The purpose of geochemical surveys changes with scale. Regional surveys identify areas where mineral deposits are most likely to occur, whereas intermediate surveys identify and prioritize specific targets. At detailed scales specific deposit models may be applied and deposits delineated. The interpretation of regional geochemical surveys must take into account scale-dependent difference in the nature and objectives of this type of survey. Overinterpretation of regional data should be resisted, as should recommendations to restrict intermediate or detailed follow-up surveys to the search for specific deposit types or to a too limited suite of elements. Regional surveys identify metallogenic provinces within which a variety of deposit types and metals are most likely to be found. At intermediate scale, these regional provinces often dissipate into discrete clusters of anomalous areas. At detailed scale, individual anomalous areas reflect local conditions of mineralization and may seem unrelated to each other. Four examples from arid environments illustrate the dramatic change in patterns of anomalies between regional and more detailed surveys. On the Arabian Shield, a broad regional anomaly reflects the distribution of highly differentiated anorogenic granites. A particularly prominent part of the regional anomaly includes, in addition to the usual elements related to the granites, the assemblage of Mo, W and Sn. Initial interpretation suggested potential for granite-related, stockwork Mo deposits. Detailed work identified three separate sources for the anomaly: a metal-rich granite, a silicified and stockwork-veined area with scheelite and molybdenite, and scheelite/powellite concentrations in skarn deposits adjacent to a ring-dike complex. Regional geochemical, geophysical and remote-sensing data in the Sonoran Desert, Mexico, define a series of linear features interpreted to reflect fundamental, northeast-trending fractures in the crust that served as the prime conduits for mineralizing fluids. At a larger scale, the linear, northeast-trending anomalies can be shown to result from a series of discrete mineralized systems with different ages and mineral assemblages. The linear pattern of anomalies disintegrates. A regional geochemical survey in the Sonoran Desert in southwestern Arizona displays a cluster of samples anomalous in Pb, Mo, Bi and W. In detail, the original regional anomaly separates into four discrete anomalous areas, each with its own distinctive suite of elements, geographic distribution and age of mineralization. A prominent regional gold anomaly in the Gobi Desert, Xinjiang, Peoples Republic of China, extends southeastward for 30 km from known lode gold deposits. Because the anomaly cuts both lithologic units and the structural grain, and because it parallels the prevailing direction of high-velocity winds, it was originally attributed to eolian dispersion. In detail, the regional anomaly consists of several east-west-trending anomalies, parallel to local lithology and structure that most likely reflect independent sources of lode gold. The regional anomaly results from smoothing of an en-echelon set of local anomalies. These examples emphasize that interpretation of regional anomalies must be tempered to consider regional-sized geologic features. Attempts to overinterpret anomalies by assigning deposit-scale attributes to regional anomalies can lead to confusion and incorrect interpretations. Potential targets that can be readily resolved only at intermediate or detailed scales of study may be overlooked. ?? 1991.