Plant Population Dynamics and Conservation in Wooded Hay-Meadows – Effects of Intensified Management

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 282

LOTTA WALLIN

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 UPPSALA ISBN 978-91-554-6829-3 2007 urn:nbn:se:uu:diva-7743 ! " # $ %&&' ()&& * * * + , - . /.,

0 , %&&', + + 1 0 2" . 3 4** * 5 * " , # , %6%, (7 , , 5/89 :'62: 277!2$6%:2(,

- * 2 . . * , 5 . ** . 2 * . 2 . /. * ; 8 / , 5 * < = . > * .2 ., - ** * * . , * . , / * ? . @ A 7& * . * , . . . B * C . , . . , 4 * *2 * . C * . . . . * , - 2 * . . * * 2 , D 5 ** * * , 5 * * . . 5 * . . * * 2 . , 8 * * . , # * . , 5 * * ,

! 2 . + /

" # $ % % $ % $ & ' ()$ $ %*+,-./ $

E 0 %&&'

5//9 $7 2$% ! 5/89 :'62: 277!2$6%:2( ) ))) 2''!( < )== ,,= F G ) ))) 2''!(> Till Hamish

List of papers

This thesis is based on the following papers, which will be referred to in the text by their Roman numerals.

I. Wallin, L. & Svensson, B. M. Effects of intensified management of hay-meadows on population dynamics of Succisa pratensis. Manuscript.

II. Wallin, L. & Svensson, B. M. Intensified management affects population growth differently in the two perennial herbs Hypochoeris maculata and Polygala amarella. Manuscript.

III. Wallin, L., Svensson, B. M. & Lönn, M. Artificial dispersal as a restoration tool in meadows: sowing or planting? Submitted manuscript.

IV. Wallin, L. & Avery, H. R. D. A timesaving, accurate method for locating and re-locating plants in ecological field studies. Ecological Informatics. In Press.

Contents

Introduction...... 9 The importance of management...... 9 Population responses in perennial plants to changes in management of semi-natural grasslands ...... 10 Aims with the thesis ...... 11 Methods ...... 13 Study sites ...... 13 Study species...... 14 Management intensification ...... 15 Demographic study ...... 15 Matrix analyses and population dynamics...... 15 Artificial dispersal ...... 18 Ultra-sound positioning system...... 18 Results and Discussion ...... 20 Intensified management decreases biomass accumulation...... 20 Higher population growth rates following management intensification ..20 Stochastic simulations demonstrate larger population sizes ...... 22 Population structure can possibly reveal population viability...... 22 Artificial dispersal successful for establishment of meadow plants...... 22 Coordinate measuring device increases speed and accuracy of data gathering...... 23 Conclusions and Management Implications ...... 24 Acknowledgements...... 26 Växtpopulationers dynamik och naturvård i lövbärande ängar – effekter av ökad hävd...... 28 Den viktiga hävden...... 28 Ökad hävd ger mindre förna och biomassa ...... 29 Växtpopulationer reagerar på ökad hävdintensitet...... 30 Artificiell spridning...... 30 Enklare mätningar med hjälp av elektroniskt instrument...... 31 Literature cited...... 32

Introduction

Since the domestication of livestock, humans have relied on semi-natural grasslands to provide fodder. Since the climate cooled in northern Europe around 2000 years ago stores of fodder have been needed when the animals had to be kept inside over winter. The tradition of mowing for hay was therefore introduced (Ekstam et al. 1988). In areas around the southern Baltic Sea, e.g., on Gotland, the wooded hay-meadow was the most common form of fodder grassland up until the beginning of the 20th century (Figure 1, from Kukk & Kull 1997). At this point changes in agricultural practices led to a drastic decrease in area of unfertilised grasslands (Ekstam and Forshed 1996; Martinsson 1999; Ohlsson 2006). Traditional management of these meadows consists of four distinct components throughout the year: raking, mowing for hay, aftermath grazing and pollarding for leaf fodder (Ekstam et al. 1988; Anonymous 1992; Martinsson 1999).

Figure 1. Historical distribution of wooded hay-meadows in Europe. The shaded areas indicate the presence of hay-meadows in the beginning of the 20th century (from Kukk & Kull 1997).

The importance of management Because semi-natural grasslands can harbour a high biodiversity (Kull & Zobel 1991; Anonymous 1994; Pärtel & Zobel 1999), they are important not

9 only when considering a cultural and historical legacy, but also from a species conservation point of view (Poschlod & WallisDeVries 2002). Over the course of the 20th century, most of the traditional meadows in Western Europe have been transformed into heavily fertilised crop fields, turned into coniferous forest plantations or overgrown into broadleaf forests, and management efforts have decreased in many of the remaining meadows. Of the 1.2 million hectares of traditionally managed hay-meadows in Swe- den at the end of the 19th century, only about 3000 hectares remain (Anony- mous 1994; Martinsson 1999), and these are often small and isolated. The reason for the high species richness in meadows has been attributed to the constant disturbance and low nutrient levels caused by vegetation removal and grazing, making it impossible for strong competitors to out- compete small or light-demanding species (Kull & Zobel 1991; Zobel 1992; Berendse 1999). Consequently, the need for a management that can preserve the biodiversity of the few remaining meadows is crucial (Oostermeijer et al. 1994; Berlin et al. 2000; Bühler & Schmid 2001; Lennartsson & Oostermei- jer 2001; Wahlman & Milberg 2002; Jacquemyn et al. 2003; Brys et al. 2004). The change in management motivation from the historic fodder production to the present conservation efforts, coupled with the increase in aerial nitrogen deposition over northern Europe, suggest that the current management intensity may not be sufficient to maintain optimal habitats to support the long-term survival of meadow plant species. A particular concern is increasing nutrient levels, which favour competitive broad-leafed grasses and other nitrogen-demanding species, which in turn leads to increasing litter production (Berendse 1999). This accumulation of biomass and litter increases the difficulties for seeds to both germinate and establish, particularly so when aftermath grazing is abandoned (Bullock et al. 1994; Ekstam & Forshed 1997; Grime 2001; Franzén & Eriksson 2003; Rook et al. 2003).

