(3.) It is true that in practice the edges of the telephone diaphragm are in nowise fixed, but merely set into a groove, or rather clamped between wooden or metallic rings, whose mass is comparable to their own; and they are, therefore, as regards elasticity, in an ill ascertained state.  Yet a diaphragm of the usual diameter (from 2 to 4 inches), and very thin (from 0.001 to 0.02 inch), clamped in this way by its edges, is capable of vibrating when a continuous series of sounds are produced near it, by means, for example, of a series of organ pipes.  But the series of sounds that it clearly re-enforces, in exhibiting a kind of complex nodal lines, is plainly discontinuous; and how, therefore, would the existence of such series suffice to explain the production of a continuous scale of isolated or superposed sounds, the chief property of the telephone?

(4.) The interposition of a plate of any substance whatever between the diaphragm and the source of the vibratory motions in nowise alters the telephonic qualities of the diaphragm, and consequently the nature of the motions that it effects—­a fact that would be very astonishing if the motions were those that corresponded to the peculiar sounds of the diaphragm.  This fact is already known, and I have verified it with mica, glass, zinc, copper, cork, wood, paper, cotton, a feather, soft wax, sand, and water, even in taking thicknesses of from 5 to 8 inches of these substances.

(5.) We can put a diaphragm manifestly out of condition to effect its peculiar scale of harmonics by placing small, unequal, and irregularly distributed bodies upon its surface, by cutting it out in the form of a wheel, and by punching a sufficient number of holes in it to reduce it half in bulk.  None of these modifications removes its telephonic qualities.

(6.) We can go still further, and employ diaphragms of scarcely any stiffness and elasticity without altering their essential telephonic properties, the reproduction of a continuous series of sounds, accords, and timbres.  Such is the case with a sheet iron diaphragm.  It is very difficult, then, to imagine a fundamental sound and its harmonics.

The conclusion from all this appears to me to be that the mechanism by virtue of which telephone diaphragms perform their motions is at least analogous to, if not identical with, that through which solid bodies of any form whatever (a wall, for example) transmit to all of their surfaces all the simple or complex successive or simultaneous vibratory motions, of periods varying in a continuous or discontinuous manner, that are produced in the air in contact with the other surface.  In a word, we have here a phenomenon of resonance.  In diaphragms of sufficient thickness this kind of motion would exist alone.  In thin diaphragms the motions that correspond to their special sounds might become superposed upon the preceding, and this would be prejudicial rather than useful, since, in such a case, if there resulted a re-enforcement of the effects produced, it would be at the expense of the reproduction of the timbre, the harmonics of the diaphragm being capable of coinciding only through the merest accident with those of the sounds that were setting in play the fundamental sound of the diaphragm.  This is what experiment clearly demonstrates.