Summary Diagnosis of allergic contact dermatitis depends upon the use of epicutaneous patch tests with suspected antigens. It has been claimed that the method is unreliable because of variability. Therefore, we have tested this by comparing responses to various forms of replicate, standardized challenge in nickel-sensitive subjects. Nickel-containing coins applied at 2-4 different sites on the back elicited similar responses in up to 95% of tests. Serial dilutions of nickel sulphate (NiS04) applied, in duplicate, gave highly reproducible responses on the back. However, the forearm skin was clearly less responsive. Examination of the effect of the quantity of nickel-containing vehicle showed that, above a certain minimal level, this had no effect on the reproducibility of the responses. In conclusion, the immune system can give highly reproducible responses when presented with standard eliciting stimuli.