September 6, 1900] 



NATURE 



46: 



liicir much more energetic character. In order to extend the 

 record as far back as possible, a similar curve (Fig. i, B.^) was 

 constructed, based on Airy's statistics from 1841-1857 {Trans. 

 Roy. Soc. , vol. cliii.). 



These magnetic curves are in perfect harmony with those of 

 the auront, and both sets bring out a remarkable and, for our 

 purpose, eminently important fact, namely, that although for the 

 most part they correspond closely with the eleven-years spot 

 period, there nevertheless appear waves which at first sight seem 

 to have nothing to do with the display of solar activity. Most 

 noticeable in this respect are the maxima of 1852 and 1864, of 

 which the former especially seems almost to contradict the 

 existence of a connection between sun-spots and aurorte. A 

 closer examination of the spot-curves, how- 

 ever, reveals the fact that in these two 

 instances there occurred, at exactly the 

 same times, peculiar disturbances or irre- 

 gularities in the exhibition of spots on the 

 solar surface, so that, after all, the waves 

 exhibited at these points in the magnetic 

 and auroral curves may be supposed to 

 have been caused by solar influence, although 

 they appear for some reason or other on a 

 greatly exaggerated scale. This is decidedly 

 the opinion of Prof. Fritz and of Mr. Ellis, 

 both of whom expressly mention instances 

 of this character, the latest having occurred 

 in 1898 following the spot-maximum of 

 1894. 



The succeeding part of this research may 

 be conveniently divided into two parts. 

 In the first, the earth-magnetic variations 

 are compared and contrasted, in relation to 

 the eleven-years period of sun-spots, with 

 certain phenomena associated more or less 

 directly with latitude-variation ; and in the 

 second with similar phenomena relative to 

 the "great" sun-spot and aurora period. 



These phenomena are : — 



I. With regard to the eleven-years 

 period — 



i. The semi-amplitudes of latitude-va ri- 

 al ion. 



ii. The periods of latitude-variation. 



iii. The changes in the values of the 

 obliquity as observed at Greenwich from 

 18 1 2- 1 896. 



iv. Mr. Thackeray's corrections to the 

 right-ascensions of stars derived from ob- 

 servations of the sun during the years 

 I 836- I 895. 



II. With regard to the '* great " period — 

 v. Dr. Chandler's long period inequality 



in the latitude-variation. 



vi. The observed residuals of the obli- 

 quity from 1 753-1896 as compared with 

 Leverrier's tabular values. 



vii. The Greenwich corrections to the 

 sun's right-ascensions relative to a funda- 

 mental system of fixed stars, according to 

 Prof. Newcomb. 



The data requisite for the formation of 

 the curve of semiamplitudes (Fig. i, curve 

 I ) was taken from Dr. Chandler's paper in 

 A.J. No. 277, and from the publications of 

 Dr. Nyren {A.N. 3166), Mr. Wanach 

 ' l.N. 3ll2)andProf. Albrecht ("Berichte 

 her den Stand der Erforschung der Breitenvariation "). It has 

 I be borne in mind, as pointed out in my previous papers, 

 iiat the latitude phenomenon lags behind the comparison 

 magnetic curves by about 1-5 years. Such a lag appears 

 mdeed to be a characteristic feature common to most terres- 

 trial phenomena which have hitherto been found to be in- 

 fluenced by solar activity. In this connection I must refer to 

 the highly important and interesting discovery made by Sir 

 Norman Lockyer some years ago, that the curve representing 

 the frequency of the "unknown" lines widened in sun-spot 

 spectra during a cycle of solar activity follows the spot-curve 

 l)y exactly the same interval, viz. 1-5 years. I have received a 



NO. J 6 10, VOL 



communication from Sir Norman, stating that this curve of 

 " unknown " lines goes excellently with my curve I of Fig. i. 

 the maxima and minima of the spectroscopic curves showing a 

 perfect synchronism with those of the curve of latitude-variation. 

 I take this opportunity of expressing my indebtedness to Sir 

 Norman Lockyer for drawing my attention to this most signifi- 

 cant and singular fact. 



The lag in the case of curve I has been allowed for by 

 shifting this curve one and a half years in the backward direction. 

 When it has thus been made to coincide with the comparison 

 curves, their correspondence becomes most striking — a remark- 

 able feature being the exactness with which certain secondary 

 maxima in 1865, 1887 :ind 1S98 are represented in each case. 



FLg.n. 



(From -Ast-HoLch:) 

 ' 3e*9. ' 



turning points of Dr. Chandler's long period ii 

 aurorae (Loomis). (The ordinates represent 

 numhers.) B, great period of sun-spots (WolO 



quality of latitude-variation ; 

 means of Loor 



eriod if 

 aurora- 



twelve-years' means of residuals of obliquity 

 (Greenwich observations) after elimination of purely secular change ; 2, twelve-years' means of 



corrections to the sun's R 

 elimination of purely 

 curve representing W( 



\. relative to a system of fixed stars (Greenwich observations) after 

 secular change ; 3, combined curve of obliquity and sun's R.A., the dotted 



olf s great period of sun-spots on a somewhat different scale 



Indeed, so accurately are the motions of the amplitude-curve 

 shadowed in those of the magnetic variations, that to any un- 

 prejudiced astronomer this tact in itself must indicate with 

 sufficient force the existence of a vera causa. 



The periods of latitude-variation afford a new and almost as 

 certain proof of the existence of an intimate connection between 

 the polar motion and earth-magnetic phenomena. 



Dr. Chandler has already shown that a small amplitude 

 corresponds to a great period and vice versA {A.J. 277, p. 98). 

 How well this influence is borne out may be seen from the 

 following analysis of the average values of r according to the 

 lengths of the period : — 



6.] 



