London, and the Samuel H. Kress Foundation and the HKI. The author
thanks Ian McClure for his support and, above all, for allowing him free-
dom in pursuing these treatments.1 Transverse grain direction.
2 Changes in MC and moisture gradients in wood are the primary causes of wood movement.
Skaar (1988:chap. 4) reviews the topic thoroughly. See also Panshin and de Zeeuw (1970:206).
3 Of course, wood movement as a proportion of cross-grain dimension (percent of movement
across the grain) remains the same, no matter what the panel size.
4 This statement refers mainly to the changes in dimensions and shape that accompany an RH
change prior to equilibration. Dimensions and shape at equilibrium also depend on such
things as the proportion of tangential to radial wood, the set of the wood cells prevailing from
past conditions, and the presence of preparation and paint layers that may influence mechani-
cal restraint and the rate of moisture permeability.
5 The effect will be less if such a plank is positioned closer to the panel’s longitudinal-
grainedges.
6 For example, as RH rises, the uncoated panel back usually swells first in response to a rising
MC. The expansion is resisted by the remaining panel thickness, which has not begun to swell.
If that remaining thickness is less rigid, such as in thin panels or in woods of lower density, the
force of swelling at the back will cause a deflection, producing a concave warp when viewed
from the front. For the same wood density, thicker panels will be more rigid and therefore
have greater resistance to the effect of the swelling.
7 Longitudinal permeability may be 1,000–10,000 times greater than transverse permeability
(Panshin and de Zeeuw 1970:217).
8 Determined by microscopic examination of a cross section.
9 Raffaello Sanzio,Transfiguration(1517–20). Oil on cherry-wood panel, 4100 32790 3 45 mm
thick (average). Vatican Museums.
10 Marette (1961:65–67) gives a frequency distribution by wood type.
11 Not shown, J. M. W. Turner, The Opening of the Wallhalla, 1842,exhibited 1843. Oil, wax, and
resin on mahogany panel, 1130 32010 3 10 (bevel) to 20 mm (middle) thick. Tate Gallery,
London (inv. N00533).
12 See an early use of American mahogany (Swieteniaspp.) in two paintings attributed to
Rembrandt’s studio of the 1640s (Bruyn et al. 1989:668–78). Though not particularly large
paintings, they are both on single planks and are therefore “large” examples in that sense.
Moreover, the planks are from the same tree, and show rather “wild” (very irregular) figure,
making them even more unusual.
13 The number of joints is smaller and the clamping spans are shorter and therefore less awkward.
14 Tommaso Manzuoli, The Visitation,ca. 1560. Oil on poplar panel, 4090 32485 3 45 mm
(original thickness). Trinity Hall, Cambridge, England. HKI treatment no. 194.
15 Vittore Carpaccio, Saint Thomas Aquinas Enthroned between SS. Mark and Louis of Toulouse,
Adored by a Youthful Donor; (above) Virgin and Child with Angels,1507. Oil on poplar (?) panel,
2640 31710 3 30–40 mm thick (visual estimate by author). Staatsgalerie, Stuttgart (inv. 136).
16 Edges roughly parallel to the axial, or longitudinal, grain direction.
17 The panel itself is usually called the primarysupport, or simply the support. A secondary,
orauxiliary,support may be defined as an original or later structure applied to the panel,
whether attached or not, to provide overall reinforcement.
18 Also known as slotted angle, such girders are found in various forms in laboratories in many
countries. They can usually be acquired in various flange widths.
19 HKI treatment no. 73. The painting is on a thin glue-based ground. The glue appears to be
casein, judging from the color, hardness, relative insolubility, and swelling characteristics of aNotes
472 Brewer