Recent experimental evidence has shown that learning occurs in the host selection behaviour of Helicoverpa armigera (Hübner), one of the world‘s most important agricultural pests. This paper discusses how the occurrence of learning changes our understanding of the host selection behaviour of this polyphagous moth. Host preferences determined from previous laboratory studies may be vastly different from preferences exhibited by moths in the field, where the abundance of particular hosts may be more likely to determine host preference. In support of this prediction, a number of field studies have shown that the ‘attractiveness’ of different hosts for H. armigera oviposition may depend on the relative abundance of these host species. Insect learning may play a fundamental role in the design and application of present and future integrated pest management strategies such as the use of host volatiles, trap crops and resistant crop varieties for monitoring and controlling this important pest species.

A cytotaxonomic study was carried out on 13 species of Encarsia Förster, known to parasitize aleyrodids and diaspidids. The chromosomes varied greatly both in number and morphology, with E. protransvena Viggiani having the lowest chromosome number (2n = 6) and E. asterobemisiae Viggiani & Mazzone the highest (2n = 20). The most common chromosome number was 2n = 10. C-banding, G-banding and silver staining for nucleolar organizer regions (NOR) provided the possibility of distinguishing between karyotypes with the same chromosome number and morphology and to identify all the pairs of homologues in a diploid set for advanced cytogenetic studies. The karyotype of Coccophagus lycimnia (Walker) was also examined. The chromosome data suggested that in Encarsia, karyotype differentiation has mainly occurred through a series of centric fusions, although other rearrangements may also have been significant. The importance of karyological data in systematic studies and in the identification of biotypes and cryptic species of economic interest is outlined.

Growing interest in biodiversity and conservation has increased the demand for accurate and consistent identification of arthropods. Unfortunately, professional taxonomists are already overburdened and underfunded and their numbers are not increasing with significant speed to meet the demand. In an effort to bridge the gap between professional taxonomists and non-specialists by making the results of taxonomic research more accessible, we present a partially automated pattern recognition system utilizing artificial neural networks (ANNs). Various artificial neural networks were trained to identify spider species using only digital images of female genitalia, from which key shape information had been extracted by wavelet transform. Three different sized networks were evaluated based on their ability to discriminate a test set of six species to either the genus or the species level. The species represented three genera of the wolf spiders (Araneae: Lycosidae). The largest network achieved the highest accuracy, identifying specimens to the correct genus 100% of the time and to the correct species an average of 81% of the time. In addition, the networks were most accurate when identifying specimens in a hierarchical system, first to genus and then to species. This test system was surprisingly accurate considering the small size of our training set.

Common house flies, Musca domestica Linnaeus, were reared to pupation with exposure to ivermectin, washed, and then exposed to parasitism by Muscidifurax zaraptor Kogan & Legner. Pupae exposed to ≥ 0.25 ppm ivermectin produced 63% fewer parasitoids than did control pupae. Pupae exposed to 0.01 ppm ivermectin produced 23% more parasitoids. Exposure to 0.10 ppm ivermectin had no detectable effect. Treatments did not affect the developmental time of male or female F1parasitoids. Treatments did not affect the production of adult F2 progeny by F1 females, nor the sex ratios of these offspring. This study documents an indirect effect of ivermectin residues across trophic levels. It identifies altered host quality as another mechanism by which faecal residues of ivermectin may affect insect activity in dung of treated cattle. The importance of this phenomenon under field conditions is undetermined.

A raster or grid-based Geographic Information System with data on tsetse, trypanosomosis, animal production, agriculture and land use has recently been developed in Togo. This paper describes the generation of area-wide digital tsetse distribution and abundance maps and how these accord with the local climatic and agro-ecological setting. Results include: (i) a spatial demarcation of ecologically distinct areas, producing a seasonal cluster map based on seasonal weather data and temporal series of satellite derived National Oceanic and Atmospheric Administration (NOAA) AVHRR and METEOSAT variables: (ii) tsetse distribution maps of Glossina tachinoides Westwood, G. palpalis palpalis (Robineau-Desvoidy), G. morsitans submorsitansNewstead, G. longipalpis Wiedemann, G. medicorum Austen and G. fusca fusca (Walker); and (iii) tsetse abundance or ‘risk’ maps, corrected for within database seasonal fluctuations, for G. tachinoides and G. p. palpalis. It is concluded that grid-based sampling is the ideal method for rapid assessment of the current vector and disease situation within any country or region, and that remote sensing has an important role to play in planning such a sampling system.

