The breeding sites used by three species of the Anopheles gambiae Giles complex were investigated in northern Natal, South Africa. Those used by the two freshwater species, A. arabiensis Patton and A. quadriannulatus (Theobald) were similar, immature stages of both being collected from the same pool at five out of seven localities from which A. arabiensis was recorded. A. quadriannulatus was the most extensively distributed species of the complex and was found in association with A. arabiensis at only five of 49 localities. The difficulty in locating the breeding sites of A. arabiensis is a product of their low density, presumably as a result of the intra-domiciliary, residual insecticide spray programme. A. merus Dönitz larvae and pupae were recorded only in water with a salinity greater than 5 p.p.t. The pH, dissolved oxygen, pool size, turbidity, shade and association with vegetation of water bodies containing A. gambiae s.l. were also recorded and showed no difference between the three species. The importance of distinguishing between winter and summer breeding sites is discussed, as well as the need for entomological consultation prior to agricultural development in endemic malarious areas.

Multivariate analysis was used to demonstrate that specimens of Planococcus collected from yams (Dioscorea spp.) in both the West Indies and Africa can be differentiated from the similar P. ficus (Signoret) which occurs on figs, grapevines and pomegranates in the Mediterranean Basin, Pakistan, South Africa and Argentina. The specimens from yams are considered here to be a distinct species, to which the name P. halli Ezzat & McConnell is applied. Examination of material of both species reared on potato tubers at various temperatures showed that the differentiation is not a host-induced effect. Independent canonical variates analysis of the field-collected and reared material each showed clear differentiation of the two groups. However, the coefficients obtained for each character differed markedly between the analysis of the field-collected material and that of the reared material, and thus could not be used to assign individuals to groups. This difference in the values of the coefficients is considered to have been due to the peculiarities of individual populations, as the field-collected material included many genotypes while only three genotypes were represented by the reared material. A key and illustrations are provided for the separation of the two species.

Four models are presented which represent the four combinations of two trap types (food-baited or pheromone-baited), each containing either insecticide or chemosterilant. These were then modelled in combination to determine which combinations might be useful for controlling arthropod pest species in which the virgin females produce sex pheromone which attracts males. The most promising results occurred in combinations involving different trap baits but the same pesticide type, i.e. where one trap contains food bait while the other contains sex pheromone, and both contain either insecticide or chemosterilant. This conforms to the principle of complementarity of control methods and at the same time avoids the interference which would occur if insecticide was used in one trap type and sterilant in the other.

The importance of colour in the attraction of Glossina palpalis palpalis (Robineau-Desvoidy) to trapping devices was investigated in the Ivory Coast. Biconical traps were tested with the normally royal blue lower cone replaced with one of 25 differently coloured cloths. Royal blues were consistently the best colours for traps; yellow and orange were the worst, catching less than 1% as many as the best royal blue. Modelling trap data from the spectral reflectivities of the colours used showed that blue wavelengths contributed positively, and ultraviolet and green-yellow-red negatively, to trap performance. The responses to screens were studied using electric nets, deployed to catch flies both circling around and landing on the screen. The overall catch (of circling and landing flies) was greatest for royal blue screens, intermediate for yellow and green ones, and lowest for screens highly reflective of ultraviolet light; the total range of variation between colours was much smaller than was the case with traps. The relationship of catch to spectral reflectivity was similar to that described for trap score. An achromatic series of screens (black, white and greys) all attracted significantly fewer females than did royal blue screens, implying that attraction to blue depends on colour discrimination and not on intensity contrasts alone. For most colours, the majority of flies circled the screen without landing; with materials strongly reflecting ultraviolet, however, a high proportion of flies (especially females) landed.

Epidinocarsis lopezi (De Santis) was found to parasitize and complete development in all developmental instars of its host, Phenacoccus manihoti Matile-Ferrero. However, parasitoid mortality was high (15·8%) when development took place in the first nymphal instar of the host. Complete development from egg to adult emergence was prolonged in smaller hosts, and developmental periods recorded were 18·8, 17·4, 16·1 and 14·9 days for the first, second, third and fourth nymphal instars, respectively. Irrespective of the age of the host parasitized (with which size is strongly correlated), E. lopezi completed only one life-cycle during one life-cycle of P. manihoti. Oviposition commenced within 24 h of emergence and lasted effectively for six days, during which 95% of its eggs were laid and 10·8 large hosts were killed through host feeding. Unmated females of E. lopezi are arrhenotokous, laying only male eggs; eggs laid by mated females produced males and females in the ratio of 1:3 in favour of females.

