ELEMENTS OF WATER BACTERIOLOGY WITH SPECIAL REFERENCE TO SANITARY WATER ANALYSIS BY SAMUEL GATE PRESCOTT Associate Professor of Industrial Microbiology in the Massachusetts Institute of Technology AND CHARLES-EDWARD AMORY WINSLOW Associate Professor of Biology, College of the City of New York, and Curator of Public Health, American Museum of Natural History, New York THIRD EDITION, REWRITTEN FIRST THOUSAND NEW YORK JOHN WILEY & SONS, INC. LONDON: CHAPMAN & HALL, LIMITED 1913 QTK 1 0- CJDO . Copyright, 1904, 1908, 1913 BY :S. C. PRESCOTT AND C.-E. A. WINSLOW THE SCIENTIFIC PRESS ROBERT DRUMMOND AND COMPANY BROOKLYN. N. V. DEDICATED TO William BY TWO OF HIS PUPILS, AS A TOKEN OF GRATEFUL AFFECTION PREFACE TO FIRST EDITION THE general awakening of the community to the importance of the arts of sanitation — accelerated by the rapid growth of cities and the new problems of urban life — demands new and accurate methods for the study of the microbic world. Bacteriology has long since ceased to be a subject of interest and importance to the medical profession merely, but has become intimately connected with the work of the chemist, the biologist, and the engineer. To the sanitary engineer and the public hygienist a knowledge of bacteriology is indis- pensable. In the swift development of this science during the last ten years perhaps no branch of bacteriology has made more notable progress than that which relates to the sanitary examination of water. After a brief period of extravagant anticipation, and an equally unreasonable era of neglect and suspicion, the methods of the practical water bacteriologist have gradually made their way, until it is recognized that, on account of their delicacy, their directness, and their certainty, these methods now furnish the final criterion of the sanitary condition of a potable water. vi PREFACE TO FIRST EDITION A knowledge of the new science early became so indispensable for the sanitary expert that a special course in the Bacteriology of Water and Sewage has for some years been given to students of biology and sani- tary engineering in the Biological Department of the Massachusetts Institute of Technology. For workers in this course the present volume has been especially prepared, and it is fitting, we think, that such a manual should proceed from an institution whose faculty, graduates, and students have had a large share in shap- ing the science and art of which it treats. We shall be gratified, however, if its field of usefulness extends to those following similar courses in other institutions, or occupied professionally in sanitary work. The treatment of the subject in the many treatises on General Bacteriology and Medical Bacteriology is neither special enough nor full enough for modern needs. The classic work of Grace and Percy Frank- land is now ten years old; and even Horrocks' valuable " Bacteriological Examination of Water " requires to be supplemented by an account of the developments in quantitative analysis which have taken place on this side of the Atlantic. It is for us a matter of pride that Water Bacteriology owes much of its value, both in exactness of method and in common-sense interpretation, to American sanitarians. The English have contributed researches of the greatest importance on the significance of certain intestinal bacteria; but with this exception the best work on the bacteriology of water has, in our opinion, been done in this country. Smith, Sedgwick, Fuller, PREFACE TO FIRST EDITION vii Whipple, Jordan, and their pupils and associates (not to mention others) have been pioneers in the develop- ment of this new field in sanitary science. To gather the results of their work together in such form as to give a correct idea of the best American practice is the purpose of this little book; and this we have tried to do with such completeness as shall render the volume of value to the expert and at the same time with such freedom from undue technicality as to make it reada- ble for the layman. It should be distinctly understood that students using it are supposed to have had before- hand a thorough course in general bacteriology, and to be equipped for advanced work in special lines. BOSTON, March 10, 1904. PREFACE TO THIRD EDITION A SECOND edition of this work was called for in 1908 and it was rewritten in that year, with the inclusion of much new material in the chapters dealing with the isolation of the typhoid bacillus and of intestinal bacteria, and with the addition of a new chapter on the bacteriology of sewage and sewage effluents. In the same year there appeared an excellent volume on Water Bacteriology by Dr. W. G. Savage, which showed the English methods of investigation and interpre- tation to be closely in accord with those used in America. In the five years which have elapsed since our second edition was published, there has again been important progress along many lines in sanitary bacteriology; and in particular the publication in 1912 of a new edition of the Report of the Committee on Standard Methods of the Laboratory Section of the American Public Health Association has made necessary a change in many details of current practice. We have, therefore, prepared at this time a some- what far-reaching revision of our book. Newer ideas ix x PREFACE TO THIRD EDITION on the effect of temperature upon the viability of bacteria in water are included in Chapter I. The recent recommendations of the Committee on Standard Methods are discussed in Chapters II and IV; in particular, Chapter IV, dealing with the 37° count, has been expanded. We cannot bring ourselves to agree with the recommendation of the committee that the 37° count should replace the 20° count; but we are entirely in accord with the resolution adopted by the Laboratory Section of the American Public Health Association at its Washington meeting that both determinations should be made in ordinary routine water examinations. Indeed, this is the position we have maintained in both our earlier editions. Chapter V, dealing with the isolation of specific pathogenes from water, has been extensively rewritten and extended. The use of the Jackson bile medium for the preliminary enrichment of the typhoid bacillus has become general since 1908 and a number of suc- cessful isolations have been reported by its use; so that this procedure promises to be of increasing importance in the future. In regard to the isolation and identification of bacilli of the colon group we feel that the time has come for a change from the usual American practice of the past. The five standard tests for " typical B. coli " established by the Committee on Standard Methods in its 1905 report have come to seem more and more illogical and unscientific to most practical water bac- teriologists. The conviction has grown that they go either too far or not far enough. For waters, in the PREFACE TO THIED EDITION xi United States, at least, it seems clear that all of the lactose-fermenting group of bacilli are significant of pollution from human or animal sources when present in considerable numbers. The 1912 report of the Standard Methods Committee apparently takes this view in one place, while retaining the five tests in another section. We have felt it best to place ourselves fairly and fully in line with the view that the whole group of lactose-fermenting bacilli is significant and that the lactose bile fermentation test is a sufficient identi- fication of the colon group for ordinary sanitary pur- poses. This broad definition is the one upon which we have based our general discussion of the colon group in Chapters VI and VII. In Chapter VIII we have discussed the subdivisions of the group as worked out by MacConkey and others and their special significance with respect to recent and remote pollu- tion as suggested by the researches of Houston and Clemesha. The growing importance of the application of bac- teriology to the sanitary study of shellfish has led us to include a new chapter dealing with this subject, based largely upon the recent report of the Committee of the Laboratory Section of the American Public Health Association. Throughout the book we have resorted freely to the use of tables of actual data for the illustration of the various points discussed, believing that ample familiarity with practical examples furnishes the only sound basis for judgment in sanitary water exam- ination. xii PEEFACE TO THIKD EDITION For the benefit of the student the chapters have been sub-divided into sections with prominent headings indicating the general topics under discussion. Massachusetts Institute of Technology, BOSTON, Mass. College of the City of New York, NEW YORK, N. Y. June i, 1913. TABLE OF CONTENTS CHAPTER I PAGE THE BACTERIA IN NATURAL WATERS i CHAPTER II THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION or WATER. . 29 CHAPTER III THE INTERPRETATION OF THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION 51 CHAPTER IV DETERMINATION OF THE NUMBER OF ORGANISMS DEVELOPING AT THE BODY TEMPERATURE 61 CHAPTER V THE ISOLATION OF SPECIFIC PATHOGENES FROM WATER 74 CHAPTER VI THE COLON GROUP OF BACILLI AND METHODS FOR THEIR ISOLATION 99 xiii xiv TABLE OF CONTENTS CHAPTER VII PAGE SIGNIFICANCE or THE PRESENCE OF THE COLON GROUP IN WATER. 140 CHAPTER VIII VARIETIES OF COLON BACILLI AND THEIR SPECIAL SIGNIFICANCE. 174 CHAPTER IX OTHER INTESTINAL BACTERIA 201 CHAPTER X THE SIGNIFICANCE AND APPLICABILITY OF THE BACTERIOLOGICAL EXAMINATION 215 CHAPTER XI BACTERIOLOGY OF SEWAGE AND SEWAGE EFFLUENTS 228 CHAPTER XII BACTERIOLOGICAL EXAMINATION OF SHELLFISH 244 APPENDIX 265 ELEMENTS OF WATER BACTERIOLOGY CHAPTER I THE BACTERIA IN NATURAL WATERS Bacteria and Their Nutritive Relations. Bacteria are the most numerous and the most widely distributed of living things. They are present not merely at the surface of the earth or in the bodies of water which partially cover it, as is the case with most other living things, but in the soil itself, and in the air above, and in the waters under the earth. Probably no organisms are more sensitive to external conditions, and none respond more quickly to slight changes in their environment. Temperature, moisture, and oxygen are of importance in controlling their distri- bution; but the most significant factor is the amount of food supply. Bacteria and decomposing organic matter are always associated, and for this reason a brief consideration of the general relation of bacteria to their sources of food supply must precede the study of their distribution in any special medium. The bacteria possess greater constructive ability than any animal organisms. They lack, however, 2 ELEMENTS OF WATER BACTERIOLOGY the power of green plants to build up their own food from compounds like carbon dioxide and nitrates which have no stored potential energy. The food require- ments of various bacterial types differ, however, widely among themselves. Fischer (1900) has divided the whole group into three great subdivisions according to the nature of their metabolism. The Proto trophic forms are characterized by minimal nutrient require- ments, including organisms like the nitrifying bacteria which require no organic compounds at all, but derive their nourishment from carbon dioxide or carbonates, nitrites and phosphates, or from inorganic ammonium salts. A second group of Metatrophic bacteria includes those forms which require organic matter, nitrogenous and carbonaceous, but are not dependent on the fluids of the living plant or animal. Finally, the Para trophic bacteria are the true parasites, which exist only within the living tissues of other organisms. These sub- divisions, like all groups among the lower plants, are not sharply defined, and the metatrophic bacteria in particular exhibit every gradation, from types which grow in water with a trace of free ammonia to organisms like the colon bacillus which normally occur on the surface of the plant or animal body, feeding upon the fluids or on the extraneous material collected upon its surface. The vast majority of bacteria belong to the second, or metatrophic group, living as saprophytes on dead organic matter wherever it may occur in nature, and particularly in that diffuse layer of decomposing plant and animal material which we call the humus, or surface THE BACTERIA IN NATURAL WATERS 3 layer of the soil. Wherever there is life, waste matter is constantly being produced, and this finds its way to the earth or to some body of water. The excretions of animals, the dead tissues and broken-down cells of both animals and plants, as well as the wastes of domestic and industrial life, all eventually find their way to the soil. In a majority of cases these substances are not of such chemical composition that they can be utilized at once by green plants as food, but it is first necessary that they go through a decomposition or transforma- tion in which their chemical nature becomes changed; and it is as the agents of this transformation that bacteria assume their greatest importance in the world of life. We may take the decomposition of a comparatively simple excretory product, urea, as an example of the part which the bacteria play in the preparation of plant food. Through the activity of an enzyme pro- duced by certain bacteria this compound unites with two molecules of water and is converted into ammonium carbonate, + 2H20 = (NH4)2C03. NH2 This, however, is only part of the process. While green plants can derive their necessary nitrogen in part, at least, from ammonium compounds it is a well- established fact that this element is often obtained more readily from nitrates, and there are other bacteria which as a further step oxidize the ammoniacal nitro- 4 ELEMENTS OF WATEE BACTERIOLOGY gen to a more available form. This process of oxida- tion is known as nitrification, and takes place in a suc- cession of steps, the organic nitrogen being first con- verted to the form of ammonium salts, and these in turn to nitrites and nitrates, the oxygen used coming from the air. Several groups of organisms are instru- mental in bringing about this conversion. It is gen- erally assumed that one group attacks the ammonium compounds and changes them to nitrites; while another group completes the oxidation to nitrates. In the latter form nitrogen is readily taken up by green plants to be built up into more complex albuminoid sub- stances (organic nitrogen) through the constructive power of chlorophyll. This never-ending cycle is illustrated in the accom- panying figure, devised by Sedgwick (Sedgwick, 1889) to illustrate the transformations of organic nitrogen in nature, the increasing size and closeness of the spiral on the left-hand side indicating the progressive com- plexity of organic matter as built up by the chlorophyll bodies of green plants in the sunlight, and the other half of the figure the reverse process, carried out largely by the bacteria. In nature there are many short circuits, as, for instance, when dead organic matter is used as food for animals and built up into the living state again without being nitrified and acted upon by green plants; but the complete cycle of organic nitrogen is as indicated on the diagram. We have dwelt thus at length upon the general relation between bacteria and organic decomposition because in this relation will be found the master key THE BACTEEIA IN NATURAL WATERS 5 to the distribution of bacteria in water as well as in other natural habitats. It is true that certain peculiar forms may at times multiply in fairly pure waters; but, in general, large numbers of bacteria are found only in connection with the organic matter upon which they feed. Such organic matter is particularly abundant in the surface layer of the soil. Here, therefore, the bacteria are most numerous; and in other media their THE SPHERE OF ORGANISMS AND THE HISTORY OF ORGANIC MATTER, numbers vary according to the extent of contact with the living earth. Classification of Waters. Natural waters, then, group themselves from a bacteriological standpoint in four well-marked classes, according to their relation to the rich layers of bacterial growth upon the surface of the globe. There are first the atmospheric waters which have never been subject to contact with the earth; second, the surface-waters immediately exposed to such 6 ELEMENTS OF WATER BACTERIOLOGY contamination in streams and pools; third, stored waters, the lakes and large ponds in which storage has reduced bacterial numbers to a state of comparative purity; and fourth, the ground-waters from which previous contamination has been even more completely removed by filtration through the deeper layers of the soil. Bacterial Content of Various Waters. Even rain and snow, the sources of our potable waters, are by no means free from germs, but contain them in numbers varying according to the amount of dust present in the air at the time of the precipitation. After a long- continued storm the atmosphere is washed nearly free of bacteria, so that a considerable series of sterile plates may often be obtained by plating i-c.c. samples. These results are in harmony with the observations of Tissandier (reported by Duclaux, 1897), wno found that the dust in the air amounted to 23 mg. per cubic meter in Paris and 4 mg. in the open country. After a rainfall these figures were reduced to 6 mg. and 0.25 mg., respectively. With regard to what may be considered normal values for rain it is difficult to give satisfactory figures. Those obtained by Miquel (Miquel, 1886) during the period 1883-1886 showed on the average 4.3 bacteria per c.c. in the country (Montsouris) and 19 per c.c. in Paris. Snow shows rather higher numbers than rain. Janowski (Janowski, 1888) found in freshly fallen snow from 34 to 463 bacteria per c.c. of snow-water. As soon as the rain-drop touches the surface of the earth its real bacterial contamination begins. Rivulets from ploughed land or roadways may often contain several hundred thousand bacteria to the cubic centi- THE BACTERIA IN NATURAL WATERS 7 meter; and furthermore the amounts of organic and mineral matters which serve as food materials, and thus become a factor in later multiplication of organisms, are greatly increased. In the larger streams several conditions combine to make these enormous bacterial numbers somewhat lower. Ground-water containing little microbic life enters as a diluting factor from below. The larger particles of organic matter are removed from the flow- ing water by sedimentation; many earth bacteria, for which water is an unfavorable medium, gradually perish; and in general a new condition of equilibrium tends to be established. It is difficult, however, to find a river in inhabited regions which does not con- tain several hundreds or thousands of bacteria to the cubic centimeter. Furthermore, heavy rains which introduce wash from the surrounding watershed may at any time upset whatever equilibrium exists, and surface-waters are apt to show sudden fluctuations in their bacterial content. Seasonal Variation of Bacteria in Surface Waters. Sharp variations in bacterial content are particularly apt to occur in the spring and fall as a result of the rain and melting snow at those seasons. The high numbers shown for various rivers in the table on page 8 illustrate this point. The rainfall is the main factor which causes these sea- sonal variations; but its specific effect differs with dif- ferent streams. The immediate result of a smart shower is always to increase contamination by introducing fresh wash from the surface of the ground. More prolonged s ELEMENTS OF WATER BACTERIOLOGY SEASONAL VARIATIONS IN BACTERIAL CONTENT OF RIVER WATERS. BACTERIA PER C.C. MONTHLY AVERAGE River. Year. Jan. Feb. Mar. April. May. June. Thames x 1905-6 2075 1,679 1,161 277 I 064 8? Lea1 1905-6 5,192 3,083 1,308 471 1,350 598 New 1 1905-6 I 455 I 304 291 I4.Q -? cr 2 108 Mississippi 2 . . . . 1900-01 972 2,871 i,795 3,597 2,152 2,007 Potomac 3 1906-7 4,400 I,OOO 11,500 3,700 7 co 77 300 Merrimac 4 14,200 14.800 10,300 3,600 1,900 9,600 Susquehanna 5. . 1906 9,510 21,228 31,326 39,905 6,187 2,903 River. Year. July. / ug. Sept. Oct. Nov. Dec. Thames 1 190=5—6 952 I 633 74-O Lea1 1905-6 1,190 7,946 2.CKO New1 1905-6 450 718 621 Mississippi 2 . . . . 1900-01 1,832 805 2,021 Potomac 3 1906-7 2,700 3,000 6,2OO 2,300 1, 800 6,9OO Merrimac 4 1905 3,900 19,500 I3-500 39,800 8,700 Susquehanna $ . 1906 685 1,637 836 7,575 26,224 37,525 1 Houston, 19060, 19066. 4 Massachusetts, 1906. 2 New Orleans, 1903. 5 Harrisburg, 1907. 3 Figures obtained through courtesy of F. F. Longley. moderate rain, however, exerts an opposite effect, and after the main impurities which can be washed away have been removed, may dilute the stream with water purer than itself. What the net effect of rain may be depends, therefore, on the character of the stream. A river of fairly good quality shows its highest numbers in rainy periods. With a highly polluted stream, on the other hand, the constant influx of sewage over- balances occasional contributions of surface contamina- tion. Thus Gage (1906) shows in the following table that the bacterial content of the Merrimac is highest when the stream is lowest, that is, when its sewage content is least subject to dilution. THE BACTERIA IN NATURAL WATERS VARIATIONS IN BACTERIAL CONTENT, MERRIMAC RIVER GAGE (1906) Flow of St-eam. Cubic Bacteria per c.c. B. coli per c.c. Square Mile of Watershed Canal. ntake. Canal. Intake. Less than i 7,500 10,800 66 .88 1—2 6 800 6 200 CQ qi 2-4 3,600 5,6oo 2O 2Q Over 4 3 AOO 2 IOO 16 2O The contrast between the two classes of rivers is well brought out in a study of the Lahn and the Wieseck, by Kisskalt (1906); and the table below, compiled from his data, gives an excellent idea of the total numbers of bacteria and their seasonal fluctuations in a stream of fair quality (the Lahn) and a highly polluted one (the Wieseck). In the former case the bacterial numbers are highest when rain brings surface pollution; in the lat- ter, when the sewage constantly present is least diluted. MONTHLY VARIATIONS OF BACTERIA IN A NORMAL AND POLLUTED STREAM KISSKALT, 1906 Date. Bacteria per c.c. Date. Bacteria per c.c. 1904. Lahn. Wieseck. 1904-5- Lahn. Wieseck. My 3l8 104 ooo December * I 22O 21 2OO July August October l . . . October l . . . 132 840 1,235 420 156,800 98,400 28,400 58,000 January l . . February 1 . March1.... April ! 3,668 5,38o I,2IO 4,025 29,920 1 1 ,9OO 8,250 5,QIO November 2 34O 7Q 2OO May C7o 14 800 November l . I,74O 5 2 ,000 June . , 686 ^o 180 December l . . 780 28,600 1 Rain or high water due to previous thaw. 10 ELEMENTS OF WATER BACTERIOLOGY Effect of Storage upon Bacteria in Water. In stand- ing waters all the tendencies which make for the reduc- tion of bacteria are intensified, and when a river passes into a natural or artificial reservoir a notable reduction in numbers occurs. The table below shows the striking effect produced upon the water of the Potomac River by its successive passage through the three reservoirs of the Washington water supply in the first nine months of 1907. We owe these figures to the courtesy of Mr. F. F. Longley, the engineer then in charge of the Washington filter plant. REDUCTION OF BACTERIA IN WASHINGTON RESERVOIRS. BACTERIA PER C.C., MONTHLY AVERAGE, 1907 Potomac River. Dalecarlia Reservoir. Georgetown Reservoir. Washington City Reservoir. January February .... March . 4,400 I,OOO II 500 2,400 95° 8 300 2,2OO I,OOO 7 2OO 950 750 •2 600 April 3.700 2 IOO I 4OO 47 ^ M!ay 7 ^O •7 tro •2 2 C June July . . 2,300 2 7OO 950 600 600 3^O IOO 1 60 August September . . . 3,000 6. 200 275 425 I.QOO 80 230 The still more striking results obtained at London are indicated in the table on page n. When the water which enters a pond or a reservoir has already undergone considerable storage and reached a comparatively stable condition, the diminution due to additional storage may be almost negligible. Thus Philbrick (1905) found that the influent water of the Chestnut Hill Reservoir of the Metropolitan Water THE BACTEEIA IN NATURAL WATERS 11 Works of Boston contained on the average during the eleven years, 1893-1903, 220 bacteria per c.c., and the effluent 179. In many individual months, and in some whole years, the effluent contained more than the influent. AVERAGE REDUCTION OF BACTERIA BY STORAGE AT LONDON (HOUSTON, 1909) Water Storage, Days. Bacteria per c.c. Gelatin 20°. Agar 37°. Bile-salt Agar 37°. Raw Thames River 4405 175 208 362 8135 67 280 34 44 52 382 II 41 2 5 8 34 i Do. stored at Staines 95 15 14 "5s" Do stored at Chelsea Do. stored at Lambeth Raw Lee River Do. stored The seasonal variations in the bacterial content of a large pond or lake follow a somewhat different course from those observed in a stream. Philbrick, in , the paper just cited, gives the figures tabulated below for the Chestnut Hill Reservoir of the Metropolitan Water Works (Boston). The averages are based on weekly analyses covering the eleven years, 1893-1903. MONTHLY VARIATIONS IN BACTERIAL CONTENT OF CHESTNUT HILL RESERVOIR, 1893-1903 Month. J. F. M. A. M. J. J. A. S. 0. | K. D. Bacteria per c.c. 82 73 7i 123 69 73 82 95 134 89 103 96 12 ELEMENTS OF WATER BACTERIOLOGY The marked increase in April and September is the notable feature of these analyses; and this is due to the effect of the spring and fall overturns which, in the months in question, stir up the decomposing organic matter at the bottom and distribute it through the reservoir. The slight, but steady, increase during the warm months from May to August is also probably significant. On the whole it may be said that the bacterial content of a lake or pond should not be more than one or two hundred per c.c. and may often be under a hundred. The student will find numerous analyses of natural waters in Frankland's classic work (Frankland, 1894). He notes, for example, that the Lake of Lucerne con- tained 8 to 51 bacteria per c.c., Loch Katrine 74, and the Loch of Lintralthen an average of 170. The water of Lake Champlain examined by one of us (S. C. P.) in 1896 contained on an average 82 bacteria per c.c. at a point more than two miles out from the city of Burling- ton. Certain surface water-supplies near Boston studied by Nibecker and one of us (Winslow and Nibecker, 1903), gave the following results: City. Number of Samp es. Average Number of Bacteria per c.c. Wakefield 7 CO Lynn 6 16 Plymouth 6 •2 r Cambridge c 04 Salem c 222 Medford . . C C.24. Taunton 4 13 Peabody ? I4.I THE BACTEEIA IN NATURAL WATERS 13 In sea-water, too, bacterial numbers are small, as noted by Russell at Naples (Russell, 1891) and Wood's Hole (Russell, 1892), and in salt as in fresh water the amount of bacterial life decreases in general as one passes downward from the surface and outward from the shore. Otto and Neumann (1904) obtained the results summarized below at various points on the high seas between Portugal and Brazil. Near the European coast numbers were much higher. BACTERIA IN THE ATLANTIC OCEAN. (OTTO AND NEUMANN, 1904.) BACTERIA PER C.C. Nearest Land. Depth ir Meters. 5 50 IOO 2OO Canary Islands 1 20 76 2O I Cape Verde Islands 58 16 64 6 St Paul Island 20 480 54 4 Pernambuco 48 168 83 14 Drew (1912) finds high numbers of bacteria in surface sea- water off the Bahamas, ranging from 13,000 to 16,000, falling off below 200 fathoms (in the cold bottom waters at 10° C. or below) to o to 17. Factors Influencing the Diminution of Bacteria in Surface-waters. The decrease in numbers which takes place when a surface-water is stored in a pond or reservoir indicates that the forces which tend to produce bacterial self-purification are important ones. It is necessary to consider in somewhat more detail just what these forces are, in order to gauge their potency in any particular instance. 