80 THE ANALYST. THE INTERPRETATION OF THE RESULTS OBTAINED UPON THE CHEMICAL AND BACTERIOLOGICAL EXAMINATION OF POTABLE WATERS. JOHN C. THRESH D.Sc. M.B. D.P.H. F.I.C. Medical OficeT of Hecdth to the Essex Cozutty Cozcncil. PRIOR to obtaining a qualification in medicine and in public health and to becoming a Medical Officer of Health I had had a considerable experience in the analysis of potable waters and like most other analysts I was accustomed to declare a water good and safe or impure and dangerous according to the results of my analyses, basing my opinion upon comparison with standards arbitrarily set up either by analysts of repute or by learned societies who had given their attention to this subject. Under the conditions of my appointment as Medical Officer of Health the analysis of drinking-waters formed part of my duties and as such has led to my having examined probably 1,000 samples of water during the last five or six years.I was not long in discovering what other Medical Officers of Health had previously discovered that my analytical results often directly contradicted the results of my observations as to the source of the water and its effects upon the health of those who used it. My standards condemned waters as impure and unsafe which had never been known to produce any ill effect and which came from sources which so far as I could discover were perfectly safe and free from any injurious contamination. I n fact if I were to condemn every source of supply in my districts which yielded an excess for example of albuminoid ammonia and nitrates I should leave many localities entirely without a supply.On the other hand waters which I have had the gravest reasons for suspecting a8 the cause of disease yielded on analysis much more favourable results than others not suspected and yet upon close observation such waters were found either actually contaminated with sewage or liable to inter-mittent pollution. Further experience has also unfortunately brought me in direct conflict with analysts who have differed from me in opinion as to the quality of a water. The analyst judging from the results of his examination only; I basing my opinion not only on the analysis but also upon an examination of the source of supply to ascer-tain the possibility or otherwise of its contamination.My experience I am sorry to say is that of other health officers and it is my desire to avoid such conflict of opinions in future that has induced me to read this paper. Let it be understood that I am not now referring to difference of opinion arising from crass ignorance as in a case recently submitted to me in which a deep well water containing a considerable amount of chlorides was stated to be probably polluted with urine. Had the analyst known anything of the nature of the water derived from the source from which this came he would not have made such an absurd statement. Of the many impurities which drinking-water may contain the organic matter only is a serious source of danger and by far the greatest risk is incurred in using waters liable to contain certain living organisms which when introduced into th THE ANALYST 81 system are capable of producing specific disease.Of the presence or absence of such organisms chemical analysis can give us no information. The presence of organic matter may be chemically demonstrated but inasmuch as the nature of this organic matter whether poisonous or innocuous is beyond the power of the analyst to reveal, it is obvious that the results of a mere chemical analysis may often be worthless or even misleading. This point cannot be too strongly emphasized since the popular impression shared alike by the ignorant and the learned that a chemist by per-forming a few mysterious experiments with a water in his laboratory can pronounce at once whether it be safe or dangerous must be dispelled.This opinion has been fostered by those analysts who do not hesitate to pass judgment upon a water merely from the results of their chemical examination from the determination of the chlorides, nitrates phosphates and ammonia of the organic carbon and nitrogen and of the oxygen consumed or of the ammonia derivable from the organic matter. All these factors are of more or less importance as an index of the degree of pollution recent or remote but their real value can in very few cases be assessed without some previous knowledge of the source of the water. The inorganic constituents can easily be determined and whether either in quantity or quality these are objectionable the chemist can safely express an opinion. But this is more medical than chemical. For our purpose therefore only those salts need be considered which by their presence tend to throw some light upon the source of the organic matter always present in greater or less quantity.These are the chlorides phosphates nitrites, nitrates and ammonia and inasmuch as their determination is often of importance, the value of each may be discussed. C?zZorides.-In the great majority of instances the only chloride present is chloride of sodium or common salt ; occasionally other chlorides (calcium magnesium etc.) are also found especially if the water be derived from wells near the sea-coast. Rain-water always contains a trace of salt the amount varying with the distance from the sea. Certain geological formations also are rich in salt and waters obtained therefrom may contain considerable quantities of it.Urine again contains nearly 1 per cent. of salt ; hence pollution with sewage will add chlorine to the water. The effluents from many manufactories alkali works mines etc. are also rich in chlorine. From these various sources therefore the chlorides found in ,waters are derived. When the geological strata contain little or no salt and there are no manufacturing or mining effluents to pollute the water the amount of chlorides present may serve roughly as an index of the extent to which it has been contaminated by sewage. In Massachusetts it has been found that the amount of chlorine in the surface waters and streams decreases in amount from the seaboard westward or inland. By the examination of waters from sources removed from all risk of contamination, the normal chlorine for such districts has been determined.