Cytological features of serotonin-containing neurons and their processes in the retina of the carp (Cyprinus carpio)

Summary The morphological features of serotonin-containing neurons (SCNs) and their processes in the retina of the carp (Cyprinus carpio) were immunohistochemically studied by applying a modified peroxidase-antiperoxidase technique using flat-mount preparations. The somata of immunoreactive SCNs were mostly located in the innermost part of the inner nuclear layer (INL). These cells were distributed at a density of 64.8 cells/mm², this being similar to the density of dopamine-containing interplexiform cells. The processes of the SCNs ramified successively into two finer branches, eventually, forming a broad, extensive network in the thin layer just subjacent to the plane of the somata of the SCNs. Processes originating from neighboring SCNs exhibited cytoplasmatic continuity with each other at the lightmicroscope level. Due to their location and cytological features these SCNs appeared to correspond to amacrine cells.In honour of Prof. P. van Duijn     

Light- and electron-microscopic study of substance P-immunoreactive neurons in the guinea pig retina

Substance P (SP) immunoreactivity in the guinea pig retina was studied by light and electron microscopy. The morphology and distribution of SP-immunoreactive neurons was defined by light microscopy. The SP-immunoreactive neurons formed one population of amacrine cells whose cell bodies were located in the proximal row of the inner nuclear layer. A single dendrite emerged from each soma and descended through the inner plexiform layer toward the ganglion cell layer. SP-immunoreactive processes ramified mainly in strata 4 and 5 of the inner plexiform layer. SP-immunoreactive amacrine cells were present at a higher density in the central region around the optic nerve head and at a lower density in the peripheral region of the retina. The synaptic connectivity of SP-immunoreactive amacrine cells was identified by electron microscopy. SP-labeled amacrine cell processes received synaptic inputs from other amacrine cell processes in all strata of the inner plexiform layer and from bipolar cell axon terminals in sublamina b of the same layer. The most frequent postsynaptic targets of SP-immunoreactive amacrine cells were the somata of ganglion cells and their dendrites in sublamina b of the inner plexiform layer. Amacrine cell processes were also postsynaptic to SP-immunoreactive neurons in this sublamina. No synaptic outputs onto the bipolar cells were observed.     

NADPH-diaphorase reactivity in adult and developing cat retinae

Summary We have examined the distribution and size of nicotinamide adenine dinucleotide phosphate (NADPH) diaphorase reactivity in adult and developing cat retinae. From late gestation E (embryonic day) 58 to adulthood, NADPH-diaphorase reactivity was detected in amacrine cells with somata located in the inner nuclear layer (INL) and ganglion cell layer (GCL) and in processes spreading in the middle strata of the inner plexiform layer (IPL). Reactivity was also present in small rounded profiles located in the outer plexiform layer (OPL) and thought to be cone pedicles. The number of NADPH-diaphorase reactive cells present in adult retinae was about 40 000; 75% of these somata were located in the GCL, the remainder in the INL. At birth, however, there was more than double this number of labelled somata (85 000), the total gradually declining to reach adult values by P (postnatal day) 25. This loss of NADPH-diaphorase reactive somata may be partly explained by natural cell death (apoptosis) or by loss of the active diaphorase from the cells. The density distributions of NADPH-diaphorase reactive cells in the INL and GCL of retinal wholemounts reached maxima in regions slightly inferior to the area centralis at all ages studied. The principal topographical difference between adult and developing retinae was that the density gradient of NADPH-diaphorase reactive cells was steeper in adults than at younger ages. During early development, the somal and dendritic field diameters of NADPH-diaphorase reactive cells at the area centralis were about the same size as those in the periphery; by adulthood, cells in the periphery were larger. The change in the somal diameter gradient apparently emerged because of a reduction in somal size of the centrally located cells. The change in the dendritic diameter gradient emerged because of a greater growth of peripheral cells as compared to central cells. We suggest that NADPH-diaphorase may have a role in the formation of synapses in the developing IPL.     

