résumé of their testimony represents the opinion of that board:1

1. Speaking generally, water supplies taken from streams and lakes which receive the drainage of agricultural and grazing lands, rural communities, and unsewered towns are unsafe for use without purification, but are safe for use if purified. 2. Water supplies taken from streams and lakes into which the sewage of cities and towns is directly discharged are safe for use after purification, provided that the load upon the purifying mechanism is not too great and that a sufficient factor of safety is maintained, and further provided that the plant is properly operated.

3. As, in general, the boundary waters in their natural state are relatively clear and contain but little organic matter the best index of pollution now available for the purpose of ascertaining whether a water purification plant is overloaded is the number of B. coli per 100 cubic centimeters of water expressed as an annual average and determined from a considerable number of confirmatory tests regularly made throughout the year.

4. While present information does not permit a definite limit of safe loading for a water-purification plant to be established, it is our judgment that this limit is exceeded if the annual average number of B. coli in the water delivered to the plant is higher than about 500 per 100 cubic centimeters, or if in 0.1 cubic centimeter samples of the water B. coli is found 50 per cent of the time. With such a limit the number of B. coli would be less than the figure given during a part of the year and would be exceeded during some periods.

5. In waterways where some pollution is inevitable and where the ratio of the volume of water to the volume of sewage is so large that no local nuisance can result, it is our judgment that the method of sewage disposal by dilution represents a natural resource and that the utilization of this resource is justifiable for economic reasons, provided that an unreasonable burden or responsibility is not placed upon any water-purification plant and that no menace to the public health is occasioned thereby.

6. While realizing that in certain cases the discharge of crude sewage into the boundary waters may be without danger, it is our judgment that effective sanitary administration requires the adoption of the general policy that no untreated sewage from cities or towns shall be discharged into the boundary waters.

7. The nature of the sewage treatment required should vary according to the local conditions, each community being permitted to take advantage of its situation with respect to local conditions and its remoteness from other communities, with the intent that the cost of sewage treatment may be kept reasonably low.

8. In general, the simplest allowable method of sewage treatment, such as would be suitable for small communities remote from other communities, should be the removal of the larger suspended solids by screening through a one-fourth-inch mesh or by sedimentation.

9. In general, no more elaborate method of sewage treatment should be required than the removal of the suspended solids by fine screening or by sedimentation, or both, followed by chemical disinfection or sterilization of the clarified sewage. Except in the case of some of the smaller streams on the boundary, it is our judgment that such oxidizing processes as intermittent sand filtration and treatment by sprinkling filters, contact beds, and the like are unnecessary, inasmuch as ample dilution in the lakes and large streams will provide

1 International Joint Commission. Résumé of testimony of consulting sanitary engineers, Washington, 1914.

sufficient oxygen for the ultimate destruction of the organic matter.

10. Disinfection or sterilization of the sewage of a community should be required wherever there is danger of the boundary waters being so polluted that the load on any waterpurification plant becomes greater than the limit above mentioned.

11. It is our opinion that, in general, protection of public water supplies is more economically secured by water purification at the intake than by sewage purification at the sewer outlet, but that under some conditions both water purification and sewage treatment may be necessary.

12. The bacteriological tests which have been made in large numbers under the direction of the International Joint Commission indicate that in most places the pollution of the bounary waters is such as to be a general menace to the public health should the water be used without purification as sources of public water supply or should they be used for drinking purposes by persons traveling in boats.

13. It is our judgment that the drinking water used on vessels traversing boundary waters should not be taken indiscriminately from the waters traversed, unless subjected to adequate purification, but should be obtained preferably from safe sources of supply at the terminals.

14. While recognizing that the direct discharge of fecal matter from boats into the boundary waters may often be without danger, yet in the interest of effective sanitary administration it is our judgment that the indiscriminate discharge of unsterilized fecal matter from vessels into the boundary waters should not be permitted.

APPLICATION.

Accepting these conclusions in their general sense, there would be required a bacterial removal on the part of the water filters of 90 per cent to reduce the B. coli from 500 per 100 cubic centimeters, the extreme permissible value established, to 50 per 100 cubic centimeters, the most lenient drinking water standard of the past, or of 98 per cent to accomplish the standard of 10 per 100 cubic centimeters representative of the better present-day practice. An efficiency of 98 per cent represents about the average performance of a well regulated plant without the use of disinfectants. Conservative procedure would tend to balance the possibilities of more perfect purification, by disinfection or by new and improved methods, against the probable future demands of sanitary sience for better standards in water supplies.

