The Metamorphic and Plutonic Rocks of Lake Manapouri, Fiordland, New Zealand—Part III.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 68, 1938-39, Page 122

The Metamorphic and Plutonic Rocks of Lake Manapouri, Fiordland, New Zealand—Part III. By F. J. Turner, University of Otago. [Read before the Otago Branch, April, 1937; received by the Editor, February 5, 1938; issued separately, June, 1938.] Contents. Introductory Summary. The Trondhjemitic Granites and Associated Gneisses. The Pomona Island Granite and Associated Gneisses, etc. Structure and Tectonics. Fabric Analysis of Foliated Pomona Granite. Acknowledgements. Literature Cited. Introductory Summary. In the first two papers of this series (Turner, 1937, 1937a) an account was given of the metamorphic and intrusive rocks exposed in the northern, western and central portions of Lake Manapouri. Subsequently, with the assistance of a grant from the research funds of the Australian and New Zealand Association for the Advancement of Science, the writer has been able to complete the survey of the pre-Tertiary rocks of Manapouri by field work along the shores of South and Hope Arms and the intervening southern coast of the lake (Fig. 1). This paper embodies the results obtained during this later work, together with a summary of the structural features of the area as a whole. That part of the lake here mapped lies mainly within the Eastern Manapouri Province as defined in Part I, and the rocks there exposed fall within the same stratigraphic units as described earlier, viz. (in order of decreasing age): (a) basal gneisses (Holmwood Island Gneisses), correlated with Professor Park's Dusky Sound Series; (b) Beehive epidiorite; (c) Pomona Island granite and hybrid derivatives; (d) Tertiary sandstones and conglomerates. However the white trondhjemitic granites of the Western Province, accompanied by minor amounts of the invaded basement gneiss, are exposed continuously along the west shore of South Arm, and there form the eastern margin of the great injection-complex described in Part 2. Nowhere has the trondhjemitic granite been observed in contact with either the epidiorite or the Pomona Island granite, so that the relative age of the three groups of rocks remains rather uncertain (cf. Turner, 1937a, pp. 244, 245). The Trondhjemitic Granites and Associated Gneisses. Along the western shores of South Arm granitic rocks greatly preponderate over the invaded gneisses, the occurrence of which is limited to large blocks often several yards in diameter enclosed locally in the intrusive member of the complex.

The latter is a white, rather fine-grained often gneissic rock usually conforming to the composition of “oligoclase-granite” as previously described (Turner, 1937a, p. 232). A typical specimen (No. 4505) consists of acid oligoclase 50% to 60%, interstitial sometimes coarsely crystalline microcline 10% to 15%, quartz 20%, brown biotite 8%, and accessory muscovite, apatite and epidote. Myrmekite is present at microcline-plagioclase junctions while the quartz and some grains of feldspar show undulose extinction. Occasionally (e.g. No. 4508) microcline may make up as much as 50% of the composition and the rock may then be classed as a microcline-granite. Typical trondhjemites are absent in this sector. Fig. 1.—Geological Map of the Southern Portion of Lake Manapouri. The invaded gneisses are oligoclase-quartz-hornblende-biotite-gneisses, generally similar to those described in Part 2 (p. 228) but poorer in the dark constituents and usually containing a considerable quantity of quartz. A typical composition (No. 4507) is plagioclase 60%, quartz. 15%, hornblende 15%, biotite 10%, epidote 1%, accessory sphene, apatite and zircon. Occasionally there are lighter-coloured bands in which biotite predominates over hornblende (No. 4510), and in one case (No. 4509) quartz is almost absent. The plagioclase is usually medium oligoclase, ranging from Ab82 to Ab72;

it occurs in equant, for the most part untwinned grains, which often show incipient alteration to sericite. In several sections (e.g. Nos. 2506, 2509) crystals of plagioclase may enclose scattered minute segregations of microcline. The biotite is a yellowish brown variety, usually partially chloritised, and in some cases is associated with aggregates of flaky prehnite (Nos. 4506, 4510). The hornblende invariably shows the intense pleochroism and deep blue-green absorption for vibrations parallel to Z, that characterise the amphibole of corresponding gneisses throughout the western injection-complex. Epidote is never as plentiful as in the rocks described in Part 2, but is the same colourless strongly birefringent type (γ–α = 0.035–0.04), and often encloses vermicular quartz; in some sections (e.g. No. 4510) it is associated with strongly pleochroic allanite (X = very pale yellow, Z = deep brown; γ–α = 0.015). The sphene and apatite are coarse and always abundant. Zircon is rare (No. 4507). As in other parts of the western injection-complex signs of assimilative reaction between magma and gneiss are few. Possibly to be attributed to this process are such unimportant details as the presence of minute quantities of potash-feldspar, slight sericitisation of plagioclase and development of allanite and prehnite in the invaded rocks. In no case is there any indication of replacement of hornblende by biotite. The Pomona Island Granite and Associated Gneisses, etc.. (a) The Invaded Gneisses. Eastward from the head of South Arm to the western border of the Tertiary covering strata, the ancient gneisses* These belong to the same group as the gneisses invaded by trondhjemitic granite further west. previously correlated with the Dusky Sound Series are the dominant rocks, though broken at intervals by tongues of Pomona Island granite. Interruption of this sort is least within two miles of the head of the South Arm, and again along the western shore of Hope Arm south of Stockyard Cove. Two broad lithological divisions may conveniently be distinguished, viz., coarse-grained gneisses and finer dark-coloured amphibolites, though there is considerable variation in each group. Certain features are common to the rocks of both groups, especially the universal association of plagioclase (acid oligoclase to medium andesine) with green hornblende, biotite or both minerals, and the constant abundance of sphene and apatite as accessory constituents. Rough foliation may usually be observed, and is parallel to the bedding when the latter is recognisable. The most widely developed members are coarse hornblendeplagioclase-biotite-gneisses, often containing between 10% and 20% of quartz and small amounts of colourless highly birefringent epidote. Many of these rocks are closely similar to gneisses of the Western Province and from the area on the north coast of the lake opposite Pomona Island (cf. Part 1, pp. 95, 96; Part 2, pp. 227, 228). Typical examples are described briefly below:—

