It is a widely accepted assumption that fish predation controls structure and functioning of aquatic food webs. In the past, however, a large part of effort was concentrated on lakes and reservoirs. Thus, the knowledge about stream ecosystems is much more rudimentary than that for lakes in this respect. The aim of this thesis, therefore, was to describe and assess the effects of fish predation in natural stream ecosystems. For that purpose a reach scale field experiment was set up using an experimental stretch with benthivorous fish and a fishless reference stretch. A wide range of effects of the fish predators on their stream invertebrates prey was studied. To discriminate between lethal and sublethal predation effects, measuring the physiological status of the organisms seemed promising. However, before it was possible to decide whether or not environmental stress, such as predation, might affect the physiological status, the internal control as well as the seasonal and species-specific variability of the energy amount stored had to be assessed. Thus, the concentration and seasonal dynamics of the major energy storage components triglycerides and glycogen were measured in two species of mayflies (Rhithrogena semicolorata and Ephemera danica) with contrasting life cycle strategies. E. danica is a burrowing, semivoltine collector-gatherer, R. semicolorata is univoltine and scrapes periphyton from stones. Although triglycerides are the major energy reserve in both species throughout the whole larval development (> 84 % of total energy storage) their seasonal dynamic differed considerably. In R. semicolorata the triglyceride concentration declined during the last weeks prior to emergence in both sexes. The same pattern was found in female larvae of E. danica, but not in male E. danica. It is suggested that females use triglycerides in the last larval stages for egg maturation, which is completed in the last larval instar. In male E. danica the triglyceride concentrations remained high until emergence, presumably due to their high energy demands as adults for their swarming flights and mating. The difference in seasonal variation of triglycerides between E. danica and R. semicolorata shows the influence of environmental factors on the dynamics of storage components. E. danica lived in a very stable environment (within the substratum). Therefore the dynamic of energy storage components was optimised with respect to maximal reproduction. R. semicolorata on the other hand, suffered from hostile environmental factors such as predation or food limitation due to low periphyton biomass after leaf sprout and following light limitation in spring. Consequently, the concentration of storage components decreased during spring. One conclusion from this study was that the measurement of storage components might reveal sublethal predation effects. However, season and sex of the organisms are important factors as well and have to be considered in the sampling design. To analyse sublethal predation effects behavioural changes due to the presence of benthivorous fish were measured. Drift as a low-energy cost means of migration may enable stream invertebrates to leave risky habitats or may even be a direct escape reaction after a predator encounter. While the control of drift activity by predators has received considerable interest from many researchers, it remains still unclear whether predators reduce or increase drift activity. Drift activity of stream invertebrates was influenced significantly by the presence or absence the two benthivorous fish species gudgeon (Gobio gobio) and stone loach (Barbatula barbatula). Contrary to previous studies gudgeon and stone loach reduced invertebrate drift density and drift activity of Baetis rhodani rather than inducing higher night-time drift. Further, species composition of the invertebrate drift differed significantly between the two stretches. A further conclusion from this study is therefore that drift is not generally a mechanism of active escape from benthos-feeding fish, as previously assumed. In addition, the reduced drift activity in the fish stretch might result in a compensation of the consumptive losses due to fish predation. Thus, in this study design the effects of fish predation on invertebrate community might be underestimated. To detect predation effects on the food web structure the reactions of the grazing mayfly Rhithrogena semicolorata and the shredding amphipod Gammarus pulex to strong predation by benthivorous fish were compared. It has been hypothesised that shredders are generally less vulnerable to fish predation and therefore less likely to be predation-controlled than grazers, because the latter are visible to the predators during their feeding on stone surfaces, while shredders may hide between leaves during foraging. Biomass of G. pulex was significantly reduced in the fish stretch while that of R. semicolorata was not. Since approximately 91 % of the annual production of G. pulex but only 12 % of R. semicolorata production was consumed by benthivorous gudgeon, the observed difference of G. pulex biomass between the fish and reference reach is likely due to a lethal predation effect. However, no sublethal predation effects such as reduced concentration of storage components (triglycerides, glycogen) or reduced reproductive success were observed for both species. Hence, in contrast to the initial hypothesis, in the studied stream the shredder was top-down-controlled, while the grazer was not. It is concluded that top-down control depends on the ecological characteristics of a specific predator-prey pair rather than on trophic guild of the prey. To assess the predation effects on the life history of merolimnic insects and its consequences on fecundity the larval development and emergence of R. semicolorata was studied. It was possible to show lethal and sublethal effects of predation by benthivorous fish (Gobio gobio, Barbatula barbatula). Predation consequently resulted in changes of larval development and population fitness. The presence of two benthivorous fish species (gudgeon and stone loach) led to slower larval development and a delayed emergence. However, no differences in the adult size and fecundity between the fish reach and the reference were observed. Nevertheless, the longer time spent in the larval phase resulted in a higher mortality and therefore in a lower mean population fitness. The presence of gudgeon alone, however, did not seem to influence larval development, growth or time of emergence and consequently fecundity. Further, strong lethal impact of gudgeon could not be detected. Thus, the population fitness measured as the product of adult density and egg number was not reduced by gudgeon alone. It is assumed that the stronger lethal impact in the combined fish experiment is caused mainly by stone loach because the proportion of mayfly consumption by stone loach to mayfly production shortly before emergence was higher than the proportion related to gudgeon. Thus another conclusion is that 1) the impact of predation seems to differ for the fish species and 2) lethal effects have a stronger impact on the population survival than life history changes. Combining the results mentioned above leads to the assumption that predation by benthivorous fish has the potential to shape invertebrate communities and food webs in streams. It was possible to show reductions of benthic densities and mean population fitness. The strength of trophic interactions seemed to be specific for the single predator-prey pairs here. Finally, it can be stated that contrary to previous assumptions consumption of the fish predators seemed to be more important for the prey populations than sublethal predation effects.

