Summary and Four Papers, Which Are Referred to by Their Roman Numerals: I

The impact of cormorant nesting colonies on plants and arthropods

Gundula S. Kolb

ISSN 978-91-7447-147-2

Printed in Sweden by Universitetsservice, US-AB, Stockholm 2010 Distributor: Department of Botany, Stockholm University

Martallen skriver inga skärgårdsdikter

Martallen vet hur dagen varit. Inget att anteckna Eller skriva skärgårdsdikter om. Kvällen kommer som den kommer. Månen tar sig fram över viken På det lokala sättet, med vrickåra. Skrivet suddas genast ut.

Werner Aspenström

Doctoral dissertation

Gundula Kolb Department of Botany Stockholm University SE-106 91 Stockholms Universitet, Sweden

The impact of cormorant nesting colonies on plants and arthropods

Abstract – Seabirds concentrate large amounts of marine nutrients on their nesting islands. This nutrient input can have large consequences for island food webs and community structure. The high nutrient load may also cause runoff into surrounding waters and affect marine communities. In my thesis, I studied the effect of cormorant nesting colonies on the stoichiometry, abundance, species richness, and species composition of plants, algae and invertebrates on land and in costal waters and investigated if differences in the elemental composition or homeostasis can explain differences in the numerical response among invertebrate groups. δ15N analysis indicated that ornithogenic nitrogen provided a significant nitrogen source for plants and arthropods on nesting islands and around high nest density islands also for brackish algae and invertebrates. Furthermore, nutrient runoff created a potential feed-back loop to spiders via chironomids. Cormorant nutrient input changed island vegetation and increased plant P and N content and epiphytic algae:Fucus ratio, but decreased plant species richness and vegetation cover. Invertebrates responded indirectly to these qualitative and quantitative changes in their food source and habitat, but also directly to cormorant subsidies. However not all taxonomic and feeding groups were affected and responses were both positive and negative. Differences in the numerical response among taxonomic groups could not be explained by differences in the level of homeostasis, since, generally, all invertebrates were strongly homeostatic. Similarly, consumer nutrient content was a poor predictor for displayed responses. I conclude that cormorant colonies have strong impacts on island vegetation and some consumer groups. However, even if they can decrease the species richness of some organism groups on their nesting islands, they increase the habitat heterogeneity in an archipelago and thus may increase the regional species diversity.

Keywords: seabirds, cormorants, islands, nitrogen, phosphorus, plants, arthropods, δ15N, Baltic Sea, ecological stoichiometry, species richness, species composition, numeric response

List of papers

This thesis is comprised of a summary and four papers, which are referred to by their Roman numerals: I. Kolb, G. S., Jerling, L. and P. A. Hambäck. 2010. The impact of cormorants on plant–arthropod food webs on their nesting islands. Ecosystems 13:353-366 II. Kolb, G. S., Jerling, L., Essenberg, C. and P.A. Hambäck. The impact of nesting cormorants on plants and arthropod diversity submitted III. Kolb, G. S., Ekholm, J. and P. A. Hambäck. 2010. Effects of seabird nesting colonies on algae and aquatic invertebrates in costal waters. Marine Ecology Progress Series accepted in Marine Ecology Progress Series IV. Kolb, G. S. and P. A. Hambäck. Ecological stoichiometry and homeostasis of plants and invertebrates on and nearby heavily fertilized cormorant nesting islands manuscript

Paper I is reprinted with the permission from Ecosystems.

Contents

Introduction  What is this all about? ...... 11  Why studying seabird islands ...... 11  Seabird as ecosystem engineers...... 13  The ecology of seabird islands ...... 15  Ecological stoichiometry ...... 19 Aims of the thesis...... 22 Material and Methods  Study system...... 24  Study organisms...... 30  Sampling ...... 32  Stable isotope analysis...... 33  Carbon-, nitrogen- and phosphorus analysis ...... 35  Statistics ...... 35 Results and Discussion  Stable isotope analysis – hypothetical food web (paper I and III) ...... 37  Effects of nesting cormorants on island ecosystem  Effect of nesting cormorants on soil chemistry ...... 40  Effect of nesting cormorants on island vegetation (paper I, II and IV) . 42  Effect of nesting cormorants on island arthropods (paper I and II)...... 43  Density response (paper I and II)...... 44  Responses in species richness (paper II) ...... 44  Responses in species composition (paper II)...... 46  Direct versus indirect effects of nesting cormorants (paper II) ...... 48  Effects of cormorant nesting colonies on algae and aquatic invertebrates in coastal waters (paper III)...... 52  My results in the context of other seabird island systems...... 54  Plant and consumer stoichiometry and homeostasis under increase nutrient loads  Effects of fertilization on plant and herbivore homeostasis and herbivore survival and body size – the Lythrum salicaria – Galerucella system (paper IV)...... 56  Ecological stoichiometry and homeostasis of plants and invertebrates on and nearby heavily fertilized cormorant nesting islands (paper IV) ..... 57 Concluding remarks...... 60 Further research...... 62 Acknowledgement ...... 63 References ...... 64 Svensk sammanfattning...... 77 Tack ...... 80

Introduction

What is this all about?

In this thesis I have investigated how cormorant nesting colonies affect terrestrial plants and brackish algae (hereafter referred to as plants) and terrestrial and brackish invertebrates on their nesting islands and in near-shore waters around the islands, in the archipelago of Stockholm, Sweden. Cormorants are fish-eating seabirds that concentrate marine nutrients from a large area on their nesting islands by depositing guano, carcasses, food scrap and reproduction by-products. Such allochthonous (= from outside the ecosystem) nutrients and energy can have strong effects on food web and community structure and on population dynamics of the recipient ecosystem. However, no former study has investigated how the nutrient and energy inputs of cormorants affect island arthropod communities. This lack of research is somewhat surprising considering that cormorants in Europe and the US gain high public interest and are often strongly disliked.

Why study seabird islands?

Islands are important habitats for many species, often endemic or endangered, and provide, as “model ecosystems”, information on basic ecological and evolutionary processes (Darwin 1859, see review in Cushman 1995, Vitousek et al. 1995). They provide perfect study systems due to their limited size and species diversity, their isolation, clear habitat borders, and worldwide prevalence which allow comparing the same processes under different abiotic and biotic factors. Not surprisingly, islands have often been used to study food web interactions (Spiller and Schoener 1994, Polis and Hurd 1995, Polis et al. 1997, Schoener and Spiller 1997), the importance of allochthonous nutrient and energy input (e.g., Polis et al. 2004, Anderson 1998, 2004, Paetzhold et al. 2008), diversity pattern (e.g., McArthur and Wilson 1963, Lomolino et al. 2006) and ecosystem functions (Wardle et al .1997, Jones 2010). The importance of cross-habitat flows of nutrient, energy and prey for ecosystem processes was realized by scientists as early as the 1920ties (e.g., Elton 1927, Polis et al. 1997a), but these fluxes were not incorporated in ecological theories until the 1990ties (Polis et al. 1997b, Power and Huxel 2004). Today it is commonly accepted that movements of energy, nutrients or prey between systems are ubiquitous, although the strength of effects from these subsidies vary among systems (Polis et al. 1997b). It has been proposed that the effects of allochthonous subsidies on consumer abundances are strongest in recipient systems with a lower productivity than the donor system (Polis et al. 2004, Barrett et al. 2005). How- ever, strong numeric responses of consumers to subsidies have also been observed

11 in high productive recipient systems (Marczak and Rickardson 2007, Marczak et al. 2007). Islands, generally, receive nutrient and energy from the surrounding aquatic system via four major pathways (Polis et al. 2004) (Fig. 1):  animals (e.g., seabirds, sea turtles, mammals) which forage in a large aquatic area, but nest and roost on islands where they deposit huge amounts of aquatic nutrients and energy (e.g., Huston 1950, Lindeboom 1984, Anderson and Polis 1999, Mulder et al. 2011)  shore drifted carrion and detrital algae (Brown and McLachlan 1990; Polis and Hurd 1996b)  emerging insects with aquatic larvae (e.g., chironomids) (Nakano and Murakami 2001, Gratton et al. 2008, Mellbrandt et al. 2010)  windblown sea foam and spray (Polis et al. 2004) Seabirds forage in large marine areas and deposit huge amounts of marine nutrients, mainly in form of guano, on their nesting and roosting islands. Further- more, seabird islands are not only characterized by the huge allochthonous nutrient load but also by a strong physical disturbance from the birds (Ellis et al. 2006, Smith et al. 2011). Through the nutrient input and physical disturbance, nesting seabirds can affect some of the major factors determining species diversity and composition (disturbance, productivity, and heterogeneity), limiting population size (nutrient and energy availability), and regulating food web structures and dynamics (resource and habitat availability) (Prime 1973, Hart and Horwitz 1991, Polis et al. 1997, Siemann 1998, Micheli 1999, Mittelbach et al. 2001, Price 2002), and thus give scientists a great opportunity to study such processes. The suitability of seabird islands as model ecosystems increases by their worldwide distribution, more than 250 species nest or roost on tens of thousands islands, in every ocean and several freshwater system and in every climatic region (Croxall et al. 1984, Mulder et al. 2011). Despite or maybe due to the large cultural and economic importance of seabird islands for people across the world, many seabird islands are threatened by human activities (Anderson and Mulder 2011). These threats include guano mining, introduction of predators, overfishing, pollution, or simply habitat destruction (Skaggs 1994, Blackburn et al. 2004, Jones et al. 2008, Anderson and Mulder 2011, Towns et al. 2011). Due to the possible effects of nesting seabirds on island ecosystems a significant decrease, or an extinction of seabirds from an island, may change island characteristics and lead to strong changes in vegetation structure, species composition and ecosystem function (Wardle et al. 1997, Maron et al. 2006, Towns et al. 2009). In order to predict the consequences of the seabird loss for island ecosystem, one needs to understand nutrient cycles, food web structures and ecosystem processes on seabird islands.

Figure 1. Marine nutrient input to terrestrial island ecosystems.

Seabird as ecosystem engineers

Birds that spend most of their life at sea are commonly called ´seabirds`, but the use of this term is highly discussed among biologists (Smith et al 2011). Generally, the term seabird includes the orders Sphenisciformes, Procellariiformes and Pelecaniformes (Gaston 2004). Totally, 222 marine bird species feed exclusively on marine food and nest near the sea, and 92 additional species are partially marine or freshwater species (Smith et al. 2011). Seabirds are wide-ranging pelagic and near-shore predators, foraging in areas as large as several km2, feeding across several trophic levels, from zooplankton to fish. Many seabirds are so called ´ecosystem engineers`, altering their nesting islands physically via their nesting activities and chemically via their deposition of marine nutrients (Jones et al. 1994, Smith et al. 2011). The strength of seabird impacts on the ecosystem of their nesting islands is strongly influenced by the seabird identity, which affect the time span of birds occupancy, the size of the colony, the nest density, the size of the bird and thus the guano volume per bird, guano nutrient content, and nesting typ. The time span of occupancy can vary from 32 days once per year, to year around in the tropics. 98% of all seabird species breed in colonies (Rolland et al. 1998) including a few up to 10 million nests. The size of the colony will directly affect the total nutrient input to the islands, whereas nest density will affect nutrient load per square meter. Seabirds transport marine-derived nutrients to their nesting and roosting islands by foraging at sea and defecating on land, as well as dropping food items, losing feathers and providing egg shell (Barrett et al. 2005). Marine nutrients and carbon are also added to the island ecosystem when birds die (Sánchez-Piñero and Polis 2000). However, guano is by far the most important nutrient subsidy on seabird islands (López-Victoria et al. 2009). For example, on

13 Malpelo Islands, Colombia, 40,000 breeding pairs of Nazca boody (Sula granti) were estimated to deposit 171.6 t marine nutrients during one breeding season (160 d), of which 170 t was guano (López-Victoria et al. 2009). Such huge ornithogenic nutrient loads have also been reported from the Rottnest Islands, Australia, where 11,700 wedge-tailed shearwaters (Puffinus pacificus) were estimated to deposit 234 kg ha-1per year. Guano is a complex mixture of undigested food residue and metabolic waste products, including uric acid as one of the main metabolic waste products (Bird et al. 2008). Fresh guano contains approximately 60% water, 7.3% nitrogen and 1.5% phosphorus (Smith and Johnson 1995). However, the exact nutrient composition differs between species and habitats depending on bird metabolism and food composition (Hahn et al. 2007, Smith et al. 2011). For example, dried cormorant guano from a French lake has been reported to contain 3.3% N and 14.3% P (Marion et al. 1994), whereas the cormorant guano from a Japanese lake was estimated to contain about 10.1% N and 9.2% P (Osono et al. 2002). The principal form of N in fresh seabird guano is uric acid, accounting for 50-80% of + guano N. The other major N-containing compound in guano is NH4 making up 8- 40%. Amino acid and proteins account only for > 8% of guano N (Lindeboom 1884, Schmidt 2004). After guano deposition on the island, ammonia volatilizes rapidly from the uric acid and cause N - losses between 25-80% of total guano-N and, as later discussed, an increase in δ15N signature in the remaining guano (Lindeboom 1884, Smith and Johnson 1995). The P in fresh guano-P consists of 54.4% PO4 (Lindeboom 1884, Smith and Johnson 1995). The second greatest factor that characterizes seabirds as ecosystem engineers is physical disturbance. The strength of this disturbance is strongly dependent on the nesting type and seabird behavior. Generally, four types of seabird nests can be found: surface nests, tree nests, crevice and cavity nests, and burrows. The strongest physical disturbance is caused by burrowing seabirds, affecting soil structure and damaging vegetation by removing twigs and leaves and damaging roots (Smith et al. 2011). Surface and tree nesters, like cormorants, can also strongly violate the vegetation. Surface nesters disturb the ground surrounding of their nests by trampling, whereas tree nesters physically damage trees by breaking branches and removing leaves to line their nests (Hobara et al. 2001). Furthermore, cormorants were also observed to eliminate recruitment of new seedlings (Ishida 1996, Hebert et al. 2005).

Pathways and expected effects of cormorant subsidies Cormorant subsidies Pathway of cormorant subsidies (δ15N) Effects of cormorant on density and Cormorants diversity (= species richness) ↑ increase, ↓ decrease ↑ density Predators ↑ diversity carcasses n ihGuano N andP rich

↑ density ↑ density Scavengers ↑ density Herbivores ? diversity Detritivores Chironomidae ↑ density ↑ diversity ↑ diversity

↑ N and P, ↑ biomass, Vegetation: ↓ species richness ↑ N and P Marine algae ↑ epiphyte biomass

↑N and P ↓pH Soil: ↓ moisture Grazer ↑ density Water: N and P N and P rich run-off ↑

Figure 2. Conceptual model for the effects of nesting seabirds on islands ecosystem.

The ecology of seabird islands

The former paragraph described the role of seabirds as ecosystem engineers. Now I will discuss the consequences the engineering can have for the island ecosystems (see conceptual model in Fig. 2). It has been show that seabirds can largely affect soils and vegetation on their nesting islands through nutrient deposition and physical disturbance (e.g., Gillham 1961, Burger et al. 1978, Sobey and Kenworthy 1979, García et al. 2002, see review by Ellis 2005, Wait et al. 2005, Fukami et al. 2006, see metaanalysis Mulder et al. 2011). The large majority of studies comparing nutrient content in soils from seabird colonies with the nutrient content in soils at a distance from colonies found increase soil N and P contents within the colonies (e.g., Ishizuka 1966, Burger et al. 1978, Mizutani and Wade 1988, Polis and Hurd 1996, see review by Ellis 2005, Hobara et al. 2005, Wait et al. 2005, see metaanalysis by Mulder et al. 2011). Furthermore, stable isotope analysis of N has indicated a considerable input of marine-derived N to the soils of seabird islands (Mizutani et al. 1986, Erskine et al. 1998, García et al. 2002, Fukami et al. 2006, Maron et al. 2006). Accordingly, the enriched δ15N signature of plants on seabird islands and their increased N and P content demonstrate that marine nutrients are taken up by island vegetation and thus enters the island food

15 webs (Erskine et al. 1998, Wainwright et al. 1998, Anderson and Polis 1999, García et al. 2002, Ellis et al. 2006, Fukami et al. 2006, Maron et al. 2006). However, even if the vast majority of studies described increased N and P content of soils and plants in seabird colonies, the strength of the nutrient increase varied strongly among studies. The increased nutrient content in the soils of seabird islands is expected not only to increase plant nutrient content but also to change plant species composition, plant species richness, and primary productivity. Fertilization experiments show that high nutrient loads decrease plant species richness, change species composition towards a few fast growing, nutrient tolerant species with high fecundity, and increase primary productivity (Tilman 1987, Wedin and Tilman 1996, Siemann 1998, Haddad et al. 2000). However, island vegetation is expected not only to respond to ornithogenic nutrient input, but also to the physical disturbance by seabird nesting activity. Strong physical disturbance should decrease plant species richness and change species composition towards taxa that exhibit high resilience (Connell 1978), whereas an intermediate level of disturbance (= low nest density) may increase plant species richness (Connell 1978). Primary productivity may similarly decrease under strong disturbance. Thus, the total impact of seabirds might be regulated by the balance of nutrient addition and physical disturbance (Mulder et al. 2011). Accordingly, Ellis (2005) found that the number of plant species is lower on seabird islands than on similar islands without seabirds. Furthermore, the plant species composition differed significantly between seabird and non-seabird islands. The families Apiaceae, Brassicaceae, Chenopodiaceae, Poaceae and Solanaceae were most frequent and species-rich on seabird islands, whereas Gentianaceae and Orchidaceae were most frequent and species-rich on non-seabird islands (Ellis et al. 2011). Primary productivity is difficult to measure in the field and therefore, has rarely been studied on seabird islands. In a couple of studies on seabird islands, aboveground plant biomass has been used as estimate for primary productivity indicating that the effect of nesting seabird on plant productivity differ by archipelago (Anderson et al. 2008, Maron et al. 2008, Wardle et al. 2007). In their role as ´ecosystem engineers` seabirds not only affect island vegetation but also island consumers (e.g., Polis and Hurd 1995, Barrett et al. 2005, Towns et al. 2009, see review by Kolb et al. 2011), primarily via their role as vectors of marine nutrients onto islands and as physical modifiers of island soils and vegetation. Through either role seabirds alter the quantity and quality of food and habitat resources of island consumers. As nutrient vectors, they subsidize the diet of consumers on different trophic levels. This effect has been shown with stable isotope analysis; consumers from different seabird islands systems, including amphipods, insects, spiders, snails, lizards, geckos, and rodents have been found to be enriched in 15N and thus included seabird-derived N in their diet (Sánchez- Piñero and Polis 2000, Stapp and Polis 2003, Wolfe et al. 2004, Barrett et al. 2005, Fukami et al. 2006, Maron et al. 2006, Hawke and Clark 2010, Kolb et al. 2011). Nitrogen and phosphorus from guano is taken up by plants which are then consumed by herbivores, detritivores, and their predators (Sánchez-Piñero and Polis 2000, Barrett et al. 2005, Fukami et al. 2006). Guano itself can be consumed

