8 



proceed so far as to produce on wearing a confluence of the crescentoid 

 surfaces. This is the case in Orotherium sylvaticum in the mandibular 

 series. In both Palceosyops and Hyrachyus those tubercles of the upper 

 molars are confluent into two Vs (more or less open, when unworn). In 

 the former the inner tubercles retain their primitive conic tubercular 

 form, but in Palceotherium, Rhinocerus, Lophiodon, Hyrachyus and Ta- 

 pirus they elongate transversely so as to meet the corresponding outer 

 tubercles (now crests) forming the familiar cross-crests of those genera. 

 If the tubercles are alternate, they produce the oblique crest of Palceo- 

 therium, if opposite, the cross-crest of Tapirus. 



If on the other hand the inner tubercles flatten like the outer, on wear- 

 ing, we have the quadricrescentoid type of Anoplotherium and the Rumi- 

 nants. 



But it is important to observe, that the lower types of Quadrumana 

 and Carnivora present the quadrituberculate crown with tendency to 

 flattening of the outer tubercles, as seen in these lowest Ungulata. In 

 the Carnivora the sectorial tooth is produced by the greater flattening 

 and partial confluence of the outer tubercles, and the entire loss of the 

 inner, the "heel" being in the dogs and cats, e.g. their only representa- 

 tive. In the Quadrumanous families, including man, the primitive quad- 

 rituberculate type of molars is preserved, the flattening of the outer 

 tubercles being finally lost. 



It is to be observed that the lines of Ungulata, Quadrumana, and Car- 

 nivora, originate in plantigrade types, a state of things quite predomi- 

 nant among the lower series, or Lissencephala. It is universal in Eden- 

 tata and very usual in Bodentia and Insectivora. The lower forms of 

 Marsupialia and all of the Monotr ernes present it. In the Marsupials, 

 Rodents, Ungulates, and Carnivores we have series whose highest ex- 

 pression is in the most highly digitigrade genera. 



The accompanying diagram is designed to express to the eye more 

 clearly the propositions made above. By comparing it with a similar 

 table published by Prof. Gill (Proceedings of the American Association 

 for the Advancement of Science for 1871, p. 295), a close resemblance 

 between the two may be observed, as well as certain differences. 



I wish to be understood that the genera named in it as ancestors, are 

 to be regarded in the light of types of groups. There is no other mode 

 of explaining the facts, than that in accordance with the law of "homolo- 

 gous groups," i. e. that several genera of one group have undergone simi- 

 lar modification into corresponding ones of a second group.* 



\ * See Origin of Genera, page 79, Prop. V. 



Published, May Gth, 1873. 



