RELATION OF SOME RUSTS TO THEIR HOSTS 209 
obtain the necessary materials for spore formation. This is evident 
in the smaller amount of starch present in the parenchyma sheaths 
in the immediate region of the pustules. Oftentimes the parenchyma 
sheaths are here entirely devoid of starch, while in the neighboring 
region starch is present to a considerable extent. Yet even at this 
time, the cells of the host do not show an injury such as one would 
expect if the protoplasm itself was attached vigorously by the fungus. 
It is only after the number of pustules have increased and spore 
formation has continued for some time that the host begins to show 
the effect of the rust's presence. The effect of the rust, even now, is 
not apparent in the tissues containing the rust, but in the neighboring 
tissues as is shown by the green color of the infected areas and the 
lighter green or yellow of the surrounding tissue. The green tissue 
of the infected areas even at this time, may contain small amounts of 
starch, but the neighboring dying regions have no indications of 
starch. It would thus appear that the rust instead of attacking and 
killing the cells of the tissue in which it is situated has a very different 
effect upon them. While it is withdrawing food, at the same time it 
stimulates the infected tissue so that this loss of food is in turn com - 
pensated by the withdrawal of food from neighboring uninfected tissue. 
It would appear that the rust thus destroys the symbiotic balance 
between the cells of the host and causes some of them to have parasitic 
relations with the rest. Marshall Ward (1902&) and Tubeuf (1897) 
observed this effect and considered it as evidence of a symbiotic rela- 
tion between the rusts and their hosts. 
As this withdrawal of food goes on the yellowing of the leaf extends 
farther and farther from the green infected area, the cells of the region 
gradually die and shrivel up, and the tissue takes on a brown appear- 
ance similar to that of cells which have died due to a decomposition 
of their contents. The infected areas, however, still remain alive for 
some time, but in these areas death results from two causes. The 
first of these is the cutting off of the food supply due to the death of 
the surrounding tissue. This, however, is not probably the principal 
cause as the green cells of the infected areas could furnish food to 
prolong their life and that of the rust until by a process of gradual 
starvation both would die. The principal cause which appears to 
bring about the death of these areas is the drying up of the leaf as a 
whole. As Sachs has pointed out the loss of water from dead tissue 
is much greater than from living. The great evaporation from the 
