globulites and numerous gas-pores, have also been noted in such rocks. The term buchite has been used in this sense by several authors including Rosenbusch (1908, pp. 1295–6), Harker (1904, pp. 245–6, 1932, pp. 70–1) and Tomkeieff (loc. cit.), the last noting the presence of new-formed crystals of plagioclase and sanidine suggesting “a certain amount of transfusion between the magma and the sediments,” a view which has been supported by Reynolds (1936) and others. Harker (1932), quoting Lemberg's analyses, stated that such semi-vitreous buchites may contain 3–5% of water rising to 10–12% in the purely glassy material. Prohaska (1886) and Flett (1908) extended the connotation of buchite to included vitrified shales and phyllites. Holmes (1920) thereunder included also phyllitic and other rocks vitrified by the effects of friction in mylonitised crush-belts, and Thomas (1922) vitrified aluminous clays from which sapphire, spinel, anorthite and mullite might crystallise. Jugovics (1933), however, restricted the use of the term to more or less vitrified arenaceous rocks as originally proposed, and applied the old term basalt-jasper to the altered argillaceous rocks, as did Richarz (1924, p. 689) and Rosenbusch (1908, p. 1298) who, however, included vitrified greywackes under this term. Tomkeieff (loc. cit.), remarking that this term is self-contradictory and is not used in Britain, made the alternative suggestion (a) to use buchite in its original sense and porcellanite for the products of thermally altered argillaceous rocks, or (b) to give to these terms a textural rather than a chemical significance, using buchite for vitreous or semi-vitreous material whether arenacous or argillaceous, and porcellanite for unvitrified but finely crystalline contact products of such sediments, with hornfels for their less finely crystallised products. Since the amount of argillaceous material in the Moeraki mudstones is variable, though usually small, and some of them are rather calcareous, the latter alternative is here adopted. Though no sharp distinction has been observed between porcellanite and buchite (von Haast used the name porcellain-jasper to cover both), the porcellanite is “a compact rock of light colour with the appearance of unglazed porcelain” (cf. Holmes, 1920), and is formed by the induration of mudstone making a semi-conchoidally fracturing rock within a few feet of the intrusion; the buchitic character appears nearer to the intrusive mass, where the curving fractures assume smaller radius, and a vitreous lustre is observable in hand-specimens, while under the microscope an isotropic base is observable which may finally make up the greater part of the rock, though enclosing small residual and often corroded quartz grains, traces of micas, which may be residual, and perhaps scanty new-formed feldspar microlites.

The grey mudstone is almost unaltered within three feet of the base of the Moeraki Point intrusion, and in the middle of the narrow (14ft) strip of sediment between the bottom of the Tawhiroko sheet and its basal spur. The latter rock (5744) is a calcareous mudstone with casts of foraminifera, radiolaria and sponge-spicules. Within two feet of the intrusive sheet (5894) or four of the basal spur (5745) conchoidal fracture is well developed, and sparse residual grains (< 0.02 mm.) of quartz with a little feldspar, biotite and

sericite occur in a dominant isotropic matrix still showing the outlines of the casts of these siliceous organisms, but rendered turbid by dust-like iron-ore and argillaceous material. The refractive index of this glass, according to Dr. Hutton, is usually 1.470 ± 0 003, but ranges from 1.465 to 1.475. A similar rock (5893) forms the margin of the huge slab of altered mudstone 5–10 feet thick and over 50 yards long standing almost vertically in the dolerite of the Main Moeraki sheet south of Okahau Point. The fragment of mudstone (5873a) included in the dolerite in the centre of the Tawhiroko sill is still more turbid, and its very abundant radiolarian casts are filled with carbonates. Similar to this is a sample (5928) obtained four inches from the basal spur of this sheet, differing from the above in the tawny colour of the isotropic material resulting from a greater content of dust-like haematite, and the rather greater proportion of small residual quartz grains. The possibility that hot magmatic and connate water acting on the finely divided quartz grains in this rock might have developed opal, and that such might be the isotropic matrix of this rock suggested investigation, the results of which are indicated below:—

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

A Solubility on boiling finely ground material in 10% NaOH for two hours.

B Water given off above 110° C.

C Refractive index or indices.

D Specific gravity.

(a) Data from Doelter (1914).

(b) After Lemberg cited by Harker (1932).

(c) Kindly determined by Dr. S. N. Slater.

(d) Kindly determined by Dr. C. O. Hutton, who notes that 1.470 ± 0.003 measured in Na light is the usual figure, though one flake gave the higher value.

(e) Data from Tomkeieff (loc. cit. p. 56). R.1. = 1.472–1.497 in Holmes (1936).

It would appear from the above that buchite may be accepted as the nature of this material (5928). The slight variation in properties will probably be due to the variation in content of lime. The rock is transversed by many irregular cracks containing calcite.

A small chip (5839) enclosed in a thin tachylytic dyke invading the Tawhiroko sheet (see Benson, 1944, p. 96, fig. 7) may represent a less rapidly cooled product of semi-fused material which has taken up material from the magma, for though it contains minute angular or embayed relics of quartz grains, the groundmass seems to have devitrified into a mass of skeletal crystals of alkaline feldspar with a general parallel elongation, and occasional flakes of biotite.

Contrasting with these is a more calcareous rock, which was probably plastic when taken up as an inclusion into (5833) the dolerite near Matiaha Head (see Benson, 1943, p. 135, fig. 8). This

in a glass containing sheaf-like bundles of augite-needles (< 0.6 mm. long), dusty or finely octahedral magnetite and rectilinearly branching or curved feldspar microlites. In the distal veinlets of this material the matrix almost is wholly glassy.

