Portal de Periódicos da CAPES

Cirsium eriophorum (L.) Scop. ( Carduus eriophorus L.; Cnicus eriophorus (L.) Roth)

1999; Wiley; Volume: 87; Issue: 3 Linguagem: Inglês

10.1046/j.1365-2745.1999.00369.x

1365-2745

Richard Tofts,

Plant Toxicity and Pharmacological Properties

Sect. Eriolepis (Cass.) Dumort. Monocarpic, with a thick tap root or poorly branched main roots. Basal leaves to 80 cm, ovate-oblong, narrowing into a short stalk, deeply pinnatifid and strongly undulate, the narrowly lanceolate distant spine-tipped (4–11 mm) segments usually in pairs with the basal lobe directed upwards and apical lobe downwards, lobes 4–11 × 0.7–2.2 cm with inrolled margins, midrib often purplish (at least towards the stem); leaves somewhat hispid-hairy above, tomentose below. Stem erect, pale green with or without longitudinal purplish streaks in lower part, to 20 mm diameter at base, densely covered with ± patent hairs to c. 6 mm long, unwinged and not prickly, furrowed, 40–160(–180) cm, corymbosely branched above. Stem leaves similar to basal, but sessile and loosely auriculate. Capitula (3–) 10–50(–61) in number, ± long-pedunculate and erect, usually with a few small subtending leaves. Involucre 4–7 × 5–7 cm (width × height), usually solitary, very cottony and white; bracts lanceolate-acuminate, innermost 43–46 × 1.5–2 mm, outermost 16–19 × 3–4 mm, ending in a long narrow purplish ciliate point, usually with a slight dilation just below its apex; the outermost spine-tipped. Corolla 2.5–4.5 cm, florets red-purple, anthers purple [see also VIII (A)]. Achenes 3.9–5.4 × 2.0–2.4 × 1.2–1.5 mm, smooth and shiny, buff flecked with short black streaks; pappus of feather-like hairs 15–33 mm long, white, attached to an ochraceous collar. A very distinctive and striking member of the British flora. On mainland Europe, the species is not so well defined, with a multitude of varieties, subspecies or closely related species having been described (particularly from the Balkan peninsula; Fl. Eur. 4). British material is regarded as an endemic subspecies (ssp. britannicum Petrak) by some authors. In Britain, Cirsium eriophorum is confined to calcareous soils and is most common on roadsides, rough grassland, pasture and amongst light scrub. It also occurs in woodland clearings and quarries and other disturbed habitats, such as spoil bank and scree habitats created by limestone quarrying (Anderson & Shimwell 1981). As a native species C. eriophorum is mainly a plant of central and western Europe, extending northwards to northern England (Teesdale) and (as a casual) in south-east Scotland (Fig. 1). It extends eastwards to Rumania, and to Spain and northern Italy in the south (Fig. 2). Its geographical distribution is broadly similar to that of C. acaule (Pigott 1968) although C. eriophorum extends further to the south-east (being present in Albania, Hungary and Greece) and not so far to the north (being absent from Denmark, Norway, Sweden and Russia). Some authors (e.g. Hegi Fl. ed. 4.6) state that the range of C. eriophorum extends into the upper Volga and other parts of Russia [as C. eriophorum ssp. decussatum (Janka) Petrak), but Fl. Eur. 4 treats this taxon as a separate species (C. decussatum Janka); the limits of its distribution seem to be set largely by temperature; see II(A)]. Records of C. eriophorum as an introduced species appear to be few. Thompson (1922) cites a record for New Zealand (the Upper Wairarapa) from 1899, noting it ‘has not been recorded since’. Information given by Webb et al. (1988) suggests, however, that New Zealand records of C. eriophorum refer to C. brevistylum Cronq. A record from Turkey also appears to be an error (Davis 1975). The distribution of Cirsium eriophorum in the British Isles. Each symbol represents at least one record in a 10-km square of the National Grid. (○) Pre-1950; (●) 1950 onwards; (+) casuals pre-1950; (×) casuals 1950 onwards. Mapped by Mrs J. M. Croft, Biological Records Centre, Institute of Terrestrial Ecology, mainly from records made by members of the Botanical Society of the British Isles. European range of Cirsium eriophorum based on records in the literature. Local in England and Wales, and absent from Ireland, C. eriophorum has been recorded as a casual in south-east Scotland for over a century (Hooker & Arnott 1860). It appears to have had a restricted distribution since records began. Ray (1724), comparing it with C. vulgare, notes ‘. . . quo etiam multo rarior est’. It is more common in the central and western parts of its English range, being probably most abundant in Oxfordshire, Gloucestershire, Avon