In a companion paper a logistic regression model of seroprevalence over time was developed on the basis of data obtained during an experimental infection of weaner pigs with classical swine fever (CSF) virus. The model was applied to seroprevalence data from three outbreaks of the 1993-1994 epizootic to test whether the model could predict correctly the day of virus introduction into the herd. It was concluded that the logistic regression model has potential as a tool to estimate in retrospect the day CSF virus was introduced into a pig herd, which in turn may assist in identification of risk factors implicated in the further spread of the disease.