Preconditioning by one or several brief periods of ischemia activates an endogenous cardioprotective program that increases the resistance of cardiomyocytes to injury by subsequent prolonged periods of ischemia. Ischemic preconditioning can be mimicked by K(+) channel openers and various other substances, a phenomenon termed pharmacological preconditioning. Initially, ischemic preconditioning has been ascribed to the opening of ATP-sensitive K(+) channels at the surface membrane of cardiomyocytes. Since 1997, numerous publications have implicated mitochondrial ATP-sensitive K(+) channels (mK(ATP)) as a major trigger and/or end effector of preconditioning. Diazoxide has been suggested to be a specific activator of mK(ATP) channels, and the substituted fatty acid 5-hydroxydecanoate (5-HD) has been suggested to be a specific inhibitor. However, diazoxide and 5-HD have multiple K(+)-channel-independent actions, and the experimental evidence for an obligatory role of mK(ATP) channels in preconditioning, or even their existence, remains inconclusive. In contrast, surface K(ATP) channels have been well characterized, and we summarize the evidence suggesting that they make a major contribution to preconditioning. We also discuss a number of other factors involved in preconditioning: (1) generation of reactive oxygen species, (2) impairment of fatty acid metabolism, and (3) opening of the mitochondrial permeability transition pore. In the light of these emerging concepts, we critically re-examine the evidence for and against a role of mK(ATP) channels in ischemic and pharmacological preconditioning.