16, 6 Trelease: The Growth of Rice 621 
nitrogen (N) content of the fertilizer mixture is apparently 
effective in determining certain differences in growth, the con- 
clusions are carried somewhat further than the experimental 
results justify. It must first be shown that the sulphate ion 
(SO.-) did not account for the improved growth. Indirect 
evidence that the sulphate ion did not determine the growth is 
furnished by the fact that increased yields were not obtained 
with increased proportions of potassium sulphate. Following 
this rather indirect line of reasoning, it is apparently safe to 
conclude that the ammonium ion (NH,*) in the fertilizer mix- 
ture was responsible for the increased growth. Since the ele- 
ment hydrogen of the ammonium radicle could hardly be sup- 
posed to bring about marked improvement in plant growth, it 
might then be concluded that the element nitrogen (N) was 
the important material in the fertilizer mixture. But the ele- 
ment nitrogen is not always beneficial in promoting plant 
growth. Indeed, several studies have shown that nitrogen in 
the nitrate (NO;-) form is of but little use for rice plants.® 
Thus, the importance of the element nitrogen should not be 
given especial emphasis without, at least, qualifying statements 
regarding the chemical combination in which the nitrogen exists; 
and, in general, it seems advisable to avoid the older assump- 
tion that the element nitrogen is important to plants. To 
illustrate this point an extreme example may be cited: Potas- 
sium cyanide (HCN) would never be regarded as efficacious in 
promoting plant growth. In view of these considerations, it 
seems best to state the main result of this study in specific terms, 
as follows: Ammonium sulphate, in the present experiments, 
brought about a marked improvement in the growth of rice 
plants in soil cultures, while primary calcium phosphate and 
potassium sulphate had little or no influence upon the plants. 
Comparison of plant data.—It seemed of interest to sum- 
marize the different kinds of quantitative plant data in a way 
that would allow a ready comparison between the relative values 
obtained with each culture. This has been done by means of 
the graphs of fig. 5, which is somewhat similar in arrangement to 
fig. 4. For each of the graphs the ordinates represent relative 
growth data, while the abscissas give the cultures arranged in 
the order of the magnitude of their grain yields, from highest 
to lowest. The uppermost graph, representing the relative 
yields of grain, has a rather uniform downward slope, from left 
to right, from 100, the highest value, to 18, the lowest. The 
® See Kellner,(6) Kelley,(5) and Espino.(3) 