16 directly by some invertebrates (Sánchez-Piñero and Polis 2000). Eggs and chicks are prey for larger predators (e.g., snakes, hawks, rodents), and, together with adult carcasses and food scraps, provide resources for scavenging insects (Polis and Hurd 1996, Sánchez-Piñero and Polis 2000). Live seabirds also support several ectoparasites, which can become prey for arthropod predators (Polis and Hurd 1995). The multiple pathways by which seabirds “subsidize” the diet of island consumers can generate different direct and indirect responses in island consumer groups, including changes in population sizes, fitness, survival and species richness (e.g., Iason et al. 1986, Wolfe et al. 2004, Polis and Hurd et al. 1995, Towns et al. 2009). Increased abundance has been reported from rotifers, enchytraeids, nematodes, snails, amphipods, several orders of insects (including all major trophic groups), chilopods, mites, spiders, and lizards (Onuf et al. 1977, Iason et al. 1986, Polis and Hurd 1995, 1996, Ryan and Watkins 1989, Markwell and Daugherty 2002, Oregeas et al. 2003, Barrett et al. 2005, Fukami et al. 2006, Sinclair and Chown 2006, Towns et al. 2009). However, not all consumers respond with increased abundances on seabird islands, some showed no response, such as soil invertebrates in Antarctica (Sinclair 2001) and 11 of 19 investigated litter-invertebrates in New Zealand (Towns et al. 2009) and in one case – herbivorous beetles in a gull colony on a French Mediterranean island (Oregeas et al. 2003) – a negative. Different consumer responses may be partly due to differences in species traits, particularly traits known to affect consumer responses to nutrient input. A further explanation for the absence of positive numeric response might be the fact that food webs are rarely simple and linear. Even though seabird effects are primary bottom-up effects, top-down forces will likely bias the overall results (Hunter and Price 1992) and in some cases top-down effects can overwhelm the benefits of an primary consumers and cause a zero or even negative consumer response (Fukami et al. 2006). Furthermore, seabird induced changes in soil and vegetation may affect the habitat quality for some consumers negatively and thus again overwhelm the bottom-up benefits from an increased nutrient availability (Oregas et al. 2003). The effects of nesting cormorants on resource availability and habitat quality might not only affect consumer abundance but also consumer species richness and composition (Towns et al. 2009, Oregas et al. 2003). For examples, seabirds can change several characters of island vegetation with consequences for island consumers. Increased plant nutrient content and biomass and decreased plant species richness, concentrate the resources in the vegetation and increase the tissue quality for herbivores and thus may, assuming bottom-up control, directly increase herbivore and detritivore abundance and indirectly the abundance of predators (Root 1973, Mattson 1980, Kyto et al.1996, Siemann 1998, Haddad et al. 2000, Barrett et al. 2005). Similar to the positive effect on consumer abundance, higher plant nutrient content and productivity may also increase consumer species richness (Mattson 1980, Waide et al. 1999, Haddad et al. 2000, Mittelbach et al. 2001). In contrast, lower plant species richness should, according to the habitat heterogeneity hypothesis, support fewer consumer species than higher plant

17 species richness (MacArthur and MacArthur 1961, Siemann et al. 1998, Haddad et al. 2000, Hart and Horwitz 1991). Summarizing, seabird islands seem, generally, to be resource-driven systems, and seabird-derived subsidies have mostly positive effects on the abundance consumer groups, while the effect on species richness remains widely unknown.

Effects of nesting seabirds are not limited to the terrestrial systems but also include adjacent aquatic systems (see review by Young et al. 2011). Seabirds can affect marine and freshwater systems both as predators via ´top-down effects` and as nutrient vectors via ´bottom-up effects`. Top-down effects have been reported from the intertidal systems where seabirds prey on limpets, snails and crabs (Ellis et al. 2007, Wootton 1992, Kurle et al. 2008), whereas it is debated to what extent fish-eating seabirds affect their prey populations (Draulans 1988). Most seabirds forage in large marine areas and share their prey with other predators (e.g., fishery, mammals, and predatory fishes) which makes it difficult to study the seabird effect on prey abundance. However, studies on freshwater systems demonstrate that waterbirds, shorebirds, and seabirds can contribute significant amounts of nutrients into the water column and sediments, and can stimulate primary and secondary productivity (e.g., Blais et al. 2005, Hahn et al. 2007, Keatltey et al. 2009, see review by Young et al. 2011). Less clarity exists on the types and magnitudes of seabird effects on marine systems, because of the huge variation in ecosystems structure and sampling method used. However, most studies indicate similar, even though weaker, effects than in freshwater systems (Golovkin 1967, Onuf et al. 1977, Bosman et al. 1986, Bosman and Hockey 1988, Litter et al. 1991, Wootton 1991). Nutrients from guano can enter water bodies near bird colonies in four ways. Firstly, birds directly deposit guano into the water (Wainright et al. 1998). Secondly, guano runs off from bird colonies into surrounding water bodies (Golovkin and Garkavaya 1975, Smith and Johnson 1995). The amount of nutrient runoff from land depends on nest density, climate, vegetation and habitat structure (Branvold et al. 1976, Smith and Johnson, Loder et al. 1996). Thirdly, guano enters marine systems via leaching of nitrogen into ground water, following subsequent dispersal during tidal oscillation (Smith and Johnson 1995). Finally, guano enters marine systems after volatilization to ammonia, which then dissolves in rain. Increased nutrient concentrations in the water column nearby seabird colonies have been reported from several systems, including rock pools, bays, coastal lagoons, coastal areas, and islands (Golovkin and Garkavaya 1975, Bedard et al. 1980, Bosman et al. 1986, Lapointe et al. 1992), but only one study combined nutrient content analysis with stable isotope analysis and thus was able to determine the actual role of guano within the systems. Increased water nutrient content should, due to their N or P limitation increase productivity of marine primary producers (Elser et al. 2007). Accordingly, algae and vascular plants near seabird colonies were shown to have increased abundances, growth rates or nutrient content (Golovkin 1967, Onuf et al. 1977, Bosman et al. 1986, Bosman and Hockey 1988, Litter et al. 1991, but see Wootton 1991). These changes in the primary producers can affect aquatic consumers via bottom-up forces and both in-

18 crease (Zelickman and Golovkin 1972, Onuf et al. 1977, Bosman and Hockey 1986, Palomo et al. 1999) and decrease (Wootton 1991) consumer abundance. To summarize and conclude, nesting seabirds can have strong effects on islands ecosystems, both on land and in the water. However, there is a huge variation between systems depending on biotic and abiotic factors like size of seabird colony, nesting density, climate, habitat structure, ecosystem nutrient status, and community structure. Far more studies are needed to develop good prediction about the effects of nesting seabirds on island ecosystems.

Ecological stoichiometry

In order to understand seabird impacts on island communities and to predict seabird effects on community structure and species composition it may be worthwhile to look closer on species traits assumed to affect organisms’ responses to nutrient input. An organism`s response to increased nutrient content in its resource should be dependent on its nutrient requirements for survival, growth and reproduction. Phytophagous insects generally face the problem of large mismatch in nutrient content between their body tissue and plant tissue; plants content in average 10-20 times lower N and P than herbivores (McNeill and Southwood 1978, Mattson 1980, Elser et al. 2000). This unbalanced elemental composition between plants and consumers is thought to cause nutrient limitation in herbivores and detritivores (White 1993, Cross et al. 2003) and to critically influence herbivore survival, growth, and fecundity (Slansky and Feeny 1977, Mattson 1980, White 1993, Joenern and Behmer 1997, Wheeler and Halpern 1999). Therefore, increased plant nutrient content (as on seabird islands) should increase survival, growth, and reproduction rates of primary consumers and their abundance. Consumers on higher trophic levels may similarly to herbivores also feed on diets lower in nitrogen content than themselves (Fagan et al. 2002). Increased plant nutrient content therefore may cascade up to higher trophic levels and increase consumer abundances via increased prey abundance or prey nutrient content (Mayntz and Toft 2001, Hunter 2003, Raubenheimer and Simpson 2004). However, consumer response to increased nutrient availability may differ among species due to differences in nutrient demand. The nutrient demand of a given taxa is determined by its metabolism, structural design, specific growth rate and is reflected in its elemental composition (Elser et al. 1996, Elsner et al. 2000, 2003, Sterner and Elsner 2002). The growth rate hypothesis (GRH) propose that differences in organismal P content are caused by differential allocation to P-rich ribosomal RNA in order to meet the protein synthesis demands of rapid growth rates (Elser et al. 2000a,b, Sterner and Elser 2002). That is, fast growing species have higher P body content and thus higher P demands and are more sensitive to P limitation than slow growing species (Urabe and Watanable 1992, Sterner and Elser 2002, Urabe et al. 2002). In nutrient rich environment, like seabird islands, therefore fast growing taxa with high N:C and P:C ratios may be more successful

19 and thus dominant over consumers with low ratios whereas the opposite may occur in nutrient poor environments (Urabe and Watanable 1992, Urabe et al. 2002). Accordingly, differences in the elemental ratios among consumer groups may explain differences in their numeric response to nutrient content. Nutrient rich consumers should be more favoured by increased resource nutrient content and display stronger numeric response on seabird islands than nutrient poor consumers. A branch of ecology studying the relationship between elemental ratios, namely carbon (C), nitrogen (N) and phosphorus (P), of organisms and ecological processes and interactions is called ecological stoichiometry. Ecological stoichiometry is defined as the study of the balance of multiple chemical substances (elements) and sometimes energy in biological processes of living systems and in ecological interactions (Sterner and Elser 2002). It simplifies natural complexity to a limit set of chemical parameters and applies the law of definite proportion and conservation of mass, in order to increase the understanding of factors limiting growth, abundance, and distribution of organismal ecological stoichiometry (Sterner and Elser 2002). Based on these principles, ecological stoichiometry explains and predicts biological processes, food web structures, and community compositions. Thus, stoichiometric homeostasis, the physiological regulation of organism elemental composition relative to their external world, is a key concept in ecological stoichiometry (Kooijman 1995, Sterner and Elser 2002). Organisms with homeostasis can keep the chemical composition of their body constant, independent of the chemical composition of their food source or the environment (Kooijman 1995). This homeostasis can be maintained by food selection and physiological regulation of elemental uptake, assimilation, or excretion (Logan et al. 2002, Anderson et al. 2005, reviewed by Frost et al. 2005). Heterotrophs (animals) are commonly seen to be more homeostatic than autotrophs (plants) (Sterner and Elser 2002, Persson et al. 2010). Based on differences in physiology; autotrophs house, in contrast to heterotrophs, large central vacuoles in which they can store highly concentrated nutrients (e.g., N and P) or nutrient- free organic matter (e.g., sugars). However, during the last decade, the generality of strict homeostasis has been questioned and several studies indicate that the C:N:P ratio of some heterotroph species can vary in dependence of resource nutrient content (e.g., DeMott 2003, Mulder and Bowden 2007, Mulder and Elser 2009, Small and Pringle 2010, reviewed by Persson et al. 2010). Plastic organisms may be better adapted and more successful to variable environments than strict homeostatic organisms; because they store nutrients for later growth (Sterner and Schwalbach 2001, Raubenheimer and Jones 2006) and can, when consuming a nutrient poor diet, increase their nutrient use efficiency by adding a lower concen- tration of nutrients to new tissue (Elser et al. 2003). Thus plastic organisms may be able to buffer variations in nutrient supply and respond less to variation in nutrient availability (e.g., cormorant versus reference islands). The degree to which organisms are homeostatic may affect ecological interactions and, furthermore, influence the magnitude and nature of their ecological impact (Elser and Urabe 1999, Kay et al. 2002, Grover 2003, Mulder and Bowden 2007), e.g., strict homeo-

20 stasis may cause a disproportional depletion of resources with scarce elements and increase recycling rates of other elements (Vanni 2002). However, despite these important consequences on both individual and community level and strongly increased anthropogenic nutrient load little research has been done on the level of elemental homeostasis in heterotrophs others than zooplankton (see review Persson et al. 2010).

21 Aims of the thesis

The overall aim of this thesis was to investigate and understand how cormorant colonies affect terrestrial plants, brackish algae and invertebrate communities on and nearby their nesting islands (Fig. 3). 1) I analyzed the food web structure using stable isotope, particularly the relative use of marine versus terrestrial C and N, on and off islands with cormorant nesting colonies and compared the food web structure with that on reference islands (paper I). 2) I investigated the consequences of the high N and P input by nesting cormorants on island vegetation: aboveground biomass, species richness, species composition, and N and P content (paper I and II). 3) I assumed that changes on primary producers have consequences on higher trophic levels and a further question was thus how cormorants directly and indirectly affect island consumer abundance, species richness and species composition (paper I and II). Do taxonomic and feeding groups differ in their responses to nesting cormorants (paper I and II)? Which consumer groups respond directly to the cormorants and which groups respond to changes in the vegetation structure (paper II)? 4) Stable isotope analysis indicated that ornithogenic N also enters brackish algae in the near-shore waters of cormorant nesting islands. Thus I investigated how cormorants affected near-shore algae, the algae N and P content, and the abundance and biomass of marine invertebrate grazers (paper III)? 5) Differences in the response of consumer groups to cormorant colonies may be caused by differences in consumer nutrient demands and elemental homeostasis. Therefore, I investigated the N:C, P:C, and N:P content of primary producers and consumers on and around cormorant islands and tested for their elemental homeostasis. We tested following hypothesis (paper IV):  Nutrient rich taxa may be more favored by the high nutrient availability on cormorant islands than nutrient-poor taxa  Autotrophs are more homeostatically plastic than heterotrophs.  Strictly homeostatic consumers may show a stronger numerical response to cormorant nutrient input than more plastic consumers.

Brackish algae Invertebrtaes Changes in Changes in • Productivity • Abundance • Species richness •Biomass • Species composition Paper III • Species richness • Epiphytic algae: Fucus • Species composition ratio • Tissue quality • Tissue quality Paper IV • Fitness •Growth rate

Possible responses Possible •Growth rate Possible responses Possible • Population dynamics • Population dynamics

Water Cormorant nesting colony Changes in Guano • N and P content Carcasses Soil Eggshells, Feathers Paper I Food scrap Changes in • Microflora/fauna Kappa • Nutrient content •ph •Moisture •Structure

Terrestrial vegetation Arthropods Changes in Changes in • Productivity/Biomass •Biomass • Species richness Paper I • Abundance • Species composition Paper II • Species richness • Tissue quality • Species composition • Seed production Paper IV • Tissue quality • Growth rate • Fitness • Population dynamics •Growth rate Possible responses Possible Possible responsesPossible • Population dynamics

Figure 3. Possible responses of soil, water, terrestrial plants, brackish algae, and terrestrial and brackish invertebrates to cormorant nesting colonies. Marked are responses studied in this thesis.

23 Methods

Study system

Figure 4. Location of the islands in Stockholm archipelago, Sweden studied in thesis. For further information see Table 1.

The studies in this thesis were conducted on and around 27 islands in the archipelago of Stockholm, Sweden (Fig. 4). The archipelago includes about 24,000 large and small islands spread over an area more than 35,000 km2. The archipelago is part of the Precambrium peneplane. It is subject to isostatic rebound after the retreat of the last icesheet, at a current rate of 0.47 cm/year (Ericson and Wallentinus 1977). The bedrock consists of granite-gneiss, on the islands mostly covered with a very thin soil layer. The yearly precipitation averages 518 mm; the mean temperature is 6.1°C. At the latitude of Stockholm the archipelago stretches about 100 km out to the east. The inner archipelago is characterized by large and small islands covered by pine (Pinus sylvestris) and spruce (Picea abies) forest harboring forest species (e.g., Vaccinium myrtillus, Convallaria majalis and Polygonatum odoratum) and some agricultural land. The central archipelago includes many small islands vegetated by deciduous trees (Alnus glutinosa and Sorbus aucuparia). Common are shrubs (Juniperus communis, Rosa canina, and Rubus idaeus), herbs, grasses and bare rocks. The outer archipelago consists of groups of sparsely vegetated

24 small islands and skerries sprinkled in wide open waters. The most widespread herbs along the shore and on non forested islands are Atriplex spp, Glaux martima, Filipendula ulmaria, Valeriana salina, Lythrum salicaria, Tanacetum vulgare, Veronica longifolia, Viola tricolor, Allium schoenoprasum, Sedum acre, Sedum telephium and the grasses Phragmites australis, Phalaris arundinacea, Festuca pratensis, and Deschampsia flexuosa. Until the 1950ties, most islands in the archipelago were used for farming and livestock. Today, the majority of buildings on the islands are summer houses and the archipelago is frequently visited by tourists for boating, swimming and sun bathing. The water in the Baltic Sea is brackish with a salinity varying between 5.5 and 7 psu in the archipelago of Stockholm. The low salinity enables only a few marine and freshwater species to occur in the Baltic Sea. The vegetation on hard bottom of the upper sublittoral zone (0.5-1m) is dominated by filamentous algae and the species composition varies between seasons. In summer Cladophora glomerata often dominates (Kautsky et al. 2000). Below C. glomerata, the marine brown algae Fucus vesiculosus (bladderwrack) (hereafter Fucus) usually dominates. Below the Fucus belt, marine red algae take over. Fucus is a key species in the Baltic Sea, supporting about 30 species of macrofauna and epiphytes and providing spawning, foraging, and nursery areas for fish (Kautsky et al. 1992, 2000). On soft bottoms aquatic phanerogams and charophytes grow. Species composition varies with depth and bottom characteristics. The water body in the inner archipelago is sheltered from the open sea and consists of shallow bays and fjords. The inner waters are strongly influenced by freshwater and land-based substances from stream and coastal catchments. During January and February large parts of the inner archipelago are often covered by ice. The outer archipelago is very exposed to wind and waves, especially in the east. Eutrophication is one of the major environmental problems in the Baltic Sea, especially in coastal regions, and the main nitrogen sources are rivers and atmospheric deposition (Elmgren 2001, Voss et al. 2006). The main nutrient inputs to the archipelago are discharges from sewage treatment plants, the outflow from Lake Mälaren in central Stockholm, and the atmospheric deposition on the surface. The nutrient load to the archipelago decreases from the inner to the outer archipelago (Aneer and Arvidsson 2003). In the 1990s, discharges from sewage treatment plants and the outflow of Lake Mälaren contributed an annual load of 5,960t nitrogen and 10t phosphorus to the inner archipelago (Stockholm Water Company 2004 in Hill and Wallström 2005). According to Swedish water-quality criteria (Swedish Environmental Protection Agency 1999 in Hill and Wallström 2005), the nutrient concentrations in the surface water were categorized to be "high" to "very high" in the inner waters and "low" to "very low" in the outer waters. Such nutrient concentrations cause extensive blooms of Cyanobacteria in the inner archipelago, and a decrease in the depth distribution of Fucus. In many parts of the archipelago, eutrophication has caused anoxic conditions on deep bottoms and killed macrofauna. Eutrophication may have caused increased densities of pikeperch (Sander lucioperca) and smaller catches of whitefish (Coregonus spp.) (Hansson and Rudstam 1990).

