The process of instantaneous recrystallization described by Demorest in this paper represents an attempt to explain the deformation of a polycrystalline mass of ice. Such a mass, in which the basal glide planes are not parallel in all the various ice crystals, is similar to a polycrystalline metal, particularly one with a hexagonal structure such as zinc or magnesium. In a metal in which mechanical twinning is not the chief deformation process, the deformation can occur in two ways, either by deformation of the various grains by slip along their respective glide planes or by relative movement of the grains. In the first process, the slip of one grain will impose forces on the neighbouring grains and these will set up a series of internal stresses, while the second process can only occur if the grain boundaries are smooth and regular or if molecules transfer from one grain to the other, thus making the grain boundary appear to move through the material.

Now it is well known in the case of metals that if after (or during) straining the temperature is raised high enough, the process known as recrystaIlization occurs. In this process (on which an excellent summary article has recently been written by Burke and TurnbullReference Burke and Turnbull ¹ ) new unstrained grains appear and grow at the expense of the strained grains, the final grain size depending on the number of nuclei formed and their rate of growth. There seems little doubt that this process occurs also in ice, and that the change of appearance of glacier ice is due to this process.

Thus the modern view agrees with Demorest’s reasoning except in so far as Demorest imagined an instantaneous appearance of the new strain-free grain, whereas recent experiments have shown that such grains grow at a definite, and sometimes quite slow, rate.