secondary connections of the dorsal and ventral facial lobes in a teleost fish, the rockling...

THE JOURNAL OF COMPARATIVE NEUROLOGY 370415-426 (1996)

Secondary Connections of the Dorsal and Ventral Facial Lobes in a Teleost Fish, the

Rockling (Ciliata mustela)

KURT KOTRSCHAL AND THOMAS E. FINGER Konrad-Lorenz-Forschungsstelle fur Ethologie, A-4645 Grunau, Austria (K.K.); Department

of Cellular and Structural Biology, University of Colorado Health Sciences Center, Denver, Colorado 80262 (T.E.F.)

ABSTRACT In the rockling, Ciliata mustela (Teleostei), a portion of the dorsal fin is a specialized

chemosensory organ possessing solitary chemoreceptor cells innervated by a recurrent branch of the facial nerve. Previous studies have demonstrated that the specialized solitary chemoreceptor cell system is represented in the dorsal segment of the medullary facial lobe (DFL), whereas the taste buds in the remainder of the facial-nerve-innervated skin are represented in the ventral division of the lobe (VFL). The carbocyanine dye DiI was used to investigate the secondary and higher order brain connections of these two distinct subdivisions of the facial lobe. Both segments of the facial lohe sent fihers into the contralateral DFL via a dorsocaudal facial commissure and to the contralateral vagal lobes and VFL via fibers arching ventrally through the reticular formation. Ascending fibers from both facial lobe segments were traced into the secondary gustatory nucleus and into the lateral superficial facial nucleus, a small area in the dorsolateral brainstem laterally adjacent to the nucleus medialis of the octavolateral complex. Additionally, the VFL had reciprocal connections with a newly described nucleus adjacent to the incoming facial nerve root. Both DFL and VFL had descending fibers reaching two portions of the funicular nuclear complex, although the VFL contribution to this area is far more extensive than the DFL input. Thus, substantial overlap exists in the connections of the two facial subsystems; i.e., the solitary chemoreceptor information is not processed in nuclei distinct from those making up the usual gustatory lemniscus. o 1996 Wiley-Liss, Inc

Indexing terms: solitary chemosensory cells, taste, nucleus of the solitary tract, parabrachial nucleus, pontine taste area

Virtually all fishes possess three specialized, independent external chemosensory systems: olfactory, gustatory, and solitary chemoreceptive (Parker, 1912). Each of these in- volves a different type of receptor cell and appears chemi- cally tuned to different spectra of chemicals (Kotrschal, 1991). The skin of fishes hosts two different types of epidermal chemosensory cells: those within taste buds (TBs) and others, the solitary chemosensory cells (SCCs), scattered among unspecialized epidermal cells (Whitear, 1965, 1971). The afferent innervation of external TBs is from facial nerves (Kotrschal et al. 19931, whereas SCCs can be innervated by either facial or spinal nerves. In all teleosts studied to date, the gustatory components of the facial nerve terminate in the facial lobe, which is the rostral end of the medullary visceral sensory column (Herrick, 1905; Finger, 1993). Evidence on central representation of the SCC system is available from only two groups of teleosts, the sea robins (Triglidae; Herrick, 1906; Finger, 1982) and the rocklings (Gadidae; Kotrschal, 1991; Whi-

tear, 1991). In the former, SCCs in the pectoral fin rays (Whitear, 1971) are innervated by spinal nerves (Finger, 1982; Morrill, 1895). Central pathways devoted to this system are similar to ascending spinal systems in other vertebrates (Finger, 1983). In the rocklings (Gadidae), the anterior dorsal fin vibratile rays contain millions of SCCs but no TBs (Kotrschal et al., 1984); these SCCs are innervated from a component of the recurrent facial nerve (Fig. 1; Kotrschal andmitear, 1988; Whitear and Kotrschal, 1988). The branch of the recurrent facial nerve that innervates the SCC system of the dorsal fin terminates in a distinct, dorsal portion of the facial lobe, which receives no other peripheral input. In contrast, the mixed-function (gustatory + SCC) branches of the facial nerve terminate in

Accepted December 22,1995. Address reprint requests to Dr. Thomas E. .Finger, Dept. Cellular and

Structural Biology, University of Colorado Medical School B-111, 4200 E. Ninth Ave., Denver, CO 80262. E-mail: [email protected]

O 1996 WILEY-LISS, INC.