Within the xenolith there was a sparse development of tridymite along cracks extending 0.5 mm. from the contact adjacent to which it has reverted to quartz, but beyond 0.3 mm. from the contact what remains has such low birefringence and refractive index that it is probably still tridymite, though in units too indistinct to display characteristic outlines. Heating to over 870° C. and chilling too rapid to permit complete reversion of the tridymite to quartz may have occurred here. In association with the quartz there are (5837)) rarely large crystals (< 1.0 × 7 mm.) of andesine (An 41) which has no augite corona, but frays out against the tholeiite into almost rectangular platelets (0 05 × 0.02 mm.) separated by films of basic magma which have entered the cleavage planes parallel to 100 and 001 as Lacroix (1893, p. 572) described as casket-structure and figured, and Knopf (1938, p. 374) has also noted. Neither of these writers has, however, noted that, as determined by Dr. F. J. Turner using universal stage methods on our rock, this marginally altered feldspar has been made over into An 58, approximately the composition of the plagioclase in the adjacent tholeiite, or has been extended by deposition of such feldspar upon it. The opening out of these cleavages, but without the injection of glass may be seen adjacent to a thin tachylitic veinlet traversing the main portion of this crystal. (See Fig. 1A).

Fig. 1.—A. Xenocryst of andesine with new-formed margin of labradorite (5837) in the tholeiitic dyke of Te Karipi, Moeraki B. Fragment of quartz schist with well-developed augite-corona around devitrified marginal zone. In dolerite dyke south of Moeraki (5834). C. and D. Xenolith of vein-quartz (5915) from “pseudo-conglomeratic” dolerite at Tawhiroko Point, Moeraki.

In view of Bowen's (1922, pp. 524–531) discussion of the reaction effects between xenocrystic plagioclase and plagioclase-bearing magma, this feature is worthy of further comment. The xenolithic andesine (part of a plagioclase-bearing quartzose xenolith the origin of which will be considered below) is much more sodic than the

solid plagioclase (about An 84) with which the plagioclase in the basaltic melt would be in equilibrium (cf. Bowen, 1922, fig. 1), and could not remain unmolten in contact with basaltic melt unless its temperature was less than 1240° C., which it almost certainly would be. There will, however, be a notable solution of the xenolithic andesine, and a concomitant crystallisation of more basic plagioclase tending towards a composition in equilibrium with that in the melt. Even though, on account of the approximate equality between the density of the xenolith and the melt, there may not be much relative movement between these, the reaction cannot proceed far, since the melt is cooling rapidly, and the marginal zone of basified plagioclase produced is narrow.

Fig. 2—Margin of xenolith of gabbro in basalt (B344) of volcanic neck at Dundas, near Sydney, New South Wales.

Contrasting conditions may be exemplified by a gabbroid xenolith (344) in the rather potassic basalt filling a neck at Dundas, near Sydney, New South Wales [see Benson (1910) Pl. 24, fig. 5, and fig. 2 herewith, which is drawn as if the transverse crack in the side were not present]. Here the normative composition of the basalt-plagioclase is An 44 or An 48, according to whether the orthoclase molecule probably present therein may or may not function as albite in the reaction.^† Such a liquidus composition would be in equilibrium with the solidus An75 or An78, which is but little more calcic than the plagioclase of the xenolith shown by Dr. Turner's universal stage measurement to be An 67. In view of the greater density-difference between the xenolith and the basalt-magma, the former may have been sinking through the latter while inagmatic corrosion was in progress, coming to rest only when the basalt consolidated. Hence, in spite of the slower cooling of the basaltic magma in this case,

some solution of the xenolithic plagioclase (but not augite) may be inferred from the nature of its boundary against the basalt, though not even a thin film of new-formed plagioclase can be detected as having been deposited on the corroded surface of the plagioclase of the xenolith.

As a final example of inclusions in dykes may be mentioned one of quartz-schist (5852) in the dyke crossing the railway line a mile and a-half south-east of Hillgrove. This small xenolith has a very well developed augite-corona separated from it by pale green chloritic material probably replacing glass. (See Fig. 1B.) The medium grain-size and holocrystalline nature of the host-rock suggest slower cooling than occurred in the tholeiitic dyke at Te Karipi, a feature transitional into the next group of xenoliths to be discussed.^*

(b) Xenoliths Near the Lower Margin of the Tawhiroko Sheet.

About 15 feet above the base of the Tawhiroko sheet there is traceable for over 50 yards through the slightly vesicular olivine tholeiite (5834) a band of varied siliceous xenoliths showing more advanced reaction. A fragment of vein-quartz (5756), illustrated in Fig. 3, shows ramifying veinlets of semi-vitreous basalt very irregular in width, locally, but by no means always, following the original intergranular boundaries and dissolving their way into the quartz grains, appearing in section as rounded patches in the centre of a grain as in Y of Fig. 3. These veinlets contain zoned prisms of andesine squarish in cross section, < 0.4 mm. long and often with skeletal extensions, and less abundant short or elongated (< 0.50 mm.) prisms of pale yellow-brown augite, and long needles of apatite (or sillimanite?) with octahedra of magnetite and platelets of ilmenite. The glassy matrix contains needles of augite, and belonites or margarites of iron ores. Around many of the feldspars the glass has devitrified with formation of a minutely micro-spherulitic structure surrounded by a zone of glass blackened by the concentration therein of dust-like particles of ore. The boundaries of these acidified veinlets are sometimes quite sharp against the quartz, but more often they are marked by a narrow palisade of platelets of tridymite reverted to quartz, either rectangular or elongated, with forked ends and twinning resembling that of sanidine. Locally they merge into a thin (0.02–0.05 mm.) ribbon of quartz extending in optical continuity for half a millimetre or more along the contact between the quartz of the xenolith and the ramifying veinlet. The reverted tridymite platelets may also project in denticles from the veinlet into the quartz grains. Elsewhere the microspherulitic material in the veinlet may extend beyond its boundary into cracks in the quartz of the xenolith, with or without accompanying reverted tridymite-platelets or carbonates. Two examples of the occurrence of tridymite are illustrated by inset figures X and Y in Fig. 3. The former illustrates the more common feature of partial solution of quartz in the magma and precipitation therefrom of tridymite tabulae

