Page:Philosophical Review Volume 1.djvu/193

No. 2.] smaller harmoniums, to one of whose notes one of the prominent notes of the song was first approximately tuned, and which was set like the Harmonical within reach of the right hand. In the critical examination it was endeavored to make and keep the tuning of this note accurate; for such a study as has been made of these reproductions takes a much longer time than that during which the above test indicates the phonograph maintains its pitch accurately. Nevertheless, having carefully tuned a prominent note of the song to one of the harmonium notes at the beginning of the examination, in several cases where the reference back to this note was omitted, the two proved indistinguishable at the end perhaps of half an hour, and only in two or three cases was there doubt thrown on the work by a noticeable alteration of the phonograph pitch. It is still to be recommended that in such examinations the note tuned to the harmonium should be revisited every few minutes, to be sure that no aberration has taken place. The phonograph being accurately tuned, the pitch of each successive note of the song which exhibited that quality distinctly was then determined by cutting off the sound from the instrument the instant that note had been struck, and immediately comparing it with notes of the harmonium. In most cases, divergencies from any harmonium semitone came to light; the proportion of these divergencies from the nearest semitones above and below was then sought to be estimated as nearly as might be by repeated trials.

In order to gain some idea of the correctness with which the course of pitch constituting a melody can be determined in this way from a phonographic reproduction when its notes diverge in various amounts from the notes of the instrument of comparison, use was again made of the Harmonical. This instrument contains between the notes C of the four lower octaves the following complicated scale:

An inscription was taken of forty-eight Harmonical notes in which each of those between c' and c" occurred several times. This was then carefully written down according to the above method, using the harmonium as the instrument of comparison. The initial note of the sequence, which was the harmonical d', was tuned in the reproduction to d# of the harmonium. These two sounds being about 125 cents apart in absolute pitch, the notes of the reproduced melody were by this means thrown in among the harmonium semitones in such a way that it was impossible to predict where they ought to have come, and the mind was left entirely free to determine where they did come. The different estimates in the case of recurrences of the same note (however far apart in the melody) proved in general closely similar. In nine out of the thirteen different notes used in the sequence, the various judgments fell within a twentieth of a tone (or ten cents) of one another; in the other