High power, low frequency components of cardiac, renal, splenic and vertebral sympathetic nerve activities are uniformly reduced by spinal cord transection

Others have reported that spinal cord transection non-uniformly affects the activities of the cardiac, renal and splenic nerves. We were unable to confirm this finding while recording the wide band (1-1000 Hz) discharges of these sympathetic nerves in baroreceptor-denervated, chloralose-anesthetized cats. Most of the power in nerve discharges was below 6 Hz. There was high coherence between the low frequency discharges of different nerves, and power below 15 Hz was essentially eliminated by spinal transection. Evidence is presented that the disparities between this and past studies are, in part, due to the use of a 30-Hz high-pass filter in the earlier studies.     

Evidence for a serotonergically mediated sympathoexcitatory response to stimulation of medullary raphe nuclei

The cardiovascular role of serotonin (5-HT) containing neurons in the midline medullary raphe nuclei was studied in anesthetized cats. High frequency electrical stimulation of nucleus (n.) raphe (r.) pallidus, n.r. obscurus and n.r. magnus produced both pressor and depressor responses. Single shock stimulation of pressor sites produced an excitatory evoked potential of sympathetic nervous discharge (SND) recorded from the inferior cardiac nerve. Conversely, single shock stimulation of vasodepressor sites resulted in a computer-summed inhibition of SND. The mean conduction velocity in the sympathoexcitatory medullo-spinal pathway to sympathetic preganglionic neurons was calculated to be 1.24 m/s. The 5-HT antagonists methysergide and metergoline blocked the excitation of sympathetic activity evoked from medullary raphe nuclei. In contrast, these agents failed to alter the sympathoexcitatory response to electrical stimulation of lateral medulla pressor sites or the sympathoinhibitory response elicited by raphe stimulation. The 5-HT uptake inhibitor chlorimipramine increased the duration of the sympathoexcitatory response evoked from the raphe but not from the lateral medulla. Finally, mid-collicular transection did not effect the excitation of sympathetic activity elicited by stimulation of medullary raphe nuclei. These data suggest that serotonergic neurons in the midline medullary raphe nuclei provide an excitatory input to sympathetic neurons in the spinal cord.     

Responses of cardiac vagus and sympathetic nerves to excitation of somatic and visceral nerves

Somato-vagal and somato-sympathetic reflex responses were studied by recording simultaneously the activity of cardiac vagal and sympathetic efferents following excitation of various somatic (and 1 visceral) nerves in chloralose-anesthetized dogs.Stimulation of pure cutaneous (infraorbital, superficial radial, sural nerves), muscle (gastrocnemius, hamstring nerves) and mixed nerves (sciatic, brachial, intercostal, spinal) with short trains of pulses inhibited the activity of cardiac vagus nerve and excited that of cardiac sympathetic nerve after a latency of approximately 40–60 ms, depending on the nerve stimulated. These responses were followed by the opposite response, i.e. excitation of vagus and long-lasting inhibition (`silent period') of sympathetic nerve activity. These biphasic reflex responses recorded from both autonomic nerves had similar latencies so that a clear reciprocal relationship was observed. In addition to the above reflex responses which were observed in most instances, two peaks of excitation of short duration were recorded from the vagus nerve, in some instances, and an ‘early (spinal) reflex’ in sympathetic nerve was also observed. Both excitatory and inhibitory responses described above in either nerve were readily evoked by excitation of Group II (Aβ), but not Group I (Aα), afferent fibers and increased in magnitude when Group III (Aδ) afferents were also excited. Group IV (C) afferent contributed insignificantly to the somato-vagal reflex. The vagus nerve discharge evoked by sinus nerve stimulation was inhibited during reflex inhibition produced by somatic nerve stimulation. The latency of such inhibition was less than 20 ms and lasted for 100 ms after sural nerve stimulation. We conclude that, as in case of the baroreceptor reflex and autonomic component of the ‘defense reaction’, the somato-vagal and somato-sympathetic reflex responses are reciprocal in nature.     

