280 CARNEGIE INSTITUTION OF WASHINGTON. 



vations on the stars a Lyrae, /3 Pe-^asi, and a Tauri showed that the 

 method is capable of giving results with average errors of the order of 

 10 per ceat or less if the seeing is good. The character of the seeing, 

 however, affects the results materially. In the case of a Lyrae, at a 

 time when the seeing was fair, an extrapolation of the visibility curve 

 showed that this should cut the axis at a distance of 75 to 100 feet, 

 while on another occasion, with poor seeing, the intersection fell at 

 50 feet. This shows the importance of a careful study of the effects of 

 seeing on the vanishing point of the fringes. The following method is 

 suggested for making measurements of the seeing, which to a great 

 extent should be independent of the personal element. 



Let a diaphragm be placed at a convenient distance within the focus 

 and the aperture reduced until the diffraction rings appear. A pro- 

 visional formula connecting the seeing S with the aperture d is 



^=10 (l-e-"0 

 where a is a constant to be determined by observation. 



With the seeing measured in this v/ay, the vanishing-point of the 

 interference fringes, and hence the stellar diameter, can be corrected 

 by application of the factor ^ ^ 



l+alog — 



The constant a is found by observation, preferably, on a star of no 

 appreciable diameter. 



Another plan, which may prove even more serviceable is to observe 

 the point of disappearance of the fringes in the auxiliary system at 

 every measurement. In this case the factor may be represented con- 

 veniently by 



1—a log cos TT- 

 2 



in which r is the ratio of the areas of the apertures in the auxiliary 



system when the fringes vanish, and a a constant to be determined by 



observation. 



Interferometer Measures of Double Stars. 



The rotating type of stellar interferometer devised by Mr. Anderson 

 has been used by Mr. Merrill to measure the position angle and separa- 

 tion of two double stars, Capella and k Ursse Majoris. 



The orbital elements of Capella derived by Mr. Anderson give a fair 

 representation of Mr. Merrill's later observations on ten nights, thus 

 showing that only small corrections to the elements will be required. 

 The residuals in distance run up to 0''0035, but the definitive orbit, 

 which Vv^ill be computed when an accurate value of the period has 

 become available from radial velocity observations, should give devia- 

 tions for a night not exceeding 0''0005. Residuals of one degree may 

 be expected in the position angles. 



The duplicity of the known double k Ursae Majoris (A 1585) was 

 independently detected by Messrs. Anderson and Merrill on March 1, 