and Wiltshire. It is rare in the eastern counties of Essex (Jermyn 1974), Suffolk and Norfolk (Petch & Swann 1968; Trist 1979; Simpson 1982) and Hertfordshire (Dony 1967), and in the north-east such as the Howardian Hills of North Yorkshire (Gulliver 1990). It also becomes scattered in the southern and western extremities of its British range, despite the presence of habitat that appears at least geologically suitable. Perhaps the most puzzling aspect of its British distribution is its scarcity or absence from the North and South Downs, and in this respect it differs markedly from C. acaule, another calcicole thistle [see II(A)]. In England it has been found at up to 260 m in Yorkshire (Fl. Br. Isl.) and to 270 m at Crickley Hill, Gloucestershire, and 310 m at Hassop Mines, Derbyshire, during the course of the present survey (Table 1, sites 10 and 16). In continental Europe, it often occurs at higher altitudes and grows at up to 2000 m in Tirol. It is classified as a subalpine species in the north-eastern part of Italy (Poldini 1991). Cirsium eriophorum appears to be a local species throughout most of its range. It is regarded as one of the rarer thistles in southern Germany, lower Austria and the Alsace by Freese (1995). Salisbury & (1952) gives the distribution of C. eriophorum in England and Wales as oceanic southern, and indicates that humid locations favour it here compared with the species in mainland Europe, which is described as continental. Ellenberg et al. (1991), however, give the status of C. eriophorum in central Europe as being between oceanic and suboceanic, and Preston & Hill (1997) classify it as European temperate. Its distribution suggests that temperature is important in determining the extent of its overall range. In the British Isles, it becomes scarce or rare in its more northerly stations, such as the limestone areas of the southern Peak District (Linton 1903; Clapham 1969), and in western parts of its range, such as the coastal regions of Glamorgan, despite the humidity of these areas being greater than that in the core of its distribution in Wiltshire, Gloucestershire and Oxfordshire. Furthermore, it is completely absent from some apparently suitable oceanic areas, such as parts of the south-western coast of Wales and north Wales. It is also absent from the Carboniferous limestone of the Craven in West Yorkshire. It is noted (Hegi Fl. ed. 4.6) that in its northern stations on mainland Europe it occurs exclusively on south-facing slopes; this contrasts with its more central stations in the Alps, where it is much less restricted to south-facing sites. This fact is consistent with temperature limitation, and is also seen in C. acaule (Pigott 1968). The rarity or absence of C. eriophorum from much of the drier counties of Kent, Sussex and Surrey contrasts with the distribution of C. acaule. This aspect of its distribution does not correlate well with low rainfall per se, but shows a better correlation with summer water balance (i.e. the difference between rainfall and evaporation), being scarce or absent from areas where an agricultural drought might be expected in more than five summers out of 10 (Atlas of Britain 1963). The limits to the European distribution of C. eriophorum in the north and east appear to be determined by low winter temperature, and its north-western limits by summer warmth. In Britain it is rare or absent outside the 60 °F (15.6 °C) mean August temperature isotherm (Clim. Atl.). This may account for its absence from Ireland, where the summers are cooler. The southern European boundary does, however, suggest that high summer temperatures and drought may restrict its distribution in the Mediterranean region, and its scarcity or absence from much of south-east England may be attributable to a similar cause. In general, the factors limiting its wide-scale distribution appear to be similar to those which constrain C. acaule. In Britain, C. eriophorum is restricted to freely draining base-rich soils over chalk and limestone. The extent to which it is favoured by particular limestone types is unclear, owing to the extent to which geology is confounded with climate. However, C. eriophorum often occurs on thin soils with its roots growing amongst lumps of limestone and chalk, and a relationship between solid geology and its distribution or abundance might therefore be expected, perhaps acting in combination with climatic factors. The principal rocks in the British Isles are Oolitic limestone and chalk, with locally frequent occurrences on the Lias in Somerset and Dorset. It