I recall a thread on this topic that was interesting until degenerating into a flame war. If stiffness is resistance to deformation and strength is ability to absorb loads without breaking, it seems logical that increased stiffness does not necessarily equal increased strength, for the reason stated below. We come back to the problem that wood has tremendous strength along the grain, but very low strength across the grain. That is, the force necessary to separate adjacent wood fibers is much less than that required to lengthen or shorten them, and that in a nutshell is the paddle shaft problem.

Perhaps one solution is not to use glass as reinforcement, which has good strength but is brittle, but rather dynel or one of the other polyesters or aramids, which allow some elongation before breaking? I seem to recall that dynel has about 5% elongation at break, vs. less than 1% for glass. Can't remember the source, so am happy to stand corrected if wrong. But the JR Sweet catalog carries dynel tube for this application, see below (scroll down to Dynel Sleeve & Cord).

: As I remember, Sam Mcfadden's experiment has proved that encasting wood in
: fiberglass does not increase the utimate breaking strength, but rather
: lower it.
: The theory is once a minor break starts, the localized stress is unable to
: redistribute to other area, thereby, concentrated in one single spot.
: Result in the utimate failure of the entire unit. If component (either
: wood or fiberglass) is designed to be able to handle the stess single
: handly by overbuilding, then, why do we need the other?
: Sam, what is your opinion?