Development of projections of primary afferent fibers from the hindlimb to the gracile nucleus: a WGA-HRP study in the rat.

The projection of primary afferent fibers to the gracile nucleus was studied during development. Injections of wheat germ agglutinin-horseradish peroxidase were made into the hindlimb of fetal, postnatal and adult rats. In most cases the sections were alternately stained for wheat germ agglutinin-horseradish peroxidase including counter stain with Neutral red and for acetylcholinesterase. At embryonic day 17 labelled fibers could be traced to the mid-cervical spinal cord but not further rostrally. At embryonic days E18 and E19 labelled fibers penetrate the rostral pole of the nucleus, which does not happen more caudally. At embryonic day E21 the caudal-most pole of the gracile nucleus still is not penetrated by labelled fibers. From postnatal day 1 onwards labelled fibers are found throughout the entire rostrocaudal extent of the gracile nucleus. These results suggest that primary afferent fibers from the hindlimb first grow to the rostral pole of the gracile nucleus and penetrate the rostral pole immediately upon their arrival. During further development more caudal parts of the gracile nucleus are gradually penetrated in a rostrocaudal fashion by primary afferent fibers of the hindlimb.     

The origin of sensory nerve fibers that innervate the submandibular salivary gland in the rat

The origin of sensory nerves that innervate the submandibular salivary gland was investigated in the rat. After application of wheat germ agglutinin-horseradish peroxidase to the cut endings of the sympathetic and parasympathetic nerve branches at the hilus of the gland, labeled cells were mainly found in the dorsal root ganglia and the trigeminal ganglion, respectively. The labeled neurons in these ganglia were of various sizes compared to unlabeled neurons, suggesting that the sensory nerves of the gland conduct various modalities of sensory information.     

Central distribution of cervical primary afferents in the rat, with emphasis on proprioceptive projections to vestibular, perihypoglossal, and upper thoracic spinal nuclei

The projections of primary afferents from rostral cervical segments to the brainstem and the spinal cord of the rat were investigated by using anterograde and transganglionic transport techniques. Projections from whole spinal ganglia were compared with those from single nerves carrying only exteroceptive or proprioceptive fibers. Injections of horseradish peroxidase (HRP) or wheat germ agglutinin-horseradish peroxidase conjugate (WGA-HRP) were performed into dorsal root ganglia C2, C3, and C4. Free HRP was applied to the cut dorsal rami C2 and C3, greater occipital nerve, sternomastoid nerve, and to the C1/2 anastomosis, which contains afferents from suboccipital muscles and the atlanto-occipital joint. WGA-HRP injections into ganglia C7 and L5 were performed for comparative purposes. Injections of WGA-HRP or free HRP into rostral cervical dorsal root ganglia and HRP application to C2 and C3 dorsal rami produced labeling in dorsal and ventral horns at the level of entrance, the central cervical nucleus, and in external and main cuneate nuclei. From axons ascending to pontine and descending to upper thoracic spinal levels, medial collaterals were distributed to medial and descending vestibular, perihypoglossal and solitary nuclei, and the intermediate zone and Clarke's nucleus dorsalis in the spinal cord. Lateral collaterals projected mainly to the trigeminal subnucleus interpolaris and to lateral spinal laminae IV and V. Results from HRP application to single peripheral nerves indicated that medial collaterals were almost exclusively proprioceptive, whereas lateral collaterals were largely exteroceptive with a contribution from suboccipital proprioceptive fibers. WGA-HRP injections into dorsal root ganglia C7 and L5 failed to produce significant labeling within vestibular and periphypoglossal nuclei, although they demonstrated classical projection sites within the brainstem and spinal cord. The consistent collateralisation pattern of rostral cervical afferents along their whole rostrocaudal course enables them to contact a variety of precerebellar, vestibulospinal, and preoculomotor neurons. These connections reflect the well-known significance of proprioceptive neck afferents for the control of posture, head position, and eye movements.     

