The existence of the hemangioblast, a common progenitor of the endothelial and hematopoietic cell lineages, was proposed at the beginning of the century. Although recent findings seem to confirm its existence, it is still unknown when and how the hemangioblasts differentiate. We propose a hypothesis about the origin of hemangioblasts from the embryonic splanchnic mesothelium. The model is based on observations collected from the literature and from our own studies. These observations include: (1) the extensive population of the splanchnic mesoderm by mesothelial-derived cells coinciding with the emergence of the endothelial and hematopoietic progenitors; (2) the transient localization of cytokeratin, the main mesothelial intermediate filament protein, in some embryonic vessels and endothelial progenitors; (3) the possible origin of cardiac vessels from epicardial-derived cells; (4) the origin of endocardial cells from the splanchnic mesoderm when this mesoderm is an epithelium; (5) the evidence that mesothelial cells migrate to the hemogenic areas of the dorsal aorta. (6) Biochemical and antigenic similarities between mesothelial and endothelial cells. We suggest that the endothelium-lined vascular system arose as a specialization of the phylogenetically older coelomic cavities. The origin of the hematopoietic cells might be related to the differentiation, reported in some invertebrates, of coelomocytes from the coelomic epithelium. Some types of coelomocytes react against microbial invasion and other types transport respiratory pigments. We propose that this phylogenetic origin is recapitulated in the vertebrate ontogeny and explains the differentiation of endothelial and blood cells from a common mesothelial-derived progenitor.