Laboratory studies on embryos of salmonids, such as the brown trout (Salmo trutta), have been extensively used to study environmental stress and how responses vary within and between natural populations. These studies are based on the implicit assumption that early life-history traits are relevant for stress tolerance in the wild. Here we test this assumption by combining two datasets from studies on the same 60 families. These families had been experimentally produced from wild breeders to determine, in separate samples, (i) stress tolerances of singly kept embryos in the laboratory and (ii) growth of juveniles during 6 months in the wild. We found that growth in the wild was well predicted by the larval size of their full sibs in the laboratory, especially if these siblings had been experimentally exposed to a pathogen. Exposure to the pathogen had not caused elevated mortality among the embryos but induced early hatching. The strength of this stress-induced change of life history was a significant predictor of juvenile growth in the wild: the stronger the response in the laboratory, the slower the growth in the wild. We conclude that embryo performance in controlled environments can be useful predictors of juvenile performance in the wild.