Afferent and efferent connections of the nucleus sphericus in the snakeThamnophis sirtalis Convergence of olfactory and vomeronasal information in the lateral cortex and the amygdalaкод для вставкиСкачать
THE JOURNAL OF COMPARATIVE NEUROLOGY 385:627–640 (1997) Afferent and Efferent Connections of the Nucleus Sphericus in the Snake Thamnophis sirtalis: Convergence of Olfactory and Vomeronasal Information in the Lateral Cortex and the Amygdala ENRIQUE LANUZA1,2 AND MIMI HALPERN1* of Anatomy and Cell Biology, Health Science Center at Brooklyn, State University of New York, Brooklyn, New York 11203 2Department de Biologia Animal, Unitat de Morfologia Microscòpica, Facultat de Ciències Biològiques. Universitat de València. Burjassot 46100 Spain 1Department ABSTRACT This paper is an account of the afferent and efferent projections of the nucleus sphericus (NS), which is the major secondary vomeronasal structure in the brain of the snake Thamnophis sirtalis. There are four major efferent pathways from the NS: 1) a bilateral projection that courses, surrounding the accessory olfactory tract, and innervates several amygdaloid nuclei (nucleus of the accessory olfactory tract, dorsolateral amygdala, external amygdala, and ventral anterior amygdala), the rostral parts of the dorsal and lateral cortices, and the accessory olfactory bulb; 2) a bilateral projection that courses through the medial forebrain bundle and innervates the olfactostriatum (rostral and ventral striatum); 3) a commissural projection that courses through the anterior commissure and innervates mainly the contralateral NS; and 4) a meager bilateral projection to the lateral hypothalamus. On the other hand, important afferent projections to the NS arise solely in the accessory olfactory bulb, the nucleus of the accessory olfactory tract, and the contralateral NS. This pattern of connections has three important implications: first, the lateral cortex probably integrates olfactory and vomeronasal information. Second, because the NS projection to the hypothalamus is meager and does not reach the ventromedial hypothalamic nucleus, vomeronasal information from the NS is not relayed directly to that nucleus, as previously reported. Finally, a structure located in the rostral and ventral telencephalon, the olfactostriatum, stands as the major tertiary vomeronasal center in the snake brain. These three conclusions change to an important extent our previous picture of how vomeronasal information is processed in the brain of reptiles. J. Comp. Neurol. 385:627–640, 1997. r 1997 Wiley-Liss, Inc. Indexing terms: chemosensory systems; vomeronasal projections; biotinylated dextran amine; reptiles In most tetrapod vertebrates, there are two main sensory systems that are capable of relaying information about the chemicals present in the environment: the olfactory system, which detects volatile chemicals, and the vomeronasal system, which is specialized to detect complex chemicals of high molecular weight (e.g., pheromones). Both chemosensory systems are very well developed in squamate reptiles (lizards and snakes), given the fact that chemical stimuli are very important in the ecology of these groups (for recent reviews, see Mason, 1992; Halpern, 1992). In reptiles, as in all other tetrapods that possess a vomeronasal system, olfactory information r 1997 WILEY-LISS, INC. and vomeronasal information are processed separately in the brain (mammals: Scalia and Winans, 1975; reptiles: Halpern, 1976; amphibians: Scalia et al., 1991). Olfactory Grant sponsor: NIH; Grant number DC00104; Grant sponsor: Valencian IVGI; Grant number: 0021076. *Correspondence to: Dr. Mimi Halpern. Department of Anatomy and Cell Biology, Box 5, Health Science Center at Brooklyn, State University of New York, 450 Clarkson Avenue, Brooklyn, NY 11203. E-mail: firstname.lastname@example.org Received 11 November 1996; Revised 14 April 1997; Accepted 17 April 1997 628 E. LANUZA AND M. HALPERN information enters the brain through the main olfactory bulb (MOB), from where it is relayed to the lateral (pyriform) cortex (LC), which is the main secondary olfactory structure in the brain. On the other hand, vomeronasal information enters the brain through the accessory olfactory bulb (AOB), which projects almost exclusively to the nucleus sphericus (NS; for a review of the olfactory and vomeronasal projections in reptiles, see Lohman and Smeets, 1993). The NS, therefore, is the main secondary vomeronasal structure in the brain of squamate reptiles. In the garter snake (Thamnophis sirtalis), it is the most prominent structure of the telencephalon, occupying almost the entire caudal half of the telencephalic hemisphere. The development of the NS in different reptilian species appears to be strongly correlated with the development of the vomeronasal system in each particular species (Schwenk, 1993). In species without a functional vomeronasal system (some agamid and chamaleonid lizards, e.g., Chamaleo chamaleo), the NS is not recognizable in the brain (Northcutt, 1978); in species with a poorly developed vomeronasal organ and a small accessory olfactory bulb, like some iguanid lizards (e.g., Iguana iguana and Dipsosaurus dorsalis), the NS is small (Northcutt, 1978); and, in species with a well-developed vomeronasal organ and a large accessory olfactory bulb, like most snakes, the NS occupies almost the entire caudal half of the subcortical telencephalon (Halpern, 1980). Therefore, knowledge of the connections of the NS is essential to understand how vomeronasal information is processed in the reptilian brain. Somewhat surprisingly, almost no information is available on the connections of the NS other than its wellknown input from the accessory olfactory bulb (for review, see Lohman and Smeets, 1993), its connection with the contralateral NS (Halpern, 1980), and what is generally believed to be the efferent projection to the ventromedial hypothalamic nucleus (VMH; Voneida and Sligar, 1979; Halpern, 1980; Martı́nez-Garcı́a et al., 1993b). However, recent studies in lizards (Bruce and Neary, 1995a; Lanuza et al., 1997) and our own observations in snakes (Lanuza and Halpern, unpublished observations) strongly suggest that the projection to the VMH actually originates in the posterior part of the dorsal ventricular ridge and the lateral amygdala rather than in the NS. Therefore, vomeronasal information is not directly relayed to the VMH. If the NS does not project to the VMH, then almost nothing is known about its efferent projections, and there are not enough data to suggest a hypothesis about how vomeronasal information is processed in the brain. An important question that the present scheme of connections of the NS leaves unresolved is whether the olfactory and vomeronasal information converge in the brain in later stages of information processing. Both kinds of chemosensory stimuli are important in several crucial behaviors, such as prey detection and mate recognition, as well as in species-typical social behaviors (for reviews, see Burghardt, 1970; Halpern, 1992). Thus, it is reasonable to suggest that olfactory and vomeronasal information should interact in the brain in order to give the animal a complete picture of its chemical environment. To fill this gap in our knowledge of the tertiary connections of the vomeronasal system, we have reinvestigated the afferent and efferent projections of the NS in the garter snake (T. sirtalis) by means of the sensitive tract-tracing technique of intraaxonal transport of the modern tracer biotinylated dextran amine (BDA). Although BDA has been reported to be mainly an anterograde tracer (Veenman et al., 1992), in our experimental conditions, it also gives rise to consistent retrograde transport. MATERIALS AND METHODS Adult specimens of T. sirtalis (both sexes), with body weights between 18 g and 75 g, were purchased from a licensed dealer. Animals were maintained in terraria under a natural day/night cycle at 22–30°C, given water ad libitum, and fed weekly. For surgery, animals were cooled prior to anesthetization with an intramuscular injection of 0.003 ml of 0.5% brevital sodium (methohexital sodium; Eli Lilly and Co., Indianapolis, IN) per gram of body weight. Iontophoretic injections of BDA (MW 10,000; Molecular Probes, Eugene, OR) dissolved 10% in Tris buffer (0.05 M), pH 7.6 (Veenman et al., 1992), were made in the nucleus sphericus (n 5 8). For controls, some injections were made in the medial (n 5 2), dorsal (n 5 3), and lateral cortices (n 5 5). Abbreviations ac ADVR AOB aot apc BNST ch CoApm DBB DC DLA DMA dRF ExA gl gr H LA LC lfb LHA LPA anterior commissure anterior dorsal ventricular ridge accessory olfactory bulb