50 



REPORT — 1861. 



holds for an increase of force. The earth therefore appears to act as a whole, as a 

 great magnet, the increase or diminution being in proportion to the iutensity at the 

 given point. This fact is wholly opposed to an explanation which would attribute 

 an increase or diminution of force to a purely local phenomenon, such as the direc- 

 tion of the wind, as may easily be shown ; for if, while relating the du-ection of 

 the wind at Makerstoun to the horizontal force at the same place, we also relate it 

 to the horizontal force at some other place where the direction of the wind is 

 known to be very different, and if the same, or nearly the same, result is obtained 

 for the horizontal force at both places, we may be satisfied that the result, tchatever 

 it may he, is unconnected with the direction of the wind. For this end I have 

 chosen as a second station Singapore, nearly on the equator (1° W/ N. lat., &" 45™ 

 long, east of Greenwich). 



In a discussion of this kind, where the results obtained by others are disputed, it is 

 necessary to state distinctly the methods employed : this I shall now do. The hourly 

 observations of the bifilar magnetometers at Makeratouu and Singapore for 1844 

 having been connected for temperature *, the monthly mean con-esponding to each 

 day in the year (that is, having that day for its middle point) was obtained for each 

 place : this monthly mean includes the annual and secular change corresponding 

 to the given day ; and wten it has been compared with the coiresponding daily 

 mean, the difference (+ if the daily mean were the greater, — if the lesser) vdll 

 depend upon other causes. These differences were obtained for each day of 1844 

 on which observations were made. The approximate mean direction and mean 

 pressure of the wind (in pounds on the square foot of surface) at Makerstoun were 

 also obtained for each day of the year. In order to render the results comparable 

 with those obtained by Father Secchi, the winds were included in the four heads, 

 South, East, North, and West ; the days of intennediate directions (as N.W.) 

 being entered under the two principal heads (as north and west) with half weights 

 only. For purposes of comparison the winds were separated into frv\'o classes — that of 

 weak winds (daily mean pressure less than | of a poimd), and that of strong winds 

 (daily mean pressure I of a poimd and upwards). 



The follovdng are the results of this discussion ; and that a comparison may be 

 made at once with those of the Roman Observatory, I shall fh-st give the number 

 of days for which the horizontal force Was greater or less than the mean for each 

 of the four winds. 



Direction 



of Wind, 



Makerstoun, 



1844. 



South 



East . . 

 North 



West 



Direction 



of Wind, 



Eome, 



1860. 



South . 

 East. . . 

 North . 



West . 



Eome bifilar, 1860. 



High or 

 rising 

 days. 



20 



9 



119 



42 



Low or 



. falling 



days. 



81 

 22 

 17 

 21 



Father Secchi's numbers are placed alongside for companson. 



It wiU be seen, first, that at Makerstoun, for a south wind the number of days 

 for a high bifilar was just equal to the number of days for a low bifilar, and 

 that nearly the same conclusion holds for the north wind ; second, that both 

 east and west winds show an excess of days with a high bifilar. The residts for 

 north and south winds, then, are quite opposed to those fi'om the Roman Obser- 

 vatory, and the only case in which a similarity exists is that of the west winds ; 

 but that this coincidence is wholly accidental is evident from the con-esponding 

 result for Singapore. Indeed the numbers for Singapore agree generally very nearly 

 with those of Makei'stoun, the differences being explicable in most cases by days 

 for which the daily mean bifilar was but slightly plus or mimes of the monthly 

 mean. ■. ' ' 



If we now consider the nuipbers imder the two heads of weak and strong winds, 

 we shall obtain other grounds for concluding that the results, such as they are, are 

 independent of the direction of the wind. 



* See Trans. Eoy. Soc. Edinb. vol. xxii. pp. 484, 550. 



