40 QUANTITATIVE AaitfCCI/f'CliAL ,\\AlA'SfS
bo carefully tested at the beginning, to give a^unmr
can safely be used for this method of procedure.
Calibration of Weights.— Mxponsive sets of weights are usually
adjusted with sufficient accuracy for most analytical work, but
with weights of the grade ordinarily available u calibration
should be made. Weights that are found to be in error may then
be either adjusted to accurate values or used with corrections.
This is a matter that is given serious attention in Jar too feir
laboratories, college or industrial. Conntwmal treujhh /m/w//////
are in error to the extent of two or three per cent and mw Ittrtjer
errors may be found, after the weights hare been in nwfur a year or
more. To ignore such errors a,s- lh<w., white, inmxtiny Hjton a hit/ft
degree of accuracy in the other phaf*c$ of the laboratory work, u
nothing less than gross inconsistency.
Calibration may be accomplished most conveniently and
with accuracy by comparison of each piece of the set with the
corresponding pieces of a standard set, whoso row'rfions an*
known. Also, if the arms of the balance* an* known to be of the
same length to within a negligible en'or (, I •* not more than
- \ /.
\
0.000011 the comparisons may be made by placing tlir piecen
on opposite pans of the balance and noting whether tin* rider
must be used to obtain equilibrium and, if HO, its norrswary
position on the beam. Thin gives, directly, the valut* of tin*
experimental piece in terms of the standard piece. Sine** f IIIH is
a direct comparison, the ssero point of the balance must In* known.
If an entire set of standard pieces i.s not available a Hiiigli*
standard piece, as a gram, may be used, when the* <*aI<*uiationH
become more complicated and the calibration less accurate.
Also if the balance arms are riot sufficiently near the name length
(found by weighing an object and then exchanging object am!
weights, and reweighing) or if nothing in known regarding thin
point, a different method of comparison inunt be* employed.
This is known as the "method of Hukstitution." In thr, mrrw
that follows it will be assumed that an en lire, standard net ix at hand
and that the method of substitution is- to \H\ uml. ThiH met hod in
safest in any case and it makes unncccHwiry a determination of
the relative length of the arms or of the zero point of thr* unloaded
balance.