To examine the implications of seasonality for the construction of a single-year calibration curve we obtained separate dates on earlywood and latewood fractions of tree rings originating from England and dendrochronologically dated between AD 1352 and AD 1442. These demonstrated that an average difference of 26±15 yr exists between earlywood and latewood and that this difference can be as high as 33±19 yr during periods of high radiocarbon (14C) production. It is argued that this difference is due to both changes in atmospheric 14C and the incorporation of stored carbohydrates into earlywood. Based on this, it was possible to separate an atmospheric and physiological contribution to this difference. Our modeling indicates that storage can produce a difference of up to 10 years between earlywood and latewood. This suggests that full-year tree rings from deciduous trees may be less appropriate for the construction of a single-year calibration curve and that specific atmospheric events can be more easily detected by measuring only latewood.