25 Large colonies of seabirds exist in the outer archipelago. About 170,000-180,000 pairs of seabirds breed in the Stockholm archipelago (Kautsky et al. 2000). Great cormorants not only nest in the outer archipelago but also in the center and inner archipelago. For our studies, we chose eleven islands with active and three with abandoned cormorant colonies and 13 reference islands without nesting cormorants (Tab 1, Figs. 4 and 5). This includes about 50% of all cormorant colonies of the archipelago. The islands differ in size (0.3 to 2.7 ha), distance from mainland (0.1 to 18 km), wave exposure, and vegetation type (forest or open grassland). The nest densities in the active cormorant colonies ranged from 0.007 to 0.070 nest/m² (2008) and the time span since colonization was two to 12 years (2009). Cormorants are sensitive to human disturbance and colonies are typically estab- lished on islands with limited human activity. To reduce the probability of con- founding results we chose reference islands in the same range of size and distance to the mainland and with similar vegetation structures as the cormorant islands.

Figure 5. Cormorant nesting islands in Stockholm archipelago, Sweden. (A) Active cormorant island with high nest density (N. Småholmen), (B) Active cormorant island with low nest density (Bergskäret), and (C) Island with an abandoned cormorant nesting colony (S. Halm- ören).

Table 1. Summary of the 27 islands studied.

Island Nest Mean nest density/m2 Vegetation Arthropod sampling Time span density/ Sp/m2 Area Wave of active sweep- Fucus m2 2004 - 2003- 2006- E D-Vac pit fall (ha) exposure coloni- Sp net sampling 2007 2007 2008 2009 bio 2007 2009 zation 2008 (2008) vegco

Abandoned Cormorant Islands

Kattören*3 0.326 0 0 0 0 0 1997 – 00 Y Y Y N N N

A 51374 St.Halmören*2 0.809 0 0.012 0.007 0 2002 – 05 Y Y Y Y Y Y B 50072

St.Träskär*2 1.573 A 52051 0 0.005 0.002 0 1996 – 05 Y Y Y Y Y Y B 51900

Active Cormorant Islands

Alörarna *1 0.233 (0.070) 0.006 0.023 0.032 2007 – 09 Y N N N Y N

A 19060 0.025 Marskärskobben*1 0.716 0.018 0.023 0.025 2003 – 09 Y Y Y Y Y Y B 18821 (0.030)

A 6785 0.063 N.Småholmen*4 0.855 0.054 0.058 0.057 2000 – 09 Y Y Y Y Y Y B 7738 (0.052)

S.Småholmen*4 0.701 0.027 0.024 0.026 0.028 1998 – 09 Y Y Y Y N N

A 9524 0.039 N.Ryssmasterna*4 0.322 0.029 0.042 0.051 2003 – 09 Y Y Y Y Y Y B 9202 (0.052)

S.Ryssmasterna*4 0.369 0.017 0.006 0.015 0.024 n.a.– 09 Y Y Y Y N N

A 16834 0.029 Bergskäret*5 2.273 0.026 0.027 0.030 1998 – 09 Y Y Y Y Y Y B 19002 (0.028)

Delö*6 0.170 0.008 0.020 0.021 0.020 2002 – 09 Y* Y Y N N N

A 8846 Skraken*7 1.688 (0.024) N N N N N Y B 8613

A 340983 Våmkubben*8 1.051 (0.041) N N N N N Y B 340983

A 181913 Fälöv*8 1.431 (0.034) N N N N N Y B 180627

Reference Islands

Fårörarna*1 0.329 A 21710 Y Y Y Y Y Y

B 21710

Norröra*2 1.248 Y Y Y N N N

A 71649 Ägglösen*2 1.736 Y Y Y Y Y Y B 77045

A 50567 V.Mellgrund*2 0.389 Y Y Y Y Y Y B 50393

Ljusstaken*3 0.553 Y Y Y N N N

A 7616 Hannasholmen*4 0.774 Y Y Y Y Y Y B 8023

A 9793 S. Nickösö*4 0.521 Y Y Y Y Y Y B 9783

N. Nickösö*4 0.520 N N N Y N N

A 17361 Mjölingsören*5 2.729 Y Y Y Y Y Y B 17361

Ostkanten*6 0.234 Y* Y Y N N N

A 11073 Fredagen*7 1.803 N N N N N Y B 10883

A 382085 Rödkläppen*8 1.285 N N N N N Y B388320

A 297421 Skorvan*8 1.688 N N N N N Y B 298047

Note: Locations are shown in Fig. 1. : Sp: total plant species richness, Sp/m2: mean species richness/m2, E: evenness, bio: aboveground plant biomass (g/0.0625 m2), vegc: vegetation cover, Y: yes, N: no. * Total plant species number on these islands were estimated with help of morphospecies for grasses and were not included in plant species richness analysis. Wave exposure was calculated as effective fetch (distance of free water) in 16 directions for both Fucus sample sides (A and B) per islands with the help of the program WaveImpact (Isæus 2004). Variables not shown in the table have not been used for analysis.

Study organisms

Cormorants Great cormorants (Phalarocorax carbo) are one of the best known wild European seabirds, belonging to the family of cormorants (Phalacrocoracidae) and to the order of web-footed birds (Pelecaniformes). Three subspecies exist in Europe: Ph. c. carbo on the Atlantic coast, Ph. c. maroccanus along the coasts of Morocco, and Ph. c. sinensis in south-eastern Europe to China. The latter is the subspecies occurring in the Stockholm archipelago. Great cormorants live in colonies of up to 10,000 pairs (nesting colonies) or 12,000 individuals (roosts) (Carss 2002). They breed and roost in trees, on bare soil or rocks and in reed beds (Fig. 6). Cormorants are opportunistic foragers, feeding primarily on benthic fish but also include pelagic fish in their diet (Gremillet et al. 1998). They are diving hunters and usually prefer small fish, either small species or young individuals of larger species. Great cormorants forage up to > 20 km distance from their breeding colony (Paillisson et al. 2004, Engström 2001).

Figure 6. Cormorant nesting on trees (A) and on the surface (B)

Cormorants are one of few seabirds with increasing numbers in Europe. Great cormorants in Sweden have a long history; 8000–13000 years old bones have been found. During the 19th century cormorant colonies existed in the south of Sweden (Skåne and Blekinge), but they disappeared in the beginning of the 20th century. They started to recolonize the south of Sweden (Kalmarsund) in 1948. The archipelago of Stockholm was recolonized in 1994. Their abundance in the archipelago increased to 5759 breeding pairs in 21 colonies by 2007, whereafter the population size seemed to have stabilized at about 5200 breeding pairs (Fig. 7). (Staav 2008, 2010). The colonies in our study area are active between April and August and span between 4 and 776 pairs (2009). During fall, most birds migrate and overwinter in the Mediterranean and the Black Sea. Cormorants are disliked in Sweden, as in other parts of Europe. They are blamed by fishermen for reduced catches and by the public for damaging the environment. However, their presence has a positive effect on threatened birds, like sea eagles (Haliaeetus albicilla) and guillemots (Cepphus grylle), by providing prey and creating favored nesting habitats. Despite the controversial public discussion and the increasing cormorant density in Europe, not much

30 scientific research has been done on the impact of cormorant colonies on ecosystems.

s 6000 5000 4000 3000 2000 1000

Number of cormorant nest cormorant of Number 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Year

Figure 7. Development of cormorant population in Stockholm archipelago (Staav 2008, 2010).

Plants and invertebrates Paper I and II focus on terrestrial plants and arthropods, paper III on aquatic algae and invertebrates and paper IV on both terrestrial and aquatic plants and invertebrates. Chironomids, emerging insects with aquatic larvae were subject of all studies. Paper I studied all abundant arthropod groups including the major trophic groups: herbivores, detritivores, scavengers and predators. Paper II focused on beetles (Coleoptera), true bugs (Heteroptera), spiders (Aranaea) and non-biting midges (Chironomidae, Diptera). Paper III deals with Fucus vesiculosus, epiphytic algae, and the most abundant grazers associated with Fucus, the mollusc Theodoxus fluviatilis, the crustaceans Jaera albifrons, Gammarus spp. and Idotea spp., and chironomids.

Lythrum salicaria – Galerucella – system (paper IV) Purple loosestrife, Lythrum salicaria L. (Lythraceae) is an abundant perennial herb on the shore-line in most areas of the northern Baltic Sea. The beetles Galerucella calmariensis L. and G. pusilla L. (Coleoptera: Chrysomelidae) are monophagous herbivores of L. salicaria (Hambäck et al. 2000).

Sampling and sample preparation

Soil Per island, three soil samples were taken spread over the islands. Sampling was carried out with a soil core (ø 3cm) at 0-10 cm depth. Samples were divided into two layers 0-5 cm and 5-10cm depth. Samples were cooled directly after sampling until sieving in the evening. All large organic and inorganic material was removed from the soil samples by sieving through a 2 mm mesh sieve. Sieved samples were frozen until further preparation. For pH analysis 1 ml soil was dissolved in 5 ml water and pH was measured under stirring (Swedish version of the international standard ISO 10 390:1995). For phosphate analysis, samples with dominantly organic material were milled in a cychlotech mill to 0.5 mm. Sandy samples were homogenized in a machine mortar for maximum 3 minutes. Inorganic phosphorous in the milled or homogenized samples was determined following extraction with 2% citric acid (1:5 soil to extract solution) (Linderholm 2007). For ammonium and nitrate analysis, soils were extracted with 2M KCl (100 g soil and 250 ml liquid) for 2 hours for sandy soils and over night for clayey soils. Soils were filtered and analyzed on a flow injection analyzed (Foss, Sweden), following the application notes AN 50/84 (Tecator, 1984) and ASN 50-01/92 (Tecator, 1992).

Vegetation (paper I and II) Vegetation cover, rock cover, aboveground plan biomass, mean plant species richness per m2, and evenness were estimated along 1 or 2 transects across the islands in July 2007. Total plant species richness was estimated during May and September 2007–2009 by inventories of all plants on islands.

Terrestrial arthropods Arthropods were sampled with suction sampling, sweep-netting, pitfall trapping, and hand collection during 2007–2009, for estimating density (paper I and II) and species richness (paper II), and to carry out stable isotope analysis (paper I), and carbon (C-), nitrogen (N-), and phosphorus (P-) analysis (paper IV) (Tab. 1). Suction samples were sorted to order, families or species (Coleoptera, Heteroptera). Spiders and coleopterans from pitfall traps and chironomids were sorted to species. All taxa were grouped in trophic groups based on information from the literature (Jacobs 1998) and from experienced taxonomists (H.-E. Wanntorp, C.-C. Coulianos and Y. Brodin).

Aquatic algae and invertebrates (paper III and IV) Fucus and its associated epiphytic algae and invertebrates were collected within near shoreline. We randomly collected 6 Fucus fronds per island, 3 fronds per sample site. T. fluviatilis, J. albifrons, Gammarus spp., Idotea spp., and Chironomidae were counted and used for further analysis. To calculate the ratio of epiphytic algae to Fucus in each sample, one branch of the Fucus frond was cleaned from all epiphytic algae by scraping.

Stable isotope analysis (paper I and III)

Stable isotope analysis has become a popular tool for ecologists to elucidate the structure of food webs, to calculate the importance of marine versus terrestrial food sources for terrestrial consumers and to trace the fate of avian nitrogen (Lindeboom 1984; Anderson and Polis 1998; Wainright et al. 1998; Barrett et al. 2005; Paetzold et al. 2008). The elements C, N, S, H and O (I focused in my thesis on C and N) all have more than one isotope and the isotopic composition of natural material can be measured with mass spectrometry. The isotopic composition is expressed as δ values (‰, parts per thousand differences from a standard): δX = [(Rsample/Rstandard)-1] * 1000, where X is the heavier isotope of the element (13C or 15N) and R is the isotopic ratio (13C/12C or 15N/14N) (Peterson and Frey 1987). The δ values measures the amount of heavy (13C and 15N) and light (12C and 14N) isotopes in a sample. Increases in δ indicate an increase in the amount of the heavy isotope. Standard references are for carbon international limestone standard Vienna PeeDee Belemnite (VPDB) (δ13C) and for nitrogen atmospheric N (δ15N). Samples enriched in the heavy isotope have a positive δ value whereas samples depleted in heavy isotopes relative to the standard have a negative δ value. The logic behind stable isotope analysis is that consumers typically reflect, with some predictable changes (fractionation), the isotopic com-position of their resources. In general, carbon isotope ratios tend to have much lower fractionation rates than nitrogen isotope ratios (McCutchan and others 2003, Vanderklift and Ponsard 2003). Carbon isotope ratios are clearly separated between marine and terrestrial plants, marine primary producers are more enriched (higher (=less negative) δ13C values) in the heavy 13C than terrestrial primary producers (lower (=more negative) δ13C values) (Fig. 8). Thus, consumers feeding on marine algae can be distinguished from consumers feeding on terrestrial plants by their higher δ13C signature. Nitrogen isotopes are also slightly enriched in ocean surface waters, but generally nitrogen isotopic ratios are more variable than carbon isotopic ratios. Consumers are typical enriched in 15N relative to their diet (McCutchan et al. 2003, Vanderklift and Ponsard 2003) (Fig. 8). This 15N enrichment is mainly a result of differential protein synthesis and excretion of the lighter isotope (Ponsard and Averbuch 1999). Thus, the tissues of marine consumer on a high tropic level, such as fish-eating seabirds, have high 15N values (Barrett et al. 2005) (Fig. 8). This 15N enrichment is further enhanced in seabird + guano through the fast mineralization of uric acid to ammonium (NH4 ), a chemical reaction that selectively removes 14N from guano (Mizutani et al. 1986, Wainright et al. 1998). Soils and plants from seabird affected areas, and consumers feeding on seabird tissue or on from seabirds fertilized plants, are therefore enriched in 15N (Anderson and Polis 1999, Wainright et al. 1998, Barrett et al. 2005). Thus, stable isotope values of nitrogen can be used to trace the fate of avian nitrogen on and around islands. In my thesis I combined δ13C and δ15N

33 signatures to trace the fate of avian N and to identify trophic pathways on and around cormorant islands. For stable isotope analysis, animals were freeze-dried, vascular plants and algae were oven-dried at 50°C for up to two weeks prior to analysis and sieved soils were air dried and finely ground. Plant and animal specimens were analyzed individually, but pooled samples were occasionally used for arthropod taxa with individual dry weights less than 0.7 mg. When possible, only legs and heads were used to avoid including gut contents. Stable isotope ratios for C and N of plant and animals samples were measured with an Isotope Ratio Mass Spectrometer type Europa integra. Stable isotope ratios of soils were analyzed on a ANCA-SL (Sercon United Kingdom) coupled to a Sercon 20-20 IRMS.

Figure 8. Isotopic signature (δ13C and δ15N) of terrestrial and brackish primary producers and consumers.

In paper I a single isotope (δ13C), two-source diet mixing model (IsoSource Version 1.3.1; Phillips and Gregg 2001) was used to estimate the relative proportion of marine and terrestrial carbon in detritivores and predators. In paper III a dual isotope (δ13C, δ15N), two-source diet-mixing model (Bayesian-mixing model) was used to estimated the relative proportion of Fucus and epiphytic algae in the diet of five aquatic invertebrates with the software MixSIR (Moore and Semmens 2008). Before applying the diet mixing models in paper III, we standardized the lipid content in the samples (both sources and consumers) with a mathematical normalisation technique (Post et al. 2007). Diet mixing models in paper III were analysed with both untreated δ13C ratios and lipid-normalized δ13C ratios because it remains unclear if this normalization is necessary (Post et al. 2007). In paper I

34 we did not correct for the variation in lipid content among organisms or tissue types because of the lack of reliable normalization techniques for terrestrial organism (Post et al. 2007), the time-consuming alternative lipid extraction method and the possibility of lipid extraction causing fractionation in δ15N (Pinnegar and Polunin 1999; Sotiropoulos and others 2004). In paper I we assumed a tropic shift of 0.4‰ for C (McCutchan et al. 2003). In paper III we calculated estimates of the carbon and nitrogen fractionation factors for each consumer by using the equations in Caut et al. (2009).

Fertilization experiment (paper IV)

For the fertilization experiments we used 225 potted plants, split into 9 treatments (= 25 plants per treatment). Each treatment was watered with a certain level of N (H4N2O3) and P (Na2HPO4*2H20) input. Newly hatched Galerucella larvae were placed on the plant (15 per plant). About 4-5 weeks later Galerucella pupae were collected from the pots.

CNP-analysis (paper II, III and IV)

Phosphorus content (%P, dry mass basis) was assayed using persulphate digestion and ascorbate-molybdate colorimetry (Clesceri et al. 1998). Nitrogen and carbon content (% N, % C, dry mass basis) was assayed by using Isotope Ratio Mass Spectrometer type Europa integra or an elemental analyzer. In this analysis, samples were oxidized and reduced to N2 and CO2, respectively, which were measured with a thermal conductivity detector and IR-detection.

