1. The area (Fig. 1) may be divided into a forested uninhabited mountain region, a region of low hillocks with between them marshy bottoms, many of which had been cleared and cultivated, and a flat coastal plain protected from the sea by a wall which leaked. Rain fell from early June till late September, and the rest of the year was mainly rainless.

2. Little is known about the mountain region. A. minimus was found infected in April, at which time, however, most potential breeding-places were dry. Larvae of A. maculatus were abundant, but the adults did not bite man.

3. Malaria was intense locally in the hillocks and coastal plain region, but uneven in distribution (Fig. 1). Evidence of active transmission in March and April, during the rains, and in October was found, there was a lull in May and June, but any other lulls there may have been were not discovered, since observations through the year were not possible.

4. Observations at night revealed A. leucosphyrus, A. jeyporiensis and A. minimus near man and rarely near cattle. A. sundaicus, A. philippinensis, A. annularis and A. hyrcanus attacked both man and cattle, though some, at least, would attack man only in the absence of cattle. A. aconitus was found near man at night, but nearly all the specimens contained old blood.

5. A. subpictus and A. vagus fed on cattle and entered houses at dawn in search of shelter for the day. The man-biters mainly sought shelter in some outdoor resting-place. The conclusion was that the standard practice of searching for mosquitoes in houses early in the morning gave results which were not only valueless, but misleading (Fig. 2).

6. Few dissections were made and sporozoites were found only in A. jeyporiensis (2/126= 1·6%), which was thought to be responsible for most of the malaria transmitted early in the year. It is suggested that the results of night catching provide valuable circumstantial evidence about vectors.

7. The night catches also gave information about time of biting, and yielded two observations of importance in control work. A. minimus, A. jeyporiensis, A. philippinensis and A. sundaicus did not attack until several hours after nightfall. A. philippinensis, in late May, when it was not raining, preferred to bite persons sleeping out of doors. Elsewhere it was observed that A. philippinensis did this and also did not rest nearby before or after feeding as other species did. It might, therefore, avoid a contact insecticide. A. sundaicus, baulked of a feed by night, attacked after sunrise.