J. A. Sisneros T. C. Tricas C. A. Luer Response properties ...faculty.washington.edu/sisneros/sisneros et al. 98.pdfORIGINAL PAPER J. A. Sisneros Æ T. C. Tricas Æ C. A. Luer Response

ORIGINAL PAPER

J. A. Sisneros á T. C. Tricas á C. A. Luer

Response properties and biological functionof the skate electrosensory system during ontogeny

Accepted: 19 February 1998

Abstract This study examined the response propertiesof skate electrosensory primary a�erent neurons of pre-hatch embryo (8±11 weeks), post-hatch juvenile (1±8months), and adult (>2 year) clearnose skates (Rajaeglanteria) to determine whether encoding of electro-sensory information changes with age, and if the electro-sense is adapted to encode natural bioelectric stimuliacross life history stages. During ontogeny, electrosen-sory primary a�erents increase resting discharge rate,spike regularity, and sensitivity at best frequency. Bestfrequency was at 1±2 Hz for embryos, showed an up-wards shift to 5 Hz in juveniles, and a downward shift to2±3 Hz in adults. Encapsulated embryos exhibit venti-latory movements that are interrupted by a ``freeze re-sponse'' when presented with weak uniform ®elds at 0.5and 1 Hz. This phasic electric stimulus contains spectralinformation found in potentials produced by natural ®shpredators, and therefore indicates that the embryoelectrosense can e�ciently mediate predator detectionand avoidance. In contrast, reproductively active adultclearnose skates discharge their electric organs at ratesnear the peak frequency sensitivity of the adult electro-sensory system, which; facilitates electric communicationduring social behavior. We suggest that life-history-de-pendent functions such as these may shape the evolutionof the low-frequency response properties for the el-asmobranch electrosensory system.

Key words Ampullae of Lorenzini á Elasmobranch áElectrorecptor á Frequency response á Communication

Abbreviations BF best frequency á CR convergenceratio á CV coe�cient of variation á EOD electricorgan discharge

Introduction

Sensory receiver systems are commonly tuned to bio-logical stimuli used in the adult life-history stage. Opti-mization of sensory systems occurs among courtshipsongs and the auditory system (Feng et al. 1975; Thor-son et al. 1982), stimuli produced by prey and themechanosensory lateral line system (Lannoo 1986;Montgomery 1989), pheromones and the olfactory sys-tem (Duval et al. 1985; O'Connel 1986), and visualstimuli and photoreceptor system (Loew and Lythgoe1978; Lythgoe and Partridge 1989). Such re®nedmatches between biological signals and receiver systemsare thought to result from selective pressures that ulti-mately increase ®tness of the individual.

One of the best-studied transmitter-receiver systemsis that of the weakly electric teleost ®shes. The Mo-rmyriformes and Gymnotiformes possess weak electricorgans that produce either pulse or continuous wavestimuli. The electrosensory system of these ®shes con-sists of two classes of receptors. Ampullary electrore-ceptors are sensitive to low-frequency stimuli of extrinsicorigin such as that produced by prey (Kalmijn 1974 ). Incontrast, tuberous electroreceptors have evolved inde-pendently to detect both self-generated electric organdischarges (EODs) and those of conspeci®cs (Bullock etal. 1982; Zakon 1986a). Both taxa have morphologicallyand physiologically distinct electric organs and tuberouselectroreceptors systems which are tuned at or near the®sh's own EOD frequency (Hopkins 1976; Bullock 1982;Zakon and Meyer 1983). The match between the fre-quency selectivity of the electroreceptors and the EODfrequency optimizes the function of this sensorimotorsystem for electrolocation (Lissmann and Machin 1958;Heiligenberg 1977; Bastian 1986) and social communi-cation (Hopkins 1972, 1974, 1981; Hagedorn andHeiligenberg 1985).

In contrast to the electrogenic teleosts, the less re-cently derived elasmobranch ®shes use ampullary elec-troreceptors for prey detection, orientation, and social

J Comp Physiol A (1998) 183: 87 ± 99 Ó Springer-Verlag 1998

J.A. Sisneros á T.C. Tricas (&)Department of Biological Sciences,Florida Institute of Technology,Melbourne, Florida, USA

C.A. LuerMote Marine Laboratory, Sarasota, Florida, USA

communication (Kalmijn 1971, 1974; Tricas 1982; Tricaset al. 1995). All elasmobranchs lack tuberous electrore-ceptors but have evolved exquisitely sensitive ampullaryelectroreceptors that can detect electric stimuli at in-tensities as low as 5 nV cm)1 (Kaljmin 1982). Theampullary electrosensory system is known to be impor-tant for prey detection (Kalmijn 1971; Tricas 1982) andtheorized to function in geomagnetic navigation (Kal-mijn 1974; Paulin 1995). Recent work on the non-elec-trogenic round stingray (Urolophus halleri) shows thatthe ampullary electrosensory system functions in socialbehavior during mating by detection of weak ionic ®eldsproduced by conspeci®cs (Tricas et al. 1995). Thus, un-like that found in the teleost electric ®shes, ampullaryreceptors must mediate electric communication inelasmobranchs.

In addition to detection of ionic ®elds of conspeci®cs,the skate ampullary system may also function to detectweak pulsed ®elds produced by the electric organs ofconspeci®cs during social and reproductive interactions(Mikhailenko 1971; Mortenson and Whitaker 1973;Bratton and Ayers 1987). The adult electrosensory sys-tem is most sensitive to sinusoidal stimuli from approx-imately 0.1 to 10 Hz (Adrianov et al. 1984; New 1990)and is similar to other batoids (Montgomery 1984; Tricaset al. 1995; Tricas and New 1998). Recently, New (1994)demonstrated that a single EOD pulse from the littleskate (R. erinacea) contains spectral information whichoverlaps with frequency sensitivity of electrosensoryprimary a�erents. However, the temporal characteristicsof the EOD train are known for only a few species(Bratton and Ayers 1987) and it is currently unclear howampullary electroreceptors would encode this informa-tion. We predict that if the ampullary electroreceptorsfunction in detection of weak electric social signals thenprimary a�erent neurons should e�ectively encode them.

The best known function of the adult skate electro-sense is for the detection of prey (Kalmijn 1971). How-ever, no previous study has addressed the function of theskate electrosense in pre-adult stages. The oviparousskate deposits fertilized eggs on the benthic substratewhich are susceptible to benthic predators such as tele-osts, elasmobranch ®shes and gastropod mollusks(McEachran et al. 1976; Taniuchi 1988; Cox and Koob1993). Thus, if functional the embryo electroreceptorsystem may be used to detect and possibly avoid pre-dators. Similar arguments can be made for the expan-sion of biological function for the electrosense thatchanges with age such as the detection of di�erent preytypes or age-dependent social interactions.

The major goal of this study was to determine theresponse properties and functions of the skate electro-sense through ontogenetic development. We investigateat what stage of development the skate electrosense be-comes functional and how the electrosensory responseproperties change with age. In addition we attempt todetermine at which stage the EOD is active, the temporalcharacteristics of the EOD pulse train, and how thisstimulus is encoded by the ampullary system.