Population responses in perennial plants to changes in management of semi-natural grasslands

Many studies have concluded that mowing is one of the most favourable management methods for preserving species-richness and viable plant populations in semi-natural grasslands (Tamm 1956; Kotiluoto 1998; Hansson & Fogelfors 2000; Lennartsson 2000; Jensen & Meyer 2001; Lennartsson & Oostermeijer 2001; Stammel et al. 2003), but additional grazing is also very important to keep nutrient levels down (Jacquemyn et al. 2003). As the number of meadows has decreased dramatically, the need for suitable man-

10 agement to preserve the species richness harboured in the few remaining meadows has become crucial. It is therefore important to know how any decrease in management intensity will affect plant populations with regard to survival, growth and reproduction. Some studies have found differences in population age structure in plant populations growing in different habitat types (Oostermeijer et al. 1994), with populations in habitats with more open vegetation and bare soil displaying higher proportions of seedlings in the population. Conversely, populations in more closed habitats consisted of a majority of old individuals, and have been termed senile. With increased isolation between meadows, populations also run a risk of never being able to naturally disperse propagules between populations (Pri- mack & Miao 1992; Soons & Heil 2002; Soons et al. 2005). Dispersal is important to maintain genetic diversity, enabling evolutionary responses to changing environmental conditions and decreasing the risks of inbreeding and genetic drift (Ellstrand and Elam 1993; Fischer and Matthies 1997; Edenhamn et al. 1999; Smulders et al. 2000; Eriksson & Ehrlén 2001) A demographic study of a population will give an overall representation of how the individuals in the population contribute to the long-term population growth rate (Ȝ) through differing levels of survival, growth and reproduction (Caswell 2001). Analytical tools can provide answers to questions concerning how the population is performing after, for example, a change in management and the demographic approach can reveal where the change will have greatest effect. Some life-cycle stages can be more susceptible to management changes, and since financial constraints limit what conservation effects can be realistically achieved, the ability to pinpoint the parts of the population that respond and contribute most to a restoration effort can prove to be valuable (Caswell et al. 1998).

Aims with the thesis My major aims with this thesis were to investigate whether it was possible to improve the conditions for vascular plant species living in wooded hay- meadows, and to develop methods for facilitating dispersal between meadows. I also wanted to evaluate two different approaches when doing a viability analysis, namely a crude but quick stage distribution analysis and a more time-consuming demographic analysis, involving both prospective and retrospective analyses as well as deterministic and stochastic modelling. Demo- graphic modelling is an important tool for investigating the response of plant populations to environmental processes, whether natural or experimentally induced. Using a demographic model, one relies on gathering data on individual plant performance over several years. In an attempt to increase the speed and accuracy of this laborious work, I wanted to develop a simple and

11 reliable method of tracking individuals between field seasons without using permanent markers. I was particularly interested in investigating the effects intensified raking and mowing would have on litter accumulation and biomass production in seemingly well-managed hay-meadows (Paper I). I was also interested in how population growth rates (O) would be affected by intensified raking and mowing (Papers I and II). If the plant populations respond to intensified management, I wanted to know which demographic transitions contributed most to the differences in population growth rate (Papers I and II). Because habitat isolation and limited dispersal are problems for many meadow species, I examined if the dispersal process could be assisted artificially by sowing seeds and planting plug-plants (Paper III). Finally, to increase the speed and accuracy of demographic data gathering, I partook in the development of an electronic coordinate measurement device (Paper IV).

12 Methods

Study sites Five wooded hay-meadows (Bölske, Fardhem, Kullands, Linde and Mull- valds) situated on the southern part of the Baltic island of Gotland, Sweden, were used in the different studies (Figure 2). The meadows are situated between 3 and 55 kilometres apart and are all actively managed meadows. Bölske, Kullands and Mullvalds have a similar management history, with all having a long management history, similar moisture levels and high species richness. Fardhem, which is a newly restored meadow with comparatively lower species richness, was also included in the study since I wanted to test growth and survival in a re-introduction study. Linde was used as seed donor of Hypochoeris maculata seeds.

3 2 4

1. Mullvalds 5 2. Linde 3. Kullands 4. Fardhem 5. Bölske

10 km

Figure 2. Study sites on the southern part of the Baltic island of Gotland, east of mainland Sweden.

13 Study species The study species (Figure 3) were chosen based on the criteria that they should be herbaceous, perennial, non-threatened meadow plants that signal continuous management and must also be easy to identify, even as seedlings. Hypochoeris maculata L. (Asteraceae) has a basal rosette with leaves that usually press down the vegetation around it. It is insect-pollinated, hermaphroditic and strictly out-crossing (Wells 1976). The fruits of H. maculata are wind-dispersed with a pappus. It can be clonal and form several rosettes originating from one individual. It is fairly common in southern and middle Sweden on dry, open sandy soils and a characteristic species in managed semi-natural grasslands (Hultén & Fries 1986; Mossberg & Stenberg 2003). Polygala amarella Cr. (Polygalaceae) is a low-growing herb that mainly occurs on calcareous grasslands. The plant has several leafy stems branching out from a rosette of basal leaves, with a tender stem below attaching it to the ground. It is insect-pollinated, hermaphroditic and self-compatible. The seeds are ant dispersed (Mossberg & Stenberg 2003). Succisa pratensis Moench. (Dipsacaceae) has a rosette of leaves in oppo- site pairs, all merged at the base of the plant. It is insect-pollinated, hermaphroditic and can produce self-seed, but the self-seed has been shown to be less viable than out-crossed seeds (Adams 1955, Vergeer et al. 2003). It has been found to be slightly clonal, as it is able to sometimes produce side- rosettes attached for the first years to the main plant. The seeds are borne in a hairy capsule, which can attach to furry animals for dispersal (Adams 1955; Römermann et al. 2005), but the hairy fruits have also been classified as wind-dispersed, albeit for rather short distances (Soons & Heil 2002). Succisa pratensis is rather common throughout southern Sweden on nutrient- poor soils and occurs in meadows and other semi-natural grasslands (Hultén & Fries 1986). However, it has been found to be in decline in meadows on Gotland in the last decade (Anonymous 2004).

Figure 3. The study species, from left, Hypochoeris maculata, Polygala amarella and Succisa pratensis. (Photo: Lotta Wallin)

14 Management intensification The effects of different management intensities on the populations of Hypo- choeris maculata, Polygala amarella and Succisa pratensis were investigated during four growing seasons, from 2003 to 2006. In 2002, I positioned 2–5 transects, with permanent plots measuring 0.71 × 0.71 m (0.5 m2), with a 0.29 m border zone between plots, in open areas in Bölske, Kullands and Mullvalds meadows. To these, four treatments were applied in a factorial design, with the treatments randomly ordered adjacent to each other within four-plot blocks along each transect, resulting in 12–14 replicate blocks per meadow. The treatments were: extra raking (R), extra mowing (M), extra raking and mowing (RM) and no extra treatment (control, C). The treatments were applied in addition to the usual meadow management and each individual plot received the same treatment over the entire study period. The intensified management treatments were applied annually in May and October, starting May 2003. The raking treatment consisted of thoroughly raking the plot with a steel garden rake in spring after the normal raking, with the litter collected, dried at 70°C for 48 hours and weighed. The extra mowing treatment in October consisted of cutting the vegetation in the plot with grass shears to a maximum vegetation height of 5 cm. The cut material was collected, dried and weighed as for the raking treatment. Because the distance between control plots and treatment plots within four-plot replicate blocks could be up to three metres, I evaluated the effect of management in the final year of the experiment by raking and mowing an additional control plot perpendicularly bordering each treatment plot. This enabled a pair-wise comparison of treatment effects (for clarity, these additional control plots will be referred to as the ‘2006 control plots’ to differen- tiate them from the control plots within the transects). I measured average vegetation height in the plots using a vegetation ruler (Ekstam & Forshed 1996) each year in July before mowing.

