General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

NoteIn 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 INFORMATIONThe 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 BD1449

2. Project title

Sustainable management strategies for creeping thistle

3. Contractororganisation(s)

NERC Centre for Ecology and HydrologyMonks WoodAbbots RiptonHUNTINGDONCambridgeshirePE28 2LS

54. Total Defra project costs £ 195,000(agreed fixed price)

5. Project: start date................ 1 April 2003

end date................. 31 March 2006

SID 5 (Rev. 3/06) Page 1 of 29

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 Summary7. 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.

1. Invasion by undesirable native and exotic species can impose severe constraints on attempts to restore biodiversity. Creeping thistle is a particular problem in grasslands under agri-environment prescriptions as it responds well to the moderate fertility and less intensive management. Herbicide application is one of the most frequently used method to control thistles. In this study we contrasted herbicide control with less environmentally damaging and more sustainable approaches.

2. Between 2000 and 2005 large-scale randomised block experiments were undertaken to determine the optimum combination of grazing management and mechanical/herbicide treatments to control creeping thistle populations within: (i) a lowland beef cattle (spring/summer) and sheep (autumn/winter) grazing system; and (ii) a semi-upland continuous sheep grazing system.

3. Both experiments compared the effects of factorial combinations of tight versus lenient spring grazing and tight versus lenient autumn grazing on thistle abundance. A further treatment examined the effects of winter sheep grazing. The semi-upland experiment also made limited comparisons of cattle and sheep grazing.

4. Nested with each grazing main treatment were commonly-used weed control sub-treatments which were applied in years 1 and 2 of the study. These included cutting in June and September, wiping with the selective herbicide clopyralid in June, and a combination of cutting in June

6. Thistle abundance increased significantly under tight grazing in the spring, autumn and winter at the lowland site, and under tight spring and winter grazing in the uplands. There was also evidence that sheep grazing increased thistle infestations compared with cattle grazing. This probably reflected a reduction in the competitive ability of the grassland community in the absence of fertiliser additions.

7. The most rapid and effective means of controlling thistles at the lowland site was clopyralid wiping in June followed by lenient spring cattle grazing and lenient autumn sheep grazing. This resulted in a decrease in mean thistle stem abundance from 21.9 ±9.50 m-2 in 2000 to 2.8 ±1.43 m-2 in 2005. At the upland site herbicide wiping and lenient spring sheep grazing resulted in a decrease from 10.8 ±4.11 stems m-2 in 2000 to 0.9 ±0.32 m-2 in 2005.

8. The beneficial effects of lenient grazing and herbicide wiping on thistle abundance had important conflicting effects on the richness of non-target forb species. However, these negative effects tended to be relatively small and did not develop until year 6.

9. Synthesis and applications. We recommend two complementary approaches to the environmentally sustainable control of creeping thistle in permanent pasture: (i) prevention of infestation by best practice management to avoid the creation of large patches of bare ground

SID 5 (Rev. 3/06) Page 2 of 29

which may trigger the initial colonisation by wind-dispersed thistle seedlings. More research is required into the factors controlling initial invasion; and (ii) control of existing infestations by combinations of weed control treatments and grazing management regimes. Lenient livestock grazing to maintain a relatively competitive sward in the spring and autumn is the key to effective long-term control of existing thistle infestations. This approach will not eliminate thistle populations, but maintain them at low levels. If necessary, severe infestations can be rapidly and effectively reduced by weed wiping with clopyralid. However, given the short-lived effect of this approach, it may be better to avoid weed-wiping at more diverse sites and manage thistles purely by grazing.

Project Report to Defra8. 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. 3/06) Page 3 of 29

1. Introduction

Attempts to restore biodiversity are frequently hampered and occasionally prevented by the colonisation of undesirable native species (e.g. Pywell et al. 2002) and invasive non-native species (e.g. D’Antonio & Meyerson 2002; Ormerod 2003). Dominance of these species can lead to the loss of desirable species, changes to community composition and ecosystem function, and irreversible alteration of the physical and chemical environment. Creeping thistle (Cirsium arvense L. (Scop.)) is both an undesirable species throughout much of its native range and an invasive non-native species over large areas of North America, New Zealand, and southeastern Australia where it has become naturalised. In both instances it is a serious perennial weed of land managed for agriculture, forestry and increasingly nature conservation (Nuzzo 2000). The UK Countryside Survey showed that between 1978 and 1990 there was a significant increase in the mean cover of C. arvense from 1% to 10% in fertile grasslands and these high levels of infestation were maintained in the 1998 survey (Bunce et al. 1999). During this period cover also increased on arable land and in boundary habitats, such as hedgerows and adjacent to watercourses, suggesting an increasing potential for invasion. Reforms to agricultural support payments will result in an extensification of grassland management over large areas of the UK (Defra 2005) and elsewhere in Europe. Evidence suggests that invasive species, such as C. arvense, are likely to be an increasing problem in such grasslands, particularly those receiving reduced inputs of fertiliser and managed by grazing alone (Tallowin, Rook & Rutter 2005; BD1440).

C. arvense is of agricultural significance because it reduces pasture production and utilisation (Simpson 1993). It is also of considerable ecological significance: being on the one hand an important food plant for invertebrates and birds (Marshall et al. 2003), and on the other being capable of invading grasslands of high conservation value (Tallowin et al. 2005), with dense infestations often associated with a local reduction in botanical diversity. Effective and environmentally-sustainable weed control strategies are therefore urgently required. The formulation of effective control strategies has been hampered by the capacity of this species to regenerate from both perennial roots and seed. In the case of grasslands, research has primarily focused on four means of control, namely: 1. mechanical (cutting), 2. chemical (herbicides), 3. cultural (grazing and fertiliser), and 4. integrated weed management (combinations of 1 to 3) (reviewed by Donald 1990; Simpson 1993). Many of these approaches require a short-term increase in the intensity of grassland management practices, most of which are known to have strongly negative effects biodiversity (reviewed by Vickery et al. 2001), and are therefore discouraged under agri-environment scheme guidelines (Defra 2005). Repeated cutting can be effective in reducing thistle infestations in grasslands managed intensively for forage production (Amor & Harris 1977). Cutting silage together with increased nitrogen fertiliser application often reduced infestation, but seldom resulted in eradication (Williams 1984). Most research on herbicide control of C. arvense has been short-term and has focused on the arable situation (Donald 1990). There are a large number of post-emergence and a small number of pre-emergence herbicides which are capable of controlling C. arvense (Marshall et al. 2001). However, only those with a high degree of selectivity in their action or those which are effective when applied through targeted application systems, such as weed wipers, would be suitable for use in grassland managed extensively for biodiversity. There have been relatively few studies which have sought to evaluate these more constrained approaches to herbicide-based thistle management. Clopyralid is considered to be one of the more effective and selective herbicides for control of creeping thistle (e.g. Singh & Malik 1992; Clay & Dixon 1998; Pywell et al. 1996). Clopyralid is also effective when applied through a weed-wiper in pastures containing species which are intolerant to sequential multiple applications of this product (Boeroom & Wwyse 1988; Peel & Jones 1997). Cultural control of C. arvense requires careful management of factors such as the timing and intensity of grazing and fertiliser application to control levels of interspecific competition in the sward. For example, the growth and survival of seedlings and shoots are directly affected by defoliation and trampling, and inhibited indirectly by competition from neighbouring plants (e.g. Mitchell & Abernethy 1993; Popay & Field 1996). To date there has been relatively little research into integrated approaches to C. arvense management. Studies in New Zealand suggest that pasture receiving lenient spring sheep grazing, a single application of the herbicide MCPA in mid-summer followed by moderately intensive rotational sheep grazing showed a 95% reduction of the C. arvense population in the following year (Hartley et al. 1984; Mitchell & Abernethy 1995). Other studies have shown that topping in mid-summer followed by tight or lenient rotational grazing reduced creeping thistle numbers by 99% over a two year period (Mitchell & Abernethy 1995).

