Providing evidence to improve practice

Action: Control weeds without damaging other plants in conservation areas Farmland Conservation

Key messages


Supporting evidence from individual studies


A series of replicated experiments in the UK from 1969 to 1971 (Baker et al. 1972) tested the suitability of the non-native herbivorous beetle Haltica carduorum as a biocontrol agent for the injurious weed, creeping thistle Cirsium arvense, and found it unsuitable due to low survival in the UK climate.



A replicated, controlled laboratory study (Speight & Whittaker 1987) found that the systemic herbicide asulam had little influence on the survival or rate of development of the green dock beetle Gastrophysa viridula when used as a contact agent on eggs, first-instar larvae or adults. However, the ingestion of asulam-contaminated broad-leaved dock Rumex obtusifolius leaves reduced female fertility by 64% (a reduction in number of egg batches rather than in numbers of eggs per batch) and extended the time to reach the adult stage by 3-4 days. There was no effect on female longevity. Presence of asulam in dock leaves had no significant influence on egg-laying or feeding site selection. Fifteen replicates of each of the following treatments were sprayed on batches of 30 eggs, 20 first-instar larvae or one adult: high (5 ml/l) or low (1.25 ml/l) asulam concentration and a water control. In addition, 45 first-instar larvae were fed dock sprayed with one of the three treatment solutions. Female beetles that were carrying eggs were also put in 20 ‘choice chambers’ with leaves sprayed with each treatment; location, area of leaf consumed, position and numbers of egg batches were recorded after 6 hours.

The same authors (Speight & Whittaker 1987) also undertook a replicated, controlled study of asulam use in unmanaged grassland in the UK and found that broad-leaved dock Rumex obtusifolius survival was unaffected by beetle grazing or spraying with a low concentration of asulam (1.25 ml/l) and at high asulam concentrations (5 ml/l) only 40% of plants were killed. Numbers of first generation green dock beetle Gastrophysa viridula were similar on treated and untreated plants, but numbers of the second generation were significantly higher on untreated than treated plants. Shoot and root dry weights of asulam-treated plants were significantly lower than untreated ones. Beetle-grazing did not further reduce the dry weight of asulam-treated plants, but did those treated with water. Second generation beetles laid four times as many eggs on untreated plots and survival from eggs to larvae was 25% compared to just 4-12% on treated plots. Four blocks of six plots (4 m²) received the three treatments, each in the presence or absence of beetle-grazing. One-month-old docks were planted (16/plot) in April 1981. In May, 144 adult beetles were released in randomly allocated ‘grazing plots’ (12/plot); herbicide treatments were applied in June. Numbers of beetle eggs, larvae and adults were counted on four plants on each grazed plot from May-October. Dock plant material was harvested in August and October 1981 and February and April 1982 and dry weights and leaf areas (untreated plots) obtained.



A replicated, controlled, randomized study of a wildflower margin at a farm in Oxfordshire, UK (Marshall & Nowakowski 1996) found that applying grass-specific herbicide did not affect overall plant species diversity, however grass diversity was reduced and broadleaved plants increased in sprayed plots. Applying a grass-specific herbicide in December eliminated black grass Alopecurus myosuroides, plots not treated with herbicide were dominated by black grass. Sown crested dogs-tail Cynosurus cristatus was eliminated by a second treatment of herbicide in April; late mowing in June also decreased this species. The wildflower/grass seed was sown on 21 contiguous margin plots (3 x 12 m). Plots were grouped into three blocks, within which they randomly received one of seven treatments: unmanaged, cut April, cut April and May, cut May, cut in May and June, cut in June or grass-specific herbicide (fluazifop-P-butyl) application in April. Cuttings were removed. Half of each plot received grass-specific herbicide in December. Plant composition of sub-plots was sampled in five 0.1 m² quadrats in July 1995.



A small, replicated study from 1992 to 1996 of four pasture plots in North Yorkshire, UK (Newborn 2000) found that Hebridean sheep grazed more purple moor grass Molinia caerulea than Swaledale sheep, but the resulting density of purple moor grass and heather Calluna vulgaris did not differ. Hebridean sheep grazed significantly more purple moor grass leaves than Swaledales (61% vs 23%). However, the density of purple moor grass leaves did not differ between Hebridean (2389 m²) and Swaledale plots (2798 m²). Overall, cover by heather did not change over time: in the Hebridean plots, heather cover doubled in the first four years (12% to 29%), then declined (22%), the overall increase was 3.7%/year. In Swaledale plots, cover increased over the first two years (3% to 7%), then declined dramatically (to 1%), with an overall decline of 1%. Two areas of pasture dominated by purple moor grass, ungrazed for two years and unburnt for ten, were divided into two plots of 2 ha, one grazed by Swaledales, one by Hebridean sheep (99 kg live weight/ha) between May and September. Numbers of grazed and ungrazed purple moor grass leaves were sampled within two quadrats (0.5 x 0.5 m) and lengths of grazed/ungrazed leaves measured at 100 points across plots following sheep removal in September 1992-1996. Plant species were sampled using a point sampling technique along a 15 m transect (45 cm intervals) using an eight pinned frame with the tallest vegetation touching each pin identified.


