Males of the oriental fruit fly, Bactrocera dorsalis (Hendel), are strongly attracted to methyl eugenol (ME), a naturally occurring compound reported from ten different plant families (Metcalf 1990). Increasing evidence indicates that this attraction reflects enhanced competitiveness in sexual selection: males that feed on pure ME (Tan & Nishida 1996) or flowers containing ME (Shelly, unpublished data) or ME-like compounds (Nishida et al. 1997) have a pronounced mating advantage over unfed males. Control efforts take obvious advantage of this association, and traps baited with ME and insecticide have been used in numerous programs of male annihilation (Cunningham 1989). In fact, ME traps were successful in eradicating B. dorsalis from entire islands in the western Pacific (Steiner et al. 1965). Despite its past successes, the effectiveness of this method is potentially constrained by at least two factors. First, individual males display variable levels of attraction to ME, and this variation appears to have a heritable component (Shelly 1997). Consequently, prolonged, low-level use of ME traps could select for non-responsiveness to the lure and thus inadvertently reduce the likelihood of eradication (Cunningham 1989). In addition, Shelly (1994) demonstrated that B. dorsalis males that fed on ME were less likely to be captured in traps than control males. Thus, the possibility exists that males that have located and fed on natural sources of ME will avoid ME traps, thereby hampering control efforts. Here, I describe field experiments that examined whether males that fed previously on flowers of Cassia fistula L., which contain ME (Kawano et al. 1968), or flowers of Fagraea berteriana A. Gray, which contain ME-like compounds (Nishida et al. 1997), were less likely to be trapped in ME traps than control unfed males. Flies used in all experiments were from a laboratory stock started with 300-400 adults reared from field-collected mangos. The colony was housed in a large screen cage (1.2 m by 0.6 m by 0.6 m) with superabundant food (a mixture of honey and protein hydrolysate) and water. Ripe papayas were provided periodically for oviposition. Room temperature was maintained at 20-22?C and 65-75% RH, and under these conditions generation time was about four weeks. The experiments were performed when the colony was 3-7 months old, and correspondingly the flies used were 3-7 generations removed from the wild. Sexes were separated within seven days of eclosion, well before reaching sexual maturity at 15-20 days (Foote & Carey 1987). Males in the treatment category were exposed to flowers either one time only (single feeding trials; both plant species) or two times (repeat-feeding trials; F. berteriana only) prior to release. The single-feeding trials were conducted as follows. For C. fistula, flowering stems were removed from a tree growing near the laboratory, put in water, and then placed into screen cages (30 cm cubes). A total of 140-160 flowers was placed in a given cage. Flowers of F. berteriana were obtained commercially as leis (a Hawaiian necklace of flowers). Flowers were used on the day of purchase (presumed fresh), and 40-50 flowers were placed in each cage. Flowers of both plant species were placed in the cages at 0800 hr and immediately thereafter I introduced 50 males (22-26 days old) per cage (along with food and water). Males had been marked