GENERAL MEETING. 29 
comes stationary ; finally it begins to rise. Fifteen minutes after 
the thermometer reaches its lowest reading this rise is about 0.°01. 
It is best-in observing the freezing point of a thermometer to put it 
in ice thoroughly saturated with water. The ice should be con- 
tained in a vessel from which the water cannot flow off. A ther- 
mometer put in ice in this condition takes on the temperature of 
0° more quickly than if put in dry ice containing a good deal of 
air. This method of siicartlaa the freezing point was introduced 
by Baudin of Paris. 
There is another correction to a normal thermometer which is 
little known and rarely applied. Let: 
V = volume of bulb at 0°. 
v = volume of tube from 0° to 100° at temperature 0°. 
= coefficient of cubical expansion of glass. 
7 = coefficient of expansion of mercury. 
T = thermometer-reading corrected for calibration, ete. 
= true temperature. 
A consideration of the construction of the thermometer leads to 
this equation : 
Vd+)+ed+M 2 -va+y. (1) 
The volume of the bulb at ¢°, plus the volume of that part of the 
tube corresponding to the thermometer reading, is equal to the 
volume of the mercury at ¢°. For ¢= 100 the equation becomes: 
V+ £100) + + £100) = V1 + 100). (2) 
Eliminating 7 from (2) by means of its value found from (1) 
we have: Way 
T=. par gg (3) 
Taking as the coefficient of cubical expansion of glass, 2, the 
quantity 0.000026, the following values are found for 7'— ¢, for 
the various readings of the thermometer from — 40° to + 100°. 
fl T—t 
oC. (a 6 
— 40 — 0.145 
— 20 — 0.062 
0 0.000 
+ 20 + 0.042 
40 + 0.062 
50. + 0.065 
60. + 0.062 
80. + 0.042 