No. 4512 (East coast of South Arm, 1 ml. from head; widely distributed). A coarsely gneissie rock with average grain-size of 3 mm. The composition (estimated by inspection) is blue-green hornblende 20%, biotite 20%, oligoclase 50%, quartz 5%, epidote 5%, abundant sphene and apatite. The biotite is partly chloritised, and also shows replacement by aggregates of prehnite. Nos. 4546 and 4547 (South side of first small bay inside western headland at entrance to Hope Arm) represent light and dark bands of the same hand-specimen. Fine-grained dark material (No. 4547), comparable in composition and structure with the amphibolites, predominates. This consists of hornblende 35%, biotite 15%, medium andesine 40–45%, epidote 5%, sphene 1%, iron-ore 1%, and accessory apatite. The average grain-size is 0.5 mm., but there are occasional large feldspars 4 mm. across; much of the feldspar shows albite twinning, an unusual feature in rocks of this type. While much of the hornblende is the usual blue-green type, there is also a variety having deep greenish-brown for the Z vibration-direction. Light-coloured bands in the same rock (No. 4546) contain less hornblende and only minor biotite. The amphibole is in coarse ragged crystals with marked sieve-structure, the central portions of which may be bleached to pale-green or almost colourless. Nos. 4550 and 4551 (Western shores of Hope Arm) are plagioclase-hornblende-biotite-gneisses of finer grain, resembling the predominating rocks of Holmwood and adjacent islands (cf. Part 1, pp. 84, 85). Sphene is plentiful and pyrite usually is present, while in No. 4451 there is about 15% of quartz. On the southern side of Stockyard Cove they give place to very coarse gneisses of dioritie aspect (No. 4548) consisting of oligoclase-andesine, coarse biotite, sieved green hornblende and accessory apatite and sphene. Associated and interbedded with the hornblende-bearing rocks are gneisses containing little or no amphibole, typical examples of which are described below:— No. 4543 (South shore half-way between South and Hope Arms) is a rather fine-grained rock consisting of untwinned basic oligoclase and biotite in subequal proportions, accompanied by minor colourless epidote, relatively abundant granular sphene and accessory apatite. The biotite is a greenish-brown variety. The rock occurs as a constituent of large masses completely surrounded by Pomona Island granite. Interbedded with it are a hornblende-plagioclase-gneiss with little biotite (No. 4542) and a coarse gneissie rock of most unusual composition consisting almost entirely of biotite and epidote (No. 4541). In the latter rock, the epidote is the usual colourless highly birefringent type and makes up about 60% of the total composition. The biotite is deep greenish brown (Z) to pale yellow (X). Accessories include sphene, apatite, pyrite and occasional aggregates of sericite enclosed in the epidote. No. 4553 (West coast of Hope Arm, 1¼ mls. from head). An unusually light-coloured phase, interbedded with plagioclase-biotite-gneiss (No. 4552) and plagioclase-quartz-hornblende-biotite-gneiss (No. 4551). The composition is oligoclase 50%–60%, quartz 30%–40%, biotite 5%, coarse granular colourless epidote 5%, accompanied

