Afferent and efferent connections of the habenula in the rainbow trout (Oncorhynchus mykiss) An indocarbocyanine dye (DiI) studyкод для вставкиСкачать
THE JOURNAL OF COMPARATIVE NEUROLOGY 372~529643(1996) Afferent and Efferent Connections of the Habenula in the Rainbow Trout (Oncorhynchus mykiss): An Indocarbocyanine Dye (DiI) Study JULIAN YANEZ AND RAMON ANADON Department of Cell and Molecular Biology, University of La Coruna, 15071 La Coruna, Spain (J.Y.); Department of Fundamental Biology, University of Santiago de Compostela, 15706 Santiago de Compostela, Spain (R.A.) ABSTRACT The habenula is a conserved structure in the brain of vertebrates. With the aim of further understanding of the evolution of the habenular system in vertebrates, we studied the afferent and efferent connections of the habenula of the rainbow trout. Experiments included application of the carbocyanine dye l,l’-dioctadecyl-3,3,3’,3’-tetramethylindoc~bocyanine perchlorate (DiI) into the habenula, telencephalon, pineal organ, posterior tubercle, and interpeduncular nucleus (IPN). The results obtained reveal a consistent pattern of habenular connections. Most afferents originate from three nuclei, one extending from the preoptic region to the rostral thalamus (the entopeduncular nucleus), the second located in the region of the hypothalamus-posterior tubercle and consisting of large bipolar cells (tuberculohabenular nucleus), and the third in the preoptic region (preoptic nucleus). A few large neurons of the locus coeruleus appeared to be labeled in some cases. The trout habenula also receives pineal and parapineal projections. Small labeled glial cells were observed in the thalamus around the fasciculus retroflexus and, sometimes, around the IPN. The most conspicuous efferents coursed in the fasciculus retroflexus to the IPN, the isthmal raphe, and the central gray. The existence of olfactohabenular or habenulotelencephalic projections is discussed. o 1996 Wiley-Liss, Inc. Indexing terms: limbic system, carbocyanine dye, fasciculus retroflexus, entopeduncular nucleus, teleosts The habenula is an evolutionarily conserved structure in the brain of vertebrates. It is located in the caudal epithalamus (synencephalon), and, in mammals, it consists of two habenular nuclei, the medial and the lateral (Herkenham and Nauta, 1977). A similar distinction between the medial and lateral habenular nuclei is found in some nonmammalian vertebrates (Distel and Ebbesson, 1981; Diaz and Puelles, 1992b). The habenular nuclei are connected to several telencephalic, diencephalic, and mesencephalic centers. In mammals, afferent fibers to the habenula arise from telencephalic regions (entopeduncular nucleus, septum, nuclei of the diagonal band, substantia innominata) and from diencephalic and mesencephalic nuclei (Herkenham and Nauta, 1977; Parent et al., 1981).The efferents of the habenula course in the fasciculus retroflexus (FR),which is also called the tractus habenulointerpeduncularis, and project to the interpeduncular nucleus (IPN) and to several regions of the mesencephalon, diencephalon and telencephalon (Cajal, 1911; Herkenham and Nauta, 1979; Contestabile and Flumerfelt, 1981). There have also been a few O 1996 WILEY-LISS, INC. experimental studies on habenular connections in nonmammals. Tracing techniques have indicated that the habenular connections of frog (Kemali et al., 1980; Kemali and Lazar, 1985) and lizard (Distel and Ebbesson, 1981; Diaz and Puelles, 1992a,b), in many respects, are similar to those of mammals. However, frog habenular afferents originate exclusively in nuclei from the telodiencephalic boundary (Kemali et al., 1980). In the larval lamprey, the pattern of habenular connections is even simpler; afferents originate mainly from the lobus subhippocampalis, and efferents course in the FR to the neuropil of the IPN (Yafiez and Anadon, 1994). Classical studies in teleosts with nonexperimental methods (Sheldon, 1912; Holmgren, 1920) indicated that the habenula receives fibers from the ventromedial and ventroAccepted April 2,1996. Address reprint requests to Dr. R. Anadh, Department of Fundamental Biology, Faculty of Biology, University of Santiago de Compostela, Santiago de Compostela 15706, Spain. E-mail: Bfanadon@usc.es J. Y ~ E AND Z R. ANADON 530 lateral parts of the telencephalon through the medial and lateral olfactohabenular tracts, which, together, form the stria medullaris. According to Holmgren (19201, the habenula receives fibers from the diencephalon (nucleus rotundus) through the tractus diencephalohabenularis. This author also mentions a tractus eminentia-habenularis, which originates in cells of the middle part of the eminentia thalami (i.e., regio posthabenularis). In addition, the teleost habenula receives fibers from the pineal and parapineal organs (Holmgren, 1920; Rudeberg, 1969; Ekstrom and van Veen, 1984). In the goldfish, Sheldon (1912) describes other afferents to the habenula from the hippocampus (i.e., dorsal telencephalon), the nucleus teniae (i.e., nucleus interpeduncularis), three parts of the nucleus preopticus, the thalamic tubercle (“Haubenwulst”), and the diencephalon. The habenular efferents run mainly in two bundles, the tractus habenulothalamicus and the tractus habenulopeduncularis or fasciculus retroflexus (Holmgren, 1920; Kappers et al., 1936; Villani et al., 1987, 1994a,b).Strikingly, and despite the numerous telencephalohabenular projections reported on the basis of classical techniques (Sheldon, 1912; Holmgren, 1920; Kappers et al., 1936),there are no modern tracing studies of such projections. However, the results of total telencephalic ablation and bulk telencephalic labeling (Striedter, 1990; Villani et al., 1994a) indicate that telencephalohabenular connections exist. The neurons of origin of these telencephalic connections are currently unknown. Although, in some species, secondary olfactory fibers cross in the habenular commissure, they do not contribute significantly to telencephalohabenular projections (Finger, 1975; Bass, 1981b; von Bartheld et al., 1984; Prasada Rao and Finger, 1984; Levine and Dethier, 1985; Sas et al., 1993). Furthermore, no diencephalic or mesencephalic nucleus, apart from the IPN (Villani et al., 1987, 1994a,b), has yet been shown experimentally to be connected with the teleost habenula. In the present study, we conducted experiments with the indocarbocyanine dye 1,l‘-dioctadecyl3,3,3’,3‘-tetramethylindocarbocyanine perchlorate (DiI), which diffuses in plasma membranes of fixed tissues, to determine both the origin of habenular afferents and the targets of habenular efferents in a teleost, the rainbow trout (Oncorhynchus mykiss, Salmoniformes). MATERIALS AND METHODS Forty rainbow trout, which were obtained from a local fish farm, were used in the experiments. Trout were deeply anaesthetized with tricaine methane sulphonate (MS-222; Sigma) and perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer. The brains were carefully dissected out of the skull and were then embedded in 3% agarose. Crystals of DiI (Eugene, OR) were then applied to the structures under investigation following a modification of Holmqvist et al.’s (1992) method. In brief, blocks were cut at the level of the habenula (or other structures) with a razor blade, crystals of DiI were applied, and the surface was covered with agarose. The brains were then left in fixative for 7-60 days at 37°C. After the transport period, transverse or sagittal sections (50-100 pm thick) obtained by using a Vibratome were examined and photographed with a Nikon fluorescence photomicroscope equipped with a rhodamine filter set and with Nomarski differential interference contrast (DIC). Three types of block were used: 1) with the habenula connected to the brain caudal to it, 2) with the habenula connected to the ipsilateral telencephalic lobe and to the caudal brain, and 3) with the habenula connected to the entire forebrain. In addition, DiI was applied to several other regions (interpeduncular neuropil, isthmal central gray, posterior tubercle, entopeduncular nucleus, and pineal organ) to confirm their habenular connections. Series of Nissl-stained transverse sections of rainbow trout brains from collections in our laboratory were also used for topographic purposes. The nomenclature for regions and nuclei of the trout brainstem was adapted from Nieuwenhuys and Pouwels (1983), that for the diencephalon was adapted from Holmgren (1920) and from Bergqvist (19321, and that for the telencephalon was adapted from Northcutt and Braford (1980). RESULTS The habenulae of the rainbow trout are similar in appearance, although the right habenula is somewhat larger. Unlike in other actinopterygians (Braford and Northcutt, 1983),there is no clear distinction between dorsal and ventral nuclei in trout (Holmgren, 1920). Application of DiI to the habenula led to intense labeling of the habenular (i.e., superior) commissure and a few fiber tracts (fasciculus retroflexus, habenulothalamic tract, tuberculohabenular tract, and stria medullaris). Labeling was most prominent on the side ipsilateral to the application site (Fig. l),but the contralateral fasciculus retroflexus, Abbreviations ADL AP AVL CA CB CH CP DC DL DM FR H HL HY iP IV Lc Nd Ne dorsolateral thalamic area pretectal area ventrolateral thalamic area anterior commissure cerebellum horizontal commissure posterior commissure dorsocentral telencephalic area dorsolateral telencephalic area dorsomedial telencephalic area fasciculus retroflexus habenula hypothalamic lobe hypophysis interpeduncular nucleus trochlear motor nucleus locus coeruleus nucleus diffusus hypothalami entopeduncular nucleus Ng Nlv Npg NPO Nsv NT Nth OT P 4 sv T TL TM TS VC VL VM nucleus pseudoglomerulosus nucleus lateralis valvulae nucleus preglomerulosus preoptic nucleus nucleus sacci vasculosi nucleus anterior tuberis tuberculohabenular nucleus optic tectum pineal optic chiasma saccus vasculosus telencephalon torus lateralis hypothalami mesencephalic tegmentum torus semicircularis valvula cerebelli ventrolateral telencephalic area ventromedial telencephalic area HABENULAR CONNECTIONS IN TROUT 531 Fig. 1. Chartings of transverse sections of the trout brain showing nine perchlorate (DiI) application to the habenula (shaded area). For the ipsilateral distribution of the habenular afferent neurons (small the contralateral connections, see the text. The fasciculus retroflexus is circles) and the fiber tracts (dashes) and fields of terminal fibers (dotted indicated by a large dot. The inset indicates the levels of the sections. For abbreviations, see list. Scale bar = 1mm. areas) revealed after l,l’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocya- J. 532 YAREZ AND R. ANADON Fig. 2. Fluorescence micrograph of section through the entopeduncular nucleus after DiI application to the habenula. This sagittal section passing through the labeled habenula (asterisk) shows the organization of the entopeduncular afferents to the habenula (thin arrow) and the origin of the fasciculus retroflexus (wide arrow). Star, entopeduncular nucleus. The orientation of the figure is indicated by the perpendicular arrows (d, dorsal; r, rostral). Scale bar = 250 pm. Fig. 3. Fluorescence micrograph of section through the entopeduncular nucleus after DiI application to the habenula. Transverse section through the ipsilateral entopeduncular nucleus showing a bundle of habenulopetal fibers (arrow).Note the highly fluorescent central region of the nucleus. Scale bar = 100 p m . tuberculohabenular tract, and stria medullaris were also labeled. the habenula, with many small bundles converging at this point (Fig. 2). In these sections, the origin of the fasciculus retroflexus from several bundles is also evident, although, on the ipsilateral side, these bundles were massively labeled, and it was difficult to observe details. One result of DiI application to the habenula was the labeling of numerous fibers in two conspicuous telencephalic tracts, the corticohabenular (lateral) and olfactohabenular (medial) tracts. The lateral fibers ran between the dorsal and ventral telencephalic areas near the meningeal surface, in the ventrolateral region rich in somatostatinimmunoreactive neurons described by Becerra et al. (1995; nucleus olfactorius lateralis of Sheldon, 1912; lateral nucleus of the area ventralis of Bass, 1981a). Medial fibers ran in the middle of the area ventralis (nucleus olfactorius medialis of Sheldon, 1912). Very interestingly, these fibers give rise to highly conspicuous terminal fields (Figs. 6, 7). Some fibers running in the olfactohabenular tract were found to end in the rostral telencephalon, and a few reached the ventral portion of the olfactory bulb (Fig. 8). No labeled neurons were observed at the end of these tracts, whereas the labeled fibers were observed to branch and give rise to terminal fields; thus, these fibers are clearly telencephalic afferents. The origin of these tracts can be traced to the entopeduncular nucleus, from which three main radiations of fibers appear to arise (rostrolateral, rostromedial, and dorsomedial). In sagittal sections, individual nerve fibers were observed to branch. Very probably, these telencephalic afferents represent collaterals of habenular afferents originated from the entopeduncular nucleus, and not habenular efferents. Afferent neurons Neurons giving rise to habenulopetal projections were present in a few areas of the ventral telencephalon, diencephalon, and mesencephalon (Fig. 1).Interestingly, only a few cells were labeled in the telencephalon rostral to the preoptic region. Following Holmgren (19201, we include this region, which extends around the preoptic recess from caudal to the anterior commissure, to the optic chiasma, as part of the telencephalon. Very numerous afferent cells were located ipsilateral to the application site in a long nucleus extending from the preoptic region caudal, to the anterior commissure, to the rostral thalamus (Figs. 2, 3), accompanying the largest telodiencephalic tract (lateral forebrain bundle). These neurons correspond to the entopeduncular nucleus (area somatica of Holmgren, 1920). The nucleus was so heavily labeled and its neurons were so densely packed that observation of the shape of individual neurons was generally not possible on the ipsilateral side. Neurons were also labeled contralaterally to the application site, although more sparsely. Here, labeled neurons of the entopeduncular nucleus show a round perikaryon and sparsely branched long dendrites (Fig. 41, which extend in several directions, including ventral, medial, rostral, and caudal. The dendrites of these neurons bear small spines in some areas (Fig. 5). In sagittal sections, the course of the axons of these entopeduncular cells was characteristic: They are grouped in small bundles, which run first caudally then, at the level of the habenula, turn dorsally or rostrodorsally to ascend to Figs. 4-7. Labeled cells and processes in the contralateral (Figs. 4, 5) and ipsilateral (Figs. 6, 7) telencephalon after DiI application to the habenula. Fig. 4. Entopeduncular neuron showing the branching of a thick dendrite (arrow). Photomontage of a sagittal section. Scale bar = 25 Fm. Fig. 5 . Detail of a neuronal process of the entopeduncular nucleus showing short lateral spines (small arrows). Scale bar = 25 Fm. Fig. 6. Transverse section through the lateral region of the telencephalic lobe showing the labeled “lateral olfactohabenular tract.” Note the areas of terminal fibers (arrow) and the absence of labeled cells. Star, meninges. Scale bar = 100 pm. Fig. 7. Transverse section through the ventral telencephalic area at the level of the anterior commissure (CAI showing intensely labeled processes in the infracommissural (star) and supracommissural (asterisk) regions. Note the absence of labeled neurons. Scale bar = 100 pm. 534 Fig. 8. Detail of terminal fibers in the caudal olfactory bulb. Scale bar = 50 um. Fig, 9. Transverse section through the telencephalon ipsilateral to the DiI labeled habenula. This section through the preoptic nucleus (asterisk) shows three DiI-labeled cells. Note processes of entopeduncular neurons (upper left). Arrow, lateral optic recess; star, preoptic recess. Scale bar = 100 um. J. YANEZ AND R. ANADON application to the ipsilateral hahenula showing the tuberculohabenular nucleus. Section showing labeled cells extending towards the medial hypothalamic lobe (star). Asterisk indicates the lateral hypothalamic lobe. The orientation of the figure is indicated by the perpendicular arrows (d, dorsal; r, rostral). Scale bar = 250 pm. Fig. 10. Sagittal section through the posterior tubercle after DiI Fig. 11. Sagittal section through the posterior tubercle after DiI application to the ipsilateral habenula showing the tuherculohahenular nucleus. Detail of the labeled cells of Figure 