A range of materials were collected from a number of ecological and environmental zones within Kenya as being likely sources of entomopathogenic bacteria. These were soils, insect frass and dead insect material. Using an isolation protocol involving use of a selective medium, over 150 Bacillus thuringiensis (B. t.) strains were recovered from the sample material. Their identity as B. t. strains was confirmed by their growth characteristics, morphology and presence of a parasporal delta-endotoxin crystal.

All strains were screened for activity against two Lepidopteran pest species, the spotted stem borer, Chilo partellus and the African army worm, Spodoptera exempta. Isolates causing over 80% mortality in the screening assays were retained for further evaluation. The selected isolates were subjected to bioassay against C. partellus and S. littoralis. Following bioassay, the most toxic isolates were selected for screenhouse trials against C. partellus and field trials against a natural outbreak of S. exempta.

In the screenhouse trials all of the B. t. treated plants showed reduced levels of damage and yields were 5 to 7 times higher than the yield obtained from the untreated, infested control plots. Three isolates, A-3, A-C-2 (isolated from insect material) and M-44–2 (isolated from soil) seemed to offer superior levels of protection. Two new B. t. isolates, K-26–21 and MF-4B-2, both isolated from soils, showed high levels of toxicity to S. littoralis and S. exempta in the laboratory. When isolate K-26–21 was applied to maize seedlings in the field, almost total control of the larval population was achieved within 48 hr at each of the concentration levels tested.