1913] SHULL—SEMIPERMEABILITY OF SEED COATS 175 
show the imbibition curves for distilled water, and for molecular, 
two-molecular, and four-molecular solutions of NaCl as the seeds 
were shifted from one strength to another at three-day intervals. 
The close agreement of the constant portion of the curves in each 
strength of solution should be noted. The character of the curve 
during the last two days of each three-day interval indicates that 
the salt does not enter in appreciable amounts. Comparison with 
the type of curve given by SCHRODER (33, p. 189) for wheat shows 
that this membrane is at least as efficient in excluding NaCl as the 
coats of the Gramineae. 
By appropriate chemical methods it has been shown that no 
passage of the salt through the membrane occurs. It was necessary, 
however, to test seeds singly, since one finds rather frequently seeds 
which have invisible defects in the coat which allow the passage 
of salts. 
Many other chemical substances, usually in molecular solutions, 
were used to determine the range of selective semipermeability. 
The permeability or non-permeability to each substance was judged 
by the amount of imbibition water taken up from such solutions of 
acids, bases, and salts. Anyone who investigates the subject will 
be convinced that this means of determining semipermeability is 
sufficiently accurate for all ordinary purposes. The results, in 
percentage of increase over air-dry weight by imbibition, are given 
in table I. ; 
From this table it is seen that the nitrates as a class, and 
especially silver nitrate, penetrate the coats. Iron sulphate slowly 
enters, while copper sulphate does not. The penetration of copper 
sulphate can be detected, if it occurs, by removal of the testa and 
examination of the embryo. Occasionally as high as 20-40 per cent 
of the seeds show slight localized penetration of CuSO, through 
defects in the coat too slight to be seen on microscopic examination 
of the dry seeds. Because of these defects the figures given for any 
salt are approximations only; but in case there is general per- 
meability it is easily recognized by the amount of imbibition. 
Since 60-80 per cent of the seeds show no penetration even after 
prolonged soaking in CuSQ,, and since they retain their vitality 
perfectly, notwithstanding the very highly poisonous character of 
