THE LARCH CANKER 63 
trees artificially unless they were first wounded in some way. 
But infection experiments may be incapable of demonstrating 
a phenomenon which is almost ubiquitous in nature, and 
the case of larch canker is one in which it is exceedingly 
difficult to repeat in the culture house a process which we 
can observe in the forest. For every time the mycelium 
successfully grows from a dead branch to the trunk, it 
must fail at least a dozen times. But experiments which 
at the luckiest were only successful in 8 per cent. of the 
inoculations would not by themselves be conclusive, espe- 
cially when we know that minute punctures, too small to 
be seen with a simple lens, may be sufficient to admit the 
fungus, and these might have been overlooked in the experi- 
ment. It is thus by no means easy to get satisfactory 
results with experiments set up to show that Dasyscypha 
can infect living trunks from dead branches, and if such 
experiments did apparently succeed, it is difficult to be 
quite certain that the fungus has not entered in some other 
way. It is for these reasons that the arguments in favour 
of this means of infection have been drawn, not from culture 
experiments, but from observation. 
Conversely, wounds which have been found necessary for 
artificial inoculation are probably of much less importance 
in the forest. I have often seen billhook wounds which 
have healed naturally even in plantations where canker was 
epidemic. And where shooting rides have been cut in larch 
woods there is not generally any increase in the frequency 
of the fungus. Again, when larches are grown as nurses 
for other trees, the branches have regularly to be cut back 
to make room for the main crop, and yet, in such cases, 
cankers are often less numerous than in other plantations 
in which there has been no pruning; indeed, in a plot 
treated in this way in Bagley Wood near Oxford, where 
larch was grown as nurses for deodar, there was no canker 
at all on these nurse trees, whereas a larch plot less than 
200 yds. away was attacked with moderate severity. 
The reason why laboratory experiments give a false 
impression of the importance of wounds is by no means 
