The transmission–virulence trade-off hypothesis is one of the few adaptive explanations of virulence evolution, and assumes that there is an overall positive correlation between parasite transmission and virulence. The shape of the transmission–virulence relationship predicts whether virulence should evolve toward either a maximum or to an intermediate optimum. A positive correlation between each of these traits and within-host growth is often suggested to underlie the relationship between virulence and transmission. There are few experimental tests of this hypothesis; this study reports on the first empirical test on a plant pathogen. We infected Brassica rapa plants with nine natural isolates of Cauliflower mosaic virus and then estimated three traits: transmission, virulence, and within-host viral accumulation. As predicted by the trade-off hypothesis, we observed a positive correlation between transmission and virulence, suggestive of the existence of an intermediate optimum. We discovered the unexpected existence of two groups of within-host accumulation, differing by at least an order of magnitude. When accumulation groups were not accounted for, within-host accumulation was correlated neither to virulence nor transmission, although our results suggest that within each group these correlations exist.