2 (third lower molar of the left 
side, I.M3). 

DESCRiPTioN.^(von Koenig.swald, 1933.1, pp. 104, lOfiJ: 
"Es liegt ein Unterkiefer Nor, dessen beide M3 noch komplet 
sind. M3 rechts ist isoliert, ein 'i'eil der starken Zenientlage isl 
abgewittert, so dass auch die letzten noch nicht angekauten Joche 
und der kleine Talon hervortreten. Der linke M3 sitzt noch auf 
dem leider etwas gequetschten Kieferast, der im vorderen Teil an- 



Dubois (1908, pg. 1256) gibt fur das Stegodon der Kendenglagen 
sogar 14 Joche an, doch ist aus seinen Ausfiihrungen (Abbildungen 
hat er bisher nicht vcroffentlicht) nicht zu ersehen ob er nicht 
etwa einen Talon mitgerochnot hat. Die Exemplare der Selenka- 
Expedition haben alio, und auch die mir vorliegenden bis auf eines, 
nur 13 Joche. Dieses eine von Lepen Alit bei Tingang zeigt nur 
X 11 X Joche (vergl. v. d. Maarel, pg. 143) und gehort aber 
dennoch einem typischen St. t. trigonocephalus. Dieser Fund zeigt, 
dass noch unter den Stegodonten der Kendengschichten als 
Ausnahme Formen mit weniger Jochen auftreten konnen. Es ist 
cine bekannte Erscheinung, dass oft bei spezialisierteren Arten 
auch noch Riickschliige auftreten." 

"Ein woiteres Kennzoichcn der Ziihne von Boemiajoe ist, 
dass der dicke Schmelz viel weniger intensiv gefaltelt ist als bei 
der typischen Art (Taf. xxvii). Das liisst auch ein Vergleich mit 
den Abbildungen bei Mahiin und Janensch deutlich erkennen." 

"Von Boemiajoe stammt audi noch ein auffallend grosser 
Stosszahn von etwa 3,50 m Liinge, der seiner ganzen Form nach 
wohl nur einem Stegodon gehort haben durfte." 



THE STEGODONTIN^: STEGODON 



897 




I'^ig. 784. Typo lower jaw witli right .second molar, r.Mo, in situ, of Parastegodon'! 
kwantocnsis Tokunaga, 1934, PI. ix, fig. 1, one-half natural size. 



Stegodon yushensis Yovuis, 1935 
Figure 78.") 
l''rom Yiishe, China. 

Stcgodon yuahensis Young, 1935. "Miscellaneous 
Mammalian Fossils from Shansi and Ilonan." Pal. 
Sinica, (C), IX, Fasc. 2, 1935, pp. 26-28. Type.— 
"A well presened upper left third molar." Hori- 
zon AND Locality. — "Pontian 'violet .sands' of the 
Yiishe formation from Yiishe," China. Type 

Figure. — Op. cit., PI. v, figs. 1, la. 

Description.— (Young, 1935.1, pp. 26-28): 
"With the exception of the fourth ridge and of a 
large part of the cingulum, the tooth is well pre- 
served, including the roots. Crown composed of five 
ridges, not including the posterior heel and the 
anterior cingulum. . . . The ridges are rather closely 
set; the vallej's not very deep; only slight traces 
of cement can be observed. Cingulum is \ery weakly indicated in 
the preserved parts." 



Parastegodon? kwantoensis Tokunaga, 1934 
Figure 784 
Kakio, Kanagawa Prefecture, Japan. Upper Pliocene (fide Tokunaga). 

[The description in Japanese by Professor Shigeyasu Tokunaga 
of his species Paradcgodon? kwantoensiti, 1934, was not reviewed by 
Professor Osborn. However, as the i)resent author regarded 
Parastegodon as either a progressive Stegodon or a primitive Archi- 
diskodon, his generic reference cannot be given in the present 
Memoir. The following resume in English is by Mr. Ushinosuke 
Narahara of the American Museum of Natural History. — Editor.] 

Parastegodoni kwantoensis Tokunaga, 1934. "Fossil Elephant 
teeth found at Yokohama and Kakio, Kanagawa Prefecture," 
Journ. Geog. (Tokyo), Vol. XLVI, No. 546, .July, 1934, pp. 
365-369. Type. — Portion of lower jaw with right second 

molar, r.Mo, in situ. Horizon and Locat.ity. — Kakio, 

Kanagawa Prefecture, .Japan. Ui)iior Pliocene (fide Toku- 
naga). Type Figure. — Op. cit., PI. ix, figs. 1-3. 

Description. — I'ortion of a hjwer jaw with right second 
molar, r.M2, in situ; eight and a half ridge-crests, probably ten if 
perfect; molar valleys deep, narrow, and filled witii cement. 

Length of jaw bone 253 mm. 
Second molar, r.Ma 

Length 190 

Breadth 76 

Height (outside) 35 

Height (inside) 31 

5,4 ridge-crests in 100 mm. 




The s])ecimen is primiti\e like Parastegodon \Stegodon] aurora' 

Matsumoto, but with certain differences, which he [Tokunaga] ^- „„. ^, ■ , ■ , ^ ., , r^ ■-, , ms c <■, i i 

' ' I & J Pjg YR-y. Third .superior molar of the left side, l.M'', ol biKjoitim i/iishenMs, 

believes warrants its assignment to a new species, namely, after Young, 193.5, PI. v, figs. T, la, two-thirds natural size. From Yiishe, China. 

Parastegodon'! kwantoensis. (Upper) Crown view. (Lower) External view. 



898 



OSBORN: THE PROBOSCIDEA 



"With the exception of the first ridge which has three roots 
(extcriKil, iiifomal, median) and of the last ridge, under which the 
roots build a continuous transversal lamella, each ridge has two 
distinct root s. The four jjosterior ones are laterally fused along the 
external side of the footli. and the three jjosterior ones along the 
internal side." 

"Dimensions: — [with some omissions) 

Maximiun length of the tooth 158 mm . 

Breadth at the third ridge [maximum] 83 mm. 

Height of the crown above the ciiigulum at the 

3rd ridge 36 mm . " 

"Both Stegodoa officinalis Hopw. and zdanskyi Hopw. are 
distinctly larger, and they have a higher number of ridges at the 
third molar. In size Stcgodon orientalis Owen (1870) and Stcgodon 
orientalis graiigeri Osborn stand closer; but they also have one 
ridge more at -\F and the cement of their teeth is more developed. 
S. sinensis Owen and S. aff. bonibifrons are described from frag- 
ments only. The former one is hardly separable from S. orientalis: 
and the latter one is too large." 

"We therefore ha\e to ileal with a new species, for \\ liich the 
name Stegodon yiisheiisis (sp. nov.) is proposed." 

Stegodon officinalis Iloijwood, 1935 

I'lKiircs 786, 7S7 
Said to have come fri)ni .S/.ccluiaii(?), Chiiiii. Horiz(jii unkiiowii. 

Young in his memoir "Miscellaneous Mammalian Fossils 
from Shansi and Honan," 1935, p. 27, compares Stcgodon officinalis 
Hopwood (ex iNIS.) with his new species S. yiishensis. 





rig. 78f>. Type of Stegodon offici- 
nalis Hopwood, coiisi.stiiig of tlio "first 
two ridges of ail unworii lower molar" 
((•;vst Amer. Mils. 21878), one-half 
natural size, .\fter Ho|)wood, 1935.1, 
PI. VII, fig. 3. 



I'lg. 787. "I.a.st ridge and 
talon of an unworn upper(?) 
molar," referred by Hopwood to 
Stegodon officinalifi, one-lialf nat- 
ural size. Cast Amer. Mils. 
21879. After Hopwood, 193.J.I, 
I'l. VII, fig. 4. 



Stegodon officinalis Hopwood, 1935. "Fossil Proboscidea from 
China." Pal. Sinica, (f), TX, Fasc. 3, 1935, pp. 73 75 (Hopwood. 
1935.1). Typk.- "The first two ridges of an unworn lower 

molar." C'a.stAmer. Mus. 21878. Hoiuzox anb Locality. — 

"Bought in a meilicinc shop, ilanchow. Said to have come from 
Szechuan." Horizon unknown. Typk l''iGT'iiK. Op. cit., 

1935.1, PI. VII. fig. 3. 



Material. — (Hopwood, 1935.1, p. 73): "The unworn last 
ridge and talon of a third molar. Probably from the upper jaw." 
Cast Amer. Mus. 21879. 

Description. — (Hopwood, 1935.1, j))). 73-75): "A Stegodon 
with the ridge-crests widely spaced, with four to five maniilla? on 
each ridge; first and second ridges of lower molar divided by 
a median cleft; conules of upper molars not united into ridges." 

"The anterior ridge is divided into two cusps by a deep 
median cleft. The post-trite cusp has two large cones with a third, 
smaller one, between them. The tips of the cones of the pretrite 
cusps have, with one exception, been broken off. The second ridge 
is also divided by a median cleft, but this is not so prominent as in 
the first ridge. The pretrite cusp has three cones, and the post- 
trite has two. The ridges are much narrower at the top than at 
the base, this is especially true of the second hence the labial and 
lingual surfaces of the tooth slope inwards to a very marked 
extent. The anterior surface of the second ridge is nearly perpen- 
dicular, whereas the posterior surface slopes away at a fairly 
steep angle. This difference in the slope of the two surfaces gives 
the ridge the appearance of being tilted forwards, a sign of deri- 
vation from a lower tooth. . . . There are no lateral cingula, nor 
are there cingules at the entrances to the valleys. No cement is 
pre.sent." 

"The second fragment consists of the last ridge and talon of 
a large tooth, which is doubtfully regarded as a third upper molar. 
The ridge consists of five cones which are not fused to one mass, 
but are sei)arated by definite clefts. One cleft is deeper than the 
others ; hence the cones are divided into two groups, one with three 
cones and the other with two. A large part of the talon is covered 
by cement, above which there rise the tips of four cones. . . The 
ridges are upright, with their anterior and posterior surfaces 
sloping at about the same angle; the valley between them is 
narrow, and partly filled with cement." 

Discussion by Hopwood. — "There is no jjroof that these two 
s])ecimens are correctly associated. At the same time it is evident 
from their preservation that they are both deri\'ed from the same 
deposit, which a])pears to have been of a lignitic, or peaty nature. 
They indicate the existence of a species which has hitherto been 
unknown and it is preferable to asstime an unproven connexion, 
instead of making two 'species', one of which might have to be 
relegated to the synonymy at a later period." 

"In placing the sjjecies with Stegodon rather than in Stcgolo- 
phodoN. 1 Im^e had rcgarfl to the characters of the ridg(!s of the 
holotype. One of the chief dilferences between the two genera is 
that whereas in Stegodon the cones are united into ridges, in 
Stegolophodon they remain discrete. .Vdniittedly the cones in the 
fragment of an upper tooth are still divided by deep clefts, but 
they show the first stages of union, and the cones in each half of 
the ridg(^ arc more closely joined to each other than the two halves 
of the ridges." 

"This species is the most primitive yet recorded from China. 
In the division of the lower ridges into two cusps, as well as in the 
division of the u|)por ridges and tlu; slight dis])lacement of the 
two halves, it retains characters which derive fioin a mastodont 
ancestry. Traces of this inheritance are never entirely lost among 
the Stegodontida^ and they persist with more or less regularity 
among the elephants. There are several species in a similar stage 



THE STEGODONTIN^: STEGODON 



S99 



of evolution among the fauna of the SiwaHk deposits of India, but 
they have not yet been described. So far as is at present known, 
the Indian species have bhinter cones than the Chinese form, and 
tiieir ridges are not so high." 

"Tlie trivial name refers to the chance that these specimens 
were purchased in the shop of a Chinese druggist." 

Stegodon zdanskyi Hojiwood, 1935 

Figure 788 
Horizon unknown. 

The name Stegodon zdanskiji Ilopwood first appeared (ex MS.) 
in Young's memoir on "Miscellaneous Mammalian Fossils from 
Shansi and Honan," 1935, pp. 27 and 28; to this species Young 
has referred "an ujjper left maxillary with M-, M^ in situ and. . . an 
isolated right M''' evidently belonging to the same individual, 
both found in the Pontian sands of Yiishe. A fragment of molar 
and some milk teeth may also belong to the same species." 

Stegodon zdanskyi Hopwood, 1935. "Fossil Proboscidea from 
China." Pal. Sinica, (C), IX, Fasc. 3, 1935, pp. 75, 76 (Hopwood, 
1935.1). Type.— "The first four ridges of a right third 

lower molar." Cast Amer. ^lus. 21872. IIokizon and 

Locality. — "Bought in a Medicine Shop, Shanghai." Horizon 
unknown. Typk Figure. — Op. cit., 1935.1, PI. vii, fig. 5. 




Fig. 788. Typi' of StrgDitoii zJnnski/i Ilupunod. "First four 
ridges of a right tliird lower molar" (i-ast Amor. Mas. 21872), after 
Hopwood, 193.5.1, PI. vii, fig. 5, one-half natural size. 

Description. — (Hopwood, 1935.1, pj). 75, 76): "A Stegodon 
of very large size, with eight to ten mammillae on each ridge-crest; 
anterior ridge still preserving the two cusps of tiie earlier mastodont 
ancestors, posterior ridges foreshadowing the structure of the plates 
found in the teeth of the P^lephantida"." 

"The first ridge of this specimen is di\ide(l into two cu.sps by 
a prominent mesial cleft. Each cusp, esjiecially the pretrite one, 
is divided into two cones by a cleft which is not so prominent as 
that which separates the cusps. The cones themselves are again 
divided at the summit into two or three mammillae by grooves, 
which do not jiroceed very far down the anterior and posterior 
surfaces. None of the other ridges show this sub-division; they 
arc all divided, but in a very different manner." 

"Each of the jjosterior ridges has a large cone at either end, 



with one or two mammillae at the summit. Between these cones 
is a long, ro(.>f-like portion with a mammillated crest. The number 
of mammillae varies with the manner of counting. If tlio main 
eminences alone are reckoned, there are from four to six; if each 
tiny sub-di\ision, (jr inilication of one, is counted, this number is 
increased." 

"The crown of the tooth is curved to the rigiit, and, .seen 
from the sitle, it is very slightly conca\'e. The ridges have a for- 
ward pitch ; that is to say, their anterior faces are steeper than the 
posterior. All these are marks of a right lower molar." 

"Anteriorly there is a relati\ely slender cingulum, which 
receives a nodular buttress passing downwards and inwards from 
the inner cone of the pretrite cusp of the first ridge. There are no 
lateral cingules at the entrances of the valleys. A small quantity 
of cement is present." 

Discussion by Hopwood. — "This species is the largest Stegodon 
yet recorded. Apparently it has nothing to do with the Indian 
species hitherto described; they are all smaller, have the ridges 
closer together, and more copious cement. An undescribed 
Indian species, in which the third upper molar is 243 mm long 
and 181 mm wide, is of about the same size, but its ridges are more 
of the type of StegolopJwdon stegodontoides (Pilgrim), and it is 
probably referable to the same genus as that sj^ecies." 

"Apart from its size, the most interesting feature of S. 
zdanskyi is the curious mixture of elephantine and mastodontine 
features in its ridges. In its composition of two cusps, each made 
up of two cones, and in the relations between the pretrite cusp, 
anterior buttress, and the cingulum, the first ridge is, fundamental- 
ly, that of any mastodont of the bunolophodont, or i)riniiti\e 
zygodont, type. All the other ridges are essentially elephantine in 
their structure. They each have a large cone at either end, and 
a long, mammillated, roof-like portion in the centre. If the two 
fissures which divide the tooth in this manner are deep enough, 
the partly worn tooth would show a tripartite enamel figure. 
This tripartite division is characteristic of the elephants, though it 
is also shown by some, at least of the Stegodon group (cf. Soergel, 
1912 [1912.2], p. 8, fig. 2). In most cases it consists of two outer 
rings with an ellipse between them; exactly the type of figure 
which could arise during the wearing down of a lower tooth of S. 
zdanskyi." 

"All the other remains of Stegodon from < 'iiiua liillicrto de- 
scribed (Owen, 1870 [1870.1]; Koken, 1885 [1885.1]; Schlosser, 
1903 [1903.1]; ct alii) belong to more advanced species, and, by 
a general consensus of opinion, the various writers, other than 
Owen, have referred them to such species as S. homhifrons, S. 
ganesa, or S. insignis." 

"The trivial name is given in honour of Dr. ( ). Zdansky of the 
Egyptian University, Cairo, who is well known for his valuable 
researches on the fossil mammals of China." 

Parastegodon [Stegodon?] sugiyamai Tokuiiaga, 1935 

Figure 78(1 

At Iruhi, in Saida village, Sliikoku, Japan. I'liper Pliocene or I<ower 
Pleistocene (/((/<■ Tokunaga). 

I'arastegodon sugiyanuii Tokunaga, 1935. "A Aew Fossil 
Elephant Found in Shikoku, Japan." Proc. Imp. Acad. Tokyo, 
\o\. XI, ]). 434. Type. — An up]ier UKjlar of the left side, 



900 



OSBORN: THE PROBOSCIDEA 



probabl}' an M'^. Horizon axd Locality. — "Mr. 'I\ 

Sugiyama recently sent the author [Tokunaga] a specimen of 
a molar-tooth, obtaineil during road construction at Iruhi in 
Saida village, Shikoku. The stratum from which it was obtained 
consists of an alternation of loose shale and coarse-grained quartz- 
sandstone, and probably referable to either the old Pleistocene or 
to the youngest Pliocene." Tvpk Figuke. — Op. cit., p. 433, 

crown and inner side view of molar. 

Type Description. — (Op. cil., pp. 432-434): "Since medi- 
aeval times numerous sjiecimens of fossil bones and teeth of ele- 
phants have been scooped up from the sea-bottom by fishermen's 
nets along the eastern part of the Inland Sea of Japan. How- 
ever, very seldom are sjieciniens discovered from strata exposed on 





Tvi'K OK I'AliA.STHCioDO.V [STEaODON?) SOlilVAMAl 

V'\^. 789. 'I'j'pc molur, proliably !iti l.M", of I'arnstegoihin fnigii/amai 
TokunaKa, 193."), text figure, p. 433, i)iic-lia!f natural .sizi-. From Shikoku. 
.lapau. 



the land surface along tliat part of the sea. The author is aware of 
only several ca.ses, namely, a few found in situ at Akashi on tlu; 
north shore of Ihc sea, and at Sue village, Ayauta-gun, and 
Saida village. Mitoyo-gun, Kagawa-kcn, both in the scnilh of the 
sea." 

"The j)resent specimen ob\iously belongs to Stcgodo/Uinnr 
in having a low crown, and in other ]K)ints, but in shape, size, 
number of ridges, and esi)ecially in the nature of its enamel 
plication the present fossil is referable to no known s|)ecies from 
Japan. It closely resembles Stegodon airdwaita Mart., and Slegodoii 



t.n'yunoci'phahts Mart, from Java; and is closest to the former. 
The following is the description of the specimen." 

"In the crow'n view, one side of the tooth is nearly straight 
and the other slightly conve.x. The grinding surface is oval and 
slightly convex, and its base is also slightly con\ex antero-poster- 
iorly, indicating that it is an upper left molar. The anterior part of 
the crown is broken off antl there [are] I3reser\-e(l a posterior talon, 
seven ridges and a part of another; most prt)bably the first ridge 
and the anterior talon are lost. The foremost, namely, second 
ridge has its anterior portion somewhat damaged and its outer 
portion broken off, and the next three ridges also lack their outer- 
most portions. All the ridges are distinctly exposed, not covered 
by cement. The posterior talon was co\'ered by cement and 
first showed itself by shaving off the cement; it is half as high as 
the adjoining ridge. The last ridge is just touched by wear and 
the other ridges were worn by grinding." 

"The present specimens, when complete, probably had nine 
ridges, an anterior and a posterior talon." 

"The length of the crown of the in-esent specimen is 120 
nun +20 mm, the latter term being the estimated length of the 
lost part. The width of the grinding surface is 60 mm at the sixth 
ridge (67 nun at its base). Tiie height of the crest is 34 mm at 
the worn eightli and 38 mm at the unworn ninth. The eighth and 
ninth ridge-crests decline forwards and are a little curved in side 
view, while the sc\enth but slightly declines forwards and is flat 
and not curved in the same view; all the other ridges stand up- 
right. In well worn stage each ridge-crest is narrowly rectangular, 
not strongly constricted, and has its enamel finely plicated; in 
a less ad\anced stage of wearing, it shows a large luuuber of con- 
strictions. In these features the present specimen resembles in 
general S. airawana; however, the detailed enamel plication of 
its crown-ridges is characteristic when coini)ared with the speci- 
mens of the latter illustrated by Soergel, Janensch, Stehlin, 
Maarel and Koenigswald. The unworn last or ninth ridge of ours 
has seven mammillae on crest and the eighth has seven constric- 
tions. In the seventh ridge the dentine area is expanded and there 
are still five constrictions, while in the sixth and all the others 
forwards there is no constriction. The enamel is finely and rather 
regularly plicated in the seventh and anterior ridges and never 
form rough irregular undulations exhibited on grinding surface by 
Maarel's Javan specimen, as his photograph shows, which is 
similar to our molar in the size of crown and grinding degree. The 
width of enamel measures 4 mm at the thickest portion in grinding 
surface." 

"In this specimen all its ridges except the posterior talon 
rise higher above the cement and the la.st or ninth is the highest." 

"The projjcrly of the enamel i)li(^alion of each ridge-crest of 
the author's specimen agrees with no (iescrihcd .Japanese species 
and differs in several i)oints from the ,)a\;ui sjx'cies. Moreover 
it presents several features different from fossil elephants of other 
lands. The present specimen evidently needs a new name, and 
is named, Parodegodon xiigiyamai in honor of the discoverer, 
Mr. Tsurukichi Sugiyama. The author is at- jjresent unable to 
affirm whether it represents the first or tiie second innl.n-, hut it 
evitlcntly belongs to one of them." 



APPENDIX TO CHAPTER XIV 

MATSUMOTO ON THE PHYLOGENY AND CLASSIFICATION OF THE JAPANESE 
MASTODONTS, STEGODONTS, AND ELEPHANTS (1924, 1925, 1926, 1927)' 

The numerous and diversified Proboscidea of Japan prove that this region from Miocene to early Pleistocene 
times had broad land connections with southern Asia, Burma, and India, with remote relations by migration to 
Europe and Africa. 

Subsequent to and independently of the conclusions reached by Osborn on the classification and phylogeny 
of the mastodontine and elephantine Proboscidea, Hikoshichiro Matsumoto published the systematic and 
theoretic results of his important revision of the Japanese Proboscidea in six papers, as follows: 

"Preliminary Note on Fossil Elephants in Japan," Journal of the Geological Society of Tokyo, Vol. 
XXXI, No. 371, September 20, 1924 (1924.2). 

"Preliminary Notes on the Species of Stegodon in Japan," Journal of the Geological Society of Tokyo, 
Vol. XXXI, No. 373, November 20, 1924 (1924.3). 

"Preliminary Notes on Two New Species of Fossil Mastodon from Japan," Journal of the Geological 
Society of Tokyo, Vol. XXXI, No. 375, December 20, 1924 (1924.4). 

"On Two New Mastodonts and an Archetypal Stegodont of Japan," Science Reports of the Tohoku 
Imperial University, Second Series (Geology), Vol. X, No. 1, 1926 (1926.1). 

"On the Archetypal Mammoths from the Province of Kazusa," Science Reports of the Tohoku Imperial 

University, Second Series (Geology), Vol. X, No. 2, 1926 (1926.2). 
"On a New Fossil Race of the Asiatic Elephant in Japan," Science Reports of the Tohoku Imperial 

University, Second Series (Geology), Vol. X, No. 3, 1927 (1927.1). 

His results are summed up in his three phylogenetic diagrams reproduced herewith (Figs. 791, 792, 793) and 
in the following systematic summary or conspectus compiled from notes included in Doctor Matsumoto's letters 
of July 14 and November 20, 1924, and from his two more recent papers of 1926 and 1927. Matsumoto's valuable 
letters and notes, intercalated in the English text of his published report of 1926, give us a clear statement of 
his observations and opinions, as compared with those of the present Memoir indicated in square brackets. 

Osborn (1927) adopts certain of Matsumoto's generic and specific terms (e.g., Palasoloxodon Mats., which 
antedates Sivalikia Osb.), but rejects others (e.g., Hemimastodon Pilgrim, Prostegodon and Parastegodon Mat- 
sumoto, which are antedated by genera of other authors). A full synopsis of Matsumoto's work (1924-1927)' is 
given herewith for monographic purposes, followed by Osborn's comments (p. 908 below). 

1. Genus Hemimastodon Pilgrim [ = Suina, Volume I, Chapter VIII] 
Hemimastodon Pilgrim. 

1. Hemimastodon anmdens sp. nov., Journ. Geol. Soc. Tokyo, Vol. XXXI, 1924.4, pp. 401, 405. See present 
Memoir, Chapter X, p. 457. [ = Serridentinus annedens.] 

2. Genus Trilophodon Falconer [=Trilophodon of the Present Memoir, Chapter VIII] 

Trilophodon Falconer. 

2. Trilophodon sendaicus sp. nov., Jom-n. Geol. Soc. Tokyo, Vol. XXXI, 1924.4, pp. 402, 408. See present 
Memoir, Chapter VIII, p. 280. [ = Triloplwdon ><cmlairus.\ 

'[For a continuaUon of Matsumoto's observations and theories, np to and including 1929, see Cliai). XIX, pp. 1289 to 1300 below.— Editor.] 

901 



< 

h 

1 
I/) 

Z) 
Or 

< 

CO 



CD 
< 

CL 



< 



UJ 

< 

X 
10 



uJ 
ct: 

UJ 

(- 
z 
o 

i: 



('Monastirtan 
^Tyrrheniaa 



Loom 



Milazrian. 
("MitteldUuviQlschotterl 



Cromerian -Sicilian. 
("Altdiluviolschotter) 



( PHocdnschotter) 
Calabrian 



/ Astian 
VPlaiSanc^an 
(=Merced -Etchec|oin) 



Pontian 

(= Santa Marqarita) 



"Taisunokuchi 



Sarmatian 
(sCierljo) 



Tbrtonian 
(=Briones ) 



Helvetian 
(="Tembl or) 



Burdiqauan 
(=Vaqueros) 




?fjA 



Census cfr. nippon 

JBison occidenfalis 

Stefociou sinensis £ cfn 

5. oricntalis 

E^Zephas, iast mutati-on 
( = 2.. naumanni ) 






E Zepha.5 , 
mu-iaizon of Th/ryo -BecZ 
1= ? L. hysu-driTidica ) 



has, F7mj.ta.t10n 



F8 



SfeaocZon orien^ahs 

Tlephas, mutation of /'^inafo 

I = 'Po.reiephas trogontkerii 

3tegodoT7 orientalis 



... I . -,^71 Z'umefopias, n. sp. 

Younqer Lignites ^ 
or ^ '''^' 



^ri^ia so. 



T- /- f ^ygoZophodon , n jp. 



( : 



py 



rertat cu-5 



type ) 



Lower Lignites 

^ ■^ Hanareyama ^ ^ T^roste^odon latidem. 
or 
Upper Saboyama 



Taqa or Shirado 



7^ 4 J?icrocerus , n.sp 



Ibqari or 
-^•^ KadonoSQwa /TJ J)e5mo5fyiu-:- /aponicu^, fype 



IYicarya 
JDesmostyZus 
ITurhinodelphib 



= TsuUvyoshi 



F2 



'Vicarya 
JDeSTTiostylus 

1 Jcliocetus, n. sp. 



Hiramaki 



^ »[ f-Zemimastodon, n.sp. 
Artchitherium -fauna 



Thirteen Fossil Mammal-beauing Formations of Japan. Matsumoto, 1921 

Fig. 790. Corivlatioii from tlio Lower Mioconc HunliKnliun to tlio summit of tlic Plristoi'oiic Montistiriati of the mariiir aiui fri'shwaliT bcils of .Iai)aii with 

till' marine ami coastal .stages of western Europe (HurdiKaliaii to Mona.Mtirian) and with the epicontinental and coastal »taK<s of the California coast of North 

Amcrioa (Monterey .shale, San Pablo, Saugus-Tulare, etc.). Kejiroduecd with slight additions after pen drawing by Doctor Matsumoto (letter, .July M, 1924). 

It will be ob-servcd that Matsumoto '.s summary of 1929 (.sec Chap. XIX, pp. 1290-1292) also Osburn's summary (p. 1292) alter both the above .sj)ecific 

nomenclature and the geologic levels in the Japanese Tertiaries. 

902 



MATSUMOTO ON JAPANESE MASTODONTS, STEGODONTS, AND ELEPHANTS 



903 



Older Pleistocene 



Pliocene 



Pontian 



f. filegodontoidtii 
(India) 



I (S. & E. Asia) 



T. scm/dicit'i 
(Japan) 



P. Idfidcns 

(S. & E. Asia) 

p. cautleyi 

(India) 



T. injitiniicnfs pats. 
(I (i^itro-fjrr man tens 



Auanct/s 
[Holarclio 6f India) 



Tftrolophodon 
(Holantic & India) 



Sai nmtiun 



T ffilroneri 



T. jiijiciKticua |)iiry, 

uuslrn-gernianicu-i 

(Europe} 



T purriiiiH *rv pni"s, ^, , .^ 

stibtvfuroutcns 



Tortoninn 



T. Jfitconeri 
(India) 



Helvetian 



BurdigaMan 



//. crcpusrnli H. »tnnirlcii\ Zyr/otopiifi>/o,i 
(India) (Japan) (Enropo) 



Aquitaninn Stampian 



T. ptjfcnaicus pari, 7*. uniju^tidtina 
siiblupir<nd(u.<: (Pal«arctie A India) 
(Europe) 



Hypothet irni 
p[0-a""cci'""'^ 



Hypothetical 
pTO-crcpusculi 



Sannoisian 



T'h iomia 

('^gypt) 



Fig. 791. Theoretic puylogeny of the Mastodoxtid.e (after Matsumoto, 1920.1, p. 3) 
It is observed in tliis phylum that: (1) Phiomia gives rise to Trilophodon {pvo-amiectens, He mi mastodon crepusculi [=Suinal, H. l=Serridentinas] annec- 
tens); that it extends into Trilophodon anguslidens of the Burdigalian and Sarmatian (of Europe), also into Tetralophodon and Ariancus (all unknown in Japan). 
(2) The second i)hylum includes: Trilophodon [ = Zygolophodon] ptjrenaicus theoretically giving rise to T. falconcri, Parastegodon latidens, P. sUgodontoides, and 
^tegodou] the possible derivation of Parastegodon [ = Stegolophodon] latidens from Trilophodon [= Zygolophodon] pyrenaicus is in accord with Schlosser and 
Osborn. 

3. Genus Parastegodon Matsumoto, 1924 [may equal Archidiskodon Pohlig, 1885, 1888, of the 
Present Memoir, Chapter XVI, or a Progressive Stegodon, this Chapter]^ 
Parastegodon gen. nov., Journ. Geol. Soc. Tokyo, Vol. XXXI, 1924, pp. 256, 257, 262(1924.2). 
— Stegodon mindanensis-Elephas aurorse group of Matsumoto. 
3. Genotype: Eleyhas aurorse, 

5. mindanensis-axirorse group. Low crowned, hence l:iiegodon-\i\<.i^] ridge.s close set and valleys very 
narrow, very acute toward the base of crown (contrast to Stegodon, even to those with rather high 
and narrow ridges of the Pleistocene). No loxodont sinus. 

Parastegodon [ = Stegodon] aurorse Matsumoto, 1924, op, cit.y p. 262. 

Mt. Tomuro, Province of Kaga — horizon unknown, possibly PUocene. 
Archidiskodon Pohhg, 1885, 1888, Matsumoto, 1924, pp. 256, 259 (1924.2). 

= Elephas plam'frons-meridionalis group. No known Japanese representative. 

HSee Matsumoto, 1929, "On Parastegodon Matsumoto and its bearing on the Descent of Earlier Elephants," Sci. Rept. Tokoku Imp. Univ., (2), Geology, 
XIII, No. 1, pp. 13-15 (in English).— Editor.] 



904 



OSBORN: THE PROBOSCIDEA 



(^^U^f'tM^'^^T''^ 



"lXi/l^t**'*^^»-'<»'*^ 



nJU-} 



p^^^ 



[^Aytry^^-t^^ 



C/iUceJi'-'U.-'^.^i^ 









lAlWU^ fo-VA-LLe^^v-V^ 



L•^,>tA■ 



J(yv'-AA.!«,.w^ 






4 






0'Cec<:r«''-o->-i^ 





























I 






Fig. 792. Theoretic phylogeny of the Stegodonts (after M.^tsumoto, MS. of November 20, 1924) 
Observe that Vrmlrgodon ineludcs five phyla, namely: (1) The phylum of latidens and slegodonloidcs [ = the Slegolophodon of Schle.'iinger]; that SUgudon 
iiiclii(k'.s (2) the pliylum of ch/a' and smcnsis, (3) tlic phylum of 6owiJ^!/rons, orte/i/afo, and orienlalis sliodocnsis, (4) the phylum of bombifrons, insignis, and 
ganesa, and (5) the i)hylum of airaivana and trigonncephalus. Thi.s .subdivision of tlie species of Stegodonts into several phyla, grouped under two genera, 
namely, (a) I'roslcgodoti [= Skgolophodun] and (b) Stcgodoii, with their numerous branches, agrees substantially with the conclusions reached by Osborn. 
This diagram is reproduced in Matsnmoto (1924.3, p. 339), with omission of text outside of black lines and with extension of diagonal line from boinhifmns 
(India) into the difli (India) cohunn. 

4. Genus Prostegodon Matsumoto, 1923, 1924 [ = Stegolophodon Schlesinger, 1917, of the Present 

Memoir, this Chapter] 
Prostegodon gen. nov., Matsumoto in Osborn, Amer. Mus. Novitates, No. 99, 1923, p. 2 (Osborn, 1923.601); 

also Matsumoto, Journ. Geol. Soc. Tokyo, Vol. XXXI, p. 325 (1924.3). 

4. Genotyi)o: Madodon latidens Clift. [This genus {Prostegodo7i = Stegolophodon^) includes also Mastodon 
cautleyi Lydekker and M . stegodontoides Pilgrim, etc.] 

Matsumoto, 1926.1, p. 9. — "Skull and mandible only, imperfectly known, brevirostral. Lower incisur-tusks 
might be absent, or al>ortiv(> if present at all. Intermediate molars four- or five-ridged, last molars five- or six- 
ridged. Grinders essentially lophodont, though their first and second ridges may show a slight tendency of buno- 
donty and of trefoil pattern of cusps; mesial longitudinal cleft evident; inner and outer cusps opposite, instead 
of being alternate; valleys widely open, free of cement. ... As clearly pointed out by Lydekker and by Pilgrim, 
the grouj) which now forms the present genus is very closely allied with the coiniiioii Stegodonts, tliough more 
archetypal th;in the .same. Notwithstanding that, it is perfectly distinct phyletically from the other two phyla of 

'[See page 839.— Editor.] 



J 



MATSUMOTO ON JAPANESE MASTODONTS, STEflODONTS, AND ELEPHANTS 



90.'^ 



H.^e 



]J*T-M*>lA.>^Vi 



(jaM^fiamicrn 



6 c-c-c-oe^/SJ-jtA^tl- 






(7t>ljtAcolu-(,e<^fyt^ 



fc3(^^><W 



,H. 



^*^~<JU.*^u^ 






iCiK*L^-ticrftZi»^^ ;/, ^,lc^ 



(JtLLctf<!ir>C(fii'Cn'i^ 



Z<?*»<W^ i 



*s^i f . w.i<u 3 



l^rVWX. *CoA<A. vx 









^_^l^'>«t*-AA,*i->i 






CftXoJ***^*^^ 







(oo*AX-iK.tA^ 









(A4*''>jCi^%JL*) 






fi^ 14. 









(?'r^) 









Cfc-i^UiT-*.) 









Fig. 793. Theoretic phylogeny ok the Elephantid* of A.sia and Eorope (afteb Matsumoto, MS. of November 20, 1924) 
Oliscivc that: (1) /'n;a.?<cffO(ton ineUides "insigni>i" (pars), aurora, and mindanensis; (2) Euelephas includes the " Irogonlherii" phylum [=l'arfbphas 
of Osfxiiiil and tho "jjrimigenius" phylum [ = .Ua/Hwo«<<'((.s of Osborn], both derived from the specie.s prntnmammonleus of Japan; (3) the Arcliiilinkodon 
phylum incliide.s the .speoies planifroiis, cnmnnnHis, meridional is, and hysudricus; (4) the Elephas phylum includes liy.sudriruHcus, nainadicua (in pari), and the 
true indicus of Japan, China, and India; (.")) the I'alxohxodon i)hylum includes three divisions (a) typical Palxoloxodon namadicus and imunmnni [equivalent 
to theSuaiifcia of Osborn], (b) the anhguus phylum of Europe, with ausonius and anliquus, (c) the narrow-toothed priscus, iokunagai, and atUinticus [ = Pil- 
grimia of Osborn, made by Osborn in 1934 a synonym of Palxoloxodon Matsumoto], and (6) Loxodonla including africanus. 



Tetralophodonts, viz., Tetralophodon (longirostral) and Anancus { = Dibxmodon), being distinguished from the 
former by being brevirostral, by the more distinct lophodonty, and by the single and feeble, but not doul)Ie and 
prominent, trefoil pattern of cusps of grinders, if distinct at all, and from the latter by the more distinct lopho- 
donty, by the simpler and feebler trefoil pattern of cusps and by the opposite, but not alternate, halves of ridges 
of grinders. Further, it is distinguished from the common Stegodonts by the lower ridge-formula, by the rather 
less distinct lophodonty, by the ridges of grinders consisting of smaller number of mammillae and by the entire 
absence of cement." 

Prostegodon latidens Clift. Mt. Hanare, Kuji District, Province of Hitachi; Kitayama, northern precinct 
of Sendai. (Letter of Nov. 20, 1924.) Lower Pontian. 

Prostegodon latidens Clift. Shiwogama, Miyagi District, Province of Rikuzen. Horizon: a bed of coarse- 
grained sandy tuff, probably to be referred to the Sawoyama Formation proper (Matsumoto, 1926.1, p. 
10). Probably of typical Pontian age. 
Note: La connection with a long description of this latter specimen, measurements are given of the tooth, 
a Dp* or an M\ 



90f) OSHORN: THE rKDHOSCIDl'JA 

5. Genus Stegodon Falconer and Cautley [ = Stegodon of the Present Memoir, this Chapter] 

Siegodon Falconer and Cautley. 

5. Stegodon clifti Falconer and Cautley, Journ. Geo). 8oc. Tokyo, Vol. XXXI, 1924, p. 329 (1924.3). 

Akira-mura, Kage District, Province of Ise — possibly Plaisancian-Astian. 

6. Stegodon sinensis Owen, op. cit., 1924, p. 328 (1924.3). 

Island of Shodo (Shodo-shima or Shozu-shima), Inland Sea — Milazzian-Tyrrhenian. 

7. Stegodon orientalis Owen (typicus), op. cit., 1924, p. 330 (1924.3). (Also as S. bombifrons, op. cit., p. 329.) 

Nagahama, Minato Town, Kimitsu District, Province of Kazusa; Euelephas protomammonteus 
zone, i. e., Calabrian. Togane Town, same province — Calabrian. Ikadachi-mura (formerly Riuge- 
mura). Province of Onii: wvAy possibly be Calabrian or Cromerian. (This locality has not yet yielded 
elephant remains. The specimen of 'E. namadicus' mistakably said to have come from this locality, 
is realh' from Yokosuka.) Okimisono, Uhe coal-field, Province of Suwo— probably Cromerian (Sicilian). 

8. Stegodon orientalis shodoensis subsp. nov., op. cit., 1924, p. 333 (1924.3). 

Islands of Alitsugo (Alitsugo-shima) and Island of Shodo, Inland Sea; off Nagasaki, Eastern Sea — 
Milazzian-Tyrrhenian. Kashiwazaki (town), Province of Echigo (?this form). 

9. Stegodon bombifrons Falconer and Cautley, 1924.3, p. 336. 

Higashi-Kanamachi (town). Province of Kazusa. Upper part of Plaisancian-Astian. 

(). Gknus Euelephas Falconer [ = Parelephas Osborn (Chap. XVII), Mammonteus Osborn (Chap. 
XVIII), AND Pal.eoloxodon Matsumoto (Chap. XIX), in part, of the Present Memoir] 

Euelephas Falconer. 

= E. trogontherii-primigenius group. Matsumoto, 1924, pp. 256, 258 (1924.2). 

The Upper Pliocene form of this phylum is ai/rora^-like in the shape of the basal parts of ridges and 
valleys. No loxodont sinus. 
10. EneJephas protomammonteus, sp.n., op. cit., p. 262, 265. 

Nagahama, Minato Town, Kimitsu District, Province of Kazusa (type loc.) ; Seki-mura, same 
district; Uehata, Akimoto-mura, same district; Kokubo, Onuki-mura, same district; Otomi, Mat- 
suoka-mura, same district. 

Horizon: Sanuki Bed, which is the lowest one of the gravel bed series (Narita Series, s. ext.) of 
Calabrian-Pleistocene. 

Smaller and more archetypal than E. irogontherii. 

Newly recognized in the Japanese Upper Pliocene fauna. I mean the oldest Elephas from Minato. 
Molars .small; crown low. Lower molars E. trogontherii-\\kc, while the upper molars are E. primigenius- 
like. I became convinced, after examining a number of upper molars, that it is a mnmmotli but not 
a Loxodont. 
Parelephas protomammonteus (Matsumoto) typicus, 1926.2, p. 43, PI. xviii, figs. 1-4 (holotype); PI. 

XIX, figs. 1-3 (paratype); Pis. xx-xxiii, figs. 1, 2 (referred). 

Nagahama, Town of Minato, Kimitsu District, Province of Kazusa (locality of tyj^e and paratype 
and one referred .specimen); Seki-mura, same district; Uehata, Akimoto-miua, same district; San- 
kawa Valley, Hosono, Matsuoka-mura, same district (referred specimens). 



MATSUMOTO ON JAPANESE MASTODONTS, STEC.ODONTS, AND ELEPHANTS <JU7 

10a. Parelephns protonmiamonteioi proximus, mut. nov., 1926, p. 48, PI. xxiv, figs. 1, 2 (1926.2). 

Isone, Kokubo, Onuki-mura, Kimitsu District, Province of Kazusa. Horizon: Prol^ahly "from 
a certain lower part of the Narita Series." Age: May belong to Lower Calabrian. 

11. Euelephas trogontherii (PohUg), 1924.2, p. 265. 

Lake Kasumiga-ura, Province of Hitachi; Hishiike, Hadsu District, Province of Mikawa (lost 
specimen). 

Horizon: Loxodonta {Palxoloxodon) namadicus naumanni zone, which corresponds to the Tokyo 
Bed, as well as the middle bed of the Narita Series, s. ext., just mentioned. 

Age: Cromerian ( = SiciUan). 

12. Euelephas primigenius typicus. Japan, Siberia, Eastern Mongoha (1924.2). 

The precise locahty of the Japanese specimen, which I examined, from the collection of the Kyoto 
Imperial University, subsequently \\'ill be determined by one of the professors of tliat University. 
{Euelephas primigenius sibiricus). Siberia, Eastern Mongolia; but not yet in Japan. 

7. Genus Elephas Linn^us [ = Elephas of the Present Memoir, Chapter XX] 
Elephas Linne, s. s. (1924.2, p. 256). 

= E. hysudrindicus-indicus group. 

Characterized by the total absence of loxodont sinus, and very fine, regular, and deep plication of 
enamel. 

13. Elephas indicus Linne, op. cit., 1924, p. 266 (1924.2). 

Between Tokyo and Kanagawa (Leith Adams); Yedobashi, Tokyo; Ninohe District, Province of 
Mutsu; Sapporo, Hokkaido; Province of Mino; Prefecture of Wakayama; besides several specimens 
from unknown localities — probabl}' Post-Monastirian and Pre-Neolithie. This species lived also in 
China. 

This true Elephas phylum was doubtless present in Japan, though so long confused with 'E. namadi- 
cus.^ E.g., Naumann's Yedobashi specimen; Busk-Lydekker's Kanagawa specimen; Province of 
Mino specimen and Prof. Tokunaga's photograph at your hand (this specimen was lost by the great 
disaster of last year of Japan). All the Japanese specimens are not at all or scarcely fossilized, occurring 
from loam-soil or surface deposits. May possibly be of Post-Monastirian warm period. 

13a. Elephas indicus Buski subsp. nov., 1927, p. 57, Pis. xxvii, figs. 2 and 3 (type); PI. xxviii, figs. 1, 2 
(referred) . 

Ninohe District, Province of Mutsu (type locality) ; precise locality of referred specimen unknown. 
Age: May belong to a very late geological age, such as the Post-Monastirian. 

8. Genus Loxodonta Cuvier [ = Pal^oloxodon Matsumoto of the Present Memoir, Chapter XIX] 
Loxodonta Cuv. s. ext., 1924, pp. 257, 260 (1924.2). 

Subgenus Palxoloxodon, subg. n., op. cit., pp. 257, 260 [antedates Sivalikia Osborn]. 
= E. anliquus-namadicus group. 

Genotype: E. namadicus naumanni Makiyama. Almost similar to Cromerian stage of E. antiqiins. 
Tokyo Bed, i.e., middle bed of Upper Musashino. Probably Cromerian. 



908 OSBOIW: THE PROBOSCIDEA 

Phylum A. K. melitensis-atlanticus phylum. 
14. Loxodonta {Palseoloxodon) tokunagai, sp. n., 1924, p. 267 (1924.2). 

Hira-mura, Higashi-Tonami District, Province of Etehu (type loc). 
Horizon and age unknown, may be either Calabrian or older Pleistocene. 



Phylum B. E. aniiqmis-namadicus phylum. 

15. Loxodonta {Palseoloxodon) namadicus naumanni (Makiyama), 1924, p. 264 (1924.2). 

Yokosuka, Province of Sagami (Naumann); Tabata, Tokyo (Tokunaga); Sahamma, Province 
of Totomi (Makiyama); Kasumiga-ura (lake). Province of Hitachi; Imba-numa (lake). Province of 
Shimosa; Hishiike, Hadzu District, Province of Mikawa; Nakao, Kiyokawa-mura, Kimitsu District, 
Province of Kazusa; Yamawaki, Tamaki-mura, same district ; Miyata, Miura Penninsura, Province of 
Sagami. 

Horizon: Tokyo Bed, as well as middle bed of the Narita Series, s. ext., above mentioned. 
Age: Cromerian ( = Sicilian). 

16. Loxodonta {Palseoloxodon) namadicus {typicus) Falconer and Cautley, 1924, p. 269 (1924.2). 

'E. namadicus,' younger type. Shozu-shima; and also younger terrace deposits all over Japan 
including Hokkaido. Probably Milazzian, or also ?Tyrrhenian. Inland Sea; Tsukiyoshi, Minami- 
Akita District, Province of Ugo (Matsumoto) ; Sagawa Town, Aki District, Province of Tosa. Hanno- 
sura. Province of Noto (Matsumoto) ; Sorachi, Province of Ishikari, Hokkaido (Matsumoto) ; Itsukairhi, 
Province of Musashi. 

This form in Japan consists of two types, the larger-molared type corresponds well to the Narbada 
type, while the smaller-molarerl type is commonest in Japan. Age: Milazzian-Tyrrhenian. 

Few-ridged Palsearctic Loxodonta [priscus-melitensis group]. 

E. pmcMs-like form. M3, ca. X13X; crown low; enamel thick. Hira-mura, Province of Etchii 
(horizon unknown) ; Shozu-shima, probably Milazzian. 

Subgenus Loxodonta, s.s., op. cit., 1924, pp. 257, 261 (1924.2). 

= E. africanus group. No known Japanese representative. 



OSBORN COMMENTS (1929) ON MATSUMOTO'S PHYLOGENY AND 

CLASSIFICATION (1924 1927) 

The very important but debatable result of Matsumoto's researches and observations above cited is the 
presence in Japan of the mastodontoid genus Trilophodon and of representatives of four' subfamilies, namely, the 
Stegodontinso, Manunontina^, LoxodontiniE, and Elei)hantiniiB. Without monographic examination and com- 
l)arison of these original materials and types of Asiatic and European species, it is difficult to form coi-reot 
judgments as to the true phyletic and generic relationships of many of these species and subspecies. 

'[Actually five subfamilies, as Professor Osborn described a new subfamily, the Stegolophodontin*, type Slegolophodon, in Volume 1 of the pip.seut 
Memoir.— Editor.] 



MATSUMOTO ON JAPANESE MASTODONTS, STEGODONTS, AND ELEPHANTS 



909 



rKOBOSCIDlOA DlSCOVEUIOt) IN .IaI'AN, BuuMA, JnDIA, AND ('lIlNA 



References by Matsumoto (1924-1927) 

Genus Hemimaslodon Pilgrim 

1. H . annectens Matsumoto 

Genus Trilophodon Falconer 

2. T. sendaicus Matsumoto 

Genus Parastegodon Matsumoto 

3. Genotypic species Parastegodon aurorx Matsumoto 

Genus Proslegodon Matsumoto 

4. Genotypic species M. latidens Clift 

Vjdxms Stegodon Falconer and Cautley 

5. Slegodon cliftii F. and C. 

6. Slegodon sinensis Owen 

7. Slegodon orienUdis Owen 



Slegodon orienlalis shodoensis Matsumoto 
Slegodon bombifrons Falconer and Cautley 

Genus Euelephas Falconer 



8. 
9. 



10. Euelephas (Parelephas) protomammonteus Matsumoto 
10a. Parelephas protomammonleus proximus Matsumoto 

1 1 . Euelephas Irogontherii Pohlig 

12. Euelephas primigenius 

Genus Elephas Linnseus 

13. Elephas indicus Linnseus 
13a. Elephas indicus Buski Matsumoto 

< ienus LoxodonUi ( 'uvier 

Subgenus Palaoloxodon ^Matsumoto 

14. Lo.rodonta {Palxoloxodon) tokunagai Matsumoto 

15. Loxodonla (Palaeoloxodon) namndica nainnanni 

Makiyama 

16. Loxodonla (Pahroloxodon) namadica Falconer and 

Cautley 



References by Osborn in the 
Present Memoir 
= Suina 
= Serridentinus anneclens Matsumoto 

= Trilophodon Falconer 

= Trilophodon sendaicus Matsumoto 

-- Archidiskodon Pohlig or progressive Slego- 
don 
-- Slegodon aurorse Matsumoto 

= Stegolophodon Schlesinger 
= Stegolophodon latidens Clift 

= Slegodon Falconer and Cautley 

= Slegodon elephanloides ( = cliflii) Falconer 

and Cautley 
■Slegodon sinensis Owen 
-■Slegodon orienlalis Owen 
= Slegodon orienlalis shodoensis Matsumoto 
= Slegodon bombifrons Falconer and Cautley 

■Parelephas Osborn (in part), Mammon- 
teus Osborn (in part), Palxoloxodon 
Matsumoto (in part) 

= Palseoloxodon protomammonteus Matsu- 
moto 
= Palxoloxodon protomammonteus proximus 
Matsumoto 

= Mammonleus primigenius Blumenbach 

= Elephas Linnseus 

= Elephas indicus Linnseus 

= Elephas [Palxoloxodonf] Buski 

Matsumoto 
= I^oxodonta Cuvier 
= Palaoloxodon Matsumoto; Sivalikia 

Osborn, a synonym 
= Palseoloxodon tokunagai Matsumoto 

= Palseoloxodon namadicus naumanni 
Makiyama 

= Palseoloxodon namadicus Falconer 
and Cautley 



Japan 


Burma India 


China 


X 








X 








X 








X 


X 


X 




X 


X 


X 




X 






X 


X 






X 


X 








X 




X 




X 








X 








X 








X 








X 




X 




X 








X 








X 






X 




X 



At present (1929) Osborn is disposed to review these Japanese proboscideans systematically (see pp. 1289- 
1301) in accordance with the nomenclature and in comparison with the types of species and genera illustrated 
in the present Memoir. 




Fig. 794. Primitive, intermofliato, and i)roKressive mandibles and grinding teeth of the Blepliantid;c. After Falconer and Cautley, 

1846 [1847, PI. xiii.AJ. One-.sixth natural size. 
Fig. 8. E. africanus[ = Loxodonla africana]. Least progressive, platycephalic mandible, with iirimitivo ridge formula an,l elongate rostrum. =Lo.xodontin.« 
Fig. 6. Elephas indicus. Slightly more progressive mandible, with somewhat more reduced ro.strum. - Ei-ErH.*.NTiN.E 

Fig. 7. Elep}ms hijsudricus l=llypselephas hjsudricus]. Still more abbreviated mandible, with reduced rostrum. =Eleph.\ntin* 

Fig. 3. Elephas primigenius [probably Parekphas trogontherii]. Very robust mandible, with vestigial ro.strum. -Mammontin^ 

Figs. 4,5. Elephas antiquus[ = Hesperoloxodon anliquus]. Relatively elongate mandible, with reduced ro.strum. .\ppoars as E. mcndiowdis ^ 

on plate; changed by Falconer on original plate in British Museum. - Loxodontin.b 

Fig. 2. Ekphas primigenius I = Mammonteus primigcniux]. Juvenile individual, with second inferior molars in sUu. greatly reduced rostrum. =Mammontin« 
Fig 1. Elephas primigenius [^Mammonteus primigenius]. Fully adult mandible, with broadly arched ro.strum. Third mferior grmders 

well worn, indicating advanced age. The most progressive, hypsicephalic, brachycephalic, and bathyccpl.alic type of mandible. =Mammontin.b 

910 



Chapter XV 

CLASSIFICATION OF THE ELEPHANTOIDEA BY THEIR DIVERGENT AND 
HARMONIC CRANIAL AND DENTAL CHARACTERS 

Application of principles of phylogenetic classification. Distinction of parallel and convergent 

FROM divergent AND ADAPTIVELY RADIATING CHARACTERS OBSERVABLE IN CRANIAL AND DENTAL MECHANISMS. 

Recapitulation of ancestral characters in course of development. Correlation of brachycephaly and 

HYPSICEPHALY WITH FEEDING HABITS AND HABITAT, WITH BRACHYODONTY AND HYPSODONTY. ThREE SUBFAMILY 

i'Hyla: mammontin.^, loxodontin.e, and elephantin.e. 

I. Introductory Section. 3. Seasonal changes in food of the mammoth. 

1. Principles and methods of phylogenetic classification 4. Summary of progression from browsing to grazing 

applied to the Elephantida?. dentition. 

Failure of previous dental classifications. m Vertebral Distinctions of Elephas, Loxodonta, Mam- 

L lassihcation by cranial and dental characters. monteus and Parelephas 

Cranial mechanics of Elephas (Weithofer, 1, Vertebral formula; in the above genera. 

Osborn, Ciregory). ^ 

Comparative cranial sections of elephant skulls. IV. Synopsis of Subfamily Classification of the Elephant- 

Ontogenetic cranial changes in Elephns indicus. Oidea. 

Generic contrasts in cranial sections: Loxodonta, 1. Subfamily Mammontinse, originally browsing, pro- 

Parelephas, Mammonteus, Elephas, and Archi- gressive to extreme grazing type. Genera: 

diskodon. Archidiskodon, Parelephas, Mammonleus. 

II. Dental and Cranial Adaptation to Prevailing Feeding ^- ^"bfamily Loxodontinae, feeding habits chiefly 

Habits the Key to Phylogenetic Classification. bi-owsing. Genera : Loxodonta, PaLroloxodon , 

,_,.,,. ... , Hesperoloxodon. 

1. Kidge-plate formulae ot primitive and progressive o r, i c •, ^i i , • i ■ a u ■ , 

genera in adaptation to prevailing h.t^^its of ^- '''"M^'^™ ^ ^^^'^''^l'^' '^'^^1^ browsing, second- 

feedine: '^'' •^ grazing. Genera: Elephas, Hypselephas, 

J^lciiclcDhas 

2. Food of the Indian and African elephants and of the 

mammoth. V. Final Summary of Chapter. 

I. INTRODUCTORY SECTION 

In the years 1902-1903 the author, with the aid of his colleague William King Gregory, devoted many month.s 
of research to the ontogenetic development of the elephantine cranium, also to the method of analyzing the 
actual forms and proportions of the elephantine cranium by means of sections made with bent wire. Only by such 
means can the cranial proportions be effectively studied and illustrated. Gregory pubhshed part of the results 
of these ontogenetic studies in his paper of 1903, entitled, "Adaptive Significance of the Shortening of the Ele- 
phant's Skull," as quoted below. The complete results obtained by the wire sectional method, as illustrated in 
figures 801, 803 to 814, are here pubhshed for the first time. 

1. PRINCIPLES AND METHODS OF PHYLOGENETIC CLASSIFICATION APPLIED TO THE 

ELEPHANTID^ 

The secret of phylogenetic classification in the family ElephantidiB is to discover characters in which the 
different branches diverge from each other. Hitherto students of phylogeny have been deceived by the more 
numerous characters in which they converge, in other words, by their parallelism. In the present chapter it is 
shown that whereas all the Elephantidae converge in their proboscis, in their tusks, and in certain progression in 
their grinding teeth (e.g., addition of ridge-plates), they also diverge in the profound proportional changes in the 
cranium, with which natiu-ally the tusks and the grinding teeth are harmonic. While outwardly similar, the crania 
in different subfamilies of elephants are inirardly profoundly divergent. These divergences are in continuous adapta- 
tion to the prevailing habitat and feeding habits. 

911 



912 OSRORN: THE PROBOSCIDEA 

Step by step it has been observed by Osborn, in comparing the crania of the Hving and fossil elephants, that 
all the dental and cranial characters are harmonic, that there is a close adaptive correlation between the form and 
position of the grinding teeth and of the component parts of the cranium. Moreover, the lines of descent, which 
are apparently dissimilar in superficial dental structure, are really profoundly related both in dental and cranial 
harmony. From these comparative observations, which will now be described in detail, has arisen the present 
phylogentic classification of the Elephantidse. After a review of the cranial and what is known of the vertebral 
characters, the three subfamilies and included genera will be fully defined in Section IV of the present chapter 
(cf. p. 932 below). 

Suborder or Superfamily : ELEPHANTOIDEA Osborn, 1921 
(See Volume I of the present Memoir, pp. 22-33, figs. 3, 4, 5, 7, 8, and PI. xi) 
Original reference: "The Evolution, Phylogeny and Classification of the Proboscidea" (Osborn, 1921.515, pp. 2 and 4). 

[The present chapter was written before Professor Osborn separated the Stegodontoidea from the Elephant- 
oidea, as fully set forth in the introductory pages of the preceding chapter. Consequently in the following citations 
the reader should disregard the inclusion of the Stegodonts in the Elephantoidea, retaining only the subfamilies 
Mammontinae (Archidiskodon, Parelephas, Mammonteus), Loxodontinae {Loxodonta, Palxoloxodon, Hesperoloxo- 
don), and Elephantinae (Elephas, Hypselephas, Platelephas) . 

It was in January, 1921, in his article on "The Evolution, Phylogeny and Classification of the Proboscidea" 
that Professor Osborn first published the term Elephantoidea (Osborn, 1921.515, pp. 2 and 4 respectively): 

IV. — Elephantoidea to include the Elephantinse, Loxodontinae, Stegodontinse, and Mammontinae. 
. . [p. 4] One prime distinction in this superfamily 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 tlevelopment 
into trefoils, so characteristic of the mastodontoids, and the early tendency to form evenly transverse, 
more or less mammillate, crests which become in the highest degree hypsodont and polylophodont in 
adaptation to chiefiy grazing habits. 

Subsequently on sectioning certain of the molars of the Stegodontoidea and the Elephantoidea it was found 
that the valleys separating the adjacent ridges were V-shaped in the former and U-shaped in the latter. 

Family: ELEPHANTID^ Gray, 1821 

Original reference: Gray, "On the Natural Arrangement of Vertebrose Animals," London Medical Repository, 1821, XV, No. 
88, p. 305. 

Syn. : Elepluisideie Lesson, 1842, p. 156; Elephantida; Bonaparte, 1838, p. 112, and 1850; Girard, 1852, p. 326; Zittel, 1891, p. 458. 
While Gray in his classification of the Proboscidea (op. cit., p. 305) assigned the terms Elephantidse and 
Mastodonadae respectively to the elephants and mastodonts, Girard (1852, p. 326) was the first to use the form 
Mastodontidae. 

Family Defi.nition (Gray, 1821, p. 305). -PROBOSCIDLE Fam. 1. ELEPIIANTIDyE.— 

Teeth, two grinders in each jaw, composed of transverse vertical lamina>, enveloped in enamel, and 
soldered together ))y a cortical substance. 

Elephant, Elephas Lin. E. Indicus Cuv. 

Osborn (1910.346, p. 558): Fam. Elephantidse. Dinotheres, Mastodons, and Elephants. 



CLASSIFICATION OF THE ELEPHANTOIDEA 913 

Osborn (1925.662, p. 28): Family: ELEPHANTID/E, distinguished by plated grinding teeth 
developing out of the more or less closely compressed, serrated ridges of Stegodon into the broadly plated 
grinders of Arcfiidiskodon, the lozenge-shaped grinders of Loxodonta, and the compressed, finely plated 
grinders of Parelephas, of Mammonteus, and of Elephas the type genus of the family. 

It will be observed, however, by referring to page 25 of Volume I (1936) of the present Memoir, that Professor 
Osborn altered his opinion as to the development of the grinding teeth of the Elephantidse (p. 942 below) "out 
of the more or less closely compressed, serrated ridges of Stegodon": "It has been assumed by practically all 
palseontologists that the Elephants were descended from the Stegodonts. This assumption now proves to be 
erroneous, for neither tlie 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." — Editor.] 

Subfamily: Mammontin^ Osborn, 1921 
The Mammontinse include the southern mammoth {Archidiskodon) , the north temperate mammoth {Parele- 
phas), the true northern mammoth {Mammonteus)} 

Genus Archidiskodon = the Southern Mammoth, 

inchitling Archidiskodon proplanifrons, siibplanifrons, planifrons, A. meridionalis, A. imperator, and other 
species, chiefly ranging in the south temperate zone. 

Genus Parelephas =the North Temperate Mammoth, 

including the species Parelephas trogontherii, P. jeffcrsonii, P. trogontherioidcs, P. columbi, usually inter- 
mediate in geographic range and climatic life zone between the southern mammoth and the northern 
mammoth. 

Genus Mammonteus^ =the Northern or Woolly Mammoth, 

typified by Mammonteus primigenius, including also M. primigenius americanus, M. primigenius alasken- 
sis, M. primigenius compressus, and other subspecies of the northern steppes and tundras. 

Subfamily: Loxodontin.e Osborn, 1918 
The Loxodontinae include the three genera Loxodonta, Palseoloxodon, and Hesperoloxodon. 

Genus Loxodonta 

including Loxodonta africana, and probably L. zulu, L. prima, L. africana var. obliqua, L. subanliqua, of Africa. 
Genus Palxoloxodon 

including especially Palxoloxodon namadicus of India; P. melitensis, P. mnaidriensis, P. falconer i, P. 

Cypriotes, and P. creticus of the Mediterranean Islands, also the Palaeoloxodonts of Africa, of .Jajjan, and 

of Java. 

Genus Hesperoloxodon 

including Hesperoloxodon antiquus, H. antiquus italicus, H. antiquus germanicus, etc. of Europe. 

Subfamily Elephantin^e: Osborn, 1910 
The Elephantinae include the typical or true elephants of India, consisting of the three genera Elephas, 
Hypselephas, and Platelephas. 

Genus Elephas 

including the typical Elephas indicus and the geographical varieties of E. indicus bengalensis and E. indicus 
ceylanicus, also E. sumatranus. 

Genus Hypselephas 

including the extinct Pleistocene species Elephas [Hi/psekphas] hysudricus of India. 

Genus PlutcUphas 

including Platelephas platycephalus of the Upper Phocene or Lower Pleistocene of India. 



HThe validity of the generic term Mammonteus is doubtful— see Chapter XXI, pp. 1365-1367, below, on the nomenclature of the Proboscidea.— Editor. 



914 



OSBORN: THE PROBOSCIDEA 




Fig. 79.5. Gener.\l Climatic Distribution op the Subfamilies of the Elephantoidea and Stegodontoidea, Including 

Theoretic Migration Lines (1938) 
noutii polak pliojection prepared by the american geographical society in 1924 

1) The Stegodonts (Stegofhu): Soutlieni and E.astern A.sia, .Jiipaii, and the East Indies. 

2) The Southern Mammoths (Archiriiskodon): Africa, Southern Europe, Eughmd, Southern A.sia, the United States, and Mexico. 

3) The Northern Mammoth.s (^fammonl(■us): Circumpohir distribution extending southward to the 40th parallel in late Pleistocene times. 

4) Troooxtherian Mammoths (I'ardcphas): Southern and Eastern Europe, the United States, Mexico, and South America; elsewhere theoretic (?) 
migration lines. 

o) The Loxodontine Elephants {Loxodonla. PaUcaloxodun, JIcKpcroluxodon): Southern Europe, Asia, and the continent of Africa. See figure 1242 for 
living African elephants. 

6) Indian Elephants {EUiihus, Ili/jiseh ijlma, T'lntdc jihax) , recent and fossil: Southern Asia and the East Indies. See also figure 1242. 



Failure of Puevious Dental Classifications 
Falccjiirr's classification (1857, p. 318, and 1868, Vol. I, p. 82 et seq.) into three subgenera was based chiefly 
on the grinding teeth: Subgcn. i. Stegodon. Hiibgen. 2. Loxodon. Subgcii. 3. Euelephas. Pohlig's classification 
(1888, p. 138) was also based solely on the grinding teeth: Polydiskodon, Archidiskodon, and Loxo{disko)don. 
We find that clas.sification of the Elephantidse by dental characters alone cannot be i)hylogenctic. 

Classification by Cranial and Dental Characters 

The phylogenetic classification in this Memoir is based on the evolution of the craiiiuin and jaws, harmonic 
with the grinding teeth, especially with the thirtl superior and inferior molai's, AP-I\l.i, w iiich pro\es to afford an 
absolute means of distinguishing generic phyla from each other. 



CLASSIFICATION OF THE ELEPHANTOIDEA 915 

CuviER.— From the first distinction by Cuvier of the African from the Indian elephant and from Elephas 
primi genius,^ all authors have given more or less attention to the external characters of the crania of the Ele- 
phantidaj, although they have not made the cranium the chief basis of classification. 

Cranial Mechanics of Elephas (Weithofer, Osborn, Gregory) 
Weithofer. — In 1890 Weithofer made a detailed comparative and functional study of the elephant's 
skull, and concluded that the ontogenetic mechanical changes in the skull, observable during individual growth, 
as affecting phylogeny, such as the fore-and-aft compression and vertical heightening and deepening of the skull, 
the wide separation of the inner and outer tabulae of the bones and the cancellous condition of the diploe, the 
forward shifting of the orbits from a point above the anterior grinders, etc., are primarily correlated with the 
prodigious development of the tusks — weapons and crowbars whose effectiveness increased with and reciprocally 
hastened the phyletic advance in body dimensions. Gregory (1903.1) did not accept Weithofer's conclusion that 
the tusks were the chief cause of fore-and-aft compression, but showed that the proboscis is another and perhaps 
the chief factor in the extraordinary process of fore-and-aft cranial compression, or cyrtocephaly. 

Osborn-Gregory. — Studies of the elephant cranium as a whole were undertaken by Osborn (1902-1903) 
with the cooperation of William K. Gregory. Gregory (1903) in his "Adaptive Significance of the Shortening of 
the Elephant's Skull" attributed the fore-and-aft compression principle to the enlargement and backward shifting 
both of the proboscis and of the tusks. His principal conclusions and explanatory figures may be freely summarized, 
without quotations, as follows: (1) First factor. Lengthening of the proboscis correlated with shortening of 
cranium and of neck pari passu with lengthening of limbs and increased height; also with downward shifting and 
elongation of tusks (vide Weithofer). (2) Second factor. Backward shifting of weighty tusks and of proboscis 
diminished the anteroposterior space (i.e., brachycephaly) for grinding teeth (Dp 2-M 3). (3) Third factor. 
Rapidly heightening grinding teeth (i.e., hypsodonty), together with large backwardly and upwardly growing 
molar-alveolar pouch. 

Harmonic with the enlargement and backward shifting of the tusks, the elongation of the proboscis, the 
fore-and-aft compression of the whole cranium, and the vertical elongation (hypsodonty) of the grinding teeth, are 
the following proportional changes in the cranium of Elephas: (a) Hard palate tilted obliquely upwards; (b) 
l)alatines reduced lateroposteriorly and widely divergent posteriorly; (c) posterior nares pushed very far back; 

(d) large vertical pterygoid wing of ahsphenoid encircles and functionally replaces posterior molar-alveolar pouch ; 

(e) foramen ovale of alisphenoid shifted backward and outward, becomes confluent with foramen lacerum medium; 

(f) presphenoid, basisphenoid, basioccipital thicken in median plane, so that basis cranii point sharply downward 
at an angle of 90°+ with occipital plane; (g) tympanic bullae flatten and become closely appressed to the skull. 

The following is cited from Gregory (1903.1, p. 391) : 

. . . the obliquely placed external portion [Fig. 797] of the orbito.sphenoid . . . has been squeezed into a long, thin process; 
internally [Figs. 798, 802] the anterior edges of the basisphenoid are directed outward and backward; both internally and 
externally the optic foramen, foramen lacerum anterius, and foramen rotundum, in the order named, are obliciucly arranged 
on descending levels from within outward and from in front backward, the whole region ha\-ing been thickened by the separation 
[diploe] of the inner and outer tabula; of all the bones, and also sharing in the upward-and-backward tilting of the nasal region 
and in the general fore-and-aft squeezing [ = cyrtocephaly] of the skull, the end result being that the foramina have been pulled 
out into long tunnels running obliquely forward, outward, and downward; especially internally the fore-and-aft extent of the 
alisphenoid proper is brief. Internally the skull has shortened up, one might almost say in bellows fashion, with the optic 
foramen on each side at the apex of the internal transverse folding [Figs. 79S, 802], the ridge of the 'lesser wing' of the human 
sphenoid. As the skull has also expanded transversely [ = brachycephaly], the general effect of the internal \iew of the skull is 
thus that of compression around the center (represented by the basisphenoid) and increa.sing expansion toward the periphery- 
somewhat recalling the conditions of the domehke hmnan skull. 

'Cuvier's keen jxTception of the cranial distinctions of Elephas primigenius, E. nfricanus, and E. iiulicus arc shown clearly in his definitions of these three 
species in the first edition of his "Ossemens Fossilcs" (1806.1, pp. 262 264), in which the three crania are figoi-ed side by side (cf. Fig. 992 of the present Memoir) 
in anterior and lateral aspects. See further comment in Chapter XVIII {Mammonkas) on Cuvier's treatment of E. primigenius. 



916 



OSBORN: THE PROBOSCIDEA 



The above principles of the ontogenetic and phylogenetic adaptations in the Elephas indicus cranium cover 
three chief cranio-mechanical factors in the order named: 

(1) The lengthening of the proboscis or trunk of first importance as affecting aUke males and females. 

(2) The lengthening and increasing weight of the tusks, as affecting chiefly the male, and to a much less 
degree, the female cranium. 

(3) The heightemng of the molar crowns, especially of M^-Mg, as affecting the male and female cranium aUke. 





Juvenile Cranium of Asiatic Elephant 
Fig. 790. Palatal view (left) of jiivcnilc iTaniiini of A.siatic elephant {Elephas i7uKcus); superior view (right) of same cranium. AftcT Osborn and Gregory 
(see Gregory, 1903.1, figs. 1 and 2, ]>]>. 390 and 391). 



P.nix. — Premaxillary. 

Mx. — M axillary . 

.V/x. p. — Maxillary iKimli for molars. 

.Ua.— Malar. 

Po.f. — Postorbital ridge of frontal. 

pal. Mx. — Palatine ledge of maxillary. 

PJ.— Palatine. 

Vo. — -Vomer. 

Ps. — Presphenoid. 

Bs. — Basisphenoid. 

Bo. — -Basioeeipital. 



p. As. — Pterygoid wing of .ali.sphenoid. 

ft.— Pterygoid. 

Sq. — Sfiuamosal. 

Ex. o. — Exoeeipital. 

pg. — Postglenoid ledge of sriuamosal. 

p.ty. — Pcst-tympanic ledge of .squa- 
mosal, which with pg. forms a 
secondary external auditory meatus. 

Ty. — Tympanic bulla. 

ly. p. — -Anterior process of tymijanic. 



Ip. h. — Tympanohyal. 

cu. — Eustachian opening of tympanic. 

a. p. f. — Anterior palatine foramina 

(canals). 
i. 0. f. — Infraorbital foramen. 
p. n. — Posterior nares. 
p. a. s. — Alisplienoid canal. 
/. I. in. — Foramen laeerum medium. 
f.ov. — Foramen ovale (confluent with 

/. I. m.). 



i. e. c. — Canal for internal carotid 

artery. 
/. sL in. — Stylomastoid foramen. 
/. I. p. — Foramen laeerum j)osterius. 
c. /. — Notch, a vestige of condylar 

foramen (?) (confluent with/. I. p.). 
i.-. — Tusk, 
p' {dm 2). — Third premolar (or 

second deciduous molar of authors). 
p* {dm 3). — Fourth premolar (or 

third deciduous molar of authors). 



Osborn (1904-1924), in comparative observations, is inclined to attribute cciually great, if not greater cranio- 
mechanical influence to the vertical elongation ( = hypsodonty+bathycephaly) of tlie posterior grinding teeth in 
relation toadajjtive radiation in feeding habits, as .shown in a (H)Mipnris()ii of (Fig. 806) llic |)iin'ly I Jiowsing African 
elei)hant {Loxodonta) with the browsing and grazing Indian elei)hant (Klcphd^) and with the cliicfly grazing north- 
ern mammoth (Mammonteus). 




Po.f. 



Ex. 0. 




Fig. 797. (Right figure) Orbitosphenoidal region, left side. From Osborn and 
Gregory (see Gregory, 1903.1, p. 393, fig. 4). Tlic view is obliquely from the 
side and from below the malar bone (ef. left figure— Fig. 3 of Gregory, 1903. 1 ). 

(Left figure) Juvenile cranium of Asiatie Elephant (.see Gregory, 1903,1, 
p. 392, fig. 3). Compare figure 796. 



Po. /.— Postorbital ridge of frontal. 

Fr.— Frontal. 

0. s. — External process of orbito- 

sphenoid. 
As. — Alisphenoid. 
Sq. — Squamosal, 
p. .4s.— Pterygoid wing of alisphenoid. 



Mx. p.— Maxillary i)oueh for molars. 

op. f. — Foramen optieum. 

/. I. a. — Foramen laeerum anterius. 

/. r. and a. a. s. — Arcade leading to 
foramen rotundum and anterior 
opening of the alisphenoid canal. 



Elephss indicus male juv, 
Amer. Mus. 44 Ret. 




process for sttac/iment 
"/ tympgnohya/ 



masfoid region of Penotic 
process o/tymp. tu//a 



InI'Antile Cu.\nium of AsrATic 
Eleph.wt 

Fig. 798. Infantile basis cranii. 
Skull of Elrphas indicu.s (.\mer. 
Mus. Dept. Mam. 44); same indi- 
vidual as that represented in figure 
799. One-half natural size. Ob- 
serve identification of basicranial 
foramina by Gregory in 1903. 

Compare infantile separation with 
the adult fore-and-aft <-ompres.sion 
and confluence, cyrtoccphaly (Fig. 
SCO), of the liasicranial foramina. 



918 



OSBORX: THE PROBOSCIDEA 



Assigning, therefore, a proportional influence to the three great cranio-mechanical factors, (1) tlie proboscis, 
(2) the tusks, and (3) the grinders, relative to the frevaiUng feeding habits, whether purely browsing, browsing and 
grazing, or purely grazing, we shall come very near a complete mechanical interpretation of the cranium of the 
elephants in the fiv^e generic phyla which we shall presently examine. 



Pa. 




IxFANTiLE Cranium of Asiatic Elephant 
Fig. 799. Rcfcrrod infantile occiput and jaws of Elcplias 
indicus (.\mcr. Mu.s. Dcpt. Mam. 44). This is a much younger 
si>ecimen than tliat figured (Fig. 797) by Gregory in 1903, p. 
392, fig. 3, whicti shows a long shallow jaw. One-fourth natural 
size. 



E. indicus ' ^ 

Afner. Mus. 3819 Ref. 




Adult Cr.\niu.m of .\si.\tic Elephant 
Fig. 800. Adult palate of Indian elephant {Klcphas indicus 

beiigalensis), Amer. Mus. Dept. Mam. 3819. This is a middle-aged 

specimen with superior molar, ?M', showing eleven to twelve worn 

plates. One-eiglitli natural size. 

Compare the fore-and-aft compression ami loiiflucnce (eyrto- 

cephaly) of the ba.sicranial foramina with the relatively primitive 

condition in the infantile cranium (Fig. 798). 



Comparative Cranial Sections of Elephant Skulls 
OsBORN. — In order to ascertain the more profound changes which have taken place in the evolution of the 
crania of the elephants, Osborn and Gregory, as stated above, employed in 1902-1903 a method of sectioning the 
skull (Fig. SOI) in four different planes, by means of copper wires. A -similar method had been used (Osborn- 
Gregory) in studying the cranial mechanics of the peri.s.sodactyl family of titanothercs. 



CLASSIFICATION OF THE ELEPHANTOIDEA 



919 



These planes were established with reference to the basicranial axes, after the manner of Huxley, Flower, 
Lankester, and others in their studies on the cranial axes of other ungulates ; also with reference to the horizontal 
grinding surfaces of the molar teeth. 





The Five Wire Section Lines (A-E) 
Fig. 801. Kc}' to Sections. Young Elephas indicus bcngalensis cranium apparently of the bioail narial 'Dauntela' variety. The 
nianner in which the cranial sections are recorded in the figures of this chapter is shown in the above key diagram to sections, also in 
diagrammatic figures 803-814. 

C, Midvertical, sagittal, nasals to vertex of the occiput. />, Para-occipitofrontal, sagittal, longitudinal rim of anterior narcs to 
occipital condyles. A, Horizontal nasal. E, Occipitohorizontal through back of occiput. B, Vertical transverse frontals, intertemporal. 

The intensive application of this method by Osborn in 1924 revealed the following Unes of phylogenetic 
divergence: (1) Progressive brachycephaly of the skull involves fore-and-aft shortening of the individual bones 
( = cyrtocephaly), also an expansion in the vertical planes ( = hypsicephaly). (2) Shortening ( = cyrtocephaly) 
and deepening ( = bathycephaly) of the temporal fossae and upward expansion of the grinders ( = hypsodonty) 
and of the alveolar pouch profoundly alter the bones and foramina of the sphenoidal region. We may summarize 
these proportional changes as follows: 

Cyrtocephaly : Fore-and-aft faciocranial abbreviation. 

Cyrtodonty: Fore-and-aft molar-crown abbreviation. 

Hypsicephaly: Vertical heightening of the cranium and jaws. 

Hypsodonty: Vertical heightening of the molar crowns. 

[Brachyodonty: Vertically low molar crowns.] 

Acrocephaly: Vertical heightening of the occipitofrontal apex. 

Bathycephaly: Vertical deepening of the basicranium, molar alveoli, and jaws. 

I5rachycephaly: Broadening of (a) the occiput, or of (b) the zygomatic arches. 

( 'yP^ocephaly : Downward flexure of the facial to the basicranial axes. 

[Orthocephaly: Lack of inclination of the basifacial to the basicranial axis (cf. p. 924 below)]. 



Ontogenetic Cranial Changes in Elephas indicus 
Api^lying the methods described above, it is desirable to examine the ontogeny of the Elephas indicus cranium, 
in accordance with the biogenetic law that youthful crania and jaws in all phyla of the Elephantidx more or less 
closely resemble each other, whereas adult crania and jaws very widely differ frorn each other. 



920 



OSBORN: THE PROBOSCIDEA 



The chief ontogenetic or growth and age changes are clearly displayed in the accompanying series of Jlgures 
(Figs. 801, 807, 809, 799, 798, 796, 797, 802) illustrating the transformation from the infantile to the adult condition 
in the cranium of Elephan indicus, which should be examined in connection with the comparative sections, also 
with the adult crania figured in the present and succeeding chapters on Mammonteus, Loxodonta, and Elepltas. 




Fig. 802. Interior view of skull of youiif^ Asiatic Elopliiint, after GrcRory, 1903, PI. xxiii. Same .skull as tliat rcpri'.si'iitwl in figures 7'J(i, 7'J7. 

Crib. p). — Cribriform jilate. f.l.p. — Foramen lacerum posterius. 

orb.sph. ridge — Orbitosjihenoid ridge. i.c.can. — Internal carotid canal. 

sel.tur.— Sella turcica. tym. — Tympanic bulla. 

Fr. — Frontal. Pa.int.tab. — Internal table of parietal. 

f.r. — Foramen rolundum. Fr.ex.tab. — Outer table of frontal. 

f.l.a. — Foramen laeenim anterius. Basisph. — Basisphenoid. 

f.l.ni. — Foramen lacerum medium. Ali.sph. — Ali.sphenoiil. 



Fronto-occipital and Frontal Growth Curves (Figs. 803, 804). — As shown in midsection D, the frontal 
profile of the infantile cranium of Elephas indicus resembles that of the adult jirohle of Loxodonta africana. The 
juvenile profile displays the rounded and expanded fronto-occipital curve characteristic of E. indicus, in contrast 
to the angular occipitofrontal curve of L. africana. The adult fronto-occijiital profile is a uniformly rounded 
dome. The old male pi'ofile exhibits an acrocejjhalic dome, which attains the same heiglit as in the highest 
Parelephas jeffersonii (Fig. 810) and is distinguished by the rounded bulbous frontals. 

The infantile frontals resemble those of L. africana. The juvenile frontal curve is elongate, (;onvex superiorly, 
concave inferiorly or convexo-concave, in contrast to the uniformly convex frontals of L. africana (Fig. 811), 
or the uniformly concave frontals of Parelephas and Mammonteus in section (Figs. 805, 800, 810, 811). 



CLASSIFICATION OF THE ELEPHANTOIDEA 



921 



Nasal Growth Stages (Fig. 807).— The nasal growth stages reflect the increasing aljl)reviation (cyrto- 
cephaly) and broadening (brachycephaly) of the cranium. The infantile nasals are short, pointed, and narrow. 
The juvenile nasals are broader and uniformly arched. The adult nasals are broad and truncated. The old male 
nasals are extremely broad, truncated, and abbreviated. 



growth staqe^ 




Fig. 803. Front o-occipital gniwtli ciirvt'S of vertex. Elei>has indicus 
cranium, one-sixth natural size. This Section D passes between tlie oeeiijital 
convexities. The increasing occipital convexities are shown in tlie passage from 
the infantile to the old male stages of growth. The section is taken at the left 
of the medial line. 



SinJicui, ijrowt^ stages 




VrrTicaf sagittal 
nasals To vsrtsx oeeiput 



Fig. 804. Growth curves of vertex. Midfrontal vertical 
Section of Elephas indicus skull, ono-sixth natural size. The 
increasing concavity of the lower part, and convexity of the upper 
part, of the forehead in the passage from the infantile to the old 
male stages of growth are shown. 



Fronto-intertemporal Growth Stages (Fig. 808).— Like the nasal growth, the fronto-intertemporal 
growth reflects increasing abbreviation (cyrtocephaly) and broadening (brachycephaly) of the cranium, as 
observed in the infantile, the juvenile, the adult, and the old male growth stages shown in Section B. As compared 
with the convex profile of L. africana (Fig. 813), the midfrontal profile of E. indicus is plane. The angle between 
the superior fronto- and laterotemporal surfaces of E. indicus (Figs. 808, 813) is sharply angular, whereas in L. 
africana (Fig. 813) and Parelephas jeffersonii (Fig. 813) the frontotemporal union is rounded or gently convex. 

