G. Barus — Displacement Interferometer. 117 



water circulation had not been installed. The temperature 

 coefficient may be found without this. If, therefore, /ul — 1 

 is expressed in terms of t the result is a = -0036, or of the 

 order expected. 



From this datum the working conditions of the apparatus 

 may be specified. For a tube 23*8 cm long the micrometer 

 displacement per degree G. is *051 scale parts or # 0005 cm each ; 

 i. e., the micrometer displacement is about *000,025 cm per 

 degree C., or about 10~ acm per degree G. per cm. of length of 

 tube. Thus a minimum of about 2° G. is directly appreciable 

 in the given case, or for a tube about half a meter long a mini- 

 mum of 1° G. should be appreciable at all temperatures. This 

 moreover would correspond to the evanescence of two rings in 

 succession, whereas in the above apparatus a little less than one 

 ring vanishes per degree G., at all temperatures. The displace- 

 ment of ellipses for an atmosphere of pressure is roughly from 

 the D, nearly to the inline. 



In this method the arms need not be of invariable length 

 except during the short period of exhaustion, as the data are 

 obtained by differences. 



To turn to the second method for obtaining the same result : 

 the displacement of ellipses is accompanied by the radial 

 motion of rings to and from the center and the number vanish- 

 ing may be counted. If X is the mean wave length between 

 the initial and final position of the rings and n is the number 

 of rings vanishing, then the equivalent micrometer displace- 

 ment would be 



AJST= nX/2 



so that the micrometer reading A i^need not be taken. Hence 



fi=l+n\/2e 



To make use of the method, a fine screw stopcock is to be 

 inserted through which dry air may be admitted, at any rate, 

 into the exhausted tube. In this way the motion of the rings 

 toward the center may be controlled, perfectly, and their evan- 

 escence specified. The experiment is very interesting. Clearly 

 the arms must be kept at invariable length while the rings are 

 being counted, i. e., until they cease to move, when the pres- 

 sure is against normal. The following figure contains an 

 example of many results of this kind. Here the abscissas 

 denote the number of rings which have vanished and the 

 ordinates the corresponding pressure, the latter increasing from 

 a few mm. to an atmosphere. The line of observations hap- 

 pens to be nearly continuous. An interruption of the count is, 

 however, of no serious consequence, as the slope of the line is 

 alone in question when the initial pressure and final pressure 



