Competitive interactions between neurons making axosomatic contacts in the leech

Axons of lateral nociceptive (N) neurons in leech segmental ganglia wrap certain somata in adjacent ganglia but no somata in their own ganglion. In adults, the N neurons, which accurately regenerate axosomatic wrappings, can be induced to sprout in their own ganglion and wrap target homologues if the ganglion is isolated by cutting the nerve cord. Manipulations that denervate the new targets without injuring the lateral N cell, including focal lesions and protease injections into other N cells, also cause sprouting within 2-4 months. In contrast, cutting the lateral N cell's axons causes little or no sprouting within the ganglion without denervation. Therefore, denervation rather than injury accounts for sprouting within the ganglion. It is concluded that lateral N cells can wrap somata in their own ganglion that are homologues of their usual targets, but they are prevented from doing so by axonal wrappings from N cells in adjacent ganglia.     

Regeneration of axons and synaptic connections by touch sensory neurons in the leech central nervous system

In studies of axonal regeneration, it has been difficult to determine (a) whether growth along the normal pathway is important for restoration of connections with previous targets and (b) whether the new synapses resemble the old in strength and location. To address these problems at the level of individual nerve cells, we have studied touch (T) sensory neurons in the leech after their axons have been severed and we have confirmed that their axons regenerate electrical connections with some of their usual synaptic targets in the central nervous system. Injections of horseradish peroxidase and Lucifer Yellow dye into separate T cells in unoperated animals showed that T cell axons typically run close to one another within single ganglia or from ganglion to ganglion. Knowledge of one T cell's arborizations thus revealed the groundplan of others in the same ganglia and the sites of apparent contact with its synaptic targets. For regenerating axons, those sprouts that encountered the normal pathway (as marked by homologous axons) grew preferentially along it. Despite the striking coincidence of old and new pathways, regenerated branching patterns within the ganglionic neuropils were usually incomplete and sometimes had atypical branches. Synaptic connections with normal targets (other T cells as well as S and C cells) were abnormally weak physiologically. The numbers of apparent contacts seen with the light microscope were also lower than normal. In addition, the strength of the synaptic potentials, normalized to the number of contacts (calculated as microvolts per contact), was generally smaller in the regenerated connections than in the controls, and smallest at earliest times, during the first 6 weeks following injury. It thus appears to be characteristic of T cell regeneration that axon regrowth is aided by the recognition of specific pathways and that successful regeneration, as assayed anatomically and physiologically, occurs frequently but usually incompletely.     

Progressive morphological abnormalities observed in rat spinal motor neurons at extended intervals after axonal regeneration.

It is generally accepted that mammalian spinal motor neurons return to normal after axotomy if their regenerated axons successfully reinnervate appropriate peripheral targets. However, morphological abnormalities, recently observed in spinal motor neurons examined 1 year after nerve crush injury, raise the possibility that delayed perikaryal changes occur after regeneration is complete. In order to distinguish between chronic and progressive alterations in neurons with long-term regenerated axons, rat spinal motor neurons and dorsal root ganglion cells were examined at 5 and 10 months following unilateral sciatic nerve crush. Neurons with regenerated axons were identified by retrograde labelling with horseradish peroxidase. The structural properties of neurons ipsilateral to nerve injury were compared to those of neurons from the spinal cord and dorsal root ganglia on the contralateral side and from age-matched control rats. At 5 months postcrush, the morphology of motor and sensory neurons ipsilateral to injury was comparable to that of control cells. However, several features of the motor neurons with regenerated axons distinguished them from control motor neurons at 10 months postcrush. Mean perikaryal area of ipsilateral spinal motor neurons was larger than the means for control motor neurons (p less than .001). Ipsilateral spinal motor neurons also appeared clustered within the spinal cord and had thicker dendrites. Dorsal root ganglion cells with regenerated axons were slightly larger than control cells at 10 months postcrush but they exhibited no other morphological changes. The present findings indicate that spinal motor neurons are progressively altered after their regenerated axons have reestablished functional synapses with their peripheral targets.     

