Search

Ampullary organs of the transparent catfish, Kryptopterus bicirrhus, are present in large numbers on the head and in a regular pattern of lines on the body and fins. The organs lie in the epidermis, and have a pore that opens to the surface. Flattened cells form a roof and walls. On the floor of the organ there are a “sensory hillock,” composed of spherical receptor cells and columnar supporting cells, and a “secretory hillock” composed of columnar secretory cells. The receptor cells are nonciliated and have only afferent innervation. The organ cavity is filled with jelly. The organs are compared with ampullary organs of the weakly electric fish Eigenmannia, ampullae of Lorenzini of Raja, and small pit organs of Amiurus. Structural characteristics of the ampullary organs of Kryptopterus make them especially suitable for electrophysiological studies.

Ampullary receptor organs of the South American weakly electric gymnotid fish Eigenmannia virescens consist of a pore at the surface of the skin, a canal through the epidermis, and the expanded basal end of the canal in the corium. The cavity of the organ contains a jelly that is filled with fine fibers. The canal wall consists of three to six layers of flattened cells that appear to be derived from the adjacent skin. Along the lumen of the organ the cells are joined by tight junctions. Usually there are four spherical receptor cells in the base of the organ. They are innervated by single neural terminals. These organs are compared to tuberous receptor organs found in the same species, and the functional significance of the fine structure observed in these cells is discussed.

Both ampullary and tuberous special cutaneous receptor organs are present in the gymnotid fish Hypopomus. The ampullary receptor organs are morphologically very much like those of the related genus, Eigenmannia, in organ structure and receptor cell ultrastructure. The tuberous organs are of three types, two of which are restricted to the anterior part of the fishes' body and one to the posterior part. The anterior tuberous organs differ from the single type found in Eigenmannia primarily in synaptic ultrastructure. The posterior tuberous organs differ in having a long canal between the organ capsule and the epidermis, and in having the receptor cells in a compartment that is separated by a continuous cellular cap from the extracellular spaces of the canal and the overlying epidermis.

A South American gymnotid Eigenmannia, changes the discharge frequency of its electric organ when a weak electric signal is applied to it with a frequency very close to that of its own discharge. Otherwise, the discharge frequency is extremely constant at a fixed temperature. When the frequency of the applied signal is higher or lower than that of the fish discharge, the response is a decrease or increase of the discharge frequency, respectively. When the two frequencies are exactly the same, the response fails to occur. The threshold of this response is very low. In one fish it was about 3 μV./cm.

It has recently become evident that sensory thresholds for certain tasks are lower than those expected from the properties of individual receptors. This perceptual capacity, termed hyperacuity, reveals the impressive information-processing abilities of the central nervous system. Although much is known about spatial hyperacuity, temporal hyperacuity has received little attention. Here we demonstrate that an electric fish, Eigenmannia, can detect modulations in the timing (phase) of an electrical signal at least as small as 400 ns. Such sensitivity exceeds the temporal resolution of individual phase-coding afferents. This hyperacuity results from a nonlinear convergence of parallel afferent inputs to the central nervous system; subthreshold inputs from particular areas of the body surface accumulate to permit the detection of these extremely small temporal modulations.

1. 1. A device was designed for accurate and automatic measurement of the frequency/temperature relationship of the electric discharge in a weakly electric fish, Eigenmannia sp. 2. 2. The observed hysteresis for increasing and decreasing temperatures was in the range of 2–4 per cent, maximal 10 per cent. 3. 3. The low values of hysteresis explains, in this respect, the negative results of Enger & Szabo (1968) who used measuring methods which were not sufficiently accurate.

The distribution of neurons and fibres containing monoamines (MA, that is noradrenaline, dopamine, and serotonin) was studied in the brain of the weakly electric fish, Eigenmannia lineata, by formaldehyde-induced fluorescence histochemistry. Control experiments using a catecholamine-precursor (Dopa), a MA-depleting (reserpine), and a MA-accumulating (nialamide) drug, together with microspectrofluorometric measurements, allowed to distinguish specific (MA) from unspecific (for example lipofuscine) fluorescence. High Performance Liquid Chromatography (HPLC) assays of MA of untreated brains yielded 313 ± S.E. 24 ng/g brain tissue noradrenaline, 96 ± S.E. 19 ng/g dopamine, and 71 ± S.E. 12 ng/g serotonin (N = 12 in each case). The highest density of MA-perikarya was found, as in other species, in the diencephalon: in the region of the nucleus praeopticus periventricularis (npp), in the nucleus ventromedialis (nvm), in the nucleus posterior periventricularis (nppv), and in the hypothalamic nucleus recessus lateralis (nrl) and nucleus recessus posterior (nrp). The mesencephalic torus semicircularis, the "super-laminated" key structure of sensory-motor integration in gymnotids, contained monoamines in the large T-cells of layer VI. Both types of cells of the medullary pacemaker nucleus (pm) of the electric organ, the smaller pacemaker and the larger relay cells, showed specific, yellow fluorescence indicating the presence of monoamines. Fluorescent MA-fibres in the frontal, medial and lateral parts of the telencephalon connected with hypothalamic and midbrain areas. Several electroreceptive layers of the torus semicircularis (layers V, VIII, IX, XI, XIII) showed a clearly recognisable and in some parts dense network of MA-fibres. Only few fluorescent fibres occurred in the massive corpus cerebelli (which is mixed electro-/mechanoreceptive), in the second order electroreceptive lobus lineae lateralis posterior and in the medulla oblongata (ventrolateral to the raphe region, especially in the frontal medulla at the level of the torus).