Demographic study

Matrix analyses and population dynamics In the first season of the study, all study plots were censused for all individuals (seedlings as well as mature plants) of Succisa pratensis, Hypochoeris maculata and Polygala amarella. In plots containing more than 15 adult individuals of Succisa pratensis, a random sample of 15 individuals was selected. For each individual several parameters were recorded, from which stage classifications could be derived. The parameters measured yearly on

15 Succisa pratensis were the number of leaves, the width of the longest leaf, the number of fruits and the width of each fruit. For Hypochoeris maculata I measured the width and length of the longest leaf in June, and counted the number of fruits in July. For Polygala amarella I counted the number of stems and the length of the fruit racemes. Each year in May (after applying the raking treatment) all plants were censused for survival and all new seedlings were recorded and added to the dataset. Individuals were classified into the following stages for Hypochoeris maculata and Succisa pratensis: seedlings, juveniles, small vegetative, large vegetative, small flowering and large flowering plants (Figure 4a). For Poly- gala amarella the individuals were similarly classified except for the ab- sence of the juvenile stage (Figure 4b).

Seedling

a21 a22 Juvenile

a32 a42

a33 a a 44 Small 43 Large vegetative a 34 vegetative

a 63 a45 a36 a54 a a a 35 53 46 a64

a15 a16 a Small 65 Large flowering a flowering a 56 a a) 55 66 Seedling

a21 a31 a41

a Small 32 Large a22 vegetative vegetative a33 a23

a a52 34 a 25 a43

a a24 42 a35 a53

a14 a15 a Small 54 Large flowering flowering a45 a a b) 44 55 Figure 4. Life cycle graphs of Hypochoeris maculata and Succisa pratensis (a) and Polygala amarella (b). Arrows denote possible transitions, aij, where a denote transition probabilities or fecundity and i is the stage in year t+1 and j is the stage in year t.

16 I analysed the demographic data using population matrix modelling (Caswell 2001, Morris & Doak 2002). Using the fate of the individuals from each different stage (where transition probabilities were calculated by combining the underlying vital rates growth and survival) between each year, I constructed transition matrices for the different treatments, pooling the years. I investigated the total effect of treatments on population growth rate (Ȝ) by comparing each of the three treatments Ȝ’s with the control Ȝ. I calculated 95% bias-corrected, confidence intervals using 10 000 bootstrapped projec- tion matrices (Caswell 2001) which were generated by randomly sampling individuals with replacement within stage classes. I used a non-parametric randomisation test (Levin et al. 1996; Caswell 2001) to assess the significance of difference in Ȝ between treatments. I compared the treated Ȝ with the control Ȝ pair-wise, by calculating the difference between them (șobs):

(treatment ) (control ) T obs O O

A test statistic (ștest) was then calculated from the fates of the individuals in the two treatments being compared by randomly permuting individuals from the combined data set with their history, maintaining the sample sizes. From the permuted data set, the difference between the randomly calculated Ȝ’s formed the test statistic (ștest). A total of 2000 permutations were performed. To account for multiple testing, I Bonferroni-adjusted the level of significance (Į) (Quinn & Keough 2002). I used a life-table response experiment (LTRE, fixed design, Caswell 2001) to find the transitions that contributed most to the differences in Ȝ between the control and treatment matrices. I used the control matrix as a baseline and compared the transition values pair-wise with the three treatment matrices. I decomposed the effect into contributions from each transition value (cij):

(control) (treatment) cij >aij aij @ wO waij Am where Am is the mean matrix between the control and the three treatment matrices, respectively (Caswell 1996, 2001). The demographic analyses were done in Matlab ver. 5.2 and 6.3 (MathWorks, South Natick, Massachu- setts) using adjusted scripts from Caswell (2001) and Morris & Doak (2002). I used stochastic simulations on four transition matrices from the different management intensities (2003–2006) in three meadows for Succisa pratensis to get long-term stochastic Ȝ and the population sizes projected 50 years into the future. All matrices were selected with equal probability (Morris and Doak 2002).

17 Artificial dispersal The effectiveness of artificial dispersal of Succisa pratensis and Hypochoeris maculata using seed sowing or plug-planting was examined by comparing the establishment success of the two introduction methods after two seasons. The seeds and plug-plants were introduced into three traditionally managed hay-meadows (Bölske, Kullands and Mullvalds). In Kullands both species were already present, in Bölske only S. pratensis occurred, and in Fard- hem neither of the species initially occurred. Seeds of both species from three different donor sites (Kullands, Mullvalds and Linde (H. maculata, or Bölske (S. pratensis) were germinated in the greenhouse at Uppsala Univer- sity Botanical Garden, and planted as plug-plants in May 2004. In each of the three recipient sites, eight 0.5 m2 plots were established along transects. Each permanent plot was subdivided into 16 subplots (180 × 180 mm), which, in turn, were divided into two quadrats that each received 50 seeds of each species (October 2003) and two plug-plants (May 2004) of the two species from all donor sites. The average number of emerging seedlings from the two sub-plots was used in the later analyses of donor site importance. For the establishment comparison between plug-planting and seed- sowing, I judged a plot as successfully established if there was at least one surviving plug-plant or seedling after two growing seasons. Before sowing, I randomly chose half of the 0.5 m2 plots and thoroughly raked the plots with a steel rake to investigate if removal of litter would improve establishment and survival of seedlings. For the sown seeds, I counted emerging seedlings in May 2004 (emer- gence in the first growing season) and the number still alive in October 2005 (survival after two growing seasons). I recorded the survival of the planted plug-plants after 18 months, i.e., in October 2005 (survival after two growing seasons).

Ultra-sound positioning system The demographic study component of my research required that I tracked several thousand individuals over five years. As the meadows are managed using raking, mowing and aftermath grazing, it was not possible to perma- nently tag each individual, so I needed to find each one at the beginning of each field season. Initially I used a tape measure to record the location of the individuals, but this proved to be error prone and time-consuming and a more accurate and speedy technique was required. As it was not possible to find any low-cost readily-available solutions for rapidly recording large numbers of coordinates in small plots, I partook in the development of a custom-made electronic pantograph that could accu-

18 rately measure the location of a measurement stylus within a defined bound- ary by using ultra-sound travel times. To assess the impact on measurement error in the dataset, I compared the number of errors in the data in the first field season (when the tape measure method was used) with the within-season errors when using only the electronic pantograph. Only errors exceeding 2 cm were considered because this margin was considered to be sufficiently accurate to identify the individuals, given the movement of the soil during winter and that the movements of the plants themselves were estimated to be of this magnitude. The accuracy of the instrument was determined by recording the location of the stylus when placed within a 0.5 m2 plot, using a grid with 10 u 10 known points. The test recordings were performed in a meadow with an average vegetation height of 5 cm. Two operators took four measurements at each grid point to give a total of 400 measurements, from which the devia- tion from the true coordinates could be determined.