occurs only rarely on the geologically older Carboniferous (e.g. Peak District and the Gower) and Magnesian (North Yorkshire) limestones. In the centre of its range in Wiltshire, it occurs most frequently on Oolitic limestone, but is also frequent on the Lower Chalk (Grose 1957). In Dorset, near the southern extreme of its British range, it is virtually confined to soils over the Jurassic limestones and almost absent from the abundant chalk (Good 1948). In Staffordshire, C. eriophorum is rare on Carboniferous limestone, but frequent on Silurian limestone (Edees 1972). On the continent, it rarely occurs on any other substrate than chalk in the northern part of its range, but it is much less restricted in its southern stations (Hegi Fl. ed. 4.6 ). von Mannagetta (1892) includes a record from sandstone. In Dorset, C. eriophorum has mostly been recorded from red-brown rendzinas (Good 1948). In the eastern counties of Suffolk (Simpson 1982) and Cambridgeshire (Perring et al. 1964) it is recorded from Boulder Clay. During the course of the present survey, C. eriophorum was recorded from clay, silty loam, sandy loam or loamy sand topsoils, over chalk or limestone lumps, with a pH range of 6.7–8.1. Its pH range in north-east Italy is given as 5.5–8 (Poldini 1991). Hegi Fl. ed. 4.6 regards it as showing little selectivity in the southern part of its European range [and Burnat (1931) records it from ‘bords des champs silice’], but Hegi Fl. ed. 4.6 does list it from heavy chalk and clay-marl soils and notes that it is nitrophilous. In central Europe it is classified as an indicator of relatively dry basic soils and chalk, growing on moderately nitrogen-rich soils (Ellenberg et al. 1991). In central Europe, C. eriophorum is the character species of the ruderal Cirsietum eriophori (Oberdorfer 1979), occurring alongside calcareous roadsides and on limestone sheep grazings. It has also been recorded from pastured Seslerietum, Festuco–Brometea and a Cirsium eriophorum–Festuca ovina association is known (Hegi Fl. ed. 4.6). Species associated with C. eriophorum include Artemisia absinthium, Asperula arvensis, Carduus nutans, Cynoglossum officinale, Marrubium vulgare, Potentilla intermedia, Reseda luteola, Stachys germanica and Verbascum virgatum. Additional associates noted by Hegi Fl. ed. 4.6 include Atropa belladonna (which also grows as an associate in parts of the Cotswolds in England), Chamerion angustifolium, Senecio jacobaea, S. erucifolius, S. viscosus, Cirsium vulgare, Euphorbia cyparissias, Silaum silaus and Galium verum. In the Alps it has been recorded from Berberis–Rosa scrub, with Rosa villosa, Allium oleraceum, Cynoglossum officinale, Seseli montanum, Pimpinella saxifraga, Achillea millefolium, Artemisia vulgaris, A. absinthium, Carlina acaulis, Salvia pratensis, Prunella grandiflora, Anchusa officinalis, Linaria vulgaris, Galium verum, Melampyrum arvense and Campanula rapunculoides. A less typical habitat is dry spruce woodland with Aegopodium podagraria and Polemonium caeruleum. Cirsium eriophorum was recorded from MG1a, MG1b, MG1d, MG5b, MG6c, CG2c, CG3a, W25 and OV23 communities/subcommunities of the National Vegetation Classification (NVC) during the present survey, with more than half the records being from MG1 grasslands. Rodwell (1992) lists C. eriophorum from CG2, CG3 and CG5 communities of the NVC. In CG2 Festuca ovina–Avenula pratensis (Helictotrichon pratense) grassland, C. eriophorum is recorded from the Holcus lanatus–Trifolium repens (CG2c) and Dicranum scoparium (CG2d) subcommunities, in both cases at frequencies of 20% or less. In CG3 Bromus erectus (Bromopsis erecta) grassland, C. eriophorum is recorded from the Knautia arvensis–Bellis perennis (CG3c) subcommunity, at a frequency of 21–40%. In CG5 Bromus erectus–Brachypodium pinnatum grassland, C. eriophorum is recorded from the typical (CG5a) and Hieracium spp. (CG5b) subcommunities, in both cases at frequencies of 20% or less. Quadrat data from British localities are given at Table 1. Cirsium eriophorum occurs most frequently in rough grassland communities, maintained either by periodic grazing or mowing. Typical associates include Festuca rubra, Dactylis glomerata and, in the roughest swards, Arrhenatherum elatius. Cirsium eriophorum is favoured by conditions with at least moderate levels of nutrients, and often grows in disturbed areas with species such as Urtica dioica. In north Wiltshire it has been described as ‘a bane to some farmers’ (Gillam 1993) and can behave as a weed in badly managed pastures (e.g. Cherhill Down and Bishopton Down, Wiltshire) where