According to the classical interpretation of Knudson's 'two-hit' hypothesis for tumorigenesis, the two 'hits' are independent mutation events, the end result of which is loss of a tumor suppressing function. Recently, it has been shown that the APC (adenomatous polyposis coli) gene does not entirely follow this model. Both the position and type of the second hit in familial adenomatous polyposis (FAP) polyps depend on the localization of the germline mutation. This non-random distribution of somatic hits has been interpreted as the result of selection for more advantageous mutations during tumor formation. However, the APC gene encodes for a multifunctional protein, and the exact cellular function upon which this selection is based is yet unknown. In this study, we have analyzed somatic APC point mutations and loss of heterozygosity (LOH) in 133 colorectal adenomas from six FAP patients. We observed that when germline mutations result in truncated proteins without any of the seven beta-catenin downregulating 20-amino-acid repeats distributed in the central domain of APC, the majority of the corresponding somatic point mutations retain one or, less frequently, two of the same 20-amino-acid repeats. Conversely, when the germline mutation results in a truncated protein retaining one 20-amino-acid repeat, most second hits remove all 20-amino-acid repeats. The latter is frequently accomplished by allelic loss. Notably, and in contrast to previous observations, in a patient where the germline APC mutation retains two such repeats, the majority of the somatic hits are point mutations (and not LOH) located upstream and removing all of the 20-amino-acid repeats. These results indicate selection for APC genotypes that are likely to retain some activity in downregulating beta-catenin signaling. We propose that this selection process is aimed at a specific degree of beta-catenin signaling optimal for tumor formation, rather than at its constitutive activation by deletion of all of the beta-catenin downregulating motifs in APC.