626 [June 18, 



columns which severally present the weekly values in the years from 

 1858 to 6182 inclusive, the mean value of the secular change corre- 

 sponding to the period comprised in the Table is deduced. A pro- 

 portional part of the secular change is then applied with its appro- 

 priate sign to each of the weekly values in the mean or typical year. 

 These should all correspond with the mean declination of the whole 

 Table (viz. 21 39' 18" -1), or should exhibit only such small and 

 unsystematic differences as might reasonably be ascribed to casual 

 errors. The final column of the Table contains these differences, 

 in which it is at once seen that they divide themselves into two dis- 

 tinct categories, distinguished by the sign in the semiannual period 

 from March 21 to Sept. 21, and by the + sign from Sept. 21 to 

 March 21. Hence the author infers the existence of a variation in 

 the declination at Kew having an annual period, and consisting of 

 a semiannual inequality with epochs coincident, or nearly so, with the 

 sun's passage of the equator the magnet being deflected towards the 

 east when the sun is north, and towards the west when he is south of 

 the equator. The amount of the semiannual inequality, as shown by 

 the Table, averages 28"- 95 in the weeks from March 21 to Sept. 

 21, and + 29"'9 in those from September to March. The whole 

 amount of the annual variation at Kew is therefore 58"*85. 



The result thus obtained from the observations at Kew is com- 

 pared with the result of an investigation of the corresponding phe- 

 nomena at Hobarton in the southern hemisphere, obtained from 

 hourly observatious of the declination during five years, commencing 

 in October 1843, and terminating in September 1848. The obser- 

 vations themselves are published in tfye 2nd and 3rd volumes of the 

 Hobarton Observations, and are treated, for the purposes of this 

 paper, precisely in the same way as those of the Kew Observatory, 

 forming a table strictly analogous to the one, previously described, 

 at Kew. The final column of the Hobarton Table exhibits the differ- 

 ences, in each of the 52 weeks of the typical year, from the mean 

 declination derived from the whole of the observations in the five 

 years. The + and signs in this column attest in as striking a 

 manner as do those at Kew, the existence at Hobarton of a semi- 

 annual inequality of which the epochs coincide, or very nearly so, 

 with the sun's passage of the equator : the direction of the deflec- 

 tion is the same as at Kew, viz., of the north end of the magnet 



