Institute Of Terrestrial Ecology Research Articles

Land cover and vegetation data from an ecological survey of "key habitat" landscapes in England, 1992–1993

Abstract. Since 1978, a series of national surveys (Countryside Survey, CS) have been carried out by the Centre for Ecology and Hydrology (CEH) (formerly the Institute of Terrestrial Ecology, ITE) to gather data on the natural environment in Great Britain (GB). As the sampling framework for these surveys is not optimised to yield data on rarer or more localised habitats, a survey was commissioned by the then Department of the Environment (DOE, now the Department for Environment, Food and Rural Affairs, DEFRA) in the 1990s to carry out additional survey work in English landscapes which contained semi-natural habitats that were perceived to be under threat, or which represented areas of concern to the ministry. The landscapes were lowland heath, chalk and limestone (calcareous) grasslands, coasts and uplands. The information recorded allowed an assessment of the extent and quality of a range of habitats defined during the project, which can now be translated into standard UK broad and priority habitat classes. The survey, known as the "Key Habitat Survey", followed a design which was a series of gridded, stratified, randomly selected 1 km squares taken as representative of each of the four landscape types in England, determined from statistical land classification and geological data ("spatial masks"). The definitions of the landscapes are given in the descriptions of the spatial masks, along with definitions of the surveyed habitats. A total of 213 of the 1 km2 square sample sites were surveyed in the summers of 1992 and 1993, with information being collected on vegetation species, land cover, landscape features and land use, applying standardised repeatable methods. The database contributes additional information and value to the long-term monitoring data gathered by the Countryside Survey and provides a valuable baseline against which future ecological changes may be compared, offering the potential for a repeat survey. The data were analysed and described in a series of contract reports and are summarised in the present paper, showing for example that valuable habitats were restricted in all landscapes, with the majority located within protected areas of countryside according to different UK designations. The dataset provides major potential for analyses, beyond those already published, for example in relation to climate change, agri-environment policies and land management. Precise locations of the plots are restricted, largely for reasons of landowner confidentiality. However, the representative nature of the dataset makes it highly valuable for evaluating the status of ecological elements within the associated landscapes surveyed. Both land cover data and vegetation plot data were collected during the surveys in 1992 and 1993 and are available via the following DOI: https://doi.org/10.5285/7aefe6aa-0760-4b6d-9473-fad8b960abd4. The spatial masks are also available from https://doi.org/10.5285/dc583be3-3649-4df6-b67e-b0f40b4ec895.

Ecological survey of the native pinewoods of Scotland 1971

Abstract. In 1971, a comprehensive ecological survey of the native pinewoods of Scotland was carried out by the Institute of Terrestrial Ecology. The survey was initiated as a consequence of growing concern about the status of the pinewood resource. Since the twentieth century, this unique habitat is widely recognised, not only by ecologists for its inherent biodiversity but also by the general public for its cultural and amenity value. The survey, utilising demonstrably repeatable methods, collected information on ground flora, soils, forest structure and also general site information from the major 27 sites of the 35 sites identified as truly native pinewoods in Scotland. The results from the survey prompted the organisation of an international symposium in 1975, which set the conservation agenda for the old Caledonian pinewoods. The data collected during the 1971 survey are now publicly available via the following DOI: doi:10/7xb ("Habitat, vegetation, tree and soil data from Native Pinewoods in Scotland, 1971"). Although the data are now 44 years old, the repeatable methods will allow for a resurvey to take place, in order to assess changes in the vegetation, habitats and tree composition in a statistically robust manner.

Norman Winfrid Moore (1923–2015)

With the death of Norman Moore aged 92, on 21 October 2015, Britain lost one of its most influential conservationists. To ornithologists, Norman was perhaps best known for his work on pesticide impacts on birds. He led the team at Monk's Wood Experimental Station that studied the biological impacts of DDT and other organochlorines in the 1960s. This work eventually contributed to a banning of these chemicals from agricultural use in Britain and elsewhere. However, he was an all-round naturalist with a particular interest in birds, dragonflies and butterflies. He became a world authority on dragonflies and their conservation, and was one of the first to appreciate the problems for wildlife of agricultural intensification and habitat fragmentation, and the value of hedgerows in modern farmland. Norman firmly believed in the importance of integrating nature conservation with other land-uses, whether agriculture or forestry. He realized that small nature reserves alone could not save biodiversity in a country as highly developed and populated as Britain. He was a founder and first Chairman of the Farming and Wildlife Advisory Group (FWAG). This was a charitable trust which aimed to find practical ways of conserving wildlife on working farms, thus bridging a widening gulf between farmers and conservationists. As a person, Norman was well-informed, charming and modest. With his naturally gentle and thoughtful manner, he became widely regarded as one of conservation's elder statesmen, much loved and respected by all who knew him. Norman was born on 24 February 1923, the son of a doctor, Sir Alan Moore. In 1934 the family moved from Lewes to near Battle in rural Sussex. From boyhood, Norman was a keen naturalist. Birds were his first love, but by his teens butterflies and dragonflies had caught his attention. His first scientific paper, ‘Rare Lepidoptera and Odonata in East Sussex’, was published in 1939, when he was aged 16. He was educated at Eton College and then Trinity College, Cambridge, where he read Natural Sciences, specializing in zoology. While in Cambridge, he became President of the University Bird Club. Like so many of his generation, his studies were interrupted by war service. He was called up in 1942 and served as a gunnery officer in the Royal Artillery, with active service in Belgium and the Netherlands. When manning observation posts, he kept two notebooks: one with reports on German military activity and the other on wildlife seen at the time. Late in 1944 he suffered a serious leg wound and was captured by German forces. He was one of only four British prisoners in a camp housing 23 000 Russians. Conditions were appalling. On a pint of mangel-wurzel soup per day and one small loaf of bread per week, prisoners were starving, and deaths in the camp numbered more than a hundred per day. On being liberated by the US 7th Armoured Division in April 1945, more than 9000 men from the camp were hospitalized and de-loused with the new wonder chemical, DDT. After his return to Britain, Norman trained gunners on Salisbury Plain before returning to Cambridge to complete his zoology degree in 1948. After a brief post-graduate expedition to The Gambia, Norman moved to teach zoology at Bristol University, where he spent much of his research time studying dragonflies, with emphasis on their territorial behaviour. His thesis led to a PhD in 1953. These insects, which became a lifetime passion, formed the subject of his first book, Dragonflies (1960), co-authored with Cynthia Longfield and Philip Corbet and published in the well-known New Naturalist Series. In 1953 Norman accepted the post of regional officer for Southwest England in the recently formed Nature Conservancy, a position in which he had sufficient freedom to do what he thought was important. Based at Furzebrook Research Station in Dorset, he documented the loss and fragmentation of the once continuous heathland as area after area was turned to farmland or tree plantation. He was among the first to realize the significance of habitat fragmentation to flora and fauna, demonstrating how species numbers dwindle as the fragments of remaining habitat become smaller and more isolated. He later examined the same phenomenon in woodland, and found similar trends in bird populations, with the rarest species disappearing first. While in the southwest, he helped to select and establish some of the first National Nature Reserves in the region, including Yarner Wood, Hartland Moor and Morden Bog. With the help of the British Trust for Ornithology, in 1954 Norman organized a nationwide survey of the Buzzard Buteo buteo, whose major prey-species, the Rabbit Oryctolagus cuniculus, was being drastically reduced by the newly introduced viral disease, myxomatosis. However, his work was to highlight another factor limiting Buzzard numbers. The resulting paper, published in British Birds, showed two maps side by side. One map depicted the density of Buzzards in different parts of Britain while the other showed the densities of gamekeepers in different parts of Britain. The one was almost a mirror image of the other, hinting strongly at what really limited Buzzard numbers in Britain at that time. Soon afterwards, from the late 1950s, came a conservation problem of quite different magnitude, as thousands of seed-eating birds and their predators were found dead and dying in newly sown cereal fields across the country. The grains that killed them had been treated with some new organo-chlorine pesticides, called aldrin and dieldrin. These events brought another major change in Norman's career. In 1960 he was appointed Head of the Toxic Chemicals and Wildlife Section at the newly created Monks Wood Experimental Station near Huntingdon. His task was to organize research on the effects of pesticides on wildlife. His colleagues included several young men soon to become well known, including Derek Ratcliffe (who discovered eggshelll thinning caused by DDT) and Ian Prestt (who studied raptors and herons, and eventually became Chief Executive of the RSPB). Moore's group began an extensive programme of monitoring the pesticide residues in a wide range of birds found dead and sent in by members of the public. This programme survives to this day in changed form, albeit no longer at Monks Wood. The group found organo-chlorine chemicals in almost every species examined, birds and others, but at highest levels in predatory species that were in severe decline, from Peregrines Falco peregrinus to Otters Lutra lutra. The team also conducted field studies on the most-affected species, including Peregrine and Sparrowhawk Accipiter nisus, Grey Heron Ardea cinerea and various seabirds. Support for the work rose with the publication of Rachel Carson's Silent Spring (1962), which heightened public concern over the effects of DDT and other organo-chlorines. Norman himself performed a crucial behind-the-scenes role in all this work, serving on committees, and tirelessly battling with representatives from the Ministry of Agriculture and the agro-chemical industry. In his book on the Peregrine (1993, p. 353), Derek Ratcliffe provided a fitting and heartfelt tribute to Norman's role: ‘In Britain the brunt of this wearisome task of convincing the unwilling fell on Norman Moore… The successful outcome, in the form of progressive restrictions and phasing out of the most harmful compounds is a tribute to Norman's skills of diplomacy and persuasion, and to his scientific authority.’ Although it took years before the use of organo-chlorine pesticides was reduced sufficiently in Britain to allow wildlife to recover, the work of Norman Moore's group attracted international acclaim, as did Monks Wood Experimental Station in which it was based. As part of his other work on farmland wildlife, Norman Moore began to study hedges, realizing that, in modern farmland, they provided an important refuge for many kinds of plants and animals. At that time, hedges were being torn out over much of the country with the help of ministry grants in order to create more cropland with bigger fields. He and his colleagues produced maps from past and present to show how the loss of hedges was transforming the lowland landscapes of eastern Britain from patchwork quilts to open prairies. Along with colleagues Max Hooper and Ernest Pollard, Norman contributed to another new Naturalist book, Hedges (1974), providing the first comprehensive study of these familiar but under-rated landscape features. The Conservative Government's ideological decision to split the Nature Conservancy's conservation and scientific arms, to form the Nature Conservancy Council and the Institute of Terrestrial Ecology, changed conservation in Britain, probably for the worse. It also changed Norman's career, breaking up the interdisciplinary team of biologists, toxicologists and chemists he had assembled at Monks Wood, to his great distress. In 1975, Norman became Chief Advisory Officer to the new Nature Conservancy Council (NCC), accepting a post created especially for him. He prepared the NCC's policy paper on Nature Conservation and Agriculture which called for a rural land-use strategy that took account of nature conservation alongside food production. The Conservative Government of the day was unreceptive to the idea. But gradually, through public pressure, the tide began to turn. At a public enquiry over the fate of Amberley Wild Brooks in Sussex, at which Norman gave evidence, nature conservation won one of its first victories over land drainage and agriculture. Public opinion eventually forced the Government to provide better protection for wild places, and Norman played an influential role in shaping the 1981 Wildlife and Countryside Act. From 1979 to 1983, during his last years in the NCC, Norman was also a Visiting Professor of Environmental Studies at Wye College, then part of the University of London. Always incredibly busy, he found time for everyone. He retired from the NCC in 1983, having been the first British biologist to have spent his entire career in nature conservation, excelling in research, planning and diplomacy. He once commented that you could get a lot done if you let others take the credit. He had seen and battled through the worst effects of agricultural intensification that occurred in Britain from the late 1950s into the 1980s. After ‘official retirement’ he continued to work with FWAG and to chair a specialist dragonfly group of the International Union for Conservation of Nature (IUCN). In retirement, Norman wrote his autobiographical The Bird of Time (1987), which won a Conservation Book Prize, arguing the case for conservation as an underlying principle of wise land-use. Under the name of sustainability, this principle has become enshrined in a limited way in public policy. He chaired the dragonfly group of the Species Survival Commission of IUCN, producing a plan for worldwide dragonfly conservation in 1997. When he retired from NCC, his colleagues paid for a small pond to be created in his garden especially for dragonflies. His final and more personal book, Oaks, Dragonflies and People (2002), describes the history of this pond, and the animals it attracted, together with further thoughts on conservation. Norman was an Honorary Fellow of the Linnean Society and of the Royal Entomological Society, and a Vice-President of the British Ornithologists' Union, the British Ecological Society and the British Naturalists’ Association. He received the BOU's Union Medal (1972) for his work on pesticides, the Zoology Society's Stamford Raffles Award (2001) for his ‘distinguished contribution to the ecology and behaviour of dragonflies’, and the Royal Entomological Society's Marsh Award for Insect Conservation (2002). His name is commemorated in several species of dragonflies and damselflies, and in an award administered by the British Dragonfly Society. In 1950 he had married a fellow Cambridge zoologist, Janet Singer, who died the year before Norman. They are survived by three children and eight grandchildren. Norman died at his home in Swavesey, Cambridgeshire, on 21 October 2015, and will remain long and fondly in the memories of all who knew him. Of all the people one meets through life, he was one of the very best.

Survey of the terrestrial habitats and vegetation of Shetland, 1974 – a framework for long-term ecological monitoring

Abstract. A survey of the natural environment was undertaken in Shetland in 1974, after concern was expressed that large-scale development from the new oil industry could threaten the natural features of the islands. A framework was constructed by the Institute of Terrestrial Ecology on which to select samples for the survey. The vegetation and habitat data that were collected, along with the sampling framework, have recently been made public via the following doi:10.5285/06fc0b8c-cc4a-4ea8-b4be-f8bd7ee25342 (Terrestrial habitat, vegetation and soil data from Shetland, 1974) and doi:10.5285/f1b3179e-b446-473d-a5fb-4166668da146 (Land Classification of Shetland 1974). In addition to providing valuable information about the state of the natural environment of Shetland, the repeatable and statistically robust methods developed in the survey were used to underpin the Countryside Survey, Great Britain's national long-term integrated environmental monitoring programme. The demonstration of the effectiveness of the methodology indicates that a repeat of the Shetland survey would yield statistics about ecological changes in the islands, such as those arising from the impacts of the oil industry, a range of socio-economic impacts, and perhaps climate change. Currently no such figures are available, although there is much information on the sociological impacts, as well as changes in agriculture.

Rural land use in England and Wales and the delivery to the adjacent seas of nitrogen, phosphorus and atrazine

Abstract. In an exploratory study land use in 27 catchments covering most of England and Wales was assessed (using the land use classification devised by the then Institute of Terrestrial Ecology, now a part of the Centre for Ecology and Hydrology) and related to loads and concentrations of total nitrogen, phosphorus (mainly as orthophosphate (Op)) and some pesticides in river flows reaching the surrounding seas in the years 1990–1993. There are good correlations between seven land use types and the concentrations of total nitrogen (r=0.83) and Op (r=0.73) entering the seas. The correlations for individual land use types are best for arable land: total nitrogen, r=0.77; Op, r=0.65; and are positive. However, except for the mixed arable and grassland, the relationships between the other five land use types (improved grassland, upland grassland, upland moorland, upland mixed, and afforested and upland) and contaminant concentrations are negative. If data from the Mersey catchment are discarded, on the grounds that contaminant concentrations in the river seem unduly high, correlations between land use and contaminant concentrations reaching the seas are improved for all land uses: total nitrogen, r=0.89; Op, r=0.83; and atrazine, r=0.78. Loads and concentrations of pollutants in rivers entering the sea could be used to validate models of losses from catchments. Models may need to take greater account of runoff and transport of soil particles than they do presently.

Biomass production of upland vegetation types in England and Wales

Abstract The objective of the study was to describe the annual and seasonal production of live biomass of six plant species or communities of the uplands of the UK –Calluna vulgaris, Vaccinium myrtillus, Nardus stricta, Molinia caerulea, Eriophorum vaginatum and Agrostis–Festuca grassland. The species or communities are important for the management of the uplands to meet both biodiversity and agriculture objectives. The annual and seasonal production of live biomass was determined for sites within six regions of England and Wales in 1995, 1996 and 1997. The six regions were North‐east England, North‐west England, North Pennines, South Pennines, South‐west England and Wales. In each region, the Institute of Terrestrial Ecology's Land Classification Scheme was used to select three or four 2 km × 2 km squares that contained a mosaic of C. vulgaris and upland grassland plant species. Uniform plots of 1 ha were selected from each 2 km × 2 km square, within which three measurements were made on each sampling occasion. Biomass of each species or community was measured within 1·0 × 0·5‐m quadrats using the same standard procedures. Measurements of annual production of current season's shoots of the pioneer, building, mature and degenerate phases of C. vulgaris and of V. myrtillus were made in October, and of the annual production of live biomass of N. stricta, M. caerulea and E. vaginatum in August. A simulated grazing regime was imposed for the Agrostis–Festuca grassland with cuts being made to a height of 4 cm at 8‐week intervals. The annual production of live biomass produced was estimated as the sum of four harvests taken during the growing season. Seasonal measurements of the amount of live biomass of all the species were also made in April, June, August and October. For C. vulgaris, there were significant (P = 0·05) differences between the pioneer, building, mature and degenerate phases in the annual production of current season's shoots, with greater amounts being produced by the phases of greater age and with higher biomass. There were significant year (P < 0·05) and region (P < 0·05) effects, with the lowest production in 1995 and in the South‐west of England. The values were higher than those observed in previous studies in Scotland, and these were interpreted as being associated with a longer growing season rather than higher summer temperatures. There was no difference between years and regions in the annual production of V. myrtillus. For the grass species, there were small differences in annual production of live biomass between years and regions. However, the annual production of live biomass of N. stricta was significantly (P < 0·01) lower in 1995 than in 1996 and 1997 and significantly (P < 0·001) lower in the South Pennines region than elsewhere. The effect of a range of weather and other environmental variables on the annual production of live biomass of each species of grass was explored, but they provided only a limited insight into the reasons for the observed differences and were found to be of limited value as predictors.

Reintroduction of the European beaver (Castor fiber) to Norfolk, U.K.: a preliminary modelling analysis

AbstractAn existing spatial population model, developed to assess the reintroduction of the European beaver Castor fiber to Scotland, was applied to predict the likely result of reintroducing beavers to Norfolk, England. A habitat analysis was conducted using the Land Cover Map (classified satellite imagery developed by the Institute of Terrestrial Ecology) to predict the location of areas suitable for beavers. The model was used to simulate population processes within these areas and dispersal between them following a range of reintroduction scenarios. To represent uncertainty regarding the probable behaviour of beavers in Norfolk, the habitat analysis and model were applied using three sets of parameter values (high, medium and low) derived from the literature. It was predicted that Norfolk could support between 18 and 40 beaver families. Using high or medium parameter values, beaver populations were predicted to expand following reintroduction and spread to new areas. Using low parameter values populations were predicted to decrease, risk extinction and not colonize new areas. This analysis is a preliminary investigation into the likely fate of beavers in Norfolk. Checking the estimates of habitat availability in the field would enable greater confidence to be placed in these predictions.

Habitat associations of breeding Meadow Pipits Anthus pratensis in the British uplands

Habitat associations of Meadow Pipits breeding in British upland 1 km and 10 km squares were investigated using data from the British Trust for Ornithology's (BTO) Breeding Bird Survey (BBS, 1994–97), The New Atlas of Breeding Birds in Britain and Ireland 1988–1991 and habitat data from the Institute of Terrestrial Ecology's (ITE) Land Cover Survey (1994). Meadow Pipit density was highest in 200 m sections of BBS 1 km transects containing unenclosed grass moor, heather and bog, but there was no difference in Meadow Pipit density between these habitat types. Meadow Pipit density (number of individuals per 1 km square) increased with heather cover up to about 7.5 200 m sections (of ten per 1 km square) of heather, but fell thereafter. There was a positive association between Meadow Pipit density and the extent of grass moor cover. The ITE Land Cover data gave results that were consistent with the BBS habitat data at both the 1 km and 10 km square scale. There was a non-linear relationship between hill grass cover and Meadow Pipit abundance, which reached a maximum when hill grass covered between 40%and 60% of a square. Our results suggest that a mosaic of heather, bog and grassland may be the optimum habitat for Meadow Pipits. Recent changes in land management practices may have led to the disappearance or degradation of these mosaics in some areas. Meadow Pipits are widespread and easily counted. We suggest that this species might be a useful bioindicator of changes in heather and grassland ratios that result from land management practices such as grazing and rotational burning, though further research is required.

Koeleria macrantha (Ledeb.) Schultes (K. alpigena Domin, K. cristata (L.) Pers. pro parte, K. gracilis Pers., K. albescens auct. non DC.)

