fish

We examined the relationship between the spectral quality of the acoustic environment, the courtship sounds and hearing ability of two freshwater gobies, Padogobius martensii and P. nigricans. Sounds of P. nigricans were more intense and had a narrower frequency range than those of P. martensii, yet sound energy peaked at around 100 Hz in both species. Audiograms, obtained with evoked potentials, exhibited thresholds and bandwidth characteristics typical of hearing generalists. Maximum sensitivity occurred at 100 Hz, and thresholds increased sharply at higher frequencies. Hearing abilities in P. martensii were superior to those of P. nigricans in terms of threshold and bandwidth. Species difference in hearing correlated with differences in the amplitude and spectral content of their sounds. Underwater recordings revealed the presence of lowest noise levels at frequencies around 100 Hz, where sound energy in both species was maximal.

Based on studies of a few specialist species (otophysans), the gasbladder is believed to aid in fish audition by transferring pressure components of incoming sound to the ears. We examined the gasbladder's auditory role in goldfish Carassius auratus, an otophysan, and two generalists: the blue gourami Trichogaster trichopterus and oyster toadfish Opsanus tau. Audiograms were obtained with the acoustically evoked potential protocol. X-rays were used to locate the gasbladder, air was withdrawn with a syringe, and tuning curves were obtained from gas-filled and then deflated fish. In goldfish, with Weberian ossicles linking the gas bladder and inner ears, thresholds increased by 33-55 dB (frequency dependent) after deflation. In blue gourami and toadfish, however, thresholds and evoked potential waveforms did not change. These results indicate a variable role for the gasbladder in fish audition and suggest that it may not serve an auditory function in typical teleosts

Paired toadfish Opsanus tau sonic muscles, situated on the lateral surfaces of the swimbladder, commonly contract around 200 Hz to produce the boatwhistle advertisement call. To determine how sounds are produced, we electrically stimulated the sonic nerve and monitored contraction of the sonic muscles with emg electrodes, a laser vibrometer to measure bladder displacement, and a microphone to relate movement to sound production. The muscle fibres are curved like the human diaphragm, and a twitch generates complicated bladder movement suggesting a quadrapole source: the sides push in, increasing internal pressure that pushes the bottom of the bladder out. The acoustic waveform mirrors this movement; a rapid peak of negative pressure (sides in) is followed by a larger positive peak (bottom out). The peak of acoustic pressure occurs in the middle of contraction when velocity is maximal and acceleration is zero. Peak sound amplitude and bladder velocity increase to almost 200 Hz, indicating a specialisation for the boatwhistle, are robust at 400 Hz, and do not completely tetanise At 500 Hz. The bladder produces more acoustic pressure per displacement at higher frequencies, indicating that peak amplitude at boatwhistle frequencies is caused by optimal muscle contraction and not bladder resonance. The swimbladder is an inefficient sound source, and we suggest rapid contraction evolved to produce louder sounds.

Juvenile sturgeons do not emit sounds under normal maintenance conditions. When fed toxic or semitoxic, juveniles of beluga Huso huso produced a series of wide band pulses. Few signals (0.1 signals/specimen/minute) were produced by fish during the second to the fifth semitoxic food feeding and during the first feeding with toxic food. The number of signals produced by fishes increased considerably up to 2-4 signals/specimen/minute during the sixth semitoxic food feeding and during the second toxic food feeding. Between toxic or semitoxic feeds, juveniles did not produce sound signals. During non-toxic food feeding, juveniles fed just with toxic or semitoxic food did not produce sound signals either. But when oxygen content decreased from 7.8 mg/l to 5.3 mg/l, juveniles fed with toxic or semitoxic food produced 1-3 signals/ specimen/ minute (a series of wide band pulses), while fishes fed with non-toxic food only did not produce any sound signals.

The acoustic activity of Russian sturgeon Accipenser guldenstadti and other sturgeon species increases considerably during the pre-spawning period. During induced maturation of Russian sturgeon, 25 records each of fifteen minutes long were made, and 1358 signals (100%) identified as sturgeon sounds were registered. The males (90%) and females (10%) produce the following signals: "Tone'' (4%), "Frequency Modulated or FM'' (50%), "Wide Band with Continuous Frequency Spectrum'' (30%), "Series of Wide Band Pulses'' (16%). The number of "FM'' signals produced by fishes correlates with the fishes' readiness to spawn and it increases after injection. The males reduced the maximum number of these signals 16-22 hours after the injection of sturgeon pituitary preparation. The females produced it 22-28 hours after the injection. The number of signals of the other types did not change noticeably. Most often Russian sturgeon produces FM signals having one component in which the frequency varies inversely as the square root of the time from 13.8-14.8 kHz to 6.8-6.9 kHz. Duration of these signals is as a rule 68-73 ms. While emitting one of such signals near to the transducer, the fish locomotor activity of both sexes increased and some fishes moved towards the transducer.