OcciPiTOHORizoNTAL Growth Stages (Fig. 809).— The extraordinary brachycephaly of the E. indicus 
cranium is most clearly displayed in the rapid lateral expansion of the occiput and deepening of the interoccipital 
space. As clearly displayed in figure 809, even the infantile cranium is relatively broad in horizontal section. In 
the juvenile stage the bulbous backward growth of the occiput begins with the deepening of the interoccipital 
space. In the adult the occipital expansion describes a uniformly convex rounded curve. In the old male the 
winglike lateral expansion results in an acute angle between the occipital and temporal faces. This shows that the 
extreme of brachycephaly is attained in the broadening of the occiput. In the comparative figure 814 it appears 
that this extreme occipital brachycephaly is much greater in the old male of E. indicus than in the adult P. jeffer- 
sonii; it also exceeds the occipital brachycephaly of the L. africana cranium. 

Palatal Abbreviation (Figs. 796, 800).— The palate of the juvenile cranium (Fig. 796, left), as compared 
with the adult palate of E. indicus bengalensis (Fig. 800) and the adult palate of L. africana (Fig. 1061), proves 
that in the transition from the juvenile to the adult stage there is a very marked abbreviation (cyrtocephaly) and 




anttrtar narta 



longitudinal 

Condifle to 

middle of n'^fiT half ef paritfala 



Fig. S05 ^.- _^- 

Pkki.imixaky Stuuv of Fronto-occipital-basilar Planes •' -^ 

Fig. 80."). Comparative para-opfipitofrontal section D (Fig. 107). Soction carried / / j 

down through tlie bonis cranii to show the widely different occipitofrontal planes of the / / ,' 

skull in the three great subfamily divisions of the Elephantidffi. Reduced to one-sixth ..--J^h ' 

natiMiil .size. For basis cranii substitute basioccipital plane (see Fig. 1192 below). ..•■' I __/ 

/•■ ',y' 

EbErHANTiN.E = £'icp/ms indiciis, Indian elephant contour. Para-occipitofrontal _^-' 

union a right angle, rounded. 

LoxODONTiN.E = Loxorfon(a africana, African elephant contour. Occipitofrontal 
union rectangular. -maU; 

I , Tn. ^orimiyefiiu.s ^ai Mus 

MAM.\iONTix.E = PnrWcp/m.s iVffcr.'ionii,' Mammoth contoin-. Occipitol'rontal union ? i , ***" 

* .* ^ _^ , mf 771^ /..africana 

nciile. ' 

Tliesc three sections are made from adult skulls in the American Mu.seum collection, 
n;iniclv, of Eli/ilias iii<licu.'<, of Loxodania africana, and of a young male skull of Parcle- ^^^^ 

l,h„s j'effersonii (Amcr. Mus. 1447.5).' Compare sections in figure 806. ' ■ '"" """ ^'•""""■'■'"^ .t.^/v,^. 

m ? T.jrfftrserxii TItf. Nat Mils. 
I . -—.. . . ■,...,..-^ ".-.TT -.V.l' /026/ 

m? m? r indtcus 

Fig. SOfi 

Later Study ok Fkonto-occipital-Molau-3 Planes 
Fig. SOO. Para-occipitofrontal soction /) (Fig. SOI) of: Loxodontina- {Loxndnrila africana), Mammontinie (Parelephas jcffcrsonii, Mammonleus 
lirimi(iniius), Elephaiitinse (KUphas indicus). Eac-h .section (D) is taken as follows: (1) To the left of the median sagittal line; (2) the inferior surface of the 
grinding teeth (M^-M') |ilaced horizontally; (3) the oci-ipital condyles superimposed. It will be ol)scrvcd that: 

(1) I,().\<M)<)NTIN.K. In A«.iY»/r)H((i n/rtca/i« the fronto-occipitocondylar angles and the molar teeth (M--M') are profoundly different from those in the 
Elephantina' and in th<' Mammontime. 

(2) Mam.montix.k. I'areli-phas jeffersonii agrees with .Mammonleus primigenius, male and female-, in the concave frontal profile; /'. jejfer.':onii differs 
from .U, primigenius in the greater anteroposterior extent, it is less compres.sed. .U. primigenius, male and female, is the most strongly compressed, i.e., 
hypsicephalic, acrocephalic. 

(3) Ei-ephantin.k. The adult KU-phas indicus skull differs widely from the loxodontine and inarnrnontine ty|)es in the frontal convexity as well as in tlie 
more vertical position of llic occipital plane. Compare .sections in figure 805. 

'[Subsequently Anu'r. Mus. 1447.5 was provisionally referred by Professor Osborn to Archidiskoitun impernlnr.—'EiWioT.] 
-Amer. Mus. 14.5,59 was made the type of Mammonleus primigenius compressus by Oshiwn in 1924 (1924. ()33, )>. .5). 

922 



CLASSIFICATION OF THE ELEPIIANTOIDEA 



923 



a moderate lateral expansion (brachycephaly) of the E. indicus cranium. As noted above, the bracliycephaly of 
the E. indicus cranium is most marked in the transverse occipital expansion at the very back of the skull, whereas 
in L. africona (P'ig. 1061) the transverse zygomatic expansion of the temporal arcades exceeds the occipital ex- 
pansion. 

Growth of the Jaw (Figs. 799, 797, left). — The elongate infantile jaw of E. indicus (Fig. 799) resembles 
the adult jaw of the Upper Pliocene Archidiskodon planifrons (Fig. 849), also to a less degree the adult jaw of L. 
africann. In brief, the juvenile jaw of E. indicus is harmonic with the mcsocephalic juvenile cranium (Fig. 797, 
left). The growth stages in the jaw are harmonic with the increasing cyrtocephaly, brachycephaly, and bathy- 

Sfcno/v A cephaly of the cranium. Consequently the proportions of the adult jaws of 
E. indicus, as well shown in figure 1204, are abbreviated (cyrtocephalic), 
deepened or depressed (bathycephalic) , and broadened (brachycephalic). 



£.indicui. {/rowth stages 



Horiiifntal rtasjls 




Fig. 807. Superior nasal growth stages. 
Section A horizontal, showing increasing 
breadth of the nasals in the i)assago from the 
infantile to the old male growth stages of 
Elephas indicu.'i. One-sixth natural size. 




Fig. 808. Transfrontal growth stages ' 
Section B. Expansion of the fronto-inter- 
temporal space in infantile to old male growth 
stages in Elephas indicus. One-sixth natural 
size. 




Fig. 809. Occipitohorizontal growth stages, Section E. Complete brachycephalic trans- 
formation of the horizontal contours of the back of the occiput of Elephas indicus in the passage 
from the infantile to the old male growth stages. One-sixth natural size. 



In the jaw evolution of every phylum of the Elephantidse, these three contemporaneous adaptations occur to fit 
the jaw into its very confined space behind the vertically placed (cyptoceplialic) maxillopremaxillary sockets of 
the tusks. 

Downward Flexure of Basis Cranii (Figs. 805, 806). — The ontogenetic progressive downward flexure of 
the basis cranii (cyptocephaly) is illustrated in the series of comparative figures (Figs. 805,806, 1192) which dem- 
onstrate the transition between the relatively horizontal L. africana basis cranii (Fig. 1192) and the most sharply 
deflected basis cranii of Parelephas and of Mammonteics (Figs. 805, 806). 

Downward Flexure of Dental Alveoli (Fig. 806).— As first observed by Leith Adams, and as again 
observed by Gregory and l)y Osborn, the downward growth of the dental alveoli is harmonic with the vertical 
extension (hypsodonty) and fore-and-aft compression (cyrtodonty) of the grinding teeth, especially of M^-Ms. 
This ontogenetic bathycephaly is not shown in the ontogenetic figures but is clearly displayed in the comparative 
figure 806, in which it appears that the second and third molars of E. indicus are much more depressed or bathy- 
cephaUc than the second and third molars of L. africana. 



924 



OSBORN: THE PROBOSCIDEA 



Generic Contrasts in Cranial Sections: Loxodonta, Parelephas, Mammonteus, and Elephas 
The sectional contours described above were jilotted as in figure 805 and almost immediately revealed the 
profound differences between the three chief phyla of elephants. 

Encouraged by this preliminary observation, which gives a new clue to the classification and phylogeny of 
the Elephantidae, a series of sections were taken in all crania available to the author and it was found that the 
other cranial sections supported the inferences derived from the axial section, as shown in figures 805 and 806. 

It is demonstrated that the crania exhibit three subfamily types, namely, loxodontine, elephantine, and 
mammontine. 

(1) Flexure (Cyrtocephaly) of Fronto-occipital-Basilar Planes (Fig. 805). — The midvertical section 
(C) combined with the para-occipitofrontal section (D) is the most important, because, as shown in figures 805, 
806, and 816, it reveals the angular flexure (cyrtocephaly) which the entire upper part of the cranium bears to 
the basis cranii and especially the angle between the occipital 'plane and the basis cranii. It appears that in Loxo- 
donta (tfricana the angle of the basis cranii to the occipital plane (100°) approaches a right angle; this is because 



SCCTION D 

YirTical longitudinal 

rim of onferior naraa To eondylfl 





Fig. 810. Skull. Mid-occipitofrontal Section D, vertical longitudinal, from 
rim (if anterior nare.« to occii)ital condyles, with the ba.sis crnnii in a uniform plane. 
Reduced to one-sixth natural .size. 

I.oxoDONTiNE type {Loxodonta ofricann, male), fronto-occipital .angulate, 
pitcliing forward. 

Elei'ii.w'ti.ne tyi)e {KUphns imliciiK, male, adult), fronto-occipital rounded, 
convex. 

M.\MM0NT1NE type (Parelephas jefersmiii, adult males, and a younfi male'), 
fronto-occipital angle concave or slightly convex. 



SfcriON C 

Virticat layttfal 

natali to vtrftn occifiut 




\ 



•5-'' iV / 

J;' -■' -/ 



I ' — 

i 



Fig. 8n. Frontal section.'!. Comparative vertical section 
of the frontal bones and occiput of the three subfamily divi- 
sions of the elephant family (('), vertical sagittal, nasals to 
vertex of occiput. Reduced to one-sixth natural .size. 

I.oxoDONTi.VE type (Loxodonta africnna, male), convex. 

Elephantine type (Elephas indiciin, adult), concave. 

Elephantine type (Elephas indicus, old male), con- 
cavo-convex. 

Mammontine type (I'arelephas jeffersonii} young m.ale), 
imiformlv concave. 



the cranium of the African elephant as compared with that of Elephas indicus is less brachycephalic, moi-e orthoce- 
phalic, and less bathycephalic or vertically deepened. In E. indicus the occipital plane forms a more open angle 
of 121° with the basis cranii, expressive of its greater bathycephaly. A still wiilei' angle (130°) , however, is attained 
ill the bathycephalic Parelephas jeffersonii, while in Mammonteus primigenius the angle rises to 136°. 



'[Subsequently .\mer. Mus. 1447.') was provisionally referred by Professor Osborn to Archidiskodmi itnpirntor. — Editor.) 



CLASSIFICATION OF THE ELEPHANTOIDEA 



925 



(2) Deepening (Bathycephaly) of the Molar-3 Plane (Fig. 8()6.)~The horizontal grinding surface of 
the superior molars below, coniiiared with the cranial profiles above (Fig. 806), reveals bathycephaly as the most 
distinctive cranio-mechanical adaptation. 

In the LoxoDONTiN^: (a) The bathycephaly is less extreme: (b) the occipitofrontal union is less elevated 
above the horizontal M'-^ grinding surface; (c) the fronto-occipital plane is at an oblique angle with the molar-3 
plane; consequently the Loxodonta skull is less hypsicephalic. 



SecTiON A 

HartjonTol no5aU 




Fig. 812. Nasal contour.s. Comparative horizon- 
tal, Section A, of nasals showing the broad nasals of 
Loxodonta africana and of Elephas indicus as compared 
with the pointed mammontine nasals of Parelephas 
jeffersoiiii, male, of P. jeffersonii,^ young male, and of P. 
jeffersonii, adult male — all in the American M\i.seum 
collections. Reduced to one-sixth natural size. 

Sfcno/v B , , 

■C a/ ricanu.*. fjule 
VerTirof Transverse^ 

frvriTal^^ ^^.-^""^ 

I ^-^~-^^ £2.''Ji::'i.^^°iylL 

/\, _£ indtcus. old malt 

/ Jl Y 

I '1 ^/.^j:Cfjj.?2.\'. /' 

\C. - :^./?//}?j:^^ji:. _ _^\ 

// ... -^ /*//*'"**"?.'/„ 

Fig. 813. Midfrontal or intertemporal forehead. 
Uniformly convex tran.sfrontal, Section B, of Loxo- 
donta africana, male, as compared with the slightly con- 
cave frontal section of an adult Elephas indicus and the 
planoconcave frontal sections of three skulls of the 
mammontine type of Parelephas jeffersonii — all in the 
American Museum collections. Reduced to one-sixth 
natural size. 

LoxoDONTiN.B = Loxorfon(a africann, uniformly 
convex. 

EhEPHANTiNX = Elephas indicus, planoconcave, 
borders angulate. 

Mammontin.e = Poreiep/tas jeffersonii, planocon- 
cave, transfrontal borders rounded. 



In the Mammontine it is observed (Fig. 806) : ( a) That the 
bathycephaly of Mammonteus primigenius and of Parelephas jeffer- 
sonii is practically the same as that of Elephas indicus, because the 
molar-2-3 plane descends to the same degree below the occipital con- 
dyle; (b) while the Mammontinae agree with the Elephantinae in 
bathycephaly, they differ widely in the fronto-occipital profiles, as 
shown in comparison of the line E. indicus with the line 'E.' jeffersonii; 
(c) the Mammontinae agree with each other in the fronto-occipital 
profiles, as shown in the lines 'E.' jeffersonii 'E.' primigenius female 
[ = type of Mammonteus primigenius compressus Osborn, 1924], 'E.' 
primigenius male ; (d) in the Mammontinae the M. primigenius male 
and female profiles seem to be relatively more hyjjsicephalic and 




Fig. 814. Oceipitohorizontal, Section E, through back of occiput at broadest portion in the 
three subfamily divisions of the elephant family. Reduced to one-sixth natural size. 

LoxoDONTiN.« = Loxo(fon/n africana, broad convex flare of occiput; deep ligamentum nucha; 
pit. Indicated by 

Ei.EPH.\iiTi^M = Elephas indicus, very broad, acute flare of occiput; ligamentum nucluv pit 
moderately deep. Indicated by - - - 

MAi,tMOiiTitix = Parelephas jeffersonii, convex flare of occiput, less expanded; ligamentum 
nucha; pit moderately deep. Indicated by . . . 



acrocephalic than those of P. jeffersonii; consequently the Mammonteus cranium suffers even greater fore-and-aft 
compression than the Parelephas cranium. 

The Elephantin/E profile is followed in the dotted line E. indicus: (a) The bathycephaly indicated by the 
M^'^ depres.sion is the same as in Parelephas jeffersonii; (b) the fronto-occipital profile of E. indicus differs widely 
from that of P. jeffersonii and still more widely from that of Mammonteus primigenius ; thus the E. indicus crani- 
um is equally bathycephalic but it is less hypsicephalic and acrocephalic than either that of Parelephas or 
Manunonteus. 

'[See footnote on opposite page. — Editor.] 



92fi OSBORN: THE PROBOSCIDEA 

(3) Hypsicephaly and Acuocki'iialy in Fronto-occipital Phofiles (Fig. 810). — In the Loxodontines, as 
.sliown also in figures 805 and 806: (a) The angular fronto-occipital profile is totally different from that of either 
Parelephas or Elephas ; (b) in Elephas, both in the adult and old male stages of E. indicus, there is a rounded dome- 
like fronto-occipital profile, which rises to a great height in the old male; (c) in the "Parelephas jeffersoyiii" 
crania, shown in the line 'E.' jeffermnii, the cranium is equally hypsicephalic, but it is not acrocephalic, since it 
reaches an acute apex through the characteristically mammontine concavity of the frontals. 

In the frontal sections of the forehead profile we readily distinguish between the uniformly convex profile of 
Loxodonta africana, the concavo-convex profile of Elephas indicus, and the concave profile of Parelephas jeffersonii. 
The concave frontal profile of P. jeffersonii relates this animal to Mammonteus. 

(4) Brachycephalic and Hypsicephalic Nasal Contours (Fig. 812). — The broad nasals of Loxodonta 
africana male are harmonic with the brachycephaly of the cranium. In the Elephas indicus adult the nasals exhibit 
a rounded profile, but in the E. indicus old male the end of the nasals is truncated. In Parelephas jeffersonii the 
crania examined exhibit uniformly acute or pointed nasals clearly distinguishable from those of either Elephas 
or Loxodonta. 

(5) Convex or Concave Transfrontal Sections (Figs. 806, 801, 805, 813). — In Loxodonta the forehead or 
frontal bone is convex in both planes, sagittal and transverse, as shown in figures 806, 805, and 813. In Elephas 
indicus the transfrontal section (Fig. 813) is seen to be plane with sharply angulate lateral borders defining the 
temporal fossae. In Parelephas jeffersonii the transfrontal section is rounded but less convex than that of L. 
africana. 

(6) Relative Brachycephaly of the Occipital Planes (Fig. 814). — Whereas the Loxodonta africana 
cranium, as shown in the plain line 'E.' africanus male, is in general more brachycephalic than that of Elephas 
indicus, the greatest lateral expansion of the occipital region is attained in the E. indicus old male, in which the 
expansion far exceeds that of Parelephas jeffersonii, as indicated in the lines 'E.' jeffersonii. The pit for the liga- 
mentum nuchse is similar in the E. indicus old male and in 'E.' jeffersonii, whereas the pit in Loxodonta africana 
is very much deeper. 

Summary. — In summing up the cranial and profile proportions of Loxodonta, of Parelephas, of Mammonteus, 
and of Elephas in respect to the .six chief regions described above, we observe that Loxodonta differs profovmdly in 
every character from Elephas, Parelephas, and Mammonteus; this is ])robably due to profound differences in 
feeding habits. The crania of Parelephas and of Mammonteus agree in their bathycei^haly with the cranium of 
Elephas, but differ widely in their acrocephaly. There appear to be clear lines of demarcation into three subfamily 
groups which are summarized below (p. 932). 

While the mechanical problem of the tusks and of the proboscis remains relatively the same in these three 
generic stages of cranial evolution, the mechanical problem of the grinding teeth changes, as follows: (1) Tlio 
relatively low-crowned grinding teeth of Loxodonta are correlated with a relatively less brachycephalic or meso- 
cephalic cranium; (2) in the opposite extreme the many plated, high-crowned, extremely hypsodont grinding 
teeth of Mammonteus are correlated with an extremely hypsicephalic and bathycephalic cranium; (3) extreme 
fore-and-aft compression of the cranium of Mammonteus (i.e., cyrtocephaly, bathycephaly) is coordinated with 
extreme shortening, heightening, and multiplication of the dental ridge-plates (i.e., cyrtodonty, hypsodonty, 
polydiskodonty) . 



CLASSIFKJATION OF THE KLRPHANTOrDRA 927 

II. DENTAL AND CRANIAL ADAPTATION TO PREVAILING FEEDING HABITS THE 

KEY TO PHYLOGENETIC CLASSIFICATION 

As among other ungulates, the choicest elephantine diet abundant in favorable seasons of the year is that most 
frequently resorted to and most dominant in cranial and dental mechanics. Seasonal or climatic and secular 
changes may have enforced changes of prevailing diet. Persistence of dental type is favored by persistent climatic 
conditions and flora, as we believe was the case in the true Stegodon, which continued to browse through the Upper 
Tertiary and into early to Upper Pleistocene time in the tropical forests of India and the islands of the East 
Indies, eastward into China and Japan. Accordingly the grinding teeth of Stegodon multiply their ridge-crests 
but do not become sensibly hypsodont as in the elephants. The cranial evolution of the Stegodonts also appears 
to be distinctive. 



1 . RIDGE-PLATE FORMULA OF PRIMITIVE AND PROGRESSIVE GENERA IN ADAPTATION TO 

PREVAILING HABITS OF FEEDING 

The ridge-crests of Stegodon are transformed into the ridge-plates of the Elephantoidea.' 
Ridge-plate Formulae.— Falconer (1868, Vol. II, p. 13) in defining his three subgenera of Elephas laid great 
stress on the number of ridge-plates in the intermediate molars, namely, Dp 4, M 1, M 2, as follows: 

Subgen. Stegodon, hypisomeris, e. g., Dp 4= 7, M 1 = 7, M 2= 8. 
" Loxodon, hypisomeris, e. g., Dp 4= 7, M 1= 7, M 2= 8. 
" Euelephas, anisomeris, e. g., Dp 4 = 12, M 1 = 14, M 2=18. 

Throughout his masterly work "On the Species of Mastodon and Elephant Occurring in the Fossil State in 
Great Britain," 1863, Parts I and II, and "On the American Fossil Elephant of the Regions Bordering the Gulf of 
Mexico," Part III, Falconer makes a profoimd study of the ridge formulae, describing the formulae in various species 
with an accuracy which cannot be challenged today. This great section of Falconer (1868, Vol. II, pp. 1-291) is 
a masterpiece of accurate observation, including very appropriately a discussion of the unity or plurality of species 
(pp. 254-271) based on the vertebral formulae, as cited below. 

2. FOOD OF THE INDIAN AND AFRICAN ELEPHANTS AND OF THE MAMMOTH 

Falconer concludes Section III of his "Palaeontological Memoirs" of 1868, Vol. II, with a philosophical 
treatise (pp. 277-291) on the "Food of Living and Extinct Elephants," which he connects very closely with the 
comparative structure of the ridge-plates in Elephas, Loxodonta, and Mammonteus. 

Food of the Indian Elephant [Falconer, 1868, II, pp. 277-280]. — The 'Sal,' or 'Taiai' Forests, which stretch at the foot 
of the Himalayahs, from lat. 30°, where the Ganges and .Jumna esca])e from the mountains, to the Brahmapootra, embracing 
a range of several hundred miles, are here selected to furnish the chief illustrations which I have to adduce. They everywhere 
aboimd with Elephants, southwards from lat. 30°, which may be regarded as the extreme northern limit of the habitat of the spe- 
cies at the present day. Forests presenting similar phy.sical characters extend along the continuation of the same range, through 
Sylhet, Chittagong, Arracan, Pegu, and the Tenasserim pro\-inces, to the point of the Malay Peninsula; they become more and 
more troi^ical in their vegetation, and, as a general rule, the IClephants improve in size, form, and vigour, according to their 
more southern habitat. ... In fact, the range of his (Indian Elephant] arboreous .selection is restricted within a narrow circle, and 
mainly to the foliage and branches of trees that abound in milky juice which is not acrid, belonging to the families of the 
M areas, Artocarpex, and Sapotacex, such as species of Ficus, Balis, Artocarpus, Bassia, and Mimusops [Footnote: 'Also Mesiia 



'[It will be noted from the following quotation (see Vokinic I, ]>. 25, of the present Memoir) that Professor Osboni subsequently altered his opinion: "It 
lias been assimied by praetieally all palaeontologists that the Elephants were descended from the Stegodonts. This assumption now proves to be erroneous, 
for neither the Stegodon gi-inding 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." — Editor.] 



928 OSBORN: THE PROBOSCIDEA 

ferrea (Nat. Onl. CliiKifwew), on the iiuthority of Tenuont's \at. Hist, of Coylon, p. 230.']. Of tlioso, by far the greater part of 
his stai)lc food is derived from the colossal fig-trees which abound in the forests of India; sucli as Finis Indica, the 'Bur,' or 
Banyan-tree; F. religiom, 'Peepul,' or 'Bodhi-droonia' (Tree of knowledge); F. venom, 'Pilkhun'; F. ranlifolia, 'Gujeena,' or 
'Assoud'; F. glomernta, 'C.oolur'; /''. Tsiela, 'Kuth-bur'; and in Assam, Ficiis eloxtira, or the 'India-rubber tree,' besides other 
more southern species of similar habit and properties. The strong partiality of the Elejihant for the.se trees is so well known 
to the natives, that the 'Obees,' or Pit-falls, for entrapi)ing the animal are invariably constructed in their neighbourhood, and 
many of their old Sanscrit names connect them specially with the Elephant [Footnote: ' "Nagbhundoo," "Koonjurashun"; 
"Gujashun," and "Gujbhukshuk"; all being to the effect of "food of Elephants." {Vide Madden, Journ. Asiat. Soc. Beng., vol. 
xvii. p. 380.)']. He tears down their branches, and crunches the twigs and leaves, stripping off the lactiferous bark of the larger 
boughs. The Elephant of the 'Sal' Forests also derives occasional food from the foliage and fruit of Artocarpus Lackoocha, 
'Dhao'; ... Of aliment derived from the roots of Dicotyledonous trees and shrubs, such as the African EIe])hant is said to 
affect, I know of but one form in the 'Si'd Forests' which the Indian species is known to touch, namelj\ the huge tuberous 
dilatation of the ligneous root of the Scanilent , Piieraria tuberosa, 'Sural.' . . . Among the monocotyledonous families, a very large 
portion of his habitual fare is derived from the Graininex, and more si)aringly from Palms; of the former, he luxuriates on the 
young shoots and tender foliage of various species of Bamboo, which occur in vast abundance, together with the fleshy albuminous 
fruit of Beesha liheedii, found in the southern forests. The 'jhils,' or swamps, to which he resorts, are slieeted with the gigantic 
reeds of Ariindo kiirka, 'Nul,' the young culms of which, together with the stems and leaves of Typha FJIephantina, 'Patela,' at 
certain seasons, constitute a favourite food of the Indian Elephant. The open glades and prairie lands are covered with species 
of Saccharum, forming what is called 'Grass Jungle,' compo.sed chiefly of i5l. sponlaneum, 'Kas,' interspersed with S.fuscum, 'Tat,' 
S. Sara, 'Surkura,' or 'IMoonj,' S. exaltatum, 'Suroo,' &c. Clumps of these grasses are twisted up by his trunk, in his journeys 
to and from the forests; they are beaten against his legs to free the roots from sand, and then subjected to mastication. The 
sand which still adheres to these grasses, together with the large quantity of silica contained in the leaves and culms of .Sac- 
chaniin spontdnciim, the most characteristic species of the grass jungle, performs an important duty in the economy of wear of the 
Elephant's molar teeth [Footnote : 'The excessive abundance of silica in the culms and leaves of S. sponlaneum is practically shown 
when it is attempted to mow it with an English scythe. After a few sweeps, the edge of the implement is rounded off, as I have 
repeatedly witnessed.'] ... It is difficult to conceive of a mechanism better adapted to the duty which they have to perform 
than is presented by the molars of the Indian Elephant. Taking the three true molars, which serve during the adult stage of 
the animal, they are composed successively of 12, 16, and 24 ridges. Each ridge has the core formed of a high wedge-shaped 
plate of ivory; ... A constant equilibrium is maintained, in the normal state, between the nature of the food, the waste of the 
crown-surface, the absorption of the fangs, the forward movement of the body of the tooth, and the replacement of the worn- 
out portion by a succession of fresh plates, jjrotruded from behind. 

Food of the African Elephant [Falconer, 1868, II, pp. 282-284]. — Our knowledge of his food is, therefore, of a vague 
and general character, being deri\ed from the cursory observation of travellers, whose attention was not specially directed to 
the subject. The molar teeth of the African Elephant are intermediate, in construction and triturating characters, between 
those of the Euelephnnles, or Elephants proper, and the fossil Slegodon.s. They present, in the three intermediate and last 
molars for the ridge-formula, the successive ciphers 7: 7, 8, 10; while E. aniiquiis presents the ciphers 10: 10, 12, 16, and E. 
piimigenius and E. Indicus, 12: 12, 16, 24. The aggregate of the series of ridges in the first amounts only to 32; in the second 
to 48; and in the two last to 64 ; involving a great difference in the triturating mechanism of the teeth. In the African form the 
molars are also shorter, narrower, and of less elevation, than in the Asiatic species. The discs of wear, instead of the narrow 
trans\erse bands seen in the latter, exhibit the well-known rhomboidal expansion characteristic of the species. Instead, 
therefore, of being adai)ted to contu.se and triturate the branches and twigs of trees, they are better suited for squeezing and 
crushing leaves, and succulent stems or roots. The habits of the animal, as observed by travellers, are in accordance with these 
indications. Besides browsing on the foliage of the Mimosas and Acacias, which abound in Southern Afiica, they tear up the 
trees of certain si)ecies of these genera by the roots, aided, according to Pringle, by their tusk, used as a crow-bar (?), and they 
devour the succulent i)arts of these roots in the inverted trees [Footnote: 'Cited in the "Library of Entertaining Knowledge." 
Menageries, vol. ii. p. 36.']. Burchell mentions a small sjiecies of Prosopis, P. Elephnntnrhiza, as yielding a favourite food to 
the lOlephant [Footnote: ^Acnciu Ehphuiilinu, Hurch. "Tra\els in South Africa," vol. i. ji. 236. Elephanlorhiza Burchellii, 
Benth.'i; and the succulent 'Spekboom' Puittdacnria Afra, or 'Tree Purslane,' is noticed by most travellers as yielding another. 
That the African Elephant, such as we now see it, formerly extended to the South of Eurojie, has been put beyond question — 
1st, by the researches of Lartet upon remains found in the neighbourhood of Madrid [I'ootnote: 'Comptes llendus. 22 fev. 1858. 
Tom. xlvi.'l ; 2nd, by the remains discovered by Baron Anca in the cave of San Teodoro in Sicily [Footnote: 'Bullet. Soc. Geol. 
de l'"rance. 2e Ser. t. x\ii. p. 684. PI. xi. figs. 5 & 6.']; 3rd, by a molar from Grotta Santa, near Syracu.se, described by the 
( 'anon Alessi (Footnote: '.\tti dell' .\ccad. di Scienz. Xatur. tom. vii. p. 2'23.'], and identified by myself; and lastly, by a molar 
exhumed by M. Charles ( laudin, in lSo8, in a cave near Palermo. ... Of the more ancient lOuropean fossil species, E. aidiquus 
is that which most resembles the African I'llephant in the mesial expansion of the discs of its worn molars. But the character is 
shown in a much less degree, and the great difference in the ridge-formula of the two si)ecies places them in two distinct sub- 
genera. 

Food of the Mammoth [Falconer, 1868, II, p. 284]. — In order to estimate the force and value of the arguments which 
have been raised on this head, it is necessary to institute a rigorous comi)arison between the mechanical conditions of the molar- 
crowns of the Indian Elei)hant and of the fossil s])ecies. The ridge-formula is the same in both, being for the four last t(;eth of 
the upper jaw 12: 12, 16, 24. The number of ridges in the three first of the.se is very constant; the last, as already stated, is 
variable within certain limits, twenty-two being the most common number. Taking the penultinuite, as in the case of the 
Indian Elei)hant, the worn surface of the crown would show sixty-four alternations of unequally hard materials. 



CLASSIFICATION OF THE ELEPHANTOIDEA 929 

Falconer points out that the teeth of the mammoth are far less adapted to browsing and especially to crush- 
ing the woody fiber and bark of trees than the teeth of the Indian elephant, and he ingeniously argues a priori 
that the food of the mammoth was widely different from that of the Indian elephant. This argument, based on 
the structure of the molar crowns, is fully sustained by the discovery that, during certain seasons of the year at 
least, the food of the mammoth consisted chiefly of grasses, as quoted in full from Felix in his description of Eleplias 
primigenius (see Felix, 1912, p. 1127 of this Memoir). 

3. SEASONAL CHANGES IN FOOD OF THE MAMMOTH 

It proves that the extremely hypsodont , finely plated teeth of the mammoth are principally adapted to the 
northern grasses which j^rcvail on the tundras and Arctic prairies during the summer season. As recently described 
by Stefansson, the summer tundra flora closely imitates that of the grasslands which we know as 'prairies' in 
temperate latitudes. In this brief season the mammoth obtained its chief food supply for the year, and, like other 
northern Herbivora, stored up large reservoirs of fat which were drawn upon during the long Arctic winter 
season. In summer a grazer chiefly if not exclusively, in winter it became a browser, feeding upon twigs and the 
branches of conifers and woody substances which contained materials of far less nutritional value. Thus its 
habits directly reversed those of the African elephant (Loxodonta) , which is chiefly a browser and incidentally, and 
for the sake of food variety, a grazer, or of the Indian elephant, which is both a browser and a grazer. Similar 
seasonal habits are doubtless characteristic of the northern types of horses with extremely hypsodont teeth, 
primarily adapted to grazing, secondarily to browsing. 

Conclusion. — The number, compression, and elevation of the ridge-plates increase in all generic phyla, as the 
primitive browsing habit develops into a grazing habit. 

4. SUMMARY OF PROGRESSION FROM BROWSING TO GRAZING DENTITION 

(1) The proboscis, the tusks, and the grinding teeth of the Elephantidse originate in the browsing and plant- 
uprooting habits of Stegodon with elongate tusks.' 

(2) As in other ungulates, the Elephantidse first passed through a purely browsing stage with brachyodont 
grinders, into a browsing stage with hypsodont grinders, typified by primitive species of Archidiskodon like A. 
proplanifrons and .4. planifrons. 

(3) There also arose the browsing and crushing stage with hypsodont and loxodont grinders, typified by Loxo- 
donta africana, in which the teeth were adapted to browsing not only on leafage but on the stems, twigs, branches, 
and roots, in which the plant-uprooting functions of the tusks were useful. There is little or no record of grazing 
in this stage. 

(4) Adaptation to seasonal habits of browsing on the more tender foliage and of grazing on the coarser weeds 
and grasses, as well as on tender shoots and buds, is typified in the still more hypsodont grinders of Elephas 
indicus, which closely parallel in the elongation and multiplication of the ridge laminae species of the genus 
Parelephas. 

(5) Finally there arose the extremely hypsodont, thin-plated, reatively smooth-crowned grinding teeth of 
Mammonteus primigenius, chiefly adapted to grazing, but dining certain seasons of the year forced into use for 
browsing purposes. 

'[But sec footnote on page 927 above. — Editor.] 



930 



OSBORN: THE PROBOSCIDEA 



The primary ridgo forinulip and the primary cranial characters in the several subfamily and generic phyla 
of the Elephantidae were probably similar. We have seen that the juvenile Elephan indicua cranium resembles in 
many respects tlie adult Loxodonta ufricana cranium ; the only primitive true elephant cranium we know is that 
of Archidiskodon planifrons, which it would be interesting to section and compare with Stegodon on the one hand 
and with Elephas indicus on the other. 

As clearly set forth in the introductory pages of this chapter, there is so much parallelism, and even conver- 
gence, in the grinding teeth characters of different phyla of elephants as they pass from the chiefly browsing into 
the browsing and grazing and into the chiefly grazing and browsing types, that it is only by the combined analysis 
of the tusks, of the grinding teeth, of the cranial axes, and of the cranial planes and profiles, that we can distinguish 
these great family and subfamily phyla from each other. 

III. VERTEBRAL DISTINCTIONS OF ELEPHAS, LOXODONTA, MAMMONTEUS, 

AND PARELEPHAS 

Divergence in vertebral and rib formulae in the dorsal, lumbar, and caudal regions will undoubtedly aid us in 
distinguishing genera, species, and subspecies of the elephants. We have not at present accurate and final observa- 
tions on the distinctive characters of the ribs and of the backbone, as shown in the following review. 

1. VERTEBRAL FORMULA IN THE ABOVE GENERA 
Falconer in his invaluable Memoir of 1863, as reprinted in full in the "Palaeontological Memoirs" of 1868, 
Vol. II, pp. 212-291, gives an excellent review of the existing and extinct species of elephants, including his own 
observations (presented in more detail in Chap. XX, pp. 1312, 1313 of the present Memoir) on the vertebral as well 
as on the ridge-plate formulae; lie sununarizes (p. 257) the conclusions of Schlegel, 1845-1867, and of Tomminck, 
1862 (see Sclatcr's translation, 1862.1, of Schlegel's paper on "The Sumatran Elephant") in the following table: 





African Elephant 


Sumatran Elephant 


Indian Elephant 


In 

Temminck 

[1862] 


Schlegel 
[1845-1867] 


In 

Teraminck 

[1862] 


Schlegel 
[1845-1867] 


In 

Temminck 

[1862] 


Schlegel 
[1845-1867] 


( 'ervical \ertebr£e 

Dorsal \crtchriE 


7 
21 

3 

4 
26 

6 
15 
21 


7 
21 

4 
26 

5 
16 
21 


20 
3 
4 

34 
6 

14 

20 


7 
20 

3 

4 
33 

5 

15 
20 


7 
19 

3 

5 
34 

6 
13 
19 


7 
19 

3 

4 
33 

5 
14 
19 


Lumbar vortebriE 

Sacral vertebrae 

( "audal vertebrae 

True ribs' 

l''alse ribs 

Pairs of ribs 



According to tliis Schlegel-Tenuninck table, the continental Elephas indicus has only 19 dorsal vertebriE 
and 19 pairs of ribs, while the insular Elephas sumatranus has 20 dorsal vertebrae and 20 pairs of ribs, the African 
elephant, Loxodonta africana, having 21 dorsal vertebra? and 21 pairs of ribs. These differences, remarks Falconer, 
if they prove to be constant, would be of considerable systematic importance. 



'"Schlegel expressly states, 'that the iiuinlxT of (rue rilis is alike in all the si)C(ri(>s, that is iiiily live;' hut I here is evidejilly a riuiiiiTic^al sli|i in Ihr 
ciphers which he immediately afterwards assigns to the false ribs, namely, lo, 14, and 13 respectively, in the three different species, which would give a 
total of 20, 19, and 18, instead of 21, 20, and 19, being the asserted aggn^gate of pairs corresponding with the assigned tninilxT of dorsal vertebral in the 
different species. (Nat. Hist. Hcview, vol. ii, p. 7o.)" 



CLASSIFICATION OF THE ELEPHANTOIDEA 93I 

Falconer's own observations and records differ from those of Schlegel and Temminck and may be summed up 
as follows: 

Elephas primigenius: In the mammoth there appear (Schlegel) to be 18 dorsal vertebra; and 18 ribs. According to Tilcsius, in the 
niaiiinioth skeleton {E. primigenius) in St. Petersburg there are 19 dorsal vertebrae and 19 ribs. 

Lo.rodonUi africana, African elci)hant: 

Congo (Cuvier, de Blainville, 1844, Laurillard, Daubenton) an imported elephant: 20 dorsal vertebra and 20 pairs of ribs; total 

formula: 7 cervical, 20 dorsal, 3 lumbar, 3 sacral, 31 caudal vertebrae; 20 pairs of ribs. 
Cape of Good Hope: 7 cervical, 20 dorsal, 3 lumbar, 4 sacral, 26-30 caudal vertebra; 20 pairs of ribs [typical capensis]. 
Algoa Bay' (Flower- Falconer) : 7 cervical, 21 dorsal, 3 sacral, 30 caudal vertebra;; 21 pairs of ribs. 
Sumatran elephant (Schlegel) : 20 dorsal vertebrae and 20 pairs of ribs. 

Indian Elephant: 

Ceylon (Camper, de Blainville, Cuvier) : 20 dorsal vertebrae and 20 pairs of ribs. 

Bengal (( 'amper, do Blainville, Cuvier) : 20 dorsal vertebrae and 20 pairs of ribs. 

Bengal (Falconer) : 7 cervical, 20 dorsal, 3 lumbar, 4 sacral vertebra;, 20 pairs of ribs; 20th dorsal small and unsymmetrical. 

Falconer concludes that the continental elephant of northern India varies in the number of its dorsal verte- 
brae from 19 to 20, as the African varies from 20 to 21. 

Vektebkal Formulae According to Falconer's Summary 

[According to Eales (1929)^ 

African Indian Sumatran African Foetus 

Cervical vertebrae 7 7 7 7 

Dorsal vertebra; 20-21 19-20 20 21 

Pairs of ribs 20-21 19-20 20 21 

Lumbar vertebrae 3 3 3 3 

Sacral vertebrae 3 [-4] 3 [-4] 3 [-4] 6 

Caudal vertebrae 26-31 26 [-34] 26] 

Vertebral Formulae According to Flower's and Osboun's Summary of 1926^ 

Loxodonta africana Elephas inrlicns MammontcKS primigenius Parelephas jeffersonii 

Flower, 1885 "Jumbo" Amer. Mus. Flower, 1885 Falconer and Amer. Felix, 1912 Type 

Dept. Mam. 3283 Mus. 14559 { = M. M. primigenius Amer. Mus. 9950 

primigenius com- 
pressus Type) 

Cervicals 7 7 7 7 7 7 

Dorsals 19 20 19-20 18-19 19 19 

Pairs of ribs 18-19 19 19 

Lumbars 4 3 5-3 4-3 5 4 

Sacrals 5 4 4 4-3 4 5 

Caudals 24+ 21 24-30+ 21 21 (Salensky)^ 12+ 



'[Citod from letter, July 2.'), 1929, from Dr. Hubert Collar, Curator, The Mu-scum, SafTron Walden, E.ssex: "I liave verified the particulans you send 
relating to the skeleton of the African Elephant in this Mu.seuni and find them to be correct, except that a few of the caudal vertebra; are missing." 

"The geographic locality is also correct according to our records, but I would iraint out that tlie terra 'imported from Algoa Bay' does not necessarily 
mean that the animal was killed at that place. .'Actually, this s|)ecimen with many otliers was sent from Port Elizabeth in 1833, and so far as I know, there 
is no record showing exactly where it was obtained. It is a fair inference, however, from indirect evidence that the i)lace was somewhere in the locality." 
—Editor.! 

'Eales (1929, pp. 223, 224) observes as to the Congo elephant: "The number of fcetal vertebra; is: Cervical, 7; Thoracic, 21; Lumbar, 3; Sacral, 6; 
Caudal, 26. . . .Six sacral vcrtcbnc bear facets for the ilium. There is, however, no fusion of vertebra; in this region in the foetus. . . . The difference between the 
African and the Indian Elephant lies in the thoracic and sacral regions. The African Eleijliaiit has a larger number of ribs but fewer sacral vertebra' than 
the Indian Elephant (adult). In the foetus, however, there are six vertebral attachments on the ilium, a number greater than any given for the Indian 
Elephant. . . . 

Foetus.— [In addition to the 21 dorsal vertebra;]: sternal ribs, 6; indirectly sternal ribs, 11; free, -1. 

Adult African (Oxford Museum). — ti, 10, 't, giving a total of 21 in t^ach case. 

Adult Indian. — Various figures, e.g. Camper, 8, 12; Perrault, 7, 13; Hluir, 8, 1 1 — the details of indirectly sternal and free ribs not being given." 

^Compare a more recent summary (1929) in Chapter XIX, p. 1227. 

[See Zalensky, Vladimir Vladimirovich, in Bibliography of Volume I of the present Memoir. — Editor.) 



o 



o 



p 



^ 6 

C < 

I^ CO 

I 



X 



:=: c 
r: < 
p &- 



2 o 

o 3 

— : X 



'X 


s 


*"^ 




>^ 




.£* 




+^ 


o 


*-*^ 




^ 

g 


o 


2 








R 




< 


«r 


"o. 








1 




3^ 


.s 




>, 


J 


■§ 


r5 






.^ 


*c 


_o 










_>. 






o 


•"2 


-C 


C 


•^ 


r^ 


E5. 




■1^ 


1 


(1; 


^ 


C 










X 


>> 


^ 


•J 




^a. 


s 






_j; 


,o 


rO"U 


>J 


'*3 




O 




c3 


1 


1 


1 


OS 1 ^ 


'S. 


& 






>. 




5 


>^ ? 



1^ c = I ^ 2 . 

z == 3 -= ~- i £ 
2 £ "= ? ^^ S -c 

X -5 ? -5 g £ ^ 

-/- **^ t^ C Ci 

''- >,_ ,^x! ■-. a 

^ o = - -5 oi 

IN ~ "5 :2 ^ ~ -!^ 
QJ — ^ -r Si y. 



I I I cc 

c ; = oj 

•7 C5 ^ Tt +^ 

iJill 

So . P bi 
»™ C '-0 S ,., 



^ "^ -^^ cJ " 



< 


-£ — . 






s 


X bi 5 


a ^ 


^^ 


•^ ^ 


x 


-1"^. 






H 


■3-« 


U 







►2 to 
u i3 




« o 


?^ 


F< M 


EU 


r'" C- 


O 


r^X 


^ 




k" 




X) 





4) ■" 



(-< rt^ K 






S 

.a M>| 

'=« a.S 

> bC"^ 

■^ c in 
•— ' ^ > 

.0? 3 



CO 



en 






"-5 C a: 
- _- cj 

3 > bC 



it bl — 

5^ M 



^ g- 



2-bt 



-i— ■ c 



> 


> bl) 




S-? 




o »^ 


5 








p 1 


o o 




■*- c 




(-* 






£ 








U r- 


^ 


C 


v: 






t/3 ~ 


-^ 


-^ -r 



Is 



bt 



M 
to 



2 ho 2 

T P*— c 

« CO ., c 

bl"^ Si 2 

fl o o ? 

-^l -^ _L:4 < 

o < — ^ .^ o 

i o o 



fl b <U c 

o ct: e 
o --as: 



>v 









<U -C bti P ■ 
" ~ C ~- . 



3 



<^ is 5 

bt p o|- 

C 



OJ K i "l"^ ^ o 



CO "1^ 

„C0 



i —"3 > "os 

. ^ OT O 



bC 

C 



s = e 



2 5 
5^ 






75 rt « ~ 

£ 2 S ^ 

bC C S. =: 

o >^^::> c 



'ii S 






^ o 



bC 

c- 



^ £ 



2 ^-^ c . 

-e'crSfo ?; 
J ^t^ I 

= S. C bt 2 

f' ST— c.~ 



•^ n '« «^ c 3. 



to 

« 

H 

Q 

»: 
O 



o =-a 

o >- 

— tJ 5 
&:s ^ • 

^.2-S^ 
J2 » g bC 

S oS 

o u S 



o . g 

CO "-^ er* 

&^ 

9 -g'.bf 

■I "» 
o u C ■ 

";=_« .2 

•-0-^01 

^^ ^ bCS 
!-[ a 53 



O O) o ;r; r-^ 

o^ o 5 ' 

-fj t^ rt r* 



C3 O- 



S ft a 



2 ^ 



a> o o o _2 
t^ >i ft i; "S 

« C 

o 



= :2,^ 



e 2 o c ti 
M ■- .2 






?^ « o S 



p 






bO 

a 
'S 

I 

en 



IB bi 

2 »s 

bfl 

cl >> 

r t. 

o-a 
c c 
lij o 



03 -Q 

o ic 
o 



.a i 



o ^ 



O o 
b£) S - 



cc 0^ ^ S: 

.1i bC e 

ja 03 O =: 

2'S*" e • 

e .a ? s, 

.n CO c? -i 

o bC £ = ; 



0) 



I 



?;£:-£ 



o • 

.- CO 

a Sh 

fl bO 
aj o 

bc ^ 

ID " 

4J ^ . 

-- bC O 

J5 C ft 

-S o 5P 
■- ^ fS 

•ago 

S .S" £ 

L. &H '-' 

<o 

M >>?^ 



09 



OS 

Eh 

n 



: O 

i.c3 
. o 



^"i-g^s .a 

"rt Eh ^ *:; ^^ bf 

^ t» bC «^c*!n.— 



i^'S ? >>-^ g 2S 



^fl _;c CO ^ 



a 



; g c 
o 






I 



S 03 



Pa 5 S2^ £ 
.2 « "^s,; £ ^ -5 

-S ,; .a r- 5 S bC d 
ft. 2 5 £ -s o , o 

_w 03 r*, ■ 

a ft aj _^ ■~ -~ 
>> 8 -^ bt-. S -S _ 

o 03 bC C -1"^ - 



.iH a. 05 qil 03 



a> 



03 'C 

o O 



o c . 
05 

o o 

.S CO -^ 

-^ a'S 






C3 t— I 



a a :>, = CO 



p: — a> i3j>-< 



03 ■^ 



( 

03 

a 

03 

o3 ■"•^ 
bC te 

■91. 



02 



3.as 



bC 

o s 



!" ==^ >> . 



g* 03 

"^ CO 

S =3 



1= ^ g) 



£ CO -a o 

■— ' b£i 03 



■ i^ — — 

pi-- 









c 

'in 

S 

o 



g to o > o • 

a > S £ S ■ 
tj 3 j= £ — ; 

-T3 -g o a 
g a "^ ., aT ■ 

CQ-C flol-. 2 

_3 -5^ o -C a . 



o 



'C ;:: ^ 



§£ Sco s 
^ M a^, is 

'§'2:§ ^ 



i :«,>="■£ -5 

J: > S 6f 03 



£^ 

ft ? 



a 0; 
' "3 

o ^ 
b£j- 



IV. SYNOPSIS OF SUBFAMILY CLASSIFICATION OF THE ELEPHANTOIDEA 

As shown in the three leading sections of the present chapter, the cranium affords the chief basis of subfamily 
classification, secondly the tusks, and thirdly the grinding teeth. The three genera Archidiskodon, Parelephas, 
and Mammonteus appear to show profound subfamily resemblances in the cranium and tusks, while they are 
widely divergent in the grinding teeth. 

In tlie succeeding chapters distinctions between these profound subfamily resemblances in the cranium, 
tusks, and grinding teeth are pointed out in detail, in addition to the accompanying synopsis (p. 932). 

V. FINAL SUMMARY OF CHAPTER 

The present Chapter XV demonstrates that profound cranial characteristics separate generic phyla iMch 
hitherto have been united by parallelism or convergence in characters of the grinding teeth. 

The principles of adaptive radiation and of phylogenetic classification as applied to the Elephantidae, and 
developed chiefly through the researches of Falconer, Weithofer, and Osborn, yield a number of very surprising 
results and run counter to all previous systematic and phylogenetic conclusions. The systematic synopsis at the 
close of the present chapter (p. 932) summarizes the conclusions arrived at in this and the following chapters, 
resulting from the most intensive and difficult research incurred in the prei^aration of this entire Memoir. 

The author hopes that the reader will suspend judgment as to this surprising divergence and polyphyly in 
the Elephantidse until the succeeding chapters of this Memoir have been thoroughly examined, namely, C'hajiters 
XVI (Archidiskodon), XVII {Parelephas), XVIII (Mammonteus) included within the subfamily Mammontinffi ; 
Chapter XIX (Loxodontinae), and Chapter XX (Elephantinae). 



933 




e 



"3 



S 
o 
a 

< 
6h 



J3 



*^ T3 



X 








6 




o 


rt 
















rn 


^ 




o 






c: 


13 




'3 


■;=; 




< 






1" 






OS 


Op 




Q 

►J 

a: 
O 














0; 








u 

-a 
c 


o 

s 
s 




a. 


X3 




'is 

S5 


s 

X 




03 


i 

c 






so 
c 




















o 








0) 






o 




o 
o 


Ph 




o3 


-a 






> 

05 




s 

o 


S 




•a 

OJ 


J3 




he 


M 
N 




o 
O 




to 


-a 

03 




C 


5 




K 


>% 




OJ 
13 


o3 




03 


Q 




O 








-0 




to 






s 


-r) 




CO 


C 




s 






2 


- 


}^ 


HH 




3 


0) 




Q 


m 

> 


.05 


s 
i 


o 




S 



o3 




X 
g 

o 


1-1 


1 


1 






3 


-1 




6 


Q 


'o 




s 

? 




C 


g 




< 

X 


J2 




St 






C 


o 




o 


U. 




CO 






0^ 
to 

OJ 

C 


o 






rt 


03 
(1> 


O 

CO 


■§ 










X! 


c: 


to 






•"^ 







« 




CO 


1 




-0 
<u 

03 

D. 


OO 




P 
O 


60 


'-3 


o 
o 
>^ 

CO 


c 

OJ 




C 
o3 




z 
o 


Lh 

G. 




1 




S 
















o 






5 


s 


OP 


OJ 


QJ 




CO 


■r: 








IS 

o 

c 
o 












o 


.2 


> 

o 


a 
3 




■2 


c 

03 

■a 




o 










S 










c 












1^ 






Ph 






■-3 


S! 


rt 


Q 


^ 




o 




c 
o 


-§ 


M 


r/2 




j3 


a 








0) 


H 




<^4-0 




'S 






1 

.a 


'^1 
















o 








o 


o 






•<:* 






-a 


5 


<: 




is 


Oh 


1 


> 






rA 






a 


































c: 




>^ 


U 




u. 


< 





^i 


03 


T3 


O 


u 




< 


a 


"S 


00 


OJ 

c 


c 

03 


^ 


p 




» 


t— ( 






Si 




1^ 


■Jl 




H 


J2 


m 


OJ 


s 






P 


V. 


01 


o 


r: 








o 


a 




oo 


PL, 


n 




,- 




-/J 


(u 




>— ' 


h-1 


E 




o 


a 
o 


55 
< 


03 




8 






< 

a; 
O 


b. 


w 




o 


H 


1 






O 

2 




c 


2 








> 


^ 
K 


1 


to 


rB 




c 

03 
60 

'a 






> 


< 




is 


G 

1 

Q 


t/3 

2; 












•s 






_o 



Chapter XVI 

THE GENUS ARCHIDISKODON (SUPERFAMILY ELEPHANTOIDEA), MOST 
PRIMITIVE MEMBER OF THE SUBFAMILY MAMMONTIN^ 

Archidiskodon, parelephas, and mammonteus, united in the subfamily mammontin^ by similar 

PROFOUND cranial CHARACTERS. WiDELY SEPARATED IN THE RIDGE-PLATE STRUCTURE OF THEIR GRINDING TEETH, 
ALSO IN THEIR HABITS AND GEOGRAPHIC DISTRIBUTION: ARCHIDISKODON IN SOUTHERN PARALLELS; PARELEPHAS 
IN TEMPERATE PARALLELS; MAMMONTEUS IN NORTHERN PARALLELS. SPECIES, PHYLOGENY, AND DISTRIBUTION 
OF ARCHIDISKODON. 



I. Historical Introduction. 

1. History of the subfamily IVIammontinae. 

2. History of the genus Archidiskodon. 

3. Order of discovery and description of twenty-two 

species of Archidiskodonts. 

4. Archidiskodonts of Eurasia and America. 

5. New Archidiskodonts and Loxodonts of Africa. 

6. Approximate ascending phylogenetic order of succes- 

sion of species of Archidiskodon and Parelephas 
(1928). 

Characters of Archidiskodon and Metakchidiskodon 
AND Included Species. 

1. Archidiskodonts of southern Eurasia. 

2. Archidiskodonts and Metarchidiskodonts of South 
Africa. 

3. Archidiskodonts of the United States and Mexico. 



II. 



Review of successive discoveries in Nebriiska, Kansas, 
Texas, CaHfornia, Florida, South CaroHna, Oklahoma, 
Indiana, and Mexico. 

Archidiskodon imperator Leidy, type characters, distinctions 
from Elephas [Parelephas] colmnbi, and comparison with 
A. nieridionalis and A. planifrons. 

Type dental, cranial, and skeletal characters. 

Observations on the characters and geographic distribution 
of ^. imperator in the United States and Mexico. 

Synopsis of A. imperator in the American, U. S. National, 
Nebraska State, City of Mexico, and other museums. 

Cranial characters of Archidiskodon in comparison with all 
the known American crania. 

Skeletal characters of Archidiskodon from a comparison of 
the various skeletal materials known. 



I. HISTORICAL INTRODUCTION 

Subfamily Classification. — In its profound cranial characters, as shown in figures 805, 806, and 816, 
as well as in its cranial profile, we observe the surprising fact that Archidiskodon is far closer to Mammonteus 
and Parelephas than to either Elephas or Loxodonta. Consequently the subfamily Mammontinae, as shown on 
page 932, appears to include three great branches, Archidiskodon, Parelephas, and Mammonteus, which exhibit 
three distinct modes of adaptation of the grinding teeth, independently developed. The author has been very 
slow in reaching this conclusion, for, as shown below, his subfamily Euelephantinse [ = Mammontinae] at first 
included the northern mammoths only. 

1. HISTORY OF THE SUBFAMILY MAMMONTIN.E 

In December, 1917, Osborn (1918.468) presented before the Palseontological Society of America his poly- 
phyletic theory of the Proboscidea, a theory more or less fully anticipated by previous authors but more radical 



935 



936 OSBURN: THEPROBOSCIDEA 

in its inclusion of numerous branches. These branches of the Elephantidae, as extended in successive years 
between 1910 and 1921, are as follows: 

Elephantine Osborn, 1910, p. 558. 'I'lie true elephants of Asia. 
StegodontinvE' Osborn, 1918, pp. 135, 136. Primitive elephants of southern Asia. 
LoxoDONTiNE Osborn, 1918, pp. 135, 136. The African, Eurasiatic, and insular loxodonts. 
EuELEPHANTiNE- Osbom, 1918, p. 136. The mammoths of Eurasia and North America. 

Mammontin* Osborn, 1921, p. 1, established "because the genus Eitelephas is invalid; the term Mammontinx 
(i.e., les mammonts, the mammoths) may be substituted." 

According to this foin-fold subfamily classification of 1918-1921, the synoptic arrangement of these branches 
is as follows: 

[Subfamily: Stegodontine,' includinK the Stegoloijhodonts^ antl true 
ELEPIIANT01D1'1\, elephant-like probu- Stegodonts, from primitive member.s of which the elephant sub- 

scideans. I families may have branched off. 

[Subfamily: Mammontine or mammoths, united by very distinctive 

Elephantide, the elephant-like probe- | cranial characters; separeLtedhy several types of dental characters. 

scideans and their ancestors, in which j Subfamily: LoxoDONTiNyE, typified by the African elephants {Loxodon- 

the ridge-crested grinders gradually I /a a/r/ra/io and the numerous living subspecies), by Pa/a?o?o.rodon 

transform into ridge-plated grind- antiquus* and P. nntnadicus, of southern Asia, and by the dwarfed 

elephants of the Mediterranean islands. 
Subfamily: Elephantine, typified by Elephas indicus {majcimus), 
such as the insular types of Ceylon and the continental types of 
southern Asia. 



ers, the lower incisive tusks abort, 
and the upper incisive tusks attain 
\ery great size. 



We may now recall the first conception by the present author of the diversities which jjrevail in the subfamily 
Manuiiontinse, by quoting from the author's paper (1921.515) on the "Evolution, Phylogeny, and Classification 
of the Proboscidea," pages 14 and 15: 

The MammontinsB. The Mammoths 

It is a striking fact that the oldest geologic appearance of a member of the true Elephantoidea is the Elephas planifrons 
occurring in the Pinjor horizon. Upper Siwaliks, Middle to Upjier Pliocene, India. All the fauna of the great Siwalik deposits 
underlying this geologic level, according to Pilgrim, contain only Stegodonts, Longirostrines, and Brevirostrines. This is signifi- 
cant of a north Eurasiatic center of adaptive radiation of both the Mammontinse and the Elephantine. The chief distinction 
between these two subfamilies lies in the flattened forehead of the Mammoths, to which the specific name planifrons refers, a 
forehead which becomes increasingly concave and compressed anteroposteriorly until it reaches the high, narrow peak of 
E. imperator. 

.\gain, the succession of species is probably polyphyletic, awaiting analysis. In descending order the main geologic succes- 
sion is as follows: 

[Heferencc in Present Memoir 

Elephas primigenius Bluiiienbach. Northern Eurasia and North Americ'a, ri)per Pleistocene = Mammonteiis primigenins 

" columbi Falconer. Middle Pleistocene, North America = Parelephascolinnbi 

" imperator IjC'iAj. Lower Pleistocene, North America = Archidiskodon imperator 

" trogontherii Pohlig. Lower Plei.stocene, Europe = Parelephas trogontherii 

" hysudricus ^'^^ Falconer. Uppermost Pliocene, India = llypselephas hysndricus 

" meridionalis Nesli. Upper I'liocenc, Val d'Arno, Italy =ArchidiskodoN mcridionalis 
" planifrons Falconer. Pinjor horizon. Middle to Upper Pliocene, India, 

also Austria and Bessarabia (Russia) = Archidiskodon planifrons] 

The position of E. hysuilricus I''' in this phylum is doubtful. The cranium referred to this species by Falconer is not of the 
mannnontine type. In 1913 Pilgrim traced ba('k E. planifrons to the Upper Miocene [Middle Pliocene] Si.egodon cautleyi, but it 
would appear at jjresent that none of the known Stegodonts gave rise to the Mammoths. lOxtreme cranial abbreviation, hyi)er- 
brachycephaly, and acroco))haly are great characteristics of all the phyla in this subfamily (excepting possibly that to which E. 



'[Removed by Prof('Ssor O.sborn, Vol. I, p. 22, of tli(^ pn'.scut Memoir to hi.s new .superfamiiy .Stegodoiitoidea. Kditor.] 

'-[Subfamily Mammontinae .sub.stituted (see Csborn, 1921.515, p. 1). — Editor.) 

'[Separated by Professor Osborn from the true Stegodonts under the new subfamily name Stegolophodnntinir and placed in the superfamily Mastodont- 
oidea, family Mastodontida- (see Vol. I, p. 700). — Editor.] 

'[Members of the 'Elephas' anliquus group subsequently regarded by Professor Osborn as belonging to his new genus llesperoloxodon (see Osborn, 1931. S46, 
p. 21).— Editor.) 

^[Included in the subfamily Elephantina; of the pre.sent Memoir. — Editor.) 



THE MAMMONTIN^: ARCHIDISKODON AND MI<:TAK('11IDISK0D0N 



937 



hjj.'^iidriciis belongs). There is a wide range of divergence in the tliickness and MiultijiHcation of the hinielliB of the grinders. 
Klephm iinperator may be derived from the A,', laeridionaliti type, with very few lamellae, composed of thick enamel bands and 
with a great coating of cement, or from the E. pUintfrom Falconer type. The E. primigeniti.s phylum presents the highest lamel- 
lar formula known, with relatively little cement; this phylum is also tlistinguished by the loss of a digit in the pes, becoming 
tetradactyl, a uniciue character among proboscideans. Very great shoulder height, estimated at thirteen feet, is attained by 
K. imperaUir in the favorable environment of the southern United .States and Mexico, as compared with the height of nine feet 
six inches attained by E. primigenius in the frigid north. 

Superfamily: ELEPHANTOIDEA Osborn, 1921 
Family: ELEPHANTID^ Gray, 1821 

Subfamily: M AMMONTI Ny€ Osborn, 1921 

Original reference: Amer. Mus. Novitates, No. 1, pp. 1 and 14 (Osborn, 1921.515). 
Includes Archidiskodontinae and Parelephantinaj of Dietrich, 1927, p. 313 (in part). 

Subfamily Characters. — United by the common cranial characters of Archidiskodon, Par- 
elephas, and Mammonteua, namely, cyrtocephaly, bathycephaly, hypsicei^haly, acrocejihaly, attain- 
ing extreme cyrtocei)haly and acrocephaly in Mammonteus. Incisive tusk alveoli vertical, tusks elon- 
gated, inciu'ved, crossing each other in old males, not serving in feeding habits. Grinding teeth pro- 
gressive from a primitive, subhypsodont stage (Archidiakodon) to an excessively progressive, hyper- 
hypsodont stage {Mammonteus). Habits and geographic distribution varying widely in the three generic 
phyla. 

The cranial unity of the Mammontinsp and the profound distinctions both from the Loxodontn nfricann and 
Elephas indicus types are fully discussed in Chapter XV and clearly set forth in the diagrams (Figs. 805, 806, and 



L-/00 



/^-/36' 




Clnierior'\ 



lo72fitudinal 

Condyle to 

midaie o/ right half of parietals 



LOXODONTINE, ELEPHANTINE, AND MaMMONTINE MiD-CRANIAL AxES 

Compare with figures 805 and 806 of Chapter XV 
Fig. 816. The liorizontal plane of the worn siirfaoe of the .superior 
grinders correspond.s with the hne from tlir niid-condyle to the anterior 
nare.s in Elephas indicus. 

Luxodonta africana, the most primitive, occipital 
pUme (L) 100°. = Loxodontin.e 

Elephas indicus, intermediate, witli rounded cranial 
section, occipital plane (E) 121°. = Elephantin.b 

Mammtmieus primigenius, the most progressive, 
liypsicephalic, acrocephalic, bathycephalic, oc(i|)ital 
jilane (M) 136\ =Mammontin.e 



816); figure 816 represents sagittal sections taken through the center of the right half of the cranium. The 
sections reveal a most surprising resemblance between the three genera of the Mammontins to each other and 
equally surprising contrasts between the mammontine and the Indian and African elephant crania. A summary 
of these characters, quoted from page 932 above, is as follows: 

(1) Subfamily Mammontine, including the three generic phyla Archidiskodon, Parelephas, and Mammonteus, which 
exhibit three distinct modes of adaptation in the grinding teeth. 

(2) Tusks and tusk alveoli progres.sively vertical; tusks elongate, incurved, crossing in old males, not serving in feeding 
habits. 



938 



OSBORN: THE PROBOSCIDEA 



(3) Grinding teeth subhypsodont to extremely hypsodont, typically broad and typically short vertically; progressively 
deepening ridge-plates; typical M 3 ["^ in Archidiskodon planifrons, multiplying to M3 ^J in Mammoiiteus primigenius com- 
pressus. [Parelephas progressus has the following ridge-plate count: M 3 j^. — Editor.] 

(4) Cranium shortened (cyrtocephalic), tleepened (bathycephalic), heightened (hypsicephalic), pointed (acrocephalic), 
reaching an extreme fore-and-aft abbreviation and elevation in Mainmoideus. 

(5) Habits varying in the three generic phyla. Originally browsing, progressive to extreme grazing type. 



A PLANIFRONS Tfe/: 
British Mu^. 



Q A. MERIDI0NALI5 7y/>e 
Florence Mwz 



f S. PINJORENSIS 7y/>e 
^ Amer Mui /9772 



n A, IMPETRATOR -Ref 
Amer Mils. .' 
14476 




1/20 Natural 



Fig. 817. Comparative profiles of riaiiia of Archidiskodon {A, B, D) ami oraiiiuni of Slegmhn (C), all one-twpiitieth natural size. 

A, Archidiskodim plnnifrons ref. (Brit. Mus. M.30G0). C, Hiegodon pinjorensis tyi)t' (Amer. Mas. 19772). 

B, Archidiakodon meridionalis type (Florence Mils.). D, Archidiskodon imperalor ref. (Amer. Mils. 14476). 

Observe in A, B, D the progres.-iive increase in size, abbreviation, heightening, and deepening of the cranium, heightening of the occipital crest, and con- 
cavity of the lengthening forehead. Contrast with the still more abbreviated, hypsice])halfc cranium of .S'/c3(«/o7t pjVy'ocen.sn's (C). 

While united by these profound cranial characters, the three genera, on the contrary, are widely distinguished 
by the specialization of their grinding teeth. It seems hazardous, therefore, to unite them in a single subfamily. 
The outstaiuling generic distinctions are as follows: 



Archidiskodon Pohlig, typified by 
ElephuK meridionalis and E. planifrons. 

Arrhidiskoilon tusks and craniiun as 
in other Maininontinae. Primitive 
grinding teeth, subloxodont, subhypso- 
dont to hyp.sodont. Ridge-plates ex- 
tremely broad plated, laminate, slowly 
progressing in number from M 3 xt+ 
(.4. planifrons) toj^'.h^ {A. imperalor). 
l']nainel thick. 



Mammoths of the south temj)erate 
regions; chiefly browsing. 



Progressive in size to the gigantic 
Archidi.tkodon imperalor maibeni. 



Parelephas Osborn, typified by Ele- 
phas jeffersonii. 

Parelephas tusk and cranial charac- 
ters as in other Mammontina;; cranium 
less cyrtocephalic than in Archidisko- 
don or Mammonteus. Convergent with 
the P>lephantinse in bathycephaly and 
multiplication of ridge-plates; grind- 
ers re.sembling those of Elephas indicus; 
ridge-plates relatively fine, progres- 
sively hypsodont, multiplying from P. 
trogontherii (M 3 H|), to P. jeffer- 
sonii (M 3 §i), to Parelephas progressus 
(M30i). Enamel medium. 

Mammoths of the mid-temperate re- 
gions of Eurasia and North America; 
browsing and grazing habits, adapta- 
tions similar to Elephas indicus. 

Intermediate in size, progressive to 
Parelephas floridanus. 



Mammonteus Camper-Osborn, typi- 
fied by Elephas primigenius. 

Mammonleus crania as in other 
Mamniontina', but excessively acro- 
cephalic, hypsicephalic, bathycephalic, 
and cyrtocephalic. CJrinders with ex- 
treme fore-and-aft compression; ridge- 
])lates ])rogre.ssively deepening and 
broadening, the most elevated and at- 
tenuated known, multiplying from 
M 3 I gin*} {M. astensis), to M3 f}- (M. 
primigenius), to M 3 fy {M. primi- 
genius compressu.'i). Enamel thin. 

Mammoths of circumpolar regions; 
habits chiefly grazing, browsing in un- 
fiivorable seasons. 



Dwarfed in size, with digits 
reduced to four. 



in pes 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 
2. HISTORY OF THE GENUS ARCHIDISKODON 



939 



Although long united with Elephas, the species which are grouped in this genus constitute a very distinct 
generic phylum for which Pohlig's name Archidiskodon, referring to the archaic molar ridge-plates, seems very 
appropriate. As shown in our "Classification of the Elephantoidea" (Chapter XV) 
above, also in figures 805, 806, 816, and 817, Archidiskodon is profoundly related 
in its cranial characters to Mammonteus and to Parelephas; consequently it belongs 
in the subfamily Mammontinae, as defined on page 937 of this Memoir. 



%-nxit. siie 



A. IMPERATOR Ref. 
Ids AnqeZes A^u-S. 



M. PRfMIQENIUS Ref. 
A/ar. Mils Bsao 



P. JEFFERSONII Type 
Amer. Mus. 99S0 



P WASHrNQTONII 

Arner. Mui- BC^ 




'y^o natural sije 



Fig. 818. Hypsiceph.^lic Cbani.\ of the Mammontin« 

A, Archidiskodon imperator, juvenile male cranium, from Rancho La Brea, California. 

B, Mammonteus primigenius, young adult male cranium, from Siberia. 

C, Parelephas jeffersonii, type cranium, aged male, from Indiana. 

D, Parelephas washingtonii, adult male cranium, from tlie state of Washington. 

Observe (1) the strong fore-and-aft cranial compres.sion and vertical elevation of the occiput (hypsicephaly); 
(2) the slightly concave forehead; (3) the very deep depression of the mandibular rami (bathycephaly); (4) the 
sharp downward flexure of the basicranial axis (cyptocephaly); (5) the disimrity in size of the M. primigenius 
skull and jaws. 



Fig. 819. Comparative Series of Superior Molars showing evolution of the hidgb in the 
Elephantoidea and Stegodontoidea. One-sixth natural size. 
E, Mammonteus primigenius compressus, tyi)e M^, showing extreme comjjression of 27 ridge-plates. 
D, Archidiskodon planifrons, type M', with 9-10 ridge-plates. 
C, Stegodon aurorie, type M", with 10 -|- ridge-crests (cf. .S'. airawana, Fig. 764C). 
B, Stegodon ganesa, type M^, with 10 -|- ridge-crests. 
A, Stegodon insignis, type M^, with 10 ridge-crests. 

Stegodon aurora; (cf. p. 892) "is either a liighly progressive Stegodon or a primitive Archidiskodon, a |«)int to be 
determined positively by the discovery of a cranium." 



M. PRIMIQENIUS 
COMPRESSUS 




Fig. 819 



PoHLiG, 1885, 1888. — To our knowledge Pohlig was the first to separate the elephants with archetypal ridge- 
plates as Archidiskodonten (1885, p. 1027): "8. Ich theile die Elephanten nach Kronenformen und Lamellen- 
zahlen der Molaren ein in Archidiskodonten {E. planifrons, E. meridionalis) , Loxodonten {E. africanus, fE. anti- 
quus) und Polydiskodonten {E. primigenius, E. indicus etc.), die Stegodonten mit Clift wieder zu Mastodon 
zahlend." Three years later PoWig (1888, p. 138) defined the genus as follows: "1. Archidiskodonten. Typus: 
E. meridionalis. Uebergang zu der folgenden Gruppe bildet E. planifrons. Tapinodiske, laticoronate, kurze und 




AMERICAN MUSEUM OF NATURAL HISTORY 



ERWIN J RAISZ A H 1827 



Central Region (see black squahb) of the Siwalik Hills, 200 milks .south and north ok Simla India 
.xnn.fr"'' T^\'^ ^v^"'^ Tertiary rauKe. 800 miles in lengtli, on the southwestern base of the Himalaya Mts., arul northeast of the Punjal, Classic ' 
uTsai and detlh "T ""T t"'.''' """ k "'^,"' .''''"''' ''"' ''"^'' "'^''^-^^ '"«'-"^-^"'' ''-^ '''<■"'• ''"'''>■ ''■ C-"".v in .827, explore, l.v Dr Hugh 



Sewalik" 

Falconer 

XXV, inap of India, 



940 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



941 



pachyganale Molaren. Parsilamellat. (Meist nur 15 Lamellen an M. III.)" In the same Memoir (p. 252) he 
introduced the generic terms: 4^ Archidiskodon 5. Polydiskodon 4''- Loxo(-disko-)don. From these citations 
we have abstracted the above generic characters. 

ScHLESiNGER, 1912, 1916. — Schlesinger (1912, pp. 98, 99, footnote) confirmed this generic description and 
amphfied PohUg's definition as follows: "Ausdriicke halte ich fUr einen MissgrifT und gebe ihre Erklarung: Tapino- 
disk = mit niedrigen Jochen (Gegensatz = hypselodisk), latikoronat = breitkronig (Gegensatz = angustikoronat), 
pachyganal = mit dickem Schmelzblech (Gegensatz = endioganal), parsilamellat = mit wenigen Lamellen (Gegen- 
satz = densilamellat)." In a later paper (1916, 

Fig. 821. Chief Lower and Upper Pleistocene localities of western Eurasia in wliieh 
occur species of Archidiskodon, Parelephas, Mammonteus, Loxodonta, and Pal^oloxodon. 
After Osborn, 1910..346, p. 391, fig. 176. See more detailed caption, figure 932, below. 



pp. 102, 103) Schlesinger cites the term as a 
subgenus: "Elephas {Archidiscodon) meridion- 
alis Nesti," and (op. cit., pp. 112, ll3)"Elephas 
(Archidiscodon) planifrons Falc." Schlesinger 
also in his "Studien iiber die Stammesge- 
schichte der Proboscidier" of 1912 recognized 
a resemblance to the broad-plated American 
species Elephas columbi and E. imperator, but 
failed to separate these species into a distinct 
generic phylum. 

Osborn (1922.555) distinguished Elephas 
columbi from E. imperator (p. 3) as follows: 
"We thus find by the characters of the type 
and neotype specimens that the real Elephas 
columbi is not the animal we have been describ- 
ing under this name; it is a dwarf form, per- 
haps a dwarf female, of the animal which we 
have been describing under the name Elephas 
imperator."^ 

He clearly defined (op. cit., pp. 3 and 4) 
the characters of the grinding teeth of E. im- 
perator, concluding (p. 5): "The cranial char- 
acters observed in three more or less complete 
skulls referred to Elephas imperator tend to 
support the direct descent of this animal from 
the E. meridionalis of the Val d'Arno, Upper 
Pliocene of Italy." 




PLEISTOCENE EUROPE. — 1 Forest Bed of Cromer (Norfolk). Sables de 2 
St. Prest near Chartres ( Eure-et-Loire) . 3 Malbaitu (Puy-de-Dome). 4 Peyrotles (Bouchea- 
du-Rhone). 5 Solhilac near Puy. Clay deposits of 6 Durfort (Card). 7 Cajarc (Lot-et-Ga- 
ronne). 8 Val d'Arno (Tuscany). 9 Leffe near Bergamo (Lombardy). 10 Riidorf near Pots- 
dam (Brandenburg). Gravels of 11 Sussenbom near Weimar. Sands of 12 Mosbach in 
northern Baden. Freshwater deposits of 13 Clacton (Essex). Sands of Mauer near 14 Hei' 
delberg (western Germany). 15 Chelles on the Mame, near Paris. 16 iS(. Acheul (Somme). 
17 Ilford and Grays Thurrock (Essex). Lignites of 18 Dilmlen and of Utznach, near Zurich. 
19 Taubach near Weimar. 20 U'iWfcr'rcWi caiic on il/on( Sanfis (eastern Switzerland). Tuffs 
of 21 the Tiber Valley, near Rome. Caves of 22 Neandertal, near Diisseldorf (western Ger- 
many), 23 Spy, near Amur (Belgium), 23a Krapina (Croatia), 24 Chapelle-aux~Sa ints (Cor- 
r^ze). Caves and alluvial deposits of 25 Temifine (or PaXikao) near Gran (Algeria), 26 Pointe 
Pescade, near Algiers (Algeria). 27 Prince's Cave (Monaco). Sandy clays of 28 Vdklinshofen 
(Alsace). 29 Saalfeld (Saxe-Meiningen). Travertines, etc., of 30 Gera, Jena (Saxe- Weimar). 
31 Leipzig (Saxony) . 32 Solutre, north of Lyons. Loess of 33 Wtirzburff (Bavaria). 3i Thiede 
near Braunschweig (Prussia). Cave of 35 Montmaurin (Haute-Garonne). 36 Chdteauneuf- 
sur-Chartnte (Charente). Caves of 37 Schweizersbild near Schaffhausen, and Kesslerlock near 
Thayngen (northern Switzerland). Remains of lake dwellings at 38 W auwyl (hucerne) , 39 Ro' 
benhausen, south of Lake Pfaffikon. 40 Concise on Lake Neuchatel (Switzerland). Peatbogs of 
41 Hassleben, near Weimar. Travertines of 42 Langensalza (Erfurt) in central Germany, 
Caves of the 43 Island of Malta, 44 Island of Crete, 45 Island of Cyprus. 



Subsequently Osborn (1924.633, p. 2) confirmed Pohlig: "We therefore confirm Pohlig's separation of the 
southern mammoths Elephas planifrons, E. meridionalis, and E. imperator into the distinct generic phylum 
Archidiskodon,'" and in the year 1925 (Osborn, 1925.662) he placed Archidiskodon among the Mammontinae, 
as shown in the diagram of that year (Vol. I, Fig. 7, of the present Memoir), concluding (p. 28) with a redefinition 
of the family Elephantidae and Race XIII, as follows: 

'[Subsequently (see pp. 997, 1001 )Professor Osborn separated Elephas imperator and E. columbi, assigning the former to tlie genus Archidiskodon Pohlig 
and the latter to Parelephas Osborn. — Editor.] 



942 



OSBORN: THE PROBOSCIDEA 



Family: ELEPHANTID.'E, distinguished by plated grinding teeth developing out of the more or less closely compressed, 
serrated ridges of Stegodon^ into the broadly plateil grinders of Archidiskodon. the lozenge-shaped 
grinders of Loxodonla, antl the compressed, finely plated grinders of Parelephas, of Mammonteus, 
and of Elephas the type genus of the family. . . . 
Race XIII. The Southern Mammoths, or Archidiskodonts. Excessively broad-plated grinders with 
abundant cement; first known in India, migrating westward into southern Europe, eastward into 
America, where arriving in late Pliocene or early Pleistocene time they finally gave rise to the Im- 
perial Mammoth, Archidiskodon imperator, the last of its race. 




Fig. 822. Geographic distriljution of the principal .speoie.s of Arcliitiiskodonts. The white dots within tlie black areas represent tlie approximate locaHties 
where the tyjies of the.se twenty-two species were discovered; these dots each carry a number in a circle representing the chronologi(; sequence of type de- 
scription. The white crosses represent some of the principal referred specimens mentioned in the present Memoir. 



3. ORDER OF DISCOVERY AND DESCRIPTION OF TWENTY-TWO SPECIES 

OF ARCHIDISKODONTS 

See Figure 822 



1. 1825 E/ep/ifls 7nmrf{ona/7s Nesti, Val d'Arno, Italy 

2. 1846 [18451 Elepfuis plunifrons Falconer and ( 'autley, Siwalik Hills, northern India 
1. 1855 E. giganlcii.s Ayiiiard (MS., also in l"'alconer, 1857, p. 321), France 

[1857-1868 Elephas columbi Falconer, CJeorgia 

3. 1858 Elephas imperator Leidy, ?Seneca, Thomas Co., Nebraska {fide Hay, 1924, 

p. 100)-^ 
[1859-1861 Elephas teaianus Owen, 1859, Blake, 1861, Texas 
1. 1889-1890 Elephas lyrodon Weithofer, Val d'Arno, Italy 
[1890 Elephas mindaneiisis Naumann, Philippine Islands 

4. 1915 Elephas hayi Barbour, Crete, Saline Co., Nebraska 
[1922 El. [Elephas] Columbi var. Felicis Freudenberg, Mexico 

5. 1922 El. [Elepha.s] Columbi var. .vlveslris Freudenberg, Ejutla, State of 

Oaxaca, Mexico 



Reference in Present Memoir 

= Archidiskodon meridionalis 
= A rch idiskodon pla n if ran s 
NouKUi nudum = Archidiskodon 
meridianalisC!) 
= Parelephas columbi (see Chap. XVII)] 

= Archidiskodon imperator 
= Parelephas columbi (see Chap. XVII)] 
= Archidiskodon meridionalis (female?) 
= Siegodon (Archidiskodon'.') inindanensis 

(.see C:hap. XIV)] 
= Archidiskodon hayi 
= Parelephas columbi felicis (see Chap. 

XVII)] 



= Archidiskodon imperator silvestris 
iiiplcte separation of the Stegodontoidea from the Elephantoidea 



'(See Chapter XIV, pp. 800, 807, where Professor Osborn's final views regarding tin 
are given. — Editor.] 

-(Exact locality not cited. Lugn and Schultz (193t.l, p. 373) give Pawnee Loup Branch of Platte Hiver= Middle Loup, probably Hooker County. — 
Editor.l 



THE MAMMONTIN.^: ARCHIDISKODON AND METARCHIDISKODON 



943 



6. 1922 El. [Elephas] Columbi var. Falconeri Freudenberg, Tequixquiac, 

Valley of Mexico 

7. 1922 Loxodonta griqua Haughton, CJriqualand West, Transvaal, S. Africa 
3. 1922 Elephas Columbi var. imperator Freudenberg, Spokam Bar, near 

Helena, Montana 

8. 1923 Elephas meridionalis, mutation cromerensis Deperet and Mayet, 

Kessingland, Suffolk, England 

9. 1924 Elephas antiquus rumanus S. Stefanescu, Tulucesti (Covurlui), Rumania 

10. 1925 Elephas scotti Barbour, near Staplehurst, Seward Co., Nebraska 

11. 1925 Elephas tnaiheni Barbour, near Curtis, Lincoln Co., Nebraska 

[1927 Archidiskodon transvaalensis Dart, Lowest Terrace, Vaal River, S. Africa 

[1927 Archidiskodon sheppardi Dart, Lowest Terrace, Vaal River, S. Africa 



Leith-Adamsia siwalikiensis Matsumoto, Siwalik Hills, India 

Elephas haroldcooki Hay, Frederick, Oklahoma 

Elephas exilis Stock and Furlong, Santa Rosa Island, California 

Archidiskodon subplanifrons Osborn, Sydney-on-Vaal, S. Africa 

Archidiskodon broomi Osborn, near Kimberley, S. Africa 

Archidiskodon sonoriensis Osborn, near Arizpe, northern Sonora, Mexico 

Lox. (Pal.) Tokunagai junior mut. Matsumoto 

Archidiskodon vanalpheni Dart, Sydney-on-Vaal, S. Africa 
Archidiskodon millelti Dart, Sydney-on-Vaal, S. Africa 
Archidiskodon loxodontoides Dart, Sydney-on-Vaal, S. Africa 
Archidiskodon yorki Dart, near Christiana, S. Africa 
Archidiskodon andrewsi Dart, ?Middle Terrace, Vaal River, S. Africa 



[1929 Archidiskodon hanekomi Dart, Vaal River, S. Africa 

21. 1932 Archidiskodon meridionalis nebrascensis Osborn, Angus, Nuckolls Co., 

Nebraska 

22. 1934 Archidiskodon proplanifrons Osborn, Cong-Gong, near the Vaal River, 

S. Africa 



2. 