A compactly tufted perennial; tufts stiff and spiky or more rarely soft and gently spreading, 4–60 (70) cm high. Culms slender, 1–3-noded, erect or sometimes slightly curved, downy towards the panicle or hairless. Leaves light grey-green to mid-green on the adaxial and mid-green to dark green on the abaxial surface. Blades curving abruptly to a hooded tip, up to 25 cm long and 1–3.5 mm wide, stiff or soft, usually flat in unstressed plants when fully expanded, ridged on adaxial surface. Both surfaces extremely variable in degree of hairiness, ranging from glabrous to densely hairy (hairs on margins up to 1.5 mm long). Abaxial surface generally more hairy than adaxial, lower sheaths sparsely to densely hairy. Ligules up to 1 mm long, membraneous, usually surrounded by hairs to 2 mm long. Culm leaves usually 2, upper to 7.5 cm long. Panicles spike-like, very dense and often interrupted in the lower part, narrowly oblong, sometimes tapering towards the top; branches spreading during flowering; 1–12 cm long and 2–10 mm wide, silvery-green or purplish, glistening; rhachillas short, usually densely hairy, but sometimes sparsely so; rhachis sparsely to densely hairy. Spikelets 3–6 mm long, 2–3 flowered; densely clustered on short pedicels, to 4 mm long on lower branches; spikelets oblong or wedge-shaped, compressed, breaking up at maturity beneath the lemmas; many remain undispersed on the culm until this falls. Spikelet number variable from around 20 to 200. Glumes persistent, pointed, sometimes with a short awn, and thin membraneous margins; silvery-green or purplish, glabrous to densely hairy; lower usually rather shorter than, but sometimes equal to the upper glume; upper glume oblong or elliptic-oblong, 4–5.5 mm long, 3-nerved; lower narrowly oblong, 1-nerved, 3–4.5 mm long. Lemmas pointed, keeled upwards, sometimes with a short awn, silvery-green or purplish, 3–5.5 mm long; glabrous to minutely hairy on keels and near tips, 3-nerved. Paleas as long as or slightly shorter than the lemmas, 3–5 mm long, 2-keeled, colourless and translucent; lodicules to 1 mm long and 2-toothed. Grain 2.5–3 mm, enclosed within the hardened lemma and palea. Mean caryopsis oven-dry weight 0.32 mg. Root system is a dense mat of gingery-brown fibrous roots, largely unbranched in the upper parts, more strongly branched in the lower; lateral roots often very crinkly. Short wiry rhizomes 1–1.5 mm diameter and to 7 cm long are sometimes present. Native. This is a complex taxon with much variation as a result of polyploidy, and with many ecotypes. The division of the macrantha group in Flora Europaea into nine or more species is based entirely on morphology, and it is considered that these may represent only minor taxa. The species macrantha at the present time is not further subdivided. 46 noted that in Britain Koeleria glauca had been separated from K. macrantha and some Floras produced since then have listed the two species for Britain. Work currently being undertaken by the present author questions this distinction. The distribution of Koeleria macrantha in the British Isles is shown in Fig. 1. The species is widely distributed on Carboniferous, Magnesian and Jurassic limestones, and on chalk, extending from the north of Scotland, including the Orkney Isles, the Outer and Inner Hebrides, and the Isles of Skye, Mull and Arran, through northern Britain and Wales, the Isle of Man and Anglesey to the south of England. It is also recorded from the Isle of Wight, Jersey, Guernsey and Alderney. In Wales it is abundant on the Lias of south Dyfed and on the calcareous substrata of Clwyd and Anglesey, and less abundantly on calcareous substrata elsewhere (4). In the north of Britain it is more common in coastal regions. The species is well distributed in the west of Ireland, less so in the east. The distribution of Koeleria macrantha in the British Isles. (O) Pre-1950; (●) 1950 onwards. Each dot represents at least one record in a 10-km square of the National Grid. 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. Koeleria macrantha is recorded throughout western, central and eastern Europe, but is noted only as a casual in Fennoscandia (Flora Europaea) although K. glauca occurs in Sweden (Skåne, Öland and Gotland) and Denmark. K. macrantha is absent from the Mediterranean islands of Sardinia, Sicily and Corsica, and the Azores; the species is present in Spain but not in Portugal. It occurs in the Baltic, Central, South-West, South-East and the Crimean divisions of Russia, being recorded from all regions in the Caucasus, Upper and Lower Volga, central and southern Urals, Carpathians, Moldavia, Black Sea, Lower Don and western and eastern Siberia to 128 °E, in the Tien Shan, Altai and Pamir mountains of the Central Asian states and in the southern parts of the Aral-Caspian area (14). 29 refer K. macrantha to the Eurosiberian Temperate floristic element. In Asia K. macrantha is recorded from Afghanistan, Iran, Lebanon, Syria, Turkey and Yemen, and in the Himalayan range from India, Kashmir, Nepal, Pakistan and Tibet. It is also recorded from the western Chinese provinces of Kansu, Shantung, Kiangsu and Hopeh, from the northern Chinese province of Shansi, from Japan, Korea, Manchuria and Mongolia. Koeleria macrantha is widespread in North America and is recorded from Arizona, Colorado, Kansas, Nebraska, North Dakota, Oregon, Utah, Washington and Wyoming (37); Louisiana, New Mexico and Mexico (30); California, Delaware, Indiana, Missouri, Ohio and Texas, and in Canada from Quebec to British Columbia (26). The species extends to 60 °N in Europe, to 62°N in Asia and to 55°N in North America. Koeleria macrantha is also recorded from central Africa – Ethiopia, Eritrea, Guinea and Cameroon, from the Transvaal and Natal areas of South Africa and from Swaziland (Vergl. Chor.). In England K. macrantha has an altitudinal range from sea level to 550 m (North Yorkshire), from sea level to 412 m in Scotland (eastern Highlands), from sea level to 457 m in Ireland (Donegal) (Alt. Range Brit. Pl.), and from sea level to around 185 m in Wales. It is found to 2700 m in the Alps (37), to 2200 m in the Pyrenees (45), to 2300 m in Turkey, to 2100 m in Afghanistan, to 3200 m in China and to 4400 m in Nepal. In North America the species is found to 3400 m and in Africa to 3700 m on Mount Kilimanjaro (Vergl. Chor.). Koeleria macrantha is exposed to average monthly maximum temperatures to around 39 °C in August in Spain, to around − 50 °C average monthly minimum temperatures in January in Siberia and to − 30 °C in January in European Russia (46), and to annual precipitation of around 2000 mm on the west coast of Scotland and in the Alps, and to conditions with no precipitation being recorded during July, August and September at Pagham in Afghanistan. No evidence in Britain of wind or insolation damage has been noted, nor from seaspray other than to make the leaves more glaucous. The species is subject to strong gales on northern coasts, such as on the islands of Canna and Sanday, Scotland (4) and to Föhn-swept slopes in Germany, and is found on steep dry slopes in the Jura Alps subject to strong insolation (14). Koeleria macrantha is found on both flat ground and steep slopes; it is common on slopes of 20–50 ° (Ecol. Atl.). In the British Isles the species shows no aspect preferences. In Europe it occurs rather less frequently on east-facing, and in Asia rather less frequently on west-facing slopes. It is a component of dry, semi-natural pastures, scree slopes, rock outcrops, old lead mines, limestone and quarry spoil (29) and also of limestone pavements. It occurs on sea cliffs and sand dunes, shelly or otherwise, particularly on partially fixed dunes, but is also present in the later successional closed turf, in both very short and tall vegetation to c. 60 cm; it is found frequently on coastal golf links and in the machair of the Outer Hebrides and Iona. The species is also found on pebble strands and shingle banks, on rocks present at the tops of beaches, and on sandy beaches where limestone runs into the sea, as in the Burren, Co. Clare; it is also recorded from limestone heaths here (37). The species has been recorded from heathy pastures in the Isle of Man, from peaty soil overlying limestone in Ireland (Bot. Irl.), from tall chalk heath at Lullington Heath, Sussex (30), from limestone heath in Somerset and Wales (26), and from sandy heath near Branderburg, near Lossiemouth, Scotland, where tufts of K. macrantha were growing up through the middle of mats of Calluna vulgaris and Empetrum nigrum. Koeleria macrantha is recorded from ungrazed, partially shaded moist soil overlying Magnesian limestone in Derbyshire (45), from dry woodland in Pakistan, under Juniper scrub in Nepal, in deep shaded ravines in Nepal, and from wet marshy land in Kashmir. It is also recorded from open woods in North America, and extensively from prairies (35). Koeleria macrantha is widespread on calcareous brown earths, brown earths and rendzinas overlying chalk and limestone in England, and is less widespread in Wales and Scotland. It is recorded from moderately base-rich brown forest soils in Scotland (28), from partially stabilized aeolian deposits in Scotland (9) and on almost pure sand, with few shell fragments, at many places around the coast of the British Isles. It is reported from gleys of moderate to high base status in Scotland (10), and from incipient podzols in England (7). The species is recorded from peaty or sandy peat substrata along the coasts of western Ireland (79), and from dry eskers in Offalay and Galway (Ir. Pfl.). It is found on soils derived from basalt, dolerite (4), serpentine (57), slate, hornfels and phyllite in Scotland, and has been observed by the author on a granite cliff, north of Aberdeen. It is found over serpentine in Cornwall, and on ‘sandstone cliffs’ in Hampshire (72). Soil analyses (methods according to Chem. Anal.) for sites supporting K. macrantha gave a range for exchangeable calcium and magnesium, extracted with M ammonium acetate (pH 9.0), from 130 to 6500 mg kg−1 and from 560 to 3880 mg kg−1, respectively, while potassium ranged from 110 to 540 mg kg−1. Phosphate phosphorus, extracted with 0.5 m sodium bicarbonate (pH 8.5), ranged from 1.2 to 32 mg kg−1, and total nitrogen from 0.028 to 1.10%, to a high 1.8% from a site on basalt. The low values for all elements were from sand dune sites. Koeleria macrantha is recorded as occurring in soils of pH between 4.5 and 8.0 and very rarely between pH 3.0 and 4.5 (Ecol. Atl.). On chalk it is found at pH values up to 8.4 (14). The species grows well in minimal soil depths on tops of walls, in limestone pavement hollows and in rock crevices, with only 2–5 cm soil, but it is also found on much deeper soils (to 60 cm +), particularly on sand dunes. In Continental Europe, in addition to soils overlying calcareous rocks, K. macrantha occurs in gritty sandy soil over porphyr (40) and over gypsum (47) in Germany, in soils over serpentine in Greece, with very high nickel and chromium contents – 4400 and 880 µg g−1, respectively (39), and over schists in the Czech Republic (54). The species is recorded from dry, acid soils in Germany (24), on sand dunes in Holland and north-west Germany, and in chalk-free, humus-rich surface soils over chalk also in north-west Germany (14). It is further recorded from zinc-rich soils, containing between 0.5% and 8.7% zinc, in the neighbourhood of Aachen, Germany and also from Belgium (14). Koeleria macrantha is recorded from gneiss and biotite muscovite in the Karakorum Mountains of north Pakistan (32). The leaf sheaths of K. macrantha are white, papery and moderately persistent, as are the leaves, decomposing slowly, and so adding to the litter layer. Koeleria macrantha is predominantly a species of calcareous grasslands and occurs with a high constancy in the following communities defined by the National Vegetation Classification (61, 62). CG1 Festuca ovina–Carlina vulgaris grassland is found on freely (often excessively) draining rendzina soils with a high base status, on steep and rocky, but stable, slopes on hard limestones, usually with a southerly to westerly aspect and so with a tendency to summer droughting. Often heavily grazed by sheep and rabbits, this community is confined to Devonian limestone in Devon and to Carboniferous limestone in Wales and the Mendips. Koeleria macrantha has an overall constancy of III in CG1 but attains a constancy of V in the subcommunities of Helianthemum canum and Festuca rubra–Scilla verna and a constancy of IV in the K. macrantha subcommunity. Constant species, as defined by the NVC, for CG1 and the following communities are given in Table 1. CG2 Festuca ovina–Avenula pratensis grassland consists of a rich mixture of grasses and dicotyledons in a closed sward and is traditionally grazed by sheep and rabbits. It occurs most frequently in relatively dry and warm lowland climates on free-draining calcareous soils derived from calcareous bedrocks; often prone to summer droughting. Koeleria macrantha has a constancy of V in this community. CG3 Bromus erectus grassland achieves maximum extent on lightly grazed or ungrazed grasslands over chalk and oolite in the south-eastern areas of Britain, and K. macrantha has a constancy of III in this community. CG4 Brachypodium pinnatum grassland is also mainly associated with lightly or ungrazed calcareous swards in south-eastern Britain, but is more generally found on the cooler and damper areas of the chalk and oolite than is the Bromus erectus grassland, which is more Continental in character. Koeleria macrantha has an overall constancy of only I in this community, but of III in the Avenula pratensis–Thymus praecox subcommunity. CG5 Bromus erectus–Brachypodium pinnatum grassland occurs where B. pinnatum is favoured by amelioration of extreme Continental conditions, but not so greatly as to exclude Bromus erectus. The community is most characteristic on the calcareous, base-rich soils of the oolite on the north-western fringe of southern, lowland limestone. Koeleria macrantha achieves a frequency of III in the typical subcommunity, but an overall frequency of II. CG6 Avenula pubescens grasslands are found on a variety of gently sloping lowland limestones, mainly in the south of England. The soils are deep and moist, though mostly free-draining alluvial rendzinas, or calcareous brown earths, and in some cases more mesophytic soils occur on flat limestones. This community is lightly grazed by cattle; on sloping limestones it is grazed by rabbits. Koeleria macrantha has a constancy of III in CG6. CG7 Festuca ovina–Hieracium pilosella–Thymus praecox/pulegioides grassland is the characteristic vegetation type of thin, stony soils over chalk, or less frequently on south-facing gentle slopes of Carboniferous limestone. The soils are dry and impoverished; this is reflected in the species composition, with grasses generally exhibiting reduced vigour and herbaceous dicotyledons playing a predominant role. The short, open nature of the vegetation is maintained by heavy grazing, mainly by rabbits. Koeleria macrantha has an overall constancy of III, a constancy of IV in the K. macrantha subcommunity and V in the Cladonia spp. subcommunity. CG8 Sesleria albicans–Scabiosa columbaria grassland is found only on free-draining calcareous, steep slopes of Magnesian limestone in Durham, principally on rendzinas rich in calcium and magnesium carbonates. The climate is cool and dry, and the community is a plagioclimax vegetation maintained by grazing of domestic animals and rabbits. Koeleria macrantha has a constancy of IV in this community. CG9 Sesleria albicans–Galium sterneri grassland is found in the northern Pennine submontane or montane climate, over shallow, freely draining but moist, calcareous lithomorphic soils on drift-free Carbonifeous limestone exposures. It forms an important part of upland farm hill-pasture and as such is frequently grazed, mainly by sheep. In CG9 K. macrantha achieves a constancy of IV. CG13 Dryas octopetala–Carex flacca heath occurs in the cool oceanic lowlands of north-west Scotland, on shallow rendzinas over the calcareous rocks of the Durness limestone, and on calcareous shell sand blown from the shore. The community is mainly ungrazed on the mainland but is quite heavily grazed by sheep and rabbits on Skye. Koeleria macrantha has a constancy of III in the Salix repens–Empetrum nigrum subcommunity and an overall constancy of II in CG13. Koeleria macrantha occurs with a lower constancy in the calcareous grassland community of Festuca ovina–Agrostis capillaris (CG10). The species is also found at a low constancy in the mesotrophic grassland community of Cynosurus cristatus–Centaurea nigra (MG5) and in the calcifuge sward of Festuca ovina–Agrostis capillaris–Rumex acetosella grassland (U1). Koeleria macrantha is also present with low constancy in the heathland community of Calluna vulgaris–Scilla verna (H7), and occurs with a higher constancy in: H6 Erica vagans–Ulex europaeus heath, a subshrub vegetation confined to the Lizard in Cornwall, on freely draining brown earth poor in calcium, but usually base-rich. The community survives most extensively over serpentine. Constant species are: Agrostis canina ssp. montana (V), Carex flacca (III), Erica cinerea (V), E. vagans (V), Filipendula vulgaris (IV), Ulex europaeus (V), U. galli (IV) and Viola riviniana (IV). Koeleria macrantha achieves a constancy of V in the Festuca ovina subcommunity and an overall constancy of II. The species is also recorded from several maritime communities. It is found with a constancy of IV in the Arenaria serpyllifolia–Sedum acre subcommunity and an overall constancy of III in the OV39 Asplenium trichomanes–Asplenium ruta-muraria crevice community, on limestone. Constant species for this community are Asplenium trichomanes (V), Asplenium ruta-muraria (V), Porella platyphilla (IV) and Homalothecium sericeum (IV). Koeleria macrantha occurs at low frequency in the following communities: OV 34 Allium schoenoprasum–Plantago maritima community, restricted to areas of seasonally damp soils over serpentine on the Lizard cliff margins; OV37 Festuca ovina–Minuartia verna in limestone areas contaminated with heavy metal spoil; MC5 Armeria maritima–Cerastium diffusum maritime therophyte community; MC9 Festuca rubra–Holcus lanatus cliff grassland; MC10 Festuca rubra–Plantago species maritime grassland; MC11 Festuca rubra–Daucus carota maritime grassland. Koeleria macrantha is also found at low constancy in the following dune communities: SD8 Festuca rubra–Galium verum fixed dune community; SD9 Ammophila arenaria–Arrhenatherum elatius dune grassland; SD10 Carex arenaria dune community and SD12 Carex arenaria–Festuca ovina–Agrostis capillaris dune grassland. Koeleria macrantha has also been recorded by the author from an extensive area of shingle mixed with sand (c. 2 m O.D.) in Suffolk. Accompanying species were Achillea millefolium, Artemisia maritima, Bromus hordeaceus ssp. hordeaceus,Catapodium marinum, Hordeum marinum, Lolium perenne, Lotus corniculatus, Ononis repens, Papaver rhoeas, Plantago coronopus, Puccinellia maritima and Reseda lutea. Koeleria macrantha occurred frequently in this community, and also in a community recorded from a raised beach (c. 3 m O.D.) also in Suffolk, with sparse, very short vegetation over almost pure sand, with very few shell fragments. The accompanying species here were: Agrostis capillaris, Dactylis glomerata, Elymus pycnanthus, Festuca ovina, Hordeum marinum, Hypochaeris radicata, Lolium perenne, Lotus corniculatus, Medicago lupulina, Plantago coronopus, P. lanceolata, P. maritima and Sedum acre. In Scotland, 9 recorded K. macrantha with a high constancy from two coastal dune communities. The Euphrasio–Festucetum arenariae association occurs on immature, brown calcareous soils over fixed dunes, with gentle to moderate slopes, particularly on the Caithness coast; the differential species are Bellis perennis, Euphrasia spp., Gentianella campestris, Ranunculus acris and Viola tricolor ssp. curtisii. The Astragalo–Festucetum association also occurs on fixed dunes with gentle to moderate slopes on the drier eastern coast of Scotland; the differential species are Astragalus danicus, Festuca ovina, F. rubra, Galium verum and Tortula ruralis. Koeleria macrantha has a constancy of IV in both these dune associations. 9 also described K. macrantha from an ultra-basic grassland community in Ayrshire. The Plantago maritima–Sieglingia decumbens community is found on freely drained, brown forest soils with a high magnesium content, derived from ultra-basic rocks. Species with a high constancy are Achillea millefolium, Festuca ovina, Trifolium repens, Viola riviniana and K. macrantha, with a constancy of IV. The species has been recorded by the author from a flat, sandy foreshore at Auchenmalg Bay, Galloway, in a tall (60 cm +), open community dominated by Raphanus maritimus. Koeleria macrantha occurred frequently in this community, which also included Anthyllis vulneraria, Arrhenatherum elatius, Cochlearia officinalis, Hieracium pilosella, Plantago lanceolata, Poa pratensis, Rumex acetosella, Tripleurospermum maritimum and Vicia hirsuta. In Ireland, 79 lists K. macrantha as a component of the Plantaginetum coronopo–maritima association. This dwarf coastal community, which is subject to frequent strong winds and sea-spray, is found on peaty or sandy-peat substrata on flat or not very steep cliff tops in western Ireland. The vegetation is heavily and persistently grazed by sheep and rabbits. On Malahide Island, north of Dublin, K. macrantha was recorded by 52 from the Festuco–Galietum maritimi association, which is found on middle dunes of grass-dominated, ungrazed or very lightly grazed, vegetation. Differential species are Anthoxanthum odoratum, Briza media, Dicranum scoparium, Euphrasia occidentalis, Koeleria macrantha, Potentilla sterilis, Pteridium aquilinum, Rosa pimpinellifolia, Thymus praecox and Trifolium repens. Koeleria macrantha is a component of the association Camptothecium–Asperuletum, characterized by Asperula cynanchica, Neotinea maculata and Gentiana verna, and described from stable, landward dunes in Galway (Ir. Pfl.) and also from Co. Clare by 37. The association Violo curtisii – Tortuletum ruraliformis occurs on dry stabilized coastal dunes and is characterized by Centaurium erythraea, Cerastium diffusum, Koeleria macrantha and Phleum arenarium. Originally described from Kerry and Co. Mayo (Ir. Pfl.) it has also been recorded from Co. Clare (37). Koeleria macrantha occurs in the Antennarietum hibernicae association (Ir. Pfl.) found in hollows between dunes on the coast of Co. Clare and also in dry eskers in Offalay and Galway. The soils have a high humus content and are not very calcium-rich. Characteristic species are Antennaria hibernica, Anthyllis vulneraria, Asperula cynanchica, Carex arenaria, C. caryophyllea, Carlina vulgaris, Danthonia decumbens, Festuca rubra, Gentiana campestris, Lotus corniculatus, Plantago lanceolata, Polygala vulgaris, Ranunculus bulbosus and Thymus pulegioides. This association was also recorded from Co. Clare by 65. 37 described K. macrantha from the Hyperico–Dryadetum association from the Burren, Co. Clare. This is a shrub-dominated community found on leached organic soil over bare limestone. Characteristic species are Calluna vulgaris, Danthonia decumbens, Dryas octopetala, Empetrum nigrum, Hylocomium splendens, Hypericum pulchrum, Hypnum cupressiforme var. ericetorum, Neckera crispa, Polygala vulgaris and Rosa pimpinellifolia. Koeleria macrantha is also found in the mesotrophic grassland Cynosurus cristatus–Centaurea nigra (MG5) which is the typical grassland of well-drained pastures over limestone, usually grazed by cattle in Ireland. In Continental Europe, K. macrantha is an order character of the Brometalia erecti and occurs in a number of associations belonging to this order throughout western Europe; these occur on dry steep slopes throughout the Alpine range, in dry rocky places in France and Belgium, the deeper moister soils of the Danube and Rhine valleys, and also on north-facing, more humid slopes of the Alps. Koeleria macrantha is also a component of some of the associations of the order Festucetalia valesiacae in various parts of Central Europe, including the Jura, Nahe Valley, Rheinhesse and the Bohemian Mountains. The species is also found in some associations of the class Arrhenatheretalia in the driest parts of the range. A range of communities containing K. macrantha, with a frequency of II and above, is given in Tables 2 and 3 to illustrate its geographical and ecological range. 15 measured shoot thrust in various herbaceous species and found that those species which were dominants in their usual habitats generated the greatest thrust. Koeleria macrantha with fine shoots demonstrated less thrust and much of this was directed horizontally through the canopy. They concluded that its fine aerial shoots might confer a high ability for ‘fine-scale’ foraging for light between the dominant species, a strategy in keeping with its usual intermediate status in lightly grazed or ungrazed grasslands. Koeleria macrantha also occupies an intermediate status in the prairies of the North American Midwest, and although its rapid recovery after severe droughts initially results in a monoculture, it is not a good competitor and is gradually relegated to its intermediate role by more vigorous species such as Andropogon spp., Poa pratensis and Stipa spp. (1). Koeleria macrantha exhibits a more dominant role in its coastal habitats on cliff tops, shingle and sand dunes where vegetation is often shorter, though often ungrazed; its flattened, tufted habit is an advantage here. 29 recorded K. macrantha as being notably absent from heavily disturbed situations and where there is intense competition, and 7 noted that it was rather less frequent on steep slopes in Derbyshire with very shallow soil and loose limestone fragments, where the dominant species was Festuca ovina, than on slopes with greater stability and less loose material and with Festuca ovina and F. rubra as co-dominants. Koeleria macrantha is classed as a good forage grass in North America (35), but 66 comments that it is considered of little agricultural value in Europe because it develops few leaves and these soon become tough. However, in long vegetation on deeper soils it tends to have longer, softer leaves than when growing on, for example, rock outcrops over shallow soils. 77 observed that K. macrantha increased in a sward over a period of 34 years, after being enclosed against domestic animals and rabbits. Other abundant species in the enclosure were Festuca ovina and Avenula pratensis which also increased; these are species also frequently found with K. macrantha in grazed calcareous grassland. Watt commented that these three species were much more vigorous inside, than outside, the enclosure. Nevertheless, K. macrantha is recorded as a component of a woodland replacement plagioclimax community in central Perthshire, Scotland, which is maintained principally by sheep grazing, but is also grazed by cattle, red deer, rabbits and hares (28), and the species has a high constancy (IV) in the NVC plagioclimax community of CG8 (Sesleria albicans–Scabiosa columbaria grassland) also maintained by the grazing of domestic animals and rabbits. However, K. macrantha does not seem able to withstand continual heavy grazing. 16 reported that the effects of overgrazing were to produce a decline in its abundance, vigour and productivity, and a simulated grazing experiment by the author, in which clipping was carried out at three height regimes, every 2 weeks for 10 weeks, resulted in the greatest production of shoot material when the plants were only lightly clipped (at 15 cm above soil level). An experiment carried out by 11 extending for 4 months during the summer with clipping every 14 days showed that K. macrantha increased the diameter of its root mass and stele as a result of the clipping, to a greater degree than Bromus inermis and Poa pratensis. 44 observed that burning stimulated inflorescence production in K. macrantha in the summer of the year following a spring burn, but by the following year no significant stimulus to flowering remained, while 16 noted that burning led to reduced yield of K. macrantha for a period of 3–5 years in Canadian prairies subject to moderate grazing. However, 1 reported that where fire was used as a means of weed control in the prairies of Nebraska, K. macrantha became temporarily widespread. Koeleria macrantha is very tolerant of trampling and has been noted, for example, at one site as being the only species present, providing almost 100% cover, apart from a small amount of Festuca rubra and Poa annua. It has been observed on many paths and tracks, particularly at coastal sites, but also on inland limestones, as very flattened, stiff little tufts, flowering on very short culms (see V (b)). Koeleria macrantha forms small tufts in cliff faces, on walls and in limestone pavement hollows. It is usually found as scattered individuals or more rarely with a number of tillers joined by short rhizomes, in short grazed and ungrazed turf, but on paths and tracks which are not too heavily trampled it often forms almost a monoculture, as it does sometimes on steep roadside banks, especially on sandy soils. In moderately tall vegetation to around 60 cm it is again usually found as scattered individuals of straggling appearance, rather than as tufts. Koeleria macrantha never produces very large tussocks and plants cultivated in a fertile garden soil produ

Densities and population estimates of breeding Skylarks Alauda arvensis in Britain in 1997

We present the results of a national survey of breeding Skylarks Alauda arvensis in Britain in 1997 carried out by the British Trust for Ornithology (BTO). Numbers of Skylarks and land-use types were recorded by volunteers in 608 1-km squares, selected using random stratification based on the Institute of Terrestrial Ecology's (ITE) Landscape Classification to avoid over- or under-sampling particular habitat types. The results suggest a maximum national population of around 1000 000 pairs. This agrees extremely well with the national population estimate of around 1046 000 pairs derived from the Breeding Bird Survey (BBS) of the BTO/Joint Nature Conservation Committee/Royal Society for the Protection of Birds, and provides further evidence for the robustness of the BBS as a national monitoring scheme for common and widespread species. The figure is approximately half that of the most recent published estimate of 2000 000 in 1988–91. However, the scale of this discrepancy is likely to reflect a bias in Atlas field methodology and site selection since there has been a real decline of approximately 9% since 1990. Arable squares supported the highest densities of Skylarks; 4.6–6.0 pairs per km2, and 46–49% of the British breeding population was associated with arable areas. Marginal upland and upland areas supported lower densities but still accounted for approximately 34% of the estimated national breeding population. Differences in density at broad habitat scales were reflected in geographical differences across Britain, with southern and eastern arable regions supporting much higher densities than western and northern ones. At a finer scale, Skylarks occurred at highest densities on, and showed highest habitat preferences for, set-aside and various types of ungrazed grassland. Winter cereal, improved grassland and set-aside held the highest proportion of the Skylark population on farmland in England and Wales; grazed pasture, winter cereals and spring cereals held the highest proportion in Scotland.