Materials and Methods

Neurophysiology experiments

Clearnose skates (R. eglanteria) were classi®ed into three groupsbased on their stage of ontogenetic development. Embryo egg caseswere collected as freshly oviposited egg cases from captive bredadults (Luer and Gilbert 1985) and maintained in a water table at20 °C until 8±11 weeks of embryonic age (�x � 11.9 � 0.6 SD cmTL, n � 10). Juvenile stage subjects were 1±8 month post-hatchage (�x � 17.4 � 3.7 SD cm TL, n � 9). Adults skates were col-lected in near-shore waters o� Longboat Key, Florida and main-tained in aquaria at 18±22 °C. All adult subjects were males>2 years of age (�x � 52.3 � 2.2 SD cm TL, n � 7) as deter-mined from growth curves (C. A. Luer, unpublished observations).Embryos were removed from the egg case and placed in a smallholding dish. Experimental animals were lightly anesthetized in0.02% tricaine methanesulfonate (MS-222) and then immobilizedby intramuscular injection of pancuronium bromide (approxi-mately 0.1 mg kg)1). Juveniles and adults were clamped lightly onan acrylic stage in a 61 cm long ´ 41 cm wide ´ 15 cm deep acrylicexperimental tank and positioned with a rigid acrylic head and tailholder. Fresh seawater (20 °C) was perfused through the mouthand over the gills for ventilation during all neurophysiological ex-periments.

Single unit discharges were recorded extracellularly from pri-mary a�erent electrosensory neurons. The dorsal branch of thehyomandibular nerve, which contains primary a�erent ®bers fromthe hyoid and mandibular ampullae of Lorenzini clusters, was ex-posed immediately behind the left spiracle. Glass micropipetteelectrodes ®lled with 4 mol á l)1 NaCl (7±35 M) were guided to thenerve surface and ampli®ed using standard techniques as describedby Tricas and New (1998). Electric ®eld stimuli were delivered asbipolar sinusoidal or cathodal square pulse uniform ®elds alongeither the transverse or longitudinal axis of the animal. Sinusoidalstimulus frequencies from 0.01 to 20 Hz were applied at intensitiesfrom 0.5 to 9.5 lV cm)1 (PTP), while square pulse frequencies from0.1 to 20 Hz were applied at an intensity of 1.2 lV/cm (PTP) and apulse duration of 20±30 ms. Electric stimuli were produced by afunction generator and stimulus isolation unit that provided outputvia carbon electrodes positioned along the transverse and longitu-dinal axes of the tank. Analog unit discharges were ampli®ed,®ltered at 300±3000 Hz and stored on tape.

Period histograms were constructed o�-line to determine theneural sensitivity and phase response across stimulus frequencies.For each stimulus frequency a minimum of 300 (for embryos) or500 (for juvenile and adults) consecutive spikes were discriminatedfor at least one stimulus cycle, and distributed into a period his-togram with 128 bins. Peak discharge rate and phase relationship ofthe unit response to the stimulus were determined for each fre-quency by Fourier transform of the period histogram data as de-scribed by Tricas and New (1998). This generated coe�cients forpeak discharge rate and the phase relationship of unit response tothe stimulus frequency. Neural sensitivity (gain) of electrosensoryprimary a�erents was calculated as the net increase in the numberof spikes per second per microvolt per centimeter. Data used togenerate maximum frequency-response curves were standardized toa relative value of 0 dB assigned to the best frequency (BF). Thephase relationship of unit response to stimulus frequency was cal-culated as the di�erence in arc degrees between peak discharge rateand peak stimulus amplitude. Resting discharge regularity wasexpressed as the coe�cient of variation (CV), the ratio of standarddeviation to mean interspike interval duration.

Behavioral experiments

EODs were recorded near the tail of adult male and female skatesfrom a group of 12 individuals engaged in mating activity in anoutdoor tank at Mote Marine Laboratory in February 1995. EODswere also evoked from resting skates by stimulation with pulsed

88

stimuli from a dipole electrode (duration � 30 ms, peak cur-rent � 35 lA, and pulse rate � 10±30 Hz). The dipole stimuluselectrode (15 cm separation) was positioned 1±10 cm above thehead or pectoral disk of the animal. EOD voltage potentials weredi�erentially recorded along the tail of skates with metal electrodesseparated by about 10 cm . EODs were ampli®ed, ®ltered at 1±3000 Hz, and stored on tape for later analyses. For waveformanalysis, analog signals were input to an A/D converter and storedas digital ®les. The mean EOD pulse rate was calculated from theaverage interpulse interval for each pulse train.

Skates of 8±12 weeks of embryonic age and free-swimmingpost-hatch skates (12±19 weeks) were monitored for spontaneousEOD activity. Embryos (n � 9) were maintained in a 41 cm long ´21 cm wide ´ 27 cm deep glass aquarium at temperatures thatranged from 19 to 21 °C. A silver wire electrode (0.5 mm diameter,4 cm length) was inserted into each egg case. A common indi�erentelectrode was positioned 12 cm from the recording electrodes.Newly hatched skates were monitored for EOD activity with a pairof carbon electrodes in a 19-l tank. We also attempted to evokeEODs from juvenile skates by direct stimulation of the electricorgan command nucleus. Anesthetized juveniles were ventilatedwith fresh seawater and the cranial V and VII nerves were tran-sected bilaterally. Various sites along the basal midline of the me-dulla were stimulated with trains of 3±10 pulses of 0.5 ms durationand 15-ms intervals delivered through a concentric bipolar elec-trode as described by New (1994).

Skate embryos normally ventilate the egg case with fresh sea-water by undulation of the tail. The ventilation behavior can beinterrupted by a ``freeze behavior'' response to extrinsic stimulisimilar to that described in the catshark Scyliorhinus (Peters andEvers 1985). In order to determine the frequency response of thefreeze behavior to weak electric stimuli, we recorded the behaviorof eight embryos (10±11 weeks) electrically stimulated with sinu-soidal uniform ®elds at frequencies from 0.02±20 Hz. Embryosencapsulated within the egg case were suspended in a 41 cmlong ´ 21 cm wide ´ 27 cm deep glass aquarium between carbonrod electrodes separated apart by 34 cm (Fig. 1). Electric stimuli

Fig. 1 Experimental tank used to record the freeze behavior inembryonic clearnose skates, Raja eglanteria. Embryonic skates weresuspended in a glass tank (41 cm long ´ 21 cm wide ´ 27 cm deep)between two carbon electrodes (E) positioned along the longitudinalaxis of the egg case. The electric stimulus was delivered as a bipolarsinusoidal uniform ®eld by a function generator (FG) and isolationampli®er (A) that provided output to the electrodes. The stimulussynch output illuminated a LED for video synchronization. Behav-ioral responses of the embryonic skates were backlit by a continuousweak incandescent light source (L) and recorded by video camera (V)

Fig. 2 Resting discharge activity of electrosensory primary a�erentneurons in embryo, juvenile, and adult clearnose skates, R. eglanteria.A 1-s duration record of resting discharge is shown for eachontogenetic stage. Note the increase with age of the primary a�erentdischarge rate and regularity

89

were applied at intensities of 0.56 lV cm)1 by a function generator,isolation ampli®er and carbon rod electrodes positioned adjacent tothe egg case along the longitudinal axis. During the behavioralobservations, stimulus frequencies were presented as continuoussinusoidal stimuli for a minimum of 10 s or at least one cycle of thestimulus for frequencies lower than 0.5 Hz. The freeze response wasde®ned by cessation of tail movement for at least one-half the totalstimulus duration. In order to avoid habituation to the electricalstimulus, an empirically determined inter-trial interval of 10 minwas used after each freeze response was observed. In cases whereno response occurred, we used an inter-trial recovery periodof P 1 min.

