BD5001: Characterisation of soil structural degradation under grassland and development of measures to ameliorate its impact on biodiversity and other soil functions

Literature Review: Use of Plant Species for Remediation of Soil Compaction

February 2011

C N R Critchley1 & F W Kirkham2

1ADAS UK Ltd, c/o Newcastle University, NEFG Offices, Nafferton Farm, Stocksfield, Northumberland NE43 7XD. 2Ecological Research & Consultancy, Far View, Nymet Rowland, Crediton, Devon EX17 6AL. Executive Summary • Plant species vary in their tolerance of soil compaction and their ability to improve soil structure. The aim of this review was to identify plant species that may have potential to remediate soil compaction in grasslands and would also have some practical application in the field. The literature was reviewed, the specific objectives being to identify: 1. Plant species and traits shown to have remedial effects in compacted soils, 2. UK grassland species that could have potential to remediate soil compaction and 3. Ease of establishment, persistence and agronomic value of the candidate species. • There are relatively few relevant studies from grasslands in the UK, most studies having been carried out elsewhere under various climatic conditions and using commercial crop species within arable rotations. Lucerne (Medicago sativa) was the most widely studied species and featured most prominently in the literature. However, these studies do point to a number of consistent traits associated with alleviating soil compaction. Members of the family Fabaceae are frequently reported as being capable of tolerating compaction and improving soil structure. Possible mechanisms for this are linked to Rhizobium bacteria in root nodules and arbuscular mycorrhizal associations. A deep rooting system capable of deep penetration and radial expansion, especially a large tap root, is also a recurrent feature. In general, dicotyledonous species tend to have more strongly penetrating root systems than grasses. Long-lived perennials are also more likely to develop larger root systems than annuals. • Successful establishment of plant species into an existing grassland sward has been linked to ruderality, percentage germination of seeds and autumn germination. Members of the Fabaceae often have poor germination unless seeds are scarified or subjected to prolonged soaking. Generalist species, especially those associated with fertile soils, are more persistent in restoration experiments than habitat specialists and species associated with infertile soils. Persistence in a sward is also related to a species’ tolerance or avoidance of repeated defoliation caused by grazing or cutting. • Nutritive value of forage plants are determined by digestibility, protein content and concentration of essential minerals. Digestibility declines with maturity, the rate varying among species. Stage of growth and ability to absorb nutrients will affect a species’ protein and mineral concentrations. Forbs (including ‘weed’ species) are recognised for their mineral content and palatability but have declined in intensive grassland systems. Fabaceae are valued for their nitrogen-fixing properties and some also have high feeding value. Livestock usually learn to select the more digestible species from a sward and mostly avoid plants with spines. Plants with a low growth habit will often escape being grazed. Plants containing undesirable chemicals and highly invasive or competitive weeds are considered to be actively detrimental to agricultural production. • A total of 14 perennial, deep-rooted forb plant species, either native or naturalised to Britain, were identified by reference to their traits as candidate species with the potential to alleviate soil compaction. Species were only included in this initial list if the published trait data relevant to this study were found. Because of the relatively limited range of information in the literature and incomplete data on rooting characteristics, a further 5 candidate species were added following consultation with

i the project Steering Group. Each of these species was further reviewed under the main headings of ease of establishment, persistence, and agronomic value. Based upon this information, 9 candidate species were identified as having potential for alleviating soil compaction in grasslands, plus 2 further species that might be suitable on specific soil types (see table below). The remaining 8 species were rejected because they either have low establishment rates, low forage quality or would be unacceptable in an agricultural sward. • Field experiments are required to test how effective these plant species are at alleviating soil compaction under grassland. The 9 candidate species identified in this review have been selected and included as a mitigation method to be tested in the field experiment phase of Defra project BD5001. The 9 candidate species will be introduced into grassland swards to assess their remedial effects on soil compaction. The other mitigation method tested in the field experimental phase of Defra project BD5001 is mechanical loosening using sward lifters at two specific depths. These experiments will therefore investigate the potential of the selected plant species to alleviate grassland soil compaction in isolation and in combination with mechanical loosening. One hypothesis is that the selected plant species could potentially exploit the mechanically loosened soil and thus increase the duration of the mechanical loosening effect and perhaps increase the soil’s resistance to re-compaction by establishing roots and mycorrhizal associations within and around loosened soil aggregates.

ii a) Suitable species: Species Remedial characteristics Establishment & persistence Agronomic value Trifolium pretense Tap-rooted legume, mycorrhizal associations. Possibly not long-term persistent in High. (Red Clover) sward and native varieties might be slow to establish. Lotus corniculatus Tap-rooted legume, high mycorrhizal Establishes better than most other High. (Bird’s-foot trefoil) associations. native forbs. Achillea millefolium Mycorrhizal associations but no tap root. Establishes and persists well but Tolerates grazing; (Yarrow) might not disperse well in productive superior mineral source swards. to grasses and average for forbs. Hypochaeris Tap-rooted, mycorrhizal associations. Less easy to establish than other Reasonably good radicata species and needs to disperse source of minerals and (Cat’s-ear) seeds to persist. selectively grazed. Trifolium repens Superior to Lolium perenne although shallower Establishes and persists well. High. (White Clover) rooting depth than other forbs identified. Cichorium intybus Deep tap root but probably does not occur Establishes well and competes well High mineral content; (Chicory) naturally in grasslands. with legumes. can increase productivity and forage quality. Centaurea nigra Tap-rooted, mycorrhizal associations but no Moderate establishment and Probably low. (Black Knapweed) data on rooting depth found. persistence. Plantago lanceolata Tap-rooted but shallower rooting depth than Establishes and persists well. Medium feed value. (Ribwort Plantain) other forbs identified. Sanguisorba minor Probably deep tap-rooted with mycorrhizal Establishes well, tolerates grazing. Good quality forage. ssp. muricata associations. (Fodder Burnet)

iii

b) Suitable species but limited by soil type: Species Remedial characteristics Establishment & persistence Agronomic value Knautia arvensis Tap-rooted, mycorrhizal associations. Difficult to establish and limited to Calcium source but low (Field Scabious) infertile, dry and/or calcareous soils in other minerals and with low competition. crude protein. Medicago sativa Tap-rooted legume, mycorrhizal associations. Mainly on light or calcareous soils Comparable to other (Lucerne) and limited persistence in grazed legumes and high grassland. mineral content.

iv Contents

Executive Summary...... i a) Suitable species:...... iii b) Suitable species but limited by soil type: ...... iv Introduction...... 1 Methods...... 1 Results...... 2 Review of remedial effects...... 2 Provisional list of species...... 5 Medicago sativa (Lucerne/Alfalfa)...... 7 Trifolium pratense (Red clover)...... 9 Lotus corniculatus (Birds-foot trefoil) ...... 11 Silene vulgaris (Bladder campion) ...... 13 Knautia arvensis (Field scabious)...... 15 Hypochaeris radicata (Cat’s-ear) ...... 17 Potentilla anserina (Silverweed) ...... 18 Achillea millefolium (Yarrow)...... 19 Primula veris (Cowslip) ...... 21 Additional species...... 21 Conclusions...... 22 Suitable species...... 22 Species whose suitability is limited to certain soil types ...... 23 Unsuitable species...... 23 References ...... 23 Introduction Plant roots can alter soil structure by penetrating compacted soil and increasing its porosity, with potentially beneficial effects of increased hydraulic and gaseous movement and improved penetrability for other species. Plant species are known to vary in their ability to grow in compacted soil and in their effects on soil structure. Some species commonly associated with compacted soil might simply be tolerant of compaction (or of the agent of compaction such as trafficking) without having any significant effect on soil structure. The aim of this review was therefore to identify plant species that are potentially capable of remediating soil compaction under grasslands. Evidence was sought for remedial effects of particular species and of the species’ traits associated with successful remediation. From this, a provisional list of candidate species to be used in the field experiments was identified. It was also necessary to select species that would have practical application in the field. The literature was also therefore reviewed to determine the palatability of the candidate species to livestock and their ability to establish and persist in existing grassland swards. The specific objectives of the review were to identify: 1. Plant species and traits shown to have remedial effects in compacted soils, 2. UK grassland species with potential to remediate soil compaction and 3. Ease of establishment, persistence and agronomic value of the candidate species.

Methods The scientific literature was reviewed using combinations of keywords in the Scopus database and search facility. Keywords used were combinations of “soil compact*”, “soil”, “plant species”, “plant community”, “grassland”, “grassland restoration”, “grassland species”, “grassland species diversity”, “legume*”, “tap root” and “root elongation” plus names of a number of key species names already suspected to be associated with soil compaction. Citation trails were also followed up and a limited number of web searches carried out for grey literature on species for which no information was found in the journal searches. Suitable papers and documents were scrutinised for information on plant species and species’ traits with potential for soil compaction remediation. Very few UK grassland species were identified directly from the literature. A further selection was therefore carried out by filtering out species that possessed traits reported in the literature as being associated with remediation. Sources of species’ trait data used were Fitter & Peat (1994), Hodgson et al. (1995), Hill et al. (2004) and Grime et al. (2007) and supplemented by individual web searches where specific data for individual species were lacking. Following consultation with the project Steering Group, a further five potential species thought to have suitable remedial characteristics were added. The candidate species were then subjected to a further review on the basis of their ease of establishment in seed mixtures, their persistence once established, and their agronomic value. This part of the review was also based upon a search of the scientific literature using Scopus, Google and Google Scholar and literature sources already in hand. In a few cases these searches identified useful non-scientific or anecdotal website information. A very wide range of key words and key word combinations were used in combination with the species name and relating to topics that included, inter alia, establishment, germination characteristics, dormancy, seed bank persistence, persistence and longevity, regeneration strategy, habitat preference, tolerance of

1 defoliation and/or grazing, palatability and grazing selection, feed value and chemical composition, and value for wildlife.

Results

Review of remedial effects Medicago sativa Several studies have focused on Medicago sativa (lucerne), which is capable of penetrating compacted soils and can have a remedial effect that benefits following crops in an arable rotation. These studies on M. sativa refer to lucerne or alfalfa (i.e. ssp. sativa) and not the UK native ssp. falcata. In laboratory experiments, M. sativa had a greater proportion of roots successfully penetrating a hard wax layer than barley (Hordeum vulgare), Cichorium intybus, lupin (Lupinus luteus) or Trifolium pratense (Löfkvist et al., 2005). Field experiments in the USA on sandy loam examined the effects of harvest trafficking and showed that water use efficiency of M. sativa was unaffected by soil compaction, although it was reduced by the direct physical effects on the plant of trafficking (Rechel et al., 1991). A similar study of tractor wheeling in Poland on silty loam showed that in compacted soil M. sativa roots were thicker and a greater proportion of dry matter was allocated to roots in the upper 10cm horizon. Significantly, the root system also improved soil structure after two years, with a decrease in soil bulk density attributed to both root and earthworm activity (Głąb, 2008). M. sativa (along with C. intybus) was the most successful species in a comparative field experiment in southern Sweden on glacial till, in which roots were able to grow in compacted soil and enhanced drainage, as indicated by a reduction in saturated hydraulic conductivity. Both these species had a greater effect on saturated hydraulic conductivity than did earthworm activity (Löfkvist, 2005). The effects of M. sativa on compacted soils can also be beneficial to other plant species. In the USA, when cotton (Gossypium hirsutum) was planted into an existing 5-year old stand of M. sativa, the infiltration rates through the soil macropore system in both compacted and uncompacted soils were much higher than normally expected from cotton alone (Meek et al., 1990). However, less convincing results were obtained when M. sativa strips were planted into cotton crops on clay soils in New South Wales, Australia. Here, soil strength was actually higher and electrical conductivity lower under M. sativa than cotton, although there were some beneficial effects of M. sativa in raising organic matter, earthworm activity and air-filled porosity of the subsoil (Hulugalle et al., 1999). Introduction of a M. sativa break crop in a durum wheat (Triticum aestivum var. durum) rotation in semi-arid Syria improved soil quality by increasing organic matter, which correlated with increases in other properties including soil structure and water infiltration rates (Masri & Ryan, 2006). Other legumes In addition to M. sativa, there is evidence that other legumes can have a remedial effect on compacted soil. For example, in their study of break crops Masri & Ryan (2006) obtained comparable results to M. sativa for faba bean (Vicia faba). Intercropping a mixture of three legumes (M. sativa, Trifolium sp. and Vicia villosa) within a continuous maize (Zea mays) system resulted in decreased dry bulk density and soil penetration resistance, although this was partly attributable to green manure incorporation as well as rooting activity (Latif et al., 1992).

