APPENDIX 473 



In the Horse-chestnut seed the peculiar drying regime is connected 

 with the closely adhering absorbent tissue of the inner coverings, which 

 readily take up the kernel's moisture, and allow it to escape through 

 the large unprotected scar. The soft, swollen, white unripe seed, 

 when passing into the resting stage, mainly dries through the scar, 

 which occupies about one-third of the surface, and is not covered 

 by the ultimately impervious skin that protects the rest of the 

 seed. Long after this outer skin has undergone browning and 

 hardening during the shrinking stage, the scar remains moist and 

 gives off water. The relative impermeability of the brown skin 

 of the resting seed is well displayed by allowing the seed to soak 

 in water for a day or two. On examination the inner coverings 

 are then found to be saturated with water, which oozes out through 

 the scar. 



It would, however, seem that the peculiarity in the drying rdgime 

 of the Horse-chestnut seed is in degree rather than in kind ; and indeed 

 it probably represents in an exaggerated fashion the behaviour of a 

 typical permeable seed. It is apparent from the researches of Becquerel 

 that the permeable integuments of peas (Pisum\ lupines, and beans 

 (Faba) after a certain amount of artificial desiccation become impervious 

 to air, except when it is saturated with moisture. This implies a 

 sluggishness during exceptionally dry atmospheric conditions of the 

 coats of permeable seeds ; and it may well be that such seeds in the 

 resting state are most pervious at the hilum or scar. If also we are 

 able to consider seeds rendered impervious to air by artificial desiccation 

 as in process of becoming impermeable to water, one might rank 

 permeability as a quality by default. 



But to return to the subject of the influence of the seed's coverings 

 on its drying in the resting state, it may be remarked that the seed 

 of the Horse-chestnut, on account of the great size of the unprotected 

 scar, is exceptional in this respect. If the coats are able to restrain the 

 hygroscopic reaction of a permeable seed they will also check its 

 surrender of water to the air ; and the capacity first named has been 

 established on pages 71 and 166. Not many of my experiments 

 were made on permeable seeds as soon as they were gathered from 

 the plant ; but I append the results of those on Citrus decumana 

 (Shaddock). Here we notice that the coats restrain the drying of 

 the seed for the first three weeks, after which it assumes a stable 

 weight affected only by the ordinary hygroscopic reaction. The 

 subject, however, will not be pursued here, as there are many factors 

 determining the results, of which the relative importance cannot now 

 be estimated. 



In the reference to this note on page 226 mention is made of the 

 case of Mammea americana ; but as the fruit is drupaceous another 



