144 Journal of Agricultural Research voi. xxrir, No. 3 
The central vacuole is obliterated, and the whole cell cavity presents a 
loose, dark granular appearance in which the plastids (some of them 
almost normal in size and shape) are faintly discernible. The nuclei 
have disappeared. Adjoining this pathological tissue are cells quite 
normal in appearance except for an occasional thickened wall. 
These thickenings of the wall occur at the surface of contact between 
two cells as at c in Plate 4, E, and but rarely, if ever, on those parts of the 
cell wall which abut upon intercellular spaces. Usually the cell on one 
side of such a wall is pathological and the cell on the other side is 
healthy. There is some evidence that these thickened contact walls are 
impervious and prevent the diffusion of substances to and from the dis¬ 
eased cell. This may serve as a protection to the healthy cell by walling 
it off from any toxic materials formed in the pathological tissue. 
In view of the fact that a vague impression exists among pathologists 
that the immune host attacked by rust somehow walls itself off from the 
fungus, thus checking its spread, it should be emphasized here that this 
formation of heavy walls between cells in no way excludes the fungus 
itself. So far as is known, rust hyphae do not enter cells nor pass from 
one cell into another. The only entrance into a cell is for the purpose of 
forming a haustorium, and this takes place from an intercellular space, 
and in Kanred at least, the host cell walls adjoining intercellular spaces 
are practically always left thin. . 
In some cases the reaction between host and parasite is more 
sluggish than in the examples described. The fungus makes a full-grown 
haustorium and gains enough food from it to enable the hypha to branch 
and grow on to the next cell before the first haustorium and the invaded 
host cell die. The new hypha makes a second haustorium and this may 
be repeated several times and further branching may occur. The inter¬ 
action of host and fungus is slower in starting and not quite so severe in 
its effects. The result is a succession of dead, discolored hyphae and 
dead host cells and an ever weakening advance of the fungus. As many 
as two dozen cells or even more may be involved before the fungus dies. 
In Plate 6, A, drawn on a larger scale than the preceding, is .shown 
the advancing growth of the fungus. The dead cell at e already has 
succumbed to the attack of the fungus. At d is a nearly empty hausto¬ 
rium mother cell which has discharged its contents into the host cell 
to form the haustorium at c. The haustorium itself looks normal, but 
it lacks the cytoplasmic envelope of host cytoplasm. Evidently the 
normal relation of host and parasite has not been established. Still 
younger haustoria of the same mycelium show this cytoplasmic layer, 
but the older haustoria usually lack it. Perhaps some substance diffuses 
from the haustorium into the host cell which either destroys or repels the 
cytoplasm surrounding the haustorium. At b in the same figure (PI. 6, A) 
is an older haustorium, the neck of which does not lie in this section. 
The advancing tip of the fungus at a has formed a typical and fairly 
vigorous haustorium-producing cell with dense red-stained contents and 
two minute nuclei, and, as usual, the hypha has branched below this cell. 
The host cell shows no further signs of disturbance. Its nucleus, which 
lies above the plane of the drawing in the same section, is unaltered in 
appearance. 
Plate 6, B, shows a slightly older part of the same mycelium. At e is 
a full-grown haustorium which is somewhat vacuolated and ragged about 
the edge and has lost its connection with the cell at d that produced it, 
