Distylous species have populations with two floral morphs bearing stigmas and anthers positioned reciprocally. This arrangement, assisted by the flower-pollinator fit, facilitates pollen deposition in different parts of pollinators' bodies, promoting outcrossing between morphs (i.e., legitimate pollen deposition - LPD). Typically, distyly occurs in one dimension (i.e., in height; 1D-heterostyly) but it can also appear as 3D-heterostyly, with stamens and styles bent and twisted. It is hypothesized that 3D-heterostyly improves reciprocity and LPD, reducing pollen loss compared to 1D-heterostyly. Variations in pollinator assemblages may lead to divergent selective pressures, resulting in different degrees of reciprocity and pollen flow patterns among populations of 1D- and 3D-heterostylous species. We tested these hypotheses in Linum tenue (1D-heterostyly) and L. suffruticosum (3D-heterostyly) using data on pollen deposition on stigmas, pollinators and reciprocity from six natural populations occupying contrasting environmental conditions in Spain. LPD was higher in L. suffruticosum than in L. tenue only when specialized pollinators predominated (Usia bee flies; Bombyliidae). Both pollinator frequency and the reciprocity of sex organs were associated with LPD in L. suffruticosum. In contrast, L. tenue showed consistently moderate LPD and high reciprocity across populations, despite variation in pollinator assemblages. However, frequent spontaneous self-pollination in L. tenue might reduce LPD on its stigmas. 3D-heterostyly appears more effective in specific pollination contexts but may be vulnerable to shifts in pollinator availability. In 1D-heterostylous populations, the independence of LPD from pollinator assemblage composition may explain the prevalence of 1D-heterostyly, although persistent spontaneous self-pollination in L. tenue could threaten the long-term maintenance of polymorphism.