“Image analysis” refers to the general family of computational techniques which are concerned with the extracting of quantitative information from images captured in digital form. Such techniques have been in use for several years in areas such as biology and metallurgy, but have been slow to find wide use in the microscopic areas of Earth Sciences. Notable exceptions have included applications in the areas of mineral processing and sedimentary petrography.

The impetus for the meeting held jointly by the Applied Mineralogy Group of the Mineralogical Society and the Geological Information Group of the Geological Society of London in September 1993, was an attempt to take stock of the spread of digital image analysis techniques into the domain of the mineralogist and petrologist.

Image analysis techniques are used to quantify finite strain in microconglomerates from the ‘Zone Houillère’ (Briançonnais Zone, French Alps) using the normalized Fry method. Two different techniques have been developed to extract the necessary parameters from quartz grains: the first uses an interactive videographic image analyser linked to a digitizer, and the second uses a semi-automatic image analyser algorithm working on numeric images. Comparison between these two techniques allows the data provided by the latter to be validated. Semi-automated image analysis is then employed to compute the characteristics of the finite strain ellipse as defined by the normalized Fry method. This has been tested on natural and simulated fabrics and gives accurate results. Finally, these techniques have been applied to samples from the French Alps, in an attempt to correlate the regional pattern of finite strain with deep seismic reflectors. This paper presents the preliminary results using finite strain data determined by image analysis processing.

Many subjective descriptive terms such as ‘smooth’, ‘coarse’ and ‘anisotropic’ are routinely used to convey properties of textures and fabrics, but these aspects of geological objects are not easily quantifiable. The use of directional semivariograms of image greylevels is assessed for this purpose. These quantify anisotropy of ‘coarseness/fineness’ by using semivariogram ranges, and ‘roughness/smoothness’ by using semivariogram sill values. These can be summarized by ellipses derived from semivariogram ranges and directions. The method will find applications in analysis of rock thin sections as well as analysis of surface texture of grains and microfossils.

Analysis and processing of electronic images generated using confocal scanning laser microscopy (CSLM) provide a new means of counting and determining the angular distribution of fission tracks in crystalline solids. Using image analysis techniques, large numbers of fission tracks in external mica detectors can be counted rapidly, while measurement of aspect ratio and area of a track entry hole at its intersection with a surface can provide a means of assessing the angle of inclination of individual tracks within minerals and plastics. CSLM imaging, combined with image processing, offers a new and powerful way of probing the surfaces and interiors of crystalline materials. The techniques might also prove useful in the study of fluid inclusions and chemical zoning patterns in minerals.

Textural development of the felsic phases in two granodioritic rocks from the zoned Linga superunit of the Peruvian Coastal Batholith has been characterized using serial thin sectioning, image analysis and three-dimensional reconstruction. The study has shown how each mineral phase contributed to the texture during the formation and development of a contiguous crystal framework and during subsequent restriction, isolation and occlusion of the melt-filled porosity. The work highlights the important factors and potential problems in the use of serial thin sections and imaging in the analysis of complex polyphase rock textures.

The application of serial lapping, imaging and image processing, and three-dimensional reconstruction techniques are discussed using the quartz-cemented Penrith Sandstone as an example. The study involves the characterization of the three-dimensional pore system microgeometry using combined back-scattered scanning electron and cathodoluminescence microscopy and focuses on the definition of post-compaction and post-cementation intergranular porosity networks. Sample preparation, section spacing and re-orientation and data presentation are described. Aspects of the application of serial section datasets and their limitations are discussed in relation to permeability studies.

A new procedure is described for locating mineral grains in geological samples prepared for microprobe analysis. This procedure uses a digital image of the entire sample surface to select points of interest. These features of interest must be visible in this external image. Once the sample is mounted on the specimen stage of the microprobe, the digital image is used as a ‘map’ to relocate minerals for analysis. The image can be recorded using either a video camera, fitted with a macro zoom lens, or a flat-bed scanner. A calibration procedure has been developed in which two indexing points are used to calculate the instrument stage coordinates from the pixel coordinates of the digital image. Following this calibration, the stage coordinates of any feature visible in the digital image can be displayed to facilitate rapid relocation. The procedure was evaluated by relocating magnetite grains visible in an optical image of a geological thin section. The repositioning accuracy in relocating 28 magnetite grains distributed over an area of about 20 × 30 mm was found to be 157 ± 102 µm (one standard deviation) from the video camera image and 48 ± 28 µm from a flat bed scanner image. The former image was thought to suffer from some optical distortion. The procedure may be applied to any microprobe instrument fitted with a digital x-y specimen stage. Future applications in relocating features visible in autoradiographs are currently being evaluated.

