62 Scientific Proceedings, Royal Dublin Society. 



instead of seeds as the dispersal-unit is to increase materially the buoyancy ; 

 while the seeds of 14 of the 29 plants sink at once (probably more tlian 14, as 

 information is not forthcoming with reference to four of the species), in tlie 

 case of branches absence of buoyancy occurs in only 2 species wlien fresh, 

 and 1 species when dry, and the average buoyancy of the branches, whether 

 fresh or dried, is considerably higher than that of the seeds. So tliat 

 this exceptional means of dispersal tends towards a wider distribution by 

 water. 



As regards a second point — the buoyancy of dried branches as compared 

 with fresh ones — it is seen that the effect of drying is usually (in 14 

 cases) to increase to quite a slight extent their power of floating. In many 

 (10) other cases drying actually diminishes buoyancy, sometimes to a great 

 extent. Two remarkable instances of this kind are displayed by Saxifraga 

 umbrosa and Erica mediterranea. In the case of 8. nmbrosa I tested several 

 varieties of the species. While some sank in from 1^ to 6 days when fresh, 

 two others — a native and a garden form respectively — remained afloat for 

 periods of 25 and 40 days. The same specimens thoroughly dried sank in a 

 few days. As regards Erica mediterranea, its buoyancy depends largely (as in 

 the case of other heaths) on air imprisoned in the witliered corollas. Probably 

 under natural conditions of dispersal, where the branches were being tossed 

 about in a river or in the sea, the air would be expelled from or dissolved 

 out of the corollas more rapidly than under tlie comparatively tranquil 

 conditions of an experimental tank. 



Even leaving out of account these two exceptional species (whose inclusion 

 would make the average buoyancy of the fresh branches considerably greater 

 than that of the dried branches), we still find that the average buoyancy of 

 the remainder is but very slightly increased by drying, the average time of 

 floating of the fresh brandies being 3'9 days, and of tlie dried branches 4'2 

 days. It will be noted that this result is much less favourable to the idea of 

 increase of dispersal-efficiency by drying than that obtained by Darwin 

 from a similar series of experiments, and quoted in the " Origin of Species," 

 chapter xii. Darwin does not give a list of the plants lie experimented upon; 

 and in my own case the number of species tested is too small to permit of any 

 generalization. 



