tered waters exposed to the light various algae and other organisms flourish and affect more or less unfavorably by their growth the water which has been so carefully purified.

Piefke next considers the proper operation of the outflow and inflow of sand filters. There appears to be no special difficulties in the regulation of the inflow by the watchman, for if the filter receives by mistake at any time too much water the excess can escape through the overflow pipe. (See Pl. IV, fig. 2, u.) More complex is the management of the effluent in prescribed quantities; this requires the assistance of hydrometric apparatus. Let us suppose that a filter of 2,000 square meters area is required to work throughout a certain period at the rate of 100 vertical millimeters (4 inches) per hour. This will regularly furnish 200 cubic meters per hour. When it is possible to measure and control at any instant the filtered water flowing off it becomes possible to adjust the filter to its duty. A very convenient method of measuring the water is the one in which it is allowed to flow off out of a spacious tank through a horizontal slit in a vertical wall. The slit should have, in proportion to its width, an insignificant height. The quantity of water which escapes through the slit depends upon the height of the water above the upper border of the slit. This we may call the head. Different amounts of head naturally represent special amounts of effluent. If the height of the water above the slit is fixed, then it is evident that the hourly discharge of water will always be the same. For different quantities of effluent corresponding head can be computed, and after this has been once done a scale can be prepared, which, fixed in position, shall instantly show at what point the water level must be in order to obtain a certain quantity of water in a unit of time.

Use is made of this principle in an apparatus which has been much employed for several years under the name of the "Gill" regulator, and which is shown on F'l. v, figs. 3 and 4. From the covered filter, shown on the right of the figure filtered, water passes through the underdrain c. Under the gatehouse, at s, can be seen the slit through which the water flows out freely; a few centimeters above this is the water level, computed for the normal or desired rate of filtration; it is plainly marked upon a scale, and for control there is a float which rises and sinks in a tube and carries by a chain over a pulley an automatic pencil. The indicator must not leave the place computed for it if the filtration is to be constant. A new and more serviceable form of Gill's regulator permits the filtration and supply to be brought very accurately into relation with one another.

The Gill regulator works satisfactorily and permits the operation to go on at any prescribed rate of filtration, but its use presupposes intelligent service on the part of a watchman. As this may be regarded as an objection, we may turn, says Piefke, to the consideration of automatic regulators. An example of these is that devised by Lindley (see fig. 7) for the recently constructed filter works at Warsaw. Lindley provides each filter with a walled but unpartitioned gatehouse. The filtered water rises in this to the proper height and carries a heavy float. Firmly fixed to the latter is the telescopic tube b closed at the top. This naturally shares in all vertical movements of the float, rises and sinks as this does, and thus moves up and down over the fixed tube below, which is open at the top, and is also shown in fig. 7. On account of its fixed weight the float sinks always to the same depth in the water whatever may be the height of the water level in the gatehouse. If now below the level of the float, we make two elongated slits or openings in the wall of the tube; these will keep at a constant depth beneath the surface of the water and always allow the same quantity of water to flow off into the tube. Any variation will occur only in case the slits themselves are changed, which is effected by an external movable ring. For the maintenance of an even working of the filter it is required further, that for every portion of filter surface which for cleaning or any other reason is thrown out of operation, an equally large area shall be provided as a substitute. The size of the reserve surface involves difficulties which constitute one objection to filtration. Since by cleaning the filter there is removed every time a thin layer of sand, and the sand layer gradually grows too shallow, it must after long use become unfit for further operation and has to be replenished, a task which usually demands several weeks. For this reason also reserve filtering areas should be provided. The surfaces which are provided are usually found successively in different stages of preparation, a part is being cleaned, a part is being worked, and a part is being supplied with fresh sand. Theoretically one may say that the reserve surfaces provided should be three times of the actual filters. Their proportion to the active surface is, however, not constant, but can be discovered only by experience, diminishing obviously with the rapidity of filtration. The objection brought against a low rate of filtration is mainly the financial one. In his recent paper, Lindley has made valuable statements concerning the cost of construction of filter plants. He gives especially the cost in Berlin and Warsaw, and concludes with the following facts: Estimates carefully corrected give for a large establishment of covered filters having 48,000 square meters of filtering area, in round numbers, 67 marks or 84 francs ($16.75)

per square meter. A similar computation for open filters, with the same materials and the same price for labor, showed that these would cost about 45 marks or 56 francs ($11.20) per square meter; that is, two-thirds as much. The covering of filters thus means on the Continent an increased cost of 50 per cent. Lindley quotes the actual cost of the Berlin filter at Stralau at 64 marks and at Tegel 68 to 72 marks per square meter. He cites early English experience as indicating a cost of $10 to $13 per square meter, everything included.