Early Proterozoic chemical sediments of the Birimian Supergroup in northern Ghana host several types of metamorphosed manganese-bearing rocks. Differences in the mineralogy and geochemistry can be attributed to facies changes in a mixed volcanic-volcaniclastic depositional environment. Manganese oxide-bearing phyllite, which is enriched in transition metals (Cu, Ni, Co, Zn), formed on the flanks of submarine volcanic edifices above an oxidation boundary. Towards the deeper basin, manganese silicate-rich gondites occur. These consist either of spessartine + quartz + ilmenite, or of spessartine + quartz + Mn amphiboles ± rhodonite ± hyalophane ± Mn stilpnomelane ± ilmenite. The Mn amphiboles are identified as manganoan actinolite, tirodite, and dannemorite. Sulphides are widespread as premetamorphic inclusions in Mn garnet grains. In the basin centre, chlorite schist containing garnet with 50–60 mol.% spessartine represents the most distal manganese-bearing rock which is highly diluted by volcanogenic background sedimentation. The origin of Mn-rich rocks is explained by heat-driven seawater convection systems active in submarine volcanic centres that provided hydrothermal solutions. Precipitation of different minerals depended on the geochemical conditions prevailing on the seafloor. During the Eburnean event (around 2000 Ma), the chemical sediments were metamorphosed to upper greenschist or lower amphibolite facies. Mineral assemblages in the gondites point to metamorphic conditions in the range of 450–500°C at 2–3 kbar.

Two types of gold can be distinguished in tetrahedrite from the Veitsch magnesite deposit, Austria. Primary gold present in unfractured massive tetrahedrite, has a grain size up to 18 μm and contains, on average, Cu 8.73, Ag 7.01, Au 78.63 and Hg 2.35 (wt.%). Secondary gold is present in fractures and is directly intergrown with digenite, covellite, Cu-thiosulphate, cuprite and chrysocolla but not with malachite or azurite. The secondary gold is up to 200 µm in size with an average composition of Cu 3.06, Ag 6.82, Au 86.41 and Hg 3.51 (wt.%). It is usually closely intergrown with thiosulphates containing up to 0.21 wt.% Au. This, together with the presence of ‘dirty’ gold with cloudy thiosulphate inclusions, directly indicates the transport and deposition of Au by a thiosulphate ligand. We believe this is the first reported direct evidence of gold transport and deposition by thiosulphate complexes in a natural environment.

Screes on high serpentinite sea cliffs in Shetland have been internally cemented by deposits of brucite, hydromagnesite and aragonite since the last glaciation. These minerals have been deposited from solutions originating as rain (often contaminated with salt spray), which fell on the serpentinite screes and dissolved Mg and minor amounts of Ca from the more finely divided and powdery scree components. As the solutions filtered down through the screes they deposited cements in the form of spherulitic layers of brucite and hydromagnesite with minor amounts of aragonite. At sea level the Mg-rich solutions mingled with sea water impregnating the scree and there cementation within the scree took place by the crystallization of plush-type layers of aragonite on the serpentinite fragments.

A design for a micro furnace and a fusion technique are described by which silicate materials (in particular picritic and peridotitic rock compositions) may be fused and quenched to a homogeneous glass without significant loss of components, for the purposes of bulk analysis. The furnace consists of a Pt wire electrical resistance heater mounted in a furnace assembly which is fitted to a microscope stage. Microscopic examination of the sample during the fusion process allows the sample to be quenched as soon as all crystalline material has been fused and thus minimizes the loss of iron and alkalis due either to over-heating or prolonged fusion time. Analysis of glass beads of a model peridotite composition (analysed independently by X-ray fluorescence analysis) shows that for typical fusion times (5–10 seconds) the bulk composition is preserved at distances of > 25 µm from the margins of the glass beads. Analysis of natural rock powders of known composition shows that the method can be used to analyse whole rock compositions. The furnace is simple to construct and cheap to run and provides a simple and rapid method of producing glass samples for bulk analysis.

