There is a problematic interaction going on between grey seals and the small scale coastal fisheries in the Baltic. A large number of seals are by-caught and drowned each year, and the viability of the fishery is threatened by catch losses caused by the seals. Traditional mitigation methods are not sufficient, or have in some cases not been properly evaluated. Available methods of quantifying and analysing the catch losses are also insufficient. This thesis consists of three parts, each studying a different angle of this conflict. In the first part, new models for estimating catch losses are presented. In addition to the commonly used method of counting the number of damaged fish in the nets, the new models also allow for an estimation of the hidden losses. Hidden losses may be fish that are completely removed from nets without leaving any traces, fish that escape through holes in the net torn by the seals, or even fish that are scared away from the fishing gear. Such losses were found to be significant, and hence it is now clear that the traditional models seriously underestimate the total losses. The new models also allow for a deeper analysis of the interaction process. The first presented model compares catches between adjacent days (day-pairs), the second uses nets that are pre-baited before deployment, and the third relies on a detailed inspection and repair of all seal-induced damage to the net meshes. In the second part, some traditional methods of mitigating the conflict are evaluated. A commercially available Acoustic Harassment Device was tested in a field trial. AHDs were deployed at several set-traps for salmonids for three consecutive years. The damage reducing effect was persistent throughout a season, as well as over the full three-year test period, and no “dinner bell” effect was observed. When seal attacks became frequent in the 1980´s, several of the traditional salmon traps were reinforced with newly developed extra strong net materials. These materials dramatically reduced the damage to the nets, and to some degree also the catch losses. However, the losses were still substantial, and the traditional gear was gradually phased out when better solutions emerged. In the third part, new methods of mitigating the conflict are evaluated. A salmon trap was built, using net meshes which were large enough to allow seal-chased fish to escape through, but which would still guide and confine non-stressed fish. The trap was fitted with a fish chamber with a double wall of very taut netting, separating the catch from the surroundings by a fixed distance. Interference by seals was significantly reduced with this construction. Field experiments revealed that seals used their above-water vision to locate and search out buoys of the type that are used in the fisheries. Larger buoys were more readily found than smaller. A set of trials was initiated where certain geographical areas were made unattractive for seals prior to their seasonal arrival to the region, by deploying stationary AHDs. Finally, aquarium experiments demonstrated that underwater vision and hearing were equally important in seals’ detection of fish in a test box. It was also found that there was a “near zone”, within which seals stayed focused on a fish and attempted to catch it by a quick thrust of the head. These studies strongly suggest that new seal-safe fishing gear and mitigation methods should be based on, and would benefit from, an in-depth understanding and analysis of natural seal behaviour.