OCTOBER 1906. Vol. XXXI. No. 367. THE ANALYST. PROCEEDINGS OF THE SOCIETY OF PUBLIC ANALYSTS. THE METHOD OF ANALYSIS OF MILK USED IN THE GOVERNMENT LABORATORY FOR SAMPLES REFERRED UNDER THE SALE OF FOOD AND DRUGS ACTS. BY H. DROOP RICHMOND F.I.C. and E. H. MILLER. (Read at the Meeting July 14 1906.) INTRODUCTION. IT is remarkable that so few investigations into the validity of this method have been made and published. It appears to have been used for more than thirty years, and was first described by Dr. James Bell the then Principal of the Somerset House Laboratory in Part 11. of his ‘‘ Analysis and Adulteration of Foods ” (see ANALYST, viii. 141). The process may be considered as consisting of two parts the maceration method of estimating fat and solids-not-fat and a correction for loss of solids by fermentative change.The maceration method has during the thirty years of w e undergone no essential modification and except in minor points of detail is used to-day as described by Bell. The correction to be made has however undergone a, radical change. As described by Bell the correction took the form of a time al1owance and this was severely criticised notably by Stokes (ANALYST xii. 226) and Allen (ANALYST xii. 231 I‘ Commercial Organic Analysis,” p. 203) and undoubtedly this correction was liable to grave errors. I t has now been superseded by a process oE correcting for loss by the estimation of some of the more important products of decomposition (alcohol volatile acids and ammonia) ; the method was published in a report to the Local Government Board (see Report of the Milk Standards Committee Blue-Book Cd.484 1901 p. 404) and again by Thorpe (Transactions of the Chemical Society 1905 p. 206). The latter paper contains in addition to experimental proof that alcohol volatile acids (acetic and butyric) and ammonia are actually formed in sour milk a further correction for butyric acid should the amount of this be appreciable. The only investigation made into the maceration method outside the Government Laboratory of which we are aware was carried out by Bevan and Hehner in conjunc-tion with one of us and the results are summarized in the Blue-Book cited above (pp. 391 and 392); the directions followed were those of Bell which are not very explicit and the method was not carried out in precisely the manner employed at the Government Laboratory.It appeared to us desirable considering the importance o 318 THE ANALYST, the method in the administration of the Sale of Food and Drugs Acts to investigate its accuracy and with the concurrence and partly even at the request of Dr. Thorpe, we commenced the present research. By the kindness of Dr. Thorpe one of us was enabled to receive personal instruction in the method from Mr. C. Simmonds and we believe that our procedure differs in no essential particular from that employed in the Government Laboratory We took up the investigation with more interest as we were aware that the method had been subjected to much criticism based for the most part on no direct experimental data and it appeared desirable to see how far the criticisms were valid.The research divided itself into two main heads : 1. The maceration method itself which again could be subdivided into-( a ) The estimation of fat ; and (b) The estimation of solids-not-fat ; and 2. The allowances for decomposition. This portion of the investigation may be considered as-(a) The alcohol correction ; (b) The volatile acid correction; (c) The ammonia correction including other corrections for proteid (d) Corrections not dealt with in the Government Laboratory. Our general conclusions are that although we have found several sources of error both in the method and in the corrections they are comparatively trivial and hardly affect the validity of the process. On the other hand we have been enabled by the analysis of the decomposed milk to make a satisfactory estimate of the composition of the original milk laore especially if the modifications of the method and the corrections we propose are employed.At the same time we must confess that the method requires more personal attention than could easily be devoted to it by public analysts and for this reason we conceive that adventitious error is more likely to occur. I. THE MACEEATION METHOD. Estimation of Fat in Fwsh MiZk.-It has been shown by Thorpe (Zoc. cit.) and by Hehner Bevan and Richmond (Zoc. cit. and 4 6 Dairy Chemistry," p. 94) that the results of estimation of fat by the maceration method agree with those obtained by other methods which are admitted to give substantially accurate results and we have therefore devoted but little time to this.We have however incidentally obtained a few results which confirm the accuracy and have been favoured by Mr. Simmonds with a few others obtained on samples of homogenized milk. change; and The results are : NO. Pat by Maceration Method. Fat by other Methods. 1. 3.92 3*93* 3-94" 3.95 2. 344,* 3.42* 3-46 3. 3-81," 3*84* 3-87 4. 3.88 3.91 3-92 5. 3-90 3-83 3-85 6. 3.86 3-86 3.9 THE ANALYST. 319 Samples 1 2 and 3 were homogenized milk and the results marked * axe the work of Mr. Simmonds. The only remark that it is necessary to make is that when the milk before extraction of the fat is neutralized using phenolphthalein as indicator the ethereal solution contains the whole of the phenolphthalein which is weighed with the fat, and from which it can be separated if desired by taking advantage of its insolubility in petroleum ether.As however the amount of phenolphthalein added need not exceed a milligramme no appreciable error is caused by neglecting this. Estimation of Fat in Sowr MiZ?c.