Neural organization of the viscero-cardiac reflexes

The reflex responses evoked in the postganglionic nerves to the heart were tested in chloralose-anaesthetized cats. Electrical stimulation of the A delta afferent fibres from the left inferior cardiac nerve evoked spinal and supraspinal reflex responses with the onset latencies of 36 ms and 77 ms respectively. The most effective stimulus was a train of 3-4 electrical pulses with the intratrain frequency of 200-300 Hz. Electrical stimulation of the high threshold afferent fibres (C-fibres) from the left inferior cardiac nerve evoked the reflex response with the onset latency of 200 ms. The C-reflex was present in intact animals and disappeared after spinalization. The most effective stimulus to evoke this reflex was a train of electrical pulses delivered at a frequency of 1-2 Hz with an intratrain frequency of 20-30 Hz. The most prominent property of the C-reflex was its marked increase after prolonged repeated electrical stimulation. We conclude that: (1) viscero-cardiac sympathetic reflexes may be organized at the spinal and supraspinal level; (2) viscero-cardiac sympathetic reflexes evoked by stimulation of the A delta and C afferent fibres from the left inferior cardiac nerve have different central organization.     

Spinal segmental sympathetic outflow to cervical sympathetic trunk, vertebral nerve, inferior cardiac nerve and sympathetic fibres in the thoracic vagus

The spinal segmental localization of preganglionic neurons which convey activity to the sympathetic nerves, i.e. vertebral nerve, right inferior cardiac nerve, sympathetic fibres in the thoracic vagus and cervical sympathetic trunk, was determined on the right side in chloralose anaesthetized cats. For that purpose the upper thoracic white rami were electrically stimulated with a single pulse, suprathreshold for B and C fibres, and the evoked responses were recorded in the sympathetic nerves. The relative preganglionic input from each segment of the spinal cord to the four sympathetic nerves was determined from the size of the evoked responses. It was found that each sympathetic nerve receives a maximum preganglionic input from one segment of the spinal cord (dominant segment) and that the preganglionic input gradually decreased from neighbouring segments. The spinal segmental preganglionic outflow to the cervical sympathetic trunk, thoracic vagus, right inferior cardiac nerve and vertebral nerve gradually shifted from the most rostral to the most caudal spinal cord segments. In some cases, a marked postganglionic component was found in the cervical sympathetic trunk. It was evoked by preganglionic input from the same spinal cord segments which transmitted activity to the vertebral nerve. These results indicate that there is a fixed relation between the spinal segmental localization of preganglionic neurons and the branch of the stellate ganglion receiving the input from these neurons.     

Separation of the medullo-spinal descending pathway for somatic and autonomic outflow in the cat

The present study investigated differences between vestibulo-somatic and vestibulo-sympathetic reflexes, along with differences between somatic and autonomic spino-bulbo-spinal (SBS) reflexes in chloralose-urethane anesthetized cats. Electrical stimulation was applied to the vestibular nerve (V) for a duration of 0.3 ms. The potential responses in the sympathetic renal nerve (RN) and somatic lumbar nerve were recorded simultaneously. Responses were recorded for a variety of conditioning stimulus to testing stimulus intervals, and the results were plotted to form a recovery curve. The recovery curve for the test response from the somatic nerve was very different from that of the sympathetic nerve. Following transection of the lateral part of the thoracic cord, in the case of the sympathetic renal nerve, recorded responses were still present on vestibular and lumbar nerve stimulation, whereas in the case of the vestibulo-somatic and somatic SBS reflexes, the reflex response had disappeared after transection. These findings suggest that sympathetic and somatic reactions as a result of vestibular stimulation have different descending pathways in the spinal cord, and that their physiological characteristics are different.     

Spinal sympathoexcitatory pathways activated by stimulating fastigial nuclei,hypothalamus, and lower brain stem in cats

In anesthetized cats electrical stimulation of a fastigial (medial cerebellar) nucleus caused discharges in sympathetic nerves which were accompanied by cardiovascular system responses, including rise in arterial blood pressure, increased pulse pressure, and tachycardia. The pathways mediating the excitatory autonomic effects of fastigial origin descend bilaterally in the dorsolateral column of the spinal cord in the region between the lateral corticospinal tract and the central grey substance. Similar influences from the hypothalamus and medullary reticular substance are also mediated through fibers descending in the same region of spinal dorsolateral white matter. Local lesions in the spinal cord abolish the discharges in inferior cardiac, splanchnic, and renal nerves which could be evoked from all three sources, and the spontaneous activity of these sympathetic nerves was also reduced. The lesions did not alter intercostal nerve-evoked activity nor prevented its spreading to segments above or below the lesions.     