416

the ventral portion of the facial lobe (Fig. 2). Electrophysi- ologic study showed that the SCCs preferentially respond to heterospecific fish body mucus and show little overlap in stimulus spectrum with TBs (Peters et al., 1987, 1991). Behaviorally, the taste buds presumably are used for feeding, as in other fishes (Herrick, 1904), whereas the SCCs in this species are reportedly involved in predator detection (Kotrschal et al., 1984).

Nothing is known about the differences in central secondary and higher order connections of the TB system vs. the SCC system. The rockling facial lobe provides a unique opportunity to investigate the secondary connections of the dorsal facial lobe (DFL), which receives only SCC input, compared to the ventral facial lobe (WL), wherein the afferent fibers from TBs and SCCs of the body terminate. Hence, the question raised herein is whether the secondary brain connections of the two facial chemosensory systems are similar despite different functional contexts.

K. KOTRSCHAL AND T.E. FINGER

MATERIALS AND METHODS Rocklings, Ciliata mustela, were trapped in the Dutch

Waddenzee, at Zeeland, and kept for 3-5 days in tanks with running seawater. The fish were deeply anaesthetized by immersion in a tricaine methanesulfonate solution (MS- 222; approximately 1:10,000) and perfused via the heart with a 4% paraformaldehyde solution in 0.1 M phosphate buffer. Perfused specimens were stored in the same fixative at 4°C for up to 12 months (Finger and Bottger, 1990).

Prior to application of the dye, h e brains were removed and washed in 0.1 M phosphate buffer a t 4°C for 24 hours. Undcr thc dissection microscope, the specimen was placed on a glass coverslip and covered with 2% agar solution, with only the facial lobe area exposed from the agar cover. The agar covering prevents the inadvertent contamination of

2Gc 2Gn 2Gt ADF BC Cb-s cc cf CrC DFL dH DLF EG Fn iaf ICP LL LSFn MLF N. VII OLA OLC PD PFn RF RFN SPC SPV Tel TeO VFL VFn VL

Abbreviations

intrinsic commissure of 2Gn superior secondary gustatory nucleus secondary gustatory tract anterior dorsal fin brachium conjunctivum cerebellar stump remaining after removal of cerebellum corpus cerebelli commissural fibers cerebellar crest dorsal facial lobe dorsal horn of the spinal cord dorsolateral fasciculus eminentia granularis funicular nucleus internal arcuate fibers inferior cerebellar peduncle lateral lemniscus lateral secondary facial nucleus medial longitudinal fasciculus facial nerve octavolateral area octavolateral complex posterior dorsal ramus of the RFN perifacial nucleus reticular formation recurrent facial nerve spinal cord spinal trigeminal tract and nucleus telencephalon optic tectum ventral facial lobe ventral funicular nucleus vagal lobe

Fig. 1. Sketch of a preparation of Ciliata mustela with the skin removed and the skull opened. Parts of the brain are shown along with major facial nerve trunks (RFN; solid lines) supplying the barbels and, via a posterior dorsal ramus (PD; broken lines), the anterior dorsal fin (ADF) as well as the trunk skin. Aposterior ventral ramus supplying the pectoral and pelvic fins is not shown.

the surface of the brain with unwanted dye crystals. A few minute crystals of DiI were applied to the DFL or VFL selectively by glass pipettes of 10-50 km tip diameters (Honig and Hume, 1989). In the case of VFL applications, the DFL, which dorsally caps the VFL (Kotrschal and Whitear, 1988), was removed, and DiI was applied to the VFL directly through its exposed dorsal surface. Because the efferent fibers of the DFL exit only along the medial and lateral edges of the VFL, the dye could be applied to the VFL without involving the DFL output system.

