2.1. VEGETATION STUDIES ON OLD FIELDS AND ABANDONED ARABLE LAND
Until the introduction of set-aside very few studies had been carried out on the development of grassland ecosystems, especially on abandoned arable land in the UK. The first studies of succession on abandoned arable land in the UK were at the Rothamsted Experimental Station in Hertfordshire. The Broadbalk and Geescroft Wildernesses were created in 1882 and 1867 respectively (Burnham, 1989).
Cornish (1954) studied the grasslands of the North Downs, and Wells et al. (1976) studied the relationships between vegetation, soils and land-use history on the Porton Ranges. Most previous studies, especially those on old fields in America, were on land which was neither in arable production for any great length of time or intensively managed. The introduction of set-aside policy allows the development of grassland ecosystems on previously intensive arable land to be examined for the first time.
2.2. VEGETATION STUDIES ON ARABLE FIELDS
Before the conservation of wildlife habitats became a popular concern, all non-crop vegetation on farmland was regarded as arable weeds. Various studies have examined weed distributions in arable land e.g. Froud-Williams and Chancellor (1982), Chancellor and Froud-Williams (1984) and Chancellor (1985).
A survey of grass weeds in central southern England in fields of winter wheat revealed that Avena spp. occurred in 32% of fields, Agropyron repens (Elymus repens) in 24%, Poa trivialis in 22%, Alopercurus myosuroides in 19% and Bromus sterilis in 9% (Froud-Williams and Chancellor, 1982). Therefore it can be assumed that these species are most likely to be the most important in the colonisation of set-aside. A second survey of cereal weeds examined dicotyledenous weeds in addition to grasses. The eight most frequent species recorded were Viola arvensis, Galium aparine, Stellaria media, Myosotis arvensis, Polygonum arviculare, Convolvulus arvensis, Bilderdykia (Polygonum) convolvulus and Lamium purpureum. However, surveys in Essex revealed that 22% of fields were infested with Convolvulus arvensis, 14% with Galium aparine, 5% with Polygonum aviculare, 3% with Lamium purpureum, 2% with Rumex obtusifolius and Myosotis arvensis, and 1% with Bilderdykia convolvulus. There was no Stellaria media or Viola arvensis.
Chancellor (1985) examined changes in weed flora of an arable field cultivated for 20 years which had previously been in permanent pasture. Maximum species richness occurred four years into the arable cultivation. Grassland weeds did not persist long under continuous arable cropping, though Trifolium repens survived the whole 20 years. Rumex obtusifolius survived 12 years, Plantago lanceolata 8 years, and of the 13 other species recorded, nine had vanished by five years. This may have some interesting implications for post set-aside management.