and to decrease with isolation. There is ample evidence of impoverished species diversity as a consequence of changes in the spatial configuration of species’ habitats (e.g. Tscharntke and Brandl 2004; Fahrig 2003). Loss of species may lead to changes in ecosystem function such as pollination (Steffan-Dewenter and Tscharntke 1999; Dupont and Nielsen 2006), decomposition (Burkey 1997), and parasitism (Komonen et al. 2000). However, most studies are based on species numbers and less is known about the functional processes such as the damage caused by herbivory or mortality caused by parasitoids (Tscharntke and Brandl 2004). According to the trophic rank hypothesis, susceptibility to habitat fragmentation should amplify in species from higher trophic levels (Didham et al. 1996; Holt et al. 1999; Kondoh 2003; Henle et al. 2004). While this is probably true, it is not always the case (van Nouhuys 2005). Species attributes that affect their sensitivity to habitat fragmentation like feeding type, dispersal ability, reproductive potential, and rarity may not necessarily correspond to trophiclevel. Beneath a decrease of individuals in isolated populations, the genetic structure of populations may be affected by spatial isolation. The advent of large-scale genotyping methods such as microsatellite markers (simple sequence repeats) facilitates the assessment of isolation effects on the genetic diversity and structure of insect populations.
populations at Golinga have spawning stock biomasses below 40% of the unfished biomass. This points to a situation of a possible ongoing recruitment overfishing of those species in the two reservoirs and suggests that a further increase in fishing effort is not advisable. To support the construction of reservoirs’ food web models, a study was conducted on the feeding characteristics of the giraffe catfish with the expectation that the population in Reservoir Tono, which has an extensive macrophyte coverage, feeds more on plant material and associated insects than their counterparts in the Reservoir Bontanga. The study showed that fish food items did not differ significantly between the two reservoirs. Insect larvae and algae dominated the stomach contents. Comparative analysis of the reservoirs showed interesting differences: the mean trophiclevel of the catch was lowest in the largest and deepest reservoir (Tono), likely due to higher trophiclevel species occupying less accessible deep ‘refuge’ habitats. In the medium-sized (Bontanga) and small shallow (Golinga) reservoirs, in contrast, a larger catch portion resembles high trophiclevel species. Lake Bontanga differs from the other reservoirs by having a lower human population impact, a significantly lower Total Primary Production to Total Respiration ratio, a higher Total Biomass to Total System Throughput ratio, a higher Finn Cycling Index, a higher Detritivory to Herbivory ratio as well as the highest gross efficiency of the catch, all indicative for a more developed ecosystem. The smallest shallow (Golinga) reservoir is more impacted by anthropogenic activities than the other two reservoirs as indicated by the high levels of dissolved organic carbon, total dissolved nitrogen bonded, nitrite-nitrogen and turbidity in the reservoir. While the smallest lake had the highest fish production (per unit area) under optimal conditions of water supply, it is most vulnerable when used for both irrigated agriculture and fisheries production. The findings of this thesis suggest that the use of man-made lakes and respective catchment areas should be assessed and managed carefully to prevent the loss of nutrition and livelihoods contributions. Finally, this thesis serves as a broad template for the development of sustainable ecosystem-based management measures not only for the three studied ecosystems but for other reservoirs exposed to human activities around the world.
mode. The close nutritional link to the epipelagic phytoplankton production was further apparent from high proportions of diatom marker fatty acids like 16:1(n-7) and 20:5(n- 3), the dinoflagellate marker fatty acid 22:6(n-3) and the marker fatty acid for carnivory in zooplankton 18:1(n-9) (e.g. Dalsgaard et al. and references therein), which characterised the storage lipid fatty acid composition of the fishes. As indicated by gut content analysis, the nutritional link to the benthopelagic fish community ran either via pelagic prey to zooplanktivorous fish species, which in turn were preyed upon by piscivorous species like T. picturatus and C. conger, or via the phytodetritus-dependent benthic food web utilised by benthivorous fish species like the flatfish A. rueppelli. Stable isotope analysis also confirmed a close link to the pelagic food source. The δ 15 N values of zooplanktivorous fish were in the range of the average enrichment per trophiclevel suggested by literature (Minagawa and Wada 1984, Vander Zanden & Rasmussen 2001, Post 2002), although exact determination was not possible due to the large range of δ 15 N values covered by the potential prey. The average trophiclevel enrichment of δ 13 C values between zooplanktivorous fishes and their potential prey ranged between 1.5 and 2 ‰ and was thus slightly higher than the generally proposed <1 ‰ trophiclevel increase (DeNiro & Epstein 1978, DeNiro & Epstein 1981, Peterson & Fry 1987, Michener & Schell 1994). The slightly higher trophiclevel enrichment might be due, on one hand, to the large range of δ 13 C values covered by the potential prey, but might also be influenced by preservation effects in fish samples and by lipid content in zooplankton samples despite correction of δ 13 C values, although similar trophiclevel differences were observed in studies from shelf areas (Davenport & Bax 2002, Bode et al. 2004).