1927 


12. 


1928 


13. 


1928 


14. 


1928 


15. 


1928 


16. 


1929 




[1929 


17. 


1929 


18. 


1929 


19. 


1929 


20. 


1929 




[1929 



= Archidiskodon imperator falconeri 
= Metarchidiskodon griqua 

= Archidiskodon imperator 

= Archidiskodon meridionalis cromerensis 

= Archidiskodon planifrons rumanus 

= Archidiskodon imperator scotti (or juve- 
nile A. imperator) 

'^ Archidiskodon imperator maibeni 

= Palxoloxodon transvaalensis (see 
Chap. XIX)] 

= Palxoloxodon sheppardi (see 
Chap. XIX)] 

= Archidiskodon planifrons 

= Archidiskodon haroldcooki 

= Archidiskodon exilis 

= Archidiskodon subplanifroiis 

= Archidiskodon broomi 

= Archidiskodon sonoriensis 

= Palxoloxodon {Archidiskodon?) tokuna- 
gai mut. junior (see Chap. XIX)] 

= Archidiskodon vanalpheni 

= A rchidiskodon miUetti 

= Archidiskodon loxodontoides 

= Archidiskodon yorki 

= Palxoloxodon? andrewsi (see 
Chap. XIX)] 

= Palxoloxodon hanekomi (see 
Chap. XIX)] 

= Archidiskodon 7neridionalis 
nebrascensis 

= Archidiskodon proplanifrons 



4. ARCHIDISKODONTS OF EURASIA AND AMERICA 

In the early part of the 19th century there was described by Nesti from northern Italy (1808) the first known 
species of this phylum, to which he later (1825) assigned the name Elephas meridionalis, or, as we may now call 
it, the 'Southern Mammoth,' in reference to its remote relationship to the 'Northern Mammoth.' In 1846 there 
was described from India a direct ancestor of E. meridionalis, namely, Elephas planifrons, the name alluding to its 
flattened forehead. In 1857 Falconer recognized as a remotely related species his Elephas columbi from Georgia, 
a species long confused with Archidiskodon. In 1858 Leidy described as a closely related species Elephas imperator 
from western Nebraska. The Elephas texianus of Owen and Blake, 1859, 1861, 1862, is a synonym of Elephas 
[Parelephas] columbi. The Elephas lyrodon of Weithofer, 1889, 1890, from northern Italy, is the female(?) form 
of E. meridionalis. 

Then a long interval followed and a very primitive member of this phylum, namely, Elephas hayi, was 
discovered in western Nebraska and described (1915) by Barbour, indicating that these animals may have 
migrated into North America directly from India, because E. hayi appears to be almost as primitive in jaw 
structure as the Elephas planifrons of India. 



944 OSBORN: THE PROBOSCIDEA 

In 1922 Freudenberg reviewed the elephants of Mexico and described foiu' subspecies of Elephas columbi 
namely, El. Columbi var. Felicis, El. Columbi var. .s//res^m, El. Columbi var. Falconeri, and El. Columbi var. 
imperator; the first {El. Columbi var. Felicis) we regard as more closely related to Falconer's species Elephas 
columbi, the remaining three to Leidy's species Elephas imperator. 

In 1923 Deperet and Mayet devoted a volume to their thorough researches of the species of eastern Eurasia, 
Elephas planifrons and E. meridionalis, and added the Lower Pleistocene subspecies Elephas meridionalis, muta- 
tion cromerensis, of Kessingland, Suffolk, England. 

The wide geographic distribution of Archidiskodon is indicated by Stefanescu's description in 1924 of Elephas 
antiquus rumanus of Rumania, by the discovery of two American species, Elephas scotti and E. maibeni, 
in Nebraska, described by Barbour in the year 1925, and finally by the most surprising and welcome discovery 
of all, the Archidiskodon subplanifrons and A. broomi in the Vaal River diggings of South Africa, described by 
the present author in 1928, as well as Dart's new species, Archidiskodon transvaalenMs and .4. sheppardi from 
the Vaal River gravel terraces, described in 1927.^ 



5. NEW ARCHIDISKODONTS AND LOXODONTS OF AFRICA 

After the present author had described and figured, but not published,'- the two names above {Archidiskodon 
subplanifrons and A. broomi), he received Prof. Raymond A. Dart's "Mammoths and Man in the Transvaal," 
Nature, December 10, 1927, with the welcome figures [Fig. 823 of the present Memoir] and description of the Vaal 
River terraces at Windsorton and Bloemhof, of which the following is an abstract: 

The Vaal River valley, near Bloemhof, in the southwestern Transvaal, belongs to a great watershed which has yielded 
Australopithecus africanus Dart, 1925 [from Taungs district, Bechuanaland], and the Boskop man, besides stone implements, 
rock engravings, and other evidences of primitive man. The watershed includes three gravel deposits of different geologic age 
(Du Toit, Ann. Rept. (Jeol. C'omm., 1906) situated on a number of terraces. 

Upper. — The highest and presumably the oldest [Pliocene?] terrace has an altitude of some 200 to 300 feet above the river 
exposed at a distance of 3}i to 6 miles. No fossils have hitherto been described from this most ancient terrace (Dart, op. cit., p. 1). 
[This is possibly the level at which Archidiskodon subplanifrons Osborn, a very primitive form, was found — Fig. 875 of the pres- 
ent Memoir. — Osborn.] 

Middle. — The middle or second [Pleistocene or Mastodon] gravel terrace is on a lower level at several points west of Barkly 
West; from this terrace numerous palaeolithic implements have been described (Hodkinson, 1926); also (Beck, 1906) a frag- 
mentary tooth of Mastodon {Bunolophodon) sp. Felix, at Waldeck's Plant, 60 to 80 feet above the river bed. Haughton described 
Loxodonta griqua [ = Metarchidiskodon — see figure 882 of the present Memoir], also a new giraffe {Griquatherium cinyulntiim) 
from this 60 to 80 foot terrace. [Doctor Broom holds that the mammoth teeth are washed in from an older geologic deposit 
(see Nature, March 3, 1928, p. 324) and are not truly associated with the flint implements, in geologic age.] 

LowEH. — A still lower gravel deposit [Pleistocene], near the level of the present river Vaal, contains Equus and Hippo- 
potamus amphibius var. robuslus. From this low river bed gravel, \Yi miles below The Bend on the Vaal River, Haughton 
determined as Elephas (cf. antiquus) a portion of a tooth recovered at a de])th of 5 feet, but which Dart regarded {op. cit., p. 42) 
as resembling E. antiquus recki Dietrich. [Apparently this same specimen is now described by (Jsborn as Archidiskodon broomi. 
the label of which bears the inscription, "3682 Mus. Kimb. The Bend. II. Else." Fig. 877 of the present Memoir.] From this 
level are recently recorded Archidiskodon Iranspaalensi.s Dart and A. sheppardi Dart.l'l 



^[Archidiskodon transvaalensis and .1. s/icppanij prove to belong to the genus Palipoloxodon (.see Osborn, 1934.92."), pp. 2iinil 11) Since this section of the 
Memoir was written, four species from South Africa have been (iescribeii by Professor Dart (1929) and one by Professor Osborn (1931); jilso two 
.species from North America have been described by Professor Osborn, one from Mexico (1929) and one from Nebraska (1932). A <'omplete list of the 
species of Archidiskodon is given on pages 942 and 943 above. — Editor.] 

^Subsequently published in Nature (Osborn, 1928.749, pp. 672, (173) under the title "Mammoths and Man in the Transvaal." 



J 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



945 



Gr^vcu te:rf*.acc Go - so feet 

/ rlvcr bcd c.ravel terrace 





DiA li-^SE 



Transvaal 




The Vaal River Gravel Terraces, South 
Africa, the Scene of the Discovery of 
Mammoths and Man in the Transvaal 

After Dart, 1927 
Fig. 823. These terraces (upper at Windsorton, 
middle and lower at Blocmliof) have yielded the 
types of Boskop man (Homo capensis), of Aus- 
tralopithecus africanus, of Loxodonta [ = Metarcki- 
diskodon] griqua Haughton, of Archidiskodon 
[ = Pal^oloxodon] transvaalensis and .4. [=P.\ 
sheppardi of Dart, and of A. subplanifrons and 
A. broomi of Osborn, also of Mastodon {Buno- 
hphodon) sp. Felix. After Dart, 1927, figs. 1 to 3. 



Old Riven bed dicoincs 



Sand House oh island 




New BlutR BED 

', Gravel bed or 

, UNKNOWN DEPTH 







Consequently a summary of the discoveries and descriptions of South African specimens (1928) is as follows: 



Original Description 
Loxodonta griqua Haughton, 1922, type, Griqualand West, Transvaal 
Elephas (Lo.rorfo«) zulu Scott, 1907, type, Zululand 

Elephas zulu, referred by Hopwood, 1926, Kaiso Bone-beds, near Lake Albert 
Elephas antiquus Necki Dietrich, 1916, type, Tanganyika Territory 
Elephas aff. meridionalts Nesti, 1825, referred by Hopwood, 1926, from Kaiso 

Bone-beds, near Lake Albert 
Archidiskodon transvaalensis Dart, 1927, type, middle Vaal River gravel terrace 
Archidiskodon sheppardi Dart, 1927, type, lowest Vaal River gravel terrace 
Archidiskodon subplnnifrons Osborn, 1928, type, middle(?) Vaal River gravel 

terrace 
Archidiskodon broomi Osborn, 1928, type, middlc(?) Vaal River gravel terrace 
Mastodon (Bunolophodon) sp. Felix, referred by Beck, 1906, Waldeck's Plant 



Reference in Present Memoir 
= M elarchidiskodon griqua 

- Loxodonta zulu 

- Loxodonta zulu 

= Palseoloxodon recki 

- M elarchidiskodon griqua [or A. planifronsf) 

- Palxoloxodon transvaalensis 

- Palseoloxodon sheppardi 

-- Archidiskodon subplanifrons 
-Archidiskodon broomi 
-- Trilophodon{'!) sp. indet. 



The above species of Archidiskodon from Africa are fully described below, following the description of the 
Eurasiatic Archidiskodonts. 



946 



OSBORN: THE PROBOSCIDEA 



6. APPROXIMATE ASCENDING PHYLOGENETIC ORDER OF SUCCESSION OF 
SPECIES OF ARCHIDISKODON AND PARELEPHAS (1928) 

Only an approximate phylogenetic order of succession of the species of Archidiskodon may be given at the 
present time. In this phylogenetic list we may inchide with some certainty fifteen of the better-known species 
of Archidiskodon and Parelephas, placed in ascending order as follows: 



ARCHIDISKODON AND PARELEPHAS OF NORTH AMERICA AND EURASIA 



listocene? level 


1857- 
1859- 


•1868 
1861 


Southern United States and 

Mexico 
Texas 




1922 




Mexico 




1922 




Mexico 




1922 




Mexico 


iver Pleistocene 


1925 




Nebraska 



?Af Ionian (fide Osborn) 
lowan or Late Pleisto- 
cene (fide Lugn and 
Schultz, 1934)' 



Lower Pleistocene 


1925 


Nebraska 


Pleistocene 


1928 


California 


Pleistocene 


1858 


Nebraska 


Lower Pleistocene or 
Upper Pliocene (?)= 
Lower Pleistocene 


1915 
1923 


Nebraska 

England, Forest Bed 



Upper Pliocene 



1825 Italy 

1846 [1845] India and southern Europe 



Parelephas columbi Falconer, originally described by Falconer as 
Elephas columbi, 1857 

Parelephas columbi Falconer, originally described as Elephas 
lexianus by (Jwen, 1859, by Blake, 1861 

Parelephas columbi felicis Freudenberg, originally described by 
Freudenberg as Elephas Columbi var. Felicis, 1922 

Archidiskodon imperalor silvestris Freudenberg, originally describ- 
ed by Freudenberg as Elephas Columbi var. silvestris, 1922 

Archidiskodon imperalor falconeri Freudenberg, originally de- 
scribed by Freudenberg as Elephas Columbi var. Falconeri, 
1922 

Archidiskodon imperalor maibeni Barbour, originally described 
by Barbour (1925) as Elephas maibeni, subsequently (1926) 
as Archidiskodon maibeni 



Archidiskodon imperalor scotli Barbour, originally described by 

Barbour as Elephas scotli, 1925 
Archidiskodon erilis Stock and Furlong, originally described by 

Stock and Furlong as Elephas exilis, 1928 
Archidiskodon imperalor Leidy, originally described by Leidy as 

Elephas imperalor, 1858 
Archidiskodon hayi Barbour, originally described by Barbour as 

Elephas hayi, 1915 
Archidiskodon meridionalis rromerensis Deperet and Mayet, 

originally described by Deperet and Mayet as Elephas 

meridionalis, mvitation cromerensis, 1923 
Archidiskodon meridionalis Nesti, originally described by Nesti 

as Elephas meridionalis, 1825 
Archidiskodon planifrons Falconer and Cautley, originally de- 
scribed by Falconer and Cautley as Elephas planifrons, 

1845, 1846 



ARCHIDISKODONTS OF AFRICA 



Pleistocene 1927 

1927 

Lower Pleistocene(?)=' 1928 

Pleistocene 1922 

Upper (?) Pliocene^ 1928 



South .\frica 
South Africa 
South Africa 
South Africa 

South Africa 



Archidiskodon [ = Pal^olo.rodon] transvaalensis 

Archidiskodon [ = Palxoloxodon] sheppardi 

Archidiskodon broomi Osborn 

Metarchidiskodon griqua Haughton, originally described by 

Haughton as Loxodonta griqua, 1922 
Archidiskodon subplanifrons Osborn 



According to the above phylogenetic arrangement, the four species Archidiskodon subplanifrons\ A plani- 
frons, A meridionalis, and A. hayi- are the only known [1928] Upper Pliocene representatives of this great phylum. 



'[Srr page 1028 below for the present author's final (lelerniination. — Editor.l 

•'[Lugn and Schultz (1934.1, table opp. p. 3.")8) regard Archvlinkndnn hayi as of Kansan ((iraiid Island) age, etpiivalent to the Lower Pleistocene. — 
Ekiitor.l 

'[Dr. C. van Riet Lowe places Archidiskodon subplanifrons in the early Pleistocene and .1. broomi in the late Pleistocene (see van Riet Lowe, 1929.1, 
table opp. p. 682).— Editor.) 



THE IVIAMMONTIN.^: ARCHIDISKODON AND METARGHIDISKODON 



947 




Family Group of Archidiskgdon imperator maibeni along the Platte River, Nebraska 
Restoration by Margret Flinsch Buba in 1935, under the direction of Henry Fairfield Osborn 
Bull in foreground l/40th natural size and bull at extreme left l/140th natural size. Cow and calf in middle foreground l/60th natural size 
Fig. 824. The bull in the foreground and the one in the distance at the extreme left were based on mea.surements of the skeleton of Archidiskodon 
imperator tnaibeni in the Nebraska State Museum (Neb. Mus. 5-9-22), from Lincoln County, Nebraska (.see Figs. 910-912, 918 of the present Memoir), 
supplemented by a fine .skull of A. imperator in the Geological Institute of Mexico (No. 212) from Tepexpan (see Fig. 902) and of a giant tusk of the same species 
(.1. imperator) in the American Museum (Amcr. Mus. 22481), from Post, Texas (see Fig. 894). The height at the withers (toj) of shoulder blade) is 3826 
mm.; adding the usual 6/3'yc gives a height in the flesh of 4068 mm. or 13 ft. 4% in. 



II. CHARACTERS OF ARCHIDISKODON AND METARCHIDISKODON AND INCLUDED SPECIES 

Superfamily: ELEPHANTOIDEA Osborn, 1921 
Family: ELEPHANTID^ Gray, 1821 
Subfamily: Mammontin^ Osborn, 1921 

Genus: ARCHIDISKODON Pohlig, 1885, 1888 

Original reference: Zeitschr. deutsch. geol. Ges., XXXVII, Heft IV, p. 1027 (Pohlig, 1885.1); Xova Acta Leop. Carol., LIII, No. 1, 
pp. 138, 252 (Pohlig, 1888). 

GenotjTDic species: Elephas meridionalis and E. ■planifrons. 

Syn.: Parastegodon Matsumoto, 1924 (in part); genotype, Elephas aurorx Matsumoto, 1924. Leith-Adamsia Matsumoto, 1927; 
genotype, Leith-Adamsia Siwalikiensis Matsumoto, 1927. 

Generic Characters (translation after Pohlig, 1888, p. 138). — Genotypes E. meridionalis, 
E. planifrons; transitional to higher groups. Molars with few ridge-plates, not exceeding fifteen [?] 
lamellae; crowns broad, short, and with thick enamel (i.e., "laticoronate, kurze und pachyganale 
Molaren") ; ridge-plates tapering to summit in sections. 



948 



OSBORN: THE PROBOSCIDEA 



Generic Characters (Osborn, 1928). — Superior tusks large, incurved, crossing in old males 
as in other Mammontinse. Cranium extremely heightened (hypsicephalic), foreshortened and broadened 
(brachycephalic), deepened (bathycephalic). Molar ridge-plates extremely broad, enamel border 
thickened, cement usually very thick. Ridge-plate formula slowly progressing from minimum, M 3 ^ 
(^4. plamfrons), to maximum, M 3 —— {A. imperator). Finally attaining gigantic size. 

This genus is given first consideration in the history of the subfamily Mammon tina^ for the reason that 
Elephas [ = Archidiskodon] planifrons of the Upper Pliocene of India is geologically the earliest of the Elephantidse 
at present known. It is a very striking circumstance that until we reach the Upper Pliocene, Pinjor horizon, of 
India we discover Mastodonts and Stegodonts only, there being no trace of the Elephantidae; consequently 
Archidiskodon planifrons appears to be a new arrival in southern Eurasia in Upper Pliocene time, probably a de- 
scendant of ancestors coming from some region of southern Africa, such as Archidiskodon subplanifrons.^ 

Osborn, 1929: Comparing Falconer's measurements of thirty-nine Siwalik specimens (Table VI, opposite page) 
with Barnum Brown's measurements of twenty-seven Siwalik specimens (p. 954 below), we find a closely similar 



range of 'ascending mutations' rising from the most primitive ridge formula, M 3 



»H 



through intermediate 



mutations, M 3 f j'^j, to the most progressive mutations, M 3 j 



K-l 2-W 



Thus in sixty-six specimens collected in 



the Pinjor horizon, of the Siwaliks, India, we observe gradual ascending mutations of the Archidiskodon planifrons 
stage toward the A. meridionalis stage. The ridge-plates throughout in M^ are more numerous than those in M'*. 

Summary. — The ascending mutations in the thirty-nine sjiecimens referred by Falconer (1868) to 'Elephas 
planifrons' are seen to compare very closely with those in the twenty-seven specimens of the collection made by 
Barnum Brown and referred by Osborn to Archidiskodon planifrons. This renders it probable that the Falconer 
collection and the Brown collection were from the same geologic horizon. The extremes in both collections are 
seen in the maximum and minimum heights of the ridge-plates : 



Uppkr Pliocene Archidiskodon planifrons 

Measurements of Ridge-plates in Millimeters 

Falconer (1868) Brown (1922) 

Max. Min. Max. Min. 



Third superior molars, M' 

Height 

I>cngth 
Third inferior molars, M3 

Height 

I.cngtli 



123 
279 

115 
323 



64 
247 

77 
224 



102 
254 

135 
313 



88e 
201 

53 
203 



Allowing for individual differences in size due to sex and geologic level, and for probable differences in mode 
of measurement, especially of the height of the ridge-plates, it appears probable that in both collections we have 
to do with a series of ascending mutations some of which attain the size and ridge-plate height of Archidiskodon meri- 
dionalis. It is noteworthy that in the A. planifrons r.M-, the length (221 mm.) and the height of the 7th ridge- 
plate (71e mm.) agree closely with the 7th ridge-plate height of M- of ^4. planifrons from the Piltdown gravels of 
Sussex, England (see Figs. 852, 853, 854). 

Falconer observes that while the ridge-plates in A. meridionalis increase in nimiber they do not increase 
correspondingly in height. In two of his figiu'cs reproduced in the j^resent Memoir (Fig. 862) the measurements are: 



Val d'Arno, r.Ms 
Norwich C'rag, r.M.i 



i'liiMirivK .\rciiii)Iskodon meridionalis [or .v. I'LAMFRON.s] 

Length lircadtli 

10 in. [ = 255 mm.] 3.4 in. [ = 87 mm. 
11.25 [ = 287 mm.] 3.8 [ = 97 mm. 



Ifoiglil 
5 in. I = 126 mm. I 
4.8 [ = 123 mm. I 



*[0r .irchidiskodon proplanifrotis, an o<iuall\- primitive species sub.sequently described by Professor Osborn from Gong-Gong, Vaal River (Osborn, 1934.925. 
p. 10).— Editor.] 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



949 



UPPER SIWALIKS MUTATIONS 

TYPICAL ARCHIDISKODON PLANIFRONS: M 3 fJt, ASCENDING TO TYPICAL A. MERIDIONALIS: M 3 

Table VI. Summary of Measurements of Thirty-nine Specimens referred by Falconer, 
1868, TO Elephas [ = Archidiskodon] planifrons 



1 3 - 1 4 
11-14 













Length 


Breadth 




Height 




Superior 


Plate 


Figure 


Molar 


in. mm. 


in. mm. 


Index 


in. mm. 


Ridge-plates 


Third Molars 


XII 


Footnote, p. 


















433. 


Unfig. 


M^ 


11 =279 


3.7-4.8 = 95-123 


34 


8th 4 . 8 = 123 


K-12-}^ 




XI 


1 




l.M' 


11 =279 


4.5=114 












4 




'r.M^ 


5.4=135 


2.5= 63 






8+ 




XIV 


8 




=r.M' 
r.M3 


10 =255 
9.7 = 247 


3.5= 89 




10th? 2.5= 63 
4 =102 


8-9-K 




X 






r.M' 


9.7 = 247 


3.5= 89 




4 =102 




Second Molars 


II 


5a 


Type 


r.M'- 


8.7 = 221 






[7th = 71e 


9 




XII 


5 




r.M= 


8 =204 


3 = 77 


37 




K-8 






5a 




1.M2 


7.5=191 


3.2= 81 


42 




y2-s-% 






6, 


6a 


r.M2 










8 


First Molars 


VI 


Footnote, p. 


















427 




Ml 


6.5 = 165 


3.4= 87 


52 


4th 3.5= 89 


K-7-K 






6 




M' 


5.5=140 


2.7= 69 


49 


4th 3 = 77 


}i-7-]i 




XII 


4, 


4a 


.Ml 


5.2=132 


2.8= 71 


55 




7? 


Fourth Deciduous 


VI 


5 




Dp< 


4 =102 


2.4= 61 


60 




'A-&-}i 


Third Deciduous 


XII 


2 




Dp' 


3.8= 97 


2.3= 59 


60 




'A-Q-'A 


Second Deciduous 




1, 


la 


Dp^- 










4 


True Premolars 


VI 


4 




l.P 


1.2= 30 


1.1= 28 


96 


.8= 21 


3 with indistinct 

front and back 

ridges 


Inferior 




















Third Molars 


XII 


Footnote, p. 


















433. 


Unfig. 


r.Ms 


11.8 = 300 


3.9 = 100 


33 


nth 4.6=115 


14-K 










I.M3 


11 =279 


3.3= 84 


30 


8th 3 = 77 


13-}^ 






13 




I.M3 


12.7 = 323 


3.6= 91 


28 


10th 4.5=114 


13-H+ 




XI 


8 




M2 [M3I 


12.1 = 308 


3.8= 97 


32 




n-% 






2 




r.Ma 


11.8 = 300 


3 = 77 


25 




}M3-K 






5 




r.M3 


10.5 = 267 


4.2=105 


39 


9th 3.5= 89 


9+ 




II 


56 Paratype 


I.M3 










9+ 




XI 


7 




I.M3 


10.2 = 259 


4 =102 


44 








XVIII.A 


1, 


la 


I.M3 


10.4 = 265 


4.1 = 104 




3.2= 81 


8 




XI 


3 




I.M3 


10 =255 


3.6= 91 


36 




7+ 




XII 


12, 


12a 


?r.M3 


10 =255 


3.5= 89 




7th 4 =102 


A-io-'A 




VIII 


2, 


2a 


r.Ms 


8.8 = 224 


3.8= 97 


43 




9-11 




XIV 


9 




r.Ms 


9.5 = 241 


3.5= 89 




6th 3 = 77 


&A 




XI 


10 




I.M3 




3.7= 95 




5th 3.8= 97 




Second Molars 




6 




I.M2 


8 =204 


2.8= 71 






9-K 






10 




l.M, 


7 =178 


3.7= 95 






8-M 


First Molars 




6 




l.Mi 


6 =153 


2.8= 71 






6+ 




XII 


10, 


10a 


l.M, 


6.7 = 171 


2.6= 66 






yr-7 


Fourth Deciduous 




8, 


8a 


l.Dp, 


4.4 = 113 


2.4= 61 






X-7-K 


Third Deciduous 




7, 


7a 


r.Dp3 


2.4= 61 


1.4= .36 






6-K 


True Premolars 




10, 


10a 


LP (frag 


) 1 = 26 


1 = 26 




1.5= 38 


Does not 






8, 


8a 


LP 


1 = 26 


.8= 21 






show ridges 



'[One of the genotypes of Leith-Adamsia Matsumoto. — Editor.) 

-Valuable because found with r.M.i (PI. xiv, fig. 9). (PI. xiv, fig. 8, represents one of tlie genotypes of Leilh-Adamsia Matsumoto. — Editor.] 



950 



OSBORN: THE PROBOSCIDEA 



1-. ARCHIDISKODONTS OF SOUTHERN EURASIA 
From our present knowledge it appears probable that Archidiskodon originated in Africa, because seven stages 
in the specific evolution of this genus have been discovered on that continent. In India Archidiskodon planifrons 
suddenly appears in the Upper Pliocene, Pinjor horizon,' while /I. meridionalis appears in the uppermost Pliocene 
and Lower Pleistocene of western Europe. The subspecific name Elephas [ = Archidiskodon] meridimialis cromeren- 
sis has been applied by Deperet and Mayet to rather primitive archidiskodont molars found in the Forest Bed of 
East Anglia. 



SYSTEMATIC DESCRIPTION OF SPECIES 



Archidiskodon planifrons Falconer and Cautley, 1846 [1845] 

Figures 81.5, 817, 819, 825, 827-850, 8.52-856, 865, 871, 876, 914, 
1208, 1231, 1239, PI. xxi 
'Sewalik Hills' of Falconer, northern India; Pinjor horizon, Upper 
Pliocene, Upper Siwaliks (Pilgrim)'; Upper Siwaliks, below Boulder Conglom- 
erate, Lower Pleistocene (Brown). 

Syn. : Leilh-Adamsia Siwalikiensis Matsumoto, 1927. 

Specific Characters (Falconer, 1846, 1863, 1868; Osborn, 
1928). — Species distinguished from Elephas [ = Stegodou] auroras 
by the more elevated or hypsodont ridge-plates ; distinguished from 
Elephas [ = Archidiskodon] meridionalis by the much less elevated 
ridge-plates; by the flattened forehead to which the specific name 
planifrons refers in distinction from the concave forehead of A. 
meridionalis; also by the more primitive ridge-plate formula, 
M 3 JY+, and by the succession of premolars, this being the last 
species of the Elephantidse in which the true premolars, P 3 and 
P 4. are erupted. Rostrum of jaw elongated and depressed (see 
Fig. 849). 

Materials. — The above comparison of measurements of 
thirty-nine specimens in the Falconer and Cautley collection 
(1830-1840) and of twenty-seven or more specimensinthe American 
Museum collection of 1922 made by Barnum Brown (see Summary 
of Mca.surements, Tables Vl and VII, pp. 949 and 954) renders it 
probable that both collections, totaling sixty-six or more speci- 
mens, came from the same general locality and represent ascending 
mutations, from the more jirimitivc typical Archidiskodon plani- 
frons, M 3 YiX' toward the more jirogressivc ^4. meridionnlis 
stage, M 3j^iEj^. We ha\c to do, therefore, with nscrndinq mu- 
tations. 

Geologic Locality and Levels. — All the chief specimens 
(numbering 39 in Table VI above) listed, described, or figured by 
Falconer and Cautley as 'Elephas planifrons' are recorded as 
simi)ly from the Siwalik Hills, since they did not distinguish be- 
tween 'upper,' 'middle,' and 'lower' Siwaliks. 

Pilgrim (1913) attributes this species only to the 'Upper 
Siwaliks,' Pinjor horizon, summit of the Pliocene; he writes (1913, 
p. 294): "There is absolutely no trace of Elephas either in the 
IMiddle Siwalik or in the Tatrot zone of the Upper Siwalik. It 
first ajipears as the species Elephas planifrons some 2,000 feet 
above the base of the Tatrot zone." 



In the collection of 1922 made by Harnuni Brown, twenty- 
seven or more specimens are now recorded as, or referred to, 
Archidiskodon planifrons; this collection (see map. Fig. 826) is 
recorded by Brown e.specially from near the towns of Kalka, 
Charnian, Siswan, Chandigarh, and Mirzapur, as shown in the 
complete li.st on page 955 below. 

None of the twenty-seven or more s])ccimens of the Brown 
collection was actually found embedded in the rock; all of them 
were loosely embedded in sand, occurring in gullies or depressions. 
Consequently the exact geologic age and horizon levels are inde- 
terminable; yet they may all be safely classed as L^pper Pliocene.' 

History. — Falconer and ( 'autley first published in the "Fauna 
Antiqua Sivalensis" the name Elephas planifrons in the legend of 
Plate II, figs. 5a, 56, of date 1845. In 1846, p. 38, Falconer wrote: 
"The next serial modification in the disposition of the three dental 
substances, and in the consequent form of the teeth, is exhibited in 
fig. 5a of the same plate, which represents a section of the penulti- 
mate upjier molar of another Indian fossil sjjecies which we have 
named E. planifrons." As usual in his descriptions the locality is 
given simply Siwalik Hills, but we know from Pilgrim (1913, j). 
294), as (luoted above, that this phylum first api)ears as the .species 
E. planifrons in the Pinjor horizon. Upper Pliocene,' some 2,000 
feet above the base of the Tatrot horizon. It also occurs in Austria 
and Bessarabia. 

As early as 1863 Falconer remarked (]). SO): "The nearest 
affinity, and that a very close one, of the I'An-o|)ean E. meridionalis 
is with the Miocene E. (Lo.rod.) planifrons of India." I'alconer 
thus led the way for our present knowledge, namely, that Archidis- 
kodon planifrons is directly ancestral to .4. meridionalis. 

E. [Elephas] planifrons Falconer and Cautley, 1846 [1845]. 
"Fauna Antiqua Sivalensis," 1846, p. 38 [1845, PI. ii, figs. 5a, 
5&]. Lectotype. — A second superior molar with nine ritlges, 

the three anterior ridges being in u.sc (Brit. .Mus. M.3068). Co- 
type. — Broken third molar of lower jaw, I.M,,, with nine ridges 
remaining. Brit. Mus. M.2010. Horizon and Locality. — 

Siwalik Hills, India, Upper Siwaliks, Pinjor horizon. Upper 
Pliocene.' Lectotype and Cotype Figures. — Falconer and 

Cautley, 1846 (1845, PI. n, figs. Tm, .56]. 



'[Recent field studies in northern India l).v Dr. Hcllni\it dc Terra and Perc Teilhard dc Cli.-irdin liavc ofTcrecl strong evidence to support the conclusion 
that the Tatrot and Pinjor horizons arc equivaUnit and are entirely of Lower Pleistocene age. The Boulder Conglomenitc ranges up to t he Middh' Pleistocene 
(see Chap. XXII below on the Geologic- Succession of the Probo.scidea for more detailed discussion). 
January, 1937.— Editor.] 



This note was prepared by Dr. E. H. Colbert, 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



951 



Description. — (Falconer and Cautley, 1846.1, p. 38): "The 
next serial modification in the disposition of the three dental 
substances, and in the consequent form of the teeth, is exhibited 
in fig. 5a of the same plate, which represents a section of the 
penultimate upper molar of another Indian fossil species which we 
have named E. planifrons. This tooth shows nine ridges, the three 
anterior of which alone have been in use, the two first being worn 
down to a single disc of ivory. The common nucleus of this 
substance is of less thickness than in the corresponding tooth of 
E. insignis, and the divisions which are continued upwards from it 
into the centre of the ridges are more elongated, with a narrower 
base, forming irregular-shaped wedges. The layer of enamel is 
diminished in thickness and is less uniform in outline, and the 
surface in contact with the cement shows a feathered or ragged 
edge, indicating superficial inequalities for the firmer cohesion 
of this latter substance. The enamel is reflected over the ridges of 
ivory, and down into the hollows zig-zag wise, exactly as in fig. 6a, 
the principal difference being that the ridges are narrower, with 



between molars of the upper and lower jaws, and of different ages 
in the same species." 

FALCONER'S NOTES OF 1868 ON TYPE .\ND REFERRED SPECI- 
MENS OF ELEPHAS PLANIFRONS 
Falconer, "Palffiontological Memoirs," Vol. I, 1868, pp. 423-442, Plate.s 
11, VI, VIII, XI, XII, and xiv of the "Fauna Antiqua SivalcnsLs." 

Falconer's observations on this species are to be found in his 
successive publications of 1845 (Plates), 1846 (Letterpress), 1857, 
1863, 1865, and 1868. In the latter year ("Palseontological 
Memoirs" of 1868) a summary of all his observations was pub- 
lished by Murchison, covering in detail thirty-nine or more speci- 
mens referred to Elephas planifrons, all presiunably collected 
from the Pinjor horizon, Upper Pliocene.' 

Elephas planifrons: [Lectotype (Burr. Mus. M.3068) and 
CoTYPE (Brit. Mus. M. 2010)]. —Plate ii, fig. 5a, M- with 9 ridges, 
cement filling valleys, enamel folds thick, length 8.7 in. = 221 
mm.; fig. 5b, Ms, vertical section, with 9 ridges. 




Lectotype (R.M") 4ND Cotype (L.M3) of Archidiskodon planifrons 

Fig. 82.'). Lectotype and cotype of Elephas planifrons Falconer and Cautley, 1846 [1845, PI. 11, figs. 5a, 56]. From the Siwalik Hills 
Siwaliks, Pinjor horizon, Upper Phocenc.' One-third natural size. See "Description of the Plates in the Fauna Antiqua Sivalensi.s" (Falconer, 
Falconer [Murchison], 1868, Vol. I, p. 423): 

"Fig. 5a. — . . . Vertical section of penultimate u|)pcr molar [r.M-|, with nine ridges, the tliree anterior of which alone have been in use, the 
worn down to a single disc of ivory." Brit. Mus. M.3068. Molar inverted to show natural ])osition. 

"Fig. 56.— "[Cotype.) Vertical section of portion of last molar of lower jaw [I.M3I, with nine ridges." Brit. Mus. M.2010. Osbor 
cotype M3 is a broken tooth which typically presents ten ridges, sometimes eleven. 

Observe in comparing the original figure (5a) with the new figure (Fig. 827) that Falconer omitted the seven ridge-plated r.M'. See also 
Lydckker. M", ridge-plates pre-concave, post-convex. M3, ridge-plates pre-convex, post-concave. 



, India, Upper 
1867, p. 3; also 

two first being 

n (1924): The 

figure 828 after 



a greater vertical height. The cement substance attains its maxi- 
mum of development in this species, completely filling up the wide 
interspaces of the ridges, over which it is continued in a thick mass. 
This tooth measures 8.7 inches in length." 

"Fig. 5b, represents a corresponding section of a portion of the 
last molar of the lower jaw of the same species, comprising nine 
ridges. This tooth had been longer in use than that of the upper 
jaw, and all the ridges are more or less worn except the two last. 
It presents the same general characters exhibited by fig. 5a, in the 
elongated cuneiform ivory ridges, unequal enamel, and abundant 
cement, the differences being merely such as constantly hold 



Ridge Formulae in Upper and Lower Grinding Teeth 
Upper Jaws. — Plate vi, figs. 4, 5, 6. Skull with premolar, 
?Pm, Dp^ M\ Left Pm with 3 ridges and 2 half-ridges. Dp< with 
6 ridges also a heel and front ridge, enamel thick. (Footnote, p. 
427): Palate with M', 7 ridges and 2 half-ridges; length 6.5 in. 
= 165 mm., width 3.4 in. = 87 mm., height at 4th ridge 3.5 in. 
= 89 mm., fifth and sixth ridges with 6 conelets each. Plate xi, 
fig. 1, aged palate with r. and l.M', 10 ridges including posterior 
half-ridge, enamel very thick; fig. 4, M^ right side, remarkable 
fragment, 6% ridge-plates, length of fragment 5.4 in. = 135 mm., 
width 2.5 in. = 63 mm. Plate xii, figs. 1, la. Dp- with 4 ridges; 



'[See footnote on previous page. — Editor.] 



952 



OSBORN: THE PROBOSCIDEA 



fig. 2, section of r.Dp', with 6 main ridges, a baciv talon and front 
heel, total 6 and 2 half-ridges, length 3.8 in. =97 mm., width 
2.3 in. = 59 mm. ; figs. 4, 4a, M' with 5 ridges and heel remaining, 
two gone, 7 probable total, length 5.2 in. = 132 mm., width 2.8 
in. = 71 mm.; fig. 5, M' [same as fig. 4, this plate], M-, the latter 
with 8 ridges and front heel, length of M- 8 in. = 204 mm., width 
3 in. = 77 mm. ; fig. 5«, palate with r. and l.IVP, 8 distinct ridges and 
front and back heel, total of 8 ridges and 2 half-ridges, conelets few 



on 4th ridge, length of Mi 6 in. = 153 mm., width 2.8 in. =71 mm., 
length of Mo 8 in. =204 mm., width 2.8 in. =71 mm.; fig. 7, left 
lower jaw, M3 very aged, length of molar 10.2 in. = 259 mm., 
width 4 in. = 102 mm., three mental foramina; fig. 8, fine lower 
jaw fragment, M2 [M3, Osborn] complete, with 11,'^ ridges, enamel 
very thick, plates wide apart, much cement, few conelets, three 




ChxuycUgarh'/f- - 



2400 




Vip,. 821). Fii\'oral)lo cxpo.surc.s, southwest of Simla, of the Archidiskodon 
planifrnns life zone, Ujipor Siwaliks, India, chiefly Pinjor horizon, from which 
were collected by Barnum Brown in 1922 twenty-seven or more specimen.s 
referable to thi.s .sjiecies, especially from Kalka, Cliarnian, Siswan, Chandigarh, 
and Mirza|iur, also transitions to .1. mcridimmlis. Compare figures 820, and 
720, PI. x.w. 



and large, enamel thick, length of l.M- 7.5 in. = 191 mm., width 
3.2 in. = 81 mm., fig. 6, M- with 8 main ridges. Plate xiv, fig. 8, 
r.M', showing 8 or 9 ridges and a heel, enamel thick, length 10 in. 
= 255 mm., width 3.5 in. = 89 mm. 

I,oWKU .Iaws. — Plate viii, figs. 2, 2«, jjcrfcct lower jaw, M3 
with a total of 911 ridges, enamel very thick, length of M.i 8.8 in. 
= 224 mm., width 3.8 in. = 97 mm. Plate xi, fig. 2, sti])erl) lower 
jaw, two mental foramina, r. and l.M.i with 13 ridges and front and 
back half-ridges, enamel very thick, length of r.Mj 11.8 in. =300 
mm., width 3 in. = 77 mm.; fig. 3, superb loft half jaw, only 7 
remaining ridges in M3, enamel thick, cement abiuidant, "Proved 
to be E. planifrons by the distance between the plates, the \ery 
low crown, thick enamel, and two mental foramina," extreme 
length of jaw 24.2 in. = 014 nmi., lengtii of M.i 10 in. = 255 mm., 
width of M,i 3.6 in. =91 mm.; fig. 5, r.Ms, enormous tooth frag- 
ment, very thick enamel, low ridges, and mesial exi)ansion, 9 ridges 
remaining, length of M., 10.5 in. =267 nun., width 4.2 in. = 105 
mm., height of 9th ridge 3.5 in. = 89 mm.; fig. 6, left jaw with 
l.Mi, ridge-plates 6-I-, M2, ridge-i)latcs 9 and .«mall heel, 5 conelets 



3'\ 
5v 



Type 
Brit. Mus. M.3068 
Right Second 
Superior Molar 
rMI. I 




New Lectotype Figure of Archidiskodon planifrons, R.M- 
Fig. 827. New lectotype figure of Elephas planifrons Falconer (Brit. Mus. 

M.3068), 9+ ridge-plated second superior molar, of the right side, r.M^, with 

worn crown of a ?7/i ridge-plated first superior molar of the right side, r.M'. 

Redrawn by Miss G. M. Woodward for this Memoir; one-third natural size. 

Compare figure 825. 

Ob.serve in comparing this figure with figure 8.54 of the same molar, that 

the ridge-plates agree in size with those of Piltdown, England (Fig. 853). 



mental foramina, length of M2 [M.i, Osliorii] 12.1 in. =308 mm., 
width 3.8 in. = 97 mm.; fig. 9, fragment of I.M3 nuich worn; fig. 
10, lower jaw, M2 with 8}^ ridges, cement abundant, enamel thick, 
no crimping, and no mesial expansion, length of M2 7 in. = 178 
mm., width 3.7 in. = 95 mm., I.M3 height of 5tli imworn ridge 
3.8 in. = 97 mm. Plate xii, fig. 7, right lower jaw, with Dpa with 
6 main ridges and small half-ridge (same formula as in Loxodotila 
(ifricniia); fig. 8, left lower jaw, with small \erticMlIy succeeding 
premolar (c), above it I)p3, also D])^ with 7 main ridges, double 
front heel and small half-ridge behind; figs. 10, 10a, first left true 
molar, l.Mi, with 7 main ridges, a small ridge in front, no heel 
behind, length of Mi 6.7 in. = 171 mm., widlli 2.3 lo 2.6 in. = 59 
to 66 mm.; figs. 12, 12rt, Vr.Ma, "apparently of a small sized 
indixidual," length 10 in. =255 nun., width in front 3.5 in. =89 
nun., height of crown at 7th jilatc 4 in. = 102 nun., ten main plates 
and a front pl.-ilc and lieel; fig. 13, left lower jaw, I.M3 entire, 
with about 13 ridges and a heel, or |)ossil)ly 14, length of M3 
12.7 in. =323 nun., witHli 3.(5 in. = 91 mm., lieiglit at 10th ridge 4.5 
in. = 114 nun. Plate x\\, fig. 9, last lower molar, right side, r.Ms, 
length 9.5 in. = 241 nun., width 3.5 in. =89 mm., height at 6th 
ridge 3 in. = 77 nun., shows 8 ritlg(>K and a heel. Plate xviii.a, 
figs. 1, 1(7, an I.M3, an enormous specimen; 8 plates, lengtii 10.4 
in. = 265 mm., w idth 4.1 in. = 104 nun., hciglil 3.2 in. = 81 mm. 



THE MAMM0NTINJ5: ARCHIDISKODON AND METARCHIDISKODON 



953 



Collective Ridge Formula (Falconer, 1868). — (1) The 
]ieiiiuinent, foiiith premolar, P 4, i)ersists, with 3 ridges. (2) From 
the specimens of Elephas plaiiifrons, ag described by Falconer (39 
of which appear in Table VI "Summary of Measurements" above), 
from tlie Upper Siwaliks, we deduce the collective ridge formula 
below; (3) the minimum numbers represent partly worn or partly 
developed teeth, while the maximum numbers represent fully worn 
and fully developed teeth; (4) half-ridges develop both in front 



regarded as the l.ypr [lectotype] (see Fig. 825). From I^ydekker's 
descriptions the following characters may be summarized : 

(1) Both the third and fourth milk molars were vertically 
succeeded by ])rem()lars, thus the dental formula is: D)V'"^ Dpo.4, 
r^"* P3-4, M'"^ M1-3. This succession is an important character 
which must be looked for in the ancestors of this species; it does 
not occur in any other species of the Elephantidae thus far known; 
nor is it found in Slegodon bombifrons. (2) The correct collective 



'/-^\ 




Elephas planifrons. — Vertical and longitudinal section of the second upper true 

molar; from the Pliocene of the Siwalik Hills. J. 

rig. 828. Lectotype nine and a third ridged upper molar, r.M-, of Elephas 
[Archidiskodon] planifrotis. Reproduced from a wood engraving by Lydekker 
(1886.2, p.l02, fig. 24). Brit. Mus. M.3068. One-third natural size. Inverted 
to .show natural position of molar. 

a, cement; 6, enamel; c, dentine. 

We observe that the cement (a) completely fills the valleys. The enamel 
(6) is extremely thick, since the sides of the enamel ridge-plates diverge to- 
wards the apex and converge towards the base; the cement interspaces 
between the ridge-plates become constantly narrower as the crown wears 
down, while the dentine within the enamel becomes constantly broader, as 
shown in the figure opposite. 



and behind the main ridges; (5) the collective formula below 
indicates the half-ridges as well as the main ridges. 

Maximum and minimum collective ridge formula [of ascending 
mutations] of Elephas planifrons: 



Dp 2- Dp 3 



H-6-H^P4«^) Dp4*^^M 1 



6 45 



i4-7 



M 2 ^''— *— 

•IVl ^ 844-9-W 



M3 



8-9-VS-K-lg4i 

8-9-M-l 0-Vi-V5-l 2-W-l 4-W 



(6) It is interesting to observe the half-ridges arising both in 
front and behind the main ridges, which seems to be a char- 
acteristic feature of this species, as compared with Elephas hysudri- 
cus. (7) A primitive or simplified formula from Falconer of the 
minimum ridges in E. planifrons from the Upper Siwaliks of India 
would be: 

Dp 21 Dp 3 ^^ (P 4 ±^) Dp 4 ?± M 1 '4 M 2 f M 3 |f±. 



SPECIFIC CHARACTERS OF ARCHIDISKODON PLANIFRONS 
(LYDEKKER, 1886; OSBORN, 1924) 

The next review of Archidiskodon planifrons is that of Lydek- 
ker (1886.2, pp. 98-107) in which the cranium (Brit. Mus. M. 3060 
— our Fig. 830) was selected as the type ; in the present Memoir the 
specimen first described and figured by Falconer and Cautley is 




Elephas planifrons.— The hinder half of the third right upper true molar 
from the Pliocene of the Siwalik Hills. %. The lower border of the 
figure is the inner border of the specimen. 

Fig. 829. Referred right M' of Elephas [= Archidiskodon] planifrons. 
After Lydekker, 1886.2, p. 102, fig. 25. Brit. Mus. M.3070. Two-thirds 
natural size. Figured by Falconer and Cautley, 1846 [1845, PI. xi, fig. 4|, 
and made one of the genotypes of Leilh-Adamsia Matsumoto, 1927, but 
regarded by the present author as a synonym of Archidiskodon planifrons. 
(See Fig. 847 below.) 

As compared with figure 828 this is a crown view of a much worn third 
superior molar in which the enamel loops in the middle of the crown are 
contiguous, the dental areas are expanded, the cement areas are contracted. 
The cement extends beyond the outer borders of the plates, a highly char- 
acteristic feature of Archidiskodon. 



ridge formula is that deduced from Falconer as his maximum- 
minimum or typical, as given above. The ridge formula of 
Lydekker (1886.2, p. 99) is incorrect. (3) Crowns of molars broad; 
cement in the interspaces frequently very great; enamel usually 
thick, frequently devoid of plication in the middle and near the 
root of the crown ; plication near the summit of the crown relative- 
ly coarse. (4) Ridge-plates subeUipsoidal, frequently with a mid- 



Fig. 830. Skull of Elephas 
[= Archidiskodon] planifrons from the 
Siwalik Hills, India, reproduced from 
Gaudry, 1878, p. 185, fig. 246, after Fal- 
coner 1846 [1845, PI. x). One-.sixteenth 
natural size. This .skull (Brit. Mus. 
M. 3060) was erroneously selected by 
Lydekker as the type of Elephas 
planifrons, and is figured in his 
"Catalogue of the Fossil Mammalia in 
the British Museum (Natural HLstory)" 
of 1886 (1886.2, woodcut fig. 23, p. 
100). Compare figures 848 and 817 of 
the same cranium. 

Deperet and Mayet regard this cranium as that of a female, which, if 
true, partly accounts for its small size (.see p. 962 of the present Memoir). 




954 



OSBORN: THE PROBOSCIDEA 



expansion, sometimes imperfectly lozenge shaped, in slightly 
worn teeth the middle portion of each ridge frequently forms an 
isolated disk. (5) Cranium characterized by flatness of the fronto- 
jiarietal region, by a small incision of the temporal fosss on the 
frontals, by a comparatively slight elevation of the vertex (hypsi- 



cephaly), by smallness of the nasal ajierture, by wide di\-ergence of 
the incisive alveoli, as in ^4. meridionalis; vertex of skull flattened 
almost at right angles to occiput, occiput pitching forward 55° 
when the grinding teeth are horizontal; tusks very stout and 
incurved. 



REVISION OF ,\RCHIDISKODON PLANIFRONS BY OSBORN (1927-1928), BASED ON THE SIWALIK COLLECTION 

MADE BY BARNUM BROWN IN 1922 

The twenty-seven specimens referable to Archidiskodon ■planijrons and its ascending mutations were found in the 'Upper 
Siwaliks,' below the Boulder Conglomerate formation, and recorded as follows: 

Table VH. Mea.surements of Twenty-seven Specimens Collected by Brown and Referred by Osborn to Archidiskodon 

planifrons, ascending to a. meridionalis 

All recorded by Barnum Brown from near Kalka, Charnian, Siswan, Chandigarh and Mirzapur (see Figs. 820, 826) 



Superior 


Figure 


Amer. 
Mus. 


Molar 


Length 
mm. 


Breadth 
mm. 


Index 

est. 


Height 
mm. 


Ridge-plates 
act. or est. 


Ridge-plates 
in 10 cm. 


Third Molars 




19882 


r.M" 


254 


96 


38 


9th 


= 102 


12+ 


5K-6 




834 


19950 


l.M= 


219 


100 


46 


7th 


= 88e 


9-}HH^=12 


4+ 






19880 


1 . M'^ 


201 


88 


44 


9th 


= 95 


11 + 


53^ 


First or Seccmd Molars 


8331 
855/ 


19821 


i.:\p 


217 


94 


43 


7th 


= 66e 


8-K-K=10 


3K-4 




832| 


19955 
19955 


r.M' 
l.M' 


172 
162 


76 

80 


44 
49 


6th 
5th 


= 70 
= 65 


K-7-K= 9 
K-6-?^= 8 + 


4-^M 
4-4K 




831 


19881 


l.M' 


165 


90 


54 


3d 


= 67 


K-6 = 7 + 


3K-4 






19910 


?r.M' 


138 


80 


58 


5th 


= 107 


J^-6 


3K-4 


Inferior 
Third Molars 


8421 
845; 
855J 


19951 


r.Ma 


313 


101 


32 


7th 


= 124 


UK =12 


3K-4K 




841 


19861 


I.M3 


250 


79 


32 


9th 


= 88e 


n% =12 


5 




846 


19879 


l.M, 


190 + 


105 + 


55 


8th 


= 135 


10+ 


5-5K 




844 


19819 


I.M3 


292+ 


95 


32 


7th 


= 88e 


10K2 =11 + 


4 




855 


19798 
19870 


r.M, 

l.M 3 


203 


80 


39 


7th 
2d 


= 53 

= 85 


11+ 


6)^ 




843 


19968 


I.M3 


213e 


95 


45 


7th 


= 66e 


9+ 


5 




840 


19965 


r.M, 


180 (inc.) 


86 


48 


7th 


= 55e 


9 


4-4M 




839 


19864 


r.M, 


194+ 


78 


40 


3d 


= 45 


K-8-K=10+ 


5 




835 


19952 


l.M, 


233e 


109 


47 


6th 


= 114 


8Ke= 9+ 


3K-4 


Second Molars 


836 


19778 


r.Mo 


140+ (inc.) 


90 + 


64 


5th 


= 72 


/2-5+= 6+ 








19S71 


1 . Mo 


167 + 


87 


52 


8th 


= 99 


8 + 








19870 


l.Mo 








10th 


= 82 






First Molars 


838 


19862 


?l.Mi 


160e 


80 


50e 


3rd 


= 45 


M-8 =9 + 


5 






19917 


r.Mi 


190+ 


78 


41 


5th 


= 91 + 


+7M 


5 






19961 


?l.Mi 


189 


90 


48 


7 th 


= 73 


7 


4 






19871 


l.M. 


127 


85 


67 






y2-6-K 






836 


19778 


r.M, 


103 


70 


68 






3+ 




Fourth Deciduous 


837 


19873 


r.Dp4 


121 


63 


52 


7th 


= 44 


%--!-%= 9 


7 



THE MAMMONTINiE: ARCHIDISKODON AND METARCHIDLSKODON 



955 



As remarked abo\-e, the large American Museum collection of 
Archidiskodon -planifrons, including examples of the entire den- 
tition, but lacking a skull, renders it highly i)robable that the 
Falconer and C'autley collection of A. phuiifroiis, figured and 
listed in detail, as above described, came from the same geologic 



localities of the Pinjor horizon level, namely, from deposits around 
Kalka (see Figs. 826, 820). The following 29 si)ecimens, belonging 
to 24 individuals, are given with measurements above (Table VII) ; 
tliey are also illustrated in great detail in the accompanying 
figures (Figs. 831-844). 



Near Kalka 

Near Charnian 
Near Siswan 



9 miles west 
6 miles west 
2 miles south 



3 miles north 
3 miles north 
3 miles north 



Near Chandigarh 3 miles west 

3 miles west 
3 miles west 
1 mile west 
1 mile west 
3 miles west 
1 mile west 
3 miles west 

Probably near Chandi- 
garh, record incomplete 

Near Mirzapur 3 miles north- 

east 



Amer. Mus. 19778 
Amer. Mus. 19798 
Amer. Mus. 19819 
Amer. Mus. 19821 
Amer. Mus. 19870 



Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 
Amer. 



Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 
Mus. 



19879 
19880 
19881 
19882 
19871 
19873 
19968 
19965 
19862 
19861 
19916 
19950 
19951 
19952 
19955 
19961 
19917 



Amer. Mus. 19967 
Amer. Mus. 19864 



Fragment of right jaw with Mi and M2 in situ (Fig. 836 for M,). 

Fragment of jaw with r.Ms in situ (Fig. 855 FOR r.Mj). 

Fragment of left jaw with I.M3 in situ (Fig. 844). 

Fragment of maxilla with l.M- in situ (Fig. 833). 

Fragment of jaw with I.M2 and I.M3 in situ, M2 incomplete, M3 partly 

erupted (not figured in ])resent Memoir). 
Fragment of jaw with l.M 3 in situ (Fig. 846). 
Left AF (not figured in present Memoir). 
Left M' (Fig. 831). 

Fragment of maxilla with r.M'' (not figured in ]iresent Memoir). 
Fragment of jaw with l.Mi and M2 in situ (not figured in pre.sent Memoir). 
Fragment of jaw with r.Dp4 in situ (Fig. 837). 
Fragment of jaw with I.M3 in situ (Fig. 843). 
Fragment of jaw with r.Ms in situ (Fig. 840). 
Fragment of jaw with l.Dp4 and l.Mi (Fig. 838). 
Fragment of jaw with I.M3 in situ (Fig. 841). 
?Right M' (not figured in present Memoir). 
Fragment of maxilla with l.M' in situ (Fig. 834). 
Right M3 (Fig. 842). 
Left M3 (Fig. 835). 
Right Ml and left M' (Fig. 832). 
?Left Ml (not figured in present Memoir). 
Right Ml (not figured in pre.sent Memoir). 

Lower jaw, left ramus. 

Fragment of jaw with r.Ms in situ (Fig. 839). 



SUMMARY OF DENTAL MEASUREMENTS OF THE AMERICAN MUSEUM COLLECTION MADE BY BARNUM BROWN IN 1922 



These measurements accord in general with those of the 
Falconer and Cautley Collection in Table VI above and exhibit 
variations in size, length, breadth, and proportions, partly attrib- 
utable to male or female sex, partly to progressive ascending muta- 
tions. The increase in number of the ridge-plates and half ridge- 
plates is probably attributable to progressive evolution or to 
ascending mutations ranging into higher geologic levels. In 
general, the smaller animals, with fewer ridge-plates, probably 
occurred in lower geologic levels than the larger animals, with more 
numerous ridge-plates and other progressive characters. 



Third Superior Molar, M' 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 

Second Superior Molar, M^ 
(Only specimen available in Amer. 
Mus. Coll., No.19821) 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 



Minimum 
201 
88 
38 
66e 
9-1- 



217 
94 
43 
66e 
10 



Maximum 
254 
109 

47 
102 

12 



First Superior Molar, M' 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 

Third Inferior Molar, M3 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 

Second Inferior Molar, M2 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 

First Inferior Molar Mi 
Length 
Breadth 

Breadth-length index 
Height 
Ridge-plates 



Minimum 


Maximum 


138 


172 


76 


90 


44 


58 


65 


107 


7 


9 


190 + 


313 


78 


105 


32 


55 


45 


135 


9 + 


12 


140-h 


167 -h 


87 


90-1- 


52 


64 


72 


99 


8 + 


8+ 


127 


190-1- 


85 


78 


67 


41 




91 + 


/2-6-K 


+VA 



^A/dTsije 




ARCHIDISKODON PLANIFRONS «»/. 



Amer. Mus. fSBBJ 

3 _4- J 



L.M 





L.M- 

ARCHIDISKODON PLANIFRONS Ktf. 

Amer. Afuj. /S>$SS 




'A A/at. sij 




I 

ARCHIDISKODON PLANIFRONS Ri-/ Amer. /^U5. /982/ 




Fig. 831 



Fig. 832 



Fig. 833 



ARCHIDISKODON PLANIFRONS «./ 




American Museum Collection 

Upper Pliocene Abchidiskodon plani- 

frons of the plnjor horizon 

(see Fig. 826) 

Figs. 831-835. Archidiskodon planifrons, 
referred superior and inferior grinding teeth of 
the Barnum Brown Siwalik Collection, listed 
with measurements in Table VII. All figures 
drawn to the same one-fourth scale. Cement 
(dotted), dentine (horizontal lining). Com- 
pare Falconer's measurements in Table VI. 

Figures: (831) First left superior molar, 
l.M'. (832) First left and right superior 
molars, I.M', r.M'. 

(833) Second left .superior molar, l.M^; 
a primitive stage as shown in midsection 
(Fig. 855); this M- agrees closely with that of 
Piltdown, Sussex (Fig. 853). (834) Third 
left superior molar, 1.M^ 12 plated. 

(835) Third left inferior molar, I.M3. 



i^ A/at. sije 



Fig. 834 




956 






Fig. 836 



Fig. 837 



Fig. 838 





Fig. 839 



Fig. 840 



American Museum. Upper Pliocene Archidiskodon planifrons of the Pinjor Horizon (see Fig. 826) 
Fig,';. 836-840. Inferior grinders of Archidiskodon planifrons (Figs. 836-840), in the Barnum Brown Siwalik CoUeRtion, fully listed with measurements in 
Table VII. All figures one-fourth natural size, excepting 837 which is one-third natural size. Cement (dotted), dentine (horizontal lining). Compare Falconer's 
measurements, Table VI. 

Figures: (836) Portion of first and .second right molars, r.Mi, r.M^. (837) Fourth deciduous iiremolar, Dp4, with 9 ridge-plates. (838) Worn crown of 
left fourth premolar, l.Dp^, and first molar, l.Mj. (839) Portion of lower jaw with third right molar, r.Ms. (840) Portion of lower jaw with third right 
ranlar, r.M.i. 

957 



jL=6=s^g=S: 




Fig. 843 Fig. 844 

American Museum Archidiskodon planifrons of the Pinjor Horizon (see Fig. 826). Inferior Molars 
Figs. 841-844. Archidiskodon planifrons referred infcuor grmdiu^ teeth and jaws in the Siwalik CoUoction of the American Museum made by narniini 
Brown, as fully listed with measurements in Table VII. Ail figures one-fourth natural size. Cement (dotted), dentine (horizontal lining). Comi)are 
Falconer's measurements, Table VI. 

Figures: (841) 12 plated left third molar, I.M3. (842) 12 plated right third molar, r.Mj, of a slightly older individual; a progressive stage as shown in 
midsection (Fig. 85oB). (843) 9+ plated left third molar, I.M3, imperfect anteriorly. (844) II plated ieft third molar, I.M3. 

958 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



959 




/Imer: Mus. /995/ 
DON PLANIFRONS Rtj 

Upper Pliocene Archidiskodon planifrons of the Pinjor Horizon 
(see Figs. 826, 820) 
Arclndiskodon planifrons ref., third right inferior molar, r.Mj 
19951), in the Barnum Brown Siwalik Collection, one-third 
Compared with the collective figures above, this specimen is 
very close to the typical ridge formula of Archidiskodon planifrons, namely, 
M 3TTm ridge-plates. I..ength 313 mm., width 101 mm., height of seventh 
ridge-plate 124 mm. (see Table VII, p. 9.';4). 



Fig. 845. 
(Amer. Mus. 
natural size. 




LEITH-ADAMSIA SIWALIKIENSIS 
Figure 847 
Lp:ith-Adamsia siwalikiensis' Matsumoto, 1927. — "On 
Leith-Adamsia siwalikiensis, a New Ceneric and Specific Name of 
Archetypal Elephants." Japanese Journ. Geol. and Geog., V, 
No. 4, Art. 12, 1 page. Type. — Two superior molars, both of 

the right side, r.M', in the collections of the British Museum 
(Bfit. Mus. M.3070 and 36695). From India. Type 

Figure. — Falconer, "Fauna Antiqua Sivalensis," Pis. xi, fig. 4, 
and XIV, fig. 8. Type Description.— (Matsumoto, 1927.2): 

"In the course of a study of fos.sil Elephants, the writer has come 
to be faced with the serious need for a ]>roper name for a certain 
type of archetypal Elephants of India. He means the small and 
narrow-molared form recorded under the name of Elephas plani- 
frons. This form appears, in all likelihood, to stand at the starting 
point of the entire phylum of the Loxodontine Elephants. As it 
would appear to occupy too important a position among the 
Elephantidse to be left unnamed, the writer proposes here to call 



^.Af./S^S7S ysA/oTstje 




Upper Pliocene Archidiskodon planifrons of the Pinjor Horizon (see Fig. 826) 
Fig. 846 (right). Archidiskodon planifrons ref., a third inferior molar of the left side, I.M3 (Amer. Mus. 19879) in the Barnum Brown Siwalik Collection, 
listed with measurements in Table VII; portions of ten ridge-plates are preserved, as displayed in the transverse vertical section of the same tooth (left). 
All figures one-third natural size. Compare Falconer's measurements Table VI, p. 949. 

'[Regarded by Professor Osborn as a synonym of Elephas [Archidiskodon] planifrons. — Editor.] 



9G0 



OSBORN: THE PllOBOSCIDEA 



it by a generic name as follows, in honour of the late Professor 
Doctor Leith Adams, the eminent i)akpontologist and specialist on 
fossil Elephants." 

" Leith- Adamsia, gen. no\-. A genus of archetypal Elephants. 
Cheek-teeth subhypsodont, narrow-crowned, with a low ridge- 
formula; lo.xodont sinus present, and of an obtuse ty])e; disks 
of well-worn ridges may be more or less lozenge-shaped." 



/■i,/ t 




profile of A. meridionalis (Fig. 817). Meanwhile the cranium 
heightens (hypsice])haly) and dee])ens (bathyceplialy), in adapta- 
tion to the enlarging and deepening of the third superior molars, 
also, as fully explained in Chapter XV, j). 915, on the mechanics of 
the proboscidean cranium, the heightening of the occipital crest 
(acrocephaly) is in adaptation to the elongating tusks and the 
strengthening of the cervical ligaments and muscles which sway 
the great tusks and proboscis. Nothing is known of the limb 
skeleton of A. planifrons, but it is inferred that the animal was 
greatly inferior in shoulder height, in length, and in i)roboscis 
development to its giant successors A. meridionalis and A. impe- 
rator. The completely preserved tusks in members of this species 
discovered in southern France (Fig. 850) lack the strong outward 
curvature and incurvature characteristic of A. meridionalis 
(lyrodon), see figure 864, reaching a supreme stage in A. imperator. 




Referred Archidiskodon planifrons 
Fig. 847. Type molars of Leith-Adamsia siwalikiensis Matsumoto, 
1927. After Falconer and Cautley, 1846 [1847, PI. xi, fig. 4, and Pi. xiv, fig. 8), 
one-third natural size, reduced to one-fourth natural size. 

(PI. XI, fig. 4) Brit. Mus. M.3070. An r.M', posterior half of crown, 
exhibiting 6^ ridge-plates (same molar as in Fig. 829, above). 

(PI. XIV, fig. 8) Brit. Mus. 36695. An r.M^ with 6^ posterior ridge-plates 
preserved of a probable total of 10 ridge-plates. 



CRANIAL CHARACTERS OF ARCHIDISKODON PLANIFRONS 

The single cranium known in the British Museum collection 
(Brit. Mus. M.3060— .see Figs. 830, 848, and 817) is extremely 
primitive, somewhat resembling that of certain species of Stegodon, 
for example, Stegodon pinjorensis type (Amer. Mus. 19772 — see 
Figs. 711, 765, and 817C), except that Archidiskodon planifrons is 
much less elevated (acrocephalic) than S. pinjorensis, as clearly 
shown in the comimrative profiles (Fig. 817). This is in adaptation 
to the relatively abbreviate 10 -|- ridge-plated crown of M' in A. 
planifrons, as compared with the extremely elongate 15 ridge- 
Ijlated crown of M' in .S. pinjorensis. Also compare the flattened 
forehead of .4. planifrons with the abbreviated and extremely 
elevated forehead of S. pinjorensis. 

This flat-faced condition, to which Falconer assigned tlie 
specific name planifrons, develojis into tlie concave-forehead 




..) i. 



..V ARCHIDISKODON 
i Xz'VoTsije 



PLANIFRONS /?e/ 

A/ter Falconer Platt$ S \ 10 



Supposed Female Cranium 
Fig. 848. Cranium of Archidislcodon planifrons in the British Museum 
(Brit. Mus. M.3060) redrawn after Falconer and Cautley, 1846 [1845, PI. ix, 
front view of skull, PI. x, side view of skull). Five views of this unique cranium 
are shown in Falconer and Cautley's plate.s ix and x; a brief descrijjtion in 
the 'Talaeontological Memoirs" of 1868, Vol. I, p. 430, is as follows: 

"Plate IX. Elephas planifrons (Falc. and Caut), from the Sewalik Hills. 
Front view of skull, one-third of natural size. The forehead of this species is 
very flat; the naso-maxillary opening very small, and the occipital fissure 
very low. . . . Plate x. . . . The last true molar is seen in germ and intact on the 
right side, and well worn on the other, so that the corresponding tooth on the 
right side of the lower jaw had probably been wanting. It has eleven ridges 
and a heel. The pterygoids are very low." 

The principal measurements given by Falconer (cf. op. oil., 1868, p. 430) 
arc the following: 

Extreme length from occiput to incisive alveoli.. .25 in. =635 mm. 

Extreme width of occiput 21 .7 =551 

Height of occiput 13.7 =348 

Occiput to anterior border of orbits 20 . 7 = 526 

Length of right M' 9.7 =247 

Width of crown 3.5 =89 

Height of crown plates 4.0 =102 



THE MAMMONTINiE: ARCHIDISKODON AND METARCHIDISKODON 



961 



ARCHIDISKODON PLANIFRONS IN WESTERN EURASIA, AUSTRIA, RUMANIA, AND RUSSIA 



The species of southern mammoth known as Archidiskodon 
meridionalis has long been known in western Eurasia. The dis- 
covery of the more primitive ancestral A. plam'frons in Austria, 
Bessarabia (southern Russia), southern France, and England, is 
relatively recent, dating from Schlesinger's announcement of the 
discovery of 'Elephas' plam'frons in Lower Austria in 1912. 

Austria. — Schlesinger in his "Studien fiber die Stammes- 
geschichte der Proboscidier," of 1912, p. 89, announced the evi- 
dence of Elephas planifrons in Lower Austria: "Um so iiber- 
raschender war es, als dem niederosterreichischen Landesmuseum 
in Wien ein Elefantenmahlzahn von ungemein primitivem Charak- 
ter zukam. Es war sehr naheliegend, das Stiick mit einer der 
beiden im Jungtertiiir Europas nicht seltenen Arten zu identifi- 
zieren. Ein nur oberflachUcher Vergleich riickte den LTrelefanten 
{E. antiqmis Fnlc.) giinzlich ausser Betracht, eingehende Studien 
aber sprachen zufolge eben der Merkmale gegen eine Bestimmung 
als E. meridionalis Nesii, welche den Zahn dem E. planifrons Falc, 
einer typischen Form der indischen Sewalik-Hills [Footnote: 'Ich 
bemerke, dass dies der von H. Falconer und P. Cautley (Palaeon- 
tological Memoirs Vol. I, pag. 31) zum erstenmal gebrauchte 
richtige Name fiir die dem Himalaya siidlich vorgelagerte Hiigel- 
kette ist.'] nahe brachten. 

Die Annahme des Vorkommens einer so ausschliesslich 
sewalischen Art in unserem Gebiete mag vorerst befremdend und 
gewagt erscheinen. Doch schwinden derartige Zweifel alsbald, 
wenn wir bedenken, dass sich die Verbreitung der Riisseltiere, 
wie die etiicher Siiugerstamme, ohne die Annahme ausgedehnter 
Wanderungen nicht begreifen lasst [Footnote: 'Vgl. Ch. Deperet, 
Die Umbildung der Tierwelt (deutsch von R. N. Wegner), pag. 
260 ff., Stuttgart 1909.']." 

This specific determination as well as the geologic age of the 
specimen very carefully described in great detail by Schlesinger 
(op. cit., pp. 94-111) were challenged by Soergel in articles entitled 
"Die Stammesgeschichte der Elephanten," 1915, and "Die Plani- 
frons-Frage," 1921, with the following conclusion, namely, that 
the teeth from Dobermannsdorf, Krems, and Laaerberg are 
actually referable to Elephas meridionalis and that the geologic age 
is not Middle Phocene but of much more recent date. Schlesinger 
replied in great detail in his "Meinc Antwort in der Planifrons- 
frage" of 1916. 

The molar referred to A . planifrons from Laaerberg, said to be 
a third inferior molar of the left side (see Schlesinger, 1916.2, p. 
119, fig.s. 6, 7) does not in Osborn's opinion sustain the specific 
reference of this specimen to "Elephas {Archidiscodon) planifrons" ; 
it is too high and narrow, the ridge formula (IO/2-II), length (280 
mm.), width (87 mm.), index (31), and height of the 4th ridge-plate 
(106 mm.) seem to rule out this third inferior molar from A. plani- 
frons and to relate it to Palseoloxodon [Hesperoloxodon] antiquus. 

Bessarabia, Russia. — In the valuable Memoir by Marie Pav- 
low, "Les Elephants fossiles de la Russie" (1910), cited by Mayet, 
she fully discusses in her description of the elephants of Tiraspol 
(p. 4) the milk and permanent dentition of Elephas planifrons and 
on page 27 she describes "El. id. planifrons Falc, PI. i, fig. 23," 



discovered in ferruginous sands near the village of Farladani 
(Bessarabia), a locality which had previously yielded remains 
of Mastodon (Zygolophodon) borsoni, and thus regarded as of 
Upper Pliocene age. Madam Pavlow concludes (p. 27) : "C'est 
une dent (m"?) tres massive (fig. 23), surtout dans sa partie 
superieure. Elle est tres jeune, a peine usee, sur une surface de 15 
cm., elle n'a que 5 lames aux contours tres irreguliers, a I'email tres 
epais et avec de larges espaces de cement. Les quatre premieres 
lames sont us6es, la 5-e tres peu; la 6-e presque intacte, ce ne sont 
que des rondelettes liees entre elles. Par derriere se trouvent 
encore 9 plaques intactes; elles sont toutes tres robustes et 
mesurent sur le cote 2 cm. entre les enfoncements qui les bornent. 
Je n'ai pu trouver aucune forme en Europe qui pourrait corre- 
spondre avec cette dent par le dessin de I'email, ainsi que par le 
nombre des lames, et c'est parmi des figures donnee par Falconer 
pour I'El. planifrons que j'ai trouve ses semblables. T. XI, f. 57. 
T. XVIII, f. 12. T. XIV, f. 8, 9 (Fauna Antiq. Sivalensis)." 

Summary. — Mayet (1920, p. 310) remarks: "Une revision des 
Elephants pliocenes d'Europe devient n^cessaire, a la suite de 
laquelle une partie des molaires regardees comme appartenant a des 
formes archaiques d'E. jneridionalis seront vraisemblablement con- 
siderees comme provenant de E. planifrons. La necessite de cette 
revision a ete entrevue par d'eminents paleontologistes. La mise 
au jour dans les sables de Chagny d'un E. planifrons I'impose. 
Deja M™'' Pavlow [Footnote: 'Marie Pavlow, Les Elephants 
fossiles de la Russie (Nouveaux Memoires de la Societe imperiale 
des Naturalistes de Moscou, t. 17, p. 2.')] a signale son e.xistence 
dans le Pliocene recent de la Bessarabie et le Dr. Schlesinger 
[Footnote: 'Giinther Schlesinger, Studien iiber die Stammesge- 
schichte der Proboscidier (Jahrbuch der k. k. Geolog. Reich- 
sanstalt, Vienne, 1912, p. 87).'] dans celui de la Basse-Autriche, 
a Dobermannsdorf." 

ARCHIDISKODON PLANIFRONS IN FRANCE 
An indubitable discovery of Elephas planifrons Falc' is that 
recorded by Mayet in his "Decouverte d'un squelette d'Elephas 
planifrons Falconer dans les sables de Chagny, a Bellecroix pres 
("hagny (Saone-et-Loire)," 1920, in the Comptes Rendus (Paris), 
pp. 308 311. 

Geologic Age. — (Mayet, 1920, j). 308): "Une faune abon- 
dante les date parfaitement : les sables de Chagny se parallelisent 
avec les sables a Mastodontes de la region du Puy, avec les alluvions 
tie Perrier et probableinent avec le Crag de Norwich. lis sont de 
I'extreme debut du Pliocene superieur. Les depots marecageux 
du cirque de Seneze sont plus recents; les gisements classiques du 
Val d'Arno superieur, du bassin de Florence, des niveaux fluvio- 
lacustres de I'Astesan, sont du meme age. Cet ensemble differe de 
celui des gisements plus recents qui appartiennent au Saint- 
Prestien: sables a E. meridionalis de Saint-Prest, tufs volcaniques 
de Saint-Martial (Hcrault), limons de Durfort, etc. Dans ces 
gra\-iers de Chagny-Bellecroix, fin mai dernier, a etc decouvert un 
squelette incomplete d'elcjjhant: base du crane (celui-ci tres 
fragmente, dcjjourvu de sa moitic antero-superieure) a\'ec deux 



'[Professor Osborn regarded this specimen as a Lower Pleistocene form (see Fig. 1239 of present Memoir) and intended to make it the type of a new species 
of Archidiskodon. The description, however, was never written. For restoration, see figure 815. — Editor.] 



962 



OSBORN: THE PROBOSCIDEA 




II 



III 



Primitive Lowek Jaws of Auchidiskouon 
Fig. 849. Siipcrpo.sitioii of three mandibles of Elepkas pUiniJroms, Imliii 
ami France, reduced to approximatrcly the same one-tenth scale, after Mayet 
and Roman, 1923, p. 81, fig. 13. See figure 914 below. 

I, Elcfihas planifrons of the Siwaliks. After Falconer. 
IT, Elr/ihas pUmifrnn.'^ of Chagny-Bellecroix, France. See also figure S5(). 
Ill, Elcphas planifrons of Senezc, France. 

We observe the uniform prolongation and beaklike depression of the 
symphysis. In describing these jaws Mayet and Roman remark (op. cil., p. 
79): "Mandibule (pi. i, fig. 1 et 2). Courte, a branches horizontalcs ba.s.ses, 
opaisscs, la mandibule donne rimpression d'etre massive en arriere, amineie, 
effiI6c dans sa partie ant6rieurc a comme projet6e en avant. Cette impression 
tiont il la direction dc la sympliyse et, pour >me grande partie, i\ rcxistcnee d'un 
'bee', d'une ,sort,e d'apophy.se mentonnidre, qui plonge obliciuement. Doja, en 
182.'), Nesti disait quo I'K. meridionaliis etait un Elephant a bee et, dans Fauna 
anliqxia sivaleMsis, Falconer insiste sur le bee d'E. planifrons. II faut toutefois 
rcmarquor, des maintenant (voir aussi 2° partie), que Ic bee de I'E. mcridion- 
alis so dirigc en avant pre-sipie horizontalement et quo colui d'^. planifrons 
jjlonge vers Ic bas on tondant a .so rapprochor do la vertiealo. Ce caractere est 
particuli^rement net sur unc mandibule de Scneze (v. pi. n)." 

'[See footnote on previous page. — Editor.] 



molaires, M'; mandibule avec deux niolaires en place, Mj, et une 
apophyse nientonniere bien developpce, deux defenses, atlas, 
onioplate, femurs, cotes, etc. Ces ossements ne justifieraient qu'- 
imparfaitement la presente Note s'ils ne se rajiportaient a une 
espece non encore identifi(5e parmi les elephants pliocenes de 
1 'Europe occidentale: Elephas planifrons Falconer." 

In the original and subsequent papers the authors compare this 
important specimen both with the typical Elephas planifrons of the 
Pinjor horizon and with Elephas meridionalis. Mayet and Roman 
(1923) devote to this subject an exhaustive review in their "Les 
Elephants Pliocenes" (Premiere Partie), in which by far the most 
profound study is made of the dentition of the Upper Pliocene 
Proboscidea of France, and Depi^ret and Mayet {pp. cit., Deuxieme 
Partie) of the chief specimens referable to this species from the 
Siwaliks (pp. 95-97), from Bessarabia (pp. 101, 102), from Austria 
(the Schlesinger discoveries, pp. 102-104), as well as specimens 
possibly referable to this species from England (p. 104), from other 
parts of France, including especially Chagny (pp. 106-110), from 
Italy (pp. 110-120), and possibly also from Africa (p. 120). From 
all these materials the following specific characters are deduced. 

Characteres Specifiques de L'Elephas planifrons (De- 
PEKET AND Mayet, 1923, PP. 121-123).— "[1] Le crd^e male adulte 
est encore inconnu. Le seul crane qui ait ete dccouvert est le 
crane des Siwaliks figure par Falconer (pi. ix et x), et reproduit 
dans tous les ou\rages. C'est im crane femelle de taille relati\-e- 
ment faible, au \ertex peu eleve, au front i)lat et non excave, peu 
echancrc sur le cote par la fosse temporale; le plan fronto-nasal 
est oblique en avant, tres peu redresse. Ce sont la des caractcres 
qui se retrouvent dans les cranes jeunes et dans les cranes femelles 
des autres especes d'elephants vivants et fossiles. [2] Les defenses, 
bien conservees dans le sujet de Chagny (1" partie, p. 77), sont 
tres rapprochees Tune de I'autre a la base, peu divergentes; elles 
sont longues (plus de trois metres), mais surtout tres epaisses 
(grand diametre, maximum 22 centimetres) ; leur incurvation est 
faible vers le haut et un peu en dehors, avec a peine une tendance 
spiralee. [3] La mandihule bien conservee des pieces de Chagny et 
de Seneze presente un caractere distinctif importavt: la symphyse 
se jirolonge en avant par une sorte de hec allonge ct recnurbe en bas. 
Ce caractere existe egalement dans la mandibule des Siwaliks 
figuree par Falconer (pi. viii, fig. 2) sous le nom erronc d'E. 
hysudricus. On voit egalement ce bee un peu brise au bout dans la 
mandibule du Serre, pres Peccioli (Musee gdologique Florence). 
[4] La dentition est complete: 3 molaires de lait et 3 arriere- 
molaires. On a observ6 dans les pieces de I'lnde la presence de 
deux premolaires inferieures de seconde dentition (fig. 15, p. 122), 
caractere unique dans tout le groupe des Elephants. Ces pre- 
molaires n'ont pas encore etc observees dans les sujets d'Europc 
et c'est la seule objection qui puisse subsister sur leur assimilation 
specifique avec I'espece des Siwaliks. Nous pensons qu'on ob- 
servera ces premolaires lorsqu'on aura des mandibules a l'6tat 
d'evolution individuelle convenable. [5] Parmi les molaires, les 
Mm sont les plus caracteristiques: la formule dentaire est de 9 
ii 10 lames en haut, et de 10 a 11 en bas, chiffres un peu inferieurs 
a ceux de V K. meridionalis. La frequence laminaire est de 3, 5, 
a 4 ])our 10 centimetres de longueur de couronne, chitTre toujours 
inferieur a. celui de VE. meridionalis (4, 5, k 5). La couronne est 
notablcment plus basse que dans cette derniere espcice; elle est 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



963 



d'un type remarquablcment brachyodonte. Les lames sont larges, 
I'email epais et generalement peu festonne, la plicature n'interes- 
sant qu'exceptionnellement la totalite du cordon d'email. On 
observe un sinus loxodonie inconstant, plus frequent sur le cote 
posterieur des lames, mais existant parfois des deux cotes. [6] La 
taille de \'E. planifrons est considerable: la hauteur du sujet de 
Chagny au garrot devait etre d'environ 3 m. 85 metres: c'est une 
dimension a peine un peu inferieure a celle d'E. meridionalis." 




Fit.. 12. — l.leph.\s pUnifrons de Ilellccroix-Chapny. 

Primitive Tusks and Jaws ok Archidiskodon of Chagny 
Fig. 850. Tu.sks, ma.xillie, condyle, and atlas of Elephas planifrons of 
Chagny. After Mayet and Roman, 1923, jj. 77, fig. 12. From the sables de 
Chagny at Bellecroix near Chagny, France. Upper Pliocene. About one- 
thirtieth natural .size. This .si)ecimen is carefully described (op. oil., 1923, pp. 
7.5-80). The tusks are massive, nearly straight when viewed from above, with 
a very feeble torsion from within outwards. The tusks measure: right 2.15 m., 
left 2.18 m. The third sujierior molar measures: ap. 264 mm., tr. 117 mm.; 
length of median lamina 26 mm. The lower molars measure: ap. 305 ram., 
tr. 108 mm., height 44 mm.; laminar formula 10 + ; sujjerior molars 9+, 
inferior molars 10+. For restoration, sec figure 815 above. 

Description of Chagny Specimen. — An animal from the 
Upper Pliocene' of Chagny, France, is characterized by Mayet 
(1920, i)p. 309, 310) as follows: "Nous pouvons ra])])rochcr notre 
clcjjhant de Bellecroix de Elephas plniiifrons du Pliocene moyen et 
superieur de I'Inde (Pinjor horizon, Siwalik-Hills) : meme formule 
^Yoif meme couronne trcs basse, meme caractcre de I'email, 
meme aspect mastodonto'idc de la mandibule, du menton, des 
defenses, il y a identite des caracteres specifiques." 