14 ELEMENTS OF WATER BACTERIOLOGY Chief of them appear to be sedimentation, the activ- ity of other micro-organisms, light, temperature, and food-supply, and perhaps more obscure conditions such as osmotic pressure. The subsidence of bacteria, either by virtue of their own specific gravity, or as the result of their attachment to particles of suspended matter, is unquestionably partly, if not largely, responsible for changes in the number of bacteria in the upper layers of water at rest or in very sluggish streams. The results of numerous investigations by different workers seem to indicate that sedimentation of the bacteria themselves takes place slowly, and that the difference in numbers between the top layer and the bottom layer of water in tall jars in laboratory experiments of a few days' duration is very slight or quite within the limits of experimental error (Tiemann and Gartner, 1889). Different species may, of course, be differently affected (Scheurlen, 1891). It must be remembered, however, that in natural streams bacteria are to a great extent attached to larger solid particles upon which the action of gravity is more important. Spitta (1903) found that from one- fifth to one-half of the bacteria in canal water may be attached to gross particles, as evidenced by their sedimentation in a few hours. Jordan (Jordan, 1900) is firmly of the opinion that in the lower part of the Illinois River, where there is a fall of but 30 feet in 225 miles, the influences summed up by the term sedimentation are sufficiently powerful to obviate the necessity for summoning another cause " to explain the diminution n\ numbers of bacteria," and he further THE BACTERIA IN NATURAL WATERS 15 adds: " It is noteworthy that all the instances recorded in the literature where a marked bacterial purification has been observed are precisely those where the con- ditions have been most favorable for sedimentation." Little is known as to the share of other organisms in hastening the decrease of bacteria in stored water. Doubtless predatory Protozoa play some part in the process. Huntemiiller (1905) after infecting water containing flagellate Protozoa with typhoid bacilli, found the Protozoa crowded with bacteria; and he observed under the microscope the actual ingestion of the living and motile bacilli. Korschun (1907) and others have obtained similar results and consider the activity of Protozoa to be an important factor in self- purification. Fehrs (1906) found that typhoid bacilli would live for 7 days in unsterilized Goltingen tap water, for 46 days in the same water sterilized, and for 13 days in water inoculated with a culture of flagellate Protozoa after sterilization. Water bacteria were of course added with the Protozoa. Stokvis and Swel- lengrebel (1911) have shown that ciliated infusoria may also consume considerable quantities of bacteria under favorable conditions as to oxygen and temperature, and Horhammer (1911) reports that certain Crustacea such as Cyclops may devour considerable quantities of typhoid bacilli when present in masses from cultures? stained with methylene blue, and suspended in water. Certain bacteriologists have held that the toxic waste products of the bacteria themselves may render water unfit for their own development. Horrocks (Horrocks, 1901), Garre (Garre, 1887), Zagari (Zagari, 1887) and 16 ELEMENTS OF WATER BACTERIOLOGY Freudenreich (Freudenreich, 1888) have shown that an " antagonism " exists when bacteria are grown in artificial culture media, such that the substratum which has supported the growth of one form may be rendered antiseptic to another. Frost (1904) has exhaustively studied the phenomenon of antagonism by exposing typhoid bacilli in collodion sacs to the action of certain soil and water bacteria growing in broth. Artificial culture media, however, offer conditions for bacterial development which are scarcely paralleled in natural waters. It is difficult to believe that under ordinary conditions poisons are produced of such power as to render a stream or lake specifically toxic for any par- ticular type of bacteria. It does appear indeed from the experiments of Jordan, Russell and Zeit (1904), and Russell and Fuller (1906), which will shortly be referred to more fully, that the life of typhoid germs is shorter in water containing large numbers of other bacteria than in that of greater purity. Horrocks (1899), too, found freshly isolated typhoid bacilli alive in sterile sewage after 60 days; while they disappeared in 5 days when B. coli was also present. These phenomena may be due, however, to a struggle for oxygen, or for food, rather than to the assumed presence of highly toxic bacterial products, of which there is no independent evidence. Many investigations conducted since the pioneer researches of Downes and Blunt (Downes and Blunt, 1877) have confirmed the results reported by them, which showed that direct sunlight is fatal to most bacteria in the vegetative state and even to spores if THE BACTEEIA IN NATURAL WATERS 17 the exposure be sufficiently long, while diffused light is harmful in a less degree. Opinions vary as to the degree to which light is active in destroying the bacteria in natural waters. Buchner (Buchner, 1893) found by experiment that the bactericidal power of light extends to a depth of about three meters before it becomes imperceptible. On the other hand, Procaccini (Procaccini, 1893) found that when sunlight was passed vertically through 60 cm. of drain- water the lower layers contained nearly as many bacteria after 3 hours' treatment as before the exposure. The middle and upper portions showed a great falling off in numbers, however. But few studies have been made of the effect of light on bacteria in flowing water. Jordan (Jordan, 1900) has investigated several Illinois streams and arrived at the conclusion that in moderately turbid water, at least, the sun's rays are virtually without action. On the other hand, Rapp has observed a considerable reduction of the bacteria in the Isar at Pullach after the period of diurnal insolation, as shown by the table on the following page. Clemesha (191 2a) attributes very great importance to the action of light in the self- purification which takes place in Indian lakes and rivers; his opinion is apparently not based on com- parative experiments including and excluding this factor, but chiefly on the greater numbers of intestinal bacteria at the bottom as compared with the superficial layers of water. It is unnecessary to dwell in detail upon the effect which the lack of nutritive elements must exert upon 18 ELEMENTS OF WATER BACTERIOLOGY intestinal bacteria and soil bacteria in waters of ordinary purity. Comparative studies of culture media, to be quoted in the succeeding chapter, will show how del- icately the bacteria respond to comparatively slight changes in their food-supply. Wheeler (1906) found that typhoid bacilli would persist in almost undimin- ished numbers in sterilized water from a polluted well containing considerable organic matter and kept in the dark at 20 degrees, while in purer water or in the light they died out in from 2 to 6 weeks. EXAMINATIONS OF THE ISAR AT PULLACH (RAPP, 1903) (A} Carried out September 26. i8p8, no rain having fallen for three weeks Temperature. Time o the Experiment. Bacteria per c.c. of the Water. of the Air. i3.o°C. 8Q O /"^ . 0 C. 7.3op.M. 146 I2.I°C. 7.o°C. 9.30P.M. 270 10. 5° C. 6.2°C. 5.00 A.M. 370 10.2° C. 8.2°C. 8.00 A.M. 320 (B) Carried out November 28, 1898, no rain having fallen for some time 5-5° C. 3-o°C. 6.OO P.M. 266 5-5°C. 2-5°C. 8.00P.M. 402 5.5°C. 2.0°C. 10. 00 P.M. 482 5.o°C. 2.0°C. 3-OO A.M. 532 4.5°C. 2-5°C. 7.30A.M. 400 Whipple and Mayer (1906) have called attention to another important factor in the general problem. They find that the presence of oxygen is essential to the per- THE BACTERIA IN NATURAL WATERS 19 sistence of typhoid and colon bacilli in water, although in nutrient media both forms may thrive under anaerobic conditions. EFFECT OF OXYGEN ON VIABILITY OF TYPHOID BACILLI IN STERILE TAP WATER WHTPPLE AND MAYER, 1906 Tubes Kept in Air. Tubes Kept in Hydrogen. Period in Days Bacteria per c.c. Per Cent. Bacteria per c.c. Per Cent. o 600,000 IOO.O 6oo,000 IOO.O 2 455;°°° 76.0 2,400 0.4 4 190,000 32.0 25 0.004 8 120,000 20. O O O.O 12 67,000 II. 0 0 O.O 18 25,000 4-2 o O.O 26 9,250 i-5 o O.O 33 2,150 0.6 0 0.0 40 132 O.O2 0 O.O 47 6 O.OOI o O.O 54 o o.ooo 0 0.0 Various inorganic constituents of the medium undoubt- edly exercise an important influence upon the life of bacteria in water; and the mutual interaction of the different substances present is a highly complex one. Thus Winslow and Lochridge (1906) report that five parts of dissociated hydrogen per million parts of tap water (0.005 normal HC1) is fatal to typhoid bacilli, while ten times as much acid is required for sterilization when i per cent of peptone is present to check the dissociation of the hydrogen. In Hazen and Whipple's study of the Allegheny, Monongahela and Ohio rivers 20 ELEMENTS OF WATER BACTERIOLOGY at Pittsburgh the antiseptic effect of acid wastes was strikingly shown. (Engineering News, 1912.) Although it is hard to estimate the exact importance of each factor, the general phenomena of the self- purification of streams are easy to comprehend. A small brook, immediately after the entrance of polluting material from the surface of the ground, contains many bacteria from a diversity of sources. Gradually those organisms adapted to life in the earth or in the bodies of plants and animals die out, and the forms for which water furnishes ideal conditions survive and multiply. It is no single agent which brings this about, but that complex of little-understood conditions which we call the environment. If any one thing is of prime impor- tance it is probably the food-supply, for only certain bacteria are able to multiply in the presence of the small amount of organic matter present in ordinary potable waters. As Jordan (Jordan, 1900) has said: " In the causes connected with the insufficiency or unsuitability of the food-supply is to be found, I believe, the main reason for the bacterial self-purification of streams." Effect of Temperature upon Bacteria in Water. The effect of temperature upon the survival of bac- teria in water varies according to this primary con- dition of food-supply which has just been discussed. When bacteria are in a medium in which they are able to grow and multiply, warmth, within reasonable limits of course, favors their development, At times this may be true even of certain intestinal bacteria in water. Thus at Harrisburg, Pa., a series of B. coli examinations THE BACTERIA IN NATURAL WATERS 21 made in the midsummer of 1906 showed positive results in 7 per cent of the samples of water entering the storage reservoir and in 27 per cent of the samples leaving it. The storage period in this case was about two days and the temperature of the water in the reservoir was nearly at blood heat (Harrisburg, 1907). Clemesha (1912*) has recently made an exhaustive study of this multiplication of coli-like microbes in warm waters and has shown that it is confined to certain particular types within the colon group. For most intestinal bacteria the conditions necessary for growth and multiplication are not realized in water and an entirely different temperature effect is manifest. When a bacterium cannot multiply, the only vital activity which can take place is a katabolic wasting away, which soon proves destructive, and the higher the temperature the more rapidly the fatal result is reached. A frog in winter lives at the bottom of a pond breath- ing only through its skin and eating not at all, but as soon as the temperature rises it must eat and breathe through its lungs or perish. It is quite true that even in ice 40 per cent of typhoid bacilli perish in 3 hours and 98 per cent in 2 weeks (Sedgwick and Winslow, 1902). Recent work has shown, however, that they die in spite of the cold, not on account of it, and that the decrease is more rapid at higher temperatures, unless of course food-supply and other conditions admit of multiplication. Houston (1911) has furnished a very clear demonstration of this temperature rela- tion by storing typhoid bacilli in water with the results tabulated on page 22. 22 ELEMENTS OF WATER BACTERIOLOGY EFFECT OF TEMPERATURE ON SURVIVAL OF TYPHOID BACTERIA IN WATER (HOUSTON, 1911) Temperature C. Percentage of Typhoid Bacilli Surviving after One Week. Period of Final Disappearance of BacilH. O 46 9 weeks c; 14 7 weeks jo O O7 5 weeks 18 O.O4 4 weeks Ruediger (1911) has shown that colon bacilli are far more abundant in the Red Lake River during the winter when the river is covered with ice than in sum- mer, although the volume of the river and the amount of sewage pollution are about the same. Typhoid bacilli in celloidin dialyzers floated down the river showed only 2.5 and 3.5 per cent surviving in 2 days and 0.51, 0.89, 2.2 and 3.2 per cent surviving in 3 days when the river was not frozen, while dialyzers suspended through the ice in colder weather showed 6.1, 10.5, 17.7, 46.8 and 62.9 per cent surviving in five different experi- ments after 2 days, 31 per cent in 3 days, 19 per cent in 7 days, and 2.5 per cent in 14 days. Ruediger attributes this greater persistence at low temperatures to the absence of poisonous waste products of other organisms and to protection from the light; but there can be little doubt that it is mainly a result of the general preservative effect of cold. From an epidemiological standpoint the conclusion that disease germs perish quickly in warm waters is amply confirmed. Almost without exception outbreaks of typhoid fever due to THE BACTERIA IN NATURAL WATERS 23 polluted water occur in cold weather and this is, in part at least, due to the greater persistence of typhoid bacilli at low temperatures. Relation between Time of Storage and Self-purifica- tion. It is obvious that the efficiency of all the agencies which tend to decrease the number of bacteria in sur- face waters will increase with the prolongation of the period for which they act. Time is the great measure of self -purification. The longer the storage the greater the improvement, and after a certain period even a fairly polluted water may be safe and potable. The absolute time necessary to produce this result varies of course according to many conditions. Food supply, light, temperature and the activity of other living forms vary widely and in depos- ited material conditions are different from those which obtain in the water itself. Jordan, Russell and Zeit (1904), in an important series of experiments, added typhoid bacilli to the unsterilized waters of Lake Michigan, the Chicago River and Drainage Canal and the Illinois River, in collodion sacs suspended in the respective bodies of water. From the relatively pure Lake Michigan water the specific organisms could be isolated for at least a week, but in the polluted waters of the rivers and the Drainage Canal they were not found after 3 days except in a single instance. Russell and Fuller, (1906) confirmed these general results, finding that typhoid bacilli would live for 10 days in the unsterilized water of Lake Mendota, while they could be isolated only after 5 days when immersed in sewage. Other observers record much greater 24 ELEMENTS OF WATER BACTERIOLOGY viability for the typhoid bacillus. Savage (1905) added a heavy dose of the organism to unsterilized tidal mud and found it living after 5 weeks. Hoffmann (1905), after inoculating a large aquarium with a rich typhoid culture, was able to isolate the germ from the water after four weeks and from the mud at the bottom after two months. Konradi (1904) reports the per- sistence of typhoid bacilli in unsterilized tap water for over a year. These last experiments deal only with the maximum survival period for a few out of great numbers of germs introduced into the water or mud, and entirely ignore the quantitative aspects of the case. When one con- siders the proportion of the original bacteria surviving, the period necessary to bring about a reasonably safe condition is found to be much shorter. Houston (1908) has shown that when water is artificially infected with .typhoid bacilli and stored, 99.9 per cent of the disease germs perish in one week, although some may persist for from i to 9 weeks. In later experiments (Houston, 1911) he finds that " uncultivated " typhoid bacilli added to the water directly from the urinary sediment of a disease carrier perish much more rapidly than the laboratory strains, usually disappearing entirely after one week and always after three. On a number of occasions Dr. Houston gave dramatic expression to his confidence in these negative laboratory findings by drinking half pint portions of water which a few weeks previously had contained millions of typhoid bacilli. We have plenty of practical epidemiological evidence, such as that THE BACTERIA IN NATURAL WATERS 25 offered in the Chicago Drainage Canal case and in the lawsuit over the condition of the water supply of Jersey City, to confirm the general conclusion that any water which has been stored for 4 weeks is practically safe. Bacteria in Ground-waters. In general we have seen that surface-waters tend continually to decrease in bacterial content after their first period of contact with the humus layer of the soil. In that other portion of the meteoric water which penetrates below the surface of the earth to join the reservoir of ground- water, later to reappear as the flow of springs and wells, this diminution is still more marked, since the filtering action of the earth removes not only most of the bac- teria, but much of their food material as well. The numbers of bacteria in the soil itself decrease rapidly as one passes downward. Kabrhel (1906) found several million per c.c. in surface samples of woodland soil, a few thousands or tens of thousands half a meter below, and usually only hundreds in centimeter samples collected at depths greater than a meter. Many observers formerly believed that all ground- waters were nearly free from bacteria, because often no colonies appeared on plates counted after the ordinary short periods of time. If, however, a longer period of incubation be adopted considerable numbers may be obtained. For convenience we may divide ground-waters into three groups, namely: shallow open wells, springs and " tubular " (driven) or deep wells. This division is important because ordinary shallow wells form a group by themselves in respect to the possibility of aerial and 26 ELEMENTS OF WATER BACTERIOLOGY surface contamination, their water often being fairly rich in bacterial life. Egger (Wolff hugel, 1886) examined 60 wells in Mainz and found that 17 of them contained over 200 bacteria to the cubic centimeter. Maschek (Maschek, 1887) found 36 wells out of 48 examined in Leitmeritz which had a bacterial content of over 500 per c,c. Fischer (Horrocks, 1901) reported 120 wells in Kiel which gave over 500 bacteria per c.c. and only 51 with less than that number. In the examination of 147 shallow farmyard wells by one of us (S. C. P.) it was found that 124 of the wells which contained no B. coli, and were therefore probably free from fecal pollution, averaged 190 bacilli per c.c. while 23 which gave positive tests for B. coli averaged 570 per c.c. The distribution of the two series of samples according to the number of bacteria present is indicated in the table below. BACTERIA IN SHALLOW FARMYARD WELLS PERCENTAGE OF SAMPLES IN EACH GROUP Bacteria per c c 0 i- ii- 21- 5i- 101- 501- IOOI- 2OOI- 10 20 50 IOO 500 IOOO 2OOO 3000 Series I. B. coli absent. 3 16 U 16 II 31 5 4 Series II. B. coli present . 5 IO 57 IO 14 5 Very similar results are reported for shallow wells used as farm water-supplies in Minnesota by Kellerman and Whittaker (1909), although the general quality of the wells examined was considerably below that of the series tabulated above. THE BACTERIA IN NATURAL WATERS 27 In the ordinary standard 48-hour period very few bacteria develop from normal spring- waters. Thus in an examination of spring-waters made by the Mas- sachusetts State Board of Health in 1900 (Massachusetts State Board of Health, 1901), of 37 springs which were practically unpolluted and had less than o.io part per 100,000 excess of chlorine over the normal, 54 sam- ples were examined and gave an average of 41 bacteria per c.c. Only 6 samples showed figures over 50. It now remains to consider the other great division of ground-waters, namely, deep, " driven," or " tubular " wells, which, if carefully constructed, should ordinarily be free from all surface-water contamination, and should show low bacterial counts. The results tabulated below obtained by Houston in the examination of a series of deep wells of high quality at Tunbridge Wells are fairly typical. BACTERIAL CONTENT OF DEEP WELL WATERS (HOUSTON, 1903) Bacteria per c.c. 36 6 9 4 i 16 17 4 3 12 2 4 10 5 2 Fifteen driven wells in the neighborhood of Boston, examined in 1903, showed at the end of 48 hours an average of only 18 colonies per c.c.; and the results of certain examinations of other wells and springs, recently made by the authors, are given in the table on page 28. 28 ELEMENTS OF WATER BACTERIOLOGY BACTERIA IN DEEP WELL AND SPRING WATERS Town Bacteria per c.c. Town. Bacteria per c.c. Worcester, Mass Waltham, Mass Newport, R.I IO 3 7 Saranac Lake, N. Y. . Ellenville, N. Y Hyde Park, Mass II o 12 It is plain that water absolutely free from bacteria is not ordinarily obtained from any source. In deep wells, however, their number is small; and the peculiar character of the organisms present is manifested in many cases by the slow development at room tem- perature (frequently no growth until the third day), the entire absence of liquefying colonies, and the abundance of chromogenic species. CHAPTER II THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION OF WATER Relation of the Medium to the Number of Bacteria Obtained. The customary methods for determining the number of bacteria in water do not reveal the total bacterial content, but only a very small fraction of it, as becomes apparent when we consider the large num- ber of organisms, nitrifying bacteria, strict anaerobes, etc., which refuse to grow, or grow only very slowly in ordinary culture media, and which, therefore, escape detection. On the one hand, certain obligate parasites cannot thrive in the absence of the rich fluids of the animal body; on the other hand, the prototrophic bacteria, adapted to the task of wrenching energy from nitrites and ammonium compounds are unable to develop in the presence of so much organic matter. Winslow (1905) in the examination of sewage and sewage effluents, found 20-70 times as many bacteria by microscopic enumeration as by the gelatin plate count. Certain special media enable us to obtain much larger counts than those yielded by the ordinary gelatin method. The Nahrstoff Heyden agar, for example, has been strongly advocated by Hesse (Hesse and Niedner, 1898) and other German bacteriologists upon this 29 30 ELEMENTS OF WATER BACTERIOLOGY ground. In this country Gage and Phelps (Gage and Phelps, 1902) showed that the numbers obtained by the ordinary procedure were only from 5 to 50 per cent of those obtained by the use of Heyden's Nahrstoff agar. For practical sanitary purposes, however, our methods are fairly satisfactory. Within limits, it is of no great importance that one method allows the growth of more bacteria than another. When we are using the quantitative analysis as a measure of sewage pollution the essential thing is that the section of the total bacterial flora which we obtain should be thor- oughly representative of that portion of it in which we are most interested- — the group of the quickly growing, rich-food-loving sewage forms. In this respect meat-gelatin-peptone appears to be unrivalled; and it is in this respect that such media as Nahrstoff agar fail. Miiller (1900) showed that the larger counts obtained by plating on the Nahrstoff medium are due to the fact that it specially favors the more prototrophic forms, among the water bacteria themselves. Intestinal organisms and even the ordinary putrefactive germs, when plated in pure culture, show no higher counts on Nahrstoff agar than on gelatin. Gage and Adams (1904) found by plating pure cultures of the common laboratory bacteria, saprophytes and parasites, that Nahrstoff counts were actually lower than those obtained by the use of gelatin. When sewage and highly polluted waters are examined counts are slightly higher on Nahrstoff media, while with purer waters the Nahrstoff numbers are far in excess of those obtained with gelatin. Winslow (1905) found the ratio of Nahrstoff agar QUANTITATIVE EXAMINATION OF WATER 31 to gelatin count to be 1.7 to i.o for sewage, and 4.8 to i.o for sand filter effluent. With waters of still better quality the ratio goes up higher, reaching a maximum when the bacteria which increase and multiply in water are most abundant. Miiller (1900) found, for example, that water which normally showed six times as many bacteria on Nahrstoff agar as on gelatin might give a Nahrstoff-gelatin ratio of 20-30 after it had been standing for some time in the supply pipes. The table below, taken from the valuable paper by Gage and Phelps (1902), shows strikingly the different Nahrstoff-agar ratios for waters of TABLE SHOWING PERCENTAGES OF BACTERIA DEVELOP- ING ON REGULAR AGAR AND NAHRSTOFF AGAR FOR DIFFERENT CLASSES OF WATERS (GAGE AND PHELPS, 1902) Regular Agar Days' Count. Class of Water. 2 3 4 5 6 7 8 Ground water o 5 6 6 6 6 6 Filtered water 6 7 7 7 7 7 7 Merrimac River. . . 6 7 7 8 8 9 9 Filtered sewage. . . . 14 17 18 iQ iQ 19 19 Sewage 34 44 46 46 46 46 46 Narhstoff Agar Ground water 6 43 78 88 93 IOO IOO Filtered water 37 69 80 92 98 IOO IOO Merrimac River . . . 29 78 93 97 97 99 IOO Filtered sewage. . . . 26 65 93 95 97 99 IOO Sewage 39 75 oc IOO IOO IOO IOO 32 ELEMENTS OF WATER BACTERIOLOGY various grades of purity. It is obvious from all these facts that the effect of using the Nahrstoff medium is to increase disproportionately the bacterial counts obtained from purer waters and thus to diminish the difference in bacterial content between normal and contaminated sources. The ordinary agar and gelatin media, on the other hand, are adapted to the growth of intestinal and putrefactive forms and, therefore, serve best the prime object of bacteriological water examination. The first requisite in a procedure for water analysis is, then, that it should be adapted to the end in view, the differentiation of pure and contaminated waters. The second and equally important requirement is that the procedure should be a standard one, so that results obtained at different times and by different observers may be comparable. In this respect the work of G. W. Fuller, G. C. Whipple, and other members of the Committee on Standard Methods of the American Public Health Association has placed the art of quantitative water analysis in this country in a very satisfactory state by contrast with the varying practices which prevail in England and Germany. The first report on this question was made in 1897 (Committee of Bac- teriologists, 1898). A permanent Committee on Stand- ard Methods was then formed which reported in 1901 (Fuller, 1902), in 1904 (Committee on Standard Methods of Water Analysis, 1905), and again in 1911 (Committee on Standard Methods for the Examination of Water and Sewage, 1912), recommending in considerable detail a standard routine procedure for the quantitative QUANTITATIVE EXAMINATION OF WATER 33 and qualitative bacteriological examination of water for sanitary purposes. These reports have had a far- reaching effect in simplifying and unifying the methods of water analysis. Similar results have followed from the work of the English Committee appointed to con- sider the Standardization of Methods for the Bac- terioscopic Examination of Water which reported in 1904, although this committee unfortunately did not consider the process of media making in great detail. The last report of the American Committee on Standard Methods (1912) will be adhered to in this and succeed- ing chapters unless otherwise specifically stated; and that portion of its report which deals with methods of making media will be found in full in the appendix. Standard Procedure for Quantitative Determination of Bacteria in Water. The procedure for the quantitative determination of bacteria in water consists, in brief, in mixing a definite amount of a suitably collected specimen of the water with a sterile, solidifiable culture medium and incubating it for a sufficiently long time to permit reproduction of the bacteria and the forma- tion of visible colonies which may be counted. The process is divided naturally into four stages — sampling, plating, incubating, and counting. Sampling. All samples of water for bacteriological examination should be collected in clean, sterile bottles with wide mouths and glass stoppers, preferably of the flat mushroom type. It is desirable that these bottles should have a capacity of at least 100 c.c. They should be cleaned thoroughly before using, by treatment with sulphuric acid and potassium bichromate 34 ELEMENTS OF WATEE BACTERIOLOGY or with alkaline permanganate of potash followed by sulphuric acid, dried by draining, and sterilized by dry heat at 160° C. for at least i hour, or by steam at 115-120° for 15 minutes. If not to be used immediately the neck and stopper should be protected against dust or other contamination by wrapping with lead-foil. For transportation the bottle should be enclosed in a suitable case or box. The greatest care must be taken that the fingers do not touch the inside of the neck of the bottle or the cone of the stopper, as the water thereby would become seriously contaminated and rendered unfit for examina- tion. It is well known that bacteria are found abun- dantly upon the skin, and Winslow (Winslow, 1903) has shown that even B. coli is present upon the hands in a considerable number of cases. In order to obtain a fair sample, great precautions must be taken, and these will vary with the different classes of waters to be examined and with local condi- tions. If a sample is to be taken from a tap, the water should be allowed to flow at least five minutes (if from a tap in regular use) or for a longer period in case the water has been standing in the house-service system. In the small pipes, changes in bacterial content are liable to occur, certain species dying and others mul- tiplying. If a sample is to be taken from a pump similar pre- cautions are necessary. The pump should be in con- tinuous operation for 5 minutes at least, and preferably for half an hour before the sample is taken, in order to avoid excessively high numbers due to the growth of QUANTITATIVE EXAMINATION OF WATER 35 bacteria within the well and pump, the bacterial con- dition of the water as it passes through the ground being what we wish to determine. Thus Heraeus (Heraeus, 1886) in a well-water which had been but little used during the preceding 36 hours found 5000 organisms per c.c.; when the well was emptied by continuous pumping, a second sample, after an interval of half an hour, gave only 35. Maschek (Tiemann and Gartner, 1889) obtained similar results, shown in the following table: EFFECT OF PUMPING ON THE BACTERIAL CONTENT OF WELL-WATER Well-water after continuous pumping for fifteen minutes . . 