Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Reduce fertilizer, pesticide or herbicide use generally Farmland Conservation

Key messages

Read our guidance on Key messages before continuing

  • Of 38 individual studies from Austria, the Czech Republic, Denmark, Finland, France, Germany, Ireland, the Netherlands, Sweden and the UK investigating the effects of reducing fertilizers, pesticides or herbicides, 34 studies (23 replicated, of which six also controlled and randomized, one review and one systematic review) found benefits to some invertebrates, plants, or farmland birds. Twenty-five studies (16 replicated, of which seven also randomized and controlled and one review) found negative, mixed, minimal or no effects on some invertebrates, farmland birds or plants.
  • Ten studies (six replicated, controlled studies of which two randomized) from three countries found positive effects of reducing or stopping pesticide applications on invertebrates, plants, or birds. Eight studies (two replicated controlled and randomized, one paired before-and-after trial) from four countries found inconsistent or no effects on some invertebrates or birds.
  • Ten studies (nine replicated, five also controlled and a European systematic review) from four countries found positive effects of reducing or stopping herbicide use on plants, invertebrates, and birds. Five replicated studies (two also controlled and randomized) from three countries found no or mixed effects on birds, invertebrates and plants.
  • Five studies (three replicated controlled of which two randomized) from four countries found positive effects of reducing or stopping fertilizer applications on invertebrates, Eurasian skylark, or plants. Four studies (three replicated, controlled and randomized) from two countries found reducing or stopping fertilizer inputs had no, or no consistent effects on some invertebrates and farmland birds. Two studies from the UK (one replicated) found plots where fertilizer inputs were not reduced tended to have higher earthworm biomass or abundance.
  • Fifteen studies (three replicated controlled of which one also randomized, five site comparisons and one review) from seven countries looked at the effects of reducing or stopping applications of two or more inputs: pesticides, herbicides, or fertilizers. Thirteen studies found positive effects of reducing two or more inputs on some or all invertebrates, plants, soil organisms, and birds studied. Seven studies found negative or no effects of reducing combinations of inputs on some invertebrates, plants or birds.


Supporting evidence from individual studies


A study of three-and-a-half-year grass/clover Trifolium spp. leys from 1950 to 1956 in the UK (Heath 1962) found that plots that were grazed rather than cut and with nitrogen applications had higher earthworm (Lumbricidae) mass in some years. Overall abundance of worms was 17-39/cubic foot. Plots with nitrogen applications had a significantly greater weight of earthworms than those without nitrogen applications in 1953 (9-10 vs 5-6 g/cubic feet), but not overall (8 vs 7 g/cubic feet). However, there tended to be lower numbers of worms in plots with nitrogen (17-30) than those without nitrogen (19-39/cubic foot). One plot of each treatment combination was established in each of the six years (the study also compared cutting and grazing management). Fertilizer was applied at 280 lb N/acre/year. Earthworms were sampled when plots were ploughed out of leys in the autumn (1953-1956). Four samples of two cubic feet of soil were sampled in each plot per year.



A replicated trial on an experimental farm in eastern Scotland (Gerard & Hay 1979) found fewer earthworms (Lumbricidae) at lower nitrogen application rates. There were 78 earthworms/m2 and 0.42 tonnes earthworm/ha in plots with no nitrogen, compared to 106 earthworms/m2 and 0.53 tonnes/ha in plots with 100 kg N/ha. The highest nitrogen treatment (150 kg N/ha) had fewer earthworms but higher biomass because there were more large-bodied species like Lumbricus terrestris (93 earthworms/m2, 0.59 tonnes/ha). Earthworm biomass decreased with decreasing nitrogen application at a rate of 0.06 t/ha for every 50 kg N/ha. Only one of the eight species recorded, Allolobophora rosea, was more abundant in plots with lower fertilizer inputs (9 earthworms/m2 at 0 kg N/ha, compared to 3.7 earthworms/m2 at the highest rate of 150 kg/ha). The experiment was replicated eight times. Spring barley crops were managed from 1967 until 1973 with either 0, 50, 100 or 150 kg N/ha added annually.



A replicated study of an arable field in Ireland (Purvis & Curry 1984) found that invertebrate abundance tended to be greater where no herbicides were applied compared to sprayed areas. A greater number of detritus feeders (2,136 vs 637-674), particularly beetles (Coleoptera) and larval and adult flies (Diptera) and herbivores (2,061 vs 174-333) were found in the unsprayed plots compared to sprayed plots, once weed populations were established. Overall predator numbers differed little between treatments (unsprayed: 2,422, sprayed: 2,142-2,356), although more predatory rove beetles (Staphylinidae) (324 vs 78-149) and parasitic wasps (Hymenoptera) (376 vs 72-87) were found in unsprayed plots towards the end of the sampling period. Ground beetles (Carabidae), which were the most numerous predators, showed no difference between treatments (unsprayed: 1,312, sprayed: 1,543-1,606). Inorganic fertilizers were applied in typical applications to sown sugar beet Beta vulgaris. Three treatments were then applied, each replicated in two plots (10 x 25 m): application of pre- and post-emergence herbicides (control: Lenacil and Phenmedipham), application of pre- and post-emergence herbicides plus farmyard manure, and no herbicide application. Percentage weed cover was estimated in five 0.09 m² quadrats/plot in June 1979. Nine pitfall traps/plot (5.6 cm diameter) were set for four 7-day trapping periods (May-September).


