V1 responses to an optimally oriented test line in the receptive field center may be modulated by placing lines of same or different orientations in the surround. While iso-orientation produces strong inhibition, cross-orientation enhances the response [Knierim, J. J., & Van Essen, D. C. (1992). Neuronal responses to static texture patterns in area V1 of the alert macaque monkey. Journal of Neurophysiology, 67, 961–980]. We looked for a perceptual correlate of neuronal texture modulation using perceived salience as well as fading and filling-in as response criteria. Two patterns by Vicario (1998) [Vicario, G. B. (1998). On Wertheimer’s principles of organization. In G. Stemberger, (Ed.), Gestalt theory, (Vol. 20, pp. 256–270). Vienna: Verlag Krammer] served as targets. One consisted of randomly oriented bars in the center and uniformly oriented bars in the surround, while the other had bars of uniform orientation in the center and bars of random orientation in the surround. In spite of identical texture contrast at the boundary, the first pattern was judged more salient than the second and its center took more time to fade. When the surround was decreased in width, fading time followed no systematic trend and filling-in was increasingly replaced by filling-out. A higher salience and longer fading time for stimuli with a uniformly as opposed to randomly oriented surround was also obtained when the bars in the center were replaced by dotted arrays. However, no asymmetry was found for the converse patterns when dots were in the surround and bars in the center. Findings are interpreted in terms of stronger surround suppression exerted by randomly oriented bars as compared to uniformly oriented bars. Modeling suggests that this suppression may mediate a statistical computation by the visual system, aimed at detecting a texture boundary between the center and surround when the center texture was likely generated by a different random process than that which generated the surround.