The relationship between microcracking and ice Strength has been examined using triaxial apparatus in which track damage can be inhibited by the imposition of confining pressure. Shear fracture in ice is observed to be a rapid, unstable process with no apparent indication of tensile crack localisation or interaction prior to failure and no accompanying large-scale volumetric changes, at least to within 1 ms of the occurrence of macroscopic failure. Shear fracture strength displays little or no dependence on confinement at moderate pressures (P = 5–20MPa), and there is no evidence of significant crack sliding before macroscopic fracture under these conditions. Where flow with distributed microcracking occurs, yield strength can also remain remarkably unaffected by confining pressure, despite reduced crack damage. Particularly under conditions where microcracks are induced by predominantly elastic strains, they may remain stable and non-interacting even at high volumetric densities.

Formation mechanisms of bent icicles and thin spikes were studied by observations of natural icicles and by artificially growing icicles in a cold laboratory, and the results are discussed with reference to the general growth process of icicles. It is shown that icicles can be bent by the displacement of their roots, change in water supply rates, and wind. Spikes were observed to grow by the freezing of a water column captured at the centre of icicles after their length and diameter growth stopped.

A model of icicle growth has been developed based on an analytical solution of the differential forms of the conservation of energy and mass. The problem has been formulated using dimensionless variables defined as the ratios of the various heat fluxes which determine the icicle’s growth. The evolution of the dimensionless icicle shape has been expressed as a function of the variation of the convective heat transfer with icicle radius. The time interval needed for the icicle to reach its maximum length and the variation of the icicle mass and drip rate are expressed in dimensionless form.

Laboratory results are presented concerning ice creep at minimum creep rate (at ~1% strain) for fine-grained, initially isotropic, polycrystalline samples. The effect on the creep rate of ice density, sample shape (aspect ratio) and size, grain-size and ratio of grain-size to sample size is examined. Provided sample density is above ~0.83 Mg m−3 (i.e. the close-off density), there is no effect of density on ice-creep rate. Results provide no evidence of a creep rate dependence on test sample length for cylindrical samples. Sample diameter, however, does affect creep rate. Over the range of sample diameters studied (16.2 to 90 mm) creep rate decreases monotonically by a factor of ~4. This effect is independent of sample aspect ratio. Experiments examining size effects in simple shear indicate no dependence of minimum flow rate on shape or size in this stress configuration. Two grain-sizes were represented within the samples tested for the effect of sample size. As expected from earlier work, no grain-size effect on minimum creep rate is evident. In addition, there was no evidence of an effect on creep rate of the ratio of grain-size to sample size.

A new theory, in which friction is interpreted as the energy flux required to form surface at contact asperities, is applied to sliding on ice and snow. The results of this theoretical investigation show that in dry friction the relevant contact areas are of almost molecular scale. The properties of the interface layer in ice and snow friction arc poorly known, so that the implications of this new theory are somewhat speculative. However, qualitative agreement with experimental data is good, and the theory provides explanations to the success of some empirically developed methods of improving the frictional properties of skis and sledges.

The failure process of a Wedge-shaped ice edge is modelled by dividing the process in crushing failures at the contact and flaking failures initiating from high pressure in the contact. This type of approach has been shown earlier by Daley to give an explanation of the physical background of the line-like contact observed in ice-crushing tests and in full-scale measurements. Daley's model is developed further by studying the stress field in ice in more detail when applying the Coulomb macroscopic failure criterion to determine the force required for flaking failures. The analysis indicates that the shape of the flaking failure surface to minimise the force required to cause failure follows a logarithmic spiral. The shape of the spiral is related to the inner friction angle of ice and to the geometry of the ice edge. The curved flaking surfaces require considerably less force than straight surfaces used previously, especially with wide wedge angles. Finally, the model developed is used to simulate the time history of a crushing force measured in laboratory crushing tests and a good correspondence is obtained.