The introduction of insecticidal transgenic varieties into the Australian cotton market in the mid-1990’s allowed the industry to substantially reduce its pesticide use but resistance continues to threaten its efficacy. Indeed, CSIRO has isolated resistance in the key targets H. armigera and H. punctigera to all three toxins (Cry1Ac, Cry2Ab, Vip3A) in the current Bollgard 3 varieties. This is set in the context of an emerging global pesticide crisis that could see novel resistant variants of these pests selected elsewhere arrive into Australia.

The industry relies on a pre-emptive strategy to slow the development of Bt resistance. This is underpinned by independent monitoring of background resistance frequencies to enable the industry to autonomously respond to emerging issues, as well as research on other high priority stewardship issues related to Bt resistance in Helicoverpa species. The project was conducted in the following three parts.

PART 1: Does multiple resistance to Bt toxins in Helicoverpa spp. pose a threat to 3 gene cotton?

There is a high chance of an insect being resistant to Cry2Ab and Vip3A. Cry1Ac declines as plants age which creates selection opportunities. Our laboratory bioassays demonstrated that it is possible to select for Cry1Ac resistance in a Cry2Ab / Vip3A background.

Experiments with multi-resistant colonies (created from resistant field colonies) challenged with field grown 3-toxin cotton suggest that they carry a fitness cost but can nevertheless survive well from the neonate to 3rd instar stage. As the larvae mature they are likely to die on 3-toxin cotton but a small proportion can survive.

PART 2: Are the frequencies of resistance to 3 gene cotton increasing?

During 2017/18, we used F1 screens to sample populations of H. armigera and H. punctigera throughout the industry and did not find evidence of increases over time in the frequencies of resistance to Cry1Ac, Cry2Ab and Vip3A. We also performed F2 screens and did not isolate any dominant forms of Bt resistance but we did isolate a new recessive Vip3A resistance in H. armigera. Our continued survey of CCA members since 2008 supports frequency estimates which suggest that Bt resistance in Helicoverpa species is not increasing.

To assist with development of the molecular tool (as part of CSE1801) we examined F2 and F1 individuals previously identified as resistant using bioassays during our monitoring program. We: (1) screened for previously identified mutations and (2) examined whole genome data for novel mutations.

PART 3: What are the characteristics of different variants of Cry2Ab resistance?

There is no indication from our characterisation work that a novel variant of Cry2Ab resistant H. armigera poses a different threat to that of the first isolated variant.

Summary: It is unclear to what degree multiple resistance to Bt toxins in Helicoverpa spp. is a threat to stacked gene cotton. Our data suggest that currently there is no reason to consider changes to the Resistance Management Plan for Bollgard 3 cotton. However, it will be important to get a more complete understanding of the characteristics of the various isolated Vip3A resistance colonies to inform future methods / tools for monitoring resistance. It is also essential to validate the molecular tools being developed with standard bioassays to translate and incorporate them into future monitoring programs. Although there currently is no evidence of increasing resistance, it is critical to pre-emptively ascertain any future changes due to, for example, incursions of novel resistances from overseas and/or changing climates driving local selection in Helicoverpa and/or other pests that may carry novel resistance genes to key technologies used in Australia (i.e., Fall Armyworm).

Helicoverpa punctigera, along with Helicoverpa armigera, are major pests in Australian cotton. They are currently controlled using “Bt cotton” which contain genes derived from the bacteria Bacillus thuringiensis, that produce proteins toxic to Helicoverpa. While most effort has focused on preventing Helicoverpa spp. developing genetic resistance to these toxins, laboratory studies have shown that larvae which are not resistant are able to tolerate low to medium levels of toxin. This “induced tolerance” could lead to larvae surviving on Bt cotton without being resistant, and it could provide a stepping stone to the development of resistance. While Helicoverpa are known to develop some tolerance after one generation of exposure to Bt toxins, we did not know whether exposure is required throughout the entire larval period or only during particular instars. The aim of this summer project was to test if exposing the larvae Helicoverpa punctigera at different larval stages to 2% or 5% toxin concentrations of the discriminating dose of Cry1Ac toxin (used by CSIRO’s Resistance Team) would affect larval development and lead to tolerance in their offspring.

