The Filtration of Natural Waters.*

The Filtration of Natural Waters.*

(Continued from Page 4.)

The surface waters used to supply London from the Thames or the Lee are filtered by the method of continuous filtration, a surface of 100 acres being required for the purpose. The thickness of sand differs with the different companies supplying the city with water, from two feet at the East London and Grand Junction companies to four and a half feet at the Chel sea Company, and the rate of filtration per hour in imperia gallons per square foot of filtering surface is two and one-sixth with the Lambeth Company to one and one-half gallons with the Southwark & Vauxhall Company. Two and one-half gallons, or five vertical inches an hour (which is seldom attained), is considered the maximum consistent with good clarification. Complete analyses are made of the water supplied to the metropolis by the different companies. Some of the determinations, as for instance color, and the amount of permanganate to oxidize the organic matter, are made daily, other chemical determinations are made weekly. The monthly determinations made by Dr. Percy F. Frankland of the bacteria in the waters of the different companies have been suspended since December, 1888. The average reduction of the number of micro-organisms present in the waters of the Thames and Lee, w’as in 1887, 97.6 per cent in the case of the Thames and 93.9 in the case of the Lee. “ If,” says the report on the metropolitan supply for December, 1888, “ these figures could be accepted as at all representing the degree of security given to consumers of the waters of those rivers by preliminary filtration, it is evident that the views on this subject acquired by a consideration merely of the results of comparative chemical analysis of filtered and unfiltered waters would have to be considerably modified and the character of the water supply would be correspondingly raised in public estimation. Further, if the results obtained from month to month could be relied on as an index to the effect of filtration in eliminating objectionable matters from the water, the bacteriological method would seem to afford a delicate and easily applied test of the working efficiency of the filter.”

The average number of bacteria in the water of the Thames is generally less than in the Spree at Berlin ; thus, during the year ending December, 18S7, the highest number in November was 81,000, and the lowest in June was 2200, the average for the year being 21,492.

The only filtering plant in this country that I know of, which at all compares with the plants in Europe, is that at Poughkeepsie, where the Hudson river water is converted into good, clear water, though not absolutely free from color. Mr. Fowler, the superintendent of the works, writes me with regard to the details of the filtration : “ Our usual rate of filtration is about six inches per hour, vertically, and this we regard as the maximum of efficiency, although we can sometimes do good work, so far as clarifying is concerned, at double that rate, and at other times are unable to do good work at one-half that rate, although the latter condition is exceptional. Very much depends upon the condition of the water in the river. The depth of water on the sand varies from one to three or four feet, and the difference of level between the surface of the water on the beds and that in the intermediate basins is usually two to four or five feet.”

The rapid filtration of water through coarse gravel is not unfrequently carried out at water-works to remove the larger particles floating in the water. When a filter of this kind is cleaned it is surprising to see the amount of fine dirt of all kinds that has been intercepted by the coarse material. Filters of this character do not pretend to purify the water in the sense of removing bacteria or in oxidizing the organic matter, but they are useful just to the extent to which they clarify the water and thereby improve its appearance.

The American system of filtering large quantities of water may be said to be the mechanical filters working under pressure. These filters are composed of four or five feet of moderately fine sand (some have also a mixture of coke) enclosed generally in boiler-iron cases. They w’ork with tremendous rapidity, sometimes over a hundred vertical feet an hour, but forty feet is said by some to be the highest rate consistent with good filtration. In this system alum is generally added to the water as a coagulent. Its effect in very small amount is quite remarkable—say a grain to a gallon, or even less—in retaining the solid matters of the water in the sand of the filter. The alum is decomposed by the carbonates in the water and hydrate of alumina is precipitated. This is a gelatinous and slimy substance and immediately surrounds the algse, clay and anything else that may be suspended in the water, and the sand retains this coagulated mass. Alumina has also the effect of taking the color out of water, so that clear, colorless water may be obtained by this process from swampy waters full of growing algce, with almost incredible rapidity.


These filters are in very general use in paper mills and other industrial w’orks where a clear colorless water is needed, and where a colored turbid river water is the only natural supply available. They have also been introduced into some cities of considerable size as, for instance, Long Branch, Chattanooga and Atlanta, and they are said to give water that is satisfactory to those who use it. The objection to the system is the use of alum. If all the alum used were decomposed in the few seconds that it takes the water to pass through the filter, so that no undecomposed alum passed into the filtered water, there might be no objection to its use, but this is not always the case. The amount of alum used is ordinarily small and It is claimed that even if it all went into the filtered water it would not injure it for drinking. This tnay be so, but the prejudice that exists against drinking water that has been treated with “chemicals” is so strong that it is not likely that any system using a coagulent in a soluble form will find general acceptance. Under some conditions when the water has high color with much suspended matter the alum has to be increased largely to give good results. I have known as high as seventeen grains to the gallon to be used with a very bad swampy water.