A replicated, randomized, controlled study of arable fields between 1982 and 1984 in England (Linzell & Madge 1986) found that the abundance of soil nematodes (Nematoda), slugs (Gastropoda) and fly (Diptera) larvae was greater in plots without pesticide (insecticide) applications. In spring 1983, numbers of nematodes were significantly higher in the plots without pesticide applications (5.5-6.5/50 g) compared to plots sprayed two weeks previously (2.5-3.5). Numbers of slugs did not differ between treatments in the first year but were significantly lower in sprayed plots in 1983-1984 (1-3 vs 10-26/tile). Overall, numbers of fly larvae were higher in plots without pesticide applications (25-75 individuals/replicate cores vs 7-65). Fertilizer did not tend to have a significant effect on soil invertebrate numbers. Four replicated, randomized blocks each comprising 10 plots (6 x 3 m) were established. Treatments were three different grass (Italian rye grass Lolium multiflorum, perennial rye grass L. perenne, existing mixed ley) and two wheat regimes (‘Norman’ and ‘Armada’ varieties), with (phorate and aldicarb; three applications) and without pesticide treatments. Fertilizer was applied to all plots (except wheat ‘Norman’ in 1982) and fungicides applied to wheat when required. Invertebrates were sampled in the spring and autumn after pesticide applications. Free-living soil nematodes and fly larvae were sampled by taking two or four randomly located soil samples (soil corer: 2.5 x 15 cm and 6.5 x 8 cm respectively) from each plot. Slugs were sampled using two wooden tiles/plot with slug pellets underneath, which were collected after 4-7 days.



A replicated study on five arable fields in Austria (Kromp 1989) found that fields with no pesticides (fungicides or herbicides) or fertilizers had a greater diversity of ground beetles (Carabidae) than those with conventional chemical applications. Wheat fields with no spraying had greater numbers of ground beetle species (43-50 species) and individuals (5-6 individuals/trap/day) than those that received conventional pesticide and fertilizer applications (species: 38-40, individuals: 2-3/trap/day). Conventionally farmed sugar beet Beta vulgaris fields had similar numbers to conventional winter wheat (1/trap/day). Fields differed in terms of weed control (mechanical or herbicides), disease control (none or fungicides) and manuring (green/compost/stone meal or mineral). One or two wheat and/or sugar beet fields were under each treatment in 1982 and 1983. Invertebrates were sampled using a line of 6-10 pitfall traps in the centre of each field from May-July.


A site comparison study of arable farmland over seven years as part of the Boxworth Project in Cambridgeshire, UK (Fletcher et al. 1992) (same study as (Vickerman 1992)) found that only one of 11 bird species declined in numbers with high pesticide inputs and none of four species had reduced breeding performance. The percentage of the total territories in the high input area for the 11 species remained fairly constant during the project (pre-treatment: 44-46%, treatment: 40-50%). Common starling Sturnus vulgaris showed a significant decline in percentage of breeding territories in the high input area relative to the low input area from 1984 to 1987 (45% to 28%), numbers recovered in 1988 (41%). There were no significant differences in breeding performance between treatment areas for tree sparrow Passer montanus or starling in terms of numbers of breeding pairs, numbers of young fledged, numbers of young fledged/pair or causes of nest losses. The only difference was that the percentage of first broods that failed to produce at least one fledgling tended to be lower in the high input area in baseline years, but increased more than the low input areas in treatment years. However, overall production of young was not reduced in the high input area. Sample sizes were small for blue tit Parus caeruleus and great tit P. major nests. Pre-treatment years were 1982-1983 and treatment years 1984-1988. The ‘Common Bird Census’ method was used to monitor birds, with 10 visits from spring to early summer. A total of 220-244 nest boxes were put up in each area, which were checked weekly during the breeding season.



A controlled study as part of the Boxworth project comparing arable farmland with high and reduced pesticide inputs over five years in Cambridgeshire, UK (Vickerman 1992) (same study as (Fletcher et al. 1992)) found that the abundance of invertebrate herbivores, carnivores and parasitoids tended to be higher in areas with reduced pesticide applications, whereas detritus-feeding invertebrates did not differ with treatment. On average, total numbers of herbivores were 50% lower, predators 53% (39-70%) lower and parasitoids 39-79% lower in the conventional area compared to the reduced pesticide areas. Numbers varied with year, and numbers of some taxa were higher in conventional areas in some years. Numbers of detritus-feeders did not differ significantly between treatments. There were two treatment areas, one with conventional and the other reduced pesticide applications (selective insecticides and slug/snail pesticides). Invertebrate density was sampled in the middle of each field using a Dietrick vacuum sampler at intervals of 7-10 days between mid-April and harvest. Each sample comprised five sub-samples (each 0.09 m²) taken 10 m apart.



A replicated, controlled study in summer 1989-1990 in eight sites on one arable farm near Bonn, Germany (Henze & Şengonca 1993) found that a 50% reduction in pesticide application could control an aphid (Aphidoidea) outbreak as efficiently as the normal application. In farming systems with no insecticide use at all, natural predators reduced aphid populations to the same low levels (<5 aphids/plant), but the population decline occurred one week later than in the systems with pesticide use. Predatory arthropod populations also declined after pesticide treatment. Predator levels remained rather low in the normal pesticide system, however in the 50% reduced pesticide system they recovered in three weeks after pesticide application. Four farming systems were compared with two replicates each: conventional farming (normal pesticide use), integrated farming (50% reduction in pesticide use), ‘minimum’ farming (no insecticides, strongly reduced herbicide use) and ‘no pesticide’ farming (no pesticide use). Aphids and their predators were counted visually and with sweep nets once a week from April.



A site comparison study at the Lovinkhoeve Experimental Farm, Noordoostpolder, the Netherlands (Ruiter et al. 1993) found a higher biomass of microbes, protozoa, nematodes (Nematoda) and earthworms (Lumbricidae), but not of mites (Acari) and springtails (Collembola), in the upper 10 cm of an arable soil with reduced fertilizer and pesticide inputs, than in a conventionally managed soil. At lower depth (10-25 cm), there were no consistent differences in soil animals. The reduced input plot had 8.9 kg C/ha of earthworms in the top 10 cm, and 4.7 kg C/ha at 10-25 cm depth. No earthworms were recorded in conventional plots. Total biomass of nematodes in the upper layer was 0.79 kg C/ha in reduced input plots, and 0.3 kg C/ha in the conventional plots. Reduced input plots had applications of 65-170 kg nitrogen fertilizer/ha/year, compared to 130-285 kg N/ha on conventional plots. They also had reduced tillage. The experiment began in 1985. Soil samples were taken from three areas of each plot under winter wheat in 1986.



