METSÄNTUTKIMUSLAITOKSEN TIEDONANTOJA 905, 2003 FINNISH FOREST RESEARCH INSTITUTE, RESEARCH PAPERS 905, 2003 Response of the understorey vegetation of boreal forests to heavy metal loading Maija Salemaa VANTAAN TUTKIMUSKESKUS - VANTAA RESEARCH CENTRE METSÄNTUTKIMUSLAITOKSEN TIEDONANTOJA 905, 2003 THE FINNISH FOREST RESEARCH INSTITUTE, RESEARCH PAPERS 905,2003 Response of the understorey vegetation of boreal forests to heavy metal loading Maija Salemaa Vantaa Research Centre, Finnish Forest Research Institute Academic dissertation in Terrestrial Plant Ecology Faculty of Biosciences University of Helsinki To be presented, with the permission of the Faculty of Science of the University of Helsinki, for public criticism in the auditorium of Arppeanum (Helsinki University Museum, Snellmaninkatu 3) on February 20 th , 2004, at 12 o'clock noon. Helsinki, 2003 Supervisors: Docent Heljä-Sisko Helmisaari Vantaa Research Center Finnish Forest Research Institute Docent John Derome Rovaniemi Research Station Finnish Forest Research Institute Reviewers: Docent Pasi Rautio Department of Biology University of Oulu Docent Timo Vuorisalo Department of Biology University of Turku Opponent Docent Kari Laine Department of Biology University of Oulu Publisher: Finnish Forest Research Institute, Vantaa Research Center, P.O. Box 18, FIN-01301 Vantaa, Finland Accepted by Jari Hynynen, Research Director, 26 November 2003 Front cover: Cuttings of bearberry (Arctostaphylos uva-ursi) exposed amounts of Cu (paper VI). Photos: Erkki Oksanen /METLA to different ISBN 952-10-1611-6 (PDF version) http://ethesis.helsinki.fi/ Helsinki 2003 Helsingin yliopiston verkkojulkaisut ISBN 951-40-1899-0 (printed version) ISSN 0358-4283 Helsinki 2003 Vammalan Kirjapaino Oy 2003 Contents List of original publications 4 Abstract 5 Acknowledgements 7 1. Introduction 9 1.1. What are heavy metals? 9 1.2. Sources and forms of heavy metals in soil-plant systems in boreal forests 9 1.3. Mechanisms of heavy metal uptake in cryptogams and vascular plants 11 1.4. Plant strategies for resisting and tolerating heavy metals 12 1.5. Aims of the study 13 2. Study areas and pollutant emissions 15 2.1. Experimental sites along the Harjavalta pollution gradient 15 2.2. Emissions from the Harjavalta smelter 15 2.3. Ecological research in the Harjavalta area 15 2.4. Background areas 16 3. Materials and methods 18 3.1. Vegetation surveys (I, II) and seed bank composition (III) 18 3.2. Chemical analyses of plants (11, IV, VI), soil (I -V) and precipitation (11, IV) 18 3.3. Compensatory growth (IV) and branching pattern (V) of dwarf shrubs in the field 20 3.4. Electrophoretic analysis of isoenzyme variation in the dwarf shrub populations 21 3.5. Experimental exposure of Arctostaphylos uva-ursi and other dwarf shrubs to Cu (VI) 21 3.6. Statistical analysis 22 4. Results 23 4.1. Heavy metal accumulation and the distribution of species (I, II) 23 4.2. Seed bank composition in relation to the existing vegetation (I, III) 25 4.3. Seedling survival in forest soil polluted with heavy metals (III) 26 4.4. Genetic structure of the dwarf shrub populations 26 4.5. Compensatory growth and branching pattern of the dwarf shrubs (IV, V) 28 4.6. Cu resistance of Arctostaphylos uva-ursi compared to the other dwarf shrubs (VI) 30 5. Discussion 3 1 5.1. Heavy metal accumulation and the relative resistance of the species 31 5.2. Revegetation potential of the seed bank 32 5.3. Sensitivity of Arctostaphylos uva-ursi to Cu compared to the other dwarf shrubs 32 5.4. Clonal diversity and genetic variation of the dwarf shrubs 32 5.5. Phenotypic plasticity as a resistance strategy in clonal dwarf shrubs 33 5.6. Applicability of the results in biomonitoring and phytoremediation 34 6. Conclusions 36 References 37 List of original publications The thesis is based on the following publications, which are referred to in the text by Roman numerals. All the publications are reproduced with the publishers' permission. I Salemaa, M., Vanha-Majamaa, I. & Derome, J. 2001. Understorey vegetation along a heavy-metal pollution gradient in SW Finland. Environmental Pollution 112: 339-350. II Salemaa, M., Derome, J., Helmisaari, H.-S., Nieminen, T. & Vanha-Majamaa, I. 2004. Element accumulation in boreal bryophytes, lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. The Science of the Total Environment (in press). 11l Salemaa, M. & Uotila, T. 2001. Seed bank composition and seedling survival in forest soil polluted with heavy metals. Basic and Applied Ecology 2: 251-263. IV Salemaa, M., Vanha-Majamaa, I. & Gardner, P. 1999. Compensatory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal polluted environment. Plant Ecology 141: 79-91. V Salemaa, M. & Sievänen, R. 2002. The effect of apical dominance on the branching architecture of Arctostaphylos uva-ursi in four contrasting environments. Flora 197: 429-442. VI Salemaa, M. & Monni, S. 2003. Copper resistance of the evergreen dwarf shrub Arctostaphylos uva-ursi: an experimental exposure. Environmental Pollution 126: 435-443. Studies I and II are based on initial suggestions provided by prof. Eino Mälkönen and Dr. Heljä-Sisko Helmisaari. Dr. John Derome is responsible for the soil and deposition data in all the papers. Field work and data processing were planned together with all co authors. Maija Salemaa is responsible for data handling and writing of the first draft of papers I and 11, and for the idea, data handling and writing the first draft of papers 111 - VI. Dr. Risto Sievänen developed the simulation model for the branching architecture of A. uva-ursi (V). Anu Blom, M. Sc., performed the electrophoresis analyses of A. uva ursi and V. uliginosum and interpreted the zymograms. Abstract The detrimental effects of heavy metals on boreal forest vegetation are most evident in the surroundings of metal smelters and mines. The structure of the understorey vegetation and soil seed banks, chemical composition of plant species and branching architecture of two dwarf shrub species (Arctostaphylos uva-ursi and Vaccinium uliginosum) were studied in Scots pine forests along an 8 km transect running SE from the Cu-Ni smelter at Harjavalta, SW Finland, during 1992 - 2000. The general aim of the study was to compare the responses of bryophytes, lichens and vascular plants to pollution, and to evaluate the possibility of re-establishing the native understorey vegetation of a damaged forest. Clonal diversity and isoenzyme variation of the two dwarf shrub species were analysed in order to investigate whether the populations in the polluted sites near the smelter were genetically differentiated from those growing at further distances away. In addition to the field studies, the effect of Cu on A. uva-ursi was investigated in an experimental exposure and the results were compared to earlier experiments carried out using other local dwarf shrub species. Over 50 years' accumulation of heavy metals (mainly Fe, Cu, Ni, Zn and Pb) and sulphur in the forest ecosystem near the smelter has drastically changed the plant communities. Vegetation was almost absent up to a distance of 0.5 km from the smelter. The total coverage and the number of plant species increased with increasing distance from the smelter. Vascular plants, being capable of restricting the uptake of toxic elements, grew closer to the smelter than lichens and bryophytes, which accumulated larger amounts of heavy metals. In general, vascular plants were more pollution-resistant than lichens, whereas bryophytes were the most sensitive group. A pioneer moss species (Pohlia nutans) was an exception to this general pattern, because it accumulated considerably higher amounts of heavy metals than the other species and still survived close to the smelter. Viable seeds were found in the soil at all the studied distances from the smelter. Although the vegetation was very scanty at 0.5 km, the soil contained germinable seeds of local plant species. At the present time, young seedlings rapidly die in the contaminated soil, which restricts the natural recolonization of the sites. The clonal diversity of A. uva-ursi and V. uliginosum was relatively high in both the polluted and clean sites. Further, the genotype frequencies of either species did not differ significantly between the sites, indicating that the populations have not differentiated genetically. It is suggested that the failure of seedling establishment partially prevents the evolution of heavy metal-tolerant ecotypes in the most polluted areas. The surviving clones of dwarf shrubs were tens of years old and may represent the most resistant genotypes of the populations derived from the time before the smelter started operating in the 1940'5. A. uva-ursi and V. uliginosum showed high regrowth after autumn clipping of current year shoots (imitating pollution-induced shoot damage) in a field experiment carried out in polluted and clean sites. In contrast, spring clipping of new shoots was extremely detrimental, especially for the evergreen A. uva-ursi, which had smaller carbohydrate and nutrient reserves than the deciduous V. uliginosum. In an experimental exposure to Cu, A. uva-ursi proved to be more sensitive than the other dwarf shrubs (Calluna vulgaris, Empetrum nigrum, V. uliginosum). However, A. uva-ursi showed high plasticity in branching when studied in four habitats with varying pollution, nutrient, light and competition levels. Further, the adult clones avoided heavy metals by extending their roots into the less toxic, deeper soil layers. It is concluded that phenotypic plasticity increases the survival of this species in contaminated sites and enables it to respond to changed resource levels according to the "reserve meristem hypothesis". The results of this thesis show that heavy metal and sulphur deposition have subjected the understorey vegetation growing in the vicinity of the Harjavalta smelter to a strong selection pressure, which is reflected as a changed species composition and disappearance of sensitive species. In addition to heavy metals, nutrient imbalances, reduced water-holding capacity of the surface soil and the accumulation of large amounts of undecomposed, dry needle litter also restrict plant recolonisation on the degraded sites. However, the considerable reduction of emissions during the last decade has opened up new possibilities for the restoration of the forest ecosystems. The recovery of the existing vegetation and realisation of the revegetation potential of local seed banks can be promoted by soil mitigation, facilitating the formation of a new organic soil layer and the planting of tolerant plant species. Keywords: bryophytes, copper, dwarf shrubs, ecotypes, heavy metals, lichens, nickel, resistance, phenotypic plasticity, tolerance 7 Acknowledgements This thesis was carried out at the Vantaa Research Centre of the Finnish Forest Research Institute. I express my warmest thanks to my supervisors Dr. Heljä-Sisko Helmisaari (leader of the project "Recovery of boreal forest ecosystem from long-term heavy metal pollution") and Dr. John Derome for their scientific guidance, support and help throughout the study. Heljä-Sisko introduced me to the field of nutrient fluxes of forest ecosystems, and John to chemistry and processes in forest soil. I am also deeply indebted to the other co-authors Peter Gardner, Satu Lyyra (nee Monni), Tiina Nieminen, Risto Sievänen, Taru Uotila and Ilkka Vanha-Majamaa for the innovative and enjoyable cooperation. The other members of the "Harjavalta Group", Hannu Fritze, Oili Kiikkilä, Hannu Nousiainen, Antti Reinikainen and Christian Uhlig, have contributed to this thesis by helping in different phases of the work. I am grateful to Anu Blom for performing the isoenzyme analyses, Leila Korpela for participating in the vegetation inventories, Jari Perttunen for making the branching simulations, and Arja Ojala and Marja Vieno for advice and practical help in sampling soil seed banks. I would like to thank all of you for the stimulating discussions, valuable comments, encouragement and friendly atmosphere. The opportunity to share your expertise in different fields of ecology has enormously widened my understanding about the functioning of Nature. I owe my sincere thanks to Prof. emer. Eino Mälkönen for the possibility to start my plant ecological research in the forest fertilisation experiments at Harjavalta. I am very grateful to Prof. Erkki Tomppo and my colleagues in the Tempos Project for the possibility to use the vegetation data of the national forest inventory as reference material in my thesis. I would like to thank Maija Jarva, the former director of the Central Laboratory of Metla, and Kerttu Nyberg, Pirkko Ronkainen, Maija Ruokolainen and Arja Tervahauta for performing the chemical analyses of the plant samples. I also thank Prof. Katri Kärkkäinen, Helena Korpelainen and Tiina Nieminen for reading and commenting on parts of the summary. I warmly thank the reviewers of the thesis, Dr. Pasi Rautio and Dr. Timo Vuorisalo, for their useful comments and constructive criticism. The encouragement of Rauno Ruuhijärvi, professor emeritus in plant ecology at the University of Helsinki, has been very important when I took my first steps along the path of science. The same positive attitude towards my studies has continued during the period of Professor Heikki Hänninen and Professor Sirkku Manninen, to whom I would like to express my sincere thanks for commenting on the manuscripts and for their speedy handling of the official papers concerning my dissertation. Many persons have provided practical assistance in the field, laboratory and office. My special thanks go to Sinikka Levula, Teuvo Levula and Pekka Suolahti for helping in the field work. Pia Reponen prepared the plant samples for chemical analyses, Esa Ek, Kaarina Pynnönen and Satu Smolander took care of the plants in the greenhouse, and Sampsa Lommi and Mikko Piirainen identified the lichen and plant samples. Hilkka Granlund and Airi Piira recorded the data, Hannu Nousiainen made the drawings, Markku Tamminen the maps, and Sari Elomaa the layouts of many of the figures. In addition to the persons mentioned above, I want to thank all the staff of the Vantaa Research Center, 8 especially the library and computer services, for their highly qualified work. John Derome has made linguistic corrections to the summary and all the papers. I wish to thank the Niemi Foundation for additional financial support. Outokumpu Harjavalta Metals Oy is acknowledged for the possibility to establish the sample plots and to collect samples in the smelter area. Finally, thanks to my family and friends for their support and care. The "Lunch Society" and the "Friday Swimming Club" have brought me considerable enjoyment over the years when I was in danger of being buried in heavy metal dust. 9 1. Introduction 1.1. What are heavy metals? The term "heavy metals" has received widespread usage for metals that are potentially toxic in high doses. The elements having a specific gravity > 4.5 g/cm 3 (sometimes defined as 5.0 or 6.0) are called heavy metals (Streit & Stumm 1993). Except for their specific gravity, heavy metals have no common chemical property or behaviour in biological or ecological systems. Therefore some authors prefer to classify the metallic group of elements into "A (oxygen-seeking), B (nitrogen/sulphur-seeking) and borderline metals", according to their ability to become a part of a chemical complex (Nieboer & Richardson 1980). Many authors use the term "trace metals" as a synonym for heavy metals. However, this term should perhaps be restricted to its original meaning, i.e. for metals required by plants in extremely small amounts, "trace" amounts. In this thesis I use the term "heavy metals" when referring to copper (Cu), nickel (Ni), iron (Fe), manganese (Mn), zinc (Zn), cadmium (Cd) and lead (Pb). Aluminium (Al), which has a specific gravity of 2.7 g/cm 3 , is a metal but, according to the definition applied here, is not a heavy metal. Of the heavy metals investigated in this thesis, Mn, Zn, Cu, Fe and Ni are essential micronutrients for plants, but are toxic at higher concentrations and disturb most of the primary physiological processes of plants (Marschner 1995). Cd and Pb are non-essential elements for plants, and may be toxic or lethal even when absorbed in small amounts. 1.2. Sources and forms of heavy metals in soil-plant systems in boreal forests Natural weathering of metalliferous rocks and anthropogenic sources provide the two major inputs of metals into ecosystems (Ross 1994 a). Natural sources of these elements in the atmosphere are volcanic activity, terrestrial dust, vegetation fires, salt spray from the oceans and biogenic sources (Nriagu 1989). Anthropogenic sources in the soil-plant systems include atmospheric deposition originating from a range of industrial activities (metalliferous mining and smelting, alloying plants, petrochemical industry, fertiliser plants, coal power plants, industrial and home furnaces), agricultural amendments and motor traffic (Alloway 1995 a). The amount of heavy metals emitted by natural processes into the atmosphere is small compared with the anthropogenic input of these elements (Ross 1994 a). Heavy metals are emitted from traffic and industrial sources into the atmosphere and spread over wide regions of the boreal forest zone. Atmospheric deposition of heavy metals has decreased considerably during the last 20 years in northern Europe (Ukonmaan aho et al. 1998), which is reflected as low concentrations in bioindicator bryophytes in background areas (Mäkinen 1994, Ruhling & Steinnes 1998, Buse etal. 2003, Poikolainen et al. 2004). However, there are many industrial areas where the long-term accumulation of heavy metals, often associated with exposure to sulphur dioxide, has damaged northern 10 forests ecosystems. For instance, there are extensive damage areas surrounding the metallurgical industry complexes on the Kola Peninsula, NW Russia (Tikkanen & Niemelä 1995, Chernenkova & Kuperman 1999, Rigina & Kozlov 2000). Heavy metal polluted areas are also to be found in the Nordic countries, primarily close to smelters, mines or steel mills (Folkeson & Andersson-Bringmark 1988, Kubin et al. 2000, Tammiranta 2000, Buse et al. 2003). In Finland, for instance, it has been estimated that there are about 800 contaminated industrial areas (also including non-metalliferous) and 35 mine tailing areas requiring restoration in the near future (Puolanne et al. 1994). The atmospheric deposition of heavy metals on ecosystems occurs mainly in particulate and aerosol forms (Ross 1994 a, Luttermann & Freedman 2000). The area affected by the deposition of heavy metal containing particles is usually much more local than that of gaseous pollutants. Rates of deposition, whether dry or wet, tend to be greatest near the pollution source. However, small particles are transported over longer distances, and have longer atmospheric residence times than larger ones (Hutchinson & Whitby 1977). The following factors may affect the fate of atmospheric metal deposition in the receiving plant - soil systems: 1) particle size, 2) solubility, 3) distance of the receiving system from the metal source, and 4) acidity of rainfall (Ross 1994 a). Heavy metal deposition shows a high affinity for adsorption to organic surfaces. In boreal forests, the canopies of coniferous trees effectively filter pollutant particles from the air (Tyler 1984, Zöttl 1985, Fowler et al. 1989). Stand throughfall and plant litter increase the load of heavy metals and sulphur on the forest floor (Heinrichs & Mayer 1980, Derome & Nieminen 1998, Nieminen et al. 1999). The humus layer in coniferous forests effectively retains heavy metals through adsorption and complexation with organic matter (Derome 2000b). The characteristics of the receiving soil (especially pH, oxidation reduction potential, the amount of particulate and soluble organic matter, clay minerals and concentrations of mineral nutrients) influence the bioavailability and toxicity of most metals (Luttermann & Freedman 2000). The phytotoxicity of heavy metals depends on their bioavailability, which is related to their occurrence in different chemical forms (Alva et al. 2000). Heavy metals in soils can exist in water soluble and exchangeable form, associated with insoluble organic matter, and as carbonates, oxides of Fe, Al and Mn, and layer silicates (Alloway 1995b). Generally, increasing acidity tends to increase the mobility and toxicity of heavy metals (Alloway 1995b). Organic and exchangeable forms of heavy metals are the major forms taken up by plant species (Alva et al. 2000). Although the above-ground biomass of the understorey vegetation of boreal forests is small in relation to that of the trees, it plays an important role in regulating the nutrient fluxes (Mälkönen 1974), hydrology and micro climate (Siren 1955, Päivänen 1966, Tolvanen & Kubin 1990). In contaminated environments, certain plant species can accumulate considerable amounts of heavy metals and protect the soil from erosion and the leaching of heavy metals into the groundwater (Vangronsveld et al. 1996). Because many heavy metals are bound on plant surfaces and tissues (Rautio & Huttunen 2003), and form stable complexes with organic matter in the soil (Alloway 1995b), they may still have a long-lasting effect on forest ecosystems after the emissions have ceased (Ross 1994 a). Before we can gain a better understanding of heavy metal fluxes in boreal 11 forests, more information is needed about the importance of the understorey vegetation in ecosystem processes. 1.3. Mechanisms of heavy metal uptake in cryptogams and vascular plants Vascular plants mainly take up elements via their roots from the soil, although the foliar uptake of gases and soluble elements may also be substantial (Marschner 1995). The uptake of heavy metals via aerial plant parts has been demonstrated in fir (Lin et al. 1995) and in many crop plants including wheat (Haslett et al. 2001). Generally, the thick epidermis and waxy cuticle of the leaves provide external protection against toxic elements in evergreen species. For instance, Monni et al. (2001b) did not find ecophysiological responses when heavy metal solutions were applied to the aboveground parts of Empetrum nigrum. In industrial areas with high airborne deposition, large amounts of metal containing dust become attached to the surface structures of the aerial parts, and particles may also become embedded in the cuticular waxes (Rautio et al. 1998, Kozlov et al. 2000, Rautio & Huttunen 2003). Heavy metals dissolved in soil water enter vascular plants via mass flow or diffusion into the free space of the root cortex (Marschner 1995). Only a small proportion of the heavy metals accumulated in roots passes through the endodermis and is subsequently distributed as organic complexes into the different plant organs via the xylem and phloem (Clemens et al. 2002). In contrary to vascular plants, cryptogams (bryophytes and lichens) have no real roots, epidermis or cuticle layer, and they absorb water and dissolved elements directly across their surface. Most of the bryophyte and lichen species obtain the majority of their water and nutrients from atmospheric deposition; some species also obtain nutrients from water that has been in contact with the substrate (Bates 1992, okland et al. 1999, Garty 2001). Lichens, which are symbiotic organisms comprising mycobiont and phytobiont partners, have many similarities with bryophytes in their element uptake. The following element fractions occur in both taxa: 1) trapped particulate matter, 2) intercellular soluble elements, 3) extracellular elements bound to the cell wall on charged exchange sites, and 4) intracellular elements (Tyler 1990, Nash 1996, Zechmeister et al. 2003). Both bryophytes and lichens (especially the mycobiont partner) have a high ion exchange capacity on their cell walls, and the dead tissues also have an ability to bind ions (Tyler 1989, 1990, Chettri et al. 1997). The toxic effects of heavy metals are manifested in a wide range of plant cellular activities including photosynthesis, respiration, mineral nutrition and membrane structure in all the plant groups (Tyler 1990, Marschner 1995, Garty 2001). Toxic concentrations have been found to cause membrane damage, ion leakage and decreased chlorophyll concentrations in vascular plants (Mocquot et al. 1996, Monni et al. 2001 a, Pätsikkä et al. 2001), as well as in bryophytes (Brown & Wells 1990, Guschina & Harwood 2002) and lichens (Chettri et al. 1998, Tarhanen et al. 1999, Hyvärinen et al. 2000). General responses of dwarf shrubs to elevated concentrations of heavy metals are leaf discoloration 12 and decreased growth of the shoots and roots (Monni et al. 2000a,b). Differences in root elongation rates in toxic solutions compared to control solutions are commonly used to determine the metal tolerance index of higher plants (Utriainen et al. 1997). 1.4. Plant strategies for resisting and tolerating heavy metals Resistance is a quantitative trait that enables a plant to survive, grow and reproduce in the presence of a particular pollutant (Baker & Walker 1989). Plant populations can become resistant to heavy metals through heritable adaptation (ecotypes), or individual plants can gradually acclimatise to an increasing heavy metal load (phenotypic plasticity) (Antonovics et al. 1971, Baker et al. 1986, Dickinson et al. 1991, Punshon & Dickinson 1997). The broadness of phenotypic plasticity is also genetically controlled (Bradshaw & Hardwick 1989, Thompson 1991). In some species, all the individuals show some degree of innative (constitutive) tolerance even though they are not exposed to heavy metals (Baker 1987). For instance, cuttings of Empetrum nigrum originating from an unpolluted area showed high survival when exposed to Cu and Ni, indicating consitutive tolerance (Monni et al. 2000 a). However, normally less than 0.1 % of the individuals in natural populations of plant species are resistant (MacNair 1987). If heavy metal concentrations in the soil increase, resistant individuals are favoured as a result of natural selection and their abundance increases. In short-lived plant species such as grasses and herbaceous species, the whole population can change to a resistant one within a few years (MacNair 1987). It has been demonstrated that pioneer bryophytes are also able to undergo rapid evolution in response to a heavy metal load in soil (Jules & Shaw 1994). The evolution of heavy metal resistant ecotypes is often considered to be the best example of evolutionary changes in plant populations (Bradshaw et al. 1990). Heritable changes take place at a slower rate in trees and dwarf shrubs owing to their longer generation times. Ecotypes are much rarer in these plants than among grasses and herbaceous species. The few known examples of metal-tolerant populations of trees occur among pioneer species, e.g. in the Betula and Salix species (Eltrop et al. 1991, Kahle 1993, Kopponen et al. 2001). High plasticity in growth and physiological characteristics, which moderates the impact of local stress, is common in clonal dwarf shrubs, e.g. in the family Ericaceae (Gimingham 1972, Shevtsova 1998). Phenotypic plasticity may also provide a mechanism that improves the survival of long-lived species in metal-contaminated environments (Dickinson et al. 1991, 1992, Turner & Dickinson 1993). The mechanisms involved in heavy metal resistance are species-specific and are usually divided into avoidance and tolerance mechanisms (Fig. 1) (Baker 1987, Verkleij & Schat 1989). Avoidance is expressed as external protection against toxic elements or as active orientation of the roots to less toxic soil (Tyler et al. 1989). The avoidance of heavy metals can also be facilitated by mycorrhizal fungi. Ericoid mycorrhizas of dwarf shrubs have the ability to accumulate large amounts of heavy metals, thereby restricting metal transport to the shoots (Bradley et al. 1981, Meharg & Cairney 2000). Heavy metal tolerant strains of mycorrhizal fungi have been found on host plants growing in heavy metal polluted sites (reviewed by Hartley et al. 1997). 13 Fig. 1. Summary of the possible mechanisms involved in the resistance to elevated heavy metal concentrations according to Baker (1987), Tyler et al. (1989) and Hall (2002). Although vascular plants have some degree of control over which elements are taken up by their roots, total avoidance of heavy metal uptake is not possible (Kahle 1993). Real tolerance is based on physiological mechanisms that result in the exclusion of heavy metal ions from important metabolic processes or which accumulate metals in detoxified forms (Baker 1987). Vascular plants have many species-specific mechanisms to restrict the cellular uptake of heavy metals and to detoxify them internally (Fig. 1). For instance, living plant cells can detoxify heavy metals by binding them in cell walls, chelating and storing them in vacuoles, or binding them with phytochelatines in the cytoplasm (reviewed by Hall 2002). Mosses and lichens, which absorb nutrients directly through their surfaces, cannot prevent ions penetrating into their tissues. A number of functional groups in bryophyte and lichen structures are capable of binding metal ions on the cell walls (Tyler 1990, Nash 1996, Onianwa 2001). Intracellular complexing of metals has been found to be based on e.g. organic acids (Sarret et al. 1998) or phytochelatins (Pawlik-Skowronska et al. 2002) in lichens, and on glutathione (GSH) synthesis (Bruns et al. 2001) in bryophytes. Bryophytes have the ability to translocate heavy metals e.g. to vacuoles (Bruns et al. 2001). 1.5. Aims of the study The general aim of this thesis is to compare the responses of bryophytes, lichens and vascular plants to heavy metal loading, and to evaluate the possibility of re-establishing the native understorey vegetation of a damaged forest in the vicinity of the Harjavalta Cu-Ni smelter. A significant decrease in emissions from the smelter, which has been operating since 1945, was achieved in the beginning of the 1990's (Section 2.2). Thus the year 1992, when I started these investigations, provides an interesting reference point to studies on the recovery of the vegetation. 14 The thesis is based on the field studies and manipulations carried out at different distances from the smelter (I - V) and on a controlled greenhouse experiment (VI). In addition, some unpublished results on the clonal diversity and isoenzyme variation in the populations of two dwarf shrub species in relation to the pollution level, are presented. This study is based on the hypothesis that the airborne deposition of heavy metals and sulphur have subjected the vegetation near the smelter to a strong selection pressure. The fundamental evolutionary question to which this thesis seeks an answer is: What is the importance of phenotypic plasticity compared to ecotypic differentiation of long lived dwarf shrubs in survival under a heavy metal load? One practical aim is to evaluate the applicability of local plant species in the phytostabilisation of polluted soil and as bioindicators in biomonitoring studies. The specific aims of the thesis are - to evaluate the sensitivity of bryophyte, lichen and vascular plant species to heavy metals according to their occurrence at different distances from the smelter and accumulation pattern of toxic elements (I, II) - to study how the species composition of the understorey vegetation and soil seed banks change along a heavy metal and sulphur gradient near the Harjavalta smelter (I, HI) - to study seedling recruitment from the forest soil in order to determine the revegetation potential of the seed bank (III) - to study the clonal diversity and isoenzyme variation of A uva-ursi and K uliginosum at different distances from the smelter - to study the importance of compensatory growth (IV) and activation of the bud reserve (V) of clonal dwarf shrubs as a resistance mechanism to heavy metals - to determine the sensitivity of A. uva-ursi to Cu in relation to that of the other dwarf shrub species in greenhouse conditions (VI) 15 2. Study areas and pollutant emissions 2.1. Experimental sites along the Harjavalta pollution gradient The study area is situated near the Cu - Ni smelter at Harjavalta (61°19' N, 22°09'E), SW Finland. The Finnish Forest Research Institute established a number of experimental plots in Scots pine stands for liming and fertilization (Mälkönen et al. 1999, Derome 2000 a), nutrient flux (Helmisaari et ai. 1995) and restoration (Kiikkilä 2002) studies along a 8 km transect running SE from the smelter (Fig. 2). A large part of the field data of this thesis has been collected from these plots or from their immediate vicinity. The experimental stands along the transect (0.5, 2, 4 and 8 km) were 40 - 55 years old (I - V), whereas the two extra study stands (1 and 3 km) were 51-67 years old (I. III). All the stands, except one peatland site (IV), were growing on dry, nutrient-poor sandy soils of the Calluna site type (Cajander 1909). A detailed description of the stand characteristics is given in I and 11. 2.2. Emissions from the Harjavalta smelter The Harjavalta Metals smelter complex is one of the largest point sources of heavy metal emissions in Finland (Melanen et ai. 1999, Tammiranta 2000, Jussila 2003). The copper smelter has been operating since 1945, and the nickel smelter since 1960. The concentrated ores contain sulphur, heavy metals and arsenic. Before the sulphuric acid plant was built in 1947, all the SO, produced during the smelting process (annually about 30 0001) was emitted into the atmosphere, causing severe damage to the surrounding coniferous forests (Helmisaari 2000). Since the beginning of the 1990'5, the emissions have been considerably reduced by the installation of new filters in 1990, 1991 and 1994. The temporal change in the emissions during 1985 - 1995 is presented in II (II: Fig. 2). 2.3. Ecological research in the Harjavalta area Many investigations focusing on different aspects of forest ecosystem processes have been carried out along the Harjavalta transect during the last decade. In addition to the fertilisation experiments (Derome & Saarsalmi 1999, Mälkönen et ai. 1999, Derome 2000 a), studies on nutrient fluxes (Nieminen & Helmisaari 1996, Derome & Nieminen 1998), distribution of radiocaesium in soil and vegetation (Outola et al. 2003), soil microbiology (Fritze et al. 1989, Fritze etal. 1996, Kiikkilä et al. 2000) and ecophysiology of dwarf shrubs (Monni et al. 2000b, 2001a,b, Uhlig et al. 2001) have generated a considerable amount of information about the factors affecting the understorey vegetation. The accumulation of Cu and Ni and other heavy metals in the soil has resulted in a severe deficit of plant-available Ca, Mg and K in the organic layer caused by the inhibition of mineralisation and the displacement of these base cations from exchange sites (Derome & Lindroos 1998). In addition to toxic element concentrations in the soil, nutrient 16 Fig. 2. Location of the three study areas. The town of Harjavalta is situated about 30 km from the coast of the Gulf of Bothnia in western Finland. The study sites at distances of 0.5, 1, 2, 3, 4 and 8 km from the Harjavalta Cu-Ni smelter (see stack on the map) have been marked on the detailed map (grey points: forest health fertilisation experiments of the Finnish Forest Research Institute). The sandpit and polluted habitats (V) have been marked by Sp and P, respectively. Hämeenkangas (Hk) and Mekrijärvi (Mj) represent background areas (II). Pohjakartta © Maanmittauslaitos lupanumero 6/MYY/03 imbalances and a decreased water-holding capacity (Derome & Nieminen 1998) have increased the stress encountered by the plants growing near the smelter. Bioindicator studies were started in the middle of the 1970's (Laaksovirta & Silvola 1975, Hynninen 1986) and they have continued up until today (Jussila 2003). They show that, despite the reductions in emissions, the effects of the Harjavalta smelters extend to a distance of at least 10 km from the emission source. Kiikkilä (2003) has given a recent overview of all the ecological studies dealing with a range of organisms (e.g. birds, insects and endophytic fungi) carried out in the Harjavalta area. 2.4. Background areas One study site situated in Hämeenkangas (61°45N, 22°40'E) and another in Mekrijärvi (62°47'N, 30°58'E) (Fig. 2) were selected as reference areas for the plant chemical composition (II). The stand in Hämeenkangas (age 44 years) was an untreated control of 17 a fertilization experiment and represented the same forest site type (Calluna type) as the study stands along the Harjavalta transect. The Mekrijärvi stand (age 45 years), in contrast, represented a slightly more fertile site type (Vaccinium type) than the other stands. Its selection was justified because the low N, S and heavy metal deposition in the area in question made it a suitable reference level especially for bryophytes and lichens. Furthermore, it offered a wide range of data dealing with nutrient fluxes in forest ecosystems (Helmisaari 1995). The background areas had no local emission sources. 18 3. Materials and methods An overview of the study sites, number of sample plots, time of sampling and the studied species and variables are given in Table 1. Detailed descriptions of the methods are given in the original articles. 3.1. Vegetation surveys (1,11) and seed bank composition (III) The abundance of the understorey vegetation was studied using the point quadrat method at six locations (0.5, 1, 2, 3, 4 and 8 km) along the study transect in August 1993. The vegetation analysis was carried out on three sample plots (30 x 30 m) at each distance, and 16 vegetation quadrats (1 m 2) were studied on each plot. The total data consists of 288 1 m 2 vegetation quadrats. The vegetation survey of the National Forest Inventory, carried out in 1995 (Reinikainen et ai. 2000), was used as a reference. The same method was applied in 1992, when the vegetation survey was carried out on the smaller number of sample plots selected for the studies on the chemical composition of plants (II). In addition to the sites mentioned above (0.5, 1, 2, 3, 4 and 8 km), the soil samples for the seed bank analyses were collected from the fertilised sites located at 0.5, 4 and 8 km from the smelter in May 1994 (III). Five soil samples were taken from the buffer zones surrounding each of the three replicate plots. The soil samples were taken from the organic layer (including the litter layer) and the upper part of the mineral soil layer at a depth of 5 - 10 cm using a 9.5 x 9.5 cm metal sampler. The germination experiment with a total of 135 soil samples, spread out on trays (18.5 x 21.5 cm), was carried out in a greenhouse. The growth substrate consisted of mixed quartz sand and peat. The emerged seedlings were counted once a week. 3.2. Chemical analyses of plants (11, IV,VI), soil (I -V) and precipitation (II, IV) Species-specific composite samples were taken in 1992 (additional samples in 1993 and 1994) for the chemical analysis of understorey bryophytes, lichens and vascular plant species growing at four distances from the smelter (0.5, 2, 4 and 8 km) and at two background sites (II). The plant material was not washed before chemical analysis and thus included the surface accumulation of elements. Total element concentrations (P, K, Mg, Ca, Fe, Zn, Mn, Cu, Ni, Cd, Pb and Al) were determined by dry digestion (HNO,/ and analysed by inductively coupled plasma atomic emission spectrometer (ICP AES). Total sulphur and nitrogen concentrations were determined on the homogenized samples on LECO S-132 and LECO CHN-600 analysers. These standard methods were also applied when determining the Cu (IV, VI) and Ni (IV) concentrations of the shoots and roots of the dwarf shrub species in the field and greenhouse material. All the analyses were performed in the Central Laboratory of the Finnish Forest Research Institute. Samples were taken from the organic (I - IV) and mineral soil (II: 0-5 cm, V: 0 - 10 cm) layers along the Harjavalta transect and in the background areas in 1992 - 1993 19 Table 1. General description of the study sites, number of sample plots, time of the surveys and sample collection, studied species and measured variables. Variables Species abundances (point frequency, %) Chemical composition of organic layer Species abundances (visual cover, %) Element concentrations in species, soil and precipitation Species abundances (point frequency, %) = Germinated seeds and seedling survival in soil samples (greenhouse experiment) Compensatory growth after shoot clipping (Clonal diversity and isoenzyme variation) Branching architecture Survival, growth and Cu accumulation Species Bryophytes Lichens Vascular plants Bryophytes Lichens Vascular plants Vascular plants Arctostaphylos uva-ursi Vaccinium uliginosum Arctostaphylos uva-ursi Arctostaphylos uva-ursi Calluna vulgaris Empetrum nigrum Vaccinium uliginosum [s Time August 1993 1995 August 1992 (1993, 1994) Aug. -Sept. 1992 May1994 1994- 1995 (1994, 1997) 2000 1999 CD CD O O) O) O CT) CD O T- T- CNJ O Q. 4— O CO X CM CN CO X CO X Sites Numbe (distance from the smelter) Harjavalta transect 6 0.5, 1, 2, 3, 4 and 8 km Reference: National forest inventory Harjavalta transect 0.5, 2, 4 and 8 km Reference: Hämeenkangas and Mekrijärvi Harjavalta transect 6 0.5, 1, 2, 3, 4 and 8 km Fertilised: 0.5, 4 and 8 km 3 Field manipulation: 2 and 8 km forest 0.5 and 4 km forest, 5 km bog Field study in four habitas: Polluted and restauration 0.5 km, sand pit 6 km and forest 8 km Greenhouse experiment References: Monnietai. (2000a) Monni et ai. (2000b) Salemaa et ai. (2003) Study > > > 20 (Hämeenkangas in 1990). Some extra soil samples were taken at a later date (V). Total N was determined on a CHN analyser. Exchangeable Ca, Mg, K, Cu, Ni, Zn, Fe, Mn, Cd, Pb and extractable P and S at the Harjavalta plots were determined by extraction with 1 M ammonium acetate (pH 4.65) + 1% EDTA, followed by analysis by ICP-AES (I, 11, IV). Ammonium acetate extraction has been extensively used for determining the plant available fraction of elements in soils, and EDTA increases the efficiency of heavy metal (especially Cu and Fe) extraction (Lakanen & Erviö 1971). The extractant used for the samples from Hämeenkangas and Mekrijärvi did not include EDTA. The extractant used in 111 and Y was barium chloride (0.1 M) + EDTA (Derome 2000b). The element concentrations in the organic layer were expressed on an organic matter basis in order to reduce the variation arising from the inclusion of varying amounts of mineral soil in the organic layer samples. Soil analyses were performed in the Central Laboratory of the Finnish Forest Research Institute and in the laboratory of the Joensuu Research Centre. Bulk deposition was collected in open areas close to the study stands using five (Mekrijärvi: 20) rainfall collectors (d = 20 cm) during the snowfree period or two snow collectors (d = 36 cm) during the winter. Stand throughfall was collected using 20 rainfall collectors located systematically inside the plots during the snowfree period and six (Mekrijärvi: 10) snow collectors (11, IV). See Derome and Nieminen (1998) for details of the chemical analysis of precipitation. 3.3. Compensatory growth (IV) and branching pattern (V) of dwarf shrubs in the field Shoot clipping manipulation was performed on A. uva-ursi in two Scots pine stands at 2 km and 8 km and on V. uliginosum at 0.5 km and 4 km distances from the smelter, as well as in a drained peatland stand at 5 km (IV). The two species did not occur in sufficient numbers at the same distances. A total of 30 clones per species were randomly selected at each site. Clonal diversity and isoenzyme variation of these clones were also studied (Section 3.4). The clones were divided into three groups, ten replicates in each: undipped controls, clones clipped in autumn (1994), and clones clipped soon after bud break in spring (1995). Shoot clipping was restricted to three randomly selected main branches on each experimental clone. All the current-year shoots of the three main branches were removed and stored for further measurements. The branches were harvested for biomass measurements at the end of July 1995. Horizontal spreading and axillary bud activation of A. uva-ursi was studied in four habitats in the vicinity of the smelter in September 2000: 1) restoration experiment (0.5 km to the S of the smelter), 2) treeless polluted area (0.5 km W, Torttila), 3) sand pit (6 km SE) and 4) pine forest (8 km SE) (V) (Fig. 2). In the restoration experiment, ten six year-old plants were removed together with the roots. Five separate established clones were randomly selected in the other habitats. Altogether 1 - 3 branches with the six youngest annually grown shoots (formed during 1995 - 2000) were taken from the periphery of each clone. Thus all the sample branches had one parent shoot, formed in 1995, from which all the daughter shoots had developed. The following variables, used 21 later in a simulation model, were recorded: the length, location and branching angle of the shoots; the number of activated and inactive buds; the age, hierarchy and terminal types of the shoots. 