Ten manual metal arc (MMA) high alloy stainless steel (SS) welders were studied during one week and the concentrations of chromium (Cr) and nickel (Ni) were determined in their urine and blood. Stationary and personal air samples were collected from the immediate work environment; they covered the entire work period. Spot urine samples were collected during the follow-up period. Whole blood and plasma samples were taken from the workers before and after one shift, and the retention rate of magnetic dust in the lungs was estimated with magnetopneumography. On the basis of the results, indices of short-term exposure to Cr and Ni were evaluated. Urinary Cr and Ni concentrations (corrected to creatinine) reflect both the body burden caused by long-term and short-term exposure to easily soluble fractions of these metals. The results indicated that the use of Cr and Ni urinary analyses as indices of short-term exposure is not as dependable as previously assumed. The Cr and Ni concentrations in whole blood and plasma did not correlate with the measured exposure, but the daily mean increase in the Cr concentration reflected exposure to total Cr and Cr(VI) very well. The large variation in the Cr concentration of the morning urine (0.01-2.7 mumol/1) and blood (0.05-1.43 mumol/1) indicated large personal variations of body burden among the exposed welders. The retention rate of magnetic dust in the lungs correlated well (P less than 0.01) with the daily mean increase of Cr in blood. Very good correlations (P less than 0.001) were found between the retention rate of magnetic dust and the personal air samples of Cr and Cr(VI).