Deafferentation-induced expansion of saphenous terminal field labelling in the adult rat dorsal horn following pronase injection of the sciatic nerve

We have examined the effect of the degeneration of sciatic nerve afferents on the distribution of saphenous terminals in the adult rat dorsal horn. Deafferentation was produced by injection into the sciatic nerve of pronase, a combination of proteolytic enzymes, which causes death of ganglion cells and degeneration of their terminal fields. The saphenous terminal fields were labelled by exposing the cut nerve to a combination of horseradish peroxidase (HRP) and wheat germ agglutinin-horseradish peroxidase (WGA-HRP). Terminals were mainly found in the superficial dorsal horn, indicating that small-diameter afferents were heavily labelled. In one group of control animals, the normal sciatic and normal saphenous terminal fields were shown to be bilaterally symmetrical. In the experimental group, the initial injection of one sciatic nerve with pronase was followed 4 months later by bilateral HRP/WGA-HRP labelling of both saphenous nerves. In each animal, the terminal field of the saphenous nerve on the lesioned side was expanded in the medial, lateral, and caudal directions. Medially and laterally, the expanded terminal field overlapped more of the sciatic territory than normal; caudally, saphenous terminals were found in the rostral portion of the L5 segment, in an area normally filled by sciatic terminals and devoid of saphenous terminals. The expansion resulted in a total saphenous area 26% larger than the control side. Electron microscopy demonstrated that the label in both the normal and expanded territories was primarily contained in axons and terminals, with minor transneuronal labelling. Labelled terminals in the expanded areas were both simple terminals with round, clear vesicles, and glomerular terminals with multiple synaptic contacts; these terminal types resemble those previously described for primary afferents in the superficial dorsal horn. Although the preexistence of "silent" synaptic terminals in the expanded areas cannot be disproven, the data support the hypothesis that primary afferents in the adult have the potential to sprout and establish synapses when the conditions of the deafferentation are favorable.     

Functional Connections Formed by Saphenous Nerve Terminal Sprouts in the Dorsal Horn Following Neonatal Sciatic Nerve Section

The rostrocaudal distribution of saphenous nerve inputs into the lumbar dorsal horn from L2 to L6 has been investigated in urethane anaesthetized rats whose left sciatic nerve was cut and ligated at birth. In normal cord, electrical stimulation of the saphenous nerve evoked dorsal horn spikes in L2 to caudal L4. Few or no spikes were evoked in L5. After neonatal sciatic nerve section, saphenous nerve stimulation evoked spikes throughout segments L2 to L6. Dorsal horn cell receptive fields were also altered following neonatal sciatic nerve section. A somatotopic map of the lumbar enlargement in normal rats was constructed from the receptive fields (RFs) of adjacent dorsal horn cells. Cells with RFs in the saphenous skin region were concentrated in L3 and rostral L4 and very few were found in L5. After neonatal sciatic nerve section, however, a substantial number of cells with low threshold saphenous skin RFs were also found in caudal L4 and throughout L5. These results show that the central saphenous nerve terminal sprouts that grow into the sciatic terminal region following neonatal sciatic nerve section (Fitzgerald, 1985, J. Comp. Neurol., 240, 414-422; Fitzgerald et al., 1990, J. Comp. Neurol., 300, 370-385) form functional connections. This results in dorsal horn cells that are not normally influenced by saphenous nerve inputs developing substantial low threshold RFs in saphenous nerve skin regions.     

Afferent fibers in the hypoglossal nerve: a horseradish peroxidase study in the cat

The existence of afferent fibers in the cat hypoglossal nerve was studied by transganglionic transport of horseradish peroxidase (HRP). Injections of wheat germ agglutinin-conjugated HRP (WGA-HRP) into the hypoglossal nerve resulted in some retrograde labeling of cell bodies within the superior ganglia of the ipsilateral glossopharyngeal and vagal nerves. A few labeled cell bodies were also present ipsilaterally within the inferior ganglion of the vagal nerve and the spinal ganglion of the C1 segment. Some of the labeled glossopharyngeal and vagal fibers reached the nucleus of the solitary tract by crossing the dorsal portion of the spinal trigeminal tract. Others distributed to the spinal trigeminal nucleus pars interpolaris and to the ventrolateral part of the medial cuneate nucleus by descending through the dorsal portion of the spinal trigeminal tract. In the spinal cord these descending fibers, intermingling with labeled dorsal root fibers, distributed to laminae I, IV-V and VII-VIII of the C1 and C2 segments. Additional HRP experiments revealed that the fibers in laminae VII-VIII originate mainly from dorsal root of the C1 segment.     