A randomized, replicated trial from 1987 to 1991 on the Oxford University Farm, Wytham, Oxfordshire, UK (Feber et al. 1994) found fewer adult meadow brown butterflies Maniola jurtina on 2 m-wide naturally regenerated field margins that were sprayed with herbicide once in summer, compared to any margins that were not sprayed. There were 3-4 meadow browns/50 m sprayed plot on average, compared to 4-13 meadow browns/50 m on unsown, uncut margins that were not sprayed. There was no difference between treatments in abundance of meadow brown larvae (3 larvae/plot on average). Two metre-wide field margins were established around arable fields in October 1987, rotavated and left to naturally regenerate from March 1988. Fifty metre-long plots were either uncut and unsprayed, subject to one of four different cutting regimes but unsprayed, uncut but sprayed once a year with herbicide (glyphosate) in late June or July. There were six replicates of each treatment. Adult meadow brown butterflies Maniola jurtina were monitored weekly along walked transects in the experimental plots from June to September 1989 and from April to September 1990 and 1991. Meadow brown larvae were sampled in spring 1991, by sweep netting and visual searching.



A replicated controlled study of seven arable fields on three farms in England (Frampton et al. 1994) found that reduced pesticide inputs tended to result in higher numbers of arthropods. Applications of chlorpyrifos spray in the conventional plots resulted in decreased numbers of ground beetle species (Carabidae: Bembidion aeneum, B. lunulatum, B. obtusum), one water scavenger beetle species (Hydrophilidae: Helophorus aquaticus), springtails (Collembola: Entomobrya multifasciata, Isotoma viridis, Sminthurides signatus, S. viridis) and money spiders (Linyphiidae: particularly Erigone spp.). Some of these species disappeared from sprayed plots and did not recover for a year. Spraying with deltamethrin also resulted in a decrease in water scavenger beetles Helophorus spp., several money spider species and one ground beetle species B. lunulatum, the latter did not recover for 15 months. Fields were divided in half with one receiving conventional pesticide applications, and the other reduced pesticides, i.e. lower herbicide and fungicide and no insecticides (1991-1996). All other practices were the same. Arthropods were monitored using a D-Vac suction sampler and pitfall traps. In each plot, four samples were taken, each comprising five sub-samples (total area 0.46 m²) between 25 and 125 m from the shared field margin. Four pitfall traps (9 cm diameter) were also located in each field half (12 m apart) and were operated for 7-day periods. This study was part of the same project (SCARAB – Seeking Confirmation About Results At Boxworth) as (Frampton 1997, Tarrant et al. 1997).



A small replicated controlled trial at two sites in Lower Saxony, Germany (Hasken & Poehling 1995) found that aphids (Aphidoidea) and their insect predators were less abundant in wheat fields not treated with fertilizers, insecticides or herbicides in 1992, compared to conventionally farmed fields or fields with low fertilizer use and no insecticide. A maximum of 80 aphids/wheat stem were recorded on plots with no chemicals, compared to 300 aphids/stem in the conventional farm system and close to 300 aphids/stem in fields with a 50% reduction in nitrogen fertilizer application (105 kg N/ha, compared to the conventional 210 kg N/ha) and no insecticide (herbicides were used). Fields with no chemical use had no more than 20 aphid predator larvae/m2; hoverflies (Syrphidae), ladybirds (Coccinellidae) and lacewings (Chrysopidae), compared with up to 60-70 larvae/m2 under conventional farming and up to 40 larvae/m2 with 50% fertilizer reduction and no insecticide. Under conventional farming, ladybirds were only recorded on plots not treated with insecticide. In farming systems with reduced or no chemical use, ladybirds were the dominant aphid predator in most months. This study was carried out on areas of 35 to 45 ha at two sites (two replicates of each farming system). Aphids and their predators were counted on 150 wheat stems twice a week and suction trapped every two weeks during the 1992 growing season. This study was part of the same project (INTEX – Integrated Farming and Extensification of Agriculture) and was carried out in partly the same research site as (Schmidt et al. 1995, Krooss & Schaefer 1998).



A replicated, controlled study in 1990-1994 on three arable farms in Lower Saxony, Germany (Schmidt et al. 1995) found significantly higher plant species diversity, weed cover and seed numbers in the seed bank in an ‘integrated’ farming system with a 50% reduction in chemical inputs (fertilizer, pesticides/herbicides) than in a conventional farming system. Species richness, weed cover and seeds in the soil were also higher in the extensive (no input) farming system than in the conventional system, but did not differ from the integrated farming. Crop cover, however, was significantly reduced only in the extensive farming system. Thus, a 50% reduction in herbicide use was the most efficient way of combining the economic interests of agriculture (crop yield) with weed protection. On three farms, field trials with three different farming systems were compared: conventional farming (normal pesticide/herbicide use and fertilization), integrated farming (50% reduction in pesticide/herbicide use, 25-40% reduction in mineral fertilization), and extensive farming (no pesticide/herbicide use, no mineral fertilization). Plants were monitored several times a year in four permanent plots (10 x 10 m) on two of the farms. Soil samples (0-5 cm and 5-30 cm depth) were taken in March 1990 and 1993 on all three farms. Seeds were germinated in the laboratory for 20 months after different growth stimulations. This study was part of the same project (INTEX – Integrated Farming and Extensification of Agriculture) and was carried out in partly the same research site as (Hasken & Poehling 1995, Krooss & Schaefer 1998).


A replicated, controlled, randomized study in arable fields in Finland (Huusela-Veistola 1996) (same study as (Huusela-Veistola 1998)) found that ground beetle (Carabidae) abundance was higher in reduced pesticide compared to conventional pesticide plots. This was true in 1993 and 1994, the opposite trend was seen in 1992. Spring species tended to be more affected by pesticides than autumn species. Overall there was no significant difference in beetle abundance between cultivation treatments: customary (deep ploughing, conventional fertilizer use, no undergrowth) vs integrated (soil treatment with cultivator only, reduced fertilizer use, undersown grass/clover Trifolium spp.). There were six replicate blocks and treatments (in 0.7 ha plots) which were fully randomized within blocks (one treatment combination/plot). Treatments were conventional pesticide applications, reduced pesticides or no pesticides (control) and customary or integrated cultivation. Beetles were sampled with pitfall traps at 12, 66 and 120 m into each crop 8-10 times (one week/sample) between sowing and harvest.


