PROBLEMS OF DISTRIBUTION AND THEIR SOLUTION. 179 



quadrupeds, a lack of carnivores, a complete separation from larger 

 areas by deep sea, and, in fact, a full representation of all the condi- 

 tions which suit these insectivores, just as conversely on the continents 

 the conditions are unfavourable to the prosperity and increase of their 

 race. We do not require to connect the Antilles and Madagascar on 

 account of these animals, any more than we need to postulate the 

 existence of a former Pacific land-connection between Asia and 

 America because the camels of the former continent are related to 

 the llamas of the latter. And when we further reflect that Madagascar 

 preserves a mouse nearly related to a New World type, and snakes 

 belonging to a typical American group, we at once note how the 

 principle of seeking to prove the former wide distribution of a race 

 of animals and its modern limitation by geological and biological 

 changes forms the best clue to many of the difficulties of this science. 

 It is a clue, moreover, which is at once originated and supported by 

 the fossil histories of the animals whose distribution is the subject of' 

 remark. 



A third case which has excited the attention of students of dis- 

 tribution is that concerning the past history of the giant tortoises 

 found in the Mascarene and Galapagos Islands the former belonging 

 to the Madagascar group, and the latter being situated 600 miles 

 from the South American coast. Of these tortoises, as Dr. G anther 

 has shown us, three chief groups exist. One of these inhabits the 

 Galapagos, a second occurs on the coral island of Aldabra to the 

 north of Madagascar, and a third, which has become extinct, inhabited 

 the Mascarene group of islands. But our difficulties are lessened in 

 this case which demands the explanation of the existence of appar- 

 ently similar forms in widely removed areas by the knowledge that 

 these tortoises, though apparently related, in reality belong to distinct 

 types, and that, therefore, the necessity for presuming a connection 

 between their distribution thus disappears. The Galapagos tortoises 

 may be presumed to have come from the American continent ; and as 

 these animals can survive long exposure to sea, and are tenacious of 

 life, their own conveyance or that of their eggs, on driftwood for 

 example, is an hypothesis involving no great demands upon a scientific 

 imagination. The Mascarene tortoises may have similarly been con- 

 veyed from Africa ; and there is no greater difficulty, therefore, in 

 accounting for the detached existence of these great reptiles, than in 

 explaining how their more diminutive kith and kin, belonging, like 

 the giant tortoises, to different groups, have acquired such an extensive 

 range over the earth's surface. 



Indeed, the case of the tortoises may serve to remind us of that 

 of Hassans, an animal formerly regarded as a kind of weasel or civet, 

 but shown conclusively by Professor Flower to belong to the racoons 

 of the New World. Bassaris, however, inhabits California, Texas, 



N 2 