Statistics

Terrestrial island ecosystem (paper I and II) We compared stable isotope signatures (δ13C and δ15N) of soil, terrestrial plants and arthropods, plant and soil nitrogen and phosphorus content, soil pH and moisture, vegetation cover, plant aboveground biomass, arthropod abundance. and plant and arthropod species richness between three island categories (active and abandoned cormorant islands and reference islands) in the free software R (version 2.10.0) using a one-way ANOVA, and Tukey HSD post-hoc tests, and also performed regressions with mean nest density as explanatory variable. Furthermore, we compared arthropod community compositions and species composition of plants, coleopterans, heteropterans, cursorial spiders and chironomids with np-MANOVA based on Bray-Curtis dissimilarities, using the adonis function of the package vegan for R (version 2.10.0, Oksanen 2009). We used 2-dimensional nonmetric multidimensional scaling (NMDS) plots to illustrate

35 the ordination of the islands. We identified indicator species for the three island categories with the help of indicator species analysis in PC-ORD (version5, McCune et al. 2002). Finally, we used structural equation modeling (SEM, sem package in R 2.10.0, Fox 2006) to separate the direct effect of mean nest density and the indirect effects via changes in the vegetation on island arthropods (Wright 1934, Mitchell 1992).

Near shore brackish water ecosystem (paper III) To investigate if algae and associated invertebrates near cormorant colonies incorporate ornithogenic N in their tissue, we compared stable isotope ratios (δ13C and δ15N) between the three groups of cormorant islands and the reference islands with linear mixed effects models. Linear mixed effects models were also used to test for differences in (a) the epiphytic algae to Fucus ratio, (b) nitrogen and phosphorus content of algae and invertebrate, (c) abundance and (d) biomass of the five invertebrate groups between reference islands and the three cormorant island categories.

Homeostasis (paper IV) We compared N:C, P:C, and N:P among herbivorous, detritivorous, and predatory arthropods with linear mixed effects models. Similarly, we compared N:C, P:C, and N:P of algae and brackish invertebrates between the three groups of cormorant islands and the reference islands with linear mixed effects models. To test for strict homeostasis we conducted regression analysis for N:C, P:C, and N:P for each arthropod groups using plants as resource for terrestrial herbivores, detritivores, and insect predators, algae (Fucus and epiphytic algae) as resource for aquatic consumers and chironomids as resource for spider.

Results and Discussion

Stable isotope analysis – hypothetical food web (paper I and III)

Web Staphylinidae Lycosidae Saldidae Coccinelidae Carabidae Formicidae Nabidae Chilopoda spiders 15.9 11.4 12.7 10.4 9.7 11.9 13.2 16.8 8.1

Herbivores Detritivores Scavengers Chironomidae Aquatic invertebrates Fungivores 11.9 (8.9-12.5) Collembola 12.6 Saprophagous 8.8 (adults) Gammarus 8.8 ? coleoptera 11.5: (larvae) Jaera 7.2 Isopoda 16.2 Brachycera 19.9 Idotea 8.1 Theodoxus 7.4

Plants 12.3 Soil food web Epiphytic Fungi algae ? ? Food Cormorant scrap bodies 11.3 Soil 12.3 Fucus 11.7 Guano

Terrestrial Brackish Low δ13C High δ13C

Ornithogenic N (enriched δ15N) pathway Ornithogenic N (enriched δ15N) and marine C (enriched δ13C)

Figure 9. Simplified arthropod food web on active cormorant islands with high nest density. Numbers indicate the increase in δ15N (Δ δ15N ‰) signature compared to reference islands. Dashed lines reflect the flow of ornithogenic N (= enriched δ15N signatures), dashed - double dotted lines reflect the flow of ornithogenic N (= enriched δ15N signatures) and marine C (= enriched δ13C signatures). Thick lines indicate main flows.

Based on the stable isotope analysis and diet mixing models we constructed a simplified food web for active cormorant islands with high nest density (Fig. 9). Most importantly, these analyses strongly indicated (1) that especially spiders feed largely on chironomids, (2) that ornithogenic N enters the terrestrial food web via several pathways and travels through the whole arthropod food web, and (3) that brackish algae and associated invertebrates incorporate ornithogenic N in their body near active cormorant islands with high nest density. In more detail, the δ13C signature clearly separated terrestrial plants and their primary consumers (herbivores) and secondary consumers (coccinelids) from marine epiphytic algae and their primary consumers (Chironomidae) and secondary consumers (web- building spiders) (Fig. 10). All other terrestrial consumers seemed to mix their diet and include both terrestrial and marine sources, directly like detritivores and

37 scavengers (isopods, collembolans, brachycerid dipterans, and saprophagous coleopterans) or indirectly by preying on these consumers (e.g. nabids, formicids, chilopods, carabids, staphylinids). The enriched δ15N signatures (Fig. 10) showed that all investigated plants and arthropods on active seabird islands contained some ornithogenic nitrogen. However, chironomids and web spiders incorporated, like brackish algae and their associated invertebrates, a significant amount of ornithogenic N in their tissue only on/nearby cormorant islands with high nest density (Figs. 10 and 11). On abandoned cormorant islands plants, herbivores, detritivores, and some insect predators showed increased δ15N, whereas brackish algae, chironomids, and spiders did not. This indicates that predators without increased δ15N signature (spiders and saldids) fed mainly on chironomids, whereas predators with increased δ15N fed at least partly on terrestrial prey. The diet mixing model suggested that the diet of some of the arthropod groups differed between the island categories. Collembolans, brachycerid dipterans, staphylinids, and saldids seemed to include more marine carbon in their diet on active cormorant islands than on reference islands, whereas web spiders, lycosids, and carabids proportional seemed to include less marine carbon in their diet. Web spiders, lycosids and nabids included the highest percentage of marine carbon in their diet on abandoned cormorant islands.

Terrestrial carbon source Marine carbon source 30 Terrestrial plants Epiphytic algae 25 Reference island 20 Abandoned cormorant island 15 Active cormorant island Low nest density High nest density 10

0 30 Herbivores Collembola Brachycera Isopoda Chironomidae 25 Sapro- Sa phages 20 Coleoptera N (‰) 15 15 δ 10

30 Coccinelidae Carabidae Chilopoda Formicidae Web spiders 25 Staphylinidae S Nabidae Lycosidae C Saldidae 20 F N L S S S W 15 L S F F W 10 N N S L F L 5 W W C S N S 0 -30 -25 -20 -15 -30 -25 -20 -15 -30 -25 -20 -15 -30 -25 -20 -15 -30 -25 -20 -15

δ13C (‰)

Figure 10. Bi-plot of δ15N and δ13C for brackish algae, terrestrial plants, and arthropods on and around islands without a cormorant colony (= reference islands), with an abandoned cormorant colony and with an active cormorant colony (means ± SE). For some groups a subdivision of active cormorant islands into islands with a low and high nest density improved the ANOVA model.

20 Reference islands C F Cormorant islands Abandoned 15 E T I Active with low nest density J G Active with high nest density N 10 F F Fucus vesiculosus 13 T δ C E E Epiphytic algae J G I F 5 C J T C Chironomidae E G T I F T Theodoxus fluviatilis J I G Gammarus spp. C E G 0 I Idotea spp. -21 -14 -7 J Jaera albifrons δ13 C

Figure 11. Bi-plot of δ15N and δ13C for brackish algae and invertebrates nearby islands without a cormorant colony (= reference islands), with an abandoned cormorant colony and with an active cormorant colony with low and high nest density (means).

Effects of nesting cormorant on island ecosystem

We found that cormorant nesting colonies affected soil, terrestrial vegetation and arthropods, brackish algae and grazer (Figs. 12 and 13)

Effect of nesting cormorants on soil chemistry

+ - Soil nitrogen content (NH4 and NO3 ) was highest on active cormorant islands and about equal on reference and abandoned cormorant islands (F = 31.7, p < 0.001, r2 = 87.2 and F = 19.1, p = 0.001, r2 = 80.1). Soil P content was lowest on reference islands and about equal on active and abandoned cormorant islands (F = 42.1, p < 0.001, r2 = 90.1). Soil pH and soil moisture did not differ between the three island categories.

Figure 12. Effects of active nesting cormorant colonies on terrestrial and near-shore ecosystems. Numbers indicate x-fold change. ↑ increase and ↓ decrease. A: Abundance, S: Species richness, B: Biomass, N: percentage nitrogen content, P: percentage phosphorus content, Fu:EP: Epiphytic algae: Fucus ratio Shown are only significant effects (p < 0.100).

Figure 13. Effects of abandoned nesting cormorant colonies on terrestrial and near-shore ecosystems. Numbers indicate x-fold change. ↑ increase and ↓ decrease. A: Abundance, S: Species richness, B: Biomass, N: percentage nitrogen content, P: percentage phosphorus content, VC: Vegetation cover, E: Evenness. Shown are only significant effects (p < 0.010).

Effect of nesting cormorants on island vegetation (paper I, II and IV)

According to our assumption, cormorant nesting colonies strongly affected island vegetation. The total plant species richness, the mean species richness/m2 (excluding grasses), and the vegetation cover were all lower on active than on reference islands, whereas the plant N and P contents were higher on active cormorant islands (Figs. 12 and 14). Abandoned cormorant islands differed neither in vegetation cover nor in species richness from reference islands, but their evenness was lower than on active and reference islands, whereas the plant aboveground biomass tended to be higher on abandoned islands than on reference islands (Figs. 13 and 14). Furthermore, vegetation cover, the total plant species richness, the mean plant species richness/m2, and the plant nitrogen content decreased with nest density. In contrast, plant biomass was negatively correlated with nest density solely when including only cormorant islands in the analysis. Active cormorant islands differ from reference islands not only in plant species richness but also in species composition (Fig. 16). Active cormorant islands were characterised by a few dominant nitrogen tolerant species like Galeopsis speciosa, Urtica dioica, Tripleurospermum martimum, and Phalaris arundinacea. Common on all active cormorant islands were Spergula arvensis and Rumex crispus, whereas Orchidaceae and Ranunculaceae were absent. The displayed responses of island vegetation to nesting cormorants were fairly similar to results from fertilization experiments with decreased plant species richness and changed plant species composition towards a few nutrient tolerant species and increased plant nutrient content (e.g., Siemann 1998, Anderson and Polis 1999, Haddad et al. 2000). But in contrast to fertilization experiments, on cormorant islands primary production was not increased and vegetation cover was decreased, due to nutrient toxication (Ellis 2005, Ellis et al. 2006).

A) B) D) F = 4.8 p = 0.024 F = 3.9 p = 0.041 C) F = 5.5 p = 0.015 F = 11.9 p < 0.001 a a a

2 a ab ab 45 ab a

b 3 b 40 50 60 Evenness

richness b richness/m

b coverVegetation (sqrt-transformed) Mean plant speciesMean plant 30 Total plant species Total plant species 0.7 0.8 0.9 012 0.20.61.01.4 RF AB CO RF AB CO RF AB CO RF AB CO E) F = 2.8 p = 0.091 F) F = 24.1 p < 0.001 G) F = 25.7 p < 0.001 b 3.0 a a

a 4.0 5.0

1.0 1.5 2.0 2.5 a (log-transformed) Aboveground biomassAboveground Plant nitrogen content nitrogen Plant 0.2 0.3 0.4 0.5 0.6 2.0 3.0 0.5 Plant phosphorus content phosphorusPlant RF AB CO RF AB CO RF AB CO

Figure 14. Responses of island vegetation to cormorant nesting colonies. In the calculation of (B) and (C), grasses were excluded. Shown are median values of the response variable with upper box showing 75th percentile, lower box showing 25th percentile and whiskers showing values within the range of 1.5 interquartile distances. F and p were derived from one-way ANOVA, with significant differences in bold. Different letters indicate significant differences (p < 0.05).

Effect of nesting cormorants on island arthropods (paper I and II)

The np-MANOVA showed that the arthropod community composition differed between island categories (total df = 17, Pseudo-F = 4.2, p = 0.001, r² = 34.6) (Fig. 15). Some but not all investigated arthropod groups responded to cormorant colonies with changes in abundance, species richness and/or species composition.

Figure 15. 2-dimensional NMDS of the arthropod community.

Density response (paper I and II) Among the major trophic groups, collected by suction sampling (herbivores, detritivores, scavengers, fungivores, omnivores, and predators), cormorant colonies only affected the total density of scavengers and predators per unit vegetated area. Scavenger densities were 9 times higher on active cormorant islands than on reference islands and were about equal when comparing abandoned and reference islands. Predator densities were 3 times higher on abandoned than on reference islands but did not differ between reference and active cormorant islands. The positive density responses of scavengers and predators were also reflected in their mass response (= total wet weight per unit area). Moreover, herbivores, in contrast to their density response, had 3 times higher mass on active cormorant islands than on reference islands. In suction samples, the density of 11 of 18 arthropod groups differed between island categories (Figs. 12 and 13, Tab. 3). Six groups had higher (herbivores: lepidopteran larvae and aphids, scavengers: brachycerid dipterans and saprophagous coleopterans, predators: parasitic hymenopterans and aphidophages) and two had lower (collembolans and lycosids) densities on active cormorant than on reference islands. On abandoned cormorant islands, two groups (lepidopteran larvae and web spiders) had higher densities than on reference islands and one group (herbivorous coleopterans) had higher density than on active cormorant islands. Chironomids responded to active cormorant colonies with increased densities only if the colonies had a high nest density. For pitfall trap samples, neither the total abundance of coleopterans nor the total abundance of spiders responded to cormorant nesting colonies. The division of coleopterans in major feeding groups (herbivores, fungivores, algaevores, saprophages and predators) showed that herbivores tended to have lower abundance on active cormorant islands than on reference islands, whereas fungivores and saprophages were more abundant on active than on reference islands (Fig. 12, Tab. 3). Chironomids showed no difference in abundance between the island categories but increased with nest density.

Response in species richness (paper II) Of the five investigated groups (coleopterans and heteropterans collected by suction sampling, coleopterans and cursorial spiders caught in pitfall traps, and chironomids caught with sweep netting) showed no significant difference in the observed species richness among island categories; but spiders tended to have lower observed species richness on active cormorant islands than on reference islands (Fig. 12, Tab. 3) Heteropterans tented to have higher observed species richness on abandoned islands than on reference islands (Fig. 13). The estimated species richness of coleopterans in pitfall trap samples was higher whereas the estimated species richness of spiders was lower on active cormorant islands than on reference islands. For suction samples, the subdivision into feeding groups showed that the highest species richness of herbivorous heteropterans and fungivores coleopterans was found on abandoned islands (Fig. 13, Tab. 3). For pitfall trap samples, the subdivision showed that fungivorous and saprophagous

44 coleopterans had higher numbers of species on active cormorant islands than on reference islands (Fig. 12, Tab. 3). The species richness of herbivorous coleopterans differed not between the islands categories but decreased, as the species richness of spiders, with increasing nest density.

Table 3. One-way ANOVA of mean abundance (log-transformed), observed and estimated species richness of arthropods on active (AO) and abandoned (AB) cormorant islands with reference (RF) islands.

Taxa df, Error df F p adj. r2 RF AB AO

DVac

Herbivores

Aphidoidae Abundance 2, 16 4.7 0.024 29.4 X = X > Y

Lepidoptera larvae Abundance 2, 16 20.0 0.000 67.8 X >Y = Y

Cicadina Abundance 2, 16 n.s. X = X =X

Abundance 6.2 0.010 36.7 Coleoptera 2, 16 XY= X > Y Ob. sp. rich. n.s

Abundance n.s. n.s. Heteroptera 2, 16 X < Y > X Ob. sp. rich. 4.8 0.022 29.9

Detritivores/Fungivores

Collembola Abundance 2, 16 7.1 0.006 40.5 X = XY > Y

Isopoda Abundance 2, 16 n.s. X = X = X

Abundance 2.8 0.090 16.7 X = X = X Coleoptera 2, 16 Ob. sp. rich. 6.7 0.008 38.7 X < Y =XY

Scavengers

Brachycera (Diptera) Abundance 2, 16 19.5 0.000 67.3 X = X < Y

Abundance 10.7 0.001 51.8 X = X < Y Coleoptera 2, 16 Ob. sp. rich. 4.3 0.032 26.9 X = XY < Y

Predators

Web spiders Abundance 2, 16 4.9 0.021 30.5 X < Y = XY

Lycosidae Abundance 2, 16 11.2 0.001 53.1 X = X > Y

Other spiders Abundance 2, 16 n.s. X = X = X

Aphidophaga Abundance 2, 16 9.2 0.002 47.6 X = X < Y

Parasitic Hymenoptera Abundance 2, 16 5.3 0.017 32.5 X = XY < Y

Abundance n.s. X = X = X Coleoptera 2, 16 Ob. sp. rich. n.s. X = X = X

Carabidae Abundance 2, 16 n.s. X = X = X

Staphilinidae Abundance 2, 16 n.s. X = X = X

Abundance n.s. X = X = X Heteroptera 2, 16 Ob. sp. rich. n.s. X = X = X

Chilopoda Abundance 2, 16 n.s. X = X = X

Chironomidae * Abundance 3, 15 3.2 0.052 27.3 X = X = X

Pitfall traps

Abundance n.s. X = X = X Coleoptera Ob. sp. rich. 2, 11 n.s. X = X = X

Est. sp. rich. 4.2 0.044 32.9 X = XY < Y

Abundance 3.0 0.093 23.3 X = X = X Herbivorous coleoptera 2, 11 Ob. sp. rich. n.s. X = X = X

Abundance 9.9 0.003 57.9 X = XY < Y Fungivorous coleoptera 2, 11 Ob. sp. rich. 6.9 0.011 47.7 X = XY < Y

Abundance n.s. X = X = X Algaevorous coleoptera 2, 11 Ob. sp. rich. n.s. X = X = X

Abundance 5.1 0.027 38.7 X = XY < Y Saprophagous coleoptera 2, 11 Ob. sp. rich. 7.4 0.009 49.5 X = X < Y

Abundance* n.s. Predatory coleoptera 2, 11 X = X = X Ob. sp. rich. n.s. X = X = X

Abundance n.s. X = X = X Araneae Ob. sp. rich. 2, 11 3.4 0.072 26.8 X = X = X Est. sp. rich. 4.4 0.039 34.6 X = XY > Y

Sweep-netting

Abundance n.s. X = X = X Chironomidae Ob. sp. rich. 2, 9 n.s. X = X = X Est. sp. rich. n.s. X = X = X

* active cormorant islands subdivided into islands with low and high nest density

Species composition (paper II) The species composition of coleopterans and spiders, but not of heteropterans and chironomids, differed between the island categories (Fig. 16). Active cormorant islands were characterised by the saprophagous families Nitidulidae and Silphidae, the mainly aphidophagous family Coccinelidae and fungivorous species. Reference islands were characterised by the spider family Tetragnathidae and the lycosids Alopecosa pulverulenta and Pardosa lugubris.

