PERMEABILITY AND CLASSIFICATION 91 



As regards the differences in the variable group, it is not 

 often that the distinction between the two types is con- 

 spicuously evident to the eye, since similar-looking seeds, as 

 Professor Ewart remarks, may differ greatly in these qualities. 

 Usually the differentiation is made in the course of experi- 

 ments on the capacity of seeds to withstand immersion in water. 

 However, there are cases, like that of Entada polystachya, where 

 the two types of seeds are easily distinguished, one small, dark- 

 coloured, impermeable, and non-hygroscopic, the other large, 

 light-coloured, permeable, and hygroscopic, and both equally 

 capable of reproducing the plant. Then again, Mr Crocker, in 

 the Botanical Gazette for October 1906, has shown that Axyris 

 amaranthoides has " dimorphic " seeds, the one kind flattened, 

 winged, permeable, and germinating readily, the other rounded, 

 relatively impermeable, and only germinating after consider- 

 able delay. Professor Ewart points out that in the course 

 of his experiments on more or less impermeable leguminous 

 seeds, as with the Acacias, these two types of seeds were often 

 differentiated, the seeds that readily swelled being larger than 

 the " hard " seeds that required the employment of artificial 

 means to produce swelling, though both attained the same size 

 when swollen for germination. 



In the following tables 1 have arranged the species of seeds 

 on which experiments were made by me, amounting to about 

 105, into the three groups : 



(1) Impermeable group, containing only plants where the 



seeds are all normally impermeable. 



(2) Variable group, including plants with both permeable 



and impermeable seeds. 



(3) Permeable group, comprising plants with only per- 



meable seeds. 



Since the general question of the frequency and distribu- 

 tion of the quality of impermeability among seeds has been 

 already dealt with in Chapter III on a far more extended 

 basis than my own materials would afford, I will merely 

 confine my remarks on these groups to a few general observa- 



