The results obtained with eight pregnancy kits on 735 urines were compared. Overall accuracies were similar and fairly high (93.4 to 95.37%), although only 78.28% of all pregnancies were simultaneously detected by all eight kits. It is suggested that the different performances of each kit could be accounted for not only by lack of product homologation, differences in kit sensitivity, and laboratorian error, but also by the presence of various hormonal components in various concentrations in the urines of pregnant women, and by the presence of different antibodies against these hormonal components in various concentrations in the antisera included in the various kits.