lates during experiments, and
must be carefully watched in every endeavour to trace the source, strength,
and variations of the voltaic current. Its effect was avoided in the
experiments already described (1036. &c.), by making contact between the
plates P^{1} and P^{2} before the effect dependent upon the state of the
solution in contact with the zinc plate was observed, and by other
precautions.

[A] Journal de Physique, lvii. p. 349.

1041. When an apparatus like fig. 98. (1017.) with several platina plates
was used, being connected with a battery able to force a current through
them, the power which they acquired, of producing a reversed current, was
very considerable.

1042. _Weak and exhausted charges_ should never be used at the same time
with _strong and fresh ones_ in the different cells of a trough, or the
different troughs of a battery: the fluid in all the cells should be alike,
else the plates in the weaker cells, in place of assisting, retard the
passage of the electricity generated in, and transmitted across, the
stronger cells. Each zinc plate so circumstanced has to be assisted in
decomposing power before the whole current can pass between it and the
liquid. So, that, if in a battery of fifty pairs of plates, ten of the
cells contain a weaker charge than the others, it is as if ten decomposing
plates were opposed to the transit of the current of forty pairs of
generating plates (1031.). Hence a serious loss of force, and hence the
reason why, if the ten pairs of plates were removed, the remaining forty
pairs would be much more powerful than the whole fifty.

1043. Five similar troughs, of ten pairs of plates each, were prepared,
four of them with a good uniform charge of acid, and the fifth with the
partially neutralized acid of a used battery. Being arranged in right
order, and connected with a volta-electrometer (711.), the whole fifty
pairs of plates yielded 1.1 cubic inch of oxygen and hydrogen in one
minute: but on moving one of the connecting wires so that only the four
well-charged troughs should be included in the circuit, they produced with
the same volta-electrometer 8.4 cubical inches of gas in the same time.
Nearly seven-eighths of the power of the four troughs had been lost,
therefore, by their association with the fifth trough.

1044. The same battery of fifty pairs of plates, after being thus used, was
connected with a volta-electrometer (711.), so that by quickly shifting the
wire