360 GEOGRAPHICAL DISTRIBUTION. Chap. XI. 



per diem) ; on this average, the seeds of -^^ plants 

 belonging to one country might be floated across 924 

 miles of sea to another country ; and when stranded, if 

 blown to a favourable spot by an inland gale, they would 

 germinate. 



Subsequently to my experiments, M. Martens tried 

 similar ones, but in a much better manner, for he 

 placed the seeds in a box in the actual sea, so that they 

 were alternately wet and exposed to the air like really 

 floating plants. He tried 98 seeds, mostly different 

 from mine ; but he chose many large fruits and likewise 

 seeds from plants which live near the sea ; and this 

 would have favoured the average length of their flota- 

 tion and of their resistance to the injurious action of the 

 salt-water. On the other hand he did not previously 

 dry the plants or branches with the fruit ; and tins, as 

 we have seen, would have caused some of them to have 

 floated much longer. The result was that ^f of his 

 seeds floated for 42 days, and were then capable of ger- 

 mination. But I do not doubt that plants exposed to 

 the waves would float for a less time than those pro- 

 tected from violent movement as in our experiments. 

 Therefore it would perhaps be safer to assume that the 

 seeds of about y 1 ^ plants of a flora, after having been 

 dried, could be floated across a space of sea 900 miles 

 in width, and would then germinate. The fact of the 

 larger fruits often floating longer than the small, is in- 

 teresting; as plants with large seeds or fruit could 

 hardly be transported by any other means ; and Alph. 

 de Candolle has shown that such plants generally have 

 restricted ranges. 



But seeds may be occasionally transported in another 

 manner. Drift timber is thrown up on most islands, 

 even on those in the midst of the widest oceans ; and 

 the natives of the coral-islands in the Pacific, procure 



