Pedometrics, essentially the application of probability and statistics to soil, has its origins in agronomy in the early part of this century. For several decades it was a tool for designing experiments and surveys and in advisory work. It burgeoned in the 1960s for predicting soil properties at remote sites, for creating and analysing classifications, and for exploring multivariate relations. The 1970s brought a reappraisal, leading to changes of classifying strategy from hierarchical to non-hierarchical and to a new appreciation of the nature of soil variation itself. Pedometricians began to treat soil properties as spatially correlated random processes and to tap the richness of geostatistics for analysis and prediction.
Image-analysing computers have brought new opportunities for analysing soil structure, especially that of the pores, their sizes, shapes, and topology. There is still a need to relate the directly observed porous structure to hydraulics and mechanical behaviour of soil. Fractals may have a role.
To a more limited degree pedometrics has helped in elucidating pedogenesis by quantifying relations between individual soil properties and controlling factors. Solving the full system of multivariate equations needed to describe the products of soil genesis in individual regions, let alone globally, remains one of the biggest challenges for pedometricians.
Advanced pedometric techniques and the theory on which they are based now exceed the comprehension of many soil scientists. It is time to teach the subject from sound and clear expository texts.