The technique of whole embryo culture has made significant contributions to understanding the mechanisms of morphogenesis in mammalian embryos, especially with respect to cranial neurulation and neural crest cell migration. This study traces the fate of two specifically mammalian structures, the preotic and otic sulci. Their formation at the 1/2- and 3-somite stages respectively, divides the hindbrain neuroepithelium into prorhombomeres A, B and C. The preotic sulcus is a deeply recessed structure that forms the rostral boundary of expression of both Hoxb-2 and the first domain of Krox-20. The otic sulcus is a shallow concavity in which the second Krox-20 domain is expressed. DiI labeling followed by whole embryo culture confirmed that the later fate of the preotic sulcus is the rhombomere 2/3 boundary, and the fate of the otic sulcus is the cranial part of rhombomere 5. Structurally, the preotic and otic sulci show no specialization with respect to actin, tubulin or proteoglycans, but their maintenance depends on contact with the subjacent mesenchyme. Their formation is inhibited by exposure of embryos to retinoic acid prior to the onset of somitic segmentation, indicating that the molecular events governing prorhombomeric subdivision of the hindbrain are retinoic acid-sensitive. The preotic sulcus may be essential for neuroepithelial cell movement towards and into the rapidly enlarging forebrain; the otic sulcus may simply delineate the caudal boundary of prorhombomere B, an area with a discrete neural crest cell population discontinuous with those rostral and caudal to it. Understanding the positional relationships of the preotic and otic sulci to later rhombomeric segments makes them useful landmarks for experimental purposes, but there is no evidence that prorhombomeres are functionally significant as the precursors of rhombomeric segments.