Dispersal and persistence of an epiphytic

in a dynamic pasture-woodland landscape

Inauguraldissertation der Philosophisch-naturwissenschaftlichen Fakultät der Universität Bern

Vorgelegt von Silke Werth von Deutschland

Leiter der Arbeit: PD Dr. C. Scheidegger Eidgenössische Forschungsanstalt WSL, Birmensdorf ZH

Dispersal and persistence of an epiphytic lichen

in a dynamic pasture-woodland landscape

Inauguraldissertation

der Philosophisch-naturwissenschaftlichen Fakultät

der Universität Bern

Vorgelegt von

Silke Werth

von Deutschland

Leiter der Arbeit:

PD Dr. C. Scheidegger

Eidgenössische Forschungsanstalt WSL, Birmensdorf ZH

Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen.

Bern, 20. Juni 2005 Der Dekan

Prof. Dr. P. Messerli

Contents

Synthesis 1 Introduction 1 Conclusions 7 Hierarchy theory and a conceptual model of epiphyte dynamics 7 Biology of pulmonaria 9 Dispersal versus establishment limitation 11 Implications for conservation of L. pulmonaria 17 Chapter 1 23 Chapter 2 49 Chapter 3 71 Chapter 4 105 Chapter 5 137 Summary 149 Appendix 1 (Population genetic sampling) 153 Appendix 2 (Transplants) 165 Appendix 3 (Snow samples) 167 Acknowledgements 169 Curriculum vitae of Silke Werth 171

Synthesis

Introduction

Traditional sylvopastoral landscapes exhibit dynamics in the form of shifting mosaics of forested and open areas if disturbances are at intermediate levels (Olff et al. 1999). If grazing pressure is too high, wooded pastures develop into grasslands; in the absence of grazing, successional change leads to climax forests (Gillet et al. 2002). Wooded pastures would not be predicted as a favoured type of habitat for epiphytes requiring long ecological continuity of the forest canopy or for dispersal-limited epiphytes with long generation times, because shifting mosaic dynamics imply limited time for fulfilling life cycles, including dispersal to new patches of habitat. However, the epiphytic lichen , which is considered to be dispersal-limited (Walser et al. 2001; Walser 2004), and has a long generation time compared with other foliose (Scheidegger & Goward 2002), has a hotspot of its occurrence in Switzerland in a sylvopastoral landscape in and around the Parc Jurassien Vaudois (Zoller et al. 1999; Rychen 2002). L. pulmonaria is widely distributed in the nemoral and boreal zones of the northern hemisphere, with some additional occurrences in the southern hemisphere (Yoshimura 1971). This tripartite lichenised is associated with the Dictyochloropsis reticulata and of the genus (Geitler 1966; Jordan 1970). The generation time in lichens is defined as the average age at which a population of a lichen produces propagules, be it soredia, isidia, thallus fragments, or ascospores. Generation time in populations of L. pulmonaria in central Europe has been estimated to be 30 years or longer (Scheidegger & Goward 2002). In Europe and North America, L. pulmonaria is mainly associated with old-growth forests (Lesica et al. 1991; Rose 1992; Radies & Coxson 2004). Owing to air-pollution and intensive forest management, the species has faced a severe decline in central and northern Europe (Hallingbäck & Martinsson 1987), where it is red-listed in several countries (Wirth et al. 1996; Aptroot et al. 1999; Scheidegger et al. 2002; Søchting & Alstrup 2002). The population of the L. pulmonaria in the pasture-woodland landscape in the Parc Jurassien Vaudois is not only one of the largest in Switzerland, but also seems to exhibit genetic variation comparable in magnitude to that of populations in the Swiss

