by Diffusion of Light. 51 



plane of vision, so that the white trace ceases to be visible 

 when the analyzer is turned so as to intercept the rays which 

 are polarized in the plane containing the eye and the pencil of 

 solar light. 



When the flame is not smoky and when complete combus- 

 tion augments its brilliance, I had found, like M. Hirn, that the 

 polarization-phenomena are not sensible ; but 1 expressed some 

 doubts of the necessity of concluding from this the absence of 

 the reflecting-power of the particles at a high temperature; 

 the dazzling of the eye and the less quantity of these par- 

 ticles in a brilliant flame, in comparison with a smoky flame, 

 appeared to me sufficient to account for the facts. 



I have since endeavoured to control this way of viewing 

 the subject by concentrating the sunlight much more than I 

 had previously done, so as to give more brightness to its trace. 



The sunlight, reflected from a silvered mirror, falls upon a 

 good achromatic lens of 72 millims. aperture and 1*5 metre focal 

 length. Lastly, when greater concentration is required, a 

 second lens, much more convergent, is added, near the focus 

 of the first. The flame is then placed at the point where the 

 image of the sun is found. 



Working thus with different flames proceeding from carbu- 

 retted substances, the trace of the sun's rays can be perceived, 

 in most cases, very distinctly, and the usual phenomena of 

 polarization ascertained. When the flame is not too bright, and 

 does not fatigue the eye, the observation is readily made, by aid 

 of a Nicol, with the naked eye ; but if the flame is dazzling, 

 there is a great advantage in looking through one or several 

 plates of blue (cobalt) glass. The portions of the flame which 

 do not receive the light of the sun appear then of a purple tint, 

 while the trace of the pencil is clearly distinguished by its blue 

 colour. If observed through the Nicol in the proper position, 

 the blue trace disappears, and the whole flame appears purple. 



I have verified these facts in the following cases : — the flame 

 of a wax candle ; flame of ordinary gas from a Bengel burner 

 with a glass chimney, or from a butterfly burner ; flame of illu- 

 minating gas strongly carburetted, butterfly burner; flame of 

 a petroleum lamp, and of a moderator lamp with oil. 



I finally tried the very brilliant flame obtained by burning 

 illuminating gas strongly carburetted, with the addition of 

 oxygen. With the process of concentration of the sunlight 

 above described, the trace and its polarization are still distinctly 

 observed as long as the oxygen is not too abundant, the bright- 

 ness being, however, already incomparably more vivid than that 

 of an ordinary flame. 



By employing more energetic means of concentration — that 



E 2 



