Plant nectar flower mixture/wildflower strips
Overall effectiveness category Evidence not assessed
Number of studies: 23
Background information and definitions
Butterflies and moths often show a preference for areas with more floral resources (Vogel et al. 2007), but such resources are absent from intensive arable landscapes. Sown nectar flower mixtures, or wildflower strips, are common components of agri-environment schemes or conservation incentive programs which aim to increase the availability of flowers for insects. However, there can be many differences in the implementation methods used. For example, the exact seed mixture of species sown often differs between geographic regions and individual studies. Excluding grasses from nectar flower mixtures may reduce competition and aid re-establishment of the flowering species, while the rotation of the location of wildflower strips every 2–3 years may also be beneficial, as not all sown species will persist for longer than this (Heard et al. 2011).
Nectar flower mixes are normally composed of perennial plants, as opposed to annual species sown in wild bird seed mixes (Pywell et al. 2008). For studies on sowing these mixes, see “Plant wild bird seed or cover mixture”.
Heard M.S., Botham M., Broughton R., Carvell C., Hinsley S., Woodcock B., Pywell R.F., Amy S., Bellamy P.E., Hill R.A., Hulmes S., Hulmes L., Meek W.R., Nowakowski M., Peyton J., Redhead J.W., Shore R.F. & Turk A. (2011) Quantifying the effects of Entry Level Stewardship (ELS) on biodiversity at the farm scale: the Hillesden Experiment. Natural England report, RP00026.
Pywell R., Hulmes L., Meek W. & Nowakowski M. (2008) Creation and Management of Pollen and Nectar Habitats on Farmland: Annual report 2007/8.
Vogel J.A., Debinski D.M., Koford R.R. & Miller J.R. (2007) Butterfly responses to prairie restoration through fire and grazing. Biological Conservation, 140, 78–90.
Supporting evidence from individual studies
A replicated, randomized, site comparison study in 1987–1991 in Oxfordshire, UK (Feber et al. 1994, same experimental set-up as Feber & Smith 1995, Feber et al. 1996) found that field margins sown with wildflower seed mix had more adult meadow brown Maniola jurtina than unsown margins, and that margin management affected butterfly numbers. More adult meadow browns were found on margins sown with wildflowers (4–52 butterflies/50 m) than on unsown margins (4–15 butterflies/50 m). Sown margins also had more butterflies if they were left uncut (13–39 butterflies/50 m), or were cut in spring and autumn (16–52 butterflies/50 m), than if they were cut in summer (4–22 butterflies/50 m). There was no difference in the abundance of meadow brown caterpillars between sown and unsown, or cut and uncut, plots (3 caterpillars/plot). There were more meadow browns on all the experimental field margins than on narrow, unmanaged field boundaries of a neighbouring farm (numbers not given). In October 1987, two-metre-wide field margins around arable fields were rotovated, and either sown with a wildflower seed mix in March 1988 or left to regenerate naturally. Within each sown and unsown margin, 50-m-long plots were managed in one of four ways, with eight replicates of each treatment: uncut; cut once in June; cut April and June; cut in April and September. Hay was collected after cutting. From June–September 1989, and April–September 1990–1991, adult meadow brown were monitored weekly. In spring 1991, meadow brown caterpillars were sampled by sweep netting and visual searching.Study and other actions tested
A site comparison study in 1988–1991 on two arable farms in Oxfordshire, UK (Feber & Smith 1995, same experimental set-up as Feber et al. 1994, 1996) reported that a farm where wider field margins (some of which were sown with grasses and non-woody broadleaved plants “forbs”) had been established, and fertilizer application excluded, had a higher abundance and species richness of butterflies than a farm with conventional field margins. Results were not tested for statistical significance. The abundance of eight species (including small skipper Thymelicus sylvestris, small heath Coenonympha pamphilus, gatekeeper Pyronia tithonus and meadow brown Maniola jurtina) was higher on a farm with wider (2-m) field margins, including those sown, than on a farm with conventional (0.5-m) margins. No species was more abundant on the conventional farm. In addition, two species (marbled white Melanargia galathea and common blue Polyommatus icarus) were only recorded on the farm with wide margins, resulting in a higher species richness (16 species) than the conventional farm (14 species). In 1988, the margins of 10 fields on one farm were extended from 0.5-m to 2-m wide, and fertilizer application was excluded. Margins were either sown with grasses and forbs or left to regenerate naturally. Margins were either left uncut, or cut in some combination of April, June and September 1989–1991. In summer 1991, butterflies were surveyed for two months on transects on this farm and on a second, intensively managed farm with conventional field margins (number not given).Study and other actions tested
