Researchers from the National Institute for Occupational Safety and Health (NIOSH) conducted an evaluation of carbon monoxide emissions and control from gasoline-powered generators on houseboats. This evaluation is part of a series of studies conducted by NIOSH investigators during the past several years to identify and recommend effective engineering controls to reduce the CO hazard and eliminate CO poisonings on houseboats. Emission and dispersion characteristics of side-exhausted generators and the performance of several emission control devices (ECDs) that were manufactured by Enviromarine LLC, were studied. Recently developed prototype and production ECDs were retrofitted onto gasoline-powered generators used on houseboats to reduce the hazard of carbon monoxide (CO) poisonings from the exhaust. The prototype ECD had previously been evaluated by NIOSH when it was new and had been used on a houseboat generator for approximately 3,000 hours since the previous testing. The production ECD had several modifications from the prototype and had not been previously evaluated or used. Study results presented in this report address CO emission and exposure performance of side- exhaust houseboats that had been rafted together and the performance of prototype and production ECDs that were rear exhausted. The majority of data gathered during ECD testing involved direct emissions monitoring rather than ambient air sampling. Tests were conducted to evaluate the performance of both ECDs during cold starts and under various loading conditions. Environmental monitoring was conducted for several air contaminants that had not been previously evaluated. This and future research reports will be shared with the U.S. Coast Guard office of Boating Safety and the American Boat and Yacht Council for potential future rulemaking and standard setting. Based upon the results of this study, with side exhausted houseboats, it is clear that any uncontrolled exhaust from a gasoline-powered generator that is close to the water and boat could potentially be hazardous. Peak and average CO concentrations on the lower deck of three rafted, side-exhausted houseboats exceeded 1,000 ppm and 140 ppm respectively. CO concentrations near the waterline were even higher. The current evaluation also showed that the ECD provided a safer environment to individuals on or near the houseboat, when compared to generators having no ECDs. Mean and peak CO (concentrations were reduced at numerous locations on the houseboat when the ECD was operating. Average and peak CO concentrations were typically reduced by greater than 75% on the houseboat's lower deck; however, because measured CO concentrations were fairly low with the generator operating without the ECD, the magnitude of the CO reduction was fairly small. Emissions testing demonstrated that both the prototype and production ECDs were able to reduce generator CO emissions by over 50% during cold-starts and by several orders of magnitude during normal operating conditions when compared to the generator operating with no ECD. Comparison of the results from the prototype and production ECDs indicated that the design of the production ECD is superior to the prototype version. Both versions were able to reduce CO concentrations, but the production model was more effective. It was also learned that brief (less than 1 minute) peak CO concentrations may occur during cold starts with the prototype or production ECD and that generator loading influences ECD performance. Based upon the results of this and earlier studies, NIOSH investigators recommend that U.S. houseboats using gasoline-powered generators, should be evaluated for potential CO exposures and poisonings near the lower rear deck. Houseboat owners should consider retrofitting the generators with engineering controls to reduce the potential hazard of CO poisoning and death to individuals on or near the houseboat. The performance of the evaluated ECDs was generally good and design modifications in the production model should enhance long term performance. NIOSH researchers continue to believe that the ECD is a promising emission control option. Because some performance complications were noted with the prototype ECD, additional testing and evaluation of production ECDs, having substantially more hours of operation, is warranted.