Abstract Female ability to exploit conspecifics in reproduction may have unusual expressions. I studied the reproductive behaviour of the golden egg bug (Phyllomorpha laciniata; Heteroptera, Coreidae) experimentally in the field and in the laboratory. Female golden egg bugs lay their eggs mainly on the backs of conspecific males and other females. Non-parental eggs are often carried. Occasionally, the eggs are laid on the food plant (Paronychia spp; Polycarpea, Caryophyllaceae) of the species but typically, those eggs survive poorly due to egg parasitism and predation. I explored the dependence of female reproduction on conspecific presence and encounter rate. I also studied female current reproductive state (which depends on if she has recently oviposited) in relation to her activity as well as male choice of a female. Female bugs preferred to oviposit on conspecifics when presented with a choice between a bug and a food plant. When alone females often did not lay eggs. Increased encounter rate with others increased female egg laying rate. Survival of carried eggs among bugs did not vary significantly although males received more eggs than females. Females with high current fecundity (mature eggs accumulated to reproductive tract) were more active than females with lower current fecundity (recently oviposited). Females with high current fecundity seemed to search for conspecifics to lay eggs on. Males also preferred to court females with high current fecundity. These females were more likely to oviposit immediately after mating, lowering the risk of female remating before oviposition. To conclude, conspecifics are important egg-laying substrates for female golden egg bugs. Conspecific availability affects female egg laying and the rate of egg production in short term. In particular, males are necessary for egg-laying females and they typically receive unrelated eggs when they court females. Sexual interactions resulting from female polyandry are crucial factors that maintain female egg laying on the backs of males and other females in the unique reproductive system of the golden egg bug.

Lodgepole pines Pinus contorta (Douglas) were treated by girdling to assess the response by an endemic population of mountain pine beetle Dendroctonus ponderosae (Hopkins) (Coleoptera: Scolytidae). Capture in 'sticky traps' indicated that the beetles were attracted while in flight towards the wounded trees. A significant difference in the landing rates of bark beetles between the treated and untreated trees indicated that the beetles were able to distinguish wounded from non-wounded trees while still in flight. Pressure chamber readings demonstrated that water stress developed in the girded trees when compared to the non-girdled trees.

Ptilothrix bombiformis (Hymenoptera: Apoidea) is a specialist bee belonging to the tribe Emphorini. The emphorine phylogeny suggests that Convolvulacea is the ancestral plant family and independent evolutionary host-switches to several unrelated plant families have occurred. The role of floral scent has been well-characterized in pollination systems involving moths, butterflies, bumblebees, and honeybees, but little is known about how specialist bees mediate host selection, or how host-choice evolved in specialist bees. This research investigates the role of floral scent in host selection by P. bombiformis. Ptilothrix bombiformis has traditionally been classified as a Hibiscus (Malvaceae) oligolege. My research shows that it can now be placed into a more detailed dietary classification as an eclectic oligolege because it also collects pure pollen loads from a distantly-related plant, Ipomoea pandurata (Convolvulaceae). Using dynamic headspace sampling and gas chromatography-mass spectrometry, I obtained floral chemical profiles for Hibiscus moscheutos ssp. lasiocarpos and Ipomoea pandurata. Both flowers contain aliphatics, aromatic compounds, monoterpenes, and sesquiterpenes. The host flowers have 14 shared compounds in their floral scent, which may be responsible for the bees' ability to recognize and utilize I. pandurata, a member or the emphorine ancestral host-plant family. Some of these shared compounds are also found in other emphorine host plants and may be responsible for their constraint in host-use.

