General enquiries on this form should be made to:randd.defra.gov.uk/Document.aspx?Document=10252_SID5... · Web viewGeneral enquiries on this form should be made to: Defra, Procurements

General enquiries on this form should be made to:

General enquiries on this form should be made to:

Defra, Procurements and Contracts Division (Science R&D Team)

Telephone No.0207 238 5734E-mail:[email protected]

SID 5Research Project Final Report

Note

In line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

· This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATION

The information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1.Defra Project code

IF0126

2.Project title

Management of habitat diversity on arable farmland to maximise control of crop pests by communities of beneficial organisms

3.Contractororganisation(s)

Rothamsted Research

(Principal Investigator: Judith K. Pell)

     

     

     

54.Total Defra project costs

£1, 009, 353

(agreed fixed price)

5.Project:start date

1 April 2007

end date

31 March 2011

6.It is Defra’s intention to publish this form.

Please confirm your agreement to do so.YES x FORMCHECKBOX NO FORMCHECKBOX

(a)When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.

Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.

In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b)If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary

7.The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Entry and Higher Level Stewardship schemes (ELS and HLS) can reduce the negative impact of intensive farming on biodiversity by providing resources such as food, shelter and over-wintering sites in managed uncropped habitats on farmland. Dual benefits could be achieved if these habitats also provide resources for natural enemies of crop pests - a win-win scenario for biodiversity and crop protection. Increasing evidence suggests that some managed uncropped habitats are beneficial for natural enemies, but there is little direct evidence of an associated reduction in crop pests. The objectives of this project were to (i) use farm scale assessments to quantify the role of uncropped habitat size, distribution and plant composition on the abundance, diversity and aphid regulatory impact of natural enemies in the crop, (ii) elucidate the mechanisms for these findings in laboratory and mesocosm experiments and (iii) assess the potential of non-food crops as alternative resources for aphid natural enemies. Together these data facilitate design and implementation improvements for uncropped land that improve delivery of aphid management services.

Aphids and their natural enemies were assessed within 100ha study areas on 14 farms in East Anglia between 2008 and 2010 using a combination of vortis sampling and visual observations. Four farms followed basic ELS requirements (farmer-managed), eight farms contained additional sown habitats (project-managed) and two were organic. The new habitats on the project-managed farms comprised four zones (floristically enhanced grass (FEG), insect rich cover (IRC), natural regeneration (NR) and winter bird seed cover (WBS)) and were distributed as strips or blocks covering 1.5 or 6ha. Samples were taken in July from the crop, other uncropped areas (predominantly grass margins) and, where present, the new sown habitats. Additional samples were also taken on four farms (two project- and two farmer-managed) in May, June and August to provide a time series. All arthropods collected were identified to family and all aphid natural enemies to genus or species. Overall 283, 128 arthropods were collected in vortis samples, 15% of which were aphids and 11% were their natural enemies. Vortis samples collected in July showed that aphid abundance in crops was greatest on farmer-managed farms whereas natural enemy abundance in crops was greatest on project-managed farms (irrespective of the size or shape of the new habitats). Similar trends were observed in the time series data where the rate of increase in aphid populations in crops over the field season was greatest on farmer-managed farms and of natural enemies in crops was greatest on project-managed farms. We hypothesise that the increased abundance of aphid natural enemies in the crop on project-managed farms was as a direct result of the presence of the new habitats as they contained approximately twice as many natural enemies as other types of uncropped land. Indeed, hymenopteran parasitoids were the most abundant group in the new habitats (65% of enemies) and comprised 57% of all natural enemies in the crop. In contrast, the most abundant natural enemy group in the other uncropped areas were spiders (43% of enemies) which only represented 22% of the natural enemies in the crop. These results demonstrate that the new habitats were a valuable resource for parasitoids, the key natural enemy group found in the crop, and suggest that parasitoids moved between the new habitats and the crops where they had a regulatory effect on aphid populations. This theory was supported by a glasshouse mesocosm experiment that showed that the parasitic wasp, Aphidius ervi, and the aphid-specific entomopathogenic fungus, Pandora neoaphidis, moved bi-directionally between the crop and a field margin where they were able to parasitise/infect aphids infesting bean or vetch plants. However, the results of the field study also showed that, because habitats provide different types and qualities of resources, care must be taken to ensure that alternative habitats on a farm are not so large or so attractive to particular groups of natural enemies that they become a resource sink rather than a source. For example, whereas there was a significant positive effect of increasing the amount of semi-natural land (relatively food/host resource poor) on parasitoid abundance in the crop, there was also a significant negative effect of increasing the total amount of uncropped land (relatively food/host resource rich) on the abundance of parasitoids and beetles in the crop. We presume these effects are modulated by the relative mobility of the species concerned and their willingness to exploit the habitats but then also move/ spill out into the crop rather than just remaining in habitats that provide resources that are always superior to the crop. Although this issue may also apply to the value of non-food crops as a resource for aphid natural enemies of crop pests, mesocosm experiments showed that flowering borage plants and aphid-infested borage/fibre nettle plants did not disrupt parasitoid foraging on adjacent bean plants. The abundance and diversity of aphid natural enemies differed between other uncropped areas, the new habitats and between the four zones of the new habitats. Similarly, this is likely to be primarily influenced by differences in the resources available. Although aphid parasitoids were one of the most diverse groups collected in vortis samples (34 different species), the identity of the three most abundant species was similar in the different areas assessed (crop, uncropped and sown habitats) and included the cereal aphid parasitoid Aphidius rhopalosiphi, which was also the most abundant parasitoid species in cereal crops and therefore a key species to encourage in uncropped habitats. However, the resources utilised by the three most commonly occurring parasitoid species differed. Whereas generalist aphid parasitoids (e.g. Praon volucre) were utilising the uncropped areas and the new habitats predominantly for non-host resources, such as nectar, the specialist cereal aphid parasitoids (e.g. A. rhopalosiphi and A. ervi) were also using these areas for reproduction, especially towards the end of the field season when resources within the crop were in decline. It is important for long term pest control potential that non-crop habitats support natural enemy reproduction as this increases their abundance c.f. just concentrating individuals from the surrounding fields. Differences in the abundance and diversity of natural enemies were also found within the new habitats; spiders were most abundant in the FEG and aphid parasitoids were most abundant in the WBC. In contrast, the abundance of beetles and predatory flies did not differ between the four zones. However, despite the abundance of natural enemies being greatest in new habitats in July, greater numbers were found in other types of uncropped land earlier in the season, a critical time for subsequent aphid regulation. This suggests that either the plant composition of the new habitats could be further optimised to provide resources earlier in the season or that some standard grass margins typical of general uncropped land should also be retained. Competitive intraguild interactions between different natural enemies may also influence their abundance and diversity in the different areas assessed. Indeed, intraguild interactions occurred in a glasshouse mesocosm experiment where plant complexity affected the reproductive success of two competing aphid natural enemies, the parasitoid A. ervi and the fungus P. neoaphidis. Whereas the reproductive success of the parasitoid was greatest in a simple field margin containing only one plant species, transmission of the fungus was greatest in a complex field margin containing seven different plant species. This difference is likely to be due to reduced aphid dispersal within the complex field margin benefitting fungal transmission such that the fungus was able to out-compete the parasitoid. It may therefore be possible to reduce negative intraguild interactions and increase the abundance of key groups of aphid natural enemies by manipulating the plant composition or the habitat planting regime (e.g mixes vs. series of single species plantings). Whilst it is not easy to determine which guilds of natural enemies are most effective, the results of the field assessment indicate the importance of parasitoids in most years and this is supported by the findings of a microcosm experiment in which aphid regulation by the parasitoid A. ervi was greater than by either predators or fungal pathogens acting alone and was similar to that of additive guilds of four and five enemies. However, it should be borne in mind that microcosm experiments were done under controlled conditions and the field studies were only over three years. With climate change predictions suggesting greater variability in weather/ seasonal conditions, encouraging a diversity of enemies would be advisable to retain function under increasingly variable abiotic conditions.

In summary, this project has demonstrated that the abundance of aphid natural enemies on farms can be increased by the addition of new habitats and that these natural enemies are able to move into the crop and have a negative impact on aphid populations. In addition, the abundance of specific groups of natural enemies can be enhanced by altering the plant composition of the habitats. However, these results also highlight several challenges that still remain, including (i) optimisation of the plant composition of uncropped habitats to support key natural enemies of a wider range of pests than just aphids, (ii) provision of season long resources for natural enemies, (iii) promoting movement from sown habitats into the crop at key times to prevent these areas becoming natural enemy sinks and, (iv) further understanding the multi-trophic and intraguild interactions that occur within habitats to minimise negative processes and maximise pest control potential.

