DISCONTINUOUS DISTRIBUTION IN A FEW 

 MAMMALIAN GROUPS. 



By Thomas F. Dreyer^ B.A., Ph.D. 



(Received for publication, October 22, igi8.) 



As our knowledge of living^ animals and of their fossil fore- 

 runners increase, our classification of them becomes ever more 

 complex. If this classification were merely to serve the purpose 

 of a catalogue, it would no doubt serve its purpose excellently ; 

 but it pretends to be more than such a catalogue. It aims at being 

 as nearly natural as our knowledge allows in that every individual 

 group is meant to contain only species having the same origin, i.e., 

 only monophyletic species. 



A good example of the methods of systematists, and one that 

 will clearly illustrate the above, is the change made in the 

 orientation of the Solenodons of the West Indian Islands. The 

 genus Solcnodon is represented by two species, one in Cuba and 

 one in Hayti, and these bear such striking resemblances to a 

 Mascarene group of Insectivores, the Centetid.?e. that 

 they were formerly included in this family. But if this 

 family is to include both the West Indian and the 

 Mascarene forms, it is to be supposed that these forms 

 are derived immediately from a common ancestor. Now, 

 if the two localities of their occurrence had been close 

 together, they would, I venture to say. have remained comfortably 

 coupled together ; but since they are so widely separated from 

 each other geographically, it had to be supposed that they are 

 " both survivors oi an ancient group of Insectivores, extinct else- 

 where." With this natural bias established, it was easy to find a 

 number of variable characters to sur«Tx>rt their separation, so that 

 now they rank as separate families, i.e., their common ancestor 

 has been pushed back by one step into the distant past. 



This procedure is sanctioned by long usage, and the separa- 

 tion, phyletically, of two such geographically widely separated 

 genera as Solcnodon and Centetis into entities of a higher order 

 than, the genus is almost certainly correct ; and it is of course best 

 to be conservative and to push back the theoretical connection by 

 the smallest possible degrees. It is, however, conceivable that 

 the innumerable twigs formed in this way may completely obscure 

 the larger branches and even the main trunk. The present method 

 seems to me to aim at defining the number of phyletic leaves and 

 to limit further enquiries to the petioles of these leaves, but to 

 take it as understood that these petioles will, if followed down 

 their twngs, end in one or other of the branches which has been 

 labelled as some mammalian order. 



What I propose to do is to attempt to find data which will 

 indicate whether or not the accepted mammalian orders originated 

 from the main mammalian stem as single or multiple branches. 

 I will not attempt to make the enquirv detailed : that, in fact, would 

 be impossible. I want to limit mvself to Creodont mammals 

 (Lemurs, Primates, Carnivora and Insectivora), and -these I want 



