How cells acquire their fate is a fundamental question in both developmental and regenerative biology. Multipotent progenitors undergo gradual cell fate restriction in response to temporal and positional cues from the microenvironment, the nature of which is far from being clear. In the case of the lymphatic system, venous endothelial cells are thought to give rise to lymphatic vessels, through a process of trans-differentiation. Upon expression of a set of transcription factors, venous cells acquire a lymphatic fate, and bud out to generate the lymphatic vasculature. In this work we challenge this view and show that while lymphatic endothelial cells (LECs) do arise in the Cardinal Vein (CV), they do so from a previously uncharacterized pool of multipotent angioblasts. Using lymphatic-specific transgenic zebrafish, in combination with endothelial photoconvertible reporters, and long-term live imaging, we demonstrate that these multipotent angioblasts can generate not only lymphatic, but also arterious, and venous fates. We further reveal that the underlying endoderm serves as a source of Wnt5b, which acts as a lymphatic inductive signal, promoting the angioblast-to-lymphatic transition. Moreover, Wnt5b induced lymphatic specification in human embryonic stem cells- derived vascular progenitors, suggesting that this process is evolutionary conserved. Our results uncover a novel mechanism of lymphatic vessel formation, whereby multipotent angioblasts and not venous endothelial cells give rise to the lymphatic endothelium, and provide the first characterization of their inductive niche. More broadly, our findings highlight the CV as a plastic and heterogeneous structure containing different cell populations, analogous to the hematopoietic niche in the aortic floor. Overall design: Following Kaede photoconversion of dorsal or ventral halves of the PCV in Tg(fli1:gal4;uasKaede) embryos at 24 hpf, 6Â embryos per group were used for FACS isolation of Kaede photconverted (red) ECs.