3.4. Electrophoretic analysis of isoenzyme variation in the dwarf shrub populations Isoenzyme variation in the populations of V. uliginosum (Scots pine stands at 0.5 km and 4 km, peatland site at 5 km) and A. uva-ursi (Scots pine stands at 2 km and 8 km) were studied by means of protein electrophoresis. Leaf samples of V. uliginosum (40 samples per site) were collected in June 1994. Current-year leaves of A. uva-ursi were collected from 20 plants at 2 km and from 25 plants at 8 km in June 1997. The material consisted of samples from all the experimental clones in the shoot clipping experiment (IV) and some additional clones. Selection of the plants was carried out at a 5 m minimum distance between adjacent plants. The area of the studied populations ranged from 2 500 to 3 000 nr. A total of 10 enzyme loci for A. uva-ursi and 12 partly different enzyme loci for V. uliginosum (p. 27) were assayed by the standard starch gel electrophoresis technique as described by Mattila et ai. (1994). The analyses were carried out at the Foundation for Forest Tree Breeding. Interpretation of the tetraploid zymograms (electrophoretic banding patterns) was performed according to Krebs & Hancock (1989). In addition to the determination of genotypes, the following genetic parameters were calculated at the population level: mean number of alleles per locus, proportion of polymorphic loci and observed mean heterozygosity (H o ) over all and over polymorphic loci. Genotype frequencies between the sites were compared using the Chi-square test. 3.5. Experimental exposure of Arctostaphylos uva-ursi and other dwarf shrubs to Cu (VI) Rooted cuttings of A. uva-ursi originating from a distance of 2 km from the smelter were grown in quartz sand in pots and exposed to five levels of Cu (1,10, 22, 46 and 100 mg/1 as CuCl2) in a greenhouse (VI). The Cu was added to a nutrient solution (Stribley & Read 1976) given to the plants once or twice a week. The total amount of nutrient solution given was 12 x 50 ml/pot. The duration of the experiments was 8 weeks. The Cu resistance was quantified by means of the following variables: 1) plant survival, 2) biomass production, and 3) Cu accumulation in shoots and roots. The Cu resistance of A. uva-ursi was compared to that of the other dwarf shrub species grown under similar experimental conditions. Earlier experiments using cuttings of E. nigrum (Monni et al. 2000 a) and seedlings of C. vulgaris (Monni et ai. 2000b) differed from the present experiment in that they lasted for only 6 weeks and Cu was given in the form of CuS04 . On the other hand, the experiment using cuttings of V. uliginosum lasted for 8 weeks, but the amount of the solution (with CuCl 2 ) applied was lower, 9 x 50 ml (Salemaa et al. 2003). 22 3.6. Statistical analysis The vegetation survey data were ordinated using global non-metric multi-dimensional scaling (DECODA 2.04 software) (I). Kruskal-Wallis non-parametric analysis of variance was used in comparing the species abundances in the Harjavalta data to the national forest inventory reference (I), and the number of germinated seeds and mortality rate of seedlings in the soil samples from fertilised and untreated plots (III). Differences in the element concentrations between the life forms and species were tested by Mann-Whitney's U tests (II). Non-parametric statistics were used because the sample number was low, and it was not possible to test the normality of the distributions. Actuarial time tables and Wilcoxon statistic were used for analysing the survival probability of the Calluna vulgaris seedlings over time (III). Linear and non-linear regressions were used when studying the abundance of the plants vs. chemical composition of their tissues (II), and the number of germinated seeds and mortality rate of seedlings (III) as a function of the distance from the smelter. Regression models were also applied when studying the effect of soil chemistry (fertilisation) on seedling mortality (III), the age dependence of the bud activation in branches of A. uva-ursi (V), and the response of A. uva-ursi cuttings to the applied Cu levels (VI). Nested ANOVA, in which the sample was nested under the plot and the plot under the distance, was used when analysing the numbers of germinated seeds and mortality rate along the Harjavalta transect (III). The effects of shoot-clipping and habitat and their interaction on the growth responses of V. uliginosum and A. uva-ursi were studied using two-factor ANOVA, followed by pairwise contrasts (only the results of contrasts are given in IV: Figs. 2 - 3). Three-factor ANOVA was used when studying the effects of habitat, terminal type and age of the shoots on the number/proportion of activated buds of A. uva-ursi (V). Tukey's tests (V) and t-tests (IV, V) were used in testing pairwise differences between different factors in ANOVAs. Two-by-two contingency tables and Chi-square tests were used in testing frequency based data in the branching morphology of A. uva-ursi (V). The branching response to different environmental conditions was simulated by means of an L-system architectural model (Prusinkiewicz & Lindenmayer 1990) based on the use of an annual time step and the demographic and morphological parameters measured in each habitat (V). 23 4. Results 4.1. Heavy metal accumulation and the distribution of species (I, II) Heavy metal and sulphur deposition during the last 50 years has drastically affected the occurrence of plant species, their relative abundances (I) and chemical composition (II) along the Harjavalta transect. The total number of plant species decreased from 30 at 8 km to 8 at 0.5 km from the smelter (I). The overall coverage of the vegetation also decreased towards the smelter (Fig. 3). Elevated N, S and heavy metal concentrations (Cu and Ni distributions in Fig. 4) were found in all life forms near the pollution source. Four damage zones were distinguished along the pollution transect on the basis of the vegetation composition and the element concentrations of the organic soil layer (I). An overview of the species occurrence (I) and the highest Cu and Ni concentrations in their tissues (II) in these four areas is given below: Area of severe damage (0.5 - I km): The understorey vegetation was almost totally absent up to a distance of 0.5 km from the smelter. Except for pioneer species, the bryophytes and lichens typical of mature forests were missing. Only a few patches of E. nigrum, V uliginosum and Carex globularis were present. A few seedlings of Pinus sylvestris and Betula pubescens were growing in the most polluted area. Some shoots of Vaccinium myrtillus, V. vitis-idaea and Ledum palustre were found at 1 km from the smelter. E. nigrum and C. globularis accumulated higher concentrations of Cu (184 - 254 (ig/g) and Ni (51 - 17 ng/g) in the current-year growth than V. uliginosum (Cu 38 ug/g, Ni 42 [xg/g) (Fig. 4e,f). However, a pioneer moss Pohlia nutans accumulated considerably higher amounts of Cu (1 397 |ig/g) and Ni (334 [xg/g) than the vascular species (Fig. 4c,d). Fig. 3. The total abundances of the bryophytes, lichens and vascular plant species along the Harjavalta transect (H0.5 - H8: 0.5 -8 km from the smelter, mean abundance of three plots per each distance in 1993), Hk = Hämeenkangas (1992) and Mj = Mekrijärvi (1992) (I, II). 24 Area of moderate damage (2-3 km): A. uva-ursi and E. nigrum began to increase, but C. vulgaris, which is the characteristic species of Calluna type forests, was still very scarce. The first specimens of reindeer lichens (Cladina spp.J were recorded, but Cetraria islandica was more abundant than the reindeer lichens. Cup lichens (Cladonia spp.) and P. nutans reached their highest abundances halfway (3 km) along the transect (I). The Cu (18 -32 fxg/g) and Ni (8 -12 [ig/g) concentrations of dwarf shrubs (current-year shoots) were clearly lower than those of C. islandica (Cu 108 [ig/g, Ni 26 [ig/g) and P. nutans (Cu 872 [ig/g, Ni 209 [ig/g) (Fig 4). Area of slight damage (4 km): The floristic composition at 4 km resembled that of normal Calluna type forests. The bryophyte layer was still poorly developed (Fig. 3). Reindeer lichens occurred in normal abundances (I), but their heavy metal concentrations were considerably higher (Cu 160 - 260 [ig/g, Ni 30 - 40 [ig/g) than the background values in Mekrijärvi and Hämeenkangas (Cu and Ni 2 - 6 [ig/g) (II). Area of minimum disturbance (8 km): The total coverage of the vegetation approached almost 90 % and all the species groups typical to mature forests were present at a distance of 8 km (Fig. 3). However, the Fig 4. Cu and Ni concentrations in (a, b) the organic and mineral (0-5 cm) soil layers (mg/kg) and bulk precipitation (mg/m 2 ), (c, d) selected cryptogam species (µg/g) and (e, f) the current year shoots of dwarf shrubs (µg/g) at different distances (km) from the Harjavalta Cu-Ni smelter (H0.5, H2, H4 and H8) and in two background areas Hämeenkangas (Hk) and Mekrijärvi (Mj). 25 abundance of C. vulgaris was still lower than in the background areas (I). The abundance of Pleurozium schreberi was also lower than normal, and the Cu (180 [xg/g) and Ni (38 ug/g) concentrations in its younger parts were considerably higher than those of the background areas (Cu 6-13 [ig/g, Ni 4 - 7 ug/g). In general, when all the species grew on the same plot, heavy metal concentrations (except Mn) tended to increase in the order: vascular plants < C. islandica < Cladina lichens < bryophytes of mature forests < P. nutans (II: Fig. 3). The accumulation of Cu in the different species (excluding P. nutans) correlated positively with the closest distance to the smelter at which the species occurred (II: Fig. 4). 4.2. Seed bank composition in relation to the existing vegetation (I, III) Viable seeds were found at all the studied distances from the smelter (III: Table 3). Altogether 1 300 seedlings germinated in the total of 135 seed bank samples. The emerged seedlings represented 15 taxa, of which 6 species were grasses and sedges, 4 dwarf shrubs, 3 trees and 2 herbs (Table 2). The most numerous species were B. pubescens and C. vulgaris. The average seedling density varied from 250 to 4 750 per m 2 at the different distances. Table 2. The frequency (%) of the sample plots in which the species emerged from the seed bank (III: Table 3) versus aboveground occurrence of vascular plants (I: Table 2). Data from the untreated plots (n = 18) along the 8 km transect. Species Seed bank Aboveground Andromeda polifolia - 5.6 Agrostis capillaris 5.6 - Arctostaphylos uva-ursi - 33.3 Betula pendula - 5.6 Betuta pubescens 77.8 27.8 Calluna vulgaris 72.2 44.4 Carex canescens 5.6 - Carex ericetorum 5.6 - Carex globularis 16.7 44.4 Deschampsia flexuosa 5.6 16.7 Empetrum nigrum 5.6 88.9 Epilobium sp. - 38.9 Festuca ovina 5.6 - Ledum palustre - 22.2 Picea abies - 33.3 Pi n us syivestris 27.8 88.9 Popuius tremula 16.7 5.6 Rumex acetoselia 11.1 - Salix aurita - 5.6 Vaccinium myrtiilus - 27.8 Vaccinium uiiginosum 22.2 38.9 Vaccinium vitis-idaea 27.8 83.3 26 The seed bank species were rather well represented in the existing vegetation. Of the 17 vascular plant species growing on the unfertilised plots, 10 were found in the seed banks (Table 2). The percentage similarity between the species in the existing vegetation and the seed banks varied from 18 %t067 % at different distances from the smelter. The total similarity in the data for the whole transect was about 70 % in the unfertilised, and 60 % in the fertilised plots. The number of C. vulgaris seedlings increased with increasing distance from the smelter, but no such trend was found for the other species (III: Fig. lb). 4.3. Seedling survival in forest soil polluted with heavy metals (III) Although germinable seeds were found even in the most polluted study area, the majority of the seedlings died at an age of a few weeks. In the life table analysis, the survival probability of C. vulgaris seedlings was the higher, the further away from the smelter the soil was collected (III: Fig. 2). The survival probability was the lowest in the soil from distances of 0.5 - 2 km (0 - 30 %), increased to over 60 % at distances of 3 - 4 km, and to 80 % at 8 km. Nutrient addition and liming increased the survival probability of the seedling slightly at 0.5 km and 4 km, but the effect was not statistically significant. 4.4. Genetic structure of the dwarf shrub populations Isoenzyme analysis verified that the studied populations of A. uva-ursi and V. uliginosum were autotetraploid and multiclonal. An autotetraploid individual has four different alleles per locus, and its genotype can be marked e.g. as 1112 according to the electrophoretic banding pattern. The majority of the sampled plants represented different genotypes (Table 3a). A. uva-ursi had 4 polymorphic loci out of 10 studied, and V. uliginosum 5 out of 12 (Table 3a). The average number of alleles per locus was 2 and 1.4, respectively. The percentage of observed heterozygous individuals over all loci was about 21 % in the populations of A. uva-ursi, and 30 % in the populations of V. uliginosum. The corresponding values over polymorphic loci were 51 % and 73 %, respectively. The clonal diversity of A. uva-ursi was the highest at the distance of 8 km and that of V uliginosum at 0.5 km (Table 3a). Both species had some samples with the same isoenzyme pattern but, owing to the long distance between the samples (> 20 m), it is more realistic to assume that they represent different genotypes. Only in the peatland population (5 km) of V. uliginosum were there some (n = 5) large clones from which branches may have been selected twice. The genotype frequencies of A. uva-ursi did not differ between the sites (Table 3b). In V. uliginosum there were differences only in two loci. The peatland population differed from both forest populations (0.5 km and 4 km) for diaphorase (DIA) (j 2 = 12.84, P = 0.05, df = 6), and the forest population at 4 km differed from those at forest 0.5 km and peatland 5 km for phosphoglucose isomerase (PGI2) (% 2 - 14.72, P = 0.06, df = 8). 27 Table 3. a) Number of genotypes, percentage of loci polymorphic (0.95 criterion) and mean observed heterozygosity over all (H 0 1) and over polymorphic (H 0 2) loci in the populations of A. uva-ursi and V. uliginosum. b) Site-specific genotype frequencies of A. uva-ursi for 10 enzyme loci and of V. uliginosum for 12 enzyme loci. Both species are autotetraploid (4X). Zymogram pattern is presented. a) A. uva- -ursi V. uliginosum Site F2 F8 F0.5 F4 P5 No of samples 20 25 40 40 40 No of genot. 16 24 39 36 31 Polym. loci % 40 40 42 42 42 H 0 1 0.21 0.21 0.32 0.31 0.29 H 0 2 0.50 0.53 0.76 0.73 0.71 b) Genotype Genotype 6PG1 1111 0.05 0.00 6PG1 1111 1.00 1.00 1.00 2222 0.95 1.00 6PG2 1111 0.10 0.16 0.15 6PG2 2222 1.00 1.00 1112 0.55 0.37 0.27 FE1 1111 0.95 1.00 1122 0.15 0.29 0.30 2222 0.05 0.00 1222 0.15 0.16 0.23 MDH 1111 0.05 0.00 2222 0.05 0.02 0.05 2222 0.90 0.64 MDH1 1111 1.00 1.00 1.00 2223 0.05 0.24 MDH2 1111 1.00 1.00 1.00 2233 0.00 0.08 MDH3 1111 0.08 0.00 0.00 ADH 1111 0.05 0.00 1112 0.33 0.22 0.35 1112 0.00 0.04 1122 0.47 0.70 0.55 1122 0.05 0.04 1222 0.12 0.08 0.10 1222 0.25 0.31 ADH 1111 1.00 1.00 1.00 2222 0.65 0.61 DIA 1112 0.00 0.03 0.00 MNR 1111 0.05 0.00 1122 0.08 0.13 0.13 2222 0.95 1.00 1222 0.50 0.42 0.17 PGM 1111 0.00 0.20 2222 0.42 0.42 0.70 1112 0.20 0.16 GOT1 1111 1.00 1.00 1.00 1122 0.35 0.12 GOT2 1111 0.05 0.10 0.03 1123 0.10 0.00 1112 0.22 0.25 0.27 1133 0.00 0.04 1122 0.30 0.21 0.13 1222 0.20 0.28 1222 0.20 0.31 0.30 1223 0.05 0.04 2222 0.23 0.13 0.27 2222 0.10 0.16 PGI2 1111 0.18 0.48 0.22 PGI1 1111 0.95 1.00 1112 0.23 0.10 0.18 2222 0.05 0.00 1122 0.43 0.40 0.42 PGI2 1112 0.20 0.17 1222 0.08 0.00 0.13 1122 0.20 0.21 2222 0.08 0.02 0.05 1222 0.40 0.33 IDH1 1111 1.00 1.00 1.00 2222 0.20 0.29 GDH 1111 1.00 1.00 1.00 IDH 2222 1.00 1.00 28 4.5. Compensatory growth and branching pattern of the dwarf shrubs (IV,V) Both the evergreen A. uva-ursi and deciduous V. uliginosum displayed a considerable ability to activate dormant meristems (axillary buds) and regrow after shoot clipping (IV). The biomass of the current-year shoots during the next growing season was at least 80 % compared to the within-clone control in both species after autumn clipping (Fig. sa). Shoot clipping in early summer was more detrimental for both species, and A. uva ursi suffered more than V. uliginosum (Fig. sa). A. uva-ursi showed overcompensation (over 100 % growth compared to the control) in the number of new shoots after autumn clipping (Fig. Sb). A similar trend was found in V. uliginosum at the peatland site after spring clipping (Fig. Sb). No berries developed on either species in the year following the autumn treatment because clipping removed all the flower buds. Spring clipping had no effect on the sexual reproduction of A. uva-ursi, but decreased the berry production of V. uliginosum. The degree of compensatory growth of both species was only slightly affected by the distance from the smelter. The state of the apical buds (living or dead) in the branches of A. uva-ursi was extremely important in regulating lateral branching in resource-poor habitats (V). Apical dominance of lateral branching was strongest in the intact shoots in the polluted (nutrient limited) and forest (light limited) habitats. However, when the apical bud of the parent shoot was dead, the disruption of apical dominance caused intensive branching in the poor habitats (V: Fig. 3). This response was demonstrated both in the shoot clipping Fig. 5. The average compensatory growth of branches of V. uliginosum and A. uva-ursi after shoot clipping performed at different distances from the smelter (F = forest, P = peatland site) in autumn 1993 and spring 1994. The compensatory growth is presented as percentages of a) the biomass production, and b) the number of new shoots compared to the control branches (IV). The dashed line indicates the 100 % level. 29 Fig. 6. Examples of the simulated branching patterns of A. uva-ursi in a) the polluted and b) the sand pit habitats (LIGNUM model). The growth cycle is 30 years in a) and 20 years in b). Annual shoot mortality increased from the 3 rd year by 15 %, 20 % and 30 % years in dominant, subdominant and nondominant shoots, respectively in a). Corresponding parameters from the 4th year are 7 %, 10 % and 15 % for b). Dead branches have dropped off. The collision detection algorithm determines free growing space for active buds within an angle of 30 degrees and a distance of 20 cm in a) and 30 cm in b). experiment (IV) (Fig. Sb) and in the analysis of the branching pattern of A. uva-ursi growing in forest or heavily disturbed polluted habitats (V). Apical dominance was much weaker in the sandpit and restoration habitats, where the nutrient availability and light level were relatively high. In these habitats branching frequency was high in both intact and terminated parent shoots (V: Fig. 3). Simulations produced a variety of branching patterns for A. uva-ursi depending on the pollution and resource levels of the habitat (V: Fig. 6). A more realistic, star-like shape of the clones was produced by adding a collision detection algorithm to the model (Fig. 6). 30 4.6. Cu resistance of Arctostaphylos uva-ursi compared to the other dwarf shrubs (VI) Growth inhibition and reduction of biomass production were the general responses of A. uva-ursi cuttings to exposure to Cu in a nutrient solution (VI). Compared to the other dwarf shrub species (C. vulgaris, E. nigrum and V. uliginosum) grown under similar experimental conditions, A uva-ursi proved to be the most sensitive. The biomass production of the four species decreased in the order: A. uva-ursi > C. vulgaris > E. nigrum > V. uliginosum (Fig. 7a). The lowest external Cu level that reduced the growth of C. vulgaris and A uva-ursi by 50 % was 10 mg/1. The corresponding critical concentration for E. nigrum was 22 mg/l. V. uliginosum did not reach a 50 % decrease in growth even at the highest Cu level of 100 mg/1. Also, the accumulation pattern of Cu in the new leaves indicated high sensitivity of A. uva-ursi to absorbed Cu. The Cu concentrations were lowest in the leaves of A. uva-ursi, followed by V. uliginosum < E. nigrum < C. vulgaris (Fig. 7b). Fig. 7. a) The decrease in the relative shoot growth (%) (standardized to the Cu level of 1 mg/l) of E. nigrum, C. vulgaris, A. uva-ursi and V. uliginosum, and b) the Cu concentrations of the new leaves (µg/g) at different Cu levels applied. The shoot growth was based on the biomass production (dry weights) in the other species, except for C. vulgaris in which the length growth was used (Salemaa et at. 2003) (VI) 31 5. Discussion 5.1. Heavy metal accumulation and the relative resistance of the species In general, vascular plants, being capable of restricting the uptake of toxic elements, grew closer to the smelter than lichens, while sensitive bryophytes began to increase at further distances from the smelter (I, II). The occurrence of the life forms which were still surviving followed, in relation to the closest distance to the smelter, the order: bryophytes of mature forests (Dicranum spp. and P. schreberi) (4-8 km) > reindeer lichens (2-4 km) > other lichens (2 km) > vascular plants (0.5 - 2 km). Pioneer mosses (dominated by Pohlia nutans) were exceptions to this general pattern. Despite the accumulation of large amounts of Cu and Ni in the living tissues, these species had surviving populations in the immediate vicinity of the emission source. The order of the species along a pollution gradient may give some indication about the general resistance level of the species against pollution (e.g. heavy metals and SOJ and other stress factors. In this respect, vascular plants were more resistant than lichens, whereas bryophytes were the most sensitive plant group. However, dwarf shrubs also apparently suffered from phytotoxic effects, which was expressed as their decreased abundance towards the smelter. Other environmental factors such as nutrient deficiencies in the soil, drought and increased illumination, frequently strengthen the selection pressure on plants in severely affected stands. It should be noted that unwashed samples have considerable amounts of dust attached to them, and actual tissue concentrations that have harmful metabolic effects are much lower than those reported in field conditions (Brown & Brumelis 1996, Bennett 1999). Therefore it is very difficult to identify the contribution of individual factors and present any maximum (toxic) limits for the survival of plant species based on field data. Similar trends in the sequence of plant species around emission sources in coniferous forests have been reported e.g. near the Cu-Zn smelter at Gusum, SE Sweden (Folkeson 1984, Folkeson & Andersson-Bringmark 1988, Tyler et al. 1989), fertiliser factories in W Finland (Huttunen 1975, Väisänen 1986), a smelter complex in Sudbury, Ontario (Amiro & Courtin 1981), and the Cu-Ni smelters in Monchegorsk, Kola Peninsula, NW Russia (Rigina & Kozlov 2000). The heavy metal concentrations were elevated especially in the older parts of the species (II: Appendices). For instance, the Cu and Ni concentrations of the dwarf shrubs were the highest in dead parts, and decreased from older to current-year shoots and berries. The Cu (13.78 ug/g) and Ni (8.64 (xg/g) concentrations in the berries of Vaccinium vitis-idaea at 2 km distance from the Harjavalta smelter were considerably higher than those measured in Lapland in 1990 (Cu: 6.23 [ig/g, Ni: 0.55 [xg/g ) but lower than the highest concentrations found near the Monchegorsk smelter on the Kola Peninsula (Cu: 33.8 |ig/g, Ni: 25.2 ug/g ) (Laine et ai. 1993). Although the heavy metal concentrations of berries are usually lower than those in the other plant parts, they may be an important link in the transfer of heavy metals into food-chains via berry-eating animals and humans. 32 5.2. Revegetation potential of the seed bank Although the understorey vegetation was almost totally absent at the distance of 0.5 km from the smelter, viable seeds of native plant species were present in the soil. However, seedling establishment in a greenhouse experiment presumably failed as a result of the phytotoxicity of heavy metals (III). One reason for the high mortality of the seedlings might be the absence of mycorrhizal infections (c.f. Bradley et al. 1981). In actual field conditions the thick layer of undecomposed needle litter and drought also hinder the germination of seeds near the smelter. Thus, the realisation of the revegetation potential of the local seed banks presupposes soil mitigation to immobilise heavy metals and facilitating the generation of a new organic soil layer. 5.3. Sensitivity of Arctostaphylos uva-ursi to Cu compared to the other dwarf shrubs Compared to the earlier greenhouse experiments, A. uva-ursi (VI) seemed to be more sensitive to Cu than C. vulgaris (Monni et al. 2000b), E. nigrum (Monni et al. 2000 a) and V. uliginosum (Salemaa et al. 2003). The ranking of the four dwarf shrub species based on survival and growth in the Cu exposure experiments was: E. nigrum (most resistant) > K uliginosum > C. vulgaris > A. uva-ursi (most sensitive). This order was the same as that found for the closest distance to the Harjavalta smelter at which the species occurred. A few patches of E. nigrum and V. uliginosum were present at a distance of 0.5 km, C. vulgaris appeared for the first time at 1 km, and A. uva-ursi at 2 km from the smelter (I). The overall accumulation of Cu in A. uva-ursi was similar to that in the other dwarf shrubs: the concentrations were the highest in roots and stems and the lowest in green leaves (Monni et al. 2000 a, b). A corresponding accumulation pattern of Cu has also been demonstrated in other vascular species (reviewed by Balsberg-Pählsson 1989). Short-term experiments using high exposure levels of heavy metals (VI) do not give a complete picture of the heavy metal resistance of long-lived dwarf shrubs, which may form extensive clones with modular subunits. In the real ecological conditions, the adult clones have to cope with low, but chronic exposure to heavy metals that have accumulated in the soil or entered the ecosystem in wet and dry deposition. The two possible ways in which plant populations can become resistant to heavy metals, 1) ecotypic differentiation or 2) phenotypic plasticity of individual plants (Section 1.3.), are discussed below. 5.4. Clonal diversity and genetic variation of the dwarf shrubs In general, outcrossed plant species have a significantly higher genetic diversity than selfed or mixed-mating species (Hamrick & Godt 1990). A. uva-ursi and V. uliginosum are predominately outcrossed insect pollinated species. Both species produced flowers and berries in the study areas (IV). Isoenzyme analysis revealed that almost all the samples were different genotypes, indicating high clonal diversity in both species. This result is 33 consistent with the overview of Ellstrand & Roose (1987), who showed that genetic variation in clonal plants is not rare. The percentage of polymorphic loci at the population level in both species (40 - 42 %) was slightly higher than the average value (34 %) reported for a large number of plant species (Hamrick & Godt 1990). The distribution of the genotype frequencies in both species was rather similar at the different sites. On the basis of the restricted number of studied enzyme loci, there was no evidence of differentiation into heavy metal-specific ecotypes. However, it is also possible that isoenzyme variation is selectively neutral (e.g. Nei et al. 1976), and there is therefore no connection between the isoenzyme pattern and the pollution level. Although vegetative production often predominates in the life of clonal plants, the existence of genetic variation indicates that the populations were originally established by sexual propagules (Ellstrand & Roose 1987). This was also true for the studied populations of A. uva-ursi and V. uliginosum. Nowadays, however, the birth of new clones in the most polluted areas is prevented, because the young seedlings die in the toxic surface soil (III). It is also possible that the failure of seedling establishment results from the absence of metal-tolerant genotypes in the study populations or that soil toxicity is too high for even the existing metal-tolerant genotypes. However, as seed banks maintain genetic diversity of plant populations (Mahy et al. 1999), the evolution of tolerant ecotypes may be possible in the future if the pollution level decreases. The studied clones of both species were tens of years old. Because vegetative spreading is characteristic of clonal dwarf shrubs, some "mother clones" may date back to the time when the smelter first started operating in the 1940'5. The abundance of dead clones near the smelter indicates that the heavy metal concentrations have been too high for the majority of the individual plants. The surviving clones most likely represent the most resistant genotypes of the earlier populations, or have rooted in clean "islands" in the polluted soil, having e.g. wood debris as substrate. 5.5. Phenotypic plasticity as a resistance strategy in clonal dwarf shrubs Phenotypic plasticity is the ability of an individual organism to alter its physiology or morphology in response to changes in environmental conditions (Schlichting 1989). Morphological plasticity via growth is possible because plant development is modular in form. Vuorisalo & Tuomi (1986) define modules as partially self-maintaining, repetitive and multicellular parts of structural individuals. Integration between modules (e.g. annually grown shoots in dwarf shrubs) moderates the impact of local, adverse selection pressures (Slade & Hutchings 1987). Several studies (reviewed by Hutchings & de Kroon 1994) have described changes in the internodal length of the stems or rhizomes, lateral branching intensity and branching angle of clonal plants in response to environmental conditions (e.g. nutrients and light). Phenotypic plasticity allows both clonal and non clonal plants to use avoidance strategies in relation to the heterogeneous distribution of a pollutant (Dickinson et al. 1991). A. uva-ursi and V. uliginosum showed high regrowth after autumn clipping of current year shoots (imitating pollution-induced shoot damage) in a field experiment carried out 34 in polluted and clean sites (IV). In contrast, spring clipping of new shoots was more detrimental to the evergreen A. uva-ursi than to the deciduous V. uliginosum. Differences in the storage reserves and sink-source mechanisms of carbon allocation between evergreen and deciduous species probably explain their distinct response, as demonstrated e.g. by Tolvanen & Laine (1997) using Vaccinium myrtillus and V. vitis-idaea as experimenal species. Contrary to the predictions, the relative amount of activated meristems was higher in A. uva-ursi than in V uliginosum. Architectural constraints (e.g. number of axillary buds) or differences in the rooting pattern might explain this. The creeping branches of A. uva-ursi could have fine adventitious roots, which made the shoots more independent of intraclonal transport of water and nutrients than shoots of V. uliginosum. It should be noted, however, that mechanical cutting of shoots may not have the same physiological effect on plants as pollution induced mortality. A. uva-ursi showed high plasticity in the lateral branching, which varied according to the pollution, light and resource levels of the habitat (V). The plasticity in the clonal morphology was an expression of foraging behaviour that enables the clones to colonize favourable microhabitats and spread the risk of shoot mortality. Strong apical dominance, observed in resource-poor habitats, maintains a reserve of axillary buds that can be used to continue growth after damage ("reserve meristem hypothesis", Tuomi et ai 1994, Aarssen 1995) or after periods with low resource levels (Jonsdottir & Callaghan 1988, Hutchings & de Kroon 1994). The reserve meristem strategy partly explains why A. uva-ursi generally survives in severely disturbed sites such as the polluted one in this study. In addition to dormant bud activation, rapid regrowth and plastic branching, adult dwarf shrub clones can avoid heavy metals by extending their roots into the less toxic, deeper soil layers. For instance, the deepest roots extended down to a depth of 50 cm in the clones of A. uva-ursi (V) and E. nigrum (Uhlig et al. 2001) growing at 2 km and 0.5 km distances from the smelter, respectively. Despite external avoidance of heavy metals, clonal dwarf shrubs express different degrees of real physiological tolerance (Fig. 1). For instance, E. nigrum has an ability to accumulate Cu in cell walls, vacuoles and cytoplasm (Monni et al. 2002). 5.6. Applicability of the results in biomonitoring and phytoremediation The deposition gradient near the Harjavalta Cu-Ni smelter was very steep, resulting in strong inter-correlations between the elemental load in bulk deposition, and the concentrations in the understorey vegetation and organic layer (II). This makes it extremely difficult to distinguish between the role of airborne deposition, wind-blown dust and elements taken up by the substrate in the chemical composition of the plants. The relationship between deposition and plant uptake seem to be strongly dependent on the local conditions and the element ranges in deposition in polluted areas, as emphasized also by Halleraker et al. (1998) and Reimann et al. (2001). Bryophytes, lichens and vascular plants showed considerable differences in their capacity to accumulate pollutants and to grow in contaminated soil at Harjavalta (I, II). Therefore information about all 35 the life forms in the understorey is needed when evaluating the state and recovery of forest ecosystems in heavily polluted areas. In contaminated environments the vegetation cover protects the soil from erosion and the leaching of heavy metals into the groundwater (Vangronsveld et al. 1996). Through the litterfall, the understorey vegetation affects the composition of the organic layer which, in turn, is an important medium for the root growth of forest trees and also the understorey itself. An ability to enrich heavy metals or to grow in contaminated soil makes some dwarf shrub species suitable for the revegetation of damaged forest areas. Although the growth rate of boreal dwarf shrubs is too low for the phytoextraction of heavy metals, they can be used for phytostabilisation of contaminated soil. E. nigrum and A. uva-ursi are, in fact, the two dwarf shrub species planted in a revegetation experiment in the vicinity of the Harjavalta smelter (Kiikkilä 2002). Both species have survived well after the high mortality during the first few years, and their clonal growth habit facilitates rapid expansion and coverage of the forest floor. 36 6. Conclusions I found the following answers to the questions and hypotheses presented in the aims: 1) Heavy metal and sulphur deposition have subjected the vegetation growing near the smelter to a strong selection pressure. The species composition has changed, sensitive species have disappeared and competitive interactions between species may also have been altered. According to the species occurrence along the pollution gradient, vascular plants were more resistant than lichens, whereas the bryophytes of mature forests were the most sensitive taxon. The capacity of bryophytes and lichens to accumulate large amounts of heavy metals made them more sensitive than vascular plants. 2) The accumulation of heavy metals has caused chronic disturbances in the ecosystem, preventing the normal succession of plant communities. The size of the soil seed bank has decreased and young seedlings rapidly die in the contaminated soil. Natural recolonization of the vegetation in heavily polluted areas is a slow process, even though emissions have decreased. Recovery of the vegetation presupposes immobilisation of the heavy metals and the generation of a functioning organic soil layer. 3) Although the dwarf shrubs had the ability to produce berries, and there was genetic variation in the populations, the failure of seedling establishment is a factor preventing the evolution of metal-tolerant ecotypes. There was no evidence that the populations of A. uva-ursi or V. uliginosum had differentiated into heavy metal-specific ecotypes near the smelter. 4) Plasticity in dormant bud activation, rapid turnover of shoots after damage and root growth into deeper soil layers may help the long-lived dwarf shrubs in avoiding heavy metals. 5) The Cu resistance level of different dwarf shrub species varied in the greenhouse experiments as follows: E. nigrum (most resistant) > V. uliginosum > C. vulgaris >A. uva ursi (most sensitive). This order was the same as that found for the occurrence pattern of the species along the pollution gradient near the Harjavalta smelter. 6) The results of this thesis reveal that understorey vegetation has great indicative value when studying the effects of environmental changes in forest ecosystems. However, there were considerable differences between the bryophytes, lichens and vascular plants in their capacity to accumulate pollutants and to grow in contaminated soil. Therefore, information about all plant groups in the understorey vegetation is needed when evaluating the state and recovery of forest ecosystems in heavily polluted areas. 37 References Aarssen, L.W. 1995. Hypotheses for the evolution of apical dominance in plants: implications for the interpretation of overcompensation. Oikos 74(1): 149-156. Alloway, B.J. 1995 a. The origins of heavy metals in soils. In: Alloway, B.J. (ed.). Heavy metals in soil. Blackie Academic & Professional, London, pp. 38-57. Alloway, B.J. 1995b. Soil processes and the behaviour of metals. In: Alloway, B.J. (ed.). Heavy metals in soils. Blackie Academic and Professional, London, pp. 11-37. Alva, A.K., Huang, B. & Paramasivam, S. 2000. Soil pH affects copper fractionation and phytotoxicity. Soil Science Society of American Journal 64: 955-962. Amiro, B.D. & Courtin, G.M. 1981. Patterns of vegetation in the vicinity of an industrially disturbed ecosystem, Sudbury, Ontario. Canadian Journal of Botany 59: 1623-1639. Antonovics, J., Bradshaw, A.D. & Turner, R.G. 1971. Heavy metal tolerance in plants. Advances in Ecological Tolerance 7: 1-85. Baker, A.J.M. 1987. Metal tolerance. New Phytologist 106: 93-111 Baker, A.J.M., Grant, C.J., Martin, M.H., Shaw, S.C. & Whitebrook, J. 1986. Induction and loss of cadmium tolerance in Holcus lanatus L. and other grasses. New Phytologist 102: 575-585. Baker, A.J.M. & Walker, PL. 1989. Physiological responses of plants to heavy metals and the quantification of tolerance and toxicity. Chemical Speciation and Bioavailability 1: 7-17. Balsberg-Pählsson, A.M. 1989. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. A literature review. Water, Air and Soil Pollution 47(3-4): 287-319. Bates, J.W. 1992. Mineral nutrient acquisition and retention by bryophytes. Journal of Bryology 17: 223-240. Bennett, J. 1999. Statistical baseline values for chemical elements in the lichen Hypogymnia physodes. In: Agrawal, S.B. & Agrawal, M. (eds.). Environmental Pollution and Plant Responses. Lewis Publishers, Boca Raton, pp. 343-352. Bradley, R., Burt, A.J. & Read, D.J. 1981. Mycorrhizal infection to heavy metal toxicity in Calluna vulgaris. Nature 292: 335-337. Bradshaw, A.D. & Hardwick, K. 1989. Evolution and stress-genotypic and phenotypic components. Biological Journal of the Linnean Society 37: 137-155. Bradshaw, A.D., McNeilly, T. & Putwain, P.D. 1990. The essential qualites. In: Shaw, J.A. (ed.). Heavy metal tolerance in plants: evolutionary aspects. CRC Press, Boca Raton, pp. 323-334. Brown, D.H. & Brumelis, G. 1996. A biomonitoring method using the cellular distribution of metals in moss. The Science of the Total Environment 187: 153-161. Brown, D.H. & Wells, J.M. 1990. Physiological effects of heavy metals on the moss Rhytidiadelphus squarrosus. Annals Botany 66: 641-647. Bruns, 1., Sutter, K., Menge, S., Neumann, D. & Krauss, G.J. 2001. Cadmium lets increase the glutathione pool in bryophytes. Journal of Plant Physiology: 79-89. Buse, A., Norris, D., Harmens, H., Buker, P., Ashenden, T. & Mills, G. 2003. Heavy metals in European mosses: 2000/2001 survey. UNECE ICP Vegetation. Centre for Ecology & Hydrology, Bangor. 45 p. Cajander, A.K. 1909. Ueber Waldtypen. Acta Forestalia Fennica 1(1): 1-176. Chernenkova, T. & Kuperman, G. 1999. Changes in the spruce forest communities along a heavy metals deposition gradient on Kola peninsula. Water, Air and Soil Pollution: 187-200. 38 Chettri, M.K., Cook, C.M., Vardaka, E., Sawidis, T. & Lanaras, T. 1998. The effect of Cu, Zn and Pb on the chlorophyll content of the lichens Cladonia convoluta and Cladonia rangiformis. Environmental and Experimental Botany 39: 1-10. Chettri, M.K., Sawidis, T., Zachariadis, G.A. & Stratis, J.A. 1997. Uptake of heavy metals by living and dead Cladonia thalli. Environmental and Experimental Botany 37: 39-52. Clemens, S., Palmgren, M. & Kramer, U. 2002. A long way ahead: understanding and engineering plant metal accumulation. Trends in Plant Science 7: 309-315. Derome, J. 2000 a. Detoxification and amelioration of heavy-metal contaminated forest soils by means of liming and fertilisation. Enviromental Pollution 107: 79-88. Derome, J. 2000b. Effects of heavy-metal and sulphur deposition on the chemical properties of forest soil in the vicinity of a Cu-Ni smelter, and means of reducing the detrimental effects of heavy metals. Finnish Forest Research Institute, Research Papers 769. 78 p. Derome, J. & Lindroos, A.-J. 1998. Effects of heavy metal contamination on macronutrient availability and acidification parameters in forest soil in the vicinity of the Harjavalta Cu-Ni, SW Finland. Environmental Pollution 99: 225-232. Derome, J. & Nieminen, T. 1998. Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Environmental Pollution 103: 219-228. Derome, J. & Saarsalmi, A. 1999. The effect of liming and correction fertilisation on heavy metal and macronutrient concentrations in soil solution in heavy-metal polluted Scots pine stands. Environmental Pollution 104: 249-259. Dickinson, N.M., Turner, A.P. & Lepp, N.W. 1991. How do trees and other long-lived plants survive in polluted environments? Functional Ecology 5: 5-11. Dickinson, N.M., Turner, A.P. & Watmough, S.A. 1992. Acclimation of trees to pollution stress: cellular metal tolerance traits. Ann. Bot. 70: 569-572. Ellstrand, N.C. & Roose, M.L. 1987. Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74: 123-131. Eltrop, L., Brown, G., Joachim, O. & Brinkmann, K. 1991. Lead tolerance of Betula and Salix in the mining area of Mechernich/Germany. Plant and soil 131: 275-285. Folkeson, L. 1984. Deterioration of the moss and lichen vegetation in a forest polluted by heavy metals. Ambio 13(1): 37-39. Folkeson, L. & Andersson-Bringmark, E. 1988. Impoverishment of vegetation in a coniferous forest polluted by copper and zinc. Canadian Journal of Botany 66: 417-428. Fowler, D., Cape, J. & Unsworth, M. 1989. Deposition of atmospheric pollutants on forests. Philosophical Transactions of the Royal Society of London: 247-265. Fritze, H., Niini, S., Mikkola, K. & Mäkinen, A. 1989. Soil microbiological effects of a Cu- Ni smelter in southwestern Finland. Biological Fertility of Soils 8: 87-94. Fritze, H., Vanhala, P., Pietikäinen, J. & Mälkönen, E. 1996. Vitality fertilization of Scots pine stands growing along a gradient of heavy metal pollution: Short-term effects on microbial biomass and respiration rate of the humus layer. Fresenius Journal of Analytical Chemistry 5219: 1-6. Garty, J. 2001. Biomonitoring atmospheric heavy metals with lichens: theory and application. Critical Reviews in Plant Sciences 20(4): 309-371. Gimingham, C.H. 1972. Ecology of heathlands. Chapman and Hall. 266 p. Guschina, I.A. & Harwood, J.L. 2002. Lipid metabolism in the moss Rhytidiadelphus squarrosus (Hedw.) Warnst. from lead-contaminated and non-contaminated populations. Journal of Experimental Botany 53: 455-463. Hall, J.L. 2002. Cellular mechanisms for heavy metal detoxification and tolerance. Journal of Experimental Botany 53: 1-11. 