The genus Sternopygus, believed to reflect ancestral traits of gymnotiform electric fish, is closely related to the more 'modern' genus Eigenmannia (Mago-Leccia 1978; Fink and Fink 1981). Sternopygus is the only known genus of electric fish that does not perform a jamming avoidance response (JAR) to minimize the potentially detrimental effects of signal interference between discharging neighbors (Bullock et al. 1972, 1975), and its ability to electrolocate objects is rather immune to jamming (Matsubara and Heiligenberg 1978). By studying the responses of midbrain neurons to stimulus regimes effective in eliciting the JAR in Eigenmannia, we found that Sternopygus has neurons capable of discriminating the sign of the difference frequency between interfering electric organ discharges (EODs). These 'sign-selective' neurons, which are believed to be important elements in the control of the JAR in Eigenmannia, may, therefore, fulfill a more general function in the detection of moving objects and conspecifics but could potentially be assembled for the evolution of a JAR in Sternopygus. The relative immunity to jamming in this genus may result, in part, from a stronger reliance upon the ampullary electrosensory system which operates in the DC and low-frequency range, outside the EOD spectrum of these fish.

Neuronal reliability and sensitivity to behaviorally relevant stimulus patterns were investigated in a higher-order nucleus of the diencephalon believed to participate in the jamming avoidance response (JAR) of the weakly electric fish, Eigenmannia. The fish raises or lowers its frequency of electric organ discharge (EOD) to minimize interference from a neighboring fish's EOD. Proper JARs require determination of the sign of the difference frequency (Df) between the neighboring fish's EOD and the fish's own EOD. Bastian and Yuthas (1984) recently described diencephalic neurons within the nucleus electrosensorius that are able to make this determination. In the present study, response properties of such neurons were compared with those of lower-level 'sign-selective' cells found in the torus semicircularis and the optic tectum (Heiligenberg and Rose 1985) as well as with properties of the intact behavior. Most sign-selective cells within the nucleus electrosensorius show a high degree of selectivity for one sign of the difference frequency; cells with either sign preference were found in approximately equal numbers. The sign preference and the degree of sign selectivity is most often independent of the spatial orientation of the jamming stimulus. In contrast, the responses of toral and tectal cells are less robust and consistent and are often highly dependent on the geometry of the jamming stimulus. Determination of the sign of the difference frequency requires the analysis of amplitude modulations coupled with modulations in phase (timing) differences between pairs of areas of the body surface. The most sensitive cells recorded in the nucleus electrosensorius can determine the sign of the difference frequency with timing differences of 1 microsecond or less, roughly comparable to the behavioral threshold of 400 ns (Carr et al. 1986). The best toral/tectal response required at least a 16 microseconds modulation. Cells within the nucleus electrosensorius thus code the sign of Df with a high degree of reliability and sensitivity. Ambiguities persist, however, which suggest that single cells at this level cannot completely account for the behavioral discrimination. Additional processing may be necessary to transform a still primarily sensory code into a motor program for control of the JAR (Rose et al. 1988).

1. A spontaneous EOD response to sound is described in two gymnotoids of the pulseElectricOrganDischarge (EOD) type,Hypopomus andGymnotus, and in one mormyrid,Brienomyrus (Figs. 2–4). 2. In all three species, the EOD response to the sound onset was a transient EOD rate increase. In the low EOD rateHypopomus (3–6 EODs/s at rest) the first, second, or third EOD interval following sound onset was significantly shorter than the average EOD interval before stimulation. The shortest latency found was 100 ms, the longest ca. 1.2 s.Gymnotus (around 50 EODs/s at rest) responded similarly, but the third interval after sound onset was the first to be affected even at highest intensities (shortest latencies approx. 60 ms; latencies > 0.5 s at low sound intensities). InBrienomyrus (4–8 EODs/s at rest) the response occurred already at the first EOD interval after sound onset. 3. An EOD sound response was recorded inHypopomus and inGymnotus up to 5,000 Hz sound frequency (in oneGymnotus individual: up to 7,000 Hz). Due to technical limitations the low frequency limit of the response could not be exactly determined: the fishes responded well even below 100 Hz.Hypopomus had its maximum sensitivity around 500 Hz (Fig. 5),Gymnotus around 1,000 Hz (Fig. 6). 4. In all three species the EOD sound response was graded with sound intensity (Hypopomus: Fig. 7). 5. No EOD response to sound was found in two gymnotoids of the wave type,Eigenmannia andApteronotus, and in the gymnotoid pulse fishRhamphichthys. A criterion is proposed by which it should be possible to predict whether or not a weakly electric fish species will show the EOD sound response. 6. It is concluded that the EOD response to sound is similar to EOD responses to other kinds of stimulation (light, touch, vibration, food, and even electrical). The possible biological function is discussed.

A total of 47 tectal neurons of the weakly electric fish, Eigenmannia , were studied physiologically and labelled by intracellular injection of Lucifer Yellow. With the exception of two cell types, all cells could be classified in accordance with the Golgi studies of Sas and Maler. 25 The dominant stimulus modality of neurons was correlated with their laminar location. Neurons of the stratum opticum only responded to visual stimuli, such as modulations of the light level or the motion of an object. They showed, however, no directional preferences for motion. Neurons of the stratum griseum centrale were predominantly driven by electrosensory stimuli, most often those associated with the movement of an object, and generally were very sensitive to the direction of motion. Integration of different sensory modalities was found in neurons with dendrites invading laminae with different sensory inputs. In addition, small axons of interneurons appear to relay information across laminae. Large multipolar neurons in the deep tectum responded to the motion of objects, often preferring a particular direction of motion. Some of these large multipolar neurons of the deep tectum also discriminated the sign of the frequency difference between a mimic of a neighbor's sinusoidal electric organ discharge and the animal's own signal. These neurons are potential candidates for the control of the jamming avoidance response. These neurons were morphologically indistinguishable from large multipolar neurons of the deep tectum that either responded to moving objects or to acoustical stimuli. Individual large cells of the deep tectum project to various targets (Fig. 1) and probably contribute to the control of different behavioral responses. This suggests that the nature of such responses would then depend upon the constitution of sets of neurons recruited by a given stimulus situation, and the role of individual tectal neurons would neither be particularly specific nor very significant.

