and other Substances radiating in a strong Magnetic Field. 29 



out a middle line, in which case it shows as a doublet. For example, 

 if the vibration is restricted to one plane, and if this plane sets itself 

 at right angles to the lines of magnetic force under the influence of 

 the magnetic field, then the component of the vibration in the direc- 

 tion of the lines of force will be zero, and the middle line will vanish 

 from the triplet. If, on the other hand, the whole vibration should 

 set itself parallel to the lines of force, then the side lines of the 

 triplet would vanish, or, in other words, the spectral line would be 

 unaffected by the magnetic field. It is clear, therefore, that the 

 study of the way in which the spectral lines are affected by the mag- 

 netic field is likely to throw light on the character of the molecular 

 vibrations. 



The substance which one would expect to present peculiarities in 

 this way is iron, which, if it retains any of its magnetic properties at 

 the high temperature of spark produced between the terminals of 

 an induction coil, should exhibit some characteristic behaviour in the 

 magnetic field. I was led to expect, indeed, that many, if not all, of 

 the iron lines would be resolved into doublets rather than triplets 

 when the spark is viewed across the field. I found it no easy matter, 

 however, with the magnetic field at my disposal to resolve the iron 

 lines into anything ; in the first place, because the effect is much 

 smaller (about one-half) for iron than for the 4678 line of cadmium 

 or zinc, and, in the second place, because it is not possible to work 

 with the pole pieces of the magnet so close together when the spark 

 is passed from a solution of a salt as when it is passed between small 

 metal electrodes. But in the case of a highly magnetic substance 

 like iron the metal becomes almost unmanageable in a powerful mag- 

 netic field, and so it happened that in my first attempts I was not 

 successful in resolving the iron lines into either doublets or triplets ; 

 yet even at this stage I observed two or three lines converted into 

 what I considered doublets. Finally, I succeeded in resolving the 

 vast majority of the lines of iron by enclosing pieces of iron wire in 

 small glass tubes so that the ends of the wire protruded slightly from 

 the glass where the spark occurred. The pole pieces were then 

 pushed up to touch the glass jackets, and the spark was thus obtained 

 in a very strong field, in fact in a field strong enough to resolve the 

 majority of the iron lines into distinct triplets, showing that these 

 vibrations possess freedom in the magnetic field (at the temperature 

 of the spark) ; but, in addition to this, other lines are observed as 

 distinct doublets, and the appearance of these doublets is not that 

 usually associated with a reversal. In addition there are other lines 

 in the spectrum of iron which are scarcely affected, if at all. 



It is important to remark that these differences of effect by the 

 same magnetic field on the different lines of the spectrum of the same 

 substance are not shown in any particular ascending or descending 



