Induction of an additional floor plate in the neural tube

The role of the notochord in the morphogenesis of the neural tube was investigated by implanting a notochord fragment laterally to the neural wall of a 1.5 day chick embryo. Embryos were sacrificed at 4 days. In the basal part of the neural tube an additional floor plate was induced in the vicinity of the implant. This floor plate was characterized by a low proliferative activity, a thin wall, spindle-like nuclei crowded peripherally and some neuroblast-like cells. It was either blending with the natural floor plate or separated from it, depending on the exact position of the implant. In the latter case neuroblasts were observed in between both floor plates. The additional floor plate was present only when the implanted notochord was less than 25 micron apart from the neural tube; at larger distance an increase of the ventral horn neuroblast area could be seen. It is concluded that the implanted notochord is able to induce a floor plate at 1.5 days of incubation. The specific influence of the notochord on the morphogenesis of the neural tube, its inductive period as well as the presence of the neuroblast-like cells in the additional floor plate are discussed.     

Notochord-derived Shh concentrates in close association with the apically positioned basal body in neural target cells and forms a dynamic gradient during neural patterning

Sonic hedgehog (Shh) ligand secreted by the notochord induces distinct ventral cell identities in the adjacent neural tube by a concentration-dependent mechanism. To study this process, we genetically engineered mice that produce bioactive, fluorescently labeled Shh from the endogenous locus. We show that Shh ligand concentrates in close association with the apically positioned basal body of neural target cells, forming a dynamic, punctate gradient in the ventral neural tube. Both ligand lipidation and target field response influence the gradient profile, but not the ability of Shh to concentrate around the basal body. Further, subcellular analysis suggests that Shh from the notochord might traffic into the neural target field by means of an apical-to-basal-oriented microtubule scaffold. This study, in which we directly observe, measure, localize and modify notochord-derived Shh ligand in the context of neural patterning, provides several new insights into mechanisms of Shh morphogen action.     

Development of the notochord in normal and malformed human embryos and fetuses.

In view of its possible involvement in early embryogenesis and teratogenesis, the developmental characteristics of the human notochord were studied by light and electron microscopy and immunohistochemistry on 20 human conceptuses (5th-22nd week). At the earliest embryonic stages examined, the notochord is closely related to both the pharyngeal endoderm and the neuroectoderm of the posterior (tail) end of the neural tube. In both regions the interspace is bridged by slender cytoplasmic processes, lined with basal lamina and filled with amorphous extracellular material containing collagen types III and IV and laminin. The notochordal cells express cytokeratin brightly and vimentin weakly. As embryonic age progresses, the notochord gradually separates from the epithelia, becomes the axis of developing spinal column and undergoes progressive cell degeneration and rearrangement within the vertebral bodies. This is associated with extensive production of extracellular material and the first appearance of fibronectin. Intracellularly, the expression of vimentin gradually increases, while that of cytokeratin slightly weakens. Changes in the notochord parallel other developmental events in axial organs. In anencephalic fetuses the course of the notochord is irregular and partly interrupted with segments outside the basichondrocranium in specimens with craniorachischisis.     

Mosaic analysis with oep mutant reveals a repressive interaction between floor-plate and non-floor-plate mutant cells in the zebrafish neural tube

The floor plate is located at the ventral midline of the neural tube in vertebrates. Floor-plate development is severely impaired in zebrafish one-eyed pinhead (oep) mutants. oep encodes a membrane-bound protein with an epiblast growth factor (EGF) motif and functions autonomously in floor-plate precursors. To understand the cell behavior and cell-cell interaction during floor-plate development, the distribution and gene expression of wild-type and oep mutant cells in genetic mosaics were examined. When mutant shield cells were transplanted into a wild-type host, an ectopic neural tube with a floor plate was induced. However, the floor plate of the secondary axis was consistently devoid of mutant cells while its notochord was composed entirely of mutant cells. This indicates that oep shield cells adopt only a notochord fate in a wild-type environment. In reciprocal transplants (wild to oep), however, grafted shield cells frequently contributed to part of the floor-plate region of the secondary neural tube and expressed floor-plate markers. Careful examination of serial sections revealed that a mutant neural cell, when located next to the wild-type cells at the ventral midline, inhibited floor-plate differentiation of the adjacent wild-type cells. This inhibition was effective over an area only one- or two-cells wide along the anteroposterior axis. As the cells located at the ventral midline of the oep neural tube are thought to possess a neural character, similar to those located on either side of the floor plate in a wild-type embryo, this inhibition may play an important role during normal development in restricting the floor-plate region into the ventral-most midline by antagonizing homeogenetic signals from the floor-plate cells.     

