Download pdf - Vasotocin innervation and modulation of vocal-acoustic circuitry in the teleostPorichthys notatus

Vasotocin Innervation and Modulation ofVocal-Acoustic Circuitry in the Teleost

Porichthys notatus

JAMES L. GOODSON AND ANDREW H. BASS*

Department of Neurobiology and Behavior, Cornell University, Ithaca, New York 14853

ABSTRACTArginine vasotocin (AVT) and its mammalian homologue arginine vasopressin (AVP)

modulate reproduction-related and other social behaviors in a broad range of vertebratespecies. These functions of AVT/AVP may be in part achieved through the modulation ofsensorimotor integration, although experimental evidence supporting this hypothesis re-mains limited. In the present experiments, we demonstrate (1) AVT innervation of candidatevocal-acoustic brain regions, and (2) AVT modulation of vocal-motor physiology in the plainfinmidshipman fish (Porichthys notatus), which uses vocalizations in both mate attraction andagonistic contexts. AVT distribution was compared with known vocally active brain regionsand to central auditory and vocal pathways. AVT-immunoreactive fibers and putative ter-minals descend almost exclusively from the preoptic area and are found in two primarycandidate sites for vocal-acoustic integration - the anterior tuberal hypothalamus andparalemniscal midbrain tegmentum. AVT immunoreactivity is also located in several othervocally active regions, including the ventral tuberal nucleus, periaqueductal gray, and para-ventricular regions of the isthmus and rostral hindbrain. The parvocellular preoptic areaitself is also vocally active, although thresholds are substantially higher than for otherregions. The functional significance of AVT input to vocal-acoustic regions was demonstratedin the paralemniscal midbrain where local delivery of AVT modulated electrically evoked,rhythmic vocal-motor output, which precisely mimicked natural vocalizations. AVT produceddose-dependent inhibitions of parameters associated with call initiation (burst latency andnumber of vocal-motor bursts elicited) but not of vocal-motor patterning (fundamental fre-quency and burst duration). Together, these findings provide support for the proposal thatAVT modulates sensorimotor processes underlying social/acoustic communication. J. Comp.Neurol. 422:363–379, 2000. © 2000 Wiley-Liss, Inc.

Indexing terms: vasopressin; auditory; midbrain; sex differences; communication; fish

The neuropeptide arginine vasotocin (AVT; nonmam-mals) and its mammalian homologue, arginine vasopres-sin (AVP) are key components of coordinated behavioralexpression, as a variety of anatomic and behavioral find-ings now implicate these peptides in the modulation of awide diversity of sex-typical and species-specific behaviors(e.g., sexual, aggressive, pair-bonding, vocal, and parentalbehaviors). AVT/AVP distributions and behavioral func-tions are also often steroid- and/or seasonally dependent.Across all vertebrate classes, AVT/AVP cell bodies arefound in the preoptic area-anterior hypothalamus (POA-AH), a sensorimotor integration center that also regulatesnumerous physiologic and hormonal processes by meansof its control of the pituitary gland (for reviews and/orextensive treatments of the findings discussed above, seede Vries and Miller, 1998; Engelmann et al., 1996; Good-

son, 1998b; Lowry et al., 1997; Moore, 1992; Moore andLowry, 1998; Young et al., 1998).

Whereas AVT/AVP distributions and behavioral func-tions have been extensively studied with regard to sex,species, and endocrine variables, potential roles for AVT/AVP in the modulation of sensorimotor functions havebeen investigated relatively less. Most extensively studied

Grant sponsor: National Institutes of Health; Grant number: F32 NS-0443; Grant sponsor: National Science Foundation; Grant number: IBN9421319.

*Correspondence to: Andrew H. Bass, Department of Neurobiology andBehavior, Mudd Hall, Cornell University, Ithaca, NY 14853.E-mail: [email protected]

Received 8 December 1999; Revised 25 February 2000; Accepted 25February 2000

THE JOURNAL OF COMPARATIVE NEUROLOGY 422:363–379 (2000)

© 2000 WILEY-LISS, INC.

is the male rough-skinned newt (Taricha granulosa), inwhich AVT has been implicated in visual-, olfactory-, andsomatosensory-guided reproductive behaviors (Rose et al.,1995; Zoeller and Moore, 1986; Zoeller and Moore, 1988).By contrast, virtually no data are available on AVT/AVP’spotential modulation of vocal-acoustic integration, al-though AVT/AVP effects on vertebrate social behavior aremost widely established for vocalization (amphibians: Dia-kow, 1978; Raimondi and Diakow, 1981; Boyd, 1992;Penna et al., 1992; Boyd, 1994; Marler et al., 1995; Prop-per and Dixon, 1997; Chu et al., 1998; Semsar et al., 1998;Tito et al., 1999; mammals: Winslow and Insel, 1991,1993; birds: Voorhuis et al., 1991; Maney et al., 1997;Castagna et al., 1998; Goodson, 1998a; and fish: Goodsonand Bass, 2000). Most relevant are findings in the greentreefrog (Hyla cinerea), which demonstrate that AVT de-creases auditory sensitivity and facilitates calling in re-sponse to playback of conspecific vocalizations (Penna etal., 1992). Brain site-specific modulation of calling has notyet been established for anurans, although AVT distribu-tion in anuran species includes both forebrain and brain-stem regions implicated in call production, auditory pro-cessing, or both (e.g., POA, pretrigeminal nucleus, andhindbrain motor regions; for review, see Emerson andBoyd, 1999).

The present experiments were conducted (1) to establishan anatomic basis for AVT modulation of vocal-acousticprocesses in the plainfin midshipman (Porichthys nota-tus), a teleost fish that uses social vocalizations in multi-ple behavioral contexts (Brantley and Bass, 1994; Bass etal., 1999), and (2) to further test the hypothesis that AVTmodulates vocalization in the midshipman by local actionin candidate sites for vocal-acoustic integration. The mid-

shipman provides an excellent opportunity for such inves-tigation, as behavioral experiments have detailed acousticparameters of vocalization which are required for signalrecognition (McKibben and Bass, 1998) and the circuit-ries of the central auditory and vocal systems have beenwell characterized (Bass et al., 1994, 2000). In addition,the present experiments are intended to provide previ-ously unavailable detail on AVT-immunoreactive (-ir)innervation of the teleost brain, as descriptions of AVTimmunoreactivity in teleosts are largely restricted toinvestigations of the POA and its innervation of theneurohypophysis (see Discussion section for extensivelist of references).

The plainfin midshipman expresses three adult pheno-types that exhibit distinct suites of neural, somatic, endo-crine, and behavioral traits (for reviews, see Bass, 1992,1996; Bass et al., 1999; also see Grober et al., 1994; Foranand Bass, 1998; Schlinger et al., 1999; Goodson and Bass,2000). Only type I “courting” males emit a long-durationrhythmic “hum” (typically .1 second, up to 1 hour), whichis used to attract females, and “growls,” which arefrequency- and amplitude-modulated vocalizations of un-specified significance (Bass et al., 1999). However, allmorphs (including type II sneak-spawning males andfemales) give short-duration agonist “grunts” (,250msec).

The divergence between morphs in social behavior isparalleled by divergence in preoptic AVT-ir neuronal phe-notype and peptidergic modulation of forebrain-evokedvocalization. Compared with type I males and females,type II males have larger AVT-ir neurons in the POA(explained by differences in body size) and have more cells

Abbreviations

ac anterior commissureAH anterior hypothalamusAP area postremaAT anterior tuberal nucleusC cerebellumCA cerebral aqueductcc cerebellar crestCm molecular layer of the cerebellumCPc compact division of the central posterior nucleusCPd diffuse division of the central posterior nucleusDF diffuse nucleus of the hypothalamusDL dorsolateral telencephalonDM dorsomedial telencephalonDP dorsal posterior nucleus of the thalamusG nucleus glomerulosusH hindbrainHd dorsal periventricular hypothalamusHoC horizontal commissureHP hindbrain paraventricular nucleusHv ventral periventricular hypothalamusHYP hypothalamusIL inferior lobe of the hypothalamusIP isthmal paraventricular nucleusIS isthmal nucleusLH lateral hypothalamusll lateral lemniscusLH lateral hypothalamusM midbrainMLF medial longitudinal fasciculusnll nucleus of the lateral lemniscusOC occipital nerveOL olfactory nerveOP optic nerve

ot optic tractP pituitaryPCo posterior commissurePGc caudal division of nucleus preglomerulosusPGl lateral division of nucleus preglomerulosusPGm medial division of nucleus preglomerulosusPHT preopticohypophysial tractPL paralemniscal midbrain tegmentumPM magnocellular preoptic nucleusPN pacemaker neuronsPOA preoptic areaPPa anterior parvocellular preoptic nucleusPPp posterior parvocellular preoptic nucleusRF reticular formationSA saccular otolithSC spinal cordSMN sonic motor nucleusSV saccus vasculosusT telencephalonTe midbrain tectumTeg midbrain tegmentumTS torus semicircularisv ventricleVg granule layer of the valvulaVL ventrolateral thalamusVm molecular layer of the valvulaVM ventral medullary nucleusVs supracommissural nucleus of the ventral telencephalonvT ventral tuberal hypothalamusIII third ventricleIV fourth ventricleVde descending tract of the trigeminal nerveVIIm facial motor nucleus

364 J.L. GOODSON AND A.H. BASS

per gram body mass (Foran and Bass, 1998). AlthoughAVT-ir neuronal phenotype differs between females andtype II males, their similarity in vocal-motor morphology,physiology, and behavior (Bass, 1992, 1996) is nonethelessreflected in their neuropeptide modulation of vocal-motorcircuitry (Goodson and Bass, 2000). Specifically, vocal-motor output may be evoked from all adult morphs byelectrical stimulation of the POA-AH (see below), and thisresponse is inhibited by local administration of AVT intype I males. However, type II males and females arelargely insensitive to AVT but sensitive to isotocin (homo-logue of mammalian oxytocin). Therefore, AVT and isoto-cin contribute to the differential expression of vocal be-havior which is tied to phenotypic divergence in theexpression of parental, courtship, and nest defense behav-iors; these are behaviors known to be modulated by pep-tides of the vasopressin-oxytocin family in other verte-brate species (see Goodson and Bass, 2000).

