The role of odors in mosquito host preferences was studied in a village near Ouagadougou, Burkina Faso. Two odor-baited entry-traps were put beside one another and a choice of host odor-laden air was blown out of them. Odors of a human and a calf (of similar mass) were drawn from two tents in which each was separately concealed. Allowances were made for trap position, differences in human-subject attractiveness, CO2 levels, and trap contamination with alternative host odors. Choices for the human-baited trap greater than the 0.5 random expectation were made by Anopheles gambiae s.l. (0.96) and An. pharoensis (0.68). The choices for the human-baited trap of Culex antennatus were significantly lower than 0.5 (0.25), whereas for the Cx. decens species group (0.56), the difference was not significant. Interpretation of the latter result was complicated by the significant effect of CO2 levels on the index. Species caught in low numbers but whose trap distribution showed a bias towards the human-baited trap were An. funestus (total numbers in the human-baited trap to the calf-baited trap = 9:0), Mansonia africana (17: 1), Aedes dalzieli (22:4), and Ae. hirsutus (13:1); species showing bias towards the calf-baited trap were An. rufipes (0:11), Cx. duttoni (0:17), and Cx. nebulosus (2:35). Mansonia uniformis was the only species distributed randomly between the two traps. Molecular identification of the An. gambiae s.l. samples revealed a marked difference in trap distribution: for the human-baited trap the ratio was 52% An. arabiensis to 48% An. gambiae s.s.; for the calf-baited trap, it was 92% An. arabiensis to 8% An. gambiae s.s.