Organization of motor neurons innervating the proboscis musculature in Drosophila melanogaster meigen (diptera : Drosophilidae)

The present study addresses the question as to how the motor neurons involved in feeding in Drosophila melanogaster Meigen (Diptera : Drosophilidae) are organized. The motor neurons have been visualized both by Golgi-silver impregnation and by intramuscular injection of horseradish peroxidase, and analyzed in light of the existing information on taste sensory system and the feeding behaviour. The motor neurons have been broadly classified into the following types: labial nerve motor neurons, pharyngeal nerve motor neurons, and accessory pharyngeal nerve motor neurons, depending on the nerve through which their axons exit. The arborization of all the motor neurons is confined to the suboesophageal ganglion (SOG). All of them have predominantly ipsilateral and some contralateral arborizations. Their dendrites predominantly occupy the ventral region of the neuropil of the SOG and partially overlap the taste sensory projections, thereby providing an opportunity for interaction with the taste sensory input. The pharyngeal motor neurons arborize more extensively in the ventral tritocerebram, anteroventral. and mid-ventral neuropil, whereas the dendritic fields of labial motor neurons are confined to the mid-ventral neuropil. There is a functional segregation in motor neuron organization: cibarial muscles involved in sucking are innervated by pharyngeal motor neurons, while the proboscis muscles involved in positioning, of the proboscis are innervated by labial motor neurons. We have also observed projections of the stomodaeal nerve in the tritocerebrum.     

Transplantation of Xenopus laevis Tissues to Determine the Ability of Motor Neurons to Acquire a Novel Target

The evolutionary origin of novelties is a central problem in biology. At a cellular level this requires, for example, molecularly resolving how brainstem motor neurons change their innervation target from muscle fibers (branchial motor neurons) to neural crest-derived ganglia (visceral motor neurons) or ear-derived hair cells (inner ear and lateral line efferent neurons). Transplantation of various tissues into the path of motor neuron axons could determine the ability of any motor neuron to innervate a novel target. Several tissues that receive direct, indirect, or no motor innervation were transplanted into the path of different motor neuron populations in Xenopus laevis embryos. Ears, somites, hearts, and lungs were transplanted to the orbit, replacing the eye. Jaw and eye muscle were transplanted to the trunk, replacing a somite. Applications of lipophilic dyes and immunohistochemistry to reveal motor neuron axon terminals were used. The ear, but not somite-derived muscle, heart, or liver, received motor neuron axons via the oculomotor or trochlear nerves. Somite-derived muscle tissue was innervated, likely by the hypoglossal nerve, when replacing the ear. In contrast to our previous report on ear innervation by spinal motor neurons, none of the tissues (eye or jaw muscle) was innervated when transplanted to the trunk. Taken together, these results suggest that there is some plasticity inherent to motor innervation, but not every motor neuron can become an efferent to any target that normally receives motor input. The only tissue among our samples that can be innervated by all motor neurons tested is the ear. We suggest some possible, testable molecular suggestions for this apparent uniqueness.     

The morphology,physiology and function of suboesophageal neck motor neurons in the honeybee

We report some of the neural and muscular circuitry that allows honeybees to control head movements. We studied neck motor neurons with cell bodies in the suboesophageal ganglion, axons in the first cervical nerve (IK1) and terminals in neck muscles 44 and 51 (muscle classification: Snodgrass in Smithsonian Misc Coll 103:1-120, 1942). We show that muscle 44 actually comprises five separate bundles of muscle fibres (subunits), while muscle 51 is split into two subunits. Eight motor neurons innervate muscles 44 and 51. Two motor neurons have cell bodies in the ventral-median cell body group (one innervates a subunit in muscle 44, the other a subunit in muscle 51). One motor neuron has a ventrally located contralateral cell body (innervating a subunit in muscle 44) and five have laterally located ipsilateral cell bodies. Of the five lateral cells, one innervates a subunit in muscle 51, three selectively innervate subunits in muscle 44 and one co-innervates a subunit in muscle 44 with the contralateral cell. Extracellular recordings revealed three types of visually driven, direction-selective cell-types in each IK1 tuned for leftward, rightward and downward motion over the eyes. The spatiotemporal tuning of the units is similar to that of other visual interneurons in the bee brain.     

