The use of six cowpea varieties highly resistant to Aphis craccivora Koch in grafts (as resistant scion and rootstock) with the susceptible variety Prima indicated the presence of a graft-transmissible chemical involved in the resistance. Wrapping the resistant part in light-proof foil indicated that photosynthesis is involved in the production of the chemical. In a third test, intergrafting is involved in the production of the chemical. In a third test, intergrafting the resistant part with and without leaves confirmed the importance of photosynthesis in the transfer of resistance. Finally, covering the resistance intergrafts with foil showed that the accumulation of the chemical in the stem before the foil was applied was the major source of the graft-transmissible resistance.

In studies in Massachusetts, the population density of Brevicoryne brassicae (L.) was measured over the lifespans of two groups of kale leaves from maturation to senescence, taking into consideration aphid spatial pattern (in colonies or as isolated insects) and aphid size as large (adults), medium-sized (nymphal instars 2–4) or small (first-instar nymphs) individuals. Total aphid densities per leaf in both leaf groups showed similar patterns of initial increase followed by a decrease as leaves aged over the 3–4-week period of leaf survival. Aphid densities reached peak values of 8·14 and 8·64 aphids per leaf for the two leaf groups studied, and 65–67% of all aphids observed occurred in colonies. Host and parasitoid recruitment to the aphid and the parasitoid immature populations were measured using a modification of the technique of Van Driesche & Bellows (1988). Total host recruitments per leaf were 43·7 and 64·6 aphids for the first and second leaf groups. Parasitoid recruitment was 6·8–8·1 for the first leaf group and 8·2–15·8 for the second. Recruitment values indicated 15·6–18.6% parasitism for the aphid cohort on the first leaf group and 12·7–24·4% for that on the second one.

Experiments were conducted to investigate the role of pubescence in the resistance of two wild cowpea varieties, TVNu 72 and TVNu 73, to feeding and damage caused by two cowpea pests, Maruca testulalis (Geyer) and Clavigralla tomentosicollis Stål. The fine structure of the trichomes was also studied by electron microscopy in order to try to determine the role played by trichomes in the resistance of these varieties to both pests. Feeding and development were deterred in both insects on pods of TVNu 72 and TVNu 73 with or without trichomes compared to the susceptible variety, IT84E-124. In M. testulalis, the same effect was observed on seeds and flowers. However, M. testulalis larvae fed and developed better, as indicated by the higher growth index values, when trichomes were removed. Growth index estimates showed a 13-fold decrease when trichomes were left intact on both TVNu 72 and TVNu 73 compared to the cultivated susceptible control. It was concluded that resistance to M. testulalis was based on trichomes in the first instance (as a first line of defence) and phytochemicals. Survival of C. tomentosicollis nymphs was not affected by trichome removal even though adult feeding damage was slightly increased. This is explained in line with previous reports which indicate that phytochemicals may be the principal factor in the resistance of TVNu 72 and TVNu 73 to this species. Glandular and non-glandular trichomes were found to be present on both the cultivated and wild cowpea. Trichomes in the two types of cowpea differ significantly only in trichome number (susceptible cultivated cowpeas have more) and non-glandular trichome length (those on wild cowpea are ca 20 times longer). Trichome length and angle to pod surface seemed to be more important than density per se.

A study was made of ants in Sri Lankan coconut plantations, especially of species foraging and nesting in palm crowns. At least 11 species were found nesting in the spadices, and a special study was made of interactions between some likely to be insect egg predators or to interfere with predation. Oecophylla smaragdina (F.) and Paratrechina longicornis (Latreille) were mutually exclusive, but some non-dominant species such as Monomorium spp. commonly nested and foraged in palms occupied by large populations of dominants such as O. smaragdina and Crematogaster sp. Eggs of the coconut pest Opisina arenosella Walker (and of Corcyra cephalonica (Stainton), which were used to simulate the former on palm fronds) were rapidly found and removed by certain ant species, notably M. floricola (Jerdan) and Crematogaster sp. Oecophylla smaragdina and P. longicornis seemed erratic as egg predators. Surveys of ants in areas where Opisina arenosella does, and does not, develop to outbreak populations showed that Monomorium spp. in particular were notably more common in the non-outbreak areas. However, other ants, notably Technomyrmex albipes (F. Smith), were sometimes common in outbreak areas and were seemingly good egg predators. The presence of thriving Oecophylla smaragdina colonies on currently attacked palms shows that this otherwise well-recognized predator has little or no effect on Opisina arenosella. It is concluded that several species of ants which nest in coconut palm crowns probably contribute importantly to the mortality of eggs of O. arenosella and help keep it in its rare endemic state in most places most of the time, even in outbreak areas.