SUMMARY The fine structure of and the synaptic connections between the nerve elements of the pacemaker nucleus (MCN) of three weakly electric fish ( Sternarchus albifrons , Eigenmannia sp., Hypopomus sp.) have been investigated. The MCN of Sternarchus and Eigenmannia contains two different types of cell (small and large), whereas in Hypopomus , an additional, second small-sized neuron was observed. In all three species the axons of the large cells leave the nucleus without branching, but these neurons do differ from each other in respect to their dendritic morphology. In contrast, small cells of Sternarchus and Eigenmannia Aave branching axons and a rich dendritic tree. Small cell axons terminante with club endings (with gap junctions in Sternarchus and with mixed synapses in Eigenmannia and Hypopomus ) on relay cells and they also interconnect with one another. On the other hand, afferent thin myelinated fibers of extranuclear origin form chemical synapses on both small and large cells. The distribution of the synaptic terminals on the neurons shows, however, certain differences. In support of earlier physiological data, it can be concluded that in Sternarchus and Eigenmannia the large (relay) cells are synaptically activated by small (pacemaker) neurons, while afferent inputs of extranuclear origin may play a modulatory role for MCN activity. In Hypopomus , the MCN displays a more complicated cellular organization and synaptic network; this is in agreement with the more complex EOD behavior of this fish.

The electrosensory system of weakly electric gymnotiform fish is described from the receptor distribution on the body surface to the termination of the primary afferents in the posterior lateral line lobe (PLLL). There are two types of electroreceptor(ampul1ary and tuberous) and a single type of lateral line mechanoreceptor (neuromast). Receptor counts in Apteronotus albifrons show that (1) neuromasts are distributed as in other teleosts; (2) ampullary receptors number 151 on one side of the head and 208 on one side of the body; (3) tuberous receptors were estimated to number 3,0004,500 on one side of the head and 3,500-5,000 on one side of the body. The distribution of each receptor type is described. Each receptor is innervated by a single primary derent. Electrosensory afferents have myelinated cell bodies in the ganglion of the anterior lateral line nerve (ALLN). The distribution of these ganglion cell diameters is strongly bimodal in Apteronotus and Eigenmannia: The smaller-diameter cells may be those which innervate ampullary electroreceptors, the largerdiameter tuberous electroreceptors. Transganglionic HRP transport techniques were used to determine the first-order connections of the anterior lateral line nerve in six species of gymnotiform fish. Small branches of the ALLN were labeled so as to determine the somatotopic organization in the PLLL. The PLLL is divided into four segments from medial to lateral, termed medial, centromedial, centrolateral, and lateral segments (Heiligenberg and Dye, '81). Representations of the head are found rostrally in each zone, and the trunk is mapped caudally in each zone. Thus there are four body maps in the PLLL. The medial segment receives ampullary input (Heiligenberg and Dye, '82) and maps the dorsoventral body axis mediolaterally, as does the tuberous centrolateral segment. The tuberous centromedial and lateral segments map the dorsoventral axis lateromedially. Thus the medial and centromedial segments meet belly to belly, the centromedial and centrolateral segments meet back to back, and the centrolateral and lateral segments meet belly to belly. Adjacent electrosensory maps within the PLLL are therefore always mirror images. The octavolateralis system of teleosts is composed of the lateral line, auditory, and vestibular systems. In teleosts with electrosensory capabilities, the lateral line includes both mechanosensory and electrosensory modalities. In the weakly electric teleosts of South America (Gymnotiformes), electroreceptors and mechanoreceptors are distributed over the body surface and are innervated by the primary afferent fibers of the octavolateralis system. These fibers enter the brain via the anterior lateral line nerve

A fish of the genus Eigenmannia responds differently to a neighboring conspecific fish with a slightly higher frequency of the electric organ discharge than its own than to one with a slightly lower such frequency than its own. When the two frequencies are beating against each other the special wave shape of the electric organ discharge leads to asymmetries of the beat pattern which are distinct for the two cases. Midbrain neurons, called "deltaF recoders," sign and magnitude of the frequency difference on the basis of these patterns. that is, in the time rather than the frequency domain.

SUMMARY We collected weakly electric gymnotoid fish in the vicinity of Manaus, Amazonas, in the Solimoes river (white water). We tried to find out whether Electric Organ Discharges (EODs) are species specific which is essential for their presumed role in recognition of conspecifics and reproductive isolation. We considered at least 43 valid sympatric species, some of them unnamed. All of these displayed stable EOD waveform patterns, most of them clearly distinct from the other species' EODs. Eleven species are of the pulse EOD type, 32 of the wave EOD type (one of the latter is intermediate). The EODs of pulse species were analysed (1) by EOD repetition rate at rest (variation from 1 Hz to 60 Hz), (2) by Fourier amplitude spectrum analysis of single EODs ( Fig. 1 ; in these spectra, frequencies of peak amplitude ranged up to 2300 Hz). There was a significant, positive correlation between both parameters ( Fig. 2 ). Identification of pairs of species with similar EODs by these parameters does not appear to be possible because of inter-individual EOD variations. In wave species there is conclusive evidence that EOD fundamental frequencies (= repetition rate of a complete EOD period) do not allow species identification: twenty-eight wave species displayed EOD fundamental frequencies from 300 to 1800 Hz ( Fig. 3 ). This yields a hypothetical frequency band of 0.09 octave to signal species identity; the actual value of EOD frequency variations in Eigenmannia is much greater (1.2 octaves). Seven species of the family Apteronotidae displayed a new signal type: the main energy of the signal was contained in higher harmonics, and not at the fundamental frequencies ( Figs. 6 and 7 ). For all wave species there was a significant, positive correlation between their dominant frequency (= the strongest signal component) and harmonic content of their EOD although individual species deviated considerably from what was predicted by the regression line ( Fig. 8 ). Thus separation of species was greatly improved compared to the criterion of fundamental frequency ( Fig. 3 ) but still appeared insufficient in a number of cases. Therefore, in both wave and pulse species still other parameters must be involved in recognition of conspecifics.

