| dc.description.abstract | Human activity has caused an increase in the fragmentation of natural habitats, especially during the last couple of centuries. This means that populations of wild species are split into smaller sub-populations often connected with each other only through migration (i.e. they become metapopulations (Hanski and Gilpin 1997)). Such fragmentation is likely to affect ecological and evolutionary processes within the populations, and ultimately affect the expected time to extinction. This indicates the importance of studies that estimate parameters that affect the spatial processes within such fragmented populations.
This thesis focuses on demographic parameters within an insular metapopulation of house sparrows (Passer domesticus) at the coast of northern Norway. We documented that weather conditions during the breeding season had a major effect on the fledging and recruitment success in the metapopulation (Paper I). Although the weather conditions were highly correlated on different islands, there was considerable spatial variation among islands in both fledging and recruitment success because the onset of breeding varied among islands. This shows a mechanism that can reduce spatial synchrony in dynamics of sub-populations within a metapopulation, even in strongly spatially autocorrelated environments.
In accordance with the large spatiotemporal effect of weather on recruitment success, we also recorded considerable spatiotemporal variation in directional natural selection on fledgling morphology within the metapopulation (Paper II). Despite the considerable effect of weather on survival, the weather barely affected the strength of selection. Thus, variation in weather does not result in predictable microevolutionary changes in this metapopulation. Similarly, we did not detect any effect of either population growth rate or density, on survival rate or strength of selection.
For evolutionary change to occur in a trait, at least some of the phenotypic variation observed in the trait has to be heritable. Six morphological traits in adult house sparrows were shown to have heritability between 0.18 and 0.68 (Paper III). Any selection acting on these traits is thus expected to result in evolutionary change. Furthermore, we documented both positive and negative genetic correlations among traits, suggesting that selection on some of these traits may be either enhanced or constrained by selection on another. There was a sexual difference in the amount of heritable genetic variation in some of the traits; generally the heritability was higher in females than in males. Accordingly, the genetic correlation matrix also differed between the sexes. Consequently, selection on males and females may result in different micro-evolutionary responses in this metapopulation.
The average lifetime reproductive success (LRS) of individuals of a certain phenotype may be a good indicator of its fitness. There was large variation in LRS among different individuals (and thus phenotypes) in this metapopulation, but no sex-specific differences in the distribution of LRS (Paper IV). Accordingly, LRS was mainly influenced by variation in annual reproductive success, and to a lesser extent variation in lifespan in both sexes. Some of the traits that were found to be heritable in adult house sparrows (Paper III) affected their lifetime reproductive success (LRS). Phenotypic characteristics explained a significant proportion of the variation in male LRS, but not in females. Bill length and badge length of males was positively related to lifespan and annual reproductive success, respectively. The effect of male morphology was through an effect on the number of female recruits, and not sons. Individual variation in morphology affects the expected fitness, and may thus influence the evolutionary processes in the population.
The large individual variation in LRS (Paper IV) and the considerable spatial variation in recruitment rate among islands (Paper I and II) indicated the presence of large spatial variation in annual reproductive success among individuals as well. Accordingly, we found large spatial variation in effective population size (Ne) (Paper V), showing large spatial variation in the rate by which genetic variation was lost through random genetic drift among islands. The ratio of Ne to census population size varied between 0.19 and 0.59 on five islands in the metapopulation, showing that Ne is usually less than 0.5 the census population size. A higher proportion of the variance in Ne/N was explained by variation in the recruitment success in females than in males, where fluctuations in population size were most important. Because we have demonstrated a large effect of stochastic variation in weather conditions on recruitment rate (Paper I), our results suggest that environmental stochasticity is a major source of loss of genetic variation in this metapopulation. | nb_NO |
