General enquiries on this form should be made to: Defra, Procurements and Contracts Division (Science R&D Team) Telephone No. 0207 238 5734 E-mail: [email protected]

SID 5 Research Project Final Report

 Note In line with the Freedom of Information Act 2000, Defra aims to place the results Project identification of its completed research projects in the public domain wherever possible. The PS2113 SID 5 (Research Project Final Report) is 1. Defra Project code designed to capture the information on the results and outputs of Defra-funded 2. Project title research in a format that is easily A framework for the practical application of publishable through the Defra website. A semiochemicals in field crops SID 5 must be completed for all projects. This form is in Word format and the boxes may be expanded or reduced, as 3. Contractor Rothamsted Research appropriate. organisation(s)

 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 54. Total Defra project costs £ 422,765 project. Defra may also disclose the (agreed fixed price) information to any outside organisation acting as an agent authorised by Defra to 01 April 2006 process final research reports on its 5. Project: start date ...... behalf. Defra intends to publish this form

on its website, unless there are strong end date ...... 15 May 2010 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.

SID 5 (Rev. 05/09) Page 1 of 24 6. It is Defra‟s intention to publish this form. Please confirm your agreement to do so...... YES NO (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. The main objective of this project was to develop a framework for the practical control of in the key combinable crops used in arable rotations, i.e., cereals, oilseed rape and legumes, based on semiochemicals. These are volatile non-toxic -produced (pheromones) or plant-produced chemicals that act as information signals between organisms and cause behaviour changes in the recipient. As such, they can be used to manipulate the distribution of both pest and beneficial populations and have the potential to reduce the need for toxicant insecticides in pest management. The objective was addressed by two sub-objectives, one focused on developing a semiochemical based management strategy for oilseed rape and the other to develop this approach for cereal and legume cropping systems utilizing Defra-recommended field margin mixtures.

Devise and test a semiochemical based management strategy for pests of oilseed rape This sub objective was addressed by five, replicated field experiments conducted in years 1-3 of the project, and by an additional 3 experiments in an extension year. The overall aim was to develop and test a „push-pull‟ strategy for oilseed rape (OSR). In such a strategy, the main crop is made as unattractive as possible to pests through the use of less attractive cultivars and repellents („push‟) while simultaneously, attractive trap crops are used to deter pests from colonizing the main crop and to concentrate them in the trap crop area where they can be more easily controlled („pull‟). In our previous work (PI0340 and PS2107) we demonstrated in a model spring cropping system that turnip rape (Brassica rapa) was more attractive to several insect pests of OSR than was the crop itself, and that a trap crop border of turnip rape could reduce pest populations in the main crop to below spray threshold levels. In the project reported here, we scaled-up field experiments from the relatively small plots (30 x 30 m) used in previous experiments to a more realistic whole-field scale. Six, 1ha crops of spring OSR three with a 9m-wide turnip rape trap crop bordering the crop and three without a trap crop, were sown in separate fields, with each at least 500m from any other treatment or OSR crop. Adult pests were sampled at regular intervals on plants selected at random in the centres and borders of each plot from the green bud stage to full-flowering. Unfortunately, two out of the three replicates without the trap crop failed to establish, so statistical power was limited for comparisons. However, it was clear that pollen , Meligethes aeneus, were more attracted to the turnip rape than the OSR plants in the border, and at the damage susceptible stage of the main crop the trap crop reduced the number of beetles per plant in the main crop compared with plots without a TR trap crop. The trap cropping strategy using the model spring OSR cropping system was successfully transferred to the winter OSR cropping system. Nine different oilseed/turnip rape lines were compared in a replicated field experiment and adult pest infestation was sampled through the growing season. Pasja fodder turnip rape was significantly more attractive than any other line tested, and flowered early, thereby showing good

SID 5 (Rev. 05/09) Page 2 of 24 potential as a trap crop. Two OSR cultivars, Grizzly (late-flowering) and Cheops (white-petalled), showed potential for use as a main crop as they were less preferred by pollen beetles than other lines in the study. These were further tested in a replicated field experiment (30 x 30 m plots) both with and without a Pasja turnip rape trap crop. Plants were randomly sampled from both the borders and centres of plots throughout the bud-flowering phase of the crop. Pasja turnip rape plants were significantly more infested than OSR plants in the plot borders. This supports our results from the screening trials and indicates that Pasja turnip rape is more attractive than oilseed rape to pests – a requirement for an efficient trap crop. Furthermore the experiment showed that Pasja performed effectively as a trap crop when the main crop was the late-flowering OSR cv. Grizzly. Numbers of beetles were significantly lower in the main crop of plots with a trap crop than in plots without a trap crop. However, this was not the case when the main crop was Cheops, possibly because it flowered earlier than Grizzly. Similar results were gained from a preliminary experiment which scaled up the trap cropping strategy for winter OSR to more realistic field scales (fields of 1ha OSR cv. Grizzly with or without 9m-wide trap crop borders of Pasja turnip rape). It was also found in these large-scale experiments that application of insecticide to the trap crop further improves its efficacy. We conclude that the trap cropping strategy can reduce to below spray threshold the numbers of pollen beetles in crops of OSR and holds great potential to reduce the number of sprays required to control the pest and as a tool in insecticide resistance management for OSR pests. We suggest that further experiments would be required to ensure that the same sized trap crops still work when field size is 10-100ha, as more commonly found in the UK. In our previous work in PS2107 we identified lavender essential oil as a potential repellent against pollen to use as the „push‟ element in the push-pull strategy. In the work reported here, a slow-release spray formulation was developed using gum acacia as a base. In laboratory experiments, an effective rate was determined (4l lavender oil + 40ml Ethylan BV and 6kg gum acacia in 200l water/ha) and the periodicity over which it was likely to be effective was determined (24h). The most effective number of applications was determined in replicated field trials on winter OSR. The lavender significantly reduced the infestation of pollen beetles on treated plots compared with untreated controls and two applications were significantly more effective than one. This work was taken forward and lavender was used as the „push‟ component of a push-pull strategy with a turnip rape trap crop „pull‟). Replicated large-scale (1ha) plots of spring OSR were established to compare the efficacy of the „push‟, „pull‟ and „push+pull‟ elements of the strategy compared with an untreated control. Unfortunately, one of our field sites failed to establish so we lacked the necessary statistical power required for robust analyses. Although it was clear that the trap crop „pull‟ was effective, there was no difference in the number of beetles per plant in control plots and plots treated with the lavender repellent. We suggest that a repeat of this experiment is necessary to demonstrate the efficacy of a push-pull strategy in oilseed rape.

Devise and test a semiochemical based management strategy for pests of cereal and legume crops in a conventional arable rotation. Large, replicated habitat management plots (HMP) consisting of Defra recommended field margin mixtures „TG2‟ (grass mixture only) and „WM2‟ (pollen/nectar mix including legumes), which had been established in project PS2107, were further trialled with winter wheat and field bean crops during the course of this project. The HMP, one of which - WM2 - had been shown in PS2107 to be a good source of some beneficial species including aphid parasitoids and predatory beetles, were considered as possible components of “push-pull” strategies, in combination with semiochemical treatments to the crops, for control of pests in an arable crop rotation. The HMP and crop plots were sampled using a range of techniques throughout the experimental period to determine the levels of pest populations and their natural enemies. The same semiochemical treatments, delivered from slow release PVC rope dispensers, were used in years 1 and 2 on wheat and beans, respectively. One treatment was the plant activator cis- jasmone (CJ), which is both repellent to pest aphids and attractive to parasitoids and predators as well as inducing defence chemistry in the treated plants, and the other was a combination of the aphid sex pheromone components nepetalactone and nepetalactol, which attract aphid parasitoids and predators. In year 3, the treatments were spray formulations of CJ and CJ plus a synergist piperonyl butoxide (PBO), which deactivates the metabolic enzymes that aphids use to detoxify natural plant defence chemistry. Results were very similar for both crops in years 1 and 2. In the wheat, numbers of the main target pests, cereal aphids, were low. Metopolophium dirhodum was the predominant species. There were no significant effects of treatment on aphids, although numbers were fewer on CJ treated plots confirming a trend observed previously in PS2107. In the beans, infestation by the pea aphid, Acyrthosiphon pisum, was large and as in the wheat, there were no significant treatment effects, but numbers in the CJ plots were lower. Also, numbers of the pea and bean , lineatus caught in treated plots were lower than in controls or the HMP. Numbers of adult aphid parasitoids, caught on yellow sticky traps, were greatest in the treated plots compared to the controls or the HMP for both crops. Additionally, there were significantly more carabid beetles in pitfall traps in the crop plots compared to the HMP. Numbers of other pest and beneficial species were too low to detect any effects. In year 3, the CJ formulation reduced overall aphid numbers consistently, but after an initial effect, aphid numbers on the PBO/CJ treatment were no different to those on untreated plots. Within the HMP, in year 1 significantly more predatory Tachyporus beetles and adult hoverflies were