39 Halleraker, J.H., Reimann, C., Caritat de, P., T.E., F., Kashulina, G., Niskavaara, H. & Bogatyrev, I. 1998. Reliability of moss (Hylocomium splendens and Pleurozium schreberi) as a bioindicator of atmospheric chemistry in the Barents region: Interspecies and field duplicate variability. The Science of the Total Environment 218: 123-139. Hamrick, J.L. & Godt, M.J.W. 1990. Allozyme diversity in plant species. In: Brown, A.H.D., Clegg, M.T., Kahler, A.L. & Weir, B.S. (eds.). Plant population genetics, breeding and genetic resources. Sinauer, Sunderland, USA, 1990, Sunderland, USA. pp. 43-63. Hartley, J., Cairney, J.W.G. & Meharg, A.A. 1997. Do ectomycorrhizal fungi exhibit adaptive tolerance to potentially toxic metals in the environment? Plant and Soil 189: 303-319. Haslett, 8.5., Reid, R.J. & Rengel, Z. 2001. Zinc mobility in wheat: Uptake and distribution of zinc applied to leaves ar roots. Annals of Botany 87: 379-386. Heinrichs, H. & Mayer, R. 1980. The role of forest vegetation in the biogeochemical cycle of heavy metals. Journal of Environmetal Quality 9: 111-118. Helmisaari, H.-S. 1995. Nutrient cycling in Pinus sylvestris stands in eastern Finland. Plant and Soil 168-169: 327-336. Helmisaari, H.-S. 2000. Function of forest ecosystems under heavy metal loading. In: Mälkönen, E. (ed.). Forest Condition in a Changing Environment - The Finnish Case. Kluwer Academic Publishers, Dordrecht, pp. 248-251. Helmisaari, H.-S., Derome, J., Fritze, H., Nieminen, T., Palmgren, K., Salemaa, M. & Vanha- Majamaa, I. 1995. Copper in Scots pine forests around a heavy-metal smelter in south western Finland. Water, Air and Soil Pollution 85: 1727-1732. Hutchings, M. & de Kroon, H. 1994. Foraging in plants: the role of morphological plasticity in resource acquisition. Advances in Ecological Research 25: 159-238. Hutchinson, T.C. & Whitby, L.M. 1977. The effects of acid rainfall and heavy metal particulates on a boreal forest ecosystem near the Sudbury smelting region of Canada. Water, Air, Soil Pollut. 7: 421-438. Huttunen, S. 1975. Studies on forest vegetation in air pollution damage area. In: Acta Universitatis Oulensis. Acta Universitatis Oulensis, Series A, Biology 2. University of Oulu, Oulu. pp. 1-37. Hynninen, V. 1986. Monitoring of airborne metal pollution with moss bags near an industrial source at Harjavalta. Annales Botanici Fennici 23: 83-90. Hyvärinen, M., Roitto, M., Ohtonen, R. & Markkola, A. 2000. Impact of wet deposited nickel on the cation content of a mat-forming lichen Cladina stellaris. Environmental and Experimental Botany 43: 211-218. Jonsdottir, I.S. & Callaghan, T.V. 1988. Interrelationships between different generations of interconnected tillers of Carex bigelowii. Oikos 52: 120-128. Jules, E.S. & Shaw, A.J. 1994. Adaptation to metal-contaminated soils in populations of the moss, Ceratodon purpureus: vegative growth and reproductive expression. American Journal of Botany 81: 791-797. Jussila, I. 2003. A bioindicator study on the effects of air pollution on forest ecosystem at Pori-Harjavalta district in SW-Finland (In Finnish, English summary). Sykesarja B 15. Turun yliopisto. Satakunnan ympäristöntutkimuslaitos, Pori. 112 p. Kahle, H. 1993. Response of roots trees to heavy metals. Environmental and Experimental Botany 33(1): 99-119. Kiikkilä, O. 2002. Remediation through mulching with organic matter of soil polluted by a copper-nickel smelter. Finnish Forest Research Institute, research papers. 57 p. Kiikkilä, O. 2003. Heavy-metal pollution and remediation of forest soil around the Harjavalta Cu-Ni smelter, in SW Finland. Silva Fennica 37(3): 399-415. 40 Kiikkilä, 0., Pennanen, T., Pietikäinen, J., Hurme, K.-J. & Fritze, H. 2000. Some observations on the copper tolerance of bacterial communities determined by (3H)-thymide incorporation method in heavy metal polluted humus. Soil Biology & Biochemistry 32: 883-885. Kopponen, P., Utriainen, M., Lukkari, K., Suntioinen, S., Kärenlampi, L. & Kärenlampi, S. 2001. Clonal differences in copper and zink tolerance of birch in metal-supplemented soils. Enviromental Pollution 112: 89-97. Kozlov, M.V., Haukioja, E.. Bakhtiarov, A.V., Stroganov, D.N. & Zimina, S.N. 2000. Root versus canopy uptake of heavy metals by birch in an industrially polluted area: contrasting behaviour of nickel and copper. Environmental Pollution 107: 413-420. Krebs, S.L. & Hancock, J.F. 1989. Tetrasomic inheritance of isoenzyme markers in the highbush blueberry, Vaccinium corymbosum L. Heredity 63: 11-18. Kubin, E., Lippo, H. & Poikolainen, J. 2000. Heavy metal loading. In: Mälkönen, E. (ed.). Forest condition in a changing environment - the Finnish case. Kluwer Academic, Dordrecht, pp. 60-71. Laaksovirta, K. & Silvola, J. 1975. Effect of air pollution by copper, sulphuric acid and fertilizer factories on plants at Harjavalta, W. Finland. Annales Botanici Fennici 12: 81-88. Laine, K., Saari, E., Kemppainen, K., Pakonen, T., Havas, P, Lajunen, L., Perämäki, P, Paama, L. 1993. Lapin metsämarjojen raskasmetallipitoisuudet. Ympäristöjä Terveys 24(7-8): 443-449. Lakanen, E. & Erviö, R. 1971. A comparison of eight extractants for the determination of plant available micronutrients in soils. Acta Agralia Fennica 123: 223-232. Lin, Z.Q., Schuepp, P.H., Schemenauer, R.S. & Kennedy, G.G. 1995. Trace metal contamination and on balsam fir (Abies balsamea (L.) Mill.) foliage in Southern Quebec, Canada. Water, Air, Soil Pollut. 81: 175-191. Luttermann, A. & Freedman, B. 2000. Risks to forests in heavily polluted regions. In: Innes, J.L. & Oleksyn, J. (eds.). Forest dynamics in heavily polluted regions. CABI Publishing, Wallingford. pp. 9-26. MacNair, M., R. 1987. Heavy metal tolerance in plants: a model evolutionary system. Trends in Ecology & Evolution 2(12): 354-358. Mahy, G., Vekemans, X. & Jacquemart, A.-L. 1999. Patterns of allozymic variation whithin Calluna vulgaris populations at seed bank and adult stages. Heredity 82: 432-440. Mäkinen, A. 1994. Biomonitoring of atmospheric deposition in Finland, Estonia and Kola Peninsula, based on the chemical analysis of mosses. Publications from the Department of Botany, University of Helsinki 19. University of Helsinki, 1994, Helsinki. 28 p. Mälkönen, E. 1974. Annual primary production and nutrient cycle in some Scots pine stands. Communicationes Instituti Forestalis Fenniae 84.5: 1-87. Mälkönen, E., Derome, J., Fritze, H., Helmisaari, H.-S., Kukkola, M., Kytö, M., Saarsalmi, A. & Salemaa, M. 1999. Compensatory fertilization of Scots pine stands polluted by heavy metals. Nutrient Cycling in Agroecosystems 55: 239-268. Marschner, H. 1995. Mineral nutrition of higher plants. Academic Press, London. 889 p. Mattila, A., Pakkanen, A., Vakkari, P. & Raisio, J. 1994. Genetic variation in English oak (Quercus robur) in Finland. Silva Fennica 28(4): 251-256. Meharg, A.A. & Cairney, J.W.G. 2000. Co-evolution of mycorrhizal symbionts and their hosts to metal-contaminated environments. Advances in Ecological Research 30: 70-112. Melanen, M., Ekqvist, M., Mukherjee, A.8., Aunela-Tapola, L., Verta, M. & Salmikangas, T. 1999. Atmospheric emissions of heavy metals in Finland in 19905. The Finnish Environment 329. Finnish Environment Institute, Helsinki. 92 p. Mocquot, 8., Vangronsved, J., Clijsters, H. & Mench, M. 1996. Copper toxicity in young maize (Zea mays L.) plants: effects on growth, mineral and chlorophyll contents, and enzyme activities. Plant and Soil: 287-300. 41 Monni, S., Bucking, H. & Kottke, I. 2002. Ultrastructural element localization by EDXS in Empetrum nigrum. Micron 33: 339-351. Monni, S., Salemaa, M. & Millar, N. 2000 a. The tolerance of Empetrum nigrum to copper and nickel. Environmental Pollution 109: 221-229. Monni, S., Salemaa, M., White, C., Tuittila, E. & Huopalainen, M. 2000 b. Copper resistance of Calluna vulgaris originating from the pollution gradient of a Cu-Ni smelter, in southwest Finland. Enviromental Pollution 109: 211-219. Monni, S., Uhlig, C., Hansen, E. & Magel, E. 2001 a. Ecophysiological responses of Empetrum nigrum to heavy metal pollution. Enviromental Pollution 112: 121-129. Monni, S., Uhlig, C., Junttila, 0., Hansen, E. & Hynynen, J. 2001b. Chemical composition and ecophysiological responses of Empetrum nigrum to aboveground element application. Enviromental Pollution 112: 417-426. Nash, T.H. 1996. Nutrients, element accumulation and mineral cycling. In: Nash, T.H. (ed.). Lichen biology. Cambridge Univerisity Press, Cambridge, UK. pp. 136-153. Nei, M., Furerstt, PA. & Chakraborty, R. 1976. Testing the neutral mutation hypothesis by distribution of single locus heterozygosity. Nature 262: 491-493. Nieboer, E. & Richardson, D.H.S. 1980. The replacement of the nondescript term 'heavy metals' by a biologically and chemically significant classification of metal ions. Enviromental Pollution 1: 3-26. Nieminen, T. & Helmisaari, H.-S. 1996. Nutrient retranslocation in the foliage of Pinus sylvestris L. growing along a heavy metal pollution gradient. Tree Physiology 16: 825-831. Nieminen, T.M., Derome, J. & Helmisaari, H.-S. 1999. Interactions between precipitation and Scots pine canopies along a heavy-metal pollution gradient. Environmental Pollution 106: 129-137. Nriagu, J.O. 1989. A global assessment of natural sources of atmospheric trace metals. Nature 338(2): 47-49. okland, T., okland, R.H. & Steinnes, E. 1999. Element concentrations in the boreal forest moss Hylocomium splendens: variation related to gradient in vegetation and local environmental factors. Plant and Soil 209: 71-83. Onianwa, PC. 2001. Monitoring atmospheric metal pollution: a review of the use of mosses as indicators. Environmental Monitoring and Assessment 71: 13-50. Outola, 1., Pehrman, R. and Jaakkola, T. 2003. Effect of industrial pollution on the distribution of l37Cs in soil and the soil-to-plant transfer in a pine forest in SW Finland. The Science of the Total Environment 303: 221-230. Päivänen, J. 1966. Sateen jakautumisesta erilaisissa metsiköissä. Summary: The distribution of rainfall in different types of forest stands. Silva Fennica 119(3): 1-37. Pätsikkä, E., Aro, E.-M. & Tyystjärvi, E. 2001. Mechanism of copper-enhanched photoinhibition in thylakoid membranes. Physiologia Plantarum 113(1): 142-150. Pawlik-Skowronska, 8., di Toppi, L.S., Favali, M.A., Fossati, F., Pirszel, J. & Skowronski, T. 2002. Lichens respond to heavy metals by phytochelatin synthesis. New Phytologist 156: 95-102. Poikolainen, J., Kubin, E., Piispanen, J. & Karhu, J. 2004. Atmospheric heavy metal deposition in Finland during 1985-2000. The Science of the Total Environment 318: 171-185. Prusinkiewicz, P. & Lindenmayer, A. 1990. The algorithmic beauty of plants. Springer, New York. 228 p. Punshon, T. & Dickinson, N.M. 1997. Acclimation of Salix to metal stress. New Phytologist 137: 303-314. Puolanne, J., Pyy, O. & Jeltsch, U. 1994. Saastuneet maa-alueet ja niiden käsittely Suomessa: saastuneiden maa-alueiden selvitys- ja kunnostusprojekti (In Finnish). Ministry of Environment. 218 p. 42 Rautio, P. & Huttunen, S. 2003. Total vs. internal element concentrations in Scots pine needles along a sulphur and metal pollution gradient. Environmental Pollution 122: 273-289. Rautio, P., Huttunen, S., Kukkola, E., Peura, R. & Lamppu, J. 1998. Deposited particles, element concentrations and needle injuries on Scots pine along an industrial pollution transect in northern Europe. Environmental Pollution 103: 81-89. Reimann, C., Koller, F„ Frengstad, 8., Kashulina, G., Niskavaara, H. & Englmaier, P. 2001. Comparison of the element composition in several plant species and their subsrate trom a 1 500 000-km2 area in Northern Europe. The Science of the Total Environment 278: 87-112. Reinikainen, A., Mäkipää, R., Vanha-Majamaa, I. & Hotanen, J.-P. 2000. Kasvit muuttuvassa metsäluonnossa. (English summary: Changes in the frequency and abundance of forest and mire plants in Finland since 1950). Tammi, Helsinki. 384 p. Rigina, O. & Kozlov, M. 2000. The impacts of air pollution on the Northern taiga forests of the Kola Peninsula, Russian federation. In: Innes, J.L. & Oleksyn, J. (eds.). Forest dynamics in heavily polluted regions. CABI Publishing, Wallingford. pp. 37-65. Ross, S.M. 1994 a. Retention, transformation and mobility of toxic metals in soils. In: Ross, S.M. (ed.). Toxic metals in soil - plant systems. John Wiley, Chichester, pp. 63-152. Ross, S.M. 1994b. Sources and forms of potentially toxic metals in soil-plant systems. In: Ross, S.M. (ed.). Toxic metals in soil-plant systems. John Wiley & Sons, Chichester, pp. 3-25. Riihling, Ä. & Steinnes, E. 1998. Atmospheric heavy metal deposition in Europe 1995- 1996. NORD 1998. Nordic Council of Ministers, Copenhagen. 15 p. Salemaa, M., Monni, S. & Nieminen, T. 2003. Dwarf shrubs (Arctostaphylos uva-ursi, Calluna vulgaris, Empetrum nigrum and Vaccinium uliginosum) in the revegetation of a Cu polluted forest area. In: Gobran, G.R. & Lepp, N.W. (eds.). 7th International Conference on the Biogeochemistry of Trace Elements, vol 11, Uppsala, Sweden, pp. 196-197. Sarret, G., Manceau, A., Cuny, D., C., V.H., Deruelle, S., Hazemann, J.-L., Soldo, Y., Eybert- Berard, L. & Menthonnex, J.-J. 1998. Mechanisms of lichen resistance to metallic pollution. Environmental Science & Technology 32: 3325-3330. Schlichting, C.D. 1989. Phenotypic integration and environmental change. What are the consequences of differential phenotypic plasticity of traits? Bioscience 39(7): 460-464. Shevtsova, A. 1998. Responses of subarctic dwarf shrubs to climate change and air pollution. Annales Universitatis Turkuensis. Ser All. 113. University of Turku, Turku. 22 p. Siren, G. 1955. The development of spruce forest on raw humus sites in northern Finland. Acta Forestalia Fennica 62.4: 1-408. Slade, A.J. & Hutchings, M.J. 1987. Clonal integration and plasticity in foraging behaviour in Glechoma hederacea. Journal of Ecology 75: 1023-1036. Streit, B. & Stumm, W. 1993. Chemical properties of metals and the process of bioaccumulation in terrestrial plants. In: Markert, B. (ed.). Plants as biomonitors. Indicators for heavy metals in the terrestrial environment. VCH, Weinheim. pp. 31-62. Stribley, D.P & Read, D.J. 1976. The biology of mycorrhiza in the Ericaceae. VI. The effects of mycorrhizal infection and concentration of ammonium nitrogen on groth of carnberry (Vaccinium microcarpon Ait.) in sand culture. New Phytologist 77: 63-72. Tammiranta, A. 2000. Study of soil contamination in Harjavalta area and assessment of need for measures (In Finnish, English summary). The Finnish Environment 418. Finnish Environmental Institute, Helsinki. 87 p. Tarhanen, S., Metsärinne, S., Holopainen, T. & Oksanen, J. 1999. Membrane permeability response of lichen Bryoria fuscescens to wet deposited heavy metals and acid rain. Environmental Pollution 104: 121-129. 43 Thompson, J.D. 1991. Phenotypic plasticity as a component of evolutionary change. Trends in Ecology and Evolution 6: 246-249. Tikkanen, E. & Niemelä, I. 1995. Kola Peninsula pollutants and forest ecosystems in Lapland. Final Report of the Lapland Forest Damage Project. Gummerus, Jyväskylä. 82 p. Tolvanen, A. & Kubin, E. 1990. The effect of clear felling and site preparation on microclimate, soil frost and forest regeneration at elevated sites in Kuusamo. Aquilo Series Botanica 29: 77-86. Tolvanen, A. & Laine, K. 1997. Effects of reproduction and artificial herbivory on vegetative growth and resource levels in deciduous and evergreen dwarf shrubs. Canadian Journal of Botany 75: 656-666. Tuomi, J., Nilsson, P. & Äström, M. 1994. Plant compensatory responses: bud dormancy as an adaptation to herbivory. Ecology 75(5): 1429-1436. Turner, A.P. & Dickinson, N.M. 1993. Copper tolerance of Acer pseudoplatanus L. (sycamore) in tissue culture. New Phytologist 123: 523-530. Tyler, G. 1984. The impact of heavy metal pollution on forest: a case study of Gusum, Sweden. Ambio 13(1): 18-25. Tyler, G. 1989. Uptake, retention and toxicity of heavy metals in lichens. Water, Air and Soil Pollution 47: 321-333. Tyler, G. 1990. Bryophytes and heavy metals: a literature review. Botanical Journal of Linnean Society 104: 231-253. Tyler, G., Balsberg-Pählsson, A.-M., Bengtsson, G., Bääth, E. & Tranvik, L. 1989. Heavy metal ecology of terrestrial plants, micro-organisms and invertebrates. Water, Air and Soil Pollution 47: 189-215. Uhlig, C., Salemaa, M., Vanha-Majamaa, I. & Derome, J. 2001. Element distribution in Empetrum nigrum microsites at heavy metal contaminated sites in Flarjavalta, Western Finland. Environmental Pollution 112: 435-442. Ukonmaanaho, L., Starr, M., Hirvi, J.-P., Kokko, A., Lahermo, P., mannio, J., Paukola, T., Ruoho-Airola, T. & Tanskanen, H. 1998. Heavy metal concentrations in various aqueous and biotic media in Finnish integrated monitoring catchments. Boreal Environment Research 3: 235-249. Utriainen, M., Kärenlampi, L., Kärenlampi, S. & Schat, H. 1997. Differential tolerance to copper and zinc of micropropagated birches tested in hydroponics. New Phytologist 137: 543-549. Väisänen, S. 1986. Effects of air pollution by metal, chemical and fertilizer plants on forest vegetation at Kokkola, W Finland. Annales Botanici Fennici 23: 303-315. Vangronsveld, J., Colpaert, J.V. & Tichelen Van, K.K. 1996. Reclamation of a bare industrial area contaminated by non-ferrous metals: physicochemical and biological evaluation of the durability of soil treatment and revegetation. Environmental Pollution 94: 131-140. Verkleij, J.A.C. & Schat, H. 1989. Mechanisms of metal tolerance in higher plants. In: Shaw, J.A. (ed.). Heavy metal tolerance in plants: evolutionary aspects. CRC Press, Boca Raton, pp. 179-193. Vuorisalo, T. & Tuomi, J. 1986. Unitary and modular organisms: criteria for ecological division. Oikos 47(3): 382-385. Zechmeister, H.G., Grodzinska, K. & Szarek-Lukaszewska, G. 2003. Bryophytes. In: Markert, 8.A., Breure, A.M. & Zechmeister, H.G. (eds.). Bioindicators and biomonitors. Elsevier, Amsterdam, pp. 329-375. Zöttl, H.W. 1985. Heavy metal levels and cycling in forest ecosystems. Experientia 41 1104-1113. Paper I Salemaa, M., Vanha-Majamaa, I. & Derome, J. 2001. Understorey vegetation along a heavy-metal pollution gradient in SW Finland. Environmental Pollution 112: 339-350. Environmental Pollution 112 (2001) 339-350 0269-7491/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved, PII: 50269-7491(00)00150-0 Understorey vegetation along a heavy-metal pollution gradient in SW Finland M. Salemaa3 '*, I. Vanha-Majamaa 3 , J. Derome b a Vantaa Research Centre, Finnish Forest Research Institute, PO Box 18, FIN-01301 Vantaa, Finland bßovaniemi Research Station, Finnish Forest Research Institute, PO Box 16, FIN-96301 Rovaniemi, Finland Received 8 February 2000; accepted 30 May 2000 "Capsule": Heavy metals and sulphur in soils decreased andfloristic diversity increased with distance from a Cu-Ni smelter Abstract Understorey vegetation of Scots pine forests was studied along a 8-km transect running SE from a Cu-Ni smelter at Harjavalta, SW Finland. Long-term accumulation of heavy metals and sulphur in the forest ecosystem has drastically changed plant commu nities. Vegetation was almost absent up to a distance of 0.5 km from the smelter. The total coverage and the number of plant species increased with increasing distance from the smelter. Ordination by global non-metric multidimensional scaling (GNMDS) indicated that the floristic composition was differentiated in response to the pollution level. The main compositional gradient of GNMDS was correlated with the heavy metal concentrations in the organic soil layer and with the size of the overstorey trees. Vascular plants were more pollution-resistant than ground lichens, whereas mosses were the most sensitive plant group. In addition to heavy metals, nutrient imbalances and the considerably reduced water-holding capacity of the surface soil also restrict plant recolonisation on the degraded sites. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Boreal forest vegetation; Heavy metals; Sulphur; Non-metric multidimensional scaling; Point quadrat method 1. Introduction Heavy metals are emitted from traffic and industrial sources into the atmosphere and spread over wide regions of the boreal forest zone. Although heavy metal deposition has decreased considerably during the last 20 years in northern Europe (Ruhling et al., 1992; Mäki nen, 1994; Berthelsen et al., 1995; Ukonmaanaho et al., 1998; Kubin et al., 2000), there are many areas where the long-term accumulation of heavy metals, often associated with exposure to sulphur dioxide, has damaged forests ecosystems. For instance, there are extensive damage areas surrounding the metallurgical industry complexes in the Kola Peninsula, NW Russia (Kozlov et al., 1993; Tikkanen and Niemelä, 1995; Chernenkova and Kuperman, 1999; Rigina and Kozlov, * Nomenclature as per Hämet-Ahti et ai. (1998) (vascular species), Koponen et ai. (1977) (bryophytes) and Vitikainen et ai. (1997) (lichens). * Corresponding author. Tel.: +358-9-8570-5565. E-mail address: maija.salemaa@metla.fi (M. Salemaa). 2000). Localised damage areas are also to be found in the Nordic countries, primarily close to smelters, mines or steel mills (Väisänen, 1986; Folkeson and Andersson- Bringmark, 1988; Ruhling et al„ 1992). Coniferous trees efficiently filter pollutant particles from the air (Tyler, 1984; Zöttl, 1985; Fowler et al., 1989). Stand throughfall and plant litter increase the load of heavy metals and sulphur on the forest floor (Heinrichs and Mayer, 1980; Derome and Nieminen, 1998; Nieminen et al., 1999). Heavy metals depress soil microbial activity (Bääth, 1989; Pennanen et al., 1996), which is seen as an increase in undecomposed needle litter on the soil surface and retarded nutrient cycling of the whole ecosystem (Fritze et al., 1997). The accumu lation of Cu and Ni in forest soil can also cause a defi ciency of base cations as a result of their displacement by heavy metals (Lobersli and Steinnes, 1988; Derome and Lindroos, 1998). The toxic effects of heavy metals and changes in the nutrient status of the soil subject the vegetation to a strong selective pressure. The above-ground biomass of the understorey vege tation is small in relation to that of the trees, but it plays 340 M. Salemaa et al. I Environmental Pollution 112 (2001) 339-350 an important role in regulating the nutrient fluxes (Mälkönen, 1974), hydrology and micro climate (Axels son and Bräkenhielm, 1980; Tolvanen and Kubin, 1990) of boreal forests. In contaminated environments, certain plant species can accumulate considerable amounts of heavy metals and protect the soil from erosion and the leaching of heavy metals into the groundwater (Van gronsveld et al., 1996). Because many heavy metals are bound on plant surfaces and tissues (Wittig, 1993), and form stable complexes with organic matter in the soil (Alloway, 1995), they may still have a long-lasting effect on forest ecosystems after the emissions have ceased (Ross, 1994). Before we can gain a better understanding of heavy metal fluxes in boreal forests, more informa tion is needed about the importance of the understorey vegetation in ecosystem processes. In this study, we describe how the structure of under storey vegetation changes along a heavy metal and sul phur gradient near a Cu-Ni smelter in SW Finland. Our aims were (1) to relate the changes in the vegetation to the element concentrations in the organic soil layer and stand structure, (2) to evaluate the sensitivity of plant species to heavy metals according to their occurrence in polluted soil, and (3) to compare the abundances of plant species in the contaminated stands with data from background areas. 2. Materials and methods 2.1. Study area The study area is situated near the Cu-Ni smelter at Harjavalta (61°19' N, 22°09'E), SW Finland. The sam ple plots were established in Scots pine stands along a 8- km transect running to the SE of the smelter. This was the only direction with an unbroken corridor of con iferous forest near the smelter. Apart from the pollution level, the tree stand and site type at the individual plots were originally relatively similar. The study transect has probably not been exposed to the heaviest deposition levels because southerly winds are somewhat dominant (18% of the time) in the area (Derome, 2000 a). All the stands were 40-50 years old and growing on dry and relatively infertile sites of the Calluna site type (Cajander, 1949) in the southern boreal coniferous zone (Ahti et al., 1968) (Table 1). The soil was fine sand and was classified as orthic podsol. A number of research projects, e.g. fertilisation experiments (Derome and Saarsalmi, 1999; Mälkönen et al., 1999; Derome, 2000b), and studies on soil microbiology (Fritze et al., 1989, 1997) and nutrient fluxes (Helmisaari et al., 1995; Nieminen and Helmi saari, 1996), have already generated a considerable amount of ecological information about the transect. The Outokumpu Harjavalta Metals smelter complex is one of the largest point sources of heavy metal emis sions in Finland. Its main products are copper, nickel and sulphuric acid. The copper smelter has been oper ating since 1945, and the nickel smelter since 1960. Before the sulphuric acid plant was built in 1947, all the S0 2 produced during the smelting process (annually about 30 000 t) was emitted into the atmosphere, caus ing severe damage to the surrounding coniferous for ests. During 1985-1990, the average annual emissions of Cu from the smelter were 104 t, Ni 50 t, Zn 177 t, S0 2 8100 t and dust 1200 t. During the past few years, however, the emissions have been considerably reduced: in 1993, Cu 50 t, Ni 7 t, S02 4700 t and dust 250 t (Helmisaari et ai., 1995). The Kemira fertiliser factory produced superphosphate and PK fertilisers at Harja valta from 1948 to 1989. Most of the heavy metals have been deposited rather close to the smelter owing to the fact that the smelter stacks were relatively low (70 m) up until 1994, when a 140-m high stack was built. Copper deposition in stand throughfall during 1993-1996 was 369 mg/m 2 at 0.5 km, 12 mg/m 2 at 4 km and 3 mg/m 2 at 8 km. The corre sponding values for Ni were 138, 2 and 1 mg/m 2 and for S 1816, 464 and 376 mg/m 2 (Derome and Nieminen, 1998). The 24-h mean S0 2 concentrations in the air have decreased from the level of 38-56 |! g/m 3 in 1987 (1 January-30 June) to 17-18 n g/m 3 in 1992 (21 January -15 May) within a 1 km radius of the smelter. However, the peak hourly concentrations in 1992 still occasionally reached 500-1000 |i g/m 3 (Saari et al., 1993). 2.2. Sampling design and vegetation analysis The understorey vegetation was studied at six loca tions (0.5, 1, 2, 3, 4 and 8 km) along the study transect in August 1993. The vegetation analysis was carried out on three sample plots (30 x 30 m) at each distance (Table 1). Each sample plot was divided into four sub plots, and stratified random sampling performed on 16 vegetation quadrats (1 m 2). Three persons assessed the coverage of plant species using a modification of the point quadrat method (Goodall, 1952). The frame (1 m 2) was divided into 10 x 10 cm squares by means of strings stretched to form grids at two levels. The intersections of the strings formed 81 points; the centre point was excluded because it was placed against the marking tube fixed in the ground. A metal pin (diameter 2 mm) was inserted ver tically into the ground at each of the 80 intersection points. Touches between the pin and the same plant species were counted. The point frequency% of each species was used in the statistical analyses. One point was equivalent to a coverage of 1.25% (1/80 x 100). Species occurring in the quadrat but not touched by the pin were recorded as 0.5%, and species occurring on the sample plot but absent in the quadrat as 0.01%. The point frequency method slightly overestimates the spe M. Salemaa et ai / Environmental Pollution 112 (2001) 339-350 341 Table 1 General characteristics of the study sites a a The stand data are from the year 1992 and the other measure ments from years 1992 and 1993. CT, Calluna type, + indicates slightly more fertile type. Ground layer variables are averages of the three sample plots. Coverages of needle litter, decaying wood and trampled area were visually estimated over the whole plot. Stand data according to Mälkönen et ai. 1999. cies coverages as compared to e.g. visual estimates (Vanha-Majamaa et ai., 2000). Vegetation data collected from the permanent sample plots of national forest inventory (NFI) in Finland (1995) were used as the reference level. The reference data comprised 12 sample plots representing 40- to 65- year-old Calluna type pine forests in the southern and middle boreal zones of Finland. Species coverages were visually estimated on 3-4 quadrats (2 m 2 each) on cir cular sample plots (300 m 2). 2.3. Chemical composition of the organic soil layer Samples were taken from the organic layer at 25 sys tematically selected points on each of the same three plots as the vegetation analysis at distances of 0.5, 2, 4 and 8 km in May 1992. Corresponding soil samples were taken from one plot at distances of 1 and 3 km in September 1993. The samples were dried and milled to pass through a 1-mm sieve. pH was measured in water. Organic matter (om) content was determined by loss in weight on ignition by ashing the samples in a muffle furnace at 550° C for 3 h. Exchangeable Cu, Ni, Zn, Fe, Mn, Cd, Pb, Ca, Mg, K and extractable P and S were determined by extraction with 1 M ammonium acetate (pH 4.65)+ 1% EDTA, followed by analysis by induc tively coupled plasma atomic emission spectrometry. Total N was determined on a CHN analyser. EDTA was used in the extractant because a chelating agent is known to be necessary when extracting high concentra tions of Cu and Fe from organic material. All element concentrations are expressed with respect to the organic matter content, in order to reduce the variation arising from the inclusion of mineral soil in the organic layer samples. Removal of the organic layer was problematic at some of the sampling sites. 2.4. Data analysis The vegetation data were ordered by GNMDS (DECODA 2.04 software of Minchin, 1991) to relate the vegetation composition to the chemical soil proper ties and stand structure [see Minchin (1987) and Kent and Coker (1992) for the method]. Species occurring only once were excluded from the analysis. The Bray- Curtis (Czekanowski) coefficient was used as a measure of dissimilarity in floristic composition between the sample plots. The ordination distances were scaled to half-change units. Two- and three-dimensional GNMDS solutions were carried out with 10 randomly generated starting configurations. All 10 iterations gave effectively identical minimum stresses (two-dimensional: 0.058, three-dimensional: 0.028) within both solutions. One iteration was arbitrarily selected for the further analysis. Maximum (canonical) correlations between the environmental variables and the ordination configura tion were calculated using the vector fitting procedure of DECODA. The significance of the correlations were assessed with Monte Carlo tests. Because the third dimension did not provide new information, the two dimensional solution was selected. The weighted avera ges of the species were calculated in the ordination space. The relationships between the distance from the smelter and the plotwise means of environmental vari ables, species richness ( S), indices of Shannon diversity (//' = £i=i(Pi x Inpi), s= number of species, pi pro portion of the /th species of the total abundance) and species evenness (E = H'/\nS), as well as the total abundance of the understorey, were calculated by Ken dall non-parametric correlations (SAS, 1994). Dom inance-diversity curves were drawn to display the log scaled abundance of the fifteen most abundant species on the sample plots against their ranking from most to least abundant. Kruskal-Wallis non-parametric analysis of variance (SAS, 1994) was used in comparing species abundances at a distance of 8 km and in the NFI refer ence data. 3. Results Clear differences in the distribution pattern of the plant species and in the chemical composition of the organic layer were observed on moving towards the smelter (Fig. 1). The total number of plant species increased from 8 at 0.5 km to 30 at 8 km (Table 2). The total coverage of the vegetation also increased with Distance from the smelter (km) 0.5 1 2 3 4 8 Stand Forest site type CT + CT + CT + CT CT CT Stand age, years 49 67 54 51 48 40 Mean pine height (m) 6.1 10.1 10.9 9.5 9.2 10.6 Mean pine diameter (cm) 10.9 13.6 13.7 12.3 10.3 13.1 Stem volume (m 3 /ha) 23.2 98.0 85.3 69.3 67.8 94.5 Number of trees/ha 1008 1048 1230 1436 1517 1552 Volume increment (1981-1990) m 3 ha" 1 year" 1 0.31 - 3.78 2.78 6.27 Ground layer Thickness of organic layer (cm) 2.5 7.2 2.3 3.3 2.2 0.8 Coverage of needle litter (%) 83.3 90.3 74.3 85.0 33.0 31.7 Coverage of decaying wood (%) 8.3 7.6 15.7 5.7 9.7 5.3 Trampled area (%) 14.7 8.7 26.7 9.3 14.3 14.0 342 M. Salemaa et ai. / Environmental Pollution 112 (2001) 339-350 Fig. 1. The mean point frequencies (%) of the plant species on the sample plots (columns) at different distances from the smelter (km). Species observed on the sample plots but not on the sample quadrats are marked with x. The line indicates SEM. The distributions of exchangeable Cu and Ni in the organic layer are given in the upper left panel. Note the different scales on the y axes. increasing distance from the smelter (Tables 2 and 4). A few species accounted for most of the abundance of the communities near the smelter (0.5-2 km), but at greater distances (3-8 km) the species distributions were more even (Fig. 2). At 0.5 km, the exchangeable Cu concentration of the organic layer was 36 times higher (7540 mg/kg om) than that at 8 km (200 mg/kg om). The gradients were not as steep for the other heavy metals, e.g. the exchangeable Ni concentrations were 527 mg/kg om at 0.5 km and 72 mg/kg om at 8 km (Fig. 1, Table 3). 3.1. Vegetation pattern related to environmental gradients In general, the ordination configuration was highly related to the heavy metal concentrations in the organic layer and to the stand structure in GNMDS (Fig. 3). The sample plots were positioned according to the dis tance from the smelter along the main compositional gradient, which runs parallel to the first dimension. The species scores were arranged in three groups: (1) sensi tive species (present only at greater distances from the smelter e.g. Pleurozium schreberi), (2) resistant species which grew on both polluted and less-polluted sample plots (e.g. Pohlia nutans), and (3) resistant species, the occurrence of which near the smelter was restricted to paludified microhabitats (e.g. Vaccinium uliginosum). The correlations of the exchangeable heavy metal (Cu, Ni, Fe, Pb) and sulphur concentrations in the organic soil layer and the ordination space were highly significant (P < 0.001), and showed an increasing trend with decreasing distance from the smelter (Fig. 3a, Table 4). An opposite trend was found for some of the macro-nutrients (e.g. Ca, K and Mg), as well as for Mn. The mean height and the number of the overstorey trees increased with increasing distance from the smelter (Fig. 3a, Table 4). The diversity and the total abundance of the plant species also increased as a function of increasing distance (Table 4). The initial heterogeneity among the sample plots was reflected in the second dimension. Mire species (e.g. Ledum palustris and Carex M. Salemaa et ai. / Environmental Pollution 112 (2001) 339-350 343 (continued on next page) Table 2 Plotwise means for point frequency% of the plant species at different distances from the smelter a ,b Distance from the smelter (km) Reference Species 0.5 1 2 3 4 8 NF1 95 Vascular plants Andromeda polifolia L. 0.00 0.00 0.00 0.00 0.01 0.00 O.lliO.ll Arctostaphylos uva-ursi (L.) Sprengel 0.00 0.00 0.47 1.95 0.00 0.01 0.04±0.04 Be tula pendula Roth. 0.00 0.00 0.01 0.00 0.00 0.00 O.OOiO.OO Be tula pubescens Ehrh. 0.01 0.01 0.01 0.00 0.00 0.01 0.15±0.1 1 Calluna vulgaris (L.) Hull 0.00 0.01 0.01 0.01 1.35 1.41 12.74±3.09 * Car ex globular is L. 0.01 13.13 0.01 0.00 0.01 0.00 0.09±0.05 Deschampsia flexuosa (L.) Trin. 0.00 0.01 0.00 0.00 0.00 0.01 1.18±0.83 Empetrum nigrum L. 0.01 1.34 3.20 3.20 6.89 0.36 1.53±0.81 ns Epilobium angustifolium L. 0.00 0.01 0.01 0.01 0.00 0.01 0.04±0.04 Ledum palustre L. 0.00 1.72 0.00 0.00 0.01 0.00 0.03±0.03 ns Picea abies (L.) H. Karst. 0.00 0.01 0.01 0.01 0.00 0.01 0.04±0.04 Pinus sylvestris L. 0.44 0.22 0.76 0.77 1.24 1.22 0.06±0.03 ** Populus tremula L. 0.00 0.00 0.00 0.01 0.00 0.00 0.06±0.04 Salix aurita L. 0.01 0.00 0.00 0.00 0.00 0.00 O.OOiO.OO Vaccinium myrtillus L. 0.00 0.52 0.07 0.00 0.00 0.03 0.79±0.47 ns Vaccinium uliginosum L. 0.01 2.84 0.00 0.00 1.04 0.00 2.52±1.61 ns Vaccinium vitis-idaea L. 0.00 1.21 3.29 7.79 7.90 6.46 7.94±1.34 ns Other species 0.09±0.06 Total abundance per plot 0.48 21.01 7.83 13.73 18.44 9.52 27.37±4.22 ns Number of species per plot 4.00 9.00 8.00 5.67 5.67 5.67 Total number of species per distance 6 12 11 8 8 10 Bryophyta Ceratodon purpureus (Hedw.) Brid. 0.01 0.05 0.01 0.00 0.00 0.00 O.OOiO.OO Dicranum polyseturn Sw. 0.00 0.00 0.00 0.00 0.00 6.65 3.67±0.93 ns Dicranum scoparium Hedw. 0.00 0.00 0.00 0.00 0.03 0.39 0.68±0.36 ns Ortodicranum montanum (Hedw.) Loeske 0.00 0.00 0.01 0.00 0.00 0.00 O.OOiO.OO Pleurozium schreberi (Brid.) Mitt. 0.00 0.00 0.01 0.00 0.07 8.72 30.39i6.06 ns Pohlia nutans (Hedw.) Lindb. 0.17 2.89 2.02 9.83 3.42 2.25 0.17i0.10** Polytrichum juniperinum Hedw. 0.00 0.31 0.01 0.00 0.01 0.12 0.92i0.66 Ptilidium ciliare (L.) Hampe 0.00 0.00 0.01 0.08 0.38 0.07 0.02i0.02 Other species 2.53il.07 Total abundance per plot 0.17 3.26 2.06 9.91 3.90 18.21 38.46±6.44 ns Number of species per plot 1.33 2.33 3.00 1.67 3.67 5.67 Total number of species per distance 2 3 6 2 5 6 Ground lichens Cetraria islandica (L.) Ach. 0.00 0.00 23.91 6.41 9.01 7.04 0.48i0.32 ** Cladina arbuscula (Wallr.) Hale&W.L.Club. 0.00 0.00 0.04 4.57 10.22 10.36 9.52i2.73 ns Cladina rangiferina (L.) Nyl. 0.00 0.00 0.08 2.37 19.14 8.45 10.21i3.37 ns Cladina stellaris (Opiz) Brodo 0.00 0.00 0.00 0.03 9.99 30.10 5.40±4.42 * Cladonia bellidiflora (Ach.) Schaer. 0.00 0.00 0.00 0.00 0.01 0.00 O.OOiO.OO Cladonia botrytes (K.G. Hagen) Willd. 0.00 0.00 0.01 0.00 0.01 0.04 O.OOiO.OO Cladonia cenotea (Ach.) Schaer. 0.00 0.00 0.00 0.00 0.04 0.00 0.01i0.01 Cladonia chlorophaea coll. (Flörke ex Sommerf.) Spreng. 0.00 0.00 0.30 3.39 2.25 0.59 0.04±0.02 Cladonia coccifera (L.) Willd. 0.00 0.00 0.00 0.00 0.01 0.00 O.OOiO.OO Cladonia coniocraea (Flörke) Spreng. 0.00 0.00 0.12 0.07 0.21 0.08 O.OOiO.OO Cladonia cornuta (L.) Hoffm. 0.00 0.00 1.03 1.99 1.60 0.47 0.04i0.02 Cladonia crispata (Ach.) Flot. 0.00 0.00 0.00 0.35 0.29 0.21 0.01 ±0.01 Cladonia deformis L. (Hoffm.) + Cladonia sulphurina (Michx.) Fr. 0.00 0.00 0.08 0.55 0.89 0.27 0.07i0.06 Cladonia digitata (L.) Hoffm. 0.00 0.00 0.04 0.30 0.13 0.05 O.OOiO.OO Cladonia fimbriata (L.) Fr. 0.00 0.00 0.00 0.03 0.05 0.05 O.OOiO.OO Cladonia furcata (Huds.) Schrad. + Cladonia turgida Hoffm. 0.00 0.00 0.03 0.14 0.03 0.01 O.OOiO.OO Cladonia gracilis (L.) Willd. ssp. gracilis (Flörke) + ssp. turbinata (Ach.) Ahti 0.00 0.00 0.31 2.24 1.02 0.29 0.03±0.02 M. Salemaa et ai. I Environmental Pollution 112 (2001) 339-350 344 Table 2 (continued) a n= 3, 16 x lm 2 quadrats per plot. b Reference data are based on the mean% coverages (± SEM) of the species at 12 sample plots (3-4 x 2 m 2 quadrats). Mean total abundances and number of species per plot are given. Abundances of the most abundant species are compared between 8 km and the reference data using Kruskal-Wallis tests ( ** =p < 0.01, *= P < 0.05, °= P < 0.10, ns = not significant). globularis) and a thick organic layer with low total N and pH were typical of the sample plots at 1 km. High exchangeable Zn, Cd and S concentrations were also found at 1 km (Fig. 3, Table 3). 3.2. The damage areas Four damage areas were distinguished along the pol lution gradient on the basis of the understorey vegeta tion, the condition of the overstorey trees and the element concentrations of the organic soil layer. 3.2.1. Area of severe damage (0.5 and 1 km from the smelter ) The sample plots at 0.5 and 1 km distances from the smelter deviated the most from the other plots (Fig. 3). The understorey vegetation was almost totally dead up to a distance of 0.5 km. Most of the overstorey was alive, but the pines were stunted (Table 1), damaged and suffering from needle discolouration. The forest floor was covered with relatively undecomposed needle litter and the organic layer was thin (Table 1). Lichens and mosses were absent, except for the resistant pioneer mosses Pohlia nutans and Ceratodon purpureus. Only a few patches of Empetrum nigrum ssp. nigrum, Carex globularis, and Vaccinium uliginosum were present (Fig. 1 and Table 2). The two latter species were frequently growing in paludified depressions. Some saplings of Pinus sylvestris and Betula pubescens were also found. The surviving vascular plants were suffering from a wide range of injury and damage, e.g. dead branches and leaf discolouration. The sample plots at 1 km differed to some extent from the other plots in having a thick organic layer and older tree stand (Fig. 3a, Table 1). One plot had almost no Table 3 pH, total N (%) and exchangeable concentrations of macro-elements and heavy metals in the organic soil layer a Results are expressed on the basis of the organic matter content (om). Averages from three plots are presented for distances of 0.5, 2,4. and 8 km, and from one plot at 1 and 4 km. vegetation cover, and resembled the plots at 0.5 km. However, the total number of vascular plant species was the highest (n= 12) at 1 km distance, Carex globularis and Vaccinium uliginosum being the most abundant species (Fig. 1, Table 2). These species, as well as Ledum palustre, indicated paludification of the site. Both 0.5 km and 1 km sites were classified in the severe damage class owing to the low number of moss and lichen species ( < 5), low total abundance of vege Species Distance from the smelter (km) 0.5 1 2 3 4 8 Reference NFI 95 Cladonia phyllophora Hoffm. 0.00 0.00 0.00 0.03 0.00 0.00 0.00±0.00 Cladonia uncialis (L.) F.H.Wigg. 0.00 0.00 0.00 0.20 0.01 0.00 0.01 ±0.01 Cladonia squamosa (Scop.) Hoffm. 0.00 0.00 0.00 0.01 0.05 0.00 0.00±0.00 Stereocaulon paschale (L.) Hoffm. 0.00 0.02 0.01 0.16 0.00 0.00 0.01 ±0.01 Total abundance per plot 0.00 0.02 25.94 22.81 54.96 58.02 26.04±5.93 ° Number of species per plot 0.00 1.00 8.33 13.00 14.67 12.67 Total number of species per distance 0 1 12 17 19 14 All species Total abundance per plot 0.65 24.29 35.97 46.50 77.38 85.77 91.86±8.20 ** Number of species per plot 5.33 12.00 19.33 20.33 24.00 24.00 Total number of species per distance 8 16 29 27 32 30 Distance from the smelter (km) 0.5 1 2 3 4 8 pH (H 20) 3.84 3.34 3.66 3.62 3.56 3.62 Macro nutrients tot N% om 1.92 1.31 1.74 1.58 1.65 1.68 P mg/kg om 845 387 274 370 196 213 K mg/kg om 168 485 445 586 623 607 Mg mg/kg om 51 215 157 180 188 163 S mg/kg om 387 402 198 176 170 176 Ca mg/kg om 594 1563 1926 1868 2000 1722 Heavy metals Fe mg/kg om 10899 5792 3800 3648 1902 1659 Cu mg/kg om 7540 3501 2238 1532 786 209 Ni mg/kg om 527 406 329 230 164 72 Zn mg/kg om 167 308 144 139 153 82 Pb mg/kg om 468 177 230 196 136 122 Mn mg/kg om 12.5 59.0 135.5 148.7 75.1 91.0 Cd mg/kg om 2.83 4.74 2.59 2.97 2.04 1.09 M. Salemaa et ai. I Environmental Pollution 112 (2001) 339-350 345 Table 4 Plotwise non-parametric Kendall correlations (tau) between the dis tance from the smelter and the environmental and plant community variables 3 a Number of sample plots, «=lB. P = statistical significance (*** = /> < 0.001, ** = P < 0.01 and * = P < 0.05). tation (< 25%) (Table 2) and high exchangeable heavy metal and extractable sulphur concentrations in the organic layer (Table 3). The Cu, Fe and Pb concentra tions at 0.5 km were about double those at 1 km. In contrast, the Zn and Cd concentrations at 1 km were strongly elevated compared to the other distances. The concentrations of exchangeable Ca and K, and espe cially Mg, at 0.5 km were much lower than those at the other sites. 