FMRFamide-like immunoreactive neurons of the nervus terminalis of teleosts innervate both retina and pineal organ

Brain Research. 460(198b~) 6b;-75 Elsevier

68 BRE 13913

FMRFamide-like immunoreactive neurons of the nervus terminalis of teleosts innervate both retina and pineal organ Peter Ekstr6m 1, Tapio H o n k a n e n I and Sven O.E. Ebbesson 2 1Department of Zoology, University of Lund, Lund (Sweden) and 21nstitutefor Marine Science, University of Alaska-Fairbanks, Fairbanks, Alaska 99701 (U.S.A.) (Accepted 5 April 1988)

Key words: FMRFamide; Pineal organ; Retina; Nucleus nervus terminalis; Nucleus olfactoretinalis; Immunocytochemistry; Teleost fish

The tetrapeptide FMRFamide (Phe-Met-Arg-Phe-NH2) was first isolated from molluscan ganglia. Subsequently, it has become clear that vertebrate brains also contain endogenous FMRFamide-like substances. In teleosts, the neurons of the nervus terminalis contain an FMRFamide-like substance, and provide a direct innervation to the retina (Proc. Natl. Acad. Sci. U.S.A., 81 [1984] 940-944). Here we report the presence of FMRFamide-immunoreactive axonal bundles in the pineal organ of Coho salmon and three-spined sticklebacks. The largest numbers of axons were observed proximal to the brain, in the pineal stalk, while the distal part of the pineal organ contained only few axons. No FMRFamide-like-immunoreactive(IR) cell bodies were observed in the pineal organ. In adult fish it was not possible to determine the origin of these axons, due to the large numbers of FMRFamide-like IR axons in the teleost brain. However, by following the development of FMRFamide-like IR neurons in the embryonic and larval stickleback brain, it was possible to conclude that, at least in newly hatched fish, FMRFamide-like IR axons that originate in the nucleus nervus terminalis reach the pineal organ. Thus, it seems there is a direct connection between a specialized part of the chemosensory system and both the retina and the pineal organ in teleost fish.

INTRODUCTION The t e t r a p e p t i d e F M R F a m i d e (Phe-Met-Arg-PheNH2) was first discovered as a cardioexcitatory peptide in molluscs 22. Subsequently F M R F a m i d e , or at least related R F a m i d e s have been d e m o n s t r a t e d in most v e r t e b r a t e classes 1'3'25'2s. In teleost fish, one of the neuronal systems known to contain F M R F a m i d e , is the so-called nervus terminalis system 24. This system consists of cell bodies, located in the olfactory bulbs or in the anterior t e l e n c e p h a l o n , possessing axons that innervate, a m o n g o t h e r areas, the retina. It has previously been shown that these neurons contain luteinizing h o r m o n e - r e l e a s i n g h o r m o n e ( L H RH)-like i m m u n o r e a c t i v e material 18. Stell and coworkers 24 have d e m o n s t r a t e d that L H - R H - I i k e a n d ' F M R F a m i d e - l i k e immunoreactivities are co-localized in all neuronal s o m a t a and axons of the nervus terminalis. Thus the axons of the nervus terminalis

that innervate the retina contain both L H - R H - l i k e ' a n d F M R F a m i d e - l i k e material. Synthetic L H - R H , F M R F a m i d e and o t h e r R F a m i d e s influence the activity of retinal ganglion cells 27. Such a direct input to the retina by afferent fibers that p r o b a b l y mediate chemosensory signals would allow a unique interaction of different sensory modalities already at lower integrative levels. On the contrary, the directly photosensitive pineal organ of teleost fishes has been considered an organ with exclusively efferent neural connections 6'9'1°'14. H o w e v e r , terminal-like structures of unknown origin and function have been o b s e r v e d in the pineal organ 8"21. These might represent axons of central origin, as have been shown in several amniotes 15. In the present study, we d e m o n s t r a t e F M R F a m i d e - l i k e immunoreactive axons of central origin that innervate the pineal organ in several teleost species.

