The Basic Igneous Rocks of Eastern Otago and their Tectonic Environment. Part IV.—The Mid-Tertiary Basalts, Tholeiites and Dolerites of North-Eastern otago. Section B:—Petrology, with Special Reference to the Crystallisation of Pyroxene.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 74, 1944-45, Page 71

The Basic Igneous Rocks of Eastern Otago and their Tectonic Environment. Part IV.—The Mid-Tertiary Basalts, Tholeiites and Dolerites of North-Eastern otago. Section B:—Petrology, with Special Reference to the Crystallisation of Pyroxene. By W. N. Benson. [Read before the Otago Branch, November 19, 1942; received by the Editor, November 15, 1943; issued separately, June, 1944.] Introduction. Descriptive Petrology. (i) The Pyroclastic Rocks.   (a) The Waiarekan Tuffs.   (b) The Deborah-Kakanui Tuffs and Breccias. (ii) The Massive Igneous Rocks.   (a) Mineralogical Features.     Plagioclase.     Anorthoclase.     Monoclinic Pyroxene.     Probability of error in Composition-determination.     Classification of Clinopyroxenes in dolerite and basalt.     The Trend of Clinopyroxene-differentiation.     Orthorhombic Pyroxene.     Hornblende.     Iron Ores.     Biotite.     Apatite.     Quartz and Chalcedony.     Zeolites.     Chlorophaeite and Carbonates.     Glassy Material.   (b) Pillow Lavas and Associated Basalt Flows.   (c) The Dykes of Moeraki Peninsula.   (d) The Tawhiroko Intrusive Dolerite Sheet.   (e) The Main Moeraki Dolerite Sheet.   (f) Other Mid-Tertiary Basalts, Tholeiites and Dolerites.     Porphyritic Basalts.     Olivine Dolerites.     Anorthoclase Bearing Olivine Basalt.     Analcite-bearing Olivine Dolerite.     Basic Analcite Syenite Pegmatoid.   (g) The Chemical Composition and Petrological Affinites of the Mid- and Late Tertiary Basic Igneous Rocks of Otago.   (h) The Absence of Pigeonite from N.E. Otago dolerites and its explanation.

Appendix I. Augite, Subcalcic Augite, Pigeonite, and the Classification of Non-alkaline Clinopyroxenes. Appendix II. The Crystallisation of Pyroxenes. By Dr. A. B. Edwards. Bibliography. List of Illustrations. Introduction. The work recorded herein has been aided by several gentlemen to whom the writer's thanks are due. Of the 120 sections studied (all in the collections of the Geological Department of the University of Otago), 14 were placed there by Dr. Marshall, 3 by Mr. O. D. Paterson, and two were from rocks obtained by Mr. D. A. Brown. Through the courtesy of the Directors of the Dominion Laboratory and of the Geological Survey, five excellent rock-analyses have been made by Mr. F. T. Seelye, A.I.C. Helpful comment on certain petrological points has been received from Dr. A. B. Edwards, of Melbourne, who has very kindly contributed an appendix to this paper, also Professor F. Walker and Dr. A. Poldervaart, of Capetown. Dr. C. O. Hutton has made many measurements of refractive indices. Dr. F. J. Turner has determined on the universal stage the composition of 96 feldspars and 27 olivines and the optic axial angles of 102 pyroxenes, and has most helpfully discussed various problems. But for his generous insistence it would have been appropriate to cite him as co-author of this section. Descriptive Petrography. I.—The Pyroclastic Rocks. (a) The Waiarekan Tuffs. No attempt has yet been made to study the petrographical features of these usually very decomposed rocks. Briefly they consist of moderately fine to coarsely granular material, often richly feldspathic, and contain more or less angular or rounded tachylitic or pumiceous fragments up to an inch or more in diameter, and larger lapilli of vesicular basalt or tachylite. Hutton (1887, pp. 417, 419; 1889, pp. 152–3) has described the more or less altered tachylite, “hydrotachylites,” or palagonite from the Mid-Tertiary rocks at Enfield (Teaneraki) and Lookout Bluff (Locs. 105, 84). (b) The Deboraii-Kakanui Tuffs and Breccias. These rocks, first noted by Mantell (1850) and most clearly developed near the mouth of the Kakanui River, were carefully studied by Thomson (1906, 1907), who recorded the presence of large isolated crystals of black or green augite, brown basaltic hornblende and (?) oligoclase, and smaller grains of garnet, diopside, diallage, smaragdite, biotite and olivine; also fragments of (a) more or less limburgitic olivine basalt, sandstone and limestone coming from the Cretaceous-Lower Tertiary formations, (b) argillites from the Late Palæozoic (?)—older Mesozoic rocks immediately beneath them, (c) quartz-mica-schist, granulite, and mica-gneiss from still deeper portions of the upper crust, and

(d) nodules of olivine-augite-spinel, olivine-hornblende-augite-spinel, olivine-augite-brown hornblende-garnet-spinel, augite-spinel, augite-spinel-garnet, augite-garnet with or without hornblende and magnetite, and hornblende-biotite-garnet probably derived from still greater depths in the crust. Most of these types of rock-fragments and nodules were also found by him as xenoliths in basalt. Thomson (1907) compared the last group (d) of rock-fragments and the isolated grains of minerals in the Kakanui tuff with those in the kimberlite breccia of South Africa, or in the ultra-basic rocks of the Ariege complex in the Pyrenees. The mass of breccia mapped as solid lava by Park (1918), which rises slightly above the general surface of the surrounding Waiarekan tuff a mile east of Round Hill (Loc. 101), probably marks the site of a vent from which Deborah-Kakanui tuff was erupted. It contains a variety of fragments of rocks and minerals, including tachylite (5715), altered olivine dolerite with strongly zoned plagioclase (An50–25) (5715a) and interstitial micropegmatite, which must have been derived from an underlying sill, of greywacke from the immediate basement rocks, and of coarsely crystalline lherzolite (5099) in which, as in that in the Kakanui breccia (6644, 6655), the olivine has been partially replaced by carbonates. Turner (1942) has demonstrated the general lack of marked parallelism of the olivine grains in the fabric of the last rock. II.—The Massive Igneous Rocks (Tachylites, tholeiites, basalts and dolerites). (a) Mineralogical Features. Plagioclase. Plagioclase is the dominant mineral. Albite and Carlsbad twinning is commonly associated with pericline twinning and much less so with the twinning on the Ala law. Twinning on other laws has been but very rarely recognised. No attempt has been made, however, to study the occurrence of twinning and to form a judgment as to whether the relative abundance of various types of twinning is a feature of the chemical composition of the feldspar (cf. Barber, 1936) or of the geological mode of occurrence (cf. Chapman, 1936). While phenocrystic tabulae occur in the tholeiites and basalts the tendency towards porphyritic development is diminished or lost in the dolerites, though subordinate interstitial relatively fine-grained plagioclase may occur in them. The larger feldspars are usually free from inclusions other than iron ores and apatite, but in rocks of medium grain-size they are often moulded on or include pyroxene, though the contrasted subophitic relation is more common and very perfect ophitic structure is displayed, especially by rocks in the lower half of the larger sills and in large dykes. Rarely (5878) the outer portions of large plagioclase tabulae contain isolated patches of optically continuous, graphically intergrown augite fraying out into shreds passing through the margin of the plagioclase into the interstitial cryptographic feldspar. Except in abnormally coarse-grained rocks (see Table VI) the larger plagioclase crystals in tholeiites, basalts and dolerites are tabulae seldom greater than 2.5 × 0.7 × 0.5 mm. and more often

about half that size. Rare instances of early formed crystals, protected from subsequent reactions by inclusion within pyroxenes, may be as calcic as An70–68. More commonly the unzoned or slightly zoned crystals are about An57–31, while those which are zoned have a core of about An62–60 occupying the greater part of the crystal and a thin margin of about An28. Where zoning is more marked the margin may be as sodic as An20. Usually, however, before such alkalinity is attained there has occurred a separation of relatively fine-grained interstitial taxitic, trachytic, suborthophyric or cryptographic patches, the plagioclase in which is about An40–32 with still more sodic rims. Plagioclase fraying from the marginal zones of phenocrysts out into the fine-grained base has been found in one case (6790) to be about An40–38. Very commonly there is no sharp distinction between earlier and later-formed crystals but a seriate transition through a series of not very zoned tabulae (An60–40). Dr. Turner's study of plagioclases in these dolerites showed that their determinative points usually fell nearer the standard Nikitin (1936) curves than did those of the plagioclases in such late Tertiary basic igneous rocks in the Dunedin district as the Figure 1. 1. Dispersion from the standard curves of Nikitin (1936) of determinative points for Otago plagioclases as measured by Dr. F. J. Turner. Black triangles indicate the positions of determinative points in the mugearites of the Dunedin district, small black circles those of the plagioclases in the Mid-Tertiary dolerites and tachylite of N.E. Otago. These points indicate:—(a) the directions (points of emergence) of the twinning-axis ⊥ [001] / (010) in Carlsbad-albite twins; (b) the directions of the twinning axis [001] in simple Carlsbad twins; and (c) the directions of the normals to the (001) planes in simple albite twins, each point being referred to the symmetry-axes α, β, and γ of the optical indicatrix of the portion of the crystal examined. The dispersion from the standard curves of these determinitive points is shown by their perpendicular distances (thin black lines) from the corresponding curves.

mugearites (Benson and Turner, 1940, pp. 188–199 and fig. 2), though exceptions occur. The general problem involved has received much attention within the last twenty years. Omitting many details we note that Barth (1931, pp. 51–72) investigating albite, oligoclase and labradorite, concluded that the last but not the first two after being heated for 300 hours at 1000° C. showed a displacement of the pole of (010) to an extent estimated by Barber (1936a, p. 351) from Barth's graphical data to be as much as 6°. During a comprehensive study of a wide range of intrusive and volcanic Tertiary igneous rocks in Lower Burma, Barber (1936, pp. 221–258) noted the general dispersion of the poles of plagioclases twinned on various laws along a narrow strip covering one or both sides of the standard Reinhard-Nikitin curves for those laws, and discussing with a wealth of citations the possible effect on such dispersion of observational errors, inaccuracy of standard curves, vicinal deflection of the crystallographic directions of reference, twinning and cleavage planes, etc., and the presence of impurities, notably potash, or physical factors, especially temperature, during crystallisation, concluded that the last has probably by far the greatest significance, the others only slight importance. In his subsequent experimental work on the heating of andesine and labradorite (1936a, pp. 343–352), however, he was unable to confirm Barth's results, and was lead to conclude the variation of the optical properties of plagioclase on heating is very slight and “far too small to be of any value as a geological thermometer.” More recently this judgment has been reversed by Continental workers—Köhler (1941, 1942), Scholler (1942) and Tertsch (1941, 1942), who have shown with a wealth of experimental data the necessity of distinguishing between high temperature and low temperature plagioclase. As but brief abstracts of these papers are available at present, tentative comments are withheld. The plagioclases in the Mid-Tertiary rock are but little altered even when enclosed in deuterie carbonates. No clear evidence of albitisation has been noted, though albite may occur in slight amount among the obscure deuteric alteration products of two analcite-bearing rocks, in which, also, little veinlets of this zeolite have been formed in the plagioclase. Anorthoclase. Potassic feldspars can be detected in the general groundmass of a few minutely granular rocks. The normative composition of the feldspar in the five analysed rocks are as follows:— Table I. Normative Feldspar in Mid-Tertiary Basic Igneous Rocks. Rock Type Locality Or. Ab. An. 5734 Porphyritic olivine basalt 97 12 52 36 5741 Porphyritic olivine basalt 87 14 49 37 5701 Tachylite 104 9 48 43 5750 Coarse dolerite pegmatoid 83 13 58 29 5860 Ophitic olivine dolerite 83 7 50 43 In Barth's (1936, p. 323) view potassic feldspar must form a phase separate from plagioclase in rocks in which the normative feldspar

contains over 7% of orthoclase, the approximate limit of its solubility in plagioclase. It is noteworthy that untwinned grains of a feldspar with 2 V = 40° (-), which is probably anorthoclase, occurs in rocks (5734) and (5741), and are abundant as idiomorphic tabulae in (5749) analcite bearing olivine dolerite near Maheno (Loc. 96). Anorthoclase is most abundant in what seems to be a pegmatoid segregation (5865) in the same intrusive sheet as that containing the above three rocks—namely, at Waimotu (Loc. 97A). Here it is the dominant constituent of this rock (a basic analcite or melasyenite) and forms Carlsbad twinned tabulae 20 × 1.5 mm. in area, with the shadowy extinction or fine multiple twinning and optic axial angle [2 V = 42°-47° (-) in the plane ⊥ (010)] characteristic of anorthoclase. In other rocks the alkaline feldspar, if present, is confined to the intergranular spaces either as small tabulae associated with laths of oligoclase-andesine, in trachytoid or orthophyric segregations, or as more or less radiating fibres fraying out into micro- or cryptographic intergrowths, possibly formed by devitrification (cf. Walker and Poldervaart, 1941, p. 136), or into interstitial glass and sometimes extending in optical continuity into a thin zone mantling plagioclase tabulae. Monoclinic Pyroxenes. In view of the significance of pyroxene composition in relation to the magmatic affiliation and differentiation of basic igneous rocks special attention was given to the determination of the optic properties of pyroxenes in the granular intrusive dolerites rather than the phenocrystic pyroxenes of porphyritic rocks which Hess (1941, pp. 520, 533) holds to be by comparison therewith rather abnormal. For this purpose 102 measurements of optic axial angles on the universal stage were made by Dr. Turner, and 32 measurements of refractive indices by Dr. Hutton. As noted by Dr. Turner (1942) measurements of the extinction angles (γ ∧ c) of dolerite-pyroxenes by the ordinary petrographic procedure are subject to considerable error, even when the orientations of the sections examined are shown by observations in convergent light to be favourable for this purpose. Consequently only such extinction angles as have been measured by the most accurate method (Nemoto, 1938) in twinned crystals on the universal stage have been considered. These are very rare among our rocks and only four angles have been thus measured by Dr. Turner. Following Barth's (1931b) assumption that, apart from minor amounts of Al2O3, Fe2O3, Tio2 and the alkalies which may be present, the essential composition of the pyroxenes may be stated in terms of CaSiO3 (= Wo), MgSiO3 (= En), and FeSiO3 (= Fs), such compositions have been deduced in the sequel for 45 pyroxenes, some of which are given on Tables IV and VI and plotted on Figures 4, 5 and 6, are stated in terms of molecular percentages in order to confirm with the normal usage of the minals Or. Ab. An. for feldspars. A similar use of molecular proportions is employed in the sequel for the constituent minals Fo and Fa in olivine. This differs from the practice of the Japanese petrologists—e.g., Kuno (1936) and Tsuya (1937), also of Winchell (1927, 1933), and Deer and Wager (1938, 1939), who use weight-percentages in stating the composition of these ferromagnesian minerals. In order to obviate the necessary recalcula-

tion the standard diagram expressing the relationship of chemical composition and optical properties given by Deer and Wager (1938, fig. 2A), based on weight percentages, has been here redrafted (fig. 2B) to show molecular percentages of the three constituent minals. Fig. 2.—Relationship of Chemical Composition to Optical Properties of Clinopyroxenes. A. Curves of 2 V, extinction angle γ ∧ c, and of refractive index γ after Deer and Wager (1938), also curves of 2 V for titaniferous augites after Kuno (1936) plotted against a background of divisions showing weight percentages of Wo, En and Fs. B. Curves of 2 V and of refractive indices α and γ, after Deer and Wager (1938) plotted against a background of divisions showing molecular percentages of Wo, En and Fs. Probability of Error in the Determination of Pyroxene-composition. (1) Chemical Analysis. Though optical data show the occurrence of a variety of pyroxenes in dolerites, either as separate mineral grains or associated in successive zones in individual composite grains, the differences between the physical properties, especially density, of the several varieties are commonly too small to permit the separation of homogeneous samples, though it has been done under favourable conditions—e.g., by Harris (1937) and Deer and Wager (1938). Hess (1941, p. 579) notes that the original analysis of pigeonite from the type locality is that of an indefinite mixture of augite and pigeonite, each mixed-crystals of variable composition and inseparable

by mechanical means. Where rhombic and monoclinic pyroxenes are associated, an estimate of the composition of the latter has been determined indirectly by Wager and Deer (1939, p. 76) by estimating optically the composition (2 V determinations) and relative proportions (Rosiwal estimations) of the rhombic portion of the analysed mixed pyroxenes, involving the uncertainties inherent in such optical measurements, but where two phases of monoclinic pyroxene are associated with rhombic pyroxene as in some Antarctic (Benson, 1916), Tasmanian (Benson, 1917; Edwards, 1942), American (Walker, 1940) and African dolerites (Walker, 1940; Walker and Poldervaart, 1941, 1942), chemical determination of the compositions of these several phases is at present almost impossible. Though these last complexities do not arise in the case of our North-East Otago rocks, and it is hoped that subsequently chemical analyses may be made of their pyroxene, for the present the estimations of their compositions have been based wholly on optical data. (2) Optical determinations. Various uncertainties have been involved in the estimation of the compositions of the pyroxenes from the standard diagram as recorded below, a statement of which is appropriate. (a) The effect of the presence of an unknown amount of minor constituents on the optical properties of the pyroxenes. Though Winchell (1935, p. 568) showed that variations of the amount of Al2O3 “has little effect on the optic axial angles, though a considerable effect … on the extinction angle” Kuno (1936) has indicated that the presence of TiO2 has a noteworthy effect on the optical axial angle, as indicated by comparison of the isogonic curves drawn by Tomita (1934) and Kuno (1936), the latter for augites containing moderate amounts of sesquioxides and titania occurring in Japanese basalts and andesites. (See Figure 2A.) Since for determative purposes Kuno employs the refractive index values of β, which are not available for our rocks, we here follow Wager and Deer, using Tomita's 2 V curves and those of R. 1's α and γ. (b) Though repeated measurements of 2 V in favourably oriented sections in Na light would have increased the accuracy of the figures obtained in the present petrographical reconnaissance, single measurements only were made in ordinary light with an average accuracy of ± 2°. Reference to Fig. 2A will show that this possibility of error affects chiefly the estimation of the content of Wo and involves an uncertainty of about ± 5% of the total amount of Wo indicated in addition to a general slight under-estimation of this constituent when the Wager and Deer curves are used as here for moderately titaniferous pyroxenes. No correction has been made for this in the estimated composition herein as there is as yet no means for estimating its amount. (c) The errors ± .003 in the measurement of the R 1's α and γ affect chiefly the ratio En: Fs and involve at most a possible error of ± 3% En or ± 3% Fs, which is halved since two independent determinations are available. The invariably