(< 0.1 mm.) extending denticularly from the residual quartz into the surrounding glass. The other Y shows one of the rounded injections of magma within a large grain of quartz, the centre of which injection is occupied by a remnant of the quartz grain shown in the upper half of the figure, the central and lower portion being occupied by platelets of reverted tridymite. No examples of sectional twinning of the tridymite were observed, but many of the lath-shaped sections had forked ends. (Cf. Knopf, 1938, p. 374.) The general appearance of the tridymite-fringed quartz-grains closely resembles the figures given by Thomas (1922, Pl. 7), Campbell, Day and Stenhouse (1932, Pl. 53) and Reynolds (1940). Surrounding this aggregate is microspherulitic material, a few grains of augite and magnetite and a little residual glass.

The carbonates, chiefly calcite (and aragonite?) with a little limonite-stained siderite, fill small (< 0.3 mm.) vesicles in the glassy material in the veinlets and may replace portions of it and also of the microspherical material, but not the augites crystallising therefrom.

The quartz grains of the xenoliths contain rarely needles of sillimanite, and gaseous bubbles sometimes arranged in bands parallel and near to the intrusive veinlets.

The effects of assimilative reaction with the xenoliths of quartz-schist are more striking. The schistosity is not always marked in the sections examined from here (5754–5759) by any regular direction of elongation of the grains in the quartz mosaic, which may vary from 0 3–1.0 mm. in average diameter in different samples. It is, however, shown (5757) by bands of shearing granulation or by the occurrence of ribbon-quartz (5758, in Fig. 3), or by the elongation of the strings of introduced minerals. The magma injected into these inclusions has penetrated along the planes of weakness forming thin wedges extending 10 mm. or more from the embayed outer contact zone. The reaction-corona of diopsidic augite prisms (with 2V = 54*deg;–49°) is not continuous, but obviously disrupted by wedges of magma penetrating through it into the xenolith. Magma, however, has diffused through the corona and melted and absorbed varying amounts of quartz, producing a brown “buchite” glass in which the bases of the slightly titaniferous augite and ilmenite plates of the corona are embedded, and beyond this are idiomorphic prisms (< 1.0 mm.) of colourless augite in which 2 V = 52° with 48° in a narrow marginal zone in one prism, 42° in another as determined by Dr. Turner. With these are tabulae (< 0.5 mm.) of plagioclase (An 52–40) often more or less skeletal in development with much included glass and microlitic terminal elongations. These may be surrounded by a very fine microspherulitic growth fading out into the surrounding ring of darkened glass. Vesicles occur in this buchite glass, and near them the pyroxene forms much smaller and elongated pale green prisms, probably indicating the entry of acmite into the pyroxene-composition (cf. Lacroix, 1893, p. 19). Indeed, some of the larger colourless augite prisms tend to have greenish mantles. The smaller prisms may project from the glassy matrix into the vesicles. The siliceous nature of this buchitic glass may be inferred from Dr.

Hutton's determination of its refractive index 1.498–1.500 which is almost that (1.495) determined by Knopf (1938, p. 374) for a colourless glass enveloping the grains of quartz in a xenolith of grandiorite enclosed in olivine basalt, or that (1.480–1.497) measured by Holmes (1936, p. 415) for glass “separated from a transfused quartzite in murambite” (a potassic basaltoid lava), or that (“slightly below 1.50” fide Tomkeieff, 1940, p. 57) found for the glassy cement between the quartz-grains of a contact-altered sandstone, now a “lithic buchite” with sanidine (?), augite and cordierite crystallised from the glass. This refractive index would be consistent with a silica content of 70–72% (cf. George, 1924, Holmes loc. cit.) or slightly higher (cf. Tilley, 1922) if, as Harker's (1932, p. 71) comment suggests, the glass contains a little water. The xenolith is, however, so extremely siliceous that probably there has been some mingling of the basalt magma with the silica-glass produced.

Where the basic magma has penetrated the reaction-corona it mingles with the buchite and there is no sharp distinction possible between the doleritic and buchitic portions, both of which have vesicles with borders darkened by dust-like magnetite. (See Fig. 3.)

From both the semi-vitreous (buchite) zone beyond the corona, and from the penetrative magma wedges, a hybrid product may extend in long threads (or veinlets) into the structure-planes of weakness of the schist. Two examples are illustrated in Fig. 3. The longer intrusive wedge extending out from the semi-vitreous zone passes from it into trachytoid material with the composition of trachyandesite, containing but little microspherulitic mesostasis. The flow-structure is marked by approximate parallelism of the feldspar laths, which are dominantly oligoclase with subordinate sanidine. With these are thin prisms (< 0.1–0.4 mm.) of augite much less abundant than in the semi-vitreous zone and aggregated near the vesicles where they have green margins, and a little ilmenite in sparse large grains (< 0.1 mm.) and minute plates, as well as dust-like magnetite also aggregated about the vesicles. This trachytoid zone is not seen in all intrusive wedges, for in many of them (e.g., the right hand wedge of Fig. 3), the microspherulitic zone may follow directly out from the semi-vitreous material. It continues, however, the same differentiation trend. While some oligoclase is present, the bulk of the feldspar laths, now smaller than at the outer contact (< 0.1–0.3 mm.) seems to consist of sanidine. The amount and size (0.05–0.20 mm.) of the pyroxene prisms are less than in the trachytic portion and their greenish colouration is more marked. The dust-like iron ore is also less abundant, and but traces of glass remain near some of the vesicles. Radially branching aggregates of feldspar-fibres occur with centres either at the edge of the intrusion or about the larger microlites within it, and such microspherulitic material dominates. It shows on a minute scale the features seen in spherulites of rhyolites (e.g., as described and figured by Iddings, 1909, p. 225–7) rather than those formed in thermally altered feldspathic sandstones (e.g., Harker, 1932, p. 68, fig. 21B) which, however, seem very similar to the “replacing veinlets of granophyre” (Reynolds, 1938, p. 60,