‘(By placing on the map of the State the amount of chlorine normally present in its unpolluted waters and then connecting the points of equal amounts lines of like chlorine contents are obtained which are called isochlors.” From the map thus prepared the normal chlorine is found to vary from -45 grain per gallon near the coast to less than -06 in the western part of the State (Board of Health Report 1892). Over any given area the amount of chlorine in excess of the normal as above ascertained can onl 82 THE ANALYST. be due to the influence of the population discharging its sewage thereupon. Assuming that 100 persons per square mile add on an average -03 grain per gallon of chlorine to the water flowing from that area the extent of the contamination can be approxi-mately calculated.I t must be remembered however that the amount of chlorine present does not necessarily signify present pollution. The organic matter which originally accompanied the salt and which alone is deleterious may have undergone complete oxidation and destruction so that organically the water may be very pure, although the amount of chlorine present indicates that at one time it was excessively polluted. This fact detracts very considerably from the importance of the chlorine determination. It affords some evidence of the previous history of the water and that is all. I n insular countries the estimation of the chlorine is of even less value, since they cannot be mapped out into isochlors.Over limited areas however the normal chlorine may sometimes be ascertained and any excess found in samples from that district will be in a measure proportionate to the present or past pollution of the water. For example in the parish of Writtle all the wells in which I recently examined the normal chlorine did not exceed 2.5 grains per gallon yet in that parish subsoil waters were found containing as much as 14.0 grains per gallon and that this was due to past and present pollution with sewage was substantiated by the excess of other substances especially nitrates which as we shallsee are also in most cases derived from the same source Unless this normal chlorine be known the deter-mination of the chlorides has no value whatever.The variation in the amount of chlorine in pure surface waters from various geological formations being known any excess over these amounts would probably point to past or present pollution and in any case would indicate that farther investigation of the source was desirable or necessary. In shallow-well waters even when pure the chlorine varies so greatly in amount that it is only in rare cases as in the one referred to above that the deter-mination affords any information of value. I n spring waters also it is difficult to decide upon the normal chlorine of any particular formation but if in any case the amount found greatly exceeds the average past or present pollution is indicated. The same remark applies to deep-well waters. I n several instances when examining water from the deep wells in Essex I have found that the chlorides present were below the normal and have been able to prove that this was due to surface water (usually impure) having gained access to the well.If the source of the water be not known, reliance upon the chlorine estimation may lead to serious error. The possibility of the excess of chlorine being derived from manufactories or mines must also be con-sidered before concluding that the water contains contaminating matter of animal origin. Nitrates and Nitrites.-The combined nitrogen found in drinking-waters may be present as organic matter or in ammonia nitrites and nitrates. Traces of all three are found in most samples of rain-water. Nitrogenous organic matter under-going putrefaction invariably produces ammonia and by oxidation this ammonia is converted by micro-organisms found in all soils into water and nitric acid the latter decomposing the carbonates present and forming nitrates of soda potash or lime.The ammonia however is not apparently converted directly into nitric acid but passes through an intermediate stage a lower oxide of nitrogen nitrous acid bein THE ANALYST. 83 ~~ first formed. The Rivers Pollution Commissioners found that whilst the organio matters contained in sewage and therefore of animal origin yielded abundance of nitrates and nitrites by oxidation no less than 97 per cent. of the combined nitrogen of London sewage being converted into nitrates by slow percolation through 5 feet of gravelly soil vegetable matters yielded mere traces of these compounds.Upland surface waters ‘* in contact only with mineral matters or with the vegetable matter of uncultivated soil contain if any mere traces of nitrogen in the form of nitrates and nitrites; but . . . as soon as the water comes in contact with cultivated land or is polluted by the drainage from farmyards or human habi-tations nitrates in abundance make their appearance.” Subsoil waters derive their nitrates in part from the oxidized ammonia of rain xater in part froin the slow decay of vegetable matter and in part from sewage matters. The amount derived from the two former is almost invariably small. Vegetable matter is not highly nitrogenous, and as a rule decomposes but slowly. Animal matter on the contrary decomposes rapidly and yields much ammonia.Nitrates serve for the food of plants and the active growth of vegetation may remove nearly the whole of these salts from a water. I n reservoirs the nitrates decrease gradually as the vegetable organisms increase. The total combined nitrogen therefore in a water may at one time exist in decaying animal and vegetable matter or in the form of ammonia ; at another in the form of nitrites and nitrates and yet again as a constituent of the protoplasm of living vegetable organisms in which latter case it is not in solution but merely suspended in the water. Whenever organic matter undergoes putrefaction in the absence of air or free oxygen not only are nitrates not formed but any nitrates present are decom-posed their oxygen being required for the formation of water and carbonic acid by combination with the carbon and hydrogen of the deconiposing substances.The nitrogen appears to be set free possibly accounting for the excessive amount of that element found in such deep-spring waters as those of Bath Buxton and Wildbad. I n this way the small amount of nitrates found in most deep-well waters is accounted for. Such being the case it is evident that even concentrated sewage may undergo such changes as would totally obscure its origin so far as the combined nitrogen is concerned. At first it would be contained chiefly in the dissolved animal impurities ; after passing through the surface soil it would exist chiefly in the nitrates formed by the oxidation of the organic matter ; later the nitrates may be decomposed and the nitrogen liberated when the water would be almost or entirely free from combined nitrogen.On the other hand certain deep-well waters especially in the chalk, contain very considerable amounts of nitrates which it is difficult to believe are derived from the oxidation of sewage matters. It has been suggested that these nitrates are derived from fossil remains or from natural deposits of nitrates or from vegetable matter ; but as no proof of these statements is forthcoming they must be received with reserve. I n the eastern counties the chalk wells yield waters which in some districts are absolutely free from nitrates (south-east Essex) whilst in other districts (Norfolk) they may contain possibly as much as 1 grain of nitric nitrogen per gallon. The following may be quoted as examples THE ANALYST.