Evaluation and interpretation of regional and site-specific hydrochemical data bases for water quality assessment

Abstract. The spatial distribution of stream water composition, as determined by the Geochemical Baseline Survey of the Environment (G-BASE) conducted by the British Geological Survey (BGS) can be successfully related under baseflow conditions to bedrock geochemistry. Further consideration of results in conjunction with site-specific monitoring data enables factors controlling both spatial and temporal variability in major element composition to be highlighted and allows the value of the survey to be enhanced. Hence, chemical data (i) from streams located on Lower Silurian (Llandovery) bedrock at 1 km2 resolution collected as part of the G-BASE survey of Wales and the West Midlands and (ii) from catchment monitoring studies located in upland mid-Wales (conducted by Institute of Terrestrial Ecology), have been considered together as an example. Classification of the spatial survey data set in terms of potentially controlling factors was carried out so as to illustrate the level of explanation they could give in terms of observed spatial chemical variability. It was therefore hypothesised that on a geological lithostratigraphic series of limited geochemical contrast, altitude and land-use factors provide better explanation of this variability than others such as lithology at sampling site and stream order. At an individual site, temporal variability was also found to be of considerable significance and, at a monthly time-step, is explicable in terms of factors such as antecedent conditions and seasonality. Data suggest that the degree of this variability may show some relationship with stream order and land-use. Monitoring data from the region also reveal that relationships between stream chemistry and land-use may prove to be strong not only at base flow but also in storm flow conditions. In a wider context, predictions of the sensitivity of stream water to acidification based on classifications of soil and geology are successful on a regional scale. However, the study undertaken here has shown that use of such classification schemes on a catchment scale results in considerable uncertainty associated with prediction. Uncertainties are due to the large degree of variability in stream chemistry encountered both spatially within geological units and temporally at individual sampling sites.

Cypripedium calceolus L.

A perennial herb with horizontal rhizome (diameter about 6 mm) up to 10 cm underground. Roots long and fleshy, diameter up to 2 mm. Stem 20–60(70) cm, glandular-pubescent with brown basal sheaths; leaves 3–5, elliptical to ovate–oblong, acute to acuminate, sparsely pubescent (oval lanceolate, somewhat pleated, large), (7)11–17(22) cm long, (3)5.5–8(10) cm wide, veins obvious, bracts leaf-like, exceeding flowers; flowers 1–2(3) large; perianth segments purple-brown; dorsal outer tepal (‘sepal’) lanceolate, suberect, 3.5–6 cm long and 1.5–2.5 cm wide; lateral ‘sepals’ joined, hanging below labellum 3.5–5.5 cm, free tips up to 5 mm; inner tepals (‘petals’) 4–7 cm long and 0.5–1 cm wide, horizontal, twisted; labellum yellow, shoe-shaped (length 3–4 cm, width 2–3 cm, depth 1.5–2 cm), reflexed, covered with viscous hairs inside, not spurred; staminode petaloid shield-shaped, white with red spots, 1–1.2 cm long, 0.7–0.9 cm wide; ovary elongate, somewhat curved, pubescent (glandular hairs), not twisted, pedicellate. Two fertile anthers, one on each side of style; stigma and staminode forming a flat appendage on the front of the column. Capsule up to 3 cm long and about 0.9 cm in diameter, containing 6000–17 000 seeds; seeds oblong about 1 mm long with a mass of a few micrograms (seed volume 0.02 mm3). From continental Europe, the all-yellow flowered var. flavum Rion has been recorded (Davies et al. 1984). This exceedingly rare orchid was formerly more widespread in northern England in limestone woodland; it is more plentiful in wooded sites in northern Europe, but everywhere has suffered from collecting. Cypripedium calceolus is mainly a boreal species. Formerly it grew in Britain in the three northern counties of Yorkshire, Durham and Westmorland (Summerhayes 1951), but nowadays is definitely native at only a single site on Carboniferous limestone in Yorkshire (Wood et al. 1984) (Fig. 1), although other doubtfully native sites exist within the same counties and north Lancashire. The distribution of C. calceolus ranges from Great Britain and Scandinavia across northern and central Europe to north-east Spain and northern Italy, and from western Europe through southern Siberia to Rebun Island (Vergl. Chor.; Borodin et al. 1984; Hultén & Fries 1986; Cribb 1997) (Fig. 2). A similar decline to that in Britain has taken place in most of Europe (Wood et al. 1984; Ramsay & Stewart 1998). The distribution of Cypripedium calceolus L. in the British Isles. (○) Pre-1950; (●) 1950 onwards; (+) introduced pre-1950; (×) introduced 1950 onwards. Mapped by Mrs J. M. Croft, Biological Records Centre, Institute of Terrestrial Ecology. The distribution of Cypripedium calceolus L. in Europe and Asia (Hultén & Fries 1986). The species is found close to sea level up to an altitude of 2700 m in Switzerland (Davies et al. 1984). In Britain the range of the altitude was 150–260 m (Alt. range Br. Pl.). Cypripedium calceolus occurs in shady deciduous and mixed woodland (rarely in full sunlight at higher altitudes) or, less frequently, on stone-strewn slopes, predominantly on calcareous soils (Füller 1981; Davies et al. 1984; Wood et al. 1984; Lid 1987; Delforge 1995). It can withstand cold winters and in northern parts of Eurasia it tends to grow in calcium-rich spring fens and on marshy grasslands (Ivanter & Kuznetsov 1995). The plant is favoured by suboceanic to subcontinental climates and is rare in regions with Atlantic and Mediterranean climate (Vergl. Chor.; Heinrich & Lorenz 1996; Terschuren 1998). Cypripedium calceolus is favoured by moderately moist nutrient-poor to moderately nutrient-rich, particularly nitrogen-poor base-rich neutral to moderately acid soils. In central Europe the pH in its habitats is higher than 7.1 (Procházka & Velísek 1983; Möller 1985, etc.). In habitats of eastern Europe, the soil pH (KCl) range is 5.4–7.7 (Moisejeva 1970; Varlygina & Matsenko 1986; Kull 1997). Closely related taxa from America are known to grow more vigorously at higher pH values (Stuckey 1967). In a soil analysis (horticultural) carried out by the Agricultural Development and Advisory Service in 1994 (L. Farrell & P. Corkhill, personal communication) from the remaining site of the plant in England, the pH was 7.9 and mean values (mg l–1) of the major elements were: nitrate-N 4, P 6, K 89 and Mg 63; conductivity was 2050 µS cm–1. Calcium carbonate ranged up to 33% and particle size was predominately 2000–200 µm. Data on comparative soil analysis from Germany (9–13) and Estonia (1–8) are given in Table 1. In Britain, C. calceolus once formed part of the relatively open Quercus, Fraxinus and Corylus woodland or scrubland flora of limestone scree slopes in northern England (Summerhayes 1951; Wood et al. 1984). The only surviving native clump in the wild in Britain grows in a small shady clearing with scrubby tree growth at a woodland edge on freely draining black proto-rendzina soil derived from limestone with little organic material and abundant limestone fragments (P. Corkhill, personal communication). Sesleria caerulea is prominent in the sward, and the wide range of associated species includes Carex panicea, Epipactis atrorubens, Galium sterneri, Helianthemum nummularium, Leontodon hispidus, Linum catharticum, Polygala amarella, Primula farinosa, Sanguisorba minor and Succisa pratensis (Farrell 1999). In the National Vegetation Classification, Rodwell (1992) describes this community as CG9c Sesleria albicans–Galium verum grassland, Carex pulicaris–C. panicea subcommunity. In Europe and Asia, it is found in a variety of habitats including both deciduous and coniferous woodland, open scrub and alpine meadows. However, the majority of its habitats are woodland, described according to the Council of Europe Palearctic typology as neutrophile (Terschuren 1998), limestone and southern beech forests, oak–hornbeam forests, ravine and riverine forests, thermophilous forests, mesophile birch and alder woods, fir forests, spruce, larch and arolla, mountain pine forests, mire conifer woods, mountain dwarf juniper scrub, calciphile thickets and thermophile forest fringes. Besides woodland, tall herb and herb-rich taiga, subatlantic Sesleria semi-dry calcareous grassland and terrestrialization zones of rich fens are also mentioned (Terschuren 1998). In Sweden and Estonia, wooded meadows are often rich in C. calceolus. In France, it occurs in communities of Gymnocarpio–Seslerietum, Violo rupestris–Seslerietum, Geranion sanguinei–Gentiano–Daphnetum laserpitietosum, Carici umbrosae–Brachypodietum, Berberidion, Sorbo–Franguletum, Carici albae–Fagetum, Carici montanae–Quercetum, Dentario–Fagetum and Aceri–Fraxinetum (Didier & Royer 1994; Heinrich & Lorenz 1996). In Poland, the species occurs in Galio–Carpinetum lathyretosum verni, Tilio–Carpinetum melittetosum, Carici–Fagetum, Mercuriali–Fagetum, and Cephalanthero rubrae–Fagetum communities (Wika & Bernacki 1984). In southern taiga, associations with Juniperus–Cypripedium–Fragaria, Betula–Cypripedium–Convallaria, Betula–Pinus–Convallaria, Betula–Pteridium–Convallaria and Betula–Centaurea jacea–Trifolium montanum are referred to as habitats of C. calceolus. In Estonia, the species occurs in alvar (habitat with thin soil layer on limestone) forests and shrublands, in mesic boreal forests (Oxalis site type), in fresh boreo-nemoral forests (Hepatica site type), in drained peatland forests (Oxalis site type) and in wooded meadows (Kull 1991, 1995). Vascular plant species growing often with C. calceolus are Aegopodium podagraria, Anemone nemorosa, Convallaria majalis, Fragaria vesca, Hepatica nobilis, Paris quadrifolia and Rubus saxatilis. In northern conditions in Karelia, sometimes the species may occur in communities with Pinus sylvestris, Phragmites australis and Sphagnum warnstorfii (Dyachkova et al. 1997). Cypripedium calceolus suffers severely from competition with tall plants (herbs, grasses) favoured by increased light conditions, e.g. after clear-cutting, the result of which is population decline (Sjöberg & Ericson 1992). As fires in low forest layers are known to be favourable for C. calceolus, it means that the species is sensitive to the presence of a dense bush layer. Increasing shade is a limiting factor for C. calceolus. Dementjeva (1985) considers 1/14–1/20 of full light optimal in the southern taiga zone and demonstrates that the lighter the locality the shorter the ramets and leaves are. The same has been shown in Poland, and also that the proportion of generative ramets is larger in light conditions (Wika & Bernacki 1984; Brzosko & Werpachowski 1991). However, in Estonian populations (Table 2) height of flowering ramets is correlated with the light conditions but not the percentage of flowering ramets. In shady localities small clones dominate and it is more probable that some of them do not have above-ground shoots in some years, the buds staying in dormant phase underground (clone dormancy exists) (Kull 1995). Being a rhizomatous plant, C. calceolus grows in smaller or larger patches that may belong to one or several clones. Clones identified by isoenzyme analysis are seldom larger than 70 cm in diameter. Clones that have reached a diameter of 20–40 cm virtually cease to increase in diameter, which can be explained by an increase in the death rate of rhizome branches in bigger (denser) clones (Kull 1997). Most populations consist of a few plants; only in a few localities can thousands of specimens be found. Counts have been carried out on several populations over different time periods. Counts over a few years have often indicated stability (Salmia 1988; Korczynski & Spiewakowski 1991; Blinowa 1993). Plants at a single station and at all stations within a limited geographical area behaved similarly in any one year (Curtis 1954). Long-term counts made in Denmark (Høst 1989; Wind & Ballegaard 1996) and Germany (Drüke 1980) showed a drastic population decrease, whereas 15-year monitoring of eight populations in Estonia showed a certain increase (Kull 1997) (Fig. 3). Since being protected, the surviving English plant has spread and more flowers have been produced, with new plants arising nearby (L. Farrell, personal communication). Dynamics of the total number of ramets of C. calceolus in permanent plots of eight populations in Estonia. The horizontal axis shows the years and the vertical axis the total number of ramets. The permanent plot sizes were: Ussisoo, 126 m2; Tooma, 600 m2; Muhu I, 30 m2; Muhu II, 72 m2; Õisu, 60 m2; Hiiumaa, 2 m2; Puhtu, 65 m2; Laelatu, 3 m2. Late frosts may affect flowers, leaving brown spots on the labellum (Eberle 1973). Severe frosts in June reduce shoot survival and flowering in clear-cut areas at high latitudes (Sjöberg & Ericson 1992), but no frost damage has been observed in Britain. Cypripedium calceolus is sensitive to drought. In particular, young seedlings require constant moderate moisture (Corkhill 1996). The creeping horizontal rhizome, with a diameter of 0.4–0.9 cm, is situated at a depth of up to 10 cm. Every year, the rhizome usually produces two apical buds. From the larger bud the following year's shoot develops, forming the new increment (average length 1 cm) of the sympodial rhizome. The position of the larger bud (left or right) alternates each year, resulting in the characteristic zigzag growth pattern of the rhizome (Kull & Kull 1991). The smaller, or dormant, bud may start growing in the year of formation or in the following year (Fig. 4). However, in some cases the smaller bud may start growing in the first year and in other cases several years later, but usually it will stay dormant until it dies (e.g. after 15–20 years). The angle between subsequent annual increments formed from the bigger buds is about 130 °. In the far north the development can be slower and it may take more than 1 year for a bud to form a new shoot above ground (Blinowa 1998). Roots with a diameter of 1–2 mm grow without branching for several years and achieve a length of 50 cm (Kull 1987a). Root hairs are usually present but seldom abundant (Rosso 1966). On the lower surface of a leaf the mean number of stomata is 29 mm–2; on the upper surface they are mostly absent. A schematic view of the C. calceolus rhizome. Circles indicate shoot scars, triangles denote dormant buds and numbers show years. Type: Orchidaceous. During the first growing season, seedlings appear to be wholly dependent on the symbiotic fungus. The extent of mycotrophy decreases as plants produce more and larger leaves. The rhizome is uninfected but roots become re-infected from the soil. Therefore isolates from mature plants could differ entirely from those of the seedling fungus (Rasmussen 1995). The percentage of cells infected by hyphae increased from 12.4% to 24.4% when soil pH decreased from 6.0 to 5.1 (Sizova & Vahramejeva 1983). As in other orchids, the surface area of Cypripedium roots, which is in contact with the soil, appears to be smaller than in many other herbs. It is not clear how the orchid compensates for this, unless fungal connections are more persistent and prevalent than they appear to be (Stoutamire 1990). Cypripedium calceolus is a geophyte. The rhizome with buds perennates through the winter. Vegetative reproduction occurs by rhizome ramification. Mean temporal distance between two branchings (on a rhizome branch) is 4.9 years. Webster (1886) suggested that the species reproduces vegetatively because its pollination system is too complicated. Indeed, in many populations vegetative reproduction plays the main role. The ratio of sexual reproduction to clonal growth in recruitment in such populations is 1 : 200. However, in populations with vigorous sexual reproduction, where juveniles make up nearly half of the total ramet number, vegetative reproduction is less important. Cypripedium calceolus is a long-lived species. Many plants are more than 30 years old, and some are more than 100 years old (Kull 1988). The vegetative period before flowering lasts at least 6–10 years (Fast 1985; Rasmussen 1995). A life cycle scheme is depicted in Fig. 5. Life cycle scheme of C. calceolus, based on ramet changes. Numerical values reveal the proportion of changes occurring per ramet per year (with SD in parentheses). Open-ended arrows denote death. ADULT ramet produces annually a larger bud (B) that may die before the following year, or produce another ramet (exceptionally, may stay dormant), and a smaller bud (b) that may die, produce a ramet or remain a bud. Juveniles (JUV, defined as ramets with one or two small leaves and a thin stalk) may also produce a smaller bud (b), but this does not develop further. Larger buds (B) in juveniles may produce a juvenile or an adult ramet the following year. They may also remain dormant, but without exact data we can only estimate this parameter. A SEED may germinate, die or remain as a seed (values of the two last parameters are also estimations). In adult ramets no distinction has been made between vegetative (V) and generative (G) ramets; the number of seeds is calculated per generative ramet. 2n = 20,22 (Delforge 1995; Cribb 1997). No information. Glandular hairs of the American relatives of C. calceolus have been noted to cause skin irritation attributed to the action of the allergen cypripedin. The irritant action of the plant on the skin increased with the development of the plant, and a maximum effect was reached during seed capsule formation (MacDougal 1895). Cypripedin (2,8-dimethoxy-7-hydroxy-1,4-phenanthrene quinone), possibly derived from the phytoalexin orchinol, and obtained from the leaves of C. calceolus, does not appear to have anti-fungal activity (Schmalle & Hausen 1979; Arditti 1992). The flowers contain the anthocyanin pigment chrysanthemin, cyanidin-3-monoglucoside (Arditti 1992). Estonian populations of C. calceolus are characterized by high levels of heterozygosity (0.40–0.53) in three loci of aspartate aminotransferase. Geographically closer populations do not show a higher genetic similarity than more distant populations. Equally high levels of genetic variability in isolated populations of a species of restricted distribution can be explained by wider occurrence in previous times (Kull 1988; Kull & Paaver 1997). Dependent on the latitude and altitude, growth in spring begins in March or April, and in plants close to the northern limit, as late as June (Vorobjeva & Moskvitcheva 1987). Flower-producing shoots appear earlier than vegetative shoots. Flowering begins in May or June, and in some places in the far north even in July (Eberle 1973; Vorobjeva & Moskvitcheva 1987). Duration of flowering is 2–3 weeks (Eberle 1973). A single flower may persist 11–17 days, but withers on the sixth day after the pollen has reached the stigma (Savina 1964). The sum of positive (above +5 °C) daily temperatures needed for flowering in north-east Europe (latitude 58°26′; longitude 26 °43′) is 178 day-degrees as a mean of 4 years (Kull 1987b). Shoots start to turn yellow at the end of August. Fruits ripen in September. By this time non-fruiting shoots have withered but others remain in a dried state until the following summer. Capsules split open and liberate seeds in September–October. A new rhizome segment starts to grow at the end of flowering. Above-ground parts of juveniles persist for considerably shorter periods than adult ramets. They appear shortly before flowering of adult ramets and wither before others. Trap flowers of C. calceolus lack nectar, and pollinators are deceived as no reward is available. Crimson spotted floral structures are false nectar guides that are important for attracting and trapping bees. Floral fragrance is dominated by acetates. Chemical correspondence with pollinators’ pheromone secretions suggests that fragrance might interfere with pheromone-controlled alighting reactions (Nilsson 1979). The pollen grains are in sulcate monads, more or less smooth-surfaced, and covered by a layer not resistant to acetolysis (Burns-Balogh & Hesse 1988). The mean size of a pollen grain is 17 × 28 µm. Pollinating bees enter the opening in front of the labellum and are prevented from leaving through the same opening by the downward projecting rim of the entrance. The escape route through the narrow base of the labellum forces the insect to crawl below and touch the stigma. Escape is possible through openings on either side of the column, and to pass one of these the insect has to force its way under one of the anthers carrying along with it some pollen mass. Dimensions of the labellum entrance, the escape route under the stigma, and the final opening under the anther are selective as to visitor size (Stoutamire 1967). Medium-sized female solitary bees of the genera Andrena, Lasioglossum and Halictus are the most frequent pollinators as far as Sweden is concerned (Nilsson 1979). In Czechoslovakia mainly large Andrena spp., e.g. A. tibialis (Kirby) and A. nigroaenea (Kirby), pollinate C. calceolus (Daumann 1968). In England A. scotica Perkins and A. haemorrhoa Fabricius have been recorded as pollinators (P. Corkhill, personal communication). Apis mellifera L. too can act as a pollinator. The flower age at pollination has an effect on seed characteristics; the proportion of seeds with full embryos increases with increasing flower age at pollination (Light & MacConaill 1998). Fertilization occurs about 20 days after pollination and embryogenesis lasts up to 50 days (Wagner & Hansel 1994). Seeds of C. calceolus are some of the biggest among temperate orchids (about 1.2 × 0.3 mm) (Fig. 6a). The seed has a firm brown fusiform testa, the outer surface of which is not easily wettable. The cells of a spherical embryo, length of 0.2 mm (Eberle 1973), are isodiametric and little differentiated (Cribb 1997). The shape, size and relatively large air volume in the testa of seeds provide an explanation for their buoyancy in the atmosphere, which is an adaptation for dispersal by wind (Arditti et al. 1979). Seed and seedling morphology of Cypripedium calceolus. (a) Seed; (b) two-leaf seedling; p, bud for the next year's shoot. Cypripedium × ventricosum Sw., a natural hybrid of C. calceolus and C. macranthos, was first described from cultivated conditions in Switzerland in 1908. Later it was reported from Siberia, Manchuria and Sakhalin and the Japanese island of Rebun (Perner 1996). Measurements of flowers of C. calceolus, C. macranthos and their hybrid are given in Table 3. The size of a pollen load was not significantly correlated with seed numbers of capsules in an American relative of C. calceolus (Proctor & Harder 1994). The overall mean percentage of fruiting observed in eight Estonian populations over 11 years was 10.5% (Kull 1998). The number of seeds in five capsules varied between 5940 and 16 700 (Kull 1997). Fruit set in the other parts of the range is: Moscow district 4–14% (Varlygina & Matsenko 1986), Byelorussia 33–57% (Stavrovskaja 1984) and Sweden up to 25% (Nilsson 1979). Fruit set is pollinator limited. However, fruit set per flower is not higher in large populations of C. calceolus (Kull 1998). The dust-like seeds are wind-dispersed. However, the valves of the capsule are reported to separate in wet weather and release the seeds in raindrops (Böckel 1972). Cypripedium species often retain some quantity of seeds during the dormant period and release them slowly (Stoutamire 1974). Cypripedium calceolus germinates presumably in spring (Fuchs & Ziegenspeck 1926; Rasmussen 1995). Delayed germination would not be expected in nature, even though it has been observed in laboratory conditions (Stoutamire 1974; Ramsay & Stewart 1998). High phosphorus content of the media suppresses seed germination (Sheviak 1983). With advancing seed maturity, germination in in vitro conditions decreases. Seeds collected 40 days after pollination germinate most successfully (Wagner & Hansel 1994). The protocorm develops during the first growing season. During the second growing season it produces two roots and a scale leaf enclosing the stem tip. During the third summer, the elongating stem produces a large apical bud, while on the lower internodes one or more additional roots begins to develop in the third autumn. An aerial leaf with a very long sheath emerges in the fourth spring (Fuchs & Ziegenspeck 1926; Curtis 1943). Irmisch (1853) and Böckel (1972) have registered the first green leaf by 1 year after germination. In the following years the elongating and enlarging rhizome produces increasingly larger leaves annually. The rhizome is orthotropic in the first years, and plagiotropic later. A seedling 5 or 6 years old from a natural population is depicted in Fig. 6b. The average time from seed germination to flowering is 12 years in C. parviflorum in America (Curtis 1943) and the time could be similar in C. calceolus. No strictly species–specific associations are known to exist between C. calceolus and other organisms. However, a number of species have been recorded being in contact with C. calceolus. The trap flower serves as a prison for many other insects besides pollinators. They can be classified as follows: (i) obligatory pollinators (bees); (ii) facultative pollinators (small Coleoptera); (iii) Diptera species that may accidentally fly in and get trapped (mainly species from the families Culicidae, Empididae, Syrphidae); (iv) bigger beetles and bumblebees that destroy flowers while visiting them; (v) caterpillars that destroy the perianth. At the end of flowering, 10–30% of flowers contain dead insects (Masing 1980). High concentrations of insects are taken advantage of by crab spiders [Misumena vatia (Cl.)] and sometimes by other spiders which make nets across the main labellum entrance. Voles may bite through shoots at ground level, which may be a serious destruction in small populations of C. calceolus, for example in the single British locality in 1991. Sheep have also eaten off shoots and flowers of the English plant (Lindop 1992). No information. Unassigned potyvirus has been reported in C. calceolus from Germany (Brunt et al. 1997). Fossil records are lacking. Linnaeus described the species in Species Plantarum in 1753, but his first records of this species date back to 1737. However, the plant was certainly known before, as pieces of art depicting this plant originate from the 16th century (Heinrich & Lorenz 1996; Cribb 1997). The first British record was made in 1640 by Parkinson (1640) in his Theatrum Botanicum, from the Ingleton area near the Lancashire–Yorkshire border (Summerhayes 1951). Excessive and rapacious collection of the plant from the mid-18th to the end of the 19th century reduced it so much that it was believed extinct in Britain from the late 1910s to 1930. Since then it has had a precarious existence but is now actively conserved (see Section XI). Cypripedium calceolus belongs to the group of flagship species of nature protection and is legally protected in all European countries and Russia. The species is also protected at the supranational level by the Convention on the Conservation of European Wildlife and Natural Habitats (Bern Convention), European Union Directive CE/92/43 on the Conservation of Natural Habitats and of Wild Fauna and Flora, and the Convention on International Trade in Endangered Species (Washington Convention) (Terschuren 1998). Nowadays the largest local populations in Europe seem to be preserved in the West-Estonian Archipelago Biosphere Reserve and in some parts of Poland. In Britain the species has been a rare plant for at least the last 100 years and the main reason for decline has been the rapacity of collectors, who have dug the plants up and transplanted them to gardens (Summerhayes 1951). Besides the cutting of flowers and digging up of rhizomes, habitat destruction with logging and habitat alteration with overgrowing grassland by dense brush or closing up of the tree canopy are mentioned as main threats through the distribution range of C. calceolus (Terschuren 1998). In Britain the species has been saved from near extinction through the combined efforts of in situ conservation and propagation from seed in the laboratory (Ramsay & Stewart 1998). Besides the permanent guarding of the surviving plant, the Cypripedium Committee has aimed to re-introduce the plant to some of its former sites, from plantlet production at Kew. This work is now co-ordinated by English Nature's Species Recovery Programme (L. Farrell & P. Corkhill, personal communication). I am grateful to Professor Arthur J. Willis, Peter Corkhill, Miss Lynne Farrell and the Associate Editors (Dr A. J. Davy, Dr A.J.C. Malloch and Dr M.C.F. Proctor) for their valuable comments on the manuscript, and to Mrs J. M. Croft for preparing the distribution map.