Figure 16. 2-dimensional NMDS of plants and arthropod species.

Direct versus indirect effects of nesting cormorants (paper II) Island consumers might have responded directly to cormorants by utilizing their dead and living bodies, food scrap, and eggshells (Polis and Hurd 1995, Sanchez- Pinero and Polis 2000) or indirectly by responding to the described changes in the vegetation (Barrett et al. 2005) or to the changes in soil chemistry, structure and subsequent changes in microfauna and -flora (Fukami et al. 2006). Thereby changes at the level of primary consumer can cascade up the food web and affect consumers on higher trophic levels (Hunter and Price 1992, Polis and Hurd 1995, Barrett et al. 2005). In order to separate direct and indirect effects of nesting cormorants I used structural equation analysis, including mean nest density and three vegetation variables (vegetation cover, plant biomass, and plant species richness). Unfortunately, I was not able to include further important factors in the analysis like plant and soil N and P content, plant species composition, fungal biomass, and amount of carcass. Plant N and P content are assumed to limit herbivorous and detritivorous insects (White 1993, Cross et al. 2003) and thus I expected to find high herbivore and detritivores densities on active cormorant islands. This expectation was only partly met by my results. I found only aphids and lepidopteran larvae to have increased densities on active cormorant islands, indicating that other vegetation variables also play an important role in determining consumer abundance. Accordingly, structural equation analysis indicated that the abundance of lepidopteran larvae, auchenorrhynchans, herbivorous, fungivorous and predatory coleopterans and web spiders increased with plant biomass (Figs. 17 and 18) consistent with results from fertilization experiments (Siemann 1988, Haddad et al. 2000). Furthermore, plant biomass mainly had a positive effect on consumer species richness supporting the hypothesis of a positive relationship between consumer species richness and plant productivity (Waide et al. 1999, Mittelbach et al. 2001). Also vegetation cover had strong effects on the abundance and species richness of several beetle feeding groups, showing that vegetation cover, may determine the assemblage of epigeal coleopterans (Frank and Reichhart 2004, Eyre and Luff 2005). In contrast to previous studies, effects of increased cover on beetle abundance and species richness were generally negative, except for herbivorous beetles, perhaps due to factors correlated with vegetation cover but not included in the models (e.g., soil chemistry). One surprising results from the structural equation analysis was that plant species richness had, despite the positive effect on the species richness of predatory coleopterans, no or negative effect on consumer species richness. This contrasts with the habitat heterogeneity hypothesis (MacArthur and MacArthur1961) and with previous empirical studies (Siemann et al. 1998, Haddad et al. 2000). However, the apparent effect of plant species richness on predatory coleopteran and spider species richness and web spider abundance was more expected and supported the view of strong importance of structural and species diversity of the vegetation for arthropod diversity (Dennis et al. 1998, Perner et al. 2003, Garcia et al. 2009). Finally, the structural equation analysis indicated that cormorants also directly affected several groups; positively (saprophagous beetles, predatory coleopterans and web spiders) and negatively

(algaevorous coleopteran). Interestingly, the model showed a postive correlation of brachycerid flies and parasitic hymen-opterans indicating a bottom-up control. One weakness in my study is that I did not include the soil food web, making it difficult to understand the response of some consumer groups. For example, it was very surprising to find the highest density of fungivorous beetles on active cormorant islands, because former studies show that high nitrogen loads due to nesting seabirds negatively affect soil fungal biomass (Wright et al. 2010). Similarly, the negative effect of cormorants on the lycosid density and on species richness of cursorial spiders was suprising because the increased abundances of scavengers and bird parasites in other seabird island systems were shown to increase spider densities (Polis and Hurd 1995). One possible explanation might be changes in soil chemisty and surface structure (Bultman and Uetz 1984, Hurd and Fagan 1992).

A) Abundance of arthropods in the herb layer (except coleopterans and heteropterans) (Dvac) X² = 27.1, df = 33, P = 0.76 Path Coefficient <0.4 0.4-0.5 Lepidopteran 0.5-0.6 larvae 0.6-0.7 0.7-1.0 Vegetation 0.56** >1.1 cover Auchchenorrhyncha 0.52*

-0.81*** 0.80*** Plant biomass -0.47** Brachycera

0.58*** Mean nest Parasitica density -0.51* Plant species richness 0.41* Aphidophaga 0.46* 0.46*

Web spiders

-0.81*** 0.80*** Plant biomass -0.47** Brachycera

0.58*** Mean nest Parasitica density -0.51* Plant species richness 0.41* Aphidophaga 0.46* 0.46*

Web spiders

C) Abundance of coleopterans and heteropterans (D-vac)

Path Coefficient Χ² = 35.0, df = 31, p = 0.29 < 0.4 0.4–0.5 Herbivorous 0.5–0.6 0.60*** coleoptera 0.6–0.7 0.7–1.0 Vegetation 0.52*** > 1.1 cover Fungivorous 0.74*** coleoptera –0.51*** –0.81*** Plant biomass Saprophagous 0.39* Mean nest –0.41** coleoptera density –0.51* Plant species 0.63** richness 0.41* Predatory coleoptera

Predatory heteroptera

D) Species richness of coleopterans and heteropterans (D-vac )

Path Coefficient Χ² = 19.3, df = 26, P = 0.830 < 0.4 0.4–0.5 Herbivorous 0.5–0.6 0.56** coleoptera 0.6–0.7 Vegetation 0.7–1.0 cover 0.31* > 1.1 Fungivorous 0.58*** coleoptera –0.83*** Plant biomass –0.56** 0.43* Mean nest Predatory –0.51* density Plant species 0.39 coleoptera richness 0.51* Herbivorous heteroptera

Predatory heteroptera

Figure 17. Path diagram for effects of cormorant nest density on vegetation variables and arthropods from suction sampling on 19 islands. Approximate path coefficients are indicated by thickness of arrows. Solid lines indicate positive relationships, broken lines negative relationships. Asterisks (*) indicate the level of significance: p < 0.05*, p < 0.01**, p < 0.001***.

A) Abundance (pitfall traps) Herbivorous coleoptera X2 = 21.3, df = 25, p = 0.67 –0.67** –0.63** Fungivorous –1.11*** coleoptera 0.42*** 0.49*** Vegetation –0.92** Algaevorous –0.78*** cover coleoptera 0.70** –0.78* Mean nest 0.37* Plant biomass 0.61* density –0.71** 0.86*** Saprophagous Plant species coleoptera richness –0.52*** 0.74*** –1.3*** Path Coefficient < 0.4 1.42*** 0.4–0.5 Predatory 0.5–0.6 coleoptera 0.6–0.7 0.7–1.0 1.0*** > 1.1 Araneae

B) Species richness (pitfall traps) Herbivorous X2 = 21.5, df = 28, p = 0.80 coleoptera 0.47 –0.84*** Fungivorous Vegetation –0.75* coleoptera cover 0.36* –0.87*** 0.35* –0.78*** Plant biomass 0.50* Algaevorous –1.11*** Mean nest coleoptera density –0.71*** Plant species –0.43 Saprophagous richness 1.08*** coleoptera 0.43* Path Coefficient 1.12*** < 0.4 0.4–0.5 Predatory 0.5–0.6 0.41* coleoptera 0.6–0.7 0.7–1.0 Araneae >1.1

Figure 18. Path diagram for effects of cormorant nest density on vegetation variables and Coleopteran and Araneae from pitfall trapping on 13. Approximate path coefficients are indicated by thickness of arrows. Solid lines indicate positive relationships, broken lines negative relationships. Asterisks (*) indicate the level of significance: p < 0.05*, p < 0.01**, p < 0.001***.

Effects of cormorant nesting colonies on algae and aquatic invertebrates in coastal waters (paper III)

The effects of nutrient input by nesting cormorants were not limited to the terrestrial system, but also included adjacent aquatic systems, or more precisely the Fucus-epiphytic algae-invertebrate complex in near shore waters (Fig. 12). These effects were only visible around active cormorant islands with high nest density. Accordingly, algae nearby cormorant islands had enriched δ15N signatures only when the colony had a high nest density. This indicates that significant nutrient runoff from cormorant islands into surrounding waters only occur if the nest density exceeds a certain level. On islands with low nest densities, vegetation takes up most ornithogenic N and P in plant biomass. The nutrient load on cormorant islands with high nest densities, especially the ammonia load, which is toxic, suppressed plant growth and caused a decreased vegetation cover. The nutrient input to the terrestrial systems exceeded the amount that can be bound in soil and vegetation and notable amounts of nutrients run off into the sea. Thus, primary producers around high nest density islands responded as in previous fertilization and mesocosm studies, with an increased epiphytic algae load (Worm et al. 2000, Råberg and Kautsky 2008) (Fig. 19). Furthermore, the N and P contents of algae near active cormorant islands with high nest density were, similar to terrestrial plants, 1.6 and 1.5 higher than near reference islands (Fig. 19). Contrary to previous studies from the Baltic Sea (Kotta et al. 2000, Worm and Sommer 2000, Kraufvelin et al. 2006, 2007) and Lake Michigan, US (Blumenshine et al. 1997), not all invertebrates responded with increased abundance. Three of five invertebrate taxa (Jaera albifrons, Gammarus spp., and Chironomidae) showed increased biomass near islands with high nest density, but only J. albifrons increased in abundance compared to reference islands (Fig. 20). The lacking response of Idotea spp. and T. fluviatilis may be the role of Fucus as a habitat, providing shelter against disturbance or predation (Orav-Korra and Kotta 2004, Wikström and Kautsky 2007, Råberg and Kautsky 2007). The large Idotea spp. may be more dependent on Fucus as a shelter than the smaller J. albifrons, Gammarus spp., and Chironomidae and thus suffer from the relative decrease of Fucus. Furthermore, the diet mixing model indicated that Idotea spp. did not prefer epiphytic algae as food source as strongly as the other invertebrates and had a relatively higher Fucus consumption. T. fluviatilis might similarly suffer from decreased shelter but also from the increased filamentous algae load that competes with microalgae, an important food source for the scraping snails (Råberg and Kautsky 2008).

A) Fucus vesiculosus Epiphytic algae

0.25 * 0.10 * %P 0.06 0.10 0.15 0.20 0.02 RF AB COL COH RF AB COL COH B) Fucus vesiculosus Epiphytic algae 5 5

** 34 * 34 %N 2 01 012 RF AB COL COH RF AB COL COH C) 1.5 RF Reference Islands ratio Cormorant islands 1.0 AB Abandoned ** COL Active with low nest density Fucus

Epiphytic algae- COH Active with high nest density 0.0 0.5 RF AB COL COH

Figure 19. Nitrogen (A) and phosphorus (B) algal content and epiphytic algae:Fucus ratio (C) in near-shore waters of four island categories. Shown are median values of the response variable with upper box showing 75th percentile, lower box showing 25th percentile and whiskers showing values within the range of 1.5 interquartile distances. (•) P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001 refer to level of significances relative to reference islands.

Chironomidae Theodoxus fluviatilis

48 100 (•)

** 36 0416 0149 Gammarus spp. Idotea spp. 144 49 48 ** 1 016 0 Jaera albifrons RF AB COL COH

Invertebrate biomass (mg dry-weight per g algae) g per dry-weight (mg biomass Invertebrate RF Reference Islands Cormorant islands ** AB Abandoned COL Active with low nest density COH Active with high nest density 0 0.25 1.0 2.25 RF AB COL COH

Figure 20. Biomass (sqrt-transformed) of marine invertebrates (number of individuals per gram dry-weight algae) collected in near shore waters of four islands. Shown are median values of the response variable with upper box showing 75th percentile, lower box showing 25th percentile and whiskers showing values within the range of 1.5 interquartile distances (•) P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001 refer to level of significances relative to reference islands

My results in context of other seabird island studies

The effect of nesting seabirds on island vegetation has been intensively studied and my results are, generally, consistent with previously studies (see metaanalysis by Ellis et al. 2011, see metaanalysis by Mulder et al. 2011). However, not much work has been done on the effects of nesting seabirds on island consumers and the majority of studies have been concentrated in two regions: the Gulf of California, Mexico and New Zealand. The intensive studies in these regions give good insights on processes in these systems, but not about the effect of nesting seabirds on island consumers in general. Both systems are unique; the Gulf of Mexico with its hot-dry climate with irregular rain-pulses (El Niño Southern Oscillation) and

New Zealand with burrowing seabirds. Therefore, I find my study to be an important contribution on the effects of nesting seabirds in other regions. The majority of studies on seabird island consumers found higher consumer abundance, survival and reproduction rates (reviewed by Kolb et al. 2011). This indicates that seabird islands are mainly bottom-up regulated systems with consumers being positively affected by enhanced resources from seabirds. However, several biotic and abiotic factors, like life history traits of seabird taxa, timing of seabird inputs, nesting density, history of seabird use of islands, climate and seasonality, inputs of other marine subsidies, trophic cascades, and apparent com-petition are likely to affect the magnitude or even direction of consumer responses. In my study systems the lack of positive numeric response might be partly caused by the lack of food limitation of the consumers, due to high amounts of shore drifted algae and high abundances of chironomids. These highly abundant food sources may shift consumer population from being limited by food to being limited by habitat. This might be the case for spiders in our systems, for examples the high abundance of web spiders on abandoned cormorant islands might be explained by the increased biomass and heterogeneous vegetation and thus increased substrate for web-building (Miyashita and Takada 2007, Chan et al. 2009). Another possible explanation for lacking positive consumer response might be top-down control. Food webs are rarely simple and linear, and the real complexity of food webs, including competition and differential predation can complicate the visible responses of consumers to seabirds. For example, in an analysis of belowground community changes on seabird islands in New Zealand, herbivorous nematodes and secondary consumers increased in contrast to primary consumer microflora compared with non-bird islands (Fukami et al. 2006). The most likely explanation for this observation is that the increase in secondary consumers may have overwhelmed the bottom-up benefits the primary consumers received from increased nutrient availability. Generally, if predator interactions are very strong there might be no increase in abundance of lower level consumers. In order to draw general conclusions about the effects of nesting seabirds on island consumers further research is needed, including controlled studies on the impact of seabird nesting density on consumer with different life history traits. To be able to make prediction of the role of abiotic factors, like climate, more compressive studies in different climate regions are necessary. Compared with terrestrial systems, relatively little research has been done on effects of seabird colonies on adjacent aquatic, especially marine, systems. My study system differs in three major points from earlier studies. First, the Baltic Sea has no real tides, second the water is brackish, and third the relatively small size of colonies. Despite these differences I found, in accordance with the majority of former studies, increased algae nutrient content, increase growth of ephemeral algae near seabird islands and invertebrate responses. However, my study is one of the few studies that used stable isotope analysis and thus could determine the ornithogenic origin of the nitrogen with great certainty. Generally, relative little consensus exists about the types and magnitudes of seabird effects on marine

55 systems because the systems in such studies vary climatically, in structure, sampling methods used, and response variable used (reviewed by Young et al. 2011). Thus more studies on the near-shore communities of seabird islands that use stable isotope analysis are needed to be able to draw conclusions about the general pattern of seabird effects on near-shore communities.

Plant and consumer stoichiometry and homeostasis under nutrient enrichment

Effects of fertilization on plant and herbivore homeostasis and herbivore survival and body size – the Lythrum salicaria – Galerucella spp. system (paper IV)

The outcome of the fertilization experiment supported the in ecological stoichiometry common, simplified assumption that autotrophs (plants) are stoichiometrically plastic while heterotrophs (animals) are strictly homeostatic (Sterner et al. 1992, Sterner and Elser 2002). The elemental composition of Lythrum salicaria varied strongly in accordance with the nutrient treatment (Fig. 21). Generally, plants fertilized with N and/or P had higher N:C and/or P:C ratios than unfertilized plants. Plant N:C varied between the treatment up to 8.3 fold, plant P:C up to 6 fold . In contrast, Galerucella spp., generally, strongly regulated their stoichiometry, showing strong N:C and N:P homeostasis and strict P:C homeostasis (1/HN:C≈ 0.08; 1/HN:P ≈ 0.06) (Fig. 21). However, contrary to the general assumption that herbivores are nutrient limited (White 1993, Cross et al. 2003) and findings of previous studies that increased nutrient plant content positively affect herbivore growth, survival and reproduction (Mattson 1080, Joern and Behmer 1997, Frost and Elser 2002) we found no strong effects of fertilization treatment on Galerucella survival and mean larvae dry-mass.

(A) (B) -0.6

) 0.16

0.12 N:C) -1

0.08 L. salicaria L. -1.4 0.04 N:C ( (Galerucella

0 -1.8 n np nP N Np NP p P no Log -1.8 -1.4 -1 -0.6 Log (L. salicaria N:C) -0.5 0.016 ) )

P:C) -1 0.012

0.008 -1.5 L. salicaria L. 0.004 -2 (Galerucella P:C ( 0 -2.5 Log n np nP N Np NP p P no

-2.5 -2 -1.5 -1 -0.5 Log (L. salicaria P:C) 60 1.8 ) 50

40 N:P) 1.4 30 1 20

L. salicaria N:P ( L. 10 Galerucella 0.6 0

n np nP N Np NP p P no Log ( 0.2 Fertilizer 0.2 0.6 1 1.4 1.8 Log (L. salicaria N:P)

Figure 21. Elemental ratios (N:C, P:C, N:P) (mean ± SE) of Lythrum salicaria under different fertilization (A) and Regressions between Galerucella spp. and L. salicaria elemental ratios (B).