Results

Resting discharge activity

Resting discharge activity was recorded from 132 elec-trosensory primary a�erent units in the hyomandibularnerve of 18 skates. Figure 2 shows representative recordsof resting discharge activity for primary a�erents fromembryo, juvenile, and adult skates. Resting dischargerates ranged from 3.2 to 21.5 spikes s)1 for 8- to 12-weekembryos, 31.4±57.1 spikes s)1 for juveniles, and 36.2±66.6 spikes s)1 for adults. We were unsuccessful at sev-eral attempts to record resting spike rate activity fromprimary a�erents in 6- to 7-week embryos. Resting dis-charge rate (Fig. 3) did not di�er among juveniles(�x � 42.6 � 6.1 SD spikes s)1, n � 65) and adults(�x � 44.9 � 7.5 SD spikes s)1, n � 20), but both wereapproximately three times greater than that of embryos(�x � 12.9 � 4.2 SD spikes s)1, n � 47; one-wayANOVA, extended Tukey test, P < 0.01).

The resting discharge patterns also varied amongontogenetic stages. Figure 4 shows representative inter-spike interval histograms for primary a�erents from

each stage. Adults and juveniles show unimodal sym-metrical interspike interval histograms while those ofembryos were skewed to the right. Embryos also havethe largest discharge variability for interspike intervalduration and discharge regularity (Fig. 5) with a meaninterspike interval of 99.7 � 61.6 SD ms and a rela-tively high CV of 0.48 � 0.27 SD (n � 38). In com-parison, juveniles have a lower mean interspike intervalof 25.2 � 3.4 SD ms and an intermediate CV of0.30 � 0.03 SD (n � 40). Adults showed the lowestaverage interspike interval (�x � 22.5 � 4.0 SD ms)and the lowest CV (�x � 0.20 � 0.03 SD, n � 20).Although adult discharges were more qualitatively sim-ilar to juveniles (Fig. 5), they have both a lower meaninterspike interval (two-tailed t-test, df � 56 P < 0.01)and a lower CV (two-tailed t-test, df � 56 P < 0.001)than that of embryos.

Fig. 3 Relationship of resting discharge rate with total length (TL) forelectrosensory primary a�erent neurons in embryo, juvenile, and adultclearnose skates, R. eglanteria. Data are plotted as means andstandard errors for each experimental animal

Fig. 4 Change in regularity of resting discharge of electrosensoryprimary a�erent neurons in embryo, juvenile, and adult clearnoseskates, R. eglanteria. Interspike interval (ISI) distributions are shownfor a representative unit for each stage. Note that discharge regularityincreases with skate size. Discharge regularity is expressed as thecoe�cient of variation (CV), a dimensionless ratio of standarddeviation to mean interspike interval duration. Bin width � 5 ms

90

Neural response to sinusoidal electric stimuli

Stimulation of embryo, juvenile, and adult electrore-ceptors with sinusoidal electric ®elds produces a modu-lation of electrosensory primary a�erent neuraldischarges. These neural discharges are excited and in-hibited by cathodal and anodal stimulation, respectively.Peak modulation of the resting discharge was an in-creasing function of stimulus amplitude below fullmodulation which is approximately twice that of theresting rate as described by Tricas and New (1998).

Sensitivity and frequency response

Within the response range below full modulation, pri-mary a�erents from all skates showed peak frequencysensitivity to sinusoidal stimuli from 0.5 to 7 Hz (Fig. 6).However, peak sensitivity of electrosensory primary af-ferents measured at BF di�ered among ontogeneticstages. Juvenile BF ranged from 4 to 6 Hz(�x � 4.8 � 1.0 SD Hz, n � 23) and was 1.5 timeshigher than the 2±3 Hz BF for adults (�x � 3.2� 1.4 SD Hz, n � 15) and twice that of the 1±2 Hz BFfor embryos (�x � 2.1 � 1.1 SD Hz, n � 26; one-wayANOVA, extended Tukey test, P < 0.01; Fig. 7).The )3 dB bandwidth was 0.2±5.0 Hz for embryos,0.4±11.0 Hz for juveniles, and 0.6±5.6 Hz for adults. Thebandwidth of juveniles was approximately twice that ofembryos and adults, while those of embryos and adultswere very similar. Phase alignment of the frequency re-sponse was consistently observed for frequencies near1±2 Hz for all size classes (Fig. 6). There was no di�er-

ence in response lag at BF among embryos (�x �33.8 � 30.8° SD, n � 32), juveniles (�x � 33.8 � 11.7°SD, n � 23), and adults (�x � 33.2 � 12.2° SD,n � 15; Kruskal-Wallis Test, P � 0.86). Data used togenerate bode and phase plots for response of electro-sensory primary a�erent neurons in embryo, juvenile,and adult skates are summarized in Table 1. The lowvariability of peak frequency response within each ageclass indicates that di�erences in peak frequency re-sponse are stage speci®c and not individual speci®c.

Neural sensitivity (gain) of electrosensory primarya�erents for all skates increased gradually from 0.1 to1 Hz but rapidly decreased above BF (Fig. 6). There wasno di�erence in the low-frequency slope of neural sen-sitivity among embryos (slope � 10.8 � 5.3 SD dB/

Fig. 6 Bode and phase plots for response of electrosensory primarya�erent neurons in embryo, juvenile, and adult clearnose skates, R.eglanteria. Peak frequency sensitivity is 1±2 Hz for embryos, 4±6 Hzfor juveniles, and 2±3 Hz for adults. Data were calculated from theperiod histogram analysis and are plotted as the mean discharge peakfor eight embryo, ®ve juvenile, and four adult clearnose skates. Inorder to control for absolute sensitivity of di�erent units, data werenormalized relative to peak response for each unit and expressed inrelative dB. Phase alignment of frequency response was observed forfrequencies near 1±2 Hz for all size classes. All data plotted as mean�1 standard error. Note some standard error bars are obscured bysymbols

Fig. 5 Relationship between discharge regularity and mean ISI forelectrosensory primary a�erent neurons for embryo, juvenile, andadult clearnose skates, R. eglanteria. Note that the high variability inISI found in embryos decreases with ontogenetic stage. Dischargeregularity is expressed as CV, a ratio of standard deviation to meaninterspike interval duration

91

decade from 0.01 to 0.1 Hz), juveniles (slo-pe � 6.4 � 1.5 SD dB/decade from 0.01 to 0.25 Hz),and adults (slope � 7.9 � 3.8 SD dB/decade from 0.01to 0.5 Hz; ANCOVA, P � 0.51). In comparison, thehigh-frequency slope of neural sensitivity for juveniles(slope � )10.8 � 8.2 SD dB/decade from 12 to 20 Hz)was approximately half that of embryos (slo-pe � )24.9 � 12.0 SD dB/decade from 5 to 20 Hz)and adults (slope � )21.1 � 5.5 SD dB/decade from 6to 20 Hz; ANCOVA, GT2 test, P < 0.001).

