The objective of this study was to examine the relationship between maternal smoking status and infant nutrition. Women delivering in 5 hospitals in the Ottawa-Carleton region of Ontario were screened for eligibility over a 6-month period in 1993. Follow-up data were collected by telephone at 3 months postpartum using a validated questionnaire. Mothers were retrospectively asked about their infant-feeding choices and their smoking behaviours. Logistic regression analyses were used to determine the significance of demographic factors and maternal smoking behaviours in relationship to 3 feeding practices: bottlefeeding at birth, discontinued breastfeeding by 12 weeks, and introduction of solids by 12 weeks. A total of 796 women participated in the longitudinal study (90% follow-up rate). Less-educated; younger; single, separated or divorced; and foreign-born mothers were more likely to bottlefeed at birth. Less-educated women more often discontinued breastfeeding before 12 weeks. Mothers who had smoked during part or all of their pregnancy and were smoking at the time of the interview were significantly more likely than non-smokers to bottlefeed at birth or to discontinue breastfeeding by 12 weeks. Current smokers were also more likely than non-smokers to have introduced solid food by 12 weeks. Maternal smoking was a significant predictor of infant nutrition, with other sociodemographic factors taken into account. Smoking status should be included in clinical screening tools for infant nutrition. The relationship between other dimensions of maternal smoking (e.g., timing of quitting attempts, degree of partner support, partner's smoking behaviours) and infant-feeding practices warrants investigation.