Viscosity of Hydrogen, Nitrogen^ and Oxygen, 587 



When the spot is reasonably stationary the recording stylus 

 is set in operation to register not only the speed o£ rotation 

 but also the form — that is, the variation or constancy — of 

 the speed. The temperature is again recorded. The spot 

 is kept at. the same point lor about fifteen minutes — for that 

 is the limiting duration of the recording device, — then the 

 temperature is again recorded, and then the clutch is dis- 

 engaged and then engaged again and again until the spot 

 is brought to its position when the outer cylinder is at 

 rest, and the zero position is recorded. The spot of light 

 not only measures the angle of deflexion but serves also 

 to indicate whether the speed of rotation of the outer 

 cylinder is constant or not ; for if the speed varies but 

 slightly the spot will begin to move. After some practice 

 it is possible to maintain the speed, by either accelerating 

 or retarding the motion of the driving apparatus, in such 

 a uniformity that the spot remains at the same point as 

 though the inner cylinder were caught rigidly in that 

 position for the greater part of the duration of an obser- 

 vation. After the zero reading is made the inner cylinder 

 is set to vibrate with its natural frequency, and the period 

 of vibration is recorded on the same sheet of recording- 

 paper. 



The Temperature. 



For the sake of uniformity, 23 degrees centigrade has 

 been chosen as the temperature at which all the obser- 

 vations are made. Furthermore, as it is still doubtful as 

 to what is the precise formula for temperature correction 

 for the various gases under consideration, all the obser- 

 vations were made at temperatures differing not more than 

 0'02 or 0'03 of one degree one way or the other from 

 23 o, 00 0. The temperature of the room was carefully 

 regulated by means of a very sensitive thermostat and 

 several electric heaters, so that the inside of the apparatus 

 very seldom indicated any variation of more than 0*01 

 or 0'02 of one degree during the whole period of one 

 observation : most of the time the temperature remained 

 the same from beginning to end. 



Computation of Results. 



In the equation for viscosity as previously derived we 

 have 



_7rl (a 2 ~b 2 )0 



n ~ ta 2 ir-t-w ' 