by sphene, apatite and much pyrite as accessories. The biotite is intensely pleochroic from pale yellow to deep red-brown, in contrast with the greenish-brown tint of the mica in associated rocks. It is partially replaced by colourless chlorite. An unusual feature shared also by the associated plagioclase-biotite-gneiss is the presence of small amounts of interstitial highly irregular orthoclase. Along the eastern shore of South Arm, from the entrance to a point about 3 mls. from the head, the predominating members of the gneissic series are fine-grained, dark-green amphibolites of almost hornfelsic aspect (Nos. 4513, 4514, 4516, 4520, 4530, 4534). Small masses of contaminated epidiorite are associated with these rocks near the entrance to the Arm, and throughout their whole extent interruption by extensive outerops of granite is frequent. Full petrographic descriptions are unnecessary since there is a close mineralogical resemblance to the amphibolites from Holmwood and other islands as already described (Part 1, p. 85). The main constituents are plagioclase (usually andesine) and deep blue-green hornblende in about equal quantity, accompanied by accessory epidote, sphene, iron-ore and apatite. Reddish-brown biotite may be present in small quantities (e.g. Nos. 4520, 4534), and in one case (No. 4534) the percentage of epidote is as high as 10%. Patches of well crystallised prehnite were noted in a single section (No. 4513). Slender prisms of yellow rutile are very plentiful in No. 4519. Pyrite was observed in most sections. In contrast with the amphibolites described in Part 1, the rocks from the South Arm usually show perfect parallelism of the hornblende crystals, which typically are prismatic and idioblastic. In this respect they may closely resemble certain of the hornblende-schists that occur on Pomona Island as major inclusions surrounded by granite (Part 1, pp. 93, 94). (b) Basic Intrusive Rocks. From a point about ½ ml. inside South Arm to the headland at the entrance of Hope Arm, basic intrusives predominate among the rocks invaded by the granites, and are associated with various members of the basement gneiss series. Some are indistinguishable from rocks outeropping on Pomona Island and the northern shore of the lake, and shown in Part 1 (pp. 90–92) to be contaminated phases of the Bechive epidiorite; others, while still retaining traces of igneous structure, have no counterpart among rocks previously described from Manapouri. There are also one or two specimens of doubtful origin, which should perhaps be included with the hornblendic members of the ancient gneissic series (e.g. No. 4539). Many of the amphibolites already described are probably derivatives of basic lavas or tuffs (cf. Part 1, p. 86), and it is even possible that some of the rocks classed above as amphibolites may in reality be completely reconstituted members of the epidiorite group. The difficulty in making a sharp distinction arises from the fact that regional metamorphism of the ancient lavas and tuffs and partial granitisation of the epidiorites both lead to the development of the mineral assemblage oligoclase-hornblende-biotite-sphene-epidote. In the amphibolites the original structure has been obliterated and a crystalloblastic structure substituted. In the contaminated

intrusives, however, the original structure tends to be preserved in part, and especially the coarse tabular crystals of plagioclase usually retain their habit even though changed in composition. In all except a few doubtful cases therefore, amphibolites and contaminated epidiorites may be distinguished on grounds of structure. Typical contaminated epidiorites are represented by Nos. 4532 (5 ch. E. of entrance to South Arm), 4535 (25 ch. E. of entrance to South Arm) and 4544 (headland at western side of entrance to Hope Arm). Coarse tabular plagioclase (medium to basic oligoclase) sometimes enclosing clusters of epidote prisms makes up 50% to 60% of the composition. Deep green hornblende is plentiful, but in more altered rocks (e.g. No. 4532) tends to be replaced by aggregates of flaky biotite, while the remaining crystals develop pronounced sieve-structure. The percentage of biotite varies from 10% to 30%. Sphene and apatite are exceedingly plentiful. The former characteristically occurs as granular rims surrounding grains of iron-ore, or else as rounded aggregates with roughly radial structure resulting from complete replacement of that mineral. Epidote is a minor constituent. In No. 4544 nests of granular quartz make up 10% of the composition and there are small amounts of interstitial orthoclase. Locally the structure may be considerably modified by shearing (No. 4533). The above rocks are all invaded or surrounded by the Pomona Island granite or its contaminated derivatives. A somewhat different type of basic intrusive rock, probably belonging to the epidiorite series, occurs in association with amphibolite and granite at a point about half a mile inside South Arm (Nos. 4528, 4529). No. 4528 is a holocrystalline rock composed principally of long prisms of hornblende (40%) and equidimensional twinned grains of plagioclase (50% to 60%). Minor constituents are pyrite, coarsely crystalline chlorite (optically +, biaxial) and apatite, together making up about 5% of the composition. The plagioclase is a medium andesine and locally is altered to kaolin or sericite. The hornblende occurs in poorly terminated subidiomorphic prisms (1 mm. × 0.5 mm.) with strong pleochroism as follows:—             X = pale yellow             Y = deep yellowish-green             Z = deep blue-green             Z > Y > X. The extinction angle Z to c = 17°. The central portions of the grains are crowded with schiller inclusions. No. 4529 is obviously related to the rock just described, but is finer in grain and contains about 5% of pale biotite. Local recrystallisation has considerably reduced the grain-size. Both rocks may be classed as diorites. From their field association with an amphibolite (No. 4530) of similar mineralogical composition but with no trace of igneous structure it is possible that all three rocks have a common origin. Alternative grouping with the epidiorites is preferred by the writer however.

