In the age before electronic devices for the production of sound, sirens were used as devices for creating and studying sounds of known properties (Helmholtz & Ellis, 1954).  The simplest version of a siren was composed of a thin disc through which a series of regularly spaced holes were punched.  That is, the holes were punched at regular intervals along a circle that was centered on the middle of the disc.  The disc could be rotated rapidly, with a pipe positioned so that air could be blown through one of the holes when it passed by the pipe.  With this mechanism, puffs of air passed through the disc at regular intervals that were determined by the spacing of the punched holes and the speed of rotation of the disc.  The result was that the siren produced a musical tone of set pitch, because it was generating a periodic wave of set frequency.  Helmholtz and Ellis describe elaborations of this design that were capable of producing strong, easily perceptible tones (e.g. their figures 2, 3, and 4 illustrate the siren of Cagniard de la Tour that was invented around 1819).  They note that sirens were used to empirically determine that the frequency of one tone was exactly half of the frequency of a tone that was an octave higher.  That is, one disc with 8 equally spaced holes will produce a tone that is an octave lower than a disc with 16 equally spaced holes when the two are rotated at the same speed.  Other fundamental relationships between musical tones and the frequency of periodic waves were established with sirens (see Helmholtz & Ellis, 1954, Chapter 1).