All my "primitive" mind can think of is the case of "pre-stressed" concrete beams. What is stressed in that case is the steel cable. It is using the tension strength of the cable inside. In "pre-stressed" bow limbs, it would be the glue/epoxy that is used to hold the strength. If we just glue the slats flat, we will be using just the normal stress on the glue. But if we use the glue to hold the slats in the reflexed shape, we will be using the stressed tension strength of the glue/epoxy. That would be like inserting a more tention-strong material inside the bow limb, just like the pre-streesed steel cable. Basically we are changing the composition of the bow limb in the process. Just a thought. But using different glues and causing failures in the glueline might be a useful experiment. Do we have a glue that has the less tension strength than the wood?
The glue line is holding a shear force between the laminates, but this force is distributed over the entire surface area of contact between the laminates. Most glues should be more than able to hold this without issue, unless the glue bond is very marginal to begin with.
In Hickman’s article, he glued up two straight laminates into a high degree of reflex to make each bow limb, and then he mounted these limbs on a handle that held the limbs in a large degree of deflex so the tips were not so far forward of the bow handle. By doing this, bracing the bow actually relieved the glued-in stress on the bow limb laminates. The result is a wood bow that can be kept strung all the time, and drawn further than a conventional design would normally allow. It was a pretty clever design.
I did some experiments with this a number of years ago. By varying the ratio of thickness of the two laminates, you can play all sorts of games. For example, you can offset excess tensile strength of a tensile-strong wood to provide relief on the compression side.
Alan