Action

Sow native grass and forbs

How is the evidence assessed?
  • Effectiveness
    not assessed
  • Certainty
    not assessed
  • Harms
    not assessed

Source countries

Key messages

VEGETATION COMMUNITY (42 STUDIES)

  • Community composition (11 studies): Five of 11 studies (10 of which were replicated and/or controlled, and three of which were site comparisons) in the UK, the Czech Republic, Norway and Germany found that sowing native grass and forb seeds increased the similarity of plant community composition to that of target communities. Three studies found no increase in community similarity to target communities. Two studies found that over time communities became more similar to those of intact grasslands. One study found that over time areas sown with native grass and forb seeds became more similar to areas that were not sown with seeds.
  • Overall richness/diversity (28 studies): Sixteen of 28 studies (24 of which were controlled and four of which were site comparisons) in Europe, North America and New Zealand found that sowing native grass and forb seeds increased overall plant species richness. Seven studies found that there was no change in plant species richness or mixed effects on plant species richness and plant diversity. Three studies found that sowing native grass and forb seeds increased plant species richness during the first year, but after 3–13 years, species richness did not differ between sown and unsown areas or was lower in sown areas. One study found that after one year, sowing did not alter plant species richness but after eight years, species richness was higher than in unsown areas. Three studies found that species richness was lower in sown areas than in nearby intact grasslands.
  • Characteristic plant richness/diversity (9 studies): Six of nine studies (eight of which were replicated and/or controlled, and two of which were site comparisons) in Europe found that sowing native grass and forb seeds increased the species richness of characteristic grassland plants. Two studies found no change in the species richness of characteristic grassland plants. One study found that sowing native grass and forb seeds increased the species richness of target forbs but not target grasses.
  • Sown/planted species richness/diversity (3 studies): Three replicated, paired, controlled studies in the UK and the Czech Republic found that sowing native grass and forb seeds increased sown species richness.
  • Grass richness/diversity (1 study): One replicated, randomized, paired, controlled study in the UK found that sowing native grass and forb seeds increased grass species richness in 54% of cases.
  • Forb richness/diversity (1 study): One replicated, randomized, paired, controlled study in the UK found that sowing native grass and forb seeds increased forb species richness in 71% of cases.
  • Native/non-target richness/diversity (1 study): One replicated, controlled study in the USA found that sowing native grass and forb seeds increased the species richness of native plants.

VEGETATION ABUNDANCE (24 STUDIES)

  • Overall abundance (8 studies): Three of eight replicated, controlled studies (four of which were randomized and paired) in Europe and North America found that sowing native grass and forb seeds increased overall vegetation cover, biomass or density. One study found that sowing native grass and forb seeds increased plant species richness during the first 2–7 years, but after eight years, species richness did not differ between sown and unsown areas. Four studies found that there was no change in overall vegetation abundance in all or most cases.
  • Characteristic plant abundance (5 studies): Three of five replicated studies (four of which were controlled, and one of which was a site comparison) in Europe found that sowing native grass and forb seeds did not alter the cover of characteristic grassland species. The other two studies found an increase in the cover of characteristic or target grassland species.
  • Sown/planted species abundance (6 studies): Five of six studies (four of which were replicated and controlled, and two of which were reviews) in Europe, North America, Africa and New Zealand found that sowing native grass and forb seeds increased the abundance of sown species in all or most cases. The other study found mixed effects on sown species abundance.
  • Grass abundance (3 studies): Two of three replicated, randomized, controlled studies (two of which were paired) in the Czech Republic and the USA found that sowing native grass and forb seeds increased the cover of grass species. The other study found no change in the cover of grass species.
  • Forb abundance (4 studies): Three of four replicated, randomized, controlled studies (three of which were paired) in the Czech Republic and the USA found that sowing native grass and forb seeds increased the cover or density of forb species. The other study found that one year after sowing, the cover of forb species increased, but after 10 years it did not differ between sown and unsown areas.
  • Native/non-target species abundance (1 study): One replicated, controlled study in the USA found that sowing native grass and forb seeds increased the cover of native plant species.
  • Individual plant species abundance (1 study): One replicated, randomized, paired, controlled study in the UK found that sowing native grass and forb seeds did not alter yellow rattle abundance.

VEGETATION STRUCTURE (0 STUDIES)

OTHER (2 STUDIES)

  • Germination/Emergence (2 studies): One of two replicated, controlled studies (one of which was paired) in the USA and Germany found that sowing native grass and forb seeds increased the number of seedlings that emerged. The other study found no change in seedling number.

About key messages

Key messages provide a descriptive index to studies we have found that test this intervention.

Studies are not directly comparable or of equal value. When making decisions based on this evidence, you should consider factors such as study size, study design, reported metrics and relevance of the study to your situation, rather than simply counting the number of studies that support a particular interpretation.

Supporting evidence from individual studies

  1. A site comparison study in 1984–1994 in a former agricultural field in Essex, UK (Gough & Marrs 1990) found that after sowing grass and forb species, plant species richness and diversity were lower than in a nearby old meadow. Four years after sowing, 19 plant species were recorded in the field sown with seeds, while 26 species were present in a nearby old meadow. The same was true 10 years after sowing (sown field: 42 species; old meadow: 57 species). Species diversity was also lower in the field sown with seeds than in the old meadow four and 10 years after sowing (data reported as Shannon-Weiner index). No statistical analysis was done in this study. In 1984, rye grass Lolium perenne and white clover Trifolium repens seeds were sown in a 3-ha field following agricultural abandonment. In June 1988–1994, vegetation cover was assessed using 40 randomly placed 0.5 x 0.5 m quadrats in the sown field and 30 quadrats in a nearby 1-ha meadow (>40 years old) under similar management (cut in July and grazed in August–October).

    Study and other actions tested
  2. A replicated, randomized, paired, controlled study in 1993–1994 in a former arable field in Hampshire, UK (Stevenson et al. 1995) found that sowing grass and forb seeds increased cover and species richness of chalk grassland species, and similarity of the plant community to that of the target community, but reduced cover of weed species. After one year, in three of four comparisons cover of chalk grassland species was higher in areas where seeds had been sown (37–98%) than in areas where seeds had not been sown (8%), while in one comparison there was no significant difference (seeded: 19%, unseeded: 8%). Species richness of chalk grassland plants was higher where seeds had been sown (20–33 species/plot) than where no seeds were sown (7 species/plot). Cover of weed species was lower in areas where seeds were sown (2–12%) than in areas where no seeds were sown (23%). Seeded plots resembled the target community more than plots where no seeds were sown (data presented as graphical analysis). Seeds were collected from two nearby chalk grassland sites using vacuum seed harvesters. In each of four blocks established at the site, four plots were sown with seeds of 47 species at a rate of 0.1 g/m2–4.0 g/m2. Cover of each species was estimated in August 1993 and 1994 using two 1-m2 quadrats/plot.

