Alternative approaches of adjusting urinary concentration of cadmium for differences in specific gravity of biological samples were assessed. The main analysis used 2922 cadmium-in-urine samples collected in the period 1968-1989 from workers at a UK nickel-cadmium battery facility. Geometric means of cadmium-in-urine, adjusted and unadjusted for specific gravity, were obtained for 21 different values of specific gravity ranging from 1.010 to 1.030. There was a highly significant positive trend (P < 0.001) of unadjusted cadmium-in-urine with specific gravity. Conventional adjustment for specific gravity led to a highly significant negative trend (P < 0.001) of adjusted cadmium-in-urine with specific gravity, SG. An approach proposed by Vij and Howell, involving the introduction of a z coefficient, led to satisfactory adjustment. Conventional adjustment of specific gravity leads to overcompensation of the confounding effects of specific gravity. An alternative method is available and should probably be adopted when interpreting urine biological samples for all chemical substances.