2. Aim and objectives

The aim of this study was to develop and test integrated weed management systems suitable for the control of C. arvense within temperate grasslands managed extensively under agri-environment agreement to promote biodiversity. In order to achieve this we tested a series of non-exclusive hypotheses in two

4

grasslands: a lowland cattle and sheep grazed system in central England, and an upland sheep-grazed system in Wales. We expected that the hypotheses and thus the ecological mechanisms governing the thistle response to management would be repeatable across these different systems:

Grazing management: (i) it is possible to control thistle populations by lenient grazing management alone but this reduces the diversity of other grassland species; (ii) tight spring grazing reduces the number of thistle shoots (ramets) by defoliation and treading; (iii) tight autumn grazing encourages infestation by reducing the competitive ability of grasses and increasing the amount of bare ground available for recruitment of thistle seedlings; (iv) this effect is increased by winter grazing; (v) tight grazing by cattle is more effective in reducing thistle abundance than sheep due to less preferential grazing and increased shoot damage.Weed control practices: (i) cutting is the most practical and effective means of reducing thistle populations by depleting resources and preventing seed production, whilst not damaging non-target flora; (ii) application of a selective herbicide in June using a weed wiper is an effective means of controlling thistle populations, but is more damaging to non-target species; (iii) a combination of June cutting to prevent seeding and encourage shoot production followed by herbicide application in September when translocation is greater will be more effective than (i) or (ii). Integrated weed control strategies: it is possible to identify synergistic combinations of weed control measures and grazing management practices for the control of C. arvense which do not cause significant damage to non-target species.

The results are discussed in relation to our current understanding of the population biology of C. arvense and the implications for practical and environmentally sustainable weed control strategies under the agri-environment schemes.

3. Materials and methods

3.1 STUDY SITES

Two large-scale complementary experiments were established in April 2000 to compare and evaluate the effects of a range of grazing management regimes and weed control measures on C. arvense populations over six years. Marsh Gibbon, Buckinghamshire (Bucks) (51° 53' N 1° 4'W) represented extensive mixed cattle and sheep systems in lowland central England (67 m asl). IGER Bronydd Mawr Research Station, Powys (51° 58'N 3° 37'W) represented extensive sheep systems in marginal semi-upland conditions in Wales (360 m asl). Both sites were semi-improved grassland with severe thistle infestations at the start of the experiment. In June 2000 the mean thistle shoot density at Bucks was 14.3 (± 1.1) per m 2 and at Powys it was 11.1 (± 0.7) per m2. The grassland at Bucks was species-poor, containing an average of 2.0 (± 0.1) non-target forb species per m2 compared with 9.1 (± 0.2) species per m2 at Powys.

3.2 EXPERIMENTAL DESIGN

3.2.1 Grazing treatments

At Bucks five grazing treatments were applied in a randomised block design with three replicates to plots of c.0.5 ha (Table 1(i)). Four treatments represented factorial combinations of tight versus lenient spring cattle grazing and tight vs lenient autumn sheep grazing. A further treatment added winter sheep grazing to spring tight cattle and autumn lenient sheep grazing. All grazing regimes included lenient summer grazing by cattle. A similar experimental design was applied to Powys with six grazing treatments applied in a randomised block design to plots of c.0.25 ha; smaller plots were used as primarily sheep were grazed here (Table 2(i)). Here, four treatments represented factorial combinations of tight vs lenient spring grazing by sheep and tight vs lenient autumn grazing by sheep. A further management regime added winter sheep grazing to spring tight and autumn lenient grazing. A final treatment comprised tight spring cattle grazing followed by tight autumn sheep grazing. All grazing regimes included lenient summer sheep grazing. The grazing treatments were implemented at both sites in 2000. However, restrictions on livestock movement imposed by the national Foot and Mouth Disease outbreak prevented successful implementation of grazing management at Bucks in spring and summer 2001. Grazing treatments were successfully re-established in autumn 2001. Finally, the tendency for prolonged winter water logging on the poorly drained soils in Bucks meant that the winter grazing treatment was discontinued after 2003 because of the concerns over animal welfare. At both sites a system of continuous stocking was employed with the target sward heights (Tables 1(i) and 2(i)) maintained by adding and removing animals on the basis of regular sward surface measurements with a HFRO sward stick (Stewart et al. 2001). In practice this resulted in stocking rates of approximately 0.75

5

LSU/ha/year (equivalent to 5 ewes/ha/year or 0.75 cattle over 24 months of age/ha per year) which is appropriate for lowland semi-improved mesotrophic grassland and semi-upland acid grassland (Crofts & Jefferson 1999).

In both experiments grazing treatment 4 was the control (i.e. the grazing intensities were close to standard grazing recommendations for the farming regions of each site). Treatment 4 was compared with treatments 1, 2 and 3 to elucidate the effects of different grazing pressures (tight vs lenient) during the grazing season (spring vs autumn) on thistle abundance. Tight spring grazing should reduce the number of thistle shoots by defoliation and treading. Tight grazing in the autumn might reduce the competitive ability of grasses in the following spring and increases the amount of bare ground available for seedling recruitment. Similarly, treatment 2 was compared with treatment 5 to elucidate the effect of stocking during the winter. This may intensify the effects of tight autumn grazing. Finally, treatment 6 at Powys was compared with treatment 1 to determine if cattle grazing was more effective in controlling thistle numbers than sheep grazing through less selective defoliation and greater damage by treading.

3.2.2 Weed control sub-treatments

Nested within each grazing treatment were a series of weed control sub-treatments applied to plots measuring 10 10 m with at least a 5 m separation to minimise possible interference effects, including the translocation of herbicide. Five of the sub-treatments were common to both sites (Tables 1(ii) & 2(ii)). One control sub-treatment (A) was sited in areas of low density infestations of Cirsium arvense at the start of the experiment to examine the rate and mechanism of thistle invasion under different grazing management regimes. At Bucks the shoot density in this treatment was 1.4±0.29 per m -2 and at Powys it was 1.9±0.30 per m-2 (means of all grazing treatments). The remaining treatments examined the control of severe infestations of thistles by herbicide and mechanical means, and the interaction with grazing management. A second control (B) was situated in high density thistle infestations to investigate the effects of severity and timing of grazing on the dynamics and control of established populations of C. arvense. At Bucks the shoot density in this treatment was 14.5±3.07 per m-2 and at Powys it was 10.1±1.01 per m-2 (means of all grazing treatments). Mechanical thistle control was by cutting in June and again in September using a tractor-mounted flail mower with the cut material left in situ (sub-treatment C). Thistle control using herbicide was tested by applying the selective product clopyralid (Dow Shield, Dow AgroSciences Ltd.) to the thistles in June (sub-treatment D) using a trailed weed wiper (C-Daxtm Eliminator, C-Dax Europe Ltd., Wymondham, Norfolk, NR18 9LR, UK). There is currently no label recommendation for wiper application of this product so the dilution chosen (1 part herbicide to 4 parts water) was based on the findings from previous research (Miller et al. 1997). Subsequent studies have found much lower dilutions of clopyralid (1 part herbicide to 19 parts water) give equally effective thistle control (Pywell et al. 2003). A further sub-treatment (E) combined June cutting and September herbicide application to examine synergistic effects of weed control measures. Evidence suggests that translocation of herbicide to the root system is more effective in the late summer, giving enhanced, long term control compared to summer application of herbicide. Finally, sub-treatment F was exclusive to Bucks. This investigated the effectiveness of the regional practice of rotational hay cutting one year in three for the control of C. arvense on heavy soils. In winter 2000 enclosures were erected around these plots and a hay cut taken in late June 2001. All of the other weed control sub-treatments were applied in 2000 and again in 2001. No sub-treatments were applied after this to enable an examination of the recovery of thistle populations following the cessation of weed control.