19 Results and Discussion

Intensified management decreases biomass accumulation Dry weights of raked litter and mowed biomass were reduced in treated plots compared to the 2006 control plots after four years of intensified management in two of three meadows. This is a sign of biomass accumulation in these meadows, supporting my concern that present-day management is in- sufficient. The reason for this accumulation may be a combination of the reduction in management intensity and the increased nutrient accumulation from aerial nitrogen deposition (Foster & Gross 1998; Berendse 1999). Re- gardless of the underlying reason, this has serious implications for seedling germination and establishment (Bullock et al. 1994; Jensen & Meyer 2001). Most of the management work is presently performed by either local in- terest groups or volunteers, and represents heavy work for a relatively small economical compensation. Therefore, increased financial incentives may have to be considered in conservation budgets from local councils and the European Union to ensure that the required increase in management intensity levels is achieved.

Higher population growth rates following management intensification Population growth rates in Succisa pratensis differed between the treated and the control plots all three meadows. The permutation tests revealed that both the mowing and the combined raking and mowing treatments resulted in higher population growth rates compared to the control plots. Increased management has been shown to increase population growth rates and persis- tence in other long-lived perennial species as well (Lennartsson & Ooster- meijer 2001; Jacquemyn et al. 2003; Brys et al. 2005; Ehrlén et al. 2005; Colling & Matthies 2006). Polygala amarella displayed a population growth rate that varied consid- erably among treatments and was below unity for all treatments and control

20 plots. P. amarella had an especially low growth rate for the treatments where raking was included, which can be explained by the plant’s morphology: older individuals have a branched growth form with a basal leaf rosette, below which only a thin stem connects them to the ground, making them vulnerable to being snagged by the rake and pulled out, a phenomenon I observed in the field. Hypochoeris maculata had a population growth rate of around unity, with very little variation among treatments and showed no response to an increase in management. The rosette of the plant is probably able to press down the surrounding vegetation to create an open space to capture sufficient light, Increasing the level of management will therefore not substantially improve the competitive ability of this species. Seedlings may, however, be favoured by increased management intensity, but reproduction was rarely observed in the field during my study. Other studies on perennial plant species living in semi-natural grasslands have generally shown a positive effect of higher management intensity (Bühler & Schmid 2001; Ehrlén et al. 2005). How- ever, Jongejans & De Kroon (2005) found that three different species responded differently to the same environmental factors even though they share similar life-history traits. Therefore one has to be cautious in conclud- ing general responses to, for example, management, drawn from results from just one species, as I have shown by the different responses to an increase in management in my studied species. Since LTRE analysis is a retrospective method of decomposing the effects of a treatment compared to a control, it cannot be used to forecast the effects of specific management changes (Caswell 1996, 2000). However, LTRE it is very useful for elucidating the reasons for any differences found. I conclude that the reason for differences in population growth rate between treatment and control plots for Succisa pratensis in Bölske was that there were higher proportions of juveniles growing into small vegetative plants, as well as a higher proportion of large vegetative plants growing to large flowering plants in the treated plots. The increased management intensity may have made these transitions easier by reducing competition from the surrounding vegetation. The sensitivities of the fecundity values in the matrices were generally not large in any of the meadows or treatments. In- stead, I found most of the high sensitivity values for transitions where adult individuals stay in the same stage or grow. There is a slight difference between meadows though, with Bölske having higher sensitivities in the older stages and Kullands and Mullvalds having higher sensitivities in the younger stages. In essence this follows Silvertown et al. (1993) and Crone (2001) who suggested that proportional changes in survival in adult stages are more important than fecundity for population fitness in perennial species. For Polygala amarella the LTRE analysis revealed that raking mainly affected transitions of juveniles to older stages negatively, again supporting that this species will not be favoured by intensified raking.

21 Stochastic simulations demonstrate larger population sizes The stochastic modelling showed positive effects on Succisa pratensis of an increase in management; all meadows displayed larger projected population sizes 50 years into the future with intensified management compared with the control. Although none of the populations in Bölske, including the control, displayed a population growth rate under unity, there was a large difference between the control and the raking plus mowing treatment, showing that population sizes clearly increase with intensified management.

Population structure can possibly reveal population viability Long-term population growth rates were around unity in all populations of Succisa pratensis, even though the population structures were quite different in Bölske, Kullands and Mullvalds, with Bölske having a high proportion of old individuals and few seedlings, and the other two sites having a higher proportion of seedlings. The negligent difference in population growth rate indicates that it is possible for a population to have a stage distribution with a majority of old individuals and yet not shrink. Perhaps we can anticipate a future decline of a population with a skewed distribution. Therefore, it seems wise to judge population viability by combining population growth rate es- timates with observations of the population structure (Oostermeijer et al. 1994; Eriksson 1996). Also, since the stochastic simulation revealed larger population sizes following intensified management, it is likely that there is a negative trend not displayed by the deterministic model.

Artificial dispersal successful for establishment of meadow plants Establishment was twice as successful for planted plug-plants compared to sown seeds. The low germination rate and subsequent survival after one and two growing seasons after sowing is a concern for the long-term survival of the introduced individuals. The seedlings that established from seed were still quite small after two seasons; I observed that the average leaf width was 5 mm for Succisa pratensis and 2 mm for Hypochoeris maculata. Their small sizes decrease the chances for their long-term survival, as they are still very vulnerable to competition. This is readily apparent in the mortality of more than half of the seedlings of both species between seasons one and two. Both seed sowing and plug-planting have been used in introduction experi-

22 ments (Walker et al. 2004) but successful long-term establishment of the introduction can be varying (Primack & Miao 1992; Turnbull et al. 2000). Seedling establishment was affected by donor site. Since the Succisa pratensis seeds from the different sources were differing in weight this may well be the overriding determinant for seedling germination. Heavier seeds, with more resources, generally have a greater germination ability compared to lighter seeds (Westoby et al. 1996). In my study, the lightest seeds of S. pratensis came from the donor site Bölske, and these seeds failed to germinate both in the greenhouse and in the field.

Coordinate measuring device increases speed and accuracy of data gathering The two measurement sets taken with the tape-measure system contained 293 measurements with coordinate differences exceeding 2 cm, from a total number of 2496 recordings (12%). The distribution of these errors was roughly even over the size of the plot. The same comparison was performed on two censuses measured with the electronic system, where there were 72 measurements with coordinate differences exceeding 2 cm from a total of 2927 recordings (2.5%). Of these errors, over half had a difference of less than 6 cm. The results clearly demonstrate the advantages of automating the measuring process. The errors remaining in the measurements reveal two important points: that the concentration of errors below 6 cm indicates that the between-season movement of the plants may be larger than estimated and that the remaining errors above 6 cm must be due to transcription errors; this provides an error baseline that can be subtracted from the tape-measure results to determine the errors directly arising from the manual measurement process. The precision of the measurement device was ±1.8 cm, which is accept- able in the study system. Even though seedlings occasionally grew in tightly clustered groups, the device was still effective because previous measurements of the relative positions of individuals within a seedling cluster proved sufficient for tracking the same individuals in subsequent seasons. For demographic studies that use large numbers of tracked individuals, the automation of coordinate measurement can result in an improvement to data reliability and a substantial reduction in the time needed for finding individuals in the field.

