Proc. NatL Acad. Sci. USAVol. 79, pp. 1303-1307, February 1982Neurobiology

Organizational principles in the peripheral sympathetic nervoussystem: Subdivision by coexisting peptides (somatostatin-,avian pancreatic polypeptide-, and vasoactive intestinalpolypeptide-like immunoreactive materials)

(neuronal organization/peptide multiplicity)

J. M. LUNDBJ RG*, T. HOKFELTt, A. ANGGARDf, L. TERENIUS§, R. ELDEV, K. MARKEYII, M. GOLDSTEIN11,AND J. KIMMEL**Departments of tHistology and *Pharmacology, Karolinska Institutet, Stockholm, Sweden; tDepartment of Oto-Rhinolaryngology, Karolinska Hospital, Stockholm,Sweden; partment of Pharmacology, Uppsala University, Uppsala, Sweden; IDepartment of Anatomy, University of Minnesota, Minneapolis, Minnesota 55455;IlDepartment of Psychiatry, New York University Medical Center, New York, New York 10016; and "Mel Laboratory, University of Kansas Medical Center,Kansas. City, Kansas 66103

Communicated by U. S. von Euler, September 16, 1981

ABSTRACT Sympathetic ganglia and some peripheral tissuesof adult guinea pig and cat were analyzed by the indirect immu-nofluorescence technique with antisera to catecholamine-synthe-sizing enzymes and some peptides. In the guinea pig, noradre-nergic neurons could be subdivided into three populationscontaining respectively (i) somatostatin-like immunoreactive ma-terial, (ii) avian pancreatic polypeptide (APP)-like immunoreactivematerial, and (iii) apparently only noradrenaline (NA; norepi-nephrine). A fourth population of sympathetic neurons was non-adrenergic and contained vasoactive intestinal polypeptide (VIP)-immunoreactive material. In the cat many noradrenergic neuronswith APP and some without this peptide were seen, but no so-matostatin-immunoreactive neurons were observed. Also a pop-ulation of non-adrenergic, presumably cholinergic, neurons con-taining a VIP-like peptide was observed. These neuron, populationsseemed to innervate different tissues with some target specificity.For example, in the nasal mucosa ofthe cat, nerves containing NA/APP-like immunoreactive material (called NA/APP nerves) werefound around small arteries and arterioles, whereas venules andsinusoids were surrounded by nerves containing only NA (calledNA nerves). Also in the submandibular salivary gland of the cat,the NA/APP nerves surrounded arteries and drterioles, whereasNA nerves were seen in relation to acini and ducts. The sympa-thetic (cholinergic) VIP-containing neurons innervated blood ves-sels and exocrine tissue in the cat sweat glands. In thecoeliac-superior mesenteric ganglion complex of the guinea pigand cat, a dense network of VIP-immunoreactive fibers was seenpreferentially around noradrenergic ganglionic cell bodies lackingAPP-immunoreactive material. Thus, adult peripheral sympa-thetic neurons can be subdivided into several categories on thebasis of specific peptides. These subdivisions may innervate spe-cific targets and may receive peptide-specific neuronal inputs.

Classically the peripheral autonomic nervous system can besubdivided into the sympathetic and the parasympathetic por-tions (1, 2). The enteric nerves have been considered to rep-resent a separate division (3). The principal transmitter in thesympathetic division is noradrenaline (NA), as discovered byvon Euler (4). However, a small population of sympatheticnerves supply sweat glands (5, 6) and skeletal muscle vesselsand are cholinergic (7).

During recent years, an extensive distribution ofpeptides inperipheral neurons has been identified (8-10). Further analysishas revealed- that some of these peptides coexist with classical

transmitters in sympathetic neurons. Thus, some peripheralnoradrenergic cell bodies in sympathetic ganglia of the guineapig, particularly prevertebral ganglia, contain a somatostatin-like peptide (11). An avian pancreatic polypeptide (APP)-likepeptide has been demonstrated in a population of peripheralnoradrenaline neurons (12). Finally, the acetylcholine (AcCho)esterase-rich, presumably cholinergic, neurons in sympatheticganglia of the cat contain a vasoactive intestinal polypeptide(VIP)-like peptide (13, 14).

In the present study, the distribution patterns ofthe differentpeptide-containing sympathetic neurons were compared byanalyzing consecutive sections and by means of elution/re-staining experiments (15) with regard to cell bodies and someperipheral termination fields. The results show that the sym-pathetic nervous system may be further subdivided on the basisof the content of a particular peptide.

