Providing evidence to improve practice

Action: Grow plants that compete with damaging weeds

Key messages

Weed weight and cover: Nine studies from Australia, Slovakia, the UK and the USA tested the effects of planting species to compete with weeds. All (including four replicated, randomised, controlled trials) found reduced weed plant weight or ground cover, although two found this only in some years or conditions.
Weed reproduction and survival: Five studies (including three replicated, randomised, controlled trials) also found that competition reduced weed reproduction, survival or both. One of these found an effect only in one year only.

Crops studied were clovers, fescues, ryegrass, other grasses and turnip.

Supporting evidence from individual studies


A randomised, replicated trial in fallow farmland in Virginia, USA (Ang et al. 1994) found that sowing plots with tall fescue Festuca arundinacea and crownvetch Coronilla varia at recommended rates reduced shoot weight of the weed creeping thistle Cirsium arvense by 96%, compared to plots with no competitor plants. Sowing competitor plants at half or double the recommended rate reduced thistle shoot weight by 84-85% and 85-86% respectively. Length and weight of thistle roots followed similar patterns. Average thistle shoot weight increased from the first to the second year of competition (6.7 vs. 44.3 g/plot), but decreased after three years of competition (11.5 g/plot). Plots were 2 x 2 m separated by 1 m, in four replicate blocks. Each block had 12 randomised treatments: 0, 0.5, 1 and 2 times the recommended sowing rate (50 and 20 kg/ha of tall fescue and crownvetch respectively) for one, two or three years. The study was part of a biological control experiment using the thistle-eating green tortoise beetle Cassida rubiginosa, which was maintained at a density of >50 adults/m². Numbers quoted were extracted from figures and converted from logarithms.


A randomised, replicated trial in farmland in Virginia, USA (Ang et al. 1995) found that sowing plots with tall fescue Festuca arundinacea and crownvetch Coronilla varia reduced shoot weight (0.1-2.7 g/plant vs. 0.0-6.0 g/plant in control plots), reproduction (0.1-1.3 vs. 0.0-2.9 plants produced/original plant) and survival (0.0-1.6 vs. 0.8-2.8 plants surviving out of three) of creeping thistle Cirsium arvense in one year out of two. In the second year of the experiment thistle shoot weight was not affected, but plots with competitor plants had lower thistle root weight (0.0-2.6 vs. 0.5-3.3 g/plant) and root length (14.9-57.0 vs. 1.3-45.1 cm/plant). The experiment also found that the thistle-eating green tortoise beetle Cassida rubiginosa reduced thistle biomass and reproduction in both years. Plots were 8 x 10 m in blocks of two, one plot without competitor plants and one plot with tall fescue and crownvetch sown at 50 and 20 kg/ha respectively. Thistles were planted in cages in a 2 x 2 m grid, and four treatments of 0, 5, 10 or 20 green tortoise beetles were applied randomly to three plants within each plot.



A replicated study in 1996 in a greenhouse in Queensland, Australia (Navie et al. 1998) found that under competition from buffelgrass Cenchrus ciliaris, the weed ragweed parthenium Parthenium hysterphorus had reduced average height (29.9 cm vs. 39.8 cm in control plots), weight (1.63 vs. 7.72 g/plant) and reproduction (373 vs. 1880 mature seed heads/plant and 1140 vs. 4970 viable seeds/plant). The experiment also found that the ragweed borer moth Epiblema strenuana reduced ragweed parthenium size and reproduction, and that the moth and buffelgrass competition together had a greater effect on seed head production than each did individually. Ragweed parthenium was planted in 15 pots with buffelgrass and 15 without. Two weeks after sowing, plants were thinned to one ragweed parthenium and three buffelgrass seedlings/pot. Within each set of 15 plants, five received 10 ragweed borer eggs 35 days after germination and five received 10 eggs 53 days after germination. The experiment ran for 120 days.



A replicated, controlled study in grassland in 1996-1997 in Berkshire, UK (Edwards et al. 2000) found that percentage ground cover of the weed creeping thistle Cirsium arvense was reduced by 70-90% by sowing wildflower seeds on ungrazed, ploughed grassland. Sowing wildflower seeds had no effect on creeping thistle cover on undisturbed grassland, or on ploughed grassland that was grazed by rabbits Oryctolagus cuniculus.The results were part of a larger experiment that used five replicated blocks of forty-eight 2 x 2 m plots. Factors in the experiment were grazing (rabbits excluded or not), insecticide (applied or not), slug and snail control (applied or not), wild flower seeds (sown or not) and three disturbance treatments: control, ploughing and rotavating to 25 cm depth and ploughing and rotavating followed by fumigation with methyl bromide for seven days. Wildflower plots were sown with 60 species of wild flower at 1000 seeds/species/m². Rabbits were excluded with 1 m high, 3 cm mesh fencing. Quoted numbers were extracted from figures in the paper.