Correspondence: P. EkstrOm, Department of Zoology, University of Lund, Lund, Sweden. 0006-8993/88/$03.50 (~ 1988 Elsevier Science Publishers B.V. (Biomedical Division)

69 MATERIALS AND METHODS Adult sticklebacks (Gasterosteus aculeatus) were caught off the Baltic coast, and brought to breed to produce stickleback embryos. Rearing conditions were such that the embryos hatch during their seventh day of life, i.e. at the age of 144-168 h 11. Atlantic salmon (Salmo salar) were purchased from a commercial hatchery in Saltviken, Sweden, while the coho (Silver) salmon (Oncorhynchus kisutch) were generously supplied by the Crooked Creek Hatchery, Alaska Dept. of Fish & Game, Soldotna, Alaska. Several fixatives were used, of which the following gave the best results. (1) 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2, (2) with the addition of 0.25% picric acid and (3) with the further addition of 0.25% glutaraldehyde. Adult fish were fixed by vascular perfusion under deep MS-222 (tricaine methanesulphonate) anesthesia; embryos were fixed by immersion. Fixation times varied between 4 and 24 h. After thorough phosphate buffer rinses, specimens were stored in Tyrode's buffer containing 25% sucrose and 0.01% sodium azide until sectioning. Brains and whole embryos were sectioned in a Reichert-Jung cryostat, collected on slides subbed with chrome alum-gelatine, and allowed to dry at room temperature for at least 1 h. The slides were then thoroughly rinsed in wash buffer (see below) before immunocytochemical processing: (1) incubation overnight in rabbit anti-FMRFamide (Immunonuclear: code 72432, Lot 8630014) diluted 1:10001:1600, followed by 3 × 5-min rinses in wash buffer; (2) incubation for 30 min in swine anti-rabbit IgG (Dakopatts, Copenhagen) diluted 1:50, followed by 3 x 5-min rinses in wash buffer; (3) incubation for 30 min in rabbit PAP-complex (Dakopatts, Copenhagen) diluted 1:50, followed by a 5-min rinse in wash buffer and 2 x 5-min rinses in 0.1 M Tris-HCl buffer (pH 7.2). All incubation steps were performed at ca. 20 °C. Wash buffer was phosphate-buffered saline (PBS) containing 0.25% Triton X-100. All antibodies were diluted in wash buffer containing 1% bovine serum albumin. Immunoreactive sites were visualized by reacting the sections for peroxidase activity with 0.05% 3,3'diaminobenzidine tetrahydrochloride (DAB) and

0.015% hydrogen peroxide in 0.1 M Tris-HCl (pH 7.2). In some cases, 0.3% ammonium nickel sulphate was added to increase sensitivity, in these cases DAB and hydrogen peroxide concentrations were lowered to 0.02% and 0,005%, respectively. Controls were performed to assess the method specificity26, i.e. incubations were performed with omission of the primary antibody, or with diluted antiserum that had been subjected to liquid-phase preabsorption with synthetic FMRFamide (1 pM; Sigma, St. Louis, MA; U.S.A.). RESULTS In all adult teleosts examined, i.e. three-spined stickleback, Atlantic salmon and coho salmon, FMRFamide-like immunoreactive (FMRFamideIR) axons were observed in the pineal organ. Here, the axons followed the bundles of the pineal tract (Fig. 1A), which mainly consists of pinealofugally projecting axons 6'9,1°. In the coho salmon, it was possible to determine that the axons that entered the pineal stalk from the brain did so via two routes, one rostral and one caudal. The rostral route encompasses axons that run along the anterior border of the subcommissural organ and enter the ventrolaterai bundles of the pineal tract (Fig. 1B; cf. Fig. 2 in ref. 20 depicting the organization of the pineal tract in another salmonid, the rainbow trout). The caudal route encompasses fibers that follow the main dorsomedial bundle of the pineal tract (Fig. 1C,D). It was not possible, at the light microscopical level, to determine the location of the terminals in the pineal parenchyma. However, the largest numbers of immunoreactive axons were invariably observed in the pineal stalk, while only scattered varicose axons were observed in the end-vesicle. No FMRFamide-IR neuronal somata were observed in the pineal organ of any species. Due to the large numbers of FMRFamide-IR axons in the adult teleost brain (data not shown), it was not possible to trace the pinealopetal FRMFamideIR axons to their origin(s). This difficulty was overcome by following the ontogenetic development of FMRFamide-IR neurons in the central nervous system of embryonic and early larval sticklebacks. In this species, FMRFamide-IR axons were detected in

70

Fig. 1. F M R F a m i d e - I R axons in the pineal organ of coho salmon smolt. A: F M R F a m i d e - I R axons (arrowheads) follow the dorsomedial main bundle of the pineal tract. B: an F M R F a m i d e - I R axon (arrowhead) enters one of the ventrolateral pineal tract axon bundles in the pineal stalk. Varicose F M R F - I R axons (arrows) are also present in the parapineal organ (asterisk). C: F M R F a m i d e - I R axons in the dorsomedial main bundle of the pineal tract (arrowheads; cf. A). Note the branching axons (arrows) that enter the parapineal organ (asterisk). D: more caudally, F M R F - a m i d e - I R axons follow the pineal tract (arrowhead) dorsal to the subcommissural organ. Frontal sections, cf. Fig. 5, bottom. Bar = 50/~m.

