38 On the Diurnal Variations of Magnetic Elements. 



Smith*. The Greenwich declination-curve is less bold than 

 that of Messrs. Robson and Smith, but in other respects 

 shows great similarity, both in form and in the position of 

 the nodal points. The deficiency in boldness may be a pecu- 

 liarity of the year under discussion, the epoch being one of 

 minimum magnetic activity. 



The periodical difference shown by Messrs. Robson and 

 Smith to exist between the Greenwich and Kew diurnal 

 inequality of declination, as deduced at Greenwich from all 

 days and at Kew from quiet days, thus appears to depend on 

 difference in the character of the records employed ; since the 

 discussion of Greenwich records alone, on the two systems 

 mentioned, leads to the same conclusion, not only as regards 

 declination, but also for the horizontal and vertical forces. 

 A difference of the same general character is shown in each 

 element in all months of the year : these particulars could 

 not well be given here, but it has been thought interesting 

 to add in the Table the summer and winter effects, taking 

 summer as including the months from April to September, 

 and winter the remaining months of the year. 



The records used in the formation of the Greenwich all days' 

 results contain many irregularities which are smoothed in the 

 way before mentioned. The quiet days, in 1889, contain 

 practically no irregularities. It might thus be conjectured 

 that something may be due to systematic bias in the perform- 

 ance of the smoothing operation ; but it could, if necessary, 

 be shown that no influence of this kind would produce differ- 

 ences so dissimilar in character for the different elements ; 

 besides which irregularities in the vertical- force record are so 

 infrequent, and in the year treated so few and trifling, that 

 practically no smoothing had to be done ; but yet the curve 

 for vertical force, as may be seen, is quite as pronounced as 

 those for either of the other elements, and indeed forms a very 

 much larger proportion of the whole diurnal inequality for 

 vertical force (see the Table) than is the case for either of the 

 other elements. 



The difference between the two methods of determining 

 diurnal inequality thus appears to arise from a physical cause, 

 the form of the diurnal inequality-curve found by including 

 all days differing in a systematic way from that given by 

 using selected quiet days, in the manner shown in the diagram. 

 The nodal points in the horizontal-force and vertical-force 



* The vertical divisions on the diagram correspond each to '00025 

 metric or '000025 C.G.S. measure, equivalent on the declination-curve to 

 0'-47 in arc, whilst those on Messrs. Robson and Smith's diagram represent 

 each O'-oO. 



