HYGROSCOPICITY 167 



hygroscopic variation of the seed in proportion to the degree 

 of loss of the protection it owes to its coats. The same 

 indications are presented below in the tabulated results for 

 five kinds of seeds belonging to the four very different families. 

 It will, however, be noticed that there is a very considerable 

 difference between seeds in respect to the restraining influence 

 of the seed-coverings on the hygroscopic reaction. In watery 

 seeds like those of the Horse-chestnut, Broad Bean, and 

 Canavalia ensiformis, it is evidently considerable, whilst with oily 

 seeds like those of Anona reticulata it may be so small as to be 

 almost negligible. In such experiments it is requisite that 

 the seed should have completed its drying process, or at all 

 events that it should be drying very slowly. 



With reference to the differentiation between the coats 

 and the kernel in this respect, such a contrast would not 

 indicate a factor of much importance in determining the 

 nature of the range, since, although the inner layer of the 

 coats, being often looser in texture than the outer skin, might 

 be expected to be more hygroscopic, it would not be exposed 

 directly to the air, as in the case of the detached seed-coverings 

 in an experiment. However, the results tabulated below show 

 that the seed in its coverings is less hygroscopic than either 

 the coats alone or the bared kernel. 



Comparison of the Hygroscopic Range of Permeable Seeds 

 with and without their coats. 



(The range is the variation stated as a percentage of the average weight. The seed 

 and its parts were in each case tested together.) 



^sculus Hippocastanum (Horse-chestnut) 

 Anona reticulata (Custard Apple) 



Canavalia ensiformis .... 



Citrus decumana (Shaddock) 

 Faba vulgaris (Broad Bean) 



Note. — These results do not represent the full range in the case of the Horse-chestnut, 

 the experiment being a short one. 