Behavior of neural crest cells on embryonic basal laminae

Neural crest cells separate from the neural epithelium in a region devoid of a basal lamina and migrate along pathways bordered by intact basal laminae. The distribution of basal laminae suggests that they might have an important role in the morphogenesis of the neural crest by acting as a barrier to migration. The experiments reported here have tested directly whether neural crest cells can penetrate a basal lamina. Isolated neural tubes, neural crest cells cultured for 24 hr, or pigmented neural crest cells were explanted onto human placental amnions from which the epithelium had been removed to expose the basal lamina. In no case did neural crest cells or crest derivatives penetrate the basal lamina to invade the underlying stroma. If crest cells were grown on the stromal side of the amnion, they invaded the connective tissue. Pigmented neural crest derivative and [3H]thymidine-labeled nonpigmented crest cells were also confronted with chick embryonic basal laminae by grafting the cells into the lumen of the neural tube at the axial levels where host crest migration had commenced. Most of the grafted cells invaded the neural epithelium and accumulated after 24 hr at the basal surface of the neural tube. A few crest cells escaped through the dorsal surface of the neural tube and entered the overlying ectoderm, presumably through the wound created during the grafting procedure. Some of these grafted cells, located initially by light microscopy, were examined at the higher magnification and resolution offered by the transmission electron microscope to determine the relationship of the grafted cells to the basal lamina. In 50% (14 total) of the cases, the crest cells never reached the basal lamina of the neural tube, but were trapped by cell junctions between the neural epithelial cells. Of the remaining grafted cells that were relocated in the TEM (50%, total 15) all were spread on the basal lamina, but were not seen penetrating it. Likewise, in the three cases where crest cells were found in the epidermal ectoderm, all were in contact with the basal lamina of the ectoderm but did not have any processes extending through it. In three cases, at the level of the light microscope, crest cells were found to extend through the basal surface of the neural tube. In all these instances, the cells followed the dorsal root nerve exiting through a region of the neural tube that is devoid of a basal lamina.(ABSTRACT TRUNCATED AT 400 WORDS)     

Early Stages of Notochord and Floor Plate Development in the Chick Embryo Defined by Normal and Induced Expression of HNF-3β

We have cloned a cDNA encoding the chick HNF-3β gene and have used RNA and antibody probes that detect HNF-3β to monitor the normal and induced expression of the gene in early embryos. HNF-3β is expressed in Koller's sickle, at the onset of primitive streak formation, and later in Hensen's node. At neural plate and neural tube stages, HNF-3 β is expressed transiently in the notochord and is then expressed by floor plate cells. Prospective floor plate cells that are located in the epiblast immediately anterior to Hensen's node prior to its regression do not express HNF-3β, providing evidence that floor plate fate is normally determined only after these cells populate the midline of the neural plate and overlie the notechord. Removal of the notochord in vivo prevents floor plate development and in this condition HNF-3β is not expressed by cells at the ventral midline of the neural tube. Notochord grafts induce ectopic floor plate development and ectopic neural expression of HNF-3 β. In vitro, neural plate explants are induced to express HNF-3β by notochord cells in a contact-dependent but cycloheximide-resistant manner, providing evidence that expression of HNF-3 β is a direct response of neural plate cells to notochord-derived inducing signals.     

Experimental analysis of the mechanisms of somite morphogenesis

Earlier studies have suggested influences on somite morphogenesis by “somite-forming centers,” primitive streak regression, Hensen's node and notochord, and neural plate. Contradictions among these studies were unresolved.Our experiments resolve these conflicts and reveal roles of the primitive streak and notochord in shearing the prospective somite mesoderm into right and left halves and releasing somite-forming capabilities already present. The neural plate appears to be the principal inductor of somites.Embryo fragments containing no somite-forming centers, node, notochord, or streak nevertheless formed somites within 10 hr. Such somites disperse within the next 14–24 hr, which may explain why others failed to see them. In these fragments, an incision alongside the streak substitutes for streak regression in releasing somite formation. All such somites form simultaneously rather than in the normal anteroposterior progression. These fragments contain neural plate, but not notochord. We believe that physical attachment of somites to notochord in normal embryos stabilizes them and prevents dispersal.Pieces of epiblast were rotated 180° putting neural plate over lateral plate mesoderm regions. Somites were induced from the lateral plate by the displaced neural plate region. This is additional evidence of the powerful ability of neuroepithelium to induce somites.     

