All musicians are familiar with the repeat conventions of classical music and most are accustomed to standard responses to them which have been built into the performing practice of our own time. It is, for example, relatively common to hear the first repeats of sonata-form movements observed, much less common – in fact remarkably rare – to hear second-half repeats. The decision whether or not to observe a repeat is shared between current convention and the caprice of the performers. In modern practice, for example, unimportant factors make repeats more likely to be observed. The existence of first- and second-time bars is persuasive, for fear that valuable music under the first-time bar may otherwise be lost, even though the essence of the music is not affected: a first-time bar is, after all, usually the result of small awkwardnesses in the cadence rather than the expression of structural necessity. Similarly, in variations, just one variation having its repeat written out in varied or ornamented form is enough to constrain the players to observe every other repeat in the set when a first instinct might have been to observe none, or first-half repeats only. Then there is the universal habit of scrupulously observing minuet or scherzo repeats on the play-through and scrupulously overlooking them on the da capo, even when, as often happens in Beethoven's scherzos, the first section of the movement is a single eight-bar phrase which makes little sense unless it is repeated. It is odd that the minuet, often the least sophisticated section of a classical work, is heard generally three times while development and recapitulation, of intrinsically higher interest, are usually heard but once.