Project Report to Defra

8.As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include:

the scientific objectives as set out in the contract;

the extent to which the objectives set out in the contract have been met;

details of methods used and the results obtained, including statistical analysis (if appropriate);

a discussion of the results and their reliability;

the main implications of the findings;

possible future work; and

any action resulting from the research (e.g. IP, Knowledge Transfer).

Background

Globally we face the significant challenge of combining food security with environmental stewardship for a population predicted to approach 9 billion by 2050, and whose ‘footprint’ per captita in terms of food and energy demands are increasing rapidly. In the UK, 80% of our land is farmed, the food and drink supply chain accounts for 7% of GDP and employs 3.7 million people. The ability of UK farming to increase productivity and deliver food security while simultaneously protecting biodiversity and the environment, is a serious challenge (Beddington, 2010; Godfray et al., 2011; Phalan et al., 2011).

While intensification over the past 50 years increased crop yields significantly it was also associated with severe declines in biodiversity and degradation of the rural environment (national Ecosystem Assessment, 2011). Reforms of the Common Agricultural Policy (CAP) and adoption of ‘The Strategy for Sustainable Farming and Food - Facing the Future’ (Defra, 2002) has moved towards correcting this by rewarding farmers for environmental management practices through their Entry and Higher Level Stewardship schemes (ELS and HLS respectively). Currently two thirds of farmers are signed up to the ELS and it is hoped that this will increase through the Campaign for the Farmed Environment (CFE) initiated in 2009. Semi-natural areas and uncropped land managed through ELS/ HLS for wildlife (with a focus on pollinators and birds) can play a vital role in maintaining biodiversity on farmland because they provide resources such as food, shelter and overwintering sites. These resources are also valuable for a wide variety of the natural enemies of herbivorous pests (parasitoids, predators and pathogens) that deliver an important pest management service central to sustainable farming systems (e.g. Pell et al., 2010; Bianchi and Wäckers, 2008; Olsen and Wäckers, 2007; Ekesi et al., 2005; Powell et al., 2003; Marshall and Moonan, 2002; Landis et al., 2000; Griffiths et al., 2008). Identifying and evaluating simple but effective options for ELS that deliver dual benefits for biodiversity and crop protection, - a win-win scenario, is an urgent priority as such measures would be taken up most widely. The results of this project inform the development of these options.

Previous research, including our own, has shown that provision of resources in managed field margins can increase the abundance of particular natural enemy species/ groups within those managed margins (e.g. Thomas et al., 1991; Marshall & Moonen, 2002; Powell and Pickett, 2003; Ekesi et al., 2005). However, it is not always clear whether this is as a result of reproduction within the habitat or only attraction to the habitat from elsewhere in the ecosystem (Marshall and Moonan, 2002). While both processes are likely to occur, the former would be more important for potential pest control in adjacent crops and begs the question of how the plant composition of field margins should or could be tailored to optimise the reproduction of aphid natural enemies and not just support those already present in the landscape. Following on from this, despite studies indicating that the abundance of natural enemies is high in field margins, there are few studies where ‘spill over’ from the margins into the adjacent crop and subsequent improved pest control have been demonstrated (e.g. Powell et al., 2004; LK0915). Margin design should consider the abundance and diversity of the entire guild of enemies and not just target groups as has been the case in the majority of previous studies. This is important because it is widely accepted that, particularly for conservation biological control, a diverse natural enemy community is essential to ensure resilient pest management in the long term (Defra-funded project AR0318; Griffiths et al., 2008). However, this provides a further challenge as enemies can interact with each other in positive and negative ways (e.g. Pell et al., 1997; Roy et al., 1998; Baverstock et al., 2009), so to retain them all in the agricultural landscape, it is essential to provide resources, through habitat manipulation, in a way that encourages co-existence and niche complementarity and not competition and exclusion. While it is important in terms of pest management that field margins provide a diversity of resources for a diversity of aphid natural enemies that allows them to co-exist and reproduce, it should not be so attractive that movement of the natural enemies into the adjacent crop becomes unlikely. For pest control to be effective these habitats must be a source of enemies and not a sink. In this project we address these issues through a series of laboratory, glasshouse mesocosm and field studies to evaluate the effect of increasing habitat complexity, (the juxtaposition of managed field margins, non-food crops and arable crops), on the species richness, diversity and abundance of natural enemies, their interactions with each other and their potential to regulate aphid populations.

Specifically we will test hypotheses on the role of habitat diversity on enemy co-existence in a series of replicated experiments at increasing spatial scales (laboratory to glasshouse mesocosms) using model natural enemy communities for which we have substantial existing data. By manipulating habitat complexity (crop and non-crop plant diversity, alternative host availability, nectar availability) we can determine the levels of habitat diversity necessary to maintain natural enemy diversity and co-existence and provide robust pest management. Arable crops and their associated aphids will be used in all experiments and borage and nettles will be our model non-food crops as they provide nectar/ alternative early season hosts respectively. In order to extend interpretation to the field scale and determine the scale, distribution and diversity of field margin habitats necessary to support within-crop natural enemy diversity and cereal aphid control we will sample in arable crops on farms from the farm4Bio LINK project that vary in the quantity and distribution of managed uncropped land. Our aim is to provide the evidence base necessary to optimise management of the evolving habitat mosaic within arable landscapes to enhance natural enemy diversity and hence improve natural pest suppression. This will reduce the necessity for insecticide inputs in arable food crop production (e.g. cereals) helping farmers to become net positive contributors to the environment and reduce the environmental footprint of food production substantially. This has clear significance for sustainable food production and security.

O1. Mesocosm studies to determine the role of plant diversity in non-crop managed field margin habitats on natural enemy community diversity and pest management ecosystem services - COMPLETED

Here we use glasshouse mesocosms as a model system to (a) assess the effect of field margin plant composition on reproductive potential and co-existence of two competing aphid natural enemies (b) determine whether field margins are sources and/ or sinks of these aphid natural enemies and c) determine the role of guild diversity and identity on aphid regulation on a crop plant

The role of non-crop field margin plant diversity on natural enemy reproduction and co-existence

It is possible that negative intraguild interactions that occur in field margins may reduce the reproductive potential of aphid natural enemies and the subsequent control of pest aphids in adjacent crops. Using glasshouse mesocosms we evaluate whether margin complexity influenced the reproductive potential of two competing aphid natural enemies; the hymenopteran parasitoid A. ervi and the entomopathogenic fungus P. neoaphidis. Previous studies have shown that the presence of P. neoaphidis can have negative effects on the reproductive potential of A. ervi on crop plants (Baverstock et al., 2009). Six glasshouse mesocosms (180cm x 120cm) containing one, two or seven plant species were prepared for each run of the experiment. Either thirty-six vetch plants (monoculture), 18 vetch + 18 Yorkshire fog (biculture) or 12 vetch + 4 each of Yorkshire fog, yarrow, stinging nettle, wild radish, cocksfoot and common knapweed (polyculture) were added to the mesocosms. All plants were approximately six weeks old and not flowering when used in experiments. One hundred healthy Acyrthosiphon pisum (Harris) were added to each of twelve vetch plants in each mesocosm. After 24 hours, 80 aphids that had been inoculated with P. neoaphidis 24 hours prior to the start of the experiment (and would succumb to infection and sporulate within five days) plus 12 female A. ervi were also released into the mesocosms. After a further 15 days the number of P. neoaphidis-sporulating cadavers and A. ervi mummies on each plant was recorded. Two blocks of three mesocosms were prepared on two occasions. A linear mixed model was fitted to these data using restricted maximum likelihood (REML) in order to assess whether margin type affected (a) the total number of fungal cadavers or parasitoid mummies either on the 12 plants that were initially infested with aphids or on all 36 plants in each mesocosm, and (b) the percentage of plants in each mesocosm on which parasitoid mummies or fungal cadavers were found.

The quality of the plants in the monoculture and polyculture was good at the end of the experiment allowing the number of fungal cadavers and parasitoid mummies to be accurately assessed. However, in the biculture the Yorkshire fog had overgrown the vetch resulting in low vetch plant quality and poor infestation by A. pisum. The biculture was therefore omitted from the analyses. Although only vetch plants were initially infested with A. pisum, fungal cadavers and parasitoid mummies were recovered from all seven species of plants. Whilst the greatest number of fungal cadavers and parasitoid mummies were found on vetch, there was a considerable difference in the number of fungal cadavers and parasitoid mummies on the remaining six plant species. There was a significant effect of margin type on the overall number of fungal cadavers (F1,7 = 64.98, p<0.001) and parasitoid mummies (F1,8= 8.60, p= 0.019) on the twelve plants initially infested with aphids, with significantly more being found in the polyculture compared to the monoculture (figure 1.1). Similar results were found for fungal cadavers when all 36 plants were assessed in each mesocosm, with significantly more fungus-infected aphids being found in the polyculture (F1,7 = 9.32, p = 0.019). However, when all 36 plants were considered, a greater number of parasitoid mummies were found in mesocosms containing the monoculture (F1,8 = 10.69, p = 0.011) (figure 1.1) than on plants in the polyculture. There was a significant effect of margin type on the percentage of plants in each mesocosm that contained parasitoid mummies (F1,3 = 307.99, p < 0.001) and/or fungal cadavers (F1,3 = 41.36, p = 0.005), with the distribution of both enemies being greater in the monoculture than the polyculture.