SID 5 (Rev. 05/09) Page 3 of 24 found in the pollen/nectar mix including legumes (WM2) plots than in the grass mix (TG2) plots, although this difference declined in year 2. In years 1 and 2, the TG2 plots were found to be sources of two pest species, the yellow wheat blossom midge, Contarinia tritici, and S. lineatus compared to WM2. However, by year 3 there were no differences between the HMP due to the changes in botanical composition of the WM2 plots, which gradually became dominated by grass species. Although the WM2 mixture showed initial promise as a source of natural enemies for possible manipulation with semiochemicals, this gradual change in botanical composition reduced its potential. Although differences were not significant, the treatments demonstrated that populations of pests and some natural enemies could be influenced by semiochemicals. Therefore, if suitable stable margin mixtures were available there could be greater scope for their inclusion in pest management strategies.

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).

SID 5 (Rev. 05/09) Page 4 of 24 A recent socio-economic study performed as part of a RELU-funded project (1) indicated that arable growers do make use of a variety of IPM techniques and many would consider using semiochemical-based methods if they were available (2). To this end, the experimental work in this project was focused on developing a semiochemical-based management strategy for oilseed rape (Objectives 1 & 3) and for cereal and legume cropping systems (Objective 2).

OBJECTIVE 1. Devise and test a semiochemical based management strategy for pests of oilseed rape (OSR) Devise and test, on a realistic agricultural scale, a semiochemical based strategy for the pests of the representative brassica crop SOSR. The effects of the strategy on non-target species, particularly natural enemies and other beneficial insects will be defined and quantified. Spring rape will be targeted initially to build on previous experience and maximise natural occurrence of coleopterous pests of the inflorescence stage, but the study will extend to winter rape in the autumn of year 1.

01/01 Milestone 1. Develop the Push-pull strategy for spring OSR Plants have natural defence systems which help to protect them from herbivory. Exogenous application of certain chemicals including salicylic acid (SA) and jasmonic acid/methyl jasmonate (MeJa) has been shown to artificially induce the defences in plants. Such technology has great potential for pest management. Previous work at Rothamsted and elsewhere has shown that application of MeJa to Brassicas induces the production of indolyl glucosinolates (defence compounds which help protect Brassicas from generalist herbivores) (3) and that application of SA induces production of alkyenyl glucosinolates (these defence compounds break down upon plant damage to produce volatile isothiocyanates, particularly 2-phenylethyl isothiocyanate, which are highly attractive to Brassica specialist pests) (4). A push-pull strategy (5) using plant defence-inducing chemicals was tested on a SOSR model. The hypothesis that application of MeJa to an OSR crop would make it less attractive to pests, in particular generalist pests, and that application of salicylic acid to turnip rape (TR) trap crops would make them more attractive, particularly to the specialist pests, was tested in the lab in PS2114 (see final report for details). These studies were scaled-up to field plot scale in the current study.

Large plots (30m x 30m) of SOSR, either with or without a 3m wide border of a TR trap crop, were established to test push and pull elements of a push-pull IPM strategy. Six treatments were tested in a randomized block design with 4 replicates: (1) SOSR untreated with no trap crop (control); (2) SOSR with no trap crop treated with pyrethroid insecticides to control pollen beetles, Meligethes aeneus, and seed , Ceutorhynchus assimilis, (standard farm practice treatment); (3) SOSR with a TR trap crop border, both untreated (standard pull treatment); (4) SOSR untreated with a border of TR treated with the plant activator, SA, to enhance attraction of the trap crop to the main coleopterous pests (enhanced pull treatment); (5) SOSR with untreated turnip rape border. The SOSR treated with the plant activator, MeJa to reduce attractiveness of the crop to generalist pests (pest reduction treatment); (6) SOSR treated with MeJa with a TR trap crop border treated SA (full push-pull treatment).

Methyl jasmonate was applied at 46g a.i. (active ingredient) per ha formulated with 0.1% wetter (Gex 1664) in a volume of 200 l/ha. This was based on the least attractive formulation used in experiments in PS2114. Salicylic acid was applied at 28g a.i. formulated with 0.1% wetter (Gex 1664) in a volume of 200 l/ha. Again, this rate was based on the most attractive rate from lab experiments in PS2114. Both plant activators were applied on 9/6/0/2006. Pyrethroid insecticide (Hallmark with Zeon technology, 75ml/200l/ha) was applied on 14/6/2006.

Assessments of numbers of adult pests on the plots were made before, and four times after treatment, to compare effects of the treatments on pest distribution. On each occasion, 20 plants were selected at random from each of the plot centres along a „W‟-shaped transect, with 5 plants sampled on each leg of the transect. Twenty plants were also sampled from the plot borders (5 plants selected at random from each of the four borders). The main raceme from each plant sampled was beaten onto a white tray and the number of pest insects (pollen beetles, stem weevils, seed weevils, and aphids) was counted and recorded. To assess direct damage by pollen beetles, 20 main racemes of OSR were sampled at random from the centre of each plot at the susceptible green-yellow bud stage of the crop. Samples were returned to the lab and the number of buds on each raceme was counted and observed for oviposition and feeding damage. Damaged buds were dissected and the number of buds that were selected for oviposition (i.e. had eggs or larvae inside) or rejected for oviposition (i.e. were empty) were recorded. The number of blind stalks on each raceme was also recorded.

Only pollen beetles were found in sufficient numbers for meaningful analysis. In the crop borders, overall there was a significant effect of treatment on the mean number of beetles per plant (ANOVA F5, 20 =4.55, P<0.001): there were significantly more beetles on TR than OSR plants in the border. This supports our previous work and suggests that TR is more attractive than OSR and therefore has potential for use as a trap crop. Spraying the TR border with SA had no significant effect. In the OSR centres of the plots, there was no difference between treatments on the first assessment (pre-treatment), or on the 2nd or 5th assessments (ANOVA P>0.05) (Fig 1). On both the 3rd and 4th assessments, at the damage-susceptible stage of the crop, significantly fewer beetles

SID 5 (Rev. 05/09) Page 5 of 24 were found on plants treated with insecticide, as expected (ANOVA F5,15 =6.82, P<0.01 and F5,15 =4.40, P<0.05, respectively; Fig 1). The other treatments were not significantly different from the control, although generally there were fewer beetles on plants in plot centres when there was a trap crop than in control plots, and fewer still beetles if the trap crop was treated with SA, or if the main crop was treated with MeJa, or both (Fig 1).

The proportion of buds used for oviposition was significantly lower in plots sprayed with insecticide than all other treatments (ANOVA F5, 2 =6.82, P<0.01); otherwise there were no other differences between treatments. There were no significant differences in the number of blind stalks per plant between treatments (ANOVA P>0.05).

