Zwolinski (1992, 1993) provided equations and a method for finding nine parameter values for gene conversion between two alleles at one locus, to study mechanisms for homologous meiotic recombination. Simpler, shorter, and easier-to-use equations have been given here, and the accuracy of the parameter values found has been analysed. Fincham's (1994) criticisms of this method have been tested in several ways and shown to be incorrect. The model's validity has been tested directly: sets of arbitrarily chosen parameter values were used to work out the numbers expected in each segregation class for a sample of 100,000 octads. People not knowing the chosen parameter values used those octad class numbers in computer programs based on the Lamb-Zwolinski method. The best-fitting sets of parameter values obtained corresponded exactly to those initially set. The occurrence of multiple fits has been demonstrated and discussed; when multiple fits occurred, the parameter values in different 'best fits' were fairly similar in absolute values and very similar in relative values.