1910.] . VARIATION OF THE SEA-ELEPHANTS. 583 



equally large quite young individuals is 5 - 6 per cent, and in the 

 other 10 - 6 per cent, of the basicranial length (conf, table of 

 measurements). 



The shape of the palate should also be different in Lydekker's 

 " subspecies '' in such a way that it ought to be " flat behind and 

 hollowed in front " in " falclandicus" " markedly hollow through- 

 out " in " macquariensis" and " almost flat " in " crosetensis." 

 The skulls from South Georgia prove that the degree of " hollow- 

 n ess" is very different in different specimens. In the largest 

 skull, for instance, the palatina are markedly less " hollow " than 

 the maxillary region of the palate, and the lateral portions of the 

 first mentioned bones are very strongly convex. In the second 

 largest skull the whole palate from the posterior border is evenly 

 and strongly vaulted. In the third largest skull the palate is 

 much less " hollow," the palatina are almost Hat and somewhat 

 convex laterally. In the fifth largest the palatina are almost flat, 

 but the remainder of the palate is very concave. In the sixth the 

 palate is almost completely fiat, and in the seventh it is only a 

 little " hollow." There is thus almost every degree of " hollow- 

 ness " represented among these seven skulls, and characteristics 

 derived from such a condition cannot be of subspecific value, 



Mr. Lydekker has also used the relative length of the ptery- 

 goid processes as a distinguishing characteristic of subspecific 

 value, saying that this process is " small " in u falclandicus" 

 " longer" in "macquariensis," and "very slender" in "crosetensis." 

 Unfortunately, the pterygoid process is subjected to just as great 

 individual variation as other portions of the Sea-Elephant skull. 

 As Lydekker has not recorded any direct measurements with 

 which a comparison can be made, it may be suitable to express its 

 relative length by measuring the distance between its posterior 

 tip and the nearest opposite point on the bulla. This distance 

 measures in the seven skulls from South Georgia respectively 

 22 mm., 10 mm., mm. (in this skull the pterygoid process forms 

 a complete bridge over to the bulla and is anchylosed with the same), 

 12 mm., 10 mm., 9 mm., and 4 mm. (the skulls are enumerated 

 in the same order as in the table of measurements). As regards 

 slenderness, the pterygoid process is always compressed and its 

 longest diameter at the tip varies as the following measurements 

 indicate: 10 mm., 13 mm., 15 mm., 19 mm., 15 mm., 12 mm., and 

 8 mm. The variability of this bone may by this be fully elucidated. 



The palatal aspect of the premaxillaries is said to be V-shaped 

 in "falclandicus " and " macquariensis," but U-shaped in " croset- 

 ensis." The present material from South Georgia proves that 

 this characteristic is just as variable as the others. In the fifth 

 and sixth of these seven skulls the V- shape of the palatal aspect 

 of the premaxillaries is quite typical, but in the second and third 

 the U -shape is just as pronounced as in Lydekker's figure of 

 " crosetensis " ; the others show intermediate features. 



The accompanying figures (text-fig. 55, A-D) show this different 

 shape of the premaxillaries in the Sea-Elephant from South 



