Search

Spontaneous nerve regeneration beyond the scar frequently occurs in fish after spinal cord lesions, in contrast to mammals. Here we examined the spatiotemporal relationship between the fibrous scar and axonal regeneration in the goldfish. Within 1 week after hemisection of the spinal cord, the open wound was closed by a fibrous scar that was demarcated from the surrounding nervous tissue by the glia limitans, which was immunoreactive for laminin. Within 1 week after hemisection, regenerating axons entered the fibrous scar, and were surrounded by laminin-coated tubular structures continuous with the glia limitans. Regenerating axons that initially entered the fibrous scar were usually accompanied by glial processes. Within 2-3 weeks after hemisection, the tubular structures became enlarged, and the regenerating axons increased in number, fasciculating in the tubules. Glial processes immunoreactive for glial fibrillary acid protein and 5-hydroxytryptamine neurons then entered the tubular structures to associate with the regenerating axons. The tubular structures developed further, creating tunnels that penetrated the fibrous scar, through which the regenerating axons passed. At 6-12 weeks after hemisection, the fibrous scar was smaller and the enlarged tunnels contained many glial processes and several axons. The findings of present study demonstrated that, following spinal lesions in goldfish, regenerating axons enter and pass the scar tissue. The regenerating axons first enter the fibrous scar with glial elements and then grow through laminin-coated tubular structures within the fibrous scar. Invasion by glial processes and neuronal elements into the tubular structures reduces the fibrous scar area and allows for more regenerating axons to pass beyond the fibrous scar., (Copyright © 2014 IBRO. Published by Elsevier Ltd. All rights reserved.)

Immunoglobulin-positive lymphocytes are present close to vasoactive intestinal polypeptide-positive (VIP(+)) nerve fibers in the lamina propria of the intestinal tract, and have an important role in mucosal defense. The number of immunoglobulin A-positive (IgA(+)) cells close to the epithelial basement membrane and nerve fibers is increased by the administration of lipopolysaccharides, which induce IgA secretion into the intestinal lumen. The relationship between immunoglobulin-positive lymphocytes and the VIP(+) nerve fibers during inflammation, such as in inflammatory bowel disease, however, is not well known. The morphological relationship between immunoglobulin-positive cells and the basement membrane or the VIP(+) nerve fibers in the colon was examined using double immunofluorescent labeling in an inflammatory bowel disease mouse model created by oral administration of dextran sodium sulfate (DSS). DSS administration induced goblet cell loss, crypt loss, intestinal epithelium deformation and infiltration of inflammatory cells in the mucosa. In the colon, the number and percentage of IgA(+) lymphocytes close to the basement membrane and the VIP(+) nerve fibers in the lamina propria increased after DSS administration, in parallel with the pathologic progress in the inflamed tissue. On the other hand, the percentage of immunoglobulin G-positive (IgG(+)) lymphocytes close to the basement membrane and the VIP(+) nerve fibers decreased, although the total number of IgG(+) lymphocytes in the lamina propria increased. We suggest that the immunoglobulin-producing lymphocytes and enteric nerve fibers in the colon normally have a close morphological relationship, and that this relationship is reinforced in a cell-specific manner during inflammation.

The sympathetic preganglionic cell pool in Xenopus laevis can be divided into four parts, i.e., the intercalated nucleus (IC) and the intermediolateral nucleus (IML) located respectively at the medial and the lateral borders of the lateral field, the lateral funiculus, and the ventral field within the thoracolumbar spinal segments. We compared the location of the preganglionic cells labeled following tracer application to the paravertebral sympathetic chain with those labeled following application to the celiac ganglion (CG), the adrenal gland (AG), and the splanchnic nerves (SNs) and found that their relative contribution differs depending on the sites. In tracer application to the paravertebral chain ganglia and the sympathetic trunk, 31.4-41.9% and 43.9-58.4% of labeled cells were detected respectively in the IC and in the IML, whereas application to the CG, AG, and on all the SNs, revealed that more than 84% of labeled cells were found in the IML and in the lateral funiculus with less than 8.6% in the IC. The contribution of the ventral field cells was less than 7.5% in all experiments. This type of topographic cytoarchitecture is a character shared with the mammalian preganglionic cell pool, but what distinguishes it from that of mammals is its systematic form throughout the entire longitudinal extent of the pool. In Xenopus, differences of mean soma areas and dendritic projections of labeled cells also suggest that the cell pools are distinguished not only by their location and axonal projections, but also by the morphology of their cells., ((c) 2009 Wiley-Liss, Inc.)

A tyrosine hydroxylase-immunoreactive cell mass is found in the caudal portion of the dorsal nerve ganglion of the red-eared slider, Trachemys scripta elegans. The ganglion appears as a flat oval structure in the horizontal plane, where the major axis runs latero-medially, and the minor axis rostro-caudally in the ventral view. A communicating branch to the sympathetic chain diverges from the top of each tubercle which lies on the caudo-lateral side of the ganglion. A tyrosine hydroxylase- immunoreactive cell mass is located in this tubercle. This cell mass exists in both sexes. Tyrosine hydroxylase-immunoreactive cells, that contain Nissl bodies in cytoplasm and are enveloped by the satellite cells, are multipolar and their neural processes are distributed in a distal direction into the spinal nerve. The range of distribution of the synapsin I-immunoreactive structures is limited to the tyrosine hydroxylase-immunoreactive cell mass. The chelonian dorsal spinal nerve ganglia are a conglomerate of the spinal nerve ganglion proper and the sympathetic ganglion., (Copyright 2009 S. Karger AG, Basel.)

In contrast to mammals, spontaneous nerve regeneration occurs in the teleost spinal cord. In the present study, we examined whether neurogenesis is involved in posttraumatic regeneration in the goldfish spinal cord. In intact fish, many spinal cells positive for both a monoclonal neuronal marker (Hu) and bromodeoxyuridine (BrdU) were observed 24 h after i.p. injection of BrdU, suggesting that constant neurogenesis occurs in the goldfish spinal cord. After hemisection of the spinal cord, the number of spinal cells positive for Hu and BrdU was significantly increased around the lesion site. The number of Hu- and BrdU-positive cells reached the maximum level 7 days after hemisection. In intact fish, spinal cells positive for both Hu and BrdU were also observed 5 weeks after BrdU injection, suggesting that newborn neurons survive for a long time. Six weeks after hemisection, the number of surviving Hu- and BrdU-positive cells at the lesion site was significantly increased as compared with that in intact fish, and some of them were also positive for 5-HT. A retrograde tract tracing study showed that the 5-HT+ neurons were close to the regenerated axons passing through the lesion site. These results suggest that adult neurogenesis occurs in the goldfish spinal cord, and that neurogenesis is activated by spinal cord lesion. The newly produced neurons survive a long time at the lesion site, and might participate in the repair of injured tissue and in the regeneration of descending long axons beyond the lesion site.

