The presence of genetic variation is a key ingredient for rapid evolution in changing conditions, but we cannot explain how genetic variation is maintained or enhanced. The role of spatial heterogeneity with limited exchange between sub-populations is potentially a potent mechanism for maintaining genetic variation, but so far most simple models come to the conclusion that the range of conditions in which genetic variation can be maintained is very limited.
Models do show that heterogeneity may shift the mutation-selection equilibrium to a higher value and thus help to maintain genetic variation. The data from the long term studies on hole-breeders are ideal material to uncover and resolve a number basic methodological problems.
On one hand, there is the disentangling of phenotypic plasticity and selection in data from heterogeneous environments and on the other hand the study of population structure and dispersal is hampered by strong effects of observer distribution. In particular the data from the island populations of great and blue tits on Vlieland offer unique possibilities for analysis, because the pedgrees of these natural populations are very well known. Two different methods have been developed to filter raw data on dispersal (the net displacement between site of birth and site of reproduction). Both methods show that the average distance moved is much greater than raw data suggest.
Thus, our understanding of population structure will have to be revised. Interestingly, this makes it even more important (and plausible) that birds somehow learn about their breeding environment in that they adapt their reproductive strategy depending on the outcome of their first breeding attempts (Grieco et al 2002). This makes sense because the displacements after the onset of reproduction are very small. The importance of population structure for the explanation of spatial patterns in clutch size is nicely illustrated by great tit data from Vlieland, where different proportions of immigrants are responsible for a substantial difference between the two sub-populations that has persisted for decades.
A new problem which is currently explored is the interaction between variation in selection intensity and variation in cohort size. Whereas this variation is absent in agricultural settings where the methods were developed, this variation is often big in natural populations. The covariance between these two can easily create a severalfold difference in cumulative response to selection.
Current group-members
