NATURAL CONTROL OF HELICOVERPA-ARMIGERA IN SUNFLOWER - ASSESSMENT OF THE ROLE OF PREDATION

In three field experiments in Kenya, the seasonal population tl end of Helicoverpa armigera (Hubner) (= Heliothis armigera) in sunflower was followed in plots from which predators were excluded and in plots fro m which predators were not excluded. In experiment 1, complete exclusi on of crawling predators (predominantly Pheidole spp. of ants occurrin g at densities of 25 per plant) resulted in H. armigera densities 3-5 times higher than in plots where ants were not excluded. Pheidole sp. had more impact on young H. armigera larvae (instars 2-3) than on olde r larvae (instars 4-6). The results of experiment 2 were less striking , because H. armigera infestation was low, and ant densities were mode rate. Here, Myrmicaria spp. and Camponotus spp. were the predominant a nts. Exclusion of ants resulted in a 1.8-fold increase in densities of large H. armigera instals. In experiment 3, the impact of predators s on H. armigera,vas studied under three conditions: exclusion of crawli ng predators, exclusion of both crawling and flying predators, and the control where no predators were excluded. To evaluate the role of pre dation in total mortality, the recruitment of H. armigera larvae was d etermined with Southwood and Jepson's graphical method, and recruitmen t of newly laid eggs was directly measured on trap plants. Because it was difficult to detect older eggs due to colour change, the direct me asurement of egg recruitment was superior to the graphical method. Ant s and Anthocoridae were the dominant predators. Mortality from egg to older larvae (instars 4-6) was 73-78%. Exclusion of ants and anthocori ds did not affect the densities of H. armigera larvae. Anthocorid pred ators increased only after the main oviposition peak of H. armigera an d, therefore, their exclusion had little impact on the pest. Ant densi ty was considerably lower than in experiment 1, and did not significan tly suppress H. armigera.