71

Fig. 2. FMRFamide-IR in the three-spirted stickleback at the time of hatching. A,B: 168 h, serial sagittal sections showing the densely packed FMRFamide-IR neuronal somata of nucleus nervus terminalis (open arrows) with neurites (arrowheads) extending into the olfactory epithelium. Note the presence of one immunoreactive neuronal soma (curved arrow) in the olfactory epithelium. C: 144 h, horizontal section, cf. Fig. 5, top. At this stage, FMRFamide-IR axons innervate predominantly the inner plexiform layer of the temporal retina (solid arrow). Nucleus nervus terminalis (open arrow). D,E: 168 h, sagittal sections showing FMRFamide-IR axons rostral and caudal to (curved open arrows) the pineal organ (P), and entering the optic nerve (solid arrow); D shows a higher magnification of the pineal area. Bars = 50~m.

72 the pineal organ and retina already around the time of hatching (144-168 h after fertilization; cf. ref. 11). At the same stage, FMRFamide-IR neuronal somata were observed in two locations in the brain: along the spinal cord (data not shown) and in the ventral telencephalon (Fig. 2A). The neurons in the telencephalon/olfactory bulb have (probably dendritic) processes extending into the olfactory epithelium (Fig. 2B), and their axon-like processes could be followed along the ventral telencephalon, caudal to the optic chiasm. Here, some axons turn ventrally and enter the optic nerves (Fig. 2C), which they follow into the inner plexiform layer of the retina (Fig. 2D). From the same point behind the optic chiasm, some axons turn dorsally and course to the pineal organ (Figs. 2C, 3A,B). Also in the adult three-spirted stickleback, numerous FMRFamide-IR axons could be observed throughout the slender pineal stalk (Fig. 3C).

The neuronal group in the ventral telencephalon seems identical to the nucleus nervus terminalis (ganglion terminalis) described in different teleost species 1s'24. At least the large majority of the FMRFamide-IR axons that reach the teleost pineal organ originate here. However, ascending axons from the spinal cord also reach the optic chiasm, and some of these axons turn dorsally to innervate, among other areas, the optic tectum. The possibility that some ascending spinal axons contribute to the innervation of the pineal organ in the larval stickleback cannot be totally excluded. The distribution of FMRFamide-IR neurons in the early larval forebrain, and the innervation of the pineal organ, are summarized in Fig. 4. For clarity, Fig. 5 depicts the planes of section for embryonic brains that were not sagittally sectioned, and the levels of section through the pineal organ of the adult salmon.

Fig. 3. FMRFamide-IR axons (arrowheads) in the pineal organ (P) of the three-spined stickleback at different ages. A,B: horizontal sections (cf. Fig. 5, top) through the pineal organ of a 144 h (A) and 168 h (B) old stickleback larvae. Habenular nuclei (H). C: adult, pineal stalk. Note that the pineal stalk of the adult stickleback is long and slender, and actuallyof smaller diameter than the larval pine-

al organ. Bar = 50/zm.

73

A

E

D

Fig. 5. Semi-schematic drawings of midsagittal sections through the brain of a hatching (ca. 144-168 h old) stickleback larva (top), and through the brain of an adult salmon (bottom; adapted from ref. 16), to show the planes of sections in Fig. 1A-D (bottom) and Figs. 2C, 3A and 3B (top), respectively. Bars = 250am (top), 5 mm (bottom). Fig. 4. At the age of 144 h after fertilization (approximately at the age when hatching starts), the FMRFamide-IR neurons of the nucleus nervus terminalis (NT) give rise to neurites that spread in the olfactory epithelium (O), the telencephalon, (T), to the retina via the optic nerve (ON), to the pineal organ (P) and the optic tectum (OT). At this age, ascending axons from neurons in the spinal cord (open arrow) reach the forebrain, and intermingle with the descending axons. The upper part shows the distribution of nervus terminalis axons reaching the pineal organ and the optic nerve, while the lower part shows the total distribution of FMRFamide-immunoreactivity in the forebrain. Semischematic drawing produced by superimposing three camera lucida drawings of consecutive serial sections. Bar = 100~tm.