Following application of the dye to either the VFL or the DFL, the specimen was covered again with an agar coat and placed in a jar with fixative for 3 weeks to 4.5 months at room temperature or, to speed up diffusion, at 37°C. In total eight brains were labelled, four of which received DiI into one DFL, the other four into the VFL. In other cases, the crystals of DiI were applied to the dorsal spinal enlargement found in gadids and that was determined to have differen- tial connections with the two facial lobules.

Before sectioning, the tissue was embedded in egg yolk; after washes in 0.1 M phosphate buffer, specimens were immersed in fresh egg yolk for up to 1 hour, put on a glass coverslip, and surrounded by prefrozen and thawed egg yolk from a syringe. These blocks were again fixed for up to 24 hours, trimmed, glued to the Vibratome stage, covered with 0.1 M phosphate buffer, and cut at 50 km. Sections

FACIAL LOBE CONNECTIONS IN ROCKLING 417

RESULTS Terminology

Three small, hitherto unnamed areas consistently received labeled fibers and contained labeled perikarya. One, caudolateral of the nucleus medius of the octavolateral complex, immediately ventral to the caudal portion of the molecular layer of the cerebellar crest (Figs. 3B-D, 4B,B’), is provisionally termed the lateral secondary facial nucleus (LSFn). It is unclear based on cytology of the area whether the LSFn is a distinct nucleus or merely a division of the adjacent n. medialis of the octavolateral complex. A second hitherto unnamed nucleus, provisionally called the perifa- cia1 nucleus (PFn; Fig. 3C,D), is dorsally and laterally adjacent to the incoming facial nerve root as it curves rostrally and medially of the ventral facial lobe (Fig. 3B-D). The third newly named site, the ventral funicular nucleus (VFn; Fig. 3G), lies in the dorsolateral brainstem, caudal to the ventral facial lobe and ventral to the funicular nucleus of Herrick (1907).

Connections of the ventral facial lobe (gustatory system)

Application of DiI into the ventral facial lobe (VFL) labeled commissural and decussating fibers, ipsi- and contralateral ascending and descending fibers, as well as perikarya projecting into the VFL (Fig. 4A’-F’). The main ascending fiber system leaves the ventrolateral aspect of the VFL to form the ascending secondary gustatory tract (2Gt; Fig. 5) . Labeled fibers leave the main bundle to terminate within the ipsilateral PFn and ipsilateral LSFn areas (Fig. 6). The majority of secondary ascending fibers terminated the secondary gustatory nucleus (2Gn; Fig. 4A’); a smaller contingent of labeled fibers continued past the ipsilateral secondary gustatory nucleus to reach the contralateral side via the intrinsic commissure of this nucleus (2Gc; Fig. 7A). Within the secondary nucleus, terminal arborizations appeared to form glomerular structures. In some brains, the bulk of VFL fibers appeared to terminate in more ventral regions of the secondary gustatory nucleus. This was, however, not a consistent finding and may have varied according to the placement of the DiI within the VFL. Overall, in most DFL or VFL cases, no clear somatotopic partitioning of the secondary gustatory nucleus could be recognized.

Ipsilateral descending fibers from the VFL joined the descending secondary gustatory tract to terminate throughout the medial funicular nucleus and continuing caudally into the dorsal horn. A few labeled fibers left the main bundle to terminate within the small ipsilateral VFn area (Fig. 8).

Two major commissural systems are present in the medulla. A prominent fiber bundle connects the facial lobes at their dorsocaudal rim; it contains commissural fibers into the contralateral VFL as well as descending decussating fibers into the DFL, continuing into the small contralat- era1 VFn area and into the contralateral side of the medial funicular nucleus.

A ventral, more diffuse system of commissural fibers arches towards the contralateral side ventral to the ven- tricle into the contralateral VFL (Fig. 4D‘), via the dorsal reticular area; some of these fibers seem to end within the ipsilateral vagal lobe. Decussating fibers of this system also turn rostralward to terminate within the contralateral PFn and LSFn and extend to the contralateral secondary gusta-

Fig. 2. Dorsal view of the brain of a rockling from which the cerebellum has been removed to reveal the underlying facial and vagal lobes. (Adapted from Kotrschal and Whitear, 1988, with permission of the publisher.)

were mounted in series on slides either with buffered glycerol or with hydromount (Fisher). The sections were examined either immediately or within days. Mounted sections can be stored for several months at 4”C, but fluorescence of the most delicate structures may suffer. Mounted sections were viewed and documented with an epifluorescence Zeiss standard microscope equipped with a 100 W mercury-illumination system and a rhodamine filter combination.

