The green mirid, Creontiades dilutus (Stal) is one of the most serious early-season pests of cotton in Australia. The cotton industry currently relies on broad spectrum insecticides for the control of Helicoverpa spp., butthis disrupts biological controlofother major cotton pests. Field experiments to evaluate the use oflucerne, Medicago sativa L., in the management of green mirids on cotton were conducted under mesh house free/no choice and commercial farm conditions at Norwood near Moree, and at Auscott and the Australian Cotton Research Institute atNarrabri in NewSouth Wales from 1992 - 94. In the mesh cage choice tests and under field conditions, lucerne was preferred over cotton by C. dilutus adults for oviposition. However, with no-choice tests, oviposition on lucerne and cotton and also the survival of mirid nymphs on these plants were not significantly different. This indicates that green mirid adults have a distinct preference for lucerne over cotton but in the absence of lucerne the female will not restrain oviposition, and will deposit the same number of eggs on cotton. In an experiment where lucerne was planted as strips within commercial cotton crops, 15 and 35 times fewer mirid adults and nymphs, respectively, were recorded on cotton compared with cotton without lucerne strips. When lucerne was interplanted with commercial cotton under an IPM regime which had no insecticide sprays against Helicoverpa spp., mirid numbers were reduced to levels similar to those achieved by nine conventional insecticide sprays. We conclude that lucerne could be incorporated into an IPM system to control green mirids on cotton.

Adoption of within-field monocultures of annual crops in modern farming systems is known to discriminate against and reduce the activity of predatory insects. In Australia, cotton fields are strictly monoculture and lack ecological diversity which could be the major cause of pest problems in the cotton agroecosystem. The utility of crops such as sunflower (Helianthus annuus), safflower (Carthamus tinctorius), sorghum (Sorghum bicolor), lucerne (Medicago sativa) and tomato (Lycopersicon esculentum) as refugia for predatory insects of Helicoverpa spp., when they were planted as strips in cotton fields, was evaluated from 1993 to 1995. Densities of beneficial insects, mainly predatory beetles, bugs and lacewings, were higher in lucerne crops than any other crop tested. In an experiment where lucerne was planted in strips within commercial cotton, the number of predators was highest in the lucerne strip and declined with increasing distance from the lucerne strip to reach their lowestlevel300 m away in the cotton crop. At the end of the study, 7.1, 6.1, 5.8 and 1.5 times more predatory beetles, bugs, lacewings and spiders per metre row respectively were recorded on the lucerne strips than on cotton. When lucerne was interplanted with commercial cotton to compare densities of predators on cotton with and without lucerne strips, 2.1, 2.5 and 1.2 times more predators were recorded on cotton with lucerne strips than cotton without lucerne strips. This indicated that the lucerne strips can serve as a refugium for predators which can then be moved into cotton fields. This study supports the findings of previous researchers that increased habitat diversity in monocultural crops by strip-cropping can increase population density of indigenous predators.

The strategy for the control of cotton pests, notably Helicoverpa spp., in Australia is insecticide-dependent with limited consideration of the role of beneficial insects. A programme for the development of an integrated pest management (IPM) approach commenced in cotton fields at Norwood near Moree in New South Wales in 1992. In this approach, an alternative crop, lucerne (Medicago sativa), was used in an interplant system plus spraying of cotton with a supplementary food (Envirofeast®), which served to attract and retain predators in treated plots. The predator complex consisted of predatory beetles, bugs, lacewings and spiders. The deployment of these two components, interplants and predator retention, reduced the number of Helicoverpa spp. eggs, very small and small larvae (first, second and third stages), medium and large larvae (fourth and fifth stages) and resulted in a cotton yield (2.72 bales ha⁻¹) that was significantly higher (p < 0.01) than the yield from the untreated control plot (0.59 bales ha -1 ). However, the yield from the interplants and food sprayed plot was significantly lower (p < 0.01) than in the insecticide-managed plot (7.12 bales ha⁻¹). Whenever the ratio of predators to Helicoverpa spp. (pests) per metre row of cotton was S 0.5, Helicoverpa pest numbers were maintained below the economic damage threshold. Though not successful in suppressing Helicoverpa spp. numbers, to avoid significant yield loss, further improvement could be made by introducing other potential components to the IPM strategy.