358 
Journal of Agricultural Research voi. xxvii, No. 6 
Stem rust first began to appear on June 16, after which it developed 
and spread rapidly and uniformly until checked by the hot, dry weather. 
The growth of the wheat plants also was checked by the drought, the 
plants in the plats treated with nitrogen plus acid phosphate being 
injured most severely. As there was some drying and premature ripen¬ 
ing, the rust did not become as abundant as it had in 1915. It will be 
noted that in 1916 the period between sowing and harvesting was only 
85 days, while in 1915 it was 116 days. The range in percentage of 
stem-rust infection on Haynes Bluestem in 1916 was from 65 to 70 per 
cent and on the resistant hybrid it was from 36 to 45 per cent. The 
narrow range may have been due partly to the fact that the plants 
ripened prematurely. It is interesting to note that, as in 1915, the 
highest percentage of stem rust on Haynes Bluestem was not in the 
nitrate plats. In fact, the average percentage of infection was lower in 
those plats which had been fertilized with nitrogen alone than in the con¬ 
trol plat and in those which had received acid phosphate and potassium 
sulphate. The average percentage of stem rust also was slightly higher 
on the hybrid in the phosphate and potassium than in the nitrogen 
plats. It is perfectly clear, therefore, that excessive fertilization with 
nitrogen failed entirely to predispose either the susceptible Haynes 
Bluestem or the resistant hybrid to stem rust; nor did phosphorus and 
potassium protect them. Whatever slight effect there may have been 
appeared to be just the opposite of what one would expect. There was 
a very definite lack of correlation between fertilization and the develop¬ 
ment of stem rust (see Table VI and fig. 3), a fact for which weather 
conditions probably were partly responsible. 
About a week after the plants had headed, the weather became very 
hot and dry. The growth of the plants, especially in the nitrate and 
phosphate plats, was severely checked and many of them began to dry 
up and crinkle. The crinkling was fairly uniform and could not be 
attributed entirely to the effect of fertilizers, although it was consistently 
worse in the nitrate plats than in any others. As would be expected, 
“ burning out/’ due to heavj fertilization with nitrogen, is aggravated 
by hot dry weather. As “burning out” really means premature ripen¬ 
ing, nitrogen actually may cause a decrease in the amount of rust in 
years when the weather is hot and dry. But in years when the growing 
conditions are favorable it may increase rust infection by prolonging the 
growing period. The degree of shriveling of seed was in direct propor¬ 
tion to the amount of drying, and the yield, in general, was inversely 
proportional to the amount of shriveling. For instance, the acre yield 
of Haynes Bluestem was less than 1 bushel in all of the plats in which 
the degree of drying was 80, while on all of those in which the drying 
was less than 60, the yield was over 1 bushel per acre. 
The yields in 1916 were deplorably low, of course, as they were gen¬ 
erally throughout the hard red spring wheat region. For instance, in 
1915, the average acre yield in Minnesota was 17 bushels, in South 
Dakota 17, and in North Dakota 18.2 bushels. In 1916, they were 7.5, 
6.3, and 5.5 bushels, respectively (35, p . 59j). This was attributed prin¬ 
cipally to the terrific stem-rust epidemic, the most destructive since 1904. 
It is interesting, however, to note that the average percentage of stem 
rust on all plats in 1915 was considerably higher than it was in 1916. 
But the yields in 1915 were very much higher than they were in 1916. 
In 1915 die lowest yield in the commercial fertilizer plats was 5 bushels 