-In the Blue-Book cited a number of comparisons of fat in fresh milk and in the milk after it has been allowed to become sour are given, and it is there shown that the fat in the sour milk varies from 0.06 per cent. more to 0.15 per cent. less than in the fresh milk and averages 0.05 per cent. less. Thorpe (Zoc cit.) reproduces most of the results and adds a few more which show substantially the same agreement. We have examined nineteen samples of milk by this method and though our limits are slightly wider than those given above find that the results on eighteen of the samples of sour milk are in very fair agreement with those obtained when fresh by Gottlieb’s method.I n one sample however, which is queried in the table of pesults the “ fat ” in the sour milk is more than 0-5 per cent high; in this case the milk had undergone a most unusual decomposi-tion and we were able to prove that the ether extract was by no means all fat. The ethereal solution was quite opalescent and when the ether was evaporated and the residue taken up with petroleum ether a considerable residue was left (the weight of this was added to the solids-not-fat) and the petroleum ether was also opalescent. After washing the petroleum ether solution with slightly alkaline water it became clear but on evaporating and drying the ‘( fat ” was turbid and on again taking up with petroleum ether a flocculent residue was left ; a second flocculent residue was obtained on evaporating the clear solution.Owing to an accident the fat after these treatments could not be weighed but there is no doubt that the figure given (641 per cent.) is too high. We shall refer to this sample again; and the results obtained on it indicate that amino acids resulting from the hydrolysis of the proteids were con-tained in the ethereal solution. We are of opinion that some of the difference between the results obtained on the fresh milk and those on the sour milk is due to the diEculty of completely redis-tributing the fat in the curdled milk. Examination of an apparently well-mixed sample of sour milk with a low-power lens shows the presence of quite large particles of cream and no amount of whisking with a wire brush appears to reduce the milk to the same homogeneous condition easily obtained with fresh milk.We have had one sample from which the label had been apparently soaked off in warm water and another substituted in which it was impossible to obtain concordant results for fat. Eight determinations by the maceration Gottlieb and Werner-Schmid methods, gave the following figures 4.18 4.05 4-03 4-51 3.67 3.73 4.76 2.78. Many samples also on mixing yield a portion of their fat in a churned condition, which adheres to the wire brush ; we have in these cases made a separate estima-tion of the churned fat: We have been quite unable to extract the fat with eight treatments with ether 320 THE ANALYST.as prescribed in the description of the method and have found twenty to thirty treatments necessary. We have also been unsuccessful in obtaining a residue from which we can obtain no more fat. We have usually stopped extracting when six to eight successive treatments yield only a mgm. or two of fat. This may be due to the fact that we have worked in aluminium basins instead of the platinum vessels used at the Government Laboratory and that this metal is too soft to permit of very efficient grinding. Estimation of Solids-not-fat in Fresh Mdk.-We have made a number of com-parisons of the solids-not-fat by the maceration method with those obtained by the Saciety of Public Analysts’ method and find that invariably the former are higher than the latter.Twelve samples gave the following results : Maceration Method. 9.14 9.58 9.68 8.37 9.33 9.11 S. P. A. Method. 8.97 9.34 9.40 8.24 9.07 S.8g Maceration Nethod. 8.54 7.96 9.28 9-04 8.91 9.05 S. P. A. Method. 8.28 7-85 9.10 8-74 8.83 8-90 The average difference was 0.20 per cent. One of us has suggested (Blue-Book Cd. 484 p. 274) that the difference was due to the presence of milk-sugar in the hydrated form but this suggestion was combated by Thorpe (idem.) and denied by Lewin (Zoc. cit. p. 350). We have therefore made a number of experiments on this point, (a) Weighed quantities of milk-sugar were dissolved in water and evaporated to a firm paste; on treatment with ether this became at once solid and was finely powdered.After drying to constant weight the quantity left corresponded neither to the weight of hydrated sugar nor of anhydrous sugar but to an intermediate weight. Hydrated Sugar taken. 1. 0.5120 2. 0.6630 Residue Calculated for obtained. Anhydrous Sugar. 0.4955 0.4884 0.6470 0,6298 ( b ) We repeated Schmoger’s experiment (cf. Veitch ANALYST xi. 141). The solids-not-fat were estimated in 10 grams (circa) of milk by the maceration method ; in another portion of the same milk of almost identical weight a weighed quantity of milk-sugar was dissolved and the solids-not-fat estimated in this ; the difference between the two weights corresponded neither to the milk-sugar added nor to the anhydrous sugar calculated therefrom.Hydrated Sugar added. 1. 0*5000 2. 0.4245 3. 0.4349 Residue obtained. 0.4775 0.4232 0.4301 Calculated for Anhydrous Sugar. 0.4750 0.4033 0.4132 A similar experiment in which the total solids by the Society of Public Analysts’ method in a sample of milk was subtracted from the total solids of the same quantit THE ANALYST. 321 of milk to which a known weight of milk-sugar was added yielded an amount that differed from the weight calculated for anhydrous sugar by less than a mgm. (c) As we have found that when hydrated milk-sugar is stirred with water (1 gram to 10 c.c.) it produces a fall of temperature of 0.55" while anhydrous milk-sugar gives a rise of temperature of at least equal extent we endeavoured to use this as a means of investigating the question.With ordinary milks the change of tern-perature was slight but the solids-not-fat of milk obtained by repeatedly extracting and finely grinding the total solids obtained by the Society of Public Analysts' method with ether tended to give about 0.1" higher temperature than the solids-not-fat obtained by the maceration method. When the temperature change was exaggerated by using the solids of the milk to which milk-sugar had been added there was a marked difference. Samples 2 and 3 of solids-not-fat of milk to which milk-sugar had been added obtained by the maceration method caused a fall of temperature of 0*3" while the solids-not-fat obtained by extracting the solids by the Society of Public Analysts' method of the milk to which milk-sugar had been added gave a rise of 0.3".We think therefore that there is no doubt that the solids-not-fat obtained by the maceration method contain a portion of the sugar as hydrated sugar but the amount of water of hydration which we estimate averages