1 Synthesis

Alps and the Swiss Plateau (Zoller et al. 1999), despite the fact that the forests of the Parc Jurassien Vaudois are and were subjected to disturbances, mainly in the form of grazing pressure on juvenile trees, mainly uneven-aged forestry, some stand-level timber logging, and also stand-replacing disturbances (Vittoz 1998; Kalwij et al. in press). Genetic diversity of populations of L. pulmonaria would be expected to be low in this system for several reasons. Firstly, disturbances affecting carrier trees such as Acer pseudoplatanus resulting in reductions of L. pulmonaria population sizes may lead to instant loss of rare alleles (Nei et al. 1975). Secondly, new populations in disturbed areas which were founded by only few individuals may be subject to genetic drift, i.e., random changes in allele frequencies, leading to fixation or further loss of alleles even if overall mean allele frequencies remain constant among populations (Hartl & Clark 1997). Whether the spatial genetic signature differed among disturbance types in L. pulmonaria interested me greatly. Large population sizes of a putatively dispersal-limited organism with high genetic variation in a dynamic landscape represent an apparent contradiction, which is why I was particularly interested in studying patterns of genetic diversity in sites differing in disturbance history, as well as dispersal and establishment in L. pulmonaria. Dispersal and establishment are key processes in the life history of organisms (Clobert et al. 2001). The there has been a debate about the role of dispersal in structuring in the composition of local communities, namely if local species assemblages are a function of dispersal ability or of site conditions (Leibold et al. 2004; Ozinga et al. 2005a; Ozinga et al. 2005b). The relative importance of different environmental factors, such as regional- scale versus local factors in structuring communities can be determined by an investigation of communities along regional gradients, while simultaneously measuring site conditions (Ohmann & Spies 1998; Peterson & McCune 2001), and subjecting the resulting dataset to gradient analysis techniques such as constrained ordination (Økland 1990; Borcard et al. 1992). This study aimed (1) to determine if and how genetic structure and diversity of L. pulmonaria are influenced by disturbance type, (2) to partition spatial genetic structure to a clonal and a recombinant component, (3) to reveal to which degree the distribution of L. pulmonaria in the sylvopastoral landscape is limited by propagule availability versus establishment, (4) to assess the significance of a putative barrier to gene flow in the

2 Synthesis sylvopastoral landscape and to determine the spatial extent of populations in the studied landscape, (5) to determine the relative importance of local site conditions vs. regional gradients in shaping epiphytic lichen communities. These issues can be summarised in five main questions: (i) What is the effect of two types of stand-level disturbances on the genetic diversity and spatial genetic structure in L. pulmonaria? (ii) Are there spatially separate populations of L. pulmonaria in the study area, and do large open pastures represent effective barriers to gene flow in this species? (iii) Is the distribution of L. pulmonaria in a sylvopastoral landscape limited by dispersal or establishment, or by both? (iv) What is the spatial extent of clonal versus recombinant genetic structure? (v) How important are local environmental conditions versus large-scale, regional gradients in structuring epiphytic macrolichen communities?

The relationship between disturbance type and spatial genetic structure and genetic diversity is analysed in the first chapter. The second chapter determines how many populations are present in the study area, and assesses the effectiveness of a putative barrier, a large pasture, to restrict past gene flow in L. pulmonaria. In the third chapter, the role of dispersal versus establishment in limiting the distribution of L. pulmonaria in the study area is investigated. The spatial extent of overall genetic structure, as well as of its clonal and recombinant component are assessed in the fourth chapter. The importance of local environmental conditions versus large-scale environmental gradients for the species composition of epiphytic macrolichen communities is investigated in the fifth chapter. We furthermore investigated the implications of stand-replacing and stand-level disturbance under different levels of local versus global dispersal for persistence of L. pulmonaria using modelling approaches. The resulting manuscript is not included in this thesis. We discuss the results gained in the light of hierarchy theory. Novel and exciting aspects of the biology of L. pulmonaria are presented. The role of dispersal limitation vs. establishment limitation in structuring the

3 Synthesis local distribution of L. pulmonaria is reviewed, and implications for conservation of L. pulmonaria are discussed.

Influence of disturbance type on genetic structure and diversity

Old-forest associated lichens are commonly assumed to be negatively affected by tree logging or natural forest disturbances. However, we found that genetic diversity of Lobaria pulmonaria depends on the type of disturbance. Using 895 thalli collected from 41 plots of 1 ha (demes) corresponding to the disturbance categories stand-replacing disturbance, intensive logging and uneven-aged forestry, we determined fragment length at six mycobiont-specific microsatellite loci. There was evidence for multiple independent colonisations of demes located in areas affected by disturbance at forest stand level. Using spatial autocorrelation methods, we determined the spatial scale of similar genetic structure discriminating among the clonal and recombinant component of genetic variation and among disturbance type. Spatial autocorrelation of gene diversity was strong at distances up to 150 m in all three disturbance categories, with the strongest autocorrelation for demes affected by stand-replacing disturbance. The spatial autocorrelation was predominantly attributed to clonal dispersal of vegetative propagules. After accounting for the clonal component, we did not find significant spatial autocorrelation. This pattern may indicate low dispersal ranges of clonal propagules, but random dispersal of sexual ascospores, indicating that there was no dispersal limitation of ascospores. Genetic diversity was highest in demes affected by intensive logging at forest-stand level, and lowest in demes affected by stand-replacing disturbance. Our study exemplified that genetic diversity of some epiphytic lichens may not necessarily be impacted by stand-level disturbances for many centuries.

