flbacmillan's Science STUDIES IN WATER SUPPLY MACMILLAN AND CO., LIMITED LONDON . BOMBAY . CALCUTTA MELBOURNE THE MACMILLAN COMPANY NEW YORK . BOSTON . CHICAGO DALLAS . SAN FRANCISCO THE MACMILLAN CO. OF CANADA, LTD. TORONTO STUDIES IN WATER SUPPLY BY A C. HOUSTON, D.Sc., M.B., C.M. Director of Water Examination, Metropolitan Water Board MACMILLAN AND CO., LIMITED ST. MARTIN'S STREET, LONDON 19*3 COPYRIGHT RICHARD CLAY AND SONS, LIMITED, BRUNSWICK STREET, STAMFORD STREET, S.E., AND BUNGAY, SUFFOLK. PREFACE FOR several years past I have been urged to write a text- book on Water Supply, but I have been unable to find the time to devote to a task that seemed so formidable. When, however, the Editor of the series in which this volume appears, was good enough to ask me to write, not a text-book, but a monograph dealing with my own personal experiences and investigations, I felt the case was different. It was also pointed out to me that this did not involve an exhaustive review of other men's work, but only gathering together in one volume an epitome of my own researches, which are necessarily scattered through a considerable number of reports and papers. London possesses the largest water works undertaking in the world ; and the Metropolitan Water Board, by instituting an elaborate system of water examination, by providing facilities for carrying out many important researches, and generally by looking at the whole subject of water supply from an imperial point of view, has shown itself fully alive to its serious responsibilities. The investigations which have been carried out are now public property, but in endeavouring to bring into one volume the results of my experience gained under such favourable auspices I incidentally pay a tribute to the VI PREFACE pioneer work and wise counsels of the Metropolitan Water Board. I am much indebted to Mr. H. F. Fermor for making the drawings illustrating this monograph, to Dr. A. Norman for the microphotographs of Algse, to Mr. J. I. Goodlet for compiling the index, and assisting in other ways, and to Dr. D. G. Sutherland, Mr. A. E. Jury, and several of my other assistants for their in- valuable help. My special and grateful thanks are also due to Prof. R. A. Gregory, the accomplished Editor of this series of monographs. His advice and experience have been of inestimable value to me, I desire to dedicate this monograph to my Staff, as a small token of my deep appreciation of their willing help and loyal co-operation during the eight years I have had the honour of serving the Metropolitan AVater Board as Director of Water Examination. A. C. HOUSTON. September, 1913. CONTENTS CHAPTER I SOURCES OF WATER SUPPLY PAGE Upland gathering grounds underground sources of supply rivers health of London Rivers Thames and Lee considered as sources of water supply V . 1 CHAPTER II RESEARCHES TENDING TO JUSTIFY RIVERS AS SOURCES OF WATER SUPPLY The streptococcus test tests for pathogenic bacteria results of the examination of raw river water for the typhoid bacillus and Gartner's bacillus results of search for the typhoid bacillus in crude sewage general observations 27 CHAPTER III THE QUESTION OF ABSTRACTION Variations in the Thames colour results during a period of six years effect of different methods of abstraction on colour results number of days during a period of six years that the colour results were 100 and over, and over 200 epidemiological considerations deterioration of water as supplied to consumers during the winter months of the year 39 CHAPTER IV SUPPLEMENTARY PROCESSES OF WATER PURIFICATION York mechanical gravity filters Chester's proposed pressure filters small pre-storage settlement reservoirs advantages of passing raw river water through small reservoirs antecedent to storage in large reservoirs New River results Sunbury results advantage of using coagulants, antecedent to the storage of raw river water in large reservoirs Sunbury results Puech-Chabal system of pre-filters 51 LIST OF ILLUSTRATIONS 1. RIVERS THAMES AND LEE. CHEMICAL RESULTS FOB THE YEARS 1906-1911 .... 8 2. RIVER THAMES. AMMONIACAL AND ALBUMINOID NITROGEN RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DECEMBER SlST, 1911 . . 9 3. RIVER LEE. AMMONIACAL AND ALBUMINOID NITROGEN RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DECEMBER SlST, 1911 . . 10 4. RIVER THAMES. PERMANGANATE AND TURBIDITY RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DECEMBER 31sT, 1911 . . 11 5. RIVER LEE. PERMANGANATE AND TURBIDITY RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DECEMBER 31sT, 1911 . . 12 6. RIVER THAMES. COLOUR RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DECEMBER 31sT, 1911 ...... 13 7. RIVER LEE. COLOUR RESULTS FOR THE SEVENTY-TWO MONTHS ENDED DKCEMBER 31sT, 1#11 14 8. RIVERS THAMES AND LEE. BACTERIOLOGICAL RESULTS FOR THE YEARS 1906-1911 17 9-11. RIVER THAMES. BACTERIOLOGICAL RESULTS FOR THE SEVENTY-TWO MONTHS, FIFTY-THREE MONTHS AGAR AND BlLE-SALT-AGAR TESTS, ENDED DECEMBER 31ST, 1911 19,20,21 12-14. RIVER LEE. BACTERIOLOGICAL RESULTS FOR THE SEVENTY-TWO MONTHS, FIFTY-THREE MONTHS AGAR AND BjLE-SALT-AGAR TESTS, ENDED DECEMBER 31ST, 1911 23,24,25 15-17. DIAGRAMS ILLUSTRATING CERTAIN POINTS AS REGARDS ABSTRACTION OF RIVER WATER FOR WATERWORKS PURPOSES . . . 41, 43, 45 18. ALL LONDON WATERS. B. COLI RESULTS FOR THE SIXTY MONTHS ENDED DECEMBER 31 ST, 1911 .... .... 49 19. DIAGRAM ILLUSTRATING THE ADVANTAGES OF SIMPLE SEDIMENTA- TION 55 20-22. DIAGRAMS ILLUSTRATING THE BENEFICIAL EFFECTS OF STORAGE 97, 98, 99 23, 24. DIAGRAMS ILLUSTRATING THE TYPHOID FEVER DEATH RATES IN CERTAIN AMERICAN CITIES .... Facing pages 122 and 123 25. DIAGRAM ILLUSTRATING THE ANNUAL DEATH RATE AMONG CHILDREN FROM DlARRH(EAL DISEASES , 129 xii LIST OF ILLUSTRATIONS FKi. 1 26. TYPHOID FEVER DEATH RATES IN CERTAIN EUROPEAN AND AMERICAN CITIES 27. REPRODUCTION OF A LABEL OF A SAMPLE BOTTLE ... 28. ILLUSTRATION OF A PAGE IN A LABEL HOOK 29. SAMPLE COLLECTION BOX . . 30. RACK AS SET UP FOR THE EXAMINATION OF A SAMPLE OF WATER 31. APPARATUS FOR KEEPING MELTED AGAR TUBES AT A UNIFORM TEMPERATURE OF 45 C 32. RACK AS SET UP FOR THE EXAMINATION OF A SAMPLE OF WATER AND THE ALTERATIONS AT EACH STAGE OF THE PROCESS OF EXAMINATION . 33. AN ILLUSTRATION OF THE AUTHOR'S METHOD OF MAKING SUB- CULTURAL TESTS FOR B. COLI 34. CLASSIFICATION OF B. COLI 35. B. COLI TEST. RAW WATERS (THAMES, LEE AND NEW RIVER) . 36. B. COLI TEST. LONDON WATERS .... 37. DIAGRAM ILLUSTRATING THE DELICACY OF THE METHOD FOR ISOLATING THE CHOLERA VIBRIO FROM WATER .... . 38-43. ILLUSTRATIONS OF SOME ALGAL GROWTHS (ASTERIONELLA, OSCTL- LARIA, TABELLARIA, DINOBRYON) . ... 195, 196, STUDIES IN WATER SUPPLY CHAPTER I SOURCES OF WATER SUPPLY THERE are three main sources of water supply, namely : (1) Upland Gathering Grounds. In most cases the The gelatin, agar and bile-salt-agar results are per O'Ol, 0*1. and 1 c.c. respectively. ui unpolluted or sparsely populated land, and may require little or no purification before distribution. Birmingham, Leeds, Wakefield, Edinburgh, Cardiff, Manchester, and Liverpool are examples of towns supplied with water from upland sources. Fifty-three samples of water collected from taps in these towns were found to contain B. coli (lactose + indol + ) in only four cases, even when as much as 100 cubic centimetres of water were submitted to such cultural tests as will afterwards be described. (2) Underground Sources of Supply. Shallow wells are presumably quite safe, or highly dangerous, according to their environment. Deep wells are usually satisfactory sources of supply, especially if the boring has been carried through an B xii LIST OF ILLUSTRATIONS FKi. PAGE 26. TYPHOID FEVER DEATH KATES IN CERTAIN EUROPEAN AND AMERICAN CITIES 133 27. REPRODUCTION OF A LABEL OF A SAMPLE BOTTLE ... .139 28. ILLUSTRATION OF A PAGE IN A LABEL KOOK 139 29. SAMPLE COLLECTION BOX . 141 30. RACK AS SET UP FOR THE EXAMINATION OF A SAMPLE OF WATER 152 31. APPARATUS FOR KEEPING MELTED AGAR TUBES AT A UNIFORM TEMPERATURE OF 45 C .153 32. RACK AS SET UP FOR THE EXAMINATION OF A SAMPLE OF WATER AND THE ALTERATIONS AT EACH STAGE OF THE PROCESS OF EXAMINATION . 155 33. AN ILLUSTRATION OF THE AUTHOR'S METHOD OF MAKING SUB- CULTURAL TESTS FOR B. COLI 163 34. CLASSIFICATION OF B. COLI 165 35. B. COLI TEST. RAW WATERS (THAMES, LEE AND NEW RIVER) . . 167 36. B. COLI TEST. LONDON WATERS .... .169 37. DIAGRAM ILLUSTRATING THE DELICACY OF THE METHOD FOR ISOLATING THE CHOLERA VIBRIO FROM WATER .... . 177 38-43. ILLUSTRATIONS OF SOME ALGAL GROWTHS (ASTERIONELLA, OSCTL- LARIA, TABELLARIA, DINOBRYON) . .195, 196, 197 STUDIES IN WATER SUPPLY CHAPTER I SOURCES OF WATER SUPPLY THERE are three main sources of water supply, namely : (1) Upland Gathering Grounds. In most cases the water is stored either naturally in lakes or lochs or artificially by damming across the outlet from a valley. Such waters are commonly derived from the drainage of unpolluted or sparsely populated land, and may require little or no purification before distribution. Birmingham, Leeds, Wakefield, Edinburgh, Cardiff, Manchester, and Liverpool fire examples of towns supplied with water from upland sources. Fifty-three samples of water collected from taps in these towns were found to contain B. coli (lactose + indol + ) in only four cases, even when as much as 100 cubic centimetres of water were submitted to such cultural tests as will afterwards be described. (2) Underground Sources of Supply. Shallow wells are presumably quite safe, or highly dangerous, according to their environment. Deep wells are usually satisfactory sources of supply, especially if the boring has been carried through an B - STUDIES IN WATER SUPPLY CHAP. impermeable bed, before reaching the water-bearing strata. The Kent deep well waters are good examples of very pure well water. In a period of seven consecutive years, out of 1565 samples examined, only 5*7 per cent, were found to contain B. coli (lactose + indol + ), even when as much as 100 cubic centimetres of water were tested. Springs are often very pure, but here again everything depends on their geological source and local surroundings. (3) Rivers. Increase of population and of trade, and the progressive growth of towns in the neighbourhood of rivers, have resulted in this source of supply being regarded with some disfavour, and, when a choice exists, frequently abandoned in favour of those previously mentioned. There can be no question that rivers are liable in a special way to contaminating influences ; and long before they reach the sea, nearly all rivers are quite unfit for domestic use in their unpurified state. Nevertheless, London, the largest and most important city in the world, derives the great bulk of its supply from the Thames and Lee, 1 both of which rivers drain populous areas and are admittedly sewage-polluted. Despite this fact, the general health of the Metropolis has for long been a source of pride to Londoners, and almost of wonder to the world. This is true, not only as regards sickness and death from all causes, but also as regards the incidence of those diseases which are liable to be water-borne. In support of this statement, the following quotations may be made from the Registrar-General's Annual Summary for Births, Deaths, and Causes of Deaths in England and Wales for 1905, 1906, 1907, 1908, 1909, 1910 and 1911 :- 1 A.bout 60 per cent, from the Thames and about 20 per cent, from the Lee. i SOURCES OF WATER SUPPLY 3 1905. Enteric fever accounted (in London) 1 for 234 deaths, and pyrexia (of uncertain origin) for 3. Thus to the continued fevers in the aggregate (referred to in the returns as " fever ") 237 deaths are attributed. The 234 deaths from enteric fever were equal to a rate of 0'05 per 1,000, or 0'07 less than the decennial average rate. The highest death rates from enteric fever were 0'09 in Finsbury and 0'12 in Hackney. (Page xviii.) 1906. Enteric fever (in London) caused 260 deaths, and pyrexia (of uncertain origin) 4, while no death during the year was referred to typhus. Thus to the continued fevers in the aggregate (referred to in the returns as "fever") 264 deaths were attributed. The 260 deaths from enteric fever were equal to a rate of 0'06 per 1,000, or 0'05 less than the decennial average rate. The highest death rates from enteric ever were O'll in Bethnal Green and 0'14 in Finsbury. (Page xxix.) In the year 1907 enteric fever caused 194 deaths and pyrexia (of uncertain origin) 4, while no deaths were referred to typhus. Thus to the continued fevers in the aggregate (referred to in the returns as " fever") 198 deaths were attributed. The 194 deaths from enteric fever were equal to a rate of 0'04 per 1,000, or 0'03 less than the average in the preceding five years. The highest death rates from enteric fever were 0'07 in Borough of Stepney, 0'09 in Hackney, and 0*10 in the City of London. (Page xxxiii.) In the year 1908 enteric fever caused 225 deaths, and pyrexia (of uncertain origin) 2, while no deaths were referred to typhus, Thus to the continued fevers in the aggregate (designated "fever" in the returns) 227 deaths were attributed. The 225 deaths from enteric fever were equal to a rate of 0'05 per 1,000 as compared with 0'06, the average rate in the preceding five years. The highest death rates from enteric fever were 0*09 in Holborn, in Shoreditch, and in Bermondsey, O'lO in Bethnal Green, and 0'12 in Finsbury. (Page xl.) In the year 1909 the deaths from enteric fever of persons belonging to London numbered 146, and those from pyrexia (of uncertain origin) 2, while no deaths were referred to typhus. Thus to the continued fevers in the aggregate (designated "fever" in the returns) 148 deaths were attributed. The 146 deaths from enteric fever were equal to a rate of 0'03 per 1,000 as compared with 0'05, the average rate in the preceding five years. In St. Marylebone and Lewisham the death rate from enteric fever was only O'Ol per J,000. The highest death rates were 0'05 in Hampstead, Shoreditch, and Deptford, 0'06 in the City of London, and 0'08 in Poplar. (Pages xxxi-xxxii.) Enteric fever. In the year 1910 the deaths from enteric fever of persons belonging to London numbered 196, and those from pyrexia (of uncertain origin) 4, while no deaths were referred to typhus. Thus to the continued fevers in the aggregate 200 deaths were attributed. The 196 deaths from enteric fever were equal to a rate of 0'04 per 1,000, which is O'Ol per 1,000 below the average rate in the preceding five years. In Fulham and in Woolwich, the death rate from enteric fever was only O'Ol 1 London as dealt with by the Registrar-General is not co-terrninous with " Water London." B 2 4 STUDIES IN WATER SUPPLY CHAP. per 1,000. The highest death rates were 0'07 in Shoreditch, 0'08 in Poplar, 0'09 in Holborn, 0'12 in the City of London, and 0'14 in Bethnal Green. (Pages xxxi-xxxii.) Enteric fever. In the year 1911 the deaths from enteric fever of persons belonging to London numbered 144, corresponding to a rate of 0'03 per 1.000, which is O'Ol per 1,000 below the average rate in the preceding five years. No death from enteric fever belonged to Chelsea, to the City of London or to Woolwich, and in Shoreditch the death rate from enteric fever was only O'Ol per 1,000. The highest death rates were 0'05 in Fulham, 0'06 in Hammersmith and in Holborn, and 0'09 in Firisbury and in Poplar. (Page xxv.) It thus becomes a matter of supreme interest to inquire minutely into the quality of the sources of London's Water Supply, the methods actually in operation for purifying the water, and those in process of being adopted, together with those which are likely to receive considera- tion in the immediate future. Such a survey ought to throw useful light on the general question of water purification and the best means to be adopted to render waters, initially of doubtful purity,. " safe " for domestic use. THE RIVERS THAMES AND LEE 1 AS SOURCES OF WATER SUPPLY. The chief arguments against the use of these rivers as sources of water supply are based on : (1) Physical, topographical, and epidemiological grounds. (2) The results of chemical and bacteriological analyses. As regards (l), against the known fact that many pollutions of undesirable sort exist on these watersheds must be set the circumstance that most of the con- taminating matters do not reach the rivers concerned without first undergoing some form of natural or artificial purification process. .In regard to the second point, which will chiefly be 1 The New River is really the Upper Lee water mixed with about 50 per cent, of deep well water. SOURCES OF WATER SUPPLY 5 dealt with here, a great amount of work has been carried out since 1905 at the Laboratories of the Metropolitan Water Board. Chemical Results. Table I. shows the chief average chemical results of the Rivers Thames and Lee for the aggre- gate of six years 1906-1911, and also for each separate year. Tables II. and III., pp. 6, 7, show the average results for each month during the six year period, 1906-1911. The accompanying diagrams, Figs. 1-7, serve to illus- trate the facts set forth in the tables. TABLE I. CHEMICAL RESULTS (PARTS PER 100,000, UNLESS OTHERWISE STATED). Year. -g of|~ 00 ^ G " H 2? c o I III gj-a J 1 .2 a -a '8 j 2 j ^|| if 3 49 a M - II 11 1 ||| |S H s.-s 3* g 6 3,s! H*S 1 ll (1906 0-0073 0-0159 0-22 1-68 0-1615 2-90 56 22-77 * on 1907 0-0069 0-0165 024 1-74 0-2158 3-29 72 22-39 5-67f ? g 1908 0-0063 0-0134 0-27 1-70 0-1734 2-30 63 23-88 6-26 * " 1909 0-0072 0-0159 0-23 1-75' 0-2198 3-05 67 23-46 6-60 *H 1910 0-0071 0-0173 0-24 1-69 0-2363 3-56 82 23-28 6-06 [1911 0-0081 0-0150 0-25 1-78 0-1752 2-49 61 2244 5-44 Averages 00072 0-0157 0-24 172 02039 293 67 2305 6 09 1906-11 1907-11) (1908-11) . f!906 0-0124 0-0170 0-26 2-05 0-1621 3-02 54 21-97 * $ 1907 0-0116 0-0157 027 2-02 0-1762 3-17 63 25-51 687f " 1908 0-0109 0-0146 0-29 2-04 0-1833 3-77 69 27-85 7-44 1909 00123 0-0172 0-27 2-18 0-2172 349 68 27-69 7-89 2 1910 0-0121 0-0175 0-26 2-03 0-2252 4-07 90 26-69 6-96 J911 0-0114 0-0151 029 2-14 0-1879 3-07 71 25-68 6-76 Averages 0-0118 0-0162 0-27 208 0-1977 3-43 69 26-42 7-27 1906-11 1907-11) (1908-11) * Not estimated during 1906. t July to December only. J Anthony's turbidimeter was used. Definite mixtures of saccharated carbonate of iron (B.P.) were made, and the results read off on the scale. Plotted out these gave a curve and figures for all intermediate readings. The method is fully described in the last edition of Thresh's treatise on "The Examination of Waters and Water Supplies," pp. 277-279. Burgess's apparatus was used: see Analyst, Vol. XXVII. , No. 319, October, 1902. TABLE II. AVERAGE RESULTS OF THE CHEMICAL EXAMINATION 'OF THE HAW THAMES RIVER WATER FOR EACH MONTH OF THE Six- YEAR PERIOD, 1906-1911. Parts per 100,000 unless otherwise stated. "3 73 1. D 2 * i i'S r ^ jg ' 3 Is g g . ;i a^ 3 1 li $ "S o i 2 g SE 1 53 73 coo "03 I| * a '3 fi-g O'fi &S | 3 1 a' J2 * S H jj jj < Slits H-2-fll "O 2 3 o^-S January . . 0-0216 0-0248 0-37 2-18 0-2920 11-74 153 27-77 _ February 0-0197 0-0172 0-39 2-07 0-1851 4-97 99 28-79 March .... 0-0088 0-0133 0-33 2-04 0-1128 1-35 40 28-68 April . May 0-0038 0-0053 0-0170 0-0160 0-27 0-20 1-91 1-85 0-1024 0-1087 1-46 1-33 30 26 26-99 23-93 1906 June .... 0-0062 0-0163 0-18 1-89 0-1253 1-15 27 23-56 July .... 0-0099 0-0170 0-16 2-01 0-1645 I'll 34 22-32 August . . 0-0086 0-0153 0-12 2-12 0-1342 0-99 32 20-35 September 0-0108 0*0146 0-16 2-26 0-1209 0-80 29 19-95 October .. 0-0120 0-0135 0-22 2-17 0-1270 1-07 31 23-76 November 0-0215 0-0201 0-38 2-11 0-2637 5-59 75 27-55 December 00210 0-0187 0-39 1'99 0-2075 4-69 73 28-30 January . . 0-0238 0-0176 0-41 1-85 0-2467 5-82 78 27-70 February 0-0200 0-0157 0-40 1-88 0-1910 5-20 77 29-82 March .... 0-0070 0-0135 0-33 1-91 0-1195 0-91 41 28-38 April . 0-0096 0-0174 0-26 2-00 0-1599 1-72 53 26-52 May 0-0093 0-0215 0-24 1-90 0-2198 2-46 61 25-50 1907 June .... 0-0082 0-0191 0-23 1-88 0-1680 1-58 40 24-02 July .... 0-0082 0-0145 0-20 1-89 0-1318 0-69 38 22-26 5-10 August . . 0-0056 0-0123 0-17 2-07 0-1115 0-21 33 22-02 6-14 September 0-0049 0-0111 0-15 2-15 0-1029 0-16 32 22-72 5-93 October .. 0-0095 0-0116 0-19 2-24 0-1279 0-88 38 23-40 6-40 November 0-0141 00138 0-28 2-34 0-1881 4-03 77 25-82 6-70 December 0-0214 0-0223 0-34 217 0-3SS(J 17-34 218 28-62 937 January . . 0-0226 0-0169 0-37 2-13 0-2742 10-82 85 30-02 8-85 February 0-0148 0-0136 0-35 0-1752 4-64 69 32-43 8-50 March .... 0-0121 0-0176 0-35 2-16 0-2694 9-15 128 30-05 9-43 April May 0-0087 0'0097 0-0189 G'0202 0-30 0'29 2-00 1'99 0-2765 G'2637 3-24 3*63 123 106 27-69 27-75 8-47 8*85 1908 June .... 0-0060 0-0185 0-21 1-80 0-1591 2-29 44 24-32 5-90 July .... 0-0094 0-0129 0-22 1-87 0-1550 1-80 45 24-95 6-20 August . . 0-0061 0-0110 0'19 1-97 0-1156 0-90 38 25-10 6-28 September 0-0064 0-0108 0-22 2-02 0-1179 1-52 42 25-55 6-22 October .. 0-0081 0-0116 0'27 2-05 0-1200 1-66 41 27-60 6-25 November 0-0101 0-0106 0-33 2-13 0-1203 1-30 38 29-40 6-48 December 0-0172 0-0130 0-37 225 0-1838 3-62 74 30-75 7-37 January . . 0-0157 0-0116 0-35 2-23 0-1501 2-41 57 31-50 9-15 February 0-0135 0-0097 036 2-19 0-1133 1-91 50 30-52 9-02 March .... 0183 0-0191 0-37 2-28 0-2679 5-71 115 28-46 9-58 April .... 0-0079 0-0206 0-23 2-17 0-1891 2-81 55 27-02 7'97 May 0-0051 0-0183 0-19 2-13 0-1563 1-83 42 25-50 6-65 1909 June 0-0122 0-0169 0'22 2-08 0-1780 1-15 46 26-38 7-22 July .... 0-0057 0-0168 0-19 2-11 0-1948 1-12 47 25-80 7'10 August . . 0-0046 0-0156 0-17 2-18 0-1744 1-09 42 24-92 6-26 September 0-0040 0-0134 0-23 2-23 0-1261 1-10 35 24-90 6-80 October .. 0-0087 0-0217 0-26 2-29 0-3332 6-95 86 27-02 7-27 November 0-0276 0-0167 0-31 2-14 0-2472 3-78 77 30-80 8-24 December 0-0238 0-0275 0-31 2-12 0-4868 12-84 171 29-50 9-50 January . . 0-0229 0-0176 0-32 2-18 0-2713 5-47 105 29-82 8-16 February 0-0207 0-0259 0-34 2-19 0-4049 12-21 187 27-00 8-90 March .... 0-0157 0-0173 0-32 2 -OS 0-2234 4-65 103 29-02 7-92 April 0-0079 0-0138 0-29 1-89 0-1549 2-32 55 27-22 7-00 May 0-0102 0-0169 0-24 1-92 0-1721 1-84 55 27-36 6-62 1910 June 0-0085 0-0169 0-20 1-85 0-1690 1-82 56 25-45 5-97 July .... 0-0067 0-0146 0-20 1-92 0-1736 1-00 47 25-43 5'48 August . . 0-0047 0-0121 019 1-95 0-1091 0-70 35 24-67 5-57 September 0-0043 0-0122 0-22 2-02 0-1111 0-77 36 2562 5-70 October .. 0-0079 0-0134 0-26 2-02 0-1178 1-10 41 26-32 5-10 November 0-0180 00208 0-27 2-16 0-2985 6-66 133 25-02 7-57 December 0-0205 0-0301 0-32 2-14 0-5211 11-46 231 26-52 9-70 Jan nary . . 0-0222 0-0153 0-35 2-10 0-2262 4-65 97 30*00 S'96 February 0-0158 0-0108 0-35 1-99 0-1332 2-78 61 29-45 7-67 March .... 0-0141 0-0184 0-28 2-17 0-2732 5-43 116 28-92 8-92 April .... 00069 0-0160 0-30 2-06 0-1786 2-46 63 26-75 6-80 May 0-0075 0-0234 0-27 2-03 0-2281 3-95 54 25-86 6-24 1911 June .... 0-0076 0-0168 0-25 1-91 0-1517 1-55 40 25-67 5-47 July .... 0-0055 0-0151 0-20 1-98 0-1295 0-79 37 22-96 4-98 August . . 0-0066 0-0126 0-16 2-07 0-1194 0-80 37 21-52 5-15 September 0-0053 0-0104 0-25 2-15 0-1031 0-97 39 20-90 4-25 October .. 0-0096 o-ono 0-26 2-18 0-1389 0-09 59 24-08 4-88 November 0-0163 0-0128 0-38 2-47 0-2061 0-64 83 27-62 8-20 December 0-0188 0-0192 0-47 2-53 0-3888 0-47 182 24-37 10-70 During 1906, this test was carried out at room temperature, 7 STUDIES IN WATER SUPPLY CHAP. 906| 1907 I 1908 | \ 9 9 I O 1911 RIVE1R THAME1S 220 LELEL 2- foo-t FIG. SOURCES OF WATER SUPPLY 1906 1907 908 | 1909 | 1910 | 1 9 < 1 ~| RIVE1R. THA/ALS A i , i jl i i ' j \ 1 \ u I \ 1 | l \ I . 1 1 ! \ ! I \ f roa THE: ALBU/nihOlp MITRQGEh U? PE.C 1Q 1 1 1906 | 1907 | 19Q& 1910 IO STUDIES IN WATER SUPPLY CHAP. 230 220 210 200 J90 180 170 160 150 140 130 120 110 100 90 80 70 60 50 no 30 20 10 S 1906 | t 9 7 I 1908 | 19 "o~9 1 9 o 6 1 9 O ~7 19O8 [1909 191O 1911 FIG. 3. SOURCES OF WATER SUPPLY n J/0 300 250 230 270 260 250 240 i.'UO 190 ISC /7C ICO 150 13 12 it 10 R1VE1K THAMLS 906 '99 FIG. 4. 12 STUDIES IN WATER SUPPLY CHAP. 340 310 320 310 MO 290 280 270 &>0 zso 240 230 220 210 200 WO 180 170 IfiO ISO 140 13,0 120 HO | "T FIG. 5. SOURCES OF WATER SUPPLY 906 | 1907 | 1908 | 1909 | i9'Q| t 9 i i ~| R.IVELR. THAAACS FOR THE! 72. MONTHS ELM? .? 51 - PE.C 9 06 1907 1 9 Q 6 | 1 9 Q 9 1' 9 1 O j 1 9 1 1 FIG. 6. i 4 STUDIES IN WATER SUPPLY CHAP. 240 230 220 210 200 190 >80 170 160 150 140 130 120 110 100 90 80 70 F,O 50 40 30 20 10 5 , 9 6 I 1 307 1908 1 9 O 9 1 9 1 FOR. THE. Y MONTHS EMpE/p 3lsT X>E.C 1 Q 1 1 [1906 |1907 | 19O& ) 1909 1 1910 | 'I9M ] FIG. 7. i SOURCES OF WATER SUPPLY 15 Standards are of variable significance, but if, for purposes of illustration, it be considered that a water ought to be objected to if it yields respectively 0*01 and O'l part per 100,000 of albuminoid nitrogen and oxygen absorbed from permanganate, it is obvious that, judged by these standards, the Rivers Thames and Lee are necessarily unfit for domestic use without previous purification. Bacteriological Results. Tables IV.-VL, pp. 15, 16, show the average bacteriological results of the Rivers Thames and Lee for the six years 1906-1911 combined, and for each separate year as a whole. Tables VII. and VIII. , pp. 18, 22, show the average bacteriological results for each month during the six-year period 1906-1911. The accompanying diagrams (Figs. 8-14) serve to illus- trate the facts set forth in the tables. TABLE IV. BACTERIOLOGICAL RESULTS. RIVER THAMES RAW WATER AT HAMPTON. Average Number of Microbes per c.c. B. coli Test (typical B. coli). (Percentage results. ) Year. *dd * .Us . -& Q + + + + + + + 0> i-- -^o^ > w ^^ 8)Q 100 10 1 o-i o-oi 0-001 o-oooi jn bes per c.c. B. coli Test (typical B. coli). (Percentage results.) * 3 45 84 20 52 3 d II ^ 4 3'2 0-4 0-8 0-8 100 c.c. /o 10 c.c. /o 1 c.c. 7 0-1 c.c. 8 o-oi c.c. 9 0-001 c.c. 10 0-4 12 2-8 6-2 4-5 0-4 0-0001 c.c. 11 1-2 o-o 0-4 ti i 3709 7362 8144 16318 35125 11757 13581 2 5 6 12 825 89-8 94-3 94-2 892 86'7 89-2 432 849 468 512 591 1-2 0-8 0-8 2-1 0-8 09 13-1 8-9 4-8 5-7 7'8 12-4 35-7 46-1 36-7 32-1 36-4 40-2 38-9 34-8 39-9 40-4 36-4 38-6 38-1 7'5 7-7 13-7 14-5 11-5 7-5 55 8-9 38-1 10-2 2'4 0-3 1909 average R. Lee, Col. 1, excludes two samples containing 1,120,000, and 1,240,000 microbes per c.c. respectively. 1909, B. Coli test :-also 0'5% in O'OOOOl c.c. and 0'5% in O'OOOOOl c.c. 1906-11, B. Coli test : 0'07% ,, ,, 0'07% 1910 average R. Lee, Col. 1, excludes two samples containing 1,300,000 and 1,400,000 microbes per c.c. respectively. TABLE VI. B. ENTERITIDIS SPOROGENES TEST. Raw Thames Water. Raw Lee Water. Year. (Jan. -Dec.) 6 d 6 6 d d d i g d d d d d d d II o r-H i i ? ' ~, ^ 7 1 O 1 + + + : + + + + + 1 Average, 1906... 57-9 377 3-9 0-4 64-7 30-1 4-7 0-4 1907... 62-9 27-5 8-4 1-1 62-9 30-5 5-9 0-6 1908.. 71-6 26-4 1-8 ! 54-7 35-8 9-4 __ 1909... 69-2 26-9 3-8 i 67-2 26-9 5-8 1910... 51-9 42-3 5-8 i 55-8 36-5 7-7 1911... 65'4 32-7 19 : 67 3 30-8 o-o 1-9 ,, 1906-11 61-4 33-1 4-9 0-5 1 63-0 31-0 54 05 SOURCES OF WATER SUPPLY 17 [ 1 9 6 | 1 9 O 7 1 9 O 8 1 9 9 | '91 | 1 9 t 1 | 'fa 140 120 -WO - 40 | 190-7 | 1QOS | 19QQ | 1910 | 1 9 Ti | RIVE.R BACTERIA RIVER LLL 1907 I 19Q8 t 90/9 1 9 I O I 19 -120 -100 -80 GO 40 20 10 i -i a o -/ | igoa I FKJ. 8. TABLE VII. BACTERIOLOGICAL RESULTS. RIVER THAMES HAW WATER AT HAMPTON. Average Number of Microbes per c.c. B. coli test (Typical B. coH). (Percentage results.) 3 d . s* ^n* I*. + + + + + + + |7 2oj%> hrj OI^O "c ' 100 10 1 o-i o-oi o-ooi o-oooi oPm i5 M AYS AT 20'-22C I 1909 I -19 FIG. 9. c 2 STUDIES IN WATER SUPPLY CHAP. 2400i 23001 2200\ 2lOO\ 21 J900\ I&OQ 1700 I 1600 1SOO\ Q8~1 1909 | | 1911 | RIVER TMAAVL5 F MICROBELS PER CC FOR THEL J"5 MONTHS MI>E.p 1911 G f\ R *T 37 G 1300 12QO\ )100\ W00\ O00\ 800\ 70 0\ 66 soot 4orA 30O\ zoo\ tool so\ |19Q7| 190& |190 FIG. 10. SOURCES OF WATER SUPPLY 21 |f907| 9 t t | RIV^R THAML5 CKQ13E1S CO FQRTHEL A GAR 19 O 6 I 1909 |-I910 FIG. 11. TABLE VIII. BACTERIOLOGICAL RESULTS. RIVER LEE RAW WATER AT PONDERS END. Average Number of B. coli test (typical B. coli). Microbes per c.c. (Percentage results.) o V + + + Ifel "8 lid 'i 3 100 10 1 o-i o-oi 0001 o-oooi ps li c.c. c.c. c.c. c.c. c.o. % c.c. c c. Cols. 1 2 3 4 5 6 7 8 9 10 11 12 Jan. ... 5192 4'5 13-6 72-8 9-1 95-5 Feb. ... 3083 5-0 30-0 60-0 5-0 95-0 March 1308 4-5 54-5 36-4 4-5 95-4 April... 471 31-6 42-1 26-3 68-4 May ... 1350 4-3 o-o 21-7 43-5 30-4 _ 73-9 June ... 598 20-0 55-0 15-0 10-0 800 1906 July ... 1190 9-1 o-o 31-8 31-8 22-7 4-5 59-0 Aug. ... 1986 18-2 45 13-6 45-4 18-2 636 Sept. ... 1499 5-0 5-0 15-0 40-0 25-0 10-0 75-0 Oct. ... 2397 . 4 '3 4-3 43-5 30-4 17'4 91-3 Nov. ... 9986 _ 9-1 72-7 13-6 4'5 99-9 Dec. ... 17792 5-9 17-6 58-8 17-6 94-0 Jan. ... 28986 31-8 54-5 13'6 99-9 Feb. ... 19115 5-0 o-o 30'0 50-0 15-0 95-0 March 3001 10-5 63-1. 21-1 52 89-4 April... 1861 61'9 33-3 4-7 99-9 May ... 1594 19-0 33-3 33-3 14-3 809 June ... 1095 20-0 45-0 35-0 800 1907 July ... 854 17 '4 52-1 30-4 82-5 Aug. ... 812 351 43 5-0 5-0 15-0 50-0 25-0 75-0 Sept. ... 1453 400 36 4-7 71-5 14-3 47 4-7 95-2 Oct. ... 2593 430 49 17-4 52-2 30-4 82-6 Nov. ... 15614 373 25 42-8 38-1 14-3 4-7 99-9 Dec. ... 11725 595 45 12-5 56-2 25-0 6-2 99-9 Jan. ... 22452 515 26 27-3 59-1 9-1 4'5 100-0 Feb. ... 14070 102 14 5-0 35-0 40-0 15-0 o-o 5-0 95-0 March 7304 347 34 36-3 45-4 13-6 4-5 99-8 April... 8999 291 16 36-8 52-6 10-5 99-9 May ... 4638 407 18 9'5 47-6 33-3 4-7 o-o 4-7 90-3 Junj ... 2333 205 34 10-0 60-0 25-0 5-0 90-0 1908 July ... 6032 509 76 4-5 o-o 18-2 45-4 27-2 4-5 95-3 Aug. .. 4895 891 223 10-0 40-0 35-0 10-0 o-o 5-0 90-0 Sept. ... 4682 257 36 4-5 40-9 31-8 13-6 9-1 95-4 Oct. .. 3780 597 42 4-5 40-9 31-8 22-7 95-4 Nov. ... 1706 287 36 4'7 4-7 38-1 38-1 14-3 90-5 Dec. ... 18589 867 46 11-7 17-6 41-2 17'6 11-7 88-1 Jan. ... 3361 299 19 21-4 64'3 14-3 _ 100-0 Feb. ... 6280 294 25 31-2 43-7 18-7 6-2 99-8 March 38469 1228 73 5-5 16-6 33-3 27-8 16-6 94-3 April.. May .. 2071 1735 174 205 22 10 7-1 21-4 28-5 50-0 64-3 28-5 92-8 78-5 iqno June.. 1759 372 39 23-5 41-2 23-5 11-7 76-4 9 July . 3565 312 50 6-2 25-0 31-2 31'2 6'2 93-6 Aug. .. 4852 642 91 50-0 31-2 12-5 6'2 _ 99-9 Sept. .. 1679 465 69 61-1 27-7 11-1 99-9 Oct. .. 17033 2276 215 5-9 29-4 29-4 17-6 5-9 o-o 94-1 Nov. ... 22954 1115 115 17-6 47'1 17-6 17-6 99-9 Dec. .. 105750 2539 250 12-5 68'7 6-2 12-5 99-9 Jan. .. 73878 544 31 15-0 30-0 40-0 15-0 100-0 Feb. .. 166131 420 20 20'0 65-0 o-o 10-0 5-0 100-0 March 72637 188 5 5-3 15-8 31-5 31-5 5-3 10-5 78-8 April.. 10676 389 7 4'7 9-5 57-1 23-8 4-7 85-6 May ... 2317 139 2 12-5 o-o 12-5 50-0 25-0 75-0 iqio June.. 1672 345 15 4-5 9-1 36-4 40-9 9-1 86-4 July .. 4258 752 38 15-8 36-8 31-6 15-8 84-2 Aug. .. 2886 275 39 4-5 50-0 36-4 4-5 4-5 95-4 Sept. .. 4248 310 16 9-1 13-6 45-4 31-8 77-2 Oct. ...I 3735 255 13 9-5 52-4 38-1 90-5 Nov. ... 42677 1228 24 4-5 36-4 27-3 31-8 95-5 Dec. ... 52322 942 28 55-5 27-8 16-6 99-9 Jan. .. 12619 360 15 9-5 19-0 47-6 23-8 90-4 Feb. ... 11490 152 5 15-0 35-0 40-0 5-0 5-0 85-0 March 24648 670 23 56-5 30-4 13-0 99-9 April... 6100 376 15 12-5 56-3 31-2 87-5 May ... 13959 871 55 9-1 59-1 18-2 13-6 90-9 iqil June... 5912 367 48 25-0 56-2 IS -7 749 July ... 5855 406 95 28-6 42-8 28-6 71-4 Aug. .. ! 4476 315 41 4-8 23-8 23-8 47-6 71-4 Sept.... 3647 555 86 4-7 28-6 61-9 4-7 | 95-2 Oct. ... 13454 492 59 13'6 31-8 50-0 4-6 86-4 Nov. ... 9754 405 72 9-1 50-0 36-3 4-5 90-8 Dec. ... 28906 1157 111 6-2 o-o 25-0 50-0 18-7 93-7 Oct., 1909, B. coli test : also 5'9% in O'OOOOl and 5 -9% in O'OOOOOl c.c. Dec., 1909, Column 1 : Average excludes two samples containing 1,120,000 and 1,240,000 microbes per c.c. respectively. 22 CH. I SOURCES OF WATER SUPPLY 22000 21000 ZOOOC 19000\ I8000\ 170 I6C 15000\ )4000\ m JZOOO\ JIL JOOOJ\ 3000- 2000 1000\ sool <7* | (906| 1907 | 1 9 O 8 1 t9Q9~| |^9O6 I 1907 I 1 9 O a I 1909 I 1910 FIG. 12. 24* STUDIES IN WATER SUPPLY CHAP. [t9Q7| 1908 I 1909 | 19 10 1911 2400 2300 2200 2100 Z000 mo 1800 1700 1600 1500 1400 7300 1200 uoo woo 900 800 700 600 500 400\ 300\ 200 1 100\ 50\ LE1E1 AVERAGE- AM C ROB ELS PELR c.c FOR. THE: *5"3 /MONTHS GAR [1907! 1908 | 1 9 -1 O | 191-\ FIG. 13. SOURCES OF WATER SUPPLY 25 909 [ > < I 19M | AVERAGE. riUMBLf\ I OF PER CC FOR THE. <5"d nON FIG. 14. 26 STUDIES IN WATER SUPPLY CH. i In the face of the foregoing chemical and bacteriological results, corroborating and measuring the known, or suspected, facts as regards sewage and other undesirable pollutions, no man of science could support the continued use of the Eivers Thames and Lee as sources of supply, in the absence of: (a) Further information tending to free the Rivers Thames and Lee from the full gravity of the charge of their being sewage -polluted rivers. (6) Evidence that the purification processes in actual operation were uniformly efficient. As regard (a), this will form the text of the next chapter, and in respect to (b), nearly the whole of this Monograph deals with the remarkable transformation of an impure raw river water actually into water of good quality, and not only so, but, when considered in relation to its source of origin, into a water of remarkable purity. CHAPTER II RESEARCHES TENDING TO JUSTIFY RIVERS AS SOURCES OF WATER SUPPLY IT will be convenient to deal in the first place with the streptococcus test and then with tests for definitely pathogenic bacteria (e.g. Gartner's bacillus and the typhoid bacillus). In 1898-9 and in subsequent years the writer reported to the Local Government Board on the significance of streptococci in water. It is of this test that primary mention may be made. Streptococcus Test. The chief reasons for considering the streptococcus test of value in the examination of water supplies, are : (l) That streptococci are superabundant in human faeces. (2) That faecal streptococci are absent or non- discoverable in a relatively large volume of pure water. (3) That fsecal streptococci do not multiply in pure water. (4) That some faecal streptococci are of feeble vitality and that the presence of such streptococci in a water, if they could be differentiated from their more robust companions, would seem to indicate pollution of recent and therefore specially dangerous sort. It is desirable to consider the subject under three headings : (1) What are the results, as regards streptococci, of the 27 28 STUDIES IN WATER SUPPLY CHAP. examination of a mixture representative of the faeces of a large number of individuals ? To solve this point and in corroboration and amplifica- tion of previous work, a visit was made to a sewage-works where the lumps of faeces reach the outfall in a fairly fresh, i.e. unbroken, condition. One hundred stabs with 100 sterile iron wires (the aggregate weight of which had previously been determined) were made into 100 lumps of faeces. The lumps of faeces had previously been consecu- tively fished out of the sewage by the aid of a wire gauze scoop. The contaminated wires were placed in a sterile test-tube of known weight. The tube plus its contents of contaminated wires was next weighed, and by a simple calculation the weight of the adhering faecal matter ascertained. Sterile water was then added to form an emulsion of faeces and water, in such proportion as rendered 1 c.c. of the mixture equivalent to O'l gramme of faeces. The wires were next rotated and moved about in the liquid so as to detach and break up the faeces. One cubic centimetre of the faecal emulsion was then added to a second tube (2) containing 9 c.c. of sterile water. 1 c.c. of (2) was then added to a third tube (3), also containing 9 c.c. of sterile water. In the same way a fourth tube (4) received 1 c.c. from tube (3), a fifth tube (5), 1 c.c. from tube (4), a sixth tube (6), 1 c.c. from tube (5), a seventh tube (7), 1 c.c. from tube (6), and so on until 10 tubes had been seeded. The next step in searching for streptococci, was to take O'l c.c. from each of the various dilutions and spread it over a series of previously prepared Drigalski and Conradi 1 1 Composition of Drigalski and Conradi's medium : Agar, 3% ; nutrose and peptone, each 1% ; sodium chloride, 0'5/ f ; lactose, 1*5% ; 1 c.c. of 0'1% solution of crystal violet (Hochst) ; 13 c.c. of litmus solution (Kiibel and Tiemann) ; 0'2 c.c. of a 10% solution of sodium carbonate in excess of amount required to render medium slightly alkaline. Beef broth (1 Ib. beef per litre of water), 100 c.c. ii TESTS OF RIVER WATER 29 plates by means of a sterile glass spreader working always from the lesser to the greater amount of faeces. After incubation at 37C., the minute colonies develop- ing on the plates were next subcultured into a lactose medium (lemco 1%, peptone 1%, sodium carbonate 0*1%, lactose 0*5%, tap water tinted with litmus solution up to 100). Only those tubes showing acidity, after two days' incubation at 37 C., were examined microscopically, and 100 of those yielding satisfactory morphological evidence were studied further, with results as follows : One streptococcus was isolated from one ten-millionth of a gramme of the faeces. Twenty-seven streptococci were isolated from one- millionth of a gramme. Seventy-two streptococci were isolated from one hundred-thousandth of a gramme. None of the streptococci reduced nitrates to nitrites l - at all events to any appreciable extent. Nor, of course, did they produce any gas in the media given below. The "types" of streptococci may be classified as follows : la = acid in a lactose medium. 2 ma = acid in a mannite medium, mi = clot in a milk medium. 3 ra = acid in a mffinose medium, sac = acid in a saccharose medium, sal = acid in a salicm medium . 1 The nitrate broth was of the following composition: Beef broth 10% (1 Ib. of beef per litre), potassium nitrate 0'1%, tap water up to 100. In testing for nitrites, a solution of 0'5 gramme of metaphenylenediamine in 100 c.c. of a 1% solution of hydrochloric acid is used. 2 The lactose, mannite, raffinose, saccharose and salicin media contained 1% of the respective fermentable substances, together with lemco 1%, peptone 1%, sodium carbonate 0*1%, and tap water tinted with litmus solution up to 100. 3 The milk medium was ordinary sterilised milk. 30 STUDIES IN WATER SUPPLY CHAP. The combinations were as under : lamirasal ... 