Morphology and distribution of Müller cells in the retina of the toad Bufo marinus

We have previously shown that an antibody against neuron-specific enolase (NSE) selectively labels Müller cells (MCs) in the anuran retina (Wilhelm et al. 1992). In the present study the light- and electron-microscopic morphology of MCs and their distribution were described in the retina of the toad, Bufo marinus, using the above antibody. The somata of MCs were located in the proximal part of the inner nuclear layer and were interconnected with each other by their processes. The MCs were uniformly distributed across the retina with an average density of 1500 cells/mm². Processes of MCs encircled the somata of photoreceptor cells isolating them from each other by glial sheath, except for those of the double cones. Some of the photoreceptor pedicles remained free of glial sheath. Electron-microscopic observations confirmed that MC processes provide an extensive scaffolding across the neural retina. At the outer border of the ganglion cell layer these processes formed a non-continuous sheath. The MC processes traversed through the ganglion cell layer and spread beneath it between the neuronal somata and the underlying optic axons. These processes formed a continuous inner limiting membrane separating the optic fibre layer from the vitreous tissue. Neither astrocytic nor oligodendrocytic elements were found in the optic fibre layer. The significance of the uniform MC distribution and the functional implications of the observed pattern of MC scaffolding are discussed.     

Intrinsic organization of snake dorsomedial cortex: an electron microscopic and golgi study.

The cellular populations present in dorsomedial cortex in the snakes Constrictor constrictor, Natrix sipendon and Thamnophis sirtalis are described at the light microscopic level using Nissl and Golgi preparations as well as at the ultrastructural level. This area plays a central role in cortical organization in snakes by participating in major commissural and association projections. Systematic analyses of Golgi preparations indicate that five populations of neurons are present in dorsomedial area and have a preferential laminar distribution. Layer 1 stellate cells have somata positioned in the center of the outermost cortical layer, layer 1. Their dendrites are confined to this layer. Double pyramidal cells have their somata loosely packed in layer 2. Their dendrites bear a moderate population of spines, ascending through layer 1 to the pial surface and descending partially through layer 3. Some double pyramidal cells have somata displaced downwards into the upper third of layer 3. These neurons closely resemble the layer 2 double pyramidal cells. Layer 3 stellate cells have somata positioned in the middle third of layer 3. Their dendrites extend in all directions throughout layer 3 and through layer 2 into layer 1. Finally, horizontal cells have their somata positioned deep in layer 3, near the ventricle, and dendrites aligned concentric with the ventricle. Comparison of the organization of the known afferents to dorsomedial area with the distribution of the five cell types suggests that the laminations of both afferent fibres and dorsomedial neurons places specific neuronal populations in synaptic contact with specific sets of afferents.     

Single cell shape and population densities of indoleamine-accumulating and displaced bipolar cells in Reeves' turtle retina

Two types of bipolar cell in the Geoclemys reevesii retina were studied quantitatively by means of specific cell labelling with an indoleamine derivative (5,6-dihydroxytryptamine, 5,6-DHT), a nucleic acid stain (4,6-diamidino-2-phenylindole, DAPI) and Lucifer yellow CH. Indoleamine-accumulating (IA) bipolar cells were selectively labelled with 5,6-DHT applied intraocularly. After the cells accumulated 5,6-DHT, the indoleamine fluorescence was photoconverted to diaminobenzidine products to allow observation of morphological details. Close examination of many cells (cell number; n = 120) showed that the IA bipolar cells consist of a single morphological type whose axon collaterals ramify sublaminae 1, 4 and 5 respectively. This terminal branching pattern corresponds to cells that hyperpolarize when their receptive field centres are illuminated (Weiler 1981). The density of IA bipolar cells was highest in the visual streak (4130 cells mm-2) and lowest at the peripheral margin (1970 cells mm-2). By applying a small amount of DAPI to the eye, nuclei located in the most proximal row of the outer nuclear layer were labelled selectively. By using selective intracellular dye injection into DAPI-labelled cells under fluorescence microscope (Tauchi & Masland 1984, 1985), these cells were found to have Landolt's clubs and single descending axons. Dye injections into more than fifty DAPI-labelled somata showed that they belonged exclusively to displaced bipolar cells. These comprised at least two subtypes that differ in the ramification pattern of their axon terminals within the inner plexiform layer: one was monostratified, whereas the other was bistratified. The displaced bipolar cell density was as high as 9400 cells mm-2 in the central retina, falling to 2000 cells mm-2 in the superior margin. In vitro Lucifer labelling revealed that the overall bipolar cell density in the central retina was as high as 39,300 cells mm-2. Both the conventionally located and displaced bipolar cells were included in this population. About 11% of the total bipolar cell population consisted of IA bipolar cells. Assuming that one half of the conventionally located bipolar cells are the centre-hyperpolarizing type, IA bipolar cells represent approximately 28% of the total. As displaced bipolar cells represent almost one quarter of the total bipolar population, the dislocation of their somata stands out morphologically, inviting investigation of possible functional correlates.     