accessory olfactory tract anterior pallial commissure bed nucleus of the stria terminalis corticohypothalamic tract posteromedial cortical nucleus of the amygdala nucleus of the diagonal band of Broca dorsal cortex dorsolateral amygdaloid nucleus dorsomedial anterior thalamic nucleus dorsal retrobulbar formation external amygdala glomerular layer of the AOB granular layer of the AOB habenula lateral amygdala lateral cortex lateral forebrain bundle lateral hypothalamic area lateral preoptic area LTM m MA MC mfb MOB NAcc Naot NS OS OT ot PDVR S sh Si St v VAA VMH VPA lateral tuberomammillary area mitral cell layer of the AOB medial amygdala medial cortex medial forebrain bundle main olfactory bulb nucleus accumbens nucleus of the accessory olfactory tract nucleus sphericus olfactostriatum olfactory tubercle optic tract posterior dorsal ventricular ridge septum septohypothalamic tract nucleus septalis impar striatum ventricle ventral anterior amygdala ventromedial hypothalamic nucleus ventral posterior amygdala TERTIARY VOMERONASAL CONNECTIONS IN SNAKES 629 Fig. 1. Camera lucida drawings showing the location of the injection sites in the eight specimens with injections in the nucleus sphericus (NS). The sections are arranged from rostral (A) to caudal (H). Hatched areas indicate the extent of the injections sites. The fiber labeling adjacent to the injection site is also depicted to give an idea of the area of diffusion of the tracer. A: Specimen 9640. B: Specimen 9620. C: Specimen 9637. D: Specimen 9616. E: Specimen 9636. F: Specimen 9610. G: Specimen 9621. H: Specimen 9607. For abbreviations, see list. Scale bar 5 500 µm. The injections were made by using glass micropipettes with inner tip diameters of 20–40 µm. A positive DC current of 3–5 µA was applied during 5–15 minutes (pulse 7 seconds on/off). After a survival period ranging from 8 to 12 days, animals were deeply anesthetized with brevital sodium and transcardially perfused with 30–40 ml of saline solution (0.9% NaCl), followed by 60–80 ml of fixative (4% paraformaldehyde) in 0.1 M phosphate buffer (PB), pH 7.4. Brains were removed from the skull, postfixed in the same fixative for 4–12 hours at 4°C, cryoprotected in 30% sucrose in PB at 4°C until they sank, and cut at 35 µm in the frontal plane on a freezing microtome. The sections were collected in three matching series. For BDA histochemical detection, endogenous peroxidase activity was first inhibited by incubating the sections for 30 minutes in a 0.3% H2O2 solution in saline Tris buffer (0.05 M, pH 7.6, 0.9% NaCl). Sections were then incubated in avidin-biotin complex, usually one series in ‘‘elite’’ ABC (ABC Elite Vectastain kit; Vector Laboratories, Burlingame, CA) and another in ‘‘standard’’ ABC (ABC Standard Vectastain kit; Vector Laboratories), and were diluted in Tris-buffered saline with 0.3% Triton X-100 as a permeating agent for 1–2 hours at room temperature. The reason for using the two types of ABC kits was that the elite ABC gave an amplified signal that made it difficult to determine the limits of the injection site and also yielded considerable background, even with very short developing times. In contrast, the standard ABC gave a more precise picture of the extent of the injection site and yielded much less background, with the disadvantage of being somewhat less sensitive. The resulting peroxidase label was revealed directly by using 3,38-diaminobenzidine (DAB; Sigma, St. Louis, MO) as the chromogen diluted at 0.02% in Tris buffer, pH 8.0, with nickel salts (0.4%) and H2O2 (0.01%). Sections were subsequently counterstained with acidic toluidine blue. The research reported herein was performed according to the guidelines of the Animal Care and Utilization Act, under the supervision of the SUNY Health Science Center Institutional Animal Committee. RESULTS Injections into the NS We made injections into different parts of the dorsoventral, rostrocaudal, and mediolateral extent of the NS (Fig. 1). All of the injections into the NS gave rise to a similar pattern of anterograde labeling. The anterograde labeling found in specimen 9637 (Fig. 1C) is described and illustrated (Fig. 2) as an example. To some extent, all of the injections into the NS also affected the overlying medial cortex (MC) or dorsal cortex (DC), depending on the rostrocaudal location of the injection (Fig. 1). Therefore, some control injections into the cortex were made to distinguish the labeling due to the cortical neurons affected by the injection. In some injec- 630 E. LANUZA AND M. HALPERN tions into the NS, only a number of labeled fibers and a few labeled neurons appeared in the cortex (Fig. 1B–E), which facilitated the recognition of the labeling due to NS neurons. In any case, all of the observed labeling is described in Results, and the fiber labeling that was probably due to the affected portion of the cortex is subtracted later (see Discussion). In the present work, we have followed the nomenclature of Halpern (1980) for the telencephalon of snakes, with the exception of the amygdaloid complex, for which the nomenclature of Lanuza and Halpern (1997) has been used. Anterograde transport Intratelencephalic labeling. Following injections centered in the NS, anterogradely labeled fibers leave the NS through three main pathways. The most important projection courses to surround the accessory olfactory tract (aot), the second projection courses via the medial forebrain bundle (mfb), and the third projection courses via the anterior commissure (ac). The projection that courses along the external border of the aot gives rise to abundant terminal-like labeling in the nucleus of the accessory olfactory tract (Naot) throughout its rostrocaudal extent, especially dorsal and medial to the tract (Figs. 2C–E, 4B,E). Most of the labeled fibers avoid the center of the tract, leaving it relatively free of fibers (Figs. 2D, 4B,E). These fibers continue rostrally to retrobulbar levels, where they surround the olfactory ventricle dorsally (Fig. 2B) and give rise to a dense plexus of terminal-like labeling in the granular layer of the accessory olfactory bulb (AOB; Figs. 2A,B, 4A). In the rostral half of the telencephalon, from the retrobulbar formation to the anterior commissure, many labeled fibers leave the aot in a dorsal and lateral direction (Fig. 2C–G). At rostral levels, these fibers reach the dorsal retrobulbar formation (dRF), where they give rise to a dense plexus of fiber and terminal-like labeling (Fig. 2C). At the same rostral levels, other fibers leave the aot in a dorsolateral direction, giving rise to a light plexus of fibers in the lateral cortex (LC) (Fig. 2C). More caudally, labeled fibers innervate the outer half of the external plexiform layer of the DC (Figs. 2D–F, 4D) and the internal plexiform and cellular layers of the LC (Fig. 2E, F, 4C). The fiber labeling is denser at rostral levels (level of the retrobulbar formation; Fig. 2C) and becomes lighter caudally. At precommissural levels (Fig. 2F,G), the labeled fibers that leave the aot in a dorsolateral direction give rise to dense, terminal-like labeling in several amygdalar nuclei: the dorsolateral amygdala (DLA), external amygdala (ExA), and ventral anterior amygdala (VAA; Figs. 2F,G, 4B). Some labeled fibers also appear in the caudal lateral amygdala (LA), mainly at the level of the injection site (Fig. 2H–J). At caudal telencephalic levels, the hilus and the mural layer of the injected NS contain a considerable amount of anterograde labeling (Fig. 2J,K). The set of projections that courses via the aot, which will be described later, is bilateral, with ipsilateral predominance. The commissural fibers cross the midline via the anterior commissure. The fibers that course through the mfb exit the NS medially and appear to innervate the bed nucleus of the stria terminalis (BNST) and the medial amygdala (MA; Fig. 2H). More medially, these fibers gather together above the lateralmost part of the ac (Fig. 2H). Then, they travel rostrally through the mfb (Fig. 2F,G), traversing the ventral telencephalon to finally give rise to a dense field of terminal-like labeling in the olfactostriatum (Figs. 2C–E, 4E; terminology of Halpern, 1980; following Durward, 1930). This area is limited dorsally by the rostralmost part of the nucleus accumbens (as identified by Smeets, 1988), laterally by the Naot, ventrally by the olfactory tubercle (OT) and the horizontal limb of the nucleus of the diagonal band of Broca (DBB), and medially by the vertical limb of the DBB. At the level of the retrobulbar formation, the labeled fibers continue rostrally just medial to the olfactory ventricle, until they reach the granular layer of the