Ultrastructural Changes in the Nerve Fiber Population ofAnastomosed Vagal and Spinal Accessory Nerves in the Sheep

PIERRE DELORME,1* ANTOINETTE ROUSSEAU,2 JACQUELINE BERNARD,1BARRY F. LEEK,3 AND JEAN-PAUL ROUSSEAU2

1Laboratoire de Neurobiologie Fonctionnelle, UST Lille, Villeneuve d’Ascq Cedex, France2Laboratoire de Physiologie de la Reproduction, C.N.R.S./I.N.R.A.,

Universite Pierre et Marie Curie, Paris Cedex, France3Department of Veterinary Physiology and Biochemistry, University College of Dublin,

Dublin, Ireland

ABSTRACT Background: The ultrastructure of the vagal and spinalaccessory nerves was studied 1) in normal sheep and 2) in sheep in whichan experimental crossed-nerve anastomosis had been made by sectioningthe supranodose vagal and spinal accessory nerves, then suturing thedistal end of the vagal nerve to the distal end of the spinal accessory nerve,and allowing time for regeneration to occur. This study was carried out inorder to analyze the modifications liable to occur when this technique isused and to specify the origin and the nature of the fibers that colonize thespinal accessory nerve.Methods: The study was performed in 4- to 5-month-old sheep. After the

surgical procedure, the animals were housed indoors during 1 year untiltheir sacrifice by fixative perfusion. Then, nerve samples were dissectedout, processed for electron microscopy, examined, and systematicallyphotographed. After printing, the diameters of the nerve fibers weredetermined.Results: In sheep, the ratios of nonmyelinated to myelinated fibers

(NF/MF) in the infranodose and supranodose vagal nerve and accessoryspinal nerve were 1.21, 1.67, and 3.21, respectively. In both parts of thevagal nerve, the myelinated fibers had a unimodal diameter distributionaround a peak of 4 mm; whereas, in the spinal accessory nerve, they weredistributed bimodally, and 53% had values of 15–18 mm. After making theabove anastomosis, the centrifugal vagal fibers degenerated, and theNF/MF ratios increased in the centripetal infranodose vagal nerve, in thereinnervating supranodose vagal nerve, and in the reinnervated spinalaccessory nerve (approximately 1.87, 1.72, and 6.04, respectively). In all ofthese nerves, the myelinated fibers had a unimodal distribution with apeak at 4 mm, as in the vagal nerve of normal sheep.Conclusions: These results reveal the large part taken by the nonmyelin-

ated fibers in the nerve fiber population of the vagal nerve and support thevagal origin of the fibers reinnervating the spinal accessory nerve. Anat.Rec. 248:129–136, 1997. r 1997 Wiley-Liss, Inc.

Key words: heterogeneous nerve anastomosis; vagal nerve; spinal acces-sory nerve; ultrastructure; nerve fiber populations

To study vagal receptor activity in conscious sheep,Rousseau (1970, 1973) conceived and successfully devel-oped a method of heterogeneous nerve anastomosisbetween the peripheral stump of a cut supranodosevagal nerve (SVN) and the peripheral stump of theipsilateral spinal accessory nerve (SAN). Allowing timefor nerve regeneration and muscle reinnervation en-ables centripetal vagal receptor activation to be re-flected in activation of motor units of a superficialskeletal muscle in the neck. With this technique (Fig.1A,B), the peripheral stump of the SVN end, sectioned

as cranially as possible, is sutured to the peripheralstump of the SAN, the SAN being themotor nerve of themastoidohumeral muscle in sheep. After an interval ofseveral months, there is partial reinnervation of thisdenervated muscle by certain centripetal vagal axons.

Received January 9, 1995; accepted October 28, 1996.*Correspondence to: Professor Pierre Delorme, Laboratoire de Neu-

robiologie Fonctionnelle, UST Lille, F.59655 Villeneuve d’Ascq Cedex,France.Contract grant sponsor:M.E.N., andCNRS;Contract grant no: V.A. 308.

THE ANATOMICAL RECORD 248:129–136 (1997)

r 1997 WILEY-LISS, INC.

