126 
EIGHTH REPORT. 
of the filter companies, and the process developed into the rapid method 
of filtration as we now know it. The ordinary chemical used is sulphate 
of aluminum. This substance is first dissolved and the solution intro¬ 
duced into the water to be purified and thoroughly mixed with it. That 
produces a precipitate which entangles the impurities and carries them 
to the bottom; a portion of the precipitate in the form of fleecy masses 
is deposited on and through the sand bed and then entangles other pass¬ 
ing impurities as does the mat in the other process. The water is 
passed rapidly through the sand and the impurities will be washed 
through it if the bed is run too long, so it must be shut down and 
washed from time to time. This washing takes place from once in 
six hours to once in forty-eight hours, dependent upon the character 
of the water, and is performed by pumping water into the under drains 
and at the same time stirring up the bed with rakes or compressed air, 
thus floating the collected impurities to the surface, whence they are 
carried away to a sewer. Early plants of this type which were put in 
were not in all cases successful. There were many of those plants 
that are claimed to have been successful which really were not so. The 
filter companies naturally did not wish to build a different machine 
every time they built one, and consequently they did not make any 
special study of the water with which they dealt. The treatment of 
water is not widely different from the treatment of disease. For suc¬ 
cess you must find out what sort of a treatment a particular water must 
have. As a result a great many of their plants were failures, and very 
few, if any, were unqualified successes. Along in the 90’s, however, 
Providence took up water purification, and undertook a series of ex¬ 
periments upon rapid filters. The results showed greater possibilities 
than had previously been anticipated, and these experiments were fol¬ 
lowed by others at Louisville, Cincinnati, and Pittsburgh, and since 
then at St. Louis. There have been constructed within the past two 
or three years some very successful plants operated on this system. The 
first was that of the East Jersey Water Company at Little Falls, N. J., 
having a present capacity of about 34,000,000 gallons per day, designed 
by Mr. George W. Fuller. The next one of importance was that at 
Ithaca, and that was followed by one at Watertown, both by Mr. Allen 
Hazen; the efficiencies being in all three about the same as for slow 
sand; so that we may say that this method has proved its right to stand 
with the older one. The difference is that with a slow sand filter the 
longer you let it alone the better work it will do, but with the mechan¬ 
ical filter the operator has to be on deck all the time. He must see 
to it that all the water receives the proper dose of medicine. With the 
rapid filter the water must always be clear to be safe; if it is at all 
turbid when it comes from the filter, that is absolute proof that the 
purification is imperfect. So that where you have a water that can be 
handled with slow sand filtration it is safer. You are less likely to have 
a slip on account of the failure of an attendant. Where you have to 
make use of the rapid filter it will give good results if you put it into 
competent hands. 
This has described what we can do with a bad water supply if we 
have to. I do not advise purifying a bad water, it is better to get a 
pure water to start with, but if we cannot get such an one naturally, 
we do not give it up. 
University of Michigan. 
