Four different palladium and platinum bismuthotellurides, one containing major mercury, have been identified in the ore from Rustenburg mine on the Merensky Reef. The composition of each phase was determined by electronprobe microanalysis, and X-ray powder data are presented for three of these phases. Quantitative spectral reflectivity values at 470, 546, 589, and 650 mμ were obtained on areas of 1–10µ. Similar phases have been located in the ore from Union mine.

Two of these phases are shown to be moncheïte and kotulskite. From two electronprobe microanalyses, kotulskite is shown to be Pd(Te,Bi). The other two phases are new minerals: analysis of one of them gives the composition (Pd,Pt)(Te, Bi)2, containing 23·1% Pd, 1·8% Pt, 50·8 % Te, 14·2 % Bi, with strongest X-ray powder lines at 2·92 (10), 2·10 (6), 2·02 (3), and 1·54Å (3), and cell parameters a 3·978±0·001, c 5·125±0·002; it is probably isostructural with moncheïte, and an isomorphous series may exist between them; the name merenskyite is proposed for this mineral, after Dr. Hans Merensky who was mainly responsible for discovering the platinoid bearing ‘reef’; it is commonly intergrown with kotulskite. For the second new phase (Mineral A), no X-ray data could be obtained, but a single electron-probe microanalysis gave a composition of Pd 27·8 %, Hg 12 %, Te 38·5 % and Bi 1·6 % suggesting the empirical formula (Pd,Hg)(Te,Bi).

The alkali feldspars in the Ardara pluton are microcline- and orthoclase-microperthites. Their compositions have been determined by the 20 X-ray method, both before and after homogenization. The bulk compositions are very rich in potassium as a result of postmagmatic recrystallization of the alkali feldspar. There is no significant variation in either the bulk composition or the composition of the potassic phase in different parts of the intrusion. The structural state of the feldspars has been studied by measurement of 2V and obliquity of the potassic phases, and the reciprocal lattice angles α* and γ* of the potassic and sodic phases in selected crystals. Most of the specimens contain near-maximum microcline, but there is a decrease in the obliquity of the potassic phase at the eastern end of the intrusion. This can be related to the subsequent intrusion of the Main Donegal Granite.

Plagioclase in mafic gneisses from Broken Hill, Australia, contains twins that can be divided into two groups on the basis of shape: firstly, lamellar twins that gradually change thickness across a grain or form lenticular terminations; and secondly, simple and lamellar twins that show angular steps in the twin interface or form abrupt, planar terminations. The former are interpreted as mechanical twins, and the latter as growth twins formed by grain growth in the solid state, arguing largely by analogy with twin shapes in experimentally produced aggregates, particularly metals. The only twin laws observed in lenticular twins are albite and pericline, these being of nearly equal abundance. Stepped twins are much less common than lenticular twins, and in rocks without lenticular twins most of the plagioclase is untwinned. Several twin laws occur among the stepped twins, but the albite law is the most common. Simple twins on the albite and albite-carlsbad laws are prominent.

Sixteen samples of crocidolite asbestos from weathered zones in mines in Cape Province, Transvaal, Western Australia, and Bolivia have been studied by thermal analysis methods and their characteristics are compared with those of fresh crocidolite from Koegas-Westerberg given by Hodgson, Freeman, and Taylor (1965). New chemical analyses of these samples are reported.

The effect of natural weathering on crocidolite tends to a removal of ¼ of the OH groups and an adjustment of the Fe2+/Fe3+ ratio to unity, but there are only minor changes in the physical characteristics of the fibre. The d.t.a. pattern of decomposition is not affected by weathering, but the endothermic peak corresponding to final decomposition is moved to higher temperatures as the Mg content of crocidolite increases. Mg-blocking of this decomposition is considered to be more important than Mg-blocking of the oxidation reaction at 400–450° C, suggested by Addison et al. (1962).

D.t.a. and other evidence shows that the Transvaal crocidolite contains intergrown amosite fibres, and suggests that the Bolivian crocidolite is admixed with anthophyllite.

Natural Al-Fe-Mn epidotes are divided into Al-Fe and Al-Fe-Mn series.

In the Al-Fe series, a, b, and c increase steadily over the range 0 to 0·9 Fe3+ per formula unit, and more slowly at higher Fe contents. Index variation diagrams show the reverse trend, with indices increasing moderately fast over the range 0 to 0·85 Fe, and faster at higher Fe contents.

