CONTEXT: A cultural preference for sons may be a factor driving recourse to abortion in India, as women carrying female fetuses may decide to terminate their pregnancies. To assess this hypothesis, more information on the incidence of abortion, and on maternal and social correlates of the procedure, is needed. METHODS: Birth order-specific abortion ratios were calculated using the birth histories of 90,303 ever-married women aged 15-49 who participated in India's 1998- 1999 National Family Health Survey. For the first four births, the association between abortion and various maternal and social variables, including the sex of the respondent's last child, was assessed using logistic regression. RESULTS: The overall abortion ratio was 17.0 per 1,000 pregnancies. The ratio increased from 5.3 per 1,000 pregnancies for first-order births to 25.8 per 1,000 pregnancies for third-order births and then declined. The strongest predictor of abortion was maternal education: Women with at least a primary education were more likely than those with no education to have had an abortion (odds ratios, 1.9-6.7). Rural residence was associated with a reduced likelihood of abortion (0.6). There was no association between the sex of a woman's previous child and the odds that she subsequently had an abortion. CONCLUSION: At the national level, it is likely that unintended pregnancy, rather than the sex of the previous child, underlies demand for abortion in India. Rising educational attainment among women may lead to an increase in the demand for abortion.