The existence of an association between aminolevulinic acid dehydratase activity in erythrocytes (ALA-d) and blood lead level (Pb-B) has been investigated in a male urban population not occupationally exposed to lead. The data show a decrease of the mean ALA-d activity when Pb-B rises but the association is not statistically ascertained in every case when the data are examined by two different procedures. It appears first that the statistical significance depends on an adequate choice of the distribution function of the ALA-d. Secondly, factors such as interindividual variability of this biological index, lack of precision in the analytical measurements of the Pb-B and the range of the environmental exposure under observation may influence the statistical significance and induce misleading statements. The data aggregation allows the avoidance of some inconsistencies. Furthermore, the estimation of mean ALA-d for the different Pb-B may also be biased by confounding risk factors (such as smoking habits). As such factors are not presently taken into consideration in public health surveys, the no-threshold Pb-B hypothesis, suggested by the observed dose-effect relationship, cannot be strongly supported.