-84 _____ -Stratford Phcmix Works . . Wimbledon . . . . . Chatham Public Supply . . . Southend . . . Witham . . . Mistley Tendring Hundred W. W. Co. . Braintree Public Supply . . . Colchester (Donyland Brewery) . . Saffron Walden Public Supply . . Norwich . . . . . f , ?, Nitric N. per gallon. Depth of Well. feet. 200 200 490 900 600 160 430 305 46 Authorities. J. C. Thresh. R.6.C. J C. Thresh. T. A. Pooley. J. C. Thresh. In none of the above examples is there any possibility of recent sewage con-tamination. Notwithstanding these facts the Rivers Pollution Commissioners considered the total combined nitrogen to be an index of previous sewage contamination. They assumed that 100,000 parts of average London sewage contains 10 parts of combined nitrogen in solution.The mean amount of such nitrogen found in a large number of samples of rain-waters examined was -032 per 100,000 After deducting this latter amount from the amount of nitrogen in the form of nitrates nitrites and ammonia found in 100,000 parts of a potable water the remainder if any they say '' repre-sents the nitrogen derived from oxidized animal matters with which the water has been in contact. Thus a sample of water which contains in the forms of nitrates, nitrites and ammonia *326 parts of nitrogen in 100,000 parts has obtained -326 - 0032 = -294 part of that nitrogen from animal matters. Now this last amount of combined nitrogen is assumed to be contained in 2,940 parts of average London sewage and hence such a sample of water is said to exhibit 2,940 parts of previous sewage or animal contamination in 100,000 parts." The Rivers Pollution Commis-sioners however point out that on the one hand the nitrates may not indicate the full extent of the previous sewage pollution since the roots of growing crops take up much of the ammonia nitrites and nitrates contained in polluted water and animal matter which decomposes without access of air destroys nitrates ; and on the other hand that the nitrates present may indicate 10 per cent.of previous sewage con-tamination in deep wells and springs and the risk of using such waters be regarded as nil providing surface pollution be rigidly excluded. This 10 per cent. of previous sewage Contamination corresponds to one grain of nitric nitrogen per gallon.Mr. F. Wallis Stoddart in an excellent paper on ' The Interpretation of the Results of Water Analysis,"" describes a series of experiments in which he passed sewage containing cholera bacilli through a nitrifying bed of coarsely-powdered chalk, and found that although the organic matter in solution was completely nitrified yet the cholera bacilli or spirilla could be detected in the effluent. From the result of his own observations and experiments he concludes that natural waters can at most obtain from one-tenth to two-tenths of a grain of nitrogen as nitrates per gallon from X Practitioner 1893 ; reprinted in ANALYST xviii. 293 ; xix. 18 39 THE ANALYST. 85 sources other than animal matter,” and ( ( that practically the whole of the nitrogen of sewage may be oxidized into nitric acid without materially diminishing the risk in-volved in drinking it.” He urges that whenever the nitrogen as nitrates exceeds half a grain per gallon it indicates (I either dangerous proximity of the well to a source of pollution or such easy communication with it that the distance separating the two points is no guarantee of purification.” Many waters the source of which I carefully examined and which were collected and analysed by myself contain more than this amount of nitric nitrogen; and I am perfectly convinced that these waters are hygienically of the highest class and can be used without the slightest risk or danger.On the other hand in the table given at the end of this paper there will be found inany waters containing very much less nitrogen as nitrates which have almost certainly (in most cases the proof was most conclusive) given rise to outbreaks of typhoid fever.If Mr. Stoddart’s maximum of -5 be accepted as proof that a water is dangerous then the public and private water-supplies of many of our healthiest districts-districts remarkably free from outbreaks of typhoid fever-must all be considered dangerous. As a matter of fact the amount of nitrates which would condemn a water from one source may be absolutely without significance in another, all of which goes to prove as will be shown in the sequel that mere chemical analysis is absolutely powerless to prove that any water is of such a quality as to be incapable of producing disease amongst those who drink it.Nitrites may result from the oxidation of ammonia or from the reduction of nitrates and as it is an easily oxidizable compound its presence indicates a condition of instability of matter undergoing change Usually this matter is of animal origin and derived froin manure or sewage the ammonia produced by their decomposition being in process of oxidation to nitrates. Where the soil is not sufficient in quantity, or is defective in quality the oxidation may be incomplete and incompletely purified and probably incompletely filtered water is the result. Usually in such cases an ex-cessive amount of ammonia is also present. But though usually this is not invari-ably the source of the nitrates and ammonia. Where nitrates are present the nitric acid may be reduced by contact with metals such as iron or lead forming tne pipes in which the water is conveyed or lining the upper portion of the well.Where such is the case a trace of the metal can always be detected in the water. Unless this fact be borne in mind-and it often appears to be overlooked-a good and wholesome water may be classed as dangerous or polluted. Certain organisms also found in water are capable of reducing nitrates to nitrites. Still the presence of nitrites always renders a water suspicious and their origin should be carefully investigated. Ammonia.