580 
PROFESSOR FORBES ON THE TEMPERATURES 
eye at A. The thermometer is then placed so that the point of the scale to be viewed 
is exactly under the cross wire ; all points so placed are therefore viewed in the same 
direction. 
In Troughton and Simms’s thermometer in question (marked J. D. F.), we have 
seen that the error of the scale at the freezing point is + 0 o *33, at the boiling point 
of water it is — 0°*40. Consequently, 
212° on the scale corresponds to 212*40 nearly. 
32° on the scale corresponds to 31*67 nearly. 
Difference . . . 180*73 
In the first experiment a column of mercury was detached, which having been 
made to fill six consecutive and adjoining spaces in the tube, commencing with 32° 
on the scale, terminated at 21 1°*25 : hence, assuming the error at the last point to be 
the same as at 212°, (with which it nearly coincides,) we have 
First reading 32°*00 
Correction — 0°*33 
True 31°*67 
Last reading 211°*25 
Correction + 0°*40 
i 
True 211°* 65 
True thermometric interval corresponding to six 1 lypo-pg 
times the length of the column employed . . j 
Length of the column, or I, . . * 30°*00 
We may then immediately compare the readings taken at the successive steps with 
the true temperatures which the preceding value of I gives us, and which will be 
31°*67 + I ; 31°*67 +2 1, &c. 
True Temperatures. 
Readings of Scale. 
Errors of Scale. 
o 
o 
o 
31-67 
32-0 
+ 0-33 
61-67 
62-2 
+ 0-53 
91-67 
91-75 
+ 0-08 
121-66 
121-75 
+ 0-09 
151-66 
151-45 
— 0-21 
181-65 
181-5 
— 0-15 
211-65 
211-25 
— 0-40 
Such experiments may be repeated with columns of different lengths, which will 
serve not only to fix the errors at a greater number of points of the scale, but like- 
wise to test one another by their mutual agreement. Experiments were made with 
1 80 1 SO 
columns containing = 18° nearly, and -p = 22°*5 nearly. They were performed 
exactly as above described, and with the following results : 
