Three major problems of grain size measurements: (1) grain, (2) size, and (3) measurement

The granulometric properties and grain size data of fine-grained clastic sediments have been used in many different ways to make paleoenvironmental interpretations. This is particularly true for the eolian environments of arid regions. However, the modern methods of grain size determination that are increasingly available to many people sometimes mislead researchers and suggest that the results of measurements are infallible.
Here we present findings that are worth considering when using granulometric data. Different chemical pretreatment methods, particle-sizing techniques, and various devices (laser diffraction devices [Fritsch Analysette 22 Microtec Plus, Horiba Partica LA-950 v2 and Malvern Mastersizer 3000], and an automated image analyser [Malvern Morphologi G3-IDSE]) were applied to measure the grain size of aeolian dust deposits and soils and to quantify the effects of different approaches on particle size data. 
(1) Grain: What does it mean grain? Single grain? Aggregated particles? Coated particles? Pretreatment techniques have a significant impact on clumped particles, aggregates, and single grain granulometric parameters. According to our findings, different widely used chemical treatments have substantial and significantly different effects on the results. The purpose of studies determines the required pretreatments (e.g. particles could have been supplied as silt-sized aggregates; then it is not very useful to decompose them to reconstruct past wind flow mechanisms or post-depositional alterations, but from a geotechnical point of view, the situation is different). Our results on particle size modification effects of five widely applied chemical pretreatment procedures demonstrate the importance of pre-measurement handling of samples.
(2) Size: What is the size of an irregularly shaped particle? Mineral particles are not spheres; their size can only be estimated by applying equivalent diameters. Irregular particle shapes also have an effect on sizing. Automated image analysis provides a wide-scale of shape parameters, which can be used as a novel approach for granulometric characterisation, transport mechanism reconstruction and offers an opportunity to develop granulometric alteration indices. 
(3) Measurement: Particle size differences can also result from measurement technique differences. Direct (image analysis) and indirect (laser scattering) measurement techniques require different prior information about the mineral material to be measured. At a starting point of laser scattering measurements, it is assumed that the analysed particles are spherical, and their complex refractive index is known. In fact, due to the shape anisotropy and mineralogical heterogeneity of particles, fulfilling these requirements is impossible in the case of sedimentary samples and soils. However, even similar approaches (e.g. laser scattering devices of different manufacturers) can provide significantly different size results applying the same optical settings. Grain size distributions were calculated using both the Fraunhofer and Mie scattering theories and a wide variety of optical settings. The unknown 3rd-dimensional size (thickness) of mineral particles is the key source of the uncertainties during image analysis. 
Support of the National Research, Development and Innovation Office (Hungary) under contract NKFIH FK138692 is gratefully acknowledged.