The mesencephalic torus semicircularis in gymnotid electric fish has been identified as a key structure for electrosensory signal processing and control of the electric pacemaker (Scheich 1974, 1977c; Bastian and Heiligenberg 1980). In terms of relative volume, it is the largest torus in any vertebrate, and in Eigenmannia it can be subdivided into 15 cell and fiber laminae (see Sect. 2). The layers from the right and left toms are continuous across the midline. Some of the dorsal layers receive bilateral input from the posterior lateral line lobe (LLLP) of the rhombencephalon, the first target of the electroreceptive afferents of the lateral line nerve (Ebbesson and Scheich 1980; Scheich and Ebbesson 1981; Carr et al. 1981). These recent anatomical accounts have also substantiated the findings of Scheich and Maler (1976) and Scheich (1977c) that units in superficial layers of the torus represent a topographic map of the fish’s electroreceptive surface. There is a point-to-point projection to the torus from LLLP where several topographic representations of the fish’s surface are believed to be present (Maler et al. 1974, Maler 1979; Carr et al. 1981).

1. 1. The frequency of the electric organ discharge (EOD) of Eigenmannia virescen changes with water temperature and with the presence of an electrical sine wave stimulus of a neighboring frequency. The latter response, called the jamming avoidance response, was observed at various temperatures ranging from 18 to 30°C. 2. 2. Jamming avoidance response (JAR) in Eigenmannia, characterized by a shift in the fish's EOD frequency away from the electrical stimulus frequency, increased with rising temperature. 3. 3. The Q10 of the resting EOD frequency and JAR were different in their absolute values and variation patterns with temperature. 4. 4. These results suggest that the neuronal mechanisms for the 2 behavioral responses are probably independent and the integration site is likely to be at the pacemaker cells.

Knifefish of the genus Eigenmannia (Gymnotiformes, Teleostei) are seasonal breeders that spawn only during the tropical rainy season. Both sexes modulate their otherwise constant wave-like electric organ discharges (EODs) in the context of courtship and aggressive behavior by “chirps,” abrupt frequency modulations with brief interruptions. Play-backs of recordings of male courtship chirps can induce spawning in gravid females (Hagedorn and Heiligenberg, 1985). The EOD, produced by a specialized electric organ, is under the control of a pacemaker nucleus (Pn) in the medulla oblongata. Injections of HRP into the Pn label only a small cluster of cells bilaterally at the boundary of diencephalon and mesencephalon, constituting the prepacemaker nucleus, PPn (Heiligenberg et al., 1981). Microstimulation experiments have shown that chirp-like EOD modulations can be elicited from a subnucleus of the PPn, the PPn-C (Kawasaki and Heiligenberg, 1988; Kawasaki et al., 1988). By retrograde HRP labeling, we investigated the dependence of the PPn's morphology upon the sexual maturity of the fish. The most prominent effect was that, during the breeding season, females developed an abundance of “varicosities,” swellings 1–4 microns in diameter in distal regions of dendrites at a density of approximately 1/10 microns. In contrast to mature females, immature females had none or only a few varicosities. Such a clear correlation between the abundance of varicosities and relative gonadal weight was not found in males, most likely because testicular and dendritic development are not strictly synchronized. After the onset of the simulated dry season, however, relative gonadal weight, as well as the number of varicosities, was reduced drastically in both sexes. This reduction in the number of varicosities is accompanied by a decrease in their diameter. Varicosities may represent growing buds or regions of synaptic input from afferent areas or both. Reconstruction of individual PPn-C neurons showed maturity-dependent changes in the pattern of dendritic proliferation in females. Such dynamic changes in the structure of neurons might subserve seasonal modifications in an animal's propensity to execute specific behaviors.

