586 DR. EINAR LONNBERG ON THE [Mar. 15, 



specimens. The middle one again (38'3) is about one per cent. 

 higher than the highest of the corresponding dimension of South 

 Georgia specimens. From this fact and that recorded above from 

 Lydekker's Crozet specimen only one conclusion can be drawn, 

 that sometimes the Sea-Elephants of the Crozet- Kerguelen- Heard 

 Islands geographic group have a greater palatal width than their 

 relatives in other places so far as is known. It is not, however, 

 a general rule and thus no distinguishing characteristic. 



With regard to other measurements also, the relative dimensions 

 of the skulls from Kerguelen and Heard Islands measured by 

 Turner vary within the same limits as the corresponding figures 

 of South Georgia skulls. As an example may be mentioned that 

 the width of the skull at the posterior edges of meatus audi- 

 torius extemus is in the South Georgia skulls from 61*9 to 68"4 per 

 cent, of the basal length (conf . table of measurements), while the 

 corresponding measurements of Turner's skulls are respectively 

 64 - 4, 61 "2, and 66 - 8. These latter measurements do not indicate 

 any greater width of the skull than the corresponding figures of 

 South Georgia skulls. 



Another example indicating a similar condition is obtained by 

 comparing the width of the maxillary portion of the skull with 

 the basal length. In this case I have measured the South Georgia 

 skulls on a level with the upper posterior premaxillary suture. 

 The percentages expressing this relation (conf. table of measure- 

 ments) vary in the adult and semiadult males from 30"9 to 

 38 - 3 per cent., without corresponding with the size or age of the 

 animals. With the above measurements may correspond fairly 

 well those recorded by Turner as indicating the " width of maxilla 

 across middle of rostrum " of Heard Island and Kerguelen speci- 

 mens. If these then are compared with the basal length of the 

 skulls (conf. above), the following percentages are obtained 

 (enumerated in the same order as before) : 34*0, 36 - 4, and 32'1. 

 All these fall evidently within the limits of variation of the South 

 Geoigia material, and they do not at all indicate any greater rela- 

 tive width of this portion of the skull in the Heard-Kerguelen 

 animals than in those from South Georgia. 



In Turner's table of measurements there are some figures 

 indicating the " smallest interfrontal width." As I do not know 

 whether this measurement coincides with the least frontal width 

 according to my way of measuring, I cannot make any direct 

 comparison, but only quote the percentages obtained. In the 

 South Georgia material there is a variation from 15'0 to 20*5, but 

 the percentages reckoned from Turner's figures are respectively 

 15-1, 14*0, and 13 - 6, thus lower than the former. If Turner's and 

 my own measurements really coincide, as I believe they do, the 

 percentages appear to indicate a narrower forehead in the Heard- 

 Iverguelen Sea-Elephan bs. 



The measurements hitherto considered show very plainly a 

 quite irregular variation which does not correspond with size or 

 age. But there is another relative dimension which exhibits a very 