The mesocosm experiments provided evidence that aphids infesting non-crop plants in field margins were utilised as hosts for reproduction by the two natural enemies of cereal/ legume aphids that we evaluated, P. neoaphidis and A. ervi, and that margin complexity affected intraguild interactions between these species. Parasitised and fungus-infected aphids were abundant in both the monoculture and the polyculture, predominantly infesting the vetch plants on which they were released. Whilst these two competing species were clearly able to co-exist in mesocosms containing either the monoculture or a polyculture, determining which margin type was optimal for each natural enemy differed depending on whether only the 12 plants initially infested with aphids (release plants) or all 36 plants (whole margin) were included in analysis. When only the release plants were assessed the greatest number of fungal cadavers and parasitoid mummies were found in the polyculture. This may have been because aphid dispersal between plants was greater in the monoculture where there were three times more vetch plants available for colonisation by A. pisum. Indeed, whilst aphid dispersal was not directly assessed, aphids that been parasitised or succumbed to fungal infection were found on significantly more plants in the monoculture whereas aphid density on vetch plants appeared considerably greater in the polyculture after 15 days. It is therefore unsurprising that a greater number of fungal cadavers and parasitoid mummies were found on the release plants in the polyculture where clustering of aphids was greatest. In contrast, when the whole margin was assessed those aphids that had dispersed from the release plants and had been subsequently parasitised by A. ervi or infected by P. neoaphidis were included in the assessment. Whilst this would have substantially increased the number of prey available to an active forager such as A. ervi, it would have had less of an effect on the reproductive success of the passively dispersed fungus (parasitoid mummies were found on approximately 20% more plants in both the monoculture and polyculture compared to fungal cadavers). Indeed, when the whole margin was assessed the number of parasitoid mummies found in the monoculture was significantly greater than in the polyculture. In contrast, the number of fungal cadavers remained higher in the polyculture where aphid density on the release plants was greatest and therefore facilitating density dependent transmission. The reduced reproductive success of A. ervi in the polyculture relative to the monoculture therefore represents a fitness cost experienced by the parasitoid when forced to compete with the fungus. However, despite the reproductive success of A. ervi and P. neoaphidis being greater on the monoculture and polyculture respectively, overall there was little difference in the total number of enemies (fungal cadavers + parasitoid mummies) in mesocosms containing the monoculture (682 enemies) or the polyculture (747 enemies). Therefore, whilst differences in the plant composition of the margin may have affected the relative reproductive success of these two competing aphid natural enemies, overall there was little numerical effect on the total number of enemies.

Figure 1.1. Mean total number of P. neoaphidis-sporulating cadavers and A. ervi mummies collected on the 12 vetch plants on which A. pisum were released and on all 36 plants within the mesocosms containing 1 or 7 plant species. Error bars represent ± 1 x SEM.

· Plant diversity affects the competitive interactions between aphid natural enemies

· Negative intraguild interactions could be reduced by managing the plant composition of field margins thereby increasing the overall abundance of beneficial species

Food Crop/ non-crop field margin interface, natural enemy co-existence and aphid population regulation `

It is important in terms of pest management, that field margins are not so attractive that they act as a sink rather than a source of natural enemies that migrate into the adjacent crop. We used glasshouse mesocosms (as described above) to assess (a) the reproduction of A. ervi and P. neoaphidis in a crop and adjacent complex field margin (polyculture as described above) and (b) whether these enemy species moved over the crop-margin interface. Each mesocosm was sub-divided into a crop zone and a margin zone. The crop zone contained 18 single bean plants (approximately 14-days-old) whilst the margin zone contained five vetch plants and three pots of ‘fallow’ soil, plus two pots each of Yorkshire fog, yarrow, stinging nettle, wild radish and cocksfoot plants (positions allocated randomly). One hundred A. pisum were released onto each of the five vetch plants and onto the five bean plants in the same relative positions on the crop side. After 24 hours ten female A. ervi and 80 P. neoaphidis-infected A. pisum were released in either the crop or the margin. After a further 15 days the number of parasitoid mummies and fungal cadavers was assessed on the five A. pisum-infested bean plants in the crop and the five vetch plants in the margin. Three blocks of two mesocosms were prepared on two occasions. ANOVA was used to assess whether there was an effect of (a) the zone (crop or habitat) in which the enemies were released, (b) the plant species on which the enemies were found (crop or margin), and (c) whether there was an interaction between these two variables, on the total number of fungal cadavers and parasitoids mummies found on the five plants that were initially infested with aphids in each mesocosm.

Pandora neoaphidis-sporulating cadavers and A. ervi mummies were found in both zones of the mesocosm in all treatments. The number of fungal cadavers was not significantly affected by the position in which P. neoaphidis-infected aphids were released (F1,5 = 0.29, p = 0.616) (Fig. 1.2), with a mean of 33.4 and 39.8 fungal cadavers (SEM = 8.38) found in the release zone and adjacent zone respectively (Fig. 1.2a). Similar results were found for A. ervi where oviposition was not significantly affected by position of release (F1,5 = 0.10, p = 0.768), with 17.6 parasitoid mummies found in the zone in which the parasitoids were released and 19.1 mummies (SEM = 3.40) in the adjacent zone (Fig. 1.2b). Fungus-infected aphids and parasitoids therefore moved from the crop to the margin and vice-versa. There was no significant effect of plant type on the number of fungal cadavers (F1,5 = 0.05, p = 0.825) or parasitoid mummies (F1,5 = 0.58, p = 0.480) (Fig. 1.2ab), with 35.6 and 37.6 fungal cadavers (SEM = 6.08) and 16.8 and 19.8 parasitoid mummies (SEM = 2.78) found on bean and vetch plants, respectively. There was no interaction between release side and plant type for fungal cadavers (F1,2 = 3.34, p = 0.209) or parasitoid mummies (F1,1 = 0.46, p = 0.567) (Fig. 1.2ab). Transmission and development of the fungus and oviposition and development of the parasitoid was therefore equally good on vetch and bean plants.

Whilst the abundance of aphid natural enemies on farms with field margins has been shown to be greater than on farms without margins (Landis et al., 2000; Meek et al., 2002; Pell et al., 2010), these studies do not provide unequivocal evidence that field margins are the source of the natural enemies nor do they show movement from the margin into the crop. It is possible that field margins can also be a sink, attracting enemies that would otherwise be distributed elsewhere in the ecosystem. The results of the mesocosm experiment described here show that whilst parasitoids and fungal pathogens were able to reproduce in the margin zone, they also moved into the crop zone where they parasitised/ infected aphids infesting crop plants. Although this supports the hypothesis that aphid natural enemies reproduce in field margins and subsequently migrate into adjacent crops, movement between the crop and margin was bi-directional, with both enemies also moving from the crop into the margin. This is a good thing as it allows enemies to utilise resources as they become available, maintaining populations over the entire season. When the crop is harvested there would be potential for the enemies to then move to margins for overwintering. In addition, the number of parasitoid mummies and fungal cadavers was similar on vetch and bean plants indicating that A. pisum infesting plants in the crop and margin were equally good hosts for the parasitoid and fungal pathogen. Whilst aphid natural enemies can migrate from field margins into the crop, the extent of this movement may be influenced by host availability.

Figure 1.2. Total number of (a) P. neoaphidis-sporulating cadavers and (b) A. ervi mummies per mesocosm (i) in the zone in which the enemies were released or the adjacent zone, (ii) on bean or vetch plants, and (iii) on bean or vetch plants when the enemies were released either in the zone containing the assessed plant (+) or into the adjacent zone (-) (error bars represent ±1 x SEM).

