1910.] VARIATION OF THE SEA-ELEPHANTS. 585 



animal must for geographical reasons be just the one which 

 Lydekker has named " falclandicus" which thus in itself unites 

 most of the characteristics of all the others. The Sea-Elephants 

 have long ago been exterminated on the Falkland Islands, and if 

 now and then at the present time such an animal should be found 

 there, it is a straying individual which has come there accidentally, 

 most probably from South Georgia — an analogy to the fact that 

 sometimes, although seldom, a Walrus appears at the Norwegian 

 coast, in spite of the fact that the Walrus is as little pelagic in its 

 habits as the Sea-Elephant. 



The Sea-Elephant of the Crozet Islands Lydekker with some 

 hesitation regards as identical with those inhabiting Kerguelen 

 and Heard Islands. From a geographical point of view such an 

 assumption appears quite probable. If, however, this assumption 

 is accepted there is some material for the further consideration 

 of the " crosetensis " form available in the literature, because 

 Turner has, in his report on the Seals in the Scientific Kesults 

 of the Voyage of H.M.S. ' Challenger,'* communicated a table of 

 measurements of Sea-Elephant skulls, and among them are two 

 male specimens from Heard Island and one male from Kerguelen 

 Island. The lengths of the Heard Island skulls from premaxillary 

 to occipital condyle are respectively 493 and 486 mm., and the same 

 measurement of the Kerguelen skull is 402 mm. As the condyle 

 has been included in these measurements the figures quoted are 

 not directly comparable with those of the accompanying table, but 

 by comparing the corresponding measurements of skulls at hand 

 with the basal lengths of the same, it is easy to reckon how great 

 a reduction is necessary to obtain the approximate basal lengths 

 of Turner's skulls. It cannot be many millimetres wrong to 

 assume the basal length of the Heard Island skulls to be 470 and 

 464 mm. respectively and that of the Kerguelen skull to be 380 mm. 

 The measurements expressing the zygomatic width and the greatest 

 width of the palate of these skulls are also recorded in Turner's 

 table. If, then, the relations between these measurements and 

 the basal length are reckoned in percentages of the latter, the 

 following figures are obtained : — 



Heard Island. Kerguelen. 

 Zygomatic width in percent, of basal length 74'4 76 - 2 73'9 

 Greatest width of palate in percent, of] qo.7 00. q 09. p 

 basal length J 



The three former percentages fall all of them quite well within 

 the limits of variation found in the South Georgia skulls as 

 quoted above ; and if these specimens belong to the " crosetensis " 

 form, the zygomatic width of the same is only in exceptional cases, 

 as in Lydekker's specimen referred to above, greater than in South 

 Georgia specimens, i. e. " falclandicus ." The zygomatic width is 

 consequently no distinguishing characteristic between these two. 



Of the percentages expressing the relation of the width of the 

 palate, the first and the last are rather low even for South Georgia 



* Zoology, vol. xxvi. p. 6. 



