268 Salmon , On Specialization of'Parasitism in the Erysiphaceae. 



The first signs of infection of an inoculated plant is the appear- 

 ance at the place where spores were sown of minute white flecks of 

 myceliiun. On this mycehum young conidiophores are at once produced, 

 and are often found to he present as soon as the first trace of mycelium 

 appears. The period of incubation, in normal cases, proved to be, 

 as a rule, 4 days; but sometimes 5 days were required, or in very 

 rare cases, even longer. Neger (4, p. 37) gives 2 — 3, or at the 

 most 4, days as the period of time betw^een inoculation and the pro- 

 duction of conidiophores; and remarks that infection foUows niuch 

 quicker in mid summer, when groups of conidiophores are sometimes 

 visible after 2 days, than in autumn, when the process of infection 

 is perceptibly lengthened, so that 4 days are often requü^ed before 

 conidia are produced. The abnormally w^et or cloudy weather with 

 low temperature (see Table 15) which prevailed dming the coui'se 

 of my experiments doubtless accounts for the longer period of incu- 

 bation that was requü^ed. Many investigators have remarked that 

 cold weather affects similarly the incubation period in the case of 

 the Uredineae. 



In one case infection was followed by a curious change of 

 colour in the epidermal cells of the plant infected. This was in 

 Exper. m\ 2, (Table 2, 2 a, 2 b) where seedling plants of B. com- 

 mutatus were infected with conidia of the Oidium on B. hordeaceus. 

 On the 14 th day after the sowing of the spores, by which time the 

 Oidium patches on B. commutatus caused by the infection were 

 well established, it was very noticeable that the epidermal cells 

 surrounding these patches were of a violet-purple or claret coloui*. 

 The colour was sometimes confined to the epidermal cells on the 

 other side of the leaf and just opposite to where the fungus was; 

 sometimes the coloured cells surroimded the Oidium patches, as weU 

 as occurring on the back. In several cases, noticeably with Oats 

 and Wheat attacked by Oidium, it was observed that when an in- 

 fected leaf began to die all its cells would lose tlieir Chlorophyll 

 and tuim yellow except those immediately surrounding the 

 2ö?«Mm-patches. This fact may be taken as showing that the 

 action of the Oidium is to stimulate the cells of the leaf for some 

 time to greater activity. Marshall Ward (8, p. 299) has pointed 

 out the same fact in connection with the Uredo of the Bromes. 



In a few^ cases some of the coutrol plants in the experiments 

 became infected. This was due to one of two causes, — either to 

 infected leaves rubbing against those of the control plants, or to the 

 powdery spores in the process of inoculation falling on to the controls. 

 Two cases may be mentioned where the former was clearly seen 

 to take place. In Exper. no. 1 (Table l) two blades of the control 

 plants o± no. 1 c became infected on the 10 th day of the experiment. 

 On the 6 th day of the experiment it had been observed that two of 

 the infected leaves, which by this time were bearing ripe conidia, 

 were through their growth brushing against two of the control 

 leaves. It was these two leaves of the controls which, as mentioned 

 above, showed on the 10 th day patches of Oidium. All the other 

 control plants further removed from the infected ones remained per- 

 fectly free, and it was quite clear that a secondary infection 



