158 Mammals inhahiting Continental and Oceanic Islands. 



the other hand, it is quite out of one's power to understand 

 tlieir present distribution, except on the old grounds of inde- 

 pendent creation, without postulating a much closer connexion, 

 than Mr. Wallace appears disposed to admit, between the 

 island groups in the Indian Ocean at a comparatively recent 

 period. 



The above-noted facts lead to the following deductions, 

 namely, that, in the first place, a chain of islands sufficiently 

 close to allow of the passage, not only of the representatives 

 of the genera of insectivorous bats referred to, but also of the 

 large slow-flying frugivorous bats, must have existed between 

 Madagascar and Australia ; and, secondly, that, at a later 

 period, a temporary connexion of a similar kind lay between 

 Madagascar and India. 



It may be said that such connexion with India would also 

 permit of the introduction of insectivorous bats ; but it must 

 be again remembered that volant insects, on which such bats 

 feed, are very scarce in oceanic islands, while tree-fruit, which 

 forms the food of the frugivorous species, is usually abundant. 

 Bearing these facts in mind, it is necessary to suppose that 

 the islands, assumed to have formed the high road for the 

 insectivorous bats between Africa and Australia, must have 

 been sufficiently large to support volant insects ; while, on 

 the other hand, a chain of small coral islands, placed not too 

 far apart, and provided only with a few fruit-bearing trees, 

 Avould have sufficed for the passage of the frugivorous species ; 

 and it appears more than probable that it was by such a chain 

 that the ancestors of the flying-foxes of India were introduced 

 into that continent. 



While considering the former geographical relations of these 

 regions it maybe well to refer to an apparently most remark- 

 able instance of discontinuous distribution which long puzzled 

 zoologists — namely, the supposed close relationship between 

 the Insectivora of Madagascar and the West Indies, depending 

 upon the presence, in the islands of Cuba and Ilayti, of one 

 or more species of the genus Solenodon, which was said to 

 belong to the family Centetida3, known elsewhere in Mada- 

 gascar only. Mr. Wallace partly gets over the difficulty by 

 referring to supposed remains of species of this family in 

 France in strata believed to be of Lower Miocene age ; but 

 this was hardly necessary, for, as I have lately pointed out*, 

 Solenodon belongs to a family less closely related to Cente- 

 tidse, than the Hedgehogs (Erinaceidse) are to the Moles (Tal- 



* ' Monograph of the Insectivora, Systematic and Anatomical/ pt. i. 

 p. 87. 