MATERIALS AND METHODSAntiserum to synthetic somatostatin (Kabi, Stockholm) conju-gated to keyhole limpet hemocyanin was raised in rabbits asdescribed (16). Antisera to highly purified APP and highly pu-rified porcine VIP were raised in rabbits as will be describedelsewhere. The antisera to the catecholamine-synthesizing en-zymes tyrosine hydroxylase [TyrOHase; L-tyrosine, tetrahy-dropteridine: oxygen oxidoreductase (3-hydroxylating), EC1.14.16.2] and dopamine ,f3hydroxylase [Df3OHase; 3,4-dihy-droxyphenylethylamine, ascorbate: oxygen oxidoreductase (,(hydroxylating), EC 1.14.17.1], purified from rat pheochromo-cytoma and cow adrenal glands, respectively, have been de-scribed (17, 18).

Male guinea pigs (body weight, 200-300 g) and adult cats ofboth sexes (body weight 2-3 kg) were used. In two cats, vin-blastine sulphate (Sigma) (0.1% in 0.9% NaCl) was locally ap-plied to the superior cervical and the L7 sympathetic ganglia24 hr prior to sacrifice. This treatment increases transmitter andenzyme content in cell bodies (19). Two alternative fixationtechniques were then used. Some animals were perfused with10% Formalin in 0.1 M phosphate buffer for 30 min, followedby 90 min ofimmersion. Alternatively, the tissues were rapidlydissected out, and slices were immersed in ice-cold2% Formalinin 0.1 M phosphate buffer with 0.25% parabenzoquinone

Abbreviations: AcCho, acetylcholine; APP, avian pancreatic polypep-tide; DBOHase, dopamine P-hydroxylase; NA, noradrenaline (norepi-nephrine); TyrOHase, tyrosine hydroxylase; VIP, vasoactive intestinalpolypeptide.

1303

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-nent" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

1304 Neurobiology: Lundberg et al.

added. Tissues processed by both methods were rinsed inphosphate-buffered sucrose for 24 hr and sectioned on a cryostat(section thickness, 10 or 15 gm). The sections were processedfor indirect immunofluorescence microscopy (20). Briefly, se-ries of four sections were incubated with diluted antiserum tosomatostatin (1:40), APP (1:100), VIP (1:100), or TyrOHase(1:200) and Df3OHase (1:80); rinsed in phosphate-buffered sa-line; incubated with fluorescein isothiocyanate-conjugatedswine anti-rabbit antibodies (1:10) (DAKO, Copenhagen, Den-mark); rinsed in phosphate-buffered saline; mounted in glycerol/phosphate-buffered saline, 3:1 (vol/vol); and examined in aZeiss fluorescence microscope. Antisera blocked with an excessof the respective peptide (5 gg of peptide per ml of diluted anti-serum) or normal rabbit serum (for TyrOHase and Df3OHaseantisera) served as controls. All sera contained 0.3% Triton X-100 (21). In some cases, sections were consecutively stainedwith two antisera by the elution/restaining techniqueof Tramu et al. (15).

RESULTS

Guinea Pig. In the coeliac-superior mesenteric ganglioncomplex and inferior mesenteric ganglion, both somatostatin-immunoreactive (Fig. 1 a and b) and APP-immunoreactive (Fig.1 c and d) ganglion cells were observed. Both had a markedregional distribution in the former complex (Fig. 1 a-d),whereas the distribution was more even in the inferior mes-enteric ganglion. In general, regions with a high density ofAPP-immunoreactive cells corresponded to areas with few somato-statin-positive cells and vice versa (Fig. 1 a-d). No immuno-reactivity like that of somatostatin or APP was seen in small in-tensely fluorescent cells. Analysis of consecutive sectionsshowed that the two somatostatin- and APP-like immunoreac-tive peptides were each localized in different cells (Fig. 1 a-d).This was confirmed by elution/restaining experiments dem-onstrating that somatostatin-positive cells appearing after elu-tion of the APP-like immunoreactive material were identical tothe previously APP-negative cells and vice versa. Incubationwith TyrOHase antiserum (Fig. 1 e and f) or D,8OHase anti-serum, or both, revealed that both the somatostatin- and APP-like peptides were present in noradrenaline cells. It is importantto note that a population ofTyrOHase- and Df3OHase-positivecells did not seem to contain somatostatin- or APP-like im-munoreactive material. In addition, after incubation withDBOHase antiserum, a dense network of strongly fluorescent,varicose fluorescent fibers was seen in both prevertebral gan-glia. No corresponding fiber networks could be observed afterincubation with somatostatin or APP antiserum; only single fi-bers of both types were encountered. Small numbers of VIP-immunoreactive cells with regional localization were seen in thecoeliac-superior mesenteric ganglion complex (Fig. lh). Theywere primarily present in areas with high numbers ofAPP-im-munoreactive cell bodies. In the same ganglion, VIP-like im-munoreactive material was also observed in very dense fibernetworks, again with a regional distribution (Fig. 1 g and h).High densities ofVIP-immunoreactive fibers were in areas richin somatostatin-positive cells (Fig. ig; cf. Fig. la). In the in-ferior mesenteric ganglion, single VIP-immunoreactive cellswere observed, in addition to a diffusely distributed, densenetwork of fluorescent fibers. In prevertebral ganglia, elution/restaining experiments revealed that VIP-immunoreactive cellsseemed to lack TyrOHase and DBOHase.-