In the experiments described above, we have taken ad-vantage of the fact that the synchronous, rhythmic dis-charge of the midshipman’s hindbrain vocal pacemaker-motoneuron circuit (PN-SMN; Fig. 1) directly establishesthe fundamental frequency and duration of actual vocal-ization (Cohen and Winn, 1967; Bass and Baker, 1990).Hence, recordings of fictive vocalization from the occipitalnerves (OC; Fig. 1), whose axons arise in part from thesonic motor nucleus and directly innervate the sonic mus-culature of the swim bladder, precisely mimic naturalvocalization. In this report, we use the midshipman prep-aration to identify brain regions involved in the produc-tion of all classes of midshipman vocalization, and, on thebasis of the distribution of AVT immunoreactivity andprojections of previously identified vocal and auditorypathways, we further establish AVT modulation of fictivevocalizations evoked from candidate vocal-acoustic cir-cuitry.

MATERIALS AND METHODS

Subjects

All subjects were collected from field sites near TomalesBay, California. Type I males (11.9–20.5 cm) used forneurophysiological experiments were captured during the1998 and 1999 breeding seasons (April–August). Animalswere held in running seawater tanks at the University ofCalifornia Bodega Marine Laboratory until shipment toCornell University, where they were maintained in artifi-cial seawater tanks at 12–16°C. Care was in accordancewith institutional guidelines of Cornell University. Ani-mals used for immunocytochemistry (four females, 12.1–13.0 cm; two type II males, 7.7–12.3 cm; three type Imales, 14.0–23.2 cm) were collected in mid-May of 1999and killed on the day of capture. An additional type IImale (8.4 cm) was handled as indicated above for physiol-ogy subjects and perfused in November 1998; results fromthis subject were similar to other type II subjects.

Immunocytochemistry

Subjects were anesthetized by immersion in tricainemethanesulfonate (MS 222; Sigma Chemical Co., St.Louis, MO) and perfused with ice-cold teleost Ringer’ssolution followed by 4% paraformaldehyde. Brains wereremoved, postfixed for 1 hour, and transferred to 0.1 Mphosphate buffer (PB; pH 7.2) for storage. After overnight

incubation of brains in 30% sucrose, 30-mm sections werecut on a cryostat and thaw-mounted on chrom-alumsubbed slides.

Immunocytochemistry (ICC) was performed by usingR-82 AVT antiserum (a gift of Dr. F. van Leeuwen, Neth-erlands Institute for Brain Research) and procedures es-tablished for use with an AVP antibody in the midship-man (Foran and Bass, 1998): 20 minutes in 0.1 Mphosphate buffered saline (PBS; pH, 7.2); 20 minutes inPBS 1 0.5% bovine serum albumin (BSA); 16–18 hours inprimary antiserum diluted 1:4,000 in PBS 1 BSA 1 0.01%sodium azide; two 10-minute rinses in PBS 1 BSA; 1 hourin secondary (Vectastain Rabbit IgG Kit, Vector Labora-tories, Burlingame, CA); two 10-minute rinses in PBS 1BSA; 1 hour in avidin-biotin complex (Vectastain RabbitIgG Kit, Vector Laboratories); two 10-minute rinses in PB.

Fig. 1. A dorsal view of the plainfin midshipman brain showingthe locations of the preoptic area (POA; the primary source of vaso-tocin fibers in the forebrain and midbrain); the ventral tuberal hypo-thalamus (vT), a vocally active region which contains a small popu-lation of arginine vasotocin-immunoreactive (AVT-ir) cells andreceives projections from the auditory thalamus (central posteriornucleus; Goodson and Bass, unpublished observations); torus semi-circularis (TS; auditory midbrain region; afferent of candidate vocal-acoustic sites; Bass et al., 2000); neurophysiologically identified vocalsites demonstrated to receive vasotocinergic and toral input (anteriortuberal nucleus, AT; paralemniscal midbrain, PL); and AVT-ir vocallyactive regions afferent to the vocal pattern generator (isthmal andhindbrain paraventricular cell groups, IP/HP; Bass et al., 1994). Alsoshown are component structures of the hindbrain vocal pattern gen-erator (ventral medullary nucleus, VM; pacemaker neurons, PN; sonicmotor nucleus, SMN; after Bass et al., 1994). The rhythmically activePN are linked bilaterally by VM and establish the fundamental dis-charge frequency of the SM. The SMN projects to the sonic swimbladder musculature by means of paired occipital nerves (OC) anddirectly establishes the fundamental frequency and duration of vocal-ization. Hence, extracellular recordings of fictive vocalization from theOC precisely mimic natural vocalization (details based on intracellu-lar recording and staining; Bass and Baker, 1990). Also shown arerepresentative fictive grunts elicited by electrical stimulation of vT(from Goodson and Bass, 2000). Lettered levels above the figurecorrespond to camera lucida drawings shown in Figure 2. For otherabbreviations, see list.

365VASOTOCIN AND VOCAL-ACOUSTIC CIRCUITRY

Label was visualized with a diaminobenzidine (DAB) re-action by incubating sections for 80–100 seconds in 0.05%DAB in PB 1 0.3% hydrogen peroxide. Sections were thenrinsed twice in PB, rinsed in deionized water, lightly coun-terstained with cresyl violet (optional), dehydratedthrough an alcohol series, cleared in xylene, and cover-slipped with Permount (Fisher Scientific, Fair Lawn, NJ).For most sections, postincubation rinses were preceded bydipping in PBS; this procedure reduced background labelto virtually undetectable levels.

The specificity of the R-82 antiserum has been repeat-edly tested in fish (van den Dungen et al., 1982; Holmqvistand Ekstrom, 1991), amphibians (Gonzalez and Smeets,1992a,b, 1997; Gonzalez et al., 1995; Lowry et al., 1997),and reptiles (Stoll and Voorn, 1985; Thepen et al., 1987;Fernandez-Llebrez et al., 1988; Smeets et al., 1990). Spec-ificity in the midshipman was additionally assessed inthree control conditions: 1) primary antiserum was omit-ted; 2) primary antiserum was preincubated for 24 hoursin 10 mM AVT; and 3) primary antiserum was preincu-bated for 24 hours in 10 mM isotocin (peptides fromBachem California, Inc., Torrance, CA).

Photomicrographs. Photomicrographs were takenwith a Nikon Eclipse E-800 microscope and were scanned,scaled, and labeled in Photoshop 5.0 for the Macintosh.Photomicrographs shown are otherwise unmanipulated.

Identification of candidate vocal-acoustic sites

To identify candidate sites for AVT modulation of vocal-acoustic processes, the distribution of AVT immunoreac-tivity was compared with 1) the pattern of label obtainedby biotin compound injections into neurophysiologicallyidentified sites in the auditory division of the torus semi-circularis (homologue of the inferior colliculus of birds andmammals; Bass et al., 2000); 2) brainstem vocal-motorcircuitry delineated by transneuronal transport of biocytinfrom the sonic occipital nerve roots (Bass et al., 1994); 3)vocal regions previously identified by electrical brain stim-ulation in batrachoidid fish (Demski and Gerald, 1972,1974; Fine and Perini, 1994); and 4) vocal-acoustic regionsidentified in the midshipman as part of ongoing experi-ments to map the descending vocal circuitry originating inthe forebrain (Goodson and Bass, unpublished observa-tions). On the basis of its extensive connectivity to othervocal-acoustic nuclei and the density of AVT-ir innerva-tion, the paralemniscal midbrain was selected for furtherphysiological investigation, as detailed in the Results sec-tion.

Neurophysiological experiments

Neurophysiological procedures were adopted from thoseused in previous studies (Bass and Baker, 1990; Goodsonand Bass, 2000). Dorsal craniotomy was conducted undergeneral anesthesia (immersion in 0.2% benzocaine; Sig-ma). The subject was placed in a Plexiglas tank, and theskull was stabilized with a chronic headholder. The brainwas kept covered with a fluorocarbon (Fluorinert, 3MCorp., St. Paul, MN). Fish were perfused through themouth with salt water at 15–16°C. During recording, pan-curonium bromide (0.5 mg/kg; Astra PharmaceuticalProducts, Inc., Westborough, MA) was used for immobili-zation and fentanyl (1 mg/kg; Sigma) was used for anal-gesia.

Tungsten electrodes insulated except at their tips(125-mm diameter; 20-mm exposed tips; 5 MV impedance)were used to deliver brief (30- to 35-msec) trains of low-amplitude stimuli (0.05- to 0.1-msec duration, 280–350Hz). The use of a stimulus frequency range allowed anaccommodation for individual variability, as peptides ex-periments required robust and reliable responses overprotracted time periods. Importantly, stimulus frequencyis not related to the discharge frequency of the sonic motorsystem (Bass and Baker, 1990). Output from an occipitalroot was monitored unilaterally with an electrode con-structed of paired, Teflon-coated silver wires with exposedball tips 50–100 mm in diameter. Both sides of the brainfire in phase so that unilateral monitoring reflects bilat-eral synchrony of the vocal circuit (Bass and Baker, 1990).Recordings were digitized by using a Power Macintosh8100 with IGOR Pro software (Wavemetrics, Inc., LakeOswego, OR; software customized by R. Wyttenbach, Cor-nell University).

AVT modulation of vocal-motor physiology in theparalemniscal midbrain was assessed by local pressuredelivery of peptide in conjunction with electrical stimula-tion (method after Goodson and Bass, 2000). Consistentvocal-motor response to paralemniscal stimulation wasobtained in 16 subjects (mean treatments per fish 5 1.5; 3maximum; experiments were conducted in September–October, 1998, and May–June, 1999). Once consistent out-put was established, baseline response was assessed bydelivering 15 trains of stimuli at 1-second intervals; base-line response was defined as vocal-motor activity elicitedduring the 15 seconds of stimulation and the subsequent45 seconds.

