76 WILLIS : GEOGKArHlCAL DISTKIBUTION 



What I claim, tlien, for tlie accejitauce of the theoiy of 

 mutation, is that it greatty siiupUfies the problems of geographi- 

 cal distribution. On tiie old theory we had to assume a vast 

 amount of destruction, and it was a matter of extraordinary 

 difficulty to trace relationships, and to decide the origin and 

 past history of any plant. 



In studying the distribution of any plant about the globe, 

 it ig evident that we shall have to take many factors into 

 consideration. For instance, plants of open country* will 

 spread more readily than plants of forests ; endemics of very 

 restricted range abound in the Ceylon forests, for instance. 

 Frequent crossing with new arrivals will perhaps tend to pre- 

 vent a newly fornled mutation from establishing itself, so that 

 we shall have to take into account the fertilization mechanisms. 

 Connected with this is the fact that I have pointed out, in a 

 paper on the flora of Ritigala,* that the endemics are mainly 

 among the plants that can only rarely arrive there, i.e., among 

 those with very poor distribution meclianisms ; whereas, as I 

 have elsewhere shown, in outlying oceanic islands they are 

 more often among the plants with good mechanisms, which will 

 enable them occasionally to arrive. 



* Willis, The Flora of Ritigala, an isolak'd mountain in the North- 

 Central Province of Ceylon. Ann. PeFad., ML, 1906, p. 271. 