Neural sensitivity of primary a�erents at BF rangedfrom 0.3 to 6.2 spikes s)1 lV)1 cm)1 for embryos, 3.8 to20.6 spikes s)1 lV)1 cm)1 for juveniles, and 7.6 to 36.8spikes s)1 lV)1 cm)1 for adults. Figure 8 shows thatneural sensitivity at BF increases with size. Neural sen-sitivity at BF in juveniles (�x � 11.1 � 5.4 SD spikess)1 lV)1 cm)1, n � 19) did not di�er from adults(�x � 17.7 � 8.4 SD spikes s)1 lV)1 cm)1, n � 13) butwas approximately ®ve and eight times greater, respec-tively, than that of embryos (�x � 2.1 � 1.6 SD spikess)1 lV)1 cm)1, n � 26; one-way ANOVA, extendedTukey test, P < 0.05).

EOD characteristics

EODs were recorded from reproductively active adultskates in order to analyze the temporal characteristics.Resting male and female skates often discharged whenapproached by a conspeci®c within approximately20 cm or touched directly. EOD trains lasted from 6 to15 s in duration (�x � 9.3 � 3.1 SD s, n � 12) with anaverage pulse rate of 2±3 Hz. EOD pulses consist of amajor monophasic waveform of mean duration of32.7 � 5.2 SD ms (n � 284).

The average EOD pulse rate produced by skatesengaged in reproductive activity matches the peak re-sponse of the electrosensory primary a�erents(Fig. 9A). EOD pulse rates ranged from 0.9 to 5.0 Hz(�x � 2.5 � 1.1 SD pulses s)1, n � 34), with 68% ofall EOD trains at the 2- to 3-Hz peak frequencysensitivity of the adult electrosensory system. Therewas no di�erence in EOD pulse rate among males(�x � 2.6 � 1.0 SD Hz, n � 14) and females (�x �2.5 � 1.2 SD Hz, n � 20; two-tailed t-test, df � 32,P � 0.80). These results demonstrate that peak fre-quency sensitivity of the adult electrosensory systemclosely matches the pulse rate of conspeci®c EODsduring social behavior.

EODs were routinely evoked from resting skates inlarge holding tanks by stimulation with weak square-pulse stimuli (10±30 Hz) from a dipole electrode posi-tioned above the head or pectoral disc. Both male andfemale skates responded to the synthesized stimulus byproducing EOD trains which varied in duration from

Table 1 Summary of the number of electrosensory primary a�er-ents sampled for each stimulus frequency in embryo, juvenile, andadult clearnose skates, Raja eglanteria. Note the number of animalsused to generate maximum frequency response curves are indicatedin parenthesis next to each speci®c size class

Frequency Embryos (8) Juveniles (5) Adults (4)

0.01 4 14 40.025 4 16 100.05 9 19 110.1 17 19 120.25 22 19 130.5 24 20 131 26 22 152 26 23 153 26 23 154 26 23 155 25 23 156 25 22 157 22 22 158 19 22 159 20 22 1510 17 22 1312 12 19 1215 8 18 1220 4 18 10

Fig. 7 Best frequency (BF) histogram for electrosensory primarya�erent neurons in embryo, juvenile, and adult clearnose skates,R. eglanteria. Sample sizes are indicated by the number of animalssampled followed by the total number of electrosensory primarya�erent neurons. Best frequency shifts from 1±2 Hz in embryos to4±6 Hz in juveniles to 2±3 Hz in adults

92

1.38 to 44.1 s (�x � 17.4 � 17.4 SD s, n � 7). One fe-male responded to square-pulse stimuli at 10 Hz with aseries of trains that lasted a total of 164 s (Fig. 9B).Thus, the behaving skates are able to detect extrinsicEOD pulse trains, and will respond with their own EODtrain.

Embryos are very active within the egg case andpromote circulation of fresh seawater by undulation ofthe tail for extended periods. Of the nine 8- to 12-weekembryonic age skates monitored for a total of 102 h, noEODs were recorded over a 4-week period. In addition,we recorded no discharges from free-swimming post-hatch juveniles from 12 to 19 weeks of age. Further, wefailed to elicit EODs in post-hatch juveniles (n � 3) bydirect stimulation of the electric organ command nu-cleus, spinal cord or electric discharge organ. Thus, wewere unable to demonstrate functionality of the EODcircuitry in either embryo or post-hatch juveniles skates.

Neural response to simulated EOD stimuli

Electrosensory primary a�erent neurons in adult skatesencode simulated EOD stimuli (pulse duration 20±30 ms) at frequencies up to about 5 Hz (Fig. 10). Abovethis stimulus frequency, electrosensory primary a�erentencoding of the square-pulse stimuli declines. At 20-Hzpulse rate primary a�erents poorly encode frequencyinformation but do respond with an elevated averagedischarge that is maintained throughout the stimulusperiod. The elevated spike rate which is due to the re-current stimulation at high frequency precludes full re-covery from the stimulus, and thus interferes with theability of primary a�erents to encode high frequencies.

At a pulse rate of 33 Hz, the stimulus has a full (100%)duty cycle, equivalent to a d.c. step, and the unit fullyadapts to the stimulus within a few seconds. These re-sults demonstrate that the electrosensory system of theadult skate can faithfully encode individual EOD pulsesin natural trains produced by conspeci®cs.

Behavioral response of embryos to sinusoidalelectric stimuli

Embryos routinely undulate their tail in an approximatesinusoidal fashion to ventilate the egg case with freshseawater. These undulations continue uninterrupted

Fig. 9 A Match between the frequency sensitivity of electrosensoryprimary a�erent neurons and electric organ discharge (EOD) pulserate produced by reproductively active clearnose skates, R. eglanteria.The tuning curve of the adult electrosensory primary a�erents in R.eglanteria shows peak sensitivity at approximately 2±3 Hz with a 3-dBdrop at approximately 0.6 and 5.6 Hz. The mean EOD pulse rate(2.5 � 1.1 SD Hz, n � 34) matches the frequency of peak sensitivityfor the electrosensory system. B A representative train of evokedEODs recorded from a free-swimming female skate stimulated withsquare pulse electric stimuli at 10 Hz. The EOD train has a durationof 164.4 s and average rate of 2.8 � 1.2 SD Hz. Top trace shows theinstantaneous ®ring frequency of EODs expressed as spikes/secondfor the series of EOD pulses. Note that averaged pulse rate is about 2±4 Hz across the entire discharge period

Fig. 8 Relationship of neural sensitivity (gain) at best frequency withTL for electrosensory primary a�erent neurons in embryo, juvenile,and adult clearnose skates, R. eglanteria. Data are plotted as meansand standard errors for each experimental animal

93

when stimulated with weak uniform ®elds delivered at orbelow 0.02 Hz. However, at higher stimulus frequenciesa freeze response was evoked (Fig. 11). The freeze re-sponse begins by a termination of the ventilatory un-dulations of the tail followed by a rapid coiling of thetail about the body. The freeze response occurred in97% (31/32) of the trials performed at 0.5 and 1 Hz anddeclined to 69% at 0.1 Hz and 50% at 2 Hz. Figure 12shows the match between the frequency of the freezeresponse and the peak frequency sensitivity of electro-sensory primary a�erent neurons in embryos. Theseexperiments show that the freeze behavior of embryonicskates is a frequency-dependent response to extrinsicelectric ®elds that can readily be evoked at frequencies ator near the 1- to 2-Hz peak frequency sensitivity of theembryo electrosensory system.

