146 THE ANALYST. THE DETECTION OF ARTIFICIAL COLOURING MATTERS IN FRESH AND SOUR MILK. BY M. WYNTER BLYTH, B.A., B.Sc., F.I.C. (Read at the Meeting, March 5 , 1902.) IN view of the recommendations” of the Departmental Committee on Preservatives and Colouring Matters in Food, 1901, we may expect the addition of colouring matters to milk to be at some time or other an offtince under the Sale of Food m d Drugs Act. On this account I think a few observations and experiments recently made by myself may be of some value to analysts. The colouring matters which have been used from time to time in milk are numerou.s, and include annatto, turmeric, saffron, carotin, caramel and coal-tar dyes, but practically at the present time in this country the only colouring matters used to any extent are annatto and coal-tar dyes.That these latter we being used more and more both in milk and other articles of food in place of the natural colouring matters is a matter of common knowledge, and personally I found during the year 1901 no less than 19 per cent. of the retail samples of milk submitted to me for analysis under the Sale of Food and Drugs Act to be coloured with coal-tar dyes. The coal-tar dye used at present is, as far as my own experience goes, methyl orange (salts of dimethylamido-azo-benzene-sulphonic acid), or dyes very closely allied to this body. Should, however, the addition of colouring matters to milk be prohibited, we may expect the adulterator, us usual, to exercise his ingenuity in order if possible to defeat the analyst, and we shall certainly find many other dyes used.For instance, salts of the aniline base-amido-azo-benzene (aniline yellow), or its dimethyl derivative (butter yellow), salts of amido-azo-benzene-sulphonic acid (acid yellow), salts of diphenyl - amino- azo - benzene - sulphonic acid (orange IV.), * Recommendation B : *‘ That the use of any preservative or colouring matter whatever in milk offered for sale in the United Kingdom be constituted an offence under the Sale of Food and Drugs Act.”THE ANALYST. 147 all impart a more or less suitable colour to milk; the other sulphonated azo dyes are not in a pure state so suitable for colouring purposes. Salts of dinitro- alpha-naphthol (Martius yellow), salts of dinitro-ortho and para-cresol (victoria yellow), and certain basic dyes, such as phosphine, might also be used, especially the latter ; and if we take into consideration mixtures of various coloure, the number of possible dyes becomes very large.So, although the colouring matters used at present are limited in number, and for the most part easy of detection in fresh milk, I have, in view of probable con- tingencies, sought for gome general method which shall cover as far as possible these future developments. In this country the detection of colouring matters in milk has not received any special attention. H. Droop Richmond (“ Dairy Chemistry,” p. 139, 1899) recommends the precipitation of the colouring matter and casein with acetic acid, and subsequent extraction of the colouring matter with alcohol. In America A. E.Leach (Journ. Amer. Chem. SOC., 1900, 207) has studied the subject, and devised an excellent general method for the detection of a few colouring matters in fresh milk. The detection of artificial colouring matters in sour or decomposed milk has, as far as I know, not been previously studied. The usual test for coal-tar dyes in milk is the addition of hydrochloric acid to the milk itself or to the separated curd, when a pink colour is obtained in the presence of some coal-tar dyes, This test may fail for two reasons: (1) The milk may not be quite fresh ;* (2) a coal-tar dye may be present which does not strike a pink colour with hydrochloric acid. THE ACTION OF DECOMPOSING MILK ON COLOURING MATTERS. A. Coal-Tar CoZours.