Intelligence and Miscellaneous Articles. 11 



Marey, have of late frequently used. The disturbance communi- 

 cated itself to the air confined in a small caoutchouc tube, and thence 

 to a membrane of goldbeater's skin. A very delicate lever fastened 

 to this membrane indicated by its oscillations the slightest movements 

 of the liquid. 



The time was calculated by help of a tuning-fork, which inscribed 

 its vibrations on the blackened sheet of a registering cylinder. This 

 tuning-fork, which was verified several times, gave, for a tempera- 

 ture of 20°, 256 vibrations in a second. 



Before experimenting on the conduit filled with water, I performed a 

 series of experiments on the velocity of the propagation of sound in air, 

 in order to ascertain the degree of correctness this method admits 

 of. The apparatus, except some slight modifications, remained the 

 same. The sound was produced by a pistol charged with about 

 one gramme of powder. The shock communicated to the air of the 

 conduit was propagated through the whole length of the tubes, and 

 thenreturned, after reflection. At each successive departure and return 

 the small style of the membrane gave very distinct indications on the 

 registering cylinder. As the initial shock and the reflected shocks 

 were observed, the causes of error due to the inertia of the register 

 were eliminated. 



The greatest difficulty consisted in determining the temperature 

 of the air enclosed in the conduit. The tubes were laid in an open 

 trench ; and whilst their upper part, heated by the sun's rays, had 

 a temperature of 40°, the part in contact with the ground had one 

 of 20° only. 



Taking these two numbers as extreme limits of the temperature 

 of the air confined in the conduit, I found for the velocity of sound, 

 reduced to zero, 



V =326'60 metres (supposing the temperature 40°), 

 V =337'50 metres (supposing the temperature 20°). 



It is certain that the first number must be nearer the truth than 

 the second ; for the part of the tubes exposed to the sun was con- 

 siderably greater than that in contact with the ground. 



I now come to the experiments made on the velocity of sound in 

 the conduit filled with water. After first assuring myself that the in- 

 terior was absolutely void of air (which is easily done by examining 

 the joints), I fitted an hydraulic-press pump to the upper part of the 

 conduit. The shock in the water was caused by forcing-in the piston 

 suddenly. With whatever rapidity the lever of the pump was lowered, 

 no shock, properly so called, was produced, but a gradual compres- 

 sion ; thus the indication of the style upon the register, instead of 

 being a well-defined zigzag, as in the case of air, traced an elongated 

 curve, of which the point of coincidence with the spiral inscribed by 

 the style at rest was difficult to determine. However, four succes- 

 sive experiments gave a mean of 345 vibrations of the tuning-fork 

 between the initial and the return shocks. The length of the conduit 

 between the two plates which closed its extremities was 603*25 

 metres; hence the distance travelled by the compression, between 