A reason for caution in the use of DiI in fixed material is the possibility of transneuronal labeling. Although we cannot rule out the possibility that some of the labeled cell bodies we observe are due to transcellular passage of the dye, we do not believe that this was a ubiquitous phenom- enon in our tissues, as evidenced by the lack of labeled somata in one of the major terminal fields (2Gn).

For purposes of comparison, 10 ym paraffin sections were stained with the Bodian technique. Photomicrographs taken from comparable areas and at the same magnification as from the DiI specimens were used for purposes of comparison and documentation.

418 K. KOTRSCHAL AND T.E. FINGER

Fig. 3. A-H Series of Bodian-stained representative cross sections through the brainstem of Ciliatu mustela. Inset at lower left is a dorsal view of the brain with the cerebellum removed, as in Figure 2, to show the levels from which the sections are taken. Scale bar = 500 pm.


Figure 3 (Continued.)

K. KOTRSCHAL AND T.E. FINGER

Fig. 4. Chartings of cross sections from two different preparations. The series at right (A'-F') is from a brain in which the DFL had been removed, with DiI then placed in the exposed VFL. The second series (A-F), to the left of the first, is from a brain in which DiI was placed in

the DFL. Inset at lower left shows a dorsal view of the brain, with the approximate levels of these sections indicated. The DiI application site is indicated by heavy shading, labeled fibers by dotted lines, and labeled cell bodies by triangles.


Fig. 5. Section from the level of Figure 3C showing labeling of fiber tracts following application of DiI to the VFL. The N. VII fibers are filled retrogradely, whereas the secondary gustatory tract (2Gt) is filled in an anterograde direction. Scale bar = 100 km.

tory nucleus via the ascending secondary gustatory tract. In addition, the secondary gustatory nucleus in the pons contains an intrinsic commissure in which many secondary gustatory fibers decussate (Fig. 7A).

Labeled perikarya were found ipsilateral to the DiI application in the VFL, LSFn (Fig. 6A), PFn (Fig. 6B), and VFn (Fig. 8) as well as in the medial funicular nucleus. In

Thus labeled fibers in these and Other may therefore be recurrent collaterals from the labeled cells rather than being output targets of the VFL.

addition? labeled somata Occur in the VFL' Fig, 6. Medullary labeling following application of DiI to the VFL. A: Lateral secondary facial nucleus (LSFn) labeling. Several cell bodies (arrows) can be seen as well as numerous fibers. B: Perifacial nucleus. Both labeled cell bodies (arrows) and fibers are apparent. Scale bar = 50 km.

Connections of the dorsal facial lobe (SCC system)

Similar to the VFL, the DFL maintains ascending, descending, and commissural connections. The chief differences between the projections of the two parts of the facial lobe were in the absence of a DFL connection to ipsilateral PFn and in the marked reduction in the descending pathway from DFL to the medial funicular nucleus.

Labeled fibers left DFL in a ventrally directed, compact bundle, coursing along the medial and lateral edges of the VFL to enter the secondary gustatory tracts (Figs. 4D, 9).

The ascending fibers reached as far as the superior secondary gustatory nucleus, with labeled fibers arching into ipsilateral LSFn en route (Fig. 7B). No labeled fibers or somata were present in ipsilateral PFn. Labelled descending fibers in this system were relatively sparse and could be traced only into the rostralmost portion of the medial funicular nucleus. Some fibers in this ventrally directed system crossed the midline to terminate in the contralateral LSFn and superior secondary gustatory nucleus.


