188 BOTANICAL GAZETTE [SEPTEMBER 
behavior of the cells will of course show their true nature, but as 
material is not available for following them through the later 
stages of germination of the seedling, it seems best to publish as a 
preliminary account the figures of the stages already found. The 
importance of their bearing on a possible method of wintering of 
wheat rust in the absence of the barberry or uredo is apparent. 
These peculiar phenomena were not confined to the teleuto- 
spores, but were frequently present in the stalk cells (figs. 2, 5, 9, 
10), and even in the mycelial region below (figs. 14, 15). Fig. 14 
represents a radial section through the base of a pustule, one 
teleutospore and neighboring stalk cells being included. As is 
seen, the growing cells of the sorus were associated more or less 
in groups, but usually interspersed with smaller empty cells. 
Sometimes they formed dense areas, where it was difficult to 
determine whether they were of hyphal or teleutosporic origin. 
Now and then faint outlines resembling distorted, multicellular 
teleutospores were seen in the mass, but in all probability at least 
some of the cells arose from the mycelium. Apparently identical 
cells were found in other parts of the pericarp remote from the 
pustules (fig. 17). Lying near were filaments composed of similar 
though usually smaller cells (figs. 16, 17, 19). These however 
were enlarged portions of a smaller mycelium, all the remaining 
cells being empty. 
Quite separate from the cells just described, fragments and often 
considerable pieces of what appeared to be living rust mycelium 
were found mixed with dead hyphae of the rust (figs. 20-22). 
They were usually in the pericarp, but often lay next to the deepest 
layer. There were occasional places outside the region occupied 
by the layer of feeding cells where they passed through into the 
cells of the scutellum and were found in considerable abundance 
within 6 or 7 cells of the growing plant itself (fig. 21). 
As noted, the fate of the teleutosporic and mycelial cells described 
above remains for future determination, as my present material 
contains no later stages. The evident suggestion is that they may 
serve as growing points for the development of new rust mycelia 
and the infection of the embryo and seedling. 
The possibility for infection of the seedling when the pericarp 