    Study and other actions tested
  3. A site comparison study in 1992 in a former arable field in Kansas, USA (Kindscher & Tieszen 1998) found that sowing grass and forb seeds resulted in lower plant species richness than that found in intact prairie sites. Plant species richness in the two sites where local seeds were sown was lower (8.7–10.3 species/quadrat) than in nearby intact prairies (14.0 species/quadrat). In April 1989, seeds of 33 prairie plant species were sown. Grasses were sown at a rate of 5.8 kg/ha and forbs at a rate of 0.06 kg/ha. In September 1992, vegetation cover was surveyed in sixty 1-m2 quadrats in the site sown with seed, and 30 quadrats in an adjacent intact prairie.

    Study and other actions tested
  4. A replicated, randomized, paired, controlled study in 1994–1996 at six improved grassland sites in the UK (Hopkins et al. 1999) found that sowing grass and forb seeds had mixed effects on grass and forb species richness. No statistical analyses were carried out in this study. In 13 of 24 comparisons, plots where seeds were sown had more grass species (5–12 species/plot) than plots where no seeds were sown (4–10 species/plot), while in 11 comparisons, grass species richness was lower or equal (seeded: 4–13 species/plot, unseeded: 5–13 species/plot). In 17 of 24 comparisons, forb species richness was higher in plots where seeds were sown (7–25 species/plot) than plots where no seeds were sown (5–15 species/plot) while it was lower or equal in seven of 24 comparisons (seeded: 3–8 species/plot; unseeded: 4–8 species/plot). In 1994, at each site, soil was disturbed and seeds sown in eight 6 x 4 m plots and four plots were left unseeded. Seed mixes contained seeds of five grass species and 18 forb species. In May/June of 1995 and 1996, three 40 x 40 cm quadrats were placed in each plot and the frequency of each species recorded.

    Study and other actions tested
  5. A replicated, randomized, controlled study in 1995–1997 in a pasture in Iowa, USA (Jackson 1999) found that sowing grass and forb seeds increased plant species richness. The results of this study are not based on statistical analysis. Plant species richness was higher in areas where seeds were sown (8.6–9.4 species/plot) than in areas where no seeds were sown (3.5–5.0 species/plot). In May 1995, in order to remove any vegetation present, glyphosate herbicide was applied to four plots, each of which had an area of at least 500 m2. In June 1995, seeds of three native grass species and one native legume were sown by hand or using a seed drill in these four plots, while four other plots were not seeded. Cover of all plant species was monitored in five to eight 0.5-m2 quadrats in each plot from 1995–1997.

    Study and other actions tested
  6. A replicated, randomized, paired, controlled study in 1993–1996 in ex-arable land in the UK (Manchester et al. 1999) found that sowing grass and forb seeds increased plant species richness. No statistical analyses were carried out in this study. Plant species richness was higher in areas where seeds were sown (36–45 species) than in areas where no seeds were sown (28 species). Seeds of grass and forb species were collected from hay harvested in a nearby intact wet meadow and sown in 10 plots (size of plots unclear), commercial seed mixes containing 11–23 grass and forb species were sown in 20 plots, while no seeds were sown in 10 plots. The survey methods used to assess species richness in this study were not clear.

    Study and other actions tested
  7. A replicated, controlled study in 1996–1998 in an area that was previously burned in Galicia, Spain (Pinaya et al. 2000) found that sowing grass and forb seeds increased vegetation cover. No statistical analysis was done in this study. After six to 16 months, vegetation cover in areas sown with seeds was higher (54–95%) than in areas where no seeds were sown (19–76%). In July 1996, a 2,000-m2 area was burned using prescribed burning methods. Following this, in September/October 1996, four plots were sown with seeds of Agrostis truncatula, A. capillaris, bird's-foot trefoil Lotus corniculatus and Lolium multiflorum, and two plots were not sown with seeds. Vegetation cover was estimated after six, 10, 16, and 19 months using a 50 x 50 cm quadrat placed in each plot.

    Study and other actions tested
  8. A replicated, randomized, paired, controlled study in 1990–1998 in improved grassland in North Yorkshire, UK (Smith et al. 2000 - same experimental set up as Smith et al. 2008) found that plots sowed with grass and forb species had more plant species than unseeded plots, but that there was no difference in yellow rattle Rhinathus minor abundance. Eight years after sowing, there were more species in sown plots (17.4 species) than unsown plots (15.6 species). However, the density of yellow rattle plants did not differ significantly between sown (0–52 plants/m2) and unsown plots (0–51 plants/m2), and there was no effect of sowing on annual hay yield (data not given). In 1990, one hundred and eight 6 × 6 m plots were established in three blocks of 36 plots. Fifty-four plots (18 random plots/block) were sown with 7 kg/ha of locally collected seeds and 0.05–1.5 kg/ha of commercial seeds (containing 19 species) each autumn from 1990–1992. Fifty-four plots were left unsown. All plots were cut annually between June and September, and parts of each plot were also grazed, fertilized and/or mowed. Plant species and cover were recorded in two 2 x 2 m quadrats in each plot in summer 1994, 1996 and 1998.

    Study and other actions tested
  9. A replicated, randomized, paired, controlled study in 1995–1998 in three degraded riparian meadow sites in Nevada, USA (Martin & Chambers 2001) found that sowing native grass and forb seeds did not lead to an increase in overall plant biomass and the establishment of sown species was low. After three years, average plant biomass did not differ significantly between plots sown with native grass and forb seeds (555–825 g/m2) and unseeded plots (580–610 g/m2). In addition, only one of the six sown species (Nebraska sedge Carex nebrascensis) was reported to be present in any of the sown plots after three years. In 1995, at each of three sites, one plot measuring 16–25 m2 was sprayed with herbicide (Round-up®), tilled to disturb the soil, and sown at a rate of 875 seeds/m2 with six riparian meadow species. One plot in each site was not sprayed with herbicide, disturbed or sown with seed. In July 1996–1998, five 0.1-m2 quadrats were randomly placed in each plot and vegetation clipped to determine biomass.

    Study and other actions tested
  10. A replicated, randomized, paired, controlled study in 1996–1997 in an experimental meadow in Norway (Losvik & Austad 2002) found that sowing grass and forb seeds increased species richness as well as richness of traditional meadow plants. Plant species richness was higher in areas where seeds were sown (23 species/plot) than in areas where no seeds were sown (19 species/plot). Species richness of plants that are indicators of traditional meadow management showed a similar pattern (seeded: 7 species/plot; not seeded: 3 species/plot). In April 1996, the soil of all plots was disturbed using a power harrow and 20 kg of calcium oxide was added. Seeds from a nearby hay barn were sown in six 16-m2 plots and no seeds were sown in six other plots. All plots were mown in August 1996. In July 1997, three 1 x 1 m quadrats were placed in each plot and the presence of all plant species recorded.

