484 Journal of Agricultural Research voi. x, No. 9 



Hosts not infected by artificial inoculation: Anthoxanthum odoraium 

 L., Bromus inermis Leyss., Hordeum juhatum L., Phalaris canariensis L., 

 Triiicum vulgar e Vill. 



It seems probable that a number of species of Agrostis are hosts for 

 P. graminis agrostis, although it was found in the field only on A. alba 

 and A. stolonijera. The grasses which were easily infected artificially 

 may harbor the rust in nature. It is quite possible that if more trials 

 had been made Calamagrostis canadensis would also have proved more 

 susceptible than it appeared to be, because the few leaves which became 

 infected developed fairly normal uredinia. Of the three cereals, barley 

 and oats are most easily infected. The uredinia are, however, always 

 very small and few in number. It seems probable that Phalaris cana- 

 riensis, Hordeum jubatum, and Anthoxanthum odoratum can be infected 

 by the rust, although the inoculations made thus far have not resulted 

 in successful infection. Negative results from a small number of trials 

 often mean little. 



Stuart, as reported by Arthur (7, p. 18), succeeded in infecting wheat 

 with urediniospores from Agrostis alba vulgaris, but there is some possi- 

 bility of error since the spores on wheat were larger than those from the 

 original host. Carleton (7, p. 18) did not succeed in infecting wheat or 

 oats with the rust from Agrostis alba vulgaris, and concludes that it is a 

 distinct form. Eriksson gives Agrostis alba and A. stolonijera as hosts. 

 Jaczewski (18, p. 353) gives Agrostis alba as host and states that the 

 rust is also capable of infecting the following : 



Secale cereale, Avena saliva, Triiicum vulgare, Hordeum vulgare, Triticum repens, 

 Dactylis glomerata, Bromus secalinus, Bromus inermis, Aira caespitosa, Apera Spica 

 venti. 



It is quite possible that the rust will be found in nature on a number 

 of grasses. Although it probably occurs commonly in nature, it is doubt- 

 ful if it is of practical importance on any of the cereals, since it can infect 

 them only very weakly. 



GENERAL DISCUSSION. 



MORPHOLOGY OF UREDINIOSPORES * 



In general the size and shape of urediniospores of different biologic 

 forms of Pticcinia graminis are similar. If, however, large numbers of 

 spores are measured and the arithmetical mean or biometrical mode de- 

 termined, it becomes quite apparent that there are appreciable and fairly 

 constant differences, provided the spores measured be taken from con- 

 genial hosts. Jaczewski (18, p. 358) briefly summarizes the previous 

 observations of similar nature, but states that he was unable to dis- 

 tinguish the different biologic forms in Russia by spores sizes. It can, 



1 Spore measurements were made by Mr. M. N. Levine. Only the dimensions of spores on congenial 

 hosts are given; the other data will be presented in a separate paper. 