· The aphid natural enemies evaluated move bi-directionally between the crop and field margin utilising all available resources

Natural enemy co-existence and aphid population regulation

There is conflicting evidence concerning the effect of natural enemy species richness on the control of pest species, with positive, neutral and negative effects of enhanced diversity on pest control being observed (Straub et al., 2008). Indeed, multiple predators may have emergent impacts on prey where the overall effect is not the sum of the constituent parts (Sih et al., 1998). The population regulation of A. pisum by individual species and guilds of enemy species was therefore assessed using microcosms. Perspex cages (0.5m2 x 1m) containing eight V. faba plants positioned around a central pot were prepared and maintained within a controlled environment room (18°C, 16L:8D). Soil was placed over the pots to allow the insects to move between plants. Forty A. pisum (2nd or 3rd instar) were released onto each of the eight plants 24 hours prior to the start of the experiment. Natural enemy assemblages were added either as single species or as guilds of four enemy species (one replicate per treatment) (figure 1.3). Two control cages containing no natural enemies were prepared along with two further cages containing all five natural enemy species. The number of aphids remaining on four plants in each cage, including two that had been infested with P. neoaphidis, was assessed after one day. The remaining plants were maintained in a controlled environment room (18°C, 16L:8D) and the number of P. neoaphidis-sporulating cadavers and parasitoid mummies assessed after a further six and eleven days respectively. The experiment was done on three occasions, with each treatment randomly allocated to a cage on each occasion. A logistic regression (GLM with binomial error and logit link) was used to assess whether there was an effect of guild size (zero, one, four or five enemies) on the proportion of aphids surviving along with whether there were differences within single enemy treatments and within treatments containing guilds of four enemies. This analysis was also used to assess the effect of guild size on the proportion of aphids that were infected with P. neoaphidis or parasitised by A. ervi. All analyses were done using GenStat (13th Edition).

Figure 1.3. Proportion of aphids surviving per cage (total of the four plants assessed) when there were no natural enemies or when A. ervi (Ae), H. axyridis (Ha), 2nd or 3rd instars C. carnea (Cc2 or Cc3) or P. neoaphidis (Pn) were introduced alone or as guilds of four and five enemies. Error bars represent ±95% confidence intervals. Dashed lines indicate the mean proportion of aphids surviving in the presence of zero, one, four or five enemies.

Aphids were recovered from all plants assessed in each cage indicating that predation was not prey limited. The mean proportion of aphids surviving differed significantly amongst guild sizes (F3, 28= 22.97, p<0.001), with a mean proportion of 0.74 (95% CI = 0.59 – 0.85) aphids surviving in the control treatment, whereas a mean proportion of 0.62 (95% CI = 0.53 – 0.70), 0.34 (95% CI = 0.26 – 0.43) and 0.30 (95% CI = 0.18 – 0.45) aphids survived in the treatments containing one, four or five enemy species, respectively (Fig. 1.3). The mean proportion of aphids surviving differed significantly between the single enemy treatments (F4, 28= 5.67, p<0.01) (Fig. 1.3), with fewest aphids surviving in the presence of A. ervi followed by H. axyridis. The remaining enemies had little effect on aphid survival when introduced alone. The mean proportion of aphids surviving also differed significantly between the guilds of four enemies (F4, 28= 3.24, p<0.05), with fewest aphids surviving in the presence of guild 3 and guild 5, indeed, there were fewer aphids surviving in these treatments than in the presence of guild 6, which contained all five enemies (Fig. 1.3). There was no significant effect of guild size on the proportion of aphids that became infected with P. neoaphidis (F2, 16= 2.68, p>0.05), with a mean proportion of 0.03 (95% CI = 0.01 – 0.07) aphids infected with P. neoaphidis when the fungus was introduced as a single enemy and 0.06 (95% CI = 0.03 – 0.10) and 0.09 (95% CI = 0.04 – 0.16) aphids infected when the fungus was introduced as part of a guild of four and five enemies respectively. There was also no significant effect of guild size on the proportion of aphids parasitised by A. ervi (F2, 16= 0.25, p>0.05), with 0.20 (95% CI = 0.09 – 0.38), 0.20 (95% CI = 0.13 – 0.30) and 0.16 (95% CI = 0.07 – 0.29) aphids parasitised when A. ervi was introduced alone or as part of a guild of four or five enemies.

Encouraging an optimal guild of enemies is essential to the success of a conservation biological control scheme. The selection of this guild requires detailed knowledge of the biology of the individual enemies and how they interact (Costamagna et al. 2008). The microcosm study described here assessed aphid survival in the presence of single species of enemies, substitutive guilds of four enemies and a guild of five enemies. Unlike previous studies, the guilds evaluated here contained all three functional groups of aphid enemies (parasitoids, predators and pathogens) thereby allowing the interactions between a taxonomically diverse guild of enemies to be assessed and predictions of an optimal guild to be made. Contrary to what was expected the parasitoid A. ervi, not the coccinellid H. axyridis (an invasive species in the UK which is regarded as being a voracious predator of aphids), had the greatest effect on survival of A. pisum after 24 hours. Indeed, when introduced alone A. ervi was able to reduce the aphid population to a level considerably lower than that of the other enemies and to a level similar to that of the guild of four enemies. Furthermore, of those aphids surviving after 24 hours, 21% were parasitised and died within 12 days. Despite previous studies showing a reduction in successful parasitisation due to asymmetric intraguild predation and competition with the pathogen (Baverstock et al. 2009; Meyhofer and Klug 2002; Powell et al. 1986), the results here show that the proportion of parasitised aphids was similar when the parasitoid was introduced alone or as part of a guild of four or five enemies. Interestingly, the impact of the parasitoid on the aphid population after 24 hours was as a result of disturbance and subsequent death due to foraging activity and not parasitism. This raises the question of whether the observed reduction in aphid abundance in the treatments containing predators is due to consumption or disrupted foraging activity of the prey. It should be noted that the aphid species used in this experiment, A. pisum, is sensitive to disturbance and readily drops from a feeding site to avoid predation or parasitism and in doing so can suffer high levels of mortality (Chau and Mackauer, 1997).

Although aphid survival decreased as guild size increased, this may be an artefact of the experimental design. When assessing the impact of multiple species on the consumption of a shared prey the experimental design can either equalise the number of predators between treatments via scaling formulae (e.g. allometric scaling of metabolic rate to body mass) that allows predator diversity to be manipulated whilst the potential predator feeding rate remains constant, or through additive designs where increased diversity results in an inevitable increase in the potential predator feeding rate. Here an additive design was used because parasitoids and pathogens do not conform to the requirements of scaling formulae. Whilst these results show that aphid survival decreased with increasing guild size (as would be expected with an additive design), they also suggest that aphids could be regulated by a single species of enemy (i.e. A. ervi). However, this would be a high risk strategy to employ given the impact of biotic and abiotic factors on the establishment and survival of aphid enemies. Therefore to successfully implement a conservation biological scheme against aphids it would be preferential to identify a guild of enemies.

These results show that the combination of species within the guild had a significant effect on aphid survival, with fewest aphids surviving in the presence of guild 3 and guild 5. However, guild 3 contained representative species from each of the three major functional groups including the two species that had the greatest impact on the aphid population when introduced alone (A. ervi and H. axyridis) along with species that function best under contrasting abiotic conditions, i.e. A. ervi forages optimally under dry conditions whereas transmissions of P. neoaphidis is greatest under damp conditions. Whilst it is unlikely that the entire guild will co-occur at all times throughout the field season, the diversity of enemies in this guild may provide resilience against control failure. In contrast, guild 5 only contained enemies from two functional groups (predators and parasitoids) and may, therefore, provide less resilient control under fluctuating abiotic conditions.

Predation by guild 4 was considerably lower than by guild 3, despite the two guilds only differing by the life stage of C. carnea (guild 3 = 3rd instar, guild 4 = 2nd instar). This suggests that a negative intraguild interaction may have occurred within this guild, presumably between 2nd instar C. carnea and one or more guild members. However, it is difficult to determine the cause of this negative effect. For example, the negative interaction may have occurred between 2nd instar C. carnea and H. axyridis, a polyphagous predator of various arthropods including C. carnea (Gardiner and Landis, 2007). Indeed, previous studies have shown that 2nd instar C. carnea are more at risk of predation from H. axyridis than 3rd instar C. carnea, presumably due to 2nd instars being smaller (Wells, personal communication). However, this scenario does not explain why there were no apparent negative interactions in guild 1 and guild 5 where these species also co-occurred. This highlights the fact that subtle changes in the identity of species within the guild may have unpredictable effects on guild function. It also identifies a potential problem: C. carnea needs to pass through the sub-optimal 2nd instar before reaching the optimal 3rd instar. However, given that C. carnea only remain as 2nd instars for approximately three days (Baverstock et al., 2011), a realistic scenario under field conditions would be the simultaneous co-occurrence of guild 3 and guild 4.

