1327 



01 23456789 10 



ir invisible 14,4 14,15 14,0 13,8 13,6 13,4 13,2 



,v 14,2 14,05 13,9 13,85 13,45 13,1513,0 12,8512,7 12,4 12,25 

 r 13,2513,0 12,8 12,5 12,25 11,9 11,75 11,75 11,5 11.3511,15 



These values were made use of iii order lo deduce the inagiiilude 

 of the star-images in the scjiiares on plate /i and .1 : the shorter 

 exposures give magnitudes decreasing by about 1'", from which 

 the dilTerence in magnitude of the successive exposures B,A^,A^,A^, 

 may be deduced. 



Classifying these differences according to magnitude, we find: 



B A^ /i-.4, corrected .4, .4, A^A, corrected 



11.40 12,32 0.92(18) 0.98 11,46 12,40 0,94(5) 1,01 

 11,88 12,76 0,88(17) 0,96 11,85 12,81 0,96(11) 1,04 



12.41 13,35 0,94(30) 0,91 12,30 13,49 1.19(17) 1,15 

 12,70 13,84 1,14(18) 1,00 12,76 13,95 1,19(17) 1,03 



r,Ö7(50) 



1,04(19) 



The differences are not merely accidental ; the fact that with all 

 of them the last value is the greatest, proves that the scale is not 

 yet wholly homogeneous. By successive approximations the following 

 deviations fiom an evenly running scale were found - 



11,42-12,32 —0,06 12,36—13,35 +0,03 



11,87-12,73 —0,10 12,75—13,86 +0,15 



These are accounted for by the following corrections to the scale: 



11,2 



By introducing these corrections, we get for the difference in 

 magnitude /i, - .4=0,95; .4, — .4, = 1,07 ; .4, — .4, = 1,04. For 

 the shortei' exposures only the brighter stars could be used; they 

 gave the result of .4, — .4, = 1,16(7) ; .4, — ^^ = 1,09(14). The 

 mean error of 1 determination of magnitude is 0"',14. 



86* 