about 0.1 per cent. is not sufficient to explain the difference between the results. Another source of error in the maceration method is due to the presence of aldehydes in the ether. It is well known that ether contains aldehyde and that when this is removed the ether after a short interval again contains aldehyde. We have found that ether free from aldehyde will give a marked Schiff reaction if shaken for a few minutes in daylight. We also find that milk solids remove the aldehyde completely from ether and this appears to be due to a condensation of the - COH group with the free amino groups of the proteids.The solids-not-fat obtained by the maceration method are always more acid than the milk and the aldehyde figure is less the increase of acid and the decrease in the aldehyde figure being within the limits of experimental error identical. These figures afford data for the estimation of the increase of weight due to the condensation of the aldehyde and assuming that it is acetaldehyde the error is almost constantly 0.03 per cent. unless freshly-distilled aldehyde-free ether be used. The solids-not-fat by the maceration method also give a faint Liebermann reaction pointing to the condensation of glyoxylic acid, but this cannot be large as if it were the increase of acid would be larger than the decrease of the aldehyde figure.Even this addition does not explain the whole of the difference between the maceration and Society of Public Analysts' methods and the marked browning of the residue in the latter method suggests that the remainder of the difference is due to the results obtained by it being too low ; this conclusion is strengthened by the fact that evaporation over a large surface whereby browning of the residue is avoided, gives slightly higher results. In sour milks the error of the maceration method due to water of hydration will be reduced as the amount of milk-sugar is less than in fresh milk ; but on the other hand a portion of the milk-sugar may be converted into hexoses which will have the same effect on the weight as hydration 322 THE ANALYST.The exact details of the method we have employed are these About 10 grams of milk are weighed into a basin of aluminium about 3 inches in diameter and a little over 1 inch high with a flat bottom provided with a flat-ended glass stirrer. Two drops of a 0.5 per cent. solution of phenolphthalein are added and approxi-mately TG strontia solution run in till a faint pink colour appears. The contents are evaporated to a damp paste on the water-bath when the basin is transferred to a hot plate and the paste mixed with the stirrer ; at a certain point in the evaporation the paste comes away from the basin and by careful manipulation both basin and stirrer can be obtained practically clean.On further evaporation and stirring the paste begins to get into a condition in which it can be broken up and rubbed into pieces, and at this stage it is removed from the hot plate and about 20c.c. of ether are added On gentle rubbing with the stirrer the solids begin to go to pieces ; the stirrer and basin are now scraped with a knife or spatula to bring any small portions of solids adhering to the sides under the ether and the solids are gently rubbed to a powder. The ether is decanted through a weighed filter 9 em. in diameter and the solids again treated with ether. The solids at this stage are in a condition in which they can be ground up to a fine powder ; the ethereal solution is allowed to settle and the ether decanted through the filter ; without any further addition of ether the solids are now ground to a very fine powder.We prefer to do this with only a very little ether in the basin as it is then easy to see the larger portions which can be ground up one by one. A further addition of ether is made and the solids further ground ; the ether at this stage looks like whitewash and the solids take some minutes to subside sufficiently to allow of decantation. After about six or eight treatments in this manner the solids are allowed to air-dry the portions clinging to the stirrer and sides of the basin scraped down and 5 C.C. of alcohol and a few drops of water are added ; the solids are well mixed with the alcohol and the basin is placed on the hot plate and evaporated till the paste begins to go to pieces when the solids are again treated as before.A second treatment with alcohol and a further six to eight extractions with ether are given ; the filter gradually becomes partially blocked with the finely-divided solids but never to such an extent that filtration stops. The solids are air-dried and then dried in the water-oven to constant weight. We have not usually weighed the solids-not-fat in a weighing-bottle ; although they are hygroscopic we have satisfied ourselves that no appreciable error is due to this cause. The final weight is known from previous weighings to within a very small amount and consequently the time of weighing is very short and we have proved that not more than a few tenths of a iogm. of hygroscopic moisture are taken up during the weighing. The top of the filter where fat collects is cut off and cut into pieces and thoroughly washed with ether; the knife or spatula is wiped on some of the pieces cut off to remove any particles of solids and the filter containing the pieces cut off is dried in the water-oven to constant weight.The ether is distilled and the residue of fat weighed; the fat is extracted with petroleum ether and the small residue insoluble in this solvent subtracted from the weight ; this usually consists of phenolphthalein and its weight may be neglected without appreciable error, The ether employed by us has been methylafed ether (specific gravity 0.720) dried with calcium chloride THE ANALYST. 323 11. THE ALLOWANCES FOR DECOMPOSITION. The AZcohoZ Correction.-We have usually taken about 75 grams for the estima-tion of alcohol.The quantity is weighed into a 300-C.C. flask and half neutralized with soda solution ; about half the volume is distilled and collected in a small flask, and neutralized with soda using litmus-paper as indicator. From this 25 C.C. is distilled and the density taken at 60" F. by a Sprengel tube. From the density the percentage of alcohol is calculated by the formula given in the Blue-Book and from this the correction is calculated. We believe that the alcohol correction is very near the truth. The Volatile Acid Correction.