Distinctive epidemiological features of the sudden infant death syndrome (SIDS) are its age at death distribution and pronounced winter excess. Whether or not these effects are independent of the month of birth of the infant is uncertain. We used a loglinear model to separate the effects of age at death, month of death and month of birth amongst 6229 infants who died from SIDS in England and Wales during the period 1979-1983. The results suggest that month of birth and month of death independently influence the risk of the infant dying from SIDS, the risk related to month of death being much greater.