Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 67, 1938
This text is also available in PDF
(3 MB) Opens in new window
– 83 –

The Metamorphic and Plutonic Rocks of Lake Manapouri, Fiordland, New Zealand—Part 1

[Read before the Otago Branch, October 13, 1936; received by the Editor, October 14, 1936; issued separately, June, 1937.]


  • Introduction

  • The Eastern Manapouri Province

  • Holmwood Island Gneisses

  • Beehive Epidiorite

  • Pomona Island Granite

  • The Contact Zones


    East Coast of Pomona Island


    North Shore opposite Pomona Island


    Summary of Contact Phenomena

  • Literature Cited in Part I


The following paper is based upon about six weeks' field-work carried out during the summers of 1934–5 and 1935–6, in the vicinity of Lake Manapouri. The lake is about 50 sq. mls. in extent; its general east-west trend is broken by three long arms, one extending in a northerly and two in a southerly direction, so that the coastline is highly indented. Tertiary conglomerates and arkosic sandstones outcrop continuously in low cliffs that border the eastern portion of the lake, except where Pleistocene moraines and gravels mantle the older rocks between the points of entry and emergence of the Waiau River. Further west, however, the mountains that rise precipitously from the lake, as well as most of the islands that dot the surface, are composed entirely of ancient gneisses and acid plutonic rocks. These form the subject of the present paper.

In spite of the heavily forested nature of the country, good rock-exposures abound along the steeply sloping shores, and are sometimes continuous for distances of a mile or more. Conditions are thus ideal for working from a small boat. The writer's observations were confined to the West and North Arms, the northern shore of the main lake west of Beehive Peninsula, the southern shore east of the Hope Arm, and the various large islands. The southern shore between the Hope and South Arms, and the shores of the South Arm itself, have not yet been visited. In addition, a rapid traverse of the track between the head of the West Arm and Deep Cove, Doubtful Sound, was made on two occasions.

For the purpose of this discussion the Manapouri region will be divided into three provinces, each of which is to be regarded as a geological unit with its own history and peculiar problems:—


The Eastern Manapouri Province includes the large islands and the northern coast east of a point opposite the northwest corner of Pomona Island.

– 84 –

The Western Manapouri Province includes the shores of the whole lake west of this latter point.


The Dashwood Valley and the area between Wilmot Pass and Deep Cove from part of a third unit here termed the Doubtful Sound Province.

The boundary between the eastern and western provinces is determined by a zone of major dislocation that outcrops on the lake-edge at the eastern end of the long boulder-beach opposite the north-west corner of Pomona Island, and probably separates the Cathedral Peaks mountain-block from the lower country further east. The eastern boundary of the Doubtful Sound Province cannot be defined accurately on account of the dense bush, but probably lies not far east of the Dashwood Valley and Wilmost Pass. Here the rocks have all been profoundly affected by shearing, and it is not improbable that a major dislocation here also separates the two provinces.

To avoid confusion each of the three provinces is treated separately below. In the final section possible correlations for the Manapouri region as a whole are discussed. To facilitate publication the paper has been divided into two parts, the first of which deals with the Eastern Province.

The Eastern Manapouri Province.

Holmwood Island Gneisses.

The oldest rocks in the province are a series of gneisses which are exposed on the chain of islands (Mahara, Holmwood, Isolde) extending in a northerly direction from the eastern side of the entrance of the Hope Arm towards the Beehive. Except for a small mass of epidiorite on Holmwood Island and sparsely scattered veins of pegmatite, these islands are composed entirely of gneiss. The strike varies from 330° to 350°, and the dip is almost vertical.

The principal types are plagioclase-biotite-hornblende-gneiss, plagioclase-biotite, gneiss and amphibolite; a plagioclase-orthoclase-biotite-gneiss is less common.

Plagioclase-biotite-hornblende-gneisses (2622, 2623, 2846–2848, 2851). These are massive grey rocks the gneissic structure of which is due to partial parallelism of biotites and hornblendes. The average composition * is plagioclase 40%, quartz 10%, biotite 25%, hornblende 20%, sphene 1%, epidote 1% and accessory apatite and sometimes magnetite. The quartz content varies from 5% to 20% in different specimens, but the proportions of the other constituents are more uniform. No. 2847 is exceptional in that epidote makes up 10% of the total composition. The feldspar occurs in equidimensional grains ranging up to 1 mm. (rarely 2 mm.) in diameter. The grains of hornblende are often shapeless and rather spongy with no tendency toward elongation parallel to c (Fig. 1); less commonly faces of the prism zone tend to be distinct. Cataclastic structures are generally absent.

[Footnote] * The composition for each slide is roughly estimated by inspection.

– 85 –

The composition of the plagioclase lies between An27 and An42, but acid to medium andesine is the usual variety. It is frequently twinned according to albite and pericline laws and is water-clear except for sericitised central cores enclosed by some grains in 2623 and 2847. Slight marginal granulation and development of a myrmekite-like intergrowth with quartz are rarely shown (e.g. 2622) by the larger crystals. The hornblende is a green pleochroic variety having

  • X = pale yellow,

  • Y = deep olive-green,

  • Z = deep blue-green;

  • Z ∧ c = 25° to 27°.

The biotite occurs in clear-cut flakes with well-developed basal planes: it is strongly pleochroic from pale yellow to deep sepia-brown sometimes with a tinge of green. The grains of quartz are usually smaller than those of the other minerals and tend to occur in nests; undulose extinction is rare. The apatite is typically idioblastic with a somewhat stout prismatic habit. The sphene, on the other hand, takes the form of pale-yellow irregular drop-like granules or occasionally coarse wedge-shaped crystals reaching 0.5 mm. in length. The crystals of epidote though coarse are seldom well-formed; a colourless or very pale yellow highly birefringent type is common, but in the epidote of 2623 the birefrigence is only 0.02.

