266 S al m on , On Specialization of Parasitism in the Erysiphaceae. 



appeared.^) Prof. Marshall Ward informs me that a very sirailar, 

 if not identical, larva is found frequently associated with tlie Uredo 

 on Bromes. 



Tlie method employed in tlie principal series of comparative 

 inoculaticns was as follows. Seeds of the various species of grasses 

 used were germinated in a greenhouse at the Cambridge Botanic 

 Gardens^), Wlien the seedlings were at an age varying from 

 6—31 days inoculation was made in the following manner. Each 

 pot, which contained from 10 — 20 seedlings, was divided across the 

 centre by a strip of wood. In one half all the leaves or some 

 marked ones, were inoculated, whilst the plants in the other halt 

 served as controls. The process of inoculation was performed as 

 foilows, The part of the leaf on which the spores were to be 

 sown was gently rubbed between flngers moistened with distilled 

 water, so that a damp leaf-surface was obtained. A little heap of 

 conidia (always taken from a powdery mass of ripe ones) was then 

 deposited by means of the point of a sterilized scalpel on this damp 

 surface. After this had been done, the point of the scalpel was 

 in every case dipped into a watch-glass containing distilled water, 

 so that the spores left adhering to the scalpel were washed off and 

 floated on the surface of the water. These conidia were examined 

 microscopically after 24 hours, and the degree of germination noted. 

 The inoculated plants were at once covered over with glass beakers, 

 the sides of which were lined with moistened filter-paper, and were 

 placed on a shelf by a north winde w in a laboratory. The pots 

 of plants remained thus covered for 48 hours, at the end of which 

 time the glass beakers were raised to the level of the top of the 

 flowev-pot by means of blocks, so that a free passage of air was 

 permitted. After 24 hours the beakers were removed. With a 

 few exceptions, two pots of each species were inoculated, and of 

 these one was placed in the laboratory. while the other was placed 

 IQ the open air at the foot of a wall facing north, and exposed to 

 rain and sunshine. No marked dilferences were observed during 

 fine w^eather between the infected plants out of doors and those in 

 the laboratory; in wet weather, however, it was observed that the 

 powdery masses of spores on the former were quickly washed away. 

 The temperature in the laboratory,, and the general meterological 

 conditions ont of doors, were kept during the coui'se of all the 

 experiments. These are recorded at Table 15; so that by reference 

 to this the temperature of the laboratory and the general weather 

 conditions during the course of any experiment can be ascertained. 



As in every instance, without exception, in which the conidia 

 of an Oidium were sown on seedlings of its own host-species, füll 

 infection resulted, it may be claimed that the conditions of infection 

 obtained by the above methods were favourable, and that conse- 

 quently füll value may be attached to the negative results obtained. 



*) In the first week of Marcli one fly emerged from its chrysalis. It 

 was sent to Dr. Sharp, of Cambridge, who reported that it belonged to the 

 genus Aqromyza, or to a genus closely allied. 



2) I am glad to have this opportunity of thanking Mr. Hoskin, 

 Foreman of the Cambridge Botanic Gardens , for the trouble and care he 

 has taken in raising these seedlings for me. 



