Contributions to the Mineralogy of New Zealand, Part IV

Transactions and Proceedings of the Royal Society of New Zealand, Volume 84, 1956-57, Page 791

Contributions to the Mineralogy of New Zealand, Part IV C. Osborne Hutton [Read before Otago Branch, October 9, 1956; received by the Editor October 11, 1956.] Abstract Rhodonite-spessartite-Rhodochrosite schists from Arrow Valley, Western Otago, are described and complete chemical analyses of the two silicates are presented. For these two minerals the following physical data have been determined; Spessartite ao = 11.62 Å., N = 1.798, S.G.21°C. = 4.15; rhodonite α = 1.723, β = 1.729, γ = 1.737, γ − α = 0.014, 2V = 70 − 73° (+), ρ < ν weak, S.G.21°C. = 3.60. It is believed that these schists are the Chlorite Zone equivalents of manganiferous sediments precipitated by submarine volcanic action. Xenotime and gahnite from Port Pegasus, Stewart Island, placer deposits are reported. For xenotime α = 1.720, γ = 1.820, γ − α = 0.100, S.G21°C. = 4.68. X-ray diffraction powder patterns of Stewart Island and Banka, Indonesia, xenotime are compared, and unit-cell dimensions for the mineral from the former locality are found to be ao = 6.895 Å., ± 0.005 Å., c0 = 6.036 Å., ± 0.005 Å. Gahnite has N = 1.793 − 1.797, S. G22°C. = 4.46, a0 = 8.104 A., ± 0.003 Å. For comparative purposes the cell-edges of gahmte, Japan, and gahnospmel, Ceylon, have also been measured; these are 8.097 Å., ± 0.004 Å. and 8.086 Å., and 0.002 Å. respectively. Occurrence and Petrography of Manganiferous Schists Rhodonite-spessartite-rhodochrosite rocks are found as narrow lenses about one mile south-east from the junction of Arrow River and Billy Goat Creek* Kawarau Survey District, Western Otago, New Zealand; see Hutton (1940) map No. 3. in outcrops, not exceeding 3ft in length by 6ins in thickness, that are intimately associated with coarsely crystalloblastic albite-epidote-chlorite (with or without actinolite) schists, piedmontite-quartz schists, and Chlorite Zone (Chl. 4 sub-zone) equivalents of quartz-plagioclase sandstones of greywacke type. These highly manganiferous rocks do not exhibit the same degree of schistosity and foliation as do the adjoining schists; in fact, the latter development is almost lacking in some instances. In view of the paucity of optical and chemical data for garnets from Chlorite Zone schists, a study of this mineral in particular was thought to be desirable, and at the same time a closer investigation of rhodonite from this environment was undertaken. Xenoblastic particles of rhodochrosite (N0 = 1.803) with highly crenulated borders, colourless in thin section but bright pink in the crushed powder, constitute about 70 per cent. of the specimens examined; grain size ranges from 0.05–1.5 mm, although 0.15–0.2 mm would include the majority of the carbonate crystals. Occasionally rhombohedral outline is evident and polysynthetic twin lamellae are seen rarely. Strings and clouds of minute particles, both transparent and opaque, are closely associated with the carbonate, and their solubility in SO2 solution suggests the possibility that some of these may be manganese oxides, but this has not been confirmed in every case. Rhodonite is intimately associated with carbonate in irregular grains, 0.3–0.4 mm in diameter, although in segregations crystals may be 2 to 3 times as large. Rhodonite is often sieved with rhodochrosite and in a number of instances very irregular or wisp-like relicts of silicate are found in pools of carbonate. Somewhat rounded crystals of spessartite, colourless in thin slice but pale yellow in pure powder, quartz, and manganian apatite complete the mineral assemblage. Garnet shows marked alteration to carbonate, rhodonite to a less extent, and such transformations may have been effected by CO2-bearing waters devoid of free

oxygen. This would result in liberation of some silica, but since very minor quartz is found in these rocks it would seem more likely that SiO2 has been removed in solution. Very thin veinlets of rhodochrosite ramify through the rocks and intersect crystals of rhodonite. Surfaces of these rocks exposed to weathering are covered with thin black films–3 mm thick–of wad-like material that is amorphous, since it does not cause diffraction of X-rays, and in which the manganese is dominantly in the quadrivalent state. Observation of sections cut normal to the manganese oxide surface films shows that complete solution and removal of rhodochrosite has occurred from within a zone 4–5 mm thick, whereas the silicates have been less affected. Oxidation of manganous solutions so produced has been rapid and resulted in almost immediate deposition of amorphous manganese oxides (Fig. 1). Fig. 1.