of primary POM consumers. The results are discussed in respect of known POM dynamics. If our hypotheses prove true, sampling and analysis strategies would have to be adjusted accordingly in order to avoid serious bias in estimates of organ- isms’ trophiclevel or the degree of omnivory within populations.
A nowadays frequently used method to characterise aquatic food webs is the analysis of stable isotope ratios (δ 15 N and δ 13 C) as a time-integrating natural tracer of predator–prey interactions (e.g. Bearhop et al., 1999, Davis et al., 2012). These analyses rely on different trophic fractionation of the isotopes when transferring organic matter to higher trophic positions (Brauns, von Schiller & Gergs, 2012). As the proportion of the stable nitrogen isotope ( 15 N) is enriched in consumers relative to their diet, on average 2.3 – 3.4 ‰ per trophiclevel (e.g. Post, 2002, McCutchan et al., 2003), it can be used to estimate trophic positions of organisms relative to a known baseline (Peterson & Fry, 1987, Kling, Fry & O´Brian, 1992, Cabana & Rasmussen, 1996, Post, 2002). By contrast, the stable carbon isotope ( 13 C) is usually less enriched with increasing trophiclevel (< 1 ‰ in most cases), but varies between different carbon sources (Peterson & Fry, 1987, Post, 2002). This fact enables to determine the origin of carbon, such as littoral and pelagic primary production (France, 1995, Hardy et al., 2010). Therefore, stable isotope metrics enable the determination of trophic structures at the community- ______________________________________________ __________________________________________________ __________________________________________________ I General Introduction
model consists of 32 groups and differs from previous models of the Peruvian system through: (i) division of plankton into size groups to account for ENSO changes and feeding preferences of small pelagics, (ii) increased detail of demersal groups and separation of life history stages of hake, (iii) incorporation of mesopelagic fishes, and (iv) incorporation of the jumbo squid (Dosidicus gigas), which has gained in importance since the last El Nino 1997-98. Results show that the El Niño 1997-98 reduced the size and organization of flows of the NHCE, but the overall functioning of the ecosystem was maintained. The reduction of diatom biomass during El Niño forced omnivorous planktivorous fishes to switch to a more zooplankton-dominated diet, increasing their trophiclevel. Consequently, the trophiclevel increased for several predatory groups (mackerel, other large pelagics, sea birds, pinnipeds) and for fishery catch. A high biomass of macrozooplankton was needed to balance the consumption by planktivores, especially during El Niño period when diatoms diminish dramatically. Despite overall lower catches, the higher primary production required-to-catch ratio implied a stronger ecological footprint of the fishery and stresses the need for a precautionary management of fishery resources during and after El Niño. Energetic indicators such as the lower primary production/biomass ratio suggest a more energetically efficient state of the ecosystem, while network indicators such as the lower cycling index and lower relative ascendency are rather indicative of a less organized state of the ecosystem during El Niño. Compared to previous models of the NHCE, this study found: (i) shrinking of ecosystem size in term of flows, (ii) slight changes in overall functioning, and (iii) use of alternate pathways leading to a higher ecological footprint of the fishery.
discrimination for 15 N in Collembola is affected by food quality (N concentration) and metabolism (starvation, life stage) (Ruess et al., 2004; Scheu and Folger, 2004; Haubert et al., 2005). The C:N ratio of the diet has a major impact on fractionation rates and consumers raised on nutrient poor diets show a lower trophic shift in G 15 N (McCutchan et al., 2003). Organisms with carnivorous, herbivorous and mixed diets displayed similar estimates (2.6 to 3.0‰) while organisms consuming detritus yielded significantly lower estimates in G 15 N (0.5‰) (Vanderklift and Ponsard, 2003). This implies primary consumers in detrital food webs to have lower fractionation rates. As the above mentioned studies investigated only few feeding links the mean fractionation for nitrogen of about 3 G units across one trophiclevel may still be a valid approximation when applied on entire food webs with multitrophic pathways and many species (Post, 2002). However, our study indicates that the use of stable isotopes to infer animal diets in belowground systems requires the confirmation by traditional approaches such as food choice experiments or gut content analysis.