The results confirmed that after exposing only one generation of larvae to low levels of Cry1Ac toxins their offspring were able to tolerate higher levels of Cry1Ac than the controls. In addition we found that larvae exposed to Cry1Ac in early instars overcompensated their growth once they fed on non-toxin diet, and those exposed as late instars actively tried to avoid the toxin and developed into smaller moths. While the offspring of larvae exposed to 5% toxin as late instars showed the most tolerance, those exposed to 2% toxin as late instars also produced significantly more tolerant offspring. These results indicate that the critical period for the development of tolerance is late in larval development.

These results have implications in respect to Bt cotton efficacy, and could have implications in respect to the placement of refuges. They suggest that larvae moving off other crops and completing their development in Bt cotton could produce offspring at least as tolerant as those completing their development within the Bt crop. Therefore ideally refuges need to be far enough away from Bt cotton to avoid older larvae moving into the cotton.

Green mirids, Creontiades dilutus (Hemiptera, Miridae), are polyphagous bugs and are endemic to Australia. These bugs feed on a variety of plant species that grow across massive expanses of subcoastal agricultural landscapes in eastern Australia, as well as in the arid continental interior. Molecular evidence, along with field surveys, have demonstrated that C. dilutus bugs move long distances between native vegetation in the arid interior and crops in the eastern states of Queensland and New South Wales. These bugs arrive in the subcoastal agricultural landscapes every summer, and are important pests of cotton (Gossypium hirsutum L.) (Malvaceae). They feed on the soft tissues of cotton plants, including the developing flowers, which results in a substantial loss of fruit (cotton bolls) and the feeding damage delays harvest through the crops taking time to compensate for these losses.

The seasonal invasions of bugs into cotton are influenced by the high mobility of these insects and their ability to use a wide variety of host plant species across arid and agricultural landscapes. Some plant species are relatively good hosts in supporting the production of high numbers of nymphs, whereas others produce few nymphs but may be used incidentally as shelter, and this may aid in the dispersal of the adults across long distances. The timing of invasions of C. dilutus bugs into cotton, and the pathways followed by them are poorly understood. Also, the general mechanisms by which these insects localize appropriate host plants have not been subject to much investigation.

Host plant availability in the arid continental interior is dependent on highly variable localized rainfall, and such areas are separated by large expanses of extremely dry regions containing few green plants. In agricultural systems, non-crop plants along roadsides and the margins of farms occasionally host low densities of bugs, but previous field surveys have not found high densities of C. dilutus that may act as a source of pests that invade cotton. Within farms several legume crops, such as lucerne (Medicago sativa) and pigeon pea (Cajanus cajan) (Malvaceae), routinely support high bug densities, whereas nearby cotton crops host substantially fewer bugs. Gut analyses conducted previously suggest that individual mirids do

iii

move from lucerne into cotton, but they also move in the other direction, despite the differential numbers of bugs across these crops. Similar movement patterns across pigeon pea and cotton “boundaries” were evaluated in this thesis.

The ambiguity in the dispersal and host use patterns of C. dilutus bugs makes it difficult for pest managers to predict invasions of these insects into cotton, with accuracy. Consequently, researchers are not able to design effective management strategies. With a better understanding of dispersal and host use patterns of C. dilutus bugs, it may be possible to reduce the number of insecticide sprays in cotton if bugs could be attracted away from cotton by planting alternative hosts (trap crops), but these alternative hosts may inadvertently become local reservoirs of pests that move into cotton. A particular aim of this study is, therefore, to investigate aspects of the dispersal of these C. dilutus bugs across crop host species and the associated host localization behaviour of these insects. The ultimate goal is to use this information to form a conceptual model for the host localization process of C. dilutus bugs and provide a realistic framework to develop effective pest management decisions in cotton systems.

Specifically this thesis presents the results of: 1) surveys that asked pest managers about their perceptions of invasion patterns, 2) field surveys across vast arid and agricultural landscapes to identify which host species are used most consistently by bugs, 3) molecular evaluations to confirm feeding and movement of individuals across different hosts, and 4) behavioural experiments to identify host-associated cues used by bugs to localize specific plants.