3.2.2. Area of moderate damage (2 and 3 km) The growth of the tree stand was considerably better and the total coverage of the understorey vegetation was higher (Table 2) at distances of 2 and 3 km than those at 0.5 and 1 km. Many species typical of dry heath forests (e.g. Calluna vulgaris, Empetrum nigrum and Vaccinium vitis-idaea) were present, but their abun Fig. 2. Dominance-diversity curves of the log-transformed abun dances (point frequency %) of the 15 most common species on sample plots at different distances from the smelter. dances were much lower than those in the NFI reference data (Table 2). The understorey vegetation was not closed (Table 2) and the coverage of needle litter was still 70-80% (Table 1). The abundance of Cetraria islandica was highest (24%) at 2 km, and Cladonia spp. (cup lichens) at 3 km (Fig. 1, Table 2). The first speci mens of Cladina arbuscula and C. rangiferina were recorded at 2 km, and of C. stellaris and Cladonia uncialis at 3 km. Ptilidium ciliare and Ortodicranum montanum were two moss species recorded for the first time. The abundance of Pohlia nutans was exceptionally high. Arctostaphylos uva-ursi appeared for the first time at 2 km. The exchangeable Cu and Fe concentrations were considerably lower than those at the closest sites, but still much higher than those at greater distances from the smelter (Table 3). The exchangeable Mn concentra tions at these two sites were at their highest values along the transect, but the macronutrient concentrations at a relatively normal level. 3.2.3. Area of slight damage (4 km) The floristic composition at 4 km resembled that of normal dry heath forests. However, typical mosses were still missing or were, as in the case of Pleurozium schre beri and Dicranum spp., very scarce (Table 2). There were no large gaps in the ground vegetation, and the coverage of needle litter was only 30% (Table 1). The highest number of cup lichen species (n = 13) was recor ded at this distance, the most abundant being Cladonia chlorophaea coll., C. cornuta, C. deformis and C. sul phurina group and C. gracilis ssp. turbinata (Table 2). Heavy metal concentrations in the organic layer, apart from Zn, were much closer to those at 8 km (Table 3). Distance from the smelter tau P Organic soil layer Thickness -0.469 0.010** PH 0.185 0.314 Element concentrations N -0.021 0.908 P -0.703 0.000*** K 0.532 0.004** Mg 0.135 0.462 S -0.547 0.003** Ca 0.433 0.018* Fe -0.845 0.000*** Cu -0.958 0.000*** Ni -0.916 0.000*** Zn -0.490 0.008** Pb -0.674 0.001*** Mn 0.419 0.022* Cd -0.504 0.006** Over storey trees Stand age -0.333 0.079 Mean diameter 0.174 0.336 Mean height 0.327 0.070 Number of trees 0.475 0.009** Stem volume 0.216 0.233 Ground layer Trampled area (%) -0.099 0.589 Needle litter (%) -0.596 0.001*** Decaying wood (%) -0.159 0.393 Plant community Total abundance 0.566 0.002** Species richness, S 0.733 0.000*** Evenness, E 0.188 0.299 Diversity, ff 0.508 0.005** 346 M. Salemaa et ai / Environmental Pollution 112 (2001) 339-350 Fig. 3. Two-dimensional solution of GNMDS: (a) Ordination of the sample plots and the fitted vectors of selected environmental variables; (b) weighted averages of the species. The arrows visualize the correla tions between the environmental variables and the sample plot ordi nation. The length of the arrow indicates the magnitude of the correlation, and the direction of the arrow the polarity. Abbreviation of species names = first four letters from generic and species names (see Table 2). km = distance from the smelter. Cup lichens (Cladcol) were combined. The most abundant cup lichens (Cladonia cornuta, C. gra cilis ssp. turbinata, C. deformis (sulphurina), C. chlorophaea and C. botrytes are depicted in the lower left corner of the species ordination figure. 3.2.4. Area of minimum disturbance (8 km) The abundance of some species such as Vaccinium vitis-idaea approached the typical coverage for dry heath forests, but that of Calluna vulgaris, which is one of the characteristic species of dry boreal heaths, was still low (< 2%). In the NFI reference material its cov erage was significantly higher (Kruskall-Wallis 1-way ANOVA: F=5.35, P < 0.05, df= 1), being over 10% (Table 2). The abundances of Dicranum spp. mosses and Cladina arbuscula and C. rangiferina were rather typical, but that of Pleurozium schreberi to some extent lower (P < 0.11). In contrast, the abundance of C. stellaris was higher (P < 0.05) than that in the NFI material. The total abundance of all species was 86% at 8 km, which was lower (P < 0.01) than that in the reference (92%). The heavy metal concentrations were much lower than at sites closer to the smelter (Table 3), but higher than those for background areas (e.g. 7.4 mg/kg Cu and 5.1 mg/kg Ni in the organic soil layer; Tamminen and Starr, 1990). 4. Discussion The understorey vegetation near the smelter was more damaged than the overstorey trees. Compared to damage areas around much larger industrial complexes in the boreal forest zone, e.g. in Sudbury, Ontario (Freedman and Hutchinson, 1980; Winterhalder, 2000) or in Monchegorsk, the Kola Peninsula (Rigina and Kozlov, 2000), where the tree stands are dead over wide areas, there was no "industrial desert area" at Harja valta. The majority of pines growing in the immediate surroundings of the smelter died during the 19405, as a result of the high SOi emissions from the new smelter. The present stand was planted about 50 years ago. Today the only treeless area, restricted in size, is to be found on the opposite side of the study transect. It is clear that, in addition to the deposition of heavy metals, high SOo emissions have also earlier had a marked effect on the vegetation. In the 19905, however, the ambient SO2 concentrations have decreased, and the heavy metals that have accumulated in the soil most likely represent the greatest threat to plant vigour. However, occasional peak S02 concentrations may damage the most sensitive plant species, e.g. mosses and lichens (Bates, 1992). Nutrient imbalances (Derome and Lindroos, 1998) and a decreased water-holding capacity of the soil (Derome and Nieminen, 1998) have strengthened the selection pressure on plants growing in the most polluted area. The increased illumination in sparse, defoliated stands has also had a strong effect on plant establishment. Owing to strong abiotic control, a few tolerant species were dominant in the ground vegetation on the most polluted sites. It is probable that between-species com petition was insignificant in these harsh conditions, where the total coverage of the vegetation was low. The increased heterogeneity of the environment offered more niches to a greater number of species further away from the smelter. Competition had a more profound effect when the vegetation was closed. As a con sequence, the total coverage was divided more evenly between the species (Fig. 2). Mosses and lichens cannot prevent the passage of heavy metals or other toxic ions into their shoots or thalli, because they absorb nutrients directly from rain water or the air, and have no protective cuticle (Nash, 1989; Tyler, 1990; Bates, 1992). In addition to elements derived from the atmosphere, mosses also take up ions M. Salemaa et al. / Environmental Pollution 112 (2001) 339-350 347 from water that has been in contact with the organic layer of the soil (okland et al., 1999). Vascular plants, in contrast, can to some extent select which elements are taken up by their roots. In many cases, however, the element concentrations inside roots closely follow those in the substrate, especially when the exclusion mechan isms break down in high concentrations (Kahle, 1993; Punz and Sieghardt, 1993; Monni et al., 2000a,b). The rooting depth of the dwarf shrubs growing in boreal forests is rather shallow, the majority of the roots growing in the organic and uppermost (0-10 cm) mineral soil layers (Makkonen and Helmisaari, 1998). Empetrum nigrum clones surving at a distance of 0.5 km from the Harjavalta smelter were an exception to this: their roots extended down to a depth of 50 cm into cleaner soil horizons (Uhlig et al., 2001). Compared to coniferous trees, dwarf shrubs generally have a more superficial root system, which may increase their expo sure to the high concentrations of heavy metals in the topsoil. Without experimental exposures, it is difficult to identify the contribution of individual pollutants to the injuries of plant species. The biological availability of heavy metals to plants depends on the soil character istics (e.g. the amount of organic matter, pH and the chemical composition) (Balsberg Pählsson, 1989; Kahle, 1993; Punz and Sieghardt, 1993). Because different metal ions have antagonistic, additive or synergistic interactions, their actual absorption rate and toxicity may be modified according to the overall element com position in the soil (Kahle, 1993). The general responses of the lichen and moss com munities along the Harjavalta transect were very similar to those described in the coniferous forests near the Cu- Zn smelter in Gusum, SE Sweden, although no SO2 was emitted from the latter point source (Folkeson, 1984; Tyler, 1984; Folkeson and Andersson-Bringmark, 1988). Our results confirm the observation that common mosses are the most sensitive plant group in boreal for ests to an increased heavy metal load (Freedman and Hutchinson, 1980; Väisänen, 1986; Mäkinen, 1994). Although the frequency of P/eurozium schreberi and Dicranum spp. began to increase at a distance of 8 km, the Cu concentrations in their tissues were considerably higher (160-180 mg/kg) than those in background areas (6-8 mg/kg) (Helmisaari et al., 1995). The reindeer lichens (Cladina spp.) were more tolerant than forest mosses, but they did not increase until a distance of 4 km from the smelter. C. stellaris appeared to suffer more from the pollutants than C. arbuscula or C. rangiferina. Cladina species did not accumulate Cu as effectively as forest mosses at the same distances (Hel misaari et al. 1995), which partly explains their better resistance. The abundance distributions of Pohlia nutans and cup lichens (Cladonia spp.) along the transect were approximately bell-shaped. Their increase in abundance indicates the lack of stronger competitors halfway (3-4 km) along the transect. A similar response of Cladonia spp., but not of P. nutans, was observed near the Gusum smelter (Folkeson and Andersson-Bringmark, 1988). These species are pioneers that typically rapidly colonise vacated or disturbed habitats. Later succession stages of ground lichen communities have been found to be more sensitive to air pollutants than younger ones also on the Kola Peninsula (Gorshkov, 1993). Stereocaulon paschale and Cetraria islandica proved to be rather tolerant lichen species, the former species being found at 1 km and the latter at 2 km from the smel ter. P. nutans and C. purpureus were the only moss species surviving in small patches on the most contaminated site closest to the smelter. P. nutans, S. paschale and Clado nia spp. have been reported to withstand high ambient SO2 levels and elevated Cu and Ni concentrations in the soil at distances of 3-8 km from the smelter complex in Sudbury, Ontario (Freedman and Hutchinson, 1980). C. islandica has been found to grow at distances of 4-30 km from the Severonikel smelter in Monchegorsk, the Kola Peninsula (Gorshkov, 1993; Chernenkova and Kuperman, 1999). Arctostaphylos uva-ursi, Empetrum nigrum ssp. nigrum and Vaccinium uliginosum are clonal dwarf shrubs, which have survived on the most polluted sites. According to the annual rings on the oldest part of the stem, the age of the clones were at least 30-40 years (Salemaa et al., 2000). Because dwarf shrubs normally regenerate vegetatively, it is probable that some "mother clones" date back to the time when the smelter first started operating in the 19405, thus representing the most resistant genotypes of the earlier populations. Although A. uva-ursi was not present until 2 km along the study transect, it was found growing at a distance of 0.5 km on the opposite side of the smelter (unpublished observation). A. uva-ursi and E. nigrum ssp. herma froditum have also been found very close to the Mon chegorsk smelter on the Kola Peninsula (Lukina et al., 1993). Compared to the vegetation inventory made at Harjavalta in the 19705, A. uva-ursi and Calluna vulgaris seem to have declined, while the abundances of E. nigrum and Vaccinium vitis-idaea have remained unchanged during the last 20 years (Laaksovirta and Silvola, 1975). Experimental Cu exposure in a green house has shown that C. vulgaris is clearly more sensi tive than E. nigrum (Monni et al., 2000 a), which explains its absence from the most polluted site. In contrast, E. nigrum, has proved to be a very resistant species, and can accumulate considerably high con centrations of Cu and Ni especially in its stems and restrict their transport to the leaves (Monni et al., 2000b; Uhlig et al., 2000). However, the many signs of dead E. nigrum clones near the smelter indicate that the levels of heavy metals and sulphur have been too high for the majority of the E. nigrum population. 348 M. Salemaa et al. / Environmental Pollution 112 f 2001) 339-350 The tolerance mechanisms of vascular species to heavy metals have been studied to a considerable extent (e.g. Verkleij and Schatt, 1989; Turner, 1994). Resis tance is achieved either by restricting metal uptake by the plant (avoidance) or by binding heavy metals in cellular compartments so that sensitive metabolic pro cesses are not affected (tolerance) (Levitt, 1980; Baker, 1987; Tyler et al., 1989). Ericoid mycorrhizas have the ability to restrict the transport of heavy metals to shoots (Bradley et al., 1981). Clonal dwarf shrubs can live for decades, and they express high phenotypic plasticity in all their characteristics. V. uliginosum and A. uva-ursi showed strong regrowth after artificial shoot clipping in a field experiment carried out at Harjavalta (Salemaa et al., 1999). It was suggested that their high ability to compensate for the lost biomass is one factor con tributing to their resistance to heavy metals. Plastic growth responses have been found to be typical also of Empetrum nigrum and deciduous Vaccinium species growing along a pollution gradient on the Kola Penin sula (Shevtsova, 1998). In general, ecotypic differentia tion into metal-tolerant races is rare or very slow in long-lived plants (Dickinson et al., 1991; Turner, 1994). Increased coverages of grasses are often found in moderately polluted zones in the vicinity of point sour ces (Huttunen, 1975; Väisänen, 1986; Rigina and Kozlov, 1999). It is known that many grass species are able to evolve rapidly into tolerant ecotypes (Baker, 1987). Along the Harjavalta transect, only Deschampsia flexuosa grew at a low frequency. On the other hand, a sedge species, Carex globularis, grew successfully at paludified points on the contaminated sites at 0.5-1 km. Nitrogen application seems to have slightly increased the abundance of D. flexuosa and Epilobium angustifo lium on fertilised sample plots (unpublished data), but the general response to fertilisation was very low. In contrast to the situation in many other smelter areas, soil acidification did not increase on moving towards the smelter at Harjavalta (Derome and Lin droos, 1998). Apart from the site at 1 km, the pH of the organic layer was higher than 3.5 at all sites along the transect. Soil paludification at 1 km was reflected in the increased acidity (pH 3.3). The high concentrations of exchangeable Zn and Cd, which are more mobile than the other heavy metals, may originate from the slag heaps located close to the site at 1 km. These elements may have been leached from the surrounding soil or been deposited as particles and accumulated in the paludified depressions. The accumulated heavy metals have brought about a chronic disturbance in the ecosystem, preventing the normal succession of the plant communities. Colonisa tion of mosses and lichens is also prevented because the large amounts of undecomposed litter suppress their growth. Although viable seeds of many dwarf shrubs and tree species have been found in the soil seed banks in the polluted areas, seedling establishment is unsuc cessful (Salemaa and Uotila, 1996). It would appear, that the closest site will continue to deteriorate, but the decreased emissions will enable the vegetation to recover gradually at further distances. 5. Conclusions The species composition of the plant communities has changed, sensitive plant species have disappeared and competitive interactions between the species have been altered as a result of 50 years' activity of the Cu-Ni smelter at Harjavalta. According to their occurrence along the pollution gradient, forest mosses were the most sensitive plant group, followed by lichens. The vascular plants included both sensitive and resistant species. In general, the understorey vegetation seemed to respond more sensitively to the pollution load than the overstorey trees, which confirms the importance of including understorey vegetation in national and inter national monitoring programmes of forest health. Acknowledgements We are grateful to Leila Korpela, Sinikka Levula, Teuvo Levula and Pekka Suolahti for helping in the field work. Markku Tamminen and Tiina Tonteri advised in the data analysis. Krister Karttunen and Sampsa Lommi checked the identification of some of the species, and Hannu Nousiainen drew the cup lichens presented in Fig. 3. Heljä-Sisko Helmisaari, Leila Kor pela and Satu Monni commented on the manuscript. We would also like to thank Professor Eino Mälkönen for the possibility to use the data collected from the forest health fertilisation sample plots and Professor Erkki Tomppo for the vegetation data from national forest inventory. References Ahti, T., Hämet-Ahti, L., Jalas, J., 1968. Vegetation zones and their sections in northwestern Europe. Annales Botanici Fennici 5, 9-211. Alloway, 8.J., 1995. Soil processes and the behaviour of metals. In: Alloway, B.J. (Ed.), Heavy Metals in Soils. Blackie Academic & Professional, London, pp. 11-37. Axelsson, 8., Bräkenhielm, S., 1980. Investigation sites of the Swedish coniferous forest project biological and physiographical features. In: Persson, T. (Ed.), Structure and Function of Northern Con iferous Forests An Ecosystem Study. Ecological Bulletins (Stockholm) 32, 25-64. Baath, E., 1989. Effects of heavy metals in soil on microbial processes and populations (a review). Water, Air and Soil Pollution 47,335-379. Baker, A.J.M., 1987. Metal tolerance. New Phytologist 106, 93-111 Balsberg Pahlsson, A.-M., 1989. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. A literature review. Water, Air and Soil Pollution 47, 287-319. M. Salemaa et ai. I Environmental Pollution 112 (2001) 339-350 349 Bates, J.W., 1992. Bryophytes and lichens as target organisms. In: Ashmore, M.R., Wilson, R.B. (Eds.), Critical Levels of Air Pollu tants for Europe. Background papers prepared for the United Nations Economic Commission for Europe Workshop on Critical Levels. Egham, UK, 23-26 March 1992, pp. 113-124. Berthelsen, 8.0., Steinnes, E., Solberg, W., Jingsen, L., 1995. Heavy metal concentrations in plants in relation to atmospheric heavy metal deposition. Journal of Environmental Quality 24, 1018-1026. Bradley, R., Burt, A.J., Read, D.J., 1981. Mycorrhizal infection and resistance of heavy metal toxicity in Calluna vulgaris. Nature 292, 335-337. Cajander, A.K., 1949. Forest types and their significance. Acta Fore stalia Fennica 56, 1-71. Chernenkova, T.V., Kuperman, R.G., 1999. Changes in the forest communities along a heavy metal deposition gradient on Kola Peninsula. Water, Air and Soil Pollution 111, 187-200. Derome, J., 2000 a. Effects of heavy-metal and sulphur deposition on the chemical properties of forest soil in the vicinity of a Cu-Ni smelter, and means of reducing the detrimental effects of heavy metals. Research paper 769, Finnish Forest Research Institute. Derome, J., 2000b. Detoxification and amelioration of heavy-metal contaminated forest soils by means of liming and fertilisation. Environmental Pollution 107, 79-88. Derome, J., Lindroos, A.-J., 1998. Effects of heavy metal contamina tion on macronutrient availability and acidification parameters in forest soil in the vicinity of Harjavalta Cu-Ni smelter, SW Finland. Environmental Pollution 99, 225-232. Derome, J., Nieminen, T., 1998. Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Envir onmental Pollution 103, 219-228. Derome, J., Saarsalmi, A., 1999. The effect of liming and correction fertilisation on heavy metal and macronutrient concentrations in soil solution in heavy-metal polluted Scots pine stands. Environ mental Pollution 104, 249-259. Dickinson, N.M., Turner, A.P., Lepp, N.W., 1991. How do tree and other long-lived plants survive in polluted environments? Functional Ecology 5, 5-11. Folkeson, L., 1984. Deterioration of the moss and lichen vegetation in a forest polluted by heavy metals. Ambio 13, 37-39. Folkeson, L., Andersson-Bringmark, E., 1988. Impoverishment of vegetation in a coniferous forest polluted by copper and zinc. Canadian Journal of Botany 66, 417-428. Fowler, D.P., Cape, J.N., Unsworth, M.H., 1989. Deposition of atmospheric pollutants on forests. Philosophical Transactions of the Royal Society of London, Series B 324, 247-265. Freedman, 8., Hutchinson, T.C., 1980. Long-term effects of smelter pollution at Sudbury, Ontario, on forest community composition. Canadian Journal of Botany 58, 2123-2140. Fritze, H., Pennanen, T., Vanhala, P., 1997. Impact of fertilizers on the humus layer microbial community of Scots pine stands growing along a gradient of heavy metal pollution. In: Insam, H., Rangger, A. (Eds.), Microbial Communities. Functional versus Structural Approaches. Springer-Verlag, pp. 68-83. Fritze, H., Niini, S., Mikkola, K., Mäkinen, A., 1989. Soil microbial effects of a Cu-Ni smelter in southwestern Finland. Biological Fer tility of Soils 8, 87-94. Goodall, D.W., 1952. Some considerations in the use of point quad rats for the analysisi of vegetation. Australian Journal of Scientific Research Ser. B 5, 1-41. Gorshkov, V., 1993. The state of moss-lichen cover in polluted and unpolluted pine forests of the Kola Peninsula. In: Kozlov, M.V., Hau kioja, E., Yarmishko, V.T., (Eds.), Aerial Pollution in Kola Peninsula. Proceedings of the International Workshop, 14-16 April 1992, St. Petersburg. Kola Scientific Centre, Apatity, Russia, pp. 290-298. Hämet-Ahti, L., Suominen, J., Ulvinen, T., Uotila, P., Vuokko, S., 1998. Retkeilykasvio (Field Flora of Finland), 4th Ed. Finnish Museum of Natural History, Botanical Museum, Helsinki. Heinrichs, H., Mayer, R., 1980. The role of forest vegetation in the biochemical cycle of heavy metals. Journal of Environmental Qual ity 6, 402-407. Helmisaari, H.-S., Derome, J., Fritze, H., Nieminen, T., Palmgren, K., Salemaa, M., Vanha-Majamaa, 1., 1995. Copper in Scots pine for ests around a heavy-metal smelter in south-western Finland. Water, Air and Soil Pollution 85, 1727-1732. Huttunen, S., 1975. Studies on forest vegetation in air pollution damage area. Acta Universitatis Ouluensis Series A 33, Biologica 2, 1-37. Kahle, H., 1993. Response of roots of trees to heavy metals. Environ mental and Experimental Botany 33, 99-119. Kent, M., Coker, P., 1992. Vegetation Description and Analysis. A Practical Approach. John Wiley & Sons, Chichester. Koponen, T., Isoviita, P., Lammes, T., 1977. The bryophytes of Fin land: an annotated checklist. Flora Fennica 6, 1-77. Kozlov, M.V., Haukioja, E., Yarmishko, V.T. (Eds.), 1993. Aerial pollution in Kola Peninsula. In: Proceedings of the International Workshop, 14-16 April 1992, St. Petersburg, Kola Scientific Centre, Apatity, Russia. Kubin, E., Lippo, H., Poikolainen, J., 2000. Heavy metal loading. In: Mälkönen, E. (Ed.), Forest condition in a changing environment the Finnish case. Kluwer Academic, pp. 60-71. Laaksovirta, K., Silvola, J., 1975. Effect of air pollution by copper, sulphuric acid and fertilizer factories on plants at Harjavalta, W. Finland. Annales Botanici Fennici 12, 81-88. Levitt, J., 1980. Responses of Plants to Environmental Stresses. Aca demic Press, New York. Lobersli, E.M., Steinnes, E., 1988. Metal uptake in plants from a birch forest area near a copper smelter in Norway. Water, Air and Soil Pollution 37, 25-39. Lukina, N.V., Liseenko, L.A., Belova, E.A., 1993. Pollution-induced changes in the vegetation cover of spruce and pine ecosystems in the Kola Peninsula. In: Kozlov, M.V., Haukioja, E., Yarmishko, V.T., (Eds.), Aerial Pollution in Kola Peninsula. Proceedings of the International Workshop, 14-16 April 1992, St. Petersburg, Kola Scientific Centre, Apatity, Russia, pp. 312-321. Mäkinen, A., 1994. Biomonitoring of atmospheric deposition in Fin land, Estonia and the Kola Peninsula, based on the chemical ana lysis of mosses. Publications from the Department of Botany, University of Helsinki No. 19. Makkonen, K., Helmisaari, H.-S., 1998. Seasonal and yearly varia tions of fine-root biomass and necromass in a Scots pine (Pinus syl vestris L.) stand. Forest Ecology and Management 102, 283-290. Mälkönen, E., 1974. Annual primary production and nutrient cycle in some Scots pine stands. Communicationes Instituti Forestalis Fen niae 84 (5), 1-87. Mälkönen, E., Derome, J., Fritze, H., Helmisaari, H.-S., Kukkola, M., Kytö, M., Saarsalmi, A., Salemaa, M., 1999. Compensatory fertili zation of Scots pine stands polluted by heavy metals. Nutrient Cycling in Agroecosystems 55, 239-268. Minchin, P.R., 1987. An evaluation of the relative robustness of tech niques for ecological ordination. Vegetatio 69, 89-107. Minchin, P.R., 1991. DECODA. Database for ecological community data, Version 2.04. Australian National University. Monni, S., Salemaa, M., White, C., Tuittila, E., Huopalainen, M., 2000 a. Copper resistance of Calluna vulgaris originating from the pollution gradient of a Cu-Ni smelter, in southwest Finland. Environmental Pollution 109, 211-219. Monni, S., Salemaa, M., Millar, N., 2000b. The tolerance of Empet rum nigrum to copper and nickel. Environmental Pollution 109, 221-229. Nash, T.H.111, 1989. Metal tolerance in lichens. In: Shaw, J.A. (Ed.), Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, FL, pp. 119-131. Nieminen, T., Helmisaari, H.-S., 1996. Nutrient retranslocation in the foliage in Pinus sylvestris L. growing along a heavy metal pollution gradient. Tree Physiology 16, 825-831. 350 M. Salemaa et ai. I Environmental Pollution 112 (2001) 339-350 Nieminen, T., Derome, J., Helmisaari, H.-S., 1999. Interactions between precipitation and Scots pine canopies along a heavy-metal pollution gradient. Environmental Pollution 106, 129-137. Okland, T., okland, R.H., Steinnes, E., 1999. Element concentrations in the boreal forest moss Hylocomium splendens: variation related to gradients in vegetation and local environmental factors. Plant and Soil 209, 71-83. Pennanen, T., Frostegärd, A, Fritze, H., Baath, E., 1996. Phospholipid fatty acid composition and heavy metal tolerance of soil microbial communities along two heavy metal-polluted gradients in coniferous forests. Applied and Environmental Microbiology 62, 420-428. Punz, W.F., Sieghardt, H., 1993. The response of roots of herbaceous plant species to heavy metals. Environmental and Experimental Botany 33, 85-98. Rigina, 0., Kozlov, M.V., 2000. Pollution impact on sub-arctic forests in the Kola Peninsula, Russia. In: Innes, J.L., Oleksyn, J. (Eds.), Forest Dynamics in Heavily Polluted Regions. CAB International, Wallingford, pp. 37-65. Ross, S.M., 1994. Retention, transformation and mobility of toxic metals in soils. In: Ross, S.M. (Ed.), Toxic Metals in Soil Plant Systems. John Wiley, Chichester, pp. 63-152. Riihling, A, Brumelis, G., Goltsova, N., Kvietkus, K., Kubin, E., Liiv, S. et al., 1992. Atmospheric heavy metal deposition in northern Eur ope 1990. Nord 12, 1-41. Saari, H., Kartastenpää, R., Pohjola, V., Voltti, U., 1993. Hiukkas-ja rikkidioksiditutkimus Harjavallassa keväällä 1992. (Particle and Sulphur Dioxide Research at Harjavalta in the Spring 1992). Ilma tieteen laitos, Ilmanlaatuosasto, Helsinki (in Finnish). Salemaa, M., Uotila, T., 1996. Kanervan taimien hengissäsäilymisen todennäköisyys raskasmetallien saastuttamassa maassa. (The survi val probability of the Calluna vulgaris seedlings in heavy-metal pol luted soil). In: Rekilä, R., Kärenlampi, L., (Eds.), sth Meeting of Finnish Plant Scientists, University of Kuopio, Kuopio, Finland, p. 118 (in Finnish). Salemaa, M., Vanha-Majamaa, 1., Gardner, P.J., 1999. Compensatory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal polluted environment. Plant Ecology 141, 79-91. Salemaa, M., Vanha-Majamaa, 1., Reinikainen, A., Nousiainen, H., 2000. Response of understorey vegetation to heavy metal loading. In: Mälkönen, E. (Ed.), Forest Condition in a Changing Environment the Finnish Case. Kluwer Academic Publishers, pp. 266-275. SAS, 1994. SAS/STAT User's Guide Version 6, 4th Edition. SAS Institute Inc, Cary, NC, USA. Shevtsova, A., 1998. Responses of Subarctic Dwarf Shrubs to Climate Change and Air Pollution. Annales Universitatis Turkuensis, Ser All, Vol. 113. Turku, Finland. Tamminen, P., Starr, M.R., 1990. A survey of forest soil properties related to soil acidification in southern Finland. In: Kauppi, P., Anttila, P., Kenttämies, K. (Eds.), Acidification in Finland. Springer Verlag, Berlin, pp. 235-251. Tikkanen, E., Niemelä, 1., 1995. Kola Peninsula pollutants and forest ecosystems in Lapland. Final report of the Lapland Forest Damage Project. Gummerus. Jyväskylä. Tolvanen, A., Kubin, E., 1990. The effect of clear felling and site pre paration on microclimate, soil frost and forest regeneration at ele vated sites in Kuusamo. Aquilo Series Botanica 29, 77-86. Turner, A.P., 1994. The response of plants to heavy metals. In: Ross, S.M. (Ed.), Toxic Metals in Soil-Plant Systems. John Wiley, Chi chester, pp. 151-187. Tyler, G.. 1984. The impact of heavy metal pollution on forests: a case study on Gusum, Sweden. Ambio 13, 18-24. Tyler, G., 1990. Bryophytes and heavy metals: a literature review. Botanical Journal of the Linnean Society 104, 231-253. Tyler, G., Balsberg Pählsson, A.-M., Bengtsson, G., Baath, E., Tranvik, L., 1989. Heavy metal ecology of terrestrial plants, micro organisms and invertebrates. Water, Air and Soil Pollution 47, 189-215. Uhlig, C., Salemaa, M., Vanha-Majamaa, 1., Derome, J., 2001. Ele ment distribution in Empetrum nigrum microsites at heavy metal contaminated sites in Harjavalta, western Finland. Environmental Pollution 112, 431-438. Ukonmaanaho, L., Starr, M., Hirvi, J.-P., Kokko, A., Lahermo, P., Mannio, J., Paukola, T., Ruoho-Airola, T., Tanskanen, H., 1998. Heavy metal concentrations in various aqueous and biotic media in Finnish Integrated Monitoring catchments. Boreal Environment Research 3, 235-249. Väisänen, S., 1986. Effects of air pollution by metal, chemical and fertilizer plants on forest vegetation at Kokkola, W Finland. Annales Botanici Fennici 23, 305-315. Vangronsveld, J., Colpaert, J.V., Van Tichelen, K.K., 1996. Reclama tion of bare industrial area contaminated by non-ferrous metals: physiochemical and biological evaluation of the durability of soil treatment and revegetation. Environmental Pollution 94, 131— 140. Vanha-Majamaa, 1., Salemaa, M., Tuominen, S., Mikkola, K., 2000. Digitized photographs in vegetation analysis a comparison of cover estimates. Journal of Applied Vegetation Science 31, 89-99. Verkleij, J.A.C., Schat, H„ 1989. Mechanisms of metal tolerance in higher plants. In: Shaw, J.A. (Ed.), Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, FL, pp. 179-193. Vitikainen, 0., Ahti, T., Kuusinen, M., Lommi, S., Ulvinen, T., 1997 Checklist of lichens and allied fungi of Finland. Norrlinia 6. Winterhalder, K., 2000. Landscape degradation by smelter emissions near Sudbury, Canada, and subsequent amelioration and restora tion. In: Innes, J.L., Oleksyn, J. (Eds.), Forest Dynamics in Heavily Polluted Regions. CAB International, Wallingford, pp. 87-119. Wittig, R., 1993. General aspects of biomonitoring heavy metals by plants. In: Markert, B. (Ed.), Plants as Biomonitors. Indicators for Heavy Metals in the Terrestrial Environment. VCH, Weinheim, pp. 3-27. Zöttl, H.W., 1985. Heavy metal levels and cycling in forest ecosystems. Experimentia 41, 1104-1113. Paper II Salemaa, M., Derome, J., Helmisaari, H.-S., Nieminen, T. & Vanha- Majamaa, I. 2004. Element accumulation in boreal bryophytes, lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. The Science of the Total Environment (in press). Science of the Total Environment xx (2003) xxx-xxx Available online at www.sciencedirect.com 0048-9697/03/$ - see front matter © 2003 Published by Elsevier B.V. doi: 10.1016/SOO4B-9697(03)00629-6 Element accumulation in boreal bryophytes, lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland Maija John Derome b , Heljä-Sisko Helmisaari I*, Tiina Nieminen", Ilkka Vanha-Majamaa" * Vantaa Research Centre, Finnish Forest Research Institute, P.O. Box 18, FIN-01301 Vantaa, Finland b ßovaniemi Research Station, Finnish Forest Research Institute, P.O. Box 16, FIN-96301 Rovaniemi, Finland Received 15 June 2003; received in revised form 28 October 2003; accepted 29 October 2003 Abstract Macronutrient (N, P, K, Mg, S, Ca), heavy metal (Fe, Zn, Mn, Cu, Ni, Cd, Pb) and A 1 concentrations in understorey bryophytes, lichens and vascular plant species growing in Scots pine forests at four distances from the Harjavalta Cu-Ni smelter (0.5, 2, 4 and 8 km) were compared to those at two background sites in Finland. The aim was to study the relationship between element accumulation and the distribution of the species along a pollution gradient. Elevated sulfur, nitrogen and heavy metal concentrations were found in all species groups near the pollution source. Macronutrient concentrations tended to decrease in the order: vascular plants > bryophytes > lichens, when all the species groups grew on the same plot. Heavy metal concentrations (except Mn) were the highest in bryophytes, followed by lichens, and were the lowest in vascular plants. In general, vascular plants, being capable of restricting the uptake of toxic elements, grew closer to the smelter than lichens, while bryophytes began to increase in the understorey vegetation at further distances from the smelter. A pioneer moss (Pohlia nutans) was an exception, because it accumulated considerably higher amounts of Cu and Ni than the other species and still survived close to the smelter. The abundance of most of the species decreased with increasing Cu and Ni concentrations in their tissues. Cetraria islandica, instead, showed a positive relationship between the abundance and Cu, Ni and S concentrations of the thallus. It is probable that, in addition to heavy metals, sporadically high SO, emissions have also affected the distribution of the plant species. © 2003 Published by Elsevier B.V. Keywords: Biomonitoring; Cladina; Cladonia\ Dicranum; Pleurozium; Empetrum\ Vaccinium 1. Introduction The understorey vegetation of coniferous forests undergoes strong changes around smelter complex es emitting heavy metals, sulfur or other pollutants "Corresponding author. Tel.: +358-10-211-2539; fax: + 358-10-211-2202. E-mail address: maija.salemaa@metla.fi (M. Salemaa). in the northern hemisphere (Amiro and Courtin, 1981; Folkeson and Andersson-Bringmark, 1988; Rigina and Kozlov, 2000; Salemaa et al.. 2001). Different plant species show varying resistance to airborne and soil-accumulated toxic elements, which is reflected in their growth, survival and occurrence along pollution gradients. However, the actual degree of exposure to toxic elements is not 2 M. Salemaa et al. / Science of the Total Environment xc (2003) xxx-xxx the same for all the plant species growing at the same distance from an emission source because of differences in element uptake mechanisms (Tyler et al., 1989; Garty, 2001; Zechmeister et al., 2003). A range of morphological, anatomical and physi ological properties affect the capacity of individual species to filter, bind and accumulate elements on their surfaces, and to take them up intracellularly. The greatest difference in the uptake mecha nisms of elements is that between vascular plants and cryptogams (bryophytes and lichens). Vascular plants mainly take up elements via their roots from the soil, although the foliar uptake of gases (e.g. N0 2, NH 3 and S02) and soluble elements may also be substantial (Marschner, 1995). The foliar uptake of heavy metals has been demonstrated in many crop plants (e.g. Haslett et al., 2001) but, in evergreen species, the thick epidermis and waxy cuticle of the leaves provide external protection against toxic elements. Large amounts of metal containing dust become attached to the leaf surfac es of trees growing near the emission sources (Kozlov et al., 2000), and particles may also become embedded in the cuticular waxes (Rautio and Huttunen, 2003). Acidic deposition causes erosion of the cuticle (Manninen and Huttunen, 1995), which may increase the leaching of ele ments and the penetration of heavy metals into the foliar tissues. The bioavailability of heavy metals in the soil is regulated by many physical, chemical and bio logical properties and processes (Ernst, 1996). The mobility and toxicity of heavy metals are strongly related to the acidity and organic matter content of the soil (Alloway, 1995). Heavy metal ions accumulate first in the cortex of the roots, from where a small proportion passes through the endo dermis and is subsequently distributed into the different plant organs via the xylem and phloem. Vascular plants have many species-specific mech anisms to restrict the cellular uptake of heavy metals and to detoxify them internally (recently reviewed by Hall, 2002). The role of mycorrhizal fungi in retaining heavy metals in the root system is important in providing resistance to the host plants (Meharg and Caimey, 2000). Cryptogams, however, have no real roots, epi dermis or continuous cuticle layer, and they absorb water and dissolved elements directly across their surface. Most of the bryophyte and lichen species obtain the majority of their water and nutrients from atmospheric deposition; some species also obtain nutrients from water that has been in contact with the substrate (Bates, 1992; Garty, 2001). The following element compartments occur in both taxons: (1) trapped particulate matter; (2) inter cellular soluble elements; (3) extracellular elemen ts, bound to the cell wall on charged exchange sites; and (4) intracellular elements (Tyler, 1990; Garty, 2001; Zechmeister et al., 2003). Both bryo phytes and lichens (especially the mycobiont part ner) have a high ion exchange capacity on their cell walls, and the dead tissues also have an ability to bind ions (Tyler, 1989, 1990; Chettri et al., 1997). The pollution gradients extending from emission sources can be used as 'field tests' for studying the relative resistance of different plant species in real ecological conditions. Non-resistant species or populations show a sudden or a gradual decrease in abundance as the pollution level increases (Tyler et al., 1989; Lepp and Salmon, 1999). However, without the use of ultrastructural analysis to iden tify the intracellular accumulation of toxic elemen ts and experimental exposures, it is almost impossible to identify the contribution of individ ual pollutants to the occurrence of a species. In this study we compare the element concen trations of unwashed samples of the understorey bryophyte, lichen and vascular plant species (later refered as life forms) growing in coniferous forests at four distances from the Harjavalta Cu-Ni smelt er, and in two background areas in other parts of Finland. Our aim is to answer the following questions: 1. Does the accumulation of elements differ between the different life forms, species and plant parts in polluted and background areas? 2. How is the accumulation pattern of heavy met als and S related to the occurrence and abun dance of the plant species along the pollution gradient? 3. What is the potential of different plant species in biomonitoring of heavy metals? M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx 3 Fig. 1. Location of the three study locations. Mekrijärvi (Mj) and Hämeenkangas (Hk) represent background areas. The town of Harjavalta is situated approximately 30 km from the coast (Gulf of Bothnia) in Western Finland. The study plots at distances of 0.5, 2, 4 and 8 km from the Harjavalta Cu-Ni smelter (see stack on the map) have been marked on the detailed map. © National Land Survey of Finland 554/myy/03. 2. Material and methods 2.1. Study areas The study plots are located in Scots pine (Pinus sylvestris L.) stands at four distances (0.5, 2, 4 and 8 km) (later called H0.5, H2, H4 and H8) to the south and south-west of the Harjavalta copper nickel smelter (61°19' N, 22°9' E). The two back ground areas, Hämeenkangas (Hk) (61°45' N, 22°40' E) and Mekrijärvi (Mj) (62°47' N, 30°58' E), are located at a distance of 60 and 420 km from Harjavalta, respectively (Fig. 1). The size of the sample plots was 30X30 m (H0.5: 26X26 m). The plots were established for nutrient flux studies (Helmisaari, 1995) and as control plots for liming and fertilization experiments (Mälkönen et ai., 1999). The plots at Harjavalta and Hämeen kangas were xeric and Mekrijärvi sub-xeric heath forests (Table 1). Although Mekrijärvi represented slightly more fertile forest than the other plots, it was selected for a clean reference area especially for bryophytes and lichens. The soil on all the plots is sorted fine sand and the soil type ferric podzol. The age of the stands ranged from 40 to 50 years. Table 1 summarizes the general stand characteristics. Carpet-forming bryophytes, bilberry (Vaccinium myrtillus) and lingonberry (V. vitis-idaea) domi nated in the understorey at Mj, whereas reindeer lichens (Cladina spp.) and evergreen dwarf shrubs were typical species at Hk and H8 (Table 2). The cover of the understorey vegetation and the species richness decreased with decreasing distance to the Cu-Ni smelter at Harjavalta (Table 2). The nearest plot to the smelter (H0.5) was almost devoid of vegetation, and the forest floor was covered with a thick layer of undecomposed needle litter. 4 M. Salemaa el ai / Science of the Total Environment xx (2003) xxx-xxx Table 1 General stand characteristics CT = Calluna type (xeric heath type), VT=Vaccinium type (sub-xeric heath type),+ indicates slightly more fertile type. Results from Harjavalta published by Mälkönen et ai. (1999), Hämeenkangas (Hk) by Mälkönen et ai. (2000) and Mekrijärvi (Mj) by Helmisaari and Mälkönen (1989). 2.2. Air pollutant emissions The Harjavalta Metals smelter complex is one of the largest point sources of heavy metal emis sions in Finland. The copper smelter has been operating since 1945, and the nickel smelter since 1960. The concentrated ores contain sulfur, heavy metals and arsenic. Before the sulfuric acid plant was built in 1947, all the S02 produced during the smelting process (annually approx. 30 000 t) was emitted into the atmosphere, causing severe dam age to the surrounding coniferous forests. At the end of the 1980s the average annual emissions of S0 2 were 8050 t and total dust 1200 t, including Cu 110 t, Ni 53 t and Zn 180 t. Since the beginning of the 19905, however, the emissions have been considerably reduced: in 1992, Cu 60 t, Ni 10 t, Zn 12 t, S0 2 4800 t and total dust 280 t (Fig. 2). The 24-h mean S02 concentrations in the air have decreased from the level of 38-56 M-g/m 3 in 1987 (1 January-30 June) to 17-18 3 in 1992 (21 January-15 May) within a 1 km radius of the smelter. However, the peak hourly concentrations in 1992 still occasionally reached 500-1000 |xg/ m 3 (Saari et ai., 199/0- The Kemira fertilizer factory produced superphosphate and PK fertilizers at Harjavalta from 1948 to 1989. The study plots have probably not been exposed to the heaviest deposition levels in the vicinity of the smelter, because they were not located under the predomi nant southern winds (Derome, 2000). The background areas (Hk, Mj) had no local emission sources. The mean annual bulk deposition of sulfate in open areas varied between 300 and 500 mg/m 2 during 1988-1996 in South Finland (Kulmala et ai., 1998). Atmospheric emissions of heavy metals have decreased substantially in Fin land during the 19905, which are clearly reflected as low heavy metal concentrations in background bryophytes in the national surveys (Poikolainen et ai ..jm#). 