THEORIES AS TO ORIGIN, ANCESTRY, AND ADAPTIVE RADIATION 



773 



The structure of the titanothere pes is typically 

 perissodactyl, whereas the structure of the titanothere 

 manus is comparable rather to that of the artiodactyl 

 (even-toed) type. 



The retention of D. V in the titanotheres is probably 

 due to the early evolution of a graviportal stage, out 

 of more primitive mediportal and subcursorial stages. 



Classification hy foot structure. — In devising the 

 nomenclature of the hind foot to signify the uneven 

 number and uneven arrangement of the digits the 

 perissodactyl branch of the Ungulata, including the 

 titanotheres, has been variously termed the Peris- 

 sodactyla by Owen (1848.1), the Imparidigitata by 

 Riitimeyer (1865.1) and by Burmeister (1867.1), 



WIND RIVER 



gested by Marsh, because while in the hind foot 

 the main weight passes through the middle toe and is 

 therefore mesaxial, in the fore foot of many but not 

 all of the titanotheres the main weight is distributed 

 between the third and fourth toes and is therefore 

 paraxial, although not so distinctly so as in the hippo- 

 potamus. 



Among the amynodonts (amphibious rhinoceroses) 

 we also observe the four-toed structure of the fore foot 

 as in certain titanotheres. To reach the real meaning 

 of this four-toed structure of the fore feet we compare 

 the titanotheres with the most primitive laiown 

 direct ancestors of the horses, tapirs, and rhinoceroses. 

 We find that these also are four toed and more or less 



UPPER UINTA 



CHADRON 



Figure 701. — Evolution of the astragalus in the titanotheres 

 Upper row, back view; lower row, front view. 



A, Lambdotherium sp., Am. Mus. 14921, lower Eocene, Wind River formation. 



B, EotUanops borealis, Am. Mus. 14888, lower Eocene, Wind River formation. 



0, Limnohyops monocomts, Am. Mus. 11089, middle Eocene, lower part of Bridger formation. 



D, Manteoceras manteoceras, Am. Mus. 1587, middle Eocene, upper part of Bridger formation. 



E, ProtUanothcrJum superbum, Am. Mus. 2030, upper Eocene, Uinta C, true Uinta formation. 



T, Menodus gigajiieiis, ,\m. Mus. SO.'i, lower Oligocenc, upper Titanotherium zone (Chndron C) . 



This series shows the progressive widening of the astragalus in the transition from cursorial to graviportal habits. The earlier members have the troch- 

 lear keels high and angulate, the neck well defined, the cuboid facet verj* small, the sustentacular facet long and narrow. 'The latest members have 

 the trochlear keels low and rounded and the head sessile, the constriction, or neck, being nearly obliterated. The cuboid at this stage is very wide, 

 the sustentacular facet short and wide. 



after the principles of classification of the French 

 paleontologist De Blainville (1816.1), and the Mesa- 

 xonia by Marsh (1884.1). 



In seeking the relations of the hoofed animals 

 (Ungulata) Richard Owen, who followed the distinc- 

 tions in foot structure first pointed out by De Blain- 

 vdle between the "pachydermes a doigts pairs" 

 (number of digits even) and the "pachydermes a doigts 

 impairs" (number of digits odd), might have found the 

 isolated fore foot of an Oligocene titanothere a 

 difficult subject. De Blainville's distinction between 

 an odd or even number of digits at first appears to 

 fail, as the fore foot of the titanotheres is four-toed, 

 or paired, broadly resembling that of the hippo- 

 potamus. It is no less difficult to apply to the fore 

 feet the distinctions mesaxonic and paraxonic sug- 



paraxonic in the fore feet, which proves that the fore 

 feet of the titanotheres are in one sense a persistent 

 primitive element, because, they retain the fully func- 

 tional outside toe of the primitive five-toed foot. 



The primitive titanothere pes is very similar to 

 that of the other primitive Eocene perissodactyls. 

 All these animals are in a cursorial or subcursorial 

 stage of evolution. The progressive titanothere pes 

 shown in Figure 701 is in the graviportal stage, in 

 which the astragalus is flattened and rests broadly 

 on the cuboid. It thus resembles the astragalus of 

 the graviportal rhinoceros rather than that of the 

 mediportal tapir, in which the astragalus has a 

 narrow footing on the cuboid, and bears no resem- 

 blance at all to that of the horse, in which the as- 

 tragalus has entirely lost its anterior footing on the 



