Providing evidence to improve practice

Action: Create beetle banks

Key messages

Natural enemies in fields: Six studies from Canada, the UK and USA (three replicated, controlled, of which two were also randomised) examined the effects on predator numbers in adjacent crops. A review found that predators increased in adjacent crops, but one study found effects varied with time and another found no effect. Two studies found small or slow movements of predators from banks to crops. One study found greater beetle activity in fields but this did not improve pest predation.
Natural enemies on banks: Four studies and a review found more invertebrate predators on beetle banks than in surrounding crops, but one of these found that effects varied with time. Eight studies from the UK and USA (including two randomised, replicated, controlled trials and two reviews) compared numbers of predatory invertebrates on beetle banks with other refuge habitats. Two studies found more natural enemies on beetle banks, but one of these found only seasonal effects. One review found similar or higher numbers of predators on beetle banks and four studies found similar or lower numbers.
Pests: A replicated, randomised study and a review found the largest pest reductions in areas closest to a beetle bank or on the beetle bank itself. One review found fewer pests in fields with than without a beetle bank.
Economics: One replicated, randomised, controlled trial and a review showed that beetle banks could make economic savings if they prevented pests from reaching a spray threshold or causing 5% yield loss.
Beetle bank design: Two studies from the UK found certain grass species held higher numbers of predatory invertebrates than others.

Crops studied were barley, field bean, maize, oats, pea, radish, rapeseed, soybean, wheat and pasture.

Supporting evidence from individual studies

1 

A replicated, randomised study in spring 1988-1990 on a beetle bank in a 7 ha winter wheat Triticum sp. field in Hampshire, UK (Thomas 1991) found that over the 1988 survey period, predatory invertebrate activity shifted from the beetle bank into the crop, although the effect was small. In 1989, the ground beetle Demetrias atricapillus was initially more abundant 0-3 m from the beetle bank (average 12.2 individuals/m², 14 April-3 May) but became more evenly distributed with an average 0.4 individuals/m² at 0-60 m from the bank (8-22 May). The rove beetle Tachyporus hypnorum did not show a consistent distribution in 1989-1990, although fewer individuals were found on the bank than the crop by the end of the 1989 survey. More money spiders (Linyphiidae) were found in the beetle bank than the crop in 1989 and in 1990 there was a slight emigration of money spiders away from the bank into the crop. In 1990 wolf spiders (Lycosidae) were found throughout the crop but were most abundant next to the beetle bank. The beetle bank (290 m long, 0.4 m high and 1.5 m wide) was created in autumn 1986 and sown with grasses. This study was part of the same experimental set-up as Thomas et al. 1991, Thomas et al. 1992, MacLeod 1994 and MacLeod et al. 2004.

2 

A randomised, replicated, controlled study in winter 1987-1988 and 1988-1989 on two beetle banks in two cereal fields on a farm in Hampshire, UK (Thomas et al. 1991) found total invertebrate predator numbers collected from turf samples and ground searching were higher on beetle banks (218-1,488 individuals/m² in turf samples, 39-188 individuals/m² in surface searches) than the surrounding crop (26-29 individuals/m² in turf samples, 16-49 individuals/m² in surface searches). Invertebrate predators included ground beetles (Carabidae), rove beetles (Staphylinidae) and spiders (Araneae). In 1989, emigration patterns of the rove beetle Tachyporus hypnorum and the ground beetle Demetrias atricapillus showed movement of individuals from the bank into the field from 14 April-22 May. From 14 April-3 May, there were 12.2 individuals/m² of D. atricapillus at 0-3 m from the bank, after which the average density was 0.4 individuals/m² at 0-60 m from the bank. By the end of the study there were significantly fewer T. hypnorum on the bank than the crop. Establishment costs were estimated at £85 in year one and £30 in following years for a 20 ha field (1990 prices). Maintaining aphid (Aphidoidea) populations below a spray threshold was valued at £300/year and £660/year if an aphid-induced yield loss of 5% was prevented. This study was part of the same experimental set-up as Thomas 1991, Thomas et al. 1992, MacLeod 1994 and MacLeod et al. 2004.