2.3. Plant sampling and chemical analysis The plant samples for elemental analysis were collected from the surrounding buffer zones of the study plots in August-September 1992. Additional samples from the Harjavalta gradient (H0.5 and H2) were taken in September 1993 and 1994. Random sampling was performed using 30 small quadrats (using a 20X20 cm frame) stratified on different sides of the plots (H4, HB, Hk and Mj). The total sampled area was 1.2 m 2 per plot. Subjective selection of the plant material had to be applied on the highly polluted plots (H0.5 and H2, see species in Appendices A and B) owing to the very low abundance of understorey vegetation (Table 2). The plant samples were cut off level with the surface of the organic layer, and thus excluded rhizomes and roots. The samples were placed in plastic bags, and stored frozen up until chemical analysis. Harjavalta gradient Background areas HO. 5 H2 H4 H8 Hk Mj Forest site type CT+ CT+ CT CT CT VT Stand age, years 49 54 48 40 45 44 Number of trees/ha 1008 1230 1517 1552 2017 2660 Mean pine height, m 6.1 10.9 9.2 10.6 9.4 7.4 Stem volume, m 3 /ha 23.2 85.3 67.8 94.5 49.8 63.1 Volume increment, m'/ha^^ear^ 0.31 3.78 2.78 6.27 3.29 5.2 Distance to the Harjavalta Cu-Ni smelter, km 0.5 2 4 8 60 420 M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx 5 Table 2 The average abundance (point frequency %) of the bryophyte, lichen and vascular plant species at the six study sites in 1992 Averages from 16 sample quadrats (each 1 m 2) per sample plot.+ = species present on the sample plot but not on the sample quadrat. Nomenclature as per Hämet-Ahti et ai. (1998): vascular plants, Koponen et ai. (1977): bryophytes and Vitikainen et ai. (1997): lichens. The plant biomass was divided according to species and age class. Current-year shoots, older living parts and dead biomass were separated on vascular plants. The upper living part of the thalli of bryophytes and lichens was separated from the lower decomposing (darker) parts on the basis of their color difference. In the case of bryophytes, the upper part consisted of 2-3 year's growth. The plant material was not washed before chemical analysis and thus included the surface accumula tion of elements. All the plant samples were handled using cotton gloves. The species-specific plant samples were com bined into composite samples per side or whole plot in order to obtain sufficient material for elemental analysis. The dry weight of the compos ite samples ranged from 0.5 to 1.0 g for bryo phytes, and from 1.0 to 3.0 g for other plants. The samples were oven-dried ( + 60 °C), weighed, homogenized and dry digested ( + 550 °C). The ash was extracted with 2-3 ml of 6 M HCI (p.a.) in a water bath (approx. +BO °C). The dry residue was diluted with 10 ml 1 M HCI for 20 min and filtered, and the filter paper rinsed with 0.1 M Harjavalta gradient Background HO. 5 H2 H4 H8 Hk Mj Bryophytes: Ceratodon purpureus (Hedw.) Brid 0.0 0.1 0.0 0.0 0.0 0.0 D. polysetum Sw. 0.0 0.0 0.0 1.2 9.9 11.2 D. scoparium Hedw. 0.0 0.0 0.0 0.5 0.5 0.2 P. schreberi (Brid.) Mitt. 0.0 0.0 0.1 3.6 9.8 77.0 P. nutans (Hedw.) Lindb. 0.1 2.4 2.2 5.2 0.8 0.0 P. juniperinum Hedw. 0.0 0.0 0.0 0.0 0.9 0.1 Ptilidium ciliare (L.) Hampe 0.0 0.0 0.2 0.3 0.0 0.0 Lichens: C. islandica (L.) Ach. 0.0 26.5 9.5 6.0 2.3 5.2 C. arbuscula (Wallr.) Hale and Culb. 0.0 0.0 9.2 19.1 31.7 3.4 C. rangiferina (L.) Nyi. 0.0 0.2 13.9 10.4 12.2 3.9 C. stellaris (Opiz) Brodo 0.0 0.0 5.2 32.0 18.6 + C. uncialis (L.) F.H.Wigg. 0.0 0.0 0.0 0.0 0.1 0.0 Cladonia spp. 0.0 3.7 5.7 6.6 4.6 1.1 Vascular plants: A. uva-ursi (L.) Sprengel 0.0 1.8 0.0 0.0 0.3 0.0 Betula pubescens Ehrh. + + 0.0 0.0 0.0 0.0 C. vulgaris (L.) Hull 0.0 + + 0.9 11.7 8.5 C. globularis L. + 0.0 0.0 0.0 0.0 0.0 D. flexuosa (L.) Trin. 0.0 0.0 0.0 + 0.2 0.0 Epilobium angustifolium L. 0.0 0.1 0.0 + 0.0 0.0 E. nigrum L. 0.1 10.3 6.8 0.1 + 0.1 Juniperus communis L. 0.0 0.0 0.0 0.0 + 0.0 Ledum palustre L. + 0.0 0.0 0.0 0.0 0.0 P. sylvestris L. + 0.6 1.3 1.6 0.5 0.0 Sorbus aucuparia L. 0.0 0.0 0.0 0.0 0.0 0.1 V. myrtillus L. 0.0 0.1 0.0 0.0 + 16.4 V. uliginosum L. 0.1 0.0 0.0 0.1 0.0 0.0 V. vitis-idaea L. 0.0 2.5 11.8 6.3 3.3 9.7 6 M. Salemaa et al. / Science of the Total Environment xx (2003) xxx-xxx Fig. 2. Cu, Ni, Zn and Pb emissions (bars) and S02 emissions (t/year) from the Harjavalta smelter during 1985-1995. Source Outokumpu Harjavalta Metals Oy. HCI. The final volume of the solution was 25- 100 ml depending on the sample weight. Total element concentrations (P, K, Mg, Ca, Fe, Zn, Mn, Cu, Ni, Cd, Pb and Al) were determined by inductively coupled plasma atomic emission spec trometer (ICP-AES) (Dahlquist and Knoll, 1978). Total sulfur and nitrogen concentrations were determined from the homogenized samples on a LECO S-132 and LECO CHN-600 analyzers. All the analyses were performed in the central labor atory of the Finnish Forest Research Institute. The laboratory included its own blank and standard samples in all the batches, and the laboratory participated, with satisfactory results, in interna tional inter-calibration exercises (e.g. during the analysis period in question, IUFRO Inter-labora tory Sample Exchange 1993). The abundances of the plant species growing on the study plots (30X30 m) were measured in 1992 using the point quadrat method (Salemaa et al., 2001). Each plot was divided into four subplots, and stratified random sampling performed on 16 vegetation quadrats (1 m 2). 2.4. Soil and precipitation sampling, and chemical analyses Samples were taken from the organic and min eral soil layers (0-5 cm) at 25 systematically selected points in 1992 (Hk in 1990). The organic layer samples were dried and milled to pass through a 1 mm sieve. The mineral soil samples were passed through a 2 mm sieve to remove stones and large roots. pH was measured in water. Organic matter content was determined as loss in weights on ignition by ashing the samples (550 °C, 3 h). Total N was determined on a CHN analyzer. Exchangeable Ca, Mg, K, Cu, Ni, Zn, Fe, Mn, Cd, Pb and extractable P and S at the Harjavalta plots were determined by extraction with 1 M ammonium acetate (pH 4.65)+ 1% EDTA, followed by analysis by ICP-AES. EDTA was used in the extractant for the samples from the polluted sites because it was known that they contained high concentrations of Cu and Fe (Dero me and Lindroos, 1998). The extractant used for the samples from Hämeenkangas and Mekrijärvi did not include EDTA due to the low levels of Cu and Fe. The element concentrations in the organic layer were expressed on an organic matter basis in order to reduce the variation arising from the inclusion of varying amounts of mineral soil in the organic layer samples. The chemical properties of the soil along the Harjavalta gradient are described in more detail in Derome and Lindroos (1998) and Derome (2000). M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx 7 A description of the sampling of precipitation and its chemical analysis is given in Derome and Nieminen (1998). 2.5. Statistical analysis Plotwise means of the element concentrations of the species were calculated in the cases where there were 2-4 composite samples per plot. Dif ferences in the element concentrations between the life forms and species were tested by Mann- Whitney's U tests. Non-parametric statistics were used because the sample number was low, and it was not possible to test the normality of the distributions. The relationship between the element accumulation (Cu, Ni and S) in different species at H8 (where all life forms were present) and the closest distance to the smelter at which the species occurred, was analyzed by Pearson's correlation coefficients. The abundances of the species were plotted against the Cu concentrations of their tissues using the data of the Harjavalta gradient and the nearest background area (Hk). 3. Results 3.1. Precipitation and soil chemistry The effect of emissions from the Harjavalta smelter was clearly reflected in the composition of bulk precipitation and stand throughfall (Table 3). The deposition of heavy metals (except Mn) was tens or hundreds of times higher at H0.5 than in the background areas. The bulk deposition of Cu and Ni increased exponentially towards the smelter (Fig. 3a,b). Furthermore, elevated N, S0 4 and Mg deposition was recorded near the smelter. Of the two background areas, Mj had lower N and S04 deposition than Hk (Table 3). The throughfall values, which also include elements leached from the canopy, were higher than those of the bulk precipitation except for nitrogen (HB, Hk) on all the study plots. The heavy metal concentrations in the organic layer increased strongly towards the smelter (dis tributions for Cu and Ni given in Fig. 3a,b), whereas those of the macronutrients (K, Mg and Ca) decreased (Table 4). The effect of emissions was also seen as elevated concentrations of N, S and Pat H0.5 (Table 4). Both the macronutrient and heavy metal concentrations were higher in the organic layer than in the mineral soil (0-5 cm). The enrichment factor (H0.5 vs. Mj) for the Cu concentration was 645 in the organic layer and 66 in the mineral soil. The corresponding value for Ni in the organic layer was 848. Mj differed from the other plots by having higher macronutrient, Fe and Mn concentrations in the mineral soil (Table 4). 3.2. Element accumulation in the plant species! 3.2.1. Variation between the life forms (bryophytes, lichens and vascular plants) When the life forms growing on the same plots were compared, the macronutrient concentrations tended to be the highest in vascular plants and decreased from bryophytes to lichens (Appendices A and B). There were some exceptions to this pattern, which depended on the element, area and age of the plant part. For instance, higher concen trations of N, P, Mg and Ca (H0.5) and S (H4) were found in Pohlia nutans compared to the other species. The average N and S concentrations of Pleurozium schreberi and Dicranum spp. (upper parts, N: 1.11%, S: 0.98 mg/g, n= 7) and vascular plants (current-year shoots, N: 1.10%, S: 1.25 mg/ g, n= 6) were relatively similar (N: [7=19.0, P 0.836; S: (7=12.0, P = 0.234) in the combined data of the background areas, whereas those of Cladina and Cetraria lichens (upper parts, N: 0.62%, S: 0.54 mg/g, n =l4) were significantly lower (N: [7=0.0, P=0.001; S: [7=1.0, P= 0.001). The heavy metal concentrations were the highest in bryophytes, followed by lichens, and the lowest in vascular plants in the Harjavalta data (distribu tions for Cu and Ni given in Fig. 3c,d). The only exception was Mn, which followed the order: vascular plants > bryophytes > lichens. Vaccinium species, especially V. myrtillus, had high Mn con centrations in Mj. The differences in the heavy metal concentrations between the life forms were greater in the polluted sites at Harjavalta than in the background areas. This was especially the case 8 M. Salemaa et al. / Science of the Total Environment xx (2003) xxx-xxx Table 3 Annual precipitation (mm) and deposition of elements (mg/m 2 ) in bulk precipitation (BP) and stand throughfall (TF) Sampling period: 6 June 1992-21 June 1993 (H0.5, H4, HB, Hk), 12 January 1992-15 December 1992 (H2), and 19 October 1992-27 October 1993 (Mj). m.d.=missing data, throughfall was not collected at H2. 2OO |xg/g) concentrations in bryophytes (composite samples of P. schreberi and Hylocom ium splendens) growing near the smelter complex es on the Kola Peninsula (10 km from Zapoljarnij and 25 km from Monchegorsk). The same pattern was found in the reindeer lichens: the maximum Cu and S concentrations were at the same level in both areas, but the Ni concentrations were higher on the Kola Peninsula (Reimann et al.. 1999). The highest Cu concentrations of P. schreberi and D. polysetum were much lower (70-80 (xg/g) near (6-7 km) the Gusum smelter (Folkeson and Andersson-Bringmark, 1988). However, the same species accumulated larger amounts of Zn at Gus um (300-330 |xg/g) than at Harjavalta (60 p.g/ g). The upper parts of C. rangiferina accumulated more Cu at Gusum (350 |JLg/g) than at Harjavalta (214 p-g/g). The variation in the maximum concentrations of elements in different geographical regions indi cates that it is very difficult to present any com mon, maximum limits for the survival of cryptogams based on field data. According to our results and values from the literature, it is amazing how high metal concentrations cryptogams can withstand in field conditions compared, e.g. to the experimentally determined toxic limits of vascular plants (see below). However, it should be noted that unwashed samples have considerable amounts of dust attached to them, and the actual tissue concentrations that have harmful metabolic effects are much lower than those reported in field con ditions (Brown and Brumelis, 1996; Bennett, 1999). P. nutans, which accumulated exceptionally high concentrations of heavy metals along the Harja valta gradient, has been considered to be a pollut ant-resistant species also in other investigations (Amiro and Courtin, 1981; Tyler, 1990). Lepp and Salmon (1999) suggested that pleurocarpous bryo phytes (those with a horizontal growth form) are more sensitive to toxic elements than acrocarpous ones (upright growth form) like P. nutans. This sensitivity may be related to differences in the water-conducting systems and uptake of soluble metals between the growth forms. As Pakarinen (1981) has presented earlier, there are between-species differences within the Cladina genus in the capacity to accumulate toxic elements (C. stellaris>C. arbuscula). This was especially evident under an increasing pollution load. These results agree well with the observations of Garty (tmo. who gave examples of how lichen species with finely divided thalli had a greater affinity to collect particulate matter than undivided ones. Accordingly, C. islandica, which is characterized by a flat, coarsely divided thallus, contained small er amounts of pollutants than Cladina spp. growing at the same location. C. islandica proved to be the most resistant lichen species at Harjavalta, and it has also been found to grow near the smelter complexes on the Kola Peninsula (Chernenkova and Kuperman, 1999). However, C. islandica, as well as cup lichens (Cladonia spp.), may benefit M. Salemaa et al. / Science of the Total Environment xx (2003) xxx-xxx 15 from the disappearance of reindeer lichens in moderately polluted areas (Folkeson and Anders son-Bringmark, 1988). 4.2.2. Maximum Cu, Ni, Zn and S concentrations in vascular plants Vascular plants were growing close to the pol lution source. However, the resistance level varied between the individual vascular plant species. For instance, C. vulgaris and A. uva-ursi were absent from the most polluted area, where a few popula tions of C. globularis, E. nigrum and V. uliginosum had survived. The ranking of dwarf shrubs in experimental Cu exposures was the same as the order of the species occurrence along the pollution gradient: E. nigrum (most resistant) >C. vulgar is>A. uva-ursi (most sensitive) (Monni et al.. 2000a,b; Salemaa and Monni, 2003). When we exclude the study plot closest to the smelter (H0.5), the maximum heavy metal concentrations measured in the current-year shoots of the dwarf shrubs (Cu 8-30 |xg/g, Ni 4-12 (xg/g, Zn 9-35 (j.g/g) were generally lower than the critical tissue values given in the literature. Marschner (1995) reported that toxic foliar concentrations of most crop species are above 20-30 (xg/ g for Cu, 10- 50 |xg/g for Ni and 100-300 |xg/g for Zn. However, there are considerable differences between the vascular plant species and ecotypes in their toxic limits (Balsberg Pählsson, 1989). The concentrations in the dwarf shrubs at Harja valta increased with shoot age, and surface accu mulation seemed to be especially high on the dead parts of vascular plants (e.g. E. nigrum had Cu> 4000 jxg/g at H0.5). The maximum Cu, Ni and S concentrations in the current-year shoots of E. nigrum growing closest to the Harjavalta smelter were lower than the values reported by Reimann et al. (1999) in E. nigrum growing at 5 km distance from Mon chegorsk, the Kola Peninsula: median (max) Cu = 408 (626) |xg/g, Ni = 329 (626) (xg/g and S = 1390 (2020) p.g/g. The Zn concentrations, how ever, were higher at Harjavalta. E. nigrum is able to accumulate very high concentrations of Cu and Ni, especially in its older stems, but also in older leaves (Monni et al., 2000 a; Uhlig et al., 2001). In contrast, V. uliginosum seems to base its resis tance on the restriction of metal transport to leaves. Uhlig and Junttila (2001) found a similar relation ship between E. nigrum spp. hermaphroditum and V. myrtillus in Northern Norway, near to the adjacent Nikel and Zapoljarnij smelters on the Russian side of the border. Relatively low Cu and Ni concentrations were also found in the leaves and stems of lowbush blueberry V. angustifolium, growing near the smelter complexes at Sudbury, Canada (Bagatto and Shoithouse, 1991). 4.3. Applicability of the results in biomonitoring Emissions from the Cu-Ni smelter and fertilizer factor at Harjavalta were clearly reflected in the elemental load in bulk deposition, and the concen trations in the understorey vegetation and organic layer. The deposition gradient was very steep, resulting in strong inter-correlations between all three compartments. This makes it extremely dif ficult to distinguish between the role of airborne deposition, wind-blown dust and elements taken up by the substrate in the chemical composition of the plants. As emphasized by other authors (e.g. Halleraker et al., 1998; Reimann et al., 2001), we conclude that the local conditions and the element ranges in deposition strongly influence the rela tionship between deposition and plant uptake in polluted areas. The heavy metal and S concentrations were elevated especially in the older plant parts, which have been exposed to deposition for a longer time. The age (Pakarinen, 1981) and the growth rate (Zechmeister, 1995) affect the element accumula tion and should be taken into account when using plants as accumulation indicators. Our results reveal that there are considerable differences between the bryophytes, lichens and vascular plants in their capacity to accumulate pollutants and to grow in contaminated soil. We conclude that information about all the life forms in the understorey is needed when evaluating the state and recovery of forest ecosystems in heavily pol luted areas. Acknowledgments We thank the staff of the Mekrijärvi Research Station, University of Joensuu, and Pia Reponen, M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx 16 M. Sci., for helping in the data handling phase of the study. The comments of Dr Pasi Rautio and Dr Timo Vuorisalo on the manuscript are greatly acknowledged. The final version of the manuscript Appendix A: was greatly improved by the suggestions and comments of Dr James Bennett and an anonymous referee, to whom we express our sincere thanks. The mean element concentrations in lichens and bryophytes. Parts: 1 = living upper part, 2 = decomposing lower part, 3 = the whole thallus. n = number of composite samples. Area and year Part n Macronutrients mg/g N% P K Mg S Ca Heavy metals and A1 |xg/g Fe Zn Mn Cu Ni Cd Pb A1 Mekrijärvi 1992 C. islandica 1 1 0.49 0.60 3.00 0.33 0.48 0.86 111.7 33.0 169.3 2.27 1.26 0.16 3.06 101.9 2 1 0.37 0.70 1.89 0.36 0.51 1.34 331.0 41.4 224.7 3.70 2.66 0.17 5.56 271.7 C. arbuscula 1 1 0.51 0.55 1.92 0.33 0.44 0.76 139.4 19.9 181.0 2.13 1.30 0.13 2.79 131.6 2 1 0.48 0.48 1.00 0.22 0.46 0.73 342.1 17.7 194.9 2.82 2.03 0.08 3.49 281.3 C. rangiferina 3 1 0.52 0.43 1.41 0.28 0.37 0.68 211.4 16.8 147.8 2.29 1.78 0.14 4.79 191.4 C. stellaris 3 1 0.35 0.34 1.12 0.18 0.28 0.47 134.7 16.6 80.5 1.69 2.01 0.13 3.52 117.3 Dicranum spp. 1 1 1.10 1.35 5.74 1.17 0.72 2.37 343.0 44.4 878.5 8.45 4.25 0.25 8.70 380.8 2 1 0.94 1.11 4.45 0.91 1.03 2.43 649.6 59.0 916.1 7.34 4.55 0.27 13.91 724.0 P. schreberi 1 3 0.96 1.22 5.23 1.08 0.89 3.20 282.6 41.0 705.3 6.48 3.65 0.20 9.44 303.6 2 3 0.80 0.84 2.96 0.57 0.85 2.96 624.3 47.8 583.1 5.70 5.05 0.30 15.02 632.0 P. nutans 3 1 1.18 1.44 4.17 0.82 - * 3.69 480.0 56.5 731.6 7.55 6.06 0.41 22.25 564.9 Hämeenkangas 1992 C. islandica 1 1 0.51 0.31 1.86 0.17 0.39 0.34 113.9 24.0 40.1 3.66 2.13 0.17 6.74 100.8 2 1 0.50 0.35 0.97 0.17 0.35 0.67 505.7 34.1 61.8 7.01 5.22 0.20 12.03 373.8 C. arbuscula 1 4 0.67 0.42 1.55 0.21 0.56 0.44 196.8 28.1 53.0 4.91 1.59 0.16 4.90 165.2 2 4 0.70 0.44 0.73 0.17 0.56 0.55 474.4 22.2 55.3 7.25 2.95 0.15 8.75 346.6 C. rangiferina 1 3 0.67 0.42 1.67 0.21 0.49 0.41 283.4 27.2 48.1 4.73 5.70 0.20 6.00 158.9 2 1 0.68 0.40 0.65 0.17 0.51 0.57 637.4 27.4 48.5 8.84 3.85 0.19 18.90 446.5 C. stellaris 1 4 0.61 0.36 1.30 0.18 0.65 0.36 248.1 26.8 37.3 4.47 3.09 0.12 5.26 182.4 2 4 0.64 0.42 0.59 0.16 0.51 0.51 576.8 22.6 48.9 8.12 3.88 0.14 9.% 409.5 Cladonia spp. 3 1 0.85 0.64 1.80 0.23 0.76 0.61 454.6 46.6 63.9 7.05 2.81 0.22 9.17 426.9 C. uncialis 3 1 0.57 0.34 1.26 0.18 0.45 0.30 229.9 32.2 33.6 4.97 1.92 0.15 5.49 180.7 Dicranum spp. 1 1 1.31 0.98 3.85 0.86 1.28 3.49 681.7 58.2 370.3 15.63 7.85 0.48 20.63 548.8 2 1 1.20 0.97 2.59 0.64 1.48 3.46 1013.4 77.1 350.3 16.85 10.17 0.61 24.73 926.3 P. schreberi 1 2 1.24 0.86 3.20 0.74 1.12 2.81 547.4 55.6 342.1 13.35 7.34 0.39 20.09 523.3 2 1 0.90 0.77 2.02 0.47 1.14 2.80 825.8 55.4 237.5 14.40 8.65 0.36 25.30 804.9 P. nutans 3 1 1.36 0.% 1.91 0.61 1.04 2.76 848.4 82.7 209.1 13.82 10.99 0.70 47.74 922.5 Polytrichum juniperinum 3 1 1.10 0.81 2.66 0.58 0.85 2.23 436.7 54.3 222.3 16.26 6.04 0.44 22.97 797.0 Harjavalta 8 km 1992 C. islandica 1 1 0.48 0.42 2.30 0.18 0.52 0.36 220.1 39.7 27.5 33.03 12.02 0.20 4.44 162.6 2 1 - 0.42 1.41 0.25 - 1.00 869.4 52.4 74.9 109.87 25.13 0.34 8.44 499.2 C. arbuscula 1 4 0.78 0.51 1.78 0.28 0.76 0.50 394.4 37.6 35.0 56.94 11.92 0.27 7.69 253.7 2 4 0.77 0.61 0.92 0.61 0.90 1.28 1179.3 42.6 59.0 162.05 29.95 0.31 14.31 609.6 C. rangiferina 1 3 0.78 0.51 1.71 0.24 0.75 0.35 448.7 37.0 39.6 60.26 13.10 0.33 10.83 284.3 2 2 0.83 0.54 0.88 0.26 0.91 0.58 1191.1 40.2 54.3 148.78 29.58 0.33 18.39 603.8 C. stellaris 1 4 0.75 0.53 1.67 0.25 0.82 0.39 426.4 40.9 29.2 63.60 11.74 0.26 9.62 264.6 2 4 0.80 0.63 0.78 0.35 0.93 0.97 1291.1 40.6 52.9 175.24 30.60 0.32 19.75 612.7 Cladonia spp. 3 1 0.85 0.64 2.07 0.27 1.00 0.61 569.0 54.4 56.6 88.41 19.36 0.33 12.01 389.3 Dicranum spp. 1 2 1.14 1.26 5.14 0.78 1.11 2.82 962.3 61.5 272.6 159.14 34.97 0.83 24.81 772.6 2 2 1.12 1.11 3.36 0.67 1.38 3.35 1684.2 87.1 296.9 231.67 65.01 0.84 41.99 1264.9 P. schreberi 1 3 1.19 1.15 4.25 0.66 1.19 2.25 1054.8 59.7 206.8 183.10 37.97 0.64 30.48 801.7 2 2 0.98 0.98 2.05 0.51 1.27 2.29 1884.9 73.5 195.8 276.49 66.41 0.86 55.24 1237.8 P. nutans 3 1 1.48 1.36 2.14 0.58 1.56 2.57 1518.9 94.1 189.2 268.72 79.64 1.20 61.35 1110.6 Haijavalta 4 km 1992 C. islandica 1 3 0.61 0.37 1.66 0.17 0.58 0.32 365.0 54.9 22.3 108.47 26.40 0.34 14.48 190.3 17 M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx Appendix B: The mean element concentrations in vascular plant species. Parts: 1= current living shoots, 2=older living parts, 3 = dead parts, 4=berries, n = number of composite samples. 2 1 1.06 0.68 1.03 0.25 1.05 0.79 1254.2 70.4 52.0 365.43 76.24 0.66 29.44 483.5 C. arbuscula 1 3 0.77 0.51 1.52 0.22 0.82 0.30 631.5 46.5 24.0 159.90 28.87 0.50 13.38 246.0 2 1 0.93 0.71 0.96 0.28 1.35 0.65 1959.4 72.8 41.3 447.93 70.15 0.71 25.93 552.6 C. rangiferina 1 2 0.81 0.54 1.60 0.24 0.85 0.37 875.7 54.7 30.2 214.29 42.36 0.53 19.68 309.2 2 1 0.97 0.69 0.89 0.32 1.44 0.64 2638.2 82.1 47.4 586.27 96.55 0.70 37.84 773.3 C. stellaris 1 2 0.85 0.62 1.55 0.26 1.05 0.40 1032.8 61.1 24.0 258.70 41.96 0.53 22.40 342.0 2 1 0.88 0.58 1.42 0.24 1.16 0.30 1215.6 62.7 22.9 324.14 48.37 0.51 26.27 383.6 Cladonia spp. 3 4 0.93 0.64 2.02 0.28 0.99 0.69 886.2 74.0 45.1 252.31 51.54 0.58 23.06 395.4 P. nutans 3 1 1.49 1.27 1.60 0.54 1.64 2.93 2576.1 141.1 165.6 872.03 208.66 1.87 82.08 1198.6 Harjavalta 2 km 1992-1994 C. islandica (-92) 3 4 0.69 0.52 2.10 0.24 0.69 0.52 890.7 114.6 31.2 338.25 56.63 0.54 21.81 239.0 P. nutans (-94) 3 1 1.57 2.19 5.68 0.92 1.62 3.55 2332.6 202.4 348.0 1054.71 291.83 2.68 57.85 708.3 Harjavalta 0.5 km 1994 P. nutans 3 1 1.46 2.61 4.22 1.65 1.72 4.24 4022.1 298.2 39.7 1397.19 334.22 3.04 71.39 271.8 Area and year Part n Macronutrients mg/g N% P K Mg S Ca Heavy metals and A1 jig/g Fe Zn Mn Cu Ni Cd Pb A1 Mekrijärvi 1992 C. vulgaris 1 1 1.25 1.16 5.97 1.87 1.63 4.24 65.4 17.0 469.9 4.38 4.21 0.06 2.65 61.5 2 1 0.72 0.74 2.65 0.97 1.10 2.81 97.2 18.1 879.8 5.87 4.79 0.11 3.38 97.0 E. nigrum 1.2 1 0.83 0.99 3.26 1.03 0.78 3.32 61.7 17.5 1023.3 9.59 5.24 0.13 4.07 68.4 V. myrtillus 1 1 1.22 1.32 5.07 1.78 1.02 2.97 49.1 18.4 2580.9 6.71 5.68 0.06 2.93 123.0 2 1 0.64 0.83 3.02 0.92 0.73 2.20 34.5 32.0 2358.9 7.07 4.82 0.14 2.25 141.9 V. uitis-idaea 1 1 0.93 1.14 4.71 1.28 1.07 2.66 44.4 22.3 1054.0 5.62 3.32 0.04 1.27 115.8 2 1 0.73 0.84 3.29 0.91 0.89 2.56 58.0 22.1 1158.4 6.22 3.46 0.08 2.42 136.8 Hämeenkangas 1992 C. vulgaris 1 1 1.29 1.02 5.15 1.17 1.55 3.94 73.3 27.4 351.2 6.57 3.98 0.02 3.81 79.0 2 1 0.81 0.69 2.72 0.69 1.04 2.69 138.8 22.3 556.6 7.05 4.45 0.09 5.14 122.3 3 1 0.80 0.62 2.54 0.64 1.22 3.03 130.3 23.1 707.0 7.07 3.36 0.18 7.02 146.6 V. vitis-idaea 1 1 1.11 0.96 4.11 0.98 1.45 3.93 60.2 33.8 529.0 5.24 2.72 0.01 2.90 120.0 2 1 0.92 0.88 3.41 0.81 1.42 3.79 97.5 47.2 684.6 7.76 3.40 0.07 7.47 169.5 Deschampsia flexuosa 1 1 0.78 1.10 4.61 0.30 2.01 0.70 295.6 40.80 218.4 4.11 11.8 0.17 6.68 186.4 Harjavalta 8 km 1992 C. vulgaris 1 1 1.30 1.13 6.32 1.11 1.80 3.44 112.6 15.8 157.5 15.52 7.52 0.02 1.28 82.3 2 0.61 0.57 3.39 0.68 1.14 2.20 159.7 13.4 537.9 34.29 11.27 0.12 4.82 119.1 E. nigrum 1 1 1.25 1.63 10.49 1.16 1.55 3.52 71.7 20.8 107.8 17.16 9.06 - 1.02 37.8 2 1 0.75 0.78 3.90 0.69 0.79 3.33 128.7 21.5 177.5 30.33 12.41 0.07 4.75 94.8 V. vitis-idaea 1 1 1.08 1.16 5.20 1.31 1.45 4.26 57.2 35.3 579.8 11.75 7.32 - 0.78 118.3 2 1 0.91 0.79 3.32 0.86 1.17 2.59 150.1 43.8 668.5 27.95 12.70 0.20 6.22 197.6 Harjavalta 4 km 1992 E. nigrum 1 1 1.15 0.60 3.18 0.52 1.16 1.83 31.0 8.8 27.5 8.30 4.08 0.05 0.91 13.8 2 1 0.59 0.56 2.53 0.53 0.73 2.61 372.5 31.4 174.4 149.32 36.42 0.38 13.97 174.4 V. vitis-idaea 1 1 0.88 0.89 4.49 0.91 0.88 4.78 91.3 30.7 396.2 30.18 10.31 0.38 2.19 105.2 2 2 0.74 0.67 3.06 0.68 1.36 3.25 192.4 47.1 477.0 79.76 21.68 0.28 10.54 189.9 Haijavalta 2 km 1993 E. nigrum 1 4 1.16 1.50 8.00 0.99 1.30 3.86 189.6 24.9 43.6 32.54 12.48 0.02 2.40 27.0 2 4 0.61 0.84 3.78 0.70 0.97 2.96 1000.0 64.7 108.6 403.40 53.57 0.55 16.69 206.4 3 1 0.56 0.55 1.22 0.53 1.06 3.16 1995.0 114.9 90.5 873.82 103.74 1.10 31.04 363.9 V. vitis-idaea 1 4 0.85 0.99 5.24 0.99 1.88 2.80 142.5 31.3 313.5 31.08 11.78 0.04 1.77 143.3 2 3-4 0.74 0.74 4.48 0.75 2.08 2.46 368.2 46.0 366.6 97.94 22.13 0.15 6.44 228.2 4 1 - 1.34 7.73 0.66 - 2.52 59.5 15.8 102.7 13.78 8.64 - 6.54 53.9 A. uva-ursi 1 4 0.79 1.35 6.87 0.95 0.72 2.95 102.5 30.3 11.0 18.28 8.49 0.03 1.16 33.4 2 4 0.55 1.02 4.88 0.86 0.53 2.10 522.4 95.1 153.8 202.77 35.49 0.35 9.50 163.7 18 M. Salemaa et ai / Science of the Total Environment xx (2003) xxx-xxx References Alloway BJ. Soil processes and the behaviour of metals. In: Alloway BJ, editor. Heavy metals in soils. London: Blackie Academic and Professional, 1995. p. 11-37. Amiro BD, Courtin GM. Patterns of vegetation in the vicinity of an industrially disturbed ecosystem, Sudbury, Ontario. Can J Bot 1981;59:1623-1639. Bagatto G, Shorthouse JD. Accumulation of copper and nickel in plant tissues and an insect gall of lowbush blueberry, Vaccinium uliginosum, near an ore smelter at Sudbury, Ontario, Canada. Can J Bot 1991;69:1483-1490. Balsberg Pählsson AM. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. A literature review. Water Air Soil Pollut 1989;47:287-319. Bates JW. Mineral nutrient acquisition and retention by bryo phytes. J Bryol 1992;17:223-240. Bennett JP. Statistical baseline values for chemical elements in the lichen Hypogymnia physodes. In: Agrawal SB, Agra wal M, editors. Environmental pollution and plant responses. Boca Raton: Lewis Publishers, 1999. p. 343-353. Berg T, Royset O, Steinnes E. Moss (Hylocomium splendens) used as biomonitor of atmospheric trace element deposition: estimation of uptake efficiencies. Atmos Environ 1995;29:353-360. Brown DH, Brumelis G. A biomonitoring method using the cellular distribution of metals in moss. Sci Total Environ 1996;187:153-161. Chernenkova TV, Kuperman RG. Changes in the forest com munities along a heavy metal deposition gradient on Kola Peninsula. Water Air Soil Pollut 1999;111:187-200. Chettri MK, Sawidis T, Zachariadis GA, Stratis JA. Uptake of heavy metals by living and dead Cladonia thalli. Environ Exp Bot 1997;37:39-52. Conti ME, Cecchetti G. Biological monitoring: lichens as bioindicators of air pollution assessment—a review. Environ Pollut 2001; 114:471 -492. Dahlquist RL, Knoll JW. Inductively coupled plasma-atomic emission spectrometry: analysis of biological materials and soils for major, trace and ultra-trace elements. Appl Spec trosc 1978;32:1-30. Derome J. Effects of heavy-metal and sulfur deposition on the chemical properties of forest soil in the vicinity of a Cu-Ni smelter, and means of reducing the detrimental effects of heavy metals. Finn For Res Inst, Research papers 2000;769:1-77. Derome J, Lindroos AJ. Effects of heavy metal contamination on macronutrient availability and acidification parameters in forest soil in the vicinity of Harjavalta Cu-Ni smelter, SW Finland. Environ Pollut 1998;99:225-232. Derome J, Nieminen T. Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Environ Pollut 1998;103:219-228. Ernst WHO. Bioavailability of heavy metals and decontami nation of soils by plants. Appl Geochem 1996;11:163-167. Folkeson L, Andersson-Bringmark E. Impoverishment of veg etation in a coniferous forest polluted by copper and zinc. Can J Bot 1988;66:417-428. Garty J. Biomonitoring atmospheric heavy metals with lichens: theory and application. Crit Rev Plant Sci 2001;20:309- 371. Hall JL. Cellular mechanisms for heavy metal detoxification and tolerance. J Exp Bot 2002;53:1-11. Halleraker JH, Reimann C, de Caritat P, Finne TE, Kashulina G, Niskavaara H, Bogatyrev I. Reliablity of moss (Hylocom■ ium splendens and Pleurozium schreberi) as a bioindicator of atmospheric chemistry in the Barents region: interspecies and field dublicate variability. Sci Total Environ 1998;218:123-139. Hämet-Ahti L, Suominen J, Ulvinen T, Uotila P, editors. Retkeilykasvio (Field Flora of Finland), Edn. 4. Finnish Museum of Natural History, Botanical Museum, Helsinki, 1998, p. 656. Haslett BS, Reid RJ, Rengel Z. Zinc mobility in wheat: uptake and distribution of zinc applied to leaves and roots. Ann Bot 2001;87:379-386. Helmisaari HS. Nutrient cycling in Pinus sylvestris stands in Eastern Finland. Plant Soil 1995;168-169:327-336. Helmisaari HS, Mälkönen E. Acidity and nutrient content of throughfall and soil leachate in three Pinus sylvestris stands. Scand J For Res 1989;4:13-28. Kashulina G, Reimann C, Banks D. Sulfur in the Arctic environment (3): environmental impact. Environ Pollut 2003;124:151-171. Koponen T, Isoviita P, Lammes T. The bryophytes of Finland: an annotated checklist. Flora Fennica 1977;6:1-77. 3 2-3 0.68 0.80 2.08 0.66 0.72 4.86 933.2 120.6 106.7 413.29 78.59 1.02 22.09 282.2 4 1 0.61 1.22 7.39 0.62 0.43 1.67 44.1 16.5 8.6 12.41 11.99 - 6.11 4.1 Harjavalta 0.5 km 1993 E. nigrum 1 3 1.38 1.72 7.00 1.33 1.72 3.40 2981.0 174.6 215.7 183.69 50.55 0.37 22.40 119.2 2 4 0.72 0.90 2.88 0.94 1.87 2.60 8632.4 446.6 276.8 1453.66 166.96 1.79 99.54 411.2 3 1-4 0.70 1.40 2.38 1.54 2.14 4.82 20570.8 1041.2 280.9 4134.83 473.12 4.12 229.31 967.4 V. uliginosum 1 4 1.43 2.05 10.89 0.82 2.51 2.67 451.7 67.8 138.9 38.07 42.25 0.62 5.68 57.2 2 4 0.66 0.71 2.07 0.38 0.88 1.83 2157.0 162.5 137.1 329.94 61.76 0.99 36.64 131.7 3 1 - 0.84 4.52 0.99 - 3.66 4037.2 235.7 60.1 561.06 86.06 1.17 43.04 157.5 4 1 - 1.57 6.86 0.41 2.44 1.12 590.4 47.6 21.6 55.49 15.56 0.20 9.87 93.8 C. globularis 1 1 - 1.41 21.29 0.56 - 2.11 1625.1 242.7 233.5 254.65 117.02 0.74 13.64 72.4 3 1 1.53 1.02 1.92 0.66 - 1.95 16984.0 844.3 140.0 3978.90 488.11 4.64 209.60 888.2 M. Salemaa et ai. / Science of the Total Environment xx (2003) xxx-xxx 19 Kozlov MV, Haukioja E, Bakhtiarov AV, Stroganov DN, Zimina SN. Root vs. canopy uptake of heavy metals by birch in an industrially polluted area: contrasting behaviour of nickel and copper. Environ Pollut 2000;107:413^t20. Kulmala A, Leinonen L, Ruoho-Airola T, Salmi T, Walden J. Air quality trends in Finland. Air quality measurements. Helsinki: Finnish Meteorological Institute, 1998. p. 91 Lepp NW, Salmon D. A field study of the ecotoxicology of copper to bryophytes. Environ Pollut 1999;106:153-156. Mälkönen E, Derome J, Kukkola M, Saarsalmi A. Improving forest condition by means of slow-release fertilisers. In: Mälkönen E, editor. Forest condition in a changing environ ment—the Finnish case. Dordrecht: Kluwer Academic Pub lishers, 2000. p. 301-323. Mälkönen E, Derome J, Fritze H, Helmisaari HS, Kukkola M, Kytö M, Saarsalmi A, Salemaa M. Compensatory fertiliza tion of Scots pine stands polluted by heavy metals. Nutr Cycl Agroecosys 1999;55:239-268. Manninen S, Huttunen S. Scots pine needles as bioindicators of sulfur deposition. Can J For Res 1995;25:1559-1569. Marschner H. Mineral nutrition of higher plants, 2nd ed. London: Academic Press, 1995. p. 889 Meharg AA, Cairney JWG. Co-evolution of mycorrhizai sym biosis and their hosts to metal-contaminated environments. Adv Ecol Res 2000;30:69-112. Monni S, Salemaa M, Millar N. The tolerance of Empetrum nigrum to copper and nickel. Environ Pollut 2000a;109:221 -229. Monni S, Salemaa M, White C, Tuittila E, Huopalainen M. Copper resistance of Calluna vulgaris originating from the pollution gradient of a Cu-Ni smelter, in Southwest Finland. Environ Pollut 2000b;109:211-219. Nieminen T, Derome J, Helmisaari HS. Interactions between precipitation and Scots pine canopies along a heavy-metal pollution gradient. Environ Pollut 1999;106:129-137. Pakarinen P. Nutrient and trace metal content and retention in reindeer lichen carpets of Finnish ombrotrophic bogs. Ann BotFenn 1981;18:265-274. Poikolainen J, Kubin E, Piispanen J, Karhu J. Atmospheric heavy metal deposition in Finland during 1985-2000 mosses as bioindicators. Sci Total Environ in prcasT Rautio P, Huttunen S. Total vs. internal element concentrations in Scots pine needles along a sulfur and metal pollution gradient. Environ Pollut 2003;122:273-289. Reimann C, Halleraker JH, Kashulina G, Bogatyrev I. Com parison of plant and precipitation chemistry in catchments with different levels of pollution on the Kola Peninsula, Russia. Sci Total Environ 1999;243-244:169-191. Reimann C, Niskavaara H, Kashulina G, Filzmoser P, Boyd R, Voiden T, Tomilina O, Bogatyrev I. Critical remarks on the use of terrestrial moss (Hylocomium splendens and Pleurozium schreberi) for monitoring airborne pollution. Environ Pollut 2001;113:41-57. Rigina O, Kozlov MV. Pollution impact on sub-arctic forests in the Kola Peninsula, Russia. In: Innes JL, Oleksyn J, editors. Forest dynamics in heavily polluted regions. Wal lingford: CAB International, 2000. p. 37-65. Ross SM. Retention, transformation and mobility of toxic metals in soils. In: Ross SM, editor. Toxic metals in soil plant systems. Chichester: John Wiley, 1994. p. 63-152. Saari H, Kartastenpää R, Pohjola V, Voltti U. Particle and sulfur dioxide research at Harjavalta in the spring 1992. Helsinki: Finnish Meteorological Institute, 1993. p. 66 (In Finnish. Hiukkas- ja rikkidioksiditutkimus Harjavallassa keväällä 1992). Salemaa M, Monni S. Copper resistance of the evergreen dwarf shrub Arclostaphylos uua-ursi: an experimental expo sure. Environ Pollut 2003;126:435-443. Salemaa M, Vanha-Majamaa I, Derome J. Understorey vege tation along a heavy-metal pollution gradient in SW Finland. Environ Pollut 2001;112:339-350. Tyler G. Uptake, retention and toxicity of heavy metals in lichens. A brief review. Water Air Soil Pollut 1989;47:321 333. Tyler G. Bryophytes and heavy metals: a literature review. Bot J Linn Soc 1990;104:231-253. Tyler G, Balsberg Pählsson AM, Bengtsson G, Bääth E, Tranvik L. Heavy metal ecology of terrestrial plants, micro organisms and invertebrates. Water Air Soil Pollut 1989;47:189-215. Uhlig C, Junttila O. Airborne heavy metal pollution and its effects on foliar elemental composition of Empetrum her maphroditum and Vaccinium myrtillus in Sor-Varanger, Northern Norway. Environ Pollut 2001; 114:461 Uhlig C, Salemaa M, Vanha-Majamaa I, Derome J. Element distribution in Empetrum nigrum microsites at heavy metal contaminated sites in Harjavalta, Western Finland. Environ Pollut 2001; 112:435 —442. Väisänen S. Effects of air pollution by metal, chemical and fertilizer plants on forest vegetation at Kokkola, W Finland. Ann Bot Fenn 1986;23:305-315. Vitikainen O, Ahti T, Kuusinen M, Lommi S, Ulvinen T. Checklist of lichens and allied fungi of Finland. Norrlinia 1997;6:1-123. Wellburn A. Air pollution and acid rain. The biological impact. London: Longman group, 1988. p. 274 Zechmeister HG. Growth rates of five pleurocarpous moss species under various climatic conditions. J Bryol 1995; 18:455 -468. Zechmeister HG, Grodzinska K, Szarek-Lukaszewska G. Bry ophytes. In: Markerts BA, Breure AM, Zechmeister HG, editors. Bioindicators and biomonitors. Amsterdam: Elsev ier, 2003. p. 329-375. Paper III Salemaa, M. & Uotila, T. 2001. Seed bank composition and seedling survival in forest soil polluted with heavy metals. Basic and Applied Ecology 2: 251-263. Basic and Applied Ecology Basic Appi. Ecol. 2, 251-263 (2001) © Urban & Fischer Verlag http://www.urbanfischer.de/sournals/baeco! 1439-1791/01/02/03-251 $ 15.00/0 Seed bank composition and seedling survival in forest soil polluted with heavy metals Maija SalemaaI *, Taru Uotila 2 'Vantaa Research Centre, The Finnish Forest Research Institute, RO. Box 18, FIN-01301 Vantaa, Finland ■Department of Limnology and Environmental Protection, University of Helsinki, PB 62 Info Centre, Viikinkaari 11 A, FIN-00014 University of Helsinki, Finland Received January 2, 2001 ■ Accepted March 26, 2001 Abstract Seedling recruitment from forest soil polluted with heavy metals was studied in order to determine the revegetation potential of the seed bank. The soil samples were collected from untreated and fertilised Scots pine stands along a 8 km transect running SE from a copper-nickel smelter in SW Finland. The composition and size of the active seed bank and the survival of the germinated seedlings were studied in greenhouse conditions. The average densities of germinated seeds of 15 species ranged from 250 to 4750 plants per m 2 at the six sites. Although vegetation was almost totally absent near the smelter, germinable seeds of Betula pubescens, Calluna vulgaris, Pinus sylvestris and Vaccinium uliginosum were found in the most contaminated soil. The number of Calluna vulgaris seedlings increased with increasing distance from the smelter, but no such trend was found for the other species. The mortality rate of the seedlings was highest in the soil samples collected near the smelter. Nu trient addition (stand-specific fertilisation with N, Ca, P and Mg) did not affect the number of germinated seeds, but liming slightly reduced heavy metal induced death of the seedlings. The sur vival probability of Calluna vulgaris seedlings decreased with proximity to the smelter. The results suggest that the recovery of Calluna vulgaris near the smelter was prevented by a low availability of seeds in the soil, unfavourable germination conditions and unsuccessful seedling establishment. We conclude that, despite the presence of viable seeds in polluted soil, revegetation from seed banks is not successful without soil mitigation to immobilise heavy metals. Das Revegetationspotential einer Samenbank in einem mit Schwermetall belasteten Waldboden wurde in einem Treibhausexperiment untersucht. Die Bodenproben stammten von insgesamt 6 Kontroll- und Diingungsflächen, die auf einem 8 km langen Transekt angelegt worden waren. Der Ausgangspunkt des Transekts war eine Kupfer-Nickel-Giefterei im Siidwesten Finnlands. Die Samenbanken brachten 15 Planzenarten hervor, deren Mittelwerte von 250 bis 4750 Exemplaren pro m 2 variierten. Trotz einer fast nicht existierenden Vegetation neben der Giefierei wies die Samen bank des belasteten Standorts Sämlinge der Arten Betula pubescens, Calluna vulgaris, Pinus sylvestris and Vaccinium uliginosum auf. Nur die Anzahl der Calluna vulgaris Sämlinge nahm zu mit wachsender Distanz zur GieSerei. Die Mortalität der Sämlinge war am höchsten in den belasteten Bodenproben neben der Giefierei. Eine standortspezifische Diingung mit N, Ca, P und Mg erzielte das gleiche Ergebnis, aber durch Kalkung konnte die Mortalität der Sämlinge geringfiigig vermindert werden. Die Überlebensrate der Calluna vulgaris Triebe verminderte sich mit abnehmender Distanz zur GielSerei. * Corresponding author: Maija Salemaa, Vantaa Research Centre, The Finnish Forest Research Institute, P.O. Box 18, FIN-01301 Vantaa, Finland, Phone: ++3sB-9-8570 5565, Fax: ++3sB-9-8570 5569, E-mail: maija.salemaa@metla.fi 252 Salemaa and Uotila Basic Appl. Ecol. 2, 3 (2001) Unsere Ergebnisse zeigen, dass die Ansiedlung der Art Calluna vulgaris in der Nähe der GieSerei durch die niedrige Anzahl von Samen im Boden, schlechte Keimungsbedingungen und erhöhte Mor talität kontrolliert wird. Eine Revegetation kann trotz keimungsfähiger Samen in mit Schwermetall belasteten Boden nicht erfolgen, solange die Schwermetalle nicht immobilisiert werden. Key words: actuarial life tables - Calluna vulgaris - fertilisation - liming - revegetation Introduction Many studies on post-fire and clear-cut succession (Moore & Wein 1977, Archibold 1979, Johnson 8c Bradshaw 1979) and gap dynamics of forests (Mlade noff 1990, McGee & Feller 1993) emphasise the im portance of soil seed banks in the regeneration of forest vegetation after a disturbance. The potential role of seed banks in restoring ecologically valuable communi ties, e.g. in heathlands (Pywell et al. 1997, Mitchell et al. 1998) and forests (Halpern et al. 1999, Onaindia & Amezaga 2000), has also been studied. However, only a few studies have focused on seed bank dynamics after a disturbance caused by airborne pollution of forest soil (Vieno et al. 1993, Komulainen et al. 1994, Huopalainen et al. 2000, Winterhalder 2000). The zonal degradation of forest vegetation has been widely documented around smelter complexes emit ting heavy metals, sulphur and other pollutants in the boreal coniferous zone (Freedman & Hutchinson 1980, Folkeson & Andersson-Bringmark 1988, Aam lid et al. 2000, Rigina & Kozlov 2000, Salemaa et al. 2001). The absence of a vegetation cover enhances the wind erosion of metal-contaminated particles, de creases the water-holding capacity of the soil and facil itates the leaching of heavy metals into the ground water (Vangronsveld et al. 1996, Derome & Nieminen 1998). Revegetation of degraded sites is therefore es sential to stabilise the soil and to reduce the adverse environmental impacts (Johnson & Bradshaw 1979, Kiikkilä et al. 2001). Recolonization of the vegetation in disturbed areas can be realised via (1) propagules (e.g. seeds, spores, buds, rhizomes or roots) stored in the soil, (2) propagules dispersed from the neighbour ing areas, or (3) vegetative or sexual reproduction of the existing remnants of the plant populations (Moore 8c Wein 1977, Komulainen et al. 1994). Although heavy metal deposition has considerably decreased during the last decades in northern Europe (Kubin et al. 2000), natural recolonization of the vege tation in heavily polluted areas still faces many prob lems. In areas where the vegetation has been absent for decades, the shortage of propagules may be a factor limiting revegetation. Furthermore, the initial develop ment of the seedlings is sensitive to metals and acidity (Bradley et al. 1981, Patterson 8c Olson 1983, Bell & Teramura 1991). Natural recovery of the vegetation in industrial barrens, e.g. in the Sudbury area, Canada, has proved to be a very slow process without soil miti gation and the sowing or replanting of tolerant plant species (Winterhalder 2000). Information about the germination capacity of the soil seed banks and the critical early phases of seedling establishment helps us to choose the most viable method for restoring forest areas suffering from the impacts of severe pollution. In this paper we evaluate the possibility of re-establishing the native understorey vegetation of a Scots pine forest damaged by the depo sition of heavy metals and sulphur. Our aims were to determine: (1) the size and species composition of the active seed bank in Scots pine stands at different distances from a Cu-Ni smelter, (2) the effect of soil fertilisation and liming on the number of germinable seeds and the mortality rate of seedlings growing in soil samples representing differ ent pollution levels, and (3) the survival of Calluna vulgaris seedlings germi nated from the seed bank samples collected along the pollution gradient. Materials and methods Study area The seed banks in forest soil were studied along a 8 km transect running SE from the Harjavalta Cu-Ni smelter (61°19'N, 22°09'E) in SW Finland. This was the only