nd steadily that the
crystals of reduced silver should not form a _metallic_ communication
beneath the surface of the fused chloride. On concluding the experiment the
positive electrode was re-weighed, and its loss ascertained. The mixture of
chloride of silver, and metal, withdrawn in successive portions at the
negative electrode, was digested in solution of ammonia, to remove the
chloride, and the metallic silver remaining also weighed: it was the
reduction at the _cathode_, and exactly equalled the solution at the
_anode_; and each portion was as nearly as possible the equivalent to the
water decomposed in the volta-electrometer.

814. The infusible condition of the silver at the temperature used, and the
length and ramifying character of its crystals, render the above experiment
difficult to perform, and uncertain in its results. I therefore wrought
with chloride of lead, using a green-glass tube, formed as in fig. 72. A
weighed platina wire was fused into the bottom of a small tube, as before
described (789.). The tube was then bent to an angle, at about half an inch
distance from the closed end; and the part between the angle and the
extremity being softened, was forced upward, as in the figure, so as to
form a bridge, or rather separation, producing two little depressions or
basins _a, b_, within the tube. This arrangement was suspended by a platina
wire, as before, so that the heat of a spirit-lamp could be applied to it,
such inclination being given to it as would allow all air to escape during
the fusion of the chloride of lead. A positive electrode was then provided,
by bending up the end of a platina wire into a knot, and fusing about
twenty grains of metallic lead on to it, in a small closed tube of glass,
which was afterwards broken away. Being so furnished, the wire with its
lead was weighed, and the weight recorded.

815. Chloride of lead was now introduced into the tube, and carefully
fused. The leaded electrode was also introduced; after which the metal, at
its extremity, soon melted. In this state of things the tube was filled up
to _c_ with melted chloride of lead; the end of the electrode to be
rendered negative was in the basin _b_, and the electrode of melted lead
was retained in the basin _a_, and, by connexion with the proper conducting
wire of a voltaic battery, was rendered positive. A volta-electrometer was
included in the circuit.

816. Immediately upon the completion of the communication w