The conduction velocity of the descending spinal pathway to the renal sympathetic nerve in the cat

The conduction velocity of the descending spinal excitatory pathway to the renal sympathetic nerve was measured in five chloralose-anaesthetised, spinal cats (C1 transection). Electrical stimuli were delivered to the dorsolateral funiculus at three levels between segments C3 and T6, and responses recorded from the ipsilateral renal nerve. Spinal conduction velocity was calculated as 4.4 +/- 0.4 m/s (mean +/- SEM), from the latency difference of renal nerve volleys to stimulation at different cord levels. A contralateral pathway to the renal nerve was identified: this also ran in the dorsolateral funiculus, and crossed below segment T5. It conducted at 4.1 and 7.9 m/s (two cats). Renal preganglionic conduction velocity (greater splanchnic nerve) was 4.1 and 5.0 m/s (two cats). As the renal sympathetic nerve is functionally homogeneous, these conduction velocity measurements are of a functionally-defined sympathoexcitatory pathway.     

Supraspinal regulation of spinal reflex discharge into cardiac sympathetic nerves

(1) In chloralose anaesthetized cats, reflex responses were recorded in inferior cardiac nerves following stimulation of intercostal nerves and hind limb afferent nerves. (2) In 80% of cats, a long latency reflex response alone was recorded, whereas, in the others, a short and long latency response was present to intercostal nerve stimulation. (3) In cats displaying only a long latency somatocardiac reflex response, damage to the ventral quadrant of the ipsilateral cervical spinal cord, through which runs a bulbospinal inhibitory pathway, resulted in the appearance of shorter latency reflexes to intercostal nerve stimulation. Lesions elsewhere in the cervical cord did not do this. (4) The characteristics of the early responses indicated that they were somatosympathetic reflexes and not dorsal root reflexes. (5) The early reflexes remained and the late reflex disappeared on subsequent complete transection of the spinal cord. The early reflexes were therefore spinal reflexes, and suppressed in the animal with cord intact. (6) Lesions at C4, which included a contralateral hemisection and a section of dorsal columns extending into the dorsal part of the lateral funiculus, abolished the inhibition of a sympathetic reflex that followed stimulation of some somatic afferent nerve fibres. These sections did not release the spinal reflex. Therefore, this reflex inhibition was not responsible for the suppression of the spinal somatosympathetic reflex. (7) The descending inhibitory influence on the segmental reflex pathway was not antagonized by strychnine, bicuculline or picrotoxin. (8) The possibility is discussed that the spinal reflex pathway into cardiac sympathetic nerves is tonically inhibited by a bulbospinal pathway originating from the classical depressor region of the ventromedial reticular formation.     

Prolonged depression of pelvic ganglion transmission — a peripheral manifestation of spinal cord transection

Spinal cord transection depressed the bladder's contractile response to pelvic nerve stimulation. This depression was prevented by prior hypogastric nerve section or guanethidine treatment, but was not reversed by the same treatments used after spinal transection. Thus the depression has a peripheral sympathetic basis but is not dependent on ongoing sympathetic nerve activity. Since responses to intra-arterial acetylcholine and dimethylphenylpiperazinium were unchanged by spinal transection, the depression of the pelvic nerve response probably occurs at a ganglion site.     

Characteristics of sympathetic reflexes evoked by electrical stimulation of phrenic nerve afferents.

In chloralose-anaesthetized cats, sympathetic reflex responses were recorded in left cardiac and renal nerve during stimulation of afferent fibres in the ipsilateral phrenic nerve. In cardiac nerve, a late reflex potential with a mean onset latency of 75.6 +/- 13.8 ms was regularly recorded which, in 20% of the experiments, was preceded by an early, very small reflex component (latency between 35 and 52 ms). In contrast, in renal nerve only a single reflex component after a mean latency of 122.1 +/- 13.1 ms was observed. Bilateral microinjections of the GABA-agonist muscimol into the rostral ventrolateral medulla oblongata resulted in a nearly complete abolition of sympathetic background activity and in an 88% reduction of the late reflex amplitude with only small effects on the latency of the evoked potentials. Under this condition, an early reflex component was never observed to appear. After subsequent high cervical spinalization, the residual small potentials which persisted after bilateral muscimol injections were completely abolished and in cardiac nerve an early reflex potential with a mean latency of 45 +/- 10 ms was observed in all but one experiment. The early reflex was therefore referred to as a spinal reflex component which, however, is suppressed in most animals with an intact neuraxis. In the renal nerve a spinal response was only observed in one experiment after spinalization. The results suggest that sympathetic reflexes evoked by stimulation of phrenic nerve afferent fibres possess similar spinal and supraspinal pathways as previously described for somato-sympathetic and viscero-sympathetic reflexes.(ABSTRACT TRUNCATED AT 250 WORDS)     

