1920 Frye; on Plant Migration 395 



area, during one night, sowed seeds of eat tail on the wet roof of a house 

 over 1% miles (2 km.) away, so thickly that there was a seed for about 

 every 3 square inches of roof. 



Scirpus, Eleocharis, Sparganium and Oenanthe are usualh^ considered 

 as plants having their seeds distributed through their floating on water. 

 These might drift ashore as the level recedes thus sowing the area ; or 

 birds might carry them on their feet; or the sunken seeds might lie on 

 the bottom and grow when uncovered. 



Bidens is plainly fitted to have its seeds carried by animals. Surely 

 it could not be thus carried as if sown, over a large area of fairly thin mud. 

 One is inclined to look for some other method of dispersal. 



How Polygonum and Juncus are distributed is somewhat uncertain. 



Tests were made on the floating power of seeds. Many different kinds 

 were gathered, specially those which could be secured from the newly 

 overgrown areas. The naturalhl attached parts were left on, so the float- 

 ing power of the seed in the condition in which it might fall into the watei 

 would be determined. These were put into dishes of water and left to 

 see how long they would float. The dishes were shaken every day to 

 simulate the natural roughness of water, and to be sure that the seeds did 

 not float by mere surface tension. Of those plants mentioned above as 

 most abundant on a wet flat, all except Juncus floated for 30 days, at which 

 time the experiment was discontinued. Juncus hufonius grew mostly near 

 the former level and along streamlets, from which one might expect dis- 

 persal by animals following the streamlets, or by flowing water. 



Since the seeds of Bidens floated for at least 30 days, it is quite 

 likely that water is one of the instruments of its dispersal. In fact Typha, 

 Bidens and Epilobium are on an equality with water-dispersed plants, and 

 have the wind or animal dispersal in addition. 



A test of the seeds of many dry land plants sliowed that almost 80 

 per cent floated for at least 30 days, while only about 5 per cent sank 

 within the first 5 minutes. Floating seeds are very much more common 

 than is usually supposed. When seeds are fitted for distribution b}' 

 animals, by wind or by some other means, it is no evidence that they may 

 not also be distributed by water. Ravn's contention that shore plants have 

 floating seeds might be extended by saying that the same is true of most 

 plants. One should not conclude that dry-land plants are wanting along 

 wet shores because the seeds have not the same advantages of distribution 

 as those of shore plants; indeed many, and perhaps most of them, liave 

 better means. Those also fitted for wind dispersal have a powerful aid 

 in getting to the water. 



