Adaptive and evolutionary significance of a reproductive thermal threshold in Barbus barbus

From 1989 to 1996, barbel in the River Ourthe started spawning under variable environmental conditions, except for water temperature. Each year, spawning was initiated when water daily minimum temperature reached or exceeded 13·5° C. Any decrease of temperature below this value later in the spawning period caused spawning to be suspended. Analyses of offspring growth provided evidence that 13·5° C was the value below which 0+ barbel stop growing. It was hypothesized that barbel trade off the lower initial probability of survival against a larger size at the onset of winter. To test empirically for this hypothesis, the adequacy of alternative—theoretical—strategies associated with other thermal thresholds (12, 15·0, 17·1 and 20·2° C) was modelled with respect to: (1) the feasibility of spawning (inhibition of sexual maturation by a decreasing photoperiod); (2) the impact of the temperature on embryonic development; (3) the effect of water level variations on the integrity of spawning grounds until the emergence of larvae; (4) the size of the offspring at the onset of winter. On an 8‐year (1989–1996) average, the present spawning strategy would have produced a higher recruitment than alternative strategies (relative adequacy of 33·23, 85·64, 93·17 and 17·62%, respectively). However, alternative strategies would have produced better annual scores on five of eight occasions in the River Ourthe environment, and a better overall score in environments 1·5 or 3·5° C warmer than now. The consistency of the thermal threshold over years, despite a low selection pressure by the environment, was interpreted as the expression of a phenotypic mechanism (thermal homing) promoting the selection of the lowest efficient thermal threshold, and enabling breeders to relay to the next generation some form of thermal stability in a variable environment.