10. Scale bar = 100 pm. In the telencephalon, labeled neurons were also observed in caudal regions but only occasionally rostral to the anterior commissure. A few pear-shaped cells were observed in the supracommissural region, dorsal to the conspicuous central area ventralis ventralis (Vv) terminal field (see below). Very few small cells were labeled ventral or lateral to the anterior commissure or in the central Vv region, and most of these appeared to be putative oligodendrocytes (see below). In the preoptic region, sparse neurons were observed along the preoptic nucleus (Fig. 91, both scattered alongside the lateral area and in the layers of periventricular neurons lateral and ventral to the magnocel- HABENULAR CONNECTIONS IN TROUT lular portion. These neurons are medium sized and are clearly smaller than the cells of the magnocellular portion. In general, fibers are scarce in the preoptic region compared with other regions of the telencephalon. No precommissural telencephalic neurons were observed in our experiments either in ventral or dorsal telencephalic areas or in the olfactory bulb. Following DiI application to the habenula, a large and conspicuous habenulopetal nucleus can be observed in the hypothalamus extending from levels lateral to the organon vasculosum hypothalami to the posterior tubercle, with some neurons even reaching the dorsal walls of the median hypothalamic lobe (Figs. 1,10,11). These cells are recognizable in Nissl-stained sections by their scattered distribution and their spindle-like or piriform shape; they are about 8 km in diameter, somewhat larger than the surrounding periventricular neurons. With DiI labeling, these cells have an elegant shape, with very long, sparsely branched dendritic processes (Fig. 11) that follow ventral, dorsal, or longitudinal courses. These cells lie medial to the preglomerular-pseudoglomerular-mammillary complex of trout, which, from rostral to caudal, shifts progressively medialwards (Fig. 1). The dendrites of labeled neurons form conspicuous dendritic fields, which extend over a considerable area. In longitudinal sections, many of these cells extend in rostrocaudal directions parallel to the major telodiencephalic bundles. In view of its habenular connection, which, to our knowledge, has not been reported previously, we denominate this nucleus the “tuberculohabenular nucleus.” Its axons run rostrally and dorsally, forming a conspicuous tuberculohabenular tract near the medial wall of the ventricle (Fig. 1); this tract ascends towards the habenula medially to the tract of the horizontal commissure. The observation that the tuberculohabenular neurons project to the habenula and not to other related nuclei was confirmed by application of DiI to the tuberal hypothalamus containing them. This experiment revealed a labeled tract ascending parallel to the ventricle and conspicuous fields of terminal fibers in the habenular neuropil (Fig. 12). A portion of these fibers cross in the habenular commissure; thus, the projection is bilateral, which is in agreement with the labeling of a few contralateral cells after DiI application to the habenula. In the rostral rhombencephalon, a few (one to three) large labeled neurons were sometimes observed in the locus coeruleus (Fig. 13). These were multipolar cells with thick dendrites. Following DiI application to the habenula, the posthabenular region was generally observed to contain a sparse population of small round cells medial to the fasciculus retroflexus, in which processes were hardly distinguishable. Similar labeled cells were likewise observed in the forebrain near the medial olfactory tract in the region of the IPN and, very occasionally, in other areas. These cells appeared to be oligodendrocytes. A few tanycytes along the posthabenular course of the fasciculus retroflexus and in the ependymal walls near the IPN were also stained. The application of DiI into the habenula led to the labeling of fibers and cells in the pineal and parapineal organs. The application of DiI to the pineal vesicle led to the labeling of pineal fibers; most of these fibers coursed to the posterior commissure, although some coursed to the habenular commissure. A small subset of these fibers entered the habenula and seemed to form a plexus at its center. Pineal fibers also ran to the pretectal area and to the dorsal 635 and ventral thalamus. Some pineal fibers ran in the fasciculus retroflexus, and a few could be followed to the neuropil of the IPN. Application of DiI to the habenula, as noted above, led to labeling of telencephalic fibers that appeared to be collaterals of habenular afferents originated from the entopeduncular nucleus. To confirm this interpretation, DiI was applied directly to the entopeduncular nucleus. Such application led to labeling of fibers in the stria terminalis and the habenular commissure and labeling of conspicuous terminal fields, but not neurons, in the habenulae. Some entopeduncular nucleus fibers ran in the outer layers of the fasciculus retroflexus to the dorsomedial thalamus, where they gave rise to branches and beaded processes; a few of them even reached the IPN. Efferent connections The habenulointerpeduncular system was intensely stained after DiI application into the habenula. Most labeled efferent fibers of the habenula were found running in the FR, which could be followed caudoventrally in the direction of the IPN. This fascicle originates from the union of several small bundles that converge caudal to the habenula (Fig. 2). The FR forms a compact tract along its posthabenular course, reaching the ventral tegmentum near the oculomotor nucleus and running caudally to the IPN (Fig. 14). The IPN consists of dorsal and ventral neuropils, separated by glia and bands of fibers. Both the dorsal and ventral neuropils of the IPN were intensely labeled (Fig. 15). In the ventral neuropil, individual labeled fibers could generally be distinguished, whereas, in the dorsal neuropil, labeling was sometimes so intense that individual fibers could not be observed (Fig. 15). In cases of faint habenular labeling, FR fibers were observed to branch and course transversal or in other directions in the IPN. Caudal to the IPN, labeled fibers were followed to the raphe, where they formed a dense neuropil (Fig. 16) from which some fibers extended caudally and dorsally to the central gray (Fig. 17) a t isthmal (pretrigeminal) levels. No labeled neurons were found along the course of the FR, indicating that this tract is formed of efferent fibers. The few cells that appeared to be associated with the FR and the IPN could be interpreted as oligodendrocytes (see below). Application of DiI to the IPN or to the fasciculus retroflexus at the level of the caudal thalamus provided confirmation of our results with regard to habenulointerpeduncular projections. Both types of application led to labeling of cells in the habenulae (Figs. 18, 19). Labeled neurons were located throughout the nucleus, indicating that all habenular regions give rise to FR fibers. Habenular neurons were small piriform or stellate cells, with labeled dendrites and processes that formed part of conspicuous neuropil areas surrounded by cords of densely packed cells. In addition to these habenular cells, these types of application led to labeling of a few large neurons in the perihabenular region caudal and lateral to the habenula; these cells had long, sparsely branched processes some with very conspicuous swellings. In addition to the fibers running in the FR and the isthmal raphe, application of DiI to the habenula led to labeling of fibers in the telencephalon, diencephalon, and mesencephalon. In the telencephalon, fibers were observed in two tracts: lateral and medial. These tracts seem to be made up of axon collaterals of entopeduncular neurons (see above), Caudally, labeled fibers were observed to run to the 536 J. YANEZ AND R. ANADON Fig. 12. Detail of the terminal field of tuberculohabenular fibers in the habenula after application of DiI to the posterior tubercle. Star, third ventricle. Scale bar = 50 pm. trance of the fascicle (black star) in the interpeduncular nucleus. Note bilaterality of the interpeduncular neuropil. Scale bar = 250 pm. Fig. 13. Transverse section through the locus coeruleus after DiI application to the hahenula showing a large labeled neuron. Small arrows point to the cell axon. Scale bar = 100 pm. Fig. 14. The fasciculus retroflexus