New vistas on the initiation and maintenance of insect motor behaviors revealed by specific lesions of the head ganglia

In insects, thoracic pattern generators are modulated by the two head ganglia, the supraesophageal ganglion (brain) and the subesophageal ganglion, which act as higher-order neuronal centers. To explore the contribution of each head ganglion to the initiation and maintenance of specific motor behaviors in cockroaches (Periplaneta americana), we performed specific lesions to remove descending inputs from either the brain or the subesophageal ganglion or both, and quantified the behavioral outcome with a battery of motor tasks. We show that ‘emergency’ behaviors, such as escape, flight, swimming or righting, are initiated at the thoracic level independently of descending inputs from the head ganglia. Yet, the head ganglia play a major role in maintaining these reflexively initiated behaviors. By separately removing each of the two head ganglia, we show that the brain excites flight behavior and inhibits walking-related behaviors, whereas the subesophageal ganglion exerts the opposite effects. Thus, control over specific motor behaviors in cockroaches is anatomically and functionally compartmentalized. We propose a comprehensive model in which the relative permissive versus inhibitory inputs descending from the two head ganglia, combined with thoracic afferent sensory inputs, select a specific thoracic motor pattern while preventing the others.     

Control of leech swimming activity by the cephalic ganglia

We investigated the role played by the cephalic nervous system in the control of swimming activity in the leech, Hirudo medicinalis, by comparing swimming activity in isolated leech nerve cords that included the head ganglia (supra- and subesophageal ganglia) with swimming activity in nerve cords from which these ganglia were removed. We found that the presence of these cephalic ganglia had an inhibitory influence on the reliability with which stimulation of peripheral (DP) nerves and intracellular stimulation of swim-initiating neurons initiated and maintained swimming activity. In addition, swimming activity recorded from both oscillator and motor neurons in preparations that included head ganglia frequently exhibited irregular bursting patterns consisting of missed, weak, or sustained bursts. Removal of the two head ganglia as well as the first segmental ganglion eliminated this irregular activity pattern. We also identified a pair of rhythmically active interneurons, SRN1, in the subesophageal ganglion that, when depolarized, could reset the swimming rhythm. Thus the cephalic ganglia and first segmental ganglion of the leech nerve cord are capable of exerting a tonic inhibitory influence as well as a modulatory effect on swimming activity in the segmental nerve cord.     

Role of central and peripheral mechanisms in control of excitability of segmental motor centers in insects

The role of central and peripheral mechanisms in control of excitability of segmental centers providing different motor acts in insects of phylogenic close orders, but differing by the level of activity of their locomotor systems has been studied in the locust Locusta migratoria and the cockroach Periplaneta americana. It was shown that the level of relative excitability of segmental centers in cockroaches seemed to be much determined by the peripheral mechanisms, but not by the central mechanisms as in locust. It is suggested that control of activity of segmental locomotor centers from the higher parts of CNS can be realized by different ways: predominantly via excitatory or inhibitory influences on activity id some particular locomotor systems depending on their role and significance in motor behavior of these animals.Translated from Zhurnal Evolyutsionnoi Biokhimii i Fiziologii, Vol. 40, No. 6, 2004, pp. 508–513.Original Russian Text Copyright © 2004 by Gorelkin, Severina.To the 100-Anniversary of A. K. VoskresenskayaThis revised version was published online in April 2005 with a corrected cover date.     

