Adaptation in the Tristichacem and Podostemacece. 543 



Now if these plants were really adapted to their habitat, one of the first 

 things one would expect to find in them would be some arrangement to 

 enable the seeds to cling to the rocks upon which they find themselves shed. 

 But there is only one species in which this is the case, F. metzgerioides, 

 mentioned above. All the rest have capsules which open in dry air and 

 shed the seeds upon the rocks. When a shower of rain conies the 

 mucilaginous layer of the seed swells, and as it dries it attaches the seed 

 firmly to the rock. This has been looked upon as an adaptation for clinging 

 to the rock by many people who have forgotten that when again wetted, 

 as for instance by the rise of the water in the rainy season, the mucilage 

 again softens and the seed -washes away. The seeds once afloat have but a 

 small chance of arriving at suitable growing places, for if dropped in quiet 

 water they will not grow, and if carried to rocky places it is very unlikely 

 that they will catch in anything to form a place for germination. Seedlings 

 I found in Ceylon to be extremely rare, excepting only in F. metzgerioides. 

 Only when a seed gets caught in a crack in a rock, or in a hole in the old 

 thallus, does it get any chance to grow. As soon as it germinates it produces 

 root hairs to fasten it to the substratum, but even these are often not pro- 

 duced quickly enough, and I have found, in trying to germinate seeds in 

 the Botanic Gardens in Bio de Janeiro, that even when the cotyledons are 

 open, and a number of root hairs attached to the rock, they may yet be 

 washed away. My experience as yet with germination trials leads me to 

 suppose that it takes at least from 500 to 1000 or more seeds to give three 

 or four seedlings, and of these, perhaps, one may come to maturity. This 

 being the case it is impossible to say that the order — except in the one 

 species F. metzgerioides — shows any adaptation in this, one of the most 

 essential things, one would think, in which it might, with its wealth of 

 variety in structure, be adapted. Even in F. metzgerioides it is by no means 

 certain that the indehiscent fruit is an adaptation. This form is the most 

 highly dorsiventral of the whole family, and the peculiar feature of the 

 indehiscent fruit may be merely another expression of the increasing dorsi- 

 ventrality which runs throughout the Podostemacese. 



After a consideration of all the features of these orders, then, we come to 

 the conclusion that any and all of the adaptations that there may be in them 

 are to be found in T. ramosissima, the least modified of all. As we pointed 

 out at the commencement of the paper, in this species (and the same features 

 appear throughout the two families) these features are — (1) a great and regular 

 development of secondary shoots upon creeping, closely attached roots ; 

 (2) the development of haptera or clinging organs to attach the plants to 

 the rock ; and (3) the absence of intercellular spaces. Beyond these there 



2 T 2 



