Abstract
BACKGROUND: Most evolutionary studies on the size of brains and different parts of the brain have relied on interspecific comparisons, and have uncovered correlations between brain architecture and various ecological, behavioural and life-history traits. Yet, similar intraspecific studies are rare, despite the fact that they could better determine how selection and phenotypic plasticity influence brain architecture. We investigated the variation in brain size and structure in wild-caught nine-spined sticklebacks (Pungitius pungitius) from eight populations, representing marine, lake, and pond habitats, and compared them to data from a previous common garden study from a smaller number of populations. RESULTS: Brain size scaled hypo-allometrically with body size, irrespective of population origin, with a common slope of 0.5. Both absolute and relative brain size, as well as relative telencephalon, optic tectum and cerebellum size, differed significantly among the populations. Further, absolute and relative brain sizes were larger in pond than in marine populations, while the telencephalon tended to be larger in marine than in pond populations. These findings are partly incongruent with previous common garden results. A direct comparison between wild and common garden fish from the same populations revealed a habitat-specific effect: pond fish had relatively smaller brains in a controlled environment than in the wild, while marine fish were similar. All brain parts were smaller in the laboratory than in the wild, irrespective of population origin. CONCLUSION: Our results indicate that variation among populations is large, both in terms of brain size and in the size of separate brain parts in wild nine-spined sticklebacks. However, the incongruence between the wild and common garden patterns suggests that much of the population variation found in the wild may be attributable to environmentally induced phenotypic plasticity. Given that the brain is among the most plastic organs in general, the results emphasize the view that common garden data are required to draw firm evolutionary conclusions from patterns of brain size variability in the wild.