624 [June 18, 



subsequent investigation of the secular, periodical, and other minor 

 magnetic variations, the process of elimination introduced by the 

 author and employed by him for several years past in the reduction 

 of the colonial observations has the advantage of separating from the 

 whole body of the observations a far greater portion of the disturb- 

 ing influence than would be gained by the simple omission of the 

 observations on the 95 days. The laws of the disturbance-diurnal 

 variation, thus found to be approximately the same whether obtained 

 from the narrower or from the wider basis of investigation, are then 

 stated, and are compared with the results of similar investigations 

 recorded in the author's previous publications the points of accord- 

 ance or of difference being severally discussed in the third section. 



The fourth section contains Tables of the " Diurnal Inequality," and 

 of the " Solar-diurnal Variation " at Kew, showing the mean values 

 at each hour and in each month. The "Diurnal Inequality " is ex- 

 plained as consisting of two principal constituents, viz. the " Disturb- 

 ance-diurnal Variation," and the " Solar-diurnal Variation." It is 

 obtained for each month by taking the differences between the mean 

 positions of the magnet at each of the 24 hours, in the month, and 

 the mean position in the month itself (viz. the mean of all the days 

 and all the hours) no omission whatsoever being made of disturbed 

 observations. 



The " Solar- diurnal Variation" is obtained by a similar process, 

 after the separation and omission of all the observations which 

 differed by a certain small and constant value from the normals of 

 the same month and hour. By this process the effects of the 

 " Casual and Transitory changes " are in a very great degree elimi- 

 nated, and a very close approximation is obtained to the systematic 

 diurnal action of the sun upon the direction of the horizontal mag- 

 net, apart from the effects of disturbances. The solar-diurnal varia- 

 tion thus obtained at Kew is compared with results similarly obtained 

 at six other stations, viz. three stations in the interior of the two 

 great northern continents, one equatorial station, and two stations in 

 the middle latitudes of the southern hemisphere thus generalizing 

 upon a very extensive scale the action of the sun in producing the 

 phenomena under notice. 



The fifth section is occupied by a similar generalization of the facts 

 which have placed in evidence the existence of a semiannual inequality 



