Uncertainty associated with data derived by the analyses of heavy metals in aquatic sediment is due to variance produced in the laboratory (precision), plus 'natural', small-scale spatial variance, (or field variance) at the sampling site. Precision is easily determined and is usually reported in contaminant studies, but field variance is poorly understood and seldom documented. It is important to have an understanding of the field variance because if small-scale spatial variance in the concentration of heavy metals is excessive, regional trends may be limited value. Similarly, if temporal change is large, the results of single synoptic surveys may be questionable and the ability to demonstrate anthropogenic contributions over time will be difficult. However, it is evident from the literature that the information needed to address problems of spatial and temporal variance in the field is beyond the resources of most researchers. Analytical precision of about 5% relative standard deviation (RSD) for heavy metal analysis is typical of a well-managed laboratory. Many studies of small-scale spatial variability made during the current investigation indicate that field variance is related to ambient energy and to the type of sedimentological environment. Total variance (analytical plus field variance) is approximately 10% RSD (mean for a suite of nine trace elements) for depositional parts of estuaries and the marine environment, but increases to about 20-35% RSD for the more dynamic parts of the estuarine environment and the fluvial system. Repeated sampling over periods of up to 7 years undertaken during the present study, indicate a similar order of magnitude for temporal variability in these sedimentological environments. A proposed scheme to provide information on field variance is to undertake small-scale spatial and temporal studies in discrete sedimentological environments in the study area after sediment sampling and characterisation has been completed. The comparatively large proportion of total variance associated with small-scale spatial and temporal variability in the field questions the often excessive cost and effort made in attempting minor reductions in analytical precision in contaminant investigations.