Trichogramma atopovirilia Oatman & Platner, 1983, Trichogrammatoidea annulata De Santis, 1972 e Trichogramma bruni Nagaraja, 1983 são espécies que ocorrem naturalmente no Brasil, e assim como outras espécies de tricogramatídeos neotropicais apresentam potencial de utilização em programas de controle biológico. Todavia, para que essa estratégia seja implementada faz-se necessário um sistema eficiente de criação massal, baseando-se, principalmente, na escolha do hospedeiro adequado para multiplicação do parasitóide com qualidade. Normalmente são utilizados os hospedeiros alternativos Sitotroga cerealella (Olivier, 1819), Anagasta kuehniella (Zeller, 1879) e Corcyra cephalonica (Stainton, 1865); a produção destes hospedeiros é equivalente a mais de 70% dos custos de produção. Um sistema de criação que independa desses hospedeiros, como a criação in vitro, seria ideal para simplificar a linha de produção de parasitóides. Os objetivos deste trabalho foram: 1) selecionar o hospedeiro alternativo mais adequado para criação de T. atopovirilia, T. annulata e T. bruni, e avaliar os efeitos da criação destes parasitóides por sucessivas gerações, em suas características biológicas, incluindo capacidade de vôo e tabela de vida de fertilidade; 2) estudar a viabilidade da criação in vitro destas espécies de parasitóides. Concluiu-se que o hospedeiro alternativo afetou as características biológicas e a capacidade de vôo das espécies estudadas. Para T. atopovirilia, C. cephalonica e/ou A. kuehniella foram os hospedeiros alternativos mais adequados para sua criação, enquanto que para T. annulata e T. bruni C. cephalonica foi o hospedeiro preferencial. No entanto, com base na atividade de vôo, T. bruni não demonstrou potencial adaptativo ao longo das gerações avaliadas em nenhum dos 3 hospedeiros alternativos. As demais espécies apresentaram potencial adaptativo aos hospedeiros preferenciais. Com base em todos os parâmetros avaliados S. cerealella foi o pior hospedeiro para as 3 espécies de parasitóides estudadas. Foi possível a criação in vitro de T. atopovirilia, desde o parasitismo até a emergência dos adultos, numa dieta artificial composta de holotecidos pupais de Heliothis virescens Fabr. (65%), gema de ovo (18%), soro fetal bovino (8,5%), hidrolisado de lactoalbumina (8,5%) e anticontaminantes (0,3%). A dieta não permitiu o desenvolvimento de T. annulata e T. bruni. Os parasitóides criados in vitro apresentaram características semelhantes aos insetos criados in vivo, quando foram transferidos para o hospedeiro alternativo. / Trichogramma atopovirilia Oatman & Platner, 1983, Trichogrammatoidea annulata De Santis, 1972 and Trichogramma bruni Nagaraja, 1983 are species that occurs naturally in Brazil and just like other species of Neotropical trichogrammatid, they have potential for utilization in biological control programs. Therefore, for this strategy to be implemented, it is necessary to establish an efficient mass rearing system, based mainly on the choice of suitable hosts for multiplication of quality parasitoids. Frequently used factitious hosts are Sitotroga cerealella (Olivier, 1819), Anagasta kuehniella (Zeller, 1879) and Corcyra cephalonica (Stainton, 1865); the cost of production of these hosts is equivalent to more than 70% of the production costs. A rearing system which does not depend on these hosts such as in vitro production could be ideal to simplify production of parasitoids. The objectives of this work were to: 1) select the most suitable factitious host for rearing of T. atopovirilia, T. annulata and T. bruni, and evaluate the effect of rearing these parasitoids for successive generations on their biological characteristics including flight capacity and fertility life table and 2) study the viability of in vitro rearing of these parasitoids. It was concluded that factitious hosts affected biological characteristics and flight capacity of the studied species. For T. atopovirilia, C. cephalonica and/or A. kuehniella were the most suitable factitious hosts for its rearing, while for T. annulata and T. bruni, C. cephalonica was a preferential host. However, based on flight capacity, T. bruni did not demonstrate adaptive potential along the evaluated generations in all the 3 factitious hosts. The other species presented adaptive potential to their preferential hosts. Based on all the evaluated parameters, S. cerealella was the worst host for all the 3 parasitoid species studied. In vitro rearing of T. atopovirilia was possible from parasitism up to adult emergency on artificial diet consisting of holotissues of Heliothis virescens Fabr. (65%), egg york (18%), fetal bovine serum (8,5%), lactoalbumin hydrolysate (8,5%) and anti-contaminants (0,3%). The diet did not permit the development of T. annulata and T. bruni. The in vitro reared parasitoids presented similar characteristics to insects reared in vivo when they were transferred to factitious hosts.

