‘Robustness’ as established plants

This characteristic derives from the combination of high tolerance of competition, longevity and low palatability to slugs as established plants. When dealing with very weedy sites or sites where management is likely to be restricted, at least a core of the species selected should possess these characteristics. These species may, however, be those most likely to naturalise beyond the site, and so caution needs to be exercised (Table 6.14).

Establishment practice

The key factor that determines successful germination and establishment of all herbaceous plants is soil moisture (Fuller 1987; Wilson and Gerry 1995; Hitchmough et al. 2003). Consequently, the optimal time for sowing generally coincides with the months of lowest soil-moisture stress that are warm enough for germination to occur. Within this generalisation there are some specific times that research or practice have shown to be optimal for specific communities and species, as shown in Table 6.15.

Site preparation

This is a fundamental requirement for successful establishment and longer-term community development. Most urban sites support large populations of aggressive weed species. If these are not controlled prior to sowing, they will eliminate many of the sown forbs that germinate. In the longer term, some weed species will be controlled in grassy, meadow-like plant communities by the mowing-grazing regime. By this time, however, many of the desired species are likely to have disappeared. The practice of eliminating one vegetation to achieve an ecologically-based successor may seem to be a contradiction. Regeneration by seed in many semi-natural ecosystems is, however, an occasional event, with most seedlings eliminated by competition from the surrounding established vegetation (Grubb 1977; Morgan 1995). The potentially beneficial effects of established plants reducing soil-moisture stress by providing shade are outweighed by the harmful effects of shade on photosynthesis and root competition for water and nutrients. In most cases, competition is uniformly detrimental to the establishment and survival of sown species (Aguilera and Lauenroth 1995; Hutchings and Booth 1996). It is, however, important to assess the existing botanical significance of sites prior to finalising

Table 6.15. Optimal sowing times for herbaceous vegetation



Sowing dates

North American prairie grasses


Seed have high-temperature requirements for germination (C4 species), but are also intolerant of soil-moisture stress. As with prairie forbs, sowings beyond July are often unsuccessful. Frost heave and surface erosion results in low seedling survival

North American prairie forbs

March-June or



A diverse group of species that germinate at lower temperatures than prairie grasses. Most species show most reliable germination from October-February sowings as the chilling requirements of species are automatically met. March-June is, however, satisfactory for many species when fridge pre-chilling in moist sand is employed, and is essential if prairie grasses are to be included

Eurasian meadow grasses and forbs

March-June, August – September or October-March

Again, a very diverse group, although many species establish from spring or early autumn sowings. Overwintering losses are often less than for prairie-type species. Species that require lengthy winter chilling for germination, for example Primula veris, Astrantia and Rhinanthus, must be sown between October and December

Annual forbs


Providing the soil is moist, time of sowing will often be determined by when a flowering display is required. Species from warm, summer rainfall climates, for example Cosmos, Cleome and Helianthus annua establish poorly when sown into

cold soils under short day conditions

decisions on replacement vegetation.

Established weedy vegetation is most effectively eliminated through application of the translocated herbicide glyphosate, which is available as a range of proprietory products, for example ‘Roundup Biactive’. This herbicide has extremely low mammalian toxicity. It can be used throughout the year, when weed foliage is present, but has no effect on dormant, leafless weeds or weed seeds in the soil. Most effective control is achieved when applied to actively growing weeds between March and October. One application will kill highly sensitive weeds, however, a second or even third application at three to six week intervals may be required to control stoloniferous or rhizomatous perennial weeds, such as couch grass, Elymus repens. Weeds must be controlled prior to moving or cultivating soil, as this inevitably complicates getting the herbicide into weeds via their foliage.

Where there are philosophical or legislative objections to the use of glyphosate, other weedcontrol techniques can be used, for example repeated cultivation, steam sterilisation and mulching with opaque sheet mulches. Some of these techniques can be reasonably successful but generally require a longer timescale and cost substantially more than herbicidal weed-control.

The most problematic weed source when sowing herbaceous plants in situ is the weed seed bank. Herbicides such as glyphosate make control of established weeds straightforward, however the process of seed bed production and raking to incorporate sown seed generates a weed germination ‘pulse’. With plant communities where there is typically a lengthy lag between sowing and germination (for example as in autumn sowings of prairie forbs that will not germinate until spring), it is possible to overspray sowings with a herbicide such as glyphosate prior to emergence. There is always a fear of harming the sown seeds, however, in an unpublished study undertaken by the author on 10 species of prairie forbs there were no apparent adverse effects on the germination of the latter. This practice substantially reduces winter weed colonisation but does not give 100% control.

On long cultivated topsoils, the weed seed bank is numerically huge. Given sufficient time, the practice of shallow surface cultivation to promote weed germination followed by secondary cultivation some weeks later to kill these seedlings can reduce the weed seed bank. This is often referred to as the ‘stale seedbed technique’. Where it is possible to do so, having a year to prepare a site is extremely helpful, but is only rarely possible. The density of the weed seed bank declines with soil depth, and is sparse or absent on many soils below 200 mm. Consequently, the use of the lower topsoil or subsoil as a sowing medium is an effective strategy, providing these soils are sufficiently well- structured to support plant growth. It is not, however, always possible to strip off areas of topsoil, but where a site has already had the topsoil removed, then a subsoil is often a better choice for direct seeding than a topsoil. Because most subsoils are composed of fine soil-particles that hold large amounts of water and readily maintain continuity of moisture films with germinating seeds, they sponsor high establishment of sown species.

Weed seed banks in topsoil can also be suppressed by blanketing with shallow layers of weed-free materials, such as subsoil or mineral aggregates. In our research in Sheffield we have found that a 40-50 mm layer of coarse sand gives very effective control of weed seed banks. These mulch layers are spread then seed is sown into them and lightly raked in. Percentage emergence is generally lower when the sand sowing technique is used due to the increased moisture-stress experienced. This can result in excellent weedcontrol but very poor establishment when dry conditions are experienced concurrent with germination. Except where irrigation is available, sand mulching is most appropriate for winter sowing. For sowings made during the spring to summer period, owing into a subsoil mulch is likely to be a better alternative. These mulch layer techniques add significantly to the cost of sowing and are most appropriate for plant communities where the aim of management is to try to permanently exclude winter-growing grasses, as in the case of prairie-type vegetation or dry meadowsteppe vegetation dominated by forbs. For moist-wet meadow species tolerant of competition from cool season grasses, they are not justified.