Immunoglobulin A (IgA) lymphocytes are present close to the nerve fibers in the lamina propria of the small intestine, and the administration of lipopolysaccharides (LPSs) increases the number of these cells and IgA secretion to the lumen. In the present study, we demonstrated that the nerve fibers immunoreactive for vasoactive intestinal peptide (VIP), neuropeptide Y (NPY), and calcitonin gene-related peptide (CGRP) were close to the IgA lymphocytes in the mouse ileum lamina propria. Three hours after intraperitoneal administration of LPSs, IgA lymphocytes close to VIP nerve fibers, those close to basement membrane, and those close to both VIP nerve fibers and basement membrane were increased in number. Further, all IgA lymphocytes seen in the ileum lamina propria expressed the receptors for VIP, VIPR1, and VIPR2. Electron microscopy revealed that varicosities were in close apposition to the lymphocyte plasma membrane. The present study suggests that VIP/NPY/CGRP neurons in the submucosal plexus have a close anatomical relationship to IgA lymphocytes, playing a role in the secretion of IgA and intestinal fluid in response to stimulation by lipopolysaccharides, pathogens, or toxins.

Following spinal transection, descending spinal projections from goldfish brainstem neurons spontaneously regenerate beyond the lesion site. The nucleus of the medial longitudinal fasciculus (nFLM), which has a critical role in swimming, also sends regenerated axons over a long distance to the ipsilateral spinal cord. To examine whether regenerated axons re-innervate the appropriate targets, we injected rhodamine dextran amine (RDA) into the nFLM of spinally transected goldfish and examined anterogradely labeled axons in the spinal cord. In intact controls, there were many RDA-labeled boutons or varicosities in the spinal cord in close apposition to both neurons positive for calcitonin gene-related peptide (CGRP), and those negative for CGRP. This suggests that the nFLM neurons project axons directly to the motoneurons and interneurons in the spinal cord. Four days after hemisection 1 mm caudal to the rostral end of the spinal cord, the number of RDA-labeled boutons in close apposition to the spinal neurons was significantly decreased on the side ipsilateral to the injection. Six to twelve weeks after spinal hemisection, regenerated axons ran through the repaired lesion site, and the number of RDA-labeled boutons or varicosities in close apposition to the ipsilateral spinal neurons had returned to the control level. These findings suggest that the midbrain-spinal pathway, critical for locomotion in fish, spontaneously regenerates beyond the lesion site to re-innervate the appropriately innervated targets after spinal lesion.

Objective: In the past the microkinetics of blood flow in the infrared pit organs of pit vipers has been studied with Doppler flowmetry using various infrared stimuli such as a human hand or soldering iron at various distances, lasers of various wavelengths, etc. Quick-acting variations in blood flow were recorded, and interpreted as a cooling mechanism for avoiding afterimage in the infrared receptors. However, the Doppler measurements provided only the summation of blood flow in a number of vessels covered by the sensing probe, but did not give data on flow in individual vessels., Methods: In the present work the authors introduced into the bloodstream of Gloydius and Trimeresurus pit vipers fluorescent microspheres labeled with fluorescein isothiocyanate (FITC) contained in a solution of FITC-dextran in physiological saline. They observed the passage of the microspheres through individual pit organ vessels with a fluorescent microscope to which was attached a high-speed video camera and image intensifier. Output of the camera was recorded before, during, and after stimulus with a 810-nm diode laser. Recording was done at 250 frames/s on high-speed video apparatus and downloaded to a hard disk. Disk files were loaded into proprietary software and particles were tracked and average velocities calculated. The data were then tested for significance by ANOVA with post hoc tests., Results: A significant (p<.05) increase in blood velocity was found at the focal point of the stimulus laser, but not anywhere removed from this point. Proximal severing of the pit sensory nerves caused degeneration of the pit receptor terminals and abolished stimulus-induced blood flow changes, but did not affect normal blood flow., Conclusions: The authors conclude that the receptors themselves are directly and locally controlling the smooth muscle elements of the blood vessels, in response to heating of the receptors by infrared radiation. They speculate that the heavy vascularization constitutes a cooling system for the radiation-encoding receptors, and further that the agent of control may be a volatile neuromediator such as nitric oxide.

The visceral reflexes of the pelvic organs are mediated by connections between primary afferents innervating the pelvic organs and parasympathetic preganglionic neurons in the intermediolateral column of the sacral spinal cord. The present immunohistochemical study revealed many varicosities expressing transient receptor potential vanilloid 1 (TRPV1) that were closely apposed to the preganglionic neuronal perikarya at embryonic day 16 in mice. Many, but not all, varicosities expressing TRPV1 in the intermediolateral column were also immunopositive for calcitonin gene-related peptide. In contrast, no nerve fibers expressing TRPV1 projected to the sympathetic preganglionic cell column in the lumbar spinal cord in prenatal stages. The results of the present study raised the possibility that the primary afferents transmit signals elicited by the activation of TRPV1 receptors to the sacral parasympathetic preganglionic neurons. Thus, the functional circuit for pelvic spinal reflexes, such as micturition induced by urine influx, might develop in the prenatal stages in mice.

We examined transient receptor potential (TRP) V1 and TRPV2 expression in calcitonin gene-related peptide (CGRP) positive (+) primary sensory neurons projecting to the skin and skeletal muscles of the rat dorsum. Among the dorsal root ganglia at the levels from C2 to Th1, 34.9% of neurons projecting to the skin were positive for CGRP, and 32.6% or 21.6% of neurons projecting to the trapezius muscle or the longissimus muscle were positive for CGRP. Of the small CGRP+ neurons projecting to the skin, 53.5% were positive for TRPV1, 11.6% were positive for TRPV2. Of the small CGRP+ neurons projecting to the trapezius or the longissimus, 53.1 or 53.2% were positive for TRPV1, 8.8 or 8.3% were positive for TRPV2, respectively. In the periphery, 29.3% of CGRP+ nerve fibers were positive for TRPV1 in the skin, whereas 65.0 or 59.8% were positive in the trapezius or the longissimus. Therefore, the present study showed that the percentage of CGRP+ neurons projecting to the trapezius is higher than that to the longissimus, and that the co-localization percentage of CGRP and TRPV1 on the sensory nerves was also higher in the trapezius than in the longissimus and the skin.

