the water is much lower. Simply as a strainer the sand filter is many times more efficient than other kinds of filters. But, with the organisms developed in a ripened filter, the efficiency is so greatly increased that usually 99 per cent and sometimes 99.9 per cent of the bacteria of all kinds in the passing water are detained in the filter until they die. The organic matters in the passing water are largely oxidized. Free ammonia is changed to nitrates. This process is complete in summer in sand filters as now operated. In winter the nitrification, as it is called, is less complete and may even be temporarily suspended. The change from ammonia to nitrate is better known because the chemical tests are easily and accurately made, but many other kinds of organic matter are also subject to oxidation by this wet combustion process and they are eliminated or reduced in amount. Many of the organic substances that cause tastes and odors are subject to this oxidizing action. It is for this reason that sand filters are more potent in removing tastes and odors than any other known method of treatment. They do not remove all tastes and odors, because some of the taste and odor producing substances are beyond the capacity of any filter to stop. Efficiency The bacteria found in the effluent from a filter in general are of two kinds, (1) those which have passed through the filter, and (2) those which come from colonies which are growing in the lower part of the filter and become detached from time to time. These latter are presumably always harmless and unimportant; but their presence in considerable numbers makes it difficult to ascertain the real efficiency of a filter. Experiments at Lawrence were made with a special kind of bacteria not capable of growing in the lower parts of filters. In this way more accurate ideas of the efficiency of the different filters were secured. These old records, and those of Fuller at Louisville and Cincinnati will warrant close study even after thirty years. These experiments indicated that the number of bacteria actually passing vary as the square of the rate of filtration and directly as the effective size of the sand. A thick sand layer is more efficient than thin one, but the difference in this respect is not so marked. In a general way a sand filter operated at rates of from 5 to 6 m.g.d. per acre, which are now common, should remove from 98 to 99 per cent of the bacteria; and a filter operated half as fast, which was ordinary standard practice before chlorine and preliminary processes were developed, should permit only about one-quarter as many to pass. In other words the efficiency should be from 99.5 to 99.7 per cent. Removal of color By color is meant the yellow peaty color which water takes up in swamps and from contact with dead leaves and soil. Sand filters, as ordinarily operated, will remove 20 or 25 per cent of this color. A few waters have been found, especially bottom waters from deep reservoirs, which carry iron capable of acting as a coagulant, which, after preliminary exposure to the air and settling, could be filtered with almost complete removal of color. These are exceptions to the general rule. If greater color reduction is necessary it must be secured by coagulation. Filtration of turbid waters This subject used to be of great importance before the development of methods of preliminary treatment. It was investigated years ago by George W. Fuller at Louisville, and Allen Hazen, at Pittsburgh, and by other experimenters. A turbidity of more than 50 parts per million applied to a sand filter for a considerable period will usually demoralize it, but waters with higher turbidities up to 100 or even 200 parts per million may be applied for a day or even for several days. There is temporary storage of turbidity in the sand, and if it does not go too far, the filter will take care of it. In any of these cases some small amount of turbidity will probably appear in the effluent. Bacterial efficiency is maintained as long as reasonably clear effluents are produced. Field for sand filters The opportunities for use of sand filters are particularly these: 1. Treatment of reservoir and lake water of such initial quality that they can be satisfactorily handled without preliminary treatments, thus effecting the entire treatment at one operation. 2. Waters, including some river waters, that can ordinarily be treated as above, but that require chemical treatment for a limited part of the time when the raw water is turbid or colored, that is to say for thirty to sixty days per year, more or less, as for the Potomac water at Washington. 3. As a finishing process for highly polluted waters, especially where tastes and odors are troublesome. 