On the other hand, we see that if the telephones with several membranes were not successful in England, this was not the case in America, for the telephones which experience has shown to give the best results in that country are those of Mr. Elisha Gray and Mr. Phelps, and these have several diaphragms. It is evident that there are details of construction in these instruments which may appear insignificant in theory, and which are notwithstanding very important from a practical point of view, and we believe that it is to this circumstance that instruments of this kind owe their success or failure. Thus, for example, it seems that the vibrations of air caused in the mouthpiece ought to be immediately directed on the surface of the diaphragms by means of distinct channels; it is necessary that the empty space round each diaphragm should be sufficiently limited to prevent echoes and interruptions, unless the case is so large that there is no danger of such effects. Above all, it is necessary that the organs should be fixed in some material unsusceptible of reverberation, and for this reason a preference is given to iron or ebonite. It is certain that when the instrument is properly made, its effects are superior to those of the Bell telephones, and it is asserted in the ‘Telegraphic Journal’ that experiments were made with one of these instruments before the Royal Society, in London, May 1, 1878, and that the intensity of sound was in proportion to the number of diaphragms. This instrument was designed by Mr. Cox Walker, of York, and possessed eight diaphragms. He considers that this is the arrangement which gives the best results. [Illustration: FIG. 35.] _Mr. Elisha Gray’s System._--Mr. Elisha Gray’s last system, which we represent in fig. 35, is one of those which have given the best effects. It is made, as we see, of two telephones, side by side, to which correspond two tubes, issuing from a common mouthpiece E. One of these telephones is seen in section in the plate, the other in elevation, and they correspond to the two branches of a nickel-plated horseshoe magnet N U S, which may serve as a suspension ring. In that part of the plate which represents the section, the induction coil is shown in B, and the magnetic core, of soft iron, in A, which is screwed to the polar end of the magnet S; the vibrating plate is in L L, and, as we see, the tube of the mouthpiece terminates on its surface. In another model there are four telephones side by side, instead of two, and the effects produced are still more marked.