Providing evidence to improve practice

Action: Incorporate plant remains into the soil that produce weed-controlling chemicals Natural Pest Control

Key messages

Weeds: Six studies (including six randomised, replicated, controlled tests) from Asia, Europe and North America examined the effect of allelopathic plant remains on weeds by comparing amended soils with weeded controls. Three studies found a reduction in weed growth, and three found effects varied between years, weed groups, or the type of weeding method in controls.
Four studies from Asia  and North America examined the effect on weeds by comparing amended soils with unweeded controls. Two studies found a reduction in weed growth, but one found that residues applied too far in advance of crop planting had the reverse effect. Two studies found that effects varied between trials, weed species or the type of residue used.
Two studies, including one randomised, replicated, controlled laboratory study, found that the decrease in weeds did not last beyond a few days or weeks after residue incorporation.
Pests: One randomised, replicated, controlled study in the Philippines found mixed effects on pests.
Crop growth: Two of three studies found crop growth was inhibited by allelopathic plant remains, but this could be minimised by changing the timing of application. One study found effects varied between years.
Yield: Three randomised, replicated, controlled studies compared yields in amended plots with weeded controls and found positive, negative and mixed effects. Three studies compared amended plots with unweeded controls, two found positive effects on yield and one found mixed effects (depending on the crop).
Profit: One study found that amending soils increased profit compared to unweeded controls, but not compared to weeded controls.


Crops studied were beans, cotton, maize, rice and wheat.

Supporting evidence from individual studies


Two randomised, replicated, controlled trials in 1989-1990 in Maine, USA (Dyck et al. 1995) found that incorporating crimson clover Trifolium incarnatum residues into soil reduced emergence of lambsquarters Chenopodium album and other weeds compared to plots treated with nitrogen fertilizer. Maize Zea mays growth was initially 31% lower in plots with clover residue but returned to fertilized plot levels over the growing season. Lambsquarters growth was significantly reduced in plots of crimson clover compared to fertilized plots, with reductions of 64-81% two weeks after emergence and 37-42% lower at the final sampling date. Less maize drymatter was lost to weeds in the crimson clover treatment than the fertilized treatment (1989: 14 vs. 36%; 1990: 0-2 vs. 19-21%). Maize was grown in 3 x 9.1 m plots, each split to contain maize only or maize with lambsquarters. Other weeds were removed. There were six treatments: crimson clover residue, no fertilizer or residue and four levels of ammonium nitrate fertilizer (45, 90, 135, 180 kg N/ha). Crimson clover was sown at 84 kg/ha in May, then mown and incorporated 10-15 cm-deep on flowering. Maize was sown within 2 days of clover incorporation. A second trial in 1989 tested the effect of crimson clover residue applied to plots of lambsquarters.


A series of replicated, randomised, controlled trials in 1989-1990 in Maine, USA (Dyck & Liebman 1995) (partly the same study as Dyck et al. 1995) found incorporating crimson clover Trifolium incarnatum residue reduced weed biomass and increased maize Zea mays growth in some years but not all. In two of four experiments, the weed lambsquarters Chenopodium album had 36-65% lower biomass in crimson clover plots than in plots receiving oat residue and mineral fertilizer, whilst the other two experiments found no difference between treatments. Number of emerging lambsquarters and other weeds was higher in crimson clover plots in one year out of two. Maize biomass was higher in clover than fertilizer plots in one out of two years, by 13-47% in weed-free plots and 50-131% in weedy plots. All plots received crimson clover or oat residue, planted in summer of the previous year and killed and incorporated into the soil in May. Clover and control plots were unfertilized, while fertilizer plots received ammonium nitrate fertilizer at 45 kg N/ha. Maize and lambsquarters were sown in May or June, together in one experiment and lambsquarters alone in the other.


