June 6, 1 872 J 



NATURE 



105 



sion that every grain of oxygen oxidised eight grains of 

 " organic matter." 



Whether this lest is to be trusted may be judged from 

 the following facts. Potassic permanganate is deoxidised 

 by ferrous salts, nitrites, sulphites, &c., much more 

 rapidly than by organic substances, so that a water abso- 

 luteh- free from any organic matter, but containing one of 

 these compounds, would be set down as requiring so much 

 " oxygen to oxidise organic matter." Secondly, in the 

 case of water to which known weights of various organic 

 compounds were added, Frankland and Armstrong 

 found that in no instance was the oxidation complete, 

 even after the lapse of six hours. In fact, even after 

 that time the amount of oxygen actually absorbed was 

 in every case a mere fraction of the quantity actually 

 required to completely oxidise the organic substance. 

 The test, though thus shown to be valueless as quantita- 

 tive, is of some value qualitatively, as it can be easily and 

 quickly applied ; and it may be said that, though it might 

 induce a person to abandon a good water, it would not 

 often lead him to use a bad one. 



All the above processes were in use up to 1868, when 

 Messrs. Chapman, Wanklyn, and Smith proposed to de- 

 termine the organic matter in water from the amount of 

 ammonia evolved when the water was treated with a 

 strongly alkaline solution of potassic permanganate and 

 then distilled, the ammonia being determined in the dis- 

 tillate byHadow's modification of the Nessler test. 



That albumin is decomposed and the nitrogen thus 

 evolved, they had shown in a paper presented to the 

 Chemical Society in the preceding year. The way in 

 which this process was applied to the water may be briefly 

 stated as follows : — A measured quantity of the water 

 was rendered alkaline with freshly-ignited sodic carbon- 

 ate, and the ammonia distilled oft' and estimated by 

 Hadow's modified Nessler process. As soon as all the 

 ammonia thus obtainable had been expelled, the alkaline 

 permanganate solution (50 cubic centimetres of a solution 

 containing 200 grammes of potassic hydrate and S 

 grammes potassic permanganate per litre) was run in. 

 The distillation was then resumed, and the ammonia esti- 

 mated as before. This last was set down as "albuminoid 

 ammonia," and as the average evolution of ammonia from 

 the following substances, gelatine, caseine, dry albumin, 

 uric acid, creatine, theine, dried fish flesh, amounted 

 to 10 per cent., it was suggested the albuminoid ammo- 

 nia multiphed by 10 gave a fair estimation of amount 

 of organic matter. 



It had at first been stated that albumin gave up the whole 

 of its nitrogen when treated with alkaline permanganate, 

 but the statement w^as subsequently modified to " It appears 

 to be two-thirds, being at any rate a constant quantity." 

 Now this process would indeed be a valuable one if the 

 10 per cent, average could be depended on, or if the albu- 

 min evolved a certain quantity, and the above substances 

 were alone found in water. Unfortunately, none of these 

 suppositions are true. With regard to the last, the authors 

 themselves recognised the difficulty, and accordingly ex- 

 amined a number of other nitrogenous organic bodies ; 

 which examination led to the publication of two lists of 

 bodies that ev'olve the whole of their nitrogen as ammo- 

 nia, and bodies that yield various fractions. Frankland and 

 Armstrong nlso made some experiments on this subject. 

 Willi regard to the list of bodies yielding half their am- 

 monia, the numbers given by the authors ^ ary from 44 per 

 cent, in the case of papaverine to 58 per cent, in the case 

 of sulphate of cinchonine ; and whilst narcotine appears 

 in Wanklyn, Chapman, and Smith's list as evolving all its 

 nitrogen, Frankland and Armstrong give it as evolving 

 about 46 per cent. Strychnine, given by the former 

 authors in their list as evolving f53 per cent., is given 

 by the latter as evolving 31^ per cent., and sulphate 

 of quinine also in the list with 50 per cent., appears 

 again with Frankland to have evolved nearly 57 per cent. 