    Study and other actions tested
  11. A replicated, randomized, paired, controlled study in 1994–1998 in five arable fields in the UK (Pywell et al. 2002) found that disturbing soil and sowing grass and forb seeds increased plant species richness in most cases and increased the plant community similarity to that of target habitats. In five of eight comparisons, plant species richness was higher in areas where seeds were sown (16.2–26.4 species/plot) than in areas where seeds were not sown (10.4–18.7 species/plot). However, in three of eight comparisons, plant species richness was lower (seeded: 10.8–14.0 species/plot; unseeded: 10.4–18.7 species/plot). Similarity to target communities was higher in areas that were seeded than in areas that were not seeded (data presented as similarity index). In September 1994, in five sites, four blocks containing five 6 × 4 m plots were established. In each block, the soil of four plots was disturbed using harrows or ploughs and sown with seeds of between six and forty-one plant species. One plot was not disturbed or sown with seeds. Vegetation was cut and removed each year in June or July, and sheep grazed between October and December at a density of 25–40 sheep/ha for six to eight weeks. Vegetation was surveyed in June of each year using three randomly placed 40 x 40 cm quadrats within each plot.

    Study and other actions tested
  12. A before-and-after, site comparison study in 1998–2001 in a former arable field and a semi-natural grassland in northern France (Vécrin et al. 2002) found that sowing grass and forb seeds did not increase plant species richness, and species richness was lower than in areas in a nearby natural grassland. After three years, the plant species richness of areas where seeds were sown did not differ significantly from that of the same areas before sowing (after: 29 species/plot, before: 26 species/plot). Plant species richness after three years was also lower in areas where seeds were sown (29 species/plot) than in nearby semi-natural grasslands (34 species/plot). Before restoration, the field had been abandoned for five years and had been recolonised by some grassland species. In 1998, the field was mowed and ploughed and in autumn was sown with a commercial seed mix containing Phleum pratense, Lolium perenne, Festuca pratensis and Trifolium repens at a rate of 30–35 kg/ha. Before sowing, nine 4-m2 plots were established and vegetation was surveyed. In 1999–2001, twenty-one plots in the field were surveyed along with 30 plots in nearby semi-natural grasslands.

    Study and other actions tested
  13. A replicated, paired, controlled study in 1993–1999 in an ex-arable field near Aberdeen, Scotland, UK (Warren et al. 2002) found that sowing native grass and forb seeds led to an increase in the number and cover of sown species. After six years, sown plots on average contained more sown species (4.9 species/m²) and had a greater cover of sown species (97%) than unsown plots (1.8 species/m²; 43%). In April 1993, twelve pairs of plots (each 20 x 40 m) were ploughed and fenced. In May 1993, one plot in each pair was sown with a native seed mix (four grass and 10 forb species sown at a rate of 20 kg/ha), while the other was left unsown. Both plots in each pair received the same grazing and/or cutting treatment each year (six treatments were applied overall; see original paper for details). In June 1994–1999, vegetation was monitored annually within 20 x 1 m2 quadrats (number of sown species) and 10 x 0.25 m2 quadrats (cover of sown species) randomly placed in each of the 24 plots.

    Study and other actions tested
  14. A replicated, randomized, controlled study in 1997–1999 in a grazed wet grassland in London, UK (Gilbert et al. 2003) found that sowing native grass and forb seeds did not increase similarity of the vegetation community to that of the target community. The similarity of the vegetation community to that of the target community did not differ significantly between areas sown with seeds and areas that were not sown with seeds (data presented as similarity coefficients). In spring 1997, seeds of eight grass and forb species collected from an existing grassland were sown onto bare soil in twenty 1-m2 plots at a rate of 40 kg/ha, and no seeds were sown in 10 other plots. Cover of all species in the plots was estimated in June 1997–1999. Similarity of the plant communities to the target community, a mesotrophic grassland, was assessed using the UK National Vegetation Classification.

    Study and other actions tested
  15. A replicated, randomized, paired, controlled study in 1990–2000 in a grazed meadow in Yorkshire, UK (Smith et al. 2003) found that sowing grass and forb seeds increased plant species richness and similarity of the plant community to the target habitat type. These results are not based on statistical analyses. In 25 of 30 comparisons, plant species richness was higher in areas where seeds were sown (17.8–26.7 species/4 m2) than in areas where no seeds were sown (15.4–20.5 species/4 m2). However, in five comparisons, species richness in sown areas (16.1–18.6 species/4 m2) was lower or equal to that found in unsown areas (17.2–19.6 species/4 m2). In 26 of 30 comparisons, similarity of the plant community was higher in areas where seeds were sown than in unsown areas, but in four comparisons similarity was equal or lower than that in unsown areas (data reported as model coefficient). The experiment consisted of three 36 × 12 m blocks, each containing twelve 6 × 6 m plots. In autumn 1990–1992, in each block, seeds of four species were sown in six plots and no seeds were sown in another six plots. Seeds were also sown in August 1998 and July 2000. From 1998, all plots were mowed annually in July. Starting in 1994, vegetation was surveyed every two years using four 2 × 2 m quadrats in each plot.

    Study and other actions tested
  16. A controlled study from 2001–2002 in a former arable field in northeast France (Vécrin & Muller 2003) found that sowing grass and forb seeds increased the number of meadow plant species. One year after sowing, there were more plant species on average in the area sown with seeds (8 species/quadrat, including 5 meadow species) than in the area that was not sown with seeds (6 species/quadrat, including 3 meadow species). In part of the field, a commercial seed mixture containing cocksfoot Dactylis glomerata, tall fescue Festuca arundinacea and common birdsfoot trefoil Lotus corniculatus was sown in June 2001 at a rate of 30 kg/ha, while in another part of the field no seeds were sown and natural regeneration was allowed. Vegetation was monitored in 50 x 50 cm quadrats at 1 m intervals along 20-m transects in June 2002. Two transects were surveyed in each of the natural regeneration and commercially seeded areas.

    Study and other actions tested
  17. A replicated, randomized, paired, controlled study in 1994–1996 in two ex-arable sites in Scotland, UK (Lawson et al. 2004) found that sowing grass and forb seeds increased the number and cover of sown species and reduced the number and cover of non-sown species. More sown species were present in areas where seeds were sown than areas where no seeds were sown, and the same was true for the cover of sown plant species (no data presented). Similarly, there were fewer non-sown species in areas where seeds were sown when compared to unsown areas, and their cover was also lower (no data presented). Before sowing, sites were ploughed and harrowed. Seeds of 18 species were sown in eight 3 x 9 m plots at a rate of 4 g/m2 and no seeds were sown in four plots at each site. Plant cover and species richness were estimated in June/July 1995 and 1996 using a 1 x 1 m quadrat placed in each plot.

