As to the performance attained by the Telephone, let it be remarked, that, with its aid, I was in a position to make audible to the members of a numerous assembly (the Physical Society of Frankfort-on-the-Main) melodies which were sung (not very loudly) into the apparatus in another house (about three hundred feet distant) with closed doors. Other researches show that the sounding-rod [i.e. the knitting needle] is able to reproduce complete triad chords (“Dreiklänge”) of a piano on which the telephone [i.e. the transmitter] stands; and that, finally, it reproduces equally well the tones of other instruments--harmonica, clarionet, horn, organ-pipes, &c., always provided that the tones belong to a certain range between _F_ and _f''_[20]. It is, of course, understood that in all researches it was sufficiently ascertained that the direct _conduction_ of the sound did not come into play. This point may be controlled very simply by arranging at times a good shunt-circuit directly across the spiral [i.e. to cut the receiving instrument out of circuit by providing another path for the currents of electricity], whereby naturally the operation of the latter momentarily ceases. Until now it has not been possible to reproduce the tones of human speech with a distinctness to satisfy everybody. The consonants are for the most part tolerably distinctly reproduced, but the vowels not yet in an equal degree. Why this is so I will endeavour to explain. According to the researches of Willis, Helmholtz, and others, vowel sounds can be artificially produced by causing the vibrations of one body to reinforce those of another periodically, somewhat after the following scheme:-- [Illustration: Fig. 26.] An elastic spring is set in vibration by the thrust of the tooth of a cog-wheel: the first swing is the greatest, and each of the others is less than the preceding one (see Fig. 26). After several vibrations of this sort (without the spring coming to rest) let another thrust be given by the tooth; the next swing will again be a maximum one, and so on. The height or depth of the sound produced in this fashion depends upon the number of vibrations made in a given time; but the quality of the note depends upon the number of variations of amplitude (Anschwellungen) occurring in the same time. Two vowels of equal pitch may be distinguished from each other somewhat after the manner represented by the curves (1) (2): while the same tone devoid of any vowel quality, is represented by curve (3). [Illustration: Fig. 27.] Our organs of speech create the vowels probably in the same manner by a combined action of the upper and lower vocal chords, or of the latter and of the cavity of the mouth.