A small, controlled study of three arable fields on two farms near Braunschweig, Germany (Büchs et al. 1997) found that arthropod numbers and species richness tended to increase with a reduction in management intensity, largely a reduction in fertilizer and pesticide inputs. Arthropod abundance, number of spider (Araneae) species, numbers of juvenile spiders, abundance and number of ladybird (Coccinellidae) species increased with a reduction in inputs. Abundance of beneficial species and length of their activity period also tended to increase with decreased fertilizer and pesticides. Specific species differed in their response to treatment and the intensity of effects depended on type of lifecycle. Reduced inputs increased the activity density of wolf spiders (Lycosidae) and decreased the proportion of pioneer species. Spider species with a wide range of ecological living conditions tended to increase with an increase in pesticides. In 1989-1992 four plots within an arable field received different management intensities: no fertilizers/pesticides, extensive or integrated cultivation with medium fertilizer/pesticide inputs and intensive cultivation with high fertilizer/pesticide inputs. In 1992-1995 four plots within an arable field received farming regimes that differed in the input of fertilizers and pesticides (high input, 30-50% reduction, none), crop rotation (three/four course), tillage, weed control (mechanical/chemical), cultivars, drilling technique and catch crops. A long-term set-aside was also sampled. Six to eight emergence traps and pitfall traps sampled arthropods within each treatment. Traps were collected every 2-4 weeks throughout the year. Results for pest species are not included here.


A replicated controlled study of three arable rotation fields on three farms in England (Frampton 1997) found that reduced pesticide inputs tended to result in higher numbers of springtails (Collembola). Numbers of the springtails Entomobrya multifasciata and Lepidocyrtus spp. were significantly greater in reduced pesticide plots compared to conventional plots. In the conventional plots, these species tended to disappear following chlorpyrifos applications in particular and Lepidocyrtus spp. numbers then remained low for five years. Sminthurinus elegans also declined after chlorpyrifos applications, but tended to recover by the following year and have greater numbers in conventional plots. Fields were divided in half with one receiving conventional pesticide applications, and the other reduced pesticides, i.e. lower herbicide and fungicide where possible and no insecticides (1991-1996). All other practices were the same. Springtails were monitored using a D-Vac suction sampler. In each plot, four samples were taken, each comprising five sub-samples (total area 0.46 m²) between 25-50, 50-75, 75-100 and 100-125 m from the shared field margin. This study was part of the same project (SCARAB – Seeking Confirmation About Results At Boxworth) as (Frampton et al. 1994, Tarrant et al. 1997).



A replicated, controlled study in summer 1991-1994 on up to 13 farms and two experimental sites in the Province of Bayern, Germany (Hilbig 1997) found that the two management types with restricted pesticide use (organic farming and controlled contract production, ‘KVA’) had a more positive effect on plant species richness (average ranges for the sites: organic: 18.4-22.6 species, KVA: 16.9-19.0 spp., controls: 12.4 to 14.6 spp.) than the Bavarian culture landscape programme or control farms (15.6 and 13.8 spp. respectively). Farms under organic or controlled contract production both had restrictions concerning pesticide use. In the Bavarian culture landscape programme, no such restrictions existed but some less common crops (e.g. flax and grass seeds) can be included in the crop rotation. Vegetation was surveyed between June and September on total areas between 100 and 400 m2. Cereal crops were surveyed yearly, cut set-asides several times a year. Note that no statistical analyses were performed on these data.



A replicated, controlled study of arable fields at three sites within the TALISMAN MAFF-funded experiment in England (Jones et al. 1997) found that seed bank density and weed density were higher with reduced (50%) herbicide applications. At High Mowthorpe, plots with reduced (50%) herbicide had significantly higher seed densities (3,181-16,231/m²) than those with conventional applications (1,764-11,300/m²). At Boxworth, the same was true for spring-cropped plots (25,824 vs 8,780/m²). At Boxworth, broadleaved plant seed weights were significantly higher with reduced compared to conventional herbicides (35-151 vs 24-91 mg/m²), treatments did not differ at High Mowthorpe. Plant density tended to be higher on plots with reduced herbicides (4-18/m²) compared to conventional herbicide applications (3-16/m²). At Boxworth, only broadleaved plant species/groups differed between treatments, whereas at Drayton higher weed numbers were consistently found on reduced herbicide plots. At Boxworth there were two replicates in two blocks, at the other two sites, there was one replicate in three blocks. Conventional fertilizer, fungicide and insecticide levels were applied. Seed banks were sampled at Boxworth and High Mowthorpe after harvest from three sub-samples (60 combined soil cores) in each plot. Weed density was sampled in 15 quadrats/plot at the three sites after harvest (August-September) and in October-November.



A replicated, controlled, randomized study on a low productive grassland and high productive fallow arable field in the Netherlands (Kleijn & Snoeijing 1997) found that decreased fertilizer applications resulted in an increase in the number of plant species, minimal effects of herbicide applications were found on fallow land. There were significantly more plant species in the plots receiving no fertilizer (grassland: 16 species/m², fallow: 23-27/m²) compared to those with 25% (grass: 14/m², fallow: 19-23/m²) and 50% (grass: 13/m², fallow: 19-23/m²) of conventional fertilizer applications. In the grassland there was no significant effect of herbicide, whereas in the fallow land there was an effect in the final assessment when 0 and 5% of conventional herbicide application plots had significantly greater species diversity than 50% of conventional herbicide application plots (24-25 vs 22/m²). The most species-rich plots were the 0% herbicide-0% fertilizer plots (grass: 15/m², fallow: 31/m²) and the 10%-0% plots (grass: 15/m²); 50-50% fertilizer plots had the least species (grass: 10/m², fallow: 20/m²). Forty-eight plots (2 x 2 m) were established on a low productivity grassland and an adjacent fallow field sown with 30 broadleaved grassland species. Fertilizer treatments were: 0, 25 and 50% of the conventional application (110 kg N/ha/year). Herbicide (fluroxypyr) treatments were: 0, 5, 10 and 50% of the standard agricultural dose (200 g/ha). Vegetation composition was assessed in April-May (grass and fallow) and September (fallow only) 1993-1996.



