Recent advances in the fields of molecular cloning and peptide purification necessitate a reappraisal of our views concerning the evolution of the genes encoding somatostatin-related peptides. The currently widely held view that the genomes of tetrapods contain only the preprosomatostatin-I (PSS-I) gene, encoding somatostatin-14, with a second preprosomatostatin gene being expressed only in teleost fish is no longer tenable. Identification of genes encoding both somatostatin-14 and the somatostatin-related peptide, cortistatin in mammals, identification of the PSS-I and PSS-II preprosomatostatin genes in amphibia, and the isolation of gene products from at least two non-allelic preprosomatostatin genes in lampreys suggests the alternative hypothesis that duplication of the PSS-I gene occurred early in evolution, predating or concomitant with the appearance of the chordates. We speculate that at least two somatostatin genes are expressed in all classes of vertebrates but these genes have evolved at very different rates. It is probable that the preprosomatostatin-II (PSS-II) gene, encoding [Tyr7, Gly10] somatostatin-14 or a related peptide, arose from a second independent duplication of the PSS-I gene in the ancestor of present-day teleost fish at a time after the divergence of the teleost stock from the line of evolution leading to tetrapods. The recent isolation of urotensin II, a peptide which contains a region of structural similarity but is not evolutionarily related to somatostatin-14, from the central nervous systems of lampreys, elasmobranchs and amphibia necessitates that we modify the accepted view that urotensin II is exclusively a product of the caudal neurosecretory system of teleost fish.