There is conflicting evidence concerning seasonal variability of patch test results and no evidence concerning the influence of season on weak-positive, possibly false-positive, irritant reactions, which was analysed in the present study. Data collected in the German Information Network of Departments of Dermatology (IVDK) 1992 to 1997 were combined with external environmental data on temperature and absolute humidity in Germany, and bivariate as well as logistic regression analyses performed concerning the association between reactivity to selected allergens and air temperature and absolute humidity on the respective days of patch testing. Between 39,239 and 41,629 patients had been tested with the 4 allergens considered here. Only formaldehyde exhibited a distinct increase in questionable or irritant as well as weak-positive reactions associated with dry, cold weather. Methylchloroisothiazolinone/methylisothiazolinone and lanolin alcohol showed only a weak, and epoxy resin no, association with climatic conditions. The results indicate that weak-positive reactions, at least to formaldehyde as a prototype of an allergen which is at the same time a marginal irritant, may sometimes be irritant rather than allergic. Patients showing these should be re-tested to improve the poor reproducibilty observed with this allergen.