in Koshu City, Japan, and their mothers. Maternal smoking during early pregnancy was the exposure studied. Main outcome measures: Childhood body mass index (BMI) and BMI z-score trajectories of the children born to the smoking and non-smoking mothers by gender. Multilevel analysis that includes both individual and age as different-level variables was used for statistical analyses. Results: The participating mothers delivered 1619 babies during the study period. Birth weight and anthropometric data were collected from 1603 (at birth, 99.0%), 1358 (at age 3, 83.9%), 1248 (at age 5, 77.1%), 1270 (at age 7-8, 78.4%) and 1274 (at age 9-10, 78.7%) of these children. The mean birth weight of both the male and female children whose mothers had smoked during pregnancy was significantly low compared with those born to non-smoking mothers (Po0.01). However, the childhood BMI at each subsequent checkup age significantly increased only among the male children born to the smoking mothers. Moreover, this increase was continuously observed after 3 years of age. The results of BMI z-score analysis were also similar to these of BMI analysis. Conclusions: Smoking by pregnant women decreases the infant birth weight irrespective of gender but increases childhood weight gain especially by male children. The results might be valuable to explore the mechanism of fetal programming.