The target specificity of the extrinsic innervation of the rat small intestine

The target specificity of the extrinsic innervation of the rat small intestine was examined by simultaneously injecting the proximal and distal small intestine with either wheat germ agglutinin-horseradish peroxidase (HRP) or fast blue. The number of single- and double-labeled cells in the nodose, dorsal root and coeliac-superior mesenteric ganglia and the dorsal motor nucleus of the vagus were counted and expressed as percentages of total labeled cells. Cells containing both HRP and Fast blue projected to both regions of the intestine. We found that the nodose and mesenteric ganglia contained significantly fewer double-labeled neurons (approximately 3 and 9% respectively) than the dorsal motor nucleus (19%) or dorsal root ganglion (20%). Presumably, a large number of double-labeled afferent or efferent neurons would limit the ability of a given component of the extrinsic innervation to control the activity of restricted regions of the small intestine (but might be important in overall regulation of intestinal function). In a separate series of experiments we examined the topography of neurons in the dorsal motor nucleus of the vagus labeled with HRP injection into either the proximal or distal small intestine. Both of these injections labeled neurons in the entire rostro-caudal extent of the nucleus, though approximately 75% of the cells were located between 720 microns caudal and 720 microns rostral to the obex. Cells in the rostral regions were found primarily in the lateral pole of the nucleus, whereas caudal regions contained labeled cells in both the medial and lateral poles.     

The ovarian innervation in the dog: a preliminary study for the base for electro-acupuncture.

The origin of the canine ovarian sensory and sympathetic nerves was studied by applying horseradish peroxidase (HRP) or wheat germ agglutinin conjugated to HRP (WGA-HRP) to the ovarian stroma and into the ovarian bursa. HRP/WGA-HRP positive neurons were found bilaterally in the dorsal root ganglia of T10 to L4 segment with the majority located in T13 to L2. In sympathetic paravertebral ganglia, labeled neurons were distributed bilaterally in ganglia from T11 to L4 with the majorities located in segments T13 to L2. Both distributions show ipsilateral predominance. Labeled prevertebral neurons were mainly located in the aorticorenal ganglion, ovarian ganglia and caudal mesenteric ganglion. No labeled neurons were found in the dorsal motor nucleus of vagus, nodose ganglia or sacral segment from S1 to S3. This study provides the possible morphological basis of electro-acupuncture concerning the somato-visceral reflex of the ovary.     

Cutaneous projections from the rat hindlimb foot to the substantia gelatinosa of the spinal cord studied by transganglionic transport of WGA-HRP conjugate

The somatotopic organization of cutaneous afferent fibers from the hindlimb foot in the substantia gelatinosa of the spinal cord was investigated in adult Sprague Dawley rats following intracutaneous injections of wheat germ agglutinin-horseradish peroxidase (WGA-HRP). The different parts of the foot were found to project in a precise manner to the medial two-thirds of the substantia gelatinosa of the spinal cord segments L3-L5. The projections from the digits were arranged in a consecutive row rostrocaudally, with the most medial digit in caudal L3 and the most lateral digit in caudal L4 or rostral L5. The plantar skin was found to project both rostral and caudal to the projection of the digits. The dorsal foot skin projects lateral to the digits, neighboured by the medial edge of the foot rostrally and the lateral edge caudally. Within the projection of a particular digit, the plantar skin was found medially and the dorsal skin laterally. A certain degree of overlap for the projections from the different foot skin areas was found.     

Cervical primary afferent input to vestibulospinal neurons projecting to the cervical dorsal horn: An anterograde and retrograde tracing study in the cat

Vestibulospinal neurons in the caudal half of the medial and descending vestibular nuclei terminate in the cervical spinal cord, not only in the ventral horn and intermediate zone but also in the dorsal horn. The purpose of the present study was to examine whether the areas containing these vestibulospinal neurons are reached by cervical primary afferents. In one group of experiments, wheat germ agglutinin-horseradish peroxidase conjugate and horseradish peroxidase were pressure injected into spinal ganglia C₂-C₈ and revealed anterogradely labeled fibers and boutons in the caudal part (caudal to the dorsal cochlear nucleus) of the ipsilateral medial and descending vestibular nuclei. This projection was verified in experiments in which wheat germ agglutinin-horseradish peroxidase conjugate was microiontophoretically injected into the caudal half of either the medial or the descending vestibular nuclei and revealed retrogradely labeled cells only in ipsilateral spina ganglia C₂-C₇, with a maximum of cells in C₃. In another group of experiments, after microiontophoretic injections of Phaseolus vulgaris leucoagglutinin or Biocytin into either the medial or the descending vestibular nuclei, anterogradely labeled fibers and boutons were present in the cervical spinal cord, mainly bilaterally in the dorsal horn (laminae I–VI) but also, to a lesser extent, in the ventral horn and intermediate zone. The existence of a loop that relays cervical primary afferent information to vestibulospinal neurons projecting to the cervical spinal cord, in particular the dorsal horn, may have implications for vestibular control over local information processing in the cervical dorsal horn. © 1995 Wiley-Liss, Inc.     

