PROTOBASIDIOMYCETES 



413 



germ pores equatorially disposed and opposite, I'pmi many hosts 

 the uredospores are produced throughout a very long season. They 

 may appear upon grain or blue grass in the early spring before the' 

 ii-cidiospore may be developed in the same region. In many cases 

 it is evident that the rust may be propagated from year to year by 

 continuous generations of the uredospores. Again, it has been 

 experimentally shown that uredospores may retain the power of 



FIG. 203. Puccixi.-t 



VRKDOSI-ORES AND TELEUTOSI-OKI.S 



germination for weeks. It is, therefore, generally possible to ac- 

 count for the appearance of rust, even in spring grain, at a dis- 

 tance from the alternate barberry host. In seeking to explain the 

 infections in spring grain as due to the uredo stage from some 

 neighboring grass stubble, it should be remembered that artificial 

 inoculation experiments have shown a very definite restriction of 

 hosts in the different physiological species. Some who have fol- 

 lowed carefully outbreaks of rust in the grain fields, year after 

 year in the Northwest, feel that the problem is not yet completely 



