1 62 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



genera are plants of this sort. In New Zealand 65 per cent (206 

 species out of 315) of the endemic element is woody, but only 

 42 per cent of the species of the non-endemic element. In Pata- 

 gonia and Fuegia, trees and shrubs comprise 52 per cent (347 out 

 of 667) of the species of those genera which are practically con- 

 fined to this region and which may fairly be called "endemic" 

 in it, as opposed to 13 per cent (120 out of 920) of the species of 

 the non-endemic genera. The same fact is observable in South 

 Africa, where 70 per cent (3,296 out of 4,686) of the species of 

 the "endemic" genera are woody, but only 42 per cent (1,369 

 out of 3,298) of the species of the non-endemic types. 



In the north temperate land area, therefore, the endemic 

 (presumably most ancient) element in the flora has a decidedly 

 higher proportion of herhs than does the non-endemic element, 

 whereas in the southern hemisphere, it has a decidedly higher 

 proportion of ivoody plants. Does this indicate that herbs are in 

 general a more ancient type of vegetation than trees and shrubs 

 in the former area but a less ancient one in the latter? Such a 

 conclusion is opposed to the evidence recently brought forward 

 from various botanical fields^ in support of the view that the 

 herbaceous type has invariably been deri^•ed from an earlier vege- 

 tation which was prevailingly woody, and the writer believes that 

 the explanation of the preponderance of herbs in the endemic 

 flora of the north temperate zone lies not in the antiquity of this 

 growth type but rather in the high degree of rapidity with which 

 plants belonging to it may undergo evolutionary change. 



We must first of all distinguish clearly between two main types 

 of endemic plants: on the one hand, the isolated and localized 

 survivors of once much more widely spread types, which we may 

 call "relict" endemics; and on the other hand, those plants which 

 owe their endemism to the fact that they have never spread 

 beyond the actual region of their evolutionary origin, and which 

 may there be named "indigenous" endemics. A list of the 

 endemic genera of any large area, arranged according to natural 

 relationships, is usually easy to divide into these two categories, 

 for genera which stand well apart from the rest, without near 

 relatives, are doubtless isolated surxixors of once much more 



' Sinnott, E. W., & Bailey, I. W. The origin and (lis[)crsal of herbaceous Angio- 

 spcrms. Ann. Bot. 28: 547-600. O 1914. 