In the Al-Fe-Mn series, b increases almost linearly over the range of (Fe, Mn) contents, with Mn causing a larger increase than Fe, whilst a and c increase over the range 0 to 0·9 (Fe, Mn), and then decrease. This curious behaviour is accounted for by the application of crystal field theory. The indices increase linearly over the range 0·8 to 1·5 (Fe, Mn), the optic axial angle depending on (Fe, Mn) content and the ratio Mn/(Fe, Mn).

Thermal and mechanical properties (conductivity, thermal expansion, specific heat, compressibility, and hardness), electromagnetic properties (dielectric constants, conductivity, magnetic susceptibility), and thermodynamic data (S) are also summarized, and estimates given where experimental data are not available.

When crocidolite (fibrous riebeckite, Na2Fe3+2Fe2+2·6Mg0·4Si8O22(OH)2 approx.) is heated in argon or nitrogen at 50–400°C, uncombined water is lost. The tensile strength drops sharply at 300–400°C. At 500–600°C under static, or 570–700°C under dynamic conditions, an endothermic dehydroxylation yields an anhydride with a structure close to that of the initial material. At about 800°C, the anhydride decomposes endothermically, giving acmite, cristobalite, a spinel, and liquid. At 950–1000°C the acmite decomposes. Melting is extensive by 1050°C.

On heating crocidolite in oxygen or air, uncombined water is again lost below 400°C. At 300–450°C (static), or 400–600°C (dynamic), hydrogen ions and electrons are lost, to give an oxyamphibole, Na2Fe3+4Fe2+0·6Mg0·4Si8O24. The process is exothermic and probably occurs by proton and electron migration, as suggested by Addison et al. (1962). At 600–950°C the oxyamphibole decomposes endothermically, and most of the remaining Fe2+ is oxidized, giving acmite, hematite, cristobalite, and a spinel. The spinel has largely disappeared by about 950°C, while at 975–1000°C the acmite also decomposes and melting begins.

When crocidolite is heated in hydrogen, no change is detectable by X-rays below 530°C, when the amphibole decomposes to give mainly a pyroxene, metallic iron, cristobalite, and liquid. The same products are detectable up to at least 850°C; the proportion of metallic iron increases with temperature.

With the exception of metallic iron, all the crystalline products formed, whether under neutral, oxidizing, or reducing conditions, show varying degrees of preferred orientation. This could be explained if the reactions occur mainly by the migration of cations through relatively undisturbed oxygen frameworks.

The electron probe was used to analyse for Fe, Ti, Mg, Al, and Mn. Vanadium could not be determined because of interference from Ti. Calcium was shown to be present only in inclusions of apatite and silicate; it does not oceur in the titanomagnetite structure.

Homogeneous grains (magnetite-ulvöspinel solid solutions) do not vary greatly in bulk composition, averaging close to 61% Fe and 8 % TiO2. Inhomogeneous grains contain variable amounts of hematite-ilmenite lamellae produced by progressive oxidation of the homogeneous grains. In the early stages there is rapid build-up of titanium in the rhombohedral lamellae, which are initially relatively poor in iron. As oxidation proceeds the almost titanium-free cubic phase becomes strongly enriched in magnesium, aluminium, and manganese, while the iron: titanium ratio in the rhombohedral phase increases. The end-product may be a homogeneous titanhematite, if initial concentrations of Mg, Al, and Mn are low. Otherwise residuals of a relatively iron-poor almost titanium-free (Mg, Al, Mn)-enriched cubic phase remain in the titanhematite. The behaviour of iron and titanium agrees with recent synthetic work in the system. Some of the oxidized grains contain transparent spinel lamellae exsolved from the cubic phase. Exsolution may commence at different stages in the oxidation process, but the reason is not clear—there are no significant differences in the composition of phases with and without spinel lamellae.

Analyses and normative compositions are presented for the different phases in each grain, where possible. The compositional changes accompanying oxidation satisfactorily explain the slope of the thermomagnetic curves : there is no need to invoke bulk compositional variations within a given ironsand sample.

Samples from an ilmenite and a non-titaniferous magnetite concentrate were also briefly examined.

The third occurrence of stokesite was discovered in a desilicated pegmatite at Věžná, Western Moravia. In this locality, stokesite accompanies columbite, cassiterite, and stibian pyrochlore-microlite, in a miarolitic cavity of albite. The X-ray data and the optical properties of the Věžná stokesite agree well with those observed by previous authors.

Cell dimensions and densities derived from least-squares refinement of X-ray powder diffractometer patterns are: orthoenstatite, a 18·2249±0·0013, b 8·8149±0·0008, c 5·1746±0·0008 Å, V 831·30±0·14 Å3, D 3·208 g/cc ; low clinoenstatite, a 9·6065±0·0011, b 8·8146±0·0007, c 5·1688±0·0006 Å, β = 108·335± 0·009, V 415·46±0·09 Å3, D 3·210 gm/cc. Because orthoenstatite transforms reversibly to clinoenstatite at high pressure and low temperature, and because a statistical test shows that orthoenstatite is less dense than clinoenstatite at the 90 % confidence level (in accord with the direction of slope of the experimental transition curve), it is no longer possible to regard low clinoenstatite as a metastable form of MgSiO3.

