Art. XXX.—Notes on the Geology of the West Coast Sounds.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 42, 1909, Page 255

Art. XXX.—Notes on the Geology of the West Coast Sounds. By R. Speight, M.A., B.Sc., F.G.S. [Read before the Philosophical Institute of Canterbury, 1st December, 1909.] The following notes are based on observations and collections made during a fortnight's steamer-cruise in the Sounds in the summer of –9. On that occasion all the Sounds were visited, but the time was too short to do more that collect specimens on the shore where landings were made. However, some of these places have not been previously visited by a geologist. The vast extent of country, the difficulty of travelling on its forest-clad and precipitous slopes, and the small number of accessible outcrops of rock on the shore-line of its deep, wall-sided fiords, render a satisfactory account of its intricate geological structure absolutely impossible at present. It furnishes perhaps the most complex of all the problems that New Zealand geologists will have to solve. Sir James Hector (4, 5, 6) and Captain Hutton (8) have already given a general description of the country, and Andrews (1, 2) has traced out the evolution of its landscape in the light of modern physiography, an account which has been generally indorsed by Marshall (14). Petrological notes have been furnished by Hutton (9, 10, 11, 12), based on a few specimens collected by himself at various times; and more extended descriptions have been made by Marshall (13, 14) as a result of collections made by him during various excursions into the region. He has drawn special attention to the magnesian rocks at Anita Bay (13), where he discovered dunite, and harzburgite, and olivine-bearing marble whose origin he ascribed to modifications in the neighbouring harzburgite, a conclusion which he based on the evidence of a series of specimens showing transitions between the two rocks. These rocks have been more recently referred to by Finlayson in an interesting paper on the origin of nephrite, read before the Geological Society (3). The present paper has been contributed in order to supplement previous articles. The author has to express his great indebtedness to Dr. Marshall for the loan of a selection of microscopic sections of rocks described by him in his papers. It has thus been possible to identify some specimens here mentioned as being the same as described by Dr. Marshall, or slight variations from them. In order to give a general idea of the petrology of the area, each locality visited will be taken in turn, and a short description will be given of the rocks collected. It is impossible at present to give any account of their field relations. The notes given are therefore unsatisfactory; but the mere indication that certain rocks occur in particular localities is considered by the author to be of some importance, considering the extent of the area, and the ignorance that at present exists regarding its geological structure. Before dealing with the petrology of the area, brief reference will be made to one or two points of physiographical interest. The first of these deals with certain shore-features. Even at the heads of the Sounds the shoaling caused by the detritus poured in by large streams is limited to the area close to their mouths. Extensive flats do occur, but they present extremely steep faces to the sea beneath water-level. At the head of Wet Jacket Arm, on sounding to get an anchorage on one of these flats, the

depth was one fathom under the bow of a small steamer of eighty tons, and thirteen under her stern. There are no strong currents or tides in the remote recesses of the Sounds to sweep away detritus far from the river-mouth, so the angle of repose is extremely steep. Very fine particles must be carried some distance, as dredgings show that the floors of the Sounds are almost invariably composed of mud. In some cases, such as at the head of Milford, shoaling is very rapid over the area of deposition: at one place, where twenty years ago a depth of fourteen fathoms was recorded, mud-flats covered with driftwood, and exposed at half-tide, now exist. Apart from these, the beaches in process of formation are of extremely limited extent, and if subjected to subaerial erosion would last but a short space of time. Their absence at higher levels, noted by Hector, is therefore no proof that the land did not suffer a depression with subsequent elevation. Captain Hutton noted marine-cut terraces in certain places, particularly at the entrance to Doubtful Sound on Secretary Island, and they are very well marked near Puysegur Point, in the south-west of the area, so that it is probable that a submergence took place, probably posterior to the glaciation; but the question does not seem to be connected in any way with the cause of that glaciation. The main landscape-forms of this area, as has already been pointed out by Andrews (1, 2), are produced by the modifying effect of glaciation on a well-developed stream-valley system. His conclusions can in general be heartily indorsed. There are, however, one or two features which may be further noticed. The first of these is the general orientation of the main valleys. Although small valleys can be found running in all directions, the general disposition of the main valleys seems to depend on some general underlying cause. An examination of the map shows that the main arms of the Sounds, and