Widespread use of minimally selective fish traps has contributed to the overfishing of Caribbean coral reefs. Traps typically target high-value fish such as groupers (Serranidae and Epinephelidae) and snappers (Lutjanidae), but they also have high bycatch of ecologically important herbivores (parrotfish (Scaridae) and surgeonfish (Acanthuridae)) and non-target species. One strategy for reducing this bycatch is to retrofit traps with rectangular escape gaps that allow juveniles and narrow-bodied species to escape; yet the effectiveness of these gaps has not been thoroughly tested. On the shallow reefs of Cura莽ao, Netherlands Antilles, I compared the catch of traditional Antillean chevron traps (the control) to the catch of traps with short escape gaps (20 脳 2.5 cm), traps with tall escape gaps (40 脳 2.5 cm), and traps with a panel of large aperture mesh. With data from 190 24-h trap sets, the mean number of fish caught was 11.84 in control traps, 4.88 in short gap traps, 4.43 in tall gap traps, and 0.34 in large mesh traps. Compared to controls, traps with short or tall gaps caught significantly fewer bycatch fish (-74 and -80% respectively), key herbivores (-58 and -50% respectively), and butterflyfish (Chaetodontidae; -90 and -98% respectively). The mean length of captured fish was significantly greater in gap traps because juveniles were able to escape via the gaps. Escape gaps reduce neither the catch of high-value fish, nor the total market value of the catch. Therefore, using escape gaps could make trap fishing more sustainable without reducing fishermen's revenues.KEY WORDS: Fish traps 路 Bycatch 路 Escape gaps 路 Parrotfish 路 Caribbean 路 Coral reef fisheries 路 Gear-based fisheries management 路 Artisanal fisheries 415: 201-209, 2010 Jamaica in the 1970s, researchers have examined the influences of soak time, trap shape, bait, habitat type, conspecific attraction, lunar phase, and local level of exploitation (e.g. Munro et al. 1971, Munro 1974, Ferry & Kohler 1987, Dalzell & Aini 1992, Wolff et al. 1999, Robichaud et al. 2000, Garrison et al. 2004). However, research on the management of trap fisheries has focused almost exclusively on the use of larger mesh sizes to reduce the catch of juveniles (e.g. Bohnsack et al. 1989, Robichaud & Hunte 1997, Sary et al. 1997, Stewart 2007. One drawback of this approach is that, given the diversity of fish caught with traps, it is impossible to select one mesh size that optimizes the yield of all exploited species (Mahon & Hunte 2001). An alternative modification to fish traps, the inclusion of escape gaps that allow juveniles and narrow-bodied species to exit, is a promising method of bycatch reduction that deserves further investigation. Munro et al. (2003) tested escape gaps (largest gap size 9.0 脳 3.3 cm) in double-arrowhead-shaped traps (an experimental trap design not used by fishermen), and showed that gaps reduce both catch quantity and catch mass relative to controls. Inspired by Cura莽ao's proposed regulation to require escape gaps in all traps, I e...