The term ‘‘centripetal vagal axons’’ is used to includeafferent axons that constitute the majority of centrip-etal axons as well as a small number of recurrentcontralateral vagal motor axons whose existence wasdemonstrated by Dalal et al. (1988).Electrophysiological studies have confirmed the func-

tional reality of reinnervation (Coget and Rousseau,1983; Falempin and Rousseau, 1983). Of the 26,000–36,000 neurons located in the nodose ganglion, only oneafferent vagal axon out of every 600 is able to reestab-lish neuromuscular junctions (Rousseau and Falempin,1984).Before determining the ultrastructure of the vagal

fibers that reinnervate the SAN, fiber populations ofthe SVN and the infranodose vagal nerve (IVN) wereanalyzed. Previously, the vagal nerve of sheep has beenthe object of light microscopic studies that were unableto reveal the contribution of the nonmyelinated frac-tion. For this reason, we have analyzed first the normalnerves (IVN, SVN, SAN; Fig. 1A) and, then, the anasto-mosed nerves [reinnervating centripetal IVN (RCIVN),reinnervated SAN (RSAN); Fig. 1B] and the vagusnerve after the degeneration of the efferent fibers[centripetal IVN (CIVN); Fig. 1C].

MATERIALS AND METHODS

Nerve anastomoses (Fig. 1B) were performed on theleft side in 4- to 5-month-old sheep of the ‘‘Prealpes duSud’’ breed following the procedure described by Falem-pin and Rousseau (1983). After the surgical procedure,the sheep were housed indoors, allowed free access towater, and fed a diet of meadow hay ad libitum plus 300g/day of pellets. One year after the anastomosis, thereinnervation of themastoidohumeral muscle was dem-onstrated in each sheep by recording the electricalactivity of reinnervated motor units, which was evokedby stimulating the anastomosed vagal nerve in itscervical. The three animals were killed immediatelyafterward. The vagus nerve in its infranodose part up tothe suture site and to the peripheral SAN stump weredissected. We checked that the suture site was entirelydisconnected from the surrounding tissues, namely,from the proximal ends of both nerves, which could stillbe easily located by the presence of clips.To determine the population of the centripetal vagal

contingent, the left vagal nerve was sectioned above thenodose ganglion (Fig. 1C) in a fourth sheep. This sheepwas 15 months old at the time of the vagal section and

Fig. 1. Vagal and spinal accessory nerves located in the neck of thesheep. Schematic illustrations of the normal anatomical structure (A),the heterogeneous anastomosed nerve after 1 year (B), and the sectionof the supranodose vagal nerve without anastomosis shown after 40days (C). The numbers show the nerve samples taken. Their analysesare reported in this order in Table 1. IVN, infranodose vagal nerve;NG, nodose ganglion; SVN, supranodose vagal nerve; SAN, spinal

accessory nerve; MHM,mastoidohumeral muscle; RCIVN, reinnervat-ing centripetal infranodose vagal nerve; RSAN, reinnervated spinalaccessory nerve; CIVN, centripetal infranodose vagal nerve. Longdashed line, vagal centripetal (motor) fibers; Dashed-and-dotted lines,spinal accessory motor fibers; solid lines, vagal centrifugal (sensory)fibers; short dashed line, degenerated fibers.

130 P. DELORME ET AL.

was born in the same breeding and during the sameperiod as the three other sheep. It was killed 40 daysafter the operation.All experiments were carried out in compliance with

the principles of laboratory animal care and the Frenchlaws on the protection of animals. In particular, opera-tive techniques were performed under general anesthe-sia. An intravenous injection (200 mg of sodium pento-barbitone and 500 mg of sodium thiopentone dissolvedin 25ml of physiological saline) induced a brief anesthe-sia, which allowed the insertion of an endotrachealtube. Then, anesthesia wasmaintained by inhalation of0.5–2.0% halothane in air.The tissues of the neck and the head were fixed by

perfusion through the left common carotid artery of 4liters of a glutaraldehyde (2.5%) paraformaldehyde(0.5%) solution in 0.2 M sodium cacodylate buffer, pH7.4. The nerves were dissected out, fixed for another 2hours by immersion in the same fixative, and then cutup to a size compatible with processing for electronmicroscopy.The fragments were washed with cacodylate buffer,