A replicated, randomized, site comparison study in 1987–1991 in Oxfordshire, UK (Feber et al. 1996, same experimental set-up as Feber et al. 1994, Feber & Smith 1995) found that butterfly abundance and species richness were higher in sown wildflower margins than in unsown, naturally generated margins. From two years after establishment, both individual abundance (21–91 individuals/50 m) and species richness (7–10 species/50 m) of butterflies were higher in sown wildflower margins than in unsown margins (abundance: 14–39 individuals/50 m; richness: 6–9 species/50 m). Additionally, in all three years, sown margins had more butterflies if they were left uncut (abundance: 49–91 individuals/50 m; richness: 9–10 species/50 m), or were cut in spring and autumn (abundance: 27–88 individuals/50 m; richness: 7–9 species/50 m), than if they were cut in summer (abundance: 21–46 individuals/50 m; richness: 7–10 species/50 m). In autumn 1987, two-metre-wide field margins around arable fields were rotovated. In April 1988, half were sown with a seed mixture (3 kg/ha) containing six grasses and 17 non-woody broadleaved plants (forbs). The rest were left to regenerate naturally. Within each sown and unsown margin, 50-m-long plots were managed in one of four ways, with eight replicates of each treatment: uncut; cut once in June; cut April and June; cut in April and September. Hay was collected after cutting. Butterflies were monitored weekly from June–September 1989 and from April–September 1990 and 1991.Study and other actions tested
A replicated, site comparison study in 1994–1996 on an arable farm in Gloucestershire, UK (Feber & Hopkins 1997) found that the abundance and diversity of butterflies was higher in margins sown with wildflowers than in naturally regenerated field margins, and that margin management affected butterfly diversity. Sown wildflower margins had a higher abundance (15–16 individuals) and diversity (6–7 species) of butterflies than naturally regenerated margins (abundance: 5–10 individuals; diversity: 3–6 species). Cutting and subsequent grazing of the sown margins decreased butterfly diversity (5.6 species) but not abundance (14.6 individuals) compared to margins which were not cut or grazed (diversity: 6.8 species; abundance: 16.3 individuals). In 1994, two-metre margins were established around two organically managed arable fields by either sowing a seed mix (containing five grasses and six wildflowers) or by natural regeneration. In 1996, half of the margins were cut in June and grazed in July. The rest were left unmanaged. From May–September 1996, butterflies were monitored weekly along transects.Study and other actions tested
A replicated, site comparison study in 1998 in two agricultural regions in the Swiss Plateau, Switzerland (Jeanneret et al. 2000, same study as Jeanneret et al. 2003) found that the species richness of butterflies in wildflower strips sown on set-aside areas was similar to winter wheat fields. Butterfly species richness in the wildflower strips (more than 6 species) differed significantly only from forest edges (fewer than 4 species). However, wildflower strips attracted some species of butterfly that were never or only rarely found in other habitats. Across two arable regions, 109 sites were composed of eight habitat types: Ecological Compensation Areas including eleven wildflower strips on set-aside land, five hedgerows, 19 extensively managed meadows, 16 low intensity meadows and eight orchard meadows, along with 20 winter wheat fields, seven intensively managed meadows and 23 forest edges. From May–September 1998, butterflies were observed for 10 minutes on each of six visits to each site (0.25 ha/site).Study and other actions tested
