270 Salm on, On Specialization of Parasitism in the Erysiphaceae. 



In Experiment no. 19 the Oidium was sown on B. molUs, and 

 on seveial of tlie species used before, and also on B. brizaeformis, 

 B. macrostachys, and B. oeluiinus^ three species belonging, like B. 

 interruptus, to the Sect. Serrafalcus. New and interesting results 

 Ttvere here obtained. Whilst failing- to infect B. arvensia, B. race- 

 mosus, B. conimutatuH and B. secalinus^ as nsual, — and also B. 

 macrostachys, — the fungus infected fully B. mollis (as usualj and B. 

 interruptus, and infected also, hut to a less degree, B. hrizaeformis 

 and B. celufinus. 



The two last-named species were aifected as follows. Of i?. hrizae- 

 formis 10 leaves were inoculated, 5 in each pot. No infectiou was 

 noticed until the 7 th day after inoculation, when in one pot 1 of the 

 inoculated leaves bore a weak patch of mycelium and a few scat- 

 tered conidiophores ; in the other pot 2 leaves bore weak tlecks of 

 mycelium only. At the same date as this, it niay be observed, all 

 the infected leaves of B. mollis and B. interruptus bore vigorous 

 mycelial patches vvith powdery masses of ripe spores. On the 9 th 

 day after inoculation 3 leaves in one pot of B. hrizaeformis bore weak 

 mycelial patches and a few scattered conidiophores; in the other 

 pot 3 of the leaves bore very small patches of scattered conidio- 

 phores — scarcely visible imder a hand-lens. This was the maxi- 

 mum growth attained by the fungus on this host; that is, 6 out of 

 the 10 inoculated leaves were weakly infected. On the 11 th day after 

 inoculation the fimgus began to die away, until by the 15 th day no 

 trace of it remained on the leaves of B. hrizaeformis, although at 

 this date the fungus was covering the infected leaves of B. mollis 

 and B. interruptus wilh powdery masses of spores. With respect 

 to B. velutimis, on the 6 th day after inoculation 1 leaf, in one pot 

 only, showed signs of infection by the presence of a few flecks of 

 mycelium. On the 8 th day 4 leaves (of the 5 inoculated) in one pot 

 bore minute flecks of mycelium and a few scattered conidiophores; 

 in the other pot 2 leaves (of the 6 inoculated) bore very small 

 patches of scattered conidiophores at the exact places only where 

 the spores were sown, and no trace of mycelial hyphae could be ob- 

 served undei- a simple lens. In both pots the numerous control plauiS 

 remained free. This was the raaximum development shown by the 

 fungus. On the 14 th day the fungus had disappeared from the plants 

 in one pot; in the other, 2 of the control leaves, as well as 2 of 

 the inoculated leaves, showed small Hecks of mycelium, but no 

 conidiophores were visible. 



Such cases as these — and we shall meet with them again in 

 other experiments — where an Oidium is apparently able to infect 

 weakly a certain species of host-plant, and to produce on it a few 

 spores, but yet seems unable to maintain its hold, will require long 

 and careful investigation before their real signiflcance can be esti- 

 mated. It may be that, in some cases, under certain favourable 

 conditions the fungus could exist permanentl}^ and increase on such 

 a host-plant; that, in fact, these cases show us the first Steps by 

 which a fungus gradually accustoms itself to a new host - plant. 

 On the other band it may even be that in some of these cases no 

 true infection takes place at all, and that the weak mycelial patches 

 and the fcAv conidiophores are produced directly by the germinating 



