Objectives: To examine the stability of the cubic (two points of inflection) exercise heart rate-music-tempo preference relationship found by Karageorghis et al. (2011) in cycle ergometry using a different exercise modality (treadmill exercise). To advance previous related studies through the inclusion of psychological outcome variables (e.g., state attention and intrinsic motivation) and post-experiment interviews. Design: A mixed-model experimental design was employed with two within-subject factors (exercise intensity and music tempo) and a between-subjects factor (gender). The experiment was supplemented by qualitative data that were analyzed using inductive content analysis. Methods: Participants (n = 22) exercised at six intensities (40-90% maxHRR) during which they were exposed to music tracks at four tempi and a no-music control. Music preference, affective valence, and perceived activation were assessed during the task. Immediately afterwards, an attentional focus item, the short Flow State Scale-2 and items from the Intrinsic Motivation Inventory were administered. A subsample of participants (n = 8) was interviewed using a schedule of open-ended questions. Results: Results did not support a cubic relationship but rather a quadratic one (one point of inflection), and there was a weak association between the optimal choice of music tempo and positive psychological outcomes. Conclusions: The range of preferred tempi for treadmill exercise (123-131 bpm) was narrower than that for cycle ergometry (125-140 bpm). Regardless of its tempo, music reduced the number of associative thoughts by similar to 10% across all exercise intensities. (C) 2013 Elsevier Ltd. All rights reserved.