close composition-estimations derived from the values of α and γ taken singly attest to the accuracy of these optical determinations by Dr. Hutton on our pyroxenes. (d) Greater possibility of error may arise from the fact that as the refractive indices were determined in powdered minerals in no case were both R. 1's and 2 V determined for any particular grain. Since most of the zoned crystals examined show the normal outward decrease of 2 V, which in other rock-series has been shown to be associated with outward increase in the content of Fs and of the R. 1's, the tentative assumption has been made that where the powdered pyroxene in any rock containing zoned crystals shows variation in R. 1's the fragments with highest R. 1's are those with lowest 2 V and vice versa, and that the two compositions estimated from these correlations will show the maximum possible (not necessarily actual) range of composition in such crystals. Figures 3 and 4 have been drafted on this assumption. That it is not far from the truth is suggested by the slight dispersion of points on these two figures. (e) A further cause for uncertainty may affect our estimated compositions. Krokström (1936) has found that for unzoned pyroxenes in dolerites the optical properties (both 2 V and R.1), and accordingly the chemical compositions, are not uniform, but, as a result of differing rates of diffusion during crystallisation, vary about a mean value in rough accord with a Gaussian curve of error. Variations in the optical properties of unzoned pyroxenes in some of our rocks (e.g. 5880) may have this explanation, though it is difficult to make allowance for it in the case of zoned pyroxenes. The asymmetric frequency-curve for all determined values of 2 V (Fig. 3) is clearly not Gaussian. It is not possible, therefore, to estimate the error that may be involved in the tentative assumption that the various values of the optical properties and deduced compositions of pyroxenes in our rocks represent, not merely such random variations, but stages in the course of pyroxene-crystallisation. It will be noted, however, that such tentatively inferred-differentiation-trends as shown in Fig. 5 are in general in the same direction as the so-called “trends” indicated by the comparison of the compositions of phenocrystic and ground-mass pyroxenes in Japanese (Tsuya, 1937) and other basaltic and andesitic rocks (Barth, 1936). (f) Determinations of chemical composition based on the extinction angle γ ∧ c and 2 V, using figure, 2A have proved, as elsewhere, to be usually very untrustworthy even when, as here, only extinction angles measured by Nemoto's (1938) relatively accurate method. Turner (1942) and others have shown the frequent inaccuracy (up to 5°) of the normal petrographic method of measuring this angle in clinopyroxenes. Apart from this, the usefulness of this method is

affected by the sensitivity of the extinction-angle to variation in the amount of minor constituents in the pyroxene (cf. Wager and Deer, 1939, p. 80), and its equivocal indication for parts of the pyroxene field when only two of the three forms of optical constants are available. The following table (II) gives comparative results for certain Otago pyroxenes:— Table II. Contrasted Estimates of Pyroxene-Composition by Different Methods. Composition. γ ∧ c Composition. Rock 2V R.1. Wo. En. Fs. (Nemoto) Wo. En. Fs. 5869 43° α = 1.698* Probable values of refractive indices estimated by comparison with those of pyroxenes in closely analogous rocks. γ = 1.721* 23 50 27 42 ± 1° 25 58 17 5880 48° α = 1.692 γ = 1.718 29 49 22 41 ½° 31 55 14 6788 47° α = 1.685* γ = 1.712* 28 50 22 42 ½° 30 54 16 6790 55° α = 1.685* γ = 1.708* 39 46 15 42° 39 47 14 As the composition of the pyroxene in (5860), which is optically nearest to that in 6788 above, is normatively Wo21 En62 Fs17 and, as the rock contains olivine, it must be modally richer in Fs than the above, the tendency of composition-determinations based on the extinction angle to under estimate the content of Fs seems to be indicated. 3. Relationship between optic axial angle and refractive indices in clino-pyroxenes in N.E. Otago dolerites, etc., also the frequency of occurrence of various values of 2 V. 4. Range of optical properties and deduced trend of composition-variation in clinopyroxenes in N.E. Otago dolerites, etc., expressed in molecular percentages of Wo, En and Fs,

While, however, making the above discussed tentative assumptions in this petrographic reconnaissance, an approximation towards a knowledge of the range and differentiation-trend of these Otago pyroxenes may have been gained, the need for more exact observations is obvious. Walker and Poldervaart's (1941, p. 144) conclusion, however, is that “both 2 V and still more γ ∧ c are singularly unreliable guides to the composition of pyroxene”. An instance of this may be cited—the carefully investigated phenocrystic augite of the picritic basalt of Haleakala, Maui, Hawaiian Islands (Washington and Merwin, 1922). Apart from 8.5% of sesquioxides, 1.9% of TiO2 and 1% of the acmite molecule, its essential constituents are Wo50 En42 Fs8 as shown by chemical analysis. The composition estimated optically using 2V and R.1. on Kuno's (1936) diagram is Wo50 En28 Fs22, and on Deer and Wager's (1938) (Figure 1B) 50:30:20, but that deduced from 2 V and γ ∧ c by Figure 2A herewith is 50:8:42. The Classification of Clino-pyroxenes in Dolerite and Basalt. This has been complicated by the widely varying usage of the term pigeonite, and is discussed in Appendix I to this paper, wherein is suggested the term subcalcic augite for those pyroxenes intermediate in composition between the most calcic pigeonites as commonly defined (i.e., with 2 V = 45°) and those with 2 V < 32° (or 30°) to which Hess (1941) and Walker (1941) would confine the term pigeonite. Whether the maximum optic axial angle (2 V = 45°) suggested for the subcalcic augites will prove the most appropriate remains to be seen. It is here selected as coinciding with the limit of “pigeonite” used in its unduly extended sense by most writers. (See Table X.) With the definitions herein adopted the pyroxenes present in the Otago Mid-Tertiary rocks are rarely diopsidic (except when phenocrystic), dominantly normal augite probably containing moderate amounts of titanium and sesquioxides, with noteworthy but subordinate amounts of subcalcic augite with margins approaching the composition of pigeonite, but true pigeonite has not yet been observed in any rock studied. No indication of the presence of exsolved lamellae of orthopyroxene could be found in the clinopyroxenes of N.E. Otago, as is generally the case in rather quickly chilled rocks (cf. Hess, 1941, p. 526). Table V shows the inferred ranges of pyroxene composition in the Tawhiroko intrusive sheet and Table III those in other rocks described herein. The pyroxene-compositions based on the observed optic anxial angle and the asterisked refractive indices inferred therefrom by Figure 2 are naturally less significant than those based on observation of axial angles and both refractive indices. In view of the pigeonitic character of the average normative composition in these N.E. Otago rocks the absence of modal pigeonite calls for discussion which is given on later pages (109–111). The Trend of Pyroxene-Differentiation. Figure 5 shows the range of composition of the pyroxenes examined plotted from the data given in Tables III and V. To this figure has been added the so-called trend-lines obtained by Barth (1936, p. 327, fig. 2) by connecting the points representing the compositions of phenocrystic and ground-mass clinopyroxenes in basalts from the Pacific, Deccan, Cape

Table III. Optical Characters and Inferred Compositions of Pyroxenes in MidTertiary Igneous Rocks of N.E. Otago other than those of Tawhiroko Sheet. Siide No. Letter on Fig. 3. Locality. Rock Type. Optics 2 V and γ ∧ c. Refr. Indices. Composition. α γ Wo En Fs 5860 A Moraki Main Sheet Olivine-bearing but over-saturated dolerite Large Cryst. 52°, 51°. Small Cryst. 52°, 51°, 49°, 47°, 44°. * Refractive Indices inferred from 2 V by Figure 3. 1.684 1.711 34 49 17 * 1.691 1.719 24 50 26 5869 B Moeraki Main Sheet Micropegmatitic dolerite without olivine 56°, 54°, 45°, 43° (γ ∧ c = 42°), 36°, 36°, * 1.680 1.706 41 48 11 * 1.699 1.726 19 48 33 5872 C Moeraki Main Sheet Coarse olivine-free dolerite with glass 56°, 56° → 44°. 56° → 38°, 46° → 50°,§ The only case of “reversed zoning” observed. (For an explanation of the occasional occurrence of “reversed zoning” see Krokström, 1936, pp. 158–9.) * 1.680 1.690 1.716 1.716 41 48 11 5878 D Moeraki Main Sheet Micropegmatitic olivinebearing dolerite 60°, 50°, 48°, 46°. * 1.677 1.703 50 45 5 * 1.689 1.714 27 50 23 5880 E Moeraki Main Sheet Coarse micropegmatitic quartz dolerite 54°, 53°, 52°, 48°, 48°, 47° (γ ∧ c = 41 ½°), 46 ½°, 44°, 40°. 1.683 1.711 38 48 14 † Commonest values. 1.692 † 1.718 28 48 24 1.698 1.725 21 49 30 5716 F North Peak Richly olivinie dolerite 51°, 50°. * 1.686 1.713 33 49 18 5727 G Mount Charles Micropegmatitic dolerite with olivine 54°, 47°, 46°, 43°. * 1.683 1.709 38 49 13 * 1.692 1.719 23 50 27 6788 H Mount Charles Micropegmatitic dolerite with olivine 49°, 48°, 42° (γ ∧ c = 42–3°) * 1.687 1.713 30 50 20 * 1.693 1.721 23 50 27 6790 I Mount Charles Richly pegmatitic dolerite 61–3° (γ ∧ c = 42°). 56°, 54°, 52°, * 1.679 1.703 50 42 8 * 1.685 1.708 34 48 18 Waimotu Anortioclase 57°. * 1.680 * 1.705 45 45 10 5865 X Loc. 97A Syenite pegmatoid

5. Trends of pyroxene-differentiation (or variation) in N.E. Otago dolerites, etc., compared with those suggested by the differences between compositions of phenocrystic and groundmass clinopyroxenes in Japanese basalts and various other basaltic rocks. Points joined a line representing the compositions of phenocrystic and groundmass pyroxene in the same rock. General trend-lines of pyroxene differentiation deduced by Deer and Wager (1938) and Hess (1941) and the suggested boundaries of the fields of variation of certain pyroxenes (slightly modified) after Hess (op. cit.) and subdivisions according to molecular percentages of Wo, En and Fs are also indicated. Verde Islands and Iceland, and also similar “trend-lines” for clino-pyroxenes in Japanese basalts and andesites based on data recorded in Tsuya's (1937) excellent and comprehensive work, which contained some of Kuno's (1936) observations. In addition the figure shows the general trend of pyroxene-differentiation inferred by Wager and Deer (1939) for ferriferous gabbros in the Skaeargaard, E. Greenland, and by Hess (1941) for basic igneous rocks in general. It will be noted that with the exception of the calcic portion of Hess' diagram (cf. Edwards, 1942, p. 600, fig. 17) these two generally arcuate trend-lines are approximately parallel to those shown by Barth's, Tsuya's, and our pyroxenes. The last show a further point of interest. Though Hess excluded the phenocrystic pyroxenes from his general discussion of pyroxenes on the ground that they were abnormal in their highly calcic composition and in other ways, it is doubtful whether, in the general absence of phenocrystic pyroxene from the tachylitic margins of dolerite sheets, we may suppose the invading magma contained such plutonic crystals. The strongly calcic pyroxenes D and I (Fig. 4) in the coarsely crystallised micropegmatitic dolerite of the Main Moeraki sheet and Mount Charles may, perhaps, therefore, have been wholly autochthonous. The early differentiation-trend as the lime-content falls from Wo50 to Wo40 involves a replacement of Wo by En and Fs in the proportion of about 1:2. Beyond this Wo is replaced by Fs only, the composition being expressible approximately by Wox Fs(51) En(40)x) over the range x = 35 to 18, En remaining approximately constant (cf. Barth, 1931, p. 208). The replacement of both Wo and En by Fs beyond this limit is indicated by Wager and Deer's and Hess' curves. Dr. A. B. Edwards has discussed the significance of clinopyroxene trends in terms of ionic radii and lattice structure in an appendix (II)

kindly contributed to this paper. The explanation of the trend exhibited in our rocks is, as Edwards indicates, probably the rate of cooling, the production of metastable clinopyroxenes in the more rapidly cooled rocks such as our relatively thin intrusions (see p. 119) and the volcanic rocks, while with slower cooling and greater approach to molecular equilibrium the trend towards concentration of iron sets in earlier during “plutonic” crystallisation. The clinopyroxenes are generally free from decomposition though locally partially replaced by carbonates and chloritic minerals, a process affecting finely granular ground-mass pyroxenes more than phenocrysts. Only rarely are the larger pyroxenes partially replaced by deep green chlorite as in two olivine dolerites (5887–8) and the Waimotu syenite-pegmatoid (5865). There is no clear evidence of deuteric uralitisation. (See below.) Orthorhombic Pyroxene. Orthorhombic pyroxene referred to enstatite (though displaying the pleochroism of hypersthene) has been noted by Hutton (1887, p. 428) in the dolerite of Mt. Charles, but has not been observed in slides from there examined by the writer. Apart from this occurrence, rhombic pyroxenes are not present in the normal tholeiites basalts or dolerites; nor could there be recognised in any of these features suggesting the presence of remnants of orthpyroxene surviving incomplete transformation to clinopyroxene. Hypersthene does occur, however, in basic rocks which have absorbed silica from quartzose xenoliths as Lacroix (1893, p. 34, Harker (1904, p. 353), and others have noted, and instances will be noted in the discussion of endo- and exomorphous metamorphism at the margins of intrusions in Section C of this paper. Whether or not this is the explanation of Hutton's observation remains to be determined. Hornblende. Hornblende has been found in a quartz-dolerite (5891) where it occurs either in short (0.4 × 0.3 mm.) or long (2.0 × 0.2 mm.) prisms, or in small scattered fibres. Its pleochroism varies, pale to stronger yellowish brown tints occurring in the course of the larger grains, pale to stronger yellowish to bluish green in the outer portions, and in the fibres the latter is a common though not invariable rule. Occasionally, where greenish fibres cross a band of limonitic decomposition products, they assume a brown coloration, and the same may hold in the proximity of the iron ores, but exceptions occur here also. As the pyroxene in this rock is perfectly fresh, without sign of uralitic alteration, and in the two cases where a close association of pyroxene and amphibole has been noted, the vertical axes of the two minerals are oblique to one another, it seems possible that the latter is the result of independent crystallisation, though its composition has been affected by deuteric processes. The variability of its colour accords with what is usual under such conditions (cf. Walker, 1941, p. 1076). It contains apatite, but is enclosed in feldspar, into which, however, may penetrate tangled or sub-radiating fibres of amphibole distributed by deuteric processes.

Olivine. Olivine is usually the most abundant and often the only coloured phenocryst in porphyritic rocks. It forms in these crystals up to 3 mm. long though usually less than half this size. In the hypocrystalline rocks it is usually idiomorphic, in those with a basaltic matrix it may be more or less corroded, with (5741) or without the development of a reaction-mantle of minutely granular augite. As is commonly the case, the larger phenocrysts are usually richly magnesian and slightly zoned, though the smaller phenocrysts (5741) may show marked zoning (cf. Walker, 1941, p. 1068). Sometimes a second generation of ground-mass olivines has been formed. The more coarsely granular olivines of the dolerites have sometimes broad zoning with a small range of composition. Comparison of normative and average modal composition of the olivine is afforded by the following table:— Table IV. No. Rock Character. Norm. Mode. 5734 Porphyritic sparsely olivinic basalt Fa31 Fa3 (Average) 5741 Porphyritic richly olivinic basalt Fa26 Fa23 (Average) 5860 Slightly over-saturated olivine dolerite Nil Fa25 (Average) The olivine in (5734) has a nearly uniform composition, two unzoned crystals being Fa0 and Fa8. A zoned crystal has a core of Fa2 with a thin mantle of Fa13. Olivines in (5741) are more variable; wide zones show core → margin Fa10–24, Fa28–38, and Fa33–50; narrow zones show Fa5–13, and in an unzoned crystal Fa28. In both these rocks the average modal olivine is (as usual) more forsteritic than the normative. In (5860), which contains normative quartz, comparison can be drawn only between the Mg/Fe ratios in the modal olivine (Fa16, Fa30, Fa30) and in the normative hypersthene (Fs21.5). The gravitational concentration of coarsely granular olivine into the lower portions of intrusive sheets and the retention of only a little of the slowly sinking smaller grains in the upper portion is well displayed at Tawhiroko Point (Fig. 6) where, however, complete decomposition prevents any estimate of the original composition of the olivine. The most abundant decomposition product of olivine is weakly pleochroic rather strongly coloured brownish-green bowlingite, occasionally associated with pale green chrysotile, and less often so strongly coloured by absorbed iron oxides that it resembles (and perhaps is) iddingsite. With these silicates are varying proportion of carbonates. In some cases these are in radiating aggregates (sphaerosiderite) or more coarsely crystalline masses of siderite which on weathering form massive or pulverent haematite. In others calcite may be associated with the siderite. Both of these may occur with varying amounts of chalcedony or quartz as a finegrained mosaic or plainly pseudomorphous after chrysotile, and with these residual bowlingite remains. Some replacement of olivine by talc has occurred in a few rocks. Rarely (5878) talc forms much of the pseudomorph.