In this connection should be mentioned the spongy iron filter of Bischof which gives most excellent results both in taking out suspended matters from the water, including the bacteria, and also decreases the hardness of the water. This filter has been used on the small scale in houses, and also on the large scale in Antwerp to decolorize and otherwise purify the w’ater of that city. The filter is composed of finely divided metallic iron made by reducing iron ore by means of carbon at a temperature below fusion. Its action was not understood for a long time, and the mystery that surrounded it was an additional recommendation for it. The rationale of its working seems to be this, namely, that the iron being in a very finely divided state is dissolved to a slight extent by the combined action of the oxygen and carbonic acid in the water, and the ferrous carbonate thus produced is further oxidized, forming hydrate of iron, and then this acts as a coagulent just as the alumina hydrate docs. The system was said to be too expensive on a large scale, and it has now been replaced at Antwerp by the Anderson process in which the dark water is made to pass through a long revolving iron cylinder in which there is a large quantity of fragments of cast iron. These fragments of jron in their friction one on the other are abraded, minute particles arc broken off which are dissolved in the way above described. The water coming from the revolving cylinder is exposed to the air, the iron oxidizes and precipitates, combines with the coloring matter in the water, encloses the solid particles, and is then filtered out through sand. The process is said to work satisfactorily and give clear, colorless water. There is no objection to this use of iron as a coagulent provided that it is all oxidized and precipitated, and none is carried in solution into the filtered water; but this takes time.

Both alum and iron salts have a tendency to sterilize water. Their action may be both direct by killing the bacteria, and indirect by removing them with the precipitated alumina or iron hydrate. If a drop of a solution of alum, or of iron chloride be added to a gallon of turbid water, it will become perfectly cle$r in the course of a few hours, the alumina, or iron hydrate, which is formed in the water, settling to the bottom and carrying all the suspended matters with them. It has been proposed to clarify Mississippi water by adding a very small quantity of an iron solution to the water in the settling basins.

* By Thomas M. Drown, Member Boston Society Civil Engineers. From the Journal of the Association of Engineering Societies.

I have laid some weight on the desirability of following nature’s processes in the purification of impure waters. Neither the American system with its mighty rush of waters or the European system with its calm, steady and deliberate flow finds any analogy in nature. In the rapid-working mechanical filters a coagulent is used to grasp and hold the suspended matter ; in the continuously working filter beds the bacteria are put to a novel use in retaining the solid matter on the sticky surface of the sand.

Nature uses these methods also, she removes color by means of clay in the soil and intercepts mechanically in the bacterialaden soil all the solid matter in the water, but she goes further than tlr.s and, giving the bacteria full play, breaks up the organic compounds and leaves no trace of their existence behind.


To do this time is needed. “The bacteria of nitrification,” as I)r. Smart has well said, in referring to the systems of rapid fiitrations, “cannot be harnessed to the work of artificial filtration.”

The rate at which nature works may lie expressed in the amount of rain-fall. If we take the rain-fall at fifty inches and assume that even half of this sinks into the ground (a very high estimate) wc have twenty-five inches yearly on a square foot of surface. The amount of water which goes through the Berlin filters, at a rate of four inches an hour, is more than 1250 times this amount. If we wish to imitate the process by which nature makes its springs we must pour the water from river or lake which we wish to filter intermittingly on the surface of ground which is favorable for this purpose. The favorable conditions are these, the ground must be sufficiently porous to allow the ready flow of water through it, and it must have such relations to the strata t>elow as will enable us to collect the water at some lower level. It would not profit us much to pour the water on to a gravel bed if we could not find it again after it had been filtered. If the natural conditions for this purpose are not favorable drains would have to be put in at suitable depths to collect the filtered water. Water purified in this way would in no wise differ from natural spring water, provided that the amount of water applied did not exceed the capacity of the filtering area.

The question of the maintenance of the purity of the water supplies of large cities, which are dependent upon surface waters, is daily becoming more urgent in this country as the population becomes more dense on the collecting areas, and the protection of streams against pollution becomes more and more difficult. The radical remedy in such cases is to take water from another and. usually, more distant region, which, it is probable, will never become thickly settled. But in filtration, both intermittent and continuous, when intelligently conducted, we have a substitute which can give as clear, colorless, and we have good reason also to suppose, safe drinking water.

