70 Mr. W. Ellis. Relation between Diurnal Range of 



Order of Epochs with reference to intensity of Sun-spot and 

 Magnetic Effects. 



In each curve the greatest and least maxima are those of 1870 and 

 1884 respectively, and the order in each case is similar excepting 

 for the epochs of 1848 and 1860, which in horizontal force are 

 reversed in position, although otherwise falling in with the general 

 order of fche other curves. The horizontal force observations have to 

 be corrected for the effect of temperature (a correction not required 

 for decimation), and in the years 1848 and 1860 the magnets were in 

 the original upper magnet room, in which the diurnal range of tem- 

 perature was considerable, rendering difficult the determination of 

 the actual temperature of the magnet (a bar 2 feet in length), and 

 this may possibly account for the apparent displacement of the 1848 

 and 1860 epochs of maximum in horizontal force. In 1864 an under- 

 ground basement was specially constructed for the magnets, in 

 which the variation of temperature is small, and in this apartment 

 the magnets have since remained. As regards the minimum points 

 of the curves, the sun-spots at the epochs of 1856, 1867, 1879, and 

 1890 practically disappear, but not so at that of 1843, at which epoch 

 the most elevated of the magnetic minima, both in declination and 

 horizontal force, occurs, being so far in harmony with what has been 

 pointed out as to the behaviour at maximum. 



Considering that the irregularities in the length of the sun-spot 

 period so entirely synchronise with similar irregularities in the 

 magnetic period, and also that the elevation or depression of the 

 maximum points of the sun-spot curve is accompanied by similar 

 elevations and depressions in the two magnetic curves, it would 

 seem, in the face of such evidence, that the supposition that such 

 agreement is probably only accidental coincidence can scarcely be 

 maintained, and there would appear to be no escape from the conclu- 

 sion that such close correspondence, both in period and activity, 

 indicates a more or less direct relation between the two phenomena, 

 or otherwise the existence of some common cause producing both. 

 The sharp rise from minimum epoch to maximum epoch, and the 

 more gradual fall from maximum epoch to minimum epoch, may be 

 pointed out as a characteristic of all three curves. The similarity of 



