Lake Vättern has been contaminated with high levels of polychlorinated biphenyls (PCBs) for decades, which could be due to the release of wastes from industries and urban communities surrounding the water system. This has especially had a negative effect on fatty fishes, which could accumulate large amounts of persistent organic pollutants (POPs) and thereby also become a source of environmental toxicants to humans through consumption. Most PCB analysis only quantify a handful of congeners, the so called indicator-PCBs (I-PCBs), but this might leave out important information. In this study, an existing analytical method was improved by supplementing with additional congeners to detect a larger set of PCB congeners in Arctic char (Salvelinus salvelinus) and European whitefish (Coregonus acronius) caught from Lake Vättern, Sweden. New pre-packed multilayer silica columns from CAPE technologies were tested and used to pretreat the fish samples prior to analysis with a Gas Chromatograph coupled to low-resolution Mass Spectrometer using Atmospheric Pressure Ionization (API GC/MS). It was found that modifications of the clean up method for PCBs were necessary, such as lowering the amount of hexane in the washing step and combining the two eluent fractions. The Arctic char and the European whitefish showed a fat content of 0.18% and 0.74%, respectively. Concentrations of detected congeners ranged from 0.5 to 1470 pg g-1 fresh weight (fw) in Arctic char and varied between 1.2 to 6550 pg g-1 in European whitefish. For Arctic char and European whitefish, the WHO2005-TEQ values were 0.4 pg g-1 fw and 0.6 pg g-1 fw, respectively. The greatest total PCB concentration of 25900 pg g-1 was measured in European whitefish. The total concentration of I-PCBs (#28, 52, 101, 138, 153, 180) was 3710 pg g-1 for the Arctic char and 13900e pg g-1 for the European whitefish. All obtained results were lower than those reported from other studies. Constructed congener profiles show that the two species have similar ratios of PCB #138 and #153. Differences are observed of PCBs with a higher chlorination grade, probably due to differences in migration patterns, habitats of the lake, diets, metabolism or bioaccumulation.

European whitefish (Coregonus lavaretus) forms reproductively isolated sub-populations with different body sizes in the lakes where northern pike (Esox lucius) exists. It has been hypothesized that the initial reproductive isolation between ecotypes arises through size-assortative mate choice or because small and large individuals arrive to the spawning grounds at different times. To test these hypotheses, I caught and measured both breeding whitefish pairs and singular whitefish on the spawning grounds over the spawning season. There was no correlation between the lengths of males and females in breeding pairs, indicating that mate selection was random with respect to body size. There were no significant differences in body length between sampling dates during the field period in November 2021, suggesting that there is no divergence in spawning time between large and small individuals. Thus, neither of the two hypotheses were supported. Interestingly, males have significantly higher amount of breeding tubercles than females according to my results, which could be a selective factor in mate selection. However, the function of the tubercles is yet poorly understood.

An important question in determining species composition in a lake is what species has had the chance to reach the lake in the first place. The aim of this study was to examine natural stream connectivity limitations for whitefish (Coregonus lavaretus) between lakes, and to contribute to the development and evaluation of GIS-based methods to answer this question. 497 lakes with identified downstream source lakes, were classified as naturally colonized or introduced whitefish populations, reflecting the possibilities for whitefish to reach the lake naturally. Data of maximum stream slope and stream length between each lake pair was generated by Stefan Blumentrath, NINA. Maximum stream slope was measured on two different resolutions of stream length for comparison, one on slope over 10 m and one over 150m. A subset of lakes was manually examined in geographic information systems and compared to maps and aerial photos to evaluate the data and compare the results with the full dataset. False slope maximums were corrected and streams with much human alteration around a possible natural slope maximum was removed from the small dataset. The two datasets were analyzed using logistic regression models. Akaikes information criterion (AIC) showed that the optimal model, for both datasets, was the one using only slope maximum as predictor, and slope over 150m gave better results than 10m. A k50-coefficient, the value of a predictor that results in 50% probability of colonization, was introduced as an approximate of when the predictor forms a connectivity hinder. The k50-corefficient was estimated to 2,08±0,22° (±standard error) for slope over 150m for the big dataset and 2,58±0,20 for the small. Weaknesses in the data were distinguished and improvements for future fish connectivity studies are suggested.