Following the partial clearance of farm scrub some termite species became locally extinct or scarcer, some were not affected, whilst others were adversely affected but became re-established and increased their activity. Wood-feeding termites in the moist semi-deciduous forest zone of Ghana consume about 0.2 kg m-2 yr-1 of woody material. This weight accounts for about 20 percent of the litter fall or about 8 to 10 percent of the annual primary production of the deciduous forest and a rather greater proportion of the lower production of more open sites. There is a spatial heterogeneity in the species distribution. Although water, clay, silt and sand contents of soil, and temperature, sunshine, and sunflecks were measured, the only significant correlations are of these parameters with Microtermes subhyalinus, which is more frequent in the moister areas with higher silt and lower sand content, and also shows a temperature preference. The lack of significance in correlation and principal component analyses is attributed to the observations being made away from the unknown location of the underground nest, and the ability of termites partially to modify their environment. The several coexisting wood-feeding species show slight micro-environment preferences leading to some spatial separation, and also show a series of size differences. The species show different responses to potential food resources, some apparently preferentially attacking wood in which fungal decomposition has begun and others not. TERMITES (Isoptera) are an important component of the fauna of tropical soils. Their efficiency in utilizing many food sources has meant that some species have become pests of crops or timber, and a few species have attacked a great variety of manmade materials. Because of their economic importance, the majority of studies of tropical termites have concentrated on the pest species and their control. The species known to attack crops have been reviewed by Harris (1961, 1969), whilst Sands (1960) has reviewed the species destroying timber and crops in West Africa. Further, Harris (1970) has listed all the species known to attack structural timbers, whilst Small et al. (1960) have described the effect of termites upon West African timbers. There have been relatively few studies on the populations of tropical termites. The relations between termites and soil have been reviewed by Lee and Wood (1971), though much of their detailed work was undertaken in Australia where the Macrotermitinae, the termites with perhaps the greatest population density in Africa, are absent. A general review of African termites is given by Bouillon (1970). Sands (1965b) attempted to sample subterranean termite populations during studies on the effect of human interference with vegetation communities in the West African savanna. However, later Sands (1972) drew attention to the extreme difficulty of obtaining estimates of the population density of truly subterranean, as opposed to moundbuilding, termite species. The work described here gives an indirect estimate of the population size and impact of a woodfeeding termite community, the majority of whose members are subterranean. These studies were undertaken during the development of a termite test site, mainly designed for conducting 'graveyard'-type tests. The Ghanaian Council for Scientific and Industrial Research wished the site to be developcd in an area of farm scrub near the village of Fumesua, about 10 km east of Kumasi (6?40'N, 1?35'W), in the Ashanti Region of Ghana. Since no previous work has been undertaken on the termite fauna of the area, the present study was designed to cover the first year of the creation of the test site, prior to any tests being carried out. It was thought important to know the ecological characteristics of the termites potentially able to attack the test samples, and the community structure of the termites that might be killed by chemical treatments of soil or wood.