PERMEABILITY AND CLASSIFICATION 103 



The difficulties a seed may have to contend with before 

 it enters its prolonged rest-period under the protection of 

 its impermeable skin were brought under my notice in the 

 case of Diodea reflexa whilst specially studying the habits of Dioclea 

 this leguminous climber in its home in the mountain forests 

 of the Grand Etang in the island of Grenada. The rainfall 



here is heavy even for the tropics, averaging: 14.0 or i CO ofimperme- 

 . , , : .. . 5 5 ... . ? ability in its 



inches in the year, and very moist conditions prevail in the seeds. 



densely shaded woods. The seeds are freed by the decay 

 of the pod ; and if the detachment of the pod from the 

 parent takes place under normal conditions, the typical 

 impermeable seed is produced, such as is at times stranded 

 by the Gulf Stream with coats intact on the shores and islands 

 of Western Europe. This tree -climber often favours a 

 station on the banks of streams and lakes, so that the seed's 

 opportunities of reaching the open sea are numerous. 



But if, as frequently happens, the pods are prematurely 

 detached during the prevalence of heavy rains and strong 

 winds, that is to say, before the drying and shrinking stage 

 of the pod and the enclosed seeds is complete, then the seeds 

 usually fail to finish the process, and, being moist and permeable, 

 they either pass on into the germinating condition, a result 

 favoured by the prevailing humidity of the forests, or decay 

 and die. These fallen pods lying on the wet, sloppy ground 

 in the deep shade of the forests readily take up water, and 

 the seeds, as just remarked, soon swell up and either die or 

 germinate. On one occasion I noted that out of 30 seeds 

 obtained from 10 fallen pods lying sodden on the ground in 

 the depths of the forest, 10 were rotting, 10 were germinating, 

 and 10 were in a resting condition ; but of these last, 7 or 8 

 were permeable, with the shrinking process markedly incom- 

 plete. Even the two or three impermeable seeds were far 

 from typical in their appearance. It was only occasionally 

 that a typical impermeable seed was found that seemed fit 

 for the Transatlantic passage in the Gulf Stream. Such, 

 then, are some of the difficulties with which a seed may have 



