The cochlear geometry as a frequency analyser.

EVIDENCE concerning frequency analysis which was drawn together from several sources allowed the suggestion to be made that the design of the cochlea—i.e. its internal geometry—was itself a frequency analyser to a first approximation (Naftalin, 1965). Two sets of biological observations which seem to be of fundamental importance from the physical (acoustic) point of view may be cited as offering support for the hypothesis that the internal geometry of the cochlea performs a frequency analysis. First, the cochlea does not increase in size in proportion to the increasing size of animal from mouse to whale. Instead, the cochlea remains within less than one order of magnitude so that relative to body mass the bats and rodents have the largest and cetaceans the smallest cochlea (Hyrtl, 1845; Spector, 1956). Secondly, in the human subject the cochlea reaches adult size by the sixth month of foetal life (Bast and Anson, 1949): thereafter not only is there no further growth but the geometry does not alter. These biological facts suggest the probability that fixed size and unchanging geometry are important in cochlear function. A brief consideration of the microanatomy of the inner ear leads to the conclusion that such importance as the geometry may have in the cochlea will not be connected with transduction of acoustic to some other form of energy, but will be related to (peripheral) frequency analysis. An obvious advantage which can be immediately pointed out, of having a fixed geometry is that, if a preliminary frequency analysis were thereby achieved, this would have permanent characteristics and a nervous system once trained within this geometry would not have to relearn its frequency appreciation and interpretation with each new period of growth, i.e. frequency analysis would be the same from birth to old age. An examination of models and sections of the cochlea from the point of view of seeking structures and arrangements capable of performing acoustic analysis by virtue of geometry led to the conjecture that the shape of the tectorial membrane as a long thin wedge (Shambaugh, 1907; Naftalin, Spencer Harrison and Stephens, 1964) could be important in this connection, and that the continuous increase in cross-section and volume of the scalae, particularly of the scala tympani, from apex to base could also be meaningful. In order to investigate these problems it was necessary to devise a vibration-sensitive probe and to make models, including one no longer 619