It is a widely accepted assumption that fish predation controls structure and functioning of aquatic food webs. In the past, however, a large part of effort was concentrated on lakes and reservoirs. Thus, the knowledge about stream ecosystems is much more rudimentary than that for lakes in this respect. The aim of this thesis, therefore, was to describe and assess the effects of fish predation in natural stream ecosystems. For that purpose a reach scale field experiment was set up using an experimental stretch with benthivorous fish and a fishless reference stretch. A wide range of effects of the fish predators on their stream invertebrates prey was studied. To discriminate between lethal and sublethal predation effects, measuring the physiological status of the organisms seemed promising. However, before it was possible to decide whether or not environmental stress, such as predation, might affect the physiological status, the internal control as well as the seasonal and species-specific variability of the energy amount stored had to be assessed. Thus, the concentration and seasonal dynamics of the major energy storage components triglycerides and glycogen were measured in two species of mayflies (Rhithrogena semicolorata and Ephemera danica) with contrasting life cycle strategies. E. danica is a burrowing, semivoltine collector-gatherer, R. semicolorata is univoltine and scrapes periphyton from stones. Although triglycerides are the major energy reserve in both species throughout the whole larval development (> 84 % of total energy storage) their seasonal dynamic differed considerably. In R. semicolorata the triglyceride concentration declined during the last weeks prior to emergence in both sexes. The same pattern was found in female larvae of E. danica, but not in male E. danica. It is suggested that females use triglycerides in the last larval stages for egg maturation, which is completed in the last larval instar. In male E. danica the triglyceride concentrations remained high until emergence, presumably due to their high energy demands as adults for their swarming flights and mating. The difference in seasonal variation of triglycerides between E. danica and R. semicolorata shows the influence of environmental factors on the dynamics of storage components. E. danica lived in a very stable environment (within the substratum). Therefore the dynamic of energy storage components was optimised with respect to maximal reproduction. R. semicolorata on the other hand, suffered from hostile environmental factors such as predation or food limitation due to low periphyton biomass after leaf sprout and following light limitation in spring. Consequently, the concentration of storage components decreased during spring. One conclusion from this study was that the measurement of storage components might reveal sublethal predation effects. However, season and sex of the organisms are important factors as well and have to be considered in the sampling design. To analyse sublethal predation effects behavioural changes due to the presence of benthivorous fish were measured. Drift as a low-energy cost means of migration may enable stream invertebrates to leave risky habitats or may even be a direct escape reaction after a predator encounter. While the control of drift activity by predators has received considerable interest from many researchers, it remains still unclear whether predators reduce or increase drift activity. Drift activity of stream invertebrates was influenced significantly by the presence or absence the two benthivorous fish species gudgeon (Gobio gobio) and stone loach (Barbatula barbatula). Contrary to previous studies gudgeon and stone loach reduced invertebrate drift density and drift activity of Baetis rhodani rather than inducing higher night-time drift. Further, species composition of the invertebrate drift differed significantly between the two stretches. A further conclusion from this study is therefore that drift is not generally a mechanism of active escape from benthos-feeding fish, as previously assumed. In addition, the reduced drift activity in the fish stretch might result in a compensation of the consumptive losses due to fish predation. Thus, in this study design the effects of fish predation on invertebrate community might be underestimated. To detect predation effects on the food web structure the reactions of the grazing mayfly Rhithrogena semicolorata and the shredding amphipod Gammarus pulex to strong predation by benthivorous fish were compared. It has been hypothesised that shredders are generally less vulnerable to fish predation and therefore less likely to be predation-controlled than grazers, because the latter are visible to the predators during their feeding on stone surfaces, while shredders may hide between leaves during foraging. Biomass of G. pulex was significantly reduced in the fish stretch while that of R. semicolorata was not. Since approximately 91 % of the annual production of G. pulex but only 12 % of R. semicolorata production was consumed by benthivorous gudgeon, the observed difference of G. pulex biomass between the fish and reference reach is likely due to a lethal predation effect. However, no sublethal predation effects such as reduced concentration of storage components (triglycerides, glycogen) or reduced reproductive success were observed for both species. Hence, in contrast to the initial hypothesis, in the studied stream the shredder was top-down-controlled, while the grazer was not. It is concluded that top-down control depends on the ecological characteristics of a specific predator-prey pair rather than on trophic guild of the prey. To assess the predation effects on the life history of merolimnic insects and its consequences on fecundity the larval development and emergence of R. semicolorata was studied. It was possible to show lethal and sublethal effects of predation by benthivorous fish (Gobio gobio, Barbatula barbatula). Predation consequently resulted in changes of larval development and population fitness. The presence of two benthivorous fish species (gudgeon and stone loach) led to slower larval development and a delayed emergence. However, no differences in the adult size and fecundity between the fish reach and the reference were observed. Nevertheless, the longer time spent in the larval phase resulted in a higher mortality and therefore in a lower mean population fitness. The presence of gudgeon alone, however, did not seem to influence larval development, growth or time of emergence and consequently fecundity. Further, strong lethal impact of gudgeon could not be detected. Thus, the population fitness measured as the product of adult density and egg number was not reduced by gudgeon alone. It is assumed that the stronger lethal impact in the combined fish experiment is caused mainly by stone loach because the proportion of mayfly consumption by stone loach to mayfly production shortly before emergence was higher than the proportion related to gudgeon. Thus another conclusion is that 1) the impact of predation seems to differ for the fish species and 2) lethal effects have a stronger impact on the population survival than life history changes. Combining the results mentioned above leads to the assumption that predation by benthivorous fish has the potential to shape invertebrate communities and food webs in streams. It was possible to show reductions of benthic densities and mean population fitness. The strength of trophic interactions seemed to be specific for the single predator-prey pairs here. Finally, it can be stated that contrary to previous assumptions consumption of the fish predators seemed to be more important for the prey populations than sublethal predation effects.