Iron Ores. Ilmenite is often the more abundant iron ore. In the more coarsely granular rocks it forms thin plates (2.5 × 0.1 — 0.3 mm.), though usually it is proportionately smaller. It is moulded on olivine and rarely on pyroxene and feldspar, with the former of which it may be intergrown. More often it is idiomorphic against both pyroxene and feldspar. In places a group of parallel plates of ilmenite may extend through the rock for several millimetres, to five times the length of any one plate in the group. Rarely a thin reaction-mantle of biotite covers the ilmenite where it extends into micrographic intergrowths or among deuteric minerals. In a few more altered rocks leucoxenisation of the ilmenite is seen as Hutton (1887, p. 428) noted at Mt. Charles, and the presence of lamellar intergrowths of titano-magnetite becomes observable (5891). In some coarsely granular dolerites, but more often in those of medium to fine grain size, approximately octahedral magnetite is abundant, and rarely minute octahedra occur in the phenocrystic olivine. In glassy material, and among the feldspathic intergrowths developed therefrom, minute plates and skeletal growths of ilmenite and tiny octahedra of magnetite, either isolated or in dendritic aggregates, may be found, and a frequent feature is the development of minute magnetite octahedra studded at intervals along the augitemicrolites. In general the proportion of iron ores to ferromagnesian silicates increases with the feldspathic content of the rocks and towards the middle of intrusive sheets, but only a qualitative significance in this regard can be attached to Fig. 6. Barth's (1931, p. 393) comment, “based on textural features, that under certain circumstances the ore minerals crystallise early, but usually relatively late, though not as late as the bulk of the alkali feldspars and quartz,” seems to hold good. Phemister (1934, pp. 40–44) has argued that the state of oxidation of iron in magmas may be affected by the reversible reaction 3FeO + H2O ⇌ Fe3O4 + H2 + 15,400 calories, and that since hydrogen can escape more easily than water-vapour through the pores of wall-rocks, the ratio Fe3O4:FeO, and, therefore, the amount of magnetite developed proportionately to that of other ferriferous minerals crystallised will tend to be increased by the exothermic reaction during the later stages of magma-cooling, when the expulsion of gas is most rapid. The pyroxene formed concommittantly therewith will accordingly tend to contain En in greater proportion to its Fs than would occur if no such oxidation took place. This process may be exemplified to some extent by the rocks, especially (5855) and (5750) of the Tawhiroko sill (Fig. 6) as well as in those cited by Phemister. Biotite. Biotite is rarely developed save as a product of reaction between iron ores and micropegmatite. It is, however, present in small (< 0.05 mm.) rare flakes in a few of the more potassic rocks—e.g., the anorthoclase-bearing basalts (5734, 5741), and more abundantly in analcite-bearing dolerite (5749) occurring a mile west of Maheno, and very abundantly in a coarsely granular basic anorthoclase analcite syenite-pegmatoid (5865) at Waimotxi, two miles S.S.W. thereof (Locs. 96 and 97A respectively). In the latter rock it has been largely replaced by chlorite.

Apatite. Apatite, generally present in small amounts only, occurs in thin needles in the feldspars, and is chiefly concentrated in the interstitial crypto or micrographic material. There is a noteworthy difference between its scantiness in the Tawhiroko Sheet and its greater abundance in the upper portions of the main Moeraki Sheet, both in the larger feldspars (where the needles may be 1 mm. long) and in the interstitial material. It is especially abundant in the analcite syenite. In the more finely granular rocks it occurs in small amount only, and is not always recognisable. Quartz and Chalcedony. Quartz occurs chiefly in interstial intergrowths, and its presence there is inferred chiefly from its refractive index. It is rarely present in the Tawhiroko rocks in granules large enough to display other determinative properties, though occasionally the quartz fibres in the intergrowth may be traced out into characteristic granules. Its concentration into the part of this sheet immediately below the upper chilled phase accords with experience elsewhere (cf. Walker, 1941, p. 1075). This is much more evident in the higher rocks of the thicker main sheet, where it appears in several forms; (a) as irregular grains up to 0μ1 mm. diameter between feldspar tabulae not directly associated with interstitial intergrowths, though sometimes continuous with the quartz fibres therein, or (b) as direct continuations of such fibres extending towards the central portions of the larger interstices and there uniting into larger grains with crystal-boundaries particularly obvious in the case of hexagonal basal sections against which the ferruginous carbonates are moulded. Such more or less idiomorphic quartz may include small apatite needles and rarely anorthoclase (?) microlites, and may only occasionally show marginal undu-lose extinction. Where, however, a later generation of quartz occurs separated from the former by ferruginous carbonates or “chlorite,” undulose extinction is generally observable, and in some rocks (e.g.; 5890) the quartz is divided into radial sectors each having its vertical crystallographic axis parallel to its length. Chalcedony may rarely (e.g., 5878) form a very thin mantle on quartz with fibres growing perpendicularly to the prism faces which it separates from the surrounding carbonatés. More commonly, however, it occurs in radially fibrous nodules within such carbonates (as in 5869). In other cases (e.g., 5722) chalcedony forms irregular patches replacing interstitial glass, and formed, as Fenner (1931) indicates, as a by-product of the change of glass to chlorophaeite. Probably as Tomkieff (1941) notes the chalcedony and the radial and undulatory quartz first consolidated as colloids. Rarely a little opal is still observable. In addition to these forms of silica there are occasionally thin or irregular veins of finely granular calcite and chalcedony traversing the massive dolerites, and probably derived from an extra-magmatic source—namely, the invaded marly sediments (cf. Leitmeier, 1909). The Otago dolerites do not seem to afford any clue to the problem of the magmatic or deuteric origin of the micro- and cryptographic interstitial quartz-feldspar intergrowths. The observed features appear consistent with the former view, though Fenner (1926, 1931)

held that it is hardly safe to assume such an origin, and that these intergrowths are often the result of post-magmatic replacement of feldspar, etc., with which Walker (1940, p. 1093), modifying his (1930) former view, is now in agreement. Zeolites. In the anorthoclase syenite-pegmatoid (5865) of Waimotu analcite forms abundant intergranular masses (< 2mm. in diameter) moulded on the feldspars, and small, almost rectilinearly bounded patches surrounded by chlorite which suggest an originally idiomorphic development. It is associated with a little carbonate which also occurs idiomorphically in radiating spherules of natrolite which, together with analcite, fills one of the vesicular (?) spaces. Analcite also replaces feldspar to a slight extent. It is rare in other rocks, though forming sparse, small (< 0.5 mm.) patches or thin veinlets in plagioclase in the more finely granular part of the same intrusion extending to Maheno (5749). A little stilbite (?) seems to be present in the olivine basalt of Lookout Bluff (5752). Chlorophaeite and Carbonates. Chlorophaeite and other chloritic minerals range from very pale green finely flaky or fibrous material with weak birefringence and pleochroism [delessite (?) or “green chlorophaeite” of Peacock (1930) and Fenner (1931)] to dark greenish yellow or brown or more deeply coloured brownish ferriferous aggregates—both weakly birefringent or isotropic chlorophaeite or possibly diabantite where strongly birefringent (Shannon, 1920, 1924). The less deeply coloured but strongly birefringent mineral is possibly chlinochlore. These substances occur with radiating fibrous carbonates, usually coating the surfaces on which the latter were subsequently deposited, though they may also occur coating such carbonates. Such variations of the order of deposition of the pale or highly coloured hydrous silicates are to be seen both replacing olivine, groundmass-augites or glass, and in vesicles.* Toinkieff (1942, pp. 10–12), citing comments by Loewinson-Lessing, Backlund and Sazonova, has suggested in regard to the trachybasalts of Rum, which contain amygdaloidal portions rich, in silica minerals, chlorophaeite, celadonite and carbonate, that these rocks are possibly solidified macro-emulsions which show incipient stages of separation into possibly immiscible “dry” and “wet” fractious. “The wet fraction on further cooling gave rise to the amygdaloidal material, which, as judged by the minerals developed in it, such as opal and chlo ophaeite, was of the nature of a gel.” He notes, however, that the hypothesis does not apply “to all amygdaloidal lavas, which in many cases may be formed by the infilling of vesicles by latemaginatic or hydrothermal products.” It seems most likely that the rocks here described fall into the latter group. The deuteric carbonates are also varied in nature and position. The earliest formed were usually siderite, replacing ground-mass—or rarely phenocrystic augite, but more commonly glass. They also occur with the interstitial cryptographic aggregate, and are so abundant that few rocks can be chosen as appropriate for chemical analysis or density-determination. Both granular and radiating patches occur often stained scarlet by oxidation (e.g., 5750). Where calcite or aragonite is present it usually crystallised after the fer-

ruginous carbonates perhaps because of the greater solubility of CaCO3 (cf. Tomkieff, 1941, p. 56). Some vermiform masses of radial siderite (?) are present in large calcite crystals replacing either olivine (?) or the general interstitial matrix. Glassy Material. Glassy material occurs in several modifications. At the chilled margin of the pillow lavas (e.g., 5701, Analysis No. 1, Table VII) it is pale yellow and transparent, with R.1. 1μ5772 ± 0μ0002 and sp. g. 2μ725, figures which, according to the investigations of Tilley (1922) and George (1924), are consistent with a content of 51–53% of SiO2, agreeing well with the analytical result 52μ92%. Since the density calculated from the norm (Washington, 1922, p. 387) is 2μ965, there will be a volume contraction of 8μ7% in completing the crystallisation of the glass which contains small labradorite laths and grains of forsteritic olivine. Similar glass is formed in the chilled margins of intrusions, and is either pale brown or almost colourless (5831, see Fig. 6, 5883), or so crowded with dustlike particles of iron ore as to be dark grey or almost black (5833, 5839, 5881). It may contain innumerable minute microlites of labradorite, some of which have skeletal extensions, or tabulae up to 0μ5 mm. long, together with small pseudomorphs after olivine and crystals of augite (2 V = 46° — 42° in 5831). Interstitial glass in the more finely granular olivine tholeiites is similarly either clear pale brown with or without dustlike or minutely platy iron ores locally in skeletal aggregates, and lathy or rod-like microlites of feldspar or augite, or it may be darkened by aggregations of particles of iron ore and pyroxene either enclosing the feldspars or occupying the whole mass of glass. In vesicular rocks, the melt, while stiffening around the vesicles and giving rise to crystallites of iron ores, feldspar and pyroxene as described above, may have been squeezed in droplets into the cavity of the vesicles as the vapour pressure therein diminished (e.g., 5871), the remaining space being filled later by deuteric minerals, or the vesicles may be completely filled by dust-darkened glass containing abundant plagioclase microlites (e.g., 5743). All stages in the filling of vesicles by partially crystallised residual magma may be seen in a single rock (e.g., 5875). Features similar to these in British tholeiites were long ago noted (Teall, 1889). In some rocks (e.g., 5885), the residual magma was squeezed into vesicles before cooling was far advanced, and crystallised wholly or partly into platy ilmenite, thin augite prisms and lathy andesine. In other rocks (e.g., 5743), residual glass remains in the vesicle, but has a much lower refractive index (1μ508) than the quickly chilled marginal tachylite glass. Again, in a coarsely crystalline dolerite high in the main Moeraki Sheet (5872), (in which cryptographic intergrowths might be expected), the interstitial material is a very pale brown partially devitrified glass containing microlitic oligoclase, augite and iron ore, and has a refractive index varying from 1μ489–1μ500 but usually 1.497. If these glasses were anhydrous the refractivity would suggest for that in (5743) a content of about 68% SiO2 and for (5872) about 72% SiO2. In view of the observations of Washington (1922, pp. 770–2) and Fenner (1931, pp. 556–7) it is more likely that

the lowered refractivity and paler colour results from the deuteric hydration of the residual glass. (b) The Pillow Lavas and Associated Basalt Flows. The tachylitic margin of the pillow lava at Boatman's Harbour and Awamoa Creek [Locs. 101, 102, slide (5701) = Anal. 1, Table VII, and (6794)] is pale yellow and transparent, and contains many small (< 0 μ 4 mm.) unzoned laths of labradorite (An69–60) and granules (< 1 μ 0 mm.) of almost forsteritic olivine (2 μ = 84°–86°). The darkening of the glass around the feldspar microlites by the abundant development of dusty magnetite and pyroxene-crystallites, and the gradual individualisation of these with advancing crystallisation may be traced in slides (6795–6) intermediate between the margins and centres of the pillows, and resemble the features described by Rosenbusch (1908, pp. 1276–80) as typical of such crystallisation to a greater extent than they do those noted by Fenner (1910) in the pillowy phase of the Watchung basalt. In the centre of a small pillow (6797–8) the plagioclase (An62–60) forms slightly zoned tabulae (< 1.0 mm.) and may have skeletal outgrowths. The associated microlites may be as sodic as An42. The glass is locally coloured red by swarms of minute deep brown translucent and opaque rods which, though not determinable (cf. Rosenbusch, 1908, p. 1278), suggest microlites of titanaugite. With these are minute skeletal iron ores. The glass is partially decomposed to pale yellow palagonite, or chloritic substances associated with a little indeterminate zeolite. Sometimes, as in the pillow lava (5868) in the Totara cutting, the vesicles have a thin lining of clinochlore. Pseudomorphs after olivine crystals are larger and more abundant in this rock than in those first cited. In the centre of a larger pillow (18 × 12 ft.) at Awamoa Creek (Loc. 102, 5753) crystallisation is more advanced. The rock is a normal olivine tholeiite* The term tholeiite is here employed as by Rosenbusch (1908, p. 1224) to denote hypocrystalline doleritic rock with intersertal structure. The relation of the Otago tholeiites to those in which recent British authors, especially Kennedy (1931, 1933), have recognised distinctive chemical features, is discussed below. not unlike the British Salen and Brunton types (Holmes and Harwood, 1929, pp. 12–17, Plate 1, Figs. 1–2). The larger feldspar crystals, about 1 μ 0 mm. long, with olivine pseudomorphs are set in a matrix of labradorite tabulae with granular or subophitic augite (2 μ = 52°, 54°) and more or less altered interstitial glass. This stage of crystallisation is represented by a rock (5725) in tuffaceous greensands beneath Waitaki limestone, near Tokarahi, 19 miles north-west of Oamaru. (c) The Dykes of the Moeraki Peninsula. The thinnest of the dykes (e.g., 5883) invading the Waiarekan breccia are tachylitic, containing decomposed phenocrysts of olivine, fresh crystals of labradorite and augite in a vesicular pale brown glass. Similar material, much richer in magnetite dust, forms the selvedges of the larger dykes and the thin dykes (5839) invading the Tawhiroko sheet, the matrix of which is almost opaque. (See Fig. 7.) The central portions of the majority of the dykes, including that forming the prominent outcrop at Te Paitu (5875), are vesicular olivine

tholeiites of rather fine grain-size, with irregular or subradiating labradorite tabulae, granular to sub-ophitic augite, platy and less often octahedral iron ores, and interstitial glass occurring in varying amount (5842, −48, −70, −71, −85). Rarely glass is lacking from the middle portions of the dykes, and the rock is in consequence a finegrained olivine dolerite (5857). Most of these dykes contain siliceous xenoliths, the features of which will be studied in Section C of this paper. Less finely granular but more ophitic and non-vesicular are the olivine tholeiites forming the thick dyke a quarter of a mile west of Te Koraki Point (5876), which is horizontally columnar, and that by the railway line a mile south of Trig. E (5832). In both of these there is but little interstitial glass. The rock forming Moeraki Point (5751) is a medium grained non-vesicular sub-ophitic olivine dolerite. That forming the largest dyke running W.S.W. of Te Koraki Point (5877) is noteworthy for its coarseness of grain-size, absence of olivine, and presence of a noteworthy amount of interstitial glass containing radial feldspar microlites, and other products of devitrification and deuteric alteration. It is closely similar to the rock (5872) forming the higher portion of the main dolerite sheet above Matiaha Head. (d) The Tawhiroko Intrusive Sheet. This forms one of the most complex examples of gravitational differentiation studied in New Zealand† Attention may be called to Marshall's (1906) account of the association of Iherzolite, pyroxenite, gabbro, and diorite at the Cow Saddle, in the region difficult of access north-west of Lake Wakatipu, which suggests the effect of gravitative differentiation in plutonic rocks, and is well worthy of fuller investigation.. Its main petrographical features indicated in Figure 6 (Benson, 1943) are revised in Figure 6 of this paper, also Figure 7 and Table VI herewith. It invades the Tahuian mudstone cutting obliquely across the stratification plane. The disposition of the various differentiates suggests that, though the western margin rises steeply through the sediments, the mass as a whole has a gentle eastward dip. As the angle of slope cannot, however, be measured or estimated, and the upper margin has been eroded away, its original thickness cannot be determined precisely, but was about 150 feet. Wave-erosion has produced a tidal platform about 8000 square yards in area, whereon the complex features of the lower half of the sheet may be studied in detail. Those of the upper half may be seen less completely in the low cliffs bounding the platform and the grassy hill slopes above. Little need be added here to the information afforded by the illustration and descriptions of the geological relations of the igneous rocks to their environment given in Section A of this paper. In regard to the data on Figure 6 it should be noted that the drawings and estimations of the proportions between the several minerals are based on inference as to the original character of the rocks prior to their deuteric alteration and weathering. They cannot be considered as more than qualitatively accurate, since in view of the abundance of secondary products often distributed beyond the limits of the parent substances, and the occasional difficulty of distinguishing between those derived from glass and those derived from

6. Petrographical features at successive approximately estimated horizons in the intrusive dolerite sheet of Tawhiroko Point. The positions of rocks illustrated are indicated by thick circles on the projected section, that of other described rocks by thin circles. The attached numbers are the last two digits in the register-numbers of the corresponding rocks in the catalogue of the Geological Museum of Otago University. Except in the case of 5743 and 5750 the first two digits are 58.