To assess any chemical changes occurring as a result of the induction treatments, volatiles were collected from TR and OSR plants in the field before and after treatments with SA and MeJa, respectively. Volatiles in head space samples were collected by air entrainment and analysed by GC-MS. Interestingly, there was no evidence of an increase in 2-phenylethyl isothiocyanate produced by the treated TR compared to the control, but levels of methyl salicylate decreased and both phenylacetaldehyde and (E-E)-α-farnesene were increased in the SA treated plant. Both of the latter compounds have been identified from TR previously and phenylacetaldehyde has been shown to be one of the main volatile attractants for pollen beetles and should therefore increase their numbers on the trap crop (6). In contrast, MeJa treated OSR showed a decrease in levels of benzaldehyde and linalool and an increase in limonene, sabinene and a benzoic acid methyl ester. None of these compounds have been shown to be particularly behaviourally active against pollen beetles and may therefore have little effect in the field.

Fig 1. Mean beetles per plant in the oilseed rape plants in the centres of plots of each treatment

01/02 Milestone 2. First trial of winter OSR varieties Small plot studies of the potential of winter sown TR as a trap crop for WOSR conducted in PI0340 gave very promising results for autumn/winter pests. The TR was significantly preferred to the WOSR by adult cabbage stem flea beetles, Psylliodes chrysocephala, and stem weevils, Ceutorhynchus pallidactylus (7). Furthermore, the TR trap crop significantly reduced the number of cabbage stem flea beetle larvae in the OSR main crop compared to plots without a TR trap crop (7). However, in the spring, there was not always a sufficient separation of the flowering times between the TR and the WOSR for the trap crop to provide significant protection from pollen beetles for the WOSR. As shown in the SOSR model, this separation of flowering times is crucial if the trap crop is to attract and arrest enough of the pest population to reduce pest damage to the OSR crop at the vulnerable green/yellow bud stage (8). If we can achieve this separation by identifying and utilising late flowering WOSR varieties and/or earlier flowering trap crops, the pest control strategy for WOSR will be more robust. Also, the identification and exploitation of WOSR varieties which are less attractive to the pests due to less potent olfactory or visual cues, but have equally good yield potential, will lead to a yet more substantial pest management strategy for WOSR. In order to select varieties suitable to go forward into future „push-pull‟ strategies for the winter sown crop, nine winter rape cultivars were screened: Tyfon fodder turnip rape, Pasja fodder turnip rape, Largo winter turnip rape, Grizzly OSR, Expert OSR, Nickel OSR (the parent of AP2 and FP2 experimental lines), FP2 (an experimental line with yellow petals near-isogenic to AP2), AP2 (an experimental apetalous line, near-isogenic to FP2), and Cheops, a white petalled commercial cultivar. In the spring, the growth stages of plants in each plot were recorded and assessments of the number of inflorescence pests and beneficial insects in were made using techniques described above in 01/01 Milestone 1, on 6 sampling occasions every 2-3 days during the green- yellow bud stage and weekly thereafter until full flowering of the OSR crops.

SID 5 (Rev. 05/09) Page 6 of 24 Table 1. The mean no. pollen beetles per plant on 9 different rape lines on 6 sampling occasions through the spring growing-season. 1 2 3 4 5 6 Rape line (16/4/06) (18/4/06) (21/4/06) (28/4/06) (5/5/06) (12/5/06)

Tyfon fodder TR 4.5 2.5 16.3 6.9 1.6 0.3 Pasja fodder TR 9.6 12.8 13.9 11.3 1.5 0.9 Largo winter TR 3.3 2.1 16.2 11.5 1.2 0.7 Grizzly OSR 1.0 0.7 10.6 8.5 3.1 0.4 Expert OSR 1.7 1.9 11.4 9.7 1.3 0.5 Nickel OSR 1.5 1.4 11.2 9.1 1.9 0.5 FP2 Full petal breeders line 1.3 1.0 12.5 10.6 2.1 0.9 Cheops white-flowered OSR 2.1 1.9 12.8 12.5 1.7 0.4 AP2 Apetallous breeders line 3.5 1.7 12.9 8.5 0.9 0.3

Pasja fodder turnip rape began flowering first, and supported the greatest number of pollen beetles early in the season, indicating that it was the most attractive line tested (Table 1). During the green-yellow bud stage of the OSR, when the crop is most susceptible to damage (sampling occasions 3 and 4), Grizzly OSR was the least infested with pollen beetles (Table 1). The number of other inflorescence pests and beneficials recorded were too small for analysis.

01/03 Milestone 3. Further development of the push-pull strategy for SOSR Results from the field trial 01/01 on plant activators did not yield results worth pursuing on a field scale in 01/03. Progress made in PS2101/PS2105 and by our collaborators in developing new formulations of lavender was not completed in time to test them. Therefore, due to the lack of repellents, this objective was deferred to the extension year (see Objective 3).

01/04 Milestone 4. Towards development of a push-pull strategy for WOSR Promising cultivars of WOSR that were found to be either late-flowering or less-preferred by inflorescence pests in 01/02 Milestone 2 were taken forward and tested in a push-pull strategy using the early-flowering fodder turnip rape cv Pasja (also identified in 01/02 Milestone 2) as a trap crop. Plots (30m x 30m) of WOSR (cv Grizzly and Cheops: with yellow petals and white with yellow centres, respectively – Fig 2) were grown either with or without a Pasja turnip rape trap crop border. The four treatments were replicated four times in a randomized block design. Plant growth stage and the number of adult pests were assessed on 7 occasions using methods described in 01/01 Milestone 1. A bud sample was taken as previously described, to assess direct damage by pollen beetles and a pod sample was also taken to assess damage by seed weevils. Two-hundred pods were sampled at random from each plot and dissected in the laboratory. The number of pods infested with seed weevil larvae was recorded.

A B C

Fig 2. (A) OSR cv. Cheops and cv .Grizzly; (B) Replicated plot trial with Grizzly with a trap crop in the foreground; (C) Cheops with trap crop in the foreground. In both B and C, plots of both OSR cv‟s without trap crops can be seen in the background.

As expected, the Pasja TR trap crops developed faster than the OSR main plots and acted as an efficient trap crop for pollen beetles: at the damage-susceptible green-yellow bud stage of the OSR cultivars, significantly more beetles were found in the TR plants bordering the plots than OSR plants bordering the plots (an average of 10.6 beetles were found per TR plant and an average of 4.4 beetles in the OSR plants in the borders (P<0.05; see also Fig 3A). Turnip rape trap crops reduced the infestation by pollen beetles in the main OSR plot compared to plots without a TR trap crop. This effect was significantly different when the yellow-petalled cv. Grizzly comprised the main crop (mean no. per plant on Grizzly with trap crop = 2.96; mean no per plant on Grizzly without a trap crop = 5.1; P<0.05). However, the reduction was not significantly different when the white petalled cultivar Cheops was used as the main crop (mean no. beetles per plant on Cheops with trap crop = 4.51, mean no. beetles on Cheops without trap crop = 5.1; P>0.05). This difference may have been due to the slightly earlier development of Cheops than Grizzly with respect to the TR trap crop (see Fig 2C cf. 2B).

SID 5 (Rev. 05/09) Page 7 of 24

Significantly fewer buds were damaged by pollen beetles in Grizzly plots with a TR trap crop than those without a trap crop (9.57% and 12.38%, respectively; P<0.05). However, the trap crop did not significantly affect the proportion of damaged buds in plots of Cheops (11.5% with a TR trap cop and 9.26% without a trap crop, respectively, P>0.05). Similarly, less pods were damaged by seed weevils in Grizzly plots with a turnip rape border than those without (2.50% and 4.63%, respectively; P<0.05) but differences were not statistically significant in plots of Cheops (3.0% with a trap crop and 2.63% without; P<0.05). Pasja TR and OSR cv. Grizzly were therefore selected for use as an early-flowering trap crop and later flowering main crop in our subsequent experiments on the winter cropping system.

