in the Brontes and their Brown Rust . 235 
I was able to show that the Hemileia on Coffea Arabica will 
infect C. Liberica. 
Since then Eriksson has elaborated a theory of adapted 
parasitism in connexion with the rusts of our cereal grasses, 
which has assumed a degree of importance so grave, and 
so full of consequence for the understanding of the epidemic 
diseases of cereals, that any investigations throwing further 
light on the matter are welcome. Eriksson found, in fact, 
that the so-called rust of wheat ( Puccinia graminis) is a col- 
lective species, the morphological characters of which are now 
very thoroughly known, which behaves differently according 
to the particular host to which it has adapted itself as 
a parasite 1 . For instance, if we take the uredospores grow- 
ing on wheat and sow them on rye, barley, or oats, the 
results are negative. Nevertheless P. graminis occurs on 
these plants, and forms uredospores on them. If, similarly, 
the uredospores from rye be sown on oats, the results are 
negative, whereas if sown on barley they infect it. 
In other words Puccinia graminis , while it infects all the 
above-mentioned plants, and preserves its morphological 
characters on all of them, is so closely adapted to the par- 
ticular host it happens to be on at the time, that the uredo- 
spores from this host can only attack successfully and directly 
either this particular host or a limited number of its im- 
mediate allies. And the same specialized parasitism occurs 
in the case of other species of rusts, growing on different 
hosts. 
It thus becomes evident that we must modify our ideas 
considerably as to the danger of infection of wheat by the 
uredospores of P . graminis growing in its neighbourhood on 
some other grass. Further research showed, for instance, 
that the uredospores of this fungus on the weed-grass Agro - 
pyrum repens will infect rye but not wheat, and similarly with 
that on Elymus arenarius . On the other hand, the uredo- 
spores — of the same Puccinia — growing on Dactylis will 
attack oats but not rye, and similarly in other cases. The 
1 See Eriksson in Bot. Gaz., 1898, vol. xxv, p. 26, for summary and literature. 