Patterns of relationship between activity of sympathetic nerves in rabbits and rats

We used frequency domain analysis (power spectra, ordinary and partial coherence and phase spectra) of simultaneously recorded activity of postganglionic sympathetic nerves to investigate the construction of their central generators in rabbits and rats anesthetized with urethane. As found earlier in the cat, power spectra of sympathetic nerve discharge (SND) consisted of a wide-band component (1 to 10 Hz in rabbits and 1 to 15-20 Hz in rats) and superimposed cardiac and respiratory related peaks. The coherence between pairs of SNDs in the cardiac, vertebral, and renal nerves was significant over a wide range of frequencies, from 0 to 6-10 Hz in rabbits, and except for a sharp peak at the heart rate, was not explained by baroreceptor feedback. In rats, the coherence between distant nerves was relatively low (<0.2) except at the cardiac and respiratory frequencies. Analysis of partial coherences for the three nerves in rabbits revealed two main patterns; one characterized by dominance of the cardiac SND generator, and the other by strong coupling of the vertebral and cardiac SNDs, as compared with renal SND. Phase spectra of distant nerves contained a well-defined transportation lag corresponding to a delay of approximately 70 ms between upper and lower thoracic spinal cord segments. At frequencies close to heart rate however, the phase was constant in most experiments indicating that different mechanisms are involved in transmitting wide band and oscillatory components of resting SND. The similarities between sympathetic oscillators in cats, studied previously in great detail, and rabbits preferred in recent behavioral studies allow the translation of knowledge between these two species.     

Analysis of reflex activity in cardiac sympathetic nerve induced by myelinated phrenic nerve afferents

Electrical stimulation of the phrenic nerve afferents evoked excitatory responses in the right inferior cardiac sympathetic nerve in chloralose-anaesthetized cats. The reflex was recorded in intact and spinal cats. The latency and threshold of the volley recorded from the phrenic nerve as well as of the cord dorsum potentials evoked by electrical stimulation of the phrenic nerve indicated that group III afferents were responsible for this reflex. The phrenicocardiac sympathetic reflex recorded in intact cats was followed by a silent period. The maximum amplitude of the reflex discharges was 800 microV, the latency was 83 ms and the central transmission time 53 ms. Duration of the silent period lasted up to 0.83 s. In spinal cats the reflex was recorded 5.5-8 h after spinalization. The maximum amplitude of the spinal reflex discharges ranged from 22 to 91 microV and the latency from 36 to 66 ms.     

Autonomic nerve and cardiovascular responses to changing blood oxygen and carbon dioxide levels in the rat

Contribution of autonomic nervous system activity to the heart rate and blood pressure responses during chemoreceptor excitations by systemic hypoxia and hypercapnia and to hyperoxia and hypocapnia was analyzed in the urethane-anesthetized, artificially ventilated rats. Systemic hypoxia induced a co-activation of two antagonistic nerves: an increase in cardiac sympathetic and in cardiac vagal efferent nerve discharges. Increased heart rate was due to predominance of the cardiac sympathetic over the cardiac vagal activation. In spite of a marked reflex increase in the renal and cardiac sympathetic nerve activities, the local vasodilator effect of hypoxia prevented consistent changes in arterial blood pressure. Bilateral section of the carotid sinus nerves (CSN) mostly abolished autonomic nerve responses and produced a profound decreases in the blood pressure during hypoxia. Hyperoxia elicited a pressor response due to peripheral vasoconstriction with no significant change in the autonomic nerve activities except for a decrease in the cardiac sympathetic nerve discharges. Hypercapnia significantly increased blood pressure and renal nerve sympathetic activity. In contrast to hypoxia, hypercapnia excited cardiac sympathetic and inhibited cardiac vagal activity. This reciprocal effect did not elicit neurogenic cardioacceleration, because it was masked by the local inhibitory action of CO2 on the heart rate. The increase in sympathetic activities and in blood pressure during hypercapnia persisted after bilateral CSN section indicating that the responses were mediated by central rather than by peripheral chemoreceptors. Hypocapnia produced a significant increase in cardiac vagal discharges yet a cardioacceleratory response occurred due to the local effect upon heart rate. The present results indicate that in the rat, autonomic nervous responses differ depending on the type, i.e. hypoxic or hypercapnic, chemoreceptor stimuli. Reflex heart rate and blood pressure responses do not follow the autonomic nerve activities exactly. Circulatory responses are greatly modified by local peripheral effects of hypoxic, hyperoxic, hypocapnic or CO2 stimuli on the cardiovascular system. Species differences characterizing the autonomic nerve responsiveness to chemical stimuli in the rat are described.     

