Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 65, 1936
This text is also available in PDF
(2 MB) Opens in new window
– 187 –

Geological Investigation of the Nephrites, Serpentines, and Related “Greenstones” used by the Maoris of Otago and South Canterbury.

[Read before the Otago Institute, October, 1932; received by the Editor, August 4, 1934; issued separately, October, 1935.]

  • Introduction.

  • Nephrites and Related Rocks.

    • General Petrography.

    • Detailed Petrography.

    • Origin.

  • Serpentines.

    • Petrography.

    • Origin.

  • Talc-Rocks.

  • Macroscopic Properties of Greenstones.

    • Colour and General Appearance.

    • Hardness.

    • Fissility and Toughness.

    • Summary.

  • Relative Abundance of Varieties.

  • Possible Sources of Greenstones.

  • Acknowledgments.

  • List of Specimens Examined.

  • Literature Cited.


In 1909, A. M. Finlayson (1909) published his classic account of the peridotites and nephrites of New Zealand, in which was given an account of the chemical, mineralogical, and petrographic characters of New Zealand “greenstone,” its occurrence and mode of origin. The present investigation was undertaken by the writer with the object of obtaining further information regarding the greenstones used by the Maoris of Otago and South Canterbury, with special reference to the following points:—

  • (a) Petrographic and mineralogical nature of different varieties.

  • (b) Geological conditions governing their genesis.

  • (c) Macroscopic distinction between different varieties.

  • (d) Sources from which the Maoris obtained their supplies of greenstone.

One hundred and twenty micro-sections of representative greenstones, mostly obtained from Maori camp-sites in Otago and South Canterbury, have been examined, and the properties of the corresponding hand-specimens have been recorded. The purely geological evidence so obtained has then been applied to the more general problem of the geographic distribution of greenstones in the South Island of New Zealand. The conclusions reached are admittedly

– 188 –

incomplete and may require modification when further material is available and additional field work has been carried out in other areas.

The term greenstone is used throughout this paper, in the sense employed by Chapman (1892, pp. 480–481), to include the finegrained rocks consisting essentially of either tremolitic amphibole or serpentine, which were used by the Maoris in the manufacture of implements and ornaments. Though such rocks are typically green in colour, they vary from greenish-white on the one hand to almost black on the other.

Petrographically the greenstones of Otago and South Canterbury fall into two main divisions: nephrites and related rocks, the essential constituent of which is tremolite H2Ca2Mg5(SiO3)8 or actinolite H2Ca2(Mg, Fe)5(SiO3)8, and serpentines composed mainly of a mineral of the serpentine group, H4Mg3Si2O9. These two classes are connected with each other by rather rare specimens in which both serpentine and tremolite are prominently developed. Occasionally talc, H2Mg3(SiO3)4, is relatively plentiful, and, as the proportion of this mineral increases, there is perfect transition towards rocks composed principally of talc, with only minor serpentine or tremolite. These talc-rocks, though not strictly to be regarded as greenstones, are nevertheless included here as a subordinate third group, since they are represented by occasional specimens among the worked Maori material collected from sites in Otago.

Nephrites and Related Rocks.

General Petrography.

In spite of the great diversity observed in their macroscopic properties, the nephrites and tremolite-rocks are mineralogically alike in that all are composed essentially of an amphibole of the tremolite-actinolite series. As seen beneath the microscope, however, the structural features of these rocks are by no means constant, and show considerable and complex variations between four pairs of extremes:—

  • (a) In any representative series of specimens there is a complete range from minute hair-like fibres between 0.01 mm. and 0.05 mm. in length, to relatively coarse prismatic crystals with well developed cross-fracture and sometimes averaging 1 mm. × 0.1 mm.

  • (b) In almost every specimen the fibres or prisms are aggregated into bundles or tufts in which the arrangement of the individual crystals may vary considerably. In some rocks the fibres comprising a bundle are only sub-parallel and may be oriented obliquely to the direction of elongation of the aggregate itself, while in other cases the individual fibres are parallel both to one another and to the direction of elongation of the bundle.

  • (c) The tufts of fibres may be mutually parallel throughout the rock (to which they then impart a linear schistosity and nematoblastic structure), or on the other hand they may be

– 189 –
  • quite unoriented, in which case schistosity is not developed. Many rocks consist partly of unoriented and partly of parallel tufts of fibres.

  • (d) In typical nephrites the bundles of fibres are twisted and thoroughly felted or interwoven with one another, producing a highly characteristic structure to which the term “nephritic” may be applied. In other specimens the felting is imperfectly developed or even completely absent.

In addition to the essential constituent, the following minerals have been observed as accessories in the nephrites and related rocks examined by the writer: chlorite, serpentine, sphene, diopside, talc, garnet, a carbonate mineral (probably dolomite or magnesite), quartz, chromite, picotite, and hornblende. Of these, chlorite, serpentine, and sphene are fairly frequently met with, while either of the first two minerals may occasionally be sufficiently abundant to rank as an essential constituent of the rock. It is probable that hornblende, chromite, and picotite, whenever present, represent remnants of the original constituents of the parent rocks from which the nephrites have been derived by metamorphism. The remaining minerals appear to constitute, together with tremolite, a paragenetic metamorphic assemblage.

The classification given below is founded essentially upon microstructure. On this basis five divisions are recognised, though these are not sharply defined, transition types being by no means rare.

(a) Non-schistose Nephrites (1800 to 1811*). The non-schistose nephrites are characterised by uniformly small size and unoriented arrangement of the component tufts of fine tremolite fibres, and by perfectly developed nephritic structure, twisting and felting of the tufts being very marked. The tufts usually vary from 0.02 mm. to 0.3 mm. in length, reaching 0.5 mm. in rather rare instances. Occasionally well-defined stout prisms or acicular crystals (up to 1.5 mm. × 0.03 mm.) of coarse tremolite may be present in small amount (e.g., in 1806); these contrast sharply with the exceedingly slender tufted fibres of the surrounding nephritic base, and indicate transition towards the seminephrites. In other specimens (e.g., 1807) gradation towards the schistose nephrites is marked by the presence of minor mutually parallel aggregates composed of rather coarser fibres than the average. Accessory constituents are confined to two sections: 1804, chromite and chrome-diopside; 1809, chlorite.

(b) Schistose Nephrites (Za 5, 1370, 1812 to 1817, 1822). The distinctive feature of the schistose nephrites is the presence of plentiful parallel tufts of tremolite fibres set in a base composed of finely felted unoriented tufts which in some cases make up only a small percentage of the total composition. This subdivision of typical nephrites into non-schistose and schistose types corresponds with the observations of Finlayson (1909, pp. 367–368), who distinguished between “fissile” and “horny” nephrites and noted the main distinction between the microstructures typical of these two varieties.

[Footnote] * Numbers refer to sections and specimens in the collections of the Geology Department, University of Otago.

[Footnote] † For description of 1370, see Turner, 1933, p. 272.

– 190 –

Though the schistosity is due to parallelism displayed by many of the fibrous aggregates, this regularity of orientation does not necessarily extend to the component fibres within any particular tuft. The individual fibres of a tuft are usually considerably interwoven and only subparallel with one another, while their average orientation is frequently inclined to that of the aggregate itself, i.e., to the direction of schistosity. The length of the oriented aggregates is usually about 0.3 mm. to 0.5 mm., while the felted tufts which make up the base are generally noticeably smaller than this. Rarely (e.g., 1822) a small amount of coarse prismatic tremolite (not more than 5% of the total composition) suggests a transition towards the seminephrites. Accessory minerals were observed in only two sections: 1815, chlorite and magnetite; 1816, colourless diopside and chlorite.

