The problems of genetic counseling in hereditary polyposis coli (HPC)(taken as the type of the age-dependent dominant mendelian trait) are discussed in some detail. They are threefold: first to formalize, for purposes of decision, the total penalty (the "fardel") typically imposed by the disorder for each case, and how it may be modified by treatment; second, to determine the logical issues involved in making probability statements in the face of the uniqueness of each case; third, to use to best advantage the information on the pattern of onset to assess the probability that a person, at risk but not so far affected, does in fact harbor the gene. The third problem points up the need for a formal model of the pathogenesis and its implications for the pattern of onset and the sensitivity of the assessment to the assumptions of the model, especially where one has to rely on cross-sectional, as distinct from longitudinal, data.