196 Journal of Agricultural Research voi. iv.no. 3 



over the stomatal slits in a perfectly normal manner, send a protoplasmic 

 process through the slit, and then form the substomatal vesicle. The 

 stimulus to entrance may be negative phototropism, since Fromme (3) has 

 found such a response on the part of uredine germ tubes. Forty-eight 

 hours after inoculation infection threads have frequently already grown 

 into the intercellular spaces and have branched quite profusely. The 

 hyphae are frequently large and very vigorous in appearance. Haustoria 

 are sometimes sent into the cells of the host; in many cases they are 

 large in size and of normal appearance. 



Within a short time after the hyphae become closely appressed to the 

 host-plant cells, there are usually unmistakable evidences of some dele- 

 terious influence upon the host cells. The chloroplasts very often seem 

 to be affected first. They may appear slightly corroded at first and 

 somewhat irregular in outline. They may retain their identity for some 

 time, but more often seem to be clumped together in more or less 

 irregular masses. This appearance may frequently be due to a shrinking 

 of the protoplast from the cell wall, especially when the chloroplasts are quite 

 numerous (PI. XXVIII, fig. 4 and 5). As the process of disintegration 

 progresses, the outlines of the individual plastids become increasingly 

 fainter until they are scarcely distinguishable. Sometimes, however, the 

 outlines are still visible , although there appears only a more or less uniformly 

 staining mass (PI. XXVIII, fig. 4 and 5). Eventually the outlines 

 of the plastids become obliterated almost entirely, leaving only a fairly 

 homogeneous, uniformly staining, nongranular mass, with little remaining 

 semblance of structure. In such dead cells, however, the very faint 

 outlines of what were probably chloroplasts may be discernible in some 

 part of the cell (PI. XXVIII, fig. 6). Sometimes the contents of dead cells 

 appear more or less finely granular, the granules being variously disposed. 

 They may occur in irregular clumps, in beadlike chains, or in various 

 other combinations. More frequently the contents are very nearly 

 homogeneous, with only a few scattering granules. 



The action does not always depend upon actual contact. It seems 

 sometimes to precede actual fungous invasion, although in no observed 

 case did it occur: very extensively far in advance of actual hyphal inva- 

 sion. The hyphae often are very closely appressed to the cell wall, and 

 in such cases the action goes on very rapidly. Quite frequently, however, 

 the chloroplasts on one side of the cell will have been destroyed almost 

 completely, while on the other side, away from the hyphae, they still 

 appear quite normal. Naturally, of course, all sorts of gradations can 

 be found. The ultimate result is, however, usually the same. The 

 chloroplasts disappear, the nucleus shows definite signs of disintegration, 

 the protoplast collapses, and the cell stands out sharply from the normal 

 ones near it. 



Variations occur also in the matter of cells attacked. It seems that 

 sometimes the hyphae grow over or past cells which apparently escape 



