PHYSICS. 593 



tion. The reed speaks or is silent for certain definite variations in the 

 length of the draw-tube; the sum of the two consecutive intervals cor- 

 responding to sound and silence being equal to the wave length of the 

 sound given by the reed. Any change in the velocity of the air current 

 modifies both the intervals, but their sum remains absolutely the same. 

 With a draw-tube which may be extended through 3 meters, a series of 

 closely concordant measurements of the lengthening and shortening 

 may be rapidly made, whose mean value is closely approximate. Since 

 Biot has shown that the proper sound of a reed is independent of the 

 nature of the gas acting upon it, it suffices simply to sound this reed 

 with different gases in order to compare directly the velocities of sound 

 in them, which are proportional to their wave lengths. The relative 

 values obtained were, for air 27'5 to 32 ; carbon dioxide, 22-5 to 25 ; illumi- 

 nating gas, 42, and hydrogen 113. The mean of nine closely concordant 

 experiments gave 40-48 for the wave length in steam at 100°, the calcu- 

 lated value being 41-089. (J. Phys., December, 1885, II, iv, 550.) 



Mercadier has continued his studies on the laws of vibration of circu- 

 lar plates. He concludes (1) that it is not possible in practice to con- 

 struct disks of iron or steel of a thickness less than one millimeter which 

 shall give sounds calculated in advance from their dimensions; and 

 (2) that the mathematical theory of circular vibrating plates is not af- 

 fected by this, since the differences between the calculated and the exper- 

 imental results are explained by circumstances difficult to introduce 

 into formulas. In all cases the more the thickness increases the closer 

 is the accord between theory and experiment. Hence, in verifying the 

 theory anew, it will be necessary to experiment on disks of small radius 

 and of sufficient thickness. {J. Phys., December, 1885, II, iv, 541.) 



Semmola has observed that if the conductors of an influence machine 

 are connected by means of two wires 5 meters long with two binding 

 screws on opposite sides of a brass plate 1 millimeter in thickness, 

 resting on an ebonite funnel, and the path of the current is broken so 

 that sparks strike across, the plate begins to sound. A Geissler tube 

 or a lead wire may be interposed in the break without destroying the 

 effect, or a wire containing the break may be led to earth from one of 

 the binding screws. The sound is louder if the upper end of this wire 

 is held at a short distance from the connecting post, so that sparks pass. 

 The wires from the machine may even be connected with a second metal 

 plate parallel to and at a distance from the other one, the sound being- 

 strengthened by connecting this plate to earth, and slightly by interpos- 

 ing a glass plate. If the wires be connected with the insulated wire of a 

 sonometer, instead of the brass plate, no sound is heard unless an ebon- 

 ite ear-trumpet is placed on the resonant box. {Beibliitter Phys., IX, 

 671; Phil Mag,, December, 1885, V, xx, 548.) 



Le Conte Stevens has suggested an improvement in the method of pro- 

 jecting acoustic curves optically. He uses the well-known method of 

 Lissajous, but instead of throwing the beam reflected from the mirro)' 



ii. Jili). i^ 66 



