This study was designed to determine whether variability in bone mineral content (BMC) at the lumbar vertebrae (L2-4), radius shaft (RS), femoral neck, and distal radius can significantly contribute to the variability observed in body density (Db) among 89 females (age = 25.1 +/- 5.3 yr) of varying activity levels and menstrual status. Theoretical differences in Db were calculated at +/- 1 and +/- 2 standard deviations of BMC (SDBMC) for the population as well as for the subgroups: eumenorrheic inactive controls (C), recreational runners (RR), collegiate runners (CR), body builders (BB), swimmers (S), and amenorrheic runners (AR). Multiple regression to predict Db yielded significant coefficients (b) for BMC at L2-4 (b = 0.0190, P less than 0.001) and RS (b = 0.0425, P less than 0.01) when added separately to the sum of four skinfolds (subscapula, abdomen, thigh, calf). The differences in % BFHW at +/- 1 and +/- 2 SDBMC for the sample mean for RSBMC were +/- 1.0% and +/- 2.0%, respectively. Variability in L2-4 contributed differences of +/- 1.3% and +/- 2.6% at +/- 1 and +/- 2 SDBMC. The subgroup % BFHW differences (due to L2-4 and RS combined) ranged from an average overestimation of 1.3% for the AR to an average underestimation of 1.4% for the BB. Estimated mean errors for remaining groups were less than or equal to 0.5%. Individual differences ranged from a 3.3% underestimation (BB) to a 3.0% overestimation (AR).(ABSTRACT TRUNCATED AT 250 WORDS)