In mouse ontogeny, neurons immunoreactive for transient receptor potential vanilloid receptor 1 (TRPV1) were observed primarily in the dorsal root ganglia (DRG) at embryonic day 13 (E13). In the embryonic period, the number of TRPV1(+) neurons decreased, but then gradually increased postnatally. Some of TRPV1(+) neurons were also immunoreactive for calcitonin gene-related peptide (CGRP). At postnatal day 7 (P7), 66% of CGRP(+) neurons were TRPV1(+), and 55% of TRPV1(+) neurons were also CGRP(+) in the L4 DRG. In the peripheral organs, TRPV1-immunorective nerve fibers were transiently observed in the skin at E14. They were also observed in the urinary tract at E14, and in the rectum at E15. Many TRPV1(+) nerve fibers in these organs were also CGRP(+). At P1, TRPV1(+) nerve fibers were observed in the respiratory organs, and to a lesser extent in the stomach, colon, skin, and skeletal muscles. The number of TRPV1(+) nerve fibers on each organ gradually increased postnatally. At P7, TRPV1(+) nerve fibers were also observed in the small intestine and kidneys. The percentage of total TRPV1(+) nerve fibers that co-localized with CGRP was greater in most organs at P7 than at P1. The present results indicate that TRPV1 expression on peripheral processes differs among organs. The differential time course of TRPV1 expression in the cell bodies might be related to the organs to which they project. Co-localization of TRPV1 with CGRP on nerve fibers also varies among organs. This suggests that the TRPV1-mediated neuropeptide release that occurs in certain pathophysiologic conditions also varies among organs.

To investigate the relation between the enteric nervous system and immunocytes, fecal immunoglobulin (Ig) A and the distribution of IgA-positive cells close to nerve fibers in the lamina propria around the crypts of the small intestine were examined after intraperitoneal injection of lipopolysaccharides (LPS). In mice, IgA in the feces 1 h (LPS1) and 6 h (LPS6) after LPS injection increased compared to that in the control group. In the ileum lamina propria of the LPS1 group, the number of IgA-positive cells close to the epithelial basement membrane was increased. In the LPS6 group, on the other hand, there was a significant increase in the number of IgA-positive cells close to both the basement membrane and the nerve fibers. Our data suggest that LPS induced an increase in the number of IgA-positive cells migrating to the nerve fibers, and that migration to the nerve fibers is as important as migration to the basement membrane for IgA production.

For the purposes of comparative anatomy, we used tracer techniques and immunohistochemistry to study the facial nerve in the pit viper Gloydius brevicaudus and obtained much new data applicable to the function of this nerve in snakes and, in particular, pit vipers. We were able to identify the superior salivatory nucleus in these snakes. Preganglionic fibers from this nucleus pass along the palatine nerve and an anterior communicating branch to reach the pterygopalatine ganglion attached to the deep branch of the trigeminal maxillary nerve. The palatine nerve also contains general somatic afferents and a very few special visceral afferents from some taste buds on the palate. In the mandibular direction, preganglionic fibers from the superior salivatory nucleus join special visceral efferents from the motor nucleus in the hyomandibular nerve, from which they pass into the chorda tympani to course together for a short distance. The special visceral efferents branch off outside the cranium, and the preganglionic fibers continue on to join the trigeminal mandibular nerve to project to small ganglia within the mandible. The chorda tympani also contains general somatic afferents from the mandibular region but no special visceral afferents. This is the first time that the superior salivatory nucleus and its adjuncts have been identified in a snake. The chorda tympani of these snakes is also distinguished from the mammalian condition by lacking any special visceral afferents and by branching outside the cranium., (Copyright 2004 Wiley-Liss, Inc.)

To examine the prenatal development of spinal visceral reflexes, primary sensory nerve fibers immunoreactive for calcitonin gene-related peptide (CGRP) were examined in the spinal cord, particularly in the autonomic preganglionic nuclei of mouse embryos. On embryonic day 16 (E16), CGRP-immunoreactive fibers were first observed in the sacral intermediolateral nucleus (IML) of the parasympathetic division as well as in the lumbar central autonomic nucleus (CA) of the sympathetic division, where they appeared in proximity to preganglionic neuronal perikarya immunoreactive for choline acetyltransferase or nitric oxide synthase. Most of the CGRP-immunoreactive varicosities were negative for substance P. Substance P-immunoreactive varicosities were scattered in these nuclei, but no appositions were seen on the preganglionic neuronal perikarya. On E18, CGRP-immunoreactive fibers were more abundant in the sacral IML and the lumbar CA. Co-expression of substance P and CGRP was frequently observed in the varicosities very close to the preganglionic neuronal perikarya on E18. CGRP-immunoreactive fibers were also observed in the lumbar IML on E18, although significantly fewer were found in this nucleus compared with the sacral IML. In contrast to the upper lumbar level, no fibers immunoreactive for CGRP were observed in the IML at the thoracic level. These results suggest that peptidergic primary sensory fibers grow to project to the selective targets of autonomic preganglionic neurons during the embryonic period. The potential direct connections between the peptidergic primary sensory fibers and preganglionic neurons innervating the pelvic viscera might provide a circuit for spinal visceral reflexes active in embryos.

Boid snakes have infrared sensing pits that resemble crotaline pits in electrophysiological function and ultrastructure, but differ in gross morphology, number, and location: boids have three or more simple pits in the labial scales vs a single facial pair with more complex morphology in the crotalines. We studied the morphology of the capillary bed and the microkinetics of blood flow in a boid snake, the ball python, Python regius, and compared them with the already known condition in crotalines. We used a Doppler blood flow recorder in conjunction with an electrocardiograph to measure blood flow and heartbeat, and resin casts, transmission electron microscopy, and laser confocal microscopy to study capillary morphology. Blood flow in response to infrared stimulus was virtually identical in the two taxa, but the morphology of the capillary bed differed drastically. In the ball python pits, the capillary bed consisted of a forest of vertically oriented loops with a characteristic dome at the top in contact with the receptor layer of the fundus. Immunohistochemical staining showed pericytes constricting the capillaries and domes with smooth muscle alpha-actin-labeled processes. Since latency of response was as short as 1 ms, the capillaries were apparently responding under local control to provide both nutrition and cooling to the heat-sensitive receptors. We concluded that mitochondria-filled receptors provided with a swiftly responding cooling system were nature's most efficient way of attaining infrared imaging.

