FIRST APPEARANCE OF RUSTS IN THE SPRING. 57 



disappeared, and from April 8, a period of about four weeks, it was 

 impossible to find a single spore. On May 6 new leaves began to 

 show a diseased appearance. On May 13 open pustules were found 

 in abundance. He states further that he has found by experiments 

 that in the cooler weather of spring the incubation period following 

 inoculation with ureclospores is lengthened to between three anoL four 

 weeks, and this explains the existence of a period with no rust after 

 the first attack. 



The winter leaves die in the early spring and with them the winter mycelium, but 

 not until it has produced uredospores which inoculate the new leaves. Then follows 

 a period of incubation which may be lengthened more or less according to the tem- 

 perature and other conditions in the spring. 



This, then, is one way to account for the spring appearance of 

 Puccinia rubigo-vera in the middle Northwest. The other way is the 

 infection of the grains from spores carried in the air from the South. 

 It has been shown in this paper that P. rubigo-vera winters in the 

 vegetative uredo stage m Kansas and Nebraska, producing spores on 

 the winter grains in March and April. This is true, also, of a large 

 part of the Atlantic Coast States, particularly Maryland and Virginia. 

 These spores may be carried by the winds farther north during the 

 months of April and May, becoming generally distributed. Inocula- 

 tion may then be cumulative — i. e., spores may fall on fields from time 

 to time during several weeks in April and May without any apparent 

 effect. Then, when moisture and temperature conditions become 

 just right, a general, though sparing, outbreak may take place over 

 large territories within a few days. After this first outbreak spores 

 will be present in abundance and the attack may spread rapidly. 



Puccinia graminis in the Middle Northwest makes its first appear- 

 ance from two to three weeks after the appearance of secidia on bar- 

 berries. From this it may be. argued that the first infection always 

 comes from the aecidiospore. Barberries are grown as hedges and 

 ornamental shrubs here and there in the Middle Northwest, and cer- 

 tainly are the cause of more or less local secidiospore infection, but 

 the appearance of P. graminis over large territories within a few days 

 is to be accounted for in other ways. The wintering of the uredo 

 in the North and also wind-blown spores from southern fields in the 

 progressive northward march of this rust are, perhaps, the most 

 important agencies in its first appearance, just as in the case of 

 P. rubigo-vera. Another possibility is the transfer of the uredo of 

 P. graminis from the wild grasses, especially Eordeum jubatum, 

 Agropyron repens, and A. tenerum. Viable uredos have been found 

 in these grasses as late as April 15, and undoubtedly occur even later 

 in the season. That the graminis form on these grasses may affect 

 wheat has been demonstrated; but the new crop of uredospores on 



216 