458 many hours 140 later 68 after continuous pumping for fifteen minutes . . 578 many hours 1 79 later 73 After a proper interval of pumping the sample of a well-water may be collected from the pet-cock of the pump or from a near-by tap. With a hand-pump, such as is found in domestic shallow wells, the water is, of course, pumped directly into the sample bottle. The difficulties in securing an average sample from this latter source are often great, since if the flooring about the pump is not tight, as is usually the case, con- tinued pumping may wash in an unusual amount of surface pollution. In sampling surface-waters, the greatest precautions must be observed to prevent contamination from the fingers. In still waters the fairest sample is one taken 36 ELEMENTS OF WATER BACTERIOLOGY from several inches down, as the surface itself is likely to have dust particles floating upon it. The method most frequently recommended is to plunge the bottle mouth downward to a depth of a foot or so, then invert and allow the bottle to fill. Whenever any current exists, the mouth of the bottle should be directed against it in order to carry away any bacteria from the fingers. If there is no current, a similar effect can be produced by turning the bottle under water and giving it a quick forward motion. In rapidly flowing streams it is only necessary to hold the bottle at the surface with the mouth pointed up-stream. For taking samples of water at greater depths, a number of devices have been employed, all of which are fairly satisfactory. The essentials are, first, a weight to carry the bottle down to the desired depth, and, second, some method of removing the stopper when that depth is reached. The student will find one good form of apparatus described in Abbott's " Principles of Bacteriology" (Abbott, 1899); an admirable one was devised by Hill and Ellms (Hill and Ellms, 1898); and Thresh (1904) figures an ingenious device for the same purpose. Miquel and Cambier (Miquel and Cambier, 1902) and other authors recommend the use of a sealed glass bulb with a capillary tube which can be broken off at the desired moment. Drew (1912) has devised an interesting sampling apparatus for use at great depths in the sea. Changes in Bacterial Numbers after Sampling. As soon as a sample of water is collected its conditions QUANTITATIVE EXAMINATION OF WATER 37 of equilibrium are upset and a change in the bacterial content begins. Even in the purest spring-waters, which contain but few bacteria when collected, and in which the amount of organic matter is infinitesimal, enormous numbers may be found after storage under laboratory conditions for a few days or even a few hours. In some cases the rise in numbers is gradual, in others very rapid. The Franklands (Frankland, 1894) record the case of a deep-well water in which the bacteria increased from 7 to 495,000 in 3 days. Miquel (Miquel, 1891) from his researches, arrived at the conclusion that in surface-waters the rise is less rapid than in waters from deep wells or springs, and that in the latter case the decrease, after reaching a maximum, is likewise rapid and steady. Just how far protection from light, increase in temperature, and a destruction of higher micro-organisms is responsible for the increase, and to what extent an exhaustion of food-supply or the formation of toxic waste products causes the succeeding decrease, we are not aware; but the facts are well established. Whipple has exhaustively studied the details of this multiplication of bacteria in stored waters and has shown in the table given below that there is first a slight reduction in the number present, lasting perhaps for 6 hours; followed by the great increase noted by earlier observers. It is probable that there is a constant increase of the typical water bacilli, overbalanced at first by a reduction in other forms, for which the environment is unsuitable. 38 ELEMENTS OF WATER BACTERIOLOGY BACTERIAL CHANGES IN WATER DURING STORAGE (WHIPPLE, 1901) Sample Initial Temper- ature. Temp, of Incu- bation of Sample. Number of Bacteria per c.c. Initial. After 3 Hours. After 6 Hours. After 24 Hours. After 48 Hours. C. C. A 7.6° 17.0° 260 215 230 900 27,000 B 7-6° 17.0° 260 245 255 720 10,850 C 7.6° 12.5° 260 270 231 600 2,790 D 7.6° 12.5° 260 270 245 710 1, 800 E 7.6° 2.4° 260 243 210 675 1,980 F 7.6° 2-4° 260 235 270 560 1,980 G 11.0° 12.8° 77 55 58 101 10,250 H 11.0° 12.8° 77 53 74 87 2,175 I 11.0° 23.6° 77 5i 52 11,000 41,400 J 6-7° 20.0° 430 375 245 385,ooo1 K 6.7° 20.0° 430 345 405 75o,ooo1 L 23-2° 23.0° 510 340 230 8,000 20,000 M 23-2° 2-5° 525 300 220 380 2,200 1 0.0005 Per cent peptone added to the water. WolfThiigel and Riedel (Wolffhiigel and Riedel, 1886) noted the dependence of this multiplication on the air-supply, vessels closed with rubber stoppers showing lower numbers than those plugged with cotton. Similarly, Whipple found that the multiplication of bacteria was much greater when bottles were only half full than when they were filled completely; and also, as shown in the very striking table on page 39, that the size of the bottle markedly influenced the growth. An important series of investigations by Kohn (1906) suggests that this phenomenon of multiplication dur- ing storage may be due in part to the solution of certain constituents of glass which favor bacterial life, since the increase is notably greater in bottles of the more soluble glasses. QUANTITATIVE EXAMINATION OF WATER 39 EFFECT OF SIZE OF VESSEL UPON THE MULTIPLICATION OF WATER BACTERIA DURING STORAGE (WHIPPLE, 1901) Sample Bottle. Temp, of Incuba- tion. Number of Bacteria per c.c. Ini- tial.1 After 3 Hrs. After 6 Hrs. After 12 Hrs After 24 Hrs. After 48 Hrs. C A i -gallon 13° 77 63 65 47 42 175 B 2-qutirt 13° 77 59 63 60 45 690 C i-quart 13° 77 63 63 47 46 325 D i-pint 13° 77 57 61 36 38 630 E 2-ounce 13° 77 55 58 47 IOI 10.250 F i -gallon 24° 77 81 97 275 290 300 G 2-quart 24° 77 92 59 62 1 80 250 H i-quart 24° 77 84 77 46 340 900 I i-pint 24° 77 51 46 100 2,950 7,O2O J 2-ounce 24° 77 5i 52 U5 11,000 41,400 1 Average of five plates. Whipple's table, quoted above, shows that the multi- plication during storage was greater at a higher tem- perature; and this is a well-recognized general rule. In order to obviate the abnormal results of storage increase it is therefore obvious that samples must be examined shortly after collection and that they must be kept cool during their necessary storage. If fairly pure waters are placed upon ice and kept between o degrees and 10 degrees, they will show no material increase in 12 hours. With polluted water, however, another danger is here introduced. Samples of such water when packed in ice show a marked decrease due to the large number of sensitive intestinal bacteria present. Jordan (Jordan, 1900) found that three samples of river-water packed in ice for 48 hours fell 40 ELEMENTS OF WATER BACTERIOLOGY off from 535,000 to 54,500; from 412,000 to 50,500, and from 329,000 to 73,000, respectively. It is, there- fore, important that even iced samples should not be kept too long; and it is desirable to adhere strictly to the recommendations of the Standard Methods Committee that the interval between sampling and examination should not exceed 12 hours in the case of relatively pure waters, 6 hours in the case of relatively impure waters, and i hour in the case of sewage. Plating. The bottle containing the sample of water is first shaken at least twenty-five times in order to get an equal distribution of the bacteria. If the num- ber of bacteria present is probably not greater than 200, i c.c. is then withdrawn with a sterile i c.c. pipette and delivered into a sterile Petri dish of 10 cm. diameter. To this is added 5 c.c. of standard 10 per cent gelatin at a temperature of about 30° C., or standard agar (7 c.c.) at 40-42° C. Should the number of bacteria per c.c. probably exceed 200, dilution is necessary. This is best accomplished by adding i c.c. of the water in question to 9, 99 or 999, etc., c.c. of sterile tap water according to the amount of dilution required. The diluted sample is then shaken thoroughly and i c.c. taken for enumeration. In order to determine the number of bacteria originally present it is only neces- sary to multiply by the factor 10, 100, or 1000, etc. When a sample of water from an unknown source is to be examined it is generally desirable to make two check plates at each of the above dilutions, select- ing those which give nearest to 200 colonies on the plates after incubation as the ones on which to rely QUANTITATIVE EXAMINATION OF WATER 41 for the count. A much smaller number will not give average figures, and if more than 200 colonies are present on a plate many bacteria will be checked by the waste products of those which first develop and the count obtained will be too low. After the addition of the diluted sample and the nutrient medium, their thorough mixture in an even layer on the bottom of the plate is obtained by careful tipping and rotation. It was formerly customary to mix the water with the gelatin in the tube before pouring into the plate, but this method is objectionable because there is always a residuum of medium remaining in the tube which will retain varying numbers of bacteria and thus interfere with the accuracy of the count. Before pour- ing the medium into the plate the mouth of the tube should be flamed to remove any possibility of con- tamination. The usual method of determining the number of bacteria in water for sanitary purposes in Germany, England and the United States has always been by the use of gelatin plates with a 2 -day incubation period at 20 degrees. The 1905 Standard Methods Report of the American Public Health Association Committee recommended this procedure, which has been universally adopted. The 1912 Report, however, suggests the use of agar with a i-day period at 37 degrees, as yielding quicker results and indicating the presence of bacteria more nearly related to pathogenic types. The com- parative value of the two methods has been well dis- cussed by Whipple (1913). The use of gelatin is not only more time-consuming, but requires the use of a 42 ELEMENTS OF WATER BACTERIOLOGY special 2o-degree incubator which is difficult to regulate. The 37-degree incubator must be provided in any case for the isolation of B. coli. On the other hand, the time seems hardly ripe for the abandonment of the 2o-degree count, which has been used for 20 years all over the civilized world, and for the interpretation of which we have very complete data. There is at present no such sound basis for interpreting the 37-degree count, and in many cases, as in the control of water nitration plants, the 37-degree numbers are too small to be of any practical value. Furthermore the 20- degree count may furnish evidence of surface con- tamination as distinguished from fecal pollution, which is often of considerable value. The authors have always urged the use of the 37- degree count along with the 2o-degree count as furnish- ing most valuable information; but this is very different from the substitution of one count for the other. The recommendation that the 2o-degree count be abandoned, with no evidence to warrant such a revolu- tionary change, and no experimental results on which to base an interpretation of the 37-degree count, has aroused vigorous opposition from a large majority of practical water bacteriologists. At the Washington meeting of the American Public Health Association in September, 1912, it was resolved " that in the opinion of the Laboratory Section of the American Public Health Association, ordinary routine examination of water for sanitary purposes, and in the control of purification plants, for the present should include the determination of the number of bacteria developing QUANTITATIVE EXAMINATION OF WATER 43 at 20 degrees and at 37 degrees and a presumptive test for B. coli in lactose bile." This action of the section responsible for the appoint- ment of the Standard Methods Committee appears to supersede the report of the committee itself and makes the combination of the 20- and 37-degree counts the standard American procedure. The 2o-degree count may be made on either gelatin or agar; but it is the 2o-degree count which will be discussed in this chapter, leaving the body temperature count for consideration in Chapter IV. The exact composition of the medium is, of course, of prime importance in controlling the number of bacteria which will develop. The figures previously cited in connection with the discussion of Hesse's Nahrstoff agar show how bacterial counts may vary with media of widely different composition. The table quoted on page 44 from Gage and Phelps (1902), shows the considerable differences which may be due to the presence or absence of meat infusion, peptone, etc., in media of generally similar character (compare the figures for plain gelatin, peptone, gelatin, and meat gelatin). Much slighter variations than this, however, are significant. The reaction of the medium was found as early as 1891 to be important, for Reinsch (Reinsch, 1891) showed in that year that the addition of one one-hundredth of a gram of sodium carbonate to the liter increased sixfold the number of bacteria develop- ing. Fuller (Fuller, 1895) and Sedgwick and one of us (Sedgwick and Fresco tt, 1895), working indepen- dently, established the fact that an optimum reaction 44 ELEMENTS OF WATER BACTERIOLOGY existed for most water bacteria and that a devia- tion either way decreased the number of colonies developing. TABLE SHOWING PERCENTAGES OF BACTERIA DEVEL- OPING ON MEDIA OF DIFFERENT COMPOSITIONS (GAGE AND PHELPS, 1902) Medium. Days' Cour t. 2 3 4 5 6 7 8 9 Nahrstoff agar IQ 60 78 8* 0 ^ QO OQ IOO Nahrstoff peptone agar Peptone agar Meat agar 10 II 8 22 16 13 26 22 16 28 23 30 24 I 7 30 24 I 7 3° 24 1 7 30 24 I ^ Plain agar. ... 8 10 I? 14 14 14 14 14 Regular agar 7 g II II 1 1 II II 1 1 Nahrstoff glycerin agar Nahrstoff meat agar 6 7 10 7 II 8 II 8 II IO II IO II IO II IO Meat gelatin 12 IQ 24 06 26 06 26 26 Peptone gelatine 7 12 18 20 20 2O 20 2O Standard gelatin . 8 IO ii 12 17 I 3 I 3 I 3 Plain gelatin i 6 12 12 I ^ I?, 13 13 Nahrstoff gelatin 5 6 g 11 13 13 12 13 Whipple (Whipple, 1902) has shown that not only the particular kind of gelatin used, but its exact physical condition as affected by sterilization and other previous treatments, will materially affect the results obtained. Gage and Adams (1904) found marked differences in counts as the result of the use of the two best-known commercial peptones. A long series of waters plated on agar made up with Merck's and Witte's peptones, respectively, showed the average relative results in the table on page 45. QUANTITATIVE EXAMINATION OF WATER 45 AVERAGE RELATIVE NUMBER OF BACTERIA ON PEP- TONE AGAR WITH DIFFERENT PEPTONES (GAGE AND ADAMS, 1904) DAYS 2 4 6 8 10 12 Merck's •2-2 ci 67 80 08 Witte's 38 r? IOO JOO IOO IOO The same authors showed that the composition of the water used exercised a marked selective action upon the development of bacteria. Agar made up with sewage permitted a maximum growth of sewage bacteria and showed no colonies when inoculated with filtered city water. On the other hand agar made up with city water showed 100 per cent of the bacteria present in city water and river water, three-quarters of those present in sewage and less than half of those present in sewage effluents. Hesse (1904) found that the number of bacteria developing on Nahrstoff agar varied with the composi- tion of the glass tubes in which the media had previously been sterilized. The more soluble glasses yielded sufficient alkali to the medium to inhibit four-fifths of the bacteria present in certain cases. All these facts make it evident that only the strictest adherence to a standard method can ensure comparable results; the ordinary nutrient gelatin or agar should then in all practical sanitary work be made up from distilled water, meat infusion, peptone and gelatin or agar, in exact accordance with the directions of the Standard Methods Committee. 46 ELEMENTS OF WATER BACTERIOLOGY Even the standard procedure fails to ensure uniformity in one important respect. The meat infusion which it calls for is in itself a highly variable quantity. Gage and Adams (1904), in the examination of fifteen lots of beef infusion, found variations of nearly i per cent in organic solids (calculated on the weight of the whole infusions), after the final filtration. The organic constituents of the meat infusion varied, therefore, among themselves by nearly the total amount of pep- tone added. It is to be hoped that the standard methods may soon be so revised as to eliminate this necessarily uncertain constituent of nutrient media. Criticisms of detail must, however, give way to the importance of securing fairly comparable results; and the con- fusion which would follow the use by individual bac- teriologists of media made without meat would out- balance the errors inherent in the standard procedure. Incubation. Incubation should take place in a dark, well-ventilated chamber where the temperature is kept substantially constant at 20 degrees and where the atmosphere is practically saturated with moisture. It has been shown by Whipple (Whipple, 1899) and others that the number of bacteria developing in plate cultures is to a certain extent dependent upon the presence of abundant oxygen and moisture. Thus, reckoning the number of bacteria developing in a moist chamber at 100, the percentage counts obtained in an ordinary incubator were as follows: 75 when the relative humid- ity of the incubator was 60 per cent of saturation; 82 when it was 75 per cent; 98 when it was 95 per cent. This source of error may be avoided by the use of ven- QUANTITATIVE EXAMINATION OF WATER 47 tilated dishes and by the presence of a pan of water in the incubating chamber. According to American and German practice, plates made for sanitary water analysis are counted at the end of 48 hours. The English Committee appointed to consider the standardization of methods for the Bacterioscopic Examination of Water (1904) fixed the time at 72 hours. French bacteriologists, and some Germans (Hesse and Niedner, 1906), still recommend longer periods, and the following table from Miquel and Cambier (Miquel and Cambier, 1902) shows that many bacteria fail to appear in our ordinary procedure. It is, however, in the main, the characteristic water bacteria which develop slowly, sewage bacteria almost without exception being rapid growers. The longer period of incubation is, therefore, not only inconvenient, but undesirable, since it obscures the difference between good and bad waters. EFFECT OF THE LENGTH OF INCUBATION OF WATER BACTERIA IN GELATIN UPON THE NUMBER OF COLONIES DEVELOPING (MIQUEL AND CAMBIER, 1902) Length of Incubation. Colonies Developed. Length of Incubation. Colonies Developed. I day 2O o days 821 2 days 136 10 days . . 8qo 7 days 2 CA 1 1 days 802 4 days. . . 187 12 days Q2I 5 days S^o i 3 days . . QCI 6 days . 6*7 Mdays 076 7 days 72CJ i ^ days . . IOOO 8 days . . 780 48 ELEMENTS OF WATER BACTERIOLOGY Counting. The number of bacteria is determined by counting the colonies developed upon the plate, with the aid of a lens magnifying at least five diameters. For convenience in counting the plate may be placed upon a glass plate ruled in centimeter squares and set over a black tile; or the tile itself may be ruled. As has already been said, it is desirable that the number of colonies should not exceed 200, for when the number is very high the colonies grow only to a small size, making counting laborious and inaccurate, and many do not develop at all. The best results are obtained with numbers ranging from 50 to 200. When it is possible to do so, all the colonies on the plate should be counted. When they exceed 400 or 500 it is often easier, and fully as accurate, to count a fractional part of the plate and estimate the total number therefrom. This should not be done, however, except in case of necessity. Ayers (1911) has suggested two counting devices which will be found very useful where a great many plates have to be handled. For getting the best possible transmitted light, he places his plate on the ground- glass top of a wooden box, 7 inches square, with one side open to admit light, which is reflected upward by a plane mirror set in the box at an angle of 45 degrees. An ordinary graduated-glass counting plate may be placed between the ground-glass and the Petri dish, and the eyes are protected from direct light by a screen rising from the open side of the box. For picking colonies from a gelatin plate in a warm room, he places between the ground glass and the Petri dish a copper QUANTITATIVE EXAMINATION OF WATER 49 box with top and bottom of glass 7 inches square and ij inches deep, through which cold water is allowed to circulate. Expression of Quantitative Results. It is customary in determining numbers to make plates in duplicate, thereby affording a check upon one's own work. Owing to the lack of precision in the method, the limit of experimental error is a wide one. It should be possible for careful manipulators to obtain results within 10 per cent of each other, but a closer agreement than this is hardly to be expected. It has been suggested by the committee of the American Public Health Association that the following mode of expressing results be adopted in order to avoid the appearance of a degree of accuracy which the methods do not warrant. NUMBERS OF BACTERIA FROM 1-50 shall be recorded to the 51-100 101-250 251-500 501-1000 1001-10,000 10,001-50,000 50,001-100,000 100,001-500,000 500,001-1 ,000,000 i ,000,001-5,000,000 nearest unit 5 10 25 50 IOO 500 1,000 10,000 50,000 100,000 The determination of numbers of bacteria in water in the field has frequently been attempted. Since the laboratory method of "plating out" is difficult to use in field work, the Esmarch tube process has often been employed. This consists in introducing into a tube of melted gelatin or agar i c.c. of the water and then 50 ELEMENTS OF WATER BACTERIOLOGY rotating the tube until the medium has solidified in a thin layer on the inner wall. Other bacteriologists have devised ingenious field kits for adapting the plate method to this purpose, of which one very good form has recently been described by Van Buskirk (1912). The opportunity for air infection in work done outside a proper laboratory is, however, always great; and it is almost impossible to secure proper conditions for incubation in any makeshift establishment. On the whole, the authors are of the opinion that laboratory examinations are to be preferred to those made in the field, if a laboratory can be reached within 12 hours or so of the time of collection of the samples. CHAPTER III THE INTERPRETATION OF THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION Standards for Potable Water. The information fur- nished by quantitative bacteriology as to the antecedents of a water is in the nature of circumstantial evidence and requires judicial interpretation. No absolute stand- ards of purity can be established which shall rigidly separate the good from the bad. In this respect the terms " test " and " analysis " so universally used are in a sense inappropriate. Some scientific problems are so simple that they can be definitely settled by a test. The tensile strength of a given steel bar, for example, is a property which can be determined. In sanitary water examination, however, the factors involved are so complex, and the evidence neces- sarily so indirect, that the process of reasoning much more resembles a doctor's diagnosis than an engineering test. The older experimenters attempted to establish arbitrary standards, by which the sanitary quality of a water could be fixed automatically by the number of germs alone. Thus Miquel (Miquel, 1891) published a table according to which water with less than 10 bac- teria per c.c. was " excessively pure," with 10 to 100 51 52 ELEMENTS OF WATER BACTERIOLOGY bacteria, " very pure/' with 100 to 1000 bacteria, " pure/' with 1000 to 10,000 bacteria, " mediocre," with 10,000 to 100,000 bacteria, " impure," and with over 100,000 bacteria, " very impure." Few sanitarians would care to dispute the appropriateness of the titles applied to waters of the last two classes; but many bac- teriologists have placed the standard of " purity " much lower. The limits set by various German observers range, for example, from 50 to 300. Dr. Sternberg (Sternberg, 1892) in a more conservative fashion, has stated that a water containing less than 100 bacteria, is presumably from a deep source and uncontaminated by surface drainage; that one with 500 bacteria is open to suspicion; and that one with over 1000 bac- teria is presumably contaminated by sewage or surface drainage. This is probably as satisfactory an arbitrary standard as could be devised, but any such standard must be applied with great caution. The source of the sample is of vital importance in the interpretation of analyses; a bacterial count which would condemn a spring might be quite normal for a river; only figures in excess of those common to unpolluted waters of the same character give an indication of danger. Fur- thermore, the bacteriological tests are far more delicate than any others at our command, very minute addi- tions of food material causing an immense multiplica- tion of the microscopic flora. This delicacy necessarily requires, both in the process of analysis and the inter- pretation of results, a high degree of caution. As pointed out in the previous chapter, the touch of a finger or the entrance of a particle of dust may wholly QUANTITATIVE BACTERIOLOGICAL ANALYSIS 53 destroy the accuracy of an examination. Even the slight disturbance of conditions incident upon the storage of a sample after it has been taken may in a few hours wholly alter the relations of the contained microbic life. It is necessary, then, in the first place, to exercise the greatest care in allowing for possible error in the collection and handling of bacteriological samples; and in the second place, only well-marked differences in numbers should be considered significant. In the early days of the science, discussion ran high as to the interpretation of bacteriological analysis; and particularly as to the relation of bacterial numbers to the organic matter present in a water. Different observers obtained inconsistent results, and Bolton (Bolton, 1886) concluded that there was no relation whatever between the organic pollution of a water and its bacterial content. Tiemann and Gartner (Tiemann and Gartner, 1889) furnished the key to the difficulty in their statement that there are two classes of bacteria, the great majority of species normally occurring in the earth or in decomposing organic matter, which require abundance of nutriment, and certain peculiar water bacteria which can multiply in the presence of such minute traces of ammonia as are present in ordi- nary distilled water. Even these prototrophic or semi-prototrophic forms, however, require a definite amount of food of their own kind. Kohn (1906) determined the minimal nutrient mate- rial requisite for certain of them and found that they