1. In the single electric organ discharge (EOD) of Eigenmannia and Sternopygus the duration of the head-positive EOD half-wave (P) was found to be shorter than the duration of the head-negative half-wave (N). There were individually different P/N ratios in both species. High P/N ratios were found in Sternopygus discharging at high repetition rates, and small P/N ratios were correlated with low EOD frequencies. Furthermore, the data presented in Figure 1 suggest a correlation between the P/N ratio and the sex in Sternopygus. Fig. 1. P/N ratios of five Sternopygus (open circles) and eleven Eigenmannia (filled squares) are plotted against the EOD frequency (F). The EOD wave shapes in both species are shown to the left. There is a sinusoidal head-positive half-wave with the duration P and a trapeziform head-negative half-wave with the duration N. The P/N ratio is 0.6 in the lower EOD and 0.9 in the upper EOD. Both signals have the same duration (EOD wave length) T. Figure options Download full-size image Download as PowerPoint slide 2. In beats composed of EODs and sine waves with frequencies slightly different from that of the EOD (1 F stimuli) or from the second harmonic of the EOD (2 F stimuli) successive beat half-waves increase their amplitudes from the beat minimum to the beat maximum, and they decrease their amplitudes from the beat maximum to the minimum. With respect to corresponding zero crossings of the EODs, zero crossings of beat half-waves come earlier in one section of the beat cycle (= phase lead), and they are delayed in the other section of the beat cycle (phase lag) (Fig. 2, A, C). In beats composed of EODs and sine waves with frequencies close to the first subharmonic of the EOD (1/2 F stimuli) successive beat half-waves alternate high with low amplitudes and phase lead with phase lag (phase alternation) (Fig. 2D). Fig. 2. Comparison of pure EODs with amplitude and phase patterns in beat waves caused by the superposition of EODs and different stimuli. EODs (solid lines) and stimuli (dotted lines) are presented in the upper traces. Beat waves are shown in the middle, and sections of phase modulation of positive beat half-waves and of negative beat half-waves are illustrated below the beat waves. The length of one beat cycle is indicated by the length of the two complementary sections of phase modulation. The temporal displacement between zero crossings of EODs and beat waves is given by vertical lines: solid lines indicate zero crossings of the EODs from positive to negative polarity, and dashed lines indicate EOD zero crossings from negative to positive polarity. Time runs from left to right. EODs are added to sine waves (A) and EOD-mimicking stimuli (B) at frequencies below the EOD frequency as well as to sine waves below the second harmonic (C) and the first subharmonic (D) of the EOD. Interactions of EODs and EOD-mimicking signals above and below the first subharmonic of the EOD are shown in (F) and (E). Figure options Download full-size image Download as PowerPoint slide 3. In beats composed of EODs and EOD-mimicking signals (the P/N ratios are smaller than 1) with frequencies close to the first subharmonic of the EOD, the duration of the head-positive stimulus half-wave (P stim) is proportional to the duration of the section of phase alternation within the head-positive and the head-negative beat cycle. The duration of the section of phase lag in the positive beat cycle and of phase lead in the negative beat cycle is proportional to T eod (EOD interval) minus P stim (Fig. 3). The quotient P stim / T eod - P stim is called the phase quotient of the beat. If the phase quotient is 1, sections of phase lead, phase lag and phase alternation have the same duration, and every second beat half-wave increases from the beat minimum to the beat maximum, and decreases from the beat maximum to the beat minimum. The more the phase quotient is smaller than 1, sections of phase lead and phase lag increase at the expense of sections of phase alternation, and the relative number of beat half-waves with nearly constant amplitudes increases at the expense of the amplitude modulated beat half-waves. The more the phase quotient is greater than 1, the more the sections of phase alternation increase, and the more successive beat half-waves alternate high with low amplitudes (Fig. 4, lower part). Fig. 3. Sections of phase modulation in a 1/2 F beat cycle are illustrated by means of stimulus phase shifts in certain phases of the EOD. Three stimuli of nearly twice the EOD wave length are shown above and below an EOD trace, respectively. Vertical lines in the upper part of the figure indicate EOD zero crossings from negative to positive polarity. Vertical lines in the lower part indicate EOD zero crossings from positive to negative polarity. The origin of negative phase sections is illustrated in the lower part, and the origin of positive phase sections is shown in the upper part of the figure: during a stimulus phase shift from the position at the upper right to the position in the upper middle or vice versa, EOD zero crossings coincide only with phases of the negative stimulus half-wave. Resulting beat waves would be delayed (lag). The phase shift is proportional to the difference between the EOD interval (T eod) and the positive stimulus half-wave (P stim). During a stimulus phase shift from the upper middle to the upper left or vice versa, every second EOD zero crossing coincides with a phase of the negative stimulus half-wave, and all others coincide with phases of the positive stimulus half-wave. Resulting beat waves would alternate, the phase lag with the phase lead (alt). The phase shift is proportional to P stim. The two complementary phase sections are proportional to T eod. (from negative to positive polarity) Figure options Download full-size image Download as PowerPoint slide Fig. 4. Upper part: JAR values in two Eigenmannia as a function of phase quotients in 1/2 F beats. Solid line: responses to seven beats/s and a positive phase shift of the stimuli (F eod = 441 Hz, (P/N) eod = 0.78). Dashed line: responses to two beats/s and negative phase shift of the stimuli (F eod = 428 Hz, (P/N) eod = 0.64). Lower part: three 1/2 F beat cycles with different phase quotients (1/4 in the left, 1 in the middle and 4/1 in the right beat). Sections of phase modulation in positive and negative beat half-waves are shown to the right of each beat. Below the beats, corresponding constituent signals are presented: EODs (solid lines) and 1/2 F stimuli (dotted lines). There is a positive phase shift of the stimuli. Time runs from bottom to top. Figure options Download full-size image Download as PowerPoint slide 4. The Jamming Avoidance Response (JAR) in Eigenmannia was found to be a maximum function of the phase quotient of beats composed of EODs and EOD-mimicking signals close to the first subharmonic of the EOD. The response was graded on both sides of a best phase quotient of about 1 (P stim was about half the EOD interval). There was no systematic response when the phase quotient was very small or when P stim was about T eod, i.e., when the stimulus matched the temporal feature of subharmonic sine and square waves (Fig. 4, upper part). 5. The temporal displacement between the head-positive and the head-negative beat maximum was found to depend not only on the direction of the relative phase shift between the EODs and the stimuli but also on the position of the positive and the negative maxima in the EOD and in the stimulus. Therefore, reliable information about whether the simulus frequency is above or below that of the EOD, will not be provided to Eigenmannia by the temporal sequence of positive and negative beat maxima alone (Fig. 5). Fig. 5. Temporal sequence of beat maxima as a function of the stimulus phase shift and the position of maxima in the interval of EOD and stimuli. An EOD trace is presented in the middle of the figure. The time between the positive and the negative maximum in the EOD is longer than the corresponding time in the stimulus shown above the EOD, and it is shorter than the corresponding time in the stimuli shown below the EOD. Vertical solid lines indicate positive and negative EOD maxima. Vertical dashed lines indicate negative stimulus maxima. The EOD-mimicking stimuli are drawn incomplete and might be close to once or twice of the EOD wave length. Sine waves are shown at the bottom of the figure. All stimuli are in antiphase to the EOD in the middle of the figure (position, yielding a beat minimum). To the left and to the right positive stimulus maxima coincide with positive EOD maxima (position, yielding positive beat maxima). Numbers indicate the temporal sequence of maxima in the beat cycle, e.g., at a positive shift of the stimulus, shown above the EOD trace, the coincidence of the negative maxima of EOD and stimulus (negative beat maximum) is earlier (1.) than the coincidence of the positive maxima of EOD and stimulus (positive beat maximum) (2.). Arrows indicate the direction of stimulus phase shifts. At a negative shift (neg. shift), the stimulus frequency is below, and at a positive shift (pos. shift), the stimulus frequency is above that of the EOD or the first subharmonic of the EOD. Figure options Download full-size image Download as PowerPoint slide