Unlike the single enemy treatments containing predators or the parasitoid, the abundance of aphids in the single enemy treatment containing P. neoaphidis was greater than in the control treatment. It is unclear why this occurred given that the aphids used were either 2nd or 3rd instars that are unable to reproduce. However, this trend has been observed in previous studies (Baverstock, 2004; Wells, unpublished data) and at present we are unable to explain why it occurs. Although not significant, transmission of P. neoaphidis was greater in the presence of predators and parasitoids, thereby supporting the findings of previous studies which show that foraging arthropods enhance fungal transmission, presumably as a result of disturbing feeding aphids and increasing their likelihood of contacting infective conidia deposited on the leaf surface (Baverstock et al., 2008; Baverstock et al., 2009; Roy and Pell 2000; Wells et al., 2011).

By assessing guilds of enemies containing species from each of the functional groups it is possible to determine the optimal species or combination of species for encouragement in conservation biological control schemes. It must be noted that although we have used a technically challenging number of enemy species in this experiment, they do represent on part of the entire guild, and within conservation biological control it would be a serious challenge to encourage a specific enemy guild in the field. However, the results do demonstrate that guild diversity is important for aphid control. Also, when identifying a guild of enemies care would be needed not to increase the abundance of species that may have negative effects on non-target species. Whereas aphid-specific enemies, such as the parasitoid and entomopathogenic fungus used in this study, are considered as being ideal biological control agents, generalist predators such as H. axyridis are capable of having negative impacts on biodiversity, particularly other aphid enemies (Roy et al.. 2005; Pell et al.. 2008), and may therefore be less desirable components of conservation biological control schemes. It must also be remembered that these experiments were made under constant conditions. Under fluctuating and, potentially extreme, climatic conditions interactions are likely to change, favouring a different suite of natural enemies.

· Guild composition affected aphid regulation

· For the species evaluated, the most effective guild in terms of aphid suppression comprised A. ervi, H. axyridis, 3rd instar C. carnea and P. neoaphidis

· Parasitoids alone regulated aphid populations as well as guilds of four and five enemies

O1 is submitted for publication as:

Baverstock J., Torrance M.T., Clark S.J., Pell J.K. (submitted) Exploiting field margins as reservoirs of entomopathogenic fungi: Mesocosm experiments to assess the transmission of Pandora neoaphidis within simple and mixed margins and over the crop-margin interface. Journal of Invertebrate Pathology, December 2011

Baverstock J., Aleman-Martinez J., Clark S.J., Porcel M., Pell J.K. (submitted) Comparison of taxonomically diverse guilds of aphid natural enemies to identify an optimal species combination for aphid regulation? BioControl, November 2011

O2. Field studies to determine the role of scale and distribution of managed uncropped habitats on within-crop natural enemy community diversity and pest management ecosystem services - COMPLETED

Arable ecosystems containing large crop monocultures are generally unfavourable environments for arthropods, except for those species that exploit the crop monocultures and become pests (Landis et al., 2000). High levels of disturbance coupled with limited nectar and pollen resources result in low colonisation rates by non-pest species (Heimpel and Jervis, 2005; Landis et al., 2000; Olsen and Wäckers, 2007; Theis and Tscharntke, 1999). Semi-natural habitats and managed uncropped field margins can, therefore, play a vital role in maintaining biodiversity on farmland as they can provide a stable environment containing food, refuges, and overwintering sites for a wide variety of insects including beneficial species such as crop pollinators and natural enemies of herbivorous pests (Landis et al., 2000; Marshall and Moonan, 2002; Olsen and Wäckers, 2007). By managing the plant composition of field margins, added value in respect to conservation biological control schemes against pest aphids in arable crops could be achievable (Pell et al., 2010; Frere et al., 2007; Gurr et al., 2003). However, although field margins have been shown to be beneficial for aphidophagous insects from the Hymenoptera, Coccinellidae, Syrphidae, Chrysopidae and Carabidae, there is little evidence of there being greater control of pest populations on farms containing field margins (Marshall and Moonan, 2002). A field study was therefore done to assess whether habitats of varying size (1.5 or 6 ha) and distribution (blocks or strips) that were planted within 100 ha areas on farms with the aim of enhancing biodiversity also increased the abundance and diversity of aphid natural enemies and whether there was an associated regulatory effect on aphid populations in crops within the 100ha area.

Aphid natural enemies within a 100 ha zone were surveyed on farms in East Anglia from the Farm4Bio LINK project (RD-2004-3137: Managing uncropped land in order to enhance biodiversity benefits of the arable farmed landscape). Fourteen farms were assessed in total, eight of which contained additional habitats (project-managed; 2 reps each of 1.5ha strips, 1.5ha block, 6ha strips, 6ha block), four farms followed basic ELS requirements for uncropped land (farmer-managed) and the remaining two farms were organic. The additional project-managed habitats comprised four zones: floristically enhanced grass (FEG), insect rich cover (IRC), an area of natural regeneration (NR) and winter bird seed cover (WBC). Arthropods were sampled in July from 2008-10 using an extension vortis sampler (as used in the sister LINK project) along eight transects in (1) the crops (2) other uncropped areas such as field margins (henceforth referred as uncropped areas) and in (3) project habitats where they were present. Transects were divided amongst crop fields and uncropped land in relation to the quantity of each crop/ uncropped type. For example if 50% of field had cereals, then 4/8 transects would be in separate cereal fields. Each transect was 50m long and 15 separate ‘sucks’ were made at regular intervals along the transect and pooled to provide a single sample per transect. Additional vortis samples were taken in the same locations on four of the farms (two farmer-managed and two project-managed; henceforth referred to as ‘time series farms’) in June 2008 and in May, June and August 2009-10. Three nettle patches were also assessed on each sampling occasion in the same way. The samples were frozen on the day of collection and subsequently identified. Visual observations of aphids and their natural enemies (parasitoids, fungal pathogens and predators) were done on the four time series farms in June and July (2008) and May, June, July and August (2009-2010). Visual observations were also made on five additional farms (including one organic farm) in July (2008-2010). In uncropped areas and project habitats these observations were made using a 25cm2 quadrat (five quadrats along a 50m transect walk). In the crop, assessments were taken at 15 positions along a 50m transect into the crop, with the amount assessed in each crop type being approximately equivalent in surface area; three cereal tillers, 1 bean plant, 2 pea stems, 2 sugar beet plants, 3 oilseed rape racemes, 3 stems of clover and vetch. Consequently the relative abundance of arthropods could be compared within sampling areas (crop or uncropped or project habitats) but not between sampling areas. In addition, parasitoid mummies and fungus-infected cadavers were collected during transect walks for identification (up to five per quadrat/ position). Sampled parasitoid mummies were maintained at 18 degrees until they emerged and could be identified whilst aphids infected with pathogenic fungus were placed on slides and stained with lacto-phenol cotton blue within 24h of collection. Data from the sister LINK project describing the proportion of land in the surrounding 3km that was arable, the amount of semi-natural land within the 100ha zone (land on which crops could not be planted), the percentage of the 100ha zone that was uncropped (land on which crops could have been planted but were not and instead managed in different ways. On project-managed farms this included the project habitats as well as other uncropped land areas) and the distribution of that uncropped land (perimeter of uncropped land / amount of uncropped land = P/A and provides an indication of how ‘strippy’ or ‘blocky’ the uncropped land was).

Summary of vortis data

Overall 283,128 arthropods were collected of which 41,061 were aphids, 30,103 were their natural enemies and 2,653 were hyper-parasitoids. The number of arthropods collected differed considerably between years, for example, 43,195 arthropods were collected in July 2008 whilst 71,713 were collected in the same month in 2009 and 63,712 in 2010. Aphid enemies were identified to genus, family or species where appropriate, which included 16 families of Araneae (spiders), 14 species of Hemiptera (bugs), 9 species of Neuroptera (lacewings), 34 species of Hymenoptera (parasitoids), 46 species of Coleoptera (beetles) and 18 species of Diptera (flies) were identified. Figure 2.1 summarises the mean numbers of aphids, Coleoptera, Diptera, Hymenoptera, Araneae and total enemies collected in vortis samples (per m2) over the three year sampling period. The remaining groups (Dermaptera, Hemiptera and Neuroptera) were not collected in large numbers and are therefore incorporated throughout, only in the total enemies’ category.

Figure 2.1. Mean number of (a) aphids in the crop, uncropped areas and habitats on organic, farmer- and project-managed farms; mean number of aphid natural enemies in (b) the crop and (c) uncropped areas on organic, famer- and project-managed farms; (d) within the habitats on project-managed farms. Values are per m2 and are means of three sampling years. Error bars represent standard errors of mean.

