176 



ANNUAL REPORT SMITHSONIAN INSTITUTION, 1921. 



Apart from this encouraging result, the measurement of stellar 

 disks seemed more and more difficult to undertake. It is easily 

 calculated that an aperture of 2.50 meters (nearly 7 feet) is large 

 enough only for stars whose diameter exceeds 0".05. There now 

 exists at Mount Wilson a telescope of that aperture, the Hooker 

 telescope. Furnished with an apparatus similar to that of Stephan, 

 it showed fringes no matter what star was examined. Stars of 

 diameter greater than 0".05 therefore do not seem to exist. 



Must we therefore renounce the measurement of stellar diameters 

 or rather await the production of an objective of 5 to 10 meters aper- 

 ture? An artifice, conceived and carried out with incomparable 

 ability by Prof. A. A. Michelson, of the University of Chicago, 

 rescued astronomy from this perplexing dilemma. 



THE STELLAR INTERFEROMETER OF MICHELSON. 



The advantages and uniqueness of the interferential method lie in 

 that we may measure stars whose diameters are smaller than the 

 resolving power of the telescope. 



In fact the problem consists in causing to interfere two bundles 

 of rays of which the initial linear separation I) is considerable. 

 But it is not necesary to employ a telescope objective of diameter D. 

 This principle guided Michelson. Therefore the interferometer 

 which he constructed gives a most simple solution to the problem. 

 Four plane mirrors (fig. 3) gather two bundles of rays and reflect 



Fig. 3. — Plan of Michelson's interferometer. 



The arrangement proposed in 1891, and used at Mount Wilson in 1920, allowed the 

 interference of two pencils of rays, separated by several meters, while using an objective 

 of ordinary dimensions. 



them parallel but nearer together into the objective of a small tele- 

 scope. It is no longer necessary to have a giant telescope but in its 

 place plane mirrors and an ordinary objective carried by a rigid mount- 

 ing. The difficulties are now purely mechanical. The problem 

 becomes approachable through the resources of modern mechanical 



