414 Influence of a Magnetic Field on Radiation Frequency. 



nicol, placed in the rajs emanating transversely to the magnetic 

 lines of force, cuts off nearly all the visible magnetic effect when 

 oriented so as to get rid of light whose plane of polarisation contains 

 the lines of force — that is, of oscillations or revolutions whose 

 electrical components are across or around the magnetic lines. That 

 it does not cut off every trace of the effect appears to be due to the 

 fact that the field of force is not strictly uniform, and so its lines are 

 not strictly parallel. 



The following is a summary of the different appearances that may 

 be seen according to the state of the flame and the strength of the 

 field:— 



At low temperature, and with the flame forward in the field, when 

 each sodium line is sharp and single, magnetism widens it, and with 

 a little more power doubles it, causing a distinct dark line down its 

 middle. The same effect occurs w T ith lithium and thallium lines. 



At higher temperature, and with the flame partially behind the 

 field, when each sodium line appears as a broad hazy-edged double, 

 magnetisation greatly widens the doubling, pushing asunder the 

 bright components very markedly; stronger magnetisation reverses 

 the middle of the widened dark band, giving a triple appearance ; 

 stronger magnetisation still reverses the middle once more, giving a 

 quadruple appearance to the line. In every case a nicol, suitably 

 placed, cuts off all the magnetic effect and restores the original 

 appearance of the line. 



A curious circumstance is that although both lines, D x and D 2 , 

 show the effect, D u i.e., the less refrangible line, shows it best 

 and most sharply. I should describe the effect on D 2 as a coarse 

 widening of considerable amount, but without very clear definition; 

 whereas the widening of B 1? though perhaps no greater in amount, is 

 decidedly better defined. There is no doubt but that, with my 

 grating, Dj is the line at which one finds oneself usually looking in 

 order to see the details of the change best ; and I can hardly sup- 

 pose this to be subjective to the grating. I hope to show the effects 

 at the soiree next Wednesday. 



[The same thing is seen when salts of lithium or of thallium are 

 introduced into the flame ; and the components of the doubled red 

 lines are more widely separated than the components of the doubled 

 green lines, the effect being proportional to wave-length. The most 

 interesting line to try was the red cadmium line, since this has been 

 proved to be of specially simple constitution by Michelson. We 

 have recently been able to get the cadmium spectrum well developed 

 by means of a sort of spark arc between the magnet poles, main- 

 tained by an induction coil excited by an alternating machine; and 

 we find that the magnetic doubling of the chief lines occurs in 