Galium aparine L.

Galium L., Section KOLGYDA Dumort. (Section Aparine Koch). An annual herb with prostrate or, more usually, scrambling-ascending diffusely branched stems 15–300 cm, stout and hairy at the nodes, the four angles very rough with downwardly directed prickles. Leaves 12–60 × 3–8 mm, in whorls of 6–9, narrowly to widely oblanceolate or elliptical, mucronate, 1-veined, glabrous or with hooked bristles above, margins rough with backward-pointing prickles. Flowers 1.5 mm diameter in 2–5-flowered axillary cymes, their peduncles topped by a whorl of 4–8 bracts; hermaphrodite, sometimes male (andromonoecious). Calyx a minute annular ridge; corolla 1.5–1.7 mm in diameter, whitish, with four acute lobes. Stamens 4, exserted; styles 2, short, connate below; stigmas capitate. Two-seeded fruits with length of longest axis 3–5 mm (excluding setae), wider than the corolla, olive or purplish, covered with white hooked setae with tuberculate bases, their stalks divaricate. Mature fruit a dry mericarp with length of longest axis 2–3 mm, and a mean air-dry mass of 9.65 mg ± 0.33 (10 samples of 100 from a hedgerow population). Salisbury (1974) reports the average mass of seed from shade-tolerant woodland and hedgerow plants as 3.7 mg, and Grime et al. (1988) record a mean seed mass of 7.25 mg for a hedgerow population. Galium aparine is a highly variable species owing to a combination of phenotypic plasticity and genecological variation. The species has different responses to the environment corresponding to different environmentally conditioned life cycles [see VI (e)]; plasticity in the timing of seed germination, productivity, growth form, freezing tolerance, and the light and temperature dependence of photosynthesis enable G. aparine to establish itself in a wide variety of habitats. Potential variability concealed in the higher polyploid conditions, especially in the widespread hexaploids [see VI (d)], enabling them to respond to changing environmental conditions, is probably also the basis of their success. Among plants of G. aparine raised from seed collected from various sites in continental Europe, Böcher et al. (1955) distinguished at least five different climatic ecotypes or races that either may be discrete or more probably ranges of a large clinal variation: mediterranean, summer-annual (seed germinates in the spring) with accelerated development from Portugal; mediterranean, biennial from Spain; southern, winter- to summer-annual from France; northern, rather late flowering winter- to summer-annual from Denmark and Poland; and northern, very late flowering winter-annual (seed germinates in autumn) from Denmark and Sweden. Ecotypes of G. aparine have been demonstrated by Berkefeld (1988), differing in shape of cotyledons, hairiness of stems, colour of stem edges and length of fruit spines in plants grown under uniform garden conditions from seed collected in several different habitats. In Britain, two discrete ecotypes of G. aparine exist, one found in arable fields or open situations and the other growing in hedgerows ( Froud-Williams 1985); the two ecotypes are very similar morphologically but differ markedly in seed dormancy and germination requirements [see VIII (d)]. Bain & Attridge (1988) provide evidence of genotypic differences in shade strategies between hedgerow and field populations [see VI (e)]. Two forms have been recognized, the rarer G. aparine f. intermedium Bonnet, with smooth or tuberculate fruits without spiny hairs, and the common form G. aparine f. aparine with spiny fruits ( Moore 1975). Variants of G. spurium with setose fruits, apart from the flower characters and the diploid chromosome number, are often not easy to distinguish from G. aparine and have been described as G. spurium ssp. infestum ( Hanf 1983). Galium aparine is native on maritime shingle beaches, in primary tall-herb fen communities, and in primary alderwood in natural hydrarch successions, but is also a very variable cosmopolitan weed found in hedges, waste places, drained fen peat, on limestone scree, roadside verges, river and stream banks, and is a major arable weed, especially of cereal crops. Galium aparine is abundant throughout the British Isles; it occurs in all vice-counties ( Fig. 1). It is widespread throughout Europe ( Fig. 2), northwards in Norway to 70 °N, but otherwise absent from the region within the Arctic Circle, and southwards to the Mediterranean, west to the Azores and east to the Ural mountains. The distribution of Galium aparine in the British Isles. Each dot represents at least one record in a 10-km square of the National Grid. Mapped by Mrs J.M. Croft, Biological Records Centre, Institute of Terrestrial Ecology, mainly from records collected by members of the Botanical Society of the British Isles. (○) pre-1950: (●) 1950 onwards. The main distribution of Galium aparine in Europe and western Asia (modified from Vergl. Chor.). The base map is reproduced by permission of the Committee for the Mapping of the Flora of Europe (the bold dashed line shows the limits of Europe). Outside Europe G. aparine occurs eastwards to western Siberia, southern Siberia, western Asia and Turkestan, and southwards to the north-west African coast from Morocco to Tunisia. It is found between 30 °N and 48 °N in the eastern United States and along the western coast of North America from California northwards to southern Alaska. The altitudinal range of G. aparine in the British Isles extends from sea level to 366 m in Yorkshire (Alt. range Br. Pl.) to an upper limit of approximately 430 m at Nenthead, Cumbria ( Halliday 1997). Galium aparine is more abundant in lowland habitats, decreasing progressively at higher altitudes in the British Isles. In Europe it has a wide climatically determined range, from boreal oceanic–suboceanic conditions in the north through the temperate zone to a mediterranean climate in the south (Vergl. Chor.). The native distribution of G. aparine is given as European temperate but now widely naturalized and circumpolar temperate by Preston & Hill (1997). It occupies a niche with a seasonally contrasting light and temperature climate. In common with Urtica dioica, a species with which it is closely associated, G. aparine thrives on soils rich in available nitrogen ( Ellenberg 1988). However, it has been demonstrated by Pigott & Taylor (1964) that the supply of phosphate in combination with nitrogen is a major factor limiting the distribution of G. aparine; in response to additions of fertilizer nitrogen and phosphorus to pots of soil collected from under dominant Mercurialis perennis in Buff Wood, East Hatley, Cambridgeshire, UK, where G. aparine rarely occurs, the mean dry weight of seedlings of G. aparine after 55 days was 257 ± 14 mg with stunted leaves and no flowers without added fertilizer, 330 ± 17 mg with added nitrogen, 264 ± 25 mg with added phosphorus, and 587 ± 86 mg with vigorous growth and both flowers and fruits developed when nitrogen and phosphorus were supplied together. Galium aparine occurs widely in a range of habitats in association with mildly acid to ± base-rich damp to fairly wet soils. Grime et al. (1988) record that G. aparine is most frequent and abundant in soils between pH 5.5 and 8.0. It is abundant in scrub and occasional in hedges on chalk soils of pH 7.8 in Wiltshire (Fl. Wilt.). Galium aparine has been used successfully as a bioassay plant in a laboratory experiment with chalk grassland soils by R.A. Reed and P.J. Grubb (unpublished data). They obtained evidence of a succession of primary limiting nutrients, first phosphate when the freely diffusible nitrate was taken up from the total available soil volume more rapidly than the immobile phosphate, then nitrate as root competition became effective for the mobile nitrate but not for the immobile phosphate (cited by Bonis et al. 1997)****[see also VI (e) for supporting evidence]. Galium aparine shows a high degree of constancy in two types of shingle beach vegetation found on the north and west coasts of Scotland ( Gimingham 1964): in Atriplex glabriuscula–Silene maritima (uniflora)–Tripleurospermum maritimum stands occurring on fringing beach composed of pebbles set in a sandy matrix with an admixture of algal debris, where shingle movement is confined to winter; and on apposition banks to landward of the foreshore in Arrhenatherum elatius stands, especially on the Isle of Arran. Galium aparine is also a component of shingle vegetation at Dungeness and in Wales ( Scott 1963). Both U. dioica and G. aparine are very abundant in fen carr in the region of the Broads in east Norfolk and in the valley fenwoods of Breckland (Tansley, Br. Isl.). Galium aparine was one of 20 species recorded most frequently in 121 Wiltshire hedges (Fl. Wilt.). From detailed botanical data on the composition of hedgerows collected during countryside surveys in Britain in 1978 and 1990 ( Barr et al. 1993 ), the cover of G. aparine in paired plots increased highly significantly (P < 0.001) between 1978 and 1990, with a mean difference between percentage cover in the 2 years of 5 ± 1.1 in a sample of 172 plots ( Bunce et al. 1994 ). The following phytosociological account is based on the National Vegetational Classification ( Rodwell 1991, 1992, 1995, 1999). The joint constants of the Matricaria maritima (Tripleurospermum maritimum)–Galium aparine strandline community (SD3), together with Atriplex glabriuscula and Stellaria media, are the most prominent and frequent elements in the generally open and often patchy strandline vegetation confined to Scotland and along sheltered shores in north-west England. Apart from the two constants in Phragmites australis–Urtica dioica tall-herb fen (S26), G. aparine is the only species that is frequent throughout; locally in the Filipendula ulmaria subcommunity, G. aparine is prominent together with the other constants Phragmites or U. dioica or F. ulmaria; the A. elatius subcommunity dominated by Phragmites has varying amounts of U. dioica and G. aparine and, beneath, tussocks of A. elatius, with scattered plants of Cirsium arvense and Heracleum sphondylium very characteristic of this vegetation. The community occurs throughout the lowlands but especially in Broadland, east Norfolk, the Fens and around the Shropshire and Cheshire meres; the Filipendula subcommunity, however, has only a scattered distribution and the Arrhenatherum subcommunity is most frequent in south-west England and south Wales. Galium aparine is a common species in three woodland and scrub communities: Alnus glutinosa–Urtica dioica woodland (W6), where it is a joint constant in the typical subcommunity, widespread but local throughout the lowlands, occurring where active alluvial deposition is taking place on more mature rivers and on the remnants of undrained flood plains and eutrophicated mires; Crataegus monogyna–Hedera helix scrub (W21), where it is very common in the Hedera helix–Urtica dioica subcommunity in which Crataegus monogyna, Rubus fruticosus agg., Hedera helix and U. dioica are constant species, widely distributed through the British lowlands, and also in the floristically richer Mercurialis perennis subcommunity, found on heavy-textured base-rich soils especially in areas with clays and shales; and the Rubus fruticosus–Holcus lanatus underscrub (W24), where it is a constant in the Arrhenatherum elatius–Heracleum sphondylium subcommunity that is consistently enriched by a very distinctive group of preferentials, A. elatius, Festuca rubra, U. dioica, H. sphondylium and Taraxacum officinale agg., which are constants, and frequently Anthriscus sylvestris and Chaerophyllum temulentum (temulum), especially characteristic of this kind of vegetation which is ubiquitous on suitable soils throughout the British lowlands. Galium aparine, together with Oxalis pes-caprae and Sonchus oleraceus, is preferential in the Cerastium glomeratum–Fumaria muralis ssp. boreai community (OV6) Vicia hirsuta–Papaver dubium subcommunity, confined to regularly disturbed fertile, light, non-calcareous soils cultivated for bulb and vegetable production in the extreme oceanic climate of the far south-west of Britain, in Cornwall and The Scillies; it is also preferential in the Stellaria media–Capsella bursa-pastoris community (OV13) in the Urtica dioica–Galium aparine subcommunity, which occurs widely on fertile loamy soils, as weed vegetation among root vegetable, salad and cereal crops, throughout the British lowlands; and is a joint constant in species-poor tall-herb vegetation dominated by densely abundant U. dioica, described as a Urtica dioica–Galium aparine community (OV24), in nutrient-rich, moist but well-aerated, disturbed soils throughout the lowlands. Galium aparine occurs frequently in two maritime communities: Hippophae rhamnoides scrub (SD18), in the Urtica dioica–Arrhenatherum elatius subcommunity, on dunes on the east coast between Kent and Fife, especially in Lincolnshire and north Norfolk; and Brassica oleracea maritime cliff-edge community (MC4) Beta vulgaris ssp. maritima subcommunity, in more maritime conditions on crumbling edges and sloping ledges of south-facing cliffs in calcareous rocks on the south coast of England. Galium aparine occurs frequently in Alnus glutinosa–Fraxinus excelsior–Lysimachia nemorum woodland (W7), in the Urtica dioica subcommunity which is widely, although locally, distributed throughout the upland fringes of the north and west; in Fraxinus excelsior–Acer campestre–Mercurialis woodland (W8), in the Geranium robertianum and Allium ursinum subcommunities, both especially characteristic of the Yorkshire and Derbyshire Dales and the Welsh Marches; in Prunus spinosa–Rubus fruticosus scrub (W22), which is of widespread distribution through the British lowlands on mesotrophic mull soils of moderate base status; and is also frequent in Pteridium aquilinum–Rubus fruticosus underscrub (W25) in the Hyacinthoides nonscripta subcommunity commonly widespread on suitable soils throughout lowland Britain. In and around roadside and field ditches throughout lowland Britain, G. aparine frequently occurs in Arrhenatherum elatius grassland (MG1) Filipendula ulmaria subcommunity; it is occasional to scarce in Arrhenatherum elatius–Filipendula ulmaria tall-herb grassland (MG2), confined to scattered localities in northern England on the Carboniferous limestone of Derbyshire and Craven; and scarce in Festuca rubra–Agrostis stolonifera–Potentilla anserina grassland (MG11), a lowland vegetation type frequent near sea level, where it occurs as extensive stands in the flood-plains of major rivers and on the upper salt-marsh. Galium aparine occurs frequently in the Papaver rhoeas–Silene noctiflora community (OV16), in light to well-drained calcareous soils, mostly among cereals, across the warmer and drier south-east of England from Dorset and Wiltshire, north-east to Lincolnshire; and in the Epilobium hirsutum community (OV26) Filipendula ulmaria–Angelica sylvestris subcommunity, common in ditches along roadsides, widespread and common throughout the British lowlands. Galium aparine is occasional to scarce in a number of woodland communities: Salix cinerea–Betula pubescens–Phragmites australis woodland (W2) Alnus glutinosa–Filipendula ulmaria subcommunity largely confined to East Anglia and some of the Cheshire and Shropshire meres; Alnus glutinosa–Carex paniculata woodland (W5), especially in the Phragmites australis subcommunity widely distributed in the Shropshire and Cheshire meres; in Fraxinus excelsior–Sorbus aucuparia–Mercurialis perennis woodland (W8), distributed in the cooler and wetter north-western parts of Britain; and Fagus sylvatica–Mercurialis perennis woodland (W12), especially in the Sanicula europaea subcommunity characteristic of the scarps of the North and South Downs, the Chilterns and the western end of the Cotswolds. Galium aparine is occasional in several open communities: Papaver rhoeas–Viola arvensis community (OV3), annual weed vegetation widespread in the southern part of Britain; Veronica persica–Veronica polita community (OV7), characteristically a weed community of cereals and other annual field crops on lighter, well-drained soils in the lowlands of south-eastern and eastern Britain; Veronica persica–Alopecurus myosuroides community (OV8), a similar weed community but on wetter heavy-textured soils, around the Fenland fringes and scattered west to Dorset; Anagallis arvensis–Veronica persica community (OV15), an annual weed community, usually dominated by ephemerals, found in cereal crops on base-rich soils in south-east England; and Parietaria diffusa community (OV41), which occurs in crevices and on small ledges in walls and on rock faces in quarries and natural cliffs in the lowlands of southern England. Galium aparine is scarce in three maritime communities: on the foreshore in Leymus arenarius mobile dune community (SD5); in Elymus pycnanthus (Elytrigia atherica) salt-marsh (SM24); and in Elymus (Elytrigia) repens salt-marsh (SM28). Galium aparine is scarce in a number of swamp communities: Carex paniculata sedge-swamp (S3); Phragmites australis swamp and reed-beds (S4), Phragmites australis subcommunity and Atriplex prostrata subcommunity, with a patchy distribution on coastal and inland salt-marshes; Glyceria maxima swamp (S5); Carex riparia swamp (S6); Carex acutiformis swamp (S7); Carex pseudocyperus swamp (S17); Carex otrubae swamp (S18); Scirpus (Bolboschoenus) maritimus swamp (S21), Atriplex prostrata subcommunity and Agrostis stolonifera subcommunity. Galium aparine is also scarce in two tall-herb fen communities: Phragmites australis–Eupatorium cannabinum tall-herb fen (S25); and Phalaris arundinacea tall-herb fen (S28), Phalaris arundinacea subcommunity and Epilobium hirsutum–Urtica dioica subcommunity. Galium aparine is scarce in several woodland communities: Salix cinerea–Galium palustre woodland (W1); Salix pentandra–Carex rostrata woodland (W3); Quercus robur–Pteridium aquilinum–Rubus fruticosus woodland (W10); and Quercus petraea–Betula pubescens–Oxalis acetosella woodland (W11). Galium aparine is also scarce in a number of weed communities: Poa annua–Senecio vulgaris community (OV10); Polygonum aviculare–Chamomilla suaveolens (Matricaria discoidea) community (OV18); Lolium perenne–Dactylis glomerata community (OV23); and Epilobium (Chamerion) angustifolium community (OV27). In continental Europe G. aparine is described as having accompanied cultivation since the Neolithic, often with U. dioica, and is a character species of the class Artemisietea (Pfl. Exk.), especially in subclass Galio-Urticenea in the order Convolvuletalia and order Glechometalia ( Oberdorfer 1983), and in the persistent nitrophilous ruderal vegetation of woodland edges in the alliance Geo-Alliarion referred by Ellenberg (1988) to the order Calystegio-Alliarietalia (Convolvuletalia); it also occurs with high frequency in the subclass Artemisienea vulgaris in the order Artemisietalia ( Oberdorfer 1983). Galium aparine occurs with high frequency in arable-weed communities on damp soils in the class Secalinetea (Pfl. Exk.; Oberdorfer 1983); and with high constancy in lowland flood plain woodlands in central Europe dominated by willow, poplar, elm or oak in the alliance Salicion albae (Pfl. Exk.; Ellenberg 1988). In the submediterranean-central European region, G. aparine is a differential species in the important scrub and hedgerow association Prunus spinosa–Carpinus betulus, referred to the alliance Prunion spinosae. It also occurs with high constancy in the Galio-Carpinetum association found on soils influenced by ground water in oak–hornbeam woods ( Ellenberg 1988). In the humid mountainous regions of mediterranean France, G. aparine is frequent in the woodland association Alneto-Fraxinetum oxycarpae referred to the alliance Alneto-Ulmion, which forms thickets along river banks where it is subjected to periodic inundations ( Braun-Blanquet 1951). In the Netherlands, Westhoff & den Held (1969) record G. aparine in a number of communities: in Oxalido-Chenopodietum polyspermi (Chenopodietea) ruderal communities in gardens, nurseries and in moist to wet riverbank situations; common in the alliance Galio-Alliarion (Artemisietea); a member of the characteristic species combination of the Angelicion litoralis ruderal community around estuaries; a member of the characteristic species combination of the Epilobium hirsutum society; in Salicetum triandro-viminalis, a differential of subassociation rubestosum, and in Salicetum albo-fragilis, a differential of subassociation cardaminetosum, both in the class Salicetea purpureae; and also in the Alno-Padion of Quercetea robori-petraeae woodlands. Whilst G. aparine is an annual dicotyledonous weed of seemingly ubiquitous occurrence in hedgerows, it has increasingly become of widespread importance as a weed of arable crops. In a survey of arable dicotyledonous weeds, G. aparine was the second most frequent broad-leaved weed present at harvest, while it was the most frequent weed of winter wheat, occurring in 12% of crops surveyed ( Chancellor & Froud-Williams 1984), and was present in 57% of 450 fields of winter oilseed rape surveyed just prior to the summer harvest in nine areas of central southern England ( Froud-Williams & Chancellor 1987). Despite the apparent association of G. aparine with early sown winter cereals established by minimal cultivation, there is no evidence to suggest that weed problems are exacerbated by direct-drilling per se, although shallow cultivation may increase germination and provide more suitable conditions for establishment; as many as 44% seedlings of G. aparine may develop following tine cultivation compared with direct-drilling (29%), while substantially fewer seedlings emerge after ploughing. Formerly, straw-burning could provide an effective means of control, destroying more than 90% of seeds on the soil surface ( Froud-Williams 1985). Ripe seeds of G. aparine collected from plants growing in a field margin were sown in soil at various depths in wooden boxes in an outdoor experiment ( Froud-Williams et al. 1984 ); emergence was virtually confined to the period October–January; compared with surface-sown seed, burial over a range between 5 and 100 mm depth increased germination and emergence of the large-seeded G. aparine. The effects of G. aparine on crop production have been reviewed by Froud-Williams (1985), namely the competitive ability of G. aparine, interference with harvesting and crop contamination; interspecific competition resulted in yield losses from winter wheat of 30%, 47% and 52% at experimental densities of 25, 100 and 520 G. aparine plants m–2, and yield losses from spring barley of 19%, 42%, 55%, 70% and 80% at densities of 35, 132, 309, 619 and 1190 G. aparine plants m–2; the effect of lodging of cereals by G. aparine can reduce yields by 30–60%; up to 21% of samples of cereal seed can be contaminated by seed of G. aparine. A replicated randomized glasshouse experiment was carried out by Lintell-Smith et al. (1992) to determine the hierarchy of competitive ability of spring wheat (Triticum aestivum cv. Tonic), Anisantha (Bromus) sterilis, G. aparine and Papaver rhoeas at four nitrogen levels at rates equivalent to 0, 75, 150 and 300 kg N ha–1 in the form of a slow release fertilizer. For each of the 16 treatments (four species × four nitrogen levels) a target plant of wheat was set in the centre of each 11.5-cm diameter pot filled with John Innes No. 1 compost, and surrounded by three individuals of either wheat or of the individual weed species. The competitive hierarchy of the four species in suppressing the total ear weight of the target wheat plant at the highest nitrogen level fell in the order wheat > A. sterilis > G. aparine > P. rhoeas. Irrespective of the species of neighbouring plant, the total ear weight of the target wheat plant increased with the level of soil nitrogen. Consideration of the relative competitive abilities of the species in relation to the strongest competitor (wheat) indicated that, as the level of nitrogen increased, the weeds became more competitive in relation to wheat. Biomass, plant size, plant density and the inequality of sizes were assessed for autumn-emerging roadside populations dominated by G. aparine during early stages of growth by Puntieri & Hall (1996). One study showed that the slope of the log–log size–density relationship for all plant species present in samples was closer to –1.5 (as predicted by the 3/2 power law of self-thinning) and that for G. aparine was closer to –1.0 in five separate populations. The size inequality of G. aparine tended to increase or to remain constant during periods of high mortality, and in early harvests it was negatively related to population density. In a second study there were simultaneous decreases of both biomass and density of G. aparine and all plant species during a period of a month soon after emergence, and a higher size inequality of G. aparine in those patches where plant density (and that of G. aparine) was lower. Labelling of seedlings of G. aparine, belonging to different cohorts of emergence, indicated density-dependent mortality and a higher probability of survival for seedlings emerging very early. Galium aparine seems to have a weak size hierarchy development and limited individual growth at high population densities. Galium aparine is effectively controlled with DNOC and MCPP herbicides ( Salisbury 1964). In cereal crops it is susceptible at the early seedling stage to the use of mecoprop salt, dichloroprop salt, dinoseb, benazolin, bentazone and MCPA mixtures (Weed Handbk.). The pre- and early post-emergence herbicidal activity of diflufenican (N-2,4-difluorophenyl-2–3-trifluoromethylphenoxy-3-pyridinecarboxamide), a novel herbicide, is reported by Cramp et al. (1987) and attention is drawn to its ability to control important weeds in winter cereals, including G. aparine, which is resistant to substituted-urea herbicides. The results of a series of 11 field experiments over two seasons ( Lutman et al. 1988 ) indicated that the control of G. aparine in winter cereals was most effective when mecoprop and a proprietary mixture of ioxynil + bromoxynil was applied between December and March (up to 90%). But when mecoprop was applied alone in April, especially when soil temperatures were > 6 °C, control was increased to 94%. Galium aparine is locally capable of developing dense populations consisting of a large number of discrete individual seedlings; it is commonly present in ephemeral assemblages of enriched disturbed land and in some woodland, scrub and tall-herb communities (see III), where it can keep pace in height growth with established perennials. Seasonal changes in the above-ground biomass of G. aparine have been determined in various stands of vegetation by Al-Mufti et al. (1977) . The mean maximum biomass values, which occurred in the month of June, in 0.25-m2 samples, as g m–2 ± standard error (number of replicates), were: 56 ± 25 (5) within a stand of U. dioica on alluvial soil in Lathkilldale, north Derbyshire; 29 ± 7.5 (4) within a stand of Chamerion (Chamaenerion) angustifolium on a brown earth in Lathkilldale; 5 ± 1.8 (5) in a woodland clearing under U. dioica on alluvium in Totley Wood, Sheffield, South Yorkshire; and 20 ± 5 (10) under the tree canopy within a stand of U. dioica on alluvium in Totley Wood. The mean aerial biomass of G. aparine sampled in late July, in nine random 1-m2 samples of roadside Arrhenatheretum in south-eastern Cambridgeshire, was 2.4 ± 1.2 g m–2 ( Grubb 1982). Plasticity of growth and of biomass allocation was investigated by Auge & Mahn (1988) in three populations of G. aparine in spring-germinating plants in their native habitats, an arable field south-west of Halle, in Fraxino-Ulmetum tilietosum forest, and in Populus×canadensis forest, both south of Leipzig, Germany. Ontogenetic drift of relative growth rate, biomass allocation as well as biomass and seed production were closely related to the light factor, differing between the three ecosystems. Allocation between vegetative and reproductive growth was characterized by a complete switch at the start of the reproductive phase in the most light-limited forest population and by a large overlap under more favourable conditions in the arable field population. Biomass allocation among the vegetative organs was strongly dependent on total biomass of the individuals. Multivariate comparisons revealed that the entire allocation patterns of the forest populations were rather similar at the date of peak vegetative biomass, but differed substantially from the arable population. Some of the observed differences between populations seemed to be due to phenotypic plasticity. The seasonal course of frost tolerance in potted autumn-emergent seedlings of G. aparine has been measured by Kutsch & Kappen (1990); frost tolerance (the temperature that would result in a 20% mean death of shoots) from mid-October to mid-November was –7 °C, increasing by the end of November to –10 °C and to a limit of –17 °C by early December, then decreasing to –9 °C in March and –2 °C in April. The net photosynthetic response to light and temperature did not shift during the period December to March, with an optimum temperature for photosynthesis of 15 °C and a net photosynthetic rate of 8.1 mg CO2 dm–2 h–1. During this winter period freezing and undercooling temperatures fluctuated around –3 °C in the assimilatory tissues. Down to –3 °C no frost-induced decrease of net photosynthesis was observed, but below this temperature ice was formed in the leaf mesophyll and photosynthesis ceased. CO2 uptake was apparent within a few hours after thawing; below –5 °C the rates were reduced depending on the severity of the preceding frost. Photosynthetic pr

