The results of recent wind-tunnel and towing-tank studies are summarized. These studies were designed to obtain basic physical understanding of flow patterns and diffusion, to provide guidance in locating sources, and to provide 'rules-of-thumb' for estimating surface concentrations when a source is located in complex terrain. In the neutral-flow, wind-tunnel studies, the effects of the terrain were evaluated through a terrain amplification factor, A, which is defined as the ratio of the maximum surface concentration that occurs in the presence of the terrain to the maximum that occurs from a similar source located in flat terrain. A's were evaluated for a series of two- and three-dimensional hills of varying slope placed in simulated atmospheric boundary layers. Sources of various heights were placed upstream, at the top, and downwind of the hills. The largest A's were observed for downwind sources with values as large as 15. Flow structure and plume behavior in complex terrain are dramatically altered by the addition of stable stratification. A dividing-streamline concept has been shown to be a useful indicator in determining whether a plume will impact on a hill or surmount the top. When a plume is emitted below the dividing-streamline height and hence impinges on a downwind hill, surface concentrations can be essentially equal to those observed at the plume centerline in the absence of the hill. Favorable comparisons are made between towing-tank simulations and field observations of plumes impinging on a three-dimensional hill. Finally, limitations of the towing tank for simulating strongly stratified flows over two-dimensional hills are pointed out.