ITS ERRORS AND CORRECTIONS, ETC. 485 
395-0 : 
1845. g= 5505 = 0°67 sc. div.; probable error = 0-09 se. div. 
The value resulting from Table XIII. is, 
1624:8 : 
gi 53133 = 0:702 se, div.; probable error =0-04 se, div. ; 
whence, 
[9] = 0-0001366. 
By hot-water experiments it was found 
1841, June, g = 0:0002686 
1841, October g = 00001667 
By applying hot sand under the suspended magnet, 
gq = 0:0002072 
The value adopted was, 
gq = 0:000214 
Therefore, 
[q] = 0-638 g. 
36. Toronto, 1842-45. The observations in 1843 till November have not been 
used in this determination, on account of a derangement of the bifilar wires.* The 
following are the results obtained :— 
1842 to 1843, February 10. & = 0:0000743. 
TABLE XIV.—Sums of DIFFERENCES OF BIFILAR MEANS. Toronto, 1842. 
Month. > At. f Month. 
DN ey — 
January, . 76-55 July, . 
February, . 72°72 August, . 

March, . . 62°77 September, . 
mpril, . . 71:46 October, . 
i 48:46 November, . 
Oune,- . . 37°83 December, . 





The final result is, 
1609-80 
= “5566 = 2°53 se, div. ;t probable error = 0:08 sc. div. ; 
whence 
[q] = 0:000188. 
* Toronto Observations, 1842-43, vols. i. and ii., printed under the superintendence of Lieu- 
tenant-Colonel E. Sain. 
+ The value employed by me in 1845, to correct the Toronto bifilar observations for the annual 
law, was obtained from a few groups of days (the sums of differences of temperature being 117°), 
and was q’ = 2-55 sc. div., almost identical with that here obtained from the whole year. 
