Fertility has declined substantially in developing countries in the period since 1960, primarily as the result of increases in contraceptiveprevalence. Little dispute isfound on this point, but considerable debate has arisen about the causes of the increase in contraceptive prevalence. One unresolved issue is the causal contribu- tion of changes in fertility desires. The sources of increase in contraceptivepreva- lence are analyzed in 22 countries in Latin America, Asia, and Africa in the period from the 1970s to the 1990s, using World Fertility Survey and Demographic and Health Survey data. Through regression decomposition, change in prevalence is at- tributed to change,in fertility preferences (“composition”) and change in rates of use within preference categories (“rates”). The rates component,can be viewed,as the increase in prevalence resulting from increased implementation,of preferences. Two fertility-preference variables are examined, the desire for another birth and the dif- ference between actual and ideal family size. The rates component,dominates,in all 22 countries, ranging between 75 percent and 90 percent in most of the countries. In only two countries does the composition component,exceed 25 percent. The results refute demand-side explanations that ignore or dismiss the potential for substantial increase in prevalence through the satisfaction of existing demand. Although changes in fertility preferences account for a small proportion of the increase in prevalence, this component,weighs more heavily in societies at the early stages of transition and in sub-Saharan Africa. In these settings, declining demand for children has been a significant underlying cause of increases in contraceptive prevalence, and further declines in demand are required if the transition is to proceed to replacement-level fertility. Since 1960, fertility has declined substantially in developing countries. Increased