422 ISLAND LIFE 



fossil remains. Whenever Ave can trace the past history of 

 any group of terrestrial animals, we invariably find that 

 its actual distribution can be explained by migrations 

 effected by means of comparatively slight modifications of 

 our existing continents. In no single case have we any 

 direct evidence that the distribution of land and sea has 

 been radically changed during the whole lapse of the 

 Tertiary and Secondary periods, while, as we have already 

 shown in our fifth chapter, the testimony of geology itself, 

 if fairly interpreted, upholds the same theory of the stability 

 of our continents and tlie jDermanence of our oceans. Yet 

 so easy and pleasant is it to speculate on former changes 

 of land and sea with which to cut the gordian knot offered 

 by anomalies of distribution, that we still continually meet 

 with suggestions of former continents stretching in every 

 direction across the deepest oceans, in order to explain the 

 presence in remote parts of the globe of the same genera 

 even of plants or of insects — organisms which possess such 

 excejDtional facilities both for terrestrial, aerial, and oceanic 

 transport, and of whose distribution in early geological 

 periods we generally know little or nothing. 



The Birds of Madagascar, as Indicating a Sup^^osed 

 Lcmnrian Continent. — Having thus shown how the distri- 

 bution of the land mammalia and reptiles of Madagascar 

 may be well explained by the supposition of a union with 

 Africa before the gi'eater part of its existing fauna had 

 reached it, we have now to consider whether, as some 

 ornithologists think, the distribution and affinities of the 

 birds present an insuperable objection to this view, and 

 require the adoption of a hypothetical continent — Lemuria 

 — extending from Madagascar to Ceylon and the Malay 

 Islands. 



There are about one hundred and fifty land birds known 

 from the island of Madagascar, of which a hundred and 

 twenty-seven are peculiar ; and about half of these peculiar 

 species belong to peculiar genera, many of which are 

 extremely isolated, so that it is often difficult to class them 

 in any of the recognised families, or to determine their 

 affinities to any living birds.^ Among the other moiety, 



^ The total number of JMadagascar birds is 238, of which 129 are 



