32 



angle less marked, and lower border much less curved ; opposite 

 sides completely anchylosed, a deep hollow under upper and inner 

 edge. Cavities existing for eight molars, the socket of the anterior 

 one being simple ; two posterior molars but partially developed. 

 Lower molars more distinctly three-ridged than the upper ones, but 

 the ridges less evidently tri- tubercular. Molars with two fangs, an- 

 terior and posterior, resembling the two external fangs of the upper 

 molars, directed downwards, flattened and expanding, especially the 

 posterior one. Molars deciduous from before backwards, seemingly 

 forced out by the gradual advance forwards of the posterior ones. 



The temporal bones being both wanting, I am unable to speak of 

 the zygomatic processes, which differ in shape in the two previously 

 known species. 



Dr. Vogel's measurements being from an entire head, while mine 

 are from the dried skull, the size of the respective animals will nearly 

 approach each other, mine being rather the smaller. In the distance 

 between the orbit and the snout, on which Prof. Owen lays stress, 

 they will be found so fairly to agree that they may be presumed to 

 belong to the same species. Let us now therefore see whether the 

 other measurements and proportions of the one we have been con- 

 sidering differ sufficiently from others to favour the presumption of 

 its being a species. In 31. Senegulensis, the contour, looking at the 

 skull from above downwards, is nearly that of an isosceles triangle, 

 closely approaching an equilateral triangle, while that of 31. australis 

 more resembles the outline of a violoncello. In the Niger specimen 

 again, the form, though more nearly resembling the former, is cer- 

 tainly of an intermediate character, the base of the triangle being 

 shorter in proportion. The profile view of 31. australis shows a 

 lengthened, rather narrow beak, while 31. Senegalensis has one 

 shorter and remarkably deep ; and here again we have an interme- 

 diate form, the shape in this case certainly more resembling 31. au- 

 stralis. The inferior border of the lower jaw of 31. australis is long 

 and straightened, while that of 31. Senegalensis is short and curved, 

 its posterior angle, also, being more massive and decided, and ap- 

 proximating to that of the Dugong. Here again the Niger Manatee 

 intervenes, the angle being more obtuse, and the curve less than in 

 the Senegal species. The proportion of the length of the nasal open- 

 ing in 31. australis is to the breadth as 3 to 1, in 31. Senegalensis 

 as 1 to f , but in my specimen as 2 to 1 . The coronal suture, sharply 

 angular in the South American and almost semicircular in the Sene- 

 gal species, is in the Niger one acutely arched. The temporal ridge 

 irregularly converges posteriorly in 31. australis, in 31. Senegalensis 

 they gently diverge, while here they run antero-posteriorly almost 

 entirely parallel. The temporal bones being, as I have remarked, 

 absent, I cannot speak of the temporal zygomatic apophyses ; but 

 the molar portions which remain would seem to indicate a continu- 

 ance of the same intermediate character. 



But in a few points the Niger skull is peculiar, and differs quite 

 from the others. Thus the superior and anterior angle of the parietal 

 bone extends much further forward than in either of the others, being 



