Search for a command to run...
Species-to-genus ratios increase with number of species in the sample, other things being equal. Comparisons of ratios have nevertheless been used in ecology and biogeography repeatedly without consideration of this statistical bias. This paper reviews European debates on the validity of species-togenus ratios. It was pointed out in the 1920s that species-to-genus ratios are not valid indicators of ecological diversity, and, further, that the ratios depend strongly on the numbers of species in the sample. Maillefer of Switzerland performed simulations in 1929 and showed that most reported values deviate little from the statistical expectation; and if there is a general trend, it is that congeners tend to coexist more often than expected on the basis of random sampling. This conclusion was opposite to previous conclusions emphasizing the importance of ecological diversity and interspecific competition. The mathematician Polya obtained an exact solution based on probability theory in 1930. These pioneering results did not penetrate to the biological community of the time because of their probabilistic nature. The prevailing paradigm in plant geography of the 1920s and the 1930s stressed the importance of physical and chemical factors in explaining plant distribution, and chance was not acceptable in the contemporary framework. The rediscovery of these papers has no doubt been delayed by the fact that they were published in obscure European journals in German and in