DISCUSSION In the present study, we have been able to demonstrate the presence of a central innervation of the photosensory pineal organ of 3 teleost species. The central innervation consists of F M R F a m i d e - I R axons that, at least at early larval developmental stages, seemed to originate from a neuronal group located in the ventral telencephalon. These findings encompass several points deserving elaboration: (1) the specificity of the immunoreaction and the identity of the immunoreactive compound, (2) the identities

of the groups of neurons contributing to the central innervation of the photosensory pineal organ, and (3) functional considerations. Concerning the specificity, our tests indicated that the criteria for method specificity z6 were met. After incubation with liquid phase-absorbed primary antiserum, or without primary antiserum, no immunoreaction was observed in the brain, retina or pineal organ. This means that our F M R F a m i d e antibody specifically reacts with (among other systems) the neuronal system(s) innervating the pineal organ and the retina, but it does not prove that the neuronal systems in question contain FMRFamide. Actually, the innervation of the goldfish retina by axons of the nucleus nervus terminalis has been subject to detailed investigations, and there are indications that this neuronal system contains more than one FMRFamide-like peptide in this species 19'23'24'27. Since it is known that antisera to F M R F a m i d e also recognize pancreatic polypeptide-like peptides as well as peptides with an amidated C-terminal sequence ArgPhe-NH 2 (refs. 5, 13, 19, 25), we can only state, without further detailed biochemical analyses, that at

74 least part of the central innervation of the pineal organ in the three-spined stickleback, the Atlantic salmon, and the coho salmon consists of axons containing an FMRFamide-like peptide. As pointed out in the Results section, we cannot conclude that all axons innervating the larval pineal organ have their origin in the n. nervus terminalis. It might well be that some axons stem from neurons in the spinal cord. There is also the possibility that the larval innervation is transient, and is later replaced by axons from other neuronal groups. However, this latter alternative does not seem likely at present. In adult teleosts the largest numbers of FMRFamide-IR somata were CSF-(cerebrospinal fluid-) contacting neurons in the periventricular hypothalamus (data not shown). These neurons were found in the area to which presumably pinealofugally projecting axons were observed after labeling of the pineal tract with horseradish peroxidase (HRP) 9"1°. However, retrogradely HRP-filled neurons were never observed in this area, although massively HRP-labeled axons could be followed here. The absence of retrogradely filled neurons in the n. nervus terminalis after application of H R P to the cut pineal tract may be explained by the relatively long distance between the frontal telencephalon and the pineal organ and the small diameter of the axons. To obtain a complete labeling of the central neural connections of the pineal organ, H R P must be applied to the pineal tract as close to the brain as possible, i.e. in the proximal part of the pineal stalk. In this region the pineal stalk is very thin (ca. 50-100 ~m). Thus, it is not possible to inject a pool of tracer + detergent (type DMSO, lysolecithin, etc.) for prolonged uptake. Rather, it is necessary to make use of the immediate uptake of tracer into the acutely lesioned axons of the pineal tract 7'17. Of several tracers tried, H R P proved the most sensitive, especially when visualized with a modified detection method 7. It should be noted, however, that the axons observed in the rainbow trout 14, three-spined stickleback 9 and the eel 1°, that were interpreted as projecting to the preoptic area or rostrai hypothalamus, might well represent incompletely retrogradely filled axons, originating in the n. nervus terminalis. There are as yet no physiological data on the effect of FMRFamide on pineal cells. In the goldfish retina, that receives a central innervation of FMRFamide-