Ecological stoichiometry and homeostasis of plants and invertebrates on and nearby heavily fertilized cormorant nesting islands (paper IV)

This study investigated the stoichiometry and homeostasis of primary producers and invertebrates, both on and around cormorant and non-cormorant islands. I then looked for causal relationships between stoichiometry and the level of homeostasis of a taxonomic group, respectively, and observed numeric responses on cormorant islands. I found in accordance to principal stoichiometric theories that heterotrophs, generally, were in contrast to stoichiometrically plastic autotrophs, strongly homeostatic (e.g.; Sterner and Elser 2002). All primary producers (terrestrial plants, brackish macroalgae, and epiphytic algae) on or nearby nutrient rich cormorant islands contained higher N:C and P:C ratios in their tissue than on or nearby reference islands, while all herbivores, detritivores, and predators, except spiders, kept their elemental body ratios strongly homeostatic (Fig. 22). A potential weakness is that I tested for homeostasis mainly within families or orders

57 and not within species since I was unable to sample the same species across all islands. Within families or orders, apparent homeostasis may not only be caused by stoichiometric plasticity of species but also by shifts in species composition from the dominance of nutrient poor species reference islands towards the dominance of nutrient rich species on cormorant islands (Elser et al. 1996, Urabe et al. 2002, DeMott and Pape 2005). However, while my focus on taxonomic levels higher than species may reduce our ability to deduce mechanism from pattern, effects on higher trophic levels should be similar, at least for generalist consumers. In accordance with previous studies (Mattson 1980, Strong et al. 1984, White 1993, Fagan et al. 2002), terrestrial herbivores and detritivores and brackish grazers generally fed on diets containing less nutrients than themselves, while only predators specialized on nutrient-poor preys (e.g., Coccinelidae) consumed nutria- ent-poor diet, relative to their tissue chemistry. Thus, I expected herbivores and detritivores to respond positively to increased plant nutrient content on cormorant islands (e.g., Slansky and Feeny 1977, White 1993, Joern and Behmer 1997, Cross et al. 2003). The numeric responses observed (paper I and III) partly supported this assumption but also detected large variation among taxonomic groups. According to stoichiometric theory and strong empirical support, mainly from aquatic systems, the nutrient demand of a given taxon is determined by its nutrient content and relative growth rate (Elser et al. 1996, Sterner and Elser 2002). Taxa (or specific life stage) with high body N and P content and high growth rates have a high nutrient demand to maintain growth rate and consequently secondary productivity and thus are more sensitive to nutrient limitation than nutrient poor, slow growing taxa (Urabe and Watanable 1992, Elser et al. 1996, Urabe et al. 2002). Hence we hypothesised that within a trophic group nutrient rich consumer groups will be more favoured by increased nutrient availability on cormorant islands and thus display a stronger positive response on these islands than nutrient poor species. However, in contrast to our hypothesis the nutrient content of herbivores and detritivores explained only numeric responses observed in one group; lepidopteran larvae had the highest P:C among terrestrial insects and had, as expected, higher densities on islands with high plant P content (= cormorant islands) than on islands with low plant P content (= reference islands). Thus, our data indicates that species traits other than nutrient content, such as e.g., conversion efficiency, respiration rates, host plant obligations, diet specialization, and habitat specialization (Hall et al. 2004, Moretti and Legg 2009) may mainly determine the success of a taxon in a certain environment. Likewise, in contrast to our assumption, variation in homeostasis among taxa was seemingly a poor predictor of numerical responses. Furthermore, since we found all terrestrial herbivores and detritivores to strongly regulate their stoichiometry, increased abundances of higher consumers (e.g., Coccinelidae and parasitic hymenopterans) were not caused by increases in prey (e.g., aphids, lepidopterans) quality but quantities.

(A) Terrestrial arthropods (B) Brackish invertebrates -0.4

-0.6

-0.8

-1

-1.2

-1.4 Log (ConsumerLog N:C) -1.6

-1.8 Herbivores/ Detritivores Predators -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 Log (Resource N:C) Log (Resource N:C) -0.5

-1

-1.5

-2 log (Consumer P:C) -2.5 Herbivores/ Detritivores Predators

-2.5 -2 -1.5 -1 -0.5 -2.5 -2 -1.5 -1 -0.5 Log (Resource P:C) Log (Resource P:C)

1.6

1.4

1.2

0.8

0.6 Log (Consumer N:P) 0.4 Herbivores/ 0.2 Detritivores Predators

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Log (Resource N:P) Log (Resource N:P)

Figure 22. 3 Regressions between consumer and resource N:C, P:C and N:P for (A) terrestrial arthropods from active, abandoned and non-cormorant and (B) brackish invertebrates nearby active, abandoned and non-cormorant. Black lines indicate regressions with p<0.1 (plastic), grey lines indicate regressions with p>0.1 (strictly homeostatic), drawn lines indicate herbivores and detritivores, and dashed lines indicate predators.

Concluding remarks

Visiting a cormorant nesting island, it becomes pretty obvious that nesting cormorants strongly affect the vegetation of their nesting islands. After a few years of colonization some or most trees have died due to ammoniac poisoning. The plant species richness and vegetation cover was decreased and plant species composition has changed. These effects on island vegetation are especially pronounced on islands with high nest density. Such islands cause some people to claim that cormorants kill their nesting islands or even the whole archipelago. Well, I found none of my study islands to be dead. The vegetation on abandoned cormorant islands were in terms of species richness similar to non-cormorant islands, but produced more biomass. Even on active cormorant islands with high nest densities I found living plants and all cormorant islands harbored a variety of arthropods. Some arthropod groups (e.g., lepidopteran larvae, herbivorous beetles) responded to the modifications in vegetation structure or nutrient content, other more ground living taxa (e.g., collembolans, fungivores beetles and cursorial spiders) were seemingly affected by alterations in soil chemistry and structure, while further groups (e.g., brachycerid flies, saprophagous beetles) responded directly to cormorant subsidies. Consumer responses included changes in species composition and both increases and decreases in abundance and species richness; however, not all taxonomic or feeding groups responded to the presence of cormorants. Differences in consumer nutrient content poorly explained numeric response observed among taxonomic group. Similarly, no causal relationship between the level of homeostasis and displayed numeric response could be found, since consumers, generally, kept their stoichiometry strongly homeostatic. This strong consumer homeostasis evidenced that bottom-up effects of cormorant nutrients on higher trophic consumer acted not via increases in prey quality but quantity. Thus, I found two likely bottom-up cascades, one via guano – increased plant nutrient content – increased aphid density – increased aphidophage density and the other via seabird carcasses and food scrap – increased brachycerid fly density – increased parasitic hymenoptera density. When evaluating the impact of cormorant nesting colonies on species diversity one should distinguish between species diversity on the nesting islands and species diversity in the whole archipelago. Even if cormorants decrease the species richness of plants and some invertebrate groups on nesting islands, they also change the species compositions and thus likely the beta-diversity (= species turnover) among islands. Furthermore, active and abandoned cormorant islands may increase the habitat heterogeneity in the archipelago by adding islands different in plant species composition, soil chemistry, resource availability, and portion of dead wood. Such increased habitat heterogeneity can be assumed to increase the total species richness of archipelago. An important factor in my thesis was, as demonstrated by stable isotope analysis, that when studying arthropod food webs and communities on small islands in the

Baltic Sea one has to include the adjacent brackish ecosystem. Emerging insects with aquatic larvae (mainly chironomids) are an important prey for major island predators (mainly spiders) and thus pose a possibility for feed-back loops between the brackish and terrestrial system. Such feed-back loops may be likely for active cormorant islands with high nest densities, where δ15N analyses of brackish algae and associated invertebrates have indicated that significant amounts of ornithogenic nutrient run off in surrounding waters and enters the brackish food web via algae and caused increased chironomid biomass. The high nutrient content in the water column nearby high nest density islands caused increases in the epiphytic algae biomass relative to Fucus biomass and thus had similar cones- quences as reported in studies investigating the effects of eutrophication in the Baltic Sea. Even if some grazers responded, like chironomids, with increased biomass, the decrease of Fucus may lead to a decreased abundance and species richness of larger invertebrates, or even fishes, which are dependent on Fucus as a habitat in a long time perspective (Råberg and Kautsky 2008). However, I do not expect seabird colonies in the Baltic Sea or other marine systems to cause serious eutrophication problems, because they rather remove nutrient by transporting them, as shown, to land and, furthermore, the area affected by seabird nutrient input is restricted to the local surrounding of the colonies.

Further research

During my PhD study I met several interesting question which fall outside the scope of my work but whose answers may nevertheless be crucial for understanding the effects of nesting cormorants on island communities in particular and the effects of increase nutrient loads on community and food web structure, dynamic and interactions of the recipient system in general. For example,  When using stable isotope analysis and diet mixing models, two factors may confound the outcome: lipid content and fractionation. The quantitative role of these factors necessitates controlled diet experiments.  In my study I did not include the soil food web and hence was not able to understand responses of several ground living taxa, and how nesting seabirds affect the interaction between below- and aboveground ecosystem.  In my work I focused on bottom-up effects and disregarded top-down effects. Thus, I would like to investigate the role of top-down effects in dependence of nutrient and resource availability.  One weakness in paper IV is that I often compared the homeostasis within families or orders and not within species. Thus, it would be interesting to collected the same species on all study islands and, additionally, use these species in controlled diet experiments.  I found that ornithogenic nutrient also entered the surrounding waters, but I did not measure quantities and qualities. It would be interesting to quantitatively measure the nutrient flows from the islands, the nutrient content in the water columns nearby seabird islands and use this information, together with data on the algal responses, to conduct ecological network analysis in order to study changes food web functioning.  Seabirds are worldwide declining dramatically in numbers. In order to understand consequences from this decrease on the island, we need far more studies on the effects of nesting seabirds on terrestrial and marine ecosystem, in a wider range of island systems.

Acknowledgement

I gratefully acknowledge my supervisors P. A. Hambäck and L. Jerling for helpful comments on this summary and, additionally, J. Hansen for help wit the Swedish summary. I also thank L. Jerling, B. Jerling, E. Jerling, S. Jerling, H. Axemar, M. Enskog, J. Ekholm, C. Essenberg A. Lindström, C. Palmborg and A. Sjösten for help in the field and in the lab and H.-E. Wanntorp, C.-C. Coulianos, and Yngve Brodin for help in identifying Coleoptera, Heteroptera, and Chironomidae. Funding was provided by the Swedish Research Council Formas (to PAH).

References

Abrams PA (1995) Monotonic or unimodal diversity productivity gradients - what does competition theory predict. Ecology 76:2019-2027 Anderson TR, Hessen DO, Elser JJ, Urabe J (2005) Metabolic stoichiometry and the fate of excess carbon and nutrients in consumers. American Naturalist 165:1-15 Andersen T, Elser JJ, Hessen DO (2004) Stoichiometry and population dynamics. Ecology Letters 7:884-900 Anderson TR, Hessen DO (1995) Carbon or nitrogen limitation in marine copepods. Journal of Plankton Research 17:317-331 Andersen T, Hessen DO (1991) Carbon, nitrogen, and phosphorus content of freshwater zooplankton. Limnology and Oceanography 36:807-814 Anderson WB, Mulder CPH (2011) An introduction to seabird islands. In: Mulder CPH, Anderson WB, Towns DR, Bellingham PJ (eds) Seabird islands: Ecology, invasion and restoration. Oxford University Press in press Anderson WB, Polis GA (1999) Nutrient fluxes from water to land: seabirds affect plant nutrient status on Gulf of California islands. Oecologia 118:324-332 Anderson WB, Polis GA (1998) Marine subsidies of island communities in the Gulf of California: evidence from stable carbon and nitrogen isotopes. Oikos 81:75-80 Aneer G, Arvidsson D (2003) Inputs of nutrients (nitrogen and phosphorus) to coastal waters of Svealand, northern Baltic Sea, in 1997 (in swedish with english summary). Stockholm County Administrative Board Report 2003:17: Barrett K, Anderson WB, Wait DA, Grismer LL, Polis GA, Rose MD (2005) Marine subsidies alter the diet and abundance of insular and coastal lizard populations. Oikos 109:145-153 Bedard J, Therriault JC, Berube J (1980) Assessment of the importance of nutrient recycling by seabirds in the St Lawrence estuary. Canadian Journal of Fisheries and Aquatic Sciences 37:583-588 Bird MI, Tait E, Wurster CM, Furness RW (2008) Stable carbon and nitrogen isotope analysis of avian uric acid. Rapid Communications in Mass Spectrometry 22:3393-3400 Blackburn TM, Cassey P, Duncan RP, Evans KL, Gaston KJ (2004) Avian extinction and mammalian introductions on oceanic islands. Science 305:1955- 1958 Blais JM, Kimpe LE, McMahon D, Keatley BE, Mattory ML, Douglas MSV, Smol JP (2005) Arctic seabirds transport marine-derived contaminants. Science 309:445. Blumenshine SC, Vadeboncoeur Y, Lodge DM, Cottingham KL, Knight SE (1997) Benthic-pelagic links: responses of benthos to water-column nutrient enrichment. Journal of the North American Benthological Society 16:466-479

Bosman AL, Dutoit JT, Hockey PAR, Branch GM (1986) A field experiment demonstrating the influence of seabird guano on intertidal primary production. Estuarine Costal and Shelf 23:283-294 Bosman AL, Hockey PAR (1986) Seabird guano as a determinant of rocky intertidal community structure. Marine Ecology Progress Series 32:247-257 Brandvold DK, Popp CJ, Brierley JA (1976) Waterfowl refuge effect on water- quality.2. Chemical and physical parameters. Journal - Water Pollution Control Federation 48:680-687 Brit VL, Brit TP, Goulet D, Cairnse DK, Montevecchi WA (1987) Ashmole`s halo: direct evidence for prey depletion by seabirds. Marine Ecology Lake Progress Series 40:205-208 Brown AC, McLachlan A (eds) (1990) Ecology of sandy shores. Elsevier, Amsterdam, The Netherlands Bultman TL, Uetz GW (1984) Effect of structure and nutritional quality of litter on abundances of litter-dwelling arthropods. American Midland Naturalist 111:165-172 Burger AE, Lindeboom HJ, Williams AJ (1979) The mineral and energy contributions of guano of selected species of birds to the Marion Island terrestrial ecosystem. South African Journal of Antarctic Research 8:59-70 Carss DN (2002) Reducing the conflict between cormorants and fisheries on a pan-European scale. Centre for Ecology & Hydrology Banchory, Hill of Brathens, Banchory Aberdeenshire, AB 31 4BW, Scotland, UK Chan EKW, Zhang YX, Dudgeon D (2009) Substrate availability may be more important than aquatic insect abundance in the distribution of riparian orb-web spiders in the tropics. Biotropica 41:196-201 Clesceri LS, Greenberg AE, Eaton AD (1998) Standard methods for the examination of water and wastewater. American Pupblic Health Assoication, NY Cross WF, Benstead JP, Rosemond AD, Wallace JB (2003) Consumer-resource stoichiometry in detritus-based streams. Ecology Letters 6:721-732 Croxall KH, Evans PGH, Schreiber RW (1984) Status and conservation of the world's seabirds. International Council for Bird Preservation, Cambrig, UK Cushman JH (1995) Ecosystem-level consequences of species additions and deletions on islands. In: Vitousek PM, Loope LL, Adsersen H (eds) Biological diversity and ecosystem function. Springer-Verlag, Berlin Dalton CM, Ellis D, Post DM (2009) The impact of double-crested cormorant (Phalacrocorax auritus) predation on anadromous alewife (Alosa pseudoharengus) in south-central Connecticut, USA. Canadian Journal of Fisheries and Aquatic Sciences 66:177-186 Darwin C (1959) The origin of species by means of natural selection of the preservation of favoured races in the struggle for life. Signet Classic, London DeMott WR, Pape BJ (2005) Stoichiometry in an ecological context: testing for links between Daphnia P-content, growth rate and habitat preference. Oecologia 142:20-27

Dennis P, Young MR, Gordon IJ (1998) Distribution and abundance of small insects and arachnids in relation to structural heterogeneity of grazed, indigenous grasslands. Ecological Entomology 23:253-264 Draulans D (1988) Effects of fish-eating birds on fresh-water fish stocks - an evaluation. Biological Conservation 44:251-263 Ellis JC (2005) Marine birds on land: a review of plant biomass, species richness, and community composition in seabird colonies. Plant Ecology 181:227-241 Ellis JC, Bellinham PJ, Cameron EK, Croll DA, Kolb GS, Kueffer C, Mittelhauser GH, Schmidt S, Vidal E, Wait DA (2011) Effects of seabirds on plant communities. In: Mulder CPH, Anderson WB, Towns DR, Bellingham PJ (eds) Seabird islands: Ecology, invasion and restoration. Oxford University Press in press Ellis JC, Shulman MJ, Wood M, Witman JD, Lozyniak S (2007) Regulation of intertidal food webs by avian predators on New England rocky shores. Ecology 88:853-863 Ellis JC, Farina JM, Witman JD (2006) Nutrient transfer from sea to land: the case of gulls and cormorants in the Gulf of Maine. Journal of Animal Ecology 75:565-574 Elmgren R (2001) Understanding human impact on the Baltic ecosystem: Changing views in recent decades. Ambio 30:222-231 Elser JJ, Bracken MES, Cleland EE, Gruner DS, Harpole WS, Hillebrand H, Ngai JT, Seabloom EW, Shurin JB, Smith JE (2007) Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecology Letters 10:1135-1142 Elser JJ, Acharya K, Kyle M, Cotner J, Makino W, Markow T, Watts T, Hobbie S, Fagan W, Schade J, Hood J, Sterner RW (2003) Growth rate-stoichiometry couplings in diverse biota. Ecology Letters 6:936-943 Elser JJ, Sterner RW, Gorokhova E, Fagan WF, Markow TA, Cotner JB, Harrison JF, Hobbie SE, Odell GM, Weider LJ (2000a) Biological stoichiometry from genes to ecosystems. Ecology Letters 3:540-550 Elser JJ, O'Brien WJ, Dobberfuhl DR, Dowling TE (2000b) The evolution of ecosystem processes: growth rate and elemental stoichiometry of a key herbivore in temperate and arctic habitats. Journal of Evolutionary Biology 13:845-853 Elser JJ, Urabe J (1999) The stoichiometry of consumer-driven nutrient recycling: Theory, observations, and consequences. Ecology 80:735-751 Elser JJ, Dobberfuhl DR, MacKay NA, Schampel JH (1996) Organism size, life history, and N:P stoichiometry. Bioscience 46:674-684 Elton C (1927) Animal ecology. MacMillan, New York, USA Engstrom H (2001) Long term effects of cormorant predation on fish communities and fishery in a freshwater lake. Ecography 24:127-138 Erskine PD, Bergstrom DM, Schmidt S, Stewart GR, Tweedie CE, Shaw JD (1998) Subantarctic Macquarie Island - a model ecosystem for studying animal- derived nitrogen sources using N15 natural abundance. Oecologia 117:187-193