3.3 MONITORING

Each September between 2000 and 2005 the number of C. arvense stems present in ten randomly placed 1 1 m quadrats was recorded within each 10 10 m sub-treatment plot. In addition, effects on non-target forb species was investigated by recording rooted frequency in three 1 1 m quadrats sub-divided into 25 cells of 20 20 cm located at random within each sub-treatment plot in 2000, 2001, 2002 and 2005. Nomenclature of vascular plants follows Stace (1997).

3.4 STATISTICAL ANALYSIS

Both mean thistle shoot counts and the mean number of non-target forb species per m2 were calculated for each sub-treatment plot for each time period. In addition, the mean rooted frequency of individual forb species present in >50% of the sub-treatment plots was calculated. Thistle shoot counts and species richness were square root transformation prior to analyses following an assessment of the normality of residuals (PROC UNIVARIATE SAS for windows version 9.1; SAS 2004). The frequency of individual species was arcsine transformed. Both experiments were analysed separately using analysis of variance

6

(ANOVA) with a split-plot design comparing grazing main treatments and weed control sub-treatments tested against the block × grazing treatment interaction using PROC GLM in SAS. Repeated measures ANOVA with polynomial contrasts were used to assess pattern of changes over time. Tukey’s multiple comparison tests were used to compare different control sub-treatments. In order to test the experimental hypotheses three separate ANOVA models were used:

Analysis 1. Seasonal grazing intensity and weed control measures. This ANOVA compared fully factorial combinations of spring lenient vs tight grazing and autumn lenient vs tight grazing (treatments 1-4) combined with weed control sub-treatments on thistle abundance and non-target forb species for each site and recording date (Tables 1 & 2). Main treatment effects provided tests of hypotheses H1(i) to (iii), sub-treatment effects provided a test of H2 (i) to (iii), and the interaction term provided a test of H3.

Analysis 2. Winter grazing. The effect of winter grazing on thistle abundance and non-target forbs was examined by comparing treatment 2 (T-L-L) with 5 (T-L-L-winter) for 2000 to 2003 for Bucks and 2000 to 2005 for Powys. This provided a test of H1(iv).

Analysis 3. Grazing animal. The effects of cattle vs sheep grazing on thistle abundance and non-target forbs was examined at Powys by an ANOVA comparing treatment 1 (sheep T-L-T) with treatment 6 (cattle T-L-T). This provided a test of H1(v).

4. Results

4.1 EFFECTS OF GRAZING TREATMENTS ON CIRSIUM ARVENSE ABUNDANCE

Analysis 1 showed there was a gradual build up of seasonal grazing effects on Cirsium arvense abundance over the six years at both sites (Table 3a,b). At Bucks polynomial contrasts showed a significant linear increase with time in the response to autumn grazing (F1,6=6.4, P<0.05). Lenient grazing in both seasons resulted in a significant (>50%) reduction in thistle abundance from 2003 (autumn) and 2004 (spring) onwards (Table 3a), whereas tight grazing in spring or autumn maintained thistle abundance at the same high level as at the start of the experiment (Fig. 1a). Importantly, lenient spring and autumn grazing alone was sufficient to prevent the increase of thistle populations in the low density thistle infestation control sub-treatment (A) plots (Fig. 2a). At Powys polynomial contrasts showed a quadratic change over time in the response to spring grazing (F1,6=10, P<0.05), suggesting a more sudden onset of grazing effects than at Bucks. Once again lenient grazing in spring resulted in a significant (>50%) reduction in thistle abundance from 2003 onwards compared with tight grazing (Table 3b). However, unlike Bucks, autumn grazing had no effect on thistle abundance and only tight grazing in spring increased thistle numbers (Fig. 1b). The effectiveness of lenient spring and autumn grazing in preventing the build up of thistles in the low density thistle infestation control was therefore less clear cut (Fig. 2b). Finally, at both sites there were no significant interactions between spring and autumn grazing which confirmed the effects of grazing treatments were additive.

Analysis 2 showed there was a linear increase over time in the effect of winter grazing on thistle abundance at the Bucks site (F1,2=18, P<0.05). By 2003 winter grazing had a similar effect to tight autumn grazing in maintaining high densities of thistles (Table 3a; Fig. 3a). At Powys there was a comparable effect of winter grazing on thistle abundance, but this did not develop until 2005 and there was no pattern over time (Table 3b; Fig. 3b). Finally, there were highly significant effects of the type of grazing animal on thistle abundance at Powys (Analysis 3; Table 3b) and polynomial contrasts showed a linear increase with time (F1,2=16, P<0.05). After 2003 tight autumn cattle grazing resulted in significant decreases (>50%) in thistle abundance compared with tight autumn sheep grazing (Fig. 4).

4.2 EFFECTS OF WEED CONTROL SUB-TREATMENTS ON CIRSIUM ARVENSE ABUNDANCE

The effects of weed control sub-treatments on Cirsium arvense abundance were also examined by Analysis 1. Polynomial contrasts for Bucks showed a significant quadratic change over time in the response to weed control measures (F5,40=9.9, P≤0.001). This reflected the rapid decrease in control effects on thistle abundance after application ceased in 2001, so that no effects remained by 2005 (Table 3a). Overall, the herbicide weed wiping sub-treatment (D) was always the most effective means of reducing thistle abundance (Fig. 5a). Both the cutting (C) and the cutting and herbicide wiping (E) sub-treatments were approximately equal in their effectiveness in controlling thistles. Finally, the rotational hay cutting (F) sub-treatment yielded

7

an average of 2.7±0.29 t Dm ha-1 in 2001. This was the least effective sub-treatment and was only slightly better than the control (B).

At Powys the polynomial contrasts showed a linear change over time in the response to weed control sub-treatments (F4,32=8.1, P≤0.001). This reflected a more gradual decline in effects after application ceased (Table 3b). However, by 2004 no significant weed control effects remained. At this site herbicide wiping (D) was also the most effective means of reducing thistle abundance (Fig. 5b). However, unlike Bucks, the cutting and wiping (E) sub-treatment was significantly more effective than cutting alone (C). In this case cutting was only slightly better than the control (B).

4.3 INTERACTIONS BETWEEN GRAZING AND WEED CONTROL

Analysis 1 for both sites showed there were no significant interactions with grazing main treatments and weed control sub-treatments either in any year or with repeated measures. This suggests that the effects of grazing management and weed control measures were additive in both situations. Therefore the most effective way of controlling thistles in the lowland sheep and cattle system in Bucks was herbicide wiping in June (D) followed by lenient spring cattle grazing and lenient autumn sheep grazing (Fig. 6a). This regime resulted in a decrease in mean thistle stem abundance from 21.9 ±9.50 m-2 in 2000 to 2.8 ±1.43 m-2 in 2005. The most effective means of controlling thistles in the semi-upland sheep system in Powys was once again herbicide wiping, but in this case followed by lenient spring sheep grazing (Fig. 6b). This regime resulted in a decrease in mean thistle stem abundance from 10.8 ±4.11 m-2 in 2000 to 0.9 ±0.32 m-2 in 2005.