1R axons from the n. nervus terminalis 24, a large number of ganglion cells are excited by iontophoretic application of FMRFamide. However, numerous ganglion cells are not affected, while some are inhibited, and some show biphasic responses 27. The strength of this effect varied with the time of year; especially the strongest excitatory effect was noted during winter and spring, i.e. the natural reproductive period of this species 27. Making the hypothesis that the n. olfactoretinalis system may directly mediate chemical stimulation, e.g. by sexual pheromones, to the retina, this input may modulate responses to visual stimuli with variable strength, depending on the time of year. It should be noted that it has so far not been unequivocally demonstrated that the terminal nerve of teleosts transmits chemosensory information from the olfactory epithelium to the brain and retina, although several lines of evidence strongly suggest this 2. It is possible that the terminal nerve functions as a central integrator: it innervates peripheral (olfactory epithelium) as well as central (telencephalon, retina, pineal organ) targets and receives a central catecholaminergic innervation 2'12'19. Granted that the terminal nerve transmits chemosensory information to the retina, direct transduction of chemical stimuli from a specialized portion of the olfactory system to the pineal organ might serve analogous, season-dependent, functions. FMRFamide could directly influence the activity of the pinealofugaily projecting neurons, i.e. the intrapineal neurons analogous to the retinal ganglion cells. By stimulating or inhibiting their spontaneous activity, relative darkness or relative brightness, respectively, would be mimicked 4. However, the exact site of termination of the intrapineal FMRFamide-IR axons remains to be determined, as well as the direct action of FMRFamide and related peptides on different types of pineal cells, before it is possible to speculate on the function of FMRFamide in the pineal organ. It is also of importance to determine whether single neurons of the terminal nerve innervate the pineal organ and retina by means of collaterals, or there are neurons innervating either the pineal organ or the retina. We are presently attempting to solve the problems in achieving double retrograde tracing (see above) combined with immunocytochemical identification to tackle this question.

75 ACKNOWLEDGEMENTS

ty of Stockholm, Sweden) for providing embryonic stickleback material, and Dr. D . R . N~issel for critical

We thank Lorin Flagg, Bob Och and Stephen Bates of Crooked Creek Hatchery, and Alaska Department of Fish and G a m e , Soldotna, Alaska, for logistical support during part of the experiment. We also thank Dr. B. Borg (Dept. of Zoology, Universi-

reading of the manuscript. The technical assistance

REFERENCES

47-69. 16 Kotrschal, K., Krautgartner, W.-D. and Adam, H., Comparative scanning electron microscopy of the ventricular surface in four Actinopterygian fishes: Acipenser ruthenus, Salmo gairdneri, Tinca tinca and Blennius sanquinolentus (Pisces), J. Hirnforsch., 26 (1985) 41-63. 17 Kristensson, K. and Olsson, Y., Retrograde transport of horseradish peroxidase in transected axons. 3. Entry into injured axons and subsequent localization in perikaryon, Brain Research, 115 (1976) 201-213. 18 Miinz, H., Claas, B., Stumpf, W.E. and Jennes, L., Centrifugal innervationof the retina by luteinizing hormone-releasing hormone (LH-RH)-immunoreactive telencephalic neurons in teleost fishes, Cell Tissue Res., 222 (1982) 313-323. 19 Muske, L.E., Dockray, G.J., Chohan, K.S. and Stell, W.K., Segregation of FMRF amide-immunoreactive efferent fibers from NPY-immunoreactive amacrine cells in goldfish retina, Cell Tissue Res., 247 (1987) 299-307. 20 Omura, Y., Light and electron microscopic studies on the pineal tract of rainbow trout, Salmo gairdneri, Rev. Can. Biol., 38 (1979) 105-118. 21 Omura, Y. and Ali, M.A., Responses of pineal photoreceptors in the brook and rainbow trout, Cell Tissue Res., 208 (1980) 111-122. 22 Price, D.A. and Greenberg, M.J., Structure of a molluscan cardioexcitatory peptide, Science, 197 (1977) 670-671. 23 Stell, W.K., Ball, A.K., Chohan, K.S., Djamgoz, M.B.A., Downing, J.E.G., Kyle, A.L., Muske, L.E. and Walker, S.E., Co-localization of neuroactive substances, and its functional significance, in the cyprinid fish retina. In E. Agardh and B. Ehinger (Eds.), Retinal Signal Systems, Degenerations and Transplants, Elsevier, Amsterdam, 1986, pp. 73-87. 24 Stell, W.K., Walker, S.E., Chohan, K.S. and Ball, A.K., The goldfish nervus terminalis: a luteinizing hormone-releasing hormone and molluscan cardioexcitatory peptideimmunoreactive olfactoretinal pathway, Proc. Natl. Acad. Sci. U.S.A., 81 (1984) 940-944. 25 Triepel, J. and Grimmelikhuijzen, C.J.P., A critical examination of the occurrence of FMRFamide immunoreactivity in the brain of guinea pig and rat, Histochemistry, 80 (1984) 63-71. 26 Van Leeuwen, F., Pitfalls in immunocytochemistry with special reference to the specificity problems in the localization of neuropeptides, Am. J. Anat., 175 (1986) 363-377. 27 Walker, S.E. and Stell, W.K., Gonadotropin-releasing hormone (GnRF), molluscan cardioexcitatory peptide (FMRFamide), enkephalin and related neuropeptides affect goldfish retinal ganglion cell activity, Brain Research, 384 (1986) 262-273. 28 Williams, R.G. and Dockray, G.J,, Immunohistochemical studies of FMRF-amide-like immunoreactivity in rat brain, Brain Research, 276 (1983) 213-229. 29 Wirsig, C.R. and Getchell, T.V., Amphibian terminal nerve: distribution revealed by LH-RH and AChE markers, Brain Research, 385 (1986) 10-21.

