A recent re-analysis of the data of Salisbury (1974) claims his data do not support the hypothesis that seeds of species from shaded habitats are heavier than those from unshaded habitats, partly because the original analysis was inappropriate and partly because of bias in the dataset. We show first that the re-analysis itself contains errors, and second that the charge of bias is based largely on a misunderstanding. We also show that analysis of a larger dataset, drawn from Salisbury (1942) and from Salisbury (1974), provides convincing support for the hypothesis and suggests that the relationship is independent of life history.