23 Conclusions and Management Implications

I have shown that intensified management (extra raking and/or extra mowing) can reduce the amount of litter and biomass even in meadows that are considered well-managed. Therefore, an increase in the present management intensity level may be necessary to maintain the habitat characteristics that are the basis for preserving long-term viable populations of meadow plants in the future. In the meadows that showed litter and biomass accumulation I found that intensified management had a positive effect on Succisa pratensis, with a combination of raking and mowing increasing the long-term population growth rate compared to the control plots that had just the normal management. For Hypochoeris maculata, there was no response to management in population growth rate, but this species was only studied in a meadow in which the levels of biomass accumulation were not affected by an increase in management intensity. Polygala amarella responded negatively in the same meadow to any management that involved extra raking due to the structure of the plant – too vigorous raking resulted in pulling the plants out of the ground. These results suggest that species with differing life-history characteristics and morphologies will react individually to management intensification. I nevertheless advise that if the amount of accumulated litter and vegetation biomass in a meadow is moderate, a more intense management will be needed, even though some species may initially react negatively. Since Po- lygala amarella displayed very dynamic population changes, with a capabil- ity for fast growth, large disturbances may not be detrimental to its survival. Artificial dispersal between meadows can be achieved either by seed sowing or plug-planting. Successful establishment was accomplished by both methods using Hypochoeris maculata and Succisa pratensis as model species. After two seasons, plug-planting had a higher establishment degree than seed sowing, making the first method especially suitable if seed sources are scarce or if the recipient habitat is highly competitive with, for example, low light levels leading to high seedling mortality. For demographic studies where large amounts of individual data are gath- ered over several seasons from permanent study plots, more reliable location data and a speedier acquisition time can be attained by using an electronic coordinate measuring device.

24 Finally, I want to point to the danger of merely studying the present-day distribution of stages and using this for assessing population viabilities. I have shown that even if a species is common and has a population distribution that appears healthy (i.e., many young individuals, many flowering plants and a high seed production), we can find other populations of the same species with a distribution consisting of mainly of old individuals that displays the same population growth rate. To gain the full picture, more data are required, preferably collected over several years, and stochastic simulations seem to be the most reliable tool when projecting the fate of a population.

25 Acknowledgements

First and foremost I would like to thank my supervisor Brita Svensson, who has supported me throughout these years, mainly in my role as a PhD- student through her patience and pedagogic and scientific knowledge, but also by caring for me as a person – this has strengthened me both in my aca- demic role as well as in my own identity. My assistant supervisor Mikael Lönn has always been helpful, and the extremely quick response to any of my questions has been remarkable. Other people who have helped shape this thesis are the scientists and PhD-students at the department, who by discussions and comments on manuscripts and talks have enlightened me to what this work really is about – to learn more! Some that need special thanks due to their commitment in the finishing phase are Håkan Rydin, Bengt Carlsson and Jon Ågren, who helped with insightful comments and suggestions to various manuscripts. Without a strong administrator a work place easily falls apart, and thanks to the always-friendly Ulla Johansson, the bureaucratic parts of my time as a PhD-student have seemed like one of the easier parts. Stefan Björklund has helped by building seed-traps and transportation boxes for plug-plants. Financial support for the study has been kindly provided by Bjurzons, Ex- tensus, E. Hellgrens, Helge Ax:son Johnsons, P.O. Lundells, B. Lundmans and Tullbergs scholarships. My years at Växtbio have not only meant a probable increase of knowledge and certainly more years at the university, but also a number of new friends. My office mates in Norra Tornet (Tord, Helena, A-K, Sylvía och Claire) have with their presence made it more fun to go to work in the morn- ings. My VäxtBIO-friends have proved that also evenings can be enjoyable at work! All the discussions with people over lunch and fika have proven that it’s not important what you discuss, but how you discuss is! My field assistants, Karin Jakobsson, Therese Eriksson, Monia Lindeberg and Ellen Flygare have been valuable co-workers in the field, and good friends when we did not work – thank you! I would like to thank the people that are managing the last remaining hay- meadows today; their work preserves a part of our cultural history as well as conserving a species richness in our harsh country that is comparable to that of the rain forests. Stort tack till Anders och Ingrid Lingwall i Gerum, Tom- my Andersson och Helena Cedergren i Fardhem, Lars-Göran Söderström och Helén Eriksson i Grötlingbo och Lasse Olsson och de andra eldsjälarna i

26 Ardre, för att jag har fått arbeta i era vackra ängen. Våren kommer inte bli densamma utan att se vitsipporna i blom på Gotland! I spent all my field seasons on Gotland, and by finding a fantastic place to stay in (stuga 4) it became both comfortable and enjoyable to live on the island. Tack till Inga-Maj och Matts på Borum för ett husrum utöver det vanliga och många trevliga samtal! Mina föräldrar, Tyra och Lars, och min bror Petter och hans Carina med familj har funnits som starka grundpelare i min tillvaro. En lördagförmiddag är inte komplett utan ett Melodikryss, och det är viktigt att äta nyttig medel- havskost varje fredag! Mina bästa vänner Malin och Emma har sedan länge känt till allt som finns att veta om mig – det är skönt att ha vänner som er! Ett särskilt tack till Emma som ställt upp med sitt kunnande var gång jag behövt hjälp med layout och illustration. Mina mostrar och morbröder har lärt mig att det är roligt med sorgsna allsånger på finska, i alla fall om man sjunger dem tillsammans! And if a sad Finnish song wouldn’t help to cheer things up, my family in New Zealand always know how to make a good time, it has just been a bit hard to see you often enough! Without one person, life truly would be harder and darker, my beloved husband Hamish has been there for me every day of this time, in the good times, and possibly more importantly managing to get me out of the bad times faster than I would myself. Luckily I will never graduate from you! The future is wide open now – like it always was, but without a thesis lin- gering in the back of my head, the horizon looks a lot clearer!

För ska det va nån nytta med vår flora så får den faktiskt inte växa som den vill....