Frequence Formule 
"Es])cces. Eapportf-. laminaire. dentaire. 

E. planifrons, Chagny 2,28 4 

E. planifrons, Siwalik- 
Hills 2 a 2, 5 4 

The number of plates in 10 cm. ("frequence laminaire") is 
four as compared with five to six in E. meridionalis. The x of 
Mayct's dcscrijjtion refers to the rudimentary plates at the front 
and back of the crown. It would ai)i)ear that the ridge formula of 
E. planifrons in comparison with the Chagny specimen may be 
written: 

M 2f M 3f^ 

'[See footnote on page 961 above. — Editor.] 



x-l 0-J 
x-l O-x 



__x-l_o-x_,, 
x-l 0-1 1-x 



ARCHIDISKODON MERIDIONALIS [AND ?PLANIFR0NS1 IN THE 

PLIOCENE RED CRAG AND NORWICH CRAG OF SUFFOLK 

AND NORFOLK, ENGLAND 

Compare page 972; see figure 871 G, H 

The Red Crag is of greater antiquity as a whole than the 
Norwich Crag (Judd, 1911, p. 185; Osborn, 1922..563, p. 436); 
in it are recorded both Archidiskodon (?) planifrons and A. meridion- 
alis. The less richly mammaliferous Norwich Crag is partly more 
recent and yields ^4. meridionalis (Osborn, op. cil., p. 437). To- 
gether these horizons represent a very long period of closing Pli- 
ocene time, in which A. planifrons may have evolved into .4. 
meridionalis. Compare similar ascending mutations in Europe 
(Deperet) and in India (Osborn). 

Possibly belonging to Archidiskodon planifrons are the Red 
Crag teeth, 'broad plated with very thick enamel,' and the femur, 
referred to "E. antiquus" and "E. meridionalis" by Leith Adams 
(1877-1881, PL XXVI, figs. 2, 3, 4). 

In the Ipswich Museum (Osborn, 1922.563) is to be found 
a typical Upper Pliocene collection recorded from the Red Crag 
horizon; it contains si)ccimens referred to Castor {Trogontherium}, 
Orca (Trichecodon) huxleyi, Hyxnarctos, Felis pardoides, Ursus 
arvernensis, Mastodon [Anancus] arvernensis (27 specimens), 
Mastodon [Zygolophodon] borsoni (20 specimens), also two 10-11 
ridge-plated molars (?)erroneously referred to E. [ = Hesperolo.ro- 
don] antiquus and E. [ = Archidiskodon] meridionalis. Compare 
figure 871 G, H, Archidi.skodon (f) planifrons. 

In three British collections and in several faunal lists (Osborn, 
1922.563, pp. 436, 437) six grinding teeth referred to 'Elephas 
meridionali.'i' are attributed to the Red Crag or to the Norwich 
Crag of East Anglia, an Upper Pliocene stage; it is possible that 
certain of these grinders are referable rather to the more primitive 
Archidiskodon planifrons. See Falconer's notes (1868, Vol. II, pp. 
130-132, and PI. viii, figs. 1 and 4) on "Elephas {Loxodon) meri- 
dionalis.'" Falconer repeatedly compares these teeth to those of 
"E. (Loxodon) planifrons"; see full abstract below (pp. 972-974) 
of Falconer's notes of 1868. 

The finely preserved 10-11 ridge-plated molar in the Ipswich 
IMuscum (Fig. 8710 below) from the 'Norfolk Bone Bed,' re- 
ferred to 'Elephas meridionalis,' appears to Osborn to resemble 
Archidi.'ikodon planifrons, as indicated in the figure; the approximate 
measurements are, length 190 mm., breadth 77 mm., index 41; 
it exhibits ten to eleven ridge-plates only (cf. Amer. Mus. 19864^ — 
Fig. 839) ; it is greatly inferior in breadth to the more recent 
Forest Bed molar (Fig. 871 F) attributed to A. meridionalis 
(Savin Mus., Cromer, No. 197), to be described below under 
Archidiskodon meridionalis cromeren.sis, also to the Forest Bed 
molar (Fig. 871 E). The measurements and other characters of 
these molars resemble very closely those of the molars described 
by Deperet and Mayet in their Memoir of 1923 and figured in 
PL IX. 

Lydekker (1886.2, p. 113) provisionally refers to yl;v7(«//sfcorfo« 
meridionalis a mandibular symi^hysis from the Norwich Crag of 
Thorpe (Brit. Mus. M.2009); four ridges of a true molar, trans- 
versely cut and polished, from the Red Crag of Felixstowe, Suffolk 
(Brit. Mus. 44895) ; also fragment of a molar, vertically and 
longitudinally cut and polished, from the Red Crag of Fakenham, 



964 



OSBORN: THE PROBOSCIDEA 



Suffolk (Brit. Mus. 44140). These specimens are probably attrib- 
utable to A. planifrons. 

Osborn, 1929: Certain of the 'Norfolk Bone Bed' or Red 
Crag molars agree precisely in measurement (length, breadth, in- 
dex) and in the number of ridge-plates with certain of the Archi- 
diskodoti planifrons molars of the American Museum collection 
made by Barnum Brown, e.g., figure 839, found three mOes north- 
east of Mirzapur, Pinjor horizon, India, indicating a common 
eastward to westward range of A. planifrons from the Siwalik 
Hills to East Anglia in Upper Pliocene time. 



GEOGRAPHIC DISTRIBUTION OF ARCHIDISKODON PLANIFRONS 

(COMPARE SCHLESINGER, 1911, 1912, 1913, DEPERET AND 

MAYET, 1923, SABBA STEFANESCU, 1927, 

HOPWOOD, 1935) 

The migration area of Archidiskodon planifrons extends from 
India to western Europe, passing through southern Russia (the 
Caucasus, Crimea, Bessarabia) the basin of the Danube, near 
Vienna, Italy (Astesan, the superior and inferior Val d'Arno), 
France (Chagny, Seneze, le Puy, Randan) to England (Crag of 
Suffolk and Norfolk, Piltdown grav^els of Sussex) , finally in northern 
Africa (Algeria). In many of these localities the geologic and 
stratigraphic level is difficult to determine precisely. The type 
horizon of the Pinjor, Siwaliks, is regarded as between Middle and 
Upper Pliocene.' In Russia and Austria the formations containing 
A. planifrons are Pliocene, but it is difficult to exactly determine 
the level. In France and Italy it is possible to determine with great 
precision the horizon of Puy, of Seneze, and of Chagny, which are 
the exact equivalents of the classic horizon of Perrier (Auvergne) 
and constitute the most ancient phase of recent Pliocene time 
(Calabrian = Villafranchian) at the limit of the 'Pliocene ancien' 
(Astian). We may therefore consider that A . planifrons character- 
izes the horizon bounded by the 'Pliocene ancien' and the 'Pliocene 
recent,' a horizon somewhat in advance of that which contains 
A. nicrididrialis, a species In which A. planifrons is related by a con- 
tiniial series of inler mediate mutations. 

\\ third suju'rior molar found in the Lower Pleistocene of 
Shansi, China, has been compared by Dr. A. Tindell Ilopwood 
(1935.1, p. 88, PI. viii) to Archidiskodon planifrons. — Editor.] 

Ascending Mutations of A. planifrons in India and in 
Wkstkkn Eurasia (Osboun, 1928). — The above conclusion of 
Deperet and Mayet that Archidiskodon planifrons passes by a con- 
tinuous scries of ascending mutations into A. nieridionalis agrees 
entirely with that independently reached above by Osborn in the 
measurement and analysis (Tables VI and Vll) of the sixty-six 
specimens in the Falconer and Cautley and Barnum Brf)wn col- 
lections of the Pinjor horizon. Upper Siwaliks, India. The A. 
planifrons ridge formula; of the specimens of C'hagny, France, 
agree closely with the tyjiical A. planifrons ridge formula* of the 
sjjecimens from the Pinjor horizon, India. This was undoubtedly 
a case of indejjcndent contemporaneous progression in India and 
in France. 




Fig. 851. Type localities of Archidislcodon proplanifrons (22) and of A. 
planifrons (2) in circles. In solid black (2), Upper Pliocene' localities of A. 
planifrons referred specimens. In oblique lines, theoretic range from the sup- 
posed African center (22) northward to France and Britain, eastward to India. 
Compare Dcporet and Mayet (1923) who give the route of migration in the 
ojjposite direction. 



ARCHIDISKODON PLANIFRONS OF THE PILTDOWN GRAVELS 
(Figs. 852 and 853) 
The teeth figured (Fig. 852) and determined by Smith Wood- 
ward as "Stegodon," as pointed out by Freudcnberg (1915) and 
Matsumoto (1918), probably belong to Elephas [ = Archidiskodon] 
planifrons. These and other fossil mammal remains in the Pilt- 
down gravels may be divided geologically as follows : 

Pleistocene(?) Cervus elaphus ref. 
Ca.^lnr (?) fiber rof. 
II ippopotamus an>phibius (?) ref. 

Pliocene(?) Koanthropus dawsoni tyjjc. 

Archidiskodon planifrons ref., "Slegodon sp." of 

Smith Woodward. 
Ananciis arvrrnensis ref., "Mastodon" of Smith 

Woodward. 
Egiiiis slenonis (?) ref. 
Rhinoceros etruscus ref. 



'[Upper Pliocene to Lower Pleistocene (see Vol. I, fig. 413). Sec also footnote on p. 950 above. — Editor.] 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



965 



Authorities differ as to whether the Piltdown gravels, contain- 
ing Ennuthropus dawsoni, are of Upper Pliocene or of Pleistocene 
age, as shown in the following citations from Smith Woodward and 
Freudenberg : 

Smith Woodward (1913, p. 123).^"It is clear that this 
stratified gravel at Piltdown is of Pleistocene age, but that it con- 
tains, in its lowest stratum, animal remains derived from some 
destroyed Pliocene deposit probably situated not far away, and 






AnCHIUISKODON PLANIKKONS OF PiLTDOWN (SuSSEx) 

Fig. 852. Fragments of a molar tooth probably referable to Archidiskodon 
planifrnns. After Smith Woodward, 1913, PI. xxi, figs. 2, 2a and 3, 3a, 
natural size. Found with the remain.s of Eoaiilhropus dawsoni in the Piltdown 
gravel.s, Piltdown Common, Fletohing (Sussex), England. 

Vertical section of supposed portion of molar representing the valleys 
between the superior (2, 2a) sixth and seventh or seventh and eiglit ridge- 
plates and inferior (3, 3a) third and fourth ridge-plates (sec Figs. 853 to 855 C). 



consisting of worn and broken fragments. These were mixed with 
fragments of early Pleistocene mammalia in a better state of pres- 
ervation, and both forms were associated with the human skull 
and mandible, which show no more wear and tear than they might 
have received in situ." 

Freudenberg (1915, p. 420). — "An der Basis fandcn sich 
Schiidel und Unterkiefer von Eoanthropus Dawsoni in unzusam- 
menhiingenden Fragmenten. Im selben Lager zwei Zahnfrag- 
mente eines etwas gerollten Molaren von Elephas cf. planifrons 
(wohl nicht Stegodonl), wie solche schon friiher aus Red Crag von 
Leith Adams in A monograph on the British fossil Elephants, 



PI. 26, Fig. 3 und 4, kiirzlich aus dem Belvedereschotter von Wien 
durch Schlesinger abgebildet wurden. Ferner fanden sich mehrere 
Ziihne von Castor fiber, der nach Newton (The Vertebrata of the 
Pliocene deposits of Britain, PI. v fig. 16) gleichfalls im Red Crag 
vorkommt. Ferner ein Fragment von Mastodon arverncn.sisf, von 
Hippopotamus major und von Rhinoceros etruscus?" 

History. — (1) Freudenberg (1915, p. 420) pointed out that 
the molar crest fragments (Fig. 852) found in the Piltdown gravels 
with the type of Eoanthropus dawsoni resemble Archidiskodon 
planifrons rather than the "Stegodon sp." to which Smith Wood- 
ward referred them (Woodward, 1913, pp. 139, 142, 144, PI. xxi, 
figs. 2, 2a, 3, 3a). (2) Matsumoto observed (1918, p. 55): "But, 
judging from his [Woodward's] figures, the real reference of his 
'Stegodon sp.' appears to the present writer, as well as to Freuden- 
berg [Footnote: 'Neu. Jahrb. f. Min. Geol. u. Pal., Bd. I., Heft 3, 
1915: — Freudenberg, referating Woodward's papers on Eoanthro- 
pus dawsoni, has pointed out that. Woodward's 'Stegodon sp.' is not 
referred to genuine Stegodon but to Elephas cf. planifrons.'], to be 
otherwise than that stated by Woodward. Woodward's material 
consists of several fragments of molars, of which the ridges appear 
to be too high and too narrow and the valleys too deep and too 
narrow to be referred to the Stegodonts. Woodward has compared 
it with molars of E. meridionalis but not of E. planifrons. The 
present writer suspects that, if one compare Woodward's material 
with molars of E. planifrons and of the Stegodonts, one may easily 
recognise its closer resemblance to the former rather than to the 
latter." (3) As remarked by Matsumoto (op. cit., 1918, p. 56): 
"As to Woodward, his Mastodon sp. and 'Stegodon sp.,' of which 
association with these human remains may perhaps be secondary, 
have very probably been due to earlier — Tertiary — strata; and 
the former species is probably to be referred to M. arvernen.'iis. 
Now, the age of European E. cf. planifrons is included in that of 
M. ai-vernensis; so that it is highly possible, that these two species 
are found from one and the same deposit either primary or second- 
ary." (4) This determination by Freudenberg and Matsumoto is 
in accord with the discovery of Keid Moir that the flint imjilements 
found in the Upper Pliocene Red Crag of Foxhall are similar to 
those found in the Piltdown gravels. 

Conclusions as to Geologic Age (Osbohn, .Ii'ne, 1928). — 
Smith Woodward and other high English authorities (1913) re- 
garded the two probable Pliocene species, Archidiskodon planifrons 
ref. and Anancus arvernensis ref., as washed in from an older 
Pliocene deposit; whereas the Piltdown skull, Eoanthropus 
dawsoni, was regarded as of the same geologic age as the Piltdown 
gra\'els, an age which has not yet been positively determined, and 
which may also include Castor fiber ref.. Hippopotamus amphibius 
ref., and Cervus elaphus ref., properly belonging in the Pleistocene. 
Osborn observes : 

(1) The dark-colored skull fragments of Eoanthropus dawsoni 
are of Pliocene age, if washed in from the same geologic stratum as 
Arrhidi.^kodon planifrons and A. arvernensis. (2) On the other 
hand, these skull fragments are of Pleistocene age, if washed in 
from the same stratum as Hippopotamus amphibius and other 
Pleistocene fossils. On this point Hopwood inclines to Pliocene 
age; he writes (letter, June 4, 1928): 

"Secondly, I am inclined to put the skull with the older fauna, 
and the Eolithic culture. To put the skull with the older objects 



960 



OSBORN: THE PROBOSCIDEA 



presents some difficulties, but I think that there is much to be said 
in favour of such a step. First, the remains are of the same dark 
chocolate colour'" as the other fossils in the same group, whereas the 
younger fauna is represented by fragments of a light ochraceous 
colour. This is not very strong evidence I admit, but it would be 

strange to find one dark specimen where all the others are light 

The find of the nasal-bones and turbinals i)resents the most serious 
difficulty to be met. They were found lying together in the same 
spot. My own idea is that the skull had been rolling along the 



Crown yiew 



Longitudinal sec tion 
of two ridqe plates 




stream entire, and that for some reason or other the facial skeleton 
became detached. The brain-case was lying apart from the face 
and the whole lot was smashed [by the workmen] when the gravel 
was being dug. It is noteworthy that the nasals were found in the 
disturbed gravel, and this would account for the fact that the 
maxillse and other facial bones were missing. I do not pretend that 
this is the whole story, but it seems to me that these are little 
circumstantial details which help to strengthen the case for putting 
the skull with the older set of fossils." 

[The foregoing text on Archidixkodon jilauifrons of the Pilf- 
down gravels was written by Professor Usborn in June of 1928. 

In the autumn of 1933, at the request of Prof. Dr. Henri 
Delsol, Secretaire, Jubilo du Professeur Begoucn, Professor Osborn 
j)repared an article for the volume in celebration of the 70tli 
birthday of Comte Henri Begouen entitled "The Geologic Age of 
the Piltdown (Eoanthrnpus) and of the Trinil {Pilhecanlhropus) 
Man," (?publishcd)'- in which he advanced certain evidence in 
support of the new hypothesis that Eoanthropus is of closing 



Cross 
section 

AnCHlDISKODON PI-ANIKRONS OK PiLTOWN, EnGI.AND, AND Ol' THE PlNJOR HoUIZON, InDFA, EXIUBITINO AI'IMIO.XIMATELY THE SAME EnAMEL lilUGE-I'LATE 

Height, namely, 71 mm. Natural size. 




Fig. 853. Eneland. lldllcd frjiKiiK'nts of Archidiskmlon molars (Brit. 
Mils. E595 lower, E(i'22 upper), i)re,sumably portion.s of molar repre.sontiiiK 
llic valley.s between the .superior (B, Bl) sixth and seventh or .seventh and 
eiKlith ridge-plates and inferior (A, Al) third an<l fourth ridge-plates, from 
the Piltdown gravels, Sussex. Redrawn natural size by Mi.ss (!. M. Woodward, 
July, 1928. Longitudinal and ero.ss-seetions; erown vicw.s. Compare figures 
852 and 855. 



Fig. 854. India. Sixth, .seventh, an<l eighth superior ridge-plates of the 
Siwalik lectotype of lilrphns [— Arrhidiskodon] idanifrons reproduced natural 
size from drawing by Miss O. M. Woodward (Fig. S.").")D), for direct eonipari.son 
with the corresponding ridge-plates (Fig. 853) of A. phinifroitJS from the 
Piltdown gravels. 



'[Doctor Hopwood in a later palter ("Fo,ssil Elephants and Man," Proc. Ccol. Assoc, XIA'I, Pt. I, 193.'), |). J8) mak(\s the following statement: "Con- 
cerning the colour, reference should here be made to a recent paper by Sir Arthur Smith Woodward (1933) in which he .states that Mr. Dawson soaked the 
first pieces in a solution of bichromate of pota.sh to harden them! The remaining jiieces were not so treated and retain I heir original colour. This explains the 
very dark ehoi-olatc tone of parts of the brain-ca.se in contrast with the lighter, .slightly more grcyi.sh colour of the remainder, but it di)i's not afTect the statement 
that the colour of the human remains as a whole agrees with tliat of the fossils in Group A rather tlian with those in Group B. The fragments of Piltdown II. 
confirm this statement." — Editor.) 

-[Sec Bibliography under Osborn (193G.951), at close of this Volume. — Editor.) 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



967 



Tertiary age, while Pithecanthropus is of middle Quaternary age — 
evidence based chiefly on the ritige-plate and ganonietric method, 
which consists of counting the number of ridge-plates and esti- 
mating the enamel length of proboscidean molars. While Dr. 
Eugen Dubois adhered to his original opinion that Pitheranthropus 
is of Upper Pliocene or Lower Pleistocene age. Dr. W. O. Dietrich 
and Professor Osborn independently reached the conclusion that 
there is strong e\'idence for assigning to Pithecanthropus a more 
recent geologic age, namely. Middle Pleistocene, owing partly to 
the associated remains of Stegodoii airdwana and Elephas [Palscolox- 
odon] hysudrindicus in the Trinil beds. As to the Piltdown fossils 
and flints, Professor Osborn regarded these remanics, or worked 
over siiecimens, as having been washed in from older horizons, 
belonging in two groups, as clearly shown in the coloring of the 
\ery fragmentary fossils. 



Type 
Brif. Mas. M.3068 



- __, Ri^hf Second 
"^ 1 1 Superior Molar 




Line of wear 
.'® 6 P e 



e e 6/-r 'i 




The following is a citation from Professor Osborii's article 
(forwarded to Dr. Delsol in October, 1933), which, to the best of 
our knowledge, gives his final ojjinion on the subject in ciuestiou: 

"The conclusions which may now be drawn from the ridge 
]jlate and ganometric and from other palseontological methoils 
of com]jarison are as follows: 

"First, Pithecanthropus crectus, by association with Stegodun 
airdwana with a ganometric measure of 482 mm., also with an 



ARCHIDISKODON PLANIFRONS Amer. Mu5. f982l Ref. 



enamel 




'"XTpll ■il^;' ARCHIDISKODON PLANIFRONS 
Amer Mas /995I Ref 

Fig.855. Comparative sections of the lectotype and referred Archidistcodoti planifrons molars, with the A. planifrons molar fragments of Piltdown (solid black). 
Reduced to a uniform one-third scale for comparison with Falconer's figure (Fig. 825). 

\, Referred eleven ridge-plated r.Mj (Amer. Mus. 19798), 203 mm. in length, maximum 10th ridge-platc height 61 mm., a small and very primitive 
molar. 

B, Referred third right inferior molar, r.Ms (.\mer. Mas. 199-51), 313 mm. in length, maximum 9th ridge-plate height 113 mm., consequently a large and 
progressive lower molar stage. Exterior of the same tooth (Fig. 84.5). 

C, Second left superior liiolar, l.M'' (.\mer. Mus. 19821), 220 mm. in length, maximum 7th ridgr-plalc licight 74 mm., somewhat exceeding the 71 nun. 
ridge-jilate height in the Piltdown specimen. 

D, Falconer's type of Elephas [Archidiskodon] planifro7is (Brit. Mus. M.3068), an r.M', 221 mm. in length, maximum 7th ridge-plate height 71 mm. 
Redrawn by Miss G. M. Woodward for this Memoir (cf. Figs. 828 and 825). 



968 



OSBORN: THE PROBOSCIDEA 



Elephas hijsudritidicus Rrinder of an estimated measure of 5459 
mm., proves to be of Middle Pleistocene age." 

"Second, by association with Arrhidiskodon planifrotis with 
eleven ridge jjlates and an enamel length of 1113 mm., Eoanlhropus 
dawsoni of the Piltdown gra\els may pro\e to be of Upper Pliocene 
age." 

"Third, thus, by the application of this ridge plate and gano- 
metric method these two famous fossil men change places, as 
follows: 

"The supposed oldest fossil man, Pithecanthropus, is dethroned 
and becomes a survival; Eoanlhropus is enthroned as the oldest 
fossil man knoivn up to the present time." — Editor.] 



Sci., Paris, Tome 179, p. 1418, December 15, 1924. Type.— 

Fragment of a third inferior molar of the left side, I.M3, in the 
Laboratory of Geology, University of Bucharest. Hohizon 

AND Locality. — Upper Pliocene. Tulucesti (Covurlui), Rumania. 
Type Figure. — Sabba Stefanescu, 1927; originally figured by 
Sava Athanasiu in 1912 [1915], PI. xvii, fig. 4, as Elephas cfr. 
mfridionalis. 

Type Description.— (S. Stefanescu, 1924, p. 1418, and 1927): 
"A I'exception de Maria Pavlow . . qui a rapporte a VElephas 
af. plauifrnns une molaire qu'elle a regue de Ferladany (Farladeni) 
en Bessarabie, personne a ma connaissance n'a indique jusqu'a 
present cette espece dans les couches geologiques de Roumanie. Or 




Fio. 856. Scene on the Ancient Riveh Ouse Illu.stkating the Osborn Theory of the Upper Pliocene Age ok Eo.^nthropus uawsoni 

Restoration by Margrct Flin.sch in 1934, under the direction of Henry Fairfield Osborn 
Eoanthrnpna (/(/»woni Woodward, a.s restored from the cranium and jaw discovered in the Piltdown gravels dej)ositcd in tlu^ cluiniic!l of tJie ancient rivoi' 
Ou.se. Bodily proportion.s and outlines from a Bushman hunter (.j feet) of tlie Kalahari Desert, We.st Africa, reduced in present restoration to a scale of 
1 "JOth natural size. The erect figure is holding tlic sliar|)ened femur of Anaricus or of Arckidinkixloii. discovered in flic Piltdown gravels. Osborn does not 
agree with his friend Henri Breuil tliat this sharpened bone is the work of a rodent. 

The mammals are (left foreground) ArchiiHxknilim plntiijrons, l/;59tli natural size, (left rear across the riv<'r) the straight-tusked mastodon, Anancus 
arvernensis, and Archidiskodon planifrons, 1 /100th luitural size, (right bank) herd o! JCqims stoionis, 1/lGOth natural size. 



ARCHTDISKODOX PLANIFRONS OF RUMANIA 
Archidiskodon planifrons rumanus S. Stefanescu, 1924 

Figure ,S.")7 
From Tulucesti (Covurlui), Rumania. Upper Plioceue. 
Elephas antiqiuis rumanus Stefanescu, 1924. "Sin- la i)rcsence 
de VElephas planifrons et de trois mutations de VElephas uutiquus 
dans les couches geologiques de Roumanie." ( 'ompt . Hend. Acad. 



l)lusieurs molaires et ([uelques nioitios de mandibules d'elephants 
de ma collection jin'senfeiit d(>s cnraclcn^s tjui rMi)])elI('iit de tres 
|)r('s cetix de V Elephas planifrons. Par consctiuent, je puis afhrnier 
(|ue cette espece est non seulement rei)resentcc mais (lu'clle est 
I'utie (les plus rcpandues dans les couches pliocenes et pleistocenes 
de notre pays. Mais n'ayant jias rcHUieilli nioi-iuenie ces fossiles, il 
lue ])arait hasardeux de pr^ciser leur age geologique. . . . D'autre 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



969 



part, vu la grande variabilite des caracteres specifiques de VElephas 
planifrons, il est absolument necessaire de ie\iser ces caracteres 
afin lie conclure, s'il y a lieu, a I'existence d'une seule ou de 
plusieurs mutations, ou meme a I'existence d'une seule ou de plu- 
sieurs especes, confondus ensemble par les paleontologistes qui 
m'ont precede. Dans ce but je i^ratiquerai ?na methode de recherchefi 
basee sur I'organisation des lames isolees des molaires, methode 
naturelle ciui tout recemment m'a permis de reconnaitre que la 
molaire provenant de Tulueesti (C'ovurlui) et rapportee par Sava 
Athanasiu [Footnote: 'Mamifere pliocene de la Tiducesii, districtul 
Covurhii, 1915.'] k VElephas cf. meridionaUs, doit etre attribuee 



a une mutation ancestrale de VElephits (uili<iuus, que j'ai ilononuiiee 
rinnanus." 

"Je n'insiste pas pour le moment sur les caracteres de cette 
mutation qui, a mon avis, est la plus rapprochee de I'origine 
mastodontide de I'espece anliquus. J'ajoute seulement que la mu- 
tation ausonius de I'ltalie lui succede et (jue la mutation de Weimar 
et Taubach (Allemagne), que je designe sous le nom germanicus, 
est la plus recente." 

"Toutes ces trois mutations ont ete trouvees en Koumanie: 
rumanus a Tulueesti (C'ovurlui), avsonius a Colintina (Ilfov), 
germanicus a Tanganu (Ilfov)." 





Type and Referred Inferior Molars of Archidiskodon planifrons rumanus 
Fig. 857. (Left) Type third left inferior molar, I.M3, of Elephas anliquus rumanus S. Stef3,nescu, 1924; 
originally figured by Sava Athanasiu, in 1912 [1915], PI. xvii, fig. 4, as "i'fepfeis efr. »nen'(HonaZis"; refigured 
by StefSnescu in 1927 as Elephas anliquus rumanus. Original in the Laboratory of Geology, University of 
Bucharest. (Right) Last two ridge-plates and talon of a referred third right inferior molar, r.Ms, stripped of 
its very thick cement (ColL of the Laboratory of Palaeontology, University of Bucharest). After Stefanescu, 
1927. 

The present author considers that tliis type is more closely related to Archidiskodon owing to tlie com- 
paratively low ridge-plates and the excessively wide cement areas between them. StefSnescu rightly attributes 
this specimen to Elephas [ = .-irchidiskodon] planifrons. He states (1924, p. 1418): "Or plusieurs molaires et 
quelques moitiiis de mandibules d'alephants de ma collection pr(5sentent des caracteres qui rappellent de tres 
pros ceu.x de VElephas planifrons. Par consequent, je puis affirmer que cette espece est non seulement repre- 
sentee mais qu'elle est I'une des plus repandues dans les couches pliocenes et pleistocenes de notre pays." 



ARCHIDISKODON MERIDIONALIS AND A. MERIDIONALIS CROMERENSIS 



Archidiskodon meridionalis Nesti, 1825 
Figures 815, 817, 858-869, 925-928, 1239, PI. xxi 

Upper Pliocene' to Lower Pleistocene of Val d'Arno suporieure, northern 
Italy, of Saint-Prest, France, and of England (Forest Bed). 

Syn.: Elephas giganteus Aymard (MS. in Lartet, 1857); Elephas lyrodon 
Weithofer, 1890, ba.sed on a female type skull. 

Early Discoveries of Elephas meridionalis by Tozzetti 
IN THE Val d'Arno. — Giovanni Targioni Tozzetti, who in 1725- 
1745 traveled through the greater part of Tuscany, described 
especially the geology of the valley of the Arno, making a sharp 
distinction between the Lower (Val d'Arno inferieure) and the 
Upper (Val d'Arno superieure). The 'Val d'Arno inferieure' he 
regarded as a sea-water deposit, which formed the hills and filled 
them with marine fossils; these fossils occur in abundance from 
Capraja to the sea. The 'Val d'Arno superieure,' with its hills of 
chalk and ochreous clay, weathered by sand and ice, shows no 
trace of marine animals; these deposits, Tozzetti believes, were 

'Possibly Lower Pleistocene (see footnote on p. 1049). — Editor. 



attributable to the fluviatOe waters of the Arno, which at some 
ancient period poured forth from natural cataracts having their 
source in the mountains between Incisa and Rignano. If it is true, 
howe\'er, that ribs of Fiseter [Physeler] were found two miles from 
Arezzo, then there is reason to believe that the deposits of the Val 
d'Arno superieure were not of fluviatile origin. 

Falconer in the "Palaontological Memoirs" of 1868, Vol. II, 
p. 104, remarks that "The 'Val d'Arno Superiore' has, from remote 
ages, been celebrated for the vast abundance of fossil remains 
found there. Huge bones and teeth of Elephants were especially 
numerous. A large collection of these was formed by Targioni 
Toretti [Tozzetti], which ultimately found its way into the Grand 
Ducal Museum at Florence; and numerous additions were made 
by Nesti, who, in 1808, soon after the publication of Cuvier's 
'Memoir of the Mammoth' (Annales du Museum, torn, viii.), ex- 
amined the Tuscan Elephantine remains, and was so satisfied of 
their difference from those of the Mammoth, that he proposed for 



970 



OSBORN: THE PROBOSCIDEA 



tliem two specific flesignalions, namely, Elephas meridionalis and 
E. minutus (or mininuts].^ Influcnceil by the fact that Ciuier luid 
laid so much stress ujion the peculiar form of the lower jaw, and 
guttered beak of the symphysis, as distinctive marks of E. prinii- 
genius, Nesti (not a professed anatomist) was naturally led to 
direct his attention, in the first instance, chiefly to the same parts 
in the Val d'Arno remains. Unluckily the specimen that presented 
the most pronounced beak had lost its molar teeth [Nesti, 1808, 
Tav. I, figs. I and ii]; Nesti assumed it to be of an Elephant. But 
this selected 'piece justificative' for his Elephas meridionalis was 
l)roved by ( 'uvier to be the lower jaw of Mastodon Anwrncnsifi, . . . 
and E. iiinndiis to be merely a young Elephant." 



of sjiecific distinction. . . . This singular conclusion is, in some 
measure, explained by the fact that hardly a specimen of a molar 
of the true Mammoth exists in the Florentine Museum for com- 
liarison. It is, perhajts, still more remarkable that the ex])prienced 
eye of ("uvier should have glanced o\er the multifarious evidence 
sujiphed by the Tuscan collections, without being convinced that 
E. meridionalis was a well-founded species, considering the ra])idity 
with which he seized, anil the logical precision with which he char- 
acterized, the distinctive marks of the Mammoth from the existing 
Indian Eleijhant." 

"Failing the teeth, Nesti drew his specific distinctions from the 
form of tlic ci'aiiiuin and lower jaw. Ample evitlence is afforded by 





Lectotype Cii.iNinM (C) of Archidiskodon meridionalis, Museum of I'lukencb 
Fig. 858. Lectotype figure after Nesti of the first described mid figured cranium (Cranium C) of Elephas meridional^ Nesti, 182.5, Tav. i, figs. 1 and 2; 
cited by Falconer (1868, Vol. II, p. 122) and selected as the Ujpe by Deporct and Mayet (1923, p. 12G). Oiie-twelftli natural size. Compare figure 861 of the 
present Memoir, also figure 865 (13). 



"After a long interval, during which ('uvier had visited the 
Tuscan collections, Nesti brought out another memoir upon the 
subject |1825|, in which, ujjon greatly extended obser\ations on 
s|)ecimens of all ages, from the foetus upwards, including crania, 
lower jaws, molars, tusks, and bones of the extremities, he upheld 
the soundness of his first inference in regard to the distinctness of 
E. meridionalis, while he admits tacitly the force of Cuvier's 
criticism upon his second species, E. minulus. . . . Another circum- 
stance, which materially damaged the authority of Nesti upon 
a question of such difficulty and imi)ortance, is that he states that, 
after examining a vast number of molars of all ages, he had found 
them to vary so much — some having thick i)lates, others thin, 
and the same tooth presenting such different {jatterns, according to 
its age and degree of wear — that he had abandoned the characters 
yielded by the molar teeth as worthless ( !) for any reliable marks 



them for establishing E. meridionalis as an independent form." 

"The Abbe Croizet, to whom palaeontology is indebted for so 
much \-aluable research on the fossil fauna of Velay, was the first 
who had the courage to question the decision of ("uvier against 
E. meridionalis. In his work upon Puy-de-D6rae, he has figured 
and described a fragment of an ui)per (?) molar (lower left of 
Croizet and .lobert) discovered at Malbattu. ... He refers to 
Nesti's researches, and sums u]) by inferring that, as there are 
two living Elephants, so there were two fossil species — the one with 
attenuated plates, being the Mammoth of Siberia, the other with 
thick plates, as seen in specimens from . . . the \'al d'.\rno. He 
considered the facts sufficient, but a.ssigned no other name to the 
second species than that of 'Ek'])hant de Malbattu.' " 

Geologic Level, Fouest Bed oh ("komerian (compake 
OsBORN, 1922.563, pp. 439, 440.— (1) The survival in the Forest 



'[Re-searches of the present author failed to substantiate the assignment by Nesti of the names Elephas minutus or E. minimus in either of his articles of 
1808 or 1825. Weithofer (1890.1, p. 134) attributes both these names to Falconer, i.e., Elephas minulus (Pal. Mem., 1868, V'ol. II, p. 104) and E. minimus 
(1846, letterpress, "Fauna .Anticjua Sivalensis," p. 13). —Editor.] 



THE MAMMONTINiE: ARCHIDISKODON AND METARCHIDISKODON 



971 



Bed of East Anglia, above the Red and Norwich Crags {Archidisko- 
don planifrons level) of warm temperate types, such as specimens 
referred to Arrhidiskodon meridionalis, Pala'olojodon [Hesperoloxo- 
don] (uiti(/uus, Hyxi/a tttn'otci, Dicerorhinus etrufscuH, Equus stenotiis, 
Machxrodus, is exactly paralleled by the same genera and species 
occurring in southern France and northern Italy (Val d'Arno) 




Specific ( 'haracters of Archidiskodon meridionalis. — As 
compared with Elephati [ = Archidisko(l(in\ pUuiifnnts, cranium of 
Arrhidiskodon meridionalis larger, more lofty or acrocephalic; 
frontonasal contour ty])ically concave instead of phiiw (Fig. 861) ; 
fronto-oceipital crest broadly truncated (Fig. 865) not rising to an 
acute a])ex as in ^4. imperator. Maximum ridgc-jjlate formula: 
Dp 4| M 1/+ M 2|f| M 3j-j.T3 a^^ compared (p. 933) 
with the typical ridge-plate formula of A. planifrons 
(after Falconer) of Dp 3 §+ Dp 4 y| M 1?^ M 2f 
M 3 ij^. Crinding teeth broad, heavily plated with 
cement; superior incisive tusks large and lyrate. 
Another progressive character in A. nwridionalis is the 
loss of true premolar (P 3-4) eruption. In brief, the 








Young Male Cotype Cranium (A) of Archidiskodon meridionalis Ne.sti, Museum of Florence 



Fig. 859. Cranium A of Nesti, cotyjic of Elephas meridionalis Nesti, 
182.5, Tav. i, fig. 3, one-eighth natural size. Figure reproduced from a plate 
which also gives the profile (fig. 2) and front view (fig. 1) of the lectotype 
specimen (Cranium C). 

After this cranium (A) was figured, as above, the incisive alveoli were 
added, as shown in figure 860 oppo.site. Falconer (1868, Vol. II, p. 122) 
observed that enormous tusks have also been atlded, yielding a diameter of 
0.26 m. or 10.2 inches. 



Fig. 860. Side view of the restored cotype skull of Elephas niendionalis 
in the Florence Museum. After Weithofer, 1890, Taf. i, fig. 2: "Elephas 
meridionalis Nesti; Cranium A; oberes Arnothal; von links." Reproduced 
about one-tenth natural size, for direct comparison with figure 859, which is 
one-eighth natural size, .showing that the peak or apex of the cranium is 
greatly elevated. Consequently the frontojiarietal plane is much more 
elongate than that in E. planifrons. 



during the long warm 1st Inlerglacial period. (2) During this 
period there also appear for the first time in Great Britain certain 
African types, like the Hippopotamus, and there became more 
abundant in Great Britain the loxodontinc type, Palseoloxodon 
[Hesperoloxodon] antiquus, as well as the Hyxna. (3) The Forest 
Bed arri\'al of tundra and northern forest types, such as Mani- 
monleus primigenius, Ovibos inoschatus, Aires latifrons, is a dis- 
tinctive feature of the northern latitude and cold climate of East 
Anglia during the period of the first Scandinavian glaciation, 
which has no ])arallel in southern France or in northern Italy. 



typical A. meridionalis of the Ipper Pliocene, lowermost Pleis- 
tocene, and Lower Pleistocene is much more progressive than the 
typical A. planifrons. 

This is the great 'southern mammoth,' well named E. meri- 
dionalis by Nesti in 1825. It is connected with Elephas planifrons 
by a series of ascending mutations, completely confirming Fal- 
coner's observations of 1863, p. 80 : "The nearest affinity, and that 
a very close one, of the European Elephas meridionalis is with the 
Miocene E. (Loxod.) planifrons of India." Nesti's type is a fine 
skull still preserved in the Florence Museum. 



972 



OSBORN: THE PROBOSCIDEA 



El. [Elfphns] meri(li.»ial/s Nesti, 1808, 1825. "Di Alcvine 
Ossa Fossili di Maiiiinifeii chc s'incontnvno nel Valdarno," Ann. 
Mils. Imp. di Fisica e Storia Nat. Firenze, 1808, I (description 
without name); "Sulla nuova specie di elefante fnssile del 
\'aldaino" (letter from Nesti to Dott. Prof. Ottaviano Targioni 
Tozzetti). Nuov. Giorn. Lett., 1825, XI, No. 
24, p. 211. Lectotype. — Cranium with 

third superior grinding teeth, the rostrum and tusks 
being broiven away ( = Cranium C of Nesti) . Cotype 
( = Cranium \ of Nesti). Horizon and Locali- 

ty. — Sui)eri()r Val d'Arno, Upper Pliocene, northern 
Italy. Lectotype Figure. — Op. rit., 1825, Tav. 
I, figs. 1 and 2. Cotype, Tav. i, fig. 3. 

Description. — In his description of 1825 Nesti 
compares the skull with that of "Mastodonte" and 
of "primigenia," concluding (p. 211): "Potrebbe a 
questa Specie imjjorsi il nome di Elephans Valdar- 
nen.si.s, o Etruscus, o anco Italicus, ma poiche le 
regioni, nelle quali questo animale viveva, non sono 
note, e d'altronde pare che fosse destinato a climi piu 
temperati e meridionali della Specie primigenia, pre- 
ferisco di appellarlo El. tneridionalis." There is an 
earlier reference to this fossil which does not in- 
clude the name. It is to be found in the Annali 
del Museo Imperiale di Fisica e Storia Naturale, 
Firenze, Tome I, 1808, mentioned above with title 
of the article. 

Lectotype Skull. — Cranium C' of Nesti's description was 
selected by Deperet and Mayet (1923, p. 126) as the type; see also 
Falconer (1868, Vol. II, p. 122). It is here reproduced from Nesti's 
original figures, front and side views, figs. 1 and 2 of his Tav. i 
(Fig. 858 of the present Memoir). Falconer {op. cit., p. 122) de- 
scribed Cranium C as follows: "It is nearly perfect in the frontal 
and occipital regions, condyles, maxillaries, and molars, but im- 
perfect in the facial portion, the border of the nasal opening being 
broken, together with the terminal portion of the incisive alveoli 
and the zygomatic arches. Since Nesti's figures were taken, this 
specimen has suffered considerable damage, the upper lamina of 
the right incisive alveolus having disappeared, together with the 
salient tip of the nasals and the lateral margin including the left 
orbit. The last molar is present on either side, far advanced in 
wear. (See PI. i. fig. 11, and PL ii. fig. 16.)" 

Cotype Skull. — Falconer also described the cotype Cranium 
A as follows (op. cit., p. 122): "4. The cranium A of Nesti's 
references, fig. 3, comjjrising the ])alatine, maxillary, and temporal 
regions, the inferior part of the occiput, and the zygomatic arches, 
the only deficiency being in the facial region. The specimen, 
which is highly ferruginous, has now joined on to it the entire 
incisive sheaths (not rei)resented in Nesti's figure) and two enor- 
mous tusks, which are spread out horizontally in the Theristocau- 
/of/o/i-manner above noticed. Nesti, in his memoir, cites the tusks 
of this specimen as yielding a diameter of 0.26'", or 10.2 inches. 
The last molar, much worn, is present on either side." 



FALCONER'S NOTES OF I8C8 ON TYPE AND REFERRED SPECI- 
MENS OF ELEPHAS [ARCHIDISKODON] MERIDIONALIS [AND 
?PLANIFRONS] OF THE NORWICH CRAG AND VAL D'ARNO 
Falconer, "PaliEontological Memoirs," Vol. I, 1868, pp. 443-447, 
Plate xiv.b' of the "Fauna Antiqua Sivalonsis" 

Falconer's observations collected in his "Palaeontological 
Memoirs" extended from the year 1863, in 
which he first published a ridge formula of the 
species Elephas mer/dioiialis, up to the year 
1868, his final observations being embodieil 
by Murchison in the "Palaeontological 
Memoirs" of that date. He closely 
i compared specimens from the Nor- 
wich Crag (in the Norwich Mu- 
seum), and from the Val 




Aged Male. Lectotype Skull (C) ok Archidiskodon mehidionalis 
Nesti, Museum of Florence 

Fig. 861. Lectotype .skull of Elephas meridionalis Nesti, as refiguretl by 
Weithofer in 1890, Taf. ii, fig. 1: "Elephax mrridionalis Nesti; Cranium C; 
oberes Arnothal; von rechts." One-seventh natural size. Compare figures 
8.58 and 86.T (7, 13, 14) of the present Memoir. 

An outline of the same skull, after Falconer's plate of 1847, appears in 
figure 86.5 of the present Memoir. All these figures agree in the peculiarly 
flattened appearance of the summit of the aged occiput, which is quite unlike 
the acrocephalic occiput of the juvenile Archidiskodon imperalor. 

d'Arno (in the Museum of Florence). In his ridge formula he 
undoubtedly confused ascending mutations of specimens belong- 
ing to Arrhidiiihodon pkniifrmif: with specimens truly representa- 
tive of ])rimiti\-e Archidiskodon meridionali.'i. 

Compare above with the ridge formulae and measurements of 
Archidiskodon plniiifro?!.^. The ridge formula; below gi\-en by 
Falconer and others embrace the A. planifrons stage as well as the 
A . meridionalis stage. 

'[A footnote by the Editor of the Palffiontological Memoirs on pp. 443, 444 of Volume 1, states that according to the corrected copy by Falconer in the 
Briti.sh Museum, all the figures in PI. xiv.b, except 10, 17, and 18, should belong to E. antiquus; that the correction, however, is incompatible with the descrip- 
tion and identification of every figure in pi. xiv.H, given in a subsequent part of the same memoir, according to which every figure in the ])late, with the except- 
ion of 16, belongs to E. meridionalis. — Editor.] 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



973 



Elephas meridionalis Nesti, 1825. Lectotype. — Cranium C, 
Tav. I, figs. 1 and 2. Cotype. — Cranium A, Tav. i, fig. 3. 

Upper Jaws. — Plate xiv.b, figs. 1, la, Dp^, ridge-plates K-6-K, 
"the crown is composed of six principal ridges, besides front and 
back talons. . . . compared with the corresponding tooth of E. 
(Loxodon) planifrons, which it resembles very closely, ... it has 
a broader crown"; figs. 2, 2a, r. Dp', ridge-plates 6, Norwich Crag. 

Lower Jaws. — Plate xiv.b,' figs. 3, 3o, lower jaw, r.Dps, 
ridge-plates YrQ-Yi, very broad in the crown relative to the length, 
discs of ridges very wide as in Italian specimens, Norwich Museum; 
figs. 4, 4a, lower jaw, l.Dp4, 8 ridge-plates, Norwich Museum, the 
ridge formula in these specimens agrees with the Italian, also the 
broad crown, the low ridges, and thick plates of enamel; figs. 5, 
5a, lower jaw, l.Mi, ridge-plates K-8-/2, eight principal ridges, with 



front and back talons, discs of first three wide and open, tendency 
to mesial expansion, crown low in reference to breadth; figs. 6, 
6a, lower jaw, l.Mi, ridge-plates %-8, the crown presents a front 
talon and eight ridges, all of them worn, discs wide and open, 
Norwich Museum; figs. 7, 7a, lower jaw, r.Ms, ridge-plates 7 + , 
Norwich Museum; figs. 10, 10a, lower jaw from Val d'Arno, 
demonstrating how exactly the English specimens agree with the 
Italian, long symphysis, gradual incUnation into a beak; figs. 
17, 17a, r.Ms, from Val d'Arno, showing 13 ridge-plates, length 
10 in. =255 mm., width 3.4 in. =87 mm.; figs. 18, 18a, Norwich, 
lower jaw, r.Ms, ridge-plates K^ll-K, "showing eleven principal 
ridges, an anterior talon, and [?] a back talon," cement decomposed 
or denuded, 4 thick denticles on 6th to 8th ridges, length 11.25 in. 
= 287 mm. 




-/^ /<fa^. 




FCg:M. 




Fig. 862. Molars from the Val d'Arno (left) and the Norwich Crag (right) referred to 'Elephas' [Archidiskodon] meridionalis by Falconer 

"Fauna Antiqua Sivalensis," Plate xiv.b, figs. 17, 17a, 18, I80, one-third natural size 
Compare "Palsontological Memoirs," Vol. I, 1868, pp. 443, 444, 447, 448 



Figs. 17, 17a. Third inferior molar of the right .side, r.Ms, Val d'Arno. 
Original in Oxford Museum, one-third natural size. 

(Op. cit., p. 447): "Figs. 17 and 17a. — Elephas meridionalis. A Val 
d'Arno lower molar of the same age, from Dr. Buckland's collection in the 
Oxford Museum, crown side. (Reproduced in Plate viii of vol. ii.) Length of 
crown, 10. in. Width of crown, 3.4 in. Height of crown, 5. in." 

Note I. — Falconer (as explained in the "Palseontological Memoirs," 
1868, Vol. I, pp. 443, 444), in lettering the plates of the "Fauna Antiqua 
Sivalensis," became satisfied at the time the plates were lettered that he had 
interchanged Elephas meridionalis and E. aniiquus. In his Memoir of 1857 
(not published until after his death) he concluded [Pt. II of his Memoir 
publi.shed in 1865, p. 281]: "I beg leave to explain now, that all the plates 
bearing the name of E. meridionalis in the 'Fauna Antiqua Sivalensis,' includ- 
ing the outline-figures of crania in plate 42, belong to E. aniiquus, while those 
that bear the latter name belong to E. (Loxodon) meridionalis. In the descrip- 
tions which follow, they will be cited as such." 

'[See footnote on opposite page. — Editor.] 



Figs. 18, 18a. Third inferior molar of the right side, r.Ma, "Mam- 
maliferous Crag" (Norwich). Norwich Museum 1570, one-third natural size. 

{Op. cit., p. 447): "Figs. 18 and 18a. — Elephas meridio7ialis. The finest 
detached molar of this species that has come under my ob.servation is a speci- 
men which was discovered in the 'Mammaliferous Crag' on the Thorpe Road, 
near Norwich, by Mr. Prestwich. ... It is now lodged in the Museum at 
Norwich, and is the specimen which first convinced me many years ago that 
the 'Crag' yielded a species of Elephant entirely distinct from the Mammoth 
and from E. aniiquus. . . . Extreme length of crown, 11.25 in. Width of crown 
in front, 3.3 in. Width at fifth ridge, where the crown is broadest, 3.8 in. 
Extreme height of ridge, 4.8 in. Width of ninth ridge, 3.5 in. Height of ninth 
ridge, 4.6 in." Reproduced in the 'Pateontological Memoirs,' Vol. II, PI. 

VIII. 

Note II. — Compare these i)rimitive .4. [planifrons] meridionalis molars 
of Italy and England with the more progressive A. planifrons molars of the 
Siwaliks, India. 



974 



OSRORN: THE PROBOSCIDEA 



Falconer, 1868, E. meridionalis: Dp 3^^^" Dp 4^ 
M 1 vi-e-ti M 2 7 4-9 M 3 is-i i-vi- 

The above ridge formula is too low, obviously because 
Falconer included within his 'Elephas' tneridimialis grinding teeth 
belonging to Archidiskodon plaiiifrons. IMore distinctive of the 
species A. meridionalis, in Osborn's opinion, is the ridge formula 
which we may deduce from Falconer's own observations (1868, 
Vol. II, p. 1 18) on nine Itahan crania in the Florence Museum : 

Falconer (Italy), E. meridionalis: Dp 2f Dp 3f Dp 4t 
M 1 /^ M 2 If: M 3 T^J^. 

Weithofer (1890, p. 173) observes that Falconer's high figure, 
M 3 Yi, rests on a single indi\idual. 

Leith Adams, 1877-1881 
Formula. — Leith Adams reviewed this important species in 
his "British Fossil Elephants," 1877-1881, and while in general 
confirming Falconer's observations by excluding E. plaiiifrons, he 
reached a nuich more accurate dental formula which became the 
standard : 

Leith Adams, 1877-1881, E. meridionalis: Dp 2f Dp 3t 
Dp4fMlf^M2|i}iM3|Hi. 

This formula of 1877-1881, together with Leith Adams' sum- 
mary of the points distinguishing the molars of Elephas ineridionalis 
from those of the narrow-toothed Elephas antiquus (p. 232), are to 
be found in his important Memoir (p. 208). He assigns the follow- 
ing characters to Elephas meridionalis. 

Characters. — (Cf. op. cit., p. 30): M^ in E. meridionalis, 
ridges nearly as broad as they are long, thick plates, grosser masses 
of intervening cement, machaerides uncrimped, X 10 X . (Cf . p. 
44) : ]\P massive, enamel and plates very thick, machaerides 
scarcely plaited, great breadth and low ridge formula, which rarely 
if ever exceeds that of M' of E. antiquus. (Cf. p. 48) : In M= the 
highest ridge formula of E. meridionalis equals lowest of E. anti- 
quus; so that in number of ridges we find E. primigenius, E. 
antiquus, E. meridionalis, and E. namadicus meeting at their 
extremes. Leith Adams (cf. pp. 129 to 144) compares the cranium 
of E. meridionalis in great detail with that of E. planifrons, E. 
hysudricus, E. bombifrons, E. africanus, etc., concluding that the 
skull and dentition of E. planifrons make the nearest approach to 
E. meridionalis (see pp. 186, 208-210, 239, 244). 

Error. — Originally Leith Adams also made the error of 
including the low ridge-plate formula (M 3 \^) of E. planifrons 
with the high ridge-plate formula (M 3 Hrfj) of E. meridionalis, 
in describing the range of evolution and variation in the ridge 
formula of the British specimens of E. meridionalis. 



Weithofer, 1890, Lydekker, 1886 
Collective Ridge Formula. — The collective ridge-plate 
formula [of A. planifrons and A. meridionalis] may also be cited 
from Weithofer (1890, p. 172) : "Als (lesammtformel fiir die Ziihne 
des El. meridionalis, soweit sie hauptsiichlich aus dem Material 
des Museums zu Florenz resultirt, ergabe sich demnach": 

Weithofer, 1890, E. meridionalis: Dp 2 7,^3 Dp 3 '.'"" 
Dp4^MlHM2f^jM3ii^^ 



Weithofer in his collective formula, like Falconer (1868), un- 
doubtedly erred by including specimens having the typical Arcki- 
di.^kodon planifrons ridge formula with those of A. meridionalis; 
thus this collective formula ranges from the formulae of Archidis- 
kodon planifrons, e.g., M 3 tt, to A. meridionalis, e.g., M 3 tI. 

Lydekker, 1886. — Lydekker in his "Catalogue of the Fossil 
Mammalia in the British Museum (Natural History)," 1886.2, p. 
107, adopts the Leith Adams formula of 1877-1881 as follows: 
(1) The molars of E. meridionalis so closely resemble those of E. 
planifrons, that if they both occurred in the same area it is more 
than doubtful if they could be specifically distinguished; (2) both 
frequently exhibit partial denudation of the enamel ridges; (3) 
there is variation in the thickness of the enamel and in the breadth 
of the ridges, in some molars the enamel being relatively thin ami 
considerably plicated; (4) the cranium is characterized by large, 
slightly curved tusks and widely diverging alveoli; (5) in general 
contour it is intermediate between the cranium of E. planifrons 
and E. hysudricus, although nearer the latter; (6) it agrees with the 
cranium of E. planifrons in the relative distance between the 
nasals and the vertex, but has the vertex more vaulted, the 
frontal profile concave, the temporal fossae intruding largely on the 
frontal aspect; (7) the species attained an enormous size, the 
height of some individuals being estimated at upwards of fifteen 
feet. 




Archidiskodon meridionalis Ref. 
Fig. 863. Rpferred Elephas meridionalis of Cliagny (Cote-d'Or), one- 
fourtli natural size, after Gaudry, 1878, p. 178, fig. 237. Tlii.s fourteen ridged 
tliird .superior molar, M^, of Chagny, is to be compared with tlie tliirteeii 
ridged molar stage of Eh'phas ineridionalis of the Val d'.\rno, i.e., M 3 ,^J -, a.s 
dedueed by Falconer (see Falconer, 18G8, Vol. II, p. 118). Observe superior 
ridge-j)late.s ])rc-concave, post-convex (ef. Fig. 827, .1. planifrons). 



Cranial and Dental Characters. — Weithofer (1890, jip. 
136, 137) refigurcs Nesti's type skull (Fig. 861 of the present 
Memoir) showing the right side (Taf. ii, fig. 1): "Elephas meri- 
dionalis Nesti; Cranium C; oberes Arnothal; von rechts." 
Falconer also rofigures this skull (1846 [1847, figs, xix of Pis. 
XLii and XLiv|), as well as Depi'ret and Mayet (1923, p. 128, fig. 
16). The three skulls in the Florence Museum exhibit the follow- 
ing characters: (1) Pointed nasals, (2) concave frontals, (3) high 
o('cii)ilal crest flattened anteroposteriorly, (4) extreme hrachy- 
cephalic cranium shortened anteroposteriorly, (5) broad narial 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



975 



openings, (6) parallel sides of the premaxillary sockets of the tusks, 
as contrasted with the broadly flaring sides of the sockets similar 
to those in the contemporary Hesperohu-odon antiquus. All 
these characters point to the cranial relationship of the Italian 
Archidiskodon meridionalis and its ancestor A. planifrons. 





Weithofer's Type Skull of Elephas lyrodon = Archidiskodon 
meridionalis female 
Fig. 864. Types of two individuals of Elephas lyrodon Weithofer, 
Florence Museum, upper Val d'Arno deposits (Weithofer, 1889, pp. 79 and 
80): "Hier sei beziiglich der zum erstenmal genannten Species Elephas lyrodon 
nov. sp. nui bemerkt, dass sie auf zwei vollstandige Schiidel sammt Stoss- 
ziihnen, sowie Schadelfragmenten mit Stossziihnen und mehreren Unterkiefern 
und isolirten Stoss- und BackenzJihnen des Museums von Florenz basirt ist." 
After Weithofer, 1890, Taf. in, fig. 2 (right); iv, fig. 2, Schadel o (left); v, 
fig. 1, Schadel a (middle). All one-twentieth natural size. Compare figure 
865 (11, 11a). 



MS. Elephas giganteus Aymard (in Falconer, 1857, p. 
321).— (Lucien Mayet, letter, December 11, 1922): "Elephas 
giganteus Aymard n'est qu'une designation portee sur des etiquettes 
de sa collection par ce paleontologiste. II y aurait lieu de retrouver 
les pieces et de les determiner ; ce serait peut-etre E. meridional- 
?'s(?) ou E. lrogontherii{1) ou E. antiquus{?) . II s'agit tres probable- 
ment la — comme pour beaucoup des ossements recueillis par 
Aymard dans le Pliocene superieur et le Pleistocene du bassin du 
Puy — de designation inscrite sur des etiquettes, sans que les 
pieces correspondantes aient jamais ete d^terminees ou aient fait 
I'objet d'une revision ulterieure. C'est un nom — celui d'Elephas 
giganteus — a faire disparaitre purement et simplement." 



SKULLS AND SKELETONS OF ARCHIDISKODON MERIDIONALIS 
In figure 865 are displayed outlines of the lectotype (Nos. 
7, 13, 14), cotype (Nos. 8, 9, 11, 11a, 15), and other skulls of Archi- 
diskodon meridionalis, male and female, of which we ha\e been able 
to find figures in the literature. 

They include the following: 

1) Male skull figured in error by Falconer and Cautley in 1847 

as Elephas antiquus: front view, PI. xlii, fig. xix; side 
view, PI. XLiv, fig. XIX. Cranium C of Nesti [lectotype of 
Elephas meridionalis Nesti, 1825, Tav. i, figs. 1, 2]. 

2) Cranium of Elephas meridionalis, side view, after Weit- 

hofer. Cranium C of Nesti. 

3) Cranium A of Nesti [cotype], palatal view after Nesti, 

1825, Tav. i, fig. 3; front and side views after Weithofer, 
1890, Taf. I, figs. 1, 2. 

4) Front and side views of female cranium, type of Elephas 

lyrodon Weithofer, 1890, Taf. v, fig. 1, Taf. iv, fig. 2. 

Characters (Osborn, 1924). — These crania agree in the 
following points: (1) Extremely broad and vertically shallow 
anterior narial openings, a feature shared by Archidiskodon im- 
perator (Fig. 896); (2) extreme cranial abbreviation (hyper- 
brachycephaly) and depth (bathycephaly) resulting in hypsicephaly 
and acrocephaly ; (3) concave forehead or frontoparietal profile; 
(4) i)arieto-occipita] crest rising high in profile (cf. Weithofer, 
1890, Taf. I, fig. 2, E. meridionalis cotype, with E. imperalor ref., 
Los Angeles, Calif.) ; (5) through hypsicephaly, orbits and occipi- 
tal condyles approximated, vertical diameters greatly exceeding 
anteroposterior diameters; there seems to be little doubt of the 
phylogenetic kinship of the Archidiskodon meridionalis with the A. 
imperalor crania; (6) cranial profiles of A. meridionalis and A. 
imperator analogous to the hypsicephalic and acrocephalic profile 
of Mammonteus primigenius; (7) in comparison with the more 
primitive, more platycephalic A. planifrons, with smaller narial 
openings (Fig. 848), the known crania of A. meridionalis are much 
larger and more progressive than the known crania oiA. planifrons. 

CRANIA IN THE FLORENCE MUSEUM 
(W. D. Matthew, September, 1920). — Here are found be- 
sides the tyi^e skull of Elephas meridionalis Nesti [Cranium C] 
several more or less incomplete skulls, many jaws and teeth, in- 
cluding those from the Upper Pliocene near Magello. A fine 
.series of E. meridionalis crania published by Weithofer (1890) 
includes the type of Elephas lyrodon Weithofer, which is obviously 
a female of Elephas [ = Archidiskodon] meridionalis in which the 
tusks are of a lyrate arrangement drawn together at the points, 
more after the manner of Hesperoloxodon antiquus but more slen- 
der. In the same beds with these crania of E. [A.] meridionalis was 
found part of a cranium of the straight-tusked elephant//, antiquus, 
which exhibits a wide spreading of the tusks at the base, a broaden- 
ed rostrum not less than 2}i feet apart at the rim of the socket. 
Contemporary with the above Archidiskodon and Hesperoloxodon 
are numerous remains of Anancus arvernensis, including a fairly 
preserved skull with tusks found in the environs of Florence, one 
tusk broken off and repointed during life. 




/--,..-.'--,--A 



E. PLANIFRONS Rel 
File, 1B47, PI. XUl. FiQ. XVI 



f . PLANIFRONS Ref 
F.ic. 1845, Pi. IX 



E. PLANIFRONS Re'- 
F«lo. 1M7, PI. XUV, F.o XVI («v.i 



E, PLANIFRONS He' 
F»lc , 1846, Pi. X, Fifl. 1 'w* 



Fig. 8G5. Crania of the MAMMONTm;E (1-19) as determined uy the 

976 



IMIESENT AUTHOR IN THE EXPLANATORY LEGEND OPPOSITE. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



977 



Elephas lyrodon Weithofer, 1889, 1890. — (Weithofer, 
1890, p. 173) : "Hier in Florenz befindet sich ein vcillig ausge- 
wachsener, leider aber nicht besonders gut erhaltener Schadel 
sammt den beiden Incisiven in situ (Falconer's Nr. 6), ein gleich- 
falls sehr altes Pramaxillarfragment mit dem rechten Stosszahn 
vollstiindig, dem linken zum grossten Theil erhalten (Falconer's 
Nr. 9), ferner ein j lingerer Schadel sammt beiden Stosszahnen und 
dem Unterkiefer, womit weiter auch die Wirbel, Rippen, Schulter- 
blatt und Becken im Zusammenhang gefunden vvorden waren 
(Falconer's Nr. 8), ein Oberkieferfragment mit dem linken 
Stosszahn und jederseits zwei Backenziihnen in situ, endlich mehre- 
re mehr oder weniger vollstjindige Unterkiefer oder Unterkiefer- 
halften; einige Stosszahne, sowie offenbar auch einigc der i.solirten 
Backenzahne werden hieher geziihlt wcrden konnen." Weithofer 
(op. cit., pp. 191, 192)observes very close kinship to E . meridionalis 
in the dentition and ridge formula: M 3 j"]^| . He points out 
numerous differences in the structure of the skull and skeleton, and 
then concludes (p. 193) : "Ich glaubte daher, durch die angefiihrten 
Umstiinde gezwungen, diese neue Form als eine distincte Species 
betrachten zu mussen, die ich nach der so iiberaus charakteristi- 



schen Form ihrcr Stosszahne Elephas lyrodon nov. spec, benannte." 
Matthew, 1920, Osborn, 1924: The type cranium of Elephas 
lyrodon may be regarded, from the very slender character of the 
tusks, as a female of Archidiskodon meridionalis. 

SKELETON OF ARCHIDISKODON MERIDIONALIS IN THE PARIS 

MUSEUM 

The affinity of Archidiskodon meridionalis to A. irnperator is 
clearly displayed in a comparison of figures 866 (.4. meridionalis) 
and 896 {A. irnperator). The superb skeleton discovered in 1869 
near the village of Durfort, Gard, is shown herewith (Fig. 866) 
through a photograph kindly furnished by Dr. Marcellin Boule and 
retouched by our own artist. 

Skeletal Characters. — Gaudry (1893, p. 12) observes of 
this specimen: "L'filcphant de Durfort n'appartient pas a la 
race primitive de VElephas meridionalis, ou les molaires ont un 
petit nombre de collines basses, enduites d'un email epais, mais a 
la race modifice de cette espece, c'est-a-dire au type du Val d'Arno 
[Footnote: 'II y a des dents du Val d'Arno qui, par leur allonge- 
ment et leurs lames nombreuses, etroites, a email mince, ressem- 



EXPLANATORY LEGEND OF MaMMONTEUS, PaRELEPHAS, AND ArCHIDISKODON CrANIA 

Key to Osborn's dptprminations in the present Memoir of the crania 
illustrated in figure 865. 

Fig. 865. The inscriptions on this figure were written before the phylogenetie relationships of these crania were fully understood and before Elephas 
Irogontherii was clearly separated from E. primigcnius. The outline figures taken from various authors are arranged in three sections: Upper, including Elephas 
primigenius and E. trogontherii; middle, including E. meridionalis, and lower, including E. planifrons. 



(A) Skulls of True Mammonteus primigenius (a); of Parelephas trogontherii (6) 

f 2. Typical MammonUus primigenius. E. primigenius ref., after Falconer, 1846 [1847, PI. XLiii, fig. xxiv|. 

(a) 'I 1. Mammonteus primigenius. E. primigenius rci., after Pohlig, 1891, p. 384, fig. 120. 

[ 5. " " E. primigenius veL, after Pohlig, 1891, p. 384, fig. 120 (rev.). 

[ 3. Typical Parelephas Irogontherii. E. primigenius trogontherii (female), after Pohlig, 1891, p. 386, fig. 121. 

(6) \ 4. " " " Side view of same cranium as above. 

0. " " " E. primigenius rcf., after Falconer, 1846 [1847, PI. xlv, fig. xxiv (rev.)]. 



A rchidiskodon meridionalis. 



(B) Skulls of Archidiskodon meridionalis; (C) Skull of ,\. imperator 

E. meridionalis Icctotype, after Falconer, 1846 [1847, PI. xLii, fig. xixj, Val d'Arno, Italy = Cranium C of Nesti. 
E. meridionalis Icctotype, after Falconer, 1846 [1847, PI. XLiv, fig. xix (rev.)]. Side view of same skull as above. 
E. meridionalis Nesti, eotype, 1825, Tav. i, fig. 3, Val d'Arno, Italy = Cranium A of Nesti. 
E. meridionalis eotype, after Weithofer, 1890, Taf. I, fig. 1 = Cranium A of Nesti, Florence Museum. 
E. meridionalis eotype, after Weithofer, 1890, Taf. i, fig. 2 = Cranium A of Nesti, Florence Museum. 
E. meridionalis Icctotype, after Weithofer, 1890, Taf. ii, fig. 1 (rev.), Val d'Arno, Italy = Cranium C of Nesti. 
Side view of same skull as that figured by Falconer above. 
11,11a. VFemale of Archidiskodon meridionalis. E. lyrodon Weithofer, type, 1890, Taf. v, fig. 1, Taf. iv, fig. 2 (rev.). 



7. 

i:i. 

9. 

8. 
15. 
14. 



12. ?Female of Archidiskodon meridioruilis. E. lyrodon ref., after Weithofer, 1890, Taf. vi, figs. 1, 2. 



(C) 



10. Archidiskodon imperator. E. imperator ref.. Museum of History, Science and Art, Los Angeles, Calif. I'Vom Rancho La Brea. 



(D) Skulls of Archidiskodon planifrons, Upper Pliocene of India 

16. Archidiskodon planifrons. E. planifrons ref., after Falconer, 1846 [1847, PI. xlii, fig. xvi], Pinjor horizon, India. 

17. " " E. planifrons ref., after Falconer, 1846 [1845, PL x, fig. 1 (rev.)]. 

18. " " E. planifrons rcf., after Falconer, 1846 [1845, PI. ix (skull); PI. xi, fig. 3 (jaw)|. 

19. " " E. planifrons ref., after Falconer, 1846 [1847, PI. xliv, fig. xvi (rev.)]. 

We observe that in its front and side cranial aspects Archidiskodon planifrons is very primitive. In the two crania of .1. mrridionalis shown in side view, we 
observe a transition from the yl. /jfam/rons type to the more typical .4. meridionalis, namely, E. meridionalis Weithofer, 1890, Taf. i, fig. 2; this skull is 
decidedly hypsicephalic; it strongly resembles in this feature the skull of -4. imperator, also the skull of the typical Mammonteus primigenius; it differs pro- 
foundly from the skull of Parelephas trogontherii which has a profile more like that of Elephas indicus. 




Abchidiskodon mbbidionalis of Durfobt, of Cromerian or Nokfolkian, Lower Pleistocene, Age 

I.'ig 866 Skolotoii (largplv restored) of Elephas [ = Archidiskodon] meridionalis, known as "I'Elcphant de Durfort," as restored and inount,ed in the tialei'ic 
de I'al.V.ntologio, Mus.mn, d'Histoirc Naturelle, Paris (Gaudry, 1893); discovered in 1869 near the village of Durfort, Gard, between Ntmes and Vigan; 
exeavati<.nr„m|)letedJune21, 1873; restored under the direction of Gervais; installed in 1885; fully compared, measured, and figured by Gaudry m 1893. 
One-thirtietli natural size. Photograph by courtesy of Dr. Mareellin Boule. 

Doctor Boule writes a,s to the parts restored, or "artifieiels," in this Durfort skeleton (letter of December 12, 1921): 'Me n'ai j.as voulu doteriorer ••etle 
photographic en marquant sur elle Ics parties restaurfies, mai.s voiei la nomenclature des principaux elements de squelette qu'un cxamen assez somraairc et 
superficid Ic scul quV.n puisse faire en cettc saison .sans avoir recours a des cchafaudages, nous a montro etre arlificicls: Crane.-Parties supencure ct 
posterieurc Alveoles des defenses, en partie. Extromite ant6rieure de la defense droite. Mandibide droite.-Branehcs montantcs, extr6nut6 de la symphysc. 
Colomie vertebrale. ^uclques vertebres eervieales, lombaires et caudalcs. Cbtes.-La plupart sont restaurees ,.u en platre. Membres ant6ricurs.-Un 
scaphoidc et 3 ou 4 i)halanges. Membres posterieurs.— T^te supiSrieure du tibia droit et p6ron6 droit." 

Compare figure 868, lateral view of the same skeleton, modified after Gaudry, 1893, Plate. 

Ilidgc-platcs (Gaudry, 1893, p. 13): "I>es molaires de I'ftlophant de Durfort ont des lames ph.s nombreuscs ct plus n.inccs ,|uc ,-cllcs dc Scniur, dc 
Chagny, dc Perols, prcs de Montpellier, dc Randan (Allier), du Monte Verde, pres dc Rome, (pic nous pos.scdon.- 
autant aux dents de V Elephas (inliquus appelces inlcrmedius tpi'aux dents dc \'E. meridiimalia." 
diskodon] meridionalis. 



ous pi)s,s(Klons au Museum. . . lOllcs vcsscniliicnt pn'S(iiic 
Referred by Gaudry to a progressive stage of Ekphas [=Archi- 



978 




Archidiskodon meridionalis of Durfort, France 
Rfsldration by Margrct Flinsfh Buba, May 23, 1930, oiu'-fifticMi natural size 
Fig. 867. Primitive staKc, drawn diroctly from skeleton in Paris Museum, represcntin;; an animal of young adult ago. Obs(-rv(' tlic lialf-grown tusk.^, 
the proljoscis elevated in order to expo.se the long slender rostrum of the lower jaw (■haraeteri.sti<- of this species. Ear drawn .small, of somewhat primitive 
contour. Ratlier drooping posterior quarters, draw n from supposed mammontiiic affinities. Digits o in the manus and 4 in the pes, ungues 5 and 4. 




Durfort Skeleton of Archidiskodon meridionalis. Lower Pleistocene Age. About one-fiftieth natural .size. Compare Figure 866 

Fig. 868. Lateral view of "I'Elephant dc Durfort" in the Galerie de Paleontologie, Museum d'Histoire Natureile, Paris (Gaudry, 1893); for details 
see legend of figure 866, for description see pages 977, 980 of the present Memoir. 

This classic skeleton, redrawn after photographic plate in Gaudry (1893), exhibits the right lateral view as compared with the oblique anterior view 
introduced in figure 866 above, and enables us to estimate clearly the height of this animal in the restoration (Fig. 867). 

According to Gaudry's measurements, the skeletal shoulder height from the ground to the top of the anterior dor.sal spine is 3830 mm. (12 ft. 6^4 in.); 
summit of occiput to the ground 4150 mm. (13 ft. 7% in.), as mounted. 

According to Osborn's measurements, tlie skeletal shoulder height is 3499 mm. (11 ft. 5?^ in.). 

The method of the present author of estimating the skeletal shoulder height is to take the standing height of the forelimb to the top of the scapula. This 
measurement was arrived at through the observation of the author, when sitting upon a living elephant, that by placing his thumbs together on the tallest 
spine between the scapulse with the extended small fingers resting on the top of each scapula, the spines were found to be on a level with the scapulae, the 
latter rising and lowering a few inches above and below the spines in the process of walking. This accounts (in the present instance) for the smaller measure- 
ment given by Osborn (3499 mm.) in the skeleton and, after adding the usual 6^ per cent, of 3721 mm. or 12 ft. 2}i in. in the flesh (see restoration to the 
same one-fiftieth scale, Fig. 867). 

979 



980 



OSBORN: THE PROBOSCIDEA 



blent plus a pelles de VEIcphas auliquits qu'a certaines dents d'Ele- 
phas meridionalis.'] et du Forest-bed ou les eollines commencent 
a se multiplier, a diminuer d'epaisseur, a augmenter de hauteur. 
EUes ressemblent presque autant aux dents de VElephas anliquus 
appelees intermedius qu'aux dents de VE. meridionalis. II me 
semble en outre que les defenses sont plus courbees et que les os 
des pattes sont moins opais que dans les E. meridionalis les plus 
anciens; en cela I'animal de Durfort marque encore une tendance 
vers les Elephants quaternaires. Les molaires de 1' Elephant de 
Durfort ont des lames plus nombreuses et plus minces que celles 
de Semur, de Chagny, de Perols, pres de IMontpellier, de Randan 
(Allier), du Monte Verde, pres de Rome, que nous poss6dons au 
Museum." 

"J'ai dernierement, avec M. Marcellin Boule, travaille a 
degager un enorme Elephas meridionalis qui a etc decouvert dans 
les sables voleaniques de Sencze, pres de Brioude, par un savant 
archeologue, M. Le Blanc. Ses dents contrastent singuliorement 
avec celles de VElephas meridionalis de Durfort par leurs eollines 
basses, tres grosses, a email epais. Elles annoncent un animal 
encore i)lus gigantesque." 

The dimensions given by Oaudry (]). 19) are as follows: 

"Hauteur du squelette, a la tete 4™15 

du squelette, au garrot 3 83 

Longueur du squelette avec les de- 
fenses, la queue n'etant 
pas allongee et etant 
placee dans sa po.sition 
naturelle 6 80" 



The associated fauna is described by Gaudry (op. cit., p. 14) 
as follows: "En realite, tous les os de mammiferes de Durfort que 
nous avons dans le Museum se rapportent seulement a 4 Elephants, 
4 Hippopotames, 5 Bisons, 4 Cerfs, 1 Rhinoceros, 1 Cheval." The 
associated flora is rich in oak trees (op. cit., p. 17) : ". . . il y a eu 
la une veritable foret de ces arbres, comprenant au moins quatre 
especes." There were also beeches. 

Archidiskodon meridionalis cromerensis 

Deperet and Mayot, 1923 
Figures 870, 87 IF 

Kessingland, Suffolk, England. Age: Cromer Forest Bed of Norfolk 
East Anglia = Cromerian or Norfo!kian stage, 1st Irilrrglacinl of authors. 
See geologic note above on the Forest Bed or Cromerian (p. 970). Lower 
Pleistoeene. 

Specific Chahacteus (compare Depehet and Mavet, 1923, 
p. 153).— M' with -I-IOK (i.e., 13?) ridge-plates; very broad, 
diameters 180 mm. X 80 mm.; laminar frequency 6-5 in 10 cm., 
greatly exceeding the typical E. meridionalis (5); enamel of 
moderate thickness, much plicated, with salient loxodont sinus on 
posterior border. 

According to Deperet and Mayet this subspecies is regarded 
as a final phase of the Lower Pleistocene Forest Bed age of Eng- 
land, a horizon which Osborn (1922.570) regards as corresponding 
with the Isl Interglacial period, the period at which the true 
Archidiskodon meridionalis makes its last a]:)pearance in southern 
Europe. 

Giant grinders of the southern mammoth, Archidiskodon meri- 
dionalis, occur in the Forest Bed fauna (Osborn, 1922.563, p. 440, 
list of vertebrata, etc.). The 12-|- ridge-plated tooth (Savin Mus. 
No. 197 — Fig. 871 F of the present Memoir) is 109 mm. in breadth, 
as compared with 80 mm. in Deperet and Mayet's type (see Fig. 




EUphas anfiquus. — The first left upper true molar ; from the Pleistocene of 
Grays, Eteex. J. The lower border of the figure is the inner border 
of the specimen. 

ReKEUIIED .\rCHIDISKODON MEIllDION.\LIS FHOM THE I,OWEIt PLEISTOCENE OI' 

OK NoHFOL 

Fig. 869. Eleven plated grinder, Vl.M', of Archidiskodon meridionalis 
illustrated and erroneously referred by Lydckker to "Elephas anliquus." 
After I.ydekker, 1886.2, ii. 12.5, fig. 26 (Brit. Mus. M.2004). Three-fourths 
natural size. 

lA'dekkcr described (p. 126) this tooth as follows: "The first left upper 
true molar in a half-worn eoudition; probably from Grays. This specimen 
(woodcut, fig. 26) agrees very closely in general characters with the .Japanese 
molar of E. nninadicus figured on page 168 |Fig. 1189 of present Memoir]; 
it is noticed by I.eith-Adams, op. cit. [1877 1881], p. 22. ,Vo histori/." Com- 
pare figure 870 op()ositc, type of the Elephas meridionalis cromerensis of 
Dcpdrct and Mayet. 



Essex (I^ekt), .\ni) Tvi-e op Archidiskodon meiudio.nalis cuomkiiensis 
K (Right) 

Fig. 870. Ty|ie l.M' of Elephas meridionalis cromerensis Dep6ret and 
Mayet, 1923, PI. ix, fig. 1, p. 220, about one-half natural size: 

"Fig. 1. — Elephas meridionalis mutation cromerensis, du Forest-Bed, 
a Kessingland. M' gauche avec fossili.sation et patine caracteristifiues (Iron- 
pan). (Voir, p. 1.52.) Ce trc^s beau document i)al(5ontologique — qui ressemble 
de fayon frappante a la dent n° 33,334 du British Museum- fait partie de la 
collection du Dr. Pontier. Photographic obligcanuncnt c(iiunuini(|uee par 
notre .savant confrere." 

(Osborn) Mrit. Mus, 33,331 from the I''i>rcst Bed of the Norfolk coiist (,se(> 
Leith Adams, 1S77 1881, PI. xxiv, fig. 2, p. 198) agrees closely with the above 
Deperet and Mayet type. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



981 



871 as compared with Fig. 870, Dcperet and Mayet's type). Also 
very numerous A. meridionalis grinders in the British Museum are 
catalogued by Lydekker (1886.2, pp. 108-113), mostly entered as 
"dredged off Happisburgh" or from the Forest Bed of Cromer, 
Norfolk. Two grinders (cf. Fig. 862) figured by Falconer agree 
in width with the measurements of Depcret and Mayet's type (80 
mm.). The superior grinders of A. planifrons from the Siwaliks 
vary in width from 88 to 100 mm. ; the inferior grinders vary in 
width from 78 to 109 mm. (see Table VII above). There is 
therefore no appreciable change in the breadth of the grinding 
teeth between A. planifrons and these specimens of ^. meridionalis 
(see Osborn's remarks above, page 970, on the Forest Bed fauna). 

Elephas meridionalis, mutation cromerensis Deperet and 
Mayet, 1923. "Monographie des Elephants Pliocenes d'Europe et 
de I'Afrique du Nord," Deuxieme Partie of "Les Elephants 
Pliocenes." Ann. Univ. de Lyon, Nou\'elle Serie, I. — Sciences, 
Medecine. Fasc. 43, pp. 150, 152, 157. Type.— A third 

superior molar of the left side, l.M^ Horizon and Locality. — 
Forest Bed at Kessingland, England, Lower Pleistocene. Type 

Figure.— Op. cit., PL ix, fig. 1. 

Type Description. — (Op. cit., p. 153): "Cette derniere piece 
(pi. IX, fig. 1) est une M' tres usee en avant ou manquent sans doute 
plusieurs lames. La couronne est large (80 mm.) pour une longueur 
conservee de 180 millimetres. On n'observe plus que 10 lames plus 
le talon posterieur. L'email est peu cpais et tres plisse sur toute 
I'epaisseur des bandelettes; il existe de sinus loxodontes assez 
saillants du cote posterieur. La frequence laminaire atteint 6, 5 
pour 10 centimetres, chiffre tres superieur a celui de I'E. meri- 
dionalis type (5) et meme a celui de I'animal de Saint-Prest (5, 5). 
C 'est le caractere essentiel qui permettra toujours de reconnaitre 
la mutation cromerensis de I'horizon du Forest-bed." 



ASCENDING MUTATIONS OF ARCHIDISKODON MERIDIONALLS 

Revision by Deperet and Mayet (1923) 

Deporet and Mayet give a detailed description (op. ril., pp. 

125-160) of all the principal known remains referable to this great 

species, showing that the archaic forms of skull, jaw, and molar 



MAMMONTEUS PRIMIGENIUS 'ASTENSIS 




M*MMONTEUS PRIMIGENIUS?ASTENSIS 




/ost./c/' Mai Forest Bed 

PARELEPHUS ■> TROCONTMERH SP? 



*13 + 

F.Tfsf Bed of Crome'', Norfolk 

PAPELEPHAS '' TffOGONTMERII SP? 

?r 




Dri.<iaea 



5ai.n Mui No 1240 Forest Bed 



Ip5.,ch Mas Forest Be. — ....,.« A. MERlDrON Alis CROMERENS-S 

ARCMIDISKOOON MERIDIONALIS CROMERENSIS f^Vi /%"/% — ""^ \x3 




Rase of Red Crgq Fglkenha. 



19 28 



H. F. O. 



Primitive Grinding Teeth of the Mammontin.e; Archidiskodon, Parelephas, M.uhmonteus 
Fig. 871. Molar ridge-plate structure and laminar frequency in the Forest Bed (Lower Pleistocene) and Red Crag (Upper Pliocene) eleiihants, after pencil 
sketches by the author in 1926 of specimens in the Ipswich and Savin (Cromer) Museums. All figures reproduced one-fourth natural size. Compare Deporet 
and Mayet, 1923, Pis. ix and xi. 

Pleistocene A, Cromer Forest Bed (Ipswich Mammonleus primigenius(l) astensis Dep6rct and Mayet, M" (?), 11 ridge-plates or laminie in 10 cm.; total 
Mas.) of 15 -f- ridge-plates. 

B, Cromer Forest Bed Mammonleus primigeniusC?) asicyisisC?) Depcret and Mayet, I.Mj, 12 ridgc-platcs or lamina- in 10 cm.; total 

of 13 -|- ridgc-plates. 

C, Cromer Forest Bed (Ijiswich Parelephas (?) Irogonlherii Pohlig, sp.(?),r.M3, 8 ridge-plates or lamina? in 10 cm.; total of l.j-l- ridge-plates. 
Mus., dredged) 

Parelephas (?) trogontherii Pohlig. sp.(?),' f)% ridge-plates or lamina? in 10 cm.; total of 21+ ridgc-plates. 