Neurons in the torus semicircularis of the weakly electric fish Eigenmannia encode phase differences between sinusoidal electrical stimuli received in different body regions. These fish normally experience time-varying phase differences when the electric organ discharge fields of two or more individuals overlap. These phase differences supply information necessary for the animal's jamming avoidance behavior.

In order to identify which brain centers are involved in the control of electric organ discharge in electric fish, HRP was injected into the medullary electromotor (pacemaker) nucleus in the gymnotoid Eigenmannia. Neurons were retrogradely labeled in only a small nucleus of the mesencephalic tegmentum, herein called the prepacemaker nucleus. The prepacemaker nucleus lies just caudal and ventral to the posterior commissure and comprises at least two types of neurons.

South-American weakly electric fish are known to use a self-generated electric field in conjunction with electroreceptors in the skin, to localize prey and to communicate. In a way similar to echolocation in other animals electrolocation in these fish relies on an active sensory system which is provided with its own energy source, various feedback mechanisms, and a functional match of components for signal processing (see reviews by Scheich and Bullock 1974, Heiligenberg 1977, Bullock 1982). In the case of the genus Eigenmannia the discharge of the electric organ, which is driven by a pacemaker in the brainstem, has a maximum intensity of a few mV and a fundamental frequency of between 200 and 400 Hz. Each fish has a stable private frequency to which electroreceptors are tuned. The geometry of the electric field is that of a distorted dipole with more resolution of field lines in the anterior half of the fish (Fig. 1).

The general organization and synaptology of the medullary command (pacemaker) nucleus (MCN) was investigated in the high frequency weakly electric fish, Apteronotus leptorhynchus. This study was undertaken in order to establish differences and similarities between the MCN of A. leptorhynchus and that of the closely related species, Apteronotus albifrons which has been studied previously. The basic morphology and synaptology of the MCN in A. leptorhynchus is similar to that of A. albifrons. The MCN of A. leptorhynchus consists of large (relay) and small (pacemaker) cells; both cell types receive synaptic input or large club endings with electrotonic gap junctions and bouton-like terminals with polarized chemical synapses. Club endings originate from thick meyelinated fibres belonging to the small (pacemaker) cells, whereas the bouton-like terminals issue from thin myelinated fibers of extranuclear origin. Via their club endings, the small (pacemaker) cells are connected both to each other and to the large (relay) cells. Besides the similarities, there are distinct and characteristic differences between the MCN of the two species, which mainly concern the synaptology of the nucleus. In A. leptorhynchus, the large (relay) cells possess long dendritic processes, covered exclusively with bouton-like terminals; the axon initial segment of large (relay) cells receives boutons, in addition to club endings. Small (pacemaker) cells have short dendritic protrusions receiving input from club endings and boutons; furthermore, the small pacemaker cells axon initial segment receives both club endings and bouton-like terminals. These differences are discussed in terms of the functional organization of the MCN in certain gymnotoids and draw attention to the fact that the morphological and ultrastructural aspects of the central command of the electric organ discharge reveal several differences not only between different gymnotoid fish (Apteronotus and Eigenmannia) but also between closely related species such as A. albifrons and A. leptorhynchus.

Structure-function relations in the electrosensory system of Eigenmannia were examined by labeling physiologically characterized neurons of the dorsal torus semicircularis. The sensitivity of cells to modulations in amplitude (AMs) and modulations in differential phase was determined. Approximately half of the 48 cell types defined by Golgi studies (Carr, C.E., and L. Maler (1985) J. Comp. Neurol. 235: 207–240) were identified in this manner. The majority of the neurons located in laminae (8a, b, c, and d, and 9 exhibited sensitivity to differential phase. In laminae 5 and 7, however, in addition to neurons which were sensitive to differential phase, many cells were found that were purely AM sensitive. Differential phase sensitivity originates in the small cells of lamina 6 (Heiligenberg, W., and G. Rose (1985) J. Neurosci. 5: 515–531), the exclusive termination site of phase-coding afferents from the electrosensory lateral line lobe. Cells that had dendritic extensions into the neuropil of lamina 6 exhibited sensitivity to differential phase, whereas neurons lacking dendrites in this lamina were only excited by AMs. These findings support the notion of a relationship between the morphology and laminar position of a neuron and its function.

