476 MR JOHN ALLAN BROUN ON THE BIFILAR MAGNETOMETER, 
22. Cape oF Goop Hore, 1841-46.—The series of published observations 
closed June 1846.* The values of g' obtained by myself from the daily mean 
uncorrected scale readings and mean temperatures were as follow :— 
Magnet R. £=0-000180 
TABLE IV.—SuMs oF SUCCESSIVE DIFFERENCES OF DaILy MEAN BIFILAR TEMPERATURE 
(2A ¢) AND SCALE READING (= A R). 








Month. DAC. DIN 185, q'. Month. ZAt 
porn: Oe, Aaa scant oa oe oll a Ee 
1841. Dee. 33:23 52°38 1:58 |1842. Dec. 29°65 
1842, Jan, 29°27 84°31 117 | 1843. Jan. 32°55 
Feb. 34°43 24:73 0°72 Feb. 39°95 
March, 23°14 24°43 1:04 March. 26:23 
April. 33°31 46:16 1:39 April. 
May. 18°81 19-10 1:02 May. 18:60 
June. 27°56 36°88 1:34 June. 35°41 
July. 25°41 44-48 1-75 July. 29°38 
Aug. 20:18 13°69 0-68 Aug. 43°60 
Sept. 16:99 9-52 0-56 Sept. 28°30 
Oct. 39°70 50°65 1-28 Oct. 33:01 
Noy. 26:48 27:44 1:04 Nov. Readjusted. 
igh als Us Wie besos tx oer ee alt lpeel at le let, oe! Loreal 13) [et be 



The resulting value is 
, 2AR, 763:34 
JUS At. 646-14 

= 1/183 sc. div. ; mean temperature, 63°6. 
As the magnet was removed for temperature experiments in April 1843, the 
value of q after the readjustment may be determined apart. We find for May 
to October 1848, 
ty 228-06 
7 = 788-30 

= 1:185 scale divisions. 
23. It should be repeated here, that the extent of agreement or disagreement 
in the partial results for g’ given in the above table, is not to be considered as 
deciding the accuracy of this method, since the variation of these partial results 
is chiefly due to days of great magnetic disturbance, and no day has been rejected, 
however great its irregularity, in these calculations. These partial results give 
the probable error of the final value of ¢ to be, 0:05 sc. div.; or g’ = 1-183 scale 
divisions, with a probable error of 0°05 sc. div. 
* Magnetic Observations, Cape of Good Hope, vol. i., printed under the superintendence of 
Lieut.-Colonel H. Sanine. 