olivine or augite, micrometric studies seemed hardly justified. Recourse was. therefore, made to the method of estimatiion by inspection, with the helpof Holmes (1921, p. 324) diagrams, the areal proportions of the several minerals in not less than 12 quadrants (sixteen or twenty for more coarsely granular rocks) of fields of view selected as typical, and calculating weight-percentages from the Table V. Optical Data Determined by Drs, Turner and Hutton and Inferred Chemical Composition of Plagioclase and Pyroxene in the Tawhiroko Intrusive Sheet. Plagioclase Pyroxene Rock Slide Index Letter Fig. 4 Estimated Height Above Base Ft. % of Anorthite (Univ.stage) Optic Axial Angle Ref. Indices Composition α γ Wo En Fs 5856 J 135 Large crystals 49°±43° 1.690 1.718 Core 62→50→47, 44°, 30 47 23 58→50. Margin Small crystals 23 47 30 59, 55±5. 5851 K 125 59, 48→43±2, 40°±3°, 38° 1.685 1.711 22 55 23 44→35±2. 37±3°, 36°. 19 55 26 5750 L 110 Large 49, 48, 45, 48° 1.690 1.714 29 49 22 40→34, 45°→40° 1.718 22 48 30 small, 32. 5855 M 100 45→36, 51°, 48°, 1.683 1.708 32 51 17 39→27. 40°, 39°, 1.692 1.721 20 52 28 5827 N 90 39, 38, 37±2. 44°±3° 1.698 1.721 25 48 27 38° 1.723 19 49 32 5838 P 85 58→40, 52°, 50°, 1.685 1.712 33 49 18 57, 51. 49° 30 50 20 5828 Q 50 69→45 (thin zone), 53°, 50°, 1.684 1.705 36 50 14 63, 59, 56, 55 ±3. 50°, 38° 1.710 1.696 1.715 18 50 32 5829 R 30 49, 49±3, 44. 50°→38°, *Refractive indices and compositions estimated from 2V by Figs. 3 and 4. 1.686 1.713 31 49 20 46° 1.697 1.723 19 50 31 5835 15 58→30. 5834 S 10 62, 60, 60→50. 50° 50° 48° 1.690 1.718 30 50 20 48°, 44° 43° 23 50 27 5836 5 60→48. * 1.689 1.716 26 50 24 5831 T 0 64, 63, 61. 46°, 44°, 42° * 1.693 1.720 22 50 28 1.683 1.710 38 48 14 5743 Spur 60, 60, 44 lath. 54° 49° * 1.687 1.715 30 49 21

averaged areal proportions of the several minerals of known composition the densities of which were obtained from Washington's (1922 pp. 586–7) table, the glass being assumed to yield crystalline material with an average sp. g. of 2.90 (or 2.95 in the more basic rocks). The densities of the Tawhiroko rocks calculated from the above data, on the assumption that they were water-free holocrystalline and non-vesicular, give the curve plotted on Figure 6, and afford an indication of the extent of gravitative differentiation. Finally, it should be noted that the projected section of the Tawhiroko sheet is a slight modification of that given as Fig. 6 of Part A of this paper. The thickness of the possibly lensoid intrusive-sheet is not available as it is very irregular and transgresses the bedding planes of the invaded formations. The elevation of the various specimens above the base of the sheet is not known with certainty, and the relative levels of 5827, 5828, 5838 and 5855 are also not quite certain. The “estimated heights” given in Table V and Fig. 6 are thos rather hypothetical. The spur from the base of the main sheet and the base of the sheet itself have tachylitic selvedges less than an inch thick (e.g., 5831), containing phenocrysts of almost unzoned labradorite, olivine and very subordinate augite set in a pale brown glass containing feldspar microlites and dusty or skeletal iron ore. Less than five feet above this is vesicular olivine-tholeiite (5836–7) with nearly homogeneous tabulae of labradorite sometimes moulded on the gravitationally segregated olivine (which is here in greater amount than in any other part of the Tawhiroko sheet), and more abundant but smaller idiomorphic to subophitic augite and minute granules plates or skeleton-crystals of iron ores and feldspar microlites in glass which occurs interstitially and in sheets 0.5 mm. thick around the vesicles. Xenocrysts of quartz occur surrounded by the usual reaction-rim of minutely prismatic augite. About ten feet above the base is the first aggregated band of such xenoliths (5825–30) to be described in Section C of this paper. The enclosing rock (5830) is similar to (5743) the middle of the basal spur, which in turn is similar to the rocks of the Te Koraki Point dyke (5742, 5748) and is probably continuous therewith. Both the basal spur and this dyke contain quartzose xenoliths. Crystallisation of these tholeiites is further advanced. Glass still remains about the vesicles but is less abundant in the general ground-mass, where is augite in thin prisms studded with crystals of magnetite. Here and there are small (< 2μ0 mm.) patches of ophitic pyroxene. Olivine crystals are naturally smaller (< 1μ0 mm.) and rather less abundant in the spur and dyke than they are in the main sheet (5830) because of their gravitative accumulation therein. The plagioclase is now well zoned and on the whole a little more sodic. Radiating carbonates with subordinate delessite (?) fill the vesicles (5830) or replace glass (5743). A few feet above this xenolithic band olivine has decreased in abundance, but a little glass is present save about the vesicles, where it has the usual decomposition products (5755, 5834–5). The probably hydrated glass has R.I. = 1μ508 (see p. 89). The bulk-density

of the rock (2μ83) is less than normal for a holocrystalline, non-porous unaltered rock of this composition. Its texture approaches that of rather finely granular ophitic dolerite. This texture is more pronounced in (5858–9) and (5873) beyond the northern end of the xenolithic band, which rocks contain vesicles wholly or partially filled with more or less devitrified glass. About fifteen feet higher in the sheet (5829) the grainsize is a little larger, the ophitic structure more pronounced and in place of most, if not all, the glass there is a finely granular rather strongly feldspathic but otherwise somewhat pilotaxitic groundmass around the vesicles or forming large, irregular patches up to 10 mm. across, containing besides granular augite, long (< 0μ8 mm.) thin augite prisms studded with magnetite, long rows of aggregated magnetite and platy ilmenite and numerous small patches of siderite. Olivine was very rarely present among the larger crystals, but pseudomorphs thereafter may be recognised. The larger grains of augite are subophitic. The vesicles are filled with sphaerosiderite which rarely contains idiomprphic but indeterminable tabulae apparently of a zeolite. This rock lies at the base of a depression in the zone of holocrystalline augite dolerite. About the same horizon or possibly a little higher, though intersertal (5839) is without such micro-pilotaxitic material, but contains a little glass and rather more olivine than (5829). After it had consolidated and cooled it was invaded by thin (1–2 feet) dykes having tachylitic margins with more crystalline but still dominantly glassy central portions. (See Figs. 6 and 7.) Rather higher, though still containing a trace of glass and a little olivine (5828) has developed irregular, finely granular, almost orthophyric or trachytoid patches 4–5 mm. across, consisting chiefly of squarish prisms or tabulae of labradorite-andesine with a minor amount of pyroxene and iron ores. The larger pyroxene crystals are normal, not very calcic augite, and are associated with or zoned by subcalcic augite. The larger plagioclases are surprisingly calcic and zonal feldspar more sodic than An50 forms only very thin mantles about the calcic cores. The explanation of this basicity is not obvious. Above this the rock (5852) is holocrystalline and free from olivine. Its grain-fabric shows but a trace of the ophitic texture. The plagioclase (< 0μ8 × 0μ4 × 0μ3 mm.) and augite (0μ4 × 0μ3 mm.) are associated with plates of ilmenite (< 1μ0 × 0μ1 mm.) and irregular patches (< 3μ0 mm.) of cryptographic material together with some siderite. A curious mingling of different phases of dolerite occurs at a horizon near, possibly a little below, that of (5852). A more or less “blotchy” effect is seen where the mingling is intimate (5841–2), and a sharper separation in other cases (5840). (See Fig. 7.) The paler portion has the general composition of the rock with the orthophyric trachytic mesostasis, but is more coarsely granular, and consists of elongated plagioclase (An50), sometimes almost phenocrystic, with marked ophitic development of the very subordinate augite and little or no olivine or glass. The darker portion contains long prisms of augite and a little olivine, with platy ilmenite and glass, abundant near the contact but almost lacking a few feet away from it where

the rock (5896) is a subophitic, holocrystalline and rich in ferriferous minerals. This constitutes the “cementing matrix” of the irregular “conglomeratic” layer of aggregated quartzose xenoliths in the middle of the tidal platform which will be described in Section C of this paper. It would seem as if rather basic dolerite magma containing abundant xenoliths had been injected into and ramified within the Tawhiroko sheet just after it had completed gravitative differentiation and consolidation, but while it was still too hot to allow the formation of a tachylitic selvedge, though not hot enough to prevent the consolidation of some residual glass therein (cf. Walker, 1940, p. 1086–7). The rapidity of injection from the magma may account for the absence of hypersthene from the dolerite surrounding the xenoliths (5838, 5896), though some has been formed in the little veinlets injected into the xenoliths, where there has been greater opportunity for solution-reaction (cf. Lacroix, 1893, p. 34; Harker, 1904, p. 353). The injection of this xenolith-charged doleritic magma was so irregular that its varying spatial relations to the several gravitationally differentiated phases of the Tawhiroko sheet could not be fully determined, and it is not clear whether (5838) is portion of this later mass. 7. Welded (5840) and sharp (5839) contacts between coarse-grained dolerites and later intrusions in the Tawhlioko Sheet. In the middle of the tidal platform (5827), which has no trace of olivine or of glass, contains interstitial alkali feldspars (anortho-clase? and sodic andesine) in cryptographic or radial intergrowths between tabulae of calcic andesine and subcalcic augite. The last occurs in long magnetite-studded prisms or in ophitic intergrowth with plagioclase. Similar to this are (5855, -4, 3, -1) occurring above the tidal platform. They differ from (5827) chiefly in the occasional trachytoid texture of the interstitial material, and the increasing content of augite and calcic nature of the plagioclase towards the top of the sheet. Iron ores seem to have a slight maximum within this cryptographic zone, the higher members of which (5854–1) contained large crystals of calcic labradorite and rare small grains of olivine which sank but a short distance out of the upper marginal portion

of the sheet. Only in 5851 are there a few grains of quartz large enough for examination in convergent light. The highest dolerite in Tawhiroko Hill (5856) is 72 feet above the tidal platform, about 10 feet above 5851, and was probably about 15 feet below the upper tachylitic margin of the sheet. In accordance with this, its plagioclase crystals are smaller and more calcic than those in 5851, the augite smaller, more calcic, and more abundant. The rock is slightly vesicular, and contains glassy and “microbasaltic” interstitial material. There was very little if any olivine present. Though the grain-size of the rock shows the normal increase from the margins towards the interior, uncertainty as to the precise thickness of the intrusive sheet, and the exact horizon therein of the specimens described prevents any quantitative study of the variation in grain-size (cf. Alling, 1936, pp. 321–327). Breaking the cryptocrystalline zone is a coarsely granular, probably lensoid mass of pegmatoid dolerite, grading imperceptibly outward into the less coarse to medium-grained enclosing rocks. The least weathered specimen (sp. gr. = 2μ75) contains much deuteric siderite which forms powdery haematite on weathering, leaving a porous rock (sp. gr. = 2μ66) a sample of which was analysed (5750, Anal. 5, in Table VI). Assuming the pyroxene has a composition approaching that determined optically the calculated mineral composition would be—Quartz and chalcedony (in part deuteric) 13μ5%, feldspar (Or13 Ab58 An20) 63μ7%, augite (Wo20 En52 Fs18* Probably the addition of a little of the normative anorthite and iron ore to the pyroxene would result in a closer approach to its modal composition.) 7.7%, apatite 1μ1%, ilmenite 5μ7%, haematite replacing siderite 8μ3%. Omitting the last mineral, the calculated sp. g. of the rock is 2.72. The illustration (Fig. 6) is that of a field containing rather more than the average proportion of coloured constituents. The larger feldspar crystals are zoned andesine, but there is approximately 20% of trachytoid, micropegmatoid or cryptographic matrix in irregular ovoid or interstitial patches consisting of lathy andesine-oligoclase and anorthoclase (?) with R.1. respectively equal to or less than that of Canada balsam. Most of the silica occurs in indeterminably fine quartz-feldspar intergrowths and as chalcedony in this matrix, but some (surprisingly few) clear grains (< 0μ3 mm.) of quartz may be seen. A little microlitic pyroxene and magnetite occurs with this felsitic matrix, which may surround vesicles filled with partially oxidised sphaerosiderite, and subordinate chlorite, quartz, chalcedony and opal. The composition of the pyroxene is transitional between normal and subcalcic. It forms relatively narrow, sometimes arcuate prisms up to 5 mm. long, though usually subophitic and zoned. The reason for such curvature is not apparent, though the feature has been noticed by several petrographers. Thus Bowen (1910) found it in Late Pre-Cambrian dolerite in Canada, Teall (1884) and Tomkieff (1929) in the Late Palaeozoic Whin Sill, Phillips (1899), Emerson (1905) and Shannon (1924) in the Triassic Palisadan dolerite sills of Eastern U.S.A., and Browne (1923) in Tasmanian and Antarctic dolerites. When present, it is nearly always associated with coarsely

granular, probably or certainly pegmafoid phases of the dolerite, a rather exceptional case being that described by Emerson in which the pyroxenes have assumed a curiously “plumose” type of spherulitic structure, and occur in a coarse-grained, but irregular narrow band near the margin of a main intrusion. In one case only there is a suggestion of a shearing movement in the consolidating rock; usually there is no evidence of this. Tomkieff's (op. cit. pp. 116–8) conclusion concerning similar but more sharply bounded coarsegrained dolerite in the Whin Sill, namely, that it was “possibly the result of ‘wet’ differentiation formed in the intercrustal basin, caught up in the ascending magma and stretched out in the form of lozenge-shaped tabular bodies parallel to the walls of the injection chamber” is not applicable to this rock-mass, which seems to have formed by differentiation in situ. (See Shannon, 1924, p. 39, as quoted below p. 111, and Phemister, 1928, pp. 162–170). In general, the vesiculation and abundance of deuteric minerals throughout this sheet indicates the abundance of fluxing volatile materials in the invading magma, and the consequent facilitation of its gravitative differentiation to a greater extent than occurs in the much thicker intrusive sheets of the Palisades (Walker, 1941), the Karroo (Walker, 1940, etc.), Tasmania (Edwards, 1942) or (so far as is known) Antarctica (e.g., Benson, 1916).* We may contrast these products of magmas not rich in fugitive constituents with the more deuterically altered and gravitationally differentiated diabase sill at Bridgehead, Ontario, which presents interesting analogies and contrasts with the N.E. Otago Sheets (Emmons, 1927, pp. 73–82); see also Phemister (1928, pp. 102–170, 185). The need of fluxing to promote gravitative differentiation in sills was stressed by Harker (1916, p. 555), who noted that “clear instances of gravitative differentiation in sills and laccolites … are all in rocks which must represent very unusually fluid magmas such as the analcime-bearing intrusions of Permian Age in Scotland,” and, we may add, the richly zeolitic theralitic intrusion at Waihola (Benson, 1942). The bulk of the water in the Tawhiroko, Main Moeraki and Mt. Charles intrusive sheets was probably magmatic, but this might have been supplemented by water derived from the plastic sediments which they invaded (cf. Leitmeier, 1909; Daly, 1917, p. 445, 1933, pp. 307–11; Shannon, 1924, p. 39), a view which Grout (1928, pp. 567–70)§ But see Grout (1932, pp. 212–3). opposed, though Day and Allen's (1925, pp. 76–84) application of Morey's (1922) experimental study of the absorption of water into silicate melts at low (even atmospheric) pressures of “meteoric water acquired in the usual way through contact with water-bearing strata or reaching the volcanic hearth under a head determined by the elevation of the crater basin,” to the explanation of the effect of downward percolating snow-water in stimulating the eruptive activity at Lassen Peak, supports this view. The general course of the formation of these Otago Sills has, however, much in common with that of the large Palisade Sill (Walker, 1940, p. 1101). Thus:— 1. “The main differentiation was effected by the sinking of early-formed olivine followed at a later stage by pyroxene.” The

larger olivine crystals accumulated above the chilled base, a few of the smaller were retained by the stiffening melt near the upper margin. 2. “Although the sinking of the two minerals probably overlapped, crystalisation of olivine ceased at an early stage, when the magma was still fluid, whereas pyroxene continued to form and to sink until brought to rest by the increasing viscosity of the magma and by crystal interference.” 3. “There was no appreciable sinking of iron ore.” 4. “Interstitial alkaline feldspars are found in the upper portion of the sheet with micropegmatite and free quartz in its higher parts,” but there is no evidence of albitisation or formation of albitic veinlets such as occur in the Palisade Sill. 5. The later stages of crystallisation were followed by pronounced hydrothermal activity resulting in formation of chlorophaeite and chloritic minerals with chalcedony chiefly at the expense of residual glass, bowlingite, quartz, chalcedony, and ferruginous and calcic carbonates, more or less in the above order. 6. Though the thickness of this sheet is so much less than that of the Palisade Sill or the Tasmanian sills (above 1000 feet) the range of average grain size of the plagioclase is but little different from that throughout the Palisade Sill and the lower 700 ft of the Tasmanian Mt. Wellington Sill. Probably this resulted from a greater content of water in the magma, which also facilitated the gravitational differentiation. 8. Olivine dolerite (5874) near the base, and quartz dolerite (5869) 200 feet above the base of the Main Moeraki Sheet. (e) The Main Moeraki Dolerite Sheet. The obscuring drift on hill-slopes prevents a detailed tracing of the differentiation-sequence in this sheet, though it may later be accomplished by a study of the high cliff half a mile north-west of the lighthouse the base of which consists of olivine dolerite and the top of quartz dolerite. (See Benson, 1943, Fig. 8.) The two end-members of the series are however, widespread. Olivine dolerite occurs always near the base of the intrusive sheet as at Tawitiatiauka (5860) and a quarter of a mile west thereof (5790) by the shore

near the Moeraki village, on the northern flanks of Trig E (5884), capping low hills half a mile east thereof (5895, 5874), off the shore by Matiaha Head (5886), and again immediately west (5887) of the lighthouse and half a mile north-west thereof (5888). On the other hand, the upper portion of the plateau extending north from the lighthouse (5879, -78, -97) and the tops of the higher hills near Trig E (5889, -90, -91), all 100–200 feet above the base of the sheet, consist of quartz dolerite, which dips eastward towards Okahau Point. Olivine dolerite (5860), which is probably the rock analysed by Seelye (Table VI, Anal. No. 2), was collected 10–15 feet above the basal content with porcellanitised mudstone. Though slightly over-saturated with silica, it is a sub-ophitic olivine dolerite.* “The norm always tends to make the rock appear more acid than it really is, and therefore a small amount of normative quartz, up to 2% say” (4.32% in Analysis No. 2) “is compatible with the normal basaltic composition” (Phemister, 1934, p. 42). The dominent mineral is labradorite (An65-55) in tabulae rarely over 1 mm. long. Augite (see Table III) is rarely half as large, the partially bowlingitised olivine crystals (Fa14-30) may be up to 1μ5 mm. long. Platy ilmenite and a little glass are also present. Other olivine dolerites from near the base of the sheet differ in minor details:—(5790) though of rather small grain-size has less glass; (5884–86), perhaps a litle further from the base, have no glass, but a little trachytoid mesotasis; (5886) some magnetite; (5884) less olivine and a rather coarser grain-size; (5874, 5895) about 30 feet above the base are very ophitic (Fig. 8), the individual pyroxene crystals sometimes extending over 3 mm. The small amount of partially decomposed glass in these rocks contains apatite needles and rare feldspar microlites. Near the lighthouse dolerites (5887–8) adjacent to porcellanite at the base of the sheet have been rendered friable by the almost complete replacement of pyroxene and olivine by dark green chlorite and bowlingite. Differing from these is (5835) a porphyritic tholeiitic olivine dolerite (cf. 5834 in Fig. 6) occurring on the east coast near the lighthouse, adjacent to large included masses of sediment to the chilling effects of which it owes its texture. It contains phenocrysts of olivine and pyroxene, small interstitial patches of finely granular feldspathic microbasalt, and glass, the last chiefly around the vesicles. Its specially interesting feature is the presence of a small minutely-granular aggregate of bytownite and spinel, resulting from the fusion and recrystallisation of a marly xenolith, as will be described in Section C of this paper. Quartz-bearing Dolerites (5869; see Fig. 8), 5878–79, -89, -90, -91, and -97) are typical and closely similar, A more coarsely granular rock (5880) occurs on the shore near Okahau Point. The pyroxene (see Table III) forms short or long prismoid or subophitic to typically ophitic plates, the last of which may be over 3 mm. across. Slight chloritisation of the pyroxene is sometimes seen, especially in (5891). Sometimes (5878) the pyroxene occurs graphically intergrown with plagioclase especially in the terminal portions of the smaller tabulae, or occurring in still finer division in the alkaline feldspars extending sub-radially from the margins of

