Potential contamination at ex-industrial sites means that, prior to change of use, it will be necessary to quantify the extent of risks to potential receptors. To assess ecological hazards, it is often suggested to use biological assessment to augment chemical analyses. Here we investigate the potential of a commonly recommended bioassay, the earthworm reproduction test, to assess the status of urban contaminated soils. Sample points at all study sites had contaminant concentrations above the Dutch soil criteria Target Values. In some cases, the relevant Intervention Values were exceeded. Earthworm survival at most points was high, but reproduction differed significantly in soil from separate patches on the same site. When the interrelationships between soil parameters and reproduction were studied, it was not possible to create a good model of site soil toxicity based on single or even multiple chemical measurements of the soils. We thus conclude that chemical analysis alone is not sufficient to characterize soil quality and confirms the value of biological assays for risk assessment of potentially contaminated soils.