Two independent signals are necessary for neural crest (NC) induction in Xenopus: a Bmp signal, which must be partially attenuated by Bmp antagonists, and a separate signal mediated by either a canonical Wnt or an Fgf. The mesoderm underlying the NC-forming region has been proposed as a source of this second signal. Wnt8 and Fgf8a are expressed in this tissue around the time of NC induction and are therefore good candidate NC inducers. Loss-of-function studies indicate that both of these ligands are necessary to specify the NC; however, it is unclear whether these signaling molecules are operating in the same or in parallel pathways to generate the NC. Here, we describe experiments addressing this outstanding question. We show that although Wnt8 expression can restore NC progenitors in Fgf8a-deficient embryos, Fgf8a is unable to rescue NC formation in Wnt8-depleted embryos. Moreover, the NC-inducing activity of Fgf8a in neuralized explants is strongly repressed by co-injection of a Wnt8 or a beta-catenin morpholino, suggesting that the activity of these two signaling molecules is linked. Consistent with these observations, Fgf8a is a potent inducer of Wnt8 in both whole embryos and animal explants, and Fgf8a knockdown results in a dramatic loss of Wnt8 expression in the mesoderm. We propose that Fgf8a induces NC indirectly through the activation of Wnt8 in the paraxial mesoderm, which in turn promotes NC formation in the overlying ectoderm primed by Bmp antagonists.