Genome-wide chemical mutagenesis screens have been successfully performed in zebrafish to identify genes involved in erythropoiesis. We have now executed a large-scale screen for genes involved in red cell production in mice by using N-ethyl-N-nitrosurea mutagenesis. Five thousand G1 mice were screened by automated haematological analysis at 7 weeks of age for abnormalities of their red cell indices that lay more than 3 standard deviations outside the normal population average. Mice with co-incident abnormalities in white cell or platelet counts were excluded. Six mutant lines were identified with microcytosis (RBC1-6) that was heritable with subsequent backcrossing. Three of these lines (RBC 1, 4, 5) were found by gene mapping and candidate gene sequence analysis to harbour different mutations in the transferrin receptor gene (TfR) that led to embryonic lethality in their homozygous state. These mutations were clustered in the sequence encoding the domain involved in receptor homodimerisation, providing, for the first time, in vivo functional correlations of the previous structural analyses. The RBC3 line was also mapped to the interval containing the TfR gene, although no mutations were present in the TfR coding region or splice donor/acceptor sites in this line, raising the possibility of a regulatory sequence mutation, which we are investigating. The RBC2 line was mapped to chromosome 8, and found to contain a splice acceptor mutation in the ankyrin gene. This mutation results in a 22 bp deletion with a resultant frame shift in the regulatory domain. Mice homozygous for this mutation are born jaundiced with massive splenomegaly and a blood film consistent with severe spherocytosis, and die within 6 weeks of progressive anemia. As such, this line represents a new mouse model of severe hereditary spherocytosis. The final line, RBC6 displays a profound hypochromic, microcytic anemia in the homozygous state at birth, with apparent normality of all other organ systems. The mutation has been mapped to a chromosomal interval containing only a few genes, none of which have been previously linked to red cell production or differentiation, and therefore represents a novel erythropoietic gene/allele.