49 lamirasacsal 44 lamamirasacsal 4 lamirasac 1 lamira 2 Total 100 Avoiding technicalities, the main point is that strepto- cocci were found in great abundance in the faecal mixture representative of the faeces of a large number of individuals. (2) What are the results as regards streptococci, of the examination of multiple samples of raw river water ? Drigalski and Conradi plates were inoculated with 1 c.c. amounts of the raw Thames, Lee and New Kiver waters, and all the minute colonies were subcultured and made the subject of attentive study The summarised results were as follows : 1908 (764 + 896 + 248) subcultures were made from 1 c.c. amounts of 156 (52 + 52 + 52) samples of raw river water (Thames, Lee and New Biver). Twenty-eight (13 + 13 + 2) of these samples contained streptococci (lactose + ), and the total number of streptococci isolated was 71 (19 + 50 + 2). It should be noted, however, that in three instances (Thames 2 samples, Lee 1 sample), the 1 c.c. plates were so crowded that it was necessary to make the sub- cultures from O'l c.c. plates. In these three samples one streptococcus was isolated from O'l c.c. ( = 10 per c.c.) of one of the Thames samples, and none from O'l c.c. ( = less than 10 per c.c.) of the other sample. Two streptococci were isolated from O'l c.c. ( = 20 per c.c.) of the Lee sample. It may therefore be desirable ii TESTS OF RIVER WATER 31 to amend the 71 figure by adding to it, either 104-0 I- 20 = 30, or 10 + 10 + 20 = 40. If we adopt the latter figure, then 71+40 = 111 streptococci. It is of interest to compare this figure with the aggregate number of bacteria (gelatine and agar) in the raw waters, and also the aggregate number of microbes growing in bile-salt agar (chiefly B. coli and excremental bacteria). The aggregate number of bacteria and streptococci in 156 c.c. of water derived in equal amounts from 156 comparable samples of raw river water (Thames, Lee and New River) collected at weekly intervals during the year 1909 was as follows : A. B. a D. Number of Bacteria. Number of Bile-salt-agar Agar at Gelatine at Streptococci. at 37 C C. 37C. 20-22 C C. Ill 6,650 59,645 2,199,190 The ratio figure is 1 streptococcus to 59 under B., 537 under <7., and 19,812 under D. (3) What are the results, as regards streptococci, of the examination of raw river water (a) in its normal condition and (b) after being purposely inoculated artificially with minute traces of human faeces (one part of fseces in one million parts of water) ? Ten samples of raw Thames water were examined, and 400 subcultures were made under (a) non-infected condi- tions, and a like number under (6) infected conditions. In the former case 36, and in the latter 204 streptococci were isolated, the preponderance under in- fected conditions being 168. The results would have been relatively still more striking had it not been for the fact that the particular samples of raw water used in the experiments happened to be exceptionally impure. The following points may be noted specially in connec- tion with these investigations : 32 STUDIES IN WATER SUPPLY CHAP. (1) The superabundance of streptococci in human faeces. (2) The relative difficulty of isolating any streptococci of a comparable sort from raw river water as exemplified by the Thames, Lee and New Kiver. (3) The ease with which streptococci can be isolated from raw river water, artificially inoculated with human faeces in the proportion of one part of faeces in 1,000,000 parts of water. As a corollary it surely follows that the wholesale con- demnation of raw river water, comparable with the Thames, Lee and New River, does not seem urgently warrantable as the result of the application of the streptococcus test. The writer has made many experiments on the vitality of streptococci in water, and has found that whereas some streptococci die very speedily, others persist for a long time, and he has so far been unable satisfactorily to associate this variability as regards vitality, with diiferences in the biological characters of the streptococci. Tests for pathogenic bacteria. The value of the streptococcus and B. coli tests, rests primarily on the assumption that these bacteria are present in abundance in human faeces, and, relatively speaking, absent from substances not exposed to excremental pollution. Hence their numerical estimation in water, is an indirect gauge of the degree of probability of their being accompanied, habitually or occasionally, by other microbes of truly pathogenic import. Until comparatively recently the search for pathogenic microbes in water was regarded as almost labour lost, for the reason that the practical difficulties attending their isolation were so great that failure to isolate them afforded no satisfactory indication of their real absence. Of late years, however, improvements in technique and methods have altered the whole complexion of affairs, and it is now possible, although still a difficult and laborious ii TESTS OF RIVER WATER 33 task, to isolate the typhoid bacillus when present in only small numbers, from a water, even when that water is swarming with other bacteria. How far this statement o is true may be gathered from the following account of the search for pathogenic bacteria in London's raw sources of water supply. In the first investigation 294 experiments in eight series were made with 156 samples of raw river water during the twelve months ended July 31st, 1908. The total number of bacteria (gelatine at 20-22C.) in the 29,400 c.c. examined was in the aggregate 135,687,500. Owing to the temperature of incubation, the composition of the media employed, and the fact of their appearing on the particular plate cultures, as coloured colonies, the great majority of these bacteria were excluded from con- sideration. The actual number subcultured and subse- quently more minutely studied was 7,329. Not one of them proved to be the typhoid bacillus. Later the inquiry was extended so as to include the search for Gartner's bacillus, and it differed from the previous research in a most important particular. Each sample of raw river water examined was divided into two equal portions of 500 c.c. (A and B). The A sample was inoculated with a very small number of typhoid bacilli and Gartner's bacilli, separately determined by Agar plate cultures in the usual manner. The B sample was not so infected, and was, therefore, normal raw river water. A and B were then examined in a strictly comparable manner. The object of this procedure was to show that the search for pathogenic bacteria in raw river water is not, as has been suggested, as hopeless as looking for a needle in a haystack. On the contrary, the results, on the average, showed that in the case of the A sample (artificially infected) the test was delicate to the extent of isolating the typhoid bacillus and Gartner bacillus, when only I was D 34 STUDIES IN WATER SUPPLY CHAP. present per about 7 c.c. of water in the first case, and per about 18 c.c. of water in the second case. It follows that failure to isolate these same microbes from the B samples (non-infected) under comparable conditions of experiment, although a negative result, acquired a positive significance. Twenty-four experiments were carried out altogether, and the results are summarised in Table IX. : TABLE IX. Total number of Colonies Sub-cultured. Typhoid part of Experiment. Gartner part of Experiment. Average number of artificially added Typhoid bacilli per c.c. of the raw river water. Average number of Typhoid bacilli recovered from : Average number of artificially added Gartner bacilli per c.c. of the raw river water. Average number of Gartner bacilli recovered from : Infected Sample A. Non- infected Sample B. Infected Sample A. Non- infected Sample B. Infected Sample A. Non- infected Sample B. Infected Sample A. Non- infected Sample B. 5451 2-242 None 14-54 14-54 None (? 1 out of 5,451 sub- cultures) 0-686 None 12-417 12-417 None (? 1 out of 5,451 sub- cultures) 2-242 = lto 6-485 0-686 = 1 to 18-10 One typhoid-like microbe and one Gartner-like microbe were isolated; excepting these two, none of the 5,451 microbes studied, bore any reasonable resemblance either to the typhoid bacillus or Gartner's bacillus. Practically the experiments showed that the typhoid bacillus cannot be uniformly present in 7 c.c. nor Gartner's bacillus in 18 c.c. of raw river water. In a still later investigation exactly the same methods were adopted, 35 additional experiments having been carried out. The results are summarised in Table X. (p. 3'5). One typhoid-like microbe 1 was isolated ; with this 1 This microbe was motile, multi-flagellated, and fulfilled all the ordinary tests for the typhoid bacillus. It was agglutinated with an anti-typhoid serum in practically the same dilutions as a stock laboratory culture of B. typhosus. It was, however, decidedly less virulent to rodents, than is usual with strains of typhoid bacilli isolated from cases of typhoid fever. II TESTS OF RIVER WATER 35 exception, none of the 7,991 microbes studied bore any reasonable resemblance either to the typhoid bacillus or Gartner's bacillus. In this series of experiments the average number of typhoid bacilli added to the A (infected) samples was 0'653 per c.c. of river water, and the average number recovered was 11*6, so that, by inference, it may be con- cluded that the test was delicate to the extent of detecting TABLE X. Total number of Colonies Sub-cultured. Typhoid part of Experiment. Gartner part of Experiment. Average number of artificially added Typhoid bacilli per c.c. of the raw 'river water. Average number of Typhoid bacilli recovered from : Average number of artificially added Gartner bacilli per c.c. of the raw river water. Average number of Gartner bacilli recovered from : Infected Sample A. Non- infected Sample B. Infected Sample A 11-6 11-6 0'653 1 to 177 Non- infected Sample B. Infected Sample A. Non- infected Sample B. None Infected Sample A. 7-86 7-86 0-633 1 to 12-4 Non- infected Sample B. None 7991 0-653 None. None (? 1 out of 7,991 sub- cultures) 0-633 1 typhoid bacillus per 17*7 c.c. of river water. The average number of Gartner's bacilli added was 0*633, and the average number recovered was 7*86, so that, by inference, it may be concluded that the test was delicate to the extent of detecting 1 Gartner's bacillus per 12*4 c.c. of river water. Practically, therefore, the experiments showed that the typhoid bacillus cannot be uniformly present in 17'7 c.c. or the Gartner's bacillus in 12*4 c.c. of river water. 1 1 The above results represent the practical, but not the absolute, limits of delicacy of the method. For instance, when the quality of the river water was good, more than 500 c.c. could have been centrifugalised ; when it was bad, more primary plates could have been used. In either case the number of subcultures could have been increased. On the other hand, the work involved in " picking off" 250 colonies from the plates, as was done in these experiments, is so considerable, that it would scarcely be practicable, greatly to increase this number of subcultures. D 2 36 STUDIES IN WATER SUPPLY CHAP. Taking all the foregoing researches together, the study of 7,329 + 5,451 + 7,991 = 20,771 specially selected colonies derived from 156 + 24 + 35 = 215 separate, raw river water samples, has resulted in the discovery of only two typhoid- like microbes. One Gartner-like microbe was found ; the number of separate samples examined having been 24 + 35 = 59, and the number of colonies studied 5,451 + 7,991 = 13,442. It may perhaps be objected that the bacilli experi- mentally added were " cultivated " specimens (i.e., grown on artificial media in the laboratory), and that similar results would not have been obtained if "uncultivated" bacilli (as present in the urine of a typhoid "carrier") had been used. This objection, however, falls to the ground ; for special experiments were subsequently made, in which the urine of a typhoid " carrier " case was added experimentally to river water, and the typhoid bacillus was successfully isolated from the mixture. In one experiment, 1 part of typhoid urine was added to 50 million parts of Thames river water, the typhoid bacillus being present in the proportion of 152 per litre of water. Yet the typhoid bacillus was recovered without difficulty from the artificially infected water. The writer's latest experiments have been with sewage, and it has been found that with samples, artificially infected, it is possible to isolate the typhoid bacillus when present in the proportion of one typhoid bacillus per O'OOl c.c. of sewage. No typhoid bacilli having been found iu duplicate non-infected samples, the conclusion is surely legitimate that the typhoid bacillus cannot be present in sewage, in the proportion of one per 0*001 c.c., though it may of course be present in larger quantities of sewage. Now, it is easy to show by separate bacterial tests (e.g., the bile-salt agar test) that the river Thames, on the average, is about 20,000 times less impure than sewage, and hence infer en tially may it not be safely ii TESTS OF RIVER WATER 37 concluded that there are usually no typhoid bacilli in 20 c.c. of Thames river water ? This is a novel, if indirect, method of gauging the quality, in relation to water-borne disease, of any river water. For instance, the number of bacteria per c.c. of the river water may be estimated by means of the bile- salt agar test, and the figure obtained used to divide the number of bacteria (bile-salt agar) in 0*001 c.c. of sewage. The product gives the number of c.c. of river water which, on the basis of the writer's experiments, cannot be expected to contain the typhoid bacillus. The conclusions which may be drawn from the foregoing observations seem, to be : (1) A watershed may be exposed to manifold pollutions, all more or less objectionable. . (2) The river draining the watershed may be impure as judged by the ordinary chemical and bacteriological tests. (3) The river water may nevertheless be shown to con- tain none or scarcely any of the microbes of water-borne disease, when tested by methods of proven value. Epidemiologists have for long marvelled at the com- parative, and oft-times long continued, immunity from typhoid fever, enjoyed by towns drinking sand-filtered water, from a polluted river source. The explanation usually forthcoming is the perfection of the filtration process, but may not the true explanation really be that the river water is never primarily so thoroughly noxious that it does not subsequently become relatively innocuous, 98 percent, of its total bacteria having been removed by sand-filtration ? These views, although tending to free rivers from the full share of blame hitherto attached to them, are not inconsistent with the writer's continued and continual advocacy of a source of supply, wherever possible, of the greatest initial purity. 38 STUDIES IN WATER SUPPLY CH. n Having now learnt a little of the inherent and adven- titious quality of river water, as illustrated by the Thames and Lee, it seems next desirable to consider the purification processes which seem absolutely necessary or at least desirable, before such a water is rendered reasonably safe for domestic use. CHAPTER III THE QUESTION OF ABSTRACTION The Question of Abstraction. In the case of upland gravitation schemes of supply, storm water courses are frequently provided so that when a flood occurs the worst water is prevented from entering the reservoirs, and flows instead along the water-courses on one or both sides of the reservoirs and so escapes down the valley. Even with a pumping scheme, assuming adequate storage capacity, a great deal can be done to safeguard the purity of the supply by closing the " intakes " during the worst period of each flood. It seems to be insufficiently realised that there is no necessary or absolute parallelism between the flow of a river and its current quality. At the commencement of a flood the results, as regards quality, are apt to be dispro- portionately bad, but the quality of the water may improve more quickly than the rate of subsidence of the flood. Further, the first autumn or early winter flood may produce a much worse water than floods occurring after the bed of the river has been well scoured. It follows that, given adequate storage accommodation, it is unnecessary and most undesirable to leave the intakes open during the worst period of a flood should there be a prospect of an oversufficiency of water still available for abstraction of far better quality at a not much later period. 40 STUDIES IN WATER SUPPLY CHAP. Of course, in the case of many water-works, owing to deficiency in storage accommodation and to there not always being sufficient water available in the river, it is impossible to place the abstraction question on a strictly scientific basis. Nevertheless, although it may be practi- cally impossible for these and other reasons to approach the ideal as regards abstraction, it is always well to aim at as high a standard as is reasonably practicable. In illustration of different methods of abstraction, the Kiver Thames colour-results for the 72 months ended December 31st, 1911, have been chosen as an example. The consumption of water has been taken arbitrarily as 100 million gallons a month, and questions of insufficency of water due to drought have been ignored. In the first diagram (Fig. 15, p. 41) it is assumed that there is no storage and that therefore the water must be abstracted uniformly at the rate of 100 million gallons per month. The remarkable fluctuations in the colour 1 are apparent, varying indeed from a maximum of 239 in December, 1907, to a minimum of 29 in April and May, 1906, the average for the whole period being 67*4. In such a case, not only would the average colour of the water to be dealt with by the filters be fairly distinct, but during certain months the colour would be so marked (even after filtration) that the water on delivery would be appreciably brown in colour. In the next two diagrams (Figs. 16-1 7) a large reserve of storage has been assumed, but the matter has been considered independently of the known effect of storage in reducing colour, and also of the circumstance, that it may really be better to store a worse water for a longer period than a better water for a shorter period. 1 The colour results have been chosen for purposes of illustration, as this is a matter which appeals at once to the sense of sight, and is commonly, although erroneously, used by the consumer as a gauge of probable quality. Burgess's apparatus was used in determining the colour of the water, in THE QUESTION OF ABSTRACTION 41 z* 136 210 20L 190 170 160 150 130 120 HO 100 SO 70 60 50 3D 20 10 5 I 9 O 6 I 1907 O 8 19O9 1Q1O 19 RAW COLOUR, ~^ B *a | 1 9 o | 1907 | 1906 | 19091 1910 119 '11 I AVERAGE. /MONTHLY COLOUR RESULTS TOR. 72 MONTHS 31 ST PE-C 1 9 11 FIG. 15. Average monthly colour results (colour x volume abstracted -f 100) for the 72 months ended December 31, 1911. Abstraction uniform at the rate of 100 million gallons per month. No storage. Average colour result for the whole period, 67 '4. Here the abstraction is constant and the colour variable. 42 STUDIES IN WATER SUPPLY CHAP. In the second diagram (Fig. 16, p. 43) the abstraction has been varied in proportion to the colour, the net result being that the colour has been kept constant at 52*6 (colour x volume abstracted -*- 100) and the abstraction varied from as much as 181'67 million gallons in April and May, 1906, to as little as 22*04 in December, 1907. To attain this end an immense reserve of water would be necessary, and it has been calculated, that starting with storage reservoirs capable of holding 611 million gallons, and containing 250 million gallons to begin with, the amount of water in store at the end of the period would be the -same, namely 250 million gallons. In October, 1906, the reservoirs would have been full and in April, 1911, they would have contained 133*91 million gallons. Here, apart from the effect of storage in reducing colour, the actual colour of the water abstracted would have remained constant at 52*3. In the next diagram (Fig. 17, p. 45) the storage has been taken as 343*7 million gallons and whenever the reservoirs became depleted and the water coincidently better than the average, the opportunity has been taken to fill them on the following basis : A water with a colour of 29 has been taken as the best water ever likely to be obtainable and in order to arrive at the permissible volume of water for filling the reservoirs, the storage capacity has been multiplied by 29 and the answer divided by the current colour. Thus with a colour of 29, 343*7 million gallons could be taken, or as much as was needed to fill the reservoirs. But with a colour of 57, only 174*87 could be taken. In cases, however, when the river water was worse than the average (67*4), the abstraction figure has been proportional to the impurity, the balance of water required for filtration purposes being borrowed from the storage reservoirs. For example, with a colour of 74, 71*19 million gallons would be abstracted, but with a colour of 150 only 35*12, in THE QUESTION OF ABSTRACTION 43 1906 I 19Q7 I 1908 | t9O9 I 19'0 I 1911 THA/AL5 190 180 170 160 150 140 130 120 110 100 90 SO 10 GO 52 30 20 10 f\ RSTRACTlOh IM AMLLlOM GALLONS PER | 1906 | 1907 [ 19Q6 | 1909 | 1910 | 1911 I REL5UL.T5 FOR 2 f*\OriTH5 EHpElp 3 1 ^ DE-C R ') 9 1 1 FIG. 16. Average monthly abstraction figures in million gallons for the 72 months ended December 31, 1911, the average for the whole period being at the rate of 100 million gallons per month. Colour constant at 52'6 (colour x volume abstracted -f'100), Here the colour is constant and the abstraction variable. Storage, 610*79 million gallons. 44 STUDIES IN WATER SUPPLY CHAP. It will be seen (see Fig. 1 7) that this leads to an average colour of 46 (colour x volume abstracted *- 100) ; the reservoirs would be full during 34 out of the 72 months, the smallest amount being 27 '90 (about 8 \ days' supply) in May, 1908, and the volume left in store at the close of the period being 244 million gallons. Of course, in the case of London it would be impossible to abstract on the lines here suggested (merely for illustrative purposes), as the volume of water required is about 244 million gallons 1 per day, whereas in the foregoing example the consumption has been taken at only 100 million gallons per month; and at the very times when it would be desirable to fill up the reservoirs owing to the purity of the river there may be no water available for abstraction purposes. Neverthe- less, there can be no question that, without risking a shortage of water, knowledge of the current quality of river water, on the average and the probable variations from that average, from time to time should assist in determining the abstraction of water to the best advan- tage from the point of view of purity. On the other hand, it must be remembered that water is now known really to improve so much under storage, that it might be contended to be better to keep the reser- voirs full even when the river water was of unsatisfactory quality, and so utilise storage to its fullest extent, rather than close the intakes and deplete the reservoirs with consequent curtailment of the number of days the water should be stored after the flood had subsided and river water was again admitted into the reservoirs. Although we know the average reduction of colour effected during storage for certain known periods, it is obviously difficult, if not impossible, to say for each individual sample of river water what its condition will be after it has traversed a reservoir. 1 A certain proportion (about 20%) of this, however, is of non-river water origin. in THE QUESTION OF ABSTRACTION 45 9 O8 | 1909 I I 9 I 1911 | RIVE1R THA/ACS 31*- VC.CR 1911 I 1906 | 1 9 O ~f \ -1906 1909 I FIG. 17. 10 I 1911 / = aver jrage monthly abstraction figures in million gallons for the 72 months ended December 31, 1911, the average for the whole period being at the rate of 100 million gallons per month. = average monthly colour results (colour x volume abstracted -f 100). Here both the colour and abstraction vary (see text), the average for the former being 46 and for the latter at the rate of 100 million gallons per month. Storage, 343'7 million gallons. 4 6 STUDIES IN WATER SUPPLY CHAP. On comparing the best and worst raw river water results with the best and worst stored water results, although the percentage reduction in the two sets of cases is not widely dissimilar, the actual results are markedly different. For example, during the 12 months ended July, 1908, the best and worst river Thames colour results were 31 and 384 respectively. The comparable figures (i.e. best and worst) for the same water after storage in the Chelsea reservoirs were 18 and 200. Judging from the writer's general experience of storage it would probably be safe to conclude, at all events, that the worst samples of river water may still be so highly coloured even after storage that it is a matter of great aesthetic importance always to aim at excluding from supply as much flood water as is reasonably practicable. The matter, indeed, should be judged, not by the per- centage improvement in the two cases, but by the actual state of the water after storage, corresponding on the one hand to flood, and on the other to non-flood, conditions of the river water. It is of interest to note (Table XL) the number of days when the colour of the River Thames was 200 or over and 100 or over * during the six years 1906-1911. TABLE XI. RIVER THAMES. COLOUR RESULTS. NUMBER OF DAYS THE COLOUR WAS 100 AND OVER AND OVER 200. Number of days the colour was 100 and over. 200 and over. 1906 38 55 49 61 85 56 57 10 20 11 6 24 5 13 1907 1908 1909 1910 1911 Average (appro .) 1 Samples are not collected on Saturdays, Sundays or certain holidays. To overcome this difficulty, the figure for the blank days has been arrived at by taking the average of the preceding and succeeding samples actually examined. in THE QUESTION OF ABSTRACTION 47 The average colour (all results included) for the fore- going period (1906-1911) was 6575, but if we exclude all samples yielding 100 (or over) and 200 (or over) the amended results become 48 and 59 respectively. It is evident from the table that even in the worst year only 24 days' storage would be required to enable the intakes to be closed whenever the colour results were 200 or over, even if the unsatisfactory results occurred in sequence. On the other hand, to close the intakes, when the colour results were 100 or over would have called for 85 days' storage in the worst year, 38 in the best year, and 57 in an average year. It is impossible to lay down a fixed method of abstraction to suit all cases, but, for purposes of illus- tration, the months might be selected, during which the filtered water was noticeably free from colour, and the average colour of the raw water taken for the same period, using the latter figure as some sort of guide as to what volume should be abstracted during the rest of the year in order to obtain the best possible final results. It will be said by the epidemiologist that colour is only of secondary importance, and that the behaviour of the microbes of water-borne disease in water of varying degrees of impurity is the essential factor for judging the question of abstraction. There is much truth in this contention, but the physical quality of water is not only intrinsically of importance, but is the only means the public have of judging the quality of a water supply. The writer's experience has been that there is not a very wide difference between the death-rate of pathogenic bacteria artificially added to river water of varying degrees of impurity. These microbes, however, are more likely to be present in river water during floods than under more normal conditions and so it seems prudent 4 B STUDIES IN WATER SUPPLY CHAP. to exclude either partially or wholly, such flood water from storage reservoirs. It must also be remembered that it is easier (for technical reasons) to isolate artificially-added pathogenic bacteria from a pure water, than from an impure water, although the difference in the two cases tends to disappear with the lapse of time. It follows, therefore, that a negative result in the former case is apt to have a greater significance than in the latter ; and perhaps the critic might even go so far as to suggest that, in the case of very impure waters, the failure subsequently to isolate artificially-added pathogenic bacteria was some- times due, not to their real extinction, but to the technical difficulties involved in their isolation under these circumstances. On the whole, however, the writer is convinced that artificially-added typhoid bacilli die fairly rapidly in stored water even when such water was of great initial impurity, but this view is not inconsistent with belief in the desirability of avoiding impure flood water for storage purposes. If one might predicate from the actual quality of water as delivered to London consumers, using the B. coli test as an indicator, there can be no doubt that during the winter flood months of the year, the results tend to deteriorate to a considerable extent. This is illustrated in Fig. 18, and the writer is of opinion that the observed results are largely due to the inclusion of flood water in the storage reservoirs. In conclusion, it must be remembered that many water works unfortunately depend almost solely on heavy floods to replenish their depleted reservoirs, and in such cases it is obvious that questions of quality must be subordinated to questions of quantity. There remain, however, it is to be feared, not a few instances of waterworks, which might improve the quality in THE QUESTION OF ABSTRACTION 49 -fej -fcj - /w Jit ;o STUDIES IN WATER SUPPLY CH. m of their supply by an applied study of this important question of abstraction. In some cases where storage is absent or inadequate it is impossible to avoid the use of flood water, and in these circumstances some supplementary process of purification ought to be interposed between the river and the storage reservoirs (if of insufficient size) or the river and the filter beds where no storage exists, or alternatively, between the reservoirs and the filter beds. 1 1 Reference may be made to the writer's Third Research Report on Storage, Metropolitan Water Board. CHAPTER IV SUPPLEMENTARY PROCESSES OF WATER PURIFICATION. THE so-called mechanical or pressure filters worked either with or without a coagulant may be used with great advantage. This, for example, is the preliminary process carried on at York (River Ouse) where, for various reasons, it was decided to select this system in preference to storage reservoirs. According to the published returns, excellent results were obtained 1 and the writer has had sufficient experience of pressure filters to have faith in this method of treat- ment. Frequent tests of the efficiency of the York mechanical gravity filters during a period of three years (without the use of a coagulant) showed the following satisfactory reduction in the number of bacteria. Reduction of bacteria per cent. No. 1 filter 86-4 2 86-9 86-2 86-5 85-9 85-1 86- Quite recently Chester has decided not only to construct storage reservoirs, but also to interpose pressure filters between the River Dee and these reservoirs. 1 Presidential Address to Association of Water Engineers, 1910 "Mechanical Gravity Filters," by W. H. Humphreys, C.E., York. 51 E 2 STUDIES IN WATER SUPPLY CHAP. Alternatively, small pre-storage settlement reservoirs may be placed between the river and the large storage reservoir proper and they may be worked with or with- out a coagulant. The advantages of passing raw river water through small reservoirs antecedent to storage in large reservoirs is seen by the results of the following experiments carried out in connection with the New River and East London (Sunbury) supplies : New River : At the Hornsey Works there is a small storage reservoir having a capacity of about seven million gallons. During the period under observation about eleven million gallons were passing daily through this reservoir, the nominal storage thus being only fifteen to sixteen hours. The average results may be summarised as follows : TABLE XII. EXPERIMENT 1. HORNSEY (NEW RTVER) EXPERIMENTS. March 10th to April 25th ; May 2nd to May 24th ; and May 29th to June 10th, 1910. Average Parts per 100.000. number of Colour /VY^Wi Description of the Sample. Bacteria perc.c. Ammo- (Agar at niacal Albumi- Oxy. abs. '' noid fr.perman- Turbidity. (mm. brown 2-ft. tube). 37 C.) | Nitrogen. Nitrogen. ganate. Inlet (New River) water to Hornsey Reservoir. (Samples : 47 chemical and 52 bacteriological) 92-2 0-0030 0-0093 0-0750 1-33 34 Outlet water from i Hornsey Reservoir. (Samples : 47 chemical 1 and 52 bacteriological) 56-0 0-0024 0-0089 0-0708 1-04 33 Percentage Improve- ment ... 39'2 200 4-3 5-6 21-8 2*9 It is apparent that the continuous flow of the New Kiver water through a reservoir, holding even less than a day's supply, produced a considerable improvement in the quality of the water, especially as judged by the reduction in the number of bacteria, and by the ammoniacal nitrogen and turbidity tests. iv PROCESSES OF WATER PURIFICATION 53 TABLE XIII. EXPERIMENT 2. HORNSEY (NEW RIVER) EXPERIMENTS. October 13th to December 2nd, 1910. Average number of Parts per 100,000. Colour Description of the Sample. Bacteria per c.c. (Agar at Ammo- niacal Albumi- noid Oxy. abs. fr.perman- Turbidity. brown 2-ft. tube). 37 C.) Nitrogen. Nitrogen. ganate. Inlet (New River), Hornsey Reservoir. (Samples : 37 chemical and 37 bacteriological) 188 0-0038 0-0068 0-0718 2-17 47 Outlet water from Hornsey Reservoir. (Samples : 37 chemical and 37 bacteriological) 108 0-0036 0-0062 0-0631 1-55 39 Percentage Improve- ment ... 425 5'26 8-82 12-12 28'57 17-02 East London (Thames) Supply: Here the reservoir has a capacity of only five million gallons. During the period covered by the experiments, about eight million gallons were passed daily through it, the nominal storage thus being only about fifteen hours. TABLE XIV. EXPERIMENT 3. SUNBURY (THAMES) EXPERIMENTS. Bacteriological : March 22nd to April 15th ; and April 27th to May 10th, 1910. Chemical : April 7th to April 15th ; and April 27th to May 10th, 1910. Average number of Parts per 100,000. Colour / Description of Sample. Bacteria per c.c. (Agar at Ammo- niacal Albumi- noid Oxy. abs. fr. perman- | ^miu. i brown Turbidity. 2-ft. tube). 37 C.) Nitrogen. Nitrogen. ganate. Inlet (Thames) water to Sunbury Reservoir. (Samples: 18 chemical and 27 bacteriological) 155 0-0026 0-0130 0-1304 154 43 Outlet water from Sunbury Reservoir. (Samples : 18 chemical and 27 bacteriological) 111 0-0021 0-0126 0-1246 1-42 41 Percentage Improve- ment 284 19-2 3-08 4-45 7-8 4-65 The improved condition of the Thames water even after only a few hours' continuous flow settlement is well shown 54 STUDIES IN WATER SUPPLY CHAP. by the foregoing results, especially as regards number of bacteria and amount of ammoniacal nitrogen. The results given in experiments 1, 2, and 3, are illus- trated in Fig. 19 (p. 55). The advantages of using coagulants, antecedent to the storage of raw river water in large reservoirs may now be considered. Apart from the questions of cost and of altering the mineral composition of a water, practically any desired result can be attained by the use of certain coagulants, e.g. alumino-ferric, as the following experiment shows : TABLE XV. ALUMINO-FERRIC. 16 parts per 100,000 parts. 16 Ibs. per 10,000 gallons. 6d. per 10,000 gallons. Average of three experiments. Quiescent settlement. Sample. Bacteria per c.c. & S S) Percentage Improvement. Colour m.m. brown 2 ft. tube. Percentage Improvement. Perman- ganate, 3 hrs. at 80 F. parts per 100,000. Percentage Improvement. Turbidity (sacchar- ated carb. of Iron) Parts per 100,000. al fa G 03 > o o If Raw Thames water before treatment... 1390 217 0-4395 8-9 _ Raw Thames water after treatment... 40 97-1 31 85-7 0-1103 74-9 0-75 91-6 It is to be noted that not only are the results excellent on the basis of percentage improvement but also that the actual state of the " treated " water was remarkably good. Next, dealing with more practical doses of coagulant, the following results are of interest : Raw Thames water was passed through the East London Waterworks Reservoir at Sunbury at the rate of about eight million gallons per day. (Nominally about 15 hours' storage.) March 22nd to May 10th. From May 23rd to June 28th (both inclusive) raw Thames water was passed through the reservoir at a slower rate, viz., about 5| million gallons per day (about 21 '8 hours' storage). iv PROCESSES OF WATER PURIFICATION 55 Sl/APLE PER-CEHT/^GE: EIXPEIR I/AQ1TS FIG. 19 56 STUDIES IN WATER SUPPLY CHAP. PERIOD 1. From March 22nd to April 15th, 1910 (both inclusive), no precipitant was used. PERIOD 2. From April 16th to April 26th, 1910 (both inclusive), alumino- ferric l was used in the proportion of about 3 parts of alumino-ferric to 100,000 parts of water (3 Ibs. to 10,000 gallons). PERIOD 3. From April 27th to May 10th, 1910 (both inclusive), no precipitant was used. PERIOD 4. From May 23rd to May 27th, 1910 (both inclusive), no precipitant was used. PERIOD 5. From May 30th to June 14th, 1910 (both inclusive), alumino- ferric was used in the following proportions : May 30th to June 5th, about 2 parts per 100,000. June 6th to June 8th, about 2'4 parts per 100,000. June 9th to June 14th, about 2 '85 parts per 100,000. That is, about two or three Ibs. of alumino-ferric were added to every 10,000 gallons of water. PERIOD 6. From June 15th to June 28th (both inclusive), the amount of alumino-ferric was raised to about 5 '7 (or roughly double) parts per 100,000 (5 '7 Ibs. per 10,000 gallons of water). During the first period when no precipitant was used the number of bacteria was reduced by simple sedimentation from an average of 178 to 129 (27'6 per cent.). During the second period, when alumino-ferric was used (about three parts per 100,000), the percentage reduction was 63 '6 (162 reduced to 59 bacteria per c.c.). During the third period, when the precipitant was no longer in use, the percentage reduction fell to 31 (117 reduced to 81 bacteria per c.c.). During the fourth period, when the rate of flow was reduced 2 and still no precipitant was used, the percentage reduction was 21 '8 per cent. (78 reduced to 61 bacteria per c.c.). During the fifth period, when alumino-ferric was again used (about two three parts per 100,000), the percentage reduction was 49*8 (395 reduced to 198 bacteria per c.c.). During the sixth period, when the dose of precipitant was doubled, the percentage reduction was 70*1 (678 reduced to 203 bacteria per c.c.). The results are shown in Table XVI. (p. 57). 1 This is really a commercial form of sulphate of alumina, containing some iron. The sample used in these experiments contained 15*99 per cent. A1 2 O 3 (including the iron). 