    Study and other actions tested
  18. A replicated, randomized, controlled study in 1998–1999 in grazed grasslands in Manawatu, New Zealand (Edwards et al. 2005) found that sowing grass and forb seeds increased plant species richness and the cover of five of eight sown plant species. No statistical analyses were carried out in this study. After one year, plant species richness in plots where seeds were sown was higher (8–9 species/plot) than in plots where no seeds were sown (5–6 species/plot). After 21 months, the average cover of five of eight sown plant species was higher in plots where seeds were sown than plots where no seeds were sown: spear thistle Cirsium vulgare (sown: 3–9 %, unsown: 0–1%); ribwort plantain Plantago lanceolata (sown: 1–7% , unsown: 0.3–0.5%); bitter dock Rumex obtusifolius (sown: 1–7%, unsown: 0–0.3%); white clover Trifolium repens (sown: 14–21%, unsown: 11–18%); perennial ryegrass Lolium perenne (sown: 74–84%, unsown: 71–80%). The three other species (greater bird’s foot trefoil Lotus uliginosus, Dallis grass Paspalum dilatatum and creeping thistle Cirsium arvense) had few or no seedlings in both sown and unsown plots. In March 1998, seeds of eight plant species were sown (1,000 seeds/species/m2) in 120 randomly located plots (each 2 x 2 m), while in another 120 plots no seeds were sown. Plant cover and species richness was estimated in all plots in December 1999.

    Study and other actions tested
  19. A replicated, randomized, paired, controlled study in 1999–2001 on former logging roads and agricultural areas in British Colombia, Canada (Burton et al. 2006) found that sowing native grass and forb seeds increased the total cover of sown plant species. These results are not based on statistical analyses. During two years after sowing, total cover of sown plant species was higher in areas where seeds were sown (4–62%) than in areas where no seeds were sown (0–10%). In autumn 1999, six blocks were established each containing twelve 2.5 × 2.5 m plots. All plots were cleared of rocks and vegetation and tilled to a depth of 12 cm. Ten plots in each block were sown with seeds of six native species at a rate of 375–6,000 seeds/m2, and two plots were not sown with seeds. In September 2000 and August 2001, vegetation cover was estimated in three randomly placed 0.5 × 0.5 m quadrats/plot.

    Study and other actions tested
  20. A replicated, paired, controlled study in 2003–2004 in a former arable field prairie restoration site in Iowa, USA (Martin & Wilsey 2006) found that sowing with grass and forb seeds increased the number of seedlings and species richness of seedlings. There were more seedlings in areas where seeds were sown (104–156 seedlings/m2) than in areas where no seeds were sown (85–139 seedlings/m2). The same pattern was seen for seedling species richness (seeded: 3.0–5.8 species/0.1 m2; unseeded: 2.5–2.8 species/0.1 m2). In June 2003 and April 2004, forty-eight 1-m2 plots were sown with seeds at a rate of 19,700 seeds/m2 and 24 plots were not sown with seed. Seedling numbers and species richness were estimated once a month during the growing season using a 50 x 20 cm quadrat. Herbivore density in the restoration site was approximately 0.1 bison/ha and 0.05 elk/ha. Prescribed burning was carried out at the site every two years and the site was mowed to control weeds. The plots used in the study were not mowed or burned in 2003 or 2004.

    Study and other actions tested
  21. A replicated, randomized, paired, controlled study in 2000–2005 in an abandoned pasture in Kansas, USA (Foster et al. 2007). After five years, total plant species richness was higher in plots where seeds had been sown (13–15 species/plot) than in plots where no seeds were sown (8–9 species/plot). The same was true for plant diversity (data reported as Shannon diversity index). In January 2000, seeds of 24 native and eight non-native grassland species were sown at a rate of 12,800 seeds/plot in thirty-four 1 x 1 m plots, and no seeds were sown in another 34 plots. Species richness and diversity were assessed from plant biomass harvested from each plot in June and September 2005.

    Study and other actions tested
  22. A replicated, randomized, paired, controlled study in 1999–2004 in semi-dry grassland in the Czech Republic (Jongepierova et al. 2007) found that sowing grass and forb seeds increased the cover of local forb and grass species. Cover of local grass species was higher in areas where seeds were sown (29–53%) than in areas where no seeds were sown (0–2%). The same pattern was seen for local forb species (seeded: 16–32%, unseeded: 0–2%). In 1999, four blocks were established. In each block, one 55 x 20 m plot was sown with a locally sourced seed mixture containing seven grass species and 20 forb species at a rate of 2 g/m2, while one plot was not sown with seeds. In June in 2000–2004, ten 1.5 x 1.5 m quadrats were placed in each plot and all species present and their cover recorded.

    Study and other actions tested
  23. A replicated, randomized, paired, controlled study in 1996–2003 in five ex-arable fields in the Czech Republic, the Netherlands, Spain, Sweden, and the UK (Lepš et al. 2007) found that sowing grass and forb seeds had mixed effects on plant species richness but increased biomass. Initially, in one of two comparisons, sowing seeds increased plant species richness (38 species/plot) compared to areas where no seeds were sown (27 species/plot), but in one of two comparisons there was no significant difference (30 species/plot). However, after eight years, species richness of areas sown with seeds (26–27 species/plot) was not significantly different to that found in areas where no seeds were sown (32 species/plot). After two years, plant biomass was higher in areas where seeds were sown (0.59–0.71 kg/m2) than in unsown areas (0.3 kg/m2), and the same was true after eight years (sown: 0.50–0.64 km/m2; unsown: 0.46 kg/m2). In each site, ten 10 x 10 m plots were sown with a seed mix of either four or 15 grass, legume and other forb species, while five plots were not sown with seeds. Grasses were sown at a rate of 2,500 seeds/m2, and legumes and other forbs at a rate of 500 seeds/m2. All plots were mown at least once a year. In each plot, the vegetation in twelve 1 x 1 m quadrats was surveyed in 1996–1998 and 2002–2003. Biomass was estimated by cutting and drying all vegetation in twelve 0.25 x 0.25 m quadrats placed within the plots.