–Crystalloblastic aggregate of rhodochrosite, rhodonite, spessartite, and manganapatite in schist from outcrop located about 1 mile south-east from junction of Arrow River and Billy Goat Creek, Kawarau Survey District, Western Otago, New Zealand. Note also the zone of alteration in which rhodochrosite has been completely, and rhodonite very slightly, replaced by wad. Magnification × 20. Mineralogy of Spessartite The schist was reduced to particle sizes of approximately 10–15 microns, and, after treatment in SO2 solution to remove traces of manganese oxide stains, pure spessartite was obtained by centrifuging the rock powder in Clerici solution. When the pure monomineralic particles of these dimensions are examined in liquid of similar refractive index occasional instances of fine zoning are apparent; attempts to fractionate these zones for analysis were unsuccessful because of the very slight difference of specific gravity involved and because of the very small quantities at either end of the range. The analysis of the garnet is listed in Table I, and a comparison with analyses of spessartites from other localities shows that Crown Range garnet is almost pure spessartite with a modal composition as follows: spessartite 89.77 per cent, almandine 4.14, pyrope 1.02, grossularite 5.07. Physical properties determined for analysed spessartite together with calculated* Data used here are those determined by Skinner (1956) for synthetic pure end-members. values are: a0 = 11.62 Å., ± 0.005 Å. (calc. a0 = 11.62 Å.); refractive index,

Table I.–Analysis of Spessartite 1 2 3 4 5 6 7 SiO2 35.87 35.76 36.06 35.04 36.52 33.36 35.10 Al2O3 20.84 21.06 20.38 21.63 21.20 14.99 19.23 TiO2 trace 0.10 0.28 trace 0.60 nil nil Fe2O3 0.06 1.78 0.54 nil 0.87 0.83 2.07 FeO 1.78 - 0.69 1.86 4.62 1.71 1.67 MnO 38.24 39.40 40.88 39.83 31.52 43.10 40.28 CaO 2.72 1.23 1.88 1.78 3.76 1.10 0.60 MgO 0.22 0.46 0.21 0.26 0.38 nil 0.02 H2O- 0.15 - - - - 0.10 - 99.88 99.79 100.92 100.40 99.47 100.08 99.75 1. Crown Range, Western Otago, New Zealand. Analyst: C. Osborne Hutton. 2. Tsilaisina, Madagascar (Menzer, 1929, analysis on p. 374). 3. Tsilaisina, Madagascar (Menzer, 1929, analysis on p. 374). Analyst: M. Bendig. 4. Bald Knob, North Carolina (Ross and Kerr, 1932, analysis on p. 16). Analyst: E. V. Shannon. 5. Kinko Mine, South-west Honshu, Japan (Lee, 1955, analysis on p. 16). Analyst: W. H. Herdsman. 6. Wodgina, Western Australia (Mason and Berggren, 1942, analysis on p. 415). Analyst: T. Berggren. Analysis includes: P2O5 4.10, Li2O 0.10, Na2O 0.21, K2O nil, H2O + 0.48. 7. Wodgma, Western Australia (Mason and Berggren, 1942, analysis on p. 415). Analyst: T. Berggren. Analysis includes P2O5 0.74, Li2O 0.04. N = 1.798 ± 0.001 for most particles, with a range of 1.797–1.801 (calc. N = 1.796); S.G. at 21°C = 4.15 with a range of 4.12–4.16 for all of the garnet in the schist (calc. S.G. = 4.158). When these physical data are plotted on Stockwell's (1927) curves, the Crown Range spessartite is found to fall very close to theoretical spessartite but just outside the field drawn to include the non-calcium group in the figure that demonstrates relationship between cell edge and specific gravity (Stockwell, 1927, p. 338); for Stockwell's S.G./R.I. and R.I./cell edge diagrams the garnet described here lies close to spessartite and within the pyralspite field. Empirical unit-cell contents for spessartite are set out in Table II. It will be noted that, as is common for members of the garnet group, only a very small proportion of aluminium occurs in the tetrahedrally co-ordinated networks. The octahedral group of cations is slightly deficient in spite of the fact that the very small number of Mg2+ ions has been grouped with Fe3+ and Al3+ rather than with ions of larger radius, although this departure from the required figure of 16 is quite minor. The size of the Mg2+ ion with a radius of 0.65 Å. lies midway between that of Al3+ (0.50 Å.) and Mn2+ (0.80 Å.), yet more closely to that of Fe3+ (0.60 Å.). Accordingly, and provided neutrality is observed, it would seem more appropriate to group Mg with Fe3+ than with the (Ca, Mn, Fe2+) group. Mineralogy of Rhodonite A pure fraction of rhodonite from the Western Otago schists has been analysed, and inspection of Table III will show that it has an exceedingly low Fe2+ and high Mn2+ content as compared with most rhodonites, although it is not as low in iron as rhodonite from Val d'Err (Table III, anal. 4), nor does it contain as high a percentage of manganese as that found in one of two specimens from Vittinge, Finland (Table III, anal. 5).