Primary sensory afferents in the inferior mesenteric ganglion and related nerves of the guinea pig. An experimental study with anterogradely transported wheat germ agglutinin-horseradish peroxidase conjugate

Peripheral visceral afferents in the guinea pig were labeled by injections of wheat germ agglutinin-horseradish peroxidase conjugate (WGA-HRP) into the L2 and L3 dorsal root ganglia bilaterally. After anterograde transport of the tracer the following areas were examined for the presence of HRP-labeled fibers: the inferior mesenteric ganglion (IMG), the inferior mesenteric artery (IMA) with surrounding tissue, the hypogastric nerves, parts of the descending and sigmoid colon as well as the urinary bladder. Large numbers of heavily labeled fibers were found in the IMG, in the colonic nerves around the IMA and in the hypogastric nerves. In the IMG, profiles suggestive of being labeled axon terminals were observed. Labeled fibers were observed in the muscle layers of the colon and in the bladder wall. The results show that anterograde tracing with WGA-HRP can be used successfully in analyzing the morphology and structural organization of visceral afferents in the periphery.     

Origin of sympathetic and sensory innervation of the elbow joint in the rat: A retrograde axonal tracing study with wheat germ agglutinin conjugated horseradish peroxidase

After injection of wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP) into the elbow joint of adult rats, labeled neurons were found in the stellate and the T2-T4 ganglia of the ipsilateral sympathetic trunk, and also in dorsal root ganglia at the C4–T4 levels. Most labeled sympathetic cells, 90% or more, were located in the stellate ganglion. The sensory innervation to the joint originated mainly from the dorsal root ganglia at the levels of C7–T1.     

The afferent and sympathetic components of the lumbar spinal outflow to the colon and pelvic organs in the cat. II. The lumbar splanchnic nerves

The cell bodies of the lumbar sensory and sympathetic pre- and postganglionic neurons that project to the inferior mesenteric ganglion in the lumbar splanchnic nerves of the cat have been labeled retrogradely with horseradish peroxidase applied to the central end of their cut axons near the inferior mesenteric ganglion. The numbers, segmental distribution, location, and size of these labeled somata have been determined quantitatively. After all the lumbar splanchnic nerves on one side of an animal were labeled, most labeled cell bodies were situated ipsilaterally in dorsal root ganglia, ganglia of the lumbar sympathetic trunk, and spinal cord segments L2-L5, with the maximum numbers in L3 and L4. A few labeled somata lay contralaterally or rostral to L2. After labeling of only one lumbar splanchnic nerve, the majority of cell bodies were found in the labeled segment, but a few were also present up to three segments rostral or caudal. These variations could always be attributed to extraspinal connections usually via the lumbar sympathetic trunk. Cross-sectional areas of labeled afferent somata were small relative to those of the entire population of dorsal root ganglion cells. Preganglionic cell bodies were labeled in the intermediate gray matter extending from its lateral border ventrolaterally across to the central canal. Two regions of high density were observed: one laterally just medial to the edge of the white matter and the other lateral to the central canal. The dorsolateral group lay somewhat medial and caudal to the usual limits of the intermediolateral column. Labeled preganglionic neurons were on the average larger than the unlabeled cells in the inferior mesenteric ganglion, with the group lying medially being larger than those that were laterally positioned. From the data, it is estimated that about 4,600 afferent axons, about 4,600 preganglionic axons, and about 2,800 postganglionic axons travel in the lumbar splanchnic nerves to the inferior mesenteric ganglion of the cat.     

