new Standard of Light* 397 



^r = ^-7tt^ or total variation of the candle about 30 per 

 Mm. 1-21' v 



cent, from the average in the very short interval of time of 



about five minutes. This needs no further comment. Some 



additional experiments were made in order to ascertain the 



variation of the light of a standard candle. 



The P. L. S. (B)* with a current=5'9 webers was used as 

 unit. 



1st candle, 7 readings in 10 minutes. 



Mean = 1-08 P. L. S. (B) ; 



max. 1*19 . . , ... ^_ ' 

 — : — = -, 7 r F r . or total variation = 17 # b per cent, 

 mm. TOO 7 r 



The maximum was obtained directly after having opened 

 the photometer, when fresh air entered. 

 2nd candle, 10 readings in 14 minutes. 



Mean = l'07P.L.S.(B) ; 



— ^-= 7 ---r, or total variation = 59 per cent, 

 mm. 0'69 ; r 



The minimum was obtained directly after freshly lighting 

 the candle. 



3rd candle, 12 readings in 24 minutes. 



Mean = 1-07 P. L. S. (B) ; 



— r— '=777^-, or total variation = 46 per cent, 

 mm. 0*81 r 



The lowest reading was obtained shortly after lighting the 

 candle. 



Ath candle, 14 readings in 22 minutes. 



Mean = 0-94 P. L. S. (B) ; 



max. 1*26 . . , ... „ a . 



— : — = tt^, or total variation = (2 per cent, 

 mm. 0'58 r 



The lowest reading cannot be accounted for. 



Two new platinum light-standards, of the same form and 

 size as the P. L. S. (A) described before, were placed in cir- 

 cuit of eight Grove's cells connected up successively and with a 

 mercury rheostat in circuit, to keep the needle of the tangent 

 galvanometer at a constant deflexion. 



These two new P. L. S., called II. and III., were placed in 

 the photometer to compare their lights and by it test the 



* This platinum standard (B) was the first made and has a different 

 form from the other (A) described. Dimensions and weight cannot be 

 accurately given now. 



