1913] SHULL—SEMIPERMEABILITY OF SEED COATS 171 
cemented over the end of a glass tube fitted with a perforated 
rubber cork, the tube being then filled with mercury, and set up as 
a sort of barometer. Although the coats were supported by strong 
cloth, they were too delicate to withstand the strain, and would 
always burst on being inverted over the mercury cup. Shorter 
columns also were tried. In one case a column of mercury 20 cm. 
in height was sustained for four days without any diffusion of 
atmospheric gas taking place through the membrane. 
Several attempts were made to test the diffusion of oxygen 
through the dry seed coat by means of an apparatus like that used 
by CROCKER (14) to test the diffusion through moist coats. These 
tests ran for only a few hours at a time and the results were negative. - 
No measurable diffusion was noted, but the slightest injury to the 
coat was found-to permit a rapid passage of oxygen. 
Much better results were obtained with the experiments on 
the permeability of dry coats to ether, chloroform, acetone, and 
absolute alcohol. Seeds of Xanthium glabratum were dried for two 
weeks at 4o° C., then put into a desiccator over phosphoric anhy- 
dride at 10 mm. atmospheric pressure for 18 days at 40° C., after 
which they were stored in a similar desiccator for use. 
_ Water-free ether was prepared by treating the best ether obtain- 
able with alcohol-washed KOH in sticks for 12 hours, and then 
distilling it with ribboned metallic sodium in both distilling and 
receiving flask. The sodium was allowed to act in the receiving 
flask until every trace of water had disappeared. 
Five lots of dry seeds were then immersed in dry ether, and kept 
at a constant temperature in a Freas thermostat for various lengths 
of time from 2 to 36 days. On removing them from the ether the 
seeds were carefully dried, soaked in water, the testas removed, 
and the embryos placed in germinative conditions; 87 per cent of 
the seeds germinated, and the young plants were just as vigorous as 
the untreated controls. In those sets of seeds soaked for 16 and 
36 days respectively, 100 per cent of both uppers and lowers 
germinated, and the average growth in length in 6 days was for the 
uppers 9.5 cm., for the lowers 11 cm. 
Seeds carefully dried were also placed in chloroform, absolute 
alcohol, and acetone without the precaution of drying the fluids. 
