MOUNT WILSON OBSERVATORY. 251 



a displacement, which has not yet been looked for, would be of the 

 order of the distance between the fringes — a very small quantity. 



In view of these encouraging results, it was decided to attempt to 

 observe interference with an interferometer consisting of two plane 

 mirrors, separated by a distance of 20 feet, which reflect the hght from 

 the star to two other planes (about 4 feet apart), whence the two pencils 

 proceed to the 100-inch surface, and thence to the convex Cassegrain 

 reflector and the eyepiece, in the focus of which the two pencils unite, 

 producing, under proper conditions, the interference fringes to be 

 observed. 



A preliminary trial showed that the adjustments provided for the 

 mirrors were much too coarse, and these were accordingly remodeled. 

 Fm'ther, instead of attempting to produce equality in path (with an 

 ordinary accuracy of a ten-thousandth of an inch, which is necessary 

 if the interference is to be observed in white light) , an adjustable double 

 wedge of glass made it possible to vary the optical path gradually and 

 continuously, until the required condition was fulfilled. A plane par- 

 allel plate of thickness equal to that of the double wedge is placed in the 

 path of the other pencil, and by appropriate inclination proves valuable 

 in securing the coincidence of the two images. 



A trial of this an-angement was made on Vega on August 6 by Mr. 

 Pease, and interference fringes were observed with a ba^e of 18 feet, 

 seeing 4 on a scale of 10, magnification 1,600. Fringes were also found 

 in white light. The observations were confirmed by Mr. Sherburne. 



On August 8, with the outer mirrors about 7 feet apart, and with the 

 aid of a direct-vision prism in front of the eyepiece, Mr. Pease found 

 fringes, which were then observed by Professor Michelson. Fringes in 

 white Hght were also found. Observing Vega again on August 9, fringes 

 were seen in white light by Messrs. Hale, Michelson, and Pease (who 

 found them) with a base of 12 feet, but, owing to an error in adjust- 

 ment, they were not observed at a greater distance. 



The appearance of the interference fringes was about as good with 

 the maximum base of 18 feet as with one of 7 feet, which seems to indi- 

 cate that it should be quite feasible to obtain interference with a base 

 of 100 feet or more. If any of the nearer stars, say with a parallax of 

 0''2, are as large as the sun, their disks would subtend an angle of 

 about 0^002, which would be manifested by a disappearance of the 

 interference fringes if the base were 200 feet long. Recent calculations 

 by Professor Eddington and by Mr. Shapley, based upon provisional 

 values of the surface brightness of a star, indicate that some of the 

 late-type giants may have diameters of the order of 20 times the value 

 specified, which would render the conditions much more favorable. 

 In addition to the determination of stellar diameters, an interferometer 

 with a long base could also be used for measurements of double stars 

 whose separations are very small — of the order of 0'' 001, for a base of 

 200 feet. 



