Infection Experiments with Various Uredinece. 189 
infecting. Inoculations were made to see if these spores infected 
this variety more readily than spores from another variety, but no 
difference was found. 
As it was possible that the unhealthy state of the plants (they 
had been growing in a hot greenhouse) caused them to take the 
rust, plants were inoculated which were somewhat starved but 
were otherwise quite healthy; the result however was the same. 
The young plants were then divided into four groups. A were 
fed as if for exhibition, B were grown normally, C were starved by 
being grown in small pots, and D were grown in a warm greenhouse. 
On inoculating them in July, about the thirteenth day after 
inoculation A shewed whitish flecks spreading over a great part of 
the leaves, B, C, and D shewed a paling of the leaves at the inocu¬ 
lated spots only. On hardening of the affected part of A and 
cutting sections of it the tissue of the leaf was found to be in a 
great part dead, the cells being shrunken till their walls almost 
touched, and the whole dead part staining very deeply. The 
mycelium was difficult to distinguish. The pale parts of B, C and 
D shewed the same features, 
From the above we see that in an almost immune plant the 
capacity to feed the rust so that it develops normally does not 
depend on the state of health of the plant, but on the season of 
the year, though a luxuriant state of growth favours the spread of 
the mycelium. 
Another species of Chrysanthemum (C . Bi-oussonetii) was 
inoculated with rust from C. sinensis, and in this case the mycelium 
made little more headway than it did when Ranunculus Ficaria was 
used as host, but considerable destruction of the surrounding tissue 
took place. A very slight amount of killing was visible in the 
neighbourhood of many of the germ-tubes whose entry was recorded 
in the first part of the paper. So that it seems that whenever a 
germ-tube of any Uredine enters any plant but its own proper host 
a struggle goes on, resulting in the death of the host locally and of 
the parasite, and the more closely related the host is to the proper 
host of the fungus the more prolonged is the struggle and the wider 
the area covered by it, being greatest where the host is an immune 
variety of the proper host; while in normal infection both host and 
parasite continue to live in union for some long time. 
An attempt to confer immunity by grafting failed. Shoots of 
a susceptible variety were grafted on to immune plants and after a 
junction was seen to have been made the new leaves which grew 
were inoculated, but these took the rust as freely as if the shoot 
were growing on its own roots. 
