In the last decade, the concept of "cancer stem cells" has emerged, recognised by the fact that only a small fraction of tumour cells appears to retain the stem cell properties of self-renewal and unlimited proliferation. At the same time, it is well known that cancer is an age-related disease developing at the limit of proliferating cell senescence. The apparent need to link senescence and the capacity for self-renewal has lead some authors to suggest that cancers develop from amongst senescing stem cells. However, an alternative solution has recently been proffered by Sundaram M, Guernsey DL, Rajaraman MM, Rajaraman R [Neosis: a novel type of cell division in cancer. Cancer Biol Ther 2004;3:207-18], who suggest that stemness may be a transient, cyclic property afforded by de-polyploidisation of senescing cells which have undergone polyploidisation. In this mini-review, we attempt to reconcile both of these views by the idea that cycling polyploidy intermitting senescence and rejuvenation may be features of a life cycle analogous to the life cycles of certain unicellular organisms. Furthermore, we suggest that mitotic catastrophe may represent a mechanism through which the cell can switch from the usual mitotic cell-cycle to this evolutionarily conserved life cycle. Intriguingly, some most recent data suggest that cell senescence may be reversible and that stem cells are tolerant to polyploidy caused by genotoxic stress.