A description of quantum systems at the time interval between two successive measurements is presented. Two wave functions, the first preselected by the initial measurement and the second postselected by the final measurement, describe quantum systems at a single time. It is shown how this approach leads to a new concept: a weak value of an observable. Weak values represent novel characteristics of quantum systems between two measurements. They are outcomes of a standard measurement procedure that fulfills certain requirements of ``weakness.'' We call it weak measurement. Physical meaning, underlying mathematical structure, and prospects of practical usage of weak measurements are explored.