OP CPYLON AND JNDIA. 
455 
would appear to have been usually a dwarfing of the secondary 
shoots, ultimately going so far as to reduce their axes almost 
to nothing, as in the Dicræas and the Farmerias. As this 
dwarfing progressed, the plants would become more and more 
suited to life in shallower (and in general also swifter) water, 
and would spread into regions of shorter vegetative season, 
into the higher regions of the hills, and into the smaller 
streams, as well as into the swifter w^ater of the larger 
streams. In this way they would escape more than ever from 
competition, and perhaps be enabled to evolve in greater 
variety of form than might otherwise be the case. 
The dwarfing at the same time would seem to have been 
accompanied by an ever-increasing dorsiventrality of the 
vegetative structure, the roots becoming more and more 
flattened and thalloid, and taking over more and more the 
functions of assimilation. This dorsiventrality would seem 
to have ultimately re-acted on the floral organs, which became 
more and more zygomorphic, Avith suppression and other 
important phenom.ena accompanying the zygomorphism . At 
the same time the flowers appear to have become more and 
more anemophilous and self-fertilized. Most, if not all, of 
the changes in floral morphology on Avhich the groupings of 
the order are largely based would seem thus to be correla- 
tive or induced, and not to have been the subject of natural 
selectiou, unless in a very minor degree, a conclusion which 
opens up interesting general considerations. 
In conclusion, I wish to thank, in addition to the friends 
mentioned at page 182 of the preceding paper, my wife, 
without whose help this paper Avould never have reached 
its present state of comparative completeness, Professors 
Warming and Goebel, and Dr. Treub, to whom I am indebted 
for references and criticisms. 
Peradeniya, July 15, 1902. 
