[i] The number of nonrecombinant and recombinant classes observed for each chromosome are listed along with the numbers of classes expected according to a Poisson distribution (in parentheses). Where the expected number is <1, no figure is given. χ² values from goodness-of-fit tests applied to each chromosome are given. All chromosomes differ very significantly from a Poisson distribution at p < 10⁻⁹. Although there appears to be a broad relationship between numbers of total recombinants per chromosome and genetic length (cM), the observed distribution deviates significantly from that expected if the number of recombinants varies in proportion to genetic length (χ² = 313.64, P < 10⁻¹⁰; 19df.). A binomial test of significance has also been applied to each chromosome to identify those chromosomes that depart most significantly from expected, and the χ²values for each chromosome are listed. We have taken account of the likelihood of false positives, i.e., chromosomes that depart from expected by chance. Monte Carlo simulations indicate that when adopting a significance level of 0.05, one false-positive chromosome will arise with a frequency of 0.6317. Reducing the false-positive rate to 0.05 the binomial test needs to be applied at a significance level of 0.002. At α = 0.002 chromosomes 1, 3, 7, 8, 11, 15, 16, 17 and 18 show a significant departure from expected (indicated by asterisk).