Cancer is essentially a somatic evolutionary process and is, therefore, effectively defined by the genetic and epigenetic changes underlying this process. An understanding of the function of these changes is fundamental to devising new approaches to prevention and treatment. Colorectal cancer (CRC), apart from its obvious importance as one of the most frequent cancers, provides an excellent model for such studies because of the availability of precursor adenoma lesions and the existence of several clear-cut familial inherited susceptibilities. These include familial adenomatous polyposis (FAP), which led to the identification of the APC gene and the importance of the Wnt pathway, and hereditary non-polyposis CRC (HNPCC), which identified the role of the mismatch repair genes in colorectal and other cancers. The presently known range of genetic and epigenetic changes in CRCs and adenomas is reviewed in this paper and the evidence against a requirement for genomic instability presented, together with a discussion of patterns of gene methylation, including especially our work on the homeobox gene, CDX1. Clearly, familial cancers, such as FAP and HNPCC, cannot account for more than perhaps 5% of the incidence of CRC. There is, however, evidence that approximately a further 25-30% have some inherited susceptibility. Based on the association of APC missense variants with multiple adenomas, we proposed that much of this may be due to the cumulative effects of low frequency, low penetrance variants, and the "rare variant hypothesis". The evidence for this from our work on multiple adenoma cases, and certain other examples, is discussed.