Analysing variation of skull shape in wild populations can give us an insight into evolutionary processes. By looking into patterns of morphological variation within populations, we can extrapolate and make assumptions on the patterns of variation on higher taxonomic levels. In this thesis, I collected data on skull shape of a wild population of Soay sheep using geometric morphometrics methods. I applied a multitude of quantitative genetics and geometric morphometrics methods to explore the factors behind the evolution of skull shape. I analysed several aspects of skull shape, including integration and modularity, predicted response to hypothetical selection, estimates of natural selection, presence of heterochrony, and fluctuating asymmetry (FA) in an ecological and genetic point of view. I found low levels of integration in the Soay sheep skull, along with the presence of two modules: face and neurocranium. Both the levels of integration and the pattern of modularity were found across different origins of variation, individual, genetic and developmental (FA), indicating a strong correspondence between levels. Although genetic integration is not strong, I found that most of hypothetical selection regimes result in as much correlated selection as direct selection responses. But the nose region has shown to be quite independent from the rest of the skull, with a retraction of the nose causing almost no correlated changes. Selection on skull shape is comparatively strong in the Soay sheep. The skull shape selected for in males differs from the shape selected in females. Males are selected to have wider nasal bones along with longer, flatter braincase, whereas females are selected towards narrower nasal bones, along with smaller braincases. But selection is not the only factor influencing changes in skull shape. Changes in temperature along the past 30 years are causing a change in developmental timing in Soay sheep, causing a decrease in overall body size in the population. And I found that skull shape is also being affected, thus indicating that this population might be undergoing heterochronic processes. Finally, I found that factors such as vegetation quality, parasite load and breeding success are related to FA levels, whereas population density and climate are not. I also found no significant heritability nor dominance in FA of skull shape. Overall, this thesis certainly contributes to the understanding of evolution of skull shape in natural populations. By successfully combining quantitative genetics and geometric morphometrics methods, I was able to tackle questions about the evolution of complex shapes, and give valuable insights into this still underexplored field.