Enclaves of diverse origin are present in minor amounts in the coarse-grained biotite granites of the Cornubian batholith, southwest England. The most common enclave type is layered, rich in biotite, cordierite and aluminosilicates, and has textures and compositions that reveal variable degrees of melt extraction from metasedimentary source rocks. Rare sillimanite-bearing enclaves represent residual material, either from the region of magma generation or its ascent path, but most such enclaves were probably derived from the contact aureole closer to the present level of exposure. These non-igneous enclaves (NIE) and their disaggregation products are present in all major plutons, comprising from < 2 to 5 vol.% of the granites. Enclaves of igneous origin are also present in all major plutons except Carnmenellis, generally comprising < 1 vol.% of the granites. The most common type is intermediate in composition, with microgranular texture, and mineral compositions and textures consistent with an origin by magma mixing. Large crystals of K-feldspar, plagioclase and quartz, common in these microgranular enclaves (ME) but absent in NIE, represent phenocrysts derived from the silicic end-member during magma mixing events rather than products of metasomatism as suggested previously. Although the composition of the mafic end-member (basaltic or lamprophyric) involved in the mixing process is poorly constrained, the presence of ME in the granites, and the preponderance of mantle-derived mafic rocks in the coeval Exeter Volcanics, indicate that mafic magma injection into the crust was a factor in the generation of the batholith. Advection of sub-crustal heat provides an explanation for large-volume crustal melting in regions of relatively thin crust such as southwest England.

The crystal structure, including hydrogen positions, of borcarite, Ca4Mg[B4O6(OH)6](CO3)2, monoclinic, a = 17.840(4), b = 8.380(2), c = 4.445(1) Å, β = 102.04(3)°, V = 649.9(3) Å3, space group C2/m, has been refined by full-matrix least-squares methods to R = 2.5% and wR = 3.0% for 1020 unique observed [|F| ≥ 5σ(F)] reflections collected using Mo-Kα X-radiation. The H positions were located on difference-Fourier maps and were refined using the ‘soft’ constraint that O-H distances are ∼ 0.96 Å. The 4:4T FBB (fundamental building block) of the borcarite structure contains four Bϕ4 tetrahedra (ϕ = unspecified ligand) which share corners to form a four-membered polyhedral ring. Borcarite is the only mineral known to contain this FBB. The FBBs do not polymerize, but each shares three anions with an Mgϕ6 octahedron on either side, forming rods of composition [MgB4O6(OH)6]4- along the c-axis. The rigidity of these rods is enhanced by hydrogen bonding, and individual rods are connected through Caϕ8 polyhedra, Cϕ3 triangles and hydrogen bonds.

Shannonite, ideally Pb2OCO3, is a new mineral species that occurs as mm-sized white porcellanous crusts, associated with fluorite, at the Grand Reef mine, Graham County, Arizona, USA. Other associated minerals are plumbojarosite, hematite, Mn-oxides, muscovite-2M1, quartz, litharge, massicot, hydrocerussite, minium, and unnamed PbCO3·2PbO. Shannonite is orthorhombic, space group P21221 or P212121, with unit-cell parameters (refined from X-ray powder data): a 9.294(3), b 9.000(3), c 5.133(2) Å, V 429.3(3) Å3, a:b:c 1.0327:1:0.5703, Z = 4. The strongest five lines in the X-ray powder pattern [d in Å (I)(hkl)] are: 4.02(40)(111); 3.215(100)(211); 3.181(90)(121); 2.858(40)(130); 2.564(35)(002). The average of eight electron microprobe analyses is PbO 89.9(5), CO2 (by CHN elemental analyser) 9.70, total 99.60 wt.%. With O = 4, the empirical formula is Pb1.91C1.05O4.00. The calculated density for the empirical formula is 7.31 and for the idealized formula is 7.59 g/cm3. In reflected light, shannonite is colourless-grey to white, with ubiquitous white internal reflections (× 16 objectives), weak anisotropy, barely detectable bireflectance, and no evidence of pleochroism. The calculated refractive index (at 590 nm) is 2.09. Measured reflectance values in air and in oil (× 4 objectives) are tabulated. Transmission electron-microscopy studies reveal that individual crystallites range in size from 10–400 nm, are platy, and are anhedral. Physical properties for cryptocrystalline crusts include: white streak; waxy lustre; opaque; nonfluorescent under both long- and short-wave ultraviolet light; uneven fracture; brittle; VHN100 97 (range 93–100); calculated Mohs’ hardness 3–3½. Shannonite is soluble in concentrated HCl and in dilute HNO3 and H2SO4. The mineral name is for David M. Shannon, who helped collect the samples and who initiated this study.

