After contraceptive use, breast-feeding duration is the major determinant of the birth interval length. Three methods of estimating births averted by breast-feeding, and the increase in contraceptive use needed to substitute for breast-feeding, are presented. Method 1 simply utilizes Bongaarts Ci, and the other two are based on mean birth intervals with and without breast-feeding. Estimates for each method are derived for six countries with DHS surveys from the mid-1990s: Burkina Faso, Uganda, India, Indonesia, Brazil and Peru. The estimated percentage of additional births that would occur if there were no breast-feeding ranged from 14% in Brazil to about 50% in Burkina Faso and Uganda, reflecting very low breast-feeding in Brazil and very high levels in the sub-Saharan African nations. Strengths and limitations of the three methods are considered. [ABSTRACT FROM AUTHOR]