Advantage of the density bottle
We have already pointed out the advantage of the idea of relative density as a step towards the accurate measurement of density. There are no volume measurements to worry about. Weighings only are required and these can be carried out with a tappan or beam-balance to a fairly high degree of accuracy. Thus, a good value for the relative density and hence the density of the liquid can be obtained by this method. Contrasted with the simple measurements of density described earlier a measuring cylinder of the size ordinarily used can be read only to within about 0.5 cm-‘; even a burette to within 0.1 cm ‘ only. Consequently, unless a very large volume of substance is used the percentage error in the volume measurement will be large. It follows that density determinations of liquids which depend on the direct measurement of volume will be less accurate than those obtained by the use of a density bottle.
These two cubes were made at the National Physical Laboratory, and used to determine the density of mercury to one part in a million. Measurements were made of the mass of mercury which filled the hollow cube and of the mass displaced by the solid one. A mean value from the double experiment eliminated errors due to surface contamination which caused a positive error in one case and a negative error in the other. The volumes of the cubes were calculated from their dimensions measured in terms of light wavelengths. It is of interest to note that the fused quartz plates of the hollow cube were made so accurately flat that they stuck together by molecular cohesion without need for cement. The solid cube was made
from a sintered mixture of tungsten carbide and cobalt