Zusammenfassung
In this thesis, I provide a description of the shape space of wild mouse mandibles with a focus on Mus musculus. Extending the comparisons to captive mice, inbred strains and some experimental populations, I try to infer which biological processes might account for observed patterns of shape variation, including genetic and developmental aspects (variability). I obtain the following results:
1) Mahalanobis distances based on CVA of Procrustes coordinates are a good measure of the global shape difference between two populations. Combined with the use of two-dimensional projections of μCT images of mouse hemimandibles, they are sufficiently robust in the face of diverse problems with sample quality and data processing, such as limitations in sample size, sampling errors with respect to sex, age and size of the animals, orientation of the samples inside the μCT scanner, preparation of bones and landmark digitization error.
2) Phenotypic plasticity as a reaction to environmental differences affects mandible shape by a smaller amount than the average distance between samples of wild-caught populations, suggesting that the shape differences between wild populations mostly have a genetic basis.
3) Various types of selection may have acted on shape. Four populations of mice from summer-dry regions cluster closely together, indicating that stabilizing selection may have conserved their shape. A M. m. domesticus sample from a site in Spain where the mice live in sympatry with a population of M. spretus is highly divergent from other M. musculus. This could represent a case of character displacement. Two populations of M. m. domesticus representing rather recent events of colonization on the subantarctic Kerguelen islands have diverged from other M. musculus in partially similar directions, which could represent an adaptation to the cold climate on these islands.
4) Inbred mouse strains are more divergent from wild mice and from each other than different species in nature, suggesting that nonadditive mechanisms of inheritance, especially epistasis, are important determinants of shape. This idea is supported by the finding that F1 of outcrosses between inbred strains look more similar to wild mice
than their parentals, i. e. their phenotype is not just intermediate, and there is some complementation of changes from the wildtype, but no complete reversal.
5) Wild-derived outbred populations kept in the laboratory diverge from wild mice over the course of a few generations, albeit less so than inbred mice. They are, however, not more divergent from each other than wild populations. This finding may point toward the existence of some epigenetically inherited mechanism of shape change which is somehow induced under laboratory conditions.
6) The Kerguelen mice, inbred strains, and wild-derived outbred populations (“derided populations”) do not diverge from wild mice in random directions. This pattern needs to be analyzed from a trait-based perspective. Geometric morphometrics alone is not suitable to dissect overall variation into individual traits. Simple alternative methods based on interlandmark distances (ILMDs) help to quantify the similarity between directions of shape change and to dissect shape changes with respect to the mandibular subregions involved.
7) Using a purpose-designed manual protocol, 20 groups of covarying ILMDs are identified, which can themselves be grouped into 5 “major traits”. These can largely be assumed to represent tradeoffs of tissue mass allocation during growth and to some degree functional coupling between parts of the mandible.
8) The 5 major traits explain large proportions of variation in several contexts: divergence of the derived populations from wild mice, variation within outbred and inbred populations (for the latter, i.e. developmental instability), “epistatic deviations” of outcross F1 from the interparental mean, and postnatal longitudinal ontogenetic shape change. This variety of contexts indicates that a large part of the genetically/developmentally generated variation is expressed via a limited number of types of shape changes. At the same time, the 5 traits are less important for the explanation of differences between populations and species.
9) Taking together the evidence for epistatic genetic architecture of shape and the results on the specific contexts in which the corresponding shape changes are observed, I hypothesize that epistatic shape variance may relate to developmental instability and provides the major part of phenotypic variance in wild populations. Stabilizing selection is unable to control this variation. Evolutionary divergence, however, happens predominantly along axes of additive variance, which provide a lower part of phenotypic variance within populations under this model, potentially due to the action of stabilizing selection.
