352 
Journal of Agricultural Research voi. xxvn, No. 6 
plats in which the plants theoretically should have been protected by the 
action of fertilizers. The percentage of infection in the plat which had 
received 1,000 pounds of sodium nitrate alone was 80, while in the one 
which had received 250 pounds it was 85. Certainly there seems to be 
no evidence that very heavy fertilization with nitrogen predisposed 
plants to stem rust, nor that acid phosphate and potassium chlorid 
protected them from it. 
There are very striking differences, however, between the degree of 
infection on Haynes Bluestem and that on Iumillo. The average 
percentage of infection on Bluestem was approximately 80, while that 
on Iumillo was about 20. It is quite evident that the resistance of 
Iumillo was not broken down by excessive fertilization with sodium 
nitrate. In fact, there was more rust on the plants in the plats which 
had received no fertilizer whatever than there was on those in the plats 
which had received 1,000 pounds of sodium nitrate. 
The results are particularly striking, as the soil on which the experi¬ 
ments were made was well supplied with nitrogen. It will be noted 
from Table III that the addition of nitrogen to the soil not only did not 
result in increased yields but actually caused a decrease, indicating 
quite clearly that the soil was not deficient in nitrogen. Any nitrogen 
which was added, therefore, was excessive. 
While there was no correlation between kind and amount of fertilizers 
applied and the amount of stem rust which developed, there was a definite 
correlation between fertilization and lodging and yield. (See fig. 2.) 
The degree of lodging was much greater in the nitrogen plats than in 
any of the others and the yields were lower. The lowest yield, 5 bushels 
per acre, was in plat 1, which had received 2,000 pounds of acid phosphate 
and 1,000 pounds of sodium nitrate, and in which the degree of rust 
infection was lower than that in any of the other plats. The highest 
yield, 31 bushels per acre, by a rather remarkable coincidence, was in 
the plat in which the rust infection was 88 per cent, or higher than that 
in any of the other plats. It is a rather striking fact that the yield 
was always inversely proportional to the amount of nitrogen applied. 
It is well known that wheat which has been overfertilized with nitrogen 
is likely to yield poorly. Not only does it often lodge badly but it 
frequently drys out, a phenomenon spoken of by farmers as “ burning 
out.” This burning out was quite conspicuous in the plats fertilized 
with nitrogen. It was supposed that there might be more rust on the 
plants which lodged badly in the nitrogen plats than on those which 
remained erect, but apparently such was not the case. 
It is worthy of attention that the amount of rust on the plants is not 
necessarily an index of the damage which may be done. It seems 
quite likely that much of the injury usually attributed to rust on plants 
heavily fertilized with nitrogen probably is due to the direct effect of 
nitrogen rather than rust. For instance, wheat growing in soil which 
had received 1,000 pounds of sodium nitrate per acre had 80 per cent of 
rust infection and yielded only 7.8 bushels per acre, while that which 
had received 500 pounds of acid phosphate and 250 pounds of potassium 
chlorid had 88 per cent of rust but yielded 31 bushels per acre. It 
will be seen by referring to figure 1 (p. 347) that 88 per cent of infection 
is extremely heavy. The fact that plants so heavily infected with rust 
can produce well simply indicates that the visible amount of rust on the 
plants is no accurate measure of the damage done. It seems clear that 