-If a small quantity of an easily-seen dye, such as methylene blue, be added to milk, and if the milk be incubated at body temperature for a few hours, or allowed to stand at ordinary temperature for some days, the blue colour will disappear, leaving the milk quite white.If the sample be now vigorously shaken with air the blue colour will slowly return. If, on the other hand, the milk be made strongly yellow with such a dye as methyl orange, and incubated in the same manner, the yellow colour disappears, and cannot be restored by shaking with air, nor can any pink colour be obtained on the addition of hydrochloric acid. This change takes place with small quantities of dyes very quickly and at very low temperature, Commercial samples of milk dyed with methyl orange may give a pink colour even on the fourth day in cold weather, but in summer they will, as a rule, give no reaction after standing twenty-four hours in the warm laboratory.The action of deconzposing milk on coal-tar dyes is exactly the same as i s the action of zinc dust a?zcZ hydyockloric acid. This may be proved by the fact that the same decomposition * To this reason, I think, must be attributed the fact that no coal-tar dye was found in any of the 296 samples examined in the Government Laboratory as recorded in the report of the Departmental Committee on Preservatives, etc., 1901. The Committee evidently recognised the difficulty of examining decomposing foods, for they state ‘‘ that the characteristic reactions for the colouring agents are frequently interfered with or obecured by the organic matter of the food itself, especially after decomposition has commenced.”148 THE ANALYST.or reduction products may be extracted from a dye incubated in milk as may be extracted from the dye when reduced by zinc and hydrochloric acid, and those dyes which are reduced to colourless bodies by zinc and hydrochloric acid, and which ape again restored to the original dye by oxidation, are also reduced to colourless bodies by incubation in milk and again restored to the original dye by shaking with air. For example, methyl orange breaks up under the action of reducing agents into dirnethyl-para-diamido-benzene and sulphanilic acid. Milk treated with methyl orange and incubated, if rendered alkaline and treated with ether, will yield up to the ether para-diamido-benzene.On evaporating the ether a rose-coloured residue will be observed, especially at the edges of the separated fat. This may be identified as para-diamido-benzene by shaking with 10 C.C. of warm distilled water, separating the fat, and adding to the aqueous solution (1) 2 drops of normal sulphuric acid, (2) 2 drops of saturated sulphuretted hydrogen solution, (3) 2 drops of very dilute ferric chloride solution, and warming, when methylene blue will be formed. By operating as above, as little as 0-00002 gramme of para-dimaido-benzene may be detected in 10 C.C. of water, if at the same time a, control-tube with the reagents and distilled water is treated in the same manner. This methylene blue reaction is easily obtained with the residue obtained from sour milk when the amount of methyl orange added in the first place is large ; but owing to the ease with which para-diamido-benzene oxidizes during the manipulations necessary to isolate, it, it is very difficult to obtain with the residue from commercial milks which contain between I part of colouring matter in from 100,000 to 300,000 parts of milk.The rose-coloured residue is, however, very characteristic, and may easily be observed, even without separating the fat, in the ether residue from 50 C.C. of commercial milk containing methyl orange. In the same manner orange IV. yields a residue of impure para-amido-diphenyl- amine, which with a drop of ferric chloride gives an intense blue-green colour changed to scarlet by strong sulphuric acid. Other dyes yield residues which are either coloured or become so on standing for a little time exposed to the air.A few examples of the nature of the decomposition and reduction products of different dyes are given in the table at the end of this paper, but the important point is, not the absolute recognition of the decomposition product of any dye, but the fact that milk which has never contained a coal-tar dye does not become coloured on shaking with air, and yields from the ether extract no coloured residue or any product which gives any colour reactions with such reagents as ferric chloride or strong mineral acids. B. Other Colouring Matters.-Annatto, turmeric, saffron, carotin, and caramel are not reduced to colourless bodies by decomposing milk (except perhaps after a very long time).THE GENEBAL EXAMINATION OF MILK FOR GOLOURING MATTER. The colour of the milk should be carefully noted, as some dyes, such as aniline yellow and acid yellow, impart a faint pink colour to the curd of sour milk; butter yellow, on the other hand, rises with the fat, which it at first colours yellow, but afterwards becomes colourless ; annatto imparts a characteristic colour to the curd, and, like saffron and turmeric, does not become colourless in decomposed milk;THE ANALYST. 