Discussion

This study is the ®rst to compare the response propertiesof elasmobranch electrosensory primary a�erent neu-rons from embryo, juvenile, and adult stages. Our aimwas to determine if frequency response of the electro-sensory system changes with age and whether the skateelectrosensory system is optimized to detect naturalbioelectric stimuli. Our results show that during ontog-eny in the clearnose skate, electrosensory primary af-ferents increase resting discharge rate and regularity,shift in both peak frequency sensitivity and bandwidth,and display an increased maximum sensitivity at BF.Additionally we demonstrate that the peak frequencysensitivity of adult Raja is aligned with the EOD pulserate produced during mating activity, while peak fre-quency sensitivity in the embryo matches sinusoidal

Fig. 10 The average ®ring frequency of electrosensory primarya�erent neurons in adult clearnose skates as a function of pulsestimulus frequency. Each stimulus (ST) was delivered as a cathodaltrain of 10 pulses (20 ms duration) delivered as a uniform ®eld acrossthe skate body at a rate of 0.5±20 Hz. Average ®ring frequency wascomputed across successive 50-ms periods. The time segment for eachrecord can be inferred from the stimulus pulse train which is alignedwith the response of the unit for each frequency. Note that at lowfrequencies each pulse is encoded as an increase in ®ring rate. Athigher frequencies primary a�erents lose the ability to encodetemporal features of the pulse train

Fig. 11 Behavioral responses of an embryonic clearnose skate(R. eglanteria) to sinusoidal uniform electric ®elds at frequencies of10 Hz, 1 Hz, 0.1 Hz, and 0.02 Hz. ST were applied at intensities of0.56 V cm)1 across the longitudinal axis of the animal. The response(R) is expressed as a change in peak to peak tail displacement of theskate within the egg case. Prestimulus tail displacement for eachrecord was 10 mm peak to peak. At 1 Hz and 0.1 Hz, note the largetail displacement that occurs during coiling of the tail around the bodyafter the onset of the electrical stimulus and a period of no tailmovement during and after stimulation. Time bars � 5 s

94

stimuli that produce a freeze response. In this discussionwe interpret our results as they relate to possible func-tions of the electrosensory system during ontogeny anddiscuss adaptations of the electric sense to facilitateadult communication during social behavior.

Raja eglanteria complete embryonic development andhatch at approximately 12 weeks at 20±22 °C followingoviposition (Luer and Gilbert 1985). We were unsuc-cessful in several attempts to record the neural activityfrom embryos <8 weeks of age. At 8 weeks, embryoprimary a�erents showed spontaneous discharges andresponded to weak sinusoidal electrosensory stimuli ofintensities of 2.9 lV cm)1. Behavioral responses to uni-form electric ®elds occurred in experimental animals 10±11 weeks of age, and con®rm a functional sensory motorsystem at that age. However, the embryonic age at whichthe electrosense becomes fully functional remains to bedemonstrated. Ultrastructural studies on the develop-ment of the hair cell synapse are necessary to con®rm thesequence of developmental events that result in activa-tion of the peripheral electrosensory system.

The dramatic threefold di�erence of resting dischargerate among embryos and juveniles most likely representsa change at the receptor-a�erent neuron synapse duringdevelopment. For example, increases in the amount orrate of neurotransmitter release by the ampullary elec-troreceptor cells would increase depolarization rates.Another possible explanation is an increased conver-gence ratio (CR) of electroreceptors to primary a�erentsthat would presumably increase the postsynaptic depo-larization of a�erent neurons. In this regard, Peters andIeperen (1989) reported that the ampullary electrore-

ceptor organs of the freshwater cat®sh (Clarias ga-riepinus) exhibit a one- to threefold increase in CRduring the ®rst 4 months of age but only a 1.1-fold in-crease in resting discharge rate. Resting discharge rate,however, was not correlated with di�erent CRs (1:1±3:1)and therefore may be a function of the developmentalmaturity of the electroreceptor organ (Teunis et al.1990). Experimental studies that characterize the quantarelease of neurotransmitter and synaptic morphology ofelectroreceptors in embryo and juvenile skates wouldprovide important insight into how changes in electro-receptors and primary a�erents may a�ect electrosen-sory function and sensitivity during development.

The resting discharge rate recorded for juvenile (42.6spikes s)1) and adult (44.9 spikes s)1) clearnose skates at20 °C is higher than that reported for most other el-asmobranch species. Resting discharge rates in batoidsrange from 8.6 spikes s)1 at 7 °C in R. erinacea (New1990) to 34.2 spikes s)1 at 18 °C in Urolophus halleri(Tricas and New 1998). These di�erences may be due tothe in¯uence of higher temperature which can decreasethe thresholds for membrane depolarization and spikeinitiation (Carpenter 1981; Montgomery and MacDon-ald 1990). Primary a�erents also exhibit a continuedincrease in discharge regularity with age with thegreatest change in the transition from embryo to juvenilestage. The large reduction in discharge variability fromembryos to juveniles also coincides with a large increasein resting discharge rate which together may ultimatelycontribute to the increase in neural sensitivity of theadult Raja electrosensory system.

Neural sensitivity at BF of electrosensory primarya�erent of juveniles and adults was approximately ®vetimes that of embryos. We could not demonstrate anincrease in sensitivity from juveniles to adults but thismay be due to low sample size. The most likely expla-nation for the increase in juveniles and adults is the in-crease in ampullary canal length. Ampullae of Lorenzinidetect potential di�erences between seawater at thesurface pore and the common internal potential of theanimal at the ampullary cluster (Bennett 1971). Thesubdermal canal and the internal lumen of the electro-receptor are isopotential with the opening of the canalpore at the skin surface. Thus, the voltage sensitivity ofthe electroreceptor is a function of canal length. Skatesapproximately double in disk width during the ®rst8 months of growth and therefore the proportionateincrease in canal length would also increase unit sensi-tivity.