The central and peripheral organization of thoracic visceral and somatic nervous elements was studied by applying dextran amines to the proximal cut ends of the thoracic splanchnic and somatic nerves in Xenopus laevis. Many labeled dorsal root ganglion cells of visceral afferents, and all somatic afferents, were located in a single ganglion of one spinal segment, and the two types of cells were distributed topographically within the ganglion. The labeled sympathetic preganglionic neurons were located predominantly in the same area of the thoracic spinal gray as in other frogs and in mammals. The labeled visceral afferents projected to Lissauer's tract and the dorsal funiculus. The visceral fibers of the tract ascended to the level of the subcerebellar area, supplying collateral branches to the lateral one-third of the dorsal horn and to the area of brainstem nuclei, including lateral cervical and descending trigeminal nucleus, and descended to the filum terminale. The visceral fibers of the dorsal funiculus were distributed to the dorsal column nucleus and the solitary tract. A similar longitudinal projection was also seen in the somatic afferents. The dual central pathway of thoracic primary afferents in the anuran spinal cord is a property held in common with mammals, but the widespread rostrocaudal projection through Lissauer's tract may be a characteristic of the anuran central nervous system. In frogs, the direct transmission of primary afferent information to an extremely wide area of the central nervous system may be important for prompt assessment of environmental factors and control of body functions., (Copyright 2003 Wiley-Liss, Inc.)

We compared the distribution of substance P (SP) and calcitonin gene-related peptide (CGRP) fibers of the superficial muscle layer (trapezius muscle), median muscle layer (rhomboideus muscle), and deep muscle layer (longissimus and spinalis muscles) of the dorsum of the rat. SP- and CGRP-immunoreactive fibers were seen along the walls of various types of blood vessels and within nerve bundles in skeletal muscles of all layers. Coexistence of SP and CGRP was evident in nerve fibers along the blood vessel walls. The total number of CGRP varicosities per millimeter square of muscle surface area was evaluated quantitatively, and CGRP varicosities were found to be significantly more numerous in the superficial muscle layer than in the deeper ones. After capsaicin treatment, most of the SP and CGRP fibers along the blood vessel walls were eliminated. These results suggest that sensory nerve fibers containing SP and CGRP are distributed more abundantly in the superficial muscle layer than in the deeper ones and that they might be involved in the regulation of local blood flow. The finding of SP- and CGRP-immunoreactive nerve fibers along the blood vessel walls connecting the trapezius muscle and the hypodermis raises the possibility that sensory stimuli to the skin affect the local blood flow of superficial muscle through collaterals of cutaneous fibers.

We used transmission electron microscopic montages to examine the composition of nerve bundles serving the infrared pit organs of two species of crotaline snakes, Agkistrodon blomhoffii and A. brevicaudus. In the three main bundles, the myelinated fibers totaled 2,200-3,700, and unmyelinated fibers 2,400. We also discovered for the first time two accessory bundles composed almost entirely of unmyelinated fibers running alongside the main bundles, containing an average total of 3,300 unmyelinated fibers vs. an average of 10 myelinated fibers. Thus, the average total of unmyelinated fibers was nearly twice that of myelinated fibers. To study the nature of the unmyelinated fibers, we did double staining immunohistochemistry with antibodies for substance P (SP) and vasoactive intestinal polypeptide (VIP) in combination with and without capsaicin pretreatment. SP and VIP immunoreactive varicose fibers ran straight toward the center of the pit membrane in parallel with arterioles and venules, and also formed a dense network around the periphery of the membrane. There were three types of fibers: fibers containing only SP, fibers containing only VIP, and fibers containing both peptides. SP-only fibers were distributed singly throughout the pit membrane and in small bundles around the periphery. SP+VIP fibers were distributed sparsely in the pit membrane and around its periphery. VIP-only fibers were distributed throughout the pit membrane and were of smaller diameter than SP and SP+VIP fibers. After treatment with capsaicin, most of the three types of varicose fibers disappeared from the central part of the pit membrane, but those around the periphery remained unaffected. The capsaicin-sensitive fibers may be unmyelinated sensory types, and the unaffected ones may be autonomic nerve fibers., (Copyright 2002 Wiley-Liss, Inc.)

The present study showed neurons immunoreactive for choline acetyltransferase (ChAT) in the cranial sympathetic ganglia lying close to the trigeminal-facial nerve complex of the filefish. In these ganglia, less than 1% of ganglion cells were positive for choline acetyltransferase. Choline acetyltransferase-positive neurons were significantly larger than the randomly sampled neurons in this ganglion. The majority of choline acetyltransferase-positive neurons were negative for tyrosine hydroxylase, but many of them were positive for galanin (GAL). Some neurons were positive for both choline acetyltransferase and tyrosine hydroxylase, but these neurons were rarely immunoreactive for dopamine beta hydroxylase, suggesting that they are not adrenergic. In the cranial sympathetic ganglia and the celiac ganglia, many nerve fibers immunoreactive for galanin were seen, and varicose terminals were in contact selectively with neurons negative for both choline acetyltransferase and tyrosine hydroxylase, but not with those positive for choline acetyltransferase or tyrosine hydroxylase. Nerve fibers immunoreactive for choline acetyltransferase were found to be present in contact with the deep layer of chromatophores, which was observed only in the labial region. These results suggest that cholinergic postganglionic neurons are present in the filefish cranial sympathetic ganglia, and that they also contain galanin. As few cholinergic sympathetic neurons express tyrosine hydroxylase and none express dopamine beta hydroxylase, they are unlikely to synthesize noradrenaline or adrenaline.

In an immunohistochemical study, the vagal motor nucleus of a teleost, the filefish Stephanolepis cirrhifer, could be divided into a rostral part and a caudal part, and the former into a dorsolateral group and a ventromedial group. The dorsolateral group consisted of neurons immunoreactive for calcitonin gene-related peptide, whereas the ventrolateral-caudal group was negative for calcitonin gene-related peptide. The latter group was retrogradely labeled after dextran amine injection to the visceral ramus of the vagus nerve, suggesting that it is a general visceral efferent column, made up of parasympathetic preganglionic neurons, whereas the dorsolateral rostral group is a special visceral efferent column. In the general visceral efferent column, a dense concentration of nerve fibers immunoreactive for serotonin, tyrosine hydroxylase, cholecystokinin-8, and substance P, and a small number of fibers immunoreactive for neuropeptide Y was observed. Perikarya in contact with varicose terminals immunoreactive for these substances were frequently seen. In contrast, in the special visceral efferent column, only a moderate concentration of neuropeptide Y-immunoreactive nerve fibers and a sparse distribution of fibers immunoreactive for tyrosine hydroxylase were observed. Perikarya in contact with varicose terminals immunoreactive for these substances were rare. These results suggest that the vagal parasympathetic preganglionic neurons might receive multiple inputs of monoaminergic and peptidergic fibers involved in the regulation of the visceral organs. On the other hand, monoaminergic and peptidergic afferent fibers might be of much less significance in the activity of the special visceral efferent component of the vagus nerve., (Copyright 2002 Wiley-Liss, Inc.)

