REVIEWS. 681 
this evidence of isolation. But while we would thus insist on the isola- 
tion of Zalophus, we would not consider it as entitled to rank other than 
are the relations between the members of that family and the groups 
which have been distinguished as sub-families in the apet and which 
we are happy to learn meet with Mr. Allen’s approbation. 
Availing ourselves now of the data that have adicit up to the 
present time, and which have been so well digested by Mr. Allen, we be- 
lieve that the relations of the Otariids may be expressed by the following 
synoptical table, in which only the most obvious and distinctive charac- 
ters are introduced. 
I. Skull with a more or less decurved front rostral profile, and with a 
sagittal groove from which are Mai ted the low ridges indicating 
the limits of the temporal mus 
A. Pelage with under-fur: vidas normal : rent 5 aee s 
above 
j Mnt dnos 
a. Snout much decurved above, and abbreviated, its length being 
, les SS than the longitudinal ver of the orbits, Callorhinus. 
in le 
ceeding the longitudinal — of the orbit, «itf . Arctocephalus. 
B. take without defined under 
bove 6-6; the Sr little remote from the preceding 
‘aad in a line with, or in advance of the transverse max! illo- 
palatine suture; bony palatal margin much nearer the 
pterygoid hamuli than the teeth; hinder feet with swimming 
membrane much produced and deeply incised, Otaria. 
b. Molars above 5-5; the last remote from the srepeding. and B 
hind the transverse maxilo palatine suture; bony palatal 
and m Eumetopias. 
II. Skull om as poma incurved fron to rostral profile, a and with. a 
st, 
solid, t ed and iu elevated sagittal cre Zalophus. 
-Although we are not inclined to place much stress on the sequence of 
forms when so many gaps remain unfilled, and when "f unknown might 
reverse the opinion that we have with more or less reason derived from 
me acquaintance wit th the seen, we are disposed to believe that the pre- 
of relationship, but such a system, especially when the genera are very 
numerous, becomes too complicated, and is of really little or no use. We 
Dpr.G i d on artificial groups, 
Tm, age "i 3. E 
are entirely differently limited. 
AMER. NATURALIST, VOL. IV. 86 
