Estimates of abundance of medium-to-large mammals by traditional mark-recapture models may be unreliable because quantity and quality of trapping data are low. The proposed closed-subpopulation model provides a flexible framework to increase the amount of data used for estimation of demographic parameters, by taking into account characteristics of the population and using ancillary non-trapping data. This model defines a subsection of the population that is known to be alive and within the study area during a certain period, regardless of which animals were actually caught. Population size is estimated from the proportion of animals in this closed subpopulation that were actually captured. We used this model to estimate size of a partly culled population of Eurasian badgers (Meles meles). Number of badgers included in the closed subpopulation was maximized by using data from trapping, road-traffic accidents, and radiotelemetry, and by assuming that no additions occurred to the population of young between trapping occasions. Probabilities of capture varied by season and age-class but not sex, trapping, or radio-tagging. Population estimates appeared reliable because estimated number of times individual badgers were trapped in a year corresponded with observed frequencies and estimated size of the young and adult populations corresponded favorably with estimates based on a mark-resight procedure.