2 The beneficial effects on following crops of including legumes in the rotation have also been demonstrated. A 2-year crop of Stylosanthes hamata penetrated a compacted layer in sandy soil and increased the macroporosity, so that a following maize crop developed a larger root system compared to a conventional 1-year rotation (Lesturgez et al, 2004). Soil strength, as measured by penetrometer resistance, was decreased more by the legumes Vicia faba, lablab (Lablab purpureus) and field pea (Pisum sativum) than others (wheat, cotton), although soybean (Glycine max) performed less well. This increased the yield of following cotton crops, which was thought to be due to their being able to develop better root systems (Rochester et al., 2001). Similarly, yields of maize and cowpea (Vigna unguiculata) were more improved following legume cover crops than following grasses, which was attributed to slightly decreased soil bulk density in the top 10cm and improved water retention and transmission properties. This appeared to be attributable to loosening of the soil and an increase in soil pores by roots, especially those of the more deeply-rooted legumes (Wilson et al., 1982). This is supported by a study in the tropics where woody legumes decreased soil bulk density and soil penetrability, with tree species being more effective in the long term than a herbaceous climber (Pueraria phaseoloides) (Salako et al., 2001). Evidence for a remedial effect of legumes therefore comes from a wide range of environments. Legumes are, however, variable in their effectiveness for alleviating soil compaction. It has been claimed that lupin is highly tolerant of compacted soils (Atwell, 1988) and it was found to have the greatest root penetration in a comparison with seven other crop species (barley, oats, Lolium rigidum, wheat, faba bean, safflower (Carthamus tinctorious) and pea) in compacted soil (Materechera et al., 1993). Similarly, lupin roots were able to penetrate a hard wax layer in a laboratory experiment (Löfkvist et al., 2005) although in field experiments by the same research team it performed no better than other species tested (Löfkvist, 2005). Cresswell & Kirkegaard (1995) also reported that lupin was unable to penetrate dense subsoil to improve its porosity. Pasture overseeded with Trifolium incarnatum had higher water stable aggregate concentrations and lower penetration resistance at 10-15 and 15-20cm depths than pasture supplied with fertiliser N, but only at specific sampling dates, which appeared to be linked to some unknown interaction with climate (Karki et al., 2009). Native species Most of the legumes studied in the literature have been commercial crop species. Trifolium pratense also features in the literature and, although most studies probably refer to agricultural varieties, the results are likely to have some relevance for the British native species. T. pratense appears to have moderate capability for improving compacted soils. In comparison with other cropping systems consisting of combinations of barley (Hordeum vulgare) or Bromus inermis, continuous T. pratense resulted in fewer large soil aggregates and more small aggregates, although aggregate stability was at an intermediate level (Broersma et al., 1997). Similarly, Löfkvist (2005) and Löfkvist et al. (2005) found T. pratense to have an intermediate capability for root penetration and improving soil structure, compared to M. sativa, C. intybus, Festuca arundinacea and barley. If T. pratense were to be used in grassland to alleviate soil compaction, the source of compaction would first need to be eliminated as it is weakly intolerant of trampling (Roovers et al, 2004). In a direct comparison with Lolium perenne, Trifolium repens was found to be beneficial to soil structure (Mytton et al., 1993). T. repens developed a more free-draining structure with more large pores although no differences in bulk density, porosity or aggregate

3 stability were detected. In contrast to T. pratense, T. repens is highly tolerant of trampling (Roovers et al., 2004). Grasses Some grasses can also tolerate compacted soils. For example, a field study in eastern USA showed that Tripsacum dactyloides was tolerant of an acidic compacted soil (Krizek et al., 2003). Beneficial effects have also been demonstrated. Over a period of 4 years, pore space in the top 30mm of a compacted soil was reduced by a Lolium perenne sward, which was thought to be attributable either to the pressure from adventitious roots or increased faunal activity (Douglas et al., 1992). A continuous grass system of Bromus inermis had the highest surface aggregate stability in a comparison with crop rotations including barley and Trifolium pratense (Broersma et al, 1997). Crop rotations including grass can benefit the following crops by reducing the effects of soil compaction. Paspalum notatum penetrated a compacted soil layer, and allowed a following cotton crop to make use of the biopores created (Dexter, 1991), while root growth of a soybean crop below a compacted layer was improved by a previous crop of Avena strigosa (Da Silva & Rosolem, 2002). In general however, dicotyledonous species are more likely to have a beneficial effect on soil structure than monocotyledonous species as they have been shown to have superior penetration in compacted soils resulting in the creation of more biopores (Materechera et al., 1993). This is supported by experimental evidence of Da Silva & Rosolem (2003) who found that grasses were more sensitive to soil compaction during early development than legumes. In addition, Löfkvist (2005) found that although roots of Festuca arundinacea penetrated compacted soil, they had little effect on its structure, attributing this to the small diameter of its roots which could use existing pores but not create new ones. Microbial associations The ability of some legumes to ameliorate compacted soil is probably related to Rhizobium bacteria associated with the root nodules, their association with arbuscular mycorrhizae (AM) and their physical rooting characteristics. Rhizobium bacteria produce extracellular polysaccharides, which improve soil structure by binding soil particles into larger units (Latif et al., 1992; Mytton et al, 1993). However, these effects are transient since the polysaccharides are unstable. Various nodular functions of Anthyllus cytisoides were maintained in compacted soils when AM were present, but not in non-mycorrhizal plants, including carbon and antioxidant metabolism and some protein contents (Goicoechea et al., 2005). In a laboratory experiment on pigeon pea (Cajanus cajan), the stress impact of compacted soil was reduced in the presence of an AM fungus, which appeared to promote root elongation (Yano et al., 1998). A different effect was identified by Li et al. (1997), in which the hyphae of an AM associated with T. pratense were able to penetrate a compacted soil layer, thus enabling the plants to obtain more phosphorus despite their own root systems being severely reduced. AM also produce glomalin- related soil protein in large quantities, which improve soil structure by helping to aggregate soil particles in water-stable macro-aggregates (Bedini et al., 2009). AM have also been shown to reduce the stress impact of compacted soil on maize (Miransari et al., 2007) and wheat (Miransari et al., 2008), suggesting that species other than Fabaceae with AM associations might also be able to tolerate soil compaction. Since microbial activity appears to be important in affecting soil structure, any agent that alters the structure of the microbial community might also affect this process. One example is the hemi-parasite Rhinanthus minor which is associated with a reduction in

4 the fungal:bacterial biomass ratio and enhanced N-cycling in grassland soil (Bardgett et al., 2006). However, it is not known whether these changes to the microbial community are likely to help or hinder the process of improving soil structure. Root systems The other major trait which enables some legumes and other species to remediate soil compaction is a strong root system capable of deep penetration and radial expansion, and in particular a large tap root (Wilson et al., 1982; Materechera et al., 1992; Whalley et al., 1995; Löfkvist, 2005; Reinert et al., 2008). Most studies have been carried out on crop species, but Reintam et al. (2008) also showed that the deep roots of Cirsium arvense were better able to penetrate compacted soil than barley, and resulted in improved soil structure. Tanakamaru et al. (1998) found that among most cereal crops, the penetration force in compacted soil tended to increase as the diameter of the seminal root increased. In contrast to other findings, however, they also found that, with the exception of Trifolium repens, the penetration force of legume crops was greater with smaller diameter of the main root. Despite this, most evidence points to plants with a large root system being more likely to tolerate and alleviate soil compaction. Results of experiments comparing different species might have to be treated with caution however, because there can be significant variation among cultivars of the same crop species in the ability of both tap and basal roots to penetrate a compacted layer of soil (Bushamuka & Zobel, 1998). Perennials with longer life span are also more likely to penetrate deeper into the soil, because they will have more time for the roots to develop than annuals and they will have more opportunity than short-lived species to grow during favourable conditions which might be seasonal (Cresswell & Kirkegaard, 1995; Löfkvist, 2005).

Provisional list of species From the literature review, the principle traits associated with remediation of soil compaction were of dicotyledonous species with perennial life history (especially family Fabaceae), tap root, deep root system and AM association. Using the 2008 version of PlantAtt1 (Hill et al., 2004), 126 perennial species associated with the Neutral Grassland Broad Habitat were first identified. Data on rooting system and depth were extracted from the Ecoflora database2 (accessed November 2009) (Fitter & Peat, 2004) and supplemented with web searches for individual species for which data were lacking. Information on mycorrhizal associations was obtained from Hodgson et al. (1995) and updated from Grime et al. (2007). An initial list of 14 candidate species was filtered (Table 1a) and prioritised in three groups comprising (i) Fabaceae with tap root; (ii) other dicots with tap root; (iii) any dicots with no tap root but mycorrhizal association. In all cases, minimum rooting depth class of 10-50cm was specified and species were sorted by descending rooting depth.

1 database of plant species’ attributes for Britain and Ireland. 2 database containing over 2200 species of higher plants that occur in the British Isles.

5 An additional species, Cirsium heterophyllum, was also selected on the basis of its rooting depth and mycorrhizal association (group (iii)). However, its range in Britain is restricted primarily to upland and northern locations (Preston et al., 2002) and so it was excluded from the initial list. Data on rooting systems and rooting depth could not be found for all species and it is possible that others might also have suitable properties for alleviating soil compaction. Examples were Lathyrus pratensis and Trifolium medium (no specific data on root system or depth found) and Centaurea nigra, C. scabiosa and Silaum silaus (no specific data on root depth found). Following consultation with the project Steering Group, an further five species were added that were considered to have potential for alleviating soil compaction (Table 1b)

Table 1. Initial list of candidate species. a) Perennial species occurring in the Neutral Grassland Broad Habitat. Taxon name Family Root Depth Root System Mycorrhizae (cm) Medicago sativa Fabaceae 200-400 Tap Yes Trifolium pratense Fabaceae >100 Tap Yes Ononis spinosa Fabaceae 50-100 Tap Yes Lotus corniculatus Fabaceae 10-50 Tap, Fibrous Yes Cirsium arvense Asteraceae >100 Tap, Adventitious Yes Rumex crispus Polygonaceae >100 Tap Intermediate Silene vulgaris Caryophyllaceae >100 Tap No Knautia arvensis Dipsacaceae 50-100 Tap Yes Hypochaeris radicata Asteraceae 50-100 Tap, Fibrous Yes Potentilla anserina Rosaceae 10-50 Tap, Adventitious Yes Senecio jacobaea Asteraceae 10-50 Tap Yes Achillea millefolium Asteraceae 10-50 Adventitious Yes Ranunculus repens Ranunculaceae 10-50 Adventitious Yes Primula veris Primulaceae 10-50 Fibrous Yes b) Additional species. Taxon name Family Root Depth Root System Mycorrhizae (cm) Trifolium repens Fabaceae 0-20 Tap, Adventitious, Yes Fibrous Cichorium intybus Asteraceae n.d. Tap Yes Centaurea nigra Asteraceae n.d. Tap Yes Plantago lanceolata Plantaginaceae 0-10 Tap Yes Sanguisorba minor Rosaceae n.d. Tap n.d. ssp. muricata n.d. = no data found.

Five species on the initial list are not considered further because it would be unacceptable to introduce them into an agricultural sward. C. arvense, R. crispus and S. jacobaea are all noxious weeds under the Weeds Act 1959 and O. spinosa has spines that are harmful to livestock. R. repens is an invasive species normally considered to be an agricultural weed.

6 In the following sections, the remaining species are reviewed individually with respect to ease of establishment, persistence and agronomic value and their value to biodiversity also summarised.

