Sir,

Our work (Reference Mellor and TestaMellor and Testa, 1969) was prompted by disbelief in the claimed Newtonian behaviour of ice at deviator stresses below 1 bar; our results appear to dispose of this claim, or at least to discredit the evidence on which it is based. Our final conclusion was that classical creep tests become impractical for stresses below 0.5 bar. Thus, to the extent of our published conclusions, we seem to be in complete accord with Weertman.

We were careful to point out that the results do not establish a firm stress/strain-rate relation for the low stress range, although they do provide a better approximation to secondary creep rates than those previously available. We did mention in a footnote the similarity between our apparent stress/strain-rate relation and a corresponding one derived from glacier flow observations, which were not subject to a serious time restriction.

While conceding the inadequacy of existing data for the low stress range, I would question the implication that a simple power relation (for ice) must necessarily be maintained over an indefinite range of strain-rates. New studies on the ductile-brittle transition (as yet unpublished) show stress tending to a limiting yield stress at high strain-rates, and many earlier investigations suggest curvature of the log-log plot at low strain-rates. Furthermore, indirect evidence (such as the temperature dependence of creep) causes one to doubt whether the straining of impure polycrystalline ice at very high homologous temperatures can be controlled by a single physical process.

19 May 1969