2 For some unexplained reason the results were less good, instead of being better, with the reduced rate of flow. iv PROCESSES OF WATER PURIFICATION 57 TABLE XVI. EAST LONDON (SUNBURY) BACTERIOLOGICAL RESULTS. Period and Description of Samples. Average number of Bacteria per c.c. Agar at 37 C. Percentage Reduction. PERIOD 1. No coagulant. Inlet Water. 17 samples Outlet Water. 17 samples 178 129 27-6 PERIOD 2. Coagulant 3 parts to 100,000 parts of water. Inlet Water. 7 samples... Outlet Water. 7 samples 162 59 63-6 PERIOD 3. No coagulant. Inlet Water. 10 samples Outlet Water. 10 samples 117 81 31 PERIOD 4. No coagulant. Inlet Water. 5 samples Outlet Water. 5 samples 78 61 21-8 PERIOD 5. From 2 to 2 '85 parts coagulant to 100,000 parts of water. Inlet Water. 16 samples Outlet Water. 1 6 samples 395 198 49-8 PERIOD 6. 5'7 parts coagulant to 100,000 parts of water. Inlet Water. 14 samples Outlet Water. 14 samples 678 203 70-1 There can be no question, therefore, that, bacterio- logically, there is considerable advantage to be gained by even less than twenty-four hours' continuous flow settlement, and that when alumino-ferric is added, still better results are obtained. The actual chemical results are shown in Table XVII. (p. 58). The percentage chemical and bacteriological improve- ment is shown in Table XVIII. (p. 59). Taking the chemical and bacteriological results together, there can be no doubt that the beneficial effects observed in connection with simple continuous flow settlement could be considerably enhanced by the use of coagulants. Unfortunately the experiments did not happen to have STUDIES IN WATER SUPPLY CHAP. been carried out during the flood months of the year. Had it been otherwise the results would in all probability have been much more striking from the point of view of percentage reduction, although doubtless it would have been necessary to increase the dose of coagulant. TABLE XVII. EAST LONDON (SUNBURY) CHEMICAL RESULTS. Parts per 100,000. Period and Description of Samples. Ammo- niacal Nitrogen. Albumi- noid Nitrogen. Oxygen absorbed from Perman- ganate. Turbidity. Colour. PERIOD 1. No coagulant. Inlet Water. 8 samples Outlet Water. 8 samples 0-0023 0-0020 0-0127 0-0118 0-1173 0-1054 164 1-41 39 37 PERIOD 2. Coagulant 3 parts to 100,000 parts of water. Inlet Water. 7 samples Outlet Water. 7 samples 0-0042 0-0033 0-0144 0-0119 0-1580 0-1193 1-96 1-33 45 36 PERIOD 3. No coagulant. Inlet Water. 10 samples Outlet Water. 10 samples . . . 0-0029 0-0023 0-0131 0-0133 0-1409 0-1399 1-46 1-42 46 44 PERIOD 4. No coagulant. Inlet Water. 5 samples Outlet Water. 5 samples 0-0055 0-0068 0-0175 0-0215 0-1651 0'1608 1-38 1-48 48 44 PERIOD 5. From 2 to 2*85 parts coagulant to 100,000 parts of water. Inlet Water. 16 samples Outlet Water. 16 samples 0-0083 0-0071 0-0211 0-0198 0-2315 0-1768 2-80 1-86 68 55 PERIOD 6. 5 '7 parts coagulant to 100,000 parts of water. Inlet Water. 14 samples Outlet Water . 1 4 samples . . . 0-0042 0-0058 0-0222 0-0216 0-3662 0-3032 2-77 1-73 78 64 As regards cost, the coagulant used in these experiments was aluinino-ferric (15'99 per cent, of A1 2 3 , including the iron), which cost 3 12s. 6d. per ton. Bought in larger quantities, no doubt the cost would be less, say 3 10s. Od. per ton. In the experiments, two to six parts per 100,000 parts were used, that is, two to six Ibs. per 10,000 gallons. iv PROCESSES OF WATER PURIFICATION 59 Each Ib. of the material cost (on the basis of 3 10s. Od. per ton) 1'5 farthings, so that the experiments cost three farthings (fd.) to twopence halfpenny (2jd.) per 10,000 gallons of water treated. As sand filtration is said to cost only Id. per 10,000 gallons, the use of a coagulant as a supplementary process would thus increase the total cost of purification to a TABLE XVIII. PERCENTAGE REDUCTION. Outlet as compared with Inlet water. Description. Ammo- niacal Nitrogen. Albumi- noid Nitrogen. Oxygen absorbed from perman- ganate. Turbidity. Colour. Per cent. Reduc- tion. No. of Bacteria (Agar at 37 C.) PERIOD 1 No pre- cipitant used 13-04 7-09 10-14 14-02 5-13 27-6 PERIOD 2. Alumino- ferric used, about 3 pts. per 100,000 ... 21-43 1736 24-49 32-14 20-0 63-6 PERIOD 3. No pre- cipitant used 20-69 Increase 0-71 2-74 4-35 31 PERIOD 4. No pre- cipitant used Increase Increase 2-60 Increase 8-33 21-8 PERIOD 5. Alumino- ferric used, about 2-3 pts. per 100,000 ... 14-46 6-16 23-63 33-57 19-12 49-8 PERIOD 6. Alumino- ferric used, nearly 6 pts. per 100,000 ... Increase 2-7 17-20 37'55 17-95 70-1 considerable extent. But for emergency purposes, a much higher cost than the above might be regarded as relatively unimportant. The question of dose is one of great diffi- culty. If too little of the coagulant is used, the beneficial effects may be practically nil, and therefore the process would be wasteful, or at all events, the improvement might be so slight as not to be commensurate with the cost involved. On the other hand, obviously any increase of the dose in excess of actual requirements is economically 60 STUDIES IN WATER SUPPLY CHAP. undesirable. On the whole, however, it is better to face the cost of a dose which is certain to be markedly effective, than by being too niggardly, to ran the risk of spending money without effecting any very tangible result. The effective dose varies not only with the amount of suspended matter in the water, but also with its character. Less than about three parts per 100,000 (3 Ibs. per 10,000 gallons) is usually ineffective with raw Thames water or not sufficiently effective to justify completely the expenditure involved. 1 With very bad waters a much larger dose may be required, but, apart from the question of cost, even the foulest water can be brought to almost any pitch of perfection by the use of enough of the coagulant. The advantages of pre-storage-settlement reservoirs, especially if so arranged as to permit of the use of coagulants would thus appear to be considerable. Al- though recommended chiefly in conjunction with storage proper, these small reservoirs would be of real value in those cases where river water is taken directly on to the filter beds, for the reason that during floods coagulants could be judiciously used to counteract the evil effects of the use of storm water for filtration purposes. The author has had no personal experience of the Puech-Chabal Multiple System of pre-filters, but remark- ably good results appear to have been obtained by their use. The principle involved is that of rapid filtration by decantation through a series of filters consisting of materials of diminishing grade (very coarse to very fine). It is claimed that complete clarification is obtained. A strong point in favour of this system is that the whole of the material in the filters is in full action and not merely 1 These remarks apply, of course, to waters about to be stored. In the case of waters about to be filtered, the dose may be considerably reduced, and yet good results obtained. Here, the coagulant and the filter particles com- bine to produce the best results. iv PROCESSES OF WATER PURIFICATION 61 the surface layers, as is solely or largely the case with most filters. Another supplementary process of water purification recently put forward by the writer is known as the "Excess Lime Method," but it seems desirable to consider this matter under the head of sterilisation processes. CHAPTER V STERILISATION PROCESSES WITH SPECIAL REFERENCE TO THE "EXCESS LIME" METHOD EARLY in 1905 the author l sterilised the Lincoln Water Supply (population above 50,000) by means of " Chloros." The dose varied according to circumstances, but was usually 1 to 100,000 (about 1 in 1 million in terms of available chlorine). The conditions were far from ideal, as the contaminated river water, which often contained much suspended matter and a good deal of organic matter, was run, practically speaking, directly on to the filter beds ; the consequence being that larger doses of chloros were required than are found to be necessary in general cases. The treatment was continued up to 1911, when a new supply was introduced. The best series of results occurred during a period of about ten months, when 62 samples collected consecutively contained no typical B. coli even in 100 c.c. of water. As regards dose (in terms of available chlorine), the usual limits are from 1 in 1 million to 1 in 5 millions. With a water containing a good deal of organic matter, and 1 In conjunction with Dr. McGowan, who controlled the chemical part of the investigation. For further information reference may be made to the Fifth Report of the Royal Commission on Sewage Disposal, Appendix IV. 2 Chloros is an alkaline solution of sodium hypochlorite containing about 10-15% of available chlorine. Bleaching powder (about 33% available chlorine) is equally suitable. CH. v STERILISATION PROCESSES 63 subject to considerable fluctuations in composition, it is not always an easy task to secure uniform sterilisation and invariable absence of objectionable taste. Well waters containing but little organic matter and of fairly constant composition lend themselves readily to treatment. The following is an example of the results that may be achieved with this class of water : A particular well in the series of experiments is situated in the Upper Chalk. The water has to be filtered, as, apart from its constant liability to contamination, it is occasionally visibly turbid. The chloros (dose 1 in 200,000 to 1 in 400,000) l was mixed with the water as it flowed on to the sand filters. The average chemical analysis of the well water, before treatment, was as follows : Parts per 100,000. Ammoniacal Nitrogen Q'0003 Albuminoid ,, 0'0022 Oxidised 0'23 Chlorine T80 Oxygen absorbed from Permanganate (3 hrs. 80F.) 0'0134 Total Hardness 28'47 Permanent Hardness 5'97 The bacteriological results, before and after treatment, were as under : During a period of one year's working, 102 samples of the water both before and after treatment (collected at approximately equal intervals) were examined bacterio- logically. Out of these 102 samples, 83 contained typical B. coli in 100 c.c. (or less) before treatment, whereas after treatment only 8 yielded positive results, and for more than Jive consecutive months the results were uniformly negative even in 100 c.c. The treatment of the Lincoln water early in 1905 has been followed since about 1907 by similar forms of 1 The chloros contained from 10-15% of available chlorine. 64 STUDIES IN WATER SUPPLY CHAP. treatment in the United States, and apparently with excellent results, although occasionally difficulty has been experienced, in obtaining successful bacteriological results, without imparting a slight mawkish taste to the treated water. It also appears to have been insufficiently realised that to obtain the best results with the smallest dose, time is required. The following extract from Hooker's treatise on " Chloride of Lime in Sanitation" (1913) shows that some at least of the American authorities are evidently not fully cognisant of what took place at Lincoln in 1905 : "The use of extremely minute quantities of chloride of lime has offered a very practical and simple solution of the sanitary troubles of nearly every city water supply. The most astonishing part of all this is that the true import of these facts has only been realised within the last four years. "As a practical process it dates from 1908, when Mr. G. A. Johnson, of New York City, was called in to remedy some serious trouble in the water purification at the Chicago Stock Yards. The filtered water of Bubbly Creek contains a large amount of sewage, and it had been purified by a process of filtration in conjunction with copper sulphate, but it was the complaint of the large stock shippers that animals drinking this filtered water made less gain in weight than when city water was supplied to them. Under pressure of a lawsuit brought by the City of Chicago against the Union Stock Yards Company, the contractors for the filter plant were, however, enabled to fulfil their guarantees by Mr. Johnson substituting chloride of lime for the copper sulphate. The treatment raised the quality of the sewage-laden water from the Creek far above that of the Chicago City water, as was shown in its low percentage of cases where B. coli were found. B. coli. found. Bubbly Creek, treated '34% of cases. Chicago City water 12'8% "The hypochlorite was added 7i hours before filtration; the addition after filtration did not give as satisfactory results. The amount of chloride of lime added was forty-five pounds per 1,000,000 gallons. "Thus a new epoch in the annals of water purification dates from Mr. Johnson's success at Chicago." In the United Kingdom, things have not gone altogether smoothly, there being among the public an almost in- vincible repugnance to the use of chemicals (except lime) in connection with water supply. v STERILISATION PROCESSES 65 It is easy to carry conservatism too far in these matters, and more than one case could be quoted where a more tolerant attitude would have saved the ratepayers large sums of money. There are, of course other methods of sterilising water, e.g., by heat, ozone, ultra-violet rays. Sterilisation by heat is the ideal method bacteriologically, but it is too costly a process to meet with general acceptance. Ozone can be successfully employed for the destruction of the microbes of water-borne disease, but it labours under some disadvantages. The capital cost involved is serious, the working cost is considerable, arid the process is not well adapted for waters containing much oxidisable and suspended matters. Sterilisation by means of ultra-violet rays has many advocates, but the feasibility of the method on a large scale has still to be demonstrated. It is convenient here to consider the " excess lime method" of sterilisation, described by the writer in his Eighth Research Report to the Metropolitan Water Board. THE EXCESS LIME METHOD In the ordinary lime softening process as commonly practised, quicklime or slaked lime is added to the total bulk of the water to be softened. It is always in amount purposely calculated to be rather less than is needed to combine with the bicarbonates present (i.e., just short of removing the whole of the temporary hardness). If on testing with phenolphthalein or silver nitrate solution, a pink or brown colour develops with these two tests respectively, this indicates an excess of lime, which is always to be carefully avoided. When, however, such an event happens, the dose of lime has either to be reduced at once or carbonic acid pumped into the water, so as to combine with the excess of free lime. F 66 STUDIES IN WATER SUPPLY CHAP. Apart from the fact that any excess of lime re-hardens the water, it is not desirable to give the consumer lime- water to drink. In short, in practice, a purely chemical standard being kept in view, the utmost care has, wittingly or unwittingly, been exercised so to arrange the process as to prevent any true bactericidal action taking place, by accurately treating the whole of the water to be softened .with less than the amount of lime required to combine with the bicarbonates, so as to form an insoluble and bacteriologically inert carbonate of lime. In the method now under consideration, part of the water is purposely overdosed with lime so as to bring about a known bactericidal effect, and then, after a suitable interval this is mixed with enough " untreated," i.e., unlimed, water to combine with the excess of lime. Anticipating what follows, it may be convenient at this stage to give an example : (1) 15 Ibs. of quicklime costing 1'5 pence (l^d.) added to 7,500 gallons of raw River Thames water (l in 5,000) would suffice to kill the B. coli present in the water within 5-24 hours. The liquid, being decidedly alkaline, must next be neutralised. (2) Not less that 2,500 gallons of adequately stored water (i.e., a bacteriologically " safe " water) * must subsequently be added for neutralisation purposes. (3) The resulting mixture (10,000 gallons) antecedent to final filtration would be epidemiologically safe, and its initial hardness would be reduced about 65 per cent. (4) Of course, any larger proportion than 25 per cent, of stored water could be used, the only difference being that the mixture of lime-treated and stored water would be " softened" to a correspondingly less extent. (5) The only essential points to remember are that 1 part of quicklime is needed to over-combine with the bicarbonates and to sterilise 5,000 parts of River Thames 1 For evidence of this, see Chapter VI. STERILISATION PROCESSES raw water, and that a sufficiency of stored water must be added subsequently to neutralise the excess of lime. Dose of Lime. As regards the dose of lime required for sterilisation purposes, this depends primarily on the degree of temporary hardness of the water, the degree of its impurity, and the duration of contact. Table XIX shows the dose necessary to sterilise Loch Katrine water, notoriously known to be an exceedingly "soft" and pure water. So as artificially to render it most impure, bacteriologically, 1 per cent, of crude sewage was purposely added to the water before lime treatment. TABLE XIX. STERILISATION or LOCH KATRINE WATER. L. Katrine water + =B. coli alive ; - =B. coli dead. ( + 1 per cent, sewage) CaO. Duration of contact. 1 hour. 5 hours. 24 hours. 1 A 1 in 5,000 ... - 10 and 1 c.c. - 10 and 1 c.c. - 100, 10 and 1 c.c. IB ... 55 55 55 2 A 1 in 10,000 ... + 10 and 1 c c. 55 J > 2B ,, ... 55 55 3 A 1 in 20,000 ... + 10-1 c.c. 3B ... - 10 and 1 c.c. 4 A 1 in 40,000 ... + 10 and 1 c.c. 4 B ,, ,, j 5 A 1 in 80,000 ... 5 + 100 and 10 -Ic.c. SB ,, 5 6 A 1 in 160,000 ... 4-100, 10 and 1 c.c. 6B ... , 55 55 It will be seen that a dose of from 1 to 10,000 to 1 to 20,000 was effective in 5 hours, and a dose of 1 to 40,000 in 24 hours. With hard waters, on the other hand, the dose of lime must of course be increased considerably, as the bicarbonates render the lime inert as a bactericidal agent. In practice, indeed, it is desirable to add even more lime than would seem to be indicated on theoretical grounds. With very soft waters the bactericidal dose is, as has been shown, very small, but the author has found with hard F 2 68 STUDIES IN WATER SUPPLY CHAP. waters (especially if impure) and with sewage, that to effect sterilisation the amount of active lime left in these liquids, after neutralising the substances which combine with the lime must be in excess of the apparent bactericidal dose. Table XX illustrates the germicidal action of quicklime (1 in 5,000) in the case of the hard raw Thames River water. TABLE XX. STERILISATION OF THAMES RIVER WATER. Duration of contact. Number of samples of raw Thames water. Number of tests made. Results as regards B. coli. + = alive. - =dead. Remarks. 1 hour 28 43 In 1 case -10 and 1 c.c. Tn 2 cases +10-1 c.c. In 40 cases + 10 and 1 c. c. B. coli alive in 42 out of 43 cases. 5 hours 20 33 In 3 cases + 10 and 1 c.c In 4 cases +10-1 c.c. In 26 cases - 10 and 1 c.c. B. coli dead in 26 out of 33 cases. 24 hours 28 43 In 1 case +10-1 c.c. In 42 cases - 10 and 1 c.c. B. coli dead in 42 out of 43 cases. It will be seen that in nearly all cases B. coli was killed either in 5 or in 24 hours. For the purpose of contrast and comparison, a further set of experiments with raw Thames water, purposely inoculated with 1 per cent, of crude sewage was also undertaken. The additional number of B. coli thus added to the river water was estimated to be about 1,000 per c.c. The results are shown below (Table XXL, p. 69). It is apparent that the effective dose was 1 in 5,000 (5 to 24 hours' contact), but that a dose of 1 in 7,500 was ineffective. Speaking generally, the "bactericidal dose with hard waters would seem to be rather less than 1 to 5,000 and with very soft waters rather more than 1 to 50,000. STERILISATION PROCESSES 69 Even so grossly impure a liquid as crude sewage can be successfully sterilised by means of lime. According to the author's experiments, London sewage requires a close varying from 1 to 1,500 to 1 to 2,500. Neutralisation of the Excess of Lime. It is obvious that the treated water requires further treatment to get TABLE XXI. STERILISATION OF ARTIFICIALLY POLLUTED THAMES RIVER WATER. Duration of contact. One part quicklime (about 75 per cent. CaO) in : 5,000 parts sewage-inoculated raw Thames water. A. 7,500 parts sewage-inoculated raw Thames water. B. 1 hour B. coli alive in 10 and 1 c.c. in all 10 experiments. B. coli alive in 10 and 1 c.c. in all 10 experiments. 5 hours B. coli alive in 10 and 1 c.c. in 2 experiments. B. coli alive in 10 but not in 1 c.c. in 4 experiments. B. coli dead in 10 and 1 c.c. in 4 experiments. B. coli alive in 10 and 1 c.c. in all the 10 experiments. 24 hours In 1 experiment B. coli alive in 10 but not in 1 c.c. ; in the remaining 9 experiments B. coli dead in 10 and 1 c.c. B. coli alive in 10 and 1 c.c. in 4 experiments. B. coli alive in 10 but not in 1 c.c. in the remaining 6 experi- ments. rid of the excess of lime. The procedure to be adopted may have to be varied according to the original quality of the water undergoing treatment. With very soft waters the excess of lime in the water, after sterilisation, is so small that it might be found best to treat the whole of the water (instead of only a part) and then to neutralise it with carbonic, sulphuric, or other suitable acid, or possibly it might suffice to filter the water through a percolating filter. With hard waters, it is easy to neutralise the excess of lime by the addition of " untreated " water. For example, 70 STUDIES IN WATER SUPPLY CHAP. in the case of the raw Biver Thames water (sterilised by adding 1 part of quicklime to 5,000 parts of water) the " treated" portion must receive not less than 25 per cent, of " untreated " water (stored water in this particular instance) to effect neutralisation. Treatment of the water used for neutralisation pur- poses. The question naturally arises as to the sterilisation of this water added for neutralisation purposes. According to the author's experience the answer is that stored (i.e., unfiltered) water could safely be used for this purpose. Alternatively, and so as to provide the neutralising water, active chlorine (hypochlorites) or ozone could be em- ployed. The usual objection to ozone is that of cost, but here only a part of the total supply would require ozonisation the total cost being reduced proportionately. The com- mon objection to "active chlorine" (hypochlorites) is the difficulty of sterilising a water containing a variable amount of oxidisable matter without giving the water a faint taste, and of knowing how best to get rid subsequently of the slight excess of chlorine. Here also, as only a proportion of the total supply would require chlorination, any question of taste would be practi- cally eliminated and any slight excess of chlorine would be absorbed by the organic matter in the " lime treated " water. The cost of the process varies according to the degree of temporary hardness of the water. Lime containing at least 75 per cent. CaO can be bought wholesale for about Id. per 10 Ibs., and to sterilise a water like London water, about 1 part would be required per 5,000 parts, or 10 Ibs. per 5,000 gallons, or 2,000 Ibs. (costing about 16s.) per million gallons. As, however, the lime sterilisation process is applicable only to a proportion of the whole (say three-quarters), the cost would be correspondingly less (say 12s.) The hypochlorite-sterilisation would be con- v STERILISATION PROCESSES 71 siderably cheaper (say Is. to 10s. per million gallons) than this, but, of course, the advocates of " water softening " might assert (rightly or wrongly) that the saving in soap and fuel consumption more than counter-balanced the apparent economy effected by the chlorine process. With very soft waters, the cost of lime would be greatly reduced, perhaps to as much as only one-fifth to one-tenth of this estimate. With very hard waters containing very little oxidisable matter the cost of sterilisation by means of hypochlorite is proportionately very small, but with very soft waters containing much oxidisable matter the lime process may turn out actually to be cheaper. This, however, is only the cost of the lime used. Account would have to be taken for tanks and usually for filtering the finally treated water together with the cost of preparation of the neutralising water. The peri- odical removal of the sludge from the tanks would also entail additional expense. These are matters which so obviously encroach on the domain of the water engineer that no apology is needed for merely mentioning them in passing. The circumstances under which the Excess Lime Pro- cess appears to be specially attractive are as follows : (1) A water, bacteriologically impure, which in any event it is proposed to soften. 1 (2) A supply hard, as well as impure, situated in a dis- trict badly circumstanced as regards alternative supplies. (3) A river supply contaminated (or liable to be con- taminated) having scarcely any available storage accommo- dation, and necessitating for waterworks purposes the occasional use of Hood water, with mere sand-filtration as a safeguard. (4) Wells not sufficiently pure to be used without some 1 As regards water supplies already undergoing softening treatment, it ought not to prove a very difficult or costly matter so to modify the existing plant as to give effect to the principles laid down in this chapter. 72 STUDIES IN WATER SUPPLY CHAP. form of treatment, and concerning which some doubt exists as to whether storage plus subsequent filtration would prove an altogether satisfactory remedy. 1 (5) Surface waters unduly " soft " and not free from the risk of undesirable pollution. The River Dee is a case in point, and it is of interest to note that the Medical Officer of Health and the Waterworks Engineer of Aberdeen were so impressed with the results of the author's experiments that they persuaded their Town Council to sanction the carrying out of an experimental trial on the whole of the Aberdeen water supply (f>f million gallons a day). The experiment lasted three weeks, the dose of CaO being 3 parts per 100,000. The results were so highly satis- factory bacteriologically that the Aberdeen Town Council has now decided on an annual expenditure of 1,270 for liming purposes. An excellent account (" Clarification of Water Supplies by the Excess Lime Method ") of the Aber- deen experiments by Dr. James Watt will be found in the Journal of State Medicine, August, 1913, No. 8, vol. xxi. (6) Mixed supplies, part being impure bacteriologically and the rest above suspicion ; the former to be lime- treated and the latter used for neutralisation purposes. Some of the advantages attending the use of lime instead of other sterilising agents may now be mentioned. One great advantage of the use of lime is that it is " hallowed by precedent " and that practically all drinking waters already contain lime salts. There is, indeed, no objection to the use of this substance either on medical grounds or on the score of taste, or for sentimental reasons. Personally, the writer sees no objection to the use, under skilled supervision, of hypochlorites, but this view is not shared by a great many persons. Further, in the case of hard waters (temporary hard- ness) it decreases markedly the " lime content " of the 1 Well water supplies for a variety of reasons are not always so amenable to purification by storage and filtration as are river water supplies. STERILISATION PROCESSES 73 finally treated water, thus rendering such waters much " softer," and in the case of very soft water, the treatment may with great advantage be used for an exactly opposite purpose, namely, to " harden " the supply. TABLE XXII. "EXCESS LIME" METHOD OF STERILISATION. (Parts per 100,000) = lbs per 10,000 gallons. To convert into grains per gallon and "degrees" of hardness, multiply by 7 and divide by 10. 3. 1. 2. Theoretical bactericidal Temporary hardness of the water as CaC0 3 Theoretical amount of CaO required to combine dose on the assumption that not less than 2 parts (soap test). with the bicarbonates.* of CaO remain uncombined in the treated water. * 22 12-32 14-32 21 11-76 13-76 20 11-20 13-20 19 10-64 12-64 18 10-08 12-08 17 9-52 11 52 16 8-96 10-96 15 8-40 10-40 14 7-84 9-84 13 7-28 9-28 12 6-72 8-72 11 6-16 8-16 10 5-60 7-60 9 5-04 7-04 8 448 6-48 7 3-92 5-92 6 3-36 5-36 5 2-80 4-80 4 2-24 4-24 3 1-68 3-68 2 1-12 3-12 1 0-56 2-56 * N.B. It will, however, be seen from the text that the figures here given may be found to be too low in practice. Table XXII may be of some little use in determining the amount of lime required for bactericidal purposes for waters of different degrees of hardness. The first column gives the temporary hardness (parts per 100,000) of the water ; the second column shows the theoretical amount of lime required to combine with the bicarbonates, and the last column gives the same figure plus 2 parts (per 100,000) 74 STUDIES IN WATER SUPPLY CHAP. of additional lime for bactericidal purposes. For various reasons this amount will^ be found to be insufficient in practice, especially in the case of hard waters, and about 0'25 part extra for each part of temporary hardness will probably be required. The following procedure is suggested : Determine the temporary hardness in the ordinary way. Inoculate the water with a trace of sewage (say 1 per cent.) to ensure the presence of large numbers of B. coli. To one bottle (A) add the amount of lime indicated in column 3 of the table. To four other bottles add 10, 20, 30, and 40 per cent, extra lime (B, C, D, E). Examine the samples (A, B, C, D, E) bacteriologically for B. coli after 1, 5, and 24 hours (1, 10, and 100 c.c. cultures). It will probably be found that after 1 hour all of the five samples (A, B, C, D, E) contain B. coli even in 1 c.c., but that in from 5-24 hours one or more of the samples may be sterile (1, 10, or even 100 c.c.) On the other hand, one or more of the samples will probably yield positive results even in 24 hours (100, 10 perhaps in 1 c.c.) The following alternative method may also be used : (l) add to the water a known amount of CaO, but considerably in excess of what is required to combine with the bicarbonates, C0 2 , etc. (2) After thorough mixture and settlement estimate the excess CaO in the liquid. (3) The difference between (1) and (2) is the amount of CaO required to combine with the bicarbonates, C0 2 , etc. , to which must be added 0*002 per cent, extra CaO as this is approximately the minimum bactericidal dose (24 hours), (4) Add amount indicated under (3) to the water previously inoculated with 1 per cent, sewage and make B. coli cultures after 1, 5, and 24 hours. (5) Repeat (4), but with slightly larger doses of CaO, until the exact bacteri- cidal dose has been determined, STERILISATION PROCESSES 75 Having determined the bactericidal dose (let us say that sample C gives the best indication of the bactericidal dose required) it next becomes necessary to find out the per- centage of " untreated " water which has to be added to it to neutralise the excess of lime. This is best done by mixing C (after decantation) and the untreated water in stoppered bottles in the following proportions : TABLE XXIII. NEUTRALISATION or EXCESS LIME. C. Treated water. Per cent. Untreated water. Per cent. C. Treated water. Per cent. Untreated water. Per cent. (*) 90 10 (/) 40 60 (6) 80 20 (g) 30 70 (c) 70 30 (h) 20 80 (d) 60 40 () 10 90 () 50 50 After a suitable interval (preferably several hours) the clear liquid (a) to (i) is decanted (or " siphoned " off) and tests made for hardness and free lime. The sample yielding the lowest results as regards hard- ness [say (d)] indicates the best proportions of untreated and treated water. This sample will usually be found to give a slight pink with phenol phthalein solution, but little or no appreciable brown colour with silver nitrate solution. On the other hand, (c), (b), (a) will be found to give an increasing pink colour with phenolphthlein, and may also show a brown colour with silver nitrate. As regards hardness, not only will (e), (f), (g), (h), (i) be found harder than (d), but (c) (b) (a) will also be harder owing to excess of CaO. There are several important, if fairly obvious, points to be borne in mind. The percentage of CaO in the sample of lime must be determined, and allowance made for its not being of full 7 6 STUDIES IN WATER SUPPLY CHAP. strength. For example, if it is of 75 per cent, strength an extra 25 per cent, must be added. Time is necessary for sterilisation purposes, and within certain limits the longer the time allowed the smaller the dose of lime required for bactericidal purposes. At least 5-24 hours' contact should be given. The author has actually succeeded in sterilising very soft waters with so minute a dose as 1 part CaO per 100,000 parts of water by prolonging the contact to from three to six days. A margin of safety should be allowed, particularly in the case of very, hard waters, because here an increase in the temporary hardness might use up the excess of lime required for bactericidal purposes, whereas the same per- centage increase in the temporary hardness of a very soft water would not materially affect the bactericidal dose. In those cases particularly, in which it is sought to reduce the permanent as well as to remove the temporary hardness, the following modification may be tried. Sterilise the whole instead of a proportion of the supply, and instead of adding carbonate of soda, use bicarbonate of soda, when the following reactions may be assumed to take place. CaO + NaHCO 3 = CaCO 3 + NaHO. NaHO + NaHC0 3 = Na 2 CO 3 + H 2 O. Na 2 CO 3 + CaSO 4 = Na 2 SO 4 + CaCO 3 . It will be seen that 1 molecule of CaO (56) requires 2 molecules of NaHC0 3 (84 + 84 = 168) so that if the water after sterilisation contains an excess of CaO of 2 parts per 100,000 (2 Ibs. per 10,000 gallons) then 6 parts per 100,000 (6 Ibs. per 10,000 gallons) of bicarbonate of soda are required theoretically for the above purpose. Apart from questions of cost, the chief advantages of this modified procedure are as follows : (1) The whole of the supply is sterilised by means of lime. STERILISATION PROCESSES 77 (2) The water is rendered " softer " owing to the removal of part of the permanent hardness. The following example illustrates what has been said : A well water had a total hardness of 30 '32 ; permanent 9'29 ; temporary 2 TO 3. The water was inoculated with 1 per cent, of sewage to increase greatly the B. coli content, and then treated with an excess of lime. Sterilisation took place in 5 hours but not in 1 hour. The excess of CaO at the end of the experiment was Q'0068 per cent. Bicarbonate of soda was next added in correspondence with the foregoing equation. The result was that the water lost practically all its caustic alkalinity (CaO), and the hardness was reduced to 1'69. When carbonate of soda was used instead of the bicar- bonate, the hardness was found to be 8 '8 6 and the alkalinity 0*00487 per cent. It is obvious, therefore, that bicarbonate of soda can be used for the double purpose of neutralising the excess of CaO, and of reducing the permanent hardness. Consideration may now be devoted to the important question of storage, which may be described as Nature's own method of sterilisation. Note: The following useful table shows the method employed by American chemists in estimating the alkalinity of waters due to bicarbonates, normal carbonates and hydrates. (Standard Methods of Water Analysis, American Public Health Association, 1912) : Bicarbonates Carbonates Hydrates P=O P>|E E E-2P O O O 2P 2P 2(E-P) O O O 2P-E E E = Erythrosine alkalinity. P = Phenolphthalein alkalinity. Phenolphthalein does not react with bicarbonates and only with half the carbonates, whilst erythrosine estimates the alkalinity due to both. [Cochineal or methyl orange might apparently, if preferred, be sub- stituted for erythrosine.] CHAPTER VI STORAGE, IN RELATION TO PURIFICATION THE processes which .make for the purification of water under storage conditions are chiefly (1) Sedimentation, (2) Equalisation, (3) Devitalisation. (1) Sedimentation. In the old days this was regarded as the chief, if not the only, factor of consequence. Though we now know that sedimentation is insufficient per se to produce the desired results, yet its importance is clearly indicated in Chapter IV, where it is shown that even less than 24 hours' settlement has a considerable purifying effect. (2) Equalisation. This factor has never received the attention it deserves. If the water destined to be stored were, to begin with, of uniform composition and contained the microbes of water-borne disease in uniform distri- bution, no equalisation in the sense here meant could take place. But, as judged by the usual chemical and bacteriological tests, the quality of a river water varies enormously from time to time ; it therefore need scarcely be said that mere storage on purely physical grounds undoubtedly smooths over (levels, as it were) abrupt fluctuations in its quality. Moreover, my observations and experience confirm me in my belief, that even sewage-polluted river waters do not uniformly, or of necessity, contain (i.e. in ascertainable 78 CH. vi STORAGE 79 numbers) the microbes generally associated with water- borne disease (e.g. the typhoid bacillus). In support of the truth of this statement reference may be made to Chapter II and to the writer's reports 1 to the London Water Board in which detailed evidence is adduced that when equal volumes (A and B) of raw Thames river water, admittedly sewage-polluted, are taken and the one part (A) is purposely infected artificially with the typhoid bacillus or Gartner's bacillus, and the other part (B) is not so infected, it is comparatively easy to isolate the aforesaid artificially added microbes from (A) whereas it is practically impossible to demonstrate their presence in (B). In the case of the River Thames raw water, while it is of course possible that the typhoid bacillus may be uniformly present, yet surely it must be in such very few numbers as to be practically non-discoverable. On the other hand, it may be that, although only occasionally present, its numbers on those occasions are nevertheless of serious import. If the latter view is the correct one, the point to be urged is that, apart from sedimentation and devitalisation, the " levelling " effect of adequate storage would probably so average the infected and non-infected waters as to render the final mixture reasonably, if not demonstrably, safe. To take a concrete example : Imagine an individual who has the misfortune to be a typhoid carrier obeying the calls of nature in such a way as unwittingly to con- taminate a river water a short distance above the adjacent intakes supplying two towns, A and B. In the case of town A, the water, let us say, is taken directly on to the only line of defence, namely the filter beds, by means of which we can hope to remove, at best, 98 per cent, of the bacteria of all sorts. On the other hand, let us assume that in the case of town B, its main line of defence is 1 Second, fifth, seventh, and ninth reports on research work. 80 STUDIES IN WATER SUPPLY CHAP. adequate storage, antecedent to and in anticipation of filtration ; it must be obvious that, even apart from sedi- mentation and devitalisation the probabilities of an epidemic resulting, would be enormously greater in the case of A as compared with B. (3) Devitalisation. This is a factor of supreme if not paramount importance. Beyond all question the de- struction of the microbes of epidemic disease (e.g. typhoid fever and cholera) in water is merely a question of time ; although the exact time necessary for their deterioration and final disintegration is open to controversy. It is known to vary with the temperature, and no doubt is also influenced by many other factors, such as the prevalence of competitive organisms or their products. According to the writer's experiments, even one week's storage would result in a larger percentage reduction in the initial numbers of typhoid bacilli and cholera vibrios than we could ever hope to achieve by the use of sand filters alone under the usual practical conditions of working. Without question, however, to secure the absolute elimination of these pathogenic bacteria, several weeks' storage may be required, but for all practical purposes, provision for 30 days' real storage is ample, that is, when dealing with sources of supply comparable to those of London. Strong confirmation of the beneficial effect of storage has lately been obtained by the experimental proof that " un- cultivated " bacilli succumb in raw river water at a more rapid rate than their " cultivated " brethren (pp. 83-86). Let us now consider the vitality of (1) the cholera vibrio in river water, next that of (2) the typhoid bacillus, and lastly, the bearing of (3) temperature on the question generally. (1) Vitality of the cholera vibrio. Eighteen experi- ments have been carried out : namely Six with raw Thames water collected at Hampton. vi STORAGE 8 1 Six with raw Lee water collected at Ponders End. Six with raw New River water collected at Hornsey. It is important to note that antecedent to inoculating (i.e., seeding) them with the cholera vibrio, the samples of water were not sterilised. Subsequent to inoculation with the cholera vibrio, the samples were tested at weekly intervals by various methods, and in different media, for the presence of this microbe. Sometimes one strain, and sometimes a mixture of different strains of cholera, was used for inoculating the samples of water. The experi- ments were necessarily carried out in the laboratory, and this fact must not be forgotten in interpreting the results. The chief results are summarised in Table XXIV. (p. 82) The table shows the number of cholera vibrios added to the water, the number one week later, the percentage reduction in one week and the length of time this microbe persisted in 1, 10, and 100 cubic centimetres of the infected water. The bottles were always vigorously shaken before being tested. There need ' be no difficulty in understanding Table XXIV. A glance at column 2 will show that the number of cholera vibrios per c.c. of the artificially infected water varied from a maximum of 13 millions to a minimum of 49,000. Column 3 shows the enormous reduction of cholera vibrios that occurred as the result of even one week's storage in the laboratory. Column 4 indicates the percentage reduction, which in all the experiments was at least 99 '9. Columns 5, 6, and 7 show the number of weeks the cholera vibrios survived in 1, 10, and 100 c.c. of the infected water, respectively. The outstanding feature was that the cholera vibrios could not be isolated from 100 c.c. of the infected water (column 7) after one week (2 out of eighteen experiments), two weeks (9 out of eighteen experiments) and three weeks (7 out of eighteen experi- ments). o 82 STUDIES IN WATER SUPPLY CHAP. TABLE XXIV. VITALITY OF THE CHOLERA VIBRIO. Experiment. Initial Number of Cholera Vibrios per c.c. of the infected water. Number of Cholera Vibrios per c.c. one week later. Per- centage Reduc- tion in One Week. Number of weeks after infection of the water vhen the Cholera Vibrio could no longer be isolated from 1, 10 or 100 c.c. of water. Range of Tempera- ture during progress of experi- ment (deg Fahr.) 1 c.c. 10 c.c. 100 c.c. Cols. 1 2 3 4 5 6 7 8 # 1 T Nov. 2/08 3,750,000 j + 1 c.c. ; \ j - 0-01 c.c. / 99-9 2 2 2 5064 2 L 3,750,000 10 99-9 2 2 3 5064 Nov. 2/08 3 N.R. 3,750,000 20 99-9 2 3 3 5064 Nov. 2/08 4T.... ... 