276 SCIENCE PROGRESS. 



free from elastic fatigue. If this could be done the method 

 would probably be the best at our disposal. 



Pringsheim has recently introduced a modification 

 (B. A. Rep., 1894) by reading the temperature immediately 

 after opening the globe, instead of reading the pressure, 

 using the resistance of a very thin strip of platinum as his 

 temperature indicator. The modification seems to have 

 improved the method considerably. 



Kundt makes use of the fact that the square of the 

 velocity of sound in a gas is approximately y times the 

 pressure divided by the density. The velocity of sound in 

 the gas is deduced from its known value for air by causing 

 the same note of high pitch to traverse two tubes, one 

 filled with air, and the other with the gas under investiga- 

 tion, and observing the wave length of the sound in the 

 two cases by means of fine dust spread along the tubes. 



Here too a correction is required if the gas does not 

 obey the laws of Boyle and Charles, but though this cor- 

 rection may reach as much as 2 or 3 per cent., it has been 

 ignored by most investigators. 1 



The method is probably the most accurate we have at 

 present, as the conditions of accuracy have been very fully 

 determined by its inventor, but it is seriously handicapped 

 by the density of the gas appearing as a factor in the ex- 

 pression by which y is calculated. This makes it necessary 

 in most cases to determine the density experimentally, and 

 renders the method very sensitive to small amounts of 

 impurity in the gas. 



A third method has been used by Miiller ( Wied. Ann., 

 18, 1894) in a large number of observations, but as it assumes 

 that alternate compressions and rarefactions of half a second 

 period in a globe of glass holding only a litre are adiabatic, 

 it must of necessity give inaccurate results, and need not 

 be further described. 



These three methods all require the amount of deviation 

 of the gas from Boyle's law to be known before the calcula- 



1 The writer has given a formula and method of application that is 

 accurate, and makes no assumption as to the form of the characteristic 

 equation {Phil. Trans., 185, p. 1). 



