The effects of endosulfan and deltamethrin, chemicals used to control Glossina morsitans Westwood (tsetse fly), were tested against Cyrtobagous salviniae Calder & Sands (Curculionidae), a biological control agent for the floating fern Salvinia molesta. The weevil was very susceptible to deltamethrin and less susceptible to endosulfan. The use of these chemicals for control of G. morsitans in Botswana, where the biological control agent is active, is discussed.

The stages of bush fly Musca vetustissima Walker, killed by the dung beetles, Onthophagus binodis Thunberg and Onitis alexis Klug, at low and high densities were measured in cattle dung of low nutritive value. A high level of M. vetustissima mortality occurred in the presence of Onitis alexis and Onthophagus binodis in experiments both in the laboratory at a constant 25°C, and exposed to summer fluctuating temperatures outside. The K values for total eggpuparia mortality increased curvilinearly with increasing log10 beetle numbers. The mortality of older larvae, k3, was most closely correlated with total mortality, K, where b = 0.720 and r2 = 0.931, while a lesser but significant correlation was obtained with mortality of young larvae, k2, where b = 0.291 and r2 = 0.645, but no correlation was obtained with egg mortality, k1, where b = −0.010 and r2 = 0.058. It is predicted that in unfavourable dung which occurs during summer, mortality of older bush fly larvae will be the key mortality caused by scarabaeine dung beetles.

Two new species of mealybug found on coffee roots are described from Africa. Planococcus fungicolasp. n. from Kenya is the species that has been known in Africa as the coffee root mealybug. This mealybug is almost always associated with the fungus, Diacanthodes novoguineensis (Hennings) Fidalgo. Planococcus radicumsp. n. is from Nigeria and Tanzania. Dactylopius coffeae Newstead is synonymized with Planococcus lilacinus (Cockerell) syn. n.

A host choice experiment was carried out to determine the host range of some potential biological control agents of yellow starthistle (Centaurea solstitialis), a Mediterranean plant which has become a serious rangeland weed in western USA. This paper describes the results of that experiment and discusses its taxonomic implications. The experiment confirmed the extreme host specificity of some species. However, two reputedly polyphagous species, Acanthiophilus helianthi (Rossi) and Urophora quadrifasciata (Meigen), attacked a narrower range of hosts than expected. The results of this and similar host choice experiments are discussed and it is concluded that this sort of anomaly can be caused by the existence of host races, unrecognized sibling species, or by varieties of plants naturally resistant to attack. These complex patterns of host relationship highlight the need to interpret host catalogue data cautiously, as they oversimplify complex biological systems. The identity of one species reared in the experiment, Terellia luteola (Wiedemann), is discussed; a key is presented for its separation from other Terellia species, and it is removed from synonymy with T. colon (Meigen).

The estimation of tsetse fly mortality rates from life-table data is central to opulation dynamics studies and to the development of tsetse fly control programmes. For a population at equilibrium with a stable age distribution, the age-specific mortalities may be estimated directly from the number of individuals in each age class, but a correction must be applied when the population is growing or declining. Furthermore, if the mortality rates are changing with time, inaccuracies will be introduced into estimates of the mortality rates derived from the age structure of the population since the population will take time to reach a new stable age distribution. In this paper we use the Euler-Lotka equation, which relates the age-specific mortality and fecundity to the overall growth rate of the population, to study the loss rate of the tsetse fly Glossina pallidipes (Austen) as a function of pupal mortality, adult mortality and mortalities applied to each age class seperately. We then present a simulation model in order to quantity and to set limits on the precision of estimates of mortalities when the mortalities are themselves changing.