116 C. Barus — Displacement Inteirferometer . 



and have found it trustworthy throughout. In the following 

 table the apparatus is adjusted with an angle of incidence of 

 I = 15°. The index of refraction is determined from the dis- 

 placement of ellipses when the air contained in a longitudinal 

 sealed tube in one of the component beams of light which 

 interfere is alternately tilled with air at pressure p and 

 exhausted, at a given temperature t. This tube was 23'8 cm 

 long and of brass, surrounded with a close fitting tubular tem- 

 perature bath. 



Table I. — Index of Refraction of Air from shift A N of ellipses. e=23*8 cm . 

 Angle of incidence 1=15°. 



Barometer Vacuum Shift Shift AiV /u t 



cm. cm. (scale pts.) cm. 



76-95 -5 13-27 -00663 1*000279 30°? 



19° 

 76-98 -4 13-48 -00674 1-000283 23° 



24° 

 77-15 -4 14-54 -00727 1-000306 2° 



21° 

 77-15 -4 13-55 -00677 1'000284 21° 



21° 



The first two columns of the table show the barometric height 

 and degree of exhaustion (residual pressure) ; the third the 

 micrometer displacement of the opaque mirror which brings 

 the centers of ellipses back to their original position with 

 reference to a given spectrum line. This reading is taken on 

 the drum on the micrometer, the scale parts being -0005 cm . 

 When electric arc light is used, the accentuated sodium line is 

 always in presence in the spectrum and makes an excellent 

 fiducial line for the centers of the ellipses. A single exhaus- 

 tion is sufficient for two readings, as the displacement occurs 

 on exhaustion (from red to green, for instance) and is measured 

 by the turn of the micrometer to bring the ellipses back ; while 

 on readmitting air the displacement is in the opposed direction 

 and is again measured by restoring the center of ellipses to the 

 position of the sodium line. If this displacement of the opaque 

 mirror on the micrometer is AiTcm. the index required is 



if e is the length of the tube. ATT must be given in cm. as 

 shown in the fourth column of the table and from this the //. of 

 the fifth is found at the temperature t in the last column. 

 They refer to the wave length of the D line, as this was taken 

 as the fiducial mark. The absolute values of //, are good but 

 here of relatively little interest, as no attempt was made to 

 standardize the screw, etc. of the measuring apparatus, and the 