4.4 EFFECTS OF GRAZING AND WEED CONTROL ON NON-TARGET FORB SPECIES

There were no significant effects of grazing main treatments or weed control sub-treatments on the overall species richness of non-target forb species in Bucks (Analyses 1 & 2; Table 4a). This probably reflected the low number of forb species present at this site. In contrast, grazing regime had a number of significant impacts on non-target forb richness at Powys (Table 4b). Lenient spring grazing had a significant negative effect on forb richness (Fig. 7), but this effect did not develop until year 6 of the experiment. Hard grazing by sheep in the autumn had a more rapid and consistent beneficial effect on forb richness compared with tight cattle grazing (Table 4b). Finally, herbicide application in June using a weed wiper had a small, but long-term negative effect on forb numbers which extended into 2005 (Fig. 8).

There were a large number of grazing and weed control effects on the rooted frequency of individual species. For simplicity, only clear, consistent effects which developed over time are described. In Bucks there were no effects of grazing treatment on individual species. However, herbicide wiping (D) and cutting and wiping (E) resulted in overall significant reductions in Ranunculus repens (L.) (F5,40=3.7, P<0.01) and Triolium repens (L.) compared with other control sub-treatments. In the case of T. repens, this effect was significant in 2002 and 2003, but not at in 2001 or 2005, leading to an overall effect (F 5,40=9.3, P≤0.001) and a control × year interaction (F15,120=3.1, P≤0.001).

At Powys the overall the frequency of Achillea millefolium (L.) was significantly decreased by lenient spring (F1,6=7.2, P<0.05) and autumn (F1,6=7.7, P<0.05) grazing. In 2005 the mean frequency under spring lenient grazing was 0.8±0.23 compared with 3.6±0.84 under spring tight grazing. Similarly, frequency under autumn lenient grazing was 1.3±0.44 compared with 3.1±0.81 under autumn tight grazing. T. repens responded in a similar way, with significant reduction under lenient spring (F1,6=6.3, P<0.05) and autumn (F1,6=6.2, P<0.05) grazing. These effects increased over time giving rise to significant spring × year (F3,18=12.7, P≤0.001) and autumn × year (F3,18=6.2, P<0.01) interactions. In 2005 mean frequency under spring lenient was 14.6±1.25 compared with 21.2±1.09 under spring tight grazing. Similarly, under autumn lenient the frequency was 16.3±1.41 compared with 19.5±1.23 under autumn tight grazing. The frequency of three other non-target forbs was significantly reduced by lenient spring grazing alone in 2005: Leontodon autumnalis (L.) (F1,6=11.3, P<0.05); R. repens (F1,6=6.7, P<0.05), with a spring × year interaction (F3,18=3.1, P<0.05); and Taraxacum officinale agg. (F1,6=9.5, P<0.05), also with a spring × year interaction (F3,18=5.4, P<0.01). The frequency of L. autumnalis under spring lenient grazing was 0.7±0.21 compared with 4.9±0.82 under tight grazing. Similarly, R. repens frequency was 3.4±0.49 under lenient grazing compared with 5.2±0.52 under tight. Finally, T. officinale frequency under lenient was 1.7±0.37 compared with 5.1±0.61 under tight grazing.

There was an overall significant effect of weed control sub-treatments on T. repens (F4,32=7.6, P≤0.001), with frequency lowest in the herbicide wiped plots. Also, the frequency of Cerastium fontanum (Baumg.) was lowest in the low infestation control sub-treatment (F4,32=4.1, P<0.01). There were no significant effects of grazing treatments or weed control sub-treatments on the rooted frequency of the following species which

8

occurred in >50% of plots at Powys: Bellis perennis (L.), Cardamine pratensis (L.), Plantago lanceolata (L.), Prunella vulgaris (L.), R. acris (L.), Rumex acetosa (L.) and Veronica serpyllifolia (L.).

5. Discussion

5.1 EFFECTS OF GRAZING ON CIRSIUM ARVENSE AND NON-TARGET FORBS

This study has shown that the abundance of Cirsium arvense in permanent pasture was significantly increased by tight grazing. In the lowland cattle and sheep system this was the result of tight grazing in spring, autumn or winter, whereas in the upland sheep system it was due to tight winter or spring grazing only. Intensive rabbit grazing in semi-natural acid grassland also caused thistles to increase (Edwards et al. 2000). Also, the biennial weed of pastures, C. vulgare (Savi) Ten., showed a similar increase in abundance in response to tight sheep grazing in the summer, winter or spring (Bullock et al. 1994; Bullock et al. 2001). Grazing may affect the population and community ecology of pasture plants through a variety of mechanisms (Duffey et al. 1974; Bullock & Marriott 2000). These include: 1) creation of vegetation gaps for seedling germination and recruitment; 2) direct damage by defoliation, trampling and dung and urine deposition; and 3) localised pulses of nutrients resulting from dung and urine allowing greater survival and growth. Defoliation will also have important indirect affects through the alteration of the intensity of competition for resources and relative competitive abilities of other grassland species (Bullock 1996).

Gap creation by tight grazing is unlikely to result in increased recruitment of Cirsium arvense seedlings in permanent pastures. Thistle seedlings are very susceptible to shading and competition (Bakker 1960) and as a result recruitment in plant communities with such dense cover is likely to be a rare event (Bostock & Benton 1983). Most authors therefore conclude that the recruitment of C. arvense seedlings is only important in the early stages of colonisation (Solé et al. 2004). This was confirmed by detailed demographic studies undertaken at the Bucks site between 2001 and 2002 which found only five seedlings in quadrats covering 135 m2, none of which were recruited into the population (Pywell et al. 2003). The rate of damage by defoliation and trampling within the thistle populations was also recorded in Bucks in June 2002 (R. Pywell, unpublished data). The proportion of trampled or grazed stems was recorded in ten 1× 1 m quadrats placed randomly within each weed control sub-plot. Proportions of trampled or grazed stems were relatively low in both tight spring grazed (grazed 0.1% trampled 0.5%) and lenient grazed treatments (grazed 0.03%, trampled 0.01%). ANOVA showed no significant difference in either damage or defoliation between grazing treatments (grazed F1,6 = 1.2, P>0.05; trampled F1,6 = 1.3, P>0.05). This suggests that even intensive cattle grazing is unlikely to have direct damaging effects on the fitness of thistle populations. Detailed demographic studies (Pywell et al. 2003) showed that grazing damage typically occurred in spring when the newly emerged thistle stems were soft and palatable, but plants often recovered and soon became tall and prickly, making them much less susceptible to grazing later in the season (Simpson 1993).

It is therefore likely that grazing management primarily affects thistle shoot abundance through alteration of the competitive ability of grasses within the sward. Both grasslands in this study were extensively managed under the agri-environment schemes and received no inorganic fertiliser or farmyard manure to boost grass growth in the spring. This would have severely impeded the ability of the grassland to recover from tight grazing, leaving it open and relatively uncompetitive at the time of thistle emergence. There were few differences in soil fertility between the two sites (Bucks Phosphorous (P) <10 mg l -1, Potassium (K) 121-180 mg l-1; Powys P <10 mg l-1, K 61-120 mg l-1). It is therefore likely that the comparatively better soil drainage of the Powys site would have allowed the grass sward to recover more rapidly from tight autumn grazing than the Bucks site. Similarly, tight cattle grazing resulted in a taller (5-7 cm) and therefore more competitive sward than tight sheep grazing (3-5 cm) which would explain the greater effectiveness of cattle grazing in controlling thistle abundance compared with sheep. Indeed studies in North America have found that sowing perennial competitive grass species were as effective in controlling creeping thistle as yearly applications of clopyralid herbicide at 0.55 kg ha-1 (Wilson & Kachman 1999).