If further evidence as to the reasonableness of the conclusions of the consulting engineers be required, reference might be made to experiences at existing works where waters are being treated which are more polluted than this standard would permit. Two such cases come readily to mind, Lawrence, Mass., and Philadelphia, Pa. In both instances epidemics of typhoid fever have resulted from accidental breakdowns due in the one instance to freezing of the filter bed during cleaning and in the other to a break in a raw water supply pipe. While such accidents are exceptional, they are in no sense rare, and can never be fully guarded against. In addition to these accidental happenings of exceptional occurrence it is readily con

ceivable that there are more frequent, if less serious, interruptions in the filter service ranging in magnitude all the way from the daily fluctuations in efficiency, so well known to filter managers, to the less frequent but more serious failures due to varieties of causes. While the net effect of such fluctuations during a single year is not great and may readily be concealed in a statement of average performance, their frequency and magnitude are of the greatest significance in determining the general wholesomeness of the water supply and its effect upon the health of the consumer, matters which can not be expressed in averages. In the case of Lawrence, it is not possible to arrive at a definite statement of the actual character of the Merrimack River from existing data. B. coli is reported in 100 per cent of the cases in 1 cubic centimeter samples. A careful study of experimental filter operation and relative numbers of B. coli and of other bacterial groups, which latter have been determined and reported quantitatively, indicates that the character of the Merrimack River is neither very much better nor very much worse than the estimated limit of pollution of 500 B. coli per 100 cubic centimeters. This pollution represents the sewage of approximately 100,000 people 10 miles above, with the resulting self-purification which comes from a mean low-water storage period of 1.2 days in a 10-mile stretch of back water behind the dam at Lawrence. On the basis of mean annual run-off and Detroit and Niagara values previously derived, this pollution factor indicates a purification of between 95 and 98 per cent in the Merrimack River between Lowell and Lawrence. No corrections are possible, however, for seasonal effect or for the very great fluctuations in stream flow. In 1913, the last year for which complete data are available, the filter effluent showed B. coli in 1 cubic centimeter 15.7 per cent of the time, indicating approximately 16 B. coli per 100 cubic centimeters in the water supplied to the city. While the typhoid fever statistics in this case do not permit of definite conclusions regarding the relation of this water supply to the typhoid fever of the city, it is noticeable that the typhoid death rate in Lawrence is exceeded by that of only one city in the State of 50,000 or more inhabitants. In the single exception a higher rate is justly attributable to an excessive number of hospital cases. Further evidence of this general character might be extended and will undoubtedly be accumulated as the importance of higher standards in domestic water supplies becomes better appreciated. For the purposes of the present report, however, it will be assumed, without further discussion, that the pollution of boundary waters can not be permitted to exceed that which will give 500 B. coli per 100 cubic centimeters as a mean cross sectional value.

The upper limit of 500 B. coli per 100 cubic centimeters was also suggested as a yearly average. In view of the extraordinary seasonal variation disclosed

by the present study of the data of the Progress Report, the extent of which was not known to the consulting engineers at the time; and in view of the absence of data covering the winter months, it has been decided to deal in the present report with the mean seasonal value for the period, May to September, inclusive, 1913, deduced from the analytical data by means of a seasonal curve.

On the basis of the 1913 data, this mean value is exceeded throughout the months of July and August, and the maximum value is, for a short period, about 2.6 times the mean.

NECESSARY EFFECTIVE DILUTION.

The mean seasonal value for the Detroit and Niagara Rivers, taken together and properly weighted by populations, is 9.5 B. coli per 100 cubic centimeters per 1,000 population. This mean is taken as the best representative value for the present purposes, although the possibility exists, and is suggested by some evidence in the Progress Report and elsewhere, that the per capita pollution increases with increasing concentration of population. The difference in the values below Detroit and Buffalo, respectively, has already been pointed out in Table VII. The value given for the St. Clair River reduces to a mean seasonal value approximately six-tenths of the Detroit result, while in the case of the small communities on the St. Lawrence the pollution index was shown to account for less than one-third the population on the basis of the Detroit values. The use of the mean DetroitBuffalo value will serve the present purpose with sufficient accuracy until a more careful adjustment of values is made possible by more complete and extensive bacteriological data. Taking as the mean runoff the round figure 210,000 second-feet for each of these rivers, it is found that with a river run-off of 1 second-foot per capita of contributing population, the B. coli will be approximately 2,000 per 100 cubic centimeters, so that there is required a dilution of 4 second-feet per capita in order that dilution alone shall reduce the index of pollution to the established limit of 500. This value of 4 second-feet per capita may be taken as the necessary effective dilution, and may be regarded as the quotient of the actual dilution by the fraction of polluting material left after sewage treatment. For example, if the actual dilution be 1 second-foot per capita, it will be necessary by proper treatment to reduce the contributing sewage to onefourth of its initial value in order to bring up the effective dilution to four.