trophiclevel . Accordingly, the application of stable isotope analyses might be helpful to elucidate the trophic relationship between a host and its associ- ated parasite. So far, the number of studies based on stable isotope analyses of host-parasite interaction re- mains scarce. Although different parasitic taxa were considered so far (e.g. [13–15]) no information on acanthocephalans is available. Furthermore, the avail- able data show contrasting patterns, with some endo- parasites being depleted in δ 15 N in comparison to their host tissues whereas others were enriched [5, 13, 14, 16–21]. However, often important factors such as de- velopmental stage, feeding strategy and localisation of the parasite within the host were not considered, which are all crucial for the trophic interaction in host-parasite associations.
1994; Jennings et al. 2002). Albeit species with a similar maximum body size can evolve to feed at different trophic levels, there are less small species feeding at high trophic levels than at low trophic levels (Jennings 2005). The overall trophic continuum across body-size classes shows that fixed (integer) trophic levels do not appropriately describe the structure of aquatic food webs (France et al. 1998). Accordingly, Jennings (2005) treats the parameter “trophiclevel” as a continuous measure. If the potential diet of a given predator is defined as a subset of the next largest predator’s diet, then a nested hierarchy of dietary niches will result, as it is widely observed in nature (e.g. Woodward & Hildrew 2002, Cushing et al. 2003). Hierarchical ordering of feeding niches is a central component of the recent niche models that have successfully reproduced many of the topological patterns seen in real food webs, including the prevalence of generalism and omnivory, from a relatively simple set of rules (e.g. Warren 1996, Williams & Martinez 2000, Cattin et al. 2004). If community niche space can be collapsed into a single axis, as suggested by these models, and if that axis is body size, then characterizing the size distributions within a food web will capture much of the biologically meaningful variation in a relatively straightforward manner (Woodward et al. 2005).
Integrated multi-trophic aquacultures (IMTA) are characterized by an increase in the diversity of the selected breeding species compared to traditional monocultures ( Chopin et al., 2012 ), where organisms occupying different trophic levels use the same limited physical space or occupy neighboring areas. They share water bodies (and with them nutrients) via water movements (open ocean) or artificial current systems (inland basins or ponds). In this way, the by-products (organic and inorganic nutrients) of one or more organisms become resources for a lower cultivated trophiclevel ( Chopin et al., 2008 ). Non-multitrophic systems, otherwise, would rely on costly biofiltration apparatuses to control water quality and avoid the release of nutrient-rich waters directly into the natural environment, which would likely affect ecosystem functioning negatively ( Mineur et al., 2015 ), for example, by favoring the increment of opportunistic species eventually promoting algal blooms ( Smetacek and Zingone, 2013 ). Additionally, a more energy-balanced ecosystem has the advantage of preventing or combating disease outbreaks, as seen in the open ocean IMTA of Atlantic salmon, blue mussels (Mytilus edulis), and kelp forest ( Skår and Mortensen, 2007 ; Molloy et al., 2011 ). Benefits were also highlighted in land-based IMTAs, such as high productivity and reduced variability of the protein content of macroalgae, which were related to the constant supply of nutrients between the different compartments in culture and reduced grazing of the algae ( Schuenhoff et al., 2003 ; Mata et al., 2010 ). Taken together, this and the cultivation of more high-value species within the system are often thought to promote economic
For debt contracts, the demand for accounting conservatism is much less for H-share companies. As to compensation contracts, the separation of ownership and control give rise to agency cost (Jensen and Meckling 1976, Fama and Jensen 1983a, 1983b). H-share companies mostly are state-owned companies with the state being the large shareholder. The state is not in a position to directly manage the company, so an agent is designated instead. Here the manager is the representative of the largest shareholder, resulting in immaterial agency problems and less need for conservative reporting to constrain manager’s opportunistic behavior. Furthermore, if the manager decides the terms of the compensation contract, non-pecuniary career development and other “grayer income” are more important. In Chinese state-owned listed companies Chen et al (2005) notes that a regulation exists to limit the size of managerial compensation. Hence, there is a need for non-pecuniary compensation as an alternative choice for managers. Tong (2005) argues that the economic reason for the executive of state-owned companies’ receiving low compensation but remaining is that position is because they can receive certain “gray income” by exercising control rights. As a result, the demand by H-share companies for accounting conservatism stemming from compensation contracts is low. Overall, from a consideration of ﬁrm-level factors, there is not much demand for accounting conservatism in H-share companies.
