in the Bromes and their Brown Rust . 295 
That it was really the high temperature, and not the 
method of infection, however, seemed clear after the following 
experiments. 
Three series (Nos. 759, 763, and 764) of the same hosts, 
&c., as in the preceding lists, were infected by placing the 
spores in the tip-drops and on the face of the leaf, in water 
only. The temperature was very high, and all failed. 
On the other hand, by adopting the precaution of carrying 
out the infection in a cool cellar, and allowing the incuba- 
tion to take place in a well-lighted but cool area, where 
the temperature did not rise above 25 0 C, I found the results 
quite comparable to those of the previous experiments before 
the very hot weather set in. See Tables XIX and XX, 
and Table XXII, p. 297. Similar successful infections were 
then obtained in a number of other cases, some of which have 
been shown in results already published 1 i and which are 
quite conclusive as to the state of affairs when the external 
conditions for infection are satisfactory. 
10. Discussion of the foregoing Results. 
There seems to be no doubt from the evidence now accu- 
mulated that the Uredospores of Puccinia dispersa are 
capable of germinating upon any leaf of any species of Bromus 
and putting forth the germ-tubes which seek the stomata. 
Here the tip of the germ-tube behaves differently according 
to the species of Bromus attacked and the species of host- 
plant whence the Uredospore has been derived. 
If these species are identical — e. g., if Uredospores derived 
from Bromus mollis find the conditions for germination on 
leaves of Bromus mollis — the tip of the Uredospore swells up 
over the stoma, and soon puts forth a branch into the inter- 
cellular spaces beneath, and there ramifies as a mycelium 
which, in ten days or so, again puts forth pustules of Uredo- 
spores. Even here, however, the success of the mycelium 
depends on circumstances. If the temperature, light, supply 
of water and minerals, &c., are unfavourable, and the leaf is 
1 Proc. Cambridge Phil. Soc., vol. xi, Pt. V, p. 307. 
