Jacques Heyman in his article Truesdell and the History of the Theory of Structures, mentions the following:
The first full-scale tests on building frames were made in the 1920s in London, and the results were, in the fullest sense of the word, astounding. The internal stress resultants (in this case, bending stresses in beams) that were measured in practice bore almost no relation to those confidently calculated by the elastic (Navier) designer. The investigators quickly found what was wrong; while the equilibrium equations were obeyed, and the elastic equations of deformation were found to be almost exact, the boundary conditions used in the analysis did not represent reality. A small error in manufacture, a tiny dimensional mismatch, an inevitable forcing together of members during construction - all these led to a great change in the values of internal stresses.
Where can I find more information about these astounding results?