Hertz fracture experiments were carried out on the two surfaces of specimens of float glass and on specimens of polished-plate glass in the condition as-received from the manufacturer and after various etching treatments in hydrofluoric acid. The results are analysed in terms of the theory of flaw statistics originated by W. Weibull. An acceptable degree of agreement is obtained between the experimental results and values predicted using simple flaw distribution functions which have been determined for each of the five surfaces investigated. Brief consideration is given to the objections which have been raised to the application of flaw statistics analysis to this type of experiment. The present results suggest that objections are not justified.
