104 MR. W. G. P. ELLIS ON A TRICHODERMA 
I cut sections through the boundary-line so as to include in 
the section diseased tissue and a part of the apparently healthy 
thallus. In these sections I found the septate mycelium in the 
cells; and while some segments were full of granular, much 
vacuolated protoplasm, the other segments contained little, if any 
(see figs. 16-17). 
As the diseased area extended over the pan, new white patches 
appeared, always at some distance—as a rule about an inch— 
behind the margin, the intervening zoue being occasionally 
covered (fig. 2) with a white cobweb-like mycelium, similar in 
structure to that contained in the sections of the thallus. In all 
cases these patches underwent the change in colour from white, 
through blue-green, to sap-green. On May 25th zones were 
well shown on the pan; thus white spore-clusters appeared au 
inch or an inch and a half behind the margin of the diseased 
area, the intervening zone being covered with mycelium ; greet 
spore-clusters formed a zone further back, 14 to 24 inches from 
the margin, and hence about an inch from the white-patch zone. 
Thus it seemed that conidia were produced only when the 
mycelium had obtained sufficient nutriment from the cells of the 
host which had been killed by it. This appearance of zones was 
by no means constant: in other words, the period usually re 
quired for the completion of the sap-green spore-clusters—eight 
days—was not invariable; and hence later on white, bluish, and 
green clusters were much closer together. In July, when almost 
the entire area was involved in the disease, green patches might 
be found almost at the margin of the diseased area; but by this 
time the apparently dead thallus first diseased had commenced 
to grow again from certain points not really killed, as will be 
described later. 
The plan adopted was to first isolate the fungus, then to 
cultivate it, and next to show how it infected the host; and the 
following pages are based on the experiments conducted with 
these objects in view. 
l I may say at once that the septate mycelium and the conidia 
indicated that the fungus was the conidial phase of some 
Ascomycete, whose resting stage I have not yet obtained. I 
the absence of phases other than conidial an exact identification 
is hardly possible; but the fungus seems closely allied to, if not 
indeed identical with Trichoderma, the conidial stage of à 