Nerve fibers immunoreactive for enzymes synthesizing catecholamines were examined in the central autonomic nucleus, a column of sympathetic preganglionic neurons, in the filefish Stephanolepis cirrhifer. Varicose nerve fibers immunoreactive for tyrosine hydroxylase were densely distributed in the rostral part, sometimes in contact with perikarya but were sparse in the caudal part of this nucleus. Fluorescent double labeling distinguished noradrenergic nerve fibers immunoreactive for both tyrosine hydroxylase and dopamine beta hydroxylase, and dopaminergic fibers immunoreactive only for tyrosine hydroxylase. In the brainstem, catecholaminergic neurons were observed in the locus coeruleus, the caudal dorsomedial reticular zone of the medulla, and the area postrema. Double labeling of tyrosine hydroxylase and dopamine beta hydroxylase showed that the neurons in the locus coeruleus were all noradrenergic, and those in the caudal dorsomedial medulla were mostly noradrenergic, whereas the area postrema contained both noradrenergic and dopaminergic neurons. No catecholaminergic neurons were found in the ventral region of the brainstem. After application of DiI to the central autonomic nucleus, retrogradely labeled neurons were seen in the caudal dorsomedial medulla but not in the locus coeruleus or the area postrema. These findings suggest that the sympathetic preganglionic neurons of the filefish may receive noradrenergic axonal projections from neurons in the caudal dorsomedial medulla. In the light of previous studies, inputs of these catecholaminergic fibers to the central autonomic nucleus may be involved in regulation of sympathetic activity of peripheral organs, together with serotoninergic and peptidergic inputs to this nucleus., (Copyright 2001 Wiley-Liss, Inc.)

In the cellular column of sympathetic preganglionic neurons (SPNs) of the filefish Stephanolepis cirrhifer, neurons containing galanin (GAL) form a distinct population projecting specifically to non-adrenergic postganglionic neurons in the celiac and cranial sympathetic ganglia. The present study showed that virtually all of the GAL-immunopositive SPNs made contact with many nerve terminals immunopositive for cholecystokinin octapeptide (CCK-8). GAL-negative preganglionic neurons made contact with only 26% of this type of nerve terminal; CCK-8-immunopositive nerve fibers appeared to project selectively to GAL-immunopositive SPNs with projections to specific targets. The CCK-8-positive nerve fibers might be of primary sensory origin, and participate in the visceral reflexes.

Immunoreactivity for substance P and cholecystokinin-8 was examined in the nerve fibers in the central autonomic nucleus, a cell column for sympathetic preganglionic neurons, in the filefish Stephanolepis cirrhifer. Substance P-immunoreactive fibers were distributed throughout the entire rostrocaudal extent, but were more abundant in the caudal part of the column, where substance P-immunoreactive varicosities sometimes made contacts with the sympathetic preganglionic neurons. Cholecystokinin-8-immunoreactive fibers were found almost entirely in the rostral part of the column, where a dense network of varicosities was in close apposition to a considerable number of the sympathetic preganglionic neurons. Double labeling immunohistochemistry showed that substance P fibers and cholecystokin-8 fibers were entirely different, and distinct from serotonin-immunoreactive fibers. By using immunoelectron microscopy, synaptic specialization was sometimes observed between the dendrites of preganglionic neurons and varicosities immunoreactive for substance P and cholecystokinin-8. Substance P- and cholecystokinin-8 fibers were seen from the descending trigeminal tract, through the dorsolateral funiculus and the ventral portion of the dorsal horn, to the central autonomic nucleus. After colchicine treatment, substance P-immunoreactive perikarya were found in the cranial and spinal sensory ganglia. These results suggest that the sympathetic preganglionic neurons of the filefish receive innervation by substance P fibers and cholecystokinin fibers, and that the former might be of primary sensory origin. Topographical distribution of cholecystokinin-8-immunoreactive terminals in the central autonomic nucleus along the rostrocaudal extent might underlie the differential regulation of sympathetic activity via a distinct population of sympathetic preganglionic neurons., (Copyright 2000 Wiley-Liss, Inc.)

The infrared sensory membranes of pit organs of pit vipers have an extremely rich capillary vasculature that forms many vascular loops, each serving a small number of infrared nerve terminals. We clarified the ultrastructure of capillary pericytes in the pit membranes by scanning and transmission electron microscopy, and examined the immunoreactivity in their cytoplasm to two contractile proteins: smooth muscle alpha-actin (SM alpha-actin) and desmin. The capillary pericytes had two major cytoplasmic processes: thickened primary processes that radiate to embrace the endothelial tube and flattened secondary processes that are distributed widely on the endothelium. Coexpression of SM alpha-actin and desmin was observed in the pericytes of entire capillary segments, and SM alpha-actin was characterized by prominent filament bundles directed mainly at right angles to the capillary long axis. This expression pattern was different from that of capillary pericytes of the scales, where SM alpha-actin was expressed diffusely in the cytoplasm. In a series of electron microscopic sections, we often observed the pericyte processes depressing the endothelial wall. We also observed a close relationship of the pericytes with inter-endothelial cell junctions, and pericyte processes connected with the endothelial cells via gap junctions. From these findings, we surmised that capillary pericytes in the pit membrane have a close functional relationship with the endothelium, and through their contractile and relaxing activity regulate capillary bloodflow to stabilize production of infrared nerve impulses., (Copyright 2000 Wiley-Liss, Inc.)

Neuropeptides in the motor nerves innervating the red and white muscles of the goldfish Carassius auratus were examined. In the tonic red muscles, varicose nerve endings immunoreactive for both calcitonin gene-related peptide and substance P were found spread over the surface of the muscle fibers, but in the twitch white muscles only scattered nerve endings immunoreactive for calcitonin gene-related peptide were found. At the electron microscopic observation, dense electron products immunoreactive for calcitonin gene-related peptide and for substance P (SP) were detected in the motor nerve endings making synapses on the muscle fibers of the red muscles. In the spinal cord, all of the motor neurons showed immunoreactivity to calcitonin gene-related peptide, but the motor neurons immunoreactive for substance P were restricted to the ventrolateral group that has been shown to project predominantly to the red muscles. These results suggest that the motor neurons innervating the red and white muscles of the goldfish are distinct in their neuropeptide content. The present study also raises the possibility that SP might be related to the unique physiological properties of the tonic type red muscles, probably by direct binding to the acetylcholine receptors.

Immunoreactivity for galanin was examined in the sympathetic preganglionic neurons in the spinal cord, adrenal glands, sympathetic ganglia, and some sensory ganglia of the filefish Stephanolepis cirrhifer. Galanin-immunoreactive neurons were found only in the rostral part, but not in the caudal part of the central autonomic nucleus (a column of sympathetic preganglionic neurons of teleosts). Many galanin-immunoreactive nerve terminals were found in contact with neurons in the celiac ganglia and the cranial sympathetic ganglia on both sides of the body. Most neurons encircled by galanin-immunoreactive nerve fibers were negative for tyrosine hydroxylase. Galanin-immunoreactive nerve fibers were very sparse in the spinal sympathetic paravertebral ganglia. No galanin-immunoreactive nerve fibers were found in the adrenal glands. No sensory neurons of the trigeminal, vagal, or spinal dorsal root ganglia were positive for galanin-immunoreactivity. These results suggest that galanin-immunoreactive sympathetic preganglionic neurons have distinct segmental localization and might project specifically to a population of non-adrenergic sympathetic postganglionic neurons in the celiac and cranial sympathetic ganglia.

