70 bulletin: museum of comparative zoology. 



The inferior surface of the maxillary fragment containing PM 2 and 

 the alveolus of PM 1 , just mentioned, is apparently complete, and indi- 

 cates that the palate, and hence the rostrum also, was quite narrow 

 behind the canines, although the posterior extension of the palatal 

 branch of the premaxillae may have added a little to the breadth. This 

 fragment is from the left side, while the piece of the maxilla containing 

 three teeth is from the right side. Hence, it is not possible to determine 

 positively whether, when the former is transposed to the right side, as 

 has been done in figure 1, the two pieces should be contiguous, or 

 whether space for another tooth should be left between them. I am of 

 the opinion that they should be contiguous. 



The fragment containing three teeth (Plate 1, figure 2) is one of 

 Gibbes's types and was well figured by him in 1845 (Proceedings of the 

 Academy of Natural Sciences of Philadelphia, 1845, Vol. 2, Plate 1, fig- 

 ure 4) and 1847 (Journal of the Academy of Natural Sciences of Phila- 

 delphia, 1847, series 2, Vol. 1, Plate 4, figure 1), though reversed, and 

 in 1847, at least, considered as belonging to the mandible. Beside the 

 three teeth it has, at the posterior end, a concavity which represents 

 the outer wall of the alveolus of a tooth about half the size of the others. 

 This was probably the last molar (M 2 ). The three large teeth are all 

 somewhat broken, but enough of them remains to indicate their original 

 form and size. A very small fragment of the palatal surface of the max- 

 illa remains attached to the second of these teeth near its anterior root. 

 It is concave and presents a small depression which may mark the posi- 

 tion of the apex of one of the mandibular teeth. As already mentioned, 

 the fragment bears at the posterior end a short, triangular process, flat 

 on top, intended to receive the anterior end of the malar bone. A small 

 piece only appears to be lacking from the process. The outer surface of 

 the maxilla above the three large teeth is convex, and there is a depres- 

 sion above the second tooth (PM 4 ) which appears to represent the ante- 

 orbital foramen. Above the third tooth (M 1 ) a short, narrow, triangular 

 ridge is developed. This ridge appears to be nearly complete, and its free 

 end is probably about in line with the posterior margin of the maxilla, or, 

 in other words, the anterior margin of the orbit. The shape of the max- 

 illa above the three large teeth (PM 8 , PM 4 , and M 1 ) and of the ridge 

 just mentioned is characteristic of Dorudon, and distinguishes it from 

 Zeuglodon brachysjyondylus, but Prozeuglodon appears to have a similar 

 conformation. 



The original height of this posterior portion of the maxilla cannot be 

 determined accurately, but from the small angle between the superior 



