Fisher’s randomization test and Darwin’s data – A footnote to the history of statistics

In the presentation of his randomization test for paired data, Fisher used Darwin’s data on the relative growth rates of cross- and self-fertilized corn to motivate the development. On reading Darwin’s description of his experiment, it appears clear that the experiment did not use true paired comparisons. Although the statistical foundation of Fisher’s randomization test is sound, it is of historical interest that it does not suit the design of the motivating experiment.