Morphological properties of tectal neurons that project to the nucleus geniculatus lateralis,pars ventralis (GLv) and the surrounding ventral thalamus in chicks

Layer 10 neurons of the chick tectum were morphologically investigated. The layer 10 neurons displayed heterogeneous immunoreactivities to calcium-binding proteins (CaBPs). Calbindin (CB)-immunoreactive (ir) neurons had pyramidal or round somata, primarily found in layers 5, 9, and 13. Parvalbumin (PV)-ir neurons were of various shapes with small to large somata (109.7 ± 48.6 μm²) that were located mainly in layers 4 and 10. Calretinin (CR)-ir neurons had small to middle-sized somata (79.3 ± 9.7 μm²) located primarily in layers 10 and 13, and most of them were similar to typical radial cells in size and shape. Two distinct types of neurons that projected to the nucleus geniculatus lateralis, pars ventralis (GLv) and ventral thalamus were demonstrated in layer 10. Type 1 cells had small to middle-sized somata (74.3 ± 33 μm²), and each cell had a single apical dendrite that ramified into bush-like branches in layer 7. These cells corresponded to CR-ir neurons and radial cells in size and shape. Type 2 cells had larger somata (124.7 ± 52.6 μm²), and their shapes were pyramidal, polygonal, or oval. They had multiple obliquely ascending dendrites that ramified into bush-like branches in layer 7. These cells often appeared similar to PV-ir neurons.     

Radial astrocytes and ependymocytes in the spinal cord of the adult toad (Bufo bufo L.)

Summary Immunohistochemical and ultrastructural techniques have been used to demonstrate glial fibrillary acidic protein (GFAP) immuno-positive cells in the adult toad spinal cord. Two types of GFAP-immunoreactive cells were observed: ependymocytes and radial astrocytes. GFAP-positive ependymocytes were scarce and contained the immunoreactive product in their processes. They showed intermediate filaments in the basal pole and in their processes when studied with the electron microscope. These immuno-positive ependymocytes represent the tanycytic form of ependymal cells because their processes ended at the subpial zone. The radial astrocytes showed a more intensive immunoreactive product in somata and processes when they were located far away from the ependymal layer. Cell bodies and processes were also associated with blood vessels, but most of the processes ended at the subpial zone forming a continuous subpial glia limitans. The GFAP-positive processes, which form this subpial glia limitans in the toad spinal cord, belong to both tanycytic ependymocytes and radial astrocytes, whose somata are located in the grey matter. These findings lead us to suggest that both types of GFAP-immunopositive cells might be the functional equivalents of mammalian astrocytes.     

GABA immunoreactivity in the main olfactory bulb of the frog Rana temporaria

Immunoreactivity for gamma-aminobutyric acid (GABA) was localized at the light microscopic level in the main olfactory bulb (MOB) of the frog, Rana temporaria. By means of free-floating peroxidase-antiperoxidase immunocytochemical technique, GABA was found in a large number of neurons in the granular cell layer, in a few small somata in the mitral cell layer and in two different types of cell somata in the glomerular layer. Individual GABA-immunopositive cells were found in the olfactory nerve layer. GABA immunostaining was also localized in cell processes and fiber fragments. There were many immunoreactive puncta in all layers of the MOB. GABA-positive punctate structures often outlined immunonegative cells in the mitral cell and glomerular layers. Rounded tightly packed groups of immunoreactive puncta were found only along ventral border of the glomerular layer. The results are discussed in comparison with data obtained on mammalian MOB in terms of MOB functional organization.     

Localization of CD15 immunoreactivity in the rat retina

Using immunocytochemistry, we have investigated the localization of CD15 in the rat retina. In the present study, two types of amacrine cell in the inner nuclear layer (INL) and some cells in the ganglion cell layer were labeled with anti-CD15 antisera. Type 1 amacrine cells have large somata located in the INL, with long and branched processes ramifying mainly in stratum 3 of the inner plexiform layer (IPL). Type 2 cells have a smaller soma and processes branching in stratum 1 of the IPL. A third population showing CD15 immunoreactivity was a class of displaced amacrine cells in the ganglion cell layer. The densities of type 1 and type 2 amacrine cells were 166/mm(2) and 190/mm(2) in the central retina, respectively. The density of displaced amacrine cells was 195/mm(2). Colocalization experiments demonstrated that these CD15-immunoreactive cells exhibit gamma-aminobutyric acid and neuronal nitric oxide synthase (nNOS) immunoreactivities. Thus, the same cells of the rat retina are labeled by anti-CD15 and anti-nNOS antisera and these cells constitute a subpopulation of GABAergic amacrine cells.     

