Renewed interest in the relationships between viability and culturability in bacteria stems from three sources: (1) the recognition that there are many bacteria in the biosphere that have never been propagated or characterized in laboratory culture; (2) the proposal that some readily culturable bacteria may respond to certain stimuli by entering a temporarily non-culturable state termed 'viable but non-culturable' (VBNC) by some authors; and (3) the development of new techniques that facilitate demonstration of activity, integrity and composition of non-culturable bacterial cells. We review the background to these areas of interest emphasizing the view that, in an operational context, the term VBNC is self-contradictory (Kell et al., 1998) and the likely distinctions between temporarily non-culturable bacteria and those that have never been cultured. We consider developments in our knowledge of physiological processes in bacteria that may influence the outcome of a culturability test (injury and recovery, ageing, adaptation and differentiation, substrate-accelerated death and other forms of metabolic self-destruction, prophages, toxin-antitoxin systems and cell-to-cell communication). Finally, we discuss whether it is appropriate to consider the viability of individual bacteria or whether, in some circumstances, it may be more appropriate to consider viability as a property of a community of bacteria.