    Study and other actions tested
  24. A paired, site comparison study in 2004 in 40 restored and 40 ancient calcareous grasslands in southern England, UK (Fagan et al. 2008) found that sowing grass and forb seeds resulted in plant communities similar to ancient grasslands and sites where natural regeneration was allowed. The plant community of restoration sites where grass and forb seeds were sown was similar to the plant communities of paired ancient grasslands and sites where natural regeneration was allowed, but was different from sites where grass seeds alone were sown (results presented as graphical analysis). Between one and >20 years prior to the study, 40 ex-arable sites were seeded with grass and forb seeds (12 sites), grass seeds only (12 sites) or no seeds (natural regeneration; 16 sites). Each site was paired with the closest ancient grassland (0–9 km away; aged >200 years). All sites were grazed, and occasionally mown. In June–August 2004, the cover of plant species was estimated within 50 x 50 cm quadrats placed at 10 m intervals along a 100-m transect at each site.

    Study and other actions tested
  25. A replicated, randomized, paired, controlled study in 2003–2004 in a former hay field in Ohio, USA (Fraser & Madson 2008) found that sowing native grass and forb seeds increased plant species richness. Average plant species richness was higher in plots where seeds were sown (7–9 species/plot) than in plots where seeds were not sown (4–5 species/plot). In March 2003, in each of six blocks, seeds of 20 native, wet meadow grass and forb species were sown at a rate of 150 seeds/species in each of two 1 x 1 m plots, while two other plots were left unseeded. Seeds were collected from nearby fields or from cultivated plants. In August 2004, vegetation in the central 0.25 m2 of each plot was harvested and used to identify plant species.

    Study and other actions tested
  26. A replicated, randomized, paired, controlled study in 1990–2004 in three agriculturally-improved grasslands in North Yorkshire, UK (Smith et al. 2008 – same experimental set up as Smith et al. 2000) found that sowing grass and forb seeds increased the number of plant species as well as the similarity of the plant community to the target community. In the four years following sowing, sown plots had more species (22 species) and higher similarity to the target plant community of upland meadow species (61% similarity) than plots that were not sown with seeds (19 species, 55% similarity). In 1990, thirty-six 6 x 6 m plots were established, with 12 plots in each of three fields. Eighteen plots (six random plots/field) were sown with 15 kg/ha of commercial seed containing bird’s foot trefoil Lotus corniculatus, quaking grass Briza media and bulbous buttercup Ranunculus bulbosus in August 1998, and 0.5 kg/ha wood cranesbill Geranium sylvaticum in September 1999; eighteen plots were left unsown. Seeded plots had previously been sown with local and commercial seeds in 1990–1992. All plots were grazed in spring and autumn, and cut in July. Vegetation cover was assessed in four 4-m2 quadrats in each plot every other summer in 1994–2004. The target plant community was defined as that associated with well-drained permanent upland meadows, characterised by sweet vernal grass.

    Study and other actions tested
  27. A review in 1996–2009 of 17 studies of semi-natural grassland restoration in Europe, North America, and Africa (Hedberg & Kotowski 2010) found that sowing grass and forb seeds had mixed effects on the introduction of sown species. Eight of 17 studies that carried out seeding to restore semi-natural grasslands reported successful introductions of sown species, while four studies reported limited success, one reported a failed reintroduction, and four studies did not report enough information to allow the success of introductions to be determined. The review used keyword searches to identify studies where semi-natural grassland restoration was carried out. All studies collected seeds from wild or cultivated plants, and often sowed them as multi-species mixes.

    Study and other actions tested
  28. A review in 2010 of 13 studies of grassland restoration in Europe (Kiehl et al. 2010) found that after sowing seeds most studies reported that more than half of sown species became established and cover of sown species tended to be high. In 14 of 16 sites, more than half of sown plant species became established (53–96%), while in two of 16 sites, establishment was relatively low (28–32%). In eight sites, cover of sown plant species was 33–96%. The studies in the review lasted for 4–21 years and were carried out in a mixture of ex-arable fields, species-poor grasslands and mining sites. The soil was disturbed by ploughing or turf removal before sowing in some sites, and some sites were mowed or grazed by livestock after sowing. The seeds of 11–41 grass and forb species were sown at the restoration sites.

    Study and other actions tested
  29. A replicated, randomized, paired, controlled, before-and-after study on a road verge in Norway (Rydgren et al. 2010) found that sowing grass and forb seeds did not alter vegetation cover or similarity of the plant community to intact grasslands, but did increase plant species richness. For three of three years, vegetation cover in areas where seeds were sown (32–71%) did not differ significantly from areas where no seeds were sown (27–57%). The same pattern was seen for plant community similarity to a nearby intact grassland (data reported as Bray-Curtis dissimilarity index). However, plant species richness was higher in areas where seeds were sown (10–11 species/plot) than areas where no seeds were sown (8–9 species/plot). In September 2004, five blocks containing six 0.5 × 0.5 m plots were established. In each block, seeds of 17 species from nearby intact grasslands was sown in three plots and no seeds were sown in three plots. In each plot, vegetation cover of each species was estimated in sixteen 12.5 × 12.5 cm quadrats in July/August 2005–2007.

    Study and other actions tested
  30. A replicated, randomized, paired, controlled trial in 1988–2000 in arable field margins in Oxfordshire, UK (Smith et al. 2010) found that plots sown with grass and forb seeds had more perennial but not annual plant species than unsown plots, but the number of species in all plots declined with time. Five months after sowing, there were more species where seeds were sown (24 species: 8 annual and 16 perennial) than in unsown plots (15 species: 10 annual and 5 perennial). After 13 years, the number of species was not significantly different between sown plots (9 perennial species and 0.5 annual species) and unsown plots (7 perennial species and 0.5 annual species). Forty-eight 50 x 1.5 m plots were established in field margins, with eight plots in each of six fields. Twenty-four plots (four random plots/field) were tilled and sown with a mix of six grass and 17 forb species at a rate of 30 kg/ha in March 1988. The other twenty-four plots were left unseeded. Most plots were cut annually. Plant species were recorded three times/year from 1988 to 1990, and once in July 2000 in three 0.5 x 1 m quadrats/plot.

    Study and other actions tested
  31. A replicated, controlled study in 1996–2005 in former arable fields in Saxony-Anhalt, Germany (Conrad & Tischew 2011) found that sowing grass and forb seeds did not increase plant community similarity to a nearby intact grassland compared to not sowing seeds. After nine years, the similarity of the plant community to a target intact grassland did not differ significantly between seeded and unseeded areas (data reported as Sørensen’s similarity index). In, 1996, 1997 and 1999, some sites were sown with seed mixes containing cultivars and non-local grass and forb species, while others were not sown with seeds (replication unclear from study). Before sowing, all sites were ploughed and harrowed. Vegetation was surveyed in 5–19 x 25-m2 plots/site in May/June 1998–2001, and again in June 2005. Twenty-one plots were sampled in species-rich mesophile grassland within the same region to allow for comparison with the target habitat.