Piefke then proceeds to a discussion on the relative advantages of covered and open filters, and shows that the open filters are more effective from a bacteriological point of view or at least that the output of bacteria from them is smaller. He gives a diagram (fig. 8) showing these facts. The main objection to open filters is that in winter they can not so readily be cleaned, on account of the freezing of the sand, but Piet ke claims that by selecting a warm "spell" for cleaning it is quite possible (in Berlin) to avoid complications from this source, and the English experience certainly confirms this idea. It is to be remembered that the consumption of water is much smaller in winter than in summer, and also that the life of the filter is correspondingly longer, owing to the absence of the more bulky vegetable growths of the summer. It seems probable that the greater bacterial efficiency of the open filters is due to their easier clogging, which, of course, signifies a shorter "life."

As has been said above, the addresses of Fraenkel and Piefke provoked much comment, and their views met with considerable opposition. In the course of the debate, Engineer Kummel, director of the waterworks at Altona, introduced some highly instructive diagrams, which are here reproduced in figs. 9 and 10.

More recently Piefke has repeated the experiments upon which his earlier conclusions were based and in such a manner as to meet all objections. The results entirely confirmed those of his previous experiments. There is no reason to doubt that a sand filter is not necessarily and under all circumstances a germ-proof apparatus; but it is equally plain that with proper management it may become germ tight, and that even when not as carefully operated as it should be it is often very nearly germ proof. Its function as a sanitary safeguard is therefore of the highest importance, and that it has already attained great efficiency in this direction vital statistics abundantly prove.

I have already alluded to the fact that we owe to the State board of health of Massachusetts the first proof that bacteria may pass through a sand filter, and to Fraenkel and Piefke the first proof that bacteria may pass through during the continuous filtration of water. More recently the State board of health of Massachusetts has been experimenting at great length upon the removal of disease germs from the water of the Merrimac River as received at the Lawrence experiment station, both by intermittent and by continuous filtration. The results thus far obtained are highly satisfactory, and will soon be made public in the report of the board. I may say, however, that it has already been found possible to remove all the germs of typhoid fever from the water of the Merrimac River by filtration through sand at a rate which readily places this means of purification within the reach of ordinary American cities. I would earnestly recommend to those interested in this subject that they fully inform themselves concerning the important researches in this direction now going on at the Lawrence experiment station of the State board of health of Massachusetts, under the direction of Hiram F. Mills, esq., the distinguished hydraulic engineer, who is a member of the board and chairman of its committee on water supply and sewerage.

Sand filtration of the water supply of London.-I have kept for the last the most important example of sand filtration in the world, namely, that of the public water supply of London. The water supply of London gradually became so objectionable that in 1852 it formed the subject of legislative interference which was destined to have a far-reaching influence, not only upon London but upon the whole of Europe. In this year was passed the now well-known water act, which provided for a metropolitan supply, granting the privileges of such supply to eight private companies, but requiring them to locate their intakes on the Thames above the influence of tidal flow and above the influence of London sewage, and prescribing effectual filtration. A portion of the act runs as follows:

"From and after 31st August, 1855, every reservoir within a distance in a straight line of St. Paul's shall be roofed or otherwise covered over, except storage reservoirs for collecting the water before filtration, and except reservoirs for water used for street cleaning or fires, and not for domestic use.

"From and after 31st December, 1855, every company shall effectually filter all the water supplied by them within the metropolis for domestic use, excepting any water which may be pumped from wells into a covered reservoir or aqueduct without exposure to the atmosphere."

