t 
RELATION OF SOME RUSTS TO THEIR HOSTS 1 83 
causing but little injury, at least up to the time of spore formation. 
Tubeuf, in the case of hypertrophies caused by rusts, thought that 
there appeared to be "a closer symbiotic relation between the fungus 
and its host branch than between the host branch and its main branch." 
He also noticed that in the case of some rusts the infected portion of 
the leaf remained alive after the death of the surrounding tissues and 
he looked upon this as a condition resembling that of some lichens. 
Ward (1890, 1902c) suggested that the relation between Puccinia 
dispersa and its host tends towards symbiosis and that the rust is not 
destroying the protoplasm of the host, but is robbing the host of its 
food supplies. Ward (1905) found that when the rust does attack 
the host so vigorously that the protoplasm is killed it brings about its 
own death and immunity for the host is produced. This condition, 
he discovered, can be duplicated by starving the host and by so doing 
starving the rust. His earlier work (1902a, 1902c) had already 
pointed towards this conclusion since he found that immunity did not 
depend upon anatomical features such as number and size of the 
stomata, hairiness, etc., and that mineral starvation, although it did 
not produce immunity, reduced the number of spores which were 
produced. 
The relation of the rusts to the carbohydrate supply has been noticed 
by a number of workers. Halsted (1898) found in the case of Puc- 
cinia Podophylli that there is a collection of starch in infected regions 
of the leaf. The centers of such areas however contained much less 
starch than the margins. Robinson (191 3) noticed much less starch 
in those areas of the leaf which are infected with Puccinia Malvacearum. 
McAlpine (1906) quotes the work of F. T. Shutt, who found that the 
grain and straw of rusted wheat contained more protein and less 
carbohydrates than the rust-free plants. Tischler (191 2), working 
with Uromyces Pisi, discovered that the portion of the host containing 
the most mycelium of the rust also contained the greatest amount of 
sugar. 
The effect of environmental factors such as soil, moisture, tempera- 
ture, and light upon the relation of the rusts to their hosts has been 
studied by a number of workers. McAlpine (1906) noticed that 
nitrogenous manures retarded the ripening of grain, while phosphate 
of lime brought about early maturity and enabled the grain to escape 
the rust to some extent. Sheldon (1905) has reported that soils favor- 
able to the host are also favorable to the rust of carnations ( Uromyces 
Caryophyllinus) . 
