The immune and the detoxication systems of animals are characterized by allelic polymorphisms, which underlie individual di¡erences in ability to combat assaults from pathogens and toxic compounds. Previous studies have shown that females may improve o¡spring survival by selecting mates on the basis of sexual ornaments and signals that honestly reveal health. In many cases the expression of these ornaments appears to be particularly sensitive to oxidative stress. Activated immune and detoxication systems often generate oxidative stress by an extensive production of reactive metabolites and free radicals. Given that tolerance or resistance to toxic compounds and pathogens can be inherited, female choice should promote the evolution of male ornaments that reliably reveal the status of the bearers' level of oxidative stress. Hence, oxidative stress may be one important agent linking the expression of sexual ornaments to genetic variation in ¢tness-related traits, thus promoting the evolution of female mate choice and male sexual ornamentation, a controversial issue in evolutionary biology ever since Darwin.

We approach the field of stress immunology from an ecological point of view and ask: why should a heavy physical workload, for example as a result of a high reproductive effort, compromise immune function? We argue that immunosuppression by neuroendocrine mechanisms, such as stress hormones, during heavy physical workload is adaptive, and consider two different ultimate explanations of such immunosuppression. First, several authors have suggested that the immune system is suppressed to reallocate resources to other metabolic demands. In our view, this hypothesis assumes that considerable amounts of energy or nutrients can be saved by suppressing the immune system; however, this assumption requires further investigation. Second, we suggest an alternative explanation based on the idea that the immune system is tightly regulated by neuroendocrine mechanisms to avoid hyperactivation and ensuing autoimmune responses. We hypothesize that the risk of autoimmune responses increases during heavy physical workload and that the immune system is suppressed to counteract this.

We have tested the importance of genetic variation in the major histocompatibility complex (MHC) class IIB in Atlantic salmon (Salmo salar) for survival after challenge with a highly virulent bacterial pathogen. Forty juvenile full siblings from each of 120 families were infected with the bacterium Aeromonas salmonicida, which causes high mortality in salmon due to furunculosis. Fishes from high-resistance (HR, < 35% mortality) and low-resistance (L,R, > 80% mortality) families were screened for their MHC class IIB genotypes using the denaturing gradient gel electrophoresis (DGGE) technique. The exon 2 sequences, encoding the major part of the peptide-binding region, were established for each DGGE fragment. One allele, e, containing a missense single base substitution was significantly more prevalent in HR families than in LR families. An odds-ratio test showed that broods carrying this allele had a 12-fold higher chance of being HR than broods without the e allele. A second allele, i, showed significantly higher frequencies in uninfected and surviving individuals than in infected dead individuals. A third allele, j, tended to more prevalent both in LR families and in individuals that had died of the infection. There was no correlation between MHC heterozygosity and resistance to A. salmonicida. Our results support the hypothesis that MHC polymorphism is maintained through pathogen-driven selection acting by means of frequency-dependent selection rather than heterozygous advantage.

The genes of the major histocompatibility complex (MHC) code for proteins involved in antigen recognition and triggering of the adaptive immune response, and are therefore likely to be under selection from parasites. These selection regimes may vary in space and time. Here we report a strong geographical structure in MHC class II B genes of a migrating bird, the great snipe (Gallinago media). Genetic differentiation in the MHC between two ecologically distinct distributional regions (Scandinavian mountain populations vs. East European lowland populations) was still present after statistically controlling for the effect of selectively neutral variation (microsatellites) using partial Mantel tests. This suggests a role for selection in generating this spatial structure and that it represents local adaptation to different environments. Differentiation between populations within the two regions was negligible. Overall, we found a high number of MHC alleles (50, from 175 individuals). This, together with a tendency for a higher rate of nonsynonymous than synonymous substitutions in the peptide binding sites, and high Tajima's D in certain regions of the gene, suggests a history of balancing selection. MHC variation is often thought to be maintained by some form of balancing selection, but the nature of this selection remains unclear. Our results support the hypothesis that spatial variation in selection regimes contributes to the high polymorphism.