Information on 869 076 singletons and 17 566 twins, born during the period 1983-1991, was obtained from the Swedish Medical Birth Registry. Data on birth weight, gestational duration, vital status, and maternal smoking habits during pregnancy were analyzed in order to investigate whether twinning potentiates the effect of maternal smoking on birth weight and perinatal mortality. The individual birth weights were expressed as percentages of mean birth weight, where mean birth weights of singletons and twins were calculated separately. The birth weight reducing effect of maternal smoking was found to be of the same magnitude among twins and singletons weighing > 90% of mean birth weight. For infants weighing < 90% of mean birth weight, maternal smoking had a significantly stronger effect on birth weight among singletons than among twins. When gestational duration was taken into consideration, this difference was less pronounced. The effect of maternal smoking on gestational duration was stronger among singletons than twins. The smoking-related risk increase of perinatal death was of about the same magnitude among twins and singletons.