EXTENT OF SEWAGE TREATMENT PROPOSED.

Upon this basis the present and future requirements of any particular case are readily computed. The flow of the Detroit and Niagara Rivers, approxi

mately 210,000 second-feet, will receive the discharge of 52,500 persons without treatment, or of twice that number with a 50 per cent removal of polluting material, or of 10 times that number with a 90 per cent removal. Taking the present population on the Detroit River at 780,000, this leads to an indicated removal by sewage treatment of 93.5 per cent, while with a future population of 1,000,000 the corresponding purification factor for the same effective dilution will be 95 per cent. Similarly, a purification factor of 90 per cent will be necessary upon the Niagara River for the present population of approximately 500,000.

While these statements are based primarily upon examinations and standards, both in terms of B. coli, it must not be overlooked that this organism is accepted merely as a convenient indicator of sewage, and that other harmful polluting agencies must be dealt with upon the same basis. Sewage treatment deals with three types of harmful pollution-organic matter which may putrefy, solid matter which may deposit or interfere with filter operation, and bacteria. In the present case putrescible organic matter gives no concern, because the dilution limits imposed by the other two factors are far beyond those necessary in dealing with this particular feature. Suspended solids, or "solids capable of settling," as used in sewage analysis, are undesirable because of their effect upon the treatment of water supplies. The extent of purification indicated in the case of bacteria will also be adopted, therefore, for suspended solid matter.

DISTRIBUTION OF RESPONSIBILITY.

In considering the burden of responsibility, it has been assumed that it is individual and to be borne equally by all concerned. The larger cities, by reason of their greater populations, are called upon to bear the greater costs of remedying conditions, so that the burden of this cost falls more heavily upon the American cities than upon the Canadian. Considering the boundary waters and communities as a whole, however, there is found to be a condition calling for remedial measures in which the individual contributions are supposedly equal and call for equal treatment. It is therefore concluded that the specified degree of treatment will be applied in all cases, leading to an approximately equal division of the costs per capita. Increased populations upon either side of the boundary will lead to increased standards of quality in the treated sewage. Only by such a procedure can the maximum development of the natural resources of the streams be attained with a minimum of imposed burdens.

ENGINEERING INVESTIGATIONS.

The degree of purification indicated in these studies as necessary for the protection of these boundary

waters has been made the basis of detailed engineering investigation as to the feasibility and cost of accomplishing the desired results.

ORGANIZATION AND SCOPE.

For this purpose engineering offices were established at Detroit, Mich., and at Buffalo, N. Y., under the immediate supervision of District Engineers Henry C. McRae and Frank C. Tolles, respectively. The district covered by the Detroit office included all cities and towns upon both sides of the St. Clair and Detroit Rivers, and that under the Buffalo office all cities and towns upon both sides of the Niagara River. The organization in each of these offices was essentially the same, and comprised a district engineer, one assistant engineer, and two draftsmen, together with a clerk. In the case of the two principal cities involved, complete studies were made of the main drainage systems for the collection of sewage at a site or sites suitable for disposal. The studies were based primarily upon "present populations," these being taken as the estimated population at such a date in the near future as would be selected in each case for the design of works to be built about 1920. In order that the plans might be comprehensive, however, estimates of future growth, both as to magnitude, character, and direction were made, and the requirements of the future are provided for as fully as possible. Since these two cities contribute the major part of the existing pollution, and since the solution of their specific problems represents the greater part of the financial outlay necessary in the protection of the boundary waters, most of the work of the district offices was spent upon them. The problems of the smaller communities, however, received as detailed and careful study as the necessities of the situation seemed to indicate, and in each case drainage and treatment plans have been prepared in sufficient detail to permit the estimate of initial costs and of annual capital charges, including costs of operation. In each instance satisfactory drainage lines have been found, and suitable sites for the necessary treatment works have been located. A detailed description of this phase of the work, with the estimates of costs involved, will be found in the reports of the district engineers submitted herewith. The present discussion will deal with these matters only in sufficient summary to indicate the nature of the conclusions arrived at.