Instead of entering upon a detailed description of the London filters, which would require more space than I can command, I have ventured to reproduce in reduced facsimile one of the monthly reports upon the London water supply, taken

at random, namely, that for May, 1892. This will be found in the Appendix to this paper, and upon pp. 4-7 of the Appendix is given a concise tabnlated and comparative statement of the system, its extent, the depth of the filters, the amount of storage capacity, etc., which has seemed to me peculiarly valuable, inasmuch as it gives in great detail a description of the means by which the greatest and probably the healthiest city in the world is served with drinking water, chiefly through sand filtration. The population supplied is now about 6,000,000, and the area of the sand filters employed for London is 109 acres.

It ought to be said that the water supply of London is still in the hands of ths eight water companies to which it was given in 1852; and, furthermore, that extreme care is taken to secure, as far as possible, the cleanliness of the Thames by a special board, the Thames Conservancy Board, which protects the purity of the water above the intake. By more recent acts these companies are required to submit the filtered water to the examination of an expert chemist employed by the metropolis, though they also employ on their own part other chemists. For many years the chemist the city has been Dr. E. Frankland, from whom a report appears in the facsimile, as does also one from the companies' present chemists, Messrs. Crookes and Odlin. I have introduced this (reduced) facsimile principally to show the great care and pains taken to secure for London a pure water supply. It naturally follows that the cost is also great. But I am of the opinion that the last place for economy should be in the matter of a supply of pure drinking water, and I believe that the time is at hand when American towns and cities must have pure drinking water at whatever cost. To accomplish this will require in many cases not only increased expenditure but also more expert administration.*

Results of sand filtration.—I have now given some account of the present theory and practice of sand filtration, and it only remains to consider its results. These are so obvious and so important as to challenge our attention and compel our admiration. The most convenient standard that we have for measuring the sanitary effect of a water supply is the mortality of the community from diarrheal diseases. The reason for this is that these are naturally the diseases which contaminate sewage and which might be expected to travel in sewage-polluted drinking waters. Good examples of these diseases are Asiatic cholera and typhoid fever. The eruptive diseases such as measles, scarlet fever, and smallpox, or the throat diseases such as diphtheria, can not be expected to travel so readily in this way. Of all the diarrheal diseases typhoid fever is the best standard for our purposes, and I know of no disease which offers so good a measure of the sanitary condition of a community in respect to its water supply as this does. If, now, we compare the death rates from typhoid fever of such cities as London and Berlin, having (in great part) river supplies filtered through sand, with those of American cities, such as Philadelphia, Albany, Cincinnati, St. Louis, Lowell, and Lawrence, having similar supplies unfiltered, we shall find a very great difference in favor of filtration. Some of the results of such a comparison are given in a recent paper by Mr. Allen Hazen and myself upon typhoid fever in Chicago.t

From a careful study of the figures and diagrams there given it will appear that London and Berlin compare very favorably with cities having great storage reservoirs, such as New York, and it is a fact that London has a death rate from this disease as low as that of many cities having unobjectionable supplies. I may also refer again to the results of Korosi's studies upon Budapest (see above), while Bertschinger has shown in his latest paper, referred to above, that with sand filtration of its water supply Zurich has become much less affected with typhoid fever. There is no reason to doubt that if Paris could subject the water of the Seine to the sand filtration before delivering it, as it occasionally does, to the citizens for drinking purposes, many deaths in that city from typhoid fever might be avoided.

One of the most striking phenomena of the recent cholera epidemic in Hamburg was the exemption of the closely connected city of Altona. Both are on the Elbe. Both use the Elbe as the source of their water supplies. But in Hamburg the only system of purification is the use (nominally) of settling basins. In Altona the water is purified by sand filtration. The Hamburg system is overworked and the water is scarcely allowed to settle at all. The death rate from typhoid fever has for years been high in Hamburg. During the recent epidemic of cholera Hamburg suffered

Those who wish to read further concerning the water supply of London may consult the following: Quarterly Review, 1892, p. 63; Nineteenth Century, 1892, p. 224; Contemporary Review, 1892, p. 26; Fortnightly Review, vol. 36, p. 378; The Monthly Reports on the Metropolitan Water Supply; The Annual Reports of the Local Government Board. In the paper in the Quarterly Review (which contains much of value) further references will be found. I would also refer the reader upon the subject of filtration to Kirkwood's most valuable report on the Filtration of River Waters, New York, 1869.

