We evaluated worker lead exposures and cleaning effectiveness during initial cleanup of 19th-century buildings with highly deteriorated lead-based paint. Eighteen rooms of similar size and condition in two university-owned buildings were selected for a pilot project to compare three methods for removing loose paint, paint chips, and dust. The methods used were: dry scraping followed by dry sweeping (no engineering or work practice controls); wet scraping and high-efficiency particulate air (HEPA) vacuuming; and the latter method with the addition of a portable HEPA-filtered exhaust fan in the room providing about 40 air changes per hour. The final step for all methods was wet-mopping once with tri-sodium phosphate solution. During a single day 18 rooms were cleaned; each of three two-person work crews cleaned six rooms, two with each method. Air and surface samples were collected before, during, and after cleaning. All of the methods were potentially hazardous to workers: 44 percent of the method-based exposures (range: 5.0-360 ug/m3) and one of five full-shift exposures exceeded the OSHA PEL (range 9.4-110 ug/m3). Lowest worker exposures were during the wet scraping and vacuuming method (mean: 24 ug/m3). Providing general ventilation in rooms did not reduce worker exposures and appeared to increase them (mean: 73 ug/m3). Overall, the mean floor surface lead levels were reduced 50 percent after cleaning (from 2,600 to 1,300 ug/ft2), but the effectiveness of the three methods in reducing floor lead levels did not differ significantly. Overall, the method, mean paint lead concentration, pre-cleaning surface lead concentration, and work crew were significantly associated with the mean worker exposures during cleaning (p=0.023), but not with the post-cleaning surface lead concentrations (p=0.13).