EXPERIMENTS WITH THE DISPLACEMENT INTERFEROMETER. 73 



obtained adequately clear columns, though the color of the glass is frequently 

 a disagreeable feature. The three columns made had the following dimensions : 



In order to keep the plates together, the top and bottom of the long column 

 AB, fig. 42, were provided with close-fitting strips of metal mm and rn'm', 



4 



ra 



PIG. 42. 



tightened by the screws ss' on both sides. It was then incased in a wood and 

 metal sheath CD, carrying a long stem 5 at right angles to the length of the 

 column, by aid of which it could be clamped in any necessary position rela- 

 tively to the interferometer. 



Over half a year was allowed for the aging of these columns, at the end of 

 which time the balsam was hard (the column having been kept under stress), 

 quite clear, and free from air-bubbles. In fact, the narrow beam AB from the 

 collimator, passed through these columns twice, came to an adequately sharp 

 focus in the telescope, and no difficulty in finding or adjusting the interference 

 rings was experienced. 



In view of the large glass path the ellipses were necessarily quite small and 

 their motion in response to the micrometer screw sluggish. They were con- 

 tinually in motion, owing to the unavoidable tremors to which the laboratory 

 is subject, indicating of course that the evanescence of rings is still commen- 

 surable with the wave-length of light, whereas the sensitiveness of the shift 

 has been reduced in proportion to the length of column. Only in one respect 

 was the behavior peculiar. It was quite common to obtain only the right- 

 hand half or the left-hand half of the set of sharp concentric-ring patterns ; 

 i.e., on one side of the vertical line passing through the center of ellipses the 



