338 



The correspondence of the two results gives a great weight to the 

 accuracy of both ; this will be more evident when it is remembered, 

 that the whole motion of the dipping-needle for the moon's varying 

 declination is included by a small circle with a diameter of little 

 more than one-tenth of a minute of space, and, that no observation 

 in the sixty thousand employed for this result has been rejected, 

 however greatly affected by disturbance ; although the graphic in- 

 terpolation to remove slight irregularities may be considered as an 

 equivalent operation. 



Diurnal Motions. — The monthly mean diurnal variations for the 

 magnetic declination and magnetic dip, from four years' observations, 

 still present irregularities, especially from 10*^ p.m. till 4^ A.M., the 

 hourly positions for this time depending on only two years' observa- 

 tions. For this reason, the values from the Tables having been pro- 

 jected, curves were passed freely among the points, and the interpo- 

 lated ordinates thus formed, were taken for the projections in Plate 

 VII. : the interpolated quantities differ very little from the actual 

 values, and this is especially the case for the summer months. 



The diurnal motions for the four winter months , November to 

 February, are of the same class, and they differ considerably from 

 those for the other months (see Plate VII.) : in each of these months 

 the motion consists of a figure of two closed loops : the north end of the 

 needle moves eastwards with little change of dip from about l** p.m. 

 till 9^ or 10'' P.M., after which it turns westwards, and begins to 

 ascend about 4'' a.m., crossing near its position at G*" p.m. ; thus 

 forming an eastern loop, which is small compared with the western 

 loop, excepting in December. After G^ a.m., the north end of the 

 needle having moved a little westwards, again descends, crossing a 

 second time the afternoon track near b^ p.m. ; still moving westwards, 

 it ascends about 11*^ a.m. till it meets the position of 1^ p.m., thus 

 completing the western loop. The eastern loop is not formed in 

 March, the north end of the needle not rising sufficiently high to 

 cross the afternoon track. The change in the figure from February 

 to March is very great ; in April and May the remains of the east- 

 ern loop are still visible, but in June and July its position is indi- 

 cated by a simple inflection in the figure ; in August and September 

 the germ of the eastern loop becomes more distinct, and in October 

 the loop is actually formed. The transition in form from autumn to 

 winter is quite gradual, unlike that from winter to spring. In the 