Differentiation of the embryonic disc, amnion, and yolk sac in the rhesus monkey

The inner cell mass of the blastocyst has differentiated into epiblast and hypoblast (primitive endoderm) prior to implantation. Since endoderm cells extend beyond the epiblast, it can be considered that both parietal and visceral endoderm are present. At implantation, epiblast cells begin to show marked evidence of polarity. They form a spherical aggregate with their basal ends toward the basal lamina and apical ends toward the interior. The potential for an internal space is formed by this change in polarity of the cells. No cytological evidence of separation of those cells that will form amniotic epithelium from the rest of the epiblast is seen until a cavity begins to form. The amniotic epithelium is originally contiguous with overlying cytotrophoblast, and a diverticulum remains in this position during early development. Epiblast forms a pseudostratified columnar epithelium, but dividing cells are situated toward the amniotic cavity rather than basally. The first evidence of a trilaminar disc occurs when a strand of cells contiguous with epiblast is found extending toward visceral endoderm. These presumptive mesoderm cells are undifferentiated, whereas extraembryonic mesoderm cells are already a distinct population forming extracellular materials. After implantation, visceral endoderm cells proliferate forming an irregular layer one to three cells thick. Visceral endoderm cells have smooth apical surfaces, but very irregular basal surfaces, and no basal lamina. At the margins of the disc, visceral endoderm is continuous with parietal endoderm and reflects back over the apices of the marginal visceral endoderm cells. This sacculation by visceral endoderm cells precedes pinching off of the secondary yolk sac from the remaining primary yolk sac.     

The notochordal sheath in amphioxus--an ultrastructural and histochemical study

The notochord and notochordal sheath of 10 adult amphioxus were investigated ultrastructurally and histochemically. The notochord in amphioxus consists of parallel notochordal cells (plates) and each plate consists of parallel thicker and thinner fibrils and numerous profiles of smooth endoplasmic reticulum situated just beneath the cell membrane. Histochemical staining shows that the notochordal plates resemble neither the connective tissue notochordal sheath nor the typical muscular structure myotomes. The notochordal sheath has a complex three-layered organization with the outer, middle and inner layer The outer and middle layer are composed of collagen fibers of different thickness and course, that correspond to collagen type I and collagen type III in vertebrates, respectively, and the inner layer is amorphous, resembles basal lamina, and is closely attached to the notochord by hemidesmosome junctions. These results confirm the presence of collagen fibers and absence of elastic fibers in amphioxus.     

Development of the basal lamina and extracellular materials in the early chick embryo

Summary Chick embryos at developmental stages up to primitive streak formation were fixed in a mixture of tannic acid and glutaraldehyde. A basal lamina was present in the unincubated embryo and consisted of a lucent lamina interna and a lamina densa. At the primitive streak stage the lamina densa showed a periodicity of stained elements. Densely stained materials were present on the cell surfaces lining the cavity between the epiblast and endoblast, and on the mesoderm cells within this cavity. Considerable amounts of extracellular material were observed in the cavity. Hyaluronidase treatment removed the cell surface and extracellular material, indicating that hyaluronic acid is a major component. This enzyme disrupted the basal lamina, leaving a fibrillar remnant with no periodic structure. It is therefore suggested that the dense periodicities consist of glycosaminoglycan built on an enzyme-resistant framework which is probably collagen. Enzyme-resistant fibrils, presumably collagen precursors, are present elsewhere within the tissue spaces.     