A

Mean no. pollen beetles per plant (borders)

0.6 0.5 Grizzly -TR Grizzly + TR 0.4 Cheops -TR 0.3 Cheops +TR 0.2 0.1

Mean Mean no. beetles/plant 0 Date 1 Date 2 Date 3 Date 4 Date 5 Date 6 Date 7 Sampling occasion

B

Mean no. pollen beetles per plant on oilseed rape plants (plot centres) Grizzly -TR 0.4 Grizzly +TR

0.3 Cheops -TR Cheops + TR 0.2

0.1

0 mean mean no. beetles/plant Date 1 Date 2 Date 3 Date 4 Date 5 Date 6 Date 7 Sampling occasion

Fig 3. Mean no. pollen beetles per plant on plants in plot borders (A - top) and plot centres (B- bottom) throughout the bud- and flowering period

01/05 Milestone 5. Large-scale evaluation of a pest management strategy for SOSR Previous work on the potential of trap cropping in OSR has used plot sizes of 30 x 30m. In this experiment we trialled the tactic at larger, more realistic field scales. Six, 1ha crops of SOSR, three with a 9m-wide Pasja TR trap crop bordering the crop and three without a TR trap, were sown in separate fields, with each at least 500m from any other treatment or OSR crop. Unfortunately two of the sites with no trap crop failed to establish. However, the remaining plots were sampled every 3-4d on 7 occasions from the green bud stage of the TR (BBCH GS 50-53) to full-flowering of the OSR (BBCH GS 67-68; by this stage the TR was at GS 79). Adult pests were sampled as described in 01/01 Milestone 1.

The TR trap crop flowered approximately 2 weeks before the main crop and pollen beetles were significantly more attracted to it than the OSR plants in the border at the damage-susceptible stage of the main crop (sampling occasions 2 and 3 Fig 4A). This supports our previous work using small plots and suggests that pollen beetles

SID 5 (Rev. 05/09) Page 8 of 24 prefer TR to OSR on full field scales. However, its performance at protecting the main crop was variable; at the damage susceptible stage of the main crop (sampling occasions 2 and 3) the trap crop performed well and reduced the number of beetles per plant in the main crop compared with plots without a TR trap crop on the first occasion, but not the second (Fig 4B). The loss of two out of three replicates of one of the treatments could have been responsible for this.

A B

Fig 4. (A) Number of pollen beetles per plant on oilseed rape (OSR) and turnip rape (TR) plants in plot borders; (B) No. beetles in oilseed rape plots with and without a turnip rape trap crop

01/06 Milestone 6. Large scale evaluation of a pest management strategy for WOSR The trap cropping strategy for winter OSR was also tested at a more realistic field scale. To avoid inter-plot interactions, large-scale (1ha) plots of WOSR either with or without a winter TR trap crop were grown in separate fields at least 500m from the nearest OSR crop/plots. The plots were established in autumn 2008 and sampled as described in 01/01 Milestone 1 in spring 2009 during the extension year to the project (Objective 3). Four treatments were tested: (1) WOSR with a TR trap crop – both untreated; (2) WOSR with a trap crop treated with insecticide; (3) WOSR with no trap crop – untreated; (4) WOSR with no trap crop – whole plot treated with insecticide. Two replicates of each treatment were established at two sites (Rothamsted and Woburn Farms).

Replication was too low to allow robust statistical analysis (more replicates will be done as part of the Defra-LINK project LK09108 „Development of an integrated pest management strategy for control of pollen beetles in winter oilseed rape‟). However, it was clear that, as in our previous experiments, the TR plants (untreated) in the plot borders were more heavily infested than the untreated OSR plants (Fig 5A). Insecticide sprays applied to the borders reduced beetle infestation as expected (Fig 5A). At the damage-susceptible stage of the crop (sampling occasions 3-5), untreated plots with a TR trap crop had fewer beetles per plant than the control plot (Fig 5B) and insecticide application to the TR trap crop further reduced the number of beetles per plant in the main crop compared to leaving the trap crop untreated (Fig 5B).

SID 5 (Rev. 05/09) Page 9 of 24 A

Plants in the borders

25

20 1. OSR-/TR-

15 2. OSR-/TR+

3. OSR-/OSR- 10 4. OSR+/OSR+

5 Mean no. beetles/plantno. Mean

0 1 2 3 4 5 6 7 8

Sampling occasion

B

Plants in the centres

14

12 1. OSR-/TR- 10 2. OSR-/TR+ 8 3. OSR-/OSR- 6 4. OSR+/OSR+

4

Mean no. beetles/plantno. Mean 2

0 1 2 3 4 5 6 7 8

Sampling occasion

Fig 5. Mean number of pollen beetles per plant in plot borders (A- top) and in plot centres (B-bottom) in plots with or without a trap crop and with or without insecticide applications.

OBJECTIVE 2. Devise and test a semiochemical based management strategy for pests of cereal and legume crops in a conventional arable rotation. Use plant-derived and other semiochemical stimuli that provide enhanced attractant and non- host/masking/repellent cues under field conditions for target pests and their natural enemies, to formulate an integrated crop protection strategy for the other combinable crops used in conventional arable rotation i.e. cereals (winter wheat chosen to utilise best existing knowledge of predatory and parasitic insects) and legumes (chosen to target both coleopterous and aphid pests). Define and quantify the effects of the strategy on non-target species, particularly natural enemies and other beneficial insects.

In previous Defra project PS2107 (see SID5 final report on Defra website), existing Defra recommended mixtures for field margins and forage/undersowing applications were investigated, under field conditions, for their suitability for inclusion in pest management strategies for target pests of cereals and legume crops. Large (576m2) habitat management plots (HMP) consisting of Defra recommended field margin mixtures „TG2‟ (grass mixture only) and

SID 5 (Rev. 05/09) Page 10 of 24 „WM2‟ (pollen/nectar mix including legumes) were established in a 6 x 6 quasi-complete Latin square design (see Fig 6). This particular design was generated at Rothamsted specifically to investigate neighbour interactions between semiochemical treatments. All treatments occur next to each other twice, in both rows and columns, thereby allowing for analysis of plot to plot interactions via simple mathematical models. The statistical analysis is most powerful with paired treatments (two margin mixes, two semiochemical treatments and two untreated controls). The design and summary of the full experiment was as follows below. Each season two from B, D, E, and F were treated with semiochemicals and the other two were untreated controls.

C E A B D F A = clover/grass mixture WM2 F D B A E C B = crop A F E D C B C = grass mix for margins TG2 E B F C A D D = crop B C D E F A E = crop D A C F B E F = crop

The work described here is a continuation of the study started in PS2107 and utilises the same established experimental plots. In each year, pest and beneficial insect populations were monitored throughout the season by a range of sampling techniques including Vortis suction sampling, water, sticky and pitfall traps and by specific pheromone traps where appropriate in the HMP and the crop plots. Trap catches were assessed in the laboratory at the end of the trial. Plant samples were taken and visual assessments were made in situ on a 5 x 5 grid pattern in the crop plots to determine temporal/spatial distribution of feeding damage and patterns of pest distribution and these were mapped and analysed by ANOVA. Yields were taken from each of the wheat plots at harvest.

Fig 6. Aerial photograph of Habitat management plot experiment Year 2 of PS2107

02/07 Milestone 7. Winter wheat – 3rd crop trialled with habitat management plots As in the first trial with winter wheat in PS2107, the semiochemical treatments were delivered from PVC rope dispensers deployed as slow release point sources tied to canes within the plots. Treatment B was the plant activator cis-jasmone (CJ), which is both repellent to pest aphids and attractive to parasitoids and predators as well as inducing defence chemistry in the treated plants (9), and was put out on 18 May 2006 releasing approximately 67.25 mg/plot/day equivalent to 11.68g/ha/day. Treatment E was a combination of the aphid sex pheromone components nepetalactone and nepetalactol, which attract aphid parasitoids and predators (10). These were released in a 1:1 ratio (1.8mg/plot/day of each), and were not deployed until aphid colonisation had begun, since these natural enemies will cease to respond to the attractants if there is no reward. Aphid populations were visually assessed on 4 tillers per sample site on numerous occasions throughout the experiment and populations of other pest and beneficial insects and spiders were assessed using a range of trapping techniques including Vortis suction samples, yellow sticky traps and pitfall traps.

Migrating aphid populations were similar in number to the previous season with a peak of approximately 1.4 aphids /tiller on 28 June (Fig 7), but arrived and peaked a week earlier than in 2005. Most of the aphids were Metopolophium dirhodum with very low numbers of Sitobion avenae and Rhopalosiphum padi. There were no significant treatment effects on aphid numbers, although aphids were fewer on CJ treated plots confirming a trend seen in 2005 and in previous experiments (9). The results of counts and trap catches on the untreated plots were combined as there were no significant differences between them.