Findings from this study indicate that most pest managers reported that the earliest seasonal infestations into cotton are associated with the proximity of cotton to legume crops and also with storms that move in from the arid regions to the west. Infestation patterns are consistent with multiple invasion events in each season and into each crop, and a gradual increase in bug numbers as nymphs develop into adults within squaring (flowering) cotton. Field surveys

iv

in the arid zone found that the highest densities of bugs were found on Cullen australasicum (Fabaceae) and Goodenia cycloptera (Goodeniaceae), at a time prior to when cotton was planted, and on lucerne and pigeon pea in agricultural systems during the flowering period of cotton.

Similarly, too, bug densities were consistently much higher on pigeon pea than on cotton. Creontiades dilutus bugs were found in the field to feed on both pigeon pea and cotton, and frequently they move back-and-forth between these crops, as found across lucerne-cotton boundaries. Behavioural tests in the laboratory revealed that these bugs are arrested in the vicinity of pigeon pea and cotton by olfactory cues, but there was no evidence that olfactory cues alone attracted bugs to either host beyond a range of 2cm. Also, this is the first behavioural study that observed an increase of insect locomotion at night, suggesting that these bugs are essentially nocturnal.

Collectively, the results of this thesis indicate that C. dilutus bugs are produced in relatively greater numbers on specific plant species than on nearby alternative species (with cotton being a relatively poor host ecologically (although not economically)). Olfactory cues that arrest bug movement to a locality appear to have a stronger influence on settling patterns across host plant species than do olfactory cues that attract bugs towards plants. A general host localization model is proposed in which high densities of bugs develop on host plant species that maintain soft tissues (after receiving rain or growing on irrigated farmlands), then move across the landscape and land on plants until recognizing cues that arrest their movement. Other host species, such as cotton, are used if their primary hosts are not available. The implications of how these bugs persist in highly variable environments are discussed, and the implications for pest management are specified.

CRDC's Strategic RD&E Plan is CRDC’s primary RD&E planning document and provides a high-level overview of CRDC’s strategic direction for the next five years. CRDC's vision is to power the success of Australian cotton through world-leading RD&E, and the Plan sets out how this will be achieved: the goals, investment approach, and planned impact. The Plan is ambitious: over the course of the five years, CRDC aims to contribute to creating $2 billion in additional gross value of cotton production through investments in RD&E. These investment will be split across the five key focus areas of the Plan: increasing productivity and profitability on Australian cotton farms; improving cotton farming sustainability and value chain competitiveness; building the adaptive capacity of the Australian cotton industry; strengthening partnerships and adoption; and driving RD&E impact.

CRDC 2013-2018 Strategic R&D Plan under its Responsible Landscape Management theme, outlines the industries desire to lead in managing natural resources and be recognized for its leadership in environmental performance. This project built on the past decade of investment in NRM research by providing a National NRM technical specialist who helped the industry meet this strategic goal through:

• developing and implementing annual national NRM campaigns,

• continuously improving the industries best practice recommendations for

NRM, and

• facilitating the capture of past and current NRM research into project

activities, outputs and outcomes.

A key measure of success under this goal as outlined within the Strategic plan is 1000km of riparian land and one million hectares of floodplain vegetation managed under best practice.

Uptake of phosphorus (P) by cotton crops demonstrated an excellent correlation with lint yield. Despite high levels of P fertiliser being applied, crops in Kununurra took up less than 15 kg P/ha and failed to produce lint yields greater than 8 bales/ha. Lint yields and P uptake at Narrabri however were considerably greater, indicating that another factor was preventing P uptake at Kununurra and thus limiting yield potential. These results confirm the P fertiliser recommendations suggested in the NUTRIpak manual.

The Cotton Field Awareness Map is an industry initiative which has been designed to highlight the location of cotton fields. The service is provided free of charge with the purpose of minimising off-target damage from downwind pesticide application, particularly during fallow spraying.

This project provides for CRDC support to the Cotton Map initiative. This initiative provides ongoing support for an online tool that enables cotton growers to communicate the location of cotton to reduce the risk of off-target damage associated with misuse of Group I herbicides.

Farmers, farm managers, resellers, consultants, agronomists, applicators and contractors are encouraged to input their cotton fields. Users can also access the Cotton Map to check the location of the paddocks they may be planning to spray to assess the proximity of the nearest cotton crop. Since the introduction of Cotton Map, reported herbicide damage to cotton has remained below 3%, compared to 11% in 2009 (before introduction of Cotton Map).

