262 Salmon, On Specialization of Parasitism in the Erysiphaceae. 



4. The conidia of E. Cichoracearum on Galium silcaticum ^) would 

 not infect Senecio vulgaris. 



5. The conidia of E. Polygoni on Heracleum SpJiondylium (Um- 

 helliferae) would not infect Aegopodium Podagraria and An- 

 thriscus syloestris {Umbelliferae); nor Hypericum montanum 

 {Hypericineae) — all host-plants of the species. 



6. The conidia of E. Polygoni on Trifolium incarnatum {Legu- 

 minosae) woiild not infect T. repens and Vicia sepium [Legu- 

 minosae), nor Hypericum montanum — all tliree known as 

 host-plants of the species. 



7. The conidia of E. Galeopddis on Galeopsis Tetrahit {Lahiatae) 

 would not infect Glechoma hederacea and Calamintha Acinos 

 [Lahiatae) — host-plants of the species. 



On the other hand it was found that the conidia of E. Polygoni 

 on Hypericum perforatum were able to infect H. montanum\ those 

 of Microsphaera Astragali on Astragalus glycyphyllos were able to 

 infect A. Cicer^); those of Uncimda Salicis on Salix purpurea in- 

 fected 'S". Caprea; and those of U. Aceris on Acer Pseudo-Platanus 

 infected A. campestre. 



We see, then, from Neger 's experiraents that this phenomenon 

 of specialization occurs in many instances in the Erysiphaceae, and 

 moreover that it is sometimes carried to a high degree, since the 

 Oidium on Trifolium incarnatum is not able to infect T. repens, 

 and that on Arfemisia vulgaris is not able to infect A. Absinthium. 



In February came the bare Statement from von Schrenk (5) 

 that ,the spores of Erysiphe Graminis on wheat are unable to in- 

 fect oats, rye, Poa, corn (= maize), roses etc.' 



This was followed in July by a paper by Marchai (6), giving- 

 the results of a nuniber of infection-experiments in which the Oidium 

 of Erysiphe Graminis was used. Marchai proved flrst that this 

 Oidium on Rye, Wheat, Barley and Oats constituted on each 

 host a ,forme specialisee,' the fungus in each case being unable to 

 infect the three other host-plants. It was found further that the 

 Oidium on Hordeum was unable to infect the species of any other 

 genus; 48 species of grasses belonging to 27 different genera were 

 used in this experiment. Conversely, it was proved that Oidia 

 which were found growing on Poa amma, P. pratensis, Agropyrum 

 repens, A. giganteum, Holcus lanatus, Festuca pratensis, Bromus 

 sterilis ^11(1 B. mollis \Yeve all unable to mfect Hordeum. Marchai 

 considers that his experiments prove the existence of seven ,formes 

 specialisees,' as follows: 



1. E. Graminis f. spec. Tritici, sur Trificuni vulgare, Spelta, 

 polonicum, turgidum, non sur T. durum, monococcum, dicoccum. 



1 ) Neger (1. c. pp. 222, 248, 251, 254) quotes me as referring the Eri/siphe 

 on Galium to the species JE. Polygoni, whereas I bave alwaj'S referred the 

 fungus in qiiestion to E. Cichoracearum. See Monograph, p.'20;3, 207—208: 

 also ,Suppl. Notes', p. 183. 



-) Contradictory statenients are made by Neger with respect to the 

 infection-powers of the Oidium thus obtained on A. Ciccr; at p. 252 it is 

 "stated that it proved to be unable to infect A. glycyphyllos, whilst at p. 254 

 it is Said to be able to do so. 