(c) Seminephrites. This group is represented by about twenty specimens (1818–1821, 1823–1837, 1442). While consisting partly of fine-grained interfelted fibrous tufts of tremolite with true nephritic structure, the seminephrites also contain abundant relatively coarsely crystalline tremolite which takes the form of well-defined acicular prisms or relatively large sheaves of parallel unfelted fibres. This coarse material may or may not show approximate parallelism throughout a section. The group of seminephrites as here defined includes also a number of transition rocks (1818–1821) in which the coarse component is developed to only a limited extent, and which therefore grade towards the true nephrites described above. Again, as the interstitial fine-grained nephritic base decreases in amount, these rocks pass in the opposite direction into the group of tremolite-rocks in which truly nephritic material is almost or completely absent. In almost every section examined, the coarse well-defined prismatic crystals are seen to fray out directly into the enclosing felted fibrous base, which itself has obviously originated by shearing and cataclasis of initially coarser crystals. This is especially well shown in 1829 and 1824. The following dimensions are typical: fibrous tufts of the nephritic groundmass, 0.05 mm. to 0.3 mm. in length; sheaf-like aggregates of long fibres and prismatic crystals, 4 mm. × 2 mm. (e.g., 1829); individual coarse acicular crystals, 1 mm. × 0.05 mm. to 3 mm. × 0.25 mm. (1823).

Accessory minerals are much more commonly present in the seminephrites than in the true nephrites, chlorite being present in six out of the twenty specimens examined; in 1825 it makes up between 10% and 15% of the total composition of the rock. Other minor constituents include serpentine (1828), chromite (1831, 1833), picotite (1826), chrome-garnet (1826, 1833, 1834), and diopside (1826).

(d) Tremolite-rocks with Linear Schistosity. Those rocks in which fine-grained truly nephritic fibrous tufts are absent or present in only minor amount are classed as tremolite-rocks. Many of these (1838–1847) are composed essentially of coarsely-fibrous or prismatic tremolite, the crystals of which show marked parallelism imparting nematoblastic texture and distinct linear schistosity to the specimen. There is gradual transition through types such as 1841 and 1845 in which parallelism of the component crystals is less marked than usual, to the second group of tremolite-rocks in which only plane schistosity

– 191 –

is developed. Accessory constituents are present in nine out of ten specimens; they include chlorite (1842–1845), opaque iron-ore (1838, 1843, 1845, 1847), serpentine (1840, 1843), sphene (1845, 1847), talc (1839, 1840), a rhombohedral carbonate (1838), and quartz (1838).

Typically the oriented crystals of tremolite are coarser than in the seminephrites (e.g., 5 mm. to 10 mm. long in 1840, and 2 mm. × 0.5 mm. in 1841), though this is not invariably the case.

(e) Tremolite-rocks with Plane Schistosity. The majority of the tremolite-rocks of this group are composed mainly of fibrous or prismatic crystals of tremolite, which in sections cut parallel to the schistosity lack regular orientation, but at the same time show no trace of the felted and twisted arrangement that gives rise to nephritic structure in other rocks. Coarse, distinct prisms of tremolite are usually plentiful, though a few specimens (e.g., 1852) appear to be composed entirely of relatively small fibres. The rough plane schistosity which is usually evident in hand-specimen is due to a marked tendency for the component crystals to lie with their directions of elongation in parallel planes. Accessory constituents are present in most specimens: 1379, diopside, serpentine, and sphene; 1425, hornblende and sphene; 1443, hornblende and serpentine; 1452, hornblende and sphene; 1849, pennine and chromegarnet; 1850, serpentine; 1851, chromite; 1852 and 1853, hornblende and sphene; 1364 and 1848, accessories absent.

Two of these specimens (1364 and 1379) were collected from boulders in the beds of streams draining from the northern end of the peridotite belt which outcrops along the north-western flank of the Olivine Range, South Westland,* while three others (1425, 1443, 1452) were obtained in situ from a small altered ultrabasic intrusion which outcrops on the northern bank of the Routeburn Stream about two and a-half miles above its junction with the Dart River, Lake Wakatipu district. The remaining six specimens are worked Maori material.

Detailed Petrography.

The minerals observed in the nephritic and related greenstones fall into two groups:—

  • (a) Minerals of metamorphic origin, including tremolite, chlorites, serpentine, talc, garnet, sphene, diopside, magnetite, a carbonate, and quartz.

  • (b) Relict constituents of the parent ultrabasic igneous rocks from which the nephrites, seminephrites, etc., have been derived by metamorphism. These include iron-ores (chromite, picotite, and opaque iron-ore), hornblende, and in one instance possibly aluminous pyroxene.

Tremolite. The tremolite is always colourless in sections less than 0.025 mm. in thickness. The extinction angle (Z to c) can be measured accurately only in the coarser prismatic crystals where it varies between 16° and 20°. The optic sign is always negative,

[Footnote] * For full petrographic descriptions see Turner, 1933, pp. 272, 273.

– 192 –

and thus precludes any possibility of the mineral being a cummingtonite. These properties agree with those recorded by Finlayson, who also demonstrated by chemical analyses that the chief constituent of the nephrites examined by him is tremolite or actinolite, the FeO content of which may in some cases reach 7%.

Chlorites. The properties of the chlorite are remarkably uniform. In almost every instance it is a colourless or pale-green faintly-pleochroic pennine with positive optic sign, negative elongation, small optic axial angle, and low birefringence. Yellowish-brown anomolous interference tints are characteristic, though the well-known violet-blue colour is occasionally seen between crossed nicols.

In several sections (e.g., 1826, 1833) radially-disposed flakes of pennine build up sharply-defined round patches 0.5 mm. to 2 mm. in diameter around central grains of chromite or picotite. In 1849 there are similar chloritic patches which however lack the spinellid nucleus, but enclose numerous minute grains and rhombic dodecahedra of pale green garnet, which are concentrated in narrow peripheral zones. In all cases these sharply-bounded areas of pure chlorite appear to have crystallised subsequently to the crystallisation and shearing of the enclosing tremolite mass. In a number of sections, however (e.g., 1818, 1824), the chlorite is developed as irregular lensoid patches stabbed through with needles and prisms of tremolite, or simply as isolated flakes in the mesh of amphibole fibres; here crystallisation of the two minerals appears to have been simultaneous.

Chlorite of a different type—probably clinochlore—occurs in a single section (1844), where it takes the form of sharply idioblastic, unoriented, almost colourless flakes, often showing lamellar twinning, scattered abundantly through a groundmass of parallel rather coarse fibres of tremolite. The latter mineral appears to have ceased crystallising prior to development of chlorite.

Serpentine. Serpentine, though not as frequently present as chlorite, was recorded in six specimens of seminephrites and tremolite-rocks. It is colourless, with negative optic sign and small axial angle, and appears to be identical with the principal constituent of the serpentine-rocks described in a later section. It may occur interstitially among the crystals of tremolite which make up the bulk of these rocks (e.g., 1379), or it may take the form of rather irregular streaks and patches of relatively pure fibrous serpentine enclosed in a matrix of tremolite (e.g., 1850).

Talc. Talc was noted in only two specimens (1839, 1840), both tremolite-rocks with linear schistosity. In 1839 it builds up sharply-bounded round patches 6 mm. to 7 mm. in diameter, which appear as bright emerald-green spots on the otherwise dull olive-green surface of the polished hand-specimen. These spots are composed entirely of unoriented relatively coarse tabular crystals (0.5 mm. × 0.15 mm.) with distinct pleochroism according to the scheme:—

  • X = pale bluish green.

  • Y = Z = pale yellowish green.

  • Y = Z > X.

– 193 –

The mineral is optically negative and distinctly biaxial, though the axial angle is small. Basal sections show undulose extinction. In this specimen the talc appears to have crystallised subsequently to crystallisation and shearing of the surrounding tremolite. In 1840 green talc having much the same properties as those recorded above occurs in small lensoid patches, intercalated among subparallel bundles of unusually coarse tremolite fibres averaging between 5 mm. and 10 mm. in length. This specimen is in fact an asbestos-rock representing the coarse-grained, non-felted extreme of the series of tremolite-rocks showing linear schistosity. The talcose patches may also contain either serpentine or tremolite in addition. In this rock there is no evidence to suggest that crystallisation of tremolite, serpentine, and talc was not simultaneous.

