Objective To study the relative importance of various determinants of total and cause-specific infant and early childhood mortality rates and their decline in The Netherlands in the period 1875–1879 to 1895–1899. Data and Methods Mortality and population data were derived from Statistics Netherlands for 16 towns and 11 rural areas. Mortality levels and their decline were estimated with a Poisson regression model. The associations of the estimated levels and declines, and determinants of infant and early childhood mortality were analysed using multivariate linear regression analysis. The causes of death studied were major contributors to infant mortality (convulsions, acute digestive disease, acute respiratory disease) and early childhood mortality (encephalitis/meningitis, acute respiratory disease, measles). Results Infant mortality rates were high in the south-western part of The Netherlands in 1875–1879. Due to a rapid decline in the western regions, this pattern changed to a north-south gradient in 1895–1899. Early childhood mortality showed an urban-rural gradient in 1875–1879 with mortality high in towns. This gradient had largely disappeared by1895–1899, due to a rapid decline in mortality in towns. Roman Catholicism was significantly associated with infant mortality (particularly from diarrhoeal disease) in 1875–1879 and 1895–1899. The association with Roman Catholicism was stronger in 1895–1899 because mortality declines were less rapid in Roman Catholic areas in 1875–1879 to 1895–1899. Urbanization was significantly associated with early childhood mortality (particularly from respiratory disease) in 1875–1879 and 1895–1899. This association weakened over time, due to the rapid decline in mortality in towns. Conclusions Different determinants of mortality (decline) were important in infant and early childhood mortality and they acted on different causes of death. Therefore, infant and childhood mortality should be studied separately. International comparison of the results showed that findings with ...