558 HENRY A. KOWLAND 



tions made at the time of measuring the coincidences; Plates 9, 10, etc., 

 refer to photographic plates; C, R, etc., refer to the series as given in 

 the final table, although they may differ very slightly from the latter, 

 as the final table contains slight corrections. Figures in parentheses 

 are the number of readings. The photographs were usually measured 

 from two to six times. 



Table III gives the first series of observations made in 1884 with 

 a 21 foot concave, 14,436 lines to the inch. The numbers taken for 

 the standards are only preliminary, and agree as nearly as practicable 

 with my Table of Preliminary Standards. As only differences are 

 finally used, they are sufficiently near. The fractions give the order 

 of the spectra observed. 



Thus, the first observation on Ji and t is worked up as follows : 



4691-590 7027-778 



Correction to standard. . . . 626 +2-785 



4690-964 7030-563 



4691-590 4690-326 7027-778 



626 +-626 +2-785 



4690-964 4690-952 7030-563 



4690-962 7030-556 



The equation 3 Ji 2 t = 11-774 then readily follows. 



Tables IV and V are from a 21 foot concave with 7218 lines to the 

 inch, used on both sides, and thus equivalent to two gratings used on 

 one side only. I have not yet determined theoretically whether the 

 minor errors are perfectly neutralized in this manner, but it would evi- 

 dently have a tendency in this direction. 



The photographic coincidences are given in the main table (X), as 

 not only the standards are compared by this process, but whole regions 