Garnet. Garnet occurs in very small amounts in four sections (1826, 1833, 1834, 1849). It always takes the form of minute pale green granules or rhombic dodecahedra (1849) enclosed in patches of pennine, and appears from its colour to be a member of the grossularite-uvarovite series. In two sections the garnetiferous patches of chlorite surround central grains of chromite or picotite, and in one of these (1826) there is marginal transition between the latter mineral and granular garnet. These facts are best explained on the supposition that after crystallisation and shearing of tremolite, there was subsequent local chloritisation of amphibole in the vicinity of relict grains of chromite or picotite which supplied the necessary Al2O3 and probably some Cr2O3 also. This reconstitution involves loss of lime, part of which appears in some cases to have entered into garnet. In 1849 the central grains of spinellid have been completely destroyed, but the garnetiferous chlorite patches with their characteristic structure still persist.

Sphene. Sphene is confined to the tremolite-rocks, and more especially to those which have only plane schistosity. It was noted in eight sections. The mineral is usually pale yellow in colour and occurs in the form of minute xenoblastic granules, which appear, in those specimens that contain residual hornblende, to have been formed during replacement of the latter mineral by tremolite.

Diopside. Colourless diopside is a relatively plentiful constituent in three sections, viz., 1816 (nephrite), 1826 (seminephrite), and 1379 (tremolite-rock), in each of which it makes up between 5% and 10% of the total composition. In contrast with the enclosing base of minute felted tremolite fibres, the crystals of diopside are large (0.2 mm. to 0.4 mm. long) sharply idioblastic and little, if at all, affected by shearing. They are clearly products of late crystallisation and appear to have been formed at the expense of the surrounding amphibole.

In a single specimen of fine-grained non-fissile nephrite (1804), green chromiferous diopside is present and is actually visible in the hand specimen as small bright emerald-green spots. The crystals are stout idioblastic prisms averaging 0.5 mm. × 0.1 mm., which typically show sharply-defined boundaries in the prism zone, but irregular often jagged terminations. The macroscopically-visible green spots are seen in section to consist of roughly radial aggregates

– 194 –

of pyroxene prisms built up round central nuclei of chromite. Theoptic sign is positive, the extinction angle (Z to c) is about 39°, and there is usually distinct pleochroism in pale shades of green and yellow. The depth of colour varies considerably even within a single crystal, and is most pronounced in crystals lying adjacent to the chromite grains in the compound aggregates referred to above. The pleochroism is rather unusual in that the colour-axes lying in the 010 plane are distinctly inclined to the X and Z vibration-directions; one axis (yellow absorption) is more or less parallel to c, while the other (green absorption) is but slightly inclined to the horizontal. The third colour-axis (green absorption) is parallel to b (Y). In the clinopinacoidal section there is faint pleochroism from pale yellow (slow) to pale green (fast); in the orthopinacoidal section, which is also nearly perpendicular to an optic axis, the pleochroic effect reaches a maximum, c being pale yellow and b (Y) being pale green; the section perpendicular to the second optic axis is almost non-pleochroic and pale green in colour. In this slide, as in those just described, the pyroxene has obviously crystallised subsequently to crystallisation and shearing of tremolite.

Magnetite. Small granules of opaque iron-ore, presumably magnetite, are present in very small quantities in several sections (e.g., 1815, 1838).

A Rhombohedral Carbonate and Quartz occur in mutual association as lensoid patches and streaks in a single section (1838) where they together constitute 10% of the total composition. The carbonate is much the more abundant mineral of the two, and appears from the lack of secondary polysynthetic twinning to be either dolomite or magnesite. The individual grains are elongated parallel to the schistosity, while the lensoid aggregates are frequently stabbed through by prisms of tremolite. Crystallisation of tremolite, carbonate, and quartz seems to have been approximately simultaneous.

Relict Iron-ores. Chromite is the commonest of the relict minerals observed, and is found in all types of nephritic greenstone examined by the writer (e.g., 1804, 1831, 1833, 1841, 1851). Picotite (1826) and rather coarse opaque magnetite (1843) are much rarer.

Relict Hornblende. The presence of residual hornblende is limited to the tremolite-rocks with plane schistosity, where it is a frequent accessory and in some instances (e.g., 1852, 1853) may even make up 10% of the total composition. The mineral may be either deep brown or green in colour, but in sections where both varieties are present (e.g., 1852) the green type appears as an intermediate product formed during replacement of brown hornblende by tremolite. The individual grains do not usually exceed 0.5 mm. in diameter, but optical continuity between adjacent grains often indicates that the parent crystals must have been of much larger size. All stages of transition from hornblende through coarsely prismatic to finely fibrous tremolite are perfectly illustrated in many of these sections, sphene often being a secondary product representing the titanium of the original amphibole.

– 195 –

Origin of the Nephritic Rocks.

As pointed out by Finlayson (1909, pp. 377–378), the formation of nephrite is a dual process involving in the first place the chemical production of crystals of tremolite, and in the second place the mechanical production of the characteristic felted microstructure. In the present discussion no attempt will be made to summarise the various hypotheses of origin which have been put forward; such summaries may be found in the accounts of Finlayson (1909, pp. 376–378) and Bauer (1914, pp. 700–704). With reference to the chemical side of the problem it may here be stated, however, that nephrites are believed by many writers to be essentially products of uralitization of pyroxenic rocks.

In his accounts of the nephrites of New Zealand, Finlayson suggests that these rocks have originated in four distinct ways, viz., uralitization of pyroxene, contact-action of peridotite intrusions upon limestone, replacement of olivine by tremolite, and deep-seated metamorphism of serpentine-talc-carbonate-rocks, intense shearing being an essential factor in every case. The evidence advanced by Finlayson in support of direct replacement of pyroxene and of olivine by nephritic tremolite is in accordance with modern views concernng the metamorphism of ultrabasic igneous rocks. The other two processes mentioned by Finlayson cannot, however, be regarded as equally well established.

Additional evidence as to the mode of origin of some of the nephrites and related rocks described in this paper may now be brought forward. In the district lying north and west of the head of Lake Wakatipu there is a series of ultrabasic rocks represented by peridotites, antigorite-serpentines, and talc-schists. In the southern portion of the area these take the form of small lensoid masses, invading the sheared greywackes, phyllites, and schists of the Caples, Routeburn, and Hidden Falls Valleys, while further north is a great intrusion which extends from the vicinity of Red Mountain to the northern end of the Olivine Range. All these rocks show petrographic and tectonic peculiarities which stamp them as belonging to a single intrusive series. Among the nephrites and related rocks which form the subject of the present investigation there is a distinctive series of specimens—mainly tremolite-rocks and seminephrites—which, as will be shown in a later section, represent a group of rocks genetically connected and associated in the field with the peridotites and serpentines of the region between Lake Wakatipu and the Olivine Range.

In a number of these rocks there are residual grains of hornblende clearly showing gradual transition into tremolite, which is sometimes accompanied by crystallisation of small amounts of sphene. The view is therefore put forward that many of the nephritic rocks from this region have been derived from hornblendic rocks such as hofnblendites, a mode of origin which, so far as the writer is aware, has not hitherto been recorded.

In this connection it may be noted that in an account of the ultrabasic igneous rocks of the Olivine Range (Turner, 1933), the writer has described dykes of hornblendite and of pyroxenite invading the main peridotite intrusion. Whereas the hornblende of these

– 196 –

hornblendites is invariably at least partially replaced by tremolite, transition from pyroxene to tremolite or other amphibole was never observed in the pyroxenites, even though there is widespread alteration of pyroxene to antigorite, talc, pennine, garnet, diopside, and magnetite in many of these rocks. Certain specimens of tremolite-rock (1364 and 1379) and of schistose nephrite (1370) described in this paper were obtained from boulders brought down from the peridotite belt of the Olivine Range.

