On the Physiology of Parasitism. 119 
of the presence in them of considerable quantities of impurities 
derived from the culture fluid. 
III. On the Entrance of the Fungus into the Host Tissue. 
Fungi of the type of Botrytis cinerea are able, in favourable 
circumstances, to attack their host when simply brought in contact 
with the external surface of the latter. Their entrance is not 
conditional upon the existence of pre-formed injuries as is the case 
with so-called “ wound ” parasites, nor do their germ tubes as a 
rule enter by way of the stomata or other natural apertures, but 
penetrate directly through the epidermis. A study of the mechanism 
of this process of penetration is obviously of fundamental importance. 
The dominant view on this question is that due to de Bary (l.c.). 
In the case of the allied fungus, Sclerotinia Libertiana, he stated that 
the fungus, when placed in contact with the epidermis, possessed 
the power of killing the underlying cells of the host before any 
actual penetration of the latter had taken place. In consequence 
of this an exosmosis of food material takes place on to the epidermis. 
This serves to promote the growth of the fungus which is now able 
to penetrate directly into the underlying cells. 
In the light of the above view it was important to determine 
whether the extract employed in the present investigation was able 
to exert any influence when placed on the outside of the host 
plant. Such action would consist in the death and disintegration 
of the underlying epidermal cells, with or without an alteration of 
the mechanical properties of the cuticle. 
These experiments were carried out in very great detail and on 
a considerable number of plants. Though in a very few cases 
traces of action by the fungal extract under these conditions could 
not altogether be discounted, the evidence obtained was over¬ 
whelmingly in favour of the opposite view. In the great majority 
of cases no action whatever on the part of the extract when laid 
on the surface of plant structures could be demonstrated. The 
striking nature of this feature may be illustrated by mention of a 
particular experiment. A fungal extract when laid on a rose petal 
for twenty-four hours produced no change whatever. After this 
lapse of time, injection of the same liquid into the same petal 
produced complete disintegration within half an hour. 
It was thus clear that the extract under consideration is unable 
to penetrate cuticularised tissue and thus cannot explain the killing- 
before-penetration effect described by de Bary. It was necessary 
