Preface. It is a pleasure and a honour to be included in this issue commemorating the centenary of Robert Stoneley's birth. I was, I believe, the last of the small number of research students supervised by Stoneley, and it gives me great pleasure that most of my research has been, by chance, in a field he initiated-(azimuthal) seismic anisotropy. Bob Stoneley was one of the first seismologists to consider azimuthally anisotropic seismic waves (specifically surface waves) in cubic (Stoneley 1955) and orthorhombic symmetries (Stoneley 1963). Although, he considered these papers 'essentially as a development in the theory of elasticity', they were invaluable references to me when I first began to calculate surface waves in an anisotropic Earth. Bob would have been gently amused that a large part of the Earth is now recognized as having orthorhombic anisotropic symmetry.
S U M M A R YThe shear-wave splitting observed along almost all shear-wave ray paths in the Earth's crust is interpreted as the effects of stress-aligned fluid-filled cracks, microcracks, and preferentially oriented pore space. Once away from the free surface, where open joints and fractures may lead to strong anisotropy of 10 per cent or greater, intact ostensibly unfractured crustal rock exhibits a limited range of shear-wave splitting from about 1.5 to 4.5 per cent differential shear-wave velocity anisotropy. Interpreting this velocity anisotropy as normalized crack densities, a factor of less than two in crack radius covers the range from the minimum 1.5 per cent anisotropy observed in intact rock to the 10 per cent observed in heavily cracked almost disaggregated near-surface rocks.This narrow range of crack dimensions and the pronounced effect on rock cohesion suggests that there is a state of fracture criticality at some level of anisotropy between 4.5 and 10 per cent marking the boundary between essentially intact, and heavily fractured rock. When the level of fracture criticality is exceeded, cracking is so severe that there is a breakdown in shear strength, the likelihood of progressive fracturing and the dispersal of pore fluids through enhanced permeability. The range of normalized crack dimensions below fracture criticality is so small in intact rock, that any modification to the crack geometry by even minor changes of conditions or minor deformation (particularly in the presence of high pore-fluid pressures) may change rock from being essentially intact (below fracture criticality) to heavily fractured (above fracture criticality). This recognition of the essential compliance of most crustal rocks, and its effect on shear-wave splitting, has implications for monitoring changes in any conditions affecting the rock mass. These include monitoring changes in reservoir evolution during hydrocarbon production and enhanced oil recovery, and in monitoring changes before and after earthquakes, amongst others.