apparent preference for culicine larvae as food is brought about by the head-downwards attitude assumed by the diving beetle when near the water surface. In this attitude it can detect larvae of Culex, but not those of the surface-dwelling Anopheles. Although anopheline larvae are normally protected from the attacks of H. litigiosus by virtue of their surface-dwelling habit, they are seized by the predator without discrimination when caused to leave the water surface.

Two dytiscids, a male and a female, were kept in a vessel containing 8 cm. (250 cc.) of boiled and filtered water. The water in this container was changed each week, but no animals other than larvae of A. farauti were admitted. Any anopheline pupae were removed as soon as their presence was noticed, and they were replaced with early instar larvae of the same species. At the end of three weeks the female Hydaticus died, its body being promptly eviscerated by the male. During all this time only thirty-five of the anopheline larvae in the beaker had been consumed by both beetles. After a further week fifty larvae of C. pullus were introduced into the beaker. The surviving water beetle immediately began to attack these larvae, eating eleven within the first five minutes. All the larvae had been consumed at the end of an hour.

It was found that an average of 11.4 pupae were consumed by a single Hydaticus adult each day. Like the predators already discussed, this diving beetle was often seen to be eluded by pupae, although larvae were seldom observed to avoid capture.

In feeding, H. litigiosus captures its prey with a sudden lunge of its forelegs, then rises slowly to the surface while guiding the victim to its mouthparts with its fore and middle legs. Usually both larvae and pupae are eaten bodily, but sometimes the more heavily chitinized parts are discarded.

Hargreaves (1932) records that an Italian dytiscid devoured larvae and pupae of all mosquito species fed to it, at the rate of twenty a day. Howard (1912) quotes Galli-Valerio and Rochaz de Jongh as stating that an adult of Dytiscus marginalis Linnaeus, although previously fed with meat, devoured great numbers of mosquito larvae and pupae. No observations on the predacity of water beetle larvae were made at Palmalmal, although many authors have referred to the voracity of dytiscid and hydrophilid larvae as mosquito destroyers.

Anopheline larvae were often present in large numbers in pools inhabited by H. litigiosus, although species of Culex were seldom found in any numbers in association with the beetle. From this circumstance, and the results of the laboratory-feeding experiments outlined above, it appears that while H. litigiosus plays a significant part in the natural control of culieine mosquito breeding at Palmalmal, this beetle is of little importance as a factor in anopheline control.

Copelatus sp. (Text-fig. 9.)

This little dytiscid is a new species as yet undescribed, but for which Balfour-Browne proposes the name of Copelatus lairdii. It was commonly encountered in permanent pools throughout the Jacquinot Bay area. The results of feeding experiments with imagines of this water beetle are tabulated below.