A replicated study of arable fields on three farms in England (Tarrant et al. 1997) found that overall earthworm (Lumbricidae) populations did not differ significantly under conventional and reduced pesticide inputs. The only significant difference between treatments was found in autumn 1993 when earthworm density was higher in reduced pesticide treatments (35-50% of normal application) than controls at two of the farms (Warwickshire: 1,529 vs 1,149/m², North Yorkshire: 409 vs 346), the reverse was true at the third farm (Nottinghamshire: 35 vs 45). Differences in earthworm densities were much greater between farms than between fields within farms. Species and age composition differed between farms but the treatment effect was not consistent between fields, even within the same farm. Seven fields over three arable farms were split in two, one half received a conventional pesticide regime and the other a reduced (35-50%) input and no insecticides (1991-1993). Earthworms were sampled in spring and autumn (1993-1994) from three 50 x 50 cm quadrats/plot by hand-digging and using 0.2% formalin solution (20 min period). This study was part of the same project (SCARAB – Seeking Confirmation About Results At Boxworth) as (Frampton et al. 1994, Frampton 1997).



A site comparison before-and-after study from 1989 to 1994 in Sussex, England (Aebischer & Potts 1998) found that survival rates of grey partridge Perdix perdix chicks were significantly higher on 21 km2 of arable farmland that received irregular insecticide applications, compared to a 7 km2 farm with insecticide applications four times a year (average 34% survival on low application farms vs 22% on high application farm). Before the start of intensive insecticide application (1970-1988), survival on the farm had been similar to, or higher than, that on the surrounding farms (27% survival on low application farms vs 30% on intensive application farm). Chick survival rates (up to the age of approximately six weeks) were calculated each year and compared between areas with intensive and irregular insecticide applications. A long-term data set (1970-1988) collected prior to this study was used to investigate chick survival prior to insecticide application on the intensive application farm.



A replicated controlled study of former arable fields at six sites in Sweden (Hansson & Fogelfors 1998) found that after 10 years, there were twice as many plant species in unfertilized compared to fertilized set-aside (30 species in the least fertile site, 10 in the most fertile). Cutting and planting a cover crop also had a positive effect on the number of plant species. At each site, two plots (10 x 20 m) were sown with a grass cover crop and two were left bare. Each year, one of each pair had fertilizer added (equivalent to 150 kg N/ha) and half of every plot was cut and cuttings removed (late July). Vegetation cover was assessed in the centre of each plot (8 x 1 m²) in 1975-1986.



A replicated, controlled, randomized study in arable fields in Finland (Huusela-Veistola 1998) (same study as (Huusela-Veistola 1996)), found that spider (Araneae) abundance was greater in reduced pesticide compared to conventional plots. This was the case in 1992 and 1994 (reduced fertilizer: peak 17-31/three traps/week, conventional: 12-23), there was no significant difference in 1993. Conventional pesticide use decreased money spider (Linyphiidae) numbers in all years (peak: 9-12 vs 10-20/three traps/week), but wolf spider (Lycosidae) catches only in 1994. Only one of the species tested (Erigone atra, money spider family) differed significantly between pesticide regimes. There was no significant difference in spider abundance between cultivation treatments: customary (deep ploughing, conventional fertilizer use, no undergrowth) vs integrated (soil treatment with cultivator only, reduced fertilizer use, undersowing with grass/clover Trifolium spp.). There were six replicate blocks and the treatments (in 0.7 ha plots) were fully randomized within blocks (one treatment combination/plot). Treatments were conventional pesticide applications or reduced pesticides and customary or integrated cultivation. Spiders were sampled with pitfall traps at 12, 66 and 120 m into each crop 8-10 times (one week/sample) between sowing and harvest.



A controlled trial of different farming systems at Reinshof experimental farm, Lower Saxony, Germany (Krooss & Schaefer 1998) found that the highest number of rove beetles (Staphylinidae) was caught under conventional farming with reduced inputs (50% reduction in nitrogen fertilizer, no insecticide, although herbicides were used) (7,897 beetles in total, compared to 6,581 in the control plot with conventional farming). The reduced input field did not have more rove beetle species than conventional farming (39 and 42 species respectively). Extensive farming with no nitrogen fertilizer, herbicides or insecticides had the lowest number of rove beetles (5,038 beetles, from 40 species). Rove beetles were monitored with pitfall and/or emergence traps throughout the year. The experiment was run from 1990 to 1994. There were three or four fields under each treatment, representing the full crop rotation. Monitoring was only in the wheat field from each system, each year. The study also included integrated farming (30% of the nitrogen fertilizer used in conventional system and 50% of the pesticides/herbicides, along with other measures) and extensive farming. The authors suggest that integrated farming without fertilizer does not create a favourable environment for beetles because plant growth is sparse. This study was part of the same project (INTEX – Integrated Farming and Extensification of Agriculture) and was carried out in partly the same research site as (Hasken & Poehling 1995, Schmidt et al. 1995).



A review of literature (Kromp 1999) found evidence that decreases in ground beetle fauna (numbers of species and individuals) caused by intensive agriculture can be reversed by reducing pesticide and fertilizer use (three European studies, including Büchs et al. 1997). Different species responded differently.


A study of spiders (Araneae) in an apple orchard in the Czech Republic (Pekar 1999) found that an integrated pest management strategy resulted in higher spider diversity than conventional pesticide applications. The number of spider species was highest on the plot with reduced fungicide and no herbicide applications and mixed planting (49 species, 1,212 spiders), followed by reduced fungicide and no herbicide applications and sown grass (45 species, 1,497 spiders), conventional spraying resulted in the lowest number of species (39 species, 1,252 spiders). Conventional applications caused much greater fluctuations in late summer spider populations and had lower numbers of spiders after winter (4/plot) than plots under integrated pest management (9-10/plot). Half of the 2 ha orchard received normal applications of fungicides and herbicides, the other half received less frequent applications of fungicides and no herbicides (integrated pest management). Half of the latter was sown with buckwheat Fagopyrum esculentum, common millet Panicum miliaceum, dill Anethum graveolens, and horse bean Faba vulgaris in 1992-1993 and coriander Coriandrum sativum in 1994-1995. The other half was sown with red fescue Festuca rubra. Spiders were sampled by tapping single branches (25 trees) over a 0.25 m² cloth and sweeping ground cover with a 0.25 m² net at weekly intervals (April-October 1992-1995). Cardboard traps (30 x 100 cm²) were also attached to 10 tree trunks in each plot overwinter at a height of 50 cm.



