Nocturnally active mosquitoes are attracted visually to large (1·5 × 2·4 m) unpainted plywood suction traps. Such traps, which discharge air horizontally in one direction only, were placed in an open field in Florida in various patterns to study the flight behaviour of mosquitoes. When four traps were placed 15 m apart in a column, the two end traps captured 83 or 53% more mosquitoes than the two inner traps, depending upon whether the air discharged from the traps was perpendicular to, or in line with, the trap column. When the traps were spaced 30 m apart, the end traps captured only 13% more than the inner traps, indicating that at this distance visual competition between traps was small. When 16 suction traps were spaced 15 m apart in a grid of four columns of four traps each, the traps at the four corners of the grid captured about 2·1 times, and traps between the corner traps along the edge of the grid 1·5 times, as many mosquitoes of most species as traps within the grid. Subsequently, four additional traps were placed 15 m beyond the corner traps, and these traps captured 3·3 times as many mosquitoes as traps inside the grid. Traps within the grid had four adjacent traps while traps at the edge, corner, and beyond the corner had three, two and one adjacent traps, respectively. The numbers captured by a trap at a particular location in the grid were inversely related to the number of adjacent traps, viz., suction-trap catches were decreased by about 33% as the number of adjacent traps were increased and thus each trap acted as a competing visual attractant on nearby traps. It was estimated that an isolated suction trap would capture approximately five times as many mosquitoes as a trap inside the grid. Assuming these differences between trap collections were caused by the distance at which mosquitoes respond to the traps and thus would be proportional to the visually unobstructed area about each trap, it was calculated that most adults responded visually to the traps from distances of approximately 15·5 to 19 m. Aedes vexans (Mg.) and Psorophora columbiae (D. & K.) were most responsive, followed by Culex nigripalpus Theo. and Culiseta melanura (Coq.) and then by Anopheles crucians Wied., P. ciliata (F.) and Uranotaenia lowii Theo. Only U. sapphirina (O.S.) and Culex quinquefasciatus Say appeared to possess visual ranges of 7.5 m or less. Visually conspicuous objects can serve as long-distance attractants. Low air velocities caused appreciable reductions in trap collections. At a distance of 15 m, the velocity of air discharged from the traps was only 0·6 m/s, but catches from a trap downstream of an adjacent trap were reduced by 17%. Catches from rows of traps downstream from one, two or three rows of traps were reduced by 15, 36, and 51%, respectively. Estimates of mosquito population densities in a natural environment will be affected by the differing visual attractiveness of objects of irregular distribution and sizes and their effect upon wind velocities.