(c) Pomona Island Granite. Extensive outcrops of granite of the Pomona Island type occur at frequent intervals along the eastern side of South Arm from a point about three miles from the head to the entrance, and thence eastward along the southern shore of the lake to Hope Arm. The individual outcrops are often five or ten chains in width, and alternate with equally extensive exposures of gneiss and epidiorite. Toward the head of South Arm the granites are represented only by sparsely distributed pegmatitic and aplitic dykes cutting the gneiss. Granites are well developed in Hope Arm only around Stockyard Cove and again on the eastern shores on a small peninsula half a mile north of the Monument. Also in one or two places on the western shore of Hope Arm massive pegmatitic dykes invade the prevailing gneiss. In composition representative specimens (Nos. 4523–4525, 4527, 4538, 4540) are similar to the Pomona Island granites described in Part 1 (p. 89): microcline, albite-oligoclase and quartz are about equally plentiful, while the dark constituents together total less than 5% of the composition. In structure there are certain differences however. Typically there is definite foliation due partly to segregation of quartz into narrow lenticles and partly to parallel development of streaky aggregates of dark minerals; this will be discussed more fully in the section dealing with fabric analysis. The quartz is coarse-grained and undulose. The habit of the feldspars varies. In most specimens from South Arm (e.g. Nos. 4523–4525) small grains of microcline and plagioclase (0.2 mm. in diameter) enclose a few larger crystals of both minerals, those of plagioclase being charged with epidote or sericite. In No. 4540 (halfway between South and Hope Arms) coarse composite crystals composed of microcline and oligoclase in subequal proportions are accompanied by a few large grains of oligoclase alone. In other instances (e.g. No. 4538) clear coarse grains of perthitic microcline and cloudy crystals of oligoclase often riddled with small flakes of sericite occur independently. The range of structures suggests that in some cases microcline and plagioclase have crystallised independently while in others microcline has replaced crystals of plagioclase at a late stage. The dark streaks observed in the hand-specimen in most cases are aggregates of black opaque iron-ore, very strongly coloured chlorite, sphene and epidote. The chlorite is uniaxial, negative and intensely pleochroic; X = yellow, Y = Z = very deep green. Apparently it is a product of hydrothermal alteration of biotite, for residual flakes and laminae of that mineral were noted in association with chlorite in one section (No. 4538). The sphene often mantles the grains of iron-ore but also occurs as independent rounded crystals. Allanite with intense pleochroism from pale yellowish-brown to deep red-brown is often associated with the epidote. In some rocks aggregates of muscovite and chlorite or muscovite, iron-ore and minor chlorite also occur (Nos. 4527, 4540). Comparison with the contaminated granites of Pomona Island (Part 1, p. 99) suggests that these dark composite streaks are largely of xenolithic origin, derived from the disintegration of granitised epidiorite and amphibolite.

The aplites and pegmatites associated with the Pomona Island granite are normal rocks worthy of only brief comment. The former (Nos. 4521, 4522) consist largely of oligoclase and quartz, with small quantities of biotite and accessory chlorite, muscovite, epidote and apatite. The pegmatites are coarse-grained rocks in which perthitic microcline, medium oligoclase and quartz are the main constituents; accessories include biotite, muscovite, apatite and sphene. Crystals of the two varieties of feldspar seem to be quite independent. (d) Contaminated Derivatives of Granite Magma. Intrusive rocks believed to be contaminated derivatives of the granite magma were collected from three localities. These occurrences are described separately below so that the evidence bearing on petrogenesis may be recorded fully. The most southerly mass of Pomona granite exposed in South Arm forms an extensive outcrop about 15 to 20 chains in length flanked on either side by amphibolite, about 3 mls. from the head of the arm. The typical granite collected from the middle of this outcrop (No. 4515) is much poorer in potash-feldspar than the normal Pomona granites. The main constituents are coarse albiteoligoclase (enclosing much epidote and sericite) 60%–70%, orthoclase with local microcline structure 10%; quartz 15%, chloritised biotite 5%, and epidote 5%. Allanite, apatite and pyrite are accessories. An unusual feature is the presence of pinkish garnet in highly irregular granular masses between the boundaries of the large feldspars. Border phases collected at the northern contact are coarse blotched rocks of dioritic appearance (Nos. 4517, 4518). No. 4517 differs from the granite just described in complete absence of potashfeldspar, greater abundance of chloritised biotite (10%) and the presence of plentiful green hornblende (20%). The biotite is almost completely replaced by chlorite and prehnite, and grains of ironore enclosed in the hornblende crystals are rimmed with granular sphene. In composition the rock is a tonalite. No. 4518 appears to be a typical diorite composed essentially of medium andesine (90%) and coarse blue-green hornblende (10%). Accessories include black iron-ore, biotite, epidote and apatite. About five chains east of the entrance to South Arm a coarsely blotched rock (No. 4531) of rather similar composition occurs as fairly extensive masses invading contaminated epidiorite and enclosing major inclusions of amphibolite. Plagioclase approximating to oligoclase-andesine makes up 75% of the rock. The remainder consists of aggregates of coarse blue-green hornblende in all stages of replacement by intensely pleochroic yellowish-brown biotite. These aggregates enclose coarse prisms of apatite and large grains of iron-ore rimmed with granular sphene. In the duplicate section from the same hand-specimen (No. 4531a) replacement of hornblende by biotite is complete and the biotite itself has locally been altered to chlorite. Rather similar rocks occur again halfway between the entrances to Hope and South Arms, as masses 50 yds. in width alternating with contaminated epidiorite. Of these No. 4537 resembles No. 4531a in consisting almost entirely of medium oligoclase (75%) and deep