1 Boer, H.H., Schot, L.P.C., Veenstra, J.A. and Reichelt, D., Immunocytochemical identification of neural elements in the central nervous system of a snail, some insects, a fish, and a mammal with an antiserum to the molluscan cardioexcitatory tetrapeptide FMRFamide, Cell Tissue Res., 213 (1980) 21-27. 2 Demski, L.S. and Northcutt, R.G., The terminal nerve: a new chemosensory system in vertebrates?, Science, 220 (1983) 435-437. 3 Dockray, G.J., Vaillant, C. and Williams, R.G., New vertebrate brain-gut peptide related to a molluscan neuropeptide and an opioid peptide, Nature (Lond.), 293 (1981) 656-657. 4 Dodt, E., The parietal eye (pineal and parietal organs) of lower vertebrates. In R. Jung (Ed.), Handbook of Sensory Physiology, Vol. VII/3B: Central Processing of Visual Information, Part B, Springer, Berlin, 1973, pp. 113-140. 5 Ebberink, R.H.M,, Price, D.A., Van Loenhout, H., Doble, K.E., Riehm, J.P., Geraerts, W.P.M. and Greenberg, M.J., The brain of Lymnaea contains a family of FMRFamide-like peptides, Peptides, 8(1987) 515-522. 6 Ekstr6m, P., Central neural connections of the pineal organ and retina in the teleost Gasterosteus aculeatus L., J. Comp. Neurol., 226 (1984) 321-335. 7 Ekstr6m, P., Anterograde and retrograde filling of central neuronal systems with horseradish peroxidase under in vitro conditions, J. Neurosci. Methods, 15 (1985) 21-35. 8 Ekstr6m, P. and Korf, H.-W., Pineal neurons projecting to the brain of the rainbow trout, Salmo gairdneri Richardson (Teleostei). In vitro retrograde filling with horseradish peroxidase, Cell Tissue Res., 240 (1985) 693-700. 9 Ekstrfm, P. and Van Veen, Th., Central connections of the pineal organ in the three-spined stickleback, Gasterosteus aculeatus L. (Teleostei), Cell Tissue Res., 232 (1983) 141-155. 10 Ekstr6m, P. and Van Veen, Th., Pineal neural connections with the brain in two teleosts, the crucian carp and the European eel. J. Pineal Res., 1 (1984) 245-261. 11 EkstrOm, P., Borg, B. and Van Veen, Th., Ontogenetic development of the pineal organ, parapineal organ, and retina of the three-spined stickleback, Gasterosteus aculeatus L. ( Teleostei), Cell Tissue Res., 233 (1983) 593-609. 12 Fernald, F.H. and Finger, T.E., Catecholaminergic neurons of locus coeruleus project to the ganglion cells of the nervus terminalis (NT) in goldfish, Soc. Neurosci. Abstr., 10 (1984) 50. 13 Grimmelikhuijzen, C.J.P., FMRFamide immunoreactivity is generally occurring in the nervous systems of coelenterates, Histochemistrv, 78 (1983) 361-381. 14 Hafeez, M.A, and Zerihun, L., Studies on central projections of the pineal nerve tract in rainbow trout, Salmo gairdneri Richardson, using cobalt chloride iontophoresis, Cell Tissue Res, 154 (1974) 485-510. 15 Korf, H.-W. and M¢ller, M., The central innervationof the mammalian pineal organ, Dev. Endocrinol.. 16 (1985)

of Ms. C. Rasmussen and Ms. I. Norling is gratefully acknowledged. This study was supported by G r a n t B-8554-102 from the Swedish Natural Science Research Council.