Adaptation in the Tristichacece and Podostemacece. 545 



Cladopus . . 

 Willisia . . 

 Griffithella 

 Farmeria . . 



Java. 



W. India, Burma. 

 W. India. 



S. India, Ceylon. 



The African species are as yet too little known to make it possible to give 

 their distribution ; the genera to which they really belong have yet to be 

 made out in many cases, but are almost certainly all, or nearly all, different 

 from the American and Indian. 



In other words, the only widespread genera are the non-specialised ones, 

 while the more specialised the genus, on the whole, the less is its area of 

 distribution. The non-specialised forms live everywhere with the specialised, 

 and are every whit as well suited to the positions, which show no differences 

 in general conditions of life. This result may well be compared with that 

 which I traced for the Dilleniaceae* in a previous paper, being exactly parallel 

 with that. 



Or, again, take the well-known fact that in the Podostemacese the species 

 are usually very local in distribution, as has frequently been pointed out.f 

 The best known instance is Castelnavia, where in the same river, the 

 Araguaya, some of the species differ at every cataract. Now, in the case of 

 different species of a genus on land, it has been customary to say that they 

 have been evolved to suit different mixtures of the conditions of life, and 

 the absence of many species of one genus in most plants of 'still water 

 has been put down to their uniformity of conditions. But here this explana- 

 tion will not hold. The physical conditions of life at all cataracts in the 

 Araguaya are the same, and there is no mixture with other forms of life at 

 all. And this must have been true since the foundation of the family. And 

 yet the Araguaya contains seven species of this one genus, a genus, moreover, 

 which is almost confined to this river, in which there occur besides only one 

 or two species of Oserya and Apinagia. It is another expression of the fact 

 to which attention has often been called,! that isolation, as isolation, 

 favours the production of new species. Why this should be so, we cannot at 

 present say, but the fact remains. These species of the Araguaya are, of 

 course, each endemic to its own few waterfalls, and to them may be applied 



* Willis, " The Geographical Distribution of the Dilleniacefe as illustrating the Treat- 

 ment of this Subject on the Theory of Mutation," 'Ann. Perad.,' vol. 4, p. 69 (1907). 



t Weddell, " Sur les Podostemacees en general et leur Distribution Geographique en 

 particulier," ' Bull. Soc. Bot. France,' vol. 19, p. 50 (1873) ; Goebel, ' Pflanzenbiologischen 

 Schilderungen,' vol. 2, pp. 331, 374 ; Willis, " Studies in the Morphology and Ecology of 

 the Podostemacese of Ceylon and India," 'Ann. Perad.,' vol. 1, p. 450 (1902). 



% Willis, " The Floras of Hilltops in Ceylon," ' Ann. Perad.,' vol. 4, p. 135 (1908). 



