then seized larvae of A. farauti and C. pullus without discrimination. Thus it seems that the bottom-dwelling habit of this odonate nymph prevents its encountering the surface-dwelling larvae of Anopheles in the normal course of events, and leads to its apparent preference for culicine larvae as food.

The number of pupae eaten daily by each nymph of O. villosovittatum averaged 10.3. That less pupae than culicine larvae were destroyed was largely consequent on the erratic movements of the former affording them some degree of protection from their formidable enemy.

Nymphs of O. villosovittatum and of other Anisoptera kept under observation in the laboratory consumed larvae and pupae of Chironomidae, odonate nymphs smaller than themselves, and young tadpoles, as well as the developmental stages of mosquitoes. Garman (1927) considers that dragonfly nymphs probably devour members of all aquatic families of insects as well as other aquatic animals outside this class. Young (1921) used nymphs of Pantala flavescens in feeding experiments at Manaos, Brazil. He records that these nymphs destroyed large numbers of larvae of Stegomyia calopus Meigen (=Aëdes aegypti Linnaeus) in the laboratory and under natural conditions.

Anisopterid nymphs must destroy a great many larvae and pupae of mosquitoes in nature. Nevertheless, larvae, particularly those of Anopheles, were often found in abundance at Palmalmal in small pools inhabited by dragonfly nymphs; and the latter consumed few anophelines as compared with culicines in the laboratory experiments. Obviously, as Hearle (1926) points out in regard to his work in British Columbia, the value of dragonfly nymphs as mosquito enemies is limited by their bottom-feeding habits.