Table II.–Empirical Unit-Cell Contents of Crown Range Spessartite Weight Per Cent. Cations Basis of 12 Oxygens. SiO2 35.87 Si 23.48 24.00 3.00 Al2O3 20.84 Al 16.08 0.52 Fe2O3 0.06 Fe3+ .03 15.56 1.97 MgO 0.22 Mg .21 15.80 FeO 1.78 Fe2+ .97 MnO 38.24 Mn 21.20 24.07 3.00 CaO 2.72 Ca 1.90 a03 = 1568.97 Å.3. Wt. of cell = 6511.225 × 10-24 gms. F = W/1.6603 × 99.73 = 39.32 Table III.–Analyses of Rhodonite. 1 2 3 4 5 SiO2 46.42 46.58 46.57 46.70 46.13 Al2O3 0.07 0.25 0.73 nil. nil. TiO2 tr.? nil nil - nt. dt Fe2O3 0.11 tr. nt dt -nil FeO 1.49 1.66 3.70 0.35 0.15 MnO 47.62 44.89 46.28 43.81 51.55 MgO 0.92 1.52 0.12 0.30 0.61 CaO 3.26 4.46 1.60 7.80 1.31 H2O 0.18*H2O- 0.58† Listed as total water. 0.72‡ H2O+. 0.09* 0.23† 100.07 100.15 99.72 100.04 100.05 1. Crown Range, Western Otago, New Zealand. Analyst: C. Osborne Hutton. 2. X Mine, Central Honshu, Japan. Analyst: W. H. Herdsman. Analysis includes ZnO 0.21, K2O + Na2O nil (Lee, 1955, p. 21, Table 6). 3. Vittinge, Finland. Analyst: N. Sahlbom (Saxén, 1925, p. 15). 4. Val d'Err, Switzerland. Analyst: J. Jakob. Analysis also includes: H2O + 0.77, P2O5 0.22; V2O5, As2O5 nil (Jakob, 1923, p. 236). 5. Vittinge, Finland. Analyst: A. Bygdén. Analysis also includes: BaO 0.07 (Sundus, 1931, p. 517, Table I, anal. 1). The cell contents, calculated on the basis of six oxygen atoms per unit cell, are set out in Table IV, together with the modal composition of the silicate. Table IV.–Unit Cell Contents of Rhodonite. Weight Per Cent Atoms. Per Cent. SiO2 46.42 1.997 Al2O3 0.07 .003 2.00 MnSiO3 86.79 Fe2O3 0.11 .003 MgSiO3 2.95 FeO 1.49 .054 FeSiO3 2.67 MnO 47.62 1.735 2.00 CaSiO3 7.51 MgO 0.92 .059 CaO 3.26 .150

The physical properties are as follows: α = 1.723 ± 0.002; β = 1.729; γ = 1.737; γ-α = 0.014;2V = 70–73° (+); ρ < ν, very weak to absent in most instances; S.G.21°C = 3.60 ± 0.01. Origin of Manganiferous Schists The most notable feature of the rhodonite-rhodochrosite-spessartite lenses and quartz-piedmontite schists is their constant association with Chlorite Zone equivalents of basic lavas and tuffs. It was this association that led Turner (1946) to suggest in conformity with views expressed by Taliaferro and Hudson (1943) and others, that the quartz-rich schists and accompanying manganiferous and ferriferous sediments, of this area at least, are the metamorphosed equivalents of silica gels contaminated by iron and manganese that had been precipitated in large amount by submarine solfataric and fumarolic action connected with the volcanism that gave rise to associated tuffs and lavas. This view is undoubtedly the explanation for the association found in Western Otago, and the discovery of richly manganiferous lenses of rhodonite and rhodochrosite as an additional member of the association adds considerable support to the suggested mode of origin. The association described here is highly manganiferous but at the same time it is especially poor in iron. This is at first surprising, since discharge of large amounts of ferruginous compounds is also evident during eruption of submarine solfatara–Santorin is only one notable instance of this. However, metamorphosed equivalents of more ferruginous* Magnetite-spessartite-actinolite schists, magnetite-spessartite-piedmontite schists, stilpnomelane-spessartite-almandme schists, ankerite schists, etc. material, although not evident in the immediate vicinity of the rocks described herein do occur elsewhere in this general region in association with metamorphosed tuffs and lavas. Notable here are outcrops on the lower Ben Lomond Ridge, the western slopes of Mt. Soho, headwaters of the Templeburn and Palnoonburn.† See Maps 2 and 3, Hutton, 1940. Xenotime and Gahnite, Stewart Island Both xenotime and gahnite have been found in sands and gravels along the courses of Pegasus and Smith's Creeks in the Port Pegasus area of southern Stewart Island, and these minerals have apparently been released by weathering and degradation of the mineral lodes and greisen zones of the Tin Range. This is the first record of xenotime from Stewart Island and the second time that this phosphate has been found in New Zealand (Hutton, 1950, pp. 685–686). Gahnite has been previously recognized and recorded (vide Morgan, 1927, p. 43), but more recently in some detail Williams (1933, 1934A, 1934B) has studied the mineral in the course of his investigation of the petrology and ore-deposits of the Tin Range and surrounding areas. Mineralogy of Xenotime Xenotime makes up about 7 per cent of four panned concentrates from Mud-town, Port Pegasus District, Stewart Island, and it occurs as very pale yellow to almost colourless, sharply euhedral crystals that are distinctly paler in colour than monazite, yet lacking the greenish tinge of epidote, both of which are associated with the yttrium phosphate. The following forms and habits were recognized: (1) Simple doubly terminated tetragonal pyramids. (2) Every gradation between crystals with {011} as the sole form, through those in which {011} is dominant and {010} less well developed, to short prismatic crystals with {010} dominant. (3) {110} and {010} about equally developed with {011}, or {111} and {011} terminations; these are quite common.