AOB. Therefore, the NS appears to reach the AOB through two different projections, a lateral one via the aot and a medial one via the mfb. The fibers that course through the mfb do not form a compact tract but, rather, are loosely arranged. Some of them enter the vertical limb of the DBB and the nucleus accumbens (NAcc; Fig. 2F). They exhibit varicosities along their way through the ventral telencephalon that may represent boutons en passant mainly on neurons of the bed nucleus of the mfb (Fig. 2F). This projection is also bilateral with ipsilateral predominance. The third intratelencephalic projection is a commissural projection that courses via the ac (Fig. 2I). After crossing the commissure, a few fibers enter the contralateral mfb and travel rostrally, innervating the olfactostriatum (Fig. 2D–F). The majority of the commissural fibers continue laterally and give rise to terminal-like labeling mainly in the external, mural, and submural layer of the contralateral NS (Fig. 2J,K), always in a position symmetric to the site of injection. In the case illustrated in Figure 2, the injection site is rostral and dorsal in the NS, and the resulting fiber labeling in the contralateral NS is most abundant in the dorsal and rostral areas (Fig. 2I–L). The fibers destined for the lateral part of the contralateral NS tend to course over the dorsal NS instead of crossing through it. Numbers of labeled fibers continue rostrally, innervating the Naot (Fig. 2F–H) and also the same amygdalar nuclei that receive a NS projection on the ipsilateral side: the DLA (Fig. 2H), the VAA, and the ExA (Fig. 2F,G). Whereas the contralateral innervation of the ExA and the VAA is very light, the innervation of the contralateral DLA is remarkable. More rostrally, labeled fibers are observed in the contralateral rostral LC (Fig. 2E) and in the granular layer of the contralateral AOB (Fig. 2A,B). Therefore, all of the intratelencephalic efferent projections of the NS are bilateral, although they have a strong ipsilateral predominance. In addition to the already described labeling, anterogradely labeled fibers and terminals appeared in the medial and dorsal cortices at the level of the injection site (Fig 2H–J) and in the septum at most of it rostrocaudal extent, but especially caudally (Fig. 2H,I). Some of the labeled fibers in the septum cross the midline through the anterior pallial commissure (apc; Fig. 2I), reaching the contralateral septum and also the contralateral dorsomedial cortex (Fig. 2J,K). This labeling was very light in the injections in which only a few cortical neurons were affected. Extratelencephalic labeling. All of the anterograde labeling found in extratelencephalic areas was located in the hypothalamus. Following injections centered in the NS, a small bundle of fibers leaves the NS in the ventromedial direction, surrounds the lfb dorsally (Fig. 2H), and turns ventrally, crossing the lateral preoptic area (LPA; TERTIARY VOMERONASAL CONNECTIONS IN SNAKES Fig. 2. A–L: Camera lucida drawings of transverse sections through the brain of a snake showing the distribution of biotinylated dextran amine (BDA)-labeled fibers and neurons (solid circles) after an injection centered into the nucleus sphericus (NS). Injection site depicted in 631 H and I (hatched area). Arrowhead in H indicates the bundle of fibers destined for the hypothalamus. A is rostral, L is caudal. For abbreviations, see list. Scale bar 5 200 µm. 632 E. LANUZA AND M. HALPERN Fig. 2I). Some scattered fibers are observed to reach the medial preoptic area (Fig. 2H,I). During its dorsoventral course through the LPA (Fig. 2I), this bundle of fibers splits apart and is no longer visible as a unit but, rather, as solitary fibers travelling in a ventral direction located lateral to the lfb (Fig. 2J, 4F). They traverse the lateral hypothalamic area (LHA; at the level of the anterior hypothalamus; Fig. 2J) and, apparently, terminate in a cell-poor area located just lateral to the ventromedial hypothalamic nucleus (VMH; Fig. 2K). This area extends caudally to the level of the mammillary bodies (Fig. 2L). Because it is the lateral area of the tuberal part of the hypothalamus, and it reaches the level of the mammillary bodies, we have called it the lateral tuberomammillary area (LTM). Most of the fibers that course through the lateral hypothalamus exhibit varicosities along their route that may represent boutons en passant on cells of the LHA. In the contralateral hypothalamus, some (although very few) fibers can be observed following the same pathway as that described on the ipsilateral side. In addition to the described labeling, some of the fibers that cross the septum continue farther ventrally and reach the hypothalamus (Fig. 2H,I), where they course both medially and laterally and, apparently, terminate in the periventricular area of the anterior and tuberal hypothalamus and in the LHA (Fig. 2J,K). Retrograde transport Following injections centered in the NS, the majority of the retrogradely labeled neurons were found in the mitral cell layer of the AOB (Fig. 2A,B) and in the Naot, mainly in its medial part (Figs. 2C–F, 4H). In addition, retrogradely labeled cells appeared scattered in the nucleus of the diagonal band (Fig. 2G), in the vicinity of the lateral forebrain bundle (Fig. 2G), and, more caudally, in the nucleus of the ac (Fig. 2H). Retrogradely labeled neurons also appeared in the DC and the MC. Control injections into the DC We describe here only the efferent projections of the DC that coincided with those described in the injection centered in the NS. Following injections into the DC (Fig. 3B), in the telencephalon, numerous labeled fibers appeared in the septum, the DBB, the DLA, and the LA (Fig. 3A,B). In addition, a few anterogradely labeled fibers appeared in the granular layer of the accessory olfactory bulb, the rostral LC, and the nucleus accumbens. In the diencephalon, anterogradely labeled fibers reach the hypothalamus via two pathways, one medial and one lateral. The medial pathway, which crosses through the septum, is actually formed by a rostral component, the precommissural fornix (Lohman and Van Woerden-Verkley, 1972, 1976), called septohypothalamic tract (sh) by Hoogland and Vermulen-Vanderzee (1989; Fig. 3A), and a caudal component, the postcommissural fornix (Lohman and Van Woerden-Verkley, 1972, 1976), called corticohypothalamic tract (ch) by Hoogland and Vermulen-Vanderzee (1989; Fig. 3C). The lateral pathway is formed by fibers that curve around the lateral corner of the lateral ventricle and course medially through the amygdaloid complex (Fig. 3A). The precommissural fornix and the lateral tract join in the lateral preoptic area (Fig. 3B) and innervate the lateral hypothalamus (lateral preoptic area: Fig. 3A,B; lateral hypothalamic area: Fig. 3C,D). The postcommissural fornix enters the hypothalamus more caudally (Fig. 3C) and innervates the periventricular and medial hypothalamus at anterior and tuberal levels (Fig. 3C,D). Injections into the LC We wanted to confirm the projection of the NS to the LC by retrograde transport after injections into the LC. Somewhat surprisingly, retrogradely labeled cells after injections in the LC consistently appeared only in the lateral wall of the mural layer of the NS and not in the medial layer (Fig. 4G). The injections into the LC were located rostrally (at the levels of Fig. 2E–G); therefore, we can be sure that the retrograde labeling in the NS was not due to medial diffusion of the tracer. DISCUSSION The results of the present work reveal (after subtraction of the projection from the cortex; see below) that the nucleus sphericus of the snake T. sirtalis has four major efferent projections: 1) a projection to the accessory olfactory bulb, rostral DC, rostral LC, and several nuclei of the amygdaloid complex that courses along the external border of the accessory olfactory tract; 2) a projection to the olfactostriatum that courses via the mfb; 3) a projection to the contralateral hemisphere that reaches the contralateral NS plus all of the targets of the ipsilateral projections through the aot (AOB, Naot, DC, LC, and amygdaloid nuclei) and through the mfb (OS). This projection crosses the midline through the anterior commissure; and 4) a projection to the hypothalamus that courses as a bundle of fibers medial to the lfb. On the other hand, all of the NS inputs arise in other vomeronasal structures: in the AOB (the major input), the Naot, and the contralateral NS. It may also receive a light projection from the DBB. This pattern of afferent and efferent projections of the NS, as will be discussed later, suggests a new scheme for the processing of