such tabulae into the interstitial micrographic material. Hornblende occurs sparsely in (5891). Its features have already been described (p. 84). Ilmenite forms plates up to 3 mm. long, and rarely seems moulded on feldspar (5878), the reverse being the more common relation. Leucoxination is well displayed in (5891). Rarely (5879) haematitic derivatives of deuteric carbonates are associated with talc in what seems to be pseudomorphs of small olivine grains. The plagioclase tabulae 2–3 mm. long, or 5 mm. in (5880) which resembles (5750) of the Tawhiroko Sheet (see Fig. 6), are strongly zoned with a core of labradorite (An60) or basic andesine and marginal zones sometimes as sodic as An60. They contain needles of apatite in general more abundant than in the Tawhiroko Sheet. The interstitial material varies even within a single rock. It may be a trachytoid aggregate of relatively large (< 0μ5 mm.) microlites of oligoclase and anorthoclase or minutely fibrous intergrowths extending radially or in curved plumose tufts (5897) from the corners of the tabulae, and both of these structures may be associated with a little bleached residual glass (5879, -91) which in the latter rock has been partly replaced by opal. Commonly it frays out through crypto- or micrographic intergrowths into crystals of quartz which may terminate idiomorphically against ferruginous carbonates. Apatite needles are very abundant in such intergrowths. The varied features and sequence of the silica minerals, quartz and chalcedony, in these rocks, and their significance have been already discussed. (See p. 87.) In the coarsely granular dolerite in the cliffs above Matiaha Head (5872) the residual interstitial magma did not consolidate with the formation of micropegmatite, but as glass now probably hydrated with R. 1. = 1μ497, but varying between 1μ489 and 1μ500, with or without the usual opaque trichites and margarites or microlites of oligoclase, augite, ilmenite and magnetite. The pyroxene in this rock was strongly differentiated during crystallisation and shows the only instance of “reversed zoning” observed among the rocks described herein. (See Table III.) A relatively small and a very large lenticular aggregate of quartzose xenoliths in this sheet occur respectively, by the shore near Matiaha Head and the hilltop 700 yards W.S.W. thereof. The very large mudstone xenoliths in this sheet are surrounded by a narrow film of olivine-bearing tachylite (5833, 5881) grading through rather finely granular vesicular tholeiite (5825) into the nearby (5885) and perfectly normal coarse-grained dolerite (5885). Here and there veins or irregular segregations of more or less calcitic chalcedony traverse the rocks of this sheet. (f) Other Mid-Tertiary Basalts, Tholehtes and Dolerites. These are widespread, but as most of them closely resemble rocks in the Moeraki Peninsula detailed descriptions are rarely necessary. Porphrytic olivine basalt with a pilotaxitic to fluidal holocrystalline fine to medium-grained ground-mass is not known in this peninsula. It forms the sheet (5752) in Lookout Bluff (Loc. 84, Trig. D, Otepopo S.D.), and the volcanic plug (5866) half a mile south of Totara railway station (Loc. 98A). It usually contains phenocrysts of olivine only. A little stilbite (?) occurs in 5752.

Porphyritic basalt (5721) with phenocrysts of olivine, augite, singly or in ophitic or glomeroporphyritic groups of labradorite, set in a ground-mass like that of (5734), invaded both the Waiarekan tuff and the Ototaran limestone at Loc. 98, a mile N.E. of Maheno. (See Benson, 1943, Fig. 2.) Medium-grained olivine dolerites with a little intersertal glass are represented by the rock (5724) forming Round Hill (Loc. 100 = Trig. K, Oamaru S.D.), and (5719) the lower portion of the igneous cap on Kauroo Hill (Loc. 100, Kauroo S.D.). (See Benson, 1942a, p. 114.) These are closely similar to rocks in the lower part of the Tawhiroko Sheet, but contain less glass than otherwise similar material in the flow west of Tokarahi (5725) and in the centre of the large pillow (5763) in Awamoa Creek. Very coarsely granular intersertal olivine dolerite forms the top of South Peak (Loc. 82) (5722) and North Peak (Loc. 81) (5716). Both are beautifully ophitic and in the latter the mesotasis is largely devitrified with formation of skeletal extensions of the major feldspar tabulae and other feldspar microlites. The olivine (Fa15) is unzoned as is also the augite (2 V = 50°, 51°) and the plagioclase mostly basic andesine (An52–42). A similar rock (6785) occurs also at Enfield (Loc. 105 = Teaneraki), but has been very greatly affected by carbonation. Coarsely granular, more or less ophitic holocrystalline dolerites with pseudomorphs after olivine were collected by Marshall (?) from various unspecified localities near Mt. Charles (Loc. 86). The most olivinie (6787) is least coarsely crystalline and without micro-pegmatite, less olivine occurred in the more coarsely granular rocks (6788, 6791) in which a little micropegmatite appears, less again where the intergrowth becomes abundant and radiating quartz with undulose extinction is enclosed within deuteric carbonates (6790). The usually unzoned augite in this rock has 2 V 7 = 61° ± 3°, 56°, 54°, 52°. Its other optical properties and inferred composition are shown in Table III and Fig. 4 (I). It is not clear that these strongly calcic pyroxenes formed under plutonic conditions prior to magma injection. The plagioclase when completely enclosed in augite is An70, but where exposed varies from An58 to An40, the last being the composition of a small excresence from a tabula projecting into quartz. The only localised specimen (5727) came from well above the base of the sheet in the cliffs adjacent to the railway, 200 yards north of the Waianakarua Bridge (Loc. 85). It contained still less olivine and less calcic unzoned pyroxene (2 V = 54°–45°). The abundant and not very finely granular micropegmatite in this rock is associated with more and larger (< 1.0 mm.) grains of optically uniform quartz than in any other dolerite described in this paper. Finally, it is to be noted that Hutton (1887, p. 428) found no trace of olivine, but some “enstatite” (hypersthene?) in rocks obtained by him from Mt. Charles. It remains to be seen whether such hypersthene is here normally magmatic or the result of local solution of quartzose xenoliths by the basic magma. The range of petrographic features displayed in this sheet suggests that it also will be found to be gravitationally differentiated.

Anorthoclase-bearing porphyritic olivine basalt (5734), containing a little indeterminable zeolite, invades the Waiarekan tuff near Maheno (Loc. 97, Trig. S, Otepopo D.) and caps Government Hill (5741) (Loc. 87, Trig. O, Otepopo S.D.) and probably also the adjacent Little Table Hill. The analyses of these rocks (Table VII, Anal. 3–4) show a greater amount of potash than those of the Moeraki Sheet, and this is expressed mineralogically by the development of anorthoclase tabulae in small amount, usually associated with sparse, minute flakes of (partly deuteric) red-brown biotite. Analcite olivine dolerite of medium grain-size (5749) occurs a mile and a-half west of Maheno (Loc. 96) and forms part of the same rather alkaline sheet as the above. The olivine (< 0μ9 mm.) is fresh and slightly zoned, the plagioclase (< 0μ5 mm.) is basic andesine (An47), and the granular augites are up to 0μ5 mm. in diameter. Ilmenite plates, usually idiomorphic but rarely moulded on olivine, may be 1μ2 mm. long. Magnetite is also present. Anorthoclase in small amount forms tabulae (< 0μ2 mm.), or is moulded between the plagioclase tabulae. Biotite flakes, rarely 0μ2 mm. in diameter, but more usually < 0μ05 mm. and probably deuteric, are associated with chlorite. Analcite forms rare patches (< 0μ4 mm.) or veinlets replacing feldspar, and there are a few relatively large (0μ6 × 0μ03 mm.) prisms of apatite as well as minute needles. The rock differs from crinanite in being less basic, having normal rather than titaniferous augite, a non-ophitic texture and a much greater amount of potassic feldspar. Coarse-grained basic analcite syenite or melasyenite (5865) occurs by the Waimotu railway station (Loc. 97A) a mile south of the last rock, and forms probably a pegmatoid segregation in the same intrusive sheet. The largest crystals (5μ0 × 0μ2 or 3μ0 × 2μ0 mm.) are of basic oligoclase, but these are subordinate to the anorthoclase. The oligoclase has been attacked by deuteric alkaline solution with the development of checkered patches of albite (?) and analcite. The anorthoclase, having the characteristic optic properties already noted (see p. 12) forms idiomorphic zoned tabulae (2 × 1μ5 mm.), and is usually transparent in the marginal portions though it may be turbid within. Both feldspars contain abundant needles of apatite. Biotite plates (1μ0 mm. long) were common and less so are thin, faintly purplish prisms of augite (2 V = 57°) which may be 1μ6 mm. long but usually much less. Sometimes short, pale, weakly pleochroic prisms occur with a greenish sodic shell. Both the augite and biotite have been more or less converted to dark green chlorite. Aggregates of octahedral magnetite (< 0μ4 mm.) and scattered plates of ilmenite (< 1μ0 mm.) are abundant, often growing out from the augite or biotite in-branch-like fashion. The interstices are filled with straight or arcuate laths of anorthoelase in a chloritic matrix. Analcite in idiomorphie (?) or rectilinearly bounded interstitial masses (< 1μ5 mm.) is common, and its late development is shown by its being moulded on the anorthoclase microlites in these areas. It contains abundant minute liquid inclusions, but is almost free from inclusions of minerals. Natrolite (?) enclosing carbonates is locally moulded on the corroded margin of analcite. This rock, so different from any Late Tertiary product, is the most alkaline derivative of the Mid-Tertiary basic rocks.

Table VI.—Rock Analyses, Chemical Composition and Petrologieal Affinities of Mid and Late Tertiary Basic Igneous Rocks of Otago. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 SiO2 52.92 50.42 49.44 47.39 55.94 52.73 52.61 50.52 49.3 48.8 52.10 50 45 46.16 Al2O3 16.35 14.09 14.46 13.35 13.61 15.14 13.93 13.76 12.2 13.6 16.37 13 15 15.32 Fe2O3 2.05 3.57 1.95 1.95 7.83 2.40 3.03 3.87 4.7 5.1 3.27 13 13 3.50 FeO 7.62 6.30 8.75 9.78 3.36 7.57 7.39 8.50 9.2 8.5 4.99 13 13 9.67 MgO 5.33 6.59 7.89 10.01 1.71 6.70 5.60 5.42 5.5 5.3 5.76 5 8 7.25 CaO 8.32 7.97 8.10 8.80 4.86 8.17 9.62 9.09 9.4 8.4 8.20 10 9 9.58 Na2O 3.36 3.11 3.39 3.02 4.03 3.34 2.42 2.42 2.1 2.6 3.03 2.8 2.5 3.74 K2O 0.96 0.66 1.21 1.20 1.33 1.01 1.22 0.96 1.1 0.8 1.76 1.2 0.5 1.34 H2O+ 1.26 1.54 1.48 1.48 1.42 — 1.05 1.51 2.7 3.2 2.5 H2O- 0.57 3.40 0.78 0.60 2.78 — 1.00 0.76 CO2 nil tr.nf. tr. tr. — 0.45 0.58 - - 1.20 - - - - TiO2 0.93 1.58 1.88 1.94 2.81 2.21 1.21 2.39 3.1 2.5 0.67 - - 2.45 P2O5 0.33 0.34 0.46 0.40 0.43 0.39 0.21 0.26 0.4 0.4 0.21 - - 0.65 ZrO2 nf. nf. nf. nf. nf. - - - - - - - - - S 0.06 0.07 0.05 0.03 0.05 0.06 0.09 - - - 0.08 - - 0.04 MnO 0.12 0.12 0.14 0.15 0.06 0.13 0.23 0.16 0.2 - - - - 0.21 BaO 0.03 0.02 0.03 0.02 0.04 0.03 0.03 - - - - - - 0.11 SrO 0.01 0.09 0.04 0.03 0.02 0.04 - - - - - - - 0.04 Cr2O3 0.04 0.04 0.04 0.06 nf. 0.04 - - - - - - - 0.03 V2O3 0.02 0.03 0.02 0.03 0.02 0.02 0.06 - - - - - - 0.01 NiO 0.02 0.03 0.02 0.03 nf. 0.02 - - - - - - - 0.02 Cl Ti. Tr. nf. nf. 0.04§ Soluble in water. Probably dried sea spray. - - 0.11* Minor Constituents. 0.2 0.8 - - - - Total 100.30 99.97 100.13 100.27 100.34 100.00 100.15 100.31 100.1 100.1 100.10 - - 100.12 Sp.gr. 2.725 2.79 2.90 2.94 1. Tachylitic margin of pillow lava (5701). BOatman's Harbour, Oamaru, F. T. Seelye Anal.; R.I. 1.5772 + 0.0002 and sp. g. by C. O. Hutton. 2. Olivine dolerite (5860?); Tawitiatauka Point. Mooraki. F. T. Seelye Anal.; sp. g. measured on (5860). Analysed specimen not available.

3. Porphyritic olivine basalt (5734); Trig. S, Maheno S.D. F. T. Scelye Anal. 4. Porphyritic olivine basalt (5741); Government Hill, half-mile S.W. of Trig. O, Otepopo S.D. F. T. Seelye Anal. 5. Pegmatoid doleritc (5750); near centre of Tawhiroko Sheet. F. T. Seelye Anal. Fe2O4 chiefly derived from oxidised siderite. 6. Average of 4 parts each of 1 and 2, 2 parts each of 3 and 4, and 1 of 5 (with adjustment for its excess of Fe4O3). Considered as the average Mid-Tertiary magma of N.E. Otago. 7. Average composition of 19 English tholeiites listed by Holmes and Harwood (1929). 8. Average composition of 6 English quartz-dolerites of the Whin Sill type listed by Holmes and Harwood (1928). 9. Average of 6 Scottish tholeiites (Walker, 1935). 10. Average of 6 Scottish quartz dolerites (Walker, 1935). 11. Average of live German tholeiites listed by Rosenbusch (1910) and Tröger (1935). 12. General composition of the tholeiitic magma type. 13. General composition of the olivine basalt magma type. Nos. 12 and 13 both after Kennedy (1931, 1933). 14. Average composition of Late Tertiary basaltic igneous rocks in regions peripheral to the Dunediu District (Benson, 1942). (g) The Chemical Composition and Petrological Affinities of The Mid- And Late Tertiary Basic Igneous Rocks of Otago. It will be seen that the average Mid-Tertiary basic magma in North-Eastern Otago (No. 6) contrasts sharply with the Late Tertiary basaltic magma of Eastern Otago (No. 14), which in turn differs from Kennedy's (1931, 1933) olivine basalt magma type (No. 13) chiefly in its greater content of alkalies. On the other hand the average Otago Mid-Tertiary basic magma resembles rather closely the average English and Scottish quartz dolerites (Nos. 8 and 10), the average German tholeiites (No. 11) and still more closely the average of these three groups. It differs from all of them, except the German tholeiites, chiefly in its greater content of soda, a distinction which would still remain if the Otago average were calculated without reference to the anorthoclase-bearing basalts (Nos. 3 and 4) which yield a syenitic pegmatoid rather than the micrographic pegmatoid as in the remaining dolerites and in typical tholeiite-dolerite associations (Kennedy, 1933, p. 244). If comparison be made with the average composition of undifferentiated marginal portions of the Palisade Sill (Walker, 1940, p. 1080), and with that similarly of the Tasmanian dolerites (Edwards, 1942, p. 465) and the averaged composition of Western Australian tholeiites (Edwards, 1938, p. 7) also of Spitzbergen (Tyrrell and Sandford, 1933, p. 312) and Antarctic dolerites (Browne, 1923, p. 253), the same essential resemblance but with greater alkalinity in the Otago rocks is again clear. The Otago Mid-Tertiary and Late Tertiary basaltic magmas may therefore be taken to represent in some measure respectively rather alkaline facies of the tholeiitic and olivine basalt magma types of Kennedy (opp. cit.). The relationships as far as they concern the Mid-Tertiary and tholeiitic rocks, are made clearer by a comparison of average normative compositions as given in Table VII, and of normative and modal developments of pyroxenes given in Table VIII.

Table VII. Normative Composition of Quartz Dolerites, Tholeiites, Plateau Basalts and Non-alkaline Pacific Basalts. Note Rock Series No. of Anal. Feldspar Or Ab An % Pyroxeuc Wo En Fs % Qtz. % Ores % Apatite % 1 Otago Dolerite and Tholeiites 5 11 51 38 58.6 18 60 22 31.2 1.4 7.8 1.0 2 Antaretic Dolerites 4 10 26 64 52.1 19 56 25 40.9 2.8 2.3 Tr. 3 Tasmanian Dolerites 6 10 27 63 54.0 23 47 30 42.2 3.9 2.0 Tr. 4 Karroo, South Africa, Dolerites 15 10 41 49 52.7 24 42 34 42.2 1.1 3.4 0.3 5 Palisadan Dolerites, U.S.A. 20 10 37 53 53.9 16 55 29 40.7 1.0 5.0 0.3 6 Watchung Basalt, Eastern U.S.A. 8 9 49 42 51.6 22 50 28 35.7 3.6 8.8 0.3 7 British Quartz Dolerites 12 11 43 46 51.5 25 51 24 30.1 6.1 11.6 0.7 8 British Tholeiites 25 16 41 43 51.4 29 49 22 33.6 5.6 8.7 0.7 9 German Tholeiites 5 16 42 42 63.8 21 62 17 25.8 4.0 7.2 0.4 10 Western Australian Tholeiites 3 6 47 47 52.7 30 45 25 37.0 3.1 6.6 Tr. 11 Spitzbergen Quartz Dolerites 4 12 34 52 51.2 31 46 23 34.7 3.3 10.5 0.3 12 Plateau Basalts, 22 10 46 44 50.9 25 43 32 35.9 3.9 8.8 1.0 13 Pacific Basalts 42 7 47 46 50.2 25 54 21 39.2 0.4 9.5 0.7 Notes to Table VII. 1. Analyses by F. T. Seclye. Average in Column 6, Table VI. 2. Average of analyses by Prior (1907), Benson (1916), Osborne and Graham (in Browne, 1923). 3. Average of analyses of undifferentiated marginal doleritc made by Edwards (1942). 4. Average of modern analyses by Herdsman and others cited in various papers, chiefly by Walker. 5. Average of 12 analyses from the Palisade Sill, and others from adjacent parts of New Jersey, Connecticutt and Virginia, chiefly by Gage and Gonyer. (See inter alia Lewis, 1907; Walker, 1940). 6. Analyses chiefly by Gage of the effusive products of the Palisadan magma. (See Lewis, 1907).

