It is well known that specific agents active in the early months of pregnancy may cause certain congenital malformations. German measles was reoognized in the 1940's as one cause; more recently thalidomide taken by mothers early in pregnancy was found to be responsible for deformities in thousands of babies. These and similar discoveries have spurred the search for other agents operating early in pregnancy.On pursuing such research, however, one may find no routine medical records of factors of possible relevance. Hence, existing sources of data concerning events early in pregnancy need to be supplemented (1). The time of collection of the data from the mother is crucial. Gathering information at the end of pregnancy may introduce serious biases. For example, the mother of a malformed child, seeking a reason for the child's defect, is more likely to recall and report unusual events during pregnancy than is the mother of a nornal child (2).Obtaining the desired information at the time of the woman's first prenatal visit to the physician should reduce these biases, but difficulties in recall would still remain. Even among women obtaining prenatal care in the first trimester, only a small proportion start such care