Heritable genomic variation and natural selection have long been acknowledged as striking parallels between evolution and cancer. The logical conclusion, that cancer really is a form of speciation, has seldom been expounded directly. My purpose is to reexamine the "cancer as species" thesis in the light of current attitudes to asexual speciation, and modern analyses of species definitions. The chief obstacles to accepting this thesis have been the asexual nature of cancer cell reproduction, the instability of the malignant genotype and phenotype, and our conditioning that speciation is an extremely rare and imperceptibly gradualistic process. However, these are not absolute barriers to the acceptance of cancers as bona fide species. Furthermore, although ongoing clonal evolution of extant cancers also results in a series of secondary speciation events, the initial emergence of a cancer requires a level of taxonomic reclassification even beyond the concept of speciation (i.e., phylogenation), and which is almost certain to provide a rich source of novel drug targets. The implications of the "cancer as species" idea may be as important for biology as for oncology, providing as it does an endless supply of observable if accelerated examples of a phenomenon once regarded as rare.From the perspective of cancer treatment, speciation guarantees the existence of causal molecular mechanisms which may have been neglected as exploitable targets for rational therapy; in particular, the mediators of metazoan life seem to have substantial overlap with components commonly deranged in cancer cells. However, the intractability of the drug resistance problem, residing as it does in the inherent plasticity of the genome, is traceable back to, and inseparable from, the very origins and nature of life.
K E Y W O R D S :Cancer, drug resistance, mutator phenotype, speciation, tumor biology.