DIURNAL INEQUALITY I ;s. 101 



average highly disturbed day the non-cyclic change was opposite in sign to that on the average all- 

 day of T:il>lo XXIII, and so the same in sign as that in the average quieter day. Also the non-cyclic 

 change on the highly disturbed days must have been numerically large, at least -0' 4 2. It must be 

 allowed that this result looks at first sight exceedingly improbable, as one would expect apriori that the non- 

 cyclic phenomena on quiet days and on highly disturbed clays would be the opposite of one another. 

 However, probable or improbable, the fact remains that exactly the same phenomenon baa been recently 

 discovered at Ke\v, where instrumental uncertainties arc too small to be of importance. At Kew the non- 

 cyclic change was found on the average highly disturbed day of the 11 years 1890 to 1900 to agree in 

 sign with that on the average quiet day, while differing in sign from that on the average day of the year 

 (and so from the secular change). The direction of the secular change of Declination has altered at most 

 places within historical times, and presumably will alter from time to time in the future. Thus here again 

 we appear on the threshold of a most suggestive line of inquiry. 



The mean non-cyclic change given by Table XXIII for the Horizontal Force, - 0'53y, would give in a 

 year of average days an apparent decrease of 194y, whereas Table VI gives a secular increase of 150y. 

 No necessary inconsistency, however, exists between these figures, because the difference between them 

 may mean nothing beyond a diminution in the magnetic moment of the Horizontal-Force magnet, which 

 would have exactly the same effect, so far as the non-cyclic change is concerned, as a real decrease of 

 Horizontal Force. A fall of 4-3 per cent, per annum in the moment of the Horizontal-Force magnet would 

 suffice to wholly account for the non-cyclic change observed. 



In February, 1903, the diurnal inequality of Vertical Force was based on only eight "days " or periods 

 of 24 successive hours, and only three of these commenced at local midnight. Thus no value is assigned 

 in that month to the non-cyclic change of Vertical Force. It was, as a matter of fact, large and negative 

 on each of the three actual days, the average amount being - 38y. Excluding February, 1903, there were 

 18 months possessing a sufficient number of complete days to be worth taking into account. The results 

 for individual months vary much, but have a decidedly negative tendency, the mean from the 18 months 

 being - 5'35y. A real change of 5 - 35y per diem in V would mean a change of '01954 per annum. An 

 apparent decrease of 5 35y per diem would, however, be equally well accounted for by a fall of about 

 2 7 per cent, per annum in the moment of the Vertical-Force magnet. This is really a smaller change 

 than that required to account for the non-cyclic change observed in the case of the Horizontal Force, 

 and it seems of by no means an improbable magnitude. 



26. The inequalities for the 12 months, the year, and the three seasons are also shown graphically in 

 figs. 1 to 13, pp. 104 to 116. 



In the case of individual months, especially in Summer, the irregularities in the hourly values arising 

 from disturbances catch the eye whether one looks at the diagrams or the tables. They are, however, much 

 less prominent than one would have anticipated from a survey of the magnetic curves or of the hourly 

 tabulations, and, when one comes to the mean seasonal results, the smoothness is not a little remarkable. 



Considering the Declination inequalities in Tables XII to XV more closely, it will be seen that, so far 

 as the general type is concerned, the differences between all and quiet days, or between the different seasons 

 of the year, are comparatively small. There is a single daily oscillation, the extreme positions being 

 reached about 9 a.m. and 6.30 p.m., the whole year round. ., 



If in the figure NS represent the geographical Meridian, N being to 

 the north, and EW the east-west line, then if ns represent the 

 direction of the Declination needle, the angle Nora was about 152 40' on 

 the average. It was largest about 9 a.m., and least about 6.30 p.m. 



Though the type of the diurnal inequality did not vary much with w 



the season, the amplitude varied largely, being greatest in Midsummer 

 (about December), and least in Midwinter (about June). Considering, 

 however, that the sun was continually below the horizon in Midwinter, 

 the wonder is not so much that the range at that season was markedly 

 less than at Midsummer as that it attained the size it actually did. 



Comparing Tables XII and XIV, it will be seen that with the exception of January, 1903, all months 



\ 



