190 BOTANICAL GAZETTE [SEPTEMBER 
based upon the figures presented in this discussion. What the 
initial capillary or surface force would be if all hygroscopic water 
could be eliminated without changing the fundamental nature of 
the embryo must be left purely to conjecture; but the figures would 
probably exceed those which RopDEWALD found for starch. He 
states that dry starch on swelling develops a pressure of 2523 
atmospheres (32, p. 227). At present we have no adequate means 
of measuring such large forces. 
III. Discussion 
The results here recorded have a bearing upon many problems 
in widely separated fields of research. It will be impossible to con- 
sider every phase of the subject, and only the more important 
matters will be considered here. 
The general occurrence of non-living semipermeable membranes 
in plant structures, especially as seed and fruit coats, is of the 
greatest interest. A large amount of work has been done upon 
seeds with coats intact, especially upon seeds showing delayed 
germination, in attempts to stimulate protoplasm to activity. 
FISCHER (16) used many different kinds of acids, alkalies, and salts 
on seeds. LEHMANN (26), BECKER (6), and many others have used 
Knop’s solution, etc., as stimulants for resting seeds. In BECKER'S 
paper especially the seeds used by him, mostly Compositae, show 
such peculiar irregularities of behavior in germination that one 
cannot escape the conviction that the coats are responsible for much 
of it. Ifsemipermeability of protective structures is as common as 
now appears to be the case, much of the work done with salts and 
other substances acting through the coats will lose considerably in 
its significance. The very salts which are assumed to enter the 
seed may be excluded by the testa of the seed. No safe conclusions 
can be drawn from studies in which the physical characters of the 
testa are not definitely known. This discovery only emphasizes 
what the writer has said elsewhere (34) regarding the necessity of 
removing seed coats when the properties of embryos are being 
investigated. The testa has physical and chemical characters which 
may enable it to modify greatly any factor entering into germination 
behavior, and the effects of these characters must be known before 
