Intelligence and Miscellaneous Articles. 171 



are so obvious that it will be surprising if it has not already oc- 

 curred to some one. Not only are the results better, but the expe- 

 riment is more easily performed. For the purpose I have employed 

 the vibrating prism manufactured by Mr. Ladd of London, to one 

 face of which I have affixed a small plane mirror. 



I am, Gentlemen, 



Yours faithfully, 



J. Loudon, 

 Prof. Math. &c Nat. Phil., 

 Univ. Coll. Toronto. 



ON THE TRANSPARENCE OF FLAMES AND OF THE ATMOSPHERE, 

 AND ON THE VISIBILITY OF SCINTILLATING SIGNAL-LIGHTS. 

 MEMOIRS BY E. ALLARD. (ABSTRACT BY THE AUTHOR.) 



I have the honour to present to the Academy several memoirs 

 relative to some experimental and theoretic researches which I 

 have been led to make in the exercise of my functions at the Depot 

 Central des Phares. 



The first memoir is concerning the transparence of flames. The 

 burners of the lamps employed in the light-house service have 

 diameters increasing from 3 up to 13 centims., and carry from one 

 to six concentric wicks. By measuring the luminous intensities 

 of the flames produced by them, it is ascertained that these inten- 

 sities augment a little less rapidly than the consumption of oil ; 

 and if compared with the dimensions of the flames, the intensity 

 per square centimetre of apparent surface is found to go on aug- 

 menting, whilst, on the contrary, the intensity per cubic centi- 

 metre of volume diminishes in proportion as the diameter becomes 

 greater. These results can only be explained by admitting that 

 the transparence of the flame is not absolute. I have sought to 

 determine the coefficient of this transparence by three series of 

 experiments : — first by measuring the intensity of different flames 

 with flat wick, seen in front or sideways ; then by means of a 

 catadioptric reflector, which sends back towards its focus the rays 

 which it receives, and thus obliges them to pass through the flame ; 

 and, lastly, by measuring the intensity of an electric light through 

 a flame of large diameter. I have been led to adopt 0-80 as the 

 mean value of the coefficient, referred to the centimetre of thickness 

 traversed. 



After establishing the theoretic formulae giving the effective 

 intensity of a flame as a function of its volume and its coefficient 

 of transparency, I perceived that, to account for the observed 

 intensities, it was necessary, with a coefficient of 0*80, to attribute 

 to the flames mean specific intensities increasing a little with 

 the diameter. Multiplying these specific intensities each by the 

 volume of its flame, I find that the total quantity of light produced 

 (or the absolute intensity) increases much more quickly than the 

 weight of oil consumed ; but, as the quantity of light absorbed by 

 the passage of the ray through the flame increases in a still greater 

 proportion, the difference between these two quantities (or the 