In the superior cervical ganglion, only single somato-statin-positive cells were observed, whereas a large proportion

TyrOHase and D/3OHase positive. A few VIP-positive cell bod-ies were observed.

Cat. APP-immunoreactive cells were observed in all sym-pathetic ganglia studied, with a high proportion in the coeliacganglion; with many cells in the superior cervical (Fig. 2a) andseventh lumbar ganglion of the sympathetic chain and withsomewhat lower numbers in the superior and inferior mesen-teric ganglia. No APP-immunoreactive varicose fibers were ob-served. VIP-immunoreactive cell bodies were present in lownumbers in all ganglia studied with the exception of the seventhlumbar ganglion. They were often seen to form small clusters,but more evenly distributed cells were also seen (Fig. 2e). Var-icose VIP-immunoreactive fibers were generally less abundantthan in the guinea pig. However, regionally high concentrationsof fibers could be seen in the superior mesenteric ganglion,primarily in areas lacking APP-immunoreactive cell bodies. TheAPP- but not the VIP-immunoreactive ganglion cells wereTyrOHase (Fig. 2b) and D,3OHase positive. Somatostatin-likeimmunoreactive material was never seen in ganglion cells butwas found in small intensely fluorescent cells and fibers in pre-vertebral ganglia. After local application of vinblastine to thesuperior cervical and seventh lumbar sympathetic ganglion,approximately the same number of APP- and VIP-immuno-reactive cells were seen as in untreated cats. However, therewas a marked increase in fluorescence intensity after vinblas-tine treatment, and numerous diluted immunoreactive axonsappeared.

In the submandibular salivary gland, APP-immunoreactivenerves were seen around small arteries and arterioles (Fig. 2c).Additional TyrOHase- and D(3OHase-positive fibers werefound around acini and ducts (Fig. 2d).

In the nasal mucosa, APP-immunoreactive fibers were onlyseen around small arteries and arterioles (Fig. 2f), the latteroften with a diameter of about 40-100 ,um, whereas TyrOHase-and D/30Hase-positive fibers in addition were found in highnumbers around venules and venous sinusoids (Fig. 2g). SingleTyrOHase- and D/BOHase-immunoreactive fibers were seenalso around acini and ducts. VIP-immunoreactive nerve fiberswere abundant both on the arterial and venous side and aroundthe exocrine elements. It was noted that VIP- and APP-im-munoreactive nerves surrounded the same blood vessels on thearterial side.

In the uterus and the urinary bladder, APP-immunoreactivenerves were mainly seen around arterioles, whereas TyrOHaseand Dj3HOase in addition were seen in fibers within the smoothmuscle layers. VIP-immunoreactive fibers were found botharound blood vessels and within the myometrium and in thetrigonum area of the bladder.

Controls, None of the immunoreactivities described abovewere observed after incubation with control sera.