yellowish-brown biotite (25%). The latter occurs in aggregated masses enclosing granular sphene, iron-ore and apatite. Small amounts of quartz occur in nests between the crystals of feldspar. No. 4536 differs from the rock just described in the presence of residual hornblende enclosed in the biotite aggregates, and in the development of microcline and orthoclase interstitially and as patches replacing the large grains of plagioclase. The latter enclose a good deal of epidote and sericite. Quartz is absent. In considering the genesis of the rocks just described certain additional facts must be borne in mind: (a) in all cases the rocks in question definitely bear an intrusive relation to the amphibolites and epidiorites; (b) exposures of typical Pomona Island granite occur at intervals between the three localities enumerated above; (c) the coarse hornblende of the “dioritic” rocks seems in all cases to be a primary product of magmatic crystallisation and could not have been derived from the adjacent amphibolites and epidiorites by marginal disintegration of these rocks; (d) the proportion of dark constituents to plagioclase is much too low to allow the possibility of derivation by contamination of rocks of the epidiorite group. It is therefore suggested that the “dioritic” rocks such as Nos. 4518 and 4531 are products of crystallisation of a locally contaminated magma derived from the parent Pomona granite magma as a result of reaction between the latter and the invaded rocks (cf. Part 1, p. 99). Replacement of hornblende by biotite, introduction of quartz, and in one case late crystallisation of microcline at the expense of earlier-formed plagioclase are attributed to the action of residual potassic silica-rich liquids emanating from the adjacent relatively uncontaminated granitic magma. The microcline-poor granite (No. 4515) associated with dioritic hybrids in South Arm is the product of crystallisation of a granite magma in the earlier stages of contamination when potash was being removed progressively to allow conversion of hornblende to biotite in adjacent rocks (cf. Part 1, p. 99). Elsewhere in the contact zone of the granite intrusion marginal reaction has affected the bulk composition of the magma to a less extent, and microcline-rich granites containing plentiful altered xenolithic clots of biotite, sphene and iron-ore make sharp contacts with the invaded rocks. An intermediate condition has already been described on the southern end of Pomona Island, where normal granite and altered epidiorite are separated by a zone of granite poorer in microcline and richer in xenolithic biotite than is usually the case (Part 1, p. 94). Structure and Tectonics. As regards structure, the area here considered has much in common with that covered in the first two papers of this series. A summary of the structural data for the Manapouri region as a whole may therefore be presented at this point. The criteria recorded are defined as follows (for a detailed discussion of their significance see Balk, 1937; Fairbairn, 1937):— Foliation: A streaky, lamellar or banded structure resulting from

aggregation of particular minerals into lenticles and bands. Cleavage (schistosity) is parallel to the foliation when present in Manapouri rocks. Lineation: Parallel arrangement of prismatic crystals or elongated microxenoliths with long axes in a common direction within the plane of foliation. Cross-joints (Q-joints of Cloos, ac-joints of Sander): Usually vertical or steeply dipping joints perpendicular to the lineation, foliation and fold axes, i.e. to the tectonic axis. Longitudinal joints: Approximately vertical joints striking parallel to the foliation in some plutonic rocks. In the older gneisses (Dusky Sound Series) exposed in Hope Arm, the eastern coast of South Arm and at intervening points on the southern shore of the lake, the foliation dips steeply and strikes between 30° W. of N. and 20° E. of N. Approximately the same range of strike is recorded for rocks of this series throughout the whole Manapouri region, though there is a definite tendency for a N.N.W. trend to predominate.* Occasionally, for example in the vicinity of the shatter zone opposite the N.W. corner of Pomona Island, the strike locally may swing as far east at 45° E. of N. Corresponding vertical or subvertical joints are sometimes well developed, e.g. along the western shore of Hope Arm. Two sets may be recognised, viz. cross joints striking between 80° and 100° E. of N., and less frequently longitudinal joints parallel to the foliation. All three may be correlated with movements across a tectonic axis the trend of which varies between 30° W. of N. and 30° E. of N. but frequently approximates to a mean N.N.W. direction. The trondhjemites and related granites of the western injection-complex show considerable variation in structure. The most noticeable and consistently developed structural elements are a steeply dipping foliation and accompanying longitudinal jointing with a strike of 5° to 25° W. of N., together with a Fig. 2.—Range of foliation and cross joints in gneisses (A), trondhjemites and allied granitic rocks (B), and Pomona Island granite (C). Longitudinal joints often accompanying foliation.

