BLUE-VIOLKT LIGHT l\ THE SOLAR CORONA ON AUGUST 30, 1905. 323 



To determine by the Method of Least Squares only ( and 17 would mean the 

 discarding of the condition (A, + .r )/(/!. + a;) = (h,\ + x)/(h' x +x) and must lead to 

 erroneous results. 



I calculate , ?/, and A< from equations which result from logarithmic differentiation 

 of the equation given above. Let t n = C+ Af.(n = 1 to 6), t : = 100-88-<,-A< a +A< 7 

 (see 6), a = 1'05 + Aa, where t n (n = 1 to 5) are assumed to equal the observed 

 values and C is arbitrarily chosen equal to 1 TOO (see G). The values F M , are those 

 appearing in 5 (<), and they sufficiently approach their true values. I start from 

 x a = 140, y n = 4, the result of a first solution. The equations of condition are 



n = 

 where 



a= - Mod (1-F,.- 1 ')(/*,+ 



b = a log F mny c m = Mod (4a/ m )~ l , and similarly for suffix n, 



but 



e = Mod(4a)~ I (t t ~ l t 7 ~ l ) in equations m = 6, n = 7, 



d = (4a) -I log F w for m 5, n = 6 and 7/1 = 6, n = 7, 



The weight p of an equation is calculated with ? and r n as given in Table III., 

 O'Ol being the error of an equation of unit weight. The calculated weights served 

 merely as a guide. The adopted weights appear in Table IV. The numerical 

 equations are : 



Photographs. 

 I. II. 



I. III. 



I. IV. 



I. V.i. 



V6. Via. 



Vlft. VII. 



n is entered in Table IV. The brackets indicate logarithms, 10 being omitted. 



On account of the defects of Photograph VII. and the uncertainty of the exposure 

 of VI., I have solved the equations appertaining to Photographs I. to Va. separately 

 from those belonging to Photographs V6. to VII., and finally have discussed the 

 whole material. 



(a) Photographs I. to Va. These determine the intensity curve from distance 60 

 to 520. In accordance with the above, two of the A< are indeterminate. I choose At 4 



2 T 2 



n = -( 



n = -(9-239)(A 1 + 140)- I f + 0-0553i + (9-090)A/ 1 -(9-111) A/ 4 



n = -(9-367)(A, 