Effects of elevated temperature on multi‐species interactions: the case of Pedunculate Oak, Winter Moth and Tits

1. The effects of temperature on the Oak–Winter Moth–Tit food chain were studied at Wytham Wood, Oxford, and experimentally in the controlled environment solardomes at the Institute of Terrestrial Ecology, Bangor. 2. Tree cores from Wytham indicated that mature Oaks grew best at high temperatures and rainfall, but with low caterpillar populations. Young trees grew less well at elevated temperature, probably because they lost more water than they gained. Elevated temperatures advanced budburst, reduced foliar nitrogen and increased leaf toughness. 3. Moth eggs laid later or maintained at cooler temperatures than average required fewer heat units to hatch. Caterpillars took up to 50 days to complete growth at field temperatures but did so in only 20 days at a constant 15 °C. 4. The mass of Tit chicks at day 15 (day 1 = egg hatch) was positively correlated with temperature and negatively correlated with rainfall during the growing period. 5. At elevated temperature, budburst and moth egg hatch were synchronized, but earlier. Late feeding larvae and larvae fed on leaves from trees grown at elevated temperature produced smaller pupae. Pupal mass was unaffected when caterpillars and trees were maintained together under the same conditions. 6. Delaying egg hatch in Tits, to simulate conditions at elevated spring temperatures, resulted in reduced chick mass, body size and fledging success. This occurred because the chicks were fed later and prey quality was poorer, because the peak of caterpillar biomass was missed. 7. We predict that moth reproductive output will be retained at elevated temperatures because both leaves and caterpillars develop faster. Brood size in birds may be reduced because they cannot lay early enough to coincide with the narrower peak of food abundance.

Glechoma hederaceaL. (Nepeta glechomaBenth.,N. hederacea(L.) Trev.)

A polycarpic, aromatic, often purplish, softly hairy or nearly glabrous perennial herb. Flowering shoots 10–30 (–60) cm long, square in section and ascending or erect. Non-flowering stems also square in section, creeping and rooting at many, but not all, nodes. Leaf laminae 4–35 (–80) × 6–40 (–80) mm, reniform to ovate–cordate, obtuse or subacute and coarsely crenate. Petioles highly variable in length from 10 mm in grassland to over 200 mm in woodland with dense canopy cover and dense ground vegetation. Bracteoles 1–1.5 mm, shortly hispid or pubescent. Flowers zygomorphic and hypogynous, borne in few (2–5)-flowered secund axillary whorls, the bracts not differing from the foliage leaves. Calyx tubular or tubular-campanulate, 15-veined, not gibbous at base, and somewhat two-lipped. Upper lip of the corolla flat with three triangular acuminate teeth that are 0.2–0.3 times as long as the tube. Lower lip with two teeth. Corolla tube narrowly obconical, straight and hairy within at the base of lower lip. Corolla (6–) 15–22 (–25) mm, usually blue–violet with purple spots on the lower lip; rarely pink or white. Stamens four, two long and two short, borne on corolla tube and lying below upper lip; anther cells at right angles to each other, each opening by a separate slit. Ovary four-celled and deeply four-lobed, each cell with one ovule. Style gynobasic, its branches equal. Hermaphrodite and male-sterile forms of flowers occur, the latter usually with smaller corollas than hermaphrodite flowers, and much-reduced, pollen-less stamens that are often included in the corolla tube. Clones usually bear either only hermaphrodite or only male-sterile flowers, but in rare cases both types are found on the same individual (Price 1991). Each flower produces a maximum of four obovoid, slightly angled, smooth nutlets with a mean air-dry mass of 0.69 mg (n = 100). Each is slightly angled owing to pressure exerted by the adjacent developing nutlets before dispersal, and each bears a grey protuberance which may be an elaiosome. The dimensions of the seed are 1.9 × 1.1 mm. Glechoma serbica Halácsy & Wettst., which has been recorded in former Jugoslavia, is probably an ecological modification of G. hederacea (Fl. Eur. 3). The southern counterpart of G. hederacea in Europe is ssp. hirsuta (Waldst. & Kit.) F. Hermann, sometimes referred to as G. hirsuta (Elias 1979, 1983). This has denser and longer hairs and longer calyx teeth than G. hederacea, and the corolla is pale blue. It has been recorded in Albania, Austria, Bulgaria, the Czech Republic, Greece, Hungary, Italy, former Jugoslavia, Poland, Rumania, central and south-west Russia, Sardinia and Sicily (Fl. Eur. 3). Intermediates, which are presumably hybrids between G. hederacea and G. hirsuta, sometimes occur (Fl. Eur. 3). In eastern Asia a more or less distinct, short-pubescent race occurs that is called ssp. grandis (A. Gray) Hara [also var. grandis A. Gray and var. longituba (Nakai) Kuprian] (Hultén 1971). In North America the variety micrantha Moricand [or var. parviflora (Benth.) House] is distinguished (Hultén 1971). Glechoma hederacea is native to the British Isles and common throughout except for western Ireland and northern Scotland, where it is rare (Fig. 1). It is not native in Orkney and Shetland, and is absent from the Outer Hebrides. It occurs almost throughout Europe, except in the Balearic Islands, Crete, the Faroes, Iceland, Spitsbergen and Turkey. It is a weed in Tadzhikistan (Hultén 1971). It is also found in western and northern Asia to Japan, in Hong Kong and in Tonkin (Hultén 1971). It has been introduced in New Zealand, where it is found in many southern localities in the North Island and in many localities throughout the South Island (Webb et al. 1988). Nutlets have not been found on any New Zealand material (Webb et al. 1988). It has also been introduced in America (Atl. Br. Fl.; Fl. Br. Isl.; Fl. Eur. 3; Hultén 1971; Fitter et al. 1985; Grime et al. 1988), where its range extends southwards through the United States to Georgia, Tennessee, Kansas and Colorado (Gill 1979). It is naturalized in Canada and ranges from Newfoundland (Lat. 47°N, Long. 53°W) to British Columbia (Lat. 54°N, Long. 123°W). The distribution of Glechoma hederacea in the British Isles. Each dot represents at least one record in a 10-km square of the National Grid. Mapped by Mrs J. M. Croft, Biological Records Centre, Institute of Terrestrial Ecology, mainly from records collected by members of the Botanical Society of the British Isles; (○) pre-1950; (•) 1950 onwards; (×) garden escapes or outcasts. The altitudinal range of G. hederacea has been given as from sea level in England, Scotland and Wales to 400 m in Weardale and 415 m at Burbage, Derbyshire, UK (Alt. range Br. Pl.). Other records are at altitudes of 410 m (Grime et al. 1988) and 295 m (Fl. Br. Isl.) in the British Isles, 350 m in Sweden (Hultén 1971), and to about 1600 m in the Alps (Hultén 1971). Glechoma hederacea is a species of temperate latitudes. Ellenberg Indicator Values (Ellenberg et al. 1991) are given as 6 and 3, and Landolt Indicator Values (Landolt 1977) as 4 and 3 for temperature and continentality, respectively, indicating that G. hederacea is primarily a species of the colline-submontane zones of temperate Europe, with an oceanic to suboceanic distribution. In the British Isles its northern limit follows the July isotherm of 12 °C (Atl. Br. Fl.). In Canada its northern limit follows the 12 °C July isotherm and the −12 °C January isotherm (Gill 1979). It occurs on sites with a wide range of slopes. In the Sheffield region it was slightly more common on south-facing slopes than on other aspects, but more abundant on north-facing slopes in unshaded situations (Grime et al. 1988). It is most abundant on sites with a moderate to high percentage of bare ground. Preston & Hill (1997) classify G. hederacea as Eurasian Boreo-temperate, but Circumpolar Boreo-temperate after wide naturalization [see also Hultén (1971); map 274 for G. hederacea and for sspp. hirsuta and grandis]. (B) SUBSTRATUM Glechoma hederacea occurs mainly on damp, heavy, fertile and calcareous soils. It is most frequent in the pH range from slightly acidic to slightly alkaline (pH 5.5–7.5), but can also be found on soils with a pH as low as 4.0 (Fl. Br. Isl.; Fitter (1978); Grime et al. 1988). Ellenberg Indicator Values for substrate (Ellenberg et al. 1991) agree with these substrate affinities, being 6 for humidity, × for soil acidity (i.e. no strong preference) and 7 for mineral nutrient availability. Landolt (1977), however, records that it does not tolerate strongly acidic soils. He also reports that it is intolerant of salinity and that it is mainly found on fine-textured soils including clays. Kreutzer & Seibert (1984) and Sinker et al. (1985) state that G. hederacea becomes abundant only when there is a good phosphate and nitrate supply and a moderately rich supply of cations including Ca2+ in the soil substrate. III. Communities Glechoma hederacea is typically a plant of shaded habitats including woodlands and hedgerows. It is frequent on shaded roadsides, in clear-felled and coppiced woodlands, scrub and derelict woodland, and in orchards. Although it is described as having only a limited capacity to persist under tall herbs (Grime et al. 1988), it is frequently a persistent associate of Urtica dioica and other species in both open habitats and at woodland edges (Sinker et al. 1985). Peterken (1981) regards G. hederacea as a characteristic species of secondary woodland (Hermy & Stieperaere 1981). It sometimes becomes locally abundant in damp woodlands (Fl. Br. Isl.). It is also found in grassland habitats growing in more or less full sunlight, beside rides and rabbit scrapes, on spoil heaps, wasteland, waysides, walls and at the edge of pastures, or encroaching and even dominating enriched areas at the margins of arable fields. Although it is generally absent from aquatic habitats (Cadbury et al. 1971; Hyde 1976; Fitter 1978; Grime et al. 1988), it is occasional in soligenous mires and on moist soil at the periphery of topogenous mires, on river banks and floodplains. Typical associates of G. hederacea include Ajuga reptans, Brachythecium rutabulum, Bryonia dioica, Galium aparine, Geranium robertianum, Geum urbanum, Hedera helix, Poa trivialis, Ranunculus ficaria, Ranunculus repens, Sambucus nigra, Stachys sylvatica, Stellaria holostea, Tamus communis and Urtica dioica (Pl. Comm. Scot.; Sinker et al. (1985); Grime et al. 1988). Species that occur in similar habitats to G. hederacea are listed by Grime et al. (1988) as Stachys sylvatica (habitat similarity 85%), Galium aparine (77%), Tamus communis (74%), Urtica dioica (70%) and Stellaria holostea (63%). In a study of the spatial variation of vegetation in an ancient oak–ash–hazel–maple woodland following coppice management (Ash & Barkham 1976), G. hederacea was listed, together with Arum maculatum, Carex sylvatica, Dactylorhiza fuchsii, Heracleum sphondylium, Listera ovata, Mercurialis perennis, Orchis mascula, Paris quadrifolia and Poa pratensis, as a species of shaded habitats rather than of persistent clearings. Glechoma hederacea is a component of several woodland communities described in the National Vegetation Classification (Rodwell 1991a). In most of these it has a recorded frequency class of I (i.e. a frequency of 1–20% in vegetation samples) but it reaches class II (21–40% frequency) in some subcommunities of W6 (Alnus glutinosa–Urtica dioica) and W8 (Fraxinus excelsior–Acer campestre–Mercurialis perennis) woodlands, and in some subcommunities of W24 (Rubus fruticosus agg.–Holcus lanatus) and W25 (Pteridium aquilinum–Rubus fruticosus agg.) underscrub communities. It is recorded as common (class III, 41–60% frequency) in the Primula vulgaris–Glechoma hederacea subcommunity (W8a) of W8 (Rodwell 1991a). Glechoma hederacea occurs in the Alnus glutinosa–Filipendula ulmaria subcommunity (W2a) of the Salix cinerea–Betula pubescens–Phragmites australis woodland community (W2; Rodwell 1991a) that often develops on abandoned mowing marshes. Like other herbs of low stature, plants of G. hederacea are scattered. It colonized the Wicken Fen woodlands described by Rodwell (1991a) as a variant of W2a approximately 50 years after mowing ceased, when cover of Phragmites australis and other tall herbs had declined, and Urtica dioica and Poa trivialis had increased. In this condition the field layer resembled that of Alnus glutinosa–Urtica dioica (W6) valley fen woodlands on moist eutrophic soils. In the diverse collection of subcommunities within W6, G. hederacea has a frequency of II both in the typical subcommunity (W6a), in which Urtica dioica is very abundant, and in the Sambucus nigra subcommunity (W6d), which is found on rather drier soils. Urtica dioica is less prominent in the latter subcommunity, and Rubus fruticosus agg., Lonicera periclymenum, Hedera helix and Dryopteris spp. often form thick underscrub and ground layers. The Betula pubescens subcommunity (W6e) also occurs on drier soils. Here, Urtica dioica is again less common, while Rubus fruticosus agg. and Lonicera periclymenum produce a bulky underscrub, and G. hederacea falls in frequency. In the Alnus glutinosa–Fraxinus excelsior–Lysimachia nemorum communities (W7) that occur on wetter mesotrophic mineral soils, G. hederacea occurs in both the Urtica dioica (W7a) and Deschampsia cespitosa (W7c) subcommunities, although its frequency and abundance are low in both. Together with Poa trivialis, Ajuga reptans and Primula vulgaris, G. hederacea is one of a group of field layer species that are consistently found in Fraxinus excelsior–Acer campestre–Mercurialis perennis (W8) woodlands on heavy, base-rich soils in south-east England, in woods where Quercus robur has a high frequency. These species are all less common in W8 woodlands in the north-east, except where the soil is locally moister. In the W8a Primula vulgaris–Glechoma hederacea subcommunity it becomes common and can occupy large areas in coppice or coppice-with-standards of Quercus robur. Rodwell (1991a) suggests that the evergreen habit of G. hederacea helps to prevent its competitive elimination by Mercurialis perennis, which is abundant in this subcommunity. Following coppicing, G. hederacea can rapidly re-invade trampled sites in this subcommunity by stolonifery, and perhaps by seed, since it often flowers abundantly following such management. Glechoma hederacea has a lower frequency in the Anemone nemorosa subcommunity (W8b) that occurs on wetter sites. It is also less abundant in the more open Deschampsia cespitosa subcommunity (W8c) that develops when woodland on moist sites is coppiced. Here it is initially outcompeted by Deschampsia cespitosa and by annual and biennial weeds that grow rapidly in the less shaded conditions following coppicing, but it rallies as the canopy grows back. Under the continuous tree and shrub canopy in which the Hedera helix subcommunity (W8d) develops, G. hederacea is restricted to those areas where H. helix is relatively sparse on the ground and not replaced by either Mercurialis perennis or Hyacinthoides nonscripta. In both the Geranium robertianum (W8e) and Teucrium scorodonia (W8g) subcommunities G. hederacea has a low frequency and abundance. Glechoma hederacea is recorded with low frequency and abundance in all subcommunities but W10e (the Acer pseudoplatanus–Oxalis acetosella subcommunity) of W10 (Quercus robur–Pteridium aquilinum–Rubus fruticosus woodlands). In W10a–d its abundance is limited by that of several other species, including Pteridium aquilinum, Rubus fruticosus agg., Lonicera periclymenum, Hedera helix (in W10c), Holcus mollis (in W10d), dryopteroid ferns and, in the vernal aspect, either Hyacinthoides nonscripta (W10a) or Anemone nemorosa (W10b). In such vegetation G. hederacea may become locally prominent on heavy, moist soil where the cover of Pteridium aquilinum is low. In the Mercurialis perennis (W12a) and Sanicula europaea (W12b) subcommunities of W12 (Fagus sylvatica–Mercurialis perennis woodlands), and in W13a (the Mercurialis perennis subcommunity of Taxus baccata woodlands), G. hederacea is a component of sparse and patchy field layers under deep canopy shade. Glechoma hederacea is a consistent component of all recognized subcommunities within W21 (Crataegus monogyna–Hedera helix seral scrub and hedge communities), being preferentially associated with the Mercurialis perennis subcommunity (W21b). It also ranges in frequency from scarce to occasional in subcommunities included in W24 (Rubus fruticosus–Holcus lanatus underscrub) and W25 (Pteridium aquilinum–Rubus fruticosus underscrub). In W25a (the Hyacinthoides nonscripta subcommunity) it can achieve ground cover values of between 50% and 75% (Rodwell 1991a). In continental Europe G. hederacea is a characteristic species of wood-margin, hedge bank and scrub communities. Oberdorfer (Pfl. Exk.) records it in Klasse Artemisietea vulgaris, Unterklass Galio-Urticenea in the Orders Glechometalia (7–66% frequency) and Artemisietalia (5–32% frequency), but it is virtually absent from Order Onopordetalia. In the Netherlands, Westhoff & den Held 1969) mention G. hederacea specifically as a characteristic species of the Galio-Alliarion Alliance (natural or semi-natural tall forb communities of hedge-sides and woodland edges on moist, fertilized soil), the Trifolion medii Alliance (scrub/wood-margin communities with, for example, Agrimonia eupatoria, Galium mollugo, Vicia sepium, Veronica chamaedrys and Dactylis glomerata), and the Order Alno-Padion (wet woodlands) of the Fagetalia sylvaticae. In mesotrophic grasslands, G. hederacea is a component of all subcommunities of MG1 (Arrhenatherum elatius grasslands) which, in the absence of mowing and grazing, are often transitional to mixed deciduous woodland (Rodwell 1992). Although it can reach ground cover values as high as 33% in the Urtica dioica subcommunity (MG1b), it falls in frequency class I in all but the Pastinaca sativa subcommunity (MG1d) that often develops on more calcareous and oligotrophic soils in the south-east when grazing is relaxed, or when arable cultivation on marginal chalkland has failed. In acid grasslands G. hederacea is an element of subcommunities U1b–U1d of the Festuca ovina–Agrostis capillaris–Rumex acetosella community (U1). All these communities have open swards with small tussocky grasses, abundant dicotyledonous herbs and substantial amounts of lichens and mosses. The open conditions, created by the death of plants owing to drought, and disturbance caused by grazing animals, probably present opportunities for colonization by G. hederacea through stoloniferous growth. Its cover reaches its highest values (up to 33% Rodwell 1992) in the typical subcommunity (U1b). Oberdorfer (Pfl. Exk.) records frequencies of 3–33% for G. hederacea in meadow communities dominated by Arrhenatherum elatius and Trisetum flavescens. It also occurs in the Lolio–Cynosuretum. Westhoff & den Held (1969) refer to G. hederacea as a characteristic species of Arrhenatherum elatioris in the Netherlands. Several ruderal communities contain G. hederacea at low frequency (Rodwell 1999). It occurs in the Arrhenatherum elatius–Rubus fruticosus agg. subcommunity (OV24b) of OV24. This Urtica dioica–Galium aparine community is species-poor, tall-herb vegetation on nutrient-rich, moist soil, and often associated with disturbance and abandoned farmland and dwellings. In OV25 (Urtica dioica–Cirsium arvense community), which has a more open and patchy cover of U. dioica than OV24, G. hederacea is a component of the Holcus lanatus–Poa annua (OV25a) and Lolium perenne–Papaver rhoeas (OV25c) subcommunities. On damp and fertile soil that has been disturbed or burnt, it occurs in the Urtica dioica–Cirsium arvense (OV27b) and Acer pseudoplatanus–Sambucus nigra (OV27d) subcommunities of the Epilobium (Chamerion) angustifolium community (OV27). Glechoma hederacea is also a minor component of the Alisma plantago-aquatica–Sparganium erectum subcommunity (S5b) of Glyceria maxima swamp vegetation (S5) and of the Epilobium hirsutum subcommunity (S26b) of Phragmites australis–Urtica dioica tall-herb fen communities (S26). Westhoff & den Held (1969) also refer to G. hederacea as a component of a community dominated by Epilobium hirsutum and Phalaris arundinacea in the detritus zone of lake margins. Finally, it is also recorded in the Equisetum variegatum subcommunity (SD15b) and the Holcus lanatus–Angelica sylvestris subcommunity (SD15d) of Salix repens–Calliergon cuspidatum dune-slack vegetation (SD15), and in the Urtica dioica–Arrhenatherum elatius subcommunity (SD18b) of Hippophaë rhamnoides scrub (SD18) (Rodwell 1991b). IV. Response to biotic factors The responses of G. hederacea to biotic and abiotic environmental factors can be fully understood only by considering the relationship between vascular anatomy and phyllotaxis in the family Lamiaceae, and the constraints which this places on resource movement within these species. Stems in the Lamiaceae are quadrangular in section. Although the xylem forms a continuous cylinder within the stems, the major vessels are concentrated near the angles of the stem. Phloem cells are confined to four discrete regions directly adjacent to the stem angles. Thus, four main vascular bundles run parallel with the four stem angles. At each node a single leaf emerges from each of two opposing faces of the stem. The faces from which leaves emerge rotate by 90° between adjacent nodes. This opposite decussate pattern of leaf emergence (Bell 1991) results in each stem or stolon having four orthostichies of leaves. Stolons of a higher branching order are in the same orthostichy as their subtending leaves. A simple set of rules (Price & Hutchings 1992a) controls the relationship between vasculature and the patterns of resource distribution between leaves and stolons that arise in different orthostichies. These determine the patterns of physiological integration and independence between leaves and ramets. They are as follows: (i) leaves in opposite orthostichies, and their axillary stolons, are largely physiologically isolated from each other; (ii) leaves in adjacent orthostichies, and their axillary stolons, exhibit partial physiological integration; (iii) leaves in the same orthostichy, and their axillary stolons, are potentially fully integrated. These rules apply within single stolons. However, as new stolon branches develop they acquire the status of integrated physiological units (IPUs; Watson & Casper 1984) that are largely physiologically independent of each other. This happens because an additional rule operates as follows: (iv) resource translocation in the stele of G. hederacea is predominantly acropetal. New axillary stolons develop from the leaf axils when nodes have been superseded by several younger nodes towards the stolon apex. These axillary stolons capture most of the translocates leaving the subtending leaf and resources from the roots supplying the same vascular bundle. Thus, few resources are distributed from older nodes to the apex of the original stolon. Instead, the apex is provided with resources by the leaves and roots at more distal nodes on the original stolon. At this stage, each of the older nodes on the original stolon has become a largely physiologically independent segment that supplies resources primarily to stolons growing from the axils of its leaves (Birch & Hutchings 1999). When the axillary stolon can support its own maintenance and growth, it retains most of the resources it produces and acquires, rather than exporting them basipetally to other stolons. Thus it achieves the status of an IPU (Price & Hutchings 1992a). Translocation of resources from older to younger ramets increases the probability of daughter ramet survival (Slade & Hutchings 1987a). Physiological integration appears to involve no costs, at least in biomass terms, to the source ramet, but confers growth benefits upon the whole clone (Slade & Hutchings 1987b). The vascular anatomy of the species constrains the possible pathways of resource movement within the clone. There is, for example, little evidence of reciprocal translocation within stolons (Stuefer & Hutchings 1994). Price et al. (1996) compared the growth and morphology of genetically identical G. hederacea clones grown in isolation and in competition with Lolium perenne. The grass was either clipped to maintain a height of 1 cm, or allowed to grow, finally reaching approximately 13 cm. In the short grass treatment G. hederacea mainly experienced competition for soil-based resources, whereas in the tall grass treatment there was competition for both soil-based resources and light. Although total biomass of G. hederacea was significantly reduced by the presence of competing Lolium perenne, it was not significantly affected by its height. Primary stolon growth was unaffected by competition, but the number and length of secondary stolons were significantly reduced in the short grass treatment, and there was a further significant reduction in competition with long grass. Internode and petiole lengths increased significantly in clones competing with tall grass, but were not significantly affected by short grass. The proportion of biomass allocated to stolons increased from 20% in clones grown without competition to 36–38% in clones subjected to short and tall grass competition (Hutchings & Price 1993). Allocation of biomass to petioles also increased, from 15% without competition to 21% with short grass competition and 35% in clones growing with tall grass. Both leaf mass and area were significantly lower when G. hederacea was grown with competition. Thus, clone morphology was considerably changed in response to competition, becoming more linear in the horizontal plane and more extensive in the vertical plane. This raised the horizontally orientated leaves to the top of the grass canopy. These responses would promote avoidance of competition in both the horizontal and vertical planes; energy invested in densely vegetated areas would be reduced, and occupation of less utilized sites enhanced. Reductions in photosynthetically active radiation (PhAR) and in the R/FR ratio incident on the plant, and raised humidity caused by the competitor, were all implicated in causing these morphological changes. Some treatments in this experiment examined the effect of habitat heterogeneity by growing different stolons of the same clone either with or without competition from either short or tall grass. Stolons treated in this way developed different morphologies, each statistically indistinguishable from the morphologies of stolons from clones growing under the corresponding homogeneous conditions. This is a consequence of the vascular structure of the clone, and the constraints imposed by this structure upon resource movement. Each stolon behaved as an IPU. This enabled each stolon to develop a morphology appropriate to its local conditions, rather than a morphological compromise for the variety of conditions experienced by the clone. Notably, the total yield of clones with one stolon competing with tall Lolium perenne and the other growing without competition significantly exceeded that of clones growing entirely without competition. While the shaded stolon produced a lower yield and fewer branches than equivalent stolons grown without competition, the growth of the unshaded stolon apparently benefited substantially from greater local humidity caused by the grass canopy shading the adjacent stolon. In the absence of accompanying shade this led to a proliferation of secondary and tertiary stolon production. The subspecies hirsuta, which is sometimes accorded species status as G. hirsuta, produces both orthotropic (vertically orientated) and plagiotropic (horizontally orientated) shoots. As in G. hederacea, shading reduces growth of plagiotropic shoots but has little effect on orthotropic shoots. Vertically orientated internodes of orthotropic shoots show a greater proportional increase in length in response to shade than horizontally orientated petioles, whereas the opposite is true for plagiotropic shoots, supporting the hypothesis that, within a species, responses to shade will depend on orientation of growth, and that vertically orientated structures should be more plastic (Huber & Hutchings 1997). The hypothesis proposes that because light gradients are more predictable in a vertical than horizontal direction, plasticity in the lengths of vertically orientated spacers will more often result in greater light acquisition. Internodes of the plagiotropic stolons of G. hederacea lengthen under shade and other adverse conditions, including low nutrient supply. This lengthening is accompanied by significant decreases in weight per unit length and stolon cross-sectional area, mainly achieved by reducing the volume of the parenchymatous tissue. The vascular system is considerably less plastic, thus conserving its capacity for resource transport (Price & Hutchings 1992a). In the field, G. hederacea has a limited capacity to persist under tall herbs or beneath a continuous tree canopy (see Section III above). Grime et al. (1988) describe it as a species that is most characteristic of lightly or patchily shaded habitats. It may become locally abundant and flower profusely after coppicing (Fl. Br. Isl.), but gradually decline in vigour, and even be excluded as the canopy regrows. A study by Ash & Barkham (1976) suggested that it has the capacity to persist for only 3 years after coppicing. Glechoma hederacea is toxic to many species, although many phytophagous invertebrates are known to feed on it. Price & Hutchings (1992b) analysed the effect on plant performance of simulated herbivory in the form of selective defoliation, to test the hypothesis that both the structural level at which G. hederacea responds to phytophagy, and the specific response elicited, will be determined by pattern of defoliation and by the extent of physiological integration between the defoliated ramets. Yield declined progressively in clones subjected to 0%, 50% and 100% defoliation. Leaves were removed in three patterns for the 50% defoliation treatments, and yield was significantly affected by removal pattern. Compared with undefoliated clones, similar reductions in yield and in several components of yield were caused when defoliation was achieved either by removing all leaves from one of two stolons and none from the other, or by consistently defoliating two of the four orthostichies on both stolons by removing one of the two leaves from every ramet. However, when consecutive pairs of ramets on each stolon were either defoliated or left intact, the effect on yield and its components was significantly smaller. Compared with undefoliated clones, this pattern of defoliation had little impact on growth at a primary structural level, although secondary growth was impaired. This treatment, which allowed some of the leaves in all four orthostichies to contribute photosynthate to the clone carbon economy, was less damaging than treatments that rendered some orthostichies completely unproductive. Analyses of the effects of stolon apex removal in G. hederacea have shown that losses in potential biomass production (estimated by comparison with the growth of clones without apices removed) and in several other growth characteristics measured at a primary structural level were almost exactly compensated for by increased growth of secondary branches (Pelling 1994). In contrast, when clones were shaded, or when rooting of ramets was prevented, some growth-related measurements either over- or under-compensated in comparison with untreated clones. Studies of the effects of leachates of G. hederacea on Raphanus sativus a