SID 5 (Rev. 05/09) Page 11 of 24

Fig 7. Mean numbers of cereal aphids on winter wheat in 2006 HMP trial

Parasitised aphids in the wheat plots were too few to assess, but adult parasitoids, caught on yellow sticky traps, peaked in the week after the aphid peak and numbers were greatest in the treated plots compared to the controls or the HMP (Fig 8). Orange wheat blossom midge, Sitodiplosis mosellana, was very late emerging and missed the susceptible growth stage of the crop, resulting in very low infestation levels and no visible treatment effects. Numbers of yellow wheat blossom midge, Contarinia tritici, females caught on sticky traps were greater than in 2005 and showed that the TG2 grass mixture plots were an important source of this pest (P<0.001 on 6 June 2006; Fig 9), possibly because some grass species are alternative host plants, but there was no evidence of an increase in their colonisation of the wheat plots. Additionally, more adult pea and bean weevil, Sitona lineatus, were caught in pheromone traps placed in the centre of the TG2 plots in autumn 2005 and spring 2006 compared to the WM2 plots (Fig 10). This was unexpected since S. lineatus feeds on vetches and clovers when peas and beans are unavailable. However, this pattern was repeated in following seasons until the gradual change in the botanical content of the WM2 plots towards a predominance of grass species resulted in no differences (Fig 10 and see Fig 18 below). The TG2 plots formed a dense layer of vegetation, which may have provided a more suitable over wintering habitat for the weevil. As found in 2005, there were again significantly more predatory Tachyporus beetles (p<0.001) and adult hoverflies (p<0.01) in Vortis samples from WM2 plots than from TG2 plots, but unlike 2005, no significant differences for spiders or ladybirds. This is the only evidence from the 2006 trial of the HMP acting as possible sources of pests or beneficial species. Otherwise, numbers of all key pest and beneficial species were lower on the margin mixes than in the wheat. There were significantly greater numbers of carabid beetles (predominantly Pterostichus spp.) in pitfall traps (P<0.05) in the crop plots compared to the HMP indicating that activity of these predators was greater in the crop both prior to the arrival of the aphids and after aphid numbers had crashed (Fig 11).

Fig 8. Mean numbers of aphid parasitoids caught on yellow sticky traps in 2006 HMP trial

SID 5 (Rev. 05/09) Page 12 of 24

Fig 9. Mean number of female Contarinia tritici on yellow sticky traps in 2006 HMP trial

Fig 10. Mean number of S. lineatus caught in pheromone traps on HMP plots 2005-2008

Fig 11. Mean number of Carabidae (all species) caught in pitfall traps in 2006 HMP trial

SID 5 (Rev. 05/09) Page 13 of 24 02/08 Milestone 8. Winter beans - fourth crop trialled with habitat management plots Despite significant efforts to protect the crop, the winter beans were badly damaged by rooks, and as a result many plots were almost completely depleted of plants. Having given the crop every chance to recover, a late decision was made in spring 2007 to plough it in and re-sow with a spring bean variety. However, herbicide residues from the winter crop prevented germination of a percentage of spring bean seed resulting in a thin crop. Due to the failure of the winter crop, the semiochemical treatments detailed in the milestone were changed. The treatments, which have not been tested in beans on this scale before, were the same as in the wheat in 2006. Treatment D plots had point sources of cis-jasmone (CJ), at an approximate release of 11.68g/ha/day, and treatment F plots a combination of the aphid sex pheromone components nepetalactone and nepetalactol, released in a 1:1 ratio (1.8mg/plot/day of each) all deployed in the crop on 22 May. Pest and natural enemy populations were monitored weekly on the beans, by tray beating 25 plants per plot in a 5 x 5 grid pattern throughout May and June until aphid numbers declined, by Vortis on 29 May and 19 July and by yellow sticky traps and pitfall traps until the end of July on all the plots. Populations were also monitored on the HMP by Vortis, pitfall traps and in S. lineatus aggregation pheromone traps.

As in 2006, counts and trap catches for untreated plots are combined as there were no significant differences between them. Due to the warm April weather, pea aphids, Acyrthosiphon pisum, arrived very early and in large numbers. Migration of black bean aphid, Aphis fabae, was also early, but populations were very small in comparison to A. pisum. With a few exceptions, numbers of key pest and beneficial species were lower on the HMP than in the bean crop. There were fewer A. pisum in the tray beating samples from CJ treated bean plots, although the difference was not significant to untreated or nepeta plots (Fig 12). This effect was confirmed by the numbers caught in Vortis samples on 29 May (data not shown). In addition, at the peak of catch, there were significantly fewer alate aphids caught on yellow sticky traps in the CJ plots compared to control and nepeta plots (P<0.05).

Parasitised aphids in the bean plots were difficult to assess, but adult parasitoids, caught on yellow sticky traps, peaked in the two weeks after the aphid peak and numbers were significantly greater in the treated plots compared to the controls or the margin mixtures (P<0.05; Fig 13). Numbers of yellow wheat blossom midge, C. tritici, females caught on yellow sticky traps in the TG2 grass mixture plots were greater than in 2006 (Fig 14), again showing that this margin mixture is an important source of this pest, possibly because some grass species are alternative host plants. As seen in 2006, large numbers of S. lineatus were caught in pheromone traps in the TG2 grass margin mixtures in spring 2007 compared to the WM2 (see Fig 10). However, although these plots were probably acting as a reservoir for the weevil, there was no evidence of higher numbers in neighbouring bean plots, possibly because the plot size was too small to show such an effect for this very mobile insect. Numbers of S. lineatus caught in Vortis samples were lower in the treated compared to untreated plots, but there were no significant differences due to large variation in sample size (Fig 15). As seen in 2006, there were significantly more predatory carabid beetles (predominantly Pterostichus spp.) in pitfall traps in the crop plots compared to the HMP throughout May and June (P<0.05; Fig 16).

Fig 12. Mean numbers of Acyrthosiphon pisum in beans in 2007 HMP trial

SID 5 (Rev. 05/09) Page 14 of 24

Fig 13. Mean number of aphid parasitoids caught on yellow sticky traps in 2007 HMP trial

Fig 14. Mean number of female Contarinia tritici on yellow sticky traps in 2007 HMP trial

Fig 15. Mean number of S. lineatus in Vortis samples in 2007 HMP trial

SID 5 (Rev. 05/09) Page 15 of 24

Fig 16. Mean number of Carabidae (all species) caught in pitfall traps in 2007 HMP trial

02/09 Milestone 9. Winter wheat - fifth crop trialled with habitat management plots On agreement with the sponsors, the treatments for the final year were altered to investigate the potential enhancement of the effects on cereal aphids of the plant activator, cis-jasmone (CJ) in combination with the synergist piperonyl butoxide (PBO) in support of Defra project PS2124 (see Defra website for PS2124 final report and full experimental details). Treated plots were sprayed with either a standard formulation of CJ with the wetter Ethylan BV (50g formulated in 0.1% EBV in 200 l/ha) or with a combination of CJ formulated with 0.1% PBO30 on the 30th May and again on the 13th of June 2008. Cereal aphids and their natural enemies were monitored weekly by counting their numbers on 4 tillers at each of 25 sampling sites on a 5x5 grid in each plot. Aphid predators were also assessed by 3 pitfall traps and a yellow sticky trap in each plot. Infestations of gout flies, Chlorops pumilionis, and orange and yellow wheat blossom midge (wbm) were determined in field collected plant samples.

Cereal aphid numbers were well below threshold, peaking early at a maximum of just over 2 per tiller on 18 June, the predominant species being M. dirhodum. As seen in previous experiments, the CJ/EBV formulation reduced overall aphid numbers consistently, but after an initial effect, aphid numbers on the PBO/CJ treatment were no different to those on untreated plots (Fig 17). Due to the low aphid numbers and their patchy distribution statistical analysis showed no significant differences and it is difficult to draw firm conclusions from this result.

