MAGNETIC DISTURBANCES AT ItEW. 291 



15. Diurnal inequalities in I were calculated for the individual months of the year, 

 but Table IX. contains only the seasonal results. Even these leave something to be 

 desired as regards smoothness. In the mean diurnal inequality for the year in I the 

 principal maximum occurs at 10 a.m. as in the case of quiet days, but the principal 

 minimum is found at 5 a.m., instead of 7 p.m. as on quiet days. The range in the 

 mean inequality for the year bears to the corresponding range from quiet days the 

 ratio 1'82 : 1, which is a little less than the corresponding values for D and H. 



In all the elements disturbance is thus associated with a marked increase in the 

 range of the diurnal inequality. The same phenomenon was described recently* at 

 the "Discovery's" Winter Quarters in latitude 77 51' S., where the omission of 

 47 per cent, of the days, retaining only those of lesser disturbance, reduced the range 

 of the mean diurnal inequality for the year in D by about 25 per cent. Some stations 

 are much more disturbed than others, thus diurnal inequalities got out from any 

 choice of days answer to different amounts of disturbance at different stations. The 

 irregular movements presented by the average day's curves in high latitudes are such 

 that their appearance in a Kew curve would inevitably lead to its inclusion amongst 

 the highly disturbed. It is thus a question what is the real significance of the very 

 large amplitude seen in the diurnal inequality at such a highly disturbed station as 

 the " Discovery's " Winter Quarters. What would the amplitude of the inequality be 

 in such a case if we had a large number of year's data at our disposal and retained 

 only those for the quietest day of each month ? 



16. If all days were quiet and alike, and if the daily maximum and minimum 

 happened to occur at exact hours G.M.T., then the inequality range would equal the 

 absolute range. The interval of time between the daily maximum or minimum and 

 the nearest hour cannot exceed 30 minutes. If the diurnal inequality could be 

 represented by a simple wave of 24-hour period, 30 minutes of time would answer to 

 a change of 7'5 in the phase angle, and the largest hourly value could not be less 

 than 0'99 of the true maximum. 



Most diurnal inequalities are of a considerably more composite character, but sharp 

 peaked forms are rare, thus non-coincidence in time of either maximum or minimum 

 with an exact hour is unlikely to reduce the range derived from hourly readings by 

 more than 3 or 4 per cent. Even on the quietest day, however, magnetic curves 

 exhibit small irregularities, and the times of maximum and minimum vary from day 

 to day. Thus the absolute range presents a limit which the inequality range 

 may theoretically attain to, but which it is practically certain to fall considerably 

 short of. 



Passing from quiet to disturbed days, we may expect a large reduction in the ratio 

 borne by the inequality range to the mean absolute range, and this for several reasons. 

 On highly disturbed days sharp peaks are not uncommon, and the maximum and 



* National Antarctic Expedition, 1901-1904, "Magnetic Observations," p. 102. 



2 P 2 



