For this study, a sampled dataset from the original METEOSAT images was taken over a seven-month period for areas with a variety of different surface types and geographical locations. This was the initial BI dataset requested for the International Satellite Cloud Climatology Project (ISCCP). Cloud amounts were derived from these raw radiances from a scheme which utilized all three METEOSAT channels. For the daylight slots the visible channel was used to discriminate between low cloud and surface. In addition, a “spatial coherence technique” was employed to detect cloud, relying on the assumption that the cloud tops do not have a uniform temperature over small distances (∼20 km). The monthly mean cloudiness and mean diurnal variability for three months (April, July and October 1983) computed from the radiance data are described here. Significant seasonal variations in cloudiness were observed, such as the latitudinal movement of the ITCZ and the enhancement of the southern subtropical jet during July. The diurnal cycle of cloudiness was observed over equatorial Africa, particularly for the high cloud coverage.