184 BOTANICAL GAZETTE [SEPTEMBER 
means of a pipette, placed upon six slides and examined micro- 
scopically. This process was repeated two or three times a 
week for nearly a month, but no uredospores of P. graminis were 
caught until uredo pustules were abundant on the surrounding 
wheat. 
Further observations were made to determine whether uredo- 
spores are commonly borne very great distances by the wind. 
On a piece of ground one-third to one-half acre in area, which we 
used for breeding rust-resistant wheat, a rust epidemic was produced 
every year. This was accomplished by plowing into the soil 
rusty wheat straw and spraying the wheat repeatedly with aecidio- 
spores of P. graminis tritici Erik. and Henn. It should also be 
mentioned that our original seed, the foundation stock, was obtained 
from the badly rusted crop of 1904. Hence there were present 
teleutospores, aecidiospores, uredospores, and probably infected 
seed. For the present, however, we are concerned only with the 
fact that rust annually appeared upon these plots in great abun- 
dance. In fact it was almost impossible to obtain any plump 
kernels of wheat from plants grown here. During at least two 
summers (1906-1907), when these plots were thoroughly covered 
with P. graminis, there was scarcely any rust on the field plots of 
wheat which lay a short distance north of the infested area and in 
the direction of the prevailing winds, although the latter passed 
over the breeding plots, often causing considerable annoyance 
while I was taking rust notes. The only possible hindrance to 
the passage of the spores was a few rows of shrubs covering a strip 
about 10 feet wide, thinly planted and varying from 6 to 8 feet 
in height, located 20-25 yards from the rust bed. However, 
there was a road about 20 feet wide running north and south 
through the shrubbery and along the west edge of the infested 
area. Hence there was ample opportunity for wind distribution 
of the uredospores, and former experiments have shown that they 
were highly viable during the summer of 1906, yet practically 
no rust appeared upon these neighboring wheat plots. The fact 
that P. graminis does not appear upon wheat in North Dakota 
in the summer until the plants are nearly 2 feet high, several 
weeks after the wheat crop is harvested in Kansas and Nebraska, 