the lakes on the eastern border of the region, usually lie parallel to two main directions, one running north-west and the other south-west, thus cutting one another at right angles. This arrangement is difficult to explain in the light of our present knowledge, but it is possible that the formation of the valleys was dependent in the first case on lines of crust-fracture, that the arms grew along these lines, and, in spite of valley-forms being modified, the main directions were always preserved. The smaller valleys running in irregular directions would be subsequent, and determined largely, if not altogether, by the way in which streams eroded the surface. It has been pointed out by J. W. Gregory that fiords are found on the margins of fractured earth-blocks, and it is extremely likely that this is the case here, although there is no positive evidence of their presence. There seems, however, to be considerable difficulty in explaining the origin of the submarine valley of Wet Jacket Arm, and the narrow Acheron Passage, running for miles behind Resolution Island, and almost exactly at right angles to the Arm. The passage is deep throughout its whole length according to the soundings on present charts, and there is little rounding of the corners where it joins with Wet Jacket Arm. The soundings do not reveal that it has been formed by the degradation of a divide by ice erosion. It can be most easily explained as being the result of a cross-fracture. Similar valleys on a less pronounced scale are found connecting Daggs Sound with Doubtful Sound, though it is possible that these may result from the destruction of a ridge by stream and glacier action. It seems fairly certain that on the east side of the area an enormous fault with a downthrow to the east or an upthrow to the west follows the line

of the Waiau River and the east side of Lake Te Anau, and may be prolonged down the Hollyford River. If there is anything in the idea that the general orientation of the valleys depends on fracture-lines, it may have been associated with the movement of the earth's crust. The general even contour of the coast-line on the west and south of the area is reminiscent of southern Norway, and may be the result of fracture, and the settling of an area in the ocean-floor. This has been referred to previously by the present author in a short paper on a “Hornblende-andesite from the Solanders.”* Trans. N.Z. Inst., vol. xli, p. 52. The other explanation of the orientation is that the valleys were originally consequent on the shape of a former land-surface. This would result if the land to the west of Lake Te Anau had been raised in the form of an oval ridge with its chief axis running north-east and south-west. The first drainage-lines established would then be north-west and south-east, across the general strike of the beds, which have a general north-westerly dip. Subsequent streams would take a line approximately at right angles, or parallel to the strike, and, by the processes of absorbing neighbours and extension of their basins, might become the most important drainage-directions. This may, perhaps, be a more satisfactory explanation than that depending on fracture, and it is possible, of course, that both causes may have been present. But if the valley-directions depend on fracturelines they must date from a time anterior to the glaciation, and cannot be intimately connected with movements that are subsequent to it. The most interesting physiographical questions are those connected with the results of glacier erosion. That glaciers do erode their beds deeply seems almost certain when the landscape-forms of the region are considered, and the present author agrees with Sir James Hector when he attributes most of the peculiar features to the overdeepening of the main valleys by powerful glacier erosion, as compared with the feebler erosion that went on in tributary valleys when glaciers were smaller. This will explain the hanging valleys, the waterfalls, the truncated and semi-truncated spurs, the alignment of the valley-walls, and the U-shaped cross-section and the longitudinal section of the Sounds, as well as the rock-bound lakebasins—i.e., the pot-holes of a glacier stream, quite analogous to the potholes of a river. These points have been fully dealt with by Andrews (1, 2). The formation of lake-basins by scouring action analogous to the potholes of a river must result in some cases. It seems hard to explain the formation of the small rock-bound lakes at the head of George Sound on any other assumption. A striking illustration of the efficacy of this action is furnished by the Rakaia River, in Canterbury. A great glacier flowed down the valley, and part turned at right angles over a low saddle in the direction of Lake Heron. At the point of turning is a well-marked truncated spur with semi-detached knob, and below it a basin nearly half a mile across, smoothed and rounded on the inside, the direction of scouring being clearly circular and horizontal. This was evidently a whirlpool in the ice stream, formed at just such a place as that at which a whirlpool would form in a river, and no doubt formed in an analogous way. This is a peculiar case, which I shall deal with more fully in a subsequent paper, but it illustrates the power of erosion possessed by glacier-ice. It seems likely that many small rock-bound lakes in glaciated country, especially those at lower levels, may arise in this way.