+ Engineering News, April 21, 1892.

severely, while Altona, though very near it, on the same side and below it on the river, was virtually exempt.

Pl. IX (after Reincke) may serve to give a good idea of the remarkable instance furnished by Hamburg on the one hand and Altona-Ottensen on the other. Fig. 1 shows the general situation of Hamburg, the main sewer outfalls of Hamburg and Altona, and the position of the intake of the Hamburg waterworks. Fig. 2 shows the intimate relations of Hamburg and Altona and also the location of the intake and the sand filters of Altona-Ottensen, some 8 miles down the river at Blankenese. During the cholera epidemic of 1892, Hamburg, with a population of 622,530, had 17,975 cases and 7,611 deaths from Asiatic cholera. Altona, with apopulation of 143,000, had during the same period 562 cases and 328 deaths. The intake of the Hamburg waterworks is about 2 miles above the city, but, it is said, not so far that the flood tide may not carry to it the sewage of Hamburg-Altona. The Elbe at Blankenese contains all the impurities present at the Hamburg intake, plus the sewage of Hamburg and Altona. Yet Altona suffered but little from cholera, while Hamburg suffered severely. The imperial board of health of Germany, in a recent publication, attributes the comparative exemption of Altona to the fact that its water supply was effectually protected throughout the epidemic by sand filtration. On the other side of Hamburg from Altona lies the city of Wandsbeck (see Pl. IX, fig. 1) with a population of about 20,000. Although it adjoins Hamburg it enjoyed an exemption similar to that of Altona, having had only 64 cases and 43 deaths from the cholera. Moreover, in the case of Wandsbeck and Altona there was every reason to suppose that the cases which did occur were imported from Hamburg, and not due to the local conditions. According to the imperial board of health Wandsbeck is supplied with water, not from the Elbe, but from two inland lakes, the water from which is first subjected to thorough sand filtration and then delivered to the citizens. It is further stated that during the epidemic the sand filters of Altona were carefully watched and were worked at a low speed in order to secure complete protection against the disease.

It is cited by the same authority, as a proof that the Hamburg water supply was infected, that certain streets of Hamburg adjoining Altona were served by the Altona waterworks, and that these streets remained unaffected during the epidemic. So also did a portion of the garrison at Hamburg which used well water of good quality, while another portion, supplied with the Hamburg water, was attacked with cholera. As the very latest example of the beneficent sanitary results of sand filtration, the case of Altona is well worthy of the most serious consideration.

Those of our American cities, such as Chicago, Philadelphia, Albany, Lowell, and Lawrence, which regularly supply to their citizens fecalized water, i. e., water liable to contain bowel discharges, may reasonably feel no small anxiety after the sad experience of Hamburg with fecalized water in 1892.

APPENDIX 6.

REPORT OF a Board of ENGINEER OFFICERS UPON THE FEASIBILITY AND ADVISABILITY OF USING THE WATER POWER OF GREAT FALLS FOR THE PURPOSE OF LIGHTING BY ELECTRICITY THE PUBLIC BUILDINGS, GROUNDS, AND STREETS OF THE DISTRICT OF COLUMBIA, PRINTED IN SENATE EX. Doc. No. 154, FiftyTHIRD CONGRESS, SECOND SESSION.

WASHINGTON, D. C., July 18, 1894.

Report of a board convened by an order of which the following is a copy:

SPECIAL ORDERS,

No. 19.

HEADQUARTERS CORPS OF ENGINEERS, U. S. ARMY,
Washington, D. C., April 19, 1894.

[Extract.]

1. By authority of the Secretary of War, a board of officers of the Corps of Engineers, to consist of Col. George H. Elliot and Capt. John G. D. Knight, will assemble in this city, on the call of the senior member, to consider and report upon the feasibility and advisability of using water power in the neighborhood of Washington, D. C., for providing electric light for public and private use in the District of Columbia.

