328 NATURAL SELECTION [CH. 



regards as forced upon them, so to speak, by that of the vege- 

 tative organs, "without any reference to advantages or dis- 

 advantages to be derived from it in the performance of the 

 functions of the floral organs themselves 1 ." He believes that 

 the dorsiventrality was first impressed upon the vegetative 

 organs, whence it spread, as it were, to the reproductive regions, 

 affecting the bracts, spathe and flower. The stamens most 

 commonly exhibit it, but, in the cases in which it is carried 

 furthest, the gynaeceum also conforms to it. The zygomorphyof 

 the flower develops concurrently with a tendency towards 

 anemophily and autogamy, whereas in most families it is associ- 

 ated with adaptation to entomophily. Willis looks upon the 

 zygomorphy of the more specialised Podostemaceae as a cha- 

 racter without survival-value, which thus cannot owe its pre- 

 sence to Natural Selection, but which originates as an inevitable 

 corollary to the dorsiventrality of the vegetative organs. In fact, 

 he even goes so far as to regard the zygomorphy of the flower 

 as a positive disadvantage, whose influence the plant seems to 

 attempt to neutralise. "However dorsiventral the flower be- 

 comes it still stands erect as long as it possesses a stalk, and 

 when at last we come to the forms without the stalk we find 

 the flower curving its ovary and stamens so as to get them as 

 erect as possible. It seems as if the flower were, so to speak, 

 struggling against the dorsiventrality to the last 1 ." 



The aspect of the zygomorphy of the Podostemad flower 

 upon which Willis dwells with the greatest emphasis, is its 

 apparent uselessness. This is one of the points which he 

 brings forward to show that, though the family as a whole is 

 probably more completely transformed than almost any other 

 from the average mesophytic type, the great variety in morpho- 

 logical structure presented by the individual members cannot 

 be explained as due to adaptation to their individual surround- 

 ings. For, though the family has become differentiated into at 

 least thirty genera and one hundred species of the most varied 

 morphological structure, the conditions under which they live 

 1 Willis,J. C. (1902). 