Medicago sativa (Lucerne/Alfalfa) There are three sub-species of M. sativa – sickle medick (Ssp. falcata), sand lucerne (Ssp. varia) and lucerne (Ssp sativa), of which lucerne is the most commonly cultivated, although sand lucerne is cultivated in parts of Europe. Sickle medick is scarce; it is native to East Anglia though naturalised elsewhere in Britain. Sand lucerne is also native, but is actually a fertile hybrid of origin from the other two sub-species (Stace, 1997). References to either lucerne or Medicago sativa hereafter refer to M. sativa ssp. sativa because it is the subspecies reported widely to have an effect on soil structure. Lucerne is a relatively tall-growing (<90 cm, Stace, 1997) legume hemicryptophyte/ geophyte (sensu Raunkiaer, 1934; Sullivan, 1992), i.e. its buds are located near the ground surface, with its exceptionally long tap root representing a below-ground storage organ. Naturalised populations normally flower between June and September (Rose, 2006). However, it is not native to Britain, originating from the near east and central Asia, although it has become naturalised from Sweden to North Africa (Anon, 2009a). Commercial varieties suitable for growth in the UK have been available for several decades, with early varieties originating from North-West Europe (NIAB, 1966). It is grown mainly for hay or silage, although it can be used for grazing under careful management (see section on Persistence). It is grown mainly on calcareous soils (Sheldrick et al., 2000) - its optimum pH being slightly higher than for clovers at >6.5 (Anon, 2009a). Lucerne is moderately valuable as a food source for wildlife, with 45 insect species identified as specifically associated with this plant, of which 3 are mono-specific associations (Mortimer et al., 2006). These numbers were about average for the 56 plant species reviewed by Mortimer et al. (2006), although an above average number (13) of the specific insect associations identified were with Red Data Book and Nationally Notable or Scarce insect species. Whilst Medicago spp (M. sativa and M. lupulina) appear in the diet of a few bird species, these species together were not identified as an important direct food source for any (Mortimer et al., 2006). Ease of establishment Between 45-75% of seeds are hard coated, requiring scarification for maximal germination, although this may be achieved during mechanical harvesting (Sullivan, 1992). It is usual to inoculate seed with the appropriate bacterium (Rhizobia meliloti) at the time of sowing – although this may be less important when introduced into existing permanent grassland than when sown into light arable soils. The benefit of inoculation has been shown to be significantly greater when incorporated within a lime seed coating (Horikawa et al., 1996). Lucerne is best sown without a companion grass, allowing chemical weed control when it is dormant in winter (Sheldrick, 2000). Some guidelines have suggested that a companion grass could be sown at a low rate (about one third of the weight of lucerne seed), although Dactylis glomerata was recommended in preference to Lolium perenne as the latter was considered to be too aggressive to allow optimum establishment of lucerne (Royal Agricultural College lecture notes, unpublished). Persistence

7 Lucerne is a relatively short-lived perennial with sown crops lasting 3-6 years (Sheldrick, 2000; Mortimer et al., 2006). It is best suited to well-drained soils and performs badly under poor drainage conditions, although it can be grown on moderately well-drained soils (Jones & Olsen, 1987). It forms a persistent seed bank, which represents its main regenerative strategy under natural conditions (Hodgson et al., 1995). Mortimer et al. (2006) class lucerne as unsuitable for grazing, but other sources state that it will persist under grazing if grazed leniently and at a late stage of maturity (Lodge, 1991). Grazing periods of 16-20 days apparently have no effect on persistence provided that a rest period of at least 35 days is left between successive grazing periods (Lodge, 1991); shorter grazing periods of 5-7 days are recommended by some authors (Collins & Fothergill, 2008). There has been considerable recent effort in plant breeding in Europe, Australia and the USA to produce grazing-tolerant varieties, particularly as tolerance of grazing aftermath growth following silage cropping is considered important (Collins & Fothergill, 2008). The persistence of lucerne may be reduced by the particular susceptibility to clover rot (Sclerotinia trifoliorum) that it shares with red clover (Lewis & Hopkins, 2000). Agronomic value Lucerne is less digestible than other forage legumes such as red clover (T. pratense) and particularly white clover (T. repens) at equivalent growth stages, differing by about 5-6 and 10 units respectively at an early stage of development and with faster rate of decline, at 2.8 units per week, compared to 2.5 and 0.8 units per week for red and white clover respectively (Beever et al., 2000). However, it has a very high potential yield of protein, due both to a high potential dry matter (DM) yield of up to 15 t DM/ha over 3-4 harvests at D-value (i.e. % digestible organic matter in the DM, DOMD) of 55-60 (Hopkins, 2000) and a high crude protein content. Crude protein content declines with increasing maturity and declining digestibility (Sullivan,1992), but although lucerne silage and hay tend to have lower metabolizable energy (ME) values than grass hay or silage, crude protein (CP) contents are much higher, with the CP of lucerne hay equivalent to good quality fresh grass (Beever et al., 2000; McDonald et al., 2002). In an experiment comparing growth rates of fattening lambs grazing either lucerne, red clover or bird’s-foot trefoil in comparison with perennial ryegrass, lambs grazing lucerne or red clover had significantly higher rates of intake than other forages (Speijers et al., 2004). Growth rates of lambs grazing lucerne were lower than for bird’s-foot trefoil but similar to those of lambs grazing perennial ryegrass or red clover. Crude protein (CP) concentration declined significantly in lucerne (-2.6 g/kg DM/day) and at a similar rate in bird’s-foot trefoil over the grazing period starting in September, but remained comparatively constant in red clover and perennial ryegrass. The total mineral content of lucerne has been shown to be considerably higher than that of grasses at all stages in development, and higher on average than all the other forbs tested except chicory (Cichorium intybus), particularly with respect to Ca content (Thomas et al., 1952). Lucerne has a high rate of biological nitrogen fixation relative to other legumes, at 300 kg N/ha/year, compared to 150-250 kg/ha/year for white clover, growing with perennial ryegrass (L. perenne) at a 30:70 ratio (Frame & Newbould, 1986), 60-138 kg N/ha/year for bird’s-foot trefoil (Lotus corniculatus), and 125-220 kg N/ha/year for red clover (Frame et al., 1998). However, high fertilizer application rates of K in particular may be needed to maintain maximal yield (Mortimer et al., 2006).

8 In a modeling analysis of the comparative productivity and profitability of grass- and legume-based silage production systems in northern Europe, Topp & Doyle (2004) predicted both white clover and lucerne to be more consistently profitable than grass- based systems receiving 200 kg N/ha/yr. Lucerne can cause bloat if ingested when fresh and young and also contains oestrogens that can affect fertility in sheep and cattle if consumed prior to breeding (Mortimer et al., 2006), whilst Beever et al. (2000) state that these phyto-oestrogens have little effect on cattle.

Trifolium pratense (Red clover) Red clover is a relatively tall (<60cm) hemicryptophyte legume with a decumbent or erect growth habit, although commercial varieties (separated as var. sativum) are more robust (Stace, 1997). Wild ecotypes flower between May and September (Rose, 2006). It is a fairly short-lived perennial and when grown commercially for silage is not expected to be productive for longer than two years (Sheldrick, 2000). However, it is a common and usually persistent constituent of a wide range of permanent pasture and meadow types, presumably due to its ability to regenerate both from freshly sown seeds and from the seed bank (Hodgson et al., 1995). Red clover is a very valuable food plant for insects and Mortimer et al. (2006) listed 90 specific insect associations for this species. Of the 56 plant species they reviewed, only six had more associations. It is a particularly valuable nectar-source and is among several key forage species whose declining abundance in the countryside is thought to be responsible for reductions in the density and range of bumblebees in recent decades (Carvell et al., 2001). It is also an important constituent, either as seed or as leaf material, of the diet of several bird species (Vickery et al., 1999, cited in Collins & Fothergill, 2008) Ease of establishment Despite being a common constituent of most mesotrophic grassland communities (Rodwell, 1992), T. pratense was included in seed mixtures in only five of the 25 studies on the restoration of species-rich grassland reviewed by Pywell et al. (2003). Pywell et al. derived performance indices (range 0 to 1) for 58 species during 4 years after sowing. T. pratense was generally slow to establish, with an index of only 0.06 for the first year, but reaching 0.15 in year 2 and declining somewhat thereafter. These values seem low, but of the 45 forbs reviewed, only 10 performed better overall. Information described below under persistence suggests that sowing date may be particularly important for this species and others with similar germination behaviour. Persistence As noted above, red clover is a common constituent of most mesotrophic permanent grassland communities, albeit at low-moderate cover and variable frequency (Rodwell, 1992). It is sensitive to increased cutting frequency (Sheldrick et al., 1986), probably as a result of both its relatively erect growth habit and its property of regenerating from freshly shed seed. It is less suitable for grazing than T. repens (Rochon et al., 2004). According to Mortimer et al. (2006), cultivated T. pratense has a lifespan of about three years, and whilst wild types may live longer, it is likely to be more persistent under silage making than either grazing or hay making. Its frequency in hay meadows also suggests better persistence in swards cut for hay rather than grazed (Rodwell, 1992). Studies in hay meadows in Somerset (Kirkham, 1996) show that management that allows both

9 maximal seed dispersal and a high proportion of shed seed to germinate in the autumn is important for maintaining a significant presence of the species, although even with an optimal cutting date in July T. pratense occupied only about 2% of vegetation cover. The mechanisms ensuring persistence of the species at low levels in meadows, where native types are adapted to traditional hay management or to infrequent defoliation by grazing, probably differ from those operating in recently established leys where commercial cultivars are a major component. Here the timing and frequency of cutting in relation to stage of physiological development, and the consequent effect on the longevity of individual plants, may be more important than any effect of allowing plants to set seed. It therefore seems unlikely that T. pratense will persist as a major constituent of a sward for more than a few years, particularly if sown in fertile soils, but its persistence as a significant, albeit minor, component of swards may be enhanced by management that allows it to set seed. This will also help to maximise its value as a food and nectar source for wildlife. Red clover is particularly susceptible to clover rot (Sclerotinia Trifolium) (Lewis & Hopkins, 2000), and can also be infected by Red Clover Necrotic Virus and stem eelworm (Ylimaki, 1966, cited in Mortimer et al., 2006). Agronomic value When grown for agronomic reasons, red clover is normally sown with a companion grass (usually Lolium perenne and/or L. multiflorum) and managed as a ley for forage conservation (Hopkins, 2000). It can be as productive as lucerne at up to16 t DM/ha/year (Mortimer et al. 2006). Its digestibility during primary growth is intermediate between that of white clover and lucerne initially, but declines at a rate only slightly slower than lucerne e.g. 2.5 units per week compared to 2.8 for lucerne and 0.8 for white clover (Beever et al., 2000). It has a high protein content, and values quoted by McDonald et al. (2002) suggest a slightly higher level at an early flowering stage than lucerne at an equivalent growth stage. Red clover tends to lose a lot of leaf material during hay-making (Mortimer et al., 2006), but the protein content of red clover hay is still high relative to good quality grass hay, e.g. 184 g/kg DM compared to 110 g/kg, though not as high as early-flowering lucerne hay (225 kg DM/kg) (McDonald et al., 2002). As noted above under lucerne, Speijers et al. (2004) showed higher rates of intake with lambs grazing lucerne or red clover compared to perennial ryegrass or bird’s-foot trefoil, and although growth rates were higher for the latter, lambs grazing red clover fattened faster than those grazing perennial ryegrass. Wilkins & Paul (2002, cited in Mortimer et al., 2006) state that red clover silage yield without N would be equivalent to a perennial ryegrass crop receiving 200 kg N/ha but lower than a grass crop fertilized to maximum yield. However, economic modeling by Topp & Doyle (2004) predicted that red clover (either as a monoculture or in a mixture with grass) would outperform all silage systems based on pure grass swards, even those receiving 400 kg N/ha/yr. Ensiling red clover is more difficult than with grass, due to a lower soluble carbohydrate concentration and low dry matter. Forage containing a high proportion of red clover generally requires a longer wilting period than grass and may benefit from chemical additives (Frame et al., 1998).

10 As noted above under lucerne, estimates of the annual rate of biological nitrogen fixation by red clover (125-220 kg N/ha) are slightly lower than those for optimum mixtures of perennial ryegrass and white clover and 25-50% lower than lucerne, but about 1.5-2.0 times higher than those for bird’s-foot trefoil (Frame & Newbould, 1986; Frame et al., 1998). In common with lucerne and white clover, red clover can cause bloat, though the risk is lower than with white clover, and the risk mainly occurs when red clover is present at more than 50% of the sward (Mortimer et al., 2006). However, the oestrogen content of red clover is higher than in the other two legumes, and sheep should be removed from pasture containing significant amounts of red clover for at least three weeks before commencement of the breeding season (Beever et al., 2000). Plant oestrogens may, however, have beneficial effects on carcass protein deposition, although Beever et al. (2000) reported that the quantitative effects had not been fully assessed.

