Sow seeds

How is the evidence assessed?
  • Effectiveness
    70%
  • Certainty
    60%
  • Harms
    0%

Study locations

Key messages

  • Five of six studies (including three replicated, randomized, controlled studies, one site comparison study and one controlled study) in the UK, South Africa, and the USA found that sowing seeds of shrubland species increased shrub cover. One of six studies in the UK found no increase in shrub cover. One replicated site comparison in the USA found in sites where seed containing Wyoming big sagebrush was sown the abundance of the plant was higher than in sites where it was not sown. One replicated, randomized, controlled study in the USA found that shrub seedling abundance increased after seeds were sown. One study in the USA found very low germination of hackberry seeds when they were sown. One replicated, randomized, controlled study in the USA found that the community composition of shrublands where seeds were sown was similar to that found in undisturbed shrublands. One randomized, controlled study in the UK found an increase in the cover of heathland plants when seeds were sown.
  • One replicated, randomized, controlled study in South Africa found that sowing seeds increased plant cover. One replicated, randomized, controlled study in the USA found that areas where seeds were sown did not differ significantly in native cover compared to areas where shrubland plants had been planted. One controlled study in the USA found higher plant diversity in areas where seeds were sown by hand than in areas where they were sown using a seed drill.
  • Two of three studies (one of which was a replicated, randomized, controlled study) in the USA found that sowing seeds of shrubland species resulted in an increase in grass cover. One randomized, controlled study in the UK found no changes in the cover of grasses or forbs.

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 replicated, randomized, controlled study in 1995–1998 in a former maritime pine Pinus pinaster plantation in the Cape Peninsula, South Africa (Holmes 2001) found that in five of nine cases sowing seeds of shrubland species increased plant cover. After three years and in five of nine cases, plant cover in areas where seeds of shrubland species were sown was higher (17–95% cover) than areas where seeds were not sown (4–79%). Seeds were sown in seventy-two 4 m x 4 m plots and in 33 plots no seeds were sown. After sowing seeds mulch was spread over plots to stop seeds from being blown away. In each plot four 1 m x 1 m quadrats were established and vegetation cover assessed five times in 1995–1998.

    Study and other actions tested
  2. A randomized, controlled study in 1997–1999 in a sagebrush scrub shrubland that had been invaded by grass and burnt by wildfires in California, USA (Cione et al. 2002) found that sowing shrub seeds increased the seedling abundance of two of seven shrub species, but did not reduce grass cover. After one year, the areas where seeds had been sown had a significantly higher number of shrub seedlings for two of seven species (3–4 seedlings/m2) than areas that were not seeded (0 seedlings/m2). There was also no significant difference in grass cover between areas where seeds had been sown (82%) and areas that were not seeded (84%). In 1997 seeds were sown in five randomly located 5 m x 5 m plots, while in five other plots no seeds were sown. In spring 1997 plots were surveyed for grasses using two 0.25 m x 0.5 m quadrats/plot and two 0.5 m x 1 m quadrats/plot for shrubs.

    Study and other actions tested
  3. A replicated, randomized, controlled study between 2002 and 2006 in two degraded moorlands in the UK (Mitchell et al. 2008) found that sowing heather Calluna vulgaris seeds increased the cover of heather. After three years, heather cover was higher in plots where heather seeds were sown (7%) than in plots where seeds were not sown (1%). At each site fifty-four 100 m2 plots were established. On half of each plot heather seeds were sown by hand at a rate of 26000 seeds/m2. Plant cover was estimated in nine 1 m2 quadrats in each plot annually between 2003 and 2006.

    Study and other actions tested
  4. A replicated, randomized, controlled study in 1994–2003 in two agricultural fields in Suffolk, UK (Walker et al. 2007) found that sowing seeds of heathland species did not increase the cover of common heather Calluna vulgaris. After nine years, cover of common heather was not significantly different in areas where heathland seeds had been sown (0%) to that in areas where seeds had not been sown (0%). In 1994–1995 a mixture of common heather, grass, and herb seeds were sown in twelve 80 m2 plots and in four other plots no seeds were sown. In April 2003 five 2500 cm2 quadrats were placed in each plot and the cover of all plant species recorded.

