1 Introduction

In population genetics, there has long been an interest in understanding and analyzing the gene frequency structure found in natural populations. Local genetic drift will tend to create genetic differences between different subunits of a population. These differences will, after a number of initial generations, be balanced by migration. One can thus, in principle, estimate the amount of migration between the different subunits from the observed gene frequencies. Most earlier approaches to this problem have, in part, relied on the assumptions of the island model. Under that model one can estimate the number of migrants $Nm$ from the estimated amount of genetic differentiation as measured by e.g. Wright's parameter $F_{st}$. Typically, these estimates are then, in further exploratory analysis, regressed against geographic distances to test for isolation by distance effects predicted by some idealized theoretical models. This manual documents the use of some S-Plus and c code I wrote to implement an alternative solution of this problem based more directly on standard statistical principles, which allows much less restrictive assumptions about the underlying pattern of migration. The general method, which is based on an idea in Felsenstein (1982), is described in two papers (Tufto et al., 1996; Tufto and Hindar, 2002; Tufto et al., 1998); preprints of these are available on my list of publications. A brief summary of the approach is also given here.

Readers who prefer to read a printed version of this manual can download the .pdf version of the manual from this url. It is assumed that the reader has access to and some familiarity with R or S-plus.