A replicated, controlled study in 1998-1999 on fallow land at the University of Göttingen, Germany (Kluth et al. 2003) found that cutting reduced creeping thistle Cirsium arvense reproductive success, but combining cutting with infection with rust fungus Puccinia punctiformis further restricted sexual reproduction and was therefore a more effective control strategy. Potted thistle plants (15 cm tall) were transferred outdoors and one of four treatments (10 replicates of each) were applied in June 1998 and 1999: cutting at 30 cm, application of a spore suspension of the rust fungus, cutting and rust application and controls. Thistle size, number of flower buds and fertile flower heads and above and below ground (1999) dry matter were measured.



A replicated, controlled study in 1998-1999 in Germany (Guske et al. 2004) found that pathogens reduced the reproductive capacity of potted creeping thistle Cirsium arvense. This was particularly the case with the fungus Phoma hedericola and a combination of P. hedericola with Mycelia sterile, Phoma destructiva and Phoma nebulosa. Infections with pathogens, both singly and in combination, tended to increase disease severity compared to control plants; this was not the case for P. punctiformis or combined P. punctiformis and P. hedericola in the first year. Experiments were conducted under semi-field conditions at Braunschweig. Target plants (7-20) were inoculated with fungal isolates: Phoma hedericola, Phoma destructiva, Mycelia sterila and Puccinia punctiformis (Phoma nebulosa in 1999) applied individually, or a combination of P. punctiformis and P. hedericola (in 1998) or combination of all four (1999). Twenty plants were controls. Plants were evaluated each week (May-August) for disease symptoms (tissue breakdown, tissue death, pale/yellow leaves) and plant development.


A replicated, randomized, controlled study in 1998-2000 on fallow land at the University of Göttingen, Germany (Kluth et al. 2005) found that fungal pathogens resulted in a decline in cover of creeping thistle Cirsium arvense. Creeping thistle cover decreased significantly (60 to 5%), with an associated increase in co-occurring species in the experimental area over three years (1998-2000) following inoculations with the fungal pathogens Puccinia punctiformis and Phoma destructiva. There was no significant difference between disease incidence of P. punctiformis on creeping thistle in plots following single and triple inoculations, or in control plots (99%). Combined treatment with P. punctiformis and P. destructiva increased the disease incidence with P. destructiva compared to control, fungicide and P. punctiformis treatments in the third year (2-7%).  Ten treatments (six replicates) and a control (12 replicates) were randomly assigned to 72 plots. P. punctiformis and P. destructiva were applied once in June, July or August or in all three months/year (1998-2000), or both were applied together in June. These were compared with a fungicide treatment (Opus Top) twice/year and untreated controls. Plots were monitored monthly (May-September) to determine the percentage of creeping thistle infected, disease severity (P. punctiformis) and the percentage cover of creeping thistle and other species.



A replicated, controlled study in an arable field in Denmark (Graglia et al. 2006) found that increased mowing and hoeing frequency tended to reduce the amount of above ground creeping thistle Cirsium arvense biomass in the subsequent year (up to 73% compared to control). The presence of a suppressive crop (grass/white clover Trifolium repens mixture or red clover T. pratense) tended to further reduce creeping thistle. Differences in barley yield were only explained by the amount of creeping thistle biomass in one experiment, where the weed was most abundant. Four adjacent sub-fields, divided into four blocks were subject to a combination of cropping (grass/clover in half of the plots, spring barley with grass/clover undersown in half of the plots, spring barley with red clover undersown in half of the plots) and mechanical (0, 2, 4, 6 passes with a mower; 0, 3, 6 or 0, 1, 3, 5 passes with a hoe) treatments over a 3-year period (between 2000-2004). Heights and dry weights of above ground shoots of creeping thistle were obtained before harvest in the third year, as was barley yield.



At Ranscombe Farm, a nature reserve managed for arable plants in the north Kent Downs, UK (Still 2007), a four hectare field sprayed with glyphosate in September 2005 produced new populations of the rare arable plants blue pimpernel Anagallis arvensis subspecies foemina and ground pine Ajuga chamaepitys the following year. The field had been managed for some years with shallow cultivation in autumn only, and had increasing abundances of perennial weeds such as docks Rumex spp. and couch grass Elytrigia repens. The entire field was sprayed with glyphosate in September 2005, then deep ploughed or ‘disced’ in February 2006.