Initiation of swimming activity by trigger neurons in the leech subesophageal ganglion

In papers I and II of this series, we described two pairs of interneurons, Tr1 and Tr2, in the leech subesophageal ganglion which can trigger swimming activity in the isolated central nervous system (CNS). In this paper, we describe sensory inputs to these trigger neurons from previously identified mechanosensory neurons. We found that: Weak mechanical stimulation (stroking) of a body wall flap attached to a segmental ganglion in an otherwise isolated CNS excites the contralateral Tr1 slightly. Strong mechanical stimulation (pinching) of a mid-body wall flap evokes a burst of impulses in the contralateral Tr1. For both means of stimulation the effects on the ipsilateral Tr1 and on the Tr2 cell pair were much weaker. Stroking a body wall flap attached to the head ganglion (supra- and subesophageal ganglia) in an otherwise isolated CNS excites both Tr1s and both Tr2s, although the effect is weaker for the Tr2s. Pinching strongly excites both trigger neurons bilaterally. Pressure and nociceptive mechanosensory neurons (P and N cells) in the subesophageal ganglion and the first segmental ganglion appear to make direct excitatory synapses with the contralateral Tr1 and Tr2. Mechanosensory interactions with the ipsilateral trigger neurons appear to be indirect. Functional inactivation of Tr1 by hyperpolarization does not prevent swim initiation either by weak mechanical stimulation of a body-wall flap or by intracellular stimulation of P cells.2+ We conclude that the trigger neurons, Tr1 and Tr2, provide an excitatory pathway by which mechanosensory stimulation can initiate leech swimming activity.(ABSTRACT TRUNCATED AT 250 WORDS)     

Initiation of swimming activity by trigger neurons in the leech subesophageal ganglion

The aim of this study was to identify neurons in the subesophageal ganglion of the medicinal leech which initiate swimming activity and to determine their output connections. We found two bilaterally symmetrical pairs of interneurons, Tr1 and Tr2, located in the first division of the subesophageal ganglion which initiate swimming activity in the isolated nervous system when depolarized with brief (1-3 s) current pulses. Tr1 and Tr2 are considered trigger neurons because elicited swimming episodes outlast the stimulus duration, and because the length of elicited swim episodes is nearly independent of the intensity with which Tr1 and Tr2 are stimulated. Tr1 and Tr2 have similar morphologies. The neurites of both cells cross contralaterally in the subesophageal ganglion, project posteriorly, and exit the subesophageal ganglion in the contralateral connective. The axons of Tr1 and Tr2 extend as far posterior as segmental ganglion 18 of the ventral nerve cord. Tr1 provides direct excitatory drive to three groups of segmental neurons which are capable of initiating swimming: swim-initiating interneurons (cells 204 and 205), serotonin-containing interneurons (cells 61 and 21), and the serotonergic Retzius cells. In addition, all Retzius cells in the subesophageal ganglion are excited directly by Tr1. These three groups of neurons are excited even if Tr1 stimulation is subthreshold for swim initiation. In contrast to Tr1, Tr2 stimulation evokes transient inhibition in swim-initiating and serotonin-containing interneurons, and has little immediate effect on Retzius cells. In addition, Tr2 indirectly inhibits several oscillator neurons, including cells 208, 33, and 60. When Tr1 is stimulated during a swimming episode the swim period decreases for several cycles, while stimulation of Tr2 during swimming episodes reliably resets the ongoing swimming rhythm. Our findings indicate that Tr1 and Tr2 are trigger neurons which initiate swimming activity by different pathways. These neurons also have functional interactions with the swim oscillator network since either Tr1 or Tr2 stimulation during swimming can modulate the ongoing swimming rhythm.     

Octopaminergic modulation of locust motor neurones

The effect of octopamine on the fast extensor and the flexor tibiae motor neurones in the locust (Schistocerca gregaria) metathoracic ganglion, and also on synaptic transmission from the fast extensor to the flexor motor neurones, was examined. Bath application or ionophoresis of octopamine depolarized and increased the excitability of the flexor tibiae motor neurones. 1 mM octopamine reduced the amplitude of the fast extensor-evoked EPSP in the slow but not the fast flexor motor neurones, whereas 10 mM octopamine could reduce the EPSP amplitude in both. Octopamine broadened the fast extensor action potential and reduced the amplitude of the afterhyperpolarization, the modulation requiring feedback resulting from movement of the tibia. Octopamine also increased the frequency of synaptic inputs onto the tibial motor neurones, and could cause rhythmic activity in the flexor motor neurones, and reciprocal activity in flexor and extensor motor neurones. Octopamine also increased the frequency of spontaneous spiking in the octopaminergic dorsal unpaired median neurones. Repetitive stimulation of unidentified dorsal unpaired median neurones could mimic some of the effects of octopamine. However, no synaptic connections were found between dorsal unpaired median neurones and the tibial motor neurones. The diverse effects of octopamine support its role in mediating arousal.     