A randomised, replicated, controlled trial in 1996-1997 in wheat Triticum sp. fields in Punjab province, Pakistan (Cheema & Khaliq 2000) found plots with sorghum Sorghum bicolor stalks incorporated into the soil had significantly fewer weeds (38-51 plants/m², 20-41% weed suppression) than unweeded controls (64 plants), similar numbers to hand-weeded controls (33 plants, 49% suppression) and more weeds than herbicide-treated controls (12 plants, 82% suppression). Wheat grain yield was 6-17% higher in sorghum residue plots than unweeded controls, 10% higher in hand-weeded than unweeded controls, and 22% higher in herbicide-treated than unweeded controls. The net benefits of sorghum residue (15,040-15,770 Rupees) were similar to those of unweeded controls (15,768 Rupees) but lower than hand-weeding (16,480 Rupees) or herbicide application (17,477 Rupees). After harvesting, sorghum was dried, cut into 2 cm pieces and incorporated 3-5 cm deep during seedbed preparation. Wheat was sown on 21 November 1996, at 45 kg/ha. Plots were 1.5 x 7.5 m with four replicates. There were six treatments: unweeded control; 2, 4, 6 t/ha sorghum residue; herbicide treatment: Chlorotololuron + MCPA at 2.5 kg/ha; hand weeding. Weed density and biomass were recorded in two 1 m² quadrats/plot, 60 or 90 days after sowing.


A randomised, replicated, controlled laboratory experiment in 1996 in Maine, USA (Ohno et al. 2000) found that root growth of seedlings of the weed wild mustard Sinapis arvensis was reduced by 20% by extracts of soil containing red clover Trifolium pratense and wheat Triticum aestivum residues incorporated eight days previously, but not at any other time after incorporation. There were two treatments, each replicated four times: incorporated wheat crop stubble residue (approximately 30 kg/ha above ground dry matter biomass); incorporated wheat stubble and red clover Trifolium pratense residue (2,530 kg/ha). Residues were incorporated on 28 May 1996. Beans Phaseolus vulgaris and wild mustard were planted 17 days later. Approximately 25 soil samples/plot were taken 12 days before and 8, 21, 30, 41, 63 and 100 days after residue incorporation. Soil water extracts (5 ml) from the soil samples were applied to 20 pre-germinated wild mustard seedlings in the laboratory which were incubated at 20°C. Rootlet length was measured after 72 hours.


A randomised, replicated, controlled trial in 1987-1988 at two sites in Mindinao, Philippines (MacLean et al. 2003) found that weight of broadleaved weeds was higher in plots of rice Oryza sativa amended with gliricidia Gliricidia sepium (averaging 3.8-51.3 g/m²) than non-amended control plots (1.9-20.5 g/m²) in 1988. No difference was found in 1987. Weight of grass weeds was similar between treatments in 1987 and varied between study sites in 1988. Amended plots had more rice seedling maggot Atherigona oryzae eggs (2.7-15.5 eggs/m crop row) than control plots (0.8-8.8 eggs) at one site in 1987-1988, but numbers were similar between treatments at the second site (0.4-25.8 eggs). White grub (Scarabaeidae) numbers were similar between amended and control treatments except at one site in 1988, when they were more abundant in the amended plots (1.7 vs. 1.0 larvae/5 m crop row). Stem borer (Lepidoptera) damage was greater in amended plots (2.4-12.3 vs. 1.4-5.5 deadhearts/m of row) in one of two sites in each year, but otherwise similar. Rice grain yields were higher in amended (0.79-1.51 t/ha) than control (0.09-0.83 t/ha) plots. Rice was planted between hedgerows at two 0.6 ha sites and amended with gliricidia (cut from hedgerows) or left without amendment. Treatments were replicated six times.


A series of four randomised, replicated trials in 1999-2001 in cut flower production systems in California, USA (Zasada et al. 2003) found incorporating broccoli Brassica oleracea and other brassica plant residues had variable success in controlling weeds. One experiment found no effect on weed survival (redroot pigweed Amaranthus retroflexus, annual bluegrass Poa annua, little mallow Malva parviflora) as broccoli material (covered by tarpaulin) increased from 4.0 to 8.4 t dry matter/ha. One experiment found broccoli residue reduced bindweed Convolvulus arvensis populations compared to controls (approximately 56% reduction), while Brussels sprouts B. oleracea and horseradish Armoracia lapathifolia residues did not. One experiment found that broccoli residues and a tarpaulin reduced the number of common purslane plants compared to other tarpaulin treatments. Addition of a tarpaulin to plots with incorporated broccoli residue generally had no effect. Broccoli plant material was collected after floret harvesting and applied at 2.6-8.4 t/ha, approximately 10-30 cm deep. Plots were left uncovered, or covered with a tarpaulin sheet. There were four replicates. Weed species were counted in 0.25 m² quadrats.