The importance of Toxoptera species as agricultural pests is briefly discussed, along with the recent discovery in Africa of T. odinae (van der Goot). Banana is recorded as an unusual new host-plant record, and a table of other hosts is presented. The genus Toxoptera is defined and a key to species given.

A presence-absence sampling procedure is presented which relates the proportion of leaves infested with the mite Mononychellus tanajoa (Bondar) at a given node of cassava to the mean number of mites per leaf at that node. The procedure incorporates Taylor's variance-mean density relationship, taking into account mite aggregation behaviour. Aggregation of M. tanajoa was greater in plants where mite predators were excluded, on a resistant (CG5-79) rather than a susceptible cassava clone (MCol 22) and on leaves arising from the lower nodes rather than the higher ones. The implications of the effect of mite mortality factors on aggregation are discussed.

Digonogastra kimballi sp. n. and D. solitaria Wharton & Quicke sp. n. are described from Texas and Mexico. They attack stalk-boring pyralids in a variety of gramineous plants, including maize, sorghum and sugarcane. Known hosts are Eoreuma loftini (Dyar) and five species of Diatraea.

Infestations of the sugarcane pest Eldana saccharina Walker have been consistently serious in one part of the cane-growing belt of Natal, South Africa. Mortality in this region over seven years was estimated from moth catches in light traps, checked against population change measured from extensive larval counts, and plotted against climatic indices. The object was to construct a predictive model to illustrate mortality in other regions of the cane belt, to which infestations have spread in recent years. The spread southwards along the coast did not seem to have resulted from a decrease in natural mortality, but the spread to higher altitudes appeared to be due to lower mortality associated with warmer and drier years. The phenology of the insect and the effect of the annual harvesting cycle on phenology are discussed.

The within-row dispersion characteristics of larvae of Antitrogus mussoni (Blackburn), A. parvulus Britton, Lepidiota crinita Brenske, L. negatoria Blackburn, L. noxia Britton and L. picticollis Lea in sugarcane were determined in studies in southern Queensland, Australia. The Poisson distribution, negative binomial distribution, Iwao's regression model and Taylor's power law analysis were used to determine the relationship between mean and variance of larval counts. All methods examined showed that the larvae were slightly aggregated. Taylor's power law generally gave equivalent or better fits to the population dispersion compared with the other models. The power law relationship for L. picticollis differed from those of the other five species, and a common relationship for those five species was determined. Relationships to determine sample sizes for fixed levels of precision and fixed-precision-level stop lines for sequential sampling were developed for both L. picticollis and the other five species. There were functional relationships between the variance and mean of untransformed population counts for all species, and the suitability of transformation functions is discussed.

Insecticidal aerosols dispersed from aircraft are used in control measures against Glossina spp., and the interpretation of spray monitoring data in terms of the likely effect on the flies requires some knowledge of variation of the collection efficiency of the flies under a range of appropriate conditions. A low speed wind tunnel was used to measure the collection efficiencies of the various component parts of an adult of G. morsitans Westwood when at rest on cylinders simulating the branches and twigs of the natural habitat. The wind speed range was from 0·25 to 1·5 m/s and monodisperse droplets 10, 15, 20 and 25 μm in diameter were used. Although certain body zones exhibited very high apparent collection efficiencies (primarily due to interception), the average overall efficiencies varied from as low as 0·8% for 10-μm droplets at 0·25 m/s up to 22% for 25-μm droplets at 1·5 m/s.

Hypatima mangiferae Sattler sp. n. is described from Kenya, where its larva is injurious to commercial mango trees. A description of its life-history, based on extensive field observations and laboratory studies, is also provided. The moth, its male and female genitalia and the damage caused by the larva are illustrated.

