1. In 1927 a suggestion was made by P. A. Buxton to introduce Megarhinus to Samoa and Fiji for the control of Aëdes scutellaris (variegatus), which transmits human filariasis.

2. In 1929, C. E. Pemberton introduced M. inornatus from New Britain into Hawaii, where it has apparently failed to become established.

3. In the same year the writer was sent to Java to obtain parasites for a coconut pest in Fiji; and in 1930 received instructions from the Fiji Government to collect and transport Megarhinus from Java to Fiji.

4. During the latter half of 1930 early stages of Megarhinus splendens—the commonest and most widespread species in Java—were collected from tree-holes, bamboo stumps and tubs at Buitenzorg.

5. Experiments were conducted at Buitenzorg to ascertain how long the larvae could be kept in a healthy condition by restricting their food supply. It was found that they matured normally on a reduced diet after a development lasting as long as two months. The most rapid development from egg to adult was found to take 24 days. Attempts to breed the adults in captivity were unsuccessful.

6. In January 1931, a colony of M. splendens consisting of 283 larvae was taken to Fiji; 45 died or emerged as adults during the voyage of 28 days.

7. On 27th February 1931, 238 M. splendens were landed at Suva.

8. Colonies for distribution to various parts of Fiji were bred in Suva during 1931 and 1932, and by April of the latter year 51 colonies, composed of nearly 3,000 individuals, had been liberated.

9. Megarhinus was bred at Suva by liberating the pupae in barrels kept out of doors. The adults which emerged laid most of their eggs in the barrels in the absence of other breeding-places in the vicinity. The larvae were collected from the barrels when they reached their third instar. They were fed in captivity until they pupated, and then liberated as pupae in colonies sent to various parts of the group.

10. By the end of 1931 Megarhinus had been recovered at six of the colony sites in natural breeding-places, and is now well established. The most favourable type of country in Fiji for Megarhinus to colonise is the swamp land with trees of Inocarpus edulis (“ivi”), which occurs near the coast on many islands.

11. An account is given of the spread of Megarhinus through an Inocarpus swamp on the island of Taveuni, where a survey was made of the Megarhinus and Aëdes content of all the tree-holes within six feet of the ground. M. splendens was present in rather more than 50 per cent, of the holes and Aëdes variegatus in about 40 per cent. The probability is that in this locality Megarhinus has caused a 20 per cent, reduction in the number of tree-holes containing Aëdes during the three months succeeding its liberation.

12. It was found that in most places, in the forest especially, Megarhinus spreads very slowly–undergrowth making an effective barrier to its flight.

13. A. variegatus breeds mostly in tree-holes, tins, holes in rocks, coconut husks, wooden gongs in native villages and crab-holes (the last not yet proved in Fiji). Of these places Megarhinus has so far bred only in tree-holes and tins (very occasionally). It will probably breed in native gongs but unlikely that it will breed in either crab-holes, rock-pools or coconut husks. Therefore at the best it can hardly be expected to reduce the numbers of A. variegatus in Fiji by more than 5 per cent. But even that would justify the trifling expenditure which has been devoted to the project.

14. Studies of the habits and life-cycle of M. splendens made in Java and Fiji revealed the following points of interest:—Incubation period always about 2 days; larval stages, minimum 16 days (food abundant), maximum 134 days (food scarce); pupal stage always about 6 days. The minimum period for a generation is about 30 days.

In spite of the lower average temperature in Fiji very little difference was noted between the rate of development in that country and in Java.

The adults are diurnal in habit. They mate when settled; not on the wing, as do most mosquitos. The eggs are dropped singly on water where they float on the surface film. They are almost round and glossy white. Oviposition continues over several weeks, and is most prolific in sunny weather. The egg capacity was not determined.

Larvae of Megarhinus fed on tadpoles, Tipulid larvae, Chironomus larvae, and all sorts of mosquito larvae. They do not object to feeding on those mosquito larvae whose bodies are very hairy. They are cannibalistic.

15. In May 1933 a small colony of M. inornatus was brought by the writer from Rabaul to Fiji, since it was reported that this species had been found in New Britain breeding in coconut husks. This species is now being bred in Suva, whence, if it survives, it will be distributed to other parts of the group. Its immature stages are of similar duration to those of M. splendens, so that it would seem very unlikely that this species will succeed in reaching maturity in coconut husks in Fiji, complete evaporation of the water taking place too quickly.

M. inornatus is structurally very close to M. splendens, and it is quite possible that the two species will eventually hybridise in Fiji.