224 Journal of Agricultural Research voi. xv. no. 4 



(9) found that it exerted such a pronounced ei^fect on the rust, enabling 

 any biologic form to increase its host range. The grasses were also used 

 to a considerable extent, because both biologic forms of the rusts men- 

 tioned are so often found associated on them in the field. If rust changes 

 rapidly, therefore, as a result of host influence, it would seem that the 

 wheat stemrust and rye stemrust, growing on any one of these hosts for 

 a number of urediniospore generations in the field, ought gradually to 

 acquire the same parasitic capabilities. 



The Agropyrons are especially interesting because the tritici form is 

 able to attack some of them virulently {A. tenerum and A. smithii), 

 while it can attack others, such as Agropyron repens, weakly or not at all. 

 Here, then, there should be an opportunity to test the theory that tax- 

 onomic relationship determines the ability of the rust to pass from one 

 host plant to another. If the stemrust of wheat {Triticum spp.) 

 can not be transferred directly to A. repens, but can be developed nor- 

 mally on A. tenerum which is obviously more closely related to A. repens 

 than to wheat, theoretically the rust should be able to pass from wheat 

 to A. repens after it had first been transferred to A. tenerum or some other 

 species of Agropyron closely related to A . repens. 



Another problem is presented when a biologic form can attack a host 

 plant weakly. If the theory of bridging hosts is a fact, certainly the rust 

 must be easily changed by the host plants. Assuming, then, that in- 

 dividual plants of a given species vary in their susceptibility to the rust, 

 it ought to be possible to increase the virulence of the rust on that par- 

 ticular host plant by successive inoculations with spores from the most 

 vigorous uredinia. Or it ought even to be possible to increase the viru- 

 lence of the rust by constant association with the uncongenial host. 

 Rye is especially favorable for study in this respect. Individual rye 

 plants vary very greatly in their susceptibility to P. graminis tritici. 

 Some are entirely immune, others are almost immune, others are mod- 

 erately susceptible, and still others are quite susceptible. Experiments 

 were made to determine whether the rust from the susceptible plants 

 when transferred to other rye plants could attack rye with greater viru- 

 lence; also, whether it was possible to increase the virulence by simply 

 transferring for a number of successive generations to rye. Similar con- 

 ditions obtain when barley is inoculated with the phleipratensis , avenae, 

 and agrostis strains, and experiments were made with other forms also. 



As indicated in the historical summary, there seemed to be some evi- 

 dence that changing the metaboHsm of the host might materially affect 

 the parasitic capabilities of the rust. The writers therefore under- 

 took further experiments along this line. The results of these experi- 

 ments will be given in a separate paper not yet published. 



If the biologic forms of P. graminis are easily changed, it seems rea- 

 sonable to suppose, as previous experiments seem to have demonstrated 



