The primary objective of this study was to evaluate the application of dispersant to spilled oil as a means of reducing adverse environmental effects of oil spills in nearshore, tropical waters. The results of numerous laboratory and field studies have suggested that dispersants may play a useful role in reducing adverse impacts on sensitive and valued environments such as mangroves, seagrasses, and corals. However, the use of dispersants has not been allowed thus far in most situations because of a lack of direct experimental data on the various effects of dispersants and the environmental trade-offs presumed to occur as a result of their application to crude oils. To accomplish this objective, a 2!-year field experiment was designed in wh ich detailed, synoptic measurements and assessments were made of representative intertidal and nearshore subtidal habitats and organisms (mangroves, seagrass beds, and coral reefs) before, during, and after exposure to untreated crude oil and chemically dispersed oil. The results were intended to give guidance in minimizing the ecological impacts of oil spills through evaluation of trade-offs in the relative impacts of chemical dispersion to tropical marine intertidal and subtidal habitats. METHODS The experimental design was intended to simulate a severe, but realistic, worst-case scenario of two large spills of fresh crude oil in nearshore waters, one treated with chemical dispersant and the other left untreated. The experimental scenarios used in this project were developed on the basis of the collective experience of the API Task Force members and the project scientists, and the oil and dispersed oil volumes selected were uniformly acknowledged as being very strong tests of the potential impacts of each. This was particularly true for the dispersed oil scenario since almost all recommended dispersant use strategies call for treatment of oil slicks in deep water after a certain amount of natural weathering has occurred. This is almost totally contrary to the experimental procedure used in this study, so it must be noted that the dispersed oil scenario represents an extreme case, such as might occur if a large (relative to tl-Je area of water), fresh oil slick were chemically dispersed in the shallow ,;raters of a slowly flushed, semienclosed bay. 1 ,930 2,235 475 755 column expressed as the dispersed and UNTREATED OIL SITE 150 166 103 165 65 106 Hydrocarbon Concentration