A 2008 review of control methods for competitive weeds in uncropped cultivated margins managed to maintain uncommon arable plant populations in the UK (Critchley & Cook 2008) found that specific management regimes can reduce abundance of pernicious weeds in margins. Abundance of pernicious weeds tended to increase if uncropped cultivated margins were not cultivated annually in two studies (Critchley 1996b, 2000a). However five studies found weeds also build up on margins cultivated annually, particularly with the same annual cultivation regime (Critchley 1996a,b, Critchley et al. 2004, Critchley et al. 2006, Still & Byfield 2007). Eight studies found that abundance of specific weed species depended on timing or method of cultivation (Marshall 1998, Critchley 2000b, Ford 2000, Critchley et al. 2004, Critchley et al. 2005, Critchley et al. 2006, Corsie 2007, Still & Byfield 2007). Rotating margin sites was found to reduce weed abundance in two studies (Davies et al. 1994, Wilson 2000) and four studies found specific timing, frequency and height of cutting decreased certain species (Marshall 1998, Carvell et al. 2004, Corsie 2007, Westbury et al. 2008). Twelve studies reported particular weed species could be targeted with specific timing of herbicide applications in different margin types (Boatman 1991, Varney et al. 1995, Wilson 1995, Marshall 1998, Boatman et al. 1999, Ford 2000, Marshall 2002, Boatman 2007, Corsie 2007, Meek et al. 2007, (Still 2007), Still & Byfield 2007). However, rare arable plant species can also be susceptible to specific management regimes (Wilson et al. 1990, Wilson 1995, Wilson & King 2003).


Additional references:

Wilson P.J., Boatman N.D. & Edwards P.J. (1990) Strategies for the conservation of endangered arable weeds in Great Britain. Proceedings of the EWRS Symposium 1990 - Integrated Weed Management in Cereals. European Weed Research Society, Wageningen, pp. 93-101.

Boatman N.D. (1991) Selective control of cleavers (Galium aparine) in conservation headlands with quinmerac. Brighton Crop Protection Conference - Weeds - 1991, 669-676.

Davies D.H.K. & Carnegie H.M. (1994) Vegetation patterns and changes in field boundaries and conservation headlands in Scottish arable fields. Pages 173-178 in: N. Boatman (ed.) Field Margins: Integrating Agriculture and Conservation, BCPC Monograph 58. British Crop Protection Council, Farnham.

Varney P.L., Scott T.A.J., Cooke J.S. & Ryan P.J. (1995) Clodinafop-propargyl – a useful tool for management of conservation headlands. Brighton Crop Protection Conference – Weeds - 1995, 967-972.

Wilson P.J. (1995) The potential for herbicide use in the conservation of Britain's arable flora. Brighton Crop Protection Conference - Weeds - 1995, 961-966.

Critchley C.N.R. (1996a) Monitoring as a feedback mechanism for the conservation management of arable plant communities. Aspects of Applied Biology, 44, 239-244.

Critchley C.N.R. (1996b) Vegetation of arable field margins in Breckland. PhD thesis, University of East Anglia.

Marshall E.J.P. (1998) Guidelines for the siting, establishment and management of arable field margins, beetle banks, cereal conservation headlands and wildlife seed mixtures. IACR report to MAFF.

Boatman N.D., Bence S. & Jarvis P. (1999) Management and costs of conservation headlands on heavy soil. Pages 147-154 in: N. D. Boatman, D. H. K. Davies, K. Chaney, R. Feber, G. R. de Snoo, & T. H. Sparks (eds.) Field margins and buffer zones: Ecology, Management and Policy, Aspects of Applied Biology 54.

Critchley C.N.R. (2000a) Ecological assessment of plant communities by reference to species traits and habitat preferences. Biodiversity and Conservation, 9, 87-105.

Critchley C.N.R. (2000b) The conservation ecology of arable plants: what role for research? Pages 80-87 in: P. Wilson & M. King (eds.) Fields of Vision. A Future for Britain’s Arable Plants. RSPB, Sandy.

Ford S. (2000) Can arable fields be managed specifically for arable plant communities? Pages 57-60 in: P. Wilson & M. King (eds.) Fields of Vision. A Future for Britain’s Arable Plants. RSPB, Sandy.

Wilson P.J. (2000) Management for the conservation of arable plant communities. Pages 38-47 in: P. Wilson & M. King (eds.) Fields of Vision. A Future for Britain’s Arable Plants. RSPB, Sandy.

Marshall E.J.P. (2002) Weeds and Biodiversity. Pages 75-92 in: R.E.L. Naylor (ed.) Weed Management Handbook (ninth edition). Blackwell Publishing.

Wilson P.J. & King M. (2003). Arable Plants - A Field Guide. English Nature/Wildguides Ltd.

Carvell C., Meek W.R., Pywell R.F., & Nowakowski M. (2004) The response of foraging bumblebees to successional change in newly created arable field margins. Biological Conservation, 118, 327-339.