Effect of the notochord on proliferation and differentiation in the neural tube of the chick embryo

After implantation of a notochord fragment lateral to the neural tube in a 2-day chick embryo, at 4 days the ipsilateral neural tube half was increased in size and axons left the neural tube in a broad dorsoventral area (van Straaten et al. 1985). This enlargement appears to coincide with an increased area of AChE-positive basal plate neuroblasts, as determined with scan-cytophotometry. The effect was ipsilateral and local: clear effects were seen only when the implant was localized less than 80 microns from the neural tube and over 120 microns from the ventral notochord. In order to investigate the expected enhancement of proliferation, the mitotic density and the number of cells at the site of the implant at 3 days was determined and the mitotic index calculated. All three parameters showed an increase. It was concluded that the cell cycle was shorter in the implant area relative to the control area, at least during the third day. At 4 days the number of cells was still increased, predominantly in the basal plate. It appeared that the numerical increase was for the larger part due to neuroblasts. The synergism of two notochords thus resulted in enhancement of proliferation and differentiation in the neural tube. It is suggested that the notochord merely regulates and arranges the surrounding sclerenchymal cells, which are the effective cells in the regulation of neural tube development.     

Consequences of neural tube and notochord excision on the development of the peripheral nervous system in the chick embryo

Notochordectomy and neuralectomy were carried out either in one- or in two-step experiments on the chick embryo. The aim of this operation was to study the influence of the axial organs (notochord and neural tube) on the development of the ganglia of the peripheral nervous system. The neural crest cells from which most peripheral ganglion cells arise were labeled through the quail-chick marker system and their fate was followed under various experimental conditions. It appeared that the development of the dorsal root and sympathetic ganglia depends on survival and differentiation of somite-derived structures. In the absence of neural tube and notochord, somitic cells die rapidly, and so do the neural crest cells that are present in the somitic mesenchyme at that time. In contrast, those crest cells which can reach the mesenchymal wall of the aorta, the suprarenal glands, or the gut survive and develop normally into nerve and paraganglion cells. Differentiation of the neural crest- and placode-derived sensory ganglia of the head which develop in the cephalic mesenchyme is not affected by removal of notochord and encephalic vesicles. These results show that the peripheral ganglia are differentially sensitive to the presence of the neural tube and the notochord. Among the various ganglia of the peripheral nervous system, spinal and sympathetic ganglia are the only ones which require the presence of these axial structures. The neural tube allows both the spinal and the sympathetic ganglia to develop in the absence of the notochord. In contrast, if the notochord is left in situ and the neural tube removed, the spinal ganglia fail to differentiate and only sympathetic ganglia can develop.     

Retention and structure of extracellular matrix in early chick embryos after quick-freezing and freeze-substitution

Early chick embryos, stages 11 to 14, were isolated, quick-frozen by immersion in isopentane/propane cryogen (-185 degrees C) and freeze-substituted for study by scanning electron microscopy. Emphasis was placed on the extracellular matrix (ECM) in the axial region of the segmental plate and developing somites. Ultrarapid freezing, followed by delicate freeze-substitution, immobilizes and retains much more ECM than chemical fixatives that include tannic acid (TA). The matrix on the dorsal surface of the neural tube is preserved as delicate filaments which are expressed bilaterally over the tube in a dorso-ventral orientation. These parallel primary ridges of ECM have a spacing of 1 to 3 micron, forming grooves on the wall of the neural tube. Interrupting this pattern are funnel-shaped ridges about 80 to 100 micron apart along the neural tube. The ridges become decorated with cross-bridges creating a dense lattice in the region of somite development, to the extent that a basal lamina composed of dense fibrillar network and amorphous mats of matrix accumulates on the lateral wall of the neural tube. Heavy strands and fenestrated lamellae of ECM interconnect the neural tube, notochord and somites, and attach the overlying epithelium to the upper surface of the somites. The pattern of ECM is complimentary to the migratory pathways of ventrally migrating neural crest cells and is the basis for suggesting that a physical substratum influencing the direction of neural crest cell migration is an idea that should be revived.     

Epiblast origin and early migration of neural crest cells in the chick embryo

Chick embryos carrying transplants labeled with tritiated thymidine demonstrate that the neural crest originates in the anterior epiblast, at the junction of areas destined for epidermis and neural tube. As the neural tube begins to fold and the axis lengthens, cells along this junction are drawn dorsomedially; at the seven-somite stage they begin to separate from the epithelium of the head, and migrate into the angle between the epidermis and the neural tube. The paraxial mesoderm already populating this angle originates in more posterior and medial portions of the epiblast than do the neural crest cells; after invagination at the primitive streak, it migrates anterolaterally, ventral to the ectoderm layer, until it too is folded dorsomedially into the angle between the epidermis and the neural tube.     