temporal preferences (but never opposite) depending on the sampling method. These facts suggest that species-realized niche traits may display a certain level of flexibility in relation to their fundamental niche. Niche plasticity may enable species to coexist via niche pre-emption where species may respond to the presence of a superior competitor by switching to an alternative, less used resource (Sanders and Gordon 2003, Ashton et al. 2010). These levels of trait flexibility are of growing concern both in plant and animal ecology (Heemsbergen et al. 2004) as functional traits may vary as much within as between species (Valladares and Niinemets 2008) and, in consequence, impair the explanatory power of trait-based analyses in predictive models (Berg and Ellers 2010). Hence, the question remains whether fixed traits can be generalized for a species with no consideration for the ecological context (Messier et al. 2010). This thesis demonstrated that temporal and dietary preferences may significantly differ between sites (chapter 2), to the extent that certain species niches may be closer to other species than conspecifics, and that the level of plasticity is dependent on the dominance hierarchy established in a community, which is, in turn, dependent on the ecological context. Hence ecological indices which base diversity on morphological characteristics or life history traits (Ilg and Foeckler 2012, Demars et al. 2012, Sternberg and Kennard 2014, Mlambo 2014) must take into consideration the trait plasticity of a species in relation to its ecological context. Especially since these are precisely the type of traits that are highly plastic in response to environmental change (Roff et al. 2002). The different levels of plasticity among species (and within species) convey differences in realized niches that emphasize the functional asynchrony of species, an essential aspect of ecosystem stability (Loreau 2010). This is a vital feature of our understanding of stability and how species within a community may respond to perturbations such as habitat disturbance or climate change. Once again our current knowledge is mostly focused on the response diversity of plants (Laliberté et al. 2010) while little is still known for higher trophic levels. Chapter 2 is only a first step in bridging this gap.
4 A General Necessary Condition
To understand the limits of level-k implementation, we start by showing how a slight strengthening of Bayesian incentive compatibility is necessary as soon as the social choice function is level-k implementable for some arbitrary be- havioral anchors in any mechanism. This has two related implications. First, level-k reasoning does not free us from incentive compatibility constraints, even if the mechanism designer had the ability to choose the anchors in each mechanism. Second, incentive compatibility is a general necessary condition that will hold when studying level-k implementation, regardless of the regular- ity restrictions one is willing to place on behavioral anchors. Of course, such restrictions may generate supplementary necessary conditions, as we will see in Subsection 6.2.
Lake Tanganyika as the second deepest lake in the world belongs to the African Great Lakes. Next to Lake Malawi and Lake Victoria it is a major reservoir of freshwater diversity in Africa, harbouring high proportions of endemic and morphologically distinct taxa, e.g. not only cichlids (for Lake Tanganyika: at least 241 species, 99.17 % of them are endemic, ) but also bivalves and gastropods . These lakes, with their surrounding rivers, have been of high interest because they provide natural laboratories for investigating the drivers of speciation under specific ecological conditions. Here, one ‘engine’ is ecological speciation by adaptive divergence [3–7], which is most likely a key feature particularly in sympatry [8–14]. Evidence for such a process can be parallel evolution, i.e. the development of the same morphologies in closely related but independent lineages under identical extrinsic conditions [8, 15–23]. In adaptive radiations convergent or parallel evolution of phenotypes is a widespread phenomenon. In most scenarios these lineages originated in geographic isolation from each other (e.g. similar ecotypes of anoles lizards on different Caribbean islands [24– 25], stickleback fish in post-glacial lakes [15, 18], spider ecomorphs on Hawaiian Islands [26–27], and cichlid fishes in the East African Lakes Malawi and Tanganyika [28–29]) before colonializing the same habitat. But there are examples of parallel evolution in sympatry as well; recent theoretical models  predict that convergence should be common within all species-rich assemblages. Specifically, these models predict this process for ‘‘species-saturated’’ communities where the number of species exceeds the number of available niches .  described multiple instances of convergence in body and trophic morphology in cichlid fish from Lake Tanganyika and proposed that the parallel phenotypes resulted from the combination of coexistence in the same habitat and the limited number of available niches. Distantly related, yet ecomorphologically similar species show convergent morphologies that can be associated with adaptations to specific, limited habitats and resources and parallel evolution therefore offers an explanation for the greatly elevated species numbers in cichlid species .
