( 9 ) 



the basisphenoid now rises at the lefl to a ureal height against tins 

 knob and, instead of following the cranial base normally, it is situated 

 here almost sagiltally: the complete basisphenoid must consequently 

 have shown a deep median groove. Analogous changes have occurred 

 at the basioccipital, which towards the knob shows a deep concavity. 

 By this the angle between basisphenoid and basioccipital has also 

 become more acute than in the walrus. 



In order to compare the hack of the skull with that of the walrus 

 we placed the fragment in such a position thai the upper edge of 

 the parietal has a slight forward inclination, as this is also the case 

 in a walrus skull, placed on the table without uisks. 



Then with both the basisphenoid rises slightly in a forward and 

 the basioccipital in a backward direction, so that their positions 

 may be considered as corresponding. 



With Alachtherium the outline of the posterior part of the skull 

 then shows one important change, caused by the strong crista 

 lambdoidea and the broad mastoid processes. For this causes the 

 back of t he skull to consist of a narrow supra-occipital and a verj 

 broad temporal part, a phenomenon which also with old walruses 

 is sometimes indicated to some extent, but never so strongly as with 

 Alachtherium. Moreover Alachtherium has a \er\ small cranium : all 

 bones are uncommonly thick. Further in Alachterium the crista 

 lambdoidea runs in a median direction far to the front and even 

 shows a tendency to pass into a sagittal ridge. As compared with 

 the walrus this phenomenon becomes very striking by the complete 

 absence of the crista occipitalis externa. Where consequently with 

 old walruses the occipital superius is strongly convex by the massive 

 crista occipitalis, it shows with Alachtherium a deep median fold. 

 Compared with Trichechus the hind skull of Alachtherium thus shows 

 three modifications: absence of the crista occipitalis externa, size of 

 the mastoid processes, and shape of the crista lambdoidea. The dif- 

 ferences at the side of the skull are not very great and we are 

 certainly not justified in stating, as van Benedkn does: that "la 

 conformation de toutes les regions est complètement différente." 

 Especially with the skulls of old male walruses Alachtherium shows 

 many points of resemblance and it would almost appear as if van 

 Beneden had only a small material for comparison at his disposal for 

 his description. 



The many changes which the skull of Alachtherium shows when 

 compared with the skull of Trichechus, have not deprived it, though, 

 of its Trichechid character. But there are three phenomena which 

 bring it nearer the Otaridae : the smallness of the bulla ossea, the 