Spatial extent of populations and barriers to gene flow

As epiphytes as are strongly affected by the population dynamics of their carrier trees, large dispersal rates and corresponding high levels of gene flow may be needed for populations to persist at a landscape scale. Here we analysed landscape-level

4 Synthesis

population genetic data for L. pulmonaria and found indeed high amounts of past gene flow at the landscape level in microsatellite data from 895 thalli of L. pulmonaria genotyped at six fungal loci. Bayesian analysis revealed the presence of at least three spatially intermingled gene pools in the studied landscape. We also assessed the presence and effectiveness of potential barriers to gene flow and show that an a priori assumed barrier in the landscape, a large unforested area, did not significantly restrict past gene flow. Our results indicate that population genetic data at the landscape level may be relevant for our understanding of the organisation of genetic diversity and for population persistence of L. pulmonaria.

Dispersal versus establishment limitation

Dispersal is a critical population process for the dynamics and persistence of metapopulations, but is difficult to quantify. We analysed 240 DNA extracts derived from snow samples by a L. pulmonaria specific RealTime-PCR assay allowing for the discrimination among propagules originating from a single, isolated source tree, and originating from other locations. Samples which were detected as positives by RealTime- PCR were additionally genotyped for six L. pulmonaria microsatellite loci. Both molecular approaches demonstrated substantial dispersal from other than local sources. In a landscape approach, we analysed 240 snow samples with RealTime-PCR and detected propagules not only in forests where L. pulmonaria was present, but also in large unforested pasture areas and in forest patches where L. pulmonaria was not found. Monitoring of soredia of L. pulmonaria transplanted to maple bark after two vegetation periods showed high variance in growth among forest stands, but no significant differences among transplantation treatments. Hence, in all likelihood not the amount of propagules is what is limiting the old-forest lichen L. pulmonaria from occurring in younger forests in localities where large populations are present, but ecological constraints at the stand level may be. Further research is needed to identify the factors responsible for this.

5 Synthesis

Spatial extent of clonal vs. recombinant genetic structure

A geostatistical perspective on spatial genetic structure may explain methodological issues of quantifying spatial genetic structure and suggest new approaches to addressing them. We use a variogram approach to (i) derive a spatial partitioning of molecular variance, gene diversity, and genotypic diversity for microsatellite data under the infinite allele model (IAM) and the stepwise mutation model (SMM), (ii) develop a weighting of sampling units to reflect ploidy levels or multiple sampling of genets, and (iii) show how variograms summarize the spatial genetic structure within a population under isolation-by-distance. The methods are illustrated with data from a population of the epiphytic lichen Lobaria pulmonaria, using six microsatellite markers, and illustrated the spatial extent of the clonal and recombinant component of genetic variation. The clonal component was large in the studied population. Variogram-based analysis not only avoids bias due to the underestimation of population variance in the presence of spatial autocorrelation, but also provides estimates of population genetic diversity and the degree and extent of spatial genetic structure accounting for autocorrelation.

Regional versus local factors shaping epiphytic macrolichen communities

To investigate the relative influences of human impact, macroclimate, geographic location and habitat related environmental differences on species composition of boreal epiphytic macrolichen communities, lists of epiphytic macrolichen species from deciduous forests recorded in Troms county in northern Norway were subjected to constrained ordination. By Canonical Correspondence Analysis with variance partitioning, the relative amounts of variance in macrolichen species composition attributable to site factors (human impact, environmental differences among sites), regional factors (macroclimate), and the spatial component were quantified. Regional factors were most important in determining epiphytic macrolichen communities, but also site conditions such as forest stand properties contributed significantly to macrolichen community composition. Epiphytic macrolichen communities were assembled along a macroclimatic gradient from the coastline to the interior of central northern Norway.

