The test for the comparison of two probabilities as proposed in 1840 by {\it L. D. J. Gavarett} (1809-1890) is only applicable when large samples are available. In 1877 the professor of medicine Carl Liebermeister (1833-1901) published a test for small samples using a model of balls drawn from two urns. It is in principle similar but slightly different from {\it R. A. Fisher}’s exact test for $2 \times 2$-tables (1922). Liebermeister’s formula is correct, its application was too cumbersome for his colleagues and so it received no attention. In 1948, {\it D. J. Finney} computed comfortable tables for the evaluation of Fisher’s exact test, they were extended by {\it R. Latscha} (1953) and once more by {\it B. M. Bennett} and {\it P. Hsu} (1963). From a footnote by Liebermeister we learn, that always at that time he read lessons on probability and its application to medicine and science. 16 references.
Reviewer: H.Grimm (Jena)