Many large, half-sib sire families are an integral component of chicken genetic improvement programs. These family structures include a sufficient number of individuals for mapping quantitative trait loci (QTL) at high statistical power. However, realizing this statistical power through individual or selective genotyping is yet too costly to be feasible under current genotyping methodologies. Genotyping costs can be greatly reduced through selective DNA pooling, involving densitometric estimates of marker allele frequencies in pooled DNA samples. When using dinucleotide microsatellite markers, however, such estimates are often confounded by overlapping "shadow" bands and can be confounded further by differential amplification of alleles. In the present study a shadow correction procedure provided accurate densitometric estimates of allele frequency for dinucleotide microsatellite markers in pools made from chicken purified DNA samples, fresh blood samples, and frozen-thawed blood samples. In a retrospective study, selective DNA pooling with thawed blood samples successfully identified two QTL previously shown by selective genotyping to affect resistance in chickens to Marek's disease. It is proposed that use of selective DNA pooling can provide relatively low-cost mapping and use in marker-assisted selection of QTL that affect production traits in chickens.