Plagioclase-biotite-gneisses (2429a, 2430a, 2625, 2792, 2852). These are rather lighter in colour than the rocks just described and differ from them mainly in the absence of hornblende. The mean composition is plagioclase 60%, quartz 20%, biotite 20%, with epidote and sphene as abundant accessories. Small amounts of apatite are always present, and usually there are few grains of iron-ore which may be rimmed with granular sphene (2625, 2852). Small cores of brown allanite are enclosed within a few of the epidotes of one rock (2625) and a little acicular pale blue-green amphibole was noted in another (2792). The composition of the plagioclase is An30 to An36 except in one specimen (2792) where a basic andesine An45 prevails.

Amphibolites (2429, 2430, 2626, 2793). Dark, fine-grained, almost non-schistose rocks of hornfelsic aspect are interbedded with the gneisses on Holmwood and Isolde Islands. The chief constituents are plagioclase and hornblende in about equal quantities (e.g., 2793), accompanied by about 2% or. 3% of iron-ores and accessory apatite, sphene and sometimes epidote. The grain is usually very even (about 0.2 mm.) and the constituent minerals show no obvious orientation (ef. Fig. 2). Typically the plagioclase is acid to medium andesine but values ranging from An25–30 (2429) to An45–50 (2793) have been recorded. The hornblende is the usual deep blue-green type with an extinction-angle (Z to c) of 27°; in one case Z to c = 18° (2626). Biotite is often present to the extent of about 10%, and in one rock (2429) there are small quantities of quartz. Iron-ore is more abundant than in other members of the gneiss group, and tends to be associated with the hornblende. Apatite is always relatively plentiful, while small very irregular sphenes are universally present.

– 86 –

Plagioclase-orthoclase-biotite-gneisses (2624, 2850). These differ from the plagioclase-biotite-gneisses mainly in the presence of 10% to 15% of potash-feldspar, which occurs as exceedingly irregular, almost dendritic interstitial grains developed between the plagioclases or quartzes, and also as sharply-defined antiperthitic inclusions scattered through the larger grains of plagioclase. The plagioclase is a medium oligoclase in 2624, an acid andesine in 2850. Quartz is unusally plentiful (35%) in the latter rock. In both rocks the quartz appears for the most part to have recrystallised under pressure, the plagioclases show traces of marginal granulation, and myrmekite has occasionally developed. In 2850 there has also been some recrystallisation of fine shredded biotite along the margins of the large feldspars. It is possible that the potash-feldspar is an original constituent of the gneiss, crystallising simultaneously with the other minerals. On the other hand, its curious interstitial development combined with the evidence of strong shearing rather favours the possibility that the potash-feldspar has subsequently been introduced along crush-zones by solutions of magmatic origin.

Origin. Any hypothesis of origin of the gneisses must explain the following points:—


The regularly banded structure and rapid alternation of rock-types across the strike.


The N.N.W. trend and steep dip of the foliation.


The chemical and mineralogical compositions of the individual rock-types represented.


The constant development of mineral assemblages consisting essentially of a combination of two or more of the phases andesine, hornblende, biotite and quartz.

These facts are consistent with the hypothesis that the gneisses of the Holmwood Island group are in the main para-gneisses—products of high-grade metamorphism of sediments of greywacke composition. On this assumption the amphibolites might be interpreted as corresponding to intercalated basic flows of tuff-beds. The latter alternative is considered more probable in view of the small thickness (a foot or two) of some of the bands concerned, and the transition with the incoming of biotite towards the plagioclase-biotite-hornblendegneisses of sedimentary origin.

Beehive Epidiorite.

The steep-sided, glaciated boss-like mass terminating the Beehive Peninsula, and the whole of Rona Island, three-quarters of a mile to the east, are composed of a dark coarse-grained epidiorite. On the west the Beehive mass is adjoined by the invading Pomona Island grantite which here and there encloses masses of partially granitised epidiorite. A similar epidiorite cuts irregularly across the gneisses on the eastern side of Holmwood Island. Its age is therefore established as being later than the gneisses but earlier than the Pomona Island granites.

Microscopically the epidiorites are coarse-grained, massive rocks, characteristically very dark in colour as the result of the deep

– 87 –

purplish tint usually displayed by the feldspars. Some specimens (e.g., 2794, 2627) have obviously suffered severe shearing. In section the principal constituents are plagioclase and hornblende in about equal amounts, usually accompanied by 5%–10% of biotite. Apatite and iron-ore are constant accessories, while minor secondary constituents include chlorite, epidote and sphene. The plagioclase varies from acid andesine to basic labradorite, some variety of labradorite being typical. It is in large subidiomorphic to allotriomorphic grains 0.5 mm. to 1 mm. in diameter which in sheared rocks may show slight peripheral granulation. Except in Section 2794 (Holmwood Island) the feldspars are densely clouded with minute opaque inclusions. In 2794 sheaves and patches of relatively coarse sericite (up to 1 mm. long) are developed here and there in otherwise clear crystals of feldspar.

The chief dark constituent (35% to 60% of the total composition) is a green amphibole which varies a good deal both in composition and habit. In 2794 (Holmwood Island) and 2628 (Rona Island) the crystals are large (1 mm. to 2mm.) homogeneous and strongly schillered (X = pale yellow, Y = deep olive-green, Z = deep bluish-green; Z > y > X; Z to c = 180°). In 2794, sections cut perpendicularly to the acute bisectrix X show a perfect basal parting developed as an apparent cleavage parallel to Y; this suggests that the hornblende is pseudomorphous after augite, a conclusion which accords with the frequent presence of schiller-inclusions already remarked upon. In other rocks the amphibole builds up composite masses with subparalled fibrous or irregular fine-prismatic structure, the central portion usually being lighter in colour than the peripheral zone (Fig. 3). For example, in 2838 (S.E. corner of Beehive) the central zone is composed of tremolitic amphibole with X = colourless, Z = pale green, Z to c = 22°, while the marginal portion consists of deep blue-green hornblende optically identical with that described above. These composite masses of amphibole are believed to have developed as pseudomorphs after augite or other pyroxene, and in one specimen (2842, Beehive) residuals of clinopyroxene in process of replacement by hornblende are still recognisable. In most of the epidiorites there is also a small quantity of fine-grained prismatic hornblende which has crystallised between the feldspars.

The biotite is pleochroic from pale yellow to deep yellowishbrown. Coarse flakes (1 mm. in length) grouped around the ironores are probably of primary origin. Typically, however, the mineral takes the form of small flakes bordering and sometimes definitely replacing hornblende, or else developed as strings along the feldspar junctions. In both cases there is a tendency toward radial disposition of biotites with regard to the enclosed crystal, and penetration in parallel position along the cleavages of feldspar is not uncommon (e.g., 2843). In a number of rocks there has been late partial replacement of biotite and to a less extent hornblende by coarse chlorite (uniaxial, positive, X = deep green, Z = pale yellow; anomolous interference tints lacking). Secondary granular sphene occasionally builds up narrow rims surrounding grains of iron-ore

– 88 –

(e.g., 2838), while a few rocks contain small quantities of granular pale-yellow epidote bordering the hornblendes (2627) or distributed along the feldspar junctions (2843).

The rocks just described are regarded as altered gabbros on account of the basic nature of the feldspar, the universal abundance of the dark constituents and the secondary origin of the hornblende as pseudomorphs after pyroxene. The mineralogical and mechanical changes observed may be summarised thus:—

augite → hornblende (usually complete);

hornblende → biotite (common, never far advanced);

ilmenite → sphene;

plagioclase → sericite (rare);

biotite or hornblende → chlorite (common);

secondary crystallisation of hornblende, biotite and epidote along cracks in and between feldspar;

clouding of feldspars (almost universal);

marginal granulation of feldspar (rare).

Narrow dykes of pegmatoid rock consisting principally of coarse oligoclase and partially recrystallised quartz cut the epidiorites at several places. In one specimen (2795, cutting gneisses and epidiorite, Holmwood Island) plentiful plum-coloured tourmaline coarsely intergrown with quartz builds up macroscopically conspicuous aggregates up to 3 cm. in diameter; the same rock contains accessory muscovite. Though potash-feldspar is absent these pegmatoid rocks can hardly be regarded as segregation-veins, but from the presence of tourmaline in 2795 must be connected genetically with the invading granites. Quartz-epidote veins carrying minor brown biotite and apatite (e.g., 2844) are must less common.

Near the north-west corner of the Beehive, where the cliffs turn westward into Clam Bay, the epidiorite encloses several large masses of dark fine-grained amphibolite, one of which is 20 yds. in length. These are fine-grained non-schistose rocks (e.g., 2845) consisting of acid andesine 50%, blue-green hornblende 35%, biotite 10%–15% and accessory apatite, iron-ore and sphene. The rock differs from the usual amphibolites of the Holmwood Island group in the spongy, sometimes fibrous structure of the hornblende, and in the presence of a few subparallel elongated crystals of feldspar 1 mm. to 2 mm. in length scattered irregularly through the usual fine-grained mosaic. Further, the feldspars show a tendency towards cloudiness, and there appears to have been some replacement of hornblende by biotite. The rock is considered to be a block of amphibolite that has been caught up within the invading basic magma, and after consolidation of the latter has been subjected to the same process of metamorphism as was responsible for the transformation of the enclosing rock from gabbro to epidiorite.

Pomona Island Granite.

A massive pink granite, without visible foliation but often splitting with relative ease under a blow from the hammer, is exposed extensively around Pomona Island and along the northern shore of the lake between Pomona and the Beehive. To the west it invades

– 89 –

foliated rocks believed to be equivalents of the Holmwood Island gneisses, while on the east it adjoins the Beehive epidiorite. Extensive blocks of contaminated epidiorite and long narrow bands of hornblende-biotite-schist probably representing altered amphibolite are frequently enclosed within the granite, especially along the coast of Pomona Island. The date of intrusion is thus later than that of the epidiorites.

Where jointing is conspicuous the strike varies between 235° and 260° and the dip is nearly vertical. The joints thus trend at right angles to the predominant strike of the Holmwood Island gneisses and of the bands of schist enclosed within the granite itself.

Though there is a good deal of variation, granites of the Pomona type are characterised by abundance of potash-feldspar (usually microcline-perthite), presence of subordinate basic albite or albiteoligoclase, low biotite content (usually 5%) and granular interstitial development of the quartz. This last feature appears to be the result of shearing during the final stages of consolidation. No. 2500 may be described as typical. The composition is potash-feldspar 50%, basic albite 15%, quartz 30%, biotite (partially chloritised) 2%–3%, sphene 1%, accessory apatite, iron-ore and rare allanite. The potash-feldspar is mainly rather coarse-textured perthite, slightly sericitised, and commonly showing undulose extinction and local development of microcline-structure or even incipient granulation. Many of the plagioclase crystals enclose cores that are crowded with rather coarse prismatic epidote and sericite, perhaps representing xenocrysts derived from the adjacent epidiorite. The biotite is a yellowish-brown variety persistently occurring as irregular streaky aggregates of rather small ragged flakes, with which grains of iron-ore in all stages of replacement by granular sphene tend constantly to be associated. Recrystallisation of quartz as an interstitial mosaic of small grains is almost complete. Apatite is locally plentiful as slender prisms enclosed in those grains of plagioclase that have suffered replacement by epidote and sericite as noted above. The allanite is deep brown, strongly pleochroic and poorly birefringent and is rimmed with a narrow border of colourless epidote. No. 2616 (N.E. corner of Pomona) is unusual in the structure of the feldspar and the presence of accessory pale pink garnet associated with the biotite. Here the feldspar is almost entirely in the form of a coarse intergrowth of microcline and a somewht less amount of albitic plagioclase. Thus in a composite crystal 2 mm. in diameter the albite patches, in optical continuity throughout, may reach dimensions as large at 0.5 mm. square and often display regular albite-twinning. This is quite different from the normal streaky perthitic structure displayed by the potash-feldspar in most of the Pomona granites. It may be merely an ex-solution phenomenon or possibly is the result of a late-magmatic replacement of plagioclase by microcline.

The Contact Zones.

(a) East Coast of Pomona Island.

Along the eastern coast of Pomona Island outcrops of granite alternate with masses of black epidiorite ten to twenty chains in length, while narrow lenses of hornblende-schist are enclosed here and

– 90 –

there with either granite or epidiorite. There is also an extensive outcrop of epidiorite on the northern shore of the lake immediately opposite the north-east corner of Pomona Island, and this may conveniently be considered with the rocks of the island itself. The rocks of the contact zones include, in addition to the normal granites described above, contaminated epidiorites, hornblende-biotite-schists and obviously contaminated granites.

Contaminated Epidiorites. Epidiorites in various stages of contamination are especially plentiful toward the southern end of the section, where they are often intricately veined by narrow veinlets of granite injected along the lines of jointing. There is perfect transition from rocks closely resembling the Beehive epidiorites to thoroughly contaminated rocks. The different stages in alteration are illustrated by the following examples:—

No. 2836 (north shore opposite N.E. corner of Pomona Island) is a massive non-foliated rock of dioritic aspect, rather lighter in colour than the typical epidiorite of the Beehive. The structure is esentially gabbroid (average grain-size 1 mm. to 1. 5 mm) and the approximate mineral composition is plagioclase 40%, quartz 15%, hornblende 35%, epidote 5%, biotite 5%. Apatite and black ironore are abundant accessories, while sphene, sericite and chlorite are present as alternation-products. The plagioclase is an acid andesine (An30–32); it usually encloses a good deal of coarse, prismatic pale-yellow epidote and locally has suffered replacement by sericite. The hornblende is a blue-green type (Z to c = 20°) occurring as large homogeneous sometimes sieved crystals, with which aggregates of yellowish-brown biotite tend to be associated. There appears to have been some replacement of hornblende by pale irregular epidote. Nests of quartz (0.5 mm. to 1mm. in diameter) have everywhere been introduced between grains of other minerals, especially between the feldspars. Apatite is unusually plentiful and commonly displays an acicular habit in contrast with the sparsely scattered stout prismatic grains of the normal epidiorite. The iron-ore is bordered by narrow rims of granular sphene. A duplicate section (2836a) differs from that just described in the presence of a much greater quantity of epidote (20% of the total composition), the almost completely chloritised condition of the biotite and the presence of calcite and coarse prismatic tourmaline (X = pale pink, Z = deep indigo). These features are attributed to local pneumatolytic action. On the other hand, the acid composition of the feldspar, the presence of quartz and the development of acicular apatite are the results of the first stages of reciprocal reaction between the epidiorite and the granite magma (see below).

A rather more advanced stage of contamination is illustrated by 2513, 2515 and 2607. Quartz is not abundant in any of these rocks, and in 2516 is absent, but the plagioclase is more sodic (An12–30) than in the rock described above, and the amount of biotite (10% to 25%) is equal to or slightly less than that of hornblende. Replacement of iron-ore by dense granular clusters of sphene is almost complete. An unusual feature is the presence in 2513 of a single very large idiomorphic grain (0.4 mm. × 0.15 mm.) of zircon or

– 91 –

cassiterite. In 2607 there are signs of incipient crushing, while in 2516 the feldspars have suffered almost complete granulation and the amphibole is a pale blue-green actionolite type sometimes nearly colourless in the central portion of the crystal. The next stage in granitisation of epidiorite (2512, 2514) is marked by almost complete absence of hornblende, corresponding increase in the content of biotite (now the principal mafic constituent), incoming of small amounts of interstitial perthitic orthoclase and great abundance of acicular apatite. The composition of 2512 is acid or medium oligoclase 40%, orthoclase 10%, quartz 5%, biotite 35%, hornblende 7%, sphene 1%, apatite 1%, epidote 1% and a little iron-ore. The epidote occurs as dense clusters enclosed in some of the crystals of plagioclase. The sphene is in rounded granular aggregates sometimes enclosing small residual grains of iron-ore and often showing roughly radial structure. The feldspars display undulose extinction or marginal granulation while the micas are often rather twisted as a result of local shearing.

Nos. 2515 (1ft. from grantite contact) and 2504 (“epidiorite” seamed with granitic veinlets) represent the end-product of hybridisation prior to mechanical disintegration of the invaded rock (see Fig. 4). The composition of 2515 is albite-oligoclase 30%, quartz 20%, potash-feldspar 20%, biotite 25%, with about 1% each of apatite, sphene and epidote and accessory interstitial calcite. The crystals of plagioclase are clearer than in the other contaminated rocks, but the central portions are often charged with prismatic epidote. The potash-feldspar is in water-clear highly irregular interstitial grains locally showing microcline or perthite structures and sometimes enclosing finely intergrown vermicular quartz. Myrmekitic outgrowths from the large crystals of plagioclase are often developed against the intervening orthoclase. The biotite has a greenish-brown absorption tint for Y and Z. Apatite and sphene have the same properties as in the rocks described above.

The progressive granitisation of the invaded epidiorites involves the following changes:—


Conversion of basic plagioclase to acid oligoclase or albiteoligoclase with separation of epidote.


Disappearance of hornblende and corresponding increase in the amount of biotite.


Conversion of iron-ore to granular sphene.


Crystallisation of abundant acicular apatite.


Introduction of quartz.


Introduction of potash-feldspar (only in the advanced stages of contamination).


Rather rare local hydrothermal alteration involving chloritisation of biotite, replacement of hornblende by epidote and crystallisation of tourmaline or calcite.

The epidiorites, like the enclosing granites, often show local effects of shearing. That this shearing was partly contemporaneous with the intrusion of the granite is shown by local contamination of epidiorite along shear-zones. Thus towards the north-east end of

– 92 –

the Pomona Island section the epidiorite is strongly gneissic. The coarser bands (e.g., 2612) retain to some extent their original structure and composition though much quartz has been introduced. But the interlaminated fine-grained bands representing completely granulated epidiorite (e.g., 2611) have reached an advanced stage of contamination (acid oligoclase 40%, quartz 40%, biotite 15%, epidote 2%, sphene 1%, accessory iron-ore and apatite). The quartz in this fine-grained phase has been introduced as continuous irregular veinlets and streaks up to 2 mm. in width and has itself suffered complete recrystallisation.

Nockolds (1933) has recently discussed the assimilative action between granitic magma and basic rocks as taking place in two stages, viz., (a) reciprocal reaction between the invading liquid and the invaded solid rock, and (b) mechanical disruption of the solid rock after equilibrium with the contaminated magma has been nearly or completely reached. The changes listed above as (1) to (4) are typical of the reciprocal reaction stage, while introduction of quartz and orthoclase indicates the commencement of the mechanical disruption stage. Extensive development of apatite and reactive replacement of hornblende by biotite testify to the importance of the part played by volatiles during reaction (Nockolds, 1933, pp. 562–569). It is interesting to note that though quartz is frequently introduced into the epidiorities during the initial stages of contamination, potash-feldspar is similarly introduced only into rocks in which development of biotite at the expense of amphibole is complete or nearly so.

Mechanical disruption on a larger scale has also commenced at the granite-epidiorite contacts, for here there is often a narrow border-zone of granite enclosing scattered xenoliths. Curiously enough these xenoliths, though largely or completely recrystallised under the influence of the enclosing granite, are distinctly different mineralogically and structurally from the immediately adjacent contaminated epidiorite. The two specimens described below (2508, 2511) were collected at the southeren end of Pomona Island, within two feet of the contaminated epidiorite 2515 described on p. —. No. 2508 is a dark hornsfelsic rock consisting essentially of acid oligoclase and greenish-brown biotite in about equal proportions, together with about 10% of coarse introduced quartz which builds up conspicuous spots 2 mm. in diameter. Epidote, acicular apatite and finely granular sphene each make up about 2% of the total composition, while deep blue-green hornblende is even less plentiful. For the most part the plagioclose has completely crystallised as rather irregular grains (0.3 mm. to 0.4 mm. in diameter) clear except for enclosed needles of apatite; but a few large crystals enclosing abundant prismatic epidote still persist. The second specimen (2511) is from a larger xenolith about 3 in. in diameter. The composition is basic olgoclase 30%, biotite 30%, hornblende 30%, epidote 7%, sphene 2%, apatite 1%, and the grain-size averages 0.2 mm. There is a tendency for the epidote and occasionally the hornblende to be concentrated in knots (up to 2 mm.) consisting entirely of the one mineral.

– 93 –

Hornblende-and Biotite-schists. Large lensoid masses of dark-coloured schists are enclosed either by the granite or within the epidorite at several points. The strike of these bands is always within 20° east or west of north and the dip is usually at high angles towards the west. Within the granite are 2501 (long band 3 ft. wide at extreme southern end of Pomona), 2619 (mass 6 ft. wide towards north end of Pomona) and 2621 (lens 20 ft. × 3 ft. on north shore opposite north end of Pomona). The remaining specimens (2611, 2613—2615, 2617, 2618) all come from a lens 10 ft. × 2 ft. enclosed in sheared epidiorite about halfway along the east coast of the island.

All except two specimens (2614, 2619) are hornblende-biotiteoligoclase-schists. Hornblende with strong pleochroism (X=pale yellow, Y=deep olive-green, Z=deep blue-green, Z > Y > X, Z to c=22°) is the dominant mineral (30% to 60% of total composition of rock). Yellowish-brown biotite usually makes up 10% to 30%, but in 2611 is only a minor constituent. The plagioclase (about 30%) is medium to basic oligoclase in the form of small untwinned grains sometimes difficult to distinguish with certainty from quartz which may be present in small quantity (e.g., 2603, 2615) or else completely absent. The only other essential mineral is a pale yellow or colourless highly birefringent epidote (typically 10% to 20%, but less in 2613). Relatively coarse often well-formed crystals of pyrite may be numerous (2621, 2615), apatite is a constant but never plentiful accessory, while drop-like grains of sphene are fairly common in some rocks. In every section the plagioclase takes the form of unoriented shapeless grains 0.05 mm. to 0.2 in diameter, while the biotite occurs as ragged flakes of comparable size arranged parallel to the schistosity. The epidote, too, is seldom coarse, but it usually has a tendency to form idioblastic prismatic crystals; in 2615 well-formed crystals ranging up to 0.5 mm. × 0.2 mm. are common, but these are a good deal larger than in other rocks of this group. The habit of the hornblende varies a good deal. Typically the crystals are slender, idioblastic indistinctly terminated prisms 0.5 mm. to 1 mm. long which often trend parallel to a particular direction in the schistosity (Fig. 6); in some such cases (e.g., 2501) there is also an obvious tendency for the orthopinacoid to lie parallel to the schistosity-plane (cf. Turner, 1936, p. 215), definitely indicating a tectonite orientation-pattern for the hornblende-fabric. In other rocks (e.g. 2603) the crystals of hornblende are sieved stout porphyroblasts, oriented irregularly in the plane of schistosity, and thus apparently the product of growth under static conditions in the direction of minimum resistance afforded by the previously developed schistosity (cf. Sander, 1930). A special type of porphyroblastic structure is illustrated by 2613 and a less extent 2611, in which there are macroscopically conspicuous spots 2 mm. to 3 mm. in diameter, consisting of unoriented stout hornblendes sometimes accompanied by a little biotite (Fig. 5).

Biotite-schists locally containing very plentiful pyrite are represented by two specimens, 2614 and 2619. The first of these is a dark-brown highly schistose rock, obviously rich in dark mica and

– 94 –

lacking the greenish colour that is so typical of the hornblendeschists. The grain is even (0.1 mm. to 0.3 mm.) and the average composition is epidote 40%, yellowish-brown biotite 30%, basic oligoclase 20%, quartz 10% and accessory pyrite, sphene and apatite. The epidote is in subidioblastic grains (0.3 mm. × 0.1 mm. or sometimes larger) with only the long edges well defined, rarely aggregated into spots. It is colourless to pale yellow, optically neutral and only moderately birefringent. No. 2619 differs from the rock just described in the greater quantity of plagioclase and much lower epidotecontent of the rock, while the structure is almost hornfelsic. The composition is biotite 35%, medium oligoclase 45%, quartz 15%, pyrite 5%, apatite 1% and accessory epidote. The biotite, occurs as large subidioblastic flakes 0.2 mm. to 0.4 mm. in length, and minute flakelets crowding the larger grains of feldspar.

The schists of the Pomona Island contact-zone cannot represent primary magmatic segregations since identical rocks are found within both granite and epidorite. They must therefore be derivatives either of the Holmwood Island gneisses or of the epidiorite; this second possibility may be discarded, however, on account of the rapid variation displayed by the schists coupled with the complete dissimilarity from any known phase of the epidiorite. Although structurally different, many of the schists are mineralogically similar to certain of the plagioclase-hornblende-gneisses of the Holmwood Island group, except that in the latter the plagioclase is more calcic and epidote much less plentiful. They are therefore believed to represent residual masses of reconstituted Holmwood Island gneiss. Except for possible introduction of pyrite and occasionally of apatite (2619) effects of contamination comparable with those shown by the epidiorites are rare; there has been no addition of potash-feldspar or quartz nor any obvious replacement of hornblende by biotite. The apparent lack of permeation of gneiss by granitic fluids is perhaps due to the unjointed condition of the rock in question. For example, near the south-east corner of Pomona well-jointed epidiorite intimately injected with veinlets of granite encloses a mass of compact, hornblende-schist against which the veins end abruptly. In general, then, the gneisses have here been reconstituted to an assemblage (hornblende-biotite-oligoclase-quartz) that can exist in a state of internal equilibrium at the temperature induced by the enclosing or invading granitic magma and under the stress occasioned by or accompanying intrusion. This latter factor accounts for the tectonite febrics of some of the schists, and is perhaps connected with replacement of relatively calcic plagioclase in the original gneiss by a more sodic variety plus epidote in the recrystallised schist. On the other hand, reconstitution of the epidiorites has been in the direction of establishing equilibrum between the new assemblage of minerals and the invading granite magma.

Contaminated Granites. Obvious local contamination of granite near contacts with epidiorite is not uncommon. Two specimens of abnormally dark granite (2502, 2503) collected a few yards distant from such a contact at the southern extremity of Pomona, consist of microcline 40%, albite-oligoclase 20%–30%, quartz 20%, biotite

– 95 –

10%, epidote, 4%, and accessories. Evidence of contamination lies in the increased biotite-and decreased microcline-content as compared with other granites in this area, and the common occurrence of crystals of plagioclase, the central portions of which enclose abundant prismatic epidote, slender apatite, ragged flakelets of biotite or some combination of these. Another contaminated granite (2507) enclosing numerous xenoliths about one foot from an epidiorite contact contains microxenoliths of green schillered hornblende and associated biotite. The veinlets of granite cutting epidiorite a few yards from contacts appear to have been modified by loss of potash, presumably used up in conversion of hornblende to biotite in the invaded rock. The resultant vein-rock (e.g., 2517) contains little or no potash-feldspar, and in consequence is white or grey in colour in comparison with the typically pink Pomona granite.

Though obviously contaminated granites such as those described above occur only locally, certain persistent petrographic peculiarities of the typical Pomona Island granites suggest the possibility that these rocks have suffered widespread contamination. Every specimen of granite examined contains irregular aggregates of small biotites, crystals of plagioclase with central epidote-and apatite-rich cores, and granular masses of sphene surrounding central grains of ironore. All three features are shown identically by the contaminated epidiorites, and it is therefore probable that the crystals concerned are actually xenocrysts derived from the epidiorites by distintegration, and drifted out into the granite magma.

(b) North Shore opposite N.W. corner of Pomona Island.

Along the northern shore of the lake the Pomona Island granite extends continuously from a point near the western end of the Beehive westward to about the middle of the strait that separates the north end of Pomona from the mainland. Here they invade a series of rather coarse grey gneisses cut by occasional dykes of granite-pegmatite. These rocks continue with uniformly N. to N.E. strike and steep eastward dip for about a mile west of the granite-contact, but the section is then obscured by an extensive boulder beach near the mouth of a stream draining the south-east face of Cathedral Peaks. At the eastern end of this beach the gneisses are cut by a zone of intense shattering along which the granitic magma has been injected irregularly. Sharply defined bodies of granite rock are seldom easy to distinguish, but there is complete and irregular transition from gneiss to almost uncontaminated granite throughout a zone about 100 yds. in width.

Uncontaminated Gneisses. The most abundant rock is a coarse, well-foliated greenish-grey gneiss (e.g., 2831, 2832) consisting of oligoclase 60%, biotite 15%, hornblende 10%, quartz 10%, epidote 1% to 3% and accessory black iron-ore, stout prismatic apatite and sometimes zircon (Fig. 7). in 2833 quartz is more plentiful (20% to 30%) and hornblende is absent. The grain-size is uniform (about 1 mm.) and as a rule is little tendency for the principal minerals to exert their crystal outlines. The biotite is a deep

– 96 –

yellowish-brown variety frequently in process of alteration to pale-green chlorite, or occasionally to prehnite. The hornblende usually occurs in irregular equidimensional grains having X=pale yellow, Y=deep olive green, Z=deep blue-green, X < Y < Z, Z to c=22°. The epidote is colourless but highly birefringent (γ-α=0.035), usually irregular, but when enclosed by feldspar sometimes idioblastic. Coarse grains of pale-yellow sphene, sometimes idioblastic but more usually having rounded or embayed outlines, are very plentiful and often reach 0.6 mm. in diameter. In many rocks the only indication of shearing subsequent to crystallisation is the slightly undulose extinction shown by the grains of quartz and occasionally by feldspar. In 2829, however, the grains of feldspar show marked bending of cleavage-lines and twin-lamellae and locally are crossed by narrow zones of incipient granulation, while in 2834 all the quartz and much of the feldspar have recrystallised as a fine-grained moasaic sometimes accompanied by a pseudo-myrmekitic intergrowth of plagioclase and vermicular quartz.

The presence of highly irregular interstitial orthoclase or microcline sometimes enclosing small rounded grains of quartz was noted in two specimens, viz., 2835 (near granite contact) and 2829 (adjacent to the western zone of shattering and granite injection). The possibility that the potash-feldspar of these rocks may be of external origin must not be overlooked; it is strengthened by the occurrence in one of the rocks concerned (2835) of plentiful rounded grains of deep-brown allanite bordered with colourless epidote, for this mineral is a constant accessory mineral in the Manapouri granites.