could develop in the presence of 198 X io~10 to I98X io"13 per cent of dextrose, 66Xio~13 to 66Xio"17 per cent 54 ELEMENTS OF WATER BACTERIOLOGY ammonium sulphate and 66Xio~13 to 66Xio~19 per cent ammonium phosphate. Similar minute amounts of organic matter are found in the purest of natural waters and under exceptional conditions certain species of bacteria may therefore multiply in bottled samples, or, at times, in a well or the basin of a spring. In normal surface-waters, such growths of the prototrophic forms do not apparently occur. Here it is found as a matter of practical experience that the number of bac- teria present depends upon the extent to which the water has been contaminated with decomposing organic matter, either by pollution with sewage or by contact with the surface of the ground. The bacterial content varies as the extent and character of the contamination varies. It measures not merely organic matter, but organic matter in a state of active decay, and like the ammonias and other features of the sanitary chemical analysis, indicates fresh organic pollution, with the added advantage that the presence of the stable nitrogenous compounds often present in peaty waters introduces no error in the bacteriological analysis. Bacterial Content of Surface-waters. In judging of a surface-water the student will be aided by reference to the figures given for certain normal sources in Chapter I; the Boston tap water with 50 to 200 bacteria per c.c. (Philbrick, 1905) and the water of Lake Zurich with an average of 71 in summer and 184 in winter (Cramer, 1885) may be taken as typical of good potable waters; and numbers much higher than these are open to suspicion, since all contamination whether contributed , by sewage or by washings from the surface of the QUANTITATIVE BACTERIOLOGICAL ANALYSIS 55 ground is a possible source of danger. The excess of bacteria in surface-waters during the spring and winter months is by no means an exception to the general rule that high numbers are significant, since the peril from supplies of this character is clearly shown by the spring epidemics of typhoid fever which at the times of melting snow visit communities making use of unpro- tected surface-waters. Streams receiving direct con- tributions of sewage exhibit a similar excess of bacteria at all times, numbers rising to an extraordinary height near the point of pollution and falling off below as the stream suffers dilution and the sewage organisms perish. Miquel (Miquel, 1886) records 300 bacteria per c.c. in the water of the Seine at Choisy, above Paris; 1200 at Bercy in the vicinity of the city, and 200,000 at St. Denis after the entrance of the drainage of Paris. Prausnitz (Prausnitz, 1890) found 531 bacteria per c.c. in the Isar above Munich, 227,369 near the entrance of the principal sewer, 9111 at a place 13 kilometers below the city, and 2378 at Freising, 20 kilometers further down. Jordan (Jordan, 1900), in his study of the fate of the sewage of Chicago, found 1,245,000 bacteria per c.c. in the drainage canal at Bridgeport, 650,000 at Lockport, 29 miles below, and numbers steadily decreasing to 3660 at Averyville, 159 miles below the point of original pollution. Below Avery- ville the sewage of Peoria enters and the numbers rise to 758,000 at Wesley City, decreasing to 4800 in 123 miles flow to Kampsville. Brezina (1906) found 1900 bacteria per c.c. in the Danube River above, and 1 10,000 at the north of the Danube canal. This number 56 ELEMENTS OF WATER BACTERIOLOGY fell to 85,000 one kilometer below, 62,000 four kilo- meters below, and 40,000 seven kilometers down the stream. Vincent (1905) records from 1000 to 46,000 bacteria per c.c. in the waters of more or less polluted French rivers. Mayer (1902), on the other side of the world, found 21 and 35 bacteria per c.c. in the Shaho River, near its source, in the vicinity of the great Chinese Wall and from 100,000 to 600,000 in the highly polluted Whangpo near its mouth. Bacterial Content of Ground-waters. In ground- waters we have seen that bacteria may occasionally be present in considerable numbers, but if so they are generally organisms of a peculiar character, incapable of development on the ordinary nutrient media in the standard time. Thus in 48 hours we often obtain counts measured only in units or tens such as have been recorded in Chapter I. When higher numbers are present, the general character of the colonies must be taken into account, since besides the slowly-growing forms certain other, water bacteria, which require a comparatively small amount of nutriment, may multiply at times in a deep well or the basin of a spring. In such a case, however, the appearance of the plates at once reveals the peculiar conditions, for the colonies are of one kind and that distinct from any of the sewage species. Thus Dunham (Dunham, 1889) reports that the mixed water from a series of driven wells gave 2 bacteria per c.c., while another well, situated just like the others, contained 5000, all belonging to a single species common in the air. Except in such peculiar cases as this high numbers in a ground-water mean contamination. QUANTITATIVE BACTERIOLOGICAL ANALYSIS 57 Bacteria in Filtered Waters. The process of slow sand filtration for the purification of unprotected surface-water is essentially similar to the action which takes place in nature when rain soaks through the ground to appear in wells and springs; and it is in the examination of the effluent from such municipal plants that the quantitative bacteriological analysis finds, perhaps, its most important application. The chemical changes which occur in the passage of water through sand at a rate of 1,000,000 or 2,000,000 gallons per acre per day are so slight as to be negligible. The bacteria present should, however, suffer a reduction of 98 or 99 per cent, and their numbers furnish the best standard for measuring the efficiency of such filtration plants. At Lawrence, in 1905, Clark found an average of 12,700 bacteria per c.c. in the raw water of the Merrimac River, while the number present in the filtered water was only 70 (Massachusetts State Board of Health, 1906). Where the number of bacteria in the applied water is smaller it is difficult to obtain so high a per- centage efficiency. At Washington, for example, pro- longed sedimentation generally reduces the bacterial numbers to less than a thousand and it is almost impos- sible to secure a 99 per cent removal. The actual numbers of bacteria in the effluent are, however, much lower than at Lawrence. The monthly average results obtained for a year at these two plants are tabulated on page 58. Mechanical filtration gives similar results. Fuller at Cincinnati (Fuller, 1899) records 27,200 organisms per c.c. in the water of the Ohio River between 58 ELEMENTS OF WATER BACTERIOLOGY September 21, 1898, and January 25, 1899, while the average content of the effluent from the Jewell filter was 400. Data with regard to the operation of mechanical niters are now abundant, since all over the world the operation of these plants is controlled by bacteriological methods. Recently Johnson (1907) has reported some interesting results from the far East. At Osaka, Japan, an average of 200 bacteria per c.c. in the raw water of the Yodo River was reduced, in 1905, to an average of 25 by slow sand niters; at Bethmangala, India, in 1906, mechanical niters treated the water of the 'Palar River, containing 4350 bacteria per c.c., and yielded an effluent with only 13 per c.c. (Johnson, 1907). The average monthly results obtained with the new mechanical filter plant at Harrisburg, Pa., are included in the table below for comparison with the figures REMOVAL OF BACTERIA BY NATURAL SAND FILTERS AND MECHANICAL FILTERS. BACTERIA PER C.C. IN APPLIED WATER AND EFFLUENT. MONTHLY AVER- AGES Month. Washington, 1906. Lawrence, 1905. Harrisburg, 1906. Applied Water. Effluent. Applied Water. Effluent. Applied Water. Effluent. January .... 1500 39 14,200 no 9,510 104 February . . . 550 16 14,800 55 21,228 298 March 650 19 10,300 55 3^326 75 April 400 22 3,600 170 39-905 42 May 65 17 1,900 12 6,187 86 June o 22O 17 9,600 9 2,903 3i July 1 60 26 3^0° 55 685 10 August 190 14 19,500 37 1,637 5 September . . I30 14 13,500 44 836 | 12 October 275 16 39,800 110 7-575 63 November . . 22O 12 8.700 70 26,224 236 December.. . . 700 45 • • ! 37,525 163 QUANTITATIVE BACTERIOLOGICAL ANALYSIS 59 recorded at Washington and Lawrence; and these may be taken as typical, since the Harrisburg plant is the latest of its type, as the Washington plant is the newest and most perfectly equipped of slow sand niters. In well-managed purification plants the bacteria in the effluent are determined daily, and any deviation from the normal value at once reveals disturbing factors which may impair the efficiency of the process. In Prussia official regulations demand such systematic examinations and prescribe 50 as the maximum number of bacteria allowable in the filtered water. In the same way the condition of an unpurified surface supply may be determined by daily bacteriological analyses and warnings of danger issued to the public, as has been done at Chicago and other cities. In general, any regular determination of variations from a normal standard furnishes ideal conditions for the bacteriological methods; and the detection by Shuttleworth (Shuttle- worth, 1895) of a break in a conduit under Lake Ontario by a rise in the bacteria of the Toronto water-supply may be cited as a classic example of its application. Often, however, the expert is called to pass upon the character of a water of which no series of analyses is available. In such cases an inspection of the location from which the water comes should be insisted on, as a sound interpretation of a water analysis can only be made with a reasonably full knowledge of the source of the sample. After a careful sanitary inspection, however, the comparison of the result of even a single examina- tion with the normal range for waters of the same class 60 ELEMENTS OF WATER BACTERIOLOGY may prove of great significance, as a few practical examples may make clear (Winslow, 1901). In the spring of 1900 the city of Hartford, Conn., was using a double supply, from the Connecticut River and from a series of impounding reservoirs among the hills. A single series of plates showed from 4000 to 7000 bacteria per c.c. in the water of the river, while the reservoir water contained 300 to 900. The abandon- ment of the river supply followed, and at once the excessive amount of typhoid fever in the city was curtailed. In the fall of 1900, Newport, R. I., experienced an outbreak of typhoid fever, and when suspicion was thrown upon the surface water-supply, chemical analysis of the latter was not wholly reassuring; but there were only 334 bacteria per c.c. in the water from the taps, while a well in the infected district gave 6100. It was no surprise to find, on a further study of the epidemic, that the well was largely at fault and the public supply was not. In the case of ground-water the evidence is usually even more distinct. At Framingham, Mass., in 1903, high chlorin content in the public supply, drawn from a filter gallery beside a lake, had led to public anxiety. Five samples from different parts of the system showed averages of i, 2, 2, 2, and 4 bacteria per c.c.; and taking this in conjunction with the other features of the bacteriological analysis, it was possible to report that any pollution introduced upon the gathering ground had at the time of examination been entirely removed. CHAPTER IV DETERMINATION OF THE NUMBER OF ORGANISMS DEVELOPING AT THE BODY TEMPERATURE Relation between Counts Made at 20° and 37°. The count of colonies upon the gelatin plate measures, as we have pointed out, the number of the metatrophic bacteria in general; and the distribution of these forms corresponds with the decomposition of organic matter wherever it may occur. In this great class, there are some species which will grow under a wide variety of conditions. These are present in most waters in small numbers, and in sources contaminated with wash from decaying vegetable matter they occur in abundance. Other metatrophic forms, however, through a semi- parasitic mode of life, have become specially adapted to the peculiar conditions characteristic of the animal body; and these bacteria possess the property of develop- ing most actively at the temperature of the human body, 37° C., which altogether checks the growth of the majority of normal earth and water forms. The determination of the number of organisms growing at the body temperature may throw light, then, on the presence of direct sewage pollution, since the bacteria from the alimentary canal flourish under such con- ditions, while most of those derived from other sources do not. Savage classifies the bacteria which may 61 62 ELEMENTS OF WATER BACTERIOLOGY be found in water under three headings: normal inhab- itants, like B. fluorescens; unobjectionable aliens (from soil), like B. mycoides, and objectionable aliens (from excreta), like B. coli. The first sort and many of the second sort are generally unable to grow at 37 degrees. This criterion is not an absolute one. Savage, (1906) reports an experiment in which unpolluted soil, which had not been manured or cultivated for at least 3 years, was added to tap water, with the result that a 20° count of 76 was increased to 1970, and a 37° count of 3 was raised to 1630. In this case most of the bacteria in the soil were capable of development at body temperature. Experience shows, however, that the numbers of such bacteria which actually reach natural waters from such sources are seldom large. The count at 37°, therefore, helps to distinguish con- tamination by wash of the soil of a virgin woodland from pollution by excreta, since in the former case the proportion of blood-temperature organisms is much smaller than in the latter. Furthermore, this method is free from much of the error introduced by the mul- tiplication of bacteria after the collection of a sample, as most of the forms which grow in water during storage cannot endure the higher temperature and conse- quently do not develop upon incubation. Recently, for example, water from a spring of good quality was shipped to the laboratory from a considerable distance. Gelatin plates showed 4200 bacteria per c.c., but agar plates at 37° were sterile. Significance of the 37° Count. A majority of the English Committee appointed to consider the standard- DETERMINATION OF ORGANISMS 63 ization of methods for water examination (1904) recommended the body-temperature count as a stand- ard procedure. The American Committee on Standard Methods in its 1905 Report did not recommend this method even for alternative use. In its last report (1912), however, it substituted the 37° for the 20° count, which was dropped out entirely. As we have pointed out in Chapter II, this course seems to us an unwise one, and it was formally condemned at the meeting of the Laboratory Section of the American Public Health Association in September, 1912, by the passage of a vote declaring that " ordinary routine examinations of water for sanitary purposes, and in the control of purification plants for the present, should include the determination of the number of bacteria developing at 20 degrees and 37 degrees." By this action the body-temperature count is placed on a par with the 2o-degree count as an integral part of sanitary bac- teriological water examination, a course which has been strongly urged in earlier editions of this book. The body-temperature count must, of course, be made upon agar plates; but otherwise the procedure is much the same as that already described for the routine quantitative bacteriological examination in Chapter II. A 1.5 per cent agar medium has generally been used, but the Standard Methods Committee in its recent report recommends only i per cent of agar. Whipple (1913) points out that this i per cent agar often gives trouble from the running together of the colonies on the weaker medium. On the other hand, a i per cent agar gives higher counts than 1.5 per cent 64 ELEMENTS OF WATER BACTERIOLOGY agar. He emphasizes the recommendation of Jackson that the agar used should be dried at 105° C. for 30 minutes, as commercial agar itself contains more or less water. The period of incubation ordinarily adopted for body- temperature counts is 24 hours. Lederer and Bach- mann (1911) find that with sewage effluents a 48-hour period at 37° may yield counts from two to six times as high as those obtained in 24 hours; it is ques- tionable, however, whether the higher counts thus given would compensate for the loss of time. The adoption of a 24-hour period by the Standard Methods Com- mittee in any case represents an almost universal practice. In using agar plates at 37° difficulty is some- times caused by the spreading of colonies of certain organisms over the surface of the plate in the water of condensation which gathers; this may be avoided by inverting the plates after the agar is once well set, or still better by the use of plates provided with earthen- ware tops, as suggested by Hill. The porous earthen- ware absorbs the water which condenses on it, the surface of the plate remains comparatively dry, and the percentage of " spread " plates is reduced from 30 per cent to i per cent (Hill, 1904). Special pains must be taken, however, to keep the atmosphere in the incubator nearly saturated with moisture or errors will be introduced by the excessive evaporation of the medium used. Use of Litmus Lactose Agar. Additional evidence as to the character of a water sample may be DETERMINATION OF ORGANISMS 65 obtained with little extra trouble by adding a sugar and some sterile litmus to trie agar medium and observing the fermenting powers of the organisms present, as first suggested by Wurtz (Wurtz, 1892) for the separation of B. coli from B. typhi. It hap- pens that the most abundant intestinal organisms, belonging to the groups of the colon bacilli and the streptococci, decompose dextrose and lactose with the formation of a large excess of acid. The decomposi- tion of the latter sugar, on the other hand, is almost entirely wanting among the commoner saprophytic bacteria, and therefore lactose is most commonly used in making sugar agar, i per cent being added to the medium just before the final filtration (between steps 15 and 1 6 in the standard process of media-making given on p. 102). In pouring the plate a cubic centi- meter of sterile litmus solution should be added. After incubation the colonies of the acid-forming organisms will be clearly picked out by the redden- ing of the adjacent agar. Only those colonies which are sharply colored should be considered as significant, since certain bacteria of the hay-bacillus group pro- duce weak acid and faint coloring of the litmus. When polluted waters are examined in this manner the number of organisms developing on the lactose- agar plate will be very high, almost equalling in some cases the total count obtained on gelatin. Chick (Chick, 1901), using a lactose-agar medium with the addition of one-thousandth part of phenol, found, of colon bacilli alone, 6100 per c.c. in the Manchester ship canal; 55-190 in the polluted River Severn, and 66 ELEMENTS OF WATEE BACTERIOLOGY numbers up to 65,000 per gram in roadside mud. In an examination of water from the Charles River above Boston, 37° counts ranging from 9800 to 16,900 have been found. The average result of 56 examinations of Boston sewage from July to December, 1903, showed 5,430,000 bacteria per c.c., at 20° and 3,760,000 per c.c. at 37°, of which 1,670,000 were acid formers. The average of 25 samples examined in July and August, 1904, showed 1,690,000 bacteria per c.c. at 20° and 1,400,000 at 37°; 429,000 per c.c. were acid formers (Winslow, 1905). In unpolluted waters not only the absolute number of organisms developing at the body temperature is less, but its ratio to the gelatin count is very different. Rideal (Rideal, 1902) states that the proportion between the two counts in the case of a London water in a year's examination was on the average one to twelve. Mathews (Mathews, 1893) in 1893 gave the following figures, the contrast between the ponds and streams, which were presumably exposed to pollution, on the one hand, and the wells, springs, and taps, on the other, being marked. Source of Water. Average Number of Colonies per c.c. Gelatin, 20°. Wurtz Agar, 37.5°. Wells springs 1664 153 296 242 273 28 43 95 24 IOI Reservoirs Ponds Taps Streams DETEEMINATION OF ORGANISMS 67 According to the English Committee appointed to consider the Standardization of Methods for the Bac- terioscopic Examination of Water (1904), the ratio of the 20° count to the 37° count in good waters is generally considerably higher than 10 to i. " With a polluted water this ratio is approached, and frequently becomes 10 to 2, 10 to 3 or even less." In 1903 Nibecker and one of ourselves (Winslow and Nibecker, 1903) made an examination of 259 samples of water from presumably unpolluted sources in Eastern Massachusetts, including public supplies, brooks, springs, ponds, driven wells, and pools in the fields and woods, with a view to testing the value of the body-temperature examination. In many cases the samples showed high gelatin counts, since some of the waters were exposed to surface wash from vacant land, but the average number of organisms developing on lactose agar at 37 degrees was less than 8 per c.c., as will be seen by reference to the table on the following page. The highest individual counts obtained were 95 in a meadow pool, 83 in a brook, and 74 in a barnyard well, the latter probably actually polluted. Only two samples in the whole series, one from the well above mentioned, gave any red colonies on the agar plates. For a series of shallow surface wells recently examined by one of us (S. C. P.) a similar relation is indicated in the table on page 69; 124 samples which showed no colon bacilli and were apparently unpolluted, gave an average of 190 bacteria per c.c. at 20° and 8 at 37° with less than one red colony per c.c.; 23 samples which did contain colon bacilli averaged 570 bacteria 68 ELEMENTS OF WATER BACTERIOLOGY RELATION OF 20° AND 37° COUNTS IN SAMPLES OF WATER FROM APPARENTLY UNPOLLUTED SOURCES (WlNSLOW AND NlBECKER, 1903) Source of Samples. Number of Samples. "3 <8 GJ3 Litmus- lactose- agar Plates, 37°. Dextrose-Broth Tubes. Average Number of Colonies. Average Number of Colonies. Plates Showing Red Colonies. . ON ON OO vo - oo 00 Ov 00 Ov vo OO % •uinoc 'H^oig 8SOJ1 9onpoj,j O co ro O co O 1O "* uot^'B5.u9uiJ91>j 'SBQ gonpojfj + M ^J- LO OO vO t>* OO 00 t^» O vo vO O 00 •Bipap^ AaoiBiiuuuoQ o^ui pa^BJnoouj ssan^in3 jo aaquift^ • (N (N H CS VO I CO I Per Cent of Cultures. •snoosjA O M ^" H H CO O I S3 B • ' p4 t^ J S3 ; VAEIETIES OF COLON BACILLI 187 On the whole, much of the English evidence tends to the assumption that the atypical forms, or " paracolon bacilli," generally represent weakened strains from the intestinal B. coli stock. As Savage says, " we know that nearly all the coli-like organisms in faeces are quite typical B. coli, that in sewage a good many atypical varieties are present, and that in contaminated water and soil the proportion present is still larger." The data tabulated on p. 186 from Gage and Phelps (1903) lead to a similar conclusion. About 60 per cent of the cultures isolated from polluted river water, filtered water, and sewage proved to be typical B. coli, while 41 and 43 per cent of those isolated from spring water and shellfish, respectively, and 48 per cent of those from ice belonged in this class. Contradictory results, indicating a higher proportion of typical forms outside the body than within it, have been obtained by Konrich (1910) in the examination of 2387 different strains isolated in about equal propor- tions from faeces, earth, and water. Of the 2387 coli- like microbes studied, 308 were excluded by micro- scopic examination (showing abnormal morphology or positive Gram reaction) or by their liquefaction of gelatin. The other 2079 strains were tested in sugar media and peptone water with the results tabulated on p. 188. We are inclined to attribute these results of Konrich's largely to the technique which he employed. It seems to be clearly stated in his paper that he obtained his faecal cultures by direct plating on solid media, while his earth and water samples were treated to preliminary 188 ELEMENTS OF WATER BACTERIOLOGY BIOCHEMICAL REACTIONS OF 2079 ORGANISMS OF THE COLON GROUP (KONRICH, 1910) Medium. Reaction. Percentage of Positive Results. Faecal Strains. Earth Strains. Water Strains. Dextrose broth Dextrose broth, 46°. . . . Lactose broth Lactose broth. Gas production Gas production Gas production Acid production .... Coagulation IOO 59 77 97 65 54 38 60 IOO 82 87 IOO 84 68 57 77 IOO 82 92 IOO 80 65 57 76 Milk Neutral red dextrose agar Peptone solution Endo agar. Fluorescence Indol production. . . . Deep red colonies with greenish luster enrichment in sugar broths. Such an enrichment would undoubtedly tend to increase the proportion of typical B. coli. Konrich's own experiments on the storage of pure cultures of colon bacilli in water showed a gen- eral, though not invariable, relative increase of atypical forms as the sojourn in water became long continued. Houston (1911) has recently pointed out the danger that the selective action of enrichment media may confuse the normal relations of the bacterial flora; and in order to obtain a more accurate idea of the relations involved he made an elaborate study of raw water, stored water, and stored and filtered water by direct plating, preceded by the use of various physical concentration methods. The results indicated in the table on p. 189 are of interest. VARIETIES OF COLON BACILLI 189 CHARACTERISTICS OF DEXTROSE FERMENTING BAC- TERIA FROM RAW, STORED, AND FILTERED WATERS (HOUSTON, 19 n) Per Cent Positive Results. , « a Gas Production. en O •g n* " Source. w § A c . a 6 i JQ U %* 1 1§ 1 1 "3 a i— i o a |§ Raw river water . 320 76 40 2Z C7 44 I c; 7 O 3 63 T r Stored water. . 232 57 40 31 35 25 6 0.4 66 6 Stored and fil- tered water . 225 69 24 34 Si 25 5 3 40 6 The strains having the combination of positive reac- tions in lactose and peptone solution made up 53 per cent of those isolated from the raw waters. 46 per cent of those from the stored waters, and 34 per cent of those from the stored and filtered waters. MacConkey's Classification of the Colon Group. MacConkey is not satisfied with this general classifica- tion of colon bacilli into typical and atypical forms, but wishes to go much further, believing that even the so-called " typical B. coli " chould rather be con- sidered a complex including a considerable number of definite individual types. After a detailed study of the reactions of the lactose-fermenting bacteria of faeces he outlined a new classification based on fermentative reactions in the rarer sugars (MacConkey, 1905). Using saccharose and dulcite he first divided the lactose- fermenting organisms into four groups as indicated in the table on p. 190. 190 ELEMENTS OF WATER BACTERIOLOGY PRIMARY SUBDIVISIONS OF THE COLON GROUP Saccharose. Dulcite. Type. I. I B. acidi-lactici B. coli (or B. communis) B. neapolitanus (or B. communior) B. aerogenes The fourth group, fermenting saccharose, but not dulcite, was further subdivided by MacConkey into the B. coscoroba type, which does not liquefy gela- tin or give the Voges and Proskauer reaction, the B. lactis-aerogenes type, which does not liquefy gelatin, but does give the Voges and Proskauer reaction, and the B. cloacae type, which liquefies gelatin and gives the Voges and Proskauer reaction. Records of the preva- lence of the four principal groups in human and animal faeces and in milk are given in MacConkey's two papers (1905 and 1909) as well as their relative numbers in a suspension of faeces in water after various intervals of time. The results do not, however, appear to us to justify any important practical conclusions. In his later paper MacConkey (1909) carried the sub-division of the colon group much further. He isolated 497 lactose-fermenting bacilli from the faeces of man and animals, from sewage, water, grains, etc. All were Gram-negative, fermented lactose, coagulated milk and reduced nitrate. They were subdivided by their action on gelatin, pepton and various fer- mentable substances and by their motility into over 100 types of which the more important have received VARIETIES OF COLON BACILLJ 191 names. The principal types of this classification are indicated in the table below: MAcCONKEY'S CLASSIFICATION OF THE COLON GROUP (MACCONKEY, Group. No. Name. Liquefaction of Gelatin. s Indol Produc- tion. Fermentation of J| II Saccharose. Dulcite. Adonit. Inulin. 