[en] DOE workplaces strive to comply with the 10 CFR 850.31(b)(1) surface concentration release criterion. The usual planning considerations for demonstrating compliance are these: how many swipes, and where; which sample preparation and analytical methods; what reporting limits; and what sample statistic to compare with the criterion. We have reviewed swipe samples from hundreds of Nevada Test Site workplaces: office buildings; experimental facilities; forward area field units; shops; and tunnels. Our experiences have led us to a critical examination of the inner workings of the measurement process itself, involving details generally taken for granted when those usual questions are asked. In this presentation we dissect the ICP-AES Be measurement process. We discuss calibration options and how they impact the distributions of analytical results. We look at distributions of blank results obtained from different labs, and discuss their relevance to determining reporting limits. We examine the way measurements are made from spectra, how that process impacts our understanding of the actual statistical distributions of Be measurements, and how interferences can affect Be measurements. Our objective is to gain sufficient confidence in the measurement process so that the usual questions will make sense and the survey results will be credible. Based on our observations, we offer these recommendations: prepare calibration samples in digested blank swipes; force the calibration line through (0,0); base reporting limits on field blank measurement distributions rather than 40 CFR 236 calculations; use, but do not believe, the usual lognormal distribution assumption; and avoid the 234.861 nm emission line