The nature and artificial production of consonant sounds

In a previous paper experiments were described in which the writer’s English vowel sounds were reproduced by means of double resonators tuned to give the two principal audible resonances heard when whispering or breathing the vowel in question. It was found that various consonant sounds could be produced by manipulating these models, as had been previously done by Kratzenstein, de Kempelen, Willis, and Wheatstone, in the case of a single resonator, but that all models did not behave similarly. Complete closure (by hand) and sudden release, under air pressure, of the ei (hay),a (calm),D (not), ou (no), and u (who) models, gaveb at low pressure andp at high pressure. Partial closure (allowing a small leakage sufficient to enable the artificial larynx to operate) followed by rapid release, gavem . On the other hand, the same manipulation of the models i (eat), I (it), e (men), A (up), gave a sound like that ofw in we, wit, wen, etc.