Distribution of inhibitory synapses on the somata of pyramidal neurons in cat motor cortex

The mechanisms by which cortical neurons perform spatial and temporal integration of synaptic inputs are dependent, in large part, on the numbers, types, and distributions of their synapses. To further our understanding of these integrative mechanisms, we examined the distribution of synapses on identified classes of cortical neurons. Pyramidal cells in the cat motor cortex projecting either to the ipsilateral somatosensory cortex or to the spinal cord were labeled by the retrograde transport of horseradish peroxidase. Entire soma of selected corticocortical and corticospinal cells were examined using serial-section electron microscopy. The profiles of these somata and the synapses formed with each of these profiles were reconstructed from each thin section with a computer-aided morphometry system. All somatic synapses were of the symmetrical, presumably inhibitory type. For both cell types, these synapses were not homogeneously distributed over the somatic membrane, but were clustered at several discrete zones. The number and density of synapses on the somata of different corticocortical and corticospinal neurons were not significantly different. However, the density of these synapses was inversely correlated with the size of their postsynaptic somata. We discuss the significance of these findings to the integrative properties of cortical neurons.     

The fine structure of the parietal retinas of Anolis carolinensis and Iguana iguana

An electron microscopic examination of the parietal retinas of Anolis carolinensis and Iguana iguana demonstrated within each retina (1) two distinct populations of neurons, (2) two populations of glia, and (3) a population of photoreceptors which could not be subdivided. A small population of very electron-dense cells, in many respects similar to photoreceptors, was also found in the iguana. Correspondingly dark processes were found in the plexiform layer of each retina. Parietal photoreceptors generally resemble cones of the lateral eye. Glial cells were sub-classified on the basis of the location of their somata and the disposition of their processes. Neurons were identified by virtue of their cytology and their reception of axosomatic ribbon synapses from unidentified plexiform layer processes. Neuronal subtypes were located on opposite sides of the plexiform layer. Neurons distal to that layer were found to project the initial segments of their processes into the plexiform layer parallel to its long axis, while neurons central to the plexiform layer projected axons centrally and dendrites radially into the plexiform layer. The existence of at least two neuronal populations and of interphotoreceptor synapses suggests that photosensory processing within the parietal retina may be more complex than previously assumed.     

Somatostatin-like immunoreactivity in the amygdala of the pig

The distribution and morphology of neurons containing somatostatin (SOM) was investigated in the amygdala (CA) of the pig. The SOM-immunoreactive (SOM-IR) cell bodies and fibres were present in all subdivisions of the porcine CA, however, their number and density varied depending on the nucleus studied. The highest density of SOM-positive somata was observed in the layer III of the cortical nuclei, in the anterior (magnocellular) part of the basomedial nucleus and in the caudal (large-celled) part of the lateral nucleus. Moderate to high numbers of SOM-IR cells were also observed in the medial and basolateral nuclei. Many labeled neurons were also consistently observed in the lateral part of the central nucleus. In the remaining CA regions, the density of SOM-positive cell bodies varied from moderate to low. In any CA region studied SOM-IR neurons formed heterogeneous population consisting of small, rounded or slightly elongated cell bodies, with a few poorly branched smooth dendrites. In general, morphological features of these cells clearly resembled the non-pyramidal Golgi type II interneurons. The routine double-labeling studies with antisera directed against SOM and neuropeptide Y (NPY) demonstrated that a large number of SOM-IR cell bodies and fibers in all studied CA areas contained simultaneously NPY. In contrast, co-localization of SOM and cholecystokinin (CCK) or SOM and vasoactive intestinal polypeptide (VIP) was never seen in cell bodies and fibres in any of nuclei studied. In conclusion, SOM-IR neurons of the porcine amygdala form large and heterogeneous subpopulation of, most probably, interneurons that often contain additionally NPY. On the other hand, CCK- and/or VIP-IR neurons belonged to another, discrete subpopulations of porcine CA neurons.     

