A comparison of haplotype frequencies between unrelated cases and controls from population isolates identifies a strong signal for the presence/absence of the discrete phenotype in an interval on chromosome 6 bounded by markers 34 and 35. We define a dissimilarity index, D, which is sensitive to differences in allele distributions and differences in the patterns of linkage disequilibrium between cases and controls. We describe two statistical methods to utilize D: a method appropriate for a single moderately sized sample and a sequential approach appropriate for multiple small independent samples.