poaching by stock has occurred. In such situations it is associated with Urtica dioica and C. arvense. Suitable mown habitats are most frequently encountered along roadside verges in the infrequently cut strip furthest from the metalled surface (e.g. close to a hedge or bank). In such circumstances it is often associated with rank grasses such as Arrhenatherum elatius and Elytrigia repens, tall herb species such as Heracleum sphondylium, Anthriscus sylvestris and Rumex spp., and more woody species such as Rubus caesius and R. fruticosus agg. Along roadsides and in similar situations, ditch clearance and the associated disturbance to the verge and spreading of soil may also help to provide suitable habitat for C. eriophorum. It is also found in incipient scrub communities, and is known as a plant of coppices in Hatfield Forest in Essex (Rackham 1989). It has been recorded from woodland clearings in Oxfordshire (J. Dunn, personal communication) and has been recorded from apparently similar situations elsewhere in Europe (von Mannagetta 1892; Hayek & Markgraf 1931). Cirsium eriophorum can survive and flower under light shading from scrub. Less frequently, C. eriophorum may be found growing amongst species characteristic of acid soils and has been recorded from stands of bracken (Table 1, site 15) in Gloucestershire. Even in such situations, however, the soil around the roots of C. eriophorum appears to be basic. It has been recorded from limestone heath (Horwood & Noel 1933). The mature leaves of C. eriophorum are protected by stiff sharp spines, and are unpalatable to most stock. The abundance of this species (and other members of the Onopordion) around central European villages in which cattle and sheep are still driven is attributed to the fact that mature plants are avoided by grazing animals (Ellenberg 1986). When present in hay, C. eriophorum is avoided by cattle, although C. arvense is much favoured (D. Millin, personal communication). The first few (typically five to six) true leaves are, however, only softly spined, and such leaves are eaten with relish by sheep. In comparison with C. vulgare, the young leaves of C. eriophorum are softer, less hispid, longer and more erect in habit. These features make C. eriophorum more susceptible to grazing damage in the early stages after initial establishment, once the plants are large enough to attract the attention of sheep but before the stiff spines are developed. In such circumstances, repeated grazing of small C. eriophorum plants may lead to much reduced numbers or elimination from a community. Observations in various parts of England (Gloucestershire, Bedfordshire, Berkshire and Wiltshire) show both C. eriophorum and C. vulgare to be present in lightly managed (cut or grazed) grassland, but only C. vulgare to be present in adjacent and closely grazed agricultural grasslands. Cirsium eriophorum may grow on downland on disturbed soil around rabbit warrens (e.g. Knap Hill, Wiltshire; Table 1, site 12), and in such situations it appears to be avoided by rabbits. During the course of fieldwork undertaken for this study, the site where C. eriophorum was found in greatest profusion (Hill End Farm, Oxon; Table 1, site 14) showed evidence of heavy rabbit activity. Light sheep grazing occurred throughout the year, with moderate cattle grazing from May to October. Cirsium eriophorum is controlled at this site by means of cutting and removing about 80% of all flowering plants each year, resulting in a relatively stable but high population (up to 45 plants in a 10 × 10 m quadrat). The snail Helix aspersa has been observed to eat both the cotyledons and the younger basal leaves of C. eriophorum. Performance in the early stages of growth is greater in areas of bare soil compared with vegetated ground. This result is similar to that for C. vulgare, although in a comparative experiment at Little Wittenham, Oxfordshire, seeds of C. eriophorum planted into a sheep-grazed sward showed significantly better establishment (up to the two true-leaf stage, and before grazing resulted in defoliation of the plants) than those of C. vulgare (up to 76% survival vs. 13%, respectively), despite the percentage germination of both species in the laboratory being the same. Once established, the surrounding vegetation probably has relatively little effect on plants of C. eriophorum other than to slow their growth, except where the vegetation causes shading; the rosettes are large (up to 1.6 m diameter) and have the capacity to weaken or kill the surrounding plants over which the rosette leaves spread. This