SITE-SPECIFIC mutation¹ and membrane reconstitution² experiments provide compelling evidence that the product of the gene which is at fault in the disease cystic fibrosis, termed the cystic fibres is transmembrane conductance regulator (CFTR)³, is a small-conductance chloride channel activated by phosphorylation⁴. As transport of chloride ions is passive, the predicted presence of two nucleotide-binding domains in CFTR seems as puzzling as a report⁵ that ATP hydrolysis is essential to activate the channel. We now find that in the sweat duct, which expresses high levels of CFTR⁶ and has a very high Cl⁻ conductance⁷, intracellular concentrations of ATP must be about normal (5 mM)⁸ for activation of this conductance, apparently by a non-hydrolytic, perhaps allosteric, mechanism. This passive dependence on ATP should mean that even a modest depletion of cell energy levels will significantly lower the energy demands of electrolyte transport by decreasing chloride conductance. We believe this direct coupling between cellular ATP levels and chloride channel activity is an adaptive mechanism to protect the tissue from damage resulting from excessive energy depletion⁹.