Trichogramma atopovirilia Oatman & Platner, 1983, Trichogrammatoidea annulata De Santis, 1972 e Trichogramma bruni Nagaraja, 1983 são espécies que ocorrem naturalmente no Brasil, e assim como outras espécies de tricogramatídeos neotropicais apresentam potencial de utilização em programas de controle biológico. Todavia, para que essa estratégia seja implementada faz-se necessário um sistema eficiente de criação massal, baseando-se, principalmente, na escolha do hospedeiro adequado para multiplicação do parasitóide com qualidade. Normalmente são utilizados os hospedeiros alternativos Sitotroga cerealella (Olivier, 1819), Anagasta kuehniella (Zeller, 1879) e Corcyra cephalonica (Stainton, 1865); a produção destes hospedeiros é equivalente a mais de 70% dos custos de produção. Um sistema de criação que independa desses hospedeiros, como a criação in vitro, seria ideal para simplificar a linha de produção de parasitóides. Os objetivos deste trabalho foram: 1) selecionar o hospedeiro alternativo mais adequado para criação de T. atopovirilia, T. annulata e T. bruni, e avaliar os efeitos da criação destes parasitóides por sucessivas gerações, em suas características biológicas, incluindo capacidade de vôo e tabela de vida de fertilidade; 2) estudar a viabilidade da criação in vitro destas espécies de parasitóides. Concluiu-se que o hospedeiro alternativo afetou as características biológicas e a capacidade de vôo das espécies estudadas. Para T. atopovirilia, C. cephalonica e/ou A. kuehniella foram os hospedeiros alternativos mais adequados para sua criação, enquanto que para T. annulata e T. bruni C. cephalonica foi o hospedeiro preferencial. No entanto, com base na atividade de vôo, T. bruni não demonstrou potencial adaptativo ao longo das gerações avaliadas em nenhum dos 3 hospedeiros alternativos. As demais espécies apresentaram potencial adaptativo aos hospedeiros preferenciais. Com base em todos os parâmetros avaliados S. cerealella foi o pior hospedeiro para as 3 espécies de parasitóides estudadas. Foi possível a criação in vitro de T. atopovirilia, desde o parasitismo até a emergência dos adultos, numa dieta artificial composta de holotecidos pupais de Heliothis virescens Fabr. (65%), gema de ovo (18%), soro fetal bovino (8,5%), hidrolisado de lactoalbumina (8,5%) e anticontaminantes (0,3%). A dieta não permitiu o desenvolvimento de T. annulata e T. bruni. Os parasitóides criados in vitro apresentaram características semelhantes aos insetos criados in vivo, quando foram transferidos para o hospedeiro alternativo. / Trichogramma atopovirilia Oatman & Platner, 1983, Trichogrammatoidea annulata De Santis, 1972 and Trichogramma bruni Nagaraja, 1983 are species that occurs naturally in Brazil and just like other species of Neotropical trichogrammatid, they have potential for utilization in biological control programs. Therefore, for this strategy to be implemented, it is necessary to establish an efficient mass rearing system, based mainly on the choice of suitable hosts for multiplication of quality parasitoids. Frequently used factitious hosts are Sitotroga cerealella (Olivier, 1819), Anagasta kuehniella (Zeller, 1879) and Corcyra cephalonica (Stainton, 1865); the cost of production of these hosts is equivalent to more than 70% of the production costs. A rearing system which does not depend on these hosts such as in vitro production could be ideal to simplify production of parasitoids. The objectives of this work were to: 1) select the most suitable factitious host for rearing of T. atopovirilia, T. annulata and T. bruni, and evaluate the effect of rearing these parasitoids for successive generations on their biological characteristics including flight capacity and fertility life table and 2) study the viability of in vitro rearing of these parasitoids. It was concluded that factitious hosts affected biological characteristics and flight capacity of the studied species. For T. atopovirilia, C. cephalonica and/or A. kuehniella were the most suitable factitious hosts for its rearing, while for T. annulata and T. bruni, C. cephalonica was a preferential host. However, based on flight capacity, T. bruni did not demonstrate adaptive potential along the evaluated generations in all the 3 factitious hosts. The other species presented adaptive potential to their preferential hosts. Based on all the evaluated parameters, S. cerealella was the worst host for all the 3 parasitoid species studied. In vitro rearing of T. atopovirilia was possible from parasitism up to adult emergency on artificial diet consisting of holotissues of Heliothis virescens Fabr. (65%), egg york (18%), fetal bovine serum (8,5%), lactoalbumin hydrolysate (8,5%) and anti-contaminants (0,3%). The diet did not permit the development of T. annulata and T. bruni. The in vitro reared parasitoids presented similar characteristics to insects reared in vivo when they were transferred to factitious hosts.