(4) Crystals and fragments with twinning on {111} represent about 2 per cent of all crystals examined. For particle sizes of 0.125–0.25 mm xenotime is attracted at 0.25–0.33 amps, when the Frantz separator is adjusted with a slope of 15° and a tilt of 8°, and the material used for analysis and determination of physical properties was attracted between 0.28–0.30 amps. The magnetic susceptibilities of the chief constituents of the Pegasus concentrates for particle sizes of 0.125–0.25 amps are as follows:–Attracted at 0.1 amps: ilmenite, martitized magnetite; 0.2 amps: ilmenite, minor spessartite-almandine; 0.25 amps: wolframite (dominant), spessartite-almandine, xenotime (minor); 0.30 amps: xenotime, wolframite, gahnite, epidote, hornblende, staurolite (pale brown, irregular fragments with α = 1.740, β = 1.746, γ = 1.753, γ - α = 0.013, 2V = 88° (+), Sp.Gr. at 20°C. = 3.71); 0.50 amps: monazite, gahnite; 1.15 amps: rutile, monazite, zircon; rejected at 1.15 amps: zircon, rutile, cassiterite, sphene, apatite. The physical properties that have been determined for Pegasus xenotime are as follows: α = 1.720 ± 0.002, γ = 1.820, γ - α = 0.100, uniaxial, positive, colourless when particle sizes are approximately 0.03 mm thick, S.G. at 21° C. = 4.68. Measurement of rotation and Weissenberg photographs, the latter calibrated with quartz, leads to the following unit cell dimensions: α0 = 6.895 A., ± 0.005 Å., c0 = 6.036 Å., ± 0.005 Å., X-ray diffraction data from powder films for Banka xenotime and the analysed mineral from Stewart Island are set out in Table V, and d-spacings calculated from unit cell dimensions are listed for the latter occurrence. The lines on the powder films are quite sharp up to about 90° 2 θ, but in the back-reflection region diffuseness is evident. Having in mind the results that Karkhanavala and Shankar (1954, Fig. 1) obtained for monazite that appears to exhibit a certain degree of metamictization, two samples of Stewart Island xenotime were heated for 6.½ hours at 1180° C., one of them in air and the other in a high-silica capillary that held an atmosphere of nitrogen at about 5 mm pressure. In neither case was there any indication of any sharpening of those lines where slight diffuseness existed in the film obtained from unheated material, and it would seem probable, therefore, that the slight diffusivity in high-angle lines is not due to any degree of metamictization. Material was carefully prepared for analysis by electromagnetic fractionation, and a small quantity of contaminating minerals removed by tedious and time-consuming hand-picking rather than by fractionation in hot Clerici solution of high density, since preliminary experiments showed that white films developed on xenotime particles after one hour in Clerici solution at 60° C., and spectrographic investigation of such particles showed that thallium was present. The analysis is listed in Table VI, column 1, and calculations of empirical unit-cell contents show an excess of rare earths, etc., over (P, Si, Al) when all of the aluminium present is grouped, as is silicon, with the (PO4) tetrahedral groups. It is not clear what part Fe3+ plays in this case, and, although it is grouped with the cations of large radii, a slightly better balance of ions is obtained if Fe3+ is shared between rare earth and phosphorus positions. The composition of Stewart Island xenotime is comparable in many ways to that of xenotime from Hvalö and from Narestö, Norway (Blomstrand, 1887, pp. 185–187), especially in the low cerium content and totals of yttrium plus erbium, but differs in that the New Zealand phosphate contains much lower quantities of thorium and uranium. The condition of water in the Norwegian minerals is not stated, but it is an unusually high figure in each instance.