Complex innervation of three neck muscles by motor and dorsal unpaired median neurons in crickets

Summary In the crickets, Gryllus campestris and Gryllus bimaculatus, the innervation of the dorso-ventral neck muscles M62, M57, and M59 was examined using cobalt staining via peripheral nerves and electrophysiological methods. M62 and M57 are each innervated by two motoneurons in the suboesophageal ganglion. The four motoneurons project into the median nerve to bifurcate into the transverse nerves of both sides. M62 and M57 are the only neck muscles innervated via this route. These bifurcating axon-projections are identical to those of the spiracular motoneurons in the prothoracic ganglion innervating the opener and closer muscle of the first thoracic spiracle in the cricket. The morphology of their branching pattern is described. The neck muscle M57 and the opener muscle of the first thoracic spiracle are additionally innervated by one mesothoracic motoneuron each, with similar morphology. These results suggest, that in crickets, the neck muscles M57 and M62 are homologous to spiracular muscles in the thoracic segments. The two neck muscles M62 and M59 (the posterior neighbour of M57) receive projections from a prothoracic dorsal unpaired median (DUM) neuron that also innervates dorsal-longitudinal neck muscles but not M57. In addition, one or two mesothoracic DUM neurons send axon collaterals intersegmentally to M59. This is the first demonstration of the innervation of neck muscles by DUM neurons.     

Lobula plate and ocellar interneurons converge onto a cluster of descending neurons leading to neck and leg motor neuropil in Calliphora erythrocephala

Summary In the fly, Calliphora erythrocephala, a cluster of three Y-shaped descending neurons (DNOVS 1–3) receives ocellar interneuron and vertical cell (VS4–9) terminals. Synaptic connections to one of them (DNOVS 1) are described. In addition, three types of small lobula plate vertical cell (sVS) and one type of contralateral horizontal neuron (Hc) terminate at DNOVS 1, as do two forms of ascending neurons derived from thoracic ganglia. A contralateral neuron, with terminals in the opposite lobula plate, arises at the DNOVS cluster and is thought to provide heterolateral interaction between the VS4–9 output of one side to the VS4–9 dendrites of the other. DNOVS 2 and 3 extend through pro-, meso-, and metathoracic ganglia, branching ipsilaterally within their tract and into the inner margin of leg motor neuropil of each ganglion. DNOVS 1 terminates as a stubby ending in the dorsal prothoracic ganglion onto the main dendritic trunks of neck muscle motor neurons. Convergence of VS and ocellar interneurons to DNOVS 1 comprises a second pathway from the visual system to the neck motor, the other being carried by motor neurons arising in the brain. Their significance for saccadic head movement and the stabilization of the retinal image is discussed.     

Regeneration studies on a crayfish neuromuscular system. II. Effect of changing the nerve entry point into the muscle field on the gradient of innervation

The superficial flexor muscle of the crayfish is a neuromuscular system in which the neurons form position-dependent connectivity patterns with the muscle fibers. This system could be formed with the help of a single medial-to-lateral gradient during development that embodies positional information. To test this gradient hypothesis we changed the nerve's normal medial entry point into the muscle by transplanting it to the middle of the muscle sheet. When all the muscle fibers were present in the target area, most of the neurons studied passed through a stage during regeneration in which they showed preference for either medial or lateral synapse formation. Those neurons that in normal animals innervated preferentially the medial fibers showed a medial preference for new contacts; the neuron that normally innervated the lateral fibers showed a lateral preference for new contacts; the neuron that normally innervated everywhere regenerated equally well into both medial and lateral fibers. Therefore, these neurons are able to detect information regarding their position within the muscle mass and respond to it by preferential synapse formation. The effect of a positional gradient could not be detected when half of the target field was removed prior to regeneration. In this instance, the neuron that innervated the missing target area now regenerated to almost all the available fibers. It is suggested that the interplay of positional cues with other factors at different points in time could determine the final connectivity patterns formed by these cells.     

Visual response properties of neck motor neurons in the honeybee

Recent behavioural studies have demonstrated that honeybees use visual feedback to stabilize their gaze. However, little is known about the neural circuits that perform the visual motor computations that underlie this ability. We investigated the motor neurons that innervate two neck muscles (m44 and m51), which produce stabilizing yaw movements of the head. Intracellular recordings were made from five (out of eight) identified neuron types in the first cervical nerve (IK1) of honeybees. Two motor neurons that innervate muscle 51 were found to be direction-selective, with a preference for horizontal image motion from the contralateral to the ipsilateral side of the head. Three neurons that innervate muscle 44 were tuned to detect motion in the opposite direction (from ipsilateral to contralateral). These cells were binocularly sensitive and responded optimally to frontal stimulation. By combining the directional tuning of the motor neurons in an opponent manner, the neck motor system would be able to mediate reflexive optomotor head turns in the direction of image motion, thus stabilising the retinal image. When the dorsal ocelli were covered, the spontaneous activity of neck motor neurons increased and visual responses were modified, suggesting an ocellar input in addition to that from the compound eyes.     