METHODS OF TREATMENT APPLICABLE. The nature of the problem as thus far outlined does not necessitate any very extensive discussion of sewage-treatment methods.

OXIDATION PROCESSES.

Since the removal of organic matter, from the standpoint of its putrifactive tendency, is not indi

cated in this case, those major processes of sewage treatment which involve the oxidation of organic matter need not be considered. The maximum useful development of the natural resources existing in these waters, alluded to in the reports of the consulting engineers, demands that the purely oxidizing capacity of these streams be taken advantage of, and it is readily apparent that this is sufficient for all present and future needs. Conditions of local physical nuisance now so noticeable about the water fronts of both Detroit and Buffalo, in common with many of the lake-front cities, indicate that the maximum advantage of diffusion has not been utilized in these cases, and that careful attention must be given to this matter in the design of outfall works.

REMOVAL OF SUSPENDED MATTER.

For the removal of the suspended matter capable of settling, which material is undesirable in the stream, not because of its organic character, but by reason of its physical condition, there are two major systems available, namely, fine screening and sedimentation. Careful consideration has been given to the respective merits of these two systems, and for the purposes of this preliminary investigation of possible remedies it has been found desirable to base estimates upon sedimentation in preference to fine screening. This course is conditioned upon several factors, one of these being lack of complete information on the possibilities of the newer developments in the direction of screening. Assuming, however, that screens may be developed to a point of efficiency approximating that which is possible with sedimentation, a claim which is not yet justified by experimental data, the comparative estimated costs of the two systems fails to justify the recommendation at the present time of a purely mechanical system, subject to all the inherent weaknesses of machinery, in preference to the better understood tank devices. In the case of the smaller communities, by reason of the disproportionate operating expenses incidental to the mechanical plant, the total annual cost item, taken as the sum of interest charges, operation, and depreciation, is distinctly in favor of sedimentation. In the two major cities costs are more nearly equal, but there is still a slight difference in favor of sedimentation. These respective costs are largely influenced by the further necessity for some form of chemical disinfection. This treatment requires storage tanks of a capacity at least one-half that required for sedimentation, unless existing conditions as to sewer outfall permit the substitution of such an outfall for the contact tanks. This fact, coupled with the conclusion that screened and unsettled sewage will require a somewhat larger dosage of disinfectant for results of equal efficiency, brings the total charge of the two processes much more

nearly together than would be the case were the disinfection not an integral part of the treatment. Further studies based upon more detailed data as to operating costs and results under conditions of American practice may lead to a reversal of the conclusions. arrived at in this comparison, which are in no sense final. As the purpose of this investigation has been to establish the major fact of the feasibility of treatment to a specified degree, and of approximate upper limits of costs involved in such treatment, and as feasible and reasonable treatment can be indicated along the more conservative lines of sedimentation and coincident disinfection, the tentative selection of this type of treatment and estimates of its costs suffices to furnish the required information.

Of the various possibilities in sedimentation devices, the Imhoff tank represents the latest development and presents marked advantages over other and earlier types of treatment. Its particular advantage lies in the comparative freshness and freedom from secondary decomposition of the sewage effluent, absence of undesirable odors about the works, and particularly the innocuous character of the sludge produced. This material may be readily dried upon sludge beds and disposed of in any convenient way without local nuisance.

REMOVAL OF BACTERIA.

In connection with bacterial removal, there is also no necessity for a discussion of alternative methods. Treatment with chloride of lime or bleaching powder has now been developed to a practical working method, and sufficient data are readily available upon which to base cost estimates. Alternative processes, such as treatment upon oxidizing filters or the use of liquid chlorine or other chemical disinfectants, are all so much more expensive that their further consideration need not be entered upon.

DRAINAGE CANALS.

GENERAL CONSIDERATION.