3 

A randomised, replicated, controlled study over three winters from 1987-1990 on two farms in Hampshire, UK (Thomas et al. 1992) (part of the same study as Thomas et al. 1991 but extended to a third winter and a third beetle bank in a 51 ha field on a second farm) found that three years after beetle bank establishment, total predator densities on beetle banks (358-764 individuals/m²) were not different to those in natural field boundaries (541-569 individuals/m²). Ground beetle and spider community composition was similar between beetle banks and field boundaries. Cock’s-foot Dactylis glomerata, a tussock-forming grass, supported highest densities of ground beetles in the third winter. Community composition of ground beetles and spiders changed during the study to species that prefer boundary or more permanent habitats. Banks were 0.4 m high x 1.5 m wide. Two were 290 m long in 7 and 20 ha fields, one was 580 m long in a 51 ha field. One field was sown with winter wheat Triticum spp. throughout the study, one field had winter wheat then fodder pea Pisum sativum and winter rape Brassica napus, and one field had spring barley Hordeum vulgare then vining peas. This study was part of the same experimental set-up as Thomas 1991, Thomas et al. 1991, MacLeod 1994 and MacLeod et al. 2004.

4 

A replicated study over seven winters from late 1987 to early 1994 on one beetle bank in Hampshire and one in Essex, UK (MacLeod 1994) found sections sown with the grasses cock’s-foot Dactylis glomerata or Yorkshire fog Holcus lanatus generally had highest densities of predatory invertebrates, but not always significantly so. Ground beetles (Carabidae) and rove beetles (Staphylinidae) had higher densities in cock’s-foot (11-110 ground beetles/m², 1-125 rove beetles) and Yorkshire fog (1-76 ground beetles/m², 2-113 rove beetles) than two other grass species (2-15 ground beetles/m², 0-79 rove beetles). Ground beetle and rove beetle densities peaked in the second and sixth winters after banks were established. The pattern was the same for spiders (Araneae) in cock’s-foot but in Yorkshire fog, creeping bent Agrostis stolonifera and perennial ryegrass Lolium perenne the densities steadily increased to a maximum in the fifth winter. The 200 m long beetle bank in Essex had a lower density of ground beetles than a nearby hedge bottom (0.7 individuals/m² vs. 2.6 individuals). The 290 m long Hampshire beetle bank was created in spring 1987 and split into six blocks, each further sub-divided into eight plots with one sown grass treatment/plot. This study was part of the same experimental set-up as Thomas 1991, Thomas et al. 1991, Thomas et al. 1992 and MacLeod et al. 2004.

5 

A replicated study in the winters of 1993-1996 in Leicestershire, UK (Collins et al. 1996) found a beetle bank had lower densities of invertebrate predators (total of all groups combined), ground beetles (Carabidae) and rove beetles (Staphylinidae) than a nearby hedge across the study period. Total predator, ground beetle and rove beetle densities increased with age of beetle bank and by the third winter there were similar total predator and ground beetle densities between the hedge and beetle bank. Spider (Araneae) densities were similar between habitats. Total predator, ground beetle and rove beetle densities on beetle banks were highest in false oat grass Arrhenatherum elatius, cock’s-foot Dactylis glomerata and timothy Phleum pratense. Densities were lowest in crested dog’s-tail Cynosurus cristatus. In the first test, one 400 m-long beetle bank sown with cock’s-foot and Yorkshire fog Holcus lanatus (2.5 m wide, 0.5 m high) in an 18 ha field was compared with a 400 m-long hedge on the field edge (both habitats divided into 100 m blocks). In the second test, two 360 m-long beetle banks in an 8.6 ha field were divided into twenty 18 m-long blocks, sown with one of nine different grass treatments or left to naturally regenerate. In both tests invertebrates were collected from 11.5 cm diameter soil samples (3-10 samples/block). This study was part of the same experimental set-up as Moreby & Southway 2002 and Collins et al. 2003.