Early events in peripheral regulatory T cell induction via the nasal mucosa

Nasal application of soluble Ags leads to Ag-specific suppression of systemic immune responses. This tolerance can be transferred to naive mice by CD4(+) regulatory T cells (T(R) cells) from the spleen, but little is known about the induction of mucosal T(R) cells in vivo. To investigate the induction of T(R) cells in the nose-draining cervical lymph node (CLN), CD4(+) T cells from DO11.10 OVA TCR transgenic mice were transferred to BALB/c recipients. Within 48 h after nasal OVA application, CD4(+) DO11.10 T cells in CLN, but not in the peripheral lymph node, had divided. Similarly, nonmucosal (i.m.) OVA application also induced CD4(+) DO11.10 T cells to proliferate in the draining inguinal lymph node (ILN), yet more vigorously and with different kinetics than the CD4(+) DO11.10 T cells in CLN. Functional analysis revealed that only proliferating CD4(+) DO11.10 T cells from CLN, and not ILN, could transfer tolerance to naive recipients. CD4(+) DO11.10 T cells from CLN were phenotypically similar to CD4(+) DO11.10 T cells from ILN, however, in CLN a higher percentage of CD25(+) proliferating CD4(+) DO11.10 T cells were detected compared with ILN. CD25 is not a discriminative marker for mucosal T(R) cells because both CD25(+) and CD25(-) CD4(+) DO11.10 T cells from the CLN could suppress delayed type hypersensitivity responses in adoptive transfer. These findings demonstrate that although striking similarities exist between the differentiation of T(R) and effector T cells, this does not include their function. We are the first to demonstrate that functional T(R) cells, which reside within both CD25(+) and CD25(-) subsets, can be isolated from CLN as early as 3 days after nasal OVA application.     

A Golgi-electron-microscopical study of the structure and development of the lamina ganglionaris of the locust optic lobe

Summary The gross structure as well as the neuronal and non-neuronal components of the lamina ganglionaris of the locust Schistocerca gregaria are described on the basis of light- and electron-microscopical preparations of Golgj (selective silver) and ordinary histological preparations. The array of optic cartridges within the lamina neuropile — their order and arrangement — and the composition of the cartridges are described. There are six types of monopolar neurons: three whose branches reach to other cartridges and three whose branches are confined to their own cartridges. Retinula axons terminate either in the lamina or the medulla neuropiles. There are three types of centrifugal neurons, two types of horizontal neuron, as well as glia and trachea in the lamina neuropile. The development of the lamina neuropile is described in terms of developing monopolar and centrifugal axons, growing retinula fibres, and composition of the developing optic cartridges.MSN was supported in part by a Fulbrights-Hays Scholarsship. We are grateful to the Science Research Council for its grant to PMJS.     

The effect of the transplanted pineal gland on the sympathetic innervation of the rat sublingual gland

We investigated the effect of the pineal on sympathetic neurons that normally innervate the sublingual gland of the rat. When the pineal gland was transplanted into the sublingual gland, it remained as a distinct mass that was innervated by sympathetic axons. Injection of the retrograde tracer, Fast Blue, into the sublingual gland labelled sympathetic neurons in the ipsilateral superior cervical ganglion (SCG). Thirty per cent of all neurons labelled retrogradely by Fast Blue injection into transplanted pineal glands were immunoreactive for both neuropeptide Y (NPY) and calbindin. This combination is characteristic of sympathetic neurons innervating the pineal gland in its normal location, but not the sympathetic vasoconstrictor neurons normally innervating the sublingual gland. This, and our previous study in which the pineal gland was shown to similarly influence the phenotype of salivary secretomotor neurons, suggests that a range of different functional classes of sympathetic neuron are able to change their phenotype in response to signals released by the pineal gland.This work was supported by Project Grant No. 145634 from the National Health and Medical Research Council of Australia     

Gustatory hairs on the mosquito, Culiseta inornata.

