276 Salmon , On Specialization of Parasitism in the Erysiphaceae. 



The Oidium was then sown (Exper. no. 32) on B. commutatus, 

 B. velutinus, B. secalinus and B. tectorum. Again it infected B. 

 commutaius, — 3 pots ; further, füll infection resulted on B. tectorum, 

 — a remarkable fact but one for which we were prepared by the 

 similar behaviour, on the same host, of the Oidium on B. inter- 

 ruptus. With regard to B, secalinus, ,subinfection' occurred in one 

 pot, as in the previous experiment (Exper. no. 18), whilst this time 

 the plants in the other pot remained untouched. On B. velutinus 

 3 (of the 4 inoculated) leaves in one pot showed minute flecks of 

 mycelium and small tufts of conidiophores. These persisted until 

 the end of the experiment, i. e. 15 days from the date of inocu- 

 lation. On the plants in the other pot no infection occurred. The 

 present case was evidently one of ,subinfection' only. 



In Experiment no. 40 the Oidium was sown on B. interrwptus, 

 B. tectorum, B, hrizaeformis and B. erectus. Füll infection again 

 resulted on B. tectorum and on B. interruptus, and ,subinfection' on 

 B. hrizaeformis; — all three results the same as in previous ex- 

 periments (nos 23, 32). No infection occurred on B. erectus as 

 before. 



In the next experiment (Exper. no. 45) the Oidium was sown 

 on B. mollis and B. patulus, and fully infected the former but fail- 

 ed to touch the latter. 



The Oidium was then tried on B. mollis and B. sterilis (Exper. 

 no. 47), and again infected B. mollis while passing over B. sterilis. 



In Exper. no. 57 the Oidium was sown on seedhngs of B. 

 commutatus and B. madritensis, and on a potted plant of B. race- 

 mosus. The last-named was a plant 9—10 weeks old, and during the 

 course of the experiment become much covered with ,rust.' In this 

 experiment B. commutatus was, as usual, infected, but neither B. 

 racemosus nor B. madritensis was touched. 



In the next experiment (Exper. no. 65) the Oidium was sown 

 on B, mollis, B. velutinus, B. secalinus, B. crinitus and B. arduennen- 

 sis var. villosus. Füll infection resulted on B. mollis, and ,subinfection'^ 

 on both pots of B. velutinus ; while no infection occurred on B. secali- 

 nus (1 pot), B. crinitus, and B. arduennensis var. villosus. 



No further experiments were made with the Oidium on B. 

 hordeaceus, but a series of plants, viz. B. gigatiteus, B. inermis, B. 

 mollis, B. hrizaeformis , B, tectorum , B. arduenyiensis var. villosus 

 and B. asper was inoculated (Exper. no. 61) with the spores of an 

 Oidium growing on a plant in the Cambridge Botanic Gardens 

 labelled ,B. hordeaceus var. glahrescens,'' and which in all probability 

 is the same as that known as B. mollis var. glahrescens. This 

 Oidium behaved in the same way as that on B. hordeaceus in in- 

 fecting B. mollis and B. tectorum, and in causing ,subinfection' on 

 B. hrizaeformis ; also in passing over B. asper and B. arduennensis 

 var. villosus; it also failed to touch B. giganteus and B. inermis. 

 In the present case the infection of B. hrizaeformis was very slight, 

 resulting only in the production of a few (2 — 4) conidiophores on 

 1 leaf in each of the three pots. As, however, the plants in this 

 experiment were kept und er Observation for only 7 days after inoc- 



