178 ANALYSIS OF DISTURBANCES. 



eight cases, we see that at Kew Nos. (1), (5), and (7) may be regarded as of the Class A of Table LXII, 

 whilst Nos. (2), (3), (4), (6), and (8) are of Class B. 



Coming to Colaba we recognise No. (1) as of the Class A of Table LXIII, and Nos. (2), (3), and (4) 

 as of Class B. So far as the angle 6 is concerned, Nos. (5) and (7) approach Class A, and Nos. (6) 

 and (8) approach Class B, but the < angle differs rather notably from that characteristic of the respective 

 classes. 



At Mauritius we can recognise No. (1) as of the Class A of Table LXIV, and Nos. (2), (3), (4), and (8) 

 as of the Class B. 



At Christchurch Nos. (1) and (5) are of the Class A of Table LXV, though No. (5) is rather outstanding, 

 and Nos. (2), (3), and (4) are good examples of Class B. After 5.42 p.m., G.M.T., the Christchurch curves 

 seemed to lose their parallelism with those at Kew entirely. 



At the Antarctic No. (1) is fairly of the type of Class A of Table LXVI, and No. (2) is fairly of Class B. 

 The turning-points, however, in the Antarctic were not very clearly marked in the H and V curves, 

 and whilst they were clearly marked in the D curve in which alone the movements stood out from their 

 neighbours they appeared to be hardly absolutely coincident in time with the movements they were 

 believed to correspond to at Kew. 



98. In the course of our discussion of Tables LXII to LXVII there have been references to the relative 

 size of the corresponding disturbances at different stations. The information on this point is summarised 

 in Table LXVIII, p. 185, which expresses the amplitude of the disturbance All at each station in terms of the 

 amplitude of the synchronous disturbance at Kew. In addition to results from individual cases the table 

 gives mean results, treating separately the sudden commencements, the special type of disturbance, and the 

 disturbances which belong to neither of these categories. Considering the comparatively limited data, too 

 much weight must not be attached to numerical resemblances which may be partly accidental. 



We have already seen that, so far as type is concerned, sudden commencements do not appear to be in 

 any way essentially different from other short-period movements such as those of July 24, 1902. The 

 similarity seems to extend to the variation in amplitude with geographical position. Taking either of 

 these types of disturbance, the amplitudes at Mauritius and Christchurch are decidedly less than those at 

 Colaba, which in their turn are less than those at Kew and Falmouth. But the movements, even at Kew 

 and Falmouth, are much exceeded by those in the Antarctic, and this seems the case habitually, irrespective 

 of whether the time of occurrence is day or night at Greenwich. 



With regard to the relatively small size of sudden movements at Mauritius and Christchurch there is 

 one possible explanation which must be borne in mind. In selecting disturbed days for comparison I had 

 before me only Kew and Antarctic curves, the latter being invariably disturbed. Thus, no doubt, it was 

 the amplitude of the Kew disturbance, or some special feature in it, which mainly determined the choice. 

 It is thus conceivable, if instead of the Kew curves I had had the curves from some station in the southern 

 hemisphere to guide me, that a different selection might have been made, and that under these circum- 

 stances there might not have been the pre-eminence in the amplitude of the disturbances at Kew and 

 Colaba as compared to those at Mauritius and Christchurch which the table shows. Some countenance to 

 this view is supplied by the results for the special type of disturbance. Corresponding movements could 

 indeed be traced at Kew, but in no case were they so outstanding as to have caught the eye if merely 

 glancing through the curves generally. In their case it will be seen that the disturbance experienced at 

 Christchurch was usually much larger than that at Kew. If, however, the selected sudden commencements 

 represented magnetic effects whose seat was mainly in the northern hemisphere, the extraordinarily large 

 size of these movements in the Antarctic would be even more remarkable than it appears to be. Thus I 

 do not think that the suggested explanation suffices, though there may be some truth in it. 



As regards the special type of disturbance, we must, I think, conclude that though the effects are felt 

 in the northern hemisphere, yet the seat of the disturbance must be mainly at least in the southern 

 hemisphere. 



99. There is one notable peculiarity about the Antarctic results that calls for comment. At Kew, 

 Falmouth, Colaba, Mauritius, and Christchurch the result of a sudden commencing movement, whether single 

 or visibly double, is normally there may be exceptions to leave the total force increased. At Kew the 