Gustatory hairs were investigated on the legs and mouthparts of Culiseta inornata (Williston) (Diptera: Culicidae). One type of hair, each innervated by four neurons, was found on the legs. Two of the neurons responded to NaCl stimulation, one neuron to water stimulation, and one neuron to sucrose stimulation. Three kinds of hairs designated Type I (T1), Type 2 (T2) and Type 3 (T3) were analyzed on the labella. The T1 hairs are innervated by one sugar neuron, one mechanoreceptor, two salt neurons and one water neuron. The T2 hairs are innervated by two salt neurons and one mechanoreceptor. The T3 hairs, located on the oral surface of the labella, are innervated by a variable number (2-5) of neurons. Precise identification of the T3 chemosensory neurons was not made because of the small size and inaccessibility of the T3 hairs. Chemosensory hairs on the tip of the labrum were tested electrophysiologically. the sequence of decreasing effeectiveness for the three salts tested was KCl greater than NaCl greater than LiCl. Labral chemoreceptors also responded positively to sucrose.     

The morphology of suboesophageal ganglion cells innervating the nervus corporis cardiaci III of the locust

Summary The nervus corporis cardiaci III (NCC III) of the locust Locust migratoria was investigated with intracellular and extracellular cobalt staining techniques in order to elucidate the morphology of neurons within the suboesophageal ganglion, which send axons into this nerve. Six neurons have many features in common with the dorsal, unpaired, median (DUM) neurons of thoracic and abdominal ganglia. Three other cells have cell bodies contralateral to their axons (contralateral neuron 1–3; CN 1–3). Two of these neurons (CN2 and CN3) appear to degenerate after imaginal ecdysis. CN3 innervates pharyngeal dilator muscles via its anterior axon in the NCC III, and a neck muscle via an additional posterior axon within the intersegmental nerve between the suboesophageal and prothoracic ganglia. A large cell with a ventral posterior cell body is located close to the sagittal plane of the ganglion (ventral, posterior, median neuron; VPMN). Staining of the NCC III towards the periphery reveals that the branching pattern of this nerve is extremely variable. It innervates the retrocerebral glandular complex, the antennal heart and pharyngeal dilator muscles, and has a connection to the frontal ganglion.Abbreviations AH antennal heart - AN antennal nerves - AO aorta - AV antennal vessel - CA corpus allatum - CC corpus cardiacum - CN1, CN2, CN3 contralateral neuron 1–3 - DIT dorsal intermediate tract - DMT dorsal median tract - DUM dorsal, unpaired, median - FC frontal connective - FG frontal ganglion - HG hypocerebral ganglion - LDT lateral dorsal tract - LMN, LSN labral motor and sensory nerves - LN+FC common root of labral nerves and frontal connective - LO lateral ocellus - MDT median dorsal tract - MDVR ventral root of mandibular nerve - MVT median ventral tract - NCA I, II nervus corporis allati I, II - NCC I, II, III nervus corporis cardiaci I, III - NR nervus recurrens - NTD nervus tegumentarius dorsalis - N8 nerve 8 of SOG - OE oesophagus - OEN oesophageal nerve - PH pharynx - SOG suboesophageal ganglion - T tentorium - TVN tritocerebral ventral nerve - VLT ventral lateral tract - VIT ventral intermediate tract - VMT ventral median tract - VPMN ventral, posterior, median neuron - 1–7 peripheral nerves of the SOG - 36, 37, 40–45 pharyngeal dilator muscles     

Central projections of the wing afferents in the hawkmoth, Agrius convolvuli

Flight behaviors in various insect species are closely correlated with their mechanical and neuronal properties. Compared to locusts and flies which have been intensively studied, moths have “intermediate” properties in terms of the neurogenic muscle activations, power generation by indirect muscles, and two-winged-insect-like flapping behavior. Despite these unique characteristics, little is known about the neuronal mechanisms of flight control in moths. We investigated projections of the wing mechanosensory afferents in the central nervous system (CNS) of the hawkmoth, Agrius convolvuli, because the mechanosensory proprioceptive feedback has an essential role for flight control and would be presumably optimized for insect species. We conducted anterograde staining of nine afferent nerves from the fore- and hindwings. All of these afferents projected into the prothoracic, mesothoracic and metathoracic ganglia (TG1, 2 and 3) and had ascending fibers to the head ganglia. Prominent projection areas in the TG1–3 and suboesophageal ganglion (SOG) were common between the forewing, hindwing and contralateral forewing afferents, suggesting that information from different wings are converged at multiple levels presumably for coordinating wing flapping. On the other hand, differences of projections between the fore- and hindwing afferents were observed especially in projection areas of the tegulae in the TG1 and contralateral projections of the anterior forewing nerve in the TGs and SOG, which would reflect functional differences between corresponding mechanoreceptors on each wing. Afferents comprising groups of the campaniform sensilla at the wing bases had prominent ascending pathways to the SOG, resembling the head–neck motor system for gaze control in flies. Double staining of the wing afferents and flight or neck motoneurons also indicated potential connectivity between them. Our results suggest multiple roles of the wing proprioceptive feedback for flight and provide the anatomical basis for further understanding of neuronal mechanisms of the flight system in moths.     