The caudal ganglion of the leech,with particular reference to homologues of segmental touch receptors

The caudal ganglion of the leech, which provides sensory and motor innervation to the posterior sucker, represents the fusion of seven embryonic segmental ganglia. Although fused, each of the seven contributing ganglia (“subganglia”) of the caudal ganglion can be distinguished morphologically and functionally. The roots from each subganglion carry the axons of mechanoreceptors homologous to “touch” cells found in the segmental ganglia and the subesophageal compound ganglion. The receptive fields supplied by the touch cells of the caudal ganglion are uniquely arranged and reveal the modified segmentation of the circular posterior sucker. Extensive overlap of sensory innervation occurs between adjacent segments of the sucker, beyond the overlap characteristic of the homologous cells of body segments. It thus appears that the touch receptors of the caudal ganglion are less restricted than receptors of the segmental ganglia with regard to their territories of innervation. The caudal ganglion has additional unique properties that establish it as a distinct integrative center of the leech CNS.     

Physiological investigation on the role of the innervation in the regulation of the prothoracic gland in Periplaneta americana

Under in vitro conditions the prothoracic gland nerve of the last larval instar of Periplaneta americana shows the same efferent nervous activity as under in situ conditions–ie, low activity at the 9th day and high activity at the 20th day of the molting interval. Isolation of the prothoracic ganglion from the subesophageal ganglion provokes an increase in this nerve activity, suggesting an inhibitory effect of the subesophageal ganglion on prothoracic gland nerve activity in vivo. Only in 20-day-old larvae does electrical stimulation of isolated prothoracic glands in vitro via the gland nerve result in a slightly increased release of ecdysteroids from the gland. This effect could not be influenced by different lengths of stimulation periods. Denervation of the prothoracic gland by transection of the gland nerve on the 13th day of the molting interval results in a complete abolition of the first peak of ecdysteroid production in the gland but has no influence on the occurrence and the amount of the main ecdysteroid peak just before the molt. The results suggest the participation of nervous activity in special periods of prothoracic gland regulation in the cockroach.     

Glutamate-like immunoreactivity in the leech central nervous system.

Using a monoclonal antibody for glutamate the distribution was determined of glutamate-like immunoreactive neurons in the leech central nervous system (CNS). Glutamate-like immunoreactive neurons (GINs) were found to be localized to the anterior portion of the leech CNS: in the first segmental ganglion and in the subesophageal ganglion. Exactly five pairs of GINs consistently reacted with the glutamate antibody. Two medial pairs of GINs were located in the subesophageal ganglion and shared several morphological characteristics with two medial pairs of GINs in the first segmental ganglion. An additional lateral pair of GINs was also located in segmental ganglion 1. A pair of glutamate-like immunoreactive neurons, which are potential homologs of the lateral pair of GINs in segmental ganglion 1, were occasionally observed in more posterior segmental ganglia along with a selective group of neuronal processes. Thus only a small, localized population of neurons in the leech CNS appears to use glutamate as their neurotransmitter.     

