AT WINTER QUARTERS. 167 



selected as representing various types of disturbance. They were first drawn by hand on tracing paper and 

 then photographed. The disturbances included in any one of these plates have at least a general similarity of 

 type. One or two hours (local time) are marked on the base lines. The curves are shown natural size, so 

 that 20 mm. of abscissa represents very nearly an hour. The principal object being to illustrate the types 

 of disturbances, numerical magnitudes were of minor importance, and scale values are thus not shown, to 

 avoid complicating the figures. One could hardly have shown the scale values clearly without greatly 

 extending the area of the plates. If, however, scale values are desired they can be obtained by reference 

 to Table I, and they can be derived approximately from the information given as to the ranges of the 

 elements. The range recorded (D in minutes of arc, H and V in terms of ly) refers, in all cases, to the 

 portion of the disturbance actually shown. 



78. Plate XXXVII contains examples of the special type of disturbance which is dealt with in detail 

 in Chapter X, and some closely allied forms. The curve of May 22, 1903, represents a simple example of 

 the normal kind. Shortly before 8 p.m the D and H curves start running up the sheet (i.e. elements both 

 diminish), the V trace being nearly straight, but with a slight tendency downwards (V diminishing). 

 This constitutes the first phase. Superposed on the general drift in the D and H curves are minor 

 oscillations. These retider it difficult to say, to the minute, when the turning-point was reached ; but all 

 the elements show one within a few minutes after 8 p.m. The second phase then begins. The D and II 

 traces come down the sheet (elements increase) at a rate very similar to that of the first phase. V goes 

 up the sheet (element increases) to a maximum, the rise being considerably in excess of the previous fall. 

 This concludes the second phase, at the end of which H has practically returned to its original value, while 

 V retains an enhanced value for some time, no conspicuous fall setting in until half an hour after the 

 maximum was reached. 



On August 31, 1902, we have again a disturbance of the special type, but the fall in V during the first 

 phase is unusually large, and the turning-point in the D curve appears delayed. The end of the special 

 type of disturbance may be put at about 9.20 p.m., when the lowest point appears on the H curve (maximum 

 of force). Thereafter the H curve rises sharply again, and the V curve comes down the sheet, a second 

 wave, as it were, rolling in. After the first 20 minutes of this second wave the H curve hesitates and 

 fluctuates, indicating presumably the action of some independent source of disturbance. This, though 

 able only to check the H movement, sufficed to alter the direction of the D movement, which seemed at 

 first to be about to follow the H. Thus we do not get a repetition of the special type of disturbance. 



The limit of registration of the H curve was exceeded though apparently only very slightly just 

 before 9 p.m. ; the range recorded is measured from this limit. As previously explained, the light was 

 really cut off through the D instrument coming in the way. 



Unlike the two other disturbances, that of August 15, 1903, is an early morning one. There are a 

 succession of waves, crests of D and II (i.e. minima of force) answering, at least approximately, to troughs 

 (minima of force) in the V curve. The movements in V and D represent nearly equal force components, 

 each fully double the visible change in H. Judging, however, by the appearance of the H trace, the 

 limit of registration was a good deal exceeded. Towards 2 a.m. a final wave will be noticed, which at 

 least approached the special type of disturbance. 



79. Plate XXXVIII shows some transitional forms having a greater or less resemblance to the special 

 type. The disturbance of June 4, 1903, is essentially of this type. H, it is true, is opposite in phase to D, 

 but that happened in an appreciable minority of cases. There were clearly, however, subsidiary distur- 

 bances, V showing a succession of small oscillations after the maximum was reached. Owing to subsidiary 

 oscillations, the crest of the D wave is not very clearly indicated, and no assistance in fixing the turning- 

 point is derivable from the H curve, as the trace was off the sheet for fully an hour. 



October 4-5 and October 11-12, 1902, are examples of midnight disturbances. In both, the H 

 movement is at least strongly suggestive of the special type. The V magnet, however, was out of action, 

 so that no corroboration is forthcoming from its trace. The D curve is of a somewhat different type. 

 There is at least a suggestion of waves, but they seem of a shorter period than those in H. On both days 

 the D and H curves appear opposite in phase when the large movement in H begins, but towards the end 

 of the interval they seem nearly in the same phase. 