DISCUSSION

Earlier studies have shown that classical transmitters such asNA and AcCho can coexist with somatostatin- (11), APP- (12),and VIP-like (13, 14) immunoreactive peptides in sympatheticneurons. The present results extend these studies by demon-strating that these three peptides occur in different neuronsand, therefore, that adult peripheral sympathetic ganglion cellscan be subdivided into at least four populations, containing (i)a somatostatin-like peptide and NA (called NA/somatostatinneurons), (ii) an APP-like peptide and NA (called NA/APP neu-rons), (iii) NA only (called NA neurons), or (iv) a VIP-like pep-tide and possibly AcCho (called AcCho/VIP neurons). The pres-ence of detectable amounts of one of these peptides seems toexclude the presence of another, lending credit to our hypoth-

Proc. Natl. Acad. Sci. USA 79 (1982)

ofthe cells were APP positive. Many cells were neitl.er s-omato-statin nor APP positive, but virtually all ganglion cells were

Proc. Natl. Acad. Sci. USA 79 (1982) 1305

I

FIG. 1. Immunofluorescence micrographs of the coeliac-superior mesenteric ganglion complex of guinea pig after incubation with antiserumto somatostatin (a and b), APP (c and d), TyrOHase (e and f), and VIP (g and h). All micrographs are from the same ganglion complex but fromtwo different parts: a, c, e, and g are consecutive sections from one part, and b, d, f, and h are consecutive sections from the other part. Note thatone region contains many somatostatin-immunoreactive (a) and few APP-immunoreactive (c) cell bodies and a very dense network of VIP-im-munoreactive nerve endings (g). A reversed picture is seen in the other region, where few somatostatin-immunoreactive (b) and many APP-im-munoreactive (d) cells and only a patchy network of VIP-immunoreactive fibers are found. Some VIP-immunoreactive cell bodies are seen in thelatter region (arrows in h). Virtually all cells in both regions are TyrOHase positive. All micrographs have the same magnification. (Bar = 50 Anm).

esis that, under normal conditions, neurons containing a certainclassical transmitter, such as NA, may be subdivided on thebasis oftheir content ofa specific peptide (see ref. 22). It should

be emphasized, however, that immunohistochemistry may notbe sensitive enough to demonstrate very low peptide levels.Thus, it cannot be excluded that more than one ofthese peptides

Neurobiology: Lundberg et aL

1306 Neurobiology: Lundberg etalP

FIG. 2. Immunofluorescence micrographs of consecutive sections of the superior cervical ganglion (a and b), semiconsecutive sections of thesubmandibular salivary gland (c and d), the seventh sympathetic lumbar ganglion (e), and semiconsecutive sections of the maxillo-turbinal areaof the nasal mucosa (f and g) of the cat after incubation with antiserum to APP (a, c, and f), TyrOHase (b and d), VIP (e), and D(3OHase (g). (aand b) Virtually all cells in the superior cervical ganglion are TyrOHase positive (b), but only some of them are APP immunoreactive (a). Arrowheads in a point to APP-negative cells, and arrow heads in b indicate some possible TyrOHase-negative cells. (c and d) In the salivary gland, APP-immunoreactive fibers (arrow heads) are only found around blood vessels, whereas TyrOHase is present in addition around the acini (arrows) andducts. a, Small artery; d, ducts. (e) In the lumbar ganglion a population of ganglion cells are VIP positive, localized either in small groups (big arrows)or scattered within the ganglion (small arrows). (f and g) In nasal mucosa only the arterioles (a) and small arteries are surrounded by APP-im-munoreactive fibers, whereas the Dp8OHase-positive fibers are seen also close to venules and venous sinusoids (v). c, Cartilage; e, surface epithelium.All micrographs have the same magnification. (Bar = 50 gm.)

is present in the nonadrenaline cells. In fact, it has been shownthat 5-HT neurons in the central nervous system may not onlycontain substance P (23, 24) but also a thyrotropin-releasinghormone (TRH)-like peptide (25).

Furthermore, the present findings in the cat indicate that thedifferent types of sympathetic neurons described innervate dif-

ferent tissue components. For example, in the cat nasal mucosa,the NA/APP neurons preferentially innervate arteries and ar-terioles, whereas the NA neurons innervate veins. Therefore,it is tempting to seek functional correlates. In functional studies,in fact, a differential nervous control of arteries and veins hasbeen indicated, because lower stimulation frequencies prefer-

Proc. Natl. Acad. Sci. USA 79 (1982)