vomeronasal information in the snake brain. This is particularly relevant, because snakes are probably the animals with the best developed vomeronasal system, and, because of this, they have been used extensively as models for the study of the structure, function, and behavioral roles of this particular sensory system (Halpern, 1992; Halpern and Holtzman, 1993). Before discussing other aspects of the results, a comparison of the projections of the NS and the DC is necessary, because the injections into the NS always affected part of the overlying cortex. Subtraction of DC projections It was easy to differentiate the telencehalic projections of the DC from those that originated in the NS. However, it was somewhat more difficult to distinguish between the corticohypothalamic projections and the projections to the hypothalamus that originated in the NS. The clearest case of anterograde labeling found after injections aimed at the NS that actually corresponded to the cortical areas affected is the case of the anterograde labeling found in the septum. Both DC and MC, as described previously (for review, see Ulinski, 1990), project to the septum, and the anterograde labeling that we found in the septum following our injections into the NS is exactly the same as the labeling found in the septum after cortical injections (see Figs. 2H,I, 3B,C). Therefore, the Fig. 3. A–D: Camera lucida drawings of transverse sections through the brain showing the distribution of BDA-labeled fibers after an injection restricted to the dorsal cortex (DC). Injection site is depicted in B (hatched area). For conventions, see Figure 2. A is rostral, D is caudal. For abbreviations, see list. Scale bar 5 200 µm. 634 Fig. 4. Photomicrographs illustrating the anterograde (A–F) and retrograde (H) labeling following BDA injections in the NS and retrograde labeling after an injection into the lateral cortex (G). A: Anterograde labeling in the granular layer of the accessory olfactory bulb (AOB). The position of the anterograde labeling is marked by arrowheads. B: Detail of the innervation of the nucleus of the accessory olfactory tract (Naot), the external amygdala (ExA), and the ventral anterior amygdala (VAA). C: High-magnification photomicrograph showing the plexus of beaded fibers that innervate the lateral cortex (LC). Arrowheads indicate some of the varicosities of the fibers. E. LANUZA AND M. HALPERN D: Detail of the innervation of the external plexiform layer of the rostral DC. The position of the anterograde labeling is indicated by arrowheads. E: Heavy anterograde labeling in the olfactostriatum (OS). F: Scattered fibers in the lateral hypothalamic area (LHA). G: Retrogradely labeled neurons in the lateral mural layer of the NS after an injection into the LC. H: Retrogradely labeled neuron (arrowhead) in the Naot after an injection into the NS. For abbreviations, see list. Scale bars 5 200 µm in A (also applies to B,D,E,G), 40 µm in C (also applies to F,H). TERTIARY VOMERONASAL CONNECTIONS IN SNAKES fiber labeling found in the septum after the injections into the NS that affected the overlying cortex should be considered part of the efferent projections of the cortex. Similarly, the fiber labeling found in the DC and the MC at the level of the injection site (Fig. 2H,I) and more caudally (Fig. 2J,K), as well as the few fibers found in the contralateral DC and MC (Fig. 2J,K) very probably arise from the cortex overlying the NS and not from the NS itself. In addition, the following structures showed anterograde labeled fibers both after injections into the NS and after injections into the DC: the AOB, LC, rostral DC, DLA, LA, DBB, and the nucleus accumbens in the ventral telencephalon. The projection from the NS to the AOB is very clear (Fig. 4A). It was described previously in the snake T. sirtalis (Halpern, 1980) and in the lizards Tupinambis nigropunctatus (Voneida and Sligar, 1979) and Podarcis hispanica (Martı́nez-Garcı́a et al., 1991). The main difference between the projection to the AOB that originates in the NS and the one that originates in the DC is that the former is relatively massive, whereas the latter is very meager. Both terminate in the granular layer of the AOB, and the appearance of the centrifugal fibers originating from both the NS and the DC is indistinguishable. The anterograde labeling found in the LC and in rostral DC following injections into the NS clearly originated in the NS. Even after very large injections into the cortex (affecting most of the caudal part of both the DC and MC), the fiber labeling found in the LC and in rostral DC is sparse. Moreover, after injections into the NS, the course of the fibers reaching the rostral DC and LC can be followed from the aot to these cortical structures (Fig. 2C–F). Furthermore, injections into the LC give rise to retrogradely labeled cells in the NS, thus confirming the projection from the NS to the LC. The DLA and LA are both densely innervated by the DC (Fig. 3). The fiber labeling in the LA found after injections into the NS is quite sparse and is limited mainly to the levels of the injection site (Fig. 2H–J). Therefore, we believe that there is no projection from the NS to the LA, and it is more likely that the fiber labeling found in the LA is due to the projection from the DC as well as to some diffusion of the tracer from the injection site. Moreover, no contralateral projection to the LA has been observed in contrast to the rest of the amygdaloid nuclei innervated by the NS, which are all bilaterally innervated. On the other hand, the DLA is probably densely innervated by both the DC and the NS. Anterograde labeling in the DLA after injections into the NS is bilateral and is also found in animals in which the injections included only the MC and the NS. The MC has not been found to project to the DLA (Olucha et al., 1988; Ulinski, 1990; Hoogland and Vermeulen-Vanderzee, 1993). Moreover, one injection centered in the DLA (which also affected the caudal DC) gave rise to retrogradely labeled somata in the NS (not charted). Regarding the projection to the ventral telencephalon, the few labeled fibers that appear in the NAcc following injections in the NS are probably due to cortical neurons affected by the injection (even in the animals with large injections into the NS, only a very few fibers appear in the NAcc, whereas, in animals with large injections into the DC, many labeled fibers appear in the NAcc). It is more difficult to distinguish the projection of the DC and the NS to the DBB. Large injections into the DC result in a heavy projection to the DBB, whereas large injections into the NS give rise to only a relatively small number of fibers in the 635 DBB. Therefore, if there is a projection from the NS to the DBB, then it is a light one. It is difficult to distinguish between the corticohypothalamic projections and the projections to the hypothalamus that originate in the NS. The corticohypothalamic projections of the DC are much more extensive than those of the NS and, apparently, also innervate most of the areas reached by the NS projection. In agreement with the results of Hoogland and Vermeulen-Vanderzee (1989) in the lizard Gekko gecko, we have found that the DC projects to the hypothalamus via two pathways: a medial pathway that courses through the septum (actually formed by the septohypothalamic tract and the medial corticohypothalamic tract; see Fig. 3) and a lateral pathway that courses through the amygdaloid complex, between the Naot and the anterior dorsal ventricular ridge (ADVR; Fig. 3A). The MC has been reported to project to the medial hypothalamus via the medial corticohypothalamic tract (Halpern, 1980), although this projection has not been found in lizards (Olucha et al., 1988; Ulinski, 1990; Hoogland and Vermeulen-Vanderzee, 1993). The projection through the medial corticohypothalamic tract terminates mainly in the periventricular areas of the anterior and tuberal hypothalamus. In our opinion, the labeled fibers that we find in the periventricular hypothalamus following injections centered in the NS are probably due to this corticohypothalamic projection. In fact, some of these fibers could be followed back to the septum. The other DC projections to the hypothalamus (through the septohypothalamic tract and through the lateral pathway) follow a similar course within the hypothalamus to one of the fibers that leaves the NS: All three projections enter the hypothalamus rostrally, innervating the lateral preoptic area, and extend ventrally, innervating the lateral hypothalamic area of the anterior and tuberal hypothalamus, including the LTM. Accordingly, the DC projects to the LHA and to the area