504 JOHNSON AND KING: THE DETERMINATION OF [Vol. 76 The Determination of the Acidity of Milk BY E. I. JOHNSON AND J. KING (Presented at the meeting of the Society on Wednesday, April 4th, 1951) A review is given of the various mlethods that have been used in deter- mining the acidity of milk. Measurements of the increase of colour with increase of pH during the addition of al.kali when phenolphthalein is used as indicator have shown that the comparative method, using a fixed concen- tration of rosaniline acetate in the milk as a colour standard, is sound in principle and gives concordant results. THE acidity of milk is of fundamental importance to the dairy industry and many attempts have been made to standardise a method for its determination so that observers in different laboratories shall obtain concordant results.The difficulty in so doing arises mainly because milk is an excellent buffer from its normal pH to values well on the alkaline side, so that the usual sudden change in colour observable when. a strong alkali is added to a strong acid in the presence of a suitable indicator does not occur. The gradual increase in colour with addition of alkali raises doubts in the mind of the operator as to whether there is any true end-point. The observed end-point usually depends not only on the acidity of the sample but also on the operator’s acuity of colour perception. This depends on whether the operator has normal or abnormal colour vision, on his state of fatigue and on the method of lighting. To minimise these effects it is customary in many laboratories to match the onset of pink colour to that of a standard amount of rosaniline acetate solution added to a given amount of the same milk.l12 Alternatively, the amount of 0.1 N sodium hydroxide required to alter the pH of, say, 10 ml of the milk to a standard pH, usually 8.3, is recorded as the “acidity,” the measurement of pH being performed elect rometrically and not colorimetrically.The effect of adding increasing amounts of phenolphthalein has also been studied, as it is now well known that, with decreasing amounts of the indicator, the pH to which it is necessary to titrate the milk before a pink colour is given rises from about 8.3 with an indicator con- centration of 0.05 g per 100 ml of milk to 8-9 when the concentration is only 0.005 g per 100ml.2,3 The addition of the higher amount of indicator in alcoholic solution to milk at its normal pH leads to precipitation of some of the phenolphthalein in an extremely finely divided form, and this acts as a “reservoir” that increases the amount of the soluble (coloured) form of indicator in the solution as the pH rises.The dilution of the milk by different normalities of alkali used in the titration displaces the end-point. Sommer and Minos,” in a review of the literature on the subject and from their own experimental work, conclude that tri-calcium phosphate is precipitated on the addition of alkali, dilution decreasing the amount affected and hence the titratable acidity. Dilution also decreases the “salt effect’’ and “protein effect” on the colour of the indicator.Kruisheer5 has described six methods, which make use of various volumes of milk, normalities of alkali and concentrations of phenolphthalein, and has compared the results given by some of them in the determination of the acidity of samples of the same milk. In 1931 the Commission Internationale de 1’Eclairage (C.I.E.) adopted certain resolutions for the measure- ment and specification of colour.6 The Lovibond - Schofield colorirneter is a convenient The measurement of colour has been intensively studied in recent years.Sept., 19511 THE ACIDITY OF MILK 505 instrument for measuring colour in terms of units that are readily convertible to the C.I.E. ~ystem.~ The conversion graph with this instrument is based on Judd’s Uniform Chromaticity System,8 which enables colour differences to be expressed in terms of numbers of “least perceptible differences.” This system has been adopted in the following work, which was carried out to ascertain the theoretical basis for the “comparative” method of determining the acidity of milk.EXPERIMENTAL APPARATUS- The Lovibond - Schofield colorimeter was used throughout, measurements being trans- ferred to the graph of equal chromaticities described by S~hofield.~ One of the blocks of magnesium carbonate was removed and a glass cell, which completely filled the aperture, was fixed in its place. The milk to which the indicator or rosaniline acetate had been added was poured into this cell and its colour measured in the normal manner. Direct comparisons without the aid of the colorimeter were also made between a milk to which the standard rosaniline acetate had been added and the milk at various stages of titration in a divided Petri dish, constructed as follows.Molten paraffin wax was poured into a Petri dish of 6.8 cm internal diameter and 1.2 cm depth until the remaining volume was about 30 ml. Before the wax had set, a thin slip of glass measuring about 6-8 cm by 1.1 cm was held centrally in the dish to divide it into two equal semi-circular cells. The wax was then allowed to solidify and the cells were made