182 ENDEMISM AND DISTRIBUTION: GENERA [pt. ii 



have 195 endemic genera in 43 families that also occur in the 

 islands of Indo-Malaya, and only 19 in the 16 families that do not. 

 They have 187 in 39 families that occur on the African islands, 

 and 27 in 20 families that do not. 



Incidentally, the close correspondence of these curves shows 

 that it is all but certain that the floras of the world, in the mass, 

 must have been distributed by land connections, and at any 

 rate those of the bulk of the islands, though some of the far out- 

 lying ones, with few endemic genera, probably were oceanic. 



Just as the endemic species belonged to the large and "suc- 

 cessful" genera in greater proportion, so the endemic genera 

 belong to the large and "successful" families, and only a very 

 few indeed to endemic families. An analysis of the above table 

 of 1582 endemic genera of islands shows that 1150 of them, or 

 72-6 per cent., are found in the 40 largest families in the world, 

 which only contain 70-6 per cent, of the total genera in the 

 world, i.e. these families contain rather more than their proper 

 proportion of endemics. The remainder occur in another 110 

 families, leaving 141 which are not represented upon islands by 

 any endemic genera at all. The largest of these latter families is 

 the Chenopodiaceae with 86 genera, and the whole number only 

 contain 890 genera, or 6 per family, against 77 per family for 

 those which have island endemic genera. The proportion of 

 endemic genera diminishes from top to bottom of the table 

 (cf. 135, p. 509). 



The further out, and more isolated, the island is, i.e. in general 

 the more ancient the date of its peopling with plants, the more 

 do the endemic genera tend to belong to the larger families. If 

 one divide the 150 families that possess them upon islands into 

 75 larger and 75 smaller, one finds that in Madagascar 62 of the 

 families with endemic genera belong to the larger, 18 to the 

 smaller. In New Zealand the proportion is 16/4, and in the 

 Haw^aiian Islands 13/1. 



If endemic genera were really largely relics, one would expect 

 that there would be a fair number of endemic families, but, as 

 a matter of fact, these are fe^v and small, and of the five that 

 are found only upon islands (Chlaenaceae, Balanopsidaceae, 

 Corynocarpaceae, Lactoridaceae, and Cercidiphyllaceae), the 

 largest is upon the largest island (Madagascar) that is also a 

 good way out from the mainland. 



Putting together all the facts about endemic genera that have 

 been given above, and which show that in the mass they behave 