MANCHESTER FIREMEN KNOW “ DR. Carter.”—One “ Dr. Carter,” claiming to be a medical examiner of the Washington (D. C.) Fire Department, employed by the census bureau to collect statistics regarding tire department life-saving apparatus, recently paid his respects to Chief Smith of Easton, Pa. Chief Smith telegraphed to Chief Parris of Washington, D. C. The latter disclaimed all knowledge of the “ doctor.” What the man’s object was has not been LEARNED.-FIRE AND WATER.

A perusal of the above clipping will call to the minds of many the fact that the same party paid our city a visit just previous to the firemen’s parade a year ago last October. The “doctor” at that time made the same claim to being medical examiner of the fire department in Washington, hut letters to the chief there failed to even locate him in the city, say nothing of his holding a position so important as that claimed. While here he entertained the boys at the Vine street station with long yarns of his connection with every known secret organization on earth, excepting the Odd Fellows, and said he had put in an application for those degrees. On the strength of his alleged connection with the Washington Fire Department he induced several of the men to purchase a medical work for which he was canvassing. It is also pretty well understood that his attentions to a young lady at the house where he boarded was the cause of threats upon the part of another man to chastise him, which fact may have led to his somewhat abrupt departure from the city.—Manchester (TV. IT.) Union.

The Filtration of Natural Waters.


The Filtration of Natural Waters.

(Continued from Vol. VIII. Page 270.)

The experiments of the Massachusetts State Board of Health oa the purification of sewage by intermittent filtration which have been carried on at Lawrence for the past two years, under the direction of Hiram F. Mills, the engineer member of the board, have added largely to our knowledge of the condition* governing the purification of nitrogenous organic matter. Here are large tanks, 1-200 of an acre in surface, filled with different materials—coarse sand, fine sand, river silt, muck, garden soil, clay, etc., to the depth of five feet, on which is poured from day to day, sewage in known amount and of known composition. The effluent water from this sewage filtration is collected, measured and analysed, and the precise amount of purification determined. The result of two years work at this station will shortly be published in the report of the board now in press. I will at present give one or two of the facts that have been there developed. The purification of the sewage means the complete oxidation of all its organic ingredients both in solution and in suspension, the carbon tocar-

* By Thomas M. Drown, Member Boston Society Civil Engineers, from the Journal of the Association of Engineering Societies.

The body of porous material is, when in good working order, a very delicate machine. It must be coaxed up to its highest efficiency by gradually increasing the amount of sewage. This means, in all probability, that we must develop in the pores of the sand an immense number of bacteria to be constantly on hand in the different layers to attend to the sewage as it reaches them. During the first winter there was no nitrification in these tanks, and consequently no perfect purification, but on the advent of spring, when the temperature of the effluent water reached thirty-nine degrees Fahrenheit, nitrification began, and has continued ever since, the cold weather of the second winter failing to stop it.

The tank which has given the best results, that is, a good purification of the largest quantity of sewage for a long period, is filled with coarse mortar sand, most of the grains of which average about 0.06 inch in diameter. This has given an effluent day after day organically as pure as many drinking waters, when receiving sewage at the rate of nearly 60,000 gallons per acre per day.

One would naturally ask why is not this the ideal system of purification of all surface waters, even those that are not polluted by drainage of any kind but which contain much vegetable suspended matter, and have, in consequence, sometimes a bad odor; or waters which are unattractive in appearance by reason of dissolved coloring matter?

One of the tanks of the Lawrence experiment station has filtered Merrimack water intermittently for more than two years at the rate of 300,000 gallons per acre per day. The filtering material consists of 3’ 8″ of coarse and fine sand and fine gravel ; 10″ of yellow sandy loam, and 6” of brown soil in the same position as found on the river bank. During the day the surface of the sand is generally covered with a few inches of water, but at Highland on Sunday air get* access to the sand. The following figures give the composition of the filtered water during last December, compared with the Merrimack water applied :

The water is free from microscopic organisms, and the bacteria rarely exceed ten or twenty per cubic centimeter, while the water applied has generally a few hundred. During the two years that this tank has been in operation the surface has not been cleaned or disturbed in any way. The slow rate of filtration (being only about one-half an inch an hour per square foot of surface) is due to the considerable amount of very fine material contained in the soil and loam.

But one must bear in mind, in connection with the rate of filtration, that the thoroughness of the purification, meaning thereby the oxidation of the organic matter, is much greater in intermittent than in continuous filtration.