Lotus corniculatus (Birds-foot trefoil) Bird’s-foot trefoil is a perennial hemicryptophyte legume native to the UK with a variable growth habit, from creeping to fairly erect, with the latter more typical of introduced varieties (Stace, 1997). Wild ecotypes flower between June and September (Rose, 2006). Commercial varieties are available for the UK, but these have been produced in other temperate zones of the world where the species is more frequently used (Hopkins, 2004). This species is particularly tolerant of low nutrient conditions (Hopkins et al., 1996) and is a common constituent of semi-natural mesotrophic and particularly of calcareous grassland communities, and of some acid grasslands, but is rarely found in improved and semi-improved grasslands (Rodwell, 1992). It forms a long-term seed bank, germination from which represents its main regeneration strategy (Hodgson et al., 1995). Bird’s foot trefoil is a very valuable insect food plant and Mortimer et al. (2006) listed even more specific insect associations for this species (104) than for red clover. In common with red clover it was among the key forage species whose decline in the countryside has been identified as a major contributing factor to declines in bumblebee populations (Carvell et al., 2001). It is also a very important host-plant for a substantial number of butterfly and moth species and a constituent (either as seed or as leaf material) in the diet of several bird species, though less importantly so than red clover (Collins & Fothergill, 2008). Ease of establishment Lotus corniculatus is a very common constituent in seed mixtures for the restoration of species-rich grasslands, establishing better than most other forbs: Pywell et al. (2003) assigned it a performance index of 0.21 for year 1, although it declined thereafter. In multi-site studies on establishment success (Hopkins 2004), L. corniculatus was influenced by site-specific features including soil drainage status. This species established relatively poorly on clay or wet peaty soils sites where the closely related species greater bird’s-foot trefoil (L. pedunculatus) performed better. Surface sowing of Lotus into an existing sward, an approach commonly used for white clover (T. repens) resulted in a relatively poor establishment compared with using either a partial or full cultivation.

11 Persistence Hopkins (2004) reported a decline in Lotus yield with age in trials where L. corniculatus and L. pedunculatus were sown together at a relatively high seed rate (total 13.5 kg/ha, equal by weight for each species) with companion grasses (Phleum pratense and Festuca pratensis, or P. pratense and Cynosurus cristatus) depending upon site. Lotus pedunculatus persisted in the sward even by year 6 in areas of high rainfall in Devon and Wales whereas L. corniculatus was a minor component by year 5, but on a calcareous clay soil in the Cotwolds reverse was true. In each case, Lotus was more persistent when cut relatively infrequently (3 times per year) compared with either cutting twice or four times a year. There was no evidence that pests or diseases affected Lotus persistence, though this possibility was not investigated formally. In other studies comparing L. corniculatus, red and white clover, perennial ryegrass and lucerne, L. corniculatus showed the greatest reduction in yield under “simulated grazing” (i.e. cut every 30 days, compared to cutting every 50 days), followed by lucerne, whereas the yields of red clover, white clover and perennial ryegrass were not adversely affected (Collins et al., 2006, cited in Collins & Fothergill, 2008). The L. corniculatus cultivar used was of an upright morphological type. Other work at The Institute of Grassland and Environmental Research, Aberystwyth has shown considerable variation in persistence between different lines of L. corniculatus, with lines with a more creeping growth habit considerably more persistent under grazing than more upright varieties (IGER, 2007). As noted earlier, L. corniculatus is particularly common in semi-natural calcareous grassland communities, both those where tight grazing results in an open sward and those characteristic of more lax grazing (Rodwell, 1992). This is consistent with a greater tolerance of deficiencies in soil fertility compared to other legumes, including low P levels and drought stress (Bologna et al., 1996; Hopkins et al., 1996), but also suggests that native ecotypes are better adapted to grazing than commercial varieties. Agronomic value Lotus corniculatus can be reasonably productive, with young L. corniculatus-dominated swards in the first two years after sowing yielding 8-9 t DM/ha/year (Hopkins, 2000), although higher yields (e.g. 13 t/ha in grass mixture and 11 t/ha grown alone) have been reported (Mortimer et al., 2006). Superior silage intake by sheep has been recorded for this species compared with other legumes and Lotus spp. are grazed preferably to perennial ryegrass, equally so with T. repens (Mortimer et al., 2006). Ensiling is difficult due to low soluble sugar content and high buffering capacity and, like red clover, needs more prolonged wilting than grass silage and may benefit form chemical additives (Frame et al., 1998). Forage of L. corniculatus has a relatively high digestibility, about equivalent to red clover, and its nutritive value is generally as high as or higher than that of lucerne (Mortimer et al., 2006). Barber (1985) quotes a crude protein content of 20.0% in June and 18.2% in July - equivalent to fresh red clover at the early flowering stage or as hay (McDonald et al., 2002) - with high levels of Na in May-June compared to the 11 other forbs listed and good levels of Ca (about average for the 12 forbs). Levels of P were low, however, possibly reflecting the low soil P availability typical of soils where L. corniculatus is most common. Lambs grazing L. corniculatus from September onwards fattened faster than those grazing lucerne, red clover or perennial ryegrass, although forage protein content

12 declined faster for Lotus than either red clover or perennial ryegrass and Lotus voluntary intake was slightly, though not significantly, lower than for red clover or lucerne (Speijers et al., 2004). Lotus corniculatus contains a cyanogenic glycoside which can result in the release of hydrocyanic acid when it comes into contact with a hydrolyzing enzyme present in the plant. This can happen when the plant is crushed or when broken down in the rumen (Cooper & Johnson, 1984). Natural populations contain both the glycoside and the enzyme, though the amounts present are genetically determined and in some non- poisonous plants the enzyme may be absent, and reports of poisoning in practice are rare (Cooper & Johnson, 1984). Estimates of annual biological N fixation are lower at 60-138 kg N/ha for L. corniculatus than for red clover (125-220 kg N/ha) and particularly compared to lucerne (300 kg N/ha) and roughly half that produced by a perennial ryegrass-white clover sward (150-250 kg N/ha) (Frame and Newbould, 1986; Frame et al., 1998). Topp & Doyle (2004) predicted that systems based upon L. corniculatus would be less economically viable than pure grass-based systems, although accurate modelling was hampered by gaps in information. The major nutritional advantage of L. corniculatus, in addition to the high digestibility and protein content it shares with other forage legumes, is the relatively high concentration of condensed tannins in its aerial parts that not only confer protection to livestock against bloat but can also improve dietary protein metabolism (Barry & McNabb, 1999; Aerts et al., 1999). There is also evidence that condensed tannins can reduce the worm burden in livestock (Molan et al., 1999). The ability of L. corniculatus to remain fairly productive under conditions of comparative water stress and low soil fertility, particularly of P, suggests a niche for this species in comparison with swards based upon white clover (Collins & Fothergill, 2008)

Silene vulgaris (Bladder campion) Silene vulgaris is a short-lived perennial hemicryptophyte. There are two sub-species of S. vulgaris – ssp. vulgaris is native throughout to most of Britain, although rare in Wales, Scotland and Ireland; whereas Ssp. macropcarpa is much less common, being introduced from the Mediterranean and naturalized only in south Devon (Stace, 1997). All information given hereafter relates to Silene vulgaris Ssp.vulgaris. Silene vulgaris flowers between June and August (Rose, 2006) and forms a long-term seed bank, germination from which is its main regeneration strategy (Hodgson et al., 1995). It is gynodioecious, i.e. it has female and hermaphrodite flowers on separate plants. Much of the scientific interest in this species relates to the tolerance of some of its ecotypes to metallic elements, particularly Cu, Zinc and cadmium (Verkleij et al., 2001), and also to genetic aspects of its biology and ecology (Bernasconi et al., 2009). Silene vulgaris is attractive to a range of pollinating and seed-eating insects, particularly moths and long-tongued bees (Petterssen, 1991; Kephart et al., 2006; Anon, 2009b). In the Eastern and Midwestern states of the USA at least, the seeds are not considered an important source of food to birds and the ecological value of this plant to wildlife is considered to be to be low (Anon, 2009b), although this might not translate to the UK situation.

13 Ease of establishment This species is rarely included in seed mixtures in the UK (probably because it is short- lived) and information specifically on establishment performance in seed mixtures is limited. According to commercial information, S. vulgaris is easily grown from seed sown at any time of the year (Emorsgate Seeds, 2009), although it apparently has an after- ripening requirement for optimal germination (Hodgson et al., 1995). Rapid germination (within two weeks), but low germination percentage has been reported (Charlesworth & Laporte, 1998), whilst in study of 16 different families germination varied from 50-95% between families, averaging 73% overall (Andersson-Ceplitis & Bengtsson, 2002). Germination percentage tends to be slightly higher in seed from out-crossed plants compared to self-fertilized ones, but with a much more marked difference in terms of seedling survival (about 90% compared to about 75% under artificial conditions) (Bailey & McCauley, 2006). Silene vulgaris is reported as having established successfully when sown in seed mixtures on new road verges in Western Norway (Nordbakken et al., 2010). When included as a minor component in a seed mixture containing 31 grasses, legumes and herbs in a species-rich grassland restoration study in the Czech Republic it established poorly initially but by the third year after sowing its establishment was classed as good (Šrámek & Kašparová, 2005). Establishment is characterized by a rapid relative growth rate initially followed by rapid floral development, with little investment in root growth until after flowering; root growth occurs continually during the second year of growth, including during floral development, but with lower relative growth rate of aerial parts (Wall & Morrison, 1990). Persistence Silene vulgaris is described as a short-lived perennial by various sources, but no specific information on the longevity of individual plants could be found. As noted above, the species regenerates primarily from the seed bank, so that long-term persistence after introduction into habitats where it is not already present will presumably depend upon management that allows it to set seed. Nevertheless, at least one source reports that S. vulgaris can be grown artificially from crown buds (Wall & Morrison, 1990). Silene vulgaris appears to be tolerant of grazing. Where the species is a typical component of species-rich mountain meadow communities in the Czech Republic, these meadows were traditionally managed by cutting for hay and grazing the aftermath (Krahulec et al., 2001). However, S. vulgaris was among a group of species that were increased by the introduction of sheep grazing in previously abandoned meadows (Krahulec et al., 2001). In other studies in mountain meadows in the Czech republic, Silene vulgaris was selected by grazing sheep but recovered well after grazing, so showed no trend in abundance in response to grazing over a three-year period (Hejcman et al., 2008). Agronomic value Information on the agronomic value of S. vulgaris is limited. As noted above, S. vulgaris is apparently palatable to sheep and is grazed selectively but with good recovery after grazing (Hejcman et al., 2008). Analyses of mineral content showed S. vulgaris to be exceptionally high in both Mg (0.52% dry matter) and K (5.14%) content; of 25 species analyzed in Crete, S. vulgaris was highest in Mg and second highest in K (Zeghichi et al., 2003). Values for Mn were

14 also relatively high but, as with H. radicata, values for Zn were low, though not particularly low compared to the other 24 species sampled (Zeghichi et al., 2003).

Knautia arvensis (Field scabious) Knautia arvensis is a perennial hemicryptophyte with a fairly erect posture, growing to a height of up to 1m, native to Britain and widespread there, though less common in Northern Ireland and North and West Scotland (Stace, 1997). It flowers between July and September (Rose, 2006). It occurs mainly in grassland and grassy places on light well-drained soils. It is a common constituent of unimproved dry and calcareous grassland communities, particularly those where grazing is lax, but is normally absent from improved or semi-improved grasslands (Rodwell, 1992). K. arvensis forms a transient seed bank, although work in Norway shows that a proportion of seeds will survive for more than two years (Vandvik & Vange, 2003 – see below under Ease of establishment). Its primary regeneration strategy is by germination of recently shed seed (Hodgson et al., 1995) although it also produces underground stolons (Vandvik & Vange, 2003). The flowers of K. arvensis attract a variety of pollinating insects and the species was identified as particularly attractive to bumblebees in chalk grassland (Lack, 1982). However, other information suggests that, where seed mixtures are sown on ex-arable land at least, K. arvensis is not as attractive as most other species sown, although apparently selected more by short-tongued bumblebees than long-tongued (Carvell et al., 2001). This may be at least partly due to poor establishment and low availability of flower heads of this species, at least in the first year after establishment (Carvell et al., 2006). However, other studies showed that, despite its low abundance compared to other important nectar sources, K. arvensis was a key attractor of several butterfly species in arable field margins sown to a wild flower seed mixture (Feber et al., 1996). Ease of establishment Pywell et al. (1999) showed relatively poor germination (41.6%) under alternating temperature of 22/18 ˚C and under 16/8 hours light/dark regime. More detailed studies in Norway showed little response to light or alternating temperatures, and a response to cold stratification under artificial conditions did not appear to be supported by field observations, although there was apparent ecotypic variation related to altitude and proximity to the coast (Vandvik & Vange, 2003). A proportion of the seeds remained un- germinated for 32 months and the half-life was estimated at 17 months (15-20 months, varying with geographic region). Seedling survival was not affected by microsite size, suggesting that seedlings survive by tolerating competition rather than avoiding it (Vandvik & Vange, 2003). The protracted germination behaviour of K. arvensis probably explains the very poor performance of this species in the meta-analysis of restoration studies (Pywell et al., 2003). Although Knautia seedlings appear able to tolerate competition in the species natural habitats (Vandvik & Vange, 2003) they may not be able to do so under more fertile conditions. Persistence This species may not persist and spread easily when sown in relatively fertile soils and where the growth of its companions is vigorous, since the main mode of regeneration for this species is by seed dispersal (Hodgson et al, 1995) and it appears to lack dormancy

