The weight percentage and mean size of the 10–63 μm terrigenous silt fraction (termed ‘sortable silt’) of fine-grained marine sediments have recently been argued to be proxies for near-bottom palaeocurrent intensity. This paper details the accuracy and precision of the Sedigraph 5100 and the Coulter Multisizer IIe in determining the weight percentage and mean size of the sortable silt at abundances characteristic of deep-ocean sediments. Three sortable silt standards and a fine silt/clay diluent were generated from natural deep-sea sediment of the Iceland Basin. These were used to produce silty clays with standard additions of sortable silt ranging from 1% to 20% of the total by weight. Accuracy was measured relative to the size of the pure spike measured by each instrument. At sortable silt abundance >5%, the Sedigraph and the Coulter Multisizer estimate both weight percentage and mean grain size precisely and accurately, while below 5%, instrumental noise degrades their performance. On the basis of the Sedigraph results, a series of percentage errors is defined, which may be applied to sortable silt mean size and weight percentage estimates used in palaeoceanographic near-bottom current reconstruction. It is suggested that, ideally, palaeocurrent reconstructions should be based on Sedigraph determinations of the 0–63 μm grain size distribution, yielding both mean and abundance, or on Coulter Multisizer determinations of the sortable silt grain size mean. The Coulter Multisizer may prove especially useful in regions of very depleted (<5%) sortable silt component.