An anomalous vagorenal reflex pathway in the cat

Although physiological investigations support the view that the innervation to the kidney is primarily sympathetic in origin, there is anatomic evidence suggesting direct vagal projections to the kidney. We examined electrophysiologically the possibility that neural connections exist between the cervical vagus and renal nerves. Electrical stimulation of the peripheral segment of the cut cervical vagus evoked electrical activity in the central segment of cut renal nerve of chloralose-anesthetized, paralyzed cats. The evoked potentials (vagorenal responses) displayed components with peak latencies of about 50, 120, and 500 ms. Another peak at about 175 ms was also seen in some cases. In addition, a period of postexcitatory depression occurred between approximately 180 and 400 ms after delivery of the stimulus. Evoked responses were recorded in the contralateral as well as the ipsilateral renal nerves. In contrast, stimulation of the central cut end of renal nerves did not elicit responses in the cervical vagus. Vagorenal responses were not altered by cutting the subdiaphragmatic vagus indicating that the abdominal vagus was not involved in this response. Electrical activity in renal nerves elicited by vagal stimulation could be eliminated by either ganglionic blockade or by cutting or cooling the splanchnic nerves. Finally, supraspinal ischemia abolished the vagorenal response. These data suggest that a vagorenal reflex pathway exists and that the potentials recorded in renal nerves are due to activation of aberrant sensory fibers traveling from the peripheral segment of the cut cervical vagus to the central nervous system, where they excite a sympathetic efferent pathway to the kidney.     

The involvement of serotonin neurones in the inhibition of renal nerve activity during desynchronized sleep

Numerous 5-hydroxytryptamine (5-HT)-containing cell bodies were visualized by fluorescence microscopy in the caudal brainstem rostal to the decussation of the pyramids in a region from which a desynchronized sleep-like pattern of sympathetic activity was obtained in a previous study. In unanaesthetized mid-collicular decerebrated cats recordings were made of sympathetic activity in a renal nerve. The inhibition of renal nerve activity occurring during desynchronized sleep-like state induced by physostigmine was attenuated significantly by procedures which interfered with the pathways from the 5-HT-containing neurones. Small cuts in the dorsolateral funiculus of the cervical spinal cord reduced the inhibition from43 ± 6%to14.0 ± 3%. Microinjection of 5,7-dihydroxytryptamine into cervical spinal cord reduced the serotonin content of the thoracic cord by 22.4% and attenuated the desynchronized sleep-like state inhibition of renal nerve activity by a similar amount. Depletion of serotonin withp-chlorophenylalanine significantly reduced the inhibition of renal nerve activity during the desynchronized sleep-like state, from42.5 ± 5%to10.0 ± 2.0%. It was suggested that serotonin-containing neurones are likely to be involved in the inhibition of renal nerve activity occurring during desynchronized sleep.     

Sympathetic neural outflow in spinal man. A preliminary report

Microelectrode recordings were made from peroneal skin or muscle fascicles in 11 patients with traumatic spinal cord transection. Spontaneous neural activity was sparse. Both in skin and muscle nerve recordings pressure on the abdomen over the bladder, electrical skin stimulation and skin pinching below the level of the lesion gave rise to single bursts of multiunit impulse activity occurring after latencies of 0.5-1 s. Bursts were never evoked by sound stimuli or by skin stimuli applied above the level of the lesion. Bursts were unaffected by local anesthetic blocks of the nerve distal to the recording site but were abolished by proximal blocks. Conduction velocity of the bursts was approximately 0.7 m/s. After a latency of 2-5 s peroneal neural bursts were followed by cutaneous vasoconstriction and/or skin resistance reduction. Durations of vasoconstrictor responses were often much longer than in intact subjects. It is concluded that the activity was sympathetic and comprised of vasoconstrictor and sudomotor impulses of spinal origin. The parallel activation of bursts in skin and muscle fascicles suggests that sympathetic outflow from the isolated spinal cord is less differentiated than when supraspinal connections are intact. This, together with the long duration vasoconstrictor responses may contribute to the increased blood pressure during attacks of 'autonomic hyperreflexia'.     

