With thinner calico or cambric it would require a greater number of layers to intercept the entire sound; hence with such cambric we should have echoes returned from a greater distance, and therefore of greater duration. Eight layers of the calico employed in these experiments, stretched on a wire frame and placed close together as a kind of pad, may be taken to represent a dense acoustic cloud. Such a pad, placed at the proper angle beyond N, cuts off the sound, which in its absence reaches _f′_, to such an extent that the flame _f′_, when not too sensitive, is thereby stilled, while _f_ is far more powerfully agitated than by the reflection from a single layer. With the source of sound close at hand, the echoes from such a pad would be of insensible duration. Thus close at hand do I suppose the acoustic clouds surrounding Villejuif to have been, a similar shortness of echo being the consequence. A further step is here taken in the illustration of the analogy between light and sound. Our pad acts chiefly by internal reflection. The sound from the reed is a composite one, made up of partial sounds differing in pitch. If these sounds be ejected from the pad in their pristine proportions, the pad is acoustically _white_; if they return with their proportions altered, the pad is acoustically _colored_. [Illustration: FIG. 4.] In these experiments my assistant, Mr. Cottrell, has rendered me material assistance.[93] NOTE, _June 3d_.—I annex here a sketch of an apparatus[94] devised by my assistant, Mr. Cottrell, and constructed by Tisley and Spiller, for the demonstration of the law of reflection of sound. It consists of two tubes (A F, B R) with a source of sound at the end R of one of them, and a sensitive flame at the end F of the other. The axes of the tube converge upon the mirror, M, and they are capable of being placed so as to inclose any required angle. The angles of incidence and reflection are read off on the graduated semicircle. The mirror M is also movable round a vertical axis. FOOTNOTES: [1] It will be borne in mind that the Washington Appendix was published nearly a year after my Report to the Trinity House. [2] That is to say, homogeneous air with an opposing wind is frequently more favorable to sound than non-homogeneous air with a favoring wind. We had the same experience at the South Foreland.—J. T. [3] Had this observation been published, it could only have given me pleasure to refer to it in my recent writings. It is a striking confirmation of my observations on the Mer de Glace in 1859. [4] Had I been aware of its existence I might have used the language of General Duane to express my views on the point here adverted to. See Chap. VII., pp. 340-341. [5] This does not seem more surprising than the passage of light, or radiant heat, through rock salt. [6] Also “Proceedings of the Royal Society,” vol. xxiii., p. 159, and “Proceedings of the Royal Institution,” vol. vii., p. 344. [7] See page 372 of this volume.