ABSTRACTStatistical data computed from analyses of present‐day sediments may be useful for the recognition of ancient environments. Phi‐based measures may be used for the separation of sands. An arithmetic technique is suitable for suites of samples with a considerable range of grain sizes. This paper presents a re‐exploration of arithmetic scales and statistical measures which may facilitate environmental recognition. To test this technique, analyses from approximately 800 samples were used. Quar‐tile and median values were substituted into Trask's arithmetic measures of quartile deviationQDaand skewnessSka. Values ofQDawere plotted against the medianM/dmm on double‐log paper for four environments. On each graph the plots indicate a linear trend and both the positions and slopes of the individual trend curves are different. There is a decrease of gradient of the curves in the sequence: aeolian‐fluviatile‐beach‐ “quiet‐water”. Values ofQDaagainstSkawere plotted similarly, the gradients of the curves decreasing in the reverse sequence. Thus each set of curves (QDavs.Mdmm;QDavs.Ska) establishes a method for comparing and differentiating sediments fr