ON COMPAKISON AND REDUCTION OF MAGNETIC OBSERVATIONS. 221 



2 P.M., the time of the day when the diurnal inequality is most in 

 evidence, the value of ^ seems to vary very little with the season. 



Curves whose radius vector is p, and vectorial angle 4', illustrate the 

 variations of diurnal inequality in a conspicuous way. Numerous curves 

 of this kind, illustrating the results obtained at Greenwich for the years 

 1841-57, were given by Sir G. B. Airy in the 'Phil. Trans.' for 1863, 

 and a smaller number, refeiTing apparently to the years 1873-7.5, were also 

 given by Airy in the 'Phil. Trans.' for 1885. Several applications of 

 these curves to Dublin results appear in Dr. Lloyd's ' Treatise.' Following 

 these examples I have drawn curves 11, 12, and 1.3, Plate VI., based on 

 the results in Table VII., for midwinter, midsummer, and the whole year. 

 The hours are stated beside the points on the curves to which they refer. 



The prominent loop on the midwinter curve should be noticed. The 

 existence of a loop on the mean winter curve and its absence in the mean 

 summer curve for Dublin are referred to by Lloyd {loc. cit., p. 187). The 

 tendency to loops in winter months is conspicuous in Airy's Greenwich 

 curves. 



Relative Intensity of the Forces to ichich the Diurnal Inequality is due 



throughout the year. 



§ 13. Without knowing the true nature of the disturbing force or 

 forces to which the diurnal inequality is due, it is difficult to suggest any 

 wholly satisfactory measure of intensity. Assuming the inequality to be 

 a composite phenomenon, the counteraction of opposing forces may produce 

 the same effect at one hour as the absence of forces at another. All the 

 phenomena seem, however, to point to greatly increased activity of dis- 

 turbins: forces in summer. The data s;iven above as to the mean values 

 of p throughout the day at midwinter and midsummer make the former 

 mean only -374 of the latter. Evidence in the same direction is supplied 

 by the following table VIII., which gives for each month, each quarter and 

 half-year, and for the whole year, the sum of the hourly departures from 

 the mean for the day, irrespective of sign, along with the range, or algebrai- 

 cal difference between the extremes. To show how comparatively little 

 depends on how the non-cyclic element is dealt with, the table, in addition 

 to the results obtained after the non-cyclic element has been eliminated, 

 gives likewise the range deduced from the uncorrected data for each 

 month. The fractions, which the table records, refer exclusively to the 

 corrected data. 



The range and the sum of the inequalities in both declination and 

 horizontal force show a distinct minimum in December. In the declina- 

 tion both the range and the sum of the inequalities present an absolute 

 maximum in August, with apparently a second, only slightly smaller 

 maximum, in May. The variation, however, especially of the range, is so 

 small from April to August that it would hardly be safe to conclude this 

 was the normal phenomenon in a year of ' quiet ' days. In the horizontal 

 force there would appear to be a single maximum, whether for range or 

 sum of inequalities, in July ; but the difference between the results and 

 those for May, June, and August is extremely small. 



§ 14. These conclusions are only partly in accord with those got out by 

 Dr. Balfour Stewart • for a long series of years, 1858-73, at Kew, and by 

 Mr. W. EUis,^ for a still longer series of years, 1841-77, at Greenwich. 



' Prnc. Roy. Soc, vol. xxvi. 1877, p. 103. 

 2 Phil. Trans, for 1880, p. 544. 