-We have in every sample estimated the volatile acids by the method given in the Blue Book but do not think that the results by this method at all represent the amount of volatile acid except when the quantity is very small.The method is in our opinion subject to several errors and is liable to be quite 50 per cent. from the truth. I n the first place carbonic acid is always present, and is driven off on evaporating the half-neutralized milk to dryness and as this is partially estimated using phenolphthalein as indicator as acetic acid an error is thereby produced. The error is a double one as not only is a correction made for carbon dioxide lost during the production of acetic acid but the carbonic acid itaelf is retained by the strontia and is weighed with the solids. The sum of these two errors we estimate as about 0.05 per cent. Against this we find that the whole of the volatile acids are not driven off on evaporation to dryness and the error due to this cause will partly compensate the error due to carbonic acid.There is another source of error due to a possible volatility of lactic acid and to the conversion of a portion of this acid into a lactone. Against this there is the formation of amino acids by hydrolysis of the proteids which would have a feeble acidity to phenol-phthalein. We aIways find that fresh milk on half neutralization and evaporation gives a very slight increase of acidity; and in one sample of sour milk we found that there was an increase of acidity on evaporating the half-neutralized milk and this result was confirmed on repetition. Probably when the volatile acidity lies between 0.1 and 0.2 as in the majority of our samples the method gives approximately correct results; in other cases it is quite erroneoug.The following experiments will show this : 0.0162 gram acetic acid added to milk 0.0162 gram acetic acid + 0.07 gram lactic acid 0.0125 gram acetic acid + O.Oi14 gram lact'i" acid added to milk+ the solution saturated with carbonic acid (contained 0.0031 gram CO,) , 0,0192 ,, I n the presence of much volatile acid we have added to the milk from which the alcohol had been distilled a quantity of acid exactly equal to the soda used for half neutralizing and have distilled this down to a small bulk and then added water in successive quantities of about 25 c.c. and distilled this off till the distillate was practically neutral. Thorpe (Zoc. cit.) has given practically this method but apparently _ found 0.0084 gram. 9 2 9 7 1 7 7 J 9 9 ., 0.0092 ,, added to milk . . , 0.0165 ,, $ 9 . . . . . 324 THE ANALYST. has not added any acid ; as he directs that the volatile acids may be estimated in the distillate by any convenient method we employed that of Duclaux which appears to us not only to be capable of considerable accuracy but which enables us to estimate the relative proportions of even three acids in a mixture and to indicate if the modification we have devised be employed what the acids present are. We have found it essential to add sufficient acid to neutralize the soda added as in some cases the volatile acid exceeds half the total acidity. Our mode of working is this The mixed distillates neutralized to phenol-phthalein are made up to 250 c.c. and an aliquot portion taken for fractional distillation; to this is added a quantity of N sulphuric acid slightly greater than is necessary to neutralize the soda used.Eight successive fractions each about one-ninth of the total volume are distilled and separately titrated with & strontia; 25 C.C. of water is added to the residue and a further 25 C.C. distilled and titrated, and if desired this treatment may be repeated. From the last titration the total quantity of volatile acid is obtained by extrapolation of the results ; the total quantity is always slightly less than the total acidity of the first distillate due no doubt to the presence of a little lactic acid As much as possible of the first distillate is made up to a convenient bulk after adding a quantity of N sulphuric acid in slight excess of that required to neutralize the soda and exactly one-third is distilled ; this is made up to a convenient bulk and exactly one-third is again distilled.The last distillate is made up to 100 c.c. and eight fractions of about 11 C.C. each are distilled and separately titrated. The residue in the flask and condenser is washed out and titrated to obtain the total quantity. For the calculation of the results we have slightly modified the formulae given by one of us (ANALYST xx. 193) to : y = x2.1 for butyric acid ; y = ~ 1 ' 1 s for propionic acid ; and y = xj for acetic acid. y = quantity of acid left in retort ; total quantity = 1. x = volume of liquid left in retort ; total quantity = 1. For each fraction me calculate the proportion of each of the acids which would distil and calculate the ratios of butyric to acetic and butyric to propionic acid, which correspond to the proportion actually distilled The ratios from first and last fractions are liable to be slightly erroneous the first because a small amount of a very volatile acid as carbonic would appear in this fraction and the last because experimental error is greatly magnified ; but we usually find that the other six ratios are practically constant for either butyric and acetic acids or butyric and propionic acids.The distillation of the third of a third is undertaken in order to decide definitely the composition of the mixture. From the formuh above, the proportions which would distil when one-third of one-third is collected are : For butyric acid .. . . 0.3285 For propionic acid . . . 0.1445 For acetic acid . . . 0.06 THE ANALYST. 325 Distilled. I Distilled. ~ ~ - -Thus a mixture of butyric and propionic acids in equal proportions should yield approximately a 2 1 mixture in the distillate while a mixture of butyric and acetic in equal proportions would yield a 5 1 mixture. I t is even possible to deduce the relative proportions of a mixture of the three acids. I n the first we have concluded fhat the acids present were acetic and butyric ; in the second propionic and butyric; and in the third a mixture of the three. We give below the results of three experiments. -No. AC ACETIC AND BUTYRIC ACIDS. 