-All rain-water contains this compound as does also melted snow. The first portions of a shower and the rain collected in the neighbourhood of towns are richest in ammonia.As an average 602 grain per gallon taken by the Rivers Pollution Commissioners is probably fairly approximate but the variation is very wide ( a 2 to -01). I n passing over or through the ground the ammonia is rapidly oxidized and by the time the water reaches a stream or the general body of subsoil water most of it has disappeared. Rain-water stored in covered cisterns however, usually retains its ammonia for a considerable period. In such waters therefore the ammonia unlesrs excessive is devoid of significance. Many deep-well waters als 86 THE ANALYST. contain much ammonia the origin of which has given rise to a good deal of surmise. The generally-accepted theory is that it is due to the reducing action of ferruginous sands on the nitrates present. This may be so in some cases but my observations lead me to believe that it is often due to the reduction of the nitrates by the metal of the bore tube pump pipe and lining of the well.I was led to this conclusion from the fact that I found the water from one and the same well at one time quite free from ammonia and at another containing as much as one part of ammonia per inillion parts of water. I n the water containing ammonia I also found a very faint turbidity which cleared up on the addition of a little acid and gave the reactions for iron. The clear ammonia-free water also when stored for a time in an iron tube, became turbid and nitrites ammonia and iron could be detected in it. Generally, however the ammonia found in river spring and well waters is derived from putres-cent animal matter-that is from manure and sewage ; but before this conclusion can be safely drawn the other possible sources must be excluded.Dr. Brown in his Report to the Massachusetts State Board of Health 1892 whilst agreeing that im-perfect oxidation of sewage matter is generally the source of the ammonia calls attention to the fact that several waters in the State free from such pollution contain a con-siderable amount of free ammonia. “ They are all associated with iron oxide and the fungus Crenotlzriz.” Such waters are found also in many swampy regions and in wells sunk in ferruginous river silt and usually become turbid from the formation and deposition of oxide of iron when exposed to the air. The odour of these waters is said to be “ often disagreeable from dissolved sulphuretted and carburetted hydrogen.” Phosphates.-Phosphates are found widely distributed in rocks and soils and traces may be dissolved by a water rich in carbonic acid.Albuminous matters whether of animal or vegetable origin give rise to phosphates by their decay hence an ex-cessive amount present in a drinking-water is held by many analysts to indicate contamination The difficulty of determining their presence and amount is con-siderable and as the value of the determination is very doubtful few analysts pay any attention to them. They may be present in a pure water and their absence is no proof that a water is free from pollution. Orgarzic Matter.-By no known process can either the quantity or quality of the organic matter in water be determined.When a known volume of water is evaporated to dryness the weight of the residue is that of the inorganic and organic substances contained therein. When this residue is ignited the organic matter is destroyed and volatilized and the loss on ignition” has been regarded as approximately expressing the weight of the organic constituents. Such however, is rarely the case since carbonic acid may be driven off from the carbonates present, and any nitrates present will be more or less completely reduced. Moreover some salts retain water so tenaciously that the whole is not driven off at the temperature used for drying and this moisture is given of€ when the residue is ignited. For these reasons chiefly the ‘( loss on ignition ’’ cannot be depended upon as an index of the amount of organic matter present.But although the total amount of the animal and vegetable substances cannot be determined the carbon and nitrogen therein can be ascertained by careful chemical analysis. Not only so but the authors of the original procees believed that with certain reservations the propor THE ANALYST. 87 tion of the nitrogen to carbon indicated whether the organic materials are derived from the animal or vegetable kingdom. In fresh peaty water the Rivers Pollution Commissioners found that N C = 1 11.93 whilst in similar waters which had been stored for weeks or months in lakes N C = 1 5.92. After such water had been filtered through porous strata N C = I 3-26 In fresh sewage the average of a large number of samples gave N C = 1 2.1.Highly polluted well-waters soakages from cesspools etc. gave N C = 1 3.126. In sewage after filtration through soil the proportion of N to C rose from 1 1.8 to from 1 4.9 to 1 7.7. Evidently there-fore the ratios of N to C (‘ in soluble vegetable and animal organic matters vary in opposite directions during oxidation-a fact which renders more difficult the decision as to whether the organic matter present in any given sample of water is of animal or vegetable origin.” All nitrogenous organic matter whether of vegetable or animal origin yields more or less ammonia when boiled with a strongly alkaline solution of permanganate of potash and the ammonia -so yielded by potable waters is called albuminoid,” or (‘ organic ” ammonia. The proportion of nitrogen in the ammonia so yielded to the total nitrogen in the organic matter is unfortunately not constant ; but the chemists to the Massachusetts Board of Health believe that when the process is performed as in their practice about one-half of the nitrogen is converted into ammonia.Albuminoid substances of animal origin contain about 16 per cent. of nitrogen but vegetable matters contain very much less ; hence the amount of ( ( albuminoid ” ammonia is no index to the amount of organic matter present in the water. Mr. Wanklyn who devised this process considers that undeniably contaminated waters always yield an excessive amount of albuminoid ammonia (over *I0 part per million) ; usually with much free ammonia (over a08 part per million). If the albuminoid ammonia distils over very slowly and is in excess but the water contains little free ammonia and very small quantities of chlorides Mr.Wanklyn considers this an indication that the contaminating matter is of vegetable origin. He adds “ The analytical characters as brought out by the ammonia process are very distinctive of good and bad waters and are quite unmistakable.” The generally accepted opinion however is that no reliance can be placed upon these determinations taken alone and in the Nassachusetts State Board of Health Report for 1890 there is quoted as an example the results of the analyses of certain of the Boston water-supplies. Reservoir No. 4 is known to contain the purest water, but the average ‘( albuminoid ammonia ” during two years was a26 per million.The water of the Mystic Lake is the worst of the Boston waters since it contains both sewage and manufacturing refuse ; yet during the same period the average albuminoid ammonia was exactly the same as in the purer water. I n the table appended many other examples will be found of the erroneous conclusions which may be drawn from a too implicit reliance upon the determination of the ammonia yielded by distillation with alkaline permanganate. Forschammer devised a process for the estimation of the amount of oxygen required to oxidize the organic matter in water. This method as improved by the late Dr. Tidy has become very popular and many attempts have been mado to render the results comparable