The determination of Parelephas (?) (C, D), corresponding to P. Irogontherii, is somewhat doubtful. 

In the Savin Museum also occurs the typical llesperoloxodon aniiquus ausonius (?) with 12 ridge-plates. 
Archidiskodon meridionalis cromerensis Deporet and Mayet, r.M', 6 ridge-plates or lamina? in 10 cm.; total 

of 12-|- ridge-plates. 
Archidiskodon meridionalis cromerensis Dep6ret and Mayet, r.M', only 5 ridge-plates or lamina; in 10 cm.; 

total of 12+ ridge-plates. 

Archidiskodon planifrons (?) Falconer and Cautley, or E. [Hesperoloxodon] aniiquus, M3, 5/2 ridge-platcs or 

laminae in 10 cm.; total of +11 ridgc-plates. 
Archidiskodon planifrons (?) Falconer and Cautley, or E. [Hesperoloxodon] aniiquus, 0/2 ridge-platcs or lamina; 

in 10 cm.; total of 10+ ridge-plates. 

Observe that .Mammonleus has very compact ridgc-plates and fine enamel, that in the ParelcphasC?) ridgc-platcs the enamel is somewhat thicker, that 
Archidiskodon has extremely broad and widely separated ridge-platcs of coarse enamel. Compare Depcret and Mayet (1923, "Les Elephants Pliocenes," Deux- 
ieme Partie, pp. 98-160) for molar diagrams of A. planifrons and A. meridionalis, type ancien, which correspond very closely with the present diagram (E, F, G, H). 



Pliocene 



D, Cromer Forest Bed (Savin 

Mus. 1240) 

E, Cromer Forest Bed 

(Trimingham) 

F, Cromer Forest Bed (Savin 

Mus. 197) 

G, Red Crag (Falkenham) 
H, Red Crag (Ipswich Mus.) 



'Teeth similar to those of Parelephas Irogonlherii were selected by Pohlig, 1891 (1892), pp. 303, 304, as the cotypos of his Elephas aniiquus Nestii, as shown 
in figure 94 1 below (Chap. XVII) ; referred by the present author to ParelephasC!) Irogonlherii neslii. 



982 



OSBORN: THE PROBOSCIDEA 



teeth approach A. plaiu'froiis, while the lii{j;hly progressive forms 
{op. ril., p. 153) approach the Lower Pleistocene stage of Archidis- 
kodon meridionalis cromerensis. 

History. — The southern mammoth is represented by superb 
specimens in the museums of France, which have recently been 
monographed by Deperct and Mayet (1923), reciting the studies 
of Nesti (1808, 1825), of Falconer (1868), of Weithofer (1890), of 
( 'uvier, of de Blainville, of Owen, and of Gervais. The southern 
mammoth (Elephas meridionalis) was long confused with the 
northern mammoth {Elephan primigenius). Deperet and Mayet 
remark {op. cil., p. 125): "("est seulement apres les voyages de 
I'alconcr en Italic et la publication des observations de ce savant en 
1868 que les caracteres \'K. meridionalis furent enfin reconnus et 
I'espece definitivement admise par les paleontologistes." 

DisTKiBiTiON. — As to geographic distribution, De]ieret and 
Mayet conclude (op. rit., )). 144): "Ij' Elephas meridionalis, dont 
le centre de dispersion principal est I'ltalie, s'est repandu sur pres- 
ijue toute la surface de TEuropc temperee: Russie du Sud, bassin 
tlu Danube, Allemagne du Sud, France, Angleterre, peninsule 
Iberique, et enfin dans I'Afrique du Nord. Les gisements sont tres 
iiombreux; aussi notre etude descriptive se limitera-t-elle aux 
l)ipces les plus import antes, notamment les Mm, et surtout a celles 
dont nous avons pu obtenir des photographies directes, grace a 
I'obligeance de divers de nos confreres, que nous citerons au cours 
(le cet expose. Comme en Italic, nous pourrons y reconnaitre trois 
mutations successives de I'espece: A. Forme archaique. B. Forme 
type. ('. Forme evoliiee oit rwetde." 

Mutations. — The chronological mutations close as follows 
{op. cil., p. 150): 

"2° Mutation de I'etage Cromerien (Sicilien) qui represente le 
debut du (^uaternaire et que nous designerons sous le nom de 
niutalion cromerensis." 

"1° Mutation de I'etage Saint-Prestien, cjue nous rapportons a 
I'extreme fin du Pliocene." 



Ascending Skhies of STUATHiiiAPHir Mitations of 

A. .MICRiniONALIS, BEGINNING OF EXTINCTION OF THE PhVLIM 

IN Entoi'E (Deperet and Mayet, 1923, v. 157) 

"I. Elephas planifrons. Couronne tres basse (hauteur d'uiie 
lame mediane moyennement usee, 50 a 60 millimetres) et large; 
lames transverses peu nombreuses (10 chez M' et 10 11 chcz Ms); 
grand ecartement des lames (frequence laminaire 3, 5 a 4 lames par 
10 centimetres de longueur de la couronne); email tres cpais, a 
iarges ondulations limitees a la parol externe des bandelettes; sinus 
loxodo.ntcs tres saillants et assez reguliers, surtout en avant." 

"II. Elephas meridionalis, mutatioji archaique. Couronne 
large et bas.se (H =60 a 75 millimetres); frequence laminaire 4, 5; 
email dpais a Iarges plis i)ou profonds; .sinus loxodontes plus 
irr^guliers." 

"in. Elephas meridionalis, mutation ttjpe. Couronne large et 
iin |)eu inoins ba.sse (H=75 a 90 millimetres); I\P a\ec 11 a 13 
lames, M.I avec 11 a 14; freciuence laminaire 5. lOmail moiiis cpais 
et a plis plus .serres qui int<>ressent presque toute l'e])aisseur (l(!s 
bandelettes; sinus loxodontes peu accuses et irrcguliers." 



"IV. Elephas meridionulis, mutation ilu Suinl-I'rcslitn (ttndu 
Pliocene). Couronne large et moins basse (H = 90 a 95 milli- 
metres) ; 1 ou 2 lames de plus aux Mm ; frequence laminaire 5, 5; 
email plus mince a plis serres affectant toute I'cpaisseur de la 
bandelette sinus loxodontes faibles et irreguliers. 

"V. Elephas meridionalis, mutationcTOfnerensis {iormo (|uater- 
naire du Forest-bed). Couronne moins large et plus haute (H = 90 
a 95 millimetres); 12 a 15 lames aux M'', de 12 a 16 aux M3 
sur les molaires entieres. Frequence laminaire 6 k 6, 5; email 
beaucoup plus mince, a plis nombreux et serres affectant toute 
I'epaisseur de la bandelette; sinus loxodontes peu accuses et tres 
irreguliers." 

"Extinction du rameau." 

Specific Characters of Archidiskodon meridionalis 
(Deperet and Mayet, 1923, p. 156). — "1° Crane. — Le crane 
connu surtout par les belles pieces de Florence, est caracterise 
par: une boite cerebrale arrondie tres large en haut et a bords 
presque paralleles; un vertex peu elevc arrondi, mais non j)ro- 
longe en dome; une region fronto-parietale excavee chez I'adulte 
une ouverture nasale reculee tres en arriere, etroite et tres etendue 
en travers, surmontee d'une epine nasale saillante; des arcades 
zygomatiques dejetees vers le bas; des alveoles des defenses rap- 
proches et subparalleles." 

"2° Mandibule. — La mandibule offre des particularites interes- 
santes: chez le jeune et la femelle, la symjihyse .se jirolonge par 
un bee presque horizontal qui continue le bord inferieur de I'os. 
Chez le male adulte, au contraire, la mandibule est depourvue de 
bee et presente une terminaison mousse et obtuse, avec i)arfois 
une toute petite pointe mediane insignifiante (scjuelette de Dur- 
fort). On a deja dit que VE. planifrons possede dans les deux 
sexes un bee mandibulaire beaucoup plus fort et dejeto presque 
verticalement vers le bas." 

"3° Defenses. — Les defenses sont, chez le jeune et la femelle, 
di\ergentes des la base et iissez fortement spiralees. Chez le 
male adulte, elles sont presque paralleles a leur base, .«e dirigent en 
bas et en dehors, decrivant ensuite une legere spirale dont les 
pointes reviennent un peu en dedans. (Jhez VE. planifron.'<, les 
defenses sont encore [ilus paralleles et out une courbure concaNC 
moins prononcee." 

"4° Molaires. — Les molaires de \'E. ineridionulis sont de 
iiieme type general que celles de VE. planifrons: couronne large et 
basse, lames tran.sverses i)e.u nombreuses, email cpais et generale- 
ment peu plissc; sinus loxodontes assez accu.ses mais inconstants." 

THE ARCHIDISKODON MERIDIONALIS I'HYMM MKIR.VrKS TO 

AMERICA 

Osborn (1924) welcomes this splendid resume by his friends 
Deperet and Mayet of specific characters and mutations of A. 
planifrons and A. meridionalis, but does not agree that there is 
adequate evidence that this phylum became extinct, for it apjiears 
quite certain that the Archidiskodon meridionalis tyi)e migrated 
from Asia into North America and became ancestral to Archidis- 
kodon impcralor of Nebraska. Nor does Osborn believe that the 
meridionalis of Italy and France is in any way related to the 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 



983 



Elephas hysudricus of the Upper Siwaliks treated in Chapter XX. 
The cranial types in these two animals are fundamentally distinct. 
If Africa should prove to be the home of Archidiskodon, as 
appears probable through the discovery of two new primiti\e 
species in southern Africa, namely, A. subplanifrons and A. broomi, 



it may indicate that members of the generic piiyluin Archidinkodon 
originated in Africa and migrated northward into Europe and east- 
ward into southern Asia rather than having followed the generally 
accepted reverse line of migration from southern Asia westward 
into Europe and Africa. 



[It will be observed that Professor Osborn wrote this portion of the Memoir in 1924. From that time onward he 
constantly sought new evidence to test his hypothesis of an African center of dispersal of the Proboscidea (see Fig. 815). 
A decade of study (1924-1934) only confirmed him in his belief and in 1934 he wrote an article in American Museum 
Novitates (Osborn, 1934.925) entitled "Primiti\^e Archidiskodon and Palaeoloxodon of South Africa," on the basis of 
which the following section has been revised but in no sense have the views of the author been changed otherwise than 
appear in his own writings. — Editor.] 




I'lG. 872. Fo.SSIL PUOBOSCIDEA RoUTE FROM SoUTHEnN EQUATORIAL AFRICA 

Compare Andrews (1911), Dietrich (1916), Haughton (1922), Hopwood 
(1926), Dart (1927), and Broom (1928). [Since 1928 other species have been 
described (see pp. 986-993 below).— Editor.] 

(2) Archidiskodon subplanifrons, Sydney-on-Vaal, Vaal River diggings, 
near Kimberley. 

(7) M elarchidiskmlon griqua, Vaal River diggings, Griqnaland West. 

(10) Archidiskodon broomi, The Bend, Vaal River, near Kimberley. 

(11) Palieoloxodon transvaalensis, Bloemhof, southwestern Transvaal. 

(12) Palsoloxodon sheppardi, Bloemhof, southwestern Transvaal. 

(3) Loxodonta zulti, Zululand. 

(9) Mastodon s|j. (?), northwest of Lake Nyassa, near Uraha Hill. 

(4) Deinotherium hobleyi, Karungu, near Victoria Nyanza, Tanganyika 
Territory. 

(6) Palseoloxodon recki, near Oldoway, Serrengctisteppe, northern 
Tanganyika Territory. 

(8) Metarchidiskodon grigtia, Kaiso Bone-beds, near Albert Nyanza. 
(.5) Mastodon sp. (?), near Khartum, Sudan. 

(1) Deinotherium, ref., Elephas ref., Lake Rudolf. 



2. ARCHIDISKODONTS AND METARCHIDISKODONTS OF SOUTH AFRICA 
(Osborn, 1934.925, pp. 1 — 10): "Every year brings fresh proof that Africa was the center of the origin and 
adaptive radiation of the Proboscidea. Since 1907 numerous more or less primitive superior and inferior grinding 
teeth have been discovered from the Vaal River terraces and other localities of the Transvaal, South Africa. The 
geologic level and localities are chiefly on the (1) higher and most ancient terrace (200-300 feet) of the Vaal River; 
(2) middle terrace (60-80 feet) of the Vaal River; (3) lowest and most recent terrace (40 feet) of the Vaal River. 



984 



OSBORN: THE PROBOSCIDEA 



Theoretically the lowest terrace may be as old as the lower levels of the middle terrace. Flint iun)lements occur 
in the middle and lower terraces only (Dart, 1929). In Osborn's opinion the Archidiskodon subplanijrons, and 
Archidiskodon pro planif runs, . . . types found in the middle terrace were washed in from an older Pliocene horizon." 

"Certain of the Transvaal grinding teeth surely belong to very primitive stages of Archidiskodon; others 
probably belong to primitive stages of Palaeoloxodon and were originally referred to Loxodo7ita, to Archidiskodon 
and to PUgrimiay^ Only by careful comparison and analysis is it possible to separate the species belonging to 
these several genera from each other." 

"Up to the present time [1934-] the nineteen species described by W. B. 8cott (1907), Raymond A. Dart 
(1927, 1929), S. H. Haughton (1922, 1932) and H. F. Osborn (1928) are provisionally referred as follows: 



Locality 

Zululand 

?]\Iiddle terrace, Vaal River 

Lowest terrace, Vaal River 

Lowest terrace, Vaal River 

?Middle terrace (lower), Vaal Ri\er 

Lowest terrace, Vaal River 

Middle terrace (lower), Sydney-on-Vaal 

Middle terrace (lower), Sydney-on-Vaal 

Middle terrace (lower), Sydney-on-Vaal 

?Middle terrace, Vaal River 

Lowest terrace, Vaal River 

Middle terrace, Vaal River 

Lowest terrace, Vaal River 

Lowest terrace, Vaal River 

Pniel Estate, ? River 

?Recent, Limpopo River 

?Recent, Steelpoort River 

?Middle Terrace, Vaal River 

Higher terrace, Vaal River 



Original Reference 

Loxodon Zulu Scott, 1907 
Loxodonta griqua Haughton, 1922 
Archidiskodon transvaalensis Dart, 1927 
Archidiskodon sheppardi Dart, 1927 
Archidiskodon subplanifrons Osborn, 1928 
Archidiskodon broomi Osborn, 1928 
Archidiskodon vanalpheni Dart, 1929 
Archidiskodon loxodontoidcs Dart, 1929 
Archidiskodon milletti Dart, 1929 
Archidiskodon andrewsi Dart, 1929 
Archidiskodon hanekomi Dart, 1929 
Archidiskodon yorki Dart, 1929 
Pilgrimia yorki Dart, 1929 
Pilgrivtia wilmani Dart, 1929 
Pilgrimia kuhni Dart, 1929 
Loxodonta prima Dart, 1929 
Loxodonta nfricana, var. obliqua Dart, 1932 

[1929] 
Pilgrimia archidiskodontoides Haughton, 

1932 
Pilgrimia stibantiqua Haughton, 1932 



Present Generic Reference 

= Loxodorda 

= Mctarchidiskodon, n.g. 

= Palaeoloxodon 

= Palaeoloxodon 

= Archidiskodon 

= A rchidiskodon 

= Archidiskodon 

= A rch idiskodo n 

= Archidiskodon 

= f Palaeoloxodon 

= Palaeoloxodon 

= A rchidiskodon 

= Palaeoloxodon 

= Palaeoloxodon 

= Palaeoloxodon 

= Loxodonta 

= Loxodonta 



= Palaeoloxodon 
= Loxodonta" 



"Sununing up these species, the ascending geologic level records (Dart, Haughton) are as follows' 
VIII of the present Memoir] : 



[Table 



Table VIII. — "Faunal Distribution on the River Terraces of the Transvaal" 



"Middle Terrace 



Middle terrace, 60-80 feet. 
Sy d ney-o n- Vaal . 
?Lower Pleistocene 



?A. subplanifrons* 
?A. proplanifrons** 
Older than the Upper Pliocene ^. 
planif rons of the SiwaUks, India 

*actuallv recorded from a depth of 
50-60 feet, middle terrace (60-80 ft.) 

**at a depth of .56 feet, middle 
terrace, Vaal River 



Upper levels: 
^4. yorki 
A. broomi 
AL griqua 

Lower levels: 
/-". andreivsi 
A. vanalpheni 
A. 7nilletti 
A . loxodontoides 
Bunolophodon ?gen. ?sp. 



Lowest terrace, 40 feet. 
With flint implements. 
?Middle Pleistocene. Bloemhof 

Upper levels: 
Bubalis baini 
Equus capensis 



Lowest levels: 
P. transvaalensis 
P. sheppardi 
P. yorki 
P. wilmani 
P. kuhni 



Recent — 4 feet. 
Pleistocene 



Levels unknown : 
'/P. hinickoini 
fL. subantiqua 
fP. archidiskodontoides 

Recent : 
L. prima 
L. africana var. obliqua" 



''' Pilgrimia Osborn (December 20, 1924) is antedated by Palaeoloxodon Matxiimoto (September 20, 1924). 

-[To tlie.se might be added Archidiskodon proplanifrons described in this article (Osborn, 1934.92.'), p. 10), see page 986, figure 873, below.— Editor.] 



THE MAMMONTINiE: ARCHIDISKODON AND METARCHIDISKODON 



985 



"By their outstanding characters these species divide into four groups as follows" [Table IX of the present 
Memoir] : 



Table IX. "Provisional Grouping of Species Referred to Four Transvaal Genera" 



"A. subplaiiifrons group 

Crowns very broad, 101 to 114 
mm. Enamel very thick. 
Transverse conelets 4-6 (A. 
pruplanifronn) to 22-24 {A. 
broonii). Cement enveloping 
crown. V-shaped cemented 
valleys at smnmits broader 
than dentinal areas. Lo.xo- 
dont sinus foldings double, 
less prominent, irregular. 
The mass of cement exceeds 
the mass of dentine. 

Cf . paratype of A . meridionalis 
Nesti of Val d'Arno, also 
Brit. Mus. M 1264 1 , M 1 2642. 



In this group are the following 
species : 

A. proplanifrons, A. sub- 
planifrons, A. milletli, A. 
yorki, A. vanalpheni, A. 
broomi. 

The ridge plate height increases 
from 55 in .4. proplanifrons 
to 62e. in A. subplanifrons, 
to 118 in A. milletli, to 129 
in A. vanalpheni and 110+ 
in A. broomi. Meanwhile 
the number of ridge plates 
in 100 mm. remains con- 
stant, namely, 3 in ^. pro- 
planifrons, ^Yi in A. plani- 
frons of India and 3 in ^4. 
broomi. 



M. griqua group 

Crowns relatively narrow, 86- 
94 mm. Enamel thick. 
Transverse crests 6-8. Ce- 
ment areas narrower, not 
enveloping crown. Valleys 
V-shaped {M. griqua). Post- 
sinus fold very prominent. 
Total enamel length un- 
known. The mass of cement 
exceeds the mass of dentine. 



Cf. A. planifrons r)nnnn)is 
Stefanescu. 



In this group may be the 
following species: 

.1/. griqua, P. andrewsi, A. 
lo.rodontoides. 

The generic relationships of 
this group are doubtful ; the 
narrow crowns separate the 
types of M. griqua and of P. 
andrewsi from the broad 
crowns of the typical Archi- 
diskodon. The U-shaped 
valley of M. griqua of Fig. 3 
is quite distinct from the 
V-shaped valley of P. an- 
drewsi; similar teeth hsLve 
been discovered in Europe. 
It is probable that these 
teeth represent a genus dis- 
tinct either from Archidisko- 
don or Palaeoloxodon, name- 
ly Metarchidiskodon. 



P. transvaalensi s group 

Crowns of M^ relatively nar- 
row, 70 mm. {P. wilmani) to 
110 mm. {P. transvaalensis) . 
Indices = 41 to 51. Enamel 
relatively thin ; conelets fine- 
ly crimped, i.e. numerous. 
Cement areas progressively 
narrower t han dentinal areas. 
Cemented valleys greatly re- 
duced. Ridge plates narrow 
and increasingly lofty, 128 
mm. {P. wilmani), 259 mm. 
{P. hanekomi). Ridge plates 
per 100 mm. 4-6. Sinus fold- 
ings extremely reduced or 
progressi\ely wanting. \a\- 
leys V-shaped {P. andrewsi). 



Loxodonta prima group 

Crowns relatively narrow, 74 
mm. (L. prima) to 92 mm. 
(L. subaniiqua). Enamel 
relatively thin, coarsely 
crimped ; conelets numerous. 
Cement thin in middle, 
thick at edge. Ridge jjlates 
jier 100 mm. = 4 (L. afrirana 
obliqua) to B'i (L. subanti- 
qua). Broad typical loxo- 
dont sinus expansion, double 
sinus foldings in contact. 
Total ridge plates 9 (L. 
prima) to 12-13 (L. zulu). 



In this group are the following: In this group are the following; 



P. [ = Pilg.] kuhni, P. 
[ = Pilg.] yorki, P. [ = Pilg.] 
wilmani, P. archidiskodont- 
oides, P. sheppardi, P. trans- 
vaalensis and P. hanekomi. 
Also possibly ?P. andrewsi. 
These seven or eight types 
are much more uniform in 
character than members of 
the M. griqua group, P. 
sheppardi and P. transvaal- 
ensis formerly being referred 
by Dart to Archidisko- 
don. P. kuhni, P. yorki 
and P. wilmani were referred 
by Dart to Pilgrimia; the 
prevailing characters relate 
them more closely to Palaeo- 
loxodon. 



L. zulu, L. [ = Pilg.] subanii- 
qua, L. africana obliqua, L. 
prima. 

These occur only on the more 
recent levels and are clearly 
related to the existing Afri- 
can elephant, distinguished 
by the above characters." 



"The above phylogenetic arrangement is provisional. Only by the longitudinal sectioning method of Falconer 
is it possible to ascertain the true structural relationships of these Proboscidean molars. This is illustrated in the 
wide difference between sections of M. griqua (Fig. 3 [ = Fig. 882]) and P. andrewsi (Fig. 5 [=Fig. 1139]), also in the 
wide difference between the sections of A. subplanifrons (Fig. 1 [ = Fig. 875] and A. proplanifrons (Fig. 2 [ = Fig. 



986 



OSBORN: THE PROBOSCIDEA 



873]). A very interesting comparison is that of the 'planijrom series of Africa and the SiwaHks of India analysed 
as follows" [Table X of the pre.sent Menioii'|: 



Table X. — "Comparative Measurements of the Primitive Species of Ahchidiskooon FiiOM the Transvaal 

AND FROM the UpPER SiWALIKS OF InDIa" 



A. prnplanifrons 
A. subplanifrofis 
A. plamfrons* 



o 
o 



KM'' 



RM3 



mu 





tc 


^ 




3 








a> 


^^ 




"o 


bC 














Mm 






•a 


4) 


f— 1 


1 


§ 
w 






'M 






^ 


3 
Ci 






1 





-2 


IS 




93 




"a 

03 


k* 
^ 




"03 

S3 


§ 

T5 


-2 


J3 


bC 


S 

3 


_M 







0) 


bC 


"H 




c 





c 



C 


5 


'A 


'V- 


w 


hJ 





H^ 


55 


5K-6 


.3 


5 


Pre- and 

post sinus 
folds 


Conelets = 
4 6. 
Foldings 


17! 


62e. 


6 


4 


4 


Pre- and 
])ost sinus 
folds. 


Conelets = 
6-8. 


15: 


55 


8e. 


4K 


.3K 


Pre- and 
post sinus 
folds, 
greatly 
reduced. 


Conelets = 
16 e. in 
ridge plate 
4. 


17' 



^3 



h-1 P5 






104 58 



101 66 



86 48 



bC 

a 



a 



690 



650 



810e." 



SYSTEMATIC DESCRIPTION OF SPECIES 



IMid- 



Archidiskodon proplanifrons Osborn, 19.34 
Figures 873, 1239, PI. xxi 
Rpcordod from Gong-Gong near the Va.al River, South Africa. 
die (?) Plio<ene— see Fig. 1239.| 

Archidiskodon proplanifrons Osborn, 1934. "Primitive Archi- 
flixkndon and Palaeoloj-odon of South Africa." Amer. Mus. 
Xovitates, No. 741, August 24, 1934, pp. 10-12 (Osborn, 1934.925). 
Type. — "Third superior molar of the right side with six 
complete ridge plates, McCregor Museum 4334, cast Amer. Mus. 
26969." Horizon and Locality. — Recorded from (Jong- 

Cong 'at a depth of 56 feet in the "Middle Terrace," under 
a boulder, at a distance of 450 yards from the Vaal River, thus 
occurring at a depth of from 10 15 feet below the level of the 
present river bed'"'. "Probably washed from older Pliocene de- 
posits into the Pleistocene terrace gravels." [Middle Pliocene — see 
fig. 1239 below.— Editor.] Type Figure.— Op. fit., p. 5, fig. 2. 

Type Description.— (Osborn, 1934.925, p. 10): "Third 
superior molar of the right side with .si,\ complete ridge plates. 
McCregor Museum 4334, cast Amer. Mus. 26969. . . . Length = 179 
mm., breadth=104 mm., index = 58. Height of 5th ridge plate 
= 55 mm. Enamel length, restored =690 mm. Average enamel 
thickness = 5 mm. Ridge plate, number = 5^-6. Postconcave, 



preconvex, with 4 5 rounded conelets in the posterior plates and 
6-64- conelets in the anterior plates, slight median foldings in 
each transverse plate." 

"According to these measurements by Osborn, 1934, and 
those of Haughton, 1932, this type third right superior molar is the 
most primitive elephant tooth thus far discovered, even more 
primitive than Archidiskodon subplanifrons; it is indubitablj' an 
ancestral Archidiskodon with widely open \-alleys, summits of 
ridge plates much more widel}' separate than in A. subpkuiifrons, 
cement bathing the entire surface of the crown, median pair of 
conelets entirely distinct and undivided, total number of conelets 
estimated in the crown 26, as compared v\ith total estimated 
number 34 in A. siibplanifrons." 

"The . . . species, Archidiskodon subplanifrons Osborn, 1928, 
and the new species Archidiskodon proplanifrons, as tested bj' these 
thickly enameled, deeply cemented, low and sjjreading ridge 
plated, relatively broadened, third superior and inferior grinding 
teeth, are totally distinct from certain of the relati\'ely thin en- 
ameled, high ridge i)lated, less deeply cemented, relatively narrow 
molar types which have been erroneously referred by Haughton, 
Dart and Osborn to Archidiskodon, but more properly belong to 
I'nlneolo.rodoii including its .synonym Pilgrimia." 



"*This i.s No. 1996.3, .American Mu.seum Brown Collection (Fig. 4 [Fig. 8701) from near Sisvvan, India. Thi.s is tlic most primitive stage found in the 
large series of the Archidiskodon lUariifronx molar.-* collected by Barnuni Brown in the I'liper Pliociiie I'injor horizon of the Siwaliks, tndia." 
'"Haughton, 1932, page 2." 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



987 



Archidiskodon subplanifrons Osboni, 192S 
Figuies 81.j, 874, 875, 123.'-), 1239, PI. xxi 

Upper (?) Pliocene. Sydney-on-Vaal, Vaal River diggings, South Afriea, 
on the banks of the riv(>r, at a depth of from M to 60 feet [Middle Plioeene — 
see Fig. 1239]. 

History. — The tyjie tooth was riiscovered by Mr. W. Millett 
in the Vaal River diggings at Sydney-on-Vaal, South Africa, on the 
banks of the river, at a depth of from 50 to 60 feet ; the geologic 
age is probably r])i)er Pliocene. Through the courtesy of Miss 
Wilnian, Curator of the McCiregor Museum at Kiniberley, and of 



, Ce/iti nitti 




/^ nat. size- McGregor Mas. 45i>-4- 

Type. Arc/iLdiskodon propLanlfrons. Os born. 1954- 

Fig. 873. Type, right third superior molar, r.M', of Archidiskodon pro- 
planifrons, after original. McGregor Museum 433-1, Kimberley, South Africa; 
cast Amer. Mus. 26969. One-half natvu-al size. Originally referred by 
Haughton (1932.1, p. 2) to Archidiskodon subplanifrons. 

B, Crown view showing 5% ridge-plates, very broad cement areas — c, c, c; 
sixth plate represented by two conelets. Bl, section of the same showing 
cement-filled ridge-plates more widely open than in A. subplanifrons. 



our correspondent Dr. R. Broom, the molar was forwarded January 
20, 1927, to the American Museum for comparison and type de- 
scription; in return for this courtesy, a number of facsimile casts 
were made, so that each specimen bears the inscription: Archidis- 
kodon mbplanifron.% McGregor Mus. 3920, type, cast A. M. 21924. 
This species was finally described and figured by Osborn in 
"Nature," April 28, 1928, pp. 672, 673, in comparison with A. 
hroomi. 

Type Description. — The original description by Osborn is as 



folhjws (Osborn, 1928.749, p. 672) : "The first typo (re|)resented in 
Fig. 1 [ = Fig. 874 of the present Memoir]) I name Archidinkodoit 
fiuhplanifrons; it is a low-crowned, broad-plated, heavily cemented 
tooth, apparently a third inferior molar of the right side (McGregor 
Mus. 3920). The specific name subplanifrons refers to the fact 
that the crown height — from 2 to 2)1 inches — is about equal to that 
of the low-crowned types of Elephas planifrons Falconer of the 
Siwalik Hills, India; in some of Falconer's Upper Siwalik speci- 
mens the crown rises from SJi to 4% inches. The present specimen 
accordingly is believed to be of Upper Pliocene age." 

Archidiskodon subplanifrons Osborn, 1928. "Mammoths and 
Man in the Transvaal." Nature, Vol. GXXI, No. 3052, April 28, 
1928, pp. 672, 673. Supplementary description: "Primitive 
Archidiskodon and Palaeoloxodon of South Africa" (Osborn, 
1934.925, p. 10, fig. 1 = Fig. 875 of the present Memoir). Horizon 
.\Ni) Locality. — Upper(?) Pliocene. [Middle Pliocene — See Fig. 
1239 below. — F^ditor.] Sydney-on-Vaal, Vaal River diggings, on 
the banks of the river, at a depth of from 50 to 60 feet i?Middle 
Terrace), South Africa. Type Figure. — Op. cil., p. 672, fig. 1 

Type. — Posterior half of a third right inferior molar, r.Ms 
(McGregor Mus. 3920, cast Amer. Mus. 21924), exhibiting four 
complete posterior ridge-plates and half of another ridge-plate, 
deeply surrounded with cement on crown and sides, also buried in 
cement is a rudimentary plate; extremely low crowned with broad, 

/^iQreqor Mui. J920 Ximier.y 3. Africa. 

r. TO. 5 




Original Type Figure of Archidiskodon subpl.^nifrons 
Fig. 874. Ty|)e of Archidiskodon subplanifrons Osborn, 1928, from the 
Upper (?) Pliocene [Middle Plioeene — see Fig. 1239], Sydney-on-Vaal, South 
Africa, one-half natural size. Third inferior molar of the right side, r.Ms 
(McGregor Mus. 3920, Kiniberley, South Africa; cast Amer. Mus. 21924). After 
original type specimen kindly loaned to the American Museum for figuring 
and description (cf. Osborn, 1928.749, fig. 1, p. 672). Compare figures 840, 
876, also an r.Mj (Amer. Mus. 19965) of ^. planifrons from the Siwalik Hills. 



988 



OSBORN: THE PROBOSCIDEA 



heavy enamel, ridse-])lates set wide apart, two of the anterior 
ridge-phxtes expanding into a 'loxodont sinus' (4th and 5th); 
strongly resembUng the Archidiskodon planifrons type (Figs. 828, 
829), hence the specific name Airhidiskodon ftuhplariifwns. 

Specific Characters. — The concavely worn superior surface 
(Fig. 874, lower) enables us to determine this as an inferior tooth 
of the right side, apparently an r.Ms, in view of the posterior con- 
vexity (Fig. 874, upper) which forbids its reference to an r.M2. 
The crown is excessively broad and short, with convex cement- 
covered sides and ridge-plates which gradually increase in height 
posteriorly, iis follows (reading backwards) : 4th, imperfect, with 
central loxodont sinus; 5th, tr. 92 mm., height 53e mm., with 
anterior and posterior loxodont sinus; 6th, with cement, tr. 97 
mm., height 53e mm., loxodont sinus less prominent; 7th, tr. 101 
mm., height 63 mm., loxodont sinus less prominent, five to .six 
conelets; 8th, with cement, tr. 95 mm., height 61 mm., summit 
of ridge-plate narrow with six conelets; after this a rudimentary 
plate buried in cement. The thick, non-crenulated enamel of the 
ridge-plates strongly reminds the observer of the ridge-plates oi A. 
planifrons (Fig. 829). Without the aid of a section it is difficult to 
determine whether the ridge-plates are lofty as in Falconer's type of 
A. planifrons (Falconer and Cautley, "Fauna Antiqua Sivalensis," 
1846 [1845, PI. II, fig. 5al— Figs. 825, 828 of present Memoir); in 




A I 1/2 not size ' McQregor Mas. 3920 

Type. Arctitdiskodon subplanifrons. Osborn.i928 

Fig. 875. Now figure of type riglit thinl inferior molsir, r.M.i, of Archidis- 
kodon xubplanifronfi O.slmrn, 1928, from tli<! middle terrace, Sydiiey-on-Vaal, 
South Africa. McOrcKor Museum 3920, Kiml)erley, South Africa; cast 
Amer. Mus, 21924. One-half natural size. 

A, crown view, c — cement, rf -dentine. Line of mid.section. At, 
samp in midsection, exiiibiting .six ridge-plates. 



external aspect the ridge-plates are less elevated. Length of type 
molar 153 mm., breadth lUl mm., index 66; enamel length, re- 
stored, 650 mm., enamel area 2,600 sq. mm.; average enamel 
thickness 4 mm. Ridge-plates (6) postconvex, jireconcave, 5-6 
subo\al to round conelets with double central folds in the ridge- 
plate; ridge-plates per 100 mm. =4. 

Specific Compariso.n with Archidiskodon planifrons in 
THE Brown Siwalik Collection. — The South African type of 
Archidiskodon subplanifrons (Figs. 874, 875) is very similar to 




//// // // %'■); JvV 5,.-; 6. 



7. 5 




the 



1996.5 

Archidiskodon 
pLaniFrons. Ref. 
Falconer. 1845 

Enamel tengtti '810 mm. est. 
Ridge plate httght = 55m 



Fig. 876. Referred third right inferior molar, r.M.-i, of Archidiskodon 
idanifnms Falconer, 184.'), from three miles north of Siswan, India. Sj;, [91 
ridge-plat.es. Iiitenial aspect (below), exhihitiiin the gently folded section of 
the entire molar with relatively low enamel ridge-plat<'s and estimated total 
enamel length of 810 mm. One-half natural size. 

Compare this molar of Archidiskodon planifrons from India with type 
molar of .1. subplanifrons (Fig. 874) from South Africa. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



989 



a referred specimen of A. planifrons (Amer. Mus. 19965) from the 
Upper Pliocene Pinjor horizon, Upper SiwaUks, India. For com- 
parison both teeth are carefully figured to a uniform scale (Figs. 
875, 876). Observe in these teeth: (1) The similar height of the 
ridge-plates, e.g., A. planifrons 55e mm., A. subplanifrons 53e 
mm.; (2) A. planifrons with 8% [9] ridge-plates, A. subplanifrons 
with 8)26 ridge-plates; (3) similar constitution of the enamel and 
cement; (4) similar obhque, outwardly facing, slightly concave 
ridge-plates; (5) the median loops, 'loxodont sinus,' are somewhat 
more prominent in A. subplanifrons than in ^4. planifrons. 

Archidiskodon broomi Osborn, 1928 

Figure 877 

Lower or Middle Pleistocene. Discovered in 1920 at The Bend, on the 
Va.al River, near Kimberley, South Africa. 

History. — Aside from the uncertainty as to the exact loca- 
tion in the jaw of this most interesting type molar of Archidiskodon 
broomi, there is no doubt as to its relationship to the gigantic, 
broad-plated genus Archidiskodon and as to its affinity to A. meri- 
dionalis, the southern mammoth of Europe, as well as to its 
American relative A. imperalor, the imperial mammoth of the 
southern United States and of Mexico. Consequently this is 
a most important and interesting case of geographic distribution 
and of independent parallel evolution of species of Archidiskodon in 
South Africa and southern Europe and in the southern United 
States. 

Archidiskodon broomi Osborn, 1928. "Mammoths and Man in 
the Transvaal." Nature, Vol. CXXI, No. 3052, April 28, 1928, pp. 
672, 673. Horizon and Locality. — Lower or Middle 

Pleistocene. The Bend, on the Vaal River, near Kimberley, South 
Africa. Type Figure. — Op. cil., fig. 2, p. 672. 

Original Type Description of Archidiskodon broomi 
(Osborn, 1928.749, p. 672).— "In the second type (Fig. 2 [ = Fig. 
877 of the present Memoir]), the anterior half of a third superior 
[inferior] molar, probably of Pleistocene age, we observe a far more 
progressive stage, with lofty ridge-plates, the sixth attaining 
a height of 5 inches, equal to that of certain specimens of Archi- 
diskodon meridionalis in which the ridge-plates equal or exceed 5 
inches. This relatively high-crowned type (McGregor Mus., 3682) 
I name Archidiskodon brooyni, in honour of Dr. Robert Broom, who, 
after the specimen was named in MS. and figured by myself, re- 
quested that one of these molars should be named after Mr. W. 
Millett, who discovered the type of A. subplanifrons at a depth of 
from 50 to 60 feet in the Vaal River diggings near Sydney-on-Vaal." 

Type. — Seven-plated anterior portion of an imperfect third 
inferior molar of the right side, bearing the inscription: "3682 
Mus. Kimb. The Bend. H. Else." Cast Amer. Mus. 21907. Re- 
produced from original specimen (Fig. 877 of the present Memoir). 
By comparison with the superior and inferior molars (Amer. Mus. 
14476, 14558) referred to Archidiskodon imperalor (as illustrated in 
Fig. 889 B, A, of the present Memoir), this type grinding tooth of 
Archidiskodon broomi agrees approximately in size with A. im- 
peralor in the maximum width of the fifth ridge-plate, namely, 107 
mm. The ridge-plates of A . broomi also agree exactly in width (107 
mm.) with those of M^ in Brit. Mus. 7436 (see Falconer and 
Cautley, "Fauna Antiqua Sivalensis," 1846 [1847, PL xiv.b, fig. 
13a]), a true Elephas [ = Hesperoloxodon] antiquus molar. 



Specific Characters. — (1) The ridge-plates present an 
anterior concavity and a posterior convexity as in ^. imperalor 
(Fig. 889A) ; (2) the ridge-plates converge towards each other on 
the external side and are somewhat farther apart on the internal 
side, consequently we are inclined to consider this molar as an 
r.]\l2 by comparison with Archidiskodon imperalor (Amer. Mus. 
14558 — Fig. 889 A), but the exact placing of this tooth is uncertain; 

(3) the fact that the fifth ridge-plate preserved is broader (tr. 
107 mm.) than the sixth and seventh ridge-plates preserved 
(tr. 105 mm.) favors the determination of this tooth as a ten or 
eleven plated r.M2 rather than a twelve to fourteen plated r.Ma; 

(4) the maximum width of the fifth ridge-plate or pentalophid is 
107 mm., practically the same as that of the corresponding penta- 




ARCHI0I5K0D0N BROOMI 
Mus. Ktmk 3682 Tyfle 



//. M cast 2/ 90/ 




2 nat. sue 



Osbgrn's Original Type Figure of Archidlskodon broomi 
Fig. 877. Type of Archidiskodon broomi Osborn, 1928. Original in 
the McGregor Mu.seuni, Kimberley, South .Africa (No. 3682), cast Amer. 
Mus. 21907. Third inferior molar of the right side, r.Ms. From The Bend, 
Vaal River, near Kimberley, South Africa. After original specimen kindly 
loaned to the American Museum for figuring and description. Compare 
Osborn, 1928.749, p. 672, fig. 2. 

Seven ridge-plates partly preserved, broadest portion of the crown at 
fifth ridge-plate, indicating that three or four posterior ridge-]jlates are 
missing and that this may represent a ten or eleven plated second molar, r.M2, 
according to the formula; of Leith Adams (1877-1881) and of Weithofer (1890), 
in wliich Archidiskodon meridionalis is clearly distinguished from .1 . planifrons. 



990 



OSBORN: THE PROBOSCIDEA 



lophid in A. iniperdlor (tr. 108 mm.) and the corresponding penta- 
l(tpli in A. men'dioiKtlis (AP); (5) in ,1. iinpenilor (Amer. ]\Ius. 
14476 — Fig. 889B) the convexity of the ridge-plates is anterior, the 
concavity is posterior — therefore, by comparison, the A. broomi 
type is an inferior molar; (6) against this interpretation, however, 
is the fact that in profile (Fig. 877, lower) the wearing surface of the 
ridge-plates is slightly convex (as in superior molars generally), 
whereas, according to rule, the profile of the inferior molars in 
Archidiskodon .should be slightly concave (.\mer. Mus. 10598 and 
14558— Fig. 892 B, A). 

The detailed specific and generic characters and measurements 
are as follows: (1) First and second ridge-plates narrow and com- 
pletely worn into confluence, third ridge-plate (tr. 86e mm.), 
fourth ridge-plate (tr. 102 mm.), fifth ridge-plate (tr. 107 mm.), 
.sixth ridge-plate (tr. 105 mm.), seventh ridge-plate (tr. 105 mm.), 
height of seventh ridge-plate (110 mm.); (2) enamel borders of 
each ridge-plate folded or ptychoid, slightly expanded in the fourth 
ridge-plate on wear into a faint 'loxodont sinus'; (3) the ptychoid 
enamel borders dLstingui.sh the sides of the ridge-plates (Fig. 877, 
lower); (4) broad cement deposit between the ridge-plates, not, 
however, covering the plates at the sides as in /I. imperaior. 

Archidiskodon vanalpheni Dart, 1929 

Figure 878 
From Sydney-on-V'aal, South Africa. Middle terrace (lower levels). 
?Lower Pleistocene. 

Archidiskodon vanalpheni Dart, 1929. "Mammoths and 
Other Fossil Elephants of the Vaal and Limpopo Watersheds." 
So. Afr. Journ. Sci., XXVI, p. 704. Type.— Third superior 

molar of the left side, l.Ml McCiregor Museum (Kimberley) 
4086 (cast Amer. Mus. 22723). Horizon and Locality. — 

"Middle terrace at a depth of 80 feet at Sydney-on-Vaal," South 
Africa. ?Pliocene [?Lower Pleistocene {fide Osborn, 1934.925, pp. 
2, 3)]. Type Figure.— Dart, op. cil., p. 704, figs. 8 and 9. 

Type Description.— (Dart, 1929.1, pp. 704-706): "This 
tooth (greatest length 234 mm., greatest width 112 mm., greatest 
height 129 mm.) has 8 plates and is incomplete posteriorly, there 
having been originally probably 9 plates present (see Figs. 8 and 
9 (Fig. 878 of present Memoir])." 

"The tooth in A. vanalpheni presents a characteristically 
rugged appearance, the massive plates being separated laterally 
l)y great valleculae approaching an inch in depth. These huge 
furrows are continued across the grinding surface of the tooth to 
a depth of % inch or more (save in the case of the space between the 
3rd and 4th plates only) and give to the organ an appearance 
simulating that characteristic of the Stegodonts. Owing to the 
fragility of the specimen, the fangs of the tooth are absent, as 
also are most of the mammillated processes of the posterior four 
plates. The first or anterior plate is talon-like in appearance, 
being aborted and only half the height of the succeeding one, to 
which it is firmly adherent, with the interposition between them of 
very little cementum." 

"The massive plates are markedly cuneiform, as opposed to the 
chevron-patterned plates of Stegodonts on the one hand, and the 
pectinated arrangement exhibited by more atlvanced elephants on 
the other hand, their antero-posterior witlth tapering from 1% 
inches at their base to %-'/i inch at the grinding edge. The contrac- 



tion of the plates is not less marked in their transverse length which 
at the ba.se is in the region of 4 inches, hut at the grinding eilge is 
not more than 2}i inches or 63 mm. These features lend to the 
tooth as a whole a pyramidal appearance which is in sharp con- 
trast to the squat semicuboidal form and exuberant cementum 
found in .4. subplanifrons. The enamel in thickness (3-4 mm.) is 
comparable with that of A. subplanifrons but is somewhat more 
folded than in that form, definite longitudinal depressions running 
e\'en along the exposed lateral and medial sides of the plates. It 
is these enamel foldings which more jiarticularly give the tooth its 
rugged character." 

"Owing to the early age of the tooth, not more than three 
plates (excluding the most anterior talon) being in wear, a clear 




Fig. 878. Type left third superior molar, l.M' (McGregor Mu.s. 4086, 
ea.st Amer. Mu.s. 22723), of Archidiskodon vanalpheni Dart, 1929, figs. 8 and 9, 
p. 704, one-third natural size. From Sydney-oii-Vaal, South Africa. 



jiicture of the grinding surface in full wear is not a\ailable. In 
the fourth plate, 6 mammillated ])oints are in slight wear, in the 
third there are 2 islets and 3 mammillated proce.sses — indicating 
in all apparently 7 mammillated processes to each plate, a number 
which appears to be ratified in the posterior plates which ha\'e 
lost their processes by fracture — while the second j)late shows four 
islets. A tendency is demonstrated especially by the anterior 
talon and the second and third plates towards their centre to 
throw out a posterior buttress. There is no evidence whate\er of 
an anterior buttress." 

"The tooth is distinguished also by the great degree of sep- 
aration between the plates owing to the presence of great wedges 
of interlamellar cementum desjiite their valleculisation. This 
.separation, which in A. subplanifrons does not exceed %-% in., is 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 



991 



%-Va in. in A. vanalpheni. There can be little doubt that we 
possess here another very ancient Archidiskodont species, which 
despite its greater number of ridge plates (A. siibplcinifrons shows 
only 5 plates but may have had more) illustrates in its cement 
characters and the ridge plate form an exceedingly primitive type. 
I have named it after Mr. van Alphen who was responsible for 
securing it for scientific investigation." 

Archidiskodon milletti Dart, 1929 
Figure 879 
From Sydney-on-Vaal, South Africa. Middle torrace (lower levels). 
?TyOwer Plei.stoocno. 

Archidiskodon milleili Dart, 1929. "Mammoths and Other 
Fossil Elephants of the Vaal and Limpopo Watersheds." So. 
Afr. Journ. Sci., XXVI, p. 706. Type. — Third superior 

molar of the left side, l.M^. McGregor Museum 4085; cast 
Amer. Mus. 22722. Horizon and Locality. — "Middle ter- 

race at a depth of 80 feet at Sydney-on-Vaal," South Africa. 
?Pliocene [?Lower Pleistocene {fide Osborn, 1934.925, p. 3)]. Type 
Figure. — Dart, o-p. cit., p. 706, figs. 10 and 11. 

Type Description.— (Dart, 1929.1, pp. 707-708): "In 
general measurements (greatest length 222 mm., greatest width 
108 mm., greatest height 118 mm.) it corresponds fairly closely 
[with A. vanalpheni] being, however, considerably .shorter, wider 
and low'cr than the foregoing. These measurements are fairly 
comparable with those given ior A. vanalpheni, seeing that here, too, 
the fangs are absent; here, too, eight plates are present including 
a similar anterior talon, which in this case, however, more closely 
approximates in height that of the second plate, and is actually in 
wear. Here also one or more posterior plates are absent by loss. 
(Figs. 10, 11 [Fig. 879 of present Memoir])." 

"But it is on still closer examination that distinctions are to 
be drawn between the teeth. While in this tooth, too, the indi\i- 
dual plates are cuneiform, the disparity between their thickness 
at base and edge is not so great, in that the plates are not more 
than % inch thick at their bases and are %-}2 inch thick at their 
edges. The degree of contraction in transvei-se length of plates 
from base to edge, is, however, virtually identical with that in 
A. vanalpheni." 

"The enamel is just as thick as that of A. vanalpheni and while 
it displays some crimping is not nearly so rugged in appearance as 
that of A. vanalpheni, while on the exposed sides of the plates 
there is no trace whatever of longitudinal grooving. Owing to the 
early age of this tooth also, which is practically identical in this 
respect with that of the preceding, there is no clear picture of its 
surface in wear. The anterior talon is in wear and shows a median 
posterior buttress as also do the second, third and fourth plates. 
The fifth, seventh and eighth plates show five digitations, the 
sixth plate six, the fourth plate six digitations and a posterior 
buttress digitation, the third plate seven worn digitations and 
a posterior buttress digitation, and the second plate two islets, the 
posterior buttress protruding from the medial one of the two 
islets. There is no evidence of any anterior buttress in any of the 
plates." 

"The degree of separation between the second and third 
plates, 'A-V/i in., is identical with what is typical for A. vanalpheni, 
but the other plates are not nearly so widely separated being 



generally only 'i inch apart, so that in .4. tnilletli a length of 5)^ 
inches includes 5 plates while in A. vanalpheni it includes but 4 
plates." 

"When in addition it is pointed out that, while there is here, 
too, an appreciable vallecuHsation of the interlamellar cementum — 
which is continued across the tooth between the posterior plates — 
there is, nevertheless, in A. milletti a far greater relative bulk of 
cementum and a far greater tendency towards the reduction of the 
valleys, it is clear that we are confronted with an advancing and 
specificall}^ distinct though closely related form of Archidiskodont. 
I have named it after Mr. Millett who has been responsible for the 
finding of the three mammoth teeth reco\ei'ed at Sydney-on-Vaal." 

"There remains, of course, the possibility that one of these 
forms with a well-marked posterior buttress is closely related to 




Fig 879. Ty|x> left third superior molar, l.M^ (McGregor Mus. 4085, 
cast Amer. Mus. 22722), of Archidiskodon milleili Dart, 1929, figs. 10 and U, 
p. 706, one-third natural size. From Sydney-on-Vaal, South Afri<'a. 



if not identical with, ^1. griqua of Haughton. I have had the 
privilege personally of examining that well-worn and fragmentary 
specimen, and, with a view to making our knowledge of the South 
African mammoths as complete as possible, have taken the 
liberty of making some additional remarks upon the material." 



Archidiskodon loxodontoides Dart, 1929 
Figure 880 
From Sydney-on-Vaal, South Africa. Middle terrace (lower levels). 
? Lower Pleistocene. 

Archidiskodon loxodontoides Dart, 1929. "Mammoths and 
Other Fossil Elephants of the Vaal and Limpopo Watersheds." 
So. Afr. Journ. Sci., XXVI, p. 709. Type.— Third superior 

molar of the left .side, l.M'. McGregor Mu.seum (Kimberley) 



992 



OSBORN: THE PROBOSCIDEA 



4087; cast. Amer. Miis. 22724. Horizon and Locality. — 

"Middle terrace at a depth of 80 feet(?) at Sydney-on-Vaal," 
South Africa. ?PHoceiie [?Lower Pleistocene, fide Osborn, 1934. 
925, p. 3]. Type Figure.— Dart, op. dt., p. 709, fig. 13. 

Type Description.— (Dart, 1929.1, pp. 709-711): "This 
tooth, also apparently a third upper molar (greatest length 149 
mm., greatest width 94 mm., greatest height 112 mm.), is less 
complete than the previous two new types described, presenting 
only 4% in. plates and lacking in addition one or more plates both 
anteriorly and posteriorly. It is impossible to state whether it 
j:)ossessed originally a ridge plate number as great as that of the 
two preceding — it probably did not. All the plates present were 
in wear and the specimen was probably from an old adult. For an 
upper molar in early wear it is not higher than one would e.xpect in 
A. subplanifrons, but it is narrower than that specimen and, 
indeed, than all the aforementioned forms. Here also the fangs 
are absent and the measurements of height consequently com- 
jiarable (see Fig. 13 [Fig. 880 of present Memoir])." 




Fig. 880. Type left third superior molar, l.M^ (McGregor Mus. 4087, 
cast Amer. Mus. 22724), of Archidiskodnn Inxodontoides Dart, 1929, fig. 13, [). 
709, oiic-tliird luttura! size. From Sydncy-oii-V'aal, South Africa. 



"The plates are not so distinctively cuneiform as in the two 
previously described forms, nor do they present so distinct a de- 
gree of contraction in length toward the grinding surface where 
they show a transverse length of 80 mm. The enamel is thick 
(3-4 mm.) and is very markedly plicated or crenated, despite its 
thickness. Longitudinal grooving of the medial and lateral 
aspects of the ridge plates is also strongly in evidence. Owing to 
the well worn nature of the tooth the pattern produced on the 
grinding surface is very clear. The most posterior or fifth plate 
shows 7 mamniillated processes which have all been broken off, 
the fourth shows eight pedunculated processes in wear, the third 
and fourth from the medial side having fused to form a small 
islet. In the third plate there are three islets, one formed by the 
fusion of the medial two mamillae and one islet by the lateral two. 
The four central mamillae are fused to form a large central islet 
and in the region corresponding to the medial two of the four 
central mamillae a pronounced anterior and posterior dilatation 



is evident, due to the presence of distinct buttresses here on both 
the anterior and posterior aspects of the plate. In the second and 
first plates the islets have completely disappeared, and in the 
second plate especially there is shown a virtually median widening 
out with the full development of the buttresses." 

"The plates, which are roughly 19 mm. wide in their more or 
less parallel portions, expand to 31 mm. in the regions of the 
buttresses with the result that the buttresses of adjacent plates are 
in direct contact with one another and together separate the inter- 
lamellar stratum of cementum, 10-17 mm. wide, on one side from 
that on the other side of adjacent buttresses, as occurs in the true 
Loxodonts, a condition not previously encountered to my knowl- 
edge in any Archidiskodont, and justifying the name given to this 
peculiar species." 

"The cementum is abundant and almost encases the ridges 
completely on the medial aspect of the tooth. Laterally, however, 
there are distinct and deep trenches ]{-% inches in depth between 
adjacent plates, indicating the probable origin of this species from 
a form comparable with those previously described, especially 
A. vanalpheni. Indeed, if it were not for the greatly reduced 
width of the tooth and reduced width of interlamellar cementum 
as well as the presence of the anterior buttress in A . luxodonloides, 
I should have been considerably more reluctant to separate it 
specifically from A. vanalpheni. It must be closely related to that 
species and the recovery of further remains of both types, especially 
a well worn molar of A. vanalpheni, will be awaited with great 
interest in this connection. It may be repeated, however, that 
there is no evidence of such an unusual anterior buttress in A. 
vanalpheni and the resultant loxodontoid form of plate in surface 
wear as exists in A. loxodo7itoides. The same feature serves to 
distinguish it equally from the A. griqua of Haughton." 



Archidiskodon yorki Dart, 1929 

Figure 881 

Vanasswegen.shoek-Bloemheuvel, near Christiana, South .Africa. Middle 
terrace (upper levels), ?Lower Pleistocene. 

Archidiskodon yorki Dart, 1929. "Mammoths and Other 
Fossil Elephants of the Vaal and Limpopo Watersheds." So. Afr. 
.lourn. Sci., XXVI, p. 717. Type. — ". . . two very frag- 

mented enamel plates of an Archidiskodont mammoth" {op. cit., 
p. 715). McGregor Museum (Kimberley) 4073. Horizon 

AND Locality. — "Middle gravel stratum of river bed gravels at 
Vanasswegenshoek-Bloemheuvel, near Christiana," South Africa. 
TMiddle Pleistocene [?Lower Pleistocene, ^rfe Osborn, 1934.925, p. 
3]. Type Figure.— Dart, op. cit., p. 717, fig. 19. 

Type Description.- (Dart, 1929.1, pp. 717, 718): "The 
plates of this tooth are of great size (fragment is 114 mm. broad 
and 193 mm. high, but the widest diameter of the tooth was at 
least 120 mm. and the height at least 200 mm.). In all their 
mea,surements these plates (see Fig. 19 [Fig. 881 of ])resent 
Memoir]) exceed considerably those of A. hroojni . . . The enamel 
plates measure 16-23 mm. across, that is, slightly narrower than 
in A. broomi, and ajjproach cio.sely in form the type presented by 
A. broomi. Laterally \iewed, howe\'er, there is a distance from 
mid-point to mid-point of succeeding enamel ridge plates in A. 
broomi of at least 32 mm. whereas in this specimen the .same 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



993 



measurement does not exceed 24 mm. thus demonstrating the 
relatively decreased amount in it of the interlamellar cementum. 
We are, therefore, dealing here with a broader and higher tooth 
than in the type of A. broonii and one with slightly narrower 
enamel plates and decreased interlamellar cementum. It must be 
specifically different but closely related thereto and I will name it, 
despite the paucity of the remains, Archidiskodon yorki in honour 
of its discoverer." 

"The finding of A. hroomi and A. yorki, which are advanced 
Archidiskodont types in gravel strata which appear to be directly 
capable of correlation with the advanced Archidiskodont gravel 
at Bloemhof, is of great interest and importance. Both of these 
mammoths are admittedly more i)rimiti\'e than the Bloemhof 
mammoths, although more advanced than A. subplanifronti, which 
came from the 50-60 feet dejjth at Sydney-on-Vaal." 

"There is no question in my mind, from the state of fossilisa- 
tion of A. yorki, from its fragmentation and from its unnatural 
stratigraphical jjosition as compared with the Bloemhof mam- 
moths, that it belongs to the more superficial part of the middle 
terrace, as I have already suggested is true also of .4. hroomi." 




Fig. 881. Type molar fragment of Archidiskodim yorki Dart, 1929, fig. 
19 (McGregor Mus. 4073), found in river bed gravels at Vanasswegenshoek- 
Bloemheuvel, near Christiana, South Africa. One-half natural size. 



SPECIES PROVISIONALLY REFERRED BY THE PRESENT 
AUTHOR TO THE GENUS PAL.a;OLOXODON 

Archidiskodon transvaalensis Dart, 1927 

From Bloemhof, Vaal River, South Africa. Pleistocene. 
A third superior molar, r.M^, described by Doctor Dart in 
1927 as Archidiskodon transvaalensis, was removed by Professor 
Osborn to the genus Palseoloxodon (see Osborn, 1934.925, pp. 2 
and 14). This species is treated fully in Chapter XIX, p. 1284, 
below (the Loxodontinse). 



Archidiskodon sheppardi Dart, 1927 
From Bloemhof, Vaal River, South Africa. Pleistocene. 
The species Archidiskodon sheppardi, type an l.M^, also de- 
scribed by Doctor Dart at the same time as his species A. trans- 
vaalensis, was regarded by Professor Osborn as belonging to the 
genus PalivoloTodon and as a consequence is treated in full in 
Chapter XIX, p. 1278 below (the Loxodontinse). 



Archidiskodon andrewsi Dart, 1929 
From Gong-Gong, ? Middle terrace (lower levels), Vaal River, South 
Africa. ?Lower Pleistocene. 

The type of Archidiskodon andrewsi Dart is a fragmentary 
third lower molar of the left side, I.M3 (McGregor Mus. 435; cast 
Amer. Mus. 26968). It was first thought by Professor Osborn to be 
referable to Archidiskodon subplanifrons, but on further study of 
the original specimen kindly loaned by Curator Wilman of the 
McGregor Museum he was inclined to regard it as belonging to the 
genus Palseoloxodon (see Osborn, 1934.925). The description and 
figure, therefore, will be found below on page 1278 of Chapter XIX 
(the Loxodontinse). 



Archidiskodon hanekomi Dart, 1929 

From the old river bed of the Vaal River at a depth of 20 feet at Deli)oort's 
Hope, South Africa. 

In 1929 Dart described a ?third right upper molar, r.M', 
in the McGregor Museum (No. 2930), which he named Archidisko- 
don hanekomi. Professor Osborn, however, in his article "Primitive 
Archidiskodon and Palaeolo.rodon of South Africa" (Osborn, 
1934.925) referred this species to the genus Palxolo.rodon. The 
description and figure will be found on page 1279, figure 1140 of 
Chapter XIX (the Loxodontinse). 



Genus: METARCHIDISKODON Osborn, 1934 

Original reference: "Primitive Archidiskodon and I'alaeoloxoilon of Soutii Africa," Anier. Mils. Novitates, No. 741, August 24, 
1934 (Osborn, 1934.925, p. 12). 

Genotypic species: Loxodonta griqua Haughton, 1922. 

Generic Characters. — (Osborn, 1934.925, p. 12): "This group [Metarchidiskodon griqua group] 
includes the fragmentary type (Fig. 3 [ = Fig. 882 of present Memoir]) of 'Loxodonta' griqua Haughton, 
1922. . . . This specimen appears to belong to a distinct form of grinding tooth to which the new generic 
name M elarchidiskodon may be applied, and distinguished from Archidiskodon as follows: (1) M 3 with 
a relatively long narrow crown; index cannot be estimated at present. (2) Deep U-shaped valleys filled 
with cement. (3) Enamel ridge plates extending to the bottom of the crown. (4) Very prominent 
post-sinus folds instead of median sinus expansion of the typical Archidiskodon. . . . Type species: Loxo- 
donta griqua Haughton, 1922." See Table IX above (p. 985) under "M. griqua group." 

Specific Characters of Metarchidiskodon griqua, by Osborn, 1934. — (Osborn, 1934.925, p. 12): "(1) 
Cement areas equal or exceed dentine areas; (2) pre-sinus folds absent or inconspicuous; very prominent post- 
sinus folds; (3) very deep U-shaped valleys extending to the bottom of the crown; this is a very important point. 
(4) These valleys are filled to the summit with cement. (5) Enamel ridge plates very deep, extending to the bottom 
of the crown, closely compressed with very narrow dentinal areas between." 

"Type figure 3 [ = Fig. 882 of present Memoir] to be compared with the relatively narrow grinding teeth of 
similar molars observed in the Val d'Arno specimens and in British Mu-seum M12641, M12642." 



Metarchidiskodon griqua Haughton, 1922 
I'igurcs 882, 883 

\'aal Rivor gravels, Griqualand West, Soutli Africa. ?Lo\ver Pleistocene. 

From these gravels Fraas (1907) described Equun cf. zebra, 
Hippopotamus amphibius var. rohuslus, Damaliscus sp. and Masto- 
don sp. ; Felix (see Beck, 1906.1, p. 49) described Mastodon 
(Biinohphodon) sp., and Haughton (1922) described a new giraffe 
{Griquatherium cingulatum). 

History. — Originally described by Haughton in 1922 as 
Loxodonta griqua this extremely fragmentary superior molar proves 
to belong to a new genus, Metarchidiskodon. The new type 
figure (Fig. 882, right) is reconstructed from the original specimen 
(McCiregor Mus. 3686) kindly loaned by Miss Wilman, Director 
of the Museum. It apparently represents three central ridge- 
plates ["Supposed third, fourth and fifth" (fide Osborn 1934.925, 
p. 8. fig. 3)] of a third left upper true molar (?1.M^), as compared 
with .'i(c/iK//.sA-of/o/i planifrons (Fig. 829); the elevation of the ridge- 
plates and approximation of the valleys are less extreme than in the 
type of A. planifrons fFig. 825); the more worn anterior plate is 
to the left, the less worn posterior plate is to the right; the concave 
side of each plate opens backward as in A. planifrons; the back- 
ward enamel loops are much more prominent than in A. subplani- 
frons; width of central ridge-plate 90e mm., height of same ridge- 
plate 90e mm. 

Loxodonta (iriqna Haughton, 1922. "A Note on Some Fossils 
from the Vaal Ri\cr (iravels." Trans. (Jeol. Soc. S. Africa, 1922, 
\'ol. .\XIV, ])p. 11 13. Typk.- -I'Vagmentary molar. Hoht- 
ZON AND Locality. — From the river gravels of the Vaal Kivcr, 
Griqualand West {op. vit., p. 11, (|Uoted from du Tdit, llth .\nii. 



Rept. Geol. Gomm., 1907, pp. 171, 172) : "The gravels are situated 
at \'arious levels as well as at varying distances from the river, 
and though they form a number of fairly distinct terraces, it 
is not always easy to determine their relati\'e age. ... On the west 
of the Vaal the terrace is finely developed at Klipdam, where it has 
an altitude of 200 feet above the river and a distance from it of 
3}2 to 6 miles." (Osborn, 1934.925, pp. 2, 3) ?Middle terrace 
(upper levels), 60-80 feet, ?Lower Pleistocene. Type 

Figure. — Haughton, op. cit., 1922, PI. i, figs. 1 and 2. 

Type Description. — (Haughton, op. cit., 1922, p. 12): ". . . 
a portion of an Elephant molar. . . . The fragment consists of 
three plates of which one side is imperfect and show.s a vertical 
longitudinal section. . . . Although the molar is incomplete, it was 
obviously large and broad and fairly low. The enamel is thick, the 
borders show no strong crimping but a considerable amount of 
unevenness. The cement wedges are large, the plates are wide 
apart and were certainly few in number. In vertical section the 
tooth has an appearance intermediate between those of E. africanus 
and E. planifron>< figured by Falconer and somewhat similar to that 
from the Red ( 'rag of Suffolk assigned by Leith Adams to E. anti- 
quus (Brit. F'oss. F^lephants, PI. xxvi., fig. 3 [2|). The cement 
wedges between the plates are as thick as the plates; and their 
bases are more rounded than wedge-shaped. The sides of the 
plates are more parallel than in planifronn, thus ap])roximating to 
those of africanu.s. The most characteristic feature of the tooth, 
however, is the shape of the wearing surface of the plates. ... [p. 
13] The present tooth seems to differ considerably from that of 
E. Zulu and also from that of E. antiqutts. These show a central 
angular notch on the enamel, but never to the same extent; while 
the strong crimi)ing of the enamel in both these forms is a distin- 



994 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



995 



guishing feature, as well as the comparative lowness and width of 
the tooth in this fragmentary specimen." 

Specific Characters of Archidiskodon [ = Metarchi- 
diskodon] griqua, by Osborn, 1928. — (1) Superior ridge-plates 
much more elevated, i.e., 3Ke in. =90e mm., than in the type of 
A. subplaiiifrons, i.e., 2}2-2% in. = 64-70 mm. (2) Central enamel 
loops much more prominent than in A. subplanifrons. (3) Width 
of central ridge-plate 90e mm. ; height of central ridge-])late 
90e mm. 

Kaiso Bonp>beds (Hopwood, 1926, pp. 33, 34). — From the 
Kaiso Bone-beds, near Albert Nyanza, Hojjwood describes a frag- 
mentary tooth (Brit. Mus. I\I12641), see figure 883, as follows: 
"The chief specimen consists of three lamellae and the loxodont 
sinus of a fourth. At the anterior end is part of the posterior wall 
of a fifth plate which has been deejily abraded. The one behind it 
has been worn in such a manner as to reveal the great depth of the 
digitations. This sjjecimen has been cut across in order to show 
the enamel figures [Fig. 883, lower]. . . . These characters, taken 
together with the lamellar frequency of four, indicate a very 
primitive form, which might be compared with E. pla/iifrons 
Falconer." 

External LM^ 
Y2 nat. size 






McQregorMus. 5686 

Type. Loxodonta griqua 
Haughton. 1922 

Type Third Left Superior Mol.\r ok Metarchidiskodon griqua 
Fig. 882. Original (upper,},4 lowor,}^ nat. size) and new (riglit pair,K nat. 
size) type figures of Loxodonta griqua Haughton, 1922, PI. i, figs. 1 and 2; both 
drawings after fractured type specimen (?1.M') in the McGregor Museum at 
Kinil)erley (McGregor Mus. 3686). From the river gravels of the Vaal, Gri- 
qualand West, South .\frica. Original type figures (left): ''Fig. 1.— Cirindiiig 
surface of fragmentary molar. . . . Fig. 2. — Profile view of same." 

(Right) New figure after Osborn, 1934.92.5, p. 8, fig. 3: Suppo.scd third, 
fourth, and fifth ridge-plates of an I.M^. Observe very deep U-shaped valleys, 
thick simply folded enamel with very prominent looped post-sinus folds. 




Srzt.Mtis M./Z6^f 




FIG. \i.—Elephaa afl. mcridionalU Nesti. Enamel pattern of larger 
frnnrneiit. Rcc'i. Ml'26-ll. 1/-2 iml. .?i/,c. 





FIG. H.—Elephas aff. meridioyxalis Nesti. Development of the enamel 
pattern in an isolated plate. A is 13.5 mm. from the top; B 
6 mm. from A: C, 8 mm. from B. und D 12 5 mm from C 
Begd. Mia642. 1/2 nat. size. 

Fig. 883. Referred Metarchidiskodon griqua (non meridionalis) 

(Upper) A fragmentary molar from the Kaiso Bone-beds of Africa, doubt- 
fully referred by Hopwood to Elephas [ = A rchidiskodoyi] meridionalis but which 
proves to be closer to Metarchidiskodon griqua. After photograph kindly fur- 
nished the present author by Dr. A. Tindell Hopwood (cf. Hopwood, 1926, PI. 
Ill, fig. 2). Brit. Mus. M12641. Natural size. 

(Middle) Mid-section of the worn coronal surface. Aft(-r Hopwood, 1926, 
fig. 13, p. 34. One-half natural .size. Brit. Mus. M 12641. Ob.serve the very 
prominent median 'loxodont sinus.' 

(Lower) Summit of the worn coronal surface. After Hopwood, 1926, 
fig. 14, p. 34. One-half natural size. Brit. Mus. M12642. 



996 OSBORN: THE PROBOSCIDEA 

This tooth, the enamel ])Iate characters of which arc best, lameHae 100 nun., breadth 72 nun., height 105 mm., as compared 

shown in figure 883 (lower), appears to resemble Metarchidi.skodon with the following corresponding measurements in the type of 

griqua more closely than Archidiskodon subplanifrons, because the M. griqua, breadth of center ridge-plate 90e mm., height 90e 

enamel ridge-plates are more elevated, the measurements given mm. This ridge-j)latc height exceeds that of the fifth ridge-plate 

by Hopwood being {op. cit., 1926, table, p. 35), length of four of /I. AM67J?nM7/70«s, namely, 53e mm., with a breadth of 92 mm. 



3. ARCHIDISKODONTS OF THE UNITED STATES AND MEXICO 
This noble genus appears to have been given origin in Africa, then to have migrated to Eurasia along the 
south temperate parallels, and finally to have reached its climax of evolution in the United States from Nebraska 
southward to the high plateau of Mexico^ 

Two quite distinct lines of descent, referable to Parelephas and Archidiskodon respectively, are found in the 
United States and Mexico, in which the low ridge formulae are very similar, namely: 

Elephas [ = Parelephas] coluinbi Falconer, 1857-1868. Upper Pleistocene of southern United States and of Mexico. 

Smaller animal than Archidiskodon, witii narrower grinders (Fig. 887), thin cement outer coating; maximum ridge- 

1 8-1 9 



plate formula, M 3 ^-grj 



i+- 



Elephas [ = Archidiskodon] imperalor, Leidy, 1858. Lower- Pleistocene of southern United States and of Mexico. 

Larger animal in size, with broader grinders (Fig. 887), very broad enamel plates, and heavy cement outer coating; 
ridge-plate formula, M 3 J^rH . 

Geologic Age. — Falconer erred in considering these animals of the same species; they are really very 
distinct. They were not geologically contemporaneous; to our knowledge the true Archidiskodon imperator and 
true Parelephas columbi are not found in the same horizons. A. imperator is a late Pliocene and early Pleistocene 
species (Nebraskan and Aftonian glacial stages, cf. Hay, 1923, p. 15), while P. columbi is probably a late Pleisto- 
cene species (during lowan to Wisconsin stages, abundant in the phosphate beds of South Carolina and Florida, 
cf. Hay, op. cit., p. 431, map 12, also pp. 430 and 155). The true P. columbi of Georgia, South Carolina, and 
Florida are not to be confused with the Parelephas jeffersonii of the northern states (cf. Hay, op. cit., map 12, p. 
431, error). Gidley held that A. imperator and P. columbi were contemporaneous. 

GiDLEY, 1928. — Gidley observes (letter, December 6) : "There can be no question that remains of Archidisko- 
don imperator and Parelephas columbi, as these two species are now defined by Osborn [in the present Memoir], 
are found associated in the No. 2 beds of Sellards at Melbourne and Vero, and it is equally certain, I think, that 
these species were contemporaneous in Florida. From my work at Melbourne and Vero last winter I conclude 
that the 'No. 2' deposit is almost entirely of wind-blown sand origin, slowly accumulated, and that stream action 
had little or nothing to do with its formation. On this theory mixtures of material of different periods would not 
be possible. ... I believe it . . . probable that descendants of A. imperator existed in Florida as well as in Mexico 
and the south, until Upper Pleistocene times." Recently described from Melbourne and Vero, Florida, by 
Gidley, are: 

Parelephas floridanus: Nat. Mus. 118U6, 118U8, and 11810 (sec Parelephas, Chap. XVII, footnote, p. 

1078). 
Archidiskodon imperator: Nat. Mus. 11805, 11814, and 11G20 (see below, p. 1005). 



'[For Professor Osborii's final opinion on the migration of Archidiskodon from the V'aal Rivor of South Africa to the Niobrara Kivor of Nebraska, see 
Chap. XXIII, p. 1580.— Editor.] 

-[Early and Middle Pleistocene (Jide Lugn and .Suhullz, 1934.1, pp. 373-37(5). — Editor.) 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 



997 



ORDER OF DISCOVERY IN AMERICA OF ARCHIDISKODON AND FARELEPHAS 
Since the opening sections of the present chapter were written several years ago, our knowledge of the two 
distinct lines of Mammontine descent found in the United States and Mexico has been greatly enriched by three 
more or less complete fossil skeletons of Parelephas discovered in Florida, also by intense and more extended 
observation on the differences between the Parelephas and the Archidiskodon crania as well as by careful compari- 
son of the Parelephas columbi grinding teeth with those of Archidiskodon imperator. 

Falconer's type of 'Elephas' columbi is the middle portion of a third inferior grinder in which the ridge-plates 
are very far apart, thus resembling the widely separated ridge-plates of Archidiskodon imperator; he concluded, 
therefore, that Leidy's 'Elephas^ imperator was the same species as his own 'Elephas' columbi. Osborn was long 
misled by these widely separated ridge-plates and erroneously concluded that the type of 'E.' columbi belonged in 
the genus Archidiskodon; whereas the newly discovered materials above mentioned prove that 'E.' columbi 
belongs within the phylum Parelephas. But, as will be shown in a subsequent chapter (Chap. XVII), the species 
Parelephas columbi Falc, 1857, is clearly distinct from the species Parelephas jeffersonii Osborn, 1922. Thus the 
two quite distinct lines of descent found chiefly in the southern United States' and in Mexico appear as follows in 
the order of discovery : 



Archidiskodon Phylum 

1858 Elephas imperator Leidy, ^ehTaska. = Archidiskodun im- 
perator 

1915 Elephas hayi Barbour, Nebraska = 4 rt'/i?f//sA'odoH hayi 

1922 Elephas Columbi var. silvestris Freudenberg, Mexico 
= Archidiskodon imperator silvestris 

1922 Elephas Columbi var. Falconeri Freudenberg, Mexico 
= Archidiskodon imperator falconeri 

1925 Elephas scotti Barbour, Nebraska = Archidiskodon impe- 
rator scotti (or juvenile A. imperator) 

1925 Elephas (Archidiskodon) niaibeyii Barbour, Nebraska 
= Archidiskodon imperator maibeni 

1928 Elephas haroldcooki Hay, Ok\&homa. = Archidiskodon har- 
oldcooki 

1928 Elephas exilis Stock and Furlong, California = .4 ;y/u'- 

diskodon e.rilts 

1929 Archidiskodon sonoriensis Osborn, Mexico = Archidiskodon 

sonoriensis 

Very broad grinding teeth; thick enamel ridge-plates; 
heavy cement outer coating ; enamel ridge-plates widely separated 
from the wearing surface of the crown downwards to the face of the 
plates. Ridge-plate formula not known to exceed M 3 iv. 



Parelephas Phylum 
1838 Elephas jacksoni Mather, Ohio = Parelephas jarksoni{'!) 

1857 Elephas columbi Falconer, Georgia. = Parelephas columbi 

1859-1861 Elephas texianus Owen-Blake, Texas = Parelephas 
columbi 

1922 Elephas Columbi var. Felicis Freudenberg, Mexico = Parele- 
phas columbi felicis 

1922 Elephas jeffersonii Osborn, Indiana = Parelephas jeffersonii 

1927 Elephas roosevelli Hay, Il\iDois = Parelephas jeffersonii 

1929 Parelephas floridanus Osborn, Florida = Pa?-ctep/ias Jlori- 
danus 

1929 Parelephas columbi cayennensis Osborn, French Cuiana, 
South America = Parefep/ios columbi cayennensis 

Narrower grinding teeth; thinner enamel ridge-plates; thin 
cement outer coating; enamel ridge-plates widely .separated or 
arcuate at the base (as in type of 'E.' columbi), more closely con- 
vergent at the summit; ridge-plates increasing from M 3 j^ 
(P. columbi) to M 3 Iff (P. progressus). 



It would be natural to suppose that from the progressive A. meridionalis stage these southern manmioths 
migrated across Asia and through the United States southward, but the present palseontologic and geologic evi- 
dence does not support this direct derivation of the imperial mammoth from the progressive A. meridionalis type,- 

'[In 1932 Professor F. H. Edmunds of the Department of Geology of the Univeraity of Saskatchewan forwarded to the American Museum for identi- 
fication a proboscidean molar from the vicinity of Wiscton, Saskatchewan, which proved upon examination by Dr. Edwin H. Colbert of the Museum staff to 
be referable to Archidiskodon imperator. This is important as establishing the most northerly range of the genus in North America. — Editor.] 

'-(Subsequently Professor Osborn became convinced that the "Elephas meridionalis" of France was the direct ancestor of the "Elephas imperator" of 
North .America, a conclusion arrived at through the discovery of a skeleton (named by Osborn Archidiskodon meridiorMlis nebrascensis), foimd one mile north- 
west of Angus, Nuckolls County, Nebraska (see Osborn, 1932.893, p. 1): "New and positive evidence of the correctness of this theory is now afforded by the 
discovery of the complete skeleton which forms the subject of the present paper. This skeleton with the lower jaw in a complete state of preservation proves to 
resemble very closely indeed in every detail the 'Elephas meridionalis' of Durfort, France, as fully described by .-Mbcrt Gaudry." — Editor.] 



998 



OSBORN: THE PROBOSCIDEA 



because since the year 1915 there have been discovered at least two specific stages very much more primitive than 
A. imperator. These are the Archidiskodon hayi Barbour, from near Crete, Saline County, Nebraska, also the 
Archidiskodon sonoriensis from Sonora, Mexico, with its prominent elongate protuberance of the rostrum (Fig. 923, 
cf. Fig. 903). Of these relatively primitive forms, A. hayi (Fig. 913, lower) has an elongate mandibular ramus, 
with a third inferior molar exhibiting from ten to eleven ridge-plates set very far apart. This not only appears 
to be the most primitive species of proboscidean thus far discovered in America, but it seems to resemble very 
closely the primitive lower jaws (Fig. 849) of A. planifrons of southern France and of India. 

The Archidiskodon sonoriensis is a far more progressive species than A. hayi, but it exhibits an elongated 
mandibular rostrum of a much more primitive character (Fig. 923) than the brevirostral mandibular symphysis 
of A. imperator (Fig. 892 A, B, 898 A, Al). 

These discoveries render it highly probable that the first migration from Eurasia to America occurred during 
the Upper Pliocene Archidiskodon planifrons phase; they also render it possible that the typical A. planifrons 
of the Pinjor horizon of India migrated to America at about the same period, and perhaps as a traveling companion 
of the ancestors of Stegomastodon mirijicus. 



SYSTEMATIC DESCRIPTION OF SPECIES 



Archidiskodon imperator Leidy, 1858 

Figures 80o, 810. 812, 815, 817, 818, 805, 884-889, 891, 892, 894- 

902, 906-909, 912, 915, 916, 920, 937, 1030, 1226, 1231, PI. xxi 

Lower Pleistocene, ('i)lsl Interglacial, or (?)Aftonian age. Central and 
Southern United States, 40th parallel southward into Mexico. Osborn, 1921- 
1924: The unrecorded level is probably equivalent to the Equus beds of Cope, 
thv Equui excelsus-Elephas imperator zone of Osborn (Osborn, 1918.473, p. 34). 

Hay, 1924, p. 100: "Seneca, Thomas County (6 [see map, Fig. 8901).— 
Somewhere along Middle Loup Fork River, probably in Thomas County, was 
found, by F. V. Hayden's party, the tooth which forms the type of Elcphas 
imperator. A more exact locality has not been determined. In the region 
about Seneca, as reported to the writer by Dr. W. D. Matthew, were found, in 
1916, remains of Equus, Camelops, Platygnnus, Canis, etc. Here, too, Hayden 
probably foimd tlie ty|ie of Stegomastodon mirificus."^ 

Synonyms or related forms (Mexico): El. Volumbi var. silvestris Frcud- 
enb., 1922, Oaxaca, Mexico; El. Cnlumbi var. Falrxmeri Freudenb., 1922, 
Tecpiixquiac, Mexico; El. Columbi var. imperator Freudenb., 1922. 

Synonyms or related forms (United States): Elfphas scnili Barbour, 1925; 
('!)Elepha.s (Archidiskodon) maibeni Barbour 1925, 1926, Elcphas haroldcooki 
Hay, 1928. See type descriptions of these synonymous or related forms below. 

This imperial species, well named Elephas imperator by Leidy 
because of its commanding size, appears to be a direct descendant- 
of Archidiskodon meridionalis and of A. planifrons of southern 
luirasia. It progresses beyond the A. meridionalis stage, the 
number of plates in its grinding teeth rising to eighteen (ma.\. 
twenty), but it retains the very thiclv enamel and the wide borders 
of cement. Inasmuch as A. meridionalis extended from Upper 
Pliocene into Lower Pleistocene time in western Kuroijc, it may by 
migration have given rise to A. imperator- which seems to char- 
acterize Lower Pleistocene (?Aftonian) time, especially in the west- 
ern and southwestern United States and in Mexico; then this 
imperial mammoth appears to liave become extinct.^ 



Specific Charactp:rs, Osborn, 1928. — (1) Relatively short 
and broad superior and inferior grinding teeth completely sur- 
rounded by a heavy layer of cement. (2) Ridge-plate formula: 
M 3 1 8-V9-2 i ^^ superior molars observed in which the ridge- 
plates exceed eighteen. (3) Brachycephalic and hypsicephalic 
portions of cranium correlated with the excessively short, broad, 
and elevated superior molar teeth (Figs. 889B, 888), also with the 
relatively broad and short inferior molar teeth (Figs. 889A, 892B, 
left, 898). (4) Lower .jaw short, deep, excessively broadened or 
swollen to accommodate the broad inferior grinding teeth (Figs. 
898, 892); inferior ridge-plates normally 16-18, may include im- 
perfect ridge-plates 19-20 (Fig. 892A). Jaws much more abbrevi- 
ate than in Archidiskodon planifrons (Fig. 865), much more mas- 
sive and swollen than in A. meridionalis or any other species of 
proboscidean; rostrum deep and short; symphysis rounded; the 
superior border of the coronoid process greatly raised above the 
grinding surface (Fig. 898). 

Elephas imperator Leidy, 1858. [On new species of Mastodon 
and Elephant from Nebraska, Mastodon mirijicus, Elephas im- 
perator.] Proc. Acad. Nat. Sci. Phila., March, 1858, Vol. X, p. 
10 (Leidy, 1858.2) ; also "Notice of Remains of Extinct Verte- 
brata. from the Valley of the Niobrara River," . . . np. cit., p. 29 
(Leidy, 1858.4). Type.— (Op. cit., 1858.2, p. 10): ". . . part 

of an upper molar tooth of an Elephant from the Niobrara; . . . The 
species he proposed lo distinguish by the name Elephas imperator." 
{Op. cit., 1858.4, p. 29): "The fragment of the tooth has been 
assumed to belong to an unnamed species from the fact that it 
was found in association witli a fauna very distinct from any 
previously noticed." Nat. Mus. 185; cast Amer. Mus. 2568. 



