Embryonic life histories in annual killifish: adapted to what?

Adaptation requires an evolving strategy and an environment. Starting from environmental patterns, we predict which strategies are adapted. Given a strategy, we want to know in which environments it might be adapted. Annual killifish embryos can arrest development, survive desiccation of temporary ponds in the soil and hatch when they are rewetted. They might implement diversified bet-hedging. However, an association between magnitudes of environmental and developmental variability across populations or species has not yet been found. Their environments also have a strongly seasonal character and small-scale variation between pond centre and edge. Using data on embryonic life histories of Austrolebias annual killifish, parameter estimates and parsimonious assumptions, a population dynamical model is constructed with explicit developmental stages. For different simulated seasonal pond filling regimes with gradual filling and drying, it is used to estimate survival in the egg bank across a year and recruitment. Survival in the egg bank is small and variable, contrary to a standard assumption of most seed bank models. Calculations aiming to determine seasonal regimes where embryonic life histories could be adapted are presented. Invasion fitness gradients of rates of development and hatching probabilities are used to search for evolutionarily singular environments ESE, where none of the traits experience directional selection. Among the seasonal annual cycles investigated, no ESE occurred. Faster development rates were always favoured. For hatching probabilities, seasonal regimes were found which made their invasion fitness sensitivities zero. However, the observed trait values did not have long-term evolutionary stability in these regimes. It is argued that tests for adaptation to uncertain environments in annual fish should focus on associations between variability in pond filling and hatching probabilities.