A 2000 literature review (Holland & Luff 2000) looked at which agricultural practices can be altered to benefit ground beetles (Carabidae). It found four European studies that examined the effect of reduced pesticide use on ground beetles. One, the UK SCARAB project (Frampton et al. 1994), found no long-term effect. The other three (Huusela-Veistola 1996, Büchs et al. 1997, Holland et al. 1998) found mixed effects.

Additional reference:

Holland J.M., Cook S.K., Drysdale A., Hewitt M.V., Spink J. & Turley D. (1998) The impact on non-target arthropods of integrated compared to conventional farming: results from the LINK Integrated Farming Systems project. 1998 Brighton Crop Protection Conference – Pests and Disease 2, 625-630.


A small replicated trial in 1997 at an experimental farm in Normandy, France (Cortet et al. 2002) (same study as (Chabert & Beaufreton 2005)) found that the biodiversity of small arthropods (mites (Acari), springtails (Collembola) and others) was not consistently higher on arable land that had reduced insecticide and fungicide use compared to conventionally managed arable land. Half of each field was managed under integrated farming techniques, with reduced pesticide use on average over five cropping years in the previous eight. The comparison was replicated on three fields. In two, the integrated management also involved no deep ploughing. Here, the difference was more consistent (significantly higher biodiversity under integrated management in five out of six monitoring months). Monitoring was between January and June 1997. The authors concluded that tillage had more influence on small soil arthropods than reduced pesticide use.


A before-and-after study in an arable field in England (Frampton 2002) found that abundance and diversity of springtails (Collembola) was significantly lower under conventional pesticide applications than reduced applications (no insecticides, minimal herbicides and fungicides). The springtail Entomobrya nicoleti disappeared from the plot with conventional pesticide application during the first year and did not recover during the three year study. There was no evidence of an effect on populations of the springtail E. nicoleti at the field edge. Lepidocyrtus spp also declined with the conventional spraying regime in the field but not at the field edge. Orchesella cincta and Tomocerus spp were found only in field edge samples. Half of a field (grass and winter wheat rotation) received conventional pesticide applications, and the other half received reduced input, insecticides were excluded from a 6 m headland around the crop (1991-1996). Treatments were reversed 1996-1999. Arthropods were monitored on three occasions/year using suction sampling (25-125 m each side of a hedgerow) and pitfall traps 75 m from hedgerow and at the field edge adjacent to a ditch beside the hedgerow.



A small replicated controlled study from May-June 1992-1998 in Leicestershire, UK (Stoate 2002) found that the abundance of nationally declining songbirds and species of conservation concern significantly increased on a 3 km2 site where pesticide use was restricted (alongside several other interventions), although there was no overall difference in bird abundance, species richness or diversity between the experimental and three control sites. Numbers of nationally declining species rose by 102% (except for Eurasian skylark Alauda arvensis and yellowhammer Emberiza citrinella). Nationally stable species rose (insignificantly) by 47% (eight species increased, four decreased).



A small replicated trial at an experimental farm in Normandy, France (Chabert & Beaufreton 2005) (same study as (Cortet et al. 2002)) found more spiders (Araneae) and ground beetles (Carabidae), but fewer rove beetles (Staphylinidae) in arable fields managed with limited use of herbicides and fungicides, and no insecticides, than in control conventionally managed fields. The experimental plots were also managed without deep ploughing, so it is difficult to separate the effects of ploughing from the effects of reducing pesticide use. However, both ground beetles and spiders were also more abundant in subplots that restricted pesticide use entirely (no fungicides) and restricted herbicide use even more, whereas rove beetles were not. The authors suggested that spiders and ground beetles were sensitive to both pesticide application and ploughing, with spiders being the most sensitive, while rove beetles are less sensitive to pesticide application and prefer deep-ploughed fields. Management was over eleven years from 1990 to 2001. Insects and spiders were monitored in May and June from 1999 to 2001.


A replicated paired sites comparison study in 2000 on 28 arable farms in County Wexford, Ireland (Feehan et al. 2005) found that wider uncultivated margins (average 181 cm-wide) with reduced agrochemical inputs (fertilizer, herbicide and pesticide) on Rural Environment Protection Scheme farms did not have higher plant or ground beetle (Carabidae) diversity or abundance than margins on non-Rural Environment Protection Scheme farms (average 145 cm). There were around 11 plant species and 21-22 ground beetle species/margin on both types of farm. Fourteen farms with REPS agreements at least four years old were paired with fourteen similar farms without agreements. On each farm, two randomly selected field margins were surveyed for plants and ground beetles. In each margin, all plant species were recorded in two 5 x 3 m quadrats, and percentage cover estimated in a 1 x 3 m quadrat. Ground beetles were sampled in four pitfall traps/field margin (8 traps/farm), set at 10 m intervals in early June and late August.



A replicated study in 1999 and 2003 on farms in East Anglia and the West Midlands, UK (Stevens & Bradbury 2006) found that five of 12 farmland bird species analysed were positively associated with a general reduction in herbicide use and conservation headlands. The study did not distinguish between conservation headlands and a general reduction in herbicide use, classing both as interventions reducing pesticide use. The five species positively associated with reducing pesticide use were corn bunting Miliaria calandra (a field-nesting species), chaffinch Fringilla coelebs, greenfinch Carduelis chloris, whitethroat Sylvia communis, and yellowhammer Emberiza citrinella (all boundary-nesting species). A total of 256 arable and pastoral fields across 84 farms were surveyed.