The results of field trials in the UK are reported, investigating the hypothesis that the disruption of proleg clasping in lepidopteran larvae by sublethal doses of the formamidine pesticide chlordimeform contributes significantly to a control mechanism. Pieris brassicae (L.) was used as a model pest. Topically dosed third-instar larvae, placed on cabbage plants, became excited and many fell. Although the behavioural symptoms were reversible within 24 h, fallen larvae rarely re-established themselves on the plants. The ED50 halving the initial population was similar to the ED50 for proleg disruption, emphasizing the importance of the latter as a major cause of the former. Arthropod or bird predation was not shown to increase following chlordimeform treatment. However, inclement weather did enhance the detaching effect of treatment and also its delayed effects on larval development and survival.

The rate at which gregarious nymphs of Schistocerca gregaria (Forskål) lose the tendency to form groups after different periods of isolation was measured. There was a marked loss of social grouping after isolation for 24 h. The fall in level of grouping was highly significantly correlated with the number of days the nymphs had been isolated. Further experiments showed that the extent of loss of grouping behaviour could be significantly modified by the length of time the nymphs spent in crowded conditions before they were isolated, but the density during previous crowding did not have a consistent effect on reduction in grouping. It had been previously considered that S. gregaria nymphs lose very little of their gregarious tendency once it is acquired; this is now shown not to be the case. The behaviour of this species is very responsive to the social environment, whether the locusts are learning to form groups or are losing that tendency.

Seasonal patterns of abundance of Musca vetustissima Walker and scarabaeine dung beetles were measured at Busselton in south-western Australia for two years before and three years after, the introduction and establishment of two species of summer-active dung beetles, Oniticellus pallipes (F.) and Onthophagus binodis (Thunberg). Previously, only a single winter-active native dung beetle species, O. ferox Harold, occurred in this area. The mean dry weight of dung beetles trapped in December (early summer) was unaffected by the establishment of the additional species, and fly populations remained high. In January, however, dung beetle dry weight increased more than 60-fold and M. vetustissima abundance fell to 12% of its previous value, essentially halving the duration of major fly problems in the area. Virtually no flies emerged from cattle dung after December, when dung beetle activity reduced egg-adult survival of M. vetustissima to only 0·5%. It is proposed that better M. vetustissima control could be achieved in much of south-western Australia by introducing additional dung beetle species that are abundant in November and December.

The development and testing of a bioassay system to evaluate the potency of floral odours as mediators of long-range floral nectar source location by mosquitoes are described. The bioassay is quantitative and behaviourally discriminating, upwind flying and landing acting as indices of behavioural activity over the 24-h light:dark cycle. In initial tests, the responses of Aedes aegypti (L.) to the flowers and floral odours of ox-eye daisy (Leucanthemum vulgare) were investigated. A. aegypti exhibited a biphasic diel cycle of nectar-feeding behaviour and landed on modified flowers that retained either their disc or ray florets. Most significantly, the mosquitoes responded in a similar way to the odour of L. vulgare in the absence of visual stimuli or nutritional cues. They did not respond to solvent extracts of ox-eye daisy flowers.

A wind-tunnel bioassay system, developed previously for the evaluation of long-range host-location stimuli, was used to investigate the responses of Anopheles arabiensis Patton to floral odours. Mosquitoes were found to respond by flight and landing to Achillea millefolium inflorescences and also to the odour of A. millefolium in the absence of visual stimuli. An air-entrainment technique was used to extract floral odours for biological and chemical testing. Both sexes of Anopheles arabiensis were found to respond to the floral odour extract. The major component of the floral odour was tentatively identified as a cyclic or bicyclic monoterpene.

The reproductive phenology of Listronotus bonariensis (Kuschel) in the north of North Island, New Zealand, was studied by dissection of adults collected at intervals from several sites. First-generation adults exhibited reproductive development and activity, followed by gonad atrophication and death. Most second-generation adults entered directly into oligopause in the autumn. Some early emergent weevils of this generation oviposited before onset of winter and reproductive dormancy. Oligopause was not intense, and reproductive maturity was gradually attained over the winter, with resumption of egg laying in early spring. These results are discussed in relation to population reproductive biology and survival.

In laboratory and field trials in the UK, larvae of Pieris rapae (L.) from the UK survived better than larvae from the Mediterranean region on Brussels sprouts, broccoli and field kale. A similar difference, associated with latitude, was found in Australia. The implications of these results are discussed.