0.48 1 ~ I Distillation of Original Distillate. ! ! 0.891 Distillation of 8 of &.-__-~ __ 1 Calculated for-0.206 --0.1215 I ' Calculated for- ~ Ratio. / I Ratio. ~ Volume 1 Distilled. _ - Acid Distilled. -0.190 0.343 0.490 0.619 0.726 o a 9 0.892 0-950 Butyric 1' Total. 1 I I -c__-0.108 0.141 ' 0.213 0.221 0-2275 I 0.406 0.335 I 0.403 1 0.5755 0.440 ' 0.515 ~ 0.708 0.559 ~ 0.6175 ~ 0.821 0-669 0.711 0.902 0.779 0.793 0.958 0.887 ~ ' 0.8705 0.9915 Butyric. 0.2225 0.397 0-566 0.701 0.8105 0.898 0.956 0.9905 -Acetic. Butyric Total. 0.78 0.78 0.77 0.785 0.79 0.795 0.805 0.805 -Acetic. -0.073 0.153 0.238 0.324 0.421 0.521 0.6345 0.780 0.077 0.148 0.2325 0,318 0,410 0.512 0.628 0.782 The quantity distilled in Q of & was 0.193.The figure calculated for a mixture of 0.495 molecules of butyric to 0.505 mole-The ratio of butyric acid in the second distillate should be 0.83. cules of acetic was 0.192. NO. 2.-PROPIONIC AND BUTYRIC ACIDS. Uistillatwn of Original Distillate. Distillation of ;5 of 6. -. _- - . Volume Distilled. -0.110 0.223 0-333 0.448 0.554 0.664 0.775 0.889 -~~ -Acid Distilled. Calcalated for- Calcdated for- Ratio. Butyric Total. 0-56 0.53 0.505 0.53 0.52 0-51 0.49 0.48 --Ratio. Butyric Total. Acid Distilled I Volume Distilled. 7-0.104 0.215 0.331 0.444 0.558 0*656 0.767 0.875 Butyric. Propionic, 0.1685 0.328 0.478 0.609 0.723 0.807 0.886 0,949 0.183 0-357 0.509 0.641 0.750 0.839 0.911 0.964 0.217 0.411 0.573 0.713 0-8165 0.899 0.9565 0.990 0.128 0.258 0.3 80 0.504 0.614 0.724 0.828 1 0.925 0.62 0.65 0.67 0.66 0.67 0.66 0.65 0.6 326 THE ANALYST+ Propionic.The quantity distilled in Q of Q was 0.242. The figure calculated for a mixture of 0.514 molecules of butyric to 0.486 mole-The ratio of butyric acid in the second distillate should be 0.70. cules of propionic was 0.239. Acetic. NO. &-ACETIC PPROPIONIC AND BUTYRIC ACIDS. Distillation of Original Distillate. I I Calculated for- Ratio. Volume Distilled. I Acid 1 Distilled. I--0.169 0.324 0.467 0.588 0.705 0.8105 0.897 0.950 Butyric B + Ac. ___~ Butj-ric + P-0.545 0.51 0.50 0.49 0.48 0.4 6 0.44 0.33 Butyric B + 2P TA.Rutyric. _____ 0.208 0.397 0-563 0-695 0.8105 0.89 7 0.955 0.99 1 Proprionic. ~ 0,1225 0.2475 0.372 0.487 0.607 0.721 0.825 0.930 Acetic. 0.071 0.148 0.231 0-314 0.410 0-514 0-627 0,7775 0.715 0.71 0.71 0.72 0.735 0.75 0.78 0.81 0.62 0.60 0.60 0.60 0.62 0.61 0.63 0.63 0.105 0.214 0.326 0.432 0.547 0.661 0.772 0.895 I t is seen that bhe ratios calculated on the supposition that the acids consist of acetic and butyric and propionic and butyric neither of them give constant figures, but when calculated for a mixture of 2 molecules propionic and 1 acetic the agree-ment is satisfactory. On this assumption the total quantity distilled in & of + should be 0.248 of the total.We actually found 0.251 and the ratio of propionic to acetic in the last distillate should be 5 to 1. Distillation of + of Q. __-Acid Distilled 0.190 0.360 0.521 0.652 0.761 0.850 0.921 0-969 Calculated for-~ Ratio. Volume Distilled. Hutyric K + Ac. Butyric n-t r. Bn tyric B+5P+A. Butyric. 0.210 0.4015 0 573 0.707 0.8165 09025 0.959 0.991 0.124 0.2505 0.380 0 *501 0.614 0-730 0.8335 0.929 0.106 0,217 0.333 0-445 0.554 0 667 0.781 0.894 0.85 0.835 0.84 0.83 0.86 0.86 0.88 0.90 0.77 0.725 0.73 0.73 0.725 0.695 0.70 0.64 0.072 0.1505 0.247 0.325 0-416 0.5225 0.637 0.776 0.79 0.75 0-76 0.765 0.765 0-76 0.76 0.74 The ratio of butyric acid in the second distillate should be 0.78.I t is noticed in every case that the ratio of butyric acid in the distillate of 8 of THE ANALYST. 327 is slightly less than that calculated and this is partly due to the fact that there is always a slight loss during the two distillations and as butyric acid is the most volatile of the compounds present the loss will be proportionately larger of this acid than of the others, We think that we may claim that the above figures show that a fairly accurate estimation of and a determination of the nature of the acids present is possible by Duclaux’s method as modified. We append a table showing the total volatile acids expressed as C.C. of N alkali per 100 grams of milk estimated by the method described in the Blue Book and Duclaux’s methods respectively in five samples which have been high in volatile acids : NO.7. 9. 11. 13. 15. Blue Book Method. 7-35 5.7 7.68 14.0 12.35 Duclaux’s Method. 12.6 12.2 21.6 20.7 9.75 Ratio. 1.71 1.71 1.59 1-54 1-68 As to the corrections to be made for volatile acids there is some doubt. When the amount has been small we have assumed them to be acetic acid alone and have applied the correction given in the Blue Book-Le. an addition of 25.5 parts for each 60 parts of acetic acid. This correction we prefer to that given by Thorpe-Le., 62 parts for each 60 parts of acetic acid-as the latter correction assumes that acetic acid was produced by oxidation and the oxygen was part of the original milk.For butyric acid Thorpe gives a correction of 92 parts for each 88 parts of butyric acid but this correction assumes that the weight of lactic acid from which it was supposedly formed was equal to the weight of the milk-sugar giving rise to it. This is the case if the milk-sugar is supposed to exist in the solids as hydrated sugar but we have shown that this is only partially the case and prefer to calculate the loss to anhydrous sugar and modify the correction to 87.5 parts for each 88 parts of butyric acid. We are however by no means satisfied that this correction is justified; it assumes that the butyric acid is derived from lactic acid. It is well known that the organisms producing lactic acid do not form more than about 1 per cent.of lactic acid as their action is inhibited by this amount of acid. In some of our samples the total acidity has been equal to nearly 3 per cent. of lactic acid and as the greater portion of this has been butyric acid it follows that it cannot have been derived from lactic acid. We assume that in every case of fermentation the milk-sugar is hydrolysed to glucose and galactose and we therefore take a hexose as our starting-point. To explain the formation of butyric and propionic acids we assume the following reaction to take place : 3C,H,,O = 2CH,CH2CH2COOH + 2CH,CH2COOH + 4C0 + H2 + 20H2 328 THE ANALYST, which gives a total loss of 94.5 for 88 parts of butyric and 74 parts of propionic acids. This is the proportion in which we have found these acids and we have therefore used this correction.For butyric and acetic acids there are so many equations which can be drawn out to explain their formation each of which gives a different correction that we have not attempted to do this but have used the corrections given above. That they are not quite correct is shown by the agreement of these samples not being quite 80 good as the others. The Ammonia Correction.