234 Royal Society. 



It appears further that the average value has a similar law of varia- 

 tion to that of the number and aggregate value. 



The second classification shows that at Hobarton the contrast both 

 in frequency and aggregate value is still between the hours of the 

 day and those of the night, the ratios being, however, in this case 

 greater than unity during the former hours, and less than unity during 

 the latter, contrary to what takes place with the easterly disturb- 

 ances : at Toronto the contrast is between the hours from noon to 

 midnight, and those from midnight to noon, the ratios being greater 

 than unity during the latter hours, and less than unity during the 

 former. In both cases the variation in the ratios appears to be de- 

 pendent on the hours of local, not on those of absolute time. 



From a table showing the ratios of easterly aggregate values to 

 westerly at Toronto, and of westerly to easterly at Hobarton, it ap- 

 pears that, at both stations, the deflection (due to disturbance) of 

 the end of the magnet of the same name as the magnetic latitude 

 is to the west during the hours of the day or from 5 a.m. to 

 5 P.M. : at a little before 6 p.m. at Toronto, and a little after 6 

 at Hobarton, the deflections pass through zei'o (or the undisturbed 

 position of the magnet) into easterly deflections of that end. The 

 magnitude of those deflections rapidly augments to a maximum at 

 9 P.M. at Toronto, and at 10 p.m. at Hobarton; they again pass 

 through zero between 4 and 5 a.m. ; and attain the westerly maxi- 

 mum at 7 A.M., the variation in the magnetic direction due to the 

 disturbances depending, like those of number and value, ou the hour 

 of local time. 



II. Inequality or variation in the number and aggregate values of 

 the disturbed observations in different months. From the tables 

 which are given, it is obvious that tliere is a systematic variation 

 in the numbers and aggregate values of the disturbances in the dif- 

 ferent months ; and at both stations the easterly and westerly ratios, 

 separately considered, diff'er little in the characters which they assign 

 to the variation, from the ratios of the two combined. The most 

 distinctly marked feature is that the disturbances are less frequent 

 and have a less aggregate value in November to Februarjr at To- 

 ronto, and in May to August at Hobarton, than in the other months 

 respectively : so that the disturbances are governed by a law de- 

 pending either on the period of the year, or on local season, not on 

 absolute time. 



III. Variation in the number and aggregate values of the dis- 

 turbed observations in different years. Taking the ratios of the 

 numbers and aggregate values of the disturbed observations at 

 Toronto and Hobarton in the different years (from 1843 to 1848), 

 to the average annual number and aggregate value respectively, it 

 appears that there is a remarkable correspondence in the variation of 

 these ratios in difterent years at the two stations ; and that at each, 

 both latios increase progressively from 1843 to 1848, with the single 

 exce])tion of 1845, in which there is a small diminution in that of 

 the number and also that of the value. Taking the mean of the 

 ratios at Toronto and Hobarton, the ratio of the number increases 