Substance P-like immunoreactivity in neurons in dissociated cell cultures of mammalian spinal cord and dorsal root ganglia

Dissociated cell cultures prepared from fetal mouse spinal cords and dorsal root ganglia were stained for endogenous substance P using the peroxidase-antiperoxidase technique. Substance P-like immunoreactivity was localized within a small percentage of rounded or multipolar neuronal somata and in varicose processes. The substance P-positive multipolar neurons were derived from spinal cord, while the small rounded neurons were possibly of spinal cord and/or sensory ganglion origin. Large dorsal root ganglion neurons were unreactive. These results are consistent with in vivo findings and indicate the feasibility of electrophysiologic studies in culture to analyze the synaptic connections between substance P neurons and their target cells.     

Distribution and morphology of nociceptive cells in the CNS of three species of leeches

The present study describes the segmental variation in the distribution and morphology of nociceptive neurons (N cells) in the central nervous system of the leech. N cells of midbody ganglia can be segregated into lateral and medial types. We show that monoclonal antibodies specific for N cells can distinguish between the two populations. The monoclonal antibodies were used to map the complete distribution of the cells along the nervous cord. There are two pairs of the medial and lateral nociceptive neurons in the midbody ganglia, one pair of the medial type in the sex ganglia (5 and 6), and a pair of the lateral type in ganglia 20 and 21. The caudal brain is without nociceptive neurons. This distribution was confirmed by electrophysiological means. The morphology of N cells in different parts of the nervous system was investigated by intracellular horseradish peroxidase (HRP) injections. In the terminal segmental ganglia the N cells showed extensive arborizations in the head and tail brains and, contrary to N cells in the midbody ganglia, their arborizations spanned more than three segments. N cells are absent in the tail brain, but the N cells of ganglia 20 and 21 were shown to innervate the entire caudal region. The basic morphology of all N-cell homologues was found to be very similar for three leech species. In the sex ganglia the pair of N-cell homologues were examined in Haemopis, Hirudo, and Macrobdella. The results showed a progressive modification in the three species of the cell's morphology, peripheral projections, and physiological responses, possibly correlated with the evolution and complexity of the sexual organs. HRP injections and monoclonal antibody staining revealed that a common feature of N-cell homologues is the presence of processes that tightly surround the cell soma of other cells. This suggests that N cells may have other functional properties in addition to being primary sensory neurons.     

The afferent and sympathetic components of the lumbar spinal outflow to the colon and pelvic organs in the cat. I. The hypogastric nerve

The cell bodies of the lumbar sensory and sympathetic pre- and postganglionic neurons that project to the pelvic organs in the hypogastric nerve of the cat have been labeled retrogradely with horseradish peroxidase applied to the central end of their cut axons. The numbers, segmental distribution, location, and size of these labeled somata have been determined quantitatively. Afferent and preganglionic cell bodies were located bilaterally in dorsal root ganglia and spinal cord segments L3-L5, with the maximum numbers in L4. Very few cells lay rostral to L3. Afferent cell bodies were generally very small in cross-sectional area relative to the entire population in the dorsal root ganglia. Most of the preganglionic cell bodies lay clustered just medial to the region of the intermediolateral column and extended caudally well beyond its usual limit in the upper part of L4. These neurons were, on the average, larger than the cells of the intermediolateral column itself, with the largest cells lying in the most medial positions. Most of the post-ganglionic somata were in the ipsilateral distal lobe of the inferior mesenteric ganglion, while some (usually less than 10%) lay in accessory ganglia along the lumbar splanchnic nerves and in paravertebral ganglia L3-L5. Postganglionic somata in the inferior mesenteric ganglion were larger than both labeled and unlabeled ganglion cells in the paravertebral ganglia. From the data, it is estimated that about 1,300 afferent neurons, about 1,700 preganglionic neurons, and about 17,000 postganglionic neurons project in each hypogastric nerve in the cat.     

Substance P antibody reveals homologous neurons with axon terminals among somata in the crayfish and crab brain.

In the search for particular neurons that stain selectively and can be identified, the cerebral ganglia (brains) of the crayfish Cherax destructor and the crab Leptograpsus variegatus were immunocytochemically treated with a monoclonal antibody raised against substance P. Four large neurons in the cerebral ganglion of the crayfish and crab label selectively with a monoclonal antibody raised against substance P. Two of the large neurons have their cell bodies in the protocerebrum and two in the deutocerebrum in both animals. Each protocerebral cell in both animals projects through the ipsilateral and contralateral olfactory lobes to end among the lateral cell somata of the olfactory lobe and not in the neuropile. Electron micrographs show the presence of synapses within the cell somata area and on the cell somata themselves. Each deutocerebral cell in both animals projects only ipsilaterally and ends within the neuropile of the olfactory lobes. The immunoreactivity to substance P antibody and the shapes and the unique projections of the four cells suggest that they are homologous in the two species. Synaptic connections between axons and cell somata are rare in the arthropods but have been found on the Kenyon cells of the mushroom bodies of Limulus. This raises questions about homologies between the crustacean olfactory lobe and the mushroom bodies of Limulus and insects.     