-We are very doubtful as to the validity of this correction. Though during the hydrolysis of the proteid a portion of the carbon and hydrogen may be converted into volatile non-acid compounds we do not think that the whole of the nitrogen is converted into ammonia with total loss of the other constituents. We are also doubtful whether the colour produced by Nessler solution is all due to ammonia as it is sometimes so markedly different in tint from that produced by the standard solution of ammonium chloride.As the whole correction is usually small we have accepted it as we have found that there are changes in the proteids on evaporation of the milk which no doubt involve a slight loss. We have in many samples estimated the aldehyde figure (cf. Steinegger, ANALYST xxxi. 46 and ourselves ANALYST xxxi. 224) both in the sour milk and in the solution of the solids obtained by evaporating the half-neutralized milk. The table below gives our results : NO. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. These Aldehyde Figure in Fresh Milk. 19.4' 19.4' 19-4" 19.4' 19.4' 19.5' 19.4" 19.4' 19.4" 19.6' 19.4' 14.4' 19.4' 14.5' Aldehyde Figure in Sour Milk.24.0' 24*1° 24.1' 21.3' 27.6" 29.8" 87.5" 35.6" 32.0" 22.7' 26.7' 23.1' 32-9O 27.1' Aldehyde Figure after Evaporation. 17.5' 15.5" 16.9' 15.1' 16.3' 13 -9' 19.4' 14.9' 16.3" 16-7" 15.7' 10.9O 18.2" 11.8' FJ H,, Correction. 0.01 0.02 --0.02 0.06 0.01 0.09 0.07 0.01 0.02 0.03 0.12 0.10 Eigures show that in all the samples of sour milk hydrolysis has taken place with the formation of amino acids but that on evaporating these amino acids condense either to form di-keto-piperazine derivatives or polypeptids ; the hydrolysis would cause an increase in weight but the condensation a loss and as the figure is always lower after evaporation than in the fresh milk this indicates that a slight additive correction should be made.I t is noticed that when there is much ammonia developed there is also a high aldehyde figure THE ANALYST. 329 We have examined one sample (No. 5) which is not included in the above list, which gave an aldehyde figure after evaporation of over 300O; on evaporation it smelt like glue and became a plastic mass. This is the sample referred to as yielding too high an ether extract. The distillate which should have contained only alcohol was found to be strongly alkaline with ammonia or an amine and the acidity was higher than it could have been had all the milk-sugar originally present been converted into lactic acid.The solids-not-fat were higher in the sour milk than in the fresh milk, and the system of corrections adopted at the Government Laboratory entirely failed to give correct results. This sample had undergone a most unusual form of decom-position but as it was not originally a normal milk we do not consider that it invalidates the system. As the unusual decomposition which was not realized till the sample had been used up prevented an accurate analysis being made we cannot speak very definitely but there are distinct indications that the further corrections which we propose would have given good results even with this sample. Corrections not dealt with in the Government Laboratory.-These corrections are three in number; the first is for the lactic acid produced. The usual equation for this is-which indicates that 342 parts of milk-sugar produce 360 parts of lactic acid.We have therefore subtracted 5 per cent. of the lactic acid from the weight of the solids-not-fat. We calculate the total acidity less the volatile acidity plus 0-2 per cent. (to represent the acidity of the milk when fresh) as lactic acid. The second correction is for the aldehyde taken up from the ether used. We find that the solids-not-fat which should of course be neutral are always acid after extraction with ether and the aldehyde figure has gone down. That this is due to condensation of aldehyde is shown by the substantial agreement of the increase of acidity with the decrease in the aldehyde figure. As the aldehyde figure is affected by the combined error of four titrations the agreement is good.C,,H,,O, + OH = 4CH,CHOHCOOH, No. 6. 8. 10. 12. 14. 16. 17. 18. 19. Increase of Acidity. 7 * 4 O 10.5' '7.6" 8-2" 4.3" 9.9" 8.2" 5.6" 5.6" 5.8" Decrease of Aldehyde Figure. 7*4O 9*1° 9.8" 6-7" 5.1" 10.0" 8.0" 5.5" 4.7" 5.5" All figures are expressed as C.C. of TG alkali per 100 grams. Assuming that the following equation takes place : NH,H coo R \ + CH,CHO = R :{:!. cH3 + OH,, it is seen that each 1" will correspond to an increase of 0-0026 per cent. of weight, When the sour milk contains a large amount of butyric acid the increase o 330 THE ANALYST. acidity does not agree with the decrease of the aldehyde figure. I n these samples an odour of butyric acid is noticed on extracting and drying and the solids-not-fat are either neutral or alkaline.I n thiscase we calculate what the increase of acidity should be from the aldehyde figure and take the difference between this figure and the figure actually found to represent loss of butyric acid by multiplying each degree by 0.0088. I n cases where the aldehyde figure of the evaporated milk is higher than 20" we should make a correction to be subtracted from the solids-not-fat for the weight of water taken up on hydrolysis. The equation of this change is : R - GO - NH - R1+ OH = R - COOH + NHZR', and each degree of aldehyde figure above 20" would represent a gain of 0.0018 per cent. of water. The abnormal sample previously referred to is the only one we have examined which would be subjected to this correction.