direction with suitable coniferous stands near the smelter. The study area is dry and infertile Scots pine (Pinus sylvestris L.) forest. The soil is sorted fine sand and the soil type ferric podsol (Mälkönen et al. 1999). The Cu smelter was founded in 1945 and Ni smelt ing started in 1960. More than 50 years' accumulation of heavy metals and sulphur has drastically changed the forest ecosystem (Helmisaari et al. 1995, Fritze et al. 1997, Derome 8c Lindroos 1998, Derome 8c Nieminen 1998). The understorey vegetation, which was originally dominated by Calluna vulgaris (L.) Hull and a well-developed moss and lichen layer, was al most totally absent up to a distance of 0.5 km from the smelter in the beginning of the 1990'5. Only some Seed bank composition and seedling survival in forest soil 253 Basic Appl. Ecol. 2, 3 (2001) Table 1. Stand characteristics, element concentrations in the organic soil layer and the coverage percentages of needle litter (visual estimation) and plant species (point frequencies). Plant-available concentrations of Cu and Ni in the organic layer were determined by extraction with NH 4 acetate + EDTA and expressed on the basis of the organic matter (om) content (for details, see Salemaa et al. 2001). Table 2. The effect of stand-specific fertilisation (SSF) on the element con centrations (mg/kg) and pH of the organic soil layer at three distances from the Harjavalta smelter in 1996 (data from Derome 2000). Plant available con centrations of Cu, Ni, Ca, Mg and P were determined by extraction with BaCl2+ EDTA. All concentrations have been recalculated per dry mass of or ganic matter. The statistical differences between the averages of the control and SSF plots (n = 3) are given by * (P < 0.05) and ° (P < 0.10) (t-tests). Nutrient applications (kg ha -1 ) in 1992: 0.5 km: 150N, 50 Mg, 1500 Ca 4 km: 150 N, 30 P, 30 Mg, 1000 Ca, 8 km: 120 N patches of the most resistant vascular plant species e.g. Empetrum nigrum ssp. nigrum L. (Uhlig et al. 2001), Carex globularis L. and Vaccinium uliginosum L. were present. The overstorey pines were alive, but they were stunted and seriously defoliated. The coverage and floristic diversity of the understorey vegetation in creased with increasing distance from the smelter (Salemaa et al. 2001). The characteristics of the tree stands, the existing vegetation and the concentrations of exchangeable Cu, Ni and total S and N in the or ganic soil layer are given in Tab. 1. Emissions from the smelter were considerably re duced in the 1990'5. During 1985-1990 the average annual emissions of Cu were 104 t, Ni 50 t and S02 8100 t. In 1993, the year before the seed bank samples were collected, the corresponding values were Cu 50 t, Ni 71 and S0 2 47001 (Helmisaari et al. 1995). Sampling and fertiliser treatments The soil samples for the seed bank analyses were col lected from untreated sites at six distances (0.5,1,2, 3, 4 and 8 km), and from fertilised sites located at 0.5, 4 and 8 km from the Harjavalta smelter on 5 May 1994. The sites at distances of 0.5,2, 4 and Bkm represented experimental plots in a forest correction fertilisation project (1992-1996) carried out to alleviate detrimen tal changes in the forest soil caused by heavy metal and S0 2 deposition (Mälkönen et al. 1999). The sam ples were collected from the plots in one (stand-specif ic fertilisation) of the four fertilisation treatments. There were three replicates of each untreated or fer tilised plot (30 x 30m). Five soil samples were taken from the buffer zone of each plot. The soil samples were taken from the organic layer (including the litter and the upper part of the mineral soil layer) to a depth of 5-10 cm using a 9.5 x 9.5 cm metal frame. The stand-specific fertiliser treatment was applied as a single broadcast dose in spring 1992. The composi tion of the fertilisers was different at the individual dis tances (Tab. 2) and was based on soil and needle analy ses in each Scots pine stand (Mälkönen et al. 1999, Derome 2000). Nitrogen was given in the form of slow release methylene urea and fast-release ammonium ni trate, and Ca and Mg as Mg-rich granulated limestone. The aim of liming and fertilisation was to reduce heavy metal toxicity and to correct nutrient imbalances in order to improve the vitality of the tree stand. Distance from the smelter, km 0.5 1 2 3 4 8 Pine stand: Stand age, years 49 67 52 51 56 40 Mean height, m 6.1 10.1 10.9 9.5 9.2 10.6 Number of trees ha -1 1008 1048 1230 1436 1517 1552 Element concentrations in the organic soil layer before fertilisation (1992): Thickness of organic 2.5 7.2 2.3 3.3 2.2 0.8 layer, cm N tot % om 1.92 1.31 1.74 1.58 1.65 1.68 S, tot mg/kg om 387 402 198 176 170 176 Cu, exc. mg/kg om 7540 3501 2238 1532 786 209 Ni, exc. mg/kg om 527 406 329 230 164 72 Coverage (%): Needle litter 83.3 90.3 74.3 850 33.0 31.7 Mosses and lichens 0.17 3.28 28.14 32.77 58.94 76.25 Calluna vulgaris 0.00 0.01 0.01 0.01 1.35 1.41 Vascular plant species 0.48 21.01 7.83 13.73 18.44 9.52 All plant species 0.65 24.29 35.97 46.50 77.38 85.77 Distance from the smelter, km 0.5 4 8 Cu control 4517.8 210.1 39.1 SSF 2520.9* 91.5» 37.4 Ni control 697.1 143.5 61.5 SSF 495.7 96.9 53.1 Ca control 904.3 2295.8 2357.7 SSF 5550.9* 5891.7* 1915.0 Mg control 105.8 256.1 316.6 SSF 557.1* 693.1* 218.0 P control 40.4 127.9 151.1 SSF 406 187.6 113.8 N tot control 2.55 1.67 1.89 SSF 2.01 1.54 1.31 pH control 3.57 3.11 3.20 SSF 4.25 3.80 3.23 254 Salemaa and Uotila Basic Appl.Ecol. 2,3(2001) Fig. 1. The relationships between a) the total number of germinated seedlings (excluding Betuja pubescens) (y) and the abundance of vascular plant species in the existing vegetation measured as point frequencies, % (x), b) the number of Calluna vulgaris seedlings (y), c) the mortality rate of seedlings of all species (in cluding Betula pubescens) (y), and d) the mortality rate of the Calluna vulgaris seedlings (y) and the distance from the smelter, km (x). The plant available concentrations of heavy metals and macronutrients in the organic soil layer four years after the treatments are given in Tab. 2. The limestone application at distances of 0.5 and 4 km considerably reduced the Cu and to some extent also the Ni concen trations in the organic layer. The pH of the organic layer increased slightly at the corresponding distances. Nitrogen application at 8 km had no significant effects on the heavy metal concentrations of the organic soil layer (Derome 2000). Ca and Mg concentrations were higher in the fertilised than in the untreated plots at 0.5 and 4 km. At 8 km, however, the concentrations were lower on the fertilised plots, presumably due to natural inter-plot variation. (Tab. 2). Only very slight changes occurred between 1992 and 1994 in the understorey vegetation after fertilisa tion. The coverage of Calluna vulgaris increased by a few percentage units. Some new patches of Epilobium angustifolium L. and Deschampsia flexuosa (L.) Trin. also appeared (unpublished results). Germination The germination experiment with a total of 135 soil samples (total surface area 1.2 m 2) was carried out in the greenhouse of the Finnish Forest Research Institute at Tuusula (60°21'N, 25°00'E). Because the samples were taken immediately after the end of winter, stratifi cation was not considered necessary. Rhizomes and large roots were removed and the samples spread out on trays (18.5 x 21.5 cm). The trays contained a 2-3 cm layer of mixed quartz sand and peat as growth sub strate. Six control trays containing only substrate were used to check whether any seeds originated from the greenhouse environment. The trays were arranged ran domly on greenhouse tables. The light conditions were natural. The temperature was maintained at +2O"C during the day and +l5°C at night. However, the tem perature occasionally reached 30° C on sunny days. The relative humidity in the greenhouse was 60-70%. The samples were watered several times a week. Seed bank composition and seedling survival in forest soil 255 Basic Appl. Ecol. 2, 3 (2001) The emerged seedlings were counted once a week, and the species were identified as soon as possible. The seedlings were left to grow on the trays, the location of each seedling having been marked on a map. The seedlings of Calluna vulgaris were monitored individu ally up until 21 weeks (from 10 May to 19 October 1994), and the other species up until 16 weeks. Some gramineous seedlings were replanted when the trial was finished and identified at a later stage. Statistical analysis The five seed bank samples from each plot were germi nated individually, thus making it possible to study both within- and between-plot variation in the re sponse variables. The statistical analyses were per formed on all species combined, and separately on Be tula pubescens Ehrh. and Calluna vulgaris when there was sufficient data. The ln(x+l) transformed numbers of emerged seedlings were compared by nested ANOVA, in which the sample was nested under the plot and the plot under the distance from the smelter (GLM proc., SAS Institute Inc. 1994). The plot means of fertilised and untreated samples were compared pairwise within each distance using Kruskal-Wallis tests based on chi-square values (NPARIWAY proc., SAS Institute Inc. 1994). The effect of the fertilisation could not be compared across the distances, because of the varying treatments. The relationships between the number of germinants and the abundance of the exist ing vegetation, the distance from the smelter and the heavy metal concentrations in the organic layer were analysed by regression models (REG and NLIN proc., SAS Institute Inc. 1994). The corresponding tests were also carried out on the mortality rates of the seedlings. The correspondence between the species composi tion of the existing vegetation and the seed bank at each distance was estimated by Serensen similarity in dices (Sorensen 1948) using the presence-absence data. Similarity = 2W/A+B, where A and B are the numbers of species occurring separately in the aboveground vegetation (A) and in the soil seed bank (B), and W the number of species common to A and B. Actuarial life tables were used for analysing the sur vival probability of the Calluna vulgaris seedlings over time (SPSS 9.0.1 software, SPSS Inc. 1999). If a seedling did not die during the experiment, the case was handled as censored. The survivor function gives the probability that a seedling will survive for a speci fied time at least, or longer, without a response (death) (McCullagh & Nelder 1989). The equality of the sur vival functions of the seedlings originating from differ ent distances from the smelter, and between the fer tilised and untreated samples, were compared using the Wilcoxon (Gehan) statistic. Results The size and composition of the seed banks Viable seeds were found at all distances from the smelter. Altogether 1300 seedlings germinated in the total of 135 seed bank samples. The emerged seedlings represented 15 taxa, of which 6 species were grasses and sedges, 4 dwarf shrubs, 3 trees and 2 herbs (Tab. 3). The most numerous species were Hetula pubescens (696 seedlings) and Calluna vulgaris (490), both of which were found at all distances from the smelter. The average seedling density varied from 250 to 4750 per m 2 at the different distances (Tab. 3). Excluding the seedlings of Betula pubescens, which were most probably derived from the previous-year seed crop, the average size of the persistent seed banks ranged from 15 to 1200 seeds per nr. There was very high varia tion in seedling numbers between the plots at the same distance, but no differences within each plot (Tab. 4, Fig. 1). No seedlings emerged from the control trays. Although aboveground vegetation was very scanty at 0.5 km, the samples contained germinable seeds of Betula pubescens, Calluna vulgaris, Vaccinium ultgi nosum and Pinus sylvestris (the last species only on the fertilised plots). Betula pubescens was the most nu merous species at 0.5 and 1 km, and Calluna vulgaris dominated in the seed bank at distances of 2 km or more. In addition to birches, a large number of Carex globularts seedlings emerged from the samples taken at 1 km. "Vaccinium vitis-idaea L. appeared for the first time in the seed bank at 1 km. Only one Empetrum ni grum seedling was found at a distance of 4 km. Seedlings of the other species, especially grasses and herbs, also appeared in low numbers and sporadically along the study transect (Tab. 3). Germination in relation to the existing vegetation, the dis tance from the smelter and the fertilisation treatments The seed bank species were rather well represented in the existing vegetation. Of the 17 vascular plant species growing on the untreated plots, 11 were found in the seed banks (Tab. 5). The percentage similarity between the species in the existing vegetation and the seed banks varied from 18% to 67% at different dis tances from the smelter. The total similarity in the data for the whole transect was about 70% in the untreat ed, and 60% in the fertilised plots (Tab. 5). The four species that were absent from the existing vegetation but present in the seed bank were Festuca ovina L., Agrostis capillaris L., Carex ericetorum Pol lich and Rumex acetosella L. In contrast, no seedlings of Arctostaphylos uva-ursi (L.) Sprengel, Ledum palustre L. and Vaccinium myrtillus L. emerged from 256 Salemaa and Uotila Basic Appi. Ecol. 2, 3 (2001) Table 3. a) The average number of germinated seeds per m 2 ±seinthesoilsamplescollectedfromtheuntreatedandfertilisedplotsat different distances from the Harjavalta smelter. The averages have been cal culated from the plot-specific data (n = 3 plots), each including 5 soil samples, b) The average mortality rate (%) of the seedlings during the 16 week germination trial. The plots where no seedlings emerged were excluded from the mortality rate data (plot number lower than 3 is given in parantheses). One seedling is equivalent to a density of 7.4 per m 2. rsj Ö rsi 00 CO m ö rn O 35 ±15(2) 24 ±5 OO ro •H m , 3 , 1 ■H i l rn 1 -H OO 1 1 1 «- 1 7.4 ±7.4 +1 Ln eri 1 rsi +1 OO i/S m ■H Ö CO "5r 8 ± 8(2) 30 ±11 •^r 7.4 ±7.4 524.5 ±372.9 14.8 ±14.8 7.4 ±7.4 7.4 ±7.4 29.5 ±19.5 7.4 ±7.4 590.9 ±394.8 598.3 ±390.4 O 10 ± 7 23 ± 15 17 ± 9 "O a n> "O ■«a- ö rs. is! LO CO rsi ib VO CO S S E di LH CO iri rn rs" ■H -H rsi p rn 00 00 VO -H rsj +1 rn u~t -H -H O 1 1 1 1 *— 1 LTI rsi 1 1 1 1 1 S rsj iri m m eri S LO rn S OO 7.4 ±7.4 7.4 ±7.4 7.4 ±7.4 1 m ■H cr» 00 1 *— 1 7.4 ±7.4 1 1 1 LH M O S rsj irt -H -=t rsi 100(1) 17 ±4 0(2) 20 ±8 "3- 4 36.9 ±26.6 1 Ml. 4 509.7 ± 315.2 7.4 ±7.4 7.4 ± 7.4 7.4 17.4 7.4 ± 7.4 14.8 ± 7.4 !.2 22.2 ±12.8 199.4 583.6 ±301.8 192.1 620.5 ± 322.9 25 ± 25(2) 31 ±1(2) 33 ±22 40 ±1 fs ■H ■H rs eri rsi -H rsi rsi rsi | -H OO m LO +4 en ö o» m cn -H rs in •H S" "H m 1 En | 1 1 tn 1 l 1 l 1 1 1 rn rsi *— rsj CO rsi ■H m 3' CO m ■H p 00 -H -H rsi +< "3" rs +1 CO -H 0 OO rsj fsj" O m m 5 ±5(2) 51 ±2 rsi rsi 1 rsi | § rn m 1 •— l rsi 1 1 N rsj rs 1 «- eri rsj rsj in rsj O in rs rs m LT» rsi ■H O vo cr, 0 oS OO VO rs1 rsj -H -H rs ct» vo ir» in eri -H CT» CO en rsi vo «— m -H -H rsj rs rsi m -H rs OO '49.6 13062.6 rsj +1 LD «— -H -H - 1 § 1 1 m rsi 1 1 1 1 rn | 1 rsj rn rs OO oo S rs ■o c H3 rsi m "O LT» O rsj ■h rs! 1 Ln 1 rs! ■H *3- 1 r< 1 3 1 1 1 1 1 ' im 7.4 ±7.4 ■a 1 14.8 ±7.4 m -H CO m in 93 ±3 100(1) 0(1) 93 ±4 Distance from the smelter, km a) Germinated seec Agrostis apillaris Betula pubescens Carex canescens Carex ericetorum Carex globularis Calluna vulgaris Deschampsia flexui Empetrum nigrum Epilobium spp. Festuca ovina Pinus sylvestris Populus tremula Rumex acetosella Vaccinium uliginosi Vacdnium vitis-idae All excluding Betuli All species b) Mortality rate, % Betula pubescens Calluna vulgaris Other species All species Seed bank composition and seedling survival in forest soil 257 Basic Appl. Ecol. 2, 3 (2001) Table 4. The effect of the distance from the smelter on a) the number of germi nated seedlings, ln(x + 1) transformation, and b) the mortality rate (%) of the seedlings. F and P values for nested ANOVA, df = degree of freedom. Data from untreated plots. Table 5. The number of vascular plant species in the aboveground vegeta tion (AV) and in the soil seed bank (SB) in the a) untreated and b) fertilised plots. Common species (C) were found both in the aboveground vegetation and in the seed banks. Percentage similarity (Sim%) was measured by Sørensen index (100% = max similarity). Data for three plots combined at each distance. the soil samples, although these species were present in the existing vegetation. The total density of the germinated seeds (Betula pubescens excluded) was positively related (r 2 = 58%, P < 0.01) to the total cover of the vascular species in the existing vegetation on the untreated plots (Fig. la), but there were no similar species-specific trends. For instance, Calluna vulgaris was absent or very scarce in the existing vegetation up to a distance of 4 km, al though the density of viable seeds varied between 7-100 per m 2 even at 0.5 km (Tab. 3). The number of Calluna vulgaris and Betula pubescens, as well as the total number of germinated seedlings, varied between different distances (Tab. 4). Calluna vulgaris was the only species in which the ger minant number increased according to increasing dis tance from the smelter (r 2 = 66%, P < 0.01) (Fig. lb). A similar trend was also found in the combined data of all species (excluding Betula pubescens) (r 2 = 59%, P < 0.01). The fertilisation treatments had no effect on the number of germinated seeds (Tab. 6). Mortality rate The mortality rate of the seedlings of all species was the highest in the untreated soil samples collected near the smelter (Tab. 3, Tab. 4). The mortality rate de creased with increasing distance from the smelter in the data of all species (r 2 = 61%, P < 0.001) (Fig. lc) and in Calluna vulgaris (r 2 = 56%, P < 0.01) (Fig. Id). The mortality rate of Calluna vulgaris was slightly lower (P = 0.121) on the fertilised and limed plots at Table 6. The Kruskal-Wallis statistics (x 2 ) for the effect of the stand-specific fertilisation on the number of germinated seeds and the mortality rate (%) of the seedlings. The means of the untreated and fertilised plots are com pared separately at each distance, because the fertilisation treatment was not comparable over the distances, n = number of untreated/fertilised plots. -= insufficient data. 'Mortality rate data from untreated plots in 0.5 and 1 km because small n 0.5 km (56%) compared to that on the untreated plots at distances of 0.5-1 km (80%). A slightly decreasing trend in the mortality rate of the Calluna vulgaris (P = 0.083) and in the data for all species (P = 0.127) was also observed in the fertilised and limed plots at 4 km (Tab. 6). On the other hand, nitrogen fertilisation alone had no effect on heavy metal and macronutrient concentrations in the organic soil layer at 8 km Source of Betula pubescens Calluna vulgaris All species variation vaiia uui i df F P df F P df F P a) Distance 5 25.70 0.000 5 5.92 0.000 5 6.55 0.000 Plot 10 3.65 0.001 10 3.19 0.002 10 2.67 0.009 (distance) Sample 12 0.71 0.736 12 0.67 0.769 12 0.73 0.716 (plot) b) Distance 5 4.90 0.006 5 4.89 0.005 5 12.12 0.000 Plot 6 10.96 0.090 5 0.67 0.653 10 2.00 0.067 (distance) Sample 12 7.05 0.021 12 0.94 0.535 12 1.22 0.308 (plot) Distance, km AV SB C Sim% a) 0.5 6 3 2 44.4 1 12 6 6 66.7 2 11 7 4 44.4 3 8 3 1 18.2 4 8 8 4 50.0 8 10 5 2 26.7 total 17 15 11 68.8 b) 0.5 5 4 3 66.7 4 7 7 3 42.9 8 10 5 4 53.3 total 12 8 6 60.0 Germinated seeds X 2 n P Mortality rate, % l 1 n P 0.5 km Betula pubescens 0.43 3/3 0.513 0.00 3/3 1.000 Calluna vulgaris 11 1.34 3/3 0.246 2.40 2/2 0.121 All species 0.43 3/3 0.513 1.19 3/3 0.275 4 km Betula pubescens 0.89 3/3 0.346 - Calluna vulgaris 0.04 3/3 0.827 3.00 2/3 0.083 All species 0.04 3/3 0.827 2.33 3/3 0.127 8 km Betula pubescens 0.89 3/3 0.346 - Calluna vulgaris 2.33 3/3 0.127 1.19 3/3 0.275 All species 1.76 3/3 0.184 0.43 3/3 0.513 258 Salemaa and Uotila Basic Appl. Ecol. 2, 3 (2001) Fig. 2. The survival functions of the seedlings of Calluna vulgaris grown a) in the substrate collected from the untreated stands at different distances (km) from the smelter, and b) - d) in the untreated and fertilised substrate representing varying heavy metal and nutrient concentrations (Tab. 2) at different dis tances from the smelter. The functions have been compared using the Wilcoxon (Gehan) tests. The same letter indicates non-significant differences between the functions in the panel a), otherwise the pairwise differences are significant at the P < 0.03 level. (Tab. 2), and had no significant effect on the mortality rates of the seedlings (Tab. 6). When explaining the mortality rate of Calluna vul garis in the combined data of the untreated and fer tilised plots (distances 0.5, 4 and 8 km, n = 14) on the basis of the 1996 soil data (Tab. 2), Cu (P < 0.001) and Ca (P < 0.04) were significant explanatory factors, but the other elements (Ni, Mg, N, P) were non-signifi cant. The mortality rate of the seedlings increased with increasing Cu concentrations, but decreased with in creasing Ca concentrations: y = mortality rate (%) of the Calluna vulgaris seedlings, x, = Cu concentration in the organic layer, x 2 = Ca concentration in the organic layer. Life tables of the Calluna vulgaris seedlings The first seedlings of Calluna vulgaris emerged dur ing the second week in the greenhouse trial, and the germination rate was highest by the fourth week. A few seedlings emerged at the end of the experiment in the 21th week. The survival probability of the seedlings was the higher, the further away from the smelter the soil was collected (Fig. 2). The death risk increased the most in young, under five-week-old seedlings and was relatively stable after this period. The survival probability was the lowest in the soil from distances of 0.5-2 km, and varied between 0-30% at the end of the trial. The survival probabili ty was over 60% at distances of 3-4 km, and 80% at 8 km. The survival distributions differed significantly between the three distances of 0.5-2 km, 3-4 km and 8 km (Fig. 2a). Although nutrient addition and liming increased the survival probability of the seedling slightly at 0.5 km and 4 km, the effect was not statistically significant (Fig. 2b-c). The survival probability was lower in the soil samples from plots given nitrogen than in the un treated soil samples at 8 km (Fig. 2d). However, the overall nutrient level was higher in the untreated than in the fertilised plots at this distance (Tab. 2). y = 33.032 + o.olBx, - 0.004x 2 , r 2 = 81%, P < 0.001, df = 13. Seed bank composition and seedling survival in forest soil 259 Basic Appl. Ecol. 2, 3 (2001) Discussion The size and composition of the seed banks The importance of soil seed banks in regenerating for est vegetation varies according to plant species and the disturbance history of the stands (Archibold 1989, Mladenoff 1990, McGee & Feller 1993). The persis tent seed banks in boreal coniferous forests are consid ered to be rather small (Archibold 1989) compared to those in e.g. temperate deciduous forests (Pickett & McDonnell 1989, Smaliidge & Leopold 1995). The densities of germinable seeds found in this study (15-1200 per m 2 excluding seedlings of Betula pubescens) were lower than the range of 900-3700 per m 2 presented by Huopalainen et al. (2000) for undrained pine bogs near the Harjavalta smelter, or the range of 90-5000 per m 2 in Calluna type pine forests in Sweden (Granström 1986). Some of the seedlings may have died before cotyledons emerged on the soil surface in our experiment. In any case, the seedling emergence method used here gives an under estimate of the actual size of the seed banks (Brown 1992). In general, the seed bank density of the present study well reflected the abundance of the existing veg etation, the pollution level in the organic soil layer and the distance from the smelter. The fertilisation and lim ing treatments did not appear to have any effect on the number of germinated seedlings. Compared to the seed bank data collected from Scots pine forests of heavily polluted areas in the Kola Peninsula, NW Rus sia (Komulainen et al. 1994), more viable seeds were found in our samples. Seedling density varied between 430 and 1200 per m 2 in the soil taken at a distance of 8 km from the Harjavalta smelter, while it was ap proximately 280 per m 2 at the corresponding distance from the Monchegorsk smelter (Komulainen et al. 1994). The number of dormant or dead seeds was not known in our study, and it is therefore impossible to determine whether heavy metals had any detrimental effects on the viability of the seeds stored in the soil. In the data from the Kola Peninsula, for instance, only 3% of the total seed pool was germinable at a distance of 15 km from the smelter (Komulainen et al. 1994). Many studies have shown that the seed germination capacity of tree and dwarf shrub species is not reduced by acidity or metals (Patterson & Olson 1983, Percy 1986, Scherbatskoy et al. 1987, Legg et al. 1992), even though Helsper & Klerken (1984) reported that the germination of Dutch Calluna vulgaris seeds decreased significantly when the pH was lower than 3.2. In our study, the pH values of the untreated plots were close to this critical value, ranging from 3.1 to 3.6 in the or ganic layer (Tab. 2). Vieno et al. (1993) observed that six years of acidic irrigation did not affect the viability of forest seed banks at a site in Finnish Lapland. A lowered seed input owing to the paucity of vege tation is a more probable reason for the reduced seed bank close to the Harjavalta smelter than soil toxicity. The total coverage of the understorey vegetation was under 1% (Tab. 1). Calluna vulgaris, which was still growing close to the smelter in the 1970's (Laaksovirta & Silvola 1975), had completely disappeared. Experi mental exposures have shown that Calluna vulgaris is more sensitive to Cu than Empetrum nigrum (Monni et al. 2000 a, b), the latter species growing in small patches in severely heavy metal contaminated sites. Seed production may also be hindered by distur bances in different stages of sexual reproduction. For instance, the clones of wind-pollinated dioecious Em petrum nigrum produced a large number of flower buds, but practically no berries in the most polluted study area in 1994 (unpublished results). Whether this was caused by poor pollen quality (Wolters & Martens 1987) or insufficient pollen for successful dis persion from male to female clones, is unknown. Ac cording to Bell & Tallis (1973), the seeds of this species need at least four years' dormancy in the soil before germination. The frequency of insect pollinators may also be lower in heavily polluted ecosystems, thus affecting the seed production of insect-pollinated species. How ever, this did not seem to be the case with Vaccinium uliginosum, the berry production of which was en hanced near to the smelter, probably as a response to the increased illumination in the damaged stand (Sale maa et al. 1999). Clonal dwarf shrubs, which are the most dominant group of vascular plants in boreal forest understorey, normally regenerate vegetatively in mature stands. For instance, the seedling recruitment of Vaccinium species is possible only after a forest fire or in disturbed mi crosites or gaps, which have a suitable substrate with high moisture and organic matter contents (Erikson & Fröborg 1996). Although Vaccinium species invest highly in seed production, their seed banks are consid ered to be small and short-lived (Granström 1982, Vander Kloet & Hill 1994). However, Vaccinium uligi nosum and V. vitis-idaea were well represented in the seed banks of our study, reflecting the composition of the existing vegetation. In contrast, V. myrtillus was absent from the seed bank, but it was also rare in the aboveground vegetation. The absence of Arctostaphylos uva-ursi from our samples, despite berry-producing clones at distances of 2-3 km from the smelter, may be due to deep dor mancy of the seeds. The berries of this species are also relatively heavy and probably fall close to the mother clones. The number of soil samples in our study 260 Salemaa and Uotila Basic Appl. Ecol. 2, 3 (2001) might be too low to reflect the patchy spatial pattern of the Arctostapbylos seeds in the soil. Also birds can eat berries and spread them over relatively long dis tances. Almost all the soil samples from the study transect had viable seeds of Calluna vulgaris. Calluna seed banks have been widely studied in heathlands (e.g. Hester et al. 1991, Legg et al. 1992, Pywell et al. 1997, Mahy et al. 1999, Valbuena et al. 2000) and forest plantations (Hill 8i Stevens 1981, Granström 1988). Compared to the other dwarf shrub species growing in boreal forests, it has large and persistent seed banks that enable it to colonise exposed soil soon after clear cutting or fire (Granström 1988). Calluna seeds have been shown to remain viable for 30-40 years under a heathland canopy (Gimingham 1972), and at least 90 years in forest soil (Granström 1988). Thus, it is possi ble that the seeds originating from the most polluted area date back decades in the past. However, our sam ples represented only the surface layers of the soil, from which the oldest part of the Calluna seed popula tion has probably drifted down to the deeper soil lay ers (cf. Bekker et al. 1998). Both Calluna vulgaris and Betula pubescens belong to the early successional species, which produce large numbers of seed with good dispersion ability. The seed crop of the Betula species was exceptional high in the year before the sampling, 1993. Because the majority of Betula seeds lose their viability within one year (Granström 8c Fries 1985), it is probable that most of the Betula seedling in our samples were derived from the previous-year seed crop. The huge number of seedlings in the samples taken 1 km from the smelter especially can be explained by the presence of a birch stand growing in the vicinity of the study plots. Seedling establishment Seedling survival was strongly affected by the heavy metal concentrations in the soil. Most of the germinat ed seedlings originating from a distance of 0.5-2 km from the smelter died at an age of a few weeks, proba bly after the depletion of the seed's nutrient reserves. Cu and Ni are known to reduce especially the root growth of many tree (Patterson & Olson 1983, Kahle 1993) and dwarf shrub species (Monni et al. 2000b). Seedlings that managed to grow their roots into the clean substrate in our experiment could have escaped the toxicity of the polluted soil. The mycorrhizas of er icaceous plants, in addition to facilitating the uptake of nutrients and water, also provide resistance to heavy metals (Bradley et al. 1981, Leake et al. 1990). One reason for the high mortality of the Calluna vul garis seedlings growing in the most polluted soil might be the absence of mycorrhizal infections. The actual germination and growth conditions close to the Harjavalta smelter are much more severe than those in the greenhouse. Drought and the thick layer of undecomposed needle litter hinder seed germina tion. If a seed successfully germinates, the roots of the young seedling still have to grow through the heavy metal enriched surface layer down into the less-pollut ed, moister soil horizons. In addition to the toxic ef fects of heavy metals, a deficiency of macronutrients such as Ca and Mg (Derome & Lindroos 1998) and the decreased water-holding capacity of the soil (Derome & Nieminen 1998) increase the stress en countered by the plant individuals. The seedlings of Calluna vulgaris, particularly, are sensitive to summer desiccation and winter browning (Legg et al. 1992). In general, protection from the wind and withering seem to increase the survival rate of seedlings growing in polluted soil (Kozlov & Haukioja 1999). The fertiliser treatments that included limestone de creased the Cu and Ni concentrations in the organic soil layer (Mälkönen et al. 1999, Derome 2000). The Ca and Mg concentrations and pH correspondingly increased, as well as the microbial activity of the soil (Fritze et al. 1996). These changes presumably con tributed to the increase in seedling survival in our sam ples. Although Calluna vulgaris is, according to its habitat preference, a calcifuge species (Gimingham 1960), its young seedlings seemed to benefit from lim ing when growing in heavy metal polluted soil. Legg et al. (1992) found that liming had no negative effect on the survival of Calluna seedlings in a transplant exper iment, but liming decreased the shoot mass by 13%. The nitrogen application without limestone had no significant effects on the nutrient status or pH of the organic layer compared to the untreated plots at a dis tance of 8 km. In contrast, the N, P, K, Ca and Mg concentrations were slightly higher in the untreated than in the fertilised plots (Tab. 2). The better overall nutrient status of the substrate seemed to explain the better survival probability of the Calluna seedlings growing in the untreated soil (Fig. 2d). The nitrogen input, although no longer reflected in the nutrient con centrations of the organic layer in 1996 (Tab. 2), might also have had a detrimental effect on the sur vival of the Calluna seedlings in 1994. For instance, Helsper & Klerken (1984) found that repeated appli cations of nitrogen inhibited the growth of Calluna, while Ca had little apparent effect. Revegetation efforts in heavy metal polluted indus trial areas have produced promising results when com bined with soil amelioration e.g. in Sudbury, Canada (Winterhalder 2000). The reduction in emissions has opened up new possibilities for the recovery of forest ecosystems close to the Harjavalta smelter. Although the understorey vegetation was almost totally absent Seed bank composition and seedling survival in forest soil 261 Basic Appl. Ecol. 2, 3 (2001) at 0.5 km from the smelter, viable seeds of native plant species were stored in the soil. Seed banks maintain ge netic diversity of plant populations (Mahy et al. 1999) which may enable evolution of tolerant ecotypes. However, seedling establishment fails as a result of the phytotoxicity of heavy metals, a thick layer of unde composed needle litter, drought and nutrient deficien cies. Our results stress that the realisation of the reveg etation potential of the local seed banks presupposes soil amelioration directed at immobilising the heavy metals and facilitating the generation of a new organic soil layer. Acknowledgements. We extend our sincerest thanks to Marja Vieno and Arja Ojala for advising on the sampling and germination procedures, to Mikko Piirainen for identi fying some difficult species, and to Heljä-Sisko Helmisaari and Satu Monni for commenting on the earlier version of the manuscript. John Derome kindly checked the English lan guage and provided the soil data. References Aamlid D, Vassilieva N, Aarrestad PA, Gytarsky ML, Lind mo S, Karaban R, Korotkov V, Rindal T, Kuzmicheva V, Venn K (2000) The ecological state of the ecosystems in the border areas between Norway and Russia. Boreal En vironment Research 5: 257-278. Archibold OW (1979) Buried viable propagules as a factor in postfire regeneration in northern Saskatchewan. Cana dian Journal of Botany 57: 54-58. Archibold OW (1989) Seed banks and vegetation processes in coniferous forests. In: Leek MA, Parker VT, Simpson RL (eds) Ecology of soil seed banks. Academic Press Inc, San Diego, pp 107-122. Bekker RM, Bakker JP, Grandin U, Kalamees R, Milberg P, Poschlod P, Thompson K, Willems JH (1998) Seed size, shape and vertical distribution in the soil: indicators of seed longevity. Functional Ecology 12: 834-842. Bell JN, Tallis JH (1973) Empetrum nigrum L. Journal of Ecology 61: 289-305. Bell R, Teramura AH (1991) Soil metal effects on the germi nation and survival of Quercus alba L., and Q. prinus L. Environmental and Experimental Botany 31: 145-152. Bradley R, Burt AJ, Read DJ (1981) Mycorrhizal infection and resistance of heavy metal toxicity in Calluna vulgaris. Nature 292: 335-337. Brown D (1992) Estimating the composition of a forest seed bank: a comparison of the seed extraction and seedling emergence methods. Canadian Journal of Botany 70: 1603-1612. Derome J (2000) Detoxification and amelioration of heavy metal contaminated forest soils by means of liming and fertilisation. Environmental Pollution 107: 79-88. Derome J, Lindroos AJ (1998) Effects of heavy metal con tamination on macronutrient availability and acidifica- tion parameters in forest soil in the vicinity of Harjavalta Cu-Ni smelter, SW Finland. Environmental Pollution 99: 225-232. Derome J, Nieminen T (1998) Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Environmental Pollution 103: 219-228. Eriksson O, Fröborg H (1996) "Windows of opportunity" for recruitment in long-lived clonal plants: experimental studies of seedling establishment in Vaccinium shrubs. Canadian Journal Botany 74: 1369-1374. Folkeson L, Andersson-Bringmark E (1988) Impoverishment of vegetation in a coniferous forest polluted by copper and zinc. Canadian Journal of Botany 66: 417-428. Freedman B, Hutchinson TC (1980) Long-term effects of smelter pollution at Sudbury, Ontario, on forest commu nity composition. Canadian Journal of Botany 58: 2123-2140. Fritze H, Vanhala P, Pietikäinen J, Mälkönen E (1996) Vital ity fertilisation of Scots pine stands growing along a gradi ent of heavy metal pollution: short-term effects on micro bial biomass and respiration rate of the humus layer. Fre senius Journal of Analytical Chemistry 354: 750-755. Fritze H, Pennanen T, Vanhala P (1997) Impact of fertilisers on the humus layer microbial community of Scots pine stands growing along a gradient of heavy metal pollution. In: Insam H, Rangger A (eds) Microbial communities. Functional versus structural approaches. Springer-Verlag, Berlin, pp 68-83. Gimingham CH (1960) Calluna vulgaris (L.) Hull. Biologi cal Flora of the British Isles. Journal of Ecology 48: 455-483. Gimingham CH (1972) Ecology of heathlands. Chapman and Hall, London. Granström A (1982) Seed banks in five boreal forest stands originating between 1810 and 1863. Canadian Journal of Botany 60: 1815-1821. Granström A (1986) Seed banks in forest soils and their role in vegetation succession after disturbance. Swedish Uni versity of Agricultural Sciences, Department of Forest Site Research, Stencil no. 6. Granström A (1988) Seed banks at six open and afforested heathland sites in southern Sweden. Journal of Applied Ecology 25: 297-306. Granström A, Fries C (1985) Depletion of viable seeds of Be tula pubescens and Betula verrucosa sown onto some north Swedish forest soils. Canadian Journal of Forest Re search 15: 1176-1180. Halpern CB, Evans SA, Nielson S (1999) Soil seed banks in young, closed-canopy forests of the Olympic Peninsula, Washington: potential contributions to understory reiniti ation. Canadian Journal of Botany 77: 922-935. Helmisaari HS, Derome J, Fritze H, Nieminen T, Palmgren K, Salemaa M, Vanha-Majamaa I (1995) Copper in Scots pine forests around a heavy-metal smelter in south-west ern Finland. Water, Air and Soil Pollution 85:1727-1732. Helsper HPG, Klerken GAM (1984) Germination of Callu na vulgaris (L.) Hull in vitro under different pH-condi tions. Acta Botanica Neerlandica 33: 347-353. Hester AJ, Gimingham CH, Miles J (1991) Succession from heather moorland to birch woodland. 111. Seed availabili- 262 Salemaa and Uotila Basic Appl. Ecol. 2, 3 (2001) ty, germination and early growth. Journal of Ecology 79: 329-344. Hill MO, Stevens PA (1981) The density of viable seed in soil of forest plantations in upland Britain. Journal of Ecology 69: 693-709. Huopalainen M, Tuittila ES, Vanha-Majamaa I, Nousiainen H, Laine J, Vasander H (2000) The potential of soil seed banks for revegetation of bogs in SW Finland after long term aerial pollution. Annales Botanici Fennici 37: 1-9. Johnson MS, Bradshaw AD (1979) Ecological principles for restoration of disturbed and degraded land. Applied Biol ogy 4: 141-200. Kahle H (1993) Response of roots of trees to heavy metals Environmental and Experimental Botany 3: 99-119. Kiikkilä O, Perkiömäki J, Barnette M, Derome J, Pennanen T, Tulisalo E, Fritze H (2001) In situ bioremediation through mulching of soil polluted by a Cu-Ni smelter. Journal of Environment Quality (in press). Komulainen M, Vieno M, Yarmishko VT, Daletskaja TD, Maznaja EA (1994) Seedling establishment from seeds and seed banks in forests under long-term pollution stress: a potential for vegetation recovery. Canadian Journal of Botany 72: 143-149. Kozlov M, Haukioja E (1999) Performance of birch seedlings replanted in heavily polluted industrial barrens of the Kola peninsula, Northwest Russia. Restoration Ecology 7: 145-154. Kubin E, Lippo H, Poikolainen J (2000) Heavy metal load ing. In: Mälkönen E (ed) Forest condition in a changing environment - the Finnish case. Kluwer Academic Pub lishers, Dordrecht, pp. 60-71. Laaksovirta K, Silvola J (1975) Effect of air pollution by copper, sulphuric acid and fertilizer factories on plants at Harjavalta, W. Finland. Annales Botanici Fennici 12: 81-88. Leake JR, Shaw G, Read DJ (1990) The role of ericoid my corrhizas in the ecology of ericaceous plant. Agriculture, Ecosystems and Environment 29: 237-250. Legg CJ, Maltby E, Proctor MCF (1992) The ecology of se vere moorland fire on the North York Moors: seed distri bution and seedling establishment of Calluna vulgaris. Journal of Ecology 80: 737-752. Mahy G, Vekemans X, Jacquemart AL (1999) Patterns of al lozymic variation within Calluna vulgaris populations at seed bank and adult stages. Heredity 82: 432-440. Mälkönen E, Derome J, Fritze H, Helmisaari HS, Kukkola M, Kytö M, Saarsalmi A, Salemaa M (1999) Compen satory fertilisation of Scots pine stands polluted by heavy metals. Nutrient Cycling in Agroecosystems 55: 239-268. McCullagh P, Nelder JA (1989) Generalized linear models. 2nd edition. Monographs on statistics and applied proba bility 37. Chapman and Hall, London. McGee A, Feller MC (1993) Seed banks of forested and dis turbed soils in southwestern British Columbia. Canadian Journal of Botany 71: 1574-1583. Mitchell RJ, Marrs RH, Auld MHD (1998) A comparative study of the seedbanks of heathland and successional habitats in Dorset, Southern England. Journal of Ecology 86: 588-596. Mladenoff DJ (1990) The relationship of the soil seed bank and understorey vegetation in old-growth northern hard- wood-hemlock treefall gaps. Canadian Journal of Botany 68:2714-2721. Monni S, Salemaa M, White C, Tuittila E, Huopalainen M (2000 a) Copper resistance of Calluna vulgaris originating from the pollution gradient of a Cu-Ni smelter, in south western Finland. Environmental Pollution 109: 211-219. Monni S, Salemaa M, Millar N (2000b) The tolerance of Empetrum nigrum to copper and nickel. Environmental Pollution 109: 221-229. Moore JM, Wein RW (1977) Viable seed populations by soil depth and potential site recolonization after disturbance. Canadian Journal of Botany 55: 2408-2412. Onaindia M, Amezaga I (2000) Seasonal variation in the seed banks of native woodland and coniferous plantations in Northern Spain. Forest Ecology and Management 126: 163-172. Patterson W, Olson J (1983) Effects of heavy metals on radi cle growth of selected woody species germinated on filter paper, mineral and organic soil substrates. Canadian Jour nal of Forest Research 13: 233-238. Percy K (1986) The effect of simulated acid rain on germina tive capacity, growth and morphology of forest tree seedlings. New Phytologist 104: 473-484. Pickett STA, McDonnell MJ (1989) Seed bank dynamics in temperate deciduous forest. In: Leek MA, Parker VT, Simpson RL (eds) Ecology of soil seed banks. Academic Press Inc., San Diego, pp 123-147. Pywell RF, Putwain PD, Webb NR (1997) The decline of heathland seed populations following the conversion to agriculture. Journal of Applied Ecology 34: 949-960. Rigina O, Kozlov MV (2000) Pollution impact on sub-arctic forests in the Kola Peninsula, Russia. In: Innes JL, Oleksyn J (eds) Forest dynamics in heavily polluted re gions. CAB International, Wallingford, pp 37-65. Salemaa M, Vanha-Majamaa I, Gardner PJ (1999) Compen satory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal pol luted environment. Plant Ecology 141: 79-91. Salemaa M, Vanha-Majamaa I, Derome J (2001) Under storey vegetation along a heavy-metal pollution gradient in SW Finland. Environmental Pollution 112: 339-350. SAS Institute Inc. (1994) SAS/STAT User's Guide Version 6. 