Understanding the genetic consequences of changes in population size is fundamental in a variety of contexts, such as adaptation and conservation biology. In the study presented here, we have performed a replicated experiment with the plant Nigella degenii to explore the quantitative genetic effects of a single-founder bottleneck. In agreement with additive theory, the bottleneck reduced the mean (co)variance within lines and caused stochastic, line-specific changes in the genetic (co)variance structure. However, a significant portion of the (co)variance structure was conserved, and 2 characters-leaf and flower (sepal) size-turned out to be positively correlated in all data sets, indicating a potential for correlated evolution in these characters, even after a severe bottleneck. The hierarchical partitioning of genetic variance for flower size was in good agreement with predictions from additive theory, whereas the remaining characters showed an excess of within-line variance and a deficiency of among-line variance. The latter discrepancies were most likely a result of selection, given the small proportion of lines (23%) that remained viable until the end of the experiment. Our results suggest that bottlenecked populations of N. degenii generally have a lower adaptive potential than the ancestral population but also highlight the idiosyncratic nature of bottleneck effects.