Serotonin-immunoreactive axonal components were observed in the central autonomic nucleus (CAN), a cell column of sympathetic preganglionic neurons in the rostral spinal cord of the filefish Stephanolepis cirrhifer. Serotonin-positive axonal varicosities were seen around neuronal perikarya through the whole rostrocaudal extent of the CAN, although their distribution pattern in the rostral CAN was different from that in the caudal CAN. Electron microscopically, serotonin-positive axonal varicosities were found to make axodendritic and axosomatic synapses on CAN neurons. Many serotonin-positive neuronal cell bodies were seen in the raphe nuclei in the lower brainstem, whereas only a few were found in the spinal cord. Thus most of serotoninergic axons within the CAN were considered to originate from the raphe nuclei in the lower brainstem.

To examine the presence of nitric oxide synthase (NOS) in the sensory system of the glossopharyngeal and vagus nerves of teleosts, nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd) activity and immunoreactivity for NOS were examined in the puffer fish Takifugu niphobles. The nitrergic sensory neurons were located in the ganglia of both the glossopharyngeal and the vagal nerves. In the vagal ganglion, positive neurons were found in the subpopulations for the branchial rami and the coelomic visceral ramus, but not for the posterior ramus or the lateral line ramus. In the medulla, nitrergic afferent terminals were found in the glossopharyngeal lobe, the vagal lobe, and the commissural nucleus. In the gill structure, the nitrergic nerve fibers were seen in the nerve bundles running along the efferent branchial artery of all three gill arches. These fibers appeared to terminate in the proximal portion of the efferent filament arteries of three gill arches. On the other hand, autonomic neurons innervating the gill arches were unstained. These results suggest that nitrergic sensory neurons in the glossopharyngeal and vagal ganglia project their peripheral processes through the branchial rami to a specific portion of the branchial arteries, and they might play a role in baroreception of this fish. A possible role for nitric oxide (NO) in baroreception is also discussed.

The infrared sensory membranes of the pit organs of pit vipers have an extremely rich capillary vasculature, which has been noted passim in the literature, but never illustrated or studied in detail. We rendered the pit vasculature visible in various ways, namely, by microinjection of India ink, by a combination of ink and succinate dehydrogenase staining, and by making resin casts for scanning electron microscope study. We also used transmission electron microscopy for identifying the types (arterioles, venules, capillaries) of blood vessels. Then we compared the pit vasculature with that of the retina and the dermis. Good visualization of the vasculature was obtained with both ink and resin injection. Arterioles, venules, and capillaries could be distinguished with all methods used. The monolayer vasculature was denser in the pit membrane than in the retina or skin. Each loop of the network enclosed a small number of infrared receptors so that all receptors were in contact with a capillary on at least one side. The forward-looking areas of the pit had a denser network than side-looking areas. Since infrared rays cause nerve impulses by raising the temperature of individual receptors, the capillary network functions not only as a supplier of energy but also as a cooling mechanism to reduce afterimages. Thus the denser network in the forward-looking areas causes these areas to be more sensitive and have better image resolution than the rest of the membrane.

The supramedullary cells (SMCs) are spinal neurons lying at the dorsal surface of teleosts. In the present study, we examined whether the SMCs of the puffer fish (Takifugu niphobles) might express gastrin/cholecystokinin-immunoreactivity, as observed in some other teleosts. All the SMCs were immunoreactive for gastrin/cholecystokinin. On the other hand, many immunoreactive varicose nerve fibers were also found terminating in the mucous glands in the skin. In addition, immunoreactive fibers were sparsely distributed in the epidermal layer. No neuronal cells other than the SMCs showed gastrin/cholecystokinin-immunoreactivity centrally or peripherally. The results suggest that gastrin/cholecystokinin-immunoreactive axons in the cutaneous mucous glands and epidermal layer are axons of the SMCs. In view of the present findings, the possible nature of SMCs was discussed.

Nicotinamide adenine dinucleotide phosphate (NADPH)-diaphorase activity was examined in the cranial sensory ganglia and brainstem of the banded dogfish, Triakis scyllia. Positive neurons were found in the vagal sensory ganglion projecting to the coelomic organs, but not in those projecting to the gills or the lateral line organs. Nerve terminals in the vagal lobe were also positive. No positive neurons were found in the glossopharyngeal, facial, or trigeminal sensory ganglia. These results suggest that use of nitric oxide in the vagal sensory transmission from the coelomic organs may have been maintained in the evolutionary process from fish to mammals.

Background: Crotaline snakes possess a pair of infrared-sensing pit organs that aid the eyes in the detection and apprehension of prey. The morphology of the receptors in the pit organs has been studied by light and transmission electron microscopy, and the ultrastructure of the receptors has been inferred from the results of this work. But this theoretical reconstruction has never been confirmed by any kind of three-dimensional imaging., Methods: We treated the receptor-containing membrane of the pit organs with potassium hydroxide to remove collagen and expose the receptors, which we then viewed by scanning electron microscopy., Results: We were able to obtain three-dimensional views of all structures previously reported to exist within the receptor-containing membrane: terminal nerve masses formed from free nerve endings, supporting Schwann cells within the nerve masses, unmyelinated and myelinated nerve fibers, a capillary bed, and vacuole cells., Conclusions: By providing the first three-dimensional views of the infrared receptors, we have confirmed that previous theoretical reconstructions of the receptors were substantially correct and have provided new evidence of the spatial arrangement of the receptors in a monolayer array.

In snakes with infrared receptors the optic tectum receives infrared input in addition to visual and general somatosensory inputs. In order to observe their tectal termination patterns in ball pythons, Python regius, we injected horseradish peroxidase (HRP) into the nucleus of the lateral descending trigeminal tract (LTTD) which mediates infrared information, the optic nerve, and the nucleus of the trigeminal descending tract (TTD) which relays general somatosensory information. Fibers from LTTD were found in layers 5-13 of the contralateral optic tectum, and were especially dense in layers 7a-8. Optic nerve fibers terminated in layers 7a-13 of the contralateral tectum, and mainly in layers 12-13. TTD fibers were few, and could be seen in only the rostral half of the contralateral tectum. These fibers were found in layers 5-7b, but mainly in layers 6-7a. Among various types of neurons stained by the Golgi-Cox method, we focused on six types of neurons whose dendritic arborization overlapped with the distribution of the terminals of these sensory afferents described above. It is possible that these different sensory modalities converge on a single neuron of the various types.