A replicated, controlled study in 2000-2001 on cereal fields of three different farms in western Germany (Wehke et al. 2006) found that both plant species richness and vegetation cover was higher in plots not sprayed with herbicide (spray windows) than in the sprayed part of the field centre. The increase in species richness in spray windows was similar for all five different plant categories considered. Whereas vegetation cover of herbs increased from 10% (field centre) to 50% (spray windows), no such increase was observed for grass cover. Note that no statistical analyses were performed on these data. Spray windows were created as unsprayed plots in the centre of arable fields on one integrated and two conventionally managed farms. Plant species richness and vegetation cover were recorded in both ‘spray windows’ and the sprayed part of the field. Plants were categorized as belonging to five different groups: Red-listed species, declining species, unthreatened arable weeds, arable ruderal species and non-arable ruderal species. Information about crop rotation and herbicide application was obtained directly from the farmers.



A randomized, replicated, controlled trial from 2003 to 2006 on four farms in southwest England (Defra 2007) found that no more foraging birds were attracted to twelve 50 x 10 m plots of permanent pasture with no fertilizer, compared to 12 control (conventionally managed) plots. Experimental plots were managed in the same way as control plots except for the lack of fertilizer, and all plots were cut twice in May and July, and grazed in autumn/winter. There were twelve replicates of each management type, monitored over four years.


A systematic review of 23 studies (Frampton & Dorne 2007) found that restricting herbicide inputs to crop edges tended to increase arthropod abundance. Studies mainly excluded or selectively used herbicides; studies excluding fungicides or insecticides separately were not available. Studies focused on ground beetles (Carabidae), true bugs (Heteroptera), rove beetles (Staphylinidae), butterflies (Lepidoptera) and grouped bird ‘chick-food’ insects. Abundance of true bugs was up to almost 13 times higher where herbicide use was restricted or where herbicides and fungicides or insecticides were restricted. For other invertebrates, restricted use generally increased abundance or had no impact. Only two species exhibited a significant decrease in abundance. In most (20 out of 23) studies, the possibility of confounding outcomes due to pesticide and fertilizer inputs could not be discounted.


A replicated, controlled, randomized study from 2003 to 2005 of arable fields at three sites in the UK (Jones & Smith 2007) found that reduced frequency applications of herbicide resulted in higher species richness and abundance of beneficial weeds and tended to increase arthropod abundance (but not always). Plant species richness and cover of beneficial weeds tended to be highest in untreated and single spring or post-emergence application plots and lowest in those with three applications. In 2004 the inclusion of a pre-emergence herbicide reduced cover of beneficial weeds compared to other treatments. Cover of undesirable weeds was higher in single pre-emergence or spring applications than combined treatments. A post-emergence application was as effective at controlling undesirable weeds as sequences of herbicides. Untreated plots tended to support more arthropods than those with herbicides, but not always. Single applications tended to reduce arthropod abundance less than sequences of herbicides, although post-emergence and pre-emergence applications were detrimental to some taxa. There were three or five replicate plots (3 or 4 x 24 m) of each treatment per site: untreated, pre-emergence, post-emergence or March applications or combinations of two/all herbicide applications. Vegetation was sampled in five quadrats (0.25 m²) in each plot (June 2003-2005). Arthropods were sampled using a D-Vac suction sampler (five sub-samples of 10s/plot) in a sub-set of treatments (June).



A replicated trial in 2004-2006 in Cheshire, Staffordshire and north Shropshire, England (Mortimer et al. 2007) found no differences in plant, insect or bird numbers between conventional and minimum input barley fields. Sixteen trial fields were sown with spring barley each on a separate dairy or mixed farm. One half of each barley field was managed conventionally, the other half managed with minimum pesticide inputs (no insecticide after 15 March, no broadleaved herbicide after 31 March, limited grass-specific herbicide). Plants, invertebrates and birds were monitored on each field, in summer 2005 and winter 2005-2006.



A replicated site comparison on 186 overwinter stubble fields in Devon, England (Bradbury et al. 2008) found that cirl bunting Emberiza cirlus, foraged at significantly higher densities on stubble fields under a ‘Special Project’ agri-environment option, compared to stubbles under standard agri-environment schemes with a conventional pesticide regime (approximately 0.45 birds/ha for 102 special project stubble fields vs 0.05 birds/ha for 52 conventional wheat stubbles and 0.15 birds/ha for 32 conventional barley stubbles). The special project stubbles were also preferentially selected to some extent by four other species of songbird. The special project was designed to encourage cirl buntings and allowed the use of fungicides, growth regulators and specified herbicides to control grass weeds, but prohibited the use of insecticides and herbicides to control broadleaved weeds.



A replicated, controlled, randomized study of undersown and conventional cereal systems in Denmark (Gravesen 2008) found that money spider (Linyphiidae) web density increased with reduction in fertilizer; the same was true for springtail (Collembola) density in conventional but not undersown crops. Money spider web density tended to be higher in undersown crops with no fertilizer (peak 250-300/m²) than low fertilizer input (200-250/m²) and in conventional crops with low fertilizer input (150-200/m²) than high-input (100–150/m²). Springtail density was significantly higher in the fertilized (2350/m²) than unfertilized undersown crops (1600/m²), but higher in the low-input (1250/m²) compared to high-input conventional crops (300/m²). Sixteen experimental plots (12 x 50 m) were established in a randomized block design. Treatments were: wheat with clover Trifolium spp. undersown, with or without nitrogen fertilization (50 kg/ha), or conventional wheat with low (50 kg/ha) or high nitrogen fertilization (160 kg/ha), only the latter received pesticide applications. Money spider web densities, vegetation density (lower layer only, i.e. clover and weed layer) were sampled between May-October 1995-1997. Money spiders and springtails were sampled in 1996.



A paired before-and-after trial in summer 2004-2006 in one arable field in central Germany (Schumacher & Freier 2008) found higher numbers of aphids (Aphidoidea) and their arthropod predators (‘predator units’) in the half field with reduced pesticide treatment than in the control (normal pesticide application) part of the same field after insecticide treatment. No clear effect of reduced pesticide use could be found on ground beetles (Carabidae) as contradicting results were found in all three years. Weed cover was very low in all years and sites (often <1% after herbicide treatment), but significantly more plants were found in the low intensity part of the field in the third study year. Pesticide use on one half of a conventionally managed arable field was reduced to 50%, whereas the other half with 100% pesticide input was used as a control. Aphids, their predators and arable weeds were monitored before and after each pesticide treatment at five points along a line perpendicular to the field edge. Ground beetles were caught weekly in six pitfall traps in each site in June and July. Plants were recorded as plants present/m2 before treatments and as plant cover after treatments. This study is also described in an additional publication (Schumacher & Freier 2006).