The tropical South American teleost Eigenmannia lineata showed a spontaneous preference for the female type, compared with the male type, of its sexually dimorphic, weak-electric organ discharge (EOD). Female and male EODs differ in waveform and harmonic content. An isolated fish was simultaneously stimulated with digitally synthesized "natural" male and female EODs of equal peak-to-peak amplitudes, at _+35 Hz frequency difference centered on its stable resting discharge frequency. The stimulus dipoles were arranged symmetrically to the right and left of the fish's hiding place. All stimulus conditions were permuted at random sequence. Among 11 fish tested, 8 showed a statistically significant preference for one stimulus, the female type, as measured by the amount of time a fish spent close to a stimulus dipole (P< 0.05 in each fish, two-tailed). Thus female EODs rather than male EODs were more attractive to adult and juvenile fish of both sexes. It was also concluded that E. lineata is capable of discriminating female from male EODs by a complex sensory capacity requiring neither amplitude nor frequency cues. The EOD waveform changed very little within the ecological range of water conductivities (approximately 10-100 gS.cm 1); the P/N-ratio (a waveform character based on zerocrossing intervals) depended only weakly, but significantly, on conductivity (negative correlation in all four fish). Also, the effect of temperature on EOD waveform was very weak: Q~o-values of the P/N-ratio were below but close to I in all fish (27_+5 ~ C). Thus, it can be concluded that the EOD waveform is remarkably stable within widely changing conditions - even beyond the variation found in the field - and is therefore potentially useful as a social cue.

Behavioral experiments show that the weakly electric fish, Eigenmannia, detects differences in timing as small as 400 nsec between electric signals from different parts of its body surface. The neural basis of this remarkable temporal resolution was investigated by recording from elements of the phase-coding system, a chain of electrotonically connected neurons devoted to the processing of temporal information. Each element of this system fires a single action potential for every cycle of the electric signal (either the fish's own electric organ discharge or a sinusoidal signal of similar frequency). For phase- coding primary afferents and midbrain neurons, the temporal resolution was determined by measuring each unit's capacity to lock its spike to a particular phase of the stimulus cycle. The jitter of a neuron's response (measured as the standard deviation of the timing of the spikes with respect to the stimulus) decreases from the level of the primary afferent (mean = 30 microsec) to the midbrain torus (mean = 11 microsec); these results can be correlated with morphological measures of convergence. The temporal resolution of single neurons is still inferior to that displayed at the behavioral level.

Animals respond to only a relatively small fraction of stimuli in the environment. A recurring question facing investigators of sensory processing is how the nervous system filters and processes information to recognize behaviorally relevant stimuli. An attractive working hypothesis is that systems are hardwired in such a way that a particular unit or class of units responds when the animal encounters the appropriate stimulus or combination of stimuli for a particular behavior. Recent studies have supported an alternate hypothesis, based upon quite different design criteria. Rather than discrete feature detectors tuned to, for instance, the fundamental frequency of a mating call, investigators have found populations of cells with a wide range of tunings but with a preponderance of cells firing in response to a particular range of the behaviorally relevant variable (R.R. Capranica, this volume).

An analysis of the nucleolar organizer region (NOR) by silver staining of five species of the order Gymnotiformes, Gymnotus carapo, Apteronotus albijrons, Sternopygus macrurus, Eigenmannia virescens, and Eigenmannia sp., is reported. The five species presented only one pair of homologs bearing NORs. In G. carapo, A. albijrons, and S. macrurus the silver-stained NORs were present as small black dots over the chromosomes. In E. virescens, the short arm of one of the largest chromosome pairs was entirely stained with silver. Eigenmannia sp. presented an intraspecific variability of the NOR-bearing pair and an increase in the length of this region, the larger one being about six times the size of the smaller. These findings suggest that an increase in the number of ribosomal genes in the NORs of certain species occurred during the evolution of this group.

During their jamming avoidance response (JAR), weakly electric fish of the genus Eigenmannia shift their electric organ discharge (EOD) frequency away from a similar EOD frequency of a neighboring fish. The behavioral rules and neural substrates for stimulus recognition and motor control of the JAR have been extensively studied (see review by Heiligenberg 1986). The diencephalic nucleus electrosensorius (nE) links sensory processing within the torus semicircularis and optic tectum with the mesencephalic prepacemaker nucleus which, in turn, modulates the medullary pacemaker nucleus and hence the EOD frequency. Two separate areas within the nE responsible for JAR-related EOD frequency rises and frequency falls, respectively, were identified by iontophoresis of the excitatory amino acid L-glutamate. Bilateral lesion of the areas causing EOD frequency rises resulted in elimination of JAR-related frequency rises above a baseline frequency obtained in the absence of a jamming stimulus. Similarly, bilateral lesion of the areas causing frequency falls resulted in a loss of JAR-related frequency falls below the baseline frequency. Whether these areas are also responsible for non-JAR-related frequency shifts is not known. The strength of response and spatial extent of the areas causing frequency shifts varied among fish and also varied in individual fish, reflecting the strength of JAR-related frequency shifts and the balance of activities in frequency-rise and frequency-fall areas. Local application of bicuculline-methiodide or GABA demonstrated a tonic inhibitory input to each area and suggests a reciprocal inhibitory interaction between the two ipsilateral areas, possibly accounting for much of the individual plasticity. The nE thus is a site for neuronal transformation from distributed, topographically organized processing within the laminated structures of the torus and tectum to discrete cell clusters which control antagonistic motor responses.

