of the derivation of acid rocks from basic magmas on other grounds, including the “impossibility” of forming discrete bodies of acidic igneous rocks by removing by filtration under gravity an acidic residual melt amounting only to 5–10% of the original volume of the parent basic magma from a liquid-filled meshwork of closely inter-locking crystals comprising the 95–90% portion of that original magma which had previously crystallised, though it might be possible to remove the acidic residuum by squeezing through intense deformation of the meshwork by externally applied stresses. It is at least open to question whether the local stress adjacent to the Shag Valley fault zone, even if active during Oligocene times, would be sufficient to meet this requirement in a region where so little deformation affected the rest of the region underlain by basic magma during this period, and the same consideration might be raised in connection with the development of the Lebombo rhyolites along a narrow warped and faulted strip through the huge, elsewhere almost undeformed area of Karroo dolerites and Drakensberg basalts. The absence of any rocks of intermediate acidity in very many associations of basic and acid rocks (our own among them) Holmes holds to be a fatal objection to the application of the Bowen hypothesis of the derivation of the latter, since it would be “completely at variance with the results of work on silicate systems.” He states that “the only hypothesis that is genuinely worthy of the name of contrasted. differentiation is that [of Fenner, viz.] distillation of a gaseous phase and its consequent separation from a liquid phase,” but notes that “We do not yet sufficiently understand the nature of magmatic behaviour to describe … the manifestation of such distillation accurately.” (Holmes, 1936b, pp. 231, 236.)

This hypothesis offers no explanation of the marked concentration of the points indicating the ratio SiO₂: NaAlSiO₄: KAlSiO₄ in rocks of rhyolitic, trachytic or phonolitic compositions, whether intrusive or effusive, into the “low temperature trough” of the triangular diagram indicating the consolidation phenomena of this experimentally investigated system, which is demonstrated by the great majority of chemical analyses of rocks within the above range of compositions (Bowen, 1937, p. 18, Fig. 9; Barth, 1939, p. 82, Figs. 54–5; Benson, 1941, p. 542, Fig. 3, and Fig. 4 of this paper) in so striking a manner that, on statistical grounds alone, it is almost impossible that the phenomena indicated could be fortuitous, but almost certainly result from some general physico-chemical evolutionary processes in crystallising magma, whether they be that set forth by Bowen, or processes as yet scarcely adumbrated involved in the alternative hypotheses qualitatively put forward by Fenner and Holmes. In the absence of any quantitative indication of such alternative evolutionary processes in magma no discussion of their possible application to the Shag Valley porphyry can here be attempted. We turn, therefore, to consider the possibility of producing acidic magma through assimilation of sialic formations by invading basic magma and the fractional crystallisation or other modes of differentiation of the resulting syntectic, reserving for the succeeding section of this paper the consideration of Holmes' (1931) hypothesis

of the formation of acidic magmas by the fusion or refusion of sialic formations with lower range of fusion temperatures than the invading basic magma, and without essential mingling of the basic and acidic melts.

There is a strong body of opinion that some assimilation of sialic material into a basic magma is a necessary preliminary to differentiation yielding an abundant richly siliceous residue. This may occur either at great depth, the familiar abyssal assimilation of Daly, or which carried out on a broad scale, according to the view of Barth (1936, p. 350, 1939, p. 80), permits the development of an acid magma where the acid rocks of the Tertiary Circum-Pacific folding zone have been kneaded into the original basaltic magma, or it may occur marginally in relatively small bodies of magma at higher levels within the earth's crust. An often cited example of the latter (following Daly, 1917) occurs at Pigeon Point, Minnesota. Bowen (1928, p. 214), while agreeing that some assimilation of quartzite into basic magma has taken place, holds that it could give rise only to increased amounts of late differentiates similar to those which might form from the original uncontaminated magma by crystallisation differentiation alone. “The rims about the xenoliths are not simply melted xenolith, but essentially normal igneous material of a late stage of the reaction series.” Grout's (1928) restudy of the area [not affected in this regard by Bastin's (1938) later work], lead independently to similar conclusions. “Though some assimilation of the quartzite is indicated, probably much less than a fourth of the granite (Table II, No. 20) is directly or indirectly due to assimilation of sedimentary rock.” “About the same amount of granite would have formed, and its composition would have been about the same had there been no assimilation.” Bowen's comments on the nature of the “rims about the xenoliths,” explains Grout's observation that the material of such rims contains twice as much alkalies as either the quartzite or the enclosing gabbro.

