328 TROPICAL NATURE, AND OTHER ESSAYS. 



There are also indications hardly less clear, of some 

 communication between India and Malaya on the one 

 hand, and Madagascar on the other ; but as these indi- 

 cations depend chiefly on resemblances in the birds and 

 insects, they do not imply that any land connection has 

 occurred. If, as seems probable, the Laccadive and 

 Maldive Islands are the remains of a large island or 

 indicate a western extension of India, while the Sey- 

 chelles, with the shallow banks to the south-east and the 

 Chagos group are the remains of other extensive lands 

 in the Indian Ocean, we should have a sufficient approxi- 

 mation of these outlying portions of the two continents 

 to allow a certain amount of interchange of such winged 

 groups as birds and insects, while preventing any inter- 

 mixture of the mammalia. 



The presence of some African types (and even some 

 African species) of mammals in Hindostan appears to 

 be due to more recent changes, and may perhaps be 

 explained by a temporary elevation of the comparatively 

 shallow borders of the Arabian Sea, admitting of a land 

 passage from North-East Africa to Western India. 



There remains to be considered the supposed indi- 

 cations of a very ancient communication between Africa, 

 Madagascar, Ceylon, Malaya, and Celebes, furnished by 

 the occurrence over this extensive area of isolated forms 

 of the Lemur tribe. The anomalous range of this group 

 of animals has been thought to require for its explanation 

 the existence of an ancient southern continent which has 

 been called Lemuria, but a consideration of all the facts 

 does not seem to warrant such a theory. Had such a 

 continent ever existed we are sure that it must have 

 disappeared long before the Miocene period, or it would 



