Brightness of the Corona of January 22, 1898. 39 



ratio of 2 01 = 1/07. (The logarithms of distance have been taken to 

 base 10 in the ordinary way.) These differences between the plates may 

 be due to any combination of the following causes : — 



(a.) Accidental error in exposure to corona. The exposures were 

 made without any mechanism, and the short ones especially may be 

 sensibly in error. Thus the difference between the 1 sec. and 20 sees, 

 exposure is 0'8. If the whole of this be due to accidental error in the 

 1 sec. exposure, it would mean that the exposure was for 1 sec. x 2 -0 ' 8 

 = 0"58sec. instead of for 1*0 sec, which is not an extravagant suppo- 

 sition. 



(b.) Accidental error in exposure to squares. This should be much 

 smaller than (a.). 



(c.) Difference in sensitiveness of the film near the edge of the plate 

 where the squares are impressed, and in the centre where the corona is 

 impresssed. There is independent evidence of sensible differences of 

 this kind, and the point is under investigation. 



(d.) Differences in the behaviour of the candle which impressed the 

 squares on the various plates. 



(e.) Climatic differences between Sahdol and Pulgaon. 



11. It becomes necessary to decide which plate to take as the 

 standard. Cause (a.) ought not to affect the 8 sees, and 20 sees, appre- 

 ciably, but cause (e.) may. They differ by 0*5, and we may perhaps 

 take the mean. The corrections to be applied to the plates are then 



Plate I II III IV 



Exposure 1 sec. 2 sec. 8 sec. 20 sec. 



Place Sahdol Sahdol Pulgaon Sahdol 



Correction ... +0*6 -0-2 +0'3 -0-2 



If any other selection is preferred, it is easily applicable as a con- 

 stant to the final numbers. 



. 12. The correction for constant illumination of the plate due to sky- 

 glare has been adopted as 2 -6 ' 4 moon, taking the moon as equal to 0'02 

 of a candle at 1 foot. If at any point the corona has a brightness 

 represented by x, meaning 2 X x moon, then the brightness measured on 

 the plate will appear as y where 



+ 2 -0 ' 4 = %y. 



A table was formed giving y in terms of x, of which the following is 

 a portion : — 



