Aug. 27. I9I7 Puccinia graminis on Cereals and Grasses 



439 



leaves which became infected, however, developed large uredinia, thus 

 furnishing ample material for the next inoculations. 



It is especially interesting to notice the results of the successive trans- 

 fers of the wheat rust from rye to rye. The rust develops very poorly, 

 and, since the amount of the inoculating material as a rule decreases 

 with each successive transfer, the rust is finally lost. It is also inter- 

 esting to note that barley, a host for both forms of rust, did not change 

 the parasitic capabilities of either. 



The diagram shows, then, that on the original Agropyron smithii both 

 P. graminis tritici and P. graminis secalis were present. For this reason 

 successful infection resulted on wheat, rye, and barley. But when 

 transfers were made from rye to barley, rye, and wheat, only the first two 

 became infected. When, on the other hand, transfers were made from 

 the rusted wheat, wheat and barley were infected, and rye only occa- 

 sionally produced small, characteristic uredinia, which are distinguishable 

 from those of P. graminis secalis both in appearance and performance. 

 Both rusts easily infected barley and A. smithii. 



Agropyron smithii is very abundant and is very often rusted. In 1916 

 it was very heavily rusted throughout the Northwestern States and 

 southern Manitoba, Canada. There has been no difficulty in finding 

 badly rusted plants during the last five years. On account of the fact 

 that the grass is a host for two distinct biologic forms, and possibly a 

 third, it is probably important in stemrust distribution. 



Table V. 



-Results of inoculations with urediniospores from Agropyron spicatzim 

 (Pursh.) Rydb. 



Only one set of inoculations was made with Agropyron spicatum from 

 east of the Rocky Mountains, although it was quite badly rusted in most 

 localities visited from Glasgow, Mont., west to the mountains. The rust 

 was ordinary P. graminis tritici. P. graminis was also collected on the 

 grass west of the mountains, but the spores were not viable when inocu- 

 lations were made. Unfortunately, therefore, the identity of the bio- 

 logic form of the rust in that region could not be determined with cer- 

 tainty (Table V). 



In every one of the 15 trials recorded in Table IV, some wheat became 

 infected, indicating that P. graminis tritici often occurs in many regions 

 on Agropyron tenerum. On the other hand, P. graminis secalis developed 

 only as a result of inoculations i, 2, and 7, and some of the uredinia in 

 No. 7 were unquestionably P. graminis tritici. 



