Females of Anopheles gambiae s. lat., most of which would have been A. gambiae s. str., were collected from houses in coastal Kenya and tested for their oviposition preferences using Petri dishes in large laboratory cages with lighting equivalent to weak moonlight. Significantly more eggs were laid overnight in water over black than over paler tones, and this difference increased as contrast with the surrounding floor was increased. Direct observation revealed that over white targets, females oviposited from a settled posture, whereas over black targets they did so from flight. The influence on this behaviour of target darkness (tone) overrode that of cage size or target size. In tests which yielded markedly fewer eggs in sea water than in tap water, no significant difference was detected when cage floors were either black or white, although a black floor might have resulted in significantly greater discrimination against sea water had more tests been conducted. All further testing was done over black cage floors. Turbid water from a natural development site received more eggs than distilled, tap or swamp water, even though the turbid water appeared paler than the others. The females did not discriminate between rearing water and tap water, or tap water with and without pupae, but the presence of larvae was repellent. Turbid water from a development site thus seemed to possess an arrestant property which overrode selection favouring darker targets, and which was not derived from prior presence of conspecific immatures. It is suggested that for A. gambiae, oviposition from a settled posture is a response to sub-optimal stimuli, possibly indicating conditions under which oviposition would not occur in nature, and hence why cage experiments using white targets have in the past yielded confusing results.