The beneficial effects of lenient grazing on thistle abundance had important conflicting effects on the richness of non-target forb species. This is unsurprising as many forb of fertile grasslands will have a similar ecology to C. arvense, in particular the ability to spread by vegetative means (Pywell et al. 2003). They would therefore be expected to respond to grazing in a similar way. However, these negative effects of grazing tended to be relatively small and did not develop until year 6 of the experiment. There are relatively few long-term studies of plant community response to changes in grazing intensity and livestock type (Bullock & Marriott 2000). Most concur with our findings that perennial grassland communities respond slowly to different grazing intensities (e.g. Smith et al. 2000; Bullock et al. 2001), and the abundance of forb species generally increases under heavier grazing, but there is little evidence of treatment effects through the colonisation of new species (e.g. Kydd 1964; Bullock et al. 1994). Overall the pattern of response of

9

individual non-target species tended to reflect the co-incidence of gap creation resulting from tight grazing and the season of germination. This would also explain the long time delay in the emergence of a grazing effect. Spring and summer/autumn germinating species, such as Achillea millefolium and Trifoilum repens (Grime et al. 1988; Burdon 1983), increased under tight spring and autumn grazing, whereas spring germinating species (e.g. Ranunculus repens, Taraxacum officinale) increased under tight spring grazing (Sarukhan & Harper 1973; Bostock & Benton 1979). Finally, the more rapid and consistently beneficial effect of hard autumn sheep grazing compared with cattle probably reflects increased gaps and lower competitive ability of the sward under this regime.

5.2 EFFECTS OF CUTTING AND HERBICIDE ON CIRSIUM ARVENSE AND NON-TARGT FORBS

The effects of all weed control sub-treatments were rapidly lost from both sites. It is possible that this is an artefact of small-plot experiments and reflects the rapid re-colonisation of thistles from the plot edges. Creeping thistle can spread vegetatively by up to 6 m per year, but most patches spread at the rate of 1-2 m year-1 (Nuzzo 2000). However, in this study casual observations suggested this was not the case. June application of the selective herbicide clopyralid using a weed wiper gave the most rapid and effective short-term control of C. arvense. This herbicide is known to be highly toxic to members of the Asteraceae family (Pywell et al. 1996; Clay & Dixon 1998) either when applied as foliar spray or through a weed wiper (Miller et al. 1997). Spring grazing followed by herbicide application in June resulted in a large differential between the grassland sward and the elongating thistle stems. This ensured that many stems received a lethal dose of herbicide whilst reducing the potential damage to non-target species (Crofts & Jefferson 1999). Other studies have found inconsistency between sites in any beneficial effects of mowing creeping thistle prior to autumn application of clopyralid (Beck & Sebastian 2000). This approach was most effective in situations where the growth of thistle roots was restricted by factors such thin soils or high water table. In this study it is probable that cutting and weed wiping was less effective because a greater proportion of thistle shoots remained as rosettes below the height of the wiper due to a combination of relatively low fertility and dry summer weather. Wiping had negative effects on richness of non-target forbs, particularly Trifolium repens and Ranunculus repens. Members of the Fabaceae are known to be susceptible to damage by clopyralid (Anon. 2003). Conversely R. repens is known to be resistant to clopyralid (Lovett-Doust et al., 1990) and the most likely explanation for this negative effect is the very high concentration of herbicide used in the weed wiper.

Cutting thistles in June and September was one of the least effect weed control measures. In theory this treatment should deplete plant resources, prevent seed production and, by removing apical dominance, cause the re-growing plant to send up a greater number of shoots than previously. Cutting of these secondary shoots in September should further deplete resources and expose the plants to increased risk of mortality during the winter months (Haggar et al. 1986; Hurrell & Bourdot 1996; Mitchell & Davis 1996). However, it is likely that repeated cutting of the grassland in the absence of fertiliser application would have had a similar effect of reducing the competitive ability of the sward, leaving it open for emerging thistle stems in the following spring. Finally, rotational hay cutting one year in three was probably insufficient to deplete the resources of the thistle and, in the absence of fertiliser or manure, the grass sward was once again insufficiently competitive in the critical early spring phase.

5.3 RECOMMENDATIONS FOR SUSTAINABLE WEED MANAGEMENT STRATEGIES

Our results suggest two complementary approaches to the environmentally sustainable control of creeping thistle in permanent grassland. The first is the prevention of infestation by best practice management to avoid the creation of large patches of bare ground which may trigger the initial colonisation by wind-dispersed thistle seedlings. More research is required into the factors controlling initial invasion of pastures, but severe livestock poaching, disturbance around water troughs, winter feeding and re-seeding are all likely to be important triggers of colonisation. The second is the control of existing infestations by combinations of weed control treatments and grazing management regimes. Lenient livestock grazing (cattle 8-10 cm; sheep 6-8 cm) to maintain a relatively dense and competitive grassland sward in the spring and autumn is the key to effective control of existing thistle infestations over a 5 year timescale. There is also a suggestion that cattle grazing may give more effective control than sheep because of the resultant taller and more competitive sward. This approach will not eliminate the thistle populations, but maintain them at low levels of about 2 stems per m-2. Severe infestations (>10 shoots m-2) can be rapidly and effectively controlled by weed wiping with selective herbicides, such as clopyralid. However, the effects of herbicide weed control measures fade very rapidly once application ceases and the lack of grazing × weed control interactions suggests that lenient grazing alone will give equally effective control in the longer term. Moreover, given the temporary effect of wiping on thistles, this suggests it may be better to avoid weed-wiping at more diverse sites and manage thistles purely by grazing. Finally, it is important to consider that any beneficial effects of lenient

10

grazing and herbicide application on thistle abundance will have important conflicting effects on several non-target forb species. However, these effects were generally small and the benefit gained by thistle control may out-weigh any negative impacts on biodiversity.

6. Future research requirements

Future research is required in the following areas:

6.1 UNDERSTANDING THE PROCESS OF GRASSLAND INVASION BY CREEPING THISTLE

A fundamental principle of weed research is that prevention of invasions is both much easier and cheaper than controlling established infestations. This is true both at the site and landscape level. However, very little is known about the process of grassland invasion by creeping thistle. Such information is fundamental to the formulation of effective preventative guidelines for sites that are not currently infested. Replicated experiments are required to investigate the role of the following factors in the invasion of grassland by creeping thistle:- Type, intensity and timing of disturbance associated with agricultural activities (e.g. poaching, feeding, re-seeding)- grazing regime- sward structure- soil nutrients- seed supply and dispersal.

Similarly, information is required on the potential for vegetative spread by lateral roots from point sources of invasion. This could be investigated in simple Garden Experiments hosted at a number of the consortium research sites chosen to represent a range of climatic regimes. This would comprise growing thistle clones taken from upland and lowland populations in large pots (mesocosm) containing soils of different nutrient composition and vegetation which would simulate a range of grassland types. This approach would enable detailed measurements of the potential for lateral root expansion and fragmentation. These data could be used to develop simple models of thistle spread in order to identify the key processes and critical stages for application of environmentally sustainable control measures.

6.2 STAKEHOLDER CONSULTATION OVER WEED CONTROL

The development of effective thistle control methods requires linking research findings with the requirements of grassland managers and other stakeholders. A survey could be conducted of stakeholder opinions in upland and lowland regions of (i) current levels of thistle infestation under different grazing systems; (ii) current weed control options both inside and outside the agri-environment schemes; (iii) novel control strategies being developed locally, and (iv) integrated weed control strategies being developed as part of this project. The survey would address practical, agronomic, ecological and animal welfare issues and aim to identify what different groups consider an acceptable level of thistle infestation. The findings of this consultation could be used with the thistle control strategies arising from the long-term replicated experiment described in this report and the controlled studies on invasion risk (6.1 above) and the potential of 3 to refine agri-environmental agreements/guidelines. These results would form the basis for recommendations for environmentally sustainable weed control strategies.

7. Technology transfer

RDS / Defra National Grassland Management Training Courses: Somerset May 2002, June 2003, March 2005; Oxford March 2006.

