Ohfei'vat'loHi on Coloufed Shadows. 103 



candle, and which is illuminated by no other light than the apparently M'hite light of the 

 heavens, blue .' I at firft thought that it might arife from the bluenefs of the Iky ; but find- 

 ing that the broad day-light refle£led from the roof of a neighbouring houfe covered with 

 the whitefl new-fallen fnow produced the fame blue colour, and, if poffible, of a ftiU more 

 beautiful tint, I was obliged to abandon that opinion. 



To afcertain with fome degree of precifion the real colour of the light emitted by a candle, 

 I placed a lighted wax-candle, well trimmed, in the open air, at mid-day, at a time when 

 the ground was deeply covered with new-fallen fnow, and the heavens were overfpread with 

 white clouds ; when the flame of the candle, far from being white, as it appears to be 

 when viewed by night, was evidently of a very decided yellow colour, not even approaching 

 to whitenefs. The flame of an Argand's lamp, expofed at the fame time in the open air, 

 appeared to be of the fame yellow hue. But the mod ftriking manner of fhewing the 

 yellow hue of the light emitted by lamps and candles, is by expofing them in the direft 

 rays of a bright meridian fun. In that fituation the flame of an Argand's lamp, burning 

 with its greatefl; brilliancy, appears in the form of a dead yellow femi-tranfparent fmoke. 

 How tranfcendently pure and inconceivably bright the rays of the fun are when compared 

 to the light of any of our artificial illuminators, may be gathered from the refult of this ex- 

 periment. 



It appearing to me very probable that the difference in the whitenefs of the two kinds 

 of light which were the fubjefts of the foregoing experiments, might fomehow or other be 

 the occafion of the difl^erent colours of the fhadows, I atteinpted to produce the fame 

 effe£ls by employing two artificial lights of different colours ; and in this I fucceeded 

 completely. 



In a room previoufly darkened, the light from two burning wax-candles being made to fall 

 upon the white paper, at a proper angle, in order to form two diftin£t fhadows of the 

 cylinder, thefe (hadows were found not to be in the leafl coloured ; but upon interpofing 

 a pane of yellow glafs, approaching to a faint orange colour, before one of the candles, one 

 of the (hadows immediately became yellow, and the other blue. When two Argand's 

 lamps were Tnade ufe of, inllead of the candles, the refult was the fame : the fliadows 

 were conftantly and very deeply coloured, the one yellow, approaching to orange, and the 

 other blue, approaching to green. I imagined that the greenifh caft of this blue colour 

 was owing either to the want of whitenefs of the one light, or to the orange hue of the 

 other, which it acquired from the glafs. 



When equal panes of the fame yellow glafs were intcrpofed before both the lights, the 

 white paper took an orange hue ; but the fliadows were to all appearance without the leafl 

 tinge of colour ; but two panes of the yellow glafs being afterwards interpofed before one 

 of the lights, while only one pane remained before the other, the colours of the fliadows 

 immediately returned. 



The refult of thefe experiments having confirmed my fufpicions that the colours of the 

 (hadows arofe from the different degrees of whitenefs of the two lights, I now endeavoured, 

 by bringing day-light to be of the fame yellow tinge with candle-light by the interpofition 

 of iheets of coloured glafs, to prevent the fliadows being coloured when day-light and 

 candle-light were together the fubjecls of the experiment; and in this I fucceeded. I was 

 even able to rcvcrfc the colours of the fhadows, by caufing the day-light to be of a deeper 



yellow 



