36 DISPLACEMENT INTERFEROMETRY. 



in the slit of the collimator) correspond to each other. In the given apparatus 

 r = 7.2 cm.; 6 = 9.7 cm.; 2 = 45; so that the constant is 



^^=0.0131 or a = 0.013 1 (AAV A/) 



9-7X40 



In the outgoing series A7V/A/ = 0.00125; or <x= io-"Xi.64; in the ingoing 

 or contracting series A-/V/A/ = 0.00130; or a= lo-^Xi.yo. 



These values are too small, doubtless for reasons connected with the dis- 

 tribution of temperature along the length of the expanding tube. That it 

 was not sufficiently constant was tested by sliding the thermometer along the 

 tube. Virtually a rod 38 cm. long, instead of 40 cm., has been exposed to the 

 given temperature conditions. The two rates are to be similarly explained. 

 The remedy would be a longer rod, or adequate jacketing. It would be well 

 worth while to do this, since the curve figure 39 shows the method itself 

 to be trustworthy. In the present place, however, these experiments are 

 incidental and merely of interest in their bearing on the magnetic work, so 

 that no further improvement was attempted. 