Eyre MD, Luff ML (2005) The distribution of epigeal beetle (Coleoptera) assemblages on the north-east England coast. Journal of Coastal Research 21:982-990 Fagan WF, Siemann E, Mitter C, Denno RF, Huberty AF, Woods HA, Elser JJ (2002) Nitrogen in insects: Implications for trophic complexity and species diversification. American Naturalist 160:784-802 Frank T, Reichhart B (2004) Staphylinidae and Carabidae overwintering in wheat and sown wildflower areas of different age. Bulletin of Entomological Research 94:209-217 Frost PC, Evans-White MA, Finkel ZV, Jensen TC, Matzek V (2005) Are you what you eat? Physiological constraints on organismal stoichiometry in an elementally imbalanced world. Oikos 109:18-28 Fukami T, Wardle DA, Bellingham PJ, Mulder CPH, Towns DR, Yeates GW, Bonner KI, Durrett MS, Grant-Hoffman MN, Williamson WM (2006) Above- and below-ground impacts of introduced predators in seabird-dominated island ecosystems. Ecology Letters 9:1299-1307 Garcia RR, Jauregui BM, Garcia U, Osoro K, Celaya R (2009) Responses of Arthropod Fauna Assemblages to goat grazing management in northern spanish heathlands. Environmental Entomology 38:985-995 Gillham ME (1961) Alteration of breeding habitat by seabirds and seals in Western Australia. The Journal of Ecology 49:289-300 Golovkin AN (1967) Effect of colonial sea birds on development of phytoplankton. Oceanology-USSR 7:521-52& Golovkin AN, Garkavaya GP (1975) Fertilization of waters off the Murmansk coast by bird excreta near various types of colonies. Soviet Journal of Marine Biology 1:345-351 Gratton C, Donaldson J, vander Zanden MJ (2008) Ecosystem linkages between lakes and the surrounding terrestrial landscape in northeast Iceland. Ecosystems 11:764-774 Gremillet D, Argentin G, Schulte B, Culik BM (1998) Flexible foraging techniques in breeding cormorants Phalacrocorax carbo and shags Phalacrocorax aristotelis: benthic or pelagic feeding? IBIS 140:113-119 Grover JP (2003) The impact of variable stoichiometry on predator-prey interactions: A multinutrient approach. American Naturalist 162:29-43 Haddad NM, Haarstad J, Tilman D (2000) The effects of long-term nitrogen loading on grassland insect communities. Oecologia 124:73-84 Hahn S, Bauer S, Klaassen M (2007) Estimating the contribution of carnivorous waterbirds to nutrient loading in freshwater habitats. Freshwater Biology 52:2421-2433 Hamback PA, Agren J, Ericson L (2000) Associational resistance: Insect damage to purple loosestrife reduced in thickets of sweet gale. Ecology 81:1784-1794 Hansson S, Rudstam LG (1990) Eutrophication and baltic fish communities. Ambio 19:123-125 Hart DD, Horwitz RJ (1991) Habitat diversity and species-area realtionship: alternative modles and tests. In: Bell SS, McCoy ED, Mushinsky HR (eds)

Habitat structure: the pysical arrangment of objects in space. Chapman and Hall, London Hawke DJ, Clark JM (2010) Incorporation of the invasive mallow Lavatera arborea into the food web of an active seabird island. Biological Invasions 12:1805-1814 Hebert CE, Duffe J, Weseloh DVC, Senese EMT, Haffner GD (2005) Unique island habitats may be threatened by double-crested cormorants. Journal of Wildlife Management 69:68-76 Hill C, Wallsröm K (2005) The Stockholm archipelago. In: Schiewer U (ed) Ecology of baltic costal waters. Springer, Berlin, Heidelberg Hobara S, Koba K, Osono T, Tokuchi N, Ishida A, Kameda K (2005) Nitrogen and phosphorus enrichment and balance in forests colonized by cormorants: Implications of the influence of soil adsorption. Plant and Soil 268:89-101 Hobara S, Osono T, Koba K, Tokuchi N, Fujiwara S, Kameda K (2001) Forest floor quality and N transformations in a temperate forest affected by avian- derived N deposition. Water, Air, and Soil Pollution 130:679-684 Hunter MD (2003) Effects of plant quality on the population ecology of parasitoids. Agricultural and Forest Entomology 5:1-8 Hunter MD, Price PW (1992) Playing chutes and ladders - heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73:724-732 Hurd LE, Fagan WF (1992) Cursorial spiders and succession - age or habitat structure. Oecologia 92:215-221 Iason GR, Duck CD, Cluttonbrock TH (1986) Grazing and reproductive success of red deer - the effect of local enrichment by gull colonies. Journal of Animal Ecology 55:507-515 Ishida A (1996) Effects of the common cormorant, Phalacrocorax carbo, on evergreen forests in two nest sites at Lake Biwa, Japan. Ecological Rearch 11:193-200 Jacobs W, Renner M (1998) Biologie und Ökologie der Insekten (in german) (3. Auflage). Gustav Fischer, Stuttgard Joern A, Behmer ST (1997) Importance of dietary nitrogen and carbohydrates to survival, growth, and reproduction in adults of the grasshopper Ageneotettix deorum (Orthoptera: Acrididae). Oecologia 112:201-208 Jones CG, Lawton JH, Shachak M (1994) Organisms as ecosystem engineers. Oikos 69:373-386 Jones HP (2010) Prognosis for ecosystem recovery following rodent eradication and seabird restoration in an island archipelago. Ecological Applications 20:1204-1216 Jones HP, Tershy BR, Zavaleta ES, Croll DA, Keitt BS, Finkelstein ME, Howald GR (2008) Severity of the effects of invasive rats on seabirds: A global review. Conservation Biology 22:16-26 Kagata H, Ohgushi T (2006) Nitrogen homeostasis in a willow leaf beetle, Plagiodera versicolora, is independent of host plant quality. Entomologia Experimentalis et Applicata 118:105-110

Kautsky L, Norbeg Y, Aneer G, Engqvist A (2000) Under the surface of the Stockholm Archipelago (in swedish) County Administrative Board. Lenanders Tryckeri of Stockholm, Sweden Kautsky H, Kautsky L, Kautsky N, Kautsky U, Lindblad C (1992) Studies on the Fucus vesiculosus community in the Baltic sea. Acta Phytogeographica Suecica 88:33-48 Kay AD, Rostampour S, Sterner RW (2006) Ant stoichiometry: elemental homeostasis in stage-structured colonies. Functional Ecology 20:1037-1044 Keatley B, Douglas M, Blais JM, Mallory M, Smol J (2009) Impacts of seabird- derived nutrients on water quality and diatom assemblages from Cape Vera, Devon Island, Canadian High Arctic. Hydrobiologia 621:191-205 Kolb GS, Young HS, Anderson WB (2011) Effects of seabirds on island consumers. In: Mulder CPA, Anderson WB, Towns DR, Belligham PJ (eds) Seabird islands: Ecology, invasion, and restoraition. Oxford University Press in press Kooijman SALM (1995) The stoichiometry of animal energetics. Journal of Theoretical Biology 177:139-149 Kotta J, Paalme T, Martin G, Makinen A (2000) Major changes in macroalgae community composition affect the food and habitat preference of Idotea baltica. International Reviwe if Hyrrobiology 85:697-705 Kraufvelin P (2007) Responses to nutrient enrichment, wave action and disturbance in rocky shore communities. Aquatic Botany 87:262-274 Kraufvelin P, Salovius S, Christie H, Moy FE, Karez R, Pedersen MF (2006) Eutrophication-induced changes in benthic algae affect the behaviour and fitness of the marine amphipod Gammarus locusta. Aquatic Botany 84:199-209 Kurle CM, Croll DA, Tershy BR (2008) Introduced rats indirectly change marine rocky intertidal communities from algae- to invertebrate-dominated. Proceedings of the National Academy of Sciences of the United States of America 105:3800-3804 Kyto M, Niemela P, Larsson S (1996) Insects on trees: Population and individual response to fertilization. Oikos 75:148-159 Lapointe BE, Litter MM, Litter DS (1992) Modification of benthic community structure by natural eutrophication: the Belize Barrier Reef. Proceedings of the Seventh International Coral Reef Symposium 1 Lindeboom (1984) The nitrogen pathway in a penguin rookery. Ecology 65:269- 277 Linderholm J (2007) Soil chemical surveying: A path to a deeper understanding of prehistoric sites and societies in Sweden. Geoarchaeology an International Journal 22:417-438 Littler MM, Littler DS, Titlyanov EA (1991) Comparisons of N-limited and P- limited productivity between high granitic islands versus low carbonate atolls in the seychelles archipelago - a test of the relative-dominance paradigm. Coral Reef 10:199-209

Loder TC, Ganning B, Love JA (1996) Ammonia nitrogen dynamics in coastal rockpools affected by gull guano. Journal of Experimental Marine Biology and Ecology 196:113-129 Logan JD, Joern A, Wolesensky W (2004) Control of CNP homeostasis in herbivore consumers through differential assimilation. Bulletin of Mathematical Biology 66:707-725 Lomolino MV, Riddle BR, Brown JH (2006) Biogeography. Lopez-Victoria M, Wolters V, Werding B (2009) Nazca Booby (Sula granti) inputs maintain the terrestrial food web of Malpelo Island. Journal für Ornithologie 150:865-870 MacArthur RH, Wilson EO (1963) Equilibrium-theory of insular zoogeography. Evolution; International Journal of Organic Evolution 17:373-37 MacArthur R, MacArthur JW (1961) On bird species-diversity. Ecology 42:594- 59 Marczak LB, Richardson JS (2007) Spiders and subsidies: results from the riparian zone of a coastal temperate rainforest. Journal of Animal Ecology 76:687-694 Marczak LB, Thompson RM, Richardson JS (2007) Meta-analysis: Trophic level, habitat, and productivity shape the food web effects of resource subsidies. Ecology 88:140-148 Marion L, Clergeau P, Brient L, Bertru G (1994) The importance of avian- contributed nitrogen (N) and phosphorus (P) to Lake Grand-Lieu, France. Hydrobiologia 280:133-147 Markwell TJ, Daugherty CH (2002) Invertebrate and lizard abundance is greater on seabird-inhabited islands than on seabird-free islands in the Marlborough Sounds, New Zealand. Ecoscience 9:293-299 Maron JL, Estes JA, Croll DA, Danner EM, Elmendorf SC, Buckelew SL (2006) An introduced predator alters Aleutian Island plant communities by thwarting nutrient subsidies. Ecological Monographs 76:3-24 Mattson WJ (1980) Herbivory in relation to plant nitrogen-content. Annual Review of Ecology and Systematics 11:119-161 Mayntz D, Toft S (2001) Nutrient composition of the prey's diet affects growth and survivorship of a generalist predator. Oecologia 127:207-213 McCutchan JH, Lewis WM, Kendall C, McGrath CC (2003) Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102:378- 390 Mellbrand K, Hambäck PA (2011). Coastal niches for prdators: a stable isotope study. Canadian Journal of Zoology in press Mitchell RJ (1992) Testing evolutionary and ecological hypotheses using path- analysis and structural equation modeling. Functional Ecology 6:123-129 Mittelbach GG, Steiner CF, Scheiner SM, Gross KL, Reynolds HL, Waide RB, Willig MR, Dodson SI, Gough L (2001) What is the observed relationship between species richness and productivity? Ecology 82:2381-2396 Miyashita T, Takada M (2007) Habitat provisioning for aboveground predators decreases detritivores. Ecology 88:2803-2809

Mizutani H, Hasegawa H, Wada E (1986) High nitrogen isotope ratio for soils of seabird rookeries. Biochemistry 2:221-247 Mizutani H, Wada E (1988) Nitrogen and carbon isotope ratios in seabird rookeries and their ecological implications. Ecology 69:340-349 Mulder C, Elser JJ (2009) Soil acidity, ecological stoichiometry and allometric scaling in grassland food webs. Gloabal Change Biology 15:2730-2738 Mulder CPH, Jones H, Kameda K, Palmborg C, Schmidt S, Ellis JC, Orrock JL, Wait DA, Wardle DE, Yang L, Young H, Croll D, Vidal E (2011) Impacts of seabirds on plant and soil properties. In: Mulder CPH, Anderson WB, Towns DR, Bellingham PJ (eds) Seabird islands: ecology, invasion and restoration. Oxford University Press in press Mulder K, Bowden WB (2007) Organismal stoichiometry and the adaptive advantage of variable nutrient use and production efficiency in Daphnia. Ecological Modelling 202:427-440 Nakano S, Murakami M (2001) Reciprocal subsidies: Dynamic interdependence between terrestrial and aquatic food webs. Proceedings of the National Academy of Sciences of the United States of America 98:166-170 Onuf CP, Teal JM, Valiela I (1977) Interactions of Nutrients, Plants Growth and Herbivory in a Mangrove Ecosystem. Ecology 58:514-526 Orav-Kotta H, Kotta J (2004) Food and habitat choice of the isopod Idotea baltica in the northeastern Baltic Sea. Hydrobiologia 514:79-85 Orgeas J, Vidal E, Ponel P (2003) Colonial seabirds change beetle assemblages on a Mediterranean Island. Ecosience 10:38-44 Osono T, Hobaraa S, Fujiwaraa S, Kobab K, Kamedac K (2002) Abundance, diversity, and species composition of fungal communities in a temperate forest affected by excreta of the Great Cormorant Phalacrocorax carbo. Soil Biology & Biochemistry 34: 1537-1547 Paetzold A, Lee M, Post DM (2008) Marine resource flows to terrestrial arthropod predators on a temperate island: the role of subsidies between systems of similar productivity. Oecologia 157:653-659 Paillisson JM, Carpentier A, Le Gentil J, Marion L (2004) Space utilization by a cormorant (Phalacrocorax carbo L.) colony in a multi-wetland complex in relation to feeding strategies. Comptes Rendus Biologies 327:493-500 Palomo G, Iribarne O, Martinez MM (1999) The effect of migratory seabirds guano on the soft bottom community of a SW Atlantic coastal lagoon. Bulletin of Marine Science 65:119-128 Perner J, Voigt W, Bahrmann R, Heinrich W, Marstaller R, Fabian B, Gregor K, Lichter D, Sander FW, Jones TH (2003) Responses of arthropods to plant diversity: changes after pollution cessation. Ecography 26:788-78 Persson J, Fink P, Goto A, Hood JM, Jonas J, Kato S (2010) To be or not to be what you eat: regulation of stoichiometric homeostasis among autotrophs and heterotrophs. Oikos 119:741-751 Peterson BJ, Fry B (1987) Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics 18:293-320

Polis GA, Sáchez-Piñero F, Stapp PT, Anderson WB, Rose MR (2004) Trophic flows from water to land: marine input affects food webs of island and costal ecosystems worldwide. Food webs at the landescape level. The University of Chicago Press Chicago and London Polis GA, Hurd SD, Jackson CT, Pinero FS (1997a) El Nino effects on the dynamics and control of an island ecosystem in the Gulf of California. Ecology 78:1884-1897 Polis GA, Anderson WB, Holt RD (1997b) Toward an integration of landscape and food web ecology: The dynamics of spatially subsidized food webs. Annual Review of Ecology and Systematics 28:289-316 Polis GA, Hurd SD (1996a) Linking marine and terrestrial food webs: Allochthonous input from the ocean supports high secondary productivity on small islands and coastal land communities. American Naturalist 147:396-423 Polis GA, Hurd SD (1996b) Linking marine and terrestrial food webs: Allochthonous input from the ocean supports high secondary productivity on small islands and coastal land communities. American Naturalist 147:396-423 Polis GA, Hurd SD (1995) Extraordinarily high spider densities on islands - flow of energy from the marine to terrestrial food webs and the absence of predation. Proceedings of the National Academy of Sciences of the United States of America 92:4382-4386 Ponsard S, Averbuch P (1999) Should growing and adult animals fed on the same diet show different delta N-15 values? Rapid Communications in Mass Spectrometry 13:1305-1310 Post DM, Layman CA, Arrington DA, Takimoto G, Quattrochi J, Montana CG (2007) Getting to the fat of the matter: models, methods and assumptions for dealing with lipids in stable isotope analyses. Oecologia 152:179-189 Power ME, Huxel GR (2004) preface. In: Polis GA, Power ME, Huxel GR (eds) Food webs at the landscape level. The University of Chicago Press, Chicago and London Price PW (2002) Resource-driven terrestrial interaction webs. Ecological research 17:241-247 Råberg S, Kautsky L (2007) A comparative biodiversity study of the associated fauna of perennial fucoids and filamentous algae. Estuarine Costal and Shelf Science73:249-258 Råberg S, Kautsky L (2008) Grazer identity is crucial for facilitating growth of the perennial brown alga Fucus vesiculosus. Marine Ecology Progress Series 361:111-118 Raubenheimer D, Jones SA (2006) Nutritional imbalance in an extreme generalist omnivore: tolerance and recovery through complementary food selection. Animal Behaviour 71:1253-1262 Raubenheimer D, Simpson SJ (2004) Organismal stoichiometry: Quantifying non- independence among food components. Ecology 85:1203-1216 Rolland C, Danchin E, de Fraipont M (1998) The evolution of coloniality in birds in relation to food, habitat, predation, and life-history traits: A comparative analysis. American Naturalist 151:514-529

Root RB (1973) Organization of a plant-arthropod association in simple and diverse habitats - fauna of collards (Brassica oleracea). Ecological Monographs 43:95-120 Ryan PG, Watkins BP (1989) The influence of physical factors and ornithogenic products on plant and arthropod abundance at an inland Nunatak group in Antarctica. Polar Biology 10:151-160 Sanchez-Pinero F, Polis GA (2000) Bottom-up dynamics of allochthonous input: Direct and indirect effects of seabirds on islands. Ecology 81:3117-3132 Schmidt MH, Clough Y, Schulz W, Westphalen A, Tscharntke T (2006) Capture efficiency and preservation attributes of different fluids in pitfall traps. Journal of Arachnology Schmidt S, Dennison WC, Moss GJ, Stewart GR (2004) Nitrogen ecophysiology of Heron Island, a subtropical coral cay of the Great Barrier Reef, Australia. Functional plant biology 31:517-528 Siemann E (1998) Experimental tests of effects of plant productivity and diversity on grassland arthropod diversity. Ecology 79:2057-2070 Siemann E, Tilman D, Haarstad J, Ritchie M (1998) Experimental tests of the dependence of arthropod diversity on plant diversity. American Naturalist 152:738-750 Sinclair BJ, Chown SL (2006) Caterpillars benefit from thermal ecosystem engineering by wandering albatrosses on Sub-Antarctic Marion Island. Biology Letters 2:51-54 Slansky F, Feeny P (1977) Stabilization of rate of nitrogen accumulation by larvae of cabbage butterfly on wild and cultivated food plants. Ecological Monographs 47:209-228 Small GE, Pringle CM (2010) Deviation from strict homeostasis across multiple trophic levels in an invertebrate consumer assemblage exposed to high chronic phosphorus enrichment in a Neotropical stream. Oecologia 162:581-590 Smith JL, Mulder CPH, Ellis JC (2011) Seabirds as engineers: Nutrient inputs and physical disturbance. In: Mulder CPH, Anderson WB, Towns DR, Bellingham PJ (eds) Seabird islands: Ecology, invasion and restoration. Oxford University Press in press Smith JS, Johnson CR (1995) Nutrient inputs from seabirds and humans on a populated coral cay. Marine Ecology Progress Series 124:189-200 Sobey DG, Kenworthy JB (1979) Relationship between herring-gulls and the vegetation of their breeding colonies. Journal of Ecology 67:469-496 Spiller DA, Schoener TW (1994) Effects of top and intermediate predators in a terrestrial food-web. Ecology 75:182-196 Spiller DA, Schoener TW (1997) Folivory on islands with and without insectivorous lizards: An eight-year study. Oikos 78:15-22 Staav R (2010) Skarv (in swedish). In: Strandfager K, Palmblad U (eds) Levande skärgårdsnatur 2010 - med rapporter från 2009. Skärgårdsstiftelsen, Stockholm Staav R (2008) Skarv (in swedish). In: Strandfager K, Palmblad U (eds) Levande skärgårdsnatur 2008 - med rapporter från 2007. Skärgårdsstiftelsen, Stockholm