Glutamate-like immunoreactivity in the leech central nervous system

Summary Using a monoclonal antibody for glutamate the distribution was determined of glutamate-like immunoreactive neurons in the leech central nervous system (CNS). Glutamate-like immunoreactive neurons (GINs) were found to be localized to the anterior portion of the leech CNS: in the first segmental ganglion and in the subesophageal ganglion. Exactly five pairs of GINs consistently reacted with the glutamate antibody. Two medial pairs of GINs were located in the subesophageal ganglion and shared several morphological characteristics with two medial pairs of GINs in the first segmental ganglion. An additional lateral pair of GINs was also located in segmental ganglion 1. A pair of glutamate-like immunoreactive neurons, which are potential homologs of the lateral pair of GINs in segmental ganglion 1, were occasionally observed in more posterior segmental ganglia along with a selective group of neuronal processes. Thus only a small, localized population of neurons in the leech CNS appears to use glutamate as their neurotransmitter.     

Localization of the site of effect of a wasp's venom in the cockroach escape circuitry

The parasitic wasp Ampulex compressa stings a cockroach Periplaneta americana in the neck, toward the head ganglia (the brain and subesophageal ganglion). In the present study, our aim was to identify the head ganglion that is the target of the venom and the mechanisms by which the venom blocks the thoracic portion of the escape neuronal circuitry. Because the escape responses elicited by a wind stimulus in brainless and sham-operated animals were similar, we propose that the venom effect is on the subesophageal ganglion. Apparently, the subesophageal ganglion modulates the thoracic portion of the escape circuit. Recordings of thoracic interneuron responses to the input from the abdominal giant interneurons showed that the thoracic interneurons receive synaptic drive from these interneurons in control and in stung animals. Unlike normal cockroaches, which use both fast and slow motoneurons for producing rapid escape movements, stung animals activate only the slow motoneuron. However, we show that in stung animals, the fast motoneuron still can be recruited with bath application of pilocarpine, a muscarinic agonist. These results indicate that the descending control from the subesophageal ganglion is presumably exerted on the premotor thoracic interneurons to motoneurons connection of the thoracic escape circuitry. Accepted: 19 December 1998     

State-dependent responses of two motor systems in the crayfish,Pacifastacus leniusculus

The expression of both swimmeret and postural motor patterns in crayfish (Pacifastacus leniusculus) were affected by stimulation of a second root of a thoracic ganglion. The response of the swimmeret system depended on the state of the postural system. In most cases, the response of the swimmeret system outlasted the stimulus.Stimulation of a thoracic second root also elicited coordinated responses from the postural system, that outlasted the stimulus. In different preparations, either the flexor excitor motor neurones or the extensor excitor motor neurones were excited by this stimulation. In every case, excitation of one set of motor neurones was accompanied by inhibition of that group's functional antagonists.This stimulation seemed to coordinate the activity of both systems; when stimulation inhibited the flexor motor neurones, then the extensor motor neurones and the swimmeret system were excited. When stimulation excited the flexor motor neurones, then the extensor motor neurones and the swimmeret system were inhibited.Two classes of interneurones that responded to stimulation of a thoracic second root were encountered in the first abdominal ganglion. These interneurones could be the pathway that coordinates the response of the postural and swimmeret systems to stimulation of a thoracic second root.Abbreviations TSR thoracic second root - epsp excitatory post-synaptic potential - ipsp inhibitory post-synaptic potential - EJP excitatory jonctional potential - PS power-stroke - RS return-stroke - INT interneurone - N1 first segmental nerve - N2 second segmental nerve - N3 third segmental nerve - A1 abdominal ganglion 1     

Flight motor patterns recorded in surgically isolated sections of the ventral nerve cord ofLocusta migratoria

