242 CARNEGIE INSTITUTION OF WASHINGTON. 



the 7-year mean until the observations at higher latitudes are fully 

 taken into account. WTien spots are plentiful and forming rapidly, 

 there is greater liability to local and superficial disturbances. The 

 significance of the higher value as indicative of a change in the period 

 of rotation is also lessened by the fact that the values for the years 

 immediately preceding and following are slightly below the average for 

 the 7 years, though in both years spots were numerous. 



The discrepancies between the earlier measures of Duner, Halm, 

 and Adams and the results from this long series are an outstanding 

 feature of the problem of solar rotation. The mean of their values 

 is 2.06 km. per second (1900-1908); the mean for Plaskett and 

 Schlesinger is 2.01 (1911-1913); the mean for the present Mount 

 Wilson series is 1.93 (1914-1920). Evidence that we are not observing 

 an actual change in the period of the sun's rotation may be stated as 

 follows : 



1. There is no progressive change in 7 years when observing con- 

 ditions remain uniform. 



2. The internal agreement for each observer is good. The results 

 for two series by Adams are 2.06 and 2.05 km. per second; for three 

 series by J. S. Plaskett they are 2.01, 2.02, and 2.01. For the current 

 7-year series the mean deviation per year is less than 1 per cent. 

 This raises the question of systematic effects depending upon the 

 observing conditions. 



3. A continuous retardation of the order indicated for the 20 

 years during which spectrographic observations have been applied 

 to the problem appears unacceptable from any point of view. 



The present program contemplates the continuance of the observa- 

 tions through a complete sun-spot cycle, the gathering of material for 

 a detailed study of rotation in the northern and southern hemispheres, 

 and the simultaneous use of duplicate instrumental equipment for 

 separating the influences of personal equation from those depending 

 upon the instruments. 



For the present, observations are in the main confined to the region 

 of X6300, in which atmospheric lines are available for eliminating in- 

 strumental disturbances, for avoiding faulty illumination of the grating 

 by the different prisms, and for studying the occurrence and influence 

 of local disturbances in the reversing layer. 



DISPLACEMENT OF SOLAR LINES AND RELATIVITY. 



Because of numerous fragmentary attacks upon' this question, the 

 situation is becoming more and more involved and unsatisfactory, as the 

 following brief summary by Mr. St. John shows: 



The Cyanogen Lines. 



From the line X4197, the unsymmetrical head of the second band, 

 P4rot (C. R., July 26, 1920, 229), after applying a correction for an 