13,000,000 20 99-9 2 3 3 5162 Nov. 16/08 5L.... ... 13,000,000 20 99-9 2 2 2 5162 Nov. 16/08 6 N.R. Nov. 16/08 13,000,000 ( + 1 c.c. ; \ \ - O'Ol c.c. / 99-9 2 3 3 5162 7 T 9,532,500 10 99-9 3 3 3 5360 Nov. 30/08 8L 9,532,500 70 99-9 3 3 3 5360 Nov. 30/08 9 N.R. 9,532,500 20 ,999 2 3 3 5360 Nov. 30/08 10 T. 1,775,000 None 99-9 1 1 1 4555 Jan. 18/09 11 L. 70,000 / + 1 c.c. ; 1 \ A.I I 99'9 2 2 2 4555 Jan. 18/09 12 N.R. ... Jan. 18/09 3,150,000 I - 01 c.c. ) ( + 1 c.c. ; V \ - 0-1 c.c. j 99-9 2 2 2 4555 13 T. 680,000 None 99-9 1 1 2 4560 Feb. 1/09 14 L. 510,000 None 99-9 1 1 2 4560 Feb. 1/09 15 N.R. ... 406,500 None 99-9 1 2 2 4560 Feb. 1/09 16 T. 49,000 None 99-9 1 1 1 4856 Feb. 15/09 17 L. 110,000 None 99-9 1 2 2 4856 Feb. 15/09 18 N.R. ... 420,000 j + lc.c. ; } \ (VI P p 1 99-9 2 2 2 4856 Feb. 15/09 \ - V I C. C. J * The letters T, L and N.R. in column 1 refer to raw Thames, Lee, and New River water respectively. In experiments 1-6 a mixture of eleven strains of cholera (known in the laboratory as strains 1-11) were used. In experiments 7-9 a mixture of strains 3 and 4 were employed. In experiments 10 and 18 strain 5 was used. In experi- ments 11 and 16 strain 6 was employed. Strain 7 was used in connection with experiments 12 and 17. In experiments 13, 14, and 15 strains 2, 3, and 4 were employed respectively. VI STORAGE 83 (2) Vitality of the typhoid bacillus. Table XXV. sum- marises the chief results obtained during an investigation of this subject : TABLE XXV. VITALITY OF THE TYPHOID BACILLUS. Number of typhoid bacilli per c.c. of Number of Initial number the infected raw river water, after storage in the laboratory for : weeks required to effect the Experiment. of typhoid bacilli per c.c. of the infected destruction of the typhoid bacillus in Weeks. raw river 100 c.c. of the water. i infected raio One. j Two. Three. Four, j Five. river water. 1 T. 40 Five 2L 40 i Five 5 L 170,00) 53 2 Five 15 N.R. 525,000 29 3 Five 3N.R. 40 Six 4T 170,000 9 2 Six 6 N.R. 170,000 40 2 Six 8 L 470,000 850 11 7 2 Seven 9 N.R. 470,000 1,430 14 7 Seven 14 L 525,000 32 2 Seven 18 N.R. 475,000 30 3 Seven 7T 470,000 480 31 5 Eight 10 T 8,000,000 3,000 30 4 Eight 11 L 8,000,000 2,900 29 5 Eight 1ST. ... . 525,000 12 1 Eight 17 L 475,000 80 11 2 o Eight 12 N.R. 8,000,000 400 22 2 .L411C-UV Nine 16 T 475,000 210 12 2 1 Nine T. = Thames, L. =Lee, N.R. = New River. These were all experiments with " cultivated " bacilli, i.e., typhoid bacilli, which, after isolation from the excreta or tissues of persons suffering from typhoid fever, had been cultivated in the laboratory on artificial media. Very different results were obtained with "uncultivated" bacilli, i.e., typhoid bacilli as they existed in the urine of typhoid " carrier " cases, and which, of course, had never previously been cultivated in the laboratory. With the object of demonstrating, at least to his own satisfaction, that his faith was equal to his experimental knowledge, it is now somewhat well-known that the writer, 84 STUDIES IN WATER SUPPLY CHAP. of set purpose, drank, without ill-effect, half a pint of Thames River water, initially contaminated with typhoid urine, known to contain myriads of typhoid bacilli. These personal experiments were made on the 24th, 25th, 26th, 27th, and 28th days, after the date of the original infection of the water in the first test, and on the 23rd, 24th, 25th, 26th, and 27th days in the second test. In these experiments, the initial number of typhoid bacilli per half-pint was computed at 218,680,000 in the first and 468,000 in the second experiment. In the first case, 770,000 typhoid bacilli per c.c. of the infected water became reduced to only 4 within one week, and subsequent cultures made with 1, 10, and 100 c.c. of the infected water on the 14th, 18th, 21st, 27th, and 29th days after the start of the experiment, all yielded negative results, i.e., as regards the isolation of the typhoid bacillus. In the second case, the initial number of typhoid bacilli was computed at 1,650 per c.c. of the infected water, and at the end of a week none could be revealed in 1 c.c. Subsequent cultures made with 1, 10, and 100 c.c. of the infected water on the 10th, 14th, 19th, 21st, and 28th day after the start of the experiment, likewise yielded negative results as regards isolation of the typhoid bacillus. Markedly different results were obtained after the previous isolation and laboratory cultivation of the typhoid bacilli derived primarily from the urine of the fore- going " carrier " case. In its " cultivated " state the bacillus was found, indeed, to survive for five weeks in river water, under otherwise comparable conditions of experiment. Inasmuch as the risk of acquiring typhoid from the drinking of polluted water is mainly due to the possible presence of "uncultivated" typhoid bacilli, these observa- tions would seem to be of far reaching importance. A.t a later date, the subject was further investigated and the chief results are summarised in Table XXVI. (p. 85). VI STORAGE TABLE XXVI." UNCULTIVATED " TYPHOID BACILLI. 43 d 43 e g 43 c 43 C 43 c -g Q o> > d [ + = smallest volume of water Jj-8 k yielding a positive result.] Date. o'S o^ . i * |1 if toq d d d d d .2 t>0 d d d d Is 8 O l-H "* l-H 1907-8. e F ( o August 9 ... 430 120 32 + 12 ... 220 160 56 -f 19 .. 500 310 146 -f 26 ... 410 90 33 -j- September 2 ... 9 ... 220 600 60 27 13 73 * 16 ... 710 166 10 -f 23 ... 800 146 23 -j- 30 ... 440 106 21 + October 7 ... 620 124 38 -j- 10 ... 210 111 5 + 14 ... 750 168 52 + 17 ... 1,800 950 152 + 21 ... 4,000 220 34 + 24 ... 1,400 204 33 + 28 ... 1,100 270 8 + 31 ... 6,100 490 185 + November 4 ... 4,300 1,120 82 -f 7 ... 1,700 176 34 + 11 ... 1,400 112 15 + 14 ... 1,600 190 71 + 18 ... 1,110 96 5 + 21 ... 1,700 165 38 -f 25 ... 1,700 226 61 + 28 ... 21,000 625 100 + December 2 ... 4,000 [10] * -f 5 ... 16,000 900 120 + 9 ... 16,000 640 74 + 12 ... 2,800 1,300 150 + 16 ... 5,000 480 46 + 19 ... 1,600 470 24 + 27 ... 530 330 60 + 31 ... 830 500 29 + January 2 ... , 6 ... 210 1,500 650 440 52 28 + + 9 ... 30,000 1,000 250 Jf. 13 ... 65,000 230 44 -f 16 .. 7,000 290 40 + 20 ... 1,800 90 19 + 23 ... 1,500 82 61 -f. 27 ... 2,000 210 22 + 30 ... 2,400 180 46 -f February 3 ... 1,600 150 [OJ + 6 .. 900 130 21 -}- 10 ... 700 73 11 + * No record. TABLE XXIX. (continued). RIVER THAMES RAW WATER AT HAMPTON BEFORE STORAGE Number of microbes per c.c. Typical B. coli (lactose + indol +). ? [ -f = smallest volume of water gl^ Q fo yielding a positive result.] Date. < 12 "S (D H o If > d [ + = smallest volume of water il d k yielding a positive result.] Date. o^co kg II S-a o^ c3 "w t-< Q3 >? d d d d d C s I $* Jl .2 kJD^ "^ o ol n o g IP T3 I! jj 1 J ^i 1 J 1 3 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1907. August 7 ... 0-0018 0-0140 *0'20 1-67 0-1751 1-31 46 *21'00 *5'9 12 ... 0-0017 0-0162 *0'17 1-82 0-1601 0-70 34 23-87 4-71 19 ... 0-0051 0-0150 r O"l7 1-77 0-1535 1-45 37 2099 5-43 ;; 26 ... 0-0039 0-0132 *0'20 1-70 0-1377 0-80 41 22-30 5-60 September 2 ... 0-0022 0-0136 0-18 1-72 0.1287 0-60 32 21-8 4 ... 0-0013 0-0126 *0'18 1-70 0-1238 0-45 33 *21-8 9 ... 0-0036 0-0120 *0'18 1-73 0-1204 0-52 36 21-4 16 ... 0-0017 0-0118 *0'19 172 0-1099 0-45 36 21-6 5-3 23 ... 0-0019 0-0120 *0'19 1-80 0-1079 [trace] 34 21-9 5-8 30 ... 0-0026 o-oioo *0'17 1-79 0-1045 M 37 21-0 5-0 October 7 ... 0-0030 0-0092 *0'16 1-71 0-1042 " 32 19-9 5-4 14 ... 0-0038 0-0096 *0'19 1-77 0-1102 [31] 21-1 5-7 21 ... 0-0151 0-0240 *0'15 1-90 0-3484 7-09 90 [17-7] 6-1 28 ... 0-0088 0-0179 *0'18 1-87 0-3596 0-88 74 22-0 6-5 November 4 ... 0-0073 0-0248 *0'17 1-70 0-4930 7-95 140 24-6 7-5 11 ... 0-0077 0-0128 *0-25 1-70 0-2089 0-80 64 26-0 6-8 , 18 ... 0-0054 0-0088 *0'24 1-65 0-1462 0-57 54 24-2 6-2 25 ... 0-0085 0-0128 0-27 1-80 0-1959 2-43 70 26-0 7-7 December 2 ... 0-0085 0-0304 023 1-71 0-5835 17-40 290 22-1 9 'It , 9 ... 0-0091 0-0300 0-19 1-65 0-6651 35-00 360 21-0 5-5 16 ... 0-0052 00248 *0'20 1-59 0-5690 19-10 38U 20-0 9-2 23 ... 0-0088 0-0126 1-58 0-2408 4-93 100 30 ... 0-0106 0-0122 *0'31 1-58 0-1879 2-72 96 26-8 6-5 1908. January 6 ... 0-0169 0-0098 *0-36 1-66 0-1246 1-40 43 26-2 6-5 13 ... 0-0120 0-0292 *0'33 1 80 0-4737 7-64 180 21-8 9-0 ,, 20 ... 0-0126 0-0124 *0'31 1-78 0-1863 2-35 50 25-2 8-0 ,, 27 ... 0-0132 0-0094 *0'33 1-63 0-13*3 2-07 50 26-9 7-2 February 3 ... O'OllO 0-0086 *0-36 1-70 0-1184 0-81 40 27-4 8-8 10 ... 0-0096 [0-0085] *0-39 1-60 0-1068 0-52 45 27-8 8-5 17 ... 0-0086 0-0088 *0-31 1-60 0-1083 1-01 24 ... 0-0066 0-0120 *0-34 1-60 0-1421 0-81 60 27-9 6-4 March 2 ... 0-0064 0-0120 *0'32 1-79 0-1329 0-95 66 26-4 6-9 9 ... 0-0122 0-0332 *0'25 1-80 0-5046 18-20 200 22-8 8-4 16 '..'. 0-0074 0-0134 *031 1-79 0-1813 1-01 80 *26'6 *8'3 23 ... 0-0040 0.0102 *0'29 1-77 0-1129 [trace] 44 *24-9 *6-l 30 ... U'0082 0-0300 *0'25 1-66 0-4867 14-10 200 *25-4 *7-0 April 6 ... 0-0053 0-0176 "0-29 1-80 0-1995 1-81 93 *261 *7'8 13 ... 0-0037 0-0140 *0'28 1-67 0-1723 0-45 64 *24'8 *6'7 23 ... [0-0006] o-oioo *0'28 1-60 [0-1024] 0-47 51 *22'9 *5'7 27 ... 0-0088 0'0228 0-24 1-62 0-2608 . 135 *19'0 *7'1 May 4 ... 0-0040 0-0236 [*0'13j [1-39] 0-4915 2-90 208 22-6 8-2 11 ... 0-0038 0-0205 *0'28 1-59 0-2736 2-48 96 24-6 6-5 ,, 18 ... 0-0030 0-0130 *0-27 1-60 0-2110 1-77 76 24-1 6-8 25 ... o-ooio 00120 *0'29 1-60 0-1340 1 50 48 *24'5 *6-9 June 1 ... 0-0030 0-0110 *0'26 1-59 0-1236 1-58 39 24-2 5-9 12 ... 0-0059 0-0142 *0-25 1-59 0-1452 2-14 47 *24'0 *5'7 15 ... 0-0016 00146 0-28 1-59 0-1392 2-14 45 217 5-8 22 ... 0-0016 0-0156 *028 1-58 0-1304 2-14 36 21-8 6-0 29 ... 0-0031 0-0156 *0'25 1-59 0-1240 2-22 38 23-2 5-4 July 6 ... 0-0050 0-0162 *0'25 1-60 0-1387 2-22 43 21-6 4-7 ,, 13 ... 0-0044 0'0116 0-23 1-46 0-1152 1-36 36 22-5 5-2 15 ... 0-0032 0-0112 *0'23 1-54 0-1230 1-72 42 *22'5 *5'2 20 ... 0-0042 0-0114 *0'25 1-55 0-1301 1 33 43 23-3 [4-3] 27 ... 0-0036 0-0124 0-21 1-56 0-1256 1-09 44 22-5 5-4 Average 0-0060 0-0153 0-26 1-67 0-2127 3-50 83 24-29 6-82 As regards columns 4, 9, 10, the tests are not always carried out on the actual samples collected on the dates shown in column 1. The figures marked thus * are, however, taken from analyses carried out on samples collected within the same week. The figures in italics and the figures in brackets refer to maximum and minimum results, respectively. 93 94 STUDIES IN WATER SUPPLY CHAP. TABLE XXXII. RIVER THAMES WATER AFTER STORAGE IN THE CHELSEA RESERVOIRS. CHEMICAL RESULTS. PARTS PER 100,000. TJ jjjri ,1 If '3 a 'S G ' O DQO >>< 03 Is Date. Ammoni; nitroge Albumin nitroge 11 Q I 6 tygen abs Q Perman hours at i.d jfj our, mm. n a 2 ft. 1 n Perman< Hardne O o M 1 a' 1907. August 7 ... 0-0016 0-0118 0-17 1-66 0-1212 [trace] 46 [19-4] 6-7 12 ... O'OOll 0-0124 0-17 1-73 0-1285 31 20-4 4-9 19 ... 0-0012 0-0126 0-19 1-72 0-1261 31 22-3 5-9 26 ... 0-0014 o-ono 0-18 1-71 0-1271 26 22-6 6-4 September 2 ... 0-0015 0-0126 [0-16] 1-73 0-1201 ,, 27 21-5 9 ... o-oon 0.0112 0-17 1-70 0-1165 27 20-9 16 ... 0-0012 0-0098 [0-16] 1-68 0-1009 29 21-7 5'5 23 ... 0-0011 0-0120 18 1 70 0-0978 t 26 21-5 5-1 30 ... [0-0006] 0-0090 0-18 1-76 0-0955 [18] 20-2 5-3 October 7 ... 0-0013 0-0088 0-17 1-71 0925 [18] 20-8 6-7 14 ... o-ooos 0-0080 [0-16] 1-73 [0-0858] }> 22 21-2 5-3 21 ... 0-0013 0-0096 0-18 1-80 0-0956 24 21-1 5-4 28 ... 0-0017 0-0112 0-18 1-78 0-1851 37 20-0 5-6 November 4 ... 0-0027 0-0126 0-18 1'80 0-2028 43 20-6 6-2 11 ... 0-0025 0-0141 0-18 1-75 0-2587 57 21-1 6-9 18 ... 0-0015 0-0115 0-21 1-67 0-2127 n 48 22-6 6-6 25 ... 0-0015 o-oioo 0-23 1-60 0-1669 " 41 24-2 5-8 December 2 ... 0-0030 0-0094 0-24 1-65 0-1439 42 24-4 7-6 9 ... 0-0035 0-0141 0-23 1-66 0-2710 2-35 94 24-2 7-1 16 ... 0-0023 0-0168 0-21 1-54 0-3976 6-82 200 21-7 7-6 23 ... 0-0026 0-0154 1-60 0-3240 4-00 124 22-2 6-7 30 ... 0-0027 0-0134 1-61 0-2577 1-60 80 24-1 8-6 1908. January 6 ... 0-0061 0-0108 0-26 1-60 0-1976 0-55 36 25-4 7-3 13 ... 0-0059 0-0122 0-29 1-63 0-2042 [trace] 50 26-8 7-0 20 ... 0-0082 0-0124 0.33 1-60 0-2263 1-35 62 25-2 6-8 27 ... 0-0054 0-0110 0-32 1-62 0-1693 0-86 50 26-9 6-4 February 3 ... 0-0058 o-oioo 0-32 1-65 0-1428 [trace] 48 26-1 7-9 10 ... 0-0048 0-0091 0-33 1 69 0-1403 46 26-9 5-6 17 ... 0-0039 [0-0063] 0-33 1-57 0-1059 H ' 26 27-2 51 24 ... 0-0018 0-0077 0-35 1-54 0-0988 31 28-3 6-2 March 2 ... 00022 0-0086 0-32 1-75 0-1206 )( 41 269 7'3 9 ... 0-0020 0-0098 0-35 1-69 0-1255 . 47 26-8 6-5 16 ... 0-0024 0-0098 0-29 1-72 0-1358 41 25-4 6-7 , 23 ... 0-0017 0-0094 0-27 1-78 0-1516 44 26-4 6-5 30 ... 0-0015 0-0096 0-27 1-70 0-1380 58 26-5 6-8 Ap-il 6 ... 0-0020 o-oioo 0-24 1-70 0-1635 0-50 46 25-9 6-5 13 . . 0-0025 o-oioo 1-58 0-1661 [trace] 47 , 23 ... 0-0009 0-0086 0-23 1-64 0-1383 46 23-4 6-4 27 .. 0-0009 0-0092 0-24 1-60 0-1420 M 50 23-8 6-3 May 4 ... 0-0012 0-0080 0-25 1-58 0-1168 M 50 23-4 5-4 11 ... 0-0016 0-0124 0-21 1-48 0-2238 0-51 62 24-5 6-5 18 ... 0-0014 0-0124 0-21 1-50 0-2110 0-25 52 23-5 6-2 25 ... 0-0014 Q-0104 0-24 1-50 0-1691 0.52 56 23-9 7-8 June 1 ... 0-0018 0-0098 027 1-49 0-1371 0-67 33 23-8 6-5 11 ... 0-0020 0-0124 0-25 1-50 0-1339 1-33 31 23-3 6-2 15 ... 0-0018 0-0104 0-25 1-50 0-1199 1-09 36 21-7 6-1 22 ... 0-0014 0-0096 0-24 1-48 0-1028 0-50 32 227 5-7 29 ... 0-0010 o-oioo 0-25 1-49 0-0935 0-52 30 24-0 6-0 July 6 ... 0-0018 o-oioo 0-24 [1-45] 0-0915 0-72 32 24-3 5-1 13 ... 0-0012 0-0104 0-21 1-52 0-0931 0-60 27 23-2 4-7 20 ... 0-0016 0-0128 0-22 1-46 0-0995 2-14 30 21 3 ! [4'4] 27 ... 0-0008 o-oioo 0-22 1-46 0-1006 077 32 21-6 ; 5'9 Average . . . 0-0022 0-0108 0-24 1-63 0-1536 0-53 45 23-49 6-28 The figures in italics and the figures in brackets refer to maximum and minimum results respectively. vi STORAGE 95 The foregoing results may be summarised as follows : TABLE XXXIII. (FiG. 20). BACTERIOLOGICAL. . . Number of bacteria per c.c. Gelatine at 20-22 C. Agar at 37 C. Bile-salt agar at37C. River Thames before storage . ; . 4465 280 41 Chelsea stored water 208 44 5 Reduction (per cent.) 95-3 . 84-3 87-8 TABLE XXXIV. (FiG. 21). B. COLI TEST (LACTOSE + INDOL + ). Per cent, of samples yielding positive results. + 100 c.c. or less. + 10 c.c. or less. + 1 c.c. or less. + 0'1 c.c. or less. + 0-01 c.c. or less. River Thames before storage 99-9 97-7 83-1 48-3 10-1 Chelsea stored water 57-2 32-5 13-4 3-3 1-1 TABLE XXXV. (FiG. 22). CHEMICAL (PARTS PER 100,000). -2 13 ., *& .- ee 5 ?n li S)4^ >* C o> . bJO c S -i 1.1 11 |2 ^"i , | i" f s SH H ,2 6 Raw Thames before storage ... 0-0060 0-0153 0-2127 3*50 83 Chelsea stored water 0-0022 0-0108 0-1536 0-53 45 Reduction per cent. 63-4 29-4 27-8 84-9 45-8 96 STUDIES IN WATER SUPPLY CHAP. Figs. 20-22 (pp. 97-99) illustrate the foregoing results. As regards Fig. 22, the difference between the outside margin of the black area, and the black upper line represents the degree of purification effected by storage. It is impossible within the compass of this work to do more than condense the chief points showing the' advantages accruing from the simple storage of raw river water : (1) Storage reduces (a) the number of bacteria of all sorts. (b) the number of bacteria capable of growing on agar at blood heat. (c) the number of bacteria, chiefly exeremental bac- teria, capable of growing on a bile-salt medium at blood heat. (d) the number of coli-like microbes. (e) the number of " typical " B. coli. (f) the amount of suspended matter, colour, am- moniacal nitrogen and oxygen absorbed from permanganate. (g) the hardness. (2) Storage alters certain initial ratios, for example : (h) it reduces the number of " typical " B. coli to a proportionately greater extent than it does the number of bacteria of all sorts. (i) the colour results improve relatively to a greater extent than those yielded by the perman- ganate test. (3) Storage, if sufficiently prolonged, devitalises the microbes of water-borne disease, e.g., the typhoid bacillus and the cholera vibrio. (4) Storage produces a marked " levelling " or " equal- ising " effect. VI STORAGE 97 orJBACTELRlA r & * c.c STORAGE. AFTR. STORAGE. R.CDUCTIOM 95"'3 PER CENT LFQREL STORAGE. /^FTLR. STORAGE. IOM 64-5 P&R CE.NJ ^l^l^-h J3EFDR.EL STORAGE. AFTELR. STORAGEL THACSRAM TO IIJJJSTRATEI THE: EIFFECT RAW THA/AELS WATELR. OF STORAGE! m THE CHEILSELA RE.SElRVQiR.5 FIG. 20. II 9 8 STUDIES IN WATER SUPPLY CHAP. (5) An adequately stored water is to be regarded as a " safe " water, and the " safety change " which has occurred in a stored water can be recognised, and demonstrated by appropriate tests. (6) The use of stored water permits of a constant check being maintained on the safety of a water supply ante- cedent to, and irrespective of, filtration. B. COLI TE1ST I 100 cc. or LESS | IQc.c . I Icc-crULgS 10-lc.c.oKiE.Sg JO-01 c.c.> FIG. 21. THAMES WATER, BEFORE AND AFTER STORAGE (B. COLI TEST). (7) The use of stored water goes far to neutralise or wipe out the gravity of any charge that a water supply is derived from polluted sources. (8) The use of adequately stored water renders any accidental breakdown in the filtering arrangements much less serious than might otherwise be the case. These are some of the chief advantages of storage, but it must not be forgotten that there are also real or potential disadvantages. VI STORAGE 99 240 230 220 210 200 190 180 170 160 ISO 14-0 130 120 no 100 so 80 70 6O 50 40 30 20 10 CHLMIGAL RE.SULT3.r. ,=, RIVE.R THAMES BEFQR.E. STORAGE. (UPPLR Uf1) RIVER THA/AE1S A FTE:R. STORAGE. CHLLSELA KESLRVQIF15 FIG. 22. H 2 ioo STUDIES IN WATER SUPPLY CHAP. Sometimes the river water may be apparently so pure that it would seem to be desirable to pass such water directly on to the filter beds instead of pumping it into the storage reservoirs. Nevertheless, from the epidemio- logical point of view, the purity on such occasions may be more apparent than real. Again, it sometimes happens that Algse develop to such an extent in storage reservoirs as. to interfere seriously with filtration processes, and in some cases the water may acquire, in consequence, a most disagreeable fishy or oily taste and odour. At the beginning of the year 1913, part of London (Hammersmith, Kensington, and Hampstead) experienced a somewhat serious " taste visitation " ; the algal growth causing the objectionable taste and smell being chiefly composed of Tabellaria, with some Asterionella. Reference may be made to Nature (No. 2266, vol. xci. p. 117, April 3, 1913) and to the writer's monthly report for March, 1913. Some figures illustrating these and other algse will be found at the end of the Monograph (Figs. 38-43, pp. 195-197). These algal troubles may, it is true, be combated or even overcome by the use of copper sulphate (dose 2 to 10 Ibs. per million gallons) but such treatment should never be undertaken in the absence of skilled supervision, and responsible advice. As a last resort in the writer's experience the objection- able taste of waters, tainted as a result of algal growths can, in certain cases, \>e effectually removed by the addition of minute doses of permanganate of potassium to the water as it passes into supply. The dose naturally varies with the oxidisability of the particular water, but speaking of the London water supply the innocuous dose of about 2 '5 to 5*0 Ibs. of permanganate per million gallons of water was found to destroy the taste within a few minutes. The very faint preliminary pink tinge imparted by the addition of the permanganate rapidly fades away, and, in practice, VI STORAGE''^ : ; ::Vi#-V 101 the consumer receives a tasteless water with no pink tint, but slightly browner in colour than the normal. Here again, such treatment is only admissible in special circum- stances, and when placed under the vigilant control of competent advisers. In conclusion, it should be noted that the results observed in connection with the storage of Thames and Lee river water must not be read as necessarily applicable to cases in general, especially if the climatic conditions are not comparable. CHAPTER VII WATER AND DISEASE. 1 APART from the question of accident, presently to be dealt with, it is difficult to conceive that a water prepared with such elaborate care as that of London could have any influence on the incidence of water-borne disease. It has been shown that the typhoid bacillus and Gartner's bacillus could not be isolated from 1 cubic centi- metre and upwards of the raw Thames water, when employing methods which yield positive results in the case of laboratory samples artificially infected with even very small numbers of these pathogenic microbes (see pp. 33-36). It has also been shown that under the conditions of storage of London water, prior to filtration, the typhoid bacillus and the cholera vibrio, when artificially added to raw river water, in numbers inconceivably greater than could ever occur in practice, dimmish rapidly in number. It is true that a few bacilli may persist for several weeks, but having regard to the very large initial number added it may be said with assurance that real storage for one month antecedent to filtration confers reasonable, if not absolute, immunity from risk of water-borne disease (pp. 80-86). Moreover, it has been demonstrated that " unculti- 1 Reference may be made to a paper by the author on this subject, "Water and Disease," Jourtu.il of State Medicine, January and February, 1912. 102 CH. vii WATER AND DISEASE 103 vated " l typhoid bacilli die more speedily in water than their "cultivated" brethren, and the danger apprehended to the consumers of impure water is solely due to the possible presence of '''uncultivated " bacilli (pp. 8386). Further, London water is not only stored for several weeks, but is also filtered, and this filtration process removes about 98 per cent. of. the largely diminished numbers of bacteria. In the United States, as will pre- sently be shown, improvements as regards water supplies have antedated a diminished incidence of general and particular diseases, but the evidence that the two circum- stances are necessarily related, in a causative sense, is not altogether conclusive. That serious epidemics 2 have been caused by impure water cannot be doubted by any sane person, but the writer is of opinion that if all the facts were known the cause would usually be found to be accidental rather than inherent in character. It is proposed at this stage to deal with the question of accident, and later with the evidence which has within recent years led certain authorities to associate in a caus- ative sense the observed diminution in typhoid fever rates with increased purification of water supplies. The aim of a waterworks authority should be to prevent the possibility of accident, and not merely to produce and distribute a water supply of excellent quality on the average. This is an aspect of waterworks policy and pro- cedure which seems to have received only a minimum of attention, owing to an incomplete understanding of the importance of the subject, and due also to the belief that 1 By "uncultivated" is meant typhoid bacilli as they occur in the discharges of typhoid patients or typhoid "carriers." By "cultivated" is meant typhoid bacilli, which as a result of and subsequent to their isolation from such discharges have necessarily been cultivated or grown in the laboratory on artificial media. 2 For example, Worthing in 1893 (1,315 cases), Maidstone in 1897 (1,847 cases), Lincoln in 1905 (more than 1,000 cases). io 4 STUDIES IN WATER SUPPLY CHAP. filtration is the panacea for all ills connected with water supply. Yet if a workman suffering unwittingly from typhoid bacilluria micturated on a filter bed, although that bed was working superlatively well, can anyone doubt the proba- bility, if not the immediate certainty, of a disastrous epidemic? For out of the millions of typhoid bacilli voided with his urine two out of every hundred would be likely to pass into supply in a fresh and presumably highly virulent condition. In truth, filtration is only trustworthy when the un- filtered water is relatively so innocuous that, when 98 per cent, of its bacteria have been removed, the finished article ceases to contain any harmful element. Strictly speaking, no workman who has ever had typhoid fever should be actively engaged on waterworks. 1 On all works, indeed, adequate and convenient lavatory accommo- dation should be provided, a watch being kept on all the employees, and notices prominently displayed, forbidding, under the most severe penalties, any nuisance. In addition, the supply everywhere must be guarded 'both antecedent and subsequent to filtration. The filtered water during its passage to the consumer must be " cut off" from the possibility of those pollutions which may occur in a variety of ways, of which the following are but a few instances. During structural alterations (e.g., laying down new or replacing old mains) the supply may be jeopardised. Filter wells, pumping wells, and service reservoirs may be so situated as not to preclude the entrance of contaminating matters. It is commonly believed that the mains are secure owing to the water being under pressure, but physicists hold that 1 The time will assuredly come when no workman, who has had typhoid fever will be engaged or retained, unless he has been certified not to be a typhoid carrier, vu WATER AND DISEASE 105 under certain conditions it is not impossible for "insuction" to occur. The danger from defective types of hydrants is so obvious as scarcely to need special mention. In February, 1906, the Edinburgh water supply was observed to con- tain insects (" Sp ringtails," Collembola) in abundance occasioning considerable alarm in the minds of the in- habitants. Apparently these minute creatures were riot really of aquatic origin, but bred in the hydrant boxes and occasionally, owing to diminished pressure, were sucked into the mains and so reached the citizens' houses. It was proved by direct experiment that, by diminishing the pressure in the street main, the insects could be sucked from the hydrant boxes into the main water supply and so into the house cisterns in the vicinity. Storage reservoirs should be protected from the inroads and deposits of trippers, picnic parties, and the public generally. Unless the circumstances are exceptional, fish- ing should be prohibited, and the aim should be to eliminate any and every possible source of contamination. The river above the intakes for waterworks purposes should be guarded from pollution as zealously as is possible under the existing or improved conditions of the law, and if the Conservancy of the river rests in other hands than the waterworks authority, constant pressure, in the interests of the Public Health, should be brought to bear on those directly responsible to fulfil their obligations faithfully and well. It is especially important that the river as it nears the intakes should receive constant care as regards prevention of pollution. At the same time, it may here be pointed out that the provision of adequate storage accommodation enormously, if not entirely, reduces the risk of disaster following upon any accidental specific pollution just above the intakes. If a typhoid carrier micturated in the river above the 106 STUDIES IN WATER SUPPLY CHAP. intakes (say from a boat) the typhoid germs, in the absence of storage, would be carried directly on to the filters, with consequences which could hardly be other than calamitous. A similar accident in the presence of adequate storage would probably, if not certainly, be unattended by serious results, owing to the levelling and devitalising effect of storage. It is the bounden duty of every waterworks authority to guard at every stage in the purification process against the possibility of accident, and the one accident above all others which excites supreme dread is the contamination of the water with the urine or dejecta of a typhoid carrier. As compared with this source of specific infection, even sewage pollution, although highly objectionable, is much less to be feared for the reason, according to the author's practical experience, that typhoid bacilli are not usually present in ordinary sewages, in any large numbers certainly fewer than one typhoid bacillus per O'OOl c.c. At present, it is the aim (if not the compellable duty) of waterworks authorities to control pollutions in general, and also to purify the water to as high a standard as is reasonably practicable. In the future, it is hoped that, superadded to all this, the endeavour will always be to protect the consumer from chance infections, and especially from that deplorable form of contamination which is associated with the specific pollution of water by the urine and dejecta of typhoid carriers. Even if water is polluted with as much as from O'OOl to 0*0001 per cent, of crude sewage, the chances of typhoid bacilli being uniformly present are extremely remote, according to the author's own personal observa- tions ; but a water polluted with extremely small traces of typhoid urine would probably suffice to produce a serious epidemic. 1 1 The author added 1 part of typhoid urine to 50 million parts of raw Thames river water, and recovered the typhoid bacillus from the mixture, but vii WATER AND DISEASE 107 In summary of the foregoing observations, the writer ventures to urge that waterworks authorities should pay special attention to the following points : Protect as far as possible the river, especially just above the " intakes," from all sources of pollution. Interpose adequate storage accommodation between the river and the filter beds, and prevent any pollution of the water in these reservoirs. Guard the filter beds, filter wells, and sand from all sources of contamination, and filter the water as efficiently as is reasonably practicable. Prevent the possibility of pollution of the water in the pump wells, service reservoirs, and distributing mains. , Remember that accidents open the road to epidemics, and that the worst form of accident is produced by con- tamination of water with the recent discharges of typhoid carriers. Although river water supplies alone are here dealt with, the principles laid down apply, in greater or less measure, to all water undertakings. The writer in confining his remarks almost solely to typhoid fever, assumes that if we render our water supplies safe in this respect, we need, generally speaking, have little fear of any other water-borne disease. Lest it should be imagined that undue stress has been laid on the duty of preventing avoidable accidental con- tamination, it seems appropriate shortly to review here the circumstances pertaining to the recent Rockford (Illinois) typhoid epidemic so graphically described by Jordan and Irons in the Journal of Infectious Diseases (Vol. XL, No. 1, July 1912). The City of Rockford, situated on the Rock River in the northern part of the State, has a population of about on the other hand, he has repeatedly failed to isolate the typhoid bacillus from 1/1000 to 1/10,000 c.c. of crude sewage when using methods which yielded positive results in the case of artificially infected samples. io8 STUDIES IN WATER SUPPLY CHAP. 50,000. The present water supply is obtained from a series of deep wells sunk in the sandstone. As drawn to the surface it is of unimpeachable quality. During January and February, 1912, 10,000 cases of enteritis and 199 cases of typhoid fever occurred in the town. So gross and explosive a winter outbreak could not be adequately accounted for by any of the contributory causes of typhoid fever acting separately or in conjunction, e.g., contact infection, fly-borne infection, polluted shell- fish, etc. Probably the only agencies capable of striking a community so suddenly and so extensively are milk supply and water supply. Milk was excluded as a primary factor because " as many as 49 different milk dealers supplied the families among whom typhoid fever had developed, and in no instance was the number of cases on any particular milk- route disproportionate to the number of customers on that route." On the other hand, the city water supply was badly implicated. The private well drinkers largely escaped contagion, whereas the incidence on those drinking city water was very marked. For example, in one group of factories using private wells the incidence of enteritis was 5*8 per cent., whereas in a second group of factories supplied with city water the incidence was 70 '2 per cent. A concatenation of circumstances pointed to the proba- bility of infection of the water about January 16th. On Monday night (January 15-1 6th), a large fire occurred. There was, in consequence, a heavy draft on the water system, and on investigation the following facts were revealed : (1) One of the wells which had been out of use for some months was pumped into supply, and used to fill up the storage reservoir, which had become partially depleted owing to the aforesaid fire. The water issuing from the well proper was quite pure, but before passing into supply vii WATER AND DISEASE 109 was held up in a pit which received contaminating water from the adjoining street. (2) The storage or balancing reservoir is partly below the ground level and the water in it on January 16th was drawn from a normal range of from 18*7 to 14 '0 feet down to 6*6 feet. Under these conditions, it was proved that some " seepage" might occur, but Jordan and Irons came to the conclusion that " the reservoir, although dangerously placed and insufficiently protected, played a very minor part, if any at all, in the outbreak." (3) The ground water in the neighbourhood of the pumping station is highly contaminated, and it was found that the suction well or pumping pit was pervious, so that whenever the pumps were " speeded up " and the level of the water lowered, ingress of polluted water took place. The porosity of the pumping pit became manifest when it was purposely emptied. " On first stopping the pumps, the pit level fell about a foot overnight, and when all the water was pumped out about six feet of refined sewage seeped in "before morning" When this discovery was made one of the City officials was led to make the following remarkable admission : " We have been sitting on a keg of powder with a sputtering pipe in our mouths. That the explosion was not greater and more harmful than it was passes comprehension. We have had an excellent water source and then handled it as they did several decades ago put our water storage in the middle of a cesspool, and trusted to chance that it would not be contaminated." The Rockford case affords a melancholy example of the futility of providing a water supply of excellent quality, unless the water is also protected from accidental con- tamination, from its source until it reaches the consumer. Attention may now be invited to current views as regards the incidence of endemic 1 typhoid fever in relation to water supply. 