1 I. i 2 3 4 5 I + ; - - 1 I - \ B. acidi-lactici B. levans B. Griinthal B. vesiculosus II. ! 34 35 B. coli communis B. Schafferi - ; : t T - t i i III. 65 68 7i 72 B. oxytocus pernicio- sus + | I B. pneumonias B. neapolitanus IV. 103 104 107 108 B. lactis aerogenes. . . B. gasoformans + ; i + •r- 1 - it =fc B. coscoroba B. cloacae In this country the MacConkey classification was first adopted by Bergey and Deehan (1908). These workers used 8 diagnostic characters, motility, indol production, liquefaction of gelatin, the Voges-Pros- kauer reaction, and the fermentation of saccharose, dulcite, adonite and inulin. They tabulated 256 different combinations of these 8 characters and in the examina- 192 ELEMENTS OF WATER BACTERIOLOGY tion of 92 colon-like bacilli from 50 samples of milk, 8 of sewage and i of kefir they found 43 of the possible combinations. Copeland and Hoover (1911) have recently urged the importance of these fermentative reactions in the rarer carbohydrates in the study of the colon group. They confirm the positive Voges and Proskauer reaction reported by other observers for B. lactis-aerogenes and B. cloacae and point out that B. lactis-aerogenes is the only form in a considerable series studied which gives a brown coloration in aesculin media in one day. On the other hand they record a positive dulcite reac- tion for B. lactis-aerogenes and B. cloaca3 which is highly confusing and makes it difficult to interpret their results. Both these names according to the usage of MacConkey, which has been accepted for the past five years, are applied to dulcite-negative saccharose- positive organisms. Still another classification of the colon group is Jack- son's modification of MacConkey's scheme in which MacConkey's four primary groups are symmetrically subdivided according to reactions in mannite and rafn- nose with motility, indol production, nitrate reduc- tion, liquefaction of gelatin and coagulation of milk as secondary differential characters (Jackson, 1911). Under each of the four groups, B. communior (MacConkey's B. neapolitanus) , B. communis (MacConkey's B. coli), B. aerogenes (MacConkey's Group IV), and B. acidi- lactici, he distinguishes four types, A (fermenting both mannite and raffinose, B (mannite +, raffinose — ), C (mannite — , raffinose -}-), and D (fermenting neither VARIETIES OF COLON BACILLI 193 mannite nor raffinose); and he indicates reactions in other media by subscript letters. These types with their subtypes are fully discussed in the last report of the Committee on Standard Methods of Water Analysis (1912). Clemesha's Investigation of Stored Waters in India. The most suggestive contribution to this subject which has been made in recent years is a book by Major W. W. Clemesha of the Indian Medical Service on The Bacteriology of Surface Waters in the Tropics (Clemesha, 1912*), in which a vigorous argument is made for the MacConkey classification in practical water work. Major Clemesha's researches show the prevalence of considerable numbers of all of MacConkey's primary types in human and bovine fasces as indicated in the table below, although the relative proportions found in England and in India do not correspond very closely. Clemesha's percentages are of special importance because they are based in each case on over 1000 colonies. RELATIVE PROPORTION OF MACCONKEY'S GROUPS IN HUMAN F/ECES AND IN COW DUNG Human Fasces. Cow Dung. Group. MacConkey. Clemesha. MacConkey. Clemesha. * I 34 53 17 40 2 38 17 25 9 3 15 7 48 16 4 12 22 12 35 Both in human faeces and in cow dung Clemesha finds the prevailing types to be B. coli, B. Grim thai, 194 ELEMENTS OF WATER BACTERIOLOGY and B. coscoroba, the three together usually making up 75 per cent all the lactose-fermenting organisms present. A very interesting point brought out in these investigations was the occurrence of " epidemics," of particular types which at certain periods become suddenly frequent, usually prevailing in human faeces, cow faeces and water supplies at the same time. (It should be noted for the benefit of anyone studying Clemesha's book that the tabular classification of the colon group at the end contains a serious misprint. B. lactis-aerogenes, B. gasoformans, B. coscoroba and B. cloacae are there given as saccharose negative, whereas they should be saccharose positive.) The discussion in the text, however, appears to refer to the orthodox MacConkey types. Clemesha (191 2a) made a number of experiments on the relative resistance of the various lactose-fermenting types by placing faecal emulsions, with or without sand, in shallow dishes in the sunlight and at various intervals isolating 10 colonies of the predominant types and working out their fermentative reactions. In general the experiments showed B. coli to be the dominant form at the beginning. It quickly disappeared, however, and after a few hours B. lactis aerogenes, B. acidi-lactici, B. cloacae and others appeared. At the end of the experiments, often on the second day, B. Grim thai or B. cloacae were generally the only forms surviving. In a long series of examinations of Red Hills Lake Clemesha obtained 138 colonies of lactose-fermenting organisms during rainy periods and of these 59 belonged to MacConkey's Group I, 10 to Group II, 14 to Group III and 55 to VARIETIES OF COLON BACILLI 195 Group IV. Of 280 colonies isolated during dry periods, 37 belonged to Group I, 22 to Group III and 221 to Group IV. When the forces of self-purification had been at work, Group II (B. coli) entirely disappeared and Group IV (B. cloacae and B. coscoroba) was pre- dominant. B. Griinthal was the commonest of the Group I forms. B. cloacae was especially prevalent in bottom samples. A study of a number of rivers in Bengal gave the results tabulated below. RELATIVE PREVALENCE OF CERTAIN LACTOSE- FERMENTING TYPES IN BENGAL RIVERS MacConkey Group. Types. Dry Weather, Dec.-June. Wet Weather, July-Nov. I. II. IV. Do. B. Griinthal and B. vesiculosus B. coii communis B. lactis aerogenes B. cloacae 41 3 7 ii 23 13 19 4 There are many irregularities in Dr. Clemesha's results. For example, B. aerogenes, as well as the other representatives of Group IV, was more abundant in Red Hills Lake during dry periods than at times of rain. On the whole, however, it does seem clear that his results justify a general classification of the lactose- fermenting organisms into three main groups accord- ing to resistance. B. coli communis and B. oxytocus perniciosus (representing MacConkey's Groups II and III, both fermenting dulcite) are sensitive organisms found in numbers only where pollution is fresh. B. lactis- aerogenes, representing the subgroup of MacConkey's 196 ELEMENTS OF WATER BACTERIOLOGY Group IV which ferments adonit, but does not form indol or liquefy gelatin, occupies a somewhat inter- mediate position, appearing in waters which have been fairly recently polluted and later disappearing again. Finally B. Grim thai and B. vesiculosus (MacConkey's Group I, negative in both saccharose and dulcite) and B. cloacae and B. coscoroba (of MacConkey's Group IV, dulcite negative and saccharose positive), are highly resistant organisms which occur in relatively high proportions in stored waters. B. cloacae is most abundant in bottom sediments and B. Griinthal and B. vesiculosus in sunned surface-waters. The moral drawn by Major Clemesha is that for Indian conditions with waters stored in warm sunned lakes and large rivers, where sensitive faecal bacteria have ample opportunity to die out and resistant faecal bacteria have an ample opportunity to multiply, it is not proper to condemn water containing any members of the colon group without distinguishing between the more and the less resistant forms. For example, he quotes 239 examinations of which only 74 showed no B. coli according to the English standard, which closely corresponds to our own, while 165 showed what we should call positive results. Of the 165, however, 69 contained only the highly resistant B. Griinthal and 59 contained mixtures of other forms not belonging to MacConkey's Group II (saccharose negative, dulcite positive). Thus of the 239 samples 31 per cent would have been passed by Houston's standard, 53 per cent would have been condemned by Houston's standard, although containing only resistant forms which Clemesha VAEIETIES OF COLON BACILLI 197 believes to be unimportant, and 16 per cent would be condemned by Clemesha as containing true B. communis. Major Clemesha does not claim that these results necessarily indicate any change of procedure in dealing with the waters of temperate climates. Indeed, the experience of English and American bacteriologists offers pretty conclusive evidence that waters so stored as to be safe do not contain large numbers of lactose- fermenting organisms of any type. In other tropical countries and perhaps in warm summer weather, the Indian conditions may possibly be duplicated (as we know they are in the case of the forms fermenting dextrose but not lactose) ; and the experiments reported in this book deserve the careful consideration of water bacteriologists and sanitarians. The results obtained by Houston (1911) in London unfortunately do not correspond at all with these Indian data. Houston studied in detail the reactions of about 800 strains of dextrose-fermenting bacteria from raw river-water, stored water, and stored and filtered water. Comparison of the relative prevalence of types from these three sources ought to furnish some confirmation of Major Clemesha's conclusions, even although the extreme conditions of warmth and sun- light are lacking. We find, however, on careful study of the figures that they do not. The Houston types corresponding to B. communis, B. Schafferi and B. neapolitanus (sensitive forms) are on the whole but little more prevalent in the raw than in the stored and filtered waters. On the other hand the types corresponding to B. Griinthal, B. vesiculosus, B. cos- 198 ELEMENTS OF WATER BACTERIOLOGY coroba and B. cloacae (Clemesha's resistant types) are less abundant in the filtered and stored than in the raw water. Houston's lactose-fermenting forms clas- sified in MacConkey's four great groups show the rela- tions indicated in the table below, which are almost the reverse of what should be expected if the dulcite- fermenting forms (Groups II and III) were indicative of recent pollution. DISTRIBUTION OF MACCONKEY'S GROUPS IN RAW, STORED, AND FILTERED WATER AT LONDON Pei centage in Group. Reactions. Type. Raw Stored Filtered Water. Water. Water. I Saccharose — Dulcite — B. acidi-lactici . . . 34 39 37 II Saccharose— Dulcite + B. coli 23 25 38 III Saccharose+Dulcite+ B. neapolitanus. . 15 26 9 IV Saccharose + Dulcite — B. lactis-aerogenes 28 IO 16 Statistical Classification of the Colon Group. From a biological standpoint, there is a twofold difficulty with such a classification as that of MacConkey and Jackson. In the first place it is enormously complex, or soon becomes so, as new investigators add new diagnostic tests. In the second place, it is entirely arbitrary in its choice of the order in which particular tests are to be used in splitting up the group. Closely related forms may be widely separated if they chance to differ in the one respect first chosen for dichotomic division. The best basis for a classification following natural bio- logical lines seems to us to be the statistical method first VARIETIES OF COLON BACILLI 199 suggested by Andrewes and Horder (1906) and Winslow and Winslow (1908) in the study of the cocci. The essential point about this method is that the characters of the organisms studied are not considered indepen- dently, but in relation to each other. The individual reactions are first studied quantitatively in a considerable series of allied strains, so that those types of reaction which are manifested by a large number of strains may be distinguished from the rarer intermediate variations. In the second place, the correlations be- tween different characters are used as a basis for group- ing the types on the assumption that a coincidence in several characters indicates a closer relationship than any single character alone. The statistical method has been applied to the colon group in two extensive investigations, neither of which has yet been published in full. Of the first by Howe, a brief, abstract has appeared (Howe, 1912). The second by L. A. Rogers, W. M. Clark, and B. J. Davis we have had the opportunity of seeing in manuscript. These two papers promise at last to lay a foundation for a sound knowledge of the relationships of the colon group. Howe (1912) in his investigation dealt with 630 strains of fresh intestinal colon bacilli. He concluded from his exhaustive study that in bacilli of this type isolated directly from stools, the characters of motility, indol formation, ammonia production, nitrate reduction, fermentation of mannite, dulcite, and starch were not sufficiently correlated with each other or with other characters to be of classificatory value. Dextrose, 200 ELEMENTS OF WATER BACTERIOLOGY lactose, saccharose and raffinose he found to constitute a natural metabolic gradient, in the order named, fermentation of any member of the series implying fermentation of those preceding it. Fifty-three per cent of his strains fermented all four sugars, 5 per cent all but rafnnose, 41 per cent attacked dextrose and lactose only, and i per cent dextrose alone. CHAPTER IX OTHER INTESTINAL BACTERIA IT would be an obvious advantage if the evidence of sewage contamination, furnished by the presence of the colon group, could be reinforced and confirmed by the discovery in water of other forms equally char- acteristic of the intestinal canal. The attention of a few bacteriologists in England and America has been turned in this direction during the past few years; and two groups of organisms, the sewage streptococci and the anaerobic spore-bearing bacilli, have been described as probably significant. Significance of the Sewage Streptococci. The term " sewage streptococci," as generally used, covers an ill-defined group, including many cocci which do not occur in well-marked chains. Those most commonly found grow feebly on the surface of ordinary nutrient agar, producing faint transparent, rounded colonies, but under semi-anaerobic conditions flourish better, giving a well-marked growth along the gelatin stab and only a small circumscribed film on the surface. They are favored by the presence of the sugars and ferment dextrose and lactose, with the formation of abundant acid but no gas. They are seen under the microscope as cocci, occurring as a rule in pairs, 201 202 ELEMENTS OF WATER BACTERIOLOGY short chains, or irregular groups. They do not show visible growth and do not form indol and nitrite in the standard peptone and nitrate solutions; most of them do not liquefy gelatin, though occasionally forms are found which possess this power. Until recently no systematic study of the various species found in the intestine had been made and all cocci giving the char- acteristic growth on agar and strongly fermenting lactose are commonly included as " sewage streptococci." Although the significance of the streptococci as sewage organisms is not established with the same defmiteness which marks our knowledge of the colon group, these forms have been isolated so frequently from polluted sources and so rarely from normal ones that it now seems reasonable to regard their presence as indicative of pollution. Although originally reported by Laws and Andrewes (Laws and Andrewes, 1894), their importance was not emphasized until 1899 and 1900, when Hous- ton (Houston, i899b, i9Oob) laid special stress upon the fact that streptococci and staphylococci seem to be characteristic of sewage and animal waste, the former being, in his opinion, the more truly indicative of dangerous pollution, since they are " readily demon- strable in waters recently polluted and seemingly altogether absent from waters above suspicion of con- tamination." In six rivers recently extensively sewage- polluted, he found streptococci in from one-tenth to one ten-thousandth of a c.c. of the water examined, although in some cases the chemical analysis would not have indicated dangerous pollution. On the other hand, eight rivers, not extensively polluted, showed OTHER INTESTINAL BACTERIA 203 no streptococci in one- tenth of a c.c., although the chemical and the ordinary bacteriological tests gave results which would condemn the waters. Horrocks (Horrocks, 1901) found these organisms in great abun- dance in sewage and in waters which were known to be sewage-polluted, but which contained no traces of Bacillus coli. He found by experiment that B. coli gradually disappeared from specimens of sewage kept in the dark at the temperature of an outside veranda, while the commonest forms which persisted were varieties of streptococci and staphylococci. In America attention was first called to these organisms by Hunnewell and one of us (Winslow and Hunnewell, 1902*), and the same authors later (Winslow and Hunne- well, i902b) recorded the isolation of streptococci from 25 out of 50 samples of polluted waters. Gage (Gage, 1902), from the Lawrence Experiment Station, has reported the organisms present in the sewage of that city, while Prescott (i9O2b) has shown that they are abundant in faecal matter and often overgrow B. coli in a few hours when inoculations are made from such material into dextrose broth. In the monograph of Le Gros (Le Gros, 1902) of the many streptococci described, all without exception were isolated, either from the body or from sewage. Baker and one of us (Prescott and Baker, 1904), found these organisms present in each of 50 samples of polluted waters. On the other hand, in the study of 259 samples of presuma- bly unpolluted waters, by the method of direct plating, Nibecker and of the authors (Winslow and Nibecker, 1903) found streptococci in only one sample. Clemesha 204 ELEMENTS OF WATER BACTERIOLOGY (191 2a) finds that streptococci in India are present in .0001 or .00001 gm. of faeces, but are rare in waters unless very grossly polluted. In a series of bottle experiments in the laboratory and in the study of an artificially polluted tank outdoors he showed that they disappear very rapidly in water, within 2 or 3 days at the most. Gordon (1904) showed that certain strep- tococci are abundant in normal saliva and are found in air which has been exposed to human pollution but not in normal air. On the whole there can be no doubt of the fact that streptococci occur on the surfaces of the human and animal body more commonly than anywhere else in nature. Isolation of Sewage Streptococci. The isolation of these organisms either from plates or liquid cultures is easy. On the lactose-agar plate, made directly from a polluted water, the colonies of the streptococci may generally be distinguished from those of other acid- formers by their small size, compact structure, and deep-red color, which is permanent, never changing to blue at a later period of incubation. Developing somewhat slowly, however, they may be overlooked if present only in small numbers. In the dextrose- broth tube, streptococci will generally appear in abun- dance after a suitable period of incubation. Prescott and Baker, in the work above mentioned, found that with mixtures of B. coli and streptococci in which the initial ratios of the latter to the former varied from i : 94 to 208 : i, the colon bacilli developed rapidly during the early part of the experiment, reaching a maximum after about 14 hours, and then diminishing OTHER INTESTINAL BACTERIA 205 rapidly. The streptococci first became apparent after 10 to 15 hours and reached their maximum after 20 to 60 hours, according to the number originally present. Applying the same method to polluted waters, similar periodic changes were observed; nearly pure cultures of B. coli were first obtained, then the gradual displace- ment of one form by the other took place, and at length RELATIVE GROWTH OF B. COLI AND SEWAGE STREPTO- COCCI FROM POLLUTED WATERS IN DEXTROSE BROTH (PRESCOTT AND BAKER, 1904) Sample Number I 2 3 4 5 6 7 8 9 10 Red colonies developing 1 from i c.c. of original sam- } pie on litmus lactose agar J 4 10 9 5 8 55 35 460 1250 105 ii B. coli o 20 68 200 185 400 130 332 420 410 hrs. Strept 0 O O o o o 0 0 0 o 16 B. coli 2OO 76 I3O 27O 220 2IO 140! 420 285 410 Number found, in millions per hrs. Strept 40 25 20 10: 45 30 20 210 75 145 cubic centime- ter, after, 23 B. coli 280 150 385 370 300 570 2OO 405 320 300 growth in dex- trose broth for hrs. Strept 140 85 280 170 300 1700 no 350 370 350 various peri- ods 30 B. coli O 0 25 no O 2IO 20 24 105 hrs. Strept 474 420 480 300 390 170 400 105 250J 63 B. coli 0 o O 0 0 12 8 0 0 o hrs. Strept 2 0 0 45 I 2 45 ISO 86 170 First gas noted after (hrs.).. . 10 10 0 0 10 8, 10 6 6 8 1 the streptococci were present either in pure culture or in great predominance as shown by the accompany- ing tables. The samples of water were plated directly upon litmus lactose agar and the plates were incubated at 37° for 24 hours, when the red colonies were counted. At the time of plating, i c.c. from each sample was also inoculated into dextrose broth in fermentation tubes, 206 ELEMENTS OF WATER BACTERIOLOGY which were likewise incubated at 37°. After various periods, as indicated by the tables below, the tubes were shaken thoroughly and i c.c. of the contents withdrawn. This was diluted (generally 1-1,000,000,) with sterile water, plated on litmus lactose agar in the usual way, and incubated for 24 hours. The colonies of B. coli and streptococci were distinguished micro- RELATIVE GROWTH OF B. COLI AND SEWAGE STREPTO- COCCI FROM POLLUTED WATERS IN DEXTROSE BROTH (PRESCOTT AND BAKER, 1904) Sample Number 18 M) 20 21 22 23 24 25 Red colonies developing from i c.c. 1 of original sample on litmus lac- j- tose agar J i 150 25 30 50 .04 O 170 . 12 O 380 128 200 • 55 o 330 80 IOO 30 1.6 0 f Number found, in mil- lions per cubic centi- meter, after growth^ in dextrose broth for various periods. . . . 7 hrs. B. coli .02 — .OI Strept. O O i? hrs. B. coli 266 TOO 88 350 Sio 160 220 300 Strept. 150 • o! 40 140 240 27 hrs. B. coli 520 610 72 700 IOOO 740 4380 7 60 35 Strept. 800 860 670 IO80 22 22 20 31 2500 36 66 70 7 52 3900 40 hrs. B. coli O o IO Strept. 252 330 260 16 38 52 hrs. B. coli IO 10 27 Strept. 40 16 3-8 4i 25 IO 30 scopically, and by difference in color and general characters. The successive growth of these two intestinal groups in the same dextrose-broth tube suggests the following method for the detection of both B. coli and sewage streptococci. Inoculate the desired quantity of water, preferably OTHER INTESTINAL BACTERIA 207 i c.c., into dextrose broth, in a fermentation tube, and incubate at 37°. After a few hours' incubation examine the cultures for gas. Within 2 or 3 hours' after gas formation, is first evident, plate from the broth in litmus lactose agar, incubating for 12 to 18 hours at 37°. If at the end of this time no acid-produc- ing colonies are found, it is probably safe to assume that there were no colon bacilli present. On the other hand, if red colonies are developed, these must be fur- ther examined by the regular diagnostic tests for B. coli. After the first plating from the dextrose broth, replace the fermentation tube in the incubator and allow it to remain for 24 to 36 hours, then plate again on litmus lactose agar. This plating should give a nearly pure culture of streptococci if these organisms were originally present in the water. Streptococci as Indicators of Recent Pollution. The comparative relation of the streptococci and the colon bacilli to sewage pollution is still somewhat uncertain. Houston (Houston, 1900) held that the former microbes imply " animal pollution of extremely recent and there- fore specially dangerous kind," and Clemesha's experi- ments led to the same conclusion. Horrocks (Horrocks, 1901), on the other hand, maintains, largely on the strength of certain experiments with stored sewage, that the streptococci persist after colon bacilli have disappeared and indicate contamination with old sewage which is not necessarily dangerous. These discordant results are probably to be explained by the different media in which the viability of the bacteria was com- pared. It seems likely that in sewage where there is a 208 ELEMENTS OF WATER BACTERIOLOGY large amount of organic food material present the streptococci may kill out the colon bacilli as they do in the fermentation tube, and as we know they fre- quently do in shellfish. This would explain Horrocks' results. On the other hand, there is good evidence that the streptococci are less resistant than B. coli to the unfavorable conditions which exist in water of ordinary organic purity. In waters of potable char- acter B. coli is frequently present without the strep- tococcus; and a negative test for streptococci has little significance. A positive test, on the other hand, furnishes valuable confirmatory evidence of pollution. This evidence is of course of special importance when through the activity of the streptococci themselves, or from any other cause the colon isolation has yielded an erroneous negative result. The English Committee appointed to consider the standardization of methods for the bacterioscopic examination of water (1904) by a majority vote rec- ommended the enumeration of streptococci, as a routine procedure in sanitary water analysis, but in this country the Committee on Standard Methods of Water Analysis (1912) has concluded that "the information afforded by the occurrence of these organisms seems to be of less value than in the case of B. coli and it is believed that for the present at least, the streptococcus test is of subordinate importance." Use of the Streptococci to Distinguish between Human and Animal Pollution. There seems some reason to hope that the streptococci may prove of assistance in the important task of differentiating human and animal OTHER INTESTINAL BACTERIA 209 pollution, a task in which all other tests have so far failed. Unlike the colon bacilli, streptococci from the intestines of cattle and men appear to belong to dis- tinct types. The recognition of- this fact we owe primarily to Gordon (1905), who made an elaborate study of the fermentative power of the streptococci in a long series of carbohydrate media. His work and that of Houston (Houston, 1904; Houston, 1905*, Houston, i905b) have made it clear that the streptococci of the herbivora differ from those found in the human body in their low fermentative power. In their review of the genus, Andrewes and Horder (1906) describe the type characteristic of the herbivora under the name, Str. equinus, and define it by its failure to ferment lactose, raffinose, inulin or mannite, or to reduce neutral red. Five other types are described from the human mouth and intestine; all of them ferment lactose, and most reduce neutral red and fer- ment raffinose. The commonest intestinal form clots milk, reduces neutral red and ferments saccharose, salicin, coniferin and mannite. The specific types of the genus Streptococcus, grade into each other by almost imperceptible degrees, and streptococci ferment- ing lactose and raffinose and reducing neutral red are sometimes found in bovine faeces; but the studies made in this country by Winslow and Palmer (1910) confirm the conclusions of the English observers that there are specific differences between the streptococci of the human, bovine, and equine intestines. The most im- portant of these results are indicated in the table below: 210 ELEMENTS OF WATER BACTERIOLOGY COMPARATIVE FERMENTATIVE POWER OF STREPTO- COCCI FROM THE HORSE, THE COW, AND MAN (WINSLOW AND PALMER, 1910) Streptococci. Percentage of Positive Results (300 Strains). Lactose. Raffinose. Mannite. Human .... 