Over the last two decades, mountain pine beetle (Dendroctonus ponderosae) populations reached epidemic levels across much of western North America, including high elevations where cool temperatures previously limited beetle persistence. Many high-elevation pine species are susceptible hosts and experienced high levels of mortality in recent outbreaks, but co-occurring Great Basin bristlecone pine (Pinus longaeva), the longest-living non-clonal organism, were not attacked. I assessed Great Basin bristlecone pine resistance to mountain pine beetle by evaluating mountain pine beetle host selection behavior and reproductive success in this species. To evaluate mountain pine beetle host selection preference for Great Basin bristlecone pine, I used no-choice 48-hour attack box experiments that confined pioneering female beetles onto pairs of living Great Basin bristlecone and limber pine (P. flexilis), a susceptible host tree species. To investigate the effect of induced tree defenses on host selection behavior, I repeated the tests on paired sections of Great Basin bristlecone and limber pines that had been recently cut, thereby removing their capacity for induced defensive reactions to an attack. Mountain pine beetles avoided Great Basin bristlecone pine relative to limber pine, suggesting that Great Basin bristlecone pine has a high level of resistance to mountain pine beetle due at least in part to stimuli that repel pioneering females from initiating attacks, even when induced defenses are compromised. To investigate mountain pine beetle reproductive success in Great Basin bristlecone pine, I compared the mating success, fecundity, and brood production of mountain pine beetle parents placed in cut Great Basin bristlecone pine bolts with that of mountain pine beetles placed in cut bolts of limber pine and lodgepole pine (P. contorta), two susceptible species. Initial reproductive development was similar in all three tree species, but nearly all brood in Great Basin bristlecone pine died before emerging. The extensive offspring mortality observed in Great Basin bristlecone pine may be a key evolutionary driver behind mountain pine beetle aversion to the species. These findings suggest that Great Basin bristlecone pine is a highly resistant species with low vulnerability to climate-driven increases in mountain pine beetle outbreaks at high elevations.

This study investigated the potential of companion planting as a strategy to aid in the management of Helicoverpa armigera in tomato and capsicum cropping systems. The investigaitons showed that a companion planting type system is unlikely to confer any benefits to the management of H. armigera in these crops. The project has contributed to the understanding of H. armigera host selection and how the pest responds to a diversified environment. Companion planting and vegetational diversity are strategies for reducing pest incidence on crops. Many theories have been devised to explain how introducing more than one plant species to a cropping space may reduce the incidence of a pest. Five ecological theories formed the basis of experiments in this study: The resource concentration hypothesis - Herbivores are more likely to find and remain on hosts that are growing in dense or nearly pure stands (Root, 1973). Associational resistance/Plant apparency - Crop plants grown as monocultures are more apparent to herbivorous insects than plants in diverse natural systems (Tahvanainen and Root, 1972). Trap cropping – the use of plants within a cropping area to attract oviposition away from the main crop (Banks and Ekbom, 1999). The ability of H. armigera to learn in regard to host seleciton (Cunningham et al., 1998a). The natural enemies hypothesis - generalist and specialist natural enemies are expected to be more abundant in diverse rather than simple systems (Root, 1973) A tall variety of sorghum (the forage variety Chopper) was tested for its ability to disrupt host location. Sorghum was planted around plots of tomatoes to act as a screen to disrupt visual cues for host location by Helicoverpa spp. No difference was found in egg numbers on tomatoes with or without sorghum companions. However, significantly higher numbers of Helicoverpa spp. larvae were found in the sorghum at flowering than in the tomatoes. This suggested the sorghum was acting as a trap crop, diverting oviposition away from the tomato crop. However, due to a short period of flowering and therefore peak attractiveness to H. armigera, sorghum was not considered to be a suitable companion plant for inclusion in commercial tomato production systems. A host preference study was conducted to determine the preferences of H. armigera for crop plants and possible companions in the hope of finding a suitable trap crop species. Tomatoes