According to the vocal similarity hypothesis (VSH; Bowling et al., 2018), harmonicity—the extent to which the partials of a complex tone approximate a harmonic series—may be associated with sensory consonance because of an evolved preference for sounds that resemble human speech. This suggests that even if individuals were exposed to chords from an entirely unfamiliar musical system, they should be inclined to evaluate such chords more favorably when these are higher in harmonicity. We tested this in a study using stimuli generated from a highly unconventional scale, the just tempered Bohlen-Pierce (BP) scale. This scale divides a tritave (representing the span of an octave plus a perfect fifth) into 13 intervals based on odd integer frequency ratios. Participants were randomly presented with and asked to rate the consonance of every possible BP dyad and triad within a tritave. Chords were presented in multiple timbres. The harmonicities of each chord were computed using algorithms devised by Bowling et al. (2018) and by Harrison and Pearce (2020). We also compared the predictive utility of these models with that of a classic spectral interference (SI) model of consonance perception (Hutchinson & Knopoff, 1978). Results showed that harmonicity was weakly correlated with consonance and that only SI uniquely predicted consonance judgments when participants were presented with chords from an unfamiliar tuning system. This lends credence to the possibility that the link between harmonicity and consonance reported by Bowling et al. (2018) may have partially resulted from confounding between harmonicity and either SI and/or familiarity. Although these results do not disconfirm the VSH, at minimum, they suggest that the extent to which vocal similarity contributes to consonance judgment may vary substantially based on the musical context.