149 caramel colours both the curd and the whey. If it be desired to make a comparison between the milk when fresh and after partial decomposition, two portions should be incubated, one containing sufficient formalin to prevent any decomposition. The following general method will be found most convenient : Take 50 C.C.or more of the milk (as much as can be spared if reduction products are specially looked for) and render it just alkaline to delicate litmus paper, evaporate to a paste, and thoroughly extract the fat with ether. Although turmeric, annatto, and such dyes as aniline yellow are all somewhat soluble in ether from an alkaline solution, yet in the presence of casein such small quantities are dissolved that they may be disregarded. (Phosphine is much more soluble in ether, and should be looked for both in the ether and alcohol extract.) Evaporate the ethereal solution to dryness, shake up the fat with a small quantity of hot distilled water in a small separating funnel, separate the water from the fat, and evaporate to dryness on the water-bath in a small flat porcelain dish ; carefully note the colour of the residue (see Table).Pure milk will give no coloured residue. Next, thoroughly exhaust the fat-free milk residue with absolute alcohol, pass the alcohol extract through a small filter, and evaporate to dryness in three or four small flat porcelain dishes ; if unreduced artificial colouring matter be present the residue will be coloured orange, yellow, or brown. Wash one of the residues into a test-tube with a small quantity of water made acid with sulphuric acid, the sulphonated azo dyes will be at once indicated by the colour of the solution; shake the solution with ether, this will divide the colouring matters into two groups, as in the table: the natural colouring matters and the non- sulphonated acid coal-tar dyes-i.e., the dyes precipitated by Weingartner’s tannin reagent-being soluble in ether, while the basic dyes and the sulphonated dyes are insoluble in ether. The ordinary tests may then be applied to the remaining portions of the dyes. For the sake of completeness I append to the table some confirmatory tests for the colouring matters most likely to be met with at present in commercial samples of milk : A. ALCOHOL EXTRACT, COLOURED ORANGE, YELLOW, OR BROWN. Take up a Portion of the Residue with Dilute Sulphuric Acid and shake with Ether. I THE ETHER DISSOLVES SOME OF THE COLOURINC, 1 MATTER. ~ MATTER. THE ETHER DOES NOT DISSOLVE THE COLOURING Natiiral Colouring Matter. Annatto Turmeric Saffron Carotin Non-sulphonated Acid 1 Basic Coal-tar Colours, 1 Sulphonated Coal-tar such as : Colours, such as : Phosphine ‘Acid yellow 1 Methyl orange 1 Orange IV.- I _- Coal-tar Colours, such as : Aniline yellow Butter yellow I Martius yellow 1 Victoria yellow -150 Colour of Residue. THE ANALYST. Probable Original Golouring Matter. B. WATER EXTRACT FROM THE FAT. Note the Colow and apply various Reagents to the Dry Residue. -- Brown Brown ~ - - Brown Yellow Yellow Rose red ____ Brown-red Yellow -- Acid yellow -- Butter yellow Aniline yellow Mar tius yellow “Victoria yellow Methyl orange Orange IV. -- -- Add a Drop t o Ferric Chloride. Dark green Dark blue- green Yellow -- Red -- Red Fugative scarlet Green -- To the Ferric Chloride add Strong H,SO@ Yellow, green on dilution -- Yellow, green on dilution -- Yellow -- Yellow Yellow -- Yellow ____- Scarlet, green on dilution -- Unreduced dyes, soluble in ether from alkaline solution Other Reactions.~ To a slightly acid solution of the colouring matter add a few drops of H S solution, then ferric chloriJe, heat-magenta colour. Strong H,SO,, or HC1, gives a delicate violet Strong HCI gives a beautiful rose Same as Martius yellow - -- ~. red colour - ------_- Treat as with butter yellow, the solution becomes a beautiful blue. Treat as butter yellow, the solution becomes dirty violet CONFIRMATORY TESTS FOR THE MORE COMMONLY USED COLOURING MATTERS. Awzatto.-(l) Take up the yellow alcohol residue with weak sodium hydroxide solution, dry a few drops on filter-paper. An orange spot changed to pink by stannous chloride solution indicates annatto.