278 S a Im o n , On Specialization of Parasitism in the Erysiphaceae. 



This Oidium on B. commidatas, then, proves to differ markedly 

 from tlie two Oidia previously experimented witli in being- unable 

 to infect either B. mollis or B. interruptus. Its other characteristics 

 — so far as the somewliat limited number of experiments (sumnia- 

 rized in Table 4) show — are seen in the apparently Ml or nearly 

 füll infection of B. secalinus and B. velutinus, the ,subinfection' of 

 B. hrizaeformis, the very slight ,subinfection' of B. tectorum (which 

 requires further investigation), and the non-infection of B. arvefisis, 

 B. racemosus^ B. maximus, B. sterilis, B. asper, B. erectus, B. macro- 

 stachys, B. madritensis, B. giganteus, B. inermis, and B. patulus, 



Thi'ee Oidia, growing on three other species of Bromus, viz. B. 

 tectorum, B. aroensis and B. racemosus were also used in a few 

 experiments. 



The Oidium on B. tectorum was sown on B. tectorum and B. 

 sterilis in two instances (Exper. nos. 25 and 38), and on B. sterilis 

 alone m another (Exper. no. 74); in all cases infection resiüted 

 (see Table 5). The result here obtained, i. e. the infection of B. 

 sterilis by the Oidium on B. tectorum, is of special interest, not 

 only because the three Oidia on B. interruptus, B. hordeaceus and 

 B. commutatus had refused persistently to touch B. sterilis^ but also 

 on account of the fact that the three Oidia mentioned are able to 

 cause füll infection or ,subinfection' when sown on B. tectorum. The 

 possible signiflcance of this latter fact is discussed below at pp. 

 24, 25. 



The Oidium on B. arvensis was sown (Exper. no. 39) on B. 

 arvetisis and B. mollis, and infected fully B. aroe?isis, but failed to 

 touch B. mollis (see Table 6). 



The Oidium on B. racemosus was sown (Exper. no. 9) on B. 

 commutatus, B. secalinus., B. maximus and B. ciliatus. Fairly 

 strong infection resulted on B. commutatus and B. secalinus, while 

 B. maximus and B. ciliatus were untouched. (See Table 7). 



No further experiments were made with these Oidia on B. 

 tectorum, B. aroensis and B. racemosus; and further investigations 

 are requü-ed to ascertain their distinctive characters. It is certainly 

 remarkable that while the Oidium on B. racemosus shows itself 

 able to infect B. commutatus^ the Oidium on B. commutatus faüed 

 in two instances (Exper. nos. 15 and GG) to touch B. racemosus. We 

 must note, too, that the Oidium on B. hordeaceus, which nearly 

 always infects B. commutatus, failed twice to touch B. racemosus 

 (Exper. nos. 18 and 57). It must be pointed out here that in every 

 instance where B. racemosus was tried as a host-plant, potted plants 

 8 — 10 weeks old were inoculated, whereas in the case of all the 

 other species youag seedlings freshly laised from seed were used. 

 Investigation is required to ascertain it tlie age of the plants of 

 B. racemosus used in these experiments accounts in any way for the 

 non-infection. 



No success attended attemps to infect grasses belonging to other 

 genera with the Oidium on species of Bromus. The Oidium on 

 B. interruptus was sown on two pots each of seedlings of Avena 

 sativa, Hordewn vulgare and Triticum vulgare^ and on potted mature 

 plants of Phleum pratense var. nodosum, Dactylis glomerata, Festuca 

 elatior var. pratensis and Poa annua; the Oidium on B. commutatus 