1 Unfortunately, owing to insufficient data, it is practically impossible, in collating statistics to separate endemic from epidemic typhoid fever. This i io STUDIES IN WATER SUPPLY CHAP. It is one thing to recognise the danger of epidemics being caused by impure water, and quite another to associate the non-epidemic prevalence of typhoid fever with water supply as a factor constantly in operation, and of direct causative significance. Within recent years it has become quite common to hear in connection with, and as a result indeed of, improve- ments in, the water supply of a town, and the coincident decline of typhoid fever, such expressions as : " The typhoid fever rate has been cut in two owing to the increased purity of the drinking water." In the United States of America, there appears to be almost a consensus of opinion that a typhoid fever death rate in excess of 20 per 100,000 population is directly and wholly attributable to impure water supply. Some writers would even seem to ascribe a definite percentage of the typhoid death rate to water supply almost irrespective of its total amount. How strong the feeling is may be judged by the follow- ing quotation, which incidentally exhibits some remarkable financial considerations l : "Cincinnati, by the improvement of its water supply, reduced its typhoid fever death rate from 64 to 13 per 100,000 population. This corresponds to a saving of $5 '10 per capita per year or one and three-fourths million dollars per year. Columbus City, by installing a water purification plant, reduced the typhoid fever death rate from 56 to 17, corresponding to a gross saving of $700,000 per year or $3 '90 per capita per year. Chicago, by extending its waterworks intake four miles, reduced the average typhoid death rate from 80 to 34, corresponding to a per capita saving of $4*60 per year, or a gross saving of seven million dollars per year." is to be regretted, as one might be prepared to accept all explosive outbreaks of typhoid fever as being water-borne, subject always to the exclusion of such agencies as milk supply and shell-fish, and yet be very chary of accepting water as the cause of the cases occurring independently of such epidemics. 1 Monthly Bulletin : Ohio State Board of Health, Jan., 1913, Vol. Ill, No. 1. vn' WATER AND DISEASE in Again, Whipple in his treatise on Typhoid Fever says l : " When a contaminated public water-supply is suddenly improved in quality by the installation of a filter plant, there is nearly always a decided fall in the typhoid fever death rate. Cities which have pure water have a generally lower death rate than those which have an impure supply. These differences may serve as a rough measure of the amount of typhoid fever due to impure public water supplies. The average typhoid death rate in Ameri- can cities is about 35 per 100,000. The cities in the North which have safe water supplies have lower rates, usually as low as 20, and frequently as low as 15 or even 10. Taking the country over, perhaps 20 may be taken as an average figure. The difference between 20 and 35 may be considered, there- fore, as being due to infected public water supplies. Of the ' residual typhoid ' the most potent causes are probably infected milk, and direct infection by contagion, by flies, etc. Oysters, vegetables and other foods really play a very insignificant part in the general typhoid death rate. " A number of years ago infected water probably caused more typhoid fever than all the other causes combined. That is not the case to-day when the country as a whole is considered, although it is still, the most important cause, and in some cities, as in Pittsburgh and Philadelphia, it still over- shadows all other causes.' 3 " The long-continued struggle for pure water is bearing fruit, and to-day in many American cities, and even in entire States, where the public water supplies are well guarded from pollution, infection by water has come to be a secondary cause of the disease. "In a general sort of way it may be said that in the cities of the United States, at the present time, about 40 per cent, of the typhoid fever is due to water, 25 per cent, to milk, 30 per cent, to ordinary contagion (including fly transmission), and only about 5 per cent, to all other causes. In cities supplied with pure well water or filtered water, the effect of water is negligible ; where the water is impure, it is still the most important cause of the disease. In the case of rural districts there are no data to show the relative effect of the infected wells, infected milk, and direct infection, but, in all probability, the * honours ' are about even. " While the care of water supplies cannot be in any degree relaxed, efforts for further reducing the disease must be directed to causes other than water." Professor Dunbar 2 of Hamburg has also thrown the weight of his authority in favour of the views represented by Whipple, Hazen, Sedgwick, and many others. 1 "Typhoid Fever: Its Causation, Transmission, and Prevention," by George C. Whipple. 2 "Reflections, Old and New, on the Condition of Surface-water Supply Systems," Journal of State Medicine, Vol. XXI., No. 2. 112 STUDIES IN WATER SUPPLY CHAP. Thus he says : "Take London, for example; here in consequence of the continuous improvements effected in the waterworks, the mortality from typhoid fever has gradually declined to 5 per 100,000 in 1905 and to 3 per 100,000 in 1909." Again : "What have been the consequences of this condition of affairs [i.e., improved water supplies] on the public health ? During the interval between 1901 to 1905 the death rate from typhoid fever in towns of North America was calculated to amount to 46 per 100,000 against 7 '6 for Germany during the same period. Fifty years ago the death rate from typhoid fever in North American towns is said to have varied, just as it did in Germany, without any exception, between 50 and fully 100 per 100,000 living. Even in the year 1900, therefore, at a time when 6 '30 per cent, of the town dwellers (counting towns with over 2,500 inhabitants) were for the first time provided with filtered drinking water, but when the remainder still consumed unfiltered suface water, the death rate from typhoid fever reached still 46'5 per 100,000. However, in 1910, when 28*20 per cent, of the town dwellers referred to had filtered water available for use this mortality had become considerably reduced. " In 1900, 24-1 deaths from typhoid fever were notified per 100,000 living in Boston City, while in 1910 only 11 '6 were registered. " The following table will show some more of the statistics concerned with this development of sanitation : Town. Death rate from typhoid fever per 100,000 living. About 50 years ago. 1900. 1910. Boston, Mass Lawrence New York City Philadelphia Washington, D.C. Cincinnati Pittsburgh Indianopolis Richmond, Va. more than 60 100 more than 70 more than 50 100 24-1 20-4 37-2 707 41-4 103-0 52-6 11-6 ca. 20-0 11-7 18-0 24-4 5-7 Less than 30 29-9 21-9 31-5 New Orleans " This comparison shows clearly that the introduction of the sand-filtration system led to a marked lowering of the mortality from typhoid fever, but that, with the exception of Cincinnati, none of the cities show any result even approximately so striking as those achieved in our British and German towns at home." vii WATER AND DISEASE 113 There can be no question that in a number of American cities improvement of water supplies has antedated a decrease of mortality, but in our present state of in- complete knowledge, to conclude from this circumstance a direct causative connection between the two is perhaps hardly warrantable in all cases. The Americans, however, attribute the decline of mortality to an " exclusion of disease germs" or an " increase of vital resistance," or a "combination of these factors." l The mere fact that there has likewise occurred a decline in the incidence of tuberculosis, pneumonia, and respira- tory diseases generally, none of which can reasonably be regarded as at all likely to be water-borne, throws some doubt on the hypothesis that an improved water supply necessarily tends to produce any marked and continuously operating decrease of mortality from general and particular diseases. Let us consider Dunbar's statement : "Take London, for example: here, in consequence of the continuous improvements effected in the waterworks, the mortality from typhoid fever has gradually declined to 5 per 100,000 in 1905, and to 3 per 100,000 in 1909 " If we accept this as wholly correct, we are faced with the grave reflection that many other cities and towns con- taining in the aggregate an immense population would seem to be yearly sacrificing many thousands of lives which might be saved by the application, at a not impractic- able cost, of the principles of water purification adopted, for example, by the Metropolitan Water Board. Indeed the critic might go further than this and say that, inasmuch as the destruction by sterilisation of all the microbes of water-borne disease is quite a feasible and 1 Massachusetts Institute of Technology, Vol. VI., " Mortality Decrease Following Water Purification," Professor Sedgwick and Dr. J. Scott McNutt. 1 1 4 STUDIES IN WATER SUPPLY CHAP. practicable measure, implicit faith in the above conclusion demands the advocacy of sterilisation for waterworks purposes as the only means of absolutely, finally, and completely safeguarding the interests of the water con- sumer. In England, the seasonal prevalence of most of the cases of typhoid fever occurs year after year, before the " worst water periods" as judged by the chemical and bacteriological tests at present in use. Assuming these tests to be trustworthy indices of quality, there does not seem prima facie evidence that water played any large part in causing the observed seasonal fluctuation in the incidence of the disease. If seasonal variations in quality of water are apparently inoperative, is it wise to conclude that improved quality of water as a result of more efficient purification is of necessity a disease-preventing factor, merely because it happens somehow to seem associated with reduction in the typhoid fever death rate ? Again, is it not true that places, the water supplies of which are above suspicion, are affected, as regards the seasonal prevalence of typhoid fever, in much the same way as places less happily circumstanced in this respect ? Further, if we attempt to form a mental picture of the water supplies in the British Isles and try to arrange them in some sort of order of merit, using for this purpose whatever knowledge is available as to source of supply, manner of purification, analytical data, etc,, and then proceed to correlate our list with the typhoid death rate pertaining to the places dealt with, we soon reach the conclusion that many towns supplied with apparently very pure water have high endemic typhoid death rates, whereas others, supplied with a less pure water, seemingly enjoy comparative freedom from this disease. It would be of interest to. know as regards America : (1) Whether the curves illustrating the seasonal quality of the water as supplied to consumers show vii WATER AND DISEASE 115 any correspondence with curves showing the seasonal incidence of typhoid fever. (2) Whether the improvements in the supplies associated with the decline of typhoid fever have always brought about a difference in the actual quality (e.g., as judged by the B. coli test) of the water much greater than the difference observed annually during the "worst" and "best" water periods, which difference in England has not, generally speaking, been observed to affect materially the question of the incidence of typhoid fever. (3) Whether the curves illustrating the actual observed quality (e.g., as judged by the B. coli test) of the water in the various towns coincide with curves setting forth the incidence of typhoid fever in the same towns. (4) Whether the improvement in sanitary matters generally which so frequently is conjoined with an enlightened policy as regards water supply may not also be a factor in the situation. The author has carried out an elaborate search for the typhoid bacillus in samples of crude sewage, and finds that it is possible to isolate the typhoid bacillus from O'Ol c.c. of sewage when this microbe has been added to it artifi- cially in the proportion of 10 per O'Ol c.c. Strictly comparable experiments, but with sewage not purposely infected, have so far yielded absolutely negative results. The conclusion would seem to be reasonable that, as it is possible to isolate the typhoid bacillus from sewage if added artificially thereto in the proportion of 10 per O'Ol c.c. or of 1 per O'OOl c.c., the failure to find it in the non- infected samples, under strictly comparable conditions of experiment, means presumably that it cannot be present in the same proportion (namely, 1 per O'OOl c.c.). Now the tested samples of sewage contained the enor- I 2 u6 STUDIES IN WATER SUPPLY CHAP. mous number of about 1,000 excrement al bacteria (as judged by the bile-salt-agar test) per O'OOl c.c. If we apply the bile-salt-agar test to a sample of water, and even if we harshly assume that all the microbes growing in this medium are truly of direct sewage origin, it becomes extremely difficult to imagine the habitual presence of the typhoid bacillus, even in samples of raw river water, owing to the fact that such waters contain so comparatively few excremental bacteria. For example, the raw Eiver Thames (admittedly a sewage-polluted river) exhibits in this medium, on the average, only about 50 of these excre- mental microbes per c.c., so that we are forced to conclude that if it were -- = 20 times worse, the typhoid bacillus would presumably still be absent from 1 c.c. ; or to put it the other way round, in its present state it cannot be expected uniformly to contain the typhoid bncillus in 20 c.c. This river supply is purified about 1,000 times before delivery to the consumer, which inferentially signifies the absence of the typhoid bacillus from 20,000 c.c. Even if the purification process improved the water only 100, nay, even barely 10 times, we still are faced with the inferential conclusion of the absence of the typhoid bacillus from 2,000 or 200 c.c. respectively of such water. Moreover, it is exceedingly difficult to believe that the ingestion of a single typhoid bacillus would be at all likely to cause typhoid fever. The author is, of course, not suggesting in the face of conclusive evidence to the contrary that serious typhoid epidemics have riot resulted from the consumption of specifically polluted water, or that any relaxation in the processes of water purification is permissible ; but he does venture to express the opinion that the evidence brought forward in favour of the direct causative relation- ship between the quality of water supplies and the degree of incidence of endemic typhoid fever is not always wholly convincing. vii WATER AND DISEASE 117 The current desire to " cut the typhoid fever death rate in two " should not lead waterworks authorities to assume, without sufficient warrant, that the danger of epidemics has been wholly removed by improved processes of purifi- cation, or to forget the vital distinction between specific and non-specific pollution, and the enormous importance of guarding against the possibility of accident. The author is strongly of opinion that a water of only a moderate degree of purity, but sheltered from accident, is safer, in the long run, than a water normally of superlative purity, but not free from the possibility of accidental specific contamination. One cannot fail to be struck with the circumstance that cases somewhat like the following have been of not infre- quent occurrence. A town, say, in 1910 has a typhoid death rate of perhaps 30 per 100,000. In 1911 it has fallen possibly to 15, subsequently to an improvement in the water service, and it is not improbably eagerly claimed that the improved quality of the water has cut the typhoid death rate in two. In 1912 the figure perhaps reverts to 30 without any proven change in the water supply having taken place. It is impossible, of course, to accept any theory which associates diminution of the typhoid fever rate as attribut- able to improved water supplies, but seems to ignore increased incidence of the disease as of water significance. In conclusion, it is apposite to quote the following extracts from the monthly (January. 1913) Bulletin of the New York State Department of Health : Reduction in Typhoid Fever Rites Resulting from Improved Water Supplies. That the public health of a community is largely dependent upon the purity of its public water supplies hardly needs demonstration in the present day of enlightenment in sanitary science. Furthermore, that of all so-called com- municable diseases the typhoid fever death rate has by experience proven to be the best index of this relation, at least in American cities, and especially when the rates are in excess of a certain amount. Owing to the fact, how- ever, that typhoid fever is traceable to other causes than impure water; and n8 STUDIES IN WATER SUPPLY CHAP.. that contamination of supplies is generally of progressive occurrence, this relationship, as is well known to students of vital statistics, does not readily nor uniformly appear in the table or charts of typhoid fever rates of municipalities having public water supplies which are or have been subject to contamination. It is of considerable interest therefore to sanitarians to find examples or illustrations in actual practice which shew this relationship, especially where improvements have been made effecting the sanitary quality of the water. The State Department of Health has for some six years or more been actively engaged in a campaign for pure water supplies through the State. Although hindered greatly by inadequate laws and lack of direct control over public water supplies it has, by means of investigations, reports, and educational measures brought strong and effective pressure to bear upon many municipalities in improving the quality of their supplies. That these efforts have borne fruit will be seen from the following table showing the typhoid fever rates in certain cities of the State where during recent years the Engineering Division has been specially active in this direction. Accompanying the typhoid statistics for each municipality is a brief state- ment of the essential facts having a bearing upon the sanitary quality of the supply. This comparison shows with marked clearness in nearly all cases, the lowering of the typhoid rates corresponding to the improvements in the water supplies responsible for them. If the data were plotted this relation- ship would be all the more apparent. Furthermore, if these cities be taken as a class and typhoid rates arranged for each year a most convincing picture is presented of not only the relationship referred to but as to what can be accomplished in a practical way in the lowering of typhoid fever rates through improvements in the sanitary quality of public water supplies, along inductive lines carried out by the Department. TYPHOID FEVER RATES. Albany. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 24-8 20-3 20'0 10'9 18'8 14-9 17'8 178 Water filtered since 1899 prior to which typhoid rate was over 80. Pre- liminary filters installed in 1910. Hypochlorite used at intervals during past three years. Note a great reduction but not marked decline since filters were installed. Amsterdam. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per '100,000) 22-2 24-8 15'9 11-9 22'1 9'0 11-6 In 1908 Department ordered city to inspect watershed, resulting in removal of violations. In 1910 as a result of Departments recommendations, one polluted watershed abandoned. Note general but irregular decline in rate since 1906. Auburn. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 23-8 12-1 6-0 46'6 17'5 8'6 7'9 11 "2 vii WATER AND DISEASE 119 In 1906 rules and regulations protecting supply were enacted by Depart- ment, since which time orders have been issued to local boards covering 170 violations. Water board maintains active sanitary patrol. In 1908 special orders were issued to the water board to remove all violations. With excep- tion of 1908 note reduction and general low rate since 1905. Binghampton. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 25-6 9-1 18-2 15-2 13-1 12'4 4-0 178 Water supply filtered since 1902, prior to which rate exceeded 50. Note as with case of Albany, great reduction but no marked decline since filters were installed. Cohoes. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 92-9 57-8 78'2 62-0 82'2 76'8 108-5 48'0 In February, 1908, the Mohawk river pollution investigation pointed out high typhoid fever rates and necessity for improved supply. In June, 1908, full investigation was made of water in connection with city investigations and water filtration was strongly urged. In 1911 filter plant was installed. Note reduction in rate in 1912. Elmira. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 45-5 44-7 28'0 30'7 33 -5 26'9 13'3 15'9 In 1896 as result of serious epidemic, filtration plant installed. In 1909 investigation by Department showed continued high rates and recommenda- tions made to abandon local wells and increase efficiency of plant. Operation improved since 1909 and hypochlorite used during past two years. Note reduction in rate since 1909. Lockport. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 47-8 67-6 50'1 60'7 49'7 1M 22-0 16'6 In 1906 as result of outbreak of typhoid fever the Department recommended abandonment of polluted Erie canal supply. Since 1909 the supply has been taken from Niagara river. In 1911 Department recommended filtration. Note reduction in rate since 1909. Middletown. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 24-1 18-8 18-8 42-1 18'1 26'1 59'1 26'5 Mechanical filter plant installed in 1900 and enlarged in 1909. Hypo- chlorite used recently. In 1908 Department ordered city to inspect and remove violations on watershed. In 1912 outbreak of typhoid due to milk indicated high typhoid rate probably due in part to other causes than water. New York. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Tvphoid rate (per '100,000) 18-6 15-5 17-4 12-8 12'7 11'6 11-0 97 120 STUDIES IN WATER SUPPLY CHAP. Prior to 1907 practically no violations reported to the Department for a number of years. Since 1907 some 267 violations of water rules examined into and necessary orders issued by Department to local health boards. Since 1907 city very active in patrol of watershed and during past two years hypochlorite plant in use. Note marked decline since 1907. Niagara Falls. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 141-5 154-5 126'0 87'1 74'9 97'9 187 71'5 An investigation of the sanitary conditions of the city of Niagara Falls was carried on by the Engineering Division during the summer of 1907, following which the Department strongly urged a new or a filtered supply. In 1910 following another investigation, public warned of danger of supply. During 1912 niters installed, with hypochlorite treatment. Note marked reduction in rate in 1912 as result of filtration. Ogdensburg. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 52-8 67-3 47' 1 26'8 26'8 37'5 30'8 36'7 Tn 1907 as result of investigation the Department urged new filtered supply. Filtration plant installed during 1912. No typhoid deaths occurred during 1912 after filters were put in continuous use in the early fall. Oswego. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 47-5 58-0 66'0 62-2 26'6 51-2 127 21-0 In 1907, as result of investigation, Department recommended new supply or purification. In 1910 new supply from Lake Ontario installed with hypochlorite used at intervals. Note decrease in rate since 1910. Peekskill. 1903-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 22-5 15-2 14-0 202-3 12'9 26'2 19'0 O'O In 1906, as result of investigation, Department urged removal of violations of water rules. In 1908, as result of typhoid epidemic, Department renewed recommendations and urged water filtration. In 1910 new filters in per- manent operation Note reduction in typhoid rate since 1910, with no deaths in 1912. Poughkeepsie. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 42-4 39-4 112 34-5 23 17'8 14-0 13'7 In March and April 1907, following a typhoid fever outbreak, Department investigated water supply and recommended improvements in operation of filters. Following special report of Mr. George C. Whipple in April, 1907, improvements undertaken and completed in 1907. Hypochlorite used for past three years. Note marked and continuous reduction in typhoid rate since 1907. Rensselaer. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) ...^_... 67-5 18-6 58-3 30'0 29'9 28'0 18'7 9'4 vii WATER AND DISEASE 121 Water supply filtered since 1901. Following recommendations of the Department, since 1907 improvements in operation have been installed and hypochlorite used for past two years. Note continuous reduction since 1907 and especially in 1912. Rome. 1900-5. 1906. 1907. 1908. 1909. 1910. 1911. 1912. Typhoid rate (per 100,000) 21-6 28-2 17'0 26'4 16'0 19'3 9'4 22'9 In 1908 Department ordered city to inspect watershed and abate violation of rules and regulations. In 1910 and 1911 special inspections made by the Department following which city urged to remove violations. Note general but irregular decline in rate since 1908. Two diagrams (Figs. 23-24, facing p. 122 and p. 123) have been prepared to illustrate the foregoing data. In Fig. 23, each place has been treated as a separate entity, in Fig. 24 the death-rates have been calculated back to the actual numbers, the total numbers have then been divided into the total population and finally expressed as a com- bined rate per 100,000 persons for the periods stated. There are considerable difficulties involved in the theory of the causative association of the diminished typhoid fever rates with improvements in the water supplies. For example, Albany does not seem to have benefited materially as a result of the remedial measures employed, yet there is no suggestion that the processes of purification were less efficient than in other cases where the lowered typhoid death rate is definitely attributed to water improvement. Auburn, despite protective measures introduced about 1906, had a rate of 12'1 in 1906, 6'0 in 1907, and actually 46'6 in 1908. Binghampton introduced filtration after 1902. In 1910 the rate was 12'4, in 1911 only 4, but in 1912, 17'8. Cohoes, in 1911 installed a filtration plant. In that year the rate was 108*5 ; it is true that it fell to 48*0 in 1912, but if the filtration plant was at all effective, it surely ought, on a water-borne hypothesis, to have fallen much lower. Middleton has an unsatisfactory and fluctuating record despite the improvements in its water service. 122 STUDIES IN WATER SUPPLY CHAP. Niagara Falls, despite filtration and hypochlorite treat- ment in 1912, had still an exceedingly high rate; almost the same, indeed, as that for the year 1909. On a water hypothesis, if filtration and hypochlorite treatment were at all effective, the rate might have been expected to be greatly reduced. Oswego in 1910 obtained a new water supply from Lake Ontario with intermittent hypochlorite treatment. The typhoid fever rate was 51*2 in 1910, and it fell in 1911 to 127 ; but in 1912 it rose to 21*0, a figure not far short of the 26*6 observed in 1909. Rome in 1908, and again in 1910-11, was pressed to improve its water supply. It is not quite clear to what extent improvement actually took place, but the typhoid fever rate in 1912 was actually slightly worse than in the period 1900-5. It may be that these and other discrepancies could be explained away convincingly by a full knowledge of all the local circumstances pertaining to each case. It is now somewhat generally accepted that gross typhoid fever rates, especially if they are dominated by explosive outbreaks, can only be attributed to impure milk or water supply. Nevertheless, full recognition of the great importance of providing pure water supplies need not blind us to the fact that it is extremely easy to over-exaggerate the part played by bad water in fostering endemic typhoid fever. The facts revealed in the Monthly Bulletin (January, 1913) of the New York State Department of Health are of extreme interest as a basis for future discussion, but in the absence of further information they cannot be accepted as finally establishing the hypothesis that the reduction in the typhoid fever rates is to be ascribed solely to the improve- ment of water supplies. Hypochlorite treatment, if properly applied, ought to be not only relatively but absolutely effective in destroying all the microbes of water-borne disease. A large number PER IOO,OOO PERSONS LIVING. NIAGARA FALLS RENSSELAER OGDENSBURG X LOCKPORT POUGHKEEPSIE I I9OO-5 I I9O6 I I9O7 | 1908 | 1909 | I9IO | 1911 | 1912 | CERTAIN AMERICAN CITIES (NEW YORK STATE). FIG. 23. VII WATER AND DISEASE 123 I T900-5 I 1906 | 1907 | 1908 | 1903 | 1910 | 1S11 1 1912 1 COMBlhELD TYPHOID TE1VLR PCflTH RATE. PER 10QOOQ PERSONS LI V IMG " H900-5- | 19O6 | 1907 | 190& | 19O9 I 191O I 1911 I 1912. I CERTAIN A/AE.R1CAN CITIELJS (NEW YORK FIG. 24. i2 4 STUDIES IN WATER SUPPLY CHAP. of American towns are now applying the hypochlorite treatment and if the typhoid death rate in these towns is above what is known as the " normal " 1 it may be suggested that either the treatment is not being properly carried out or the water-borne hypothesis must be abandoned. As regards this matter, it would not be either a very difficult or a very costly matter to carry out a demo- graphic test on a vast scale. A number of large cities might be chosen in the United States in which the typhoid death rate was so considerably above the normal as to give prima facia support to the water-borne hypothesis. Each city could then be divided into two (or more) sections comparable as regards the age, work, and pecuniary and other conditions of its inhabitants, and generally in respect to sanitary matters. By arrangement one section could still receive its old water supply, and as regards the other an absolutely sterile water could be supplied year in and year out at a practicable cost, by the hypochlorite or other approved method of sterilisation. If the water con- tained so large an amount of organic matter and suspended matter as to render sterilisation difficult with moderate doses of hypochlorite, it might first be " rough-filtered " then oxidised by chemical methods or by the use of percolating filters, again filtered through fine material, next treated with hypochlorite in tanks holding not less than several hours' supply? and, finally, if found necessary or desirable, passed through filters containing charcoal or other de- chlorinating material. Another method would be to use alum antecedent to filtration so as to render the filtrate relatively free from oxidisable matter and then apply the hypochlorite treatment after filtration. 1 The "normal " is that death rate which is assumed to be of non- water origin and caused by such agencies as shell-fish, milk, flies, &c. 2 Never less than 1 hour and preferably 5, 10, or even 24 hours. This is very important, as it ensures absolute sterilisation with a minimum dose of hypochlorite. vii WATER AND DISEASE 125 Alternatively, comparable cities might be chosen and the water supplies of a selected number of them sterilised. It must not be supposed that this is a chimerical or impracticable notion. In the first place, we have from the United States the definite suggestion that the formula D = 275 (T N) appraises the monetary loss due to impure water, where D = loss in dollars per million gallons of water used. T = typhoid death rate per 100,000 and N = normal typhoid death rate, i.e., from causes other than water (see Chapter VIII. ). Secondly, from the same quarter we are informed that a vast population in the aggregate are in the position of having T greatly in excess of N. Thirdly, the argument is adduced that, apart from humanitarian considerations, it is much cheaper to purify water efficiently than to bear the burden of disease and death. Fourthly, it is also averred that there is a well marked causative relationship between improvement of water supplies and diminished incidence of general and particular diseases. If these contentions are really sound, the logical outcome is that the difference between T and N could be and should be " wiped out" by absolute sterilisation. CHAPTER VIII THE FINANCIAL VALUE OF A PURE WATER SUPPLY MOST persons will agree that a pure water supply is not only a great boon, but a sound public asset. It seems an unworthy task to attempt to place the value of human life on a monetary basis, or to argue that the supreme gift of health and attendant happiness can be reduced to a numerical value. Nevertheless, if the result of such calculations is to bring home to negligent waterworks authorities, that an unsatisfactory water supply may not only cause sickness and death, but also lead to financial instability, a good purpose will have been served; if it can also be shown that to remedy matters is a good in- vestment, even from a utilitarian point of view, a good result will have been obtained. Within recent years a great deal of work has been done on this subject, and though many of the conclusions arrived at are in the highest degree problematical, a very strong case has been established in favour of the financial advantages of a pure water supply. Perhaps the most ingenious and daring writer on the value of pure water has been Whipple, 1 and the author, although not necessarily endorsing all the views of that authority, feels that the matter is worthy of more than merely sympathetic consideration. 1 "The Value of Pure Water/' by Geo. C. Whipple. 126 CH. vni VALUE OF PURE WATER SUPPLY 127 Sanitary value of water. In estimating the sanitary or health value of water, typhoid fever has been taken as the chief water-borne disease ; Whipple, however, considers that improved water supplies also exercise a markedly beneficial effect on diarrhoea! diseases. Albany and Troy are given by him as examples because they are neighbouring towns, and because the former greatly improved its water supply, an example not fol- lowed by the latter town. The percentage reduction of death-rates over comparable periods in the two cases was as follows : Albany. Troy. Typhoid Fever 75 Diarrhceal Diseases 57 12 Children (under 5) 49 18 Total Deaths 17 6 Although not dissenting from the view that impure water may cause diarrhceal diseases, it must not be for- gotten there are also other and perhaps more powerful factors which determine the incidence of diarrhoaa in this country. In Fig. 25 (p. 129) is shown the annual death rate per million living, among children, under five years of age from diarrhceal diseases (1891-1900) for England and Wales as a whole, and for twelve towns the water supply of which is derived chiefly from moorland or upland gathering grounds, which, speaking generally, are remote from the possibilities of sewage pollution. It is surely apparent that we have to look elsewhere than to water supply for the true explanation of this disproportionate incidence of diarrhoea in these northern towns. A number of American authorities, however, associate, not only typhoid fever and diarrhoeal diseases with impure water, but pulmonary and certain other diseases, which we have hitherto always regarded as having no relation to water supply. 128 STUDIES IN WATER SUPPLY CHAP. The following extract (being a quotation from a paper by Hyde " On the Sterilisation of Water Supplies by the Use of Hypochlorites ") from Hooker's treatise on ." Chloride of Lime in Sanitation " shows how strong are the views of certain American sanitarians on this point. "Messrs. Mills (1893), Reincke (1893), Hazen (1904), Sedgwick (1910), and others have shown that when a pure water supply has replaced an impure one in a community, the general death rate therein is generally reduced in a considerably greater degree than would be acccounted for by the reduced prevalence of Typhoid Fever and other recognised typical water-borne diseases. A study of the vital statistics of numerous places where the quality of the public water supply has suddenly been changed from bad to excellent, as for instance, by the construction and proper operation of adequate purification works, has shown that for every person thus saved from death from typhoid fever, approximately three other persons are saved from death from other causes, many of which have probably never been thought to have any direct connection with, or to be especially affected or influenced by the quality of the public water supply. This numerical statement of the reduction in death rate more or less directly due to improved water supplies has recently become known as the HAZEN THEOREM, because Mr. Allen Hazen in 1903-05 was the first to announce in definite terms this interesting and most encouraging phenomenon. Even such unexpected diseases as tuberculosis, pneumonia, bronchitis, and a series of disturbances causing undue mortality among infants seem to be decidedly affected by such changes in the quality of the water supply. From general principles it is to be inferred that the drinking of a polluted and insanitary water supply must surely tend to lower the vital resistance. On the other hand an improved water supply must mean a real improvement in the general health tone of the community, a real uplift and reinforcement, rather than an impairment of the vital resistance of the con- sumer of such supplies." Reverting to the less disputable ground of typhoid fever, Whipple calculates that each typhoid death costs the community $10, 000 1 and that all typhoid deaths beyond the "normal" are attributable to impure water. The " normal" is a suppositional figure arrived at by inferring the probable number of cases caused by such agencies as milk, uncooked food, shell-fish, flies, etc. In America 20 is usually taken as the " normal," but this is 1 To convert dollars into pounds, divide by five. The above sum includes the cost to the community of the non-fatal cases of typhoid fever. Speaking very broadly, each typhoid death represents usually about ten cases of the disease. vin VALUE OF PURE WATER SUPPLY 129 MOO 12.000 1 1 000 9.000 8000 6000 5.000 AhhUALPEIATH RATE. (PELR. AM L LI ON LIVING-) 'A/AOM G CH1LPRE1M UMP^R FlVE^YElflRS FRO/A 3)1ARKHCALPISE:ASE:5 1691- 1900 CHICfLY-MOORLAMp -OK U PLAhp- GATHERl MG- G ROUhpS WHICH (SPEAKINQ-GELME-RflLLY) ARE RLA\OTC FRQ/A-THE1 S OF SEWAGEL-POLLUTlOh FIG. 25. 