Judging from the shape of the floors of the Sounds, from their constricted and shallow entrances, with, in certain cases, shallower water in the open sea, it seems certain that glaciers have enormous power of excavation. It is possible that more exact soundings may show a continuance of some of the valleys seaward across the submarine shelf. Depths exceeding 300 fathoms are found immediately off the entrance to some of the Sounds on a sea-floor which averages less than 100 fathoms deep. It seems hardly necessary, however, to postulate enormous thicknesses of ice to perform the work. Because ice-action is seen on the sides of an ice-eroded valley high up above its floor, it no more proves that ice filled the valley up to that level than high river-terraces prove that the river formerly filled the valley from its present channel up to those levels. Ice has power to produce rockcut terraces which are analogous to stream-cut terraces of primary excavation (I have seen instances in the glaciers of the Southern Alps), and many of the signs of glacier-ice high above present valley-floors are relics of the action of the ice streams when they ran at a higher level. The evidence for the existence of glaciers five and six thousand feet thick sometimes demanded in such cases as this seems to me to be very weak, and such enormous thicknesses are quite unnecessary in order to produce the results achieved. There are one or two land-forms, however, which result from glacier erosion of a former well-developed valley which I do not remember to have seen noticed previously. The first is the special shape of the ends of truncated spurs. These are almost invariably cut off, and reduced to a facet. The uniformity and regularity of this shape is striking. This can be seen very well indeed in the broader Sounds, where the pre-glacial valley-sides were not so steep, were more regular, and where the planing action of the ice was not so intense. The phenomenon is almost absent in Milford. I have seen it well marked in some of the valleys of Canterbury, where the pre-glacial drainage system was more mature. The spur sometimes shows that truncation was repeated, and that, after one strip had been planed off, another one was begun, and a shelf or terrace left half-way up the facet. These faceted spurs seem to be stable landscape forms, and persist when other signs of glaciation are disappearing. They do not receive any drainage except that which falls directly on their surface, and from their even nature it is long before streams establish themselves on them. When streams do, they are usually straight, and take the most direct course downhill, and for this reason they are numerous and small. All this tends to perpetuate the original form of the facet. These truncated spurs are formed in a moderately mature drainage system where the lateral spurs are not strongly developed, and the axis of the valley is tolerably straight. When the direction of the valley is subject to marked variations, and turns through an angle approaching a right angle, then the ice overrides the spur. If the angle turned is greater than a right angle, the action is still more marked, and crouching-lion forms ensue. One feature resulting from this action is that of a rounded hill lying off the end of a spur. This has been moulded by the erosive action, which is more marked near the valley-wall, so that a deep notch is sometimes formed behind the knob. Some of the isolated rounded hills and rocks found in glaciated countries may arise in this way, but many must result from the destruction of longitudinal spurs dividing adjacent subparallel valleys. Transitional forms in all states of development can be seen in the mountain district of Canterbury.

Petrological Notes. The remainder of this paper consists of notes on the petrology of the region. These tend to show that the rocks are in all probability not truly Archæan, but are metamorphosed igneous rocks, with perhaps occasional metamorphosed sedimentaries. This conclusion is based on the following evidence:— (1.) The rocks never appear to exhibit that profound metamorphism which characterizes most Archæan rocks. (2.) Frequently the only sign of metamorphism is the presence of strain and cataclastic effects: gradations in these can be traced from rocks practically without them to those which exhibit them to a marked degree. (3.) The rocks do not show the effects of heat on their mineralogical character as they would if they were truly Archæan and had experienced subcrustal changes. They belong to the upper or middle zone of Grubenmann. (4.) Rutile is a prominent constituent of the rocks, and this suggests that some may be altered sedimentaries. Mica-schists, however, are very seldom met with. (5.) The marbles occurring at Milford and elsewhere seem to be the result of alterations in igneous rocks (see Marshall), and are not due to change in detrital or organic limestone. It will be seen that the rocks are generally of somewhat acid character, and consist of true gneisses and diorite gneisses, with oligoclase as the dominant feldspar. Quartz and epidote are prominent constituents, titanite and rutile minor ones, the two latter being occasionally derived from ilmenite. The ferro-magnesian minerals are chiefly hornblende and biotite, although hypersthene and augite occur freely at times; garnets are very common. Amphibolites are of frequent occurrence, and perhaps are basic segregations from the hornblende gneisses, which may represent metamorphosed diabases. In the northern part of the area peridotites and allied and associated rocks are strongly developed. Preservation Inlet. Specimens were collected at Kisbee Bay, on the south-east side of the inlet; some from boulders on the beach. Diorite Gneiss (K 1).