Summary In the locust,Locusta migratoria, the pairs of connectives between the three thoracic ganglia and in the neck were transected in all possible combinations. Each of these preparations was tested for the production of rhythmic flight motor activity, with sensory input from the wing receptors intact and after deafferentation. The motor activity elicited in these preparations was characterized by intracellular recordings from motoneurons and electromyographic analyses.The motor patterns observed in locusts with either the neck or the pro-mesothoracic connectives severed (Figs. 2, 3, and 4) were very similar to the flight motor pattern produced by animals with intact connectives. The activity recorded in mesothoracic flight motoneurons of locusts with either only the meso-metathoracic connectives cut or both the meso-metathoracic and the neck connectives transected were similar to each other. Rhythmic motor activity could be observed in these preparations only as long as sensory feedback from the wing receptors was intact. These patterns were significantly different from the intact motor pattern (Figs. 5, 6, and 7). Similar results were obtained when the mesothoracic ganglion was isolated from the other two thoracic ganglia, although the oscillations produced under these conditions were weak (Fig. 8 upper). In the isolated metathorax no rhythmic flight motor activity could be recorded (Fig. 8 lower), even when wing afferents were intact.Considering the differences between the motor patterns observed in the various preparations these results suggest that the ganglia of the locust ventral nerve cord do not contain segmental, homologous flight oscillators which are coupled to produce the intact flight rhythm. Instead they support the idea that the functional flight oscillator network is distributed throughout the thoracic ganglia (Robertson and Pearson 1984). The results also provide further evidence that sensory feedback from the wing sense organs is necessary for establishing the correct motor pattern in the intact animal (Wendler 1974, 1983; Pearson 1985; Wolf and Pearson 1987 a).Abbreviations CPG central pattern generator - EMG electromyogram     

Musculature and innervation of the neck of the desert locust,Schistocerca gregaria (Forskål)

The innervation of each of the muscles involved in mediating head movement in the desert locust Schistocerca gregaria is described in detail. The number of motor neurones to each muscle and the neutral pathway and ganglion of origin of each are deduced from both histological and electrophysiological evidence. Only two of the muscles are, on histological evidence, innervated by as few as four different neurones, while several receive more than ten, and one at least 13. Individual muscles are shown physiologically to receive, in a few cases, as many as six different motor neurones. At least six muscles are innervated by motor neurones originating in more than one ganglion. One group of four muscles consisting in total of less than 100 muscle fibres receives more than 20 different motor neurones from three different ganglia through three or four different nerve roots. In these muscles, many single muscle fibres receive innervation from at least two different ganglia. It is concluded that the segmental nature of an insect muscle can not be deduced solely from a knowledge of the ganglion of origin of the motor innervation to that muscle. The innervation patterns that exist today must reflect past evolutionary development, but changes in the peripheral distribution of motor neurones, or migration of motor neurone cell bodies from one ganglion to another, or the development of additional motor neurones, or several of these factors together, must have formed a part of that development.     

Neuromodulation for behavior in the locust frontal ganglion

Neuromodulators orchestrate complex behavioral routines by their multiple and combined effects on the nervous system. In the desert locust, Schistocerca gregaria, frontal ganglion neurons innervate foregut dilator muscles and play a key role in the control of foregut motor patterns. To further investigate the role of the frontal ganglion in locust behavior, we currently focus on the frontal ganglion central pattern generator as a target for neuromodulation. Application of octopamine, a well-studied insect neuromodulator, generated reversible disruption of frontal ganglion rhythmic activity. The threshold for the modulatory effects of octopamine was 10^–6 mol l^–1, and 10^–4 mol l^–1 always abolished the ongoing rhythm. In contrast to this straightforward modulation, allatostatin, previously reported to be a myoinhibitor of insect gut muscles, showed complex, tri-modal, dose-dependent effects on frontal ganglion rhythmic pattern. Using a novel cross-correlation analysis technique, we show that different allatostatin concentrations have very different effects not only on cycle period but also on temporal characteristics of the rhythmic bursts of action potentials. Allatostatin also altered the frontal ganglion rhythm in vivo. The analysis technique we introduce may be instrumental in the study of not fully characterized neural circuits and their modulation. The physiological significance of our results and the role of the modulators in locust behavior are discussed.Abbreviation CPG central pattern generator - FG frontal ganglion - JH juvenile hormone - STNS stomatogastric nervous system     

