12 AlS^NALS ^'IJW YORK ACADEMY OF 8CIEXCES 



certain oceanic islands, in which it appeared impossible to explain the 

 characters of the mammal fauna upon any reasonable hypothesis involv- 

 ing continental union. 



AYhen it comes to applying the rafting theory to lower vertebrates and 

 invertebrates, Dr. Barbour is quite right in insisting that one should 

 have a special knowledge of the particular habits of each group in order 

 to judge of possible or probable methods of its transportation. Because 

 I lack that especial knowledge and field experience I avoided discussing it. 

 I do not know how far the rafting theory is applicable to account for 

 their distribution, and how far it may be more reasonably explained in 

 other Avays. But I do notice that Barbour appears to consider only the 

 transportation of the adult animals, and says almost nothing about the 

 possible transportation of their eggs. It is just because mammals do not 

 lay eggs that their presence in oceanic islands seems peculiarly difficult 

 to account for, save through former union with continents. Wallace, it 

 will be remembered, for this reason regarded their presence as a depend- 

 able criterion of such union. Like most students of geographic distribu- 

 tion, I started from Wallace's views as a basis, and if I have modified 

 them ill an opposite direction from many of my confreres it is because I 

 found that in certain particulars they did not fit the details of distribu- 

 tion, 23a st and present, in the groups with which I am best acquainted. If 

 mammals laid eggs, and especially if the eggs were numerous and of very 

 small size, I should find it far easier — indeed quite too easy — to account 

 for their presence on oceanic islands; the difficulty would be to account 

 for their general absence. It is because they do not that I was compelled 

 to discuss a rather complex hypothesis to account for their presence in 

 Cuba, j\Iadagascar and elsewhere. There are numerous other accidental 

 means of transport which might be and have been invoked to account for 

 dispersal of lower animals, but the ones I discussed were the only ones 

 that, so far as I could see, would be possible for mammals. And for very 

 large terrestrial mammals these seemed to me physically impossible. But 

 no such mammals closely allied to continental ones are found on any 

 oceanic islands;^ those remotely related are derivable from the much 

 smaller and more generalized ancestors which we find living in older 

 epochs on the continents, and it is these small common ancestors whose 

 dispersal must be accounted for. 



It is the above considerations that underlay the remark to which Bar- 

 bour takes exception on page 2. I do not think his objection is war- 

 ranted, for while I do not question his judgment as to the improbability 



^ There are certain exceptions to this statement — Malta, Cyprus, Crete and Celebes, 

 These exceptions do appear to call for continental union. 



