It is widely accepted that changes in the environment affect mean trait expression, but little is known about how the environment shapes intra-individual and intra-population variance. Theory suggests that intra-individual variance might be plastic and under natural selection, rather than reflecting developmental noise, but evidence for this hypothesis is scarce. Here, we experimentally tested whether differences in intrinsic environmental predictability affect intra-individual and intra-population variability of different reproductive traits, and whether intra-individual variability is under selection.Under field conditions, we subjected Onobrychis viciifolia to more and less predictable precipitation over 4 generations and 4 years. We analysed effects on the coefficient of intra-individual variation (CVi-i) and the coefficient of intra-population variation (CVi-p), assessed whether the coefficients of intra-individual variation (CsVi-i) are under natural selection and tested for transgenerational responses (ancestor environmental effects on offspring).Less predictable precipitation led to higher CsVi-i and CsVi-p, consistent with plastic responses. The CsVi-i of all studied traits were under consistent stabilizing selection, and precipitation predictability affected the strength of selection and the location of the optimal CVi-i of a single trait. All CsVi-i differed from the optimal CVi-i and the maternal and offspring CsVi-i were positively correlated, showing that there was scope for change. Nevertheless, no consistent transgenerational effects were found in any of the three descendant generations, which contrasts with recent studies that detected rapid transgenerational responses in the trait means of different plant species. This suggests that changes in intra-individual variability take longer to evolve than changes in trait means, which may explain why high intra-individual variability is maintained, despite the stabilizing selection.The results indicate that plastic changes of intra-individual variability are an important determinant of whether plants will be able to cope with changes in environmental predictability induced by the currently observed climatic change.