Although the flower is the unit of sexual reproduction in angiosperms, a plant’s reproductive success is determined by its entire floral display. A trade-off between the size and number of flowers produced is an assumption of models of floral display evolution, but this trade-off is often not observed. We tested three hypotheses for why a trade-off between flower size and number is not observed, using 32 populations of Mimulus guttatus and 83 genera from the California flora. We found support for the hypothesis that high variance in resource acquisition masks a trade-off between flower size and number. In contrast, we did not find support for the hypothesis that trade-offs between current and future reproduction mask negative correlations between flower size and number in lineages with a perennial life history. We also did not find support for the hypothesis that negative flower size-number correlations are less likely to be observed in lineages with lower variation in flower number because the correlation between two variables is weaker when the range of values is narrower. Overall, our results support the life history evolution literature in suggesting that high variance in resource acquisition relative to allocation can often explain why traits that are predicted to trade off are instead positively correlated.