Summary 1.Between 1997 and 2000 a plague of migratory locust in Madagascar triggered largescale control operations. The widespread use of insecticides in a biodiversity hot-spot raised concern about hazards to wildlife. The aim of this study was to inform locust control policy by monitoring the effects of insecticides on key non-target organisms. The focus was on the harvester termite Coarctotermes clepsydra and its vertebrate predators. 2. Effects on Coarctotermes clepsydra were studied in two areas. At Ankazoabo in 1998, we monitored rainy season barrier treatments with fipronil and triflumuron. At Malaimbandy in 2000, we monitored early dry season, full-cover sprays with fipronil and deltamethrin and also studied effects on the lizards Chalarodon madagascariensis and Mabuya elegans , the lesser hedgehog tenrec Echinops telfairi and their non-termite arthropod prey. 3. At both sites, fipronil caused a strong reduction of the activity of Coarctotermes clepsydra colonies, culminating in high mortality. At Ankazoabo 10 months post-spray, colony mortality was 44·9%, 7·7% and 4·7%, respectively, in fipronil, triflumuron and unsprayed sites. Mortality within spray barriers was 90·7% and 10·8% for fipronil and triflumuron, respectively. Similar results were obtained at Malaimbandy 6 months post-spray, where fipronil caused 80·5% colony mortality, compared with 3·5% and 3·7% in unsprayed and deltamethrin plots. 4. We observed a significant decline in the relative abundance of Chalarodon madagascariensis and Mabuya elegans in fipronil plots at Malaimbandy, with effective reductions (compared with pre-spray levels) of 52·7% and 45·2%, respectively. Echinops telfairi was not found in fipronil plots, while being frequent in unsprayed and deltamethrin plots. 5. Termites proved to be important dietary components of all vertebrates studied, especially during the dry season. Echinops telfairi , the large-eared tenrec Geogale aurita and Mabuya elegans were more termite-specialized than Chalarodon madagascariensis . 6. The abundance of lizards and Echinops telfairi was positively correlated with the density of live Coarctotermes clepsydra colonies. This suggests that reduced abundance of these insectivorous species in fipronil plots resulted from deprivation of their principal prey. 7. Synthesis and applications . The present study provides evidence of food chain perturbations induced by fipronil. This has important implications for the management of locusts in Madagascar and other locust-affected countries. In recognition of the precautionary principle, we recommend that fipronil be withheld from full-cover spraying Correspondence: Ralf Peveling,

Curdioderma hangs from habitual perches to wait for mainly terrestrial prey. Their broad wings provide the lift at low speeds necessary to capture such prey from the ground. Male‐female pairs were found in the long dry season, a period when reduced insect availability led to lower body weights and males used song to delineate their feeding territories. Males moved between perches more frequently during the first half‐hour of nightly singing. An enlargement of foraging area was associated with earlier singing, as well as an increase in song rate, movement between perches and pair contact‐vocalizations. Females are larger than males and the sex ratio favoured the latter. There was evidence of long‐term territories and breeding throughout the year. These results are discussed in relation to pair‐bonding and environmental seasonality.

The social organisation of the polyoestrous bat, Tadarida pumila, was studied in northern Ghana (W. Africa). Although this small (8.5 g) free‐tailed bat is not sexually dimorphic in size, adult males develop an odorous inter‐aural crest of long hairs. A social system based on female defense polygyny was found which, however, also involved elements of resource defense. Harems, each generally established in the roof space of separate dwelling houses with restricted access, consisted of an adult male and up to 21 females, with their young, and the larger harems were held by heavier males. Harem composition was stable and both harem males and females showed high site fidelity over the study span of 16 months. Some female young were recruited to their natal harems, at a sufficient rate to replace the annual loss of harem females. However, most young, which were born in three successive cohorts during the rainy season, apparently dispersed over the dry season, following early sexual maturation. The potential causes and benefits of female associations are discussed.

Spraying with the organophosphate fenthion has been the predominant means to control the red-billed quelea (Quelea quelea), a major bird pest throughout semiarid regions of sub-Saharan Africa, for more than forty years. A review of known effects of fenthion and other control measures used against Q. quelea on nontarget organisms, provides a basis for comparisons with any alternative chemical control agents. Birds of prey, owls and passerines have been commonly reported casualties of spraying with fenthion over land. Moreover, fenthion is known to have negative effects on aquatic invertebrates, in particular on populations of crustacea, which predicates against its use near water bodies. Non-target species may be affected directly by spraying, but predatory birds, scavenging birds and even mammals can be contaminated by secondary poisoning when they eat Quelea carcasses found up to 20 km or more from the primary control site. To avoid secondary poisoning, where possible quelea carcasses should be removed from a site after spraying. The use of explosions to kill roosting quelea birds has similar effects to those of fenthion in terrestrial habitats, killing non-target species including birds of prey and owls. Another control method, harvesting quelea as a source of protein, is benign and could contribute to local nutritional and economic needs. However harvesting is unlikely to reduce quelea populations substantially. Integrated pest management (IPM) approaches are recommended to minimize environmental damage, but until these are successfully adopted, standardized procedures for comparative assessments of the effects of quelea control are required.