212 
E. B. MAINS 
consequently more abundant spores. A humid atmosphere also 
lessens the transpiration from the dying portions of the leaf and the 
evaporation from the dead areas so that the infected leaves in a moist 
atmosphere have a much longer life. In drier air, the infected leaves 
dry up and the plants become rust free. 
The effect of mineral starvation upon the host has been investi- 
gated principally by Marshall Ward (1902c). Sheldon (1905) and 
McAlpine (1906) have made some observation upon the effect of soils 
and manures upon the development of some of the rusts. 
Results similar to Marshall Ward's were obtained in Experiments 
6 and 7. Infection was obtained upon some plants in all of the mineral 
nutrients. Table I shows that the best infection was obtained upon 
plants which were furnished with a full mineral nutrient. A deficiency 
of an element, however, does not bring about immunity for the host, 
but it causes a smaller number of plants to be infected and a lessening 
of the amount of rust as shown by the number of pustules. 
That this effect is produced upon the rust is due partly to a lack 
of these elements for the host and a consequent slow mineral starvation 
of the rust. This, however, only explains a portion of the effect pro- 
duced, since the rust can probably obtain these elements in small 
quantities from the host as long as the host is alive. A portion of the 
effect produced is at least to be referred to the effect upon the physi- 
ology of the host and its greatly reduced ability to manufacture the 
proper food materials for the rust. That this is true is shown in 
Table V. In this table it is seen that in solutions in which the host 
was supplied with a mineral nutrient solution, but was not supplied 
with carbohydrates and was kept in the dark to prevent their manu- 
facture, the number of plants infected were few. That even this 
number was infected was due to the fact that the host was not ex- 
hausted of soluble carbohydrates, such as the sugars, before infection. 
The infection of plants supplied with carbohydrates in all cases far 
outnumber that of the plants without carbohydrates. 
The need of carbohydrates is also shown in Experiments 8-21 
conducted upon the effect of light upon the development of the rusts. 
Of the two rusts Puccinia coronata upon oats shows the closest relation 
in this regard. Thus, in Table II, it is shown that the retardation in 
infection approximates closely the period that the host was left in 
darkness and consequently the period during which carbohydrates 
were not being formed. In all cases where the host was kept contin- 
uously in darkness after inoculation there was no infection. 
