ABSTRACT A measure of grain shape is needed for incorporation in calculations of the behaviour of grain populations (for example during transport by fluids). Many shape measures have been proposed, most of them for application to single grains rather than to populations. In this paper three such shape parameters are evaluated for samples taken by size fraction from each of three parent sands. The chosen parameters are the maximum projection sphericity of Sneed & Folk (based on triaxial measurements made on the grains), the dynamic shape factor of Briggs (based on settling velocity in water), and rollability, after Winkelmolen (based on rolling behaviour in a specially mounted rotating cylinder).
It is shown that the Sneed & Folk parameter and rollability both discriminate clearly between the shape characteristics of the three sands over the size range 150-500 μm. Moreover the discrimination of the two parameters is mutually consistent. However, dynamic shape factor gives results which for sizes smaller than 300 μm are inconsistent with those of the other two methods and which do not discriminate reliably between the populations. This is inevitable because the differences between drag on spheres and on other shapes become very small at Reynolds Numbers corresponding to those which obtain in settling tests on grains smaller than 300 μm.