were shown to be a highly preferred plant making it difficult to find a compatible trap crop species that would be more attractive than the main crop. However, capsicums were less prefered and therefore more suitable for this type of experimentation. Marigolds were also found to be a highly preferred plant and formed the basis for further investigations into trap cropping systems. Field experiments were conducted in successive years in two geographical locations to assess the suitability of marigolds as a trap crop for capsicums in a field situation. However, in both years H. armigera incidence was extremely and unusually low leading to inconclusive results. A glasshouse experiment was performed to assess if the ratio of capsicum plants to marigolds plants had an effect on which species H. armigera would choose for oviposition. It was found that as more capsicum plants were introduced, moths became less likely to oviposit on the supposedly more preferred marigold plant. The diminishing attractiveness of a more preferred host in the presence of other hosts is a new observation of H. armigera behaviour. This result suggests that marigolds would be unlikely to be a successful trap crop in a field situation. The ability of H. armigera to learn in regard to host selection and the influence of this on host selection in the field was investigated. No evidence of learning was found. This was the first study investigating the effect of learning in a field situation; previously the behaviour had only been investigated in a laboratory situation (Cunningham, et al. 1998a). The ability of the Australian assassin bug, Pristhesancus plagipennis (Walker) (Hemiptera: Reduviidae) to predate H. armigera larvae on capsicum plants was investigated. This predator has been sucessfully been used for H. armigera management in cotton (Grundy, 2000b), but had not been investigated in horticultural crops. Significant reductions in larvae were achieved in treatments where assassin bugs were introduced. This predator warrants further investigation for inclusion in commercial integrated pest managment programs for capsicums. Pest repellent plants were also considered. Previous to this study, very little research work had been conducted on repellent plants for H. armigera. The herbs investigated were catnip (Nepta cataria), tansy (Tanacetum vulgare), basil (Ocium basilicum) and coriander (Coriander sativum). An olfactometer system was designed and built to test H. armigera moths’ responses to odours from the herbs. Glasshouse experiments were also conducted. No repellent activity was recorded with any of the herbs tested. The results of this study support the theory that Helicoverpa spp. employ a strategy of passive host selection as suggested by Firempong (1986). This means that all available plants in an area may be oviposited on with successful ovipostion and larval development leading to the future utilisation of a species as a host plant. The implication of this finding is that in-field trap cropping is unlikely to be a successful strategy for reducing H. armigera oviposition in tomato or capsicum production systems in Australia. In cotton crops in Australia, trap cropping is used as part of an area wide management strategy and aims to reduce the total Helicoverpa spp. population of a region. Such a strategy may also be successful in horticultural crops but only with the full participation of the growers of all crops that are attractive to H. armigera in a horticultural growing region. Due to the diverse nature of horticultural production this may be a difficult task.

This study investigated the potential of companion planting as a strategy to aid in the management of Helicoverpa armigera in tomato and capsicum cropping systems. The investigaitons showed that a companion planting type system is unlikely to confer any benefits to the management of H. armigera in these crops. The project has contributed to the understanding of H. armigera host selection and how the pest responds to a diversified environment. Companion planting and vegetational diversity are strategies for reducing pest incidence on crops. Many theories have been devised to explain how introducing more than one plant species to a cropping space may reduce the incidence of a pest. Five ecological theories formed the basis of experiments in this study: The resource concentration hypothesis - Herbivores are more likely to find and remain on hosts that are growing in dense or nearly pure stands (Root, 1973). Associational resistance/Plant apparency - Crop plants grown as monocultures are more apparent to herbivorous insects than plants in diverse natural systems (Tahvanainen and Root, 1972). Trap cropping – the use of plants within a cropping area to attract oviposition away from the main crop (Banks