Numbers of gout flies and wheat blossom midges were also very low. The CJ/EBV treatment and to a lesser extent the CJ/PBO treatment reduced the incidence of orange wbm, yellow wbm and gout fly, but again the low numbers were not significantly different. There were no obvious differences in counts or trap catches of aphid natural enemies between treatments.

In the HMP, numbers of pest and beneficial species were assessed by pitfall, pheromone and yellow sticky traps and by Vortis suction samples. Unlike in previous years there was little difference in numbers of S. lineatus caught in pheromone traps in WM2 and TG2 plots in spring 2008 (see Fig 10) probably due to the increase in grasses in the former. Additionally, much lower numbers of female yellow wbm were caught on yellow sticky traps in the TG2 grass plots this season and as for S. lineatus there was no difference between these plots and WM2 plots.

SID 5 (Rev. 05/09) Page 16 of 24

Fig 17. Cumulative mean numbers of cereals aphids on winter wheat in 2008 HMP trial

Changes in the composition of HMP plots Vegetation surveys showed notable changes in the botanical composition of the WM2 habitat management plots since establishment. A decline in legume cover (legumes accounted for 43% of total cover in 2004 on average, compared with 8% in 2006) was accompanied by an increase in overall grass cover (from 34% to 88% of total cover in the period 2004-2006). These changes were primarily attributed to a decline in the cover of alsike and red clovers, which were the dominant legumes in 2004, whilst cover of red fescue, timothy, rough-stalked meadow grass and bent grasses had simultaneously increased. The botanical composition of the TG2 plots remained more stable, although some changes did take place. An increase in the cover of timothy was accompanied by a decline in the cover of tall fescue, whilst cover of cocksfoot, red and meadow fescues remained reasonably constant throughout 2004-2006. There was a marked recovery in broadleaf cover in the WM2 plots in 2008 (Fig 18) however, this was not due to the clovers, but to an increase in crested dogstail, black knapweed and birdsfoot trefoil, which did not support the levels of beneficial species seen in these plots originally. This change in the botanical composition of the WM2 plots is the most likely cause for the decline in numbers of beneficial species in these plots. As a consequence of this finding, and those above, it is necessary for the composition of the mixes to be designed not only according to botanical development and interaction, but also to the chemical ecology of the components. Since the botanical composition of the WM2 pollen rich habitat management plots has shifted over time and a few grass species had become predominant, ultimately there was little difference between WM2 and TG2 the grass mix and, after consultation with the sponsors, it was decided that sufficient data had been acquired and that this sub-objective would not be taken forward in the extension year of the project.

SID 5 (Rev. 05/09) Page 17 of 24

Fig 18. Changes in the botanical composition of the Habitat Management plots 2006-2008

Conclusions and suggestions for further work Due to the gradual shift in botanical composition of the margin mixtures, the levels of pests and beneficial organisms in the HMP changed continuously. Therefore, it was not possible in this project (and PS2107) to determine the potential use of these particular field margin mixtures as part of a wider pest management or push- pull approach for control of pests of crops in arable rotations. However, it was clear that if stable, mixtures rich in flowering legumes such as clovers could support aphid parasitoids and predatory syrphids, coccinellids and staphylinids. Many of these species are attracted to semiochemicals associated with crop damage and could in theory be manipulated to contribute towards pest management if present in margins early enough and in sufficient numbers. The pifall traps demonstrated clearly that predatory carabids are more active within crop than in the margin mixtures. The grassy margin mix TG2 was a source of at least two pest species. One, the yellow wbm, a small delicate insect, which showed no evidence of invasion of nearby crop plots in this project, but has the potential to increase in numbers in long term margins. The other, S. lineatus used the grass margin mixture to provide over winter cover and was quick to take advantage of the field bean crop in the rotation.

Low numbers of cereal aphids confounded both of the wheat trials and the bean crop was poor. However, the cis- jasmone treatment performed consistently if not significantly better than the other treatments. Without a stable population of natural enemies in the HMP it was not possible to use other semiochemical treatments towards the development of a more targeted pest management approach. However, since this trial was established, more mixtures have come onto the market and these could be tested in additional experiments.

OBJECTIVE 3 A 1-year extension to the project (1/4/2009-31/3/2010) was granted by the sponsors to: „develop and test on a realistic field scale, a full push-pull strategy for the key insect pests of oilseed rape and assess the impacts of the strategy on promoting biological control by their natural enemies (primarily parasitic wasps), with particular attention on the development of repellent/deterrent components of the strategy‟.

03/01 Milestone 1. Investigate in laboratory and small-scale semi-field arena bioassays, the potential of repellent semiochemicals and synergists on suppression of pests in oilseed rape. A slow release formulation of lavender oil was developed for field assessment using a readily available source of gum acacia, Instantgum BB (obtained from Colloides Naturels (UK) LTD. Manchester.). To determine an effective rate, and the periodicity over which it was likely to be effective, experimental formulations with different levels of oil, gum and wetter (Ethylan BV) were sprayed onto oilseed rape plants and the volatiles released were collected by air entrainment of the headspace around one leaf at various intervals after application. The samples collected were analysed by gas-chromatography and two lavender related components, linalool and linalyl acetate, which had been shown to be repellent to pollen beetles previously (11), were identified. The behaviourally active components were readily detectable up to 5 hours after application, but by 24h very little remained. The most effective field rate was determined as 4kg lavender oil, formulated with 40g wetter and 6kg of gum in 200 l/ha.

SID 5 (Rev. 05/09) Page 18 of 24 03/02 Milestone 2. Investigate repellents for the ‘push’ element of the strategy in a small plot field trial. On agreement with the sponsors, Milestone 01/03 „Further development of a push-pull strategy for spring oilseed rape‟ was deferred from year 2 and addressed as Milestone 2 of the Extension year. The most effective number of treatments of lavender necessary to repel insect pests of oilseed rape was tested in a replicated field experiment using plots (24 x 24 m) of WOSR (cv. Grizzly). Four treatments were compared: untreated controls, and plots receiving either: 1, 2 or 3 sprays of lavender oil (4 litres oil (=~4kg) + 40ml Ethylan BV and 6kg gum acacia (instantgum BB) in 200l water/ha), as determined from experiments in 03/01. Treatments were replicated 5 times in randomized blocks. The first two treatments were applied at green/yellow bud stage (31 March and 3 April) for pollen beetles and the last one was applied on 22 May post flowering against seed weevils.

Assessments of adult pests on plants were made as described in 01/01. Plants were sampled on 5 occasions, the first on 30/3/9 before any of the treatments were applied. The second count was done 3h after the first treatment on 31/3/9 and an additional count was done the following day on 1/4/9. The forth count was done 3h after the second spray on 3/4/9. The 5th count was done on 6/4/9. There was no count done following the 3rd spray as preliminary scouting indicated that there were very few weevils in the plots. On 8/4/9, after 2 of the lavender sprays, a direct assessment of bud damage was done as described in 01/01.

Fig 19. Mean (transformed) and estimated standard errors (from REML analysis) of the number of pollen beetles per plan on oilseed rape plots through time either untreated or treated with 1 or 2 sprays of lavender oil.

Only enough pollen beetles were sampled for meaningful analysis. On the 1st assessment date before any treatments were applied, as expected there was no significant difference between the number of pollen beetles per plant on each plot (ANOVA 3,12 F=1.23; P>0.05; Fig 19). On the 2nd assessment, after the first application of lavender, there were significantly fewer pollen beetles per plant in the lavender treated plots compared with the controls (ANOVA 1,14 F=20.38; P<0.001; Fig 19). On the 3rd assessment the number of beetles in the plots had increased, but there were still significantly fewer in the lavender treated plots than the control plots (ANOVA1,14 F=15.96; P<0.001; Fig 19). There was no significant difference in the number of beetles/plant between treatments after the second application of lavender (ANOVA 2,13 F=2.46; P>0.05), but there were fewer beetles per plant in plots treated with lavender 72h after treatment (ANOVA 2,13 F=9.87; P<0.01) (Fig 19). This indicates that lavender can reduce or slow the infestation of pollen beetles into the crop, and that two applications are better than one at achieving efficacy. However, the lavender treatment did not affect the direct damage to the crop: there was no significant difference between treatments in the number of blind stalks (ANOVA 2,13 F=0.15; P>0.05) nor the proportion of buds that were damaged (ANOVA 2,13 F=0.02; P>0.05).

