CHESTNUT BARK DISEASE, 435 
that apparently enter into their success that we had not in mind 
when the experiments were undertaken, namely:—length of 
time the fungus has been in artificial cultivation, age of the 
particular spores used, and time of year of the inoculation. 
This makes it difficult to judge of the results of certain of these 
inoculations, since two or more of these factors may have been 
involved. The final results of our inoculations were determined 
about the second week in October. Of course this gave some 
of the earlier inoculations made in May a much longer time to 
develop than those made in July, although these latter had 
plenty of time to show whether or not they were successful. We 
will consider the results briefly under the following headings. 
Endothia gyrosa versus var. parasitica. Ordinarily it takes 
about a month to determine whether or not an inoculation has 
taken, and even then it is sometimes doubtful, since the tissues 
around the wound often die back for a short distance as the 
result of the mechanical injury. The sum total of our experi- 
ments brings out quite clearly the difference in the parasitic 
nature of these two fungi. For instance, 151 out of all of our 
324 inoculations with var. parasitica, from all sources on all 
hosts, produced more or less evident cankers, that is, 47 per 
cent. were successful; while of the 148 similar inoculations with 
E. gyrosa only 2 took, or about 1 per cent. Of these two, one 
showed only a comparatively small dead area, with fruiting 
pustules, around the point of inoculation, but did not seem to 
continue its growth, while the other was on a dead seedling 
whose roots had been cut off by mice, which no doubt weakened 
it, allowing the fungus to make an excellent growth, and even 
to produce its ascospores. If we take into consideration only 
our inoculations of var. parasitica originally obtained from 
chestnut and inoculated into chestnut sprouts and seedlings, we 
find that out of 232 inoculations 132, or 57 per cent., took, as 
compared with entire failure of E. gyrosa under the same con- 
ditions. None of the 228 check trees in all our experiments 
showed any signs of infection, thus proving that the wounding 
alone was not harmful when protected from infection. 
With the check trees the cutting usually killed a little bark 
on either side, especially if the knife was run under between 
the bark and the wood. This never grew larger, and the callus 
of new tissue formed in the wound was always healthy. With 