A sympathetic projection from sacral paravertebral ganglia to the pelvic nerve and to postganglionic nerves on the surface of the urinary bladder and large intestine of the cat

Anatomical and electrophysiological experiments have demonstrated a prominent projection from the sacral sympathetic chain via the pelvic nerve to postganglionic nerves on the surface of the urinary bladder and the large intestine of the cat. Retrograde labeling studies revealed that the pelvic nerve, which is generally believed to carry primarily parasympathetic axons, has a considerable population of sympathetic fibers originating mainly from the S1-S3 paravertebral ganglia. The number of sympathetic neurons projecting to the pelvic nerve (2,100) was about 75% of that projecting to the pudendal nerve (2,900), a somatic nerve which would be expected to carry a large sympathetic fiber constituent. Sympathetic neurons projecting to the pudendal nerve were located primarily in the L6-S2 ganglia. Electrophysiological studies confirmed the presence of a sympathetic pathway from the paravertebral ganglia to the pelvic viscera. Electrical stimulation (thresholds 1.5-3 V) of the lumbar sympathetic chain evoked firing in the pelvic nerve and in postganglionic nerves on the surface of the colon and bladder at latencies of 60-150 msec. The responses were unaffected by cutting the chain one segment rostral to the site of stimulation, but were abolished by the administration of a ganglionic-blocking agent (tetraethylammonium). The responses on the colon and bladder postganglionic nerves were also abolished by transection of the pelvic nerve. The conduction velocity in the sympathetic postganglionic axons was approximately 1 m/second. In summary, these studies indicate that the pelvic nerve, like somatic nerves, receives a prominent projection from the sympathetic chain ganglia. The function of this sympathetic paravertebral pathway and its relationship with prevertebral innervation of the pelvic organs remains to be established.     

Respiratory modulation of sympathetic activity

Sympathetic activity recorded from cardiac and renal nerves was correlated with phrenic and internal intercostal nerve activity under normocapnea and hypercapnea. Cats were anesthetized with halothane for surgery switching to chloralose for recording. Both vagal and carotid sinus nerves were cut, animals were paralyzed and artificially ventilated. We found that sympathetic activity followed the rhythmic pattern of phrenic nerve discharge fairly closely except in two important respects: first, sympathetic activity was significantly depressed during early inspiration and second, it reached a minimum during post inspiration while phrenic activity was decaying but still active. These effects were accentuated when PACO2 was raised. In one cat early inspiratory depression was the only manifestation of respiratory modulation of sympathetic activity superimposed on an otherwise tonic pattern. In 4 cats sympathetic activity increased in an augmenting fashion in parallel with the augmenting discharge of expiratory alpha motoneurones. We suggest that respiratory-related, excitatory and inhibitory inputs modulate sympathetic activity at the brainstem level. Inspiratory and possibly expiratory interneurones may be the source of activation, and inhibitory inputs may derive from early inspiratory and postinspiratory interneurones. The inhibitory effects may be the only manifestation of respiratory modulation during strong tonic drive of the sympathetic activity.     

Spinal segmental preganglionic outflow to cervical sympathetic trunk and postganglionic cardiac sympathetic nerves

The aim of this study was to verify in which spinal cord segments the preganglionic neurones projecting to the cervical sympathetic trunk or converging onto the somata of the postganglionic cardiac sympathetic neurones are located in cats. The thoracic white rami T1 to T5 were electrically stimulated and the evoked responses were recorded in the cervical sympathetic trunks and postganglionic cardiac nerves. The responses were mostly evoked by electrical stimulation of group B preganglionic fibres. The maximum amplitude of evoked responses in the cervical sympathetic trunk was obtained when the T2 white ramus was stimulated and decreased gradually when followed by the stimulation of T1, T3, T4 and T5 white rami. In most cases the maximum amplitude of evoked responses in the left inferior cardiac nerve, right inferior cardiac nerve and left middle cardiac nerve was obtained when the T3 white ramus was stimulated. The size of the responses decreased when more cranial and caudal white rami were stimulated. It was found that the somata of the postganglionic neurones of the right and left inferior cardiac nerves were placed in the right and left stellate ganglion, respectively. Somata of the postganglionic neurones with axons in the left middle cardiac nerve were mainly located in the left middle cervical ganglion and some in the left stellate ganglion.