SID 5 (Rev. 05/09) Page 19 of 24 03/03 Milestone 3. Demonstrate, on a realistic field scale, the effectiveness of a push-pull strategy against insect pests of oilseed rape and assess effects on their natural enemies. The full push-pull strategy for oilseed rape was tested using SOSR as a model system. The effectiveness of each component of the strategy, i.e., the push (lavender); the pull (turnip rape trap crop border) and both push and pull (OSR with lavender repellent and a turnip rape trap crop) was tested against a control (single block of OSR). One replicate of each of the four treatments was established in a 2 x 2 block on each of 3 sites to give a balanced neighbour design by block. Each plot was 75 x 75 m and the distance between blocks and rows was 12m. Unfortunately, one site failed to establish. On the remaining two sites, two applications of lavender as determined from 03/02 (4l oil (=~4kg) + 40ml Ethylan BV and 6kg gum acacia in 200l water/ha) were sprayed onto the push and push-pull plots on 2 and 12 June 2009. The plots were sampled on a grid pattern on 36 sample points evenly spaced across each plot. Adult pests on plants were sampled from three plants at each sampling point on four occasions, once before the lavender application, twice after the first application and once after the second application. As a direct measure of damage by pollen beetles, on 8/6/9, three OSR racemes were sampled from each of the 16 central sampling points in each plot after the first application of lavender and dissected as described in 01/01. Green-coloured water traps were placed at each sample point in each treatment to assess the effect of the treatments on parasitoids and non-target insects. They were set out on 3/6/9 and changed weekly over three weeks until 23/6/9 (i.e. 3/6/9-9/6/9; 9/6/9-16/6/9; 17/6/9-23/6/9): this period encompassed both applications of lavender. The samples were collected by passing the water through a sieve and transferring the captured insects into a labelled container filled with ethanol. These were later sorted to species including: pollen beetle, flea beetle (Phylotretta atra), seed weevil, stem weevil, aphids and parasitoids.

Push-Pull Push

1 1

2 2

3 3

4 4

5 5

6 6

1 2 3 4 5 6 1 2 3 4 5 6

Control Pull

1 1

2 2

3 3

4 4

5 5

6 6

1 2 3 4 5 6 1 2 3 4 5 6

Fig 20. Example of the spatial distribution of pollen beetles on experimental plots. Plots with lavender application (push); plots with turnip rape trap crop (pull); plots with both lavender and a trap crop (push-pull) and plots with neither push nor pull (control). Shaded areas show the mean no. pollen beetles per plant (n=3) on 36 sampling sites across the plots on one field site on the 2nd sampling occasion, approx 15h after the first application of lavender. Shading range: white – black = 0-10 beetles/plant.

Only pollen beetles were counted from plants in sufficient numbers for meaningful analysis. As in our previous work, we found strong evidence that TR is more attractive to pollen beetles than OSR. The presence of turnip rape had a significant effect on the number of beetles found in the borders of plots, with more beetles being found on TR plants than OSR plants on the first two assessments (ANOVA, F1,3 = 109.48; P=0.002 and F1,3 = 34.92; P=0.01; See also Fig 20 from the 2nd assessment). As in our previous work, when the OSR came into flower this preference was lost (ANOVA P>0.05). This work also shows that turnip rape can act as an efficient trap crop at the larger scale. On the first two sample dates, the number of beetles in the crop centre was significantly reduced in plots with a „pull‟ element compared to plots without a trap crop (ANOVA F1,3 = 9.26; P=0.05 and F1,3 = 8.87; P=0.05; see also Fig 20). In contrast to our work on the winter crop in the same year, the „push‟ component in this experiment had no statistically significant effect (ANOVA P>0.05 on all 3 sampling occasions after application) so it followed that there was no push-pull interaction (see also Fig 20). This is likely to be due to insufficient statistical power in the analysis as one of our field sites was lost.

SID 5 (Rev. 05/09) Page 20 of 24 There was less damage to buds in the experimental than control plots (Table 2). The effect of the lavender „push‟ bordered on statistical significance (ANOVA 1,3 F=8.33; P=0.063). Similarly, the effect of the trap crop „pull‟ borders significance (ANOVA 1,3 F=6.56; P=0.083). Had we not lost one of our replicates, these differences may have been clearer. There was no interaction between push and pull elements of the strategy (ANOVA 1,3 F=1.43; P=0.318). There was no significant difference between the plots in the number of blind stalks.

Table 2. Mean proportion ± SE of damaged buds on plots with lavender application (push); plots with turnip rape trap crop (pull); plots with both lavender and a trap crop (push-pull) and plots with neither push nor pull (control).

Push-pull Push Pull Control 0.49 ± 0.019 0.56 ± 0.017 0.56 ± 0.019 0.59 ± 0.018

There were too few individuals of most species caught in water traps to analyse meaningfully non-target catch. However, there were sufficient numbers of parasitoids of the oilseed rape pests when grouped together (Blacus nigricornis, Brachyserphus parvulus, Tersilochus heterocerus, Phradis spp. and Diospilus spp). Phradis spp were by far the most numerous caught. TR was significantly preferred over OSR on the first two trapping dates (Fig 21A), and the presence of TR had a positive effect on numbers in the OSR crop centres: OSR parasitoids were significantly more common in OSR plots with a TR trap crop border than in plots without on the 2nd and 3rd assessments (Fig 21B), indicating that TR could have a positive effect on biocontrol in the main crop.

A B

Fig 21. Mean no. parasitoids of oilseed rape pests caught in water traps in plot borders (A) and in the oilseed rape crop in the plot centres (B).

03/04 Milestone 4. Collate data and prepare final report. Completed. We are in the process of preparing manuscripts for publication.

Conclusions and suggestions for further work We consistently showed, using a spring oilseed rape model, that early-flowering turnip rape trap crops are more attractive to oilseed rape pests, particularly pollen beetles, and can reduce efficiently pest infestations in the crop centres. We showed that this strategy is effective at realistic field scales (1ha). We successfully transferred the trap cropping strategy to the winter cropping system. In preliminary trials Pasja turnip rape as a trap crop with the late-flowering oilseed rape cv. Grizzly effectively reduced pollen beetle infestation at a realistic field scale (1ha). We suggest that further experiments are required to test that both the spring and winter system perform at the 10- 100ha scales more commonly used in the UK.

Turnip rape trap crops can be used as efficient „pull‟ components of a push-pull strategy. A formulation of lavender was developed and showed promise as a repellent against pollen beetles. However, in large-scale trials using 1ha plots we failed to show an effect, possibly due to the loss of one of our replicates. We suggest that a repeat of this experiment is necessary to confirm the efficiency of lavender at realistic field scales.

A Defra and Industry supported LINK project, working towards an IPM strategy to reduce the likelihood of the development of insecticide resistance within UK pollen beetle populations has now started. As part of this project the scale of the trap cropping approach for pollen beetle control will be extended for WOSR.

SID 5 (Rev. 05/09) Page 21 of 24 References cited in the report

(1) Rural Economy & Land Use (RELU) project RES-224-25-0093-A End of award report (2009). Re-bugging the system: promoting adoption of alternative pest management strategies in field cropping systems (overcoming market & technical obsticles).

