For more than half a century it has been generally believed that cancer cells are more acidic than normal cells. This dogma arose from the studies of Warburg and co-workers (Warburg, 1930) who showed that tumour cells preferentially convert glucose and other substrates to lactic acid, even under aerobic conditions. Since lactic acid has a pK of 3.7 it seemed obvious that the intracellular fluid would become acidic. Studies on cultured tumour cells or spheroids often showed low intracellular pH (pHi) values, but this could have been due to their highly artificial situation. What would be the pHi of cells in a solid tumour in a patient? The development of pH electrodes small enough to be inserted into living tissues led to the apparent confirmation of the prevailing wisdom. Numerous studies (reviewed by Wike-Hooley et al., 1984 and Vaupel et al., 1989; see also Figure lb) showed significantly more acidic pH (pHpoT) in tumours than that in normal tissues. Microelectrodes that can be used on solid tumoursin vivo are usually quite large in com-parison to a tumour cell (see Wike-Hooley et al., 1984, Table II), and they mainly measure the pH of the extracellular fluid (pHe) rather than pHi (Vaupel et al., 1989). For most purposes, the parameter of interest is pH1, the pH of the water in the cancer cell itself, but it was generally expected (and then tacitly assumed) that pHi would also be acidic. This supposed cellular acidity intumours became part of the mental wallpaper of oncologists and cancer researchers, even though there was, for many years, no practical wayto measure pHi of intact human tumours. It had clinical consequences, too, since anticancer treatments were often designed to takeadvantage of a low pH,(for a review, see Wike-Hooley et al., 1984). It was argued, for instance, that anticancer drugs would be more effective if they contained ionising groups that would cause them to be trapped in acidic environments (Wike-Hooley et al., 1984), or that radioresistant hypoxic cells would have a particularly low pHi and might therefore be especially sensitive to treatments such as hyperthermia which are known to act preferentially on isolated cells in acidic media (Freeman et al., 1981). There were also numerous attempts to lower pHi still further by administration of glucose, and thereby enhance the action of various pH-sensitive therapies (Ross, 1961, reviewed by Wike-Hooley et al., 1984). In general, these ideas have had little clinical success, but they are still the subject of active research (see, for instance, Tannock & Rotin, 1989). Within the last 10 years a non-invasive intracellular pH meter-the Nuclear Magnetic Resonance Spectrometer-has become widelyavailable; it can be used on living tumours in situ, both in experimental animals and in man. The results obtained with these instruments have been surprising. Instead of having the expected acidic pHi, the cells of intact tumours turned out to be neutral, or a little alkaline, both in experi-mental animals (Griffiths et al., 1981; Iles et al., 1982) and man (Griffiths et al., 1983). Indeed, several studies by 31P