(2) Dissolve the yellow alcohol residue in a little water, add a little alcohol and a few drops of ammonia. Introduce a bundle of white cotton fibres and evaporate the liquid nearly to dryness on the water-bath ; wash the fibre, which will be yellow, and then immerse it in citric acid solution, when it will become pink. This excellent test is due to A. Leys, who, however, isolates his annatto in a different manner. (3) A convenient method for dealing with a large number of samples is that proposed by Leffmann and Beam (“Select Methods of Food Analysis,” p. 217, 1901). The milk is rendered alkaline by acid sodium carbonate, and a slip of filter-paper allowed to soak in it all night. Annatto causes a reddish-yellow stain on the paper.THE ANALYST.151 Turmeric.-The usual test with boric acid is sufficient. Caramel.-This is one of the most difficult of all colouring matters to detect in small quantities in milk. Whether it is used in this country is doubtful; it is, how- ever, used in America to some extent (A. E. Leach, Thirtieth and Thirty-first Annual Reports, State Board of Health of Massachusetts). The following modification of the test proposed by A. E. Leach (Jozmt. Amer. Chern. Soc., xx., 207) I have found to give the best results. I t is useless to use any evaporation procest3, as even with great care some caramel may be formed from the milk-sugar. Obtain a milk known to be free from caramel as a control. Take 50 C.C. of the suspected milk and the same quantity of the control, coagulate each by the addition of acetic acid, strain off the whey from the curd by meam of pieces of fine muslin ; carefully compare the colours of the whey from both samples.Place the curds in two white porcelain basins and just cover them with strong hydrochloric acid. Compare the colours after they have stood several hours ; caramel ill be indicated by a brownish-violet colour, but the pure curd will also develop a similar colour after standing for a long time, so that great care must be exercised in coming to a conclusion. The use of the control milk greatly increases the delicacy of the test. SuZphonated Axo Dyes.-These all strike characteristic colours with strong acids ; their presence may be confirmed by allowing the milk to go sour and extracting the decomposition products, or, better, by reducing the extracted dye by zinc and hydrochloric acid, making alkaline with ammonia, extracting the ammoniacal solution with ether, and examining the ether extract.NoTE.-since reading the foregoing paper I have fully examined a number of milks for colouring matters. The following results were obtained during the month of March : A. Retail town samples B. Wholesale f a r rn er s’ samples ... ... Total Xxamined. 50 29 Coal- tar only. 9 None COLOUREI). Annatto OIllY. 14 None Annntto and Coal-tar . 6 None Not Coloured . 21 29 These samples being few in number and representing only one month, are, nevertheless, interesting, in that they show the enormous difference between the town and farmers’ samples. These observations are being extended so as to cover the whole year.DISCUSSION. Dr. RIDEAL thought that probably the reducing action to which the author had referred was due to micro-organisms. Dr. Pakes had pointed out about two years previously that organisms of the coli group produced hydrogen, and, if the reducing action was attributable to organisms of the coli group, the occurrence of the bleaching effect also indicated the presence of an organism which ought not to be present in R milk supply. In regard to the possible effect on health of the coal-tar colours152 THE ANALYST. mentioned, it seemed reasonable to look upon them as certainly less injurious than some preservatives were. They were present in very small quantities only, and even phosphine, which many people believed to be very poisonous, had been shown, in some experiments conducted by Professor Ray Lankester and himself, to be capable of being added to the food of small animals in considerable quantities and for con- siderable periods of time without producing any injurious effect.Mr. RICHMOND said that the original object of adding colouring matters to milk was to preserve a continuity of colour between the milk produced from grass in the spring and summer and that produced from artificial food in the winter. Un- fortunately, however, the practice had developed of preserving, not only continuity of colour in milk all the year round, but continuity between milk and skim milk. Certainly in a large number of commercial milks the colouring was completely over- done, and, to one accustomed to the appearance of milk without added colouring matter, a true effect was not obtained.A very simple test for added colouring matter was to let the milk stand. I n the case of genuine high-coloured milk