After baseline establishment, the tungsten electrodewas removed and a glass pipette (tip diameter approxi-mately 20 mm) was inserted for pressure delivery of AVT(0.1, 10, and 1,000 ng; n 5 3, 4, and 3, respectively), anAVP V1 receptor antagonist ([b-Mercapto-b, b-cyclopenta-methylenepropionyl1,O-Me-Tyr2,Arg8]-vasopressin) (0.1,10, and 1,000 ng; n 5 2, 3, and 2, respectively), isotocin(1,000 ng; n 5 3), or vehicle control (teleost Ringer; i.e., 0ng; n 5 3). Substances were delivered in 0.2-ml vehicleover a 10- to 15-second period. The tungsten electrode wasimmediately reinserted for collection of posttreatmentdata. Accurate placement in the previous site was readilyachieved by using micromanipulator coordinates and mul-tiple surface landmarks. For subjects used in multipleexperiments, a minimum of 2 hours was allowed betweentreatments; multiple treatments were of different sub-stances in all but two subjects, which received two differ-ent doses of the same substance (random order). Peptideswere obtained from Sigma (St. Louis, MO) and BachemCalifornia (Torrance, CA; isotocin only).

Stimulation sites were marked in 11 fish by inducingsmall electrolytic lesions (n 5 2), pressure injection ofIndia ink (n 5 7), or iontophoretic application of biotincompound tract tracers (10% dextran biotin or 5% neuro-biotin in 3 M KCl; n 5 2). Standard histologic procedureswere used (Bass et al., 1994, 2000).

Nomenclature

The nomenclature of Braford and Northcutt (1983) wasadopted for descriptions of the preoptic area and dienceph-alon; works by Striedter (1990a,b) were additionally con-sulted. The cytoarchitecture of the midshipman POA hasbeen described previously (Foran and Bass, 1998; Foran et


al., 1997). Nomenclature for the mesencephalon andrhombencephalon is consistent with that used for previousdescriptions of vocal-acoustic pathways in the midship-man (Bass et al., 1994, 2000).

RESULTS

Distribution of AVT immunoreactivity

The R-82 AVT antiserum produces excellent labeling ofpreoptic neurons and of processes throughout the brain.Labeling is highly specific: deletion of primary antibody orpreadsorption of the primary antibody with 10 mM AVTcompletely eliminates label, whereas preadsorption with10 mM IT produces no obvious label reductions. However,the R-82 antiserum produces a different pattern of neuro-nal label outside of the POA than that previously obtainedin the midshipman by using an AVP antibody (Foran andBass, 1998), likely due to differential specificities of theseantibodies (see Discussion section). The general distribu-tion of AVT immunoreactivity was similar for all midship-man morphs. A clear sex difference was observed only inthe dorsal midbrain tegmentum, as described below.

Forebrain. Consistent with a previous investigationof POA neuronal phenotypes in the midshipman (Foranand Bass, 1998), dense clusters of AVT-ir subependymalneurons are found in the anterior parvocellular (PPa),posterior parvocellular (PPp), and magnocellular (PM)preoptic nuclei (Figs. 2A–C, 3A–G). In addition, a fewsmall, round AVT-ir neurons are found in the ventrome-dial portion of the ventral tuberal hypothalamus (vT; Figs.2C, 3H), often embedded in the preopticohypophysial tract(PHT; Fig. 2C). The majority of parvocellular AVT-irperikarya are located in the PPa (Figs. 2A,B, 3A,B,E,F); asmall number of more weakly labeled cells are located inthe PPp just ventral to the PM (Fig. 2C). AVT immunore-activity in the PM is mainly found in the magnocellulardivision (Figs. 3C–G); these neurons stain more intenselythan those in the PPa and PPp. The caudal PM divisionalso includes gigantocellular neurons (Foran et al., 1997);few of these are labeled in comparison to magnocellularPM neurons and AVT-ir gigantocellular somata arepresent in only 4 of the 10 animals studied (all morphsrepresented). A small number of AVT-ir magnocellularand gigantocellular neurons are observed in a migratedposition among cells at the caudal aspect of the telenceph-alon (T, Fig. 3G).

AVT-ir processes extend from POA neurons in a primar-ily lateral direction, although a modest number of fibersinnervate periventricular regions dorsal and ventral tothe AVT-ir cell clusters. Laterally extending fibers withlarge varicosities, beads, or both, are most concentrated atrostral levels of the PPa (Figs. 2A,B, 3A). Adjacent to themagnocellular division, a more modest number of large-caliber, beaded and varicose processes are present (Figs.2C, 3C); most AVT-ir fibers in this area course caudal andventrolateral to form the dense PHT tract (Fig. 2C). Fibersfrom AVT-ir neurons in the ventral tuberal region minglewith those of the PHT and other fibers originating in thePOA. Few fibers innervate basal telencephalic regionsoutside of the POA. At the rostral-most levels of the PPa,a few beaded axons can be found adjacent to the anteriorcommissure and extending into the supracommissural nu-cleus of the ventral telencephalon (Vs; Fig. 2A). In somebrains, scattered fibers are also observed in the dorsaltelencephalon.

The majority of descending fibers from the POA courselaterally through the hypothalamus and sweep down ad-jacent to the medial and lateral preglomerular nuclei(Figs. 2C–E, 3E). A portion of these fibers maintain acaudal course in this lateral position, whereas others de-scend to the hypothalamic floor and either continue to thepituitary or exit the PHT dorsomedially to innervate tu-beral regions of the hypothalamus (Figs. 2C,D, 3E,H,I).Innervation of the tuberal hypothalamus is modest atrostral and intermediate levels (i.e., ventral tuberal [vT]and lateral tuberal regions, respectively), although inthese regions multiple fibers with large varicosities can befound associated with the PHT (Figs. 2C, 3H). A denseinnervation of the anterior tuberal nucleus (AT) and lat-eral hypothalamus (LH) is observed at the caudal extentof the PHT (Figs. 2D, 3I).

A modest number of fibers continue dorsocaudally toinnervate the posterior tuberal region and scattered fibersand putative terminals are present throughout the medialaspects of the dorsal diencephalon (Fig. 2D). Dorsal to theanterior tuberal nucleus, beaded fibers are occasionallyfound in the auditory thalamus (diffuse and compact di-visions of the central posterior nucleus, CPd and CPc,respectively; Fig. 2D; see Bass et al., 2000, for cytoarchi-tecture). Scattered beaded fibers are also located in re-gions just ventral and lateral to the posterior commissure(PCo; Fig. 2D). Ventrally, a small number of labeled fibersis consistently found and appears to terminate in a cell-poor zone between the dorsal and ventral nuclei of theperiventricular hypothalamus (Hd, Hv; Fig. 2E). Finally, aloose aggregation of AVT-ir fibers (which forms the dorsaldescending tract to the midbrain; see below) is found inthe dorsal pretectum (Fig. 2D).

Midbrain. AVT-ir fibers descending to the midbraincourse in two tracts. The first of these is composed of fibersthat split from the PHT, maintain a position along thelateral edge of nucleus glomerulosus (G, Fig. 2E), andcourse ventromedially along the tegmental floor en routeto the caudal midbrain (Fig. 2F). The second tract is sex-ually dimorphic (see below) and is composed of a diffuseassemblage of very fine-caliber, beaded AVT-ir fibers thatcourse caudally in a cell-poor region just ventral to theperiaqueductal cell zone (arrows, Fig. 2E,F). Fibers de-scending through the dorsal diencephalon aggregate toform this tract at the level of the posterior commissure(PCo, Fig. 2D) and continue into the mesencephalon in aposition just medial to the ventral extension of theperiventricular cell band of the torus (which delimits themedial boundary of the auditory torus; Fig. 2E,F; see Basset al., 2000, for cytoarchitecture). The bundle maintainsthis position in the cell-poor zone until caudal midbrainlevels.

At caudal midbrain levels, the beaded fibers from thedorsal tract course ventrally into periaqueductal and dor-sal tegmental regions, including an isthmal nucleus (IS;Fig. 2G) that forms part of the ascending auditory path-way (Bass et al., 1994, 2000). At this same level, fibersfrom the “lateral” bundle, which now occupy a positionventrally, sweep dorsally and medially throughparalemniscal regions of the tegmentum where putativeterminals are often dense (PL; Figs. 2G, 3J). The fibersfrom the two tracts mingle as they continue caudally,where beaded fibers innervate periventricular cell groupsof the isthmus and rostral hindbrain (Figs. 2H, 3K). MostAVT-ir fibers disappear abruptly at the isthmus, although


a small number of scattered fibers from the lateral bundlecontinue in the reticular formation and along the tegmen-tal floor into the hindbrain (Fig. 2H,I).

The density of the descending tracts to the midbrain isquite variable between subjects. This is particularly pro-nounced for the dorsal tract. Whereas the dorsal tractranges from very diffuse to modestly robust in male sub-

jects (both morphs), virtually no AVT-ir fibers can be lo-cated in three of the four female subjects. This qualitativedimorphism in the number of descending fibers is evidentalso in the density of AVT-ir label in the dorsal tegmen-tum (isthmal nucleus, nucleus of the lateral lemniscus)and periaqueductal regions of the caudal midbrain, appar-ent sites of termination for the dorsal tract.

Fig. 2. A–K: Distribution of arginine vasotocin-immunoreactive(AVT-ir) cells (large dots in A–C), fibers, and putative terminals(small dots). Camera lucida drawings correspond to lettered levelsshown in Figure 1. Virtually all AVT-ir label in the forebrain, mid-brain, and rostral hindbrain (A–F) appears to originate in the POA

(PPa, PPp, and PM; A–C); a small number of cells are also present inthe ventral tuberal region (vT; C). Dense AVT-ir fibers and putativeterminals in the caudal hindbrain (H) concentrate in the area pos-trema (AP); the source of these fibers is not yet specified. For otherabbreviations, see list. Scale bar 5 1 mm.


