234 Ward. — On Relations between Host and Parasite 
results, since they appear not only of considerable interest 
and importance by themselves, but promise to throw ad- 
ditional light on some vexed questions of parasitism and 
immunity, and on the relations between the host-plant and 
its invading or attacking fungus. 
I shall here confine myself entirely to the genus Bromus 
and the Uredo of the Brown Rust — Puccinia dispersa — leaving 
any questions concerning the Teleutospores and Aecidia for 
further investigation, except in so far as it may be necessary 
to make reference to them in the form of notes in the 
sequel. 
The plan of the work, as thus extended, concerns the 
characters and behaviour of the ‘ seed 5 and the seedlings 
of the species of Bromus examined, together with observa- 
tions on the leaves of young plants, and the systematic 
relationships of these species : the description of methods of 
culture by which these grasses can be grown and kept 
growing absolutely free from danger of accidental infection 
by rust fungi : the means by which pure cultures of the 
Uredo in question can be transferred to such pure cultures 
of the Bromes, and the comparison of infected and non- 
infected plants compared ; and, lastly, the behaviour of the 
Uredo itself on, and in, the different species. 
The facts and experiments soon raised a number of ques- 
tions in their turn, among which those of specialized para- 
sitism, the persistence of the Uredo-form and the delimitation 
of species, and Eriksson’s ‘ mycoplasm * theory are important 
In my work on Hemileia in 1881-2, the uredine which 
caused the coffee-leaf disease in Ceylon, I was much im- 
pressed by the probability, amounting almost to a certainty, 
that the species Hemileia vastatrix on Coffea had originated 
by the special adaptation of a certain species, Hemileia 
Canthii , wild in the native forests on Canthium , to the culti- 
vated coffee. When once this fungus had learned, so to 
speak, to attack the coffee-plant, it proved a disastrous 
parasite capable of becoming epidemic on its new host. 
Complete experimental proof of this was not obtained, but 