Post-embryonic cytological development may alsoexplain increased sensitivity of the skate electroreceptorsystem during growth which is known to occur in othersystems. For example, during the ®rst 4 months of de-velopment in the cat®sh, Clarias gariepinus, proliferationof the ampullary organs results in a 3:1 primary a�erentconvergence ratio and a corresponding 3.6-fold increasein sensitivity (Peters and Ieperen 1989). Several speciesof gymnotiform ®sh also show increases in the numberof ampullary receptor organs per primary a�erent with

Fig. 12 Freeze response of embryonic clearnose skates, R. eglanteria,to weak sinusoidal uniform ®elds. Behavioral responses (open dots) areshown as percentage total stimulus presentation to 0.02±20 Hz. Peakneural sensitivity of electrosensory primary a�erents (solid dots) was at1±2 Hz and shows similar high- and low-frequency roll o�s. Note thealignment of 0.5±1 Hz behavioral response peak with that of peakneural sensitivity

95

size and presumably age (Zakon 1984, 1987). A 20-folddi�erence in the number of ampullary organs per a�er-ent exist among large and small specimens of Stern-opygus dariensis (Zakon 1984). In addition, the meannumber of sensory receptor cells per organ also increaseswith size which presumably increases the sensitivity ofthe system as the ®sh grows (Zakon 1987). In compari-son, primary a�erent sensitivity increased in the tuber-ous electrosensory system of S. macrurus as a result ofincreased receptor organs/axon during postembryonicontogeny (Sanchez and Zakon 1990). Furthermore, adramatic increase in threshold sensitivity is reported forthe auditory system of skates (R. clavata) where in-creased sensitivity is correlated with the addition of newsensory receptor cells (Corwin 1983). Increased sensi-tivity may also be explained by changes in the postsyn-aptic membrane at the ampullary electroreceptorsynapse. Fields and Ellisman (1985) reported a correla-tion between synaptic morphology, i.e., depth of thepostsynaptic trough membrane, and an increase in sen-sitivity in elasmobranch electroreceptors. Similar studiesthat detail the age-related development of the electro-receptor synapse are necessary to determine the cause ofincreased electrosensory sensitivity during developmentof the skate.

At least some of the ontogenetic changes that a�ectthe frequency response properties of the skate electro-sensory primary a�erents probably occur at the pe-ripheral ampullary receptor organ. One possible meansthat could a�ect the frequency selectivity of electrore-ceptors is an age-related change in the properties of ioncurrents in the ampullary epithelium. In general, thefrequency selectivity of hair cell receptors is thought toresult from the electrical resonance of receptor poten-tials (Fettiplace 1987; Hudspeth 1989). This electricalresonance is caused by the interaction between inwardcalcium and outward calcium-dependent potassiumcurrents which produces an electrical oscillation of re-ceptor potentials along the receptor epithelium (Fetti-place 1987; Roberts et al. 1988). The kinetics of these ioncurrents are known to be the key elements responsiblefor tuning hair cell receptors to a speci®c frequency (Artand Fettiplace 1987). Similar electrical resonance isknown to occur in electroreceptors cells (Clusin andBennett 1979; Viancour 1979; Meyer and Zakon 1982;Zakon 1986b). In tuberous electroreceptors of weaklyelectric teleosts, the resonant frequencies of the electricaloscillations is strongly correlated with BF of the elec-trosensory primary a�erents and is presumed to be re-sponsible for the frequency response. It is unclear howelectrical resonance a�ects a�erent tuning in the el-asmobranchs since the resonant frequencies of ampullaryelectroreceptors (21±33 Hz) in the little skate, R. erin-acea, do not match BF of primary a�erents (5±8 Hz) butrather the resting discharge rate (New 1990; Lu andFishman 1995). Alternatively, changes in the cellularmorphology of the ampullary organ could alter themembrane resistance and capacitance of the receptorepithelium or change the ionic membrane properties of

electroreceptor cells. In addition, ampullary structuressuch as the canal wall, lumen, and alveoli form tightjunctions that prevent leakage of transmembrane ioniccurrents. Changes in the cable properties of such highlyresistive structures could alter the high-pass tuningcharacteristics of the ampullary organ. One other meansthat may alter a�erent tuning in skates is the action ofsex steroids on the ion channel properties of electrore-ceptor cells. Recent studies show that steroids alterspeci®c ionic conductances in excitable electrocytes andin electrosensory neural circuits that are important insocial communication during reproduction (Zakon1993; Ferrari et al. 1995; Dunlap et al. 1997). Similarstudies that detail the properties of ion currents associ-ated with the activation of the ampullary electroreceptorthrough development will be necessary to determine howthe response properties of ampullary organs changeduring ontogeny.

Embryos of egg-laying elasmobranchs are known tooccur in the natural diet of some teleost ®shes, sharks,skates, marine mammals and molluscan gastropods(McEachran et al. 1976; Stillwell and Kohler 1982, 1993;Taniuchi 1988; Ebert 1991; Cox and Koob 1993). Late-term skate embryos are extremely active and undulatetheir tail in one corner of the egg case to ventilate freshsea-water (Luer and Gilbert 1985) as do embryos of thedog®sh Scyliorhinus canicula (Peters and Evers 1985).This behavior in the skate results in the streaming ofwater from one horn of the egg case at velocities ofapproximately 7 cm s)1 (unpublished data) which cre-ates a localized vortex around the egg case that mayattract or facilitate location by potential predators.Batoid rays use their mechanosensory lateral line to lo-calize buried bivalve mollusks by water streams from theexcurrent siphon as do teleost scorpion®sh (Scorpaen-idae) to detect ventilatory water currents created byadult crabs (J. Montgomery, personal communication).Thus, the freeze behavior produced by embryonic skateswill stop ventilatory streaming and function to decreasethe likelihood of mechanoreceptive detection by poten-tial predators.

The freeze behavior can be elicited by multiple formsof external stimuli and in skate embryos by sinusoidalelectric ®elds at 0.5 and 1 Hz. This frequency bandcorresponds to the natural ventilatory phasic signalsproduced by large predators (Tricas et al. 1995; T.C.Tricas, unpublished data), and may also re¯ect theperceived low-frequency modulation of a d.c. ®eld pro-duced by an approaching predator as it moves relative tothe embryo (sensu Kalmijn 1988). This match betweenelectric ®eld modulation at frequencies which elicit thefreeze response and peak frequency sensitivity of elec-trosensory primary a�erents indicates that the electricsense of embryonic skates is important in the detectionand avoidance of potential predators.

We found no evidence that embryos or post-hatchjuveniles discharge their electric organs. Embryonicskates of 8±12 weeks of age were not electrogenicallyactive within the egg case under lab conditions. Like-

96

wise, free-swimming post-hatch juveniles, which showconsiderable aggressive and other social interactions inholding tanks, did not discharge during day or night. Incontrast, spontaneous EOD activity occurs in adultclearnose skates and also little skates, R. erinacea, thatare kept either in isolation or in groups (Bratton andAyers 1987). It is currently uncon®rmed that the electricorgans of embryonic skates are fully developed andpotentially functional before hatch. Our inability toevoke EODs by direct stimulation of the electric organcommand nucleus in post-hatch juveniles leads us tobelieve the system is not fully functional in these stages.Electric organs may become functional later in the ju-venile stage, or perhaps during sexual maturation.

Electrosensory primary a�erents of juvenile skatesexhibit the highest peak frequency sensitivity at 4±6 Hzand greatest bandwidth of 0.5±11 Hz among the threeskate size classes. In addition, juvenile skates also showan increased neural sensitivity at BF with size fromembryos to juvenile and adults. In terms of foragingbehavior, increased sensitivity within broadened elec-trosensory bandwidth could allow young skates to betterdetect higher-frequency information from prey and atgreater distance. Other functions might be for socialinteractions and or avoidance of predators. Clearly,more information is needed on the natural predatoryand social behavior of juvenile skates.

