Specialization of Parasitism in the Erysiphaceae. 111 
Lolium temulentum, Festuca clntior, Bromus unioloides and B. sterilis 
were followed by negative results. 
The Oidium on Agropyron repens (Table 2) when sown on A. 
tenerum caused weak but apparently true infection, and apparently 
true infection when sown on A. caninum. No infection followed the 
inoculation of A. glaucum var., and A. acutnm, although in the case 
of the latter species one control leaf bore a powdery Oidium- patch 
on the nineteenth day of the experiment. 1 
The Oidium on Poaprntensis (see Table 3) caused “subinfection” 
on one of the nine inoculated leaves of P. annua, and on one of the 
three inoculated leaves of P. nemoralis. With regard to the 
“subinfection” of the latter species, the few conidiophores that 
were produced died away by the twenty-fifth day of the experiment, 
and although these plants (now flowering) were kept subsequently 
for over a month by the side of plants of P. pratensis covered with 
powdery Oidium- patches, no trace of any further infection occurred 
on P. nemoralis. The present case is of interest as showing that in 
some cases at least plants on which inoculation is followed by the 
phenomenon of “ subinfection ” are, notwithstanding, to be re¬ 
garded as being for all practical purposes “immune” to the 
respective fungus, since these plants prove able, under conditions 
in which virulent infection of the original host-species occurs, to 
complete their full growth safe from all further attacks. 
The inoculation of fifty-three leaves of thirteen species 
belonging to ten other different genera was attended by negative 
results only (see Table 3). 
The Oidium on Dactylis glomerata (Table 4) proved unable to 
infect Avena sativa , Secalc cerealc, Triticum vulgare, Lolium temu¬ 
lentum, Hordeum vulgare, and Agropyron repens. It is worthy of 
note, as bearing on the interesting question of the possession of 
distinctive colour-characters by certain “ biologic forms ” that the 
conidia of the Oidia on Poa pratensis and Dactylis glomerata 
remain permanently snow-white. 
II.— Inoculation-Experiments with the Conidia of 
Sphaerotheca Humuli (DC.) Burr., Erysiphe Cichoracearum 
DC., and E. Galeopsidis DC. 
In the experiments with S. Humuli, the Oidium on Potentilla 
reptans was used (see Table 5). Inoculation of leaves of P. reptans 
was invariably followed by more or less virulent infection, but no 
1 I am indebted to Dr. O. Stapf for kindly determining the 
species of Agropyron used in the experiments. 