'(Lugn and Schultz (1934.1, p. 373) give the type locality of Eleplias [Archidiskodon] imperator as: "Pawnee. Loup Briuicli of Plaltc River = Middle 
Loup, probably Hooki^ Co. [Nebraska], 1858." — Editor.] 

'-[See footnote 2 on previous page — Editor.] 

'[See footnote 2 on page 9% above.^Editor.] 



THE MAMMONTINyE: ARCHIDISKODON AND METARCHIDISKODON 



999 



Horizon and Locality. — (Leidy, 1858.2): "Valley of the 
Niobrara river, Nebraska." (Leidy, 1869, p. 254) : "The fragment 
of a molar tooth originally referred to a .species with the name of 
Elephas imperator was obtained by Prof. Hayden on the Loup 
Fork of the Platte River." (Hay, 1914, pp. 421, 422) Type col- 
lected from Loup Fork of Platte River, now in the National 
Maseum, designated Nat. Mus. 185. (As quoted above from Hay, 
1924, p. 100) Probably Seneca, Thomas County, Nebraska. 
(Lugn and Schultz, 1934.1, p. 373) "Pawnee Loup Branch of 
Platte River = Middle Loup, probably Hooker Co. [Nebraska], 
1858." Type Figure.— Leidy, 1869, PI. xxv, fig. 3. 



a peculiar fauna, an associate of the Mastodon mirificus, as the 
ordinary E. americanus was of the M. americanus. . . . The specimen 
assigned to E. imperator is represented in figure 3, |)late xxv, one- 
third the diameter of nature. It exhibits the characters attributetl 
by Dr. Falconer to E. Columbi, compared with the supposed 
American variety of E. primigcnins. The specimen is the fore part 
of an upper molar, probably the fifth. The triturating surface, 
extending the breadth of the fragment, is nearly five inches at its 
widest part. The breadth on the less broken side is about seven 
and a half inches, and contains only as many ridges, — that is to 
say, one ridge to an inch of breadth. There is, however, a thick 



E, imperator 
Nal Mus. 185. Type (shaded! 
Amer, Mus. I187t. rev. foutlin 






Type Third Right Superior Molar of Archidiskodon imperator, and 



Fig. 884. Type figure of Elephas imperator 
Leidy, 1858. After Leidy, 1869, PI. xxv, fig. 3. 
One-third natural size. Fragment of a third 
right superior molar, r.M', from Nebraska (Nat. 
Mus. 185; cast Amer. Mus. 2568). Inverted. 

The anterior surface is to the left; ridge-crests 
1-8 are preserved (cf. Fig. 886), ridge-plates 
concave posteriorly. 



Fig. 885. Type molar, r.M', of Elephas 
imperator, crown view. After Osborn, 1922.555, 
I). 4, fig. 4. 

Tliis tootli is in the same position as in 
Leidy's type drawing (Fig. 884), namely, anterior 
surface to the left, posterior surface to the right. 



Reconstruction 

Fig. 886. Leidy's type molar (shaded) and 
Osborn's neotype (outline) combined, both be- 
longing to M^ of the right side, .\fter Osborn, 
1922.555, p. 4, fig. 4. Neotype (Amer. Mus. 
11871) from Guadalajara, Jalisco, Mexico. See 
figure 888, showing eighteen ridge-plates. 



Type Description, 1858.— (Leidy, 1858.4, p. 29): "The 
Niobrara collection also contains the anterior portion of an upper 
molar tooth [Figs. 884-886] of an Elephant of larger proportions 
than any which are known to us. The triturating surface is 
within a line or two of five inches in breadth, and within a space of 
seven inches only eight enamel folds or double plates exist. In the 
most thick plated variety of teeth of the Elephas americanus which 
we have seen, in the same space ten folds were counted. As in the 
latter, E. primigenius, and the recent Elephant of India, the enamel 
plates become worn on the triturating surface into transverse, 
strongly crenulated ellipses." 

Second Description, 1869. — (Leidy, 1869, pp. 254, 255) : 
"The fragment of a molar tooth originally referred to a species with 
the name of Elephas imperator was obtained by Prof. Hayden on 
the Loup Fork of the Platte River. I was led to refer it to a 
species different from the more ordinary American Elephant, from 
its greater size, the comparative coarseness of the constituent 
elements, together with the fact that it appeared to be a member of 



talon in front, and the ridges curve considerably backward on the 
less broken side of the tooth. The four more perfect ridges of the 
specimen at the middle of the triturating surface occupy a space of 
a little over three inches, including the three intermediate plates of 
cementum. The widest ridges measure four and a half inches. The 
enamel is thick and strongly crimped, and the dentinal tracts are 
slightly dilated at their middle. Elephas imperator, if not regarded 
as a peculiar species of a peculiar fauna, may be viewed, together 
with those teeth which have been referred to E. Cohimhi, as belong- 
ing to the Elephas americanus." 

(Osborn, 1924: (1) In his second description Leidy correctly 
designated the type horizon as "the Loup Fork of the Platte 
River," Nebraska. (2) Doubtless influenced by the high authority 
of Falconer (1863, pp. 58, 59 cited below), Leidy was inclined to 
abandon the name Elephas imperator and to refer this type to 
Elephas columbi, and finally (see Leidy, 1869, pp. 251, 252, 255; 
1871, p. 359; 1877, p. 213) to E. americanus De Kay, a species 
representing the American variety of Elephas primigenius, namely. 



1000 



USBORN: THE PROBOSCIDEA 



Mnmmonteun priinigctn'us nmcricnnuf: of the jiresent Memoir (see p. 
1156). This was de.arly a scries of unfortunate errors on Leidy's 
part, due to lack of material. 

FALCONER ERRONEOUSLY UNITES (1863) ELEPHAS IMPERATOR 
WITH ELEPHAS COLUMBI 

(Falconer, 1863, pp. 58, 59, 67): "The second case is more 
remarkable and important, being that of the fossil Elephant of the 
Pliocene Fauna of Niobrara, an affluent of the Missouri River, in 
Nebraska, the account of which, by Dr. Leidy, has excited much 




Types of Archidiskodon imperatou (upper) and Parelepiias 

COLUMBI (lower) 

lis- 887. Elephas columbi (lower figure), fragment, of a third right inferior 
molar, r.M.^, two-sevenths natural size, and Elephas imprrnlor (ujiper figure), 
fragment of a third right superior molar, r.M', two-fifths natural size, photo- 
graphed directly from the t.yp(^ easts of PJ. columhi {Kmov. Mus. 1747) and 
E. imperator (Anier. Mua. 2568). Compare figures 918 and 881. 



T.,eidy adds, 'that the fragment of (he tooth has been assumed to be- 
long to an unnamed species, from the fact that it was found in as- 
sociation with a fauna very distinct from any previously noticed.' " 
In a subse(iuent passage in the same ])aper (1863, p. 67), 
Falconer observes that in his opinion there arc "but two well- 
determined species of fossil Elephant known in North America," 
namely: (1) E. pn'migenius (syn. E. americaniis). (2) E. cohimbi 
(syn. E. primigenius, pro parte; E. lexianus). He concludes (p. 
67): "The same, with our present knowledge, must be said of the 



E, Imp^ratoi 
Neotype 




'34 567 

E. imperatop 
Amer. Mus. 11B71 
Neotype 

B 



1/4 not. size 




Neotype of Archidiskodon imperator 

Fig. 888. From Guadalajara, .laliseo, Mexieo. B 2, Neotype superior 

molar of Archidiskodon iiripcrnlor, M'' of the right, side (Amer. Mu.s. 11871), 

inner view, eighteen ridge-|)lates. B, The jsame, crown view. One-fourth 

natural size. After Osborn, 1922..').').'), j). 5, tig. o. See figure 886 Al, in outline. 



interest and surprise among Palseontologists. [Footnote: 'Proceed. 
Acad. Nat. Scien. Philadelp., 1858, p. 20, et .seg.'] ... 1 Mastodon 
of the sub-genus Telralophodon, M. mirificus, Leidy, and a huge 
Flephant, E. [Eueleph.) imperator, Leidy. The published descrip- 
tive details of this Elephant are as yet but very meagre. One speci- 
men only is mentioned, being the anterior i)ortion of an upper 
molar, of larger dimensions than any known to the author. The 
crown is stated to be 'within a line or two of five inches in breadth, 
and within a space of seven inches, only eight enamel folds or 
double i)lates exist.' This would give an a\crage of nearly ninc- 
tenths of an inch to each ridge, corresponding closely with the 
proportions yielded by E. Columbi. The ridges are described as 
becoming worn into transverse strongly crenelated ellipses. Dr. 



E. imperator of Leidy, from Niobrara. Until a i)erfect molar is 
figured and described, no satisfactory opinion can be formed as to 
what the species is. Dr. Leidy, as already stated, iissumed it to be 
distinct, and gave it the name upon the asstuuption." 

That Falconer erred as to synonymy is clearly demonstrated in 
the accomi)anying figure (Fig. 887) in which we observe that 
while the enamel ridges are equidistant in Archidiskodon imperator 
anil in Parelepha.s colnmbi, antl that while the riilge formula 
is aijproximately similar in the two species, namely, M 3 J^tH (*4- 
imperator), ^r^.'^^j^ (P- columbi), we also observe that the molar of 
A. imperator (u])pcr) is much broader and is surrounded with 
a broad layer of cement, while the molar of P. columbi (lower) is 
much narrower and is lacking in cement. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1001 



OSBORN SEPARATES (1922) ELEPHAS IMPERATOR LEIDY FROM 
ELEPHAS COLUMBI FALCONER 

In reexamining and comparing the fragmentary molar types of 
Elephas columbi and E. imperator, Osborn (1922.555, pp. 1-7) 
determined with great care the exact position and character of both 
type fragments as shown in figures 949, 951A, 951B (for E. co- 
lumbi) , and in figure 886 (for E. imperator) of the jjresent Memoir. 
The E. imperator tyjje iiro\'es to be the anterior portion of a third 
right superior molar containing portions of the first to the eighth 
ridge-plates greatly worn, anil broken away from ridge-plates nine 



us to determine exactly to what portion of the complete neotype 
tooth (Fig. 5 [ = Fig. 888 of the present Memoir]) this ancient and 
much battered type belonged; the eight ridge-plates of the type 
[Nat. Mus. 185] which are jireserved, in comparison with those of 
the neotype (Amer. Mus. 11871), constitute the anteroposterior 
portion of a much-worn molar, M^ of the right side, in which thir- 
teen ridge-phites were in use out of an estimated total of seventeen 
[eighteen]. Of these plates five occupy a line 100 mm. long; this is 
because the ridge-plates are arcuate and widest apart in the middle 
portion of the crown. The neotype tooth (Amer. Mus. 11871), 



E. imperator 
Amor. Mus. 14558 Rel 




1/4 nat. sizo 



Fig. 889. Superior (B) and inforior (A) molars of .Irc/iirfisAvK/iwf i'mpprator, reforred Amerioan Museum .speeimcns, one-fourth natural size. Tlie.se two 
individuals are believed to be of e(jrr('spoiiding age. They exhibit, (M^, M3) meclianieal reversal of the convex and concave .surfaces both in the crown 
contours, crown surfaces, and ridge-plates. After Osborn, 1922.555, p. 0, fig. 6. 

A, Inferior molars (Amer. Mus. 14558), from Ness County, Kan.sas, with 14-15 ridge-plates in u.se, a total of 19 vi.sible in left molar. See also figure 892A, 
where a total of 20 ridge-plates are visible owing to artificial exposure. 

B, Superior molars (Amer. Mus. 14476), from Victoria, Victoria County, Texas, with 14-15 ridge-plates in use, a total of 154-- Inferior view of this skull 
shown in figure 897. See also figure 896. 



to seventeen, as clearly shown in outline in the accompanying 
figure 886 Al. At the same time Osborn selected as a neotype 
the third superior molar tooth (Amer. Mus. 11871) from Guadala- 
jara, Jalisco, Mexico, from which the outline in figure 886 Al is 
drawn. 

Characters of the Archidiskodon imperator Type and 
Neotype.— (Cf. Osborn, 1922.555, pp. 3-5) : "We are indebted to 
the National Museum for the loan of the Elephas imperator type 
specimen (Fig. 4 [=Fig. 886 of the pre.sent Memoir]), enabling 



from Guadalajara, Jalisco, Mexico, appears to attain the full size of 
the superior grinders of this species of mammoth ; the ridge formida 



may be written M 3 



[M 3 



This accords with the 



1 8-1 9 L-^ -^ '-' 1 8-1 9-20J 
actual average count of the ridge-plates in E. imperator by Hay 

(1914) and by Osborn (1921-1922) in individuals which can with- 
out question be referred to E. imperator. ... In the neotype (Fig. 
5 [ =Fig. 888 of the present Memoir]) thirteen plates were in use; 
in the referred skull (Amer. Mus. 14476) fifteen plates were in use 
(Fig. 6B [ = Fig. 889 B of the present Memoir]); in the referred 



1002 OSBORN: THEPROBOSCIDEA 

lower jaw (Anier. Mus. 14558) fifteen plates were in use (Fig. 6A Meanwhile Hay (1914) made a series of observations on the 

[ = Fig. 889 of the present Memoir)). The total ridge-plates in M3 type and other grinding teeth truly referable to E. [ = A.] imptrator 

attain nineteen [twenty], as clearly shown in Fig. 6A [ = Figs. 889A in various museums, from which the following citations may be 

and 892A of the jiresent Memoir]." We have, therefore, the fol- made, 
lowing ritlge formulae (cf. specific definition above). 

Upper Pleistocene. Elephas [ = Parelephas] columbi of the HAY'S (1914, 1923, 1924) OBSERVATIONS ON ELEPHAS IMPERATOR 

southern United States: M 3 r^.\-a'+- Osborn summarizes below (pp. 1087, 1088) Hay's observations 

Lower Pleistocene'. Elephas [ = Archidiskodon] imperatnr of on Elephas [ = Parelephas] columbi and geographic distribution 

the southeastern and western United Slates: M 3 isMa-l o- records, as well as of E. [ = Arrhidlsk(Hlon] imperator in the region 

Upper Pliocene to Lower Pleistocene. Elephas [=Archidisko- west of the Mississippi River (Fig. 890); also Osborn points out 

do>i\ men'diunalis, ancestral, of western Europe: M 3 titH- (P- 699) the clear distinction between the true E. [ = P.] colwmbi 

(Osborn, 1922.555, p. 5): "The cranial characters observed (as limited by Osborn) and the true i?. [.4.] /7«pe?-a<or of Leidy. 

in three more or less complete skulls referred to Elephas imperator Dkntal Characters (Hay, 1914, p. 421). — It now remains 

tend to support the direct descent of this animal from the E. to summarize Hay's observations of 1914 on the grinding teeth 

meridionalis of the ^'al d'Arno, Upper Pliocene of Italy." discovered near the type locality of Nebraska and on those found 

Osborn, 1924: Thus £. [ = Archidiskodon] imperator appears in Oklahoma and Kan.sas (see Fig. 890) : "Tooth formula not well 

to be firndy established in its dental characters and ridge-plate known; the hindermost molars having apparently from sixteen 

formula as a probable successor of £■. [ = ^4.] meridionalis, whereas to twenty ridge-plates; the teeth large, with thick ridge-i)lates and 

E. { = Parelephas] columbi is clearly distinguished as a smaller form thick enamel; the ridge-plates often concave on their hinder face 

with narrow molar teeth which lack the very broadly encircling and more or less warped ; those of the hinder half of the lower teeth 

layer of cement but exhibit practically the same ridge formula, leaning strongly forward. This is a species not yet well known, 

Freudenberg (1922) has suggested that the true E. [ = P.] columbi although various parts of its skeleton have been collected." The 

is geologically a successor of the true E. [ = .4.] imperator stage. chief materials observed and enumerated by Hay are: 

Xat. Mus. 185 M^ fragment (Loup Fork of Platte River, Neb.), eight anterior ridge-plates. {Op. cit., p. 422): "The breadth of the 
Type grinding face is 125 mm., but of this about 5 mm. on each side belong to the cement. There are hardly five ridge- 

plates in a line 100 mm. long and crossing the plates at right angles. The enamel plates are each nearly as thick 
as the layer of dentine enclosed by them. The plates of cement intervening between the ridge-plates are somewhat, 
but not greatly thicker than the plates of dentine. The face of each enamel plate which is directed toward the 
l)late of cement is moderately striated from the base to the summit, but this striation is not deep enough to be 
called crimping. At their inner and outer margins the ridge-plates are turned backward so as to make each one 
deeply concave on the hinder face, convex on the front face." 

Nat. Mus. 2216 M^ (Afton, Okla.). {Op. cit., p. 422): "The surface of wear extends back to the ninth plate. There are counted 
sixteen plates. The anterior talon is mi.ssing on account of wear, and the posterior one is broken off. It seems not 
unlikely that at least two ridge-plates are missing, either in front or behind. The plates are very thick, there 
being hardlj' five of them in a 100 mm. line." 
{Op. cit., p. 422): "In another tooth, which is broken, it is seen that the hinder faces of the plates are considerably 
dish-shaped. It will be ob-served likewise, that the plates converge toward their summits. The base of the tooth 
is very convex. The length of the tooth, from the base of the first ridge-plate to that of the hindermost, is 350 mm. 
The thickness is 126 mm., more than one-third the length of the tooth. The height of the ninth plate is 250 mm. 
As will be noted, the plane of wear, at the stage represented by this tooth, strikes the summits of the plates very 
obliquely." 

Nat. Mus. 2217 M3 (Afton, Okla.). {Op. cit., p. 423) : "The length from the base of the anterior plate present to that of the hinder one, 
is 297 mm. The greatest width, taken at one-half the height of the tenth jilate, is 125 mm.; the height of the 
ninth ])Iate is, in a straight line, 170 mm. At the middle of the inner and outer faces there are only three and 
a half ridge-plates crossed by a line 100 mm. long. At the rear there are five plates in such a line. There are to be 
counted sixteen ridge-plates and a posterior talon. These are about as thick as the cement plates. The enamel is 
thick and considerably crimped." 

Phil. Acad. M3 (Wellington, Kan.?). {Op. cit., p. 424): "It is i)retty certain that one ridge-plate is missing in front, and one or 

more at the rear. Eighteen plates are present. From the front of this tooth to the base of the hindermost plate 
is 435 mm., a little more than seventeen inches. The height of the tenth plate, in a straight line, is 180 mm. . . . 
On the hinder end of the tooth, on the grinding surface, six plates outcrop in a distanc^e of 100 mm.; but on the 
side of the tooth, at one-half of the height, there are only three and one-half plates in 100 mm." 

'[See footnote 2 on page 996 above regarding the Early and Middle Pleistocene age of A. imperator. — Editor.] 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1003 



Nat. Mus. 6662 Dps (Afton, Okla.), length 135 mm., width 69 mm.; ridge-plates 7+; 5% or 6 ridge-plates in 100 mm.; enamel thick 
(see Hay, op. cit., Pi. lxi, figs. 7, 8). 

Nat. Mus. 6663 Dp'' or M' (Afton, Okla.), doubtfully referred. 12 estimated ridge-plates (Hay, op. cit., p. 414). 




WiKonain 



lowan 



lUinoian 



CD 

Orifliut 



Fig. 890. Distribution of .\rchidiskodon imperator west of 

AND EXPLAN.\TI0N OF MaP, P 

Nebraska. (6) Near type locality "Loup Fork of Platte River" [probably 
Seneca (fide Hay)), al.so eight other localities, Holmes- 
ville, Fairbury, Reynolds, Wauneta, Cody, Grayson, 
Hebron, and Sutton. [Lugn and Scliultz (1934.1, p. 
373) give "Pawnee Loup Branch of Platte River = 
Middle Loup, probably Hooker Co.," as the type 
locality. — Editor. ) 

Oklahoma. (1) Afton, Ottawa County; type region of the Aftonian or 
Ist Interglacial stage, in which age A. imperator was 
very abundant. Also three other localities, Kinlock, 
Okeene, and Kingfisher. Kingfisher not on map. .1 



lUnsM Pre4Visconsin 

THE Mi.ssissippi River. .After Hay, 1924, Map 10, p. 337, 
336 (Texas omitted). 
Kansas. Localities 1 to 3, Wellington, Ness City, Russell (Russell not 
on map). 

Colorado. Localities 1 to 6, Laveta, Colorado Springs, Denver, Boulder, 
Fort Collins, Glenwood Springs. 

Iowa. Localities 1 and 2, Mt. Pi.sgah and Mapleton. 

Wyoming. (1) Powder River. 

Montana. Localities 1 and 2, Helena and Valier. 

These grinding teeth are recorded as exhibiting the specific characters of 

rchidislcodon imperator or of A . tiayi. 




1 4 nat. size 



Cranium of a Young Male of ARCHiorsKODON imperator of Texas in the American Museum 
I''ig. 891. Rilfncd young male skull of Archidiskodon imperator Leidy (Amer. Mus. Cojie Coll. 14475), discovered near Dallas, Texas. Compare figures 
895, 896 (restored), also figure 805. 

.\, Frontal aspect. One-eighth natural size. Al, Left lateral aspect. One-eighth natural .size. A2, Partly worn left M-. One-fourth natural size. 

Tliis young male skull was described by Cope in 1889 (1889.2) as "Elephas pritnigenius coliunbi Fale." and figured by him in PI. xiv, also on p. 208, fig. 9. 
It was erroneou.sly referred by Osborn (1922..5.55, p. \!i) to Elephas [ = Parelephas] jeffersonii. Both the jirofile and front views as well as the wire mid-.section 
(Fig. 80.'), Chapter XV) of this cranium in<licate that it is more elo.sely related to Arcliidiskodoii than to Parelephns. The single molar tooth preserved, l.M", 
enables us to determine this individual as probably belonging to A rchidiskoilon imperator. 

Second left superior molar, I.M", length 140e 

Ridge-plates estimated (? 12); expo.sed S + 

Width at broadest ridge-plate 80 

Ridge-plates in 10 em. 6 

Tip of left tusk to occipital condyle l,400e 

Alveolus of left tu.sk to occipital condyle 960e 

Breadth acro.ss cranium, posterior portion of occiput 032 

Breadth acro.ss jugals tilOe 

Breadth between orbits 484 

Breadth between temporal fenestrje 330 

We observe that this Texas skull (Amer. Mus. Cope Coll. 14475) belongs to a young male mammoth in which the superior tusks are partly erupt<'d; it was 
at first regarded as a female .skull and was so labeled, but the diameter of the tu.sks forbids such a reference. In vertical .section (Fig. 810) this cranium closely 
corresponds with the two other crania in the Ami^ican Museum collection certainly referable to th(! s]«;cics Parilephas jefferxonii. The young grinding 
tooth shown herewith (.-^2), a left .second superior molar, I.M^, does not .so clearly display the characters of /'. jefft-monii as to preclude the possibility 
of tliis cranium belonging to a juvenile Archidiskodon imperator; con.sequently the reference is somewhat doubtful. The skull was found in the Texa.s 
geographi<' region characteri.stic of A. imperator in early Pleistocene times, a region which was also traversed by the migrations of I'.jeffersottii in late Pleistocene 
times. 



1004 



THE MAMMONTIN^: ARCHIDISKOnON AND METARCHIDISKODON 



1005 



Geographic Distribution and Nomenclature (Hay, 1914, 
1923, 1924).— Hay (1923, 1924) has made interesting and important 
obser\'ations on the geographic distribution of the remains of 
Elephas [ = Archidiskodon] imperator in the central region of the 
United States which are summarized on pages 434 and 435, map 
and legend, 1923, and pages 336 and 337, map and legend, 1924 
[ = Fig. 890 of the present Memoir]. 

Unfortunately (1) some of the specimens referred in Hay's 
work (1914, 1923, 1924) to E. imperator may belong to the true 
Elephas \ = Parelephas] columbi and vice versa, owing to the con- 
fusion which has existed in the minds of practically all American 
writers up to July 8, 1922 (Osborn, 1922.555), when Osborn clearly 
distinguished the true E. columbi from the true E. imperator. (2) 
In Hay's volume of October, 1924, the northerly mammoths 
previously (1914-1923) named by him "Elephas columbi," as 
plotted on map 5, p. 327, and map 6, p. 329, are partly named by 
him "Elephas boreus," a name preoccupied by Parelephas jeffersonii 
Osborn; also, on maps 7 to 9, are partly named Elephas columbi. 
As above noted. Hay (1924, pp. 57-84, "Finds of Elephas columbi 
in the Middle Region of North America") fails to recognize the 
now well-determined typical ridge formula of Parelephas columbi, 
namely, I\I 3 riTe+- 

ARCHIDISKODON IMPERATOR IN THE U. S. NATIONAL MUSEUM 
FLORIDA COLLECTION 

Specimens collected for the U. S. National Museum by James 
W. Gidley in the locality of Melbourne, Florida, and presumably all 
from the "No. 2" bed of Sellards, are as follows: 

Nat. Mus. 11805. A third inferior molar, M3, with 15-18 
ridge-plates; laminar frequency 5 ridge-plates in 
10 cm. 

Nat. Mus. 11814. A much worn third inferior molar, M3; 
laminar frecjuency 5 ridge-plates in 10 cm. Total 
ridge-plates 12+ (tooth worn to base). 

Nat. Mus. 11620. The Venice mammoth from near Mel- 
bourne, Florida. Aged individual, with third superi- 
or and inferior molars much worn, also jaw, por- 
tions of skull, right tusk, and hind foot. Third 
superior molar, l.M', total ridge-plates estimated at 
18)2-19, 3 to 4 anterior worn off, length 200 mm., 
breadth 106 mm., 5}i ridge-plates in 10 cm.; maxi- 
mum breadth 106 mm. as compared with 125 mm. 
in Leidy's Nebraska type (Fig. 884). Third inferior 
molar, r.Ms, estimated ridge-plates 16-17, 3 to 4 



anterior worn off, length 228 mm., breadth 97 mm. 

Gidley (letter, Oct. 27, 1928) adds that the specimen from 
Venice has an extremely deep jaw but otherwise is like .4. impera- 
tor; that both A. [ = Parelephas] columbi and ^4. imperator occur 
at Melbourne and at Vero. "I think, therefore, there can be no 
doubt that these two species were contemporaneous in Florida, 
and perhaps also in other localities of the southern and south- 
western United States." 

Osborn, 1922-1928: This is a very important case of the ap- 
parent association of Archidi.skodon imperator ref. and of Parelephas 
columbi ref. in the same geologic locality, if not actually from the 
same level. 

ARCHIDISKODON IMPERATOR IN THE AMERICAN MUSEUM 

COLLECTION 
See figures 885, 886, 888, 889, 891, 892, 896-898, 900, 906-908, 912, 920 
of the present Memoir. 

The type cast, the neotype, and the finest referred specimens 
referable to Archidiskodon in the American Museum collection are 
listed herewith (bottom of this page). These superb materials are 
illustrated in various figures of the present Memoir and add 
greatly to our knowledge of this imperial species of southern mam- 
moth. They display clearly the specific characters enumerated 
above. 

COMPARISON OF THE JAW AND DENTITION WITH OTHER 
PROBOSCIDEANS 

A great deal of study has been given to comparison of the jaws 
of various types of the Proboscidea, especially in correlation with : 
(1) Abbreviation (brachycephaly) and elevation (hypsicephaly) of 
the cranium, also with (2) insertion of the third inferior molar 
tooth which naturally is the chief functional organ of the body. 
It is seen in the accompanying diagram (Figs. 892, 893) that the 
progressive jaws of Archidiskodon imperator (Fig. 892 A, B) are 
profoundly different from the relatively elongate jaws represented 
in figures 893, A, B, C (Archidiskodon hayi, Parelephas washing- 
tonii, Loiodonta africana). The jaws of Archidiskodon imperator 
(Fig. 892 A, B) are also readily distinguished both from those of 
Parelephas jeffersonii (Fig. 892 C, D) and of Elephas indicus (Fig. 
893, D, E). 

It is observed that in the Texas jaw (Amer. Mus. 10598 — see 
Fig. 892 B) M3 exhibits sixteen ridge-plates, while in the Kansas 
jaw (Amer. Mus. 14558 — see Fig. 892A) M3 exhibits five accessory 
ridge-plates, namely, 16, 17, 18, 19, 20; this difference cannot be due 



Amer. Mus. 2568 

United States 
Amer. Mus. 11871 

Mexico 
Amer. Mus. 14476 

Texas 
Amer. Mus. 14558 

Kansas 

Amer. Mus. 10598 

Texas 
Amer. Mus. Gope 

Goll. 14475 

Texas 



Type r.M^ of Elephas imperator, cast (Amer. Mus. 2568) after original in U. S. National Museum (Nat. Mus. 185). 

See figures 884, 886, 887. 
Neotype (Osborn) of Elephas imperator, right third superior molar, r.M^ from Guadalajara, Jalisco, Me.xico. 

See figures 886 Al, 888. 
Lower portion of cranium with tusks, here represented in figures 889B, 896, 897, 906. From Victoria, Victoria 

Gounty, Texas. 
Imperfect lower jaw, right and left third inferior molars. Crown view (Fig. 889A); side view (Figs. 892A, Al). 

Total of 19-20 ridge-plates, of which the 19th and 20th are extremely short and rudimentary (see Fig. 892). 

From Ness County, Kansas. 
Lower jaw with partly worn M3 (see Figs. 892, 898), also right forelimb (see Figs. 906, 907, 908, 912). From near 

Tule Cafion, Briscoe County, Texas. 
Referred young male skull of Archidiskodon imperator, with tusks and l.M'-, discovered near Dallas, Texas, and 

described by Cope in 1889 (1889.2) as "Elephas primigenius columbi Falc." (see Fig. 891). 



Elepnas jetlersonn 
Amer Mus 13225 




Jaws of Archidiskodon imperator in comparison with Parelephas je 

Fig. 892. Internal aspect of the lower jaw.s and M3 of Elephns [ =Parele- 
■phas\ jfffersonii (paratype and type) and Elephns [ = Archvliskodojt] imperator, 
al.so sections Dl, Cl, Bl, Al, cut at point indicated by the dotted line S. 
( )iie-eighth natural size. 

D, Elephtts [ = Parelephas] jeffersonii ])aratype (Amer. Mus. 13225); M3 
in situ with fifteen worn plates and eight unworn plates, a total of twenty-three 
plates. [See Fig. 960 for final count of twenty-four ridge-platcs in this |)ara- 
type specimen.] 

C, Type of Elephas [ = Parelephas] jeffersonii (Amer. Mus. 99.J0), a very 
aged individual, M3 with seventeen worn jilates and three additional unworn 
plates in situ, a total of twenty plates. This, as exi)lained elsewhere, is a 
very aged individual and it is probable that one or two of the cxtreine an- 
terior i)lates have been worn off. 

B, Jaw of Elephas [= Archidiskodim] imperator ref. (.\iner. Mus. l()r)98), 
M3 with fourteen worn plates and a total of sixteen plates in situ; a very 
robust individual. From Tule Canon, Texas. 

A, Elephas [ = Archidiskodon] imperator ref. (Amer. Mus. 14558), from 
Ness County, Kansas, M3 with fourteen to fifteen (est.) worn i)lates and five 
additional, a total of twenty plates. See also figure 889A, crown view of same 
tooth. 



FFERSONII, P. WASHINOTONII, LoXODONTA AFRICANA, AND ElEPHAS INDICUS 

Fig. 893. Internal views of jaws of Elephns indicus, Loxodonta africana, 
Elephas [= Parelephas] washinglonii, and E. \ = Archidiskodon] hnyi, for com- 
parison with figure 892. One-eighth natural size. 

E, Elephns indicus, a fully developed M3 of the right side with twenty- 
seven plates. Drawn after de Blainville, 1839 1804, PI. ix, fig. ti'' (reversed). 

D, Elephas indicus, right jaw (Amer. Mus. Dept. Mam. .54201); M3 with 
nineteen plates fully formed and five imperfect plates in the jaw, a total of 
twenty-four plates. 

C, Loxodonta africana jaw (Amer. Mus, Dept. Mam. 39083, Akeley Coll.); 
?M2 with twelve plates developed. 

B, Type jaw of Elephas ( = Parelephas] washingtonii Osborn (Amer. Mus. 
S081A), with twenty-one plates developed in M^. Compare figure 909, type of 
Parelephas progressus. 

A, Type jaw of Elephas [ = A rchidiskodon] haiji Barbour (Neb. Mus. 
23 0-14)." 



1000 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1007 



to age or to sex; it is more probably due to the progressive addi- 
tion of ridge-plates in the Kansas specimen. Majestic as are these 
imperial mammoths from Texas the largest pair of tusks known 
is from Mexico and is described below, with change of name, by 
Osborn (Fig. 894).' 

Apposition of the Grinders. — In the American Museum 
collection there is no case of skull and jaws being found together, 
so that we cannot positively determine the apposition of the 
grinders ; it is almost certain that an equal number of ridges are in 
use at the same time in animals of the same age. Consequently 
the superior and inferior molars represented in figure 889, in each of 
which fifteen ridge-plates are in use above and below, probably 
represent animals of the same age, although they may not be ex- 
actly in the same stage of dental progression. The teeth from 
Kansas (Amer. Mus. 14558— Fig. 889A, 892A) have 14-15 ridge- 
plates in use, also the maxillary teeth from Texas (Amer. Mus. 
14476 — see Fig. 889B) have fifteen ridge-plates in use. These 
grinders (Fig. 889) are examples of typical A. imperator molars 
exhibiting the following characters: (1) Inferior ridge-plates con- 
cave anteriorly, partly worn summits of ridge-plates exhibiting 



Incisive Tusks. — The characters of the tusks are superbly 
disi)layed in the cranium (Amer. Mus. 14476) as shown in figure 
896; these tusks curve downwards, outwards, upwards, inwards, 
until finally they cross each other on the median line, forming 
a complete superior arch, as in the aged type specimen of Elephas 
[ = Parelephas] jeffersonii. The tusks are parallel but not closely 
opposed to each other where they issue from the alveoli, as in the 
specimens referred to this species in the Nebraska State Museum 
(Fig. 899) found at Hay Springs, Sheridan County, Nebraska. 

This comparative study of the superior and inferior teeth and 
jaws enables us to define the species A. imperator much more 
closely than before. 

Record Proboscidean Tusk 
A record proboscidean tusk (Amer. Mus. 22481) was presented 
to the American Museum in 1934 by Mr. (Jeorge D. Doughty of 
Post, Texas. This tusk was found in the vicinity of Mr. Doughty's 
home, namely, Post, Garza County, Texas, in which region the 
Imperial Mammoth (Archidiskodon imperator) seems to have 
flourished in Pleistocene time. This gift was first noted in the 
January, 1935, number of Natural History, page 84, followed in 
the April number, page 357, by a brief description with compara- 
tive estimated measurements. Since that time the tusk has been 
restored conservatively to conform to the measurements given by 
Mr. Doughty, who was unable to save the anterior end; in life it 



Fig. 894. A right superior tusk of 
record size (16+ ft.) of Archidiskodon 
imperator (Amer. Mus. 22481), from 
Post, Texas, one-twentieth natural size, 
in comparison with the superb left tu.sk 
(13 ft. 9K in.) formerly in the National 
Museum, Mexico, after photograph of 
original taken by Barnum Brown in 
1910. A letter from Sr. Francesco Con- 
treras of the Universidad Nacional de 
Mexico, May, 1935, states that the orig- 
inal tusk w.is unfortunately destroyed. 



-i'mf»m>'-»«^- 



7-8 conelets; (2) superior molars, \M\ ap. 290 mm. (restored), 
tr. Ill mm., index 38; r.M', ap. 295 mm. (restored), tr. 120 mm., 
index 41 ; (3) M', ridge-plates concave posteriorly, thus following 
the mechanical principle of reversal; (4) breadth-length index, 
M' = 41. 







ite' 






«^HJB-'^&. 






j^^Kr 










^^31 


HL« 






^^^^^^^^^^^"^^^E^^k 




U 


^^^^^^^^^^^^^^^ 



Fig. 895. Referred young male(?) skull of Archidiskodon imperator from 
the tar pools of Rancho La Brea, California. Los Angeles Museum 3800-1 
(skull), 3801-1 (jaws). 

Although partly restored at the summit of the occiput, this is one of the 
most highly characteristic skulls ever found, in the sharp concavity of the 
forehead and the peaked cranium, resembling closely the skull of Mammonteus. 



'[This statement was true at the time it was written about eight or ten years ago. In 1934, however, Mr. George D. Doughty of Post, Texas, presented the 
American Museum with a giant tusk of Archidiskodon imperator (Amer. Mus. 22481), the size and weight of which establishes the record for this species, or, 
as a matter of fact, for any proboscidean, namely, a length of 16+ feet. This gift was noted in 1935 in the January number of "Natural History," page 84, 
followed in the April number (p. 357) by a brief description with measurements (see Fig. 894).— Editor.] 




Fig 890 Skull of ArchidUkodon imperator (Amer. M.is. 14476), found at Victoria, Texas. The upper portion of tins skull is entire- 
ly restored. The prema.xillaries, the palate, and the .superior grinding teeth are perfectly preserved. The palate and grinding teeth of 
this specimen are also represented in figures 897 and 889. The tusks are complete and nat,ural, without restorat,ion; they measure 4 m. 
21 5 cm. or 13 ft. 10 in., as compared with 4 m. 20 cm., est., or 13 ft. 9>i in., the measunnK'nt of the tusks of A. imperator in the Geologi- 
cal Institute of the City of Mexico (Fig. 894), a.s remeasured by Barnum Brown. [See footnote on preceding page which give.s the 
length of the record tusk of this species (Amer. Mus. 22481) as 16+ ft.— Editor.) 

1008 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1009 



was evidently longer than now restored, namely, 
16+ feet. Associated with this giant tusk were 
two superior molars (M-), probably belonging to 
the same individual. 

CRANIAL CHARACTERS OF ARCHIDISKODON 
Compare Chapter XV, pp. 915-926, also figurrs 86.5, 
891, 895, 896, 897, 902, and 906 of the present Memoir. 

In the progress from youth to maturity, the 
occipital summit and profile of the Archidiskodon 
cranium changes profoundly with growth, with sex, 
and with the development of the enormous' superior 
tusks; these progressive stages are dis]3layed in a 
comparison of the following eight more or less com- 
plete crania in the Los Angeles, American, and 
Nebraska State museums, and the Geological In- 
stitute of Mexico: 

Fig. 895. California. Primitive stage. Young 
male(?) skull from Elephant Pit No. 9, Rancho La 
Brea tar pool (Los Angeles Mus. Nos. 3800-1 (skull) 
and 3801-1 (jaws). [ = Archidiskodon imperator.] 

Fig. 891. Texas. Young male stage. Skull 
from near Dallas (Amer. Mus. 14475). 

[= Archidiskodon imperator.] 

Fig. 902. Mexico. Adult male stage. Skull 
from Tepexpan (Geol. Inst. 212). 

[ = Archidiskodon imperator.] 

Fig. 897. Texas. Aged stage. Skull (summit 
restored) from Victoria, Victoria County (Amer. 
Mus. 14476). [ = Archidiskodon imperator.] 

Fig. 918. Nebraska. 'Lincoln County Mam- 
moth.' Type skull, mandible, and tusks from near 
Curtis, Lincoln County (Neb. Mus. 5-9-22). 
[ = Archidiskodon imperator maibeni.] 

Nebraska. 'Adams County Mam- 
moth.' Not figured in present Memoir. 
Skull, mandible, and teeth from sandpit 6 
miles due south of Hastings (Neb. Mus. 
11-3-13). [ = Archidiskodon imperator.] 

Nebraska. 'Howard County Mam- 
moth.' Skull, mandible, teeth, and tusks, 
from near Dannebrog (Neb. Mus. 2-7-17B). 
[ = Archidiskodon imperator.] 

Nebraska. 'Custer County Mammoth.' 
Skull, tusk, and teeth from Callaway (Neb. 
Mus. 16-6-16). [ = Archidiskodon imperator.] 

Texas, Victoria, Skull (Amer. Mus. 
14476— Figs. 896, 897, 906).— Only the 
lower portion of the cranium, including the 
palate, condyles, and paired grinding teeth, 
is preserved, the upper portion, as shown 
in figure 896, being entirely restored 
above the white dotted line; as the summit 
of the Victoria, Texas, skull (Amer. Mus. 
14476) is restored, we have not ventured to 
give this hypsicephalic character full expres- 
sion; it characterizes all the young speci- 





Fig. 897. Skull of Archidiskodon imfjcralnr from Texa.s, as largely restored and mounted in llie 
.\merioan Museum. 

Skull (Amer. Mus. 14476) found near Victoria, Victoria County, Texas; upper part of cranium 
entirely restored. Side view one twenty-fourth natural size; palatal view one-sixth natural size. 
See also figure 889 for view of palate; figure 896 for front view of same skull. 



1010 



OSBORN: THE PROBOSCIDEA 



mens of the species /I. imperator; but an aged Mexican skull (Fig. 
902) has a broad, massive occiput. 

The hypsicephaly or peaked contour of the superior crest is 
based upon a beautiful cranium of a young male (?) in the Los 
Angeles Museum, from Rancho La Brea (Figs. 865, No. 10, and 
895); we observe a remarkable similarity between the ])rofile 
of this Los Angeles cranium and that of the true Mammonleus 
primigenius, namely: (1) Forehead concave; (2) parieto-occipital 
union acute, hypsicephalic ; (3) occipital condyles and orbits 
approximate, i.e., brachycranial ; (4) depth extreme from peak 
of cranium to base of lower jaw. Reverting to figure 865, these 
distinctive cranial characters of Archidiskodon imperator are 



Amer. Mus. 10598 Ref. 




1/B nat. size 



Fig. 898. Jaw of Archidiskodon imperator rof. (Amor. Mils. 10.598), found 
at Tiilc Canon, Texas, by the American Museum Exi)edition of 1899. One- 
eighth natural size. Section and inside view of the same jaw, completely 
exposing M3, shown in figure 892B, Bl. 

\, Higlit rsunus, outer aspect ; jaw uptilted. 

k\, Top view; tooth gn-atly shortened by perspective. 

A2, Same tooth in perpendicular view of crown, sliowing sixteeii ridge- 
platcs, fourteen of which arc worn, indicating that tliis was a fully adult 
animal. 

See figure 907 for forelimb of same individual. 



closely analogous if not genetically related to those of the true 
mammoth (Mannnonleun primigenius), while they arc widely 
distinct from those of the trogontherian elephant {Parelephas 
trogontherii) or from the typical elephant (Elephns indirus). 

Comparing the profile and palate of this cranium with thai of 
Archidiskodon planifrons and of A. meridionalis, we observe the 
close similarity in fore-and-aft compression (cyrtocephaly) and 
corresponding vertical elevation (hypsicephaly, acrocephaly). 

FLATTENED CRANIA OF ARCHIDISKODON IMPERATOR FROM 
HAY SPRINGS, NEBRASKA, IN THE NEBRASKA AND 
AMERICAN MUSEUMS. AFTONIAN AGE. 

As provisionally determined by Matthew (1902.1, pp. 317, 
318, and 1918.1, pp. 226, 227) from the American Museum col- 
lections of 1893 and 1897 (modified by Hay and by Osborn), also 
by Frick (1929.1, p. 107, and 1930.1, p. 79), [and finally by Barbour 
and Schultz who have published a preliminary list of the Hay 
Springs fauna (1937.1, pp. 3-6)], the following includes the species 
so far recorded by the above-mentioned authors: 

Edentata 

Mylodon garmani Allen 
Mylodon nebrnscensis (Brown) 
Megalonyx leidyi Lindahl 

RODENTIA 

Cynomys niobrarius Hay 

Geomys sp. 

Thomamys sp. 

Castoroides ohioensia nebraskennis Barbour 

Castor sp. 

Ondatra nebrascensis (Hollister) 

Microtus? 

Carnivora 

Canis latrans? Say 

Canis (Aenocyon) dirus nebrascensis Frick 

Arctodus simus nebrascensis Frick 

Mustela vison? Schreber 

Smilodon nebrascensis Matthew 

PnOBOSCIDEA 

Archidiskodon imperator (Leidy) 

Perissodactvla 

Equus excelsus Leidy 

Equus excelsus niobrarensis Hay 

Equus calabatiis nebrascensis Frick 

Artiodactyla 

Platygoiitis vrtus Leidy 
Camelops kansnnus Leidy 
Camelops vitakerianus? (Cope) 
Tanupolama amcricanus (Wortman) 
Odocoileus sheridanus Frick 
Capromeryx furcifer Matthew 
Telrnmcryx (Ilayoccros) falkcnbarhi l*'rick 
Bovid 

It is a striking fact that two extremely flattened crania have 
been discovered in the deposits of Hay Springs, Sheridan County, 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1011 



Nebraska, which are considered of Lower Pleistocene (Aftonian) 
age. The first is the flattened cranium (Fig. 899) in thc^ Nebraska 
Museum; the second is the flat cranium (Fig. 900) in the Ameri- 
can Museum of Natural History. The flattening of these crania 
is a proof of the heavy geologic pressure to which these beds were 
formerly subjected. 

Nebraska Museum. — In figure 899 is represented the palate 
of a flat skull referred by Prof. E. H. Barbour to Elephas imperator 
(Neb. Mus. 1-11-8-17E); it was found in 1917 at Hay Springs, 
Sheridan County, Nebraska, in a bed of diatomaceous earth, 
flattened to four inches in thickness. The broad-plated grinding 
teeth enable us to confirm this reference, the molars presenting 
marked resemblance to those of the typical Archidiskodon im- 
perator. The geologic age of this specimen is very important, 
since the Hay Springs fauna is now regarded as of Lower Pleistocene 
(Aftonian) age.' In describing this skull, Barbour observed ten 




exposed ridge-plates, also three on the deeply worn anterior 
portion of the tooth, thirteen worn ridge-plates in all; the total 
number of ridge-plates is unknown. The tusks are very massive in 
transverse section and closely approximated where they issue from 
the alveoli; when found the left tusk was 9 feet long, but it was 
partly destroyed by careless collecting. 

American Museum Crushed Skull (Fig. 900). — This 




Fig. 899. ArckiiliskodoH imperator cranium of aged male (Neb. Mus. 
1-11-8-17E) foimd in a bed of dJatomaceou.s earth, flattened to four inches in 
thickness; discovered in 1917 at Hay Springs, Sheridan County, Nebraska. 

'[Middle Pleistocene (see Barbour and Schultz, 1937.1, p. 3).— Editor.] 



Fig. 900. Archidiskodon imperator cranium of adult male (,\mer. Mus. 
17355), extremely flattened, exhibiting small posterior nares and cranial 
foramina. Discovered in 1916 at Hay Springs, Sheridan County, Nebraska. 



1012 



OSBORN: THE PROBOSCIDEA 



crushed cranivim (Ainer. Mus. 17355) closely resembles in size and 
character that in the Nebraska Museum (Neb. Mus. 1-11-8-17E). 
It was discovered in 1916 by Albert Thomson in the famous Hay 
Springs quarry, Sheridan ( 'ounty, Nebra-ska. The cranium belongs 
to a fully adult male, somewhat less aged than the Nebraska 
Museum cranium (Fig. 899), because the third superior molars, 
r.M', I.M', are less worn, displaying anterior ridge-plates. Besides 
the anomalous crushing, this specimen finely displays the char- 
acters of the palate, especially the 'backward' posterior nares, 
which contrii-sts with the very large 'forward' posterior nares of 
the superb Parelephas jeffersonii cranium (the Franklin County 
Mammoth) in the Nebraska Museum (Neb. Mus. 1-4-15 — see Fig. 
963), which for a time was erroneously referred to E. [ = Parelephas] 
rnlitmbi. This palate of Archidiskodon imperalor (Amer. Mus. 
17355 — Fig. 900) should be compared with that of Parelephas 
jc_lfersonii, which more nearly resembles that of Elephas indicus 
bengalensis (Fig. 800). The backwardly placed opening of the 
posterior nares appears to be a very distinctive character of 
Archidiskodon. 

Hay Springs, Nebuaska, Fhick Collection. — The follow- 
ing specimens, collected under the direction of Mr. Childs Frick, 
serve to confirm Archidiskodon imperalor as the characteristic 
proboscidean of the Hay Springs horizon. 

Amer. Mus. numbers: 25501 A, fragment of superior molar, 
riflgc-iilate maximum height 244 mm.; 25501, posterior jiortion 
of inferior molar, ridge-plate maximum height 164 mm.; 25501B, 
anterior portion of extremely worn superior molar, maximum 
width 111 mm. ; 25500A, middle portion of an aged inferior molar, 
width 97 mm., laminar frequency 6K in 10 cm. ; 25500, portion of 
half-worn superior molar, width 107 mm., laminar frequency 6 in 
10 cm.; 25505 A-D, portions of inferior milk molars; 25506, por- 
tion of first milk molar. 

CRANIAL MATERIAL IN THE NEBRASKA STATE MUSEUM 
The finest Archidiskodon collection is in the Nebraska State 
Museum as recently listed by Director Barbour (1925.3, pp. 117, 
118) in connection with his description of the "Columbian Mam- 
moth Elephas Maibeni." This list is entitled "Columbi Material 
in the State Museum"; as revised by Barbour to July, 1925, it is 
as follows: 
Skulls 

11-3-13 Skull, mandible, and teeth from sandpit 6 miles 
due south of Hastings, Adams ('o., Nebraska. 
Adams County Mammoth. 

[ = Archidiskodon imperalor.] 
1-4-15 Skull, mandible, and tusks complete, of extraor- 
dinary size. Campbell, Franklin Co., Nebraska. 
Franklin County Mammoth (compare Figs. 
963 and 964 of present Memoir). 

[ = Parelephas jeffersonii.] 
16-6-16 Skull, tusk, teeth, and femur. Callaway, Custer Co., 
Nebraska. Cluster (Bounty Mammoth. 

[ = .4 rchidiskodon intpcrator. \ 

2-7-17B Skull, mandible, teeth, tusks, two tibise, scapula, 

and ribs. Found 7 miles south of I''arwell, near 

Dannebrog, Howard Co., Nebraska. Howard 

County Mammoth. [ = ^4 rchidiskodon imperalor.] 



5-9-22 Skull, mandible, tusk, forelimbs, six cervicals, 
several dorsals, four lumbars, sacrum, femur, 
part of pelvis, ribs, etc. Lincoln Co., Nebraska. 
Described as Elephas {Archidiskodon) maibeni. 
Lincoln County Mammoth. 

[ = Archidiskodon imperalor 7naibeni.] 
1-4-26 Skull, teeth, and tusks of a form apparently identi- 
cal with A. maibeni; the same type of teeth and 
the same size and curvature of tusk (see Barbour, 
1926.1, 1). 122). From Lingle, Wyoming. 

1 = Archidiskodon imperalor maibeni.] 
[1-11-8-17E Palate of a flat skull referred by Barbour to 
Elephas imperalor, Hay Springs, Sheridan Coun- 
ty, Nebraska. Sheridan County Mammoth. 

[ = Archidiskodon imperalor.] 

Mandibles 

19-9-17 Mandible, left half, with teeth. Powell, Jefferson 
Co., Nebraska. [ = Archidiskodon imperalor.] 

29-25-11-18 Mandible and teeth. Inland, Clay Co., 
Nebraska. [ = Archidiskodon imperalcrr.] 

2K-3-8-19 Mandible and teeth. Republican City, Harlan 
Co., Nebraska. [ = Archidiskodon imperalor.] 

8-7-08 Mandible and teeth. Young. Benkleman, Dundy 
Co., Nebraska. [ = Archidiskodon imperalor.] 
18-2-22 Mandible with teeth. In Aftonian gravels. Staple- 
hurst, Seward Co., Nebraska. Type of Elephas 
scotli Barbour, 1925. [ = Archidiskodon im- 
peralor scotli, or juvenile A. imperalor.] 
23-6-14 Mandible with teeth. From Crete, Saline Co., 
Nebraska. Type of Elephas hayi Barbour, 1915. 

[ = Archidiskodon harji.] 

Grinding Teeth 

Largely as plotted by Hay (1924) under Elephas [ = Archi- 
diskodon] imperalor, thirty-two or more upper and lower grinding 
teeth, usually associated, from the following counties in Nebraska, 
Kansas, Indiana, and Wyoming: 

Dawes, Furnas, Jefferson, Platte, Richardson, (lagc, Cass, 
Thayer, Howard, Adams, Fillmore, Buffalo, Cheyenne, 
Valley, Harrison, Clay, Butler — of Nebraska; also 
Herndon — of Kansas, Tipton — of Indiana, and (loshen — 
of Wyoming. 

Neb. Mus. 1-4-26 Archidiskodon imperalor maibeni ref. of 
Lingle, (Joslien County, Wyoming. 
Walls of cranium \ery sloping; maxi- 
mum transverse measurement of pal- 
ate 30 in., breadth of nasal opening 
20 in.; circumference of tusk 25 in.; 
ridge-plates as thick as those in A. 
imperalor. 

5-11-20 Arrhidi.skodon imperalor ref., from Bell- 
wood, Butler County, Nebraska (cast 
Amer. Mus. 20069). Posterior half 
of a third right inferior molar, r.Ms, 
with eight broad ridgc-i)lates; six 
arched ridge-plates in 20 cm. as com- 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1013 



pared with nine arched ridge-plates in 
20 cm. in the corresponding tooth of 
Parelephascolumbi (Amer. Mus. 13707, 
I.M3); breadth 4)-^ in. = 112 mm. This 
is the most broadly ridge-plated molar 
on record. 
4-12-13 Portions of inferior molars and femur. 
Arched laminiE of inferior molar, four 
in 10 cm., as compared with three 
arched laminse in 10 cm. (Neb. Mus. 
5-11-20). Femur 56 in. long or 1422 
mm.; same measurement as that of 
Neb. Mus. 13-24-10-14 (see Fig. 908). 




Fig. 901. Referred superior molar, r.M', of Archidiskodon imperator in the Geological 
Institute of the City of Mexico (No. 207). After photograph kindly furnished by Senorita 
Reyes. One-half natural size. Compare Reyes, 1923, fig. 3. This tooth was described 
by Senorita Reyes (1923, p. 229) as follows: "Fig. 3. Escala: Vt. Ejemplar num. 207 del I. 
Geologico. Elephas imperator, Leidy. Molar derecho." Locality: Zumpango, Mexico. 

Observe that there are b% ridge-plates in 10 cm. as compared with 9}i in P. colurnhi 
felicis. Length of worn surface 220 mm., breadth 82 mm. 



ARCHIDISKODON IMPERATOR IN THE MUSEUMS 

OF THE CITY OF MEXICO 

See Parelephas colurnhi below (Chap. XVII) 

We owe to Osborn (1905.270), to Freudenberg (1922), and 

finally and most fully to Senorita Reyes (1923) descriptions of the 

superb materials referable to Archidiskodon imperator and to 

Parelephas colurnhi in the museums of the City of Mexico. 

Progressive Stages (Mexico). — As described by Osborn 
(1905.270, p. 931): "The elephant remains in the National Mu- 
seum have usually been ascribed to Elephas columbi; but they in- 
clude molar teeth not only of this species, but of the much larger 



form, Elephas imperator. In the collection of the Geological Survey 
of Mexico in the new survey building are the skull and tusks of an 
E. imperator of magnificent proportions, the tusks measuring 5 m. 
10 cm., or 16 feet 10 inches [4 m. 20 cm. est., or 13 ft. 9% in., as re- 
measured by Barnum Brown] in length ; this specimen was secured 
during the excavations for the great drainage canal of the Mexican 
Valley." 

In the collection of the Geological Institute of Mexico, under 
the direction of Dr. Jose G. Aguilera, there are also several single 
teeth of Archidiskodon imperator, molars of the Parelephas columbi 
type from the \'illage of Zacapii in Michoacan and of A . imperator 
from the valley of Puebla (cf. Osborn, op. cit., p. 931). 

Reyes, 1923. — The specimens described and figured (1923) by 
_ Seiiorita Reyes, are as follows: 

Elephas imperator 

Escuela de Ingenieros No. 1. Jaw containing 
M 3, with 6 ridge-plates e.xposed. From Tequix- 
quiac. Figs. 1 and 2, p. 228. 

Geol. Inst. 207. Palate with r.M^ 13 ridge- 
plates exposed; length 220 mm., breadth 82 mm. 
From Zumjjango. Fig. 3, p. 229 (see Fig. 901 of 
the present Memoir). 

Geol. Inst. 210. R.M», length 300 mm., 
breadth 115 mm., 12)i: ridge-plates in 25 cm. From 
Tepexpan. Fig. 4, p. 230. 

Geol. Inst. 212. Cranium of young [adult] 
individual, a beautiful specimen well conserved. 
It is without tusks. Enormous prominence of 
occipital crest. Associated lower jaw. Dimen.sions 
of second superior molar, M- : length of worn area 
168 mm., breadth 108 mm. From Tepexpan. Fig. 
7, p. 233, Fig. 8, p. 234, and Fig. 9, p. 235 (see 
Fig. 902 of the present Memoir). 
Posterior view (Fig. 7). 

Anterior view (Fig. 8), forehead concave. 
Summit of occipital crest broadly rugose. 
Crown view of lower jaw (Fig. 9). 
Geol. Inst. 211. Left ramus of jaw' with I.M3, 
12;^ ridge-plates in 25 cm., length 220 mm., breadth 
98 mm.; length of femur 1360 mm. From Tepex- 
pan. Fig. 5, p. 231, Fig. 6, p. 232 (see Fig. 903 of 
the present Memoir) . 
Archidiskodon (?)hayi Ref. in Mexico. — A mandible' (Fig. 
903) from Mexico (Geol. Inst, v-211), referred by Seiiorita Reyes 
to Elephas hayi, supports the evidence afforded by Barbour's type 
of E. hayi that an Upper Pliocene stage of Archidiskodon, apjjarent- 
ly similar in progression to the 4. planifrons of Asia and of southern 
France, entered North America. 

Osborn, 1924: This mandible from Mexico is intermediate 
between the E. hayi type of Barbour and the E. imperator type of 
Leidy; it shows a prolonged rostrum (see Fig. 903) ; it appears to 
us like a progressively modified rostrum derived from an ancestor 
with a jaw like that of .4 . planifrons. While this mandible has been 
referred to Elephas hayi by Seiiorita Reyes, it seems to Osborn to 



'Osborn, 1929: .\ similar mandible is described below (p. 1033) as Archidiskodon sonoriensis, to which species tliis specimen may also be referable. 



1014 



OSBORN: THE I'ROBOSCIDEA 





Ckanium ok an Adult Male ok Auchidiskodon impehatou is the Geological Museum ok Mexico 
mature cranium of Archiili.'ikoihii impcrator in the Geological Institute of the City of Mexico (Geol. Inst. 212), 



Locality: Tepexpan, Mexico. 



reduced to one- 
Compare Reyes, 1923, figs. 



rig. 902. Keferrei 
twelfth natural size. After photograph kindly fiu-nished the i)resent author by Scnorita Reyes 
7, 8, and 9. 

This robust male cranium lacks the sharply peaked, acrocephalic, occipital crest structure seen in the young male(?) cranium from Rancho La Brca (Fig. 
89.>). Judging by the elongated and massive alveoli and rugose exoccipital muscular attachments this cranium supported an enormnus pair of tusks. The 
adult male tusks are estimated at 13H 16+ feet in length (see Fig. 894) 



represent rather an intermediate and distinct sj^ecies in which there 
is a much larger number of ridge-plates than in the t yi)e of E. hniji. 
It is another proof of how much we have still to learn regarding 
the characters and migrations of Arrhidhhodon in North America. 
Doubtless future excavation will re\eal additional material 
of great value. We turn to Frcudenberg (1922) for his views 
regarding the relationshi]) of the true Elcphas coliiinbi Falconer to 
the true E. imperator of Leidy (see j). 1017 below); also for his 
remarks on the relationship of A. impcrator to the A. meridimiaUs 
of Europe and to the A. planifrons of Asia and southern France. 

AnCESTKV of ElKPHAS IMPKHATOll (FnKUDIONBEKG, 1922, 

SoERGEL, 1915').— (Freudenberg, 1922, p. 171) : "Die Mastodonten 
wanderten fri'iher nach Amerika iiber, die Elefanten s])ater. Die 
innerasiatischcn Hochliinder, die den meisten Saugetierstiimmen 
den Ursprung gaben, sind auch liier in diesem Fall als Ileimat der 
mexikanischen Arten anzusehen. [I'Dotnote: 'V'gl. W. D. Matthew, 
( 'limatc and I'A-olution. Annals of th(! New York Acad, of Science. 
\'ol. 24. pag. 171-318.']. Fiir die Elefanten gilt das mit ziemlicher 
Sicherheit. Eine Ableitung des El. iinprraior von El. nicridioruilis 
Euroi^is, wie Soergel |l'"ootnote: '\V. Soergei, Die Stammesge- 
schichfe der I'^lefanten. I\'. Die anierikanischen Elefanten. 
Cenfralhl. f. Mineralog., Oeolog. u. Paliiontolog. 1915. No. 9. 
pag. 27S 2S:}.'| (las will, ist gcsiicht. El. plnii/frons isl weit eher 
der Stammvater alter spiilerer Eiefaiil idcii als der /s7. in<riili(iii(ilis, 

'Sec citation below, page 101."). 




Fig. 903. Mandible of Elc/ikax [ = ArchUIUkodon] hayi{l) ref. in the collec- 
tion of the Geological Institute of the City of Mexico (v-211). Photograph 
through the courtesy of Scnorita Reyes (conipaie Reyes, 1923, figs. ."> and (i). 
About ono.seventli natural size. The prolongation of the .symphy.seal rostrum 
suggests a remote resemblance to the type of Arrliiiliskniliiii hiii/i Barbour 
(l'"ig. 913 of th(; present Memoir); this feature, however, is more character- 
istic of A. sonoriensis. [See footnote on p. 1013. — Editor.) 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1015 



der auch sicher niit Unrecht als der Ahne des EI. antiquus von 
gcnanntem Autor angesprochen wird." 

Ancestry of Elephas imperator, Soergel, 1915, p. 281. — 
"Wir iniissen annehmen, dass die nach Anierika iiberwandernden 
Fornien der Meridionalis-Trogontherii-Reihe schon beim Uber- 
wandern resp. kurz vorher besondere Charaktere gegeniiber dem 
europaischen El. trogontherii merid7onali.s ausgebildet hatten, 
Charaktere, die in der weiteren Entwicklung sich zum 'Inipemtor- 
Typus' steigerten. Es war also im altesten Diluvium im Kreis der 
kontinentalen Elephantenformen eine Variationsbreite mit 2 Polen, 
El. trogontherii ineridionalis im Weston und dem tlirekten Vor- 
fahren des El. imperator im Osten \orhanden. Es ist klar, dass 
als direkter Vorfahr beider Pole die nachstJiltere Mutation in der 
kontinentalen Reihe zu gelten hat, das ist El. meridionalis des 
altesten Diluviums, vielleicht auch des Oberpliociin. Als Vorfahr 
des El. imperator Leidy hat also jedenfalls EI. meridionalis Nesti 
zu gelten." 

In our judgment Freudenberg was less fortunate in his treat- 
ment of EIepha.s imperator as a subspecies of E. rolumbi. For 
reasons originally stated by Osborn (1922.555) the type of Elephas 
columbi Falc. is readily distinguishable from the type of Elephas 
imperator Leidy; consequently the treatment of E. imperator as 
a subspecies of E. rolumbi is invalid; it should stand E. [ = Archi- 
diskodon] imperator. 

Synonyms of Archidiskodon imperator. — The four sub- 
species of Elephas rolumbi proposed by Freudenberg (1922) in this 
manner are redetermined by Osborn as follows: 

El. Columbi var. Felieis Freudenberg, 1922, p. 147, Taf. xvi 
(viii), fig. 4, Tecamachalco, Puebla, Mexico 

= Parelephas rolumbi felieis. 

El. Columbi var. silvestris Freudenberg, 1922, p. 152, fig. 19, 
Ejutla, Oaxaca, Mexico =Arrhidi.'<kndon imperator silve.'ttris. 

EI. Columbi var. Fakoneri Freudenberg, 1922, p. 153, fig. 21, 
Tequixquiac, Mexico = Archidiskodon imperator fakoneri. 

El. Columbi var. imperator Freudenberg, 1922, pp. 160-171 

= Archidiskodon imperator. 

The subspecific value of Archidiskodon imperator silvestris 
and of .4. imperator falroneri remains to be determined. We may 
here quote in full Freudenberg's description of these subspecies and 
reproduce his type figures. 

Archidiskodon imperator silvestris Freudenberg, 1922 

Figure 904 
Ejutla, vState of Oaxaca, Mexico. Probably Lo\ver(?) Pleistocene. 
The type of this subspecies, related by Freudenberg to the 
species Elephas [ = Parelepha.s] columbi, is obviously referable as 
a synonym or subspecies of Archidiskodon imperator. The type 
(Fig. 904), with at least 16 ridge-plates, is too elevated to be re- 
garded as an M-; it appears rather to be a left M'; the six ridge- 
plates exposed (Fig. 904 B) resemble those of A. imperator rather 
than of P. columbi; this Ejutla type has the very broad outer 
coating of cement characteristic of A. imperator and lacking in P. 
columbi. 

El. Columbi var. silve.slri.'i Freudenberg, 1922. "Die Siiugetier- 
fauna des PliocJins und Postpliociins von Mexiko." (!eol. und 
Palaeont. Abhand., N. F., Band XIV, Heft 3, pp. 152, 153. Type. 



Second [third] superior molar of the left side, M- [l.M'']. Pal. Coll. 
Univ. Leipzig 4402. Horizon and Locality. — Ejutla, 

Estado de Oaxaca, Mexico. Type Figure. — (Op. cit., p. 146, 

fig. 19): "Fig. 19. a. El. Columbi va,r. silvestris. Kronenansicht 
des zweiten oberen Molaren, in % nat. Gr. b. Seitenan,sicht in 
etwa Js nat. Grosse. Original in Leii)zig, von Ejutla, Estado de 
Oaxaca in Mexiko. Coll. Felix, wohl aus jiingerom Diluvium. 
Subtropische Waldform." 




B 




1'reudenberg's Type of Archidiskodon imperator silvestris 
Fig. 904. Type, left M', of Eleplias Columbi var. silvcslris Freudenberg, 
1922, I). 146, fig. 19, reduced to one-third natural size. From Ejutla, Estado de 
Oaxaca, Mexico. After photograjjlis sent by Doctor Freudenberg. Original in 
Leipzig (Pal. Coll. Univ. Leipzig, 4402), Coll. Felix. B, Crown view; A, side 
view. 



Type Description. — (Op. ril., p. 152): "3. Dor Zahn von 
Ejutla {El. Columbi var. silvestris Freudenberg). Wenden wir uns 
genauer dem oberen Molaren von Ejutla (Estado de (_)axaca) zu, 
welchen Felix als 'El. Columbi' etikettiert und nur mit ^'orbe- 
halt wegen der relativ enggestellten Lamellen unter El. primi- 
genius Blumenbach var. angefiihrt hat. Die.ser Zahn triigt jctzt 
die No. 4402 der jialiiontologischen Sammlung der UniversitJit 
Leipzig. Es ist cin zweiter obercr Molar der linken Seite. Ich 
bilde ihn ab von der Seite als Textfig. 19b und a und von der Kau- 
Hiiche aus. Das Hinterende weist eine merkliche Einwiirtsdrehung 
iler Lamellen auf und zugleich eine \'erticfung in 5 cm Entfernung 



1016 



OSBORN: THE PROBOSCIDEA 



unter der letzten angekauten Lamelle. Diese Grube halte ich 
trotz ihrer Rauhigkeit fur eine Pressionsmarke, hervorgebracht 
voin nachdrangenden M^ Die vordersten 4 Lamellen sind abge- 
brochen; einige davon diirften abgekaut sein. Dadurch ist die 
Lamellenzahl im Minimum = 16." The author rightly contrasts 
this species as identical in the ridge formula of M' with that of the 
subspecies Parelephas columbi felicis. He remarks (op. cil. , p. 153) : 
"In der Art der Abkauung gleicht dieser Zahn sehr dem von 
Pohlig— Nova Acta Acad. C. L. C. G. Nat. Cur. Vol. 57 abgebil- 
deten M^ des El. Americae Pohlig aus Chihuahua, Mexiko. Es 
wird auch einM'von dort an gleicher Stelle abgebildet mit dersel- 
ben schiefen Stellung der Lamellen zur Liingsachse des Zahns wie 
an den Molaren von Ejutla und ganz verschieden von den Hoch- 
plateauformen der Mesa central von Mexiko oder Puebla. Es 
empfiehlt sich aus oben genannten Griinden nicht, Pohligs Namen 
El. Americae [^Elephas americanus De Kay, 1842] fur die Wald- 
form des El. Cnlumbi anzuvvenden, dessen Verbreitungsgebiet die 
siidwestlichen Randgebirge des mexikanischen Hochplateaus 
warcn. Ihnen gegeniiber stehen die als El. Columbi var. Felicis 
aus der Mesa central bezeichneten Steppenformen. 

Archidiskodon imperator falconer! Freudenberg, 1922 

Figure 905 
Tequixquiac, Valley of Mexico; probably Lower(?) Pleistocene. 
This Tequixquiac jaw (Fig. 905) represents the cotype and 
is undoubtedly related to Archidiskodon imperator rather than 
to Parelephas columbi. From the figure it is difficult to give the 
characters of this jaw, but it appears to be somewhat longer and 
more i)rimitivc than the type of A. imperator. The lower grinding 
teeth apparently present the following formula: M 3igTT7- In 
this jaw the 16+ ridge-plates exposed in the I.M3 are too widely 
separated to be related to P. columbi; the I.M3 measures ap. 175+ 
mm., tr. 95 mm., the dimensions a])parently equal those of P. 
columbi, namely, neotype I.M3 ap. 298 mm., tr. 91 mm. The 
locality of Tequixquiac yields both A. imperator and P. columbi, 
according to Reyes. 

El. Columbi va,T. Falconeri Freudenberg, 1922. "Die Siiugetier- 
fauiia des Pliociins und Postpliocans von Mexiko." Geol. und Pala;- 
oiit. Abhand., N. F., Band XIV, Heft 3, pp. 153-160. Lecto- 

TYPE. — First left lower molar, l.Mi, Mus. Royal Coll. Surg. 
741a.' Cotype. — Lower jaw with both third molars M3 in 

situ. Horizon and Locality. — Tequixquiac, Mexico. Lec- 

TOTYPE Figure. — After Falconer, 1863.1, PI. 11, fig. 1. Cotype 
I'iGURE. — Freudenberg, 1922, p. 154, fig. 21. 

Description. — {Op. cil., p. 153): "Als Typus der Form gilt 
mir der von Falconer abgebildet* Mi von Mexiko (London) und 
ein M3, den ich auf S. 54 [154], Fig. 21 abbilde. Dieser M' [M3] 
gehrirt wohl der linken Seite an. Er stammt vermutlich aus dem 
\'alle de Mexico. Die Hinterseite des Zahnes ist abgebrochen. Die 
Zahl der fehlendcn Disken lasst sich nicht mehr bestimmen. 
l'-hcns()\vcnig weiss man, wie viele Lamellen vorn durch Abkauung 
verschwunden .sind. Die Kaufliiche ist intakt und wurde in nattir- 
lichcr CJrosse abgebildet. Die Liinge der Kaufliiche misst 175 mm, 
die Breite in der Mitte, einschliesslich des Zements, misst 95 mm. 
Die Kaufliiche ist also etwa doi)pelt so lang wie breit. Der Ver- 
schmelzungstypus der eben erst angekauten Disken (im Sinne der 

'[By inference this is Professor Osborn's IcctotyiJC. It is not figured, however, in tlie present Memoir 



Soergelschen Nomenklatur) ist median lamellar, lateral anular. 
Besonders an den drei ersten Jochen ist dieser Bau der Schmelz- 
pfeiler zu beobachten. Ein mir vorliegender M'' von El. trogon- 
therii Pohlig [Footnote: '1) Abgebildet in meiner Arbeit: Die 
Siiugetiere des iilteren Quartiirs von Mitteleuropa etc. Geolog. u. 
palaont. Abh. N. F. Bd. 12. Heft 4/5. Jena 1914. t. 3. f. 5. Siehe 
auch unsere Taf. ix [xvii].'] von .Jockgrim in der Pfalz hat eine 
Kaufliiche von 195 mm bei einer maximalen Breite von 82 mm. Das 
Verhilltnis von Liinge zu Breite ist also iiber 2: 1 (Breite=l). 
Wilhrend bei El. Columbi die Verhaltniszahl f iir die Lange (Breite = 
1) unter 2 ist." 

Characters. — (Op. cil., p. 154): "Als altquartiir haben sich 
jene Formen herausgestellt, welche wir nach Falconer als Elephas 
Columbi var. Falconeri bezeichnen miissen. Ihre Lamellen sind 
kiirzer und stehen isolierter (wie z. B. an dem Taf. viii [xvi], Fig. 
1 abgebildeten Molaren) verglichen mit der var. silvestris und erst 
recht mit der var. Felicis. Bei der typischen (altciuartiiren) 




CLCPHis PRiaicemus. 



CoTYi'E OK Ahchidi.skodon imperatok falconehi Fhemdenbeho 
I'^ig. 905. Cotype jaw (one-sixth natural size) of Eleplias Columbi var. 
Falconeri I'Veudenberg, 1922, p. 154, fig. 21; crown view one-half natural size. 
From Tequixquiac, Mexico. Originally figured by Villada, 1903, lAm. vm. 



-Editor.] 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 



1017 



Columbi-Form. der gerade auch das Original Pohligs aiis Mexiko 
zugehort, kommen am oberen INP vier Lamelien auf 53 mm, 
wiihrend deren 6 sich auf die gleiche Streeke verteilen bei dem jung- 
diluvialen Zahn von Mexiko Taf. viii [xvi], Fig. 4. Es mag nun 
dieser weniger Waldform als Steppenform gewesen sein." 

EI. Columbi var. imperator Leidy 
[ = Archidiskodon imperator] Freudenberg, 1922 
Spokam Bar near Helena, Montana. 

Under the true specific name Archidiskodon imperator should 
apparently be included the El. Cohimbi var. imperator of Freuden- 
berg, 1922, in the Geological Mu.seum of Bonn. 

El. Columbi var. imperator Leidy Freudenberg, 1922. "Die 
Saugetierfauna des Pliociins und PostpliocJins von Mexiko." Geol. 
und Palaeont. Abhand., N. F., Band XIV, Heft 3, pp. 160-171. 

Freudenberg frequently introduces the subspecific name 
Elepha.^; columbi var. imperator Leidy (op. cit., pp. 155, 156) and 
finally definitely employs it on page 160. Type Description. — 
{Op. cit., p. 160): "5. El. Columbi var. imperator Leidy. Wegen 
der vielfachen Verwechslungen und wegen der mutmasslichen Ab- 
stammung des El. Columbi von El. imperator ist es niitig, naher auf 
diese Form einzugehen. Es ist das um so notwendiger, als H. F. 
Osborn [Footnote: 'Recent Vertebrate Palaeontology. Fossil 
mammals of Mexico. Science. Vol. 21. 1908 [1905]'.] diese Art aus 
Mexiko anfiihrt. Osborn griindete seine Bestimmung wohl in 
erster Linie auf den gewaltig langen Stosszahn, der 16 Fu.ss misst 
und somit der langste Stosszahn ist, von dem eine Kunde vorliegt. 
Aber warum sollte nicht ein El. Columbi gelegentlich riesige Dimen- 
sionen erreichen, zumal da diese gute Art in eine ganze Anzahl von 
Unterarten einmal zerlegt werden muss, sowie Matschie 4 Arten des 
afrikanischen Elefanten heute unterscheidet." 

Characters. — After quoting Leidy, Osborn, Lull, Lucas, 
Holmes, Pohlig, Soergel, and Cope {op. cit., pp. 160-169), Freuden- 
berg concludes (p. 170): "'El. imperator' gewinnt mit seinem 
Aufsteigen ins jiingere Diluvium immer mehr Aehnlichkeit mit den 
englamelligen, schmalkronigen, eigentlichen 'Columbi-Rnssen.' 
Vielleicht losen sich die letzten Herde des waldlebenden Kaiser- 
elefanten auf im Co/uw6/-Hauptstamm als sogenannte var. 
silvetitris. Das hiingt doch wohl mit dem immer weiter ziiruck- 
weichenden Seen- und Waldklima zusammen, das offenbar der 
Existenz der Imperator-Rasi^e in der Sonorischen Faunenprovinz 
giinstig war. Die zunehmende Austrocknung der diluvialen Seen- 
gebiete forderte die Entstehung von manmiutahnlichen Columbi- 
Rassen, die wohl vorwiegend auf die sparliche Koniferenkost und 
auf harte Steppenpflanzen angewiesen waren. Die im tropischen 
Laubwald iisenden Imperator-Hassen, die, wie El. indicus auf 
Ceylon gute Bergsteiger gewesen sein mochten, haben bei Af ton wie 
im Becken von Puebla, Chihuahua und Mexiko als deutlichen 
Hinweis auf einstige Holzasung tief ausgehohlte Zementintervalle 
zwischen den ausserst kraftig gebauten Dentinpfeilern und ihrer 
oft stark gefalteten Schmelzhiille." 

SKELETAL CHARACTERS OF ARCHIDISKODON 
Much remains to be done in establishing the skeletal characters 
of Archidi.'ikodon imperator. In the flu\'iatile sand and gravel 
deposits of early Pleistocene time, in which these imperial mam- 
moths occur, skulls, jaws, and skeletal parts are widely scattered. 



The principal partly or fully associated materials hitherto described 
are the following: 

Forelimb from Tule Canon, Briscoe 
County, Texas (Amer. Mus. 10598). 

Femur from near Reynolds, Jefferson 
County, Nebraska (Neb. Mus. 13-24- 
10-14). 

Forelimbs associated with numerous ver- 
tebrae, skull, jaw, one tusk, and parts 
of hindlimbs — the finest material thus 
far discovered [ = Archidiskodon im- 
perator maibeni] from near Curtis, 
Lincoln County, Nebraska (Neb. Mus. 
5-9-22). 



Figs. 907 and 906. 
Fig. 908. 

Fig. 912. 




Archidiskodon imperator. Aqed Male 
Fig. 906. Outline reeon.struotion from right forelimb in the Ameriean 
Museum (Amer. Mus. 10598 — Fig. 907) and from skull (Amer. Mus. 14476 — 
Fig. 896), one-sixtieth natural size. 

The forelimb (Amer. Mus. 10598) measures 348 2mm. or 11 ft. .') in., as 
shown in figure 912. This gives a shoulder height in the flesh of .3702 mm., 
whirh may be taken as a conservative estimat<^ of the shoulder height in the 
flesh of a nine-tenths grown Archidiskodon imperator. Compare figure 912, 
also figure 907, lateral and front views of the same forelimb, one twenty-fourth 
natural size. The diagram of the forelimb, found in 1899 near Tule Canon, 
Texas, is taken from the mounted forelimb in the Ameriean Museum (Amer. 
Mus 10598). The diagram of the skull is taken from Archidixliodon imperator 
(Amer. Mus. 14476), found at Victoria, Texas. 

Fore- and Hindlimbs of A. imperator (Fig. 907). — For- 
tunately the complete right forelimb (Amer. Mus. 10598), dis- 
covered in 1899 near Tule Canon, Briscoe County, Texas, along 
with other parts of the skeleton belonging to a single individual, 
gives us the means of estimating the height of Archidiskodon 
imperator, namely, 12 ft. Wa in. =3702 mm. at the withers, 13 ft. = 
3960 mm. at the top of the head when elevated. This elevation is 
shown diagrammatically in figure 906 which combines the skull 
(Amer. Mus. 14476) from Victoria, Victoria County, Texiis, with 
the forelimb (Amer. Mus. 10598) from Tule Canon, Briscoe 
County, Texas. 



1018 



OSBORN: THE PROROSCIDEA 



In the licli collections made by Professor Barbour for tiie 
State .Miiseiini, I'liiversity of XclHaska, are a femur (Fis. 908) and 
a humerus which may be comiJaretl with corresjjondinf^ limb ele- 
ments in the skeleton of "Jumbo" {Loxorlonta afrirana oxyotis) 
in the American Museum. The specific reference is somewhat 
uncertain. The fenun- probaby does not represent a full-grown 
A>-chidisko(Ion imperator, which would greatly exceed the specimen 
here represented in the pliotoKraph (Fifi. 90S) kindly furnished by 
Professor Harbour. 





I'oltKT.lMB OK ArCHIDISKODON IMI'EIIATOK. A.VIEHirAN MusEUM 

Compare Figure 912 

rin- iH)7. Rislit forclimb roforrod to ArclwHskudon imperator, a.s moiintcil 
in tlio .AmiTicaii Museum (.Amer. Mus. 10.")98). one twenty-fourth natural 
-.ize, found l)y .\ll)an Stewart in 1899 nearTule Canon, Bri.seoe Covnity, Texas, 
;il()nK with other parts of the skeU'ton belonging to a singh' in(livi<luah 

The same forelimb is represented, vertically extended, in figure 012. The 
vertieal measurements of the separate .segments are: Scapula 1017 mm., 
humerus 109.') mm., radius 890 mm., manus 480 mm., sum of total vertical 
height 3482 mm. or II ft. .'> in. .Xs the limb is always somewhat flexed, the 
cartilages, foot pads, and the flesh and skin above the shoulder give this 
animal a total height of 3702 mm., or 12 ft. ]\ in., a conservative estimate of 
the height of this animal. 




•ig. 908. l-iMH Bones of Ahchiuiskodo.n impek.^toh (Vouncj Adui.t) and 

l.OXODONTA AFRICAN A OXYOTIS COMPARED 

(bower) Three views (E) of femur of referred Arrhidiskmlon imperalor 
(Neb. Mus. 13 24 10 14) from Reynolds, .leffer.son County, Nebraska, one- 
tenth natural size. Height of fenuu- 1422 mm. =.")G in. (Barbour, 1925.3, p. 
llfi), .same measurement as the fenuw (Neb. Mus. 4-12-13), namely, 4.8 feet. 

(Upper) The same femur (Neb. Mus. 13-24- 10 14), one-twentieth 
natural size, compared with (.\, B) femur and humerus of Loxndonta tijri- 
cana oxyotis (skeleton of ".Jumbo") in the American Museum; also with 
(C, D) humerus and femur (latter computed) referred to Arc)niHifkn(Um im- 
peralor (Amer. Mus. IO.jOS). Compare figure 907 oi)posite. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1019 



HAUHonirS DESCRIPTION (1925) OF THE SKELETON OF 

ARCHIDISKODON MAIBENI 

(Soo type description of skull, jaws, and dentition below.) 

As above noted, this is the most nearly complete skeleton of 
Archidiskodon thus far discovered. The following is a free rather 
than a literal citation from Professor Barbour's \aluable paper of 
August, 1925 (Barbour, 1925.3), entitled, "Skeletal Parts of the 
Columbian Mammoth, Elephas maibeni, sp. nov." For reasons 
shown in the systematic treatment of E. (Archidiskodon) maibeni 
below, there appears to be no doubt that this skeleton is properly 



flocks of itoultry on the Karriger farm, before it was realized 
that they were out of the orilinary, after which the remaining 
parts were dug out and bared for with unusual appreciation and 
discernment. . . . The bones of Elephas maibeni were found pro- 
jecting from a loessial wall at the bottom of a small canyon. The 
general thickness of the loess at this point is about 100 feet. In an 
attempt to find additional material the writer, aided by Mr. 
William Hall and Mr. H.8. Karrigcr, blasted out many cubic yards 
of the loessial wall." 