The prey consumption by benthivorous fish predators can have profound top-down effects in stream food webs. To analyse this effects in small temperate stream ecosystems, a long-term field experiment was conducted in two streams in South-eastern Germany, Gauernitzbach and Tännichtgrundbach, from 2004 to 2011. The densities of two small-bodied benthivorous fish species, gudgeon (Gobio gobio) and stone loach (Barbatula barbatula), were manipulated following a Before-After-Control-Impact design. The top predator regime affected the benthic community composition of the streams mainly in pools, whereas the total benthic invertebrate biomass was not affected in any mesohabitat. The present work describes a causal analysis of the observed food web effects using additional field analyses and laboratory experiments, with a special focus on the habitat use and foraging behaviour of the fish as top predators. The probably most important of the analysed mechanisms was mesohabitat-specific predation by the fish. Three 24-h field video surveys combined with benthic invertebrate sampling showed that constraints in habitat use, especially for gudgeon, induced a differential predator-prey habitat overlap which resulted in a higher predation risk for the invertebrate prey in pools than in riffles. Another important mechanism was selective predation of both fish species. Their prey selectivity was largely explained by a small number of prey variables being connected to the partly non-visual foraging mode of these benthic predators. Besides the traits body size and feeding type, long-term mean abundance played a central role, small and highly abundant invertebrates, grazers and sediment feeders being preferred by gudgeon and stone loach. The preference for small and abundant prey taxa (chironomids) exceeded purely opportunistic feeding and probably facilitated resource partitioning between the two fish species having very similar diets. In addition to active selectivity, different predator avoidance strategies of the invertebrates analysed in laboratory experiments explained the passive selectivity of the fish predators for certain prey taxa in the streams. This could be shown for two abundant taxa being consumed by the fish predators in very different quantities, Gammarus pulex and Hydropsyche instabilis. These three mechanisms, although probably interacting with several other factors, could explain a large part of the effects the top-down food web manipulation had on the benthic community, especially the observed high degree of mesohabitat and species specificity. Confirming this, quantitative characteristics of predation food webs, for instance the importance of intraguild predation, differed markedly between pool and riffle mesohabitats. From the results of this study it can be concluded that the benthivorous fish affected benthic community structure mainly by mesohabitat-specific and selective predation. A manipulation of this (native) top predator type therefore will probably have such rather subtle but not catastrophic consequences in ecosystems with a high biotic diversity and a rich natural habitat structure and connectivity.

Biomanipulation im Sinne einer top-down Steuerung des Nahrungsnetzes wurde in Standgewässern intensiv erforscht und wird als Mittel zur Verbesserung der Wasserqualität bzw. Verringerung der Eutrophierungserscheinungen eingesetzt. Ebenso könnte die Idee der Biomanipulation genutzt werden, um die Eutrophierung in kleinen Fließgewässern zu verringern. Eine Förderung benthischer herbivorer Schlüsselorganismen (z.B. Eintagsfliegenlarven) durch die Reduzierung des Fischfraßdruckes könnte helfen, Algenmassenentwicklungen zu vermeiden. Studien zur Biomanipulation von Nahrungsnetzen in Standgewässern zeigten allerdings eine zunehmende Entwicklung von planktischen invertebraten Räubern bei Abwesenheit planktivorer Fische. Die Bedeutung des Fraßdruckes invertebrater Räuber in fischfreien Fließgewässernahrungsnetzen ist dagegen noch unbekannt. Aufgrund von letalen und subletalen Fischeffekten könnte die Konsumtion räuberischer Invertebraten bei Fischabwesenheit so stark gefördert werden, dass das Ziel der Biomanipulation verfehlt wird. Invertebrate Räuber haben insgesamt einen relativ hohen Anteil an der Biomasse der Benthosgemeinschaft in Fließgewässernahrungsnetzen. Sie könnten somit potentiell einen ähnlichen Fraßdruck wie vertebrate Räuber (z.B. benthivore Fische) auf die unteren trophischen Ebenen ausüben. Auch der Anteil der Omnivorie ist in Fließgewässernahrungsnetzen hoch. Ziel dieser Arbeit war es deshalb, die Bedeutung invertebrater Prädation im Nahrungsnetz eines kleinen Fließgewässers (Gauernitzbach) unter Berücksichtigung der obligaten Räuber sowie des räuberischen Potentials Omnivorer abzuschätzen. Außerdem wurde der Einfluss benthivorer Fische auf die Populationsentwicklung und das Ernährungsverhalten der wichtigsten invertebraten Räuber und Omnivoren untersucht. Dazu wurde ein Bachabschnitt, der mit benthivoren Fischen (Gobio gobio, Barbatula barbatula) besetzt war, mit einem fischfreien Abschnitt über zwei Jahre verglichen. Nach der Identifizierung wichtiger Räuber- und Omnivorenarten anhand ihrer Biomasse und ihrer trophischen Positionen mittels Isotopenanalyse wurde ihr Gesamtfraßdruck abgeschätzt. Invertebrate Prädatoren im Nahrungsnetz des Gauernitzbaches waren Rhyacophila fasciata und Plectrocnemia conspersa (Trichoptera), Isoperla grammatica (Plecoptera) sowie Dugesia gonocephala (Turbellaria). Wichtige Omnivore aufgrund ihrer hohen Biomassen waren Gammarus pulex (Amphipoda) und Hydropsyche spp. (Trichoptera). Anhand eines Mischungsmodells (ISOSOURCE), das auf der Basis der stabilen Isotopensignaturen d13C und d15N die einzelnen Ressourcenanteile einer gemischten Signatur berechnet, konnte das ausgeprägte räuberische Potential der Omnivoren gezeigt werden, da sie einen hohen tierischen Anteil in der