This project has two purposes. As the statistics core, the project will provide statistical support to other Superfund projects. This will typically involve assistance in experimental design or in the analysis and interpretation of data. In addition, this project proposes the development and refinement of statistical methods and algorithms for the analysis of toxics measurement data and for the creation of new and improved measurement techniques. Many analytical methods can be made more efficient and effective by careful statistical design and analysis of the data. This may be important to human health since its allows more frequent monitoring of hazardous sites for the same cost, and since it aids in the development and use of analytical techniques to detect toxic substances both clinically and in the field at lower levels and with greater accuracy than existing methods. A particular emphasis is on statistical contributions to the development of improved ELISA protocols and to the better analysis of data from existing protocols. Techniques such as empirical Bayes estimation, nonlinear optimal design, maximum pseudo likelihood, and M estimation will allow these methods to deal with complex problems of nonlinear calibration, nonconstant variance, possible outliers, and values near or below detection limits.