Hindbrain. AVT-ir label in most of the hindbrain isrestricted to a small bundle of fibers that extends alongthe lateral edge of the brainstem from caudal levels of thecerebellum into the rostral spinal cord, where our tissuesectioning ended (Figs. 2I–K). These fibers concentrate inthe area postrema (AP), located immediately caudal to thefourth ventricle and at rostral levels of the sonic motornucleus. Although no AVT-ir terminals or fibers enter thesonic motor nucleus proper, labeled terminals are juxta-posed to a band of small cells that surrounds the dorso-lateral aspect of the nucleus at rostral levels (e.g., arrow,Fig. 3L). Extremely dense, beaded AVT-ir fibers courseventrolaterally from area postrema along the dorsolateraledge of the sonic motor nucleus and join the lateral tract ofAVT-ir fibers (Figs. 2J, 3L,M). No AVT-ir somata areobserved in this region and the source of AVT-ir fibers inthe area postrema, therefore, remains to be determined.

AVT-ir innervation of candidate vocal-acoustic regions

With the exception of the area postrema, the most denseconcentrations of putative AVT-ir terminals in the mid-shipman brain show close correspondence to the locationsof six neurophysiologically identified and anatomically

verified vocal sites: the ventral and anterior tuberal nu-clei, the paralemniscal midbrain tegmentum, the periaq-ueductal gray, and the isthmal/hindbrain paraventricularregion. The sixth vocal region is the PPa, although re-sponses in this region are not reliably obtained and re-quire stimulus amplitudes that are approximately twicethat of other sites. Therefore, we exclude the PPa as a“primary” vocal-acoustic region, although our data do sug-gest a potential involvement in vocal processes. Theselocations are consistent with those previously identified byelectrical brain stimulation in Opsanus (Demski and Ger-ald, 1972, 1974; Fine and Perini, 1994) and Porichthys(Bass and Baker, 1990), and all are components of ana-tomically characterized vocal circuitry in the midshipman(Bass et al., 1994; Goodson and Bass, unpublished obser-vations). Fictive vocalizations evoked by electrical stimu-lation of two of these areas, the anterior tuberal nucleusand the paralemniscal midbrain tegmentum, are shown inFigure 4, and a typical paralemniscal stimulation sitemarked with dextran biotin is shown in Figure 5A.

The ventral tuberal region, anterior tuberal nucleus,and paralemniscal midbrain tegmentum are primary can-didate sites for vocal-acoustic integration, as these sitesreceive projections from identified components of ascend-

Figure 2 (Continued)


Figure 3

Figure 3 (Continued) (Overleaf)


ing auditory circuitry in the midshipman (Bass et al.,2000; Goodson and Bass, unpublished observations).Thus, biotin injections into the auditory torus label fibersand putative terminals in the anterior tuberal nucleus(Bass et al., 2000) and biotin compound injections into theventral tuberal site retrogradely label neurons in the au-ditory thalamus (central posterior nucleus; Goodson andBass, unpublished observations); AVT modulation ofvocal-motor activity in the ventral tuberal region has al-ready been demonstrated (Goodson and Bass, 2000). Fi-nally, the paralemniscal region is extensively connected toother vocal-acoustic nuclei, as discussed below; therefore,we chose to focus our neurophysiological experiments onthis region for two important reasons: (1) Such experi-ments allow a comparison of AVT effects at both forebrain(Goodson and Bass, 2000) and midbrain levels, and (2) thevocal-acoustic connectivity of the paralemniscal region farexceeds that of the anterior tuberal hypothalamus (Good-son and Bass, unpublished observations).

AVT-ir fibers and auditory torus projections converge onthe paralemniscal tegmentum at the rostral level of the

nucleus of the lateral lemniscus (nll), a structure that capsthe lateral lemniscus at caudal midbrain levels (Fig. 2G).Projections from the auditory torus are particularly densewithin the nll and more modest medial to the lemniscus,whereas AVT-ir label follows an opposite pattern. How-ever, biotin injections in the PL at this level reveal strongconnections to the nll and other vocal-acoustic nuclei (Fig.5), including the isthmal nucleus and paraventricular cellgroups of the isthmus and hindbrain. The isthmal nucleuslies dorsomedial to the nll and is a more expansive region.Modest input to the isthmal nucleus is provided by theauditory torus and putative terminals are also located inthe isthmal nucleus after transneuronal biotin labeling ofthe hindbrain PN-SMN circuit (Bass et al., 1994). Theparaventricular cell groups of the isthmus and hindbrainalso contain somata transneuronally labeled by means ofthis motor circuit. Thus, the entire paralemniscal region(inclusive of the isthmal nucleus and nll) is a point ofextensive vocal-acoustic convergence. This is shown sche-matically in Figure 6, which diagrams the overlap of AVTimmunoreactivity and projections from the auditory torus.Another study in preparation will consider the entire pat-tern of forebrain and brainstem connectivity to the PN-SMN circuit (Goodson and Bass, unpublished observa-tions).

AVT modulation of midbrain vocal-motor physiology

The majority of fictive vocalizations evoked from theparalemniscal tegmentum were grunt-like (77%). Hum-and growl-like output was not obtained in all subjects andinspection of the data revealed no differential peptideeffects upon the vocal types. Thus, all data were combinedfor analyses.

Pressure delivery of AVT (0, 0.1, 10, and 1,000 ng; n 53, 3, 4, and 3, respectively) into the paralemniscal mid-brain tegmentum produced a significant dose-dependentreduction (from baseline levels) in the number of vocal-motor bursts produced by electrical stimulation of thesame site (a marked stimulation site is shown in Fig. 5A),and also produced a significant dose-dependent increasein vocalization latency (Fig. 7A and B, respectively), sug-

Fig. 3 (Overleaf). Brightfield photomicrographs of argininevasotocin-immunoreactive (AVT-ir) label in Nissl-stained, coronal sec-tions of the midshipman brain. AVT-ir label appears as a brownreaction product. Photomicrographs in A–F are from a type II male;remaining photomicrographs are from a type I male, with the excep-tion of the inset in H, which is from a female. A: Low-power view ofAVT-ir cells in the anterior parvocellular division of the preoptic area(PPa), laterally projecting beaded fibers (inset), and the preopticohy-pophsial tract (PHT). The arrow in A points to the cells shown in B.B: PPa neurons and putative terminals (arrow). C: MagnocellularAVT-ir neurons in the magnocellular division of the preoptic area(PM) and laterally projecting fibers. Inset: A varicose fiber in closecontact with an immunonegative neuron dorsolateral to PM. D: High-magnification view of the magnocellular neurons shown in C (leftside). E: AVT-ir at an intermediate level of the POA, corresponding tothe caudal aspect of the horizontal commissure. A single bipolarmagnocellular neuron of the PM (arrow) is seen ventrally (rostral-most extension of the PM); other AVT-ir perikarya belong to the PPa.F: High-magnification view of the bipolar magnocellular neuronshown in D and adjacent, smaller AVT-ir parvocellular neurons of thePPa. G: AVT-ir gigantocellular and magnocellular neurons of the PMin a migrated position dorsolateral to the caudal PM, entering the

medial margin of the caudal telencephalon. Putative terminals (bead-ed AVT-ir fibers) are seen in the PM (small arrow, left) and telenceph-alon (large arrow, right). H: Beaded fibers (e.g., small arrows) and anAVT-ir neuron (large arrow) in the ventral tuberal region. As shown,fibers in this area are often closely juxtaposed to, or embedded in, thePHT. Inset: An AVT-ir neuron in the ventral tuberal hypothalamus.I: Fine-caliber AVT-ir fibers and putative terminals (e.g., arrows) inthe anterior tuberal nucleus; medial is to the left. J: Fine-caliberAVT-ir fibers and putative terminals (e.g., arrows) in the paralemnis-cal midbrain tegmentum; the lateral lemniscus lies just medial (left)to the area shown. K: High-magnification photomicrograph of AVT-irfibers and putative terminals (e.g., arrows) in the isthmal paraven-tricular cell group. L: Beaded AVT-ir fibers in the area postrema (AP)extend laterally to join the lateral AVT-ir fiber bundle (small arrow) ofthe medulla and spinal cord. The large arrow shows the position ofputative terminals juxtaposed to a band of small cells, which surroundthe dorsolateral margin of the rostral sonic motor nucleus (SMN).M: High-magnification view of AVT-ir fibers extending from the AP(asterisks in L and M denote corresponding positions). For otherabbreviations, see list. Scale bars 5 200 mm in A,C,L; 25 mm inB,F,I,J; 10 mm in D,H inset; 500 mm in E; 50 mm in A inset, C inset,G,K,M; 100 mm in H.

Fig. 4. A–C: Oscillograms of fictive vocalizations evoked by elec-trical stimulation of the anterior tuberal nucleus in a single type Imale midshipman: short-duration, grunt-like output (A), long-duration vocal-motor activity typical of a hum (B), and a growl-likefictive vocalization showing characteristic amplitude and frequencymodulations (C). D,E: Oscillograms of fictive vocalizations evoked byelectrical stimulation of the paralemniscal midbrain region in a singletype I male midshipman: grunt-like fictive vocalizations (D) and afictive hum (E). (For natural vocalizations, see Brantley and Bass,1994, and Bass et al., 1999.) Scale bar 5 100 msec (applies to A–E).


gesting that AVT inhibited call initiation. SignificantSpearman rank correlations for number of bursts (correct-ed for ties; i.e., correlations between peptide amount andvocal response as a percentage of baseline) were obtainedat both 10 minutes after treatment (Fig. 7A) and 30 min-utes after treatment (r 5 20.641, P 5 0.033), but not at 2minutes after treatment (r 5 20.273, P 5 0.344). Signif-icant correlations between dose and response latency werealso obtained at two time points (2 minutes post, r 50.764, P 5 0.030; 10 minutes post, see Fig. 7B). Effects onlatency were no longer apparent at the 30-minute timepoint. One subject (1,000 ng AVT) was dropped from thelatency analyses as baseline response overlapped with the

stimulation. This deletion produced a conservative bias,as this subject’s mean response latency increased to ap-proximately 200 msec at posttreatment time points.