    Study and other actions tested
  32. A replicated, site comparison study in 2008 in ex-arable fields in the Czech Republic (Lencová & Prach 2011) found that sowing grass and forb seeds did not alter the number of meadow species, but older fields had vegetation communities more similar to those of undisturbed grasslands. The number of meadow or ruderal species in fields that were sown with seeds were not significantly different from those found in fields where no seeds were sown (no data reported). However, in both sown and unsown fields the similarity of vegetation communities to nearby undisturbed grasslands was higher in older fields (data reported as Czekanovski index). Twenty-six fields that were sown with commercial seed mixes containing grass and forb seeds and nine fields that were not sown with seeds were selected for study. Sowing rates in sowed fields were 30–35 kg/ha. All fields were grazed and/or mowed. In June/­July 2008, data on plant species was collected in all fields and nearby intact vegetation (replication in surveys unclear).

    Study and other actions tested
  33. A replicated, randomized, paired, controlled study in 2000–2009 in a former mine in Saxony-Anhalt, Germany (Baasch et al. 2012) found that sowing native grass and forb seeds increased cover of target grassland species, reduced cover of non-grassland species and initially increased total plant cover and species richness. After nine years, the cover of target grassland plant species was higher in areas where seeds were sown (72%) than areas where no seeds were sown (29%). The opposite was true for non-grassland species (seed: 14%; no seed: 47%). Total plant cover was higher in areas where seeds were sown after two to seven years (43–97%) than where no seeds were sown (6–66%). However, after eight years there was no longer any significant difference in plant cover (seed: 72%; no seed: 66%). After one year, plant species richness was higher in areas where seeds were sown (seed: 65 species/plot; no seed: 43 species/plot), but the number of species declined in seeded plots so that after five years there were fewer species in seeded areas (seed: 18 species/plot; no seed: 48 species/plot). After nine years, there was no longer any significant difference in species richness between seeded (41 species/plot) and unseeded areas (48 species/plot). In September 2000, three blocks were established, each with two 70 × 18 m plots. Seeds of 21 grass and forb species were sown at a density of 860 seeds/m2 in one plot in each block, and no seeds were sown in the other plot. A mulch layer (5 cm thick) was also added to sown plots. In June 2001–2009 (except for 2003), the vegetation cover of all species was estimated using three 25-m2 quadrats in each plot.

    Study and other actions tested
  34. A replicated, randomized, controlled study in 2009–2010 in an agricultural field in southern Spain (Ballesteros et al. 2012) found that sowing native grass and forb seeds increased plant density and species richness. In plots where seeds were sown, plant density was higher (1–11 plants/m2) than in plots where no seeds were sown (0 plants/m2). A similar pattern was seen for plant species richness (seeded: 0.2–1.8 species/0.25 m2, unseeded: 0 species/0.25 m2). In November 2009, locally collected seeds of seven native grass and forb species were sown in 20 randomly placed 5 x 5 m plots, and 20 plots were left unsown. Four different bedding materials were applied to plots prior to seeding (see original paper for details). Plant density and species richness were estimated in July and October 2010 using fifteen randomly placed 0.5 x 0.5 m quadrats/plot.

    Study and other actions tested
  35. A replicated, randomized, controlled, paired, before-and-after study in 2001–2010 in a calcareous grassland previously affected by shrub encroachment in Tuscany, Italy (Maccherini & Santi 2012) found that sowing grass and forb seeds did not alter species richness. Nine years after sowing, there was no significant difference in species richness between areas where seeds were sown (31 species) and areas where no seeds were sown (31 species). In 1999, shrubs were removed from the entire grassland. In spring 2001, blackthorn Prunus spinosa plants across the site were cut. In October 2001, four 3 × 5 m plots were sown with locally collected grass and forb seeds at a rate of 4 g/m2 while four plots were not sown with seed. In June/July 2001–2010, sixteen 2 × 1 m quadrats were placed in each plot and a point quadrat used to estimate cover of each plant species.

    Study and other actions tested
  36. A replicated, randomized, paired, controlled study in 1999–2009 in semi-dry grassland in the Czech Republic (Mitchley et al. 2012) found that sowing with grass and forb seeds increased sown forb and grass species richness and initially the cover of local forb and grass species, but over time forb species cover declined. After one year, species richness of sown forbs (7.9 vs 0 species/plot) and of sown grasses (3.9 vs 0 species/plot) was higher in areas where seeds were sown than in areas where no seeds were sown, and this remained true for both forbs (8.0 vs 3.5 species/plot) and grasses (4.1 vs 1.9 species/plot) after 10 years. The plant community composition of areas sown with seeds was also more similar to ancient hay meadows than that of areas where no seeds were sown. After one year, cover of sown forb species (17% vs 0%) and sown grass species (18% vs 0%) was higher in areas where local seeds were sown than where no seeds were sown. After 10 years, this pattern was still seen for sown grass species (52% vs 9%) but the cover of sown forb species in areas where seeds were sown was no longer significantly different from areas where no seeds were sown (8.7% vs 6.1%). Four 55 x 20 m plots were sown with a locally sourced seed mixture containing seven grass species and 20 forb species at a rate of 2 g/m2, while four plots were not sown with seeds. In June 2000–2004 and 2009, ten 1.5 x 1.5 m quadrats were placed in each plot and all species present and their cover recorded.

    Study and other actions tested
  37. A replicated, randomized, controlled study in 2008–2010 in 11 urban wasteland sites in Hellersdorf, Germany (Fischer et al. 2013) found that sowing grass and forb seeds increased plant species richness and the number of local conservation priority species. Average plant species richness in plots where seeds had been sown (43–50 species/plot) was higher than in plots where no seeds were sown (28–31 species). The percentage of vegetation consisting of local priority grassland species was also higher in plots where seeds were sown (29–46%) than in plots where no seeds were sown (7–9%). In autumn 2008, at each of 11 sites, seeds were sown in one 4 x 4 m plot, while no seeds were sown in another paired plot. All plots were mown and tilled prior to sowing. Seed mixes contained 27 species from the study region. In spring, early and late summer 2009–2010, a 3 x 3 quadrat was placed in the centre of each plot and vegetation cover mapped.

