52 M. J.-L. Soret on Polarization 



is to say, by causing the sunlight, reflected by the large mirror 

 of a siderostat, to pass through an objective of 8 inches aper- 

 ture, and then through a lens of short focus, I could perceive 

 the trace on this flame of carburetted gas fed by a larger pro- 

 portion of oxygen than in the preceding case ; but when the 

 oxygen became too abundant, the trace was no longer visible — 

 which, beside difficulties of observation, may be explained, 1st, 

 by the fact that, the flame having become quite white and even 

 bluish, there is no longer any difference of tint between the 

 part which receives the solar rays and the parts which do 

 not receive them, the trace can only manifest itself by a differ- 

 ence of intensity more difficult to perceive ; 2nd, by the carbon 

 particles being immediately consumed at the moment of their 

 formation, and consequently the reflecting matter becoming re- 

 latively much rarer. 



In short, these experiments show that carbon retains its re- 

 flecting-power at very elevated temperatures, which tempera- 

 tures it would nevertheless be difficult to state precisely. 



Further, these facts appear to me to have some interest be- 

 cause they confirm, at least for ordinary flames, the theory of 

 Davy, which has recently been strongly contested ; in fact, a 

 pencil of solar light is reflected by diffusion and polarized in 

 precisely the same manner, whether it falls on a very brilliant 

 flame, or whether it illuminates non-incandescent smoke, in which 

 the presence of carbon 'particles is incontestable. 



The Cause of the Illumination of Transparent Bodies and 

 of Diffusion. 



In my former publications I have maintained the opinion 

 that the illumination of transparent bodies traversed by a pencil 

 of rays must be attributed to a defect of homogeneity in Qie 

 medium — a defect consisting most frequently in the dissemina- 

 tion of foreign particles of great tenuity, but may also result 

 from differences of refrangibility in the component parts of the 

 medium, or, in the case of a solid body, from minute voids or 

 fissures. In other terms, illumination is for me only a parti- 

 cular case of diffusion of light. 



My learned friend M. Lallemand attributes this phenomenon 

 to the molecules themselves of the transparent body ; be regards 

 the illumination as a lateral propagation of the incident lumi- 

 nous motion, caused by the condensation of the sether around 

 each molecule. Thus, for him, a pencil of light traversing a 

 body whieh is transparent, not fluorescent, and absolutely homo- 

 geneous, must in general give rise to a trace visible laterally, 

 and the phenomenon must depend essentially on the nature 

 itself of the medium in which it is produced. 