7. Average composition of six English and six Scottish dolerites selected by Walker. Analyses by Harwood and others. (See Holmes aud Harwood, 1928; Walker, 1935.) An earlier average by Tyrrell and Sandford (1933) including 24 analyses of the Whin Sill dolerite gives Or12Ab38An50, 54.3%, Wo22En50Fs28 38.0%, quartz 7.3%, iron ores 9.9%, apatite 0.7%. 8. Average composition of 19 English and 6 Scottish tholeiites by various analysts, chiefly Harwood and Herdsman. (See Holmes and Harwood, 1920; Walker, 1935.) 9. Average of five analyses of German tholeiites listed by Bosenbusch (1910), Roseubnsch-Osaun (1923), and Tröger (1935). 10. Average of three analyses by Edwards (1938). 11. Average of four analyses by Herdsman, Harwood and others selected out of the eight cited by Tyrrell and Sandford (1933). 12. Average of 16 analyses of Deccan basalt and six of Oregon basalt, chiefly by Washington. (See Washington, 1922; Tyrrell and Sandford, 1933.) 13. Average of analyses, chiefly by Washington and Keyes, omitting oligoclase basalts and other more alkaline types. (See Washington, 1923; Washington and Keyes, 1920, 1928; and Lacroix, 1927.) Table VII. Comparison Between Normative and Modal Pyroxene in Representative Rock Series. Notes Rock Series No. of Anal. in Av. Normative Pyroxene Modal Pyroxenes Wo En Fs Augite 2V>46° Subcalcic Augite Pigeonite 2V<30° Ortho-Pyroxene X Japanese Andesite 42 8 58 24 Abun. Subord. Rare Abun. Y Japanese Basalt 40 16 48 36 Domnt.? Subord. Subord. Subord. 1 Otago Dolerite, etc. 5 18 60 22 Domnt.? Subord. Absent Absent? 2 Antarctic Dolerite 4 19 56 25 Abun. Subord. Abun. Subord. 3 Tasmanian Dolerite 6 23 47 30 Abun. Abun. Subord. Abun. 4 Karroo Dolerite 11 21 43 36 Abun. Abun. Abun. Subord. 5 Palisadan Dolerite 4 21 46 33 Abun. Abun. Abun. Abun. 6 Watchung Basalt 8 22 50 28 Subord. Abn. Absent? Absent? 7 British Dolerite 25 29 49 22 Dominant Rare? Rafe 8 British Tholeiite 25 29 49 22 See Below 9 German Tholeiite 5 16 42 42 Dominant Not Reported 10 Western Australian Tholeiite 3 30 45 25 Subord.? Subord.? Domnt.? Absent 11 Spitzbergen Dolerite 4 31 46 23 Abun. Abun? Subord. Absent 12 Plateau Basalt 22 25 43 32 Abun. Domnt.? Rare Absent 13 Pacific Basalt 42 24 54 21 Abun. Subord.? Rare Rare

Notes to Table VIII. X and Y. Mineralogical data from Kuno (1935) and Tsuya (1937). Optical determinations of clinopyroxene compositions plotted on Figure 9 herewith. The non-alkaline types of basalt are alone considered. Analyses chiefly by Tanaka. 1. Data as given above. 2. Mineralogical data from Prior (1907), Benson (1916) and Browne (1923). 3. Mineralogical data from Edwards (1942) and papers cited by him. 4. Normative composition kindly supplied by Professor Walker based on analyses (some unpublished) of undifferentiated marginal phases. Mineralogical data from Walker and Poldervaart (1941, 1942). 5. Normative composition of pyroxene communicated by Professor Walker and calculated from selected analyses of undifferentiated marginal phases. Mineralogical data chiefly after Walker (1940). 6. Mineralogical data after Lewis (1907) and Fenner (1910). 7. Mineralogical data after Holmes (1928), Tomkieff (1929), and Walker (1935). “Pigconite is most common in rocks poor in hypersthcne” (Holmes). “Clinopyroxene abundant; orthopyroxene common in coarse varieties but subordinate to clinopyroxene; pigeonite less common” (Walker, priv. com.). 8. Mineralogical data after Holmes (1929) and Walker (1935). “Orthopyroxene or pigeonite completely serpentinised. No determinations of 2 V less than 45°” (Walker, priv. com.). 9. Mineralogical data from Rosenbusch (1908), Rosenbusch-Osann (1923), and Tröger (1935). 10. Mineralogical data after Edwards (1938). Pigeouite (2 V = 0°-5°) sometimes abundantly present, together with clinopyroxene with rather higher 2 V, in some instances about 45° (subcalcic augite?) or with “fairly large 2 V,” probably diopsidic. 11. Mineralogical data from Tyrrell and Sandford (1933). 12. Mineralogical data from Washington (1922), and Fermor (1925). 13. Mineralogical data chiefly from Barth (1931c). The first comment to be made on these two tables is that while the normative compositions of pyroxenes have been calculated in terms of the three standard molecules, they cannot represent adequately the average modal composition. The presence of olivine in many of the rock-series, notably X, Y, 1, 4, 7, 10 and 12, would make the normative pyroxene usually appear less ferruginous than the average modal composition, since, except in the case of the Japanese rocks, we are dealing with normal, not unusually ferruginous rocks.* In the latter it is possible for modal olivine to be richer in iron than the accompanying clinopyroxene. (For description, discussion and explanation of an instance of this, sec Smith, 1941). The same would result from the presence of quartz, tridymite and/or cristobalite, e.g., in X and Y and occasionally in 12, and from the presence of quartz and olivine together in rocks of several of the series of dolerites. Further, as Hess (1941, p. 587) indicates, the presence of Fe2O3 and TiO2 in modal pyroxenes which appear as iron ores in the norms, has the same general effect. The presence of Al2O3 in the modal pyroxene would involve the entry into the latter of CaO normatively alloted to anorthite. Hence En is often higher, Wo and Fs lower in the normative than in the modal pyroxene.† The reverse relation seems to hold, however, in regard to the average normative and modal compositions of the pyroxene in the in part olivine-bearing Karroo dolerites. (See Fig. 9, p. 117.) The average modal pyroxene composition of the Otago clinopyroxene as optically determined, viz., Wo26 En52

Fs22 when contrasted with the normative average Wo18 En60 FS22 affords a partial example of this. It may further be added that comparison of the normative compositions of the Palisadan dolerites and their effusive equivalents, the Watchung basalts, gives an interesting example of the rule that the latter, while containing about the same amount of total alkalies are richer in soda and also in iron. These differences “probably indicate that the basalts represent the parent magma more closely than the dolerites, the latter having advanced a stage in differentiation” (Tyrrell, 1933, p. 311). On the other hand, the composition of the Oamaru tachylite (No. 1), except for its low content of TiO2, resembles very closely the calculated average composition (No. 6) of the Mid-Tertiary basic magma from which it was derived. Nevertheless the development of these Mid-Tertiary Otago igneous rocks is not precisely in accord with that specified by Kennedy (1931, 1933) as typical of tholeiitic rocks. Not only is there a greater development of olivine in the more basic members, but pyroxene phenocrysts (not, however, markedly titaniferous) are common in the perphyritic members, and the trend of pyroxene-differentiation is marked by almost constant content of En with Fs increasing at the expense of Wo as in the olivine basalts rather than by decreasing En as specified by Kennedy for tholeiitic rocks, and by the absence of true pigeonite. Moreover, the close association of rocks yielding analcitic anorthoclase syenite pegmatoid with those developing quartz granophyre pegmatoid shows that here rocks with features resembling those of derivatives of the olivine basalt magma-type are not as sharply separated from rocks with features held to be distinctive of the tholeiitic magma-type as Kennedy held to be always the case. (h) The Absence of Pigeonite from the Dolerites of Northeastern Otago and its Explanation. If the available analyses of these Mid-Tertiary igneous rocks of Otago, averaged in the proportions tentatively assumed, yield a sufficiently representative average bulk normative composition, it is noteworthy, in view of the occurrence of several phases of pyroxene in other rock-series containing more calcic average normative pyroxene (Palisade, Karroo, Tasmania), that there should occur in the Otago dolerites only the single phase of clinopyroxene ranging uninterruptedly from diopsidic to calcic and subcalcie augite, except for such rare occurrences of hypersthene as are (or perhaps may be) the product of reaction between the magma and its contained quartzose xenoliths. The extent and nature of the crystallisation of the magma prior to its eruption, and especially the rapidity of its subsequent consolidation, appear to be the main factors in determining this contrast in mineralogical rather than in chemical features. In the Palisade and Tasmanian dolerites, orthopyroxene occurs as microphenocrysts in the chilled marginal phase, where it was the first femic mineral to form after olivine if the latter be present (cf. Walker, 1940, p. 1072; Benson, 1917, p. 33; Edwards, 1942, p. 587). In the basaltic contact phase of the Palisade sill it occurs “in minute

granules surrounding olivine, and is obviously a reaction-product.” In that of the Karroo sills, and indeed in the Karroo dolerites in general, it occurs only in the absence of olivine, with which it again may have a reaction-relationship (Walker and Poldervaart, 1941, pp. 140,143). In all three of these regions the magma was injected, usually in large volumes, into sub-aerially formed sandstones and mudstones sufficiently long after their deposition to allow such compaction as would have materially reduced the amount of their contained water.‡ The decrease of pore-space was probably less marked in some Triassic sediments when invaded by the Triassic Palisadan magma (see Shannon, 1924, p. 39, quoted below) than it was in the Permian to Jurassic sediments when invaded by the Karroo and Tasmanian dolerite magma, or in the marine and continental Carboniferous to Jurassic sediments when invaded by the Cretaceous dolerite magma of Spitzbergen (Tyrrell and Sandford, 1933). The specific heats and diffusivities of the invaded sediments were thus relatively low, so that large intrusive masses cooled sufficiently slowly to allow the development of more or less stable equilibrium-phases of pyroxene through a prolonged series of inversions and reactions—e.g., olivine → bronzite → early → pigeonite → augite → late pigeonite → carried to varying stages of completion, and the formation of exsolution lammellae in one or more of them (cf. Walker and Poldervaart, 1942, pp. 438–9). Conditions were quite otherwise as regards the North-East Otago rocks. The marginal tachylites and consequently the invading magma carried crystals of plagioclase of usually some olivine, but little or no pyroxene, which, in the case investigated (5831), was normal augite (2 V = 54° -49°). It was injected for the most part in comparatively small volumes into recently deposited and often still plastic marine sediments which were at the time below the sea.* Exceptions to this are afforded by the dolerites of North and South Peák (5716,5722) which invade the Upper Cretaceous sandstone 500–1000 or more feet below the level into which the Mt. Charles and Moeraki Sheets were injected. They contain, however, only normal augite (2 V = 51°–50°), and for various reasons do not afford a critieal test of this explanatory hypothesis. They would have a high content of water, and consequently relatively high specific heat and diffusivity (Clark, 1942, p. 258). Cooling of the magma was therefore rapid, and would be hastened by the latent heat of vapourisation of water absorbed into the magma (cf. p. 98). As a result, though the magmas contained enough water to permit as coarse crystallisation as, and a greater degree of gravitational differentiation than that in the much thicker sills of less aqueous magma injected into the sediments enclosing the thick Palisade, Karroo and Tasmanian sills, consolidation was too rapid to permit the attainment of equilibrium by reactive partition between the several types of pyroxene, and the course of crystallisation, starting at an abnormally low temperature, passed relatively quickly and uninterruptedly from the formation of stable normal augite to that of outer zones or individual crystals of largely metastable subcalcic augite with increasing content of Fs, following the trend explained by Dr. Edwards in Appendix II hereto.

Shannon (1924, p. 39) has described an exceptional sill in shales of the Newark series belonging to the general Palisadan assemblage. “Shales are highly hydrated rocks, and the most conspicuous features of the shales adjacent to the intrusive, both at Goose Creek and elsewhere, is a loss of shaly structure and a compacting and hardening doubtless due to loss of water. A body of molten magma of diabasic composition, surrounded on all sides by hydrous shales, would certainly tend to increase its content of dissolved water by solution of the highly heated water of the adjacent shales. In sandstones there would be less necessity for the water to dissolve in the magma, since it would be more free to move outward from the heated zone. This may explain the greater frequency of the occurrence of differentiation and other aqueous effects in shales than in those in sandstones or relatively anhydrous rocks. … If the magma crystallises from the early cooled walls inward, there must be a concentric inward expulsion of water, which in the ideal case would result in centrally placed pegmatite. …” It is worthy of remark that the pyroxene in the pegmatoid of the Goose Creek sill (op. cit., p. 11) is a uniform augite with “2 V medium”, Wo31 En40 Fs29 by chemical analysis, and without a trace of exsolution lamellae or indication of the presence of pigeonite. On the other hand, the pegmatoid phases of the less aqueous Tasmanian and Spitzbergen dolerites contain both normal augite and pigeonite (Edwards, 1942, p. 479; Tyrrell and Sandford, 1933, p. 306). The contrast exhibited by British rocks between the varied development of pyroxene in quartz-dolerites and the usually (?) less diversified range in the more quickly chilled tholeiites (see Table VIII) may be noted, also the rarity or absence of orthopyroxenes in the plateau basalts, etc. Appendix I. Augite, Subcalcic Augite, Pigeonite and the Classification of Non-Alkaline Clinopyroxenes. Discussions as to the rǒle of pigeonite in basic igneous rocks have been confused by the wide diversity of significance attached to the term pigeonite by various authors. The history of its usage (in part displayed in Table IX) has been summarised by Fermor (1925), Barth (1931a) and Hess (1941), the last of whom has put forward a revised classification of the pyroxenes of normal mafic magmas, which has been interestingly discussed by Walker (1943). There has also been a diversity of usage of the term hypersthene-augite, and of opinion as to the existence in rock-magmas of partial immiscibility between “augite” and pigeonite (with 2 V less than 32°) as most workers have believed, or of complete immiscibility under all conditions of cooling, as Hess (1941, pp. 518–519, 588) holds in the case of clinopyroxenes containing En and Fs in such proportions as occur in normal basic igneous rocks. “In practically all of them,” he remarks, “there is a gap in optic axial angle between 2 V = 30° and 2 V = 40° as Kuno (1936) observed.” The present writer has attempted an independent checking of this generalisation (Table X) in which he has been greatly helped by Professor Walker, who kindly furnished unpublished details of optic axial angles of pyroxenes in Scottish dolerites, and those of

the Palisade sill, and with Dr. Poldervaart, a large number of angles measured by them for pyroxenes in the Karroo dolerites. Moreover, Professor Walker has kindly provided the following additional details of measurements of 2 V made chiefly by the Mallard method with accuracy of only ± 4° Palisade Dolerite: 50° (1); 50°–45° (23); 45°–40° (14); 40°–35° (5); 35°–30° (3); 30°–20° (9); 20°–10° (7); 0° ± (10) with 28 measurements of orthopyroxene about as abundant as clino-pyroxene in the lower third of the sill, though subordinate in the higher portions. Scottish Dolerites and Tholeiites. 2 V > 52° (0), 51°–46° (14), 45°–36° (3), 35°–25° (1), 24°–0° (5). Orthopyroxene, though subordinate to clinopyroxene, is common in the coarse varieties; pigeonite is less common. Table IX Comparative Classifications of Clinopyroxenes with Molecular Ratio En: Fs. < 80/20 > 50/50 Note.—Wager and Deer's (1939, p. 242 and footnote) definition of pigeonite was based on chemical composition (Wo 10%–40%) and not on optical axial angle, for which the above table suggests a mean value. It foreshadowed the possibility of subdivision into metastable pigeonites and stable plutonic pigeonites. In addition to the above grouped data it may be noted that (with the exception of those for Karroo dolerites) recorded numbers of individual measurements between 40° and 15°, parallel to (010) are:—40° (20), 39° (7), 38° (13), 37° (3), 36° (8), 35° (2), 34° (6), 32° (1), 30° (15), 29° (1), 27° (1), 26° (1), 22° (2), 20° (4), 18° (1), 17° (5), 16° (2), 15° (1). The notable peak at 30° with few immediately adjacent records suggests that the rarity of such records immediately above 30°, from which Hess inferred the existence of a total immiscibility gap between augite and pigeonite, may arise at least in part from the natural tendency to estimate a value to the nearest round number when the means for determining are not very precise. Moreover, as Hess (1941, p. 524, Fig. 2A and B) indicates, omission of reference to the orientation of the optic axial plane in the case of measurements of 2 V near 30° may obscure the relations

Table X. Optic Axial Angles Recorder for Pyroxenes in Representative Volcanic and Hypabyssal Basic and Intermediate Igneous Rocks. Rocks Semi-alkaline and Alkaline Basalts, etc. Normal and Olivine Basalt Dolerite Andesite 2V Ap Ag Bc Bm Cp Cg Dp Dg Ep Eg Fp Fg Gp Gg H I J K Lp Lg M Totals* The totalled number of determinations of optic axial angles in the several intervals cannot represent the actual relative abundance of the corresponding pyroxene, since very commonly only the values of 2V in cores and in margins are given in the literature. Intermediate values of 2V, especially in the group 2V = 50°–45° must, therefore, be proportionately more abundant than the several totals in the table would indicate. > 57° DTS 14 16 5 4 4 6 3 0 9 8 5 3 8 10 2 3 0 1 11 0 5 107 57°–55° Augite 10 15 3 0 1 10 2 1 6 5 2 1 0 0 6 0 0 1 12 0 4 82 55°–50° Augite 17 5 5 1 1 6 15 8 2 1 1 3 6 8 22 0 11 21 18 0 ? 150 50°–45° Augite 16 8 7 3 0 1 15 12 0 6 4 0 5 9 32 10 77 30 49 6 5 295 45°–40° Subcalcic Augite 4 2 1 3 0 0 9 7 1 0 0 0 0 0 16 4 59 10 26 5 6 153 40°–35° Subcalcic Augite 1 1 0 1 0 0 3 10 0 6 0 0 0 0 14 4 18 0 4 5 4 71 35°–30° Subcalcic Augite 0 0 0 0 0 0 2 3 0 0 0 0 0 0 1 1 18 0 0 0 0 25 30°–25° Pigeonite 0 0 0 0 0 0 2 6 0 4 0 0 0 0 0 0 — 0 1 3 1 17 25°–0° Pigeonite 0 0 0 0 0 0 5 8 0 0 0 0 0 0 0 0 — 0 4 10 2 29 0° ± Pigeonite 0 0 0 0 0 0 0 7 1 8 0 0 0 0 0 1 43 0 2 5 0 67 0°–30° Pigeonite 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 19 0 0 0 23 > 30° Clh. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Orthop. 0 0 0 0 0 0 13 5 R 0 0 0 0 0 ? 5 38 0 11 7 1 Abun.