—This is a distinctly foliated rock, grey in colour. Under the microscope it appears composed of tolerably even-grained elements, consisting of much plagioclase (andesine), a little quartz, greenishbrown hornblende, biotite, grains of ilmenite, occasional titanite and epidote, frequent needles of apatite, and a few zircons. Diorite (K 2).—A dark-grey rock formed chiefly of hornblende, showing cleavage-surfaces plainly. In section it shows much hornblende in large green crystals and plates exhibiting fibrous structure. The feldspar (oligoclase-andesine) is in smaller proportion; a small amount of quartz is present; biotite occurs in laths, epidote in grains and aggregates derived from hornblende. There appear to be intergrowths of hornblende with mica; ilmenite is common in grains. Granite.—Granites of various degrees of coarseness are met with, the coarser types with pink feldspar crystals; some of the finer rock types are pinkish and reddish as a whole. The coarse-grained ones contain both

orthoclase and plagioclase, both micas, interstitial quartz, and titanite in grains derived from ilmenite. In some of the specimens there is very little muscovite, and an occasional garnet and zircon; epidote occurs sparingly. The Preservation Inlet granites have been described by Hutton (10) as syenites. Chalky Inlet. The only rock from here is a gneissic granite from North Port. It is dark grey in colour, very indistinctly schistose, with biotite feldspar and quartz plainly visible. Under the microscope it shows occasional straineffects. The feldspars are of both kinds, containing much sericitic material and at times micrographically intergrown with quartz; more biotite is present than in Preservation Inlet granites, and the flakes are frequently bent; muscovite is an important constituent; apatite needles, small garnets, and zircons are also present. Dusky Sound. Collections were made at Pigeon Island, at Duck Cove on Resolution Island, and at Pickersgill Harbour on the south side of the Sound, where Captain Cook established in 1768 a temporary astronomical station. Gneiss (D 1).—From Pigeon Island. In the hand-specimen the rock is whitish, showing indistinct foliation; quartz, feldspar, and biotite, and some garnet, are plainly visible. Under the microscope it appears composed of much quartz in variable-sized grains, and feldspar (orthoclase) showing the effects of strain; plagioclase and microcline are both present; some of the feldspar exhibits schiller effects; there is much biotite, some epidote and hornblende, and sphene in idiomorphic crystals occasionally included in mica and hornblende; a little apatite and zircon are present; the structure is at times cataclastic. Gneiss (D 2).—From Pigeon Island. In the hand-specimen the rock shows marked schistose structure. It is composed of microcline, plagioclase, quartz, muscovite, and biotite; granophyric growths are common; the structure is markedly cataclastic. Garnet Pyroxene Gneiss (D 3).—From Duck Cove. This is the rock (G 24) described by Marshall (14). It shows abundant garnets, as well as a green F.M. mineral associated with a white feldspar. In section it is composed of plagioclase, a little orthoclase, a fair amount of quartz, and garnets in grains with borders of pale-green secondary monoclinic pyroxene derived from the garnet, which shows marked signs of corrosion. A good deal of hornblende also occurs, a small amount of biotite, and much strongly pleochroic secondary epidote; rutile in small grains is of frequent occurrence. Gneiss (D 4).—From Duck Cove. A schistose rock showing epidote freely. It is composed of partially sericitised feldspar, partly plagioclase, a small amount of quartz, greenish hornblende, and strongly pleochroic secondary epidote (much of the epidote so frequently met with in the area appears to be primary). Ilmenite occurs in grains, with much secondary sphene in grains and aggregates derived from it, as well as apparently primary sphene in idiomorphic crystals; apatite needles are common. Amphibolite (D 15).—From Duck Cove. A dark rock formed chiefly of hornblende in rude parallel arrangement. In section it is composed almost wholly of green-brown hornblende, much rutile in grains and idio-

