1887.] NATURAL SCIENCES OF PHILADELPHIA. 283 



close structural similarity between genera from widely separate 

 localities, and the absence of types which are wholly distinct from 

 the British, even in this greatly extended field of exploration. In 

 other words, tlie Mesozoic genera appear to have had a very wide 

 geographical distribution, and many scattered forms, while generi- 

 cally distinct, are closely connected by family and ordinal characters. 

 The second result, which has been attained much more slowly, is that 

 these mammals, with a few exceptions, can now be provisionally 

 classified, as they fall into small groups with carnivorous, omnivor- 

 ous, insectivorous, and herbivorous dentition, each with clearly 

 marked characters. The third result which rests upon more debat- 

 able grounds is, that while many of these Mesozoic families became 

 extinct in succeeding geological periods, many others belong in or 

 near the ancestral lines of families still existing among the Marsu- 

 pials and Insectivores. 



Before entering the subject of classification I will give a brief 

 resume of my observations upon the British genera in so far as they 

 differ from, or are additional to those of Professor Owen. It appears, 

 from the dentition of a mandibular ramus belonging to Amphitherium 

 which has been recently added to the British Museum collection, 

 that three distinct genera have hitherto been included under this 

 genus, which may be separated by the number of premolars, by the 

 conformation of the molar crowns and the form and positions of the 

 condyle, as Amphitherium, Amphitylus and Aviphilestes} (2) Phas- 

 Golotherium. The formula is i 4 c 1 pm m 7. It seems more 

 probable that this genus has lost its premolars, (which formerly may 

 have filled the diastema), than that they should all have become 

 molariform, a modification which is always slowly acquired. (3) Tri- 

 conodon. The upper canine of this genus has a double fang. The most 

 interesting fact brought out is the evidence the various specimens 

 furnish of relationship to the Marsupials in the succession of the 

 teeth.* The fully adult molar formula is pm 4 m 4. The fourth 



1. The Amphitherium (A. Prevostii) molar is bicuspidate with a low posterior 

 heel. The Amphitylus {gen. nov.) molar has three blunt cusps, and an internal 

 cingulum; form,// 6, w 6. The Amphilestes molar has three prominent cusps 

 and a pronounced cingulum, encircling the crown ; pm 5, m 6. After a personal 

 examination of the types, Mr. Lydekker writes me (April 16th) that he finds the 

 formulae of both Amphilestes and Amphitylus, as follows : i ^^ c\ pm\ m 7. 



2, Assisted by Mr. Lydekker, I have confirmed the suggestions of Professor 

 Owen and Professor Flower. 



