AT WINTER QUARTERS. 169 



like movements, possessing the regularity and considerable period of the special type of disturbance, but 

 instead a constant succession of short-period oscillations or ripples in all three elements. The movements 

 are so large, and the intervals between turning-points so short, that it is hardly possible to decide the exact 

 relationship of the changes in the three elements. Judged by the ranges recorded, the disturbances are 

 not very large ; but this merely affords an illustration of how unsatisfactory a criterion of disturbance 

 range alone may be. Thus, on August 15, 1903, during the 2 hours considered, the D range was 

 only 2 9'; but at least 80 turning-points can be recognised on the curve, and the aggregate of the 

 changes in D, taken irrespective of sign, is about 12. 



Rapid as the oscillations are in the curves of Plate XLIII, they are, of course, infinitely slow compared to 

 the magnetic changes that accompany the sudden creation or stoppage of an electric current under normal 

 conditions. If they represent the waxing and waning, or the change in distance and direction of electric 

 currents, these changes must presumably present nothing approaching to a discontinuity, in the mathe- 

 matical sense. 



It should be explained that there is no reason to suppose that the magnets under these rapid oscillations 

 behaved in any but the most dead-beat way. There is no suggestion in the curves of vibrations possessing 

 their period. 



