PERMEABILITY AND CLASSIFICATION 99 



equally in colour, size, and weight. The contrast, though 

 marked, is not quite so great as between typical seeds of the 

 impermeable and permeable groups, because it is not possible 

 to use a test that would completely exclude the other type 

 from a sample without invalidating the experiment, and one 

 has to select merely by the external characters. 



The essential characters of the A type as a permeable 

 seed are its hygroscopicity, its relatively large water-contents, 

 and the inability to increase its weight when exposed in a 

 broken condition to the air, ordinary hygroscopic variation 

 being sufficient to explain most of the small increase in weight 

 (r6 per cent.) noted in the table. On the other hand, B as 

 an impermeable seed is practically non-hygroscopic, has a 

 relatively small water-percentage and a counterbalancing large 

 capacity of absorbing moisture when exposed in the broken 

 condition to the air. The behaviour after being subjected 

 to a temperature of 100° C. is also distinctive, the permeable 

 seeds failing by 5 per cent, to regain their original weight 

 after an exposure of some days to the air ; whilst the 

 impermeable seeds ultimately exceed their original weight by 

 about I per cent. But, as observed in the note to the table, 

 the full extent of the absorptive capacity after heating was * 



not determined. These results, however, bring the two 

 types of seeds into contrast sufficiently well, when we reflect 

 that each sample of seeds probably contained a small percentage 

 of the other kind, less easily distinguished than usual by their 

 external characters. 



When seeking for an explanation of these two sets The clue to 

 of seeds in the same plant, one has not far to look, the thetw?"° 

 uncompleted process supplying the clue. This we recognise ggedg-jj^hg 

 in a check to the shrinking and drying process that ushers same plant 

 in the rest-period; and Professor Ewart's surmise (p. 197) check to the 

 respecting the differences between Acacia seeds of the same proc"s?^ 

 species, that the impermeable seeds are smaller " because they 

 are drier," goes towards the root of the matter. With the 

 larger, pale-coloured seeds the shrinking process has been 