A replicated study in 1996–1997 on an experimental farm in Hertfordshire, UK (Carreck & Williams 2002) reported that sown nectar flower mixtures were used by six species of butterfly and moth. Five species of butterfly (small skipper Thymelicus sylvestris (3 individuals), common blue Polyommatus icarus (4 individuals), small tortoiseshell Aglais urticae (17 individuals), painted lady Cynthia cardui (4 individuals) and small white Pieris rapae (18 individuals)) and one moth (silver Y Autographa gamma (327 individuals)) used nectar flower mixtures sown with six plant species. Phacelia Phacelia tanacetifolia and borage Borago officinalis attracted the highest diversity of butterflies and moths (5 species), but individual species preferred different plants (see paper for details). From mid-April to mid-July 1996 and 1997, one plot/month (22 × 14 m in 1996, 20 × 13 m in 1997) was sown with a seed mixture containing borage, phacelia, buckwheat Fagopyrum esculentum, cornflower Centaurea cyanus, mallow Malva sylvestris and marigold Calendula officinalis at 91 kg/ha (1996) or 22 kg/ha (1997), and then harrowed. Flower-visiting butterflies and moths were recorded on 34 days from June–October 1996 and 21 days from June–November 1997 by walking around the edge of each plot.Study and other actions tested
A replicated, controlled study in 1999–2000 in four arable sites in North Yorkshire, UK (Meek et al 2002) found that butterfly abundance was higher along margins sown with a grass and wildflower seed mix or a combination of grass and wildflower seed mix and tussocky grass seed than along margins left to regenerate naturally or those sown with cereal crop, but there was no difference in butterfly species richness between any of the treatments. More butterflies were seen along margins sown with a grass and wildflower seed mix (average: 44 individuals) or where half their width had been sown with grass and wildflower seed mix and half with tussocky grass seed (average: 33) than along margins which regenerated naturally (average: 14) or those sown with cereal crop seed (average: 9). However, there was no difference in butterfly abundance along margins sown with a grass and wildflower seed mix or a split margin of grass and wildflower seed mix and tussocky grass seed than along margins sown with tussocky grass seed alone (average: 21). There was no difference in species richness between any of the treatments (see paper for details). Four 6 m wide margins of winter cereal fields (all adjacent to hedges) on two farms were split into 72 m long plots and sown in September 1999 with either a tussocky grass mix, a grass and wildflower mix, half the width tussocky grass and half grass and wildflower mix, cereal crop or left to regenerate naturally. Butterflies were surveyed weekly from May–September 2000 using walking transects (21 surveys/field in total).Study and other actions tested
A replicated, site comparison study in 1998 in the arable region of Rafz, Swiss Plateau, Switzerland (Jeanneret et al. 2003, same study as Jeanneret et al. 2000) found that butterfly species richness was higher in wildflower strips than in intensively managed wheat fields. Wildflower strips planted as Ecological Compensation Areas had more species of butterfly than intensively managed wheat fields (data not presented). Eleven wildflower strips and 20 wheat fields were sampled. Butterflies were observed for 10-minute periods on 0.25 ha of each site, on five occasions from May–August 1998, between 10:00–17:30 h on sunny days with temperatures of at least 18 °C.Study and other actions tested
A replicated, paired, controlled study in 2003 on an arable farm in North Carolina, USA (Forehand et al. 2006) found that plots sown with one of three commercial seed mixes had a higher abundance of adult moths than the other two mixes or single species plots, and all three mixes had a higher species richness of caterpillars than single species plots. Plots sown with “Border Patrol” seed mix were visited by more adult hawk moths (Sphingidae: 1.8 individuals/minute) and noctuid moths (Noctuidae: 2.1 individuals/minute) than plots sown with “Beneficial Insect Mix” or “Good Bug Blend” (hawk moths: 0.1; noctuid moths: 0.8–1.1 individuals/minute) or single species of cut flowers or herbs (hawk moths: 0.0; noctuid moths: 1.0–1.6 individuals/minute). The species richness of non-pest herbivores (including geometrid moth (Geometridae), brush-footed butterfly (Nymphalidae) and skipper (Hesperiidae) caterpillars), was similar in all three seed mixes (8 species/plot) and higher than in the single species plots (5–6 species/plot). In March 2003, seeds from three commercial mixes were separated and sown in a greenhouse, along with seeds of three cut flowers/herbs (fennel Foeniculum vulgare, common zinnia Zinnia elegans, cockscomb Celosia cristata). In May 2003, plants were transplanted to field plots (6.0 × 1.2–2.1 m), arranged in three blocks, in their original mixes and relative abundance. Dead plants were replaced for two weeks, and hand-weeded. Plots were separated by 1.5-m millet strips. All areas had been pesticide-free for ≥3 years. On eight days in June–August 2003, insects were collected with a D-Vac vacuum sampler, and two aerial nets, for 1 min/plot. On four nights in July–August 2003, moths were observed visiting each plot for one minute, three times/night, in the hour after dusk.Study and other actions tested
