188 THE ROYAL SOCIETY OF CANADA 



in the western half. Three of the forms, XV, XIX and XXX, were 

 each found in but one place; but, in all other cases, the same form 

 was found in two or more different places. This is what we would 

 expect in the prairies, where the high winds, the absence of natural 

 barriers between the wheat-growing areas, and the similarity of the 

 wheat varieties grown make it extremely unlikely that any strain 

 will be confined to one small fixed area. This, of course, is advanced 

 merely as a tentative hypothesis, to be tested by further survey. 

 It is possible that climatic factors may determine to some extent 

 the geographical limits of these biologic forms, and the areas invaded 

 by them may be more or less fixed. 



A graphic representation of frequency of occurrence and virulence 

 is given in Fig. 3. The comparative frequency of the forms is in- 

 dicated by the height of the columns. In Fig. 3, A. the forms are 

 arranged from left to right in order of virulence toward all test wheats. 

 A sub-grouping of strains of approximately equal virulence is indicated 

 by the subdivisions. A, B, C, D, and E. The same arrangement 

 with respect to their reactions toward the test wheats of the bread 

 group {i.e., excluding durum varieties, spelt and einkorn) is given 

 in Fig. 3, B. Particular interest attaches to sub-group B, which 

 includes the 6 strains referred to in the distribution map. Fig. 1, B. 

 The fact that this, the most important sub-group, is six places removed 

 from the most extreme type of virulence, is in itself matter for en- 

 couragement. 



In Table IV is given a summary of the information relating to 

 the forms isolated which, from the practical standpoint, appears to 

 be of greatest importance. This includes a statement of the number 

 of times each form was isolated, the districts in which collected, the 

 effect on common and durum wheats, and remarks suggestive of the 

 probable effect of each form in the field. 



Infection of Grasses 



Some of the most virulent of the forms of stem rust isolated 

 were collected on wild grasses. Stakman and Piemeisel (59) found 

 P. graminis present on about 35 species of grasses in the United States, 

 26 of which species they were able to infect artificially with P. graminis 

 tritici. There is absolutely no doubt that the wild grasses have a 

 marked effect upon rust epidemics. In late fall and early spring 

 hey present suitable host tissue for the fungus when the spring 

 wheat crop is not available. They may originate seasonal infection 

 by permitting the overwintering of the rust in the mycelial or uredinio- 