Table V.–X-Ray Diffraction Data for Xenotime. Camera diameter 114.59 mm; Cu radiation with CuKα = 1.5418 Å. and CuKaI = 1.54050 Å. Mudtown, Pegasus District. Bankatinwinning, Indonesia hkl d. meas. Å. d. calc. Å. I d. meas. Å. d. calc. Å. I 101 4.53 4.5 4 4.51 4.54 4 200 3.45 3.44 10 3.43 3.44 10 211 2.745 2.745 3 2.739 2.740 3 112 2.555 2.564 6 2.557 2.560 6 220 2.427 2.435 5 2.429 2.437 6 202 2.262 2.267 2 2.269 2.270 2 301 2.144 2.144 6 2.146 2.149 7 103 1.927 1.930 3 1.928 1.932 3 321 1.818 1.820 5 1.822 1.823 6 312 1.763 1.765 8 1.765 1.767 8 400 1.720 1.722 7 1.722 1.724 7 213 1.680 1.683 1 1.682 1.685 1 411 1.608 1.609 2 1.611 1.611 2 420 1.540 1.540 5 1.541 1.542 5 004 1.508 1.508 < 1 1.509 1.510 < 1 332 1.428 1.430 5 1.429 1.431 5 204 1.378 1.381 2 1.382 1.382 3 501, 431 1.343 1.343 4 1.344 1.344 4 224 1.282 1.282 4 1.282 1.283 4 251 1.251 1.251 < 1 1.251 1.252 < 1 512 1.233 1.232 5 1.234 1.234 5 440 1.217 1.217 2 1.220 1.219 1 600 1.148 1.148 4 1.151 1.150 3 404 1.135 1.134 3 1.138 1.135 2 215 1.122 1.123 < 1 1.124 1.124 < 1 611 1.113 1.113 < 1* Diffuse lines. 1.115 1.114 < 1* 532 1.100 1.099 4* 1.102 1.101 4* 620 1.090 1.089 4* 1.090 1.090 4* 424 1.078 1.0775 4* 1.079 1.080 4* 631 1.013 1.012 2* 1.013 1.013 1* 116 .988 .984 2* .986 .9855 2* 640 .956 .954 4* .957 .956 2* 543 .948 .950 .949 2* .949 1* 444 .947 .949 552 .928 .927 5* .930 .930 4* 604 .913 .915 .914 3* .914 2* 316 .912 .914 703 .883 .884 .883 3* .885 2* 624 882 .883 732 .867 .865 4* .868 .866 2* Intensities (I) were measured visually.

Table VI.–Analyses of Xenotime 1 2 3 Wt. % Atomic Ratios. Empirical Unit Cell Contents. Wt. % Wt. % SiO2 1.67 .0278 .225 1.77 2.36 TiO2 tr. - - - - Al2O3 0.26 .0050 .040 0.36 0.28 Fe2O3 1.33 .0186 .150 1.88 2.01 Y2O3 44.29 .3922 3.193 38.91 30.23 Er2O3 16.63 .0868 .706 17.47 24.34 Ce2O3 0.57 .0034 .027 1.22 0.96 ThO2 0.87 .0032 .026 3.33 2.43 UO2 0.11 .0004 .003 - 3.48 (UO3) CaO 0.61 .0010 .008 0.34 1.09 MnO tr. - - 0.13 - MgO 0.32 .0079 .064 - 0.26 P2O5 32.68 .4604 3.734 32.45 29.23 H2O+ 0.13 - - 1.03 1.77 H2O− 0.09 - - 1.03 1.77 SO3 tr.? - - - - 99.56 100.05 100.31 W = 1342.961 × 10-24 gm; F = 8.14; Σ P, Si, Al = 3.99; Σ Fe3+, Y, Er, Ce, Th, etc. = 4.17. 1. Pegasus District, Stewart Island. Analysts: Rare earths, U, and Th by Drs. Pajakoff and Getoff, Vienna; remainder by C. Osborne Hutton. 2. Hvalö, Norway. Analysis also includes PbO 0.21, ZrO3 0.76, SnO2 0.19 (Blomstrand, 1887, p. 185). 3. Narestö, near Arendal. Analysis includes ZrO3 1.11, SnO2 0.08, PbO 0.68 (Blomstrand, 1887, pp. 186–187). N.B.–Björlykke (1939, p. 71) in citing Blomstrand's analyses 2 and 3 above gives incorrect values for PbO in each case and lists PbO 0.21% in a third analysis of xenotime from Hidra, when the original reference lists none; accordingly his summations are incorrect. If the figures for PbO in Björlykke's table of analyses of xenotime are transposed one column to the right the inaccuracies disappear. Mineralogy of Gahnite Zincian spinel, which makes up about 15 per cent. of the concentrate from which it was separated, exhibits a wide range of forms and habits, from sharp euhedra on the one extreme to highly irregular fragments with marked conchoidal fracture surfaces on the other. Euhedra include the following: (a) Simple octahedra that are occasionally flattened parallel to {111} so that octahedron faces are alternately dominant or nearly suppressed. (b) Rhombic dodecahedra are never dominant but every gradation occurs from slight bevelling of edges of octahedra to those instances where {110} and {111} are equally developed. (c) About 5 per cent. of the crystals were observed to be twinned, usually simply, but occasionally repeatedly. (d) Rare instances were found with {111} slightly modified by {113}. A distinct range of colour is evident, but dark green with a distinct bluish hue is typical. Gahnite has the following physical properties: Refractive index: 1.793–1.797, but 1.795–1.797 was found for the bulk of the material. The lower values were found, in general, for the paler coloured particles,

