The effect of chromium(VI) on the lifespan of laboratory-reared guppies (Poecilia reticulata) has been studied both in the absence and in the presence of the antioxidant D-mannitol, and it has been compared with that produced by vanadium(V). The three substances used as additives exhibited either a weak (D-mannitol), a moderate (chromate) or an acute (vanadate) toxicity to fish. Vanadate, with LC50 (7 days) = 3.84 x 10(-5) mol/L, was about ten times more toxic than chromate, with LC50 (7 days) = 3.42 x 10(-4) mol/L as a single additive and 4.27 x 10(-4) mol/L in the presence of d-mannitol. An increasing effect on the maximum lifespan of males was observed when the additives studied were used at low concentrations, either alone or in a binary combination, following the sequence: vanadate (14%) < D-mannitol (41%) < chromate + D-mannitol (57%) < chromate (69%). Of these substances, only chromate increased also the maximum lifespan of females (72%). The maximum lifespan showed a strong, positive correlation with the concentration of chromate for males (P = 0.00008) and a weaker, positive correlation (P = 0.116) for females. These results suggest the existence of a chemical-hormesis phenomenon that might be subjected to sexual-genre variability. Both the toxicity and the chemical-hormetic effect provoked by chromate were substantially decreased when it was used in combination with d-mannitol, and the possible causes for this double inhibition are briefly discussed.