morphic needles, occasional grains of sericitised feldspar packed in between the hornblende; a small amount of yellowish epidote also occurs. Gneiss (D 16).—From Duck Cove. A markedly schistose rock, showing the effects of pressure plainly in the hand-specimen. The structure is completely cataclastic, the rock being composed of grains of quartz, microcline, and feldspar showing strain-effects; brown mica occurs in well-marked parallel arrangement. Gneiss (D 17).—From Duck Cove. This is the same as Marshall's G 13. It is composed of much quartz in irregular grains, plagioclase and microcline, biotite and muscovite in about equal proportions, the latter sometimes crowded with needles of sillimanite; the structure is cataclastic. Gneiss (D 18).—From Duck Cove. A greenish-white rock completely crushed to a fine aggregate of feldspar and quartz, with a few larger partially sericitised crystals of feldspar; there are a few larger quartz grains still left, and numerous flakes of muscovite in parallel arrangement. Mica-schist (D 5).—From Pickersgill Harbour. This is a dark compact rock, not showing schistose structure in the hand-specimen. In section it appears as a fine-grained rock composed of brown mica in small irregular flakes and small grains of quartz, with muscovite laths in rough parallel arrangement; small black grains of iron-ores in aggregates and strings are extremely common. This rock closely resembles a mica hornfels, and is probably due to the effect of a granite intrusion. Gneiss (D 6).—From Pickersgill Harbour. The rock exhibits wellmarked schistose structure. It is composed of much quartz; feldspar, both twinned and untwinned; occasional microcline; dark and light mica, sometimes intergrown; aggregates of epidote grains; ilmenite; apatite in short needles. Granophyric structure is common, and the whole rock shows the effect of pressure. Daggs Sound. The only type of rock found on this occasion in Daggs Sound was a gneiss (G 1) exhibiting marked schistose structure, white in colour, but speckled with greenish hornblendes. Under the microscope it proves to be a diorite gneiss. It contains the following minerals: Much oligoclase, frequently showing strain-effects, hornblende in corroded crystals, green in colour, and altered to aggregates of epidote and chlorite. Epidote also occurs along the cleavage-planes of the feldspar and as numerous grains and aggregates. Ilmenite occurs freely as grains, and at times forms irregular granular masses. Occasional mica and interstitial quartz also occur. The structure is moderately cataclastic. Thompson Sound. Collections were made at Wood Head, at the head of Gaer Arm, and at Deas Cove. Diorite Gneiss (G 2).—From the head of Gaer Arm. In the hand-specimen this is a yellowish-white rock with speckles of dark-green hornblende, indistinctly foliated. Under the microscope it exhibits an even-grained texture, with occasional strain-effects. The feldspar is an albite-oligoclase, often micrographically intergrown with quartz; quartz is present in quantity, sometimes intergrown with hornblende; hornblende, greenish-brown in colour, is a very prominent constituent; biotite occurs in broad plates,

hypersthene in grains strongly pleochroic in pink and green, and a little augite. Epidote Gneiss (G 3).—From head of Gaer Arm. This is a rock composed of feldspar frequently in foliæ, with dark foliæ of hornblende and mica, and very prominent yellowish-green epidote grains. In section, epidote in large grains and needles appears to be the most prominent constituent, greenish hornblende and brown mica, with much sphene, in grains and idiomorphic crystals. There is much feldspar (oligoclase) and some quartz, but some of the apparent quartz is clear colourless feldspar; apatite occurs in short needles. Feldspar Quartz Rock (G 4).—From the head of Gaer Arm. This is a white rock, which appears in section to be composed of plates of oligoclase, large grains of quartz exhibiting undulose extinction, and small amounts of muscovite and epidote. Mica-schist (G 5).—From Wood Head, at the junction of Doubtful and Thompson Sounds. In the hand-specimen this appears as a dark-grey schistose rock with lustrous mica surfaces, not very distinctly foliated. Under the microscope it shows well-marked parallel arrangement of the elements, which are almost wholly quartz, with brown and white mica; a little turbid feldspar occurs, and grains of ilmenite and hæmatite. This rock is in all probability an altered sediment. Diorite Gneiss (G 5).—From Wood Head. A rock composed of feldspar with speckles of dark-coloured constituents, schistose in structure. The microscope shows the presence of feldspar frequently clear and colourless and resembling quartz, a small amount of quartz being undoubtedly present; the other constituents are greenish mica, much greenish hornblende, epidote in grains, ilmenite passing into leucoxene, and anhedral rutile in sufficiently large grains to show a uniaxial interference figure; apatite needles also occur. Hornblende Schist (G 8).—From Deas Cove. A dark-coloured rock showing in section that four-fifths of the rock is yellowish-green hornblende slightly pleochroic; some feldspar occurs, and a little quartz in grains, but the amount is quite subordinate; grains of rutile are very common. This rock is evidently the same as Marshall's G26. Gneiss (G 10).—From Deas Cove. This is a white rock composed of quartz in rounded grains, much microcline, oligoclase, numerous shreds of muscovite, and very occasional garnets. Areas exhibiting granophyric intergrowths occur, strain-effects are common, and morter structure is well shown in places. Another specimen (G 9) from the same locality is very similar, but contains biotite and a little muscovite and less microcline. This appears to be like Marshall's A 9. George Sound. The specimens were collected at the head of the Sound, near the track to Lake Te Anau. Gneiss (G 6).—A light-grey rock, not distinctly foliated. In section it is composed of clear feldspar (oligoclase), with some untwinned feldspar probably orthoclase; much quartz; much epidote in large grains: both muscovite and biotite occur, and sphene in grains. The structure is slightly cataclastic.

