The sedimentological, metamorphic, petrographic and geochemical characteristics of the Armina Formation, part of the Paleoproterozoic Greenstone Belt of Suriname in South America, are described, based on field, geochemical and petrographic evidence obtained through fieldwork along the Marowijne River and study of diamond drill cores from Rosebel Gold Mine (RGM). The metagreywackes show characteristic features of deposition by turbidity currents: coarse-grained, poorly sorted graded greywackes, covered by fine-grained, parallel-laminated phyllitic beds, often with convolute structures and climbing ripples. Their immature character and composition suggest deposition in an arc-trench environment. In the Marowijne River three different facies of metagreywackes are distinguished: (1) the greyish Bonnidoro Falls facies, characterised by common red millimetre-sized pseudomorphs after siderite in the finer beds, (2) the green Paroe Tabiki metagreywacke facies, with decimetre-sized calcsilicate nodules, both metamorphosed in the lower greenschist facies with chlorite as the main mafic mineral, and (3) the grey Armina Falls metagreywacke facies, geochemically similar to the Bonnidoro type but of higher metamorphic grade with biotite as the main mafic mineral. The metagreywackes from the Marowijne River show a predominance of quartz, plagioclase and lithic (tonalitic) clasts, suggesting exhumation of tonalite–trondhjemite–granodiorite plutons before deposition of the turbidites. There is a slight increase in maturity from (1) to (3), suggesting increasing weathering in the source areas. The metagreywackes of the RGM (JZone) have a predominantly metavolcanic origin, suggesting that they have a different provenance area than the Marowijne metagreywackes. Geochemically the spread in composition within each facies is larger than between the facies because of the wide range in grain sizes in each turbidite sequence. A large part of the rocks from the RGM, classified by previous authors as arenites, are geochemically and petrographically metagreywackes. Only a few RGM samples are real arenites, and plot as a separate cluster in geochemical factor score plots because of their low Fe and Na contents.

A version of the Rao–Blackwell theorem is shown to apply to most, but not all, stereological sampling designs. Estimators based on random test grids typically have larger variance than quadrat estimators; random s-dimensional samples are worse than random r-dimensional samples for s < r. Furthermore, the standard stereological ratio estimators of different dimensions are canonically related to each other by the Rao–Blackwell process. However, there are realistic cases where sampling with a lower-dimensional probe increases efficiency. For example, estimators based on (conditionally) non-randomised test point grids may have smaller variance than quadrat estimators. Relative efficiency is related to issues in geostatistics and the theory of wide-sense stationary random fields. A uniform minimum variance unbiased estimator typically does not exist in our context.