420 ISLAND LIFE 



Now these ancient African mammals are Lemurs, Insecti- 

 vora,.and small Carnivora, chiefly Yiverrid^ ; and all these 

 groups are known to have inhabited Europe in Eocene and 

 Miocene times ; and that the union was with Europe 

 rather than with America is clearly joroved by the fact that 

 even the insectivorous Centetidse, now confined to Mada- 

 gascar and the West Indies, inhabited France in the Lower 

 Miocene period, while the Viverridse, or civets, which form 

 so important a part of the fauna of Madagascar as well as 

 of Africa, were abundant in Europe throughout the whole 

 Tertiary period, but are not known to have ever lived in 

 any part of the American continent. We here see the 

 application of the principle which we have already fully 

 proved and illustrated (Chapter IV., p. 60), that all ex- 

 tensive groups have a wade range at the period of their 

 maximum development ; but as they decay their area of 

 distribution diminishes or breaks up into detached frag- 

 ments, which one after another disappear till the group 

 becomes extinct. Those animal forms which we now find 

 isolated in Madagascar and other remote portions of the 

 globe all belong to ancient groups which are in a decaying 

 or nearly extinct condition, while those which are absent 

 from it belong to more recent and more highly-developed 

 types, which range over extensive and continuous areas, 

 but have had no opportunity of reaching the more ancient 

 continental islands. 



Anomalies of Distribution and How to Ex2:)lain Them. — If 

 these considerations have any weight, it follows that there 

 is no reason whatever for supposing any former direct 

 connection between Madagascar and the Greater Antilles 

 merely because the insectivorous Centetid^e now exist only 

 in these two groups of islands; for we know that the 

 ancestors of this family must once have had a much wider 

 range, which almost certainly extended over the great 

 northern continents. We might as reasonably suppose a 

 land-connection across the Pacific to account for the camels 

 of Asia having their nearest existing allies in the llamas 

 and alpacas of the Peruvian Andes, and another between 

 Sumatra and Brazil, in order that the ancestral tapir of 

 one country might have passed over to the other. In both 



