Laboratory experiments intended to test the effects of a purported ovipositional deterrent pheromone of Anopheles gambiae Giles sensu stricto (Diptera: Culicidae) led instead to the finding that egg cannibalism by fourth-instar larvae explained the ‘deterrent effect’ rather than chemical signals from larvae. Virtually no eggs were counted from cups containing fourth-instar larvae, after oviposition. Numbers of eggs from cups containing younger, conspecific larvae, or pupae did not differ significantly from water only control. Video recording of larval behavior revealed that fourth-instar larvae consumed entire eggs in one continuous ingestive bout. Egg output in the presence of fourth-instar larvae with their mouthparts sealed was similar to the larvae-absent treatment, but nearly 100-fold greater than that counted in the presence of fourth-instar larvae that could eat freely. Thus, the huge egg reduction in eggs counted in the presence of fourth-instar larvae was due to egg cannibalism, not because presence of these larvae deterred oviposition as has been reported in previous studies.Third and especially fourth-instar larvae also cannibalized newly hatched first-instar larvae. Artificial food availability did not eliminate but did reduce cannibalism during first several hours of exposure. The rate at which fourth-instar larvae consumed conspecifics rose significantly and linearly with prey density, generating a non-plateauing, type I functional response curve.These results clearly show that care must be taken such that mosquito ovipositional experiments are not confounded by presence of cannibalizing larvae.