I

Proc. NatL Acad. Sci. USA 79 (1982) 1307

entially activate veins. Furthermore, the maximal capacitanceresponse is reached at lower frequencies than the maximal re-sistance response (26, 27). In the cat submandibular salivarygland, the blood vessels receive NA/APP nerves, whereas theexocrine acini and ducts are innervated by NA nerves. Emmelinand Engstrom (28) have provided evidence that the noradre-nergic vasoconstrictor nerves can be activated separately fromthe noradrenergic nerves stimulating salivary secretion. A sim-ilar functional specialization may be present in the uterus andurinary bladder with regard to control ofblood flow and smoothmuscle activity. Although NA/APP nerves are associated withthe vascular system, this association is not absolute because theymay innervate other types of target organs. Thus, the ganglioncoeliacum, which contains high numbers of NA/APP neurons,innervates not only blood vessels but also the ganglionic entericplexuses. Furthermore, NA/APP nerves also seem to innervatesmooth muscle in the vas deferens (12).The AcCho/VIP neurons constitute a small part of the sym-

pathetic system in the cat. Functionally, these neurons havebeen associated with regulation of sweat secretion (13, 14). TheVIP-immunoreactive nerves, in contrast to the APP-immuno-reactive ones, are present both around blood vessels and aciniof sweat glands and in parasympathetic neurons innervatingother exocrine glands-e.g., nasal mucosa and submandibularsalivary gland (14, 29).The second species analyzed in this study, the guinea pig,

contains somatostatin in some NA neurons in contrast to the cat.The terminal areas of the NA/somatostatin neurons have-so farnot been identified. In the coeliac-superior mesenteric gan-glion complex of the guinea pig, the NA/somatostatin (11) andthe NA/APP neurons have a clear regional distribution. Thismay be related to the subdivision ofthe complex into the coeliacand superior mesenteric ganglia because, in the cat, almost allcells in the coeliac ganglion contain APP-like immunoreactivematerial.

Also, the dense networks of VIP-immunoreactive fibers inthe coeliac-superior mesenteric ganglion complex in the guineapig show a marked regional distribution with high concentra-tions around NA/somatostatin and NA cells and few around NA/APP cells, which mostly lack such fibers. Thus, only part oftheganglion complex is controlled by VIP fibers. Also, in the catthe VIP nerves preferentially surround NA cells lackingAPP-i.e., cells in the superior mesenteric ganglion. Becausethere is evidence that VIP fibers in prevertebral ganglia orig-inate in the gastrointestinal wall (13), it may be suggested thatthese VIP nerves, together with the NA/somatostatin (in theguinea pig) and the NA (in the cat) neurons, are part ofthe reflexarcs between the mid- and distal gut and prevertebral ganglia(30, 31).To understand the functional significance of a subdivision of

sympathetic neurons by peptides and especially the exact phys-iological role of these coexisting peptides, more extensive stud-ies will be required. With regard to the AcCho/VIP nerves inexocrine glands, the two mediators seem to have a comple-mentary action in causing secretion and vasodilation (14, 29).Here the occurrence of the peptide VIP has explained the well-known atropine-resistant part of the vasodilation (32).

In conclusion, peripheral sympathetic neurons in adult catsand guinea pigs can be subdivided into several subcategorieson the basis of the presence or absence of specific peptides. Aneuron population with a particular peptide may innervate spe-cific target tissues, for instance blood vessels or secretory ele-ments. Also, the neuronal input to a particular ganglion may becharacterized by a specific peptide and may have a topograph-ical relation to the presence of a particular peptide.

We thank Ms. Annika Edin, Ms. Waldtraut Hiort, and Ms. AnnePeters for skillful technical assistance and Ms. Birgit Frideen and Ms.Anita Lundmark for expert secretarial help. The present study was sup-ported by grants from the Swedish Medical Research Council (04X-2887, 17X-5438, and 04X-3766), Knut och Alice Wallenbergs Stiftelse,Magnus Bergvalls Stiftelse, Sven och Ebba-Christina Hagbergs Stif-telse, Ollie och ElofEricssons Stiftelse, the Swedish Tobacco Company,the National Institute of Mental Health (02717), and the National In-stitute of Neurological and Communicative Disorders and Stroke(06801).