RDS Weed Control Discussion meeting, Bury St. Edmunds October 2005

Natural England Training Course, Bakewell 2006

RDS / Defra visit to the Upper Tames Tributaries ESA, 14 June 2002.

RDS / Defra R&D workshop at Bronydd Mawr July 2003.

11

IGER Bronydd Mawr open days 2003-2005

Masters, G. J. & Ward, N. L. (2003) Galls on thistles. Biologist 50 (1), 25-28.

Ward, N.L. & Masters, G.J. (2002) Interactions between spring and autumn grazing on creeping thistle stem borer abundance; importance of inter-site and inter-annual variation. BES Winter Meeting Proceedings.

8. Acknowledgements

This study was funded by commissions (BD1437 and BD1449) from the Department for Environment, Food and Rural Affairs. The authors are grateful to Roy Lambourne, Derek Smith and John Benfield for assistance with managing the experiment in Buckinghamshire. We also thank Geraint Evans and the staff at IGER Bronydd Mawr for managing the Powys experiment. We wish to thank Lindsay Maskell, Phil Croxton, Ibby Moy and Rob Large for assistance in the field. Finally, we are grateful to Simon Smart for analysis of Countryside Survey data.

9. References

Amor, R.L. & Harris, R.V. (1977) Control of Cirsium arvense (L.) Scop. by herbicide and mowing. Weed Research, 17, 303-309.

Anon. (2003) Dow Shield Leaflet Label. Dow AgroSciences Limited, Hitchin, Hertfordshire, UK.

Bakker, D. (1960) A comparative life-history study of Cirsium arvense (L.) Scop., Tussilago farfara L., the most troublesome weeds in the newly reclaimed polders of the former Zuiderzee. The Biology of Weeds (ed. J.L. Harper), pp. 205-22. Blackwell Scientific Publications, Oxford.

Beck, K.G. & Sebastian, J.R. (2000) Combining mowing and fall-applied herbicides to control Canada Thistle (Cirsium arvense). Weed Technology, 14, 351-356.

Boeroom, C.M. & Wwyse, D.L. (1988) Selective application of herbicides for Canada Thistle (Cirsium arvense) in birdsfoot trefoil. (Lotus corniculatus). Weed Technology, 2, 183-186.

Bostock, S.J. & Benton, R.A. (1983) Dry weight costs and establishment of seeds and vegetative propagules. Acta Oecologia, 4, 61–69.

Bostock, S.J. & Benton, R.A. (1979) The reproductive strategies of five perennial Compositae. Journal of Ecology, 67, 91-107.

Bullock, J.M., Clear Hill, B., Dale, M.P., Silvertown, J. (1994) An experimental study of vegetation change due to sheep grazing in a species-poor grassland and the role of seedling recruitment into gaps. Journal of Applied Ecology, 31, 493–507.

Bullock, J.M. (1996) Plant competition and community dynamics. The Ecology and Management of Grazing Systems (eds A. Illius & J. Hodgson), pp. 69–100. CAB International, Wallingford, UK.

Bullock, J.M. & Marriott, C.A. (2000) Plant responses to grazing and opportunities for manipulation. Grazing Management (eds A.J. Rook & P.D. Penning), pp. 17–26. British Grassland Society, Reading, UK.

Bullock, J.M., Franklin, J., Stevenson, M.J., Silvertown, J., Coulson, S.J., Gregory, S.J., & Tofts, R. (2001) A plant trait analysis of responses to grazing in a long-term experiment. Journal of Applied Ecology, 38, 253-267.

Bunce, R.G.H., Smart, S.M., van de Poll, H.M., Watkins, J.W. & Scott, W.A., (1999) Measuring Change in British Vegetation. ECOFACT volume 2. London: Department of the Environment, Transport and the Regions.

Burdon, JJ (1983). Biological flora of the British Isles. Trifolium repens. Journal of Ecology, 71, 307-330.

12

Clay, D.V. & Dixon, F.L. (1998) The susceptibility of Compositae weed species to Clopyralid. Annals of Applied Biology, supplement, 132, 32-33.

Crofts, A. & Jefferson, R.G. (1999) The Lowland Grassland Management Handbook, 2nd edn. English Nature and the Wildlife Trusts, Peterborough, UK.

D’Antonio, C. & Meyerson, L.A. (2002) Exotic plant species as problems and solutions in ecological restoration: a synthesis. Restoration Ecology, 10, 703-713;

Defra (2005) http://www.defra.gov.uk/erdp/schemes/els/default.htm

Donald, W.W. (1990) Management and control of Canada thistle (Cirsium arvense). Reviews of Weed Science, 5, 193-250.

Duffey, E., Morris, M.G., Sheail, J. Ward, L.K., Wells, D.A. & Wells, T.C.E. (1974) Grassland ecology and wildlife management. Chapman & Hall, London, UK.

Edwards, G.R., Bourdot, G.W. & Crawley, M.J. (2000) Influence of herbivory, competition and soil fertility on the abundance of Cirsium arvense in acid grassland. Journal of Applied Ecology, 37, 321-334.

Grime, J.P., Hodgson, J.G. & Hunt, R. (1988) Comparative Plant Ecology: A Functional Approach to Common British Species. Unwin Hyman Ltd., London, UK.

Haggar, R.J., Oswald, A.K. & Richardson, W.G. (1986) A review of the impact and control C. arvense in grassland. Crop Protection, 5, 73-76.

Hartley, M. J., Lyttle, L. A. & Popay, A. I. (1984) Control of Californian thistle bygrazing management. Proc. 37th N.Z. Weed and Pest Control Conf., 24-27.

Hurrell, G. A. & Bourdot, G. W. (1996) Sclerotinia sclerotiorum and mowing independently reduce Californian thistle in a sheep pasture. Proc. 49th N.Z. Plant Protection Conf., 225-228.

Kydd, D.D. (1964) The Effect of Different Systems of Cattle Grazing on the Botanical Composition of Permanent Downland Pasture. Journal of Ecology, 52, 139-149.

Ovett-Doust, L., Ovett-Doust, J. L. & Roth Th L. G. (1990) The biology of Canadian weeds: 95. Ranunculus repens . Canadian Journal of Plant Science, 70, 1123-1141.

Marshall, E.J.P., Brown, V.K., Boatman, N., Lutman, P. & Squire, G. (2001) The impact of herbicides on weed abundance and biodiversity. PN0940. A report for the UK Pesticides Safety Directorate. IACR-Long Ashton Research Station, UK, 134 pp.

Marshall, E.J.P., Brown, V.K., Boatman, N.D., Lutman, P.J.W., Squire, G.R., & Ward, L.K. (2003) The role of weeds in supporting biological diversity within crop fields. Weed Research, 43, 77-89.

Miller, P.C.H., Pywell, R.F., Lane, A.G., Walker, K.J. & Nowakowski, M. (1997) The control of injurious weeds of herb-rich grassland in Environmentally Sensitive Areas. OC9321. A report to the Ministry of Agriculture, Fisheries and Food, UK, 71 pp.

Mitchell, R. B. & Abernethy, R. J. (1993) Integrated management of Californianthistle in pasture. Proc. 46th N.Z. Plant Prot. Conf., 278-281.

Mitchell, R. B. & Abernethy, R. J. (1995) The effect of topping and repeat grazings on Californian thistle and pasture production. Proc. 48th N.Z. Plant Prot. Conf., 189-193.

Mitchell, R. B. & Davis, L. T. (1996) Effect of control treatments on the root reserves of Californian thistle (Cirsium arvense). Proc. 48th N.Z. Plant Prot. Conf., 229-233.

