278 CARNEGIE INSTITUTION OF WASHINGTON. 



a Orionis and a Scorpii to be 5.75X10"^ cm., and for a Bootis to be 

 5.6X10"^ cm., the resulting angular diameters are: 



a Orionis 0':047, a Scorpii 0':040, a Bootis 0r022. 



A definite decrease in visibility, without complete disappearance of 

 the fringes, has been observed for a Ceti, a Tauri, /S Geminorum, 

 7 Draconis, and /? Pegasi. /3 Persei, /3 Orionis, y Orionis, a Leonis, 

 jS Leonis, a Lyrae, a Ophiuchi, a Aquilse, a Cygni, and a Pegasi have been 

 used as check stars, and in all cases show high visibilities with "^-he 

 maximum separation of the mirrors (19.6 feet). 



At first all mirror settings were made by hand and much time was 

 consumed in making the adjustments, but the subsequent installation 

 of motor-driven screws, which move the outer mirrors simultaneously, 

 has facilitated gi-eatly the operation of the instrument. Larger mir- 

 rors are now being made for the interferometer, which will extend 

 the range of observation to stars as faint as magnitude 3.5, and also 

 insure more accurate settings on stars brighter than the present limit 

 of 2.5 magnitudes. 



Four important points should be mentioned in connection with the 

 measurement of star diameters: 



1. Adjustment of the instrument. — If fringes are observed on a closely 

 neighboring check star, equality of path in the two pencils is certain; 

 but if the check star is too far away (30° or more), flexure of the instru- 

 ment may alter the adjustment in moving from one star to the other, 

 and the equality of paths is no longer certain. For isolated stars a 

 series of visibility measures is made on the star whose diameter is to be 

 dete.niined, beginning with the mirrors close together and increasing 

 their distance, step by step, until the fringes vanish, and then returning 

 a step or two until they reappear. The alterations in adjustment in 

 this case are relatively small and the method is often used. 



2. Seeing. — For final measures it is essential that the seeing be very 

 good, as poor seeing leads to a disappearance of the fringes for sep- 

 arations of the mirrors less that corresponding to the true diameter of 

 the star. On nights of bad and variable seeing, the visibility of the 

 fringes may vary as much as 50 per cent or more in the course of a 

 few minutes; and perfectly definite disappearances will be obtained 

 on one night at mirror separations quite different from those determined 

 on other nights, even when the decrease in visibility on each night is 

 uniform. 



3. Light. — Experiments with variable apertures show that for faint 

 stars the fringes vanish for mirror separations that are systematically 

 too small. 



4. Fringes. — If the fringe pattern is too fine the fringes disappear, 

 with increasing separation, sooner than when the fringes are wider. 



Mention should be made of the patient and helpful assistance given 

 by Mr. John Kimple in the troublesome adjustment of the mirrors. 



