XXX INTRODUCTION. 
SED] fe) 
ah, (7 + 2¢—¢) +300-0 
The adopted value of g=0-000294. See § 6. 
e=0000103 e’=:0000106. 
TABLE 11.—Values of & and gq for the Bifilar Magnetometer, and the periods to 
which they apply. 
Pericds to which the Values apply. Value of k. Value of q. 
1841. 
a. hh. dy ihe 
July 11 20—July 0.0001185 
Aug. 4 20—Sept. 0.0001522 
Se t. 7 20—Sept. ¢ 0.0001464 
Oct. 6 20—Oct. 0.0001303 
Oct. 19 23—Dec. 5) 
and during 1842 Onno tes 
33. AsI have had much doubt of the exactness of the values of an T have 
in all cases given the temperatures along with the corrected observations, so that 
each magnetician may have it in his power to apply a more accurate correction, if 
such is to be found, 
: : ; Am : 
It is obvious that, supposing the value of 4” to be well determined, the accu- 
Mm 
racy of the corrections will depend on the temperature of the magnet being well 
ascertained. 
The error in the observations of the scale will be generally less than 0-1 scale- 
division ; the error in reading the thermometer may be 0”1 Fahr.; the consequent 
error in correction may, therefore, be upwards of 0-2 scale-division. 
The thermometer, by Ross, has a bulb 0-5 inch in diameter ; it is inclosed in a 
glass tube, open at the bulb, which fits into the lid of the magnetometer-box, leay- 
ing the stem and scale above, and the bulb below. The box, like that of the Declino- 
meter, was not well closed, and it was soon evident that during a rapid rise or fall of 
temperature, the thermometer would indicate more or less than the temperature of 
the magnet. 
The fire lighted in the morning, during the winter months, conspired with the 
temperature of the day to make the increase more rapid than it would otherwise 
have been; and there is little doubt that the indications of the thermometer would 
frequently exceed the temperature of the magnet by 1°-0 Fahr. 
In 1843, having formed inner boxes, and had the whole well closed, Messrs 