The presence of residual grains of chromite or picotite in a number of the nephritic greenstones examined would appear to indicate that such rocks originated in most cases from pyroxenite or peridotite rather than from hornblendite, though both chromite and hornblende were in one instance observed in the same specimen (1853).

The essential differences between nephrites, seminephrites, and tremolite-rocks, as here defined, concern microstructure and grain-size, which themselves depend upon the physical conditions which obtained during the crystallisation of the rocks. Several cases may be suggested:—

  • (a) The original rock (pyroxenite, hornblendite, or peridotite) undergoes chemical reconstitution under simple shearing stress, unaccompanied by mechanical movement of the constituent crystals. The result is a tremolite-rock, usually coarse-grained, with distinct plane or linear schistosity.

  • (b) A coarse-grained tremolite-rock such as the above is later subjected to shearing stress accompanied by actual mechanical movement and shearing of the component crystals, without further recrystallisation, the process being mainly cataclastic. The resulting rock is a fine-grained tremolite-rock, such as 1853, in which the constituent sheaves of fibres, though unoriented, show no trace of felted nephritic structure.

  • (c) When a coarse-grained tremolite-rock is subjected to mechanical shearing accompanied by recrystallisation of the sheared material, i.e., if shearing and growth of fresh crystals are contemporaneous, lensoid streaks of thoroughly felted fibrous tremolite with nephritic structure are developed throughout the coarsely crystalline mass, giving rise to a typical seminephrite.

  • (d) When shearing of the original tremolite-rock and concomitant growth of fresh fibres have reached an advanced stage, the rock becomes a true nephrite either of the felted or the schistose type, sometimes enclosing small residual masses of coarse material. There is abundant evidence that most of the nephrites and seminephrites described in this paper have originated in this way.

  • (e) Nephrites may also be formed by intense shearing and mechanical breaking down of the original parent rock (pyroxenite, hornblendite, etc.), with concomitant chemical reconstitution and growth of fresh crystals.

– 197 –

Crystallisation of the tremolite has been accompanied, in many of the rocks considered, by crystallisation of various minor constituents such as sphene, pennine, serpentine, talc, carbonate, quartz, and possibly diopside. In some rocks, however, recrystallisation of certain minerals appears to have taken place subsequently to the main period of tremolite-formation, after shearing movements within the rock had almost or quite ceased. Examples of chemical reconstitution of this latter type have already been described, and may here be summed up as follows: crystallisation of sharply-bounded patches of pennine or coarsely crystalline talc, the flakes of pennine sometimes being roughly radially arranged around a central nucleus of chromite or picotite; crystallisation of idioblastic relatively large crystals of clinochlore (1844); crystallisation of idioblastic acicular prisms of diopside which in rare cases surround a nucleus of chromite, the pyroxene then being chromiferous; late crystallisation of coarse acicular idioblasts of tremolite, and growth of minute veinlets of transversely fibrous tremolite along fracture planes cutting across the felted groundmass.

Grubenmann (1910) places nephrites along with hornblende-schists and actinolite-schists among the products of dynamo-thermal metamorphism of magnesian rocks within the middle or meso-zone of metamorphism. There appears to be strong evidence, however, indicating that some if not all of the nephrites of New Zealand have crystallised as such under conditions typical of the upper or epi-zone of Grubenmann. The nephrites of the northern part of Westland occur, according to Bell and Fraser (1906, pp. 69, 70), in the highly altered magnesian rocks of the Pounamu Formation, which invade biotite-bearing schists which from the descriptions given by Morgan (1908, pp. 83, 121) appear to be typical meso-zone rocks. In the southern area between the head of Lake Wakatipu and the Olivine Range of South Westland, tremolite-rocks, semi-nephrites, and at least one specimen of true felted nephrite have been obtained from sheared intrusions of peridotite and serpentine which invade sheared greywackes, phyllites, and schists, some of which are typical meso-zone rocks, but many of which have suffered only relatively mild metamorphism within the epi-zone. In a recent paper the writer (Turner, 1933) has shown that the period of peridotite invasion in this southern area certainly postdates the period of intense dynamothermal metamorphism when the invaded schists assumed their present state, and, further, that the peridotite masses merely suffered partial metamorphism in the epi-zone shortly after intrusion. It was in this latter, probably Lower Cretaceous, period of metamorphism therefore that the nephritic rocks of South Westland and North-west Otago were formed. It appears probable that the same sequence of events also holds for North Westland, though this area has not yet been examined by the writer. In the second place, the reconstituted minerals of the nephrites and related rocks—tremolite, pennine, clinochlore, serpentine, talc, sphene, and a rhombohedral carbonate—constitute a low-grade metamorphic assemblage typical of Grubenmann's epi-zone, though the occasional presence of diopside in addition is admittedly anomolous. Finally, the microstructures of the rocks themselves indicate that cataclasis,

– 198 –

the process that characteristically predominates in epi-zone metamorphism, has played a most important part in the formation of these rocks.



Among the greenstones used by the Maoris of Otago serpentine-rocks appear to be almost, if not quite, as plentiful as nephrites and related rocks.

The optical properties of the mineral itself are remarkably constant. Typically it is colourless, with a fairly high refractive index (often greater than 1.60), and a birefringence of about 0.01. It is finely fibrous in habit, with straight extinction parallel to the length of the fibres, positive elongation, small to moderately large optic axial angle and negative optical character. As pointed out by Finlayson (1909, p. 362) these properties agree well with those of antigorite except that the characteristic flaky habit and “thorn-structure” of the latter mineral are completely lacking. Although true antigorite-rocks are known to be plentiful in the serpentine and peridotite belts of Nelson and Westland, only a single specimen with typical “thorn-structure” was found by the present writer among the Maori greenstones examined by him.

Accessory minerals are seldom plentiful in the serpentine-rocks and are completely lacking in most cases. They include talc, epidote, a rhombohedral carbonate, pennine, tremolite, magnetite, and chromite.

The microstructure of the serpentinous greenstones varies considerably and may conveniently be used as a basis for subdivision of these rocks into six classes:—

  • (a) Antigorite-serpentines with typical thorn-structure as defined by Bonney and Miss Raisin (1905, p. 702).

  • (b) Fibrous serpentines with linear schistosity due to perfect parallelism of the component fibres throughout the rock.

  • (c) Fibrous serpentines consisting entirely or for the most part of parallel fibres, crossed by minutely spaced very narrow zones of incipient strain-slip, which give rise to a striking regularly-banded appearance when the section is viewed between crossed nicols.

  • (d) Fibrous serpentines with imperfect linear schistosity, due to the partial development of lenses and patches consisting of unoriented felted sheaves of fibres, interspersed with areas where the serpentine fibres have parallel disposition.

  • (e) Fibrous serpentines with felted microstructure analogous to that of true nephrites.

  • (f) Plane-schistose fibrous serpentines which in sections parallel to the foliation appear to be composed of minute, unoriented, but in no way interfelted bundles of parallel fibres.

The beautiful translucent variety of serpentine (tangiwai) which was obtained by the Maoris from Anita Bay, Milford Sound, has been identified by Finlayson (1909, p. 361) and others as bowenite. The characteristic properties of this rock, as defined for example by Finlayson (1909, pp. 361, 362), Leitmeier (1914, p. 385), and

– 199 –

Dana (1922, p. 573), are its unusual hardness and macroscopic resemblance to nephrite. The microstructure appears, however, to be by no means constant. Thus Bonney (1908, pp. 169–170) describes the bowenite of Afghanistan as being essentially similar to antigorite-serpentine and possessing marked thorn-structure, while Finlayson (1909, p. 362), on the other hand, states that the bowenite of Milford Sound is a fibrous serpentine somewhat resembling nephrite in the closely felted arrangement of the fibres. The series of sections of Milford Sound bowenite examined by the present writer show all degrees of transition from fibrous serpentines of class (b) composed of regularly oriented fibres, to thoroughly felted serpentines of class (e) in which the microstructure resembles that of nephrite.