62 8 52 6 4 28 28 2 6 Equine Bovine The rarity of lactose-fermenting streptococci in the horse makes it probable that this group can be used for distinguishing pollution by street washings from that due to domestic sewage; and the fact that a considera- bly larger proportion of human strains attack mannite and a considerably larger proportion of bovine strains ferment raffinose should make it possible to use the ratio between results in these two media to distinguish between the wash from pastures and cultivated land and sewage. Clemesha (i9i2a) in India has, however, obtained very different results. Out of 115 strains of streptococci from human faeces 92 per cent belonged to the "lamirasacsal" class of Houston (acid in lactose, clot in milk, acid in raffinose, saccharose and salicin), and none acidified mannite. Of 39 strains from cow dung all belonged either to this same " lamirasacsal " class or to the " larasacsal " class (differing only in failing to clot milk). Nevertheless, in view of the importance of distinguishing between human and animal pollution and the hopelessness of doing so by means of the colon group these different types of streptococci well deserve further study. OTHER INTESTINAL BACTERIA 211 The Anaerobic Spore-forming Bacilli. The English bacteriologists have ascribed much importance as indicators of sewage pollution to another group of organ- isms, the anaerobic spore-forming bacilli, of which the form described as B. aerogenes capsulatus (Welch and Nuttall, 1892) and now called B. welchii, and the form isolated by Klein (Klein, 1898; Klein, 1899) in 1895 in the course of an epidemic of diarrhoea at St. Bartholomew's Hospital, described under the name of B. enteritidis sporogenes (now called B. sporogenes) are types. The procedure originally described by Klein for isolating B. sporogenes is as follows: a portion of the sample to be examined is added to a tube of sterile milk, which is then heated to 80° C. for 10 minutes to destroy vegetative cells. The milk is next cooled and incubated under anaerobic conditions, which may be accomplished most conveniently by Wright's method. A tight plug of cotton is forced a quarter way down the test-tube, the space above is loosely filled with pyrogallic acid, a few drops of a strong solution of caustic potash are added, and the tube is tightly closed with a rubber stopper. After 18 to 36 hours at 37° the appearance of the tube will be characteristic if the B. sporogenes is present. " The cream is torn or altogether dissociated by the development of gas, so that the surface of the medium is covered with stringy, pinkish-white masses of coagulated casein, enclosing a number of gas-bubbles. The main portion of the tube formerly occupied by the milk now contains a colorless, thin, watery whey, with a few casein lumps adhering here and there to the sides 212 ELEMENTS OF WATER BACTERIOLOGY of the tube. When the tube is opened, the whey has a smell of butyric acid and is acid in reaction. Under the microscope the whey is found to contain numerous rods, some motile, others motionless." Since this organism is not present in very large num- bers, even in sewage, the test of a water-supply must be made with large samples, and the concentration of at least 2000 c.c. of water by nitration through a Pasteur filter is recommended by Horrocks as a necessary prelude (Horrocks, 1901). The Committee on Standard Methods of Water Analysis (1912) recommends the following enrichment procedure for the isolation of B. sporogenes which avoids physical concentration. Vari- ous dilutions of the water to be tested are incubated in fermentation tubes containing liver broth for 24 hours at 37°. If B. sporogenes is present gas will be evolved and a characteristic " vile odor " will be produced. If this reaction is obtained the contents of each posi- tive tube is transferred to an Erlenmeyer flask or large test-tube and heated at 80° C. for 10 minutes to destroy vegetative cells. One c.c. of broth containing sediment is withdrawn from the bottom of each flask and enriched once more in a fresh liver broth tube. B. sporogenes will now usually be present in pure culture showing large sluggishly motile bacilli containing spores. A gelatin stab culture made from these 24-hour liver broth tubes will show after 48 hours incubation at 20° a dis- tinct liquefying anaerobic growth beginning about 2 cm. below the surface with gas bubbles at the top of the liquefied area. In order to obtain absolutely pure cultures it is necessary to fish from liver broth tubes OTHER INTESTINAL BACTERIA 213 only 3-5 hours old as only young vegetative cells will grow on plates. Transplants from the closed arm of such tubes will grow on dextrose liver agar plates incu- bated under anaerobic conditions. The organisms of the B. sporogenes group are large stout bacilli often occurring in chains. They liquefy gelatin vigorously and on agar produce fine discrete gray colonies. They vigorously ferment dextrose, lac- tose and saccharose, producing acid and gas, and in sugar agar each colony will be marked by one or more gas bubbles surrounded by a delicate whitish fringe. The organism is strongly pathogenic for guinea pigs, by which character it is distinguished from the B. butyricus of Botkin. B. welchii differs from B. sporogenes chiefly in lacking motility and in forming spores with less read- iness (Klotz and Holman, 1911). The researches of Klein and Houston (Klein and Houston, 1898, 1899) have shown that the B. sporogenes occurs in English sewage in numbers varying from 30 to 2 200 per c.c. and that it is often absent in considerable volumes of pure water. In Boston sewage it may usually be isolated from .01 or .001 of a c.c. (Winslow and Belcher, 1904). Since the spores of an anaerobic bacillus may persist for an indefinite period in polluted waters, their presence need not necessarily indicate recent or dangerous pollution. Vincent (1907) and other French observers consider the determination of the total number of anaerobic bacteria as significant, since the decomposition of organic matter is accompanied by anaerobic growth. It is not claimed, however, that bacteria of this type are characteristic of 214 ELEMENTS OF WATER BACTERIOLOGY animal more than of vegetable decompositions, and the total anaerobic count apparently adds nothing of impor- tance to the information gained by the ordinary gelatin plate method. The property of liquefaction was for- merly believed to be of significance, inasmuch as the liquefying bacteria were regarded as indicative of pollu- tion. This position is, however, no longer tenable, since many bacteria, typical of the purest waters, may cause liquefaction. As Savage says in summing up this question: " The number of different species of organisms in sewage is very great, and it is highly probable that many of them occur in all specimens of ordinary sewage; but, except for B. coli, streptococci, and B. enteritidis sporo- genes, their presence has not been ascertained with sufficient constancy, nor has their numerical occurrence been sufficiently investigated to make them of value as indicators of sewage pollution." (Savage, 1906.) CHAPTER X THE SIGNIFICANCE AND APPLICABILITY OF THE BACTERIOLOGICAL EXAMINATION Sanitary Inspection and Sanitary Analysis. The first attempt of the expert called in to pronounce upon the character of a potable water should be to make a thorough sanitary inspection of the pond, stream, well or spring from which it is derived. Study of the pos- sible sources of pollution on a watershed, of the direc- tion and velocity of currents above and below ground, of the character of soil and the liability to contamina- tion by surface-wash are of supreme importance in interpreting the analyses to be made. In many cases, however, the results of the sanitary inspection will be found to be by no means conclusive. If house or barnyard drainage or sewage is actually seen to enter a water used for drinking purposes it is obviously unnecessary to carry out delicate chemical or bacteri- ological tests to detect pollution. On the other hand, no reconnoissance can show certainly whether unpurified drainage from a cesspool does or does not reach a given well; whether sewage discharged into a lake does or does not find its way to a neighboring intake; whether pollution of a stream has or has not been removed by a certain period of flow. Evidence upon 215 216 ELEMENTS OF WATER BACTERIOLOGY these points must be obtained from a careful study of the characteristics of the water in question, and this study can be carried out along two lines, chemical and bacteriological. Sanitary Chemical Analysis. A chemical examina- tion of water for sanitary purposes is mainly useful in throwing light upon one point — the amount of decom- posing organic matter present. It also gives an his- torical picture which may be of much value. Humus- like substances may be abundant in surface-waters quite free from harmful pollution, but these are stable compounds. Easily decomposable bodies, on the other hand, must obviously have been recently introduced into the water and mark a transitional state. " The state of change is the state of danger," as Dr. T. M. Drown once phrased it. Sometimes the organic mat- ter has been washed in by rain from the surface of the ground, sometimes it has been introduced in the more concentrated form of sewage. In any case, it is a warn- ing of possible pollution, and the determination of free ammonia, nitrites, carbonaceous matter, as shown by " oxygen consumed," and dissolved oxygen yield important evidence as to the sanitary quality of a water. Furthermore, nitrates, the final products of the oxida- tion of organic matter, and the chlorine introduced as common salt into all water which has been in contact with the wastes of human life, furnish additional infor- mation as to the antecendents of a sample. The results of the chlorine determination are indeed perhaps more clear than those of any other part of the analysis, for chlorine and sewage pollution vary together, due allow- BACTERIOLOGICAL EXAMINATION 217 ance being made for the proximity of the sea and other geological and meteorological factors. Unfortunately, it is only past history and not present conditions which these latter tests reveal, for in a ground-water completely purified from a sanitary standpoint such soluble con- stituents remain, of course, unchanged. Thus, in the last resort, it is upon the presence and amount of decom- posing organic matter in the water that the opinion of the chemist must be based. Information Furnished by Bacteriological Examina- tions. The decomposition of organic matter may be measured either by the material decomposed or by the number of organisms engaged in carrying out the proc- ess of decomposition. The latter method has the advan- tage of far greater delicacy, since the bacteria respond by enormous multiplication to very slight increases in their food-supply, and thus it comes about that the standard gelatin-plate count at 20° roughly corresponds, in not too heavily polluted waters, to the free ammonia and " oxygen consumed," as revealed by chemical analysis. If low numbers of bacteria are found, the evidence is highly reassuring, for it is seldom that water could be contaminated under natural conditions without the direct addition of foreign bacteria or of organic matter which would condition a rapid multiplication of those already present. The bacteriologist in such cases can declare the innocence of the water with justifiable certainty. When high numbers are found the interpreta- tion is less simple, since they may exceptionally be due to the multiplication of certain peculiar water forms. Large counts, however, under ordinary conditions, 218 ELEMENTS OF WATER BACTERIOLOGY when including a normal variety of forms indicate the presence of an excess of organic matter, derived in all probability either from sewage or from the fresh wash- ings of the surface of the ground. In either case danger is indicated. A still closer measure of polluting material may be obtained from the numbers of colonies which develop on litmus-lactose-agar at 37°, since organisms which thrive at the body temperature, and particularly those which ferment lactose, are characteristic of the intestinal tract and occur but rarely in normal waters. Gage (Gage, 1907) has shown that by counts at 20, 30, 40, and 50° C., information may be quickly obtained which is of great assistance in judging the character of the water. " Modern methods of bacterial examination of water, consisting usually of determinations of the numbers of bacteria by means of plates incubated at room tempera- ture, and of tests for the presence or absence of one or two specific types, occasionally lead to an erroneous interpretation of the quality of a water, owing to the fact that they do not yield adequate data by which abnormal and inaccurate results may be separated from those which are truly indicative of purity or pollution. Furthermore, as several days must elapse before the bacterial tests can be completed, the results when obtained may have passed their usefulness. If, however, we can so modify our procedure that the varied char- acter of the bacteria in waters of different classes may be quickly and accurately recognized, the value of bacterial water analysis will be enormously increased. BACTERIOLOGICAL EXAMINATION 219 Much of this information may be obtained by the use of selective media, selective temperatures, or by a proper combination of the two. " By the use of litmus-lactose-agar in place of agar or gelatin we obtain similar counts of total bacteria, and in addition are able to separate those bacteria into two groups, which do and do not produce acid fermenta- tion of lactose, and the numbers of the two classes of bacteria so obtained indicate more completely the character of the water than would the numbers of either class alone. By incubating our plates at temperatures of 30 or 40° C. we are able to obtain counts in 12 to 18 hours, which counts, while smaller than those on plates incubated for a longer period at a lower temperature, appear to be fully as significant. If we increase our number of determinations by incubating duplicate plates at two or more temperatures, the various results and the ratios between them furnish a check upon one another in addition to increasing the available data upon which to base an interpretation." (Gage, 1907.) Finally, the search for the Bacillus coli furnishes the most satisfactory of all single tests for f a3cal contamina- tion. This organism is preeminently a denizen of the alimentary canal and may be isolated with ease from waters to which even a small proportion of sewage has been added. On the other hand, it is never found in abundance in waters of good sanitary quality, and its numbers form an excellent index of the value of waters of an intermediate grade. The streptococci appear to be forms of a similar significance useful as yielding a certain amount of confirmatory evidence. 220 ELEMENTS OF WATER BACTERIOLOGY The full bacteriological analysis should then consist of three parts, the gelatin-plate count, as an estimate of the amount of organic decomposition in process; the total count, and the count of red colonies, on litmus-lactose-agar, as a measure of the organisms which form acids and thrive at the body temperature; and the study of a series of lactose bile tubes for the isolation of colon bacilli. Special Advantage of the Bacteriological Examination. The results of the bacteriological examination have, in several respects, a peculiar and unique significance. First, this examination is the most direct method of sanitary water analysis. The occurrence of nitrites or free ammonia in a small fraction of one part per million, or of chlorine in several parts per million, do not in themselves render a water objectionable or dangerous. They merely serve as indicators to show that germ-containing and germ-sustaining organic mat- ter is present. By a determination of the chlorine and study of the relations of carbon and nitrogen, it is possible to determine with some degree of accuracy whether this organic matter is of plant or animal origin, and hence to rate its objectionable or dangerous char- acter. By the bacteriological examination, on the other hand, we are able to determine directly whether particular kinds of organisms characteristic of sewage are, or are not, actually present in the water. What we dread in drinking-water is the presence of pathogenic bacteria, mainly from the intestinal tract of man, and it is quite certain that the related non-pathogenic bacteria from the same source will behave more nearly BACTERIOLOGICAL EXAMINATION 221 as these disease germs do than will any chemical com- pounds. In the second place, the bacteriological methods are superior in delicacy to any others. Klein and Houston (1898) showed by experiment with dilu- tions of sewage that the colon test was from ten to one hundred times as sensitive as the methods of chemical analysis; and studies of the self-purification of streams have confirmed their results on a practical scale. Thus in the Sudbury River it was found that while chem- ical evidences of pollution persisted for 6 miles beyond the point of entrance, the bacteria introduced could be detected for 4 miles further (Woodman, Winslow, and Hansen, 1902). The statement is sometimes made that while bac- teriological methods may be more delicate for the detection of pollution in surface-waters, contamination in ground-waters may best be discovered by the chemical analysis. That such is not the case has been well shown by Whipple (Whipple, 1903) who cites the fol- lowing two instances in which the presumptive test revealed contamination not shown by the chemical analysis : " A certain driven- well station was located in swampy land along the shores of a stream, and the tops of the wells were so placed that they were occasionally flooded at times of high water. The water in the stream was objectionable from the sanitary standpoint. The wells, themselves were more than 100 feet deep; they pene- trated a clay bed and yielded what may be termed arte- sian water. Tests for the presence of Bacillus coli had invariably given negative results, as might be naturally 222 ELEMENTS OF WATER BACTERIOLOGY expected. Suddenly, however, the tests became positive and so continued for several days. On investigation it was found that some of the wells had been taken up to be cleaned, and that the workmen in resinking them had used the water of the brook for washing them down. This allowed some of the brook- water to enter the system. It was also found that at the same time the water in the brook had been high, and because of the lack of packing in certain joints at the top of the wells the brook-water leaked into the suction main. The remedy was obvious and was immediately applied, after which the tests for Bacillus coli once more became negative. During all this time the chemical analysis of the water was not sufficiently abnormal to attract attention. On another occasion a water-supply taken from a small pond fed by springs, and which was practically a large open well, began to give positive tests for Bacillus coli, and on examination it was found that a gate which kept out the water of a brook which had been formerly connected with the pond was open at the bottom, although it was supposed to have been shut, thus admitting a contam- inated surface-water to the supply." Whipple also calls attention to the report on the Chemical and Bacteriological Examination of Chichester Well-waters by Houston (Houston, 1901), in which the results of chemical and bacteriological examinations of thirty wells were compared. It was found that the bacteri- ological results were in general concordant and satis- factory. The wells which were highest in the number of bacteria showed also the greatest amount of pollu- tion, as indicated by the numbers of B. coli, B. sporo- BACTERIOLOGICAL EXAMINATION 223 genes, and streptococci. On the other hand, the chlorine and the albuminoid ammonia showed no correspondence with the bacteriological results. Vincent (Vincent, 1905) cites an interesting case of the detection of progressive pollution of a ground- water by bacteriological methods. The well of a military camp in Algeria showed 200 bacteria per c.c. before the arrival of a regiment of troops. Its sub- sequent history is indicated in the table below: PROGRESSIVE POLLUTION OF A WELL (VINCENT, 1905) Bacteria per c.c. Bacillus coli per c.c. Before arrival of troops 2OO O 6 days after arrival 77O o 14 days after arrival 41 days after arrival 4,240 6 060 I 2 60 days after arrival 14,900 IO Thirdly, negative tests for Bacillus coli and low bac- terial counts may be interpreted as proofs of the good quality of water, with a certainty not attainable by any other method of analysis. Many a surface-water with reasonably low chlorine and ammonias has caused epi- demics of typhoid fever; but it is impossible, under any natural conditions (except perhaps in a well polluted with urine) that a water could contain the typhoid bacillus without giving clear evidence of pollution in the bile tube or on the lactose-agar plate. In the examination of springs, especially those used for domestic supplies at country houses, the authors have found that the bacteriological examination offers a 224 ELEMENTS OF WATER BACTERIOLOGY more delicate and more certain index of the quality than may be obtained by chemical analysis. In a number of instances, springs located in pastures have become slightly polluted by animals, but to so small an extent that the chemical examination gave no indi- cation of trouble. The bacteria, however, increased greatly in number, and colon bacilli could be readily isolated from 75 per cent of the i-c.c. samples of a long series used in making the presumptive test. A single case may suffice as an illustration. This was a spring located on a hill in Hopkinton, Mass. The chemical analysis was as follows: Color None Turbidity None Sediment None Odor (hot) None Odor (cold) None Parts per Million. Total solids 33 . oooo Loss on ignition 7 . oooo Fixed residue 26 . oooo Hardness 1 1 . oooo Chlorine 10 . oooo Nitrogen as — Albuminoid ammonia o . oooo Free ammonia o . oooo Nitrites o . oooo Nitrates o . oooo The bacteriological examination showed a total count of 375 bacteria per c.c. and a 37° count of 350 per c.c. The presumptive tests for Bacillus coli showed that gas-producing organisms were present in a majority of i -c.c. samples, and typical colon bacilli were isolated. In this case the contamination was brought about by cattle gaining access to the area immediately surround- BACTERIOLOGICAL EXAMINATION 225 ing the spring; but the same conditions might easily have led to infection from human beings. Fromme (1910) cites several interesting examples of temporary pollution detectable only by bacteri- ological tests. The most striking case was that of an artesian well. Its average bacterial content had been 38 per c.c. and colon bacilli were absent from 200 c.c. In May, 1908, this well became polluted from a broken stable drain 10 meters away. The number of bacteria rose to 4370 and colon bacilli were found in 10 c.c. sam- ples. The source of pollution was removed, but the well water in July still contained 7100 bacteria and B. coli in i c.c. In September the number had fallen to 105 and colon bacilli were present in 200 c.c. In Novem- ber the bacteria numbered 120 and colon bacilli were absent from 200 c.c. At no time did chemical tests give any indication of danger, while the bacteriological data obviously measured very delicately a comparatively slight but real pollution and its gradual disappearance. Similar results have been reported by Savage and Bulstrode (Savage, 1906) in the examination of the water-supply of Bridgend. It seems to the writers that the real application of chemistry begins where that of bacteriology ends. When pollution is so gross that its existence is obvious and only its amount needs to be determined, the bacteri- ological tests will not serve, on account of their exces- sive delicacy. In studying the heavy pollution of small streams, the treatment of trades wastes, and the purification of sewage, the relations of nitrogenous compounds and of oxygen compounds are of prime 226 ELEMENTS OF WATER BACTERIOLOGY importance. In other words, when pollution is to be avoided, because the decomposition of chemical sub- stances causes a nuisance, it must be studied by chem- ical methods. When the danger is sanitary and comes only from the presence of bacteria, bacteriological methods furnish the best index of pollution. In the study of certain special problems the para- mount importance of bacteriology is generally recognized. The distribution of sewage in large bodies of water into which it has been discharged may thus best be traced on account of the ready response of the bacterial counts to slight proportions of sewage, particularly since the ease and rapidity with which the technique of plating can be carried out make it possible to examine a large series of samples with a minimum of time and trouble. The course of the sewage carried out by the tide from the outlet of the South Metropolitan Dis- trict of Boston was studied in this way by E. P. Osgood in 1897, and mapped out by its high bacterial content with greater accuracy than could be attained by any other method. Some very remarkable facts have been developed by similar studies as to the persistence of separate streams of water in immediate contact with each other. Heider showed that the sewage of Vienna, after its discharge into the Danube River, flowed along the right bank of the stream, preserving its own bacterial characteristics and not mixing per- fectly with the water of the river for a distance of more than 24 miles (Heider, 1893). Jordan (Jordan, 1900), in studying the self -purification of the sewage discharged from the great Chicago drainage canal, BACTERIOLOGICAL EXAMINATION 227 found by bacteriological analyses that the Des Plaines and the Kankakee Rivers could both be distinguished flowing along in the bed of the Illinois, the two streams being in contact, yet each maintaining its own indi- viduality. Finally, the quickness with which slight changes in the character of a water are marked by fluctuations in bacterial numbers renders the bacteri- ological methods invaluable for the daily supervision of surface supplies or of the effluents from municipal nitration plants. In the commoner case, when normal values obtained by such routine analyses are not at hand, the problem of the interpretation of any sanitary analysis is a more difficult one. The conditions which surround a source of water supply may be constantly changing. No en- gineer can measure the flow of a stream in July and deduce the amount of water which will pass in February; yet the July gauging has its own value and significance, so a single analysis of any sort is not sufficient for all past and future time. If it gives a correct picture of the hygienic condition of the water at the moment of examination it has fulfilled its task, and this the bacteriological analysis can do. The evidence fur- nished by inspection and by chemical analysis should be sought for and welcomed whenever it can be obtained, yet we are of the opinion that, on account of their directness, their delicacy, and their certainty, the bacteriological methods should least of all be omitted. CHAPTER XI BACTERIOLOGY OF SEWAGE AND SEWAGE EFFLUENTS Bacteriological and Chemical Examination of Sewage. The first object of modern sewage disposal is the oxida- tion of putrescible organic matter. Chemical, rather than bacterial, purification is usually the prime requisite; and chemical tests therefore serve best as criteria of the results obtained. Bacteria are the agents in the process of sewage purification; but the most generally useful measure of the work accomplished is the chemical oxidation attained. " To employ a simile, it is a case of the saw and the 2 -foot rule — the saw will do the cutting, but the rule will measure the work cut." (W. J. Didbin.) In certain cases, however, bacterial as well as chemical purity must be effected, in view of special local require- ments. The sewage from a contagious disease hospital, for example, should be freed from infectious material as a factor of safety. Sewage discharged into a body of water adapted for bathing may well be so treated as to protect those using the water. In the case of seaboard cities where sewage effluents are likely to contaminate oyster beds and other layings of edible shellfish the problem assumes great importance. Where bacterially impure effluents are discharged into streams used for 228 BACTERIOLOGY OF SEWAGE 229 sources of water-supply the town taking water may protect itself by nitration. It should so protect itself, at any rate, from the pollution necessarily incident to surface waters; and, unless the bacterial condition of a stream or lake is made very materially worse by the discharge of sewage effluents, it is fair that the respon- sibility of purification should rest on the water works, rather than on the sewage purification plant. Shell- fish, on the other hand, cannot be purified. Either pollution must be prevented, or the industry abandoned. Under such circumstances sanitary authorities may rightly demand, as they have demanded at Baltimore, that bacteria, as well as putrescible organic matter, shall be removed in sewage treatment. Under such circumstances the bacterial control of purification plants is as essential as in the case of water filters. Methods of Bacteriological Examination of Sewage and Effluents. In England, considerable attention has been devoted to this subject, and numerous methods have been recommended as furnishing valuable criteria of the bacterial quality of sewage effluents. Houston (i902b), for example, suggests various tests involving the use of litmus milk, peptone solution, gelatin tubes, and neutral-red broth, as well as the inoculation of animals. He considers the determination of the num- bers of B. coli and B. sporogenes as of greatest moment, while the identification of streptococci is of value in certain cases and the enumeration of liquefying bacteria, spore-forming aerobes, thermophilic bacteria, and hydro- gen sulphide producing bacteria is of subsidiary impor- tance. Rideal (1906) has recently recommended a some- 230 ELEMENTS OF WATER BACTERIOLOGY what less extensive series of tests, including aerobic and anaerobic counts, both at 20 and 37°, with the determination of the number of liquefiers and the num- ber of spore-formers. The results attained do not seem to warrant any such elaborate procedure. As far as the authors are aware, the determination of liquefying bacteria, anaerobic bacteria and thermophilic bacteria does not add any information of material importance to that obtained from the total count. Some test for specific sewage organisms is of course desirable. Here again, however, the determination of B. sporogenes and sewage streptococci tells the observer little more than can be learned from the routine use of the colon test. In the United States the practise of sewage bacteriologists is crystallizing around the total count and the estimation of B. coli. In the absence of evidence as to the specific value of other data, the routine control of filter plants may well be limited to these two determinations. The total count of bacteria should be made, as in the case of waters, at 20°. Determinations carried out in duplicate at 37° give additional information of considerable value. The ratio of the 37° count to the 20° count varies with different sewages. At Boston the body temperature count is 70 to 80 per cent of the total count; at Lawrence it appears to be propor- tionately much lower (Gage, 1906). In using either medium, it is well to add lactose and litmus and note the number of red colonies, as a check on the enumera- tion of B. coli. It should be borne in mind, as Lederer and Bach- BACTERIOLOGY OF SEWAGE . 