located lateral to the VMH in the lizard G. gecko (Hoogland and Vermeulen-Vanderzee, 1989). However, anterogradely labeled fibers in the LHA and the LTM appeared in all of the animals with injections into the NS, including those in which only the NS and the MC were affected (see Fig. 1D,E,H). Therefore, part of the innervation of the lateral hypothalamus (LPA, LHA, and LTM) probably arises in the NS. Intratelencephalic projections of the NS The major intratelencephalic projections of the NS reach the AOB, the Naot, and the olfactostriatum (OS). These structures appear to form a heavily interconnected circuit for the processing of the vomeronasal information. The connections between the AOB and the NS are massive and reciprocal (Halpern, 1980; Lanuza and Halpern, 1997), as are the connections between the Naot and the NS (this report) and between the Naot and the AOB (Lanuza and Halpern, 1997). The projection from the NS to the OS is also massive, although it is not reciprocal. The OS, in turn, projects to the AOB (Lanuza and Halpern, 1997). Therefore, several possibilities exist for feed-back regulation of vomeronasal input, and it seems that the vomeronasal information is highly processed within this circuitry before it is integrated with other sensory modalities (e.g., in the LC and the amygdala). A projection from the NS to the NAcc is reported but is not illustrated in Martı́nez-Garcı́a et al. (1993b). However, in agreement with our data, González et al. (1990) did not 636 find retrogradely labeled cells in the NS after Fluoro-Gold injections into the NAcc of the lizard G. gecko. In the work of Voneida and Sligar (1979), the projections from the NS are described based on an injection clearly centered in the DLA (see their Fig. 4), rostral and dorsal to the NS. Anterograde labeling after this injection is located in the caudal NAcc, that is, caudal to the olfactostriatum (Voneida and Sligar, 1979). Therefore, we believe that the NS in lizards probably also projects to the olfactostriatum (medial and ventral to the NAcc) and not to the NAcc itself. The basal forebrain in snakes is a cytoarchitectonically complex area, and its connections and neurochemistry are still poorly studied. The olfactostriatum, as a result, has been considered to be part of the striatum, without specifically naming it, by most authors (Ulinski, 1978; Smeets et al., 1986a). We have chosen to keep the name ‘‘olfactostriatum,’’ because it was already present in the literature (Durward, 1930; Halpern, 1980). However, a study of the distribution of dopamine in the brain of another snake (Python regius) suggests that the OS may be part of a complex dopaminergic area that is probably equivalent to the NAcc/striatum of other squamate reptiles (Smeets, 1988). Neurophysiological studies in the green iguana (I. iguana) demonstrated that electrical stimulation of neurons in the olfactostriatum elicited strong tongue-flicking responses (Distel, 1978). The same results were obtained following stimulation of the NS. However, these data should be interpreted with caution, because stimulation of the NAcc also elicited tongue-flicking responses. It is known that garter snakes respond to the presence of significant chemical stimuli in the environment (e.g., prey extracts) with an increase in tongue-flick rate (Halpern and Kubie, 1983). The projection from the NS to the olfactostriatum is probably implicated in this behavior, which has been extensively used as a way to monitor vomeronasal stimulation in behavioral tasks (for review, see Halpern, 1992). On the other hand, tongue-flicking activity is observed in a wide variety of behaviors; therefore, it is not surprising that it can be elicited by stimulating other nuclei, such as the NAcc, that are not directly implicated in the vomeronasal system. The other efferent intratelencephalic projections of the NS reach the rostral DC, the rostral LC, and several nuclei of the amygdaloid formation: the ExA, VAA, and DLA. The projections to the olfactory areas (LC, ExA, and VAA) will be discussed below in the section about the convergence of olfactory and vomeronasal information. The NS projection to the DLA is strong and bilateral (Fig. 2G,H). It has also been reported in the lizard P. hispanica (Martı́nez-Garcı́a et al., 1993b). In both lizards and snakes, the DLA does not receive (directly) either olfactory or vomeronasal input (Martı́nez-Garcı́a et al., 1991; Lohman and Smeets, 1993; Lanuza and Halpern, 1997), and, in P. hispanica, it has been shown to project bilaterally to the striatum, including the NAcc, and to receive a putative cholinergic afferent from the basal forebrain (Martı́nez-Garcı́a et al., 1993b). On the basis of these projections plus the strong acetylcholinesterase reactivity of its neuropil, the DLA has been compared to the mammalian basolateral nucleus of the amygdala (Martı́nez-Garcı́a et al., 1993b; for a review of the amygdaloid projections in mammals, see Price et al., 1987). In mammals, this nucleus has been implicated in stimulus-reward associations, especially via its projections to the NAcc of the ventral striatum (Everitt and Robbins, 1992). The DLA of T. sirtalis also appears to project to the E. LANUZA AND M. HALPERN NAcc (Lanuza and Halpern, unpublished observations). Moreover, vomeronasal stimuli in snakes are known to have an important reinforcing value, so that they can be used as unconditioned or reinforcing stimuli to condition a response to an arbitrary stimulus (Halpern, 1988). Therefore, it seems likely that the projection from the NS to the DLA is providing vomeronasal information to the amygdalostriatal circuit, which is probably implicated in mediating the reinforcing value of vomeronasal stimuli. The projection to the rostral part of the DC (Fig. 4D) has not been shown in other reptiles, but its presence was suggested in the lizard P. hispanica (Martı́nez-Garcı́a et al., 1986; Martı́nez-Garcı́a and Olucha, 1988). In the lizard G. gecko, the projection from the LC to the MC has been suggested to give rise to boutons en passant in the DC (Hoogland and Vermeulen-Vanderzee, 1995). If this is the case, then the rostral DC could be integrating olfactory and vomeronasal information that is received directly by the secondary olfactory and vomeronasal centers, respectively. No other sensory input to the DC is known in any reptile, with the notable exception of the geniculocortical projection in turtles (for review, see Ulinski, 1990). These cases seem to be examples of divergent evolution in different reptilian orders. The projection from the NS to the DC may be an important pathway for vomeronasal information to reach the hypothalamus, because the NS does not project to the VMH, and the projection to other hypothalamic areas (LHA and LTM) is sparse (see Results). The DC has strong projections to both the periventricular and the lateral hypothalamus (see Results). Furthermore, in the lizard G. gecko, most of the projection from the DC to the hypothalamus originates in the rostral part of the DC (Hoogland and Vermeulen-Vanderzee, 1989). Alternative pathways through which vomeronasal information may reach the hypothalamus will be discussed below. On the other hand, the DC receives an important input from the dorsolateral anterior thalamic nucleus (Bruce and Butler, 1984) that is probably multimodal that does not carry chemosensory information (Hoogland, 1981, 1982; Martı́nez-Garcı́a and Lorente, 1990; Martı́nez-Garcı́a et al., 1993a; Kenigfest et al., 1997). The DC also receives other inputs from the hypothalamus and brainstem (Bruce and Butler, 1984). Therefore, the DC should not be considered to be a tertiary area in the vomeronasal system but, rather, a limbic cortex (Hoogland and Vermeulen-Vanderzee, 1989) or a ‘‘general’’ cortex (nonhippocampal, nonolfactory; Bruce and Neary, 1995b) that receives a tertiary vomeronasal input from the NS and a multimodal input from the dorsal thalamus. Descending projection to the hypothalamus The main discrepancy between this study and the previous reports (Voneida and Sligar, 1979; Halpern, 1980; Martı́nez-Garcı́a et al., 1993b) is that we have not found a projection from the NS to the VMH but, instead, a meager projection to the LHA and to the LTM, a cell-poor area located lateral to the VMH. Voneida and Sligar (1979) and Martı́nez-Garcı́a et al. (1993b) reported that the NS projected to the shell (the outer rim) of the VMH, whereas Halpern (1980) found that the projection reached the whole VMH. The projection to the VMH actually originates in the structures located medial and lateral to the NS, the posterior part of the dorsal ventricular ridge (PDVR; medially adjacent to the NS) and