STANDARDS OF CAPACITY Theoretically, the unit of capacity should be the same as the unit of volume—that is, the volume of a cube each of whose sides is equal to the unit of length. In practice, however, it is extremely difficult to construct such a cube with accuracy, and still more difficult to measure the internal volumes of vessels of different shapes in terms of the unit of length. Practical neces sity has therefore ordained the use of a unit of capacity based on the unit of mass, rather than on the unit of length.
In the metric system the kilogramme was originally intended to be the mass of one cubic decimetre of pure distilled water when at its temperature of maximum density (4° Centigrade). Although great care was taken in the construction of the origi nal kilogramme from this definition, and the result attained was closer than might have been anticipated in view of the difficulty of the problem, it has been found by very careful experiment that the litre, which is now defined as the volume occupied by I kilogramme of water at 4° C, actually equals 1.000028 cubic decimetres. The gallon is somewhat similarly defined as the volume occupied by io lb. of pure distilled water at 62° F when weighed in air at a barometric pressure of 3o in. of mercury, against brass weights. As the weighing has to be made in air, and the density of the brass weights is not prescribed by the Act (Weights and Measures Act 1878) there is a certain ambiguity about this definition. On reasonable assumptions it has been calculated that i gallon = 4.5459631 litres.
It will be noted that in the case of the litre the definition refers to the kilogramme mass—that is to say, the weighings are to be reduced to vacuum, by applying a suitable correction for the difference of the air buoyancy on the distilled water and on the weights used for weighing it. This entirely eliminates the difficulty mentioned as regards the exact definition of the gallon. Much confusion of thought has, however, arisen from the pref erence of some chemists for the use of the so-called "Mohr's litre," which is defined, on the same lines as the gallon, as the volume occupied by i kg. of distilled water, when weighed in air, against brass weights, at a temperature which has never been explicitly laid down. Roughly, a Mohr's litre equals 1.002 true litres. It is very unfortunate that the term litre should have been appropriated to a unit so vaguely defined and which differs from the true litre by an amount that, although small, is too great to be negligible. For most purposes the difference between the millilitre and the cubic centimetre can safely be neglected.
form of cylindrical metal vessels with flat brims, and are filled exactly to their brims by the aid of flat glass "strikes." Com parisons with other vessels are made by transfer, the vessel under comparison being first filled and emptied to compensate for the amount of water left behind in the standard when the latter is emptied into the vessel being tested.
For scientific apparatus, such as flasks, burettes, pipettes, etc., where higher accuracy is needed, it is found necessary as a rule to base the verification directly on the original definition by actually weighing the quantity of pure distilled water contained in, or delivered from, the vessel, making due allowance for the temperature of the water at the time of weighing and for the buoyancy of the surrounding air. In the case of vessels, such as burettes and pipettes, which are intended to deliver, and not to contain, definite quantities of liquid, it must be remembered that after delivery the walls are left wet to an extent which depends not only on the rate at which delivery is effected and the time allowed for subsequent drainage but also on the viscosity and surface tension of the liquid being measured. The rate of delivery is of more importance than the drainage time, and provided it is sufficiently slow the quantity of liquid delivered will be reasonably constant. Suitable delivery times and error allowances are sched uled in a pamphlet on testing volumetric glassware issued by the National Physical Laboratory.