: Weak glue could not explain the observed behavior. The loading curves were
: completely different for the two orientations tested. In some cases,
: failure did occur along glue lines but in the other cases glue-lines
: remained unbroken. Buckling of the skin in compression was common to these
: specimens.

One thing that I find very interesting is that in the tests you cite, the glass failed on the compression side, yet in the tests cited in Kayakcraft, the glass always failed on the tension side. It would seem that there must be substantial differences in the test methodologies involved. Considering that the stress on a flexed composite panel is primarily in the tension side, I wonder if the recent tests simply applied pressure over a very small area (a sharp point or edge), causing crush failures on the compression side? Does anyone out there have more details on the exact method of stress application in these two sets of tests?