the lateral amygdala (laterally adjacent to the NS). This has been demonstrated recently in the lizards G. gecko (Bruce and Neary, 1995a) TERTIARY VOMERONASAL CONNECTIONS IN SNAKES and P. hispanica (Lanuza et al., 1997), and we have confirmed it in our own laboratory in the snake T. sirtalis (Lanuza and Halpern, unpublished observations). It can also be observed in the work of Voneida and Sligar (1979), in which the injection sites located in the DVR (anterior or posterior) revealed a projection to the core of the VMH (see their Figs. 2 and 3), whereas the projection to the shell of the VMH was found after an injection that, in our opinion, affected the DLA and the LA (see their Fig. 4) more than the NS. The PDVR and the LA were very likely affected by the lesions of Halpern (1980; see her Fig. 7) and by the horseradish peroxidase injections of Martı́nez-Garcı́a et al., 1993b; F. Martı́nez-Garcı́a, personal communication; see also their Fig. 2). Therefore, the main efferent projection of the NS is not to the hypothalamus, as was previously believed, but, rather, to the Naot, the amygdala (DLA, ExA, and VAA), rostral DC and LC, and the olfactostriatum (see Results and Fig. 2). The projection to the hypothalamus is actually quite meager, and this fact changes to an important extent our previous conception of how vomeronasal information is processed in the brain of squamate reptiles. If the vomeronasal information was directly relayed to the VMH, then the NS would probably serve mainly as a simple relay, and the VMH nucleus would be dominated very strongly by vomeronasal input. This appeared to be a logical conclusion, because vomeronasal information is very important in the ecology of most squamate reptiles. However, the telencephalic projection to the VMH appears to originate to an important extent from two structures that receive direct input from neither the main nor the accessory olfactory bulbs: the PDVR and the LA; therefore, the descending projection to the VMH from these nuclei does not appear to carry any direct chemosensory information. There are several pathways through which vomeronasal information reaches the hypothalamus. One possibility, as discussed above, is through the projection from the NS to the DC, which, in turn, projects to the periventricular and lateral hypothalamus. A more direct pathway would be via the projections of the Naot and medial amygdala (MA) to the hypothalamus. These projections are not well studied, but retrogradely labeled neurons appear in the Naot and MA after injections into the VMH in the lizards G. gecko (Bruce and Neary, 1995a) and P. hispanica (Lanuza et al., 1997). The Naot receives a strong projection from the AOB (Lanuza and Halpern, 1997) and from the NS (this study); therefore, its neurons are strongly driven by vomeronasal stimuli. The MA receives a light projection from the AOB (Lanuza and Halpern, 1997) and the NS (this study), but it projects massively to the VMH (Bruce and Neary, 1995a; Lanuza et al., 1997), therefore influencing strongly the VMH activity. The BNST also receives a projection from the NS, and it may project to the hypothalamus, although the projections of this nucleus are still unexplored in reptiles. Another possible pathway for vomeronasal information to reach the hypothalamus is through the projection from the NS to the LC. The ventral part of the LC projects to the PDVR in T. sirtalis (Lanuza and Halpern, unpublished observations) and in the lizard G. gecko (Hoogland and Vermeulen-Vanderzee, 1995). Although the NS projects mainly to the dorsal part of the LC, this structure may be relaying integrated olfactory and vomeronasal information to the PDVR. The PDVR of lizards has been shown to receive a convergent projection from the different sensory areas of the ADVR (Font et al., 1995; Andreu et al., 1996); 637 therefore, it would be receiving visual, somatosensory, and auditory information from the ADVR and olfactory and vomeronasal information from the LC. Therefore, two putative multimodal projections reach the hypothalamus, both of which may carry integrated vomeronasal information: the projection from the DC and the projection from the PDVR. We want to point out that the two projections terminate in different hypothalamic areas in a complementary fashion. The DC innervates mainly the periventricular and lateral hypothalamus, whereas the PDVR innervates mainly the VMH. The projection from the NS to the LHA and to the LTM appears to be quite sparse (see Fig. 2). However, we should keep in mind that the NS in T. sirtalis is a very large structure, and our tracer injections, although they were relatively large, always affected only a fraction of the cells that form the mural layer of the entire NS (Fig. 1). Therefore, it is possible that the whole projection is more important than it appears. Convergence of olfactory and vomeronasal information in the rostral LC and olfactory amygdala It is well known that the olfactory projection in reptiles terminates in the external half of the external plexiform layer of the LC (for review, see Lohman and Smeets, 1993). The bilateral vomeronasal projection that originates in the NS terminates mainly in the internal plexiform layer and in the cell layer of the LC. Therefore, there is not a direct overlap between the two chemosensory projections. However, it is likely that at least some neurons of the LC integrate both types of chemosensory information. The morphology of the neurons of the LC in snakes has been studied by Ulinski and Rainey (1980), and there are two main neuronal types in the rostral part of the LC, ‘‘bowl’’ neurons and double pyramidal neurons. The bowl neurons, so called because their dendritic trees curve toward the pial surface in a configuration that resembles a bowl, have their somata located in the cell layer of the LC. It has been demonstrated that they receive olfactory input on their distal dendrites in the external plexiform layer (Ulinski and Rainey, 1980). Many (probably most) of these neurons project their axons to the MC. The other type of cell, the double pyramidal neuron, has a somata located either in the inner part of the cell layer or in the internal plexiform layer of the LC. The dendrites of the double pyramidal neurons originate from both poles of the somata, forming pyramidal dendritic fields located in both plexiform layers of the LC. The dendritic fields of these neurons span the whole depth of the LC (see Fig. 12 in Ulinski and Rainey, 1980). We retrogradely labeled some of these double pyramidal cells after an injection that affected the DC and the dorsal part of the LC (Fig. 5A; only the proximal dendrites are drawn, because there is a dense plexus of fiber labeling that made it impossible to distinguish the distal dendrites). The double dendritic morphology allows these neurons to receive direct olfactory input on their external dendritic tree and vomeronasal input from the NS on the internal dendritic tree, as represented schematically in Figure 5B. In accordance with this hypothesis, the double pyramidal neurons are more frequent in the dorsal and rostral part of the LC (Ulinski and Rainey, 1980), where the vomeronasal projection from the NS mainly terminates. In any case, electron microscopic studies are needed to confirm the possibility of olfactory and vomeronasal convergence onto individual double pyramidal cells of 638 E. LANUZA AND M. HALPERN Fig. 5. A: Camera lucida drawing of a double pyramidal neuron of the LC that was retrogradely labeled after an injection affecting both the DC and the LC. The distal dendrites could not be drawn, because a dense plexus of labeled fibers occupied the plexiform layers. Numbers 1, 2, and 3 designate the external, cell, and internal layer of the LC, respectively. The approximate location of this neuron is shown in B. B: Schematic drawing of the rostral LC as seen in a frontal section showing the terminal field of the projection from the main olfactory bulb (hatched area) and the terminal field of the projection from the NS (represented by arrowheads). A double pyramidal neuron is probably receiving both kinds of chemosensory inputs. For abbreviations, see list. Scale bars 5 40 µm in A, 200 µm in B. the LC. The targets of the axons of these neurons are unknown. The ventral part of the LC, as stated above, projects heavily to the PDVR in the lizard G. gecko (Hoogland and Vermeulen-Vanderzee, 1995). It will be important to investigate whether the double pyramidal neurons project to this structure, which, in turn, projects to the VMH. A bilateral projection from the NS to the LC also has been found in the lizards P. hispanica (Martı́nez-Garcı́a et al., 1986, 1993b) and