52 MASS. EXPERIMENT STATION BULLETIN 202. 



duction of parench:^Tiiatous plugs which fill the stomatal cavity, and to 

 lenticels composed of cork cells through which the hyphse cannot pene- 

 trate. (2) On the other hand, the resistance of plants to disease has 

 often been regarded as due not to morphological characteristics, but 

 rather to physiological or chemical factors. Myoshi (1895) concluded 

 that many fungi respond to chemical attraction. According to Massee 

 (1904) infection depends on the presence of positive chemotactic sub- 

 stances in the plant cell. Klebahn (1896) concludes that infection is 

 a kind of conflict between host and parasite. Bolley (1908) attributed 

 resistance to chemical agents, such as toxins, which arise as a result of the 

 fungous attack upon the host. These few citations from the extensive 

 literature on this subject illustrate two views as to the resistance of plants 

 to disease. 



The work done by the writer indicates that the resistance of some 

 varieties of snapdragons to rust is due to morphological characteristics 

 rather than to physiological differences. The relative susceptibility 

 of forty-six varieties of snapdragon to rust has already been given. The 

 inoculated plants developed uredinia, some in large numbers and some 

 in small numbers. But on both resistant and susceptible plants the 

 sori were developed in the same length of time, and there was no apparent 

 difference in the vigor of the sori after they had once broken through 

 the epidermis. This seems to the writer to indicate that the difference 

 in susceptibility is not due to chemical factors within the host cells, 

 but rather to mechanical factors preventing infection. The most sus- 

 ceptible plant is the one infected in the most places, that is, the one into 

 which the most germ tubes enter. 



Infection of snapdragons by P . Antirrhini is always through the stomata. 

 The AM'iter sprayed urediniospores on the living leaves, and eight hours 

 later examined the leaf surfaces microscopically. This was done re- 

 peatedly, but at no time was infection seen to occur anj'ivhere except 

 through the stomata. The plants used were kept in both Ught and 

 darkness, with stomata both open and closed. The germ tubes were 

 protruded, wandered about slightly, and then bent into the nearest stoma, 

 or, if the water on the leaf dried too soon, they shriveled vip and never 

 reached a stoma. But no germ tubes were found which had penetrated 

 or were penetrating the walls of the epidermal cells. 



The mycelium within the leaf and stem is local; therefore the number 

 of sori on a leaf or stem depends on the number of infections, and, since 

 infection is only through the stomata, it was interesting to determine 

 the connection between the number of uredinia (the index of relative 

 susceptibility) and the number of stomata. 



Leaves were taken from three-months-old snapdragons of susceptible 

 and resistant varieties. The number of stomata on the upper epidermis 

 per unit area of leaf was determined. In each case ten countings go to 

 make up the average given for each variety. Ten susceptible and ten 

 resistant varieties were used. The result of these counts is given in 

 the following table: — 



