In shaded, unsprayed arabica coffee on Kilimangaro, fluctuations of numbers of the leafminers Leucoptera meyricki Ghesq. and L. cafeina Washb. (Lepidoptera: Lyonetiidae) follow a fairly regular seasonal cycle. Analysing the daily catches of adult moths taken in a suction trap, Bigger & Tapley (1969) have shown that the curve of numbers per generation plotted against time is cyclical, a basic period of 1 year being overlain by a 2-year cycle. This results in a deeper or shallower trough in alternate years and with L. caffeina a deep trough is followed by a low peak followed by a shallow trough followed by a high peak. With L. meyricki, which is now the dominant leaf-miner species, the peaks do not alternate regularly. Bigger & Tapley (1969) also showed that if linear regression lines were fitted to the periods of increase and decrease, the slopes of the lines (on a logarithmic scale) were highly correlated with the size of the generation at the base of the preceeding trough. If the trough was deep, the rate of increase was high and the subsequent rate of decrease low; if the trough was shallow then the rate of increase was slower and the subsequent rate of decrease faster. It was postulated that the regulatory mechanism must be relatively simple to give such regular fluctuations and that parasites were probably the main factor involved. In this paper the interactions of the two host species and their complex of larval parasites is considered.