THEORIES AS TO ORIGIN, ANCESTRY, AND ADAPTIVE RADIATION 



799 



ANCESTRAL TITANOTHERES SUBCURSORIAX 

 The limbs of LamidotheTium are unquestionably of 



cursorial animals, and that the mediportal and 

 graviportal forms were secondary, a generalization 



primitive cursorial type, and those of Eotitanops are 1 that is indicated also by the skull and dentition. 



'^ID^ 



D 



Figure 721. — Evolution of the pelvis in the titanotheres 



Four progressive stages of adaptation: A, Ttie subeursorial Eotitanops borealis, Am. Mus. 14887, lower Eocene; B, the mediportal Palaeosyops major. 

 Am. Mus. 13116, middle Eocene (lower Bridger); C, the subgiaviportal Manteoceras manteoceras, Am. Mus. 2358, upper Eocene of Washakie 

 Basin; D, the graviportal Brontotherium gigas, Am. Mus. 492, lower Oligocene. All one-eighth natural size. In the first stage, Eotitanops, the 

 pelvis is long and narrow. 3<: in other primitive perissodactyls. In succeeding stages the gluteal blade of the ilium becomes transversely 

 expanded, so that in the final stage the transverse diameter across the ilia greatly e-xceeds the anterosposterior diameter. All this constitutes 

 an adaptation to progressively graviportal habits, as in the elephants. 



of subeursorial type, but these genera point alike to 

 a more primitive cursorial ancestry. These facts 

 appear to warrant the generalization that all the 

 earlyjtitanotheres (see pp. 733, 736) were small, light. 



A significant fact, to which W. K. Gregory first 

 called attention and which is commented on early in 

 this chapter, is that of all the early Eocene peris- 

 sodaotyls laiown, the oldest representatives of three 