Retinal ganglion cell axons regenerate in the presence of intact sensory fibres

A novel allograft paradigm was used to test whether adult mammalian central axons regenerate within a peripheral nerve environment containing intact sensory axons. Retinal ganglion cell axon regeneration was compared following anastomosis of dorsal root ganglia grafts or conventional peripheral nerve grafts to the adult rat optic nerve. Dorsal root ganglia grafts comprised intact sensory and degenerate motor axons, whereas conventional grafts comprised both degenerating sensory and motor axons. Retinal ganglion cell axons were traced after 2 months. Dorsal root ganglia survived with their axons persisting throughout the graft. Comparable numbers of retinal ganglion cells regenerated axons into both dorsal root ganglia (1053+/-223) and conventional grafts (1323+/-881; P>0.05). The results indicate that an intact sensory environment supports central axon regeneration.     

The morphological and physiological properties of a regenerating synapse in the C.N.S. of the leech.

Regeneration of an electrical synapse between particular interneurons in the medicinal leech was traced physiologically and morphologically using intracellular recording the horseradish peroxidase (HRP) injection. The synapse between S-cell interneurons lies in the connective midway between segmental ganglia, so crushing near one ganglion severs only one S-cell's axon. The severed distal stump remains connected to the adjacent uninjured S-cell and continues for weeks to conduct impulses. The injured cell regenerates, while its uninjured "target" neuron in the next ganglion does not grow. After the sprouts of the regenerating neuron cross the crush, one or a few branches grow along the surviving distal stump toward the original synapse. After about one month when the region of original synapse has been reached, regenerating neurons form electrical junctions and stop growing. Thereafter electrical coupling improves in stages. Within two months the regenerated neuron attains full caliber, the stump degenerates and function is normal. In some instances within days or weeks of crushing, the regenerating neuron forms a basket of synapses upon its severed distal stump and then continues growing to synapse with the target. When this occurs, electrical coupling and subsequent impulse transmission between S-cells rapidly resumes. These experiments indicated that the regenerating neuron is guided to its proper synaptic target by recognizing and following its severed distal stump. Sometimes the distal stump itself becomes an intermediate synaptic target.     

Morphological features of spiking and nonspiking cells in the paratracheal ganglion of the ferret

The present series of experiments was designed to study details of the morphology and connectivity of functionally identified cells located in the paratracheal ganglia of the ferret. The morphology of 11 spiking (AH cells) and seven nonspiking (type B cells) ganglion cells was examined. Intra-axonally injected horseradish peroxidase (HRP) was used as the label. Each spiking and nonspiking cell was identified by intracellular recording prior to the HRP injection. "Whole mount preparations" were processed for HRP histochemistry with diaminobenzidine as the chromogen. HRP-labeled cell bodies of both the spiking AH and nonspiking type B neurons demonstrated similar morphological features. Both types of ganglion cells showed axons arising from a small, ill-defined axon hillock which exited from the cell as single or multiple branches of equal diameter and coursed unidirectionally through the interganglionic nerve trunk to an adjacent ganglion; short, fine, tapering processes (presumptive dendrites) in the immediate vicinity of the injected cell; and processes extending out of the ganglion cell perpendicular to the interganglionic nerve trunk which could be followed into the smooth muscle. Extraperikaryal injections of HRP into a ganglion retrogradely labeled perikarya in the adjacent ganglia. These results demonstrate that in airway ganglia the morphology of spiking and nonspiking neurons is remarkably similar despite electrophysiological differences. In addition it appears that ganglion cells project to adjacent ganglia and to smooth muscle by means of independent axonal processes. These morphological features of the ganglion cells in airways and the trajectories of their axons correspond to known features of their physiology: i.e., the axon of a ganglion cell travels unidirectionally toward the adjacent ganglion and arborizes there, providing anatomical evidence of communication between ganglia via the interganglionic nerve trunk; and the spiking and nonspiking neurons possess similar morphological features that are typical of ganglion cells described in other systems, such as in the myenteric plexus.     