A randomised, replicated, controlled trial in 2008-2009 in Iran (Mafakheri et al. 2010) found incorporating rye Secale cereale plant material into the soil resulted in a significant increase in weed density when material was incorporated 54 days before sowing maize Zea mays, but a significant reduction in weed density when incorporated 12 or 34 days before maize. Plots with material incorporated 54 days before maize showed a 1.1% decrease in maize grain production compared to unweeded controls, while incorporating material nearer the time of maize sowing increased maize grain production (4.2-7.9% increase). However, grain production in weeded controls was 39.6% higher than unweeded controls. Rye was sown as a cover crop in November 2008 at three different seeding rates and cut down 21-28 days before the plant material was incorporated. Controls for testing weed density were not sown with rye. Maize was sown on 12th June with controls divided into weed free and unweeded plots. Treatments were tested in 3 x 4 m plots replicated four times. Weed biomass and density was surveyed in 50 x 50 cm quadrats 4, 6 and 8 weeks after planting. The study does not separate the effects of growing a cover crop and incorporating plant material into the soil.


A greenhouse experiment and a replicated, controlled field trial in 2000-2001 in Cambodia (Pheng et al. 2010) found incorporating rice Oryza sativa crop residue into the soil suppressed weed germination and growth, but also suppressed growth of the following rice crop. Greenhouse pots with amended soil had lower weed germination and establishment than non-amended pots (17-47% vs. 71-75%). In field plots in 2000, rice crop residues reduced the dry weight of barnyardgrass Echinochloa crus-galli by 70-93%, depending on rice variety used. However, the rice crop dry weight was also suppressed by 66-85%. In 2001, a smaller amount of rice crop residue incorporated earlier in the season suppressed barnyardgrass by 21-32%, small umbrella sedge Cyperus difformis 15-23% and water primrose Ludwigia octovalvis 20-32%. Rice dry weight suppression was 1-6%. The field experiment ran in January-March 2000 and 2001. Residues of eight rice varieties were incorporated 0-10 cm deep. In 2000, barnyardgrass or rice was sown one week after 6 kg/plot crop residue was incorporated. In 2001, three weed species and one rice crop were sown two weeks after 4 kg/plot of crop residue. The greenhouse experiment used 16 plant lines and one non-residue control.


A set of three randomised, replicated, controlled field trials in central California, USA (Stapleton et al. 2010) found that incorporating residue of a sorghum-sudangrass hybrid (Sorghum bicolor x S. sudanense ‘sudex’) into the soil reduced weed growth, but that this effect was temporary. In the first experiment, growing and incorporating sudex reduced weed growth by 35% (136 g dry weight weed biomass vs. 208 g in control plots). In the other two experiments, weed growth was reduced by 61-89% compared to control plots 50 days after treatment, but after 57 days in the second experiment and 106 days in the third experiment this difference had disappeared. Sudex was planted in six rows in raised beds and shredded at 1.4-2.0 m tall. Experiment 1 had three treatments with four replicates in 1 m-long plots: sudex grown, shredded and left as a mulch; grown, shredded and incorporated; no sudex grown or residue added. Experiments 2 and 3 had four replicates in 4.5 x 1.5 m plots. Treatments included those from experiment 1, plus two additional treatments: shreddings added to fallow plot where no sudex had been grown; shreddings removed but roots and 3-5 cm stubble left in plots. Weed biomass was calculated by removing material from a 0.093-1 m² area.


A randomised, replicated, controlled trial in 2005-2007 in northern Greece (Vasilakoglou et al. 2011) found that incorporating oregano Origanum vulgare into the soil reduced the abundance of three weed species in cotton Gossypium hirsutum and maize Zea mays. In cotton, green manure reduced numbers of the weed common purslane Portulaca oleracea by 30-55% (55-85 vs. 121 plants/m²), barnyard grass Echinochloa crus-galli by 48-52% (23-25 vs. 48) and bristly foxtail Setaria verticillata by 43-86% (1-4 vs. 7). Maize plots with green manure had 0-45% fewer common purslane (71-128 vs. 129), 38-46% fewer barnyard grass (7-8 vs. 13) and 60-80% fewer bristly foxtail (1-2 vs. 5). The cotton yield was significantly lower in green manure treatments than in a weed free control, but not different to (and in once case higher than) an unweeded control. Maize silage and grain yields were similar between treatments. There were four oregano green manure treatments (plants from four locations, selected for high concentrations of potential allelopathic chemicals) and two controls without green manure (one weeded) replicated four times in 9 x 5 m plots. Oregano was incorporated 8-10 cm deep before flowering. Cotton and corn were planted five days later.

Referenced papers

Please cite as:

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