6 

A replicated, randomised, controlled study in 1996-1997 in one field at the Michigan State University Entomology Farm, Michigan, USA (Carmona & Landis 1999) found that raised refuge strips did not affect the activity density of ground beetles (Carabidae) in surrounding cropped subplots in both years (numbers not provided). However raised refuge strips had seasonally higher ground beetle activity densities from May-August 1996 (average 4-16 beetles/trap) and May, July and August 1997 (2-6 beetles/trap) than surrounding crops (0.5-8 beetles/trap in 1996; 1-5 in 1997). There were eight 30 x 30 m plots, each divided into two 30 x 15 m subplots. Four pairs of subplots were separated by a 3.3 m-wide, 0.10 m-high refuge strip and four pairs had no refuge strip. The field was in a three-year crop rotation of soybean Glycine max, oats Avena sativa and maize Zea mays. The central 0.3 m section of refuge strips was planted with three perennial flowering plant species and a grass-legume seed mix was sown on either side of the flowering plants. Ground beetles were sampled in May-October in three pitfall traps/refuge strip or control area and six traps/subplot in the surrounding crop area.

 

7 

A 2000 literature review (Dent 2000) found two studies from the UK and USA showing natural enemy populations were larger in beetle banks than the surrounding crop (Rodenhouse et al. 1992) or other field margin habitats (Thomas et al. 1991, the same study as above). There were fewer potato leafhoppers Empoasca fabae in fields with grass corridors (Rodenhouse et al. 1992). One study from 1988 (and updated in 1994) calculated that establishing a beetle bank in a 20 ha field could save £660 (US$1,090) if an aphid-induced yield loss of 5% was prevented and £300/year (US$495) in pesticide and labour costs if natural enemy populations kept aphid (Aphididae) numbers below a spray threshold (Wratten 1988, Wratten & van Emden 1995). Economic costs of establishing a beetle bank in a 20 ha field were approximately £85 (US$140) in year one based on: labour cost (1-2 days), yield loss from land taken out of production (assuming an average yield of 6 t/ha at £110/t, or US$180/t) and cost of grass seed (£5 or US$8). Gross yield lost in subsequent years because of the beetle bank taking up production land was calculated at £30 (US$50).

Additional references:

Rodenhouse N.L., Barrett G.W., Zimmerman D.M. & Kemp J.C. (1992) Effects of uncultivated corridors on arthropod abundances and crop yields in soybean agroecosystems. Agriculture, Ecosystems and Environment, 38, 179-191

Wratten S.D. (1988) The role of field boundaries as reservoirs of beneficial insects. Pages 144-150 in: A.J. Burn, T.H. Coaker & P.C. Jepson (eds.) Environmental Management in Agriculture: European Perspectives. EEC/Pinter Publishers, London.

Wratten S.D. & van Emden H.F. (1995) Habitat management for enhanced activity of natural enemies of pests. Pages 117-145 in: D.M. Glen, M.P. Greaves & H.M. Anderson (eds.) Proceedings of the 13th Long Ashton Symposium, England: Ecology and Integrated Farming Systems. John Wiley & Sons, London.

8 

A literature review in 2000 (Holland & Luff 2000) found three studies that showed higher numbers of ground beetles (Carabidae) in fields adjacent to beetle banks (Thomas et al. 1991, Riedel 1992, Collins et al. 1996).

Additional reference:

Riedel, W. (1992) Hibernation and spring dispersal of polyphagous predators in arable land. PhD thesis, Aarhus University.

9 

A paired, replicated, controlled study in winters 1997-1998 and 1998-1999 and summer 1999 on five farms in the UK (Thomas et al. 2000) found fewer rove beetles (Staphylinidae) on beetle banks (approximately 320-480 individuals/m²) than in field margins (560-680 individuals) in both winters, however ground beetle (Carabidae) and spider (Araneae) numbers were similar between beetle banks (200-240 ground beetles/m² and 360-440 spiders/m²) and field margins (200-280 ground beetles/m² and 400-500 spiders/m²). Ground beetle and spider diversity was slightly higher in beetle banks than field margins and rove beetle diversity was higher in field margins. Of the other invertebrates sampled (not specifically listed as natural enemies or pests), soldier beetles (Cantharidae), typical bugs (Heteroptera), other Auchenorrhyncha (excluding leafhoppers (Cicadellidae), planthoppers (Delphacidae) and bugs (Hemiptera)), other spiders, small flies (Diptera) and ants (Formicidae) were significantly more abundant on field margins than beetle banks. Total invertebrate abundance was also higher on field margins than beetle banks (averaging 64.7 vs. 46.7 invertebrates/sweep net). Predatory invertebrates were sampled on five beetle banks in winter 1997-1998 and 1998-1999. Other invertebrates were sampled on 22 beetle banks on five farms in summer 1999. Banks were paired with a neighbouring field margin. This study was part of the same experimental set-up as Thomas 2001 and Thomas 2002.

10 

A replicated, randomised study in 1996 in a winter wheat Triticum sp. field in Leicestershire, UK (Collins et al. 2002) found significantly more cereal aphids Sitobion avenae 83 m away from a beetle bank than 8 m away during the peak infestation period. Aphid numbers were 34% higher at the peak infestation period when predators, including ground beetles (Carabidae), rove beetles (Staphylinidae), money spiders (Linyphiidae) and wolf spiders (Lycosidae), were excluded. Ground beetle species typical of open field habitats were most abundant near the beetle bank before the peak aphid infestation period, while species typical of boundary habitats were most abundant near the beetle bank in April, showing a slow movement from the bank into the crop. The 400 m long (2.5 m wide x 0.5 m high) beetle bank was established in 1992 in 7.48 ha of an 18.3 ha field. The field was divided into four 100 m blocks containing a control area, and a predator-exclusion area (both 7 x 8 m) at 8, 33, 58 and 83 m from the bank. Aphids were counted twice a week on 20 labelled wheat tillers in each area (2 July-16 August) and on 10 ears of wheat each week (25 July-19 August). Arthropod predators were also counted on the 10 ears of wheat, and in three pitfall traps/area once a week from April-July.

11 

A replicated study in 1998 at two arable sites in Hampshire, UK (Thomas 2001) found that numbers of ground beetles (Carabidae) known to overwinter in boundary habitats were highest near beetle banks and declined further into the crop field in March. Beetles were more evenly spread across the field in the following months until June, when they were again clustered near the beetle bank. Ground beetles known to overwinter in the field were patchily distributed and concentrated towards the centre of the field. Beetle banks were studied in barley Hordeum vulgare and wheat Triticum aestivum fields. The study used 10 transects from the beetle bank into the field, with pitfall traps at 5, 25, 50, 75, 100 and 150 m from the beetle bank edge. The first site had 20 transects in two fields either side of a single beetle bank and the second had 10 transects from a bank positioned along the edge of a single field. Ground beetles were categorised depending on whether they spend the winter in field boundaries or in the fields themselves. This study was part of the same experimental set-up as Thomas et al. 2000 and Thomas 2002.

12 

A 2002 review (Evans et al. 2002) of two reports (Wilson et al. 2000, ADAS 2001) evaluating the effects of the Pilot Arable Stewardship Scheme in two regions of the UK (East Anglia and the West Midlands) from 1998-2003 found that grass margin options (including beetle banks) benefitted bugs (Hemiptera) and sawflies (Symphyta) but not ground beetles (Carabidae). The review does not specify whether bugs and sawflies were natural enemies or pests. The grass margin set of options included sown grass margins, naturally regenerated margins, beetle banks and uncropped cultivated wildlife strips. The review does not distinguish between these. None of the beneficial effects were pronounced on beetle banks. The effects of the pilot scheme on invertebrates were monitored relative to control areas over three years. Grass margins were implemented on total areas of 361 and 294 ha in East Anglia and West Midlands respectively.

Additional references:

ADAS (2001) Ecological evaluation of the Arable Stewardship Pilot Scheme, 1998-2000. ADAS report.

Wilson S., Baylis M., Sherrott A. & Howe, G. (2000) Arable Stewardship Project Officer Review. F. a. R. C. Agency report.

13 

A study in 1995-1999 in arable land in Leicestershire, UK (Moreby & Southway 2002) found that spiders (Araneae) and some groups of bugs (Homoptera) were consistently more abundant in uncropped strips than in four crop types or in grazed pasture. Other bug groups (Heteroptera) were most abundant in uncropped strips in four out of five years. Abundance of other groups in different crop types varied between years. The experiment sampled insects from six habitats: wheat Triticum aestivum, barley Hordeum vulgare, oilseed rape Brassica napus and field bean Vicia faba crop fields, grazed pasture fields and uncropped strips. The uncropped strips included both beetle banks and strips sown with wild bird cover mix, and the study did not differentiate results from these two habitats. Insect sampling used a "D-Vac" suction sampler. The study did not indicate whether insect groups were pests, natural enemies or neutral. This study was part of the same experimental set-up as Collins et al. 1996 and Collins et al. 2003.

14 

A paired, replicated, controlled study on five arable estates in Hampshire and Wiltshire, UK (Thomas 2002) found that ground beetle (Carabidae) density and species diversity were higher on beetle banks than field margins in summer but not winter. In spring and summer, ground beetle density and species diversity were higher in beetle banks (averaging 75 individuals/m² in spring, 90 individuals/m² in summer) than field margins (45 and 60 individuals, respectively). In winter there was no difference in ground beetle density (approximately 200-300 individuals/m²), species richness (15-22 species) or diversity between beetle banks and field margins, but species richness increased with age in beetle banks. In summer, beetle banks had higher average cover of grass weeds but grass and broad-leaved weed cover was highly variable in both habitats. Ground beetles were surveyed on five beetle banks on one estate in January-February, May, August and February the following year. Vegetation was surveyed on 22 beetle banks (including those surveyed for beetles) on five estates in January-February (nine banks) and July (22 banks). Banks were 1-13 years old. Each bank was paired with a conventional permanent margin in the adjacent field. This study was part of the same experimental set-up as Thomas et al. 2000 and Thomas 2001.

15 

A replicated study in 1994-1998 assessing two beetle banks in arable land in Leicestershire, UK (Collins et al. 2003) found higher invertebrate predator densities in false oat grass Arrhenatherum elatius (2,045 individuals/m²) than in red fescue Festuca rubra (1,492 individuals), crested dog’s-tail Cynosurus cristatus (1,380 individuals) and naturally regenerated vegetation (1,060 individuals). Rove beetles (Staphylinidae), were the dominant predator family, and showed the same significant pattern (1,716 individuals/m² in false oat grass through to 834 individuals in naturally regenerated vegetation). Spider (Araneae) density was higher in cock’s-foot (177 individuals/m²) compared with red fescue (119 individuals) and naturally regenerated vegetation (107 individuals). Ground beetle (Carabidae) density was 2.5-3.5 times higher in cock’s-foot than all other treatments. Boundary-type ground beetles dominated all treatments but were also more abundant in cock’s-foot (328 individuals/m²) compared with the other five treatments (69-126 individuals). Beetle banks created in spring 1993 were situated in an 8.6 ha clay soil field. Six treatments (five grass species and naturally regenerated vegetation) were established with two replicates/bank. Invertebrates were collected from soil samples gathered in January-February 1994-1997. Vegetation was examined visually and measured with a graduated board. This study was part of the same experimental set-up as Collins et al. 1996 and Moreby & Southway 2002.

16 

A randomised, replicated study of a beetle bank over seven winters from early 1987 to early 1994 on a mixed arable estate in Hampshire, UK (MacLeod et al. 2004) found that ground beetle (Carabidae) and rove beetle (Staphylinidae) densities were often highest in blocks sown with cock’s-foot Dactylis glomerata or Yorkshire fog Holcus lanatus (0.6-110.4 ground beetles/m², 1.2-125.4 rove beetles), although numbers were not always significantly higher than in creeping bent Agrostis stolonifera (3.1-15.4 ground beetles, 0.3-66.7 rove beetles) or perennial ryegrass Lolium perenne (2.1-11.5 ground beetles, 2.1-78.8 rove beetles). Densities of money spiders (Linyphiidae) and wolf spiders (Lycosidae) were also higher in cock’s-foot and Yorkshire fog, although not always significantly. Ground beetle species composition changed from species typical of open fields to species of field boundaries over the study period. Field boundaries were sampled in the last three winters and had lower densities of predatory invertebrates than the beetle bank, but this was not tested statistically. One 290 m-long beetle bank was divided into six blocks into which eight sowing treatments/block were applied (this study examined only four single-species grass treatments). Predator communities were sampled through ground zone searching and destructive sampling November-February. This study was part of the same experimental set-up as Thomas 1991, Thomas et al. 1991, Thomas et al. 1992 and MacLeod 1994.

17 

A replicated, controlled trial study in 2003-2004 at three organic mixed vegetable farms in British Columbia, Canada and Washington, USA and a series of replicated, controlled field cage experiments at a research station in Washington, USA (Prasad & Snyder 2006) found fields with beetle banks had higher beetle (Coleoptera) activity densities than fields without banks (figures not given). However predation rates of housefly Musca domestica eggs were not associated with activity densities of either small beetles (< 1 cm long ground beetles (Carabidae) and rove beetles (Staphylinidae)) or the large ground beetle Pterostichus melanarius. Small beetle activity densities were reduced when P. melanarius individuals were added to 2 x 2 x 2 m caged areas of a radish Rhaphanus sativus field and the number of housefly eggs predated was significantly reduced. The number of housefly eggs predated was lower when alternative aphid (Aphididae) prey were present. Beetle banks 1.5 x 30-60 m (two banks 50 cm high, two field level) sown with orchardgrass Dactylis glomerata were established in April-June 2002. Five housefly eggs were placed on a 1 cm² peat block and covered with 0.5 cm soil at plant bases, five times/field.

18 

A review (Tillman et al. 2012) described one study (Collins et al. 2002, summarised above) which found that natural predators reduced aphid (Aphidoidea) numbers up to 58 m from a beetle bank, but with greatest reductions at 8 m from the bank. Another study (Thomas 1990) found reductions were highest on the beetle bank itself. Three studies (Thomas 2001 and Macleod et al. 2004, summarised above, and Collins et al. 2003) found between 18 and 2,180 natural predators/m² in beetle banks between 1987 and 1998, including 11-423 ground beetles (Carabidae), 1-1,550 rove beetles (Staphylinidae) and 6-470 spiders (Araneae)/m². Predator numbers on beetle banks (maintained for up to 10 years) were similar to or higher than numbers in field margins. Another study (Holland et al. 2004) found total numbers of predators varied from 11 to 29 individuals/m² (in July and June respectively) in a cereal field without a beetle bank. In 2002 a beetle bank cost £975/ha to establish and £2/ha in income lost (with each subsequent crop) through land being occupied by the beetle bank (Collins et al. 2002).

Additional references:

Collins K.L., Boatman N.D., Wilcox A.W. & Holland J.M. (2003) A 5-year comparison of overwintering polyphagous predator densities within a beetle bank and two conventional hedgebanks. Annals of Applied Biology, 143, 63-71.

Holland J.M., Winder L., Woolley C., Alexander C.J. & Perry J.N. (2004) The spatial dynamics of crop and ground active predatory arthropods and their aphid prey in winter wheat. Bullentin of Entomological Research, 94, 419-431.

Thomas M.B. (1990) The role of man-made grassy habitats in enhancing carabid populations in arable land. Pages 77-85 in: N.E. Stork (ed.) The Role of Ground Beetles in Ecological and Environmental Studies, Intercept Ltd., Andover, UK.

Referenced papers

Please cite as:

Wright, H.L., Ashpole, J.E., Dicks, L.V., Hutchison, J., McCormack, C.G. & Sutherland, W.J. (2017) Some Aspects of Enhancing Natural Pest Control. Pages 359-381 in: W.J. Sutherland, L.V. Dicks, N. Ockendon & R.K. Smith (eds) What Works in Conservation 2017. Open Book Publishers, Cambridge, UK