and Ekbom, 1999). The ability of H. armigera to learn in regard to host seleciton (Cunningham et al., 1998a). The natural enemies hypothesis - generalist and specialist natural enemies are expected to be more abundant in diverse rather than simple systems (Root, 1973) A tall variety of sorghum (the forage variety Chopper) was tested for its ability to disrupt host location. Sorghum was planted around plots of tomatoes to act as a screen to disrupt visual cues for host location by Helicoverpa spp. No difference was found in egg numbers on tomatoes with or without sorghum companions. However, significantly higher numbers of Helicoverpa spp. larvae were found in the sorghum at flowering than in the tomatoes. This suggested the sorghum was acting as a trap crop, diverting oviposition away from the tomato crop. However, due to a short period of flowering and therefore peak attractiveness to H. armigera, sorghum was not considered to be a suitable companion plant for inclusion in commercial tomato production systems. A host preference study was conducted to determine the preferences of H. armigera for crop plants and possible companions in the hope of finding a suitable trap crop species. Tomatoes were shown to be a highly preferred plant making it difficult to find a compatible trap crop species that would be more attractive than the main crop. However, capsicums were less prefered and therefore more suitable for this type of experimentation. Marigolds were also found to be a highly preferred plant and formed the basis for further investigations into trap cropping systems. Field experiments were conducted in successive years in two geographical locations to assess the suitability of marigolds as a trap crop for capsicums in a field situation. However, in both years H. armigera incidence was extremely and unusually low leading to inconclusive results. A glasshouse experiment was performed to assess if the ratio of capsicum plants to marigolds plants had an effect on which species H. armigera would choose for oviposition. It was found that as more capsicum plants were introduced, moths became less likely to oviposit on the supposedly more preferred marigold plant. The diminishing attractiveness of a more preferred host in the presence of other hosts is a new observation of H. armigera behaviour. This result suggests that marigolds would be unlikely to be a successful trap crop in a field situation. The ability of H. armigera to learn in regard to host selection and the influence of this on host selection in the field was investigated. No evidence of learning was found. This was the first study investigating the effect of learning in a field situation; previously the behaviour had only been investigated in a laboratory situation (Cunningham, et al. 1998a). The ability of the Australian assassin bug, Pristhesancus plagipennis (Walker) (Hemiptera: Reduviidae) to predate H. armigera larvae on capsicum plants was investigated. This predator has been sucessfully been used for H. armigera management in cotton (Grundy, 2000b), but had not been investigated in horticultural crops. Significant reductions in larvae were achieved in treatments where assassin bugs were introduced. This predator warrants further investigation for inclusion in commercial integrated pest managment programs for capsicums. Pest repellent plants were also considered. Previous to this study, very little research work had been conducted on repellent plants for H. armigera. The herbs investigated were catnip (Nepta cataria), tansy (Tanacetum vulgare), basil (Ocium basilicum) and coriander (Coriander sativum). An olfactometer system was designed and built to test H. armigera moths’ responses to odours from the herbs. Glasshouse experiments were also conducted. No repellent activity was recorded with any of the herbs tested. The results of this study support the theory that Helicoverpa spp. employ a strategy of passive host selection as suggested by Firempong (1986). This means that all available plants in an area may be oviposited on with successful ovipostion and larval development leading to the future utilisation of a species as a host plant. The implication of this finding is that in-field trap cropping is unlikely to be a successful strategy for reducing H. armigera oviposition in tomato or capsicum production systems in Australia. In cotton crops in Australia, trap cropping is used as part of an area wide management strategy and aims to reduce the total Helicoverpa spp. population of a region. Such a strategy may also be successful in horticultural crops but only with the full participation of the growers of all crops that are attractive to H. armigera in a horticultural growing region. Due to the diverse nature of horticultural production this may be a difficult task.