G. gecko (Bruce and Butler, 1984). Martı́nez-Garcı́a et al. have suggested the possibility of integration of olfactory and vomeronasal information in the LC. Therefore, the integration of olfactory and vomeronasal information in the LC appears to be a general feature of squamate reptiles. This fact changes considerably our previous picture of how olfactory and vomeronasal information are processed in the reptilian brain. The LC should not be considered to be a pure olfactory struc- ture (the olfactory cortex) but, rather, a chemosensory cortex that integrates both kinds of chemosensory information. On the other hand, the main olfactory bulb gives rise in T. sirtalis to a light (bilateral) projection to the two nuclei of the rostral amygdala, the ExA and the VAA (Lanuza and Halpern, 1997). Given the strong innervation of these two nuclei by the NS, the olfactory and vomeronasal projections do overlap directly in these areas. The situation is opposite to that found in the LC: In the ExA and the VAA, the predominant input is vomeronasal (from the NS), and there is only a light olfactory input. In the LC, the olfactory input is much stronger than the vomeronasal input. Nothing is known about the connections of the ExA, but the VAA appears to be reciprocally connected with the LC (Martı́nez-Garcı́a et al., 1986) in the lizard P. hispanica, and it has been recently shown in P. hispanica to project (relatively lightly) to the lateral hypothalamus (Lanuza et TERTIARY VOMERONASAL CONNECTIONS IN SNAKES al., 1997). Therefore, this nucleus would be a direct relay of olfactory and vomeronasal information to the hypothalamus, avoiding the multimodal areas (PDVR and DC). Heterogeneity in the NS Curiously enough, retrogradely labeled cells in the NS after injections into the LC consistently appeared only in the lateral part of the NS, thus suggesting that the projection from the NS to the LC originates only in neurons of the lateral (not of the medial) mural layer of the NS. Similar results have been reported for the lizard G. gecko (Bruce and Butler, 1984). This is not the only case of differences between the lateral and medial part of the NS. In the present study, one injection centered in the DLA nucleus (that also affected the DC) gave rise to retrogradely labeled neurons only in the medial and dorsal parts of the NS (not illustrated). Although we only made a single injection into the DLA, it is worth noting that retrograde labeling did not appear in the part of the NS closest to the injection site (the dorsolateral quadrant). In the lizard P. hispanica, injections of BDA into the preoptic hypothalamus give rise to retrogradely labeled cells located mainly in the ventromedial part of the NS, whereas injections into the anterior hypothalamus (that affect the stria terminalis) give rise to retrogradely labeled cells mainly in the lateral mural layer of the NS (Lanuza et al., 1997). These results also suggest that different aspects of the NS have differential projections. More data are needed to confirm this hypothesis. There are also some chemoarchitectonic differences between the medial part and the lateral part of the NS of T. sirtalis. Only the ventral and medial parts contain steroid hormone-accumulating cells (Halpern et al., 1982), and only the medial half of the hilus is positive for acetyl cholinesterase reactivity (Lanuza and Halpern, unpublished observations). In the snake P. regius (Smeets, 1988) and the lizard G. gecko (Smeets et al., 1986b) the dorsal part of the NS has a dense dopaminergic innervation, whereas the ventral part receives only a sparse dopaminergic innervation. The commissural projections that link the two, as we have demonstrated, are organized in a topographical fashion (i.e., they show a rough point-to-point topography). In conclusion, the NS should be considered to be a heterogenous structure, although we do not know yet the functional significance of these intranuclear differences. The question of the origin of the dopaminergic innervation of the dorsal NS is intriguing, because we have not found retrogradely labeled cells in any dopaminergic nucleus after tracer injections into the NS. In fact, all of the retrograde labeling was restricted to the telencephalon, where there are no dopaminergic cells except for the main and accessory olfactory bulbs (Smeets, 1988). However, the dopaminergic cells of the bulbs are not mitral cells but are mainly periglomerular neurons; therefore, it is unlikely that they are the origin of the dopaminargic innervation of the NS. This suggests that the retrograde transport of BDA may not be sensitive enough to detect relatively sparse projections, such as the one that is the origin of the dopaminergic innervation of the NS. Although no topography has been found in the projection from the AOB to the NS (Lanuza and Halpern, 1997), it should be kept in mind that, in mammals, the vomeronasal sensory epithelium has two spatially separated populations of receptor cells (Shinohara et al., 1992; Halpern et al., 1995) that project differentially to the AOB (Jia and 639 Halpern, 1996). It is not known whether the same is true for reptiles. However, because the vomeronasal system appears to be conserved through evolution, the compartmentalization of the NS suggests that the AOB and the vomeronasal epithelium may also be heterogeneous structures in reptiles. If such a heterogeneity exists in squmate reptiles, then this would suggest that different attributes of the vomeronasal stimuli may be processed in different parts of the NS. Comparative remarks The NS of reptiles has been considered to be homologous to the posteromedial cortical amygdaloid nucleus (CoApm) of mammals (Ulinski and Peterson, 1981; Martı́nez-Garcı́a et al., 1991, 1993b) on the basis that both nuclei receive a vomeronasal input from the AOB, project back to the AOB, and have commissural projections with its contralateral equivalent (for the connections of the mammalian CoApm, see de Olmos et al., 1985; Price et al., 1987; Canteras et al., 1992). The CoApm was also believed to project to the shell of the VMH (Krettek and Price, 1978; de Olmos et al., 1985), as was the NS, and this feature also sustained the homology between the two structures. This work and recent studies in the lizards G. gecko (Bruce and Neary, 1995a) and P. hispanica (Lanuza et al., 1997) have shown that the NS does not project to the VMH, and, based on this result, Bruce and Neary (1995a,b) suggested that the NS is a specialization of the squamate brain and does not really have a homologue in mammals. However, Canteras et al. (1992) have shown that the CoApm actually does not project to the hypothalamus. Therefore, the putative homology between the NS and the CoApm appears to be well sustained. Moreover, Canteras et al. (1992) have also found that the CoApm provides an appreciable input to the piriform and entorhinal areas, which may be equivalent to the projection from the NS to the LC. ACKNOWLEDGMENTS This research was supported in part by an NIH grant (DC00104) and by a grant from the Valencian IVEI (Institut Valencià d’Estudis i Investigació; 002/076). EL was supported by a grant from the ‘‘la Caixa’’ Foundation. LITERATURE CITED Andreu, M.J., J.C. Dávila, M.A. Real, and S. Guirado (1996) Intrinsic connections in the anterior dorsal ventricular ridge of the lizard Psammodromus algirus. J. Comp. Neurol. 372:49–58. Bruce, L.L., and A.B. Butler (1984) Telencephalic projections in lizards. I. Projections to cortex. J. Comp. Neurol. 229:585–601. Bruce, L.L., and T.J. Neary (1995a) Afferent projections to the ventromedial hypothalamic nucleus in a lizard, Gekko gecko. Brain Behav. Evol. 46:14–29. Bruce, L.L., and T.J. Neary (1995b) The limbic system of tetrapods: A comparative analysis of cortical and amygdalar populations. Brain Behav. Evol. 46:224–234. Burghardt, G.M. (1970) Chemical perception in reptiles. In J.W. Jonhston, D.G. Moulton, and A. Turk (eds): Advances in Chemoreception, 1. Communication by Chemical Signals. New York: Appleton-CenturyCrofts, pp. 241–308. Canteras, N.S., R.B. Simerly, and L.W. Swanson (1992) Connections of the posterior nucleus of the amygdala. J. Comp. Neurol. 324:143–179. de Olmos, J.S., G.F. Alheid, and C.A. Beltramino (1985) Amygdala. In G. Paxinos (ed): The Rat Nervous System, Vol. 1, Forebrain and Midbrain. Sydney: Academic Press, pp. 223–334. Distel, H. (1978) Behavioral responses to the electrical stimulation of the brain in the green iguana. In N. Greenberg and P.D. MacLean (eds): 640 Behavior and Neurology of Lizards. Rockville: U.S.N.I.M.H., pp. 135– 147. Durward, A. (1930) The cell masses in the forebrain of Sphenodon punctatum. J. Anat. 65:8–44. Everitt, B.J., and T.W. Robbins (1992) Amygdala-ventral striatal interactions and reward-related processes. In J.P. Aggleton (ed): The Amygdala: Neurobiological Aspects of Emotion, Memory and Mental Dysfunction. New York: Wiley-Liss, pp. 401–429. Font, C., M.G. Belekhova, E. Lanuza, N. Kenigfest, A. Martinez-Marcos, and F. Martinez-Garcia (1995) Does amygdala of lizards include the caudal DVR? Eur. J. Neurosci. Suppl. 