Preliminary analysis revealed very similar pattern of results for the 4 scales. Consequently, they were collapsed into a broader cat- egory of shape roundness/angularity in order to assess any differ- ence in taste-shape correspondences as a function of the taste word under consideration. A one-way repeated measures analysis of variance (RM-ANOVA), using the Greenhouse Geisser correction, revealed a significant effect of taste word, F(1.857, 46.422) = 17.508, p < .001, g2partial = .412. Pairwise comparisons (Bonferroni corrected) revealed that sweet was rated as rounder than any of the other tastes (p 6 .001, for all comparisons, see Fig. 2A).In addition, one-sample t-tests were performed in order to determine whether the ratings on each taste differed from the mid-point of the scale (50). This analysis revealed that participants rated bitter, t(25) = 3.711, p = .001, salty, t(25) = 3.257, p = .003, and sour, t(25) = 2.921, p = .007, significantly toward the angular end of the scales, and sweet significantly toward the round end of the scales, t(25) = 5.421, p < .001. These results therefore suggest that rather than there being taste specific associations to roundness/ angularity, the correspondence between shape and taste only seems to capture the distinction between sweetness and the other three most commonly mentioned basic tastes (bitter, salty, and sour)