red, it is of great
importance to keep it well away from any optical instruments, as the
most minute trace of vapour in the air will produce a highly destructive
corrosion of any glass surfaces.

_Methods of Calibration._--In the case of apparatus for volumetric work,
this is usually carried out by weighing, although some of the smaller
subdivisions are often made by measurement. When the subdivisions are
made in this way it is of importance to see that the walls of the tube
or vessel to be calibrated are parallel. Great errors arise in some of
the commercial apparatus from neglect of this precaution. A convenient
method of testing for parallelism, in the case of a wide tube, is to
close one end and to weigh in successive quantities of mercury. An
observation of the length occupied by each successive quantity will
indicate any change in the bore. In the case of capillary tubes, it is
convenient to introduce an unweighed quantity of mercury, measure its
length accurately, and then to move it along the tube in stages, either
by tilting the tube or by the application of air pressure. A measurement
of the length at each stage will indicate whether the bore is
approximately parallel or not. Neither of these methods is to be relied
on without a careful examination of the tube, as it may happen that
there are local irregularities in the bore which compensate for each
other, and do not, therefore, affect the volume of a given length.
Obviously, the smaller the quantity of mercury with which the test is
carried out and the greater the number of observations made, the less
risk will there be of such an error. A liquid, such as water or alcohol,
which wets the glass is not suitable for such a test, unless special
precautions are taken.

When, however, a pipette or burette has to be calibrated to deliver a
certain volume of water, the final calibration must be made with this
liquid. Thus, the burette would first be calibrated by weighing in
definite quantities of mercury of say 13.54 grammes (1 cc at 15 deg.C.),
each of the 1 cc divisions should be marked by some temporary marking.
The burette is now filled with a solution of potassium bichromate and
sulphuric acid and allowed to soak for some time; the bichromate is
washed out and distilled water is put in. Successive quantities of water
are run out of the jet, a fixed time being allowed for draining, and the
weights of the quantities delivered are noted. This procedure will give