4 th Edition. SAS Institute Inc., Cary, North Carolina, USA, Scherbatskoy T, Klein RM, Badger GJ (1987) Germination responses of forest tree seed to acidity and metal ions. En vironmental and Experimental Botany 27: 157-164. Smallidge PJ, Leopold DJ (1995) Watershed liming and pit and mound topography effects on seed banks in the Adirondacks, New York, USA. Forest Ecology and Man agement 72: 273-285. SPSS Inc. (1999) SPSS 9.0.1. software. SPSS Inc., Chicago, USA. Sorensen T (1948) A method of establishing groups of equal amplitude in plant sociology based on similarity of species content. Det Kongelige Danske Videnskabernes Selskab, Biologiske Skrifter, Bind V, Nr 4, Copenhagen. Uhlig C, Salemaa M, Vanha-Majamaa I, Derome, J 2001. El ement distribution in Empetrum nigrum microsites at heavy metal contaminated sites in Harjavalta, western Finland. Environmental Pollution 112: 435-442. Seed bank composition and seedling survival in forest soil 263 Basic Appl. Ecol. 2, 3 (2001) Valbuena L, Tarrega R, Luis-Calabuig E (2000) Seed banks of Erica australis and Calluna vulgaris in a heathland sub jected to experimental fire. Journal of Vegetation Science 11: 161-166. Vander Kloet SP, Hill NM (1994) The paradox of berry pro duction in temperate species of Vaccinium. Canadian Journal of Botany 72: 52-58. Vangronsveld J, Colpaert JV, Van Tichelen KK (1996) Recla mation of a bare industrial area contaminated by non-fer rous metals: physico-chemical and biological evaluation of the durability of soil treatment and revegetation. Envi ronmental Pollution 94: 131-140. Vieno M, Komulainen M, Neuvonen S (1993) Seed bank composition in a subarctic pine - birch forest in Finnish Lapland: natural variation and the effect of simulated acid rain. Canadian Journal of Botany 71: 379-384. Winterhalder K (2000) Landscape degradation by smelter emissions near Sudbury, Canada, and subsequent amelio ration and restoration. In: Innes JL, Oleksyn J (eds) Forest dynamics in heavily polluted regions, CABI Publishing, Wallingford, pp 87-119. Wolters JHB, Martens MJM (1987) Effects of air pollutants on pollen. The Botanical Review 53: 372-414. Paper IV Salemaa, M., Vanha-Majamaa, I. & Gardner, P. 1999. Compensatory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal polluted environment. Plant Ecology 141: 79-91. Plant Ecology 141: 79-91,1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands. 79 Compensatory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal polluted environment M. Salemaa 1 ,1. Vanha-Majamaa 1 &P. J. Gardner 2 1 Vantaa Research Centre, The Finnish Forest Research Institute, FIN-01301 Vantaa, Finland (e-mail: maija.salemaa@metla.fi); 2 Environment Agency, Ridgeworth House, Liverpool Gardens, Worthing, BNII IYH, UK Key words: Bearberry, Bog bilberry. Branching architecture. Metal resistance, Overcompensation, Phenotypic plasticity Abstract The effect of artificial shoot clipping on the vegetative growth and sexual reproduction of the evergreen bearberry, Arctostaphylos uva-ursi, and the deciduous bog bilberry, Vaccinium uliginosum, was studied in the vicinity of a copper-nickel smelter in SW Finland. According to the research hypothesis, heavy metal induced shoot death breaks the apical dominance in the clones growing in a polluted environment. This causes activation of dormant axillary and adventitious buds and an increase in branching on the older parts of the stem. Regrowth after shoot death was studied by clipping off all the current-year shoots from experimental branches in autumn (1994) and spring (1995). Within-clone and between-clone control branches were used to test the data. Both species displayed a considerable ability to activate dormant meristems after the damage. Regrowth of the current shoots during the next growing season (1995) was about 80% compared to the within-clone control in both species after autumn clipping. Shoot clipping in early summer was more detrimental for both species, and the regrowth of A. uva-ursi was less than that of V. uliginosum. Differences in the storage reserves and source sink mechanisms of carbon allocation between evergreen and deciduous species probably explain their distinct response. When the removed biomass was added to the living biomass of the branches, there was overcompensation in the total dry weight of A. uva-ursi after autumn clipping, and the weight was almost 90% of the control after spring clipping. The total dry weight of V. uliginosum also equalled that of the control when the removed biomass was added. No berries developed on either species in the year following the autumn treatment, because clipping removed all the flower buds. Spring clipping had no effect on the sexual reproduction of A. uva-ursi, but decreased the berry production of V. uliginosum. The degree of compensatory growth of both species was only slightly affected by the distance from the smelter. It is suggested that dormant bud activation, rapid regrowth and plastic branching contribute to the resistance mechanism to heavy metals. Introduction The ability of plants to compensate for lost biomass is an important means of recovering from different kinds of biotic and abiotic damage (Belsky et al. 1993). Compensatory growth has most often been described following herbivory (Marquis 1996), but it can also be considered as a generalised response of plants to all kinds of damage. Heavy metal induced changes in vegetation are common in the surroundings of metal processing factories, smelters and on metalliferous waste. Although elevated heavy metal levels generally injure plants, resistant populations are known, espe cially among vascular plants (Ernst 1974). Resistance to heavy metals can be achieved through either avoid ance or tolerance strategies (Levitt 1980; Baker 1987; Verkleij & Schat 1989). Avoidance is expressed as an external protection against toxic elements or as ac tive orientation of the roots to a less toxic substrate (Balsberg 1982; Tyler et al. 1989). Tolerance is usu ally understood as a physiological mechanism, which 80 results in the exclusion or accumulation of heavy met als and their conversion into a detoxified form (Baker 1981). Rapid turnover of the roots or above-ground parts of plants may also serve as a resistance strategy, which makes it possible to exclude heavy metals from living tissues (Tyler et al. 1989) or to recover from biomass loss (Belsky et al. 1993). Release of apical dom inance after damage makes regrowth possible from activated dormant buds in metameric plant species (Aarssen 1995; Marquis 1996). In long-lived trees and clonal dwarf shrubs, ecotypic differentiation into metal-tolerant races is rare or very slow (Dickinson et al. 1991). In these species, phenotypic plasticity in morphological and physiological characteristics is a more probable mechanism for surviving in a polluted environment. In this paper we study two clonal dwarf shrubs, the evergreen Arctostaphylos uva-ursi L. (bearberry) and the deciduous Vaccinium uliginosum L. (bog bil berry) in a heavy metal polluted environment. Both of the studied species express high resistance to heavy metals (Laaksovirta & Silvola 1975; Väisänen 1986; Salemaa & Vanha-Majamaa 1993) and have a good re growth ability after damage. High plasticity in growth response is common in the Ericaceae family (Gim mingham 1972; Chester & McGraw 1983; Shevtsova et al. 1995), but different species display different de grees of compensation, depending on the life form and the type of damage (Tolvanen & Laine 1997). It is proposed that deciduous Vaccinium species are more resistant to herbivory and other tissue loss than evergreen Vaccinium species. This is due to their larger carbohydrate and nutrient reserves, more rapid movement of nutrients to the leaves, and weaker archi tectural constraints on shoot growth (Tolvanen et al. 1993, 1995). Also, the time of damage and environ mental conditions strongly affect the regrowth ability of plants (Chapin 1980 a; Maschinski & Whitham 1989; Hjälten et al. 1993). If damage occurs late in the growing season in Vaccinium species, the regrowth response is absent or lower than that after damage tak ing place early in the growing season (Tolvanen & Laine 1997). Vigorous regrowth is suggested to oc cur in resource-rich environments where plants are not subjected to stress or competition (Belsky et al. 1993). The object of this study is to compare the short term ability of V. uliginosum and A. uva-ursi to recover after artificial shoot clipping along a heavy metal gra dient. Special attention is paid to the significance of compensatory growth as a tolerance mechanism to heavy metals. Three hypotheses were tested as follows: - Deciduous V. uliginosum is more resistant to shoot clipping than evergreen A. uva-ursi. - Autumn shoot clipping at the time when growth has already ceased has a less deleterious effect on regrowth than the clipping of new growth in the spring. - Shoot clipping is more deleterious in a highly polluted than in a less polluted environment. Materials and methods Life histories of V. uliginosum and A. uva-ursi Both deciduous V. uliginosum and evergreen A. uva ursi are clonal, widespread circumpolar members of the Ericaceae. Their global range extends from the Arctic Circle down to the northern temperate zone, and their main distribution area is in the boreal vegeta tion zone (Tutin et al. 1972; Packer & Denford 1974; Jacquemart 1996). V. uliginosum grows in southern Finland mainly on peatlands and, further to the north, also on min eral soils. It has horizontal subterranean rhizomes and grows laterally by producing erect 30-100 cm long shoots and may form large, pure stands. It generally grows on nutrient-poor acid soils and, as a helio philious species, avoids substantially shaded habitats (Jaquemart 1996). It is regarded as cold-tolerant, and the limiting factor in its appearance in southern re gions seems to be relatively high temperatures (Stew art & Bannister 1973). V. uliginosum has a sympodial branching habit with the terminal bud producing flow ers and the uppermost lateral buds producing shoots (Figure la). The leaves are annual but the stem and buds overwinter. Fertility varies according to the site, and the berries are produced exclusively after cross pollination (Fröborg 1996). However, vegetative re production appears to be much more important than that via seeds (Eriksson & Fröborg 1996). Prostrate A. uva-ursi grows throughout Finland. It colonises open sandy heathlands and moraine slopes, eroded banks, and a variety of habitats in the bo real forest; however it does not grow on peat bogs, as opposed to the situation in North America (Rem phrey et al. 1983). It has a deep tap and numerous adventitious roots which make it a good coloniser of disturbed sites. The branching habit of A. uva-ursi is predominantly sympodial, but some horizontally 81 Figure 1. Branching habits of the studied dwarf shrubs, (a) Vac cinium uliginosum: (1) sympodial vegetative shoots with annual leaves, (2) sympodial vegetative shoot developed from the basal bud on the old stem (elongated and thick rejuvenation shoot), (3) terminal short-shoots with flowers, and (4) old dead shoots, (b) Arc tostaphylos uva-ursi: (1) monopodially produced leading vegetative shoot with an active terminal bud, (2) sympodial lateral vegetative shoot with an active terminal bud, (3) sympodial lateral generative shoot with a terminal inflorescence, (4) sympodial generative shoot with berries, and (5) sympodial lateral shoots with a dormant termi nal bud. Stages 1-3 represent current-year, and 4-5 previous-year shoots. elongated shoots are produced monopodially from the terminal bud of the parent shoot (Figure lb). The occurrence of neoformed growth (produced by the apex of the current-year shoots) is common and sug gests an opportunistic growth strategy (Remphrey & Steeves 1984 a). Terminal inflorescences overwinter and berries develop in the next year. The shoots do not grow during the year that they produce berries. A. uva-ursi regenerates rather poorly from seeds and spreads mainly vegetatively (Remphrey et al. 1983). Activation of lateral buds on creeping branches (Rem phrey et al. 1983; Remphrey & Steeves 1984b; Bowles 1983), or latent buds on roots (Tiffney et al. 1978), produces new shoots and aids re-establishment after disturbance. Study sites The study area is situated near the Cu-Ni smelter at Harjavalta (61°19'N, 22°9'E), SW Finland, in the southern boreal vegetation zone. A. uva-ursi was stud ied in two heath forest (Pinus sylvestris L.) stands at distances of 2 (F2) and 8 km (F8) from the smelter. V. uliginosum was studied in two heath forest stands at 0.5 km (F0.5) and 4 km (F4) and also in a drained peat land stand at 5 km (P5). The two species did not occur in sufficient numbers at all the distances. The 0.5-2 km zone represents highly polluted sites and the 4-8 km zone less polluted ones. General characteristics of the stands and data on deposition and element concentra tions in the organic layer and vegetation are presented in Table 1. The mean annual temperature was 5°C in 1995 and the annual precipitation 698 mm at the nearby weather station of the Finnish Meteorological Institute. Copper production at the smelter started in 1945 and that of nickel in 1960. Emissions from the smelter have considerably decreased since the end of the 1980s (Cu 110, Ni 53 and SO2 8000 t/yr) to the middle of 1990s (1994: Cu 40, Ni 6 and S02 5000 t/yr), when the study was started. However, accumulation of heavy metals in the soil during the last 50 years has changed the element fluxes of the ecosystem and dam aged the vegetation (Helmisaari et ai. 1995; Salemaa & Vanha-Majamaa 1993). Experimental design and measured variables A total of 30 clones were randomly selected at all sites. The distance between each clone was at least 5 m, and only visually separate patches were sampled. Unpublished data on isoenzyme variation revealed that almost all the studied clones represented different au totetraploid genotypes. The age of a few typical clones of both species growing at km distance from the smelter was determined by counting the number of an nual rings on the oldest part of the stem. The age of the clones ranged from 30 to 40 years, and it is probable that some 'mother clones' date back to the time when the smelter was founded in the 19405. The clones were divided into three groups, ten replicates in each: undipped controls, clones clipped in autumn (14—16 September 1994) and clones clipped in spring, soon after bud break (V. uliginosum 6-7 June, A. uva-ursi 14 June 1995). Clipped clones are later called experimental clones. Shoot clipping was restricted to three randomly selected main branches on each experimental clone. All the current-year shoots were removed and stored for further measurements. In addition, three branches were randomly selected for a within-clone control and three for an undipped con trol. This enables compensatory growth of the clipped branches to be compared with the undipped branches of the same clone and with the undipped control 82 Table 1. Site characteristics, element concentrations in the organic layer and in bulk precipitation, and Cu and Ni concentrations in A. uva-ursi (A. u-u.) and V. uliginosum (V. u.) tissues. Exchangeable Cu, Ni and Fe concentrations in the organic layer were determined with NH4acetate + EDTA. All organic layer concentrations have been calculated per dry mass of organic matter. Data from Finnish Forest Research Institute: peatland (Veijalainen, unpublished), stand data on heath forest (Kukkola, unpublished), concen trations in forest humus and deposition (Derome, unpublished), concentrations in plant tissues (Salemaa & Vanha-Majamaa, unpublished). clones. All the branches were harvested for biomass measurements on 25-27 July 1995. A section was removed from the lowest part of the stem of each branch for age determination by den drochronology in the laboratory. The length of the branch was measured and all living current shoots, dead shoots and berries were removed. The numbers of shoots and berries were counted and the dry weight of each separated biomass category and the remain ing branch was determined. The number of clipped shoots was counted and their dry weight measured. Six branch-specific response variables to shoot clip ping were recorded: (1) the total living biomass of the branch including the removed biomass, (2) the total biomass of the current shoots in 1995, (3) the number of current shoots, (4) the average weight of the current shoots, (5) the total biomass of dead shoots, and (6) the total biomass of berries. In bush-like V. uliginosum the studied branches represented the oldest branches and were directly con nected to the common stem or rhizome. In A. uva ursi the studied branches were only distal parts of Site Forest (Scots pine) F0.5 F2 F4 F8 Peatland (Scots pine) P5 Distance from smelter, km 0.5 2 4 8 5 Direction from smelter S SE SE SE NW Forest site type Calluna Calluna Calluna Calluna Dwarf shrub pine bog Stand age, yrs 49 52 56 90 30 Basal area, m 2 /ha 9.0 14.8 15.9 - 9.8 Thickness of humus, cm 2.5 2.3 2.2 - 3.0 Vegetation coverage, % 1 24 77 100 100 Element concentrations in organic layer(1991): N, tot % 1.92 1.74 1.65 1.68 1.49 S, tot mg/kg 387 198 170 176 219 Exc. Cu, mg/kg 7540 2238 786 209 376 Exc. Ni, mg/kg 528 329 164 72 122 Exc. Fe, mg/kg 10899 3800 1902 1659 1240 Deposition in bulk precipitation (1994): N, tot mg/rri 2 449 447 490 396 - S04-S, mg/m" 628 427 399 311 - Cu, mg/m 119 30 9 3 - Ni, mg/irr 26 5 2 0.8 - Fe, mg/m 2 38 17 10 7 - Element concentrations in current leaves (1994): V.u. A.u-u. V.u. A.u-u. V.u. Cu, mg/kg 42.4 13.5 22.3 4.4 14.0 Ni, mg/kg 40.5 10.7 19.3 3.5 12.8 in fine roots (1994): Cu, mg/kg 718.7 313.0 102.3 39.7 102.8 Ni, mg/kg 163.3 119.4 49.3 29.7 48.4 83 the creeping branches which grew from the com mon rhizome. Thus the size and age of the branches represented the real population in V. uliginosum but the sampled branches of A. uva-ursi were arbitrarily clipped to a length of about 24 cm. The definition of compensation is according to Belsky (1986). Overcompensation occurs when the cumulative total dry weight of clipped plants (includ ing removed tissue) is higher, exact compensation when it is equal to, and undercompensation when it is less than the total dry weight of the control plants. Data analysis Branchwise data were used in analysing the responses to shoot clipping, and in comparing the between-site differences in the undipped control clones. The effect of site and treatment and their interaction was tested by two-factor ANOVA (SAS Institute Inc. 1994). The response to clipping depended on branch size, which varied between the sites (Table 2). Therefore covari ates which eliminated the effect of branch size were added to the variance model. The relationship between the response variables and branch size of undipped control clones was studied by means of regression analysis. Depending on the variable or the species, either branch weight or length gave the best coefficient of determination. Both branch weight and length were therefore selected as covariates in the model. The model for the response is: where /z = the overall grand mean, a; = site effect, i = 1,2 (A.u-u.) or i = 1, 2, 3 (V.u.), fij = treatment effect, j = 0,..., 4, yij = site by treatment interac tion, a = the coefficient of the covariate x (weight), b = the coefficient of the covariate z (length), ey = random error. Comparison between the treatment means within the sites was carried out using pairwise contrasts by ANOVA (F-values). No statistics for the clipping re sponse of berries is presented because of the negative values given by the model. Undipped controls were compared between the sites by 1-way ANOVA and pairwise contrasts. The degree of compensation was studied by calcu lating the difference between the means of the control and of the clipped branches within each experimen tal clone. Between-site comparison of the degree of compensation was made by 1-way ANOVA and pair wise contrast analysis, the differences in within-clone branch weight and length being used as covariates. Results Undipped controls at different sites compared to within-done controls Despite the similar lengths, the branches of A. uva-ursi were heavier and one year older at site F2 than at site F8 (Table 2a). The average length of the branches of V. uliginosum was highest at site P5, followed by F0.5 with F4 having the shortest branches (pairwise con trast, p < 0.01). The weight and age of the branches were highest at site F0.5 (p < 0.05) (Table 2b). The means of the response variables predicted by the model of the undipped and within-clone controls are presented in Figures 2 and 3, and the original data in Table 2. The total biomass of current-year shoots of A. uva-ursi was similar at both sites (pairwise con trasts, p = 0.752), but the average shoot weight was significantly higher (p = 0.004) at site F8 than at site F2 (Figures 2b and 2d). In contrast, the total weight of berries was higher at site F2 than at site F8 (p = 0.002, Table 2a). The number of current shoots or the biomass of dead shoots did not differ between the sites (p = 0.965 and p = 0.574 respectively) (Figure 2c,e). The total biomass of current-year shoots of V. uliginosum was lower (p < 0.05) at site F4 than at the other sites (Figure 3b). The number of current year shoots was similar at all sites (Figure 3c). The average weight of the current shoots, the biomass of dead shoots and that of berries were highest at site F0.5 (p < 0.09, Figures 3d, e, Table 2b). Although the model predicted number of current shoots did not differ between the sites in either species, there were differences when the shoot number was cal culated per cm of stem length. In A. uva-ursi the shoot number was slightly higher at site F2 (0.3 shoots/cm) than at F8 (0.2 shoots/cm) (p = 0.122). In V. ulig inosum the number was highest (0.5 shoots/cm) at site F0.5, followed by site P5 (0.4 shoots/cm) and F4 (0.3 shoots/cm) (p < 0.01) (Table 2a, b). In general, the growth of the undipped controls was slightly better or similar to that of the within clone controls in V. uliginosum, but the differences were statistically significant only in a few cases (Fig ure 3b, c). The autumn clipping of A. uva-ursi at site F2 was exceptional, because both the control and ytj =M+a; + Pj + Yi) + axij + bZij + eij> 84 Table 2. Means ± standard errors of five treatments for branch size, removed biomass and response variables of a) A. uva-ursi and b) V. uliginosum. n = number of branches. Total biomass = branch weight + removed biomass + current shoots. Average shoot weight = current shoots / number of shoots. Shoot weight for removed biomass of V. uliginosum includes only stems in treatment 2. Treatment codes as in Figure 2. Berries, 00 £ 204±46 177±44 O 207±57 260±57 21±6 12±6 O 27±13 25±11 Dead branches, g o Ö -H m O O 0.05±0.02 0.19±0.07 0.05±0.02 0.09±0.04 0.01±0.01 0.04±0.01 0.09±0.03 0.05 ±0.02 0.02±0.01 Av. shoot weight, mg 81±7 77±8 6 1 ±8 62±5 27±3 115±10 97±10 37±3 86±12 36±5 •2 O 00 »O r-; NO (N «O (N NO x> E 3 z c c J= <4-1 o 5 00 5 CO 5 »n 5 ON 5 NO ON O -H co »n 5 ON NO 5 00 ON 5 r~; ö -H r-; >n Response variables Total Current biomass, g shoots, g 2.90±0.22 0.64±0.10 CM -H r- ro m -H o r- HH •n •n -H r- «n On -H n sO -H NO -H «n r- ■f -H ö nO nO -H o r- nO -H sC 00 ro -H (N ro 1 E 3 Z ! o m (N -H tN (N o tn -H q 00 tN ro 5 nO NO V = i c* "c3 (2 00 £ 1 o IS o m O -H «n iri ro «n ö -H o> (N K ö -H r-» •n 00 m r- ö Tl O m r-' o o ö -H ro O) 00 r- o -H s (N m (N O -H s (N ro ro d •i o CO tN (N ro d -H 00 fN O ro O -H 00 (N c E u oi 8 J= C/O oo -H op '53 £ S ö - On «n o 00 c ö - nO r- © m o o -H nO O (N O d -H o o in o o -H c rf O O -H ON d .S? £ 00 O -H oo ro «n ö -H 't «n nO Ö -H ro NO NO ö -H On in 00 ö HH O -H (n «S O -H ■*t ro O - nO fO d -H oo rs d -H 00 NO d -H ro ro O -H ro ro in o -H (N n d -H 00 ro nO d -H »n »n £ M C u J E £ 5 m (N «n 5 r- n r- 5 00 r- tN -H >n •t »n ON tN -H (N «n n ON 5 ON 't ON 5 o «n N x: o u 00 < to Ö H -O K ö -H m o -o ö -H (N O ■*t ö o On T}- ö -H On oo en O HH NO m d -H n to d HH 00 in ro O -H n >n ro O -H NO ro O -H NO o -fl ro o ti o 00 ro O H On nO ro O -H nO 5 s "5 g £ e O tn 00 (N O ro O ro O ro O to ON (N On (N o ro O ro O ro (N tN O ro o ro •2? "S J •S c5 o H c o £ o - tN ro ■*t O - (N ro O - tN ro p S" c/5 C o u. 1 •n o c £ E Tt cS p CL "O c ?3 E m 86 Figure 2. Model predicted means ± standard errors of five treatments of A. uva-ursi at two sites (F2 and F8) on a) total biomass of the branch including the removed biomass, b) total biomass of current shoots, c) total number of current shoots, d) average weight of current shoots and e) total biomass of dead shoots. Treatments: 0 = undipped control, 1 = within-clone control for autumn clipping, 2 = autumn clipping (shaded), 3 = within-clone control for spring clipping and 4 = spring clipping (shaded). Pairwise contrast analyses are presented with the letters inside the bars, the same letter indicating non-significant differences (p > 0.05) between the means. Three between-site comparisons are presented in the boxes (contrast analysis, p < 0.150 given): (1) undipped controls, (2) the degree of compensatory growth (within-clone difference between the means of the control and tratment branches) in autumn and (3) in spring, ns = non-significant (p > 0.05). treatment branches grew better than undipped control (p < 0.05). No differences in growth between the un dipped and within-clone controls of A. uva-ursi were found at site F8 (Figure 2b, c). Vegetative growth after shoot clipping compared to within-clone control Both species displayed a good ability to compensate for lost biomass during the growing season follow ing the autumn treatment. Overcompensation occurred in the total biomass of the branches of A. uva-ursi when removed biomass was added to the living bio mass of the branches. The total biomass of the clipped branches, including removed biomass, was about 30% higher than that of the control (pairwise contrast analy sis, p < 0.001) at both sites (Figure 2a). In V. ulig inosum the total biomass of the branches, including removed biomass, was similar to that of the control at all sites after autumn clipping (Figure 3a). The total 87 Figure 3. Model predicted means ± standard errors of different treatments of V uliginosum at three sites (F0.5, F4 and P5). Explanations as in Figure 2. biomass of the branches equalled or was very close to that of the control after the spring clipping for both species (Figures 2a, 3a). The total biomass of current-year shoots of A. uva ursi was 21% lower at site F2 (p = 0.061) and equalled to that of the control at site F8 (p = 0.912) after the autumn treatment (Figure 2b). The corre sponding biomass of current-year shoots of V. uligi nosum was about 20% lower than that of the controls at all sites (significant difference only at site F0.5, p = 0.039) (Figure 3b). Shoot clipping in the spring had more detrimental effects on the regrowth ability. In A. uva-ursi the total biomass of the current-year shoots was only 12% of that of the control at site F2 (p = 0.001) and 25% at site F8 (p = 0.001) (Figure 2b). Regrowth of V. uliginosum was slightly better than that of A. uva-ursi after the spring treat ment, being 20% compared to the control at site F0.5 (p = 0.001), 68% at site F4 (p = 0.054) and 47% at site P5 (p = 0.001) (Figure 3b). The number of current-year shoots on the clipped branches increased in A. uva-ursi (F2: p = 0.114, F4: p = 0.003) but decreased in V. uliginosum (F0.5 and P5: p = 0.001, F4: p < 0.05) after autumn clip ping (Figures 2c and 3c). The shoot number of the clipped branches did not approach the control after 88 spring clipping in A. uva-ursi (p < 0.05). Compensa tion in the shoot numbers of V. uliginosum was almost exactly the same after spring clipping at sites F0.5 and F4. and slight overcompensation occurred at site P5 (p = 0.002). The average weight of the current-year shoots of A. uva-ursi was lower in the clipped branches than in the control after both autumn and spring clipping (p < 0.03) (Figure 2d). Accordingly, when the number of shoots increased, the average weight of the shoots de creased after autumn clipping. Both the shoot number and the average shoot weight decreased (p < 0.02) after spring clipping in A. uva-ursi. The opposite re lationship was found in V. uliginosum after autumn clipping. The number of shoots was lower but the av erage weight of the shoots higher than those in the control (significant difference only at F0.5: p < 0.03), (Figure 3c, d). After the spring clipping the average shoot weight in V. uliginosum was also lower than that in the control (all sites: p < 0.004). The total biomass of dead shoots increased at site F2 after autumn clipping in A. uva-ursi (p = 0.002) (Figure 2d). There were no statistically significant dif ferences in the biomass of dead branches between the treatment and the control in V. uliginosum within the sites (Figure 3d). Sexual reproduction after shoot clipping All the flower buds were lost when current-year shoots were clipped in the autumn in both species. Conse quently, no berries developed after the autumn treat ment (Tables 2a, b). Spring clipping did not have any effect on the berry production of A. uva-ursi, and only a few flowers of V. uliginosum developed into berries following spring clipping (Table 2a, b). Effect of the distance from the smelter on the degree of compensation In general, the distance from the smelter did not sig nificantly affect the degree of compensatory growth of either species. Although the regrowth level of the current shoots of A. uva-ursi was slightly lower at site F2 than at site FB, this difference was not statistically significant (p = 0.335) (Figure 2b). Only the aver age weight of shoots decreased more at site F8 than at site F2 (p = 0.056) after autumn clipping (Fig ure 2d). In V. uliginosum the degree of regrowth of the current-year shoots was slightly higher at site F4 than at the two other sites (p < 0.04) after spring clipping (Figure 3b). In addition, the decrease in the number of current-year shoots was highest at site F0.5 (p < 0.100) after autumn clipping (Figure 3c). Discussion The effect of life form and the time of damage on the response Mechanical cutting of shoots may not have the same effect on plants as heavy metal induced shoot mor tality. Heavy metal stress damages all the primary metabolic processes of plants (Balsberg-Pählsson 1989), but the actual damage mechanism is unknown at the cellular level in the studied species. V. ulig inosum had a high abundance of dead branches in the most polluted environment. Both species also had a higher number of current shoots per cm of stem than clones growing further away from the smelter. This suggests that the breakage of apical dominance is somehow connected to shoot death caused by heavy metal toxicity. Many of the current shoots of V. ulig inosum grew from old wooden stems and were elon gated and thicker than the shoots in the apical parts of the branches (rejuvenation shoots, Figure la). Be sides direct toxic effects on the living tissues of plants, heavy metals may also accentuate natural stress factors such as frost (Sutinen et ai. 1996) or desiccation in the spring when the snow cover no longer protects the understorey vegetation. Autumn clipping did not appear to damage either A. uva-ursi or V! uliginosum seriously. Although only the short-term response was studied, the current-year growth of the clipped branches was about 80% of that of the within-clone control. It is probable that the resource loss caused by autumn clipping was rather small at the time when the growing season was already over. Thus clipping merely activated the dormant meristems and resulted in vigorous lateral branching the following summer. Contrary to the predictions, the relative number of activated meristems was higher in A. uva-ursi than in V. uliginosum. This kind of flexi ble branching seems to be very typical of A. uva-ursi after damage (Remphrey et al. 1983, Bowles 1983). However, the slight overcompensation observed in the total dry weight of A. uva-ursi was not exceptional, but probably represented only the upper range of nor mal growth (cf., Belsky et al. 1993). Aarssen (1995) suggests that overcompensation in terms of increased fitness can be expected when species that usually ben efit from apical dominance are found in conditions 89 where apical dominance affords little or no selective advantage. In the case of A. uva-ursi, it seems that the strength of apical dominance is rather plastic and that the 'reserve of dormant meristems' (Tuomi et ai. 1994; Aarssen 1995) enables the species to re cover after damage. While autumn-clipped branches of V. uliginosum produced a few heavy shoots, A. uva ursi produced many lighter ones. Chester & McGraw (1983) have also observed the tendency of V. uligi nosum to increase shoot size as a response to nitrogen fertilization, in contrast to V. vitis-idaea L. in which many lateral meristems are activated after fertilization. The weight of dead shoots after autumn clipping in creased in A. uva-ursi, suggesting that the reserves in the older plant parts were depleted as a result of nutrient translocation to the new shoots. Both species showed a marked reduction in shoot growth after spring clipping. The effect was more detrimental to A. uva-ursi than to V. uliginosum. This result is in accordance with many earlier studies, in which deciduous species are reported to recover faster from early summer damage than evergreen species (Chapin 1980 a; Archer & Tieszen 1980; Tolvanen & Laine 1997). For instance, deciduous V. myrtillus L. produced many shoots with a low carbohydrate content, while evergreen V. vitis-idaea produced less shoots with an increased carbohydrate content after ar tificial herbivory (Tolvanen & Laine 1997). It has been emphasised that the distinct responses after damage observed between evergreen and deciduous species are a consequence of differences in their storage reserves and source-sink aspects of carbon and nutrient alloca tion (Chapin 1980 a; Archer & Tieszen 1980; Tolvanen & Laine 1997). In addition to reallocation of resources after damage, plant architecture and sectoriality and the number and distribution of meristems determines the general range of tolerance (Marquis 1996). A. uva-ursi and V. uliginosum differ from each other in three essential factors affecting their regrowth ability. Firstly, the belowground storage of carbohy drates and nutrients is generally greater in deciduous than in evergreen species (Chapin 1980 a). For in stance, the proportion of the belowground biomass of V. uliginosum was about 79% whereas that of V. vitis idaea was about 51% in a subarctic area (Karlsson 1987). The proportion of the latter species is of course not comparable with that of A. uva-ursi which has a tap and many adventitious roots. While the car bohydrate and mineral nutrient stores of deciduous shrubs are located in stems or roots, the leaves are important as storage sites in evergreen species (Chapin 1980b, 1983; Karlsson 1985). Secondly, the photo synthesis and growth rates of deciduous species are higher than those of evergreen species (Johnson & Tieszen 1976; Chapin 1980 a; Karlsson 1989). Al though the photosynthesis of evergreen species starts earlier (Karlsson 1989), their growth begins later than that of deciduous species. The shoot growth of V. ulig inosum is based on both current-year photosynthesis and the stored resources, whereas, e.g., V. vitis-idaea mostly uses early-summer assimilates produced by old leaves (Karlsson 1985). It is probable that A. uva-ursi resembles V. vitis-idaea in this respect. Finally, the absorption and translocation of nutrients is faster in deciduous than in evergreen species (Chapin 1980 a). The spring decrease in the nitrogen content of the belowground parts of deciduous shrubs suggests that much of the nitrogen in their leaves is translocated from belowground stores (Chapin 1980b). Current growth in V. uliginosum is observed to be a strong sink for nitrogen (Chester & Oechel 1986). The branches of A. uva-ursi and V. uliginosum ap peared to be rather autonomous in carbon allocation because there was no marked decrease in the growth of within-clone control branches compared to the un dipped control. This observation is in agreement with results reported for other woody plants, e.g., for Pi nus sylvestris (Honkanen & Haukioja 1995). It is known that the sectoriality of plants causes resources, in particular carbon, to flow more freely within mor phological units (e.g., branches) than between such units (Watson & Casper 1984; Marquis 1996; Mar shall 1996). The development of fine roots on the creeping branches of A. uva-ursi made the annual growth segments (modules) more independent of in traclonal transport of water and nutrients than modules of V. uliginosum. The developing fruits are known to be a strong sink for resources (Watson & Casper 1984; Marshall 1996). V. uliginosum had started blooming when the young current shoots were clipped in the spring. Because berries were obviously supported by the assimilates of the neighbouring leaves, the removal of all the shoots dramatically decreased the number of berries. In con trast, overwintered flower buds in the previous-year shoots of A. uva-ursi (Figure lb) were not affected by the removal of new current shoots. The berries of this species were evidently supported by the older leaves of the same undipped shoots. 90 The effect of the distance from the smelter The degree of compensatory growth was not clearly affected by the distance from the smelter in either species, although the biomass of dead branches was highest at the most polluted sites. The prediction of a higher regrowth level in a low-stress environment (Belsky et al. 1993) was only partially fulfilled. A. uva ursi showed some signs of decreased growth after autumn clipping near the smelter, but this reduction was not statistically significant. The degree of re growth of V. uliginosum was also slightly lower near the smelter compared to the other sites after spring clipping. Both species accumulate heavy metals in roots and this may restrict their translocation to leaves (Table 1). In addition to the ability to prevent heavy metals passing into assimilating tissues, the low com petition for resources resulting from the low number of other plant species growing near the smelter, may partly explain the unexpectedly good compensation. Conclusions Contrary to predictions, A. uva-ursi showed as vigor ous regrowth as V. uliginosum after shoot clipping in autumn. Both species suffered more from spring than from autumn clipping, and the compensatory growth of A. uva-ursi was lower than that of V. uliginosum after spring clipping. The degree of compensatory growth was only slightly lower in the highly polluted environment near the smelter than at greater distances. Heavy metals and sulphur have subjected the vegeta tion growing nearest to the smelter to a strong selec tion pressure. The surviving clones probably represent the most resistant genotypes of the populations estab lished at least 30-40 years ago. It is suggested that the vigorous regrowth resulting from activation of the dormant buds and the rapid shoot turnover have signif icantly improved the survival of the studied clones in this polluted environment. In long-lived dwarf shrubs, phenotypic plasticity in morphological and physiolog ical responses probably provides resistance to heavy metals when the evolution of tolerant ecotypes is restricted by the failure of seedling establishment. Acknowledgements Päivi Lindholm, Pekka Suolahti, Taru Uotila and Outi Välilehto helped in the field and laboratory work. Yrjö Sulkala determined the age of the branches and Airi Piira assisted with the data and figures. Erkki Tomppo advised in statistical analyses. John Derome and Satu Monni commented on the manuscript. Hannu Nousi ainen made the drawings of the species. We wish to thank all the above-mentioned persons and the For est Health Fertilization Project of the Finnish Forest Research Institute for the opportunity to use the data collected on the sample plots of the project. References Aarssen, L. W. 1995. Hypothesis for the evolution of apical domi nance in plants: implications for the interpretation of overcom pensation. Oikos 74: 149-156. Archer, S. & Tieszen, L. L. 1980. Growth and physiological re sponses of tundra plants to defoliation. Arct. Alp. Res. 12: 531-552. Baker, A. J. M. 1981. Accumulators and excluders - strategies in the response of plants to heavy metals. J. Plant Nutr. 3: 643-654. Baker, A. J. M. 1987. Metal tolerance. New Phytol. 106: 93-111 Balsberg, A.-M. 1982. Plant biomass, primary production and litter disappearance in a Filipendula uimaria meadow ecosystem, and the effects of cadmium. Oikos: 72-90. Balsberg-Pählsson, A.-M. 1989. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. Water, Air, Soil Pollut. 47: 287-319. Belsky, A. J. 1986. Does herbivory benefit plants? A review of the evidence. Am. Nat. 127: 870-892. Belsky, A. J., Carson, W. P., Jensen, C. L. & Fox, G. 1993. Over compensation by plants: herbivore optimization or red herring? E vol. Ec 01.7: 109-121. Bowles, J. M. 1983. Burl sprouting in Arctostaphylos uva-ursi as a response to trampling damage. Can. J. Bot. 61: 3543-3545. Chapin, F. S. 11l 1980 a. Nutrient allocation and responses to defoliation in tundra plants. Arct. Alp. Res. 12: 553-563. Chapin, F. S. 11l 1980b. The mineral nutrition of wild plants. Ann Rev. Ecol. Syst. 11:233-260. Chapin, F. S. 11l 1983. Nitrogen and phosphorus nutrition and nu trition cycling by evergreen and deciduous understorey shrubs in an Alaskan black spruce forest. Can. J. For. Res. 13: 773-781. Chester, A. L. & McGraw, J. B. 1983. Effects of nitrogen addition on the growth of Vaccinium uliginosum and Vaccinium vitis-idaea. Can. J. Bot. 61: 2316-2322. Chester, A. L. & Oechel, W. C. 1986. Effects of leaf nitrogen availability and leaf position on nitrogen allocation patterns in Vaccinium vitis-idaea and Vaccinium uliginosum. Oecol. (Berl.) 69: 121-125. Dickinson, N. M., Turner, A. P. & Lepp, N. W. 1991. How do tree and other long-lived plants survive in polluted environments? Funct. Ecol. 5: 5-11. Eriksson, O. & Fröborg, H. 1996. 'Windows of opportunity' for recruitment in long-lived clonal plants: experimental studies of seedling establishment in Vaccinium shrubs. Can. J. Bot. 74: 1369-1374. Ernst, W. H. O. 1974. Schwermetallvegetation der Erde. Gustav Fischer Verlag, Stuttgart. Fröborg, H. 1996. Pollination and seed production in five boreal species of Vaccinium and Andromeda (Ericaceae). Can. J. Bot. 74: 1363-1368. 91 Gimmingham, C. H. 1972. Ecology of heathlands. Chapman and Hall, London. Helmisaari, H.-S., Derome, J., Fritze, H., Nieminen, T., Palm gren, K., Salemaa, M. & Vanha-Majamaa, I. 1995. Copper in Scots pine forests around a heavy-metal smelter in south-western Finland. Water, Air Soil Pollut. 85: 1727-1732. Hjälten, J., Danell, K. & Ericson, L. 1993. Effects of simulated her bivory and intraspecific competition on the compensatory ability of birches. Ecology 74: 1136-1142. Honkanen, T. & Haukioja, E. 1995. Why does a branch suffer more after branch-wide than after tree-wide defoliation. Oikos 71:441-450. Jacquemart, A.-L. 1996. Vaccinium uliginosum, L. Biological flora of the British Isles. J. Ecol. 84: 771-785. Johnson, D. A. & Tieszen, L. L. 1976. Aboveground biomass allo cation, leaf growth, and photosynthesis patterns in tundra plant forms in Arctic Alaska. Oecol. (Berl.) 24: 159-173. Karlsson, P. S. 1985. Patterns of carbon allocation above ground in a deciduous (Vaccinium uliginosum) and an evergreen (Vaccinium vitis-idaea) dwarf shrub. Physiol. Plant. 63: 1-7. Karlsson, P. S. 1987. Micro-site performance of evergreen and de ciduous dwarf shrubs in a subarctic heath in relation to nitrogen status. Holarctic Ecol. 10: 114-119. Karlsson, P. S. 1989. In situ photosynthetic performance of four co existing dwarf shrubs in relation to light in a subarctic woodland. Funct. Ecol. 3: 481-488. Laaksovirta, K. & Silvola, J. 1975. Effect of air pollution by copper, sulphuric acid and fertilizer factories on plants at Harjavalta, W. Finland. Ann. Bot. Fenn. 12: 81-88. Levitt, J. 1980. Responses of plants to environmental stresses. Academic Press, New York. Marshall, C. 1996. Sectoriality and physiological organisation in herbaceous plants: an overview. Vegetatio 127: 9-16. Marquis, R. J. 1996. Plant architecture, sectoriality and plant tolerance to herbivores. Vegetatio 127: 85-97. Maschinsky, J. & Whitham, T. G. 1989. The continuum of plant re sponses to herbivory: the influence of plant association, nutrient availabilty, and timing. Am. Nat. 134: 1-19. Packer, J. G. & Denford, K. E. 1974. A contribution to the taxonomy of Arctostaphylos uva-ursi. Can. J. Bot. 52: 743-753. Remphrey, W. R., & Steeves, T. A. 1984 a. Shoot ontogeny in Arctostaphylos uva-ursi (bearberry): the annual cycle of apical activity. Can. J. Bot. 62: 1925-1932. Remphrey, W. R„ & Steeves, T. A. 1984b. Shoot ontogeny in Arc tostaphylos uva-ursi (bearberry): origin and early development of lateral vegetative and floral buds. Can. J. Bot. 62: 1933-1939. Remphrey, W. R., Steeves, T. A. & Neal, B. R. 1983. The mor phology and growth of Arctostaphylos uva-ursi (bearberry): an architectural analysis. Can. J. Bot. 61: 2430-2450. Salemaa, M. & Vanha-Majamaa, I. 1993. Forest vegetation change along a pollution gradient in SW Finland. Proc. First Finnish Conference of Environmental Sciences. Kuopio Univ. Pub. C. Nat. Env. Sci. 14: 101-104. SAS Institute, Inc. 1994. SAS/STAT User's Guide Version 6. 