60 Scientific Proceedings, Royal Dublin Society. 



I am not prepared to account for these cases of variability, nor to say 

 bow many of the discrepancies will be reduced or ruled out by further 

 experiments on the species in question. Some of them may be due to immature 

 or unsound seed (though trouble was taken to eliminate this source of error). 

 But, as recognized by Guppy, considerable variability does exist in certain 

 species. To determine its limits and its causes, the experiments on the variable 

 species would have to be considerably extended. The fact that Guppy 's 

 buoyancy periods are usually greater than mine suggests that the difference 

 may be due to his (presumably) using salt water, while my tests were mostly 

 made in fresh water. But several of my batches, tested in salt water, gave 

 results differing but slightly from the fresh-water results. 



Buoyancy of Fruiting Branches. 



It is evident that, even if the seeds of a species sink at once, wide 

 dispersal may still be effected if branches or crowns with fruit attached 

 possess a considerable power of floating. This was pointed out long ago by 

 Darwin. Accidents of one sort or another — storms, the subsidence of over- 

 hanging banks, the trampling of animals — occasionally precipitate plants or 

 portions of them into rivers. The buoyancy of branches or fruiting crowns, 

 both fresh and after thorough drying, was tested in the case of a few plants 

 of different habit, with the result shown below. In the table (in which the 

 numerals represent days) the buoyancy of the seeds, and of fresh and dry 

 fruit where a succulent fruit occurred, is added for comparison wlierever tlie 

 information was available. 