6 Synthesis

Conclusions

Hierarchy theory and a conceptual model of epiphyte dynamics

Landscape ecological research has pinpointed the importance of temporal and spatial scales in ecology (Wiens 1989; Keitt et al. 1997; Bunnell & Huggard 1999; Li & Wu 2004; Wu 2004). Hierarchy theory assumes that processes operating at large spatial scales, the so-called “forcing functions”, shape processes at smaller scales. In the same way, processes at fine spatial scales influence processes at the next lower scale (Allen & Starr 1982; O'Neill et al. 1986; Allen & Hoekstra 1990; Muller 1992). From a hierarchy theory point of view, one may view topography, climatic conditions, the natural disturbance regime, landscape management and the level of acidic precipitation in the study area as the large-scale forcing factors shaping the dynamics of tree populations (Fig. 0-1). The spatial configuration and dynamics of the carrier tree population crucially influences epiphyte populations (Snäll et al. 2005a; Snäll et al. 2005b; Wagner et al. in prep.). Ultimately, processes that influence the carrier tree population, or habitat properties of carrier trees, determine the population dynamics of epiphyte communities, and thus of single epiphyte species. Species are assembled in communities along ecological gradients (Legendre & Legendre 1998). Macroclimatic gradients drive tree species composition on a regional scale, whereas topographical gradients are decisive at the forest stand level (Ohmann & Spies 1998; Bugmann & Solomon 2000). However, our results from epiphytic macrolichen communities in north Norway show that this is not only true for tree species composition. Macroclimatic gradients were found to be most important in structuring community composition of epiphytic macrolichens, relative to environmental factors at the level of forest stands, or spatial structure. Under different climatic conditions, other tree species than Picea abies, which is not an important carrier tree of L. pulmonaria in the study area, might have dominated the landscape. Had the forest in the study area been composed predominantly by a late successional tree species which provided more suitable habitat for L. pulmonaria, the effect of stand-replacing disturbances on the spatial occurrence and genetic diversity of L. pulmonaria might have been markedly different. In the latter scenario, the largest population sizes and highest genetic diversity of L. pulmonaria would be expected to

7 Synthesis

occur in old-growth forest stands. Further studies are needed to determine if this scenario is true.

Figure 0-1. Conceptual model of epiphyte population dynamics in central European temperate forests.

Microsite parameters like bark suitability or bark chemical and textural properties may influence the lower level processes in the hierarchy, e.g., the dynamics of epiphyte communities. The chemical properties of bark substrates may be influenced by wet acidic deposition, lowering the bark pH (Bates 1992; Gauslaa & Holien 1998). However, pollution influence is most likely constant across the study area. Macroclimatic conditions may affect the breadth of the ecological amplitude of epiphytic lichens. At their northern distribution limit in the northern Norway, some oceanic species like Lobaria pulmonaria, Lobaria hallii and Degelia plumbea occur only in the most continental sites in the interior, characterised by high summer temperature sums and comparatively low annual precipitation (see chapter five), and in the interior sites, they were predominantly occurring on calcareous schist rock substrates. At the northern outposts of their distribution, these species’ ecological amplitude was narrowed

8 Synthesis

to the climatically most favourable sites and to a favourable substrate which was rare in the study area in northern Norway. Some lichen species which are habitat specialists and confined to a narrow range of host trees in central Europe may show a wider ecological amplitude in a coastal oceanic climate, allowing them to colonise tree species which are not suitable habitat in most of central Europe, such as Picea abies. In boreal spruce forests of central Norway, an oceanic lichen species assemblage including L. pulmonaria frequently grows on P. abies branches (Haugan et al. 1995; Tønsberg et al. 1996; Holien 1998). Low bark pH may lead to a narrowing of the ecological amplitude of epiphytes, confining some epiphytes species to the most alkaline substrates (Gauslaa 1995; Wirth 1995; Wirth et al. 1996). This effect has been observed for L. pulmonaria, which used to grow on acidic substrates such as P. abies twigs in interior areas of Norway in the past, but has not been found on this substrate type for the last 60 years (Tønsberg et al. 1996).

Biology of Lobaria pulmonaria

The clonal component of genetic variation is very important in the sylvopastoral landscape in Switzerland, particularly on a local scale (chapter 1, chapter 4). However, the absence of significant linkage among the majority of loci showed that also recombination may be important in the genetic structure of the L. pulmonaria population from the Parc Jurassien Vaudois, a result which supports that L. pulmonaria is an outcrossing species (Walser et al. 2004). A further evidence of recombination is that fertile thalli of L. pulmonaria occur in the Parc Jurassien Vaudois, even though they are not frequent; during our field work, we found apothecia on a total of 15 out of 298 trees (5 %). However, the presence of four significant pairwise associations of loci points towards some non-random (i.e. assortative) mating in the total population. Also the analysis of population structure using assignment tests indicated assortative mating, as it revealed the presence of three gene pools in the study area, all of which occurred spatially intermingled. These results point towards heterothallism in L. pulmonaria, i.e. mating occurring only among compatible mating types. However, further studies, e.g. of spore multilocus genotypes, or of mating type genes, are needed for proving whether or not L. pulmonaria is a heterothallic fungus. In general, the conditions under which ascospores