A replicated, site comparison study in 2001–2005 in an arable region in Basel, Switzerland (Jacot et al. 2007) found that wildflower strips sown with a ‘locally adapted’ mix of grass and flower species (species local to the area) had a higher abundance and species richness of butterflies than conventionally cropped margins. The ‘locally adapted’ field margins had more butterfly species and individuals than standard wildflower strips, and 40 times more species and individuals than conventional cropped margins (data not presented). In 2001, seven field margins (5 x 120 m) were sown with seeds of up to 38 native grass and wildflower species (‘locally adapted’ mix). Half of each margin was cut lengthwise, alternately, in late August each year. Butterflies were counted five times from May–August 2003 and 2005 on these ‘locally adapted’ margins, 10 standard wildflower strips, and 10 conventional cropped margins.Study and other actions tested
A replicated, randomized, controlled study in 2002–2006 on three farms in eastern England, UK (Pywell et al. 2007) found that field margins sown with a flower mix designed for pollinating insects did not support more butterflies than floristically-enhanced grass margins, but both supported more butterflies than grass-only margins. In margins sown with either a pollinating insect mix or a grass and wildflower mix the abundance (18–20 individuals/plot) and species richness (6 species/plot) of butterflies was similar, but both were higher than in grass-only margins (abundance: 12 individuals/plot; richness: 5 species/plot). Management of the margins did not affect either the abundance or species richness of butterflies (data not presented). Field margin plots (6 × 30 m) were established in 2000–2001 using one of three seed mixes: a mixture of grasses and wildflowers designed for pollinating insects (four grass species, 16–20 wildflowers, sown at 35 kg/ha), a floristically-enhanced “tussock grass mix” (seven grass species, 11 wildflowers, sown at 35 kg/ha), and a grass-only “Countryside Stewardship mix” (seven grass species, sown at 20 kg/ha). Margins were managed in spring from 2003–2005 with one of three treatments: cut to 15 cm, soil disturbed by scarification until 60% of the area was bare ground, treated with grass-specific herbicide in spring at half the recommended rate. There were five replicates of each treatment combination on three farms. No further details provided.Study and other actions tested
A replicated, randomized, controlled study in 2006–2007 on a farm in Warwickshire, UK (Pywell et al. 2008) found that butterfly abundance and species richness differed between three plant species commonly sown in nectar flower mixtures. In 2006, more butterflies were found in plots sown with lucerne Medicago sativa (6.3 individuals/plot) than plots sown with borage Borago officinalis (0.3 individuals/plot), chicory Cichorium intybus (0.8 individuals/plot) and sainfoin Onobrychis viciifolia (0.8 individuals/plot). More butterfly species were found in lucerne plots (3.5 species/plot) than in borage, chicory, sainfoin and fodder radish Raphanus sativus (0.3–0.5 species/plot). There was no significant difference in abundance (1.0–3.3 individuals/plot) or species richness (1.0–2.8 species/plot) between the other plant species. In 2007, red clover Trifolium pratense had more butterflies (3.3 individuals/plot) than chicory (0.0 individuals/plot), but the abundance on all other plant species was similar (0.3–2.3 individuals/plot) and there were no significant differences in species richness (0.0–1.8 species/plot). In May 2006, three perennial species sown in pollen and nectar mixtures (chicory, red clover, sainfoin) and 10 small-seeded crop species commonly sown in wild bird seed mixtures (borage, buckwheat Fagopyrum esculentum, crimson clover Trifolium incarnatum, fodder radish, linseed Linum usitatissimum, lucerne, mustard Brassica juncea, phacelia Phacelia tanacetifolia, sunflower Helianthus annuus and sweet clover Melilotus officinalis) were sown individually in 6 × 4 m plots, replicated four times. Annual species were re-sown May 2007. Butterflies were surveyed six times/year, from July–September 2006 and May–September 2007.Study and other actions tested