No other examples will be necessary to show the extreme 

 uncertainty of the process. If the authors had enabled us 

 to ascertain the absolute error on the quantity taken in- 

 stead of the per-centage error, by giving us the quantities 

 from which the results were taken, it would no doubt be 

 much more apparent ; the results given al)ove in the case 

 of those from Frankland and Armstrong's paper are abso- 

 lute errors. 



But it may be urged in defence of the method that none 

 of these bodies are found, or are likely to be found, in 

 natural waters. Let this be granted, and the process 

 must be defended on the albumin and other bodies men- 

 tioned, and on the list of bodies giving up all their am- 

 monia. The average from this is 1 1 '82 per cent. ; from 

 the albumin list, 9^92 ; or, taking the two, lo'S/. So far 

 the lists hold good. But it must be borne in ir.ind that 

 we have three different statements about the ammonia 

 evolved from albumin — first, that all is evolved ; secondly, 

 " two thirds, or, at an)- rate, a constant quantity ; " 

 thirdly,* tliat 100 parts of albumin give 10 parts of 

 ammonia.f The inconsistency of these statements 

 needs but this comment, that they can only be caused 

 by the extreme uncertainty of the process ; in fact, 

 the amount of nitrogen converted into ammonia will 

 be influenced by the nature of its previous combination, 

 the degree of concentration of the solution, and the 

 amount of heat applied to the retort, and consequent rate 

 of distillation and time to which the solution is exposed 

 to the action of the alkaline permanganate. That this is 

 the case is proved by the fact that water which has been 

 distilled from alkaline permanganate, and gives no trace 

 of reaction with Nessler's test, will evolve ammonia if 

 again boiled with the permanganate. Lastly, let any one 

 take a water which has been largely contaminated with 

 sewage and then filtered, such as the effluent water from 

 a sewage farm. Such water, as a rule, contains much 

 nitric and nitrous acid, and comparatively small quanti- 

 ties of organic nitrogen. A water of this character con- 

 tinues to evolve albuminoid ammonia till boiled nearly to 

 dryness, and not unfrequently the retort requires to be 

 filled up with pure water, and the operation carried on. 

 The process is thus not only rendered tedious, but the 

 necessity of repeatedly taking samples of the distillate 

 and estimating the ammonia in them introduces an 

 amount of experimental error which becomes serious when 

 calculated out in milligrams per litre, though its actual 

 amount on each cylinder of distillate may be very small. 



When it is added that Mr. E. T. Chapman has included 

 in the second edition of Wanklyn's " Water Analysis " 

 a process for the estimation of volatile organic matter, 

 founded on the fact that water largely contaminated with 

 sewage evolves volatile bases when boiled with potassic 

 hydrate, the question is still further complicated ; for it 

 cannot be doubted that some portion of these bases would 

 be driven off by the action of the alkaline permanganate 

 before it had time to act on and destroy them. J Whether 

 this is the case ought at once to be determined by those 

 who use the process. 



In a second article we shall consider Frankland and 

 Armstrong's process for the analysis of potable water, and 

 also those determinations, such as nitrous and nitric acids 

 and chlorine, which are of great value as enabling iis to 

 trace back the history of a water, and to tell from whence 

 it is derived and what it has received in the way of animal 

 contamination before it came into our hands. 



* "Water Analj'sis," 2nd edition, p. 66. 



t If Licberliiihns formula for albumin H NaC?^ H„„N,3 S O,^, H„Obe 

 taken as the true one, lo parts of ammonia (N H3) from 100 parts of albumin 

 will be equal to 8*23 parts, or little more than half the total nitrogen, which 

 is for that formula i5'25 per cent. If the sodium in the above formula be re- 

 placed by hydrogen, the discrepancy is still greater. 



X An action of this nature apparently occurs when sewage is treated with 

 alkaline permanganate, as nearly the whole of ti e ammonia comes ofl" in the 

 first TOO cub. cent, of the distillate, and after that the evolution soon stops. 

 Waters that evolve small quantities of albuminoid ammonia seem almos 

 always to require a long time for its liberation. Is this because the org.'mic 

 bases are so diluted that they cannot be driven out, and so slowly decompose 