the bulk of the colour rose with the cream and the skim milk remained white, whereas with artificial colouring matter the bulk of the colouring matter would remain in the skim milk, unless the dye happened, as might occasionally be the case, to be one that was prac- tically insoluble in water-such, for instance, as the dye of which methyl orange was the sulphonated derivative. Dr. J. M. H. MUNRO inquired whether the author had tried the experiment of keeping a sample of sterilized milk at the incubating temperature, and if so, whether decolorization had been observed to take place, and whether decolorization took place after the addition of formalin.He would also like to hear whether the author could give any information as to the proportion which the milk samples coloured with coal-tar dyes bore to those coloured with annatto. I t would naturally be imagined that annatto would be more largely used in the country, and coal-tar colours in and around London. He would like to hear what the author considered to be the easiest test for aunatto. The simple test mentioned by Mr. Richmond was, of course, only applicable to whole milk, and not to separated milk. Dr. SCHIDROWITZ said that caramel, although not an injurious substance, some- times had to be tested for-in spirits, for instance, when it was desired to ascertain whether a spirit was cask-coloured or coloured with caramel.The method genera.lly recommended was that of Amthor, but his own experience with that test was by no means satisfactory. In the case of spirits that certainly were not coloured with caramel, he had obtained the precipitate with paraldehyde which Amthor described ; while, on the other hand, a plain spirit, which he had coloured with caramel made by himself, yielded no precipitate whatever. There existed also an idea, which, however, was fallacious, that while naturally coloured spirit was decolorized on the addition of lead acetate or albumin, caramel-coloured spirit was not decolorized. I n neither case was the decolorization perfect, although it was somewhat more marked in the case of the cask-coloured spirit.Mr. WYNTER BLYTH said that he believed he was correct in stating that a pure culture of the typhoid organism in litmus milk did not decolorize the litmus, so that it would appear scarcely possible, from the fact that no decolorization occurred, toTHE ANALYST. 153 draw a safe conclusion that objectionable organisms were absent. He had found that if a poisonous colour, such as Martius’s yellow, were added to milk, the poisonous colour acted as a preservative-an interesting fact which might possibly afford a means of detecting poisonous colouring matters. It would be interesting to see if phosphine acted as a preservative in that way. He was somewhat doubtful whether the simple test mentioned by Mr. Richmond would apply to small quantities of methyl orange.His experience was that methyl orange decomposed very quickly, and probably, on a warm day, by the time the cream had risen properly there would be very little methyl orange left. He had found that butter-yellow rose and coloured the cream, but he believed that this colour was not used very largely, owing to its insolubility in water. He did not think that sterilized milk had any action on arti- ficial colouring matters. He had kept litmus-which decolorized very rapidly in decomposing milk-for several months in sterilized milk without any decolorization ; and the same would be the case, of course, if formalin had been previously added to the milk. He thought that probably the reducing action was due to the evolution of hydrogen by micro-organisms, as Dr. Rideal had suggested. The effect wa8 just the same a~ that of treatment with zinc and hydrochloric acid. It did not occur in the presence of formalin. He could not give any precise information as to the proportion of milk samples containing annatto (see note), but he thought it possible that annatto was used to a considerable extent by country milkmen. Out of two hundred samples taken at railway-stations, as the milk arrived from the farmers, only one W ~ S found to contain coal-tar colouring matter. The 19 per cent. referred to repre- sented town samples only. The easiest teat for annatto was the one proposed by Leffmann and Beam, and consisted in soaking a piece of filter paper in the milk made alkaline with sodium carbonate, when, if annatto was present, a brown colour was obtained on the filter paper after twenty-four hours.