A replicated, controlled study in 1998–2004 in an arable farmland region in central Switzerland (Aviron et al. 2009) found that wildflower strips contained similar numbers of butterfly species to crop fields. The estimated number of butterfly species on wildflower strips (19 species) was the same as on conventional crop fields (19 species). The study sampled 78 wildflower strips (sown with 20–40 plant species) and 72 crop fields. From 1998–2004, butterflies were surveyed every two years between May and September, using five 10-minute observation periods across 0.25 ha/field.Study and other actions tested
A replicated, randomized, controlled study in 2002–2006 on four lowland farms in Devon and Somerset, UK (Potts et al. 2009) found that plots sown annually with mixes including legumes had a higher abundance and species richness of butterflies, but a lower abundance of caterpillars, than grassland plots. In the first two years, plots sown with legume mixes had a higher abundance (4–10 individuals/transect) and species richness (2–4 species/transect) of adult butterflies than extensively (abundance: 3–5 individuals/transect; richness: 2 species/transect) or conventionally managed (abundance: 1–2 individuals/transect; richness: 1 species/transect) grassland. However, there were fewer caterpillars in the sown plots (0–3 caterpillars/transect) than the extensively (1–8 caterpillars/transect) or conventionally managed (0–7 caterpillars/transect) grassland. In April 2002, experimental plots (50 × 10 m) were established on permanent pastures (>5-years-old) on four farms. There were nine treatments, with three replicates/farm. Two legume-sown treatments comprised barley Hordeum vulgare undersown with seven grasses and five legumes, and a mix of six crops and four legumes. Two extensive grassland treatments had minimal disturbance during summer and five conventional grassland treatments included modifications to conventional silage management (reducing fertilizer application, cutting and grazing). From June–September 2003–2006, butterflies were surveyed once/month on a 50-m transect through the centre of each plot. In April, June, July and September 2003–2006, caterpillars were counted (but not identified) on two 10-m transects/plot using a sweep net (20 sweeps/transect).Study and other actions tested
A replicated, site comparison study in 2007 at four arable farms in south Sweden (Haaland & Gyllin 2010) found higher abundance and species richness of butterflies in sown wildflower strips than in grass margins (greenways or ‘beträdor’). In wildflower strips, the abundance of butterflies (10.4 individuals/100 m) was higher than in grass margins (0.6–1.4 individuals/100 m). In total, 86% of the recorded butterflies were found in the wildflower strips compared to 14% in the grass margins. Four species of butterfly were only found in the wildflower strips. Margins with more field scabious Knautia arvensis had higher species richness and abundance of butterflies (data presented as model results). At one farm, six wildflower strips (total 2.9 km) were sown in the mid-1990s using either a commercial mix of wildflowers and grasses, or hay from a nearby meadow, and were cut once a year at the end of July. At three farms, 14 grass strips (total 6.8 km) were sown with a mixture of grass species in the 1990s, 2004 and 2005, and were cut several times a year. Butterflies and the abundance of key flower species were recorded on transects five times from June–September 2007.Study and other actions tested
A replicated, site comparison study in 2000–2004 in an arable landscape in the Swiss Plateau, Switzerland (Aviron et al. 2011) found that wildflower strips contained a higher abundance and species richness of generalist but not specialist butterflies than other arable habitats. For generalist butterflies, both the average abundance (24.0 individuals) and species richness (7.0 species) were higher in wildflower strips than in conventional grassland (abundance: 12.0, richness: 5.0) or wheat, maize and root crop fields (abundance: 2.6–3.7, richness: 1.8–2.2). However, for specialist butterflies there was no significant difference in abundance or richness (wildflower: abundance = 2.4, richness = 1.0; grassland: abundance = 0.6, richness = 0.5; crops: abundance = 0.4, richness = 0.2). Species richness of generalists was also higher in fields with more wildflower strips in the surrounding area (data presented as model results). From 1994–2004, within an 822-ha arable landscape, wildflower strips were sown with buckwheat as ground cover, and 30–40 wild plant species. They received no fertilizer or pesticide, and were not cut between 15 March and 1 October. In 2000, 2002 and 2004, butterflies were surveyed in five habitats: wildflower strips, conventional grassland, wheat fields, root crops and maize fields. Each year, 37–39 fields were sampled with 5 × 10-minute surveys every 2–3 weeks between May and August. The surrounding land cover (200-m radius) was mapped from aerial photographs. Generalist and specialist species were determined based on the number of caterpillar food plants.Study and other actions tested
A replicated, site comparison study in 2008 in a lowland agricultural landscape in Kanton Fribourg, Switzerland (Haaland & Bersier 2011) found that sown wildflower strips had a higher abundance of butterflies, but a similar species richness and a different community composition, compared to extensively managed meadows. The abundance of butterflies in sown wildflower strips (0.29 individuals/m) was higher than in extensively managed meadows (0.12 individuals/m), but the species richness was similar in wildflower strips (0.07 species/m) and meadows (0.05 species/m). The species composition was different between the two habitats, with seven of 25 species occurring only in wildflower strips, and six species observed most frequently in the meadows. None of the five rarest species in the region were recorded in wildflower strips or meadows. See paper for details on individual species. Twenty-five wildflower strips (0.15–1.16 ha) were sown with a standard seed mixture of 24 plant species, and were 1–7 years old. Eleven meadows (0.21–1.64 ha) were cut at least twice/year after mid-June. From May–September 2008, butterflies were surveyed once/month on a transect through the middle of each wildflower strip (70–450 m) or meadow (85–310 m).Study and other actions tested
A replicated, randomized, paired, controlled study in 2005–2011 on an arable farm in Buckinghamshire, UK (Heard et al 2011) found that land managed under an agri-environment scheme, including sowing nectar flower mixtures, had a higher abundance, but not species richness, of butterflies and micro-moths than conventional farms, but there was no difference in abundance or species richness of other moths. Butterfly abundance was higher under enhanced Entry-Level Stewardship (ELS) (5,400 individuals/60 ha) and standard ELS (2,000 individuals/60 ha) than under conventional farming (1,400 individuals/60 ha). Micro-moth abundance was also higher under enhanced ELS (79 individuals) than standard ELS (32 individuals) or conventional farming (20 individuals). However, the abundance of macro-moths and threatened moths was similar under enhanced ELS (macro: 126; threatened: 6 individuals), standard ELS (macro: 79; threatened: 5 individuals) and conventional farming (macro: 79; threatened: 6 individuals). Species richness of all groups was similar under enhanced ELS (macro: 20; micro: 11; threatened: 3 species), standard ELS (macro: 20; micro: 8; threatened: 2 species) and conventional farming (macro: 18; micro: 5; threatened: 2 species) (butterfly data not presented). In 2005, a 1,000-ha farm was divided into five 180-ha blocks. Three 60-ha areas/block were assigned to three treatments: enhanced ELS (5% land removed from production, field corners sown with four grasses and 25 non-woody broadleaved plants (forbs), nectar flower mixtures sown with four legumes); standard ELS (1% land removed from production); conventional farming (see paper for other details). From May–August 2006–2011, butterflies were recorded four times/year on one 50-m transect/60-ha area, passing through all available habitats. In late-May 2007–2011 and late-July 2006–2011 moths were surveyed using Robinson light traps. One block was surveyed/night, with one trap/treatment.Study and other actions tested
Referenced paperHeard M.S., Botham M., Broughton R., Carvell C., Hinsley S., Woodcock B., Pywell R.F., Amy S., Bellamy P.E., Hill R.A., Hulmes S., Hulmes L., Meek W.R., Nowakowski M., Peyton J., Redhead J.W., Shore R.F. & Turk A. (2011) Quantifying the effects of Entry Level Stewardship (ELS) on biodiversity at the farm scale: The Hillesden Experiment. Natural England report, CEH report to Natural England, no. MA01031 (Project:RP00026).