The greatest numbers of aphids within the crop were on farmer-managed farms and the fewest were on project-managed farms (figure 2.1a). In contrast, the greatest numbers of aphid natural enemies within the crop were on project-managed farms (figure 2.1b). Although Hymenoptera were considerably more abundant in crops on project-managed farms than on organic and farmer-managed farms there was little difference in the abundance of Coleoptera, Diptera and Araneae (figure 2.1b). Hymenoptera were the most abundant group of natural enemies in the project habitats (figure 2.1d). The abundance of Hymenoptera and total enemies in project habitats was approximately double that in uncropped areas on organic, farmer- and project-managed farms irrespective of the size and shape of the habitat (figures 2.1cd). Unlike the crop and project habitats, in uncropped areas Hymenoptera were not the dominant group of aphid natural enemies, indeed, the abundance of Araneae was slightly greater than that of Hymenoptera (figure 2.1c). The abundance of each group of natural enemies was similar in uncropped areas on organic, farmer- and project-managed farms (figure 2.1c).

Figure 2.2. Mean number of aphids and their natural enemies per m2 in each zone of the habitat (FEG = floristically enhanced grass; IRC = insect rich cover; NR = natural regeneration; WBC = winter bird cover).Values are means of three sampling years. Error bars represent standard errors of mean.

1st

2nd

3rd

%

Organic

Crop

A. ervi

P. volucre

A. rhopalosiphi

52.6

Non-crop

E. plagiator

D. rapae

A. rhopalosiphi

53.9

Farmer

Crop

A. rhopalosiphi

P. volucre

A. ervi

74.2

Non-crop

P. volucre

A. asychis

A. rhopalosiphi

56.4

Project

Crop

A. rhopalosiphi

P. volucre

A. ervi

84.2

Non-crop

A. rhopalosiphi

A. ervi

P. volucre

67.6

Habitats

P. volucre

A. ervi

A. rhopalosiphi

56.9

FEG

A. ervi

P. volucre

A. rhopalosiphi

68.4

IRC

P. volucre

A. ervi

A. rhopalosiphi

55.5

NR

P. volucre

A. rhopalosiphi

D. rapae

52.8

WBC

P. volucre

D. rapae

A. rho / A. ervi

58.8

Table 2.1. Summary of the three most abundant parasitoid species found on organic, farmer and project-managed farms and in the four zones within the habitats on project managed farms. The percentage of the total number of parasitoids accounted for by the three most common species within each category is also shown.

Within the project habitats there was little variation between the four zones in the abundance of Coleoptera and Diptera. However, aphids and Hymenoptera were considerably less abundant in the FEG than in the remaining three zones. In contrast, Araneae were most abundant in FEG. The abundance of Hymenoptera in the WBC was nearly double that in the IRC and NR, and nearly three times greater than in FEG. Overall 66% of the aphid natural enemies found in the project habitats were Hymenoptera. However, hymenopteran parasitoids that oviposit in aphids vary in their host specificity and not all of the parasitoid species that were collected and categorised as aphid natural enemies were also enemies of aphids found on crops. However, the three most abundant parasitoids collected from the crop, uncropped areas and project habitats on organic, farmer and project-managed farms were all enemies of cereal aphids (table 2.1).

· Aphid abundance in the crop was greatest on farmer-managed farms whilst natural enemy abundance was greatest on project-managed farms

· Hymenoptera were the most abundant group in the crop and project habitats, with the three most common species being enemies of cereals aphids

· The abundance of aphid natural enemies in project habitats was approximately double that on other uncropped areas, with most in the WBC.

Analysis of vortis data: Part 1 - Abundance

Unlike the sister LINK project which assessed 28 farms, this project was limited to sampling only 14 farms in the east of England. Replication was therefore insufficient (two replicates of each of the seven treatments) to carry out a statistical analysis using the factorial treatments implemented on the farms. In fact, the high variability associated with field studies has also prevented the sister LINK project from using a factorial analysis. Therefore, a GLMM analysis that contained both factorial and variate information describing the land use both within and surrounding the 100ha block was used by both projects to determine which parameters affected the abundance of aphids and their natural enemies (model = year, organic, arable land in the surrounding 3km, amount of semi-natural land within 100ha zone, percentage of 100ha zone that was uncropped, management (project, farmer/ organic) and distribution of uncropped land (Perimeter/Area = P/A). Table 2.2 summarises the outcome of the GLMM for the crop data.

Crop

Year

Organic

Arable

Semi-natural

% uncropped

Management

P/A

Aphids

p<0.05 (10)

-

-

-

-

p<0.05 (F)

-

Coleoptera

p<0.001 (09)

-

-

-

p=0.054 (-ve)

-

-

Diptera

-

-

-

-

-

-

-

Hymenoptera

p<0.001 (09)

p<0.05 (S)

-

p<0.05 (+)

p<0.05 (-ve)

p<0.01 (P)

-

Araneae

p<0.05 (09)

-

-

-

-

-

-

Total enemies

p<0.001 (09)

-

-

-

-

p=0.075 (P)

-

Table 2.2. Summary of GLMM (distribution = Poisson; link = logarithm; random variable = site) for abundance on the crop. Significant results are shown where appropriate for individual terms of the model. Where significant the year (08, 09, 10), farm type (organic = O; standard = S) and management (farmer/ organic = (F); Project = (P) are indicated for factorial terms whilst positive (+ve) or negative (-ve) effects on abundance are shown for variate terms. High P/A values indicate strips and low values blocks.

Crop: There were significantly more aphids in the crop in 2010 than in the previous two sampling years whilst Coleoptera, Hymenoptera, Araneae and the total number of enemies were all more abundant in 2009. Aphids in the crop were strongly influenced by how the farm was managed with significantly more present on farmer-managed and organic farms than project-managed farms. In contrast, total enemies in the crop approached being significantly more abundant on project-managed farms than farmer- managed and organic farms. There were no effects of the remaining terms in the model on the abundance of these two groups. Of all the groups of natural enemies the terms in the model best described the abundance of Hymenoptera with significantly more collected on crops from project-managed farms with high levels of semi-natural land and low proportions of total uncropped areas (i.e. project habitats plus other uncropped land) compared with farmer-managed and organic farms. There was also a significant negative effect of increasing the total proportion of uncropped land on the abundance of Coleoptera in the crop.

Uncropped areas: Aside from year, there were no other significant effects of the terms in the model on the occurrence of aphids and their natural enemies in uncropped areas. The only exception was for Hymenoptera which, as in the crop, were positively correlated with an increase in semi-natural land.

· Aphids were significantly more abundant on crops on farmer-managed farms

· Hymenoptera were most abundant on project-managed farms in the crop when there were high levels of semi-natural land and low proportions of total uncropped land

· There was little effect of farm management or land use within the 100ha study site or the surrounding 3km on aphid natural enemies in uncropped areas

Analysis of vortis data: Part 2 - Species richness and diversity

Both the species richness (S) and Shannon diversity (H’) was calculated for the major groups of arthropod natural enemies. Values were calculated at the level of family for Araneae and species for the remaining groups. The results are summarised in table 2.3. From this summary it can be seen that the species richness of Hymenoptera and Araneae is greatest in the project habitats and uncropped land respectively. These both exceed their species richness in crops. The GLMM analysis described above was used to determine which terms affected the species richness of aphid natural enemies in the crop and uncropped land.

Crop: Aside from year, there were few significant effects of the terms in the model on the species richness (S) of aphid natural enemies in the crop. The exceptions were Diptera, whose species richness was significantly greater on farmer-managed farms than the other farms and for Hymenoptera which were negatively correlated with the amount of uncropped land. There was little effect of farm management or the remaining terms in the model on diversity (H’) in the crop. However, the diversity of Hymenoptera in the crop decreased as the amount of semi-natural land increased

Araneae

Coleoptera

Diptera

Hemiptera

Hymenoptera

Neuroptera

S

H’

S

H’

S

H’

S

H’

S

H’

S

H’

Crop

Organic

3.50

0.89

4.50

0.94

1.67

0.69

1.67

0.48

5.50

1.52

0.67

0

Farmer

3.00

0.73

3.50

0.93

1.33

0.49

0.75

0.31

6.17

1.37

0.67

0.18

Project

2.63

0.55

3.25

0.85

0.58

0.11

0.58

0.23

6.04

1.16

0.63

0.15

Uncropped

Organic

6.50

1.46

3.0

0.76

2.33

0.71

2.67

0.77

6.67

1.44

0.83

0.14

Farmer

5.42

1.27

2.17

0.66

1.75

0.69

1.50

0.54

5.42

1.30

0.50

0.27

Project

5.96

1.33

3.67

1.04

0.92

0.41

1.50

0.46

6.21

1.23

0.42

0.30

Habitats

1.5ha block

4.00

0.97

3.00

0.58

1.00

0.27

2.67

0.97

9.50

1.61

0.83

0

1.5ha strip

4.17

0.98

3.00

0.78

1.17

0.27

2.17

0.46

8.67

1.36

0.50

0

6ha block

4.00

0.90

2.67

0.83

1.33

0.48

1.67

0.54

8.83

1.50

0.67

0

6ha strip

5.83

1.30

2.83

0.73

1.33

0.17

2.83

0.66

8.00

1.70

0.33

0

Table 2.3. Mean richness (S) and Shannon diversity (H’) of families of Araneae and species of Coleoptera, Diptera, Hemiptera, Hymenoptera and Neuroptera collected from the crop and uncropped areas on organic, farmer and project managed farms and from the habitats on project managed farms.

