A recent crosslinguistic study of positional privilege in vowel harmony observes that although prominent positions may uniquely block harmony, they are not singled out for transparency (Kaplan & Walker to appear). This typological gap is not predicted under a traditional set of theoretical assumptions in Optimality Theory (OT). If transparent vowels are skipped by harmony, a grammar is predicted in which positional faithfulness for the stressed syllable and a spreading imperative constraint are enforced at the cost of the constraint that penalizes skipping: ˈσ-Faithfulness >> Spreading Imperative >> *Skip. In this scenario, harmony would operate among unstressed vowels but skip those that are stressed. To address this unwanted prediction, this paper proposes an analysis that employs gestural representations (Browman & Goldstein 1986, 1995), centering on two key points. First, spreading does not actually skip vowels. Transparency results when the goal state of the harmonizing gesture is not achieved in a vowel due to blending with a concurrently active opposing gesture (Benus 2005, Gafos & Benus 2006, Smith 2018). Second, stressed vowels resist transparency because they are a locus of hyperarticulation due to a prosodic μ-gesture (Katsika & Tsai 2021), a context hypothesized to disfavor concurrently active opposing gestures. In the Gestural Harmony Model (Smith 2018), transparent vowels violate a version of a constraint *Overlap(Gest-x, Gest-y) that prohibits overlap of opposing gestures. Extending this approach, it is proposed that *Overlap constraints may incorporate prosodic μ-gestures, which are associated with a metrically prominent position: *Overlap(μ-Gest, Gest-x, Gest-y). The prosodically sensitive *Overlap constraint stands in a stringency relation with the general *Overlap constraint, which can give rise to asymmetric transparency in unstressed vowels but not stressed vowels. The proposal is illustrated in application to laxing/RTR harmony in the Eastern Andalusian Spanish variety of Granada (Jiménez & Lloret 2007, 2020, Lloret & Jiménez 2009), in which unstressed vowels may optionally exhibit harmony but stressed vowels consistently harmonize.