As regards structure, mineralogical composition and relation to the granitic rocks, the gneisses just described resemble certain rocks of the injection-complex of the Western Manapouri Province so closely that the identity of the two sets of rocks cannot be doubted. It is probable also that they are equivalent to the Holmwood Island gneiss, and in the accompanying map they have been shown as such. The plagioclase is uniformly more sodic than is usual in rocks of the Holmwood Island type, but this is compensated by the presence of epidote and sphene, and the mineral assemblage as a whole is regarded as the product of recrystallisation at the temperature resulting from intrusion of the Pomona Island granite.

Contaminated Rocks. In the great shatter-zone at the western end of the section under consideration three principal classes of rocks may be distingushed, viz., hydrothermally altered gneisses, contaminated gneisses and contaminated granites. Most show at least minor effects of shearing, while in some (e.g., 2814, 2820) a highly cataclastic structure has been developed.

Rocks which have suffered only hydrothermal alteration are relatively rare (2825–2828). They are oligoclase-hornblende-biotite-gneisses with minor quartz, epidote and sphene and accessory apatite and iron-ore. Chloritisation of biotite is almost or quite complete, and partial sericitisation and kaolinisation of plagioclase are characteristic. Prehnite has developed sometimes at the expense of biotite, sometimes as irregular interstitial patches and occasionally as transgressive veinlets (2827). The usual colourless epidote is often

Picture icon

Fig. 1.—Plagioclase-biotite-hornblende-gneiss (2831); biotite, black; a large grain of epidote near the centre of the field. × 45. Fig. 2.—Amphibolite (2430) showing hornblende (dark), mino [ unclear: ] biotite, and plagioclase. × 45. Fig. 3.—Epidiorite (2837): on the left an aggregate of pale amphibole [ unclear: ] with a narrow border of deeper colour; on the right deep-brown biotite: note the cloudy condition of the feldspar. × 45. Fig. 4.—Contaminated epidiorite (2515) 1 ft. from granite contact. Pomona Island; the dark crystals are biotite; clear areas in bottom left corner are introduced quartz. × 45.

Picture icon

Fig. 5.—Spotted hornblende-biotite-schist (2613) enclosed in epidiorite, Pomona Island. × 45. Fig. 6.—Hornblende-biotite-schist (2301) enclosed in granite, Pomona Island. × 45. Fig. 7.—Coarse plagioclase-hornblende-biotite-gneiss (2832), north shore opposite N.W. corner of Pomona Island: the hornblende encloses quartz and sphene. × 45.

Picture icon

Geological Map of Lake Manapouri. [Note.—(1) Localities where the Pomona Island granite encloses large masses of Epidiorite are indicated as Contaminated Epidiorite abundant in Granite. (2) The area marked Injection-complex consists of Trondhjemites, Oligoclase-granites, etc., invading gneiss; where the latter forms an important element in the complex the sign Gneisses prominent in Complex is employed. (3) The gneisses of the Holmwood Island and Wilmot Pass Groups are shown as ! Dusky Sound Series.]

– 97 –

imperfectly rimmed with a border, having a deep yellow colour and strong pleochroism. Incipient replacement of hornblende by tremolite or epidote was observed in several rocks, but is more likely a result of shearing than of hydrothermal action.

The contaminated rocks form a continuous series ranging from partially granitised gneisses to contaminated granites. The transition is illustrated below by details of three stages. No. 2813 is typical of a group of greenish non-schistose rocks of rather blotched appearance with macroscopically conspicuous large pink feldspars (1 mm. to 3 mm.) that give the rock a “porphyritic” appearance. The constituents are albite 70%, quartz 20%, chloritised biotite 7%–10%, epidote 2%, and accessory apatite, iron-ore, prehnite and rare allanite. Most of the albite is in the form of equidimensional grains 0.5 mm. to 1 mm. in diameter, but there also large subidioblastic crystals (1.5 mm. to 3 mm.) containing central sericitised cores, and a good deal of finer-grained albite resulting from granulation of initially coarser material. Drop-like or vermicular inclusions of quartz are occasionally intergrown with the feldspar. Nests of rather coarsely crystalline quartz with undulose extinction appear in part at least to have been introduced from the granite.

A further stage in contamination is represented by 2817, a relatively light-coloured greyish rock with conspicuous large pink feldspars 3 mm. to 8 mm. in diameter. It differs from the rock just described (2813) in its noticeably coarser grain, greater abundance of quartz and the presence of coarse orthoclase or microcline. The latter occurs as large irregular clear crystals (the pink “phenocrysts”) with local perthitic structure enclosing scattered cloudy albites, or as smaller interstitial masses, its external origin in both cases being obvious. Veinlets of prehnite often accompanied by yellow granular epidote and pale chlorite have developed between the grains of albite and quartz and along shear-zones traversing these minerals. Local aggregates of muscovite sometimes enclosing vermicular quartz (cf. Turner, 1933, pp. 207, 208; Benson and Bartrum, 1935, p. 125) were noted. Coarse colourless epidote, rimmed irregularly with deep yellow, often encloses rounded cores of brown allanite. The micro-strcture shows obvious effects of shearing, while in 2824, which represents about the same stage of contamination, the structure is thoroughly cataclastic.

Nos. 2818 and 2823 represent the extreme product of assimilative reaction in this series. These rocks from definitely intrusive bodies, and are thus to be regarded as contaminated granites though they grade insensibly into rocks such as 2824 that are more appropriately considered as granitised gneisses. The hand-specimens are light-coloured rocks of granitic appearance differing from the Pomona granite mainly in the uneven distribution of the ferromagnesian minerals, which tend to be concentrated in small dark patches. The average composition is potash-feldspar 40%, quartz 25%, plagioclase 30%, with small amounts of chloritised biotite and epidote and accessory muscovite, apatite, allanite and magnetite. The potash-feldspar builds very large water-clear crystals (4 mm.) enwrapping or completely enclosing the plagioclase; much of it shows microcline

– 98 –

structure, and perthitic striping is fairly common. The plagioclase (basic albite or albite-oligoclase) occurs mainly as equidimensional grains (0.5 to 1 mm.) clouded with kaolin or fine sericite, enclosed partially or completely by the large crystals of microcline and quartz. This type has obviously been derived directly from the disentegrating invaded gneisses. Indeed, distinct micro-xenoliths of clouded albite with interstitial biotite, chlorite or epidote are clearly observable in some sections. There are also clear crystals of albitic plagioclase that have crystallised directly from the magma. The effects of post-crystalline shearing are always conspicuous and reach a maximum in 2820 and 2821 in which more than half of the micro-section is composed of fine-grained recrystallised quartz and feldspar enclosing partially shattered porphyroclasts. Here, too, swarms of colourless idioblastic prisms of epidote and aggregates of coarse muscovite have crystallised along the lines of shattering, while cracks in the relict feldspars have been injected with finely crystallised quartz (especially in 2821).

