The majority of physiological processes proceed most favourably when O(2) is in plentiful supply. However, there are a number of physiological and pathological circumstances in which this supply is reduced either acutely or chronically. A crucial homeostatic response to such arterial hypoxaemia is carotid body excitation and a resultant increase in ventilation. Central to this response in carotid body, and many other chemosensory tissues, is the rapid inhibition of ion channels by hypoxia. Since the first direct demonstration of hypoxia-evoked depression in K(+) channel activity, the numbers of mechanisms which have been proposed to serve as the primary O(2) sensor have been almost as numerous as the experimental strategies with which to probe their nature. Three of the current favourite candidate mechanisms are mitochondria, AMP-activated kinase and haemoxygenase-2; a fourth proposal has been NADPH oxidase, but recent evidence suggests that this enzyme plays a secondary role in the O(2)-sensing process. All of these proposals have attractive points, but none can fully reconcile all of the data which have accumulated over the last two decades or so, suggesting that there may, in fact, not be a unique sensing system even within a single cell type. This latter point is key, because it implies that the ability of a cell to respond appropriately to decreased O(2) availability is biologically so important that several mechanisms have evolved to ensure that cellular function is never compromised during moderate to severe hypoxic insult.