In sharp contrast to the number of bursts and latencymeasures, parameters associated with vocal-motor pat-terning (duration and frequency) were not sensitive toAVT administration. Burst duration: 2 minutes post, r 50.000, P 5 0.999; 10 minutes post, see Fig. 7C; 30 minutespost, r 5 20.382, P 5 0.205. Frequency: 2 minutes post,r 5 20.231, P 5 0.487; 10 minutes post, see Fig. 7D; 30minutes post, r 5 20.314, P 5 0.346.

Isotocin treatment (1,000 ng; n 5 3) produced no observ-able effects relative to control (n 5 3; all Mann-Whitney P

Fig. 5. Vocal-acoustic regions labeled by iontophoresis of dextranbiotin into a neurophysiologically identified vocal site in theparalemniscal midbrain tegmentum (PL). A: Injection site in the PL.Fibers from the injection extend along the margin of the laterallemniscus (ll) and retrogradely label neurons (arrows) in the isthmal

nucleus (IS). B,C: Biotin-labeled cells, fibers, and putative terminals(e.g., arrows) in the isthmal nucleus (B) and hindbrain paraventricu-lar region (HP; C). D: Biotin-labeled fibers and putative terminals(e.g., arrows) in the nucleus of the lateral lemniscus (nll). Scale bars 5100 mm in A; 50 mm in B–D.


values . 0.10); e.g., isotocin treatment values at 10 min-utes post are (shown are median and range): number ofbursts (111/100–229% baseline), response latency (103/81–348% baseline), burst duration (129/38–169% base-line), frequency (21.4/22.3 to 0.7 Hz difference from base-line); see Figure 7 for control data. At the 10-minute timepoint, responses to 1,000 ng isotocin and 1,000 ng AVTwere significantly different for the number of bursts (P ,0.05, Mann-Whitney; see Fig. 7 for AVT data). These datasuggest that AVT effects on call initiation were specificand did not occur by action at the isotocin receptor. Sim-ilar contrasts for response latency are unfortunately notinformative, because data for this parameter are availablefor only a single AVT subject at 10 min post (see Fig. 7).

Finally, no effects of the AVP antagonist were obtained(0, 0.1, 10, 1,000 ng; n 5 3, 2, 3, and 2, respectively; allSpearman rank tied P values . 0.10; see Table 1 for10-minute posttreatment data).

DISCUSSION

We have here demonstrated that AVT-ir fibers and pu-tative terminals are located in candidate brain regions forvocal-acoustic integration in the plainfin midshipman fish.Candidate regions receive afferents from a variety ofacoustic structures (Bass et al., 2000; Goodson and Bass,unpublished observations) and electrical stimulation ofcandidate areas evokes rhythmic vocal-motor activity thatprecisely mimics natural vocalizations (i.e., grunts, hums,and growls). We have additionally shown that AVT ad-ministration to one of these sites, the paralemniscal mid-brain tegmentum, selectively modulates vocal-motor pa-rameters associated with fictive call initiation. We firstcompare the anatomic results with those obtained forother fishes and tetrapods and conclude with consider-

ations of AVT/AVP innervation of vocal-acoustic regionsand AVT/AVP modulation of vocal behavior.

Comparison with AVT immunoreactivity inother vertebrate taxa

Fishes. AVT immunoreactivity in the POA and neu-rohypophysis of teleosts has been extensively investi-gated, and the organization of AVT-ir cells reported herelargely conforms to an evolutionarily conserved patternexhibited in other species (Gill et al., 1977; Goossens et al.,1977b; Reaves and Hayward, 1980; Schreibman and Halp-ern, 1980; van den Dungen et al., 1982; Cumming et al.,1982; Batten, 1986; Yulis and Lederis, 1987; Moons et al.,1988, 1989b; Olivereau et al., 1988; Batten et al., 1990;Holmqvist and Ekstrom, 1991, 1995; Foran and Bass,1998). Similar data on hypophysial innervation are avail-able for agnathans (Goossens et al., 1977a; Nozaki andGorbman, 1983), dipnoids (Goossens et al., 1978), andelasmobranchs (Vallarino et al., 1990). Available informa-tion on AVT receptor distribution (Moons et al., 1989a) isalso generally consistent with the present findings. How-ever, a couple points of divergence are noteworthy.

First is the presence of a small number of AVT-ir so-mata in the ventral tuberal region of the anterior hypo-thalamus. Some neurons of the ventral tuberal region inPorichthys project to the pituitary (Marchaterre and Bass,unpublished observations); thus, these neurons could po-tentially be hypophysiotropic. In addition, these AVT-irneurons may influence vocal behavior by means of localprojections, descending projections, or both (Goodson andBass, 2000; Goodson and Bass, unpublished observations).

Second, in contrast to a previous investigation in mid-shipman in which an AVP antibody was used (Foran andBass, 1998), we identified no AVT-ir somata in the gan-glion of the terminal nerve or in the pineal. Given that theconditions of study were similar for these investigations,the difference is likely due to differential specificities ofthe antibodies used. No labeled fibers in the previousinvestigation were observed to extend far from the gan-glion of the terminal nerve or pineal; thus, the AVT-irfibers reported here for the forebrain and midbrain areconsidered to be mainly of preoptic origin.

The dense innervation of the anterior tuberal nucleusfound for Porichthys (present study) is not consistent withany other reports to date. An AVT-ir innervation of thenucleus of the posterior recess and the nucleus of thelateral recess (i.e, ventral and dorsal periventricular hy-pothalamic nuclei of Braford and Northcutt, respectively,see Fig. 2E) was not found for Porichthys but is reportedfor Poecilia latipinna (Batten et al., 1990) and Salmogairdneri (van den Dungen et al., 1982). In addition, thecaudal pathways and putative brainstem terminationzones of fibers are not fully described for other species; inthe midshipman these AVT-ir fibers seem to terminatediffusely in the dorsal tegmentum and periaqueductal ar-eas of the caudal midbrain, in paraventricular zones of theisthmus and hindbrain, and in the isthmal nucleus.

The dorsal mesencephalic pathway is found in both typeI and type II male midshipman but is virtually absent inthree of the four females examined. At present, othersexual dimorphisms in teleost AVT systems are limited topreoptic neuronal phenotype and preoptic gene expression(Grober and Sunobe, 1996; Hiraoka et al., 1997; Foran andBass, 1998, 1999; Ota et al., 1999). However, quantitativestudy of AVT-ir fiber pathways in teleosts may reveal

Fig. 6. A schematic sagittal view of the midshipman brain show-ing arginine vasotocin-immunoreactive (AVT-ir) pathways (black ar-rows), which originate almost exclusively from the preoptic area(POA), and projections of the auditory torus (TS; gray stippled ar-rows). Projection density is denoted by arrow size. AVT-ir fibers in-nervate six vocally active brain regions: the POA, ventral tuberalhypothalamus (vT); anterior tuberal hypothalamus (AT); paralemnis-cal midbrain tegmentum (PL) and the adjacent isthmal nucleus andnucleus of the lateral lemniscus (IS/nll); periaqueductal gray (PAG);and paraventricular cell groups of the hindbrain and isthmus (HP/IP).AVT-ir fibers and putative terminals are also found in the area pos-trema (AP) at the rostral aspect of the sonic motor nucleus (SMN); thesource of these fibers is not yet specified. The PL/IS/nll and AT areprimary candidate sites for vocal-acoustic integration, as AVT-ir andTS projections converge in these sites. The vT also receives projectionsfrom the auditory thalamus (central posterior nucleus; not shown).For other abbreviations, see list.


additional sex differences, given that sexual dimorphismsare widely reported for tetrapods (Moore and Lowry,1998). The present dimorphism may not be relevant todifferences in vocal behavior between females and type I

males, as type II sneaker males share the AVT-ir dorsaltract profile of type I males but are female-typical in theirvocal-motor morphology (e.g., Bass and Marchaterre,1989; Bass and Baker, 1990; Bass et al., 1996), vocal

TABLE 1. Effects of an AVP Antagonist on Vocal-Motor Response to Electrical Stimulation of the Paralemniscal Midbrain 10 Min Post Treatment1

Response parameter

Antagonist dose

r0 (n 5 3) 0.1 (n 5 2) 10 (n 5 3) 1000 (n 5 2)

# of bursts 100/93,125 132/114,150 93/85,108 75/56,94 2.502Response latency 93/30,107 118/111,125 113/103,127 186/84,289 .406Burst duration 78/62,112 109/93,123 98/93,140 167/93,241 .406Frequency change 1.6/22.4,3.0 0.3/21.0,1.5 0.8/20.4,1.6 20.1/21.1,0.9 2.251

1 Scores are shown as percentage baseline response, except frequency change, which is given as Hz change from baseline; given are median/minimum, maximum. Spearman rankr values are given for correlations between antagonist dose and response parameters (all P . 0.10).

Fig. 7. Effects of arginine vasotocin (AVT) on vocal-motor responseto electrical stimulation of the paralemniscal midbrain. Data werecollected 10 minutes after AVT administration and are represented asSpearman rank correlations (corrected for ties) between AVT dose (0,0.1, 10, and 1,000 ng; n 5 3, 3, 4, and 3, respectively) and responseparameters. The location of peptide delivery and stimulation corre-sponds to the location of paralemniscal AVT-ir fibers and putativeterminals shown in Figure 2G (i.e., medial to the lateral lemniscus atcaudal midbrain levels); a marked stimulation site is shown in Figure5A. AVT produced significant effects on response parameters associ-ated with call initiation: (A) Number of vocal-motor bursts obtained

and (B) call latency; both parameters are represented as a percentageof pretreatment (baseline) response. AVT produced no effect on pa-rameters associated with vocal-motor patterning: (C) Burst duration(shown as a percentage of baseline response) and (D) fundamentalfrequency (F0) of fictive vocalization (shown as post F0 - baseline F0).Each subject is represented by an open circle. Subjects not respondingat a given time point are excluded from latency and frequency analyses;an additional subject (1,000 ng) was deleted from latency measures dueto vocal response overlap with electrical stimulation during baselinemeasurements; this is a conservative deletion, as posttreatment laten-cies exceed 200 msec. See text for results from other time points.


behavior (Brantley and Bass, 1994; Bass et al., 1999), andforebrain neuropeptide modulation of vocalization (Good-son and Bass, 2000). Similarly, the dimorphic pattern ofPOA AVT-ir neuron size and relative neuron number (Fo-ran and Bass, 1998; see introduction) does not reflectdivergence in vocal-motor phenotype.

