Aug.3. J920 Daily Development of Kernels of Han7ichen Barley 421 



become common. After this date results are not consistent in such short 

 periods as 12 hours. 



The original purpose of the 12 -hour interval was to discover, if possible, 

 whether or not growth occurred during the night. For this reason, the 

 periods are not quite equal. The day period consisted of about 13 hours 

 at the beginning. As the days grew shorter this was reduced slightly. 

 This period was thought to include the hours of effective sunlight. In 

 figure 13 the gains and losses for the day and night periods are indicated 

 graphically. The disadvantage of such presentation lies in the magnifi- 

 cation of the fluctuations. For instance, the night sample of July 26 shows 

 a decrease of 120 points, not because it is smaller than the sample of July 



Fig. ij. — Graph showing dry matter per kernel at 12-hour intervals from flowering to maturity. 



25, but because it is the same size and, therefore, no gain is registered. 

 For the first lo days after flowering the day and night gains appear to be 

 nearly equal. From this time until maturity the day gain is obviously 

 greater. The author has no interpretation to suggest, but there are two 

 facts which may be noted. The night gains are most prominent before 

 starch infiltration has become very active. The temperatures, after the 

 first 10 days, are lower, the first night without gain being recorded on 

 July 26, when the mean temperature first falls to 70° F. It is not known 

 whether these facts have any essential relation to the results obtained or 

 not. During the latter part of the growth period, the variation of 

 individual spikes makes the results inconclusive. 



The significant features of the data on dry-matter content are (i) the 

 long period of daily gains following the completion of length growth, 



