. 
16, 6 Trelease: The Growth of Rice 615 
grain yields were not obtained when potassium sulphate made 
up more than 0.3 of the total salt value, nor when primary 
calcium phosphate comprised more than 0.5 of the total osmotic 
value of the salts supplied, nevertheless, high yields do not 
appear to have been determined by the proportion of either 
one of these two salts. 
The highest grain yield was obtained with culture R1C8, and 
the same culture gave highest top yields. This culture, it will 
be remembered, is characterized by having the highest propor- 
tion of ammonium sulphate tested in the series, and by having 
the lowest proportions of potassium sulphate and primary cal- 
cium phosphate tested. The remarkable improvement in grain 
yield resulting from proper application of fertilizer salts is shown 
by the fact that the control cultures, which received no addition 
of mineral nutrients, gave a. yield of grain only 18 per cent 
as high as that obtained with this best fertilized culture, R1C8. 
Low grain yields were secured with practically the same sets 
of salt proportions as those that gave low top yields. As seen 
in fig. 3, low yields of grain were produced in the cultures de- 
noted by the left-hand margin of the triangular diagram. Cul- 
ture R6C3, as in the case of top yields, is apparently excep- 
tional, since it gave low grain yields, although it occupies a posi- 
tion of the triangle where medium yields are indicated. It is 
apparent that low grain yields are correlated with low propor- 
tions of ammonium sulphate, which supplies the element ni- 
trogen. With the exception of culture R5C1, low grain yields 
were obtained with all cultures in which ammonium sulphate 
formed only 0.1 of the osmotic value of the salt mixture. Cul- 
ture R3C2, having 0.2 due to ammonium sulphate, and culture 
R6C3, having 0.3 due to this salt, also gave low grain yields. 
The proportions of the other two salts, potassium sulphate and 
primary calcium phosphate, do not appear to be related to the 
low yield of grain, since low yields were obtained with salt 
mixtures throughout the entire range of proportions of these 
two salts. 
Lowest yields of grain were obtained with three cultures— 
R1C1, R2C1, and R3C1. As already mentioned, the first and 
third of these cultures also produced lowest top yields. The 
salt mixtures added to all three of these cultures contained the 
lowest proportion of ammonium sulphate tested, and low pro- 
portions of primary calcium phosphate, but contained high pro- 
portions of potassium sulphate. A very slight, but probably not 
significant, improvement resulted from the addition of these 