The typical hornblendes of metabasic rocks in the Willyama Complex change from bluish-green to deeply coloured brownish varieties with increase in metamorphic grade. Chemical analyses and optical properties of 23 representative hornblendes are interpreted as showing this colour change to reflect steady increase in titanium content of the hornblendes with increase in grade. High-grade hornblendes also differ from those of lower grade in that tetrahedral aluminium in the Z sites is balanced to an increased extent by alkalis entering the A site rather than by octahedral aluminium entering the Y site. Comparison between theoretical and measured densities of Willyama hornblendes suggest that a calculation based on 23 oxygen anion provides closer approximations than a 24 (O, OH, F) calculation to their structural formulae.

A new occurrence of varlamoffite is described, and the status of this species is considered; the term is best employed for cryptocrystalline yellowish material consisting predominantly of Sn and H20 that may yield a diffuse powder pattern comparable to, or closely resembling that of cassiterite, but does yield a powder pattern identical with that of cassiterite after heat-treatment at about 1200° C. A new mineral, malayaite, CaSnSiO5, is briefly described; it is isostructural with sphene.

The thermal disintegration of crocidolite has been investigated with X-ray powder and rotation photographs. Initial disintegration products include monoclinic pyroxene, iron oxide phases, and eristobalite as a pseudomorph of the original fibre bundles but at about 1000° C fusion occurs due to incongruent melting of the pyroxene. The products are strongly oriented along the fibre axis and full orientation relationships have been deduced with the aid of structural models. From these relationships it is suggested that regions of the amphibole structure on (100) are retained in the product phases without change in orientation.

The basic formula of an amphibole is completely defined by stating the number of sodium atoms in X positions and the numbers of aluminium atoms in Y and Z. These three numbers are coordinates in a compositional space that can easily be represented. The alkali amphiboles are re-defined as having at least one atom of sodium in X in the basic formula unit. A systematic nomenclature for the calciferous and alkali amphiboles is suggested using existing end-member names.

The values of 2θ 01 for some homogenized lime-rich fine microperthites from the Kangerdlugssuaq alkaline intrusion, East Greenland, are investigated and found to be consistently high when compared with the values used in the standard curves for estimating the Or content of alkali feldspars. This deviation, which is approximately proportional to anorthite content, results in anomalously low Or percentages being obtained and is ascribed to a distortion of the structure due to the unusually high calcium content of the microperthites.

The temperature at which dehydroxylation occurs in amphiboles is thought to be largely dependent on the type of cation that occupies the M3 and M1 sites. Four amphibole types have been studied in which these sites are occupied, in various proportions, by Mg2+ or Fe2+. The temperature of dehydroxylation, in an inert atmosphere, is shown to rise with increasing Mg2+ content in these sites.

The necessary conditions for optical pleochroism in orthopyroxenes are: ordering of Fe2+ ions, with preferential occupation of M2 positions; presence of small, high-valence cations (Al3+, Fe3+, Ti4+) in M1 positions; replacement of Si by Al; and a sufficient Fe2+ content. These factors favour an intensification of the charge-transfer absorption bands in the ultra-violet that is more marked for the α vibration direction than for β and γ. This leads to absorption in the blue and a red colour in transmitted α-polarized light, while β and γ, remain green.

Uvarovite and its relationships with grossular and andradite have been studied in order to understand better the geochemistry of uvarovite. Synthese have been carried out under both dry and hydrothermal conditions; a solid solution with grossular was obtained under both conditions, but with andradite only under hydrothermal conditions. The results are discussed in relation to the chemistry of uvarovite.

Variation diagrams for the optic properties and densities of natural specimens within the series have been determined. They focus attention on two composition regions at about 45 % Cn and 65 % Cn. The significance of the first of these is uncertain, but it is suggested that the discontinuity at 65 % Cn is related to structural changes that take place at this composition.

X-ray and thermal studies in the system MnO-O-H2O show that the transformations groutite(α-MnOOH)-ramsdellite(MnO2)-pyrolusite(MnO2) are topotactic, with the a, b, and c directions of all three corresponding. The reactions proceed by migration of manganese atoms and, where appropriate, protons. Pure single crystals of ramsdellite could not be prepared by heating groutite, even though the two structures are very similar.