postfixed for 1 hour in 1% buffered osmium tetroxide,then dehydrated in graded alcohols and propyleneoxide, and finally embedded in Epon 812. Semithin orultrathin sections were cut on a Reichert OM U3Ultrotome. Semithin sections were stained with tolu-idine blue to allow location of the fiber bundles (Fig.2A). Ultrathin sections collected on 75 mesh gridscoated with Formvar film were contrasted with uranyl

acetate and lead citrate and examined with a SiemensElmiskop 1A electron microscope.The unavoidable presence of the grid wires (Fig. 2B)

masking areas of greater or lesser importance did notallow the complete study of the nerves: Only thebundles showing a sufficiently free area were analyzed.The analysis was completed by observations of theadjacent sections. Therefore, the choice of the analyzedbundles was purely arbitrary. The visible parts of thebundles were first carefully drawn and were thensystematically photographed at an initial magnifica-tion of 32,000, the photographs partly overlapping inorder to facilitate the reconstruction of the nerve sec-tion. After printing, the bundles were reconstructed,and the diameters of the nonmyelinated fibers (NFs)and the myelinated fibers (MFs) were determined,taking into account the final magnification of 34,000.Because the section of a fiber was rarely round, wedecided to determine its diameter by taking the averageof a long and short axis, including the myelin sheath, tomake our results more comparable with those obtainedfrom light microscopic observations (Tiao and Blake-more, 1976).Nerve samples of the three sheep with nerve anasto-

mosis were observed in the electron microscope, but, inconsideration of the vast amount of work, morphomet-ric studies were only performed from IVN, SVN, SAN,RCIVN, and RSAN samples of one of the sheep; thesamples from the other two sheep were used only for

Fig. 2. Ademonstration of the extent to which the nervous tissue sections may bemasked by the wires ofa support grid. A: Semithin section from a spinal nerve observed under light microscopy. B: The samesection as it appears when it is displayed on a grid.

131NERVE FIBER POPULATIONS AFTER HETEROGENEOUS NERVE ANASTOMOSIS

qualitative observations. Morphometric studies werealso carried out from the CIVN sample of the fourthsheep, including the supranodose section of the vagusnerve, to compare its composition and distribution ofthe centripetal fibers with those of the RCIVN samplein the infranodose level.

RESULTSQualitative Aspects

The distribution of the nerve fibers in the normal IVNor SVN (Fig. 3A) is not homogeneous. This heterogene-ity is due not only to the differences between the fiberdiameters but also to the repartition of these variousfibers among the different nerve bundles and insideeach bundle. The relatively small diameter of mostfibers that compose this nerve results in a high-densityfiber population. On the contrary, in the normal SAN(Figs. 2A, 3B), the distribution appears much morehomogeneous. Density of fibers is relatively low, as thefibers of larger diameter are the most numerous. Thisexplains why, for nearly equivalent nerve areas ana-lyzed, the fiber count was very different between sam-ples of normal vagus nerve and normal SAN (Table 1).Observation of the sectioned then anastomosed SAN

(Fig. 1B) reveals a very collagenous endoneurium di-vided into small fascicles by the long, thin processes ofendoneurial cells (Fig. 3D). Inside these thin-walledmuffs, small- and middle-diameter vagal fibers, bothMFs and NFs, alone or in groups of two or three,colonize the peripheral stump of the SAN, which is nowemptied of its own axons. Similitude of observations inthe samples of the three sheep, including reinnervationof vagal fibers into the SAN, led us to carry out themorphometric studies on the same animal.

Quantitative Aspects

Normal vagal nerve

The normal vagal nerve was sampled central to thenodose ganglion (i.e., SVN) and peripheral to the gan-glion (i.e., infranodose ganglion; IVN). Analysis of theIVN indicated that, of the 7,589 fibers measured, 3,443weremyelinated (i.e., 45% of the total; Table 1). TheMFdiameter distribution was unimodal (Fig. 4A). Thisfiber population was particularly concentrated between2 µm and 7 µm. The NF distribution was also unimodal:The diameters of the NF fibers ranged from 0.25 to 2µm, with a peak close to 1.5 µm (Fig. 4A).In the SVN (Fig. 4A), the proportion of MFs was

slightly lower than in the IVN, because these fibersrepresented only 37% of the total population (3,720 of9,935; Table 1). The diameter ranges over which bothfiber types were distributed were nearly the same as inthe infranodose part, although the MF peak was againlocated at 4 µm, whereas the NF peak had shifted to0.75 µm.

