50 Scientific Proceedings, Royal Dublin Society. 



of buoyancy is determined by the permeability of the outer coverings. 

 Sometimes the air is contained in spongy tissue in the peripheral wrappings, 

 or in the albumen ; sometimes its presence is due to the incomplete filling of 

 the fruit by the seed, or of the seed by the albumen, or by the embryo. 

 Guppy has some interesting notes on these points (loc. clt., pp. 115, 116). 



Although seeds vary considerably as regards the rate at which they sink 

 when heavier than water, or the period during which they retain buoyancy, 

 none are eventually very much heavier than water. And although it is 

 extremely improbable that, when they have sunk, they could be conveyed by 

 currents or other means across arms of the sea or lakes of any depth, they 

 might still be transported for long distances along the beds of rivers, and 

 similarly by currents over shallow sea-bottoms. Professor Oliver tells me 

 that at Blakeney the greater part of the seed-drift takes place along the 

 bottom, ofl'-shore, the seeds being eventually cast up in quantities. 



BUOTAKCY AS DISPLAYED IN THE LEADING NATURAL OeDERS. 



If we study the general table from the point of view of the buoyancy 

 of related plant-groups, we find that floating seeds are characteristic of 

 certain orders, while in others absence of buoyancy is almost universal; 

 in others, again, great variability in this respect exists. For instance, 

 the seeds of Papaveraceae, Cruciferae, Caryophyllaceae, Geraniaceae, 

 Leguminosae, Crassulaceae, Saxifragaceae, Primulaceae, Gentianaceae, 

 Scrophulariaceae, and Juncaceae in almost all cases sink at once ; it is 

 to be noted that in almost all these orders it is the seed itself which is 

 the unit of dispersal. On the other hand, Eanunoulaceae, Compositae, 

 Orobanchaceae, Orchidaceae, Oyperaceae, Gramineae are mostly possessed 

 of buoyancy ; in most of these the unit of dispersal is either a one-seeded 

 indehiscent fruit containing or retaining air, or a seed with a loose test in 

 which air is entangled. Among the orders which display wide variability 

 as regards buoyancy are Rosaceae, Umbelliferae, Rubiaceae, Ericaceae, 

 Boraginaceae, Labiatae, Polygonaceae. 



We shall now look more closely into this point, and at the same time 

 add notes on certain species which claim attention. The results given 

 will be taken from my own table as given above, reference being made 

 to Guppy's more generalized results where they supply additional 



evidence. 



Ranunculaceae. 



Species tested, 26. Almost all possess some floating power — | to 5 days. 

 H. pm-viflorus, Trollius, Eranthis, Aquilegia, and Delphinium alone sink at 

 once. Guppy found B. sceleratus and H. repens to float for 6-12 months; 