It is therefore probable that most, if not all, of the serpentinous greenstones here described, with the possible exception of the antigorite-rock 1854, should be classed as bowenites, since the hardness is typically high (sometimes greater than that of steel, even when the test is carried out on a rough surface), and there is invariably a strong superficial resemblance to nephrite.

Class (a). To this class belongs a single specimen (1854) of bright apple-green rather soft serpentine, which in thin section is seen to be a typical antigorite-rock with perfect thorn-structure, consisting of unoriented plates and blades 0.1 mm. to 0.3 mm. in length. Small interstitial grains of opaque iron-ore are locally abundant.

Class (b). Eight specimens (1855 to 1861, and 1866) belong to this group. Typically the only constituent is serpentine, which though apparently homogeneous when seen beneath the low-power objective, appears under strong magnification to be composed of long, very slender, perfectly parallel fibres, the orientation of which is approximately constant throughout the section. As a result of aggregate polarisation of the similarly oriented fibres, good interference-figures may be obtained in convergent light; these show a negative optic character in all cases, the optic axial angle being somewhat variable. The elongation of the fibre is invariably positive, so that the acute bisectrix must be normal to the elongation. This latter property, which has constantly been observed in the serpentines described in this paper whenever the fibres are sufficiently regularly oriented to yield aggregate interference-figures, is in perfect accordance with the optic orientation of antigorite but cannot be reconciled with that of chrysotile in which the fibres are elongated parallel to the acute bisectrix. The effects of incipient shearing subsequent to crystallisation are seen in the undulose extinction displayed by much of the serpentine (due to slight irregular distortion of the fibres), in the development of rare lensoid zones of shearing where the fibres have been twisted and fractured to a noteworthy degree, and in the presence of sparsely scattered, regular, parallel narrow zones (0.01 mm. wide) which go into extinction simultaneously while the main mass of the rock is still illuminated. These are interpreted as parallel zones of incipient strain-slip in which the fibres have suffered slight but uniform distortion. Talc and epidote are present as accessory constituents in 1855. The former

– 200 –

mineral occurs as swarms of minute flakes (0.04 mm. long) approximately parallel to the length of the serpentine fibres, and also as coarser ragged crystals (0.2 mm. in diameter) which are concentrated either in the sheared lenses of serpentine or else in small secondary veinlets consisting largely of epidote. The epidote is a colourless, highly birefringent variety and is confined entirely to sparsely scattered secondary veinlets 0.2 mm. in width which trend parallel to the direction of the strain-slip bands. The epidote and much of the talc therefore appear to have crystallised subsequently to the crystallisation of the serpentine which makes up the bulk of the rock.

Class (c). The rocks of this group (1862 to 1865) are serpentines consisting of originally parallel fibres, which are crossed, however, by minutely spaced parallel strain-slip cleavage-planes. Along any particular strain-slip plane the actual displacement is always very small, and the individual fibres have thus suffered distortion through a uniform angle without being fractured. As a result, the section appears between crossed nicols to be crossed by regularly arranged, parallel narrow bands (perhaps 0.03 mm. in width), which fall into alternating series, the members of each series extinguishing simultaneously in a manner somewhat comparable with the alternating lamellae of a plagioclase twin. That the fibres have been thrown only into a series of minute undulations without suffering actual fracture or even sharp bending is indicated by the undulose fashion in which the wave of extinction moves from the dark to the formerly illuminated bands as the stage is rotated. The extinction-angles measured with reference to the trend of the bands vary considerably in any one section, though always constant within the limits of the field of view; three typical measurements (1862) are 20° and 2°; 26° and 0°; 40° and 9°. The orientation of the bands is not necessarily uniform throughout the section as a whole, though such may often be the case.

As the effects of shearing become more pronounced, fracturing of the fibres along the shear planes takes place and irregularly streaky areas consisting of unoriented sheaves of fibrous serpentine are developed, so that there is gradual transition through rocks such as 1865 to serpentines of class (d). Similarly there is no sharp line of distinction between the serpentines of class (c) and those of class (b), in some of which (e.g., 1856) widely-spaced planes of strain-slip cleavage may be developed and may even be visible in hand-specimen.

Class (d). The rocks of class (d), while invariably consisting partly of finely fibrous serpentine with parallel structure as in class (b), also show a noteworthy development of lenses, streaks, and irregular patches, consisting mainly of felted, twisted, unoriented tufts and sheaves of fibres, the arrangement of which recalls the structure of nephrite. This felted material has obviously originated by shearing and breaking down of the regularly oriented fibrous serpentine. All stages of transition are thus shown, between these rocks and those of class (b) on the one hand and class (e) on the other.

Accessory minerals are usually lacking, though pennine may occasionally be present (e.g., 1870), while talc and epidote are

– 201 –

plentiful in a small group of rocks (1495, 1867, 1868) which are the partially sheared equivalents of the series represented by 1855 to 1857 in class (b). The talc occurs both as minute isolated flakes and also intimately associated with the epidote which is usually concentrated in streaky patches consisting of talc, epidote, serpentine, and sometimes magnetite. In 1868 talc and epidote together make up 10% of the section. Attention may here be drawn to the fact that all the talc-epidote-bearing serpentines examined by the writer are typical tangiwais, including specimens obtained in situ from Anita Bay (Milford Sound) and the Pyke River (Red Hills country) as well as worked Maori material.

The specimens grouped in class (d) include 1495 and 1867 to 1884. They are by far the most numerous of the serpentines used by the Maoris of Otago and South Canterbury.

Class (e). Five specimens of fibrous serpentine (1885 to 1889) have suffered such severe shearing that but little of the original parallel structure remains, and the rock now consists for the most part of unoriented interfelted bundles of fibres, the arrangement of which closely resembles that shown in the felted nephrites. Specimens of this type are therefore very difficult to distinguish microscopically from nephrites, for it is often impossible to demonstrate straight extinction in the twisted and interwoven fibres, while there seems to be no striking difference between the refractive indices of the two minerals. Nevertheless, when parallel fibres, cut perpendicular to the acute bisectrix, yield an aggregate interference-figure the relatively small optic axial angle serves to distinguish the serpentine from tremolite. Furthermore, in most of these rocks the passage from fibrous serpentine with parallel structure to unoriented felted serpentine is obvious in localised areas of the sections. In one specimen (1886) colourless pennine is present to the extent of about 5% both as scattered flakes and in well-defined streaks and lenses between 0.1 mm. and 0.5 mm. in width. Small grains of opaque iron-ore are sometimes enclosed in these chloritic areas.

Class (f). Two specimens (1890 and 1891) were found to consist almost entirely of small, straight, unoriented bundles of serpentine fibres (ranging up to 0.5 mm. × 0.05 mm.) with no trace of the twisting and felting always shown to some extent by the rocks of the previous class.

Class (g). In addition to the above six classes of serpentine based upon microstructure, a seventh group of minor numerical importance represented by specimens 1893 to 1895 may be recognised. These rocks are characterised by the presence of tremolite as an essential constituent in addition to serpentine, and may therefore be regarded as in some respects intermediate between the serpentines on the one hand and the nephritic greenstones on the other.

As a typical example, 1893 may be described in greater detail. About half of the section consists of finely fibrous serpentine with regularly parallel structure exactly similar to the serpentines of class (b). There are also irregular patches of felted nephritic, fibrous tremolite with distinctly higher birefringence than the serpentine

– 202 –

and enclosing definite, long, relatively coarse prisms of tremolite with markedly oblique extinction and the characteristic basal cross-fracture.

Origin of the Serpentine-rocks.

Though residual silicate minerals were not observed in any specimen of greenstone serpentine, it may safely be assumed that these rocks have been formed like most other serpentines by hydration of magnesian igneous rocks such as peridotites and enstatite-pyroxenites. Finlayson's suggestion that the bowenites of Milford Sound are dynamically metamorphosed derivatives of the associated talc-schists (Finlayson, 1909, p. 363), appears, in the light of modern conceptions of metamorphism, to be extremely unlikely.

