Ahlburg (1983, 1986) tested a simple version of Easterlin's relative-cohort-size model of fertility on the basis of U.S. and Canadian post-war data. His conclusion was that the Easterlin model fits the data very well and can therefore be used for calculating forecasts. However, the model he estimated is oversimplified. In this paper an alternative specification is presented. The model is applied to Dutch fertility data. The Easterlin effect is found to affect the movement of births in the Netherlands during the period 1950-85, but a declining long-term trend in average family size proves far more important in explaining post-war births. The model forecasts a rise of births until 2000.