9 Synthesis

are produced in L. pulmonaria are still very little understood and further research is needed, which could focus on a wide range of subjects including paternity analysis of ascospores, dispersal ranges of ascospores, and ecological conditions favouring the development of apothecia. In this study, in addition to the transplanted soredia, eight thallus fragments of thalli containing soredia, but not apothecia were transplanted onto each of the 55 trees which we transplanted soredia of L. pulmonaria to. When we revisited some of the sites three years after experiment establishment, some fragments had developed apothecia. What would be needed in the future is a pairwise transplantation study involving overlapping transplantation of fragments of the same microsatellite multilocus genotype onto unoccupied potential host trees at different distances, versus transplantation of very different multilocus genotypes. Long-term monitoring of the fragments could then determine under which conditions apothecia are developed, and if presence of another multilocus genotype is necessary for ascomata development in L. pulmonaria. Fifteen of the 895 analysed thalli had two alleles at one or more microsatellite loci. Most of these occurred on a single tree and were of the same multilocus genotype. Also a further five thalli collected from a single tree in the surroundings of the plot where snow samples had been collected (point source approach) showed two alleles at a microsatellite locus. There are three likely reasons for this pattern. (1) The occurrence of several alleles in the same thallus may simply represent a genome duplication affecting a microsatellite locus. A contradiction to this view is that in a few thalli, two alleles were found at two loci. This makes genome duplication less likely, unless the two microsatellite loci would be located next to each other on a chromosome. As the microsatellite loci were not significantly linked, they seemed to lie on different linkage groups, so that genome duplication is not a very likely explanation (Walser et al. 2004; Werth et al. in prep.). (2) Dikaryotic and diploid hyphae may have been extracted, which contained the allele of the “mother” thallus plus that of a “father”. What is strange, however, is that all thalli from two trees contained exactly the same alleles. (3) Another, more likely reason might be that the thalli with more than a single allele per locus do not contain a single mycobiont, but several mycobiont individuals, which may have spread clonally within the tree. A more detailed study is needed to be conclusive on this subject,

10 Synthesis

genotyping multiple extracts from the thalli where several alleles were found, perhaps in combination with genotyping of sterile cultured thallus hyphae. Together with information on heterothallism, this kind of study may provide very interesting insights into the biology of L. pulmonaria.

Dispersal versus establishment limitation

Recruitment in plant or lichen populations is the number of new individuals entering a given population per generation. Recruitment is a key process for population dynamics, to which both dispersal and establishment contribute (Begon et al. 1999; Nathan & Müller-Landau 2000). Dispersal has to be sufficient to provide diaspores at a given site within a population. In the following, limited availability of propagules at a particular habitat will be termed dispersal limitation (Nathan & Müller-Landau 2000). Establishment may be influenced not only by biotic factors like competition, but also by abiotic factors such as site conditions (Begon et al. 1999). Inability of diaspores to establish at a given habitat is in the following referred to as establishment limitation (Nathan & Müller-Landau 2000). Dispersal limitation and establishment limitation are not mutually exclusive, but may simultaneously reduce recruitment in a population. There has been an ongoing debate on whether local species pools are a function of ecological niches or of the dispersal ability of species (Leibold et al. 2004; Ozinga et al. 2005b). Also for epiphytic lichen communities, these concepts have been implicitly adopted in the form of the dispersal limitation hypothesis, stating that limited availability of propagules in young forests is the major reason why many epiphytic lichens are confined to old-growth forests (Sillett et al. 2000b; Hilmo 2002). According to an alternative hypothesis, the ecological gradient hypothesis, which corresponds to the niche assembly view of local communities, old-growth dependence in lichens is a function of favourable ecological conditions (Dettki et al. 2000). Both hypotheses are not necessarily mutually exclusive. It is however important to distinguish between the two hypotheses, because each one leads to a different conservation strategy. If the ecological gradient hypothesis applies, habitat quality of primeval forests would be equal to that of secondary forests once comparable microhabitats had developed. The ultimate conservation strategy