    Study and other actions tested
  5. A controlled study in 1998–2006 in a fynbos site previously invaded by non-native trees in Western Cape, South Africa (Pretorisu et al. 2008) found that sowing seeds of fynbos species increased shrub cover, but did not alter the cover of non-native species, or native forbs. After eight years, cover of shrub species was higher in areas where seeds had been sown (20%) than where seeds had not been sown (2%). However, cover of native forbs and non-native species did not differ between areas that had been seeded and those that had not (no data provided). In 1997 all non-native trees were felled. Subsequently, a wildfire burned the area in 1998. In 1998 seeds of species typical of fynbos were sown in five 50 m2 plots, while in five plots no seeds were sown. Four 1 m2 quadrats were placed in each plot and vegetation cover assessed in 2006.

    Study and other actions tested
  6. A site comparison study in 1994–2003 in sagebrush scrub habitat that had previously been burnt by wildfire in Utah, USA (Gardner et al. 2009) found that sowing seeds increased shrub and perennial grass cover, but decreased cover of weeds. After nine years, shrub cover in areas where seeds had been sown was higher (10%) than in areas where seeds had not been sown (0%), but not significantly different to that found in undisturbed shrublands (13%). Perennial grass cover was also higher in areas where seeds had been sown (5%) than in unsown areas (0%) but was also higher than that found in undisturbed areas (1%). Cover of weeds was lower in areas where seeds had been sown (2.3%) than in unsown areas (55%), and it was also lower than in undisturbed areas (12%). In 1994 a wildfire burned part of the study area. Part of the burned area was subsequently seeded with a mixture of native and non-native shrubs and grasses, while another part was not seeded. A nearby unburned area was also used for comparison. In 2003 twenty 0.25 m2 quadrats were randomly located in each area and vegetation cover estimated.

    Study and other actions tested
  7. A replicated, randomized, controlled study in a degraded sagebrush scrubland habitat in California, USA (DeSimone 2011) found that sowing seeds did not increase cover of native plant species compared to planting, or a combination of planting and sowing seeds. Native plant species cover in areas where seeds were sown (2–9%) was not higher than in areas where plants were planted (2%) or areas where plants were planted and seeds sown (7–14%). Six randomly located 1 m2 plots were sown with seeds of shrubland plants, while six plots were planted with California sagebrush Artemisia californica plants, and another six plots were planted with plants and sown with seeds. Plant cover was recorded every year in May-July in the 1 m2 plots.

    Study and other actions tested
  8. A randomized, controlled study in 1988–2005 in former heathland in Dorset, UK (Pywell et al. 2011) found that the addition of shoots and seeds from a nearby heathland increased the cover of heathland species, but did not reduce the cover of grasses or forbs. After 17 years, cover of heathland species was higher in plots where shoots and seeds were spread (44%) than in plots where they were not (16%). There was no significant difference in the cover of grasses or forbs in plots where shoots and seeds were spread (grasses: 36%, forbs: 54%) and plots where they were not (grasses: 59%, forbs: 27%). In 1989 shoots and seeds harvested from a nearby heathland were spread on three 500 m2 plots and three plots were seeded. In 2005 the cover of plants was recorded in four 1 m2 quadrats which were randomly placed in each plot.

    Study and other actions tested
  9. A replicated, randomized, controlled study from 2002 to 2010 in a former heathland in mid-Wales, UK (Critchley et al. 2013) found that in the majority of cases sowing heather Calluna vulgaris seeds increased heather cover. After eight years, sowing of heather seeds increased heather cover in eight of twelve cases (1–30% cover) when compared to areas where no sowing had been carried out (0–7% cover). Nine paddocks 5–7 hectares in size were selected. Paddocks were grazed with cattle, sheep, or left ungrazed. In each paddock six 10 m x 10 m plots were established. Plots were randomly selected to be disturbed using a rotavator, trampled by cattle, or left undisturbed. These plots were divided into four 5 m x 5 m plots with one plot bring grazed but not seeded, one plot being grazed and seeded, one plot not being grazed nor seeded and one plot being seeded but not grazed.

    Study and other actions tested
  10. A replicated, randomized, controlled study from 2009 to 2012 in a sagebrush ecosystem that had previously been burnt in Oregon, USA (Davies et al. 2014) found that sowing grass and sagebrush Artemisia tridentata seeds led to increased grass and sagebrush cover. After three years, sagebrush cover was higher where grass and sagebrush seeds were sown (6%) than where grass seeds or no seeds were sown (both 0%). Grass cover was higher in blocks where grass seeds (8%) or grass and sagebrush seeds were sown (6%), compared to blocks where no seeds were sown (4%). Fifteen 15 m x 30 m plots were randomly assigned to be sown with perennial herbaceous plant seeds, sagebrush and perennial herbaceous plant seeds, or no seeds. Vegetation cover was measured annually between 2010 and 2012 using sixty 0.2 m2 quadrats.