Tail morphogenesis in the ascidian, Ciona intestinalis, requires cooperation between notochord and muscle

We present evidence that notochord and muscle differentiation are crucial for morphogenesis of the ascidian tail. We developed a novel approach for embryological manipulation of the developing larval tissues using a simple method to introduce DNA into Ciona intestinalis and the several available tissue-specific promoters. With such promoters, we misexpressed the Xenopus homeobox gene bix in notochord or muscle of Ciona embryos as a means of interfering with development of these tissues. Ciona embryos expressing bix in the notochord from the 64-cell stage develop into larvae with very short tails, in which the notochord precursors fail to intercalate and differentiate. Larvae with mosaic expression of bix have intermediate phenotypes, in which a partial notochord is formed by the precursor cells that did not receive the transgene while the precursors that express the transgene cluster together and fail to undergo any of the cell-shape changes associated with notochord differentiation. Muscle cells adjacent to differentiated notochord cells are properly patterned, while those next to the notochord precursor cells transformed by bix exhibit various patterning defects. In these embryos, the neural tube extends in the tail to form a nerve cord, while the endodermal strand fails to enter the tail region. Similarly, expression of bix in muscle progenitors impairs differentiation of muscle cells, and as a result, notochord cells fail to undergo normal extension movements. Hence, these larvae have a shorter tail, due to a block in the elongation of the notochord. Taken together, these observations suggest that tail formation in ascidian larvae requires not only signaling from notochord to muscle cells, but also a "retrograde" signal from muscle cells to notochord.     

The effect of tenascin and embryonic basal lamina on the behavior and morphology of neural crest cells in vitro

We have investigated the morphology and migratory behavior of quail neural crest cells on isolated embryonic basal laminae or substrata coated with fibronectin or tenascin. Each of these substrata have been implicated in directing neural crest cell migration in situ. We also observed the altered behavior of cells in response to the addition of tenascin to the culture medium independent of its effect as a migratory substratum. On tenascin-coated substrata, the rate of neural crest cell migration from neural tube explants was significantly greater than on uncoated tissue culture plastic, on fibronectin-coated plastic, or on basal lamina isolated from embryonic chick retinae. Neural crest cells on tenascin were rounded and lacked lamellipodia, in contrast to the flattened cells seen on basal lamina and fibronectin-coated plastic. In contrast, when tenascin was added to the culture medium of neural crest cells migrating on isolated basal lamina, a significant reduction in the rate of cell migration was observed. To study the nature of this effect, we used human melanoma cells, which have a number of characteristics in common with quail neural crest cells though they would be expected to have a distinct family of integrin receptors. A dose-dependent reduction in the rate of translocation was observed when tenascin was added to the culture medium of the human melanoma cell line plated on isolated basal laminae, indicating that the inhibitory effect of tenascin bound to the quail neural crest surface is probably not solely the result of competitive inhibition by tenascin for the integrin receptor. Our results show that tenascin can be used as a migratory substratum by avian neural crest cells and that tenascin as a substratum can stimulate neural crest cell migration, probably by permitting rapid detachment. Tenascin in the medium, on the other hand, inhibits both the migration rates and spreading of motile cells on basal lamina because it binds only the cell surface and not the underlying basal lamina. Cell surface-bound tenascin may decrease cell-substratum interactions and thus weaken the tractional forces generated by migrating cells. This is in contrast to the action of fibronectin, which when added to the medium stimulates cell migration by binding both to neural crest cells and the basal lamina, thus providing a bridge between the motile cells and the substratum.     

Expression of gamma-adducin is associated with regions of morphogenetic cell movement in the chick embryo

The adducin family of cytoskeletal proteins are known to mediate actin driven cell movements. Here we describe the cloning and expression pattern of a gene encoding gamma-adducin in the chick embryo. Expression of this gene is strikingly restricted to the epithelial cell layer (with a few exceptions including emerging notochord and lateral mesoderm). Gamma-adducin is detected at particularly high levels in cell populations undergoing important morphogenetic movements, such as epiblast approaching the primitive streak, regressing spinal cord primordia and closing neural tube.     

Tissue interactions affecting the migration and differentiation of neural crest cells in the chick embryo.

