
182 BOTANICAL GAZETTE [SEPTEMBER 
as the concentration increases, the action is very rapid. On the 
other hand, acids do not dissolve tannins, but rather precipitate. 
them from solution, according to REICHARD. 
The barley grains used by REICHARD did not seem to be as per- 
fectly semipermeable as those used by Brown. This variation in 
semipermeability is attributed to differences in the tannin content 
from differences in the ripening process. He believes that the 
tannin is an ‘‘assimilation” product of chlorophyll, and that in 
unripe seeds the tannin has not been fully deposited in the seed 
coats. Or one can assume that the tannin for some unknown reason 
has failed to reach the proper chemico-physical condition for 
complete semipermeability. 
While it seems entirely possible that the resistance of dry seeds 
to alcohol and other reagents such as ether, chloroform, etc., might 
be related in certain cases to tannin in the seed coats, it does not 
seem probable to me that semipermeability of these membranes 
can be. explained on these grounds. 
There is a considerable amount of tannin in the testa of Xan- 
thium, as shown by treatment with ferric chloride, ferric sulphate, 
potassium dichromate, etc. But the tannin in this case does not 
seem to form so definite a layer as REICHARD reports for Hordeum. 
It seems to be scattered through the tissue in minute granules, 
occurring especially in the outer and middle layers of the coat with 
only very minute quantities in the inner layer. As has been shown, 
treatment of the coats with molecular NaOH, which is a strong 
tannin solvent, according to REICHARD, does not destroy its semi- 
permeable character. For these reasons it is the writer’s opinion 
that the property belongs to the cellulose, rather than to substances 
associated with it in the coat. Further microchemical studies upon 
the seed coat of Xanthium are in progress, and in time it may be 
possible to analyze the conditions which cause the exclusion of 
certain salts and the passage of others through the semipermeable 
coat of seeds. 
4. SEMIPERMEABILITY A WIDESPREAD PHENOMENON 
In addition to the work on Xanthium, a number of seeds in 
various families have been tested. The results show that the 
BY pe ec a 