A replicated, randomized, paired, controlled study in 2003–2005 in an arable farm in Yorkshire, UK (Pywell et al. 2011) found that margins sown with complex seed mixes did not have a higher abundance or species richness of butterflies than simple mixes, but the timing of cutting and removal of cuttings did affect butterfly numbers. The abundance and species richness of butterflies was similar in plots sown with simple (abundance: 6–11 individuals/150 m2; richness: 3–4 species/150 m2) or complex (abundance: 6–8 individuals/150 m2; richness: 3 species/150 m2) seed mixes. In the first year of management, butterfly abundance (1–2 individuals/150 m2) and species richness (1 species/150 m2) were lower in plots cut in June than in plots cut at other times of year (abundance: 17 individuals/150 m2; richness: 4–5 species/150 m2). However, in the second year, there was a higher abundance, but not species richness, of butterflies in plots cut in April and June (9 individuals/150 m2; 4 species/150 m2) or October (8 individuals/150 m2; 5 species/150 m2) than plots cut in April and October (3 individuals/150 m2; 2 species/150 m2). Plots where cuttings were removed in April and June had a higher abundance and species richness of butterflies than plots where cuttings were left, but plots where cuttings were removed in October had a lower abundance and species richness than plots where cuttings were left (data not presented). In April 2003, two margins (200 × 6 m) were established in each of two cereal fields. One margin/field was sown with six legumes, common knapweed Centaurea nigra, and six grasses at 20 kg/ha (£140/ha), and the other was sown with four legumes and three grasses at 20 kg/ha (£55/ha). Margins were cut three times in 2003 with cuttings removed. In April 2004, each margin was sub-divided into eight 25 × 6 m plots, which were randomly assigned to one of eight treatments: cut in October, cut in October and April, cut in October and June, or cut in April and June, each with cuttings left in place or removed. From May–September 2004–2005, butterflies were counted 7–8 times/year on a 25-m transect through the middle of each plot.Study and other actions tested
A replicated, controlled study in 2007–2010 in six arable fields in Jokioinen, Finland (Korpela et al. 2013) found that sown wildflower strips had a higher abundance of habitat specialist butterflies and total species richness of butterflies, moths and bumblebees (Bombus spp.) combined than either grass or cereal fields, or permanent field margins. Three years after sowing, the abundance of habitat specialist butterflies (1.5–3.1 individuals/strip) and total species richness of butterflies, moths and bumblebees (16–21 species/strip) were higher in wildflower strips than in reed canary grass Phalaris arundinacea (butterflies: 0.1–0.6 individuals/strip; richness: 2–7 species/strip), spring cereals (butterflies: 0 individuals/strip; richness: 1 species/strip) or permanent margins (butterflies: 0.9–1.2 individuals/strip; richness: 10–12 species/strip). Neither the diversity of the sown seed mixture, nor the shape, location or orientation of the wildflower strip, affected butterfly abundance or total species richness (see paper for details). In May 2007, six wildflower strips were sown in each of six fields. Five strips/field were sown with five wildflower species, and one was a monoculture of brown knapweed Centaurea jacea. Five strips/field were 5 × 50 m, and one was 10 × 25 m. Strips were located at the field edge, either adjacent to another field or to forest, or in the centre of the field. From May–August 2007–2010, butterflies, moths and bumblebees were surveyed seven times along one 5 × 50 m transect/wildflower strip, and in four strips/field within the surrounding crop (reed canary grass or spring cereals) and two strips/field in permanent, unsown field margins.Study and other actions tested
