1. In blue tit (Parus caeruleus L.) populations the numbers of surviving young from a clutch generally declines the later in the season it hatches, but hatching too early also reduces productivity. 2. The extent to which seasonal changes in offspring survival were caused by environmental changes was assessed experimentally by cross-fostering clutches of differing age between pairs of nests in Bagley Wood, Oxfordshire. This manipulation had the effect of advancing or delaying the hatch date experienced by parents. Productivity of experimental nests could then be compared with the productivity of control pairs breeding at the same time, and with seasonal trends in long-term data from a blue tit population nesting in a similar habitat at Wytham, Oxfordshire. 3. The number of young which survived from experimental clutches was significantly predicted by the hatch date of the clutch from which the young originated (adopted clutch), rather than by the hatch date of the clutch actually produced by their foster parents (fostered clutch). This suggests a causal link between hatch date and offspring survival. The predictive model generated using experimental nests was in agreement with the model based on control nests and the model based on the long-term data from Wytham. 4. Mechanisms underlying the relationship between hatch date and offspring survival are discussed, as well as the implications for the evolution of breeding date in blue tits.