response is probably not owing to the effects of shading alone, but also caused by the pressure exerted by the rosette leaves on the surrounding vegetation and ground (Salisbury 1952). Almost invariably moderately to highly gregarious in Britain over a scale of a few tens of metres, C. eriophorum is also reported as gregarious on mainland Europe (Hegi Fl. ed. 4.6;WP ***font***Ellenberg et al. 1991). Plants usually occur in groups of a few individuals in suitable areas of grassland, often where some obvious disturbance has occurred (e.g. old rabbit scrapes). Observations suggest that C. eriophorum is more prolific in areas that are ungrazed or only lightly grazed during the early stages of establishment, although a closed sward reduces the number of individuals germinating and surviving through the early stages. The species is therefore most abundant where some form of disturbance has created suitable germination sites in a grassland (e.g. rabbit scrapes, poaching by stock) after which the disturbance regime has been relaxed, allowing establishment. Cirsium eriophorum is occasionally found in light scrub and woodland clearings. It is favoured by conditions of high light flux, and only exceptionally occurs where shading reduces light levels to 40% of the amount received in unshaded conditions (Ellenberg et al. 1991). It seldom occurs under mown conditions except where mowing is very infrequent (once every few years), although plants that are cut in the spring immediately prior to stem elongation can produce (relatively small) flowering stems later in the same year. The rosettes of C. eriophorum may remain green until late winter, at least in southern England, although the large, distant-lobed leaves of the previous autumn die off during hard weather (see also VII). Ellenberg et al. (1991) classify C. eriophorum as a summer-green species. It develops a substantial root, and larger plants seem seldom to be affected by drought. Seedlings are, however, much more susceptible, and may perish during prolonged periods of dry weather. There are cytological distinctions between the palisade cells of the upward-pointing and horizontal leaf segments (Hegi Fl. ed. 4.6), with the former being longer and particularly chlorophyll-rich. This observation has been confirmed on British material. Stomata are present on both upper and lower leaf surfaces. On the basal leaves, stomatal densities range between 29 and 91 mm−2 (mean 63 mm−2) and 128–333 mm−2 (mean 216 mm−2) on the upper (adaxial) and lower (abaxial) surfaces, respectively. Plants examined in November 1996 showed abundant evidence of colonization by V-A mycorrhizal fungi. Normally classified as a biennial semi-rosette hemicryptophyte [although Huxley (1967) describes it as biennial to perennial], with reproduction entirely by seed. Cirsium eriophorum takes 2 years to complete its life cycle when grown in garden conditions (nutrient-rich soil with no close neighbours), but field observations indicate that at least a 3-year period (and possibly often much longer) is normally needed in order to complete the life cycle. 2n = 34 in British plants (Fl. Br. Isl.) and on the continent (e.g. Strid & Tan 1991 and references therein). Cirsium eriophorum has a lower growth rate than C. vulgare during the early stages of development (after reaching the two true-leaf stage), achieving only 38% of the biomass of C. vulgare after 49 days of growth in 9-cm pots, despite the larger seed mass of C. eriophorum. This suggests that C. vulgare may have a strong competitive advantage in agriculturally improved grasslands compared with C. eriophorum, perhaps accounting in part for its greater abundance even in calcareous areas, despite apparently performing more poorly in terms of initial germination [IV(B)]. No information for C. eriophorum. Biochemical data for the closely related C. vulgare are provided by Klinkhamer & De Jong (1993). A tiny rosette of leaves often persists under cover of dead vegetation during much of the winter, and rosette leaves expand during the spring, the earliest usually reaching full size by May. Some large rosette leaves usually persist into January, even during conditions of heavy frost. They subsequently die back during the later part of the winter. Stem elongation normally begins during mid-June in southern England, and the first flower buds appear during late June or the beginning of July. The first flowers typically open in mid-July, and fruit is formed during August or early September. Smaller buds in the lower leaf axils may flower later in the season (September), although these