758 



TITANOTHERES OF ANCIENT WYOMING, DAKOTA, AND' NEBRASKA 



progressive Meseutheria, which includes the Am- 

 blypoda and Condylarthra, theoretical branches of 

 relatively archaic and conservative stock (see Greg- 

 ory, 1910.1, p. 458). This question is, however, still 

 an open one and is of so much importance and inter- 

 est in connection with the whole problem of the 

 evolution of the hoofed mammals that it is desirable 

 to review it in more detail in the light of our present 

 knowledge. The final solution must await the discov- 

 ery, probably in Asia, of the actual Upper Cretaceous 

 and basal Eocene ancestors of the perissodactyls. 



CARPUS OF CREODONTA, AMBLYPODA, CONDYLARTHRA 



Our present (1918) knowledge of the structure of 

 the carpus in these primitive orders of mammals is 

 displayed in the accompanying Figure 687. All are 

 pentadactyl, possessing five more or less spreading 

 digits. The creodont Dissacus, the amblypod Panto- 

 larribda, and the primitive condylarths Euprotogonia 

 and Meniscoiherium are more or less plantigrade. The 

 condylarth Phenacodus is secondarily digitigrade and 

 tending toward tridactylism. 



Figure 687. — Carpus of Creodonta, Amblypoda, and Condylarthra, showing the primitive "alternating" 

 placental type and the specialized "serial" type 



Pentadactyl 

 Creodonta: 



A, Dissacus carnifex. Four-ninths natural size. A primitive placental type. Magnum small, supporting centrale and halt oi lunar; lunar 



wedge-shaped, resting on magnum and unciform. 

 Amblj'poda: 



B, Pantolambda bathmodon, Am. Mus. 2547. Natural size. Magnum small, supporting centrale and part of lunar; lunar wedge-shaped, 



resting on magnimi and unciform. 

 Condylarthra; 



C, Mmiscotherium sp. Natural size. Magnum small, supporting centrale and lunar; lunar flattened, resting on magnum and unciform. 



D, Euprotogonia. Four-thirds natural size. Magnum small, truncate, partly supporting scaphocentrale; chiefly supporting lunar; lunar 



truncate, retaining contact with unciform. 



Mesaxonic 



E, Phenacodus. Two-thirds natural size. Serial type. Magnum enlarged, truncate, wholly supporting lunar; lunar truncate, resting 



chiefly on magnum, with small contact with unciform. 



HYPOTHETICAI ORIGIN OF THE PERISSODACTYLA FROM 

 THE CONDYLARTHRA 



HISTORY OF OPINION 



Cope (1884-1898) regarded the Condylarthra, 

 typified by Euprotogonia and Phenacodus, as repre- 

 sentives of the long sought five-toed ancestors of the 

 perissodactyls. Osborn (1890.51) at first accepted 

 this view and with Cope regarded the manus of 

 Phenacodus as close to the primitive foot of the Peris- 

 sodactyla and Ungulata; but in 1893 (1893.82, p. 89) 

 he abandoned Cope's view and proposed the theory 

 that the Condylarthra were part of an archaic ungu- 

 late radiation not directly ancestral to the perisso- 

 dactyls. Subsequently Osborn developed the idea 

 (1894.89) that the source of the Perissodactyla was to 

 be sought far back of the Condylarthra. 



The serial carpus of Phenacodus, as first observed by 

 Matthew, is obviously secondary when compared 

 with that of Euprotogonia. This important fact 

 compels us to abandon Cope's great hypothesis that 

 the serial or taxeopod type was ancestral to that of all 

 the Ungulata; certainly the Amblypoda and Con- 

 dylarthra did not spring from a "serial" type. 



This fact also modifies many of the conclusions 

 which Osborn reached in his "Evolution of the un- 

 gulate foot" (1890.51, p. 531), a treatise based upon 

 the adoption of Cope's theory that the common 

 protungulate foot was serial. 



Matthew (1897.2, pp. 309-310) showed that the 

 taxeopod, serial foot of Phenacodus is not primitive 

 but secondary, because the foot of its ancestor 

 Euprotogonia has partly alternating or displaced 

 carpals. He reached the following conclusion (op. 

 cit., pp. 300, 308, 309): 