fig. 4) invading a quartzite xenolith in lamprophyre, and are so regarded by her. The margin of this thin wedge or veinlet of microspherulitic material is commonly sharply marked against the invaded quartz, is bulbous or embayed, but may locally fray out into irregular groups of microlites at the interstices of the quartz-mosaic of the schist. Vesicles, rarely rimmed with dark glass, are present, and the crystallisation of the feldspar-microlites must have been from solution or hydrous “melt” replacing, rather than wedging apart the more permeable sheared portions of the schist. Rosenbusch (1908, pp. 1295–7) describing similar occurrences notes that with increasing distance from the contact-surface the (buchite) glass decreases in amount and in its place there occurs a microlitic felt which leads out finally into the normal cement of the invaded sandstones. Analogous features were observed (Benson, 1914, pp. 450–1) in a xenolith of quartz schist (569) enclosed in a basalt dyke at Gerringong, New South Wales, and attention was called to the decrease in the amount of the coloured minerals in the veinlet with increasing distance from the contact, and its greater abundance in the centre rather than the sides of the veinlet, features also displayed in the Moeraki xenoliths. Miss Reynolds (1936. pp. 373–4; 1938, pp. 51–76) remarks similarly on the sharpness of the lobate boundary of the veinlets found in quartzite xenoliths in hornblendite and also in vosgesite, adding that such sharpness is to be expected when replacements result from compound-formation dependent on diffusion though a solid.

After the crystallisation of the feldspar the composition of the magmatic emanations was changed, and in the distal portions of the intrusive wedge the feldspar-fibres were in part replaced by finely crystallised siderite now forming close packed granules about 0.05 mm. in diameter, while solutions depositing similar material extended into the planes of weakness of the schist and developed replacement-streaks of irregular width penetrating into the crevices in the quartz-fabric and containing residual patches of microspherulitic feldspar. Metasomatic replacement of portions of the glass of the semi-vitreous zone also occurred, though the plagioclase and pyroxene resisted the solutions and may remain sharply bounded against the surrounding carbonates. The vesicles also were lined by sphaero-carbonate. Its yellowish coloration resulting from incipient oxidation facilitates its distinction from calcite (or aragonite), the deposition of which immediately followed and filled the remaining spaces of the vesicles and continued the replacement of glass. Some of the colourless carbonate in the vesicles, in which division into radial sectors is most strongly marked, may be aragonite rather than calcite, but no systematic attempt has been made to distinguish between the two forms of Ca CO 3, though some of the more finely granular colourless carbonate replacing glass has a clearly rhombohedral form. Several other specimens of quartz-schist from the band of xenoliths near the base of the Tawhiroko Sheet show long strings of irregularly rounded microspherulitic aggregates with carbonates running generally along the schistosity planes and doubtless representing the partially replaced distal portions of wedges or laminae of invading magma.

C. More Slowly Chilled Xenoliths.

About the middle of the Tawhiroko intrusive sheet—i.e., about 80 feet above the base, are several patches of dolerite so extremely rich in xenoliths as to resemble a conglomerate (see Pl. 21B), and some of these display features differing from those described above. Two of those studied have been permeated with vaguely bounded injections of magma, which have reacted strongly with the minerals of the xenoliths. They, however, are peculiar in their abundant content of coarsely granular andesine, in which they resemble the probably similarly derived coarsely granular quartz-andesine rock (5837) of the Te Karipi dyke described above (p. 293), but contain in addition coarsely crystalline orthoclase, and hypersthene. It is difficult to determine from the small amount of material available for examination what were the conditions of origin of these xenoliths. Tentatively it is suggested that two phases of reaction may have been concerned, a deep-seated stage when coarsely schistose vein-quartz reacted with the parent magma of the Tawhiroko sheet under plutonic conditions, and a subsequent stage when reactions occurred under differing environments as the enveloping dolerite magma was carrying the xenoliths up into and cooling within that sheet. The more coarsely granular specimen (5917) is illustrated by Fig. 4, the less coarsely (5919) is generally similar but differs in minor details. The most abundant minerals are quartz and andesine varying a little in relative amount from place to place. The quartz forms roughly lenticular or nearly equi-dimensional grains, with a general parallelism of elongation, and reaching a maximum length of about 8 mm. The grains show embayed or rounded outlines, in part resulting from magma-corrosion, and there is little trace of optical strain. The largest plagioclase crystal is about 10 × 8 mm., but the majority are considerably smaller. The larger grains rarely show well marked twinning and may indeed be untwinned. Commonly, as a result apparently of subsequent thermal metamorphism, they are more or less broken up into a mosaic of rounded or prismoid grains a millimetre or less in diameter (cf. Harker, 1932, p. 109), and these may show strain-produced twinning sufficiently clearly marked to permit determination of the composition of the feldspar. Dr. Turner has found by universal stage methods the composition of three unzoned grains is An 38, 40, 44 agreeing with the determination (An 42) of the large grain in the Te Karipi dyke. Orthoclase (possibly microperthitic) is smaller in amount, less abundant than plagioclase, and makes rather smaller and more irregularly-shaped grains, and is difficult to determine save by its refractive index, because of the alterations that have taken place. A little micropegmatitic quartz occurs in the orthoclase adjacent to the large quartz-grains of the xenolith. Hypersthene forms a few crystals only, roughly equidimensional the original maximum diameter being up to 3 mm. The pleochroism is of moderate intensity, with little indication of zoning, but according to Dr. Turner's measurements of optic axial angles the composition is about Fs 20–40. The various changes in mineral-composition, and in the nature of the permeating hybridised veinlets will be described below.

