SA^.3 Reprinted from The Journal of Infectious Diseases, 1907, Supplement No, 3, pp. 41-49 SANITARY CHEMICAL EXAMINATION OF WATER BACTERIA/ Andrew Watson Sellards. (State Water Survey, University of Illinois, Urbana, Illinois.) In view of the modern theory of disease the problem of sanitary water analysis concentrates itself on the bacteriological rather than on the chemical condition of the water in question, whatever the method may be of arriving at this knowledge. All waters would be perfectly safe for drinking purposes, with certain well-characterized exceptions in the case of mineral constituents, if only the pathogenic bacteria were removed. Our investigations are, therefore, directed toward those bacteria that are pathogenic to man, are capable of being borne by water, and have the possibility of infecting the human subject through the alimentary tract. The routine sanitary chemical tests, all of which are for substances perfectly harmless in themselves, are merely an indirect method of determining the bacteriological condition of the water. Naturally with such an indirect method there is a very wide range of possibilities in the interpretation of the results. The analytical data of a sanitary chemical analysis indicate merely the presence, recent presence, or possible future presence of pathogenic bacteria. It follows accord- ingly that by current methods the sanitary chemical analysis of pure distilled water to which a little sterile bouillon had been added would indicate a dangerously polluted water. Naturally, with the advancement of bacteriology, more direct methods of analysis have been undertaken. These are of course limited to quantitative and qualitative investigations of the bacteria. One does not need to emphasize the difficulty of establishing a stand- ard for the maximum number of bacteria permissible in a potable water, or of the uselessness of searching the Mississippi River or the Great Lakes for a typhoid bacillus. > This work was carried on under co-operative agreement between the Illinois State Water Survey, State Geological Survey, Engineering Experiment Station of the University of Illinois, and the Division of Hydro-Economics of the United States Geological Survey. 41 19o!0 42 Andrew Watson Sellards The only qualitative work attempted at present is the identification of intestinal forms. The presumptive coli tests are in a very uncertain and unsatisfactory condition. The quantitative estimation of coli is well-nigh impossible in a laboratory where a large number of sam- ples are received for daily routine examination. Though possible presence or absence can be quite satisfactorily made out, even when many samples require attention at the same time, an opinion cannot be based on a qualitative test alone. It is essential to know the quantity of coli bacteria present. Of the various routine methods of procedure, the direct chemical analysis, being the broadest of all, has one advantage over all, or one disadvantage as the case may be. Such an analysis does not depend on the presence of living active organisms as do all bacterial methods ; but the pollution may still be detected where bacteria have died in large quantities after exhausting their food material. Also the chemical analysis will detect the pollution where soil-filtered sewage reaches a water-supply. Probably the most necessary factor in inter- preting the sanitary chemical analysis is a thorough knowledge of the source of the water under examination. In this section of Illinois one is utterly helpless in the interpretation of the analysis of a well water, unless he knows whether the well is deep or shallow and in drift or in rock. The results obtained by analyzing the water from two wells in Urbana will illus- trate the difficulty. Amounts are stated in parts per million. No. I No. 2 Total residue on evaporation . • 415- 365- Chlorine in chlorides .... 3-5 7-5 Oxygen consumed .... 5.2 2 .0 Nitrogen as free ammonia 3.6 0.054 Nitrogen as albuminoid ammonia . 0.136 0.056 Nitrogen as nitrites .... 0 . 040 0.030 Alkalinity • 353 - 245 - I c.c. 1 c.c. Bacteria at 20° C 200 . 1000 . B. coli communis .... Absent Present No. I is from a well i8o feet deep and has high oxygen consumed and high free and albuminoid ammonia, which are characteristic of deep wells in the drift in central Illinois. Bacterial tests show that the water is in good condition. No. 2 is a water from a 30-foot dug well, and chemically, according to usual methods of interpretation, it is a better water than No. i. Bacterially it contains 1,000 bacteria per C.C., and the reaction is positive to the presumptive test for coli. Generally speaking, the chemist must know the normal constituents of a given locality before he can determine the amount due to pollution. Chemical Examination of Water Bacteria 43 In addition to the sanitary chemical analysis and the quantita- tive and qualitative bacteriological examination, a further hne of procedure is suggested to the sci^tist, that is, the chemical analysis of bacterial cultures obtained by the inoculation of water samples into artificial media. Our experiments upon this principle are based on the supposition that, by the inoculation of water into artificial sterile media, we could, in a way, imitate the changes that would ordinarily be brought about by bacteria in a water