The locust frontal ganglion: a multi-tasked central pattern generator

The locust frontal ganglion (FG) constitutes a major source of innervation to the foregut dilator muscles and thus plays a key role in control of foregut movements. This paper reviews our recent studies on the generation and characteristics of FG motor outputs in two distinct and fundamental locust behaviors: feeding and molting. In an in vitro preparation, isolated from all descending and sensory inputs, the FG was spontaneously active and generated rhythmic multi-unit bursts of action potentials, which could be recorded from all efferent nerves. Thus the FG motor pattern is generated by a central pattern generator within the ganglion. Intracellular recordings suggest that only a small fraction (10-20%) of the FG 100 neurons demonstrate rhythmic activity. The FG motor output in vivo was relatively complex, and strongly dependent on the locust's physiological and behavioral state. Rhythmic activity of the foregut was found to depend on the amount of food present in the crop; animals with full crop demonstrated higher FG burst frequency than those with empty crop. At the molt, the FG generates a distinct motor pattern that could be related to air-swallowing behavior.     

FMRFamide-like immunocytochemistry in the brain and subesophageal ganglion of <Emphasis Type="Italic">Triatoma infestans</Emphasis> (Insecta: Heteroptera). Coexpression with β-pigment-dispersing hormone and small cardioactive peptide B

The distribution of FMRFamide (FMRFa)-like immunoreactivity (LI) was studied in the brain and subesophageal ganglion of Triatoma infestans, the insect vector of Chagas disease. The neuropeptide displayed a widespread distribution with immunostained somata in the optic lobe, in the anterior, lateral, and posterior soma rinds of the protocerebrum, and around the antennal sensory and mechanosensory and motor neuropils of the deutocerebrum. FMRFa-immunoreactive profiles of the subesophageal ganglion were seen in the mandibular, maxillary, and labial neuromeres. Immunostained neurites were detected in the medulla and lobula of the optic lobe, the lateral protocerebral neuropil, the median bundle, the calyces and the stalk of the mushroom bodies, and the central body. In the deutocerebrum, the sensory glomeruli showed a higher density of immunoreactive processes than the mechanosensory and motor neuropil, whereas the neuropils of each neuromere of the subesophageal ganglion displayed a moderate density of immunoreactive neurites. Colocalization of FMRFa-LI and crustacean pigment-dispersing hormone-LI was found in perikarya of the proximal optic lobe, the lobula, the sensory deutocerebrum, and the labial neuromere of the subesophageal ganglion. The distribution pattern of small cardioactive peptide B (SCP_B)-LI was also widespread, with immunolabeled somata surrounding every neuropil region of the brain and subesophageal ganglion, except for the optic lobe. FMRFa- and SCP_B-LIs showed extensive colocalization in the brain of this triatomine species. The presence of immunolabeled perikarya displaying either FMRFa- or SCP_B-LI confirmed that each antisera identified different peptide molecules. The distribution of FMRFa immunostaining in T. infestans raises the possibility that FMRFa plays a role in the regulation of circadian rhythmicity. The finding of immunolabeling in neurosecretory somata of the protocerebrum suggests that this neuropeptide may also act as a neurohormone.This work was sponsored by the Facultad de Ciencias Biomédicas, Universidad Austral. Part of this work was performed at the Division of Neurobiology, Arizona Research Laboratories (Tucson, Arizona) with the support of a Fulbright Research Award to B.P.S.     