DTS = Diopside, Calcic Titanaugite and Salite. Clh. = Clinohypersthene. The following are the rock series included in the above:— A. Phenocrystic (p) and groundmass (g) titanaugites in Late Tertiary basaltic rocks of E. Otago, N.Z. Benson and Turner (1939, 1940), Benson (1942a), Paterson (1942). The so-called orthopyroxene of these papers is olivine. B. Titanaugite (c = cores, m = margins) in homogeneous or zoned crystals in olivine theralite, Waihola, E. Otago, Turner and Benson (1942). C. Augite in semi-alkaline and alkaline basaltoid rocks of Pacific and Kurile Island: (p), (g) as above. Barth (1931), Kuno (1935). D. Augite is often iron-rich basalts of Huzi zone, Japan, sometimes containing olivine with quartz, tridymite, and/or cristobalite: (p) and (g) as above Tsuya (1937). See also Sugi (1937). E. Augite in normal basalts and olivine andesites of the Pacific, (p) (g) as before (R = rare). Barth (1931). (For more detailed analysis see Edwards, 1935, p. 18.) F. Augite (p and g) in basalts and olivine andesite of Banks Peninsula, N.Z. Turner, in Benson and Turner, 1939. (No orthopyroxene, see A.) G. Augite (p and g) in olivine basalt. Auckland, N.Z. Turner, in Benson and Turner (1939). No orthopyroxene. H. Augite and subcalcic augite in dolerites, etc., N.E. Otago. I. Pyroxenes in Tasmania dolerites. Edwards (1942). J. Pyroxenes in the Karroo dolerites measured chiefly by Dr. A. Poldervaart (priv. com.).§ Values of 2V in the ranges 31°–25° and 25°–0° in Karroo pyroxenes are here omitted as only five such angles, all doubtfully measured, were noted by Poldervaart. Including some of these and others possibly less accurately observed by earlier workers 10 angles in the range 25°–0° are recorded in the literature. Though 38 measurements of 2V in orthopyroxenes were made, the number (fide Professor Walker) over-estimates the quantitative abundance of this pyroxene, K. Unzoned pyroxene in marginal olivine dolerite. Hallefors, Sweden, showing Gauss variation about the mean value 49°. Krokström (1936). L. Pyroxene (p) and (g) in Japanese andesites, often containing quartz, tridymite and/or cristobalite and olivine in some cases. Kuno (1936), Tsuya (1937). M. Pyroxene in Cascade Andesites of Oregon. Bogue and Hodge (1940). of pigeonite to “augite.” Though the data collected into Table X do not lead to as sharp a separation between them as that shown in Hess's Fig. 2B, they support the inference which may be drawn from Walker (1943, p. 518) cited below, that the continuity in the series of clinopyroxenes is least marked beneath rather than above the value 2V = 30° in the plane parallel to (010). Table IX illustrates the varying usage that has been attached to the term pigeonite. The conventional usage as denoting clino-pyroxenes with 2V = 0°–45° is that which has been followed by such authors as Winchell, Lacroix, Ford, Kuno and Tsuya. Wahl, Rosenbusch, Flett, Holmes and Fermor are among those who have used enstatite-, bronzite- or hypersthene-augite in about the same significance. While limiting pigeonite to pyroxenes with 2V < 32° or 30°, Walker (1940, 1941) used “hypersthene-augite” for the more calcic pyroxenes with 2V = 30°–50° or 52°, which had been included by others within the term pigeonite, but he later (1942) abandoned it. It seems, however, that a new term for such pyroxenes would be useful, and subcalcic augite is here suggested to include augitic pyroxenes with 2V less than 45°—i.e., all pyroxenes with

2V > 30° hitherto termed pigeonite by most petrologists other than Walker and Hess. An alternative upper limit, 2V < 48°, suggested but not stressed by Professor Walker in his very generous and helpful letter, would include within the subcalcic augites nearly all the Karroo clinopyroxenes other than pigeonite, and the majority of those in the Tasmanian dolerites other than pigeonite. Time will show which of these or other alternative limits is the most useful, if the recognition of the subcalcic augites as a petrologically significant group should obtain general acceptance. A further suggestion concerning the minimum value of 2V in subcalcic augites is advanced with diffendence. Walker (1943, p. 518) “is quite prepared to accept 32° as the maximum value for 2V (in pigeonite), but considers that the limit should be without reference to the orientation of the optic axial plane. In his experience pigeonites with the optic axial plane parallel to (010) are most frequently found as the margins of strongly zoned augite crystals with which they are in perfect optical continuity, and their optic axial angle does not fall much below 32°. They are the result of extreme zoning and show complete gradation into normal augite. Such pigeonites are strongly ferriferous, and lie just below the ferroaugite field in the triangular diagram. Pigeonites with the optic axial plane perpendicular to (010) are of quite a different nature. They may precede augite in the crystallisation of diabases and dolerites, or may come after it, but in both cases the boundary between pigeonite and augite is sharply defined under the microscope, indicating a discontinuity. In the first case the mineral is magnesian, and may be a reaction product of magnesian olivine or of bronzite, in the second it is ferriferous, being a lime-poor equivalent of ferroaugite.” As it is desirable where it is practically possible, that mineral nomenclature should express genetic features, and the orientation of the optic axial plane is generally determinable in pyroxenes with 2V within a few degrees of 30°, it is tentatively suggested that the pyroxenes of the first group distinguished by Walker, including those with 2V = 30° should be classed with the subcalcic augites, or subcalcic ferroaugite. This may mean that some of the rare pyroxene with 2V in the plane parallel to (010) is a degree or two less than 30° would be classed as subcalcic augite, and would mean that the composition-range of this group of pyroxenes extends across the narrow, sparsely-tenanted interval immediately above the line of 2V = 30° assumed by Hess (1941) to be the field of total immiscibility separating augite from pigeonite, and also that the boundary between subcalcic augite and pigeonite lies in the much wider and more sparsely tenanted interval immediately below 2V = 30° which was assumed by Hess to form portion of the field of continuous variation of pigeonite. The suggestion here advanced as to the upward limit of the pigeonite-field accords, however, very closely with Hess's alternative proposal to accept the line of Wo = 15% as that limit. (See Fig. 5 on p. 83.) The pigeonites as here limited might be divided on statistical grounds into magnesian pigeonites, pigeonite, and ferropigeonite according as the molecular ratio En:Fs therein is > 2:1, 2:1 < > 1:1 or < 1:1, but any division should be considered on chiefly genetic grounds. The first two members of the above trio are represented

respectively by the early and late pigeonites of Walker and Poldervaart (1941, Fig. 5. See also our Fig: 9), the second and third occur (the third but rarely) in the groundmass of the unusually furruginous Japanese basalts (Tsuya, 1937), but the second also forms phenocrysts or more probably xenocrysts (Kuno, 1936, p. 125) in a Japanese andesite and ferropigeonite (Wo 9 En40 Fs51 in molecular ratios, M on Fig. 8) forms rounded phenocrysts in the semi-vitreous dacitic inninmorite of Mull (Hallimond, 1914; Thomas and Bailey in Anderson and Radley, 1916, p. 209). The conditions under which iron-rich pigeonites may have formed have been discussed by Holmes and Harwood (1928, pp. 507–8), Bowen and Schairer (1935, p. 203), Kuno (1936, p. 149), Wager and Deer (1939, p. 254), and Hess (1941, pp. 589–591), who (p. 581) concludes with Phemister (1934, p. 58) that in general “hypersthene with oriented plates is characteristic of slowly cooled intrusives, and pigeonite” (as here recognised) “is absent in them. Pigeonite is characteristic of rapidly cooled extrusives, and hypersthene with oriented plates is absent in them. Whereas the intermediate type, the fine grained intrusives, may show both forms. In those rare cases where pigeonite began to crystallise in depth and the magma containing it was suddenly extruded, pigeonite phenocrysts will be found.” Walker and Poldervaart (1941, p. 132) remark that “there must be instability boundaries” [other than those defined on the basis of chemical composition by Wager and Deer (1939, p. 255)] “in the intervening zones between the volcanic rocks in which all compositions of pigeonite” (as here) “are stable, and the plutonic rocks, in which only the iron-rich varieties are to be found.” Edwards' (1942) views are re-stated below in his generously contributed Appendix II. Dr. F. J. Turner has suggested in conversation with the writer, that the relations of pigeonite, subcalcic augite and augite, in so far as they are not affected by magmatic water (see below) may in some measure be likened to those of orthoclase, anorthoclase and albite, and may perhaps be expressible by an equilibrium-diagram resembling that drawn by Barth (1939, p. 24, Fig. 8) for these feldspars on the basis of Schairer and Bowen's work. Such an equilibrium-diagram for clinopyroxenes would show on either side in the lower (plutonic) temperature-range the boundaries of the stability-fields of pigeonite and (diopsidic?) augite respectively, there possessing only limited miscibility. Traced to higher temperatures these boundaries would converge and eventually join, possibly at a “dry melt” temperature higher than the natural crystallising temperatures of basalt. Above this junction the stability-fields of the three clinopyroxenes would merge continuously into one another. Below it, at temperatures at which most basalts and dolerites crystallise, though subcalcic augite is often formed, it is metastable. This suggestion seems to accord with the observations both of Bowen and Schairer (1935) and of Hess (1941), and to afford a synthesis of their divergent explanations. The application of the above suggestions to the classification of clinopyroxenes in several rock-series is illustrated by Fig. 9, in which the compositions of representative clinopyroxenes are plotted in weight percentages, the molecular percentages being indicated by the

9. Nature and relation of pyroxenes in several series of rocks. For explanation see text. divisions along the margins. The limits drawn by Hess (1941, p. 518) for the fields of diopside, endiopside, salite, ferrosalite and heden-bergite are omitted. The small crosses indicate the optically determined compositions of phenocrysts in Japanese non-alkaline basalts and andesites, showing their normally augitic to subcalcic nature and lack of continuous variation towards the rare pigeonite phenocrysts indicated by the two larger asterisks marked by p. Triangles indicate the optically determined compositions of groundmass pyroxenes in the same Japanese rocks and demonstrate the continuous variation from the lower calcic range of augite through subcalcic augite to pigeonite and rarely ferropigeonite, for these rocks are abnormally rich in iron. (Data from Tsuya, 1937.) Black circles show the average normative compositions of the representative rock series considered in Table VIII as follows:—A = Antarctic dolerites; B = Plateau basalts; G = German tholeiites; K = Karroo dolerites; O = Otago dolerites, basalts and tholeiites; P = Palisadan dolerites; S = Scottish and North of England dolerites and tholeiites; Sp = Spitz-bergen dolerites; T = Tasmanian dolerites; W = Watchung basalts, the effusive products of the Palisadan magma. The black rings indicate the chemically determined composition of:—D = the pyroxene from the Deccan basalt submitted to X-ray spectrographic examination (see below) and M = the ferro-pigeonite in the inninmorite of Mull. I, II, III and IV show the fields of composition-range for successive generations of pyroxene in the Karroo dolerites as determined optically by Walker and Poldervaart (1941, Fig. 5) being respectively I = initial hypersthene, II = early magnesian pigeonite, III = “hypersthene augite”—i.e., subcalcic augite (with a little augite poor in lime), IV = late-formed pigeonite. It will be noted that the average composition of the modal pyroxene in the Karroo dolerites is less ferruginous than the average normative composition K. It will further be seen that the sequence I-II-III is almost exactly that found by Kuno (1933) in a Japanese cristobalite-bearing olivine, augite-hypersthene basalt and plotted by Alling (1936, p. 218), whose “Pigeonite No. 2″ (with 2V = 48°–50°) in Kuno's and our view [and indeed on his own classification (Alling, 1936, p. 102)] is an augite poor in lime. It lies on Alling's (op. cit., pp. 100, 218) co-tectic line (CT—CT on our Fig. 9) near the turning point in the

differentiation-trend of the Karroo pyroxenes from an increasingly to a decreasingly calcic composition, the physico-chemical significance of which has apparently not yet been discussed. There seems little indication in Fig. 9 that Tsuboi's (1932, p. 75) “Two pyroxene line” (PP-PP) has significance in regard to the development of these dolerite pyroxenes. The field and arrow N.Z. on Fig. 9 show the composition-range and differentiation-trend as determined for the pyroxenes of the Mid-Tertiary dolerites, tholeiites and basalts of Otago. In spite of the close approach of the average bulk composition of these rocks and of average normative compositions of their pyroxenes to those of the rock-series taken by Kennedy (1931, 1933) to be typical of his tholeiitic magma-type (see Tables VII and VIII above), the differentiation-trend of the Otago pyroxenes is that of the pyroxenes in the basic members of Kennedy's olivine basalt magma-type, and is perpendicular to the differentiation-trend I-II-III of the pyroxenes in rock-series taken by him to be typical of his tholeiitic magma-type, a feature which remains true even if we consider only the quartz-bearing dolerites among the Otago rocks. It would seem, therefore, that some significant factors have been omitted from Kennedy's discussion, and it suggested that the effect of high content of water in crystallising magmas may be among these. Phemister (1934, p. 1963) suggested that since “the nature of the equilibrium in the pyroxene group changes with the approach to plutonic conditions of crystallisation of the magma” (see above), “and since also the amount of water in the magma as it reaches saturation varies with the position of the mass in the crust, the petrological evidence leads to the suggestion that water may be the controlling influence in pyroxene-equilibrium.” Bowen and Schairer (1935, p. 201) doubted “whether the necessary quantity of volatiles could have been present and yet leave no evidence of their presence in the formation of hydrous phases such as amphibole, which are often entirely lacking.” But, as has been noted, though amphibole is extremely rare in our rocks, there is abundant independent evidence of the former presence of much magmatic water. It may, perhaps, be this which, at a late stage in magma-consolidation, has determined the almost rectangular change in the direction of differentiation-trend of the Karroo pyroxenes. Appendix II. A Note on the Crystallisation of Pyroxenes. By A. B. Edwards. “The progressive changes in composition developed in pyroxenes crystallising from a differentiating magma are governed by two dominant factors: (1) the composition of the residual magma at any stage, and (2) the stability, or otherwise, of the molecular lattice structure of the pyroxene concerned. “Recent studies by Hess (1941) and by Edwards (1942) indicate that under conditions of slow cooling, which permit the development of stable pyroxenes, the molecular structure of the pyroxene is the more important factor of the two in deciding the trend of change of composition. The difficulties of substituting Fe++ (ionic radius 1.83A°) for Ca++ (ionic radius 1.06A°) in the pyroxene structure without undue distortion of the crystal lattice, leads to the formation

of two immiscible series of pyroxenes, which crystallise side by side. The width of the immiscibility gap between these two series is a maximum in slowly cooled magmas, and decreases as cooling becomes more rapid, until in rapidly cooled rocks such a thin lava flows, sills and dykes, it is reduced to zero. “In rapidly cooled rocks, the single pyroxene that forms is in a metastable condition. Under such conditions Fe++ can substitute for Ca++ in the lattice structure of the pyroxene without any undue distortion resulting. As a result the molecular structure of the pyroxene ceases to be the dominant factor controlling its progressive changes in composition; and changes in composition shown by the pyroxene simply reflect the changing composition of the residual magma. Examples of the change induced by a sudden increase in the rate of cooling on the composition of the pyroxene have been provided by Tsuboi (1932) and C. N. Fenner (1938). Tsuboi records instances where augite and orthopyroxene crystallising as pheno-crysts have given place to a single “pigeonite” of intermediate composition crystallising in the groundmass. Fenner described an instance where magnesia-rich orthopyroxene crystallising as phenocrysts has given place to a groundmass pyroxene which is lime-rich augite and which more nearly reflects the composition of the residual magma as represented by an analysis of the aphanitic groundmass, than does the earlier formed orthopyroxene. “Since the prevailing change in composition in most residual magmas of basaltic character is towards an enrichment in iron relative to magnesium, the change of composition of the metastable pyroxene is generally towards an enrichment in iron in the marginal zones of the crystals, regardless of whether the original pyroxene is calcium- or magnesium-rich. This is shown by the results obtained by Barth (1931, 1936), by Kuno (1935) and still more recently by the data recorded in the foregoing paper.” To this the writer (W.N.B.) may add that Dr Edwards (1942, p. 599) recalls that the formation of metastable subcalcic pyroxene on relatively rapid cooling is in accord with the results of Bowen and Schairer's (1935) experimental work, and that there is a tendency for such pyroxenes gradually to change with separation and inversion to enstatite of the MgSiO3 in excess of diopsidic proportions. That this may occur submicroscopically seems indicated by X-ray powder spectrography. An apparently homogeneous pyroxene separated from the Deccan basalts and found by chemical analysis to have the composition Wo25 En37 Fs38 (D in Fig. 9 herewith) was examined by X-ray spectrography (Wyckoff, Merwin and Washington, 1925). “Its predominant line-pattern is diopside-like. Other lines are present, however, which could be the principal ones of an enstatite-like structure. It is thus compatible with the observed measurements, though these measurements do not furnish final proof that this Indian augite is an intimate mixture of materials having these two structures. The density of such a mixture could be calculated on the assumption that the diopsidic content is normal with respect to calcium, and that the (Fe, Mg) in excess of the amount necessary to combine with it is present as hypersthene. The density calculated on this basis is close to the observed density.”