Editorial – Clonal plants and environmental heterogeneity – space, time and scale

The fifth Clonal Plant Workshop entitled 'Clonal plants and environmental heterogeneity – space, time and scale' was held at the University of Wales, Bangor, from the 9th to the 14th September 1997. The workshops began in 1988, when a small group of plant scientists with different backgrounds met in the Netherlands to discuss plant clonality. This meeting was followed by workshops in 1990, 1992 and 1995 in Sweden, the Czech Republic and Hungary, respectively. The purpose of the Bangor meeting was to continue to integrate knowledge from a wide range of biological disciplines to attempt a better understanding of clonal plants in the context of the heterogeneous environments in which they occur. This special issue is a collection of a number of the presentations at the Bangor workshop. The effects of spatial and temporal environmental heterogeneity on individuals, populations and communities are, of course, diverse, and this collection of papers is not intended to provide a synthesis of the ecological consequences of environmental heterogeneity. Rather, this special issue more accurately reflects progress in various areas of clonal plant research including population dynamics of ramets and genets, mathematical modelling and molecular techniques. We wish to thank all participants of the workshop for their contribution and trust that they found the meeting as stimulating and enjoyable as we did. We are indebted to the authors and the many referees for their friendly cooperation and to Plant Ecology for publishing the proceedings. We also gratefully acknowledge the financial support of the following that made the workshop possible: The British Ecological Society, The Tansley Fund of the New Phytologist Trust, The Countryside Council for Wales, The Institute of Terrestrial Ecology (Bangor) and Skye Instruments Ltd, Wales.

Vulpia ciliata Dumort. ssp. ambigua (Le Gall) Stace &amp; Auquier (Vulpia ambigua (Le Gall) More, Festuca ambigua Le Gall)