Nahrung aufwiesen (20-90 %). Im Nahrungsnetz des Gauernitzbaches waren Omnivore als Räuber damit ebenso bedeutsam wie obligat räuberische Invertebrate. Der Gesamtfraßdruck der invertebraten Räuber und Omnivoren auf die Beuteorganismen (bis 16 g TM m-2 a-1) in der Fischstrecke war höher als der Fraßdruck der benthivoren Fische (~3 g TM m-2 a-1). Dieser Unterschied lag hauptsächlich in höheren täglichen Konsumtionsraten bei ähnlich hohen Biomassen der Invertebraten begründet. Durch die Analyse der stabilen Isotope von Kohlenstoff und Stickstoff konnten die trophischen Beziehungen im Nahrungsnetz des Gauernitzbaches identifiziert werden. Dabei stellten allochthone Kohlenstoffquellen im Frühjahr und Herbst die bedeutendste Ressource für die Primärkonsumenten dar, während autochthone Ressourcen nur im Frühjahr für wenige Herbivore relevant waren. Dies spiegelte sich ebenso in der Nahrung der Omnivoren wider, die neben Organismen hauptsächlich Detritus nutzten. Ebenso zeigten die invertebraten Räuber eine deutliche Abhängigkeit vom allochthonen Kohlenstoff über die Nutzung der Primärkonsumenten. Letale Effekte der Fische auf die räuberischen und omnivoren Populationen im Gauernitzbach fanden hauptsächlich in Abhängigkeit von den bewohnten Habitaten statt. Die Pool-Arten (P. conspersa, D. gonocephala, G. pulex) entwickelten geringere Biomassen in der Fischstrecke, während die Biomassen riffle-bewohnender Invertebraten (R. fasciata, I. grammatica, Hydropsyche spp.) nicht deutlich reduziert wurden. Dabei waren allerdings meist hohe Biomassen in der Fischstrecke weniger stark ausgeprägt als in der fischfreien Strecke. Die Adulten der untersuchten Arten wurden oft stärker letal beeinflusst als die Juvenilstadien. So war die Emergenzproduktion der räuberischen Köcherfliegen und die Abundanz großer und eiertragender Weibchen von G. pulex in der Fischstrecke deutlich verringert. Subletale Effekte der Fische traten im Gauernitzbach in Abhängigkeit von den Ernährungstypen auf. Während die adulten räuberischen Köcherfliegenarten R. fasciata und P. conspersa keine negative Beeinflussung der Fitness und Größe zeigten, emergierte die omnivore Köcherfliege H. instabilis in der Fischstrecke kleiner als in der fischfreien. Außerdem unterschieden sich die Antworten der omnivoren Arten vermutlich aufgrund ihrer unterschiedlichen Lebenszyklen. G. pulex zeigte als ausschließlich aquatisch lebende Art keine subletalen Kosten (unveränderte Fitness) bei Fischanwesenheit wie dies für die omnivore merolimnische Insektenart H. instabilis beobachtet wurde. Die mittlere Fitness der Populationen (gemessen an der Eiproduktion pro Zeit- und Flächeneinheit) war für alle untersuchten invertebraten Räuber und Omnivoren bei Fischanwesenheit deutlich geringer als in der fischfreien Strecke. Für die Räuber (R. fasciata, P. conspersa) und G. pulex waren vermutlich letale Fischeffekte für den geringen Reproduktionserfolg verantwortlich. Dagegen wurde die Fitness der Population von H. instabilis durch die Kombination aus der geringeren individuellen Körpergröße (subletal) und den schwachen letalen Effekten auf die Emergenz reduziert. Der Einfluss der Fische auf die räuberischen Larven von R. fasciata und P. conspersa resultierte in einem Wechsel der Beutezusammensetzung, nicht in der Reduzierung der täglichen Fraßaktivität. Dies war vermutlich auf eine Aktivitätsänderung des Räubers R. fasciata bzw. auf die veränderte Beuteverfügbarkeit für P. conspersa unter Fischeinfluss zurückzuführen. Eine grundsätzliche Verringerung der Fraßaktivität oder der tierischen Anteile in der Nahrung der omnivoren Arten unter dem Einfluss benthivorer Fische wurde nicht festgestellt. Dagegen änderten sich die Relationen der Ressourcen Omnivorer zueinander, so konnte in der fischfreien Strecke eine flexiblere Ressourcennutzung nachgewiesen werden. Die jährliche Gesamtkonsumtion der räuberischen und omnivoren Invertebraten wurde durch die Anwesenheit der benthivoren Fische um 20-50 % reduziert, wofür hauptsächlich fischinduzierte Biomasseänderungen verantwortlich waren. In der fischfreien Strecke konnte dagegen die erhöhte Konsumtion der Invertebraten den fehlenden Fischfraßdruck mehr als kompensieren. Allerdings änderte sich die beutespezifische Konsumtion der Räuber in der fischfreien Strecke durch den verhaltensinduzierten Wechsel in der Beutezusammensetzung. Damit fand ein erhöhter Fraßdruck auf detritusfressende Arten (z.B. Diptera) statt, während die vermuteten Schlüsselarten einer potentiellen Biomanipulation (grazende Eintagsfliegen) in der fischfreien Strecke relativ unbeeinflusst blieben. Da die invertebrate Prädation in Fließgewässern sehr bedeutend sein kann, ist Biomanipulation nur unter Berücksichtigung aller Populationen räuberischer Invertebraten möglich. Nur durch Einstellung einer optimalen (relativ hohen) Biomasse geeigneter benthivorer Fischarten kann eine ausreichende Hemmung der invertebraten Räuber erreicht werden, so dass grazende Arten indirekt profitieren. / Biomanipulation (top-down control of the food webs) in lakes was intensively studied and can today be regarded as an important tool for improvement of the water quality and eutrophication. Similarly, the idea of biomanipulation can be used to reduce eutrophication phenomena in streams. The enhancement of benthic herbivorous key species (e.g. mayflies) by reduction of fish predation pressure could prevent exploding benthic algae biomass. Studies dealing with trophic cascades in lakes suggested that biomass of planktic invertebrate predators is increasing in the absence of planktivorous fish. But until now the importance of invertebrate predation pressure in fishless stream food webs is hardly known. Due to lethal and sublethal fish effects the consumption of predatory invertebrates could be enhanced by the absence of fish to such an extent that biomanipulation could be off target. In general, the proportion of