[ 338 ] 



XXXVIII. On the Application of Interference -Methods to Spec- 

 troscopic Measurements. — I. By Albert A. Michelson *. 



[Plate VII.] 



IN a paper on " Measurement by Light-Waves " f it was 

 shown that by substituting a refractometer for a micro- 

 scope or telescope or spectroscope, the accuracy of measure- 

 ment may be increased to a remarkable degree, but at the 

 cost of resolution and definition. When the object viewed is 

 smaller than the limit of resolution, however, then, even in 

 the ordinary sense, these instruments possess scarcely any 

 advantage ; but if by " defining-power " is meant not the 

 actual resemblance of the image to the object, but the ac- 

 curacy with which the form or the distribution of light may 

 be inferred, then the refractometer has remarkable advantages 

 in both defining-power and accuracy. 



This has been pointed out in the preceding paper (supra, 

 p. 256) in the case of telescopic observations. It was shown 

 that by a proper interpretation of the " visibility-curve " the 

 form and the distribution of light might be correctly inferred, 

 even when the object is beyond the resolving-power of the 

 telescope. 



In the present paper it is proposed to show that the same 

 principle holds in spectroscopic observations, and even with 

 greater force. 



In fact, a spectral line might be viewed by a telescope con- 

 verted into a refractometer, and the " line " could be studied 

 in precisely the same w 7 ay as a nebulous star. Such a method, 

 however, would multiply the necessary imperfections of lenses 

 and prisms or gratings. 



Fortunately both spectroscope and telescope may be dis- 

 pensed with, by substituting the refractometer in any form 

 which allows a considerable and steady alteration in the dif- 

 ference in path between the two interfering streams of light 

 from any approximately homogeneous source. 



One of the first applications of this method J showed that 

 the red hydrogen-line is a very close double ; and the same is 

 also true for the green thallium-line. Another instance of 

 the class of problems which the method may help to solve is 

 the effect of the velocity of the vibrating molecule in the line 

 of sight in broadening the spectral lines. 



* Communicated by the Author. 



t American Journal of Science, vol. xxxix., Feb. 1890. 



X Phil. Mag. Dec. 1887. 



