Experimental measurements of velocity and tracer gas concentration are taken in a straight tube to evaluate the effectiveness of mixing in achieving conditions as required by ANSI N13.1-1999 for single point extractive sampling from stacks and ducts of nuclear facilities. Mixing is evaluated for inlet turbulent intensities of 1.5%, 10%, and 20%, achieved by introducing various bi-plane grids, and for conditions generated by a commercial static gas mixer. The data obtained (at Reynolds number = 15,000) highlight the importance of inlet turbulence intensity in the process of turbulent dispersion of a dilute gas. The gas mixer does not introduce significant pressure losses and unlike bi-plane grids, the turbulence downstream of the mixer is not homogenous. A judicious choice of the release location that uses the large scale eddies and inhomogeneity of the turbulence ensures that the specified ANSI N13.1-1999 criteria are attained within 7 diameters downstream of the duct inlet. This is significantly more effective than a bi-plane grid where even with 20% inlet intensity the criteria are met only at 21 diameters downstream. The predictions of a proposed semi-empirical correlation match favorably with data. For example, at 18 diameters downstream with inlet intensities of 1.5% and 10%, the predicted coefficients of variation (COVs) of 150% and 65% are close to the actual values of 154% and 50%; where the COV of a set of measurements is the ratio of the standard deviation of the set to its mean value. The corresponding results obtained using commercially available software are 141% and 12%. Results from a particle-tracking model show good qualitative trends, but they should not be used to determine compliance with the requirements of the ANSI standard.