William Brown. 
1 20 
therefore to examine the early stages of infection by the actual 
fuugus in order to check the statements of de Bary. 
De Bary’s view necessitates the existence of a toxic substance 
capable of diffusing through cuticularised membrane. This con¬ 
sideration postulates a considerable diffusive capacity on the part 
of this substance so that it should become uniformly distributed (or 
nearly so) throughout the limits of the small drop of nutrient in 
which the fungal spores are placed on the plant. Thus if such a 
toxic substance be present, the underlying tissue should be killed 
as near as may be simultaneously over the whole surface of contact 
of the drop and the plant. In a large number of cases however it 
was shown that the first appearances of attack showed a definite 
pattern within the drop, commencing as a marginal ring and 
spreading centre-wards. Again, on the same hypothesis, the time 
for the attack to develop should depend directly on the concentration 
of spores present, and inversely on the size of the nutrient drop, the 
larger the latter the longer the time required till an effective 
concentration of the toxic substance is reached. Nevertheless it 
was found that within very wide limits the time required for 
attack to manifest itself was independent of the number of 
spores present and of the size of the drop. Also, when drops 
in which infection had just been established were taken up, cleared 
of spores and tested on fresh pieces of plant tissue, no killing action 
whatever was demonstrable. 
Further convincing evidence on this subject was obtained by 
use of a conductivity method. Drops of water with or without 
spores were laid on certain plant structures, and the electrical 
conductivity of the drops determined from time to time. In the case 
of both spore-free and spore-containing drops, a gradual and equal 
increase of conductivity was shown up to a stage when macroscopic 
discolouration of the underlying tissue in the latter case was 
observable. The conductivity in the latter case then increased with 
great rapidity, thus showing that a strong exosmosis was taking 
place from the host cells in virtue of the action exercised by the 
fungus. The drops in which this rapid exosmosis had begun when 
subjected to microscopic examination invariably showed penetration 
by the fungus to have already taken place. Indeed, in a number of 
fortunate cases, early stages of penetration could be demonstrated 
in drops in which the rapid rise of conductivity had not yet begun. 
Thus it is clear that the rapid exosmosis of soluble cell contents 
which is necessary according to the de Bary hypothesis to enable 
