In modern analyses of pop/rock music, a song is parsed into a succession of section categories drawn from a limited set of standard labels, e.g., verse, chorus, and bridge. Despite the common currency of these labels, scholars to date have provided only very general, if not conflicting, theories as to what sort of musical parameters engender or correlate with our perception of these categories. For example, Everett (2001) states that verse and chorus sections nearly always prolong the tonic, while Endrinal (2008) and Stephan-Robinson (2009) associate an emphasis on tonic more strongly with chorus than verse sections. Furthermore the supporting evidence for such statements has been limited mostly to a few exemplars rather than any large body of data. One notable exception is Summach (2012), who provides statistics on tonal closure in various section types. 
This paper continues the investigation of how harmonic traits align with typical section categories using an empirical approach. In this study, I examine the interaction of form and harmony in rock music within a corpus of 200 rock songs, drawn from Rolling Stone magazine’s list of the “greatest songs of all time”. Each song was encoded by two separate musicians, so first I appraise the extent of subjectivity for this type of formal analysis in general. I then present statistics on chord quality, chord inversion, and root distribution for individual song sections, which I compare to the global data as reported in my corpus study of rock harmony with Temperley (2011). Other results include information on chord durations and chord transitions, both for the corpus as a whole and for individual section types. It is shown that individual section types display significant harmonic differences from rock harmony overall, and these differences may help us understand our perception of form in this style.