PHENOMENA. OF CONVERGENCE. 29 



quently. The Linnsean conception of things, to- 

 gether with the idea of type, and finally the more 

 recent idea of homology as agreement by reason 

 of derivation, have, however, been the cause that 

 approximations in general, especially within the 

 given ' natural group,' were more or less neglected, 

 or, at least, that merely surmised homologies 

 such as the breathing apparatus of the lung-snails 

 turned out to be convergences. Within the 

 group of mammals we meet with a very evident 

 series of convergences which have already been 

 spoken of above. Still, all circumstances con- 

 sidered, convergence can explain only the smallest 

 portion of the phenomena. That the Marsupials 

 should show agreements with certain orders of the 

 higher mammals is, at least, in part approxima- 

 tion. That they form an unity among one another 

 is another thing. Why are cats and dogs classed 

 together as beasts of prey ? Why are pigs, oxen, 

 and deer classed as hoofed-animals ? In short, 

 setting aside the above instances, in the other cases 

 of agreement convergence is improbable a priori, 

 and the explanation of all the other, the main por- 

 tion, must be looked for in the doctrine of descent. 

 It contains the greater amount of probability, a 

 probability which often borders upon certainty. 



