Petroduric and 'petrosepiolitic' horizons in soils of Namaqualand, South Africa .

Indurated, light-coloured 'sepiocrete' horizons have been found in Namaqualand Calcisols and Durisols. These horizons resembled calcrete but were non- to only mildly calcareous, resisted slaking in acid and alkali, and often broke with a conchoidal fracture. The presence of elevated quantities of sepiolite in the bulk-soil was confirmed by XRD analysis. The degree of induration in some these horizons suggested cementation by silica, and so in this paper the slaking properties, bulk chemistry, mineralogy and micromorphology of these horizons are compared with the typical silica-cemented, reddish-brown petroduric/duripan (dorbank) encountered in the region. 'Sepiocrete' horizons are chemically, mineralogically and morphologically distinct from the petrocalcic and petroduric horizons with which they are commonly associated. Micromorphology of the petroduric horizons revealed prominent illuviation in the form of oriented clay parallel to grains and crescent coatings on voids, a red matrix due to iron oxides, and translucent, isotropic amorphous silica coatings on grains and voids. In the 'sepiocrete' horizons, sepiolite appeared as a matrix of interlocking, sub-parallel fibres while the amorphous material was localised. The amorphous material was silica-rich with prominent aluminium and lesser magnesium; light brown under plane polarised light; not completely isotropic and had a lower birefringence than the sepiolite. The calcite was usually micritic, but also appeared as loose granules and as elongate crystals in a sepiolite matrix. The presence of the laminar Si-Al -rich areas on the sections suggested at the least localised duric properties and so mutual reinforcement of sepiolite and silica is possible. However, the 'sepiocrete' horizons did not meet the slaking requirements of the petroduric (dorbank) horizons and are distinct in appearance to the typical petroduric horizons in the region. They contained more MgO than the region’s typical petroduric, and too little SiO₂ to be silcrete. While the '-crete' terminology provides a useful expression of the cemented nature of the horizon, in order to fit existing soil classification and description schemes the terms 'sepiolitic' and 'petrosepiolitic' (in the same sense as 'calcic' and 'petrocalcic') are proposed and defined. The term 'sepiolitic' would be useful in the adjectival form in petrocalcic or petroduric horizons where sepiolite is significant but not the primary cement. The genesis of the 'petrosepiolitic' horizons is likely to be essentially similar to that of petrocalcic and petroduric horizons, except for chemical differences in the matrix solutions from which secondary minerals were precipitated, dictated by the pH and evaporative evolution of the soil solution.