Carrying Electrical Current. 103 



ing the tripod (below G), by similar screws acting vertically. 

 The latter, moreover, were advantageous in bringing (raising 

 or lowering) the center of ellipses into the center of the field, 

 after an approximate adjustment had been made. Similarly the 

 foot of the revolving arm, which supports the telescope at D, 

 could be used for the same purpose with advantage. 



An open collimator at C, i. e., objective and slit each in 

 opaque screens, is specially convenient, inasmuch as it allows 

 the image of the slit S reflected at M to be visibly reproduced 

 on the jaws of the slit. The occurrence of parallel light and 

 of a beam of light normal to M are thus both put in evidence. 

 If m is cut off, the same applies to the mirror JST and beam n, 

 though this is liable to be too dark and the rough and fine 

 adjustments are best made at the telescope, using a wide slit 

 first. The collimator C should be long, in other words, the 

 lens of long focal distance, so that the beam m passing through 

 the tube may remain very short or spot-like in its vertical 

 dimensions. Otherwise too much is cut off by the tube and 

 reflection from G before the beam enters the telescope. 



With the arc lamp the spectrum may be darkened, so that 

 the sodium lines stand out clearly, by raising or lowering the 

 arc behind the black screen, with a hole about one centimeter in 

 diameter for the passage of light. A small swiveled screen in 

 front of the objective telescope, and approached from one side, 

 greatly sharpens the interferences and the Fraunhofer lines by 

 cutting off undesirable light and particularly the stationary 

 interferences. Possibly the latter would disappear if the plate 

 of the grating were made very thick. When water circulation 

 is used, some time must elapse before the temperature condi- 

 tions are adequately stationary. With distances m and n as 

 long as the above, the ellipses are rarely perfect, and frequently 

 appear coarse and distorted. It is nevertheless easy to adjust 

 the center of ellipses with an error not larger than *000,05 cen- 

 timeter, at the micrometer screw M, i. e., to about the mean 

 wave length of light. 



3. The Same. Results. — The sealed glass tube, t, of length 

 e —- 89*4:9 centimeters within, having been inserted, the data 

 for the refraction of air found by alternately exhausting and 

 filling it were about as given in Table I. 



Table I. 



Glass tube with plate-glass ends; diameter, 4"5 cm ; length, 89'49 cm . 

 Barometer, 75"59 at 17"5° ; temperature, about 20°. 



Po 



P 



10 5 AJV 



/"-I 



75-59 



•32 



2505 



280 







■10 



2495 



279 