invertebrate predators to the biomass of the benthic community in stream food webs is relatively high. Therefore, invertebrate predation on benthic prey can be as important as vertebrate predation (e.g. by benthivorous fish). Furthermore, omnivory occurs frequently in streams. Therefore, the aim of this thesis was to assess the importance of invertebrate predation in a food web of a small stream (Gauernitzbach) including the obligatory predators as well as the predatory potential of omnivores. Furthermore, the impact of benthivorous fish on population dynamics and feeding behaviour of the most important predators and omnivores was studied. Therefore, a reach stocked with benthivorous fish (Gobio gobio, Barbatula barbatula) was compared with an upstream fishless reach over a two-year period. After identification of important predators and omnivores based on their biomass and trophic position by stable isotope analyses their total consumption was estimated. Rhyacophila fasciata and Plectrocnemia conspersa (Trichoptera), Isoperla grammatica (Plecoptera) as well as Dugesia gonocephala (Turbellaria) were the invertebrate predators in the food web of Gauernitzbach. Because of their high biomasses the most important omnivores were Gammarus pulex (Amphipoda) and Hydropsyche spp. (Trichoptera). Using a mixing model (ISOSOURCE) that estimates the proportions of single resources on a mixed signature on the basis of the stable 13C and 15N isotopes, the predatory potential of both omnivores was pronounced by a high animal food proportion (20-90%). In this stream food web the predation impact of omnivores was comparable to this of obligatory predacous invertebrates. The total invertebrate predation pressure (up to 16 g TM m-2 a-1) in the fish reach was higher than the vertebrate predation pressure (~3 g TM m-2 a-1) mainly caused by the higher daily consumption rates and similar biomass compared to fish. Using stable isotope analyses of carbon and nitrogen the trophic pathways in the food web of Gauernitzbach could be identified. The most important resources for primary consumers in spring and autumn were allochthonous carbon sources. Autochtnonous carbon sources were only utilized by a few herbivores in spring. This was also reflected in the food of omnivores which assimilated mainly detritus beside animal material. Similarly, the predators depended distinctly on the allochthonous pathway by feeding primary consumers. Lethal fish effects on the predatory and omnivorous invertebrate populations were mainly depended on their preferred habitats. Pool-dwelling species (P. conspersa, D. gonocephala, G. pulex) showed decreased biomass in the presence of fish. Contrary, the biomass of riffle-dwelling invertebrates (R. fasciata, I. grammatica, Hydropsyche spp.) was not significantly reduced in the fish reach. Thereby, the highest biomass values in the fish reach could not reach the same level as in the fishless reach. Adults of the investigated species were influenced more lethal than the young stages. Hence, the emergence production of the predatory caddisflies and the abundance of the biggest and gravid females of G. pulex were reduced in the fish reach. Sublethal fish effects existed in dependence on the feeding groups of the invertebrates in Gauernitzbach. Whereas the adult predatory caddisflies R. fasciata and P. conspersa did not show disadvantages in fitness and size, the omnivorous caddisfly H. instabilis emerged with smaller size in the fish reach than in the fishless reach. Different responses of both omnivorous species were observed because of their different life cycles. The exclusively aquatic living G. pulex was not influenced sublethal by fish as the merolimnic insect species H. instabilis. The average fitness of population (measured as egg production per unit time and space) of the studied invertebrate predators and omnivores was distinctly reduced in the fish reach compared to the fishless reach. Probably, lethal fish effects were mainly responsible for the low reproductive success of the predatory species (R. fasciata, P. conspersa) and G. pulex. In contrast, the reduction of population fitness of H. instabilis was caused by the reduced individual adult size (sublethal effect) in combination with the weak lethal effect on emergence. The impact of fish on the feeding behaviour of the predatory larvae of R. fasciata and P. conspersa resulted in a switched prey composition, whereas a reduction in the daily feeding activity was not observed. That was supposed because of changes in the activity pattern by the predator R. fasciata or the changed prey availability for P. conspersa in the fish reach. A general reduction in feeding activity or animal food proportion of the omnivores was not assessed in the presence of fish. In contrast, the relations between the resources switched to a more variable utilisation in the fishless reach. The annual total consumption of invertebrate predators and omnivores was decreased by 20-50% in the fish reach, mainly caused by fish induced biomass reduction. The missing fish consumption could be at least balanced by higher invertebrate consumption in the fishless reach. Contrary, the prey specific consumption of invertebrate predators changed by behavioural-induced switch in the prey composition between the stream reaches. Therefore, a higher predation pressure on detritus feeding species (e.g Diptera) was found in the fishless reach, but not on the supposed key species of the potential biomanipulation (e.g. grazing mayflies). Due to the high importance of invertebrate predation in stream food webs, the aim of biomanipulation can only be reached by considering the populations of predatory invertebrates. Controlling the biomass of suitable benthivorous fish to an optimum (relatively high) level could sufficiently reduce the impact of invertebrate predators leading to an indirect enhancement of grazing species.