Stapp P, Polis GA (2003) Marine resources subsidize insular rodent populations in the Gulf of California, Mexico. Oecologia 134:496-504 Sterner RW, Elser JJ (2002) Ecological Stoichiometry. The biology of elements from molecules to the biosphere. Princeton University Press, Princeton, New Jersey, UK Sterner RW, Schwalbach MS (2001) Diet integration of food quality by Daphnia: Luxury consumption by a freshwater planktonic herbivore. Limnology and Oceanography 46:410-416 Sterner RW, Elser JJ, Hessen DO (1992) Stoichiometric relationships among producers, consumers and nutrient cycling in pelagic ecosystems. Biochemistry 17:49-67 Thingstad TF (1987) Utilization of N, P, and organic C by heterotrophic bacteria.1. Outline of a chemostat theory with a consistent concept of maintenance metabolism. Marine Ecology Progress Series 35:99-109 Tilman D (1987) The importance of the mechanisms of interspecific competition. American Naturalist 129:769-774 Tilman D, Wedin D, Knops J (1996) Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379:718-720 Towns DR, Wardle DA, Mulder CPH, Yeates GW, Fitzgerald BM, Parrish GR, Bellingham PJ, Bonner KI (2009) Predation of seabirds by invasive rats: multiple indirect consequences for invertebrate communities. Oikos 118:420- 430 Urabe J, Kyle M, Makino W, Yoshida T, Andersen T, Elser JJ (2002) Reduced light increases herbivore production due to stoichiometric effects of light/nutrient balance. Ecology 83:619-627 Urabe J, Watanabe Y (1992) Possibility of N-limitation or P-limitation for planktonic cladocerans - an experimental test. Limnology and Oceanography 37:244-251 Vanderklift MA, Ponsard S (2003) Sources of variation in consumer-diet delta N- 15 enrichment: a meta-analysis. Oecologia 136:169-182 Vanni MJ (2002) Nutrient cycling by animals in freshwater ecosystems. Annual Review of Ecology and Systematics 33:341-370 Vitousek PM, Adsersen H, Loope LL (1995) Introduction - Why focus on islands? In: Vitousek PM, Loope LL, Andersen H (eds) Islands: biological diversity and ecosystem function. Springer-Verlag, Berlin Voss M, Deutsch B, Elmgren R, Humborg C, Kuuppo P, Pastuszak M, Rolff C, Schulte U (2006) Source identification of nitrate by means of isotopic tracers in the Baltic Sea catchments. Biogeoscienc 3:663-676 Waide RB, Willig MR, Steiner CF, Mittelbach G, Gough L, Dodson SI, Juday GP, Parmenter R (1999) The relationship between productivity and species richness. Annual Review of Ecology and Systematics 30:257-300 Wainright SC, Haney JC, Kerr C, Golovkin AN, Flint MV (1998) Utilization of nitrogen derived from seabird guano by terrestrial and marine plants at St. Paul, Pribilof Islands, Bering sea, Alaska. Marine Biology 131:63-71

Wait DA, Aubrey DP, Anderson WB (2005) Seabird guano influences on desert islands: soil chemistry and herbaceous species richness and productivity. Journal of Arid Enviroments 60:681-695 Wardle DA, Bellingham PJ, Fukami T, Mulder CPH (2007) Promotion of ecosystem carbon sequestration by invasive predators. Biology Letters 3:479- 482 Wardle DA, Bonner KI, Nicholson KS (1997) Biodiversity and plant litter: Experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos 79:247-258 Wheeler GS, Halpern MD (1999) Compensatory responses of Samea multiplicalis larvae when fed leaves of different fertilization levels of the aquatic weed Pistia stratiotes. Entomologia Experimentalis et Applicata 92:205-216 White TCR (1993) The Inadequate Environment: Nitrogen and the Abundance of Animals. Whittaker RJ, Heegaard E (2003) What is the observed relationship between species richness and productivity? comment. Ecology 84:3384-3390 Wikstrom SA, Kautsky L (2007) Structure and diversity of invertebrate communities in the presence and absence of canopy-forming Fucus vesiculosus in the Baltic Sea. Estuarine Coastal and Shelf Science 72:168-176 Wolfe KM, Mills HR, Garkaklis MJ, Bencini R (2004) Post-mating survival in a small marsupial is associated with nutrient inputs from seabirds. Ecology 85:1740-1746 Wootton JT (1992) Indirect effects, prey susceptibility, and habitat selection - impacts of birds on limpets and algae. Ecology 73:981-991 Wootton JT (1991) Direct and indirect effects of nutrients on intertidal community structure - variable consequences of seabird guano. Journal of Experimental Marine Biology and Ecology 151:139-153 Worm B, Lotze HK, Sommer U (2000) Coastal food web structure, carbon storage, and nitrogen retention regulated by consumer pressure and nutrient loading. Limnology and Oceanography 45:339-349 Worm B, Sommer U (2000) Rapid direct and indirect effects of a single nutrient pulse in a seaweed-epiphyte-grazer system. Marine Ecology Progress Series 202:283-288 Wright DG, van der Wal R, Wanless S, Bardgett RD (2010) The influence of seabird nutrient enrichment and grazing on the structure and function of island soil food webs. Soil Biology and Biochemistry 42:592-600 Wright A (1934) The method of path coefficients. Annals of Mathematics and Statistics 5:161-215 Young HS, Hurrey L, Kolb GS (2011) Effects of seabird-derived nutrients on aquatic systems. In: Mulder CPH, Anderson WB, Towns DR, Bellingham PJ (eds) Seabird islands: Ecology, invasion and restoration. Oxford University Press. in press Zelickmaea, Golovkin AN (1972) Composition, structure and productivity of neritic plankton communities near bird colonies of northern shores of Nnovaya Zemlya. Marine Biology 17:265-26

Svensk sammanfattning

Den här avhandlingen handlar om hur skarvkolonier påverkar dels växter och leddjur (insekter, spindlar och kräftdjur) på häckningsöar och dels alger och associerade ryggradslösa djur omkring öarna. Jag har använt mig av så kallad stabil isotopanalys (kol och kväve) för att följa resan av kväve från skarvarna genom näringsvävarna på öarna och i närliggande vatten. Jag undersökte hur individantal, artrikedom och artsammansättning av växter och djur förändras av skarvarnas närvaro, genom att jämföra öar med häckande skarv med öar utan. Dessutom undersökte jag näringsinnehållet i djuren och deras förmåga att hålla kol:kväve:fosfor (C:N:P) proportionerna konstanta (homeostatis) och testade om skillnader i näringsinnehåll och homeostasis kan förklara hur individantal av de olika taxonomiska grupperna ändrade sig under påverkan. Skarvar är fiskätande sjöfåglar som häckar i kolonier upp till 10 000 par. De har återkoloniserat Stockholms skärgård sedan 1994 och deras antal ökade fram till 2007. Därefter verkar populationsstorleken ha stabiliserat sig på omkring 5200 par fördelat på 20 kolonier. Det finns en stor svensk opinion som ogillar skarvens återkomst i svensk natur och anser att skarven påverkar skärgårdens ekosystem negativt. Argumenten för detta är att skarven konsumerar stora mängder fisk samt förstör naturen vid häckningsöarna. Konstigt nog har ingen vetenskaplig undersökning gjorts för att utreda hur skarvarna påverkar skärgårdens ekosystem, och detta är mig veterligen den första studien. Skarvar transporterar, liksom andra fiskätande sjöfåglar, näring från vattnet till land. De jagar fisk inom ett stort område och häckar eller vilar på land, ofta på öar, där de lämnar stora mängden av näringsrik spillning (guano), kadaver, döda fiskar och äggskal. Detta näringsbidrag gödslar jorden och ökar näringsinnehållet i växterna, samt förändrar växternas artsammansättning och minskar antalet arter. Dessutom gynnas många leddjur direkt eller indirekt av skarv, genom att dels konsumera det som skarvarna lämnar efter sig och dels genom växternas ökade näringsinnehåll. Individantalet av leddjur ökar därför ofta på skarvöar men den minskade artrikedomen av växter reducerar också artantalet av växtätare. Näringen från fågelspillning påverkar inte bara ekosystemet på öarna utan även genom avrinning ekosystemet i det närbelägna vattnet. Ökade närsaltskoncentrationer i vattnet kan påverka vattenlevande organismer på motsvarande sätt som på land. En gren inom ekologin som kallas ekologisk stökiometri har skapat en teoretisk ram för relationen mellan artegenskaper, näringsbehov och populationstillväxt. Många tillväxtmodeller är byggda på antagande att djur håller kvoten C:N:P konstant oberoende av näringsinnehållet i sin föda (homeostas), i motsats till växter. Senare studier har dock visat att inte alla djur är strängt homeostatiska. Jag antog att djur som inte är homeostatiska kan dämpa förändringar i närings- innehållet av sin föda och därför visa mindre skillnader i individantal mellan skarv och icke-skarv öar. Tidigare studier tyder på att djurens C:N:P-kvot bestämmer deras näringsbehov. Näringsrika djur behöver mer näringsämnen och är känsligare

76 för att bli begränsade av näringsämnen än näringsfattiga djur. Därför borde näringsrika djur påverkas mer positivt av skarvödsling och visa starkare ökning i individantal på skarvöar än näringsfattiga djur. Jag observerade, liksom tidigare studier på fågelöar, att kväve och fosfat- innehållet i växter var högre, medan vegetationens täckningsgrad och artrikedomen var lägre på skarvöar än på icke-skarvöar. På övergivna skarvöar var artrikedomen lika hög som på icke-skarvöar, men växtbiomassan tenderade att vara högre., Artsammansättningen av växter skilde sig mellan skarv och icke-skarvöar, men var lika mellan övergivna skarvöar och icke-skarvöar. Några, men inte alla, grupper av ryggradslösa djur reagerade på skarvnärvaro genom förändringar i individantal, artrikedom och artsammansättning. De kvantitativa förändringarna kunde vara både positiva och negativa. Några växtätare, svampätare och kadaver- ätare blev positivt påverkade av förhöjd födokvalitet eller kvantitet och ökade i individ- eller artantal. Andra växtätande, detritusätande och rovlevande djurgrupper missgynnades av förändringarna i vegetationsstruktur och jordkemi och var på skarvöar mindre individ- eller artrika än på icke-skarvöar. Skillnaden i förändringar av individantal mellan taxonomiska grupper kunde inte förklaras av skillnader i djurens näringsinnehåll, i motsats till mitt antagande. Jag fann vidare inget samband mellan homeostasnivån och den observerade förändringen i individantal mellan djurgrupper. Vill man avgöra om skarvkolonierna ökar eller minskar artrikedom måste man skilja mellan lokal artrikedom (=artrikedom på en ö) och regional artrikedom (=artrikedom i ett skärgårdsområde). Även om skarvar kan minska artrikedom av växter och några leddjursgrupper på häckningsöarna, ökar de också den regionala habitatheterogeniteten, genom att ändra artsammansättningen hos växterna, jordkemi, vegetationsstruktur och resurstillgång. På grund av en ökad habitatheterogenitet kan den regionala artrikedomen öka eftersom olika arter trivs i olika habitat. En viktig observation i mitt arbete är vidare den täta kopplingen mellan näringsväven på land och den i vattnet. Genom att fjädermyggor som har larvstadier i vatten och vuxenstadier på land och marina alger spolas upp på land blir föda åt landlevande djur kopplas näringsvävarna ihop. Analyser av stabila isotoper har visat att fjädermyggor är en mycket viktig resurs för spindlar och några rovinsekter på land och detta öppnar möjligheten för en återkoppling mellan ekosystemet på land och det i vattnet. Dessutom indikerade den stabila isotopanalysen att näring från skarvspillningen rann ned från öarna till omgivande vatten, att denna näring togs upp av alger och sedan fortsatte upp i näringsväven till växtätande ryggradslösa djur. Ökade närsaltskoncentrationer stimulerande produktionen av påväxtalger vilket ledde till högre biomassa av några djurgrupper, däribland fjädermyggslarver. Ökad biomassa av vuxna fjädermyggor som flyger in över land återkopplar därmed till näringsväven på land. Mina studier visade inte att den lokalt ökade närsaltskoncentrationen runt skarvöarna främjar övergödningen av Östersjön. Tvärtom tar skarvarna upp näring från vattnet (fiskar) och

77 transporterar denna till land, och endast en liten del av denna näring rinner tillbaka till vattnet.

TACK!

När jag började min doktorandtjänst trodde jag att fyra år var en ganska lång tid men de fyra år rasade förbi tack vare ett omväxlande arbete i fält, i laboratoriet och framför dator, artikel- och bokläsning, kurser, möten och doktorandresor. Under dessa år som doktorand har jag lärt mig många saker; att fika är mycket viktigt i Sverige, att allt tar längre tid än man tror, att det kan vara bra att tänka efter lite innan man gör något (och även innan man säger något), att man kan skriva om ett stycke 1000 gånger och att det ändå inte blir bra, att fältdata och statistik inte alltid går bra ihop, att man ska spara noga och regelbundet, att ett litteraturprogram kanske inte är så dumt, att man ska ha ett system in sina datorfiler (det vill säga ett bra system), att man inte bör ta kritik på sin arbete personligt, att åka båt inte är detsamma som att åka bil, att abborre och sill smaka faktisk bra. (Självklart lärde jag mig också mycket om forskning, metoder, teorier, statistik, planering och sådant). Men framför allt gav mig min doktorandtjänst möjligheten att upptäckta hur fantastiskt vacker Stockholms skärgård är, hur olika den kan se ut beroende på vädret och årstid och det är jag mycket glad över.

Alla dessa lärorika år hade jag många hjälpsamma människor på min sida och utan dem skulle jag aldrig ha klarar att genom- och slutföra min avhandling. Alla dessa människor vill jag passa på att tacka här. Först vill jag tacka min handledare Peter, för att du ordnade med resurser så att jag kunde genomföra mitt projekt, hade assistenter och kunde åka på konferenser och kurser. Tack för att du gav mig mycket frihet i mitt fältarbete och tidsplaneringen, tack för att du alltid läste mina texter så snabb utan att tröttna, tack för att du hjälpte mig med statistiken, vetenskapliga och språkliga frågor och tack att jag fick ringa dig även på helgen eller under din semester. Tack även min biträdande handledare Lenn och hela familjen Jerling (samt Eva) för underbara somrar på Norröra som jag redan har börjat att sakna och som jag aldrig kommer att glömma., Tack för båtkörning (utan er skulle hela projektet aldrig ha fungerat), assistans i fält, matlagning, diskning, tömma dassen (Bitte). Tack för trevligt sällskap och för att ni alltid fick mig att känna mig välkommen på ert sommarställe! Tack Lenn för att ta hand om båten, berätta så mycket om skärgården och tack för ett kram och uppmuntrande ord när det behövdes.

Tacka vill jag också alla mina doktorandkolleger på Botan för en bra stämning, bokdiskussioner och hjälp när man behövde det. Framförallt tackar jag min rumskompis Martin; utan dig skulle de fyra åren ha varit så mycket tristare.0 Tack för allt vetenskapligt och privat utbyte och för att du lyssnade på mina utbrott. Tack också till dig Jocke, för din hjälp och för att lyssna, särskilt i början och nu på slutet av doktorandtiden. Tack Johan D. för statistiskhjälp. Tack Didrik för uppmuntrande ord. Tack Kajsa för hjälp i fält. Tack Petter för roliga samtal. Tack Tove för hjälpen med engelskan. Bryndis för din energi. Tack Helena för virkningstips. Tack alla doktorander som var med i Spanien på två roliga resor. Tack Sonja för hjälpen med Fucus-frågor.

Ett stort tack också till mina fält- och labassistenter Maria, Hanna, Annika och Caroline för allt utmärkt jobb ni har gjort och trevligt sällskap ni har varit på Norröra och på Botan. Tack också till dig Jessica för att jag alltid har kunnat fråga dig om hjälp och prata några ord. Tack Sofia för hjälpen men paper III. Tack Kristoffer för hjälpen med paper II. Tack Johan E. för att vara trevlig och roande. Tack Cissi för jorddata. Mitt tack gäller även den övriga personalen på Botan; tack Peter och Ingela för att ni tagit hand om mina fackelblomster och för att gestalta så fina rabatter att jag alltid blir glad när jag kommer till jobbet. Tack Johan K. för datorsupporten. Tack Ahmed för att städa och alltid le när man ser dig. Tack till den administrativa personalen och tack Gisela för all hjälp när jag började på Botan.

Thanks also to the SEAPRE network for inspiring meetings and real interest in my work.

But I would have never been able to do a PhD if my parents wouldn`t have paid me my education: Vielen lieben Dank Mama und Papa für die Finanzierung meines Studiums und seelische Unterstützung. Und auch Dir Oma Friedel vielen lieben Dank. Danke auch Euch meinen Freunden in Deutschland für Besuche, Telefonate, mails och Skypsupport .

Tack min svenska familj och alla mina svenska vänner som gör att jag känner mig hemma i Sverige och visar mig vad det är som är verkligen är viktig i livet.

Sist men inte minst vill jag tacka dig min älskade Magnus. Tack att du stod ut med mig under dem senaste veckorna, att du lyssna på mitt gnäll och tack för all assistans i fält. Tack att du finns, utan dig vore mitt liv så mycket fattigare.