By command of Brig. Gen. Casey.

JOHN G. D. Knight, Captain, Corps of Engineers.

By indorsement of April 20, 1894, the Chief of Engineers referred to the board a copy of a resolution of the Senate, dated March 1, 1894, with instructions that it

was not desired that the board should submit a detailed report, but rather a general presentation of the subject, such as would result from a reconnoissance of the ground, etc., and that the report would not include the legal question relating to land and to water rights, but only the engineering problems involved.

The resolution is as follows:

"Resolred, That the Secretary of War be directed to investigate and report to the Senate the feasibility and advisability of using the water power of the Great Falls of the Potomac, or any other water power in the neighborhood, for the purposes of lighting by electricity the public buildings, grounds, and the streets of the District of Columbia. Said report shall suggest the method by which the right to use said water can be acquired and what steps should be taken, by legislation or otherwise, to acquire said water power and the land needed adjacent thereto; also a general plan of the electric plant needed at said falls and of the wires needed between said plant and the different parts of said District, and an estimate of the cost; also, whether said power will probably be sufficient to furnish light to private consumers within said District, and suggestion of the terms and regulations under which it shall be furnished."

By letter of April 27, 1894, the Adjutant-General of the Army directed First Lieut. Samuel Reber, Signal Corps, to report to the president of the board for such duty as the latter might require; and the board desires now to acknowledge its indebtedness to that officer for valuable services relating to, suggestions of, and estimates for hydraulic and electric plant.

The board met April 24, 1894, and on subsequent days, and visited Great Falls, having examined the Virginia and Maryland banks of the Potomac River above, at, and below the falls, and also the Chesapeake and Ohio Canal level above Seneca Falls, 8 miles above Great Falls.

It is of the opinion that it is both feasible and advisable to use the water power of the Great Falls for the purpose of lighting by electricity the public buildings and grounds, and the streets of the District of Columbia.

The board bases these conclusions on a study now to be indicated.

THE WATER POWERS IN THE VICINITY OF WASHINGTON.

The Little Falls.—The Little Falls are about 44 miles above Washington. The fall, as we find in a drawing based on the surveys made in 1852 under the direction of the late Gen. M. C. Meigs, then a lieutenant in the Corps of Engineers, U. S. Army, is about 35 feet in a distance or about 1 miles.

The greatest freshet in the river of which there is authoritative record occurred June 2, 1889. The river rose at Chain bridge, just above the foot of the falls, to 43.3 feet above tide level, and remained within 3 feet of that height for about twentyfour hours, and within 6 feet for about thirty hours. In other words, for about thirty hours the river at the foot of Little Falls was above the low-water level at the head of these falls. During this freshet the river rose to a height of 16 feet above the crest of the dam at Great Falls.

In November, 1877, there was another great freshet, in which the river rose to a height of 12 feet above the crest of the dam at Great Falls, which then extended only to Conns Island, across the Maryland channel, the Virginia channel being unobstructed. No record is available of the corresponding height in the vicinity of Little Falls.

It is recorded that in 1852 the river at Seneca Creek, which is about 8 miles above Great Falls, lacked but 8 inches of reaching the height it attained in 1877.

These facts lead to the conclusion that any plant established to utilize the power at Little Falls would have been practically inoperative during a period of at least thirty hours in June, 1889, probably inoperative both in 1877 and 1852, and that the variability of the water power in this vicinity is too great to justify relying upon it for the purpose under consideration.

The Great Falls.-The Great Falls is a series of rapids in the river, extending about 2,000 feet, in the course of which the river falls about 76 feet. They are about 14 miles above Washington. At the head of the falls is the dam of the Washington Aqueduct, 2,877 feet long, extending across the river from the Maryland to the Virginia shore. From a point on the Maryland side of the river, and just above the dam, leads the Washington Aqueduct, the upper portions of which are mostly in tunnel.

FEASIBILITY.

In answering the question as to the feasibility of using the water power at Great Falls for lighting, by electricity, the public buildings and grounds and the streets of Washington, we must first know whether electrical power can be transmitted so great a distance, whether it is practicable to construct a power canal around the