Populations of Nysius vinitor Bergroth were studied from 1979 to 1982 in two weed hosts, Arctotheca calendula and Polygonum aviculare, and eight irrigated sunflower crops in a summer cropping area of northern Victoria, Australia. The spring generation began with the adults colonizing flowering A.calendula plants in September and concluded with the rapid development of late stage nymphs and an exodus of adults from these plants from mid-November to December. Gradual invasion of sunflowers occurred mostly in late December and reached a peak at flowering, after which nymphs appeared. P. aviculare attracted adults from February and hosted a number of overlapping generations until winter. The weed sustained diminishing numbers of adults through the winter, except in 1982, when a further generation produced an early spring peak. Immigrant populations were regarded as a common source of adults for initiating the spring and summer generations. The rate of development of N. vinitor in spring was more rapid than that predicted by phenological simulation based on ambient temperatures and laboratory-derived day-degree estimates. This was attributed to increased body temperatures through absorption of solar radiation, and the simulation model was adjusted by increasing daily minimum and maximum temperatures by 1·3 and 5·5°C for young and older instars, respectively. This suggested that older nymphs have lower developmental thresholds or are better able to optimize body temperatures.

A series of host-transfer trials using both laboratory-cultured and field-collected individuals of the aphid parasitoids Aphidius ervi Haliday and A. rhopalosiphi De Stefani Perez were done in order to clarify inconsistent results from several previous studies. A. ervi cultured on Acyrthosiphon pisum (Harris) produced very few mummies when confined with Microlophium carnosum (Buckton), whereas those cultured on M. carnosum produced as many mummies on A. pisum as they did on their original host. Mummy production was correlated with the attack rate of adult parasitoids on potential hosts. The production of mummies on M. carnosum by parasitoids reared on A. pisum was often greatly improved if their male parent had been reared on M. carnosum, suggesting that genotype strongly influences host preference. Aphidius rhopalosiphi from laboratory cultures produced significantly more mummies on Metopolophium dirhodum (Walker) than on Sitobion avenae (F.) regardless of their original host, but this preference was not shown by parasitoids from field populations. It is concluded that inadvertent selection occurs in laboratory cultures of aphid parasitoids as a result of low founder numbers, genetic drift and genetic bottlenecks and that this can strongly influence experimental results in biology and behaviour studies, which has important implications for biological control workers.

Rastrococcus invadens Williams has been the most important polyphagous pest of horticultural crops since 1982 in some West African countries. It originated in South-East Asia and was probably introduced on infested plant material. The pest has been reported up to 150 km north of the coast of Benin and 500 km north from the coast in Togo. Although the species is polyphagous, mango, citrus, breadfruit, banana, frangipani (Plumeria alba) and species of Ficus are among the most attacked hosts. Several indigenous natural enemies were identified, but their ability to regulate the populations of R. invadens is very low. The most important among the predators were the coccinellids Chilocorus nigrita (F.), Exochomus promtus Weise and E. troberti Mulsant and the lycaenids Spalgis spp. In Togo, an indigenous parasitoid, Anagyrus sp. ?nr aurantifrons Compere, has become adapted to the pest.

Rastrococcus invadens Williams, a pest of fruit and ornamental trees in West Africa, was introduced from the Oriental Region. In 1986, a search was initiated in India for natural enemies of Rastrococcus spp. with a view to selecting potential biocontrol agents for introduction into West Africa. The results of the survey, laboratory breeding of the hosts and their natural enemies and host-specificity of the natural enemies are reported. Based on host-acceptance tests, two encyrtids, Gyranusoidea tebygi Noyes and Anagyrus sp., have been sent to UK for further screening.

Rastrococcus invadens Williams, a serious pest of fruit crops in West Africa was studied under controlled conditions. The different stages of the mealybug are described and data given on development times, adult lifespan, fecundity, sex ratio and mortality rates. Similar data are also given for a primary parasitoid, Gyranusoidea tebygi Noyes. The parasitoid completed almost two generations for each one of the mealybug and had a female-biased sex ratio compared with a male-biased one for R. invadens. The parasitoid did not parasitize other species of mealybugs offered. G. tebygi was considered a suitable candidate for a biological control programme and was introduced into Togo in October 1987.

The frequency of DDT resistance in Anopheles culicifacies Giles in Sri Lanka remained relatively constant from the cessation of DDT use for malaria control in 1975 until 1982. From 1983 to 1986, the frequency of resistance decreased slowly but significantly. In contrast, the resistance gene frequency in A. subpictus Grassi declined between 1980 and 1983, after which the trend was reversed, and there was a significant increase in DDT resistance in 1984–86. This rise in DDT resistance cannot be attributed to positive DDT selection pressure as there has been little or no agricultural use of DDT during this period. It is also unlikely that the rise in resistance is due to pyrethroid selection as the quantity of pyrethroids used on the island is low and there is no evidence of pyrethroid resistance from WHO susceptibility tests. Studies with the synergists FDMC (4-chloro-α-(4-chlorophenyl)-α-(trifluoromethyl)benzenemethanol) and piperonyl butoxide indicated the involvement of DDT-dehydrochlorinase rather than oxidase in the resistance. This was confirmed by metabolism studies, as DDT-resistant adults of both A. culicifacies and A. subpictus showed an increase in the rate of DDE production, which was also correlated with an increase in glutathione transferase activity.