with those obtained by Frankland’s proces8 in whic 88 THE ANALYST. Uplend Surface Water.the amount of organic carbon and nitrogen is ascertained by combustion but with only partial success. The results when compared with those obtained by the “ albuminoid ammonia ” process prove that there is no relation between the amount of ammonia yielded by a water when distilled with an alkaline solution of permanganate of potash and the amount of oxygen absorbed when the water is digested with an acid solution of the same salt. This process tells us little or nothing of the nature of the polluting material ; it does not even distinguish between organic matter of vegetable and animal origin and it affords us no evidence of the amount of such substances present. The presence of certain bodies of mineral origin (sulphuretted hydrogen nitrites the lower oxides of iron etc.) also cause absorption of oxygen and unless great care is taken to ascertain the absence of these or to ascertain the exact amount of oxygen consumed by them if present serious errors may be introduced.When these corrections are made the oxygen process is still open to all the objections which have been urged against the albuminoid ammonia process. It may condemn a perfectly harmless water as polluted and pass as of good quality a water of most dangerous character. The following table was devised by Drs. Tidy and Frankland : Water Other than Surface Water. AMOUNT OF OXYGEN ABSORBED BY 1,000,000 PARTS OF WATER. Not more than 1.0 9 I f 3.0 9 9 9 4.0 More than . 4.0 Water of great organic purity . , medium purity . . l doubtful purity .. . . Impure water . . . . Not more than -5 9 1 7 9 -15 9 1 ? *2 More than . 2.0 When the quality of a water is considered from the biological side instead of the chemical the absurdity of dividing waters into classes of pure medium doubtful purity and impure is obvious. A water containing a poisonous amount of typhoid bacilli might upon analysis be brought within any of these classes according to the quantity and quality of the accompanying impurities. I n the analyses given below there are instances of waters coming within Tidy’s limit of ‘‘ great organic purity,” yet which proved to be capable of causing disease. I have examined many such waters myself and have also passed many waters as perfectly safe for domestic purposes which a mere reference to the above standards would have condemned as doubtful or impure.Many other special processes for determining whether a water be safe or dangerous have been devised but inasmuch as they are rarely used it may safely be inferred that they possess no advantage over those to which we have already ref erred. Sir Charles Cameron in his ‘ I Address on Public Health,” delivered before the British Medical Association at Bristol (1894) says I ( We have not yet determined, in relation to chemical analysis the line which divides a bad from a good water no TEE ANALYST. 89 do I believe that such a line can be drawn. One that contains very little organic matter may be dangerous to drink whilst on the contrary water with a large amount of organic matter has been drunk with impunity for years.I t is unfortunate that there is no accurate chemical method for determining the precise nature of the organic matter found. As a rule the larger the percentage of nitrogen contained in the organic matter the more likely is it to have been derived from an animal source. Analysis certainly shows the amount of organic matter present and although such matter is usually innocuous it is sometimes the reverse. The greater the amount of organic matter or rather of albuminoid matter the greater the chance that it may include infective substance. When we have a water with a very small amount of nitrogenous organic matter there is no likelihood that typhoid dejections will be present. Under the ordinary conditions affecting our water-supplies it is improbable that typhoid stools would be mixed with them.They are thrown into water-closets, privies ashpits manure heaps the soil ditches rivers etc. From all these media they may find their way into rivers or wells furnishing potable water but with rare exceptions there must under such circumstances be large quantities of other kinds of organic matter associated with the typhoid excreta and carried with it into the water. I n short the passage of unmixed excreta into potable water must be a circumstance of rare occurrence. In the ordinary cases of pollution by typhoid or other dejecta the analysis of water proves its unfitness for use.” The latter statement is directly contradicted by the results of the chemical examination of very many samples of water which have given rise to epidemics of typhoid fever.A reference to the table already referred to will show that many most extensive epidemics have been attributed to the use of waters which chemists of eminence had declared to be of a high degree of purity. I n these c8ses also repeated analyses failed to prove ‘‘ unfitness for use.’’ Whilst no single determination will enable the analyst to certify that a water is free from danger or that it is so polluted as to be dangerous to health the deter-mination of several constituents may enable him to pronounce it to be polluted and dangerous but will never justify him in certifying that it can be used absolutely without risk. As the freedom from all dangerous polluting material is the informa-tion usually sought from the analyst it follows that if this cannot be ascertained by analysis a chemical examination is in most cases quite useless Where a water is known to be contaminated with sewage or known to be liable to such pollution an analysis is superfluous.When we also consider that many sources of supply are only subject to intermittent pollution and that waters from the same reservoir or from the same well (vide Analyses Nos. 24 25 and 26 27) may vary considerably in composition according to the depth from which the samples are taken the character of the season etc. it is obvious that the chemical examination of a water is a matter of comparatively trifling importance compared with the thorough examination of the source of the water and a knowledge of its history. Frequently waters are sent for analysis and the analyst is wilfully kept in ignorance of their source lest the information should prejudice his report yet without this knowledge he is not justified in expressing the opinion that the water from any source can be used with safety.In commenting upon a recent pa8per in which I expressed these views a writer i 90 THE ANALYST. the Chemist and Druggist says “ It would seem therefore that we are face to face with the question ‘Is water analysis a failure?’ It has been so exclusively the province of chemical analysts to pronounce judgment upon domestic waters and they generally have given so little attention to the large issues attached to analysis, and so very much to sets of standard figures for chlorine nitrogen hardness and so on that the attack from the medical health side is not unexpected.There has been more wrangling over water analyses than over anything else in chemistry-and for what ? Some figure in the second or third place of decimals probably and in regard to what this ammonia or that ammonia implies when a visit to the source of the water and an inspection of the sewage trickling into it might settle everything. That is what Sir George Buchanan and Dr. Thresh advocate.” The Royal Commission on Metropolitan Water Supply received evidence proving that waters containing very large amounts of organic matter were drunk continuously by a population with perfect impunity whilst other waters containing so little organic matter BB almost to defy chemical detection had proved time after time to be of the most poisonous character.For these reasons they conclude that the water question has passed from the domain of chemistry into that of biology. This however is not strictly correct. The biological problems involved in the investigation of water supplies are numerous and complex and as yet but imperfectly understood. At the present time it is doubtful whether a biological examination really tells us more than a chemical analysis and very often it cannot tell us a8 much. The reason will be explained shortly . With the discovery of the fact that such diseases as typhoid fever and cholera are due to the introduction into the system not of dead organic matter but of actual living organisms faith in the chemical analysis of waters began to be shaken.When still more recently the actual microbes causing these diseases had been identified, and processes had been devised for isolating them froin the multitude of other organisms found in water it seemed as though the examination of water for sanitary purposes had passed from the domain of the chemist to that of the bacteriologist. The study of the number and character of the bacteria it was hoped would enable the biologist to definitely pronounce whether a certain water was capable of causing disease or whether it was perfectly harmless in character. Up to the present time such hopes have not been realised and the results of an ordinary bacteriological examination are as likely to be misleading as those of a chemical analysis. The reason for this is not difficult to explain when the significance of certain of the discoveries made by bacteriologists is thoroughly understood.An enormous number of species of bacteria have already been discovered although the science is in its infancy. They are almost ubiquitous abounding in the air water and nearly all articles of food and drink Of this immense variety very few appear to be capable of causing disease ; the remainder are perfectly harmless to human beings whilst nzany are already known to discharge most important functions in the economy of nature. Upon their presence the fertility of soil in a great measure depends; they break down the dead organic matter into the simpler forms which can be assimilated by the roots of plants. By their action the foul organic constituents of polluted water are converted into carbonic and nitric acids which in combination with the minera THE ANALYST.91 bases form innocuous carbonates and nitrates. They are in fact nature’s scavengers, consuming the foul and effete and producing therefrom food for the living. [The remainder of this paper and the discussion which ensued on it and the preceding paper wzll be ptbbliyhted in our next issue.] Purification of Streams by Green Plants by Bacteria and otherwise.-Researches on this question by Dr. T. Bokorny (Amh. f. Hygieiz 1894 xx. No. 2) and Duclaux (Ann. Inst. Pasteur 1894 viii. pp. 117-127) are abstracted in Royal Micro. SOC. Jozmz. 1894 p. 714. Bokorny remarks on the well-known fact that green plants can assimilate organic substances and the apparent certainty that aquatic bacteria play a very important part in the purification of water containing a con-siderable quantity of organic matter ; and then recounts his own experiments, chiefly with algz and especially with diatoms.Volatile fatty acids amido-acids, indol skatol phenylacetic acid urea etc. were exposed in suitable solutions-( a ) with free access of light and ( b ) with exclusion of light-to the action of a l p . The results showed that algs were capable of using as nutriment volatile fatty acids, such as butyric and valerianic and also glucose leucin and tyrosin. Even urea, when properly diluted could be assimilated. I t seems clear therefore that many organic impurities that are drained into rivers are destroyed by water-plants as well as by bacteria.Duclaux’s work concerns the purification of river-water by bacteria and its ‘‘ spontaneous purification,” only. After alluding to the condition of the Seine above Paris and the quantity of impurities deposited in the river from the city he shows that lower down the river has again become practically pure He quotes Prausnitz’s researches on the water of the Isar which when entering Munich, contained 305 germs per c.c. but seven kilometres below the city 12,600 per c.c.; whilst thirty-three kilometres below Munich (eight hours’ flow) the number of germs had been reduced to one-sixth. The causes of this purification according to Duclaux are both physical and vital-particularly notable among the former being the action of the ground-waters which mingle with the stream along its course and which influence purification partly from their lower temperature in summer partly from their chemical composition.Another prominent physical mode of purification is the deposition of floating particles all of which on account of their density tend to fall to the bottom and drag down other matters with them this action being aided by the molecular adhesion of water and the particles; but the purification is effected more surely when an organic or mineral precipitate is formed than when these bodies are left simply to the action of gravity or of molecular adhesion. The principal agent in the purification is however the microbe; the r6le of the physical agencies being subsidiary thereto. F. H. P. C 92 THE ANALYST. ___ ._ _-___ ~ ~ On the Micro-Chemical Examination of Quinine.I€. Behrens. (Reczceil des Travaux Chzim. des Pays-Bas xiii. 1 pp. l-l2.)-For the detection of impurities in quinine a series of crystallizations is employed and the form of the crystals examined under the microscope : 1. Crystallization qf Quinine Bisdphate.-From 0.05 to 0.2 grammes of the impure sample is warmed with two parts of water and sufficient H,SO to retain three-quarters in solution and crystallized at as low a temperature as possible. If less than two-thirds of the solid taken be obtained the mother liquor must be slightly concentrated on the water-bath and cooled. The crystals obtained are pure bisul-phate while the other alkaloids are left in the mother liquor. The crystals are separated by pressure with a platinum spatula and washed with several drops of water.2. Crystallization of Nornzal Quinine Su1phate.