INTRODUCTION The most environmentally desirable response to an oil spill on water is to contain the oil and recover it for reprocessing. Unfortunately@ this may also be the most logistically difficult response@ since it involves the transport of bulky containment equipment to the spill site. This equipment must be capable of holding the oil slick during the recovery operations when the water will likely be moving in some manner relative to the oil. Currently available containment devices are known to fail when the oil and water move relative to each other at low speeds. Generally@ oil cannot be held on a quiet stream flowing at more than about 0.25 m/s (0.5 knots) nor is it possible to tow an oil filled boom at greater than this speed without oil leakage@ even when the water is calm. Although the difficulty of containing oil moving water has been recognized for many years. Very little research on the problem has been conducted. A compute search of the NTIS and API abstract indices identified only 130 entries since 1969 concerning oil spills on rivers; all but about 20 of these described spill incidents. Most boom development occurred in the 1970s during the time that a general understanding of the physics of containment failure was developed. Papers by Wilkinson (1973). Hale et al. (1974)@ Lau and Moir (1979)@ and Delvigne (1989) are examples which describe and review containment hydrodynamics or containment failure. A few novel boom designs intended to overcome containment failure in moving water have been suggested (Wicks@ 1969; Ayers@ 1976). These booms all operate at higher water speeds by utilizing some mechanism to decrease the differential speed between the water and oil. Neither of the two designs cited above have been actively developed. No containment studies have addressed the problem of trapping and recovering very viscous. Dense oils which may not be floating on the water surface.