Nuzzo, V. (2000) Element stewardship abstract: Cirsium arvense. Weeds on the Web: The Nature Conservancy Wildland Invasive Species Program. Available at http://tncweeds.ucdavis.edu/esadocs/cirsarve.html

13

Ormerod, S.J. (2003) Restoration in applied ecology: editor’s introduction. Journal of Applied Ecology, 40, 44-50.

Peel S. & Jones A.T. (1997) Control of C. arvense on an ancient grazed common. Grassland Management in Environmentally Sensitive Areas (ed. R.D. Sheldrick), pp. 276-277. Occasional Symposium No. 32, British Grassland Society, Reading, UK.

Popay, I. & Field, R. (1996) Grazing animals as weed control agents. Weed Technology, 10, 217-231.

Pywell, R.F., Nowakowski, M, Allchin, E.A. & Aldous, E. (1996). Preliminary studies of the effects of selective herbicides on wild flower species. Aspects of Applied Biology, 44, 149-156.

Pywell, R.F., Bullock, J.M., Hopkins, A., Walker, K.J., Sparks, T.H., Burke, M.J.W. & Peel, S. (2002) Restoration of species-rich grassland on arable land: assessing the limiting processes using a multi-site experiment. Journal of Applied Ecology, 39, 294-310.

Pywell, R.F. Bullock, J.M., Hayes, M., Tallowin, J.B., & Masters, G. (2003) Effects of grazing management on creeping thistle and other injurious weeds and integration of grazing with weed control . BD1437. A report to the Department of Environment, Food and Rural Affairs, UK, 38pp.

Sarukhan, J. & Harper, J.L. (1973) Studies on plant demography: Ranunculus repens L., R. bulbosus L. and R. acris L. 1. Population flux and survivorship. Journal of Ecology, 61, 675–716.

SAS (2004) SAS/STAT 9.1 users guide. SAS Institute. Carey, North Carolina.

Simpson, N. A. (1993) A summary review of information on the autecology and control of six grassland weed species. English Nature research report No. 44, Peterborough, UK, 165 pp.

Singh, S. & Malik, R.K. (1992) Evaluation of clopyralid against Cirsium arvense. Annals of Applied Biology, supplement, 120, 46-47.

Smith, R.S., Shiel, R.S., Millward, D. & Corkhill P. (2000) The interactive effects of management on the productivity and plant community structure of an upland meadow: an 8-year field trial. Journal of Applied Ecology, 37, 1029-1043.

Solé, M., Durka, W., Eber, S. & Brandl, R. (2004) Genotypic and Genetic Diversity of the Common Weed Cirsium arvense (Asteraceae). International Journal of Plant Sciences, 165, 437–444.

Stace, C. (1997) New flora of the British Isles. Cambridge University Press, Cambridge, UK.

Stewart, K. E. J., Bourn, N. A. D., Thomas, J.A. (2001). An evaluation of three quick methods commonly used to assess sward height in ecology. Journal of Applied, 38, 1148-1154.

Tallowin, J.R.B., Rook, A.J. & Rutter, S.M. (2005) Impact of grazing management on biodiversity of grasslands. Animal Science, 81, 193-198.

Vickery, J.A., Tallowin, J.R., Feber, R.E., Asteraki, E.J., Atkinson, P.W., Fuller, R.J. & Brown, V.K. (2001) Distribution and abundance of breeding and wintering birds and their food resources with special reference to recent changes in fertiliser, mowing and grazing practices on lowland neutral grasslands in Britain. Journal of Applied Ecology, 38, 647-664.

Williams, R.D. (Editor) (1984) Crop protection handbook – Grass and clover swards. British Crop Protection Enterprises, UK, 105 pp.

Wilson, R.G. & Kachman, S.D. (1999) Effect of Perennial Grasses on Canada Thistle (Cirsium arvense) Control. Weed Technology, 13, 83-87.

14

11. Tables and Figures

Table 1. Grazing management main treatments (a) and weed control sub-treatments (b) at Bucks.

a) Grazing management main treatments

Spring Summer Autumn Winter1 Cattle tight (5-7 cm) Cattle lenient (8-10 cm) Sheep tight (3-5 cm) No2 Cattle tight (5-7 cm) Cattle lenient (8-10 cm) Sheep lenient (6-8 cm) No3 Cattle lenient (8-10 cm) Cattle lenient (8-10 cm) Sheep tight (3-5 cm) No4 Cattle lenient (8-10 cm) Cattle lenient (8-10 cm) Sheep lenient (6-8 cm) No5 Cattle tight (5-7 cm) Cattle lenient (8-10 cm) Sheep lenient (6-8 cm) Yes

b) Weed control sub-treatments

A No additional treatment, low thistle infestationB No additional treatment, high thistle infestationC Cutting June, cutting SeptemberD Herbicide wiping JuneE Cutting June, herbicide wiping SeptemberF Hay cutting (3 year rotation)

15

Table 2. Grazing management regimes (a) and weed control treatments (b) at Powys.

a) Grazing management main treatments

Spring Summer Autumn Winter1 Sheep tight (3-5 cm) Sheep lenient (6-8 cm) Sheep tight (3-5 cm) No2 Sheep tight (3-5 cm) Sheep lenient (6-8 cm) Sheep lenient (6-8 cm) No3 Sheep lenient (6-8 cm) Sheep lenient (6-8 cm) Sheep tight (3-5 cm) No4 Sheep lenient (6-8 cm) Sheep lenient (6-8 cm) Sheep lenient (6-8 cm) No5 Sheep tight (3-5 cm) Sheep lenient (6-8 cm) Sheep lenient (6-8 cm) Yes6 Cattle tight (5-7 cm) Sheep lenient (6-8 cm) Sheep tight (3-5 cm) No

b) Weed control sub-treatments

A No additional treatment, low thistle infestationB No additional treatment, high thistle infestationC Cutting June, cutting SeptemberD Herbicide wiping JuneE Cutting June, herbicide wiping September

16

Table 3. ANOVA F-values for grazing main treatments and nested thistle control sub-treatment effects on Cirsium arvense shoot numbers between 2000 and 2005 at (a) Bucks and (b) Powys.

a) Bucks

d.f. 2000 2001 2002 2003 2004 2005 Repeated measures1. Seasonal grazing intensity Across years Year interactions (d.f.)Spring tight vs lenient 1,6 1.3 2 0.3 0.6 7.6* 4.9* 3.1 0.6 (5,30)Autumn tight vs lenient 1,6 <0.1 <0.1 4.6 8.2* 7.1* 7.7* 12* 6.5*** (5,30)Weed control sub-treatment 5,40 28*** 21*** 23*** 9.5** 3.5* 1 12*** 8.5*** (25,200)2. Winter grazingAutumn vs autumn + winter grazing 1,2 0.2 13 1.4 520** - - 7.9 8.1* (3,6)Weed control sub-treatment 5,20 14*** 18*** 13*** 6.8*** - - 16*** 5*** (15,60)

b) Powys

d.f. 2000 2001 2002 2003 2004 2005 Repeated measures1. Seasonal grazing intensity Across years Year interactions (d.f.)Spring tight vs lenient 1,6 1.9 0.7 0.9 21** 10* 6.1* 9.9** 12*** (5,30)Autumn tight vs lenient 1,6 0.1 2.5 5.2 1.9 0.3 <0.1 2.1 2.1 (5,30)Weed control sub-treatment 4,32 24*** 25*** 27*** 5.9** 2.3 1 14*** 4.9** (20,160)2. Winter grazingAutumn vs autumn + winter grazing 1,2 <0.01 <0.1 <0.1 0.3 7.5 20* 1.3 2.2 (5,10)Weed control sub-treatment 4,16 9.9*** 14*** 12*** 4.6* 4.4* 1.8 8*** 4.5*** (20,80)3. Grazing animalCattle vs sheep 1,2 1.2 0.4 <0.1 29* 101** 20* 1.6 5.9** (5,10)Weed control sub-treatment 4,16 17*** 7.9*** 8.9*** 5** 2.7 2 5.6** 12*** (20,80)

* P<0.05, ** P<0.01, *** P<0.001, blank – no significant difference

17

Table 4. ANOVA F-values for grazing main treatments and nested thistle control sub-treatment effects on non-target forb species richness in 2000, 2001, 2002 and 2005 at (a) Bucks and (b) Powys.

a) Bucks

d.f. 2000 2001 2002 2005 Repeated measures1. Seasonal grazing intensity Across

yearsYear interactions (d.f.)

