S a Im n , On Specialization of Parasitism in the Erysiphaceae. 287 



and Hieracium murorum as supportiüg liis view. Now, in the case 

 of the Erysiphe on tlie foiir host-plants quoted, a productiun of peri- 

 thecia does take place. I have found perithecia somewhat fre- 

 quently on Senecio vulgaris, and very frequently indeed on Sym- 

 phyümi, while otlier observers have recorded theni on the Cala- 

 mintha and the Hieracium. Now, Neger states that on certam 

 individuals of these hosts observed by him no formation of peri- 

 thecia occurred, and yet the same plants, or plants of the same 

 species in the same locality, became reinfected in the spring. In 

 face of the obvious possibijity that the reinfection may have 

 been caused by ascospores pioduced, not on other species of host- 

 plants, as Neger assumes, but on some other individuals of the 

 four host-species in question, the facts cannot, I think, be regarded 

 as having the signiiicance attached to them by Neger, as quoted 

 above. Although 1 consider that the phenomena cited by Neger 

 give no certain evidence on the subject, yet it certainly does seem 

 probable on a priori grounds, that the ascospores may prove to be 

 able to infect plants which the conidia of ,biologic forms' are unable to 

 do. Marchai, it may be noted, inclines to the opposite opinion. 

 This author remarks, in connection with the biologic forms of E. 

 Graminis studied by him, (6, p. 212), ,Comment se comportent les 

 ascospores des diverses races physiologiques etudiees? Contribuent- 

 elles ä fixer d'une faQon plus profonde, definitive, dans la descen- 

 dance, l'etroite adaptation parasitaire acquise? Ou bien permettent- 

 elles, ce qui est moins probable, ä V Erysiphe Graminis d'etendre son 

 aire de dispersion sur d'autres hotes'? 



It will be of interest here to compare the specialization of 

 parasitism whicli has been observed in the case of the parasitic 

 fungus Puccinia dispersa Erik SS. und Kenn, on Bromes with that 

 which has taken place in the Oidium on the same grasses. In a 

 recent paper by Marshall Ward (8) the results of a large number 

 of infection-experiments with this fungus are given. We find here 

 (1. c. p. 287) the Statement that ,the spores of the brown JJredo 

 (P. dispersa) if grown on B. molhs are able to infect successfully 

 practically any member of the group Serrafaicus, and but rarely 

 or not at all any member of the other sections so far examined of 

 the genus Bromus; whereas the same fungus grown on B. sterilis 

 readily infects members of its own grow^' Stenohromus — e. g. B. 

 sterilis, B. madritensis, and B. maximus — but is debarred from the 

 Serrafalcus group. ^ It was found, for instance, that the spores of 

 the Puccinia on B. mollis caused füll infection of B. arcensis and 

 less readily infected B. commutatus, B. secalinus, B. racemosus, and 

 B. mlutinus. It would seem therefore, that specialization of para- 

 sitism has been carried to a higher degree — with the consequent 

 result that more , biologic forms' liave been evolved (cfr. Tables 8, 9) 

 — in the case of the Oidium on Bromus than in that of the 

 Puccinia. 



The question may here be asked, is there any relation between 

 liability to infection or power of resistauce and the visible structural 

 features of the leaf of the species of Bromus used in the above 

 experiments with the Oidium of E. Graminis? This question has 

 been asked by Marsh all Ward (1. c.) in connection with the 



