ANALYSIS OF DISTURBANCES. 179 



change is usually greatest in II, though sometimes in D. Taking the H change as the most notable, we can 

 usually see with certainty only an increase, but sometimes a smaller decrease for a relatively short time is 

 distinctly visible, and on other occasions, though not clearly visible, it is suggested by the appearance of 

 the curve. In the Antarctic there were six occasions May 8, August 20, and November 6, 1902; and 

 April 5, August 25, and December 13, 1903 (Greenwich dates) when sudden movements were detected 

 synchronous in time with sudden commencements at Kew and elsewhere. All these six were distinctly 

 double or oscillatory movements, and in all the second movement was the larger. So far there is no certain 

 difference from the phenomena at Kew, as we are not in a position to say with certainty that when only 

 one movement was seen at Kew it was unaccompanied by a previous very small and short-lived 

 diminution in H. But of the six Antarctic double movements, three had the first movement a decrease 

 in the elements, while three had it an increase. The three occasions which showed the increase first 

 were May 8 and November 6, 1902, and December 13, 1903. 



On May 8, 1902, we have an Antarctic disturbance A followed by a larger disturbance B, the two 

 synchronous with a Kew disturbance A. On August 20, 1902, we have an Antarctic B disturbance, 

 unrepresented elsewhere, followed by an A disturbance which synchronised with an A disturbance at 

 Kew. On November 6, 1902, we have an Antarctic A disturbance synchronising with a Kew B 

 disturbance, and then an Antarctic B disturbance synchronising with a Kew A disturbance. On April 5, 

 1903, we have an Antarctic B disturbance synchronising with a Kew A disturbance, being followed by a 

 larger A disturbance not represented at Kew. On August 25, 1903, we have an Antarctic B disturbance 

 synchronising with a Kew B disturbance, there immediately following a larger A disturbance at both 

 stations. On December 13, 1903, we have an Antarctic A disturbance not apparently represented 

 elsewhere, followed by a B disturbance which covered the time occupied by a B and A disturbance at 

 Kew, the latter the larger. 



All of the special disturbances present the B type first in the Antarctic, and this is true in three of the 

 four eases at Kew. Between 2.10 and 2.28 p.m., G.M.T., on May 8, 1902, we have a B disturbance at Kew, 

 but an A disturbance in the Antarctic. On July 24, 1902, the A and the B disturbances at the two 

 stations correspond. 



It would thus appear that whilst A and B movements in the Antarctic are just as opposed to one 

 another in type as they are elsewhere the one representing an increase, the other a decrease in the 

 elements which vary most the order in which they occur shows a variability which is at least unusual, 

 and either may synchronise with an A movement at the co-operating stations. 



100. The subject of the coincidence in time of the commencement of the magnetic storm of May 8, 

 1902, and the eruption of Mont Pel^e has been already referred to (57), and a further discussion was 

 promised. We see from Tables LXII to LXV that at Kew, Colaba, Mauritius and Christchurch and the 

 same is true of Falmouth the movement was of the same type A as other sudden commencements of 

 storms. The only peculiarity was that at Kew and Falmouth and possibly at Colaba there was a short 

 suspension of the upward movement in H, only just recognisable in the curve. In the Antarctic the 

 movement, it is true, was represented by an oscillatory A and B movement instead of by a simple A 

 movement but a similar phenomenon occurred on December 14, 1903. There seems thus to be nothing 

 of an outstanding character in the type of the commencing disturbance of May 8. As regards the relative 

 amplitudes of the disturbance at different stations, there is nothing at all outstanding in the ratio recorded 

 on this occasion in Table LXVIII. The Colaba disturbance was certainly relatively larger than usual, 

 though not so large relatively as two hours later in the same day, and the Mauritius disturbance was also 

 a little above average ; but there is nothing at all abnormal in the figures at these stations. Moreover, 

 the fact, that relatively considered the Kew and Falmouth disturbances are somewhat less than usual 

 compared to those at Colaba and Mauritius, is the reverse of favourable to the view that the disturbance 

 was directly due to the Mont Pele'e eruption. If an eruption, which consists of a vertical movement of 

 material, causes a magnetic disturbance, one would certainly expect it to be of a more or less symmetrical 

 character round the vertical at the place, and one would unquestionably expect the disturbance to fall off 

 rapidly as the distance increases. Now, somewhat curiously, Colaba, Mauritius and Christchurch are not 

 far from equidistant from Martinique, all being at an angular distance fully double that of Kew and 



2 A 2 