whereas dark green and dark greenish-brown varieties have values at the higher end of the range. The calculated value is 1.797. Specific gravity: For eight fragments in the fraction concentrated for analysis a value of 4.46 ± 0.02 at 22° C. was found (calc. 4.46). Cell edge: 8.104 Å. ± 0.003 Å. (calc. 8.096 Å.). The calculated physical properties are based on the following data* Winchell (1951, p. 82) incorrectly infers that W. Hugill (Iron and Steel Inst., London, Special Paper, 26, 201–204, 1939) lists the specific gravity of hercynite as 1.83.: Spinel 8.080 Å., 3.55, 1.719; hercynite 8.119 Å., 4.39, 1.800; gahnite 8.0848 Å., 4.62, 1.805, for cell edge, specific gravity, and refractive index respectively. The unit-cell dimensions for spinel and gahnite are the values recently recorded for these compounds by Swanson and Fuyat (1953, pp. 38, 41). In this particular connection it is perhaps surprising that Kordes and Becker (1949) should find that artificially prepared mix-crystals of MgAl2O4 and ZnAl2O4 exhibit unit cells of constant size no matter what the composition may be, whereas on the other hand density and refractive index show linear relationships. During concentration of gahnite by electromagnetic fractionation a distinct range of susceptibilities was evident. With the Frantz separator set with a slope of 15° and a tilt of 10° a few very dark green and brownish-green particles were attracted at 0.325 amps, and the least susceptible pale green to dark bluish-green varieties at 0.375 amps. However, 85 per cent of the total gahnite fraction was concentrated at 0.35–.36 amps. The analysis was made with material very carefully prepared by electromagnetic fractionation followed by flotation in thallous malonate and formate melts at 60–70° C. No indications of any reaction between melt and spinel was evident, although white films developed on the few particles of xenotime and monazite that contaminated the magnetically fractionated material; the phosphates were discarded and were not employed in any aspect of this work. In the course of the analytical work the writer was unable to obtain any significant solution of the finely powdered gahnite in HF and concentrated H2SO4. As a result iron was determined as Fe2O3 and this entire quantity was recalculated as FeO in the belief that it would be chiefly present in that oxidation state; a satisfactory summation of the weight per cent of oxides would seem to support this assumption. Table VII–Analyses of Gahnite A B C SiO2 0.03 - 1.08 Al2O3 58.02 58.60 57.72 FeO 14.53 14.30 8.88 Theoretical composition: ZnO 25.37 22.80 29.46 Per cent. MgO 1.52 3.96 2.79 MgAl2O4 6.8 CaO nil - - FeAl2O4 36.7 H2O– 0.23 - - ZnAl2O4 56.5 TiO2 tr - - MnO tr - - 99.70 100.97 99.95 A: Ferroan gahnite, Mudtown, Pegasus District, Stewart Island, New Zealand. Analyst: C. Osborne Hutton. B: Migiandone; analysis No. 2, Doelter (1926, p. 527). Total includes Fe2O3 1.31. C: Goyamin Pool, S. W. Division, W. Australia, Simpson (1937). In the analysis reported in Table VII it is interesting to note that the percentage of silica is extremely low, although this is not always so with members of the spinel family of minerals (vide Rankin and Merwin, 1918, p. 307; Tilley, 1938, p. 83; Pehrman, 1948, p. 330, etc.). If the analysed spinel is quite free from extraneous material it would seem most likely that the silicon present enters the structure during

crystallization and not at some post-crystallization period, in view of Fréchette and Andrews' (1944, p. 200) study, which demonstrated an absence of solid-state reaction between spinel and forsterite at 1400° C. Furthermore, the silicon ions, on account of their size, would presumably be located in the A sites with 4-co-ordinated ions, that is with Mg2+, Zn2+, and Fe2+, and not with Al3+ in the B sites, since the aluminates, except NiAl2O4, appear to be normal and not inverse types of spinel (Bacon, 1952; Bacon and Roberts, 1953; Gorter, 1954, pp. 10–11). The composition of the Stewart Island gahnite has a medium content of iron, and only two similar analyses were found in a search through the literature (Table 7, analyses B and C); zincian spinels with total iron content in excess of the figure found for the Stewart Island mineral are, however, not rare (cf. analysis No. 15, Palache, Berman, and Frondel, 1946, p. 691). Empirical unit-cell contents of Stewart Island gahnite are set out in Table VIII, and it should be noted that the numbers of aluminium ions are in slight excess of the required figure, whereas the (Mg, Fe, Zn) group of ions is negligibly deficient. Although the balance between the two groups is not as satisfactory as it might be, at least the assumption that iron is present in the lower state of oxidation receives additional support here. Table VIII.