TRANSACTIONS OF SECTION A. 789 



The sensitive part consists of a loop of pure platinum wire, one-tbousandtli of an 

 inch in diameter, suitably protected in a tube or other enclosure. 



On account of its great sensitiveness, and the larjre change of resistance, the 

 instrument is best adapted for measuring small variations of voltage over a limited 

 range, such as 100 to 120 volts. This gives a scale of 1 cm. per volt with a scale 

 of approxiirately equal parts over the range covered. 



On Radiation from a Source of Light in a Magnetic Field. 

 By Professor T. Preston, F.R.S. 



9. On the Discovery by Righiofthe Absorption of Light in a Magnetic 

 Field. By Silvanus P. Thompson, D.Sc, F.R.S. 



An intense horizontal beam of siiulight or arc-light is passed along the axis of 

 a Ruhmkorti's electromagnet. It is polarised at its entrance by a Nicol's prism, 

 and by a second Nicol turned at right angles to the first it is extinguished at its 

 exit. Let a sodium liame be now placed between the poles of the magnet. When 

 the electromagnet is excited there appears in the field of vision a yellow spot of 

 light, which, on being examined by a direct- vision spectroscope, shows the emission 

 spectrum of sodium. If the observation is made without a spectroscope, and the 

 analyser is turned, the yellow light does not disappear, but turns into a white 

 Light of increasing brightness. Its occurrence cannot therefore be accounted for on 

 the basis of the liiraday phenomenon of the rotation of the plane of polarisation. 



This result may be considered upon the following basis. Let Ji^ be the frequency 

 for a single absorption line in the spectrum of the body under examination. If the 

 field-magnet is excited, then in consequence of an action the converse of that which 

 exists in the Zeeman phenomenon the body no longer absorbs light of the fre- 

 quency Wj, but instead absorbs two kinds of light — namely, one having a right- 

 handed circular polarisation and of frequency w.,, the other having a left-handed 

 circular polarisation of frequency n.^\ the numbers «., and 'n^ being so related that 

 one of them is slightly less than Wj and the other slightly greater. Optically the 

 result is that there is now present circularly polarised light, some right-handed of 

 one frequency, some left-handed of a different frequency, which, being circularly- 

 polarised, cannot be cut of!" in any position of the analysing Nicol. The latter 

 allows only the transmission of components parallel to its principal plane of 

 section ; and the yellow light which appears on the stimulation of the electromagnet 

 is a proof that an effect converse to the Zeeman effect occurs in those cases where 

 the body is absorbing instead of emitting light. 



The intensity of the light which becomes visible on the stimulation of the 

 magnetic field is obviously proportional to the intensity of the source of light from 

 which emanates the beam that travels along the axis of the apparatus. A com- 

 paratively weak magnetic field suffices to produce the phenomenon. A field of 

 intensity 300 C.G.S. units suffices ; and with the ordinary pattern of Ruhmkorff 

 magnet a single bichromate cell is adequate. The author has found no difficulty in 

 I'epeating this fundamental etl'eet in the laboratory of University College, Bristol. 



If lithium is used, the light which appears is of course red ; if thallium is used, 

 the light is green. 



Apart from its great sensitiveness the new method of observation possesses 

 other features. What was said above about the beam having frequency Wj also 

 holds good for any and every other colour absorbed by the body, even in cases 

 where these colours follow one another continuously and the absorption spectrum 

 is no longer a line spectrum. The new method permits the proof to be given that 

 the Zeeman effect occurs in bodies in which its occurrence could not be demon- 

 strated in any other way. For, in order to carry out the observation of Zeeman 

 in its original form, it was necessary that the spectrum, whether an emission or an 

 absorption spectrum, should consist of sharply defined lines, since otherwise the 

 splitting of the lines could not be demonstrated. 



