It has been suggested that benzene metabolites might be good indicators of smoking. Moreover, benzene could stimulate the neutrophil lineage while depressing the lymphocytic and erythroid lineages, possibly by an interference with cytokines. The effect on the neutrophil lineage could explain the smokers' leukocytosis, the mechanism of which is presently unknown. Therefore, the usefulness of benzene metabolites as indicators of smoking was compared to that of cotinine and thiocyanate, and the relationships between benzene metabolites, the hematological parameters of smokers, and interleukin 1a production were examined. The results show that benzene metabolites are not better indicators of smoking status than cotinine or thiocyanate. Furthermore, it seems unlikely that the smokers' leukocytosis is benzene induced.