Abstract
A model of delayed response, developed within the frame of a general sociological analysis, is applied to oscillations in aggregate legitimate fertility, considered as delayed responses to variations in the marriage rate. The appropriate lag function is derived on the basis of an assumed fertility schedule by duration of marriage. It is shown that if oscillations were caused solely by marriage fluctuations they would be very smooth with an ‘average period’ of 25 years and upwards. In Swedish data from the period 1830–1879 this smooth pattern is discernible as one component, overlaid by short-term fluctuations for which some other cause must be found. After a discussion of alternative explanations it is submitted that the cause is the practice of birth control on a fairly extensive scale, long before the secular decline in marital fertility, a conclusion which ties in with the results of earlier analyses of nineteenth-century fertility.