In this paper, we described the idea of opin- ion frames as a representation capturing discourse level relations that arise from related opinion tar- gets and which are common in task-oriented di- alogs. We introduced the alternative relations that hold between targets by virtue of being opposing in the discourse context. We discussed how our opinion-frame scheme and discourse relations go hand in hand to provide a richer overall interpreta- tion. We also illustrated that such discourse level opinion associations have useful benefits, namely they help gather more opinion information and help interdependent interpretation. Finally, we showed via our machine learning experiments that the presence of opinion frames can be automati- cally detected.
The analyses showed a significant negative relationship between uncertainty avoidance practices and full-time entrepreneurial activity, but a significantly less negative link with part- time entrepreneurial activity. For full-time entrepreneurship, this result is in line with previous research on how uncertainty avoidance influences entrepreneurial activity (Autio et al., 2013; Shane, 1993) and underscores the idea that the acceptance of uncertainty and risk-taking are supportive of entrepreneurial actions (Hayton and Cacciotti, 2013; Holm et al., 2013; McMullen and Shepherd, 2006). In line with the theoretical reasoning, however, this result also points to the fact that it is not entrepreneurial activity, per se, that is disdained when strong tendencies to avoid uncertainty are present on a societal-level. Instead, it is the high level of investments put at risk when pursuing entrepreneurial endeavors on a full-time basis. This interpretation is also consistent with recent results presented by Raffiee and Feng (2014), which illustrated that individual-level risk aversion more negatively aligns with entrepreneurial activity on a full-time, rather than on a part-time, basis. At least partially, this finding may help to reconcile the previous inconclusive findings on the link between societal- level uncertainty avoidance and entrepreneurial activity (Autio et al., 2013; Hayton and Cacciotti, 2013). More specifically, it suggests that research may come to different conclusions, when focusing on countries where entrepreneurial activities are dominated either by full-time or by part-time entrepreneurs.
As we construct this controller we hope to improve our knowledge of some structures of the brain such as the motor cortex, the pre-frontal cortex, the striatum or the cerebellum. Models of all these structures and other are used in the model here developed. The researches aim especially to better under- stand the influence of each structure on the global behavior of the robot as well as the synergies that emerge from the cooper- ation between structures and to create a new type of humanoid robot where all parts from the technology, through the low level control to the high level control is thought in the optic of realis- tic interactions with humans.
street-level activities such as demonstrations. When centre-right and centre-left parties take up the Islamophobia issue (often in coalition with right-wing parties), this leaves little space for movements that are dedicated to Islamophobia. Hence, a political opportunity framework explains why Pegida-like rallies have not materialized. Islamophobia has become central to mainstream politics in the Visegrád Four, which does not allow for other protests to emerge. It has however allowed several of the ruling governments to direct popular antagonism towards Muslims away from themselves. This is distinct from Western Europe, where Islamophobia is most aggressively used by the right-wing opposition against the political establishment. (The rise of the Lega and the Five Star Movement in Italy, however, invites comparison with the Visegrád Four.)
Die wesentliche Idee der Level-Sets besteht darin, dass Konturen oder Oberflä- chen nicht explizit in ihrer Lage beschrieben werden, wie es bei den Snakes der Fall ist, sondern implizit durch eine Höhenlinie einer höherdimensionalen Funk- tion φ : R dim(I) → R. Diese Funktion ist in der Regel eine vorzeichenbehafte- te Distanzfunktion. Wenn die Nullhöhenlinie als Rand eines Objektes definiert wird, dann ist dessen Kontur implizit durch den Nulldurchgang der Funktion φ beschrieben. Für eine ausführliche Darstellung sei hier auf Kapitel 3 verwiesen. Die Level-Set-Methoden befassen sich mit der iterativen, zeitlichen Entwicklung der Distanzfunktion und damit mit der Deformation der beschriebenen Oberflä- chen in Form von partiellen Differentialgleichungen. Die ursprüngliche Formulie- rung wurde von Osher und Sethian 1988 veröffentlicht [Osh88], im gleichen Jahr wie die aktiven Konturmodelle von Kass et al. [Kas88]. Begründet auf der Änh- lichkeit der Modelle wurden die Level-Sets zunächst von Caselles in [Cas93] im Bereich Computer Vision für die Kontur- bzw. Oberflächenmodellierung verwen- det. Ähnliche Ansätze folgten durch Malladi in [Mal93]. Eine Überführung der Modelle ineinander erfolgte wiederum über die geodätischen aktiven Konturen von Caselles [Cas97].