122 PODOSTEMACEAE [CH. ix 



not be overlooked that the data are still highly incomplete, for, 

 as a recent writer has pointed out, we probably know only a 

 small proportion of the existing species belonging to these 

 families 1 . It was recorded a decade ago, for instance, that the 

 examination of a few kilometres of a river in Venezuela 

 hitherto unexplored in this respect yielded no less than four 

 species of Podostemaceae new to science 2 . The extremely local 

 distribution of many forms, their anomalous morphology and 

 progressive dorsiventrality, and the great variety of types of 

 structure which they present, offer every incentive to specula- 

 tion. Dr Willis has put forward certain far-reaching theoretical 

 views, based on his study of the group, and to these and related 

 questions we shall return in Chapter xxvn, when we are 

 touching upon the problem of Natural Selection. 



1 Went, F. A. F. C. (1910). 2 Matthiesen, F. (1908). 



