This study examines mothers' and fathers' smoking patterns in different kinds of smoking households, and assesses their relative contribution to infants' exposure to environmental tobacco smoke. It uses data from a cross-sectional survey of 314 smoking households (infants: mean age 10 weeks) in Coventry and Birmingham, England, examining reported tobacco consumption and objective measures of exposure: the study infant's urinary cotinine:creatinine ratios and their mother's salivary cotinine. The study shows that both mothers' and fathers' tobacco smoke make substantial contributions to infant exposure to tobacco smoke. Households were more likely to contain a smoking father than mother, with over two-thirds of households including a smoking father. In households where both parents smoke, fathers' tobacco consumption was found to be significantly higher than in households where only the father smokes. This suggests that the interaction between parents needs to be considered rather than focusing on mothers' or fathers' smoking behaviour in isolation. The implications for health promotion programmes are discussed, particularly the need to place more emphasis on tackling fathers' smoking. Currently, fathers' smoking receives far less research or health promotion attention than mothers' smoking. Protecting infants from fathers' as well as mothers' smoking is key to reducing environmental tobacco exposure in early infancy, when the risk of Sudden Infant Death is highest.