The nervous system of the tube-worm Chaetopterus variopedatus (Polychaeta)

Silver impregnation of serial histological sections of the tubeworm Chaetopterus variopedatus revealed the presence of a subepidermal nervous system. The anterior nervous system is delimited by the first 11 segments and comprises (1) two dorsolateral cerebral ganglia and lateral instead of ventral nerve cords which are widely separated and thus connected by unusually long commissures, (2) a pharyngeal ganglion in the fourth segment which is connected to the cerebral ganglia by pharyngeal nerves and constitutes along with the pharyngeal plexus a stomatogastric or enteric nervous system, and (3) small, presumably segmental ganglionic swellings along the lateral nerve cords from which emerge commissures and parapodial nerves. No subesophageal ganglion or periesophageal connective could be identified. The lateral nerve cords converge toward the midline in the 12th segment to form the posterior nervous system comprising a pair of ventromedian nerve cords with their repetitive segmental ganglia from which emerge numerous short commissures and three segmental nerves coursing toward the dorsal and ventral regions of parapods and toward the neuropod. Light and electron microscopic investigations of cerebral and segmental ganglia showed an arrangement of inner neuropile and of unipolar neuron somata at the periphery. The neuropile comprises numerous neurites ranging in diameter from 0.5 to 10 μm and making polarized or symmetrical synaptic junctions with each other. The pharyngeal ganglion consists of a similar neuropile and of a large mass of cell bodies which is traversed by an elaborate network of sinuses and harbors three types of neurosecretory cells in addition to the conventional neuron somata. These findings are interpreted in the framework of the highly specialized morphological features and habits of Chaetopterus, and the welldeveloped stomatogastric system is considered to be related to control of the feeding activities.     

Neurosecretion in the scorpion Heterometrus swammerdami

Gross morphology, staining characteristics and mapping of the diversity of the neurosecretory cell types in the brain and subesophageal ganglion of the scorpion Heterometrus swammerdami are reported. Special neurosecretory cell groups whose product is stainable with orange-G, acid fuchsin and Heidenhain's hematoxylin are present in the brain. In many of the living isolated neurosecretory cells, the secretory material appears luminous when viewed with dark ground illumination and granular when observed with phase contrast microscope. In the subesophageal ganglion the metameric arrangement of neurosecretory cells is distinct. Neurosecretory product accumulating in specific regions of subesophageal ganglion, and its axonal transport into the dorsal nerves and their termination in cephalic blood vessels apparently representing a storage and release organ of neurosecretion is reported.     

Termination of leech swimming activity by a previously identified swim trigger neuron

Cell Tr2 is a neuron in the subesophageal ganglion of the leech that can trigger swim episodes. In this report, we describe the ability of Tr2 to terminate ongoing swim episodes as well as to trigger swimming. Stimulation of Tr2 terminated ongoing swim episodes in nearly every preparation tested, while Tr2 stimulation triggered swim episodes in only a minority of the preparations. We suggest that the primary role of Tr2 is in the termination rather than the initiation of swimming activity.The swim trigger neuron Tr3 and a swim-gating neuron, cell 21, hyperpolarized during Tr2-induced swim termination. Another swim-gating neuron, cell 204 was sometimes slightly excited, but more often, hyperpolarized during Tr2-induced swim termination. In contrast to these cells, Tr2 stimulation excited another swim-gating neuron, cell 61. The responses of the swimgating cells were variable in amplitude and sometimes not evident during Tr2-induced swim termination. Hence, the effects of Tr2 stimulation on swim-gating neurons seem unlikely to be the direct cause of swim termination.Oscillator cells examined during Tr2-induced swim termination include: 27, 28, 33, 60, 115, and 208. The largest effect seen in an oscillator neuron was in cell 208, which was repolarized by up to 10 mV during Tr2 stimulation. Tr2 stimulation did not produce any obvious synaptic effects in motor neurons DI-1, VI-1, and DE-3. Our findings indicate that other, yet undiscovered, connections are likely to be important in Tr2-induced swim termination. Therefore, we propose that cell Tr2 is probably a member of a distributed neural network involved in swim termination.Abbreviations DP dorsal posterior nerve - Mx midbody ganglion x - Rx neuromere x of the subsesophageal (rostral) ganglion - DE dorsal excitatory motor neuron - DI dorsal inhibitory motor neuron - VI ventral inhibitory motor neuron