JuxY 20, 1906.] 



SCIENCE. 



91 



manner of their implantation, but by the more 

 complex structure of the crown of the fourth 

 superior premolar as compared with that of 

 the modern seals. This tooth may be said to 

 be implanted by three roots, of which the one 

 supporting the inner cusp is not entirely dis- 

 tinct at the base, at least. The crown ex- 

 hibits the remains of three distinct cusps 

 corresponding in every way to those of the 

 less specialized sectorial of many of the 

 creodonts. It is somewhat simpler in struc- 

 ture than the corresponding tooth of Patrio- 

 felis, but its derivation from a tooth of that 

 character is clearly evident. The single molar, 

 both above and below, is missing from the 

 specimen, so that the organization of the 

 crowns can not be determined. They were 

 each implanted by two distinct roots. The 

 dental formula is I. 3/2, C. 1/1, PM. 4/4, 

 M. 1/1. 



One character which is most unusual for a 

 seal is the deep heavy horizontal ramus of the 

 lower jaw, together with a relatively power- 

 ful symphysis reminding one at once of the 

 corresponding parts of Oxycena and Pairiofelis 

 among the creodonts. The impression which 

 one immediately receives upon exaraination 

 of the specimen is that of a short heavy- jawed 

 animal. I mention this character in partic- 

 ular for the reason that the heavy jaw and 

 powerful symphysis of Pairiofelis was made 

 the basis of an especial objection, on the part 

 of Osborn, to any possibility of affinity be- 

 tween the two. The jaws of the modern seals 

 are relatively weak and slender. 



In the cranium the more important char- 

 acters to be noted are the rudimental post- 

 orbital processes, the marked constriction of 

 the postorbital region, the relatively heavy 

 zygomatic arches and the peculiarly distinctive 

 creodont organization of the otic region of 

 the skull. The mastoid is of moderate pro- 

 portions and the widely separated paroccipital 

 process is unusually large and massive and 

 projects outwards and backwards. This ar- 

 rangement and unusual size of the paroccipital 

 is one of the most distinctive characters of the 

 creodont skull and is one not fovmd in any of 

 the modem Carnivora. I may add likewise 



that its condition in this ancient seal is al- 

 most identical with that seen in Pairiofelis 

 and Oxycena. 



The base of the skull had not been suffi- 

 ciently freed from the inclosing matrix to 

 permit of an accurate determination of the 

 characters of this region, but I think the 

 presence of a small rugged uninflated bulla 

 and an alisphenoid canal may be assumed. 

 The occipital crest is low and inconspicuous 

 and does not overhang the occiput as in the 

 creodonts. The brain case is large and roomy 

 and the brain was considerably convoluted. 

 In fact the brain case and occipital region of 

 the skull resemble that of the modern Phoca 

 closely. 



From a consideration of the foregoing char- 

 acters of this important specimen the follow- 

 ing conclusions appear to be firmly and con- 

 clusively established: (1) the seals have been 

 derived from ancestors in which the molars 

 were tri tubercular ; (2) in which there were 

 but two pairs of incisors in the lower jaw; 

 (3) in which the true molars were early re- 

 duced; (4) in which the lower jaw was stout 

 and heavy with a strong, heavy symphysis; 

 (5) in which the mastoid region of the skull 

 was identical with that of certain of the creo- 

 donts; (6) in which the postorbital process 

 was rudimentary or wanting. 



Less firmly established characters of the 

 ancestors of the seals may be enumerated as 

 follows : (1) Double tongue and groove articu- 

 lations of the lumbar vertebrae; (2) early de- 

 velopment of aquatic habits with consequent 

 modification of limbs. 



If next we direct our attention to the appli- 

 cation of these characters to the known fossil 

 Carnivora which preceded them in time, we 

 find that it is only among the creodonts that 

 these characters are to be met with. It, there- 

 fore, follows that a creodont ancestry is the 

 only possible or logical solution of the prob- 

 lem. As to the particular family of creodonts 

 from which the seals have been derived the 

 evidence is by no means so clear nor con- 

 clusive. As far as now known the choice 

 seems to be restricted to either the Hysenodon- 

 tidae or the Oxyaenidse. Of these the latter are 



