Published online by Cambridge University Press: 31 July 2008
A function is proposed for estimating the proportion of women who resumed menstruation at a given time post-partum on the basis of the proportion of women breast feeding at that time. The function was fitted to 25 series of incidences from nine countries (236 time points), most countries providing three contrasting groups: urban (well-to-do), urban (poor) and rural. The average error of the estimate within series was 0·08. At any given time post-partum variation in breast feeding proportions [explained] on average 85% of between series variance in the proportions of menstruating women. Some uses of the function are briefly considered.