–Empirical Unit Cell Contents of Gahnite, Stewart Island, New Zealand Wt. % No. of ions Theoretical ratio. Al2O3 58.02 16.36 16 FeO 14.53 2.90 ZnO 25.37 4.48 7.92 8 MgO 1.52 0.54 Cell weight (W) = 2373.739 × 10-24 gm. Factor (W / 1.6602 S) = 14.378. Formula: 4 [(Mg, Fe, Zn)0.99 Al2.04O4]. The measured and indexed X-ray powder pattern of Stewart Island gahnite is set out in Table IX, and these data are compared there with the d-spacings calculated from rotation and Weissenberg films; the latter were standardized with quartz. Relative intensities have been determined visually. If the specific gravity and refractive index data found for Stewart Island gahnite are plotted on Fiumano's (1947, p. 58) diagram both points are found to lie much above the curves proposed to illustrate relationship between ZnO and these physical properties, even when the incorrect labelling along the ordinate of Fiumano's diagram is recognized and taken into account. This is also the case with gahnite from Finland (Pehrman, 1948, p. 330) and Japan (Omori and Hasegawa, 1953, p. 149). It would seem clear then that the relationships refractive index/ ZnO and specific gravity/ ZnO are not simple linear ones, and much more must necessarily be taken into account than the question of the percentage of ZnO in spinel. When these data are plotted on the ternary diagram MgAl2O4-ZnAl2O4-FeAl2O4 (Fig. 2), reasonably good coincidence is achieved between experimentally determined and graphically estimated values for refractive index and specific gravity. Determinations of unit cell dimensions of analysed zincian spinels are conspicuously lacking in the literature, and accordingly the writer has measured the cell edges of analysed gahnite from Japan (Omori and Hasegawa, 1953), and Ceylon (Anderson and Payne, 1937), in addition to the spinel from Stewart Island, New Zealand. These are as follows: 8.097 Å. ± 0.004 Å (Japan), 8.086 Å. ± 0.002 Å. (Ceylon), and 8.104 Å. ± 0.003 Å. (New Zealand). If the a0 values

Table IX-X-ray Diffraction Data Yielded by Stewart Island Gahnite CuKα = 1.5418 Å., CuKα1 = 1.5405 Å. Camera diameter 114.59 mm. d. meas. Å d. calc Å I hkl. 4.65 4.68 <1 111 2.863 2.866 9 220 2.442 2.444 10 311 2.326 2.340 <1 222 2.021 2.026 2 400 1.856 1.860 2 331 1.649 1.654 3 422 1.555 1.560 4 511, 333 1.431 1.433 5 440 1.369 1.369 <1 531 1.280 1.281 2 620 1.235 1.236 3 533 1.2205 1.221 < 1 622 1.169 1.170 1 444 1.134 1.135 1 711, 551 1.0825 1.0825 2 642 1.055 1.055 3 731, 553 1.0125 1.013 2 800 .9885 .987 <1 733 .955 .954 2 822, 660 .936 .936 3 751, 555 .929 .930 <1 662 .905 .906 <1 840 .8895 .890 <1 911, 753 .864 .8635 1 664 .850 .850 2 931 .827 .827 3 844 .795 .795 2* Diffuse lines. 10.2.0, 862 .783 .783 2* 951, 773 for pure spinel, hercynite, and gahnite quoted earlier are employed in the ternary diagram, and the zincian spinels of Japan, Ceylon, and New Zealand plotted thereon, they would appear to have unit cell dimensions smaller than the values actually found experimentally–0.007 Å. smaller for the first two minerals and 0.004 Å. for the latter. Although these data are too few to permit any accurate estimation of the situation, they suggest that the value at present accepted for the dimension of the unit cell of hercynite may be too low. Acknowledgments I wish to acknowledge the opportunities provided for this research by the award to me of a John Simon Guggenheim Memorial Foundation Fellowship during 1953–54. Professor K Omori, of Tohuku University, Japan, very kindly allowed me to have a specimen of analysed gahnite, Professor W. P. de Roever, of the University of Leiden, Holland, made available a xenotime concentrate from Banka, Indonesia, and Dr. G. F. Claringbull, of the British Museum (Natural History), through the good offices of Mr. B. W. Anderson, provided me with fragments of gahnospinel from Ceylon. To all of these gentlemen I wish to express my sincere appreciation for their help.