"The skeletal parts preserved are the skull, mandible, one 




Fig. !MHI. Im|ii'iiul Mammoth (Archvlukodon iiii/xralor) of Nebraska and Texas. After restoration by Osborn and KniKJit, 190S. .\bout one-fiftieth 
natural size. 



referable to Archidiskodon rather than to Elephas; also that it is 
closely related to Archidiskodon imperator rather than to Parelephas 
colmnbi. The narrative of discovery is as follows: 

"The last and most remarkable specimen of the columbi 
[imperator] type was found in IJncoln County, about 16 miles 
north of Curtis, on the Karriger farm. It was discovered by Mr. 
and Mrs. H. S. Karriger, and was dug out antl preserved by them. 
Later it was procured of them for the palaeontological collections of 
Mr. Hector Maiben, who, next to Mr. Charles H. Morrill has been 
the most generous contributor of funds for the purchase and pres- 
ervation of choice Nebraska specimens. ... It should be recorded 
in connection with this specimen that an unknown number of bones 
and jjurts of bones were poimded uj) to furnish lime for the large 



tusk, the atlas, axis, and four other cervicals, several thoracics, 
lumbars, and the sacrum, ribs, and double ribs, both fore limbs and 
parts of the hind limbs. Both fore limbs are practically complete 
and are essentially ])erfect save that but one foot bone was found, 
hence the feet must be supplied. The hind quarters are represented 
by parts of the peh'is, the shaft of a femur, and the major portion of 
a fibula. The dentition is perfect. The molars have 14 ridges 
bonded together by an uncommon thickness of cement, which is 
a character of Columbian elephants. Even the great Columbian 
elephant from Franklin Count}' [referred to Parelephas jeffersonii 
in the present Memoir] seems surpassed in size. Heretofore, the 
tusks of the Franklin County elephant [Neb. Mus. 1-4-15] have 
been considered the largest repoi'ted, namely about ISli feet long 



1020 



OSBORN: THE PROBOSCIDEA 



with a maximum circumference, near the incisive sheath, of 29 
inches. The skull and mandible of Klephas maibeni is noticeably 
larger. The incisive sheath shows that the tusk had a diameter of 
lO'i inches and the incredible circumference of 33 inches." 

"The outstanding characters of Elephas [ = Aickidiskodon\ 
maibeni are size, extreme curvature and divergence of tusks and 
incisive sheaths, unusual shortness of centra coupled with great 
width. The tusk must have lain in a plane or nearly so and must 
have de.scribed a circle, the radius being 28 inches (711 mm.). 
The diameter of the tusk i.s 6^i inches (165 mm.) at the tip, 7U 
inches (190 mm.) four feet back of the tip, and IOJ2 inches (297 
mm. [267]) at the incisive sheath. Originally it was a magnificent 
jiiece fif ivory." 

"The fore quarters have bones unexpectedly large and mas- 



to note that the cancellous portion is uniform and continuous and 
is without partitions or vestiges of the dual origin. From this it 
may be inferred that they had been in coalescence for ages and 
that the character may have become fixed. At any rate it is not 
a ca.se of pathology. The great fore limbs, several vertebrae, and 
their corresponding ribs, have been mounted as an arch, a palaeo- 
zoologic arch [Fig. 910 of the present Memoir], through which all 
students and visitors must ptiss on entering the main floor. It is 
but a temporary mount which must be dismantled and moved into 
the new museum sometime in 1926, where the complete skeleton 
will be carefully articulated and properly installed. So many 
skeletal parts are at hand that this huge elephant when ready for 
exhibition will seem complete. Its proper installation demands 
a ceiling 18 feet high." [See Fig. 911 for present moimt. — Eflitor.] 




Fig. 910. I'Virolimbs of typo .skolefun of 
ArcMdinkoilon imperator maibeni as moimted 
in the Nebraska State Museum (Neb. Mus. 
.") 9-22) in the year 192.'). Professor Barbour 
is .standing in the background. The dimen- 
sions of this sii|)erb skeleton are shown (Fig. 
912) in direet comparison with the forclimbs 
of .1. imperator from Te.xas, of Loxmlontn 
(ifricana, and of Elephas indiciis. 

The cranium and tu.sks in the back- 
ground to the left, from Franklin County, 
Nebraska, belong to Parelephas jeffersoiiii. 
Barbour notes (192.').3, legend to Fig. .')8): 
'The great .skull and tusks in the background 
are tho.se of Elcjjhas columbi from .Jefferson 
|Franklin| County." 

[For |)resent mount, see figure 911 on 
following page. — Editor.] 



sive, especially the humerus. The humerus is huge beyond the 
visualization of those who have not .seen it, hence must judge of it 
from figures and measurements. In the hind quarters the bones 
are, if anything, less massive than might be exi)ected. They seem 
in contrast to those of the fore limbs, .lutlgiiig from the very short 
vertebriE the body must have been unduly foreshortened." 

"The centra of the vertebrae are very short, comi>ared with 
their width, ('onsecjuently certain ribs, presumably the fifth and 
sixth pairs, came in contact and became completely fused into one 
shaft with a double head and double tubercle, making huge and 
peculiar mammal ribs. A slight longitudinal depression is a vestige 
of the original boundary between the two shafts. It is interesting 



"Mammoth scapulae are large, heavy, and very thin in portions, 
so it seems the more remarkable that the two huge shoulder blades 
should have been preserved practically without blemish. The 
right humerus is likewise perfectly preserved; in the left the head 
is wanting but has been modeled on from the right humerus. The 
right ulna is perfect save that the distal epiphysis is missing. This 
has been modeled after the left ulna in which the ej)iphysis is 
l)resent but the shaft missing. . . . The fore liivibs, four vertebrae, 
two pairs of single ribs, and one i)air of double ribs of this ex- 
cej)tional mammoth are mounted in approximate position and 
make an impressive arch, the height of which is 13 feet from the tip 
of the toe to the top of the spine, see fig. 58 [Fig. 910 of the present 



THE MAMMONTINiE: ARCHIDISKODON AND METARCHIDISKODON 



1021 



Memoir; see also Fig. 912]". 

"In the flesh the height of Elephas [ = Archidiskodon] maibeni 
at the shoulder must have been about 13 feet, and the top of the 
head of this magnificent beast must have been about 14 feet above 
the ground." 

"From the tip of the toes to the top of the scapula is 11 feet, 6 
inches." 

"This specimen is believed to hold the record for size amongst 
the Columbian group of mammoths." 

"In point of .size, Elephas [ = Archidiskodon] maibeni was 
a rival of the Imperial elephant itself, which stood 13Ji feet high. 



The tallest living African elephant stands 11 feet high and the 
average elephant of the menagerie and circus 8 to 9 feet." 

Referring to figure 912 of the present Memoir, we observe 
that A 1 (right) is foreshortened, while A, B, C, D represent ortho- 
gonal full length projections of each limb segment, with the actual 
measurement of each segment in millimeters. This affords an 
absolutely reliable comparison of the ascending height of these 
four animals. 

The following entirely consistent comparative measurements 
appear to demonstrate that Archidi.ikodon rmperntor and the more 
primitive giant species A. maibeni towered in height far above the 
largest existing elephants : 



Fig. 912, A, Archidiskodon imperator maibeni 

B, Archidiskodon imperator 

C, Loxodonta africaiiri oxyoiis (".lumbo") 

D, Elephas indicus 

Fig. 1083, Hesperoloxodon antiqiiiis (Upnor) 





Mamm 


Forearm 


Humerus 


Scapula 


297 


(as mounted) 


1085 


1251 


1066 + 


424e 


(fully extended) 








480 


(fully extended) 


890 


1095 


1017 


361 


(fully extended) 


830 


1078 


925 


391 


(fully extended) 


744 


1008 
1290 


864 
1170 




Fig. 911. Present mount of the type skeleton of Archidiskodon imperator maibeni in Morrill Hall of the Universi- 
ty of Nebraska (Neb. Mus. 5-9-22): for dimensions of the forelimbs of this superb specimen, see figure 912, which 
was based on figure 910 showing the forelimbs as first mounted in 1925. In the left foreground is a skeleton of 
a juvenile Elephas indicus. .\fter original photograph kindly furnished the present author by Professor Erwin H. 
Barbour. 



r-^OOO 



Imperial Mammoth (mailreni) 
Imperial Mammoth Est height 

'^ —3702 



Est. height 



African Elephant Max. height 

-3568 



Indian flephant 
Max. height 

-3200 



300O 



'■3007 



-2000 V/ 



o 
\ 2 



1000 



.3/94- 




INDICUS 


L. AFRICANA 


A. IMPERATOR 




"s7u 771 2rO " 


A.M. lossa 


B'-^oya" 


10-5^4" 


n-5" 


D 


C 


B 



A. MAIBENI 



A'eZr.A7u.S. S-9-22 

12-6%" 

A Al 

I'iK- 012. Shoulder Heights of Living and Extinct Elephants. Compare Figure 1 191 

[Profo.^sor O.sborn'.s mothofl, iiscfl in the present Memoir, of e.itimatinK the heiRlit in the flesli i.s to add six and a third per cent, to tlie .skeletal height.— 

E<iitor.| 

Flesh 

3200 mm., 10 ft. fi in. e.st.' 
;539(; mm., 11 ft. l\ in. est. 

3702 mm., 12 ft. \\ in. 



Elephas inilicus: 

Standard .skpleta! height of full-grown male 

iMXodonia africana oxijotis ("Jumbo"): 

Adult male, extended .skeletal height at shoulder 

Arehuliskoflon imperator (Amer. Mu.s. 10.')98); sex unknown: 
Exteiid.'d skeletal height 
Estimated height in flesli 

Archidiskodon imperator maiheni (Neb. Mas. 5 -9 22): 

(Riglit) Perspective drawing after mount in the Nebraska Museum. 
(liCft) Right forelimb, also bony projection of each segment. 

Skeletal height: 

Fully extended forelimb 
Estimated, at shoulder, in flesh 



Skeleton 
3007 mm., 9 ft. 10^^ in. 

3194 mm., 10 ft. o'l in. 

3482 mm., II ft. .'> in. 



382f. mm., 12 ft. li^,; in. 



40t)8 mm., 13 ft. 4), in. 



■(Rowland Ward's "Records of Big Game," 1922, p. 408. A subsequent record (Ward, edition of 1928, p. 451) gives a height of 10 ft. 8 in. F:dit(ir.| 
-[The record skeletal height of the Afri<-an elephant (Lnxodonln afriainn) is 3290 mm. =10 ft. Ojj in. ("set up forelimb in Quex Museum," Kiichington, 

England— letter of January li, 1931, from Major 1'. H. (i. Pdwell-Cotton to Mr. J. W. Walker of Michigan) or 3.V20 mm. = 1 1 ft. f.K in., in the flesh, as 

measured on the sjime specimen in the field by Major Powell-Cotton {op. cil.. Ward, 1928, p. 4.')t)). Editor.] 



1022 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1023 



The principal measurements of the type skull and skelet 


ton of 


Archidiskodon imperator maibeni (see Barbour, 


1925.1, pp. 


114- 


117) are a.s follows: 






Skull 


Mm. I 


nches 


Occipital condyle, greatest diameter 


125 


5 


Molar (r.M^), length of grinding surface 


245 


^A 


Molar (r.M^), breadth 


110 


4K 


Breadth-length index 


45 




Molar (r.M'), greatest thickness of cement 


13 


'A 


Mandible 






Molar (M3), length of grinding surface 


240 


9K 


Molar (M3), greatest breadth 


95 


2,% 


Breadth-length index 


40 




Transverse ridges 14 






Scapula 






Total height 


1,257' 


49'i 


Extreme width suprascapular border 


889 


35 


Length of glenoid border 


556 


22 


Humerus 






Total length 


1,226' 


48)'4 


Transverse diameter of head 


330 


13 


Ulna 






Total length 


1,080' 


42K 


Fore Llmb 






Total height 


3,505' 11 ft. 6 in.' 


In the flesh this huge creature must 






have stood not less than 13 feet high 






at the shoulders [13 ft. 4'i in. est.] 






Tusk 






Diameter 8 inches from tip 


165 


%){ 


Diameter at incisive sheath 


267 


lOM 


Atlas 






Total width 


440 


17K, 


Length 


140 


5M 


Axis 






Height 


336 


13?^ 


Cervical 5 






Posterior centrum, vertical diameter 


186 


7'4 


Posterior centrum, transverse diameter 


192 


7)i 


Cervical 6 






Centrum, height 


175 


6'^ 


Centrum, length measured at center 


39 


VA 


Cervical 7 






Posterior centrum, vertical diameter 


185 


7K 


Posterior centrum, transverse diameter 


192 


lA 


Dorsal 9 






Centrum, vertical diameter, anterior 


167 


QYi 


Centrum, transverse diameter, anterior 


144 


5% 


Dorsal 10 






Posterior centrum, vertical diameter 


175 


6"s 


Posterior centrum, transverse diameter 


156 


^i 


Lumbar 






Centrum, length 


80 


^% 


Centrum, height 


131 


f>% 


Average length of 4 lumbars 


83 


3K 



Sacrum 
Length 

Greatest breadth 
Centrum, anterior, transverse diameter 

Femur 
Circumference at slimmest point, about middle 
Transverse diameter at narrowest point of shaft 
Anteroposterior diameter at narrowest point of 
shaft 



325 


12)i 


296 


l\% 


205 


8 


432 


17 


157 


6K 



115 



4K 



Archidiskodon hayi Barbour, 1915 
Figures S93.\, 903, 913, 91.^ 
Crete, Saline County, Nebra-ska. Lower Pleistoeene. 

This appears to be the most primitive species of proboscidean 
thus far discovered in America, distinguished by its shallow jaw 
and elongated rostrum. It obviously belongs to the broad-plated 
Archidiskodon phylum, but with 10-11 ridge-plates like A. plani- 
frotjs. The relatively long, shallow jaw, the prominent rostrum, 
the low coronoid proce,ss are totally different from the adult 
Archidiskodon imperalofr, also from the adult A. meridionalis, and 
there is certainly a very strong resemblance to the jaws of ^. plani- 
frons as figured by Mayet and Roman (Figs. 894 above and 914). 
A juvenile jaw of A. imperator would have this shallow character 
but Professor Barbour, who discovered this important specimen, 
remarks: (Letter of .January 17, 1923) : ". . . we both examined it 
[the type] as well as we know how, and count it an old not a young 
individual. I feel quite sure of this. [E.] hayi is very distinctive." 

Elephas hayi Barbour, 1915. "A New Nebraska Mammoth, 
Elephas hayi." Amer. Journ. Sci., (4), XL, No. 236, pp. 129- 
134 (1915.2). Type.— Mandible and teeth. Associated with 

this jaw were fragments of a large tusk. (Barbour, op. cit., 1915, p. 
132) : "The distinguishing character on which this new mammoth 
must depend is derived, first of all, from the teeth. Especial care 
was exercised to determine whether the teeth in the jaw of E. hayi 
are penultimate or ultimate molars. If penultimate, a successsor 
should be in e\idence in each ramus, but not a fragment of a tooth 
or plate could be found in the cavities, which were filled with com- 
pact sand and gravel ; nor could any such fragments be found in the 
surrounding gravels when screened. Undoubtedly the two teeth 
are the sixth molars, a point of consequence in this connection." 
Collections of Hon. Charles H. Morrill, the Nebraska State Mu- 
seum (Neb. Mus. 23-6-14). Type Locality and Horizon. — 
(Op. cit., p. 129): "Hurlbert sand pit at Crete, Nebraska, eight 
blocks east and three blocks north of the center of the town. . . . 
Aftonian. ... 11 feet below the surface." Type Figure. — 
Barbour, op. cit., 1915, p. 130, fig. 1, p. 133, fig. 3, p. 134, fig. 5d. 

Type Description. — (Barbour, op. cit., p. 129): "The chief 
distinguishing characters of Elephas hayi are: unusual length of 
mandible; the last molar small, narrow, and anterior to the coro- 
noid; transverse ridges 10 to 11; angle distinct and sharp pos- 
teriorly; coronoids uncommonly prominent, deeply pitted, and set 
very obliquely. . . . The mandible . . . measures 29,'2 inches (750°"") 
from the tip of the symphysis to the angle, .... The dejith of the 
jaw at the coronoids is 9/2 inches (241 "^'"), .... The coronoid 

'Letter of E. H. Barbour, May 8, 1929, give.? corrected measurements of Barbour, Colbert, and Shanafelt, namely, scapula 1066+ mm. =42 + in.; 
humerus 12."j1 mm. =49/8 in.; "'na 108.') mm. =42% in.; forolimb 3826 mm. =12 ft. 6^^ in.; length of grinding.surfaeeof left lower molar 238 mm. 



1024 



OSBORN: THE PROBOSCIDEA 



process is conspicuously robust, boing 2?^ inches (70""") through 
near its base, and an inch (25"'"') near the summit. It stands 4 
inches (102""") above the superior mandibular border, and 2 
inches (51 """) above the crown of the teeth. It is set more 
obliquely than in other mammoths. Its inner surface is deeply 
pitted, and extends from the outer to the inner alveolar border. . . . 
The teeth are those of a mature individual, with the crowns well 




Barbour's Type .Iaw of Archidiskodon hayi 
Fig. 913. Ty|)e of Ekphas haiji Barbour, 1915, p. l.W, fig. 1 (Nob. Mus. 

23-G-14), one-eighth natural .size. Ramus perfect to summit, condyle only 

restorpd. 

Compare jaw of Archidixkoilon jiltinifrona of Chagiiy, France (I'ig. 914). 
Profe.s.sor Barbour (letter of .January 17, 1923) remarks: ". . . we both 

examined it [the type] as well as we know how, and count it an old not a young 

individual. I feel quite .sure of this. [E.\ haiji is very distinctive." 

worn. Though well cemented and strong, the teeth of E. hayi are 
noticeably small. The postero-anterior diameter is but 9 inche.s 
(229 "'"'), and the greatest transverse diameter 3 inches (76 "'"'). 
. . . The dimensions of these teeth agree more closely with those 
of our earlier Nebraska mastodons than with those of our mam- 
moths. The number of transverse plates is noticeably reduced, for 



there are but 10 in one tooth, and 11 in the other, with no plates 
missing. . . . In E. hayi, there are 4 and a fraction transverse 
enamel ridges to the decimeter. The valleys are deep and bordered 
by highly crenulated enamel ridges. The great anterior prong 
branches widely and carries 3 plates. The teeth lack the sym- 
metrical development common to mammoths. They are notice- 
ably constricted back of the anterior prong, and taper posteriorly 
to 1)2 inches (38"'™) . . . there are but 11 transverse ridges at most 
[to each molar tooth], the last being small, perhaps a heel. This 
form seems to be an earlier and more primitive type of mammoth 
than any other known to the State [Nebraska]. The inferior dental 
foramen is small, and has a circular border, while in E. imperator 
it is very large and deeply notched, as shown in the accompanying 



/^// to ya /Vat. sije 



J From 5iwaZi7<5 \ 




From 5 en 676 



ARCHIDISKODON PLANIFRONS 



Fig. 914. Primitive mandibles of Ehphas [Archidiskodon] planifrons of 
the Siwaliks, India, Chagny-Bellecroix and Seneze of France, after Mayet 
and Roman, 1923, p. 81, fig. 13, in.serted for compari.son with type figure of 
E. [Archidiskodon] hayi (Fig. 913 opposite). See caption to figure 849, p. 
962 above. One-eighth natural size. Ob.serve .similarity to th<! Chagny 
mandible. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1025 



figures. Although inferior dental foramina differ in individuals, 
and even between opposite sides of the jaw, the differences shown 
by the cuts are significant. The ascending rami of our probo- 
scideans also vary between wide limits." 

Specific Characters. — (1) Mandibular ramus elongate, de- 
pressed, coronoid process low. (2) Third inferior molars broad in 
center, with ten to eleven ridge-plates set very far apart; laminar 
frequency i/i in 10 cm. ; molars narrowing posteriorly. (3) Ramus 
of jaw shallow, moderately expanded; rostrum relatively promi- 
nent. (4) Agreeing with Archidiskodon planijrons in ridge formula, 
differing in greater width of M3. (5) Dimensions of M3, length 229 
mm., maximum breadth 76 mm., index 33. 

Comparison with Archidiskodon planifrons (Osborn, 
1924). — It is remarkable that the type of Archidiskodon hayi re- 
sembles closely in the profile of the jaw the Archidiskodon plani- 
frons of India and of southern France, apparently justifying Bar- 
bour's statement that A. hayi seems to be an earlier and more 
l)rimiti\e type of mammoth than any other known to the state 
[Nebraska], and suggesting the possibility that we have to do with 




Archidiskodon imperator scotti Barbour, 1925 
[ =A. imperator ?juvenile] 
Figures 916, 917, PI. xxi 
Five miles south of Staplehurst, Seward County, Nebraska. Lower 
Pleistocene, Aftonian gravels. 

Awaiting further evidence, Osborn is inclined to regard the 
type of 'Elephas scolti' as representing a young individual of 
A rchidiskodon imperator. 

This Archidiskodont, discovered in 1922 in the Aftonian 
gravels, 5 miles south of Staplehurst, Seward County, Nebraska, 
and 20 miles south of Crete, Saline County, where the type of 




Fig. 915. Barbour's Type of Archidiskodon hayi compared with A. imperator Ree. 



(Left lower) Type right mandible of Archidiskodon hayi, sectioned at three points, grinding teeth with lO-U ridge-plates. 
(Left upper) Archidiskodon imperator ref. Same aspect of right mandibular ramus, grinding teeth with 18 ridge-plates. 
(Right upper and lower) Archidiskodon imperator with 18 ridge-plates, lateral and superior aspects of mandible. 
Both specimens in the Nebraska State Museum, Morrill Collection. After Barbour, 1915.2, figs. 2, 3, and 4. 



the arrival in North America in the late Pliocene or early Pleisto- 
cene of a primitive proboscidean, displaying some of the chief A. 
planifrons characters, namely: (1) Jaw long and shallow; (2) 
rostrum prominent; (3) coronoid relatively depressed; (4) grind- 
ing teeth broad with 10-11 transverse ridge-plates. Supplementary 
figures and sketches of the jaw and teeth of A. hayi (jaw. Fig. 
893 A) serve to emphasize the wide contrast between this type 
jaw and the adult jaw of ^. imperator (Fig. 892 B, A). 



Elephas [ = Archidiskodon] hayi was found (1914), is regarded by 
Barbour (1925.1, p. 22) as a mature individual, as primitive as 
A. hayi, if not more so. This statement rests upon the identifi- 
cation of the small 8 ridge-plated molars as representing third 
inferior molars, M3, which appears to Osborn very doubtful, 
especially as the jaws of the type of 'Elephas scotti' (Fig. 916) are 
very robust in section with greatly abbreviated symphysis, alto- 
gether different from the relatively slender, elongated mandibular 



1026 



OSBORN: THE PKOBOSCIDEA 



rami of A. hayi. Consequently it seems probable that they are 
second inferior molars, M2. 

Elephas srolli Barbour, 1925. "Elephas scotti, A New- 
Primitive Mammoth from Nebraska." The Nebraska State 
Museum, Bull. 2, Vol. I, April, 1925, pp. 21-24 (Barbour, 1925.1). 
Type. — Mandible with last [second?] lower molar of each side 
(Neb. Mus. 18-2-22, the Maiben Palaeontological Collections. 
Cast of type right M3 [r.:M2], Amer. Mus. 14610). Horizon 

.\ND Locality. — Discovered on farm of Mr. E. J. Hartman, five 
miles south of Staplehurst, Seward County, Nebraska. Type 




ric- I'I'i. .Juvenile jaw of Archiilinkndnn impiralur ref. (Neb. ^I^l.>^. 
IS 2-22), friini Sew;ir(l Coiinly, Ni'bniskji. After plKitonnnih by Prof. 
E. II. Hiirbiiiir wlm in l!(2.') niiuie lliis tlie type of Eliplmx xmlli. ('ompMre 
figure !(I7. Seiile about one-sixth natural size. 

Tlio.se 8 ridge-platfd griiulinK teetli, deseribed as third ii)f<'rior molars, 
M3, are regardi^d by Osborn as seeond inferior molars, Mj (ettst of r.M'j, .\nier. 
Mus. 14610). 



FiGUKE. — Op. cii., text figures 7-10. 

Type Description.— (Barbour, 1925.1, pj). 21-24): "On 
February 18, 1922, the mandible of an unusually primitive mam- 
moth was secured for the palaeontological collections of IMr. Hector 
Maiben by Mr. E. T. Engle. . . . The peculiarities of this mam- 
moth seem to entitle it to a position as a di&tinct species, for which 
we are proposing the name, Elephas scotti, named for Profes.sor 
William D. [B.] Scott. It cannot be compared with the later and 
more advanced mammoths, such as imperator, jeffersoni (columbi), 
or jjrimigenius. It is comparable instead, with the earlier and more 
conservative mammoth, Elephas hayi. . . . The new mammoth is 
as primitive as Elephas hayi, if not more so. Like hayi, it is un- 
doubtedly a mature individual. Its teeth are taken to be last 
[second] molars. The enamel plates, which are highly crenulatcd, 
incline noticeably backward, and are worn with extreme obliciuity. 
At the same time the valleys, or dental spaces, are so deeply in- 
dented, as to still further heighten and exaggerate the effect. There 
are but five pronounced ridges, and in all, but eight and a cone. 
Two of the anterior ridges are so confluent that the count is ren- 
dered somewhat uncertain, as shown in figure 10 [our figure 917A, 
Al, A2]. The transverse ridges in E. hayi are eleven. The molars 
of scotti measure 219 mm. (8\ in.) in length, by 117 mm. (4!'8 in.) 
in extreme width. They are short and abrujjtly expanded in the 
middle. In the mammoths the number of enamel ridges to the 
decimeter serves, in a general way, in the recognition of s])ecies. 
In E. scotti there are three and a fraction, transverse, grinding 
ridges to the decimeter; in E. hayi four and a fraction; in E. 
imperator five to six; in E. jeffersoni (columbi) six to eight; and 
in E. primigenius nine to ten. It is a noteworthy feature that the 
robust jaws of Elephas scotti come within three-fourths of an inch 
of meeting on the middle line, as is plainly shown in the figures 
[see Figs. 916 and 917A of the present Memoir]. This is not due to 
crushing, as far as can be learned, for the specimen in hand is 
essentially perfect. The coronoids of the earlier and the later 
mammoths differ widely and are worthy of notice. Those of scotti 
and hayi are much more robust, thick, and heavy, and flare out- 
wardly, and are posterior to the molars. The inner wall is broader 
and more heax'ily roughened and pitted for ligamentous attach- 
ment. Each ramus, measured back of the molar, has a width of 
185 mm., (7X6 in.) and a depth of 180 mm., (7's in.). On the middle 
line the jaws are but 19 mm., (:'.| in.) apart." 

Osborn, 1928: The Seward (^ounty jaw (Neb. Mus. 18-2 22 
— see Fig. 916, and as seen from abo\c. Fig. 917.\) in Osborn's 
opinion resembles a juvenile jaw of .4. impcrnlor rather than the 
type jaw of E. hayi Barbour. In superior view (Fig. 916) the jaw 
appears fairly robust, less swollen than that of A. iinpcralor, with 
highly characteristic outwardly flaring coionoid i)r(>cesscs. In 
lateral view (Fig. 916) it is relatively short and deej), the rostrum 
is short and dejiressed, in wide contrast to the long, shallow jaw of 
the tyjje of A. hai/i. The single grinding toolii, in which 8 9 ridge- 
plates appear, may represent M2 of a young A. imperator; the 
tooth is obli(|uoly worn and conseciuenfly the dental space between 
the broad enamel ridges a|)i)e;ns to be much greater than it actually 
is; the disparity in the actual distance between th(^ ridge-plates 
(Fig. 917 A2) and the a|)pai-eiit distance due to obliriuity of wear 
(Fig. 917 Al) are clearly shown in this diagrammatic represen- 
tation. 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1027 



Archidiskodon imperator maibeni Barbour, 1925 

(Sec full (l('SiTij)tion above, page 1019, of the skeleton of .1. maibeni) 
Figures 815, 824, 910-912, 917, 918, 1239, PI. xxi 

Tiineolii County, about sixteen miles north of Curtis, Nebraska. 
Aftoniaii,' loes.s 100 feet in general tliiekness. Upper Pleistocene (see 
I'ig. 1239, also PI. viii, Vol. I). 

The type of this species is an unusually complete skeleton, 
entitled the 'Lincoln County Mammoth,' in the Nebraska State 
Museum, described by Barbour in 1925 (1925.3) as the "Columbian 
Mammoth Elephas maibeni, sp. nov." and subsequently (June, 
1926) transferred by him to Archidiskodon maibeni, now mounted 
in the new Museum, Morrill Hall, University of Nebraska. To 



thoracics, lumbars, and the sacrum, ribs, and double ribs, both fore 
limbs and parts of the hind limbs. . . . The hind quarters are 
represented by parts of the pelvis, the shaft of a femur, and the 
major portion of a fibula. The dentition is perfect." Neb. Mus. 
5-9-22. Horizon and Locality. — Discovered by Mr. and 

Mrs. H. S. Karriger about sixteen miles north of Curtis, Lincoln 
County, Nebraska, on the Karriger farm. Type Figure. — 

Op. ciL, Figs. 58-60, 63-70, 72, 74, 76-87. 

Type Description.— (Barbour, 1925.3, pp. 97-111): "Con- 
fusion has long surrounded the Columbian and jeffersonian 
mammoths. But the one under consideration is undoubtedly of 
the true Columbian type. The bones of Elephas maibeni were 




Neb. Mus. 18-2-22 





I"ig. 917. Comparison of : (B) Mandible of the t yp<' of Elephas (Archidiskodo7i) maibeni [ = A rrhidiskodon iiiiperalor maibeni] (Neb. Mus. 5-9-22), with 
third inferior molars, M3, in plaee, exhibiting 1.')+ ridge-plates, 14 exposed; and (.\) type jaw of Elephas scolli [= Archidiskodon imperator scolti or ?juvenile 
A. imperator] (Neb. Mu.s. 18-2-22), containing second inferior molars, M2, exhibiting 8 ridge-plates, and measuring ap. 219 mm., tr. 117 mm., i.e., short and 
abru])tly expanded in the middle; \l, A2, enlarged views of right second inferior molar of the type of 'E. scotii' (cast Amer. Mus. 14610). 



the above full description of the skeleton (p. 1019) by Barbour 
(1925.3) may now be added the type description and characters of 
the skull and dentition (Barbour, 1926.1), as follows: 

Archidiskodon maibeni Barbour, 1926. Professor Barbour in 
his supplementary description ("Archidiskodon maibeni," Nebras- 
ka State Museum, Bull. 11, Vol. I, June, 1926, pp. 119-122) states: 
"Archidiskodon maibeni was first described in Bulletin 10 of the 
Nebraska State Museum under the title 'Skeletal Parts of the 
Columbian Mammoth, Elephas Maibeni, sp. nov.'" Under this 
designation I^arbour gives additional measurements and comments 
on the skeleton. He also mentions the discovery of another skull 
(Neb. Mus. 1-4-26) referable to A. maibeni. 

Elephas maibeni Barbour, 1925. "Skeletal Parts of the 
Columbian Mammoth, Elephas Maibeni, sp. nov.," Nebraska 
State Museum, Bull. 10, Vol. I, August, pp. 95-1 18. Type.— 

(Op. cit., p. 98): "The skeletal jjarts ])reserved are the skull, 
mandible, one tusk, the atlas, axis, and four other cervicals, several 



found projecting from a loessial wall at the bottom of a small 
canyon. The general thickness of the loess at this point is about 
100 feet. . . . The outstanding characters of Elephas maibeni are 
size, extreme curvature and divergence of tusks and incisive 
sheaths, unusual shortness of centra coupled with great width. 
The tusk must have lain in a plane or nearly so and must have 
described a circle, the radius being 28 inches (711 mm.) The 
diameter of the tusk is 6)2 inches (165 mm.) at the tip, 7'; inches 
(190 mm.) four feet back of the tip, and 10).^ inches (297 [267] mm.) 
at the incisive sheath. Originally it was a magnificent piece of 
i^'ory. . . . The humerus is huge beyond the visualization of those 
who have not seen it, hence must judge of it from figures and 
measurements. In the hind quarters the bones are, if anything, 

less massive than might be expected Judging from the very 

short vertebrae the body must have been unduly foreshortened. 
. . . The skull is broken into several large, and numerous small 
pieces, which have not been set permanently in place. . . . The 



'[Lugn and Schultz (1934.1, p. 376, also Tabic A) regard it as of lowan, late Pleistocene age. — Editor.) 



1028 



OSBORN: THE PROBOSCIDEA 



accompanying free hand sketches were made with the pieces set 
approximately in position, and give our impressions of the skull. 
. . . The head must have presented a bull-dog effect or aspect. 
It might be called the bull-dog mammoth. . . . The mandible is 
well preserved, and in point of size passes the largest in our col- 
lections. The rami are widely divergent and the condyles far 




apart. . . . The anterior borders of the coronoids are excessively 
roughened; they are even nodular, and we find no parallel. . . . 
The lachrymal process, which is uncommonly large, prominent, 
and acorn-shaped, is dissimilar to all others available for study and 
comparison. . . . The fore limbs, four vertebrae, two pairs of single 
ribs, and one pair of double ribs of this exceptional mammoth are 
mounted in approximate position and make an impressive arch, 
the height of which is 13 feet from the tip of the toe to the top of 
the spine, see fig. 58 [Fig. 910 of the present Memoir]. In the flesh 
the height of Elephas maibeni at the shoulder must have been 
about 13 feet, and the top of the head of this magnificent beast 



Bl 




y. 




B 





BaKBOUr's TyI'E of AnCHIDIKKODON IMI'EUATOU MAIBENI 

Ji'lg. 918. .Irc/iidwfcodontmperatorjKaibem', tho 'Lincoln County Mammoth,' .superior and inferior dentition of (lie type skeleton (Neb. Mus. 5-9-22)- 
.\fter pliotograph.s kindly furnished the present author by Prof. E. H. Barbour (compare Barbour, 1925.3, figs. 64, 65, 67, 68). 

.\, MandibU' of .1. imprrator maibeni from above, about one-eiglith natural size. 

Right and left inferior molars, r.Ms, I.M3, with 15-|- ridge-plates, 14 exposed, more or less worn, broad < inieiit, somewhat .sinuous, but slightly 
concave posteriorly, crown externally plane, internally convex. 

Al, The same in lateral view. One-eighth natural size. 

B, Palate containing right and left third ,su|«Tior molars, r.M'', I.M''. One-eighth natural size. 
Bl, Right third suijcrior molar, r.M'. One-fourth natural size. 

C, Extremity of tusk 4 feet, in lengtli measured mi outer curve. About (ine-tcnth natural size. 

Obs<'rve in B, Bl, 16+ ridge-plates, of which I 13 show .signs of wear, (|uite .strongly concave posteriorly with heavy border of cement; crown externally 
convex, internally slightly (concave. Very similar in contour and ridge formula to A. imperalor (.\mer. Mus. 14470 — Fig. SSOB), from Victoria, Texas. 
Observe in A, Al, 15-f- ridge-plates, 14 ex|«jsed. 



THE MAMMONTIN^: ARCHIDISKODON AND METAIICHIDISKODON 



1029 



must have been abuut 14 feet above the ground. From the tip of 
the toes to the top of the scapula is 1 1 feet, 6 inches. This specimen 
is believed to hold the record for size amongst the Columbian 
group of mammoths. In point of size, Elephas maibeni was a rival 
of the Imperial elephant itself, which stood l.S}^ feet high. The 
tallest living African elephant stands 11 feet high and the average 
elephant of the menagerie and circus 8 to 9 feet." 




Type of Archidiskodon hakoldcooki 
Fig. 919. Typo mandibk' and tliird inferior molar of tlii' right .side 
in silu of Elephas haroldcooki Hay, 1928. After Hay and Cook, 1930, PI. iii, 
fig. 1. One-fifth natural .size. 



Archidiskodon haroldcooki Iluy, 1928 

Figu"' 919 

Found in Holloman's gravel quarry, Frederick, Oklahoma. Aftonian? 
gravels. 

Elephas haroldcooki Hay, 1928. "Preliminary Descriptions of 
Fossil Mammals Recently Discovered in Oklahoma, Texas and 
New Mexico." Proc. Colo. Mus. Nat. Hist., Vol. VIII, No. 2, 
Pt. 1, February 2, 1928, p. 33. Type.— Nearly complete 

lower jaw containing last right and left molars in situ (Colo. Mus. 
1057). HouizoN AND Locality. — Holloman's gravel quarry, 



at Frederick, Oklahoma. Type Figure. — Hay and Cook, 

1930.1, Pis. Ill, fig. 1, v, fig. 1, XIII and xiv. 

Type Description.— (Op. cit., 1928, p. 33): "Elcphan harold- 
cooki Hay. Based on a nearly complete lower jaw containing the 
right and left last molars. Anterior fang and 2 ridge-plates absent 
through wear. 12 ridge-plates and the rear talon present. 4.4 
ridge-plates in 100 mm. Enamel thick, moderately folded. A 
loxodont expansion present in some of the ridge-plates. The crown 
very high. Found in Holloman's gravel quarry at Frederick, 
Oklahoma. No. 1057, Colorado Museum of Natural History, 
Denver." 

Supplementary Description (Hay and Cook, 1930.1, pp. 
32, 33). — "This jaw was embedded in the cemented gravel very 
close to the Permian red clay ... It will be seen that nearly all of 
the jaw behind the teeth is lacking. These teeth are in fine con- 
dition and show that the animal was somewhat beyond middle age. 
The height of the jaw where the ascending ramus arises is a little 
more than 200 mm. Its thickness, where greatest, at the middle of 
the tooth is 155 mm.; the width of the bone, taken at the rise of 
the ascending rami, is about 390 mm." 

"As will be observed from the views of the jaw, it is wholly 
without a beak in front; and does not turn downward, also the 
symphysis is short, about 60 mm. in length. The teeth present are 
])arts of the hindmost molars. This is shown by the arrangement 
of the rear ridge-plates, only partly shown in the figures. In front 
of each tooth (plate iii, figure 1 [=Fig. 919 of present Memoir]), 
especially of the one of the right side, is seen a cavity in which was 
lodged the anterior fang. The tooth had been worn down to the 
bar of bone separating this fang from the part behind it and the 
fang had fallen out. With it went three worn-out ridge-plates. 
Behind the bar of bone may be counted on the grinding face 11 
ridge-plates, the tenth being represented by a small circle of enamel, 
the eleventh by a dot of enamel. The length of the abraded surface 
of the molar (plate v) is 182 mm. ; its width, with the cement, is 
85 mm. ; without it, 80 mm." 

"Behind this eleventh ridge-plate is a mass which may be 
regarded as a talon, a ridge-plate as yet undeveloped. Adding 
now to these eleven ridge-plates the two supported by the anterior 
fang and the one by the bar of bone behind it, we have 14 ridge- 
plates for the hindmost molar of the species. From plate xiv it is 
seen that the ridge-plates are very thick. Measured at one-half 
their height three are spanned by a line just a little less than 75 mm. 
Measured at one-half their height 4.44 plates are crossed by a 100 
mm. line. The hind end of the crown is almost 6 inches high, 
nearly as high as the grinding surface was long." 

"The view (plate xiv) of the inner face of the tooth shows 
that the ridge-plates are at first directed strongly forward, then 
arc turned abruptly upward and (in relation to the grinding face) 
somewhat backward. It is due to this oblic|uity that the front 
enamel plate of the hinder ridge-plates is more exposed to view 
than they are farther forward. It will be seen that the outer ends 
of the loops of enamel are rather strongly turned forward. The 
enamel, as usual in primitive elephants, is thick, here about 
3 mm. It is moderately crimped. A feature of interest is the 
presence of a lozenge-shaped expansion at the middle of each 
ridge-plate. This is a characteristic of the earlier elephants, as 
E. planifrons, E. meridionalis, and E. imperator." 




ARCHIDISKODON /MPERATOR /?e/" 

DW.,.,.- I.vs.,.,. ,S...OKS OP AkCH.„,SKODON- ..OM,.U,0„ „,..„ A. ,M,.K,UT„K 

1030 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1031 



Archidiskodon exilis Stock and Furlong, 1928 
Figures 815, 920, 921, PI. xxi 

Santa Rosa Island, California. Plcistocrno. 

This dwarfed insular form of imperial mammoth, estimated 
at 6-8 feet in height measured at the shoulder, as compared with 
12 ft. 1% in., distinctive height of Airhidiskorlon imperator. is of the 
greatest interest. There is little doubt from the drawings recei\ed 
(Fig. 920) that this is a diminutive insular form related either to 
A. imperator or to Parelephas coliimhi. 

Elephas exilis Stock and Furlong, 1928. "The Pleistocene 
l^lci)hants of Santa Rosa Island, California," Science, Vol. LXVIII, 
No. 1754, p. 140. Type. — "A skull and mandible including 

four cheek-teeth and two tusks" (Calif. Inst. Tech. Coll. Vert. 
Pal. 14). Horizon and Locality. — Santa Rosa Island, 

California, the second largest of four islands separated from the 
mainland by the Santa Barbara Channel. Pleistocene. Type 



recently summarized the available information and recognizes the 
presence of Elephas imperator and of an undetermined species. 
During the past year Dr. Spencer Atkinson and Mr. J. A. Barbieri, 
of Pasadena, secured a fragmentary elephant skull on Santa Rosa 
Island and presented the specimen to the California Institute of 
Technology. Through the courtesy of the ^'ail Company of Los 
Angeles, owners of the island, the California Institute, with the 
cooperation of the Carnegie Institution of Washington, have been 
given the oi)]iortunity to investigate the occurrence, and facilities 
were kindly made a\ailable to collect further remains. . . . The 
collections secured by the California Institute include a number of 
teeth, parts of skulls and skeletal material. Occasionally several 
skeletal elements and teeth are found associated in the deposits. 
Usually the remains are scattered. One curious feature of the 
occurrence is the ai)parent total absence of associated mammalian 
types. The proboscidean remains are referable to the genus 




Restoration of Archidiskodon exilis of Santa Rosa Island 
Fig. 921. Fai-ial [lortion of tlio skull, with tusks and lower jaw, of Elephas exilis Stock and Furlong, 1928. From 
Quaternary deposits, Santa Rosa Island, California. After unpublished photograph kindly furnished by Dr. Cliester 
Stock. 



Figure.— Stock, 1935.1, p. 210, fig. 6. Supplementary De- 

scription. — Stock, "Exiled Elephants of the Channel Islands, 
California," Scientific Monthly, 1935, XLI, September, pp. 205- 
214, text figs. 1-10. 

Type Description. — (Stock and Furlong, 1928.1, pp. 140, 
141): "W. G. Blunt's discovery of fossil teeth of an elephant on 
Santa Rosa Island, one of the Channel Islands off the coast of 
southern California, was recorded by Stearns [Footnote: 'Stearns, 
R. E. C., Proc. Calif. Acad. Sci., Vol. 5, p. 152, 1873.'] in 1873. 
Since that time this interesting and significant occurrence has been 
referred to by several authors. Hay [Footnote: 'Hay, O. P., 
Carnegie Inst. Wash. Pub. 322B, pp. 42, 43 and 51, 1927.'] has 



Elephas. The individuals exhibit considerable variation in size, 
and this is undoubtedly to be a.scribed in part to differences in age. 
A survey of the collection as a whole yields the impression rather 
strongly that the elephant types were of relati\-ely small size. 
Some of the forms may have a height of six to eight feet as measur- 
ed at the shoulder. The larger individuals are perhaps comparable 
in size to the American mastodon and are certainly smaller, 
possibly considerably smaller, than the Pleistocene mammoths of 
the southwestern United States. While tlie Santa Rosa Island 
elephant has been determined as representing the species Elephas 
primigcniiis Blumenbach and E. (Archidiskodon) imperator Leidy, 
the difference in size, coui)led with differences noted in the skull and 



1032 



OSBORN: THE PROBOSCIDEA 



dentition, seem quite clearly to distinguish the island form as 
a distinct species for which the name Elephas exilis is here pro- 
posed." 

Supplementary Description (Stock, 1935.1, pp. 206, 207, 
212-214). — "Remains of extinct elephants are now known to 
occur on three of the Channel Islands, namely, on San Miguel, 
Santa Rosa and Santa Cruz (see Fig. 1 [ = Fig. 922 of present 
Memoir]). The first material was found on Santa Rosa more than 
sixty years ago, and this island has furnished by far the largest 
collection of fossil specimens representing the.se types. Similar 
material has been brought to light on San Miguel. In contrast to 
the rather numerous finds of elephant remains in Quaternary de- 
l)osits of Santa Rosa, the presence of elephants on Santa Cruz is 
known thus far by only two fragmentary enamel plates of a cheek- 
tooth." 

"San Miguel: Although this island is wind-swept and shifting 
sand dunes mantle much of the area underlain by sediments of 
Tertiary and Quaternary age, the incision of the present land sur- 
face by ravines and gullies and the constant though gradual re- 



Scale 



Submarine depths -n fathoms 

nta Barbara 

Carpinterja 



Ventura 




Fig. 922. "Map of coastal province of southern California in vicinity of .Santa Barbara. 
I><)iation of some occurrences of fossil elephants on Channel Islands shown by x. Dotted 
line indicates hypothetical border of land during Pleistocene time, after Chancy and Mason." 
Reproduced from Stock, 193.J.1, p. 206, fig. 1. 



cession of the ,sea-cliffs develop exposures on which occasionally the 
weathered-out materials of fossil mammals have been discovered. 
Several tusks and cheek-teeth of elephants were found in a thin 
series of (Quaternary alluvial deposits lying beneath a table-like 
surface and exposed in the sides of gullies near the northwest end 
of San Miguel. Scattered proboscidean teeth have been found 
from time to time elsewhere on this island. Among tlie fossil 
materials are specimens which clearly point to the fact that the 
San Miguel elephants are among the largest types to be obtained 
in the insular region." 



Conclusions (pp. 212-214). — "Whether or not more than one 
species of elephant is present among the island forms remains to be 
definitely determined. In this connection, it should be recognized 
that an interesting and perhaps significant difference may exist 
between tho.sc forms on Santa Rosa and the types of San Miguel." 
"Numerous cheek-teeth and tusks, fragmentary jaws and 
skeletal elements comprise the bulk of the collections obtained on 
Santa Rosa. Individuals of all ages are preserved, from an unborn 
type to fully grown adults. The youngest specimen, evidently 
belonging to a foetus, is represented by a lower jaw (Fig. 6 [not 
figured in present Memoir]) in which the enamel plates had not firm- 
ly consolidated to form the lower cheek-teeth and had not erupted 
through the gums. One fairly complete skull repre.sents an adult in- 
dividual and furnishes valuable information as to the specific char- 
acters of the island elephants. Wlien found in Quaternary strata, 
exposed in the sea-cliff near the mouth of the Caiiada Corral, only 
the weathered cranial portion was visible. . . . Excavation revealed 
the rest of the skull and upper tusks . . . with the lower jaw in 
position below the palate. [This is the type.] Illustrations of this 
specimen and of a young adult skull of the im- 
perial mammoth (Airhidiskodon impemlor), drawn 
to the same scale, are shown in Figure 9 [not fig- 
ured in ])resent Memoir]." 

"Comparison of fossil remains of elephants 
found on Santa Rosa with comparable materials 
occurring on the mainland establishes clearly the 
fact that the island forms were smaller in stature 
than their relatives of the mainland. Consider- 
able variation in size exists among the island types, 
but the difference in stature between island and 
mainland forms remains a notable feature. . . . 
While the elephants of the mainland ranged in 
height from a])i)roximately IO/2 feet to 13/2 feet 
as measured at the shoulders, those of the islands 
presumably never exceeded 8 or 9 feet in height 
and the smaller individuals were probably no 
taller than 6 feet. Thus, the smaller size of these 
elephants presents a character wherein they re- 
semble the fossil or subfossil, dwarfed elephants 
described from the Maltese Islands of the Medi- 
terranean. The diminution in size, however, has 
not been carried so far in the Channel Island 
elephants as in the Maltese species." 

"As mentioned before, the elephants of San 
Miguel are among the largest types to be recorded 
from the island region. Tusks of these forms have 
been found which measure 5 feet in length and 
6 inches in diameter at the base. While some of the fossil materials 
on Santa Rosa likewise indicate the former presence of relatively 
large individuals, it is possible that the average size of the San 
Miguel elephants was larger than that of the Santa Rosa types. 
Were this ultimately established to be the case, on the basis of 
a comparison with more extensive collections than are now avail- 
able from San Miguel, it is interesting to speculate whether the 
difference may not have been the result of an earlier extinction of 
elephants on the smaller of the two islands." 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODON 



1033 



Archidiskodon sonoriensis Osborn, 1929 
Figure 923 
One mile east of Arizpe, northern Sonora, Mexico. Lower Pleistocene. 

Archidiskodnn sonoriensis Osborn, 1929. "New Eurasiatic and 
American Proboscideans," Amer. Mus. Novitates, No. 393, Dec. 
24, 1929, p. 18. Type.— "Nearly complete skeleton,' of 

which the palate with third superior molar, M', of both sides, 
right lower jaw (lacking ascending ramus), with third inferior 
molar, r.Ms, in situ, also .symphysis, are in the American Museum." 
Amer. Mus. 22637 (Osborn, 1929.797). Horizon and 

Locality. — "One mile east of Arizpe, northern Sonora, Mexico, on 
the Sonora River, 60 miles southeast of Cananea and approximate- 
ly 100 miles north of La Prietas and San Jose de Pimas. . . . The 



Archidiskodon meridionaHs nebrascensis Osborn, 1932 
Figures 815, 924, 927, 928, 1239, PI. xxi 

One mile northwest of Angus, Nuckolls County, Ncbrask.a. Lower to 
Middle Pleistocene. 

Archidiskodon meridionnlis nebrascensis Osborn, 1932. "The 
'Elephas meridionalis' Stage Arrives in America," Proc. Colo. 
Mus. Nat. Hist., XI, No. 1, Sept. 7, 1932, pp. 1-3 (Osborn, 
1932.893). Type. — Skeleton, lacking cranium also tusks, ex- 

cepting mid-portion of left tusk, with lower jaw in comjilete state 
of preservation. Colo. Mus. 1359. Horizon and Locality. — 

Found "one mile northwest of Angus, Nuckolls County, Nebras- 
ka, some fourteen or fifteen years ago." Lugn and Schultz (1934.1, 
Table A) regard this species as of Yarmouth (Upland) age, equiv- 




Fig. 923. Archidiskodon sonorietisis, anterior portion of type mandible and maxilla showing r.Ms, rM^ (Amer. Mus. 
22637), one-sixth natural .size. Compare O.sborn, 1929.797, p. 18, fig. 18. 



Arizpe horizon is regarded by Barnum Brown as Lower Pleistocene 
(lake deposit)." Type Figure. — Op. cii., p. 18, fig. 18. 

Specific Characters. — "Mandibular rostrum prolonged 
obliquely downwards, with downturned beak, as seen both in 
front and side views; length from symphysial groove to tip of 
rostrum 230 mm., exposed length of M' 246 mm., of Ms 346 mm.; 
depth from third unbroken plate to bottom of jaw 244 mm. A 
total of ll + 2(?) exposed ridge-plates in M=, of 2(?) + lH-3 in 
Mo." 



alent to Lower to Middle Pleistocene. Type Figure. — 

Osborn, 1932.893, figs. 1 and 2. 

Characters. — "The inferior grinding teeth similar in charac- 
ter and in ridge formula to the 'Elephas meridionalis' of Durfort, 
but somewhat broader with much thicker surrounding layer of 
cement. . . . Mandible: (1) A very prominent rostrum. (2) A rel- 
atively elongate and shallow ramus. (3) Measurements as follows : 

Length mandibular condyle to symphysis 943 mm. 

Depth below M3 of mandibular ramus 220 mm. 



'[Remainder of skeleton unintentionally destroyed by discoverer. — Editor.] 





Fig. 924. Type Mandible of Archidiskodon meridionalis nebrascensis (Colo. Mns. 1359). Both figures one-fifth natukal size 
(UpiM'r) Cro-.vn view of mandihlc with nglit and left third molurs in silu. After Osborn, 1932.893, fifj. 1. 
(Lower) Lateral view of mandible and ro.struin. -Vfter Osborn, 1932.893, fig. 2. 

1031 



THE MAMMONTIN.E: ARCHIDISKODON AND METARCHIDISKODON 



1035 



Drawn from cast Amer Mas 2!Q95 




yARCHIOISKOljON MERI0I0NALI5 Ref. of DurfcrT France 




Drawn from cast Amer Mu'^ 21394 



Fig. 925. Referred superior and inferior molar.s found as.sooiated witli 
Durfort skeleton (Archidiskodon meridionalis) in the Mu.seum d'Histoire 
Nafurelle, Paris, after casts kindly furnished the present author by Dr. 
Marcellin Boule in January, 1930. Both figures one-third natural size. 

(rp|)cr) Right second superior molar, r.M- (cast Amer. Mus. 2189')), 
with +8 worn ridge-plates; ^Yi ridge-plates in 10 cm.; coronal surfa(^e length 
162 mm., maximum breadth 81 mm. 

(Lower) Summit of crown of left third inferior molar, I.M3 (cast Amer. 
Mus. 21894), with HM ridge-plates, 6 partly worn; o anterior ridge-plates 
in 10 cm. ; maximum length 276 mm., maximum breadth 83 mm. 



I.pnfjth of r.M3, Third Inferior Mohir 289 mm. 

Width of r.M3, Third Inferior Molar 84 mm. 

R.M3 total ridge-erest.s + 13 + 

L.M3 total ridge-crests 'a 13 }i 

Ridge-crests in 10 cm b]i 

Widest ridge-crest 73 mm. 

The above ridge-crest formula, together w ith the dimensions of the 
thin! grinders, agree closely with those which ]irevail among most 
of the specimens referred to 'Elephas tneridionalis' in the British 
Museum as described and figured by Falconer." 

"Unfortunately, the cranium was the first part of this animal 
to be exposed and was completely weathered out. Only the e.\- 
tremity of one of the superior incisive tusks remained; this was 
lost; a mid-portion of the left incisive tusk remains. Fortunately, 
every othei' part of the skeleton was preserved in absolutely com- 
plete condition on one side of the animal or on both sides so that 
the skeleton is now superbly mounted and becomes a classic in 
all its dimensions as follows: 

Vertebral column: 

7 cervical vertebrae measuring 543 mm. 

19 dorsal vertebrae measuring 1640 mm. 

4 hunbar vertebrae measuring 400 mm. 

Sacrals not preserved. 
7 caudals only preserved. 
Height dorsal spine to ground (as 

mounted) 3695 mm. 




R.M- 



L.M 



ARCMIOlSKOOOh MERIDIONAUS ^e/ of J,urfo»^ rr^,^ 
Mrot^n ^rom cast AmerMuj 2/gs/ 



ARCHIDISKODON MERIDIONALIS ffcf of Hur/jrr Trance 
J)T-ah'n from cast Amer A/us 2/89/ 



L.M 



'/^ /Vafural si'73 





Fig. 926. Superior and inferior molars of Durfort skeleton {Archidiskodon mmdionalis) in the Museum d'Histoire Naturellc, Pari.s, after ca.sts kindly 
furnished the present author by Dr. Marcellin Boule in April, 1930. All figures one-third natural .size. Compare with figure 924, showing type mandible (if 
.4. meridionalis nebrascensis of Nebraska, with third inferior molars in situ. 

(Ix'ft) Left and right second and third superior molars (M-"'), after cast (Amer. Mus. 21891). 

(Right) Right second and third inferior molars (r.M2.3), same as opposite figure, and left .second and lliird inferior molars (l.M-.-a), after cast (Amer Mus 
21891). 



1036 



OSBORN: THE PROBOSCIDEA 



Fore and Hind Limbs: 

Height fore-limb scapula to ground 3454 mm. 

Scapula, length of 1020 mm. 

Humerus, length of 1220 mm. 

Ulna, length of 910 mm. 

Radius, length of 980 mm. 

jMedian metacari)us HI 200 mm. 



Pelvis, length of os innominatum 1350 mm. 

Pelvis, width of os innominatum 1754 mm. 

Femur, length of 1390 mm. 

Tibia, articular length of 840 mm. 

Pes: astragalus to tip of Mts. Ill 475 mm. 

Pes: depth of Mts. Ill 150 E." 




Fig. 927. Restoration of Archidiskodon meridionalis nehrascensi.i Osborn, 1932, from the complete skeleton and m.andible with lower portion of the tu.sk.s, 
lacking only the cranium. One-fiftieth natural size. 

The grinding teeth of thi.s |>riceless specimen are in the .same stage of evolution as those of the famous 'FAeplmn' niiridionalis of Durfort, France, a.s 
established by casts of the Durfort grinding teeth kindly sent to the American Museum by JJirector Marcellin Boule (see I'^igs. 92."), 92()). 



THE MAMMONTIN^: ARCHIDISKODON AND METARCHIDISKODOX 



1037 



TABLE XI. Measurements of Molars of Archidiskodon meridionalis of Durfort (cf. Figs. 925,926, also Fig. 924, 

Mandible and Third Molars of A. meridionalis nebrascensis) 



Cast Amer. Mus. 21891 



Associated molars 

Cast Amer.Mus. 21894 
Cast Amer. Mus. 21895 





Length 


Breadth 


L.M2 


152 


94 


R.M2 


157 


93 


L.M" 




85+ 


R.M3 




82+ 


R.Ma 


140 


86 


L.Ma 


143 


92 


R.M3 




90 


L.Ms 




85 


L.M3 


278 


83 


R.M2 


157 


81 



5% ridge-plates in 100 mm. 



Ridge-plates 

}i-7 

Yi-l 5}2 ridge-plates in 100 mm. 

4 partly exposed, 
heavy cement 

3 exposed 

K-6 5 ridge-plates in 100 mm. 

K"-6 5 ridge-plates in 100 mm. 

4 exposed 

4 only exposed 



yr-1-% 



bYi ridge-plates in 100 mm. 
5K ridge-plates in 100 mm. 



[This species was described by Professor Osborn in 1932, 
while visiting the Colorado Museum of Natural History at Denver. 
In his article in the Proceedings of the Colorado Museum (Osborn, 
1932.893) he states: "In the years of study which the present 
writer has devoted to the evolution of the fossil elephants, he 
became convinced that the 'Elephas meiiclionahs' of France is the 
direct ancestor of the 'Elephas imperator.' New and positive evi- 
dence of the correctness of this theory is now afforded by the dis- 
covery of the complete skeleton which forms the subject of the 
present paper. This skeleton with the lower jaw in a complete 
state of preservation proves to resemble very closely indeed in 
every detail the 'Elephas meridionalis' of Durfort, France, as fully 
described by Albert Gaudry. ... In August, 1931, it [the skeleton] 
was brought to the attention of Director Figgins of this Museum 
[Colorado Museum], who immediately took steps not only to 



rescue the fragments from careless visitors but to institute the 
complete excavation by the most skillful modern methods. The 
result will be most gratifying to palaeontologists all over the world, 
mainly for two reasons; first, because it is by far the most com- 
plete and perfectly preserved skeleton of Archidifskodon ever found, 
lacking only the cranium; .second, because it enables us to record 
the early migration of 'Elephas meridionalis' to North America, 
and thereby establish a direct ancestral relationship of the Durfort 
form to the present Nebraska mammoth." 

"Fortunately, the writer had recently secured from Director 
Boule of the Paris Museum a series of casts of the upper and lower 
grinding teeth of the Durfort specimen [Figs. 925 and 926 above]. 
Placed side by side with corresponding teeth of the Nebraska 
specimen, there can be no doubt that the two forms are closely 
related."— Editor.] 





ARCHIDISKODON MERIDIONALIS 3721 mm, I2'2 /s" 
1931 S. FRANCE DURFORT 



ARCHIDISKODON MERIDIONALIS NEBRASCENSIS 357£mm.,l2'>S* 
l<333 S. NEBRASKA 



Fig. 928. Restoration by Margret Flinsch Biiba of Archidiskodon meridionalis of Durfort and A. meridionalis nebrascensis of Nebraska, under the 
direction of Henry Fairfield Osborn. One-fiftieth natural size. 



Chapter XVII 

THE GENUS PARELEPHAS (SUPERFAMILY ELEPHANTOIDEA). OF THE 
SUBFAMILY MAMMONTINyE, INTERMEDIATE BETWEEN ARCHIDISKODON 
AND MAMMONTEUS, DISTRIBUTED IN THE NORTH TEMPERATE ZONE 

OF EURASIA AND NORTH AMERICA 



Profound cranial and incisive tusk resemblances to archidiskodon and mammonteus. Clear dis- 
tinctions FROM these genera IN GRINDING TOOTH STRUCTURE. FiRST APPEARANCE IN THE UPPER PLIOCENE^ 
AND LOWER PLEISTOCENE OF EUROPE. MIGRATION THROUGH ASIA MINOR AND ASIA TO THE UNITED STATES — 
WASHINGTON, NEBRASKA, TEXAS, ILLINOIS, INDIANA, OHIO, SOUTH CAROLINA, GEORGIA, FLORIDA — AND FRENCH 
GUIANA, SOUTH AMERICA. EXTINCTION IN UPPER PLEISTOCENE OR P0STGLACIAL(?) TIMES. 



1. European north temperate origin. History of separation from 

other extinct proboscideans. 
Complete separation by Osborn of the phylum Parelephas. 
Order of discovery and description of species of Parelephas. 
Phylogenetic order of succession of Parelephas. 
Distinctive cranial characters of Parelephas. 

2. Systematic description of European and Asiatic species in 

ascending progressive order. 
Parelephas Irogonlherioides, Italy. 
Parelephas trogontherii, Germany. 
Parelephasil) Irogonlherii neslii, England. 
Parelephas armeniacus, Asia Minor. 
Parelephas unrecorded in China and Japan. 
Parelephas intermedins, France. 
Parelephas wiisli, South Russia. 



3. North and South American species of Parelephas. 

Parelephas jacksoni, Ohio, e.xact locality unrecorded. 

ParelephasC^) mississippiensis(1), Indiana. 

Parelephas culumbi, Georgia. 

Parelephas columbi felicis, Mexico. 

Parelephas columbi cayennensis, French Guiana, 

South America. 
Parelephas jeffersonii, Indiana. 

Elephas roosevelti (synonym of P. jeffersonii), Illinois. 
Parelephas progressus, Ohio. 
Parelephas ivashinglonii, Washington. 
Parelephas eellsi, Washington. 
Parelephas floridanus, Florida. 



1. EUROPEAN NORTH TEMPERATE ORIGIN. HISTORY OF SEPARATION FROM OTHER 

EXTINCT PROBOSCIDEANS 

The generic i)hylum separated as Parelephas by Osborn in the year 1924 is briefly mentioned in Chapter II of 
the present Memoir and distinguished especially in its cranial characters in Chapter XV, in which it appears 
that Parelephas is linked with Archidiskodon and Mammonteus in its cranial resemblances and great incurved 
incisive tusks, while in its grinding teeth and ridge-plate formulae it is so nearly intermediate between these two 
genera as to have been mistaken for an actual connecting link. In the present chapter it is shown to be an entirely 
distinct generic phylum which during the more temperate interglacial periods (Fig. 795) occupied the same geo- 
graphic range as that of the true woolly mammoth (Mammonteus) during the glacial periods. 

History. — The sixteen species which are grouped in the genus Parelephas constitute a very ancient and 
distinct generic phylum for which Osborn's name {Parelephas) seems appropriate, in reference to the convergence 
or parallelism of the grinding teeth in this phylum with those of the true Elephas. For more than half a century, 
owing to the similarity in appearance of the grinding teeth to those of the true mammoth (Mammonteus), all 
English and American authors, including Falconer, Leith Adams, and Lydekker, confused the grinding teeth of 
specimens which Osborn now refers to Parelephas with those of the northern mammoth Elephas [Mammonteus] 
primigenius, and this accounts for the great discrepancies in the collective ridge formulae attributed to the species 
E. primigenius, i.e., M 3 \^ as given by Leith Adams (1877-1881, p. 127) and by Lydekker (1886.2, p. 175), and 
copied by Hay (1914, p. 395). This unfortunate confusion in the ridge formula arose despite the fact that Falconer 

'[See footnote 1 on page 1049 below. — Editor.) 

1039 



1040 



OSBORN: THE PROBOSC'IDEA 



as early as 1863 (p. 65) had correctly defined the ridge-plate formula of E. primigenius as M 3 ff, a formula which 
is fully confirmed by Osborn's researches for this Memoir (see Chapter XVIII). The lower ridge-plate count 
(M 3 tI) attributed by Adams and Lydekker to Elephas primigenius really belongs to grinding teeth of primitive 
Middle Pleistocene species of Parelephas, such as P. trogonlherii, which exhibits M 3 J g+ ridge-plates. As described 
in detail below, Jourdan (1861) was the first to separate from E. primigenius one of these species of Parelephas 
under the name Elephas intermedius, a stage which exhibits M 3 |f ridge-plates; he was followed by Pohlig (1885) 
who clearly defined the more primitive Elephas trogontherii. 




I- 

it 

i- 

L 

Restoration of the Type of Parelephas jeffersonii 
One-thirtieth natural size 

Fig. 9.30. This painting hy Charles R. Knight, in the year 1909, was taken direetly from the type skeleton of Parelephas jeffemonii in the 
American Museum of Natural History. The characters of the typical /'a;f/cp/i(i,s cranium and tusks are particularly well shown, with short concave 
foreiiead an<l prominent convex occipitofrontal ere.st. The sli.ape of the ears i.s entirely conjectural. The hairy covering, unlike the hairy and 
woolly covering of Mainmoidius /jriiiiigeuius, is a wholly conjectural character, because no remains of the iiair have been discovered, hut the presence 
of this Jeffersonian mammoth in north temperate regions, appearing in post(?)-Wisconsin times, furnishes indirect evidence of a hairy if not of 
a woolly coat. 



In the present chapter, chiefly from the researches of Deperet, Mayet, and Osborn, it is shown that the 
phylum Parelephas constitutes a long line of progressive ascent wholly distinct from that of Mammonteus primigenius. 
Its first appearance is in Upper Phocene time in Italy. Its final appearance in IV Glacial and Postglacial times, 
principally on or near the 40th parallel of the United States, in the species Parelephas jeffersonii (M 3 H), is 
followed by the closing stage Parelephas progressus (M 3 M). 



THE MAMMONTINiE: PARELEPHAS 



1041 



It has also been a long and difficult matter both in Europe and America to clearly separate the members of 
the generic phylum Parelephas from the newer phylum of the broad-toothed, narrow-plated true mammoth 
(E. [ = Mammonteus] primigenius) on the one hand, and from the older phylum of the broad-toothed, broad-plated 
southern {Elephas meridionalis) and imperial {E. imperator) mammoths of the genus Archidiskodon on the other. 
The chief grounds of separation are as follows: (1) The cranium in Elephas intermedius, E. irogontherii, and E. 
jeffersonii is now known to be readily distinguishable from the crania of either E. [ = Mammonteus] primigenius or 
E. [ = Archidiskodon] imperator; (2) the grinding teeth are intermediate in form and in the number of plates, as 




Second Figure of the Type Skeleton of Parelephas jeffersonii 
One-thirtieth natural size 
Fig. 931. Second figure of the aged type skeleton of Parelephas jeffersonii Osborn, 1922, p. 11, fig. 10, as mounted in the American Museum 
(Amer. Mus. 9950). For further information about this type skeleton, see legend of figure 966, also the description below. 



observed by Jourdan in 1861 in applying the name Elephas intermedius; (3) it is noteworthy that both in France 
and Germany the grinding teeth of Parelephas have been independently described as intermediate in structure 
between those of E. [ = Archidiskodon] meridionalis and E. [=Mammonteus] primigenius, as shown in the full 
historic notes below. Elephas columbi proves to be Parelephas columbi. 

France: Jourdan, 1861. — Despite the early separation by Jourdan of Elephas intermedius from E. primi- 
genius and description of the grinding teeth as intermediate between E. primigenius and E. meridionalis (1861), 



1042 



OSBORN: THE PROBOSCIDEA 



Fig. 032. Chief Lower to Upper Pleistocene localities in wliieh occur species of 
Archidiskodon, Parelephas, Mammonteus, Loxodonla, and Palseoloxodon (syn. Pilgrimia), 
after Osborn, 1910.346, p. 391, fig. 176. 

Fossils attributed to the ijhylum Parclephas, and of large size, are especially abun- 
dant in the Middle Pleistocene of Siis-senborn (Fig. 932,11), of Mosbaeh (12), and of 
Taubach (19). At Mosbaeh (in Isl! Inlcrghcial times) they occur in the same layers, 
but with much greater frequency, than those of E. [Hesperoloxodon] antiquum, namely, 
three to one (Soergel, 1912, p. 32); more recently it is estimated (Schmidtgen, 1926) 
that in the Mosbaeh sands E. [Parclephas] trogonthtrii is about ten times as abundant as 
E. [Hesperoloxodon] antiquus; in Mosbaeh also this elephant attains its greatest size. 



Gaudry, who was very conservative in the matter of applying new names to species, continued to describe (1876, p. 
40, also PI. ix) these animals as "Elephas primigenius a lames ecartees." The steps in the gradual separation of 
this phylum by palaeontologists of France are clearly described by Deperet and Mayet (1923, pp. 176-190). 

Italy: Falconer, 1868. — As early as 1868 
(see below under Parelephas armeniacus) , Fal- 
coner entertained a strong suspicion that a 
form closely related to E. armeniacus occurred 
at St. Paolo, Italy, which he identified (1868, 
II, pp. 249, 250) as resembling E. armeniacus, 
but which we now know belonged to Parelephas 
irogontherioides (see p. 1055). This is a fine ex- 
ample of Falconer's unerring sense of form. 

Germany: Pohlig, 1885.— In 1885 Pohlig, 
on discovering two grinding teeth of one of 
these mammoths in the Interglacial sands of 
SUssenborn near Weimar, proposed the specific 
name Elephas irogontherii; he maintained that 
both geologically and zoologically this species 
was a link, or intermediate, between Elephas 
primigenius [i.e., Mammonteus] and Elephas 
meridionalis [i.e., Archidiskodon]. To quote 
Pohhg's own language {op. cit., 1885, p. 1027): 
"Unter der Bezeichnung 'Elephas Irogontherii 
Pohl.' fUhre ich in meiner Monographic eine 
europaische Molarenform auf, welche zwischen 
denjenigen des E. primigenius und E. meridi- 
onalis zoologisch, wie ihrer geologischen Lager- 
stiitte nach, in der Mitte steht." 

Pohlig also observed the relationship of E. 
Irogontherii to the species which Falconer named 
Elephas armeniacus in 1857. At the same time 
Pohlig erred in suggesting the relationship of E. 
trogonlherii to Elephas [Palxoloxodon] natnadicus 
Falc. ; he also erred in suggesting that both in 
craniology and dentition there was a direct 
phylogenetic or ancestral succession between 




PLBISTOCBNX EUROPE. — 1 Forrst Bed of Cromer (Norfolk). Sables de 2 
St. Prest near Chiirtres (Eure-et- Loire). 3 Maihnitu (Puy-de-Dorae). 4 Peyrolles (Bouches- 
du-Rhonc). 5 .Holhilac near Puy. Clay deposits of 6 Uurforl (Card). 7 Cnjarc (Lot-et-Ga- 
ronne). 8 Vol d'.irno (Tuscany). 9 Leffe near Bergamo (Lombardy). 10 Rixtlor/ near Pots- 
dam (Brandenburg). Gravels of 11 SUssenborn near Weimar. Sands of 12 Mosbaeh in 
northern Baden. Freshwater deposits of 13 CUicton (Essex). Sands of Mauer near 14 Hei- 
delberg (western Germany). 15 Chelles on the Mame, near Paris. 16 SI. Acheul (Sonime). 
17 Il/ord and Grays Thurrock (Essex). Lignites of 18 Diimlen and of Utznach, near Zurich. 
19 Tauhach near Weimar. 20 WtWirrcWi cooe on A/on( 5an(is (eastern Switzerland). Tuffs 
of 21 the Tiber Valley, near Rome. Caves of 22 Neandertal, near Dusseldorf (western Ger- 
many), 23 .Spy. near .\mur (Belgium), 23a Krapina (Croatia), 24 Chapelle-aux-Soints (Cor- 
r^zc). Caves and alluvial deposits of 25 Temifine (or Palikao) near Gran (Algeria). 26 Pointe 
Peacade, near Algiers (Algeria). 27 Prince's Cave (Monaco). Sandy clays of 28 Vdklinshofen 
(Alsace). 29 Saalfcld (.Saxe-.Meiningen). Travertines, etc., of 30 Gera, Jena (.Saxe-Weimar). 
31 L«ipii'fl (Saxony). 38 .So/u(rc, north of Lyons. Loess of 33 Wurjfcurff (Bavaria). H Thicde 
near Braunschweig (Prussia). Cave of 35 Montmaurin (Haute-Garonne). 36 Chdteauneuf- 
sur-Chartnle (Charente). Caves of 37 .Schweizershild near Schaffhausen, and Kesslerloch near 
Thayngen (northern Switzerland). Remains of lake dwellings at 38 Wauwyl (Lucerne), 39 Ro- 
benhausen, south of Lake Pfaffikon. 40 Concise on Lake Neuchatcl (Switzerland). Peatbogs of 
41 Hassleben. near Weimar. Travertines of 42 Langenaalza (Erfurt) in central Germany. 
Caves of the 43 Island ol Malta, 44 Island of Crete, 45 Island of Cyprus. 



INot on map: 46 San Paolo Je Villafranca (Piedmont), Italy, 
loess. Lyon, France. 49 Tira.spol, .S. Kuasia) 



47 Eizcruni, .\rnienia. 48 Plateau 



E. meridionalis, E. Irogontherii, and E. primigenius (i.e., "directer Verwandtschaft"). 

In 1886, p. 181, Pohlig remarks: "Under the name of Elephas trogontherii, Pohlig, I have described European 
molars which hold a middle place, both zoologically and geologically, between those of E. primigenius and E. 
meridionalis, most closely apprf)ucliiiig those of PJ. antiquus in the ridge-fonnula, but differing more from them than 
from the other two in the form of the crown. The position of E. Irogontherii with regard to E. armeniacus, Falc, 
and E. namadicus, Falc, still remains to be investigated." 



THE MAMMONTIN^: PARELEPHAS 1043 

France: Deperet, 1923. — In 1923, Deperet and Mayet in their invaluable review of "Les Elephants 
Pliocenes," Deuxieme Partie, out of respect for Pohlig's clear distinction and definition of the species E. trogon- 
therii, decided not to adopt the term Elephas intermedius of Jourdan, but to name this phylum "Rameau de 
VElephas trogontJierii." They recognized its remote kinship to Elephas primigenius (see "III Groupc des Elephas 
irogontherii et E. primigenius (Mammouths)") but established it (p. 176) as a distinct branch of the "groupe des 
Mammouths." Osborn believes this to be its true phyletic position, resting on two chief characters, namely: 
(a) The intermediate character of the grinding teeth, observed by Jourdan, by Gaudry, by Pohlig, and by Deperet 
and Mayet; (b) the dome-shaped cranium, figured and observed by Pohlig and by Deperet and Mayet. 

In their analytical treatment of the geologic succession of species, Deperet and Mayet clearly pointed out that 
the type locality {j3d Interglacial plateau loess near Lyons) of Elephas intermedius Jourd. is of more recent geologic 
age than the type locality {Sd rnterglacial sands) of Elephas trogontherii Pohl.; moreover in seeking a Pliocene 
ancestor of this phylum, they believe they have discovered it in the Upper Pliocene subspecies of Italy, named by 
Zuffardi in 1913 Elephas antiquus var. frogontherioides. They thereby confirmed Falconer's observation of 1868 
that an elephant closely similar to E. armcniacus occurred in northern Italy (see Parelephas armeniacus descrip- 
tion, p. 1060 below). 

Consequently the actual geologic succession of the three types of Parelephas thus far discovered in France and 
Italy is as follows : 

Pleistocene {Sd Inkrglaciul plateau loess), Lyons, France Elephus intermedius Jourdan, type 

Pleistocene (2d Interglacial sands), Siissenborn near Weimar, (lermany Elephas trogontherii Polilig, type 

Upper Pliocene (Villafranchian), Italy [see footnote, p. 1049 below. — Ed.] Elephas antiquus var. trogontherioides Zuffardi, type 

America: Osborn, 1922. — Among the abundant remains of members of this phylum in America, a precisely 
similar confusion between specimens belonging to Archidiskodon, Parelephas, and Mammonleus arose in the minds 
of American palaeontologists (Cope, Osborn, and Hay). Cope (1889.2, pp. 208, 209) referred the remains of a very 
fine skull of Archidiskodon (Fig. 891) from Texas to "Elephas primigenius columbi Falc." Osborn also saw only 
resemblances to "Elephas columbi" in the fine skeleton found in Indiana (type of Parelephas jeffersonii Osborn, 
Figs. 961, 931); whereas Hay (1914) referred the same skeleton to Elephas primigenius. No stronger proof 
could be given of the truly intermediate character of members of this phylum in America than this alternate 
reference of the same skeleton by Osborn to Elephas columbi and by Hay to Elephas primigenius. In Hay's 
exhaustive summary (1914) of grinding teeth and other remains from various parts of the United States (Alaska 
to Florida) which he referred to "Elephas columbi," he includes teeth which certainly belong to Parelephas columbi 
as well as to the truly intermediate form (P. jeffersonii) . 

Osborn (1922.555) was the first to separate these intermediate animals, previously described by Cope, Hay, 
and himself as "E. columbi" and "E. primigenius," under the new specific name of Elephas jeffersonii, a species 
which he subsequently (1924.633, p. 4) made the genotype of Parelephas. This separation was based principally 
upon cranial characters (Osborn, 1922.555, p. 15): "Still more obvious are the differences between the relatively 
long, broad, and shallow crania of E. jeffersonii and the relatively short, narrow, and deep crania of E. primigenius^ 
proportions which are correlated respectively with the corresponding proportions just described and figured in 
the teeth." 

Phylogenetic Relationships (Pohlig, Soergel). — Inasmuch as the same species of animal had previously 
been erroneously described by Cope and Osborn under the name of "Elephas columbi," we must credit Soergel 
(1921, p. 60) with the prior observation that this supposed "Elephas columbi" of America { = Elephas [Parelephas] 
jeffersonii Osborn) shows phyletic relationships to the Elephas [ = Parelephas] trogontherii of Pohlig. 



1044 OSBORN: THE PROBOSCIDEA 

Schmidtgen and Freudenberg (1926, pp. 62, 68) describe and discuss Elephas [Parelephas] Irogontherii of 
Mosbacli, especially the supposed phylum ("Stammreihe") Elephas meridionalis — E. Irogontherii — E. primi- 
genius, which in unbroken succession lived in the region of Wiesbaden. 

Osborn, 1924-1928: It has remained, however, for Osborn in the present Memoir to institute a close compari- 
son between all the known characters of the Elephas Irogontherii phylum of Europe and the Elephas jeffersonii of 
America and to establish the fact that there is a close phyletic relationship which justifies the linking of the 
European and American species in the new and distinct genus Parelephas. This aflfinity is most clearly indicated 
in the cranium, as shown in the resemblance of several crania, erroneously figured as "Elephas primigenius" by 
Falconer, to those of E. trogontherii as figured by Pohlig, and to the cranium of E. jeffersonii as figured in the 
present Memoir; there can be no doubt that this relatively broad, elongate, and rounded cranium is entirely 
distinct from the extremely short, compressed, and peaked cranium of the true mammoth Elephas primigenius. 

Habitat of Parelp:phas trogontherii (Soergel, 1912). — In his masterly monograph of 1912, entitled 
"Elephas trogontherii Pohl. und Elephas antiquus Falc," Soergel observes (pp. 105, 106) that the Steppe [plains] 
and \^'ald [forest] faunas lived contemporaneously in central Europe; as observed in the Id Interglacial period, 
they are found in the lower sands of Mosbach, the "kiese" [gravels] of Mauer, the sands of Petersdorf bei Gleiwitz 
in Schlesien, the "Tone" [clays] and "grauen Rheinsande" [gray sands of the Rhine] near Jockgrim in Pfalz, and 
the gravels of Sussenborn (which persist into II Glacial time).' He divides the plains or "Steppe" and forest 
or "Wald" faunas of the 1st Interglacial period as follows: 

Chiefly Inland Steppes or Plains Chiefly Ocean Borders and Forests 

E. [Parelephas] Irogontherii Pohl. E. [Hesperoloxodon] anliquus Falc. 

Very abundant in Mosbach; absent in Mauer. Very abundant in Mauer. 

Dicerorhinus etruscus F'alc. Dicerorhinus etruscus Falc. 

Equus stenonis Cocchi, E. siissenbornensis Wlist, Equus stenonis Cocchi, E. mosbachensis v. Reich. 

E. mosba£hcnsis v. Reich. 
Leptobos etruscus Falc. 

Bison priscus Boj. Bison priscus Boj. 

Cervus elaphus Linn., C. elaphus trogontherii Pohl., Cervus elaphus antiqui Pohl., C. capreolus Linn., 

C. capreolus Linn., Alces lalifrons .lohns. Aloes lalifrons .Johns. 

Ursus arvernensis Croiz. and Job., U. Deningeri v. Ursus arvernensis Croiz. and Job., [/. Deningeri v. 

Reich. Reich. 

Felis lea fossilis Goldf. Felis leo fossilis Goldf. 
Hysena arvernensis Croiz. and Job. 

Canis neschersensis Croiz. and Job. Canis neschcrsensis Croiz. and Job. 

[Homo heidelbergensis Schoet. (lower sands of Mauer)]. 

Certainly Hesperoloxodon antiquus, a forest loving form with coarser teeth, occurred nearer the ocean borders, 
whereas Parelephas trogontherii, with finer teeth, frequented both the steppes (plains) and forests. Since animals 
of both habitats are buried together in river sands, gravels, and clays, there is an intermingling of plains and forest 
faunas, as shown in Mosbach and Mauer (Osborn). 

Of the habitat and geographic distribution, Soergel {op. cit., p. 110) observes that whereas Hesperoloxodon 
antiquus preferred a warmer ocean bordering climate, not under the direct influence of the northern inland ice 
masses, Parelephas trogontherii sought the cooler northern-northeastern continental localities. Accordingly in 
Italy, Spain, and Greece it is almost entirely wanting; in France P. trogo7itherii is less abundant than H. antiquus, 
at least, it appears in less characteristic forms. The homeland of the P. Irogontherii type appears to be confined to 
England and Germany, perhaps al.so to Russia. During the arid 1st Interglacial peroid we find P. Irogontherii 
distributed from southern England over middle Germany and eastward to southern Russia, as shown in the 
Forest Bed of Cromer, the sands of Mosbach, the gravels of SUssenborn, the sands of Petersdorf in Schlesien, and 
the clays of Jockgrim in Pfalz. Already in this Isl Interglacial period migration towards the east may have begun. 

'[See p. 1056 below, where Parekphas Irogontherii type is placed by Professor Osborn in the 2d Interglacial. — Editor.] 



THE MAMMONTINiE: PARELEPHAS 



1045 



In the 3nd Interglacial period we observe practically the same geographic distribution of Parelephas tro- 
yontherii; its southernmost appearance in Germany at this time is at Steinheim on the river Murr; its northern- 
most appearance is at Rixdorf near Berlin. In Thuringia it is found in the Ilm gravels below the older travertines 
of Taubach, also in the stream gravels at Vieselbach near Erfurt, etc. 

For the better understanding of the geologic as well as the geographic distribution of the Parelephas trogon- 
therii phylum in Germany may be cited the following geologic correlation by Soergel (op. cit., p. 106) : 



Dem I. 'Interglazial' gehoren an: 
die Sande von Mosbach [n. 

