C. GENERAL PHYSICS. 165 



hundred atmospheres to the square inch ; this tube was four- 

 teen inches long, and bent at a right angle. A very large 

 electric flash being sent through the tube, it was split by the 

 first discharge, and the pieces thrown to a distance of several 

 feet. The inner surface of the tube was, in fact, completely 

 pulverized, as though it had been struck by a hammer. Rey- 

 nolds estimates that the pressure must have been more than 

 one thousand atmospheres. 19 (7, 1874, 147. 



MEASUREMENT OF GALVANIC CURRENTS. 



Mr. Latimer Clark states, in reference to a common source 

 of error in the measurement of currents of short duration, 

 that, w r hen using galvanometers with shunts, a certain dis- 

 crepancy has been noticed, which is usually attributed to 

 some peculiarity of the vibration of the needle of the gal- 

 vanometer, and on endeavoring to ascertain the cause of this 

 discrepancy, he discovered that the results given by the use 

 of resistance coils were correct, and that those obtained by 

 the use of condensers were wrong. The cause of the error 

 he traces up to the fact that the movement of a magnetic 

 needle tends to induce in the coil within which it is suspend- 

 ed a current in the opposite direction to that producing its 

 deflection. A larger proportion of the whole current conse- 

 quently passes by the shunt, which is not subject to the in- 

 fluence of the needle. If, therefore, two galvanometers are 

 used, the movements of the needles within their coils being 

 similar, each of them counteracts the other's influence, thus 

 concealing the errors of the process. This source of error 

 does not exist when the measure is made by a differential 

 galvanometer. De Sauty's method of comparing measures 

 is also free from this source of error, and, in general, he con- 

 cludes that it is better, in testing submarine cables, to avoid 

 the use of condensers for comparing batteries with standard 

 cells, and to make use rather of very high resistance coils; 

 and he has generally recommended the use of a resistance as 

 high as two hundred and fifty thousand ohms. Journal Soc. 

 Tel. Engineers, II., 1C. 



THE CONDUCTIVITY OF METALLIC VAPORS FOR ELECTRICITY. 



The conductivity of flame for galvanic currents is known 

 to be greatly exalted by the presence of metallic vapors, and 



