4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 66 



far less than i per cent. Hence it seems probable that the influence 

 of changes in the inclination of the solar equator* as affecting our 

 investigations is inappreciable. 



Figure 2 (pi. i ) shows a number of drift curves at selected regions 

 of the spectrum chosen to indicate the differences of form which are 

 found, depending on wave lengths. As discovered by earlier investi- 

 gators, and hitherto abundantly confirmed by ourselves, the contrast of 

 brightness between the center and the edge of the sun is greatest for 

 short wave lengths, and diminishes as one comes to the red and 

 infra-red. 



We have two principal objects in this research : First, to repeat our 

 earlier determination of the distribution of light along the diameter 

 of the solar image ; second, to detect fluctuations of distributions 

 from year to year and from day to day, if any, and to compare them, 

 if found, with fluctuations of the radiation of the sun as determined 

 in our solar-constant investigations. 



On each day of observation we are accustomed to take 14 drift 

 curves, two each at seven different wave lengths, as follows : 0.3737/x, 

 0.4265/x, 0.5062/x, 0.5955/x, 0.6702/^, 0.8580/i, and 1.008/X. In the rediu:- 

 tion of observations, we proceed as stated on pages 154 and 155 of 

 Vol. 3 of our Annals, from which we quote as follows : 



" We have determined the rate of descent in the plate carrier of 

 the photographic plates on which the curves are recorded, and have 

 determined from the Nautical Almanac the time required for the 

 sun's disk to pass the meridian. From these data we have determined 

 the distance along the plate corresponding to the width of the sun's 

 diameter. This distance has been regarded as the true width of the 

 U-shaped curve, and all the measurements of ordinates of this curve 

 have been made at certain round-numbered fractions of the corre- 

 sponding solar radius. 



" To illustrate : On June 8, IQ08, the sidereal time required for the 

 sun's semidiameter to pass the meridian was i'" 8.72^ The corre- 

 sponding mean solar time for the passage of the diameter is 2.284'". 

 On this date the photographic plate descended 3.978 centimeters per 

 minute. Flence, the diameter of the sun expressed on the photo- 

 graphic plate is 9.086 centimeters. Measurements were made at the 

 center and at 10 places on either side of it, making 21 places in all, 

 at distances from the center of the U-shaped curve which correspond 

 to certain fractions of the solar radius from the center of the sun's 

 disk. 



" In further reduction of the observations, the mean values of the 

 measurements on each curve for the advancing and following limbs 



