Table 1

In this way a table, or curve, of corrections may be constructed, and the error at any scale reading of the thermometer may be easily ascertained.

"Normal" thermometers may now be purchased ; they are compared with a standard thermometer before graduation, and true temperatures are said to be registered by them.

2. Redetermination of Zero Point. - The zero point of a thermometer should be redetermined from time to time, as it is subject to changes which, in the case of the cheap soda glass thermometers, may be considerable. These changes are of two kinds: (a) If a thermometer be graduated shortly after the bulb has been blown, the zero point will be found to rise, at first with comparative rapidity, then more and more slowly, and the elevation of the zero point may go on for many years. If the thermometer be kept at a high temperature - especially, as shown by Marchis, if there are periodical, slight fluctuations of temperature - the rise of the zero point takes place with much greater rapidity, and up to, at any rate, 360°, and probably 450°, the higher the temperature the more rapid is the rise and, apparently, the higher is the final point reached. A rise of more than 20° has several times been observed in the case of soft German glass thermometers on being subjected to prolonged heating at 360°. In all cases the rise, which is rapid at first, becomes slower and slower, and it seems doubtful whether, at any given temperature, actual constancy of zero point has ever yet been attained. If, however, a thermometer has been heated for many hours to a given high temperature and then allowed to cool very slowly, subsequent heating to lower temperatures has very little effect on the zero point. The best thermometers, as first recommended by Crafts, are kept at a high temperature for a long time before being graduated.

(b) If a thermometer - even after its zero point has been rendered as constant as possible - be heated and then cooled very rapidly, a slight fall of zero point will be observed ; but after a day or two the greater part of this fall will be recovered, and the remainder after a long period.

3. Volatilisation of Mercury in Stem of Thermometer. - In the cheaper thermometers there is a vacuum above the mercury and, when the mercury in the stem is strongly heated, volatilisation takes place, the vapour condensing in the cold, upper part of the tube ; when, therefore, the temperature is really constant it appears to be gradually falling. The better thermometers, which are graduated up to high temperatures, contain nitrogen over the mercury, a bulb being blown near the top of the capillary tube to prevent too great a rise of pressure by the compression of the gas ; but thermometers which are only required for moderate temperatures, say, not higher than 100° or even 150°, are not usually filled with nitrogen. If, however, such thermometers are used for the distillation of liquids boiling at so low a temperature as 100°, or even 80°, a quite perceptible amount of mercury may volatilise and, after prolonged heating, errors amounting to 0-2° or 0-3° may occur. It would be much better if all thermometers required to register temperatures higher than 60° were filled with nitrogen.

4. Correction for Unheated Column of Mercury. - As already mentioned, the thermometer should, if possible, be so placed in the apparatus that not only the mercury in the bulb but also that in the stem is heated by the vapour of the boiling liquid ; otherwise the following correction, which, at the best, is somewhat uncertain, must be applied :To the temperature read, add 0.000143(T-t)N, where T is the observed boiling point, t the temperature of the stem above the vapour, and N the length of the mercury column not heated by the vapour, expressed in scale divisions.

The coefficient 0.00016 - the difference between the cubical expansion of mercury and that of glass - is very frequently employed, but it is found in practice to be too high ; and Thorpe has shown that the value 0.000143 gives better results.

Table 2, on page 12, given by Thorpe 1 may be found useful.

1 Thorpe, " On the Relation between the Molecular Weights of Substances and their Specific Gravities when in the Liquid State," Trans. Chem. Soc, 1880, 37, 159.