A study in 2009 on a peanut-cotton farm in Georgia, USA (Tillman & Carpenter 2014) reported that tropical milkweed Asclepias curassavica plants placed between peanut and cotton fields were used by monarch butterflies Danaus plexippus. Milkweed plants were visited by 0–0.03 monarch butterflies/plant/observation. Monarch caterpillars were also observed feeding on milkweed plants and developing in to pupae. In 2009, four plots (23 × 61 m) were established between a 10-ha peanut field and 9-ha cotton field (each planted in May). Two weeks before cotton bolls appeared, 25 potted, greenhouse grown, flowering tropical milkweed plants/plot were placed 1.2 m apart along a 1-m-wide strip of bare ground between the crops. On eight days in August 2009, each milkweed plant was observed for 15 seconds/day to record adult monarchs feeding on the flowers, and the presence of caterpillars was noted.Study and other actions tested
A replicated, randomized, controlled study in 2007–2010 on 28 arable farms in Wessex and East Anglia, UK (Holland et al. 2015) found that farms with enhanced agri-environment scheme (AES) habitats, including wildflower strips, had a higher abundance of some butterfly species than farms with simpler AES habitats. In early summer, farms with enhanced AES habitats had a higher abundance of blue (Lycaenidae: 0.05 individuals/100 m) and white (Pieridae: 0.46 individuals/100 m) butterflies along boundaries than farms with Entry Level Scheme (ELS) habitats (blues: 0.04; whites 0.21 individuals/100 m), but a lower abundance of skippers (Hesperiidae) in the AES habitat itself (enhanced: 0.00; ELS: 0.02 individuals/100 m). In mid-summer, enhanced AES farms had a higher abundance of white butterflies (0.69 individuals/100 m), but a lower abundance of brown butterflies (Satyridae: 0.16 individuals/100 m) in the AES habitat, and a lower abundance of blue butterflies (0.05 individuals/100 m) along boundaries than ELS farms (whites: 0.38; browns: 0.49; blues: 0.11 individuals/100 m). In spring 2007, twenty-four farms (12 in East Anglia and 12 in Wessex) were randomly assigned to two treatments: 16 farms with enhanced AES habitat (1.5–6.0 ha of wildflower strips, floristically-enhanced grass mixes, wild bird seed mixes and natural regeneration by annual cultivation); and eight farms with ELS habitat (1.5–6.0 ha of grass margins and game cover (usually maize)). Two additional ELS farms/region, already managed organically with 1.5 ha of ELS habitat, were also studied. From 2008–2010, butterflies were surveyed twice/year on 11 fixed 100-m transects, in mid-May–mid-June and mid-July–early August. Eight transects/site were located in AES habitat, and three transects/site were located on field boundaries away from the AES habitat.Study and other actions tested
A replicated, controlled, before-and-after study in 2010–2015 at two arable sites in Baden-Württemberg, Germany (Buhk et al, 2018) found that five years after starting to sow wildflower strips butterfly species richness increased at one site, but not at the other, compared to before sowing, and there was no difference in species diversity between before and after sowing at either site. At one site there were more butterfly species five years after sowing wildflower strips than before sowing (maximum species after: 21; before: 10), and on average there were seven more species in the areas sown with wildflowers than in areas not sown with wildflowers. At the other site there were no differences. There were no differences in species diversity between before and after sowing wildflowers in the areas that had been sown at either site. From 2011–2015, mixtures of annual seeds were sown in spring in strips distributed across 50 ha areas on two farms. From 2012–2015 perennial and winter-hardy seeds were also sown in autumn. On each farm, a 50 ha area with no wildflower sowing was also monitored as a control. From 2010–2015, butterflies were surveyed along 250 m transects in the most florally diverse areas of the treatment and control areas (5 transects in the treatment areas, 4 in the controls, in both farms). Each transect was surveyed twice daily on five days between April and August each year.Study and other actions tested