Sectioned SVN

The vagal nerves were sectioned central to the nodoseganglion with and without creating an anastomosis. Inboth cases, the centrifugal efferent fibers degenerated,so that the peripheral stump sampled distal to thenodose ganglion would contain only centripetal fibers.In the CIVN (Fig. 4C), the MFs represented 35% of

the fiber population. It was slightly lower than the 37%

of the IVN population. Their distribution was alwaysunimodal, but the peak shifted to 3 µm. On the otherhand, the NF population displayed a greater range(0.25–4.0 µm), and its peak was shifted toward slightlygreater diameters (1.5–1.75 µm).In the RCIVN (Fig. 4B), the MF diameter distribu-

tion was like that in the IVN, but the most importantcomponent of the MF population, although it was stillcentered at 4 µm, fell from 32% to 24% of the total.Moreover, the NF/MF proportion was increased from1.2 to 1.72 (2,020 of 5,514 RCIVN fibers analyzed; Table1). The NF distribution ranged from 0.25 to 2.5 µm, butit did not have a well-defined peak.

Normal SAN

Analysis of the SAN demonstrated that the MFcomponent was smaller than the NF component (Fig.5A): Indeed, it represented only one-fourth (24%) of thefibers measured (Table 1). The total fiber population ofthe SAN can be estimated as about 9,600, becauseFalempin (1981) estimated the total number of MFs tobe about 2,300. The MF diameter spectrum was bimo-dal, with a lower peak (16% of theMFs) of 6–8 µm and ahigher peak (53% of the MFs) of 15–18 µm. The NFswere distributed between 0.25 and 1.25 µm, with a peakof 0.75 µm.

RSAN

After anastomosing the distal stump of the sectionedvagal nerve to the distal stump of the sectioned SAN,the regenerating axons of the vagal centripetal neuronsinvaded the empty endoneurial fascicles of the degener-ated spinal accessory motor nerve fibers (RSAN; Fig.5B). In the RSAN, the NF spectrum was a littlemodified; its peak of 0.5 µm had slightly shifted com-pared with the NF peak in the normal SAN.The MF spectrum had changed substantially. The

bimodal MF distribution in the normal SAN becameunimodal in the RSAN, with a peak at 4 µm. This peakhad the same location as that of the SVN, IVN, andCIVN. Comparison of the MF sizes in the differentphotomicrographs of Figure 3 illustrates this fact. Themodal MF percentage in the normal SAN (24%) is evenlower in the RSAN and corresponds to only 14% of thetotal fiber population (Table 1).

DISCUSSION

The aim of this study was to analyze and to comparethe populations of MFs and NFs in the normal vagalnerve and SAN with nerves in which a peripheralvagoaccessory nerve anastomosis had been made (Fig.1B) as a means of recording afferent vagal activity inconscious sheep (Rousseau, 1973). The relatively heter-ogeneous distribution of the various nerve fibers in the

Fig. 3. Ultrastructural aspect of the myelinated (MFs) and nonmy-elinated (NFs) nerve fibers in bundles from the normal vagal nerve(A), the normal spinal accessory nerve (B), the reinnervating centrip-etal infranodose vagal nerve (C), and the reinnervated spinal acces-sory nerve (D). Sometimes, Schwann cells (S) are observed near nervefibers that are surrounded by abundant collagen (Co). The arrow in Dindicates the thin processes of the endoneurial cells that form thesmall endoneurial fascicles. The MF diameters of the RSAN (D) arelike those of the SVN and RCIVNmyelinated fibers (A,C). Scale bars52 µm.

132 P. DELORME ET AL.

Fig. 3.

133NERVE FIBER POPULATIONS AFTER HETEROGENEOUS NERVE ANASTOMOSIS

vagal nerves induced us to analyze more bundles in thisnerve than in the SAN, the fiber distribution of whichseemedmore homogeneous. To obtain a reasonably truefiber representation, all the workable bundles thatwere not masked by the grid wires were photographedand were then reconstituted before analysis.