15 and germination mechanisms that would allow it to exploit gaps in vegetation (Vandvik & Vange, 2003). Information gained from the introduction of pot-grown or plug plants in the restoration of species diversity to improved grassland suggests that once established in such situations K. arvensis persists well over a period of five years (Davies, 2002). Pot plants survived better than plugs and the survival of K. arvensis was good compared to most of the other species tested. There was no effect of cutting twice in spring and three times in autumn in addition to a hay cut in July, compared to normal July hay cutting with spring and aftermath grazing either by sheep or dairy cows, although all three treatments included a July hay cut (Davies, 2002). The usual absence of Knautia arvensis from plant communities associated with tight grazing (Rodwell, 1982) suggests an intolerance to hard grazing or frequent defoliation and the confinement of the species to infertile unimproved grasslands suggests a lack of competitiveness under fertile conditions, although the species is classed as CSR (Hodgson et al., 1995), i.e. occupying an intermediate position in terms of all three established strategies: competitiveness, stress-tolerance and ruderality (sensu Grime, 1979). It seems likely that the species will persist in well-drained soils of low-moderate fertility as long as management allows it to set and disperse seed at least occasionally. Agronomic value There is little information available on the agronomic value and yield potential of K. arvensis. Analyses of the chemical composition of shoots harvested at the commencement of flowering in Swiss alpine meadows (Lebreton et al., 2001) show relatively low levels of P, K and Ca (0.23%, 1.33% and 0.99% respectively) compared to values shown for other candidate species presented above. The values for P and K were also below average for the18 forbs tested in the same study (0.32% and 2.38% respectively), whilst the value for Ca was slightly above the average for all 18 species of 0.95%. However, the value for P for Knautia was equivalent to that of 0.23% quoted for Achillea millefolium in the same analyses (Achillea being the only other of the present candidate species that was included) which was lower than the value for Achillea of 0.40 reported by Barber (1985), whilst the Ca values for Achillea were similar in the two analyses (1.08% compared to 1.13%). The crude protein value for Knautia was also particularly low at 10.8% compared to the average of 14.8% for all 18 species - the latter figure being more typical of forbs in general (Barber, 1985). The crude protein value for Knautia was nevertheless higher than the value of 9.1% given for Achillea in the same analyses. Note, however, that samples in the Lebreton et al. (2001) analyses were taken at an equivalent growth stage for each species, i.e. at initiation of flowering. In the UK at least, this would be July for Knautia and probably June for Achillea, though exact timing varies between geographical locations. Though this might not entirely explain the apparent discrepancies between results from different sources it does highlight the difficulties of making such comparisons. For Knautia it is probably fair to say that it is a reasonable source of Ca and a poor source of both P and crude protein. Although in the case of Ca and P at least this might reflect the calcareous soils on which the species most commonly grows, only one of the five sites from which it was sampled appeared to be on calcareous soil, judging by the pH values given (Lebreton et al., 2001).

16 There is evidence that inclusion of K. arvensis or its extracts in the diet of ruminants might improve N use efficiency by enhancing protein degradation (Hoffmann et al. 2007).

Hypochaeris radicata (Cat’s-ear) Hypochaeris radicata is a fairly short-lived perennial rosette-forming hemicryptophyte. It is native to Britain and very widely distributed in grassland and grassy places in the British Isles and in many other temperate regions of the world (Turkington & Aarssen, 1983; Stace, 1997). It flowers between June and September (Rose, 2006). It produces a very wiry flowering stem <60cm in height that is very resistant to grazing damage and can frequently survive mowing, although maximum reproductive weight occurs under lax grazing (Turkington & Aarssen, 1983). H. radicata shows a relatively modest number of specific insect associations (28) compared to other forbs, and a similar number of mono-specific associations to S. jacobaea and C. arvense (6); it is unremarkable as a direct or indirect source of bird food (Mortimer et al., 2006). Ease of establishment H. radicata seeds have no innate dormancy, although light is required for germination, and Grime et al. (1981) showed more than 90% germination of freshly collected seeds, 50% of which germinated within 14 days with no effect of dry storage for up to 6 months, though germination was reduced to 31% after 12 months. Ho (1964, cited in Turkington & Aarssen, 1983) found poorer germination of both fresh seed (68%) and seed stored dry for one month (58%) and only 4% after two months of dry storage. Field experiments have shown variable establishment success in a range of artificially created microsites, and that H. radicata established particularly poorly in drier and more exposed microsites (Oomes & Elberse, 1976, cited in Turkington & Aarssen, 1983). This species was included in seed mixtures in nearly two-thirds of the studies reviewed by Pywell et al. (2003). Initial establishment tended to be fairly poor compared to the other candidate species Achillea millefolium and L. corniculatus, although its performance was comparable to the latter species by the third year after sowing. Persistence Individual plants are relatively short-lived, with longevity quoted as between 2-5 years (Botanic Gardens Trust, undated) and no more than 10 years (Soons & Heil, 2002), but can reproduce vegetatively from buds in the root crown to produce clones (Turkington & Aarssen, 1983). Plants can also develop very quickly from seed, producing flowering plants within two months under favourable conditions (Ho, 1964, cited in Turkington & Aarssen, 1983). H. radicata forms only a transient seed bank (Thompson et al., 1997) and is very dependant upon seed dispersal to maintain populations (Hartemink et al., 2004) since it tends to exist as isolated individuals or small groups, possibly due to autotoxicity (Newman & Rovira, 1975; Turkington & Aarssen, 1983). Its allelopathic properties may help to explain its ability to persist in the face of competition from other species (Turkington & Aarssen, 1983), including in semi-improved and improved grasslands (Rodwell, 1992). H. radicata tends to increase once established from seed mixtures, at least until the third year after sowing (Pywell et al., 2003). It is a common component of a wide range of permanent grasslands in the UK, though seldom more than a minor component (Rodwell, 1992).

17 In grassland it is associated primarily with those that are frequently cut or grazed, particularly in more fertile conditions, and is intolerant of shade (Turkington & Aarssen, 1983). Agronomic value Milton (1933) identified Hypochaeris as one of the more palatable to livestock (of unspecified type) of the “miscellaneous herbs” of lowland grassland, with animals apparently selecting the species, particularly in winter. Values for crude protein during the summer in Britain, New Zealand and the USA North- Pacific region fall within the range from 9-14% (Turkington & Aarssen, 1983; Partridge et al. 2001). This is somewhat lower than those for L. corniculatus, R. obtusifolius and C. arvense (Barber, 1985), but equivalent to several other forbs described by Barber (1985). A value of 11.7% for vegetation in the North-Pacific region (Partridge et al., 2001) was low compared to Trifolium repens and for grasses (unspecified) sampled at the same time (17.4% and 16.8% respectively), as was the value for gross energy (4.3 kcal/g compared to 4.4 and 4.5 kcal/g). However, apparent digestible energy (ADG) was intermediate between T. repens and grasses, although closer to T. repens (i.e. 33.9 kcal/g compared to 40.0 and 18.4 kcal/g for T. repens and grasses respectively), due to low total dietary fibre of both H. radicata (50.5% of DM) and T. repens (45.2%) compared to grasses (63.9%). Analyses of leaf material of H. radicata plants sampled in Crete (Zeghichi et al., 2003) show relatively high values for Ca (1.79% dry matter), Mg (0.35%), Na (1.70%) and Mn (100 mg kg-1). Most of the values fall within the ranges quoted for this species from sources in New Zealand and Britain (Turkington & Aarssen (1983), with the exception of K, Zn and Cu. Values from Zeghichi et al. (2003) for K and particularly of Zn fall below the ranges quoted by Turkington & Aarssen (i.e. 2.1-4.3% and 19-41 mg/kg respectively) and are also low compared to those for other candidate species (Barber, 1985). Zeghichi et al.’s value for Cu (20.5) is considerably higher than the Turkington & Aarssen range of 7.2-7.3 mg/kg. However, some caution is needed in comparing data from different sources. For example, the values for Cu quoted by Turkington & Aarssen (1983) were taken from fewer samples than those for other species, and also seem low compared to other candidate species (Barber, 1985). On the other hand, whilst the values from Zeghichi et al. (2003) for Zn of both H. radicata and S. vulgaris seem low, none of the 25 species analysed by them exceeded 70.1 mg/kg Zn, although the value of 15.7 for H. radicata was in fact the lowest of all the species.

Potentilla anserina (Silverweed) Potentilla anserina is a perennial hemicryptophyte with long creeping stolons bearing terminal leaf rosettes, native to Britain and widespread in waysides, waste places, pastures and sand dunes (Stace, 1997). It flowers between June and August (Rose, 2006). P. anserina is a common constituent of several unimproved and semi-improved mesotrophic, mostly grazed, grassland communities but is most prominent in Festuca- rubra-Agrostis stolonifera-Potentilla anserina (MG11) grassland (Rodwell, 1992). This grassland type is characteristic of moist but free-draining soils that are frequently inundated with fresh or brackish surface water, especially frequent near sea level (Rodwell, 1992).

18 P. anserina forms a short-term seed bank (1-5 years) (Thompson et al., 1997) and its main regeneration strategy is by vegetative growth (Hodgson et al., 1995) with an unusual ability to “find” micro-areas of low surrounding vegetation density (Eriksson, 1986a). No review could be found of the number of insect or bird associations for this species, although it is very attractive to bees (Denisow & Vrzesień, 2007) and has been listed as a valuable food plant for butterfly larvae (Smart et al., 2000). Ease of establishment This species is seldom if ever included in seed mixtures and little information is available on its germination characteristics or ease of establishment. Seeds collected from a Californian salt marsh germinated most effectively after soaking in fresh water for 24 hours and storage at 4˚C for 30 days (Heimbinder, 2001), suggesting that germination will occur mainly in spring under natural conditions. Persistence A. anserina reproduces mainly by stoloniferous spread and the formation of daughter ramets, but also produces seedlings. However, seedlings remain juvenile for at least four years, whereas daughter ramets are capable of both flowering and stolon production in the year following establishment (Eriksson, 1986b). Increased intensity of competition increases the plants allocation to flower production as opposed to vegetative spread (Rautiainen et al., 2004). The mean longevity of clones of P. anserina has been estimated at 21 years in alpine fellfields in Colorado (Forbis & Doak, 2004) although a population turnover of 32 years was recorded in a Baltic seashore meadow (Eriksson, 1986b). Sexual reproduction is relatively unimportant in the maintenance of populations, due to a high mortality rate in seedlings (Eriksson, 1986b), but when competition stress is high seed dispersal and seedling establishment becomes more significant (Rautiainen et al., 2004). Potentilla anserina increased under grazing management in coastal meadows in Finland (Jutila, 1999), possibly due to avoidance by grazing animals (Milton, 1933). Agronomic value The small amount of information found on the agronomic value of P. anserina suggests it to be low. It is appears to be unpalatable to livestock - almost none of the small amounts present in either the lowland grasslands or the cereal stubbles that Milton (1933) surveyed was grazed by livestock. Information presented in relation to wetlands in Idaho identifies P. anserina as low in both protein and energy value when sampled between flowering and the following early spring compared to other forbs present, although no values were given (Hall & Hansen, 1997). However, work at the Rowett Institute of Nutrition and Health suggests that, due to its concentration of phenolic compounds, it may have potential for controlling rumen proteolysis, thereby increasing protein utilization efficiency (RINH, 2008).

Achillea millefolium (Yarrow) Achillea millefolium is a perennial semi-rosette forming hemicryptophyte native to Britain and very widely distributed both here and elsewhere (Stace, 1997). It is a common constituent of a range of acid, calcareous and mesotrophic permanent grassland communities, though not in wetter communities or those subject to seasonal flooding (Rodwell, 1982). It flowers between June and August (Rose, 2006).