    Study and other actions tested
  11. A controlled study in 2010–2011 in sagebrush scrub invaded by non-native grasses in California, USA (Kimball et al. 2014) found that sowing seeds by hand resulted in a higher density of shrub seedlings and plant species indicative of disturbance than using a seed drill. In two of two trials the densities of shrub seedlings were higher where seeds had been sown by hand (11–17 seedlings/quadrat) than where they were sown using a seed drill (3–5 seedlings/quadrat). In one of two trials the densities of seedlings of plant species indicative of disturbance were higher where seeds had been sown by hand (6 seedlings/quadrat) than where they had been sown using a seed drill (3 seedlings/quadrat). In December 2010 nine 18.5 m2 plots were sown with native seeds by hand, and in another nine plots seeds were sown using a seed drill. Seedling density was measured in June 2011 in 25 cm x 25 cm quadrats placed in each plot.

    Study and other actions tested
  12. A replicated, randomized, controlled study between 1984 and 2009 in a sagebrush steppe ecosystem in Colorado, USA (Hoelzle et al. 2012) found that sowing seeds of late successional shrubland plants produced a plant community which was similar to undisturbed shrublands after 25 years. Sowing seeds of late succession shrubland plants resulted in a plant community that was similar to undisturbed shrublands (no data presented). However, sowing seeds of early succession shrubland plants resulted in a plant community that resembled plots that had been left unseeded (results presented as model results). In 1984 four blocks consisting of six 500 m2 plots each were established. Within each block all vegetation and the top 5 cm of soil were removed. Plots were randomly assigned to be sown with late successional shrubland seeds, early successional shrubland seeds, or to be left unseeded. Adjacent to blocks a number of shrubland plots were left undisturbed.

    Study and other actions tested
  13. A study in 2014–2015 in a shrubland in Utah, USA (Stevens et al. 2016) found that sowing seeds of the shrub hackberry Celtis reticulata led to very low germination rates. After eight months, 41 of 2600 (1.6%) hackberry seeds had successfully germinated. However, after 12 months only 19 of the 41 seedlings had survived. In October 2014 seeds were collected from hackberry plants. Later in October 2014 2600 of the seeds were sown under boulders, with 100 placed under each boulder. The site was visited to check for signs of germination in May-November 2015.

    Study and other actions tested
  14. A replicated site comparison study in 1984–1999 in eleven sagebrush shrubland sites affected by wildfire in Nevada, USA (Eiswerth et al. 2009) found that sowing seed containing Wyoming big sagebrush Artemisia tridentata wyomingensis increased the density of Wyoming big sagebrush plants. Areas where Wyoming big sagebrush seed was sown had a higher density of Wyoming big sagebrush than areas where their seed was not sown (results presented as model results). All study sites burned in wildfires in 1984-1997 and seed mixes containing Wyoming big sagebrush seed sown at an unspecified number of sites, while other sites were either not sown with seed or sown with mixes that did not contain Wyoming big sagebrush seed. Wyoming big sagebrush density was recorded at least once at all sites following the sowing of seed.

    Study and other actions tested
Please cite as:

Martin, P.A., Rocha, R., Smith, R.K. & Sutherland, W.J. (2020) Shrubland and Heathland Conservation. Pages 483-525 in: W.J. Sutherland, L.V. Dicks, S.O. Petrovan & R.K. Smith (eds) What Works in Conservation 2020. Open Book Publishers, Cambridge, UK.

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Shrubland and Heathland Conservation

This Action forms part of the Action Synopsis:

Shrubland and Heathland Conservation
Shrubland and Heathland Conservation

Shrubland and Heathland Conservation - Published 2017

Shrubland and Heathland synopsis

What Works 2021 cover

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, mammals, 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 21

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 ProgrammeRed List Champion - Arc Kent Wildlife Trust The Rufford Foundation Mauritian Wildlife Supporting Conservation Leaders
Sustainability Dashboard National Biodiversity Network Frog Life The international journey of Conservation - Oryx Cool Farm Alliance UNEP AWFA Bat Conservation InternationalPeople trust for endangered species Vincet Wildlife Trust