~ i s c u s s i o ~ of Results.-In the table on p. 331 we give the analytical figures for nineteen samples of milk which were examined when fresh by the Society of Public Analysts' method and when sour by the method used in the Government Laboratory. Taking the whole of the samples it is seen that the fat estimated by the macera-tion method in the sour milk varies from 0.10 above to 0.18 below the fat estimated by the Society of Public Analysts' method in the fresh milk and averages 0.03 per cent. less. The solids-not-fat of the whole of the samples vary when corrected by the methods used in the Government Laboratory from 0.28 above the solids-not-fat in the fresh milk determined by the Society of Public Analysts' method to 0.27 per cent.below ; while if the corrections that we propose are used in addition the variation is from 0.32 above to 0.20 below ; the differences average +0*08 per cent. and +0.06 per cent. respectively. For milks that have undergone the normal decomposition-b. in which chiefly lactic acid is produced-the figures are : Maximum. Minimum. Average. Government Laboratory . . . . 0.28 - 0.11 + 0.14 R. and M. . . . . 0.17 - 0.19 + 0.07 Even when there is a marked alcoholic fermentation or a strong butyric fermen-tation the agreement is still fair and does not exceed the limit laid down by Thorpe (0.3 per cent.) (Zoc. cit.) as being sufficiently satisfactory. We have included two samples which contained about 25 per cent. of added water and here the agreement is satisfactory.These samples were made up by mixing known weights of the milk immediately preceding them in the table and of water and the amount of water calculated from the analysis of the sour milk after correction corresponded to within a few tenths of 1 per cent with that actually added. As showing what different fermentations milk may undergo we may mention that Nos. 8 and 9 and Nos. 10 and 13 were each taken from the same bulk of milk, though at slightly different times yet the character of the decomposition varied widely. It is of interest to mention that in the two samples in which butyric fermenta-tion had taken place and which give results for solids-not-fat on correction rathe THE ANALYST. 331 I 332 THE ANALYST. higher than the original the ratio of butyric to acetic acid was 1 1 while in those which are low the ratio was 3 1.This may be a coincidence and it is evident that many more samples must be examined before the proper corrections can be established. We have also examined some solutions of milk-sugar to which a small quantity of sour milk was added to supply the organisms necessary for fermentative change. The milk-sugar solution used contained 4.77 per cent. total solids calculated from the amounts of milk-sugar and milk taken and calculating the sugar as anhydrous sugar and 4-74 per cent. by evaporation to dryness and drying to constant weight. The following are the analytical figures : Volatile Acid Total Acidity Correction. Correction. as Lactic Acid. Alcohol Age. Total Solids.12 days . 4.56 0.35 0.03 0.20 37 days . 4.84 0.20 0.03 0.20 70 days . 4.70 0-15 0.04 0.29 85 days . 4.17 0.78 0.09 0-51 It is evident that the samples had undergone but little change; the volatile acid, which was found to be almost wholly carbonic acid accounted for about half the total acidity. It is seen that even making a minus correction for the carbonic acid the results are still too high and these experiments afford further proof that in the maceration method a portion of the milk-sugar is weighed as hydrated sugar. In conclusion we think we may claim to have shown that-1. The maceration method for fresh milks gives results slightly too high for solids-not-fat and about to the same extent that the Society of Public Analysts' method is low. 2. By the method used in the Government laboratory a satisfactory determina-tion of the composition of the orginal milk can be made the results except in cases of high butyric fermentation and other abnormal decompositions not being more than 0-2 per cent.from the truth. 3. The method for the estimation of volatile acids used at the Government Laboratory is not a good one and requires modification if any appreciable amount of these acids be present. 4. Certain small additional corrections may be made with advantage, Considering that in the course of an examination of nineteen samples of milk we have found at least two forms of decomposition (a proteolytic and a butyric-propionic fermentation) which require special corrections it is evident that many more experiments should be made before the analysis of sour milk can be considered t o be on a satisfactory basis.We would emphasize the fact that our comparatively large proportion of samples showing the butyric fermentation is partly due to a selection of these samples in preference to others and partly to the special precautions taken to prevent the bottles from bursting. With the normal conditions under which reference samples are kept, we doubt whether more than one of -the five would have survived to reach the Government Laboratory THE ANALYST 333 Firially we would express our thanks to Dr. Thorpe and Mr. Simmonds of the Government Laboratory for the kind way in which they have assisted us in this investigation. DISCUSSION. The PRESIDENT (Mr. Bevan) in inviting discussion said that his own experience was that the fat determinations made in the Government Laboratory in samples of sour milk were generally lower than those obtained by the public analyst on the fresh milk.Mr. Richmond some seven or eight years ago had considered the maceration process to be unsatisfactory in the case of sour milk but had now apparently changed his view. I n ‘‘ Commercial Organic Analysis,” Mr. Allen had said in referring to the Government method ‘‘ Some points strongly invite criticism, but the method must be welcomed as an ingenious attempt to effect a satisfactory solution of a very difficult problem.” And it seemed now to be generally agreed that the method was an ingenious and fairly satisfactory one. MR. F. J. LLOYD said that he had done a good deal of work on the examination of sour milk and should be interested to see what light these improved methods would throw on the changes which took