19 A. millefolium is a valuable food source for insects, with 74 specific and 13 monospecific associations identified by Mortimer et al. (2006). Only nine species of the 56 species reviewed showed more specific associations, and only two showed more monospecific associations. Ease of establishment A. millefolium appears to be easy to establish in seed mixtures. It can germinate in spring and autumn, requiring a temperature of 18-25 ˚C and exposure to light, so that germination is inhibited by burial deeper than 0.5 cm or so (Warwick & Black, 1982; Thompson, 1989). This species was sown in 19 of the 25 studies included in the meta- analysis of species establishment and persistence (Pywell et al., 2003). It was one of the most successfully establishing species in that analysis, second only to Leucanthemum vulgare among the forbs, and superior to more than half the grasses included (Pywell et al., 2003) - and considerably more successful than the remainder of the present candidate species that were included. Persistence In the Pywell et al. (2003) meta-analysis, it not only established well but also showed a highly positive trend in performance, implying good persistence and spread, at least over the four years following sowing. Hodgson et al. (1995) identify this species as having no seed bank persistence, although other sources suggest that its persistence may have been underestimated (Warwick & Black, 1982). However, it forms stolons from which it regenerates effectively, whilst it also regenerates from recently sown seed (Hodgson et al., 1995). A combination of these characteristics presumably explains its colonising success, since individual plants are relatively short-lived (Warwick & Black, 1982). A. millefolium tolerates grazing, although summer grazing reduces seed production (Mortimer et al., 2006). This tolerance is partly due to the plants ability to modify it’s morphology in response to grazing by forming a more prostrate rosette structure (Herms & Mattson, 1992, cited in Mortimer et al., 2006). Trampling by livestock may also cause the fragmentation of rootstock thereby stimulating the production of daughter plants, but A. millefolium does not compete well in tall foliage where it becomes excluded (Mortimer et al., 2006). Agronomic value Palatability studies in Welsh lowland grasslands suggest that A. millefolium is eaten by livestock all year round, though it is apparently more palatable in autumn and winter than in spring and summer (Milton, 1933). The crude protein content of A. millefolium is moderate compared to a wider range of forbs (Barber, 1985). Whilst digestibility in May is about average for forbs the subsequent decline appears to be smaller (Barber, 1985), although other data suggest a marked decline between May and June (Isselstein, 1993, cited in Mortimer et al., 2006). Ca and P levels are moderate to good compared to other forbs and Cu levels are higher than most (Barber, 1985), whilst the content of these and most other minerals is higher for Achillea than for grasses (Thomas et al., 1952; Trzasko, 1994, cited in Mortimer et al., 2006). A. millefolium ensiles well compared to Lolium perenne and is as productive in monoculture (Isselstein, 1993, cited in Mortimer et al., 2006), but makes less good hay due to difficulty with drying (O’Beirne-Ranelagh, 2005).

20 Primula veris (Cowslip) Primula veris is a perennial hemicryptophyte native to Britain although rare in much of the north, found in grassy places usually on light calcareous or base-rich soils (Stace, 1997). It is early flowering, normally April-May (Rose, 2006). P. veris has a chilling requirement for germination but shows no dormancy beyond the spring after sowing and its main regeneration method is vegetative, through the production of lateral buds (Tamm, 1972; Hodgson et al., 1995). Very few (4) specific insect associations were identified for P. veris, none of which were monospecific, it appears to have no value as a direct food source for birds and its value as an indirect food source is very limited (Mortimer et al., 2006). Ease of establishment Though frequently included in seed mixtures sown for grassland restoration or in arable field margins, establishment is generally poor (Pywell et al., 2003), often completely failing to establish (e.g. Kirkham et al., 1999; De Cauwer, 2005). If established, however, it tends to increase over the first 3-4 years at least (Pywell et al., 2003). This poor establishment performance is consistent with its lack of two of the properties relating to colonizing ability: autumn germination and a persistent seed bank (Pywell et al., 2003). However, Milberg (1994) showed that seeds did not germinate in darkness and therefore could become dormant again in late spring and early summer and dormancy could be broken again in the following winter. Persistence Seedling emergence and survival is generally low in established vegetation, but vegetative reproduction through side rosettes sometimes occurs (Tamm, 1972; Hodgson et al., 1995). Leaf removal by grazing in early spring reduces current seed production and subsequent growth and grazing at the fruit production stage can affect performance in the following year (García & Ehrlén, 2002; Brys et al., 1994). Growth and persistence is also reduced by lack of defoliation (Brys et al., 1994). And whereas mowing in mid July, or commencement of grazing then, increases flowering and seed set compared to grazing from early spring, mowing in October is the most beneficial because the resulting low vegetation density enhances seedling establishment in the following spring (Brys et al., 1994). Once established, and under favourable conditions, individual adult plants can live for several decades (Inghe & Tamm,1988 and Ehrlén & Lehtilä, 2002, both cited in Brys et al., 2004). Agronomic value No relevant information could be found on the agronomic value of P. veris, other than that it is little grazed by livestock, probably due to its prostrate growth habit (Mortimer et al., 2006).

Additional species Trifolium repens has shallower rooting depth (0-20cm) than other species selected but has been shown to be beneficial to soil structure in direct comparison with Lolium perenne (Mytton et al., 1993; Holtham et al., 2007). It is also highly tolerant of trampling (Roovers et al., 2004) and its high agronomic value is widely documented (e.g. Beever et al., 2000; Rochon et al., 2004; Mortimer et al., 2006).

21 Cichorium intybus is only recorded by Hill et al. (1999) as being associated with linear and boundary habitats and not with any grassland Broad Habitats. Neither is it recorded in any grassland communities in the National Vegetation Classification (Rodwell et al., 1992) but Grime et al. (2007) do record it as ‘widespread in pasture’. It has a deep tap root according to the Ecoflora database although no actual rooting depth data have been found. Under laboratory conditions, its penetrability of compacted soil was inferior to M. sativa but in field experiments it was equal to M. sativa and superior to T. pretense (Löfkvist, 2005; Löfkvist et al., 2005). It has a high mineral content and high value for forage production, especially under rotational grazing in dry summer conditions, and the first commercial cultivar was approved in New Zealand in 1985 (Barry, 1998). It is now sometimes introduced into grasslands for its agronomic value as it can increase productivity and forage quality (e.g. Høgh-Jensen et al., 2006). Centaurea nigra is tap-rooted and has mycorrhizal associations (Grime et al., 2007) but no other data were found on its root system. It has a moderately high establishment/persistence index (Pywell et al., 2003) and high biodiversity value but probably low feed value (Mortimer et al., 2006). Plantago lanceolata has a tap root but shallower rooting depth (0-10cm) than other forbs identified. It has a high establishment/persistence index (Pywell et al., 2003) and medium feed value (Mortimer et al., 2006). No data on rooting characteristics for Sanguisorba minor ssp. muricata were found. However, ssp. minor has a tap root of depth 50-100cm (Ecoflora database) and mycorrhizal associations (Grime et al., 2007). This suggests that ssp. muricata might have the same root properties, which could be beneficial for alleviating soil compaction. In New Zealand, S. minor ssp. muricata produces a useful mass of forage, comparable to Lotus corniculatus and Medicago sativa (Douglas et al., 1990). It also responds to successive defoliation by increasing number and area of leaves and can adjust to water depletion (Douglas et al., 1994). It has also been shown to produce good quality forage for sheep in the French Mediterranean (Viano et al., 1999).

Conclusions Based upon the foregoing reviews, the candidate species are grouped below into those that appear suitable for inclusion in experimental seed mixtures for alleviating soil compaction, those whose suitability is limited to certain soil types, and those that are unsuitable.

Suitable species Of the Fabaceae, T. pratense and L. corniculatus are tap-rooted, have mycorrhizal associations, and high agronomic and biodiversity value. T. pratense may be slow to establish, although commercial varieties may perform better in mixtures than the native varieties normally sown in biodiversity mixtures. Both species prefer periodic defoliation as opposed to continuous grazing. A. millefolium does not have a tap root but does have mycorrhizal associations, is a common constituent of biodiversity seed mixtures and establishes and persists well. It is a very valuable food source for insects. It tolerates grazing and, as a source of minerals is about average for forbs and superior to most grasses. H. radicata has a tap-root and mycorrhizal associations although less easy to establish than the above species and needs to be able to disperse seeds in order to persist. It is

22 reasonably good source of minerals and is apparently selected by grazing animals, but is of modest wildlife value. Trifolium repens, Cichorium intybus, Centaurea nigra, Plantago lanceolata and Sanguisorba minor ssp. muricata were not included in the original selection of species using published trait data but are likely to be suitable for inclusion in experimental seed mixtures.

Species whose suitability is limited to certain soil types M. sativa is tap-rooted, has mycorrhizal associations, high agronomic value and moderate biodiversity value. It has a deep root system but it is not certain how it would perform in a biodiversity mixture, and is likely to be most appropriate for calcareous soils. M. sativa ssp. sativa is a non-native, commercial crop plant so its long-term persistence in a semi-natural sward would not be desirable. It prefers periodic defoliation as opposed to continuous grazing. K. arvensis is also tap-rooted and with mycorrhizal associations but is generally difficult to establish in seed mixtures and likely to be useful only in infertile dry and/or calcareous soils where competition is slight. There is little detailed information on its agronomic or wildlife food value. It appears to be a reasonable source of calcium for livestock but below average both for other minerals - especially P - and for crude protein. It is apparently attractive to butterflies and short-tongued bumblebees.

Unsuitable species S. vulgaris has a deep tap root but does not have significant myccorhizal associations. It is seldom used in seed mixtures and information on ease of establishment is somewhat contradictory, but suggests that establishment from seed mixtures may be slow. On the other hand, S. vulgaris is a good source of several minerals, is apparently selected by grazing sheep, and recovers well from grazing. P. veris is considered unsuitable here because it is very difficult to establish in seed mixtures, though it does have mycorrhizal associations and once established can be very long-lived. It has generally low wildlife value and appears to have little or no agronomic value, although information is limited. P. anserina is tap-rooted and has mycorrhizal associations. It has some wildlife value but no significant agronomic value and may be difficult to establish in seed mixtures.

References Atwell, B.J. (1988) Physiological responses of lupin roots to soil compaction. Plant & Soil 111, 277-281. Aerts, R.J., Barry, T.N. & Mcnabb, W.C. (1999) Polyphenols and agriculture: beneficial effects of proanthocyanidins in forages. Agriculture, Ecosystems and Environment 75, 1-12. Andersson-Ceplitis, H. & Bengtsson, B.O. (2002) Transmission rates and phenotypic effects of mitochondrial plasmids and cytotypes in Silene vulgaris. Evolution 56, 1586-1591. Anon (2009a) Tropical forages: Medicago sativa fact sheet. Available at: http://www.tropicalforages.info/key/Forages/Media/Html/Medicago_sativa.htm

23 Anon (2009b) Encyclopedia of life: Silene vulgaris (Moench) Garke. Available at: http://www.eol.org/pages/581986 Bailey, M.F. & McCauley, D.E. (2006) The effects of inbreeding, outbreeding and long- distance gene flow on survivorship in North American populations of Silene vulgaris. Journal of Ecology 94, 98-109. Barber, W.P (1985) The nutritional value of common weeds. In: J.S. Brockman (Ed). Weeds, pests and diseases of grassland and herbage legumes. British Grassland Society Occasional Symposium no. 18, pp104-111. British Grassland Society, Reading, UK. Bardgett, R.D., Smith, R.S., Shiel, R.S., Peacock, S., Simkin, J.M., Quirk, H. & Hobbs, P.J. (2006) Parasitic plants indirectly regulate below-ground properties in grassland ecosystems. Nature 439, 969-972. Barry, T. N. (1998) The feeding value of chicory (Cichorium intybus) for ruminant livestock. The Journal of Agricultural Science 131, 251-257. Barry T.N. & McNabb W.C. (1999). The implications of condensed tannins on the nutritive value of temperate forages fed to ruminants. British Journal of Nutrition 81, 263-272. Bedini, S., Pellegrino, E., Avio, L., Pellegrini, S., Bazzoffi, P., Argese, E. & Giovannetti, M. (2009) Changes in soil aggregation and glomalin-related soil protein content as affected by the arbuscular mycorrhizal fungal species Glomus mosseae and Glomus intraradices. Soil Biology and Biochemistry 41, 1491-1496. Beever, D.E., offer, N., & Gill, M. (2000) The feeding value of grass and grass products. In: A. Hopkins (ed). Grass: Its Production and Utilization, pp140-195. Blackwell Science, Oxford, UK. Bernasconi, G., Antonovics, J., Biere, A., Charlesworth, D, Delph, L.F., Filatov, D., Giraud, T., Hood, M.E., Marais, G.A.B., McCauley, D., Pannell, J.R., Shykoff, J.A., Vyskot, B., Wolfe, L.M. & Widmer, A. (2009) Review: Silene as a model system in ecology and evolution. Heredity, 1–10. Bologna, J.J., Rowarth, J.S., Fraser, T.J. & Hill, G.D. (1996) Management of bird’s foot trefoil (Lotus corniculatus L.) pastures for productivity and persistence. Proceedings Agronomy Society of New Zealand 26, 17-21. Botanic Gardens Trust (undated) Hypochaeris radicata. Available at: www.rbgsyd.nsw.gov.au/science/Research/Ecology_of_Cumberland_Plain_Woodl and/woodland_plants Broersma, K., Robertson, J.A. & Chanasyk, D.S. (1997) The effects of diverse cropping systems on aggregation of a Luvisolic soil in the peace river region. Canadian Journal of Soil Science 77, 323-329. Brys, R., Jacqemyn, H., Endels, P., De Blust, G. & Hermy, M. (2004) The effects of grassland management on plant performance and demography in the perennial herb Primula veris. Journal of Applied Ecology 41, 1080-1091. Bushamuka, V.N. & Zobel, R.W. (1998) Differential genotypic and root type penetration of compacted soil layers. Crop Science 38, 776-781.