Tribe Poeae. An annual grass, small, solitary or loosely tufted, 5–30 (–45) cm high. Culms slender, erect, sometimes bent and branched at or near base; 1–2 noded, sheath enclosing scape to base of panicle or higher. Branching intravaginal. Leaves (1–10 cm) green, purplish below, especially in seedling, tightly inrolled or opening out and up to 2 mm wide. Sheaths smooth, rounded on the back; very short membranous ligules. Blades rough on the margins and minutely hairy on the upper surface. Panicles erect or slightly nodding, very slender, linear or lanceolate, one-sided, 3–13 (–20) cm long; green, purplish or reddish. Inflorescence little branched and raceme-like in the upper part or throughout; branches short and erect. Pedicels up to 1 mm, swollen. Spikelets overlapping, wedge-shaped or narrowly oblong, 5–7 mm long excluding awns, 12–18 mm including awns. Florets, 3–7, lower 1–2 (–3) fertile, upper (2–) 3–6 (–7) sterile or male only. Glumes awnless and much shorter than lowest lemma, persistent. Lower glume ovate or oblong, veinless, 0.2–1.0 mm long, sometimes longer on spikelets that are terminal on the inflorescence or on its branches. Upper glume linear to lanceolate, 1-veined, 1.5–3.0 mm long, the fragile, membranous tip often breaking to give a blunted appearance. Florets separated by rhachilla segments, the lowest ± sessile within the glumes. Lemmas narrowly lanceolate in side view, firm, minutely scabrid, overlapping at first and later with the margins incurved, with 3–5 inconspicuous nerves, tapering apically into a fine awn up to 10 mm long. Fertile lemmas 3 (–5) veined, 4–5 mm long excluding awn, 11–17 mm long inclusive of awn. Sterile lemmas up to 6 mm long with shorter awns. Small horseshoe-shaped callus at base of fertile lemmas, 0.1–0.2 mm in longitudinal axis. Paleas of fertile florets almost as long as lemmas, 4.0–4.9 mm with two rough keels. Palea of sterile floret absent or much reduced. Anthers 1 (–3), 0.4–0.6 mm long. Ovary tip glabrous. Caryopsis, linear, slender, 3.5–4.5 mm, enclosed by toughened lemma and palea and firmly attached to the latter, breadth-length quotient c. 1:7 (Cotton & Stace 1977). Short embryo visible at base of dorsal surface. Abaxial surface rounded and inrolled to form groove on adaxial surface. Narrow hilum extends almost entire length of caryopsis. Plants exhibit considerable phenotypic plasticity, varying in height, the amount of tillering and spikelet number. In an unheated glasshouse, exceptional plants have reached a height of 70 cm, but on very poor soils plants may attain only a few cm with one or two spikelets. Growth habit is generally consistent across sites in the British Isles. Populations from Santon Downham in the Breckland and Holme next the Sea in Norfolk at the northern edge of the species range are genetically differentiated in relation to both biomass and seed production (Norton 1996). The plants from Santon Downham grown in a controlled environment facility produced 25% more seeds than those from Holme next the Sea. There were no significant population × environment responses in relation to either temperature or carbon dioxide. There are two subspecies of Vulpia ciliata (Stace & Auquier 1978). Subspecies ambigua (Le Gall) Stace and Auquier (V. ambigua (Le Gall) More), the plant reported here, is devoid of long hairs on the inflorescence, although the lemmas, glumes, rhachis, etc. are scabrid with prickle hairs. In contrast, ssp. ciliata is loosely hairy on the lemmas; the fertile lemmas are pubescent on the dorsal midline and sparsely ciliate on the margin, whereas the sterile lemmas are densely ciliate on the margin. Subspecies ambigua also has a distinctive appearance with a stiffly erect, very narrow inflorescence and a purplish colouring at fruiting. Otherwise, the two taxa differ only by more or less overlapping quantitative characters; ssp. ciliata is generally larger in all its parts with spikelets mostly 7–10.5 mm, fertile lemmas 5–7.5 mm, and sterile lemmas up to 8 mm, all measurements excluding awns. The two subspecies are chiefly allopatric (Fig. 1) with ssp. ambigua restricted to England, Wales and a few localities in Belgium and north-western France (Cotton & Stace 1976). The degree of pubescence in ssp. ciliata is extremely variable (Cotton 1974) and glabrous individuals occur sporadically throughout much of this range and are referred to as V. ciliata var. imberbis (Vis.) Halácsy (Stace & Auquier 1978); these plants can be distinguished from ssp. ambigua by the slightly larger size of their floral parts. The distribution of Vulpia ciliata in Europe. Hatched area is the distribution of V. ciliata ssp. ambigua in the British Isles, Channel Islands, northern France and Belgium. The numbered sites in France are: 1, Le Kernével and Larmor; 2, La Guimorais; 3, Châtelaillon; 4, Brenne; 5, Amboise. Unhatched area is the distribution of V. ciliata ssp. ciliata. Reproduced from Cotton & Stace (1976) as modified by Stace & Auquier (1978). Subspecies ciliata is occasionally found as a casual in England (Hubbard 1984) and was formerly naturalized at railway sidings in Ardingly, East Sussex (Stace & Auquier 1978; Stace 1997). Vulpia ciliata ssp. ambigua is a native species of maritime or submaritime sand and shingle and is also found on roadsides and by tracks in sandy areas inland. All subsequent references to V. ciliata in this paper are to ssp. ambigua. Although formally regarded as ‘scarce’ in a national context (Stewart et al. 1994), the plant may be locally abundant. A key feature in distinguishing ssp. ambigua from other British Vulpia species is the relative length of the two glumes. These are very unequal and the lower is less than one-quarter (0.1–0.25) the length of the upper. Vulpia ciliata belongs to the Mediterranean–Atlantic element of the British flora (Preston & Hill 1997). Distribution in the British Isles centres around the inland, sandy heaths of Norfolk and Suffolk known as the Breckland, and along parts of the southern and eastern coasts extending as far north as the Wash (Fig. 2). In East Anglia, the main range of the species extends from the north Norfolk coast southwards through the eastern counties of Cambridgeshire and Essex. Recent sightings of the species have also been made at Bicester in Oxfordshire and Shoebury in Essex (P.D. Carey & K.J. Adams, personal communication). It is sparsely distributed along the south coast of England, in Devon, Dorset, Hampshire, the Isle of Wight and Sussex, where it occurs on golf links and coastal dunes and inland on sandy roadsides, cinder paths and similar dry places. Vulpia ciliata also occurs in Kent where it is found on maritime and inland shingle. The distribution of Vulpia ciliata ssp. ambigua in the British Isles. Each dot represents at least one record in a 10-km square of the National Grid. Mapped by Mrs J.M. Croft, Biological Records Centre, Institute of Terrestrial Ecology, Monks Wood. (○) pre-1970; (•) 1970 onwards; (+) introductions pre-1970; (×) introductions 1970 onwards. The most northerly occurrence of the species was in north Yorkshire at West Tanfield in 1954. No subsequent records exist of plants at this location. The current two most northerly records are Crosby Warren in north Lincolnshire at the entrance to a quarry and Holme next the Sea in north Norfolk, where there are two small patches of the plant on the edge of a golf course at the back of the dunes. The species’ distribution in western Britain is confined to a few locations in Devon and Somerset and to one site in Gwynedd on the coastal golf course at Aberdyfi. The recorded world distribution of ssp. ambigua is limited to England, Wales, the Channel Islands (Jersey, Guernsey and Alderney) and a few locations on the neighbouring coasts of continental Europe (Fig. 1). There are records of populations along the coast of Belgium and in northern France in the départements Pas de Calais and Somme (Lambinon 1958; Auquier 1977; Stace & Auquier 1978). The subspecies has, however, spread in Belgium in recent years (C.A. Stace, personal communication), and is probably under-recorded in France, where it may well be more widespread in Brittany to the south of the distribution limit shown in Fig. 1; note that the subspecies was first described from Morbihan in southern Brittany within the range of subspecies ciliata (see Section X). Five outlier populations have been recorded within the range of V. ciliata ssp. ciliata (Fig. 1) and both subspecies co-occur on dunes at Châtelaillon in northern France (Stace & Auquier 1978). Subspecies ambigua is a lowland grass in Britain and generally occurs at altitudes between sea level and 60 m (Table 1). In contrast, ssp. ciliata is widespread throughout southern and central Europe (Fig. 1), extending south to North Africa and west at least as far as north-western India (Cotton & Stace 1976). It occurs from sea level to about 2000 m and has a much wider ecological amplitude than subspecies ambigua (Stace & Auquier 1978). Subspecies ciliata is found as a pioneer species on well-drained, open or broken ground, both acidic and calcareous, open grassland, waste places, wood margins or clearings and dunes, usually on sandy or rocky substrata. It is one of the commonest Vulpia species in Europe (Cotton & Stace 1976) and has been introduced into Australia (Willis 1970). The grass is most abundant in the dry eastern counties of England. Throughout its range in the British Isles it is associated with areas of low annual rainfall, typically <900 mm per annum, and at the middle of the species’ range in the Breckland <650 mm per annum (Table 1). The mean annual rainfall across the sites listed in Table 1 is 713 mm and the mean annual minimum and maximum temperatures 6.9 and 13.3 °C, respectively. Norton (1996) has investigated the effect of climate on the survival and fecundity of V. ciliata within and outside its current range by sowing seeds in pots at 22 sites throughout Britain, spanning Orkney in the north to Exeter in the south. The experiment was carried out in 1993–94 and it was found that germination was high at all sites but took longer under the cooler temperatures and shorter days in northern Britain; germination occurred 15 days later in the north. There were no differences in survival between the sites but the long, warm days associated with southerly latitudes shortened the time to seed maturity and increased seed production. Vulpia ciliata produced seeds at all sites in the transplant experiment (Norton 1996), and it was only in Orkney that the finite rate of population increase was less than one, i.e. the number of seeds produced fell below the number of seeds sown. Seed production at a number of sites in the west and north exceeded that within the native range of the plant, and the fundamental climate niche for V. ciliata appears to extend both to the north and west of the present species’ range. The realized climate niche has been modelled by Carey et al. (1995) using multiple regression. They found that the current distribution of the plant could best be modelled in terms of three climatic factors [sunshine hours in March–May; temperatures in winter (December–February) and autumn (September-November)] and concluded that the species is found where there are sunny springs, warm autumns and cold winters. The transplant experiment of Norton (1996) indicates that the requirement for cold temperatures is not absolute, and given that the plant can perform very well outside of its current range, it appears that the present distribution of the plant is more limited by the availability of suitable habitats (Carey et al. 1995). The model produced by Carey et al. (1995) relates primarily to the factors that determine the northern and western limit of the subspecies; different or additional climatic variables may be correlated with the southern edge of its distribution. Vulpia ciliata is often found on small local ridges of slightly elevated ground, for example on banks and track edges, and occasionally on walls. However, the species is not confined to a specific aspect and is also found on flat ground and moderate slopes. At beach locations such as at Snettisham in Norfolk, it can tolerate exposure to high winds, salt spray and occasional inundation by high tide seawater. Vulpia ciliata grows on maritime dunes, especially fixed and semi-fixed dunes, over most of its range (Stace & Auquier 1978), but it also occurs on maritime shingle in Kent and East Anglia and on inland sandy heathlands, especially in the Breckland of East Anglia. Within the Breckland it is found on disturbed sand, on sparsely colonized verges of forest tracks, open areas on heath tracks, on wind-blown low sand-banks under pine belts, open sides of pits, arable headlands in areas of thin crops such as rye, on derelict arable land, in loose sand at the base of wire fences, and in disturbed ground on the sides of furrows (Trist 1979). The locality at Brenne, within the range of ssp. ciliata in France, resembles the Breckland sites in several ways (Stace & Auquier 1978). The subspecies shows a strong preference for dry sands and shingles (Carey et al. 1995) that are well drained with a relatively low organic matter content. Carey et al. (1995) found that within the UK, populations were particularly associated with Brown sands, Brown calcareous sands and Sand-pararendzinas by overlaying the current distribution of the plant with the dominant and subdominant soil types on a 1-km scale for England and Wales. The mean pH of soils on which the plant grows in East Anglia is 6.8, but the species occurs in soils with pH values as low as 3.4 and as high as 7.9 at Santon Downham and Red Lodge, respectively (Table 2). While ssp. ambigua is frequently regarded as a plant of inland, acidic sands, data in Table 2 indicate that it is more characteristic of soils that are neutral to slightly acidic. Soils colonized by V. ciliata are usually nutrient-poor, typically with NH+4-N, NO−3-N, P and K contents of <9.0, <13, <50 and <130 mg kg−1, respectively (Table 2). In west Norfolk and Suffolk, the most extensive populations are found on soils of low moisture content and low organic matter, phosphorus and potassium. Vulpia ciliata is commonly confined to small areas and is seldom found widely dispersed at any location. It may, however, be locally dominant, particularly in disturbed areas, where it is particularly noticeable at the time of fruiting when the infructescence turns its characteristic reddish colour. A list of higher plant species that are commonly associated with V. ciliata in East Anglia is given in Table 3. It is a species of open communities, where perennial cover is typically low and usually includes a range of other annuals, in particular Aira praecox, Anisantha sterilis, Arenaria serpyllifolia, Bromus hordeaceus, Myosotis ramosissima and Vulpia bromoides. Achillea millefolium, Erodium cicutarium, Festuca spp., Koeleria macrantha, Plantago coronopus, P. lanceolata, Poa pratensis, Rumex acetosella, Sedum acre, Taraxacum spp. and Trifolium campestre are also frequently found in association with V. ciliata. In the Breckland, Trist (1979) recorded that V. ciliata was frequently found in association with Arenaria serpyllifolia, Rumex acetosella and Sedum acre. These species are all characteristic of sandy heaths, dunes and shingle. On the last it is also found in association with characteristic shingle species such as Glaucium flavum, Rumex crispus and Silene uniflora. On dunes, it does not grow on such disturbed sand as V. fasciculata, but at Holme next the Sea in Norfolk there is one small patch of dune vegetation at the back of the local golf course where V. bromoides, V. ciliata, V. fasciculata and V. myuros have all been found growing together. In the Breckland, V. bromoides, V. ciliata and V. myuros are occasionally found together. In such vegetation, V. ciliata typically occurs on bare ground in a matrix of perennials dominated by forbs (e.g. Achillea millefolium and Plantago lanceolata), sedges (e.g. Carex arenaria) and grasses (e.g. Festuca rubra and Koeleria macrantha). The species is also commonly found growing with mosses such as Brachythecium albicans, Ceratodon purpureus, Hypnum cupressiforme var. lacunosum, Polytrichum piliferum and Tortula ruralis ssp. ruraliformis (Newsham et al. 1994; 1995). Locally V. ciliata may be found where there is complete moss cover. Germination does not appear to be adversely affected by such cover and may even be facilitated by the improved microclimate within the cushion where the seeds germinate (Newsham et al. 1995). Subsequently the plants root through the moss cushion and leaves rapidly emerge above its surface. Plants are also commonly found growing amongst lichens. For example, at Mildenhall in Suffolk, V. ciliata occurs with Cladonia rangiformis, C. fimbriata, C. squamosa and Diploschistes scruposus (Newsham et al. 1995). It is difficult to place the communities in which V. ciliata is found within the National Vegetation Classification (Rodwell 1992; Rodwell in press) as many of the patches are small, disturbed and influenced by the seed rain from adjacent vegetation. The sample descriptions in Table 1 are, however, consistent with the species being found in Rumex crispus–Glaucium flavum shingle communities (SD1) and communities that are typically transitional between Festuca ovina–Agrostis capillaris–Rumex acetosella grassland (U1), semi-fixed and fixed sand dune types (e.g. SD7 and 8), and disturbed communities of mesotrophic neutral pH conditions (e.g. MG7f, OV23 and OV27). Subspecies ciliata is described by Braun-Blanquet et al. (1952) as a member of the Brachypodietum ramosi association, the richest in species in the French midi. It is one of the characteristic species that splits the western Provence race from the Languedoc race and gives its name to the subassociation Vulpietosum ciliatae. The competitive ability of V. ciliata has been studied by Carey (1991) in a series of experiments in the glasshouse and field. The effects of intraspecific competition on the fecundity of the species was investigated by sowing a monoculture at five densities in trays containing soil from Red Lodge, Suffolk, in an unheated glasshouse. Non-linear regression analysis showed that there was a negative relationship between the mean fecundity (seed number) of individual plants and the density of V. ciliata (fecundity = 1123 [1 + density]−0.849). The effect of local crowding on the grass was also studied using polygon and neighbourhood analysis. Data pooled from all density classes showed that 64% of the variation observed in an individual plant's fecundity in monoculture was related to the number of neighbours surrounding it within an optimal radial distance (the distance over which a plant can influence its neighbour) of 43 mm. The response of V. ciliata to interspecific competition from perennials (Festuca rubra and Plantago lanceolata) and other annuals (Aira praecox, Anisantha sterilis and Vulpia bromoides) was also investigated in an unheated glasshouse by Carey (1991). The mean individual fecundity of V. ciliata within areas bordered by Festuca rubra was reduced by 60–90% when grown within gaps, but no correlation between the density of V. ciliata and its mean fecundity could be discerned in these areas. Further experiments (A.R. Watkinson, L. Forrester & R.P. Freckleton, unpublished data) in an unheated glasshouse have investigated the effects of gap size and gap type on the performance of V. ciliata sown over a range of densities. Gaps were created by planting small bunches of Festuca ovina tillers around square gaps and also by planting mixtures of (i) Hypochaeris radicata and Plantago lanceolata, (ii) F. ovina and Koeleria macrantha and (iii) F. ovina, H. radicata, K. macrantha and P. lanceolata. The competitive pressure exerted on V. ciliata plants grown in 20 × 20 cm gaps in a matrix of F. ovina was equivalent to c. 8000 V. ciliata plants m−2. Competitive pressure exerted by F. ovina increased rapidly as gap size decreased; within 10 × 10 cm gaps, the competitive effect rose to 13 500 plant equivalents m−2. However, decreasing the gap size to 5 × 5 cm had no further measurable effect on the competitive pressure exerted on V. ciliata. The competitive pressure exerted by the three types of gap was approximately equivalent. The implication, in the field, is that the mean plant fecundity of V. ciliata will decline as perennial cover increases (see V (b)). However, at the level of the individual gap in the field, where the gaps are often small and diverse in nature, variation in gap size will have little effect on fecundity as the competitive pressure exerted by the perennials will be relatively constant. Also, variation in the density of V. ciliata can be expected to have little effect on mean individual fecundity as the competitive pressure exerted by the perennials is so large. This accords with the observation that manipulating the densities of V. ciliata in the field has no impact on fecundity (Carey et al. 1995) and with the observation made by Watt (1971) that V. ciliata belongs to a group of small annual plants with low competitive ability. Vulpia ciliata is commonly situated on paths and roadsides and is consequently subjected to trampling by humans and domestic animals. Disturbance can favour the annual by preventing the establishment of perennial species and maintaining an open habitat. Trampling may also aid dispersal through disturbance (Carey & Watkinson 1993); minute upward and forward pointing prickle hairs on the awns of the lemma may also help lodge the diaspores in clothing and fur. Heavy trampling during flowering may damage inflorescences and reduce seed production. Large and small scale disturbances may both destroy and create opportunities for populations. Ant nests, rabbit scrapes and vehicular disturbance have all been observed to create gaps within populations that are then recolonized; for example, seedlings may be found aggregated in depressions in the soil made by tyres. At a larger scale, coastal erosion and the maintenance of sea defences have been observed to destroy populations on shingle, ploughing and tree felling activities have been observed to destroy populations on headlands and forest rides, respectively, while the abandonment of sand and gravel pits has been seen to lead to the establishment of new populations. After the British Museum archaeological dig at High Lodge, Mildenhall in 1970, thousands of V. ciliata had established by 1976 (Trist 1979); a small population still occurred on the site in 1996 in the grazed turf, isolated from other populations by the surrounding pine plantation. No direct evidence of herbivory on V. ciliata has been recorded, even in areas grazed by sheep, deer and rabbits. This is in contrast with the closely related V. fasciculata, which is often found eaten by rabbits during reproductive stages (Watkinson 1978). Grazing of the species may be controlled by the presence of clavicipitaceous fungi, hyphae of which are frequently observed in leaf sections of mature plants. These fungi, which colonize the above-ground plant parts of many grasses, deter feeding by animals and insects by producing alkaloids in leaves (Clay 1990). Some populations, for example at Bodney, Norfolk, and Mildenhall, Suffolk, are mown in late May or early June to tidy up the vegetation close to buildings and along tracks. This rarely has a significant impact on population size in the subsequent year, as a sufficient number of inflorescences escape mowing and/or sufficient seed has generally been set by this time of year. There is generally no production of secondary tillers after mowing. Vulpia ciliata usually forms small, patchy populations, many of which persist over a number of decades (see X). Perennial plants probably limit the size of patches and the density of V. ciliata within them (see IV (a)). Owing to the short seed dispersal distance of the species (see VIII (c)), seedlings are often found aggregated around the remains of inflorescence stalks from the previous year. The largest population in East Anglia extends in a 10-m wide strip intermittently for c. 3 km along an exposed, disturbed shingle beach at Snettisham. More usually, patches tend to range in size from one to a few thousand m2; the distribution of population sizes as measured by area is approximately lognormal, with the area of most populations being c. 100 m2. The area of the four populations studied by Carey et al. (1995) in East Anglia ranged from 184 to 5340 m2, with average densities of plants in the populations ranging from 300 to 878 m−2. The numbers of plants in these populations were estimated to be between 20 000 and 4 700 000. At the edge of its range, and as perennial cover increases, V. ciliata populations occur at densities below 60 plants m−2. Within a population at Mildenhall, where the densities of plants were recorded in contiguous 10 × 10 cm quadrats within four plots of 2 × 1 m, the densities of plants ranged from 0 to 7000 m−2, but most of the densities fell within the range of 500–3500 m−2, with a mode of just under 2000 m−2. An analysis of population performance in the 10 × 10 cm quadrats at Mildenhall showed that density had no effect on mean plant fecundity, but that population growth rate was density-dependent as a consequence of density-dependent recruitment (A.R. Watkinson, R.P. Freckleton & L. Forrester, unpublished data). Population growth rates were reduced by c. 80% as a consequence of density-dependence. Population growth rates were also reduced by 30% as a result of perennial plant cover, but effects could potentially be much higher if the competitive effects of perennials were completely removed (see IV (a)). Analysis of the effects of cover on the growth rate of populations indicated that the effects were not mediated by resource competition, but rather through seed survival through to germination. Vulpia ciliata reproduces freely at all suitable sites throughout its range in the British Isles, and beyond (see II (a)). A study in 1994 examining the relationship between seed number per plant of V. ciliata and soil characteristics at 14 sites in East Anglia indicated that fecundity was higher on soils with relatively low soil moisture (r = −0.73, P < 0.01), organic matter (r = −0.68, P < 0.01) and magnesium (r = −0.70, P < 0.01) content and relatively high nitrogen content (r = 0.55, P < 0.05). There was no evidence to suggest that pH or the concentrations of phosphorus or potassium were associated with plant fecundity. The fecundity of plants also increased with the amount of bare ground (r = 0.74, P < 0.01) in the communities studied. Carey et al. (1995) investigated the population dynamics of V. ciliata at four locations in East Anglia situated along a 75-km transect from the Norfolk coast in the north to Mildenhall, Suffolk, in the south. Plant performance varied both between and within sites. The basic reproductive rate of V. ciliata was higher at the core of the species’ range than at its northern limit. This was attributed to decreased fecundity at more northerly locations, although mortality did also increase to a small extent at these sites. Within sites, high density zones of V. ciliata had earlier seedling emergence and higher survival and fecundity. However, the primary reason for the decline in abundance towards the margin of populations was a decline in the fecundity of individuals, which was strongly correlated with the finite rate of population increase. In a further study of the population dynamics of V. ciliata carried out over 3 years at five contrasting sites (A.R. Watkinson & L. Forrester, unpublished data), it was found that the pattern of population flux was similar although there were significant variations in detail. These sites included coastal dunes, maritime shingle, Breckland heath and the verges of forest tracks. The net rate of population increase varied from 0 to just under 2 in 1993–94 across the five sites; there was local population extinction on the verge of the forest track resulting from forestry operations. In contrast to the findings of Carey et al. (1995), the highest rate of population increase was at the most northern site (Holme next the Sea), while at the other sites the rate of population growth varied from 1 to 1.5. Except on the verge of the forest track, where conditions had changed very substantially as a result of tree felling, the net rate of population increase was < 1 at all the other sites in 1994–95. This was owing to drought in spring 1995, which resulted in higher mortality of vegetative plants than in the previous year. Variation in population density between sites was greater after the dry spring than at any other time. The net rate of population increase across the sites over two years was positively correlated with plant fecundity, supporting the conclusion of Carey et al. (1995) that variations in the finite rate of population increase and abundance between sites are largely a result of variation in the number of seeds produced per plant. At the northern edge of the species’ range at Holme next the Sea, the population was found to have a high population growth rate in one year but negative growth rates in two other years, including the one monitored by Carey et al. (1995). It is clear that the persistence of populations at the northern edge of the species’ range depends critically upon the balance of the finite rate of population increase from year to year. Although disturbance had a major impact on the population by a forest track at Santon Downham, data from Sandringham in Norfolk and a number of other sites indicate that overall population structure is fairly stable with high density patches remaining high and low density patches low over considerable periods of time. Th

Issues in GIS development: adapting to research and policy-needs for management of wet grasslands in an Environmentally Sensitive Area

Geographical Information Systems (GIS) represent a rapidly changing technology, and awareness of their capabilities outside the GIS arena has grown rapidly during the last few years. As part of a wide-ranging ecological research programme on wetlands in Environmentally Sensitive Areas (ESAs) the Institute of Terrestrial Ecology (ITE) developed a GIS for the examination of lowland wet grassland landscapes, and the potential for their restoration (the 'Wetlands GIS'). A flexible approach to the development of the GIS was necessary to accommodate changes in technology and in the needs and interests of the key 'stakeholders' in the project: the research ecologists who supplied much of the data for the GIS and the policy-makers concerned with the application of research findings to land-management problems. This paper explains the rationale behind the use of GIS in the context of ESA-management and its evolution over a three year period. Examples of output are used to demonstrate the benefits of map displays for encouraging communication between developers and users, the data-storage, handling and analytical capabilities of the GIS and its role in matching the needs of researchers and policy-makers.

Scope for strategic ecological assessment of trunk-road development in England with respect to potential impacts on lowland heathland, the Dartford warbler (Sylvia undata) and the sand lizard (Lacerta agilis)

Environmental impact assessment (EIA) has a potential role in sustainable development. However, its commonest application has been at the project level, where it fails to ensure adequate consideration of potentially serious trans-boundary, widespread, indirect, cumulative and synergistic ecological effects. The fact that ecological effects cannot be predicted or evaluated effectively if EIA is confined to single development actions and contrained by artificial boundaries is an important justification for moves towards the strategic application of EIA. The need for a more strategic approach to ecological assessment can be illustrated by reference to road development in the UK, where project-level EIA has failed to quantify the overall impact of new road development on biodiversity. Some form of strategic ecological assessment (SEcA) is required to ensure that proposed new road development is compatible with international obligations to conserve protected habitats and their associated species. In common with all forms of EIA, the effectiveness of SEcA depends on the ability to define the proposed action or set of actions and to characterize the receiving environment (baseline conditions). The ability to quantify potential impacts and to estimate their risk of occurrence is strongly dependent on the availability, accuracy, reliability and resolution of national data on the distributions of habitats, species and development proposals. This paper summarizes the results of a SEcA carried out to determine whether currently proposed road developments, when considered collectively, would be compatible with the safeguard of an internationally important and protected wildlife habitat (lowland heathland) and associated protected species. The study was carried out using a version of the Countryside Information System developed for the UK Department of the Environment by the Institute of Terrestrial Ecology. Areas of lowland heathland are identified together with «hot spots» where two protected lowland heathland species (sand lizard,Lacerta agilisand the Dartford warbler,Sylvia undataalso occur. Possible risks attributable to proposed new road development are quantified for 1 km squares and evaluated on a county basis.1998 Academic Press Limited

Andromeda polifoliaL.