1 30 STUDIES IN WATER SUPPLY CHAP. sometimes reduced to 15 or lower where the sanitary con- ditions are good and increased to 25 or more where the converse holds good. Obviously everything depends on the correctness or otherwise of whatever figure is chosen to represent the "normal." In London the total typhoid death-rate is only about 4 per 100,000 ; so that either we should have to conclude that water played no part in causing the disease, or reduce the American " normal " five times or more. In order to arrive at the loss in dollars per million gallons of water used, the following formula is applied : T = Typhoid death rate per 100,000. N= " Normal " rate. Daily consumption taken at 100 gallons per capita. (T-N) 10, 000 = loss to community in dollars for 365x100x100,000 gallons of water, or ( T "'-^j 1 ' 000 = 2'75 (T - N), where 3o5 D stands for loss in dollars per million gallons of water used. The application of the formula is of considerable interest. Taking the average death-rate from typhoid fever in American cities as 35 per 100,000 and assuming a value of 20 for N. D = 2-75 (35 -20) = $41 "25. On this basis American cities are losing, owing to unsatisfactory water supplies, $41*25 per million gallons or about $15,000 per annum for each million gallons a day of supply. Whipple gives Pittsburg as an example l : "In Pittsburg for example, the typhoid death-rate for several years has averaged 120. Here according to formula (1), D = 275(12020) = $ 275 per million gallons. This is figured, however, on a per capita water consumption of 100 gallons a day. The actual consumption is about 250 gallons per capita 100 per day ; hence D should be taken as - of $275, or $110 per million gallons. *)\J 1 Whipple's book is dated 1907. Many changes have occurred since then, and this fact must not be forgotten in judging the present day value of all the figures and quotations given in this chapter. vin VALUE OF PURE WATER SUPPLY 131 Each million gallons of polluted Allegheny water pumped to Pittsburg has therefore reduced the vital assets of the community by $ 1 10. This for a population of 350,000, amounts to $3,850,000 per year a sum enormously greater than the annual cost of making the water pure." In the British Islands the consumption of water per head is much less than 100 gallons, so that D must be corrected on the basis of .p^ ~ : . Taking a general Daily Consumption average, the consumption would only be about one-third, which involves multiplying D by 3. Perhaps no very serious exception can be taken to the figure of $10,000 for each death from typhoid fever, but the crux of the whole position is fixing the highly deba- table figure for N. For example, the typhoid death-rates for Edinburgh, London and Glasgow are about 2, 4, and 8 respectively. It would take a very bold controversialist to suggest here the figure 2 for N and to carry into the foregoing formula the figure 2 for London and 6 for Glasgow (T - N). It is only necessary to compare the widely different typhoid death-rates in European and American cities to realise how delicate a matter it would be to suggest a number for N. The following diagram (Fig. 26, p. 133) has been con- structed from some figures given by McLaughlin in a paper entitled " The Eradication of Typhoid Fever." 1 The population of all the cities dealt with exceeds 300,000. A glance at the diagram enables one to realise how difficult, if not impossible, it would be to suggest a non- controversial figure for N. For example, taking Paris the worst of the European cities given in the diagram as the normal (namely 5*6), the writer has calculated that on this basis the aggregate 1 Boston Medical and Surgical Journal, Vol. CLXVI., No. 21, pp. 764- 771. May 23rd, 1912. K 2 132 STUDIES IN WATER SUPPLY CHAP. number of deaths'from typhoid fever in the 1 5 American Cities, in excess of the assumed normal of 5 '6, totals 1,495, and probably the case rate would be ten times greater. To attribute so much suffering and waste of life to impure water would be a serious step to take. If it be assumed that bacteriological tests, although indirect and subject to certain limitations, afford a not untrustworthy means of comparing the potential quality of different waters in relation to disease, then the cities dealt with in the diagram ought to show on bacteriological examination results in correspondence with the curve exhibited in the diagram in order to fit in with the water- borne hypothesis. Such an investigation, if carried out by competent observers using precisely the same methods and media, would be of extreme interest. In the absence of an international standardisation of methods and media, it is impossible to institute many comparisons of the gravest importance, alike to bacteriologists and epidemio- logists. Physical Quality of Water. Too little attention is paid to the physical characters of water from a financial point of view. A water supply which is too highly coloured or turbid, or has an objectionable taste, drives many people to use household filters or to increase the demand for bottled waters. It may be that a supply unsatisfactory in these respects may be hygienically quite safe, but it is very difficult to persuade complainants that such is the case, and the almost invariable retort is how can it be pure if it is dirty-looking or " smelly " ? The formulae suggested by Whipple for calculating loss due to colour, turbidity, and smell are admittedly open to criticism, and it may suffice to quote the effect of their application to particular instances : ( ' For instance, such a water as that now supplied to New York city from the Croton river has a depreciation of 1 11 per million gallons, or nearly a viii VALUE OF PURE WATER SUPPLY 133 THE: TYPHQ1P TEIVEIR PE1ATK RATE1S IQQ.QQQ ) m EIUROPElAh AMElRlCAh EllVKQPElAn C1T1E1S C1TIEL5 FIG. 26. i 3 4 STUDIES IN WATER SUPPLY CHAP. million and a half dollars a year for a daily supply of 350 million gallons. At 4 per cent, this represents the interest on about $35,000,000, a sum several times as large as the cost of filtration. An algae-laden water like that of Ludlow Reservoir at Springfield, Mass., has a depreciation of more than $ 20 per million gallons, because of its. odour and turbidity. A coloured water like that of the Black River at Watertown before filtration has a depreciation of $ 11, while a turbid water like that of the Mississippi River at St. Louis gives $25." Hardness of Water. As regards the hardness of water the greatest diversity of opinion exists ; and the consensus of medical opinion seems to be that for the general run of people the matter is relatively unimportant. London, the largest, and one of the healthiest cities in the world, has a decidedly hard water supply. Glasgow, the second largest city in the British Isles, is exceedingly proud of its very soft supply, derived from Loch Katrine. On the other hand, a hard water supply undoubtedly means a loss due to increased consumption of soap, waste of fuel, etc. Whipple calculates that for every increase of one part per million of hardness, the cost of soap increased about $10 per million gallons of water completely softened. All the water used by a community is not completely softened, and Whipple considers that a conservative estimate is one gallon per capita. On this basis, the depreciation of water, on account of its hardness, is D = H/10, in which H = the hardness of the water in parts per million, and D the depreciation in dollars per million gallons. Table XXXVI. (p. 135) is given by Whipple in illustration of the application of the formula. The Aftermath of Water Epidemics. Whipple does not appear to have taken into account the ultimate heavy financial loss accruing when a town loses its fair reputation as a result of a water-borne typhoid epidemic. It may be doubted whether Worthing and Maidstone have yet fully recovered from the effects of their typhoid epidemics in vin VALUE OF PURE WATER SUPPLY 135 1893 and 1897, respectively. Lincoln, similarly afflicted in 1905, despite its magnificent new supply introduced in 1911, is likely to suffer financially for many years to come. The old saying, " Give a dog a bad name," applies with almost brutal force and persistency to the future prosperity of a town which has lost its reputation for health. It is almost useless to urge that the old water supply has been efficiently purified, or even that a new and perfect supply has been obtained. A lost reputation as regards health TABLE XXX VL DEPRECIATION DUE TO "HARDNESS" (WHIPPLE). State. City or Town. Source of Supply. Total hard- ness (parts per million). Depreciation per million gallons. Maine Massachusetts New York . . Augusta Waterville . . Boston Cambridge . . Pittsfield . . New York . . Kennebec River ... Messalonskee River Sudbury and Nashua Rivers Storage Reservoir 5) < Croton River ... 20 15 12 33 50 40 $2-00 1-50 1-20 3-30 5-00 4-00 > j> Pennsylvania Ohio England Albany Oswego . . Philadelphia Toledo Columbus . . Warren . . London . . > Hudson ,, Oswego ,, Schuyekill River .. Maumee ,, Scioto . ,, Mahoning ,, Chelsea Company ... East London ,, ... 64 191 179 200 335 578 215 243 6-40 19-10 17-90 20-00 33-50 3350 21-52 24-30 affects a town long after the introduction of improved sanitary measures, and persists, despite the publication of reassuring vital statistics. Summer and holiday resorts, and educational centres, are always very hardly hit, as their welfare depends on what may be called an "optional" population. Nearly all classes of trade suffer either directly or in- directly to a greater or less extent, and the burden may not be lightened for many years. No formula can be suggested for estimating the financial : 3 6 STUDIES IN WATER SUPPLY CHAP. loss attributable to the aftermath of a typhoid epidemic, but that it is very serious cannot be doubted. Table XXXVII. has been constructed from figures given by Whipple to show to what extent, in his opinion, the sanitary value of a polluted public water supply is increased by an efficient system of filtration : TABLE XXXVII. INCREASE IN SANITAIIY VALUE (IN DOLLARS). Place. Per million gallons. Per year. Per year per capita. Lawrence, Mass Albany, NY. 652 130 665,000 450 000 9-50 4'75 Binghampton, N. Y Watertown N Y 65 120*34 160,000 175 000 3-80 6'90 As regards physical quality, Whipple calculates that by improving their water supplies, Lawrence, Albany, Yonkers, Poughkeepsie, Binghampton, Watertown, Little Falls, and Brooklyn increased the value of their water supplies by 6'0, 8'4, 9'0, 14'2, 10'4, 12'4, 9'2, 9'0 dollars per million gallons respectively. Again, the same writer claims that Winnipeg (Manitoba) and Oberlin (Ohio), by softening their water improved its value by 3870 and 12*20 dollars per million gallons respectively. The figures refer only to water used for domestic purposes ; if industrial uses were also considered, the gain would be materially greater. As regards the cost of purification (i.e., by filtration), Hazen calculates that : "As a general average, with a well-designed modern plant adapted to its work, the cost of filtering water, exclusive of pumping, but including all costs of operating the filters and furnishing the supplies required, and including the interest on the cost of the works and a reasonable allowance for repairs and depreciation, will amount to about $ 10 per million gallons or one cent per thousand gallons of filtered water." "In a general way, the purification of the water adds from 10 to 20% to the entire cost of furnishing and supplying water to an American city." viii VALUE OF PURE WATER SUPPLY 137 In England the average cost of sand filtration is usually placed at id. per 10,000 gallons. It is open to anyone to challenge and criticise the facts, figures, formulae, and inferences here presented, but it is impossible to gainsay the broad and general conclusion that, even if we disregard all humanitarian considerations, the provision of a pure water supply is a sound financial proposition. CHAPTER IX BACTERIOLOGICAL ROUTINE METHODS IT is proposed in this chapter to deal with the labelling of samples, registration and classification of results, explanation of symbols and variations occurring in connection with the so-called "hektograph" sheets, collection of samples, decimal mode of dilution, the use of racks, coloured wools, labels, and so on. Sample Collection boxes. The wooden boxes used for this purpose are made to carry 4, 6, 8, 10, and 12 bottles, each bottle having a capacity of about 180 cubic centi- metres. 1 Figure 29 (p. 141) represents an eight-bottle box and is self-explanatory. Labels. The author uses a very simple method for labelling samples. Each sample collector is known by the first letter of his surname, and in cases where two or more names begin with the same letter, some simple device is used to distinguish them, for example, taking the first and last letters of the surname. Every collector begins each year by labelling his samples 1, 2, 3, and so on to the end of the year, prefixing the number with the first letter of his surname. An actual example may be given (Fig. 27, p. 139). 1 The American Public Health Association recommends two-ounce bottles, but this small size precludes the possibility of 100 c.c. cultures being made. CH. ix BACTERIOLOGICAL METHODS 139 The labels are contained in a book and are printed in triplicate on each page with a perforated division to allow > ^^- FIG. 27. Facsimile of a Label used for a water sample. of two of them being detached, as is indicated in the following figure (Fig. 28) which is of course purposely reduced in size. The description of the sample is the same on A, B, C. B is torn out and stuck on the sample bottle, C is torn off after the sample collector . reaches the Laboratory and is pasted, together | with those handed in by *> the other sample collectors at the end of the day's collection, into a book ou specially kept for this 2g purpose. The labels in this book are then given " running " numbers, so that the total number of samples collected up to any date can at once be found. As a matter of fact, inasmuch as samples for chemical examination have also to be collected, it is the custom to have red " running " numbers for chemical samples, blue for the bacteriological and black for the aggregate number. The sample collector retains counterfoil A in his book so 140 STUDIES IN WATER SUPPLY CHAP. as to guide him in the collection of further samples. In this way, each sample collector is responsible for his own labels and numbers, and if any question arises about any sample, no conflict of opinion can occur between the different sample collectors as to misunderstood instructions, the inquiry being necessarily limited to what a particular man has done or left undone. Some notes on the registration and classification of results, with particular reference to the B. coli test. A full copy of the label, together with all the subsequent results, are entered in a book known as the "parent" laboratory book. This book contains all the permanent records necessary for future abstraction and classification purposes. As the " parent" laboratory book contains, opposite the space left for the results of the analysis of each sample, a full copy of the label referring to it, together with its letter and number, it is only necessary to label all the cultures and sub-cultures relating to the sample with the sampler's number prefixed by the first letter of his surname. Loose working sheets are also used and on these the results obtained are " pencilled " opposite the letter and number from day to day, and so when the final results come out one worker reads out the figures, and another worker records them neatly in the " parent" laboratory book. In addition, records are abstracted on to loose sheets, using hektograph ink for this purpose, and these are arranged under their proper headings, e.g., raw Thames water, raw Lee water, etc., etc. At the end of the month several copies are " pulled " off and used for filing, Committee, and other purposes. The foregoing description takes no account of the numerous other ways in which the results are abstracted for special purposes, but the whole system is based on a carefully thought-out plan, which aims at simplicity, accuracy and rapidity. When more than 1,000 samples a month have ix BACTERIOLOGICAL ROUTINE METHODS 141 REMOVABLE COPPER RECEPTACLE. FOR ICE R.EAAOS/A&UE. TOR- BOTTLES RA\OVA9LE- COPPER RECEPTACLE;- FOR '\ce /O o o /O O "-"-> C /O :"-> o o x o o c-j c x t* FAL.SE. DOT TO /A -REMOVABLE FELT CASE. FOR BOTTLES. FIG. 29. Sample Collection box. 142 STUDIES IN WATER SUPPLY CHAP. to be examined minutely, it is a matter of great im- portance so to organise the work as to prevent errors in entries or any accumulation of arrears of work. Examples of the "hektograph" sheets already referred to are here given (p. 148). These have been purposely reduced in size, and slightly modified so as to allow of reproduction on a single page. The symbols on the page will be un- intelligible to the uninitiated reader, but they enable the observer to register what is technically known as the " farthest out + " for the different B. coli results (presump- tive, confirmatory, and typical) for each sample examined. This finally allows percentages to be struck with perfect accuracy owing to the self-checking nature of the method. By the " farthest out + " is meant the smallest volume of water which yields a positive result on a presumptive, confirmatory and typical B. coli basis. Thus the B. coli results would in practice be summarised in the manner shown on the hektograph sheets, and the results (taking sheet 1 as an example) finally tabulated as follows : TABLE XXXVIII. B. COLI TEST, LAMBETH DISTRICT, GENERAL WELL, JAN., 1907. (Summary from Sheet 1). M Number of Samples. - 100 c.c. of water. +100 c.c. of water. +10 c.c. of water. +1 c.c. of water. d > 1 2 Con- firma- tory. Typical. 9 f & s z 1 7 Con- firma- tory. Typical. Presumptive. Con- firma- tory. Typical. Presumptive Con- firma- tory. Typical. Gl or Ag. $ 4 if 6 B 1 6 9 & 8 8 (i <5 & . s.S a j &* 'S> <3 Gl or Ag. tc ^ It JL C u Gl or Ag. bi 3 s_ a .| P i 3 9 10 11 12 13 14 15 16 17 18 19 20 21 22 % 13-6 3 & 5 & 9 J-4 10 & 16 J 10 % 40-9 9 &* 9 % 36-3 8 % 9-1 2 % 40-9 9 & 8 18-2 4 18-2 4 ilt 4 % % % % % It will be noted that the sum of the figures in the following columns corresponds to the number of samples ix BACTERIOLOGICAL ROUTINE METHODS 143 examined, and the sum of the percentages of course in each case is 100. Presumptive columns 2, 7,12, 17. Confirmatory {s lora S columns 3, 8,13,18. lag columns 4, 9,14,19. m . * /Sagin or agin ... columns 5, 10, 15, 20. "(Agin columns 6, 11, 16, 21. To take the presumptive test as a single example : 22 samples of water were examined, 100, 10, 1, OT and 0*01 c.c. cultures of each sample having been made. 3 Samples (13 '6%) yielded negative results in all cultures (column 2). 10 Samples (45*4%) yielded positive results with 100 c.c. but not with 10 c.c. of water (column 7). 9 Samples (40*9%) yielded positive results with 100 c.c. and 10 c.c. but not with 1 c.c. of water (column 12). 22 Samples 99 '9%. The same principle governs the classification of the confirmatory and typical B. coli results. It will also be noted that the B. coli test yields results which may be classified under five separate headings. Test 1. Presumptive B. coli test, mixed culture gaseous fermentation of primary medium, 48 hours at 37-40 C. Test 2. Confirmatory B. coli test, on a general basis. Presumptive test has been confirmed by the isolation of either a glucose fermenting ("gl")> or a lactose fermenting ("ag ") coli-like microbe. Test 3. Special confirmatory B. coli test, on a lactose basis (" ag "). Test 4. Typical B. coli test, on either a "sagin" or " agin " basis. The microbe isolated in pure culture is indistinguishable (on the basis of the tests employed) from the typical B coli of the human intestine. Here the term *' sagin " represents a microbe which forms gas in saccharose and in lactose media, and produces indol in peptone water i 4 4 STUDIES IN WATER SUPPLY CHAP. culture. " Agin," indicates fermentation of lactose, and formation of indol, negative result with saccharose. Test 5. Specially typical B. coli, on an " agin " basis. Here saccharose fermentation excludes an otherwise typical B. coli, from being accepted as specially significant of undesirable pollution. The two examples of hektograph sheets (see end of chapter) are necessarily at first a little difficult of com- prehension, but if the reader follows the explanation of the symbols given the matter will become comparatively simple. It may assist somewhat if all the actually occurring variations of results recorded on the two hekto- graph sheets are explained in detail. Variation 1. (M3, M12, M20 (Sheet 1) ) Here the presumptive test (black dots) yielded positive results with 100, 10, but not 1 c.c. of water (see columns 4, 9 and 14). Hence the black dots are placed on the farthest out + , namely, in column 9. The confirmatory test on a general basis (blue bars) yielded positive results with 100, 10, but not 1 c.c. of water (see columns 5-6, 10-11, 15-16). Hence the bars are placed on the farthest out +, namely, between cols. 10 and 11. As regards the special confirmatory test (blue dots on right of blue bars, positive results were obtained with 100, 10 but not 1 c.c. of water (see columns 6, 11, and 16). Hence the dots are placed on the farthest out +, namely, in column 11. No attention need here be directed to the dots on the left side of the bars, as a record of these need only be taken when the more significant special confirmatory test fails. Next, as regards the typical B. coli test on a general basis (red bars) positive results were obtained with 100, 10 but not 1 c.c. of water (see columns 7-8, 12-13, 17-18). The bars are here placed on the farthest out + , namely, between columns 12-13. In respect of the specially typical B. coli (red dots on right side of bars) positive results were obtained with 100, 10 but not 1 c.c. of water (see columns 8, 13, 18) so the dots are placed on the farthest out +, namely, in column 13. No attention need be directed to the dots on the left side of the bars, as a record of these need only be taken when the more significant specially typical B. coli test fails. Variation 2. M60 (Sheet 1), SS64, SS67 and SS 72 (Sheet 2). The foregoing description is equally applicable here except that the samples instead of yielding positive results with 100 and 10, only yielded positive ix BACTERIOLOGICAL ROUTINE METHODS 145 results with 100. Hence the dots and bars occur under the columns pertain- ing to 100 c.c. instead of 10 c.c. Variation 3. M68, M75, M123 and M171 (Sheet 1). Here the only difference from Variation 2 is that no specially typical B. coli could be isolated, "sagin" (col. 7) microbes being obtained instead. The red dot therefore occurs in col. 7 and in column 1 is marked a red dot on the right side of a red bar, signifying that no specially typical B. coli were found even in 100 c.c. of water. Variation 4. M27 (Sheet 1). The symbolic representation here is exactly the same as Variation 1, except that in connection with the 100 c.c. culture a "sagin" instead of an "agin" microbe was isolated, so a + has to be entered in column 7 instead of column 8. This, it is true, does not affect the classification of the B. coli results, and therefore the dot and bar pattern is unaffected ; but it does affect the number of specimens (of one or another type) isolated from the different volumes of water dealt with, and of which a record must be kept for a different purpose altogether. Variation 5 M36 (Sheet 1). Here we are dealing with a new pattern because the sample although yielding positive results with the presumptive test in 10 c.c. (black dot column 9) failed to yield any typical B. coli in this amount and indeed was only con- firmed on a lower " gl " plane. Hence the blue bar although placed between cols. 10-11 has the blue dot to the left (col. 10) and we have to fall back on the 100 c.c. culture to obtain an "ag" microbe (see blue dot in col. 6). Further, the 100 c.c. culture had to be fallen back upon to obtain typical B. coli and as it turned out to be an "agin" microbe the red bar is placed between cols. 7-8 with the red dot in col. 8. Variation 6. M140, M147 (Sheet 1) SS. 60 (Sheet 2). The pattern here exactly corresponds to variation 5 except that the failure to isolate an "agin " microbe necessitated the red dot being placed in column 7 instead of 8 and a red bar with right hand red dot in column 1 which may really be regarded as the negative 100 c.c. column. Variation 7. M156 (Sheet 1) SS 68 (Sheet 2). In these samples the failure to isolate typical B. coli of any kind, even in 100 c.c. of water, calls for a red bar, with red dots on either side of it, being placed in column 1. Otherwise the pattern corresponds to variation 2. L 146 STUDIES IN WATER SUPPLY CHAP. Variation 8. M44, M51 (Sheet 1) This corresponds with variation 7 except that as a " gl " instead of an " ag " microbe was isolated, the blue dot comes on the left side of the blue bar (col. 5 instead of 6) and a blue bar with right side blue dot has to be introduced into column 1. Variation 9. M168, M116 (Sheet 1). These samples yielded presumptive positive results in 100 c.c. (black dot column 4) but no confirmation was obtained on sub-culture, so a red triangle is introduced into the picture and the No. 1 column receives a blue and a red bar, each with two dots. Variation 10. M132 (Sheet 1). This sample yielded positive presumptive results in 100 and 10 c.c., so the black dot appears in column 9. Confirmation was not obtained in the 10 c.c. culture, hence the red triangle. A "gl" microbe, however, resulted from the 100 c.c. culture, so a blue bar appears between cols. 5-6 with a blue dot on its left hand side (col. 5) and the No. 1 negative column receives a blue bar with the blue dot on the right side. As no typical B. coli of any kind were found it became necessary to place in column 1 a red bar with red dots on either side of it. Variation 11. M164 (Sheet 1). This sample yielded similar results except that confirmation was obtained in the 10 instead of the 100 c.c. cultures with corresponding alteration of the pattern. Variation 12 SS 65 (Sheet 2). In this sample, both the presumptive and special confirmatory results were positive so far out as the 1 c.c. culture thus leading to the black dot being placed in col. 14, the blue bar between cols. 15-16 and the blue dot in col. 16. In none of the cultures, however, could a specially typical "agin" B. coli be found, hence the red bar with right hand red dot in column 1. A "sagin" microbe resulting from the 100 c.c. culture, a red bar appears between cols. 7-8 and a red dot in col. 7. ix BACTERIOLOGICAL ROUTINE METHODS 147 Variation 13. SS 70 (Sheet 2). Here the presumptive test yielded positive results even in O'l c.c., hence the black dot in col. 19. A " gl "confirmatory result was obtained from the same culture, so a blue bar appears between cols. 20-21 and a blue dot on its left hand side, namely, in column 20. An " ag " confirmatory microbe being isolated from the 1 c.c. culture, a blue dot is placed in col. 16. Typical B. coli was isolated from the 1 c.c. culture, so a red bar appears between cols. 17 and 18 ; but as it was a "sagin" not an "agin" microbe, the red dot necessarily falls into col. 17. As however an "agin" microbe was obtained from the 10 c.c. culture, a red dot is placed in column 13. Variation 14. M84, M92, M99 (Sheet 1) and SS62 (Sheet 2). ' Here the samples yielded completely negative results, as judged by all the tests, so there will be found in the negative No. 1 column a black dot, a blue bar with blue dots on either side of it, and a red bar with red dots on either side -of it. The foregoing description may appear to be very com- plicated ; in practice, however, this method of recording results is not only extremely simple, but it reduces the chances of error practically to nil. Decimal mode of dilution. The author has used for so many years the decimal mode of dilution and has described it so often in Ms reports that the briefest reference need only be made to it here, A series of tubes each containing 9 c.c. of sterile water are inoculated from the sample of water as follows : 1 c.c. of water is added to the first tube (l) 1 c.c. is taken from tube (l) and added to a second tube (2) , 1 c.c. is taken from tube (2) and added to a third tube (3). 1 c.c. is taken from tube (3) and added to a fourth tube (4) and so on according to the quality of sample being dealt with. Thus with sewages, it may be necessary to go as far as the tenth dilution to ensure negative results at the extreme* end of the scale. L 2 i 4 8 STUDIES IN WATER SUPPLY CHAP. 1 c.c. from the various dilutions is then added in the reverse order to appropriate media and their values in terms of the original material are obviously as follows : 1 c.c. of tube (10) = 0-0000000001 of sample ( 9) = 0-000000001 ( 8) = 0-00000001 ( 7) = 0-0000001 6) = 0-000001 5) = 0-00001 ( 4) = O'OOOl ( 3) = 0-00 ( 2) = 0-01 0-001 ( 1) = In addition 1 c.c., 10 c.c., and 100 c.c. cultures are also made and in certain cases the writer has used 1000 and 10,000 c.c. cultures as well. It may be, and indeed has been, objected that there is a wide gap between 100 and 10, between 10 and 1, between 1 and O'l, and so on. This is true, but if reference be made to the author's original reports the suggestion will be found that, if time and media allow, inter- mediate cultures may in special cases be made as follows : 100, 90, 80, 70, 60, 50, 40, 30 and 20 c.c. cultures. 10, 9, 8, 7, 6, 5, 4, 3, and 2 c.c. cultures. 1, 0'9, 0'8, 07, 0'6, 0'5, 0'4, 0*3, 0'2, c.c. cultures, and so on. It is obvious, however, that in routine work this pro- cedure is impossible, nor if it were attempted would the information obtained be likely to prove of much additional value. As some bacteriologists condemn waters on a 1 c.c. basis and others on a 10 c.c. and yet others again on a 100 c.c. basis, it is apparent that with such wide dis- crepancies of opinion prevalent, the time is scarcely ripe for insisting on meticulous refinements. The attempt to do so usually ends in curtailing the range of the cultures, so that the results are all positive with none negative or all negative with no positives. The writer, however, has often EXAMPLE OF HEKTOGRAPH SHEET.- No. 1. Lambeth District,- General Wells. (January, 1907.) B. Coli test. 4 100 c.c. 10 c.c. 1 c.c. Description of Sample. i 8 of Bacteria per c.c. (gelatine at 20-22 C). 1, Confir- matory. Type. w t Confir- matory. Type. c- i Confir- matory. Typ . . . . % g '& i '& M i S gl. g. 1 (9 I * ag. \ 3 | gl. g. Cote. 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 M. 3. 2 23 + + + + 1 + 1 + M. 12. 3 21 + + + + 1 \ 1 \ M. 20. 4 11 + * + + 1 + 1 \ M97 . . At . M. 36. 8 6 + + 1 + ; ^ 1- -|. M. 44. 9 I + +i o 1- ! M. 51. 10 8 + + 1 M. 60. 11 2 + *+ 1 '+ 1. M. 68. 14 2 + ^ \\ 1. M. 75. 15 1 + "+ \\ -I. M. 84. 16 12 . -|. M. 92. 17 2 . -I- M. 99. 18 5 . 1. M. 108. 21 5 X ^ o * M. 116. 22 6 j /* "V o M. 123. 23 1 + \ +1 |. . M. 132. 24 4 . \ + X" x o . M. 140. 25 3 + + +l + +1 . M. 147. 28 4 + ; +i ; ;i o - . M. 156. 29 8 -f + 1- . M. 164. 30 6 + s s^ -f ^.j . M. 171. 31 11 f + ^-1 (22) Samples 3, 5. 9. 10. 16. 10 9 9 8 2 9 8 4- Explanation of Symbols on Hektograph Sheets Nos. 1 & 2. Note. (1.) Furthest out presumptive on positive side, or negative presumptive. (2.) -| Furthest out positive " gl." in 100 o.c. or less. When the furthest out positive confirmatory is an " ag." this symbol ie not required. |3.) |. Furthest out positive " ag." in 100 c.o. or less,or negative "ag." If furthest out positive in this confirmatory test was " gl." then blue dot is placed on furthest out "ag." if present. This completes the symbol [ the bar portion of the symbol having been previously allocated in the "gl." symbol. In the event of no positive "ag." being found in any of the cultures, then symbol l is placed in the negative column, at left side of sample number. 4.) ( Negative "gl." and "ag." 5.) 'I Furthest out positive "sagin " in 100 c.o. or less. When the furthest out positive typical is an " agin " this symbol is not required. 6.) ! Furthest out positive "agin " in 100 c.c. or less, or negative "agin." If furthest out positive in this typical test was a " sagin " then red dot is placed on .the furthest out " agin " if present. This completes the symbol '! , the bar portion of the symbol having been pre- viously allocated in the " sagin " symbol. In the event of no positive " agin " being found in any of the cultures, the symbol ! is placed in the negative column, at left hand side of the sample number. 7.) M- Negative "sagin" and "agin." 8.) XX = Failure. This indicates that although the primary culture formed acid and gas, failure was experienced in the attempt to isolate coli- like microbes from secondary cultures. 9.) Where the results were negative in 100 c.c. or less, as regards these tests, the symbols occur on the negative side, at the left of the sample number. Each of the symbols must occur in number, corresponding to the number of samples examined. 10.) PRESUMPTIVE TEST: Add number of black dots in each column on positive side, and number on negative side. 11.) CONFIRMATORY TEST. " Gl." and " ag." basis : Add number of blue bars I in each column on positive side and number of *! symbols on negative side. " Ag." basis : Add number of in each column on positive side ("ag." column), and number of ! and '! on negative side. 12.) TYPICAL TEST. u Sagin " and " Agin " basis : Add number of red bars I in each column on positive side and number of "I- on negative side. " Agin " basis : Add number of in each column on positive side, ( ll aon-n " fif\\-nmr\\ anrl -niTml^v rvP on/? !. r 2321 EXAMPLE OF HEKTOGRAPH SHEET. No. 2. River Thames water after Storage in the Staines Reservoirs. (June, 1909.) 17510 Number of Microbes per c.c B Coli teat. .a *" 1 100 c.c. 10 c.c. 1 C.C. 1 c.c. 01 c.c. fl 4 Description of Sample. ' s i Gelatine at 20-22 c. coun on third day. Agar, at 87 c. counle 20-24 hours. Eebipelagar, at 37 c. count* 20-24 hours. Presumptive. Confir- matory. Type. Presumptive. Confir- matory. Type. Presumptive. Js! 8 Type. Presumptive. Confir- matory. Type. Presnraptive. V, I i T, *-> I AGIH. 1 i o d AGIN. 1 f AGIK. a gl ag. gl ag. gl. I gl- ag gl. I Cola. 1 2 8A 3B 3C 4 5 7 s 9 10 11 12 13 14 15 17 18 19 20 21 22 23 ti 25 26 27 a . 1. ^.8. 60. 3 220 20 5 + + + -rJ >! -1. S.s. 62. 7 270 120 7 S.s. 64. 10 210 34 -f + f I- S.s. 65. 14 200 22 1 + + + + + + 1 S.s. 67. 17 72 4 + I + + 1. S.s. 68. 21 120 19 + I + S.s. 70. 24 340 88 4 + + + 4 + f + +1 + '4 S.s. 72.- 28 80 10 f 1 + -f . *! * 1 I- i 1 1 1 1 1 1 1 i 1 M. 5 2 1. 5 1 3 I ix BACTERIOLOGICAL ROUTINE METHODS 149 made ten 10 c.c. cultures instead of one 10 c.c. and one 100 c.c. culture, and obtained useful results, although it does not necessarily follow that this is really equivalent bacteriologically to making 100, 90, 80, 70, 60, 50, 40, 30, 20, and 10 c.c. cultures. Racks, coloured wools, labels and media, wires, etc, It is desirable, as is done at the Metropolitan Water Board Laboratories, to set up racks (Fig. 30) containing the various dilutions and media. Coloured labels are very convenient, e.g., blue for 100, brown for ten, green for 1, pink for O'l, yellow for O'Ol red for 0*001, and maroon for O'OOOl c.c. cultures, and so on. Coloured wools should also be used to distinguish the various media, or the media may be tinted with litmus, neutral red, etc. Beads also of various colours may be used in the media itself. Instead of platinum wires and loops, iron wires may be used with great advantage. These may be sterilised in bulk, and their use saves a great deal of time in sub- cultural work. The straight iron wires used by florists and sold in bundles are best for this purpose (about 175 mm. long and about 1 mm. thick). With a suitable pair of round-nosed pliers, it is easy to make a loop at one end of a number of these wires, and these loop wires then serve admirably for all purposes where loop " cultures " are required. CHAPTER X. BACTERIOLOGICAL ROUTINE METHODS (continued). IN this chapter will be found a detailed description of the exact method followed in the examination of a sample of raw river water. The work to be carried out each day is fully explained, and the composition of the media used in connection with the various tests is also given. Examination of a Sample of Water. First day. It will perhaps be best to take the examina- tion of a sample of raw river water as an example, it being understood that when dealing with more impure samples the dilutions, etc., must be carried further, the converse holding good in the case of filtered water samples. The rack shown in Fig. 30 (p. 152) may be used for this examination, and the dilutions, culture tubes, etc., arranged in it in the way explained in the accompanying description. At this stage it may be convenient to describe a piece of apparatus used for keeping ordinary agar and bile-salt-agar cultures at the proper temperature so as to be available at any time for " pouring " purposes. To start with, the tubes are first placed in the retainer shown in Fig. 31 (p. 153) ; this is then lifted out of the water-bath, and placed in another copper vessel, containing a sufficiency of water, and heat applied by means of a Fletcher burner. After the water has boiled foi some time so as to ensure that all the 160 CH. x ROUTINE METHODS 151 agar in the tubes has been liquefied, the water is cooled down (but not below 50 C.), the retainer is lifted out and transferred back to the copper water-bath (see Fig. 31), the water in which is maintained at a temperature of 45 C. by means of a capsule which, by its expansion and contraction, governs the supply of gas to the bunsen used for heating purposes. Before very long the contents of the tubes and the surrounding water become of the same temperature of 45 C., and the tubes are then always ready for " pouring " purposes, either on the same day or subsequent days. One large and six small sterile Petri dishes are placed conveniently on the laboratory bench and labelled as follows : 1 large dish with brown 10 c.c. label for bile-salt-agar. 2 small dishes with green 1 c.c. labels (1 for ordinary agar and 1 for bile-salt-agar. 2 small dishes with pink O'l c.c. labels (Ifor ordinary agar and Ifor gelatine). 