This study investigated the potential of companion planting as a strategy to aid in the management of Helicoverpa armigera in tomato and capsicum cropping systems. The investigaitons showed that a companion planting type system is unlikely to confer any benefits to the management of H. armigera in these crops. The project has contributed to the understanding of H. armigera host selection and how the pest responds to a diversified environment. Companion planting and vegetational diversity are strategies for reducing pest incidence on crops. Many theories have been devised to explain how introducing more than one plant species to a cropping space may reduce the incidence of a pest. Five ecological theories formed the basis of experiments in this study: The resource concentration hypothesis - Herbivores are more likely to find and remain on hosts that are growing in dense or nearly pure stands (Root, 1973). Associational resistance/Plant apparency - Crop plants grown as monocultures are more apparent to herbivorous insects than plants in diverse natural systems (Tahvanainen and Root, 1972). Trap cropping – the use of plants within a cropping area to attract oviposition away from the main crop (Banks and Ekbom, 1999). The ability of H. armigera to learn in regard to host seleciton (Cunningham et al., 1998a). The natural enemies hypothesis - generalist and specialist natural enemies are expected to be more abundant in diverse rather than simple systems (Root, 1973) A tall variety of sorghum (the forage variety Chopper) was tested for its ability to disrupt host location. Sorghum was planted around plots of tomatoes to act as a screen to disrupt visual cues for host location by Helicoverpa spp. No difference was found in egg numbers on tomatoes with or without sorghum companions. However, significantly higher numbers of Helicoverpa spp. larvae were found in the sorghum at flowering than in the tomatoes. This suggested the sorghum was acting as a trap crop, diverting oviposition away from the tomato crop. However, due to a short period of flowering and therefore peak attractiveness to H. armigera, sorghum was not considered to be a suitable companion plant for inclusion in commercial tomato production systems. A host preference study was conducted to determine the preferences of H. armigera for crop plants and possible companions in the hope of finding a suitable trap crop species. Tomatoes were shown to be a highly preferred plant making it difficult to find a compatible trap crop species that would be more attractive than the main crop. However, capsicums were less prefered and therefore more suitable for this type of experimentation. Marigolds were also found to be a highly preferred plant and formed the basis for further investigations into trap cropping systems. Field experiments were conducted in successive years in two geographical locations to assess the suitability of marigolds as a trap crop for capsicums in a field situation. However, in both years H. armigera incidence was extremely and unusually low leading to inconclusive results. A glasshouse experiment was performed to assess if the ratio of capsicum plants to marigolds plants had an effect on which species H. armigera would choose for oviposition. It was found that as more capsicum plants were introduced, moths became less likely to oviposit on the supposedly more preferred marigold plant. The diminishing attractiveness of a more preferred host in the presence of other hosts is a new observation of H. armigera behaviour. This result suggests that marigolds would be unlikely to be a successful trap crop in a field situation. The ability of H. armigera to learn in regard to host selection and the influence of this on host selection in the field was investigated. No evidence of learning was found. This was the first study investigating the effect of learning in a field situation; previously the behaviour had only been investigated in a laboratory situation (Cunningham, et al. 1998a). The ability of the Australian assassin bug, Pristhesancus plagipennis (Walker) (Hemiptera: Reduviidae) to predate H. armigera larvae on capsicum plants was investigated. This predator has been sucessfully been used for H. armigera management in cotton (Grundy, 2000b), but had not been investigated in horticultural crops. Significant reductions in larvae were achieved in treatments where assassin bugs were introduced. This predator warrants further investigation for inclusion in commercial integrated pest managment programs for capsicums. Pest repellent plants were also considered. Previous to this study, very little research work had been conducted on repellent plants for H. armigera. The herbs investigated were catnip (Nepta cataria), tansy (Tanacetum vulgare), basil (Ocium basilicum) and coriander (Coriander sativum). An olfactometer system was designed and built to test H. armigera moths’ responses to odours from the herbs. Glasshouse experiments were also conducted. No repellent activity was recorded with any of the herbs tested. The results of this study support the theory that Helicoverpa spp. employ a strategy of passive host selection as suggested by Firempong (1986). This means that all available plants in an area may be oviposited on with successful ovipostion and larval development leading to the future utilisation of a species as a host plant. The implication of this finding is that in-field trap cropping is unlikely to be a successful strategy for reducing H. armigera oviposition in tomato or capsicum production systems in Australia. In cotton crops in Australia, trap cropping is used as part of an area wide management strategy and aims to reduce the total Helicoverpa spp. population of a region. Such a strategy may also be successful in horticultural crops but only with the full participation of the growers of all crops that are attractive to H. armigera in a horticultural growing region. Due to the diverse nature of horticultural production this may be a difficult task.