That the process of serpentisation took place under conditions of intense shearing stress is suggested by the fact that the chief constituent of all specimens is a variety optically identical with the typical “stress-mineral” antigorite. It is not yet possible to speculate as to the conditions determining the fibrous habit of the serpentine, in contrast with the “thorn-structure” normally displayed by antigorite-serpentines both in New Zealand and elsewhere.

As will be seen from the petrographic account given above, the various microstructures observed in the different classes of serpentinous greenstone [except class (a)] are all interpreted as resulting from more or less intense secondary shearing of serpentine which originally crystallised under shearing stress as parallel or subparallel fibres.


Fine-grained schistose rocks containing abundant talc are occasionally found among Maori greenstones and are here represented by three specimens (1657, 1896, 1897).

In 1896 about 95% of the section consists of minutely crystalline talc in unoriented flakelets about 0.03 mm. long, rarely reaching dimensions as large as 0.1 mm. × 0.05 mm. Tiny flakelets of pale-green pennine build up well-defined patches 0.2 mm. to 0.6 mm. in width, occasionally enclosing small grains of iron-ore. In 1897, serpentine, chlorite, and talc are all plentiful. Though often mutually associated, these minerals also tend to build up lensoid monomineralic areas reaching 7 mm. in length. In 1657 the section consists entirely of rather coarse pale green talc and finely granular quartz. The rock is in fact a talc-quartz-schist, but the hand-specimen has a most striking deep emerald-green colour due to the unusually deep colour of the talc.

Macroscopic Properties of Greenstones.

The diagnostic physical properties of greenstones most readily observed in hand-specimens include colour, hardness, fissility, and mode of fracture.

Colour and General Appearance.

Colour is one of the most variable properties of greenstone, and most of the classes founded on petrographic grounds are found to include specimens ranging from greyish and even whitish tints to

– 203 –

various shades and depths of green. Nevertheless, colour is often an important indication of the petrographic class to which a rock belongs, and is especially useful in distinguishing serpentines from nephrites.

In general the colour appears to be more uniform in the serpentines as a class, yellowish sea-green and bluish-green tints being especially characteristic, although occasionally observed also in the nephritic greenstones. Serpentines, too, seem on the whole to be more translucent than the nephritic rocks, and, furthermore, often have a characteristic finely-banded appearance on polished surfaces, not unlike the grain of polished wood. On the other hand, most of the opaque greyish-green specimens are tremolite-rocks and semi-nephrites, while deep grass-green, subtranslucent, homogeneous types are usually true nephrites. Among the serpentines many of the tangiwais have a fairly deep green tint; but these can usually be distinguished from nephrites by their greater translucency and mottled appearance. The above generalisations are subject to many exceptions, but nevertheless serve as useful guides in hand-specimen determination.

Finlayson (1909, pp. 369 to 371) has discussed fully the causes of variation in colour of nephrites. He concludes that the depth of green is proportional to the amount of ferrous iron replacing magnesium in the constituent tremolite or actinolite. It may here be noted that the brilliant emerald-green spots observed in very rare specimens of nephrite (e.g., 1804) have been found to be due to the presence of aggregated crystals of chrome-diopside, while in one or two specimens (e.g., 1839) patches of green talc produce a somewhat similar appearance.


In his discussion of the physical properties of bowenitic serpentines (the Maori tangiwai) and nephrites, Finlayson (1909, pp. 361, 362, 367) gives the following values for hardness: for bowenite, estimated on rough surfaces 3 to 4.5, on polished surfaces 5 to 5.5, amounting at times to 6; for nephrites, on rough surfaces 4.5 to 6, on polished surfaces uniformly 6.5. The values obtained in the present investigation are somewhat different and may be summarised as follows:—

Nephrites: 6 to 6.5 (always > steel) on polished surfaces; 5.5 (< steel) to 6 (> steel) on rough surfaces.

Seminephrites: 5.5 to 6 on polished surfaces; 3.5 to 5.5 (rarely 6) on rough surfaces.

Tremolite-rocks: 5 to 6 on polished surfaces; 3 to 5.5 (in one case 6) on rough surfaces.

Serpentines: 5.5 to > 6 (rarely as low as 5) on polished surfaces; 4.5 to 6 (rarely > 6) on rough surfaces.

Thus, as a whole, the serpentines are little inferior in hardness to the nephrites and related rocks. Out of about 100 specimens of greenstone tested, only 20 were harder than steel as estimated on rough fractured surfaces, and of these half were serpentines and half nephritic rocks. While the hardest specimens of nephrite appear to have a hardness slightly greater than that of the hardest

– 204 –

serpentines, some of the latter definitely have a hardness > 6 as determined by careful comparison with the standard mineral orthoclase. Hardness is therefore of little use in distinguishing betwen serpentines and nephritic greenstones.

In the nephrites and allied rocks, the great hardness is due to the characteristic felting and twisting of the constituent tufts of fibres, since a marked increase in hardness accompanies the transition from poorly felted rocks such as tremolite-rocks and seminephrites, to the true nephrites. In the serpentines, however, this correspondence between hardness and microstructure is not so pronounced, and the hardness is less variable than in the nephritic greenstones. Corrugation of the serpentine fibres such as is responsible for the “banded” structure of serpentines of class (c) seems, however, to have a notable effect in increasing the hardness which in all specimens of this type has a value of 6 or more even on rough fractured surfaces.

Fissility and Toughness.

The fissility of a greenstone, i.e., the ease with which it splits in one direction (parallel to the schistosity), depends upon the degree of parallelism of the constituent crystals of serpentine or tremolite. Thoroughly felted rocks such as the felted nephrites are thus non-fissile and so tough that it is sometimes almost impossible to chip the specimens even with a heavy hammer. As would be expected from the fact that thoroughly felted microstructure is seldom developed in serpentinous greenstones, the latter are on the whole more fissile and less tough than the nephritic rocks. Again, among the members of the nephritic group, true nephrites are much tougher and less fissile than seminephrites and tremolite-rocks. It is this extreme toughness of nephrites combined with great hardness, that rendered these rocks so valuable to the Maoris in the manufacture of such implements as adzes, chisels, and drill points.

Summary of Macroscopic Properties.

Discrimination between different varieties of greenstone must always be based ultimately upon microscopic examination of a thin section. Since, however, such an examination is often impracticable in dealing with ethnographic material, in order to facilitate hand-specimen determination, the following generalised summary of the diagnostic properties of the more important varieties has been drawn up. Such a scheme is by no means infallible, but may be followed with success in perhaps 80 cases out of 100.

Nephrites: colour various shades of green, deep green predominating; translucent to opaque; fracture finely flaky; hardness 6 to 6.5 (always > steel) on polished surfaces, 5.5 to 6 (sometimes > steel) on rough surfaces; usually very tough; schistosity poorly developed or absent in the “felted” nephrites, well marked in “schistose” nephrites.

Seminephrites and Tremolite-rocks: colour greyish-white to green, grey or greenish-grey tints being characteristic though by no means universal; streaks and flaws are commoner than in true nephrites; often opaque or poorly translucent; hardness 5 to 6 (often < steel) on polished surfaces, 3 to 5.5 (always < steel) on fractured surfaces; fracture is typically rather coarsely flaky; usually fissile.

– 205 –

Fibrous Serpentines: colour variable, sea-green and bluish-green tints being highly characteristic; usually translucent, often perfectly so; polished surfaces often show parallel marking recalling the grain of polished wood; hardness 5.5 to 6 (> or < steel) on polished surfaces, 5 to 6 or rarely > 6 on fractured surfaces; fracture finely to coarsely flaky, sometimes splintery; fissility usually well marked.

