268 li. M. Bceleii—The Viscosity of Ice. 



and the coefficient of viscosity 7 is one-third of this 



= 27-83 X 10'=. 



Owing to the unfortunate uncertainty of the mean temperatures at 

 which Main's experiments were made, a perfectly satisfactory com- 

 parison cannot be made of the viscosities in Table I. 



On the Diagram, however, havebeenplotted the viscosities calculated 

 from Main's figures. They show, but not very regularly, an increasing- 

 viscosity with falling temperature. The curve gives about the mean 

 values of his results ; but probably gives too high a viscosity, as his 

 temperatures were taken during the day and maj^ have been lower 

 during the night. From the curve on this diagram the viscosity at 

 0'0° C. would appear to be about 6 x 10'-. 



I have already attempted to ascertain the viscosit)' of some of the' 

 Swiss glaciers from their thickness, slope, and rate of flow.^ The 

 results obtained are given in Table II. 



TABLE II. 



Mer de Glace. Mokteratsch. 



At Trilaporte . . 29-48 x 10'- High up . . 292-2 x 10''- 



At Les Fonts . . 27-51 x 10'= Middle . . 143-4 x 10'- 



Above Montauvert . 27-64 x 10= Near end . . 92-31 x 10^= 



Below Montauvert . 24-16 x 10'* 



Great Aletsch. Lo'sver Grindelwald. 



162-2 X 10'- 3-274 X 10'= 



Mr. Connell found that ice crystals sheared readih' without fracture 

 in a direction at right angles to the optic axes. From his experiments 

 I calculated that the viscosity in this direction was only 2 X 10'".= 



Main's experiments, we have seen, were made with artificially 

 frozen ice. Ice frozen in a metal mould is generally fine-grained, 

 the mass consisting of needles of ice radiating from the centre to the 

 sides. In such cases the optic axes of the crystals would be at riglit 

 angles to the direction of flow, and the sliear which ice can undergo 

 at right angles to the optic axis would not assist the bar to lengthen 

 very appreciably, even though the viscosity of ice in this direction be 

 as low as 2 X 10'°. This arrangement of the individual crystals 

 would tend to give a high value to the viscosity of the ice Main used ; 

 but the smallness of the crystal grains or needles would perhaps have 

 a greater tendency to make it low. 



Glacier ice, unlike the frozen ice Main used, consists of crystalline 

 granules, varying from the size of a pea to that of a walnut, or even 

 larger. The optic axes are also at all angles, and the ease with which 

 shear takes place at right angles to this axis no doubt tends to lower 

 the viscosity of the mass. 



All things considered it would seem that the ice tested by Main 

 had a somewhat lower viscosity than glacier ice. For the viscosity 

 at 0° C. we get from the Diagram 6 X 10'=, whereas the calculations 

 I made from glacier ice led me to take 125 X 10'= as a mean, or about 

 twenty-one times as great as those given by Main's experiments. 



' Proc. Eoy. Soc, A, vol. Ixxxi, p. 250, 1908. 

 = Ibid., p. 256. 