Ultrasound, in the form of a 200-ms pulse (duty cycle 1:1) consisting of a finite number of harmonic cycles, was presented to a gizzard shad school (n = 10) to determine whether or not it would elicit an avoidance response and significantly affect the quadrant distribution in a test chamber. Three frequencies were presented to the fish in four goo-second trials. Quadrant distribution was recorded 22 times during each trial. A chi-squared analysis (alpha = 0.05) was used to compare 8sh distribution in each quadrant to random distribution and distribution observed during control. Although no dramatic avoidance responses were observed, at 60 kHz (150 dB: re 1 microPa) both Control and Test fish distributions were significantly different to random distribution and from each other, at 80 kHz (162 dB: re 1 microPa) both Control and Test distributions were significantly different to random distribution but not to each other, at 120 kHz (152 dB: re 1 microPa) both Control and Test distributions were not significantly different to random distribution but were to each other, and at 120 kHz (160 dB: re 1 microPa) both Control and Test distributions were significantly different to each other with only the Test distribution different to random distribution. These results indicate that gizzard shad, like other Clupeids, may be able to detect some high-frequency sounds.

Fish hearing specialists (e.g. goldfish, holocentrids, moromyrids) are known to evolve specialised structures, e.g. Weberian ossicles; gas-filled bullae, swimbladder diverticulae to enhance their hearing abilities. The anabantoid fish have gas-filled suprabranchial chambers (SBC) situated in close proximity to inner ears. It is hypothesized that the gas bubbles inside the SBC can modulate hearing abilities of anabantoides. Three sound producing gouramis (blue gouramis Trichogaster trichopterus; kissing gouramis Helostoma temminckii; dwarf gouramis Colisa lalia) were used to test the hypothesis. The baseline audiograms (tested at 300, 500, 800, 1500, 2500, 4000 Hz) were obtained by the auditory brainstem response (ABR) protocol. The air bubbles inside the SBC were then driven out by water currents and the audiograms were measured again. Significant changes (i.e., thresholds) of audiograms were observed in all three species. The results support the hypothesis that air bubbles inside the SBC can affect hearing abilities of gouramis. However, the absence of air bubbles inside the SBC did not reduce hearing frequency ranges which indicated that swimbladder might also play a role in modulating hearing abilities of gouramis. The best hearing frequency of each species matched with the dominate frequency of the sound produced. This is an indication of co-evolution of sound production and hearing ability in anabantoids.

Within labyrinth fishes (Anabantoidei), croaking gouramis Trichopsis vittatus and pygmy gouramis T. pumila vocalize regularly during agonistic interactions. These vocalizations are broad-band pulsed sounds with high- pitched dominant frequencies (1-2.5 kHz), which contrast with the generally low-frequency hearing abilities of perciform fishes. Utilizing a recently developed auditory brainstem response recording technique, the auditory sensitivity was measured and compared to sound characteristics. The dominant frequencies of sounds were 1-2 kHz and 1.5-2.5 kHz for croaking and pygmy gouramis, respectively. Maximum low-frequency sensitivity for both species was 0.1-0.2 kHz. Maximum high-frequency sensitivity occurred between (3.8 and 2 kHz for croaking gouramis, which closely matched the best frequency of hearing. However, the best hearing frequency of pygmy gouramis was below 1.5 kHz, which indicated a mismatch between the two parameters. The correlation between sound production and perception in these fishes is likely to be facilitated by the suprabranchial airbreathing cavity, which lies close to the hearing and sonic organs and enhances both sound perception and emission at its resonant frequency. Additional examination of Macropodus opercularis, Trichogaster trichopterus and Colisa lalia revealed that all five gourami species could perceive sounds up to 5 kHz. The high-frequency hearing abilities of labyrinth fishes qualifies this group as hearing specialists.

Repeatability, that is the intra-class correlation coefficient (Sokal & Rohlf 1981), describes the degree to which variation within individuals contributes to the total variation in a group of individuals. In this study the repeatability for several parameters of the tonal and the drumming sounds emitted by breeding males of Padogobius martensii was measured to examine (1) which parameters of the sounds were more repeatable (i.e. stereotyped), (2) the contribution of temperature and male size to between-male variation in sound parameters, and (3) the potentiality for individual assessment/recognition by acoustical means in this species. Sounds were collected from single bouts of calling by 16 males differing in body size and ambient water temperature. All the sound parameters examined showed a remarkable inter-male variability. Water temperature explained most of between-male variation in the properties of the tonal sound, while body size was the main factor explaining the between-male variation in the properties of the drumming sound. After temperature control, however, the two types of sound showed a similar pattern of variation in the acoustic properties. Parameters of the tonal sound were less repeatable than their counterparts in the drumming sound. In general, high repeatability in a sound parameter was associated with a larger contribution of male size to the between-male variation in that property (after correction for temperature). Because of higher repeatability and stronger relationship with male size, the drumming sound appears to be adapted to convey information about the quality of the emitter (assessment signal).

Reactions of Japanese anchovies Engraulis japonicus to 100 to 700 Hz underwater pure tone signals were observed. The mean body length of the anchovies was 11 cm and mean weight was 10 g. Seven hundred fish were kept in a net enclosure which was 2 m in diameter and 1.5 m in depth. The anchovies ordinarily swam in a circle at almost same speed in the net enclosure. The startle response behaviour was defined as follows: more than a half of the fish changed their behaviour just after the sound projection, for example, the acceleration of the swimming speed, dispersion or concentration of the fish group. When no or few anchovies changed their ordinary behaviour, it was defined as no response. The up/down staircase method was used to determine the threshold level of the startle response. The sound pressure level ranged from 130 to 160 dB and was changed by 5 dB steps. The startle response levels were 154.5 dB at 100 Hz, 153.3 dB at 200 Hz, 146.8 dB at 300 Hz and 153.8 dB at 500 Hz. The 700 Hz signal did not affect the fish behaviour up to 158 dB.