and its terminal fields in the interpeduncular nucleus after DiI application to the habenula. En- Fig. 15. The fasciculus retroflexus and its terminal fields in the interpeduncular nucleus after DiI application to the habenula. The interpeduncular nucleus at a middle level showing its dorsal and ventral neuropils. In this case, the high intensity of fluorescence in the dorsal neuropil obscures individual fibers. Arrows indicate the midline. Scale bar = 100 pm. posthabenular dorsomedial thalamus, where they formed a conspicuous terminal field, and to terminal fields in the pretectal area at the level of the posterior commissure and the dorsal tegmentum. Because this type of application labels the pineal tract, and because DiI application to the pineal vesicle (see above) led to labeling of pineal fibers that HABENULAR CONNECTIONS IN TROUT Fig. 16. Habenular projections to the isthmal raphe and central gray revealed after DiI application to the habenula. Dense field of terminals in the raphe (star) just caudal to the interpeduncular nucleus. Scale bar = 100 mm. Fig. 17. Habenular projections to the isthmal raphe and central gray revealed after DiI application to the habenula. Labeled habenular fibers ascending through the raphe to the isthmal central gray. Asterisk, aqueduct; stars, medial longitudinal fascicle. Scale bar = 100 km. 537 Fig. 18. Labeling of the habenula and fasciculus retroflexus after DiI application to the interpeduncular nucleus. Transverse section through the habenulae (stars) showing the bilateral labeling of the fasciculus retroflexus, which, at this caudal level, is formed of many small bundles (wide arrows). Asterisk, optic tectum. Scale bar = 250 Fm. Fig. 19. Labeling of the habenula and fasciculus retroflexus after DiI application to the interpeduncular nucleus. Detail of the left habenula of Figure 18 showing labeled habenular neurons. Scale bar = 50 km. J. YANEZ AND R. ANADON 538 form conspicuous terminal fields in these areas, it seems very probable that these projections were not habenular. context, because most habenular connections in trout are unmyelinated. Subdivision of the habenulae DISCUSSION To our knowledge, this is the first experimental study of the afferent and efferent connections of the trout habenula. Our results show, first, that most habenular afferents arise from only a few nuclei (notably, the entopeduncular nucleus, the parvicellular preoptic nucleus, and a habenulopetal nucleus of the posterior tubercle) and, second, that the fasciculus retroflexus contains most, if not all, of the habenular efferents. Methodological considerations First, the DiI method used led to labeling of fibers that branched to form conspicuous terminal fields; this indicates that the transport times used (7-60 days) were sufficient for complete labeling of fibers. Labeling was highly reproducible provided that care was taken to ensure that crystals did not contact neighboring regions. The dye diffused both anterogradely and retrogradely along processes, as was shown in reciprocal labeling experiments, and it labeled axon collaterals regardless of the point of application. This explains why DiI labeling revealed the complex pattern of projections of the entopeduncular nucleus after application into the habenula (see below). This characteristic of the labeling procedure used, which is similar to that suspected to occur with wheat germ agglutinin (WGAI-conjugated horseradish peroxidase (HRP) labeling (Sas et al., 19931, should be borne in mind when interpreting the present results. For instance, the terminal fields observed in the ventral and dorsolateral telencephalon after application of DiI to the habenula should not be interpreted as habenulotelencephalic projections. Second, it should be stressed that both axon membranes and myelin sheaths in the region of dye application were labeled (in transverse sections appearing as fluorescent rings). Thus, in cases of habenular application, the posthabenular region contained small, round, labeled cells medial to the fasciculus retroflexus, in which the presence of processes was scarcely distinguishable. Likewise, similar cells were occasionally observed near the medial olfactory tract, in the region of the IPN, and (less frequently still) in other regions, always close to tracts with conspicuously labeled myelinated fibers. Most (if not all) of these round cells are probably oligodendrocytes that were labeled because their membranes are continuous with myelin sheaths. The possible labeling of oligodendrocytes was also remarked upon by von Bartheld et al. (1990),and it may lead to such cells being identified mistakenly as neurons. Interestingly, in areas containing putative oligodendrocytes (close to the fasciculus retroflexus along its thalamic course and dorsal to the IPN), a few stained tanycytes were generally observed. The association of the two types of cell (oligodendrocytes and tanycytes) with tracts containing heavily labeled myelinated fibers strongly suggests that DiI diffuses from myelinated sheaths to tanycytes, probably via oligodendrocytes that contact tanycyte processes through gap junctions, as was found recently in a teleost (Diaz-Regueira et al., 1992). Moreover, the variable number and location of these cells are further arguments in favor of their nonneuronal character. Regardless of these considerations, this methodological problem is of limited concern in the present Holmgren (1920) reported that the habenulae of salmonids are asymmetric, the right habenular nucleus being larger than the left and the neurons of the left being more compactly grouped than those of the right; our observations are in agreement with these findings. Because our DiI applications affected all of the habenula, distinction between the connections of dorsal and ventral parts was not possible. Labeling of habenulorecipient nuclei did not reveal differences in innervation between dorsal and ventral parts; however, we cannot rule out the possibility that different areas within the habenulae have different connections. The left habenula of coho salmon contains a serotoninimmunoreactive subnucleus (Ekstrom and Ebbesson, 1988); in the right habenula, on the other hand, serotoninimmunoreactive neurons appear only transiently during development (Ebbesson et al., 1992). Moreover, only the ventral portion of the trout habenula contains a significant number of somatostatinergic fibers (Becerra et al., 1995), and a heterogeneous distribution of other types of fiber has also been observed (unpublished results). Habenular afferents In trout, most habenular afferents come from specific nuclei in the telodiencephalic boundary (entopeduncular nucleus, preoptic nucleus) and the posterior tubercle (tuberculohabenular nucleus). However, afferents from other regions (supracommissural telencephalon, pineal, parapineal, locus coeruleus) were also observed. In amphibians, reptiles, and mammals, habenular afferents originate from several telencephalic, diencephalic, and mesencephalic nuclei (amphibians: Kemali et al., 1980; reptiles: Hoogland, 1982; Diaz and Puelles, 1992b; mammals: Herkenham and Nauta, 1977; Parent et al., 1981; Van der Kooy and Carter, 1981; Phillipson and Pycock, 1982; Swanson, 1982; Risold et al., 1994). In what follows, we compare our results with those reported previously in fish and in other vertebrates. In particular, we discuss the possible roles of the habenular circuitry in descending and ascending forebrain pathways. Our results in trout reveal a major habenulopetal nucleus rostral to the optic chiasma. This nucleus correspond to the “area somatica” of Holmgren (1920) and to the entopeduncular nucleus of Sheldon (19121, Bergqvist (19321, Northcutt and Davis (19831, and Gomez-Segade and Anadon (1988). To our knowledge, this is the first demonstration of habenular projections of the entopeduncular nucleus in a teleost. In trout, Vigh-Teichmann et al. (1983) and 01ivereau et al. (1984) identified a somatostatinergic nucleus located at the level of the anterior commissure and rostral to it as the entopeduncular nucleus (Becerra et al., 1995); this nucleus probably corresponds to the nucleus olfactorius lateralis of Sheldon (19121, the nucleus olfactorius lateralis of Holmgren (19201, or the lateral nucleus of the area ventralis of Bass (1981a). Early studies suggested that as many as seven telencephalic tracts project to the habenula in brook trout (tr. olfactohabenularis, tr. hippocampohabenularis pars lateralis, tr. corticohabenularis, tr. teniae, tr. olfactorius lateralis habenulae rectus et cruciatus, tr. hippocampohabenularis pars medialis, and tr. olfactorius medialis habenulae; Holmgren, 1920). Sheldon (1912) likewise described several telencephalohabenular tracts and, in addition, a tractus HABENULAR CONNECTIONS IN TROUT preopticohabenularis (with partes anterior, medialis, lateralis, and posterior) in the goldfish. No tracts of this type were detected in the present study. Indeed, our results after application of dye in the habenula and ventral telencephalon suggest that the “habenulopetal” tracts of Holmgren (19201, in fact, are entopedunculofugal projections to the ventral and laterodorsal telencephalic areas. Experimental studies of olfactory bulb efferents have indicated that olfactory projections that cross in the habenular commissure are present in some teleost species (Finger, 1975; Bass, 1981b; Northcutt and Davis, 1983; von Bartheld et al., 1984; Levine and Dethier, 1985; Sas et al., 1993) but not in others (Scalia and Ebbesson, 1971; Davis et al., 1981; Ebbesson et al., 1981; Murakami et al., 1983; Prasada Rao and Finger, 1984). Like these latter authors, we did not find evidence of olfactohabenular fibers in trout. Even in species in which olfactory fibers use the habenular commissure, it is not clear whether they make any contacts in the habenula, although habenular terminal fields have been reported (Sas et al., 1993).The experiments of Herkenham and Nauta (1977) in the rat indicated that the stria medullaris contains passage fibers from several olfactory structures that do not terminate in the habenula. Axons of some telencephalic lobe neurons in lamprey also cross the habenular commissure, although they do not seem to contact habenular neurons (Northcutt and Puzdrowski, 1988; Yaiiez and Anadon, 1994). Some studies of the bulk telencephalic efferents in a few teleost species have indicated the existence of habenulopeta1 projections (Striedter, 1990; Villani et al., 1994a). However, Vanegas and Ebbesson (19801, Echteler and Saidel (1981), and Murakami et al. (1983) do not mention telencephalohabenular projections. In trout, we could not find any evidence of such telencephalohabenular projections: specifically, no labeled cells were found either in the olfactory bulb or in the precommissural telencephalon. Moreover, our results suggest that the entopeduncular nucleus is the only nucleus with a significant habenular projection, although a small habenulopetal nucleus is also present in the commissural region. In view of its position, the entopeduncular nucleus of the trout appears to correspond to the main habenulopetal nucleus of lamprey, the subhippocampothalamic nucleus (Yafiez and Anadon, 1994); thus, these nuclei can be considered as homologous. In the habenula of land vertebrates, most telencephalic afferents arise from nuclei intercalated in the telencephalic tracts. In frog, afferents originate ipsilaterally from three nuclei of the telodiencephalic boundary (the entopeduncular nucleus, the nucleus dorsomedialis anterior thalami, and the adjacent periventricular area; Kemali et al., 1980). These nuclei may correspond to the cells of the posterior pole of the telencephalon and striatal complex found to project to the habenula in newts (Clairambault et al., 1986). In lizards, the several telencephalic nuclei projecting to the medial habenula (nucleus of the posterior pallial commissure, nucleus septalis impar, nucleus eminentia thalami, and nucleus of the stria medullaris) and to the lateral habenula (bed nucleus of the anterior commissure, diagonal band nucleus, and lateral preoptic area) are related to the main telencephalic tracts (Diaz and Puelles, 1992b). In mammals, most telencephalic afferents to the habenula come from nuclei intercalated in the main telencephalic tracts: the entopeduncular nucleus (Herkenham and Nauta, 1977; Parent et al., 1981; Van der Kooy and Carter, 1981; Araki et al., 1984), the nucleus of the diagonal band 539 (Herkenham and Nauta, 1977; Parent et al.,1981; Contestabile and Fonnum, 19831, the inner segment of the pallidum (Parent and De Bellefeuille, 1982; Hazrati and Parent, 19911, the bed nucleus of the anterior commissure (Staines et al., 19881, and the nuclei of the posterior (supracommissural) septum (n. septofimbrialis, n. triangularis; Herkenham and Nauta, 1977; Parent et al., 1981; Contestabile and Fonnum, 1983; Fonnum and Contestabile, 1984; Staines et al., 1988; Kawaja et al., 1990). The two habenular nuclei of mammals (the medial and the lateral) can be distinguished by their different afferents (Herkenham and Nauta, 1977), and similar results have been obtained in reptiles (Diaz and Puelles, 1992b). The lack of distinction between medial and lateral habenular nuclei in trout and the presence of a single main habenulopetal nucleus in the telencephalon might be related to the compact organization of the actinopterygian telencephalon, which is strikingly different from that of other vertebrates (see Northcutt and Braford, 1980). By application of DiI to either the habenula or the pineal organ, we were able to confirm the existence of pineal projections to the trout habenula (Hafeez and Zerihun, 1974). Thus, it is conceivable that habenular function is influenced by light. Pinealohabenular projections have also been reported on the basis of experimental studies in other teleosts (Ekstrom, 1984; Ekstrom and van Veen, 1984). This contrasts with the apparent lack of pineal projections to the habenula in lampreys (Puzdrowski and Northcutt, 1989; Yadez et al., 1993) and in some amphibians (Paul et al., 1971; Eldred et al., 1980). Application of DiI to the habenula or the parapineal organ of trout (Yaiiez et al., 1996; present results) confirms the existence of parapinealohabenular projections (Rudeberg, 1969). The posterior tubercle of trout contains a well-developed habenulopetal nucleus (the tuberculohabenular nucleus). This nucleus, which is located rostral or somewhat lateral to the posterior tuberal nucleus itself, extends along the course of the main telencephalohypothalamic tracts. Thus, it probably corresponds to the habenulopetal cells located in the lateral hypothalamic and lateral mammillary areas of the lizard (Diaz and Puelles, 199213) and, possibly, to the habenulopetal cells reported from the medial hypothalamus (Parent et al., 1981), the arcuate nucleus (Sim and Joseph, 1991), and the mammillary region (Veazey et al., 1982) of mammals. In the lungfish, neurons in the paraventricular organ seem to project to the habenula (von Bartheld and Meyer, 1990); the location of these cells in the tuberal region suggests that they might correspond to the trout habenulopetal nucleus, but it should be noted that they are cerebrospinal fluid contacting. The periventricular distribution of neurons in the brain of the lungfish, which is totally unlike the more complex distribution observed in teleosts, might explain this difference. Habenulopetal cells have not been detected in this region either in lamprey (Yafiez and Anadon, 1994) or in amphibians (Kemali et al., 1980; Clairambault et al., 1986).These results suggest that a tuberal population of habenulopetal cells might have appeared independently in teleost, lungfish, and amniote lineages. However, experimental studies in other vertebrate groups are needed to rule out the possibility that the lack of a tuberal habenulopetal system in amphibians is a derived characteristic. J. YANEZ AND R. ANADON 540 Possible origins of chemically identified afferent fibers A number of chemically identified afferent fiber types have been observed in the salmonid habenula, including somatostatinergic (Becerra et al., 1995), catecholaminergic (Manso et al., unpublished observations), serotoninergic (Ekstrom and Ebbesson, 1988), and FMRF-amide fibers (Ekstrom et al., 1988; Ostholm et al., 1990). Interestingly, habenulopetal somatostatinergic (Palkovits et al., 1982), dopaminergic (Phillipson and Griffith, 1980; Phillipson and Pycock, 1982; Skagerberg et al., 1984), and serotoninergic (Steinbusch, 1984) fibers have also been demonstrated in mammals. Owing to the small number of these fibers observed in the trout habenula, with the exception of somatostatin, they probably do not form part of a specific habenulopetal system, but they may form part of diffuse projection systems to a number