231 mann (1911) have recently pointed out, that the sampling error is a very serious one with sewage. Duplicate tests made at i-minute intervals for a period of 10 minutes in their experiments gave extreme values of 190,000 and 550,000 per c.c. The determination of the number of colon bacilli in sewage and effluents should furnish an integral part of bacteriological sewage analysis, since it is important to know whether the decrease of intestinal bacteria in the process of purification is proportional to the reduc- tion of total bacteria. The State Sewerage Commis- sion of New Jersey has adopted this procedure in its supervision of the disposal plants in that State; and the results seem amply commensurate with the labor involved. As in the case of polluted waters the enumera- tion of B. coli may be carried out, either by the study of the red colonies which appear on litmus-lactose-agar plates inoculated with the sample directly, or by the use of a preliminary enrichment process. The com- plete identification of B. coli seems unnecessarily tedious, however, where the organisms are present in such abundance. Some approximate presumptive method is indicated here, if anywhere; and the experience with polluted water, reviewed in Chapter VI, points to the Jackson bile medium as the most promising one. Experience at the Sewage Experiment Station of the Massachusetts Institute of Technology has shown that this presumptive test in general yields good results. As pointed out above, a 48-hour incu- bation period at 37° is required. All tubes showing 20 per cent gas at the end of this time may be con- 232 ELEMENTS OF WATER BACTERIOLOGY sidered positive tests for the colon group, without serious error. Numbers of Bacteria in Sewage. The total number of bacteria and the number of colon bacilli naturally vary widely in the sewages of different cities and towns. European sewages, being more concentrated, show as a rule higher numbers than are found in America. Results compiled from various sources show from 1,000,000 to 5,000,000 bacteria in the sewages of Essen, Berlin, Charlottenburg, Leeds, Exeter, Chorley, and Oxford, 2,000,000 to 10,000,000 in the sewages of Lon- don, Walton, and W. Derby and over 10,000,000 in the sewages of Paris, Ballater and Belfast (Winslow, 1905). The number of colon bacilli in English sewages varies from 50,000 to 750,000. In American sewages, on the other hand, bacteria are somewhat less numerous. At Lawrence the determinations made from 1894 to 1901 showed on the average 2,800,000 bacteria per c.c. At Worcester, Eddy reported 3,712,000 in 1901 (Eddy, 1902); at Ames, Iowa, Walker (1901) found 1,248,256 in the same year. At Columbus, Johnson (1905) reports an average of 3,600,000 bacteria per c.c.; the individual numbers varied from 320,000 to 27,000,000. The number of colon bacilli varied from 50,000 to 1,000,000 and averaged 500,000. Day samples of Boston sewage collected three times a week, from October, 1906, to April, 1907, showed an average of 1,200,000 bacteria per c.c. In the summer months numbers are notably higher than at other seasons in many sewages. Thus in 1903, Boston sewage con- tained 2,995,000 bacteria in July, 4,263,600 in August, BACTERIOLOGY OF SEWAGE 233 11,487,500 in September, 3,693,000 in October, 587,100 in November, and 712,000 in December (Winslow, 1905). There is also a marked diurnal variation in the bacterial content of sewage, since the flow con- tains a smaller proportion of intestinal matter at night than at other times. For example, a series of hourly samples at the Sewage Experiment Station of the Massachusetts Institute of Technology showed the following results: BACTERIA IN BOSTON SEWAGE— AVERAGES FOR EACH FOUR-HOUR PERIOD. AUGUST 13-14, 1903 (WINSLOW AND PHELPS, 1905) Period. Bacteria per c.c. 7:30-11:30 A.M 1 1 :3O A.M.-3 :3o P.M I,8oo,OOO 3,2OO,OOO 3:30-7:30 P.M. 4 600 ooo 7 :3o-i i :3O P.M ii '30 P M -3 '30 A.M 3,500,000 I OOO OOO 3 :3o— 7 :3O A.M 400,000 It is evident that many published results of bacterial examinations of sewage are in excess of the average values, since they refer in most cases to day samples only. Bacterial Content of Sewage Effluents. The bacterial content of sewage effluents varies widely according to the process of purification adopted and the efficiency of the particular plant. The only process which yields a notably purified effluent from the bacteri- ological standpoint is that of filtration through sand. Processes of this type when operated with care may give a bacterial purification well over 99 per cent as 234 ELEMENTS OF WATER BACTERIOLOGY shown by bacteriological examinations at the Brockton (Mass.) filters, reported by Kinnicutt, Winslow and Pratt (1910) as follows: BACTERIA IN SEWAGE AND EFFLUENTS AT BROCKTON, AVERAGE OF FOUR EXAMINATIONS, AUTUMN OF 1908 Bacteria per c.c. Gelatin 20°. Colon Bacilli per c.c. Lactose Bile. Sewage 3,1 ^O.OOO 1 50 ooo Effluent A 1,900 4.OO B 6 300 j r D I2Z o E 1,400 r F •? OOO j Such high efficiencies as this table indicates are often not realized under the actual working condi- tions of a municipal plant. At Vineland, N. J., for example, the intermittent niters show a reduction of 90 to 95 per cent in total bacteria and a somewhat higher reduction of B. coli. The results of three examinations made in 1906 are given below. BACTERIA IN SEWAGE AND SAND FILTER EFFLUENT AT VINELAND, N. J. (N. J. STATE SEWERAGE COMMISSION, 1907) Bacteria per c.c. B. Coli in Sewage. Effluent. Sewage. Effluent. March 2 480,000 20,000 O.OOOI C.C. O.OI C.C. July 26 496,000 . 6l,000 O.OOOI C.C. O.OOI C.C. July 26 511,000 38,000 O.OOOOI C.C. O.OOI C.C. BACTERIOLOGY OF SEWAGE 235 The newer bacterial processes, contact beds, and trickling niters naturally show a much less satisfactory bacterial removal than sand nitration beds. In the Columbus experiments, Johnson (1905) found from 1,000,000 to 2,000,000 bacteria in the effluents of con- tact beds and from 750,000 to 1,900,000 in the effluent from trickling niters. At the experiment station of La Madeleine, in Lille, Calmette (1907), reports 5,000,000 bacteria per c.c. in the crude sewage, 2,900,000 in the second contact effluent and 800,000 in the effluent from the trickling bed. Of 20,000 B. coli per c.c. applied to the filters, the contact system delivered 4000 and the trickling bed 2000 per c.c. The average results of examinations made three times a week at the Sewage Experiment Station of the Massachusetts Institute of Technology, during two different periods, were as follows: BACTERIA IN SEWAGE, SEPTIC EFFLUENT AND TRICKLING EFFLUENT AT BOSTON (WINSLOW AND PHELPS, 1907) Bacteria per c.c. B. Coli. Positive Tests in o.oooooi C.C.* July-Sept., 1906. Oct., I9o6-April, 1907. July-Sept., 1906. No. Per Cent Reduc- tion. No. Per Cent Reduc- tion. Per Cent. Sewage Septic effluent. . . Effluent from 1,300,000 1,650,000 1,200,000 750,000 38 65 66 Inc. trickling bed . . Septic tank and trickling bed . . 750,000 750,000 42 42 200,000 180,000 83 85 35 35 Jackson bile test. 236 ELEMENTS OF WATER BACTERIOLOGY The following average data for two of the largest trickling filter plants in the United States are cited by Kinnicutt, Winslow and Pratt (1910). BACTERIAL CONTENT OF SEWAGE AND EFFLUENTS FROM TRICKLING FILTERS Place. Period. Bacteria per c.c. Screened Sewage. Septic Effluent. Filter Effluent. Reading, Pa Columbus, Ohio. . . 1908-1909 1909 3,100,000 2,370,000 I,8oo,OOO 1 ,050,000 600.0OO 560,000 It is obvious that effluents of this character cannot be considered satisfactory from the standpoint of bacterial purification. As Houston concluded, after a careful review of the subject, " The different kinds of bacteria and their relative abundance appear to be very much the same in the effluents as in the crude sewage. Thus, as regards undesirable bacteria, the effluents frequently contain nearly as many B. coli, proteus-like germs, spores of B. enteritidis sporogenes and streptococci, as crude sewage. In no case, seemingly, has the reduc- tion of these objectionable bacteria been so marked as to be very material from the point of view of the epidemiologist" (Houston, 1902^. Experimental studies with specific bacteria have confirmed these conclusions. Houston (igo4b) found that B. pyocyaneus appeared in the effluent of a trickling bed 10 minutes after application to the top and con- tinued to be discharged for 10 days. In septic tanks and contact beds, the same germ persisted for 10 days. BACTERIOLOGY OF SEWAGE 237 Rideal (1906) quotes experiments by Pickard at Exeter, which show that typhoid bacilli may persist for 2 weeks in a septic tank and that contact bed treatment only effects a 90 per cent removal of these organisms. Disinfection of Sewage Effluents. Where bacterial purity is required, some special process of disinfection must be combined with the contact bed or the trickling filter. For this purpose treatment with chloride of lime or other chemicals is rapidly gaining ground as an important adjunct to bacterial disposal plants; and in connection with this process bacteriological control is an essential. Rideal (1906) first showed at Guildford that 30 parts of available chlorine per million would reduce the number of bacteria in crude sewage from several mil- lions to 50,000, while 50 parts would reduce their number to 20 per c.c. Colon bacilli were reduced from one million per c.c. to less than one per c.c. by 30 parts of chlorine. In septic effluent 25 to 44 parts of chlorine per million reduced B. coli from two and a half to four and a half million per c.c. to less than one per c.c. With contact effluents smaller amounts of chlorine proved efficient, The primary effluent required 20 parts per million, the secondary effluent 10.6 parts per million and the tertiary effluent 2.5 parts per mil- lion to reduce the number of B. coli so that this organism could not be isolated in 5 c.c. In this country Phelps and Carpenter (1906) demon- strated the practical usefulness of bleaching powder disinfection, at the Sewage Experiment Station of the Massachusetts Institute of Technology. As indicated 238 ELEMENTS OF WATER BACTERIOLOGY in the table below smaller amounts of chlorine than were used by Rideal will give good results with more dilute American sewages. BACTERIA IN TRICKLING FILTER EFFLUENT BEFORE AND AFTER TREATMENT WITH CHLORIDE OF LIME (5 PARTS PER MILLION AVAILABLE CHLORINE) (PHELPS AND CARPENTER, 1906) T-Jai_ Bacteria per c.c. B. Coli, Jackson Bile Test. Before. After. Before O.OOOOOI C.C. After I.O C.C. IQ06 August 1 1 .... 270,000 69 + o + o 13.... 630,000 41 O 0 + o ' 14 135,000 406 + + + o ' IS--.- 230,000 21 o o 0 O 16.... 250,000 37 + o 0 0 18 HO,OOO 40 0 O + o 20.... 90,000 54 + o 0 0 21.... 220,000 22 0 O 0 0 23. . + o 0 0 Average 24O,OOO 86 33% + 22% + Average removal .... 99.96% 99-993% The success of chemical disinfection varies with the character of the sewage or effluent treated, since the organic matter present consumes a certain amount of the disinfectant and renders it inoperative. Dis- cordant results are therefore reported from different sources. An important series of experiments tarried out in Ohio by Kellerman, Pratt, and Kimberly (1907) showed good results with sand filter effluents and BACTERIOLOGY OF SEWAGE 239 contact effluents. Septic sewage, on the other hand, required large amounts of chlorine to produce a rea- sonable bacterial reduction. The table on page 240 shows the results obtained at Marion, Ohio. In Germany, on the other hand, Schumacher (1905), Kranepuhl (1907), and Kurpjuweit (1907) found larger amounts of chlorine necessary, in the neighborhood of 60 parts per million parts of sewage. Their tests were somewhat severe, however, the criterion of success being the absence of B. coli in a large proportion of liter samples. Standards for Sewage Effluents. The science of sew- age bacteriology is in its infancy; and it is difficult to give any general rules for the interpretation of bac- teriological examinations designed to indicate whether disposal plants are successful or not. Houston stated provisionally that the 20° count should be under 100,000 and the 37° count under 10,000, while B. coli should be absent from .001 c.c. and B. sporogenes from .1 c.c. (Houston, i902b). This standard seems to us far too lenient. Either organic purity alone is necessary, as at many sewage disposal plants, or a higher grade of purity than this should be attained. It seems wisest at the present time to avoid fixing any general standards of purity for sewage effluents. Each case should be judged intelligently on its own merits. In general, however, where bacterial purification is indicated at all, it seems fair to demand that the effluent should be of such a quality as not to increase materially the bacterial content of the body of water into which it is discharged. 240 ELEMENTS OF WATER BACTERIOLOGY BACTERIA IN SEPTIC EFFLUENT, CONTACT EFFLUENT, AND SAND EFFLUENT AT MARION , O., BEFORE AND AFTER TREATMENT WITH CALCIUM HYPOCHLORITE (KELLERMAN, PRATT, AND KIMBERLY, 1907) Date. Effluent. Average Available Chlorine. Parts per Million. Bacteria per c.c. 20° C. 37° C. Total Count. Untreated. Treated. Untreated. Treated. 1907 Apr. ii Apr. 12 Apr. 15 Apr. 28 Apr. 29 Apr. 30 Mar. 21 Mar. 22 Mar. 26 Septic Septic Septic Contact Contact Contact Sand Sand Sand 4-3 6.2 7-6 2.9 5-0 4.4 3-8 3-o i-S 850,000 4,400,000 6oo,OOO IIO,OOO 65,000 500,000 49,OOO 56,000 70,000 I,IOO,OOO 550,000 400,000 2,500 1, 600 800 570 140 4,OOO 1,200,000 850,000 450,000 240,000 260,000 190,000 73,000 160,000 9,800 7,000 20,000 370 400 ISO 60 1 60 Date. Effluent. Average Available Chlorine. Parts per Million. Bacteria per c.c. 37° C. Red Colonies. B. Coli. Untreated. Treated. Untreated. Treated. 1907 Apr. ii Apr. 12 Apr. 15 Apr. 28 Apr. 29 Apr. 30 Mar. 21 Mar. 22 Mar. 26 Septic Septic Septic Contact Contact Contact Sand Sand Sand 4-3 6.2 7.6 2.9 5-o 4-4 3-8 3-0 i-5 55,000 6o,OOO IOO,OOO 7,400 15,000 51,000 20,000 I5,OOO 2O,OOO I,OOO 2,000 2,000 Not in o . 5 " 0.5 " I.O " I.O " I.O In i .0 IO,OOO 21,000 1,300 800 4,000 O 3 0 O I BACTEEIOLOGY OF SEWAGE 241 Bacteriology of the Sewage Filters Themselves. Before leaving the subject of sewage bacteriology, brief reference must be made to the importance of bacteri- ological studies in relation to the processes of sewage purification which bring about the removal of the organic matter itself. Nothing is more necessary to the development of the present art of sewage disposal than knowledge of the micro-organisms concerned and of the conditions which favor their activity; but such knowledge is woefully deficient. Something is known of the nitrifying organisms long ago discovered by Winogradsky. More recent work, like that of Schultz- Schultzenstein (1903), Boullanger and Massol (1903) and Calmette (1905), has cleared up many points concerning these forms; but much remains to be done. In regard to the reducing action of bacteria in the septic tank and contact bed we are almost wholly in the dark. Septic tanks work well with some sewages and badly with others; and the presence or absence of the right bacteria is probably largely responsible for the different results. In some cases, as at Plainfield, N. J., the seeding of a tank with cesspool contents has produced a material improvement in septic action. Knowledge of the kinds of bacteria involved would make it possible to substitute scientific control for such empiricism and might well lead to improved methods of a more intensive character than are yet available. The work already done upon a laboratory scale furnishes promise of such results. The student 242 ELEMENTS OF WATER BACTEEIOLOGY who wishes to follow out this line of investigation will find a good summary of what is already known of the hydrolysis and denitrification of nitrogenous bodies and the decomposition of cellulose and other carbohydrates in Rideal's " Sewage and the Bacterial Purification of Sewage " (1906). Gage (1905) has made a suggestive study of the bacteria which carry on the reducing changes in sewage which deserves the study of all who are interested in the more theoretical aspects of sewage treatment. His method consisted in plating sewages and effluents, isolating typical cultures and determining their power to decompose peptone and nitrates with the produc- tion of ammonia and free nitrogen. The rate of gelatin liquefaction, the amount of nitrate reduced, the amount of free ammonia formed, and the amount of nitrogen liberated were quantitatively determined for each culture thus isolated. The numerical values obtained, multiplied by the number of bacteria, apparently of the same type, observed in the plates, gave coefficients of the liquefying, denitrifying, ammonifying, and nitrogen-liberating power of the effluent; and these coefficients may be considered as measures for a given sample of the tendency of the bacterial flora to set up certain changes. The results of further studies made by Clark and Gage (1905), on sewages and on sand, contact, and trickling effluents, show that there may be important differences between various sewages in this respect which must render their purification more or less easy. They indicate that the effluents obtained from intermittent sand BACTERIOLOGY OF SEWAGE 243 filters in cold weather contain larger numbers of ammo- nifying and denitrifying bacteria than appear at other seasons, which may help to explain the poorly nitrified effluents obtained in the winter season. Along these lines research work in sewage bacteriology promises to be fruitful of results. CHAPTER XII BACTERIOLOGICAL EXAMINATION OF SHELLFISH Shellfish and Disease. The pollution of areas devoted to the growing of shellfish and the consequent pollution of the shellfish themselves is a matter of much sanitary importance. Oysters, clams and mussels are the shellfish commonly used as food, and since they are likely to be eaten in an uncooked or partially cooked condition, it is important to be assured as to their character from the bacteriological standpoint. In their normal habitats, in clean sea-water, or in river estuaries free from pollution, shellfish are unquestionably free from dangerous bacteria, although their feeding habits make it probable that the types of bacteria indigenous to the waters in which they are found might be present in considerable numbers. With the pollu- tion of streams by unpurified sewage the areas in \vhich oysters and clams develop may easily become infected by organisms of intestinal types, and there is, therefore, offered an easy means for the typhoid bacillus and other pathogenes to pass from the sewage directly into the intestinal tract of the consumer of the raw oysters or clams. The history of this subject is well summarized by Newlands and Ham (1910), from whose excellent report the following paragraphs are adapted: 244 EXAMINATION OF SHELLFISH 245 Attention was first drawn to the danger from shell- fish by the remarkable outbreak of typhoid fever which occurred in Middletown, Conn., in 1894, as a result of the serving of raw oysters at college fraternity banquets. The oysters used in this case were all derived from a certain portion of Long Island Sound, where they had been put down, or planted, in order to fatten. Investigation showed that the stream entering the Sound at this point was highly polluted, and furthermore, that at a nearby house there were two severe cases of typhoid fever from which the intes- tinal discharges were turned into the drain and thence into the stream without disinfection. The course of the passage of the bacteria from the patient suffer- ing with the disease to the oyster and so on to the young men at the banquets was, therefore, traced out in a most complete and thorough \vay. This investigation, which was conducted by Prof. H. W. Conn, of Wesleyan University, caused immediate invest- igations to be set on foot in England and in this coun- try. Two years later there followed a report by the Local Government Board of Great Britain dealing with pollution of shellfish along the English coast, and the matter has also received much attention in this country. A study of the literature reveals only a few references to oysters as carriers of disease germs previous to 1880. In that year Cameron, in a paper entitled " Oysters and Typhoid Fever," read before the British Medical Association, suggested that outbreaks of typhoid fever and cholera might be caused by eating oysters. 246 ELEMENTS OF WATER BACTERIOLOGY In 1893 Thorne-Thorne, in a report to the Local Gov- ernment Board, wrote that, in his opinion, certain cases of cholera which had occurred that year at various inland towns in England were due to eating contam- inated oysters from beds at Grimsby, where there had been a small cholera epidemic. Following ' the suggestions embodied in this report the English Govern- ment began a series of investigations which have made many important additions to our present knowledge of the subject. In 1902 the famous oyster epidemics at Winchester and Southampton, England, were proven beyond reasonable doubt to have been caused by contami- nated oysters taken from grounds at Emsworth. Here again we have to deal with banquets given in different cities where the only common source of infection appears to have been contaminated oysters. Of the 267 guests at these banquets 118 were attacked with intestinal disorders and 21 cases of typhoid fever developed, 5 of which were fatal. Although a great many sensational attacks have been made against oysters as carriers of disease germs which have been based on little or no evidence, the above-mentioned investigations and others, among which might be mentioned those of Thresh, Marvel, and Soper, have brought out sufficient trustworthy evidence to show that contaminated oysters must be considered as a real factor in the dissemination of typhoid fever and other water-borne diseases. An esti- mate of the extent to which such illness is due to oysters would be impossible at the present time. The EXAMINATION OF SHELLFISH 247 Royal Sewage Commission after an extensive investi- gation on this subject came to the following conclusion : " After carefully considering the whole of the evidence on this point, we are satisfied that a considerable num- ber of cases of enteric fever and other illness are caused by the consumption of shellfish which have been exposed to sewage contamination; but in the present state of knowledge, we do not think it possible to make an accurate numerical statement. " Moreover an examination of the figures which have been placed before us as regards those towns in which the subject has been most carefully studied shows that there may be occasional errors. Indeed the witnesses themselves recognized that absolutely accurate figures were not obtainable. " We are far from denying that isolated cases may have been due to contaminated shellfish, but we must remember that the possibility of some of them being due to other causes cannot be altogether excluded." In the above-mentioned cases, where oysters have been proven or reasonably suspected of being the cause of disease, it was found that the oysters in ques- tion had been floated or grown in heavily polluted water where direct contact with specific infection could be proven or readily assumed. The Wesleyan epidemic is a case in point. Oysters had undoubtedly been floated in the contaminated waters at Fair Haven for a number of years previous to 1894 without any noticeable effect on the health of persons eating them, but specific infection of the water from two patients in a house near by was followed by a serious epidemic. 248 ELEMENTS OF WATER BACTERIOLOGY Valuable studies of the relation between shellfish and disease have recently been published by Bulstrode (1911) and Wilhelmi (1911) and Stiles (1912). Effect of Cookery upon Polluted Shellfish. It should be noted that it is unfortunately not only raw shellfish which are responsible for the spread of disease. Most of the processes of cookery to which these foods are subjected are insufficient to destroy pathogenic germs. Clark (1906) found that clams and oysters in stews and fried and scalloped in the usual manner were generally free from colon bacilli and streptococci. With steamed clams, however, the bacteria present could not be destroyed except by a temperature high enough and prolonged enough to ruin the clams for eating. Rickards (1907) confirmed these results as to the danger from steamed clams, while he found fried clams and clams in chowder and scalloped oysters to be practically sterilized. Oyster stew, however, is not exposed to long continued heat as is clam chowder, and fried oysters are less thoroughly heated than fried clams in the ordinary processes in use. Oysters in both of these forms and fancy roast oysters still contained colon bacilli and streptococci. Buchan (1910) finds that the ordinary methods of cooking mussels do not remove the risk of typhoid infection. Bacteriological Examination of Shellfish. Without further discussing the general sanitary aspects of the subject it is important to consider just how one may determine whether the oysters from a given region are polluted or not. The methods which have been developed for this work are essentially modifications EXAMINATION OF SHELLFISH 249 of the methods used in water examination, involving sometimes total counts of bacteria at different tem- peratures, but especially the application of the various tests for the determination of the colon bacillus, since here, as in water examination, this organism may be taken as an index of pollution and its occurrence in considerable numbers must be looked upon not merely with suspicion, but as a practical proof that the supernatant waters are polluted and that the shell- fish themselves may contain organisms of pathogenic importance, such as B. typhi, B. dysenteries, B. sporo- genes and others. Determinations of the pollution of the water above the beds are sometimes made as bearing indirectly and inferentially on the possi- bility of the pollution of the shellfish contained therein. Results of the two determinations are not always in close agreement, however, owing to the rapidly changing local conditions due to tide, etc. The gen- eral relations and the individual variations between water and shellfish determinations are well illus- trated in the table on page 250 from the report by Newlands and Ham (1910) on conditions in New Haven Harbor. Study of the methods of examination of shellfish has been conducted with great care at the Lawrence Experiment Station by Gage, at the Sanitary Research Laboratory at the Institute of Technology by Phelps, at Brown University by Gorham, and in New York by Pease. Other officials of the Shellfish Commissions of different States have also carried out investigations upon this subject. The Lawrence Experiment Station 250 ELEMENTS OF WATER BACTERIOLOGY BACTERIA IN WATER AND SHELLFISH, NEW HAVEN HARBOR Water. Oysters. Station. Samples Taken. Av. Number Bacteria per c.c. Average Number B. Coli * per c.c. Average Number B. Coli * per c.c. Number Oyster Samples Character, of Bottom. 