Uncropped

Year

Organic

Arable

Semi-natural

% uncropped

Management

P/A

Araneae

p<0.001 (10)

-

-

-

p<0.05 (+ve)

-

-

Coleoptera

p<0.001 (09)

-

-

p<0.001 (+ve)

p=0.052 (+ve)

-

-

Diptera

p<0.001 (09)

p=0.066 (O)

-

-

-

p<0.05 (F)

-

Hemiptera

-

-

-

-

-

-

p<0.05 (+ve)

Hymenoptera

p<0.001 (09)

-

-

-

-

-

-

Neuroptera

p<0.001 (09)

-

-

-

-

-

-

Table 2.4. Summary of GLMM (distribution = Poisson; link = logarithm; random variable = site) for species richness in uncropped areas. Significant results are shown where appropriate for individual terms of the model. Where significant the year (08, 09, 10), farm type (organic = O; standard = S) and management (farmer/ organic = F; Project = P) are indicated for factorial terms with the greatest species richness whilst positive (+ve) or negative (-ve) effects on species richness are shown for variate terms.

Uncropped areas: The terms in the model were better able to describe species richness (S) in uncropped areas (table 2.4). There was a positive effect of increasing the percentage of uncropped land on the species richness of Araneae and a negative effect of increasing P/A on the species richness of Hemiptera. There was a positive effect of increasing the amount of semi-natural land and uncropped land on the species richness of Coleoptera. As in the crop, the species richness of Diptera was significantly greater on farmer-managed farms. There was little effect of farm management or the remaining terms in the model on diversity (H’) in uncropped areas. However, the diversity of Hemiptera in uncropped areas increased as P/A increased.

· Parasitoid diversity was greatest in new project-managed habitats

· There was little effect of farm management or the remaining terms in the model on species richness or diversity in the crop or uncropped areas

Analysis of vortis data: Part 3 - Time series

Crop: Although time constraints only allowed monthly vortis samples to be taken over the field season, clear trends can be observed in the crop data (figure 2.3). On farmer-managed and project-managed farms the abundance of aphids in the crop increased at a similar rate in May and June, however, in July the rate of increase on farmer-managed farms increased substantially whilst on project-managed farms the rate of increase slowed. In contrast, the abundance of Hymenoptera and total enemies showed the opposite trend, with numbers on project-managed farms increasing at a greater rate in July whilst on farmer-managed farms the rate of increase slowed (figure 2.3). However, this trend was not observed for Araneae which increased at a similar rate on farmer- and project-managed farms throughout the field season.

Figure 2.3. Mean number of aphids and their natural enemies per m2 (log10 transformed) on the crop on two farmer managed farms and two project managed farms.

Uncropped areas: In uncropped areas on project- and farmer-managed farms the abundance of aphids and their natural enemies remained stable throughout May, June and July before aphid populations crashed in August (figure 2.4).

Project Habitats: In the project habitats the abundance of aphids and their natural enemies increased considerably during May, June and July and, whilst the aphid population crashed in August, the abundance of natural enemies remained stable. In addition, whilst the abundance of aphids and total enemies was considerably lower in the project habitats than in uncropped areas in May, numbers were similar to that of uncropped areas in June and by July the abundance was greater than that in uncropped areas.

Figure 2.4. Mean number of aphids and their natural enemies per m2 (log10 transformed) in uncropped areas and in the habitats on two farmer-managed and two project-managed farms.

· Aphid abundance increases at a slower rate and natural enemy abundance at a faster rate in crops on project-managed farms compared with farmer-managed farms over the field season

· There was no effect of farm management on the abundance of Araneae in crops over the field season

· The abundance of aphids and their enemies was lower in habitats than in uncropped areas at the start of the field season but was greater in July

Visual observations

Project- and farmer-managed farms: The results of the visual observations made in July 2008-10 where sampling was done using quadrats are summarised in figures 2.5 to 2.7 for the four project- and four farmer-managed farms. Visual observations were only taken on one organic farm so these data are summarised separately. Overall the trends observed in the visual observations were similar to those in the vortis data with one major exception: whereas the vortis data shows that the abundance of Hymenoptera in the crop was greater on project-managed farms than on farmer-managed farms, the opposite was shown in the visual observations, with Hymenoptera being more abundant in both the crop and uncropped areas on farmer-managed farms (figure 2.5bc). Unlike the other groups of natural enemies in which the same life stage was assessed with both sampling techniques, Hymenopteran mummies (pupae) were assessed when doing visual observations whereas only adult Hymenoptera were assessed when using the vortis. The vortis data therefore shows the abundance of parasitoids visiting or emerging from the crop whereas the visual observations and emergence data shows their reproductive activity in the crop. It is therefore unsurprising that the abundance of Hymenopteran mummies was greatest on farmer-managed farms where there were also more aphids (figure 2.5a). To take into account the number of hosts available the ratio of hymenoptera to aphids was calculated as an indicator of parasitism efficiency. The ratio of Hymenoptera to aphids in the crop on project-managed farms (0.7) was greater than on farmer-managed farms (0.35), which supports the findings of the vortis data and suggests aphid population regulation is occurring. However, in uncropped areas the ratio of Hymenoptera to aphids was greater on famer-managed farms (0.47) than on project-managed farms (0.19).

Organic farms: As with the farmer- and project-managed farms, the results of the visual observations on the organic farm are similar to those in the vortis data. However, the visual observations showed that the abundance of Hymenoptera on the crop was even greater on organic farms than on farmer-managed farms, in particular on the clover and legume crops where aphids were also extremely abundant. Indeed, 97% of Hymenoptera and 96% of the total enemies observed in the crop on organic farms were in clover fields.

Figure 2.5. Mean number of (a) aphids in the crop, uncropped areas on farmer and project managed farms and within the habitats on project managed farms; mean number of aphid natural enemies in (b) the crop and (c) uncropped areas on farmer and project managed farms; (d) within the habitats on project managed farms. Values are per m2 and are means of three sampling years. Error bars represent standard errors of mean.

Parasitoid emergence: Part 1 – Direct counts

In total 800 mummies were collected from the nine farms in May, June and July over the three sampling years, 59% of those that emerged contained primary parasitoids whilst 41% contained hyperparasitoids. Figure 2.6 shows the number and species of parasitoids that emerged and where they were collected from. It should be noted that sampling effort was not the same for all sites and this would have influenced the number of parasitoid mummies collected i.e. more cereal fields were sampled than legume fields therefore the greater number of mummies collected in cereal fields may have been due to the increased sampling effort, not the crop per se. Therefore in the count data the relative abundance of each species cannot be compared between crop types or areas of uncropped land but can be compared between different zones of the project habitats where sampling effort was always the same. To assess the effect of crop and uncropped habitat type on the emergence of cereal aphid parasitoids, the raw data were then adjusted to show the percentage emerging if sampling had been equal in each crop or area of uncropped land type (figure 2.7).

Overall the species of parasitoids that emerged from mummies collected in the crop, uncropped areas and habitats were similar to those that were most abundant in the vortis samples. However, there are some exceptions, for example, Praon volucre (Haliday) was one of the most common species collected in uncropped areas and project habitats in vortis samples but was relatively uncommon in the emergence data. This may suggest that either P. volucre uses trophic resources in uncropped areas and project habitats but does not reproduce there, or that signs of reproduction in the form of mummies were not yet evident at the time of visual sampling. In contrast, Aphidius funebris (MacKauer), a parasitoid of aphids infesting non-crop plants such as thistles and knapweed, was the most common species emerging from mummies in the project habitats and third most common in uncropped areas but was not one of the three most common species in the vortis samples. This too might be down to the timing of sampling and temporal asynchrony between adults and pupae or it might mean that while this species is the most common species reproducing in the habitats, numbers emerging and collected in the vortis are small in comparison with the large numbers of other Hymenopteran species caught in the vortis while visiting the habitats for trophic resources.

