10 



DISPLACEMENT INTERFEROMETRY BY 



where AN corresponds to A* in figure 4. Hence 



and equations (4) and (7) apply as before. This method also gives good 

 results even for short distances, r. 



3. Observations. The use of the apparatus, figure i, with the strip of 

 glass g to be tested sliding up or down, did not at first give satisfactory results, 

 because the mirror mm' was too thin (2 mm. thick). It was found however, 

 that on breaking contact at e during the sliding of g between successive posi- 

 tions, or by gently tapping the bar or standard G, very fair results were obtain- 

 able. There would have been no difficulty in using a thick glass mirror mm' 

 (0.25 inch or more), in which case the annoyance of flexure would have been 

 negligible. The following is an example of results obtained, the position of the 

 glass strip g being read off on a parallel vertical millimeter scale : 



Position 



io 3 AAf 



41.1 

 12.9 



37-5 

 10.6 



34-3 

 8.6 



31.6 

 7.2 



3-4 

 6.8 



32.7 

 7-9 



34-7 

 9.0 



38.7 



II. 2 



40.8 cm. 

 ii. 8 cm. 



These data are shown in figure 5 and the direct and return series are con- 

 sistent. The average slope of the strip, which is not quite uniform, may be 

 estimated at 



A7V=5XicT 4 cm. 



per centimeter of length, so that 



per centimeter of length. 



tfLv 



Using the ocular micrometer, it was found that one scale-part corresponded 

 to one-fortieth of AN = io~ 3 cm. Tests along a single centimeter of the glass 

 strip gave the results 



35-5 

 64.0 



34-5 cm. 

 45-3 



Position 34.55 



Ocular scale-parts 44.3 



i.e., a difference of 19.2 ocular scale parts per centimeter of length, so that 



A7V=i9.2(io~ 3 /4o) = 4.8Xio~ 4 cm. 

 a result virtually identical with the preceding, as no refinement was attempted. 