place in such samples.Those changes it might be mentioned were never constant. He had found that in genuine milk changes occurred which depended entirely on the nature of t6e bacteria present ; and when a sample was adulterated with water other changes were brought about by the organisms peculiar to water. E e had studied the action of various micro-organisms on sterilized milk and had found that while it was fairly easy to obtain accurate results with samples which had only been acted on by the lactic acid bacillus it was very difficult when gas-producing organisms (such as BaciZZz~s coZi cowmunis) were present in large quantities to obtain all the products of the decomposition and calculate the composition of the original milk.In the majority of cases he had found that the loss was proportional to the amount of acid produced being almost always half the total acidity calculated as lactic acid. The supposition that what took place in the conversion of milk-sugar into lactic acid was merely an addition of the elements of water to the milk-sugar and a splitting up into exactly four parts of acid was incorrect. As a matter of fact only about 60 per cent. of the milk-sugar was actually converted into lactic acid other compounds being formed from the remainder. Mr. FAIRLEY said that he could endorse what Mr. Lloyd had said as to the varying effects produced in milk by different bacteria.Possibly some of the abnormal results stated in the paper might be thus explained. Mr. RICHMOND said that it was quite true that his opinions had changed as to the extent to which the fat could be obtained from eour milk by the maceration process. During the past few years he had learned a good deal more about the process and had come to the conclusion that formerly he had not applied it in the best way. I t was iiiiportant to dry the milk to the right extent and to grind it thoroughly. He quite agreed with Mr. Lloyd as to the difficulty caused by gas-producing organisms. Nevertheless he thought that even in that case by determining a sufficient number They had found about a day and a half’s hard grinding necessary 334 THE ANALYST, of compounds the composition of the original milk might be fairly accurately deduced.From the figures they had given it would be seen that their experience was quite at variance with Rlr. Lloyd’s as to the total loss being equal to half the quantity of lactic acid. He thought that merely to correct for lactic acid would by no means be sufficient. With regard to the conversion of milk-sugar into lactic acid it was of course the case that even in stimples that had undergone what they had called normal decomposition small quantities of volatile acids and alcohol were always produced and therefore the very simple formula referred to would not strictly speaking represent what took place. Nevertheless for purposes of correction when each of the other substances was duly taken into account it did not involve a serious error to apply the correction in the simplest way by assuming that the residue of the lactic acid was derived from milk-sugar according to the formula.Dr. VOELCKER suggested that if the authors of this paper and also Mr. Lloyd, continued their observations it would be advantageous to pay attention more particularly not to normal samples which gave the public analyst no difficulty, but to samples which were mixed or believed to be mixed with smaller or larger quantities of water. Mr. ESTCOURT said that according to the experiment at the Government Laboratory the loss of solids-not-fat which occurred on keeping a milk was mainly due to the formation of alcohol. Indeed the alcohol formed was said to represent seven-tenths of the total loss by keeping.The figures given by the authors of the paper do not bear out this theory. How does Mr. Richmond account for this? Mr. RICHMOND said that the statement that the alcohol correction amounted to seven-tenths of the total correction was to be found in the Appendix to the Report of the Departmental Committee. They did not endorse it; it was not generally correct. Mr. LLOYD remarked that he had never found anything like 2 per cent. of acid, and only once over 18 per cent. even in samples that had been kept for months. The PRESIDENT said that in some samples which he had analysed some years ago he had found 13 per cent. of total acidity including GO,. Mr. RICHMOND said that in one of their samples as would be seen the total acidity was 22 per cent. and in two others it was even a little more. That of course was the total acidity from all sources calculated as lactic acid. The quantity of volatile acids in those three cases was considerable and there was very little actual lactic acid. If all the acidity were due to lactic acid he quite agreed with Mr. Lloyd that it would not be likely to reach 18 per cent. It was only when some unusual form of decomposition had occurred in the course of which large quantities of other acids were produced that these high acidities were found. NOTE BY DR. THoRPE.-Mr. Droop Richmond has courteously afforded me the opportunity of perusing the foregoing paper and of appending any comments that it might seem desirable to make. Anything like detailed criticism however would here be out of place; and beyond congratulating the authors upon having dealt in an instructive way with an interesting matter I think I need only remark that the principal matter on which Messrs. Richmond and Miller’s procedure differs fro THE ANALYST. 335 that of the Government Laboratory is the method of dealing with the volatile acids. In actual practice however we find the cases extremely rare in which the volatile acids exceed 0.2 per cent. so that the extended process described in the paper would scarcely ever be applied. On the whole it appears that although there are certain points open to discus-sion the conclusion of Messrs. Richmond and Miller is that in general a substantially accurate determination of the original solids of the milk can be made by the method in use at the Government Laboratory. That in certain cases a somewhat closer approximation can be obtained by elaborated processes is no doubt true and the authors are to be congratulated upon their careful investigation of the question