Fig. 1. Two models for the action of Nodal signals in patterning the mesendoderm along the dorsoventral axis in zebrafish embryos. In each case, the gradient of shading at the margin represents the putative distribution of Nodal signals, with the darkest shade indicating the highest concentration. The fate maps shown in the second column are loosely based on those of Kimmel et al. (Kimmel et al., 1990). Mesendoderm is shaded red, green or blue, depending on the position along the dorsoventral axis, and regions generating both mesoderm and endoderm shaded darker than regions producing mesoderm alone. (A) High levels of Nodal signals specify dorsal mesendodermal fates, intermediate levels specify lateral mesendodermal fates and low levels determine ventral mesendoderm. In this model, Nodal signals act in a dorsal-to-ventral gradient to pattern the mesendoderm. The gradient shown here is only one of many such gradients that could be drawn consistent with the evidence. However, all versions of the gradient model predict that reductions in the level of Nodal function would result in the transformation of dorsal marginal cells to more ventrolateral fates (illustrated in the right-hand panels in A). (B) Nodal activity is uniformly distributed along the dorsoventral axis, but a gradient of Nodal signals along the animal-vegetal axis patterns the germ layers. Independent dorsalizing factors pattern the mesendoderm along the dorsoventral axis (represented by the red arrow). This model predicts that endodermal cells (darker colors near the margin) are transformed to more animal fates as levels of Nodal signals are reduced (right-hand panels in B). Dorsalizing factors remain to establish dorsoventral pattern. Ne, neuroectoderm; D, dorsal mesendoderm; L, lateral mesendoderm; V, ventral mesendoderm, Y, yolk.