This is, legally, spirit of the strength denoted as proof by Sikes's hydrometer. Another legal definition makes proof spirit 'that which at the temperature of 51° F. weighs exactly J 12/13th parts of an equal measure of water." This matter is dealt with more fully a little further on. Meanwhile, it is convenient to state here what has been found to be the exact composition of proof spirit, with a few brief historical notes.

In the year 1847, Drinkwater, who was then a student at University College, London, and subsequently became a Collector of Excise, published the results of a very careful investigation which he had made upon the composition of proof spirit.1 He prepared alcohol as nearly free from water as he could get it, and concluded that "the number 0.79381 expresses the specific gravity of absolute alcohol at 60° F. within a very close degree of approximation." He then prepared proof spirit according to the above definition, using this absolute alcohol, and concluded from his experiments that the proof spirit consisted of 49.24 per cent. by weight of absolute alcohol and 50.76 per cent. of water, that its specific gravity at 60°; F. was 0.91984, and that the strength of the absolute alcohol was 75.25 degrees over proof.

1 Phil. Mag., 1848, 22, 123.

Subsequent work has only modified Drinkwater's figures very slightly. According to a careful revision of data based upon the most trustworthy published results of several investigators, the specific gravity of proof spirit at 60° F., in air, is 091976, water at the same temperature being taken as the unit of reference. Proof spirit contains 49.28 per cent. of alcohol by weight, or 57.10 per cent. by volume at 60° F.

The table on p. 267 shows the percentage of proof spirit corresponding with each integral indication-degree of Sikes's hydrometer at 60° F. The complete tables, which are published in book form, are arranged for each degree of temperature from 30° F. to 100° F., and for each fifth of an integral degree of indication.

A specimen of the tables for three different temperatures is here adduced*.

Specimen of Sikes's tables. Spirits at different temperatures.

All the strengths shown at the first temperature (50° F.) are "over proof" (op.); those at the second (60° F.) are partly over and partly under proof; and those at the third are all "under proof" (u.p.).

If an over proof strength is added to 100, the sum represents the

Proof spirit strength corresponding with the indications of Sikes's hydrometer. Temp. 60° F.

volume of spirit at proof strength which 100 volumes of spirit at that particular over-proof strength would make.

If an under proof strength is subtracted from 100, the remainder shows the volume of proof spirit which is contained in 100 volumes at that particular under-proof strength.

The sum and the remainder show, in fact, the percentages of alcohol, calculated as proof spirit, in the stronger and the weaker spirits, respectively. From this it is easy to find the equivalent proof quantity of any given volume of over-proof or under-proof spirit.

Suppose, for instance, that we have 120 gallons of spirit strength 6.5 o.p. Then the equivalent proof gallons are: 106.5 per cent. of 120 = 1.065 x 120 = 127.8 proof gallons.

If the strength had been 6 5 under proof, the equivalent proof gallons would have been: 93.5 per cent. of 120 = 0.935 x 120 = 112 2.

Sikes's tables are so constructed as to show, for a given spirit, the same strength at whatever temperature (within the limits of the tables) the strength is taken. A spirit, for instance, which shows 62.0 over proof at 60° F. will show the same strength at 65° F., though its indication will be different. This, no doubt, was often a convenience to revenue officers in identifying spirit in transit with the particulars furnished on its "permit."

There is, however, one defect of the system. It fails to take account of the change in volume due to alterations of temperature. For example, 100 gallons of proof spirit at 60° F. would become 100.5 gallons at 70° F.; but as its strength is still shown as " proof," the quantity of spirit on which duty may be levied is greater than before, though the actual quantity of alcohol is the same. Conversely, at a lower temperature the quantity chargeable will be less. This error in the evaluation of spirit has never been provided against in this country, though proposals have been made to that end. The differences which may arise from this defect, and which will be sometimes in one direction and sometimes in the other, are not considered to be so great as to make the question one of much practical importance.

Sikes's table, it will be seen, expresses the indications of the hydrometer in terms of alcoholic strength, not of specific gravity. The ' indications ' themselves, as already noted, are constructed upon an arbitrary scale. In the year 1833, however, a Committee of the Royal Society, which had been appointed to inquire into the question of spirit valuation, reported in favour of a hydrometer with a stem graduated in terms of specific gravity at a given temperature - 62° F. being recommended. The suggestion was not adopted; but, instead, a number of experiments were made to determine the specific gravities which corresponded with the indication.numbers of Sikes's hydrometer, and the results were embodied in a "Table for determining the Weight per gallon of Spirits by Sikes's Hydrometer." This table was shortly afterwards embodied in legislation authorising its use in evaluating spirits, and will be found as a schedule to the Spirits Act, 1880. It was employed by revenue officers in ascertaining the quantity of spirits in cask by the method of weighing. A revision of this original table was published in 1916, and its use is legalised by Sec. 19 (3) of the Act 5 & 6 Geo. V., cap. 89. The revised table is reproduced here for each integral indication.number: the complete table includes also the values for each fifth: -