37° C. 20° C. Ferry St. Bridge 12 2 IO I26O 43 Soft Tomlin- son Br 15 9IO 2650 34 « * No. i ... 15 510 1680 Si " No. 2 15 375 9IO 73 1 < No. 3. . . 16 255 835 9 72 16 " Buoy 10 15 155 450 10 ' * No. 4. . ! IS 160 1720 9 •• No. 5. . . i? 615 1340 74 308 13 " Buoys. • 23 3iS 715 15 * * Buoy 8. . IS 205 4IO 8 • ' No. 6. . . 16 145 485 8 37 "o" Seaweed No. 7. . . 21 215 74° 29 425 ii Hard No. 96. . II 220 260 7 64 10 * ' No. QA. . 13 100 185 9 46 IO 11 No. 9. . . 12 195 200 17 37 IO 1 * No. 7 A . II I2O 240 IO 255 II * ' No. 8. . . 16 180 270 7 370 6 No. 10. . 23 300 615 9 100 i Soft No. ii .. II 405 5io 8 IO 4 * * Buoy 6. . 21 815 1690 9 . „ . . Buoy 3 . . 17 175 590 6 291 Hard No. 12. . 14 620 1190 4 6 5 * * No. 13. . 7 240 I2O IO 10 8 " No. 14. . 12 285 I 100 i — 7 8 ' ' Buoy 4. . 7 375 I4OO 4 No. 15- . 7 455 1680 i 45 IS Hard No. 16. . 14 280 1025 — i — 3 Soft No. 17. . 14 300 I26O Hard No. 19. . i 800 300 No. 20. . IO 135 860 _ IO 8 Hard Buoy 2. . II 375 90S No. 22. . 8 305 560 — 7-3 3 Hard Buoy i . . 6 us 995 — 4 12 No. 18. . IO 255 675 — No. 24. . 4 710 1340 9 10 Hard No. 23 . . 6 450 240 — ' * No. 25 . . 2 130 IOOO — No. 26. . 5 130 465 — No. 27. . IO 630 695 Soft No. 28. . 4 415 1400 — 4 3 Mud and sand No. 29. . s 370 1700 — Soft No. 30. . 5 185 440 — — IS Hard No. 31. - 7 320 130 — — IS * * No. 32. . S 70 1050 — 15 " No. 33-. 4 485 405 — — IO * * No. 34- - 4 535 495 — 10 ' ' No. 35- • 4, 120 270 ~ IS Sticky * Jackson's lactose bile presumptive test used. Minus sign after figure i indicates that the average was less than i. EXAMINATION OF SHELLFISH 251 method was published in the Massachusetts State Board of Health Report for 1905 (Clark, 1906). This method consisted in the total counts of bacteria de- veloping at 20° and 37° and the fermentation reaction in dextrose broth. Experience indicated that it was not merely necessary to examine the stomach contents of the oysters but the " shell water " as well was sub- jected to examination. With the advent of lactose bile as a better medium for the development of B. coli without interference with other types of bacteria, the substitution of this medium for dextrose broth was commonly made, and this is now one of the standard media employed for the determination. It has been noted that the superiority of lactose bile to dextrose broth is greatest in water examinations when the water is most polluted. In the study of shellfish the danger of overgrowths is even greater than in polluted waters, since the organic matter in the oyster and its surrounding shell water furnishes a culture medium for many bacteria. Streptococci are particularly abundant. As pointed out in Chapter IX, streptococci die out more rapidly than colon bacilli in potable waters, but where organic matter is present in abundance the former may survive the latter. We have compiled the table from results given by Clark (1906). It will be noted that in all cases except in that of the shell water there is a consider- able difference between the dextrose fermentation tests and the colon isolations, indicating an overgrowth by streptococci and other forms, of colon bacilli originally present. The B. sporogenes is also very frequently 252 ELEMENTS OF WATER BACTERIOLOGY responsible for such anomalous results in shellfish examinations. COLON BACILLI AND STREPTOCOCCI IN DIFFERENT PORTIONS OF THIRTY CLAMS Per Cent of Samples Showing Fermenta- tion in Dextrose B. Coli. Strepto- cocci. B. Coli and Strepto- Broth. Shell water 90 83 47 40 Gills 77 53 o- I ;; Stomach (intestine).. . . 55 35 22 12 Rectum (intestine) 82 45 43 13 Liver 77 18 15 7 Visceral tissue 18 8 7 2 It will be noted from Clark's table above that the shell liquor is not only freer from overgrowths than the portions of the body of the clam, but that the propor- tion of positive reactions is in each case higher. Since the shell water is of course easier to examine than the macerated animal, this is now generally adopted as the standard material for examination. Self -purification of Shellfish. In connection with the bacteriological examination of shellfish for colon bacilli certain investigations have been carried out which are of great importance from the commercial as well as from the sanitary standpoint. Phelps (1911) has shown that oysters which develop in waters sub- ject to sewage pollution may be purified or entirely freed from colon bacilli by the removal of the oysters themselves to waters of purer character, when, after EXAMINATION OF SHELLFISH 253 sufficient time has elapsed, the oysters will have cleansed themselves through their metabolic processes and become entirely safe even for consumption in the raw state. It is of considerable importance to determine the length of time necessary for this self-purification to take place. Obviously, from the commercial stand- point it is desirable to make it as short as possible, while from the sanitary standpoint it must be long enough to insure a thorough and satisfactory removal of all traces of polluted matter. Oyster beds which are free from pollution or which are sufficiently good for the re-laying for polluted oysters are difficult to find and limited in areas because of their nearness to sources of pollution. The investigations in question were conducted by Phelps in the Providence River and the upper part of Narragansett Bay. The oysters were removed from heavily polluted regions and car- ried to waters which were practically free from pollu- tion, where they were planted. Examinations were made from day to day in order to determine the length of time that these particular oysters showed pollution and it was found that within 4 days the organisms of the colon type were practically all eliminated. It must be borne in mind that, if shellfish are care- lessly opened and handled, they may suffer a considerable additional pollution in the process, and may therefore be much worse instead of better than when they were taken. This is well brought out by the table on page 254, taken from a report by Stiles (1911) in which shucked market oysters show much worse pollution than market oysters in the shell. 254 ELEMENTS OF WATER BACTERIOLOGY COMPARATIVE BACTERIOLOGICAL CONDITION OF MAR- KET OYSTERS, SHUCKED AND IN THE SHELL (STILES, 191 1). Number of Samples. Average per c.c. Liquor. Bacteria. B. Coli. Strepto- cocci. Plain Agar. Bile Salt Agar. 25° 37° 37° Shell oysters Shucked oysters . . 36 33 6,000 867,000 I,OOO 268,000 2OO 45,000 7 74,000 8000 Seasonal Variation of Bacteria in Oysters. It has been observed by Gorham (1912) and others that the examination of oysters from certain regions made in the' summer failed to agree with the similar analyses from the same beds made in the winter. With the advent of cold weather there seems to be a great improvement in the sanitary quality, so that oysters taken from beds in close proximity to the outfalls of large sewers show in the colder months entire absence of any evidence of contamination, judged solely by the bacteriological data. Thus Gorham found in the sum- mer of 1910 that all oysters on the beds in the Prov- idence and Warren Rivers and the upper part of Narra- gansett Bay were so badly polluted by sewage as to be unfit for jood. Colon bacilli were found in the " shell water " of every oyster in amounts as small as .01 of a cubic centimeter or less. Chemical and bacteriological examination of the waters over these EXAMINATION OF SHELLFISH 255 beds showed them to be heavily sewage polluted. In December of the same year the analyses of the oysters were strikingly different, although the condition of the water was apparently unchanged. In the examina- tion five oysters were selected in each case and the average total number of bacteria per cubic centimeter was determined and the presence of colon bacilli was tested by the bile tube and subsequent isolation and identification of the organisms. The table on page 256 shows the numbers of bacteria found, and the propor- tion of the five oyster samples in which colon bacilli were present in cubic centimeter amounts and also in o.i and o.oi of a cubic centimeter. The conclusions arrived at by Gorham are that during the cold weather the oysters assume a condition of rest or hibernation, during which time ciliary move- ment ceases and the process of feeding is suspended. No organisms are therefore taken in from the outside water and those inside the oyster are gradually elim- inated, so that the total number of organisms is reduced very considerably and the oyster becomes practically free from colon bacilli. Standard Methods for the Examination of Shellfish. The examination of shellfish for pollution is regarded as of such importance by the American Public Health Association (1912) that a committee was established to report upon methods of examination and estima- tion of the numbers of colon bacilli found. The following abstract of the second report of this committee gives the recommendations for standard methods for bacteriological examination of shellfish 256 ELEMENTS OF WATER BACTERIOLOGY SEASONAL VARIATION IN THE BACTERIAL CONTENT OF OYSTERS (GORHAM, 1912) _"£ u, " Proportion of Five Date. till Oysters Showing B. Coli in Score. B. Coli Present in Water in Tempera- ture of Water > rt •*-! •+. <; w o o I C.C. O.I C.C. o.oic c. BED No. PROVIDENCE RIVER Dec. 20, 1910 IOOO 3 i o 4 0 01 C.C. -1° Jan. 14, IQII 750 5 3 i 4i Jan. 2<. . 80 4 3 o 23 O.OI C.C. 1° Tan 27. 23 c 3 o 32 Feb. 10 *o 130 o 2 2 o O * 4 I .O C.C. 0.1° Feb. 28 140 O 0 0 o O.OOOI C C. 1° Mar. ii 2OO 5 4 o 41 O OI C.C. 1.75° April 14 275 5 2 o 23 O.OI C.C. -*• i O 8-5° April 28 700 5 5 4 410 O.OOOI C.C. 12-5° May 12 I7OO s < < =;oo 0 0001 C.C. 15° BED No. 44. PROVIDENCE RIVER o Jan. 7, 191 1.. 425 5 5 i 140 0.25° Feb 10 2 Baton, 107, in, 117. 184, 193, 210. Bachmann, 64, 231. Cramer, 54. Beckmann, 143. Conn, 245. Belcher, 213. Conradi, 76. Belitzer, 140. Copeland, 102, 105, 192. Bergey, 180, 191. Bettencourt, 141, 142, 147, 172. Biffi, 176. Davis, 118, 199. Blachstein, 144. Deehan, 180, 191. Blunt, 16. Dibdin, 228. Bolton, 53. Ditthorn, 85. Borges, 141, 142, 147. Doebert, 77. Boullanger, 241. Dolt, 130. Braun, 122. Downes, 16. Brezina, 55. Drew, 13, 36. Brotzu, 140. Drigalski, 76, 85. Brown, 151. Duclaux, 6. Bmns, 144. Duggeli, 147- Buchan, 248. Dunbar, 97. Buchner, 17. Dunham, 56, 96. Bulstrode, 225, 248. Durham, 93, 95. Burri, 150, 183. Dyar, 140. 307 308 INDEX OF AUTHORS Eddy, 232. Egger, 26. Eijkman, 120. Ellms, 36. Eisner, 75. Endo, 76. English Committee, 136. Escherich, 99. Eyre, 140. Fehrs, 15. Ferguson, 116. Ferreira, 102, 141. Ficker, 80, 84. Fischer, 2, 26, 90. Flatau, 90. Fox, 91. Frankland, 12, 37, 98. Fremlin, 140. Freudenreich, 16, 150. Fromme, 113, 119, 121, 142, 163, 171, 225. Frost, 1 6, 179. Fuller, 16, 32, 43, 57, 116, 182. Gaehtgens, 77. Gage, 8, 9, 30, 31, 43, 44, 45, 46, 69, 70, 72, 107, 108, 109, in, 122, 133, 146, 157, 162, 166, 177, 186, 187, 203, 218, 230, 242, 249. Garre, 15. Gartner, 14, 35, 53, 159, 170, 171. Gautie, 172. Gildemeister, 85. Gordan, 147, 209. Gorham, 249, 254, 256. Hachtel, 79, 91, 97. Hale, 116, 126, 127, 129. Ham, 244, 249, 250, 264. Hammerl, 151. Hammond, 113, 115, 116. Hankin, 75. Hansen, 221. Harrison, 129. Hazen, 19. Heider, 226. Heraeus, 35. Hesse, 29, 45, 47. Hilgermann, 120. Hill, 36, 64. Hiss, 78, 89. Hoffmann, 24, 80. Holman, 213. Hoover, 102, 192. Horder, 199, 209. Horhammer, 15. . Horrocks, 15, 26, 87, 153, 182, 203. 207, 212. Horta, 102, 141. Houston, n, 21, 22, 24, 27, 96, 115, 141, 146, 150, 152, 158, 159, 162, 178, 185, 189, 197, 202, 207, 209, 213, 222, 236, 239- Howe, 175, 177, 180-199. Hunnewell, 115, 118, 203. Huntemiiller, 15. Irons, 106, 109, 120, 122. Jackson, 64, 78, 81, 91, 122, 123, 131, 192. Janowski, 6. Johnson, 58, 142, 182, 235, 232. Jordan, 14, 16, 17, 20. 23, 39, 55, 153, 155, 226. Kabrhel, 25. Kaiser, 163. Keith, 140. Kellerman, 26, 238, 240. Kimberly, 238, 240. Kinnicutt, 234, 236. Kisskalt, 9. INDEX OF AUTHORS 309 Kloumann, 80. Klein, 75, 81, 146, 211, 213. Kligler, 148. Klotz, 82, 213. Koch, 97. Kohn, 38, 53. Konradi, 24. Konrich, 101, 102, 121, 141, 147, 148, 163, 169, 170, 171, 176, 183, 187, 188. Korschun, 15. Kranepuhl, 239 Kruse, 143, 169. Kiibler, 90. Kurpjuweit, 239. Laws, 92, 202. Lederer, 64, 231. LeGros, 203. Lemke, 77. Lentz, 77. Levy, 144. Lochridge, 19. Loeffler, 77, 96. Longley, 107, in, 117, 165. Lubenau, 87. MacConkey, 122, 175, 177, 178, 180, 184, 189, 190, 191. Makgill, 122. Marshall, 176. Marvel, 246. Maschek, 26, 35. Mass. State Board of Health, 136, 161. Massol, 241. Massini, 183. Mathews, 66, 105. Mayer, A., 18, 19. Mayer, G., 56. McWeeney, 175. Melia, 81, 91, 116, 126, 127, 129. Miquel, 6, 37,47, Si, 55- Moore, 140. Moroni, 144. Muer, 131. Muller, 30, 31, 36, 84, 183. Neufeld, 90. Neuman, 13, 120, 147. Newlands, 244, 249, 250, 264. Nibecker, 12, 67, 68, in, 133, 204. Niedner, 45, 47. Nieter, 85. Nowack, 77, 121. Nuttall, 211. Orlandi, 161. Osgood, 226. Otto, 13. Pakes, 153. Palmer, 209. Papasotiriu, 146. Paredes, 102, 141. Parietti, 74. Park, 89. Pease, 249. Peckham, 182. Penfold, 183. Petruschky, 168. Phelps, 30, 31, 43, 44, 71, in, 113, 115, 116, 122, 133, 177, 186, 187, 233, 237, 249, 252. Philbrick, 10, n, 54. Poujol, 144. Pratt, 234, 236, 238, 240. Prescott, 12, 26, 43, 67, 107, 146, 203, 204. Procaccini, 17. Pusch, 168. Rapp, 17, 18. Refik, 143. Reinsch, 43. Remlinger, 93. 310 INDEX OF AUTHORS 168 Reynolds, 119. Rickards, 248. Rideal, 66, 229, 237, 242. Riedel, 38. Rivas, 1 80. Rogers, 118, 199. Rondelli, 161. Roth, 80. Rothberger, 122. Ruediger, 22, 23. Russell, 13, 16. Sauerbeck, 183. Savage, 24, 62, 122, 141, 145 214, 225. Sawin, 124, 125. Schepilewski, 83. Scheurlen, 14. Schneider, 93. Schottelius, 96. Schuder, 84. Schultz-Schultzenstein, 241. Schumacher, 239. Sedgwick, 4, 21, 43, 95, 105. Shell Fish Commission, A.P.H.A 255- Shuttleworth, 59. Smith, 101, 106, 140, 150, 180. Soper, 246. Spitta, 14. Stamm, 117. Starkey, 86, 95, 96. Sternberg, 52. Stiles, 248. Stokes, 79, 91, 97, 122, 127. Stokvis, 15. Stoner, 127. Swellengrebel, 15. Thoman, 75. Thorne-Thorne, 246. Thomson, 75. Thresh, 36, 95, 246. Tiemann, 14, 35. Tietz, 77. Tully, 73. Twort, 183. Vallet, 84. Van Buskirk, 50. van der Leek, 129. Vincent, 56, 172, 213, 223. Walker, 101, 147, 180, 232. Wathelet, 92. Weissenfeld, 145. Welch, 211. West, 180. Wheeler, 18. Wherry, 178. Whipple, 18, 19, 38, 39, 41, 44, 46, 63, 71, 72, no, 118, 181, 221. Whittaker, 26. Widal, 89. Wilhelmi, 248. Willson, 80, 84, 90. Winslow, 12, 19, 21, 29, 30, 34, 60, 66, 67, 68, 71, 101, in, 118, 133, 147, 148, 155, !8o, 199- 203, 209, 213, 221, 232, 233, 234, 236. Wolffhugel, 26, 38. Woodman, 221. Wright, 140. Wurtz, 65, 104. Zagari, 15. Zeit, 16, 23. SUBJECT INDEX Acid-forming organisms, 65. Acid wastes, antiseptic effect of, 20. Aesculin, 129. Aesculin bile medium, 129. Aesculin medium, preparation of, 279. Aesculin test, 129. Agar, drying of, 78. Agar, for body temperature count, 63. Agar, preparation of, 270. Agglutination of typhoid bacilli, 81, 82, 83. Anaerobic bacteria, 201. Anaerobic spore-forming bacilli, 211. Anglo-American procedure, 172. Aniline dyes, use of, 77. Antagonism, 16. Anthrax, occurrence in water, 98. Apparatus, treatment of, 279. Arbitrary standards, fallacy of, 51. Atmospheric waters, 5. Atypical colon bacilli, 135. Azolitmin, 272. Bacillus, anaerobic spore-forming, 211. Bacillus acidi lactici, 189. Bacillus aerogenes, 189. Bacillus alcaligenes, 94. Bacillus anthracis, 98. Bacillus coli, 94, 99. cultural features, 100. discovery, 99. distribution in waters, 152. effect of temperature on, 113. fermentation reactions, 101. importance of numbers, 149. index of pollution, 219. in cold-blooded animals, 142. in ground waters, 113. in sewage, 230. in soils, 152. isolation of, 104, 132. isolation by Conradi-Drigalski medium, 106. isolation by Endo medium, 106. isolation by synthetic media, 130. mutations in, 183. occurrence, 99. occurrence in animals, 140, 141. occurrence on plants, 146. pathogenicity, 100. positive isolations, no. preliminary enrichment of, 106. quantitative test, 138. standard tests for, 115. streak characteristics, 134. ubiquity of, 143. Bacillus coli and sewage strepto- cocci, 205, 206. 311 312 SUBJECT INDEX B. communis, 181. B. communior, 181. B. dysenteriae, 89, 94. B. enteritidis, 94. B. mycoides, 133. B. sporogenes, 211. characteristics of, 213. growth in liver broth, 212. in sewage, 213. isolation of, 211. occurrence of, 212. B. typhi, 94. identification of, 87. comparison with B. coli, 87. in oysters, 91. isolation from water, 89, 90. isolation of, 90. small numbers in water, 92. B. welchii, 129, 134, 138, 211. Bacteria, affected by temperature, 20. as agents of decomposition, 3. count of as index of efficiency of niters, 59. counts of on different media, 44. development at high tempera- tures, 69. distribution, i. diurnal variation of in sewage, 233- effect of composition of media on, 19. effect of light, 16, 17. effect of rainfall on, 7. effect of storage on, 10. estimation of, 48. expression of counts of, 49. factors influencing diminution, 13- field determinations of, 49. food requirements, 2. in contact effluents, 240. in disinfected effluents, 240. Bacteria in driven wells, 27. in dust and air, 6. in earth, 6. in filtered waters, 57. in ground waters, 25. in lakes and ponds, 12. in ocean, 13. in oysters, seasonal variation of, 254, 256. in polluted streams, 7. in rain and snow, 6. in sand effluents, 234, 240. in septic effluents, 240. in sewage, 232. in sewage effluents, 233. in shallow wells, 26. in springs, 26. in surface waters, 54. in trickling filter effluents, 236. in unpolluted streams, 7. in water, 5. in water and shellfish, 249, 250. metatrophic, 2. microscopic enumeration of, 29. mineral nutrients for, 53. multiplication in stored waters, 37- nitrifying, 4. number of as index of purity, 60. numbers of in sewage, 232. occurrence, i. paratrophic, 2. pathogenic in water, 74. proto trophic, 2, 29. relation to character of food, 29. quantitative methods of deter- mination, 29. seasonal variation of, 7. sedimentation of, 14. Bacteriological examination of shellfish, 244, 248. Bacteriological examination of sewage, methods of, 229. SUBJECT INDEX 313 Bacteriological examinations of water, significance of, 217. Bacteriological methods for super- vision of filtration plants, 227. Bacteriological methods for super- vision of water supplies, 227. Bacteriological methods in detect- ing sewage distribution, 226. Bacteriology of sewage, 228. Bacteriology of sewage effluents, 228. Bacteriology of sewage filters, 241. Bacteriological examination, 220. advantages of, 220. certainty of, 223. delicacy of, 221. Bile, importance of, 127. Bile media, 81, 122. Bile salts, 122. selective action of, 123, 128. Body temperature count, 42, 61. relation to hot weather, 70. Caffein, selective action of, 80. Calcium hypochlorite, 238. Carbon dioxide, absorption of, 117. Chemical disinfection, 238. Chlorine disinfection, relation to counts, 72. Cholera red reaction, 97. Cholera spirillum, isolation from water, 96. media for, 96. Clams, 244, 263. Cold, action of on bacteria, 21. Colon bacilli, 64, 99. as index of pollution, 168. as index of self-purification, 153- atypical, 135, 177. colonies of, 133. comparison with B. typhi, 87. Colon bacilli, distribution in waters, 152. effect of temperature on, 22. " flaginac," 185. importance of numbers, 149. increased survival in cold weather, 22. in cultivated soil, 170. in dust, 148. in filter effluents, 166. in filtered waters, 163. in foods, 147. in fruits, 147. in grains, 146. in ground waters, 161. in sewage, 231. in sewage effluents, 231. in shallow wells, 162. in shellfish, 252. in soil, 148. in surface waters, 157. in unpolluted waters, 155. isolation of, 104. isolation by bile media, 122. on plants, 146. standard tests for, 115. " typical," 176. ubiquity of, 143. varieties of, 174. viability at different tempera- tures, 128. Colon group, 99. characteristics, 104. distribution in water, 185. Jackson's classification of, 192. McConkey's classification of, 189. statistical classification, 198. tests for, 174. variations in, 181. Colon test, 101. Colon typhoid group, 94. reactions of, 95. 314 SUBJECT INDEX Comparison of sand and mechan- ical niters, 58. Composition of medium, impor- tance of, 43. Confirmatory tests, 136. Conradi-Drigalski medium, 76. for B. coli, 106. preparation of, 275. Contact beds, 235. Counting, 48. Crystal violet, 76. Culture media, ingredients for, 265. preparation of, 265. reaction of, 267. sterilization of, 266. titration of, 267. uniform methods for, 265. Deep wells, bacteria in, 27. Dextrose broth, 107. advantages of, 108. comparison with lactose bile, 125. disadvantages of, 108. Dextrose test, failure of, 116. . Diluting samples, 40. Disinfection of sewage, 237. Disinfection of sewage effluents, 237- Distribution of types of colon group in waters, 184. Division of colon group, 1 74. Dunham's solution, 97. Dysentery, spread by water, 95. Eijkman test, 120. Endo medium, 76. for B. coli, 106. preparation of, 276. Examination of shellfish, standard methods for, 255, 257. Examination of shell water, 251. Expression of quantitative re- sults, 49. Fermentation of lactose, 114. Fermentation test, 105, 109, 179. effect of temperature on, 112. exceptions to, in. interpretation of, no. Field kits, 49. Field methods, 50. Filter plants, routine control of, 230. Filtered waters, bacteria in, 57. Filtration in Japan, 58. " Flaginac " B. coli, 185. Food supply, importance of, 20. Gartner bacillus, 95. Gas-forming bacteria, growth in liver broth, 131. Gas production in vacuo, 118. Gas ratio, 100, 109. effect of age on, 117. unreliability of, 118. Gelatin liquefaction, 177. Gelatin plates, use of, 41. Gelatin, preparation of, 270. Green plants, food requirements, 3- Ground waters, 6. bacterial content of, 56. B. coli in, 113. Hesse medium, 78. drying of, 78. preparation of, 275. High temperatures, significance of, 69. Hiss agar medium, 78. preparation of, 277. Hog cholera bacillus, 94. Incubation, 46, 280. SUBJECT INDEX 315 Incubation period, 41, 47. for body temperature count, 64. Incubator, necessity for moisture in, 46. Indol test, 176. Infusoria, destruction of bacteria by, 15- Interpretation of results, 51. Intestinal bacilli, 103, 201. Isolation of B. coli, 104, 132. by bile media, 122. Isolation of cholera spirillum, 97. Isolation of streptococci, 203. Lactose bile, 81, 122, 126. action of bacteria on, 102. advantage of, 127. decomposition of, 64. comparison with dextrose broth, 125. preparation of, 269. Lactose fermentation, 114. importance of, 115. Lactose fermenting bacilli, 102. effect of storage on, 115. " Lamirascsal," streptococci, 210. " Larasacsal " streptococci, 210. Light, destructive effect on bac- teria, 16, 17. Litmus lactose agar, 64, 104. counts, 65. incubation of, 132. preparation of, 272. Liver agar, preparation of, 274. broth, 131. preparation of, 273. Liver gelatin, preparation of, 274. Malachite green agar, 77. McConkey's group in filtered water, 198. McConkey's groups in raw waters, 198. McConkey's groups in stored waters, 198. Mechanical filtration, 57. Methods of estimation, 48. Middletown outbreak, 245. Milk, preparation of, 278. Mussels, 244. Mutations in B. coli, 183. Nahrstoff agar, ratio, 31. Nahrstoff gelatin, ratio, 31. Nahrstoff-Heyden agar, 29. Neutral red, 122. Neutral red reaction, 121. Neutral red test, 178. Nitrate broth, preparation of, 287. Nitrates, 3. Nitrite test, 177. Nitrogen cycle, 4. Nitroso-indol reaction, 97. Nutrient broth, preparation of, 268. Nutrose, 76. Obligate parasites, 29. Overgrowth, effect of, 119. Oysters, 244. B. coli in, 259. bacterial counts of, 258. bacteriological examination of, 257- opened, examination of, 262. rating for B. coli, 260. sampling, 257. Oysters and typhoid, 246. Para-colon bacilli, 187. Para- typhoid bacilli, 94. 316 SUBJECT INDEX Pathogenes in water, 74. Pathogenic bacteria, 74. Peptone, importance of, 44. Petri dishes, 105. Phenol agar, 105. Phenol broth, 119. Phenol dextrose broth, 108. Plate method, 105. Plating, 40. Polluted shellfish, effect of cook- ery on, 248. Polluted waters, isolation of B. coli from, 107. Pollution, progressive, 223. of shellfish, 245. temporary, 225. Porous tops, 64, 105. Precipitation of typhoid bacilli, 81, 83, 84. Preliminary enrichment, 106. Presumptive test, no. Presumptive tests, various, 126. Presumptive tests with bile media, 124. Progressive pollution, detection of, 223. Prototrophic bacteria, 29. Protozoa, reduction of bacteria by, 15. Pumping, effect of on bacteria, 35. Quantitative bacteriological de- termination, 29. interpretation of, 51. Quantitative results, expression of, 49. Rainfall, effect of on bacteria, 7. Reaction, importance of, 43. Reaction of culture media, 267. Reaction optimum, 43. Reducing bacteria, importance of, 242. Relation between room tempera- ture and body temperature counts, 61, 67, 71. Room temperature counts, 61. Saccharose, fermentation of, 180. Samples, dilution of, 40. Samples, icing of, 39. Sampling, 33. Sand nitration, 57. Sanitary chemical analysis, sig- nificance of, 216. Sanitary inspection, 215. importance of, 172. Seasonal variation, 72. Sedimentation, 7. Sedimentation of bacteria, 14. Self-purification, 20. Selective media, 219. Selective temperatures, 219. Sewage, bacteria in, 214. bacteriology of, 228. colon bacilli in, 231. Sewage effluents, bacteriology of, 228. colon bacilli in, 231. standards for, 239. Sewage examination, use of bile medium in, 231. Sewage sampling, error of, 231. streptococci, 134, 201. as index of pollution, 202. isolation of, 204. Sewage streptococci and B. coli, growth in dextrose broth, 205, 206. detection of, 206. Sewage and sewage effluents, 228, 229. Shallow wells, bacteria in, 26. SUBJECT INDEX 317 Shallow wells, body temperature count in, 67. Shellfish and disease, 244. bacteriological examination of, 248, 251. bacteria in shucked, 254. careless handling of, 253. colon bacilli in, 252. self-purification of, 252. standards of interpretation, 264. streptococci in, 252. Shell water, examination of, 251. Significance of 37° count, 62. Specific sewage organisms, tests for, 230. Springs, bacteria in, 26. Standard msthods, 32, 33, 41. necessity of, 45. Standard reaction, 268. Standards for sewage effluents, 239- Staphylococci in sewage, 202. Storage, effectiveness of, 25. effect of on bacteria, 10, 21. effect of duration of, 23. effect on lactose fermenters, 115. Storage of samples, effect of, 37. Stored waters, 6. Streptococci, 64, 133. antagonism to colon bacilli, 208. | comparative fermentations by, 210. from different animals, 209. in sewage, 202. in saliva, 204. in shellfish, 252. in polluted waters, 203. in stored sewage, 207. indicative of recent pollution, 207. index of pollution, 220. isolation of, 203. on animal bodies, 204. Streptococci, varieties of, 208, 209. Streptococci, " lamiracsal," 210. Streptococci, " larasacsal," 210. Streptococcus equinus, 209. Sugar broths, preparation of, 269. Sugar reactions, 179. Sugars, action of bacteria on, 102. Surface waters, 5. bacterial content of, 54. Swimming pools, bacteria in, 72. Synthetic media, 130. Temperature, effect on B. coli, 113. effect on bacteria in water, 20. effect on fermentation test, 112. Temporary pollution, detection of, 225. Titration of culture media, 267. Toxic products, effect of, 15. Trickling filters, 235. Typhoid, occurrence in cold weather, 22. Typhoid and shellfish, 245. Typhoid bacilli, agglutination of, 81, 82, 83. artificial infection of water with, 24. developing on malachite green media, 77. effect of oxygen on, 19. enrichment in caffein media, 79. enrichment of, 75. examination of water for, 74. in polluted waters, 23. in pure culture, 16. in tap water, 24. in unsterilized waters, 23. isolation by lactose bile, 81. isolation of, 7.6. media for, 76, 77, 78, 79. precipitation of, 81, 83, 84. 318 SUBJECT INDEX Typhoid, preliminary enrichment of, 79. separation by motility, 85. small numbers in water, 92. summary of isolation methods, 86. uncultivated strains in water, 24. viability in mud, 24. viability in sewage, 16. viability in water, '15. " Typical" B. coli, 176. Unpolluted waters, body tempera- ture count in, 68. Urea, decomposition of, 3. Voges-Proskauer reaction, 180. Waters, classification of, 5. Wells, bacteria in, 26, 56. B. coli in, 26, 162, 163. deep, 27.