(c) Summary of Contact Phenomena.

The observed effects of reaction between the Pomona Island granite-magma and the invaded rocks in the Eastern Manapouri Region may now be summarised briefly as follows:—


Effects shown by Gneisses. For some distance from the gneiss-granite contacts the Holmwood Island gneisses have been reconstituted to the uniform mineral assemblage oligoclase-quartz-biotite-hornblende-epidote-sphene under the influence of the temperature induced by the intrusion. Close to the contact where the gneisses have been permeated by magmatic liquids, the composition of the plagioclase has become even more sodic (basic albite or albite-oligoclase), and large crystals of “porphyritic” aspect have developed through the rock (cf. Nockolds, 1932, pp. 443–448). The schistosity of the original rock has at the same time been obliterated while hornblende has been replaced by biotite, and quartz has been introduced from the magma. * The next stage is marked by introduction of potash-feldspar (often very coarse) and still more quartz, with consequent disruption of the contaminated gneiss, as single crystals and micro-xenoliths are isolated and drift out into the magma. The presence of allanite in contaminated gneisses is also considered to be a minor effect of reaction with magmatic liquids. Hydrothermal alteration connected with invasion by granite is responsible for widespread chloritisation of biotite, sericitisation and kaolinisation of feldspar and crystallisation of prehnite and deep-yellow epidote in certain gneisses on the north shore of the lake. The pyrite noted in some of the relict masses of gneiss on Pomona is perhaps due to a similar cause.

[Footnote] * The lenses of altered gneiss in the Pomona Island section have not been permeated intimately by magmatic liquids, and therefore do not show these advanced effects of contamination. In these rocks, too, the schistosity was intensified by reconstitution.

– 99 –

Effects shown by Epidiorites. The clouding universally shown by the feldspar of the epidiorites even at distances of more than a mile from observed contacts with the granite is probably to be attributed to the influence of the uprising acid magma (ef. McGregor, 1931; Williamson, 1936, p. 149). To what extent the other changes involved in transformation of the original gabbro to epidiorite are due to the same cause is uncertain. In the contactzone, however, contamination of epidiorite has resulted in conversion of plagioclase to albite-oligoclase with simultaneous separation of epidote, conversion of hornblende to biotite, replacement of ilmenite by sphene, introduction of coarsely crystalline quartz and development of swarms of apatite needles. In the immediate vicinity of the contact potash-feldspar may also be introduced into the epidiorite, but only after replacement of hornblende by biotite is complete. Hydrothermal action is responsible for local crystallisation of chlorite, calcite and tourmaline in some rocks.


Effects shown by Granites. Within a few yards of the epidiorite or gneiss the granites have been noticeably modified by reciprocal reaction, and assume a darker colour and rather heterogeneous aspect in comparison with rocks further from the contacts. The process of contamination is two-fold. In the first place xenolithic crystals or aggregates of plagioclase. biotite, epidote, sphene and rarely hornblende have been strewn thickly through the contaminated facies. At the same time the proportion of potash-feldspar has diminished, and has even fallen to zero in the extreme case of granitic veins extending some distance into the surrounding rocks. For example, the coarse granitic veins and dykes penetrating the gneisses and epidiorites on the Beehive and the various islands, are composed of rocks in which orthoclase is nearly or completely absent, though their otherwise normal composition and the occasional presence in them of tourmaline point clearly to their granitic origin. This impoverishment in orthoclase is probably the result of progressive removal from the magma of potash required for replacement of hornblende by biotite in the adjacent rocks.

Apart from the marginal development of obvious hybrids, there appears to have been some general contamination of the granite magma as a whole, for xenolithic plagioclase, sphene, biotite, and epidote are constantly present in the typical Pomona granites. Further, if the bands and lenses of schist now enclosed in the granites at long distances from the contact represent relict masses of altered Holmwood Island gneiss, considerable assimilative reaction and contamination of the granite must have taken place.

Note.—The geology of the other two provinces and problems of correlation will be discussd in the second part of this paper (now in the press).

– 100 –

Literature Cited in Part I.

Benson, W. N., and Bartrum, J. A., 1935. The Geology of the Region about Preservation and Chalky Inlets, South-west. Fiordland, New Zealand, Part III, Trans. Roy. Soc. N.Z., vol. 65, pp. 108–152.

Harker, A., 1932. Metamorphism, Londop, Methuen & Co.

Mcgregor, A. G., 1931. Clouded Feldspars and Thermal Metamorphiam, Min. Mag., vol. xxii, pp. 524–538.

Nockolds, S. R., 1932. The Contaminated Granite of Bibette Head, Aldernay, Geol. Mag., vol. lxix, pp. 433–452.

— 1933. Some Theoretical Aspects of Contamination in Acid Magmus, Jour. Geol., vol. xli, pp. 561–589.

Sander, B., 1930. Gefügekunde der Gestcine, Vienus, J. Springer.

Turner, F. J., 1938. The Metamorphic and Intrusive Rocks of Southern Westland, Trans. N.Z. Inst., vol. 63, pp. 178–284.

— 1936. The Significance of Schistosity in the Rocks of Otago, New Zealand, Trans. Roy. Soc. N.Z., vol. 66, pt. 2, pp. 201–224.

Williamson, W. O., 1936. Some Minor Intrusions of Glen Shee, Perthshire, Geol. Mag., vol. lxxiii, pp. 145–157.