Ice observed on the surface of Blue Glacier is classified texturally into three types: coarse bubbly ice, coarse clear ice, and fine ice. The three types occur intercalated to form the observed foliated structure of the bulk glacier ice. Petrofabric study of fine ice reveals consistently a broad maximum in the density of c-axis orientations, centered about the pole of the foliation plane. This single-maximum fabric is in some respects similar to fabrics of stressed ice from polar glaciers, and the textures of fine ice and polar ice are similar. The fine-ice layers, also resemble layers that have recently been produced by rapid shearing deformation of glacier ice in the laboratory. It is inferred that the fine-ice layers in the glacier constitute zones that are undergoing (or have recently undergone) rapid mechanical plastic flow, and that the adjacent coarse-ice layers originate by recrystallization from fine ice and are not now deforming rapidly by mechanical plastic flow (basal glide). Whether the fine-ice layers have predominately a tectonic origin or whether they originate predominately as in-fillings of snow in crevasses in the icefall is not known for certain. Coarse bubbly ice fabrics generally show more than one maximum in the density of c-axis orientations. The statistical significance of multiple-maximum fabrics is tested by a comparison of several independent fabrics from given stress situations, and it is shown that the basic four-maximum pattern is reproducible, though subject to unexplained fluctuations in orientation. The ‘diamond-shaped’ four-maximum pattern is characteristic of ice subjected to long-continued shear stress of persistent orientation, and the long axis of the ‘diamond’ is (approximately) parallel to the direction of the stress vector that acts across the persistent plane of maximum shear stress. It is inferred that the basic features of the pattern develop at some depth within the glacier, and that subsequent deformation has affected the pattern to some extent. The results of recent experimental studies of the origin of ice fabrics are in moderately good agreement with the Blue Glacier observations. Recent theoretical treatments are in sufficient disagreement to be ruled out. A new method of presenting orientation data allows statistical inferences to be drawn directly from the fabric diagrams.