1 small dish with yellow O'Ol c.c. label (for gelatine). 1 small dish with red O'OOl label (for gelatine). After first shaking the sample bottle, the lip is flamed, the stopper withdrawn and held in such a way as to avoid contamination, or it may be placed temporarily on any sterile surface. If reference be made to Fig. 32 (p. 155), the following description will be easily followed. One c.c. of water is withdrawn from the bottle by means of a sterile pipette and transferred to dilution (l), the tube being moved one place to the left. One c.c. from dilution (1) is then transferred to dilution (2), 1 c.c. from (2) to (3), and, lastly, 1 c.c. from (3) to (4), the tubes on each occasion being shifted one place to the left (see 2 in Fig. 32). One c.c. amounts from dilution (4), representing O'OOOl c.c. of the sample, are next taken and added to the tube behind it in the rack, namely, the O'OOOl bile-salt- peptone tube, with maroon label, and whilst this is being 152 STUDIES IN WATER SUPPLY CHAP. o o $ Row 4. The open circles are empty and Row 3. the dark circles contain tubes. o o Row 2. o Row 1. FIG. 30. Rack, as set up for the Examination of a Sample of Raw River Water. As seen in plan. Heading always from left to right. Row. 1. First hole empty. The remaining four contain 9 c.c. dilution tubes. Row 2. First hole contains a 10 c.c. double strength bile-salt peptone tube (blue wool). The 2nd hole is empty. The 3rd, 4th, 5th, 6th, and 7th holes contain 10 c.c. single strength bile-salt peptone tubes (pink wool) for the 1, 0-1, 01, O'OOl, and 0001 c.c. cultures. B. coli test. The 8th hole is empty. Row 3. The 1st, 2nd, and 3rd holes contain gelatine tubes for the (H, O'Ol, and 0*001 c.c. cultures. Test for numbers. The 4th hole is empty. Row 4. First hole contains a 40 c.c. milk tube for 10 c.c. culture. The second hole is empty. The 3rd and 4th holes contain 10 c.c. milk tubes for the 1 and O'l c.c. cultures. The milk tubes are previously boiled to expel oxygen and then rapidly cooled in cold water. These are for B. enteritidis sporogenes test. The last hole is empty. The big hole (right top corner) contains a 50 c.c. treble strength bile-salt-peptone tube, and is for the 100 c.c. B. coli culture. The 100 c.c. label is blue, and the rest as follows: 10 c.c. (brown), 1 c.c. (green), O'l c.c. (pink), O'Ol c.c. (yellow), O'OOl c.c. (red), and O'OOOl c.c. (maroon). The ordinary agar and bile-salt-agar cultures are made by adding the water direct to the plates and subsequently adding the melted media (see text). x BACTERIOLOGICAL ROUTINE METHODS 153 done each tube is shifted one place to the right so as to reduce to an absolute minimum any possibility of a mistake (see 3 in Fig. 32). One c.c. amounts from dilution (3), representing 0*001 c.c. of the sample, are next taken and added successively to the tubes behind it in the rack, namely, the 0*001 bile- salt-peptone and gelatine tubes with red labels, the shifting from left to right being again practised (see 4 in Fig. 32). O o o o o oo oSpo q jJspo o|^oo o [^ SC COPPER / WATER BATH FIG. 31. Apparatus for Agar Tubes. One c.c. amounts from dilution (2), representing 0*01 c.c. of the sample, are next taken and added successively to the appropriate tubes, namely, the 0*01 c.c. bile-salt-peptone and gelatine tubes with yellow labels, the shifting process being repeated (see 5 in Fig. 32). One c.c. amounts from dilution (1), representing 0*1 c.c. of the sample, are then added successively to the 01 c.c. bile-salt-peptone, gelatine, and milk tubes with pink labels, with the usual shifting operations. In addition, 1 c.c. from the same dilution ( = 0*1 c.c. of sample) is added to 154 STUDIES IN WATER SUPPLY CHAP. the agar plate with the pink 0*1 c.c. label (see 6 in Fig. 32). We have now come to the end of the dilutions, and have to deal directly with the water in the sample bottle. 1 c.c. amounts are added successively to the 1 c.c. bile- salt-peptone, gelatine, and milk tubes with green labels, each tube, as before, being shifted from left to right as soon as it has been inoculated. In addition, 1 c.c. amounts are added to two of the plates, bearing the green labels, one for ordinary agar and the other for bile-salt agar (see 7 in Fig. 32). Next, 10 c.c. amounts of water are added successively to the 10 c.c. bile-salt-peptone (double strength) and milk tubes bearing the brown labels and shifted from left to right, and as well, 10 c.c. are added to the plate with the brown label for bile-salt-agar (see 8 in Fig. 32). Lastly, the single 100 c.c. (blue label) culture is made, by pouring the water out of the sample bottle into the tube until it reaches the file mark. The file mark is so placed that the capacity of the tube between the level of the bile-salt-peptone medium (50 c.c. of treble strength) and the file mark is 100 c.c. The appearance of the rack at each stage of the process is shown in the Fig. 32 (p. 155). The contents of the bile-salt-agar and ordinary agar tubes, maintained at 45C., are next poured into the Petri dishes already inoculated with water, and by means of a circular and swaying motion, the medium and the water thoroughly mixed together. After the medium has " set " the following drying operations are required : The covers are tilted by means of suitably twisted wire and placed upside down in the incubator at 37C. until ready for counting- The milk tubes are placed in a copper vessel containing water at a temperature of 80 C. and left for ten to twenty minutes' heating at 80C., cooled and then placed in the incubator at 37C. for two days. x BACTERIOLOGICAL ROUTINE METHODS 155 1. O*o o :: o o o 2. o o o 090 o o o 3. C o o OO O * o 4. C o o 0*0 0*0 8. FIG. 32. Rack, during progress of Examination of a Raw River Water. As seen in plan. The open circles are empty and the dark circles contain tubes. 1. Rack, as set up. 2. After making dilutions. 3. After making O'OOOl c.c. culture. 4. After making '001 c.c. cultures. 5. Af ter making '01 c.c. cultures. 6. After making O'l c.c. cultures. 7. After making 1 c.c. cultures. 8. After making 10 c.c. cultures. There remains only the making of the 100 c.c. cultures. 156 STUDIES IN WATER SUPPLY CHAP. The gelatine tubes are placed in warm water (about 37C.) until the gelatine is all melted. After shaking, the contents are poured into the Petri dishes waiting to receive them, and which have been already labelled in corres- pondence. After the gelatine has " set," the plates are placed in the cool incubator (20-22 C.) for three days, after which the colonies are counted. The liquid bile-salt cultures are placed in the hot incubator; some bacteriologists prefer a temperature of 37C. whilst others rely on 42C., others, yet again, on some intermediate temperature. This concludes the work for the first day, apart, of course, from the work falling due, on account of samples previously examined. The second day's work (1st day after) may be described as follows : The colonies in the bile-salt-agar and ordinary agar plates are counted. The results are expressed per c.c., so that if the count is made from a plate bearing the 10 c.c. brown label the numbers must obviously be divided by ten. If the count is from a plate with a 1 c.c. green label, the numbers stand as counted. Again, if the count is from a plate having a O'l c.c. pink label, the numbers must be multiplied by ten, and so on. It is obvious that attention to the simple pre- cautions now being described greatly reduces the chances of error. It may be objected that counting after one day is un- desirable inasmuch as all the colonies may not have had sufficient time to develop. On the other hand, if the ordinary agar plates are not counted at this early stage, there is grave risk, despite all drying operations, of the sporing bacteria spreading all over the plates, and render- ing accurate counting altogether impossible. No such danger occurs in connection with the bile-salt- agar plates, but it is not unimportant for comparative purposes that the ordinary agar and bile-salt-agar plates x BACTERIOLOGICAL ROUTINE METHODS 157 should both be counted on the same day, and we do know that with sewage-polluted waters the excremental bacteria really require no longer period for their development. Whether any further work falls due on this day depends on circumstances. If for the B. coli work we are dealing with glucose or lactose bile-salt cultures, which require 48 hours 1 in the incubator, there is none, but if we are using the " drop " glucose method associated with a lactose bile- salt medium, then the following procedure is adopted. All the lactose-bile-salt liquid cultures are removed temporarily from the incubator, which in connection with this method is kept at 42 C. Those tubes showing no redness or no development of gas associated with the redness, even on smartly tapping the tubes with a flat piece of wood, receive one or more drops of a concentrated sterile glucose solution added by means of a sterile pipette and are then placed in another incubator maintained at 37 C. The idea underlying this method is either that enfeebled lactose fermenting microbes may be coaxed into activity in the presence of glucose and a lower temperature or, failing this, that the glucose fermenters are, as a last resort, given a chance of asserting themselves. Speaking very broadly, this method is perhaps most useful during the colder months of the year, say October to March. The remaining tubes, namely, those showing gas formation, are replaced in the 42 C. incubator. The third day's work (2nd day after) will now be described. The milk tubes are examined and those show- 1 Some bacteriologists prefer 20-24 hours or even less. If the writer curtailed the incubation period in the case of the London waters he might lay himself open to the charge of using methods whicn appreciably affected the results in the direction of placing too favourable an opinion on the quality of the Metropolitan Water Supply. 158 STUDIES IN WATER SUPPLY CHAP. ing the B. enteritidis sporogenes phenomenon (gas form- ation, separation of the whey from the casein with precipitation of the latter) are registered as positive results. It may be objected that a variety of microbes may produce these changes, arid the writer has always frankly admitted that he regards the test primarily as a test, and not necessarily as proving the presence of a particular microbe, and that microbe alone. The above changes in milk cultures were described by Klein as characteristically typical of the growth of B. enteritidis sporogenes, and so it is both complimentary and convenient to retain the name, although the word test should be added as signifying that, in recording a positive result, one is not necessarily committed to the view that the characteristic changes are produced solely by one microbe, and that the one described by Klein. The prac- tically important point is that sewages commonly yield a positive result with O'OOl c.c., and pure water a negative result with 10 c.c. or even more. The B. coli cultures, whether from a glucose or lactose primary medium (or lactose supplemented with glucose after 24 hours' incubation), are now removed from the incubator, and all those showing gas formation are entered as " presumptive positives." It is desirable to tap smartly with a flat piece of wood all tubes which show redness but no apparent gas formation, as a small proportion of these may be induced to show visible bubbling. Each one of the presumptive positive tubes is next subcultured as follows : A loopful from the tube is transferred by means of a sterile loop wire into 10 c.c. of sterile water. Alternatively, a straight wire may be used if experience shows that the loop carries with it too much of the culture. In either case, a sterile loop is used to inoculate two slops bile-salt- neutral-red-agar tubes, one of which also contains x BACTERIOLOGICAL ROUTINE METHODS 159 lactose and the other cane-sugar. The loop should not be allowed to touch the agar until it has reached the base of the slope ; the wire is then worked upwards towards the operator slowly and in a to and fro fashion, which ensures separate colonies if carried out with reasonable care. Of course, plate cultures may be used instead, but the slope cultures are so convenient that no one who has once fairly tried them is ever likely to abandon their use. In making these cultures, it is desirable to use special racks and to shift the tubes one place as soon as each one has been inoculated. Separate racks should be provided for the 100, the 10, and the 1 c.c. (and less amounts) cultures. It is unnecessary to illustrate these racks, as their exact shape and the size of the holes for the tubes, are matters which may be left to the judgment of bacteriologists individually. It is im- portant to note that the agar slope cultures should be labelled with blue, brown, green, etc., labels in corres- pondence with the "presumptive positives" from which they were derived. After inoculation, the agar slope cultures are placed in the incubator at 37 C. The fourth day's work (third day after) comprises the following :- The gelatine plates are counted and the results expressed as number of colonies (or microbes) per c.c. of the sample. The question whether the numbers counted are to be divided by ten or remain as they are, or be multi- divided by ten, one hundred, or more, involves no possible error, as the labels being of a distinctive colour at once determine this point. Many bacteriologists prefer to count the plates after four, five or more days' incubation. This, however, involves risk of the plates being utterly ruined for counting purposes by the development and spread of liquefying colonies. The lactose and cane-sugar slope cultures are next 160 STUDIES IN WATER SUPPLY CHAP. examined, the object being to secure one or more red colonies (if present) from the lactose tube and one or more colourless colonies (if present) from the cane-sugar tube. The reason for this procedure is that the red colonies on the lactose tube are likely to be, not only " acid-producers," but " gas formers " in a lactose medium, and the colourless colonies on the cane-sugar tube may possibly be non-gas producers in a cane-sugar medium, a negative property associated with the B. coli communis as first described by Escherich. It might perhaps be surmised that this involved risk of losing a lactose fermenter, but the writer's experience is that as many lactose fermenters are derived from the cane-sugar slopes as from the lactose slopes, and there is an increase in the chances of isolating a microbe corresponding to B. coli communis (Escherich). Of course, if the lactose slope shows only colourless colonies, and the cane-sugar slope only coloured colonies, these are sub-cultured, and at the worst one is almost certain to secure at least a glucose fermenter from such slopes. The number and variety of sub-cultures to be made from the colonies depends on circumstances and the particular views of the individual bacteriologist, but the writer suggests that to save waste of time and to avoid oppor- tunities for error, the principle of his automatic method of inoculation should be practised. 1 This depends essentially on touching the colony to be sub- cultured only once with a sterile iron wire, transferring the now inoculated wire into a small tube, containing a drop of saline solution, subsequently adding to that tube as many more sterile wires (all of which will be auto- matically inoculated) as it is desired to make cultures. For example, if it is desired to follow the author's flaginac 1 Described in the Author's Report on the Condition of the Metropolitan Water Supply during the month of January, 1907. Published by the Government Printers. x BACTERIOLOGICAL ROUTINE METHODS 161 method, four extra wires would be required for the following cultures : Neutral-red-broth ... ... ... fl. = fluorescence. Lactose peptone... ... ... ... ag. = acid and gas. Peptone water in. = indol. Litmus milk ... ... ... ... ac. = acid clot. Cane-sugar-peptone cultures may be used as well and the symbols C.S.N. or C.S.P. added, according to whether or not there is gas formation. According to this classification, a flaginac microbe is ac- cepted as a typical B. coli or as a microbe characteristic of excremental matter. Inasmuch as a majority of human fsecal B. coli either do not ferment cane-sugar at all or act on it but feebly, there would seem to be at least arguable reasons for considering a C.S.N. microbe more objection- able than a C.S.P. microbe. Of course, one could obviously use a peptone water tube instead of a saline tube (as used in the automatic inocula- tion method) and incubate this until multiplication had proceeded far enough to allow of further sub-cultures being made from it, but this involves practically a wasted day and it is assumed here that any unnecessary delay should be avoided. The method at present used by the writer may now be described, and one of its essential features is that gelatine, solid media instead of liquid media are used, the chief reason being that the former are much more delicate and rapid indicators of gas production. The two little saline tubes are taken from their parent tube (see Fig. 33, p. 163) and placed in the holes pro- vided in the rack shown in the same figure. A red (by preference) colony from the bile-salt-agar lactose slope is picked off with a sterile iron wire, and the wire placed in the saline tube on the left-hand side (a) and "twirled" round so as to spread the bacilli throughout the medium. M 1 62 STUDIES IN WATER SUPPLY CHAP. Similarly, a white (by preference) colony from the bile- salt-agar-cane- sugar slope is transferred to the other saline tube (b). Two extra iron wires are placed in (a) and three in (b). The two little peptone tubes are next taken from their parent tube and one marked by means of a red grease pencil with a transverse bar, and inoculated with one of the wires from the (a) saline, the other one is inoculated from (b) and relates of course to the cane-sugar slope, just as the (a) one belongs to the lactose slope. They are then returned to their parent tube. The five little gelatine tubes are then removed from their parent tube (see Fig. 33) which contains one tube of glucose (not coloured), two tubes of lactose (both coloured with litmus), one of cane-sugar (coloured with neutral red), and one of dulcite (coloured with bismarck brown). These are then inoculated as follows : One of the lactose tubes is " barred " to correspond with the " barred " peptone tube for indol and receives one of the wires from the (a) saline tube, and the remaining wire is used to inoculate the glucose tube. The three wires from the (b) saline tube are placed separately in the cane-sugar, dulcite, and remaining lactose tubes. All the five wires are then withdrawn from the five gelatine tubes, which are returned to their parent tube and incubated at 37 C. for three hours, and thereafter at 20-22 C. for one or two days. The peptone tubes are, of course, incubated for a similar period but throughout at 37 C. In practice, it is customary to inoculate a whole series of little saline tubes from the slope cultures and to reverse (a) with (b), for the reason that when the rack is full it is given a complete half- turn, which, as it were, undoes the reversal and enables the observer to work clear of an army of wires which otherwise would prove a source of em- barrassment. Special racks are provided for the slope cultures and x BACTERIOLOGICAL ROUTINE METHODS 163 a . a PARENT TUBE I OR n ALLY UYino INSIPE. 'ARthT Tu&E_ J TUB E: .... -- - ^ -. '"r J 3 U2 1^5 s 'jJ I of 3 U) tf} jl >nt ^Q iS x o "I 55 < p, JU i 3 5 r D <0 o 1 NORMALLY BAttDElp TOGETHER AMP LYI/^GIM PARE1MT Tu B FIG. 33.- B. Coli Tests (The Parent Tubes in Practice, are plugged with Cotton Wool.) M 2 164 STUDIES IN WATER SUPPLY CHAP. the parent saline peptone and gelatine tubes, and the same principle of shifting a space as each inoculation process is completed is resorted to as has been previously explained. It is far quicker to work in this methodical manner, and it reduces the chances of a mistake to a minimum. On the fifth or sixth day (4 or 5 days after) the final records are obtained. The fifth day (one day so far as these particular tests are concerned) is certainly not too soon for the glucose, lactose, and cane-sugar results, and suffices, although only barely, for the indol test, but is doubtfully long enough for the dulcite test. In a labora- ,tory where much water work is undertaken, it is desirable to record all the results of the sub-cultural tests on the same day and at the earliest moment ; and many, for these reasons, may decide to leave out such tests as demand extra time, particularly when those tests do not after all yield information materially affecting the judgment of the operator, in pronouncing on the hygienic qualities of the water under examination. This is a matter which must be left to the individual bacteriologist, although most practical workers will agree that those who have suggested a great variety of tests have not brought forward convincingly any intelligible classification enabling one to grade the microbes isolated and studied in the order of their undesirability in a water supply. At the Water Board Laboratories, the tests are carried out on the fifth day in cases of special urgency and on the sixth day if a rapid diagnosis is not required. The peptone water cultures are tested for indol by the para- dimethylamidobenzaldehyde method, the paradimethyl- amidobenzaldehyde solution being added first and then the potassium persulphate solution, a pink colour develop- ing after the lapse of a few minutes, indicating the presence of indol (see page 173). As regards the gelatine tubes, the presence of gas is x BACTERIOLOGICAL ROUTINE METHODS 165 readily seen owing to the medium being split up and fissured with gas bubbles. The results are then classified as under . Typical B. co,i ^ A) } Saccharose + ......... (^ C = lactose + indol + ; other tests not determined. 1 = glucose + , lactose 0. 2 <= lactose + , indol 0. + = gas (or indol). = No gas (or no indol). This is shown diagrammatically in Fig. 34 : B coli test:. 0-rw gas (or no indol) (\ = glucose + lactose 0) = lactose + indol 0) Typical J3 Coti (Lactose +, indol+) C- lactose-^. indol +, other tests net determined. Saccharose + . B s ^Saccharose 0. _ , , _ _ n r ~i A = dulcit 0. (A}= dulcib + B - dulcit 0. (B) =dalcib -h . FIG. 34. As a matter of practice, the rule is to glance at the (6) peptone tube first, and if this shows indol, to see whether the non-barred lactose tube is positive. If positive, it remains only to look at the saccharose and dulcite results to determine whether the microbe is to be classed as A (A), B or (B). If however, the lactose result is negative, attention is next turned to the (a) indol tube and its corresponding " barred " lactose tube. If both are positive the entry is C, which implies a typical B. coli, but one the characters of which, as regards saccharose and dulcite, remain undetermined. It very rarely happens that resort has to be made to this entry owing to a failure to isolate an A, (A), B, or 1 66 STUDIES IN WATER SUPPLY CHAP. (B) microbe. Suppose however, the (6) peptone tube had been negative, attention would equally be directed to the (a) series in search for a (C) microbe. But if all these tests failed, then both the lactose tubes would be examined and if either yielded a positive result the entry would be classed as " 2." This also having failed, the glucose tube would be observed, and the result if positive would be entered as " 1." It will be gathered that the object is (l) to isolate if possible a " lactose + indol + microbe " with determined characters as regards saccharose and dulcite. (2) Alterna- tively to isolate a " lactose 4- indol + microbe," but with undetermined characters in respect of saccharose and dulcite. (3) Failing these to obtain a " lactose + microbe," and (4) lastly, as a final resort, to fall back upon a " glucose + microbe." Figs. 35 and 36 (pp. 167, 169) show the results of the application of the B. coli test to the London waters. For full information on " the varieties and significance of B. coli in water supplies," the reader is referred to a paper bearing this title read by the writer at the annual meet- ing of the British Medical Association, 1912 (British Medical Journal, Sept. 21, 1912). It may be worth while here to reproduce the summary of conclusions : Typical B. coli (lactose + , indol + ) is present in enormous numbers in excremental matters, and is absent from, or present only in small numbers in, substances free from undesirable pollution. Typical B. coli is a decadent microbe when divorced from the animal body ; hence its presence in a water in any number probably points to fairly recent pollution. Pure waters, generally speaking, contain no typical B. coli in 100 c.c. in a majority of representative samples ; and incidentally it may be stated that impure waters can be so purified as to yield similar results at a not unreason- able cost. Even so lenient a standard as "no typical B. coli in 1 c.c." of water in a majority of representative samples implies that the supply is not habitually contaminated with one gallon of sewage or its bacteriological equivalent, even in 100,000 gallons of water. The 10 c.c. and 100 c.c. standards (that is, no B. coli in these amounts) obviously connote the absence of j 1 ^ and x BACTERIOLOGICAL ROUTINE METHODS 167 I 1 68 STUDIES IN WATER SUPPLY CHAP. T ^ gallon of sewage respectively from a like volume of water. With too severe standards there is danger of condemning reasonably safe supplies, and with too lax standards there is the possibility of passing waters which ouyht properly to be condemned. Always to steer safely between the Scylla of the one and the Charybdis of the other may be difficult or impossible without the aid of a pilot familiar with all the local conditions. The B. coli test ought primarily to be regarded as a quantitative decimal enumeration of "lactose +, indol + " microbes, but the subsequent group- ing, according to certain attributes of the organisms thus obtained, into fairly stable or apparent varieties, may be of real practical or diagnostic importance. Standards are worse, than useless if they are not interpreted with discretion and in relation to local and other conditions. Instances could be quoted where even similar B. coli results ought properly to lead to dissimilar con- clusions as regards quality and safety. The B. coli test even alone is of the greatest value, but a water supply should be finally judged on a summation of verdicts (geological, topographical, physical, bacteriological, and chemical). The B. coli test is by far the most reliable and speediest method of judging the degree of efficiency of the particular water purification process under investigation, and when a sterilisation treatment is in operation, the certified destruction of B. coli should afford absolute proof, practically speaking, of the devitalisation of all the microbes of epidemic water-borne disease. The B. coli test having now survived the various vicissitudes of an earlier time, stands to-day as the most practical and delicate and rapid test for ex- cremental filth, and may surely be taken as the most reliable indicator in its positive aspects of possible danger, and in its negative aspects of the almost certain absence of microbes associated with epidemic water-borne disease, and, generally speaking, as the one test which, above all others, it is least excusable for a water analyst to omit. All the counts in the indictment charged against the B. coli test when fully marshalled amount to no more than this: the test is only, or mostly, of relative, not absolute, value, and therefore the results should be interpreted with discretion. The working schedule may be summarised as follows : First Day : Examination of Sample. Second Day (first day after). Agar and bile -sal t-agar plates counted and glucose added to the lactose B. coli cultures if this method is in use. Third Day (second day after}. Presumptive B.coli and B. enteritidis sporogenes results recorded, and bile- salt-agar slopes made. Fourth Day (third day after}. Gelatine plates counted and colonies from slopes sub-cultured, x BACTERIOLOGICAL ROUTINE METHODS 169 /to /JO '20 //O /oo 93 80 70 60 SV 40 30 20 so O /O 2O 30 K) SO 60 70 SO 90' /OO //O FIG. 36. B. Coli Test. London Waters. Years 1910-11. 14, 149 Samples. i yo STUDIES IN WATER SUPPLY CHAP. Fifth or Sixth Day (fourth or fifth after). Final B. coli results placed on record. The composition of the media used in connection with the routine work may be of interest and is here set out as follows : COMPOSITION OF MEDIA. Enumeration of microbes (plate cultures). Gelatine. (a) Standard. Gelatine (yellow gold label) 12 per cent. Peptone (Witte) 1 per cent. Sodium Chloride 0'5 per cent. Beef broth (1 Ib. beef to 1,000 c.c. of water) to 100 c.c. (b) Modified. Gelatine 12 per cent. Peptone 1 per cent. Lemco 0'5 per cent. Beef broth 25 c.c. Water to 100 c.c. Made faintly alkaline, till blue litmus paper is rendered slightly more blue, by means of a 20 per cent, solution of potassium hydroxide. About 10 c.c. in test-tubes 6" x f" with white wool plug for 1 c.c. amounts of water, and 40 c.c. in test-tubes 8" x I" with white wool plug for 10 c.c. amounts of water. Agar Agar 2 per cent. Peptone (Witte) 1 per cent. Sodium Chloride 0*5 per cent. Beef broth to 100 c.c. Make slightly alkaline by means of a 20 per cent, solution of potassium hydroxide 40 c.c. in test-tubes 8" x 1", with white wool plug, for use in Petri dishes 4j" diameter. x BACTERIOLOGICAL ROUTINE METHODS 171 Rebipelagar 1 (neutral red, bile-salt, peptone, lactose, agar). Agar 2 per cent. Peptone 2 per cent. Bile salt 0*5 per cent. Lactose 1 per cent. Water tinted with neutral red up to 100 c.c. (4 c.c. of a 1 per cent, sterile aqueous solution of neutral red per litre of medium.) No alkali required. About 40 c.c. in test-tubes 8" x l" with white wool plug. B. Coli test PRIMARY CULTURES Dilution Water (decimal method) 9 c.c. of water in tubes 6" x f ", tinted with fluorescin (l part in a million). White wool plugs. Bile salt peptone water (MacConkey) Peptone (Witte) 2 per cent. Bile salt 0'5 per cent. Lactose 1 *0 per cent. Tinted with 10 c.c. of a 10 per cent, sterile litmus solution. Water to 100 c.c. Above is single strength for 1 c.c. cultures. About 10 c.c. of medium in tubes 6" x f " with inverted inner tube 2" x J". Pink cotton wool plug. Double strength for 10 c.c. cultures: Use twice the percentage of above solid constituents with same size of tubes, but white wool for plugs. 1 This term is merely used for convenience ; it will be understood that to MacConkey belongs the credit of the introduction of bile-salt media. 1 72 STUDIES IN WATER SUPPLY CHAP. Triple strength for 100 c.c. cultures : Three times the percentage of above solid constituents. 50 c.c. of medium in tubes 8" x lj", with inner tube 3" x I". White wool plug, covered with sterile paper. SECONDARY CULTURES Dilution water. About 9 c.c. distilled water in tubes 6" x f ". White wool plug. Rebipelagar. see supra. Medium sloped in tubes 6" x I". White wool plug. Rebipesagar. Similar composition to rebipelagar, 1 per cent, saccharose being substituted for 1 per cent, lactose. Medium sloped in tubes 6" x I". Ked wool plug. TYPE OF B. COLL Attributes determined by (A) " FLAGINAC " TEST, with additional glucose and cane-sugar test. fl = fluorescence within two days at 37 C. in neutral red broth cultures. Peptone 1 per cent., sodium chloride 0*5 per cent., beef broth to 100 c.c. [2 c.c. of a sterile 1 per cent, aqueous solution of neutral red per litre of medium]. Make faintly alkaline with a 20 per cent, solution of potassium hydroxide. Tubes 6" x 1", inner tube 2" x ". White wool plug. a<7 = acid and gas within two days at 37 C. in lactose peptone cultures. Peptone 2 per cent., lactose 1 per cent., 10 c.c. of a 10 per cent, litmus solution, water to 100 c.c. Tubes 6" x f", inner tube 2"xf. Blue wool plug. in = mdol within five days at 37 C. in peptone water cultures. Peptone 1 per cent., sodium chloride 0*5 per cent., water to 100 c.c. 1 c.c. of medium in tubes 6" x -f ". White wool plug. x BACTERIOLOGICAL ROUTINE METHODS 173 ac = acid and clot within five days at 37 C. in litmus milk cultures. 10 c.c. of milk tinted with litmus. Tubes 6" x f ". White wool plug. In addition the two following tests may be applied : gl = gas in glucose gelatine (shake cultures). Standard gelatine plus 1 per cent, glucose. 8 c.c. in tubes 6 " x f " white wool plug. cs = acid and gas within two days at 37 C. in cane sugar peptone cultures. Peptone 2 per cent., saccharose 1 per cent., 10 c.c. of a 10 per cent, litmus solution. Water to 100 c.c. Tubes 6" x f ", inner tube 2" x J". Red wool plug. (B) "GLAGINS" TEST, with additional dulcite test. One wide tube 3" x 1", with white wool plug, containing four small tubes 2" x J" unplugged. Saline solution : Two small tubes each containing one drop of 0*5 per cent, sodium chloride solution. Peptone water medium for indol test : Two small tubes each containing about 0*5 c.c. of medium. Peptone 1 per cent., sodium chloride 0*5 per cent. water to 100 c.c. Composition of solutions used for testing for indol (Bohme, Centr.f. Bakt. Bd. XL. 1905) :- (1) 8 grammes paradimethylamidobenzaldehyde, 160 c.c. hydrochloric acid, 760 c.c. absolute alcohol. (2) Saturated cold water solution of potassium per- sulphate. One wide tube 3" x I", containing five small tubes (2" x ") of gelatine sugar media of the following stock composition. Peptone 2 per cent. ; gelatine 7 '5 per cent. ; 1 c.c. of a 5 per cent, solution of potassium hydroxide per 100 c.c. of medium. 174 STUDIES IN WATER SUPPLY CH. x Glucose gelatine medium : One small tube of above stock medium to which 1 per cent, glucose is added. Medium not tinted. Lactose gelatine medium : Two small tubes containing stock medium to which 1 per cent, lactose has been added. Medium tinted with litmus solution. Saccharose gelatine medium : One small tube contain- ing stock medium to which 1 per cent, saccharose has been added. Medium tinted with neutral-red. Dulcite gelatine medium: One small tube containing stock medium to which 0'5 per cent, dulcite has been added. Medium tinted with bismarck brown. B. Enteritidis Sporogenes test. 10 c.c. of whole milk in 6" x %' tubes } for 1 c.c. cultures. , A c i i -n o" -,'/ , i r White wool plugs. 40 c.c. of whole milk in 8 x 1 tubes I for 10 c.c. cultures. CHAPTER XI BACTERIOLOGICAL SPECIAL METHODS IT will be convenient next to describe some of the special methods used by the writer in searching for such patho- genic microbes as the cholera vibrio, the typhoid bacillus, and Gartner's bacillus in samples of water. The Cholera Vibrio. No serious cholera epidemic has occurred in this country for many years, and the isolated cases of this disease have been few in number. Although the protection of our water supplies may be thoroughly safeguarded owing to the vigilance of the Local Govern- ment Board, Port Sanitary Authorities, Rivers Boards, and Medical Officers of Health, yet the ravages of cholera, in the past forbid the belief that our insular position and temperate climate afford us absolute immunity from an importation of this disease. It is of importance, therefore, to consider what method or methods can be relied on to isolate the cholera vibrio from water if present therein. Without prejudice to the value of other methods, the writer has found the old peptone method, with certain modifications, of great value in the case of impure raw river water, purposely inoculated with only a few cholera vibrios, and so, when the method is successful under these conditions, it is quite certain that with less impure waters still better results could be obtained. 175 i 7 6 STUDIES IN WATER SUPPLY CHAP. Thirty-seven samples of raw river water (Thames, Lee, and New River) were purposely inoculated with cholera vibrios, and in 23 cases a positive result was obtained. The number of cholera vibrios artificially added to the infected water (in the 23 successful experiments) was as follows : Per c.c. of infected water. 10-2400 or about 10 5-2000 5 4-6400 5 3-5200 4 3-5200 4 2-5600 3 2-4960 3 2-4800 3 1-5840 2 1-4400 1 1-2800 1 0-5000 or fewer than 1 0-4760 0-3760 0-3540 0-3520 0-2700 0-2380 0-2020 0-1280 0-1060 0-0273 0-0240 The delicacy of the test is illustrated by the accompany- ing diagram (Fig. 37, p. 177). It will be seen that in 12 out of the 23 positive experiments the number of arti- ficially added vibrios was considerably less than 1 per c.c. (about 1 per 4 c.c.) of raw river water. Of course, it might be said by anyone not conversant with the whole facts that positive results were only ob- tained when the numbers of cholera vibrios ranged from 10 '2 4 to 0*024 per c.c. of the water, and that a draught of water (say half-a-pint) is equal to about 284 c.c. But the author was dealing with raw river water samples, and the B. coli results indicate that the water as supplied to London is at least 1,000 times purer than the raw river water, as judged by this test. For example, nearly as many samples of filtered water contain no typical B. coli in 100 c.c. of water, as there are samples of raw Thames water containing typical B. coli in 1 c.c. or less of water. In other words, to prove the absence of the cholera vibrio from 1 c.c. of unpurified river water is, one may venture to say inferentially, equivalent to proving its absence from 1,000 c.c. of the same water after its purification. xi BACTERIOLOGICAL SPECIAL METHODS 177 0, y^g.