We studied the association between multivitamin use during the periconceptional period and the occurrence of neural tube defects using data from the Atlanta Birth Defects Case-Control Study. There were 347 babies with neural tube defects who were live born or stillborn to residents of metropolitan Atlanta from 1968 through 1980. The 2829 control-babies born without birth defects were randomly selected through birth certificates. Periconceptional multivitamin use was defined as reported use for each of the three months before conception through the first three months of pregnancy. Mothers who reported not using multivitamins any time during the six-month period were defined as nonusers. Fourteen percent of mothers reported periconceptional multivitamin use and 40% reported nonuse. Multivitamin users were different from nonusers in a number of demographic, health-related, and life-style characteristics. We found an overall apparent protective effect of periconceptional multivitamin use on the occurrence of neural tube defects, with a crude estimated relative risk of 0.40 (95% confidence interval, 0.25 to 0.63). At this time, it is not possible to determine whether this apparently lower risk is the direct result of multivitamin use or the result of other characteristics of women who use multivitamins.