488 Journal cf Agricultural Research ' voi. x, No. 9 



PARASITISM OF BIOLOGIC FORMS 



The parasitism of some of the biologic forms is surprisingly similar. 

 Often the reaction of only one or two host plants distinguishes one form 

 from another. By reference to Tables XXVIII to XXXIII it is quite 

 apparent that P. graminis iritici, P. graminis tritici compacti, and P. 

 graminis secalis fall naturally into one group, so far as parasitism is con- 

 cerned. In the same way P. graminis avenae, P. graminis agrostis, and 

 P. graminis phleipratensis fall into another. 



There are a number of common hosts for P. graminis tritici, P. graminis 

 tritici compacti, and P. graminis secalis. From the results thus far 

 obtained it is very evident that a number of species of the genera Agro- 

 pyron, Elymus, and Hordeum are about equally congenial hosts for these 

 three biologic forms. The first two can be distinguished from each other 

 definitely, so far as present knowledge goes, only by their action on 

 Triticum vulgare, while P. graminis secalis differs from both in its effect 

 on Secale cereale and Agropyron repens. These differences are, however, 

 very distinct and apparently quite constant. The matter of the con- 

 stancy of the forms will be dealt with fully in a subsequent paper. 



P. graminis avenae, P. graminis agrostis, and P. phleipratensis are also 

 similar parasitically. Koeleria cristata, Holcus lanatus, Alopecurus pra- 

 tensis, A. geniculatus, and Dactylis glomerata are common hosts for all 

 three. P. graminis agrostis and P. phleipratensis attack Avena sativa in 

 about the same degree. The infection is always weak, the uredinia 

 nearly always being very minute. Hordeum vulgare and Secale cereale 

 are infected weakly by all three. Exhaustive experiments on these three 

 forms are very desirable and may yield interesting results. 



All of the biologic forms studied can develop, at least to a limited 

 extent, on Hordeum, vulgare, Secale cereale, and possibly on other hosts. 

 The degree of infection varies widely, however, as previously noted. 



RELATION OF WILD GRASSES TO RUST EPIDEMICS 



From a practical standpoint the effect of the wild grasses on cereal- 

 rust epidemics is important. The influence of the grasses would seem to 

 be possibly significant in three ways: (i) In permitting the overwintering 

 of the rust in the mycelial or urediniospore stage, (2) in aiding in the 

 dissemination of rust, and (3) in possibly breaking down the specializa- 

 tion of the biologic forms in nature. 



The writers have done preliminary work on the overwintering of rust 

 on wild grasses, but are not prepared to draw definite conclusions. Cer- 

 tainly urediniospores of P. graminis phleipratensis can survive the winter 

 in some localities. Johnson (19, p. 12-13) showed this to be true at 

 Arlington, Va., and Hungerford (17) showed the same to be true in 

 Wisconsin. Mercer (21) states that in North Dakota urediniospores of 

 timothy rust are hard to find after the first hard freeze. Observations 



