396 M. H. Becquerel's Experimental Investigations 



is greater than that of the numbers obtained for the magnetic 

 rotations of the gases. 



II. Want of Homogeneity in the Luminous Source. — A 

 shaded polariscope, formed of a divided nicol, such as the one 

 employed by us, might be used for studying light of any 

 colour whatever, provided that the luminous rays of different 

 wave-lengths experienced the same action. Under the in- 

 fluence of magnetism, on the contrary, the planes of polari- 

 zation of the different rays are unequally deflected, the image 

 from the polariscope is tinged with various colours, and before 

 any measurements are possible it is necessary to make use of 

 a monochromatic source of light. On the other hand, our 

 apparatus needed an extremely brilliant source of light, on 

 account of the very long distance generally traversed by the 

 luminous rays before reaching the eye : the passage through 

 the glasses and the successive reflections are all causes of loss 

 of light ; and equality of tint is obtained by almost extin- 

 guishing the images, and preserving not more than about 

 0'0019* of the intensity of the incident ray. 



I first attempted to direct the blast of a compound blow- 

 pipe onto a large piece of fused chloride of sodium. This 

 light is extremely bright for a moment; but the salt begins to 

 melt very rapidly, without giving time for a measurement to 

 be completed by reversing the direction of the electric current. 

 It is true that the light might be maintained by progressively 

 displacing the piece of chloride of sodium, which melts only 

 at one point ; but this gives rise to a very grave cause of 

 error: the luminous point is displaced; and this displacement 

 causes corresponding deviations of the plane of polarization, 

 which completely disturb the phenomena due to magnetic 

 influence. I was therefore obliged to relinquish the employ- 

 ment of this source of light. 



When the piece of chloride of sodium just spoken of is re- 

 placed by a piece of quicklime, the light emitted by the in- 

 candescent lime is composed of nearly all the rays which 

 form the visible part of the solar spectrum, but the luminous 

 point has the advantage of remaining almost perfectly fixed. 



As the rotations observed in the apparatus are very slight, 

 the images are only very faintly coloured, and the measure- 

 ments of equality of tint can be made with great exactitude. 

 They become still clearer if a coloured screen, which allows 



* This number is obtained as follows : — The position of the analyzer 

 which corresponds to the equality of tint is at 2° 30' from those which 

 correspond to the extinction of each half of the image from the polari- 

 scope. In this position, therefore, the luminous intensity is represented 

 by cos 2 87° SO'zsO'OOlOOO, the intensity of the incident ray being taken as 

 unity. 