1. Gaskell, W. H. (1916) The Involuntary Nervous System (Long-mans Green, London).

2. Kuntz, A. (1953) The Autonomic Nervous System (Lea and Fe-biger, Philadelphia), 4th Ed.

3. Langley, J. N. (1921) The Autonomic Nervous System (W. Hoef-ter and Sons, Cambridge), Part 1.

4. von Euler, U. S. (1946) Acta Physiol Scand. 12, 73-97.5. Langley, J. N. (1891)J. Physiol (London) 12, 347-374.6. Dale, H. H. & Feldberg, W. (1934) J. Physiol (London) 82,

121-128.7. Bulbring, E. & Burn, J. H. (1925) J. Physiot (London) 83,

121-128.8. H6kfielt, T., Johansson, O., Ljungdahl, A., Lundberg, J. M. &

Schultzberg, M. (1980) Nature (London) 284, 515-521.9. Lundberg, J. M., Hokfelt, T., Anggird, A., Uvnis-Wallensten,

K., Brimijoin, S., Brodin, E. & Fahrenkrug, J. (1980) in NeuralPeptides and Neuronal Communication, ed. Costa, E. & Trabuc-chi, M. (Raven, New York), pp. 25-36.

10. Schultzberg, M. (1980) Dissertation (Karolinska Institutet, Stock-holm, Sweden).

11. H6kfelt, T., Elfvin, L. G., Elde, R., Schultzberg, M., Goldstein,M. & Luft, R. (1977) Proc. Natl Acad. Sci. USA 74, 3587-3591.

12. Lundberg, J. M., Hokfelt, T., AnggArd, A., Kimmel, J., Gold-stein, M. & Markey, K. (1980) Acta. Physiol Scand. 110, 107-109.

13. Lundberg, J. M., Hokfelt, T., Schultzberg, M., Uvnas-Wallen-sten, K., Kohler, C. & Said, S. J. (1979) Neuroscience 4,1539-1559.

14. Lundberg, J. M. (1981) Acta Physiol. Scand. Suppl. 496, 112,1-57.

15. Tramu, G., Pillez, A. & Leonardelli, J. (1978)J. Histochem. Cy-tochem. 26, 322-324.

16. Arimura, A., Lundqvist, G., Rothman, J., Changy, R., Fernan-dez-Durango, R., Elde, R., Coy, D. H., Meyers, C. & Schally,A. V. (1978) Metabolism Suppl. 1 27, 1139-1144.

17. Markey, K. A., Kondo, S., Shenkman, L. & Goldstein, M. (1980)Mol Pharmacol 17, 79-5.

18. Goldstein, M., Anagnoste, B., Freedman, L. S., Roffman, M.,Ebstein, R. P., Park, D. H., Fuxe, K. & Hokfelt, T. (1974) inFrontiers in Catecholamine Research, eds. Usdin, E. & Snyder,S. (Pergamon, New York), pp. 69-78.

19. Dahlstrom, A. (1971) Acta Neuropathol, Suppl 5, 226-237.20. Coons, A. H. (1958) in General Cytochemical Methods, ed.

Danielli, J. F. (Academic, New York), pp. 399-422.21. Hartman, B. K., Zide, D. & Udenfriend, S. (1972) Proc. Nati

Acad. Sci. USA 69, 2722-2726.22. Hokfelt, T., Lundberg, J. M., Schultzberg, M., Johansson, O.,

Ljungdahl, A. & Rehfeld, J. (1980) in Neural Peptides and Neu-ronal Communication, eds. Costa, E. & Trabucchi, M. (Raven,New York), pp. 1-23.

23. Chan-Palay, V., Jonsson, G. & Palay, S. L. (1978) Proc. NatlAcad. Sci. USA 75, 1582-1586.

24. Hokfelt, T., Ljungdahl, A., Steinbusch, H., Verhofstad, A., Nils-son, G., Brodin, E., Pernow, B. & Goldstein, M. (1978) Neuro-science 3, 517-538.

25. Johansson, O., H6kfelt, T., Pernow, B., Jeffcoate, S. L., White,N., Steinbusch, H., Verhofstad, A,, Emson, P. C. & Spindel, E.(1981) Neuroscience, in press.

26. Mellander, S. (1960) Acta Physiol Scand. Suppl 176 50, 1-86.27. AnggArd, A. & Edwall, L. (1974) Acta Oto-Laryngol 77, 131-139.28. Emmelin, N. & Engstrom, J. (1960)1. Physiol (London) 153, 1-8.29. Lundberg, J. M., AnggArd, A., Fahrenkrug, J., Hkfielt, T. &

Mutt, V. (1980) Proc. Natl. Acad. Sci. USA 77, 1651-1655.30. Kuntz, A. (1938) J. Comp. Neurot 69, 1-12.31. Szurzewski, J. H. (1981) Annu. Rev. Physiol 43, 53-68.32. Heidenhain, R. (1872) Pfluegers Arch. 5, 309-318.

Neurobiology: Lundberg et aL