We examine cohort trends in premarital first births for U.S. women born between 1920 and 1964. The rise in premarital first births is often argued to be a consequence of the retreat from marriage, with later ages at first marriage resulting in more years of exposure to the risk of a premarital first birth. However, cohort trends in premarital first births may also reflect trends in premarital sexual activity, premarital conceptions, and how premarital conceptions are resolved. We decompose observed cohort trends in premarital first births into components reflecting cohort trends in (1) the age-specific risk of a premarital conception taken to term; (2) the age-specific risk of first marriages not preceded by such a conception, which will influence women's years of exposure to the risk of a premarital conception; and (3) whether a premarital conception is resolved by entering a first marriage before the resulting first birth (a "shotgun marriage"). For women born between 1920-1924 and 1945-1949, increases in premarital first births were primarily attributable to increases in premarital conceptions. For women born between 1945-1949 and 1960-1964, increases in premarital first births were primarily attributable to declines in responding to premarital conceptions by marrying before the birth. Trends in premarital first births were affected only modestly by the retreat from marriages not preceded by conceptions-a finding that holds for both whites and blacks. These results cast doubt on hypotheses concerning "marriageable" men and instead suggest that increases in premarital first births resulted initially from increases in premarital sex and then later from decreases in responding to a conception by marrying before a first birth.