Gneiss (G 11).—A white rock with a few large flakes of biotite. The rock is almost all feldspar (oligoclase), some quartz, and muscovite in ragged flakes. The feldspar has schiller inclusions regularly arranged parallel to basal and brachypinacoids, and contains irregularly disposed areas of microcline. Diorite Gneiss (G 18).—A dark-coloured, rather fine-grained, distinctly foliated rock. Under the microscope the rock appears even-grained, more than half composed of hornblende and biotite, with both, especially the mica, in rudely parallel arrangement. There are colourless grains of epidote, and sphene in numerous grains. The quartz and plagioclase are in about equal proportions. Diorite Gneiss (G 7).—Similar to G 18, but more basic. This is a schistose rock with distinct foliæ of feldspar and dark minerals. The feldspar is oligoclase, accompanied by a little quartz. There is much hornblende and biotite, the former more important; occasional epidote; and numerous grains of sphene, both dark and pale coloured. Bligh Sound. These rocks were collected on the north side of the river entering the sound. Hypersthene Diallage Gneiss (B 1).—A pinkish-grey rock, distinctly foliated. The feldspar is clear; there are grains of quartz, hypersthene with characteristic pleochroism and straight extinction, also diallage, and a little biotite and ilmenite. Dr. Marshall has pointed out to me the resemblance between this rock and a mica norite from the Darran Mountains, Milford Sound (14). Both rocks contain the same minerals, but on comparing slides his specimen appears more basic, and is without the gneissic structure so markedly exhibited by the rock from Bligh Sound. This occurrence may, however, be a gneissic variety of the mica norite. Diorite Gneiss (B 2).—An indistinctly foliated rock. The feldspar is oligoclase in broad plates with faulted and bent lamellæ, at times crowded with inclusions of epidote needles and crystals, similar to the rock from the Bowen Falls mentioned by Marshall. There are corroded crystals of hornblende, sometimes with schiller-like inclusions parallel to the base, and filled with grains of quartz (quartz de corrosion); a small amount of other quartz also present, and numerous large individuals of epidote and flakes of biotite. There are grains of ilmenite surrounded by aggregates of sphene which are evidently formed from it. Hypersthene Amphibolite (B 3).—This is a dark-green rock, which appears under the microscope to be composed principally of greenish-brown strongly pleochroic hornblende, with schiller-like inclusions; some of this may be altered augite. There is a considerable quantity of hypersthene, a little olivine and augite. A small amount of oligoclase occurs, and much rutile in grains. The rock was not found in situ, and it may be a basic variety of a diorite found in the locality, but it shows no foliation or parallel arrangement of the elements. The rock is related to a cortlandtite; but there is little olivine present, and this is not in pœcilitic grains as in a typical cortlandtite, but wedged in between the hornblende elements. Diorite (B 4).—A rock formed of equal parts of feldspar and dark constituents, not distinctly foliated. In section it is composed of oligoclase, a little quartz, much epidote, a fair amount of greenish-brown hornblende, some biotite, and grains of sphene. It is an even-grained directionless rock, with little, if any, sign of pressure.

Mckinnon Pass. Two rocks from this locality, a wehrlite and a gneiss, have been recorded by Marshall. The present specimens, with one exception (the first enumerated), were collected on the western side of the pass. Hornblende Schist (C 1).—From the Clinton Valley, near the Pompelona Huts. In the hand-specimen this is a dark-greenish rock, with cleavage-faces of hornblende plainly visible. Under the microscope more than half of the rock is of brownish-green and bluish-green hornblende, with numerous inclusions of rutile and ilmenite. The feldspar is oligoclase, and some quartz is present, as well as a few shreds of biotite and muscovite. Gneiss (M 1).—This is a distinctly foliated whitish rock, with quartz and muscovite clearly visible to the eye. In section it is composed of much quartz, plagioclase, and microcline, biotite and muscovite, epidote in grains, and sphene. The elements have a rudely parallel arrangement, strain-effects occur, and the structure is partly cataclastic. Diorite Gneiss (M 2, and M 4–8, –15).—These specimens, all from the western side of the top of the pass, include a number of distinctly and indistinctly foliated rocks, showing hornblende freely, and occasional garnets. In section they are composed of brownish-green and blue-green hornblende; biotite in broad plates quite subordinate to the hornblende; occasional muscovite; and much epidote, both colourless and yellow, in crystals, grains, and granular aggregates. The feldspar is oligoclase, usually clear, with quartz in varying amount, but sometimes an important constituent. Rutile is plentiful as brown grains and needles, titanite is also present at times, and an occasional garnet. Cataclastic structure is frequent. Hornblende Epidote Schist (M 3).—A dark-grey indistinctly foliated rock, composed of greenish-blue and greenish-brown hornblende, biotite subordinate in amount to the hornblende, muscovite, a little feldspar, much quartz, and a large amount of epidote; rutile grains are common. This rock is apparently a variety of the diorite gneiss in which feldspar is practically absent. Amphibolite (M 9).—A dark rock composed chiefly of hornblende. In section it is composed of the usual hornblende, greenish-brown and -blue in colour, forming seven-eighths of the rock; occasional muscovite; much epidote; and brown grains and needles of rutile. Gabbro (M 10).—A dark-green rock, distinctly schistose, with greasy feel. In section it appears markedly cataclastic, and composed of greenish hornblende much crushed and fibrous, some apparently derived from the pyroxene. There occurs also greenish faintly pleochroic diallage; enstatite; with serpentine and talc, which may have been derived from the olivine, but undoubted evidence of this is lacking. Granular sphene is common, and a considerable amount of feldspar, untwinned, and probably of secondary origin, being quite clear, but crowded with needle-like crystals of pale-green colour and with oblique extinction, in all probability actinolite. A part of this clear mineral may be secondary quartz. Unaltered feldspar occurs in small amount, the grains being completely crushed. The rock appears to be a crushed gabbro or wehrlite. Anita Bay (Milford Sound). Captain Hutton (11) has described hornblende diorites and enstatite gabbros from Milford Sound, specially from Harrison's Cove, and Dr.