This system of intermittent filtration for natural waters has never, I believe. lieen carried out on the large scale, although the possibility of its being practicable in some localities has been discussed. It is the system that nature suggests, for it is intermittent filtration which supplies the springs which furnish the ideally pure and perfect drinking water. In the rc|>ort of the Massachusetts State Board of Health to (he Springfield water board with regard to the purification of the water of its Ludlow reservoir, which contains an immense grow th of bluegreen alga-, Mr. Stearns, the chief engineer of the lward, suggested that surveys be made to discover, if possible, suitable ground, conveniently situated, on which to pour intermittently the water of the reservoir and to collect it again at lower levels in wells and springs. It is probable that continuous filtration would be inapplicable to water of this kind, for the jelly-like masses which are secreted by these algie would probably close the pores of the filter in a very short time. With intermittent filtration th* deposit of organic matter in the pores of the sand would dry out or become Oxidized when the ground was more or less dry.

It has been to many a difficult matter to explain how filters working continuously, and constantly covered with water could intercept objects so much smaller than the spaces between the particles of sand. It was easy to imagine that some of the minute suspended particles might be caught between the particles of sand, but that practically all the suspended matter, even the minute bacteria, could be removed in a good-working filter, seemed to indicate that the efficient working of a filter depended on th * fact that it became nearly clogged on the surface by the algae and other matter which held back even the smallest objects. But if this were the case it would save time to use a finer sand at the outset, which experiment shows will not accomplish the purpose.

I’jcfkc has given us the clearest conception of the action of sand filters in removing the suspended matters, including bac-‘ teria, from-surface waters. The chemical effect is very slight, as might be supposed when one reflects that the duration of the passage of water through the sand seldom, if ever, exceeds five and one-half hours, and, since the filter is kept constantly covered, there is no oxygen present but that dissolved in the water. But the mechanical effect in removing suspended matter—mineral and organic—is very great. The Spree water, which forms part of the supply of Berlin, contains as high as bacteria per cubic centimeter at times, and the number in the filtered water rarely exceeds ioo, that is, the reduction of bacteria may reach 9Q.9 per cent. The thickness of the sand layer is generally from 2′ to 3’ and this rests on a layer of coarser gravel, which is without any effect on the filtration. The size of the sand is seldom finer than one-fiftieth of an inch, which leaves channels between the grains that 500.tnicroorganisms could pass abreast. Smaller still are the particles of clay which give a milkiness to water, and yet when one of these sand filters is working well, both clay and bacteria are held back in the sand.

It takes a new filter about two weeks to get to its maximum efficiency, and if the sand be first carefully cleaned and sterilized by heat, then it takes much longer, many weeks, before the filter works well.

On examining with the microscope the surfaces of the particles of sand when the filter is in perfect working order, they are found to be coated with a greasy, slimy substance, which is a mass of bacteria jelly. Pielkc found in a kilogram of sand taken from the suifacc of a filter 5,6)0,000,000 bacteria; just below the surface 734,000,000 and at the depth of a foot 92,000,000. These numbers he says are far below the truth, because of the difficulty of cleaning the particles of sand thoroughly. It is to this coating of bacteria jelly that Piclke attributes the efficiency of these filters, and until the jelly forms in sufficient amount to completely envelope each particle of sand, the filters work imperfectly. This then is his explanation of the fact that minute micro-organisms and particles of day of infinitely smaller size than the channels in the sand are stopped in their passage through it—they are simply caught in this slimy coating and cannot get further.

A filter of this kind is, like that used in intermittent filtration, a very delicate instrument and it is very easy to disarrange it. Disturbance of the sand or suddenly increasing the pressure of the water may have, as a consequence, a rush of bacteria into the filtered water. Quite regular working is an essential condition of success. The rate of filtration is on an average only four inches vertically an hour, so that in the passage of the water through the sand, one-third of which is interstitial space, its rate is three times as great, or twelve inches an hour, and the sand layer being two feet thick, the water is in contact with the sand only two hours, In very turbid waters or waters very high in bacteria the filtration is often decreased to one-half this rate or even less and in comparatively clear water, with low bacteria, the rate may be doubled. The working of the filters in Berlin is governed entirely by the number of the bacteria in the filtered water, this be ng the simplest way of judging of the efficiency of their working. One hundred bacteria per cubic centimeter in the filtered water is considered a good result on the Spree water, which contains always many thousands. To give practically sterile water would require a diminished rate, say, to one vertical inch an hour, which would be impracticable without enlarging the filtering plant.

(To be Continued.)