This study examines the relationship between fertility decline in India and the evidence of an increase in sex bias. Data were obtained from the 1981 and 1991 Registrar General of India and from Khanna re-study villages in the Punjab (moderate-fertility population) and in Kerala (low-fertility population) in 1991. Total fertility declined by 20% during 1981-91. The number of sons desired by women who were childless declined by only 7.4%. Among the Khanna sample, reduced fertility led to a decline in excess mortality of women from 9% to zero when the sex bias was unchanged and fertility level varied. When fertility was kept constant at a low level and the sex bias varied, excess mortality of women increased from zero to 25%. The findings suggest that changes in birth distribution by parity outweigh intensified sex bias at lower parities. The sex ratios of children during 1981-91 rose in all states, including the south where the sex ratio has generally been more balanced. Data indicate that over a million additional girls aged 0-6 years were missing during 1981-91. Declining fertility led to a reduction in excess mortality of adult females. The sex ratios of age-specific deaths remained constant during 1979-81 and 1990-91. The author estimates how much the sex ratio of children can be attributed to sex-selective abortions. During 1981-91 about 4.2 million excess postnatal deaths occurred among girls, or 4 excess postnatal deaths for each excess prenatal death among girls (1 million aborted female fetuses). This suggests that sex-selective abortion accounted for the missing girls. The sharpest rise in the sex ratio at birth with parity was in Punjab for 1990-91. The model of the relationship between the decline in total fertility in each major state with a change in the sex ratios of children during 1981-91 indicated that the sex ratio increased more in northern states with less fertility decline.