542 



Dr. J. C. Willis. On the Lack of 



the most important morphological character in the classification of the 

 order, is a direct result of, or in direct correlation with, that of the vegeta- 

 tive organs, being greater the greater the dorsiventrality of the latter." 

 And again " it seems then not unreasonable to conclude that the dorsi- 

 ventrality of the floral organs has been, so to speak, forced upon them by 

 that of the vegetative organs or by their position upon the latter without 

 any reference to advantages or disadvantages to be derived from it in the 

 performance of the functions of the floral organs themselves. The only 

 demand made upon them, so to speak, has been that they should not cease 

 to set seed." 



A comparison of the characters used for separating the genera of 

 Podostemacese proper, e.g. by Warming in Engler and Prantl's ' Natiirlichen 

 Pflanzenfamilien,' shows at once that practically all these characters, e.g., 

 suppression of the stamens of the upper side of the flower, obliquity of the 

 gynceceum, mode of opening of the spathe, etc., are simply expressions of, 

 and involved in, the degree of dorsiventrality of the flower. As I regard 

 this, as above reiterated, as being due to a large extent, at any rate, to the 

 vegetative dorsiventrality, I have consequently employed vegetative characters 

 in addition to the floral in determining the Indian genera, and in the work 

 which I am carrying out upon the South American forms I propose to do 

 the same. 



By no stretch of imagination can the dorsiventrality of the flowers be 

 regarded as adaptational, and I have elsewhere shown good reason to suppose 

 that in general it follows that of the vegetative organs. But in this paper 

 we have seen reason to believe that that of the vegetative organs also is 

 not adaptational. No gain whatever comes to these plants, in other words, 

 from all their wonderful morphological changes, and the whole differentiation 

 of the two orders into nearly 30 genera with numerous species is almost 

 entirely an expression of the dorsiventrality forced upon them by their 

 plagiotropic growth. The least dorsiventral and least modified species of 

 the least modified genus of the Tristichacese can and do live in nearly all the 

 places occupied by the two families, and the whole differentiation of the 

 families is merely de luxe, and without any adaptational signification what- 

 ever. 



Finally, let us consider the seed. In most of the plants of these orders 

 there are very numerous (200-600) minute seeds, whose outer coat becomes 

 mucilaginous when wetted, but in some the number is much reduced, 

 especially in Farmeria, where in one species there is a dehiscent fruit with 

 about four seeds, in the other, F. metzgerioides Willis, an indehiscent fruit 

 with only two, which germinate in situ. 



