A soil geochemical survey in the Maritime Provinces of Canada and part of the Northeast United States was completed for the North American Soil Geochemistry Landscapes Project. Soil samples, derived largely from unsorted glacial till, were collected over 349 sites, from 0 to 5 cm depth (regardless of horizon), A-, and C-horizons. The 0 to 5 cm depth interval represents the soil of interest in health risk assessments and is termed the Public Health (PH-) layer. The <2 mm fraction of each sample was analysed for a broad suite of major and trace elements using a near-total four-acid digestion, and major mineralogical components were determined by quantitative X-ray diffraction. Multivariate statistical analyses of the logcentred soil geochemistry from the PH-layer and the two soil horizons, and of the soil mineralogy from the A- and C-horizons, reveal distinctive inter-element relationships from deeper soil (represented by the C-horizon) upwards into topsoil (represented by the A-horizon and PH-layer). Statistical dispersion of several elements increases upwards in the soil profile. Maximum data dispersion occurs in the PH-layer and A-horizon soils. Elements including S, P, Pb, Hg, Cd, Se, Mo, Sb, Bi and Sn are relatively enriched in the PH-layer and A-horizon, and are positively correlated with increasing organic carbon contents. The relative enrichment of groups of elements in the C-horizon, in contrast to those elements in the A-horizon and PH-layer, suggests a composition that reflects the geochemistry of the glacial till that is derived from the local bedrock. Elements such as Ni, Mg, Cr, V, Co, Fe and Sc, represent a mafic component of the parent material, and relative enrichments of K, Rb, Zr, rare-earth elements, Li and Al indicate a more felsic component. The patterns revealed by the application of multivariate methods to the soil chemistry and mineralogy are attributed to underlying geology, soil-forming processes, and anthropogenic activity, or combinations of all three factors. Both the soil geochemistry and mineralogy were tested in their ability to predict soil horizon and underlying bedrock lithology or time-stratigraphic assemblages. The geochemistry and mineralogy of the soils are both good for predicting soil horizon; however, the soil geochemistry is better for predicting the underlying lithologies/assemblages than the soil mineralogy.