Crotaline snakes, which have infrared-sensitive pit organs, provide a good model for linking neuron morphology with sensory modality. In the trigeminal ganglion of the habu, Trimeresurus flavoviridis, cells positive for calcitonin gene-related peptide-like (CGRP) immunoreactivity were found to be of two types, darkly stained and lightly stained. They were pseudo-unipolar, having an axon divided into stem, peripheral branch, and central branch, all of which were 1 micron or less in diameter. Other, CGRP-negative cells in the ganglion were also pseudo-unipolar, but much larger. In configuration, some of the positive cells were similar to the neurons with A-delta fibers, and others to the neurons with C fibers that have been reported by other workers. On the basis of their distribution and density, and physiological studies by other workers, the CGRP-positive cells were judged to be not part of the infrared-receptive system, but to be involved in the transmission of nociception in small fibers.

We demonstrate the existence of nerve fibers possessing substance P (SP) and calcitonin gene-related peptide (CGRP) immunoreactivity in the mouse cervical ventral roots. The distribution of the SP and CGRP fibers was similar, but CGRP fibers were generally more numerous. Both types entered the ventral pia mater or formed hairpin loops, but they did not enter the spinal cord directly through these roots. SP and CGRP fibers in the ventral roots were thin and had many varicosities. We suggest that these SP and CGRP fibers are involved not only in a sensory mechanism, but also in other functions, via the release of SP and CGRP from varicosities in the ventral roots.

We report the successful use of a cortico-cancellous bone graft to repair a cleft of the secondary palate in a 6-month-old Akita dog. The cleft extended from the incisal papilla to the posterior border of the soft palate. It was 10 mm wide in the hard palate and 16 mm in the soft palate. The cleft was repaired using the "push-back" method with single-pedicled mucoperiosteal flaps to close the cleft in the soft tissues, and a free bone graft from the tibia (cortex with cancellous bone attached) to fill the defect in the hard palate. Two and one-half years after the operation the graft was indistinguishable from surrounding bone on radiographs. Maxillary growth appeared to have progressed normally, resulting in normal occlusion.

In snakes with infrared receptors, the optic tectum receives input from both the visual and the infrared senses. We investigated the infrared and optic fiber terminations in the tectum with a combination of horseradish peroxidase and degeneration labeling. In addition to synapses by visual and infrared fibers onto individual neurons, we were able to observe for the first time visual and infrared synapses on one and the same dendrite.

Anterograde and retrograde transport of horseradish peroxidase was used to examine the afferent and efferent projections of the glossopharyngeal-vagal nerve in the hagfish Eptatretus burgeri. Anterogradely labeled ganglion cells are scattered in the glossopharyngeal-vagal nerve trunk, in the saccular ganglion, and in the brainstem. Afferent fibers of the glossopharyngeal-vagal nerve terminate in both the vagal lobe and the fasciculus communis. Close observation showed no morphological differentiation between these two structures, indicating that they are not separate entities, but a single, continuous structure that is homologous with the nucleus and tractus solitarius of other vertebrates. The median part of this structure (the commissura infima) is displaced more rostrally than the same part of the solitary nucleus in many other vertebrates. Some of the afferent fibers invade the ventral portion of the trigeminal sensory nucleus, which receives the maxillo-mandibular nerve fibers, and terminate there. Our study showed that the hagfish has only one nucleus in the vagal motor system, i.e., the vagal motor nucleus, which contains both parasympathetic and branchiomotor neurons. The dendrites of the vagal motor neurons in the hagfish are more highly developed than those in other vertebrates. This suggests that the motor reflex arc of the glossopharyngeal-vagal nerve in hagfishes may be simpler than in other vertebrates.

The lateral line sensory system of Lampetra japonica is innervated by the anterior and posterior lateral line nerves. The anterior lateral line nerve innervates all electroreceptors throughout the body and mechanoreceptors of the head. The posterior lateral line nerve innervates trunk mechanoreceptors. The anterior lateral line nerve consists of two ganglia (anterior lateral line and intracapsular) and four major peripheral branches (superficial ophthalmic, buccal, hyomandibular, and recurrent nerves). The posterior lateral line nerve has one posterior lateral line ganglion and one peripheral branch. The location and central projection patterns of the primary sensory neurons of these branches of the lateral line nerves were studied with the aid of horseradish peroxidase labeling. The ganglion cells of the buccal nerve were found in the rostral half, and those of the hyomandibular nerve were found in the caudal half of the medial part of the anterior lateral line ganglion. The lateral part of the anterior lateral line ganglion contains ganglion cells of the recurrent nerve and the superficial ophthalmic nerve. The rostral half of the intracapsular ganglion contains ganglion cells of the recurrent, hyomandibular, and buccal nerves. The ganglion cells of the posterior lateral line nerve were found in the posterior lateral line ganglion. The buccal nerve afferents terminated mainly in the lateral part of the ipsilateral mechanoreceptive medial nucleus. The peripheral part of the electroreceptive dorsal nucleus also received several afferents. The hyomandibular afferents terminated ipsilaterally in the central part of the medial nucleus and in the dorsolateral part of the dorsal nucleus. Some afferents of the hyomandibular nerve ascended and descended in the descending nucleus of the trigeminal nerve near its dorsal margin. The ventral nucleus, the primary nucleus of the VIIIth nerve, received a few fibers of the buccal and hyomandibular nerves. In the recurrent nerve, the fibers of the lateral part of the anterior lateral line ganglion terminated throughout the entire dorsal nucleus, and the fibers of the intracapsular ganglion projected to the dorsolateral part of the nucleus. The afferents of the posterior lateral line nerve terminated in the medial part of the ipsilateral medial nucleus and in the lateral part of the contralateral medial nucleus. In the cerebellar area, afferents of the anterior lateral line nerve were located laterally to those of the posterior lateral line nerve. Several fibers terminated in some branchiomotor nuclei, the cerebellar crest, and the dorsal gray near the obex level. No efferent cell bodies were found in the place where efferent neurons of the VIIIth nerve have been previously reported.

The VIIIth cranial nerve projections in the hagfish, which has only one circular canal in the ear, were studied by transganglionic HRP transport. This nerve has two branches, the nervus utricularis (N. utr.) and the nervus saccularis (N. sac.), each with its own ganglion, the ganglion utriculare (G. utr.) and the ganglion sacculare (G. sac.), respectively. Although the G. sac. has uniformly small cells, the G. utr. consists of two separate cell masses, a ventral mass of large cells and a dorsal mass of small cells. The small cells were labeled in both ganglia after horseradish peroxidase (HRP) injection into the endolymphatic space. The greater part of the terminal areas of these two branches overlapped in the ventral nucleus of the area acoustico-lateralis, but the terminals of the N. sac. extended slightly further in a caudal direction. No projections to the primordial cerebellum and no retrogradely labeled cells in the brain were found. The large cells in the ventral part of the G. utr. seem to be general cutaneous neurons, and the dorsal part of the area acousticolateralis seems to receive lateral line input.