Serotonin-immunoreactive neurons in the suboesophageal ganglion of the caterpillar of the hawk moth Manduca sexta

Summary Serotonin-immunoreactivity is mapped in wholemounts and slices of the suboesophageal ganglion (SOG) of larval Manduca sexta by means of immunocytochemistry. An extensive meshwork of serotonin-immunoreactive nerve fibres on some peripheral nerves of the SOG has been demonstrated. This meshwork appears to belong to a serotonergic neurohemal system, probably supplied by two pairs of bilateral serotonin-immunoreactive neurons with big cell bodies on the dorsal side near the midline in the mandibular neuromere. Intracellular recording and staining revealed their physiology and morphology. These neurons produce long lasting (50 msec) action potentials, which suggest that they are neurosecretory cells. Two pairs of bilateral serotonin-immunoreactive interneurons similar to those of other insects are stained in the labial and maxillar neuromeres, but not in the mandibular neuromere. Their ventrolaterally located cell bodies project through a ventral commissure into the contralateral hemiganglion and then cross back again through a dorsal commissure. The axons project into the contralateral circumoesophageal connective.     

Absence of competitive interactions among axon terminals of regenerating motor neurons

Competition among axon terminals is usually considered to contribute to the formation of patterned synaptic connections. During axonal regeneration of motor neurons in the cockroach, leg muscles initially become innervated by appropriate and inappropriate motor neurons. All axon terminals from inappropriate neurons eventually are eliminated, resulting in the reformation of the original innervation pattern. Destruction of an identified motor neuron by the intracellular injection of pronase did not prevent the elimination of inappropriate axon terminals in the muscle normally innervated by that motor neuron. Therefore, competition does not play a role in the reinnervation of the leg muscles. This indicates a major role for specific cell-cell recognition.     

Octopamine neuromodulatory effects on a social behavior decision-making network in Drosophila males

Situations requiring rapid decision-making in response to dynamic environmental demands occur repeatedly in natural environments. Neuromodulation can offer important flexibility to the output of neural networks in coping with changing conditions, but the contribution of individual neuromodulatory neurons in social behavior networks remains relatively unknown. Here we manipulate the Drosophila octopaminergic system and assay changes in adult male decision-making in courtship and aggression paradigms. When the functional state of OA neural circuits is enhanced, males exhibit elevated courtship behavior towards other males in both behavioral contexts. Eliminating the expression of the male form of the neural sex determination factor, Fruitless (Fru(M)), in three OA suboesophageal ganglia (SOG) neurons also leads to increased male-male courtship behavior in these same contexts. We analyzed the fine anatomical structure through confocal examination of labeled single neurons to determine the arborization patterns of each of the three Fru(M)-positive OA SOG neurons. These neurons send processes that display mirror symmetric, widely distributed arbors of endings within brain regions including the ventrolateral protocerebra, the SOG and the peri-esophageal complex. The results suggest that a small subset of OA neurons have the potential to provide male selective modulation of behavior at a single neuron level.     

Electrical activity in the chemoreceptors of the blowfly. I. Responses to chemical and mechanical stimulation

The electrical responses of the neurons associated with the various types of chemosensory hairs of the blowfly, Phormia regina Meigen, following stimulation by chemical and mechanical means have been studied. The singly innervated chemosensory hairs on the ovipositor, maxillary palpi, and antennae respond vigorously to chemical stimulation, but not to mechanical stimulation. The triply innervated chemosensory hairs on the labellum, tarsus, and wing have two neurons which respond only to chemical stimuli. The third neuron responds only to mechanical stimulation. The differential responses of the two chemosensory neurons to various chemical stimuli following the removal of the tip of the hair suggest that the structures responsible for chemoreception are located throughout the distal processes of these neurons. The response of the third neuron to mechanical stimulation is similar to the response recorded from the neuron associated with one type of tactile hair which responds to motion and not to steady deformation. Recordings have been made from the neurons associated with purely tactile hairs using the cut hair as an extension of the micropipette. The mechanosensory neuron of the wing chemosensory hair is capable of responding at the rate of at least 600 impulses per sec. and may serve to indicate changes in air flow over the wing surfaces during flight to enable the fly to correct the wing camber and attack angle.     

