Growth and Structure of the Wood of Larix. 639 
spring zone, may it be supposed that growth had ceased for the year 
before the due time had come for the seasonal peculiarity of autumn 
wood formation with its attendant reduction in the size of the cells. 
It has been remarked that the first year to show the effects of defolia- 
tion has no thick-walled autumn wood, at least at its ultimate boundary, but 
as a rule the cells here are of much smaller size than those formed in 
the spring or early summer (Figs. 1 and 3). We have, as it were, un- 
thickened autumn wood, at the close of a fairly broad ring. Subsequent 
years, however, show much narrower rings, with or without a few rows 
of cells with more or less strongly thickened walls and flattened shape, and 
exhibiting that sharp demarcation from the spring wood that is general in 
the narrower rings of Coniferous timber (Figs. 1, 2, 4, 7, and cp. Fig. 13 from 
a part of a tree dating back to before the commencement of the attack). 
Possibly this recovery means that defoliations were less severe after the first 
two years of the attack ; the only records available were found to be 
inaccurate with regard to individual trees, however they may apply to whole 
plantations. The exceptionally strong development of autumn wood in the 
outermost two rings of the whole of tree D and of the lower parts of trees 
E and F has already been referred to the thinning out of the Thirlmere 
plantations from which these three trees came, but even here it is not very 
clear why a tree can once more begin to produce well-thickened autumn wood 
before its rings have been able to reattain their former breadth. Probably 
the spring wood is formed out of the reserves stored up over winter, while the 
autumn wood is supplied from the assimilated products of the current year. 
Consequently the defoliation endured in any particular year would have 
a greater effect on the autumn wood, — indeed, the spring wood is largely 
formed before the defoliations begin in June. It may be pointed out also 
that, other things being equal, decreasing breadth of ring up to a certain 
point means an increasing proportion of well-formed autumn wood, so that 
slow grown trees often yield the harder timber. 1 But that the hardness 
absolutely depends on the percentage of autumn wood would be an 
erroneous conclusion. 2 Even in different parts of the same tree great 
variations of hardness are to be met with that cannot be referred to differ- 
ences in the ratio of thick-walled to unthickened cells. As a rule it was 
remarked that in the larches investigated the hardest sections, judging 
from the wear of the razor used to cut them, were those where the autumn 
wood was very sharply demarcated. 3 
(c) Resin Duct Formation, 
Finally, attention may be drawn to the presence of abnormal resin- 
cavities in some of the rings that are deficient in autumn wood. Fig. 15 
1 Cp. Schwarz, p. 350 and the foot-note on p. 348. 2 Cp. Hartig (1), p. 40. 
3 Cp. Record, p. 40. 
