ual to 20,000 gallons an hour could be obtained.

After a number of experiments had been carried out it was decided to run
the sewage as rapidly as possible through electrodes, six cells or two
rows in series fixed in a long channel or shoot, for experience showed
that the motion of the liquid acted on reduced the back E.M.F. and
hastened the formation of the precipitate.

A channel is kept at the bottom of the electrodes for the silt to
collect, with a culvert at side to flush it into, so as to prevent any
block occurring; the advantage of this is obvious. The plates in each
section may be from half an inch to an inch thick, and can be of any
length up to 6 ft. It may possibly be objected that a large number of
plates is required. This may be so, but the larger the number of plates,
the less the engine power required, and the longer they last. In each
section the electrodes are in parallel, and any one section is in series
with the other, the arrangement being exactly like that of a series of
primary battery cells.

By actual experience I have been able to prove that at least 25 sections
of electrodes should be in series and across any one of these sections
the potential difference need not be greater than 1.8 volts, the current
being of any desired amount, according to the surface of plates used.

The electrical measurements taken by Dr. John Hopkinson during these
experiments for the Electrical Purification Association, to whom I had
sold my patents, entirely corroborated my contentions as to E.H.P. used,
and agreed with the measurements of the managing electrician, Mr.
Octavius March.

The process was then thoroughly investigated by Sir Henry Roscoe, who
had control of the works for one month. He reports as follows:

"The reduction of organic matter in solution is the crucial test of the
value of a purifying agent, for unless the organic matter is reduced,
the effluent will putrefy and rapidly become offensive.

"I have not observed in any of the unfiltered effluents from this
process which I have examined any signs of putrefaction, but, on the
contrary, a tendency to oxidize. The absence of sulphureted hydrogen in
samples of unfiltered effluent, which have been kept for about six weeks
in stoppered bottles, is also a fact of importance. The settled sewage
was not in this condition, as it rapidly underwent putrefaction, even in
contact with air, in two or three days.

"The results of this chemical investigat