251 0. C. Harsh — Discovery of Cretaceous Mammalia. 



but the elevations are crescentic, with the convexity for- 

 ward. In a third series, the two outer rows are composed 

 of tubercles, and the middle one of crescents only. In the 

 smaller upper molars, the middle row and the outer one are 

 complete, but the inner row is less developed. In the second 

 series, where the true molars have but two rows of elevations, 

 these are either conical, or distinctly crescentic, the former 

 being especially seen in the smaller forms. 



Among the very large number of Allotheria teeth found in 

 the localities explored, not one has yet been secured corre- 

 sponding to the molar described by Cope, and referred with the 

 type to Meniscoessiis. The nearest approach to it is the upper 

 molar described by the writer in Part II, as the type of Selen- 

 acoclon hrevis (Plate VII, figures 9-12), and this tooth is smaller 

 in size. The two specimens are from widely distant localities, 

 and perhaps from distinct horizons. In the molar described 

 by Cope, one end of the crown was lost, and this fact was 

 stated in the original description, and shown in the first figure 

 published.* 



The various Allotheria remains represented in the accom- 

 panying plates include several that appear to be different from 

 those previously described, but in view of their fragmentary 

 nature, some have been referred to species already named. 

 The material now secured, when fully investigated, will test the 

 accuracy of these references. A few of the smaller forms 

 which appear to be quite distinct are briefly described below, 

 and their main characteristics shown in the plates ; accurate 

 figures being the best description, as every paleontologist 

 knows. 



Cimolodon parvus, sp. nov. 



One of the smallest forms of this genus is represented by 

 various isolated remains, among which is the anterior portion 

 of a lower jaw with the incisor in place, which may be 

 regarded as the type. This specimen is shown on Plate VI, 



* Prof. H. F. Osbom lias since given a figure of this tooth, which represents 

 the crown as complete (Am. Nat, July, 1891. p. 598). In the same paper, he has 

 a figure of the skull of Bolodon, of Owen, which is even more misleading. This 

 figure is made by putting together two different specimens not found associated, 

 and adding a strongly marked suture, which neither shows. Nothing is said in 

 the text in explanation of these indefensible reconstructions, which are mild 

 examples compared with other composite work in the same article, made out of 

 specimens I have described. Such use of my original figures, without permission, 

 while my preliminary work was still in progress, cannot be justified. 



Ou the same page with the Bolodon figure is another of a supposed premolar, 

 which, as I have shown elsewhere (Proc. Ac. Nat. Sci. Phila.. 1891, p. 240), is 

 almost certainly made up of portions of Hybodus fish teeth. Additional evidence 

 on this point makes it sure that other similar specimens are in existence, and 

 also suggests that the type of D'iglyphus may be a like production. 



