CH. XVI] ENDEMISM AND DISTRIBUTION: GENERA 179 



only, or between these two classes, are the same, and that they 

 are the same whether we consider the areas occupied by the 

 species of the genera, or the numbers of species in the genera 

 themselves. The same type also appears in the population of a 

 country by its flora, whether some of it is endemic or not. In 

 all cases of distribution, whether it be distribution by areas 

 occupied — geographical distribution or distribution in space — 

 or by numbers of species in the genera — evolution or distribution 

 in time — the distribution seems to have been determined largely 

 by time. If age alo7ie were operative^, one would get much the 

 same distribution as at present exists, when one allows for geo- 

 logical happenings, and the action of barriers. Among these 

 latter, of course, ecological barriers are of great importance, but 

 the general evidence goes to show that their action is principally 

 negative, like that of physical barriers. 



Just as with species, endemic genera have been regarded as 

 (1) locally adapted — a view which has largely died out, especi- 

 ally since it was realised how difficult it would be to find anything 

 to which such a list of genera as those given above for New 

 Zealand (p. 171) could be adapted, and a view upon which it is 

 impossible to explain such an arrangement of genera in order of 

 size as we have just seen to be the rule; (2) as survivals; and 

 {3) as in general new genera beginning life as such. 



As islands have always been regarded as the typical location 

 in which to look for endemics — species, and still more genera — 

 we may do well to consider them. 



Now if the endemic genera of islands be in reality survivals, 

 one would expect that they would at least show a tendency to 

 belong to families that are small or of broken distribution, i.e. 

 such families as we have been accustomed to look upon as more 

 or less moribund. And in any case, one would not expect the 

 great bulk of them to belong to the large and "successful" 

 families. If, on the other hand, Age and Area hold good, they 

 should be found to occur upon islands (provided the connection 

 was mainly by land) in proportions not dissimilar to the pro- 

 portionate sizes of existing families. 



In order to test this question thoroughly (135). I have added 

 up from my Dictionary (1 ) all the endemic genera of all the islands 

 in the world, (2) all the endemic genera of West Australia, South 

 Africa, and Brazil, three areas very rich in endemics, and with 

 much variety of habitat, (3) all the genera confined to Australia, 

 1 I.e. if the average speed of dispersal of a species were constant. 