    Study and other actions tested
  38. A replicated, randomized, controlled study in 2006–2008 in a former agricultural field in Illinois, USA (Goldblum et al. 2013) found that sowing grass and forb seeds increased plant species richness, but had mixed effects on plant species diversity and vegetation cover. In 12 of 12 comparisons, plant species richness was higher in plots where grass and wildflower seeds were sown (5–10 species/plot) than in plots where no seeds were sown (2 species/plot). In five of 12 comparisons, plant species diversity was higher in plots where grass and wildflower seeds were sown than where seeds were not sown, but in seven comparisons there was no significant difference (measured as Shannon-Weiner index). In eight of 12 comparisons, there was no significant difference in plant cover between areas where seeds were sown (90–143%) and not sown (62–136%), but in four comparisons plant cover was higher in sown areas (sown: 105–120%, unsown: 62–76%). In December 2006, fifteen 9 x 9 m plots were sown with a mixture of grass and forb seeds of 128 species (applied at four different rates: 11.2, 33.6, 56.0 and 78.5 kg/ha), and three plots were not sown with seed. In June, July and August 2008, 1 x 1 m quadrats were placed in each plot and plant cover estimated.

    Study and other actions tested
  39. A replicated, randomized, controlled study in 2006–2008 in formerly forested grasslands in Slovakia (Novák et al. 2013) found that sowing grass and forb seeds did not alter plant species richness or cover compared to not sowing with seeds. In each of three years, plant species richness did not differ significantly between plots that were sown with grass and forb seeds (18–26 species) and plots that were not sown (16–25 species). The same pattern was true for vegetation cover (sown: 18–90%; unsown: 47–82%). In spring 2006, all shrubs and trees were cut and wood removed from the site. In three 2.2 x 2.2 m plots, bare ground was raked and seeded with a mixture of four grass and two legume species at a rate of 12.6 kg/ha, while three plots were left unseeded. All plots were grazed by cattle from May to October each year. In May 2006–2008, the cover and species richness of plants was visually estimated within a 1 x 1 m quadrat placed in the centre of each plot.

    Study and other actions tested
  40. A replicated, site comparison study in 2009 in 34 restored grassland sites and 20 intact grassland sites in the Czech Republic (Prach et al. 2013) found that sowing seeds of grass and forb species produced a different plant community to that of intact grasslands, but older sites were more similar to intact grasslands. Sites where seeds were sown had different plant communities to intact grasslands (data reported as ordination analysis). However, sites that were older had a community composition more similar to that of intact grasslands (data reported as ordination analysis). Thirty-four ex-arable sites where a mixture of grasses, legumes and other forbs had been sown 1–12 years previously were selected for study. These sites were compared to nearby species-rich grasslands. In June 2009, at each restored site, three 5 x 5 m plots were established and cover of all plant species estimated visually. Information on intact grasslands was taken from a national database. All sites were mowed once annually in June–August.

    Study and other actions tested
  41. A replicated, randomized, paired, controlled study in 1998–2010 in a former arable field in Minnesota, USA (Grygiel et al. 2014) found that sowing native grass and forb seeds increased native forb density where seeds were sown, as well as in surrounding areas. After 10–12 years, the average density of sown forb species was higher in plots where seeds were sown (29–40 plants/m2) than in plots where seeds were not sown (1 plant/m2). Sown forb density was also higher in areas 1–3 m away from plots where seeds were sown (5–19 plants/m2) than in areas a similar distance from plots that were not sown with seeds (<0.1–14 plants/m2). In autumn 1998, in each of five blocks, five 4 x 3 m plots were tilled and sown with a seed mixture containing four native grasses and 12 native forbs at a rate of 400 seeds/m2, while one plot was tilled but no seeds were sown. Some of the sown plots also received additional treatments (fertilizer, heat and/or carbon addition). In July–August 2005–2010, density of sown forb species was estimated in each plot, in a 1-m area surrounding each plot, and in three 0.5 × 0.5 m quadrats within each of four 3 x 3 m areas located 3 m away from each plot to the north, east, south and west.

    Study and other actions tested
  42. A replicated, site comparison study in 2009–2011 in 47 restored grassland sites and 25 intact grassland sites in the Czech Republic (Prach et al. 2014) found that sowing grass and forb seeds did not alter plant species richness, or the number or cover of grassland species compared to not sowing seed. After seven years, plant species richness in areas where seeds were sown (31.2 species/plot) did not differ significantly from that found in areas where no seeds were sown (47.8 species/plot), but both had fewer species than nearby intact dry grasslands (62.4 species/plot). The number of grassland species followed a similar trend (seeded: 19.7 species/plot, unseeded: 24.6 species/plot, intact grassland: 55.6 species/plot) as did the cover of grassland species (seeded: 74.3%, unseeded: 68.7%, intact grassland: 102.7%). Between one and 11 years before the start of the study, 35 restored sites were sown with a seed mix containing 44 local plant species at a rate of 17–20 kg/ha and 16 sites were not sown with seed. Twenty-five intact dry grassland sites were also used as a comparison. In 2009–2011, three 5 x 5 m plots were placed in each site and cover of all plant species visually estimated.

    Study and other actions tested
  43. A replicated, randomized, controlled, paired study in 2003–2006 in an unvegetated, former grassland in Norway (Auestad et al. 2015) found that sowing native grass and forb seeds initially increased plant species richness compared to areas where seeds were not sown, but this difference declined over time, and species composition of areas where seeds were sown became more similar to that of areas where seeds were not sown. After one year, the number of species in areas where seeds were sown (19 species/plot) was higher than in areas where seeds were not sown (15 species/plot), but after three years there was no longer a significant difference (seeded: 14 species/plot, unseeded: 12 species/plot). The species composition of sown and unsown areas became more similar over time (result based on ordination analysis). In 2003, all vegetation was removed from the site. Seeds were collected from road verges <10 km from the study area. In sixteen 0.5 x 0.5 m plots, seeds of two grass and 13 forb species were sown at a density of 2,300 seeds/m2, and in 16 plots no seeds were sown. Plant abundance and species richness were recorded in June–August 2004–2006 in each plot.

    Study and other actions tested
  44. A replicated, controlled study in 2008–2013 in a former arable field in Massachusetts, USA (Neill et al. 2015) found that sowing native grass and forb seeds increased the cover and species richness of native plants compared to unsown areas.  During 1–5 years after sowing, plots sown with native grass and forb seeds had on average greater cover (24–59%) and species richness (10–11 species/plot) of native plants than plots not sown with seeds (8–16%, 2–3 species/plot). The same was true for total plant species richness (sown plots: 17–23 species/plot; unsown plots: 10–11 species/plot). Statistical analyses carried out in this study did not test for significant differences between sown and unsown plots. Five 5 x 5 m plots were tilled to a depth of 16cm in June and August 2008 to remove non-native vegetation. In November 2008, the five plots were sown with seeds of 26 locally collected native species at a rate of 38 kg/ha. Fifteen other plots were not tilled or sown with seeds. Vegetation was surveyed annually in a 3 x 3 m quadrat placed in the centre of each plot in July and August 2007–2013.