FMRF-amide-like substances in the leech. I. Immunocytochemical localization

FMRF-amide-like immunoreactivity (FLI) was localized to approximately 50 neurons in each segmental ganglion of the medicinal leech using immunocytochemical techniques. Although most of these neurons were iterated in each segmental ganglion, some were more restricted in their segmental distribution. The head and tail ganglia likewise contained numerous FMRF-amide-like immunoreactive cells. In addition to cell bodies, many nerve processes and varicosities were also immunoreactive throughout the ganglion. All labeling of FLI was blocked by preabsorption of the anti-FMRF-amide antiserum with synthetic FMRF-amide. Using a combination of Lucifer Yellow cellular injection and indirect immunofluorescence techniques, we identified several of the neurons possessing FLI. Identified neurons included excitatory motor neurons (HE, RPE, LPE, AE, and L), the HA modulatory neuron, interneuron cell 204, and cells of unknown function (AP). The processes of HE motor neurons and HA modulatory neurons which innervate the heart tubes were also immunoreactive. These results indicate a role for FMRF-amide-like substances as neurochemical signals in the leech.     

Projection patterns of posterior dorsal unpaired median neurons of the locust subesophageal ganglion

Six neurons in a group of dorsal unpaired median (DUM) neurons with cell bodies in the posterior part-maxillary and labial neuromeres-of the subesophageal ganglion of locusts have two axons each that descend into both the left and the right halves of the ganglia of the ventral nerve cord. None of the neurons has peripheral axons, so they are interneurons. Electrophysiology shows that the axons of at least four neurons project to the terminal abdominal ganglion to which they conduct spikes at a velocity of 0.5-0.6 m. second(-1). In the somata, the spikes have a smaller amplitude and briefer duration at half height than the spikes of thoracic, efferent DUM neurons. Each neuron has bilaterally symmetrical branches within the subesophageal ganglion and in the thoracic ganglia. On the basis of the specific patterns of branches, and the neuropiles, tracts, and commissures in which they occur, three types of neurons (DUM SD 1-3) can be recognized. DUM SD 1 and 3 project to ventral regions of neuropile in the thoracic ganglia in which the efferent DUM neurons of these ganglia have no branches. DUM SD 2 projects to dorsal neuropiles. The projection patterns of these putatively octopaminergic neurons suggest that they could be the source of the octopaminergic modulation of networks underlying sensory processing and motor pattern generation within these ganglia. Within this group of posterior DUM neurons, two additional cells were stained that have axons ascending to the brain.     

Neuronal organization and cell interactions of the cat stellate ganglion.

The functional structure of the cat stellate ganglion (SG) and, in particular, its extra- and intraganglionic connections and neuronal organization, were investigated using histochemical, immunohistochemical, morphological and histological methods. Retrograde axonal transport of horseradish peroxidase was used to determine most of the extraganglionic interactions. Of the targets tested, the most extensive efferent connections of the SG were with the stemocleidomastoid muscle, trachea, esophagus and heart. Neurons of the SG also send a small number of postganglionic efferents to the thyroid and stomach. Furthermore, ganglion cells send axons to the spinal ganglia. Several afferent connections of the SG were determined. Sympathetic preganglionic neurons of segments C8-T10 of the spinal cord, sensory neurons of C8-T9 spinal ganglia, intramural ganglia of the thoracic viscera and the reticular formation of the medulla oblongata send their axons to the SG. Intraganglionic interactions of intemeurons with principal ganglionic cells were assumed to occur, based on the presence of interneurons immunoreactive to GABA and substance P. GABA- and substance P-immunoreactive fibers located around a small number of postganglionic neurons were also identified. Morphological study revealed asymmetry between the left and right ganglia: the right ganglion is larger than the left and contains more cells. This asymmetry was also reflected in basic structural parameters of neurons, such as average neuronal area and average diameter of cell somata. The present data has been used to develop a scheme for the basic inter- and intraneuronal connections of the cat SG. This ganglion is a true nervous center, with postganglionic neurons, some of which might be performing sensory functions, and interneurons. The ganglion is connected not only with the spinal cord and spinal ganglia, but also with neurons of the intramural ganglia and, by direct links, with efferent neurons of the medulla oblongata. Thus, the SG may play an essential role in viscera-visceral reflexes.     

