The shape of the vocal cord wave is an important parameter in a true analog representation of the vocal mechanism, since the final speech wave is a function of both the generator wave shape and the transfer impedance of the vocal tract. The shape of this wave has been determined by two independent methods: (1) by the derivation and use of a network having a characteristic which is the inverse of the first vocal resonance, and (2) by measuring the area of the vocal cord opening as a function of time through the use of motion picture studies. Harmonic analysis of some of the typical shapes obtained indicate that a uniform distribution of harmonic amplitudes is a rarity; instead there is a tendency for the distribution to have a cyclical variation of the form (sinkn/kn 2). The results obtained through the use of the inverse network indicate that the main excitation of the higher resonances occurs at the point of vocal cord closure and that the magnitude of this excitation can be controlled by the talker over wide ranges.