610 
Philippine Journal of Science 
1920 
measurements are given in the fifth column of Table 3; each 
one represents the average of the two measurements made for 
each culture on this date. To facilitate comparison the values 
are expressed as relative numbers, on the basis of the largest 
average considered as 100; but the actual value of the largest 
average is given in the table, in parentheses. It may be sup- 
posed that the values give an approximate idea of the vegetative 
growth that had been made by the plants at the time of flower- 
ing, although of course they do not show the number of stools. 
This method gives only a somewhat rough approximation of the 
size of the plants ; a more accurate method for comparing vege- 
tative growth in such cultures might show the average stem 
height and the total leaf area for each culture. The relative 
height values have been classified in Table 3 into three groups, 
characterized as high, medium, and low, following a method sim- 
ilar to that used by Shive. The highest nine values are desig- 
nated as high, indicated in the table by the letter H following 
the value; since two other values were the same as the lowest 
of the nine, these also are designated as high. Similarly, the 
lowest nine values are designated by the letter L. Medium 
values are without special designation in the table. The 
highest relative value is shown in bold-faced type. The relative 
values obtained with each of the three control cultures, and also 
their average, are given at the bottom of the column. 
This division of the yield values into three large groups allows 
ready comparison to be made between these values and the 
salt proportions tested, and eliminates complexity that may 
result from small variations, due to unknown or uncontrolled 
conditions, such as hereditary variability of the plants. Only 
general conclusions, based upon large differences in plant meas- 
urements, may be drawn from the rather limited series of data 
here available. 
The air-dry weights of tops are shown in the sixth column of 
Table 3, where they are presented as relative average values 
derived from two plants, classified as high, medium, and low, 
as in the case of the height values. The actual yield, in grams, 
of the culture giving the highest yield of tops is shown below 
the relative value (100) for that culture; actual yields for any 
of the other cultures can of course readily be calculated from 
these data. In order to bring out relations between the top 
yields and the proportions of the fertilizer salts employed, these 
data have been plotted on the triangular diagram of fig. 2, 