Composition of the Normal Vagus Nerve

Estimation of the total numbers

The MF distributions in the SVN and the IVN werenear enough to those that were described by Iggo (1956)or Dussardier (1960) from light microscopy observa-tions. These authors found that the great majority ofthe MFs (about 80%) had diameters between 2 and 4µm, and, in our study, a similar percentage pertained tothe fibers, with a peak of 4 µm. Such a unimodal distri-

bution of vagal MFs is not exceptional, because it is alsofound in other species, such as the cat (Burgh-Daly etal., 1953;Mei et al., 1980) and the ferret (Asala andBrower,1966), whereas Bronson et al. (1978) have described anMF bimodal distribution in the rat vagal nerve.The originality of this study lies mostly in the fact

that the NF component in normal vagal nerves of sheepwas estimated. The NFs represent 55% of the fibers inthe infranodose part and represent 63% in the suprano-dose part (Table 1). These proportions are considerablygreater than the 10–30% of the NFs estimated inthoracic vagal nerve by Habel (1956). The only previousstudies (Iggo, 1956; Dussardier, 1960) on vagal nervesin sheep used light microscopy and led to an under-estimation of the NF component of these nerves. How-ever, this component in sheep is less important than

TABLE 1. Populations of nerve fibers in the various nerves studied

No. ofsample inFigure 1 Analyzed nerves

No. ofanalyzedbundles

Total no. (TN) ofmeasured fibers

Nonmyelinatedfibers (NF)

Myelinatedfibers (MF)

NF/MFNumber % of TN Number % of TN

1 Normal infranodose vagus nerve (IVN) 5 7,589 4,146 55 3,443 45 1.22 Normal supranodose vagus nerve (SVN) 5 9,935 6,215 63 3,720 37 1.673 Normal spinal accessory nerve (SAN) 3 1,386 1,057 76 329 24 3.214 Reinnervating centripetal IVN (RCIVN) 4 5,514 3,494 63 2,020 37 1.725 Reinnervated SAN (RSAN) 5 3,799 3,260 86 539 14 6.046 Centripetal infranodose vagus nerve (CIVN) 16 16,613 10,842 65 5,771 35 1.87

Fig. 4. Comparisons of the distributions (expressed as percentages ofthe total fibers) of the myelinated (MFs) or nonmyelinated (NFs) nervefibers between the normal infranodose (IVN) and supranodose (SVN)vagal nerves (A), between the IVN and the reinnervating centripetalinfranodose vagal nerve (RCIVN; B), between the IVN and the

centripetal infranodose vagus nerve (CIVN; C), and between theRCIVN and the CIVN (D). Each pie chart indicates the number andthe percentage of the analyzed MFs and NFs for each nerve. Thenumbers above the pie charts summarize the total number of fibersanalyzed in each nerve studied.

134 P. DELORME ET AL.

that observed by electron microscopy in the cat vagalnerves, where it represents 68–85% of the infranodosefibers and 86–90% of the supranodose fibers (Mei et al.,1980). The relative wealth of MFs in the vagal nerves ofsheep could be accounted for by the large innervation ofthe complex stomach of ruminants, the sequentialmotility of which is triggered once every 45 seconds bythe medullary gastric centers through the vagal nerves.From our observations, the total number of fibers in

the cervical vagal nerve could be determined onlyapproximately. The number of bundles that we haveanalyzed corresponded to about one-sixth of the wholearea of the IVN section. Therefore, we can estimate thetotal nerve fiber population of the IVN to be about45,000 fibers. Dussardier (1960) estimated that therewere 10,000–12,000 MFs present in each dorsal tho-racic vagal nerve and 8,000–10,000 in the commonventral thoracic vagal trunk. Taking into account the45% of MFs that we found in the IVN and the observa-tion that the IVN contains the fibers from one of thedorsal thoracic vagal nerves and one-half of those fromthe ventral thoracic common vagal trunk, we cancalculate from Dussardier’s results the total populationof the IVN to be 31,000–38,000 fibers. Consequently,the results of Dussardier (1960) represent an under-estimation, probably because of the difficulty in count-ing the NFs by light microscopy.