24 Carvell, C., Pywell, R.F., Smart, S.M. & Roy, D. (2001) Restoration and management of bumblebee habitat on arable farmland: literature review. DEFRA Contract Report BD1617. Centre for Ecology and Hydrology, Monks Wood, Cambridgeshire, UK. Carvell, C., Westrich, P., Meek, W.R., Pywell, R.F & Nowakowski, M. (2006) Assessing the value of annual and perennial forage mixtures for bumblebees by direct observation and pollen analysis. Apidologie 37, 326–340. Charlesworth, D. & Laporte, V. (1998) The male-sterility polymorphism of Silene vulgaris: analysis of genetic data from two populations and comparison with thymus vulgaris. Genetics 150, 1267–1282. Collins, R.P. & Fothergill, M. (2008) Potential for greater use of forage legumes in UK grassland systems for improved biodiversity, soil quality, and livestock nutrition and health. Final report for Defra project IF0122. Available at: http://randd.defra.gov.uk/Default.aspx?Menu=Menu&Module=More&Location=Non e&Completed=1&ProjectID=14942 Cooper, M.R. & Johnson, A.W. (1984) Poisonous plants in Britain and their effects on animals and man. Her Majesty’s Stationary Office. London, UK. Cresswell, H.P. & Kirkegaard, J.A. (1995) Subsoil amelioration by plant-roots – the process and the evidence. Australian Journal of Soil Science 33, 221-239. Da Silva, R.H. & Rosolem, C.A. (2002) Soybean root growth as affected by previous crop and soil compaction. Pesquisa Agropecuaria Brasileira 37, 855-860. Da Silva, R.H. & Rosolem, C.A. (2003) Early development and nutrition of cover crop species as affected by soil compaction. Journal of Plant Nutrition 26, 1635-1648. Davies, A. (2002) The floristic restoration of agriculturally improved grassland. PhD Thesis. University of Sheffield. Relevant chapter available at: http://www.karisgarden.com/lusty/Chapter7.pdf De Cauer, B. (2005) Biodiversity and agro-ecology in field margins. PhD Thesis. University of Gent, The Netherlands. Available at: https://biblio.ugent.be/input/download?func=downloadFile&fileOId=490084&record OId=471523 Denisow, B. & Wrzesień, M. (2007) The anthropogenic refuge areas for bee flora in agricultural landscape. Acta Agrobotanica 60, 147–157. Dexter, A.R. (1991) Amelioration of soil by natural processes. Soil & Tillage Research 20, 87-100. Douglas, G.B., Robertson, A.G., Chu, A.C.P., Gordon, I.L. (1990) Establishment and growth of sheep's burnet in the lower North Island of New Zealand. New Zealand Journal of Agricultural Research 33, 385-394. Douglas, G. B., Robertson, A. G., Chu, A. C. P. & Gordon, I. L. (1994) Effect of plant age and severity of defoliation on regrowth of sheep's burnet during substrate moisture depletion. Grass and Forage Science 49, 334–342. Douglas, J.T., Koppi, A.J. & Moran, C.J. (1992) Changes in soil structure induced by wheel traffic and growth of perennial grass. Soil & Tillage Research 23, 61-72. Emorsgate Seeds (1009) Silene vulgaris – Bladder campion. Available at: http://wildseed.co.uk/species/view/131

25 Eriksson, O. (1986a) Mobility and space capture in the stoloniferous plant Potentilla Anserina. Oikos 46, 82-87. Eriksson, O. (1986b) Survivorship, reproduction and dynamics of ramets of Potentilla anserina on a Baltic seashore meadow. Plant Ecology 67, 17-25. Feber, R.E., Smith, H. & Macdonald, D.W., (1996) The effects on butterfly abundance of the management of uncropped edges of arable fields. Journal of Applied Ecology 33, 1191 – 1205. Fitter, A.H. & Peat H.J. (1994) The Ecological Flora Database. Journal of Ecology 82, 415-425. Forbis, T.A. & Doak, D.F. (2004) Seedling establishment and life history trade-offs in alpine plants. American Journal of Botany 91, 1147–1153. Frame J. & Newbould P. (1986). Agronomy of white clover. Advances in Agronomy 40, 1-88. Frame J., Charlton J.F.L. & Laidlaw A.S. (1998) Temperate Forage Legumes. CAB International, Wallingford, UK. García, MB & Ehrlén, J. (2002) Reproductive effort and herbivory timing in a perennial herb: fitness components at the individual and population levels. American Journal of Botany 89, 1295–302. Głąb, T. (2008) Effects of tractor wheeling on root morphology and yield of lucerne. Grass & Forage Science 63, 398-406. Goicoechea, N., Merino, S. & Sánchez-Díaz, M. (2005) Arbuscular mycorrhizal fungi can contribute to maintain antioxidant and carbon metabolism in nodules of Anthyllis cytisoides L. subjected to drought. Journal of Plant Physiology 162, 27-35. Grime, J.P. (1979) Plant strategies and vegetation processes. Wiley, Chichester, UK. Grime, J.P., Hodgson, J.G. & Hunt, R. (2007) Comparative plant ecology: a functional approach to common British species. Second Edition. Castlepoint Press, Dalbeattie, UK. Grime, J.P., Mason, G., Curtis, A.V., Rodman, J., Band, S.R. Mowforth, M.A., Neal, A.M. & Shaw, S. (1981) A comparative study of the germination characteristics in a local flora. Journal of Ecology 69, 1017-1059. Guil, J. L., Torija, M. E., Giménez, J. J., Rodríguez-Garcia, I. & Giménez, A. (1996). Oxalic acid and calcium determination in wild edible plants. Journal of Agricultural and Food Chemistry 44, 1821-1823. Hall, J.B. & Hansen, P.L. (1997) A preliminary habitat type classification system for the Bureau of Land Management districts in Southern and Eastern Idaho. Technical Bulletin 97-11. Idaho Bureau of Land Management. Available at: http://www.blm.gov/id/st/en/info/publications/technical_bulletins/TB_97-11.html Hartemink N., Jongejans, E. & De Kroon, H. (2004) Flexible life history responses to flower and rosette bud removal in three perennial herbs. Oikos105, 159-167. Heimbinder, E. (2001) Revegetation of a San Francisco coastal marsh. Native Plants Journal 2, 54-59.

26 Hejcman, M., Žáková, I., Bílek, M., Bendová, P.,Hejcmanová, P., Pavlů. V. & Stránská, M. (2008) Sward structure and diet selection after sheep introduction on abandoned grassland in the Giant Mts, Czech Republic. Biologia 63, 506—514. Høgh-Jensen, H., Nielson, B. & Thamsborg, S.M. (2006) Productivity, quality, competition and facilitation of chicory in ryegrass/legume-based pastures under various nitrogen supply levels. European Journal of Agronomy 24, 247-256. Hill, M.O., Preston, C.D. & Roy, D.B. (2004) PLANTATT. Attributes of British and Irish Plants: Status, Size, Life History, Geography and Habitats. Centre for Ecology and Hydrology, Monks Wood (updated 2008). Hodgson, J.G., Grime, J.P., Hunt, R. & Thompson, K. (1995) The Electronic Comparative Plant Ecology. Chapman and Hall, London. Hoffmann, E.M., Selje-Assmann, N. & Becker, K. (2007) Dose studies on anti-proteolytic effects of a methanol extract from Knautia arvensis on in vitro ruminal fermentation. Animal Feed Science and Technology 145, 285-301. Holtham, D.A.L., Matthews, G.P. & Scholefield, D.S. (2007) Measurement and simulation of void structure and hydraulic changes caused by root-induced soil structuring under white clover compared to ryegrass. Geoderma 142, 142-151. Hopkins, A. (2004) Research to explore the potential of Lotus. Final report for Defra project LS1307. Available at: http://randd.defra.gov.uk/Default.aspx?Menu=Menu&Module=More&Location=Non e&Completed=0&ProjectID=7712 Hopkins, A. (2000) Herbage production In: A. Hopkins (Ed). Grass: Its Production and Utilization, pp90-108. Blackwell Science, Oxford, UK. Hopkins, A., Martyn, T.M., Johnson, R.H., Sheldrick, R.D. & Lavender, R.H. (1996). Forage production by two Lotus species as influenced by companion grass species. Grass and Forage Science 51, 343-349. Horikawa, Y., Iwabuchi, K. & Ohtsuka, H. (1996) Establishment and yield of Alfalfa (Medicago sativa L.) sward using lime-coated seeds. Grassland Science 42, 211- 215. Hulugalle, N.R., Entwhistle, P.C & Mensah, R.K. (1999) Can Lucerne (Medicago sativa L.) strips improve soil quality in irrigated cotton (Gossypium hirsutum L.) fields? Applied Soil Ecology 12, 81-92. IGER (2007) Developing selection criteria for forage legumes that balance production, biodiversity and reduced environmental pollution. Final report for Defra project LS3643. Availanble at: http://randd.defra.gov.uk/Default.aspx?Menu=Menu&Module=More&Location=Non e&Completed=0&ProjectID=11917 Jones, J.H & Olsen, F.J. (1987) Alfalfa Establishment and Production on Soils with Different Drainage Characteristic. Agronomy Journal 79, 152-154. Jutila, H. (1999) Effect of grazing on the vegetation of shore grasslands along the Bothnian Sea, Finland. Plant Ecology 140, 77-88. Karki, U., Goodman, M.S. & Sladden, S.E. (2009) Nitrogen source influences on forage and soil in young southern-pine silvopasture. Agriculture, Ecosystems & Environment 131, 70-76.

27 Kephart, S., Reynolds, R.J., Rutter, M.T., Fenster, C.B. & Dudash, M.R. (2006) Research review - Pollination and seed predation by moths on Silene and allied Caryophyllaceae: evaluating a model system to study the evolution of mutualisms. New Phytologist 169, 667–680. Kirkham, F.W. (1996) The agricultural ecology of hay meadows within the Somerset Moors and Levels Environmentally Sensitive Area. Unpublished PhD thesis, University of Plymouth, UK. Kirkham, F.W., Sherwood, A.J., Oakley, J.N. & Fielder, A.G. (1999) Botanical composition and invertebrate populations in sown grass and wildflower margins. Aspects of Applied Biology 54, 291-298. Krahulec, F., Skálová, H., Herben, T., Hadincová, V., Wildova, R. & Pecháčková, S. (2001) Vegetation Changes Following Sheep Grazing in Abandoned Mountain Meadows. Applied Vegetation Science 4, 97-102. Krizek, D.T., Ritchie, J.C., Sadeghi, A.M., Foy, C.D., Rhoden, E.G., Davis, J.R. & Camp, M.J. (2003) A four-year study of biomass production of eastern gamagrass grown on an acid compact soil. Communications in Soil Science & Plant Analysis 34, 457-480. Lack, A.J. (1982) The ecology of flowers of chalk grassland and their insect pollinators. Journal of Ecology 70, 773-790. Latif, M.A., Mehuys, G.R., MacKenzie, A.F., Alli, I. & Faris, M.A. (1992) Effects of legumes on soil physical quality in a maize crop. Plant & Soil 140, 15–23. Lebreton P., Jeangros B., Gallet C. & Scehovic J. (2001) Sur l'organisation phytochimique de formations prairiales permanents. Canadian Journal of Botany 79, 510-519. Lesturgez, G., Poss, R., Hartmann, C., Bourdon, E., Noble, A. & Ratana-Anupap, S. (2004) Roots of Stylosanthes hamata create macropores in the compact layer of a sandy soil. Plant & Soil 260, 101-109. Lewis, G.C. & Hopkins, A. (2000) Weeds, pests and diseases of grassland. In: A. Hopkins (Ed). Grass: Its Production and Utilization, pp119-139. Blackwell Science, Oxford, UK. Li, X-L., George, E., Marschner, H. & Zhang, J-L. (1997) Phosphorous acquisition from compacted soil by hyphae of a mycorrhizal fungus associated with red clover (Trifolium pratense). Canadian Journal of Botany 75, 723-729. Lodge, G.M. (1991). Management practices and other factors contributing to the decline in persistence of grazed lucerne in temperate Australia: a review. Australian Journal of Experimental Agriculture 31, 713–724. Löfkvist, J. (2005) Modifying soil structure using plant roots. Doctoral Thesis, Swedish University of Agricultural Sciences, Uppsala. Löfkvist, J., Whalley, W.R. & Clark, L.J. (2005) A rapid screening method for good root- penetrating ability: Comparison of species with very different root morphology. Acta Agriculturae Scandinavica Section B: Soil and Plant Science. Masri, Z. & Ryan, J. (2006) Soil organic matter and related physical properties in a Mediterranean wheat-based rotation trial. Soil & Tillage Research 87, 146-154.