containing natural media. One great advan- tage would lie in the ease with which the artificial media could be analyzed, and the accuracy with which the changes due to the water could be determined. By this means it is possible that local factors which affect the interpretation of the results of analysis may be removed. There arises the possibility of establishing an absolute standard for the maximum limits of impurities. To establish the value from the sanitary standpoint of the analysis of the cultures of water, our experiments are carried out with a view to securing constancy of results. After a few preliminary tests (see Table i) we chose for a medium an ordinary meat-extract broth of double concentration, to which 2 per cent of gelatin was added. The presence of sugars prevented decomposition of the nitrogenous constituents. When it was desired to study the nitrogenous decom- position products, sugars were not added to the medium, although the extremely small amount of sugar present in ordinary acid meat extract was not removed. For our first experiments we chose two types of water, one from a small creek, known to be seriously polluted, the other from deep, driven wells of unquestionable purity. We measured accurately 5 c.c. of each medium into test-tubes, inoculated with i c.c. of each sample of water, and incubated at 37° C. In these preliminary tests, the cultures, when they showed appreciable differences, were steamed in an Arnold sterilizer 20 minutes, diluted to 1,000 c.c., and subjected to sanitary analysis. A priori it was expected that it would be neces- sary to work with very young cultures, presuming that for old cul- tures the ultimate analysis of a pure and of a polluted water would be very similar. The waters tested thus far have given decidedly the contrary result. Table i is characteristic of the effect of 44 Andrew Watson Sellards variations in media and age of culture. The results are expressed in parts per million of the diluted culture, 5 c.c. of the inoculated medium diluted to 1,000 c.c. with pure distilled water. TABLE I. Comparisons of the Action of Pure and Polluted Water on Media of Various Com- positions FOR Different Periods of Time. Sample Oxygen Consumed Acidity Ammonia Total Solids Nitrates Chlo- rides Media Free Albuminoid Hours. . . 24 89 24 89 24 89 24 89 24 89 24 89 89 hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. hrs. Tap .... 102 75 12.13 12.13 0.64 1. 12 8.00 13-6 212 0.48 0.40 37-0 Creek... . 90 55 12.13 21.82 4.60 13-20 8.00 4.8 124 0.40 0.44 35-5 Broth Control . . 75 7.27 0.64 6.80 254 ■ 0.48 37-5 Tap 165 170 16.97 15-52 0. 72 1 . 00 8.80 14. 386 0.40 36-0 Creek — 145 145 36.37 54-80 0. 76 1 .00 7.60 12.0 309 0. 72 36-0 Glucose Control. . 165 14-55 0.64 6.4 433 0.52 35-0 Broth Tap 94 92 14-55 19.40 0.52 1.04 18.80 23-2 428 308 ■ 56 34-5 Creek... . 104 60 24-25 67.90 3 - 6 o 23.20 18.80 10.4 397 168 0.64 32.5 Gelatin Control. . 85 9.70 0.64 13-2 359 0.52 32-5 Broth Tap i8s 183 21.82 26.67 0. 76 1 . 12 18.00 23.20 612 489 0.72 0.48 36.5 Gelatin Creek .... 195 163 33-95 60.62 0.84 1.36 19. 20 22.4 574 400 0.80 0.60 36.0 Glucose Control. . 180 16.97 0.64 13-60 557 0.44 37-0 Broth As the most striking changes took place in the nitrogenous constituents, we decided to secure a series of preliminary tests deter- mining only free ammonia. We were sometimes able to nesslerize directly, though more frequently, on direct nesslerization, we secured a greenish color, not comparable with a true nessler color. In an attempt to establish a maximum limit of free ammonia, allowable, we obtained samples from deep wells of unquestionable purity, from the best available shallow wells, and from chemically pure water artificially polluted. In the following table the cultures were analyzed at different ages as a sufficient variety of waters had not yet been analyzed to determine the most favorable age of culture. In Table 2 the following facts are of special interest. The previous statement that the chemical analysis of pure, deep, and shallow ground waters is entirely different is verified. Noticeable examples are Nos. 18 and 19. In No. 18 and No. 19 the results of the bacteriological analysis are practically the same and also the chemical analysis of the cultures of the two agree very well. This strengthens our hope that a universal standard of purity may be established. Comparison of Sanitary Analyses of Waters, from Various Sources, with the Ammonia that is Obtained from Media by Inoculation Chemical Examination of Water Bacteria 45 Analysis of Culture, P. M. 1-200 Dil. Age Cult. in hrs. Tj- Tf . vo 0"0 O' - 0000 - irifo -ooooococoooooooooooo -oooo OIO0 loo •• •• • ooo fOVO fO'O • • Alka- linity r'CO'O itt^MINOO rj-... co^ (Ti TtvO M H O M tT ... lor^ COM o ... POPO WW fO N(SCOWPOWWP<)POPO ... Oxygen Con- sumed m* lolo • loo^ooO'oioio Mvo • wot^Oiot^r^'^sOt^ O'Cco lOO • MNN mmmvomO 8J?8 8 8 8 8;? S88 OO OO O OOOHOOOOOO ooo d d d d d ’ ' d ’ d Albumi- noid 0.136 0.64 0.336 0. 144 O.OQ 4 0.048 .056 .016 0. 128 0.64 .056 .082 0.184 0.094 0.64 0.05 0.016 0.010 Free 0 "T 0 't ’too 0 N N 00 w 00 W M PI M lO'C '^00 <000 0 <0 PI CO W lOOO MO 0 OOt^PlOOO'OOOO ooo fo PI dd 0/-^ 0 MO pooci ooo CJ 0 0 Bacterial Analysis Coli in one c.c. -f (abov (abov (abov (abov -f ?-f + (abov + + + + No, per c.c. 00 . . , .OOOMO . OOOOOOVPOIOPSO ooo 00 oOOOChO