Supranodose vs. infranodose

The different composition of the SVN and the IVNmust be emphasized. The SVN contains more NFs thanthe IVN, the NF/MF ratio being 1.67 against 1.2 (Table1). The difference between these two levels is plainlyless marked than in the cat, in which the NF/MF ratiosare 7 to 9 and 4 to 6 for the supranodose and infrano-dose parts of the vagal nerves, respectively (Mei et al.,1980). This comparison confirms the more importantmyelinization of cervical vagal nerve fibers in sheep.Similar to cat vagal nerves, the unimodal distributionof the NFs presents different maxima when we considerthe whole SVN (0.75 µm) on the one hand and the wholeIVN or its centripetal contingent (1.25–1.5 µm) on theother hand.In the infranodose region, the CIVN remaining after

supranodose nerve section and the degeneration of thecentrifugal (motor) fibers contains relatively fewer MFsthan the normal IVN (35% vs. 45%). The presence of thecentripetal fibers in the normal IVN, therefore, isresponsible for the higher percentage of MFs, suggest-ing that efferent vagal fibers are proportionally moremyelinated than afferent vagal fibers. If the 16,613centripetal fibers (Table 1) that we counted in the CIVNrepresented about two-thirds of the area of this nerve,then we can estimate the infranodose afferent compo-nent to be 25,000 fibers, i.e., about 55% of the estimatedtotal IVN population of 45,000 fibers. Taking intoaccount the whole population, if 45% of all fibers areMFs and 35% of all centripetal fibers are myelinated,then the calculation shows 60% of the infranodoseefferent component to be NFs, i.e., 27% of the totalpopulation in which there are 45% efferent fibers.These estimations are in line with those given byPrechtl and Powley (1990) for the rat. Those authorsconsider that the efferent fiber component of the vagalnerves is more important than it has been proposedpreviously, namely, 10% (Agostini et al., 1957; Evansand Murray, 1954; Kemp, 1973). Compared with theafferent MF spectrum of the CIVN, the IVN MF spec-trum is shifted to the right (Fig. 4C), suggesting thatthe myelinated centrifugal fibers generally have agreater diameter that the centripetal fibers.

Reinnervation of the SAN

The new population, which consists principally ofsmall-diameter MFs, differs markedly from the popula-tion of the normal SAN (compare Fig. 3D with Fig. 3B)and rather resembles the population observed in thenormal vagus nerve (Fig. 3A) or in the reinnervatingIVN (Fig. 3C). The relative similarity of the fiberdistribution (Fig. 4D) and of the MF percentages, 35%and 37%, respectively (Table 1), in the CIVN and inthe RCIVN, is not surprising, because, in both cases,the nerves are deprived of their efferent component.The MF component changes from being bimodal in theintact SAN to becoming unimodal in the RSAN. Distrib-uted around a peak at 4 µm, the spectrum of theirdiameters is identical to that observed in the RCIVN.On the other hand, the NF diameter is smaller in theRSAN than in the RCIVN (Fig. 5B). In other respects,they are plainly more numerous, because, in the RSAN,the NF/MF ratio is 3.5 times greater than in theRCIVN. This situation could be due to the fact that, inthe RSAN, many axons from vagal sensory neurons,

Fig. 5. Comparisons of the distributions of the myelinated (MFs) andnonmyelinated (NFs) nerve fibers between the normal spinal acces-sory nerve (SAN) and the reinnervated spinal accessory nerve (RSAN;A) and between the RSAN and the reinnervating centripetal infrano-dose vagal nerve (RCIVN;B). For details, see the legend to Figure 4.

135NERVE FIBER POPULATIONS AFTER HETEROGENEOUS NERVE ANASTOMOSIS

which colonize the endoneurial fascicles of the recipientnerve, are in the process of maturation and have not yetreached their final size.It is possible to compare the diameter spectrum of the

RCIVN fibers with the distribution of the conductionvelocities measured between two points of stimulationon this nerve (Falempin and Rousseau, 1983). In sheep,themaximum of the evoked responses (44%) is obtainedby the stimulation of fibers with conduction velocities of6–12 meters/second corresponding to fiber diameters ofabout 1–2 µm. Eighteen percent of the responses areevoked by stimulation of fibers of diameters less than 1µm, and 25% are evoked by stimulation of fibers of 2-4µm in diameter. In other words, the diameter spectrum,which is deduced from the distribution of conductionvelocities, is shifted to the left compared with thoseestablished by direct histological observation. Thismeans that, among the vagal centripetal fibers runningbelow the nodose ganglion, about 62% of those that hadreinnervated the neckmuscle are either slightlymyelin-ated or nonmyelinated, with diameters equal or lessthan 2 µm. They would correspond to the peripheralprocesses of sensory neurons, because retrograde label-ing of nerve cells was evidenced in the ipsilateralnodose ganglion following successful reinnervations(Falempin and Rousseau, 1983). They could also becentrally directed recurrent collateral branches of pre-ganglionic efferent gastric fibers from the contralateralvagus nerve, as suggested by subsequent results (Dalalet al., 1988).