With the peroxidase-antiperoxidase immunohistochemical method we ascertained the presence of substance P-like immunoreactivity (SPLI) in fibers and cell bodies of the trigeminal sensory system of the pit viper, Agkistrodon blomhoffi. There are a few SPLI fibers each in the principal sensory nucleus and the main neuropil of the lateral descending nucleus (i.e., the infrared sensory nucleus); a moderate number in the descending nucleus; and a large number in the caudal subnucleus, the medial edges of the interpolar subnucleus, and the marginal neuropil of the lateral descending nucleus. About 30% of the cell bodies in the ophthalmic and maxillo-mandibular ganglia show SPLI, and of the two craniocervical ganglia, the proximal ganglion has many more cells with SPLI than the distal ganglion. The SPLI distribution in the common trigeminal sensory system is similar to that of mammals, and suggests that the function of this system is also similar. In the infrared that the function of this system is also similar. In the infrared sensory system, the differing distribution in the main and marginal neuropils suggests separate functions for these two structures in the system.

Substance P-like immunoreactive nerve fibers were identified in the central retinal artery of the rabbit using the peroxidase-antiperoxidase method. The fibers were seen to encircle the central retinal artery throughout its course in the main trunk of the optic nerve. No labeled fibers were seen in the central retinal vein or in the retinal blood vessels. It appears, therefore, that the central retinal artery and the retinal blood vessels are innervated by different nerve systems: the central retinal artery by one of the peripheral nerves, and the retinal vessels by the central nervous system.

Anterograde and retrograde transport of horseradish peroxidase was used to examine the afferent and efferent projections of the VIIIth cranial nerve in the lamprey Lampetra japonica. Ganglion cells of the VIIIth nerve are classified into three types on the basis of their morphology. The central processes of these ganglion cells enter the medulla in two groups: the anterior group (mostly thick fibers) and the posterior group (mostly thin fibers). Afferent fibers mainly terminate within the ipsilateral ventral and octavomotor nuclei of the octavolateralis area and within the granular and molecular layer of the cerebellum. Some fibers terminate in the contralateral cerebellum, the medial and dorsal nuclei of the octavolateralis area, the descending nucleus of the trigeminal nerve, some cranial motor nuclei, and the lateral octavus nucleus, which has not been described previously. This small nucleus is located beneath the descending nucleus of the trigeminal nerve near the obex. Within the ventral nucleus, thin fibers occupy the dorsal part and thick fibers occupy the ventral part. The basic projection pattern of the primary afferents of the VIIIth nerve in the lampreys was similar to that of gnathostome fishes that have been studied to date. Cell bodies of the efferent vestibular neurons are located between the ipsilateral trigeminal motor nucleus and the facial motor nucleus. The lateral location of these cell bodies differs from that of all other fish species that have been studied.

Trigeminal neurons were labeled by inserting HRP into holes cut in the pit receptor membranes of a crotaline snake, Agkistrodon blomhoffi brevicaudus. Neurons were labeled in the ophthalmic ganglion and the maxillary division of the maxillo-mandibular ganglion, and the HRP was further transported across the ganglia and through the lateral descending trigeminal tract (dlv) to label axon terminals exclusively in the dlv nucleus (DLV). In 6 successful preparations, 7.1-19.3% of totals of 5568-5986 cells in the maxillary division of the ganglion were labeled, but none at all were labeled in the mandibular division. Only a few or none at all were labeled in the ophthalmic ganglion. Cells in the two ganglia ranged in size from 10 to 55 micrometers, but large cells (greater than or equal to 40 micrometers) were scarce (4.9% of the total population). All HRP-labeled neurons fell in the median range of 20-39 micrometers. We concluded that these ganglion cells were infrared neurons, and were therefore the origin of the A delta fibers in the pit membrane. There were no HRP-labeled neurons above or below this range, in spite of the fact that smaller cells (less than or equal to 19 micrometers) made up 35.8% of the total population. In normal Nissl preparations we found both light- and dark-staining cells, but the size range of neither corresponded to the size range of infrared neurons.

We discovered in the brainstem of infrared-sensitive vampire bats, Desmodus rotundus, a specific nucleus not known in other species of bats. Because it corresponded in location and histological features to the infrared nucleus of infrared-sensitive snakes, we suggest the probability of its being part of the infrared processing system of vampire bats.

The trigeminal and glossopharyngeal ganglia of the adult mallard were studied following HRP injections into the principal trigeminal nucleus (PrV). The PrV consists of the principal trigeminal nucleus proper (prV) and the principal glossopharyngeal nucleus (prIX). After an injection into the prV, the labeled cells were found in the ipsilateral trigeminal ganglion. After an injection into the prIX, labeled cells were found in the ipsilateral distal glossopharyngeal ganglion, but not in the proximal ganglion of the IX and X cranial nerve (pGIX + X). In Nissl preparations, two types of ganglion cells in the trigeminal ganglion, pGIX + X, and distal ganglion of N IX could be distinguished: larger light cells and smaller dark cells. We could not determine whether the HRP-labeled cells belonged to both types or to one of them; but because all the labeled cells were over 20 microns, we concluded that the smallest cells (10-19 microns) in the trigeminal ganglion and distal ganglion of N IX did not project to the PrV. The labeling of the cells in the distal ganglion of N IX (average 34.5 microns) was uniformly moderate. In the trigeminal ganglion there were two types of labeled cells: heavily labeled cells (average 29.1 microns) and moderately labeled cells (average 35.1 l microns). These two types of labeling (moderate and heavy) may reflect two types of primary sensory neurons: cells with ascending, nonbifurcating axons, and cells with bifurcating axons. We speculate that the former are proprioceptive neurons and the latter tactile neurons. Labeled bifurcating axons in the sensory trigeminal complex gave off collaterals to all parts of the descending trigeminal nucleus except to the caudalmost laminated spinal part.

The infrared receptor neurons of Python reticulatus pit organs were all found to have bimodal sensitivity, responding to both infrared and touch stimuli with fairly rapid adaptation. The majority (22 of 29 neurons) had no background discharges at any temperature between 20 and 33 degrees C. The receptive areas were 150-250 micrometers in diameter and identical for both modalities. There was only one receptive area for each neuron. These facts suggest the possibility that some kinds of temperature sensitive neurons can also function as touch neurons and vice versa, not only in this species, but also in other animals.