36 
Journal of Agricultural Research 
Vol. XXI, No. i 
place at this time occurs under the conditions of plot 8 two or three days 
earlier than in plots i to 4. This change may mark the end of the addi¬ 
tion of new cells, the abandonment of the central cells for active deposi¬ 
tion, the entering of the second stage of starch formation where numerous 
new grains are started, or it may be some fundamental change the sig¬ 
nificance of which is not yet recognized. 
The kernels from plots 5, 6, and 7 show departure from the rate of 
deposit in those of plot 8 at later dates. There is a departure in all cases; 
and even in plot 7, irrigated as late as July 29, the kernels are significantly 
greater in size and their maturation date is postponed by the irrigation. 
The status of the water supply on succeeding days is more apparent 
in figure 18. The course of the water content of kernels from plot 1 is 
here seen to be very regular. From this plot four samples were taken 
instead of two, and the more favorable irrigation caused a more uniform 
development in the plants. If some allowance is made for plot 2, the 
effect of time of irrigation is quite apparent. By August 1, plot 2 has 
begun to suffer from lack of water, while 3 and 4 are enjoying the full 
benefit of a sufficient supply. By August 7, maturation is well advanced, 
the kernels from plot 5 being equal to those from plots 3 and 4 in water 
content and those from plots 6 and 7 showing obvious benefit from the 
late irrigation. 
The trend of the percentage of ash is very uniform. As can be seen 
from the table, there is a gradual daily decrease from about 8 per cent at 
flowering to less than 1 per cent at maturity. When expressed in milli¬ 
grams per kernel, the ash content bears a direct though not exact rela¬ 
tionship to quantity of the water used. This is apparent in figure 19, 
where the ash per kernel is averaged in 3-day periods. There is some 
evidence that the application of water retards the deposit of ash for 
two or three days following irrigation. The curves of July 31 and August 
6 are strongly suggestive of the availability of water. It is to be sup¬ 
posed that plots 2 to 4 would use more water than those irrigated later, as 
they had the use of water for a longer period. Plots 3 and 4 received 
much less water than plot 1, but they contained fewer culms and the 
water per culm seems to have been sufficient for an equal deposit of ash. 
The nitrogen determinations were not satisfactory. As may be seen 
in the table, the percentages varied from day to day more than seems 
natural At first this is due to extremely small samples, but the varia¬ 
tion is just as great at maturity. No definite conclusions can be drawn, 
but from figure 20 it appears that a plentiful water supply may hinder the 
deposit of nitrogen. As much nitrogen apparently is deposited in the 
kernels in plots suffering for water as in those amply provided with it. 
Of course the percentage of nitrogen is higher in the smaller kernels. 
