Antarctic krill, Euphausia superba, and the salp Salpa thompsoni are the most abundant macrozooplankton in the Southern Ocean and thus among the most important grazers of phytoplankton. Both species play an important role in biogeochemical cycles by the production of large, carbon-rich and fast sinking faecal pellets, but the extent and magnitude of their contribution to the flux of particulate organic carbon (POC) remain uncertain. Krill and salps occupy very different ecological niches and differ markedly in the way they fuel the lower and higher food webs. Krill is a key species in the Southern Ocean ecosystem connecting phytoplankton and apex predators, including whales, seals, and sea birds. Salps are a less important food source for apex predators, but can feed on a wider food size-range than krill and control phytoplankton blooms. Over the past decades, krill and salps have shown dramatic changes in their distribution and abundance in the Southern Ocean. In response to warming temperatures and declining sea ice, salps have moved southward and expanded into previously krill dominated areas. In some regions along the Western Antarctic Peninsula (WAP), salps have replaced krill as the dominant grazer species. A long-term shift in the dominance from krill to salps is expected to have cascading effects on the food web with unknown consequences for biogeochemical cycles. However, studies directly comparing the role of krill and salps were so far lacking. This dissertation provides a first-time direct comparison of the role of krill and salps in the carbon cycle at the Antarctic Peninsula by comparing the diet and faecal pellet composition of krill and salps, as well as the contribution of their faecal pellets to the total flux of POC. To shed light on the diet and faecal pellets composition of krill and salps, their stomach contents were compared to the ambient plankton community and to the composition of the faecal pellets produced by either species using 18S metabarcoding. Krill are suggested to be selective filter feeders, while salps have been previously assumed to feed indiscriminately on a wide size range of particles. The direct comparison of the stomach content to the in situ plankton community showed that neither species’ diet reflected the composition of the plankton community, suggesting that in contrast to the accepted paradigm not only krill, but also salps are selective feeders. In addition, the comparison of the stomach contents showed that krill and salps had similar diets, which could enhance the competition between both species. This competition could be augmented by salps’ ability to rapidly reproduce in favourable conditions, posing further pressure on krill populations, which are targeted by commercial fishery and are already affected by climatic changes and are more susceptible to warming temperatures than salps. In contrast to the stomach contents, the composition of the faecal pellets produced by krill and salps showed significant differences. Krill pellets consisted for the most part of flagellate remains, while salp pellets were composed to equal shares of diatom and flagellate remains, implying that salp faecal pellets sink at higher velocities than krill pellets due to the ballasting effect of the diatom derived opal. In addition, the higher proportion of diatoms in the faecal pellets of salps compared to their stomach content suggests a low assimilation efficiency of diatoms. In addition, in this thesis several methodological aspects of diet composition studies were assessed. Molecular analyses such as 18S metabarcoding require a rapid inactivation of digestive enzymes and DNA nucleases and therefore depend on suitable sample processing and preservation methods. To investigate this, the effect of different processing procedures on the prey community of the stomachs of krill and salps was compared, showing that in species with highly active and efficient digestive enzymes, like krill, the detection of rapidly digested prey organisms, e.g. gelatinous plankton, is negatively affected by a prolonged sample processing time. In addition, different variable regions of the 18S rRNA gene were compared to assess differences between the two marker regions. This comparison of the 18S regions V4 and V7 showed that the composition of prey communities retrieved from either marker region differ significantly, suggesting that different regions of the 18S rRNA gene might not be suitable for a direct comparison within or between studies. Finally, the signals of dietary fatty acid markers from stomach and tissue samples of krill and salps were compared. Fatty acid markers differed significantly between stomach and tissue samples, indicating that for the differentiation of short- and long-term signals and to detect differences in the assimilation efficiency, it is critical to separate digestive tract and tissue for the analysis of fatty acids. To further investigate the role of krill and salps in the biological carbon pump at the Antarctic Peninsula, a direct comparison of the contribution of krill and salp faecal pellets to the total flux of carbon was conducted. Consecutively deployed sediment traps were combined with vertical profiles of an in situ particle camera system at a high temporal resolution. These in situ measurements were referenced with on-board experiments of the production, carbon content, and microbial degradation of krill and salp faecal pellets, and estimates of the in situ biomass of krill and salps. On average, the potential faecal pellet production of krill was lower than that of salps, while both species contributed equally to the carbon flux at 200 and 300 m. Thus, krill pellets were exported at higher efficiencies, while the majority of salp faecal pellets were retained in the upper 200 m of the water column. However, the presence of large, dense krill swarms, which were observed near some of the deployed sediment traps, increased the flux of krill faecal pellets by almost two orders of magnitude. These results support the notion that krill swarms can produce ‘rains’ of faecal pellets that exceed the capacities of detritus feeders and therefore lead to disproportionately high fluxes of krill faecal pellets and low attenuation rates. Microbial degradation rates of salp pellets accounted for only about one percent per day and could not explain the high turnover rate of salp pellets. Thus, fragmentation by zooplankton seems to be a key process controlling the flux of salp pellets. Due to the fragile structure of salp faecal pellets, the organic matter therein might be more easily accessible for scavengers and suspension feeders compared to krill pellets. Increased recycling of salp pellets might therefore lead to an additional release of nutrients in the surface ocean, benefiting zooplankton and microorganisms. This dissertation provides new insights into the role of krill and salps in the carbon cycle at the Antarctic Peninsula and expands the knowledge on the potential consequences of a long-term shift from krill to salps in the Southern Ocean. Opposite to the thus far accepted paradigm, salp faecal pellets seem to be less efficient in exporting carbon out of the euphotic zone to the deep ocean, but are mostly retained in the surface layer. This could lead to a less efficient biological carbon pump, but fuel the food web in the mixed layer, including microorganisms and zooplankton. Thus, on the long-term less carbon could be exported and sequestered in the Southern Ocean, which would be augmented by the increasing salp abundances and the poleward shift of krill. In addition, similar diets of krill and salps could lead to competitive removal of prey, which might be a critical factor driving krill and salps apart on spatial and temporal scales.