The prey consumption by benthivorous fish predators can have profound top-down effects in stream food webs. To analyse this effects in small temperate stream ecosystems, a long-term field experiment was conducted in two streams in South-eastern Germany, Gauernitzbach and Tännichtgrundbach, from 2004 to 2011. The densities of two small-bodied benthivorous fish species, gudgeon (Gobio gobio) and stone loach (Barbatula barbatula), were manipulated following a Before-After-Control-Impact design. The top predator regime affected the benthic community composition of the streams mainly in pools, whereas the total benthic invertebrate biomass was not affected in any mesohabitat. The present work describes a causal analysis of the observed food web effects using additional field analyses and laboratory experiments, with a special focus on the habitat use and foraging behaviour of the fish as top predators. The probably most important of the analysed mechanisms was mesohabitat-specific predation by the fish. Three 24-h field video surveys combined with benthic invertebrate sampling showed that constraints in habitat use, especially for gudgeon, induced a differential predator-prey habitat overlap which resulted in a higher predation risk for the invertebrate prey in pools than in riffles. Another important mechanism was selective predation of both fish species. Their prey selectivity was largely explained by a small number of prey variables being connected to the partly non-visual foraging mode of these benthic predators. Besides the traits body size and feeding type, long-term mean abundance played a central role, small and highly abundant invertebrates, grazers and sediment feeders being preferred by gudgeon and stone loach. The preference for small and abundant prey taxa (chironomids) exceeded purely opportunistic feeding and probably facilitated resource partitioning between the two fish species having very similar diets. In addition to active selectivity, different predator avoidance strategies of the invertebrates analysed in laboratory experiments explained the passive selectivity of the fish predators for certain prey taxa in the streams. This could be shown for two abundant taxa being consumed by the fish predators in very different quantities, Gammarus pulex and Hydropsyche instabilis. These three mechanisms, although probably interacting with several other factors, could explain a large part of the effects the top-down food web manipulation had on the benthic community, especially the observed high degree of mesohabitat and species specificity. Confirming this, quantitative characteristics of predation food webs, for instance the importance of intraguild predation, differed markedly between pool and riffle mesohabitats. From the results of this study it can be concluded that the benthivorous fish affected benthic community structure mainly by mesohabitat-specific and selective predation. A manipulation of this (native) top predator type therefore will probably have such rather subtle but not catastrophic consequences in ecosystems with a high biotic diversity and a rich natural habitat structure and connectivity.

Biomanipulation im Sinne einer top-down Steuerung des Nahrungsnetzes wurde in Standgewässern intensiv erforscht und wird als Mittel zur Verbesserung der Wasserqualität bzw. Verringerung der Eutrophierungserscheinungen eingesetzt. Ebenso könnte die Idee der Biomanipulation genutzt werden, um die Eutrophierung in kleinen Fließgewässern zu verringern. Eine Förderung benthischer herbivorer Schlüsselorganismen (z.B. Eintagsfliegenlarven) durch die Reduzierung des Fischfraßdruckes könnte helfen, Algenmassenentwicklungen zu vermeiden. Studien zur Biomanipulation von Nahrungsnetzen in Standgewässern zeigten allerdings eine zunehmende Entwicklung von planktischen invertebraten Räubern bei Abwesenheit planktivorer Fische. Die Bedeutung des Fraßdruckes invertebrater Räuber in fischfreien Fließgewässernahrungsnetzen ist dagegen noch unbekannt. Aufgrund von letalen und subletalen Fischeffekten könnte die Konsumtion räuberischer Invertebraten bei Fischabwesenheit so stark gefördert werden, dass das Ziel der Biomanipulation verfehlt wird. Invertebrate Räuber haben insgesamt einen relativ hohen Anteil an der Biomasse der Benthosgemeinschaft in Fließgewässernahrungsnetzen. Sie könnten somit potentiell einen ähnlichen Fraßdruck wie vertebrate Räuber (z.B. benthivore Fische) auf die unteren trophischen Ebenen ausüben. Auch der Anteil der Omnivorie ist in Fließgewässernahrungsnetzen hoch. Ziel dieser Arbeit war es deshalb, die Bedeutung invertebrater Prädation im Nahrungsnetz eines kleinen Fließgewässers (Gauernitzbach) unter Berücksichtigung der obligaten Räuber sowie des räuberischen Potentials Omnivorer abzuschätzen. Außerdem wurde der Einfluss benthivorer Fische auf die Populationsentwicklung und das Ernährungsverhalten der wichtigsten invertebraten Räuber und Omnivoren untersucht. Dazu wurde ein Bachabschnitt, der mit benthivoren Fischen (Gobio gobio, Barbatula barbatula) besetzt war, mit einem fischfreien Abschnitt über zwei Jahre verglichen. Nach der Identifizierung wichtiger Räuber- und Omnivorenarten anhand ihrer Biomasse und ihrer trophischen Positionen mittels Isotopenanalyse wurde ihr Gesamtfraßdruck abgeschätzt. Invertebrate Prädatoren im Nahrungsnetz des Gauernitzbaches waren Rhyacophila fasciata und Plectrocnemia conspersa (Trichoptera), Isoperla grammatica (Plecoptera) sowie Dugesia gonocephala (Turbellaria). Wichtige Omnivore aufgrund ihrer hohen Biomassen waren Gammarus pulex (Amphipoda) und Hydropsyche spp. (Trichoptera). Anhand eines Mischungsmodells (ISOSOURCE), das auf der Basis der stabilen Isotopensignaturen d13C und d15N die einzelnen Ressourcenanteile einer gemischten Signatur berechnet, konnte das ausgeprägte räuberische Potential der Omnivoren gezeigt werden, da sie einen hohen tierischen Anteil in der Nahrung aufwiesen (20-90 %). Im Nahrungsnetz des Gauernitzbaches waren Omnivore als Räuber damit ebenso bedeutsam wie obligat räuberische Invertebrate. Der Gesamtfraßdruck der invertebraten Räuber und Omnivoren auf die Beuteorganismen (bis 16 g TM m-2 a-1) in der Fischstrecke war höher als der Fraßdruck der benthivoren Fische (~3 g TM m-2 a-1). Dieser Unterschied lag hauptsächlich in höheren täglichen Konsumtionsraten bei ähnlich hohen Biomassen der Invertebraten begründet. Durch die Analyse der stabilen Isotope von Kohlenstoff und Stickstoff konnten die trophischen Beziehungen im Nahrungsnetz des Gauernitzbaches identifiziert werden. Dabei stellten allochthone Kohlenstoffquellen im Frühjahr und Herbst die bedeutendste