1917] 



Stock: Skull Structure of Mylodont Sloths 



171 



or the inner half of this lobe may be reflected anteriorly, assuming a 

 position more nearly at right angles to the long axis as in no. 645. The 

 posterior lobe may be at right angles to the long axis, as already noted, 

 or it may also be decidedly oblique, as in no. 645. In the latter speci- 

 men, which in length of skull exceeds somewhat Mylodon garmani, 

 the fourth inferior tooth of the right side has practically the same 

 anteroposterior length as in that species, while the tooth of the left 

 side is somewhat shorter. 



The third inferior tooth may occasionally be quite broad from the 

 antero-internal to the postero-external side, as in no. 696. As a rule, 

 the proportions of this tooth are more like those in the type of M. 

 harlani, or the tooth may be narrower in Rancho La Brea skulls. 



The fifth superior tooth generally is comparatively small. In no. 

 642 the anteroposterior diameter of this tooth is somewhat longer than 

 is usual, and thus is more like the corresponding tooth in Brown's 

 specimen. 



MYLODON HAELANI TENUICEPS, n. subsp. 

 Plates 3 and 4 



Type specimen no. 642, a skull contained in the Rancho La Brea collection of 

 the Museum of History, Science and Art of Los Angeles. 



Subspecific characters. — Skull more constricted behind postorbital processes 

 and cranium more slender than in M. harlani from Rancho La Brea. Associated 

 with these characters is a very narrow palate posterior to the fifth tooth. In 

 possessing a slender cranium the subspecies approaches M. garmani more than 

 does the typical M. harlani from the asphalt deposits, differing, however, from 

 both in distinctly narrower palate. It resembles the Nebraska skull, no. 2780, 

 Amer. Mus. Coll., described by Brown in degree of constriction behind the post- 

 orbital processes, but differs in having a narrower palate. 



The extensive range of variation exhibited by the skull and denti- 

 tion of Mylodon harlani makes it difficult to establish a type absolutely 

 distinct from the parent species. It does not seem possible to establish 

 a series showing gradations between one extreme, as represented by 

 no. 642, and the other extreme as exemplified by no. 21158 Univ. 

 Calif. Coll. Palae., because of difficulties encountered in endeavoring 

 to bridge the gap between the narrow and broad types of skulls. 



No. 644 was considered at, first to supply a very necessary link in 

 establishing a series of nearly perfect gradation, but after further 

 comparison its affinities appeared to be rather more with the broader 

 skulls. In least width of palate and least width of skull behind post- 

 orbital processes (see table of measurements 8 ) no. 644 is undoubtedly 



8 Compare with table of measurements in Univ. Calif. Publ., Bull. Dept. Geol., 

 vol. 8, p. 326, 1914. 