of brain regions. The size and distribution of DiI-labeled cells in the preoptic nucleus is similar to that previously reported for somatostatinergic cells in adult trout (Becerra et al., 19951, which suggests that the somatostatinergic fibers observed in the habenula may originate from the preoptic nucleus. Although the entopeduncular nucleus gives rise to habenulopetal somatostatinergic projections in mammals (Vincent and Brown, 1986), the trout entopeduncular nucleus does not contain somatostatin-immunoreactive cells (Becerra et al., 1995),and it is currently unclear whether the entopeduncular nuclei of mammals and teleosts are homologous. For some of the other fiber types, our DiI applications likewise indicate a possible nucleus of origin. The cells labeled in the locus coeruleus are very similar in location and appearance to the large tyrosine hydroxylase-immunoreactive neurons present in this region of the trout brain (Manso et al., 1993). In teleosts, locus coeruleus cells form a diffuse noradrenergic projection system to most brain regions (Ekstrom et al., 1986; Hornby and Piekut, 1990; Meek et al., 1993; Ma, 1994). Serotoninergic fibers were scarce in the trout habenula. Six areas containing serotoninergic cell bodies have been demonstrated in the diencephalon and mesencephalon of rainbow trout (Frankenhuys-van den Heuvel and Nieuwenhuys, 19841, and the pineal organ also contains serotoninergic cells (van Veen et al., 1984).Although the posthabenular region, the nucleus anterior tuberis, the periventricular region of the tuberal periventricular region, and the raphe nuclei of teleosts contain a rich population of serotoninergic cells (Frankenhuys-van den Heuvel and Nieuwenhuys, 1984; Meek and Joosten, 1989; Johnston et al., 19901, we did not observe DiI-labeled neurons in any of these locations. The present demonstration of pinealohabenular projections strongly suggests that the serotoninergic fibers in the rainbow trout habenula arise from the pineal organ. Although serotoninergic cells have been observed in the left habenula of coho salmon (Ekstrom and Ebbesson, 19881, such cells were not detected in rainbow trout (Frankenhuysvan den Heuvel and Nieuwenhuys, 1984). Scarce FMRF-amide-immunoreactive fibers were observed in the trout habenula (Ekstrom et al., 1988); however, it is not clear whether these terminate in, or simply pass through, the habenula. There are a number of FMRFamide-immunoreactive nuclei in the telencephalon, diencephalon, and mesencephalon of teleosts (Fujii and Kobayashi, 1992; Vecino and Ekstrom, 1992);in the present study, however, application of DiI to the habenula did not lead to labeling of neurons in any of these previously reported locations. In salmon, gonadotropinreleasinghormone (GnRHIimmunoreactivefibers are also present in the habenula (Amano et al., 1991), although, again, it is not clear whether they originate in terminal ganghon or in telencephalic or preoptic GnRH-immunoreactiveneurons. The fact that application of DiI to the trout habenula did not lead to labeling of known FMRF-amide- or GnRH-immunoreactive nuclei may simply reflect the scarcity of this type of fiber in the habenula. Habenular efferents Similar to mammals (Herkenham and Nauta, 19791, frogs (Kemali et al., 1980), and lamprey (Yafiez and Anadon, 1994),the habenular efferents of trout run in the FR. This is in good agreement with the results of immunocytochemistry and habenular ablation in goldfish (Villani et al., 1987, 1991, 1994a,b). The trout FR seems to project not only to the IPN (Villani et al., 1987, 1991, 1994a,b) but also to the isthmal raphe. A similar distribution of the FR, with fibers extending caudally into the isthmal raphe, has been reported in lampreys (Yanez and Anadon, 1994), amphibians (Kemali and Lazar, 1985), lizards (Distel and Ebbesson, 1981; Diaz and Puelles, 1992a),and mammals (Herkenham and Nauta, 1979). Interestingly, the habenular projections to the interpeduncular and raphe nuclei appear to correspond only to the core of the FR of reptiles and mammals, which originates from the medial habenular nucleus (Distel and Ebbesson, 1981; Diaz and Puelles, 1992a; Herkenham and Nauta, 1979; Contestabile and Flumerfelt, 1981; Hamill and Jacobowitz, 1984; Contestabile et al., 1987). In lizards and mammals, the FR also projects fibers to mesencephalic, diencephalic, and telencephalic nuclei (mammals: Herkenham and Nauta, 1979; lizard: Distel and Ebbesson, 1981; Diaz and Puelles, 1992a). Interestingly, projections to the periaqueductal gray (Beat et al., 19901, substantia nigra (Christoph et al., 19861, and ventral tegmental area (Christoph et al., 1986; Oades and Halliday, 1987) arise from the lateral habenular nucleus and run in the mantle of the FR; similar observations have been made in lizards (Distel and Ebbesson, 1981). Together, these results suggest that there is no nucleus homologous to the lateral habenular nucleus either in trout (present results) or in lamprey (Yafiez and Anadon, 1994).Interestingly, the lateral habenular nucleus of mammals receives afferents from telencephalic nuclei that do not project to the medial habenula, like the entopeduncular nucleus (Herkenham and Nauta, 1977; Van der Kooy and Carter, 1981; Araki et al., 1984; Vincent and Brown, 1986) and the globus pallidus (Parent and De Bellefeuille, 1982; Hazrati and Parent, 1991). Again, similar observations have been made in lizards (Hoogland, 1982; Diaz and Puelles, 1992b). Thus, the lateral habenular nuclei of lizards and mammals might be a relatively recent development in vertebrate phylogeny, perhaps reflecting the increasing importance of cortical circuits. If so, it seems likely that the land vertebrate homologue of the principal habenulopetal nucleus in trout (the entopeduncular nucleus) is not the entopeduncular nucleusiglobus pallidus, but, rather, it may be the septa1 nuclei, which, in lizards and in mammals, project to the medial habenula. A further problem when analyzing the habenular projections of trout after DiI labeling is the interpretation of fibers observed in some thalamic, pretectal, and tegmental regions. In experiments in which DiI was applied to the HABENULAR CONNECTIONS IN TROUT pineal vesicle, all of these areas received a rich innervation, in agreement with results of Hafeez and Zerihun (1974). The pattern of fibers in these regions is similar regardless of whether DiI is applied to the habenula or to the pineal vesicle, which suggests that the fibers in thalamic, pretectal, and tegmental regions found when DiI is applied to the habenula are pineal fibers that become labeled through their habenulopetal collaterals. In conclusion, our observations suggest that the habenular system of fishes forms a system of projections between nuclei of the caudal telencephalon and posterior tubercle and the basal rnesencephalonirostral medulla. This system is simpler than that found in amniotes. Comparison of teleosts with amniote species in which the habenular system has been studied experimentally indicates that their pattern of connections is similar only to the medial habenular nucleus of reptiles and mammals. These land vertebrates appear to have acquired additional habenular circuitry with the appearance of the lateral habenular nucleus. Further study of habenular connections in other groups of fishes may shed light on whether the absence of a lateral habenular nucleus homologue in teleosts is a primitive or a derivative characteristic. ACKNOWLEDGMENTS The authors thank Mr. G. Norman for his stylistic suggestions for improving the paper and Miss C. Farifia for providing the trouts. This work was supported by grants from the DGICYT (93-0527CO201) and the Xunta de Galicia (XUGA20001B93). LITERATURE CITED Amano, M., Y. Oka, K. Aida, N. Okumoto, S. Yawashima, and Y. Hasegawa (1991)Immunocytochemical demonstration of salmon GnRH and chicken GnRH-I1 in the brain of masu salmon, Oncorhynchus mosou. J. Comp. Neurol. 314:587-597. Araki, M., P.L. McGeer, and E.G. McGeer (1984) Retrograde HRP tracing combined with a pharmacohistochemical method for GABA transaminase for the identification of presumptive GABAergic projections to the habenula. Brain Res. 304:271-277. Bass, A.H. (1981a) Organization of the telencephalon in the channel catfish, Ictaluruspunctatus. J. Morphol. 169:71-90. Bass, A.H. (1981b) Olfactory bulb efferents in the channel catfish, Ictalurus punctatus. J. Morphol. 