A particularly striking feature of Porichthys is the pres-ence of an extremely dense plexus of heavily beadedAVT-ir fibers that extends along the dorsolateral marginof the sonic motor nucleus (i.e., at caudal levels of thefourth ventricle) into the area postrema. These fibers donot make direct contact with the somata of sonic motoneu-rons and virtually no label is present more ventrally in thedendritic arbor field of the vocal pacemaker neurons (formorphology of pacemaker neurons, see Bass and Baker,1990; Bass et al., 1994). However, these fibers do notappear continuous with other AVT-ir fibers, and theirsource, therefore, remains to be determined. Similar labelis not reported for other teleosts, although Poecilia latip-inna does exhibit AVT-ir fibers in the nucleus of the soli-tary tract (Batten et al., 1990).

Tetrapods. AVT-ir neurons of the magnocellular andparvocellular preoptic nuclei in fish are considered homol-ogous, respectively, to AVT/AVP neurons of the supraopticand paraventricular nuclei of amniotes and to AVT neu-rons of the magnocellular and posterior POA in amphibi-ans (Lowry et al., 1997; Moore and Lowry, 1998). In addi-tion, tetrapods share numerous features of AVT/AVP-irfiber distributions, which are now shown to be present inPorichthys as well. Thus, across vertebrates classes, AVT/AVP -ir fibers are found in the POA, anterior hypothala-mus, and lateral hypothalamus. In addition, most speciesexhibit a fairly dense AVT/AVP-ir innervation of the peri-aqueductal midbrain (not recognized in birds) and adja-cent tegmentum, ventral tegmentum (particularly theventral tegmental area), and isthmal regions (periven-tricular isthmus, nucleus isthmi, locus coeruleus). Hind-brain label typically includes the lateral aspect of themedulla and a concentration of AVT/AVP -ir fibers inviscerosensory regions such as the nucleus of the solitarytract and/or the area postrema (no review is currentlyavailable, see these recent whole brain descriptions: Prop-per et al., 1992; Lakhdar-Ghazal et al., 1995; Lowry et al.,1997; Hilscher-Conklin et al., 1998; Panzica et al., 1999).Homologies between teleosts and tetrapods for specificAVT/AVP-innervated structures of the anterior and lat-eral hypothalamic regions are not firmly established,whereas the relationships between other regions contain-ing AVT/AVP-ir fibers are more clear (POA, periaqueduc-tal gray; Nieuwenhuys et al., 1998; Northcutt, 1981).

In the majority of tetrapod species across tetrapodclasses, AVT/AVP-ir fibers also innervate the amygdala,septum, and bed nucleus of the stria terminalis (or regionsof the dorsal diencephalon). Similarly, a few AVT-ir fibersin the midshipman are observed in the ventral and supra-commissural nuclei of the ventral telencephalon (proposedseptal and amygdalar homologues, respectively; North-cutt and Braford, 1980). This innervation in midshipmanis exceptionally sparse compared with most tetrapods,however, and AVT/AVP-ir pathways to these forebrainregions in tetrapods arise largely from cell groups outsideof the POA-AH.

AVT/AVP modulation of acousticcommunication: anatomic, physiological, and

behavioral evidence

Vocal-acoustic integration. As shown here, the mostdense concentrations of AVT-ir fibers in the midshipmanare located primarily in six neurophysiologically identifiedvocal regions: the ventral tuberal hypothalamus, anteriortuberal hypothalamus, anterior parvocellular preopticarea (PPa), periaqueductal midbrain, paralemniscal mid-brain tegmentum, and paraventricular cell groups of theisthmus and hindbrain. Four of these areas, the PPa,ventral tuberal hypothalamus, anterior tuberal hypothal-amus, and paralemniscal midbrain tegmentum, receivedirect or indirect projections from neurophysiologicallyidentified auditory sites within the torus semicircularis(Bass et al., 2000; Goodson and Bass, unpublished obser-vations), and all but the PPa, therefore, are consideredprimary candidate sites for AVT modulation of vocal-acoustic integration. Although the PPa likely participatesin vocal-acoustic integration as well (suggested by thedemonstrated importance of AVT in vocalization), thevocal-motor responses obtained from this region are sub-stantially less reliable and require much greater stimulusamplitudes. Finally, anatomic experiments confirm theconnectivity of AVT-ir vocal nuclei and cell groups to eachother, to the vocal pacemaker circuitry of the medulla, orboth (Goodson and Bass, unpublished observations). Thus,all AVT-ir candidate vocal-acoustic sites in the midship-man are also components of anatomically characterizedvocal-motor circuitry in this species.

A review of available studies shows that an associationof AVT/AVP-ir cells and fibers with vocal brain regions isa common feature of taxa that use acoustic communication(see the above references for general AVT/AVP-ir distri-butions in these groups and the following references forconsiderations of vocally active sites: mammals [Jurgensand Ploog, 1981]; anuran amphibians [Schmidt, 1966;Wada and Gorbman, 1977]; birds [Brown, 1971; Wild,1997]; fish [Demski and Gerald, 1972; Fine and Perini,1994; present study]; reptiles [Kennedy, 1975]). Vocallyactive sites such as the amygdala, septum, and POA arewell-recognized sites of sensory integration (Emerson andBoyd, 1999), and auditory-evoked activity is obtained inbrainstem vocal sites of anurans (pretrigeminal nucleus;Aitken and Capranica, 1984) mammals (periaqueductalgray and adjacent tegmentum, nucleus of the lateral lem-niscus, paralemniscal tegmentum; Kirzinger and Jurgens,1991; Covey, 1993; Metzner, 1996) and birds (nucleusintercollicularis; Cheng and Havens, 1993). Thus, AVT/AVP may in fact modulate vocal-acoustic integrationacross a broad range of vertebrates.

Vocal behavior and physiology. Modulation of vocal-ization is the most common behavioral effect of AVT/AVPobserved in vertebrates. AVT/AVP modulates vocaliza-tions in rats (Winslow and Insel, 1993), primates (Win-slow and Insel, 1991), oscine and non-oscine birds (Voor-huis et al., 1991; Maney et al., 1997; Castagna et al., 1998;Goodson, 1998a), fish (Goodson and Bass, 2000; presentstudy), and anuran amphibians (Diakow, 1978; Raimondiand Diakow, 1981; Boyd, 1992, 1994; Penna et al., 1992;Marler et al., 1995; Propper and Dixon, 1997; Chu et al.,1998; Semsar et al., 1998; Tito et al., 1999). The majorityof experiments have used intracerebroventricular or pe-ripheral administration of AVT/AVP; delivery into specific


neural sites has been conducted in only a small number ofexperiments (Goodson, 1998a; Goodson and Bass, 2000;present study).

The site(s) of AVT action remains to be determined formost species thus far examined, although the availableevidence suggests that AVT may modulate vocalizationwithin multiple neural loci (for discussion see Emersonand Boyd, 1999; Goodson, 1998a; Maney et al., 1997). Thisfinding is indeed the case for the plainfin midshipman,because evidence now shows that AVT produces divergenteffects on vocalization in forebrain and midbrain regionsof this species (Goodson and Bass, 2000; present study).Local AVT administration to the ventral tuberal hypothal-amus of type I males inhibits vocal-motor output producedby electrical stimulation of this site, primarily by reducingburst duration. These effects are reversed by a V1 antag-onist, indicating that forebrain AVT action is a necessarycomponent for the establishment of vocal-motor pattern-ing. In contrast, similar experiments conducted here inthe paralemniscal midbrain tegmentum reveal that AVTdoes not modulate vocal-motor patterning but selectivelyinhibits fictive call initiation (i.e., reduces the number ofvocal-motor bursts obtained and produces a concomitantincrease in response latency). These effects are not re-versed by an antagonist, suggesting that AVT action inthe paralemniscal site is not tonic or necessary for normalcall production. Therefore, we suggest that paralemniscalAVT likely inhibits vocalization in a conditional manner.

CONCLUSIONS

Findings across a broad range of vertebrate species im-plicate AVT/AVP in the integration of sensory and endo-crine stimuli and in the generation of complex sex-typicaland species-specific patterns of behavior. These neuropep-tides modulate vocalization in multiple vertebrate classes,yet little information is available on specific target areasfor AVT/AVP modulation of vocal-acoustic communica-tion. We have here shown that in the plainfin midship-man, AVT immunoreactivity is extensively associatedwith neurophysiologically identified vocal sites of the fore-brain, midbrain, and rostral hindbrain; three of these lociare candidate sites for vocal-acoustic integration. Further-more, AVT modulates different vocal-motor parameters intwo of these candidate sites, the POA-AH (burst duration,Goodson and Bass, 2000) and the paralemniscal midbraintegmentum (physiological indices of call initiation;present study). AVT also modulates vocalization differ-ently between morphs which express divergent reproduc-tive tactics (Goodson and Bass, 2000). Comparison ofAVT-ir in Porichthys with AVT/AVP immunoreactivity intetrapods reveals some extensive similarities, particularlyin the innervation of forebrain and midbrain regions in-volved in vocal communication. Thus, the midshipmanmay provide useful insights into the AVT/AVP modulationof vocal-acoustic communication across a wide range ofvertebrate taxa.

ACKNOWLEDGMENTS

The authors thank Matthew Weeg, Margaret Marchat-erre, and Robert Wyttenbach for extensive technical as-sistance and Jessica McKibben for field support. We alsothank two anonymous reviewers for their useful sugges-

tions. Logistical support provided by the University ofCalifornia Bodega Marine Laboratory. J.L.G. received apostdoctoral grant from the National Institutes of Health,and A.H.B. received a grant from the National ScienceFoundation.

