1076 
Journal of Agricultural Research 
Vol. XXVIII, No. 11 
was still growing vigorously and no flower buds had appeared. Commencing on 
December 20, plat two was illuminated the same as plat one and in a few days the 
flower buds which had appeared began to show signs of being stunted. Many 
of these flowers never opened. All the plants in this plat were stimulated to 
further growth and no more flowers appeared until the period of illumination was 
shortened. The use of artificial illumination was discontinued on January 13, 
and on January 21 the hemp in plat two again showed flower buds. The appear¬ 
ance of buds in plat one occurred at this time also. Thus by merely giving the 
plants a little longer period of light each day it was possible to delay the time of 
flowering a month. By giving a longer exposure to light than was given in this 
experiment it is easily possible to delay blooming even more than this so that 
hemp can be made to flower in 1 month, 2 months, or 3 months from the time of 
planting, as the necessities of the case may require. 
The differences in the sex ratios of the plants in the different day lengths may 
be the result of so many other causes that a serious consideration of the dis¬ 
crepancies from the standpoint of the effect of relative length of day and night 
can not be entertained. It is true that the excess of carpellate plants was greatest 
in the 5-hour exposure and decreased more or less regularly as the length of daily 
exposure to light was increased, but when the factors of death and differential 
germination of seed are considered, such differences cease to have any significance. 
If it could be assumed that all of the seedlings which died were staminate, the 
differences in sex ratios could be accounted for, but such an assumption does 
not seem justifiable here. Since the germination of the seed was only about 65 
per cent, there is considerable chance that the excess of carpellate plants was 
caused by a differential germination of the seed. The sex-ratio differences could 
also be accounted for in each lot on the assumption of an unconscious selection 
of seedlings at the time of potting, but since all of the seedlings were saved, this 
factor does not affect the sex ratio of the total number of plants used in the experi¬ 
ment. This ratio was 117:100. 
The effect of relative length of day and night upon flower development was one 
of extent rather than character. Plants which received less than 7 hours of light 
daily produced very few flowers but practically all of these were normal. This 
result is somewhat surprising since the appearance of hermaphrodite flowers on 
hemp which is grown in the greenhouse during the short days of winter has been 
reported by Tournois ( 8 ) and by Schaffner (5, 6 , 7 ) who have observed that 
either sex may revert to the other under such conditions. Although the writer 
has observed similar phenomena in plants grown in the greenhouse, cases of such 
sexual modifications were very rare among plants grown in short days in the dark 
house. All the plants were examined very carefully several times but only two 
cases of such spontaneous “sex reversal” were found among the 650 hemp 
plants used in the experiment. One of these, and by far the more pronounced of 
the two, occurred on a staminate type which was growing in the normal length 
of day for that time of year. This plant produced a few hermaphrodite flowers 
on one of its branches. The other case was the appearence of a few abnormal 
staminate buds on a carpellate plant in the 10-hour exposure room. Only a few 
abnormal anthers were produced and these dried up before any pollen was liber¬ 
ated. These were the only cases of “sex reversal” among plants grown in the 
dark house and not otherwise treated. It seems quite clear, therefore, that a 
short exposure to light is not in itself the determining cause of sexual modification 
in hemp. 
