UROMYCES ON GRASSES 69 



5. f. Glyceriae — on G. aquatica 



6. f. Agropyri — on A. repens 



7. f. Epigaei — on Calamagrostis epigeios (also, but rarely, on Arena 

 sativa) 



8. f. Hold — on Holcus lanatus. 



To these Miihlenthaler (1910) adds a ninth form, on several 

 species of Bromus. These results agree pretty well with those 

 of Klebahn, but not with those of Carleton. According to the 

 latter, the only host of P. Lolii in nature, in the United States, 

 is Avena sativa : but in artificial cultures it can be foisted on 

 other species because of the unnatural conditions, especially 

 on account of the employment of very young and non-resistant 

 plants. In any case, however, there is a general agreement 

 that the form of P. Lolii on Avena sativa cannot be transferred 

 to Wheat, Barley, or Rye. In fact, with the exceptions 

 mentioned, it was found by all experimenters more or less in 

 all these cases, that attempts made to transfer the fungus from 

 the host of one " special form " to those of the others were 

 unsuccessful from some unknown cause. 



The state of things in regard to the two common Uromyces 

 species, found upon Grasses, is more perplexing. There is no 

 agreement whatsoever between the various authors who have 

 experimented upon them. The latest results are, perhaps, the 

 following : 



Krieg (1909) divides Uromyces Dactylidis into two biological 

 races or " format speciales " : 



1. f. sp. with peciclium on Ranunculus bulbosus, B. repens 



2. f. sp. with aecidium on several non-British species of Ranunculus. 



Juel (Svensk Bot. Tidskr. ii. 169, 1908) divides Uromyces 

 Poae into nine biological races, of which the following seven 

 may be British : 



1. f. sp. Ficariae-nemoralis 



2. f. sp. Ficariae-trivialis 



3. f. sp. Ficariae-pratensis 



4. f. sp. repentis-nemoralis (also on R. bulbosus) 



5. f. sp. repentis-trivialis (also on Poa annua) 



6. f. sp. repentis-pratensis 



7. f. sp. azirico/ni-pratensis. 