Antigorite-serpentine with “thorn-structure”: colour deep apple-green; subtranslucent; hardness 5 to 5½ on rough and polished surfaces, thus distinguishing it from true nephrite; fracture finely flaky; non-fissile.

Talc-rocks: colour very variable, greyish to bright emerald-green; subtranslucent to opaque; hardness typically 1 to 3 on rough surface, this extreme softness being distinctive; usually fissile.

Relative Abundance of Varieties.

In making a rough estimate of the relative abundance of the different types of greenstone used by the Maoris of South Canterbury and Otago, those specimens collected from the three regions from which the Maoris obtained the raw greenstone—i.e., the Teremakau area of North Westland, the Lake Wakatipu-South Westland region and Milford Sound—have been omitted from consideration. The estimate given below therefore applies only to material obtained from Maori camp-sites in the vicinity of the east coast:—

Nephrites (felted and schistose) 16
Seminephrites 16
Tremolite-rocks 7
Total 39
Serpentines of Class (a) 1
" " Class (b) 5
" " Class (c) 4
" " Class (d) 17
" " Class (e) 5
" " Class (f) 2
Total 34
Serpentine-tremolite-rocks [Class (g)] 3
Talc-rocks 3
Total number of greenstone 79

The above figures admittedly do not accurately represent the relative abundance of the different petrographic types among the greenstones of Otago and South Canterbury as a whole; for example, true nephrites are probably slightly more abundant than is indicated above. Nevertheless, it is obvious that serpentines are much more plentiful than has hitherto been supposed by previous writers on the subject (e.g., Finlayson, 1909, p. 361), many of whom assert that the greenstones used by the Maoris were mainly nephrites, serpentines being of minor importance only. It should be noted that this latter statement does hold true as far as the translucent serpentines known as tangiwai are concerned, for among the serpentine-greenstones sectioned by the writer only one specimen of typical tangiwai was found in the east-coast area.

– 206 –

Possible Sources of Greenstones.

Petrographic evidence bearing upon the problem of the sources from which the Maoris of Otago and South Canterbury obtained greenstone has been brought together from observations made upon a number of specimens collected both in situ and from stream-boulders, in the three main districts where greenstone is known to occur, viz., the Teremakau-Arahura area of northern Westland; the region between the head of Lake Wakatipu and the coast of southern-most Westland; and the district around the mouth of Milford Sound long known as the principal source of tangiwai. The observations now put forward are admittedly incomplete, and can be extended only by much additional field-research. Nevertheless, the following facts elicited during the present investigation may now be placed on record as a partial solution to the problem:—

  • (a) The greenstones from the Arahura-Teremakau district, with a single exception (1844), are true nephrites (both schistose and felted), and it is probable that most, if not all, of the superior nephrites used in Otago came from this northern area. A single specimen (1370) of true schistose nephrite was nevertheless obtained from a boulder in the bed of the Jackson River ten miles above its junction with the Arawata in the far south of Westland, and was doubtless derived from the northern end of the great peridotite belt which forms the north-western flank of the Olivine Range. Judging from the fact that Maori implements (tangiwai) have been found at Smoothwater Bay, not more than one day's journey from the South Westland peridotite belt, it is not improbable that this source of greenstone was known to the Maoris, and some of the nephrites of Otago may ultimately have been obtained from this area.

  • (b) A number of adzes and partially worked pieces of poor-grade seminephrite and tremolite-rock with distinctive petrographic peculiarities, such as the frequent presence of pennine, serpentine, sphene, chromite, and especially residual hornblende, were obtained some years ago from the Dart Valley, Lake Wakatipu, by Mr C. Haines, of Glenorchy, who presented representative specimens to the Otago Museum. Identical rocks have recently been found by the writer in the Routeburn Valley, some two and a-half miles above its junction with the Dart, both as stream boulders and in situ in association with serpentine and talc-schist. It has thus been definitely esablished that the Dart Valley greenstones are of local origin.

  • (c) Tremolite-rocks (e.g., 1364 and 1379) somewhat similar to those of the Dart and Routeburn Valleys have been collected by the writer from boulders in tributaries of the Cascade River, which drains the peridotite area of the Olivine Range (Turner, 1933).

  • (d) Many of the seminephrites obtained from Maori sites in Otago and South Canterbury have certain distinctive petrographical features in common with the seminephrites, etc., of the Lake Wakatipu-South Westland region, e.g., the common occurrence of a particular variety of pennine and the rarer development of sphene, chromite, and, above all, diopside, a mineral not hitherto recorded in New Zealand nephrites. It is not unlikely, therefore, that much of the seminephritic greenstone of Otago and South Canterbury

– 207 –
  • came from this southern source,* though further field-work in northern Westland is necessary before a definite conclusion can be reached in this connection.

  • (e) The source of the clear translucent serpentine known to the Maoris as tangiwai has long been known to be Milford Sound. Much of the Milford Sound greenstone contains a characteristic and otherwise unusual pair of accessory minerals, epidote and talc, which sometimes make up as much as 10% of the total composition. The same minerals were also found in a typical tangiwai obtained in situ by Mr C. Yunge, of Stewart Island, from the South Westland peridotite belt in the upper part of the Pyke Valley. It is therefore possible that this area formed an additional source of the Maori tangiwai.

  • (f) The single specimen of antigorite-serpentine with “thorn-structure” obtained from a Maori site at Greenstone Island, South Canterbury (1854) bears a striking resemblance, both in hand-specimen and beneath the microscope, to the hard antigorite-rocks which occur plentifully as boulders in the Cascade Valley and which were originally recorded by Ulrich (1890, p. 629) as “hard nephrite-like serpentine (? bowenite).” Antigorite-serpentines are known, however, from many of the peridotite areas of the South Island, and it is therefore possible that the Greenstone Island specimen was obtained elsewhere than in the Cascade Valley region.

  • (g) It is impossible at present to conjecture what may be the exact source of many of the greenstone serpentines. Some of these have persistent peculiarities in microstructure which stamp them as being distinct from anything yet recorded in situ. At present it can only be stated that the source or sources probably lie in the serpentine intrusions that cut the schists and older sedimentary rocks at many points along the western flank of the Main Divide of the South Island.


This research was originally undertaken at the suggestion of Mr H. D. Skinner, Lecturer in Ethnology at the University of Otago. The writer is greatly indebted to Mr Skinner, not only for his assistance in selecting suitable Maori greenstones for investigation, but also for much helpful discussion with reference especially to ethnographic data bearing upon the question of the sources from which the Maoris of Otago obtained greenstone.

The thanks of the writer are also due to the Council of the University of Otago and to Dr W. B. Benham, Curator of the University Museum, for permission to use most of the specimens on which this research is based; to Mr Gilbert Archey for assistance in the examination of greenstone in the Auckland War Memorial Museum; and to Mr David Teviotdale for assistance and information of various kinds.

[Footnote] * Mr H. D. Skinner informs the writer that he has collected ethnographic evidence conclusively showing that the Lakes country of West Otago was a source of much Maori greenstone. These data are placed on record in a paper (now in the press) to be published later in this volume of the Trans. Roy. Soc. N.Z.

– 208 –

List of Specimens Examined.

In the following list the localities, from which the various specimens examined were obtained, are given. The left-hand column refers to the numbers of the micro-sections and hand-specimens in the collections of the Geology Department, Otago University. In the middle column are given the corresponding numbers of hand-specimens lodged in the Otago University Museum, from which most of the material forming the subject of this research was initially obtained. In the right-hand column, localities written in italics refer to greenstones obtained as river-boulders or in situ; otherwise the localities recorded are Maori camp-sites.