Assuming that the xenolith before its removal from the magma reservoir was of the above described nature, the problem of its origin is difficult. Three hypotheses, none of them without objection, suggest themselves. Normative mineral-compositions of (A) a hypersthene ortho-gneiss at Penig in the granulites of the Erzgebirge (Rosenbusch-Osann, 1923, p. 677), (B) quartz-charnockite (“Hypersthene adamellite”) from Ellesmere Land, and (C) quartz-charnockite (“Hypersthene granite”) from Norway, the last two analyses being cited by Washington (1916, p. 325), here tabulated are comparable

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

with the composition of our xenoliths, though there the grain-size is apparently rather smaller than that of the xenoliths, and there is commonly a little modal diopside and biotite in them. In contrast with the earlier conception of the purely igneous origin of such “charnockitic” rocks, Ghosh (1941) summarises the alternative views that have been advanced concerning them as:—(a) complete recrystallisation of plutonic rocks under deep-seated conditions (Stillwell, Groves, Harker), (b) extreme metamorphism of volcanic rocks (Vredenburg), (c) assimilation of aluminous sediments by basic magma and subsequent differentiation (Evans), and (d) deep-seated metamorphism of impure sediments containing silica, lime, iron and magnesia, with gradual influx of alkalies from granitic magma (Ghosh). Though Thomson (1906) recognised fragments of quartz-mica-schists, “granulite” (nature undefined), and garnet gneiss as well as ultrabasic plutonic rocks and xenocrysts derived therefrom, among the inclusions in the basic tuffs coeval with the Moeraki sills and occurring 14 miles north-east thereof, nothing comparable with charnockites or granulites have been found in xenoliths in other igneous rocks in Otago, nor is there other indication of the presence of such deeply metamorphosed material beneath Southern New Zealand.

Alternatively diffusion (“transfusion”) of alkalies with some calcium, aluminium, iron and magnesium from basic and ultrabasic magmas into quartzose xenoliths,^* with the production of alkali-feldspars therein replacing the quartz as described by Holmes (1936), Reynolds (1936), and perhaps suggested by the veinlets of finegrained feldspar in xenoliths already described (see e.g., Fig. 3), or the more coarsely crystallised granophyric material formed about siliceous xenoliths in dolerite or gabbroid sills and laccoliths described by Fenner (1926), Jones (1930), Mountain (1936), Walker and Poldervaart (1942) and others, might be thought to indicate that

under deep-seated conditions it would be possible for coarse grained feldspar to be formed replacing quartz in this fashion, an hypothesis put forward with some hesitation. Such new-formed feldspars are sometimes accompanied by new-formed femic minerals, but those mentioned in the available literature are biotite, hornblende and clinopyroxene, and there is no mention of hypersthene as the sole femic mineral so formed. Moreover, if it were so formed, the conclusion of several of the above-mentioned authors that iron is in general more readily diffused than magnesium, would suggest, that if hypersthene were so formed it is likely to be more ferriferous than that in our xenoliths.

Again it is perhaps noteworthy that each of the hypersthene grains in our rocks lies surrounded in the more or less variolitic consolidation-product of partially hybridised basic magma that has permeated in many veinlets through the xenolith, which material now contains small, thin, prisms of hypersthene. If such material represents consolidation of magma filling crevices in the xenoliths which were open while they lay in the deep-seated magma-basin, it is perhaps conceivable that the coarsely granular hypersthene may have been formed by the direct reaction between the xenolith and the undersaturated adjacent magma. The hypersthene in a veinlet of acidified basaltic material in a quartz xenolith (5919) from the same mass of “pseudoconglomerate” has a composition Fs₂₆, practically that of the hypersthene in these coarse-grained quartz-feldspar-hypersthene xenoliths, and showing the same proportion between En and Fs 63:27 as exists in the average dolerite magma of North-Eastern Otago. (Benson, 1944, p. 106.) Moreover, though in some veinlets of acidified basaltic material in quartzose rocks (e.g., 5919), the hypersthene is associated with subordinate augite, no clinopyroxene could be detected in the acidified veinlets that have been injected into the schistosity planes of the xenolith illustrated in Pl. 21A (5931–5). It seems possible, therefore, to extend the analogy of the effects of acidification of thin veinlets and laminae of basaltic material injected into quartzose xenoliths at high levels in the crust to account for the development of coarse-grained hypersthene (and feldspars?) in xenoliths at depth under plutonic conditions. It is noticeable that the hypersthene grains have been attacked and dissolved by magma presumably under relatively low pressure, and the pyroxene crystallising therefrom in the variolite is dominantly augite though associated with hypersthene. While some combination of the two latter hypotheses seems to be a less improbable method of origin of these quartz-feldspar-hypersthene xenoliths than their derivation from charnockitic masses in a region of intense deep-seated dynamic metamorphism, no final conclusion can be reached at present.