A randomised, replicated, controlled study in 1999 in a greenhouse in Pennsylvania, USA (Tracy & Sanderson 2004) found that the weed curly dock Rumex crispus did not grow in pots where turnip Brassica rapa was used as a pasture species. Curly dock plant weight was 0 g/m² when grown with turnip compared to 80-89 g/m² with other pasture species and 191 g/m² when grown alone. Curly dock also had a reduced germination rate when grown with turnip (19%) compared to other pasture species (31-38%) or when grown alone (60%). The experiment used 30 litre pots which each received 100 turnip or other pasture species seeds and 100 curly dock seeds. The control treatment was not sown with pasture species. Plants were harvested 65 days after planting and weight of above ground plant matter was measured for each species.



A replicated, randomised, controlled trial in a greenhouse in Tasmania, Australia (Davies et al. 2005) found that ryegrass Lolium perenne competition reduced average shoot weight of invasive gorse Ulex europaeus by 96%. Ryegrass competition used together with a biological control agent (gorse thrips Sericothrips staphylinus) or with simulated grazing also increased gorse seedling mortality by 23 and 33% respectively, and by 93% when all three were combined. However, gorse seedling mortality was not affected by ryegrass, thrips or simulated grazing alone. Gorse seedlings were grown in boxes of six in a greenhouse at 20°C. Treatments were 1.5 g/m² ryegrass seeds, 10 thrips/plant and simulated grazing by cutting with scissors to 3 cm height, plus all possible combinations of these three. Each treatment was replicated five times. Seedling mortality and shoot weight after 123 days were recorded.



A randomised, replicated, controlled trial in 2005 in a greenhouse in Colorado, USA (Norton et al. 2008) found reduced growth of diffuse knapweed Centaurea diffusa (an invasive weed) when grown in competition with prairie sagewort Artemisia frigida (diffuse knapweed weight of 1.5 g/plant) or blue grama grass Bouteloua gracilis (0.5 g/plant), compared to growing diffuse knapweed alone (2.2 g/plant). Diffuse knapweed also reduced yield of prairie sagewort by 58% and of blue grama by 35% compared to growing either species alone. The experiment used 2 litre pots with one diffuse knapweed plant and two prairie sagewort or blue grama plants, and controls with each species individually. Pots containing diffuse knapweed also received one of four different treatments with herbivorous insects used for biological control. Each treatment was replicated 12 times.



A randomised, replicated, controlled study in a glasshouse (Fauzi 2009) found that the weed ragweed parthenium Parthenium hysterophorus had 12% lower plant height, 20% lower plant weight and 22% lower seed production when grown in pots containing one buffelgrass Cenchrus ciliaris plant, compared to when grown without competition. Plants were grown from seed in trays, and transplanted into 20 cm diameter plastic pots after 14 days. Pots were kept in a naturally lit greenhouse with 13 hours of daylight, at 26-30°C. Competition pots had one ragweed parthenium and one buffelgrass plant, while control pots had only one ragweed parthenium. Each treatment was replicated six times, as part of a larger experiment on biological control. Plant height and weight was measured after 16 weeks. Study location and date are not given.



A controlled study in 2004-2008 in pasture land in the Strážov Hills, Slovakia (Vozár et al. 2009) found that the proportion of pasture covered by stinging nettles Urtica dioica was reduced by 91% after one season of cutting and reseeding with cock's foot Dactylis glomerata and white clover Trifolium repens. Cutting without reseeding reduced stinging nettle cover by 70% if cuttings were left as a mulch, or by 51% if cuttings were removed. Stinging nettles increased by 11% in an uncut and unseeded treatment. By the end of year five of the study, stinging nettles were rare in the cut and reseeded treatment, covered 1-2% of the pasture in the cut, unseeded treatments and covered 93% of the pasture in the uncut, unseeded treatment. Grass cover in the cut and reseeded plots was 84% by the end of the second year and remained between 68-92% for the rest of the experiment, compared to 3.7-43% in the cut, unseeded treatments and 0.3-6% in the uncut, unseeded treatment. Cut plots were cut every fifth week, starting when the ground cover was 250-300 mm high. The paper gives no further details of the study setup.


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