NO. 2H3. EXTINCT SIRENIAN DE8M08TYLUS HESPERUS— HAY. 387 



this process extends backward on its upper edge to the outer end of 

 the occipital ridge. The lower border of the process extends back- 

 ward and inward to the buUa-Uke tympanic ridge. Between these 

 two borders, or roots, of the zygomatic process is a large tympanic 

 cavity. At the front of this cavity the process is 33 mm. high; at 

 the rear of the malar bone, two-thirds of the length from the rear, 

 it is 53 mm. high and 14 mm. thick. The glenoid fossa is quite 

 different from that of either Triclieclius or Halicore. In these genera 

 the articular surface for the lower jaw is elevated (as seen with the 

 palatal surface held upward) above the surrounding bone. In 

 Tricltechus, especially, there is a deep transverse groove behind this 

 surface, and behind the groove is a ridge. In Desmostylus the articu- 

 latory surface is flat and on a level with the rest of the bone. It 

 measures 39 mm. from side to side; 22 mm. from front to rear. 



The auditory organ ts very different from that seen in Tricheclius 

 and Halicore. In the former there is an oval opening 74 mm. wide 

 and 53 mm. long between the exoccipital and the alisphenoid, and 

 this is occupied mostly by the otic bones. It is smaller in Halicore, 

 but stiU large. In Desmostylus what coiTesponds to the same open- 

 ing extends obliquely forward and inward a distance of about 60 mm. 

 and fore and aft about 20 mm. Included in this opening is, out- 

 wardly, what is probably the stylomastoid foramen; mesiaUy, the 

 foramen lacerum posterius. In front of these openings, running 

 obHquely forward and inward, is a ridge 50 mm. long, about 12 mm. 

 high, and 7 mm. thick at the base. It is wedged in between the glen- 

 oid fossa and a part of the alisphenoid in front and the exoccipital 

 behind. It evidently corresponds to the tympanic bulla of the ox. 

 In the deep tympanic cavity the tympanic ring is probably not 

 ankylosed to the suiTounding bones, but the sutures can not be 

 distinguished. 



In Tricheclius and Halicore there is a cleft between the exoccipital 

 and the squamosal which is partly fiUed up by the mastoid portion 

 of the petrous bone. In Desmostylus the mastoid portion is rela- 

 tively much larger and forms a prominent portion of the hinder lower 

 angle of the skull. Seen from behind it forms a strip of bone 60 mm. 

 or more high and 28 mm. wide, tightly wedged in between the par- 

 occipital and post-tympanic processes. Seen from below it passes 

 between the processes nearly to the stylomastoid foramen. In a 

 yoimg musk-ox I find a very similarly disposed mastoid, but it is only 

 55 mm. high and 11 mm. wide. 



As akeady stated the alisphenoid does not come into contact with 

 the parietal. The pterygoid processes are feeble in comparison 

 with those of Tricheclius. They are thin — about 8 mm. or 10 mm. 

 thick, and they descend from the level of the basisphenoid only 

 about 20 mm. In a skull of Trichechus thev descend a distance of 