This project supports the ongoing development of IPM in cotton by targeting emerging pest issues, and inappropriate management which may threaten IPM. Key outcomes were:

1. Green vegetable bug (GVB) uses broad leaf weeds as hosts on cotton farms and in refuge areas. GVB prefer to feed and oviposit in legume crops such as mungbean, pigeon pea and soybean. Management of these weeds and crops on farms could influence risks of problems in cotton. Parasitism rates by the egg and nymphal/adult parasites are generally low. 2. Information summarising effects of the new registered compounds (e.g. Shield) and the lower rates of dimethoate has been incorporated into the ‘Impact of insecticides and miticides on predators in cotton’ table in the Cotton Pest Management Guide 2010-11. 3. Leaf damage resulting in reduced leaf area at or after cutout is unlikely to affect yield unless it is high – probably > 50% leaf loss in the upper canopy (top 6-9 nodes). Damage in the boll fill period before cutout may reduce yield. A tentative leaf loss threshold of 30% to 40% could be used. Results are relevant in assessing effects of leaf loss due to locusts and cluster caterpillar. 4. The efficacy and IPM fit of two fungal biopesticides BC639 and BC667 was evaluated. Both reduced abundance of aphids compared with the control by about 10-50% but the results were erratic and slow. However, the bio-pesticides are more selective than most commercial options – hence the conservation of beneficials may be greater. 5 The spread of CBT from the transplant colonies (= ratoon plants) was greater than from the inoculation colonies (= influxes from host outside the field). Transmission rate increased from < 10% with 1-2 aphids to > 50% with 5-15 aphids. If single aphids infest plants the latent period is 3 to 3.5 weeks but could be a little as 9 days with greater infestations. In the latter case, early management of aphids would be required to reduce the risk. 6. Pale cotton stainer (PCS) females are more damaging than males or mating couples. Females caused up to 50% yield loss and reduced germination success when feeding on young bolls. Feeding on older bolls did not reduce boll weight, but did affect boll opening, harvestability and germination. 7. Bemisia tabaci B-Biotype dominated whitefly populations during 2008-09, with virtually no B. tabaci Eastern Australian Natives, few greenhouse whitefly (Trialeurodes vaporariorum) and no B. tabaci Q-Biotype found. Volunteer and ratoon cotton, sowthistle, marshmallow, turnip weed, noogoora burr and paddy melon are hosts through winter. This project provides new information to make better decisions about management of emerging pests. Many outcomes have been delivered to industry through presentations, published resources and the WWW. Benefits to the industry are more rational decisions on the need to control pests and awareness of risks for different control options to obtain a better management balance between control and environment.

In pesticide science, research and modelling by private companies exceeds that carried out by public organisations (e.g. USDA & EPA in the USA) – e.g. Paul Hendley (Syngenta) presented or was a co-author on 10 papers at the ACS meeting. Also consulting firms (e.g. Waterborne Environmental, Inc) now are major providers of environmental research. Aerial movement of pesticides and non-active components of pesticide formulations (e.g. solvents) have become important areas of environmental assessment. Models (eg Pesticide Root Zone Model (PRZM), , Soil Water Assessment Tool (SWAT)) have become standard tools in risk assessments, evaluating management and in regulation.

Assessing water quality benefits of investments in improved land management (e.g. incentives to land holders) is a high priority (e.g. for USDA), just as is occurring in the Great Barrier Reef (GBR) catchments in Australia. Combined stream monitoring, catchment modelling and remote sensing are evolving as the preferred method, just as they are in the GBR catchments.Attend American Chemical Society (ACS) Agrochemicals Division International Research Award Symposium, ‚Don Wauchope and Friends--Reflections on the Future of Pesticide Environmental Chemistry‛ Washington DC and present a paper, and visit various USDA offices.

Publicise the outcomes of the LWA/CRDC/MDBC ‘Pesticides in the Riverine Environment’ program and the cotton industries BMP program, and learn about the state-of-the-art in pesticide science and modelling for application in Australia. Specifically:

- To attend & present an invited paper (‚What drives pesticide runoff: An empirical journey to a framework for pesticide runoff using some of Don's ideas‛) to the ACS Symposium,

- To attend a symposium at the ACS AGRO meeting on ‘Simulation modeling of pesticides in the environment’, in memory of US-EPA's Larry Burns.

- To visit & present seminars with USDA pesticide scientists at Beltsville Maryland and Tifton Georgia.