Ressource für die Primärkonsumenten dar, während autochthone Ressourcen nur im Frühjahr für wenige Herbivore relevant waren. Dies spiegelte sich ebenso in der Nahrung der Omnivoren wider, die neben Organismen hauptsächlich Detritus nutzten. Ebenso zeigten die invertebraten Räuber eine deutliche Abhängigkeit vom allochthonen Kohlenstoff über die Nutzung der Primärkonsumenten. Letale Effekte der Fische auf die räuberischen und omnivoren Populationen im Gauernitzbach fanden hauptsächlich in Abhängigkeit von den bewohnten Habitaten statt. Die Pool-Arten (P. conspersa, D. gonocephala, G. pulex) entwickelten geringere Biomassen in der Fischstrecke, während die Biomassen riffle-bewohnender Invertebraten (R. fasciata, I. grammatica, Hydropsyche spp.) nicht deutlich reduziert wurden. Dabei waren allerdings meist hohe Biomassen in der Fischstrecke weniger stark ausgeprägt als in der fischfreien Strecke. Die Adulten der untersuchten Arten wurden oft stärker letal beeinflusst als die Juvenilstadien. So war die Emergenzproduktion der räuberischen Köcherfliegen und die Abundanz großer und eiertragender Weibchen von G. pulex in der Fischstrecke deutlich verringert. Subletale Effekte der Fische traten im Gauernitzbach in Abhängigkeit von den Ernährungstypen auf. Während die adulten räuberischen Köcherfliegenarten R. fasciata und P. conspersa keine negative Beeinflussung der Fitness und Größe zeigten, emergierte die omnivore Köcherfliege H. instabilis in der Fischstrecke kleiner als in der fischfreien. Außerdem unterschieden sich die Antworten der omnivoren Arten vermutlich aufgrund ihrer unterschiedlichen Lebenszyklen. G. pulex zeigte als ausschließlich aquatisch lebende Art keine subletalen Kosten (unveränderte Fitness) bei Fischanwesenheit wie dies für die omnivore merolimnische Insektenart H. instabilis beobachtet wurde. Die mittlere Fitness der Populationen (gemessen an der Eiproduktion pro Zeit- und Flächeneinheit) war für alle untersuchten invertebraten Räuber und Omnivoren bei Fischanwesenheit deutlich geringer als in der fischfreien Strecke. Für die Räuber (R. fasciata, P. conspersa) und G. pulex waren vermutlich letale Fischeffekte für den geringen Reproduktionserfolg verantwortlich. Dagegen wurde die Fitness der Population von H. instabilis durch die Kombination aus der geringeren individuellen Körpergröße (subletal) und den schwachen letalen Effekten auf die Emergenz reduziert. Der Einfluss der Fische auf die räuberischen Larven von R. fasciata und P. conspersa resultierte in einem Wechsel der Beutezusammensetzung, nicht in der Reduzierung der täglichen Fraßaktivität. Dies war vermutlich auf eine Aktivitätsänderung des Räubers R. fasciata bzw. auf die veränderte Beuteverfügbarkeit für P. conspersa unter Fischeinfluss zurückzuführen. Eine grundsätzliche Verringerung der Fraßaktivität oder der tierischen Anteile in der Nahrung der omnivoren Arten unter dem Einfluss benthivorer Fische wurde nicht festgestellt. Dagegen änderten sich die Relationen der Ressourcen Omnivorer zueinander, so konnte in der fischfreien Strecke eine flexiblere Ressourcennutzung nachgewiesen werden. Die jährliche Gesamtkonsumtion der räuberischen und omnivoren Invertebraten wurde durch die Anwesenheit der benthivoren Fische um 20-50 % reduziert, wofür hauptsächlich fischinduzierte Biomasseänderungen verantwortlich waren. In der fischfreien Strecke konnte dagegen die erhöhte Konsumtion der Invertebraten den fehlenden Fischfraßdruck mehr als kompensieren. Allerdings änderte sich die beutespezifische Konsumtion der Räuber in der fischfreien Strecke durch den verhaltensinduzierten Wechsel in der Beutezusammensetzung. Damit fand ein erhöhter Fraßdruck auf detritusfressende Arten (z.B. Diptera) statt, während die vermuteten Schlüsselarten einer potentiellen Biomanipulation (grazende Eintagsfliegen) in der fischfreien Strecke relativ unbeeinflusst blieben. Da die invertebrate Prädation in Fließgewässern sehr bedeutend sein kann, ist Biomanipulation nur unter Berücksichtigung aller Populationen räuberischer Invertebraten möglich. Nur durch Einstellung einer optimalen (relativ hohen) Biomasse geeigneter benthivorer Fischarten kann eine ausreichende Hemmung der invertebraten Räuber erreicht werden, so dass grazende Arten indirekt profitieren. / Biomanipulation (top-down control of the food webs) in lakes was intensively studied and can today be regarded as an important tool for improvement of the water quality and eutrophication. Similarly, the idea of biomanipulation can be used to reduce eutrophication phenomena in streams. The enhancement of benthic herbivorous key species (e.g. mayflies) by reduction of fish predation pressure could prevent exploding benthic algae biomass. Studies dealing with trophic cascades in lakes suggested that biomass of planktic invertebrate predators is increasing in the absence of planktivorous fish. But until now the importance of invertebrate predation pressure in fishless stream food webs is hardly known. Due to lethal and sublethal fish effects the consumption of predatory invertebrates could be enhanced by the absence of fish to such an extent that biomanipulation could be off target. In general, the proportion of invertebrate predators to the biomass of the benthic community in stream food webs is relatively high. Therefore, invertebrate predation on benthic prey can be as important as vertebrate predation (e.g. by benthivorous fish). Furthermore, omnivory occurs frequently in streams. Therefore, the aim of this thesis was to assess the importance of invertebrate predation in a food web of a small stream (Gauernitzbach) including the obligatory predators as well as the predatory potential of omnivores. Furthermore, the impact of benthivorous fish on population dynamics and feeding behaviour of the most important predators and omnivores was studied. Therefore, a reach stocked with benthivorous fish (Gobio gobio, Barbatula barbatula) was compared with an upstream fishless reach over a two-year period. After identification of important predators and omnivores based on their biomass and trophic position by stable isotope analyses their total consumption was estimated. Rhyacophila fasciata and Plectrocnemia conspersa (Trichoptera), Isoperla grammatica (Plecoptera) as well as Dugesia gonocephala (Turbellaria) were the invertebrate predators in the food web of Gauernitzbach. Because of their high biomasses the most important omnivores were Gammarus pulex (Amphipoda) and Hydropsyche spp. (Trichoptera). Using a mixing model (ISOSOURCE) that estimates the proportions of single resources on a mixed signature on the basis of the stable 13C and 15N