ranging between 75° and 100° E. of N. Along the southern shore of the lake between West and South Arms, and in the latter inlet itself, foliation and longitudinal joints trending 50° to 55° W. of N. prevail, though even here the N.N.W. strike may locally assert itself. On the other hand a N.E. foliation has been observed at several points in West and North Arms. Thus for the western injection-complex as a whole the prevalent trend of the tectonic axis is N.N.W., the local deviation to N.W. or N.E. in particular areas may have a tectonic significance or on the other hand may be the result of local variation in direction of magmatic flow, a condition to be expected in the upper parts of the bathylith where roof-pendants are large and numerous. The Beehive epidiorite is typically a non-foliation well-jointed rock, but the directions of jointing show little regularity when compared over wide areas. In the Pomona Island granite, foliation is inconspicuous or absent except in the vicinity of contacts. Here, especially in the contact-zone of South Arm, where granite and invaded gneiss alternate rapidly for several miles along the eastern shore, a vertical foliation marked by parallel arrangement of small dark xenolithic clots and aggregations of quartz into narrow streaks is often conspicuous. In all such cases it lies parallel to the regional trend of the foliation in the adjacent gneiss. Longitudinal joints and narrow xenolithic bands of dark schist often many yards in length trend in the same direction as the foliation, and may also be observed in non-foliated massive granite far from visible contacts and in all parts of the intrusion. All these structures have therefore the same tectonic significance, and whether occurring together or alone define the plane of flow of the congealing granite magma and the tectonic axis of contemporaneous earth movements. This direction invariably lies between 15° W. of N. and 20° E. of N. Correlated cross joints with a trend of 75° to 100° E. of N. constantly occur throughout the whole mass (Fig. 2). In addition to the cross and longitudinal joints just deseribed there may locally be a third set of subvertical joints the strike of which varies between 105° and 120° E. of N. For example, all three sets are well developed in the massive granite that outcrops continuously along the northern shore of Pomona Island. Possibly these may indicate minor movements across a tectonic axis 15° to 30° E. of N.; on the other hand they may have no obvious tectonic significance. Special structural peculiarities characterise the Pomona granite in the contact zone of South Arm. The southern margin of the intrusion lies in the roof region where pendants of invaded rock occupy as much space as the intervening intrusive granite and its hybrid derivatives. This contrasts with the conditions observed on the northern half of Pomona Island and the opposite northern shores of the lake, where major inclusions are much more sparsely distributed and rarely exceed 20 yds. in length by three feet in maximum thickness, and the granite itself is non-foliated. In the marginal zone of South Arm a vertical foliation with a strike of 5° to 10° E. of N. is well defined in most outcrops of granite; junctions between gneiss and granite also dip steeply and frequently show the same

trend as the foliation. Here also two sets of mutually perpendicular joints dipping north and south respectively at angles of 40° to 50° are perfectly developed and share a common strike perpendicular to the tectonic axis, i.e. 95° to 100° E. of N. A marked lineation in the plane of foliation here dips southward more or less parallel to one set of joints, so that only the north-dipping set may be interpreted as cross joints (cf. Fig. 3). According to Cloos, Balk and others (e.g. see Balk, 1925, 1937) foliation and lineation in granitic rocks are primary structures developed by flow of the partially crystallised magma during the later stages of intrusion, and their directions are therefore perpendicular to the direction of contemporaneous compression. Lineation in relation to magmatic flow has been interpreted variously: near the margin of the intrusion where differential movement is strong lineation, usually dipping steeply subparallel to the contact, represents the direction of magmatic flow; but in the more central portions of the intrusive mass, where the velocity of the rising magma was subject to little lateral variation in velocity, differential movement was small and lineation therefore has developed horizontally, i.e. across the direction of flow (Fairbairn, 1937, p. 111). In the granites of the South Arm of Manapouri, lineation may safely be considered to represent direction of flow, for the dip is steep (45°) and away from the centre of the intrusion and the rocks-themselves lie in the contact zone. The northern margin of the Pomona intrusive lies in the heavily forested region beyond the northern shore of the lake and is therefore inaccessible for structural investigation. In the first two papers of this series (Turner, 1937, p. 88; 1937a, p. 244) the Pomona Island granite was considered, from its generally massive character as displayed throughout the region then mapped, to be a late intrusion the emplacement of which occurred after compressional movements had almost ceased. Certainly the evidence of such movement is much less striking, and developed to a more limited extent, than in the older trondhjmite-granite intrusion of the western Manapouri Province. It is now evident however from the present study that upward flow of the Pomona granite magma was affected by compressional movements acting across a N.N.W. to N. tectonic axis, an axis shared also by the invaded gneisses and the trondhjemitic rocks alike during their deformation. Since only a part of the Pomona intrusion is exposed around the lake shores it is impossible to reconstruct its detailed structure. But the nearly vertical disposition of the cross joints throughout the main mass, in contrast with their inclined orientation and the complementary southward dip of the lineation near the southern margin, suggests an arched arrangement of flow-lines such as has commonly been observed in bathylithic masses forced upward under pressure during the closing stages of orogeny (cf. Balk, 1937, pp. 69–86). This reconstruction of the conditions accompanying the uprise of the Pomona granite magma accords with the generalisations recently summarised by Bucher (1933, pp. 285–290), e.g.: “Observation teaches us that as far back as the record is clear, granite magmas had accumulated in sufficient size to rise effectively in the crust only by the time that orogenic stress was diminishing” (p. 289). “On

