If for any reason the specific gravity of the alcohol is not taken at the standard temperature, it is necessary to include a correction to compensate for the deviation. The correction is greater at higher strengths than at lower, as will be seen from the table given below, which is used in the following manner: -

The difference between the standard temperature and the actual temperature of the observation is multiplied by the appropriate factor, taken from the table. If the actual temperature is higher than the standard, the product is added to the observed specific gravity; if it is lower, the product is subtracted. The unit throughout is water at the standard temperature (60° F.).

Table of temperature correction's.  Example: - At the temperature 65° F., the sp. gr. of a specimen of diluted alcohol is 0 9475, referred to water at 60° F. as unity. As this lies between 0 946 and 0 949, the appropriate factor is

000036, and the correction is 5 X 000036 = 00018. Hence the sp. gr. of the sample at the standard temperature, 60° F., is 09475 + 00018 = 09493.

Had the temperature of observation been 55° F. instead of 65° F., the product 0 0018 would have been subtracted, and the sp. gr. at 60° would then have been 0 9475 - 00018 = 09457.

It may be pointed out that where the greatest accuracy is required the temperature of the spirit must be very carefully adjusted in taking the specific gravity. For ordinary purposes it suffices if the sp. gr. is correct to one unit in the fourth place of decimals, which corresponds with about 0.1 per cent. of proof spirit. To obtain a result correct to one unit in the fifth place of decimals it is necessary to adjust the temperature to 001° C. or 002° F.; and in general for fairly accurate work the temperature of the alcohol should be correctly adjusted to within 01° F. This is requisite for a result accurate within five units in the fifth decimal place of the specific gravity, corresponding with about 0 05 per cent. of proof spirit at medium and lower strengths. When special accuracy is required, the desired temperature should be obtained by means of a thermostat, and the form of pyknometer employed should be one in which the level of the liquid is adjusted to a mark on the neck, not the form with perforated stopper. Fig. 33. - geissler's form of pyknometer.

With ground-in thermometer and capped side-tube.