Variation form, as we saw on pages 118–19, entails the repetition of a clearly defined melodic unit, the theme, with various changes at each repetition. In the Baroque era, the theme was usually a bass pattern (sometimes called a ground bass). The same basic principle is at work in Classical variation form, but now the theme is a tune in the upper register.

We can understand why the Baroque era, which developed the idea of the basso continuo supporting harmonies from below, would have cultivated variations on a bass pattern, whereas the Classical era, with its emphasis on simple melody, preferred variations on short tunes in the upper register.

The point of variations is to create many contrasting moods with the same theme, which is transformed but always somehow discernible under the transformations. In principle, nothing distracts from this process, at least until the end, where composers usually add a coda. There are no contrasting themes, transitions, cadence sections, or development sections, as there are in sonata form movements (and in many rondos).

A Classical theme and variations movement begins with a theme that is typically in |: a :||: b :| or |: a :||: ba :| form. This miniform nests within the larger variation form:

Variations were part of the stock-in-trade of virtuosos of the Classical era. At a musical soiree, someone might suggest a popular opera tune, and the pianist would improvise variations on the spot, for as long as his or her imagination held out. Twelve was a common number for these variations when they were published; virtuosos piled them up for maximum effect. In symphonies and concertos, theme and variations movements are less extended, since they have to fit into a time scale with all the other movements.