Fig. 8. Labeled fibers and cell bodies (arrow) in the ventral funicular nucleus (VFn) after dye application to the VFL. Scale bar = 50 bm.

occurred within the ipsilateral VL, LSFn, and VFn areas but not within either the ipsilateral PFn or the funicular nucleus as was the case after VFL labeling (see Fig. 12).

Some interconnections exist between the DFL and the VFL. Following application of DiI to the DFL, labeled fibers were seen in both the ipsi- and contralateral VFL (Fig. 11A,B). In addition, a few labeled somata were found in the ipsilateral VFL and contralaterally in both the VFL and the DFL.

In summary, the VFL (taste) and DFL (SCC) systems have largely similar connections within the brainstem. The biggest difference between the two systems is in the degree of interconnection with the funicular nuclei ( ~ i ~ . 12). The VFL system has far more extensive connections with the

centers than does the DFL system. Ascending connections from the two systems are essentially identical.

Fig. 7. Labeling in the secondary gustatory nucleus. A: L&eled fibers in the secondary nucleus (2Gn) and its intrinsic commissure (2Gc) following application of DiI to the VFL. B: Labeled fibers in the secondary gustatory nucleus following application of DiI to the DFL. This photomicrograph shows a region of the 2Gn more lateral than that shown in A.

The majority of decussating fibers from the DFL, however, crossed the midline via the dorsocaudal facial commissure to enter the contralateral DFL (Fig. 10). This dorsocaudal commissure lies at the caudal end of the DFL, and fibers originating from the rostral part of the DFL course along the surface of the DFL to reach the commissure.

No labeled fibers reached the ipsilateral PFn area, and only a few DFL secondary fibers were found to terminate within the medial funicular nucleus. Labeled perikarya

DISCUSSION Despite the difference in functional input to the DFL and

VFL, both have similar ascending and descending projection systems targeted predominantly to the superior secondary gustatory nucleus and a portion of the funicular nuclear complex, especially the VFn. These findings are somewhat different from those of Herrick (1907) who examined the brain of the codfish, a related gadid species. Herrick described the major output of the facial lobe in gadids as being


Fig. 9. A Bodian-stained cross section at a level between the levels shown in Figure 3D and E. The approximate area shown in B is indicated by the box. B Efferent labeled fibers following DFL injection. The major outflow from the DFL appears in the center top of this figure as it heads toward the secondary gustatory tract. A few commissural fibers (co can be seen turning medially from the major outflow system. These commissural fibers include those connecting to the contralateral reticular formation, lateral superficial facial nucleus, vagal and facial lobes, and superior secondary gustatory nucleus.

directed to the medullary reticular formation and spinal funicular complex. He suggested that the contribution to the ascending secondary gustatory nucleus from the facial lobe was minimal. Although we do see a slight projection to the medullary reticular formation, the bulk of the facial lobe output does reach the superior secondary gustatory

Fig. 10. Labeling in the DFL following DiI application to the contralateral DFL. This section is just anterior to the intrinsic commissure of this lobule; the root of the commissure appears at upper right. Scale bar = 100 wm.

nucleus. Although this might be attributable to a species difference, preliminary studies in cod (Finger, unpublished observations) indicate that the facial lobe in this species too has a major ascending output directed to the superior secondary gustatory nucleus.

In the present study, the VFL system was found to maintain much more extensive connections with the ventral funicular nucleus compared with the DFL system. This might be attributable either to functional differences (e.g., SCC vs. taste buds) or to somatotopic differences. The funicular complex is reportedly a “correlation” center (Herrick, 1907) for spinal somatic sensations and facially mediated gustatory ones. The DFL contains only the sensory representation of the vibratile dorsal fin, which is not used extensively as an exploratory, tactile organ. In contrast, the VFL contains the representation for widespread areas of the body, including the specialized pelvic fins, which are used for active exploration of the substrate. Accordingly, the more extensive interconnections between the VFL and the funicular complex may be related more to the manner in which the different specialized fins are used rather than to the type of chemoreceptor cells innervated in each area.