Critchley C.N.R., Fowbert J.A. & Sherwood A.J. (2004) Botanical assessment of the Arable Stewardship Pilot Scheme, 2003. ADAS report to Defra.

Critchley C.N.R., Fowbert J.A. & Sherwood A.J. (2005) Re-assessment of uncropped wildlife strips in Breckland Environmentally Sensitive Area. ADAS report to Defra.

Critchley C.N.R., Fowbert J.A. & Sherwood A.J. (2006) The effects of annual cultivation on plant community composition of uncropped arable field boundary strips. Agriculture, Ecosystems and Environment, 113, 196-205.

Boatman N.B. (2007) Potential effects of Environmental Stewardship on arable weeds: uptake of options relevant to conservation of the arable flora and weed control issues. 44th BCPC Annual Weed Review, 2007. BCPC, Reading.

Corsie C. (2007) Worcestershire Wildlife Trust Important Arable Areas Project. Report to Plantlife International. Worcestershire Wildlife Trust, Worcester.

Meek W.R., Pywell R.F., Nowakowski M. & Sparks T.H. (2007) Arable field margin management techniques to enhance biodiversity and control barren brome, Anisantha sterilis. Pages 133-141 in: C. Britt, A. Cherrill, M. le Duc, R. Marrs, R. Pywell, T. Sparks, I. Willoughby (eds.) Vegetation Management, Aspects of Applied Biology 82.

Still K. & Byfield A. (2007) New priorities for arable plant conservation. Plantlife, Salisbury.

Westbury D.B., Woodcock B.A., Harris S.J., Brown V.K. & Potts S.G. (2008) The effects of seed mix and management on the abundance of desirable and pernicious unsown species in arable buffer strip communities. Weed Research, 48, 113-123.


A replicated, controlled, randomized study in 2004-2005 in a former agricultural field near Bern, Germany (Wandeler et al. 2008) found that creeping thistle Cirsium arvense could be infected with a systemic rust fungus Puccinia punctiformis using the weevil Ceratapion onopordi as a disease carrier. There was a significantly higher rust incidence within 1 m of weevil-treated thistle shoots (34 shoots infected) compared to controls (1 infected). Overall, within a radius of 1 m, 27% of weevil-treated shoots had rust infections compared to 3% of control shoots. There was no significant effect of the treatment within radii of 0.3 m or above 1 m. Therefore, rust infections could be induced between 0.3-1 m from weevil-treated thistles. The field had been sown with a mixture of wildflower seeds, grass and clover Trifolium spp.. In April 2004, 60 thistle shoots (? 1 m apart) in the wildflower strip were randomly assigned as either infected (with one female weevil powdered with rust spores (1000 spores/female)), or controls. Weevils were confined to shoots for 72 hours using a cylinder sealed at the top; controls received only the cylinder. Systemically infected thistles were located and assigned to the nearest experimental shoot, within radii of 0.3, 1, 2 or 3 m in April-July 2005.


A replicated, randomized, controlled study in 2000-2005 on the effectiveness of control strategies on creeping thistle Cirsium arvense numbers at two pastoral farms in England and Wales (Pywell et al. 2010) found that lenient grazing was most effective for long-term control. In a lowland cattle and sheep system and an upland sheep system, thistle numbers decreased under lenient grazing (cattle grazed to 8-10 cm, sheep 6-8 cm) whereas they remained constant or increased under heavy grazing regimes (cattle grazed to 5-7 cm, sheep 3-5 cm). Herbicide wiping gave the most rapid and effective control and cutting was one of the least effective measures, however, the effects of all weed control sub-treatments were short-lived at both sites. Six grazing treatments (a combination of lenient and heavy grazing) were applied to plots in a randomized block design with three replicates. Five weed control sub-treatments (thistle cutting, herbicide wiping, cutting and wiping, controls) were undertaken within grazing treatments using a split-plot design with replication at the block level; hay cutting on a three year rotation was also undertaken at one site. Data on thistle shoot density and effects on non-target broadleaved plants (rooted frequency) were obtained within each sub-treatment plot (2000-2005). This study is also described in a 2004 Defra report (Pywell et al. 2004).


Additional references:

Pywell R., Tallowin J. & Masters G. (2004) Effects of grazing management on creeping thistle and other injurious weeds and integration of grazing with weed control. Defra report BD1437.

Referenced papers

Please cite as:

Dicks, L.V., Ashpole, J.E., Dänhardt, J., James, K., Jönsson, A., Randall, N., Showler, D.A., Smith, R.K., Turpie, S., Williams D.R. & Sutherland, W.J. (2017) Farmland Conservation Pages 245-284 in: W.J. Sutherland, L.V. Dicks, N. Ockendon & R.K. Smith (eds) What Works in Conservation 2017. Open Book Publishers, Cambridge, UK.