A standard practice in the analysis of contaminant concentrations is to conduct the statistical analyses in the logarithmic scale. This practice ยฎnds support in the empirical fact that many contaminant concentration data appear to be lognormally distributed. However, regulatory rules such as those followed by the US Environmental Protection Agency require that risks should be characterized in terms of the mean contaminant concentration. Furthermore, recent studies suggest that the lognormal distribution may exhibit heavier tails than required in practice for adequate description of contaminant concentrations. Most of the studies on contaminant concentration data deal with one-sample or several sample problems. In this article, we examine the above issues in situations where the contaminant concentration data are supplemented with measurements on concomitant variables such as environmental factors. We use log-regression with arbitrary error distributions as the underlying models and develop estimation and inferential methods for the moments of the contaminant concentration at a new set of experimental conditions. We focus on maximum likelihood, bias correction, minimum variance unbiased estimation, nonparametric estimation and conยฎdence intervals. To illustrate the methods we provide a detailed analysis of data on calcium and magnesium concentrations in samples of water from the Fraser River, British Columbia, Canada, as well as some comparisons using a small sample of the year production and price of ground nuts and cotton in Israel.