EARTH'S MAGNETISM TO THE SOLAR AND LUNAR PERIODS. 103 



eludes from the Toronto observations for 1842 (corrected by the usual method), that 

 there is a maximum in June, and a minimum in December. I have projected 

 the monthly means of the Toronto obsei-vations as corrected by the usual method. 

 From these Colonel Sabine draws his conclusion. Under it I have projected the 

 temperatm-e of the magnet in a broken line, and below both, the means from the 

 two-hourly observations, as corrected approximately by myself. (See Curves, 

 No. 5, Plate IV.) These will shew how much depends on the accm-acy of the cor- 

 rection in arriving at sound conclusions. I conceive that the consistency of the 

 results at which I have arrived, independently of other considerations, will leave 

 little doubt as to which method of obtaining the corrections should be adopted. 



12. It has been already mentioned (10), that the apparent secular change con- 

 sists of a considerable increase of horizontal intensity. Throughout the whole period, 

 the rapidity of increase has been diminishing, and it is much less in 1845 than in 

 any of the previous years. Of all the puzzling problems in terrestrial magnetism, 

 that of connecting the secular change with some known or observed phenomenon 

 has been the most difficult ; any fact, therefore, tending to this, wUl have interest. 

 One of the first questions which I proposed to myself, connected with it, was 

 whether all hours of the day were equally affected by the secular change? In order 

 to answer this more distinctly, the annual period was eliminated from the monthly 

 means, or, which is nearly the same thing, the mean of each month was reduced to 

 the straight line passing through January and December 1844.* I then found 

 that the mean horizontal force in the first six months of the year 1844, was almost 

 constant one hour after the period of the morning maximum, and also that it was 

 almost constant for the last six months, one hour before the period of the even- 

 ing maximum. When the diurnal curve for each month was projected, I found 

 the curves for the first six months to pass through a space of 3-4ths of a scale 

 division in the ordinate of Q^ 40"" a.m., with the exception of the cm-ve for Feb- 

 ruary, which is very iiTegular there. The curves for the last six mouths pass 

 through a space of 1^ scale divisions, in the ordinate of 4'' 40"" p.m. ; the increase 

 of horizontal force from January till December was 18 scale divisions. I next elimi- 

 nated all the larger disturbances from the monthly means of each hoiu", but this 

 neither aflfected the periods of the nodes, nor the values of the ordinates in which 

 they were contained. In this way, then, the horizontal force in its secular pro- 

 gress, seems to rest one foot dmlng the fii'st half of the year about an hour after 



* The line should have been drawn through January 1844 and January 1845, but that there is an 

 irregularity in the progress of the horizontal force from December 1844 to January 1845, comjjared 

 with the previous years. I have, however, also reduced the means to the line passing through Janu- 

 ary 1844 and January 1845, and find the ordinate of the morning node slightly increased, but that 

 for the evening node diminished. 



VOL. XVI. PART II. 2 D 



