Experiments using natural populations have provided mixed support for thermal adaptation models, probably because the conditions are often confounded with additional environmental factors like seasonality. The contrasting geothermal environments within Lake Mývatn, northern Iceland, provide a unique opportunity to evaluate thermal adaptation models using closely located natural populations. We conducted laboratory common garden and field reciprocal transplant experiments to investigate how thermal origin influences the life history of Radix balthica snails originating from stable cold (6 °C), stable warm (23 °C) thermal environments or from areas with seasonal temperature variation. Supporting thermal optimality models, warm origin snails survived poorly at 6 °C in the common garden experiment and better than cold origin and seasonal snails in the warm habitat in the reciprocal transplant experiment. Contrary to thermal adaptation models, growth rate in both experiments was highest in the warm populations irrespective of temperature, indicating cogradient variation. The optimal temperatures for growth and reproduction were similar irrespective of origin, but cold origin snails always had the lowest performance, and seasonal origin snails often performed at an intermediate level compared to snails originating in either stable environment. Our results indicate that central life-history traits can differ in their mode of evolution, with survival following the predictions of thermal optimality models, whereas ecological constraints have shaped the evolution of growth rates in local populations.