Maternal effects occur when maternal environment or phenotype influence offspring phenotype, in addition to genetic contribution of the mother. As maternal effects often influence phenotypes that are under natural selection, they hence have evolutionary consequences. Further, the expression of both maternal effects and evolutionary potential has been argued to depend on environmental conditions, but the evidence of this dependency for the process of adaptation has been inconclusive. The main objective of this thesis was to investigate evolutionary consequences of maternal effects and stressful or variable environmental conditions.

I started by performing a meta-analysis of quantitative genetic studies that investigated expression of additive genetic, maternal, and residual variance under both stressful and benign environmental conditions (Paper I). Data spanning over many animal taxa and stress types revealed that high levels of environmental stress correlated with increased expression of genetic and residual variances. However, against our predictions, maternal effects were relatively unaffected by stress.

In Paper II and III, I explored the evolutionary divergences of traits previously shown to be under maternal control. Specifically, in Paper II, I performed a second meta-analysis, that investigated if parents of common frogs (Rana temporaria) influenced offspring development time to mediate the effects of time constraints, across a latitudinal cline. I found that reproductive delay in the parental generation correlated with decreased development time in tadpoles of northern R. temporaria populations, suggesting that parental effects may further decrease development time in populations from time-constrained environments.

In Paper III, I used an annual killifish system, to explore if environmental unpredictability, measured by variation in precipitation during rainy season, correlated with maternally mediated variation in embryo development time (bet-hedging). Although I found significant among-species differences in variation in development time, there was no clear linear relationship between variation in development time and precipitation. The results suggest that either bet-hedging is not important for persistence in the unpredictable annual killifish habitats, or that other ecological factors, rather than precipitation unpredictability, influenced evolution of variation in development times.

Lastly, I investigated if occurrence of placenta correlated with increased offspring brain size among poeciliid fish (Paper IV). In contrast to our prediction, I did not find any consistent differences in relative brain size between the fry of placental and non-placental species. It is possible that either the poeciliid placental structures do not have a sufficient capacity to transfer resources necessary for increased brain development, or that other factors, such as sexual selection, or differences in food abundance and competition, shaped brain evolution among poeciliids.

In conclusion, the results of this thesis suggest that environmental stress may influence evolutionary potential by increasing genetic variation available for selection, that time-constrained habitats may be conducive to evolution of parental effects on offspring development times, and that maternal influence on offspring traits may be difficult to detect, as many ecological factors may potentially influence evolution of life-history and morphology traits.