This study investigated the potential of companion planting as a strategy to aid in the management of Helicoverpa armigera in tomato and capsicum cropping systems. The investigaitons showed that a companion planting type system is unlikely to confer any benefits to the management of H. armigera in these crops. The project has contributed to the understanding of H. armigera host selection and how the pest responds to a diversified environment. Companion planting and vegetational diversity are strategies for reducing pest incidence on crops. Many theories have been devised to explain how introducing more than one plant species to a cropping space may reduce the incidence of a pest. Five ecological theories formed the basis of experiments in this study: The resource concentration hypothesis - Herbivores are more likely to find and remain on hosts that are growing in dense or nearly pure stands (Root, 1973). Associational resistance/Plant apparency - Crop plants grown as monocultures are more apparent to herbivorous insects than plants in diverse natural systems (Tahvanainen and Root, 1972). Trap cropping – the use of plants within a cropping area to attract oviposition away from the main crop (Banks and Ekbom, 1999). The ability of H. armigera to learn in regard to host seleciton (Cunningham et al., 1998a). The natural enemies hypothesis - generalist and specialist natural enemies are expected to be more abundant in diverse rather than simple systems (Root, 1973) A tall variety of sorghum (the forage variety Chopper) was tested for its ability to disrupt host location. Sorghum was planted around plots of tomatoes to act as a screen to disrupt visual cues for host location by Helicoverpa spp. No difference was found in egg numbers on tomatoes with or without sorghum companions. However, significantly higher numbers of Helicoverpa spp. larvae were found in the sorghum at flowering than in the tomatoes. This suggested the sorghum was acting as a trap crop, diverting oviposition away from the tomato crop. However, due to a short period of flowering and therefore peak attractiveness to H. armigera, sorghum was not considered to be a suitable companion plant for inclusion in commercial tomato production systems. A host preference study was conducted to determine the preferences of H. armigera for crop plants and possible companions in the hope of finding a suitable trap crop species. Tomatoes were shown to be a highly preferred plant making it difficult to find a compatible trap crop species that would be more attractive than the main crop. However, capsicums were less prefered and therefore more suitable for this type of experimentation. Marigolds were also found to be a highly preferred plant and formed the basis for further investigations into trap cropping systems. Field experiments were conducted in successive years in two geographical locations to assess the suitability of marigolds as a trap crop for capsicums in a field situation. However, in both years H. armigera incidence was extremely and unusually low leading to inconclusive results. A glasshouse experiment was performed to assess if the ratio of capsicum plants to marigolds plants had an effect on which species H. armigera would choose for oviposition. It was found that as more capsicum plants were introduced, moths became less likely to oviposit on the supposedly more preferred marigold plant. The diminishing attractiveness of a more preferred host in the presence of other hosts is a new observation of H. armigera behaviour. This result suggests that marigolds would be unlikely to be a successful trap crop in a field situation. The ability of H. armigera to learn in regard to host selection and the influence of this on host selection in the field was investigated. No evidence of learning was found. This was the first study investigating the effect of learning in a field situation; previously the behaviour had only been investigated in a laboratory situation (Cunningham, et al. 1998a). The ability of the Australian assassin bug, Pristhesancus plagipennis (Walker) (Hemiptera: Reduviidae) to predate H. armigera larvae on capsicum plants was investigated. This predator has been sucessfully been used for H. armigera management in cotton (Grundy, 2000b), but had not been investigated in horticultural crops. Significant reductions in larvae were achieved in treatments where assassin bugs were introduced. This predator warrants further investigation for inclusion in commercial integrated pest managment programs for capsicums. Pest repellent plants were also considered. Previous to this study, very little research work had been conducted on repellent plants for H. armigera. The herbs investigated were catnip (Nepta cataria), tansy (Tanacetum vulgare), basil (Ocium basilicum) and coriander (Coriander sativum). An olfactometer system was designed and built to test H. armigera moths’ responses to odours from the herbs. Glasshouse experiments were also conducted. No repellent activity was recorded with any of the herbs tested. The results of this study support the theory that Helicoverpa spp. employ a strategy of passive host selection as suggested by Firempong (1986). This means that all available plants in an area may be oviposited on with successful ovipostion and larval development leading to the future utilisation of a species as a host plant. The implication of this finding is that in-field trap cropping is unlikely to be a successful strategy for reducing H. armigera oviposition in tomato or capsicum production systems in Australia. In cotton crops in Australia, trap cropping is used as part of an area wide management strategy and aims to reduce the total Helicoverpa spp. population of a region. Such a strategy may also be successful in horticultural crops but only with the full participation of the growers of all crops that are attractive to H. armigera in a horticultural growing region. Due to the diverse nature of horticultural production this may be a difficult task.