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

Specimen No. (Geology Dept., Otago University). Specimen No. (Otago Museum). Locality.
1800 D 33. 1382 Otago Beaches.
1801 D 30. 1746 Milford, South Canterbury.
1802 D 30. 1689 Morven, South Canterbury.
1803 D 21. 447 Warrington.
1804 D 33. 1394 Otago Beaches.
1805 D 33. 1813 Teremakau River.
1806 D 33. 1395 Otago Beaches.
1807 D 30. 1739 Greenstone Island, S. Canterbury.
1808 D 33. 1410 Westland.
1809 D 33. 1380 Otago Beaches.
1810 33. 1378 Shag River.
D 29. 2377
1811 D 33. 1406 Greenstone Island, S. Canterbury.
D 30. 1712
1812 D 33. 1419 Tercmakau River.
1813 D 33. 1409 New River, North Westland.
1814 D 30. 1739 Greenstone Island, S. Canterbury.
1815 D 33. 1400 Shag River.
D 22. 2314
1816 D 30. 1745 Milford, South Canterbury.
1817 D 33. 1393 Greenstone Island, S. Canterbury.
D 30. 1705
1818 D 33. 1377 Murdering Beach, Otago.
1819 D 21. 496 Centre Island, S. Canterbury.
1820 D 30. 1742 Greenstone Island, S. Canterbury.
1821 D 30. 1693 Milford, South Canterbury.
1822 D 29. 2373 Shag River.
33. 1403
1823 D 33. 1414 Shag River.
1824 D 33. 1381 Murdering Beach, Otago.
1825 D 30. 1693 Milford, South Canterbury.
1826 D 19. 368 Dart Valley, Western Otago.
1827 D 30. 1737 Greenstone Island, S. Canterbury.
1828 D 29. 2415 Shag Valley.
1829 D 33. 1386 Murdering Beach, Otago.
1830 D 33. 1402 Murdering Beach, Otago.
1831 D 30. 1702 Greenstone Island, S. Canterbury.
1832 D 29. 6039 Willowbridge, South Canterbury.
1833 D 19. 365 Dart Valley, Western Otago.
1834 D 19. 237 Dart Valley, Western Otago.
1835 D 22. 526 Warington, Otago.
1836 D 33. 1389 Mudering Beach, Otago.
1837 D 30. 1709 Greenstone Island, S. Canterbury.
1838 D 19. 243 Dart Valley, Western Otago.
1839 D 33. 1375 Little Papanui, Otago.
– 209 –

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

1840 D 29. 2417 Shag River.
1841 D 30. 1703 Greenstone Island, S. Canterbury.
1842 33. 1405 Dart Valley, Western Otago.
D 19. 231
1843 D 19. 244 Dart Valley, Western Otago.
1844 Arahura River.
1845 D 33. 1415 Dart Valley, Western Otago.
1846 D 30. 1696
1847 D 19. 240 Dart Valley, Western Otago.
1848 D 19. 122 Dart Valley, Western Otago.
1849 D 19. 232 Dart Valley, Western Otago.
1850 D 33. 1392 Otago Beaches.
1851 D 33. 1416 Otago Beaches.
1852 D 19. 230 Dart Valley, Western Otago.
1853 D 19. 363 Dart Valley, Western Otago.
1854 D 30. 1687 Greenstone Island, S. Canterbury.
1855 D 22. 455 Smoothwater Bay, S. Westland.
1856 D 18. 388 Milford Sound.
1857 D 29. 1326 Anita Bay, Milford Sound.
1858 D 33. 1412 Little Papanui, Otago.
1859 D 33. 1408 Murdering Beach, Otago.
1860 D 33. 1390 Warrington, Otago.
D 22. 518A
1861 D 29. 2419 Goodwood Beach, Otago.
1862 33. 1391 Little Papanui, Otago.
D 30. 530
1863 D 29. 180 Murdering Beach, Otago.
1864 D 33. 1411 Little Papanui, Otago.
1865 D 22. 521 Warrington, Otago
1866 D 33. 1398 South Canterbury.
D 30. 1710
1867 D 33. 1404 Otago Beaches.
1868 D 33. 1385 Milford Sound.
1869 D 33. 1417 Willowbridge.
1870 D 22. 246 Sandflat S. of Kaik, Otago Heads.
1871 D 30. 1743 Greenstone Island, S. Canterbury.
1872 Otago Beaches.
1873 D 29. 5605 Little Papanui, Otago.
1874 D 30. 637 Little Papanui, Otago.
1875 D 33. 1399 Warrington, Otago.
1876 D 30. 1718 South Canterbury.
1877 D 30. 1713 South Canterbury.
1878 33. 1418 South Canterbury.
1879 D 33. 1383 South Canterbury.
D 30. 1758
1880 D 30. 1692 Milford, South Canterbury.
1881 D 33. 1379 Otago Beaches.
1882 D 30. 1387 Greenstone Island, S. Canterbury.
1883 D 33. 1387 Murdering Beach, Otago.
D 22. 553
1884 D 33. 1376 Otago Beaches.
1885 D 22. 553 Murdering Beach, Otago.
D 33. 1374
1886 D 29. 6044 Willowbridge, South Canterbury.
1887 D 33. 1397 Shag River.
D 29. 2380
1888 D 23. 107 Murdering Beach, Otago.
1889 D 33. 1386 Otago Beaches.
1890 D 33. 1388 Otago Beaches.
1891 Otago Beaches.
– 210 –

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

1892 D 30. 1704 Greenstone Island, S. Canterbury.
33. 1407
1893 D 29. 2395 Shag River.
1894 D 30. 1750 Milford, South Canterbury.
1895 D 29. 2379 Shag River.
1896 D 30. 1124 Greenstone Island, S. Canterbury.
1897 Otago Beaches.
1364 Jackson River, South Westland.
1370 Jackson River, South Westland.
1425 Routeburn, Western Otago.
1443 Routeburn, Western Otago.
1452 Routeburn, Western Otago.
1495 D 33. 1396 Pyke River, Western Otago.
1657 Southland.
Y 18 Milford Sound.
Za 5 Teremakau River.

Literature Cited.

Baur, M., 1914. Nephrit und Jadeit, C. Doelter's Handbuch der Mineralchemie, vol. 2, pt. I, Dresden, pp. 649–704.

Bell, J. M., and Fraser, C., 1906. The Geology of the Hokitika Sheet, North Westland Quadrangle, N.Z. Geol. Surv. Bull., No. 1 (new series).

Bonney, T. G., 1908. On Antigorite from the Val Antigorio, with Notes on Other Serpentines containing that Mineral, Q.J.G.S., vol. 64, pp. 152–170.

Bonney, T. G., and Raisin, C., 1905. The Microscopic Structure of Minerals forming Serpentine, and their Relation to its History, Q.J.G.S., vol. 61. pp. 690–715.

Chapman, F. R., 1892. On the Working of Greenstone or Nephrite by the Maoris, Trans. N.Z. Inst., vol. xxiv, pp. 479–539.

Dana, E. S., 1922. A Text Book of Mineralogy, 3rd ed., John Wiley and Sons, New York.

Finlayson, A. M., 1909. The Nephrite and Magnesian Rocks of the South Island of New Zealand, Q.J.G.S., vol. 65, pp. 351–381.

Grubenmann, U., 1910. Die Kristallinen Schiefer, 2nd ed., Berlin.

Leitmeier, H., 1914. Serpentin, C. Doelter's Handbuch der Mineralchemie, vol. 2, pt. I, Dresden, pp. 385–435.

Morgan, P. G., 1908. The Geology of the Mikonui Subdivision, North Westland, N.Z. Geol. Surv. Bull., No. 6 (new series).

Turner, F. J., 1930. The Metamorphic and Ultrabasic Rocks of the Lower Cascade Valley, South Wesland, Trans, N.Z. Inst., vol. 61, pp. 170–201.

Turner, F. J., 1933. The Metamorphic and Intrusive Rocks of Southern West-land, Trans. N.Z. Inst., vol. 63, pp. 178–284.

Ulrich, G. H F., 1890. On the Discovery, Mode of Occurrence, and Distribution of the Nickel-Iron Alloy Awaruite on the West Coast of the South Island of New Zealand, Q.J.G.S., vol. 46, pp. 619–632.