By the rotation of a perforated pasteboard disk, it has been proved to you that a musical sound is produced by a quick succession of puffs. Had we any means of registering the number of revolutions accomplished by that disk in a minute, we should have in it a means of determining the number of puffs per minute due to a note of any determinate pitch. The disk, however, is but a cheap substitute for a far more perfect apparatus, which requires no whirling table, and which registers its own rotations with the most perfect accuracy. I will take the instrument asunder, so that you may see its various parts. A brass tube, _t_, Fig. 26, leads into a round box, C, closed at the top by a brass plate _a b_. This plate is perforated with four series of holes, placed along four concentric circles. The innermost series contains 8, the next 10, the next 12, and the outermost 16 orifices. When we blow into the tube _t_, the air escapes through the orifices, and the problem now before us is to convert these continuous currents into discontinuous puffs. This is accomplished by means of a brass disk _d e_, also perforated with 8, 10, 12, and 16 holes, at the same distances from the centre and with the same intervals between them as those in the top of the box C. Through the centre of the disk passes a steel axis, the two ends of which are smoothly bevelled off to points at _p_ and _p′_. My object now is to cause this perforated disk to rotate over the perforated top _a b_ of the box C. You will understand how this is done by observing how the instrument is put together. [Illustration: FIG. 26.] [Illustration: FIG. 27.]