The peak frequency response of 2±3 Hz for electro-sensory primary a�erents in adult R. eglanteria is similarto the reported range of 1±5 Hz for the little skate, R.erinacea (New 1990), but higher than the 0.1- to 0.5-Hzrange reported for the black sea skate, R. clavata(Adrianov et al. 1984). While these di�erences in peakfrequency sensitivity among species may re¯ect di�er-ences related to their natural ecology and behavior, thelow-pass characteristics likely represent physiologicalconstraints of the ampullary electroreceptor system. Thepeak EOD pulse in adult R. eglanteria is about 30 ms induration; thus, a burst of 10 pulses s)1 would representan average duty cycle of 30%. We show that the skateampullary system can encode 30-ms-duration pulsedstimuli, which are similar to EODs produced by con-speci®cs, when delivered at pulse rates as high as about5±10 Hz. The poor encoding of frequency informationat higher frequencies is at least partially due to theproportional increase in average duty cycle that ac-companies increased pulse frequency. Given the relativelong time constant of 3±4 s for batoid elasmobranchelectrosensory primary a�erents (Montgomery 1984;Tricas and New 1998) the encoding of electrosensoryinformation should rapidly decrease with increasedEOD pulse rate.

An important ®nding of this study is that the peakfrequency sensitivity of adult skates (2±3 Hz) is alignedwith mean pulse rate of EODs (2.5 Hz) produced duringsocial and mating behaviors. This match may representan adaptation of the adult electrosensory system to fa-cilitate communication during social interaction andmating. Similarly, the EOD pulse rate for the little skate,

R. erinacea, is about 5 Hz during interactions withconspeci®cs (Bratton and Ayers 1987) and the electro-sensory primary a�erents have a peak frequency sensi-tivity near 5±7 Hz (New 1990); thus, encoding of pulserate information should also be important for that spe-cies. However, New (1994) suggested that the electro-sensory system in R. erinacea was adapted to detect thespectral components of individual EOD pulses. This isapparently not the case for R. eglanteria because thesupracutaneous EOD is composed primarily of amonophasic negative pulse about 30 ms in durationwhich would require a considerably higher low-pass cuto� to be optimally detected by the electrosensory sys-tem. Unfortunately, we could not analyze the spectralcomponents of single EOD pulses from R. eglanteria dueto electrical interference in the holding tank and low-frequency noise induced by the movements of the elec-trode or tail during EOD recordings. The relevance ofother spectral components of the EOD train and indi-vidual pulses should also be considered in future inv-estigations. An analysis of the EOD waveform and trainfrequency spectrum needs to be performed to determineif other components of the discharge may be detectableand possibly sex speci®c during the social behavior forthis species.

In summary, we have shown that the embryo elec-trosensory system is functional during the pre-hatch lifeof the clearnose skate, detects weak external ®elds suchas those produced by potential predators and mediatesthe freeze behavior which can potentially increase sur-vival. For adults, there is also a strong physiologicalbasis that the electrosense is well suited to detectconspeci®c EOD trains. This leads us to suggest that life-history-dependent functions shape the evolution of low-frequency response properties of the skate, and possiblyother elasmobranch electrosensory systems.

Acknowledgements We thank Karen Maruska for art work, JoseCastro and Patricia Blum for photographic assistance, Jay Bradleyand Karen Maruska for ®eld assistance, Jonathan Trumbull forlaboratory assistance, and Bernd Fritzch for comments on anearlier draft of this manuscript. This study was supported by agrant from the Whitehall Foundation to TCT and a MARC pre-doctoral fellowship from the National Institutes of Health to JAS.Partial support to CAL was provided through a grant from theDisney Wildlife Conservation Fund. Experimental procedures fol-lowed NIH guidelines for the care and use of animals and wereapproved by the Institutional Animal Care and Use Committee atFlorida Institute of Technology.

References

Adrianov GN, Broun GR, Ilyinsky OB, Muraveiko VM (1984)Frequency characteristics of skate electroreceptive centralneurons responding to electric and magnetic stimulation.Neurophysiology 16: 365±376

Art JJ, Fettiplace R (1987) Variation of membrane properties inhair cells isolated from the turtle cochlea. J Physiol (Lond) 385:207±242

Bastian J (1986) Electrolocation: behavior, anatomy, and physiol-ogy. In: Bullock TH, Heiligenberg W (eds) Electroreception.Wiley, New York, pp 577±612

97

Bennett MVL (1971) Electroreception. In: Hoar WS, Randall DJ(eds) Fish physiology, vol 5. Academic Press, New York,pp 493±574

Bratton BO, Ayers JL (1987) Observations on the electric dischargeof two skate species (Chondrichthyes: Rajidae) and its rela-tionship to behavior. Environ Biol Fishes 20: 241±254

Bullock TH (1982) Electroreception. Annu Rev Neurosci 5: 121±170

Bullock TH, Northcutt RG, Bodznick DA (1982) Evolution ofelectroreception. Trends Neurosci 5: 50±53

Carpenter DO (1981) Ionic and metabolic bases of neuronal ther-mosensitivity. Fed Proc 40: 2808±2813

Clusin WT, Bennett MVL (1979) The oscillatory responses of skateelectroreceptors to small voltage stimuli. J Gen Physiol 69: 121±143

Corwin JT (1983) Postembryonic growth of the macula neglectaauditory detector in the ray, Raja clavata. Continual increasesin hair cell number, neural convergence, and physiologicalsensitivity. J Comp Physiol 217: 345±356

Cox DL, Koob TJ (1993) Predation on elasmobranch eggs. Envi-ron Biol Fishes 38: 117±125

Dunlap KD, McAnelly ML, Zakon HH (1997) Estrogen modi®esan electrocommunication signal by altering the electrocyte so-dium current in an electric ®sh, Sternopygus. J Neurosci 17:2869±2875

Duval D, Muller-Schwarze D, Silverstein RM (1985) Chemicalsignals in vertebrates, vol 4. Plenum, New York

Ebert DA (1991) Diet of the sevengill sharkNotorynchus cepedianusin the temperate coastal waters of southern Africa. S Afr J MarSci 11: 565±572

Feng AS, Narins PM, Capranica RR (1975) Three populations ofprimary auditory ®bers in the bullfrog (Rana catesbeiana): theirperipheral origins and frequency sensitivities. J Comp Physiol100: 221±229

Ferrari MB, McAnelly ML, Zakon HH (1995) Individual variationand androgen modulation of the sodium current in electricorgan. J Neurosci 15: 4023±4032

Fettiplace R (1987) Electrical tuning of hair cells in the inner ear.Trends Neurosci 10: 421±425

Fields RD, Ellisman MH (1985) Synaptic morphology and di�er-ences in sensitivity. Science 228: 197±199

Hagedorn M, Heiligenberg W (1985) Court and spark: electricsignals in the courtship and mating of gymnotid ®sh. AnimBehav 33: 254±265

Heiligenberg W (1977) Principles of electrolocation and jammingavoidance in electric ®sh. A neuroethological approach. In:Braitenberg V (ed) Studies of brain function. Springer, BerlinHeidelberg New York

Hopkins CD (1972) Sex di�erences in signalling in an electric ®sh.Science 176: 1035±1037