do not generally appear to produce good seed. Florets tubular, hermaphrodite (Fig. 3). The corolla comprises a tube 23–28 mm long expanding into an unequally five-lobed limb 12–15 mm in total length, the limbs being 3.5–9 mm long. The five stamens are epipetalous, 14–17 mm long, the anthers being united into a cylinder 9.5–10 mm long with five teeth at each end, enclosing the central part of the style. The style is 20–26 mm long. In Britain and continental Europe, flowers are commonly visited by bumble bees and butterflies (Table 2). The proboscis of the bumble bee must be at least 7–8 mm long in order to reach the nectar in the floret base (Knuth, Poll. 2). The flowers are reddish-purple, with white-flowered forms being rarely recorded by a number of authors (Hegi Fl. ed. 4.6; Rouy Fl. 9; Willkomm & Lange 1870; Thompson 1912; Huxley 1967). (a) Achene; (b–e) seedling stages from germination to 14 days growth in a glasshouse; (f) floret; (g) part of floret with corolla tube removed, showing stamens and style; (h,i) pollen grain, meridian and polar view. The pollen grain is spiny, circular to weakly three-angled and 40–70 μm in diameter in polar view, circular to slightly ellipsoid in meridian view (Fig. 3). The hybrid Cirsium eriophorum × C. vulgare = C. × grandiflorum Kittel (C. × gerhardtii Schultz-Bip) has rarely been recorded in England from north Somerset, north Wilts, Dorset, north Essex, Oxon, Cambs, south Lincs and south-east Yorks (Hyb. Br. Isl.). It is also known from Austria, the Czech Republic and Slovakia, France, Germany, Switzerland, Hungary, Italy and the former Yugoslavia. It is intermediate in stem wingedness and capitulum characters and is partially fertile (Stace 1997). Two further hybrids [C. eriophorum (L.) Scop. × C. palustre (L.) Scop. = C. × dominii M. Schulze, and C. arvense (L.) Scop. × C. eriophorum (L.) Scop. = C. × sennenii Rouy] are known from Austria, the Czech Republic and Slovakia, and from France, respectively. The number of achenes produced per plant is highly variable, varying from plant to plant within a population and between populations. Salisbury (1952) notes that up to about 300 fruits may be produced per head (an observation that accords well with the results of this study), but observes that many fruits are eaten by beetle larvae. During the present survey, 101 out of 102 freshly fruiting heads (i.e. those in which the majority of fruits had just reached maturity) showed signs of achene predation by insect larvae, and one head was found to contain nine larvae. From a study of seven sites, the median number of undamaged fruits per head varied from 0 to 50, with a maximum observed number of 282 in one head. Figure 4 shows the frequency distribution of undamaged seeds per flowering head for C. eriophorum at Little Wittenham, Oxon. The median number of fruits per head varies greatly between populations (P < 0.001 for the seven sites, Mood’s median test), but Fig. 4 shows the general form of the distribution that prevails at all sites. The number of flowering heads produced per plant is typically between 10 and 50. An estimate of the number of fruits usually produced per plant is therefore up to about 2500. As seed predation owing to insect larvae living in the thistle heads is almost universal, and typically acts to reduce the number of fruits released by at least 80% and often much more, the activities of insect larvae may control the abundance and distribution of C. eriophorum over at least parts of its range. Fruiting heads of C. eriophorum close during damp weather, and in such circumstances the population of insect larvae inside the heads may consume many achenes, destroying them before they can be dispersed when the heads reopen in drier weather. Fruits of C. eriophorum growing in areas with cooler summer temperatures may therefore mature more slowly, ripen later and be subject to greater retention during wet conditions, suffering higher predation by insect larvae before they can be released. Number of undamaged achenes per flower head for 33 heads from a population of Cirsium eriophorum at Little Wittenham, Oxfordshire. Mean achene (cypsela) weights (excluding pappus) vary from 7.9 to 10.3 mg per 100. For collections made at approximately the same time, plants from near the central part of the British range (Little Wittenham, Berkshire) were found to have a significantly heavier mean achene weight than those from near the western extreme (Wick, Glamorgan), with a 100 seed weight of 9.9–10.3 mg compared with 7.9–8.3 mg (P < 0.01, U-test). The western plants also produced fewer