Wedge-a?id-JDiaphragm Photometer. 27 



it follows that the observed light contains proportionally 

 I more I oran g e than the paraffin-flame. 



Thus with the white-glass pane daylight in the laboratory 

 was balanced by a reading which the table indicated to be 

 equal to 55 candles at 1 metre. Observed through the orange 

 glass, however, the reading indicated the orange ingredients 

 in the daylight to be equal to 42 candles; that is to say, with 

 the paraffin-flame the light which passed through orange 



55 

 glass was -rx =1*3 times the corresponding ingredient of the 



same coloured light in daylight. The light of an incan- 

 descent lamp through white glass was 28 candles, while 

 through the orange glass it was 26 candles ; therefore the 



26 

 incandescent lamp had ^ = 094 time the corresponding 



colour in the paraffin-flame; and as the paraflin-flame had 1*3 

 as much as daylight, the incandescent lamp had 1*3 x 0*94 = 1*2 

 as much as daylight. In other words, if we had a similar quan- 

 tity of light obtained from daylight and from incandescent 

 lamps, we should have in the latter about 20 per cent, more 

 orange-tinted light. 



This colour-comparison is of course only very rough; but it 

 enables the excess of orange light in incandescent lamps, and 

 its occasional deficiency in some arc lamps, to be approxi- 

 mately compared. 



I have been led to employ in practice a selected portion of 

 a paraffin-flame in preference to a standard candle, partly in 

 consequence of the fact that the paraffin-flame can be enclosed 

 in a light-tight lantern, but chiefly on account of its superior 

 steadiness and constancy. As an instance, I have obtained a 

 balance between selected portions of two paraffin-flames, which 

 balance has been maintained for hours at a time without the 

 slightest variation. On the other hand, when a standard 

 candle was compared with the bright part of a paraffin-flame, 

 the photometer had to be readjusted continually, showing 

 several-per-cent. variations in the burning of the candle. I 

 therefore prefer to take a long series of readings once for all 

 in determining the value of the light admitted to the compa- 

 rison-field of the photometer by each diaphragm, to assume 

 means of these as constants of reduction to the nominal stand- 

 ard, and to depend upon the constancy and steadiness of the 

 lights given through the diaphragms by the paraffin -flame 

 afterwards. After working practically with this system for 

 some time, my opinion is that a selected area of the bright 



