21 Tack is the property (present in some modern vinyl resin adhesives but not in casein glues)
that holds the parts to be joined together while the adhesive is still fresh.
22 Many observations of disassembled supports, as well as expert opinion, confirm that
splinesfulfilled an alignment function only and were not intended to support or reinforce
theconnection.
23 X-shaped cleats (tasselli a doppia coda di rondine,“double dovetail cleats”; or farfalle, “butter-
flies”) were usually mortised into the boards as deep as one-half of the board’s thickness,
withtheir grain running crosswise to the board’s grain, to hold adjacent boards or parts of a
fissured board tightly together.
24 In polyptychs, predelle were often painted on horizontal boards, whereas other sections were
on vertical boards; polyptychs should not be considered as single panels, however.
25 On the contrary, Bomford and coworkers report two relatively small painted crosses made
from vertical boards by the Master of Saint Francis. The work (92.1 3 71 cm) in the National
Gallery, London, was cut from a single plank of poplar, whereas the side terminals of the one
in the Louvre were constructed separately and then attached with wooden dowels (Bomford
et al. 1989).
26 The need to “hold together” the panel does not exist only while it is on display and therefore
subject to the stresses imposed by its own weight and other internal or external static stresses.
Other situations—such as transport, earthquakes, explosions, etc.—may impose exceptional
stresses on panels. Del Serra describes damages, possibly due to transportation, found on
Duccio’s Maestà(Del Serra 1990).
27 Nailed joints do not behave in the way that some widespread but incorrect ideas suggest.
Joints perpendicular to the nail axis may yield significantly, both because of bending of the
nail shaft and because of the giv e ofthe wood (lowstrength perpendicular to the grain).
When appropriately clinched, however, they will resist large pullout forces. Thus, nailed cross-
beams restrain somewhat the transversal shrinkage and swelling of the panel and at the same
time prev ent it from warping (detaching from the crossbeam) (Uzielli 1994).
28 Preboring of nail holes was carried out in at least the first of the two parts to be joined, in
order to to guide the slender nail correctly (the nail was prone to deformation, especially at its
thinner point) and to prev ent fissures from forming in the seasoned wood.
29 Even though evidence shows no general rule, Castelli and coworkers report that two or three
nails could be driven into each board, depending on its width. Nails could be either inserted
from the back, in the more complex back frames such as the “lattice structures” of large
crucifixes and altarpieces, or inserted from the front, in polyptychs; in some supports, the
crossbeams were placed along the edges (making it possible to cover nails on the front face
with engaged frames and predelle) (Castelli, Parri, and Santacesaria 1992).
30 Wooden plugs and plaster proved to be the best insulation against rust, since neither parch-
ment nor cloth proved able to block rust. Parchment also proved to be an unstable basis for
the ground layer.
31 See Buck 1972; Castelli, Parri, and Santacesaria 1992; and Uzielli 1994, among many others.
The use of sliding crossbeams rather than nails is an attempt to provide adequate freedom for
the panel to undergo shrinkage and swelling without generating concentrated and potentially
harmful stresses.
32 Glued joints are stiffer; nailed joints are more yielding. If a joint is both glued and nailed, dur-
ing its normal working life it will not differ from one that has been glued only; the presence of
nails will not increase strength or stiffness.
33 Even more than in the case of the “sliding” crossbeams, the property of sliding applies for
only a limited number of the dovetailed crossbeams. In fact, the higher contact pressure pro-
duced by the inclined walls of the dovetail and by the longitudinal taper generate even higher
friction, which opposes sliding.
34 Obviously such action will hold only as long as the edges of the mortised groove—the weak-
est part ofthe system—are not damaged by insect galleries, decay, or pure mechanical stress.
Evidence shows that many ofthese grooves, whether original or made during later restora-132 Uzielli