A model-based assessment of adaptation in embryonic life histories of annual killifish

Abstract Annual killifish embryos can arrest development in different diapauses and survive desiccation of temporary ponds in soil egg banks. They seem to implement diversified bet-hedging, generally seen as an adaptation to uncertain environments. Using a pragmatic eco-evolutionary model tailored to annual fish, it is investigated whether rates of embryonic survival and development and hatching could be adapted to pond filling regimes with within-year variation only. In the model, ponds fill and dry gradually. Locations within a pond where embryos are buried can therefore differ in seasonal dryness patterns. Results from lab experiments on the Argentinian pearlfish Austrolebias bellottii are used to model embryonic life histories and hatching probabilities. Maturation, survival and reproduction of fish are modelled in agreement with lab experiments and field data on the same or related species. Invasion fitness sensitivities of four stage-specific rates of development, five mortality rates and two hatching probabilities were estimated. Faster development rates and lower mortalities were always favoured. Only trade-offs or constraints could keep them at their current values in the simulated environments. For hatching probabilities, pond filling regimes were found which made their invasion fitness sensitivities zero or, for hatching at the first rewetting, with different signs in different pond filling regimes. Between-year variability might render development rates adapted if faster rates would increase population size fluctuations which would be selected against. For hatching at first rewetting, an alternation between years with positive and negative selection will contribute to their adaptedness.