Properties of the serotonergic cerebral ganglion neurons of the gastropod mollusc, Philine aperta

The characteristics of a pair of identified neurons found in the cerebral ganglia of the gastropod mollusc Philine aperta have been examined. Because they appear to contain serotonin, and since they probably also use serotonin as a neurotransmitter, these neurons were named the serotonergic cerebral neurons (SCNs). Each SCN sent an axon out of the ipsilateral cerebro-buccal connective to the buccal ganglia. The SCNs also had extensive projections to all the ipsilateral, and most of the contralateral, buccal nerve trunks. Stimulating the SCNs produced the hyperpolarization of a pair of identified buccal ganglion mechano-sensory neurons (the S-cells), and had an excitatory action on the electrical activity of acinar cells of the salivary glands. A comparison of the properties of the Philine SCNs with those of similar serotonin-containing cerebral ganglion neurons in other gastropod molluscs provides evidence of homology with these neurons.     

Phe-X-Pro-Arg-Leu-NH(2) peptide producing cells in the central nervous system of the silkworm,Bombyx mori

Members of the neuropeptide family having Phe-X-Pro-Arg-Leu-NH(2) (FXPRLamide; X=Ser, Thr, Val, or Gly) at the C-terminus serve as regulators of oviduct and visceral muscle contraction, sex pheromone production, and diapause induction. Antibody raised against Bombyx mori diapause hormone recognized a variety of FXPRLamide peptides. Using this antibody, the antigen was immunocytochemically localized in the central nervous system (CNS) of the silkworm, Bombyx mori. Immunoreactive somata were observed in all ganglia of the CNS including the brain. Twelve somata localized at the midline of the suboesophageal ganglion (SG) were most intensely stained, and their neurite projections reached the retrocerebral complex. Thus, these cells in the SG exhibited typical features of neuroendocrine neurons. Marked reduction in immunoreactivity was observed in a pair of neurosecretory cells in the labial neuromere in SG of diapause type pupae, which indicates an active release of FXPRLamide peptides from these cells. No clear connection to neurohemal sites were observed in immunoreactive cells in the brain, thoracic or abdominal ganglia, suggesting that the immunoreactive peptides in these organs are likely to serve as neurotransmitters or neuromodulators.     

Gaba sensitivity of the somata and dendrites of pyramidal cells in isolated slices of rat hippocampus

During experiments on isolated slices slices of rat hippocampus the inhibitory action of -aminobutyric acid (GABA) was investigated on the excitation of field CA, pyramidal neurons, together with the effects of bicuculline, penicillin and thiopentone on this process. It was found that GABA effectively and reversibly reduced the amplitude of the antrodomic population spike in the area of both the somata and the dendrites of these cells. The sensitivity of apical dendrites to GABA exceeded that of the somata by one order, increasing in a proximal-distal direction. The somata of pyramidal neurons were marked by pronounced desensitization to GABA. Bicuculline and penicillin acted as GABA antagonists at all the levels of CA, pyramidal cells investigated. Bicuculline blocked the effects of GABA on somata and dendrites in almost equal measure. The antagonistic effects of penicillin were 10 times greater in the pyramidal layer than in the dendritic region. Thiopentone reinforced the inhibitory effects of GABA. The potentiating effects of thiopentone were exerted most strongly on the dendrites. It is postulated that the membrane of field CA, neurons contain two types of bicuculline-sensitive GABA receptors, differing in their location (mainly on the cell body or dendrite), their pharmacology, and degree of desensitization to GABA.A. A. Bogomolets Institute of Physiology, Academy of Sciences of the Ukrainian SSR, Kiev. Translated from Neirofiziologiya, Vol. 17, No. 6, pp. 737–746, November–December, 1985.     

The morphology and fine structure of the giant interneurons of the wood cricket Nemobius sylvestris

The structural and ultrastructural characteristics of giant interneurons in the terminal abdominal ganglion of the cricket Nemobius sylvestris were investigated by means of cobalt and fluorescent dye backfilling and transmission electron microscopy.The projections of the 8 eight pairs of the biggest ascending interneurons (giant interneurons) are described in detail. The somata of all interneurons analyzed are located contralateral to their axons, which project to the posterior region of the terminal ganglion and arborise in the cercal glomerulus. Neuron 7-1a is an exception, because its arborisation is restricted to the anterior region of the ganglion. The fine structure of giant interneurons shows typical features of highly active cells. We observed striking indentations in the perineural layer, enabling the somata of the giant interneurons to be very close to the haemolymph. The cercal glomerulus exhibits a high diversity of synaptic contacts (i.e. axo-dendritic, axo-axonic, dendro-axonic, and dendro-dendritic), as well as areas of tight junctions. Electrical synapses seem to be present, as well as mixed synapses. The anatomical organization of the giant interneurons is finally discussed in terms of functional implications and on a comparative basis.     

