phenocrysts the change in the percentage of anorthite was observed to vary as follows: central zone, An44 varying to An40 on the outer rim of this area; intermediate zone An44–An40; and for the peripheral zone the variation was more pronounced with An44–An34. In the hybrid type with very calcic plagioclase—viz., bytownite, zoning of this type is also present but in occasional crystals where this does not occur a very thin sodic envelope has been precipitated around the anorthite-rich core. The cause of the development of the normal type of zoning is clear enough, and the course of crystallization can be readily determined from the albite-anorthite phase diagram (Bowen, 1913) but the normal oscillatory type, on the other hand, presents a more difficult problem. Phemister (1934) has advanced the theory of diffusion-reaction control, whereby crystals more calcic than the initial liquid separate, resulting in the production of sodic liquid surrounding the crystals. These crystals are then presumed to react with this sodic liquid causing a “concurrent rise of the An. Ab. ratio in the liquid both through reaction and through diffusion.” Finally a renewed deposition of calcic plagioclase results. Hills (1936, p. 50), in the present writer's view, correctly drew attention to a fallacy in Phemister's argument, particularly the latter's hypothesis that on account of a rise of temperature resulting from crystallization, reaction between plagioclase and the sodic enveloping liquid would produce a less sodic zone to the crystals. Hill's view (1936, p. 52) of the origin of what he terms oscillatory—normal zoning does seem to fit the case more accurately. He considered that crystallization, having commenced under conditions of equilibrium, proceeds gradually and produces, through diffusion, a crust surrounding the crystal itself, and a zone in the adjoining liquid that is enriched in soda. Concomitantly with this, the liquid beyond the sphere of reaction becomes supersaturated with lime. Crystallization of plagioclase is now slow or may have ceased altogether; but, when diffusion has increased the concentration of lime in the zone immediately adjacent to the crystals, calcic-rich material will again be precipitated. When the twinning axes of measured plagioclase phenocrysts from some of the New Plymouth rocks are plotted on Nikitin's (1936, Pl. VII) useful diagram an important and constant angular variation from the standard Reinhardt-Nikitin curves was found to exist. This displacement of the poles was most noticeable in the hybrid rocks (P. 9339) of Mataora Island. The position of poles of a number of plagioclase phenocrysts twinned according to the albite law have been plotted in Text fig. 3 in order to show their position relative to the ⊥ (010) curve, and four points are also plotted for crystals in which pericline twinning is developed. This shift of the poles from the standard curves has been noted previously in a number of cases. In Otago, Benson and Turner (1940) considered a pronounced displacement of the twinning axes in the plagioclase of a series of mugearites as due to considerable orthoclase in solid solution. Both Barth (1931) and Barber (1936) showed that intense heating would cause a shift or dispersion of the twin poles of plagioclase, although Barber (1936a) after some later work, somewhat reversed his opinions. However, a number of European workers—viz., Kohler