BLACK RUST 65 
restrict themselves to a particular species of grass; in course of 
years they became more and more closely adapted to this host, 
and in so doing grew less and less able to infect other species. 
But possibly they have seldom completely lost this power, 
as is shown by the existence of “bridging” species of which 
Ward produced the best evidence in P. bromina; these will be 
referred to later. 
It -is assumed that all the forms of P. graminis will infect 
the Barberry; the restrictions are confined to the alternate 
host. As a result of his experiments, Eriksson divided the 
species into the following “special forms,” which are here called 
“biological” races :— 
f. Secalis—on Rye 
f. Avenae—on Oat 
f. Tritici—on Wheat 
f. Atrae—on Aira 
f. Agrostidis—on Agrostis 
6. f. Poae—on Poa. 
Sot re Oates 
Race 1 grows not only on Rye, but also on Hordeum vulgare, 
H. murinum, Agropyron repens, A. caninum, Elymus arenarius, 
Bromus secalinus etc. (In all these enumerations non-British 
species will be omitted.) 
Race 2, on Oat, and on Arrhenatherum elatius, Dactylis 
glomerata, Alopecurus pratensis, Miliwm effusum, Bromus 
arvensis, B, madritensis, Festuca Myurus, F. sciuroides, F. ovina 
(tenurfolia). 
Race 3, on Wheat, but also though more rarely on Barley, 
Oat, and Rye. 
Race 4, on Aira caespitosa. 
Race 5, on Agrostis canina, A. stolonifera, A. vulgaris. 
Race 6, on Poa compressa, P. caesia, P. pratensis. 
A seventh Race, f. Horde, is sometimes added, though 
Eriksson included it under his f. Tritici or f. Secalis, 
Jaczewski (1910) from numerous inoculation-experiments 
arrived at somewhat different results: he found that he 
could infect Rye only from Agropyron repens, A. caninwin, 
Bromus secalinus, and Dactylis glomerata; Wheat only from A. 
repens, Festuca gigantea, and Lolium perenne; Oat only from 
GU. 5 