Designs for insecticide-impregnated screens based on combinations of two different colours were evaluated for their attractiveness to Glossina palpalis palpalis (Robineau-Desvoidy) in the Ivory Coast. Screens were tested with flanking panels of mosquito netting (flanked screens) and without such panels (simple screens), using electric nets to catch flies contacting cloth of various colours or mosquito netting. No flanked screen caught more flies than an all-blue single-coloured screen; flanked screens without blue portions generally had lower catches than those with blue portions. The majority of flies was always caught over the netting panels rather than over the cloth screens, although the proportions varied for different colours. Among simple screens, one half pthalogen blue and half ultraviolet-reflecting white caught 2·4 times as many female flies as an all-blue screen and 3·6 times as many as an all-white (ultraviolet-reflecting) screen. The basis for this synergistic effect is that blue attracts flies strongly to a position near the target (e.g. circling), while the white induces landing responses; in the blue-and-white simple screens, 77% of all tsetse landed on the white surface. Several other colour combinations in simple screens gave similar results to the blue-and-white screen for females flies, for example black-and-white, and one of the blue-and-light-blue combinations. For males, the best combination for simple screens was blue and black. In tests of different configurations of blue and ultraviolet-white on simple screens, most flies were caught using a diagonally-divided screen, the lower triangular portion being white and the upper one blue.

Three carbon-dioxide-baited silhouette traps of different shape were used to investigate aspects of trap-finding by Austrosimulium bancrofti (Taylor) in the Australian Capital Territory. Highly significant effects were included in a model which generated fitted estimates of the number of flies captured under defined conditions. Fitted captures indicated higher numbers were collected on two of the 15 trapping days. Low wind and high cloud cover increased trap catches. Temperature and solar flux had a quadratic effect, with fitted catches indicating that the highest numbers were captured at 19°C and 620·6 W/m2. The results were compared with trap captures of A. bancrofti 1300 km north in Queensland.

Triks of the group of Rhipicephalus sanguineus (Latreille) in 423 total monthly collections taken from camels, cattle, donkeys, sheep and goats between September 1978 and August 1979, and in a few miscellaneous collections, from the Yemen Arab Republic were examined. Adult ticks of three species in the group were identified: R. camicasi Morel, Mouchet & Rodhain, R. turanicus Pomerantsev and R. sanguineus s.s. Host relationships and ecological affinities were analysed. R. sanguineus s.s. occurred rarely, only on dogs. R. turanicus occurred on all host species including dogs but more frequently on cattle and camels. It was collected from both dry and wet habitats but was taken regularly at only one wetter site. R. camicasi, the most common species in the group, occurred at most collecting sites on all host species except dogs. Overall, the numbers of R. sanguineus group ticks were low throughout the country and no seasonal patterns could be detected.

The human helminth parasite burdens on cyclorrhaphan flies was studied in an aboriginal settlement in Malaysia in 1987–88. Chrysomya megacephala (F.), which formed 79.5% of total fly population trapped in this settlement, had eggs of the roundworm Ascaris lumbricoides, the pinworm Trichuris trichiura and hookworm (Necator americanus and/or Ancylostoma duodenale) on the external body surface and gut lumen. There was a close correlation between monthly rainfall and the mean numbers of C. megacephala caught per two trap-hours. Eggs (but no larvae) of the parasites were found on flies when the monthly rainfall and the C. megacephala population were relatively high, in April-June 1987, October-November 1987 and March 1988. Human helminth eggs were also isolated from C. rufifacies (Macquart) and Hemipyrellia tagaliana (Bigot), being recorded from the latter in Malaysia for the first time.

The value of damage thresholds in determining the need for insecticide sprays against Helicoverpa armigera (Hübner) on cotton was evaluated in 1984 and 1985 in eastern Tanzania. Spraying at thresholds of 10, 20, 40 and 80% damaged squares was compared against spraying on a recommended prophylactic regime of six fixed-interval sprays. In both years, spraying on damage thresholds reduced the number of insecticide applications from six to between one and five without significantly affecting control of H. armigera, seed cotton yield or overall levels of damage to cotton.