Allan Edwall

27 Växtpopulationers dynamik och naturvård i lövbärande ängar – effekter av ökad hävd

Den viktiga hävden Ängshävd har funnits i Sverige ända sedan järnåldern och skapade ett unikt landskap med en mycket hög rikedom på arter. På ängarna producerades det foder som gjorde att kreaturen kunde överleva vintern inomhus innan de fick komma ut i skogen igen på sommarbete. Vinterfodret var indirekt också väldigt viktigt för åkerbruket, då de stallade djuren producerade gödsel som spreds på åkrarna, därav uttrycket ”äng är åkers moder”. De öppna gräsmarker som en gång i tiden täckte stora delar av södra och mellersta Sverige, skapades av den årlig hävden som innefattade flera mo- ment. Dessa var tillsammans mycket viktiga för att maximalt med höfoder skulle produceras. På våren räfsades (fagades säger man på Gotland, där jag gjort mina undersökningar) ängen för att avlägsna grenar och gammalt fjol- årsgräs och mossa som annars hindrade gräsväxten. Denna del av hävden var utomordentligt viktig, och kunde ta flera veckor att genomföra. Man satte igång med arbetet då vitsippan (fagningsblomma eller fagningsros på got- ländska) stod i blom och med sitt bladverk lyfte upp fjolårsgräset. Efter fag- ningen lämnades ängen i fred fram till slåttern. På Gotland skedde slåttern normalt från mitten av juli till början av au- gusti, beroende på hur varm och torr sommaren varit. Höet slogs med lie och låg sedan kvar tills det var helt torrt. Dessutom vändes höet flera gånger under denna tid för att det torka jämnt och utan röta, men det innebar också att ängsväxternas frön föll ur sina fröställningar. Ett par veckor efter slåttern lät man ofta djuren komma in på efterbete, och ta tillvara den växtlighet som hunnit komma upp efter höskörden. Djurens närvaro var också mycket viktig för frögroningen: klövarnas tramp skapade luckor i grässvålen där fröna kunde gro och växa till. Sensommaren var också en period då lövtäkt, så kallad hamling eller –på Gotland – klappning kunde äga rum i lövbärande ängar. Eftersom behovet av vinterfoder ofta var stort drygade man ut det genom att skörda lövade grenar från framförallt alm och ask. Detta innebar att mer ljus nådde till markvege- tationen, men också att de trädrötter som dog och förmultnade till följd av att grenar togs bort, kunde ge en gödslingseffekt som var gynnsam för gräsväx-

28 ten. Under vintermånaderna städades det i ängarna, man plockade nedfallna grenar och gallrade i buskar och träd. Alla komponenter i hävdcykeln skapade tillsammans den näringsfattiga och ljusrika naturtyp där många arter kunde samsas på en begränsad yta. Den mänskliga hävden skapar alltså förutsättningar för många växtarter att sam- existera genom att snabbväxande och kvävande (det vill säga starkt konkurrenskraftiga) arter inte tillåts ta över. Änghävden såg i stort sett likadan ut från järnåldern fram till slutet på 1800-talet, då introduktionen av konstgödsel och kemiska bekämpningsme- del ledde till att man kunde få fram betydligt större fodermängder utan att behöva sköta sin äng med alla de arbetskrävande delarna. I och med denna förändring blev allt större arealer ängsmark omvandlad till åkermark, barr- skogsplantager eller bara övergivna och så småningom naturligt övervuxna av buskar och lövträd. Idag finns endast 3 promille kvar av den ängsareal som fanns i slutet av 1800-talet. Cirka 3000 hektar äng hävdas idag i Sverige på det traditionella viset. På Gotland finns ungefär 350 hektar.

Ökad hävd ger mindre förna och biomassa Ett ökat atmosfäriskt kvävenedfall över Nordeuropa och Sverige under de senaste femtio åren har lett till en höjning av näringsnivåerna i de naturligt så näringsfattiga ängsmarkerna. Detta kan leda till en ökad gräsväxt och en ökning av förekomsten av konkurrenskraftiga arter, såsom hundkäx – en av ängshävdare illa sedd art. Eftersom det idag inte finns någon ekonomisk avkastning för ängshö är den mesta av hävden bekostad av olika ekonomiska bidrag från EU och landets länsstyrelser. Den arbetsinsats som krävs för att genomföra alla delar av ängskötseln täcks ofta inte av bidragen. Kombina- tionen ökade näringsnivåer i ängarna och en svagare hävd tror jag kan leda till att dagens skötselnivå inte är tillräckligt intensiv för att bevara ängarnas höga artrikedom. Genom att under fyra säsonger experimentellt öka hävdintensiteten i äng- ar som är välhävdade med dagens mått mätt, kunde jag undersöka om det blev någon skillnad mot den naturliga hävden. Jag använde fasta provrutor, där jag antingen räfsade extra noga på våren, klippte ner vegetationen till lägre än 5 cm på hösten, både räfsade noga och klippte, eller lämnade utan förstärkt hävd, det vill säga dessa provrutor fick den hävd som resten av ängen fick. Jag undersökte hur mängden ihopräfsad förna på våren samt mängden vegetation klippt på hösten påverkades av den extra intensiva häv- den. Jag fann att förnamängden minskade i en av ängarna och vegetations- mängden minskade i två av tre undersökta ängar, vilket styrkte min misstan- ke om att hävden i dagens ängar kan vara för svag.

29 Växtpopulationer reagerar på ökad hävdintensitet Genom att märka upp individer av tre växtarter (slåtterfibbla, rosettjungfrulin och ängsvädd) i provrutor med olika hävdintensitet och sedan följa dem under fyra säsonger, kunde jag få ett svar på hur växtpopulationerna svarar på den förstärkta hävden. Jag använde mig av matrismodellering. En matrismodell bygger på att man följer ett antal individer över tiden, och be- räknar sannolikheter för överlevnad, tillväxt och reproduktion (så kallade demografiska data). I slutändan får man ett mått på om populationen växer, krymper eller är oförändrad till sin storlek. Man får också reda på vilka delar av växtens livscykel som kan tänkas svara positivt eller negativt på den för- stärkta hävden. Med detta analysverktyg kunde jag se att den experimentellt ökade häv- dintensiteten gav en högre populationstillväxt hos ängsvädd i de ängar som också visat en ökad ansamling av förna och vegetation. Det är alltså möjligt att förbättra situationen för ängsvädd genom att förbättra hävden. Slåtterfibb- la reagerade däremot inte alls på den ökade hävden, utan hade samma popu- lationstillväxt i alla provrutor, oavsett hävdintensitet. Rosettjungfrulin reagerade negativt på den förstärkta hävd som innebar extra räfsning, vilket jag tror berodde på att artens växtsätt medför att den lätt fastnar i räfsan och blir uppdragen. Den har nämligen endast en tunn stam som fäster den i marken, med en tämligen grenad blad- och grenrosett ovanför. De andra två arterna har mer robusta bladrosetter nära marken och undgår därmed åka med räf- san. Rosettjungfrulin kan å andra sidan tillväxa mycket snabbt under goda ljusförhållanden, och en ökad dödlighet till följd av en ökad hävd kan möjli- gen kompenseras av snabbare tillväxt från groddplanta till vuxen individ. Man har funnit att vuxna individer i vissa växtpopulationer av fleråriga arter kan överleva i en sämre miljö under långa perioder, vilket däremot inte deras fröplantor kan. Dessa populationer kan stå kvar länge, och kanske misstänker man inte att de kan vara på väg att sakta försvinna på grund av att det inte finns någon föryngring. Populationer med en tyngdpunkt på gamla individer har kallats senila, och jag ville undersöka om detta var fallet i de ängar som ingick i mina studier. I en av mina ängar fann jag att ängsvädd hade en populationsstruktur med mest gamla individer. Däremot innebar detta inte att populationstillväxten var minskande här, utan den var stabil.

