186 SPECIAL TYPE OF DISTUEBANCE. 



CHAPTER X. 

 SPECIAL TYPE OF DISTURBANCE. 



102. Reference has already been made, 69, to a " special type " of disturbance in the Antarctic, 

 examples of which are afforded by the curves of June 19, 28 and 29, July 26, and August 17, 1903 

 (see Plates XXVII-XXXI). The close resemblance of the disturbances of June 28 and 29 first drew my 

 attention, and, in order to judge of the real significance of the apparent repetition of a disturbance at a 

 nearly 24-hour interval, it appeared necessary to examine the curves in detail. Examination soon showed 

 that the type of disturbance was of somewhat frequent occurrence during certain hours of the day. 



The essential part of the phenomenon may be regarded as consisting of two phases. During the first 

 there is normally a slight fall in V, during the second there is always a rise, usually considerably larger 

 than the preceding fall. 



During the first phase D practically always decreases, and H usually, though not always, does the same. 

 During the second phase the D and H changes are in the opposite direction to those during the first phase 



On the crests, so to speak, of the D and H waves minor oscillations usually occur (cf. Plate XXIX), also the 

 slope of the V curve near its lowest point is sometimes very slight. Thus the time when the first phase 

 ends and the second begins is usually uncertain to a minute or two, and a greater uncertainty often 

 attaches to the beginning of the first phase and the ending of the second. The main movement was 

 usually accompanied by the minor variations which were of such persistent occurrence in the Antarctic. 

 But these minor movements and tie- -various- uncertainties which enter into individual cases will, it is 

 believed, have but little influence on the general conclusions reached below. 



What happened after the end of phase 2 varied a good deal. June 29, 1903, represents the most 

 usual order of events. Here V remains nearly uniform for some time. Not infrequently, however, after 

 reaching a maximum V made some further oscillations, and then diminished somewhat rapidly, as in the 

 curve of June 28, 1903 (Plate XXVIII). 



In some cases the to-and-fro movements in D and H were closely alike ; in other cases there was a 

 good deal of asymmetry. In some cases the return wave, so to speak, was checked by what seemed to be 

 a second wave surging in before normal conditions had been restored. In a few instances there were two 

 complete examples of the phenomenon in a single day, and once or twice in immediate sequence to one 

 another. 



As will be seen presently, the great majority of the occurrences took place between 6 and 10 p.m. 

 (local time), no occurrences being noted from 2 a.m. to 1 p.m. There is, however, more than one possible 

 explanation of this fact. At most seasons of the year the type of the disturbed movements varied 

 throughout the 24 hours. During the afternoon, up to 10 or 11 p.m., the curves exhibited a marked 

 tendency to rounded swelling movements, like a succession of irregular waves with a considerable interval 

 between the crests. There were incessant minor oscillations superposed on these, but they did not usually 

 obliterate the long-period waves. In the early forenoon, on the other hand, slow-swelling movements 

 were rarely recognisable, while short-period oscillations became dominant, the curves appearing tumultuously 

 irregular and highly disturbed. 



The hour marks in the D, H and V curves were seldom exactly in a line transverse to the hour lines, 

 and in the case of these short-period oscillations it was exceedingly difficult to decide whether crests on 

 the different curves answered exactly to one another. It is by no means improbable that microscopic 

 examination of the curves would have shown a number of cases in the morning where phenomena of the 

 type now being discussed occurred, only with the phases much shortened. In many morning disturbances 

 it was clear that a turning-point answering to the minimum value of D during a rapid to-and-fro move- 

 ment answered, at least very approximately, to the lowest point of a sharp V depression. There were, 

 however, usually a succession of crests and depressions on the V traces at such times, the successive waves, 

 tending to become confused. 