Drusy granophyric patches occurring in the Skaergaard gabbros, interpreted as being formed by recrystallisation of fused xenoliths of the grey gneiss (with 68% SiO₂) invaded by the gabbro, are surrounded by more coarsely granular hybrid granophyres, and the same fused product of gneiss-xenoliths is considered to be the source of the material forming small veins of granophyre extending for a few metres at most through the gabbros, not chilled against but merging into them. But “it is unlikely that the source of the magma of the larger transgressive veins of acid granophyre” (see Table II, No. 26) which are unchilled at their contact with the gabbro, “was so derived, but was a late differentiate from the magma forming the highest unlaminated layered rocks” (basic hedenbergite granophyre with 59% of SiO₂), which they also intersect. Fenner's (1931, p. 549) view that fractional crystallisation would tend to produce a residuum relatively enriched in iron is supported in some measure by the features of the rocks at Kangerdlugssuak formed up to the middle stage of differentiation, but thereafter the differentiation suddenly takes a new direction, and in the latest stage filter-press action has produced a final highly acid residuum now found in veins and sills

cutting all the earlier differentiates. An amount of acid gneiss estimated as being not more than one per cent. of that of the whole intrusive complex, was assimilated into the magma, but it is considered that the granophyre would have formed without such assimilation of acid gneiss, and that such assimilation did not radically affect the nature of the late differentiates but only their amount. “The complete incorporation of acid material into the Skaergaard magma, during the protracted period of cooling, complicates the interpretation of the late stages of differentiation, but it may be safely assumed that it did not significantly modify the early and middle stages” (Wager and Deer, 1939, pp. 185–212, 239, 306–9).

Mountain (1936, 1937, 1944), Polderwaart (1944), and Walker (1940, 1941, 1942) have described many occurrences of granophyric rims about inclusions of sediments in the Karroo dolerites, or at the margin of sills and of veinlets of granophyre derived therefrom injected into or beyond the dolerite originating from the reaction between the basic magma and the sediments—a process of mobilisation or rheomorphic change of sedimentary material. Analyses of these syntectic granophyres are given and are plotted on Fig. 3 and are stated in normative form in Table III.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

A few of them containing nearly 80% normative sialic minerals are indicated by asterisks on Fig. 4 and lie close to the normal curves of Fig. 3 and in or near the “low temperature trough” of Fig. 4. Nevertheless, if Poldervaart's (1944, p. 110) view of the origin of at least one of these granophyres be true of all, they cannot be considered as differentiates from syntectic magma. Walker (1942) and

Poldervaart (1944) consider the differences between their compositions and those of the sediments from which they were severally derived are the result of “transfusion” of various elements from the invading or enclosing magma, following the work of Drs. Reynolds (1936, p. 403) and Holmes (1936, p. 417). Dr. Reynolds found the general order of diffusibility from a lamprophyre magma into quartzite to be: Al₂O₃, K₂O, Na₂O, with minor amounts of P₂O₅, S, NiO, BaO, and SrO. Holmes (loc. cit.) found the order of relative diffusivity from an alkali basic magma into quartzite to be: SO₃, H₂O, K₂O, MnO, Al₂O₃, Na₂O, TiO₂, FeO, CaO, MgO, with minor variations in this sequence in the different areas and samples studied. What is produced is “neither fused silica nor a solution of silica in the material of the enclosing lava. It is a metasomatic replacement product of quartz due to the introduction into the latter of various constituents in proportions surprisingly different from those in which they could have been present in the magmatic part of the lava.” (Holmes). Dr. Reynolds (1936, p. 39) noted that “the circumstances which contributed to the present distribution and relative concentration of the elements in the transfused xenoliths include (a) sequence of introduction into xenoliths, (b) relative rate or power of diffusion through the xenolith, (c) sequence both in time and space of fixation by the xenoliths.” Variation in permeability adds a further cause for the extreme irregularity of the amount of concentration of the several elements transfused from the invading magma into the more or less granophyric, or less often glassy (cf. Holmes, 1936) rocks produced. Concerning South African rocks, Poldervaart (1944, pp. 109–110) comments: “the responsible magma was highly differentiated. Hence … the emanations derived from this magma would (probably) be of a constantly changing character. With all these variable factors it is scarcely surprising that the metasomatic granophyres are of such different compositions.” It is noteworthy, however, that the above list and relative abundance of the typical non-gaseous elements transfused from magma into sediments resembles but is not identical with the list, “Si, Na, K, Fe, Ti, Al, most prominent in the vapour phase, from boiling pegmatitic liquid, among which Ca and Mg should hardly be expected at all.” (Bowen, 1933, pp. 119–120.) Walker (1942) and Poldervaart (1944), however, found that the materials transfused into the sediments in order of relative abundance appeared to be (a) K, Fe, Mg, Al; (b) Ca, Fe, Mg; (c) Ca, Fe, Mg; (d) Fe, Ca, Na, Al, Mg, Ti; and (e) Fe, Ca, Ti, Na, Mg, P respectively in five cases studied, Mg and, in general, Ca being concerned in transfusion, Si and and K being the chief elements lost from the xenoliths. Hence, power to form volatile compounds may not be the chief factor in determining “transfusibility”. We may here recall the development of minute veinlets or sheets of (“transfused”)? alkaline feldspar extending from the margin of the enclosing igneous rock into the quartzose xenoliths in the dolerites of Moeraki (Benson, 1945, p. 297, fig. 3).

But all these “transfused” rocks are associated with relatively large developments of basic igneous masses, are themselves relatively small, and in general holocrystalline. Of one granophyre it is stated