-Small quantities of NaHCO, solution are added to the mother liquor with warming and shaking until neutral, leaving the liquid clear. On cooling a large quantity of the sulphate is deposited in long needle-shaped crystals which are separated from the mother liquor as before. 3. Second Crystallization of Normal Su1phate.-The neutral mother liquor is evaporated to dryness over H,SO, or over a water-bath larger crystals being obtained by the former process. The crystals dissolve '.very slowly in cold water, while the sulphates of quinidine cinchonine and cinchonidine may be extracted by adding several drops. A small quantity of this liquid may be warmed a drop of NaHCO solution added and the crystals formed examined.Lumpy crystals or rhombs point to the probable presence of quinidine short needles to cinchonine and cinchonidine while a powdery deposit changing on boiling into microscopic drops, denotes a considerable residue of quinine. In the last case the quinine should be removed by precipitation with potassium oxalate before proceeding. 4. Fractional Precipitation.-Quinidine is first removed by adding a saturated solution of KI which forms a cloudy precipitate in a few seconds. Cjnchonine and cinchonidine are not precipitated by this reagent. After removal of the greater part of the quinidine fractional precipitation is employed. The liquid warmed to 40" C., is stirred with a platinum wire dipped in NaHCO solution and the deposit examined under the microscope.If rectangular rod crystals of cinchonine are observed on cooling they should be left a few minutes to develop. If outside these there are needle-shaped crystals of cinchonidine the whole should be warmed several times to boiling the evaporated water being replaced After this large well-developed crystals of cinchonine may often be found at the edge of the drops. When undoubted crystals of cinchonine have been obtained the mother liquor is allowed tocool at one side and the cinchonine crystals fixed to the glass washed with a drop of water. A second precipitation is then made in the same way when mixed crystals are often obtained some of them being the characteristic needles of cinchonidine. Only in the presence of a large quantity of cinchonine is a third crystallization necessary.The author employed the method successfully in the examination of a mixture consisting of 0.1 gramme impure quinine sulphate containing 0-5 per cent. cinchonine, 0.3 per cent. cinchonidine and 0.2 per cent. quinidine. C. A. M THE ,4NALmY ST. 93 The Analysis of Varnishes. Parker C. McIlhiney. (Jour. Amer. Chem. SOC., 1894 348.) After criticising the various methods which have been hitherto proposed for the separation and estimation of the turpentine and other volatile constituents of a varnish the author proposes the following. I t is based on the well-known fact that on distilling a mixture of two immiscible liquids the boiling-point is lower than that of the more volatile constituent.On distilling together 100 C.C. of water and 5 C.C. of turpentine the first 95 C.C. of the distillate contains all the turpentine which readily separates from the water. Experiment showed that 90 C.C. of water either dissolves or holds in suspension about 0.3 gramme of turpentine. On these principles the following process was devised Twenty-five grammes of varnish are weighed into a flask of 400 C.C. capacity in which has been placed a piece of granulated tin or its equivalent to prevent bumping and 10 C.C. water. The contents of the flask are now submitted to distillation the distillate being caught in a tapped separator. When from 90 to 95 C.C. of water have come over the distillation is stopped and the turpentine allowed to separate. Should the contents of the flask still retain the odour of turpentine more water is added and the distillation resumed.The separated turpentine is weighed in a tared flask and a correction of 0.3 gramme for every 90 C.C. of water made. The results obtained were satisfactory. In some experiments salt was added to the water but no apparent advantage was gained except when the non-volatile residue was heavier than water. I n this case the salt prevented the residue from sinking to the bottom of the flask and causing bumping by greasing the flask. If it is desired to examine the residue the remaining water is poured off, alcohol added this distilled off and the last traces of alcohol removed by evaporation with a little ether. The residue is not heated above the boiling-point of water ; the turpentine is weighed as wch and may be subjected to further examination.For the determination of petroleum products in turpentine the method of Burton gives the best results. It depends upon the conversion of the turpentine into acids soluble in water by the action of nitric acid while the other products remain unaffected. The analysis of the residue which consists of linseed oil and one or more resins all more or less altered by heat presents an extremely difficult problem, which as yet remains unsolved. The author has applied to these residues the same tests as have been appliid to oils-viz. the acid figure the Kcettstorfer figure the iodine and bromine figures the bromine addition figure and the solubility in ether, alcohol etc.* He arrives at the following conclusions The differences in the solubility between linseed oil and melted resins are but slight and of little value in effecting separations.The Hub1 figure is evidently useless as a means of quantitative analysis for the figures of linseed oil and kauri the most frequently occurring resin, are almost identical. The difference in the Kettstorfer figure is only about 75 and even supposing that the figures of both oil and resin are accurately known which is not often the case it would be difficult to make the analysis so carefully that the percentages of oil and resin would be correct. The acid figure also fails when applied to a varnish though it proves correct with a mixture of common resin and linseed oil which has not been heated too long. The bromine absorbed is of no more * A long list of these figures for a large number of reuina and the correaponding ones for raw and boiled linseed oil are given in the paper 94 THE ANALYST.use than the Rub1 figure for its quantity depends largely on the conditions under which the analysis is performed. As the bromine-addition figure of resin is 0.0 and that of linseed oil 102 this promised to be of some use; but it completely broke down on account of the changes made by the process of manufacture in the properties of oil and resin. Gladding’s process for the estimation of common resin in mixtures with fatty acids when tried on varnish also failed to give satisfactory results. The author concludes that it is evident the processes used in oil analysis are not adapted to the analysis of varnishes.The oil and resin react upon one another in some way not understood giving rise to new compounds and it is necessary that something further should be known as to the nature of these before this analytical problem can be solved. W. J. S. Tho Viscosimetrical Examination of Butter. Dr. C. Killing. (Zea’t. fur angcwan. Chemie 1895 Heft 4 pp. 102 103).