Artificiell spridning I och med att ängarna idag är få och ofta ligger långt från varandra kan frö- spridning mellan lokaler vara svårt. Spridning är viktigt för att populationer- na ska kunna bibehålla en genetisk variation som är grunden för evolution och anpassning till förändrade miljöförhållanden. Eftersom många ängsväx- ter inte kan sprida sig längre sträckor, måste vi hjälpa dem för att populatio-

30 nerna inte ska isoleras. Jag testade att artificiellt sprida slåtterfibbla och ängsvädd med hjälp av två metoder, frösådd och plantering av småplantor. Båda sätten fungerade väl, men metoden med småplantor gav en högre eta- bleringsgrad två säsonger efter introduktionen. Troligtvis är detta en säkrare metod i och med att växterna har hunnit växa till sig i en konkurrensfri miljö (i mitt fall i växthus), medan de plantor som ska etablera sig från frö startar i en mycket hårdare miljö. Om man vill vara säker på att få etablering ska man alltså välja att introducera större plantor.

Enklare mätningar med hjälp av elektroniskt instrument Eftersom min demografiska undersökning krävde att jag följde samma individ under flera växtsäsonger, innebar det att jag behövde en metod som gjorde att jag enkelt och snabbt kunde hitta tillbaka till rätt individ. Jag kunde inte märka plantan (exempelvis med en färgad tråd), och heller inte lämna markeringar i jorden, eftersom jag var tvungen att avlägsna alla spår efter undersökningen innan slåttern och efterbetet. Genom att utveckla ett elektroniskt mätinstrument som med hjälp av ultraljud kunde mäta koordinater i mina fasta provrutor, kunde jag göra mätproceduren snabbare och samtidigt minimera antalet felmätta koordinater. Instrumentet kallar jag för ”plantas- tic”!’

31 Literature cited

Adams, A.W. (1955) Succisa pratensis Moench. Journal of Ecology 43:709-718. Anonymous (1992) Ängs- och hagmarker på Gotland. 1992. Länsstyrelsen Gotland, Visby. In Swedish. Anonymous (1994) Monitor 14. Biologisk mångfald i Sverige - En landsstudie. Naturvårdsverket, Solna, Sweden. In Swedish. Anonymous (2004) Återinventering av gotländska ängen. Rapport nr 3 från Länssty- relsens miljöenhet, Länsstyrelsen i Gotlands län. In Swedish. Berendse, F. (1999) Implications of increased litter production for plant biodiversity. Trends in Ecology and Evolution 14:4-5. Berlin, G.A.I., Linusson, A-C. & Olsson, E.G.A. (2000) Vegetation changes in semi-natural meadows with unchanged management in southern Sweden, 1965– 1990. Acta Oecologica 21:125-138. Brys, R., Jacquemyn, H., Endels, P., De Blust, G. & Hermy, H. (2004) The effects of grassland management on plant performance and demography in the perennial herb Primula veris. Journal of Applied Ecology 41:1080-1091 Brys, R., Jacquemyn, H., Endels, P., De Blust, G. & Hermy, M. (2005) Effect of habitat deterioration on population dynamics and extinction risks in a previously common perennial. Conservation Biology 19:1633-1643. Bühler, C. & Schmid, B. (2001) The influence of management regime and altitude on the population structure of Succisa pratensis: implications for vegetation monitoring. Journal of Applied Ecology, 38:689-698. Bullock, J.M., Clear Hill, B., Silvertown, J. & Sutton, M. (1994) Gap colonization as a source of grassland community change: effects of gap size and grazing on the rate and mode of colonization of by different species. Oikos 72:273-282. Caswell, H. (1996) Analysis of life table response experiments II. Alternative parameterizations for size- and stage-structured models. Ecological Modelling 88:73-82. Caswell, H., Brault, S., Read, A. & Smith, T. (1998) Harbor porpoise and fisheries: an uncertainty analysis of incidental mortality. Ecological Applications 8:1226- 1238. Caswell, H. (2000) Prospective and retrospective perturbation analyses: their roles in conservation biology. Ecology 81:619-627. Caswell, H. (2001) Matrix population models, 2 edn. Sinauer Associates, Inc., Sun- derland. Colling, G. & Matthies, D. (2006) Effects of habitat deterioration on population dynamics and extinction risk of an endangered, long-lived perennial herb (Scor- zonera humilis). Journal of Ecology 94:959-972. Crone, E.E. (2001) Is survivorship a better fitness surrogate than fecundity? Evolu- tion 55:2611-2614. Edenhamn, P., Ekendahl, A., Lönn, M. & Pamilo, P. (1999). Spridningsförmåga hos svenska växter och djur (Dispersal ability in Swedish plants and animals). Elanders Gotab. Naturvårdsverket, Stockholm. In Swedish.