Cordierite is a characteristic mineral of many peraluminous felsic igneous rocks. A combination of T-P-X parameters, which overlap the stability conditions for felsic magmas, control its formation. Critical among these parameters are relatively low T, low P, and typically high (Mg+Fe2+), Mg/Fe2+, A/CNK, aAl2O3, and fO2. Spatial and textural information indicate that cordierite may originate in one of three principal ways in felsic igneous rocks: Type 1 Metamorphic: (a) xenocrystic (generally anhedral, many inclusions, spatial proximity to country rocks and pelitic xenoliths); (b) restitic (generally anhedral, high-grade metamorphic inclusions); Type 2 Magmatic: (a,b) peritectic (subhedral to euhedral, associated with leucosomes in migmatites or as reaction rims on garnet); (c) cotectic (euhedral, grain size compatibility with host rock, few inclusions); (d) pegmatitic (large subhedral to euhedral grains, associated with aplite-pegmatite contacts or pegmatitic portion alone); and Type 3 Metasomatic (spatially related to structural discontinuities in host, replacement of feldspar and/or biotite, intergrowths with quartz). Of these, Type 2a (peritectic) and Type 2c (cotectic) predominate in granitic and rhyolitic rocks derived from fluid-undersaturated peraluminous magmas, and Type 2d (pegmatitic) may be the most common type in fluid-saturated systems.

Petrological and mineral chemical data are presented for two new occurrences of co-existing borosilicate minerals in the Larsemann Hills, East Antarctica. The assemblages contain kornerupine and the rare borosilicate, grandidierite (Mg,Fe)A13BSiO9. Two distinct associations occur: (1) At McCarthy Point, 1–10 mm thick tourmaline-kornerupine-grandidierite layers are hosted within quartzofeldspathic gneiss; and (2) Seal Cove, where coexisting kornerupine and grandidierite occur within coarse-grained, metamorphic segregations with Mg-rich cores of cordierite-garnet-spinel-biotite-ilmenite and variably developed plagioclase halos. The segregations are hosted within biotite-bearing, plagio-feldspathic gneiss. Textural relationships from these localities indicate the stability of co-existing kornerupine and grandidierite.

The grandidierite- and kornerupine-bearing segregations from Seal Cove largely postdate structures developed during a crustal thickening event (D2) which was coeval with peak metamorphism. At McCarthy Point, grandidierite, kornerupine and late-tourmaline growth predates, or is synchronous, with F3 fold structures developed during a extensive granulite grade, normal shearing event (D3) which occurred prior to, and synchronous with, near-isothermal decompression. Average pressure calculations on assemblages that coexist with the borosilicates at Seal Cove, indicate the prevailing conditions were 5.2–5.5 kbar at ∼ 750°C for formation of the grandidierite-kornerupine assemblage.

Penfieldite is a lead hydroxychloride mineral with composition Pb2Cl3(OH). It belongs to the hexagonal system, space group P, a = 11.393(3), c = 4.024(1) Å. The 4 Å c parameter corresponds to the basic sub-cell, whereas modulated structures are known with a true c axis 12 times longer. The average crystal structure of penfieldite has been solved with direct methods and refined to Rw = 0.041 for 871 refections collected with Mo-Kα radiation. The chemical and structural relationships between penfeldite and laurelite, Pb2F3(F,Cl,OH), are briefly discussed. An electron diffraction study of penfieldite revealed the occurrence, besides the common modulated structure with C = 12c, of domains with a 15c periodicity. Moreover, a 9c periodicity has been observed in crystals heated at 180°C Penfieldite is quickly destroyed above 200°C.