Solitary chemoreceptors and taste buds In many fishes, the taste buds and solitary chemorecep-

tors are innervated by different types of nerves terminating


I I

Fig. 11. Labeling in the VFL following application of DiI to the DFL. A: Clusters of labeled fibers, reminiscent of glomeruli described by Herrick (1905,1907) from the ipsilateral VFL. B: Similar, but sparser, labeled fibers in the contralateral VFL. C: Labeled somata in the ipsilateral VFL.

in separate functional areas of the central nervous system: special visceral and general cutaneous, respectively (Finger, 1993). The cutaneous taste buds are innervated by the facial nerve, which projects into the rostral portion of the medullary visceral sensory column. The innervation of SCCs appears less consistent. In sea robins, Prionotus, the fin-ray SCCs are innervated by cutaneous components of

I

VFL

I

DFL Fig. 12. Summary diagrams showing the projection patterns of the

DFL and VFL as determined in the present study. Shared, minor reciprocal connections with the LSFn and VFn are not hown for either lobule. The two lobules have largely similar central connections, although the VFL maintains substantially larger interconnection with the ventral funicular nucleus.

spinal nerves that project into the general somatic zone in the dorsal horn (Finger, 1982). In the rockling, however, the SCCs of the specialized dorsal fin and dorsal trunk skin are innervated by the facial nerve rather than by trigeminal or spinal fibers (Whitear and Kotrschal, 1988; Kotrschal et al., 1993). As was shown previously (Kotrschal and Whi- tear, 19881, these fibers terminate not in the general somatic nuclei of the brainstem but in the facial lobe, which is of the visceral sensory column. Thus, the facial lobe in rocklings contains two anatomically distinct lobules-the DFL, in which only SCC inputs terminate, and the VFL, in which mixed-function taste and SCC input terminates.

Results from the present study indicate that both parts of the facial lobe have similar connections within the central nervous system and that these are essentially identical to facial lobe (gustatory) connections in other teleosts (Kan- wal and Finger, 1992). Thus, for rocklings, the dorsal-fin SCC system appears to be related to the external taste system.

In sea robins, which have a specialization of the SCC system of the pectoral fin (Scharrer et al., 1947; Whitear, 19711, the SCCs are innervated by spinal nerves that terminate in the dorsal horn (Finger, 1982). The central connections of the sea robin SCC system do not involve the visceral sensory column but are part of the ascending somatosensory system, which after a relay in the funicular nuclei reaches the contralateral diencephalon (Finger, 1983). One interpretation of the very different central connections and pattern of innervation of the rockling and sea robin SCC systems is that the SCC systems of these two fish are not homologous, i.e., that, despite similarities in receptor morphology between the two systems, the SCC system of rocklings may be unique in terms of being secondarily derived from the gustatory system. In other words, the SCC of the rockling dorsal fin are a unique specialized form of


taste buds and do not represent a modification of the ancestral SCCs found in virtually all aquatic anamniote vertebrates (Kotrschal, 1991, 1995; Whitear, 1991). Thus the SCC system in rocklings may not be representative of the SCCs in other vertebrates. Conversely, the SCCs of the distantly related sea robins may be the derived condition. The concept of nonhomologous receptors serving similar sensory capacities is not unique to the SCC system. For example, electroreceptors in various fishes are not homologous (Bullock et al., 19831, although they detect similar physical stimuli.

Despite the similarity in central connections of the gustatory and dorsal fin SCC system in rocklings, behavioral and electrophysiological studies (Peters et al., 1987, 1989, 1991; Kotrschal et al., 1993) indicate that the two systems are used to elicit quite different behaviors and are tuned to different types of substances. Typically, the gustatory system in vertebrates is considered to be a contact chemoreceptor system (i.e., requiring direct contact with the stimulus) involved in regulation of feeding. In fishes such as catfish, however, the external gustatory system does not require contact with foodstuffs but can act as a long-range chemosense utilized for localization of a distant food source (Atema, 1971). Nonetheless, throughout the vertebrate lineage, taste is utilized for food acceptance and rejection, not for social activities or predator avoidance (Atema, 1971; Finger, 1987). In contrast, the SCC system of rocklings is reported to respond to substances in the mucus of heterospecifics (Peters et al., 1991) and to bile salts (Kotrschal and DGving, unpublished observations); chemical stimulation of this system decreases locomotor activity and causes cessation of respiratory movements, a response perhaps related to predator avoidance (Kotrschal et al., 1993). Thus, despite sharing central connections with the gustatory system, the SCC system of rocklings is not involved in food detection and triggers nonfeeding behaviors.

