PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 421 
As I have not seen Von Schrenck’s original account, I do not feel sat- 
isfied that he positively intended to affirm that the roots of the second 
premolars are bifurcate, although he is so interpreted by Allen.* 
From my examination of the skulls at hand I should be inclined to 
doubt that they are ever so. The form of the roots of the first true 
molar is such as to afford strong ground for a difference of opinion. 
There is invariably a longitudinal groove along the root, but the tip of 
the latter may or may not be sufficiently cleft to admit of the use of the 
term bifurcate. The posterior accessory cusps are apparent in the supe- 
rior true molars, but almost or quite disappear from the premolars. The 
anterior accessory cusps are wanting superiorly. In the lower jaw both 
accessory cusps are traceable in all the grinding teeth, but are strongest 
in the true molars. The crowns of the premolars are considerably di- 
rected backward ; the true molars less so or not at all, although they 
bear that appearance when worn. The molars are so placed in the ma- 
jority of specimens of both Histriophoca fasciata and Phoca vitulina that 
they rub against each other. This results in the obliteration of the an- 
terior accessory cusps of the superior molars when the teeth are worn. 
The skeleton; vertebral characters.—The vertebral formula is as fol- 
lows: 0. 7; D. 15; L. 5; 8S. 3; Ca, 14=44. The atlas is low and 
broad, with wide transverse processes. The axis is high, with a large 
and long odontoid process. The superior border of its neural spine is 
parallel with the long axis of the centrum. In Pusa it forms an angle 
of about 45° with that axis. All the processes of the remaining cervi- 
cal vertebre are short and stout. The low neural spines of the dorsal 
and lumbar vertebre are subequal in height, but increase in breadth 
(in the anteroposterior direction) from the first backward. They are 
lowest at about the middle of the series. They do not bend backward 
so strongly as in Pusa. The hypapophysis becomes apparent in the 
thirteenth dorsal, and increases in size posteriorly, becoming strongest ° 
in the second lumbar. The transverse processes of the lumbar verte- 
bree increase slightly in length posteriorly and are subequal in breadth. 
Only the first two caudals have complete neural arches. The last 
twelve decrease in size very gradually posteriorly. 
The ribs and sternum.—The first rib is very broad and short, but less 
so than in Pusa. The first thirteen have articular facets, but that of 
the twelfth is rudimentary, and does not reach the centrum. Nine are 
connected with the sternum, the first at the base of the first of the 
sternebre. In Pusa ten ribs reach thesternum. The sternum consists 
of nine sternebre, the posterior, or ninth, nearly two and a half times 
the length of any of the others. The first is the smallest, somewhat 
conical, with the apex directed forward and compressed laterally; the 
others, except the ninth, are subequal. 
The scapula and pelvis.—The scapula is falciform, being prolonged 
*ALLEN, North American Pinnideds, 1880, p. 675. 