Assuming tentatively that the xenoliths originated through modification in some such manner in the deep-seated reservoir of basic magma, further changes by reaction with the permeating magma may be considered to have occurred at much smaller depth. This magma invaded the intergranular crevices, and its products now show various stages of hybridisation. Between the corroded quartz and feldspar grains the injected material in veinlets widened by

solution of these minerals is usually a pale buff colour, and is largely devitrified so that under polarised light, what in ordinary light seemed glassy, proves to contain a tangled association of skeletal or semi-radiating crystallites of alkaline feldspar.^* Often cores of the corroded feldspar crystals remain, from the margins of which parallel or radiating microlites of feldspar grow out into the glass. These features are indicated by the fine-line microspherulitic pattern in Fig. 4, but traced away from these areas of replaced (chiefly alkaline) feldspar towards the variolitic injections, perfect transition from the leucocratic material just described to the typical variolite darkened with abundant finally-divided iron ore may be observed. Locally, as shown particularly in the inset in Fig. 4, the mass of devitrified feldspathic material contains abundant small graphic patches of quartz suggesting (but not proving) that alteration of micropegmatite has here occurred. An alternative and perhaps more probable explanation is that the micrographic (micropegmatic) intergrowth of quartz and feldspar was not original, but has resulted from recrystallisation. (cf. Harker, 1932, p. 69.) Here and there the large grains of orthoclase show marginally a separation into minute cleavage-bounded platelets fraying out into the devitrified material, and may be compared with Guppy and Hawkes' (1925, Pl. 20, fig. 4) illustration of the “finger-print” structure of a xenocryst of soda-orthoclase in dolerite. The central portions of the orthoclase grains are usually characterised by a rippled “moirée” microperthitic (?) appearance, and irregular distribution of turbid kaolinised areas. A further change is the breaking down of the originally large orthoclase grains into a mosaic of smaller rounded or prismoid grains of varying optical orientation. These features are illustrated in the upper left-hand corner of Fig. 4 and in the inset thereto.

The change of the plagioclase grains is also varied, and is illustrated well by the large grain in the lower left-hand corner of this figure. Much of the central portion of this grain is unaffected. The first indication of change is the development of a “peculiar cloudiness due to the development of a multitude of very minute opaque inclusions” (cf. MacGregor, 1931) which, near the margin of the grain, are seen to increase in size and to be films of basic magma which has entered into and opened out the cleavages parallel to (001) and (010), dividing the grain into platelets as in the andesine xenocrysts in the Te Karipi dyke (Fig. 1A), and, as in the latter case, there has crystallised about the original crystal a narrow band of newformed clear feldspar from which microlitic tabulae extend out into the adjacent variolitic rock. A further change affecting much of the andesine grain, particularly well seen around the jet of basic magma in its centre, and about its margins, is the replacement of the originally continuous material of the grain by a mosaic of irregularly sutured or prismoid grains separated from one another or rendered cribriform by the injection of thin films or jets of basic magma.

Residual portions of such smaller grains, where surrounded by devitrified melt, have generally a fringe of microlites extending out from them either in parallel or radiating bundles into the glassy material.

The hypersthene grains have also been attacked by the injected melt, and show corroded or embayed outlines, and small grains among the devitrification products of this melt remain in optical parallelism with the larger adjacent grains with which they were formerly continuous.

The material of the injected veinlets may be in large part tholeiitic (5919) containing tabulae and laths of labradorite An₆₀, here and there grouped in a skeletal arrangement of hook-like crystallites in optical parallelism with the adjacent tabulae, set in a glassy matrix darkened by dust-like particles of iron ores, and containing granular augite and often arcuately bent thin prisms of strongly pleochroic hypersthene with skeletal branches or extensions studded with small (0 01–0.02 mm.) octahedra of magnetite, and occasionally large (< 0.4 mm.) tabulae of ilmenite with attached small skeletal plates. In the thinner portions and ends of such injections of basic magma, an approach to a variolitic structure is produced. The latter^* is markedly displayed in the veinlets in 5917 (Fig. 4) by the sheaf-like arrangement of the laths of plagioclase, among which are earlier formed unoriented idiomorphic prisms of pale hypersthene (< 0.15 × 0.03 mm.) and rarely of augite with minute fibres of feldspar growing perpendicularly to the surface of the pyroxene grains. Scanty dendritically-branching plates of ilmenite and abundant platelets, dust-like particles and trichites of iron ores occur in the pale brown glass of the matrix, which fades outwards into colourless aggregates of radiating and sheaf-like aggregates of fibres of alkaline feldspar among residual corroded grains of clear, cribriform or clouded grains of feldspar and of quartz, and extends into thin, now partly microspherulitic veinlets traversing intergranular crevices, and in such pale devitrified material there is very little iron ore. The features suggest that the injected veinlets of basic magma have become greatly acidified by solution of the material of the xenolith as they worked their way into the intergranular crevices. It is not always possible to distinguish a boundary between the acidified injected material and the devitrification product of the fused feldspar of the xenolith, though the difficulty of representation of such delicate structures in a line-block has rather emphasised this approach to complete transition in Fig. 4.

Expulsion of carbonate solutions at a late stage from the invading magma already noted earlier in this paper is here expressed by the development of faintly pleochroic siderite and more calcite (and aragonite?) replacing irregularly bounded portions of the variolite, of the thin intergranular veinlets of acidified magma, and of the minerals of the xenolith, whether unaltered or (in the case of feldspar) fused and devitrified.

Fig. 5.—Parallel impersistent laminae of slightly acidified basic material in a xenolith of quartz-schist (5930) in “pseudoconglomeratic” dolerite in the centre of the Tawhiroko dolerite sheet.