Baden], 
die Kiese von Mauer [b. Heidel- 
berg], 
die Sande von Petersdorf b. 

Gleiwitz in Schlesien, 
die Tone und grauen Rhein- 
sande b. Jockgrim i. d. 
Pfalz, 
die Kiese von Siissenborn [bei 
Weimar], die allerdings 
bis in die \l. Eiszeit hin- 
eingehen. 
[Plains ("Steppe") and forest 
("Wald") faunas, as listed above.) 

The Siissenborn deposits extend into 
H Glacial time {fide Soergel, 1912.) 



Dem II. 'Interglazial' gehoren an: 
die Bachkiese bei Vieselbach, 
die Saalekiese von Uichteritz 

b. Weissenfels, 
die Ilmkiese unter dem alteren 

Travertin von Taubach- 

Ehringsdorf [b. Weimar], 
die Schotter von Steinheim 

(II. Glazial-Interglazial). 



die Sande von Rixdorf b. Berlin 

fiir eine Ablagerung des II. 

'Interglazial.' 

[Plains and forest faunas not clearly 

separated ; Parelephas trogontherii and 

Hesperoloxodon antiquus occur in the 

same horizon.] 



Dem III. 

der 

dcr 

der 
und 



Interglazial gehoren an: 
Travertin von Taubach- 
Ehringsdorf- Weimar, 
Tra\ertin von Burg-Cirae- 
fentonna, 

Travertin von Bilzingslcbcn 

verschiedene andere Tra\-- 

ertinvorkommen Thiiringcns. 



[Plains and forest faunas again api^ear 
partly separated, as in Mauer and 
Mosbach. 

Hesperoloxodoiiantiquussurvives Par- 
elephas trogonlherii, which disappears. 

In the tundra fauna appears Mam- 
monteus pr imige nius.] 



Complete Separation by Osborn of the Generic Phylum Parelephas 
Osborn finally (1924.633, p. 2) concluded to cut the Gordian knot and terminate this confusion by distinguish- 
ing the generic phylum Parelephas throughout from both Mammonteus and Archidiskodon, as shown in the follow- 
ing citation : 

[p. 2] Much more difficult has been the separation of the third generic series of the Mammontinae, which hitherto has been 
referred to the genus Elephas but which we now remove to the new generic phylum Parelephas. 

[p. 4] The eight or ten species included within this genus constitute a very ancient and wholly distinct generic phylum, for 
which the name Parelephas seems appropriate, because in certain characters these animals parallel the true Elephas, although in 
profound cranial and dental structure they are closely related to and convergent with the mammoths Archidiskodon and 
Mammonteus. 

Generic Characters. — A phylum of the Mammontinae. Elephas jeffersonii, genotypic species, E. armeniacus, E. 
intermedins, E. trogontherii, E. trogontherioides. Craniimi intermediate in form between that of Archidiskodon and that of 
Mammonteus, namely, brachycephalic, acrocephalic. Frontals concave, occipital crest elevated; occiput more or less convex. 
Molars in the Upper Pliocene and Lower Pleistocene stages with relatively few ridge-plates, i.e., M 3 -}-f+; progressive Upper 
Pleistocene stages with multiple ridge-plates, i.e., M 3 f§. Ridge-plates compressed to 7-8-9 in 100 mm. Alolar crowns broad, 
RP short, with enamel of intermediate thickness, more or less crimped or sinuous. 

The progressive ridge formulae in Parelephas are distinct throughout from those of Mammonteus, and the final ridge formulae 
in the two generic phyla are different, namely: 

Mammonteus primigenius comprcssus, M 3 Tf = final stage. 

Parelephas jeffersonii progressus, M 3 ff = final stage. 

The crania of Parelephas throughout are readily distinguishable both in frontal and lateral aspects, and especially in vertical 
section, from those of Mammonteus, as can be seen in all of Falconer's beautiful plates of E. primigenius and in Pohlig's excellent 
figures of P. trogontherii. 

The jaws of Parelephas and of Mammonteus are less readily distinguishable, but by more profound study they can also be 
separated from those of Mammonteus. The contrasts in the crania of the two genera may be summed up as follows: 

Mammonteus. — Cranium and jaws extremely compressed fore-and-aft (cyrtocephalic) ; extremely elevated and pointed 
above (hypsicephaHc) ; extremely depressed and foreshortened below (bathycephalic). 

Parelephas. — Cranium moderately compressed fore-and-aft (cyrtocephalic); moderately elevated occipitofrontal borders 
(acrocephahc) ; moderately depressed molar-grinding area (bathycephalic) . 

Thus, while the intermediate forms of crania and teeth of Parelephas and of Mammonteus may prove difficult to separate, 
the two finally progressive forms are readily separable, namely, Parelephas jeffersonii progressus and Mammonteus primigenius 
compressus. 



1046 OSBORN: THE PROBOSCIDEA 

To this phylum certainly belongs the Elephas armeniacus of northern Asia Minor, as first observed by Fal- 
coner, as well as many specimens to the eastward described under other names. The connection of this jihylum 
with the Parelephas jeffersonii of late Glacial times in the United States seems highly probable; the reseml)lance 
both of the teeth and cranium has been repeatedly observed between E. trogontherii and the northerly tyi)c form- 
erly known as "Elephas columbi," now known as Parelephas jeffersonii. 

The southerly type (Georgia, South Carolina, Florida) of Elephas [ = Parelephas] culumbl, with rather 
primitive ridge formula (M 3 rirp), has recently been reinforced by the large and more progressive Parelephas 
Jloridanus (M 3 ^l), apparently a late Pleistocene stage distinguished by its larger size, more robust tusks, and 
coarser ridge-plates from the northerly (Indiana) Parelephas jeffersonii type. 

A branch of the Parelephas cohnnbi stage appears to have migrated to South America, as mentioned Ijy Lull 
1908, p. 204) and Freudenbcrg (1922, p. 159) on the authority of Lartct (1859, pp. 500, 505). 

I.AiriKT, \H'->9. VI'. .'iOO' and 505. — L'existcncc des Ck'phants n'cst encore indicnit'e dans rAnierifiue nieridionalc ((ue ])ar un 
fragment de molaire a lames epaisses, rapporte de Cayenne i)ar le capitaine Ferret et donne par lui au Miisec de Marseille. . . . 
Dans rAniprique dii Sud, deux formes si)ecifiques du genre Mastodonte se montrent dans des depots posl-pliocenes; mais ]ieiit- 
etre se sont-elles aussi retrouvees dans des formations plus anciennes et raijportables a I'age precedent on pliocene, (^uant au 
type Elephant, il n'y est encore indique que par le seul fragment deja cite d'une molaire a lames epaisses, rapports de Cayenne 
par le capitaine Ferret. 

iMiErDENBERC, 1922, PP. 159, 160. — Wir haben Lartets Autoritat fiir den Fund eines Zahnes von Elephas im unteren 
Fleistociin von Cayenne (Franzosisch-Cniayana). Nach der Reschreibung der dickriickigen Schmelzplatten ist das Excnijilar 
offcnbar von El. impcrator. Wahrscheinlich vollzog sich eine Wanderung nach dem Siiden, bevor die Redingungen geschaffen 
«aren. wclche spater die Wanderung der Proboscidier nach Siidamerika ermoglichten. . . . Ks ist das einzige mir bekannt 
gewordene Beispiel eines wahren Elefanten im Siiden des Hochlands von Mexiko. 

Osborn, 1929: Through the extreme courtesy of Mm. Laurent and Repelin of the Faculte des Sciences de 
Mar.seille, this priceless molar fragment from French Guiana has been located in the Museum d' Histoire Naturelle 
fie Marseille, photographs taken, and casts made, especially for the purpose of specific and generic identification 
for the present Memoir. As shown in figure 957, on a one-half scale, this specimen consists of three ridge-plates 
of coarse enamel, as described by Lartet, remotely resembling in crown and side views the posterior ridge-plates 
of a superior grinding tooth of Parelephas columbi (Fig. 952). In 1929 the present writer (Osborn, 1929.797, p. 20) 
made this the type of a new subspecies, namely, Parelephas columbi cayennensis. 

Th(> rcH-entlj- disco\-('red Parelephas Jloridanus includes parts of three full skeletons collected by the Holmes 
Florida Expedition of the year 1929. In this stage, intermediate in ridge-plate fornuila (M 3 ff^) between Elephas 
[Parelephas\ columbi (M 3 \j+) and P. jeffersonii (M 3 H), the ridge-plates are broad at the base and compressed 
at the summit ; the inci.sive tusks extremely massive and relatively short. The males attain very large size. The 
femora measure, according to age, 1230 mm. to 1393 mm., as compared with 1250 mm. in the adult Indiana type 
of P. jeffersonii. 

Summary of Migration 

The restudy of long-known specimens like Elephas jacksoni (1838), Elephas columbi (1857), Elephas armen- 
iacus (1857), and Elephas intermedius (1861), together with the recent recognition of Parelephas columbi cayen- 
nensis in French Guiana, builds up a Parelephas phylogeny and migration second only to that of Archidiskodon 
(Chapter XVI). We await the discovery (probably in northern Africa) of an ancestral stage more i)rimitive than 
Parelephas trogontherioides of the Uppei- Pliocene- of Italy. 

'Dr. Richard S. Lull (letter, Sept. 21, 1928) wrilt's: "rruin this (icworiptiori, inentioiiinE the fragment of ji molar with 'thick plates,' I drew my own 
conclusion as to the genus and species of that elephant. It seemed to me that that it would be Elephas [ —Archidiskodon] iiuj/cralor." 
-[See footnote 1 on page 1019 below.— Editor.) 



THE MAMMONTIN^: PARELEPHAS 



1047 




Fig. 933. Geographic distribution (aecording to tlie numbers in the accompanying hst) of species of Parelephas. The white dots within the black 
areas represent the approximate localities where the types of these species were discovered. The white crosses represent referred specimens. It will be 
noted that Professor Osborn regarded EUpkas texianus Owen, Blake (1859-1861) as a synonym of Parelephas columbi, and Elephas roosevelli Hay (1922) as 
a synonym of Parelephas jeffcrsonii. 



ORDER OF DISCOVERY AND DESCRIPTION OF SPECIES OF PARELEPHAS 

See Figure 933 



1838 1. Jackson County, Ohio 

1857-1868 2. Brunswick Canal, Georgia 

1857 3. Armenia, Asia Minor 

1859-1861 4. San Felipe de Austen on Brazos 

River, Texas 

1861 5. France, 3d Interglacial loess 

1872 6. Indiana 



1872 
1885 

1891 



1910 


9 


1913 


10 


1922 


11 


1922 


12 


1922 


13 


1923 


14 


1924 


15 


(1924 




[1926 




1926 


16 


1929 


17 


1929 


18 



6. Indiana 

7. Siissenborn near Weimar, Germany, 

2d Interglacial sands 

8. Cromer Forest Bed, East Anglia, 
England 

Tiraspol, southern Russia 
Piedmont (Villafranchian), Italy 

Jonesboro, Grant Co., Indiana 
Mexico, Tecamachalco, Puebla 
Ashland, Cass Co., Illinois 
Pine Creek, Whitman Co., 

Washington 
Zanesville, Ohio 
Province of Kazusa, Japan 

Province of Kazusa, Japan 

Port Williams, Clallam Co., 

AVashington 
Florida, near Bradenton, Manatee 

County 
Cayenne, French Guiana, South 

America 



Original Name 

Elephas jacksoni Mather 

E. {Euelephas) Columbi Falconer 

Elephas armeniacus Falconer 

Elephas texianus Owen, 1859, Blake, 

1861 
Elephas intermedius Jourdan 
Elephas Indianapolis Foster. Later in 

same year changed to Elephas missis- 

sippiensis — see below 
Elephas mississippiensis Foster 

Elephas trogontherii Pohhg 

Elephas {antiquus) Nestii Pohlig 

Elephas Wilsti Pavlow 

Elephas antiquus var. trogontherioides 

Zuffardi 
Elephas jeffersonii Osborn 
El. Columbi var. Felicis Freudenberg 
Elephas roosevelti Hay 

Elephas ivashingtonii Osborn 
Parelephas jeffersonii progressus Osborn 
Euelephas (Parelephas) prolomammon- 

teus Matsumoto 
Parelephas protomammonteus proximus 

Matsumoto 

Elephas eellsi Hay 
Parelephas floridanus Osborn 
Parelephas columbi r.ayennensis Osborn 



Specific Refekence in 
Present Memoir 

= Parelephas jacksoni 
= Parelephas columbi 
= Parelephas armeniacus 

= Parelephas columbi 
= Parelephas intermedius 

= Nomen nudum 

= Indeterminate, possibly Parele- 
phasiy.) mississippiensisiy.) 

= Parelephas trogontherii 

= Parelephas{?) trogontherii nestii 
= Parelephas wiisti 

= Parelephas trogontherioides 
= Parelephas jeffersonii 
= Parelephas columbi felicis 
= Parelephas jeffersonii 

= Parelephas washingtonii 

= Parelephas progressus 

= Palxoloxodon protomammonteus — see 
Chap. XIX] 

= Palseoloxodon protomammonteus prox- 
imus — see Chap. XIX] 

= Parelephas{1) eellsi 

= Parelephas floridan us 

= Parelephas columbi cayennensis 



1048 



OSBORN: THE PROBOSCIDEA 



Sui>erfamily: ELEPHANTOIDEA Osborn, 1921 

Family: ELEPHANTID^ Gray, 1821 

Subfamily: Mammontin^e Osborn, 1921 

Genus: PARELEPHAS Osborn, 1924 
Original reference: Osborn, Amer. Mus. Novitates, No. 152, pp. 4 and 5 (Osborn, 1924.633). 

Generic Characters (cf. Osborn, 1924.633, p. 4). — Genotypic species, Elephan jeffersonii Osborn, 
of Indiana; species referred to this genus, Elephas armeniacus, E. intermedius, E. trogonlherii, and E. 
trogontherioides. Cranium intermediate in form between that of Archidiskodon and that of Mammonteus, 
namely, bathycephalic, acrocephalic. Frontals concave, occipital crest elevated; occiput more or less 
convex. Molars in the Upper Pliocene to Middle Pleistocene stages with relatively few ridge-plates, 
i.e., M 3 14+; progressive Upper Pleistocene stages (P. progressus) with multiple ridge-plates, i.e., 
M 3 26. Ridge-plates compressed to 6-6.5-7-8-9 in 100 mm. Molar crowns broad, M^ short, with enamel 
of intermediate thickness, more or less crimped or sinuous. Adapted to continental plains or steppe 
environment, grazing and browsing. 



I Pl)er Pleistocene 


Ohio 


i'leistocene 


Ohio 


Upper Pleistocene 


Indiana 


Pleistocene 


Illinois 



APPROXIMATE PHYLOGENETIC ORDER OF 



Parelephus progressus Osborn 



SUCCESSION OF PARELEPHAS 

Ridge Formula Length Breadth Height 



Upper(?) Pleistocene 



Ui)per(?) Pleistocene 


Florida 


rp])er(?) Pleistocene 


(leorgia 


Pleistocene 


Washington 


Pleistocene 

3rl Interglnrinl 
U])per Pleistocene 
Pleistocene 


THEORE 
France 

S. Russia 
Asia Minor 


Middle Pleistocene 

-^f/ Intcrglacial 
[Abundant, altaining 
greatest size (Mosbacl: 


CJcrmany 
(Siisscnborn) 

i)J 


Lower Pleistocene 
Isl I liter glacial 


( lermany 
(Basal Mosbauli 


Lower Pleistocene 
Isl Inlerglacial 

Upper Pliocene 
Villafranchian' (type) 
English Crag (ref.) 


lOngland 
Forest lied 
(Cromer) 
Italy 



Parelephas jacksoni Mather 
(Imperfectly known) 

Parelephas jeffersonii Osborn 

Elephas roosevelti Hay = syn. 
Parelephas jeffersonii 
French Guiana Parelephas columbi cayennensis 
So. America Osborn 

Parelephas floridanus Osborn 
Parelephas columbi Falconer 
Parelephas washingtonii Osborn 

THEORETIC PERIOD OF MIGRATION TO NORTH AMERICA 



Type: 


M3 


30 
2 6 


205 
221 


109 
97 


203 
161 


Type: 


M3 
M2 


i^ 








Type: 


M3 


+ 1 7 
+ 2 


203 + 
208 


108 
86 


168 


Type: 


M3 


25 
2 1 


282 
304e 


98 
92 


202 
154 


lype: 


M 3 superior — fragmentary 






Type: 


M3 


22 + 
21 + 


320 
290 


88 
79 


235 
180 


Ref.: 


M3 
M3 


18-19 


249 
24 le 


96 

82e 


207 


Type: 


15-1 6 + 


152e 


Ref.: 


M3 










Typo: 


M3 


23 
2 I 


348 


88 





Parelephas intermedius Jourdan 

Parelephas ivusli Pavlow 
Parelephas armeniacus F'alconer 

Parelephas trogonlherii Pohlig 



Parelephas trogonlherii 

) ("kleincre Var.") Freudenberg 

Parelephas{1) trogonlherii nestii 



Purelrphiis trogontherioides 
Zuffardi 



Ref.?: M 3 

Cotypes: M 3 
Type: M 3 

Type: M3 



Ref. : M 3 



Cotypes: M 3 



2 0- 2 2 
30-21 



22 

+ 1714 



15 + 

16 + 



270 

340e 

299 

225 

270 + 



92 

lOOe 

105 

101 

75 



Not available 



Lnclotype: 
('otyi)e: 



M3 



la 

14 + 



M-14-W 



302 

243 + 

210 + 
340 



80 
83 

110 
148 



150 

198 
181 
138 
100 



167 

103 

150 

188 



'[See footnote 1 on opposite page. — Editor.] 



THE MAMMONTIN^.: PARELEPHAS 1049 

Only an api)roximate phylogenetic order of succession may be given at the present time. In this phylogenetic 
list we include with some doubt the imperfectly known Elephas jacksoni Mather, also the recently described 
Elephns washingtonii Osborn. In phylogenetic order it would seem that this Parelephas phylum first appears in 
Elephas trogontherioides of the Upper Pliocene (Villafranchian)' of Italy. Doubtless there were many intermediate 
forms or ascending mutations between this Pliocene stage and the Mid-Pleistocene Elephas trogontherii of Pohlig 
at the close of the 2d Interglacial period when this animal was quite abundant. The type of E. trogontherii of 
Germany is in fact in about the same stage of evolution of the third upper molar as the type molar of the E. 
armeniaais of Asia Minor. 

The phylum is next represented in the 3rd Interglacial in the plateau loess of southern France in the form of 
Elephas intermedins of Jourdan. Deperet and Mayet consider the type of Elephas intermedins as more progressive 
than the type of E. trogontherii. Pohlig also states that a form related to E. trogontherii survived into 3rd Inter- 
glacial times in Germany and was then replaced by Elephas primigenius; whereas Deperet and Mayet consider 
that the Elephas trogontherii-intermedius phylum of Europe became extinct before 3d Interglacial time. This 
3d Interglacial period may possibly mark (Fig. 795) its time of migration across Asia into North America.^ 

Priority of Deperet in separating the E. trogontherii Phylum.— To Charles Deperet and his colleague 
Lucien Mayet belongs the credit of clearly distinguishing the origin and geographic distribution of the "Rameau 
de V Elephas trogo7itherii," to which phylum Osborn has assigned the generic name Parelephas. Deperet and Mayet 
(1923, pp. 176-183) first distinguished the "A. Rameau de VElephas trogontherii Pohlig" from "B. Rameau de 
VE. primigenius Blumenbach," both of which they included within the "Groupe des Elephas trogontherii et E. 
primigenius (Mammouths)." They credit Jourdan with the recognition of E. intermedins as distinguishable from 
E. primigenius, followed by Gaudry (1876) and by Leith Adams (1877) who traced this form back to the Forest 
Bed of England, then by Pohlig (1889 [1888]) who described E. trogontherii of Sussenborn near Weimar, but which 
Deperet and Mayet remark is absolutely identical with the E. intermedius of Jourdan. Finally they trace the 
"Rameau de VElephas trogontherii" down into the Pliocene superieur (Villafranchian of Italy), confirming a note 
by Pohlig (1889 [1888], p. 208) as to the tooth subsequently made by Zuffardi (1913, p. 174, Tav. xii [vi], figs. 3a, 36) 
the type of Elephas primigenius var. trogontherii. Deperet and Mayet observe (p. 177) that the molars of this 
E. trogontherii var. are readily distinguished from those of the contemporary E. [Archidiskodon] meridioncdis 
Nesti and E. [Hesperoloxodon] ausonius Major-Deperet. Zuffardi also described and figured another molar under 
the name oi Elephas antiquus var. trogontherioides (Zuffardi, Tav. ix [iii], fig. 6), and two other molars of the 
E. trogontherii form (PI. ix [in], figs. 4 and 5a, 56). They state (p. 179) that these four molars of E. trogontherii 
described from the "Villafranchien de I'Astesan" are the only ones thus far observed in Italian Museums, although 
there are probably others, and that they are clearly distinguishable from 'E. meridionalis' and 'E. antiquus, ' 
concluding (p. 180) as follows: 

On peut done dire seulement que la forme pliocene de VE. trogontherii possede des molaires de dimensions plutot un peu 
rcduites, sans qu'il puisse etre question d'une veritable mutation de petite taille de I'espece. On doit s'attendre a decouvrir un 
jour dans le Pliocene ancien de quelque region eloignee les types ancestraux nains du rameau de l'^^. trogontherii. 



'[There is a tendency among many geologists (see Chap. XXII of this Memoir by Dr. Edwin H. Colbert, where the views of Haug, Hopwood, Pinkley, et al are 
cited) to regard the Villafranchian as of Lowest Pleistocene (or Quaternary) rather than of Upper Pliocene age. Apparently one of the most reliable factors 
in the determination of the boundaries between the Pliocene and the Pleistocene, aside from the geological evidence, is a change in fauna, for example, the 
presence of Bos (including Bison), Elephas, or Equus, would indicate the Pleistocene age of the deposit in which they were found. Compare also Haug, 1911.1, 
pp. 1701 and 1767, Hopwood, 1935.2, pp. 46 and 47.— Editor.] 

^[See page 1071 below where Professor Osborn expresses the opinion that migration to America may have taken place during Sd Interglacial, and possibly 
as early as 1st Interglacial time. — Editor.) 



1050 



OSBORN: THE PROBOSCIDEA 



They also recognize (p. 180) the "Rameaii de VElephas trogontherii" in the Lower Pleistocene Forest Bed of 
Cromer, but find no trace of it in the Pliocene superieur Crags of England. 

From the Forest Bed they trace (p. 181) the phylum E. trogontherii, with gradual increase in size, through 
the terraces of 30 meters (Chellean or Tyrrhenian stage) in the valley of the Thames, also at Gray's and at Ilford 
where Lydekker (1886.2, pp. 127, 134) mistook E. trogontherii for E. antiquus. 

In Germany observations fail to indicate the Pliocene age of E. trogontherii, but this species appears in a 1st 
Interglacinl stage corresponding to the Forest Bed (Sicilien, Lower Pleistocene), also according to Wiist (1901, p. 
61) in the ancient gravels of the Ilm at Sussenborn near Weimar, where it was very abundant. This level they 
regard as contemporaneous with the 1st Interglacial sands of Mosbach near Wiesbaden and of Mauer near Heidel- 
berg (Chellean). They also state (p. 182) that Pohlig records E. trogontherii from SUssenborn, Heinthurm and 
Denstedt, near Weimar; Brucksdorf near Halle; Suiza, Angelhausen, near Armstadt; Apolda, Mosbach, near 
Wiesbaden; the lower gravels of Taubach, and of Rixdorf near Berlin; the gravels of the terrace of 30 meters at 
Mauer near Heidelberg, and the travertines of Taubach and of Weimar. It became extinct, without extending 
liigher, in the Quaternary. 

Osborn, 1929: Observe the constantly ascending ridge formula in M 3 ^il (in the Upper Pliocene) to M 3 M 
(in the Upper Pleistocene), a period of perhaps 1,250,000 years. Observe also that during 3d Interglacial times 
Parelephas disappears in western Europe and in early Pleistocene times it appears in North America. 




r. PH!MIGCNIUS Atf 
P&'.-B. I89I p 30*. Fifl t20 



PRIMIGENIUS Rtr 
falo , 1847. Pi, XLV, F.g XXIViftv." 



E PRIMIGENIUS Rel 
Fi'c. IW, Pi. Xtlll, Ftg. XXIV 



E. PRIMIGENIUS TROGONTHERI 
Polil'9. 1691. p 368. Fig 121 

Fig. 934. Cranial Profiles of Mammonteus primioenius (1,2,5) .\nd Parelephas troqontherii (3,4,0) of Europe 
One-twentieth natural size. The plate lettering is incorrect (compare Fig. Sfi."), ii[)i)pr row, ami koy, p. 977) 

3. Parelephas trogontherii (female); front view of cranium. After Pohlig, 1891. 

4. Parelephas trogontherii (female); side view of same eraiiium. After Pohlig, 1891. 
G. Parelephas trogontherii; adult male cranium in .side view. After Falconer, 1847. 
i). Mammonteus primigenius; adult cranium. After Pohlig, 1891. 

1. Mammonteus primigenius; front view of cranium (after Pohlig, 1891), wliiili resembles Parelephas in certain characters. 

2. Mammonteus primigenius, front view (if cranium (after Falconer, 1847), which apjjears to be tyi)ical of 'E. primigenius.' 

Observe that the Parelephas trogontherii cranium, both in the male (6) and in the female (3, 4), is relatively lower, broader, and less compressed fore and aft 
than the Mammonteus primigenius male {'i); in other words, it is less hy[).sicephali<' and less bathycephalie, with broader space between condyle and orbit. 
Compare esjiecially (o) M . primigenius and ((1) /'. trogontherii, both males. 



THE MAMMONTIN^.: PARELEPHAS 



1051 



Distinctive Cranial Characters of Parelephas 
Up to the time that Pohlig separated the grinding teeth of Parelephas trogontheru from tlie grinding teeth of 
Mammonteus primigenius the crania in these two distinct lines of generic descent were confused, as well as the 
grinding teeth, yet, as shown in figures 865, 934, 935, and 937, the crania of Parelephas are readily distinguishable 
from those of Mammonteus. 

The principal characters by which we distinguish the Parelephas crania are the following: (1) In frontal 
aspect the crania of Parelephas are relatively broader, more spreading, and more brae hy cephalic than those of 
Mammonteus, which are deeper and more bathycephalic ; (2) in lateral aspect (a) the orbit is more widely separated 
from the occipital condyle, (b) the occiput is much more convex, thus throwing the occipitoparietal apex farther 
forward, (c) the height from the occipital apex to the superior molar crowns is less deep, i.e., less bathycephalic, 
(d) the apex formed at the summit of the cranium is less acute, (e) the facial front is shorter and more deeply 
concave, (f) the maxillo-premaxillary sockets are less vertical and the tusks emerge in a less vertical plane; (3) in 
frontal aspect (a) the premaxillary sockets are more expanded at the extremities, whereas in Mammonteus they 
are more elongate and more closely compressed, (b) the transverse diameter of the frontals is relatively broader 
than in Mammonteus, (c) the anterior nares are proportionately broader transversely and less deepened vertically ; 
(4) in brief, the jirojiortions of the cranium of Parelephas throughout are harmonious with those of the grinding 
teeth, i.e., less compressed anteroposteriorly, less bathycephalic and less hyp.sicephalic than those of Mammonteus . 

Profiles (Fig. 935) of juvenile, adult, and middle-aged crania of Parelephas reveal a contour which is readily 
distinguishable from that of the larger Archidiskodon cranium, on the one hand, and that of the smaller and much 
more compressed Mammonteus cranium, on the other; but it is not at all surprising that Falconer confused the 
Parelephas and Mammonteus crania, because they present so many points of subfamily (i.e., Mammontinae) 




Left Lateral Profiles of Seven Species of Parelephas with Prooressive Ridge Formula 

One-thirtieth natural size 

Fig. 93.5. This series is remarkably uniform in its progressive characters: 

1) M 3 x^ to M 3 l-f ; superior ridge-plates exceeding the inferior ridge-plates in number. 

2) The orbits lie just below the level of the occipital condyles, with relatively wide separation. 

3) The occipitofrontal profile is restored in P. inlermedius (C), P. columbi (E), and /'. floridanus (F), but is perfectly preserved in P. trogonlherii (A, B), 
female and male of Mid-Pleistocene time, al.so in P. washinglonii (D), P. jeffersonii (G), aged type, and P.jeffersonii (H), giant male in the Nebraska Museum. 

4) All the profiles present a concave forehead, a moderately elevated summit, and a rather angulate occipital contour. 

5) As shown by comparison with figure 834, the hypsicephaly, acrocephaly, and bathyeephaly of the cranium are much less acute than in Mammonteus. 



ior)2 



OSBORN: THE PROBOSCIDEA 



resemblance. The relation (Figs. 806, 816) of the basis cranii to the fronto-occipital contours reveals a very 
jirofound difference between the fully adult Parelephas jeffersonii (Nat. Mus. 10261) and the adult Elephas 
iitdicus profile, on the one hand, and the Loxodonta africana profile, on the other. In all stages Parelephas agrees 
both with Arc/iidiskodon and Mammonteus in the concave frontal plane or forehead. 




PARELEPHAS TROGONTHERIl 4500mni..l4'9/a" e 

GERMANY MOSBACH 



PARELEPHAS FLORIDANUS 272] mm.,\2'2>/t" e 

FLORIDA 




PARELEPHAS TROGONTHERII 4054mm..l3'3>^" e 

GERMANY TAUBACH 




PARELEPHAS JEFFERSONII 

INDIANA 



3402mm.,ll'lV 





ELEPHAS INDICU3 BENOALENSIS 
INDIA 



3200min„10'6" 



PARELEPHAS COLUMBl 3b4omm.,l I'l l/e" 

GEORGIA 



Fig. 936. Parelephas of Europe and America in comparison with Elephas indicus benoalensis 
Rostoration.s to a one-fiftieth scale by Margret Flin.scli Bul)a, under tlic diieetion of Henry Fairfield Osborn. 



The crania conform in size with the three skeletons more or less fully known, namely, of Parelephas inter- 
medhis of France (Fig. 944), .said to attain a skeletal height of 3750 nun., or 12 ft. 3^^ in., of P. jeffersonii in the 
American Museum (Fig. 966), with an estimated height of 3200 nun., or 10 ft. 6 in., of the giant P. jeffersonii of 
the Nebraska Mu.seum (Neb. Mus. 1-4-15), the skeleton of which is unknown, and of P. columhi of the Amherst 
Museum (Fig. 955), with an estimated height of 3430 imu., or 1 1 ft. 3 in., finally of the giant P.Jloridanus, known 



THE MAMMONTIN^: PARELEPHAS 



1053 



from an associated humerus and femur in the American Museum to attain an estimated height of 3584 mm., or 
11 ft. 9 in., probably influenced by the favorable climatic conditions of Florida in contrast to the more severe 
northern conditions which produced P. jeffersonii. For estimated height in the flesh, compare figure 936 on 
opposite page. 



A. IMPERATOR Kef. 
/.OS AngeZes Mu-s. 



M. PRIMiqENIUS Fief. 
Var. /^u5 SS30 



P. JEFFERSONII 7")/oe 
Amsr. Mu:,. 9 9 SO 



.- WAaHIN^TONII 
Araer. /Vtts. 66 J 




^20 natural sijS 



Fig. 937. Cranial Profiles of the Mammontin.« of America 

One-twentieth natural size 

(A) Archidiskodon imperator, young male in the Los Angeles Museum. 

(B) Mamnwnieus primigenius, adult male, dwarfed, in the U. S. National Museum. 

(C) Parelephas jeffersonii, type aged male in the American Museum of Natural History. 

(D) Parelephns washingionii, referred young adult male in the American Museum of Natural History, the least progressive of the series. 

Observe extreme progressive hypsicephaly, cyrtocephaly, and bathycephaly as we pass from right (D) to left (B, A); thereby the occipital crest 
is elevated, the grinders are depressed, the orbits are approximated to the condyles (B), the face is flattened or concave. 





LeCTOTYPE and CoTYPES of PaRELEPHAS TROQONTHERIOIDBS 

About two-fifths natural size 

Fig. 938. Lpctotype and cotypes of Elephas anliquus Falc. var. trogoniherioides Zuffardi, 1913, Tav. ix (iii), figs. 3a-G6, namoly: Lectotype from Nizza 

dolla Paglia (Astiisan), third left superior molar, l.M' (Ha, 6'6), with 14+ ridgc-platcs; cotypes from near Sail Paohi de Villafranea, third riglit inferior molar, 

r.M3(4), with +14+ ridKe-i)latcs, third left inferior molar, I.M3 (50,56), withWH '■; ridge-plates and talon, and. seeond left inferior molar, I.M2 (3(i, 36), withal least 

10 ridge-plates. Comjjare Zuffardi, op. cil., pp. 130, 155-158, also Dep6ret and Mayct, 1923, pp. 177 170. These figures arc designated by the author as follows: 

"|I'ig.] 3a. — Molare probabilmonte quinto inferiorc sinistro (G), — pag. 155 [35]. 
" 36. — Lo stes.so visto dal fianco osterno. 

" 4 . — Molare ultimo inferiore destro ('8'), — pag. 156 [36). [Mus. Palais Carignan, Turin.] 
" 00. — Molare ultimo inferiorc sinistro ('9'), — pag. 156 [36|. 
" 56. — Lo stesso visto dal fianco intcrno. 

6a.- — Molare ultimo supcriore sinistro ('5'), — p.ag. 1.58 [381. [Geol. Mus. Turin] 

06. — Lo stes.so visto dal fianco csterno. [Geol. Mus. Turin.]" 



1054 



THE MAMMONTIN^: PARELEPHAS 



1055 



2. SYSTEMATIC DESCRIPTION OF EUROPEAN AND ASIATIC SPECIES IN ASCENDING 

PROGRESSIVE ORDER 



Parelephas trogontherioides Zuffardi, 1913 
Figure 938 

Type: Upper Pliocene (Villafranchian stage),' Piedmont, northern Italy, 
Nizza della Paglia and San Paolo de Villafranoa (Astesan). Referred: Red 
and Norwich Crags of East Anglia. 

Third superior and inferior molars, M', M3, described as 
a mutation of small size of Elephas trogontherii; grinding teeth of 
reduced dimensions; ridge-plates of similar character to E. 
trogonlherii; laminar frequence 6 to 6.5 in 10 cm.; total esti- 
mated ridge-plates in M 3 *tili*? as preserved in the lectotype 
(Fig. 938, 6a, 66), in the cotype M 3 ^^tttzjs (Fig. 938,4), in the 
cotype M 3 ^^rnr^j (Fig. 938, 5a, 56) ; estimated ridge-plate formula 
M 3 virTT^fl somewhat inferior to that of E. trogontherii, as we 
should anticipate. We may depend upon the observations of 
Falconer, Pohlig, Zuffardi, and Deperet as to the resemblance of 
these cotypes to E. [= Parelephas] trogontherii, which, if cor- 
roborated by future disco^'ery, make this the most primitive 
species. 

As to size, Osborn observes that whereas the typical molars of 
Parelephas trogontherioides e.xhibit fewer ridge-plates (M 3 qij^), 
they are about equal in size (length, ^P, 2104-mm., breadth 110 
mm., height 150 mm.) to Pohlig's type (Fig. 939) of P. trogon- 
therii (M*, length 225 mm., breadth 101 mm., height 138 mm.), 
also superior in size to the smallest AF of P. trogontherii from 
Siissenborn near Weimar (length 213 mm., breadth 74 mm., 
height 118 mm.); consequently P. trogontherioides cannot be 
described as a "mutation de petite taille"; it is, however, inferior 
in size to one of the largest superior molars of Siissenborn, namely, 
length 317 mm., breadth 76 mm., height 183 mm. (see Soergel, 

1912, PI. viii). 

E. [Elephas] antiquus var. trogontherioides Zuffardi, 1913. 
"Elefanti Fossili del Piemonte." Palaeont.-Italica, XIX, pp. 
130, 155-158. Lectotype and Cotypes. — Lectotype, third 

left superior molar, l.M' (Fig. 938, 6a, 66) ; cotypes, third right 
inferior molar, r.Mj (4), third left inferior molar, I.M3 (5a, 56), 
and second left inferior molar, I.M2 (3a, 36). Hokizon and 

Locality. — Upper Pliocene (Villafranchian stage),' Nizza della 
Paglia and San Paolo de Villaf ranca (Astesan) , Piedmont, northern 
Italy. Lectotype and Cotype Figures. — Zuffardi, op. cit., 

1913, Tav. IX (ill), figs. 3a-66. 

Description. — (Zuffardi, 1913, p. 130): "Pero pensando che 
il Jourdan considerava la sua specie E. intermedius come specie 
di passaggio dall' E. antiquus all' E. primigenius, per evitare 
dannose confusioni, proporrei ora il nome di E. antiquus var. 
trogontherioides alia varietii in questione, in ricordo del nome 
specifico dapprima dato al materiale attualmente ad essa attribuito. 
... [p. 155] Ricordo prima un dente [cotype I.M2] (Tav. ix [iii], 
fig. 3a, 36) che si trovava in raccolta senza indicazione alcuna e che 
distinguo colla lettera G. La corona larga con fianchi non molto 
convessi, termina anteriormente con una faccia ristretta e arroton- 
data, e posteriormente e tronca. La superficie di abrasione e 



concava e di forma ovale. . . . Lunghezza della corona m. 0,156. 
T>arghezza massima della corona, m. 0,105. Altezza massima della 
corona m. 0,094. Indice dentale 0,015. Frequenza laminare 6^. 
Simbolo [ridge-plate formula] a- 10 — ." Zuffardi states on page 
156 that the two third inferior molars [cotyi)es] (Tav. ix, figs. 

4, 5a, 56) were [erroneously] referred to Elephas antiquus 
Falconer and Cautley, from the Pliocene lacustrine of Astigiana 

5. Paolo, according to the label attached to the specimens in the 
Turin Museum, numbered 8 and 9 respectively. No. 5 [lectotype] 
of the collection (Tav. ix, figs. 6a, 66), also [erroneously] referred 
to Elephas antiquus, is from Nizza della Paglia (cf. p. 158); 
length of crown 210 mm., maximum breadth of crown 110 mm., 
maximum height of crown 150 mm., laminar frequence 6M to 7, 
ridge-plate formula — 15 x. Deperet and Mayet (1923, p. 178) 
estimate the total ridge-plates of this type specimen at about 20. 
Zutt'ardi's full type description is as follows: 

(Zuffardi, 1913, p. 158) "La stessa discussione puo farsi in 
riguardo di un altro grosso molare superiore (Tav. ix [iii], fig. 
6a, 66). Esso porta I'indicazione '5' Euelephas antiquus Cautl. 
et Falc. Nizza della Paglia — Dono del sig. cav. Pio Corso Bosen- 
asco' ed e accompagnato da un biglietto cosi concepito 'n.'' 20 — 
Rinvenuto nel 1819 in un podere della famiglia Corso Bosenasco 
dal sig. Pio Corso Bosenasco Sindaco di Nizza Lapaglia, a Nizza 
Lapaglia.' fi un dente alto e massiccio, troncato davanti, arro- 
tondato e assottigliato in dietro. I fianchi della corona, ancora 
rivestiti dal cemento, sono verticalmente quasi diritti, e longitu- 
dinalmente convesso I'interno, concavo I'esterno. La superficie 
di abrasione, quasi rettangolare-arrotondata, e composta di dodici 
lamine, tortuose, larghe come gli interspazi di cemento e tendono 
ad allargarsi bruscamente nel mezzo in modo da originare una 
espansione ovolare. Le coma piegano piii meno sentitamente 
air indietro, le lamelle sono leggermente pachiganali e fittamente 
increspate. Della 1.* lamina e presente solo la meta interna della 
lamella di smalto posteriore, e della 2." manca la meta esterna della 
lamella anteriore. Sino all'8.^ compresa, le lamine hanno figura 
completa; la 9", 10", 11." hanno tre elementi laminaroidi di cui il 
mediano piu sviluppato, e assai meno I'estremo interno ridotto 
quasi a un semplice anulo. La 12." e composta di quattro cercini 
appena usati. Segue I'altra parte della superficie di abrasione non 
usata che forma con la precedente un angolo di poco superiore al 
retto. Essa quantunque ricoperta ancora da cemento, dalle costo- 
lature lateral! puo ritenersi constituita da cinque lamine delle quali 
I'ultima, strettissima e arrotondata, puo considerarsi come tallone 
prossimale." 

"Le radici, per la massima parte perdute, si iniziano in un 
brusco restringimento laterale della corona e vi si vedono i tronconi 
basali di un paio di robusti rami anteriori probabilmente liberi, e i 
resti dell'ammasso delle altre radici. Dall'andamento di queste e 
dalla forma della corona sembra manchino al massimo altre due 
amine." 



'[See footnote 1 on p. 1049 above regarding the possible Lowest Pleistocene age of the Villafranchian. — Editor.) 



1056 



OSBORN: THE PROBOSCIDEA 



"Valori metrici: earth colored by fire ("rotgebrannte Erde"), also bones belonging 

^ , , ,, „r.,r> to an ancient canipina; place of priniiti\e man which inieht well 

Lunehezza della corona m. 0,210 , , , , , ,, , , ,, e ,i i • 

, . , 11 i< ,> ,,,A have beioneed to Uunui fiei<ltlberqeii.sis; oi the same geologic age 

Larghozza niassima della corona 0,110 .. /, i^ * r, j u- u • -^ ii ^u j -^ r 

. , -,11 ,< ,> , r,. iis the Cromer I'ore.st Bed which is quite as old as the deposits of 

Altezza massima della corona 0,150 , , ■ /t^ i i M\f\n\ t-.u ■ ^ u ^ • ^• 

^ ,. , , , ,1 ,^,^-,r Jockgrim (1* reudenberg, 1909). 1 he very ancient characteristics 

Indice dentale 0,015 , ^/! u c < <j. •• . .■ / c ^u 

. . „,^ „ of this small E. Irogoidneni mutation at once eonhrm the great 

' IK" antiquity of this basal Mosbach sands horizon, which also contains 

the Etruscan rhinoceros (Dicerorh-inus eiruscus), Trogontheriuin 
"fi d'uopo notare pero che il Falconer ([40), vol. II, pag. cuvieri, and Hippopotamus. In this lower Mosbach stage of this 
249-250) si riferiva for.se anche a questo dente quando parlando small E. trogontherii mutation, the third inferior molar, Ms, ex- 
del suo E. nrmeiiiacus, che credeva ravvisare in un grosso molare hibits 8 ridge-plates in 10 cm., e.\actly as in the primitive Irogoii- 
segnato col n.° 7 nella nostra raccolta, e che ci occupera ])iu avanti Iherii molars of Jockgrim, whereas the true Elephos primigeiiiiis M3 
(pag. 174 [54]), affermava che altri ve n'erano tra i molari di S. exhibits 7 ridge-plates in 10 cm. These basal sands of Mosbach 
I'.aolo." cannot be more recent than the 1st Interglacial (Giinz-Mindel) 

„ T /i~v ^ AT ,rvr>n , -/. Stage of thc Alps, or the Elephant Bed of Cromer (Forest Bed 

Pliocene of Italy (Deperet and Mayet, 1923, pp. 1<6- . , „^.,, . ^ V, , , . , \ t,, ■> < 

,„„, . ,. , ,. , /,D/.Dx Ti 11- /,oo,A r,oooiN I SBrics). Still uiorc aucicnt arc the Underlying laycrs at Mosbach 

183).— According to lalconer (1868), Pohhg (1889 1888), and , . . „, , , ^ .•.•,. 1 ■, , , tt 

r, «- ,■ /,r^,oN .. ■ r. I L . .. • • I I 1 r- containing tlephas [ = AirnidisKodon\ mcndionahs and an Upper 

Zuffardi (1913), this rarelephas trogontnerioides branch i.e., „,. . .,, .,,,,/ , • , !, 

, , , , , , c , • il TT ni- /TT-ii 1 liocene launa with Mastodon \ = Aiiancus\ nrrernensis and the 

I)hyluiii] although rare hrst appears in the L pper Pliocene (Villa- „ , ,, 

fninchian) of Italv. As discovered near .San Paolo de Villafranca ,, . ,„„'o' n^, , ., i t-v ^i-il 1 t^ j i 

, . „ , ,," . ., ., , 1 ,, ,, , , ,,, Osborn, 1928: Ihe above note by Dr. Wilhelm I'reudenberg 

this lectotypej M-" exhibits so closely all the characters of the cor- . , 1 . , ,. . i- . ^ ^^ ..-,,/• 

1- / ^L r n. L . .L -.i . i 11 . IS deeply interesting as tending to eonhrm the great antiquity of 

responding tooth ot A/emflas <7'ooon</!e?» that it can only be separat- ,1 r. , , ,1 ■ , r^ ■ ,, ..1 

,' . ,, , ,,n^ox .• 1 r. Ti 1 \. 1 the Parelepnas phylum in western Europe and the po.ssible co- 

ed as a variety, lalconer (1808) mentioned ban Paolo as one of the ., , T .,. . , r. > l /o . „ ■ j \ 

,...,,, ^ . . , ,, 1,. ,1 , , r, existence of a primitive species of f'amepna.s ( i^sp. <rono«//(e)7o/ae.s) 

localities m Italy containing teeth resembling the type 01 E. .,1 „ 1 n /w 1 ■ 1 .1 ■ . 

.,.,,. ILL 1- 1 f 1 . 1 with Heidelberg man (Homo heidetberqensts). 

armemacus of Asia Minor, which he accordingly referred to the 

Armenian species (seep. 1060 below). ■■ r,,,- ,oor 1000 lom 

,, ,,.' ,\, T, ,,,. ,,,,,, c. Parelephas trogontherii Pohlig, 1885, 1888-1891 

ror this and other I pper Pliocene teeth found at or near San 
ni ij 11 ic lutuu'/o-j- luT^-.. Figures 794, 865, 871, 929, 934-936, 939, 940, PI. XXII 

Paolo and described and hgured both by Zurfardi and by Deperet 

, ^, , ,1 I -n m L J- i .1 ■ Type: 2d Interglacial deposits at Siissenborn near Weimar (Fig. 932, 

and Mavet, the subspecihc name A/ep/io.s a?i^(/«MS var. //ooonZ/jeri- ,,, ' , ^ „, j ^ . ^, , , , •, t> . ,• /Toot <> 

' , , 1 1 11)> northern Germany. Referred: Isl-M Interglacial. Pohhg (1887.2, 

oides was given, the characters being summed up as follows ,,.274). "vom mir zuerst in den alten Flusskiesen ThUringens zu Sii.s.senborn 

(Deperet and Mayet, 1923, p. 179): "Ces molaires presentent tous etc. bei Weimar in grosserer Menge aufgefundene altdiluviale Ele|)hantpn- 

les caracteres essentiels de VE. trogontherii quaternaire, savoir: zahnform beschrieben." 

une couronne tres large et tres haute, avec des lames ^paisses et This very important species, widely distributed in 2d Inter- 

a.ssez ecartees (frequence laminaire de 6 a 6, 5) ; un email tortueux glacial times in western Europe, is, with the more progressive 

et fortement plisse, orne sur sa longueur de quelques sinus aigus Elephas intermedius of Jourdan, a member of the distinct phyletic 

irreguliers, mais ne ressemblant en rien au sinus loxodonte median series Parelephas. Parelephas trogontherii belongs to a phylum by 

des E. meridionalis et antiquus. ... [p. 180] On pent done dire itself which appears sparsely in Europe at the close of Pliocene 

seulement que la forme pliocene de VE. trogontherii possede des time and reappears in the /.s/ and .^rf/Htec^/arm/ periods. We do not 

molaires de dimensions plutot un peu n'duites, sans qu'il puisse agree with Pohlig's opinion that the species i?. ^ro^ow/Zier// is a con- 

etre question d'une veritable mutation de petite taille de I'espece. necting ancestral link between Elephas meridionalis and E. primi- 

t)n doit s'attendre a d^couvrir un jour dans le Pliocene ancien de genius. With Deperet we regard it rather as a distinct phylum, 

(juelque region eloignee les types ancestraux nains du rameau de The estimated ridge-plate count of the type, M^ (Fig. 939, upper) 

VE. trogontherii." is {-;, somewhat exceeding the ridge-plate count of E. [P.] trogon- 

FiRST Interglacial of Germany (Freudenberg, 1926). — //ienoirfe.s Zuffardi. 
In the basal layers of the Mosbach sands, near Wiesbaden (lug. Soergel (1912.2, Tab. vii, viii) records three inferior and 

932, 12), Freudenberg records small mutations ("kleinere Vari- five superior molars from the type locality of Sussenborn {1st to 

ante") which he refers to Elephas trogontherii, associated with 2d Interglacial deposits), in the liebling and Weimar Museum 

artificially broken off ("kiinstlich abgetrennte") portions of the collections, which vary widely in length, breadth, and height, as 

horns of Cervus verticornis, charred wood ("Holzkohlen"), and follows: 

Sussenborn: Reeling and Weimar Mus. Ridge formula Length Breadth Height 

R.M3 No. 46 (Weimar) xl8x(xl7.x) 270 62 160 

17 (Weimar) 16 x 255 83 105 

10 (Weimar) x21x 316 73 137 

R.M» No. 80 (Rebling) xl8x 302 91 192 

69 (Hebhng) 20 x 291 88 160 

32 (Weimar) x20x 296 66 137 

89 (RebUng) xl8x 264 71 162 

87 (Rebling) x20x 274 85 171 



THE MAMMONTIN/E: PARELEPHAS 



1057 



Elephas lnty<intherii Pohlig, 1885. "Ueber eine Hipparionen- 
Fauna von Maragha in Nordpersien, iiber fossile Elephantenreste 
Kaukasiens und Persiens und iiber die Resultate einer Mono- 
graphie der fossilen Elephanten Deutschlands und Italiens." 
Zeitschr. deutseh. gaol. Ges., 1885, XXXVII, Heft IV, p. 
1027. Type. — Apparently a last superior molar, M', of the 

right side, also a last inferior molar, M3, of the same side. Hori- 
zon AND Locality. — 2d Inlerglacial deposits, SiJssenborn near 
Weimar, northern Germany. Type Figure. — Pohlig, 1888, 

p. 193, fig. 79, and p. 195, fig. 82 (Fig. 939 of the present Memoir). 
Schwabe Coll., Weimar. Referred Specimens. — The Siis- 

senborn stage of the upper and lower grinding teeth of Elephaa 
trogontherii, illustrated by Wiist, 1901, Taf. 11, figs. 4-12 (Fig. 940), 
appears in comparison with the stage erroneously referred to 
'Elephas' [Hesperoloxodon] antiquus in figures 3 and 18. 

RELATIONSHIPS OF ELEPHAS TROGONTHERII POHLIG (1885, 
1887, 1888-1891, 1907, 1912) AND SOERGEL (1921) 

Relationships.— (Pohlig, 1885, p. 1027): "6. Unter der 
Bezeichnung 'Elephas trogontherii Pohl.' fiihre ich in meiner 
Monographie eine europiiische Molarenform auf, welche zwischen 
denjenigen des E. primigenius und E. meridionalis zoologisch, 
wie ihrer geologischen Lagerstiitte naeh, in der Mitte steht, dem 
E. antiquus in der Lamellenformel am nachsten kommt, aber durch 
die Kronenform und geologische Lagerstatte von dieser Art weit 
scharfer gesondert ist als von den anderen beiden Species. Das 
Verhaltniss von E. trogontherii zu E. armeniacns Falc. und audi 
E. namadicus Falc.-Cautl. bleibt noch genauer zu untersuchen. — ■ 
E. meridionalis und E. primigenius stehen nach Kranologie und 
Dentition iiber E. trogontherii hin in director Verwandtschaft." 

Description (Pohlig, 1887.2, p. 274) : "Unter diesem Namen 
Elephas trogontherii habe ich in meiner Monographie eine, wie es 
scheint, in Europa allgemein verbreitete, von mir zuerst in den 
alten Flusskiesen Thiiringens zu Siissenbom etc. bei Weimar in 
grosserer Menge aufgefundene altdiluviale Elephantenzahnform 
beschrieben, welche geologisch sowohl als zoologisch die beiden 
Arten E. meridionalis Nesti aus dem Pliocaen und E. primigenius 
Blum, aus dem Plistocaen vollig mit einander verkettet: in der 
Form der Krone und Kaufliiche und in dem Charakter der ein- 
zelnen Schmelzfiguren auf letztercr stimmt jene altdiluviale 
Rasse mit den zuletzt genanntcn beiden Arten iiberein, steht 
jedoch, was die Anzahl der Schmelzlaniellen und die Dicke der 
Schmelzwande anbetrifft, zwischen diesen beiden Species in der 
Mitte, — in den unterplistocaenen Schichten mehr dem E. meri- 
dionalis, in den mittelplistocaenen Lagerstatten mehr dem Mam- 
muth in dieser Beziehung angenahert." 

Pohlig (1885-1887) from the first expressed the opinion that 
Elephas trogontherii forms a connecting ancestral link between 
Elephas meridionalis and Elephas primigenius, its earlier represen- 
tatives resembling the former, its more recent representatives re- 
sembling the latter, accordingly he wrote the specific name at the 
time E. primigenius trogontherii. He held that this species entered 
Europe during the second glaciation (Saxonian-Mindel) and sur- 
vived during the ,^r/ Inlerglacial period (Helvetian, Mindel-Riss) ; 
thus E. trogontherii was contemporary with the appearance of the 
Chellean industry in France (Penck). Pohlig (1888-1891, p. 458) 
pointed out that while Elephas trogontherii resembles E. antiquus 



in ridge formula, in the structure of the crowns of 
it is sharply separated, because the grinders of E. 
different in proportion from those of E. antiqu 



the molar teeth 

irogontherii are 

us; the enamel 




Fig. "3. Letzter Maxillenmolar von Elephas trogontherii. Original zu Weimar 

(Dr. Schwabe), '/a 




Fig. 82. Letzter Mandibelmolar von Elephas trogontherii. 

(Dr. Schwabe), '/,. 



Original zu Weimar 



Types or Parelephas trogontherii. Compare Soeroei,, 1912.2, Tab. vii 
Onr-third natural size 
Fig. 939. Type third suijorior and inferior niolar.s of Elephas trogontherii 
Pohlig, 1885. After Pohlig, 1888-1891, p. 193, fig. 79, and p. 195, fig. 82. 
These arc the specimens first mentioned by Pohlig (1885, p. 1027) and figured 
by him in his Memoir of 1888-1891: "Fig. 79. Letzter Maxillenmolar von 
Eleplias Irogonlhrrii. Original zu Weimar (Dr. Schwabe), ):)." This is apparent- 
ly a last superior molar of the right side, r.M'; it is eomijlcte and displays 15 
ridges. "Fig. 82. Letzter Mandibelmolar von Elfplms trngntdhrrii. Original 
zu Weimar (Dr. Schwabe), }:1." This is a last inferior molar of the right side, 
r.Ms; it is broken in front and displays 16+ ridges. The incomplete ridge 
formula of these types of Elephas trogontherii is: M ^\%X- See text below. 



1058 



OSBORN: THE PKOROSCIDEA 



bands rise above the cement not vertically but obliquely, after the 
manner of (he Asia<icelci)hant. PJ. /lof/o/f/An// hesi)caks !i rolatixo- 
ly moist climate but is also occasionally minj^lcd w itli remains of 
the reindeer (linngifrr) where (he forest bordered on the open 
country, as a((os(c(l by (he appearance of bo(h animals in Siisson- 
born andS(einhoim. Pohlig (18S8-1891, p. 380, fig. 121) figures the 
female skull as Elephas {primigenius) Irogontherii (reproduced in 
Fig. 865, Nos, .3, 4 above), in which the superoccipital region is 
imjierfect and restored. lV)hlig's opinion as to the (mnxitiomd 
character of this animal between E. mcridionalts and E. primi- 
genius is indicaied in (he names he has successively assigned to it, 
namely : 

Elephas Irogontherii I'ohlig, 1885, p. 1027. 

Elephas (primigenius bczw. mericlin)mlis) irogontherii, 1887, 
p. 274. 

Elephas (primigenius) irogontherii Pohlig, 1887 11888'J, i). 799. 

Soergel (1921.2, \). 68) in his comi)rehensive review eniided 
'Elephas rulumbi I'alconer" likewise adopts J'ohlig's opinion and 
places Elephas meridionalis-trogontherii-primigenius as successive 
jjhyletic stages in the same west iMiropean phylum, fie assigns 
(op. oil., J). 55) the following collective ridge formula to (he Elephas 
Irogontherii specimens of Sussenborn and Mosbach: 



Dp2i Dp 3^'- Dp4-^V Mllfffj M2H-:f= M3f 



16-2^1 



Soergel is of the opinion that Elephas trogoDtherii (western 
luu-ope) and the tyjjical Elephas eolumbi (southern U.S.A.) are in 
similar stages of e\olution. He points out that the typical Elephas 



eit., pp. 69-97) concurs with Pohlig in the opinion (hat Elephas 
irogontherii is ancestral to Elephas primigenius. 

Haditai'. — In a letter of ,lune 30, 1922, Soergel states: 
According to my interi)retation E. trogontherii lived during a period 
of limited rainfall, in a half-arid climate, not migral-ing during 
glacial times inio the colder ice-dad regions of middle lOurope 
(compare also Soergel: 'Die Ursachen der diluvialen Aufschot- 
terung und Erosion,' Verlag Porntrager, 1921). E. trogontherii 
therefore occupied the region la(er traversed by the mammoth. 
We still regard E. trogontherii as the direct ancestor of E. primi- 
genius. We possess in the museinns of I'^urope all the transition 
forms in large numbers and are unable to make any sharp dis- 
tinction between the two species. If one seeks a complete under- 
standing of the chief literature of value on this subject, it may be 
found in the following works: M. Pavlow, 'Les elephants fossiles 
de la Husse,' Nouv. M<'>m. d. 1. Soc. Imp. des Nat. de Moscou, 
'J\ XVII, 1910; W. Soergel, 'Elephas trogontherii vmd Elephas 
antiquus etc' Palieonlographica, 1912; 1']. Wiist, 'Das Plioziin 
und das iilteste Pleistoziin Thuringens. Abhand. d. nalurf. 
(iesellsch. zu Halle, Bd. XXIII, 1900. 1 rejoic'C that in tlie iirinci- 
pal points we agree as to the diluvial elephanls and their phylogeny. 

PHVLOciENKTK; ( "oNcusioN (OsBOHN, 1924). — Osborn does 
not agree wi(h either Pohlig or Soergel that the typical Elephas 
trogontherii is a descendant of A'. [Archidiskodon] meridional is; 
nor does he agree that it should be regarded as transitional or 
ancestral to the typical E. [Mammonleus] primigenius. The pro- 
found difference between the broad-plated molars of the Aixhi- 



Nr. 18 



Nr. 6 




Fig. 040. Hcfcrrcd molars (r.M.T with 18+ ridRc-pliitc's) of 'Kltphns Irogoittlirrii' Pohlig from tho 2<l I tilrrglacinl of Sii.sscnborii aii<l Woiinar, Oormany. 
After Wiist, 1901, Taf. ii. Wiist. orronroiisly refers r.M'' (Nr. 3) and r.M" (Nr. 18) to Elrphas antiquus; they are Parvlcphas trngonllicrii molars from the 
type locality (Siissenhorn near Weimar). 

"Nr. 3. Klephas antiquus Falc. M.III max. dextr. Siissenborn. . . . 
" 6. Desgl. M.III max. dextr. Siis.senborn. . . . 
" 18. Elcplias antiquus Fate. M.II max. dextr. Weimar." 



meridional is may hv regarded as the direet aneestor of the Elephas 
eolnmbi-imperator phylum: (Op. eit., 1921, p. 68) "Diese Tatsache 
lj<'reclitigt zu der .\nnahme, 'd;iss die nach .Vmerika iiberwandern- 
den I'ormen der .Meridionalis-trogontherii-llcihc schon beim 
Ueberwandcrn resp. kurz vorher besondere Charaktere gegeniiber 
dem europiiischen El. trogontherii meridionaJis ausgebildet batten, 
Charaktere, die in der weiteren ICntwicklung sich zum 'Imperaior- 
Typus' 8teiger(en. Es war also im iiltesten Diluviimi im Krcis der 
knntincnt;ileii l'',lefantenf()rmen eine ^'aria(i^Ilsbrcilc iiii( zwci 
Polen, El. trogontherii mrriilionalis im Wcsicii und ilem direktcn 
N'orfahren des El. imperalor im t)s(en vorhanden." Soergel (op. 



di.slcodon planifron.s-mcridionalis-imperalor phylum widely disl-in- 
guishes this phylum from Elephas trogontherii; Osborn accepts 
rather (he view of Depi'ret and Mayet (Iiat E. trogontherii repre- 
sents a distinct branch of the mamniolhs. !\Ioreo\'cr, the extreme 
hypsicephalic, acroccphalic, bathycephalic cranium of Elephas 
primigenius, with corresponding hypsodon(. finely compressed, 
and laminated molars, cannot be derived from (he cranium and 
molars of the Elephas Irogontherii type. Osborn agrees entirely 
\\i(h Dep(''r('( aii<l Maye( (ha( we ha\e (o do wi(li throe (lis(incl 
|)hyla, certainly separate from (he clo.sc of Pliocene (im(\ .\gaiii 
(Jsborn does not accept the ojiinion of Soergel that E. trogontherii 



'In the Bibliography of the present Memoir (p. 795), it is possible that 1887.1 should road 1888. 



THE MAMMONTIN^: PARELEPHAS 



10o9 



is related to the true E. { = Parelephas] colutnbi of the southern 
United States, although it may well prove to be related to tht; 
anunal described by Cope, Hay, Osborn, and others as "Ekphas 
cohtmbi" but which now proves to represent a distinct species, 
namely, the Parelephas jejfernonii of Osborn. Consequently, 
Osborn, as cited above, has placed E. Irogontherii in the distinct 
phylum Parelephas, the generic characters of which are defined bj' 
cranial and dental characters in the chapter on the "C^lassification 
of the Elephantoidea" (Chap. XV). 

Parelephas '?j trogontherii nestii I'uhlig, 1891 
Figure 911 

Typo loc-alitj': ".JunKplior'aoiicii," Forest Bed (Norfolk), Walton (Essex), 
and Soutliwold (Suffolk), England. Lower Pleistocene. 

Regarded by Ueporet and Mayet (192.'j, j). 103) as a synonj-in of EU/jlut.s 
[JlesperiiUjiodon] aumniua F. Major. Cotype.s regarded by C. W. .\ndrews as 
belonging to E. [= Parelephas] Irogontherii. 

Pohlig (op. cil., ]). 303) in reviewing I>citli .Vdams' Monograj)h 
on Elephas aniif/iiiis ob.served that certain teetii of the 'riianie.s 
Valley (Adams, 1877-1881, p. 33), of Norfolk (p. 40) and of South- 
wold (p. 38), ilo not agree with the continental molars [of < Icnnany] 
referred t(j E. aidiqims. Pohlig continues, that j)art of the "jung- 
l)liocaenen . . . I'ore.stbed stammenden" molars do belong to the 
typical E. aiiUquus, but that the majority l>elong to an inferior 



/j/Vaiu; 




race; he then separates this race as E. Nedii. I'ohlig's lypes or 
(•(jlj/pcs of E. Ne.slii are therefore molars of Forest Ji(!d age. 

CoTYPES. — As shown above, the subspecies Elephas anliquns 
Xeslii was erroneously based on cotype specimens selected from 
Leith .\danis' .Monograph of 1877 1881, pp. 37 and 38, namely, 
Prit. Mus. 33327, a dredged lower molar from Happisborough with 
+ 16 + ridge-plates in 11. o inches [catalogued by Lyrlekker, 1886.2, 
I). 128, as a "right third upper true molar"]; P.rit. .\lus. 27915 
[given as No. 27515 by I>eith Adams], a 14+ ridge-plated third left 
upper true molar from Walton (lOssex), with 13+ ridge-jjlates in 
8.5 inches, and Hrit. Mus. 39463, a half-worn third left lower true 
molar from Southwold (Suffolk), compare Lydekker's "Catalogue 
of the Fossil Mammalia in the Pritish Museum," 1886.2, pp. 128, 
129, and 135. .Ml these cotyjies are of Lower Pleistocene Cromeri- 
an or Forest Jled age; their generic and specific reference to 
'E. aidiquus' is doubtful; C. W. Andrews refers Prit. Mus. 27915 
and 39463 to 'E. tro(ji)tilherii; Ilopwood (letter, August 6, 1929) 
kindly furnished i)hotograi)hs and detailed measurements wliich 
tend to confirm ,\ndrews' reference to 'E. Irogonlherii.' 

Hai)i)isborough (Forest Ped), England. Prit. Mus. 33327, 
r..M', length 268 mm., breadth 92 mm., height 196 mm.; ridge- 
plates + 16+, 6.3 in 10 cm. 

Walton (Essex), England. Prit Mus. 27915, \.W, length 
302 mm., breadth 80 mm., height 167 mm.; ridge-plates 18, G'A 
in 10 cm. (referred to 'E. trogontherii' by C. W. .\ndrews). 

Southwold (Suffolk), England. Prit. Mus. 39463, I..M3, 
length 243+ mm., breadth 83 mm., height 103 uim., ridge-plates 




^ A/atu^roJ szje 



,2 „ 10.3 8 7 

j3 ^ j: /■■ f 



Bntfpus. 39463 




COTYPKH OF Ei.EPHAS AXTIQUUS nestii PoHMO now TR.\NSI-EUIIKD TO pAHEI-fil'HAsf?) TItOGONTHEKlI .N'KSTII 

I'ig. (Ml. Cotype.s or syntypes of A^cp/iu* art(t9«u« ,N>«(u Pohlig, 1891, after photographs kindly furnished by Dr. A. Tindell Ilopwood of the Uritish 
.Museum, August 6, 1929. From the Lower Pleistocene of eastern England. One-third natural size. 

(f>eft) Brit. Mus. 2791.J. Left third superior molar, I.M', from Walton (Essex), England, exliibiting I 18 ridge-plates; also crown view of same. 

(Right) Brit. Mus. 3946.3. Incomplete left third inferior molar, I.M3, from Southwold (.Suffolk), England, exhibiting 1-14 posterior ridge-plates; anterior 
ridge-plates broken off. Lateral and crown views of same. 



1060 



OSBORN: THE PROBOSCIDEA 



14+, 7H to 8 in 10 cm. (referred to 'E. trogontherii' by C. W. 
Andrews). 

Unfortunately these specimens were not figured by Lcith 
Adams or Pohlig, but they were described by Leith Adams as 
belonging to his 'Ekphas antiquus.' It is to be regretted that the 
name Elephas antiquus Nestii has been adopted and widely quoted 
in the literature, for we cannot agree with Deperet and Mayet 
(1923) that 'Elephas antiquus Nestii' Pohlig, erroneously based on 
Lower Pleistocene cotypes, is synonymous with 'Elephas auso7iius' 
Major-Depcret-Mayet, correctly based on an Upper Pliocene type 
of the Villafranchian stage.' 

History. — In 1891 Pohlig imperfectly established this sub- 
species (named after Filippo Nesti) as of Upper Pliocene age, 
whereas the cotypes selected by him, as mentioned by Leith Adams 
in his Monograph of 1877-1881, are said to be from Walton, South- 
wold, and Norfolk, the latter from the Cromer Forest Bed level 
which is of Lower Pleistocene age. This subspecies, founded on 
a misconception as to its Pliocene age, is invalid ; it is certainly not 
in the same geologic stage as the Upper Pliocene 'Elephas' ausonius 
Major-Deperet-Mayet, from the Villafranchian of Italy.' Although 
not figured, the ridge formula was said by Pohlig (1891, p. 305) to 
agree with that of E. antiquus: "In der Lamellenformel wiirde E. 
Nestii sowohl mit E. antiquus lypus, als auch mit E. trogontherii 
iibereinstimmen." 

E. Nestii Pohlig, 1891. "Dentition und Kranologie des 
Elephas antiquus Falc. mit Beitragen iiber Elephas primigenius 
Blum, und Elephas mcridioiialis Nesti. . . . Nachtriige," pp. 303, 
304, 465. On page 465 the name appears as Elephas antiquus 
Nestii n.f . Type.— (Letter, Pohlig, Sept. 10, 1924) : "As type 

specimens of E. Nestii from the forest bed are named in my mono- 
graph (p. 304) those of L. Adams' [monograph], p. 38, from Walton 
and Southwold, and the Nr. 33327 of the British Museum, from 
Norfolk." Locality and Horizon. — Forest Bed (Norfolk), 

Walton, and Southwold, England; Lower Pleistocene. Type 
FiciiRE. — Not figured. See figure 941 of the present Memoir after 
photographs. Referred skull of northern Italy figured by Pohlig, 
(ip. cit., 1891, p. 350 = ('raiiiTmi B of Nesti (Elephas vicridionalis); 
referred to Elephas antiquus by Weithofer 1890, and erroneously 
referred by Pohlig to his subspecies E. antiquus Nestii. 

Type Description.— (Pohlig, 1891, pi). 303 305): "b. 
Ein Theil der letzteren — die aus dem jungpliocaenen [Footnote: 
'Vgl. u. (sub E. meridionalis) .'] Forestbed stannnenden — lassen 
(lie Zugehorigkeit wcnigstens zu der Species E. antiquus zu, 
nianche mcigen selbst als typische Molaren derselben gelten; die 
meistcn haben jcdoch Abweichungcn von der typischen Gleich- 
miissigkeit der Kronenbreite niclit nur, sondern auch von dem 
charakteristisclicn Lamellenbavi der diluvialen Ausgangsform, be- 
sondcrs von dem ausgcpriigten Loxodontismus, der complicirteren 
I'estonirung und der grtisseren Dicke des Schmelzes Ix^i jener. Da 
nun diesc Eigenthiimlichkeiten zum Theil mit den im ersten Ab- 
schnitt erorterten, ebenfalls jungpliocaenen Erfunden von der 
S|jecies aus dem .\rnothal iibereinstimmen, so scheint es mir nun- 
mehr docli geboten, diesen pliocaenen E. antiquus von dem dilu- 
vialen durch cine Rassenbczei(;hnung, ctwa E. Nestii, vorltiufig zu 
(rennen. . . . Es mijsste danach die in dem ersten Abschnitt aus- 



gesprochene Vermuthung, dass Loxodon von unhckannten Masto- 
dontenformen sich ableite, fallen gelassen werden; denn ganz wie 
E. Nestii zum typischen E. antiquus verhiilt sich nach Mitgetheil- 
tem das Wenige, was man vom E. priscus weiss . . . , zum modernen 
E. africanus, — es zeigt den Charakter des letzteren minder markirt, 
mit Eigenthiimlichkeiten archidiskodonten Geprages, wenn auch 
in geringem Grade, vermengt. ... In der lamellenformel wiirde 
E. Nestii sowohl mit E. antiquus lypus, als auch mit E. trogontherii 
iibereinstimmen; und da an mehreren Punkten der englischen 
Ostkiiste, wie bei Mundesley-Cromer, C'lacton etc. interglaciale 
Schichten mit E. antiquus und E. primigenius — an erstercm sogar 
mit ganz iihnlichem Erhaltungszustand — die jungpliocaenen mit 
E. meridionalis, E. Nestii und E. trogontherii iiberlagern, also 
Molaren aller dieser Formen in die Brandung und von da in die 
Museen gelangen: so war die Schwierigkeit, alles das nunmehr 
Vermengte in der rechten Art wieder voneinander zu sondern, 
anfangs begreiflicher Weise selbst fiir den Geiibten zu gross. Und 
doch sind eben grade diese Verhaltnisse von grosstem Interessc." 

Osborn, 1930: We are therefore obliged to regard Elephas 
antiquus Nestii as comparable to the species 'Elephas' [ = Par- 
elephasi"!)] trogontherii Pohlig (cf. Parelephas trogontherioides 
Zuffardi). 



Parelephas armeniacus Falconer, 1857 
Figure 942 

Pleistocene, near Khanoos, Province of Erzcrum, Armenia. 

We note that Falconer in 1868 (Vol. II, pp. 187, 192, 193, 248, 
249, 250) observed a strong resemblance between his type of E. 
(Euelephas) armeniacus and Upper Pliocene molars from the 
locality of San Paolo, or near it, Nizza della Pagha, Italy, recently 
chosen by Zuffardi as the types of Elephas antiquus var. trogon- 
therioides. It is probable, therefore, that the species Elephas 
armeniacus of Falconer belongs in the generic phylum Parelephas, 
but we cannot be certain of this reference until the cranial charac- 
ters of the species become known. The specific distinctions are: 
Ridge-plates of M 3 "^^^-^ (Fig. 942) , worn anteriorly. This api)ears 
to be a more progressive stage than Parelephas trogontherii and 
much more progressive than P. trogontherioides. 

E. (Eueleph.) Armeniacus I'alconer, 1857. "( )n the S|)ecies <if 
Mastodon and Elephant occurring in the fossil state in Great 
Britain. Part I. Mastodon." Quart. Journ. Geol. Soc. London, 
Vol. XIII, Synoptical Table op]). p. 319. Type. — (Falconer, 

1863, p. 74): ". . . two last upper molars in fine preservation, and 
a portion of a lower, all apparently of the same individual." 
Brit. Mus. 32250, 32251, and 32252; also fragment of tusk, head 
of femur, and portion of tibia (Brit. Mus. 32256, 32254, 32253), 
which Lydekker (1886.2, p. 174) regards as belonging to the same 
individual as the teeth. Horizon an» Locality. — Pleisto- 

cene. Near Khanoos, Province of Erzerum, Armenia. Type 
Figure.— Falconer, 1863, PI. ii, fig. 2. 

History. —I'^alconer first named thissi)cciesin his "Syn(>|)(ical 
Table" of 1857, facing i)age 319: "Spec. 14. E. (Eueleph.) 
Armeniacus . . . Armenia: Erzeroom ... In the Brit. Mus. Coll. 
Discovered between Erzeroom and Moosh in 1856. The molar 



'[See footnote 1 on page 1049 above regarding tlie possible Lowest Pleistocene age of the Villafranchian. — Editor.) 



THE MAMMONTIN^: PARELEPHAS 



1061 



plates closely aijproxiinatctl, and the cnainel-edges very undulat- 
ed." In 1863, pp. 74-76, he renewed his description as follows: 
"These Khanoos molars are intermediate in character, between 
the Mammoth and the Indian Elephant, but more nearly allied to 
the latter. ... In the synoptical table appended to my Memoir 
on the 'Species of Mastodon and Elephant, &c.,' the Khanoos 
fossil form is ranged between E. indicus and E. primigenius, under 
the provisional name of E. Armeniacus. [Footnote: 'Quart. 
Journ. Geol. Society, 1857, vol. xiii, p. 319.'] . . . [p. 76] while E. 
Armeniacus, as stated above, approaches nearer to the existing 
Indian species." Falconer also speaks of the remains of fossil 
elephants on the banks of the Bosphorus and the northern shores of 
the Black Sea and of the Sea of Azof {op. cit. , p. 75) . On page 74 he 
gives what may be called the type description. 

Type Description (Falconer, 1863, p. 74). — "The speci- 
mens presented by Colonel Giels to the national collection, consist 
of two last upper molars in fine preservation, and a portion of 
a lower, all apparently of the same individual. These molars 
strike a practised eye, at the first glance, as presenting something 
intermediate between the Mammoth and the existing Indian 



monly seen in the existing species, less (jpen and less parallel 
across. The crown is broad, and the enamel plates are high. To 
render these descriiitive details more appreciable and available for 
comparison, I append the princiiml dimensions [l.M']:^ 



[In.] [Mm. 

11.75 299 



9.5 



242 



5.7 


146 


6.1 


156 


4.1 


105 


3.7 


95 


7.1 


181J' 




One of the Type Molaks of Parelephas armeniacus 
One-half natural size 
Fig. 942. Third .su]5crior molar, l.M', of Elephas Armeniacus Falconer, 
Falconer, 1863, PI. ii, fig. 2, p. 1 14: "Represents the crown-aspect of the last true 
upper jaw, left .side, of E. Armeniacus, from a si)ecinien in the British Museum, 
procured by Col. Giels, in the province of Erzeroum, in Armenia." 

Elephant. The case is of so much interest, that I shall venture on 
some of the details. The left upper molar (m.3, being No. 32,250 
Museum Kegist. Palseont. Gallery) is entire from behind the large 
front fang, the portion borne upon which had been ground down by 
protracted wear. [Footnote: 'PI. ii, fig. 2, shows the crown-view 
of the tooth.'] The anterior part of the crown to the extent of 2.7 
inches is also worn out, presenting merely a smooth surface of 
ivory, behind which there are seventeen ridges and a posterior 
talon. Of these, fifteen are more or less worn. The anterior nine 
form transverse narrow discs; the next six are divided nearly 
equally by two rather wide longitudinal channels into three 
divisions, consisting each of a flattened elliptical disc. The trans- 
verse discs, in their general character, bear a close resemblance to 
those of the Indian Elci)hant, the enamel-plates being rather thick, 
with very pronounced chjse-set crimping in the middle, but dimin- 
ishing towards the cornua. These discs are narrower than is com- 
'[See footnote 1 on page 1049 above. — Editor.] 



Extreme length of crown 

Length of crown surface in use (partly 

worn out) 

Space occupied by the anterior ten discs 

measured at top of crown 

Ditto ditto, at base of crown 

Width of crown at 3rd ridge (greatest). . 

Ditto at nth ditto 

Height of crown at 12th ridge 



(?) E. ARMENi.\cus IN Italy (Falconer, 1868). — The same 
description is found in the "Palaeontological Memoirs," 1868, Vol. 
II, p. 247, in which Falconer again treats Elephas armeniacus as 
intermediate between Elephas primigenius and E. 
indicus but more nearly allied to the latter. He adds 
(p. 249): "I now entertain a strong suspicion that 
remains of E. Armeniacus, or of a form closely allied 
to it, occur in Italy. This impression is founded 
upon specimens which I ob.served in the Natural His- 
tory Mu.seum of Turin, in the Unixersity Museum of 
Pisa, in the private collection of the Marchese Carlo 
Strozzi at Leghorn, and in those of Professor Ponzi 
and Signor Ceselli at Rome, the satisfactory specific 
identification of which puzzled me greatly. . . [p. 250] 
The space occupied by the twelve discs of wear, meas- 
ured along the summit of the crown, is 7 inches, yield- 
ing an average of about .6 inch to each, which comes 
very near that indicated above in the E. Armeniacus 
of Khanoos, i.e. .bl . The principal dimensions are: — 
Length of crown (not quite entire), 13.75 in. Extreme 
width of crown, 4.5 in. Height of crown at 12th ridge, 
8. in. Space occupied by the 12 discs of wear, 7. in. 
I have detailed notes of numerous other molars, ex- 
humed on the same occasion, from the same locality, St. Paolo, or 
near it, 'Nizza della Paglia,' which yield similar characters. ... I 
have been led to identify the 'Khanoos' and St. Paolo molars as 
being of the same species, E. Armeniacus, and to consider that 
they are not referable either to E. primigenius or E. anliquus[¥oot- 
note: 'See arUea, pages 187 and 192, note. — [Ed.]']." 

Osborn, 1928: (1) It is now known that these Upper Pliocene' 
Italian specimens thus referred by Falconer to E. armeniacus actual- 
ly belong to Elephas antiquus var. trogontherioides Zuffardi, as 
confirmed by Deperet and Mayet (see above). (2) Leith Adams 
(1877-1881, p. 241) mentions E. armeniacus but does not add to 
our knowledge. (3) Lydekker (1880.1, p. 285) erroneously gives 
the ridge formula of E. armeniacus as M 3 ^, whereas Falconer's 
type figure shows: M 3 ±i^. (4) Lydekker (1886.2, p. 174) 
remarks: "Elephas armeniacus. Falconer [Footnote: 'tjuart. 
Journ. Geol. Soc. vol. xiii, table facing p. 319 (1857).']. Syn. 



18.57. After 

molar, (m.3) 

No. 32,250, 



1062 



OSBORN: THE PROBOSCIDEA 



Euelephn.s armoiianis, auct. This species is definitely known only 
by the following [six] specimens, which belong to an animal of 
\ery large size. The structure of the molars indicates a species 
intermediate between E. pn'inigeniiin and E. indicus, which is 
probably an ancestral form closely connected with both [Footnote: 
76/V/. vol. xlii, p. 174 (1886).']. The crowns of the molars are 
extremely wide, the ridges api)roximated, with their extremities 
curxing backwards, the enamel is moderately thick and much 
l)licated, and the wear of the crown irregular. In their great width 
these teeth agree with those of E. priniigeniun, but in the other 
characters with those of E. indicus [Footnote: 'See Leith- Adams, 
British Fossil Elephants (Mon. Pal. Soc), p. 241.']." 

(.5) Lydekkcr (1886.3, p. 174) remarks: "From strata of un- 
known age at Erzerum, in Armenia, Dr. Falconer many years ago 
described some elephant molars under the name of Elephan 
arnieiiiaciis; and as Erzerum is comjiaratively near to Tabriz, it 
may be suggested that some of the Maragha elephants' teeth may 
not imjjrobably belong to this species; but be this as it may, the 
Erzerum and Maragha faunas may be geographically grouped 
together." 

Phylogenetic Conclusion (Osboun): The extremely able 
and acute observations of Falconer (18.57-1868) on this species 
from Armenia and on allied forms from Italy, confirmed by 
ZufTardi's observations on the resemblance of this animal to 
Elephna trogontherii of Pohlig, as signified by the specific name 
trogonlheiioides, also the recent observations by Deperet and 
Mayet (1923), all emphasize the intermediate character of this 
species and its phyletic relationship to the Elephafs inlermedius 
of .Jourdan and to the E. trogontherii of Pohlig. 

The ridge formula, M 3 ±ili*, indicates that E. armeniacus 
I'alconer is in about the same stage of evolution as the type of E. 
trogonlhrrii Pohlig, in which the ridge formula is said to be M 3 Yi+- 
\\ hilo I'alconer repeatedly described the type of E. armeniacus as 
iiitennediate in character l)etween E. primigenina and E. indirns, 
hut approacliiiig the latter species more clo.sely, Osborn regards it 
rather as belonging to the same phylum as E. inlermedius and E. 
trogontherii and refers it to the indei^endent genus Parchphas. 

PARELEPHAS ERRONEOUSLY RECORDED IN CHINA 
AND JAPAN 

I'>roneous arc the records of Farelephos in ( "hina and Japan. 
The supi)osed Elcphns nrmeniarus (Brit. Mus. 29007) is related 
(Lydekker, 1886.2, p. 169, Hopwood, letter, 1928) to Elephas 
[Pula'oloxodon] namadicus. Matsumoto's types of Parelephas 
protomnmmonleus and P. protomammontcns pro.rirnu.i of .Japan also 
iielong to Palnvloxodon. The following literature may be (luotcd: 

LvDKKXKK (1886.3, p. 174).^"There is in the British Mu- 
seum an elejihant's molar from China (No. 29007), which has been 
suggested to belong to this siiecies [Elephan armcniacux]; and if 
this were correct it would seem that the range of E. armeniacuH 
was somewhat the .same as that of Rhinoceros Hlonfordi, i.e. that 
it extended from western Asia through the regions lying to the 
north of India and China; I am, however, disposed to refer the 
specimen to E. nnnwdicus." 

1'akki,ki'Has AUMENiACi's KKK. (Lcitli Aiiauis and (1. Busk, 
1868, pp. 496 499): "In the British Museum there is a large part 

'[See page 1334 below.— Editor.) 



of a fossil molar tooth from China, which, on comi)arison with the 
teeth presented by Colonel (nils [(Jiels] from .\rmenia, so exactly 
resembles them in every respect, that no doubt can be entertained 
as to its belonging to the same species, namely, E. Armeniacus of 
Dr. Falconer; and as such, I