28 Materechera, S.A., Alston, A.M., Kirby, J.M. & Dexter, A.R. (1992) Influence of root diameter on the penetration of seminal roots into a compacted subsoil. Plant & Soil 144, 297-30. Materechera, S.A., Alston, A.M., Kirby, J.M. & Dexter, A.R. (1993) Field evaluation of laboratory techniques for predicting the ability of roots to penetrate strong soil and of the influence of roots on water sorptivity. Plant & Soil 149, 149-158. McDonald, P., Edwards, R.A. & Greenhalgh, J.F.D. (2002) Animal Nutrition, 6th edition. Pearson Education, Harlow, UK. Meek, B.D., Detar, W.R., Rolph, D., Rechel, E.R. & Carter, L.M. (1990) Infiltration rate as affected by an alfalfa and no-till cotton cropping system. Soil Science Society of America Journal 54, 505-508. Milberg, P. (1994) Germination ecology of the polycarpic perennials Primula veris and Trollius europaeus. Ecography 17, 3-8. Milton, W.E.J. (1933) The palatability of the self-establishing species contributing to different types of grassland. Empire Journal of Experimental Agriculture 1, 347- 360. Miransari, M., Bahrami, H.A., Rejali, F., Malakouti, M.J. & Tarabi, H. (2007) Using arbuscular mycorrhiza to reduce the stressful effects of soil compaction on corn (Zea mays L.) growth. Soil Biology & Biochemistry 40, 1197-1206. Miransari, M., Bahrami, H.A., Rejali, F. & Malakouti, M.J. (2008) Using arbuscular mycorrhiza to alleviate the stress of soil compaction on wheat (Triticum aestivum L.) growth. Soil Biology & Biochemistry 40, 1197-1206. Molan, A.L., Waghorn, G.C. & Mcnabb, W.C. (1999) Condensed tannins and gastro- intestinal parasites in sheep. Proceedings of the New Zealand Grassland Association 61, 57-61. Mortimer S.R., Kessock-Philip R., Potts S.G., Ramsay A.J., Roberts S.P.M., Woodcock B.A., Hopkins A., Gundrey A., Tallowin J., Vickery J. & Gough S. (2006). Review of the diet and micro-habitat values for wildlife and the agronomic potential of selected grassland plant species. English Nature Research Reports No. 697. English Nature, Peterborough, UK. Mytton, L.R., Cresswell, A. & Colbourn, P. (1993) Improvement in soil structure associated with white clover. Grass & Forage Science 48, 84-90. Newman, I.E. & Rovira, A.D. (1975) Allelopathy among British grassland species. Journal of Ecology 63, 727-737. NIAB (1966) Varieties of Lucerne. Farmers leaflet No. 4, National Institute of Agricultural Botany, Cambridge, UK [no longer available] Nordbakken, J.-F., Rydgren, K., Auestad, I. & Austad, I. (2010) Successful creation of species-rich grassland on road verges depend on various methods for seed transfer. Urban Forestry & Urban Greening (In Press). O’Beirne-Ranelagh, E. (2005) Managing grass for horses. J.A. Allen, London, UK. Partridge, S.T., Nolte, D.L., Ziegltrum, G.T. & Robbins, C.T. (2001) Impacts of supplemental feeding on the nutritional ecology of black bears. Journal of Wildlife Management 65,191-199.

29 Pettersson, M.W. (1991). Flower herbivory and seed predation in Silene vulgaris (Caryophyllaceae). Effects of pollination and phenology. Holarctic Ecology 14, 45– 50. Preston, C.D., Pearman, D.A. & Dines, T.D. (2002) New Atlas of the British and Irish Flora. Oxford University Press, Oxford. Pywell, R.F., Bullock, J.M., Roy, D.B., Warman, L., Walker, K.J. & Rothery, P. (2003) Plant traits as predictors of performance in ecological restoration. Journal of Applied Ecology 40, 65-77. Pywell, R.F., Bullock, J.M., Walker, K.J., Myhill, D. Warman, L., Barratt, D., Sparks, T., Hopkins, A., Burke, M. & Peel, P. (1999) Multi-site experiments on the restoration of species-rich grassland on ex-arable land in the ESAs. Final report for MAFF project BD1404. Centre for Ecology and Hydrology (Unpublished). Raunkiær, C. (1934) The Life Forms of Plants and Statistical Plant Geography, being the collected papers of C. Raunkiær. Oxford University Press, Oxford. Rautiainen, P., Koivula, K. & Hyvärinen, M. (2004)The effect of within-genet and between-genet competition on sexual reproduction and vegetative spread in Potentilla anserina ssp. egedii. Journal of Ecology 92, 505–511. Rechel, E.A., DeTar, W.R., Meek, B.D. & Carter, L.M. (1991) Alfalfa (Medicago sativa L.) water use efficiency as affected by harvest traffic and soil compaction in a sandy loam soil. Irrigation Science 12, 61-65. Reinert, D.J., Albuquerque, J.A., Reiehert, J.M., Aita, C. & Andrada, M.M.C. (2008) Bulk density critical limits for normal root growth of cover crops. Revista Brasileira de Ciencia do Solo 32, 1805-1816. Reintam, E., Trukmann, K., Kuht, J., Toomsoo, A., Teesalu, T., Koster, T., Edesi, L. & Nugis, E. (2008) Effect of Cirsium arvense L. on soil physical properties and crop growth. Agricultural & Food Science 17, 153-164. RINH (2008) Rumen-Up – Ruminal metabolism enhanced naturally using plants. Final report of the European Commission Contract No. QLK5-CT-2001-00992. Available at: http://www.rowett.ac.uk/rumen_up/ Rochester, I.J., Peoples, M.B., Hulugalle, N.R., Gault, R.R. & Constable, G.A. (2001) Using legumes to enhance nitrogen fertility and improve soil condition in cotton cropping systems. Field Crop Research 70, 27–41. Rochon J.J., Doyle C.J., Greef J.M., Hopkins A., Molle G., Sitzia M., Scholefield D. & Smith C.J. (2004) Grazing legumes in Europe: a review of their status, management, benefits, research needs and future prospects. Grass and Forage Science 59, 197-214. Rodwell, J.S., ed. 1992. British Plant Communities. Volume 3. Grassland and Montane Communities. Cambridge University Press, Cambridge, UK. Roovers, P., Baeten, S., Hermy, M. & Gulinck, H. (2004) Plant species variation across path ecotones in a variety of common vegetation types. Plant Ecology 170, 107- 119. Rose, F. (2006) The Wild Flower Key. Frederick Warne – Penguin Books Ltd. London, UK.

30 Salako, F.K., Hauser, S., Babalola, O. & Tian, G. (2001) Improvement of the physical fertility of a degraded Alfisol with planted and natural fallows under humid tropical conditions. Soil Use & Management 17, 41-47. Sheldrick, R.D. (2000) Sward establishment and renovation. In: A. Hopkins (ed). Grass: Its Production and Utilization, pp13-30. Blackwell Science, Oxford, UK. Sheldrick, R.D., Lavender, R.H. & Tewson, V.J. (1986) The effects of frequency of defoliation, date of first cut and heading date of perennial ryegrass companion on the yield, quality and persistence of diploid and tetraploid broad red clover. Grass and Forage Science 41, 137-149. Smart, S.M., Firbank, L.G., Bunce, R.G.H. & Watkins, J.W. (2000) Quantifying changes in abundance of food plants for butterfly larvae and farmland birds. Journal of Applied Ecology 37, 398-414. Soons, M.B & Heil, G.W. (2002) Reduced colonization capacity in fragmented populations of wind-dispersed grassland forbs. Journal of Ecology 90, 1033-1043. Speijers, M.H.M., Fraser, M.D., Theobald, V.J. & Haresign, W. (2004) The effects of grazing forage legumes on the performance of finishing lambs. Journal of Agricultural Science 142, 483-493. Šrámek, P. & Kašparová, J. (2005) Evaluation of different methods for restoration of species-rich grasslands. In: E. Molina Alcaide, H. Ben Salem, K. Biala, P. Morand- Fehr (Eds) Sustainable grazing, nutritional utilization and quality of sheep and goat products. Options Méditerranéennes, Series A, No. 67, pp123-127. CIHEAM/FAO/CSIC, Zaragoza, Spain. Stace, C. (1997) New flora of the British Isles. 2nd edition. Cambridge University Press, Cambridge, UK. Sullivan, J. (1992). Medicago sativa. In: Fire Effects Information System, [Online]. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory. Available at: http://www.fs.fed.us/database/feis/plants/forb/medsat/all.html Tamm, C.O. (1972) Survival and flowering of perennial herbs. III. The behaviour of Primula veris on permanent plots. Oikos 23, 159-166. Tanakamaru, S., Hayashida, S., Mochizuki, T. & Furuya, T. (1998) Specific difference in root penetration into the compacted soil cakes in crop plants. Japanese Journal of Crop Science 67, 63-69. Thomas, B., Thompson, A., Oyenuga, V.A., & Armstrong, R.H. (1952) The ash constituents of some herbage plants at different stages of maturity. Empire Journal of Agriculture 20, 10-22. Thompson, K. (1989) A comparative study of germination responses to high irradiance light. Annals of Botany 63, 159-162. Thompson, K., Grime, J.P. & Mason, G. (1977) Seed germination in response to diurnal fluctuation in temperature. Nature 67, 147-149. Thompson K., Bakker J.P. & Bekker R.M. (1997) The Soil Seed Banks of North West Europe: Methodology, Density and Longevity. Cambridge University Press, London, UK.

31 Topp C.F.E. & Doyle C.J. (2004). Modelling the comparative productivity and profitability of grass and legume systems of silage production in northern Europe. Grass and Forage Science 59, 274-291. Turkington, R. & Aarssen, L.W. (1983) Biological Flora of the British Isles. Hypochoeris Radicata L. (Achyrophorus Radicatus (L.) Scop.). Journal of Ecology 71, 999- 1022. Vandvik, V. & Vange, V. (2003) Germination ecology of the clonal herb Knautia arvensis: Regeneration strategy and geographic variation. Journal of Vegetation Science 14, 591-600. Verkleij, J.A.C., Van Hoof, N.A.L.M., Chardonnens, A.N., Koevoets, P.L.M., Hakvoort, H., ten Bookum, W. M., Schat, H. & Ernst, W.H.O. (2001) Mechanisms of heavy metal resistance in Silene vulgaris. Plant Nutrition 92, 446-447. Viano, J., Masotti, V. & Gaydou, E.M. (1999) Nutritional Value of Mediterranean Sheep's Burnet (Sanguisorba minor ssp. muricata). Journal of Agricultural and Food Chemistry 47, 4645–4648. Wall, D.A. & Morrison, I. N. (1990) Phenological development and biomass allocation in Silene vulgaris (Moench) Garcke. Weed Research 30, 279-288. Warwick, S.I. & Black, L. (1982) The biology of Canadian Weeds. 52. Achillea millefolium s.l. Canadian Journal of Plant Science 62, 163-182. Whalley, W.R., Dumitru, E. & Dexter, A.R. (1995) Biological effects of soil compaction. Soil & Tillage Research 35, 53-68. Wilson, G.F., Lal, R. & Okigbo, B.N. (1982) Effects of cover crops on soil structure and on yield of subsequent arable crops grown under strip tillage on an eroded alfisol. Soil & Tillage Research 2, 233-250. Yano, K., Yamauchi, A., Iijima, M. & Kono, Y. (1998) Arbuscular mycorrhizal formation in undisturbed soil counteracts compacted soil stress for pigeon pea. Applied Soil Ecology 10, 95-102. Zeghichi, S., Kallinthraka, S., Simopoulos, A.P. & Kypriotakis, Z. (2003) Nutritional composition of selected wild plants in the diet of Crete. In: A.P. Simopoulos & C. Gopalan (eds) Plants in Human Health and Nutrition Policy. World review of Nutrition and Dietetics 91, pp22-40. S. Karger, Basel, Switzerland.

32