Experiments were conducted in a cotton field in north-western New South Wales to determine the effect of pheromone trap design and placement on the catch of Helicoverpa punctigera (Wallengren) and H. armigera (Hübner) adults in paired traps. A cone trap design was found to give consistently larger catches of moths than two funnel trap designs. Furthermore, lower efficiency of catch of H. punctigera tended to bias relative catches in the funnel traps towards H. armigera. In comparison to the ratios of the two species in egg samples from the field, the cone trap did not show much bias to either species of moth. Traps located 40 m inside a cotton field showed larger catches and reduced variation compared with egg species ratios than traps placed at the edge of the field. Wind speed and direction and insecticide application also affected the catch in the traps. The catch data tended to become unreliable after the end of February, possibly due to changes in crop attractiveness for oviposition. It was found that the less numerous moth species at a given time still tended to be over-represented in pheromone traps in comparison to the ratios in egg counts. A linear time trend was introduced to compensate for the bias. The use of the traps shows promise in determining the relative abundance of the two species for use in resistance management strategies and other purposes.

Attempts to establish a laboratory colony of Glossina pallidipes Austen from adults and puparia of Uganda origin (Lugala) sent from a colony previously established in Amsterdam have been highly successful. Flies are maintained at 25 ± 0·5°C, L:D 12:12 and 85% RH. They are fed six days per week on whole defibrinated pig blood collected aseptically and irradiated at 1·5 kGy before storage at 4°C. The blood is presented through silicon rubber membranes at 37°C. Virgin female flies are mated when 7–9 days old with equal numbers of 9–12-day-old males in a 30 × 30 × 10-cm cage covered with white Terylene netting and illuminated with a fluorescent light from above, providing 300–500 lux on the cage bottom. The sexes are left together for 40–48 h, during which time they are offered food twice. After separation of the sexes, 30 females are housed in a cage 15 cm in diameter and two males are added. Flies are kept for 90 days before discarding. The establishment of this colony is discussed in terms of the mating behaviour of the species, highlighting differences in the laboratory between G. pallidipes from Uganda and Zimbabwe and between G. pallidipes and G. morsitans morsitans Westwood.

Studies were carried out at Jilaal Moogi, south-eastern Somalia, of the responses of Glossina pallidipes Austen to traps and targets. F3 traps caught three times as many tsetse as a biconical trap. Baiting F3 traps with acetone (released at 5–50 000 mg/h), butanone (10–50 mg/h) or octenol (0·05–5 mg/h) either alone or as mixtures did not affect the catch significantly. The catch was increased by 1·6 times (P<0·05) by releasing a mixture of 4–methylphenol (at 0–2 mg/h) and 3-n-propylphenol (0·04 mg/h), and by four times by releasing a mixture of acetone (500 mg/h), octenol (0·5 mg/h) and the two phenols. Baiting a target enclosed in an electric net with the combination of acetone, octenol and phenols increased the catch only 1·3 times, (P>0.05). Baiting an electric net plus target with natural ox odour increased the catch 1·8 times (P<0·05). A pthalogen blue (peak reflectivity = 40% at 450 nm) target caught 1·7 times as many tsetse as a black target (P<0·01) and a standard blue (reflectivity = 29%) target caught 7·4 times the catch of a yellow one (P<0·001). The results are compared with published data from Zimbabwe, Kenya and Mozambique, and it is noted that the level of response of G. pallidipes in Somalia to host odours is lower than elsewhere.

Field observations on the foraging behaviour of Hodotermes mossambicus (Hagen) showed that workers did not exhibit any marked preference for a particular grass species in a degraded savanna area in the north-western Orange Free State, South Africa. Two commercial baits, with carbaryl and sodium fluosilicate as their active ingredients, are used for H. mossambicus control; these baits were shorter, heavier and wider than the natural plant material collected by the workers. Thus, they are probably not transported as readily into the nest as the natural food. Minor workers were observed to be mainly responsible for transporting plant material, while the major workers were primarily responsible for cutting it. Observations on laboratory colonies indicated that workers collected litter first and thereafter harvested a tussock of grass from the top downwards. Foraging stopped once the tussock had been trimmed to a height of approximately 3 cm.