8:102. González, A., F.T. Russchen, and A.H.M. Lohman (1990) Afferent connections of the striatum and the nucleus accumbens in the lizard Gekko gecko. Brain Behav. Evol. 36:39–58. Halpern, M. (1976) The efferent connections of the olfactory bulb and the accessory olfactory bulb in the snakes, Thamnophis sirtalis and Thamnophis radix. J. Morphol. 150:553–578. Halpern, M. (1980) The telencephalon of snakes. In S.O.E. Ebbesson (ed): Comparative Neurology of the Telencephalon. New York: Plenum Press, pp. 257–294. Halpern, M. (1988) Vomeronasal system functions: Role in mediating the reinforcing properties of chemical stimuli. In W.K. Schwerdtfeger and W.J.A.J. Smeets (eds): The Forebrain of Reptiles: Current Concepts of Structure and Function. Basel: Karger, pp. 142–150. Halpern, M. (1992) Nasal chemical senses in reptiles. In C. Gans and D. Crews (eds): Biology of the Reptilia Vol. 18: Hormones, Brain and Behavior. Chicago: University of Chicago Press, pp. 423–523. Halpern, M., and D. Holtzman (1993) Chemosensing and chemosignaling in reptiles. Brain Behav. Evol. 41:118–268. Halpern, M., and J.L. Kubie (1983) Snake tongue flicking behavior: Clues to vomeronasal system functions. In R.M. Silverstein and D. MüllerSchwarze (eds): Chemical Signals in Vertebrates. New York: Plenum Publishing, pp. 45–72. Halpern, M., J.I. Morrell, and D.W. Pfaff (1982) Cellular [3H]-estradiol and [3H]-testosterone localization in the brain of garter snakes: An autoradiographic study. Gen. Comp. Endocrinol. 46:211–224. Halpern, M., L.S. Shapiro, and C. Jia (1995) Differential localization of G proteins in the opossum vomeronasal system. Brain Res. 677:157–161. Hoogland, P.V. (1981) Spinothalamic projections in a lizard, Varanus exhanthematicus: An HRP study. J. Comp. Neurol. 198:7–12. Hoogland, P.V. (1982) Brainstem afferents to the thalamus in a lizard, Varanus exhanthematicus. J. Comp. Neurol. 210:152–162. Hoogland, P.V., and E. Vermeulen-Vanderzee (1989) Efferent connections of the dorsal cortex of the lizard Gekko gecko studied with Phaseolus vulgaris-leucoagglutinin. J. Comp. Neurol. 285:289–303. Hoogland, P.V., and E. Vermeulen-Vanderzee (1993) Medial cortex of the lizard Gekko gecko: A hodological study with emphasis in regional specialization. J. Comp. Neurol. 331:326–338. Hoogland, P.V., and E. Vermeulen-Vanderzee (1995) Efferent connections of the lateral cortex of the lizard Gekko gecko: Evidence for separate origins of medial and lateral pathways from the lateral cortex to the hypothalamus. J. Comp. Neurol. 352:469–480. Jia, C., and M. Halpern (1996) Subclasses of vomeronasal receptor neurons: Differential expression of G proteins (Gia2 and Goa) and segregated projections to the accessory olfactory bulb. Brain Res. 719:117–128. Kenigfest, N., A. Martı́nez-Marcos, M. Belekhova, C. Font, E. Lanuza, E. Desfilis, and F. Martı́nez-Garcı́a (1997) The lacertilian dorsal retinorecipient thalamus: A reinvestigation in the Old-World lizard Podarcis hispanica. Brain Behav. Evol. (in press). Krettek, J.E., and J.L. Price (1978) Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat. J. Comp. Neurol. 178:225–254. Lanuza, E., and M. Halpern (1997) Efferents and centrifugal afferents of the main and accessory olfactory bulbs in the snake Thamnophis sirtalis. Brain Behav. Evol. (in press). Lanuza, E., C. Font, A. Martı́nez-Marcos, and F. Martı́nez-Garcı́a (1997) Amygdalo-hypothalamic projections in the lizard Podarcis hispanica: A combined anterograde and retrograde tracing study. J. Comp. Neurol. (in press). Lohman, A.H.M., and W.J.A.J. Smeets (1993) Overview of the main and accessory bulb projections in reptiles. Brain Behav. Evol. 41:147–155. Lohman, A.H.M., and I. Van Woerdey-Verkley (1972) Some cortical connections of the tegu lizard (Tupinambis teguixin). Brain Res. 45:325–344. E. LANUZA AND M. HALPERN Lohman, A.H.M., and I. Van Woerdey-Verkley (1976) Further studies on the cortical connections of the tegu lizard (Tupinambis teguixin). Brain Res. 103:9–28. Martı́nez-Garcı́a, F., and M.J. Lorente (1990) Thalamo-cortical projections in the lizard Podarcis hispanica. In W.K. Schwerdtfeger and P. Germroth (eds): The Forebrain in Nonmammals. New Aspects of Structure and Development. Berlin: Springer-Verlag, pp. 93–102. Martı́nez-Garcı́a, F., and F.E. Olucha (1988) Afferent projections to the Timm-positive cortical areas of the telencephalon of lizards. In W.K. Schwerdtfeger and W.J.A.J. Smeets (eds): The Forebrain of Reptiles: Current Concepts of Structure and Function. Basel: Karger, pp. 30–40. Martı́nez-Garcı́a, F., M. Amiguet, F.E. Olucha, and C. López-Garcı́a (1986) Connections of the lateral cortex in the lizard Podarcis hispanica. Neurosci. Lett. 63:39–44. Martı́nez-Garcı́a, F., F.E. Olucha, V. Teruel, M.J. Lorente, and W.K. Schwerdtfeger (1991) Afferent and efferent connections of the olfactory bulbs in the lizard Podarcis hispanica. J. Comp. Neurol. 305:337–347. Martı́nez-Garcı́a, F., P.V. Hoogland, and E. Vermeulen-Vanderzee (1993a) An anatomical analysis of the sensory inputs of the reptilian hippocampus. Eur. J. Neurosci. Suppl. 6:62. Martı́nez-Garcı́a, F., F.E. Olucha, V. Teruel, and M.J. Lorente (1993b) Fiber connections of the amygdaloid formation of the lizard Podarcis hispanica. Brain Behav. Evol. 41:156–162. Mason, R.T. (1992) Reptilian pheromones. In C. Gans and D. Crews (eds): Biology of the Reptilia, Vol. 18: Hormones, Brain and Behavior. Chicago: University of Chicago Press, pp. 114–228. Northcutt, R.G. (1978) Forebrain and midbrain organization in lizards and its phylogenetic significance. In N. Greenberg and P.D. MacLean (eds): Behavior and Neurology of Lizards. Rockville: U.S.N.I.M.H., pp. 11–64. Olucha, F., F. Martı́nez-Garcı́a, L. Poch, W.K. Schwerdtfeger, and C. López-Garcı́a (1988) Projections from the medial cortex in the brain of lizards: Correlation of anterograde and retrograde transport of horseradish peroxidase with Timm staining. J. Comp. Neurol. 276:469–480. Price, J.L., F.T. Russchen, and D.G. Amaral (1987) The limbic region II: The amygdaloid complex. In A. Björklund, T. Hökfelt and L.W. Swanson (eds): Handbook of Chemical Neuroanatomy, Vol. 5: Integrated Systems of the CNS, Part 1. Amsterdam: Elsevier, pp. 279–388. Scalia, F., and S.S. Winans (1975) The differential projections of the olfactory bulb and accessory olfactory bulb in mammals. J. Comp. Neurol. 161:31–56. Scalia, F., G. Gallousis, and S. Roca (1991) Differential projections of the main and accessory olfactory bulb in the frog. J. Comp. Neurol. 305:443–461. Schwenk, K. (1993) The evolution of chemoreception in squamate reptiles: A phylogenetic approach. Brain Behav. Evol. 41:124–137. Shinohara, H., K. Sato, and T. Asano (1992) Differential localization of G proteins, Gi and Go, in the accessory olfactory bulb of the rat. J. Neurosci. 12:1275–1279. Smeets, W.J.A.J. (1988) Distribution of dopamine immunoreactivity in the forebrain and midbrain of the snake Python regius: A study with antibodies against dopamine. J. Comp. Neurol. 271:115–129. Smeets, W.J.A.J., P.V. Hoogland, and A.H.M. Lohman (1986a) A forebrain atlas of the lizard Gekko gecko. J. Comp. Neurol. 254:1–19. Smeets, W.J.A.J., P.V. Hoogland, and P. Voorn (1986b) The distribution of dopamine immunoreactivity in the forebrain and midbrain of the lizard Gekko gecko: An immunohistochemical study with antibodies against dopamine. J. Comp. Neurol. 253:46–60. Ulinski, P.S. (1978) Organization of the anterior dorsal ventricular ridge in snakes. J. Comp. Neurol. 178:411–450. Ulinski, P.S. (1990) The cerebral cortex of reptiles. In E.J. Jones and A. Peters (eds): Cerebral Cortex, Vol. 8A. New York: Plenum Press, pp. 139–215. Ulinski, P.S., and E.H. Peterson (1981) Patterns of olfactory projections in the desert iguana, Dipsosaurus dorsalis. J. Morphol. 168:189–227. Ulinski, P.S., and W.T. Rainey (1980) Intrinsic organization of the snake lateral cortex. J. Morphol. 165:85–116. Veenman, C.L., A. Reiner, and M.G. Honig (1992) Biotinylated dextran amine as an anterograde tracer for single- and double-labeling studies. J. Neurosci. Methods 41:239–254. Voneida, T.J., and C.M. Sligar (1979) Efferent projections of the dorsal ventricular ridge and the striatum in the tegu lizard Tupinambis nigropunctatus. J. Comp. Neurol. 186:43–64.