the other hand Cloos' detailed studies have proved that the regional stress had not completely ceased while most granitic intrusive bodies studied by him rose to their final lodging place. He could show that generally the direction of ‘stretching,’ that is, the direction of drawing out and thinning out of the solidifying magma recorded by the parallel orientation of linear elements among the crystals and inclusions, is quite independent of irregularities of shape of the intrusive body. This remarkable uniformity in the inner texture must be due to the regional stress which ultimately caused and controlled the discordant uprise of the magma. In most cases this regional stress has the same direction as that which produced the last folding” (p. 286). Petrofabrie analysis shows that the same compressional forces continued to operate even after the Pomona granite magma had completely solidified. Fabric Analysis of Foliated Pomona Island Granite. For investigation of the microfabric of the Pomona Island granites a well foliated specimen from the southern contact zone was selected.* The megascopic data (cf. Fig. 3a) are as follows:— Foliation (ab) vertical; strike 5° E. of N. Rift (direction of easiest cleavage) parallel to foliation. Lineation (b) dips southward at 45°. Cross joints (ac) strike 95° E. of N. and dip northward at 45°. Joints subparallel to bc well developed; strike 95° E. of N. Polished surfaces of the hand-specimen show that the foliation is due principally to aggregation of quartz into discontinuous foliae about 0–5 mm. to 1 mm. in thickness, which are more clearly defined on the bc than on the ac surface. Flattened aggregates of flaky chlorite and granular iron-ore of xenolithic origin, not exceeding 1 cm. in length, lie with their greatest dimensions parallel to the foliation. As seen on ah and bc surfaces they are arranged along “flow lines” parallel to b thus clearly defining the latter direction. Fig. 3.—A. Diagram showing structural features of granite (represented by specimen No. 4525) halfway along east coast of South Arm; invaded amphibolite and gneiss indicated by dotted areas (the actual exposures are often 100 yds. in width). Joints in granite are indicated by full lines; broken lines represent lineation. Foliation vertical, parallel to plane of the page. B. Section block showing relative positions of a, b and c axes in relation to lineation and foliation. Directions in A and B correspond. In section the principal constituents are sodic oligoclase, quartz, microcline and chlorite. Both feldspars build up a mosaic of small

grains averaging 0.2 mm. in diameter, in which larger tabular crystals (1 mm. to 1–5 mm.) are enclosed; the latter tend to lie parallel to the foliation. Quartz occurs mainly as relatively coarse crystals (0–5 mm. to 2 mm.) concentrated in streaks and lenticles trending parallel to ab, but is also represented by small grains in the feldspar mosaic. Undulose extinction is almost universal and is developed on a perfect scale. Closely spaced cracks and lines of dusty inclusions oriented parallel to ac are clearly observable in ab and be sections. Fig. 4.—Orientation curve for quartz in ab section of specimen No. 4525-(375 grains).

Fig. 5.—Orientation curve for quartz in be section of specimen No. 4525 (500 grains) Fig. 6.—Orientation curves for quartz in ac section of specimen No. 4525; broken curve represents 450 grains in a single rock section; full curve represents 900 grains in two different sections, including that upon which the broken curve is based.

Sections perpendicular to each of the three fabric axes were cut (Fig. 3B) and the quartz fabric was analysed following the method described by the writer in previous papers (Turner, 1936, 1938). This involves measurement of the angle Z° to b (or a) for a large number of quartz grains in each section, using an ordinary petrographic microscope fitted with a simple mechanical stage. The results are graphically represented in Figs. 4 to 6. The main features of the fabric are clearly indicated and are summarised below:— (1) A pronounced minimum parallel to b (Figs. 4 and 5). (2) Strong concentration of the quartz axes (Z) in the ac plane (Figs. 4 and 5). (3) A second minimum parallel to c (Fig. 6). This is borne out by the prominence and sharpness of the maximum in the ab curve (Fig. 4) compared with the less pronounced and broader zone of concentration perpendicular to b in the bc curve (Fig. 5). Prevalence of grains showing low double refraction in the be section also indicates that the majority of the grains are oriented with their optic axes not greatly inclined to a. (4) Presence of three maxima in the ac concentration zone, situated respectively at angular distances of 25° and 145° from a and parallel to a itself. To test the validity of these maxima a second ac section was cut and the component quartz grains measured. The two sets of results showed a very close correspondence as will be seen in Fig. 6 by comparing the broken curve (representing a single section) with the full curve (combining the results obtained from the two sections). Figs. 7a and 7b show the positions of the deduced maxima in stereographic projections upon ab and ac respectively. In Fig. 7c a contoured orientation diagram, such as might correspond to the orientation curves of Figs. 4 to 6, has been constructed in stereographic projection. The positions of contours are not represented accurately, but their general pattern has been deduced broadly from consideration of the orientation curves. Thus the points where the a axis is cut by contours corresponding to the percentages 2, 4, 6, 8, 10 and 12 the ac curve of Fig. 5 were constructed stereographically in the contoured diagram. Some degree of elongation of one or both of the principal maxima toward the negative end of the b axis is indicated by the form of the bc curve, and is shown accordingly in Fig. 7c. The reconstructed figure though obviously far from exact does indicate the main features of tectonic significance, viz. the positions of the main maxima and minima, the broken ac girdle, and the general pattern of the quartz orientation. Further it allows rapid comparison with the standard contoured diagrams constructed from results obtained with a universal stage.

