1. A general model is presented to analyse size-dependent reproductive output in plants. The model formulation makes it possible to test for both a minimum size for reproduction and a non-linear relationship between size and reproductive output. 2. The model is statistically analysed with the likelihood-ratio test. In contrast to previously proposed models, this model admits a hump-shaped relationship between reproductive effort (RE) and plant size. 3. Data for three species analysed here show that RE is not always monotonically increasing or decreasing with plant size. 4. For a realistic data set the statistically required sample size is calculated.