The use of molecular polymorphisms identifiable at the level of DNA, such as restriction fragment length polymorphisms (RFLPs) and minior microsatellites (Love et al. 1990; Hearne et al. 1991; Montagutelli et al. 1991) together with the possibility to breed interspecific hybrids, has greatly contributed to the devel.~ opment of the mouse genetic map. Currently, well aZ over 1000 markers have been assigned to a specific o locus, and a map with an average resolution at the level of 0.3-0.5 cM has become a reasonable perspec>, tive for the near future. However, the chromosomal localization of new a, mutations, identified only by their behavior or external phenotype, and the localization of the genes involved in the determinism of complex phenotypes, remain a problem because, by definition, they require that specific crosses be set up and specific markers be used for each mutation or genetic trait to be mapped. Strategies based on the use of a limited number of DNA probes each identifying moderately repeated genomic sequences, such as ecotropic leukemia proviruses (Taylor and Rowe 1989) or the set of pseudogenes characteristic of the ornithine decarboxylase gene (Odc; Siracusa et al. 1991) or of the triose phosphate isomerase gene (Tpi), facilitate chromosomal assignments but do not completely resolve the problem, since only about 70% of the mouse genetic map is covered with these markers. In addition, most of the " allelic DNA sequences revealed by those probes ideno tifying several copies have first to be localized, with ~ respect to a bona fide anchor locus on a particular o chromosome, before they can be used. In the same way, the discovery of an increasing .~ number of microsatellites will certainly contribute in the future to make this mapping both more efficient and less expensive, but, again, some regions of the mouse genome are not covered by such markers. In this paper we report the isolation of a set of