(2) Bailey AS, Bertaglia M, Fraser IM, Sharma A, Douarin E (2009) Integrated pest management portfolios in UK arable farming: results of a farmer survey. Pest Management Science 65:1030-1039

(3) Bodnaryk RP (1994) Potent effect of Jasmonates on indole glucosinolates in oilseed rape and mustard. Phytochemistry 35: 301-305

(4) Kiddle GA, Doughty KJ & Wallsgrove RM (1994) Salicylic acid-induced accumulation of glucosinolates in oilseed rape (Brassica napus L.) leaves. Journal of Experimental Botany 45: 1343-1346

(5) Cook SM, Khan ZR & Pickett JA (2007) The use of push-pull strategies in integrated pest management. Annual Review of Entomology 52: 375-400

(6) Cook SM, Rasmussen HB, Birkett MA, Woodcock CM, Murray DA, Pye BJ, Watts NP & Williams IH (2007) Behavioural and chemical ecology of the success of turnip rape (Brassica rapa) trap crops in protecting oilseed rape (Brassica napus) from the pest Meligethes aeneus. -Plant Interactions 1: 57-67

(7) Barari H, Cook SM, Clark SJ & Williams IH (2005) Effect of a turnip rape (Brassica rapa) trap crop on stem- mining pests and their parasitoids in winter oilseed rape (Brassica napus). Biocontrol 50: 69-86

(8) Cook SM, Smart LE, Martin JL, Murray DA, Watts NP & Williams IH (2006) Exploitation of host plant preferences in pest management strategies for oilseed rape (Brassica napus). Entomologia Experimentalis et Applicata 119: 221-229

(9) Bruce TJA, Martin JL, Pickett JA, Pye BJ, Smart LE & Wadhams LJ (2003) cis-Jasmone treatment induces resistance in wheat plants against the grain aphid, Sitobion avenae (Fabricius) (Homoptera: Aphididae). Pest Management Science 59: 1031-1036

(10) Powell W & Pickett JA (2003) Manipulation of parasitoids for aphid pest management: progress and prospects. Pest Management Science 59: 149-155

(11) Mauchline AL, Birkett MA, Woodcock CM, Pickett JA, Osborne JL & Powell W (2008) Electrophysiological and behavioural responses of the pollen beetle, Meligethes aeneus, to volatiles from a non-host plant Lavender, Lavendula angustifolia (Lamiaceae) Arthropod-Plant Interactions 2: 109-115

Knowledge Transfer

S.M. Cook, L.E. Smart, M.P. Skellern, N.P. Watts, I.H. Williams (2006) Development of a push-pull strategy for control of oilseed rape pests. Proceedings of the International Symposium on Integrated Pest Management in Oilseed Rape 3-5 April, 2006, Gottingen, Germany

Cook, S. M. (2006). „Push-pull‟ strategies for pest management in agroecosystems: A plenary lecture. VIII European Congress of Entomology, September 17-22 2006, Iznir, Turkey, Abstract Book pp 7-10

Professor Pickett gave the Royal Society Croonian Prize Lecture, “Plant and Communication”, on 3 June 2008 and has referred to this work in many other presentations over this time period.

Dr Cook gave a talk entitled “Development of an integrated pest management strategy for pollen beetles in oilseed rape''at a joint Rothamsted Research Association (RRA) /HGCA workshop “Controlling Insect Pests” 21 April 2009.

Ms Smart gave a talk entitled "Semiochemicals in field experiments: aiming at improved pest control" at the PlantComMistra meeting in Alnarp, Sweden on 22 October 2009.

Demonstration plots of the trap cropping principle were presented at Cereals 2009

SID 5 (Rev. 05/09) Page 22 of 24

References to published material 9. This section should be used to record links (hypertext links where possible) or references to other published material generated by, or relating to this project.

SID 5 (Rev. 05/09) Page 23 of 24 Refereed Scientific Papers S.M. Cook, L.E. Smart, J.L. Martin, D.A. Murray, N.P. Watts and I.H. Williams (2006) Exploitation of host plant preferences in crop protection strategies for oilseed rape (Brassica napus). Entomologia Experimentalis et Applicata 119, 221-229.

Cook SM, Rasmussen HB, Birkett MA, Woodcock CM, Murray DA, Pye BJ, Watts NP & Williams IH (2007) Behavioural and chemical ecology of the success of turnip rape (Brassica rapa) trap crops in protecting oilseed rape (Brassica napus) from the pest Meligethes aeneus. Arthropod-Plant Interactions 1: 57-67

S.M. Cook, Z.R. Khan and J.A. Pickett (2007) The use of push-pull strategies in integrated pest management. Annual Review of Entomology 52, 375-400.

Cook, S. M., Jönsson, M. Skellern, M. P. Murray, D. A. Anderson, P. & Powell, W. (2007). Responses of Phradis parasitoids to volatiles of lavender, Lavendula angustifolia - a possible repellent for their host, Meligethes aeneus. BioControl 52(5):591-598

Bruce, T.J.A., Hooper, A.M., Ireland, L.A., Jones, O., Martin, J.L., Smart, L.E., Oakley J. & Wadhams, L.J. (2007) Field trapping trials of pheromone baited traps for orange wheat blossom midge, Sitodiplosis mosellana. Pest Management Science 63: 49-56.

Pickett, J.A., Birkett, M.A., Bruce, T.J.A., Chamberlain, K., Gordon-Weeks, R., Matthes, M.C., Napier, J.A., Smart, L.E. & Woodcock, C.M. (2007) Developments in aspects of ecological phytochemistry: The role of cis-jasmone in inducible defence systems in plants. Phytochemistry 68: 2937-2945.

Moraes, C.M.B., Birkett, M.A., Gordon-Weeks, R., Smart, L.E., Martin, J.L., Pye, B.J., Bromilow, R.H. and Pickett, J.A. (2008) cis-Jasmone induces accumulation of defence compounds in wheat, Triticum aestivum. Phytochemistry 69: 9-17.

Bruce, T.J.A., Matthes, M., Chamberlain, K., Woodcock, C.M., Mohib, A., Webster, B., Smart, L.E., Birkett, M.A., Pickett, J.A. and Napier, J. (2008) cis-Jasmone induces Arabidopsis genes that affect the chemical ecology of multitrophic interactions with aphids and their parasitoids. PNAS 105: (12) 4553-4558.

Mauchline AL, Birkett MA, Woodcock CM, Pickett JA, Osborne JL & Powell W (2008) Electrophysiological and behavioural responses of the pollen beetle, Meligethes aeneus, to volatiles from a non-host plant Lavender, Lavendula angustifolia (Lamiaceae) Arthropod-Plant Interactions 2: 109-115

Van den Berg, J., Torto, B., Pickett, J.A., Smart, L.E., Wadhams, L.J. and Woodcock, C.M. (2008) Influence of visual and semiochemical cues on field trapping of the pollen beetle, atromaculatus (Col: ). Journal of Applied Entomology 132: 490-496.

Bruce, T.J.A., and Smart, L.E. (2009) Orange wheat blossom midge, Sitodiplosis mosellana, management. Outlooks on Pest Management 20: (2) 89-92.

Other publications JA Pickett, TJA Bruce, K Chamberlain, A Hassanali, ZR Khan, MC Matthes, JA Napier, LE Smart, LJ Wadhams & CM Woodcock. (2006) Plant volatiles yielding new ways to exploit plant defence. In “Chemical Ecology: from Gene to Ecosystem.” Eds. M Dicke & W Takken. Springer, Wageningen, The Netherlands. Ch. 11. pp. 161-173

Barari, H., Cook, S. M. & Williams, I.H. (2006). Rearing the larval parasitoids of Psylliodes chrysocephala and Ceutorynchus pallidactlylus from field-collected specimens. IOBC/wprs Bulletin 29 (7): 215-223.

Cook, S. M. (2006). „Push-pull‟ strategies for pest management in agroecosystems: A plenary lecture. VIII European Congress of Entomology, September 17-22 2006, Izmir, Turkey, Abstract Book pp 7-10. Cook, S. M. & Denholm, I. (2008). Ecological approaches to the control of pollen beetles in oilseed rape. EPPO Bulletin 38: 110-113

Williams, I. H. & Cook, S.M. (2010). Crop location by oilseed rape pests and host location by their parasitoids In: Biocontrol-Based Integrated Management of Oilseed Rape Pests, Ed. Ingrid H. Williams, Springer Science+Business Media B.V.

SID 5 (Rev. 05/09) Page 24 of 24