This apparent contrast between the central anatomy (gustatory-like) and behavior (predator avoidance) is puz- zling. One interpretation of these data is that the predator- avoidance behavior is a form of inhibition of feeding. A large body of literature exists on feeding deterrents in fish (e.g., Mackie and Mitchell, 19821, and much of this deterrence may be mediated by the gustatory system. If the SCC system in rocklings is a modified taste system, then the SCC-mediated predator avoidance may be one form of feeding deterrence. This in itself would be an unusual way of using a (modified) gustatory system; the external (facial) gustatory system in other fishes appears relatively insensi- tive to taste-aversive compounds such as quinine (Daven- port and Caprio, 1982), and gustatory-mediated feeding deterrence is likely mediated through intraoral taste buds innervated by the glossopharyngeal or vagus nerves (Kan- wal and Caprio, 1983; Lamb and Finger, 1995).

Alternatively, the rockling SCC system may be representative of the SCC system in other vertebrates. In that case, the sea robins, in which the SCC system is used to drive feeding behaviors, may be the unusual situation. Insofar as no functional studies of the SCC systems in other fishes have been performed, virtually no information is available regarding behaviors elicitable by these systems. Thus further study is required to determine the ancestral situation in aquatic vertebrates.

ACKNOWLEDGMENTS The authors thank Barbel Bottger for technical assis-

tance and Drs. Charles Lamb and William Carr for useful comments and discussion on aspects of this paper. This work was supported by NIH grant DC00147 to T.E.F. and by the Konrad-Lorenz Forschungsstelle fur Ethologie.

LITERATURE CITED Atema, J. (1971) Structures and functions of the sense of taste in the catfish

(Ictulurus natalis). Brain Behav. Evol. 4.273-294. Bullock, T.H., D.A. Bodznick, and R.G. Northcutt (1983) The phylogenetic

distribution of electroreception: Evidence for convergent evolution of a primitive vertebrate sense modality. Brain Res. Rev. 6:2546.

Davenport, D.E., and C. Caprio (1982) Taste and tactile recordings from the ramus recurrens facialis innervating flank taste buds in the catfish. J. Comp. Physiol. 147.217-229.

Finger, T.E. (1982) Somatotopy in the representation of the pectoral fin and free fin rays in the spinal cord of the sea robin Prionotus carolinus. Biol. Bull. 163:154-161.

Finger, T.E. (1983) Ascending spinal pathways in the teleost fish Prionotus carolinus. SOC. Neurosci. Abstr. 9.244.

Finger, T.E. (1987) Organization of chemosensory systems in the brain of fishes. In A. Popper, R. Fay, J. Atema, and W. Tavolga (eds): Sensory Systems in Aquatic Animals. New York: Springer-Verlag, pp. 339-363.

Finger, T.E. (1993) What's so special about special visceral? Acta Anat. 148t132-138.

Finger, T.E., and B. Bottger (1990) Transcellular labeling of taste bud cells by carbocyanine dye (DiI) applied to peripheral nerve in the barbels of the catfish, lctalurus punctutus. J. Comp. Neurol. 302.884-892.

Herrick, C.J. (1904) The organ and sense of taste in fishes. Bull. US. Fish. Comm. 22237-272.

Herrick, C.J. (1905) Central gustatory paths in brains of bony fishes. J. Comp. Neurol. 15:375456.

Herrick, C.J. (1906) On the centers for taste and touch in the medulla oblongata of fishes. J. Comp. Neurol. 16t403-439.

Herrick, C.J. (1907) A study of the vagal lobes and funicular nuclei of the brain of the codfish. J. Comp. Neurol. 17337-87.

