RICHARDS AND MARK. — THERMAL EXPANSION OP HYDROGEN. 125 



gether at 0<> was 307.118— 1.944 = 305.174 milliliters. This volume 

 was sought at 32.38°, however, not at 0°, therefore the coefficient of 

 expansion of the glass must be determined and the appropriate correction 

 applied. It is easily calculated that the slight difference between the 

 expansion of the projecting platinum points and the glass is wholly neg- 

 ligible, causing a possible error in the volume of only about 0.00002 

 milliliter. Therefore the coefficient of expansion of glass alone need 

 be considered. 



For the sake of simplicity of manipulation the measurement of the 

 coefficient of expansion of glass was made upon the side tube which had 

 been cut off as already explained. The assumption here made that the 

 glass of the bulb and of the side tube was of the same quality seemed 

 admissible both because of the homogeneous appearance of the glass 

 itself at the line of welding and because of the consideration that such a 

 union could be made only between pieces of glass of approximately the 

 same coefficient of expansion. 



After a cross had been scratched with a diamond a few centimeters 

 from each end, the tube was mounted on a firm support the height of 

 which could be regulated by a finely threaded screw. Connections were 

 made so that a rapid stream of water from a large reservoir could circu- 

 late through the tube and a thermometer for reading the temperature of 

 this water was inserted into the middle of the tube. The tube as well as 

 the connection to the reservoir were wrapped in cotton-wool except for 

 the spaces about the crosses and the position opposite the thermometer 

 scale. 



The reservoir was now filled with ice and water, and after the tem- 

 perature of the stream of water had come to equilibrium, the lower cross 

 on the tube was adjusted by means of the fine screw of the support to 

 the same level as the cross-hair in the eyepiece of the cathetometer tel- 

 escope. The position of this cross-hair was kept fixed throughout the 

 experiment, and was made the permanent point from which changes in 

 height were measured. The position of coincidence of the crosshair of a 

 second telescope with the image of the upper cross on the tube was read 

 off on the dial of the micrometer which bore this cross-hair. When this 

 position had been determined by the average of several settings, the tube 

 was readjusted in respect to the lower cross and the height of the upper 

 cross was again read. This performance was repeated several times 

 until the tube until the exact distance between the upper cross and the 

 lower cross was well established. 



Warm water was then substituted for the ice-water and a similar set 