Bibliography Alling, H., 1936. Interpretative Petrology. McGraw Hill, New York. Anderson, E. M. and Radley, E. G., 1916. The Pitchstones of Mull and their Genesis. Quart. Journ. Geol. Soc., vol. lxxi, pp. 205–217. Barber, C. T., 1936. The Tertiary Igneous Rocks of the Pakokku District. Mem. Geol. Surv., India, vol. lxviii, pp. 121–292. —– 1936a. Effect of Heat on the Optical Orientation of Plagioclase Feld-spars. Min. Mag., vol. xxiv, pp. 345–352. Bartii, T. F. W., 1931. Permanent Changes in the Optical Orientation of Feld-spars Exposed to Heat. Norsk. Geol. Tidsskrift, vol. xii, pp. 57–72. —– 1931a. Crystallisation of Pyroxene from Basalt. Amer. Miner., vol. xvi, pp. 195–208. —– 1931b. Proposed Changes in Calculation of Norms of Rocks. Min. Petr. Mitt., vol. xxxxii, pp. 1–7. —– 1931c. Mineralogical Petrography of Pacific Lavas. Amer. Journ. Sci., vol. xxxi, pp. 377–405, 491–530. —– 1936. The Crystallisation Processes of Basalt. Amer. Journ. Sci., vol. xxxi, pp. 321–351. —– 1939. In Barth, T. F. W., Correns, C. W., and Eskola, P., Die Entstehung der Gesteine. Julius Springer, Berlin. Benson, W. N., 1916. Report on the Petrology of the Dolerites Collected by the British Antarctic Expedition, 1907–9; B. A. Exped. 1907. Rept. Sci. Invest. Geology, vol. ii, pp. 153–160. —– 1917. Notes on the Geology of the Cradle Mountain District. Papers and Proc. Roy. Soc. Tas. for 1916, pp. 29–43. —– 1941. Basic Igneous Rocks of Eastern Otago and Their Tectonic Environment. Part I. The General Tectonic Environment. Trans. Roy. Soc. N.Z., vol. lxxi, pp. 208–222. —– 1942. a and b Ibid. Part II. Distribution, Petrography and Chemical Composition of the Late Tertiary Igneous Rocks. Loc. cit., vol. lxxii, pp. 85–110. Part III, The Olivine Theralite of Waihola, a Gravitationally differentiated Sill. Loc. cit., vol. lxxiii, pp. 116–185. —– 1943. Ibid. Part IV, Section A. The Mid-Tertiary Basalts, Tholeiites and Dolerites of North-Eastern Otago. Loc. cit., vol. lxxiii, pp. 116–138. Benson, W. N. and Turner, F. J., 1939. Mineralogical Notes from the University of Otago, no. 2 Trans. Roy. Soc. N.Z., vol. lxix, pp. 56–72. —– 1940. Mugearites in the Dunedin District. Trans. Roy. Soc. N.Z., vol. lxx, pp. 188–190. Bogue, R. and Hodge, E. T., 1940. Cascade Andesites of Oregon. Amer. Miner., vol. xxviii, pp. 627–665. Bowen, N. L., 1910. Diabase and Granophyre of the Gowganda Lake, Ontario. Journ. Geol., vol. xviii, pp. 658–674. —– 1913. The Melting Phenomena of the Plagioclase Feldspars. Amer. Journ. Sci., vol. xxxv, pp. 557–599. Bowen, N. L., and Schairer, J. F., 1935. The System, MgO-FeO-SiO2 Amer. Journ. Sci., vol. xxix, pp. 151–217. Browne, W. R., 1923. The Dolerites of King George Land and Adelie Land. Sci. Rept. Australasian Antarct. Exped. 1911–1914, Series A, vol. iii, pp. 246–258. Chapman, W. H., 1936. A Study of Feldspar Twinning in a Differentiated Sill: Amer. Miner., vol. xxi, pp. 33–47. Clark, H., 1942. Table 17.4.1 Handbook of Physical Constants. Special Paper Geol. Soc. Amer., no. xxxvi, p. 256. Daly, R. A., 1917. The Geology of Pigeon Point, Minnesota. Amer. Journ. Sci., vol. xxxxiii, pp. 423–448. —- 1933. Igneous Rocks and the Depths of the Earth. McGraw Hill, New York, pp. 307–311. Day, A. L. and Allen, E. T., 1925. The Volcanic Activity and Hot Springs of Lassen Peak. Carnegie Inst. Washington. Publ. no. 360. Deer, W. and Wager, L. R., 1938. Two New Pyroxenes Included in the System Clino-enstatite, Clino-ferrosilite-Diopsode-Hedenbergite. Min. Mag., vol. xxv, pp. 15–22. Edwards, A. B., 1935. The Olivine Basalt-Trachyte Provinces and Some Theories of Petrogenesis. Proc. Roy. Vic., vol. xxxxviii, (N.S.), pp. 13–26.

—– 1938. Tertiary Tholeiite Magma in Western Australia. Journ. Roy. Soc. W. A., vol. xxiv, pp. 1–12. —– 1942. Differentiation in the Dolerites of Tasmania. Journ. Geol., vol. 1, pp. 451–480, 579–610. Emerson, B. K., 1905. Plumose Diabase and Palagonite from the Holyoke Trap Sheet. Bull. Geol. Soc. Amer., vol. xvi, p. 91. Emmons, R. C., 1927. Diabase Differentiation. Amer. Journ. Sci., vol. xiii, pp. 73–82. Fenner, C. N., 1910. Crystallisation of a Basaltic Magma from the Standpoint of Physical Chemistry. Amer. Journ. Sci., vol. xxix, pp. 221–253. —– 1926. The Katmai Magmatic Province. Journ. Geol., vol. xxxiv, pp. 675–771. —– 1929. The Crystallisation of Basalt. Amer. Journ. Sci., vol. ccxvi, pp. 221–253. —– 1931. The Residual Liquids of Crystallising Magmas. Min. Mag. vol xxii, pp. 339–360. —– 1938. Olivine Fourchite from Raymond Fosdick Mountains, Antarctica. Bull. Geol. Soc. Amer., vol. xxxxix, pp. 367–400. Fermor, L. L., 1925. Note on the Enstatite- Augite Series of Pyroxenes. Rec. Geol. Serv. India, vol. lviii, pp. 323–9. —– 1931. A Further Note on Chlorophaeite and Palagonite. Geol. Mag., vol. lxviii, pp. 266–274. Ford, W. F., 1932. Dana's Textbook of Mineralogy; 4th Edit. John Wiley, New York, p. 358. George, W. O., 1924. The Relation of the Physical Properties of Natural Glasses to their Chemical Composition. Journ, Geol., vol. xxxiv, pp. 353–374. Grout, F. F., 1928. Anorthosite and Granite as Differentiates of a Diabase Sill on Pigeon Point, Minnesota. Bull. Geol. Soc. Amer., vol. xxx, pp. 555–578. —– 1932. Petrography and Petrology. McGraw Hill, New York, pp. 210–13. Hallimond, A. F., 1914. Optically Uniaxial Augite from Mull. Min. Mag., vol. xvii, p. 97. Harker, A., 1904. The Tertiary Rocks of Skye, Mem. Geol. Surv. United Kingdom Grt. Brit., p. 353. —– 1916. Differentiation in Intercrustal Magma Basins. Journ. Geol., vol. xxiv, pp. 354–358. Harris, N., 1937. A Petrological Study of the Portrush Sill, and its Veins. Proc. Roy. Irish. Acad., vol. xxxxiii, Sect. B, pp. 95–113. Hess, H. H., 1941. Pyroxenes of Common Mafic Magmas. Amer. Miner., vol. xxvi, pp. 515–535, 573–579. Holmes, A., 1921. Petrographic Methods and Calculations. T. Murby and Co., London, p. 324. Holmes, A. and Harwood, H. F., 1928. The Age and Composition of the Whin-Sill and related Dykes of the North of England. Min. Mag., vol. xxi, pp. 493–542. —– 1929. The Tholeiitic Dykes of the North of England. Min. Mag., vol. xxii, pp. 1–52. Hutton, F. W., 1887. The Geology of the Country Between Oamaru and Moeraki. Trans. N.Z. Inst., vol. xix, pp. 415–430. —– 1889. The Eruptive Rocks of New Zealand. Journ. Proc. Roy. Soc., N.S.W., vol. xxiii, pp. 102–156. Kennedy, W. Q., 1931. The Parent Magma of the British Tertiary Province. Sum. Prog. Geol. Surv. Gt. Brit. for 1930, Part 2, pp. 62–67. —– 1933. Trends of Differentiation in Basic Magma. Amer. Journ. Sci., vol. ccxxv, pp. 239–256. Kohler, A., 1941. Die Abhangigkeit der Plagioklas vom vorgangen Warmverhalten. Die Existenz einer Hoch- und Tief-temperaturoptik. Min. Petr. Mit. (Tschermak), vol. liii, pp. 24–49. (Abst. in Min. Mag. (Min. Abst.) 1943, vol. viii, p. 313.) Krokström, T., 1936. The Hallefors Dolerite Dyke and Some Problems of Basaltic Rocks. Bull. Geol. Unst. Univ. Upsala., vol. xxvi, pp. 115–263. Kuno, H., 1933. Preliminary Note on Basalt from Aziro, Idu. Bull. Earthquake Research Inst. Tokyo., vol. xi, pp. 391–401. —– 1935. Petrology of Alaid Volcanoes, North Kurile. Jap. Journ. Geol. Geog., vol. xii, pp. 153–162.

—– 1936. Petrological Notes on Pyroxene Andesites … Hakone Volcano, etc., Ibid., vol. xiii, pp. 107–140. —– 1936a. On the Crystallisation of Pyroxene from Rock Magmas, with Special Reference to the Formation of Pigeonite, Ibid., pp. 141–150. Lacboix, A., 1893. Les Enclaves des Roches volcaniques. Protat Freres, Macon. —– 1927. La Constitution Lithologique des Iles Volcaniques de la Polynésie Australe. Mem. Acad. Sciences. tom. lix, pp. 1–80. Leitmeier, H., 1909. Der Basalt von Weitendorf in Steiermark und die Mineralien seine Holraume. Neues Jahrb. für Min. Beil. Bd. xxvii, pp. 219–260. Lewis, J. V., 1907. Petrography of the Newark Igneous Rocks of New Jesey. Ann. Rept. State Geol. N.J. for 1907, pp. 98–169. Mantell, G. A., 1850. On the Remains of the Dinornis and the Fossils and Rock Specimens Recently Collected by G. W. Mantell in the Middle Island of New Zealand, etc. Quart. Journ. Geol. Soc., vol. vi, pp. 319–343. Marshall., P., 1906. Geological Notes on the Country North-West of Lake Waka-tipu. Trans. N.Z. Ins., vol. xxxviii, pp. 561–568. Morey, G. W., 1922. Development of Pressure in Magmas as a Result of Crystallisation. Journ. Wash. Acad. Sci., vol. xii, pp. 219–230. Nemoto, T., 1938. A New Method of Obtaining Extinction Angles of Monoclinic Minerals … etc. Journ. Fac. Sci., Hokkaido. Imp. Univ., vol. iv, pp. 107–111. Nikitin, W., 1936. Die Federow Methode. Gebruder Borntrager, Berlin Paterson, O. D., 1942. The Geology of the Lower Shag Valley, N.E. Otago. Trans. Roy. Soc., N.Z., vol. lxxi, pp. 32–58. Peacock, M. A., 1930. The Distinction Between Chlorophaeite and Palagonite. Geol. Mag.., vol. lxvii, pp. 170–178. Phemister, T. C., 1928. A Comparison of the Keweeuawan Sill-Rocks of Sud-bury and Cobalt, Ontario, Trans. Roy. Soc. Canada, Third Series, vol. xxii, Sec. 4, pp. 121–196. —– 1934. The Role of Water in Basaltic Magma. Tschermaks Min. Petr. Mitt., vol. xlv, pp. 19–77, 99–132. Phillips, A. H., 1898. Structure and Composition of the Trap of Rocky Hill, N.J. Amer. Journ. Sci. 4th Ser., vol. viii, pp. 267–285. Prior, G. T., 1907. Report on the Rock Specimens collected during the “Discovery” Antarctic Expedition, 1901–4. National Antarctic Exped. Nat. Hist., vol. i, Geology, Part 2, chapter 5. Rosenbusch, H., 1907–8. Mikroscopische Physiographie der Massigen Gesteine. Bd. 2, Schweizerbartsche Verlagshandiung. Stuttgart. —– 1910. Elemente der Gesteineslehre, 3rd Edit., p. 400. Schweizerbartsche Verlagsbuchhandlung, Stuttgart. Rosenbusch, H. and Osann, A., 1923. Elemente der Gesteinslehre; 4th Edit. Schweitzerbartsche Verlagsbuchhandlung, Stuttgart. Scholler, H., 1942. Versuche zur Temperatur-abhängigkeit der Plagioklas Optik, Min. Petr. Mit. (Tschermak), vol. liii, pp. 180–221. (Abst. in Min. Mag. Min. Abst., vol. viii, 1943, p. 313.) Shannon, E., 1920. Diabantite, Stilpnomelane and Chalcedony of the Trap. Quarries of Westfield, Massachusetts. Proc. U.S. Nat. Mus., vol. lvii, pp. 397–403. —– 1924. The Mineralogy and Petrology of the Intrusive Diabase at Goose Creek, Loudon County, Virginia. Ibid., vol. lxii, pp. 1–86. Smith, A. P., 1941. Olivine and Pyroxene of Mt. Tripyramid, New Hampshire. Amer. Min., vol. xxvi, pp. 202–3. Sugi, K., 1937. Note on the Occurrence of Zoned Pigeonitic Augite in the Dolerite from Semi, near Sinzyo, Uzen. Journ. Geol. Soc. Jap., vol. xliv, pp. 255–261. Teall, J. J. H., 1884. On the Chemical and Microscopical Characters of the Whin Sill. Quart. Journ. Geol. Soc., vol. xxxx, pp. 640–657. —– 1889. On the Amygdaloids of the Tynemouth Dyke. Geol. Mag., Dec. 13, vol. vi, pp. 481–3. Tertsch, H., 1942. Zur Hochtemperatur Optik der Plagioklas. Zentbl. Min., for 1942, pp. 137–144. (Abst. in Min. Mag. Min. Abs., vol. viii, p. 313.) Thomson, J. A., 1906. The Gem Gravels of Kakanui, with Remarks on the Geology of the District. Trans. N.Z. Inst., vol. xxxviii, pp. 481–494.

—– 1907. Inclusions in Some Volcanic Rocks. Geol. Mag., Dec. 5, vol. iv, pp. 490–500. Tilley, C. E., 1922. Density, Refractivity and Composition of Some Natural Glasses. Min. Mag., vol. xix, pp. 275–294. Tomita, T., 1934. Variations in Optical Properties According to Chemical Composition in the Pyroxenes of the Clino-enstatite Clino-hypersthene-Diopside-Hedenbergite system. Journ. Shanghai Inst., vol. i, pp. 41–58. Tomkieff, S. T., 1929. A Contribution to the Petrology of the Whin Sill, Min. Mag., vol. xxii, pp. 100–120. —– 1941. Metasomatism in the Basalt of Heddenrig Quarry, near Kelso, and the Veining of the Rocks Exposed There. Ibid., vol. xxvi, pp. 45–59. —– 1942. The Tertiary Lavas of Rum. Geol. Mag., vol. lxxix, pp. 1–13. Tröger, E., 1935. Spezielle Petrographie der Eruptivgesteine. Deutsches Min. Ges. Berlin. Tsubot, S., 1932. On the Course of Crystallisation of Pyroxenes from Rock Magmas. Jap. Journ. Geol. Geog., vol. x, pp. 67–82. Tsuya, H., 1937. On the Volcanism of the Huzi Volcanic Zone. Bull. Earthquake Res. Inst., Tokyo, vol. xv, pp. 216–357. Turner, F. J., 1940. Determination of Optic Axial Angle in Pigeonites, Amer. Min., vol. xxv, pp. 281–285. —– 1942. Determination of Extinction Angles in Monoclinic Pyroxenes and Amphiboles. Amer. Journ. Sci., vol. ccxxxxvi, pp. 571–583. —– 1943. Preferred Orientation of Olivine Crystals in Peridotite. Trans. Roy. Soc. N.Z., vol. lxxii, pp. 280–300. Tyrrell, G. W. and Sandford, K. S., 1933. Geology and Petrology of the Dolerites of Spitzbergen, Proc. Roy. Soc., Edinburgh, vol. liii, pp. 284–321. Uttely, G. H., 1918. The Volcanic Rocks of Oamaru, with Special Reference to Their Position in the Stratigraphical Succession. Trans. N.Z. Inst., vol. 1, pp. 106–117. Wager, L. R. and Deer, W. A., 1939. The Petrology of the Skaergaard Intrusion, Kangerdlugssuaq. East Greenland. Meddel. om Gronland, vol. cv, no. 4. Wahl, W., 1907. Die Enstatitaugite. Tschermaks Min. Petr. Mitt., vol. xxxvi, pp. 1–131. Walker, F., 1935. The Late Palaeozoic Quartz Dolerites and Tholeiites of Scotland. Min. Mag., vol. xxiv, pp. 131–159. —– 1940. Differentiation of the Palisade Diabase of New York. Bull. Geol. Soc. Amer., vol. li, pp. 1060–1105. —– 1943. Note on the Pyroxenes of Basalt Magmas. Amer. Journ. Sci., vol. ccxli, pp. 517–520. Walker, F. and Poldervaart, A., 1941. The Karroo Dolerites of the Calvinia District. Trans. Geol. Soc. S. Africa, vol. xliv, pp. 127–148. —– 1941a. The Hangnest Dolerite Sill, S.A. Geol. Mag., vol. lxxviii, pp. 429–450. —– 1942. The Petrology of the Karroo Dolerites between Sutherland and Middelburg, C.P. Ibid., vol. xlv, pp. 55–64. Washington, H. S., 1922. Isostasy and Rock Density. Bull. Geol. Soc. Amer. vol. xxxiii, pp. 375–410. —– 1922a. Deccan Traps and Other Plateau Basalts. Ibid., pp. 763–803. —– 1923. Petrology of the Hawaiian Islands Nos. 1–4. Amer. Journ. Sci., vol. v (5th Series), pp. 465–502; vol. vi, pp. 99–126, 338–367, 409–423. Washington, H. S. and Keyes, M. G., 1926. Ibid. no. v, Loc. cit., vol. xii, pp. 336–352. —– 1928.Ibid. no. vi, Loc. cit., vol. xv, pp. 199–220. Washington, H. S. and Merwin, H. E., 1922. Mineralogy. Augite of Halea-kaaka. Maui, Hawaiian Islands. Amer. Journ. Sci., vol. iii, pp. 117–122. Winchell, A. N., 1927. Elements of Optical Mineralogy, Part 2, John Wiley, New York, 2nd Edit., pp. 181–2. —– 1933. Ibid. 3rd Edit., pp. 221–4, 7. —– 1935. Further Studies on the Pyroxene Group. Amer. Miner, vol. xx, pp. 562–8. Wyckoff, R. W. G., Merwin, H. E. and Washington, H. S., 1925. X-ray Diffraction Measurements Upon the Pyroxenes. Amer. Journ. Sci., vol. x, 5th Series, pp. 383–397.