Fig. 2.–Ternary diagram spinel-gahnite-hercynite illustrating relationships between R.I., S.G., cell edge, and chemical composition. Key to locations of gahnite employed: A = Japan, B = Ceylon, C = Stewart Island, New Zealand. References Anderson, B. W., and Payne, C. J., 1937. Magnesium-zinc-spinels from Ceylon, Mineral. Mag., 24, 158, 547–554. Bacon, G. E., 1952. A neutron-diffraction study of magnesium aluminium oxide Acta. Cryst., 5, 5, 684–686. —— and Roberts, F. E., 1953. Neutron diffraction studies of magnesium ferrite-aluminate powders, Acta. Crysta., 6, 1, 57–62. Björlykke, H., 1939. Feldspat V. De sjeldne mineraler pa de Norske granittiske pegmatittganger, Norges Geol. Undersök. 154, 5–78. Blomstrand, C. W., 1887. Analys af cer- och ytter-fosfater fran Södra Norge, ett bidrag till fragan om dessa mineraliers kemiska byggnad, Geol. Förens i Stockholm Forh. 9, 3, 160–187. Doelter, C., and Leitmeier, H., 1926. Handbuch der Mineralchemie, 3, 2, esp. 527–532, Dresden und Leipzig: Theodor Steinkopff. Fiumano, E., 1947. Lo spinello zincifero di Tiriolo (Catanzaro), Messina Univ. Ist. di Min. Notiz di Min., 1, 55–63. Frechette, V. D., and Andrews, A. I., 1944. Investigation of reactions of simple magnesia spinels with alkaline earth orthosilicates in the solid state, Jour. Amer. Ceram. Soc., 27, 7, 197–202. Gorter, E. W., 1954. Saturation, magnetization, and crystal chemistry of ferrimagnetic oxides, Univ. of Leyden thesis, 1–111. Hutton, C. O., 1940. Metamorphism in the Lake Wakatipu Region, Western Otago, New Zealand, N.Z. Dept. Sci. Ind. Res. Geol. Mem., No. 5, 1–90. ——, 1950. Studies of heavy detrital minerals, Geol. Soc. Amer., Bull. 61, 635–716. Jakob, J., 1923. Vier Mangansilikate aus dem Val d'Err (Kt. Graubinden), Schweiz. Minü. Petr. Mitt., 3, 227–237. Karkhanavala, M. D., and Shankar, J., 1954. An X-ray study of natural monazite; I, Proc. Indian Acad. Sciences, 40, 67–71.

Kordes, E., and Becker, H., 1949. Spinellmischkristalle des Systems MgAl2O4–ZnAl2O4, Zeits. Anorg. Chem., 258, 227–237. Lee, D. E., 1955. Mineralogy of Some Japanese Manganese Ores, Stanford University Publications, University Series, Geol. Sciences, 5, 1–64. Mason, B., and Berggren, T., 1942. A phosphate-bearing spessartite garnet from Wodgina, Western Australia, Geol. Förens. Stockholm Forhand. (for 1941), 63, 1, 413–418. Menzer, G., 1929. Die Kristallstruktur der Granate, Zeits. Krist, 69, 3–4, 300–396. Morgan, P. G., 1927. Minerals and mineral substances of New Zealand, N.Z. Geol. Survey, Bull. 32 (N.S.), pp. 1–110. Omori, K., and Hasegawa, S., 1953. The first occurrence of gahnite in Japan, Science Repts. Tohoku Univ., 3rd ser., 4, 3, 147–150. Palache, C., Berman, H., and Frondel, C., 1946. The System of Mineralogy, 7th ed., New York: John Wiley and Sons, Inc. Pehrman, G., 1948. Gahnit von Rosendal auf Kimito, S. W. Finnland, Geol. Inst. Upsala, Bull. 32 329–336. Rankin, G. A., and Merwin, H. E., 1918. The ternary system MgO–Al2O3–SiO2, Amer. Jour. Sci., 45, 268, 301–325. Ross. C. S., and Kerr, P. F., 1932. The manganese minerals of a vein near Bald Knob, North Carolina, Amer. Mineral., 17, 1–18. Saxen, M., 1925. Om mangan-jarnmalmfyrdigheten i Vittinki, Fennia, 45, 3–40. Simpson, E. S., 1937. Contributions to the mineralogy of Western Australia, Series X, Jour. Roy. Soc. W. Austr. (for 1936–37), 3, 17–35. Skinner, B. J., 1956. Physical properties of end-members of the garnet group Amer. Mineral., 41, 5–6, 428–436. Stockwell, C. H., 1927. An X-ray study of the garnet group, Amer. Mineral., 12, 7, 327–344. Sundius, N., 1931. On the triclinic manganiferous pyroxenes, Amer. Mineral., 16, 11, 488–518. Swanson, H. E., and Fuyat, R. K., 1953. Standard X-ray diffraction powder patterns, U. S. Nat. Bur Standards, Circ, 539, 2, 1–63. Taliaferro, N. L., and Hudson, F. S., 1943. Genesis of the manganese deposits of the Coast Ranges of California, California State Div. Mines, Bull. 125, 217–275. Tilley C. E., 1938. Aluminous pyroxenes in metamorphosed limestones, Geol. Mag., 75, 2, 81–86. Turner, F. J., 1946. Origin of piedmontite-bearing quartz-muscovite-schists of north-west Otago. Trans. Roy. Soc. N.Z., 76, 2, 246–249. Williams, G. J., 1933. The tin-tungsten deposits of Stewart Island, New Zealand, Mining Mag. (March), 3–8. ——, 1934A. The auriferous tin placers of Stewart Island, New Zealand, N.Z. Jour. Sci. and Tech. 15, 5, 344–357. ——, 1934B. A. granite-schist contact in Stewart Island, New Zealand, Quart. J. Geol. Soc., 90, 322–353. Winchell, A. N., 1951. The Elements of Optical Mineralogy, Part 4, 4th. ed. New York: John Wiley and Sons, Inc. Prof. C. Osborne Hutton, School of Mineral Sciences, Stanford University, Stanford, California.