Recent work has identified a growing body of evidence that subtle changes in noncoding sequences can result in significant pathology. These mutations, which would have been called silent polymorphisms in the past, affect gene transcription and mRNA splicing and lead to drastic changes in gene expression. Previous work from our lab has characterized the murine flexed-tail (f) mutation, which encodes Smad5, a transcription factor that functions downstream of the receptors for bone morphogenetic proteins (BMPs). f/f mice are unable to rapidly respond to acute anemia. Our analysis of these mice led to the development of a new model for stress erythropoiesis, where BMP4 expression in the spleen leads to the Smad5-dependent expansion of a specialized population of stress erythroid progenitors during the recovery from acute anemia. f/f mutant mice exhibit a defect in Smad5 mRNA splicing in the spleen such that the majority of Smad5 transcripts are two misspliced mRNAs. One of these mRNAs encodes a truncated form of Smad5 that inhibits BMP4 signaling when overexpressed. Here we show that a mutation in a poly(T) element in intron 4 causes the splicing defect in f/f mutant mice. This subtle mutation (loss of 1 or 2 Ts in a 16-T element) results in defects in splicing throughout the Smad5 gene. Furthermore, we show that this mutation results in tissue-specific splicing defects, which may explain why f/f mice are viable when Smad5-/- mice are embryonic lethal.