All the framing labels correspond to the labels on the keel, and there are no anomalies. This strongly supports the conclusion that all of La Belle's frames were placed in their location on the keel for the first time in 1684. If any of the labeled framing pieces or complete frames were replaced individually prior to 1684, there would be no reason for the shipwrights to label or relabel them. The extensive and consistent labeling on La Belle only makes sense—serves a purpose—in the context of a onetime construction event or the preparation of timbers for a kit ship for future assembly. It is difficult to imagine, as Carrell proposes (Carrell 2003:216–221, 407–408), that "a ship with an existing older method of design" was so extensively rebuilt in 1684 that complete frame sets from other older vessels were incorporated and all the components were relabeled and marked with surmarks. It must be kept in mind that a vessel of La Belle's small size was unlikely to garner such attention within the context of the French naval establishment.

Replacing worn-out planking is fairly straightforward, but replacing parts of the framing situated under ceiling planking and stringers is a massive undertaking. Such work would leave extensive and unambiguous evidence—for example, differences in fastening patterns, differences in frame layout, old fastening holes, damage edges, areas of cut-out rot, or repair pieces. Quite the contrary, when first seeing La Belle's framing timbers, it was striking how crisp the edges were and how little wear and tear the timbers displayed. La Belle's remains in no way gave the impression of being assembled from components of older ships.

Even if one still presumed such extensive rebuilding work was done in 1684, it would not explain the scattering of sampled felling dates throughout the first half of the seventeenth century and earlier. All but five of the sampled timbers have felling dates earlier than 1666, the date of the establishment of the naval dockyard at Rochefort. Even disregarding the samples with felling dates from the sixteenth century and the two samples with fifteenth-century felling dates, the other samples have felling dates in each decade of the seventeenth century between 1613 and 1683. Since none of the sampled timber felling dates group nicely before an "original" building date, the dendrochronological dates themselves do not provide a reason for shifting the known building date of La Belle.

Why some of the sampled timbers have such early felling dates and why the felling dates are scattered over so many decades are legitimate questions. The dendrochronological study also revealed that almost all the sampled timbers came from the greater Rochefort/Charente region (Bruseth and Turner 2005:80; Carrell 2003:254), and I suspect the answer to the mystery of the timber dates lies in understanding the buildup of the Rochefort arsenal's timber stockpiles from the time of its establishment in 1666. NEXT