The hearing thresholds of two adult manatees were measured using a forced-choice two alternative paradigm and an up/down staircase psychometric method. This is the first behavioral audiogram measured for any Sirenian, as well as the first underwater infrasonic psychometric test with a marine mammal. Auditory thresholds were obtained from 0.4 to 46 kHz, and detection thresholds of possible vibrotactile origin were measured at 0.015-0.2 kHz. The U-shaped audiogram demonstrates an upper limit of functional hearing at 46 kHz with peak frequency sensitivity at 16 and 18 kHz (50 dB re: 1 microPa). The range of best hearing is 6-20 kHz (approximately 9 dB down from maximum sensitivity). Sensitivity falls 20 dB per octave below 0.8 kHz and approximately 40 dB per octave above 26 kHz. The audiogram demonstrates a wider range of hearing and greater sensitivity than was suggested from evoked potential and anatomical studies. High frequency sensitivity may be an adaptation to shallow water, where the propagation of low frequency sound is limited by physical boundary effects. Hearing abilities of manatees and other marine mammals may have also been shaped by ambient and thermal noise curves in the sea. Inadequate hearing sensitivity at low frequencies may be a contributing factor to the manatees' inability to effectively detect boat noise and avoid collisions with boats.

The resonant frequency of a bubble attached to a rigid infinite plane in an infinite incompressible liquid was determined theoretically taking into consideration the angle of contact of the bubble with the plane. The resonant sizes at 33 and 36 kHz were determined experimentally by plotting the relative acoustic pressure (required to cause bubbles to jump one bubble diameter horizontally) versus the bubble radius and were found to be 76 and 67 tz, respectively. The experimental resonant sizes were found to be in good agreement with the theoretical resonant sizes. A small minimum on the plot can possibly be explained by using a Mathieu equation.

Underwater noise radiating from dredging can effectively obscure or mask biological and other important sounds. This study recorded underwater acoustic characteristics of hopper dredging in the St. Johns River, Jacksonville, FL, to evaluate noise impacts in the waterway with respect to the endangered West Indian manatee. Of particular interest was the extent and range that dredging noise may mask the sounds of approaching commercial and recreational vessels. Vertical hydrophone arrays and a multi-channel PC-based recording system were used to measure dredging noise at various distances. Ambient noise surveys, active propagation of calibrated sources, and controlled boat noise measurements were conducted along the waterway. These data were integrated with behavioral hearing data to estimate zones of masking surrounding dredging. Three discernable noise sources that masked boat noise were (1) cavitation from dredge propellers, (2) draghead vacuuming, and (3) noise from submerged slurry pipelines. Sustained high ambient noise levels from dredging can significantly increase the risk of manatee-boat collisions by masking the sounds of approaching vessels over large radii (up to 2.5 miles). Mitigations suggested include ship quieting, reducing propeller cavitation, insulating or elevating slurry pipelines, and minimizing transects to pump out stations. [Work funded by the City of Jacksonville Waterways Commission.]

Wavenumber-frequency calibration of underwater, planar, receiving arrays requires the ability to generate single-wavenumber pressure fields over the surface of the array. When the wavenumber-frequency region of interest is evanescent, transmitting arrays previously constructed have been found to generate fields contaminated with harmonics, acoustic wavenumbers, and nonacoustic wavenumbers from the excitation of antisymmetric Lamb waves. An array that greatly reduces contamination has recently been constructed using a sheet of polyvinylidene fluoride (PVDF) with independent rectangular electrode stripes. The array operates in the frequency range of 500 Hz to 2 kHz and generates evanescent waves with phase speeds between 30 and 150 m/s. Contamination due to the excitation of antisymmetric Lamb waves is eliminated by shifting the phase speed of the Lamb wave out of the region of interest. This is accomplished by bonding the thin sheet of PVDF directly to a thick plate of LEXAN. Contamination from harmonics and acoustic wavenumbers is eliminated by driving the electrode stripes with suitably chosen shading coefficients generated by a simple-source numerical algorithm. Measured pressure fields and phase velocities compare favorably with numerical calculations.

No abstract