232 CHRISTCHURCH TERM-DAY OBSERVATIONS. 



ratio borne by the range in Midwinter to that in Midsummer is considerably less than at Kew. The only 

 seasonal variation of type that appears at all marked is that the extreme westerly position is reached 

 earlier in Summer than in Winter. 



Horizontal-Force changes, as already stated, were less regular than those of Declination, thus the 

 inequality figures for the former element in Table IA are less smooth than those for the latter. Still, 

 they suffice to bring out the general character clearly. As with 1), there is a distinct double daily period, 

 even in Midwinter. The principal minimum is, as at Kew, more conspicuous than the maximum. This 

 minimum occurs near noon later in the day than at Kew and it seems some two hours later in 

 Midwinter than in Midsummer. There is a morning maximum some 5 hours before the principal 

 minimum, and a second maximum about 5 p.m., but while the former maximum considerably predominates 

 in Midwinter, the latter is the larger in Midsummer. 



The variation in the hour of occurrence of the principal minimum reduces the range in the diurnal 

 inequality for the year, which falls appreciably short of that for the equinoctial months. The difference 

 between the ranges during Midsummer and Midwinter, though well marked, is less conspicuous than in the 

 case of the Declination. 



This brief summary will suffice to show the main features of the diurnal variation. 



7. Looking at inequalities such as those in Table IA, and still more so when looking at inequalities 

 based on a very large number of days, one sees a magnetic element rising persistently to a maximum, and 

 falling as steadily to a minimum. But if one takes the more familiar case of a tide in the sea, one knows 

 that while the combination of the observations of a large number of days produces the semblance of 

 continuous rise for hours, followed by continuous fall, on any individual day either rise or fall consists of 

 numerous ebbs and flows in the height, dependent partly on the waves. The number and amplitude of 

 these subsidiary movements may, for all I know, differ widely even on the same day at two spots on the 

 same coast no great distance apart. Very probably the phenomena on any given day are largely 

 dependent on the wind and on the distance of storm centres. Still, I am disposed to think it is probable 

 that the minor phenomena in the rise and fall of the tides in different oceans, or different parts of the 

 same large ocean, possess characteristics of a distinctive character, though these may not be apparent in tidal 

 data calculated in the usual way. Now it is probably the same with the magnetic diurnal variation, with 

 this difference that purely local features have little if any effect. These considerations encouraged the 

 hope that a systematic study of the changes shown by the Christchurch data might disclose phenomena 

 whose comparison with corresponding phenomena at other stations might some day throw a valued light 

 on the nature of the diurnal variation. If similar data should be got out on a future occasion for 

 Christchurch from a year of large sun-spot frequency the comparison could hardly fail to be instructive. 



8. The magnetic changes recorded by ordinary magnetographs rarely if ever occur quite suddenly. 

 Even with slow-run curves the movements usually appear continuous. The fact that the value of H is 

 given on March 1 as '22672 at Oh. Om. and as '22671 at Oh. 3m. does not mean that the value was exactly 

 22672 at Oh. Om., Oh. 1m. and Oh. 2m., but exactly -22671 at Oh. 3m. Even supposing the photographic 

 trace perfection and the tabulator unerring, it may mean a number of different things. The values at 

 four successive minutes might for instance be 



22672 3 , -22672 , -22671 7 , -22671 4 , 

 or 



22672 0) 22671 8 , -22671 , -22671 4 , 



and there might have been numerous oscillations in value too small to be detected. 



The conclusions one would draw as to the number of changes of force occurring in a given time would 

 certainly depend somewhat on the sensitiveness of the instrument, even when one required a given 

 minimum amplitude to count as a change. The number of changes recorded would inevitably vary with 

 the size of the minimum accepted and with the interval between successive readings. On the other hand, 

 I do not think the number of changes in Declination or Horizontal Force (I cannot say the same of 

 Vertical Force) would depend much on the type of the magnetographs (if reasonable damping exists) as 

 distinct from their sensitiveness. In the absence of artificial sources of disturbance, such as persistent air 



