In jazz theory and pedagogy, the so-called chord-scale theory (CST) is a widely-used system to describe and prescribe associations of scales to chords and vice versa. Despite its dominance, a large-scale empirical evaluation of its significance for jazz history has not yet been carried out. In particular, the exact nature of mappings between chords and scales is debated, in particular regarding the inherent tension between vertical (chordal) and horizontal (melodic) tone coordination as well as the influence of larger tonal areas on scale selection. Here, we evaluate CST empirically by drawing on a corpus of 456 jazz solos contained in the Weimar Jazz Database. We give a comprehensive description of pitch choices in (tonal) jazz solos with regard to underlying chord types and their scale degree, and determine the most commonly used scales over the most relevant chord types using suitable matching algorithms. Finally, we present three case studies that address important phenomena in jazz harmonic theory in more depth: 1) tritone substitutions of dominants, 2) fully diminished chords and their underlying scale(s), and 3) the role of the sixth scale degree in pre-dominant chords. Our study demonstrates how musicological theories can be evaluated empirically and provides the first comprehensive overview of actual pitch and scale use in jazz history. Results indicate that scale choices for chords with a diatonic interpretation are influenced by their local scale degree and thus by local (or global) key contexts, whereas inherently non-diatonic chords allow for more freedom. Generally, the presence of chromaticism beginning in the early jazz styles blurs the mapping of scales to chords and renders the question of the prevalence of melodic and chordal thinking. This suggests the need for a more integrated approach to adequately describe chord-scale relationships.