    Study and other actions tested
  45. A replicated, randomized, paired, controlled study in 2003–2006 in eight prairie restoration sites in Iowa, USA (Wilsey & Martin 2015) found that sowing grass and forb species did not alter plant species richness. Plant species richness did not differ significantly between areas where seeds were sown (10.7 species/plot) and areas where seeds were not sown (9.8 species/plot). In each site, four 1 x 1 m plots were sown with seeds and two plots were not seeded. All plots were burned in April 2005. In 2003–2006, a 1 x 0.4 m quadrat was placed in each plot and species richness estimated by eye.

    Study and other actions tested
  46. A replicated, randomized, controlled, paired study in 2004–2012 on a road verge in Borgund, Norway (Auestad et al. 2016) found that sowing native grass and forb seeds initially did not alter plant species richness, but after eight years there was an increase in species richness. After one year, plant species richness in plots where seeds had been sown did not differ significantly from that found in plots where no seeds were sown (seeded: 15 species/plot, unseeded: 13 species/plot). However, after eight years, species richness was higher in seeded plots (21 species/plot) than in unseeded plots (16 species/plot). In September 2004, in each of five blocks, seeds of 11 grass and forb species were sown at a density of 1,900 seeds/m2 in three 0.5 x 0.5 plots, while in one plot/block seeds were not sown. All plots were paired. Seeds were harvested from donor sites (3–16 km away) in late summer/early autumn 2004. The site was mown in July/August 2007–2010. Vegetation cover was recorded in each plot in July–August 2005–2007 and 2012.

    Study and other actions tested
  47. A replicated, controlled study in 2009–2015 in a species-poor grassland near Wittenberg, Germany (Baasch et al. 2016) found that sowing grass and forb seeds increased the species richness of target grasses and forbs but did not alter their cover. After six years, plots sown with seeds had more target grass and forb species than unsown plots (data reported as statistical model results). However, the cover of target grass and forb species did not differ significantly between sown and unsown plots (data reported as statistical model results). In 2010–2015, plots sown with seeds had on average 9–19 target grass and forb species/year (6–20% cover), whereas unsown plots had 3–7 target species/year (4–13% cover). In 2009, three 30 x 6 m plots in each of six blocks were rotovated and rolled. In each block, two plots were sown with seeds (obtained by threshing with or without a regional seed mixture added), and one plot was left unsown. On-site threshing was carried out at a meadow 3 km away. All plots were mulched twice and mown once in 2009, and mown twice/year in 2010–2015. Vegetation was recorded annually within a 4 x 4 m quadrat in each of the 18 plots in 2010–2015.

    Study and other actions tested
  48. A replicated, randomized, paired, controlled study in 2009–2014 in a species-poor grassland near Wittenberg, Germany (Engst et al. 2016) found that sowing grass and forb seeds increased the species richness of target forbs, but did not alter target forb species cover or the species richness of target grasses. After five years, plots sown with seeds had on average more target forb species (4.6–6.5 species/plot) than plots not sown with seeds (2.9 species/plot). However, there was no significant difference in average target forb species cover (seeds: <1–3%; no seeds: <1%), or the average number of target grass species (seeds: 1.2–1.4 species/plot; no seeds: 0.8 species/plot). In 2009, six blocks each with three plots measuring 30 × 6 m were established. In each block, two plots were tilled, rolled and sown with seeds (one plot with seeds obtained by threshing only, and the other with seeds obtained by threshing and combined with a regional seed mixture), while no seeds were added to the third plot. Between 2010 and 2014, the study site was repeatedly flooded and mown twice a year. On-site threshing was carried out at two nearby sites, which were also regularly flooded and mown. Vegetation in each plot was recorded annually from 2010 to 2014 using 4 x 4 m quadrats.

    Study and other actions tested
  49. A replicated, controlled study in 2014–2015 in 73 agricultural grasslands in Brandenburg, Thuringia and Baden-Württemberg, Germany (Klaus et al. 2016) found that sowing native grass and forb seeds increased seedling species richness but did not alter the number of seedlings. After 7–19 months, plots that were sown with grass seeds had on average more plant seedling species overall (5 species/quadrat) and more grassland seedling species (data not reported) than plots that were not sown with seeds (all plant species: 4 species/quadrat; grassland species: data not reported). The average number of plant seedlings/quadrat did not differ significantly between sown (94 seedlings) and unsown plots (67 seedlings). Two 7 x 7 m plots were established in each of 73 grasslands. One plot was sown with a mix of native grass, legume and herb seeds (47–66 region-specific species), the other was left unsown. Seeds were mixed with sand and crushed soybean, which was also added to the unsown plots. Sowing was carried out in November 2014 and March 2015. Vegetation was monitored within a 2 x 2 m quadrat in each of the 146 plots on three occasions in May–June 2015.

    Study and other actions tested
  50. A replicated, randomized, controlled study in 2008–2015 in an agricultural field in California, USA (Werner et al. 2016) found that sowing grass and forb seeds increased cover of native forbs, decreased cover of non-native plants, and had no effect on grass cover. After eight years, cover of native forb species was higher in areas where seeds were sown (33%) than in areas where no seeds were sown (17%). However, grass cover did not differ significantly between areas where seeds were sown (9%) and areas where no seeds were sown (2%). Cover of non-native plants was lower in areas where seeds were sown (56%) than areas where no seeds were sown (81%). In 2008, a mix of native grass and forb seeds common to Californian grasslands was sown in twenty-five 1.5 x 1.5 m plots at a rate of 800 seeds/m2 and 25 plots were not sown with seeds. Two weeks before sowing all plots were tilled. In November–June of 2008–2010, all plots were weeded. In May 2015, quadrats measuring 1 x 1 m were placed in the centre of each plot and vegetation cover was visually estimated.

    Study and other actions tested
Please cite as:

Martin, P.A., Ockendon, N., Berthinussen, A, Smith, R.K. and Sutherland W.J. (2021) Grassland Conservation: Global evidence for the effects of selected interventions. Conservation Evidence Series Synopses. University of Cambridge, Cambridge, UK.

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Grassland Conservation

This Action forms part of the Action Synopsis:

Grassland Conservation
Grassland Conservation

Grassland Conservation - Published 2021

Grassland Synopsis

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 18

Go to the CE Journal

Discover more on our blog

Our blog contains the latest news and updates from the Conservation Evidence team, the Conservation Evidence Journal, and our global partners in evidence-based conservation.


Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape Programme Red List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Bern wood Supporting Conservation Leaders National Biodiversity Network Sustainability Dashboard Frog Life The international journey of Conservation - Oryx British trust for ornithology Cool Farm Alliance UNEP AWFA Butterfly Conservation People trust for endangered species Vincet Wildlife Trust