> With the beam method, the bending stresses start at zero at the point of
> loading, and increase linearly until the clamped end, where the forces
> should be the greatest. This is what we would expect for homogeneous
> materials, such as a steel beam, but variations in wood may cause failures
> at places other than at the clamped end. A kayak bounced off a rock will
> see the highest stresses at the point of contact, and this is better
> represented with the 3-point bending scheme. In this case, the bending
> stresses are the highest at the point of loading and decrease to zero at
> the supported edges.

I'm going to talk myself into a circle: Maybe it makes no difference where the bending stresses occur, at the point of loading or at the clamp. All that matters is that the stresses are caused by bending, and that the tests are administered repeatably.

Dean