Population regulation in Gyrodactylus salaris - Atlantic salmon (Salmo salar L.) interactions: testing the paradigm.

Background: Gyrodactylus salaris is a directly transmitted ectoparasite that reproduces in situ on its fish host. Wild Norwegian (East Atlantic) salmon stocks are thought to be especially susceptible to the parasite due to lack of co-adaptation, contrary to Baltic salmon stocks. This study i) identifies whether time- and density-dependent mechanisms in gyrodactylid population growth exist in G. salaris-Atlantic salmon interactions and ii) based on differences between Norwegian and Baltic stocks, determines whether the 'Atlantic susceptible, Baltic resistant' paradigm holds as an example of local adaptation.

Methods: A total of 18 datasets of G. salaris population growth on individually isolated Atlantic salmon (12 different stocks) infected with three parasite strains were re-analysed using a Bayesian approach. Datasets included over 2000 observations of 388 individual fish.

Results: The best fitting model of population growth was time-limited; parasite population growth rate declined consistently from the beginning of infection. We found no evidence of exponential population growth in any dataset. In some stocks, a density dependence in the size of the initial inoculum limited the maximum rate of parasite population growth. There is no evidence to support the hypothesis that all Norwegian and Scottish Atlantic salmon stocks are equally susceptible to G. salaris, while Baltic stocks control and limit infections due to co-evolution. Northern and Western Norwegian as well as the Scottish Shin stocks, support higher initial parasite population growth rates than Baltic, South-eastern Norwegian, or the Scottish Conon stocks, and several Norwegian stocks tested (Akerselva, Altaelva, Lierelva, Numedalslågen), and the Scottish stocks (i.e. Conon, Shin), were able to limit infections after 40-50 days. No significant differences in performance of the three parasite strains (Batnfjordselva, Figga, and Lierelva), or the two parasite mitochondrial haplotypes (A and F) were observed.

Conclusions: Our study shows a spectrum of growth rates, with some fish of the South-eastern Norwegian stocks sustaining parasite population growth rates overlapping those seen on Baltic Neva and Indalsälv stocks. This observation is inconsistent with the 'Baltic-resistant, Atlantic-susceptible' hypothesis, but suggests heterogeneity, perhaps linked to other host resistance genes driven by selection for local disease syndromes.