Objectives:: Individuals with fetal growth restrictions may become unhealthy as adult if they live in a nutritionally rich environment. Yet, there is little evidence for the counterpart that profuse resources during early development may prove detrimental in an impoverished adult environment. Using birth season as a proxy for nutrition in utero, this study tests whether early life effects on longevity are conditional on the environment in which people reside during adulthood. Methods:: I used Cox proportional hazard models of mortality after the age of 60 years among historical cohorts of French Canadian women born before 1750. Models are stratified by siblings to adjust the results for unobserved factors shared by family members. Results:: A birth during winter conferred the best‐survival prospects in the south of the Saint‐Lawrence River but the worse prospects in the north, where a birth during the fall was associated with the lowest mortality. Women who migrated to live on the other side of the river lost whatever advantage or disadvantage they had from their birth season and faced increases or decreases in risk that were specific to their new location. Conclusions:: Based on Hales and Barker's thrifty phenotype hypothesis, and adding the “hopeful phenotype” counterpart, this study suggests that cues from the external environment during development are highly specific to that environment and that the slightest change of location may affect the chances for survival into old age. I finally address the role of fertility as a possible mediator or modifier of the effect of birth season on longevity. Am. J. Hum. Biol., 2012. © 2012Wiley Periodicals, Inc.