Marshall Ward . — On a lily -disease. 323 
destroyed (Fig. 6), but in the green parts around there are no 
hyphae. From the densely packed hyphae in the epidermis, 
thin branches descend almost vertically through the tissues 
below (Figs. 6 , 7), and emerge at length through the epidermis 
of the inside of the perianth-lobe, cross the slight inter- 
space between this lobe (sepal) and the edges of the petals 
which it overlaps, and so infect the interior of the bud. As 
I shall have occasion to show later (though I did not know it 
till after infections had been artificially carried out), the 
spread of these hyphae is facilitated by the poisoning action of 
the hyphae on the tissues around them. 
Sections through still younger spots, e. g. a little more 
advanced than Fig. 2, show that the hyphae are as yet entirely 
confined to the cell-walls (Figs. 55 > 5 ^)» * n the swollen 
substance of which they are growing and branching in all 
directions, but especially in a plane parallel to the surface of 
the organ. 
In the neighbourhood of the mycelium, e. g. at the margin 
of the diseased area in Fig. 3, the cell-walls bounding the 
lacunae, and those of the epidermis and guard-cells of the 
stomata, are often found to be swollen and turning brown and 
granular (Fig. 8). This was a phenomenon which greatly 
puzzled me until I found that it is due to the action of a 
soluble ferment excreted by the fungus itself, and which 
slowly diffuses around and kills the cells. 
The mycelium in the tissues is richly-branched, septate, and 
colourless, excepting that with age the cell-walls assume a pale 
sepia-tint. The branches which come to the exterior to form 
conidiophores are also at first colourless : as they grow older 
the cell-walls quickly turn brownish (Fig. 9), as also do the 
ripening conidia. All the parts are filled with a dense fine- 
grained or minutely vacuolated protoplasm, in which I have 
seen no definite nuclei, at any rate with ordinary staining 
reagents. At the same time, it should be mentioned that no 
special search for nuclei has been undertaken in detail. 
The ripe conidium is of an ovoid form, and usually pale 
sepia in colour, and very large for a Botrytis ; its average size 
