How predictable is rapid evolution?

Abstract

Although evolution is historically considered a slow, gradual process, it is now clear that evolution can occur rapidly over generational timescales. It remains unclear both how predictable rapid evolution is and what timescales are ecologically relevant due to a paucity of longitudinal studies. We use a common garden approach to measure genetic-based change in complex, fitness-associated traits that are important for climatic adaptation in wild Drosophila over multiple timescales: an estimated 1-16 generations within each year and 48-89 generations over five consecutive years. Evolution is fast and pervasive with parallel patterns of rapid evolution in three distinct locations that span 4{o} latitude. Developmental time evolves consistently across seasons with flies collected in spring developing faster than those collected in autumn. The evolutionary trajectory of stress traits (heat knockdown and starvation) depends on the severity of the preceding winter: harsh winters result in a predictable evolutionary trajectory with high stress tolerance in spring flies that declines in the subsequent generations across the summer. Flies collected after mild winters do not evolve in a predictable pattern but may utilize an alternative strategy as plasticity for chill coma recovery and starvation evolves across seasons. Overall, winter severity determines the predictability of rapid seasonal evolution, but there are also long-term shifts in the phenotypic correlations and allele frequencies that indicate long-term population changes that have broader implications for how organisms respond to the changing climate. Significance StatementAdaptive tracking may result in rapid evolution over short timescales, but the repeatability and predictability of rapid adaptation is less well resolved without long-term, multi-year analyses. Here, we collect wild flies at regular intervals across five years to determine what traits evolve consistently over seasons and which environmental variables predict this rapid evolution. Traditional temperate seasonal patterns of harsh winters are crucial for normal selection patterns, although independently changing phenotypic and genetic correlations help the populations respond to long-term shifts over years, particularly in response to heat stress. This has the implication that populations may be flexible within certain genetic constraints to adapt to changing climatic temperatures.