Odor-baited traps (OBTs) are increasingly being considered for use in sampling, surveillance and control of malaria mosquitoes. Most odor baits contain carbon dioxide, which apparently enhances trap catches given its role as a mosquito activator. Unfortunately, organic CO2, which is the most locally amenable source of the gas, must be replaced after each experimental night. This precludes the application of carbon dioxide-based odor baits for prolonged area-wide use. This study explored the possibility of replacing organically-produced CO2 with 2-butanone in odor blends targeting malaria mosquitoes. Experiments were conducted under field and semi-field conditions in western Kenya. A fresh preparation of organic carbon dioxide was made on each experimental night. The ketone 2-butanone was impregnated on nylon strips and used repeatedly for entire durations of the experiments. In preliminary investigations it was observed that the numbers of laboratory mosquitoes attracted to a reference blend (MB5) were significantly more than those attracted to MB5 without its CO2 component (P < 0.001), CO2 only (P < 0.001) or an unbaited trap (P < 0.001). Whereas the unbaited trap caught significantly fewer mosquitoes than all the baited traps (P < 0.001), the trap containing CO2 only caught more Anopheles gambiae sensu stricto (hereafter referred to as An. gambiae) mosquitoes than the one containing MB5 minus CO2 (P < 0.001). In all cases the reference blend attracted a significantly higher number of mosquitoes than its variants containing the different dilutions of 2-butanone used to replace CO2 (P = 0.001). The highest catches were associated with the 99.5% and 1.0% concentrations of 2-butanone. The reference blend formed the best barrier for reducing house entry of An. gambiae. Although more wild female An. gambiae sensu lato, An. funestus and Culex spp. were attracted to a variant of MB5 containing pure 2-butanone, the catches did not differ significantly from those due to the intact reference blend (P = 0.450, P = 0.090, P = 0.075, respectively). When compared to existing sampling methods, female An. gambiae s.l. mosquitoes were highly attracted to a CDC light trap than to a human subject (P = 0. 001), MB5 (P = 0.001), or MB5 with its CO2 component replaced with pure 2-butanone (P = 0.001). This study, which demonstrates that 2-butanone can serve as a good replacement for CO2 in synthetic mosquito attractants, further underscores the possibility of using OBTs for monitoring and surveillance of malaria and other mosquito vectors.