Statistical analysis of mixed fossil populations

This paper outlines a procedure to analyze statistically a mixture of two populations that cannot be Separated by usual methods. In such instances, frequency histograms of taxonomically important characters that are normally distributed show the presence of two modes revealing the possible existence of two populations. As a first step in the analysis, the joint distribution of a suitable character is resolved by one of the two, Pearson or Rao, methods of moments; this gives estimates of the mean and standard deviation of that character for each of the two constituent populations. A numerical example explains the computational procedure. Rao's method is easier to use but is applicable only if the two standard deviations are assumed equal. If the distribution of the characters have less than 10-percent overlap, it is possible to make further statistical analysis on the basis of results obtained by either of the two methods. In paleobiometrical work the problem of having a mixed population is fairly common. The procedure outlined in this paper may be utilized in such situations for a better evaluation and interpretation of data.