Abstract
A test has been made of Bliss's hypothesis that when infected tree roots are fumigated with carbon disulphide, Armillaria mellea is killed not by direct fungicidal action, but through the agency of Trichoderma viride, which is relatively tolerant of the fumigant and becomes dominant in the treated soil. Bliss's observation that T. viride becomes much more abundant in soil after fumigation was confirmed, though higher dosages of carbon disulphide tended to select still more resistant fungi, such as ascosporic penicillia of the P. luteum series. When small woody inocula of A. mellea had been incubated for 3 weeks in soil that had been previously fumigated with carbon disulphide at the rate of 828 p.p.m. and then kept for 3 weeks before use, 30% of the inoculum segment-ends failed to produce rhizomorphs when placed in tubes of fresh soil for a viability test. Nevertheless, this indirect effect of soil fumigation cannot wholly account for the total loss of viability when inocula of A. mellea are directly fumigated in soil. It seems, therefore, that carbon disulphide must directly damage at least the peripheral mycelium of A. mellea in the woody host tissues, and so facilitate invasion by T. viride and/or other fungi. With this additional postulate, Bliss's hypothesis seems to account well for all the observed facts.