Did Mendel falsify his data?

Abstract

In 1936 Professor R. A. Fisher (1890–1962) published a 26-page article1 criticizing Mendel's publication of 1866 entitled Versuche uber Pflazen-Hybriden 2 (Experiments in Plant-Hybridization). He accuses Mendel (1822–84) of falsifying his data. Fisher had worked through most of Mendel's results over the 8-year period of experiments to calculate the deviations of expected numbers of different progeny types from those actually observed, based on the relevant sample sizes. For example, he calculated from the F2 hybridization experiments of the year 1858 that 258 plants gave 8023 seeds of which 6022 were yellow compared with 2001 that were green. Fisher states that the deviation (that is 0.009) from the expected ratio of 3:1 is less than the standard error of random sampling; therefore there may have been manipulation of the data sets by someone. Nor did Mendel test the significance of the deviations from the expected ratio of 3:1 but states it as 3.009:1 without giving any probable error, either absolute or relative. Fisher points out that tests of significance of deviations from expectations in a binomial series were familiar to mathematicians since the middle of the 18th century. (Incidentally Fisher's article failed to comment on a paper by Tschermak3 of 1900 where the difference between observed and expected ratios was even smaller than Mendel's, the observed ratio being 3.008:1 giving a difference of 0.008 from expected4). As a result of Fisher's calculations, and without access to Mendel's laboratory notebooks, he directs a stream of pejorative remarks at Mendel's work. To give …