Historical air-sampling data measured within 320 km (200 mi) of the Nevada Test Site (NTS) have been reviewed for periods following atmospheric nuclear tests, primarily in the 1950s. These data come mostly from high-volume air samplers, with some from cascade-impactor samplers. Measurements considered here are for beta radiation from gross fission products. The resulting air-quality data base is comprised of almost 13,000 samples from 42 sampling locations downwind of the NTS. In order to compile an accurate air-quality data base for use in estimating exposure via inhalation, raw data values were sought where possible, and the required calculations were performed on a computer with state-of-the-art algorithms. The data-processing procedures consisted of (1) entry and error checking of historical data; (2) determination of appropriate background values, air-sampling volumes, and net air concentrations; and (3) calculation of integrated air concentration (C) for each sample (considering fallout arrival times). Comparing C values for collocated high-volume and cascade-impactor samplers during the Upshot-Knothole series showed similar lognormal distributions, but with a geometric mean C for cascade impactors about half that for the high-volume air samplers. Overall, the uncertainty in C values is about a factor of three. In the past, it has been assumed that C could be related to ground deposition by a constant having units of velocity. In our data bases, simultaneous measurements of air concentration and ground deposition at the same locations were not related by a constant; indeed, there was a great amount of scatter. This suggests that the relationship between C and ground deposition in this situation is too complex to be treated adequately by simple approaches.