S alm o n , On Specialization of Parasitism in the Erysiphaceae. 273 



In the next experiment (Exper. no. 35) the Oidium was sown 

 on B. mollts, B. commutatus^ B. veluimus, B. brizaeformis and also 

 on B. tectorum and B. madritensts, two species belonging to the 

 section Stenobromus. The inclusion of B. tectorum gave rise to a 

 most interesting and unexpected result, since füll infectlon of this 

 species occurred. Of the other species, B. madritensis and B. commu- 

 tatus were passed over, while füll infection occurred as usual on 

 B. mollis and .subinfection on B. brizaeformis; of the two pots 

 of B. velutinus one remained untouched, whilst the other presented 

 a doubtful case of ,subinfection.' The following details may be 

 noted in this experiment. 



First, with regard to the infection of B. tectorum. On the 

 5 th day after inociüation, 2 (of the 3 inoculated) leaves in one 

 pot, and 3 (of the 4 inoculated) leaves in the other pot, bore dis- 

 tinct flecks of mycelium with young conidiophores. On the 6 th 

 day all the inoculated leaves in both pots were clearly infected, 

 and on the 9 th day the infected leaves bore powdery masses of 

 spores. As was shown by the rapidity of the development of the 

 fungus, and by the production of powdery masses of spores, the 

 present case is clearly to be regarded as one of füll infection. 

 The special interest of this infection of B. tectorum lies in the 

 fact that this plant belongs to the section Stenobromus of Bromus, 

 while B. ruf errupf US, the host- plant in nature of the Oidium 

 used in the experiment, belongs to the section Serrafalcus. The 

 infection was all the more remarkable since previous experiments 

 had shown clearly that this Oidium is unable to infect a number 

 of species of Bromus belonging to the section Serrafalcus. The 

 possible significance of the infection-power here shown will be dis- 

 cussed later. 



Of the two pots of B. velutinus used, the 3 inoculated leaves 

 of one pot showed no signs of infection, while of the 4 inoculated 

 leaves m the other pot only 1 showed, on the 6 th day, a very minute 

 fleck of mycehum bearing about 6 conidiophores. By the 9 th day 

 tliis Single leaf showed a few more weak mycelial flecks, and about 

 10 conidiophores. On the 14 th day a few similar patches of 

 scattered conidiophores without visible mycelium were seen on several 

 of the control plants. 



With regard to the behaviour of the fungus on B. brizaeformis, 

 similar results to those of the previous experiment were obtained. 

 Slight signs of infection occurred on the 5 th day after inoculation ; 

 on the 6 th day 2 of the 4 inoculated leaves bore minute flecks of 

 mycelium with minute scattered tufts of conidiophores ; the two other 

 inoculated leaves had unfortunately withered. No further develop- 

 ment occurred ; by the 9 th day the fungus was dying away, and by 

 the 14 th had disappeared. We find again, in fact, only ,subinfection' 

 of this plant. No further experiments were made with B. brizae- 

 formis, and for the present it must remain doubtful as to what is 

 the exact degree of infection the Oidium on B. interruptus possesses 

 as regards this species. 



The next experiment (Exper. no. 44) confirmed the belief that 

 the Oidium on B. interruptus is unable to attack B. commutatus\ 

 füll infection occurred as usual on B. mollis. 