isotopes, the predatory potential of both omnivores was pronounced by a high animal food proportion (20-90%). In this stream food web the predation impact of omnivores was comparable to this of obligatory predacous invertebrates. The total invertebrate predation pressure (up to 16 g TM m-2 a-1) in the fish reach was higher than the vertebrate predation pressure (~3 g TM m-2 a-1) mainly caused by the higher daily consumption rates and similar biomass compared to fish. Using stable isotope analyses of carbon and nitrogen the trophic pathways in the food web of Gauernitzbach could be identified. The most important resources for primary consumers in spring and autumn were allochthonous carbon sources. Autochtnonous carbon sources were only utilized by a few herbivores in spring. This was also reflected in the food of omnivores which assimilated mainly detritus beside animal material. Similarly, the predators depended distinctly on the allochthonous pathway by feeding primary consumers. Lethal fish effects on the predatory and omnivorous invertebrate populations were mainly depended on their preferred habitats. Pool-dwelling species (P. conspersa, D. gonocephala, G. pulex) showed decreased biomass in the presence of fish. Contrary, the biomass of riffle-dwelling invertebrates (R. fasciata, I. grammatica, Hydropsyche spp.) was not significantly reduced in the fish reach. Thereby, the highest biomass values in the fish reach could not reach the same level as in the fishless reach. Adults of the investigated species were influenced more lethal than the young stages. Hence, the emergence production of the predatory caddisflies and the abundance of the biggest and gravid females of G. pulex were reduced in the fish reach. Sublethal fish effects existed in dependence on the feeding groups of the invertebrates in Gauernitzbach. Whereas the adult predatory caddisflies R. fasciata and P. conspersa did not show disadvantages in fitness and size, the omnivorous caddisfly H. instabilis emerged with smaller size in the fish reach than in the fishless reach. Different responses of both omnivorous species were observed because of their different life cycles. The exclusively aquatic living G. pulex was not influenced sublethal by fish as the merolimnic insect species H. instabilis. The average fitness of population (measured as egg production per unit time and space) of the studied invertebrate predators and omnivores was distinctly reduced in the fish reach compared to the fishless reach. Probably, lethal fish effects were mainly responsible for the low reproductive success of the predatory species (R. fasciata, P. conspersa) and G. pulex. In contrast, the reduction of population fitness of H. instabilis was caused by the reduced individual adult size (sublethal effect) in combination with the weak lethal effect on emergence. The impact of fish on the feeding behaviour of the predatory larvae of R. fasciata and P. conspersa resulted in a switched prey composition, whereas a reduction in the daily feeding activity was not observed. That was supposed because of changes in the activity pattern by the predator R. fasciata or the changed prey availability for P. conspersa in the fish reach. A general reduction in feeding activity or animal food proportion of the omnivores was not assessed in the presence of fish. In contrast, the relations between the resources switched to a more variable utilisation in the fishless reach. The annual total consumption of invertebrate predators and omnivores was decreased by 20-50% in the fish reach, mainly caused by fish induced biomass reduction. The missing fish consumption could be at least balanced by higher invertebrate consumption in the fishless reach. Contrary, the prey specific consumption of invertebrate predators changed by behavioural-induced switch in the prey composition between the stream reaches. Therefore, a higher predation pressure on detritus feeding species (e.g Diptera) was found in the fishless reach, but not on the supposed key species of the potential biomanipulation (e.g. grazing mayflies). Due to the high importance of invertebrate predation in stream food webs, the aim of biomanipulation can only be reached by considering the populations of predatory invertebrates. Controlling the biomass of suitable benthivorous fish to an optimum (relatively high) level could sufficiently reduce the impact of invertebrate predators leading to an indirect enhancement of grazing species.

A good ecological status of streams and rivers is crucial for maintaining ecological functionality of running waters. Worldwide eutrophication threatens to change structure and function of freshwater ecosystems (Dodds et al., 2008). To reduce the symptoms of eutrophication in streams and rivers an additional approach, besides the reduction of external nutrient inputs from catchment areas, is needed. Therefore the goal has been set to transfer the approach of biomanipulation, which is widely accepted as tool in water quality management in lakes and reservoirs, to streams. The objective of this study was accordingly to analyse and evaluate some crucial preconditions for top-down control of stream food webs. For that purpose the present thesis examined effects of fish predation (stone loach and gudgeon) on grazer-periphyton interaction in small streams by assessing predator avoidance by benthic grazers, effects of benthic grazers on periphyton community composition during fish presence/absence and the possibility of top-down control on algal biomass by benthivorous fish.

A good ecological status of streams and rivers is crucial for maintaining ecological functionality of running waters. Worldwide eutrophication threatens to change structure and function of freshwater ecosystems (Dodds et al., 2008). To reduce the symptoms of eutrophication in streams and rivers an additional approach, besides the reduction of external nutrient inputs from catchment areas, is needed. Therefore the goal has been set to transfer the approach of biomanipulation, which is widely accepted as tool in water quality management in lakes and reservoirs, to streams. The objective of this study was accordingly to analyse and evaluate some crucial preconditions for top-down control of stream food webs. For that purpose the present thesis examined effects of fish predation (stone loach and gudgeon) on grazer-periphyton interaction in small streams by assessing predator avoidance by benthic grazers, effects of benthic grazers on periphyton community composition during fish presence/absence and the possibility of top-down control on algal biomass by benthivorous fish.