SCIENCE. 



[Vol. XXIL No. 564 



invertebrates is constantly and easily demonstrable. 

 Chatin, in the previous paragraph, referred to osmic acid; 

 it is natural to suppose that the organisms and globules 

 were submitted to that treatment, a method which, at 

 least in my hands, has i^roved extremely uncertain in its 

 results. 



In conclusion, allow me to request some of your very 

 numerous correspondents to inform us if the spectroscope 

 would give any material assistance in the solution of the 

 true nature of these markings. (I, of course, mean the 

 diffraction spectrum), my acquaintance with the instru- 

 ment being limited to test fluids. 



Since writing the above my colleague, W. F. Pentland, 

 has persuaded me not to be too dogmatic with regard to 

 the reticulation of the invertebrate corpuscles and indi- 

 vidual (esjiecially conjunctival) cells of invertebrates till 

 after next spring, as in the meantime he intends working 

 up the subject. 



THE BACTEEIOLOGICAL ANALYSIS OF WATEE. 



BY J. H. STOl.LEK, DHION COLLEGK, SCHENECTADY, N. Y. 



When, in 1881, Koch announced the gelatine culture 

 method for bacteria devised by him, it was believed that 

 one of its most important applications would be in the 

 examination of waters with reference to their i^otable use. 

 This method, as is now well known, renders possible an 

 exact determination of the abundance of bacteria in water. 

 But it w as soon discovered that the mere demonstration 

 of the presence of bacteria was of little value in estimat- 

 ing the qualities of waters, inasmuch as waters of unques- 

 tionable suitability for potable use often contained bac- 

 teria in considerable abundance. However, the general 

 result was established that the numbers of bacteria are 

 in relation to the amount of putrescible organic matter 

 in the water. 



The ideal value of the gelatine culture method not 

 having been realized, it is jDrobable that its true useful- 

 ness in water analysis has not been estimated as highly 

 as it deserves. An experimenter who has familiarized 

 himself with the distribution of bacterial life in waters 

 will be able to form definite and reliable conclusions up- 

 on the basis of numbers of bacteria. This is especially 

 true in the case of river water subject to j)olution by sew- 

 age from towns. Numerical determinations of abun- 

 dance of bacteria having been made of samples taken at 

 various points from the same river, a fair judgment may 

 be formed of the amount of sewage polution at any re- 

 quired point. The first step requisite to be taken is to 

 determine, for use as a standard, the numbers of bacteria 

 in unpolluted water in the stream under investigation. 

 Comparisons made with this standard give reliable quanti- 

 tative indications of polution. Any access of sewage raises 

 the number of bacteria above the normal for that stream 

 and the excess is a definite indication of the extent to 

 which the water has suffered polution. The standard is 

 obtained by testing the water, both at such points and at 

 such times as give the condition approaching nearest to 

 purity for that stream. In general, samples taken from 

 the head waters of the river, above the first town from 

 which sewage polution is received, and at a time of con- 

 tinued fair weather when the water is free from rain- 

 wash, are best suited for the control tests. In regard to 

 the effects of surface washings from the land by rains, as 

 indicated by turbidity of the water, it is necessary to el- 

 iminate them from all tests by taking samples only when 

 the water is clear. This rule being observed, comparisons 

 of results give indications of the extent of contamina- 

 tion due to sewage. 



It should be added that there are other conditions 

 which enter in a minor degree as factors in the results of 

 numerical determinations of bacteria. These are tem- 

 perature of water, depth at which the sample is 

 taken, point at which the sample is taken with 

 reference to rifts and pools in the stream, free ex- 

 posure to air and light (prevented in winter by ice), etc. 

 Consideration should always be given to these conditions 

 and as far as possible samples should be taken under 

 similar conditions throughout in order to render the re- 

 sults comparable. 



The writer, working in association with Prof. C. C. 

 Brown, consulting engineer for the New York State 

 Board of Health, in furtherance of his work in investigat- 

 ing rivers as sources of water supply, has made numeri- 

 cal determinations of bacteria for some six hundred sam- 

 ples of water from the Hudson and Mohawk rivers. A 

 statement of the results of this work is given in the an- 

 nual reports of the State Board of Health of New York for 

 the years 1891 and 1892. 



It naturally occured to us, early in the work here al- 

 luded to, that a method of differentiating sewage bacteria 

 from ordinary water bacteria would be of great value as 

 affording a more exact means of ascertaining the degree 

 of sewage pollution than is possible by the method out- 

 lined above. Dr. Theobald Smith, of Washington, D. C, 

 was then consulting bacteriologist for the New York State 

 Board of Health and upon submitting the idea to him he 

 informed us of a method of differentiating gas-producing 

 bacteria from others which he had devised and published 

 some time j)reviously {Centralblatt fur Bakteriologie, Vol. 

 VII., p. 302 and Vol. XII., p. 367) and which he believed 

 was applicable to the end sought by us. 



The method thus placed at our disposal consists in the 

 use of a culture fluid of which sugar (glucose) is a com- 

 ponent and which is placed for inoculation in tubes simi- 

 lar in princij)le to the ureometer employed by chemists. 

 Bacteria capable of causing sugar-fermentation when in- 

 troduced into such culture-tubes give rise to a gas the 

 quantity and composition of which can be ascertained. 

 In the apj)lication of this method to the bacteriological 

 analysis of water its value rests upon the fact that the 

 most common species of bacteria present in feces are gas 

 generators. As is well known the most constantly occur- 

 ing species of bacteria in feces is Bacillus coli cmimune; and 

 for some time our experiments related to the determina- 

 tion of the abundance of this species in the waters under 

 investigation by means of the characteristic quantity and 

 composition of the gases which it generates in the fermen- 

 tation-tubes. Later others of the more common fecal bac- 

 teria were isolated and studied with reference to their gas- 

 generating character. In this way a method was elabor- 

 ated by which, it is believed, there can be determined with 

 approximate exactness the numbers of prevailing species 

 of fecal bacteria in a unit quantity of water. This' deter- 

 mination is taken as a definite indication of the amount 

 of sewage pollution. 



In the practical use of this method the procedure is 

 as follows : The saccharine culture fluid contained in a 

 set, say eight, of fermentation-tubes is inoculated with a 

 measured quantity of water from the source of supply 

 under investigation. The tubes are immediately jjlaced 

 in an incubator and kept at a temperature of thirty-eight 

 degrees centigrade for forty-eight hours or somewhat 

 longer. (This is favorable to the develo23ment of fecal 

 bacteria and ptrobable destruction of the greater number 

 of ordinary water bacteria.) Those tubes in which gas 

 has been developed are then examined withreference to 

 the amount and composition of the gases j)resent and 

 note is taken of those which agree in these respects with 

 the effects produced by known fecal bacteria. Finally 