Compensatory projection of primary nociceptors and c-fos induction in the spinal dorsal horn following neonatal sciatic nerve lesion

The sciatic nerve was cut in newborn rats, and prevented from regenerating for 8 weeks. The number of dorsal root ganglion (DRG) neurons in L4 and L5, the distribution of central axon terminals of primary nociceptors, and the activity of secondary nociceptors were examined in the lumbar dorsal horn. The neonatal sciatic lesion caused about 60% reduction of DRG neurons. The central terminal field of the sciatic primary nociceptors negatively labeled by in situ binding of Bandeiraea simplicifolia isolectin B4 (BsIB4) markedly shriveled. Instead, the central representation of the saphenous nerve and the posterior cutaneous nerve of the thigh (PC) expanded. The laminae I/II neuropil in the medialmost (1/4) of the L3 dorsal horn and in the second lateral (1/4) around the L4/5 junction was occupied by the BsIB4 binding sites derived from the saphenous and the PC primary neurons, respectively. Noxious stimuli applied to the receptive fields of the saphenous and the PC nerves induced c-Fos-like immunoreactivity in many neurons in the expanded central terminal fields of the nerves. The collateral sprouts of uninjured primary nociceptors did not only invade the deafferented area of the dorsal horn but also established functional synaptic connections.     

The use of anterogradely transported wheat germ agglutinin-horseradish peroxidase conjugate to visualize cutaneous sensory nerve endings

Small amounts of wheat germ agglutinin-horseradish peroxidase conjugate were injected into the L5 spinal ganglion of adult rats. The conjugate was transported peripherally by the sensory neurons to their cutaneous receptors in the hind paw. Nerve plexa in the skin were heavily labeled, and different types of nerve endings could be clearly distinguished. This method gives new possibilities to study peripheral fibers and nerve endings originating in single sensory ganglia.     

Intramucosal distribution of the glossopharyngeal sensory fibers of cats

In order to determine distribution of the sensory fibers of the glossopharyngeal nerve (IX) in the pharynx of cats, wheat germ agglutinin-horseradish peroxidase was injected into the superior and inferior ganglia of IX. Results were as follows: Labeled peripheral sensory nerve fibers in the pharynx were recognized ipsilaterally. The pharyngeal branch of IX innervated the nasopharyngeal mucosa at the level of the torus tubarius. The tonsillo-lingual branch was divided into four rami. The first ramus innervated the caudal one-third of the tongue and the vallate papillae. The second ramus innervated the palatine tonsil and the caudal half of the soft palate. The third ramus supplied a part of the radix linguae, the vallecula epiglottica and the lingual aspect of the epiglottis. The fourth ramus supplied the hypopharyngeal mucosa rostral to the middle level of the aryepiglottic fold.     

Collateral sprouting of sensory axons in the glabrous skin of the hindpaw after chronic sciatic nerve lesion in adult and neonatal rats: a morphological study

The anterograde transport of wheat germ agglutinin-horseradish peroxidase conjugate was used to study the normal distribution of sensory nerve axons in the plantar skin of the rat hindlimb and at various times after chronic sciatic nerve injury in adult and neonatal rats. In adults, thin saphenous nerve axons were found in a small area laterally to the normal saphenous nerve territory 2–24 weeks after sciatic nerve lesion. In neonatal rats, at 6 and 10 weeks after sciatic nerve injury thin sapherous nerve axons were found almost or all over the sole of the foot, respectively, and in all 5 toes. At longer survival times, the area innervated by saphenous nerve axons became smaller. However, this area was now occupied by thin, as well as coarse axons. When adult animals were subjected to saphenous nerve crush simultaneously with the sciatic nerve lesion, thin, as well as coarse, nerve axons were found laterally to the normal saphenous nerve territory. The findings indicate that thin cutaneous sensory axons of adult mammals can extent collateral sprouts in glabrous skin for a short distance. This capacity appears to be greater in neonatally lesioned animals, where it is present for coarse cutaneous sensory axons as well. However, after neonatal nerve injury collateral sprouts seem to disappear from the initially most distally reinnervated area. Regenerating sensory axons in adult rats seem to have a greater capacity for collateral sprouting than intact axons. Coarse and thin cutaneous sensory axons could be found in this instance. In all instance a great part of the plantar skin remained denervated, suggesting that there is an upper limit for the territory which can be maintained by cutaneous sensory neurons reinnervating glaborous skin.     

