founded on the Principles of Bouguer. 445 



be employed, in order to insure a very close approximation to 

 the truth. Take any number of observations, turning the instru- 

 ment round at each time, and the mean of these will give a re- 

 sult, perhaps as accurate as the nature of the case admits ; at 

 least, it will be sufficiently accurate for all ordinary purposes. 



When the colours of the flames are different, it is very diffi- 

 cult to ascertain the place of equal illumination. We can, how- 

 ever, as before, find the space over which the instrument moves, 

 before we discover an obvious difference between the illuminated 

 halves of the oiled or white paper. We must then take the 

 middle of this space, which will, even in that difficult case, give 

 us a very good approximation to the truth. The same method 

 was also used by M. BOUGUER, and found to be the best in simi- 

 lar cases. * But still this method is of very difficult application, 

 when one of the lights is of a fine white, and the other of a 

 dusky red or blue colour. In this case, I prefer the following 

 contrivance. 



Procure a piece of fine white paper, and get it printed with a 

 small distinct type. Paste it on the rectangular opening in the 

 instrument, which, in this case, may be somewhat enlarged. 

 Brush over the paper with fine transparent oil, or, if the paper 

 be very fine, this will be unnecessary. Place the instrument be- 

 tween the flames, and cause two assistants move them in either 

 direction, till you can just read continuously along the paper 

 with the same ease, and the squares of the distances will then 

 afford a good approximation to the truth. If the second form 

 of the instrument be used, the printed slip of paper must be 

 pasted on the faces of the mirrors, or smooth wood, and read 

 directly through the opening in the lid. 



* Traite d'Optique, page 50. 