REFERENCESAitken PG, Capranica RR. 1984. Auditory input to a vocal nucleus in the

frog Rana pipiens: hormonal and seasonal effects. Exp Brain Res 57:33–39.

Bass A, Baker R. 1990. Sexual dimorphisms in the vocal control system ofa teleost fish: morphology of physiologically identified neurons. J Neu-robiol 21:1155–1168.

Bass AH. 1992. Dimorphic male brains and alternative reproductive tacticsin a vocalizing fish. Trends Neurosci 15:139–145.

Bass AH. 1996. Shaping brain sexuality. Am Sci 84:352–363.Bass AH, Marchaterre MA. 1989. Sound generating. sonic. motor system in

a teleost fish (Porichthys notatus): sexual polymorphisms and generalsynaptology of sonic motor nucleus. J Comp Neurol 286:154–169.

Bass AH, Marchaterre MA, Baker R. 1994. Vocal-acoustic pathways in ateleost fish. J Neurosci 14:4025–4039.

Bass AH, Horvath BJ, Brothers EB. 1996. Nonsequential developmentaltrajectories lead to dimorphic vocal circuitry for males with alternativereproductive tactics. J Neurobiol 30:493–504.

Bass AH, Bodnar DA, Marchaterre MA. 1999. Complementary explana-tions for existing phenotypes in an acoustic communication system. InHauser MK, Konishi M, editors. The design of animal communication.Cambridge, MA: MIT Press. p 493–514.

Bass AH, Bodnar DA, Marchaterre MA. 2000. Midbrain acoustic circuitryin a vocalizing fish. J Comp Neurol 419:505–531.

Batten TFC. 1986. Ultrastructural characterization of neurosecretory fi-bres immunoreactive for vasotocin, isotocin, somatostatin, LHRH andCRF in the pituitary of a teleost fish, Poecilia latipinna. Cell Tissue Res244:661–672.

Batten TFC, Cambre ML, Moons L, Vandesande F. 1990. Comparativedistribution of neuropeptide-immunoreactive systems in the brain ofthe green molly Poecilia latipinna. J Comp Neurol 302:893–919.

Boyd SK. 1992. Sexual differences in hormonal control of release calls inbullfrogs. Horm Behav 26:522–535.

Boyd SK. 1994. Arginine vasotocin facilitation of advertisement calling andcall phonotaxis in bullfrogs. Horm Behav 28:232–240.

Braford MR Jr, Northcutt RG. 1983. Organization of the diencephalon andpretectum of the ray-finned fishes. In: Davis RE, Nortcutt RG, editors.Fish neurobiology. Ann Arbor: University of Michigan Press. p 117–164.

Brantley RK, Bass AH. 1994. Alternative male spawning tactics and acous-tic signals in the plainfin midshipman Porichthys notatus Girard (Te-leostei, Batrachoididae). Ethology 96:213–232.

Brown JL. 1971. An exploratory study of vocalization areas in the brain ofthe red-winged blackbird (Agelaius phoeniceus). Behaviour 34:91–127.

Castagna C, Absil P, Foidart A, Balthazart J. 1998. Systemic and intrace-rebroventricular injections of vasotocin inhibit appetitive and consum-matory components of male sexual behavior in Japanese quail. BehavNeurosci 112:233–250.

Cheng MF, Havens MH. 1993. Auditory-responsive units in the midbrainvocal nuclei in the ring dove (Streptopelia risoria). Brain Res Bull30:711–715.

Chu J, Marler CA, Wilczynski W. 1998. The effects of arginine vasotocin onthe calling behavior of male cricket frogs in changing social contexts.Horm Behav 34:248–261.

Cohen MJ, Winn HE. 1967. Electrophysiological observations on hearingand sound production in the fish, Porichthys notatus. J Exp Zool 165:355–370.

Covey E. 1993. Response properties of single units in the dorsal nucleus ofthe lateral lemniscus and paralemniscal zone of an echolocating bat.J Neurophysiol 69:842–859.

Cumming R, Reaves TAJ, Hayward JN. 1982. Ultrastructural immunocy-tochemical characterization of isotocin, vasotocin and neurophysinneurons in the magnocellular preoptic nucleus of the goldfish. CellTissue Res 223:685–694.

Demski LS, Gerald JW. 1972. Sound production evoked by electrical stim-ulation of the brain in toadfish (Opsanus beta). Anim Behav 20:507–513.


Demski LS, Gerald JW. 1974. Sound production and other behavioraleffects of midbrain stimulation in free-swimming toadfish, Opsanusbeta. Brain Behav Evol 9:41–59.

de Vries GJ, Miller MA. 1998. Anatomy and function of extrahypothalamicvasopressin systems in the brain. Prog Brain Res 119:3–20.

Diakow C. 1978. Hormonal basis for breeding behavior in female frogs:vasotocin inhibits the release call of Rana pipiens. Science 199:1456–1457.

Emerson SB, Boyd SK. 1999. Mating vocalizations of female frogs: controland evolutionary mechanisms. Brain Behav Evol 53:187–197.

Engelmann M, Wotjak Carsten T, Neumann I, Ludwig M, Landgraf R.1996. Behavioral consequences of intracerebral vasopressin and oxyto-cin: focus on learning and memory. Neurosci Biobehav Rev 20:341–358.

Fernandez-Llebrez P, Perez J, Nadales AE, Cifuentes M, Grondona JM,Mancera JM, Rodriguez EM. 1988. Immunocytochemical study of thehypothalamic magnocellular neurosecretory nuclei of the snake Natrixmaura and the turtle Mauremys caspica. Cell Tissue Res 253:435–445.

Fine ML, Perini MA. 1994. Sound production evoked by electrical stimu-lation of the forebrain in the oyster toadfish. J Comp Physiol A 174:173–185.

Foran CM, Bass AH. 1998. Preoptic AVT immunoreactive neurons of ateleost fish with alternative reproductive tactics. Gen Comp Endocrinol111:271–282.

Foran CM, Bass AH. 1999. Preoptic GnRH and AVT: axes for sexualplasticity in teleost fish. Gen Comp Endocrinol 116:141–152.

Foran CM, Myers DA, Bass AH. 1997. Modification of gonadotropin releas-ing hormone (GnRH) mRNA expression in the retinal-recipient thala-mus Gen Comp Endocrinol 106:251–264.

Gill VE, Burford GD, Lederis K, Zimmerman EA. 1977. An immunocyto-chemical investigation for arginine vasotocin and neurophysin in thepituitary gland and the caudal neurosecretory system of Catostomuscommersoni. Gen Comp Endocrinol 32:505–511.

Gonzalez A, Munoz A, Munoz M, Marin O, Smeets WJAJ. 1995. Ontogenyof vasotocinergic and mesotocinergic systems in the brain of the SouthAfrican clawed frog Xenopus laevis. J Chem Neuroanat 9:27–40.

Gonzalez A, Smeets WJAJ. 1992a. Comparative analysis of the vasotocin-ergic and mesotocinergic cells and fibers in the brain of two amphibi-ans, the anuran Rana ridibunda and the urodele Pleurodeles waltlii.J Comp Neurol 315:53–73.

Gonzalez A, Smeets WJAJ. 1992b. Distribution of vasotocin and mesotocin-like immunoreactivities in the brain of the South African clawed frogXenopus laevis. J Chem Neuroanat 5:465–479.

Gonzalez A, Smeets WJAJ. 1997. Distribution of vasotocin- and mesotocin-like immunoreactivities in the brain of Typhlonectes compressicauda(Amphibia, Gymnophiona): Further assessment of primitive and de-rived traits of amphibian neuropeptidergic systems Cell Tissue Res287:305–314.

Goodson JL. 1998a. Territorial aggression and dawn song are modulatedby septal vasotocin and vasoactive intestinal polypeptide in male fieldsparrows (Spizella pusilla). Horm Behav 34:67–77.

Goodson JL. 1998b. Vasotocin and vasoactive intestinal polypeptide mod-ulate aggression in a territorial songbird, the violet-eared waxbill (Es-trildidae: Uraeginthus granatina). Gen Comp Endocrinol 111:233–244.

Goodson JL, Bass AH. 2000. Forebrain peptide modulation of sexuallypolymorphic vocal circuitry. Nature 403:769–772.

Goossens N, Dierickx K, Vandesande F. 1977a. Immunocytochemical dem-onstration of the hypothalamohypophysial vasotocinergic system ofLampetra fluviatilis. Cell Tissue Res 177:317–324.

Goossens N, Dierickx K, Vandesande F. 1977b. Immunocytochemical lo-calization of vasotocin and isotocin in the preopticohypophysial neuro-secretory system of teleosts. Gen Comp Endocrinol 32:371–375.

Goossens N, Dierickx K, Vandesande F. 1978. Immunocytochemical studyof the neurohypophysial hormone producing system of the lungfish,Protopterus aethiopicus. Cell Tissue Res 190:69–77.

Grober MS, Fox SH, Laughlin C, Bass AH. 1994. GnRH cell size andnumber in a teleost fish with two male reproductive morphs: sexualmaturation, final sexual status and body size allometry. Brain BehavEvol 43:61–78.

Grober MS, Sunobe T. 1996. Serial adult sex change involves rapid andreversible changes in forebrain neurochemistry. Neuroreport 7:2945–2949.

Hilscher-Conklin C, Conlon JM, Boyd SK. 1998. Identification and local-ization of neurohypophysial peptides in the brain of a caecilian am-phibian, Typhlonectes natans (Amphibia: Gymnophiona). J Comp Neu-rol 394:139–151.

Hiraoka S, Ando H, Ban M, Ueda H, Urano A. 1997. Changes in expressionof neurohypophysial hormone genes during spawning migration inchum salmon, Oncorhynchus keta. J Mol Endocrinol 18:49–55.