Generation of new cerebral ganglion neurons in the snail Melampus: An ultrastructural study

Reports in the literature have established that reconnection of central neural tracts occurs following commissurotomy and cerebral ganglion excision in the primitive pulmonate snail Melampus bidentatus and have suggested the possibility that long-term regeneration might result in the appearance of new neurons in the ganglion bud. We have used electron microscopy to examine the ganglion buds that form by reconnection of cerebral nerves, commissure, and connectives following cerebral ganglion excision in adult Melampus. The buds were examined from 2.5 to 12 months postoperatively. By 2.5 months, ganglion buds consist of a mixture of axon tracts that travel through the bud region and some dendritic processes; a few synaptic contacts can be identified at this stage, scattered throughout the bud. By 5–6 months, some of the most advanced ganglia have undifferentiated cells that are distinct from glia. By 7 months, differentiated neurons with clear, small dense-core or neurosecretory vesicles are present, although these cells are not all concentrated in a rind on the ganglion surface. Another cell type, the pigment-sheath cell, is present by this stage. By 11–12 months, the most advanced regenerating ganglia have neurons which form a cell rind on the ganglion surface. The gross appearance of a regenerated ganglion at this stage is similar to that of the intact contralateral cerebral ganglion, although the regenerated ganglion is smaller. One 12-month ganglion was found to possess fairly normal intraganglionic morphology, with lobes and cell types that were recognizable. Hence, nerve cell regeneration can occur in the absence of body part regeneration in adult members of one species of pulmonate snail.     

Regenerating afferent fibers stimulate the recovery of mauthner cell dendritic branching in the axolotl

In the medulla of the axolotl (Ambystoma mexicanum), Mauthner cells (M-cells) occur as a pair of large identifiable neurons at the level of entry of the vestibular nerve (nVIII). Each receives synapses from the ipsilateral nVIII; the terminals can be identified as club endings and are restricted to a specific set of M-cell dendritic branches. We have examined these branches for morphologic changes following long-term deafferentation in the presence and absence of nerve regeneration. Deafferentation was brought about in a group of young larvae by unilaterally severing nVIII. The nerve was allowed to regenerate in half of the larvae. In those remaining, the nVIII ganglion was damaged to preclude or limit nerve regeneration. The contralateral side served as control. After 3 months survival, the larvae were killed and the medullae prepared for microscopy. To estimate the extent of nerve regeneration, axons in the experimental nVIII tract were counted and compared with the number in the control. The mean number of axons in the nVIII tract ipsilateral to intact ganglia indicated that 69% of the fibers had regenerated. In contrast, only 31% regenerated in larvae with damaged ganglia. Electron microscopic analysis of selected sections revealed that the mean number of nVIII terminals per section through M-cells ipsilateral to destroyed ganglia was significantly less than the mean number in analogous sections through either control cells or cells ipsilateral to intact ganglia. Control and experimental M-cells were reconstructed from serial sections. Deprived M-cells had significantly reduced dendritic branching patterns in the region that normally receives nVIII input. On the other hand, the extent of branching on cells receiving regenerated afferents from intact ganglia was like that of their contralateral controls. The distribution of dendritic branches on many reinnervated M-cells, however, was broader than that on control cells. Electron microscopic examination of the displaced dendritic branches (those extending into adjacent tracts) revealed that they received vestibular synapses. Thus, in some animals, regenerated vestibular fibers were not restricted to the nVIII tract. Deafferentation of the M-cells results in a reduction of dendritic branches in the region deprived of vestibular contacts.(ABSTRACT TRUNCATED AT 400 WORDS)     

Secretoneurin-like immunoreactivity in rat sympathetic, enteric and sensory ganglia