169:91-111. Beat, P.M., R.J. Summers, J.A. Stephenson, C.J. Cook, and M.J. Christie (1990) Excitatory amino acid projections to the periaqueductal gray in the rat: A retrograde transport study utilizing DL3Hlaspartate and ["IGAEiA. Neuroscience 34r163-176. Becerra, M., M.J. Manso, I. Rodriguez-Moldes, and R. Anadon (1995) Ontogeny of somatostatin-immunoreactive systems in the brain of the brown trout (Teleostei).Anat. Embryol. 191:119-137. Bergqvist, H. (1932) Zur Morphologie des Zwischenhirns bei niederen Wirbeltieren. Acta Zool. 1357-304. Braford, M.R., Jr., and R.G. Northcutt (1983) Organization of the diencephalon and pretectum of the ray-finned fishes. In R.E. Davis and R.G. Northcutt (eds): Fish Neurobiology, Vol. 2, Higher Brain Areas and Functions. Ann Arbor, MI: The University of Michigan Press, pp. 117-163. Cajal, S.R. (1911) Histologie du Systeme Nerveux de 1'Homme et des Vertebres. Vol. 2. Paris: Maloine. Christoph, G.R., R.J. Leonzio, and K.S. Wilcox (1986) Stimulation of the lateral habenula inhibits dopamine-containing neurons in the substantia nigra and ventral tegmental area of the rat. J. Neurosci. 6:613-619. Clairambault, P., C. Pairault, and A. Fasolo (1986) Habenular connections in the brain of the newt, Triturus cristatus carnifex Laurenti. J. Hirnforsch. 27:ll-119. Contestabile, A,, and B.A. Flumerfelt (1981) Afferent connections of the interpeduncular nucleus and the topogaphic organization of the ha- 541 benulo-interpeduncular pathway: An HRP study in the rat. J. Comp. Neurol. 196r253-270. Contestabile, A., and F. Fonnum (1983) Cholinergic and GABAergic forebrain projections to the habenula and nucleus interpeduncularis: Surgical and kainic acid lesions. Brain Res. 275287-297, Contestabile, A., L. Villani, A. Fasolo, M.F. Franzoni, L. Gribaudo, 0 . Oktedalen, and F. Fonnum (1987) Topography of cholinergic and substance P pathways in the habenulo-interpeduncular system of the rat. An immunocytochemical and microchemical approach. Neuroscience 21.953-270. Davis, R.E., R. Chase, J. Morris, and B. Kaufman (1981) Telencephalon of the teleost Macropodus: Experimental localization of secondary olfactory areas and of components of the lateral forebrain bundle. Behav. Neural Biol. 33357-279. Diaz, C., and L. Puelles (1992a) In vitro HRP-labeling of the fasciculus retroflexus in the lizard Gallotia galloti. Brain Behav. Evol. 39:305-311. Diaz, C., and L. Puelles (199213) Afferent connections of the habenular ganglion in the lizard Gallotiagalloti. Brain Behav. Evol. 39:312-324. Diaz-Regueira, S., M. Becerra, and R. Anadon (1992) Light- and electron microscopic study of oligodendrocytes in the lateral line area of the medulla in Chelon labrosus. J. Hirnforsch. 33:477485. Distel, H., and S.O.E. Ebbesson (1981) Habenularprojections in the monitor lizard (Varanus senegalensis). Exp. Brain Res. 43:324429. Ebbesson, S.O.E., L.M. Meyer, and H. Scheich (1981) Connections of the olfactory bulb of the piranha (Serrasalmus nattereri). Cell Tissue Res. 216:167-180. Ebbesson, S.O.E., B. Holmqvist, T. Ostholm, and P. Ekstrom (1992) Transient serotonin-immunoreactive neurons coincide with a critical period of neural development in coho salmon (Oncorhynchus kisutch). Cell Tissue Res. 268:389-392. Echteler, S.M., and W.M. Saidel(1981) Forebrain connections in the goldfish support telencephalic homologies with land vertebrates. Science 212:683685. Ekstrom, P . (1984) Central neural connections of the pineal organ and retina in the teleost Gasterosteus aculeatus L. J. Comp. Neurol. 226:321-335. Ekstrom, P., and S.O.E. Ebbesson (1988) The left habenular nucleus contains a discrete serotonin-immunoreactive subnucleus in the coho salmon (Oncorhynchus kisutch). Neurosci. Lett. 91:121-125. Ekstrom, P., and T. van Veen (1984) Pineal neural connections with the brain in two teleosts, the crucian carp and the European eel. J. Pineal Res. 1245-261. Ekstrom, P., M. Reschke, H. Steinbusch, and T. Van Veen (1986) Distribution of noradrenaline in the brain of the teleost Gasterosteus aculeatus L.: An immunohistochemical analysis. J. Comp. Neurol. 254297-313. Ekstrom, P., T. Honkanen, and S.O.E. Ebbesson (1988) FMRF-amide-like immunoreactive neurons of the n e w s terminalis of teleosts innervate both retina and pineal organ. Brain Res. 460t68-75. Eldred, W.D., T.E. Finger, and J. Nolte (1980) Central projections of the frontal organ of Rana pipiens, as demonstrated by the anterograde transport of horseradish peroxidase. Cell Tissue Res. 211:215-222. Finger, T.E. (1975) The distribution of the olfactory tracts in the bullhead catfish, Ictalurus nebulosus. J. Comp. Neurol. 165:125-142. Fonnum, F., and A. Contestabile (1984) Colchicine neurotoxicity demonstrates the cholinergic projection from the supracommissural septum to the habenula and the nucleus interpeduncularis in the rat. J. Neurochem. 432381-884. Frankenhuys-van den Heuvel, T.H.M., and R. Nieuwenhuys (1984) Distribution of serotonin-immunoreactivityin the diencephalon and mesencephalon of the trout, Salmo gairdneri. Anat. Embryol. 169:193-204. Fujii, K., and H. Kobayashi (1992) FMRF-amide-like immunoreactivity in the brain and pituitary of the goldfish, Carassius auratus. Ann. Anat. 174.917-222. Gomez-Segade, P., and R. Anad6n (1988) Specialization in the diencephalon of advanced teleosts. J. Morphol. 197:71-104. Hafeez, M.A., and L. Zerihun (1974) Studies on central projections of the pineal nerve tract in rainbow trout, Salrno gairdneri Richardson, using cobalt chloride iontophoresis. Cell Tissue Res. 154:485-510. Hamill, G.S., and D.M. Jacobowitz (1984) A study of afferent projections to the rat interpeduncular nucleus. Brain Res. Bull. 13527-539. Hazrati, L.N., and A. Parent (1991) Contralateral pallidothalamic and pallidotegmental projections in primates: An anterograde and retrograde labeling study. Brain Res. 567.212-223. Herkenham, M., and W.J.H. Nauta (1977) M e r e n t connections of the habenular nuclei in the rat. A horseradish peroxidase study, with a note on the fiber-of-passage problem. J. Comp. Neurol. 173:123-146. 542 Herkenham, M., and W.J.H. Nauta (1979) Efferent connections of the habenular nuclei in the rat. J. Comp. Neurol. 187t19-48. Holmgren, N. (1920) Zur Anatomie und Histologie des Vorder- und Zwischenhirns der Knochenfische. Hauptsachlich nach Untersuchungen an Osmerus eperlanus. Acta Zool. 1:137-315. Holmqvist, B.I., T. Ostholm, and P. Ekstrom (1992) DiI tracing in combination with immunocytochemistry for analysis of connectivities and chemoarchitectonics of specific neural systems in a teleost, the Atlantic salmon. J. Neurosci. Methods 4245-63. Hoogland, P.V. (1982) Brainstem afferents to the thalamus in a lizard, Varanus exanthematicus. J. Comp. Neurol. 210:152-162. Hornby, P.J., and D.T. Piekut (1990) Distribution of catecholaminesynthesizing enzymes in goldfish brains: Presumptive dopamine and norepinephrine neuronal organization. Brain Behav. Evol. 35:49-64. Johnston, S.A., L. Maler, and B. Tinner (1990) The distribution of serotonin in the brain of Apteronotus leptorhynchus: An immunohistochemical study. J. Chem. Neuroanat. 3:429-466. Kappers, C.U.A., G.C. Huber, and E.C. Crosby (1936) The Comparative Anatomy of the Nervous System of Vertebrates, Including Man. New York: MacMillan. Kawaja, M.D., B.A. Flumerfelt, and A.W. Hrycyshyn (1990) Synaptic organization of septal projections in the rat medial habenula: A wheat germ agglutinin-horseradish peroxidase and immunohistochemical study. Synapse 6:45-54. Kemali, M., and G. Lazar (1985) Cobalt injected into the right and left fasciculi retroflexes clarifies the organization of this pathway. J. Comp. Neurol. 233:l-11. Kemali, M., V. Guglielmotti, and D. Gioffre (1980) Neuroanatomical identification of the frog habenular connections using peroxidase (HRP). Exp. Brain Res. 38:341-347. Levine, R.L., and S. Dethier (1985) The connections between the olfactory bulb and the brain in the goldfish. J. Comp. Neurol. 237:427-444. Ma, P.M. (1994) Catecholaminergic systems in the zebrafish. 11. Projection pathways and pattern of termination of the locus coeruleus. J. Comp. Neurol. 344:256-269. Manso, M.J., M. Becerra, P. Molist, I. Rodriguez-Moldes, and R. Anadon (1993) Distribution and development of catecholaminergic neurons in the brain of the brown trout. A tyrosine hydroxylase immunohistochemical study. J. Hirnforsch. 34:239-260. Meek, J . , and H.W.J. Joosten (1989) Distribution of serotonin in the brain of the mormyrid teleost Gnathonemuspetersii. J. Comp. Neurol. 