Spring tight vs lenient 1,6 0.5 0.3 3.2 1.7 0.1 1.5 (3,18)Autumn tight vs lenient 1,6 2.8 0.2 1.8 1.2 3.6 1,5 (3,18)Weed control sub-treatment 5,40 0.3 2.2 0.9 0.6 1.9 0.9 (15,120)

b) Powys

d.f. 2000 2001 2002 2005 Repeated measures1. Seasonal grazing intensity Across years Year interactions (d.f.)Spring tight vs lenient 1,6 2.4 1 2.6 6.6* 2.9 1.3 (3,18)Autumn tight vs lenient 1,6 0.3 0.1 <0.1 0.3 <0.1 1.2 (3,18)Weed control sub-treatment 4,32 2.5 1.9 1.6 5.7** 3.1* 2.1*2. Winter grazingAutumn vs autumn + winter grazing 1,2 37* 7.7 29* 49* 31* 2.7 (3,6)Weed control sub-treatment 4,16 0.9 6.7** 1.4 3.1* 3.8* 1.7 (12,48)3. Grazing animalCattle vs sheep 1,2 2.8 0.8 1.5 2.6 1.9 0.1 (3,6)Weed control sub-treatment 4,16 1.9 0.7 0.9 4.9** 2.1 1.8 (12,48)

* P<0.05, ** P<0.01, blank – no significant difference

18

0

2

4

6

8

10

12

14

16

18

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

2000 2001 2002 2003 2004 2005

This

tles

m-2

Tight

Lenient

0

2

4

6

8

10

12

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

Spr

ing

Aut

umn

2000 2001 2002 2003 2004 2005

This

tles

m-2

Tight

Lenient

Fig. 1. Effects of factorial combinations of tight vs lenient spring and autumn grazing on Cirsium arvense shoot numbers between 2000 and 2005 at (a) Bucks and (b) Powys. Grazing standard errors by taking blocks as samples.

a) Bucks

b) Powys

19

0

2

4

6

8

10

12

14

16

2000 2001 2002 2003 2004 2005

Year

This

tles

m-2

Spring T Spring L Autumn T Autumn L

0

2

4

6

8

10

12

14

2000 2001 2002 2003 2004 2005

Year

This

tles

m-2

Spring T Spring L Autumn T Autumn L

Fig. 2. Effects of spring and autumn grazing on Cirsium arvense shoot numbers in the light infestation control sub-treatments between 2000 and 2005 at (a) Bucks and (b) Powys.

a) Bucks

b) Powys

20

0

2

4

6

8

10

12

14

16

18

2000 2001 2002 2003

Winter grazing

This

tles

m-2

None Grazed

Fig. 3. Effects of winter grazing on Cirsium arvense shoot numbers at (a) Bucks between 2000 and 2003 and (b) Powys between 2000 and 2005. Grazing standard errors by taking blocks as samples.

a) Bucks

b) Powys

0

2

4

6

8

10

12

14

16

2000 2001 2002 2003 2004 2005

Winter grazing

This

tles

m-2

None Grazed

21

0

2

4

6

8

10

12

14

16

2000 2001 2002 2003 2004 2005

Grazing

This

tles

m-2

Sheep Cattle

Fig. 4. Effects of cattle vs sheep grazing on Cirsium arvense shoot numbers between 2000 and 2005 at Powys. Grazing standard errors by taking blocks as samples.

22

0

5

10

15

20

25

30

2000 2001 2002 2003 2004 2005

This

tles

m-2

LI HI Cut Wipe Cut & Wipe Hay

Fig. 5. Effects of sub-treatments on Cirsium arvense shoot numbers between 2000 and 2005 at (a) Bucks and (b) Powys. Data presented are means from Analysis 1 comparing grazing treatments 1-4 only.

a) Bucks

Weed control sub-treatment

2000 2001 2002 2003 2004 2005

A. Low Infestation C CD BC BC AB AB. High Infestation A A A A A AC. Cut AB BC CD C AB AD. Herbicide wipe C D D C B AE. Cut & herbicide wipe B CD D C AB AF. Hay cut B B AB AB AB A

Sub-treatments are statistically different where they have no letters in common (P<0.05).

23

0

2

4

6

8

10

12

14

16

18

20

2000 2001 2002 2003 2004 2005

This

tles

m-2

LI HI Cut Wipe Cut & Wipe

b) Powys

Weed control sub-treatment

2000 2001 2002 2003 2004 2005

A. Low Infestation BC B B ABC A AB. High Infestation A A A A A AC. Cut CD A A AB A AD. Herbicide wipe D C C C A AE. Cut & herbicide wipe B B B BC A A

Sub-treatments are statistically different where they have no letters in common (P<0.05).

24

Fig. 6. Additive effects of most effective main grazing treatments and weed control measures on Cirsium arvense shoot numbers at (a) Bucks and (b) Powys between 2000 and 2005.

a) Bucks

0

10

20

30

40

5013

/06/00

30/12

/0018

/07/01

03/02

/0222

/08/02

10/03

/0326

/09/03

13/04

/0430

/10/04

18/05

/0504

/12/05

This

tles

m-2

1 (Cattle T-L; Sheep T) B (Control high infestation)

1 (Cattle T-L; Sheep T) D (Herbicide)

4 (Cattle L-L; Sheep L) B (Control high infestation)

4 (Cattle L-L; Sheep L) D (Herbicide)

b) Powys

0

5

10

15

20

25

30

This

tles

m-2

1 (Sheep:T-L-T) B (Control high infestation)

1 (Sheep: T-L-T) D (Herbicide)

4 (Sheep: L-L-L) B (Control high infestation)

4 (Sheep: L-L-L) D (Herbicide)

25

0

2

4

6

8

10

12

14

16

18

Spring Autumn Spring Autumn Spring Autumn Spring Autumn

2000 2001 2002 2005

Forb

s m

-2

Tight

Lenient

Fig. 7. Effects of tight vs lenient spring and tight vs lenient autumn grazing by sheep on non-target forb species richness at Powys in 2000, 2001, 2002 and 2005. Grazing standard errors by taking blocks as samples.

26

0

2

4

6

8

10

12

14

16

18

2000 2001 2002 2005

Forb

spe

cies

m-2

LI HI Cut Wipe CW

Fig. 8. Effects of sub-treatments on non-target forb species richness at Powys in 2000, 2001, 2002 and 2005.

Grazing treatment1. Seasonal grazing

intensity2. Winter grazing 3. Grazing animal

Weed control sub-treatment

2000 2001 2002 2005 2000 2001 2002 2005 2000 2001 2002 2005

A. Low Infestation A A A A A A A A A AB A AB. High Infestation A A A AB A A A AB A BC A ABC. Cut A A A A A A A A A ABC A ABD. Herbicide wipe A A A B A A A B A C A BE. Cut & herbicide wipe A A A A A A A A A A A AB

Sub-treatments are statistically different where they have no letters in common (P<0.05).

27

References to published material9. 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.

28

29