4th Edition. SAS Institute Inc., Cary, North Carolina, USA. Shevtsova, A., Ojala, A., Neuvonen, S., Vieno, M. & Haukioja, E. 1995. Growth and reproduction of dwarf shrubs in a sub arctic plant community: annual variation and above-ground interactions with neighbours. J. Ecol. 83: 263-275. Stewart, W. S. & Bannister, P. 1973. Seasonal changes in carbohy drate content of three Vaccinium spp with particular reference to V. uliginosum L. and its distribution in the British Isles. Flora Bd. 162: 134-155. Sutinen, M.-L., Raitio, H., Nivala, V., Ollikainen, R. & Ritari, A. 1996. Effects of emissions from copper-nickel smelters on the frost hardiness of Pinus sylvestris needles in the subarctic region. New Phytol. 132(3): 503-512. Tiffney, W. N., Benson, D. R. & Eveleigh, D. E. 1978. Does Arc tostaphylos uva-ursi (Bearberry) have nitrogen-fixing nodules? Amer. J. Bot. 65: 625-628. Tolvanen, A. & Laine, K. 1997. Effect of reproduction and artificial herbivory on vegetative growth and resource levels in deciduous and evergreen dwarf shrubs. Can. J. Bot. 75: 656-666. Tolvanen, A., Laine, K., Pakonen, T. & Havas, P. 1995. Recov ery of evergreen clonal dwarf shrub Vaccinium vitis-idaea after simulated microtine herbivory in a boreal forest. Vegetatio 116: 1-5. Tolvanen, A., Laine, K., Pakonen, T., Saari, E. & Havas, P. 1993. Ef fect of habitat and time of clipping on the recovery of the bilberry (Vaccinium myrtillus). Ann. Bot. Fenn. 30: 15-20. Tuomi, J., Nilsson, P. & Äström, M. 1994. Plant compensatory re sponses: bud dormancy as an adaptation to herbivory. Ecology 75: 1429-1436. Tutin, T. G., Heywood, V. H., Burges, N. A., Moore, D. M., Valen tine, D. H., Walters, S. M. & Wegg, D. A. 1972. Flora Europaea. Vol. 3. Cambridge University Press, Cambridge. Tyler, G., Balsberg Pählsson, A.-M., Bengtson, G., Bääth, E. & Tranvik, L. 1989. Heavy metal ecology of terrestrial plants, mi cro organisms and invertebrates. Water, Air, and Soil Poll. 47: 189-215. Verkleij, J. A. C. & Schat, H. 1989. Mechanisms of metal tolerance in higher plants. Pp. 179-193. In: Shaw, J. A. (ed.), Heavy metal tolerance in plants: evolutionary aspects. CRC Press, Florida. Väisänen, S. 1986. Effects of air pollution by metal, chemical and fertilizer plants on forest vegetation at Kokkola, W Finland. Ann. Bot. Fenn. 23: 305-315. Watson, M. A. & Casper, B. B. 1984. Morphogenetic constraints on patterns of carbon distribution in plants. Ann. Rev. Ecol. Syst. 15: 233-258. Paper V Salemaa, M. & Sievänen, R. 2002. The effect of apical dominance on the branching architecture of Arctostaphylos uva-ursi in four contrasting environments. Flora 197: 429-442. Flora (2002) 197,429-442 http://www.urbanfischer.de/journals/flora 0367-2530/02/197/06-429 $ 15.00/0 FLORA (2002) 197 429 The effect of apical dominance on the branching architecture of Arctostaphylos uva-ursi in four contrasting environments Maija Salemaa* & Risto Sievänen Vantaa Research Centre, The Finnish Forest Research Institute, RO. Box 18, FIN-01301 Vantaa, Finland * e-mail corresponding author: maija.salemaa@metla.fi Submitted: Mar 28,2002 ■ Accepted: Jul 19,2002 Summary We studied horizontal spreading and axillary bud activation of colonizing branches of a clonal dwarf shrub, Arctostaphylos uva-ursi, in SW Finland in four habitats with varying pollution, nutrient, light and competition levels. A. uva-ursi showed high plasticity in the proportion of released buds, which changed according to the stress and resource level of the habitats. Apical dominance of lateral branching was the strongest in the resource-poor habitats (low soil nutrient or low light levels). However, when the apex of the parent shoot was terminated (due to mortality of an apical bud or the formation of an inflorescence) the disruption of apical dominance caused intensive branching in the poor habitats. Apical dominance of the dominant shoots was much weaker in the resource-rich habitats. When nutrient availability and light level were relatively high, branching frequency was high in both intact (expanded) and terminated parent shoots. Shoot mortality and the proportion of terminated shoot apices was the highest in the polluted habitat, as a result of soil toxicity. This caused intensive branching in all shoot hierarchies, which may enhance the plant's survival in heavy metal polluted soil. We conclude that plastic branching enables this species to recover after damage, or to respond to changed resource levels according to the 'reserve meristem hypothesis'. The branching response to different environmental conditions was simulated by means of an L-system architectural model based on the demographic and morphological parameters measured in each habitat. Key words: bearberry, clonal growth, heavy metals, L-system architectural model, meristem activation, revegetation Introduction Modular growth enables clonal plants to respond to environmental variation by modifying their shape (e.g. Hutchings & de Kroon 1994; de Kroon & Hut chings 1995). Several studies have described changes in the internodal length of the stems, lateral branching intensity and branching angle of clonal plants in re sponse to environmental conditions (Slade & Hut chings 1987; Callaghan et al. 1990; de Kroon & Knops 1990; Dong et al. 1997; Moen et al. 1999). It has been demonstrated that simple computer algorithms can generate large differences in clonal morphology (BELLetal. 1979; Bell & Tomlinson 1980; Suther land & Stillman 1990). The role of apical dominance is crucial in determining the form of many plant species (Hutchings & Mogie 1990). The control exerted by the shoot apex over the outgrowth of lateral buds causes directional growth of elongated primary shoots. Death or removal of the shoot apices as a result of damage disrupts apical dominance and enables regrowth through lateral branching (Cline 1991, 1997). A number of hypotheses on the selective advantages of apical dominance have been proposed, e.g. competition for light, reduction in within-plant competition (Aarssen 1995; Irwin & Aarssen 1996), keeping dormant meristems in reserve as a strategy against possible damage (Tuomi et al. 1994; Lortie & Aarssen 2000), and the function of dominant apices as a metabolic sink under nutrient and water stress (Lortie & Aarssen 1997). The strength of apical dominance varies between species and between different environ mental conditions (Hutchings & de Kroon 1994; Bonser & Aarssen 1996). The growth of the plant may be limited by the number and activity of meristems, which affect its ability to respond to damage or changes in resource availability (Harper 1977; Salomonson et al. 1994; Tolvanen & Laine 1997). 430 FLORA (2002) 197 Fig. 1. Morphological structure of a) a compact clone, and b) a colonizing branch of A. uva-ursi, with three age classes of shoots (yearly grown segments). Current-year shoots: la) monopodial dominant (D) shoot with an intact apical bud, lb) sympodial subdomiant (SD) shoot with an inflorescence, 1 c) monopodial SD shoot with a dead apical bud, 1 d) sympodial nondominant (ND) shoot with an intact apical bud. Previous-year shoots: 2a) monopodial expanded D shoot, 2b) sympodial SD shoot with berries, 2c) sympodial expanded SD shoot. 3 rd year shoots: 3 monopodial expanded D shoot. Shoot hierarchy levels: D domi nant, SD subdominant and ND nondominant. Location %: the distance between the attachment point of the lateral shoot from the apex of the parent shoot / the total length of the parent shoot. Branching angle (a) between the lateral and parent shoots is marked. Bonser & Aarssen (1996) presented general pre dictions of the optimal pattern of axillary meristem allo cation of herbaceous plants (here only polycarpic perennials) along light (L) and nutrient (N) gradients. They propose that 1) Where soil nutrients are abundant, but light is limit ing due to crowding from neighbours (N+L-), keeping axillary meristems inactive (strong apical dominance) is favoured in order to enable vertical growth/extended horizontal axes. 2) Where both nutrients and light are abundant (N+L+), releasing axillary meristems (weak apical dominan ce) is favoured in order to maximize resource uti lization through intensive branching. 3) Where nutrients are limiting, but light is abundant (N-L+), keeping axillary meristems in an inactive state (strong apical dominance) is favoured in order to extend access to richer nutrient patches in the soil or to hold inactive meristems in reserve, from which growth can be increased if a nutrient rich patch is encountered or if apical meristems are destroyed (Aarssen 1995). In this paper we test the above predictions using a clo nal evergreen dwarf shrub Arctostaphylos uva-ursi L. (bear berry) as the study plant (Fig. 1). This species is suitable for studying the state of meristems (i.e. keeping them inactive, or activating them to form vegetative or reproductive shoots) because of existing detailed analyses of its architecture (Remphrey et al. 1983 a, b; Remphrey & Steeves 1984 a, b), and because it is rela tively easy to determine the terminal status of the shoots and to identify the axillary buds on leaf axils. We address the hypotheses by considering the pro strate growth habit of A. uva-ursi in four habitats with varying pollution, nutrient, light and competition levels. We designed a model for the branching architecture using the results of measurements. The purpose of the model is to summarize the empirical results and to study qualitative features of the branching patterns by simula tions. FLORA (2002) 197 431 Material and methods The study plant The xeromorphic evergreen dwarf shrub, Arctostaphylos uva ursi (L.) Spreng., grows circumpolarly in boreal and temp erate forests and in dry subalpine sites above the forest limit (Schutt & Lang 2000). It colonises open sandy heathlands and moraine slopes, eroded banks, sand dunes and a variety of other habitats in the boreal forest (Remphrey et al. 1983 a; Salemaa et al. 1999). It is a pioneer of light, dry and nutrient poor habitats, but occurs also in the understorey of dry Scots pine forests in its Nordic range. It is one of the few understorey species which survives in polluted soil close to non-ferrous smelters (Väisänen 1986; Lukina et al. 1993; Salemaa et al. 2001). The clones growing in open sites form compact patches, with a deep tap root (50-100 cm) in the centre and prostrate branches (stems) spreading radially around the plant (Fig. 1 a). The stems may produce superficial adventitious roots as they get older. A. uva-ursi regenerates rather poorly from seed and spreads mainly vegetatively. In this study we follow the terminology presented by Remphrey et al. (1983 a, b) on the morphological units of A. uva-ursi. The outer zone of the patch consists of branches with extended shoots and a linear growth pattern (colonizing complexes), and the inner zone shorter branches with more upright orientation (maintaining complexes) (Fig. la). The growth segments formed during one year (later on called shoots) are divided into three hierarchy types: dominant (D) shoots are long and low, whereas subdominant (SD) and non dominant (ND) shoots are shorter and grow with higher ele vation angles (Fig. lb). The differences between the lengths of D, SD and ND shoots were statistically significant (p < 0.05) in most cases in the study material (Fig. 2a). Each leaf axil has a single meristem (axillary bud), which is initially dormant but may be released in subsequent years. Shoot apices are terminated either by vegetative buds or by inflorescences. Intact apical buds of parent shoots expand monopodially into daughter shoots in the next growing season or remain dormant. If apical buds die or produce inflores cences, branching continues sympodially from axillary buds during the next year (Fig. lb). Thus, in contrast to the meri stem allocation model of Bonser & Aarssen (1996), repro ductive shoots of A. uva-ursi have the capacity to continue lateral growth after flowering. Terminal inflorescences over winter and berries develop during the following year. The shoots do not grow during the year that they produce berries. Study sites The study area is a dry, infertile Scots pine (Pinus sylvetris L.) forest of the southern boreal coniferous zone, located close to the Harjavalta Cu-Ni smelter (61°19'N, 22°09'E) in SW Fin land. The soil is sorted fine sand and the soil type ferric pod zol (Mälkönen et al. 1999). The forest ecosystem near the smelter has been drastically changed by heavy-metal and Table 1. Site characteristics and element concentrations in the growing substrate in four habitats of A. uva-ursi. Plant available (exchangeable) element concentrations in the mineral soil layer (at 0-10 cm depth) for the polluted (data 0.5 km S from the smelter in the year 1999), sand pit and forest habitats (year 2001) and in the mulch (year 1998) added to the restoration habitat. All element concentration in sand were determined with BaCl2 + EDTA. Concentrations in the mulch were deter mined with BaCl 2 and calculated per mass organic matter. The relative competition level for light and belowground resources was subjectively estimated according to the overall density of the ground vegetation. Habitat Polluted Restoration Sand pit Forest Distance from smelter, km 0.5 0.5 6 8 Direction from smelter W S SE SE Shading of overstorey none open canopy none dense canopy Vegetation coverage, % <1 <10 <10 100 Substrate sand mulch sand sand Element conc. K, mg/kg 12.6 375.6 25.5 29.5 Mg, mg/kg 2.8 649.0 4.5 10.6 Ca, mg/kg 20.8 7888.0 62.8 73.0 P. mg/kg 3.8 87.6 4.4 7.4 Cu, mg/kg 84.2 190.7 17.7 8.3 Ni, mg/kg 11.3 11.8 2.6 2.3 Pollution level + + Nutrient level ++ + + Light level + ± + Competition level _ _ _ + 432 FLORA (2002) 197 sulphur emissions during the past 50 years (Helmisaari et ai. 1995; Fritze et ai. 1997; Derome & Lindroos 1998; Derome & Nieminen 1998; Salemaa et ai. 2001). During 1985-1990, the average annual emissions of Cu were 104 t, Ni 501 and S02 81001. By 1999, the corresponding values had decreased to Cu 6 t, Ni 1 t and S02 3 400 t (data from Outo kumpu Harjavalta Metals smelter). Since 1995 the smelter has been emitting considerable amounts of gaseous NH 3 , which has been reflected as elevated pH values in precipitation and increased nitrogen deposition (Derome 2000). The long-term accumulation of Cu and Ni in the soil close to the smelter has resulted in a severe deficit of plant available Ca, Mg and K in the organic layer (Derome 2000). The understorey vegetation is almost totally absent up to a distance of 0.5 km. Visible toxic effects of heavy metals on the vegetation extend up to a distance of 3-4 km (Salemaa et al. 2001). The distances of 6 and 8 km represent "clean" sites in this study, although even at these distances the heavy metal concentrations in the organic layer exceed background levels (Derome 2000). The general characteristics, element concen trations in the upper soil layers and the relative light (L), nutri ent (N), pollution and competition levels of the four habitats are given in Table 1. The polluted habitat (N-L+) is a severely disturbed Scots pine stand at a distance of 0.5 km to the W of the smelter. The forest floor is almost completely devoid of vegetation and covered with undecomposed needle litter. Only a few patches of A. uva-ursi, Empetrum nigrum L. and resistant pioneer mos ses survive there. The clones of A. uva-ursi may be relicts and date back to the time before the smelter was founded in the 1940'5. The soil is very dry and the nutrient status poor, and the toxic effects of heavy metals restrict nutrient uptake by the roots. Light availability is high in this open, damaged stand. Interspecific competition is insignificant because the site is almost empty of species. The restoration habitat (N+L+) is a revegetation experiment in an open Scots pine stand 0.5 km to the S of the smelter. The experiment was established in spring 1996 by spreading a mulch consisting of compost and wood chips onto the pollut ed forest floor (Kiikkilä et ai. 2001). A. uva-ursi was one of the species that were planted on one-species, 5 x 5 m experi mental quadrats in attempts to revegetate and stabilise the pol luted soil. The rooted cuttings of A. uva-ursi were of local Harjavalta origin (taken in April 1996), and had one shoot (formed in 1995) when planted. The light level is relatively high because the stand is seriously defoliated. The nutrient rich mulch pockets (see Table 1 for chemical data from 1998) ensure high nutrient availability to the roots and protect them against heavy metals in the soil. The understorey is not closed, indicating a low competition level between the plants. The sand pit habitat (N+L+) 6 km to the SE of the smelter represents one of the most favourable sites for A. uva-ursi. At this site the light, nutrient and moisture conditions are all suf ficiently high to allow maximum biomass production. The clones have regenerated naturally decades ago along the open edges of the abandoned sand pit. The pioneer mosses and lichens and Scots pine seedlings have low competitive effects on A. uva-ursi. The forest habitat (N+L-) 8 km to the SE of the smelter is relatively fertile and half-shaded by the overstorey Scots pines. The naturally regenerated A. uva-ursi clones may be decades old. The A. uva-ursi clones are intermixed within the closed understorey (mosses, Cladina lichens and other dwarf shrubs e.g. Vaccinium vitis-idaea L., Calluna vulgaris (L.) Hull). Between-species competition for light and space is probably high. Branch sampling We randomly selected five separate clones of A. uva-ursi in the polluted, sand pit and forest habitats in September, 2000. Altogether 1 -3 branches with the six youngest shoots (formed Table 2. Number of studied clones, branches and shoots in different habitats, n, = number of all shoots (age classes 1996-2000) (Fig. 2a,b),n2 = number of lateral shoots (1996-2000) (Fig. 2c, d), n 3 = number of shoots with intact apex (1996-1999) (Fig. 3) and n4 = number of shoots with termi nated apex (1996-1999) (Fig. 3 ). Shoot hierarchy types: D = dominant, SD = subdominant and ND = nondominant. Habitat Clones Branches Type n i n2 n 3 n4 Polluted 5 13 D 101 37 55 28 SD 104 52 45 21 ND 125 68 7 14 Restoration 10 10 D 120 34 77 14 SD 338 151 166 30 ND 483 126 135 24 Sand pit 5 5 D 42 11 27 6 SD 100 52 48 14 ND 232 140 89 25 Forest 5 9 D 78 27 41 17 SD 108 49 53 15 ND 62 35 20 12 FLORA (2002) 197 433 Fig. 2. Shoot-specific means (± se) of morphological variables measured on living shoots in the 1996-2000 age classes in four habitats (Poll. = Pollut ed. Rest. = Restoration, Sand = Sand pit and Forest = Forest habitats). Pairwise comparisons between the means were performed by Tukey's tests. The same letter indicates that there are no signi ficant differences (p > 0.05) between the means. Shoot hierarchy types: D = dominant, SD = subdominant and ND = nondominant. The number of clones, branches and shoots given in Table 2 (n1 = number of all shoots and n 2 = number of lateral shoots). during 1995-2000) were taken from the colonizing zone of each clone. In the restoration habitat we removed ten plants, each comprising six shoots (1995-2000) together with the roots. Thus all the sample branches had one parent shoot, formed in 1995, from which all the daughter shoots had devel oped. In the sand pit habitat, two branches had the parent shoot from 1996. Detailed morphological analysis of the branches was carried out in the laboratory. This included measuring the length, location of attachment and branching angle of the shoots, counting the number of activated and inactive buds, and determining the age, hierarchy and terminal types of the shoots. The total number of branches and shoots analysed is given in Table 2. Statistical analyses We calculated shoot-specific mean values for the morpholo cial variables from the data representing the age classes 1996-2000 (1996-1999 for lateral branching because cur rent-year shoots do not branch). The pairwise differences in the mean values over the habitats and shoot hierarchy types were tested by Tukey's tests. The effect of the habitat, termi nal type (intact or terminated apex) and age of the shoots, and their interaction in the number and proportion of released buds on D shoots, were tested using three-factor ANOVA. The dif ferences between the intact and terminated parent shoots were compared by two-sample t tests and between the habitats with in the same shoot hierarchy type by Tukey's tests. The time delay in the production of released buds was studied by means of linear regression models. All the tests were carried out using SAS software (SAS Inc. 2000). The frequency of dead and living shoots and terminal types were calculated separately for the current-year shoots (2000) and for the combined data of the older (overwintered) shoots (1996-1999). The mortality rates and frequencies of the ter minal types were tested pairwise between the habitats using two-by-two contingency tables and between intact or termi nated parent shoots using chi-square tests (Statistix 7, Ana lytical Software 2000). Model for the morphological development of A. uva-ursi We constructed a morphological model in order to summarize the experimental results and to gain an understanding of the processes affecting the branching pattern of A. uva-ursi. We 434 FLORA (2002) 197 applied the modelling approach of Remphrey et al. (1983b). Our model is based on the use of an annual time step and a set of rules, derived from measurements, to control the production of new shoots. The plant is a collection of D, SD, and ND shoots. The shoot may be intact or terminated. In a growth cycle of one year a parent shoot produces a number of daughter shoots. We used the field measurements from the four habitats to parameterize the model. It is presented in detail in Ap pendix. We evaluated the validity of the model by comparing the rate of increase (r) in the number of living shoots during the first five years, calculated from the mean of ten simulations with the parameters measured in the field. We linearized the exponential growth equation (N, = Noe", in which N 0 = initial number of shoots, t = time in years, r = rate of increase) as In N, = In N 0 + rt. Differences in the values of r parameters (regression slopes) between the observed and model-simula ted data were tested by t-tests (Statistix 7, Analytical Soft ware 2000). Results Basic morphology The length of the D shoots of A. uva-ursi followed the order: sand pit > restoration > forest > polluted (Fig. 2a). The SD and ND shoots were also the longest in the sand pit, but there were no differences between the other habitats. D shoots were clearly longer than SD and ND shoots within the habitats, except in the polluted one where the length of D and SD shoots did not differ sig nificantly (Fig. 2a). There were no statistical differences in the lengths of intact or terminated D shoots. The total number of axillary buds (Fig. 2b) correlated positively with the shoot length (r = 0.774, p < 0.0001, n = 371 in the combined data of D shoots for all sites). The branching angles of D, SD and ND shoots were relatively similar, varying between 26-36 degrees in all the habitats, except for the ND shoots in the forest which had a wider angle (52°) and upright orientation (Fig. 2c). The relative location of attachment of most of the lateral shoots was near to the apex of the parent shoot (Fig. 2d). However, the lower the shoot hierarchy, the further away from the shoot apex they grew. The number of shoots with green leaves was the high est in the restoration habitat where the four youngest year classes (1997-2000) had leaves. The other sites followed the order: polluted (3 green year classes) > sand pit (2.5) > forest (2). The length of the main roots was about 30 cm in the plants excavated from the resto ration experiment, but no adventitious roots had yet developed. In the other habitats the branches had occa sional adventitious roots along the branches in the 1995-1999 year classes. A number of second flushes (sensu Remphrey & Steeves 1984 a) produced by the current-year shoots were found in the restoration habi tat. Lateral branching The most important factors that affected lateral branch ing of the D shoots were the habitat, terminal type and age of the parent shoots (Table 3). The effects of the last two factors varied according to the resource level of the habitat (significant interaction terms in ANOVA). The overall pattern of lateral branching was similar, irre spective of whether it was calculated on the basis of the number or the proportion of released buds. In general, apical dominance of the intact D shoots on lateral branching was stronger in the resource-poor (pol luted and forest) than in the resource-rich (restoration and sand pit) habitats (Fig. 3a). The proportion of re leased buds was significantly lower (p < 0.001) on intact than on terminated D shoots in the polluted and forest habitats (Fig. 3a). However, when the apex of the parent shoot had been terminated, the breakage of apical domi nance caused intensive branching in the poor habitats. Apical dominance of D shoots was much weaker in the resource-rich habitats. This was expressed in the similar Table 3. The effect of the habitat, terminal type (intact or terminated) and age of the dominant shoots on the number of released buds and the proportion of released buds. F and p values for three-factor ANOVA. Only significant interactions given, df = degree of freedom. Data from years 1996-1999. Source of df Number of released buds Propo rtion of released buds variation F P F P Habitat 3 9.68 <0.0001 4.93 0.0024 Terminal type 1 17.90 <0.0001 28.16 <0.0001 Age 3 5.89 0.0007 8.13 <0.0001 HxT 3 4.34 0.0053 11.73 <0.0001 Hx A 9 3.87 0.0001 3.59 0.0003 TxA 3 - - 2.77 0.0422 Error 234 FLORA (2002) 197 435 (p>0.05), high proportion of released buds in both intact and terminated D shoots in the restoration and sand pit habitats (Fig. 3a). The proportion of released buds of intact D shoots was higher (p < 0.05) in the resource-rich habitats than in the resource-poor habitats (Fig. 3 a). In contrast, the proportion of released buds on terminated D shoots was relatively similar between the habitats. Intact SD and ND shoots had a significantly lower (p<0.05) proportion of released buds than terminated ones in all the habitats (Fig. 3b, c). There were no great between-habitat differences in lateral branching of SD and ND shoots, although the proportion of released buds in terminated ND shoots tended to be higher in the pol luted (19%) than in the other habitats (9-15%). In gene ral, the proportion of released buds tended to decrease according to shoot hierarchy type in the order D > SD > ND. The proportion of released buds increased with the age of the intact parent D shoots (significant positive regres sion slopes in the restoration, sand pit and forest habitats) (Fig. 4). The age trend was not as clear in the terminated D shoots, although the proportion of released buds in creased from 1- to 3-year-old shoots in the other habitats except for the polluted one. The decrease in the branching in the 4-year-old shoots may be an artefact caused by the small amount of data and the problems in identifying points at which dead branches have disappeared. Shoot mortality and terminal types The mortality rate of the current-year shoots (2000) was lower than that of the older over-wintered age classes (1996-1999) (Table 4a, b). Most of the dead shoots represented the ND hierarchy type. The mortality rate of the older ND shoots was higher in the resource-poor (57% in the polluted and 30% in the forest habitats) than in the resource-rich habitats (7% in both the restoration and sand pit habitats) (Pearson's X 2 = 70.76, p < 0.001, n = 390) (Table 4 b). The current-year shoots terminating in the inflores cences were the most abundant in the polluted and sand Fig. 3. The mean proportion (%) (± se) of released buds in a) dominant, b) subdominant and c) nondominant shoots. The intact (white bar) and terminated (hatched bar) shoots compar ed by two-sample t-tests (significancies p < 0.001 = ***, p < 0.01 = **, p < 0.05 = *, ns = not significant). The statisti cal differences (Tukey's tests) between the habitats within the same terminal type marked with letters: the same letter indi cates that there are no significant differences (p > 0.05) be tween the means. The number of shoots in each shoot catego ry in Table 2 (n 3 = number of intact shoots and n4 = number of terminated shoots, 1996-1999). Abbreviations as in Fig. 2. 436 FLORA (2002) 197 Fig. 4. The proportion (%) of released buds (ẋ ± se) on the intact (I, white circles) and terminated (T, black circles) parent D shoots as a function of the shoot age (data from the years 1996-1999). The significant slopes (b) of linear regressions are given. pit habitats where the light level was high (Table 4 a). However, the frequency of shoots bearing berries was much lower than that assumed according to the number of inflorescences. The frequency of dead apices in the older D shoots was higher in the resource-poor habitats (polluted and forest combined 32%) than in the re source-rich habitats (restoration and sand pit combined 16%) (x 2 = 8.88, p < 0.0029, n = 265) (Table 4b), The occurrence of shoot apices with dormant buds was con siderable only in the polluted habitat. Hierarchy types of the daughter shoots The terminal daughter shoot that expanded from the api cal bud mainly represented the same hierarchy type as the parent shoot. The type of lateral daughter shoot was never higher than that of the parent. Thus D parents pro duced all types of hierachy (D, SD and ND), but ND parents only ND daughters. When the apex of the parent shoot had been terminated, the proportion of lateral shoots with a higher hierarchy increased (Fig. 5). The increase was the most significant in the polluted (D shoots: x 2 = 10.8, p = 0.004, df = 2) and the restoration (D shoots: x 2 = 13.2, p = 0.002, df = 2) habitats. Simulations of the branching pattern Simulations produced a variety of branching patterns to the colonizing complexes of A. uva-ursi depending on the pollution and resource levels of the habitat (Fig. 6). The between-habitat differences in the size, shape and lateral spreading of the branches were clearly expressed already during the five-year growth cycle. When the fre quency of terminated D shoots and shoot mortality rate were decreased in the polluted habitat in order to predict the possible effect of decreasing emissions, allocation to lateral branching also decreased (Fig. 6a, b). Both the observed and simulated number of living shoots increased exponentially over time in all the habitats. The observed / simulated values of the rate of increase (r) were: polluted 0.62 / 0.65 (t = 1.19, df= 1,116, p = 0.277), restoration 0.89/0.91 (t = 0.56, df = 1, 96, p = 0.458), sand pit 0.91 / 0.89 (t = 0.10, df = 1, 71, p = 0.748), and forest 0.63 / 0.63 (t = 0.01, df = 1, 91, p = 0.927). The close correspondence be tween the simulated and observed r values showed that the model is able to produce the basic features of branch ing dynamics during the first five years. Discussion The general predictions of optimal meristem allocation presented by Bonser & Aarssen (1996) were valid for the branching pattern of D shoots of A. uva-ursi. Rela tively high allocation in lateral branching (weak apical control) in order to maximize growth was found when light and nutrients were abundant. In contrast, a rela tively high proportion of meristems were kept inactive (strong apical control) under light or nutrient limitation. However, the latter pattern was expressed only in the 437 FLORA (2002) 197 Table 4. The number of living and dead shoots, the frequency (%) of dead shoots and the frequency (%) of terminal types of living shoots in a) current-year shoots (2000) and b) the combined data of the over-wintered shoot age classes 1996-1999. intact shoots of plants growing in resource-poor sites. The predictions were also confirmed by the simulations (Fig. 6). The frequency of dead shoot apices in the polluted habitat was high owing to the direct and indirect effects of air pollution (Fig. 6a). This caused abundant lateral branching in all shoot hierarchy types. The number of lateral shoots per unit length was high because the parent shoots were short, increasing the "branchy" appearance of the clones at the polluted site. However, high annual mortality decreased the shoot density. Rapid shoot turnover may serve as a resistance mechanism, which enables the plant to exclude heavy metals from the living biomass and to recover from biomass loss (Salemaa et al. 1999). Inactive buds were kept in reserve on the intact shoots. We predict that the fre quency of terminated shoots and shoot mortality will decrease in the future as a consequence of decreasing pollution emissions. However, nutrient levels in the soil will still remain low and the light level high. Although the overall increase in the vigour of the plants is expect ed to weaken apical dominance, the deficiency of nutri ents affects oppositely. In this situation the simulations generated a plant architecture with linearly extended branches (Fig. 6b). The rate of increase of the living shoots was the high est in the restoration experiment. Addition of the mulch has increased plant available nutrients and decreased heavy metal concentrations (Kiikkilä et ai. 2001). The plants responded to the nutrient addition (from the a) Apices of living shoots, % Habitat Type Living Dead Death % Intact Dead Dormant Flower Polluted D 18 0 0.0 83.3 5.6 0.0 11.1 SD 38 0 0.0 60.5 5.3 0.0 34.2 ND 104 5 4.6 76.9 5.8 0.0 17.3 Restoration D 2 1 3.3 89.7 10.3 0.0 0.0 SD 14 0 0.0 93.0 6.3 0.0 0.7 ND 324 6 1.8 94.4 5.6 0.0 0.0 Sand pit D 9 0 0.0 100.0 0.0 0.0 0.0 SD 38 0 0.0 68.4 5.3 0.0 26.3 ND 118 0 0.0 78.8 2.5 0.0 18.6 Forest D 20 0 0.0 95.0 0.0 0.0 5.0 SD 40 0 0.0 100.0 0.0 0.0 0.0 ND 30 0 0.0 100.0 0.0 0.0 0.0 b) Apices of living shoots, % Habitat Type Living Dead Death% Intact Dead Dormant Berry Polluted D 83 3 3.5 66.3 32.5 0.0 1.2 SD 66 5 7.0 68.2 30.3 1.5 0.0 ND 21 28 57.1 33.3 42.9 23.8 0.0 Restoration D 91 0 0.0 84.6 15.4 0.0 0.0 SD 196 1 0.5 84.7 15.3 0.0 0.0 ND 159 13 7.6 84.9 13.2 1.9 0.0 Sand pit D 33 0 0.0 81.8 18.2 0.0 0.0 SD 62 0 0.0 77.4 17.7 0.0 4.8 ND 114 9 7.3 78.1 18.4 1.8 1.8 Forest D 58 0 0.0 70.7 29.3 0.0 0.0 SD 68 6 8.1 77.9 22.1 0.0 0.0 ND 32 14 30.4 62.5 37.5 0.0 0.0 438 FLORA (2002) 197 Fig. 5. The proportion (%) of different hierarchy types of the lateral daughter shoots produced by intact (I) and terminated (T) parent shoots. D = dominant, SD = subdominant and ND = nondominant shoot. Significant differencies between I and T parent shoots (x 2 test): p •) as a function of the applied amount of Cu (x) in older stems, fine roots, aboveground parts and the whole plant. by = 0.12 x° 6S , r 2 = 99%, P< 0.001. cy = OJIx 066 , r 2 = 99%, />< 0.001. dy = 0.92x°- 6 \ r 2 = 99%, P< 0.001. ey= 1.67 x° 65 , r* = 99%, P C. vulgaris > A. uva-ursi (most sensitive). Acknowledgements Kaarina Pynnönen and Satu Smolander helped in the greenhouse work. Hilkka Granlund and Airi Piira assisted in weighing the samples and recording the data. Heljä-Sisko Helmisaari, Tiina Nieminen and Pasi Rautio commented on the manuscript, and John Derome checked the English language. We express our sincere gratitude to all. References Aarssen. L.W., 1995. Hypothesis for the evolution of apical dom inance in plants: implications for interpretation of over compensation. Oikos 74, 149-156. Baker, A.J.M., 1987. Metal tolerance. New Phytology 106, 93-111 Baker, A.J.M., Walker, P.L., 1989. Physiological responses of plants to heavy metals and the quantification of tolerance and toxicity. Chemical Speciation and Bioavailability 1, 7-17. Balsberg Pählsson, A.-M., 1989. Toxicity of heavy metals (Zn, Cu, Cd, Pb) to vascular plants. Water, Air, and Soil Pollution 47, 287-319. Bradley, R., Burt, A.J., Read, D.J., 1981. Mycorrhizal infection and resistance of heavy metal toxicity in Calluna vulgaris. Nature 292, 335-337. Bradley, R., Burt, A.J., Read, D.J., 1982. The biology of mycorrhiza in the Ericaceae. VIII. The role of mycorrhizal infection in heavy metal resistance. New Phytologist 91, 197-209. Bradshaw, A.D., Hardwick, K., 1989. Evolution and stress—geno- typic and phenotypic components. Biological Journal of the Lin nean Society 37, 137-155. Chernenkova, T.V., Kuperman, R.G., 1999. Changes in the forest communities along a heavy metal deposition gradient on Kola Peninsula. Water, Air and Soil Pollution 111, 187-200. Cumming, J.R., Tomsett, A.8., 1992. Metal tolerance in plants: signal transduction and acclimation mechanisms. In: Adriano, D.C. (Ed.), Biogeochemistry of Trace Metals. Boca Raton Publishers, Lewis, pp. 329-364. Dahlquist, R.L., Knoll, J.W., 1978. Inductively coupled plasma atomic emission spectrometry: analysis of biological materials and soils for major, trace, and ultra-trace elements. Applied Spectro scopy 32, 1-30. Derome, J., 2000 a. Effects of heavy-metal and sulphur deposition on the chemical properties of forest soil in the vicinity of a Cu-Ni smel ter, and means of reducing the detrimental effects of heavy metals. Finnish Forest Research Institute, Research papers 769, 77. p. Derome, J., 2000b. Detoxification and amelioration of heavy-metal contaminated forest soils by means of liming and fertilisation. Environmental Pollution 107, 79-88. Dickinson, N.M., Turner, A.P., Lepp, N.W., 1991. How do tree and other long-lived plants survive in polluted environments? Functional Ecology 5, 5-11. Ernst, W.H.0., Nelissen, H.J.M., 2000. Life-cycle phases of zinc- and cadmium-resistant ecotype of Silene vulgaris in risk assessment of polymetallic mine soils. Environmental Pollution 107, 329-338. Galli, U., Meier, M., Brunold, C., 1993. Effects of cadmium on non mycorrhizal and mycorrhizal Norway spruce seedlings [Picea abies (L.) Karst.] and its ectomycorrhizal fungus Laccaria laccata (Scop, ex Fr.) Bk. & Br.: sulphate reduction, thiols and distribution of the heavy metal. New Phytologist 125, 837-843. Galli, U., Schuepp, H., Brunold, C., 1994. Heavy metal binding by mycorrhizal fungi. Physiologia Plantarum 92, 364-368. Hall, J.L., 2002. Cellular mechanisms for heavy metal detoxification and tolerance. Journal of Experimental Botany 53, 1-11. Helmisaari, H.-S., 2000. Areas polluted by heavy metals. In: Mäl könen, E. (Ed.), Forest condition in a changing environment—the Finnish case. Kluwer Academic, pp. 248-251. Hutchings, M., de Kroon, H., 1994. Foraging in plants: the role of morphological plasticity in resource acquisition. Advances in Eco logical Research 25, 159-237. Kahle, H., 1993. Response of root of trees to heavy metals. Environ mental and Experimental Botany 33, 99-119. Karlsson, P.S., 1985. Patterns of carbon allocation above ground in a deciduous (Vaccinium uliginosum) and an evergreen (Vaccinium vitis-idaea) dwarf shrub. Physiologia Plantarum 63, 1-7. Kiikkilä, O. 2002. Remediation through mulching with organic matter M. Salemaa, S. Monni / Environmental Pollution 126 (2003) 435-443 443 of soil polluted by a copper-nickel smelter. Finnish Forest Research Institute, Research Papers 831, 57p. Levitt, J., 1980. Responses of plants to environmental stresses. Aca demic Press, New York. Lukina, N.V., Liseenko, L.A., Belova, E.A., 1993. Pollution-induced changes in the vegetation cover of spruce and pine ecosystems in the Kola Peninsul. In: Kozlov, M.V., Haukioja, E., Yarmishko, V.T. (Eds.), Aerial Pollution in Kola Peninsula. Proceedings of the International Workshop, 14-16 April 1992. St. Petersburg, Kola Scientific Centre, Apatity, Russia, pp. 312-321. Marschner, H., 1995. Mineral nutrition of higher plants, 2nd ed. Academic Press, London. Meharg, A.A., Cairney, J.W.G., 2000. Co-evolution of mycorrhizal symbiosis and their hosts to metal-contaminated environments. Advances in Ecological Research 30, 69-112. Monni, S., Salemaa, M., Millar, N., 2000 a. The tolerance of Empetrum nigrum to copper and nickel. Environmental Pollution 109, 221-229. Monni, S., Salemaa, M., White, C., Tuittila, E., Huopalainen, M., 2000b. Copper resistance of Calluna vulgaris originating from the pollution gradient of a Cu-Ni smelter, in southwest Finland. Environmental Pollution 109, 211-219. Mälkönen, E., Derome, J., Fritze, H., Helmisaari, H.-S., Kukkola, M., Kytö, M., Saarsalmi, A., Salemaa, M., 1999. Compensatory fertili zation of Scots pine stands polluted by heavy metals. Nutrient Cycling in Agroecosystems 55, 239-268. Perotto, S., Girlanda, M., Martino, E., 2002. Ericoid mycorrhizal fungi: some new perspectives on old acquaintances. Plant and Soil 244, 41-53. Read, D.J., 1983. The biology of mycorrhiza in the Ericales. Canadian Journal of Botany 61, 985-1004. Remphrey, W.R., Steeves, T.A., Neal, 8.R., 1983. The morphology and growth of Arctostaphylos uva-ursi (bearberry): an architectural analysis. Canadian Journal of Botany 61, 2430-2450. Salemaa, M., Sievänen, R., 2002. The effect of apical dominance on branching architecture of Arctostaphylos uva-ursi in four contrasting environments. Flora 197, 429-442. Salemaa, M., Vanha-Majamaa, 1., Gardner, P.J., 1999. Compensatory growth of two clonal dwarf shrubs, Arctostaphylos uva-ursi and Vaccinium uliginosum in a heavy metal polluted environment. Plant Ecology 141, 79-91. Salemaa, M., Vanha-Majamaa, 1., Derome, J., 2001. Understorey vegetation along a heavy-metal pollution gradient in SW Finland. Environmental Pollution 112, 339-350. Salemaa, M., Vanha-Majamaa, 1.. Reinikainen, A., Nousiainen, H., 2000. Response of understorey vegetation to heavy metal loading. In: Mälkönen, E. (Ed.), Forest Condition in a Changing Envir onment the Finnish Case. Kluwer Academic, pp. 266-275. Scannerini, S., Bonfante-Fasolo, P., 1983. Comparative ultrastructural analysis of mycorrhizal associations. Canadian Journal of Botany 61,917-943. Schiitt, P., Lang, U.M., 2000. Arctostaphylos uva-ursi (Linne) Spreng., 1825. In: Schiitt, P., Weisgerber, H., Schuck, H.J., Lang, U.M., Roloff, A. (Eds.), Enzyklopädie der Holzgewächse. Handbuch und Atlas der Dendrologie, 19. Erg.Lfg. 3/00. Shainberg, 0., Rubin, 8., Rabinowitch, H.D., Tel-Or, E., 2001. Loading beans with sublethal levels of copper enhanches condition ing to oxidative stress. Journal of Plant Physiology 158, 1415-1421. Stribley, D.P., Read, D.J., 1976. The biology of mycorrhiza in the Ericaceae. VI. The effects of mycorrhizal infection and concen tration of ammonium nitrogen on growth of cranberry (Vaccinium microcarpon Ait.) in sand culture. New Phytologist 77, 63-72. Turner, A.P., 1994. The response of plants to heavy metals. In: Ross, S.M. (Ed.), Toxic Metals in Soil-plant Systems. John Wiley, Chi chester, pp. 151-187. Uhlig, C., Junttila, 0., 2001. Airborne heavy metal pollution and its effects on foliar elemental composition of Empetrum hermaphrodi tum and Vaccinium myrtillus in Sor-Varanger, northern Norway. Environmental Pollution 114, 461-469. Uhlig, C., Salemaa, M., Vanha-Majamaa, 1., Derome, J., 2001. Ele ment distribution in Empetrum nigrum microsites at heavy metal contaminated sites in Harjavalta, western Finland. Environmental Pollution 112, 435-442. Verkleij, J.A.C., Schat, H., 1989. Mechanisms of metal tolerance in higher plants. In: Shaw, J.A. (Ed.), Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, FL, pp. 179-193. Väisänen, S., 1986. Effects of air pollution by metal, chemical and fertilizer plants on forest vegetation at Kokkola, W Finland. Annales Botanici Fennici 23, 305-315. Weckx, J.E.J., Clijsters, H.M.M., 1996. Oxidative damage and defence mechanisms in primary leaves of Phaseolus vulgaris as a result of root assimilation of toxic amounts of copper. Physiologia Plan tarum 96, 506-512. Winterhalder, K., 1984. Environmental degradation and rehabilitation in the Sudbury area. Laurentian University Review XVI (2), 15-47. Zak, 8., 1976. Pure culture synthesis of bearberry mycorrhizae. Canadian Journal of Botany 54, 1297-1305. ISBN 951-40-1899-0 ISSN 0358-4283