The Concept of Path Coefficient and Its Impact on Population Genetics

The method of path coefficients was first published by Professor Sewall Wright thirty-five years ago. In 1921 there appeared in the Journal of Agricultural Research (Vol. 20) a general account of the method and the relationship between correlation and path coefficients, together with some examples of application; and in Genetics (Vol. 6) a series of five papers dealing exclusively with the application of path coefficients to genetic problems. Previously known results of various mating systems, obtained by laborious arithmetical procedures, were confirmed by the more elegant method, and many new results were reached, some of which were later corroborated by the method of matrix algebra while others are still difficult to obtain by any other method today. These classical papers, together with the pioneer work of Fisher (1918), still constitute the basic readings for students of population genetics, although the method has since taken a more sophisticated form and the field of application has been widened. However, one must admit that the method of path coefficients, as powerful and flexible as it is, was not immediately very popular among geneticists, still less so among professional statisticians. It was much later that its usefulness became gradually and generally appreciated. Path coefficients can be treated at various mathematical levels. The most important properties, however, can be deduced and studied by standard statistical tools. To understand the method requires little more than a knowledge of multiple and partial regression and correlation. It is a special type of multi-variate analysis a method of dealing with a closed system of variables that are linearly related. (For nonlinearly related variables, an appropriate transformation of scales may