undamaged achenes per head (median = 6, n = 30) compared with the Berkshire ones (median = 46, n = 33) and generally fewer heads per plant (9–15 compared with 10–21). Achenes are attached to a pappus, but field observations suggest that the pappus often becomes detached from the achene before it leaves the flower head. Even in cases where this is not so, most of the fruits appear to land within a distance of a few metres from the plant. Figure 5 shows the results of a dispersal experiment in which seeds (with pappus attached) of C. eriophorum were released during a period of natural seed dispersal at Little Wittenham, Oxon. A comparison made at the same time showed that C. vulgare disperses seed significantly further than C. eriophorum (median = 4.9 m vs. 3.9 m, n1 = 100, n2 = 100, P < 0.05, U-test). The maximum dispersal distances recorded in this experiment were 60.5 m for C. eriophorum and > 100 m for C. vulgare. Dispersal distances (metres) for 100 achenes of Cirsium eriophorum released experimentally at Little Wittenham, Oxfordshire. Some predation of seed owing to small mammals has been observed. This appears to be very localized but may be severe, with upwards of 90% of seeds in a patch (c. 100 cm2) being consumed. Seeds have no cold requirement. Germination is more rapid in the dark, with the first signs of germination occurring after 3 days at 20 °C. Seed stored at room temperature for 1 month after collection and then germinated in the dark at 20 °C showed t50 of c. 5 days. Germination in the dark and light (20 °C, 100 μmol m−2s−1, standard fluorescent lighting) was 93% and 85%, respectively. Seed stored for 12 months showed germination of 81% in the dark. In nature, most germination appears to take place in the autumn, although some seeds may germinate in the following spring. Stages in seedling development are shown in Fig. 3. Cotyledons obovate, 15–20 × 8–12 mm. First true leaves lanceolate, some slightly spathulate, tomentose-woolly below, with weak buff-coloured spines along margin. Freese (1995) recorded 11 phytophagous species from the stems and roots of C. eriophorum, but concluded that more extensive sampling would reveal many more. For all phytophagous species combined, Freese (1995) calculated a mean number of 6.6 individuals per stem, and 93% of C. eriophorum plants were attacked. As in C. vulgare, a complex insect fauna is present in C. eriophorum, with herbivores and their associated predators and parasitoids, together with species which use the plants simply for shelter. Wasps (Vespidae) have frequently been observed to remove the ‘wool’ from small areas around the bracts, and to drink the liquid flowing from small incisions made in the bracts themselves. Table 2 lists the insect species recorded from C. eriophorum. Lousley (1950) includes a record of Orobanche reticulata on C. eriophorum, although this rare plant cannot be considered an important parasite of C. eriophorum in Britain. Ellis & Ellis (1985) list Puccinia cnici Mart. (a rust, on living plants) and Phomopsis cirsii Grove (on dead stems and leaves) from C. eriophorum. The hyphomycete Periconia cookei Mason & M.B. Ellis has also been seen on dead stems of C. eriophorum from Croughton, Northamptonshire. Hegi Fl. ed. 4.6 lists Erysiphe cichoracearum DC., Puccinia caricis frigidae E.Fisch., P. cirsii eriophori Jacky and Ophiobolus acuminatus (Sow.) Duby. No remains are recorded from British deposits of any age (Godw. Hist.), although the fruits of C. eriophorum are distinctive. First recorded in Britain by Lobelius in 1570 ‘Tomentosus Carduus Angliae . . . Frequens in Angliae collibus strigosis agri Sommerseti juxta aedes . . . ,’ Clarke (First Rec.). I am very grateful to many people who have helped with this work, in particular to Dr J. Silvertown who supervised my PhD during which time the present work was undertaken. I am indebted to David Clements, Revd Dr D. J. L. Agassiz, Dr R. R. Askew, P. J. Chandler, P. J. Hodge, Dr M. R. Shaw and Dr I. M. White who provided insect data. I would like to thank Dr A.J.C. Malloch for allocating the vegetation samples in Table 1 to the communities of the NVC, and Professor A.J. Willis and the associate editors for their comments on earlier drafts. Others who have helped in various ways are Mrs N. Nijhof-Vaags, J. Duckworth, Miss J. Dunn, C. Hora, D. Millin, J. Newton, the staff of the Plant Sciences Library and Fielding-Druce herbarium at Oxford, and the staff of the nature reserve at Little Wittenham. I also thank Mrs J.M. Croft for Fig. 1. This work was undertaken whilst I was in receipt of a BBSRC grant.