32 Ehrlén, J., Syrjänen, K., Leimu, R., Garcias, M.B. & Lehtilä, K. (2005) Land use and population growth of Primula veris: an experimental demographic approach. Journal of Applied Ecology 42:317-326. Ekstam, U., Aronsson, M. & Forshed, N. (1988) Ängar: Om naturliga slåttermarker i odlingslandskapet. Lts förlag. Naturvårdsverket, Stockholm. In Swedish. Ekstam, U. & Forshed, N. (1996) Äldre fodermarker. Naturvårdsverket, Stockholm. In Swedish. Ekstam, U. & Forshed, N. (1997) Om hävden upphör – Kärlväxter som indikatorar- ter i ängs- och hagmarker. (If grassland management ceases – Vascular plants as indicator species in meadows and pastures.) Naturvårdsverket, Stockholm. In Swedish with English summary. Ellstrand, N.C. & Elam, D.R. (1993) Population genetic consequences of small population size: Implications for plant conservation. Annual Review of Ecology and Systematics 24:217-242. Eriksson, O. (1996) Regional dynamics of plants: a review of evidence for remnant, source-sink and metapopulations. Oikos 77:248-258. Eriksson, O. & Ehrlén, J. (2001) Landscape fragmentation and the viability of plant populations. In Silvertown, J. & Antonovics, J. (eds.) Integrating ecology and evolution in a spatial context, pp. 157-175. Blackwell, Oxford, UK. Fischer, M. & Matthies, D. (1997) Mating structure and inbreeding and outbreeding depression in the rare plant Gentianella germanica (Gentianaceae). American Journal of Botany 84:1685-1692. Foster, B.L. & Gross, K.L. (1998) Species richness in a successional grassland: Effects if nitrogen enrichment and plant litter. Ecology 79:2593-2602. Franzén, D. & Eriksson, O. 2003. Patch distribution and dispersal limitation of four plant species in Swedish semi-natural grasslands. Plant Ecology 166:217-225. Grime, J.P. (2001) Plant strategies, vegetation processes, and ecosystem properties. John Wiley & Sons Ltd, Chichester, England. Hansson, M. & Fogelfors, H. (2000) Management of a semi-natural grassland; results from a 15-year-old experiment in southern Sweden. Journal of Vegetation Science 11:31-38. Hultén E. & Fries, M. (1986) Atlas of North European vascular plants: north of the Tropic of Cancer 2. Koeltz Scientific Books, Königstein. Jacquemyn, H., Brys, R. & Hermy, M. (2003) Short-term effects of different management regimes on the response of calcareous grassland vegetation to increased nitrogen. Biological Conservation 111:137-147. Jensen, K. & Meyer, C. (2001) Effects of light competition and litter on the performance of Viola palustris and on species composition and diversity of an abandoned fen meadow. Plant Ecology 155:169-181. Jongejans, E. & De Kroon, H. (2005) Space versus time variation in the population dynamics of three co-occurring perennial herbs. Journal of Ecology 93:681-692. Kotiluoto, R. (1998) Vegetation changes in restored seminatural meadows in the Turku Archipelago of SW Finland. Plant Ecology 136:53-67. Kukk, T. & Kull, K. (1997) Puisniidud (Wooded Meadows). Estonia Maritima 2:138-146. In Estonian with English summary. Kull, K. & Zobel, M. (1991) High species richness in an Estonian wooded meadow. Journal of Vegetation Science 2:715-718. Lennartsson, T. (2000) Management and population viability of the pasture plant Gentianella campestris: the role of interactions between habitat factors. Eco- logical Bulletins 48:111-121.

33 Lennartsson, T. & Oostermeijer, J.G.B. (2001) Demographic variation and population viability in Gentianella campestris: effects of grassland management and environmental stochasticity. Journal of Ecology 89:451-463. Levin, L., Caswell, H., Bridges, T., DiBacco, C., Cabrera, D. & Plaia, G. (1996) Demographic responses of estuarine polychaetes to pollutants: life table response experiments. Ecological Applications 6:1295-1313. Martinsson, M. (1999) Böisårkar u daldargras – Naturvärden och vård i gotländska odlingslandskap. Länsstyrelsen i Gotlands län. In Swedish. Morris, W.F. & Doak, D.F. (2002) Quantitative Conservation Biology: Theory and practice of population viability analysis, 1 edn. Sinauer Associates Inc., Sunder- land. Mossberg, B. & Stenberg, L. (2003) Den nya nordiska floran. Wahlström & Wid- strand, Stockholm. Ohlsson, E.W. (2006) Det gotländska änget (The wooded hay-meadow on Gotland). Fälth & Hässler, Värnamo. In Swedish. Oostermeijer, J.G.B., Van't Veer, R., & Den Nijs, J.C.M. (1994) Population structure of the rare, long-lived perennial Gentiana pneumonanthe in relation to vegetation and management in the Netherlands. Journal of Applied Ecology 31:428-438. Poschlod, P. & WallisDeVries, M.F. (2002) The historical and socioeconomic per- spective of calcareous grasslands - lessons from the distant and recent past. Bio- logical Conservation 104:361-376. Primack, R. B. & Miao, S. L. (1992) Dispersal can limit local plant distribution. Conservation Biology 6:513-519. Pärtel, M. & Zobel, M. (1999) Small-scale plant species richness in calcareous grasslands determined by the species pool, community age and shoot density. Ecography 22:153-159. Quinn, G.P. & Keough, M. J. (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge, UK. Rook, A. J., Dumont, B., Isselstein, J., Osoro, K., WallisDeVries, M.F., Parente, G. & Mills, J. (2003) Matching type of livestock to desired biodiversity outcomes in pastures - a review. Biological Conservation 119:137-150. Römermann, C., Tackenberg, O. & Poschlod, P. (2005) How to predict attachment potential of seeds to sheep and cattle coat from simple morphological seed traits. Oikos 110:219-230. Silvertown, J., Franco, M., Pisanty, I., & Mendoza, A. (1993) Comparative plant demography - relative importance of life-cycle components to the finite rate of increase in woody and herbaceous perennials. Journal of Ecology 81:465-476. Smulders, M.J.M., van der Schoot, J., Geerts, R.H.E.M., Antonisse-de Jong, A.G., Korevaar, H., van der Werf, A., & Vosman, B. (2000) Genetic diversity and the reintroduction of meadow species. Plant Biology 2:447-454. Soons, M.B. & Heil, G.W. (2002) Reduced colonization capacity in fragmented populations of wind-dispersed grassland forbs. Journal of Ecology 90:1033- 1043. Soons, M.B., Messelink, J.H., Jongejans, E. & Heil, G.W. (2005) Habitat fragmentation reduces grassland connectivity for both short-distance and long-distance wind dispersed forbs. Journal of Ecology 93:1214-1225. Stammel, B., Kiehl, K. & Pfadenhauer, J. (2003) Alternative management in fens: response of vegetation to grazing and mowing. Applied Vegetation Science 6:245-254. Tamm, C.O. (1956) Composition of vegetation in grazed and mown sections of a former hay-meadow. Oikos 7:144-157.

34 Turnbull, L.A., Crawley, M.J. & Rees, M. (2000) Are plant populations seed- limited? A review of seed sowing experiments. Oikos 88:225-238. Vergeer, P., Rengelink, R., Copal, A. & Ouborg, N.J. (2003) The interacting effects of genetic variation, habitat quality and population size on performance of Suc- cisa pratensis. Journal of Ecology 91:18-26. Wahlman, H. & Milberg, P. (2002) Management of semi-natural grassland vegetation: evaluation of a long-term experiment in Southern Sweden. Ann. Bot. Fen- nici 39:159-166. Walker, K.J., Stevens, P.A., Stevens, D.P., Mountford, J.O., Manchester, S.J. & Pywell, R.F. (2004) The restoration and re-creation of species-rich lowland grassland on land formerly managed for intensive agriculture in the UK. Bio- logical Conservation 119:1-18. Wells, T.C.E. (1976) Hypochoeris maculata L. (Achyrosphorus maculatus (L.) Scop.). Journal of Ecology 64:757-774. Westoby, M., Leishman, M., Lord, J., Poorter, H. & Schoen, D.J. (1996) Compara- tive ecology of seed size and dispersal. Philosophical Transactions of the Royal Society B. London 351:1309-1318. Zobel, M. (1992) Plant species coexistence – the role of historical, evolutionary and ecological factors. Oikos 65:314-320.

35 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 282

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-7743 2007