transported or relocated to a new environment, would be an excellent way
to detect changing conditions that might eventually result in dimensional-
change problems. The relative amount and rate of weight gain or loss
could signal developing problems.
With time, the dimensional response of wood may lessen slightly,
in part because hygroscopicity of the wood may decrease or because of the
mechanical effects of repeated shrinkage/swelling cycles or stress setting of
the wood. Nevertheless, experiments with wood taken from artifacts thou-
sands of years old have shown that the wood has retained its hygroscopicity
and its capacity to dimensionally respond to changes in MC. The assump-
tion should therefore prevail that wooden objects, regardless of age, can
move dimensionally when subjected to variable RH conditions.Restrained swelling and compression shrinkage
An important consequence of dimensional behavior occurs when wood is
mechanically restrained from the swelling that would normally be associ-
ated with increased MC. If transverse swelling of wood is restrained, the
effect is that of compression by the amount of the restraint. The conse-
quence is therefore best understood in terms of the mechanical properties
of wood in compression perpendicular to the grain.
As shown in Figure 11, the elastic limit of transverse compressibil-
ity ofwood is typically between 0.5 and 1%, and compression beyond this
elastic limit results in permanent strain, or set. The importance of the low
elastic limit is evident when it is compared quantitatively to typical values
of free swelling ofwood subject to common variation in RH, with its
resultant MC change. For example, consider a panel prepared from tangen-
tially cut boards of poplar with an average tangential shrinkage percentage
(from Table 1) of 8.5%. Suppose further that the panel had been prepared
from wood in equilibrium at 20% RH and mounted into a frame that
would confine it from swelling along its edges, and that the panel were
later subjected to a humidity of 80% until EMC was reached. As shown in
Figure 10, example 2, a change from 20% to 80% RH would be expected to
produce a swelling (negative shrinkage) of approximately 3% in an unre-
strained panel. However, given our restrained panel with an elastic limit of
less than 1%, at least two-thirds of its restrained swelling is manifested as
compression set. If the panel is eventually reconditioned to the original
20% humidity, it would recover only its elastic strain and would shrink to
a dimension some 2% or more smaller than its dimension at the original
MC. This loss of dimension from cyclic moisture variation under restraint
is called compression shrinkage. This mechanism is a very common cause—
and perhaps the one most often incorrectly diagnosed—of dimensional
problems in wooden objects. Too often any loss of dimension of a wooden
component is interpreted simply as “shrinkage,” with the assumption that
MCmust be lower than it was originally.
Cracksand open gaps in painting panels that areattributed to simple
drying and shrinkage may in fact be traceable to compression shrinkage
induced by restrained swelling. The elastic limit in tension perpendicular
tothe grain is of similar magnitude to that in compression—0.5–1%.
However, the compression set accumulated by excessive restrained
swelling cannot simply be reversed by continuing the restraint of the panel
during the drying/shrinkage phase of the cycle, because the amount of ten-
sile strain is limited to about 1.5%, whereupon failure occurs. Therefore, if18 Hoadley
0 1 2
Strain (%)StressABP (^9) +
set
Figure 11
Diagrammatic stress/strain relationship plot-
ted for a wooden element compressed tangen-
tially beyond its proportional/elastic limit (P 9 )
topoint A. When unloaded, strain recovers
only to point B, resulting in permanent set.