A tracer gas method employing sulfur-hexafluoride (SF6) was developed to estimate the efficiency of industrial exhaust hoods controlling gaseous and vaporous emissions, and a study was performed to determine the minimum sampling distance from the exhaust hood at which the sampling probe should be located to minimize error caused by the nonhomogeneous dispersion of SF6 in the duct. SF6 was measured with a gas chromatograph equipped with an electron capture detector. It was found that the sampling probe should be located in the straight duct at a minimum distance of approximately 50 duct diameters, where SF6 was considered to be dispersed uniformly across the duct cross section, and sampling could be conducted at any depth in the duct. With a straight duct side branch combination, the SF6 dispersion was somewhat accelerated and the minimum sampling distance was reduced to 25 and 40 duct diameters at Reynolds numbers of 120,000 and 200,000, respectively. For a straight duct with one elbow, the minimum distance was significantly reduced to 7 duct diameters. When two elbows were present, a minimum distance of approximately 4 duct diameters was found.