Somatotopically organized transient projections from the primary somatosensory cortex to the cerebellar cortex

The organization of transient projections from the primary somatosensory cortex (S-I) to the cerebellar cortex in neonatal kittens was examined using orthograde intraaxonal labeling techniques. Tritiated amino acid injections into face, forelimb and hindlimb areas of representation in S-I labeled mossy fiber-like terminals of cerebrocerebellar axons in different areas of the cerebellar cortex bilaterally. The hindlimb area of S-I projected to lobules I-IV in the anterior lobe and to ventral folia of the paramedian lobule (PML). Injections into forelimb areas of S-I labeled terminals in lobules IV and V and in intermediate and dorsal folia of the PML. The face area of S-I projected to the lobules V and VI, to medial folia in the ansiform and simplex lobules and to dorsal PML folia. Labeled terminals were more numerous in the cerebellar cortex contralateral to the S-I injections, except in lobules I and II and the ventral PML where the density of hindlimb input was approximately the same on both sides. These observations were supplemented by findings that small wheat germ agglutinin-horseradish peroxidase (WGA-HRP) injections into the dorsal or ventral PML resulted in retrogradely labeled layer V pyramidal neurons in lateral (face and forelimb) and medial (hindlimb) areas of S-I respectively. The somatotopic organization of transient S-I cerebrocerebellar projections is very similar to the topography of cerebellar somatosensory afferent pathways in adult cats.     

Auditory belt and parabelt projections to the prefrontal cortex in the Rhesus monkey

Recent anatomical and electrophysiological studies have expanded our knowledge of the auditory cortical system in primates and have described its organization as a series of concentric circles with a central or primary auditory core, surrounded by a lateral and medial belt of secondary auditory cortex with a tertiary parabelt cortex just lateral to this belt. Because recent studies have shown that rostral and caudal belt and parabelt cortices have distinct patterns of connections and acoustic responsivity, we hypothesized that these divergent auditory regions might have distinct targets in the frontal lobe. We, therefore, placed discrete injections of wheat germ agglutinin-horseradish peroxidase or fluorescent retrograde tracers into the prefrontal cortex of macaque monkeys and analyzed the anterograde and retrograde labeling in the aforementioned auditory areas. Injections that included rostral and orbital prefrontal areas (10, 46 rostral, 12) labeled the rostral belt and parabelt most heavily, whereas injections including the caudal principal sulcus (area 46), periarcuate cortex (area 8a), and ventrolateral prefrontal cortex (area12vl) labeled the caudal belt and parabelt. Projections originating in the parabelt cortex were denser than those arising from the lateral or medial belt cortices in most cases. In addition, the anterior third of the superior temporal gyrus and the dorsal bank of the superior temporal sulcus were also labeled after prefrontal injections, confirming previous studies. The present topographical results suggest that acoustic information diverges into separate streams that target distinct rostral and caudal domains of the prefrontal cortex, which may serve different acoustic functions. J. Comp. Neurol. 403:141–157, 1999. © 1999 Wiley-Liss, Inc.     

Retrograde tracing shows that CGRP-immunoreactive nerves of rat trachea and lung originate from vagal and dorsal root ganglia

The origins of sensory innervation of the lower respiratory tract are thought to be principally the nodose and jugular ganglia of the vagus nerve. It has been suggested and partially demonstrated that there is also a component arising from dorsal root ganglia, but the segmental levels involved are not known precisely. We have therefore investigated the origins of sensory nerves within the rat respiratory tract, particularly those containing calcitonin gene-related peptide (CGRP), using the technique of retrograde axonal tracing combined with immunohistochemistry. Injections of True blue were made into extra-thoracic trachea (n = 4 rats) and percutaneously into the right and left lung (n = 4 each). Retrogradely labelled neuronal perikarya were detected in vagal and dorsal root ganglia, and sympathetic chain ganglia. CGRP-immunoreactive cells were seen only in vagal and dorsal root ganglia. Tracheal innervation arose bilaterally in the vagal sensory ganglia but those on the right side represented the principal source; the majority of CGRP-containing neurons occurred in the jugular ganglion. A very small component of labelling occurred in spinal ganglia at levels C2-C6. The sensory innervation of the lungs was seen to arise predominantly from the ipsilateral dorsal root ganglia (45% of cells CGRP-immunoreactive) at levels T1-T6. In contrast to the trachea, the contribution of vagal sensory neurones to the lungs appeared to be less than that of the spinal ganglia. These results show that the sensory innervation of the rat lungs has a major origin in the dorsal root ganglia, in which almost half of the involved neurons contain CGRP, and confirm that most CGRP-immunoreactive nerves in the trachea arise in the right jugular ganglion.