281:206224. Meek, J., H.W.J. Joosten, and T.G.M. Hafmans (1993) Distribution of noradrenaline-immunoreactivity in the brain of the mormyrid teleost Gnathonemus petersii. J. Comp. Neurol. 328: 145-160. Murakami, T., Y. Morita, and H. Ito (1983) Extrinsic and intrinsic fiber connections of the telencephalon in a teleost, Sebastiscus marmorutus. J. Comp. Neurol. 216:115-131. Nieuwenhuys, R., and E. Pouwels (1983) The brainstem of actinopterygian fishes. In R.E. Davis and R.G. Northcutt (eds): Fish Neurobiology, Vol. 1, Brainstem and Sense Organs. Ann Arbor, MI: The University of Michigan Press, pp. 25-87. Northcutt, R.G., and M.R. Braford (1980) New observations on the organization and evolution of the telencephalon of actinopterygian fishes. In S.O.E. Ebbesson (ed): Comparative Neurology of the Telencephalon. New York: Plenum Press, pp. 41-98. Northcutt, R.G., and R.E. Davis (1983) Telencephalic organization in ray-finned fishes. In R.E. Davis and R.G. Northcutt (eds): Fish Neurobiology, Vol. 2, Higher Brain Areas and Functions. Ann Arbor, MI: The University of Michigan Press, pp. 203-236. Northcutt, R.G., and R.L. Puzdrowski (1988) Projections of the olfactory bulb and nervus terminalis in the silver lamprey. Brain Behav. Evol. 32:96-107. Oades, R.D., and G.M. Halliday (1987) Ventral tegmental (A101 system: Neurobiology. 1.Anatomy and connectivity. Brain Res. 434:117-165. Olivereau, M., F. Ollevier, F. Vandesande, and J. Olivereau (1984) Somatostatin in the brain and pituitary of some teleosts. Immunocytochemical identification and effects of starvation. Cell Tissue Res. 238:289-296. Ostholm, T., P. Ekstrom, and S.O.E. Ebbesson (1990) Distribution of FMRF-amide-like immunoreactivity in the brain, retina and nervus terminalis of the sockeye salmon parr, Oncorhynchus nerka. Cell Tissue Res. 261~403-418. Palkovits, M., L. Tapia-Arancibia, C. Kordon, and J. Epelbaum (1982) Somatostatin connections between the hypothalamus and the limbic system of the rat brain. Brain Res. 250:223-228. J. YMEZ AND R. ANADON Parent, A,, and L. De Bellefeuille (1982) Organization of efferent projections from the internal segment of the glohus pallidus in primate as revealed by fluorescence retrograde labeling method. Brain Res. 245201-213. Parent, A,, S. Gravel, and R. Boucher (1981) The origin of forebrain afferents to the habenula in rat, cat and monkey. Brain Res. Bull. 6:23-38. Paul, E., H.G. Hartwig, and A. Oksche (1971) Neurone und zentralnervose Verbindungen des Pinealorgans der Anuren. Z. Zellforsch. 112:466-493. Phillipson, O.T., and A.C. Griffith (1980) The neurons of origin for the mesohabenular dopamine pathway. Brain Res. 197:213-218. Phillipson, O.T., and C.J. Pycock (1982) Dopamine neurons in the ventral tegmentum project to both the medial and lateral habenula. Exp. Brain Res. 45:89-94. Prasada Rao, P.D., and T.E. Finger (1984) Asymmetry of the olfactory system in the brain of the winter flounder, Pseudopleuronectes umericenus. J. Comp. Neurol. 225492-510. Puzdrowski, R.L., and R.G. Northcutt (1989) Central projections of the pineal complex in the silver lamprey Zchthyomyzon unicuspis. Cell Tissue Res. 225:269-274. Risold, P.Y., N.S. Canteras, and L.W. Swanson (1994) Organization of projections from the anterior hypothalamic nucleus: A Phaseolus uulguris-leucoagglutinin study in the rat. J. Comp. Neurol. 348:140. Rudeherg, C. (1969) Structure of the parapineal organ of the adult trout, Salmo gairdneri Richardson. Z. Zellforsch. 93:282-304. Sas, E., L. Maler, and M. Weld (1993) Connections of the olfactory bulb in the gymnotiform fish, Apteronotus leptorhynchus. J. Comp. Neurol. 335:486507. Scalia, F.,and S.O.E. Ebbesson (1971) The central projections of the olfactory bulb in a teleost (Gymnothorax funebris). Brain Behav. Evol. 4:37&399. Sheldon, R.E. (1912) The olfactory tracts and centers in teleosts. J. Comp. Neurol. 22: 177-339. Sim, L.J., and S.A. Joseph (1991) Arcuate nucleus projections to brainstem regions which modulate nociception. J. Chem. Neuroanat. 4:97-109. Skagerberg, G., 0. Lindvall, and A. Bjorklund (1984) Origin, course and termination of the mesohabenular dopamine pathway in the rat. Brain Res. 307:99-108. Staines, W.A., T. Yamamoto, K.M. Dewar, P.E. Daddona, J.D. Geiger, and J.I. Nagy (1988) Distribution, morphology and habenular projections of adenosine deaminase-containing neurons in the septal area of rat. Brain Res. 455:72-87. Steinbusch, H.W.M. (1984) Serotonin immunoreactive neurons and their projections in the CNS. In A. Bjorklund, T. Hokfelt, and M.J. Kuhar (eds): Handbook of Chemical Neuroanatomy, Vol. 3, Part 11, Classical Transmitters and Transmitter Receptors in the CNS. Amsterdam: Elsevier, pp. 68-125. Striedter, G.F. (1990) The diencephalon of the channel catfish, Zctalurus punctatus: 11. Retinal, tectal, cerebellar and telencephalic connections. Brain Behav. Evol. 36:355-377. Swanson, L.W. (1982) The projections of the ventral tegmental area and adjacent regions: A combined fluorescent retrograde tracer and immunofluorescence study in the rat. Brain Res. Bull. 9:321-353. Van der Kooy, D., and D.A. Carter (1981) The organization of the efferent projections and striatal afferents of the entopeduncular nucleus. Brain Res. 211:15-36. Van Veen T., P. Ekstrom, L. Nyberg, B. Borg, I. Vigh-Teichmann, and B. Vigh (1984) Serotonin and opsin immunoreactivities in the developing pineal organ of the three-spined stickleback, Gasterosteus aculeutus L. Cell Tissue Res. 273.559-564. Vanegas, H., and S.O.E. Ebbesson (1980) Projections of the teleostean telencephalon. In S.O.E. Ebbesson (ed): Comparative Neurology of the Telencephalon. New York: Plenum Press, pp. 117-127. Veazey, R.B., D.G. Amaral, and W.C. Cowan (1982) The morphology and connections of the posterior hypothalamus in the cynomolgus monkey (Macueafascicularzs).11.Efferent connections. J. Comp. Neurol. 207:135156. Vecino, E., and P . Ekstrom (1992) Colocalization of neuropeptide Y (NPY)like and FMRF-amide-like immunoreactivities in the brain of the Atlantic salmon LSalmo salar). Cell Tissue Res. 27Ot435-442. Vigh-Teichmann, I., B. Vigh, H.W. Korf, and A. Oksche (1983) CSFcontacting and other somatostatin-immunoreactive neurons in the brains of Anguilla anguilla, Phoxinus phoxinus and Salmo gairdneri (Teleostei). Cell Tissue Res. 233t319-334. Villani, L., S. Battistini, R. Bissoli, and A. Contestabile (1987) Cholinergic projections in the telencephalo-habenulo-interpeduncularsystem of the goldfish. Neurosci. Lett. 76:263-268. HABENULAR CONNECTIONS IN TROUT Villani, L., T. Guarnieri, U. Salsi, and D. Bollini (1991) Substance P in the habenulo-interpeduncular system of the goldfish. Brain Res. Bull. 26:225-228. Villani, L., L. Dipietrangelo, C. Pallotti, P. Pettazzoni, I. Zironi, and T. Guarnieri (1994a) Ultrastructural and immunohistochemical study of the telencephalo-habenulo-interpeduncularconnections of the goldfish. Brain Res. Bull. 34r1-5. Villani, L., T. Guarnieri, and I. Zironi (1994b) Choline acetyltransferase and NADPH-diaphorase localization in the goldfish habenulo-interpeduncular system. Neurosci. Lett. 173:67-70. Vincent, S.R., and J.C. Brown (1986) Somatostatin immunoreactivity in the entopeduncular projection to the lateral habenula in the rat. Neurosci. Lett. 68t160-164. von Bartheld, C.S., and D.L. Meyer (1990) Paraventricular organ of the lungfish Protopterus dolloi: Morphology and projections of CSFcontacting neurons. J. Comp. Neurol. 297:410-434. 543 von Bartheld, C.S., D.L. Meyer, E. Fiebig, and S.O.E. Ebbesson (1984) Central connections of the olfactory bulb in the goldfish, Carassius auratus. Cell Tissue Res. 238:475-487. von Bartheld, C.S., D.E. Cunningham, and E.W. Rube1 (1990) Neuronal tracing with DiI: Decalcification, cryosectioning and photoconversion for light and electron microscopic analysis. J. Histochem. Cytochem. 38t725733. Yariez, J., and R. Anad6n (1994) Afferent and efferent connections of the habenula in the larval sea lamprey (Petromyzon marinus L.). An experimental study. J. Comp. Neurol. 345t148-160. Yadez, J., R. Anadon, B.I. Holmqvist, and P. Ekstrom (1993) Neural projections of the pineal organ in the larval lamprey (Petromyzon marinus L.) revealed by indocarbocyanine dye tracing. Neurosci. Lett. 164:213-216. Yariez, J., H. Meissl, and R. Anadon (1996) Central projections of the parapineal organ of the adult rainbow trout (Oncorhynchus mykiss). Cell Tissue Res. (in press).