panel had been thinned by sawing (Fig. 3). The relief, now in several
pieces, had to be glued together again.
As the relief was not solid enough to support itself in its frame, a
new structure had to be built. There were many options. The relief is very
irregular: heavy and thick, especially at the left and right margins, and thin
in the center for almost the whole height. A light support was required,
capable of adapting to the potential movements of the original, including
swelling, shrinking, and convex and concave warping. Moreover, the back
of the panel is very uneven. After a few weeks, during which the relief
was left flat without constraints, a cradle design was selected: it was to be
made ofsmall balsa-wood pieces, 10 cm long, 4 cm wide, and 1.5 cm thick,
glued in two staggered layers, with the grain direction following that of the
relief. The size chosen for the pieces was related to the width, height, and
thickness of the relief, as well as to the irregularities of the surface.
The back of the panel was very uneven and had many holes, which
needed to be filled in to even out the surface to some extent. Sheets of
limewood veneer, with the edges thinned down and the angles and edges
rounded, were adhered to the panel with Keimfix and clamped. After sev-
eral attempts to fill other holes with various glues, these other cavities were
filled with sifted limewood sawdust mixed with ethyl cellulose glue in a
toluene solution: this produced a fine, soft, and easily worked elastic paste.
Next, the back ofthe panel was coated with a solution of 10% Paraloid
B72in toluene, to isolate the panel from the wax used to attach the cradle,
thereby preventing penetration of wax into the panel’s wood. This wax is a
50–50 mixtureof beeswax and Lascaux 443-95 adhesive wax (pure beeswax
would not have been strong enough; the adhesive wax would have been too
strong). The wax mixture was heated in a double boiler and brushed on the
back of the panel; the mixture was then warmed with an industrial-type
heat gun to spread it evenly in a thin layer.
The cradle was started in a vertical line in the center of the panel.
The balsa pieces were dipped in the hot wax and arranged side by side as
one might build a wall. However, before they were actually glued, they
wereset into place to see how well theyfit. Ifthere was a gap betweenT C R A A M S 361Figure 3
Martin Schaffner (attrib.), Annunciation.Saw
marks on the back ofthe relief.