Sensory modalities conveyed in the hindlimb somatic afferent input to nucleus tractus solitarius.

To determine the somatic sensory modalities conveyed by hindlimb somatic afferent inputs, the discharge of neurons in the nucleus tractus solitarius was recorded in anesthetized rats after electrical stimulation of either the contralateral sciatic nerve or L(6) spinal nerve, which innervates the hindlimb. The discharge of seven of eight cells was increased (P < 0.05) by capsaicin injected into the arterial supply of the hindlimb. Discharge was unaltered in 19 neurons tested for sensitivity to nonnoxious (40 degrees C) and noxious (47 degrees C) heating of the hindlimb skin. In contrast, lightly stroking the skin elicited discharge in 2 of 14 cells, whereas noxious pinching increased activity in 4 other cells. Rhythmic (1- to 3-s) muscle contraction (MC) increased (P < 0.05) discharge in >60% of neurons tested (11 of 18). Static (10- to 30-s) MC significantly (P < 0.05) increased discharge in four cells, two of which were also responsive to rhythmic MC. Rhythmic and sustained muscle stretch increased discharge (P < 0.05) in three of eight neurons tested. These data indicate that nucleus tractus solitarius neurons receive input from low- and high-threshold cutaneous mechanoreceptors, respond to capsaicin delivered into the hindlimb arterial supply, lack thermal sensitivity, and respond to activation of mechanosensitive as well as metabosensitive endings in skeletal muscle.     

Neuroanatomy of the sucking pump of the moth, Manduca sexta (Sphingidae, Lepidoptera)

Knowledge of the neuroanatomy of the sucking pump of Manduca sexta (Sphingidae) is valuable for studies of olfactory learning, pattern generators, and postembryonic modification of motor circuitry. The pump comprises a cibarial valve, a buccal pump, and an esophageal sphincter valve. Cibarial opener and closer muscles control the cibarial valve. Six pairs of dilator muscles and a compressor muscle operate the buccal pump. The cibarial opener and one pair of buccal dilator muscles are innervated by paired neurons in the tritocerebrum, and the cibarial opener has double, bilateral innervation. Their tritocerebral innervation indicates that these muscles evolved from labro-clypeal muscles. The remaining paired buccal dilator muscles each are innervated by an unpaired motor neuron in the frontal ganglion. These motor neurons project bilaterally through the frontal connectives to dendritic arborizations in the tritocerebrum. These projections also have a series of dendritic-like arborizations in the connectives. The cibarial closer and buccal compressor muscles are also innervated by motor neurons in the frontal ganglion, but only the closer muscle neuron projects bilaterally to the tritocerebrum. The innervation of the pump muscles indicates that they are associated with the stomodaeum, and, therefore, the buccal pump evolved from the anterior stomodaeum rather than from the cibarium.     

Sprouting of aberrant neuropeptide Y-immunoreactive sympathetic nerves into neonatally denervated smooth muscle.

Sympathetic nerves normally project ipsilaterally to lateral cranial targets. Following unilateral superior cervical ganglionectomy in neonatal rats, however, neurons from the contralateral superior cervical ganglion sprout into the denervated region. In the present study we examined neuropeptide Y immunoreactivity (NPY-ir) of neurons comprising ipsilateral (control) and denervation-induced contralateral pathways to the superior tarsal smooth muscle of the eyelid. Fluoro-Gold injection of the control muscle retrogradely labelled 133 +/- 18 neurons in the ipsilateral superior cervical ganglion; of these, 21 +/- 3% displayed detectable NPY-ir. Fluoro-Gold injections of the reinnervated muscle labelled 20 +/- 4 neurons in the contralateral superior cervical ganglion, of which 85 +/- 3% contained detectable NPY-ir. Examination of the control tarsal muscle revealed DBH-ir noradrenergic nerves throughout the muscle and vasculature, while NPY-ir nerves were present primarily around blood vessels. In the reinnervated preparation, NPY-ir fibers innervated both blood vessels and tarsal muscle in a pattern similar to that of DBH-ir innervation. Acute excision of the remaining superior cervical ganglion eliminated all DBH-ir fibers bilaterally; NPY-ir was reduced markedly in the reinnervated preparations, though some fibers remained. We conclude that, following neonatal denervation, the tarsal muscle is reinnervated by a subpopulation of sympathetic neurons that differs in neuropeptide phenotype from that of the normal ipsilateral innervation.