16, 6 Trelease: The Growth of Rice 623 
lower graphs in this figure show the corresponding data of 
relative weight of tops, relative number of panicles, and relative 
height of tops, respectively, plotted with the same abscissas. 
The general slope and degree of fluctuation of the lower graphs 
give an idea of the degree of correspondence of the other kinds 
of data with those of the grain yield. By comparing the two 
upper graphs, it will be observed that in a general way they 
correspond in slope. The uppermost graph, showing grain 
yields, falls in approximately a linear manner from the value 
of 100 to 18, since the order of arrangement is that of grain 
yield. The graph below this, for top yields, is somewhat irreg- 
ular, but has the same general downward slope as the graph 
for grain yields; in fact, the uppermost graph represents very 
well the smoothed graph that would be drawn from the graph 
of top yields. It may be concluded, therefore, that in a general 
way yields of grain and of tops were proportional to one an- 
other; both were affected in a similar way, and to about the 
same degree, by the same alteration in the various additions of 
fertilizer salts. 
When the next graph, showing the relative number of pani- 
cles, is compared with the graph for grain yield, a smaller 
degree of correspondence is seen; still, the two graphs cor- 
respond in general form at least. Though the variations are 
numerous and in some cases great, the number of panicles is 
in a very general way proportional to the yield of grain. 
The graph at the bottom of the page, representing the height 
of tops, fluctuates considerably, but differs from the graph for 
grain yield mainly in the fact that its slope is much less steep. 
The graph for height of tops shows a general downward slope 
from left to right; but the total range in relative values lies 
between 100 and 57, whereas in the case of the other graphs 
here considered the range is from 100 to approximately 20.. 
Evidently, the various salt mixtures had a much less-pronounced 
influence upon the height of the plants than upon the number 
of panicles, the weight of the tops, or the yield of grain. This 
suggests that increased size in rice plants might be more defi- 
nitely shown by an increase in the number and size of the stools 
than by an increase in the stem or leaf height. 
General considerations.—The plant and the soil in which it 
grows form an exceedingly complex physico-chemical system, 
in which the processes are determined by many variable condi- 
tions. The development of scientific agriculture requires care- 
ful investigation of the general principles determining the rates 
