The Structural Conservation of Panel Paintings

was not considered necessary to alter the lattice, and the project was com-

pleted by mounting the assembly in a tray with a spring-bridge support

behind the lattice. The overall result appeared to be perfectly adequate even

though the original design data were so limited.

This method of estimating the lattice flexibility has since been

used successfully on other panels; therefore, although it may appear to be

an arbitrary assessment, the results justify its use until a better method of

calculation can be found.

Panels that have been cradled have frequently been thinned or have had

some surface preparation to enable the cradle to be fitted. While such pre-

vious changes may have contributed to harmful effects suffered by the

panel, they also make the attachment ofanother auxiliary support rela-

tively straightforward.

Recently, conservation work was undertaken on a panel for which it

was appropriate to use a flexible attached auxiliary support. The panel had

not, however, been cradled or thinned, and consequently, the attachment of

the support to an irregular surface presented some difficult problems.

Description

The seventeenth-century Flemish painting Death ofOrpheus,by Alexander

Keirincx and Roelant Savery,^10 measures 1.4 3 2.03 m; it is made up of six

oak boards with doweled and glued horizontal joins. Early in its history,

following some poor board rejoins, an attempt was made to flatten the

panel. Four rigid poplar battens, each 100 mm wide, were glued into

trenched rabbets across the grain of the boards. Shrinkage of the boards

had then caused partial disjoins and some fracturing. In a misconceived

attempt to prevent further damage, butterfly cleats were inserted across

the board joins, while the cross-grain battens were left in place. These

cleats were deeply recessed, with their grain perpendicular to that of the

boards. As would be expected from these contradictory interventions, fur-

ther damage had occurred in the form of fractures at the outer edges of

the butterflies.

Some of the small butterfly cleats had been removed and even

larger ones inserted, causing further fracturing. When the glued surface

joins of the battens failed, the battens were reglued and their ends screwed

to the outer edges of the panel. In one area on the bottom board, this had

recently caused a severe fracture 35 cm long (Figs. 7–12).

At various times during these conservation attempts, areas of the

boards had been crudely thinned, particularly where the large butterflies

were inserted. Otherwise, the boards retained their original thickness,

varying between 6 mm and 10 mm, with consequent steps of up to 4 mm

at the joins. When the panel arrived for treatment, it showed signs of being

highly stressed. When viewed from the front, it was concave, and some

fractures were held open, indicating severe tension.

Before any structural work could be carried out, the panel was

first kept in an environmental enclosure at 75% RH. When equilibrated, its

profileindicated that much ofthe high stress was relieved. The battens,

along with twenty-eight small butterfly cleats and five large ones, were

then removed so that rejoins could be made. The recesses from which the

Second Case Study:

Support for a Panel with

an Irregular Surface

392 Marchant