^H-r , I A G-R AM -ILLUSTRATING THE. -"DELICACY THE: /nEiThcrp-uSEip I.&.THE:- MUAABEIR.- PE.R. C.C. or ARTIFICIALLY \ W- R1VE1R- WATE1R IM TMEI 2>"5 E.X PE1F11 A\ E.HTS-IM- WHICH- A POSITIVE^ KC.SU LTw/\s OBTAIHE1P 1 78 STUDIES IN WATER SUPPLY CHAP. The author has worked successfully with as much as 10,000 c.c. of water, but, of course, 1,000 c.c. is a more practicable and yet satisfactory volume to deal with. 1,000 c.c. of the sample are poured into a flask containing 100 c.c. of concentrated peptone water. The culture is incubated at 37 C. and surface agar plate cultures made after 8, 16, and 24 hours, or more frequently, if this be found to be practicable. The plates are also incubated at 37 C. and examined after 24 hours. The author has also worked with bile-salt-agar and Drigalski and Conradi's medium, but his preference is for ordinary agar. The cholera colonies are rather small, and have a characteristic translucent appearance, which, although not exactly diag- nostic, is of some prima facie value in selecting colonies for sub-culture. It is a good plan to make the sub-culture primarily into cane-sugar peptone water (tinted with litmus). All cultures, which, after 24 hours' incubation at 37 0., show gas formation or absence of acidity are rejected. But those yielding acid, without gas formation, are grown in peptone water. Duplicate cultures should be made, and one set tested with pure nitrogen-free sulphuric acid, for the cholera-red reaction in 24 hours. If the results are negative all the tubes are rejected. But when positive, further peptone cultures are made and the test reapplied after another 24 hours to the original tubes, which by this time have been incubated for 48 hours. If the test is now negative, all the cultures are discarded ; but if positive, further tests must be applied. The reason for this procedure is that some harmless " water microbes " give a slight cholera red reaction in 24 hours, but none in 48 hours ; whereas the cholera vibrio gives a reaction as strong (or stronger) in 48 hours as in 24 hours. 1 It might be supposed that this simple procedure would leave a large proportion out of the total number sub- cultured still requiring to be studied. 1 Tried brands of peptone and sulphuric acid are essential for this test. xi BACTERIOLOGICAL SPECIAL METHODS 179 The contrary, however, is the case, as the following statement clearly shows : 3120 sub-cultures made from the primary plate cultures. r ~T~ 1 1,790 rejected because 549 provisionally accepted 781 were rejected be- no acid change oc- as they produced acid cause they gave rise curred in a cane- (without gas formation) to gas formation in a sugar medium. in a cane-sugar medium. cane-sugar medium. 532 rejected because they failed 17 provisionally accepted, to produce the cholera red but finally rejected by a reaction in peptone water. . combination of tests. It will thus be seen that, by following the above simple method, it is quite easy to reject 3103 out of 3120 colonies subjected to examination. Of course it will be understood that the author is here dealing with what may be called " normal " cholera vibrios. If any worker desires to include aberrant forms, which do not give acid in a cane-sugar medium, and fail to give the cholera red reaction, but which are nevertheless, rightly or wrongly, considered to be of significance, then he must modify the above methods in such a way as to fit in with his particular plan of campaign. The few microbes left over after submitting them to the cane-sugar and cholera-red tests may be readily eliminated by employing a combination of tests. In limiting the number of these extra tests, the author does so, because in his experience he has found them trustworthy in practice. Thus he would reject all microbes, which did not conform to the following type. Acid formation (but no gas) in saccharose, dextrin, and glucose media. Decided cholera red reaction both in 24 and 48 hours at 37 C. in peptone water cultures. No apparent change in salicin media. Eeduction of nitrates to nitrites. N 2 i8o STUDIES IN WATER SUPPLY CHAP. Feeble growth and little or no appreciable liquefaction within 48 hours at 20-22 C. on a gelatine (12 per cent.) slope culture. Minute red colonies on glucose (usually also on saccharose) neutral red bile-salt-agar slope cultures. Of course, if a microbe did fulfil these characters the likelihood of its being the true cholera vibrio would be so great that resort would naturally be had to all other known tests until its acceptance or rejection ceased to be a matter of speculation. Of these, the following may be mentioned merely in passing. " Saturation " and " fixation " tests as well as Pfeiffer's reaction and the hsemolytic and agglutination tests. Morphological appearances, and, in specimens stained for flagella, the presence of only one flagellum. The composition of the chief media used in searching for the presence of the cholera vibrios in water is as follows : Primary peptone cultures. Peptone 11 per cent. Sodium chloride 5 '5 per cent. Water up to 100. 100 c.c. of this medium per 1000 c.c. of the sample of water. Agar plate cultures. Agar 2 per cent. Peptone 1 per cent. Sodium chloride 0'5 per cent. Beef broth up to 100. The beef broth is made in the ordinary way, 1 Ib. of beef being used for every 1,000 c.c. of water (rendered slightly alkaline). Peptone sugar media, tinted with litmus. Peptone 2 per cent. 1 per cent, of whatever sugar or fermentable substance is required, e.g., cane-sugar, glucose, salicin, dextrin, etc. Water up to 100. xi BACTERIOLOGICAL SPECIAL METHODS 181 Peptone ivater, for cholera-red reaction. Peptone 1 per cent. Sodium chloride 0'5 per cent. Water up to 100. Nitrate broth for testing reduction to nitrites. Potassium nitrate 1 per cent. Beef broth 10 per cent. Water up to 100. The beef broth is made in the ordinary way, 1 Ib. of beef being used for every 1,000 c.c. of water (rendered slightly alkaline). JBile-salt-agar media (MacGonkey). Agar 2 per cent. Bile-salt 0'5 per cent. j^o 1 ution red } 0-4 c.c. of a 1 per cent, solution of this substance. Peptone 2 per cent. 0*5 per cent, of whatever sugar or fermentable substance is required, e.gr., cane-sugar, glucose, etc. Water up to 100. Gelatine cultures. Gelatine 12 per cent. Peptone 1 per cent. Sodium chloride 0'5 per cent. Beef broth up to 100. The beef broth is made in the ordinary way, 1 Ib. of beef being used for every 1,000 c.c. of water (rendered slightly alkaline). For additional media and fuller information, the reader is referred to the author's Fourth and Fifth Research Reports to the Metropolitan Water Board. The Typhoid bacillus and Gartner's bacillus. The author has tried many methods, but, in his experience, the method about to be described yields the best results. Any method will yield positive results if a sufficient number of these microbes are added to water, especially when the water is pure ; but there are few methods which will stand the test of successful isolation when only a very few microbes are added to an impure water. Three series of experiments with artificially infected samples of Thames and Lee raw river water yielded results as follows : 1 82 STUDIES IN WATER SUPPLY CHAP. 1st Series, 24 samples. (a) Average number of artificially added typhoid bacilli per c.c. of raw river water =2 '242 (b) Average number of typhoid bacilli recovered = 14'54 l (c) Average number of artificially added Gartner bacilli per c.c. of raw river water = 0*686 (d) Average nmmber of Gartner bacilli recovered =12 '417 1 It should perhaps be explained that the reason why apparently more microbes were isolated than were actually added, is, that for purposes of securing a positive result there is no limit, beyond that of practicability, to the number of cubic centimetres subjected to exhaustive examination. 2nd Series, 35 samples. (e) Average number of artificially added typhoid bacilli per c.c. of raw river water = 0*653 (f) Average number of typhoid bacilli recovered =11 '6 (g) Average number of artificially added Gartner bacilli per c.c. of raw river water = 0*633 (h) Average number of Gartner bacilli recovered = 7 '86 3rd Series, 42 samples. (i) Average number of artificially added typhoid bacilli per c.c. of raw river water = 0*993 (j) Average number of typhoid bacilli recovered =7 '524 (k) Average number of artificially added Gartner bacilli per c.c. of raw river water =0*728 (1) Average number of Gartner bacilli recovered =20'45 If we divide (b) by (a) ; (d) by (c) ; (f ) by (e) ; (h) by (g) ; (j) by (i) and (1) by (k) we get an approximate idea of the number of cubic centimetres of river water, which might be expected to yield a positive result, even if only one typhoid bacillus and one Gartner's bacillus were present. 1st Series. 1 typhoid to 6'485 c.c. of river water. 2nd Series. 1 typhoid to 17*7 c.c. of river water. 3rd Series. 1 typhoid to 7 '58 c.c. of river water. 1st Series. 1 Gartner to 18 '10 c.c. of river water. 2nd Series. 1 Gartner to 12*4 c.c. of river water. 3rd Series. 1 Gartner to 28*10 c.c. of river water. xi BACTERIOLOGICAL SPECIAL METHODS 183 Inferentially, it may be concluded that failure to isolate these microbes under comparable conditions of experiment suggests the absence of the typhoid bacillus from about 7 to 18 c.c. of water, and the absence of Gartner's bacillus from about 12 to 28 c.c. of water. The significance of this will be readily understood, when the reader is reminded that the raw river water before de- livery to the consumers is purified about 1,000 times, as judged by the B. coli test. 500 c.c. of the sample of water are centrifugalised and the deposit spread over 16 malachite green bile-salt agar plates (Medium A). If the water does not contain an appreciable amount of suspended matter, a little alumino-ferric (5 parts per 100,000 parts) may be added just before centrifugalisation. After 24 hours' incubation at 37 C., 250 of the colour- less colonies are picked off, if so many are obtainable, and sub-cultured into Proskauer and Capaldi's medium No. 2 (Medium B). After 48 hours' incubation at 37 C., the tubes are examined, and those showing no change are discarded. Those tubes showing gas formation are followed up on the Gartner side, and those showing acidity, without gas formation, are followed up on the typhoid side. Gartner part of experiment. The cultures showing gas formation (Medium B) are sub-cultured into a dulcite medium (Medium C). After incubation for five days at 37 C., they are dealt with as follows : All those showing no change or no gas formation are discarded. Those showing gas formation are next sub- cultured into a gelatine medium (Medium E). After in- cubation at 37 C. for three hours, and thereafter for about 45 hours at 20-22 C., the tubes are examined. All those showing gas formation are discarded. The remainder are sub-cultered into peptone water, litmus whey, and litmus milk (Media H. F. G.) Those cultures giving rise either to indol formation in i8 4 STUDIES IN WATER SUPPLY CHAP. peptone cultures or to acidity (instead of alkalinity) in litmus whey (or litmus milk) cultures, after five days' incubation at 37C., are discarded. The author's experience, founded on the examination of many thousands of colonies, leads him to assert that these few simple tests suffice for the exclusion of all, or practi- cally all, ordinary "water microbes." Of course, in the event of a microbe passing the barriers set up by these tests, every known test for Gartner's bacillus should be superadded (e.g., the agglutination test, Pfeiffer's reaction, etc. Typhoid part of experiment. The cultures showing acid, but no gas formation (Medium B) are next sub-cul- tured into a gelatine medium (Medium D). After incuba- tion for three hours at 37 C., and thereafter for about 21 hours at 20 22 C., they are dealt with as follows : All those showing gas formation are discarded. Those showing acid formation without gas production are further studied. Not uncommonly cultures are met with which fail to produce gas, but are otherwise not characteristic. These are given the benefit of the doubt, and further sub- cultured. Comparatively so few microbes are left at this stage that usually it is found best to employ the following multiple tests for their final acceptance or exclusion, viz. : (a) Neutral red broth medium (I) purplish tint = satisfactory ; fluorescence or gas production allows of final rejection. (6) Peptone water for indol, medium (H). No indol = satisfactory ; indol formation allows of final rejection. ; (c) L.S.D.S. medium (J). No change = satisf actory ; acid or bleached change or' gas formation allows of final rejection. (d) P and C No. 1 (modified) medium (K). No change = satisfactory ; acid or bleached change or gas formation allows of final rejection. (e) Litmus milk medium (G). Slight acid change, but no clot = satisf ac- tory ; alkaline change or clot allows of final rejection. The foregoing tests (2 days at 37 C.) in nearly all cases suffice, but it is desirable to supplement them with two extra tests. Growth on gelatine slope cultures and on potato cultures [see media (M) and (L)]. Occasionally cultures pass the ordeal of tests (a), (b), (c), (d), (e). In all, or nearly all, cases these will be found xi BACTERIOLOGICAL SPECIAL METHODS 185 either to liquefy gelatine or to produce a non -characteristic bluish opaque (instead of a filmy reddish transparent) growth on gelatine slope cultures [medium (M)] ; usually also, they produce a dirty fawn-coloured growth on potato cultures [medium (L)] instead of a transparent invisible growth. Speaking of thousands, and even tens of thousands, of cultures of water microbes, the foregoing procedure is, for all practical purposes, absolutely satisfactory, for the purpose of exclusion, but as a counsel of perfection, tests for agglutination, etc., should also be employed. The chief reason why this method is so effective as a means of excluding " water microbes " is very simple. Microbes which yield acid in liquid sugar media, but no apparent gas, are constantly to be found in water, and they bear a superficial resemblance to the typhoid bacillus, morphologically and culturally. In nearly all cases, they can be shown to be feeble gas producers, by using a gelatine sugar medium (Medium D). In the past these microbes were undoubtedly at times confused with typhoid, and not improbably elaborate experiments were sometimes under- taken to prove their identity ; and it is even to be feared that in some cases the final diagnosis was erroneous. Assuming, as the author believes, it to be true that these tests and plans of procedure do not exclude any microbes which can strictly be classed under the terms typhoid and Gartner, the methods recommended have the merit of simplicity and they have stood the test of practical experience. The composition of the various media is as follows : (A) Bile-salt agar medium for .plate cultures (both for typhoid and Gartner). Agar, 2 per cent. Peptone, 2 per cent. Bile-salt, 0*5 per cent. Lactose Saccharose Adonite \ 0*2 per cent, of each. Raffinose Salicin 1 86 STUDIES IN WATER SUPPLY CHAP. Water tinted with neutral red (4 c.c. of a 1 per cent, solution of neutral red per litre of medium) up to 100. Just before pouring the plates, malachite green is added in the proportion of 1 per 10,000 of medium. Remarks : Both typhoid and Gartner form colourless colonies. (B) Proskauer ani CapaldVs No. 2 medium (both for typhoid and Gartner). Peptone, 2 per cent. Mannite, 0*1 per cent. Distilled water tinted with litmus up to 100. Remarks : Typhoid, acid, no gas ; Gartner gas formation. (C) Dulcite medium (for Gartner only). Peptone, 2 per cent. Dulcite, 0'2 per cent. Water tinted with litmus up to 100. Remarks : Gartner produces gas. (D) Glucose gelatine medium (for typhoid only). Peptone, 2 per cent. Lemco (Liebig's Extract of Meat), 1 per cent. Gelatine, 7 '5 per cent. Glucose, 1 per cent. (Potassium hydrate, 10 c.c. of a 5 per cent, solution per litre of medium). Water tinted with litmus, up to 100. Remarks : Typhoid produces acid but no gas. (E) Mixed sugar gelatine medium (for Gartner only). Peptone 2 per cent. Gelatine 7 '5 per cent. Glycerine Raffinose Saccharose Adonite 0'2 per cent, of each. Salicin Lactose (Potassium hydrate, 10 c.c. of a 5 per cent, solution per litre of medium). Water tinted with litmus, up to 100. Remarks : Gartner bleaches but produces no gas. (F) Litmus whey medium (for Gartner only). Milk clotted with Hydrochloric acid, filtered, heated and neutralised with caustic soda tinted with litmus and filtered. Or the milk may be clotted with rennet, strained, heated, filtered, and tinted with litmus. Remarks : Gartner produces faint acid, and then an alkaline change. (G) Litmus milk medium (both for typhoid and Gartner). Milk steamed allowed to cool and siphoned off to avoid the coagulum. Tinted with litmus. Remarks : Typhoid, faint acid no clot ; Gartner, faint acid, later alkaline change, no clot. xi BACTERIOLOGICAL SPECIAL METHODS 187 (H) Peptone water medium for indol test (both for typhoid and Gartner). Peptone 1 per cent. Sodium Chloride 0'5 per cent. Water up to 100. Remarks : Typhoid, no indol ; Gartner, no indol. Composition of solutions used for testing for indol (Bohme, Centr. /. Bald. Bd. XL. 1905) : (1) 8 grammes paradimethylamidobenzaldehyde, 160 c.c. hydrochloric acid, 760 c.c. absolute alcohol. (2) Saturated cold water solution of potassium persulphate. (I) Glucose neutral red broth medium (for typhoid only). Peptone, 1 per cent. Sodium Chloride, 0'5 per cent. Lemco, 0'5 per cent. Glucose, 0*5 per cent. Water tinted with neutral red (2 c.c. of a 1 per cent, solution per litre of medium) up to 100. Remarks : Typhoid, a purplish tint, no gas, and no fluorescence. (J) L.S.D.S. medium (for typhoid only). Peptone, 2 per cent. Lactose Saccharose Dulcit 0'25 percent, of each. Salicin Distilled water tinted with litmus up to 100. Remarks : Typhoid, no change. (K) Proskauer and Capaldi's No. 1 (modified) medium (for typhoid only). Asparagin 0'2 per cent. Calcium chloride 0-02 per cent. Galactose 0'2 per cent. Glucose 0'2 per cent. Laevulose 0'2 per cent. Magnesium sulphate O'Ol per cent. Maltose 0'2 per cent. Mannite 0'2 per cent. Potassium monophosphate 0'2 per cent. Sodium chloride 0'02 per cent. Distilled water tinted with litmus, to 100. Remarks : Typhoid, no change. (L) Potato medium (for typhoid only). Half cylinder of potato cut obliquely, placed in tubes and sterilised. Remarks : Typhoid, colourless growth. i88 STUDIES IN WATER SUPPLY CH. xi (M) Gelatine medium, slope cultures (for typhoid only). Peptone, 10 grammes. Gelatine, 100-120 grammes. Sodium Chloride, 5 grammes. Sorbite, 2 grammes. Beef broth up to 1,000 c.c. (1 Ib. beef per litre). Medium rendered faintly alkaline and tinted with litmus solution. Remarks: Typhoid, filmy transparent growth, no liquefaction ; the medium acquires a faint pinkish tinge. MISCELLANEOUS INFORMATION. Weather of London (Camden Hill). Average temperature and rainfall for 50 years equals 50 F. and 25 inches respectively. "Absolute drought" = more than 14 consecutive days, no one of which is a rain day. (A rain day = 0*01 inch or over ; anything over 0*005 being calculated as 0*01, and 0'005 and under being ignored.) " Partial drought " = more than 28 consecutive days, the mean rainfall of which does not exceed O'Ol inch per day. One inch of rain per acre = 22,622 gallons (about 101 tons). Some extremes of rainfall attained in the British Isles Symons's "British Rainfall," 1884. 0*55 inch in 5 minutes. 1*10 inches in 15 ,, 1-25 30 1*50 inches in 45 minutes. 1-80 60 2-20 120 HYDRAULIC EQUIVALENTS. One. Gallons. Litres. Cubic Centi- metres. Lbs. Fluid Ounces. Cubic Inches. Cubic Feet. Gallon ... Litre 1 0-22 4-5459 1 4545-9 1000 10 2-2 160 35-196 277-463 61-024 0-16 0-0353 Cubic cent. Pound Ounce Cubic inch Cubic foot 0-00022 o-i 0-00625 0-0036 6-228 o-ooi 0-454 0-0284 0-0164 28-317 1 454-6 28-412 16-387 28317 0-0022 1 0-0625 0-036 62-3 0-0352 16 1 0-5766 1000 0-0610 27-746 1-728 1 1728 0-000035 0-016 0-001 0-00057 1 United States gallon = 231 cubic in. = 37854 litres. 1 pint = 20 fluid oz. = 0'568 litre. 1 centimetre = 0*3937 inch. 1 litre = 1*759 pints = 35 fluid oz. 1 inch = 2*54 centimetres. 1 Ib. = 7,000 grains=453*59 grammes. 1 cwt. = 112 Ibs. 1 ton = 2,240 Ibs. 1 micron = 10 ~ 6 metre = 0*001 mm. 189 190 STUDIES IN WATER SUPPLY CONVERSIONS : Grammes (or c.c.) into grains, ounces, or pounds multiply by 15*432, 0-03528, and 0*0022046 respectively. Grains, ounces, or pounds into grammes (or c.c.) multiply by 0*0648, 28*35, and 453*6 respectively. Metres into inches, feet, or yards, multiply by 39*3701, 3*2808, and 1*0936 respectively. Inches, feet, or yards into metres, multiply by 0*0254, 0*3048, and 0*9144 respectively. Degrees Centigrade into degrees Fahrenheit, multiply by 9, divide by 5, and add 32. Degrees Fahrenheit into degrees Centigrade, subtract 32, multiply by 5, and divide by 9. Parts per 100,000 into grains per gallon, multiply by 7, and divide by 10. Grains per gallon into parts per 100,000, multiply by 10, and divide by 7. "Hardness," parts per 100,000 (CaC0 3 ) into degrees of hardness = Clark's scale = grains per gallon, multiply by 7, and divide by 10. Cubic capacity of cylinder = square of diameter x 0*7854 x length. Circumference of circle = diameter x 3*1416. Area of circle = square of diameter x 0*7854. Capacity in gallons of rectangular vessels : Multiply length by breadth, by depth in inches, and divide by 277'463 (say 277*5). Capacity in gallons of cylindrical vessels : Square of half the diameter in inches, multiplied by 3*1416, multiply product by depth in inches, divide answer by 277*463 (say 277*5). Rate of nitration : 1 inch per hour = 2 feet per day = 2 cubic feet per square foot per day = 12'5 gallons per square foot per day = 544,500 gallons per acre per day. Number of gallons per square foot per hour x 1,045,440 = number of gallons per acre per 24 hours. The following values are given by Baldwin-Wiseman (Proc. Inst. C.E. 1910 for 86 installations in Great Britain : Maximum rate of discharge, 1*444 cubic ft. or 9*028 gallons per sq. ft. per hour Minimum ,, 0*083 ,, ,, 0*521 ,, ,, Mean 0*372 2*322 METROPOLITAN WATER BOARD. STATISTICAL INFORMATION (1913). Population supplied, about 6,630,000 (census, 1911). Daily supply, 244 million gallons. Gallons per head per day, 36*5. Total storage capacity, 13,000 million gallons. About 60 per cent, of London water is derived from the River Thames, about 20 per cent, from the River Lee, and the remainder from wells, etc. MISCELLANEOUS INFORMATION 191 CLASSIFICATION OF ODOURS CAUSED BY ORGANISMS IN POTABLE WATERS (after WHIPPLE). Group. Organism. Natural Odour. AROMATIC ODOUR. Diatomacew : Asterionella Cyclotella Diatoma Meridion Tabellaria Aromatic geranium fishy Faintly aromatic )> 5J Aromatic 9 > Protozoa : Cryptomonas Mallomonas Candied violets Aromati c violets fishy GRASSY ODOUR. Cyanophycece : Anabsena Rivularia Clathrocystis Ccelosphserium Aphanizomenon Grassy and mouldy green cornnastur- tiums, etc. Grassy and mouldy Sweet, grassy Grassy FISHY ODOUR. ChlorophycecK : Volvox Eudorina Pandorina Dictyosphserium Fishy Faintly fishy ?) >5 J J Protozoa : Uroglena Synura Dinobryon Bursaria Peridinium Glenodinium Fishy and oily Ripe cucumbers bitter and spicy taste Fishy, like rock weed Irish moss salt marsh fishy Fishy, like clam-shells Fishy Figs. 38-43 show the microscopic appearances of certain algal growths (Asterionella, Oscillaria, Tabellaria, and Dinobryon). Copper sulphate treatment for Algal growths. Usual dose 0'2 to 1 part per million parts (Ibs. per 100,000 gallons). BROAD "STANDARDS" for guidance and working purposes only. Bacteriological. River-derived samples (London). It is assumed that conclusions are based on the examination of a sufficient number of representative samples. Less than 20 microbes per c.c. (Gelatine at 20-22 C., counted on 3rd day). ,, 5 (Agar at 37 3 C., counted after 24 hours). ,, 0-5 ,, (Bile-salt Agar at 37 C., counted after 24 hours). 192 STUDIES IN WATER SUPPLY Speaking generally, these standards may be considerably relaxed if the B. coli results are favourable. In striking the gelatine and agar averages it is customary in the case of the London waters to exclude samples containing 100 or more microbes per c.c. Less than one-half the samples should contain typical B. coli (lactose +, indol +) in 100 c.c. Chemical (parts per 100,000). River-derived samples (London). Oxygen absorbed from per- 1 , ,-, manganate at 80 F. J J [ /O'l in 4 hours. \ 0'038 in 5 minutes. Albuminoid nitrogen, less than O'Ol. Ammoniacal nitrogen, less than O'OOl. Colour, less than 20 mm. brown in a 2-foot tube. WATER-BORNE EPIDEMIC DISEASES. TYPHOID FEVER, CHOLERA, POSSIBLY DlARRHGEAL DISEASE. AVERAGE DEATH-RATE FROM ENTERIC FEVER (TYPHOID FEVER) PER 100,000 OF POPULATION. 1871-80. 1881-90. 1891-00. 1901-10. 1911. England and Wales... London 33 24 20 19 17 14 9 6 7 3 Seasonal incidence of typhoid fever. Minimum about May, June, and July. Maximum, about October and November. Incubation period of typhoid fever, about 10-14 days. For information as to suspected connection between water supplies and outbreaks of enteric fever (1867-1892), extracted from reports of Inspectors of the Local Government Board and others, see Appendices to " Evidence of Royal Commission on Metropolitan Water Supply," 1893, page 532. Chief English typhoid epidemics of recent date attributed to water supply. Worthing, 1893 (1,315 cases); Maidstone, 1897 (1,847 cases); Lincoln, 1905 (more than 1,000 cases). Hypochlorite sterilisation. Usual dose in terms of available chlorine, from 0'2 to 1 part per million parts (Ibs. per 100,000 gallons). METROPOLITAN WATER BOARD. Notable Events. Appointed Day (Amalgamation of old Companies). June 24th, 1904. Staines Reservoirs. Water first used, Dec. 28th, 1904. Combined capacity, 3,338 m. g. Walton Reservoirs. Opened June 10th, 1911. Combined capacity, 1,198 m.g MISCELLANEOUS INFORMATION 193 Island Barn Reservoir. Opened Nov. 4th, 1911. Capacity, 922 m. g. Beachcroft Service Reservoir at Honor Oak. Opened May 5th, 1909. Capacity, 56'3 m. g. First used, March 14th, 1910. Metropolitan Water Board (New Works) arid Thames Conservancy Acts, 1911. Humphrey pumps installed, and Chingford Reservoir opened by His Majesty King George the Fifth, March 15th, 1913. Capacity, 3,000 m. g. Now known as King George's Reservoir. LIST OF PUBLISHED WATER BOARD REPORTS. The following is a list of the published reports relating to the quality of the Metropolitan Water Supply, since the work commenced on November 1st, 1905 : (1) A report each month from November, 1905, onwards. [An unabridged version of the November, 1905, report, giving all the results hi great detail was published by the Water Board.] These monthly reports form part of the Government Water Examiner's Monthly Report on the Metropolitan Water Supply, and are published by the Government Printers. All other reports are published by the Water Board. (2) First Annual Report, for the year ended March 31st, 1907. (A report dealing with the results for the 12 months ended October 31st, 1906, had previously been issued.) (3) Second Annual Report, for the year ended March 31st, 1908. (4) Third Annual Report, for the year ended March 31st, 1909. (5) Fourth Annual Report, for the year ended March 31st, 1910. (6) Fifth Annual Report for the year ended March 31st, 1911. (7) Sixth Annual Report for the year ended March 31st, 1912. (8) Seventh Annual Report for the year ended March 31st, 1913. (9) First Report on Research Work. The Vitality of the Typhoid Bacillus in artificially infected samples of raw Thames, Lee, and New River Water with special reference to the question of storage. (10) Second Report on Research Work. The Negative Results of the examination of samples of raw Thames, Lee, and New River Water for the presence of the Typhoid Bacillus. (11) Third Report on Research Work. The storage of raw River water antecedent to filtration. (12) Fourth Report on Research Work. The vitality of the Cholera Vibrio in artificially infected samples of raw 'Thames, Lee, and New River Water, with special reference to the question of storage. (13) Fifth Report on Research Work. I. The Results of the examination of samples of raiv Thames and Lee Water for the presence of the Typhoid Bacillus and Gartner's Bacillus. II. The Results of the examination of the raw River Waters (Thames, Lee, and New River) for Faecal Streptococci. III. The Results of the Examination of the Pro-filtration Water, i.e. repre- sentative samples of practically all the London Water (raw, stored, gravel, well, and mixed water) antecedent to filtration. IV. The Negative Results of the Examination of Samples of raw Thames, Lee, and New River Water for the presence of Morgan's (No. 1) Bacillus. V. The Isolation of Cholera Vibrios from samples of raw River Water artificially infected with only a few Vibrios. 194 STUDIES IN WATER SUPPLY (14) Sixth Report on Research Work. The Comparative Vitality of "Uncultivated" and "Cultivated" Typhoid Bacilli in artificially infected, samples of raw River water with special reference to the question of storage. (15) Seventh Report on Research Work. I. Search for Pathogenic Microbes in raw River water. II. The comparative vitality of " Cultivated" and "Uncultivated" Typhoid Bacilli in artificially infected samples of raw River water. III. The Comparative Vitality of the Typhoid Bacillus in raw Thames water at different temperatures. IV. On the hiological characters of B. coli in raw, stored and filtered water. V. On the advantages of passing raw River water through small reservoirs antecedent to storage in large reservoirs. VI. On the advantages of occasionally using precipitation methods, antecedent to the storage of raw River water in large reservoirs. (16) Eighth Report on Research Work. The softening, purification, and sterilisation of water supplies. (17) Ninth .Report on Research Work. Search for certain pathogenic microbes in Raw River Water and in Crude Sewage. SOME ROYAL COMMISSIONS, BOOKS, AND REPORTS ON THE METROPOLITAN WATER SUPPLY Scratchley, P. A. " Bolton's " London Water Supply, 1888. Sisley, R. The London Water Supply, 1899. Shad well, A-- The London Water Supply, 1899. Richards and Payne. London Water Supply, 1899. For a complete bibliography of 629 principal reports and papers (up to 1903) relating to the Water Supply of London, see "London Water Supply" issued by the London County Council, 1905. The following are the titles of a few reports containing information specially concerned with the quality of the Water Supply : 1850. Report by the General Board of Health on the Supply of Water to the Metropolis. 1851. Report by the Government Commission on the Chemical quality of the Supply of Water to the Metropolis. 1868. Rivers Pollution, Royal Commission of 1868 (Sixth Report, 1874). The Domestic Water Supply of Great Britain. Royal Commissions Reports, Minutes of Evidence, and Appendices : 1867. Shortly cited as Duke of Richmond's Commission, reported 1869. 1892. Shortly cited as Lord Balfour of Burleigh's Commission, reported 1893. 1897. Shortly cited as Lord Llandaff s Commission, reported 1899. ALGAL GROWTHS 95 \ \ FIG. 38. Asterionella x 320. ll>. FIG. 39. Oscillaria x 250. 196 STUDIES IN WATER SUPPLY FIG. 40. Tabellaria (front view) x 280. -v, FIG. 41. Tabellaria (front and side view) x 280. ALGAL GROWTHS 197 I FIG. 42. Dinobryon x 180. FIG. 43. Dinobryon x 450. INDEX ABSTRACTION, different methods of, 40 knowledge of current quality as guide to, 44 relation of colour results to, 40-47 Accidental pollution, importance of safeguards against, 104 Agar, medium, composition of, 170 plates, method for drying, 154 tubes, constant temperature water bath for, 150-151 Albuminoid nitrogen, in Chelsea stored water, 94-95 in River Lee water, 5-7 in River Thames water, 5-6, 93 Algse, development of, in storage reservoirs, 100 microphotographs of, 195-197 remedy for destroying growth of,. 100, 191 remedy for taste due to, 100 Alumino-ferric, composition of, 56 cost of, 58 dose required as coagulant, 59-60 results of experiments with, 54-58 American Cities, typhoid death rate in, 110-112, 117-121, 133 Asterionella, a cause of taste and smell in water, 100, 191 microphotograph of, 195 Automatic inoculation, method of, 160 BACILLI, "uncultivated" and "culti- vated," definition of, 103 Bacteria enumeration of, average standards for, 191 in River Lee, 16, 22 in River Thames, 15-16, 18, 89-90 in sewage, 116 in stored water, 91-92, 95 media for, 170-171 Bacteria, excremental, classification of, 161, 165 significance of, 166, 168 Bacterial efficiency of mechanicalfilters, 51 of sand filtration, 103 Bacteriological examination, bottles and sample cases for, 133 media rack used in, 152, 155 quantity of water required for, 138 routine methods for, 151 et seq. time reqiiired for, 168, 170 Bacteriological results, methods of registration of, 142 of examination of River Lee, 16, 22 of examination of River Thames, 15-16, 18, 89-90 of examination of stored water, 91-92, 95 Bacteriological standards, 168, 191 Bacillus coli, isolation of by "flaginac" method, 161 isolation of by " glagins " method, 161-164 media used in isolation of, 170-4 significance of in water, 166, 168 types of, 165 Bacillus Coli test, classification of results, 165 description of presumptive, confirma- tory and typical, 143 results of examination bv, of filtered water, 1, 49, 169 of Kent wells, 2 of River Lee, 16, 22 of River Thames r 15-16, 18, 89-90 of stored water, 91-92, 95 Bacillus Enteritidis Spororjenes test, description of, 154, 158 media for, 174 Bacillus Typhosus, See Typhoid. Bacteria pathogenic, delicacy of method used in search for, 35 methods for isolation of, 33 results of experiments, 34-35 search for, in raw water, 33 in sewage, 36 Bile-salt peptone water, composition of medium, 171 Bleaching powder, composition of, 62 Boxes, for sample collecting, 138, 141 199 200 INDEX CHELSEA reservoir, results of storage in, 91-92, 94-95 Chemical examination, of Chelsea stored water, 94 of River Lee, 5, 7 of River Thames, 5, 6, 93 Chemical standards, 15, 192 Cherwell, temperature of, at Oxford, 87 Chicago water supply, sterilisation of, 64 Chloros, composition of, 62 dose for sterilisation, 62 results obtained at Lincoln, 62 with wells, 63 sterilisation by, 62-63 Cholera Vibrio, characters of, 178-188 delicacy of method employed in isolation of, 176 media for isolation of, 180-181 method for isolation of, 178 vitality of in river water, 80-81 Classification of odours caused by organ- isms in potable water, 191 types of B. Coli, 165 Coagulants, advantages of use antece- dent to storage in large reser- voirs, 54 dose required, 59-60 results of experiments with, 54-59 Coli. B, see B. Coli Collembola, in hvdrants at Edinburgh, 105 Colour results : comparison of river and stored water, 46 Chelsea stored water, 94-95 method used for determining, 40 relation to abstraction, 40-47 River Lee, 5, 7 River Thames, 5, 6 ; diagram of, 41 tables showing number of days results were over 200 and over 100, 46 Confirmatory test for B. Coli, definition of, 143 Conradi and Drigalski's medium, com- position of, 28 Copper sulphate treatment for Algal growths, 100, 191 Cultivated bacilli, definition of, 103 DEATH rate, definition of normal, 124 average from enteric fever, England and Wales, 192 Decimal mode of dilution, description of, 147-148 Definition of drought, 189 normal death-rate, 124 "uncultivated" and "cultivated " bacilli, 103 Devitilisation as a factor in purification by storage, 80 Diarrhoeal diseases, annual death rate from, 129 Dilution, description of decimal mode of, 147-148 Dilutions, coloured labels for, 149 Dinobryon, a cause of taste and smell in water, 191 microphotograph of, 197 Drigalski and Conradi's medium, com- position of, 28 Drinking experiment as proof of devi- talisation, 83-84 Drought, definition of, 189 ENDEMIC typhoid fever, current views on, 109-114, 116-117 English typhoid epidemics of recent date, 192 Enteric fever, average death rate, Eng- land and Wales and London, 192 Enteritidis tiporogeiies B test, descrip- tion of, 154, 158 media for, 174 Enumeration of microbes, media for, 170-171 Epidemic, typhoid, description of Rock- ford (Illinois), 107-109 water borne diseases, 192 Epidemics, caused by impure water, 103 financial loss from water, 134-135 Equalisation, nature of, as a factor in purification of water, 78-80 Equivalents, table of hydraulic, 189 "Excess lime," advantages of process, 72 cost of process, 70-71 dose for hard and soft waters, 67-68 method for determining amount of lime required, 74-75 method of sterilisation, 65-77 FAECES, results of examination for streptococci in, 29-30 Filtered water, possible causes of pollu- tion of ; 104 Filters, mechanical, see Mechanical filters Filtration of water, cost of sand, 59, 136-137 efficiency of, 103 rate of, 190 trustworthiness of, 104 Financial advantages of pure water supply, 126 et seq. Fishing, prohibition of, 105 Fishy odour, list of organisms causing, 191 "Flaginac" method for isolation of B. Coli, description of, 161 media for, 172-173 Flood water, quality of, 39 reason for excluding, 47-48 INDEX 2OI Flow of river, in relation to quality, 39 Fluorescent water, composition of, 172 Formula, for monetary loss due to im- pure water, 125 GARTNER'S bacillus, characters of, 183- 184 delicacy of method used, reasons for, 185 media for isolation of, 185-187 method of isolation, 183 results obtained, 182 Gelatin shake test, composition of medium for, 173 "Glagins" method for isolation of B. Coli, description of, 161-164 media for, 173-174 Glucose, drop method for isolation of B. Coli, 157 HARDNESS, financial loss from, 134 method for reduction of permanent, 76 permanent, removal of, 76-77 of River Lee, 5, 7 ; of River Thames, 5-6, 93 ; of stored water, 94-95 table of, in American Cities and London, 135 Heat, as a method of sterilisation, 65 Hektograph sheets, description of symbols on, 144-147 examples of, 148 et *<