ACKNOWLEDGMENTS

The authors thank Yolande Dodey and Sophie Duclosfor their secretarial help and Claudine Guichard for herphotographic assistance. This research was supportedby grants fromM.E.N. and from C.N.R.S. (U.A. 308).

LITERATURE CITED

Agostini, E., J.E. Chinnock, M.D. Burgh-Daly, and J.G. Murray 1957Functional and histological studies on the vagus nerve and itsbranches to the heart lungs and abdominal viscera in the cat. J.Physiol., 135:182–205.

Asala, S.A., and A.J. Brower 1986 An electron microscope study ofvagus nerve composition in the ferret. Anat. Embryol., 175:247–253.

Bronson, R.T., H. Bishop, and E.T. Hedley-White 1978 A contributionto the electron microscopic morphometric analysis of peripheralnerves. J. Comp. Neurol., 178:177–186.

Burgh-Daly, M.D., and D.H.L. Evans 1953 Functional and histologicalchanges in the vagus nerve of the cat after degenerative section atvarious levels. J. Physiol., 120:579–595.

Coget, J., and J.-P. Rousseau 1983 Reinnervation of striated muscle byperipheral vagal fibers cut above or below the nodose ganglion inthe cat and rabbit. J. Physiol., 335:481–493.

Dalal, A., B.F. Leek, and J.-P. Rousseau 1988 Gastric centripetal vagalunitary activity in conscious sheep. J. Physiol., 406:137P.

Dussardier, M. 1960 Recherches sur le controle bulbaire de lamotricitegastrique chez les Ruminants. These Doct. Sci., Paris, 199 pp.

Evans, D.H.L., and J.G. Murray 1954 Histological and functionalstudies on the fiber composition of the vagus nerve of the Rabbit.J. Anat., 88:320–337.

Falempin, M. 1981 Contribution a l’etude des afferences vagalesdigestives chez l’animal eveille. These Doct. Sci., Lille, 211 pp., 79illustrations.

Falempin, M., and J.-P. Rousseau 1983 Reinnervation of skeletalmuscles by vagal sensory fibers in the sheep, cat and rabbit. J.Physiol., 335:467–479.

Falempin, M., and J.-P. Rousseau 1984 Activity of lingual, laryngealand oesophageal receptors in conscious sheep. J. Physiol., 347:47–58.

Habel, R.E. 1956 A study of the innervation of the ruminant stomach.Cornell Vet., 46:555–628.

Iggo, A. 1956 Central nervous control of gastric movements in sheepand goats. J. Physiol., 131:248–256.

Kemp, D.R. 1973 A histological and functional study of the gastricmucosal innervation in the dog. I. The quantification of the fibercontent of the normal subdiaphragmatic vagal trunks and theirabdominal branches. Austral. NZ J. Surg., 43:288–294.

Mei, N., M. Condamin, and A. Boyer 1980 The composition of thevagus nerve of the cat. Cell Tissue Res., 209:423–431.

Prechtl, J.C., and T.L. Powley 1990 The fiber composition of theabdominal vagus of the rat. Anat. Embryol., 181:101–115.

Rousseau, J.-P. 1970 Contribution a l’etude de la reinnervation et del’eructation chez le mouton. These Doct. Sci., Aix-Marseille,156 pp.

Rousseau. J.-P. 1973 Reinnervation du muscle mastoido-humeral parles axones des neurones sensitifs vagaux. J. Physiol. (Paris),67:308A.

Rousseau, J.-P., and M. Falempin 1984 Reinnervation of a striatedmuscle by vagal sensory axons. J. Autonom. Nervous Syst.,10:217–224.

Tiao, Y.C., and C. Blakemore 1976 Regional specialization in goldenhamster’s retina. J. Comp. Neurol., 168:439–458.

136 P. DELORME ET AL.