The results of calculations are easily verified using prepared
sample battens and weights. It is not suggested that support battens be
specifiedpurely by theoretical calculations but rather that calculations may
serve as a useful shortcut to produce sample sections for empiric evalua-
tion. It then becomes a question of judgment based on experience to
decide whether, or by how much, to alter such a batten to suit the particu-
lar requirement.
It should also be stressed that even for those with no understand-
ing of structural design theory, there is at least one important relationship
included in the equations that should be recognized. This is the correlation
between section thickness and flexibility (as discussed above). In the design
ofa secondary support, or even in the thinning of battens to ease an exist-
ing cradle, the result of reducing thickness by what may appear to be only
a small amount can have a very dramatic effect on the flexibility of the
support. Conversely, it is very easy to produce an auxiliary support many
times more rigid than is necessary to perform its function—with a conse-
quent risk of damaging the panel.Method of attachment
With the form that the flexible battens should take having been established
in principle, the next problem to consider was the method of attachment
to the panel.
The main factors to consider were as follows: It should not be
possible for the battens to seize, thus restricting dimensional changes in
the panel. The attachment ofretaining points to the panel should be
achieved without the creation of rigid glue areas that are larger than nec-
essary or that extend too far across the grain, as this could contribute to
the characteristic “washboard” effect and the tendency to fracture at the
transition edges ofglue areas. And it would be an advantage if the means
ofattachment allowed for removal of the battens.
All of these basic requirements were achieved by the use of
flexible retaining strips against the surface of the battens, held in place
parallel to the panel grain with slotted retaining blocks glued to the panel.
The blocks were made narrow in the cross-grain direction, and their size
was limited according to the number used. The greater the number of
blocks, the less tension each had to bear individually and the smaller the
glue area needed for safe attachment. For compatibility with the panel,
the blocks were made of oak. Evo-Stik polyvinyl acetate (PVA) wood-
working adhesive was used for the glue joins.^4 The number ofretaining
strips, and hence the distribution of blocks, is determined by such factors
as the number of boards making up the panel,^5 surface irregularities that
may make attachment points difficult, areas of weakness that should be
av oided, and original features that one would prefer to leave unobstructed.
Using retaining strips against the face of the battens instead of
anchoring the battens directly to the panel ensured that there was little
risk of seizure occurring. However, it was also necessary to stop the indi-
vidual battens from moving and becoming misaligned. This was done by
linking them together in an accurately spaced configuration, with thin, flat
timber strips used to create a lattice.
Finally, a supporting timber section was made to fit under the bot-
tom edge of the panel. This skid strip was joined to the tips of the lattice.
It provided protection for the weak load-bearing edge, as well as providing388 Marchant