542 A NATURALIST IN THE PACIFIC 



results were produced in my experiments in 1892 on Cornish seeds, and 

 in 1904 on Devonshire seeds. In the great contrast between their floating 

 capacity in sea-water and in fresh water the seeds of this plant defy the general 

 rule that seeds that float a long time in sea-water float also a long time in 

 fresh-water. According also to Sernander the seeds float a long time in 

 the sea. He says that the capsules float, but since they ultimately dehisce 

 this could scarcely be efficacious in dispersal. Floating portions of the 

 plant also aid in its dispersal, according to the same authority (p. 174). 

 The plant forms great extended masses on the pebbly shores of Spitzbergen 

 (Ekstam, p. 28). 



Beta maritima. Thuret found that the dried fruits of this plant 

 floated only two or three days in sea-water ; whilst in my sea-water 

 experiments the freshly gathered fruits floated only one or two days. 

 Sernander speaks of them as fitted for dispersal from shore to shore ; but 

 this could only be to a limited extent. Martins and Thuret established 

 by experiment the capacity of the germination of seeds of other species of 

 Beta after long immersion in sea-water ; and the first seems to imply that 

 those of Beta vulgaris float for many weeks ; but I am inclined to think 

 an error lies here. 



Cakile maritima. The fruits, even after long drying, float, as a rule, 

 only a week and sink within ten days, the same results being afforded in 

 my sea-water experiments in 1893 on fruits from Cornwall, and in 1904 on 

 fruits from Devonshire. The fruits are common in the stranded drift on 

 the north coast of Devonshire and may often be seen germinating there. 

 They are also frequent in the beach drift of the Scandinavian coasts 

 (Sernander, p. 156). 



Crambe maritima. The fruits were kept floating by Sernander more 

 than 13 days (p. 165). Martins implies that they floated for 45 days. 

 Darwin says that they germinated after 37 days' immersion in sea-water, 

 but does not specify that they floated all the time. 



Crithmum maritimum. The ripe fruits readily separate into the two 

 carpels, which are very buoyant and float in sea-water for months. In my 

 experiments, 95 per cent, remained afloat after 10 months. It is remark- 

 able that whilst in sea-water the spongy covering of the carpels retains its 

 vitality, in fresh-water it becomes sickly and decays and the carpels lose 

 their floating power, so that they float weeks instead of months as in the 

 sea-water. The carpels are extremely light, being washed up in the spray 

 and blown up by the wind amongst the lightest of the stranded drift of the 

 Devonshire beaches. In a moderate gale they are often blown off the 

 beach and up the cliff-faces. 



Convolvulus soldanella. From 40 to 50 per cent, of the seeds float 

 after six months in sea-water, and about 30 per cent, float after eighteen 

 months, retaining up to the end their germinating capacity. Sernander 

 implies that the plant is not found on the Scandinavian coast to the north 

 of Nissum Fjord in Denmark. It is known, however, to occur in the south 