-Since the publication of his former communications,” the author has examined a large number of viscosity apparatus, and has found that two may give identical times for water at 20’ without doing so for One and the same fat. The latter will only be similar when the body of the pipette delivery tube etc. are of exactly the Bame dimensions. From this it follows that the time taken by water to run out cannot be taken as a standard for fats and, accordingly the viscosity number which depended on this relation falls to the ground.This however makes no difference in the use of the method but only in the manner of stating the results. Each apparatus before being sent out is now standardized for the mean ‘( running out” time of butter as well as for that of mar-garine and is thus ready for immediate use. Since lard has occasionally been used to adulterate butter the author gives the time taken by it in running out and also that of beef fat. I n an apparatus giving for butter the mean time of 3 min. 43.5 sec. and for margarine 4 min. 19 see. the follow-ing results were obtained : Lard. min. aec. 1. 4 29-00 2. 4 26.75 3. 4 30.25 4. 4 26-25 Beef Fat. min.aec. 1. 4 36-50 2. 4 33.50 3. 4 31.00 As these both take a longer time readily than the latter. margarine they would be detected even more C. A. M. Note on the Adulteration of Commercial Peptones. BE. I,. Hugounenq. ( J o ~ r t de Pharm. et Chim. 1885 pp. 179 180.)-Two samples of peptone examined by the author gave an abundant reduction with Fehliug’s solution. The solution of the peptone when warmed on the water-bath for an hour with a mixture of sodium acetate and phenylhydrazine chlorhydrate deposited on cooling a magma of needle-shaped crystals which after purification by washing and recrystallization from boiling water melted at 205°C. This melting-point of the osazone pointed to the * See ANALYST xx. p. 66 THE ANALYST. 95 presence of ordinary glucose.But the absence of alcoholic fermentation in the SUS-pected liquid and the fact that the compound with phenylhydrazine although closely resembling phenylglucosazone in the form of its crystals its melting-point and behaviour towards organic solvents was yet distinguished by its solubility in boiling water showed that the substance was probably lactose. This was confirmed by Riibner’s reaction in which when a dilute solution of lactose is saturated with a hot neutral solution of lead acetate and ammonia added drop by drop to the boiling solu-tion the liquid changes to yellow then to orange and finally to red. With pure peptone this does not occur while in the two adulterated samples it did so very markedly. The amount of milk sugar estimated with Fehling’s solution was found to be upwards of 40 per cent.in each case. A good peptone from the Lyons hospital yielded 12.72 per cent. while in the two adulterated peptones only 8.34 per cent. was found. The Kjeldahl determination of the nitrogen was also very conclusive. C. A. M. Chemical Process for Recognizing Horse-flesh. M. Humbert. Recueil de medicine veteriwaire (through Jown. de Pharm. e t Chem. 1895 pp. 195-198).-1n the year 1892 over 20,000 horses were slaughtered in Paris for the purpose of food. Much of their flesh was used in sausages the vendors of which were compelled t o indicate the nature of what they sold. I t is thus a matter of great importance to be able to detect horseflesh when mixed with other kinds of meat. The Belgian chemists, Brautigam and Edelmann have recently shown that it could only be confounded with the flesh of the human fcetus and the fcetus of animals.For its detection the reaction of iodine on broth obtained from the flesh is used both by these chemists and the author. The latter proceeds as follows ‘‘ About 50 grammes of the muscular tissue are cut into sinall pieces and boiled for an hour with 200 C.C. of water. After cooling, nitric acid is added in the proportion of 5 C.C. to 100 C.C. of broth and the liquid filtered. To some of the filtrate in a test tube iodine water is added drop by drop so as not to mix the liquids and a reddish violet circle will show that horseflesh was present in the original sample. The iodine water is prepared by saturating hot water with iodine. When the reaction is doubtful it is advisable to replace water by a 3 per cent.solution of KOH. The reaction is more pronounced when the iodine water has been recently prepared. Of ten samples of horseflesh obtained from different dealers in Paris seven showed the colour very plainly in two it was less pronounced but still clear while in the last it was doubtful. I n no case was there any coloration with beef veal mutton or pork. Beef broth left in contact with the iodine for ten days showed no signs of change. The flesh of the ass also gave a negative result but with that of the mule the reaction was the same as with horseflesh. A mixture of equal parts of horseflesh beef veal mutton and pork showed the coloration but it was less pronounced than with horseflesh alone. C.A. M. The Determination of Phosphoric Acid. H. Pemberton Jun. (Jour. Amer. Chem. Soc. 1895 xvii. pp. 178-181.)-The author cites several precautions to be observed in using his method : I n standardizing the alkali phenol-phthalein pzot methyl orange must be used a 96 THE ANALYST. the indicator. Since KOH as bought usually contains alumina the neutral reaction is consequently shown with methyl orange before it appears with phenol-phthalein. When precipitating the phosphoric acid with the aqueous solution of molybdate, the liquid must be removed from the flame as soon as the boiling point is reached, and no more heat appZied. The temperature on adding the molybdate will then be about 70° and no water-bath is required. In order to obtain the precipitate in a granular state the molybdate solution should be added very slowly from the pipette the flow being interrupted by closing the end with the finger and the solution stirred as it goes in. The washing of the precipitate which is a difficult matter is best done on a Hirsch funnel or in a Gooch’s crucible with the aid of a suction pump. A pre-cipitate containing about 50 milligrammes of phosphorus pentoxide is about the largest quantity it is advisable to work on and where suction is not used 30 to 40 milligrammes should be the limit. A precipitate containing 40 milligrammes is thoroughly washed on the Hirach funnel by 300 C.C. of water. C. A. M