Other xenoliths from the “pseudoconglomeratic” masses of dolerite in the centre of the Tawhiroko sheet give so strong a suggestion of the injection of magma into the planes of schistosity and the possibility that the basic material may incorporate albite-muscovite-chlorite-epidote laminae in the schist, that they have been subjected to petrofabric analysis by Dr. F. J. Turner with results that have been incorporated in a separate note appended hereto. The small, sharp, “folded” bands of basic material (5915) illustrated with small and large magnification in Fig. 1 C and D indicate that such appearance of dependence on original folding of now invaded schistosity-planes may be illusory. Though there may have been some slight influence on the course of the injection-laminae by the directions of the intergranular boundaries of the quartz-grains of the xenoliths, these are in general unrelated to the position of the “folded”-laminae. The optic orientation of the quartz-grains is uniform throughout the slide, and is that of vein-quartz rather than of a schist fabric. In 5930 (Fig. 5) the parallelism of the directions of elongation and of the optical orientation of the quartz grains and approximation thereto of the trend of the injected laminae of basic material are more strongly marked, and the same is true of a more coarsely granular rock (5916). There are, however, no features in

the character of the basic material suggesting that it is the devitrification product of fused chloritic laminae in schist, rather than that of a somewhat acidified basic magma. The rock of (5930) was almost wholly quartz-schist, though containing a few grains of feldspar. The basic injections formed short lenticular masses of irregular dimension varying in width up to 1.0 mm. but usually less, extending for 8 mm. or more with irregular off-settings from one parallel line to another, or breaking up into strings of oval or lerticular masses a fraction of a millimetre in length. There is no sign of a corona about their pyroxenes. The broader portions are holocrystalline with a pilotaxitic structure, composed chiefly of laths (< 0.4 × 0.04 mm.) of labradorite and prisms of about the same size of hypersthene and subordinate granular augite, platey ilmenite and much interstitial micropegmatite, in which the sanidine may extend with optical continuity over about 2 mm. The hypersthene may form granules locally moulded on the plagioclase, or thin (0.8 × 0 05 mm.) prisms commonly arranged in radiating groups especially in the pilotaxitic portions. Locally the basic injections may swell about a vesicle and here commonly contain dust-darkened glass, and needles of apatite may project from the basic material into the carbonate-filling of the vesicle. In such glassy material parallel microlites of sanidine may be grouped into rectangular patches or stand perpendicularly to the plagioclase tabular microlites, or sanidine fibres may radiate from the plagioclase microlites, or from the corroded margins of andesine grains which may have been derived from the quartz-schist. In very small ovoid patches of basic intrusive material the microspherulitic habit of the feldspar is rather marked and the small elongated prisms of hypersthene are often arcuate. Sphaerosiderite and calcite fill abundant vesicles which may extend across the whole width of an intrusive band, or like the basic material itself may occur in veinlets running irregularly across the general schistosity-plane. In (5916)—cf. Fig. 4—though the general features are similar to the above, the grainsize of the intrusive material is larger, and there is a more striking development of rectangular patches of micropegmatite, and of the intimate branching of the microlites of sanidine between the plagioclase tabulae. The hypersthene is almost colourless with 2V = 74° (—) indicating the composition is Fs₂₆. In (5915), as noted above, the distinctive feature is the fold-like curvature of the parallel, but irregular layers of intrusive material, the greater abundance of rather titaniferous augite relatively to the hypersthene, the optical parallelism shown by groups of grains of hypersthene, and the frequently marked development of micropegmatite between the plagioclase tabulae.

The last rock to be described is that illustrated in Pl. 21A, fig. A, a block of schist from the richly xenolithic “pseudoconglomerate” from the centre of the tidal platform at Tawhiroko Point. The enclosing dolerite (5896) which is sub-ophitic and holocrystalline, becomes a little finer in grain-size adjacent to the xenoliths from which it is separated by a thin augite corona, through which rare, thin jets of basalt extend across the schistosity of the xenolith, and send thin sheets of magma into the laminae of the schist. The schistosity of the xenolith is very strongly marked, as shown in the

separated laminae, two in section (5931) and one in section (5932). The three corresponding partial diagrams which have been combined into the collective diagram, Fig. 6, are closely similar in pattern, though not identical as regards location of maxima. In fact they show just the slight degree of heterogeneity usually found in the quartz fabric of Otago schists. The pattern of Fig. 6 is that of a typical B tectonite, with a strongly defined minimum at b, and is generally similar to patterns previously recorded for quartz in the schists of Eastern Otago (Turner, 1940). The same pattern differing only in minor detail appears in Fig. 7, which illustrates the orientations of optic axes in 150 grains of quartz measured in the transverse vein in section (5932).

It is concluded, therefore, on petrofabric grounds, that although the basaltic magma penetrated freely between the quartzose laminae of the original schists (incorporating into itself, in addition to quartz, any albite, muscovite, chlorite, etc. that may originally have been present in the xenolith), the essential megascopic and microscopic characters of the B tectonite fabric of the schist have been preserved unmodified. The fabric conforms in pattern to that typical of the quartz-albite-muscovite schists of Eastern Otago. The late stage at which the pattern of the preferred orientation characteristically developed during rock-deformation is once again illustrated by the close correspondence between the quartz-fabric of the laminated schist and that of the transgressive vein which must have originated later than the laminated structure of the host-rock.

2. Coarse-grained Non-schistose Quartz-rocks.

Xenoliths of grey-white quartz-rock without obvious schistosity are plentifully represented among the inclusions in the Moeraki dolerite. Two specimens (5930) and (5915) were sectioned for fabric analysis.