Marshall (14) notes, with descriptions, gneisses and hornblende schists from Milford, but he has devoted special attention to the basic rocks at Anita Bay, on the southern side of the entrance to the Sound. He has recorded from there hornblende schists and gneisses, bowenite, dunite, and harzburgite, the last-named passing into marble. Finlayson (3) has also referred to the bowenite, and the matrix in which it occurs. This locality, owing to its remoteness and the thick covering of bush, is very imperfectly known, and the present author was able to collect a number of other rocks mostly from boulders which can only have come from the adjacent hillsides; but the marble passing into harzburgite was found in position as veins in a dark indistinctly foliated rock at the western point of the bay. The following types were met with, and will serve to emphasize the remarkable variety of rocks occurring there, and the difficulty of fully establishing their true relationships. Hornblende Garnet Gneiss (An. 1).—This is a markedly schistose rock, with many garnets visible to the eye. Under the microscope it is composed of large irregular reddish garnets, with quartz (probably quartz de corrosion) and rutile inclusions. There is a considerable amount of greenish hornblende, and some flakes of brown mica. Much quartz occurs in rounded grains with undulose extinction, and feldspar (andesine) frequently exhibiting wavy and broken twin lamellæ. Zoisite occurs, very occasional epidote and muscovite, and much brown rutile in grains included in the garnets and mica. Hornblende Garnet Gneiss (An. 9).—A dark, even, fine-grained rock, showing in section much garnet in grains, and the minerals of the preceding rock, but the texture is much finer-grained and shows the effects of pressure. The biotite flakes are bent round the larger elements in pseudofluxion structure. Garnet Gneiss (An. 10).—This is a pink-grey distinctly schistose rock. It contains much garnet in grains, with inclusions of quartz and feldspar sometimes closely resembling graphic intergrowths; the garnet is cracked, and has both muscovite and biotite freely developed in the cracks. Flakes of both these micas occur, and much quartz and feldspar twinned and untwinned; apatite needles are occasionally to be seen. Pressure granulation is very prominent. Diorite Gneiss (An. 14).—A rock showing much dark-green hornblende and feldspar, distinctly foliated. It contains much greenish and bluishgreen hornblende, passing into chlorite and epidote; plagioclase, and some quartz. Another variety (An. 15) of this gneiss contains long needles of epidote, and flakes of muscovite included in the feldspar. Hornblende Garnet Gneiss (An. 11).—This is a distinctly foliated rock, composed of much hornblende frequently intergrown with garnet, garnet in grains, some faintly pleochroic epidote, columnar aggregates of zoisite, a little brown mica and muscovite, quartz and much plagioclase, brwon grains of rutile. Dunite (An. 5).—This is similar to Marshall's rock, but it contains occasional enstatite as well as diopside. The structure is markedly cataclastic. Harzburgite (An. 3, 4, 6, 7).—Some of my specimens are similar to those described by Marshall (13), but others contain much augite with occasional diallage structure. The olivine is partially serpentinised, and apparently passes into talc, as flakes and aggregates of that mineral occur frequently, and specially as inclusions in the carbonate present in the rock.