Distribution of secretoneurin-like immunoreactivity (SN-LI) was studied in the rat sympathetic ganglia/adrenal gland, enteric and sensory ganglia by immunohistochemical methods. SN-LI nerve fibers formed basket-like terminals surrounding many of the postganglionic neurons of the superior cervical, stellate, paravertebral chain ganglia, coeliac/superior mesenteric and inferior mesenteric ganglia. Postganglionic neurons of the superior cervical and other sympathetic ganglia exhibited low-to-moderate levels of SN-LI. In all these sympathetic ganglia, clusters of small diameter (<10 μm) cells, which may correspond to the small intensely fluorescent (SIF) cells, were found to be intensely labeled. Surgical sectioning or ligation of the cervical sympathetic trunk for 7–10 days resulted in a nearly total loss of SN-LI fibers in the superior cervical ganglia, whereas immunoreactivity in the postganglionic neurons and small diameter cells remained essentially unchanged. In the thoracolumbar and sacral segments of the spinal cord, SN-LI nerve fibers were detected in the superficial layers of the dorsal horn as well as in the intermediolateral cell column (IL_p). Occasionally, SN-LI somata were noted in the IL_p. SN-LI nerve fibers formed a delicate plexus underneath the capsule of the adrenal gland, some of which traversed the adrenal cortex and reached the adrenal medulla. While heavily invested with SN-LI nerve terminals, chromaffin cells seemed to express a low level of SN-LI. In the enteric plexus, varicose SN-LI nerve fibers and terminals formed a pericellular network around many myenteric and submucous ganglion cells; the ganglionic neurons were lightly to moderately labeled. A population of ganglion cells in the dorsal root, nodose and trigeminal ganglia exhibited moderate-to-strong SN-LI. The detection of SN-LI in nerve fibers and somata of various sympathetic ganglia, enteric plexus and adrenal medulla and in somata of the sensory ganglia implies an extensive involvement of this peptide in sympathetic, enteric and sensory signal processing.     

Substance P in the vagal sensory ganglia: Localization in cell bodies and pericellular arborizations

The distribution of Substance P-like immunoreactivity in the jugular and nodose ganglia of rabbits and pigeons has been studied using immunocytochemical staining techniques. Substance P-like immunoreactivity is localized to neuronal cell bodies and processes in the jugular and nodose ganglia, and to pericellular fiber plexi in the nodose ganglia of both species. The numbers and sizes of cells which exhibited Substance P-like immunoreactivity in each ganglion were determined using quantitative morphometric techniques. The distribution of Substance P-like immunoreactivity in the rabbit and pigeon vagal sensory ganglia is characterized by several general features. In most of the ganglia, immunoreactive neurons factor into discrete types which can be distinguished from one another, and from non-immunoreactive neurons, by size. In addition, immunoreactive nodose and jugular ganglion cells, respectively, are distinguishable on the basis of size. Finally, a considerably higher percentage of immunoreactive neurons is found in the jugular ganglion than in the nodose ganglion. Substance P-like immunoreactivity was also seen in pericellular fiber plexi which encircle individual neurons in the nodose ganglion of rabbits and pigeons. These plexi are composed of varicose fibers which appear to terminate as boutons on the surfaces of the cells which they encircle. The distribution of Substance P-like immunoreactivity within the vagal sensory ganglia is discussed with respect to the possible peripheral targets and functions of Substance P-containing vagal afferents. Our findings suggest that Substance P-containing vagal sensory neurons are involved in a variety of visceral and somatic afferent functions.     

Novel,secondary sensory cell organ in ascidians: in search of the ancestor of the vertebrate lateral line

A new mechanoreceptor organ, the "coronal organ," located in the oral siphon, is described by light and electron microscopy in the colonial ascidians Botryllus schlosseri and Botrylloides violaceus. It is composed of a line of sensory cells (hair cells), accompanied by supporting cells, that runs continuously along the margin of the velum and tentacles of the siphon. These hair cells resemble those of the vertebrate lateral line or, in general, the acoustico-lateralis system, because they bear a single cilium, located centrally or eccentrically to a hair bundle of numerous stereovilli. In contrast to other sensory cells of ascidians, the coronal hair cells are secondary sensory cells, since they lack axonal processes directed towards the cerebral ganglion. Moreover, at their base they form synapses with nerve fibers, most of which exhibit acetylcholinesterase activity. The absence of axonal extensions was confirmed by experiments with lipophilic dyes. Different kinds of synapses were recognized: usually, each hair cell forms a few afferent synapses with dendrites of neurons located in the ganglion; efferent synapses, both axo-somatic (between an axon coming from the ganglion and the hair cell) and axo-dendritic (between an axon coming from the ganglion and an afferent fiber) were occasionally found. The presence of secondary sensory cells in ascidians is discussed in relation to the evolution of sensory cells and placodes in vertebrates. It is proposed that the coronal organ in urochordates is homologous to the vertebrate acoustico-lateralis system.