372 BOTANICAL GAZETTE 2 [may 
rather difficult. I believe, however, that for the present purpose 
these experiments will serve as the best illustration. 
Before starting an experiment, the wheat grains were always 
sterilized: R, an Erlenmeyer flask (of 1oocc. for the experiments 
with 70 seedlings, tooocc. when 200 seedlings were used) was pro- 
vided with a rubber stopper and glass tubing which could be 
detached at D. The bottom of the flask was covered with about 
an inch of glass wool, and the flask and tubing were then sterilized 
in the autoclave. The wheat grains were placed in a small cheese- 
cloth sack, immersed for three minutes in a concentrated aqueous 
solution of chloroform, and shaken to free the seeds from adhering 
air bubbles. As much as possible of the chloroform solution was 
removed, and the contents quickly emptied into the sterile flask. 
Dry, carefully filtered air was then drawn through the flask for 
at least 24 hours until the seeds were perfectly air dry, so that no - 
trace of chloroform could remain in the flask. Sterile water was 
then added through one of the tubes; these were then connected 
at D. Material treated in this way very rarely developed any 
growth of fungus during the course of the experiment, nor did it 
show any differences in growth as compared with untreated seeds. 
It is highly improbable, therefore, that the seed coats were pene- 
trated by the chloroform solution. It was found that the wheat 
could also be sterilized by means of ultra-violet light. This is an 
exceedingly convenient method, but as it was found that the sub- 
sequent growth of the seedlings was somewhat affected, the method 
was not used for these experiments. 
The deionizing apparatus used consisted of a brass tube (Y) 
five feet long and one inch in diameter. Into this was concen- 
trically placed an iron rod, the same length and one-quarter inch 
in diameter, and held by fiber supports. The tube and rod were 
attached to the opposite poles of a series of batteries of 50 volts . 
and 800 amperes. To one end of the brass tube was connected a 
glass tube containing cotton, to act as a filter. The deionizing 
apparatus was connected at J with E. 
Blank experiments were, of course, swiss run in order to test 
the apparatus. In experiments 1, 2, 4, and 5 it will be noted that 
the night periods are longer than the day periods. It is conceivable 