A low, glabrous, evergreen shrub, (5)–10–25–(40) cm tall. The flowering stems are decumbent to erect, sparingly branched, arising from a creeping, woody rhizome (Warming 1908; Fl. Eur. 3; Fl. Br. Isl.; Hegi Fl. 1, ed. 3; Rameau et al. 1993; Stace 1997). The leaves are narrow, alternate, subsessile, coriaceous, oblong to linear, dark green above, silvery-glaucous beneath, 10–40 × 2–8 mm (Barvaux 1958; Fl. Eur. 3; Fl. Br. Isl.; Rameau et al. (1993); Stace 1997). They are mucronulate (Anderson 1961) or acute to apiculate (Fl. Eur. 3); their margins are usually revolute (Fl. Br. Isl.; Rameau et al. 1993). The flowers are pentamerous (Barvaux 1958; Anderson 1961; Hegi Fl. 1, ed. 3), in very short, umbel-like terminal racemes of 2–7–(12) (Warming 1908; Fl. Eur. 3; Rameau et al. 1993). However, Warming (1908) reported 4-merous flowers in Finnmark and in Greenland. The pedicels are pinkish, 2–4 times as long as the corolla (Barvaux 1958; Fl. Br. Isl.; Hegi Fl. 1, ed. 3); they are 7–15 mm long and slender (Fl. Eur. 3). The sepals are persistent, triangular, reddish, acute, about 1.5 mm long (Anderson 1961; Fl. Eur. 3) and connate at base (Fl. Eur. 3). Two bracteoles lie at the extreme base of the pedicel, concealed by a bract (Fl. Eur. 3). Corolla globose-urceolate with 5 recurved teeth; drooping (Webb 1967), inclined or pendent (Fl. Eur. 3), bright pink at first, fading almost to white. It reaches 5–8 mm long (Warming 1908; Barvaux 1958; Anderson 1961; Fl. Br. Isl.; Hegi Fl. 1, ed. 3; Rameau et al. 1993) and it continues to grow during flowering (Warming 1908). The interior of the corolla is covered with erect, usually somewhat upturned, hairs that extend to the margin of the limb (Warming 1908). The stigma lies just at or immediately within the throat of the corolla, and the 10 anthers, which are at the level of the middle of the pistil and the corolla, bend inwards towards the style (Warming 1908; Barvaux 1958; Hegi Fl. 1, ed. 3; Rameau et al. 1993). The filaments are strongly swollen just above their slender short bases and are covered with fairly long hairs along the greater part of their length (Warming 1908; Anderson 1961). The anthers are covered with small protuberances; they open at the apex by two pores of irregular shape (Warming 1908; Hegi Fl. 1, ed. 3). The two apical appendages are as long as the anthers, glabrous and slightly curved (Warming 1908; Anderson 1961; Fl. Eur. 3). Pollen tetrad diameter ranges between 36 μm and 60 μm (Oldfield 1959; Paquereau 1959; Visset 1971; Hebda 1979; Warner & Chinnappa 1986). The mean number of tetrads per anther reaches 317 (range 277–428; Hebda 1979) or 403 (range 325–468; A.-L. Jacquemart, unpublished data). The stigma is not always distinctly lobed, but sometimes there are five small, rounded projections upon it covered by a thick mucilage that it secretes (Warming 1908). The nectar is secreted by swellings at the base of the ovary (Warming 1908; Fl. Br. Isl.). The superior ovary is 5-celled, with a mean of 46.0 ovules in Belgium (± 1.1, n = 10; A.-L. Jacquemart, unpublished data) or 47.2 in Sweden (Fröborg 1996). Baranec et al. (1996) counted a mean of 40 oval-shaped ovules, with 8 ovules per carpel. The fruit is an erect, loculicidal (Fl. Eur. 3), subglobose, 5-valved and glaucous capsule (Anderson 1961; Hegi Fl. 1, ed. 3). The capsule is 2.2–4.5 mm wide and 1.4–4.7 mm in height (Fl. Eur. 3). Seed number per capsule ranges between 1 and 44 (A.-L. Jacquemart, unpublished data). The seeds are oval, smooth (Fl. Br. Isl.), bright, elliptic, brown (Rameau et al. 1993), 0.4–0.8 × 0.7–1.5 mm (Hegi Fl. 1, ed. 3). In Belgium, the fresh mass of 100 seeds averages 21.97 ± 0.49 mg (n = 20) (A.-L. Jacquemart, unpublished data). Native, mainly on Sphagnum-rich ombrogenous bogs and rarely on other wet, acid habitats, mostly lowland (Webb 1967; Fl. Eur. 3; Fl. Br. Isl.). The distribution of A. polifolia in the British Isles is shown in Fig. 1. This species is in decrease (Stewart et al. 1994). It is now virtually absent from southern and much of central England as well as some parts of Scotland. It is extinct in Somerset, Norfolk and Hunts (Fl. Br. Isl.) but it has been re-introduced into Somerset (transplants from Cumbria). In Ireland, it occurs in the Central Plain, and extends to Antrim, East and West Mayo, Limerick and Wicklow (Fl. Br. Isl.; Webb (1967); Foss et al. 1984). Thus it has been reduced drastically in England and in Ireland, as shown in Fig. 1, where a considerable number of records are pre-1950 only (Birse 1980; Foss et al. 1984; Stewart et al. 1994; Stace 1997). The undisturbed raised mire habitat has suffered severely from drainage, afforestation, heather burning and peat digging in recent years, and A. polifolia has correspondingly become more local (Stewart et al. 1994). Often the extinction in West Europe results almost certainly as a consequence of drainage of bogs, particularly in association with peat extraction (Mennema et al. 1985; Rodwell 1991; B.D. Wheeler, personal communication). The distribution of Andromeda polifolia in the British Isles. Each dot represents at least one record in a 10-km square of the National Grid. (○) Pre-1950; (•) 1950 onwards. Mapped by the Biological Records Centre, Institute of Terrestrial Ecology, using Dr A. Morton's DMAP program, mainly from data collected by members of the Botanical Society of the British Isles. The species is present in all circumboreal regions (Anderson 1961; Hegi Fl. 1, ed. 3; Oberdorfer 1983; Hultén & Fries 1986; Billings 1988). It occurs in North and Central Europe from Scandinavia (not Iceland) to Normandy, the Auvergne, South Alps, northern Spain, northern Italy, North and South Germany, Central Switzerland (low places), East Carpathians and Central Russia (Fig. 2), mainly in the mountains in the southern part of its range (Fl. Br. Isl.; Hegi Fl. 1, ed. 3; Hultén 1971; Fitter 1978; Welter & Ruben Sutter 1982; Hultén & Fries 1986; Haeupler & Schönfelder 1989). According to Hegi (Fl. 1, ed. 3), it has disappeared from Spain, Croatia and Romania. It is considered as a rare species and it is protected in France (occurs only in the Vosges, Jura, Savoie, Massif Central, Pyrénées & Normandie; Rameau et al. 1993) and in the Netherlands (Mennema et al. 1985). It is also found in North Asia to Japan; Greenland; North America south to New York and Idaho (Fl. Br. Isl.; Fl. Eur. 3; Hultén & Fries 1986). The northern limit seems to lie a long way north of the Arctic Circle to about 71°7′ or even 74° (Hegi Fl. 1, ed. 3; Hultén & Fries 1986). The European distribution of Andromeda polifolia based on information in the literature. (£) Common; (!) less common. In the British Isles, A. polifolia occurs from sea level up to about 530 m in England and Wales (Ward's Stone and Bottom Head Fell), 735 m on Mount Keen and about 460 m in Ireland (Alt. range Br. Pl.; Stewart et al. 1994). In continental Europe, it may be found up to 1700–2000 m (Hegi Fl. 1, ed. 3; Rameau et al. 1993). This species is considered as boreo-alpine (Barvaux 1958), circumboreal (Rameau et al. 1993) and circumpolar boreal-montane (Preston & Hill 1997). It is circumpolar in northern areas, although its range extends to Central Europe and it is absent from the oceanic Atlantic coast; it occurs chiefly in the arctic and taiga biomes (Lindholm 1982; Hultén & Fries 1986). In Britain, Rodwell (1991) gives A. polifolia at mostly low altitude, between 10 and 530 m. These areas have between 800 and 1200 mm annual precipitation, with around 140–180 wet days year−1 (Rodwell 1991; Schaminée et al. 1995). It occurs at greatest shoot frequency in low hummocks and lawns of ombrotrophic bogs (Lindholm & Vasander 1981). It is successful at the alpine-montane level in oceanic and suboceanic regions of mainland Europe (Rameau et al. 1993). Andromeda polifolia is found on acid soil types. The humus type is peat, dysmoder or mor; all the soils are poor in bases and in nutrients (Small 1972; Rameau et al. 1993). It occurs on virgin surfaces where there is consistent waterlogging, with a high water-table, which favours the luxuriant growth of peat-building Sphagna (Small 1972; Rameau et al. 1993). These sites are uniformly acidic at the surface, with a pH between 3.8 and 4.3 in UK and West Europe (Rodwell 1991; Schaminée et al. 1995) or 4.6–5.0 in Scandinavia (Malmer & Nihlgaård 1980; Heikkilä 1987). Andromeda polifolia is not especially a hygrophilous species; it can grow well in quite dry soils and can persist in drained bog vegetation long after Sphagnum species have disappeared. Apart from the fact that drainage is usually associated with some other, more detrimental, activity (e.g. peat removal, afforestation), thus greatly augmenting its impact, when drying is detrimental to Andromeda this has probably more to do with species interactions than any direct influence of low water levels (B. D. Wheeler, personal communication). It grows thus on elevated Sphagnum hummocks. Where the drier vegetation has little Sphagnum (when it is strongly dominated by Calluna), A. polifolia is often scarce or absent, although when it does occur individual plants are often considerably more robust than in the case of Sphagnum lawns. The habitat of A. polifolia in Britain and Ireland is best described as oligotrophic with low pH (3.1–5.0), few exchangeable cations and little available nitrogen and phosphorus (Small 1972; Malmer & Nihlgaård 1980; Malmer & Wallén 1986; Heikkilä 1987; Cullings 1996). This species is considered as hygrophilous–acidophilous by Rameau et al. (1993). The chief habitat of the plant is the main dome of raised bogs over much of North West Europe. It extends into wet hollow communities, wooded bogs and occasionally blanket bogs. The following treatment of its phytosociological behaviour in Britain is based on the National Vegetation Classification (Rodwell 1991). Andromeda polifolia occurs typically in the Erica tetralix–Sphagnum papillosum mire (M18). This mire community is generally dominated by Sphagna, with ericoid subshrubs and monocotyledons often playing a more or less important role, Calluna vulgaris, Erica tetralix and Eriophorum vaginatum being the most common species. Andromeda polifolia and Vaccinium oxycoccos are distinctive of the Sphagnum magellanicum–Andromeda polifolia subcommunity (M18a); these two continental northern species have their British distributions largely centred in the regions where this community is found. This community is characteristic of waterlogged ombrogenous peats, typically at low altitudes in the moderately oceanic parts of Britain (lowlands of Wales, north-west England and Scotland, with less frequent occurrence in blanket bog at higher altitudes (> 300 m) and with a few outlying sites in lowland eastern England). Greatest frequencies of A. polifolia are usually associated with a Sphagnum-rich carpet, and one that usually contains a variety of Sphagnum species, often including Sphagnum magellanicum (Tüxen 1980; Doyle & Moore 1980; Rodwell 1991). Andromeda polifolia also occurs in the Sphagnum cuspidatum/recurvum bog pool community (M2), especially in the Rhynchospora alba–Sphagnum cuspidatum subcommunity (M2a) but not in the wettest locations. Andromeda polifolia can sometimes also be found in the Sphagnum recurvum subcommunity (M2b), with Erica tetralix, Eriophorum angustifolium, Vaccinium oxycoccos and Drosera rotundifolia. Nevertheless, Rhynchospora is absent and Andromeda is much less abundant in this second subcommunity. The community is dominated by extensive soft wet carpets of Sphagnum recurvum; vascular plants typically occur as scattered individuals of low total cover. This community is found in pools and hollows on very wet and base-poor ombrogenous mires in the less oceanic parts of Britain (Wales, the Borders, south-west Scotland and some localities in north-east Scotland). Andromeda polifolia does extend into these Rhynchosporion communities but becomes increasingly sparse with increased wetness. In Ireland, the species is fairly common in ombrotrophic peatland but is of restricted distribution in Atlantic blanket bog (Doyle & Moore 1980; Doyle 1990; Schouten 1990). Braun-Blanquet & Tüxen (1950) reported the species only in the subassociation with A. polifolia of the Pleurozia purpurea–Erica tetralix (Lid 1929) Br.-Bl. & Tx 1950 association (in the Oxycocco-Sphagnetea class). It is also worth noting that in Britain and parts of mainland Europe (as opposed to Fennoscandia), A. polifolia is almost exclusively a plant of ombrogenous peatlands. In Britain, it is more closely confined to these than any other vascular plant. There are a few sites in which it also grows in weakly minerotrophic conditions, but these are usually alongside ditches cut into an ombrotrophic surface, or instances where a once Sphagnum-rich ombrotrophic surface has been subject to recent, mild enrichment (B. D. Wheeler, personal communication). Outside Britain, A. polifolia is characteristic of Sphagneta with Vaccinium oxycoccos, Carex spp. and Vaccinium uliginosum, as well as Drosera rotundifolia, Ledum palustre, Betula nana and Calluna vulgaris (Hegi Fl. 1, ed. 3). In Fennoscandia, the species is rather common in poor fens and in transitional rich fens (Eurola 1962; Dierssen & Dierssen 1980). For example, in Finland, this oligotrophic species occurs in meso-eutrophic mud-bottom flark fens and Calliergon stramineum flark fens. It also appears at lower frequency in mesotrophic mud-bottom flark fens, eutrophic Sphagnum obtusum fens, meso-eutrophic Molinia fens and eutrophic pine and spruce swamps (Heikkilä 1987). In Swedish mires, A. polifolia occurs on hummock-lawns (Kvillner & Sonesson 1980) with Eriophorum vaginatum, Empetrum hermaphroditum and principally mosses such as Sphagnum balticum. It is present in the Empetrum hermaphroditum–Vaccinium microcarpum (= Empetro-Sphagnetum fusci Du Rietz 21 emend. Dierssen 82) association, mainly in three variants: the Cetraria nivalis–Dicranum elongatum variant (100% frequency), the Dicranum elongatum–Sphagnum balticum variant (92% frequency) and the Eriophorum vaginatum–Sphagnum russowii variant (100% frequency) (Sonesson & Kvillner 1980). It also occurs in the Carex rotundata–Drepanocladus schulzei association (in all variants with a frequency of 17–86%) (Sonesson & Kvillner 1980). In general, in western Europe, A. polifolia occurs in two phytosociological classes: the Oxycocco-Sphagnetea Br.-Bl. & Tx. 43 and the Scheuchzerietea DenHeld, Barkman & Westhoff 69 (Oberdorfer 1957, 1983; Tüxen 1980; Dierssen 1980; Schaminée et al. 1995). It occurs principally in the Oxycocco-Sphagnetea, which is typical of ombrogenous raised peat bogs. The species occurs in the Sphagnetalia-magellanici (Pawlowski 28) Kästn. & Flösn. 33, in the Oxycocco-Ericion (Nordh. 36) Tx. 37 in two associations: the Erico-Sphagnetum magellanici Osvald 23 and the Sphagno palustris-Ericetum Meltzer 45. It is also found in the Erico-Sphagnetalia Schwick. 40 emend. Br.-Bl. 49, in wet heaths (= Ericion tetralicis Schwick. 1933) in one association, the Ericetum tetralicis Allorge 22 (Oberdorfer 1957, 1983; Lecointe & Provost 1977; Schaminée et al. 1995). In the second class – represented in bog pools and poor fens, including floating mats – it is present in two alliances, the Rhynchosporion albae Koch 26 within three associations (the Sphagnetum cuspidato-obesi Tx. & vs. Hübschm. 58, the Sphagno-Rhynchosporetum Osvald 23 and the Caricetum limosae Osvald 23), and the Caricion lasiocarpae Vanden Berghen 49 in the Eriophoro-Caricetum lasiocarpae Osvald 23 (Oberdorfer 1957, 1983; Schaminée et al. 1995). In Germany as well as in the Netherlands, Oberdorfer (1957, 1983) noted its occurrence in the Pino-Ledion Tx. 55 and in the Piceon-Pinion uncinatea Tx. 55. The annual height increment of the moss layer is usually in the range 1–5 cm year−1 on hummocks and more than 10 cm year−1 in pools or wet depressions (Lindholm 1982; Malmer et al. 1994), and most of the dry matter in A. polifolia is beneath the moss layer (Lindholm 1982). Such growth rates often equal the annual shoot growth rate of the dwarf shrub species co-occurring with the Sphagnum mosses (Malmer et al. 1994). Owing to the continuous upward growth of Sphagnum, the vascular plants growing in bogs need to put resources into vertical growth to keep pace with the Sphagnum carpet (Malmer et al. 1994; Svensson 1995). A minimal yearly average increment in Sphagnum fuscum of 2 mm (Sonesson 1972 cited by Wallén 1986: probably a minimum value on the Swedish site studied) implies a yearly embedding of at least 4 g m−2 of wood of A. polifolia (Wallén 1986). For a steady-state above-ground biomass, this implies a minimum annual embedding of about 20% of the wood above the moss surface (Wallén 1986). The influence of the vascular plants on the growth and development of Sphagnum is considered to be of less importance than the reverse (Malmer et al. 1994). Both the vascular plants and Sphagnum mosses must be regarded as highly stress-tolerant, with a highly conservative strategy with respect to the acquisition and use of mineral nutrients (Malmer & Wallén 1986; Stewart & Nilsen 1995). As the two groups of plants are dependent on different sources for this resource, it cannot be shown that they compete directly for it (Malmer et al. 1994). On the other hand, the presence of Sphagnum obviously reduces the supply of mineral nutrients to the vascular plants and, in that sense, Sphagnum competes with the other plants by capturing a limiting resource before the vascular plants can do so (Malmer et al. 1994). Andromeda polifolia can recolonize peat workings on ombrogenous bogs: it rapidly establishes on drier soil but is often only a late recolonist of wetter, revegetating workings (Smart et al. 1989). There were 1220 ± 209 live tillers m−2 (n = 42 quadrats) in a Swedish mire, with a mean of 1.21 ± 0.06 shoots per individual tiller (Flower-Ellis 1980). The tillers are sparsely branched and each has approximately eight (7.83) leaves distributed between four leaf generations (Flower-Ellis 1980). There were 3.71 ± 0.28 current-year leaves per shoot, 1.22 ± 0.04 1-year leaves, 0.73 ± 0.08 2-year leaves and only 0.12 3-year leaves (Flower-Ellis 1980). The mean number of flowers or capsules is very variable; not all tillers flower and flowering individuals have several flowers, exceptionally 20–30. Abundant flowering is associated with reduced current shoot production, consequently with reduced total leaf number and area (Flower-Ellis 1980). Tillers with numerous flowers therefore have both a smaller assimilating area, relatively older leaves and a larger sink for assimilate above-ground, than purely vegetative tillers of equivalent age and size (Flower-Ellis 1980). The rhizome must be assumed to be common to both flowering and non-flowering tillers (Flower-Ellis 1980). The leaves remain green on the stems throughout at least one winter (Warming 1908) and 4 years at the most (Flower-Ellis 1973, 1980). Not more than about 5% of the current-year leaves appear to die in the interval between their appearance and that of the next year's leaves (Flower-Ellis 1980). By the end of their second growing season, a mean of 59% of those initially present have fallen. By the end of the third season, 62% of the original number have fallen. By the end of the fourth season, 96% have fallen. Some older leaves (more than 3 years) may persist until the end of the fourth growing season on the elevated areas in northern Sweden, but are exceedingly uncommon (Flower-Ellis 1980). The leaves of plants growing in hummocks are larger (72.9 cm2 vs. 59.4 cm2) and heavier (14.7 mg vs. 10.0 mg fresh mass) than those from the hollows (Simonis & Hirsch 1962). Flower-Ellis (1980) showed the same trend: leaves from hummocks are longer (7–13 mm vs. 7–11.5 mm) and the capsules are heavier (5–16 mg vs. 5–13 mg) than those of plants in hollows. Moreover, current-year leaves contain about 0.51% fresh weight nitrogen in hollows and 0.65% in hummocks (Simonis & Hirsch 1962). The same higher proportion for plants in hummocks was also detected for 1-year leaves (0.49% and 0.58% fresh weight N, respectively) (Simonis & Hirsch 1962). The mean dry weight of the aerial biomass is about 10 g m−2 in southern Finland (Lindholm & Vasander 1981) and about 28 g m−2 (Flower-Ellis 1973, 1980) or averaging 17 ± 2 g m−2 (Wallén 1986) in northern Sweden. Both the number of branches and the number of flowers vary with age and with microsite (Flower-Ellis 1980). The intermediate slopes of the elevated areas, and hummocks in small depressions, have both the most abundantly branched tillers and the largest numbers of flowers (Flower-Ellis 1980; A.-L. Jacquemart, unpublished data). Andromeda polifolia suffers more from frost on more open communities (Lindholm & Vasander 1981). Plants on the high hummock community withstand frost damage better than those of open communities (Lindholm & Vasander 1981). There is also a clear difference between the morphology of the tillers of the individuals of the high hummock community and of the other more open habitats. The tillers of the former habitat were upright and long compared with the tillers of the other habitats, which were stunted and low (Lindholm & Vasander 1981; Lindholm 1982). Fröborg (1996) estimated the fruit set in central Sweden to be about 65%. Flower-Ellis (1980) considered that most of the flowers in a northern Swedish mire develop into capsules, although not all capsules ripen. The limit of the potential resistance of ericaceous dwarf shrubs to low temperatures during the growing season is generally about –5 °C (Lindholm 1982). Many arctic and boreal plants are tolerant of low temperatures during the dormant phase, but during the active phase they are sensitive even to brief periods of low temperatures (Lindholm 1982). Frost caused considerable loss of production in destroying current shoot tips and flowers in southern Finland (Lindholm 1982). Since the flowers are self-compatible and easily self-pollinated, the seed production is primarily reduced by spring frosts but also perhaps by the lack of pollinators (Lindholm 1982). The size of the yield depends on the number of flowers. As flowers are formed in the year preceding fruiting, the yield is also dependent on the weather conditions in the previous year (Warming 1908; Flower-Ellis 1973, 1980). The coriaceous leaves, their silvery-glaucous lower surface and their revolute margins could indicate frost and drought resistance (Barvaux 1958). The slender, far-reaching subterranean rhizomes have long internodes, bear scale-leaves, and have a straight apex. The roots produced by the rhizomes are very slender and abundantly branched; they arise from the axils of the leaves (Warming 1908). The year's shoots have scale-leaves at their base. The axillary buds are partly protected by the erect petioles. The more vigorous of the lateral shoots arise from the axils of the uppermost leaves, i.e. those below the terminal inflorescence, and the uppermost shoot often forms a sympodium together with the parent stem (Warming 1908). On average, the below-ground biomass in mire ecosystems is known to be considerably higher than that above-ground (ratio 1:3–1:6; Flower-Ellis 1980; Wallén 1986; Malmer & Wallén 1986; Backéus 1990). Roots of A. polifolia may reach a depth of 45 cm in northern Sweden, with a ratio closer to 1:6 than to 1:3 (Flower-Ellis 1980). Andromeda polifolia has even more of its biomass below-ground: it reaches 75% (Malmer & Nihlgaård 1980; Malmer & Wallén 1986) and Wallén (1986) reports up to 98%. The erect-growing above-ground shoots develop from the terminal end of sparingly branched, vertically orientated rhizomes. Rapid shoot growth soon after the beginning of the growth period (May) was observed (Lindholm 1982). The mean above-ground biomass of 28 g m−2 would give a root and rhizome biomass of 168 g m−2 to a depth of 45 cm assuming a 1:6 ratio (Flower-Ellis 1980). The plant shows no measurable fine root activity below 10–15 cm. At the 2.5–5 cm level as much as 50% of total dry mass was vascular plant material, most of which was of fine root mass (up to 370 g m−2) (Wallén 1986). The fine roots thus formed 24 ± 4% of the total below-ground biomass (Wallén 1986). The mean length of leaves of all four generations was 8.53 ± 0.14 mm (Flower-Ellis 1980). There were in all 7200 ± 1844 leaves m−2 (Flower-Ellis 1980). The leaf area index (LAI) was 0.0855 ± 0.0244 in mid-season in northern Sweden (Flower-Ellis 1980). With 1008 live tillers m−2, this would represent 0.85 cm2 per tiller, shared mainly between leaves from 2 years (older leaves being few in number) (Flower-Ellis 1980). Fresh weight per leaf is 10.0–14.7 mg with a leaf area of 59.4–72.9 cm2 m−2 (Simonis & Hirsch 1962). Dry weight reaches 2–4 mg for current-year leaves and 3–4.5 mg for older leaves, with a leaf area ranging from 0.25 to 138.5 cm2 m−2 (Flower-Ellis 1980). LAI at the end of the growing season is as follows: current-year leaves 0.0444, 1-year leaves 0.0180, 2-year leaves 0.0020, assuming a mean leaf length of 8.53 mm with corresponding projected area as low as 11.9 mm2 in northern Sweden (Flower-Ellis 1980). The water content of the leaves reaches 54.1% (Tamas et al. 1975); values given by Simonis & Hirsch (1962) are 55.0–57.2%. The 2-year leaves are both less active photosynthetically and heavier than the younger leaves; the weight difference is most probably the result of the storage of reserves, possibly also of structural and other changes (Flower-Ellis 1980). One-year leaves from flowering tillers were lighter (at least 7%) than those from non-flowering tillers (Flower-Ellis 1980). The leaves are devoid of trichomes, but are covered with thick epicuticular waxes (Warming 1908). Stomata occur only on the lower surface (543.2 ± 24.6 mm−2; n = 12) of the leaves (A.-L. Jacquemart, unpublished data). Most Ericaceae are characterized by the presence of ericoid mycorrhizal infection (Cullings 1996), as recorded in A. polifolia (Harley & Harley 1987). The fungus normally involved in the infections is the ascomycete Hymenoscyphus ericae (Read) Korf & Kernan or its anamorphs (Pearson & Read 1973a; Harley & Harley 1987; Read & Kerley 1995). There are numerous, largely unsubstantiated, claims of the involvement of other fungi in the formation of ericoid mycorrhiza. Members of the genus Bisoascus and its anamorphs Oidiodendron are most frequently reported to be so involved, and there is evidence to support the view that they may be mycorrhizal under some circumstances (Read & Kerley 1995). Two fundamental roles of the mycorrhiza can be recognized, one involving detoxification of the soil environment, the other acquisition of essential nutrients, primarily nitrogenous (Pearson & Read 1973b; Kerley & Read 1995; Read & Kerley 1995; Cullings 1996) and phosphatic compounds (Harley & Harley 1987). Chamaephyte. The aerial shoots branch and the buds over-winter. Andromeda polifolia is under-represented in the seed banks of soils wherever it is common. Extremely rare germinants were observed in England (J.P. Grime, personal communication) and in Belgium. The basic chromosome number in Andromeda species is 2n = 48 (Fl. Eur. 3; Fl. Br. Isl.; Oberdorfer (1983); Uotila & Pellinen 1985). Andromeda polifolia showed positive photosynthesis from immediately after snowmelt in May, until mid-September in Sweden (Johansson & Linder 1975). The current-year leaves developed during late June, and from the middle of July their photosynthetic rates exceeded those of the 1-year-old leaves, both on a leaf area and on a dry weight basis (Johansson & Linder 1975). The photosynthetic assimilation has been calculated to be 0.6–0.8 mg CO2 h−1 g−1 dw for the 1-year-old leaves at the end of the growing season (1.0–3.6 mg CO2 h−1 g−1 dw for the current-year leaves) and 2.6–3.4 mg CO2 h−1 g−1 dw and 0.7–2.9 mg CO2 h−1 g−1 dw, respectively, for old and current-year leaves at the beginning of the season (Johansson & Linder 1975). The seasonal photosynthetic assimilation of evergreen leaves has been calculated to be 4–10 g CO2 (g leaves)−1 year−1 (Karlsson 1986). The seasonal carbon gain could thus be estimated to 0.06–0.14 g CO2 (g biomass)−1 (Karlsson 1986). Starch varied between 4.5 and 9.5% of dry weight and total carbohydrates between 9.1 and 14.9% in 1-year-old leaves (Flower-Ellis 1980; in the same Swedish site). Starch maxima occurred at noon, indicating that there was some accumulation in the leaves, mainly during the first half of the day (Flower-Ellis 1980). Cumulative diurnal dry weight equivalents of CO2 fixation reached 0.126 mg standard leaf−1 day−1 for current-year leaves, 0.132 for 1-year leaves and 0.106 for 2-year leaves (Flower-Ellis 1980). There was a simultaneous increase in the starch content of the 1-year leaves (0.25 mg leaf−1, 174 mg dm−2) and a small increase in the sugars, together (0.26 mg leaf−1, 180 mg dm−2) commensurate with the dry weight equivalent of assimilation during the period (0.25 mg leaf−1, 172 mg dm−2) (Flower-Ellis 1980). Thus 49% of the change in weight was accounted for by change in carbohydrate content. Similarly, variation in starch alone accounted for about 36% of the variation in dry weight (Flower-Ellis 1980). Therefore dry weight change and changes in starch and carbohydrates are associated, but the greater part of the weight variation cannot be accounted for by changes in starch and total carbohydrates. The residual leaf weight continued to show variation with time, over a range of 0.68 mg leaf−1 (472 mg dm−2). This difference should be attributable to translocation. This translocation from the leaf might be estimated as 0.44 mg leaf−1 (305 mg dm−2). During the growing season, current-year leaves gained 0.12–0.15 g m−2, 1-year leaves lost 0.73–0.97, 2-year leaves lost 0.14–0.29 (Flower-Ellis 1980