Worker exposures to metalworking fluids were characterized at a plant that produced air compressors. Full-shift, side-by-side air samples (n = 147) were collected and analyzed for total particulate matter, extractable total particulate matter, thoracic particulate matter, and extractable thoracic particulate matter. The thoracic particulate matter geometric mean of 0.32 m/m(3) was below the National Institute for Occupational Safety and Health (NIOSH) recommended exposure limit (REL) of 0.4 mg/m(3). The total particulate matter geometric mean of 0.52 mg/m(3), however, was above 0.5 mg/m(3), the total particulate matter concentration offered as a surrogate REL in the NIOSH Criteria for a Recommended Standard for Occupational Exposure to Metalworking Fluids. Of the 83 total particulate matter results that were at or above smash 0.5 mg/m(3), only 50 (60%) of the corresponding thoracic particulate matter results were at or above 0.4 mg/m(3). These data indicated a conversion factor of 1.65 between thoracic particulate matter and total particulate matter concentrations and 1.40 between thoracic extractable particulate matter and total extractable concentrations. These factors were significantly different from the 1.25 used to compare total particulate matter with thoracic particulate matter concentrations in the NIOSH Criteria Document (p < 0.01) and call into question the validity of a universal conversion factor. The authors conclude that thoracic particulate matter exposure assessment should be done directly. In terms of protecting the worker, however, the 1.25 conversion factor appeared to be conservative since each time a total particulate matter result was below 0.5 mg/m(3), its paired thoracic particulate matter measurement was below 0.4 mg/m(3).