DIURNAL INEQUALITIES. 103 



at Winter Quarters. The contribution from II to 81 is usually much the larger. Thus the curves in fig. 12 

 have a considerable resemblance to those in fig. 10, with the sign of the ordinates reversed. The 

 Midsummer curve, however, in fig. 12 shows a slackening in the rate of fall of I about 7 or 8 a.m., which is 

 due to the influence of the V contribution. 



Winter Quarters was only about 400 miles from the position which Commander CHETWYND has deduced 

 for the south magnetic Pole. Thus, if the diurnal inequality is due to electrical currents in the upper 

 atmosphere, we should on the whole expect no great difference to exist between the amplitudes of the 

 diurnal inequality of I at Winter Quarters and in the immediate neighbourhood of the Pole itself. This 

 would imply, of course, a considerable diurnal variation in the actual position of the Pole. The probable 

 nature of this movement may be derived from the consideration that an increased Inclination at Winter 

 Quarters is presumably equivalent to a diminution in its distance from the magnetic Pole. Inclination 

 was largest at Winter Quarters about 4 or 5 a.m. in the morning, so that the magnetic Pole was presumably 

 at that hour displaced towards Winter Quarters, i.e. was to the SE of its mean position for the day. 

 Inclination at Winter Quarters was lowest from 2 to 3 p.m., so that the magnetic Pole at that hour is 

 probably to the NW of its mean position for the day. 



30. From a mathematical standpoint much is to be said for treating the diurnal variation of the 

 components of force in and perpendicular to the geographical Meridian as fundamental rather than those 

 of the Declination and Horizontal Force. The arguments in favour of this course are in some respects 

 stronger for Winter Quarters than for an ordinary station, because the magnetic Meridian there was so 

 highly variable. Again, if the causes to which the diurnal inequality is due are related to the Earth's axis of 

 rotation rather than to the position of the magnetic Pole, the diurnal inequalities of the components S and 

 W, directed to geographical south and west, are likely to be much less variable round a parallel of latitude 

 than those of D and H. If the orientation of the spot relative to an adjacent magnetic Pole is of none, or 

 of but secondary importance, the diurnal inequalities of D and H at Winter Quarters probably differ largely 

 from those at other stations only 50 or 60 miles to the west, whereas the diurnal inequalities in S and W 

 are unlikely to differ much. On the other hand, any element of uncertainty that may attach to the 

 absolute value of the Declination enters into the diurnal inequalities of S and W, whereas it hardly enters 

 into the diurnal inequality of D itself. Another drawback is that the S and W inequalities each depend 

 on both D and H, whereas the days contributing to the D and H inequalities differ to some extent. If the 

 D and H contributions to S and W had been derived strictly from the same days, the results would 

 probably have been slightly different from those actually obtained. This last source of uncertainty might 

 of course have been avoided, but, considering all the circumstances, it seemed hardly worth while to 

 recalculate D and H inequalities from a common series of days. 



The methods of calculating the inequalities in Tables XXI and XXII have been already explained 

 in 23. Both S and W show only a single daily period. The extreme values of W occur distinctly earlier 

 in the day than those of S. As appears either from the ranges or the sum of the hourly differences from 

 the mean, the amplitude of the daily changes of W is more variable throughout the year than is that of S. 



The inequalities of S and W are combined and shown graphically in the vector diagrams of figs. 13. In these 

 NS and EW are drawn respectively in and perpendicular to the geographical Meridian, and the ends of the 

 lines denoted by the letters N, S, &c., are each at a distance from the point of intersection which represents 

 lOy on the scale to which all the diagrams are drawn. The crosses and the numbers attached answer to 

 the hours of the day counted from 0, local midnight. The line drawn from the intersection of NS and 

 EW to a particular cross represents in direction and magnitude the horizontal component of the force 

 acting on the north pole of the magnet to which may be ascribed the departure of the magnetic elements 

 at the hour indicated from their mean value for the whole 24 hours. The diagrams are all described 

 anti-clockwise, and so in the opposite direction to corresponding diagrams for English stations. 



Being based on less than 2 years' data, the diagrams are naturally not very smooth, but their general 

 form is remarkably symmetrical. The difference of type between the diagrams for the different seasons is 

 unusually small. Another exceptional feature is the comparative smallness of the difference between day 

 and night, the angular velocity of the vector being only very slightly less during the 12 hours centring 

 at midnight than during the corresponding " day " hours. 



