Human and Ecological Risk Assessment of Chemical Mixtures207Acknowledgment that mixtures may cause risks that are not fully covered by single compound evaluations does not automatically imply that mixture assessments should be performed for all potential mixture exposures. This depends on the problem at hand, the specific exposure situation, and the available information. If a regulator must make a decision about the remediation of 1 contaminated soil plot, and it is already known that one of the mixture components exceeds the remediation threshold, and this fact results in a need for remedial action itself, then a mixture assessment is redundant. Mixture assessment for contaminated soils (and other compartments and exposure routes) is useful if the known individual components do not exceed their respective thresholds, but if it is suspected that the overall mixture may still cause unacceptable adverse effects, and in case risk managers have a limited budget for a large number of contaminated sites, that is, when they have to prioritize the most hazardous sites to be remediated first, while other sites should possibly be subject to simple risk reduction measures. The situation is different for mixture emissions. Here, potential mixture effects should always be addressed because the aim was and is to establish an emission level that is lower than a certain (acceptable) effect level. Compared to per-compound risk assessment, mixture risk assessment is based on relatively little data, relatively weaker scientific underpinning, and relatively few and simple null models. In contrast to these impler as compared to�characteristics is the presence of all kinds of mixtures in the field, with exposure levels ranging from very low to very high, and pertaining to an innumerable set of possible mixture compositions. Therefore, it is hard to provide proof that mixture risk assessments are scientifically fully justifiable for a case, and it is hard to obtain the relevant data for the most justifiable approach. Despite this, mixture assessments are frequently requested, andhen it is adopted as principle that risk assessors should provide the best possible answer to a problemhe question is whether the approaches that are proposed are etter�than neglecting mixtures, and handling each situation compound by compound. Our firm belief, founded in both review of the field and practical experience, is that executing mixture risk assessments is feasible, at least partially, and often indeed provides better answers than those generated by applying a per-compound approach. This particularly holds for relatively well-characterized common mixtures such as PCBs, PAHs, and coke oven emissions. It also holds for the default models of the component-based approaches, that is, CA and RA, particularly if the MOA of the chemicals is known. But even if the MOA is unknown, there are mathematical resemblances between predictions generated by both mixture models, as long as the slopes of the separate concentrationffect curves are more or less similar (Drescher and Bdeker 1995). For problems where a good assessment is crucial, it is often even feasible to apply multiple approaches, and follow a multicriteria-like analysis, like in TRIAD (see Section 5.3.2.1). When absolute predictions of mixture risks might be out of reach (since those require proof that is not there), elative answers�may be of high value to the risk manager. In many cases the relative answer is already sufficient for improving the risk management, for example, to determine the remediation priority of a set of contaminated soils. Notwithstanding the useful application of current mixture assessment methods, it must be acknowledged that current methods are not always accurate and estimates