164 RECORDS OF DISTURBANCES 



successive days at about the same hour until the sensitiveness hail disappeared. I have seen in the Kew 

 curves a considerable number of instances of the kind referred to by Senor CAPELLO and Dr. STEWART, 

 but in few, if any, has there been so close a resemblance taking all the elements into account as that 

 presented by the Antarctic curves of June 28 and June 29, 1903. The interval between the two 

 occurrences, as already stated, was only approximately 24 hours, but in this respect there is no departure 

 from what has been observed of apparent repetitions at Kew and elsewhere. In considering whether more 

 than chance enters into the occurrence regard must be had to the greater or less frequency of the particular 

 type of disturbance, and as to whether it shows a preference for particular hours of the day. It is shown 

 later, p. 189, that this special type of disturbance was in reality of somewhat frequent occurrence in the 

 Midwinter months, and it happened much oftener during the four hours 6-10 p.m. than throughout 

 the remaining 20. Thus the fact that two occurrences should follow one another at nearly a 24-hour 

 interval hardly possesses the significance one is disposed to give it at first sight. 



On June 28 the second phase seems to have been hardly completed before the arrival of a second 

 wave-like disturbance of much shorter period. After the passage of this short-period disturbance the 

 V curve comes down the sheet for a time more rapidly than was usual in the special type of disturbance. 

 The disturbance of June 29 was more normal in this respect. 



71. Plate XXX shows another example of the special disturbance occurring on July 26, 1903. The end 

 of the first phase occurred about 5.55 p.m., L.T. In this instance the H trace indicates first an increase, 

 then a decrease, the element being so to speak opposite in phase to the other two, in which the decrease in 

 value comes first. As already stated, this opposition in phase in H sometimes happened, though in but 

 a small minority of cases. The movement in the Antarctic synchronised with the first portions of the 

 disturbance shown in the three lower Christchurch curves of Plate XVII. As Plate XXX shows, there 

 were later disturbances on the same date in the Antarctic, and one or two of the movements show some 

 resemblance to the principal one described above. The V trace throughout its entire length shows 

 a tendency to wave-like swelling forms, and it will be noticed that the hollows in these tend to synchronise 

 more or less closely with peaks on the D and H curves. 



The interval between the disturbances of June 29 and July 26 is approximately 26 days 21| hours, and 

 so is not far from the period 27- days which Mr. MAUNDER believes his researches show in magnetic 

 storms at Greenwich. It cannot be said that there is a closer resemblance between the storms of July 26 

 and June 29 than between those of the former date and the storms of June 28 and June 19. So that the 

 evidence is equally strong for a considerable choice of periods. 



72. Plate XXXI shows another example of the special type of disturbance occurring on August 17, 1903. 

 The end of the first phase was about 7.54 p.m., L.T. The disturbance is not large, but it occurred during 

 a very quiet period. The slowness with which V diminished after attaining its maximum is shown with 

 especial clearness, because the temperature was unusually steady, so that no serious complications arise 

 through it. 



73. Plate XXXII shows a somewhat highly disturbed state of matters in the Antarctic on August 22-23, 

 1903, synchronising with the disturbances at Christchurch shown in Plate XVIII. The H curve was off 

 the sheet during a considerable portion of the time, first on the one side, then on the other. Probably 

 the range in the element very largely exceeded that shown, 150y. The D curve went also off the sheet, 

 but only for a short time, a little after 4 p.m., L.T. ; the range shown is 3 45'. Even the V curve shows 

 continual oscillations ; the range shown by the trace, 200y, is only slightly affected by the temperature 

 variation, which was small. There is at least an approach to the special type of disturbance between 

 3.55 and 4.50 p.m., L.T., but much irregular disturbance existed at the time. 



74. Plates XXXIII and XXXIV show respectively the concluding portion of the trace of August 

 25-26, 1903, and the commencing portion of the trace of August 26-27. The days intermediate between 

 August 22 and August 25 showed also a good deal of disturbance. During the latter part of August 26 

 the V magnet was out of action and its trace docs not appear in Plate XXXIV. Plates XXXIII 

 and XXXIV cover the time shown by the disturbances at the co-operating stations which appear in 

 Plate XIX. At the co-operating stations the storm had a sudden commencement, consisting of to-and-fro 

 movements the latter much the larger occupying from 10.57 to 11.3 p.m., G.M.T., on August 25. 



