THERMOMETRY
���The thermometer is fixed in the central tube, which contains a sufficient quantity of pure water. The whole apparatus is cooled a few degrees below zero ; then a part of the supercooled water is induced to crystallise. The temperature rises, and the ascension of the mercury stops when the meniscus indicates the true zero.
To verify the 100° point, the apparatus shown in fig. 84 is used. The water used must be pure, and nearly the whole thermometer must be suspended in the vapour, the surface of the mercury column extending only a very short distance out of the apparatus. N.B. — The operation will only give an exact in- dication if it be carried out under a pressure of 760 mm. of mercury. For each millimetre higher or lower than this the boiling point changes by -fo of a degree. The error at the 100° point affects the lower degrees proportio- nately to their distance from the zero of that at 100° ; at true zero it is nil.
Let us admit for the present that the points 0° and 100° on the scale have been found correct. The equidistant marks denoting intermediary temperatures will only be exact if the internal diameter of the capillary is quite uniform. This condition is not generally fulfilled, and by the calibration of the instrument we get a new list of corrections (see Ostwald's 'Manual of Physico-chemical Measurements/ English translation by Walker).
It is easy to combine all the corrections in one table, or even to give them a graphic representation (correction curve).
The temperatures denoted by a thermometer experi- mentally corrected do not absolutely agree with those indicated by an instrument which has been graduated
��I
�� �
