Aims: To investigate infant deaths in Cumbria, 1950-93, in relation to individual and community level socioeconomic status. Methods: Retrospective birth cohort study of all 283 668 live births and 4889 infant deaths in Cumbria, 1950-93. Community deprivation (Townsend score) and individual social class were used to estimate socioeconomic status. Logistic regression was used to investigate risk of infant death (early neonatal, neonatal, and postneonatal) in relation to social class and Townsend deprivation score, adjusting for year of birth, birth order, multiple births, and stratified by time period, 1950-65, 1966-75, 1976-85, 1986-93. Results: The risk of infant death in all categories was higher in the lower social classes and more deprived communities, although inequality in risk of neonatal death declined after 1975 to such an extent that there was no significant difference in neonatal death rates by socioeconomic status in the most recent time period. By contrast, there was no narrowing in socioeconomic inequality in postneonatal death risk over the study period. Community deprivation was associated with a significant increased risk of postneonatal death after adjusting for individual level socioeconomic status. Conclusions: Postneonatal deaths remain higher in the most deprived communities and in the more disadvantaged social classes. The social, lifestyle, and environmental determinates of adverse health outcomes for children need to be fully understood, and interventions should be designed and targeted at the more socially deprived sectors of our community. Aiming to investigate infant deaths in Cumbria, 1950-93, in relation to individual and community level socioeconomic status. Methods: Retrospective birth cohort study of all 283 668 live births and 4889 infant deaths in Cumbria, 1950-93. Community deprivation (Townsend score) and individual social class were used to investigate risk of infant death (early neonatal, neonatal, and postneonatal) in relation to social class and Townsend deprivation score, adjusting for year of birth, birth order , multiple births, and stratified by time period, 1950-65, 1966-75, 1976-85, 1986-93. Results: The risk of infant death in all categories was higher in the lower social classes and more deprived communities, although inequality in risk of neonatal death declined after 1975 to such an extent that there was no significant difference in neonatal death rates by socioeconomic status in the most recent time period. By contrast, there was no nar rowing in socioeconomic inequality in postneonatal death risk over the study period. Community deprivation was associated with a significant increased risk of postneonatal death after adjusting for individual level socioeconomic status. Conclusions: Postneonatal deaths remain higher in the most deprived communities and in the more disadvantaged social classes. The social, lifestyle, and environmental determinates of adverse health outcomes for children need to be fully understood, and interventions should be designed and targeted at the more socially deprived sectors of our community.