Exposures to extremely-low-frequency magnetic fields were assessed by taking personal measurements with a dosimeter calibrated at 50 Hz with a bandwidth of 40-400 Hz. The study group was a population-based random sample of 1,098 Swedish men. Exposures were determined as workday mean, median, maximum, and standard deviation, and the time fraction of the day when exposures exceeded 0.20 µT. For workday means, the 50th percentile was 0.17 µT, and the 75th percentile was 0.27 µT. For median values, the 50th percentile was 0.11 µT and the 75th percentile was 0.16 µT. The strongest correlation (Spearman rank correlation = r&infs;) found was between the workday mean and the fraction of time above 0.20 µT (r&infs; = 0.89). The authors used the same data to estimate exposures for the 100 most common occupations according to the 1990 Swedish census. A minimum of four independent measurements for each occupation was required. Among occupations with low workday mean values were earth-moving machine operator, health care worker, and concrete worker. Among occupations with high workday mean exposures were welder and electrical or electronics engineer or technician. High exposure levels were also found in occupations outside the study base, such as train engine driver and glass, ceramic, or brick worker. Exposures to magnetic fields vary widely, since levels of exposure are strongly affected by factors such as duration of exposure and distance from the source. Large variations often found between individuals within occupations could reflect variations in tasks across different workdays for the particular occupations and/or local conditions such as tools and installations, and/or how the work is organized and performed.