The reference specifically refers to the use of drainage canals as a possible remedial measure. The construction of such canals must obviously be conditioned by other uses than that of mere sewage disposal. This matter has therefore been given careful and detailed consideration, with reference not to the desirability of constructing a canal for purposes of sewage disposal but to the possible advantages of utilizing a canal to be constructed for other purposes, power development and shipping, as an alternative to the river for the disposal of sewage. In no case save that of the American cities on the Niagara frontier has there been any special study of a canal project, and no attention will be given to this particular phase

of the reference save in this single case. The primary function of this canal project is the development of water power, but incidental to such development there are certain possibilities of improved shipping facilities and of the utilization of the canal waters for purposes of sewage disposal. This project obviates the difficulties arising out of the treaty obligations in so far as the river itself is concerned. To merit consideration, however, it must also remedy the situation in Lake Ontario and must accomplish this result with at least as small a burden of expense and in as satisfactory a manner in all other respects as can reasonably be shown to be the case with the alternative project of river disposal.

CONDITIONS WITH TREATMENT.

The disposal of sewage by dilution in a running stream results under ideal conditions in a practically complete oxidation of the putrescible organic matter, and destruction of sewage bacteria. If conditions. permit sedimentation, there is an added effect due to this cause. The natural agencies at work in the stream require time for the completion of the purification, and the deposit of material upon the bed of the stream has the effect of increasing the time elemen and permitting more perfect purification. This is noticeably the case in Lake Erie, where the sewage of the Detroit district is so completely destroyed as to leave in the center of the lake no trace of its existence. In a canal the situation is somewhat different. Sedimentation results in an initial purification which is only apparent and temporary. The deposited material still makes its demands upon the oxygen supply of the water above and, as in this case the supply is limited and, as the long-time demand of the sludge exceeds many times the shorter time demand of the flowing unsettled sewage, conditions as to physical nuisance are aggravated, not improved, by sedimentation. Thus, while sedimentation assists in the ultimate disposal of sewage, it does so at the expense of the limited oxygen supply of the stream. For this reason the prevention of nuisance in a sluggish stream demands a much greater initial dilution of the sewage than in the case of a rapidly running stream.

In the Chicago Drainage Canal for many miles above the outlet there is during warm weather a condition of complete exhaustion of the oxygen and of foul anaerobic decomposition. Under these conditions further purification by oxidation can not proceed, and it is only when this foul stream is further diluted in the shallower Des Plaines that oxidation by natural agencies is resumed.

With a dilution of 6,000 second-feet, and with velocities of from 1.9 to 2.5 feet per second, such as are indicated in the preliminary plans of the projected drainage canal between Lake Erie and Lake Ontario, and with the present tributary population, sedimentation

30568-18-2

would take place and local nuisance arise, especially in warm weather. Ultimate oxygen exhaustion would probably not occur, so that the discharge into the lower lake would probably not be foul and offensive in a physical sense. In the course of 17 hours, the estimated time of passage from the principal sewer outfalls of Buffalo to the lake, a certain improvement in bacterial content would be noted, but owing to the greater concentration of the sewage, the discharge into the lake would be under far less favorable conditions than at present, with thorough dispersion in the large volume of river water. The resultant pollution of that portion of the lake waters near the point of discharge would be serious, and by reason of the intensity of the pollution, the effect of winds and current would carry undesirable pollution to greater distances and in greater intensity than is now observed. Aside from the possibilities of transboundary effects, the conveying of the sewage of this district in much more concentrated form than at present directly to the shores of Lake Ontario, and at a point outside the main current of the Niagara River, would result in an injury to American riparian and lake interests greater than those now existing. Nor can the possibility be excluded that this more concentrated pollution will at times cross the international boundary.

Thus far the discussion has dealt with present populations. A project of this magnitude can not be entered upon without full consideration of future necessities. With a future population of 1,000,000, conditions as regards lake pollution would be aggravated not only in proportion to the increase, but to a far greater extent because of the overtaxing of the dilution capacity of the canal itself. In such a case there can be no question as to the character of the resultant stream. Throughout the warmer months it would be devoid of oxygen, dark in color, and foul smelling. Its character would be such that some change would be urgently demanded. Further dilution or sewage treatment before discharge into the canal would be the only available remedies. Since the dilution is limited, treatment alone may be considered.

TREATMENT REQUIRED.

The problem presents too many indefinite factors to permit of any exact statement of treatment requirements. It can be stated, however, with assurance that these will not be less than those outlined for the present conditions of river discharge. In the case of the canal interest attaches primarily to putrescible organic matter, which involves the most difficult and expensive form of treatment. In other words, by adopting a canal project, utilization of the purifying capacity of the river upon this material will have been discarded and added responsibility placed upon the lake. The removal of some putrescible matter by