1885.] Terrestrial Magnetic Force in the Horizontal Plane. 257 



as " magnetic inequality " of that ordinate for that hour ; producing 

 twelve measures of " inequality " for each day. When the photo- 

 graphic system was introduced, the elevation of a pencil curve drawn 

 by eye so as to smooth down the irregularities of the photographic 

 trace above a photographic base was measured for every hour, 

 producing twenty-four measures of "inequality." 



In the instances of excessive and rapid disturbances of the magnets 

 during magnetic storms, no measures of ordinates were taken for the 

 present purposes. 



Thus the daily measures at each hour or two hours were obtained. 



The next step was to collect for each month all the daily measures 

 on corresponding hours through each month, and to take their mean. 

 These are the measures for the hours which are actually treated in the 

 present memoir. By combining (for each month) the inequality of 

 magnetic horizontal force at every two hours or each hour, as 

 abscissa, with the inequality of magnetic decimation (on the same 

 scale of measure) at the same two hours or hour, as ordinate, points 

 were defined in every monthly curve representing completely the 

 mean diurnal changes of magnetism for each month. On the recom- 

 mendation of the Board of Visitors of the Royal Observatory, reduced 

 photographic copies of these curves were prepared by the Astronomer 

 Royal for publication with the volume of Greenwich Observations for 

 1884. 



The number, and the character, of the curves produced uninter- 

 ruptedly on this plan, and the circumstance that they are intended 

 for publication in the Greenwich Observations, appear to render 

 them unfit for dissemination in the Royal Society's Transactions. I 

 have, therefore, decided on the following course. With the permis- 

 sion of the Astronomer Royal, I have adopted the three years 1863, 

 1864, 1865, for partial exhibition of results. (Any other years would 

 have answered equally well, for general exhibition.) For each of 

 these years I have attached to this paper the curves for the months 

 January, April, July, October, which suffice for showing generally the 

 characteristic changes of magnetism for the several months. But 

 some general account may be given, for which this is perhaps a 

 suitable place. 



The form of the curves, and the position of the points on them 

 corresponding to hours of solar time, leave no doubt that the diurnal 

 inequality is due mainly — and, as far as I can judge, entirely — to the 

 radiant heat of the sun ; and, it would seem, not to the sun's heat on 

 the earth generally, but to its heat on parts of the earth not very 

 distant from the magnets. In the hot months of the year, the curve, 

 though far from circular, surrounds the central point in a form which, 

 as viewed from that central point, never crosses itself ; and is, 

 roughly speaking, usually symmetrical with regard to B. and W. 



S 2 