Hopkins CD (1974) Electric communication in the reproductivebehavior of Sternopygus macrurus. Z Tierpsychol 35: 518±535

Hopkins CD (1976) Stimulus ®ltering and electroreception: tuber-ous electroreceptors in three species of gymnotid ®sh. J CompPhysiol 111: 171±207

Hopkins CD (1981) On the diversity of electric signals in a com-munity of mormyrid electric ®sh in West Africa. Am Zool 21:211±222

Hudspeth AJ (1989) How the ear's works work. Nature (Lond)341: 397±404

Kalmijn AJ (1971) The electric sense of sharks and rays. J Exp Biol55: 371±383

Kalmijn AJ (1974) The detection of electric ®elds from inanimateand animate sources other than electric organs. In: Fessard A(ed) Handbook of sensory physiology, vol. III/3. Springer,Berlin Heidelberg New York, pp 147±200

Kalmijn AJ (1982) Electric and magnetic ®eld detection inelasmobranch ®shes. Science 218: 916±918

Kalmijn AJ (1988) Detection of weak electric ®elds. In: Atema J,Fay RR, Popper AN, Tavolga WN (eds) Sensory biology ofaquatic animals. Springer, Berlin Heidelberg New York, pp507±528

Lannoo MJ (1986) Vision is not necessary for size-selective zoo-planktivory in aquatic salamanders. Can J Zool 64: 1071±1075

Lissman HW, Machin KE (1958) The mechanism of object loca-tion in Gymnarchus niloticus and similar ®sh. J Exp Biol 35:451±486

Loew EW, Lythgoe JN (1978) The ecology of cone pigments inteleost ®shes. Vision Res 18: 715±722

Lu J, Fishman HM (1995) Localization and function of the elec-trical oscillation in electroreceptive ampullary epithelium fromskates. Biophys J 69: 2458±2466

Luer CA, Gilbert PW (1985) Mating behavior, egg deposition,incubation period, and hatching in the clearnose skate, Rajaeglanteria. Environ Biol Fishes 13: 161±171

Lythgoe JN, Partridge JC (1989) Visual pigments and theacquistion of visual information. J Exp Biol 146: 1±20

McEachran JD, Boesch DF, Musick JA (1976) Food divisionwithin two sympatric species-pairs of skates (Pisces: Rajidae).Mar Biol 35: 301±317

Meyer JH, Zakon HH (1982) Androgens alter the tuning of elec-troreceptors. Science 217: 635±637

Mikhailenko NA (1971) Biological signi®cance and dynamics ofelectrical discharges in weak electric ®shes of the Black Sea (inRussian). Zool Zh 50: 1347±1352

Montgomery JC (1984) Frequency response characteristics of pri-mary and secondary neurons in the electrosensory system of thethornback ray. Comp Biochem Physiol 79A: 189±195

Montgomery JC (1989) Lateral line detection of planktonic prey.In: Coombs S, Gorner P, Munz H (eds) The mechanosensorylateral line: neurobiology and evolution. Springer, Berlin Hei-delberg New York, pp 561±574

Montgomery JC, MacDonald JA (1990) E�ects of temperature onnervous system: implications for behavioral performance. Am JPhysiol 259: 191±196

Mortenson J, Whitaker RH (1973) Electric discharge in freeswim-ming female winter skates (Raja ocellata). Am Zool 13: 1266

New JG (1990) Medullary electrosensory processing in the littleskate. I. Response characteristics of neurons in the dorsal oc-tavolateralis nucleus. J Comp Physiol 167: 285±294

New JG (1994) Electric organ discharge and electrosensory reaf-ference in skates. Biol Bull 187: 64±75

O'Connell RJ (1986) Chemical communication in invertebrates.Experientia 42: 232±241

Paulin MG (1995) Electroreception and the compass sense ofsharks. J Theor Biol 174: 325±339

Peters RC, Evers HP (1985) Frequency selectivity in the ampullarysystem of an elasmobranch ®sh (Scyliorhinus canicula). J ExpBiol 118: 99±109

Peters RC, Ieperen SV (1989) Resting discharge and sensitivityof ampullary electroreceptors in Clarias gariepinus related toconvergence ratio during ontogeny. Brain Behav Evol 34: 43±47

Roberts WM, Howard J, Hudspeth AJ (1988) Hair cells: trans-duction, tuning, and transmission in the inner ear. Annu RevCell Biol 4: 63±92

Sanchez DY, Zakon HH (1990) The e�ects of postembryonic re-ceptor cell addition on the response properties of electrorecep-tive a�erents. J Neurosci 10: 361±369

Stillwell CE, Kohler NE (1982) Food, feeding habits, and estimatesof daily ration of the short®n mako (Isurus oxyrinchus) in thenorthwest Atlantic. Can J Fish Aquat Sci 39: 407±414

Stillwell CE, Kohler NE (1993) Food habits of the sandbar sharkCarcharhinus plumbeus o� the U.S. northeast coast, with esti-mates of daily ration. Fish Bull US 91: 138±150

Taniuchi T (1988) Aspects of reproduction and food habits of theJapanese swell shark Cephaloscyllium umbratile from Choshi,Japan. Nippon Suisan Gakkaishi 54: 627±633

Teunis PFM, Bretschneider F, Peters RC (1990) Convergence ratioand transduction in cat®sh electroreceptive organs. CompBiochem Physiol 97: 405±410

Thorson J, Weber T, Huber F (1982) Auditory behavior of crickets.II Simplicity of calling song recognition in Gryllus, and anom-alous phonotaxis at abnormal frequencies. J Comp Physiol 146:361±378

98

Tricas TC (1982) Bioelectric-mediated predation by swellsharks,Cephaloscyllium ventriosum. Copeia 1982: 948±952

Tricas TC, New JG (1998) Sensitivity and response dynamics ofelectrosensory primary a�erent neurons to near threshold ®eldsin the round stingray. J Comp Physiol 182: 89±101

Tricas TC, Michael SW, Sisneros JA (1995) Electrosensory opti-mization to conspeci®c phasic signals for mating. Neurosci Lett202: 129±131

Viancour TA (1979) Electroreceptors of a weakly electric ®sh: II.Individually tuned receptor oscillations. J Comp Physiol 133:327±338

Zakon HH (1984) Postembryonic changes in the peripheral elec-trosensory system of a weakly electric ®sh: addition of receptororgans with age. J Comp Neurol 228: 557±570

Zakon HH (1986a) The electroreceptive periphery. In: Bullock TH,Heiligenberg W (eds) Electroreception. Wiley, New York,pp 225±256

Zakon HH (1986b) The emergence of tuning in newly generatedtuberous electroreceptors. J Neurosci 6: 3297±3308

Zakon HH (1987) Variation in the mode of receptor cell addition inthe electrosensory system of gymnotiform ®sh. J Comp Neurol262: 195±214

Zakon HH (1993) Weakly electric ®sh as model systems forstudying long-term steroid action on neural circuits. Brain Be-hav Evol 42: 242±251

Zakon HH, Meyer JH (1983) Plasticity of electroreceptor tuning inthe weakly electric ®sh Sternopygus dariensis. J Comp Physiol153: 477±487

99