Some aspects of sampling family data in human genetics are discussed. Two well-investigated ascertainment models, the complete ascertainment model and the single ascertainment model which have been derived using restrictive sampling assumptions, are shown to hold under 4 more general sets of assumptions.