Laminar Distributions of Choline Acetyltransferase and Acetylcholinesterase Activities in the Inner Plexiform Layer of Rat Retina

Choline acetyltransferase and acetylcholinesterase activities were measured in samples taken at 7-micron increments through the inner plexiform layer of rat retina. These enzyme activities were not uniformly distributed through the depth of the inner plexiform layer. Peaks of choline acetyltransferase activity occurred at about one-third and peaks of acetylcholinesterase activity at about one-fifth of the depth into the inner plexiform layer from either side. The positions of the two peaks of choline acetyltransferase activity most likely correspond to the locations of processes from cholinergic amacrine somata in the inner nuclear layer, which spread in sublamina a, and processes from cholinergic amacrine somata "displaced" in the ganglion cell layer which spread in sublamina b of the inner plexiform layer. The peaks of acetylcholinesterase activity may in addition correspond to the processes of cholinoceptive amacrine and ganglion cells. The magnitudes of choline acetyltransferase and acetylcholinesterase activities are as high as found anywhere in rat brain, emphasizing the important role of cholinergic mechanisms in visual processing through the rat inner plexiform layer.     

Intrinsic organization of snake medial cortex: an electron microscopic and Golgi study.

The intrinsic organization of medial cortex in snakes, primarily of the genera Natrix and Boa, was studied using Golgi and electron microscopic techniques. The area has three distinct layers, each containing a characteristic population of neurons. Stellate cells comprise a relatively small population of neurons with their somata and dendrites restricted to layer 1, the most superficial layer. Their axons course horizontally in layer 1. Candelabra cells form the largest population of neurons in medial cortex. Their somata lie densely packed in layer 2 and are joined by specialized junctions. Ascending dendrites extend from the somata into layer 1. They consist of spine-free proximal segments and spine bearing distal segments. Descending dendrites extend from the somata into the upper half of layer 3. The proximal segments bear few spines but branch into several tapered, distal segments which have a moderate covering of spines. One or two axons originate from the descending dendrites and descend through layer 3. The axons bear collaterals in the deep half of layer 3 and eventually bifurcate in the alveus. The medial branches run into the septum; the lateral branches course through other cortical areas. The axons bear frequent varicosities within medial cortex. Periventricular cells lie in the deep half of layer 3, either singly or in clusters. Their ascending dendrites extend radially into layer 1 where they branch into distal segments which resemble those of the candelabra cells. Their descending dendrites arborize horizontally in the alveus and bear a moderate covering of spines. Ependymal cells line the ventricular surface and send radial processes through the area's depth bearing lamellate processes.     

Immunohistochemical profiling of the ultimobranchial remnants in the rat postnatal thyroid gland

Ultimobranchial (UB) remnants are a constant presence in the thyroid throughout rat postnatal life; however, the difficulty in identifying the most immature forms from the surrounding thyroid tissue prompted us to search for a specific marker. With that objective, we applied a panel of antibodies reported to be specific for their human counterpart, solid cell nests (SCNs), using double immunohistochemistry and immunofluorescence. Our results demonstrated that cytokeratin 34βE12 and p63 are highly sensitive markers for the immunohistologic screening of UB‐remnants, independently of their maturity or size. Furthermore, rat UB‐follicles (UBFs) coincided with human SCNs in the immunohistochemical pattern exhibited by both antigens. In contrast, the pattern displayed for calcitonin and thyroglobulin differs considerably but confirm the hypothesis that rat UB‐cells can differentiate into both types of thyroid endocrine cells. This hypothesis agrees with recent findings that thyroid C‐cells share an endodermic origin with follicular cells in rodents. We suggest that the persistence of p63‐positive undifferentiated cells in UB‐remnants may constitute a reservoir of basal/stem cells that persist beyond embryogenesis from which, in certain unknown conditions, differentiated thyroid cells or even unusual tumors may arise.