Many species of insects associated with cultivated rice do not over-winter in Korea and Japan, but migrate into these areas each year. To understand better the origins of these immigrations as well as the geographic structure of rice pests in Asian rice growing regions, intraspecific variation in two species of delphacid planthoppers, Nilaparvata lugens (Stål) and Sogatella furcifera Horvath, was examined. An 850 base pair region of mitochondrial DNA cytochrome oxidase-I (CO-I) was sequenced from a total of 71 individuals collected from 11 localities in seven countries: Korea, Philippines, China, Bangladesh, Malaysia, Vietnam and Thailand. In N. lugens, three haplotypes were found and all populations sampled shared a dominant haplotype. Localities in Korea contained two haplotypes and localities in China and the Philippines contained three. However, in samples from the Indochina peninsula no variation was detected either within or between populations, consistent with a hypothesis of regular migration and gene flow. These populations did not contain some haplotypes found in Korea, suggesting they were not the source of yearly immigration into Korea and, by extension, Japan. Populations from China did share haplotypes with Korea, which was consistent with the hypothesis that China was the source for yearly immigration into Korea. There was insufficient resolution to distinguish among populations in China. For N. lugens, the data suggested that populations south of the Red River Valley in Vietnam experienced regular mixing and were distinct from populations to the north which contributed to yearly immigrations. In S. furcifera, there was less differentiation among populations. Two haplotypes were found in all populations except Malaysia. The results for both species were consistent with seasonal weather data and indicated that more detailed analysis of DNA sequence data will be fruitful.

Three new prototype traps, (S1–S3), were developed during studies of the behavioural ecology of Glossina swynnertoni Austen in Kenya and Tanzania. The traps were compared in latin square experiments relative to the regular biconical trap as a standard and a selection of other conventional tsetse traps. Observations were also made on fly behaviour in the vicinity of traps using electric nets and sticky materials. When baited with acetone and 1-octen-3-ol, the S1 trap was 3.5 times as effective in catching G. swynnertoni in Kenya as the biconical trap. In Tanzania, the relative performance of the S1 and biconical traps differed; also, both traps were found to be inferior to an all-black, sticky 1-m2target. A second prototype (S2) performed slightly better than the biconical trap, but was still inferior to the black target. The final prototype (S3) was 2.9 times as effective as the biconical trap and performed as well as the black target. The potential for further improvement of traps for capturing G. swynnertoni and flies of the G. morsitans Westwood group is discussed.

A survey was conducted on 84 pig and cattle farms in Denmark between August and October 1996 and 1997. The aim was to describe the activity and relative abundance of parasitoids (Hymenoptera: Pteromalidae and Ichneumonidae) that attack puparia of Musca domestica Linnaeus and Stomoxys calcitrans (Linnaeus). In total, seven pteromalid species were recovered: Spalangia cameroni Perkins, S. nigripes Curtis, S. subpunctata Förster, Muscidifurax raptorGirault & Sanders, Pachycrepoideus vindemiae (Rondani), Urolepis rufipes(Ashmead) and Nasonia vitripennis (Walker), an ichneumonid Phygadeuon fumatorGravenhorst, a diapriid Trichopria sp., and a staphylinid Aleocharasp. This is the first time that U. rufipes has been recorded in Europe. Spalangia cameroni and M. raptor were the most frequently recorded species in all regions of the country, and accounted for the main parasitism of Musca domesticaand Stomoxys calcitrans puparia. The overall rate of parasitism per farm was low: 12.9% of the total number of fly puparia collected. Direct ordination, used to assess the habitat distribution of the parasitoids, showed that Muscidifurax raptor mainly seeks fly puparia in outdoor manure heaps and especially in manure from pigs rather than from cattle, whereas Spalangia cameroni mainly seeks fly puparia indoors, irrespective of livestock.

In recent decades, aphids identified as Cinara cupressi (Buckton) have seriously damaged commercial and ornamental plantings and native stands of Cupressus, Juniperus, Widdringtonia and other Cupressaceae in Africa, Italy, Jordan, Yemen, Mauritius and Colombia. Aphids currently identified as Cinara cupressiappear to belong to a species complex. This work was undertaken to identify the pest, its likely area of origin and any specific natural enemies, to facilitate a biological control programme. The biogeography of the species complex was studied using morphometric analysis of specimens from all the major regions where they have been collected, and analysis of host-plant ranges, parasitoid records and host association information. The cause of tree damage is shown to be an unnamed species, here described as C. (Cupressobium) cupressivora Watson & Voegtlin sp. n., which probably originated on Cupressus sempervirens in a region from eastern Greece to just south of the Caspian Sea. The remainder of the complex, Cinara cupressisensu lato, apparently consists of three morphologically similar species (C. sabinae(Gillette & Palmer) and C. cupressi (Buckton), probably originating in North America, and an unidentified species). There is insufficient data available to separate these taxa, or to determine the status of C. canadensis Hottes & Bradley, so C. sabinae and C. canadensis are left in synonymy under C. cupressi (Buckton). Lachniella tujae del Guercio and L. juniperi signata del Guercio, thought to be part of the C. cupressi complex, were found to be junior synonyms of C. tujafilina (del Guercio); these new synonymies are proposed and lectotypes of L. tujae and L. juniperi signataare designated.