160 ARMILLARIA MELLEA, 
required in devising experiments to elucidate the problems 
connected with it which require solution. But there are 
some general considerations which may lead to the adoption 
of a reasonable view, in keeping with the present state of 
our knowledge. 
No conidial form of reproduction occurs anywhere in the 
life-history of Armillaria, so that the fungus is confined to 
two methods of attack—by spores and by rhizomorphs. In 
my opinion the role of these two organs is quite distinct, 
as will be shown by the following analysis. In the discussion 
the fungus will be regarded as an enemy whose sole object 
is to attack living trees. Actually, this is not quite true to 
fact, as no doubt the fungus is just as happy living sapro- 
phytically on dead stumps as when sapping the life of 
growing trees ; but if it confined its attention to a saprophytic 
existence, we should have no quarrel with it, and it is only 
in its career of depredation that we, as pathologists, are 
interested in it. 
Infection by spores has never, to my knowledge, been 
experimentally demonstrated, but it is safe to formulate 
the following postulates with respect to spore-infection : 
(i) Spores are incapable of infecting a tree through 
healthy bark of root or stem. 
(ii) Infection is not known to take place through stems 
or roots which are so young as not to have 
developed a bark. 
(iii) Consequently infection can only be effected through 
wounds or dead roots or dead stems. 
Further, dead branches are not a suitable medium for 
infection by a root-attacking fungus like Armillaria, pro- 
bably because the surface is not kept sufficiently damp for 
the germination of the spores. And dead roots can only 
very occasionally become exposed to the air-borne spores. 
Thus it is fairly safe to assume that spores can only attack 
a living tree through wounds in the rootstock just at the 
surface of the soil. Owing to mice and other rodents such 
wounds. are not uncommon, but, as they soon become 
