A survey was conducted in the four regions of Tanga, Morogoro, Coast and Dar es Salaam in eastern Tanzania to determine the distribution and species composition of cereal stemborers and their natural enemies. A total of 22 fields were sampled in Tanga, 28 in Morogoro, 17 in Coast and one in Dar es Salaam. Chilo partellus was found to be the predominant species, accounting for 80 % of all stemborers collected. Busseola fusca, Sesamia calamistis and Chilo orichalcociliellus made up 15,4 and less than 1 % of the stemborers collected respectively. Infestation levels varied with the age of the plants, whereby younger plants were more severely infested (infestation ranging from 20–40 %) than mature plants (infestation ranging from 5–15 %). Two larval parasitoids, Cotesia sesamiae and Cotesia flavipes, were the main natural enemies collected. However, the levels of parasitism of the stemborers were very low. The exotic natural enemy Co. flavipes may have spread from the Kenya coast (where it is established) into eastern Tanzania.

Difficulties of identification of natural enemies of cereal stemborers for ‘redistribution’ in Africa are discussed. Tritrophic level studies on wild and cultivated habitats of borers and beneficial species are needed to judge the impact of a natural enemy species in the cropping system. Areas with low pest pressure which are climatically favorable for pest development may be selected for study. Based on the results of various countrywide surveys to map the relative importance of Sesamia calamistis Hampson, Eldana saccharina (Walker) and Busseola frisca (Fuller) in western Africa, recommendations are given for sites for tritrophic level studies. It is hypothesised that because maize is not always present in the field and because of its high susceptibility, natural biological control has to come from wild habitats. This emphasises the importance of the knowledge on the wild host-plant range. Survey results complemented with oviposition and life-table studies in the laboratory showed that, rather than being reservoirs for pests, most wild grass species act as trap plants causing mortalities of 100%. A comparison of light trap catches with pupae found on wild hosts and the scarcity of known wild hosts in areas with high pest pressure suggest gaps in our knowledge of the range of host plant species. Based on comparison of known natural enemy complexes in East and West Africa, the scelionid egg parasitoid Telenomus isis and an East African strain of the braconid larval parasitoid Cotesia sesamiae (Cameron) are proposed for redistribution against B. fusca and S. calamistis, respectively. Telenomus isis has never been reported from East Africa whereas G sesamiae is common in East and southern Africa and scarce in western Africa, suggesting that G sesamiae is probably not adapted to S. calamistis and B. fusca in this region.