The stress-strain curves generally exhibited a type of stress saturation. This is different from that for either basal or non-basal glide of single crystals of ice, with which a large yield drop or strain-hardening effect were observed respectively. The saturated value of the stress on the curves was considered as the maximum stress or the yield value. The relationship between the strain-rate Ė and the maximum stress σ was expressed as

This relationship can be drawn as a straight line with inclination n on a log-log plot and the softness or the degree of deformability is expressed by the value of A or by the level of strain-rate on the line at the same stress level. Although there was not much difference between L and T specimens as regards the softness for a sample from relatively shallow depth, which exhibited no strong preferred orientation of c-axes, there appear significant differences among T and L specimens with a sample of moderately oriented c-axes from deeper depth: the L specimens are easier to deform than the T specimens. And for samples of very strong preferred orientation (a single maximum in the fabric diagram), the I specimen deforms much easier than the L and T by one or two orders of magnitude of ἐ at the same stress level, in addition to the same tendency between the L and T as stated above. These results are well interpreted by the feasibility of basal glide in polycrystalline aggregates with respect to the preferred orientations in ice specimens.

Our recent experiments on the deformation behaviour under hydrostatic pressure (up to 300 atmospheres) with Antarctic deep-core ice revealed that the strain-rate decreases with increasing pressure while the fracture stress increases. This tendency must be caused by the reduction of mobile dislocations in ice crystals due to closure of cracks. It is already known that a high density of dislocations around the cracks should give rise to the production of efficient sources of mobile dislocations. This explains the softening effect in ice which contains many cracks.

The mechanical properties of the core ice with strong preferred orientation obtained above were applied to gain an understanding of the results of measurements of bore-hole tilting at Byrd Station obtained by Reference Garfield and UedaGarfield and Ueda (1976). The velocity profile was calculated on the assumption of laminar flow, with estimated shear stresses under Byrd Station and an activation energy 66 kJ/mol for considering the temperature effect. For the upper half of ice sheet (< 1 200 m), our predicted velocity agreed well with the relative velocity component in the direction S. 40° W. obtained by Reference WhillansWhillans (1977). However, for the lower part of ice sheet, the velocity predicted without considering the effect of hydrostatic pressure on the strain-rate is much higher than that found by Whillans. If we use results in which the effect of hydrostatic pressure is considered, then the predicted velocity is lower than that by Whillans at middle depths (1 200-1 500 m). To explain the large difference between the value of the velocity at the bottom extrapolated from our estimated profile of velocities and the real bottom velocity estimated from the surface velocity at Byrd Station measured by repeated Doppler satellite fixes, it seems necessary to consider the possibility of basal sliding. This may be not unreasonable, because the existence of water layer at the bed is proved by water flooding into the hole to a height of approximately 50 m above the bed when the bore hole was drilled.