Page 56 - Fall2020

DICK FAY AND GOLDFISH
psychophysical results for several common measures of hearing for all vertebrates covered in his book. By put- ting all of these data together in one graph, Dick showed that although there are clear differences across species in measures of the thresholds of hearing, there are also some very clear similarities. He also demonstrated that all terrestrial species have a spectral region where their thresholds are lowest; although these spectral regions vary greatly across species, the sound pressure level of a species’ best sensitivity is remarkably the same (i.e., within approximately ±15 dB of 0 dB sound pressure level [SPL] re 20 μPa). Given that 183 species are represented and the range of thresholds across frequency for many species is 100 dB or more, these similarities are striking.
Many of Dick’s contributions after his Databook was published were built on the theme of commonali- ties among the auditory function of vertebrates even when they had evolved different methods for process- ing sound. Two such related auditory functions and commonalities among species began to occupy Dick’s thoughts: sound source localization and its relationship to sound source processing (i.e., often called auditory scene analysis). In the remaining available space, we describe his work on sound source localization and refer to some of his articles on sound source processing (e.g., Fay, 2008, 2009a,b).
Sound Source Localization
One of the most important features of the vertebrate audi- tory system is the ability to determine the direction of a sound source around the animal. However, although it is clear that terrestrial vertebrates can localize, the ability of fishes to localize sound has been debated and studied for almost 100 years, without providing a full understand- ing of the capabilities and mechanisms (reviewed in Fay, 2005; Hawkins and Popper, 2018)
Dick became very intrigued by sound source local- ization, and he was convinced that fish must also be able to localize sources. However, data showing that fish did so and how that ability depended on the properties of sound were almost nonexistent. Dick’s Databook included only three papers that had exam- ined directional hearing (Chapman and Johnstone, 1974; Hawkins and Sand, 1977; Buwalda, 1981). Each of these papers dealt with the same species, the Atlantic cod (Gadus morhua).
In the articles cited by Dick, the investigators conditioned fish (e.g., via conditioned heart rate) to respond to a dif- ference in the angular separation of two sound sources and determined the minimum audible angle (MAA) that the fish could discriminate. The MAA is not a direct mea- sure of sound source localization, and Dick was not sure which cues fish, especially the relatively small goldfish, might use to localize a sound source.
As pointed out by van Bergeijk (1964), the traditional cues used to explain mammalian sound source localization seem to be unlikely for a fish. For mammals, sound arrives at the two ears with an interaural time difference (ITD) that pro- vides a cue for the azimuthal location of a sound source and/or the sound at the ear farthest from the source is less intense than that at the other ear, generating an interaural level difference (ILD) due primarily to the head producing a sound shadow for the sound arriving at the far ear. And sound is differently attenuated as a function of its frequency and the location of the sound source relative to the torso, head, and pinna. Thus, for mammals, spectral changes can provide a cue for the relative location of a sound source, especially in the vertical plane (elevation). The width of the goldfish head and the good impedance match between water and the fish’s head, as well as the approximately 4.8 times faster speed of sound in water than in air, reduce the ITD and ILD cues to negligible values. Fishes have no pinna, are sensitive only to low-frequency sounds whose wavelengths are very large relative to the size of most fish, and have very poor spectral resolution, so spectral changes are almost certainly not a cue for fish sound source localiza- tion. So, if fishes do localize sources based on sound, what do they use to do so? This is the sort of challenge Dick loved, and he became aware that European scientists had suggested that it was the detection of particle motion, a vector quantity, by fishes that allowed them to determine the direction of sounds (reviewed by Hawkins, 2014).
The Fay Shaker Table
The details of what Dick did can be found in an article that he published in Science (Fay, 1984). Dick argued that hair cell sensitivity to sound is directional, in that the axis of vibration of hair cell stereocilia is perpendicular to the direction of the inner ear vibratory motion. So, because fish hair cells are oriented in different directions within the macula of each inner ear organ, the sensitivity of the hair cells and the nerve fibers innervating the hair cells might also be directionally dependent.
56 Acoustics Today • Fall 2020