birds fed upon. A Drongo-cuckoo, Surniculus lugubris, had 50 caterpillars in its stomach. Healthy cocoa trees were sparsely inhabited by birds but areas that were obviously infested with insects literally swarmed with birds. Dead shade trees in the cocoa plantings also provided food for woodpeckers, with four species being found utilizing these dead trees.

Abaca.—The last of the habitat types that I recognized at Quoin Hill was a small grove of Abaca, Musa textilis, and wild bananas, Musa sp. This habitat type was frequented by spiderhunters (Arachnothera sp.) of the family Nectariniidae.

Kalabakan.—We worked at three localities in this area: 12 miles north of Kalabakan, Brantian Estate, and Kalabakan.

Primary forest.—We were fortunate in being able to work on the very edge of the primary forest 12 miles north of Kalabakan. The composition of the primary forest was much like that at Quoin Hill and will not be discussed further. About a month after we arrived at our forest camp, logging crews moved in and cut the commercial timber near our area. The only immediately noticeable difference in the makeup of the avifauna after destruction of the forest canopy was the appearance of the drongo Dicrurus aeneus. This drongo was seen in areas where the trees had been cut, sitting on limbs and darting out after insects.

Secondary forest.—The area around Kalabakan proper was in secondary forest, which was almost impossible to penetrate. At Kalabakan, Cymbirhynchus macrorhynchus, Cecropsia striolata, and Macronous ptilosus were taken and I did not see them elsewhere. Kalabakan is situated on the Kalabakan River at the upper tidal limit. The Nipa-Mangrove association, not investigated, lies immediately below Kalabakan.

The Brantian Estate area was mostly in secondary forest and was situated on the Brantian River. There were some fairly large areas of grass with water buffalo wallows in them. These grassy areas were favorite haunts of the Painted Quail, Coturnix chinensis.

Tenom.—The first locality that we investigated was 5.5 miles southwest of Tenom, approximately 4,000 feet elevation, in moss forest. A few days were spent collecting in the area of Tenom itself.

Moss forest.—The lower altitudinal limit of the moss forest was about 3,600 feet. The trees on top of the mountain were mostly oaks (Quercus) and were festooned with ferns, orchids, and other epiphytes. The area had been used as a triangulation station by a survey team and a small area on top of the mountain had been cleared earlier. At the time of our visit this small area had grown to secondary vegetation, mostly Pandanus. The oaks in the primary forest surrounding this disturbed area were generally about 30 to 50 feet high and there was little undergrowth in virgin stands. This area was usually swathed in fog from three o'clock in the afternoon until eleven o'clock the next morning. One morning of our fourteen there was clear.

Paddy.—The area visited at Tenom itself consisted mostly of old paddy grown to grass and scrub. Forest did occur but was of secondary nature in the immediate vicinity of Tenom.

Siamil Island.—This island is about one mile in circumference and the highest point is about 300 feet above sea level. The island has high bluffs on three sides but slopes gently to the sea on the other. There were patches of forest left on the island, one on the north side and one on the south. The sheer bluffs on the east side of the island were covered with Pandanus sp. The undergrowth of the north forest had been cut, leaving extensive bare areas. The principal undergrowth was rattan. The natives are clearing and planting more of the island to coconuts and hope eventually to clear it completely.

Karindingen Island.—This island, about half a mile in circumference and between 10 and 20 feet above sea level at its highest point, was surrounded by extensive coral reefs and sand; the principal vegetation was mangroves.

ECOLOGICAL AFFINITIES OF THE AVIFAUNA AT QUOIN HILL

More time was spent at Quoin Hill than at any other locality. Fifty five of the more common and hence best-known birds are listed in Table 1 together with their primary and secondary preferences of habitat. The habitat distribution of the birds shows the amount of secondary utilization of habitats by birds that occurred predominantly in one habitat. Cocoa was utilized by 6.2 per cent of the birds of the primary forest, and 88.8 per cent of birds of the secondary forest. This indicates that cocoa is an effective substitute for secondary forest for some birds. Of the species of the primary forest, 18.7 per cent occurred also in secondary forest; thus, three times as many species of primary forest utilized secondary forest as utilized cocoa. This too might be expected, since "secondary" forest is of frequent natural occurrence and an ancient feature while the comparatively simple cocoa plantings are new and artificial.

TABLE 1.—Habitat preferences of 55 Quoin Hill birds.