Honig, M.G., and R.I. Hume (1989) DiI and DiO: Versatile fluorescent dyes for neuronal labelling and pathway tracing. Trends Neurosci. 12333.

Kanwal, J.S., and J. Caprio (1983) An electrophysiological investigation of the oro-pharyngeal (IX-X) taste system of the channel catfish, Ictalurus punctutus. J. Comp. Physiol. 150:345-357.

Kanwal, J.S., and T.E. Finger (1992) Central Representation and Projec- tions of Gustatory Systems. In T.J. Hara (ed): Chemoreception in Fishes, 2nd Ed. New York Elsevier, pp. 79-102.

Kotrschal, K. (1991) Solitary chemosensory cells-Taste, common chemical sense or what? Rev. Fish Biol. Fisheries 1.3-22.

Kotrschal, K. (1995) Ecomorphology of solitary chemosensory cell systems in fish: A review. Environ. Biol. Fish 44.143-155.

Kotrschal, K., and M. Whitear (1988) Chemosensory anterior dorsal fin in rocklings (Guidropsurus and Ciliutu, Teleostei, Gadidae): Somatotopic representation of the ramus recurrens facialis as revealed by transgangli- onic transport of HRP. J. Comp. Neurol. 268:109-120.

Kotrschal, K., M. Whitear, and H. Adam (1984) Morphology and histology of the anterior dorsal fin of Gaidropsarus mediterruneus (Pisces, Teleos- tei), a specialized sensory organ. Zoomorphology 104t365-372.

Kotrschal, K., R. Peters, and 3. Atema (1993) Sampling and behavioral evidence for a mucus detection in a unique chemosensory organ: The anterior dorsal fin in rocklings (Ciliatu mustelu: Gadidae, Teleostei). Zool. Jb. Physiol. 97:47-67.

Lamb, C., and T.E. Finger (1995) Gustatory Control of feeding behavior in goldfish. Physiol. Behav. 57.483-488.

Mackie, A.M., and A.I. Mitchell (1982) Further studies on the chemical control of feeding behavior in the Dover sole, Soleu soleu. Comp. Biochem. Physiol. 73Ar89-93.

Morrill, A.D. (1895) The pectoral appendages of Prionotus and their innervation. J. Morphol. 1 lt177-192.

Parker, G.H. (1912) The relation of smell, taste and the common chemical sense in vertebrates. J. Acad. Nat. Sci. 15.219-234.


Peters, R.C., G.W. Van Steenderen, and K. Kotrschal(1987) A chemoreceptive function for the anterior dorsal fin in rocklings (Gaidropsarus and Ciliata: Teleostei: Gadidae): Electrophysiological evidence. J. Marine Biol. Assoc. U.K. 672319423.

Peters, R.C., K. Kotrschal, W.-D. Krautgartner, and J. Atema (1989) A novel chemosensory system in fish: Electrophysiological evidence for mucus detection by solitary chemoreceptor cells in rocklings (Czliata mustela, Gadidae). Biol. Bull. 177:329.

Peters, R.C., K. Kotrschal, and W.-D. Krautgartner (1991) Solitarychemore- ceptor cells of Ciliata mustela (Gadidae, Teleostei) are tuned to mucoid stimuli. Chem. Sens. 16:3142.

Scharrer, E., S.W. Smith, and S.L. Palay (1947) Chemical sense and taste in the fishes Prionotus and Trichogaster. J. Comp. Neurol. 86:183-193.

Whitear, M. (1965) Presumed sensory cells in fish epidermis. Nature 208:703-704.

Whitear, M. (1971) Cell specialization and sensory function in fish epidermis. J. Zool. London 163:237-264.

Whitear, M. (1991) Solitarychemoreceptor cells. In T.J. Hara (ed): Chemore- ception in Fishes, 2nd Ed. New york Elsevier.

Whitear, M., and K. Kotrschal(1988) The chemosensory anterior dorsal fin in rocklings (Gaidropsarus and Ciliata, Teleostei, Gadidae): Activity, fine structure and innervation. J. Zool. London 216:339-366.

secondary connections of the dorsal and ventral facial lobes in a teleost fish, the rockling...

Documents