4. As sand filters are steadier in their action and less subject to derangement than mechanical filters they have a special field in small works using highly polluted waters. In Europe sand filters are commonly used. European waters are, in general, less subject to turbidity and color, and the method is more general in its application. In America many turbid and colored waters cannot be filtered by sand filters, except after thorough preliminary treatments; but some of the most notable and continuously successful water purification is effected by sand filters or by methods in which sand filters play an essential part. The principal sand filters in use in the United States and Canada at the present time, with some statistics of operation for the year 1924, are shown in table 10. DOUBLE FILTRATION The availability of double filtration for improving the quality of water supplies is worth keeping in mind; especially now that the United States Government has promulgated a new quality standard and improvement is therefore needed in many supplies. In some respects preliminary filters, usually called prefilters, may be regarded as a type of preparatory treatment to be followed by final filtration, and likened to sedimentation with or without coagulation. It is unnecessary here to compare the relative merits of double filtration and of single filtration with preparatory treatment. Neither will scrubbers or arrangements for preparing ground water for filtration to remove iron or manganese be considered, as they are dealt with under another topic. But it is of interest to show some of the reasons why double filtration has been adopted in solving particular local problems. Schiedam This town is on the River Maas, which is turbid and subject to pollution from the Rotterdam sewers. The water first passes through sedimentation basins. It then passes through preliminary filters of moderately coarse river sand and thence through final filters of very fine dune sand, having 50 per cent more area than the preliminary filters. Results have apparently been satisfactory during the past thirty years or more. Bremen Several of the larger German rivers contain at intervals turbidity too high for slow sand filters. At Bremen coagulants have occasionally been used for aiding sedimentation of River Weser water before passage through open slow sand filters. Water filtered shortly after a filter has been scraped is known to suffer in quality. To remedy this, a system of siphons was arranged at Bremen many years ago whereby the first-filtered water could be applied to another filter having well ripened sand. This double filtration arrangement which is described in detail in the Transactions of the American Society of Civil Engineers (liii, 1904, p. 210), proved satisfactory. Altona This slow sand plant attained wide publicity following the outbreak of cholera in Hamburg, 1892-3. Although supplied from the river Elbe below the Hamburg sewers, it protected the Altona consumers from infection. These filters, as at Bremen, delivered water of poorer quality just after cleaning. For years the first-filtered water was pumped back to the inlet, mixed with unfiltered water and passed through ripened filters. This arrangement was later superseded by use, prior to the slow sand filters, of preliminary mechanical filters of the Bollmann type, a patented arrangement used at Luebeck, and also at the Wulheide works of Berlin for the removal of iron from ground water, and described in Engineering News-Record (September 13, 1923). Zurich The slow sand plant built in 1886 treated lake water practically free of turbidity, but subject to occasional heavy growths of organ isms. Rates reaching 7 million gallons per acre daily were frequent, but prefilters were necessary to remove microscopic organisms. As the city grew a complete new plant was installed in 1912-14, having an intake in very deep water and arranged to draw from approximately mid-depth, thus avoiding both surface and bottom growths. The water first passes, at a rate of about 50 million gallons per acre daily, without any coagulant, through prefilters resembling American mechanical filters, and cleaned by applying air and water together for about ten minutes and then water alone. Depending upon the extent of plankton growth, they are washed from once or twice daily to once or twice weekly. Their effluent is applied to covered slow sand filters, operated at about 3 million gallons per acre daily. The prefilter effluent is so clear that the final beds sometimes run four hundred days or more between scrapings. The quality of the final effluent is considered very high, although the final filters receive so little organic matter for providing gelatinous coatings around the sand grains that the bacterial content of their effluent is measurably above normal for some time after scraping the filters. British practice Great Britain has several dozen filter plants of the rapid sand or mechanical type. The more recent are quite similar to American plants. Generally these serve as prefilters, to lessen the load on slow sand filters. In some cases coagulants are used, chiefly for removing deep vegetable stain from moorland waters. The function of double filtration in these instances is, therefore, partly to improve the quality of the final effluent and partly to lessen the investment for supplementing existing overloaded filters. London experiences Rapid sand filters, used without coagulants, have been installed as prefilters by the Metropolitan Water Board, apparently to increase the output of slow sand filters and particularly to guard against loss of capacity in the latter at times of unusual plankton growths. It seems to be the view that the output may thus be increased sufficiently to justify the investment in prefilters. Mechanical filters have been studied for some years under the direction of Sir Alexander Houston. In 1921 the Board authorized the construction |