11 Synthesis would be to prolong the rotation cycle. Forestry practices would not matter directly, as long as forests had sufficient time for regrowing to mature or even old-growth, and for old-growth dependent lichens it would not make a difference whether a forest was clearcut or whether remnant trees were retained at final harvest. If the dispersal limitation hypothesis holds, on the other hand, it would be of primary importance to retain as many old-growth forests as diaspore source for old-growth associated species as possible, and instead of clearcutting whole forest stands, old trees should be retained at harvest to facilitate recolonisation (Sillett et al. 2000b). An important question is why L. pulmonaria is not found in many parts of the study area in Switzerland where seemingly suitable habitat is present. The distribution of L. pulmonaria could be either limited by propagule availability, or by establishment. One may argue that dispersal of every species on the planet is limited because the number of its propagules is finite, hence limiting distribution. Decreasing propagule density with distance from source is a common phenomenon (Ouborg et al. 1999; Cain et al. 2000; Nathan & Müller-Landau 2000; Clobert et al. 2001) and this alone does not allow the conclusion that dispersal limits the distribution of a given species. Furthermore, the term dispersal limitation is meaningless unless an explicit spatio-temporal context is given. Most transplantation experiments of propagules concluding that lichen species are dispersal limited have not defined what they mean with dispersal limitation. Studies involving transplantation of propagules do actually test establishment limitation, i.e. the ability of propagules to establish successfully at a given site (Nathan & Müller-Landau 2000). Hence, such studies allow limited conclusions on recruitment limitation in terms of dispersal limitation, i.e. the availability of propagules at a particular site (Nathan & Müller-Landau 2000). Molecular, population-genetic approaches as the variogram method employed in this study (chapters one and four) only allow inference of “realised” dispersal, i.e. dispersal events which were followed by successful establishment of propagules, and thus they integrate propagule limitation with establishment limitation (Nathan & Müller-Landau 2000). I propose the following definition for dispersal limitation: a given epiphytic lichen species is dispersal limited if it does not manage to colonise a forest stand with suitable habitat within three times its generation time. Three times the generation time is the time scale at which population decline of endangered

12 Synthesis

species is assessed for conservation purposes such as in the IUCN criteria for Red Lists of threatened species (IUCN 2001), and may therefore also be a relevant time scale for a definition of dispersal limitation. The generation time in L. pulmonaria is 30 years or more for vegetative propagules, and even probably longer for ascospores (Scheidegger & Goward 2002). Our model organism has been able to colonise an area affected by stand- replacing disturbance within 132 years after disturbance, and has already reached a rather large population size; the disturbed area has been colonised multiple times and has by now reached a comparatively high genetic diversity, given the strength of the disturbance event. This alone makes it unlikely that L. pulmonaria is dispersal limited in the study area in the Parc Jurassien Vaudois in terms of our rather wide definition. Assuming that the main carrier tree species of L. pulmonaria in the study area, Sycamore maple (Acer pseudoplatanus), needs to be about 50 years old to be suitable for colonisation by L. pulmonaria, this would leave 82 years or about three generations for L. pulmonaria to colonise the stand. L. pulmonaria has managed to successfully colonise the disturbed area within these three generations. Our data provide therefore evidence that L. pulmonaria is not dispersal limited. This is not very surprising. Most lichens have small, lightweight propagules, and thus one would expect many lichens to be efficient dispersers (Bailey 1976; Galloway & Aptroot 1995). Lichen colonisation of newly emerged volcanic islands three years after the last phase of eruption (Kristinsson 1972) and rapid colonisation of anthropogenic substrates by the lichen Lecanora dispersa (Bailey 1976) are only two examples of efficient dispersal and establishment in lichens. As compared to e.g. seeds of most forest trees, propagules of L. pulmonaria are lightweight and hence well suited for wind dispersal.

Also our results concerning the number of colonisation events to found demes disagree with dispersal limitation. There was evidence of multiple colonisations of demes affected by stand-replacing disturbances, and the minimum number of colonisation events was not related to disturbance level. Furthermore, ANOVA showed that there was no relationship between disturbance and genetic distance, suggesting that demes of L. pulmonaria were independently colonised several times from different sources after a disturbance event.

13 Synthesis

Throughout our study area, however, we found 1-ha plots which were not colonised by L. pulmonaria even though they contained maple trees which seemed to be suitable for colonisation (Kalwij et al. in press). An alternative definition of dispersa