Holmqvist BI, Ekstrom P. 1991. Galanin-like immunoreactivity in thebrain of teleosts: distribution and relation to substance P, vasotocin,and isotocin in the Atlantic salmon (Salmo salar). J Comp Neurol306:361–381.

Holmqvist BI, Ekstrom P. 1995. Hypophysiotrophic systems in the brain ofthe Atlantic salmon: neuronal innervation of the pituitary and theorigin of pituitary dopamine and nonapeptides identified by means ofcombined carbocyanine tract tracing and immunocytochemistry.J Chem Neuroanat 8:125–145.

Jurgens U, Ploog D. 1981. On the control of mammalian vocalization.Trends Neurosci 4:135–137.

Kennedy MC. 1975. Vocalization elicited in a lizard by electrical stimula-tion of the midbrain. Brain Res 91:321–325.

Kirzinger A, Jurgens U. 1991. Vocalization-correlated single-unit activityin the brainstem of the squirrel monkey. Exp Brain Res 84:545–560.

Lakhdar-Ghazal N, Dubois Dauphin M, Hermes MLHJ, Buijs RM, Beng-elloun WA, Pevet P. 1995. Vasopressin in the brain of a desert hiber-nator, the Jerboa (Jaculus orientalis): presence of sexual dimorphismand seasonal variation. J Comp Neurol 358:499–517.

Lowry CA, Richardson CF, Zoeller TR, Miller LJ, Muske LE, Moore FL.1997. Neuroanatomical distribution of vasotocin in a urodele amphib-ian (Taricha granulosa) revealed by immunohistochemical and in situhybridization techniques. J Comp Neurol 385:43–70.

Maney DL, Goode CT, Wingfield JC. 1997. Intraventricular infusion ofarginine vasotocin induces singing in a female songbird. J Neuroendo-crinol 9:487–491.

Marler CA, Chu J, Wilczynski W. 1995. Arginine vasotocin injection in-creases probability of calling in cricket frogs, but causes call changescharacteristic of less aggressive males. Horm Behav 29:554–570.

McKibben JR, Bass AH. 1998. Behavioral assessment of acoustic param-eters relevant to signal recognition and preference in a vocal fish. JAcoust Soc Am 104:3520–3533.

Metzner W. 1996. Anatomical basis for audio-vocal integration in echolo-cating horseshoe bats. J Comp Neurol 368:252–269.

Moons L, Cambre M, Marivoet S, Batten TFC, Vanderhaeghen JJ, OllevierF, Vandesande F. 1988. Peptidergic innervation of the adrenocortico-tropic hormone, ACTH, and growth hormone Gh-producing cells in thepars distalis of the sea bass, Dicentrarchus labrax. Gen Comp Endo-crinol 72:171–180.

Moons L, Cambre M, Batten TFC, Vandesande F. 1989a. Autoradiographiclocalization of binding sites for vasotocin in the brain and pituitary ofthe sea bass Dicentrarchus labrax. Neurosci Lett 100:11–16.

Moons L, Cambre M, Ollevier F, Vandesande F. 1989b. Immunocytochem-ical demonstration of close relationships between neuropeptidergicnerve fibers and hormone-producing cell types in the adenohypophysisof the sea bass, Dicentrarchus labrax. Gen Comp Endocrinol 73:270–283.

Moore FL. 1992. Evolutionary precedents for behavioral actions of oxytocinand vasopressin. Ann N Y Acad Sci 652:156–165.

Moore FL, Lowry CA. 1998. Comparative neuroanatomy of vasotocin andvasopressin in amphibians and other vertebrates. Comp BiochemPhysiol C 119:251–260.

Nieuwenhuys R, ten Donkelaar HJ, Nicholson C, editors. 1998. The centralnervous system of vertebrates. New York: Springer-Verlag.

Northcutt RG. 1981. Evolution of the telencephalon in nonmammals. AnnuRev Neurosci 4:301–330.

Northcutt RG, Braford MR Jr. 1980. New observations on the organizationand evolution of the telencephalon of actinopterygian fishes. In: Ebbes-son SOE, editor. Comparative neurology of the telencephalon. NewYork: Plenum. p 41–98.

Nozaki M, Gorbman A. 1983. Immunocytochemical localization of soma-tostatin and vasotocin in the brain of the Pacific hagfish, Eptatretusstouti. Cell Tissue Res 229:541–550.

Olivereau M, Moons L, Olivereau J, Vandesande F. 1988. Coexistence ofcorticotropin-releasing factor-like immunoreactivity and vasotocin inperikarya of the preoptic nucleus in the eel. Gen Comp Endocrinol70:41–48.

Ota Y, Ando H, Ueda H, Urano A. 1999. Differences in seasonal expressionof neurohypophysial hormone genes in ordinary and precocious malemasu salmon. Gen Comp Endocrinol 116:40–48.

Panzica GC, Plumari L, Garcia-Ojeda E, Deviche P. 1999. Central


vasotocin-immunoreactive system in a male passerine bird (Juncohyemalis). J Comp Neurol 409:105–117.

Penna M, Capranica RR, Somers J. 1992. Hormone-induced vocal behaviorand midbrain auditory sensitivity in the green treefrog Hyla cinerea.J Comp Physiol A 170:73–82.

Propper CR, Dixon TB. 1997. Differential effects of arginine vasotocin andgonadotropin-releasing hormone on sexual behaviors in an anuranamphibian. Horm Behav 32:99–104.

Propper CR, Jones RE, Lopez KH. 1992. Distribution of arginine vasotocinin the brain of the lizard Anolis carolinensis. Cell Tissue Res 267:391–398.

Raimondi D, Diakow C. 1981. Sex dimorphism in responsiveness tohormonal induction of female behavior in frogs. Physiol Behav 27:167–170.

Reaves TA, Hayward JN. 1980. Functional and morphological studies ofpeptide-containing neuroendocrine cells in goldfish hypothalamus.J Comp Neurol 193:777–788.

Rose JD, Kinnaird JR, Moore FL. 1995. Neurophysiological effects ofvasotocin and corticosterone on medullary neurons: implications forhormonal control of amphibian courtship behavior. Neuroendocrinol-ogy 62:406–417.

Schlinger BA, Creco C, Bass AH. 1999. Aromatase activity in the hindbrainvocal control region of a teleost fish: divergence among males withalternative reproductive tactics. Proc R Soc Lond B 266:131–136.

Schmidt RS. 1966. Central mechanisms of frog calling. Behaviour 26:251–285.

Schreibman MP, Halpern LR. 1980. The demonstration of neurophysin andarginine vasotocin by immunocytochemical methods in the brain andpituitary gland of the platyfish, Xiphophorus maculatus. Gen CompEndocrinol 40:1–7.

Semsar K, Klomberg KF, Marler C. 1998. Arginine vasotocin increasescalling-site acquisition by nonresident male grey treefrogs. Anim Be-hav 56:983–987.

Smeets WJAJ, Sevensma JJ, Jonker AJ. 1990. Comparative analysis ofvasotocin-like immunoreactivity in the brain of the turtle Pseudemysscripta elegans and the snake Python regius. Brain Behav Evol 35:65–84.

Stoll CJ, Voorn P. 1985. The distribution of hypothalamic and extrahypo-thalamic vasotocinergic cells and fibers in the brain of a lizard Gekkogecko presence of a sex difference. J Comp Neurol 239:193–204.

Striedter GF. 1990a. The diencephalon of the channel catfish, Ictaluruspuntatus. I. Nuclear organization. Brain Behav Evol 36:329–354.

Striedter GF. 1990b. The diencephalon of the channel catfish, Ictaluruspuntatus. II. Retinal, tectal, cerebellar and telencephalic connections.Brain Behav Evol 36:355–377.

Thepen T, Voorn P, Stoll CJ, Sluiter AA, Pool CW, Lohman AHM. 1987.Mesotocin and vasotocin in the brain of the lizard Gekko gekko: animmunocytochemical study. Cell Tissue Res 250:649–656.

Tito MB, Hoover MA, Mingo AM, Boyd SK. 1999. Vasotocin maintainsmultiple call types in the gray treefrog, Hyla versicolor. Horm Behav36:166–175.

Vallarino M, Viglietti Panzica C, Panzica GC. 1990. Immunocytochemicallocalization of vasotocin-like immunoreactivity in the brain of the car-tilaginous fish Scyliorhinus caniculus. Cell Tissue Res 262:507–514.

van den Dungen HM, Buijs RM, Pool CW, Terlou M. 1982. The distributionof vasotocin and isotocin in the brain of the rainbow trout. J CompNeurol 212:146–157.

Voorhuis TAM, De Kloet ER, De Wied D. 1991. Effect of a vasotocin analogon singing behavior in the canary. Horm Behav 25:549–559.

Wada M, Gorbman A. 1977. Mate calling induced by electrical stimulationin freely moving leopard frogs Rana pipiens. Horm Behav 9:141–149.

Wild JM. 1997. Neural pathways for the control of birdsong production.J Neurobiol 33:653–670.

Winslow J, Insel TR. 1991. Vasopressin modulates male squirrel monkeys’behavior during social separation. Eur J Pharmacol 200:95–102.

Winslow JT, Insel TR. 1993. Effects of central vasopressin administrationto infant rats. Eur J Pharmacol 233:101–107.

Young LJ, Wang ZX, Insel TR. 1998. Neuroendocrine bases of monogamy.Trends Neurosci 21:71–75.

Yulis CR, Lederis K. 1987. Co-localization of the immunoreactivities ofcorticotropin-releasing factor and arginine vasotocin in the brain andpituitary system of the teleost Catostomus commersoni. Cell Tissue Res247:267–273.

Zoeller RT, Moore FL. 1986. Correlation between immunoreactive vasotocinin optic tectum and seasonal changes in reproductive behaviors of malerough-skinned newts Taricha granulosa. Horm Behav 20:148–154.

Zoeller RT, Moore FL. 1988. Brain arginine vasotocin concentrations re-lated to sexual behaviors and hydromineral balance in an amphibian.Horm Behav 22:66–75.
