the 3·75% corundum in the norm, and may be due to the presence of some argillite fragments in the analysed sample. The presence of this igneous breccia is of particular significance for deciphering the conditions under which the syenite itself crystallized. The composition of the syenite and the composition of the groundmass of the igneous breccia correspond so closely that they must have formed from the same magma. The presence of the igneous breccia is believed to indicate that the syenite crystallized under a comparatively thin and weak roof. As the syenite crystallized the remaining magmatic liquid became more highly charged with volatiles (chiefly water) until the pressure was sufficient to cause an explosive shattering of the roof (partly greywacke and argillite and partly previously solidified syenite). The rapid expulsion of the remaining magma into this shattered mass cemented it together into this igneous breccia, at the same time giving rise to the remarkable porosity and friability of the syenite. Nature and Origin of the Rocks The chemical affinities of the rocks are perhaps most simply and clearly expressed by the alkali-lime index, as introduced by Peacock (1931). By plotting the analyses of the igneous rocks of this area the alkali-lime index is found to be 52·5, which would place these rocks in Peacock's alkali-calcic group, not far from the boundary with alkalic group, which is at an alkali-lime index of 51. This correlates well with the mineralogy of the rocks themselves. In general terms the alkalic group and the alkali-calcic group are distinguished mineralogically by the occurrence of feldspathoids in the former and their absence in the latter. The Mandamus-Pahau intrusive rocks are free from feldspathoids (except for the sodalite phonolite, which probably represents an extreme differentiate) but are distinctly alkaline in character, as evidenced by the biotite and the small amount of anorthoclase in the gabbro, and the alkaline pyroxenes and amphiboles in the syenite and trachytes. The relationship of the Mandamus-Pahau intrusives—geological, mineralogical, and chemical—suggests that they have differentiated from a common parent magma. The visible amount of the syenite is very much greater than that of the gabbro (it is of course dangerous to draw conclusions regarding the absolute amounts of the different intrusives on the basis of surface exposure). With these circumstances in mind, the writer suggests that the parent magma of the intrusive rocks of the Mandamus-Pahau area was intermediate in composition between the syenite and the gabbro, and that the gabbro represents an accumulation of early-formed crystals from this magma. This accumulation of early-formed crystals was separated from the remaining liquid by pressure due to crustal movements, and this liquid was injected into higher levels in the crust and formed the syenites and trachytes. This concept of the origin of the syenite and trachytes as the products of crystallization of a magmatic liquid from which the products of early crystallization had been removed is supported by certain criteria developed by Bowen (1937). He points out that the laboratory study of silicate systems indicates that the residual liquids from the fractional crystallization of complex silicate magmas must be enriched