A series of microsurgical operations was performed in chick embryos to study the factors that control the polarity, position and differentiation of the sympathetic and dorsal root ganglion cells developing from the neural crest. The neural tube, with or without the notochord, was rotated by 180 degrees dorsoventrally to cause the neural crest cells to emerge ventrally. In some embryos, the notochord was ablated, and in others a second notochord was implanted. Sympathetic differentiation was assessed by catecholamine fluorescence after aldehyde fixation. Neural crest cells emerging from an inverted neural tube migrate in a ventral-to-dorsal direction through the sclerotome, where they become segmented by being restricted to the rostral half of each sclerotome. Both motor axons and neural crest cells avoid the notochord and the extracellular matrix that surrounds it, but motor axons appear also to be attracted to the notochord until they reach its immediate vicinity. The dorsal root ganglia always form adjacent to the neural tube and their dorsoventral orientation follows the direction of migration of the neural crest cells. Differentiation of catecholaminergic cells only occurs near the aorta/mesonephros and in addition requires the proximity of either the ventral neural tube (floor plate/ventral root region) or the notochord. Prior migration of presumptive catecholaminergic cells through the sclerotome, however, is neither required nor sufficient for their adrenergic differentiation.     

Light- and electron-microscopic observations of the tail bud of the larval lamprey (Lampetra japonica), with special reference to neural tube formation

Serial transverse and horizontal sections of the tail of the 26-day larval lamprey, Lampetra japonica, were observed by light and electron microscopy. The axial structures in the tail of the larval lamprey seem to differentiate from the prospective materials derived individually from the tail bud. The latter consists of two closely adjoined cell populations (C1 and C2). C1 is a small cell cluster located posterior to the other group (C2) and consists of loosely arranged polymorphous cells. The cell cluster extends cranially as a cell sheet on the ventral surface of C2; somites differentiate from this cell sheet. C2 is composed of cells elongated mediolaterally and stacked horizontally to form a compact cell mass which is covered on each lateral surface by a basal lamina. The upper one-third of C2 seems to differentiate into the neural tube, anteceding other axial structures. The middle one-third of C2 seems to develop into the notochord, and the lower one-third into the subchord and the undefined cell cord. The central canal develops in the upper one-third of C2 through the following events: 1) appearance of cilia and a small cavity between adjoining cells; 2) appearance of microvilli in the cavity, in addition to cilia; and 3) development of junctional complexes along the luminal borders of cells surrounding the cavity. Together with these events, cells surrounding the cavity increase in number, acquiring apicobasal polarity and radial arrangement. The cavity itself enlarges by incorporation of periciliary clefts and fusion of cavities with each other to establish the central canal. Near the caudal end of the neural tube, the central canal is directly confluent with the connective-tissue space through an opening in the dorsal wall of the neural tube.     

Origin of the dorsal surface of the neural tube by progressive delamination of epidermal ectoderm and neuroepithelium: implications for neurulation and neural tube defects

Knowledge of the morphogenetic events involved in the development of the dorsal portion of the neural tube is important for understanding neural tube closure, neural crest cell formation and emigration, and the origin of neural tube defects. Here, I characterize the progressive development of the tips of the neural folds during fold elevation in the trunk of mouse and chick embryos and the events leading to formation of the dorsal portion of the neural tube as the epidermal ectoderm (EE) and neuroepithelium (NE) separate from each other. The nature and timing of appearance of collagen IV, laminin and fibronectin were analysed by immunofluorescent and immunogold labelling, and ruthenium red and tannic acid were used to enhance staining for proteoglycans and glycosaminoglycans. As the neural folds elevate, the NE and EE delaminate progressively beginning at the basal surface of the lateral extremes of the neural plate. Nevertheless, the two epithelia remain connected across the zone of delamination by their previously existing basal laminae. In each fold, proteoglycan granules appear at the interface between the NE and EE before delamination begins, and then an (interepithelial) space begins to open and propagate dorsally. Other extracellular matrix (ECM) molecules appear within the space a short distance behind its tip and basal lamina deposition begins shortly thereafter. As fusion occurs, the interepithelial spaces of the two folds coalesce and the final separation of the EE from the NE is accomplished. These observations suggest that the previously recognized delay in deposition of ECM and basal lamina on the dorsal portion of the neural tube and on the overlying EE is a direct consequence of the delamination of the two epithelia and the establishment of two new basal surfaces. The observation that the surface of the dorsal third of the neural tube forms by delamination rather than by juxtaposition of previously existing basal surfaces of the two epithelial is discussed in terms of possible implications for models of neurulation and the origin of neural tube defects.