182 BOTANICAL GAZETTE [SEPTEMBER 
evidence that it ever spread to the wheat. Hence it appears 
that the forms upon these two species are distinct. 
The wintering of P. graminis as mycelium in plant tissues in 
North Dakota is extremely doubtful, as there are no winter cereals 
and the uredo stage does not appear upon the grasses until very 
late in the spring, when they are quite large. To test the supposi- 
tion, however, that the fungus might pass the winter in occasional 
plants under shelter and produce a new outbreak of uredo the fol- 
lowing season, I placed heavily rusted plants of Agropyron tenerum, 
A. repens, Hordeum jubatum, Elymus virginicus, and E. canadensis 
in large pots three successive falls (1904-1907) and transferred them 
to the greenhouse where they remained until summer, but no fresh 
uredo pustules ever appeared on any of them. Furthermore, 
in collecting data relative to the appearance of P. graminis upon 
cereals and grasses in the spring of 1905, a piece of low meadow 
containing Hordeum jubatum, Agropyron tenerum, and A. repens, 
which was flooded by the city in the winter and used as a skating 
pond, was carefully observed, but no uredo pustules appeared here 
until they were found on the grains and grasses elsewhere. 
The origin of spring infections has frequently been attributed 
to over-wintered uredospores, although this is merely a hypothesis. 
In order to determine with some accuracy the duration of the ger- 
minative capacity of the uredospores of P. graminis, the following 
experiments were made. Bundles of rusty straw which ha 
stood in the shock until late in the fall of 1904 were placed on the 
ground. Others were tied to trunks of trees, and some stored 
in the attic of one of the college buildings, where the temperature 
was below freezing but much warmer than the outside atmosphere. 
Rusty wheat straw was also put in manila envelopes and in test 
tubes, and these laid in pasteboard boxes on the ground. All the 
material placed upon the ground was covered by snow the greater 
part of the winter. To still vary the conditions, test tubes of rusty 
wheat straw were attached in an inverted position to stakes out- 
side, 2-3 feet above the ground, while packets were buried in ice 
at the ice house and others kept in the laboratory. From the 
middle of September to the following July germination tests of the 
uredospores were made once a week from all these sources. At 