Additional reference:

Schumacher K. & Freier B. (2006) Impact of low-input plant protection on functional biodiversity in wheat and pea. Bulletin OILB SROP, 29, 121-124.


A site comparison study in 1998 and 2003 of ten 1.1 km² plots in Austria (Wrbka et al. 2008) showed that grasslands managed for extensive mixed agriculture or intensive livestock farming contained a greater number of plant species when the use of pesticides and fertilizers was reduced. On arable farmland, reducing pesticide use had no effect on the number of plant species present, except for on mixed extensive arable land where fields with no agro-chemicals applied during critical periods had significantly more plant species than traditionally managed fields. For areas of mixed arable farmland in mountainous areas, fields without any agro-chemicals had a greater number of plant species than fields where the use of agro-chemicals was merely reduced. The number of broadleaved plant species in each plot was determined according to the relevés method of sampling vegetation during field surveys in April-September of 1998 and 2003.



A replicated, controlled, randomized study of arable fields over two years in England (Eyre et al. 2009) found that crop protection measures (normal or no pesticide applications) had less impact on insect and spider (Araneae) abundance than type of fertilizer. Wolf spider (Lycosidae), ground beetle (Carabidae), ladybird (Coccinellidae) and true bug (Hemiptera) abundance tended to be higher in plots with organic (compost) compared to inorganic fertilizers and those with no pesticides. In contrast, rove beetles (Staphylinidae), money spiders (Linyphiidae), hoverflies (Syrphidae) and parasitoid wasps (Braconidae) tended to be more abundant in plots with conventional fertilizer and/or pesticide applications. Ground beetles were more abundant in no pesticide vegetable plots in both years, but more abundant in conventionally sprayed bean plots in one year. Effects depended on crop type (grass/clover Trifolium spp., cereals, vegetables) and year. There was no effect of treatments on net-winged flies (Neuroptera) and Proctotrupoidea (parasitoids). In both years the organic fertilizer and conventional pesticide combination had the greatest effect on invertebrates; the organic fertilizers and no pesticide combination also had a significant effect. A field was divided into four blocks (122 x 122 m), each with 32 plots (24 x 12 m). Treatments were: conventional or organic (no) pesticide applications, and conventional (inorganic) or organic (none or compost) fertilizers. Invertebrates were sampled using five monthly samples from five pitfall traps/plot from May-September and three 1-minute suction samples/plot in the first week of July, August and September 2005 and 2006.



A controlled study in 2000-2005 on 61 ha of farmland in Bedfordshire, England (Henderson et al. 2009) found that both winter and summer densities of most farmland bird species and ground beetles (Carabidae) were higher on areas with no pesticide input, compared to areas with conventional levels of pesticides (higher summer densities with no pesticides for 10 of 14 species, although only Eurasian skylark Alauda arvensis, yellow wagtail Motacilla flava and linnet Carduelis cannabina showed a significant increase; all songbirds and 16 of 19 species recorded in winter were at higher densities on zero-pesticide fields). Skylarks were also significantly higher on areas with no fertilizer inputs, but no other species were affected by fertilizer reduction.



A replicated study of autumn-sown and spring-sown barley on four farms in Scotland (Douglas et al. 2010) found that arthropod abundance was higher with fewer herbicide applications. Peak season (July) counts of total arthropod abundance in autumn and spring-sown barley were significantly higher in fields that received one herbicide application (28/sample) than fields receiving two applications (18-21/sample). This was also the case for many individual orders, particularly beetles (Coleoptera) (spring barley one application: 14/sample, two applications: 12; autumn barley one: 12, two: 4) and spiders (Araneae) (spring barley one application: 1.5, two: 0.75; autumn barley one: 2.5, two: 1.75). A total of five spring and five autumn barley fields were selected from four farms (two of each crop type). No insecticides were applied, but fields received one or two herbicide applications. Arthropods were sampled on five occasions in each field (April–August 2004) using a leaf vacuum (15 cm diameter). Sampling was undertaken at intervals (5 or 30 m) along 2-5 parallel transects (100 m apart) across the width of each field.



A replicated site comparison study from 2004 to 2008 in England (Ewald et al. 2010) found that reduced chemical inputs in combination with overwinter stubbles were associated with smaller grey partridge Perdix perdix brood sizes. However, year-on-year partridge density was positively associated with this combination of interventions. There was no relationship between reduced chemical inputs in combination with overwinter stubbles, and grey partridge overwinter survival or the ratio of young to old birds. Spring and autumn counts of grey partridge were made at 1,031 sites across England as part of the Partridge Count Scheme.



A replicated, controlled study from April-July and November-February in 2004-2006 on 16 livestock farms in the West Midlands, England (Peach et al. 2011) found that there were no differences in bird usage of barley fields between fields sprayed with only a narrow-spectrum herbicide (amidosulfuron, at 25-40 g/ha) and those sprayed with both a narrow- and a broad-spectrum herbicide. Broadleaved plant cover was higher on plots treated with only a narrow-spectrum herbicide, but only in the first year of barley production. Invertebrate biomass did not differ between treatments. Insect-eating songbirds and crows (Corvidae) showed reduced use of broad-spectrum-sprayed fields in summer and late summer respectively, but all other groups used fields at equal rates. Barley fields on the farms were split, with half being used for each treatment. Narrow-spectrum herbicide was applied in April-May and broad spectrum in July. All plots were treated with mineral fertilizer, many received fungicide applications but very few received insecticides.

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. (2019) Farmland Conservation Pages 291-330 in: W.J. Sutherland, L.V. Dicks, N. Ockendon, S.O. Petrovan & R.K. Smith (eds) What Works in Conservation 2019. Open Book Publishers, Cambridge, UK.