The cytoarchitecture of the dorsal nucleus praeeminentialis in two families of weakly electric fish (Eigenmannia viriscens and Apteronotus albifrons) was examined in both Nissl and Golgi material, and an attempt was made to correlate this information with our data from HRP studies on the afferent and efferent connections of this nucleus. The n. praeeminentialis is an isthmic structure located dorsolateral to the lateral lemniscus and anterior to the eminentia granularis-a subdivision of the archicerebellum of fish. The n. praeeminentialis can be divided into a large dorsal portion concerned with electroreception and a small ventral portion involved with the lateral line mechanoreception. The dorsal n. praeeminentialis consists of three parts: a pars medialis, a large pars principals (p.P.) and a narrow pars lateralis. The p.P. presents three zones: a dorsal, a central, and a ventral zone, which are reciprocally and topographically connected with the zones of the electroreceptive lateral line lobe (ELLL), medial ELLL with ventral zone, central ELLL with central zone, and lateral ELLL with dorsal zone. Several types of projection cells are present in the n. praeeminentialis: (a) neurons that show preferential orientation of their long dendrites in relation to the afferent fiber systems, (b) cells with wide dendritic fields radiating in all directions, and (c) cells with small polarized dendritic fields toward the incoming ELLL afferents. Interneurons are also identified, showing different axonal ramification patterns. The afferent and efferent fiber systems linking the n. praeeminentialis to ELLL, lobus caudalis (L.C.), and torus semicircularis (T.S.d.) point to the important position of this nucleus in the feedback loop of the electrosensory pathway. The complex processing within this nucleus is reminiscent of the feedback loops in the auditory system.

Using a frequency difference (Δ F) clamp to maintain a stimulus and frustrate the normal escape from a jamming frequency, the response is found to be a characteristic function of the ΔF between stimulus and fish (Figs. 2, 3, 4, 6). It is graded on both sides of a best ΔF of about 3 Hz (= 0.3% in Sternarchus, 1.0% in Eigenmannia). There is no systematic response when F stimulus = F fish, regardless of phase.

THE PROPOSITION THAT neurons can encode information about the intensity of stimuli in other ways than by the familiar gradation of intervals between impulses has been put forward (5, 6, 12). Perkel and Bullock (23) catalog the known and the candidate codes by which neurons can represent information in streams of nerve impulses. For none of these, other than the familiar frequency coders, is a quantitative characterization available of the behavior of the impulse train as a function of intensity.

The tuberous organ, a specific cutaneous organ of Gnathonemus , represents one electroreceptor type in these weakly electric fish of Mormyridae. The structure of the sensory cell, as well as the presence of specific accessory cells, differentiates them from the common sense organs bearing hair cells as sensory elements. In the tuberous organ each sensory cell lies with almost its entire surface free in an intra-epidermal sensory chamber while the basal cell membrane is in contact with the subsensory platform through an opening of the sensory chamber. The neurosensory junctions, the only ones found between sensory cell membrane and nerve endings, have the same characteristics as the afferent sensory junction described in other sense organs. The epithelial cells of the epidermis overlying the tuberous organ have a strongly vacuolated cytoplasm and a particular organization like those described in Eigenmannia . The subsensory platform is composed of four types (I–IV) of cells: three of them display a concentric organization around the nerve axis located between the basement membrane and the sensory cell. The cells of types I and II appear to have an excretory function in relation to the contents of the sensory chamber. The electrophysiological results are discussed on the basis of these morphological data.

Special cutaneous receptor organs of fresh-water weakly electric fish are grossly similar to other receptor organs of the acoustico-lateralis system, but differ in numerous cytological details. They are presumed to function as detectors of weak electric currents. Those of Eigenmannia are of two types that have been designated tuberous and ampullary receptor organs. Parts of the lateral and basal surfaces of the receptor cells are in contact with accessory cells. A single relatively large, almost spherical, nerve ending is inserted between the basal end of each receptor cell and the subjacent accessory cell, lying within a hemispherical concavity in the base of the receptor cell.The presumed synaptic areas between each receptor cell and the nerve ending associated with it are not contiguous, but are morphologically discrete structures, separated by nonsynaptic areas of membrane (Fig. 1). Approximately thirty synaptic areas are found on each receptor cell in both kinds of organs. The surface contours of the apposed cell membranes are irregular in the nonsynaptic areas, and hence the extracellular space is variable in width. In the synaptic areas the membranes of both cells appear slightly thickened (Fig. 2), and are separated by a constant gap of about 200 Å. Fine fibrillar material spans the synaptic gap.

VARIOUS species of fish are known to emit and perceive weak electrical signals1–4. Whereas the physiological mechanisms involved in the production and perception of such signals have been investigated in detail (ref. 3, pages 347 and 493), less is known about their behavioural relevance. Lissman2 proposed and gave evidence for a role in object detection, by the distortion of the fish's own electrical field due to either dielectric or conducting objects within range. This idea has been supported by later work (refs 4–6 and personal communication from A. Kalmijn and R. Adelman). The study reported here was undertaken as a first effort to devise ways of measuring the performance of the object detecting system in respect to its discrimination of particular parameters such as distance, size and motion. A primary goal was to find methods based on built-in responses, rather than time-consuming conditioning procedures, which could be used on any individual on a given day to assess its skill in relation to species norms and would thus facilitate the quantification of changes in performance after nerve transections, brain lesions and such like.

L'hormone melanophoretiqueβ-MSH, a des concentrations de 0.1 a 0.2µg/ml dans l'eau de l'aquarium, produit une diminution de la frequence des modifications spontanees de l'amplitude des decharges electriques du poissonG. eigenmannia. Ces resultats sont en accord avec les observations anterieures montrant divers effets deβ-MSH sur l'activite electrique du systeme nerveux central chez les mammiferes.