Fig. 7.—(a) Stereographic projection upon ab showing the points of maximum concentration of quarts ‘axes’ (Z) in Pomona Island granite, No. 4525. (b) Stereographic projection upon ac showing relative positions of S1, S, and Sa, In specimen No. 4525. (c) Stereographic projection upon ab with hypothetical contoursrs indicating the type of orientation that could account for the data recorded on the orientation curves in Figs. 4, 5 and 6 (quarts in specimen No. 4525). In the previous section it was shown that the lineation in the granites of the southern marginal zone was developed by flow of the partially crystallised magma parallel to the lineation itself. The quartz fabric just described however, is not obviously related to such a moverdent arid must have originated during continued deformation after cessation of magmatic flow. The universal undulose extinction exhibited by the grains, now generally interpreted as a result of gliding in crystalline grains on (1010) in a direction parallel to the vertical crystallographic axis (Fairbairn, 1937, p. 38), is itself an indication that the present orientation of the grains is due to “plastic” deformation of the solid rock and not to “viscous” deformation of a partly crystalline magma. From the orientation diagrams it follows that the direction of gliding was perpendicular to the lineation. On one interpretation it may have occurred in two principal planes (S1 and S3 of Fig. 7b) inclined respectively at 25° and 145° to the foliation (S2), and

probably also to some extent in the plane of foliation itself. Sub-horizontal compression acting more or less perpendicularly to the foliation in the closing phases of orogeny could give rise to such a movement in S1 and S3 and, need not have involve much actual transport (cf. Sander, 1934, p. 42). On the other hand, following Schmidt's interpretation of the significance of maxima in B-tectionite girdles, the orientation diagrams just described may indicate operation of a compressive force acting parallel to the a fabric axis (cf. Schmidt, 1932, fig. 49; Phillips, 1937, p. 598). In either case there is an intimate relation between the linear and foliated structures produced by magmatic flow, and the quartz fabric resulting from later deformation of the solid rock, though the actual directions of the movements concerned in the earlier and the later phases are mutually perpendicular. Fairbairn (1937, pp. 112, 113) summarises other instances where the fabrics of foliated intrusive rocks show that the latest movements occurred across the flow-lines after solidification, thus giving rise to a B-tectonic girdle perpendicular to the principal lineation. In conclusion, attention may be drawn to the similarity between the reconstructed contoured diagram representing the quartz fabric of the Manapouri granite, and the orientation diagrams for quartz in certain foliated pegmatites discussed by Sander (1930, pp. 184, 185; D.30, 31, 33, pp. 307, 308). Acknowledgements. The writer gratefully acknowledges the financial assistance of the Australian and New Zealand Association for the Advancement of Science in defraying the cost of field work. He is also indebted to Dr. L. H. Briggs for assistance in field work and to the Murrell brothers and Miss Murrell, of Manapouri, for their hospitality. His thanks are also extended to Professor W. N. Benson for helpful advice during laboratory investigations. Literature Cited. Balk, R., 1925. Primary Structure of Granite Massive, Bull. Geol. Soc, Am., vol. 36, pp. 679–606. —1937. The Structural Behaviour of Igneous Rocks, Mem. Geol. Soc. Am., no. 5. Bucher, W. H., 1933. The Deformation of the Earth's Crust, Princetown University Press. Fairbairn, H. W., 1937. Structural Petrology, Queen's University, Kingston, Ontario. Phillips, F. C., 1937. a Fabric Study of some Moine Schists and Associated Rocks, Q. J. G. S., vol. xciii, pp. 581–620.

Sander, B., 1930. Gefükekunde der Gesteine, Vienna, J. Springer. — 1934. Petrofabrics and Orogenesis, Am. Jour. Soi., vol. xxviii, pp. 37–50. Schmidt, W., 1932. Tektonik und Verformungslehre, Berlin. Turner, F. J., 1936. Interpretation of Schistosity in the Rocks of Otago, New Zealand, Trans. Roy. Soc. N.Z., vol. 66, pp. 201–224. — 1937. The Metamorphic and Plutonic Rocks of Lake Manapouri, Part 2, Trans. Roy. Soc N.Z., vol. 67, pt. 1, pp. 83–100. — 1937a. The Metamorphic and Plutonic Rocks of Lake Manapouri, Part 2, Trans. Roy. Soc. N.Z., vol. 67, pt. 2, pp. 227–249. — 1938. Petrofabric Investigations of Otago Schists, No. 1, Trans. Roy– Soc. N.Z., vol. 67, pt. 4, pp. 443–462.

The late Edward Meyrick.