This is well shown in the specimens from veins on the west side of the bay, where the rock approaches a sagvandite (Rosenbusch's “Elemente der Gesteinslehre,” p. 549). Large quantities of carbonate are found in the rocks containing much diallage, as well as in the normal harzburgite, showing that the enstatite in other ultrabasic rocks passes into carbonate; unless, indeed, the process is a reverse one, and this marble is passing into a magnesian silicate rock owing to the effect of an intrusive mass of dunite. However, in some cases here the alteration enstatite to talc certainly occurs, as crystals occasionally show a change in that direction. Besides the above, large boulders of rocks of somewhat abnormal type were found. They cannot have come from a distance, and they show connections with the rocks found in position. They are distinctly foliated, and contain a large amount of garnet. The first is a greenish-grey rock with a tinge of pink, close-grained, showing faint schistose structure: it is composed principally of reddish garnet, with augite, olivine, titanite, and chromite, some quartz of secondary origin, and much carbonate (calcite or dolomite); the garnet grains are in strings and aggregates, often in rudely parallel arrangement, and show marked signs of crushing;- the whole structure is cataclastic. The second specimen is a green rock, with cleavage-surfaces of green augite strongly showing: it is composed of greenish augite occasionally showing diallage structure, grains of olivine which are clear and only occasionally passing into serpentine, occasional enstatite, and much crushed garnet of slightly reddish tint in grains distributed irregularly but sometimes collected in masses; the structure is markedly cataclastic. The first of these rocks is a eulysite or garnet peridotite, in connection with which it may be observed that Rosenbusch, in his work referred to previously, has noted the association of carbonates with rocks of this type. The second rock contains much less garnet, and is more closely related to the dunite. As it contains a considerable amount of monoclinic pyroxene it should perhaps be called a garnet-wehrlite. Both of these rocks seem to be modifications of the ultrabasic intrusions which have penetrated the gneisses of this area, and belong to that great series of intrusions which occur up the western side of the Southern Alps to Nelson and reappear in New Caledonia. The date of the New Zealand intrusions was first of all considered as Devonian, but the consensus of opinion at the present would make them much younger, probably Upper Jurassic or Lower Cretaceous; however, according to the reports of Messrs. Pelatan and Piroutet, the serpentines and allied rocks of that island, which is so closely connected geologically with New Zealand, are certainly post-Cretaceous, and there is no reason why the New Zealand rocks of like facies should not be of the same date. An inference drawn from two widely separated localities is certainly somewhat dangerous; but the almost entire absence of fragments of ultra-basic rocks in detrital deposists in this country is certainly suggestive, although explanations of their absence can be readily put forward, and of course they may be discovered after a further examination of the deposits. The date of the intrusion of the New Zealand peridotites may be as recent as Tertiary times and possibly late Tertiary times. In concluding, the author wishes to acknowledge the imperfect and unsatisfactory character of these petrological notes; they will serve, however, to mark the complex character of the area, and to emphasize Captain Hutton's conclusion that the rocks are not truly Archæan, but metamorphosed igneous, with perhaps metamorphosed sedimentary rocks included among them

Bibliography. 1. Andrews, E. C. “Some Interesting Facts concerning the Glaciation of South-western New Zealand.” Report Aust. Assn. Adv. Sci., vol. x, 1904. 2. Andrews, E. C. “New Zealand Sound Basins.” Journal of Geology, vol. xiv, 1906. 3. Finlayson, A. M. “On the Nephrite and Magnesian Rocks of New Zealand.” Quart. Jour. Geol. Soc., vol. lxv, 1909. 4. Hector, Sir James. “Geological Expedition to the West Coast Sounds of Otago.” Otago Provincial Gazette, 1863. 5. Hector, Sir James. “On the Geology of the Province of Otago.” Geol. Mag., vol. i, 1864. 6. Hector, Sir James. “On the Geology of Otago, New Zealand.” Quart. Jour. Geol. Soc., vol. xxi, 1865. 7. Hector, Sir James. “Origin of Rock Basins in New Zealand.” Geol. Mag., vol. ii, 1865. 8. Hutton and Ulrich. “Geology of Otago.” 1875. 9. Hutton, F. W. “On a Hornblende Biotite Rock from Dusky Sound.” Quart. Jour. Geol. Soc., vol. xliv, 1888. 10. Hutton, F. W. “The Eruptive Rocks of New Zealand.” Jour. Roy. Soc. N.S.W., vol. xxiii, 1889. 11. Hutton, F. W. “On the Foliated Rocks of Otago.” Trans. N.Z. Inst., vol. xxiv, 1892. 12. Hutton, F. W. “Corrections of the Names of some New Zealand Rocks.” Trans. N.Z. Inst., vol. xxxi, 1899. 13. Marshall, P. “Magnesian Rocks at Milford Sound.” Trans. N.Z. Inst., vol. xxxvii, 1905. 14. Marshall, P. “Geological Notes on the South-west of Otago.” Trans. N.Z. Inst., vol. xxxix, 1907.