In (5915) the hand-specimen shows what appears to be a folded structure, the microscopic detail of which is indicated on Fig. 1, C and D and, somewhat diagramatically, in the inset circle of Fig. 8, which is 10 mm. in diameter. These figures show that the “fold” pattern is determined solely by sub-parallel veinlets of basaltic material which cut indiscriminately across the larger grains of quartz and seem to be little if at all related to the fabric of the xenolith itself. The latter consists entirely of coarse, sometimes slightly undulose, interlocking grains of quartz 0.3–5 mm. in diameter, with a local tendency to elongation within 15° of a mean direction shown by the broken line LL in Fig. 8. The orientations of the optic axes of 50 grains were measured (in a microsection cut perpendicularly to the “fold” axis) for each of the three sectors of the “fold”—namely, for the two limbs and the intervening arch along the “fold” axis. The three corresponding partial orientation-diagrams were found to be identical and have, therefore, been combined in the single collective diagram, Fig. 8. As would be expected from the identical character of the partial diagrams, Fig. 8 is completely unrelated to the “fold” pattern, and the latter is thereby shown to be a secondary feature accidentally developed during the penetration of the magma into the xenolith, and having therefore no tectonic significance. The

degree of preferred orientation of the quartz-lattice depicted in Fig. 8 is very high, considerably higher than in any Eastern Otago schist which the writer has yet investigated. The pattern is that familiar in B ⊥ B' tectonics, with two girdles (broken arcs in Fig. 8) intersecting at about 80°. Such fabrics are common in granulites and certain highly deformed quartzites in many parts of the world. The maximum at the point of intersection of the girdles is a character commonly recorded in fabrics of this type, in conformity with which the intersection is tentatively identified as the pole of the a fabric axis. The presence of two sub-maxima at points about 90° apart on the bc circle (dotted arc in Fig. 8) is also characteristic of many B ⊥ B' tectonics (cf. for example Sahama, 1936, p. 32, D 26, p. 58, No. 7).

Fig. 8.—Orientation-pattern of optic axes of 150 grains of quartz in a quartz-rock (deformed quartz vein?), xenolith (Section 5915). Contours 12%, 8%, 4%, 2%, 0.7% per 1% area. Maximum concentration, 18%. The line LL is the mean direction of elongation of the quartz grains. The “fold” pattern of the injected basaltic material is sketched in inset circle 10 mm. in diameter and more precisely in Fig. 1, C and D.
Fig. 9.—Orientation-pattern of optic axes of 50 grains of quartz in quartz-rock (deformed quartz vein?) xenolith (Section 5930). Contours 20%, 12%, 8%, 2% per 1% area. Maximum concentration, 35%. The line LL is the mean direction of elongation of the quartz grains. The microstructure of this xenolith is shown in Fig. 5.

The second xenolith was selected as typical of a group in which the microscopic streaks of basaltic magma have been injected along regular sub-parallel plane surfaces, which as in (5915) cut irregularly across the larger grains of quartz, which are up to 4 mm. long, and in many cases are somewhat elongated more or less parallel to the trend of the invading streaks of basalt (LL in Fig. 9). Fig. 5 illustrates the microscopic structure of this xenolith, Fig. 9 the orientations of the optic axes in 50 grains of quartz therein. The small number of measurement results from the wide spacing of the measuring traverses necessary to avoid including more than one subindividual in any one of what may be superindividual grains of quartz. The pattern of Fig. 9 is almost identical with that of Fig. 8, from which it differs in the incomplete development of girdles and the very high concentration of quartz axes at the principal maximum.

There can be little doubt that the two xenoliths (5930 and 5915) have been derived from the same parent rock—a rock in which a

relationship to the fluorapatite of Waihola and older formula used by Samojloff (1918)—viz., 2Ca₅F (PO₄)₃ CaCO₃ etc. may, however, be cited here.

Fig. 10.—A. Refractive indices in the Fluorapatite-Chlorapatite system, and of the apatite in a theralite at Waihola. B. Variation of birefringence with chemical composition in non-alkaline clinopyroxenes deduced from Wager and Deer's curves of maximum and minimum refractive indices.

B. The Birefringence of Non-alkaline Clinopyroxenes.

Study of the curves showing the variation of the refractive indices and with variation in the chemical compositions of clinopyroxenes expressible in terms of Wo, En and Fs as given by Deer and Wager and combined in Fig. 2B of Part IVB of this series of papers (Benson, 1944, p. 77) makes possible estimation of the double refraction (γ — α) for all possible compositions of such clinopyroxenes and to express it graphically as in Fig. 10B herewith. Comparison of this with Figs. 5 and 9 (pp. 83 and 117) of Part IVB shows that the common clinopyroxenes along the main path of the differentiation trend for such minerals in dolerites are almost exclusively those with high birefringence (γ — α = .022–.028) as the text-books of optical mineralogy indicate (e.g., Winchell, 1933, Figs. 140 and 14, pp. 221 and 223). Where, however, a petrologist has not available the refractive indices of a pyroxene the optical axial angle of which he knows, he could utilise the method of estimating its chemical compositions from the values of 2V and the extinction angle, if these optical measurements were thoroughly characteristic of the pyroxene-composition, which is seldom the case. When using this method Fig. 10B herewith might aid in selecting the more probable composition out of the two indicated by the double intersection of the characteristic curves, by reference to the double refraction of the pyroxene, which is readily measured by comparison with that of an adjacent grain of plagioclase of known composition, preferably made by a Berek compensator. Thus, e.g., a uniaxial pigeonite with extinction angle of 40° might be either approximately Wo₁₅En₇₂Fs₁₃, in which case its double refraction would be about .018, or Wo₉En₃₇Fs₅₄, in which the double refraction would be .027, which would be easily distinguished from the lower value. Unfortunately, in addition to the small variation in optic angle with variation in the content of