158 Professor Thomas Preston [May 12, 



magnetic triplet. The frequency of the vibration in one of these 

 lines is the sum, and the frequency of the other is the difference, of 

 the two frequencies mentioned before— namely, the frequency of the 

 revolution of the ion around its orbit, and the frequency of the pre- 

 cessional revolution of the orbit round the lines of force. The centre 

 line of the triplet has the frequency of the original vibration, and 

 this frequency disappears completely when the light is viewed along 

 the lines of force — that is, through axial holes pierced in the pole- 

 pieces. In this direction, too, a further peculiarity arises, for not 

 only does the triplet drop its central member and become a doublet, 

 but each member of this doublet is not plane-polarised, as the 

 members of the triplet are. They are each, on the contrary, cir- 

 cularly polarised — that is, the vibration is circular instead of being 

 rectilinear. 



This all follows as the expectation of the simple theory which 

 supposes that the ions are free to describe their elliptic orbits undis- 

 turbed by any forces other than the magnetic field. But it is only to 

 be expected that other perturbing forces must come into play in the 

 assemblage of ions which build up incandescent matter of the source 

 of light. We know, for example, that the other members of the solar 

 Bystem perturb the earth's motion, so that it deviates from the simple 

 elliptic motion predicted by the simple theory which did not take 

 these perturbing forces into account. Hence, if any such perturbing 

 forces exist, and we should be surprised if they did not exist, the 

 tripling pure and simple of the spectral lines will be departed from, 

 and other types will arise. From the character of these new types 

 we may infer the nature of the perturbations which give rise to 

 them, and hence by the study of these types we obtain a view of what 

 is going on in matter when it is emitting light, which we should not 

 possess if such perturbations did not occur. These deviations from 

 pure tripling are consequently of more importance almost, in regard 

 to our future progress, than the discovery of the tripling itself. To 

 give you some idea of the influence of such perturbations in modi- 

 fying the triplet form, I may mention that it follows from simple 

 theoretical considerations, that if the perturbing forces cause the 

 orbit to revolve in its own plane, or cause it to change its ellipticity 

 periodically, then each line of the triplet produced by the magnetic 

 field will be doubled, and a sextet will result, and other oscillations 

 of the orbit will give rise to other modifications of the normal triplet 

 type. It is not quite easy to see at once, however, what the per- 

 turbing forces are exactly, for we do not know the way in which the 

 ions are associated in matter ; but if we regard an ion as a charged 

 element of matter describing an orbit, it will be analogous to a closed 

 circuit, or to a magnetic shell, and will be urged to set in some 

 definite way in the magnetic field. In coming into this position it 

 may oscillate about the position of equilibrium, and thus introduce 

 an oscillation into the precessional motion of the orbit, which may 

 have the effect of doubling or tripling the constituents of the pure 

 precessional triplet. 