If we focus on cereals as the primary crop, of those primary parasitoids emerging in uncropped areas and habitats, 59% and 35% were enemies of cereal aphids respectively. Although parasitoids were more abundant in the winter bird seed cover than the remaining three habitats in the vortis samples, only 20% of the parasitoids emerging from mummies collected in this zone were enemies of cereal aphids whilst 92% of parasitoids emerging from mummies collected in the FEG were enemies of cereal aphids (Figure 2.6c). The most abundant parasitoid species in cereal crops, and therefore a key species to encourage in habitats, is A. rhopalosiphi. It was also the most abundant parasitoid species in reproducing in uncropped areas but in the habitats it was only common reproducing in FEG and overall only the third most common species behind A. ervi, an enemy of cereal and legume aphids that was relatively uncommon in cereal crops, and A. funebris, which was absent from cereal crops. Within the uncropped land A. rhopalosiphi was specifically associated with unmown grass margins (fig. 2.6 and 2.7) suggesting that it is important to have some unmown grass for reproduction of this species. Because sampling effort was not the same it is difficult to determine whether unmown grass or FEG supported more A. rhopalosiphi mummies. The most abundant parasitoid species in legume crops was A. ervi and in brassica crops was D. rapae and P. volucre. While A. ervi and P. volucre can also reproduce on cereal pest aphids they are clearly more important as regulators in legume crop aphids. In terms of which habitats supported reproduction of these species, D. rapae reproduced in the WBC and A. ervi in the FEG.

Hyperparasitism was a common occurrence with 41%, 50% and 35% of parasitoid mummies collected in the crop, uncropped areas and habitats respectively containing hyperparasitoids. The majority of hyperparasitoids collected in the crop and all of those collected in uncropped areas and project habitats were capable of parasitising cereal aphids that had been parasitised by primary parasitoids.

Parasitoid mummies were also collected from nettle patches. Aphidius microlophii (Pennacchio & Tremblay) emerged from 33 of the 47 mummies whilst Aphidius urticae (Haliday) emerged from three mummies. The remaining 11 mummies contained hyperparasitoids (Asaphes vulgaris (Walker) = 5, Dendrocerus carpenteri (Curtis) = 2, Alloxysta = 1, Dendrocerus aphidum (Rondani) = 1, Dendrocerus dubiosus (Kieffer) = 1, Coruna clavata (Walker) = 1). None of the primary parasitoids emerging from aphids collected in nettle patches were able to parasitise cereal aphids whilst all of the emerging hyperparasitoids, with the exception of D. dubiosus, were able to attack parastised cereal aphids.

Figure 2.6. Number and species of primary and hyperparasitoids emerging from aphids collected in (a) the crop, (b) uncropped areas and (c) habitats (* = primary parasitoids in uncropped areas and habitats that have been recorded as being enemies of cereal aphids in the UK; ** = hyperparasitoids of cereal aphids). Values are totals of three sampling years.

Parasitoid emergence: Part 2 – Adjusted data

These results support the findings of the direct counts and show that the most important parasitoid reproducing in pest aphids on cereal crops is A. rhopalosiphi, and that P. volucre and A. ervi (the two other most common parasitoids collected from crops in vortis samples) reproduce much more on legume and brassica crop aphids, and less so on cereal aphids (figures 2.6a & 2.7a). In uncropped areas the count data and the adjusted data were similar (figures 2.6b & 2.7b).

Figure 2.7. Percentage emergence of cereal aphid parasitoids from (a) the crop and (b) uncropped areas. The data has been adjusted to account for unequal sampling between crop types or areas of uncropped land.

Parasitoid emergence: Part 3 – Direct counts vs. vortis samples

By comparing the vortis data and the direct counts of parasitoid emergence it is possible to determine which parasitoid species were utilising the crop, uncropped areas and habitats for non-host resources such as nectar but not for reproduction (i.e. species that were present in the vortis samples but were not present in the emergence samples). Aphidius rhopalosiphi, a cereal aphid specialist, was found in the vortis samples and emerging from mummies collected from the crop, uncropped areas and habitats. This indicates that A. rhopalosiphi utilised all these areas for reproduction and potentially for non-host resources (figure 2.8a). In addition, the number of A. rhopalosiphi emerging from mummies collected in the habitats increased in July suggesting that these areas may provide suitable host aphids when the crop and uncropped areas are in decline. The planted habitats may therefore have a positive effect on the number of A. rhopalosiphi that are able to successfully overwinter. Similar trends were found for A. ervi, a semi-specialist parasitoid of cereal aphids. Although the generalist parasitoid P. volucre was also found in the vortis samples and emerging from mummies collected from the crop, it did not emerge from mummies collected from uncropped areas and was only found emerging in low numbers from samples collected from the habitats in July (figure 2.8b). Praon volucre therefore appears to use uncropped areas and the planted habitats for the non-host resources they provide and rarely for reproduction. This suggests that whilst uncropped areas and planted habitats could increase the abundance and overwintering success of specialist cereal aphid parasitoids they may have little effect on the abundance of generalist parasitoids.

Figure 2.8. Number (log n+1) of A. rhopalosiphi and P. volucre in vortis samples or emerging from mummies collected from the crop or uncropped areas on the four time series farms and from the habitat on the project managed farms

Fungal cadavers

Overall fungal cadavers were most abundant on farmer-managed farms, with the abundance in the crop in July being greater than that in the uncropped areas (figure 2.9a). As with Hymenoptera, when sampling fungal cadavers it is the reproductive activity of the fungus that is being assessed and this is dependent on the abundance of aphids. Indeed, the abundance of fungal cadavers in the crop and habitats follow the same trends as aphid abundance in these areas (figures 2.5a, 2.9a). Over the three month sampling period fungal cadavers were most abundant in nettle patches and least abundant in uncropped areas (figure 2.9b). However, the most rapid increase in abundance of fungal cadavers occurred in the habitats between June and July. Although the number of fungal cadavers on nettles on project-managed farms was less than on farmer-managed farms in May, numbers continued to rise over the three month sampling period on project-managed farms whereas a decrease in abundance was observed on farmer-managed farms in July (figure 2.9b). The abundance of fungal cadavers in the crop over the three month sampling period was similar on farmer- and project-managed farms (figure 2.9b).

Figure 2.9. Mean number of fungal cadavers (logn+1) in (a) the crop and uncropped areas on farmer- and project-managed farms and the habitats on project-managed farms; (b) on farmer- (dashed lines) and project-managed farms (solid lines) and the habitats on project-managed farms (dotted line) over three months. Values are means of three sampling years. Error bars in (a) represent standard errors of mean.

Having both vortis data and visual counts and identifying to species level really aids interpretation. It is important to know which species may largely only be using habitats for trophic resources (e.g. P. volucre) and which also reproduce in some habitats. This can guide selection of plants for use in managed habitats to provide trophic and reproductive resources for key enemy species. Pooling all Hymenoptera within ‘Parasitica’ can be misleading as not all Parasitica are enemies of pest aphid species, as is the case for A. funebris reproducing on aphids in WBC.

· The trends observed in the visual observations are similar to those in vortis data for most arthropod groups assessed

· Although parasitoid mummies and fungal cadavers were more abundant on crops from farmer-managed farms where aphid abundance was also greater, the ratio of parasitoids to aphids was greater on project-managed farms

· Parasitoids emerging from mummies collected in uncropped areas and crop were also capable of parasitising cereal aphids, however, hyperparasitism was common

· The habitats were utilised by cereal aphid parasitoids for reproduction whereas some parasitoid species only appear to utilise them for non-host resources

· Project habitats and nettles are a good reservoir of fungal cadavers

Redundancy analysis

Whilst the GLMM of the July vortis samples allows multiple parameters to be included in a model, it is limited to assessing a single response variate e.g. aphids per m2. However, when assessing aphids and their natural enemies the presence or absence of one group of insects may have an impact on a second group e.g. the presence of aphids may affect whether coccinellids are present. Therefore a redundancy analysis of the vortis data is being done to assess multiple response variates (aphids and their natural enemies) along with factorial terms (management and crop combinations) and variate terms (farm characteristics). Figure 2.10 shows the output of the analysis for the crop data for 2010. In this example two dimensions (RDA-1 and RDA-2) explained 69.5% of the variance. Several trends can be observed. For example, Hymenoptera and Coleoptera are associated with each other and not associated with Diptera and spiders. Coleoptera and aphids are associated with each other whilst spiders and aphids are disassociated. High levels of aphids are most closely associated with particular farms, all of which are either farmer- or organically-managed farms. In contrast, high levels of Hymenoptera are most closely associated with three different farms, all of which are project-managed farms. The binary centroids for project and organic-managed farms are similar but differ considerably to farmer-managed farms. As part of future research the analysis could now be modified to combine the data sets from 2008, 2009 and 2010 and the factorial and optimise variate terms to allow more detailed and accurate interpretation of the data.

Figure 2.10. RDA ordination plot of the crop vortis data for the 14 farms sampled in 20010. Aphids (Aph) are shown along with spiders (Spi), Hymenoptera (Hym), Coleoptera (Col) and Diptera (Dip). The farm characteristics are shown as environmental variables whilst the management and crop combinations are shown as centroids of binary variables.

Summary of field studies

Th