water-tight by sealing at the ends of the partition with paraffin wax. Molten paraffin wax was then poured into each com- partment until each volume was about 12 ml. For the electrometric determination of pH the meter used was of normal construction with a glass electrode; it could be read to 0.02 pH.The readings of the meter were checked at a range of pH 8.0 to 9-0 with standard buffer solutions. METHODS- For direct comparisons, 10-ml portions of milk were delivered into both halves of the divided Petri dish. To the contents of one cell was added 1 ml of a 0.5 per cent. solution of phenolphthalein in 50 per cent. v/v alcohol and to the other 1 ml of a freshly-prepared 040024 per cent. solution of rosaniline acetate in 50 per cent. v/v alcohol, slightly acidified with acetic acid. A 0.1 N solution of sodium hydroxide was then added t o the cell containing phenolphthalein, while stirring until the colour matched that of the contents of the rosaniline- acetate cell.With a little manipulation it was possible to bring the surfaces of the liquids to the top of the glass dividing partition by dropping in glass beads or by further addition of paraffin wax so that the two fields were practically touching each other with a sharp dividing line. I t is possible to judge a colour match more accurately in this way than by comparing the contents of separate dishes and also more accurately than in the small fields of the colorimeter. It is therefore possible to discern the first slight change of colour at a point somewhat before that necessary to record “one least perceptible difference” by measure- ment in the colorimeter. For measurement of the colours by the Lovibond - Schofield colorimeter, 50 ml of milk were used, the amount of the phenolphthalein and rosaniline acetate solutions being increased correspondingly.This bulk was necessary to provide sufficient material for a viewing cell of sufficient size and for the simultaneous measurement of pH as the titration proceeded. A 0.1 N solution of sodium hydroxide was added in increments of about 0.4ml at a time and the colour and the pH were measured as rapidly as possible, because the colour fades on standing, especially as the pH approaches 9. It was found to be very difficult to measure the pink colour at a pH below 8.2 with accuracy by means of the instrument, but differences in the Petri cells were observable down to a pH of about 8.0. The milk containing the rosaniline acetate was then poured into the cell and its colour measured in the same way as described for the milk plus phenolphthalein.The readings from the Lovibond - Schofield colorimeter were transferred to a scale of equal chromaticity and are shown in Fig. 1. The corresponding pH of the milk for each colour reading is also shown. The lines joining the central point to “Blue,” “Yellow” and “Red” represent the chromaticities of the blue, yellow and red Lovibond glasses on the equal chromaticity scale. The line on the right represents the locus of saturated colours,506 JOHNSON AND KING: THE DETERMINATION OF [Vol. 76 i.e., of monochromatic wavelengths from about Ei65 mp to 585 mp. The colour of the milk before adding indicator (or rosaniline) is represented by a point on the line joining the central point to yellow, showing that this milk is on the yellow side of white by about two “least perceptible differences.” As the alkali is added to the milk containing phenolphthalein, the milk becomes at first slightly pink and then a deeper red with increasing amounts of blue, the final colour being a red - magenta.Starting from the yellow line it will be seen that an increase of pH at first brings about very little change in chromaticity, so that there is little perceptible colour change. From about: a pH of 8.4, however, there is a rapidly increasing colour change with increasing pH. This rate of colour change has been plotted in Fig. 2 to a scale of least perceptible differences, with the corresponding pH and additions of 0.1 N sodium hydroxide. The colour of the milk containing rosaniline is also shown on the same figure, and it is obvious from an inspection of the curves that it occurs at a point where the rate of change of colour with increasing pH is rapidly increasing.Below this pH, 0 5 10 15 -1 Scale of least Perceptibk Differences 570 mp Rosanillne Colour 1 8.08 8.16 8.3 I 836 848 8.56 8.74 8.84 1 Y CHOW Red 580 mi Fig. 1. The dotted line shows alteration of colour with changing pH. Titration of 50 ml Of milk containing Ei ml of 0-5 per cent w/v phenolphthalein. Locations of colours measured on Schofield’s Uniform chromacity chart. The colour changes slowly a t first with changing pH, from a pale yellow to pale pink, the red colour rapidly increasing with increasing pH. the colour change is so gradual that the operator may be in doubt when to cease adding alkali, but at higher pH values the rate of change of colour is so great that a very small addition makes an appreciable colour change.It. is therefore easy to obtain a colour match at the “rosaniline point.” DISCUSSION OF RESULTS It has been a matter of dispute among laboratories as to whether more consistent results are obtained by comparison with the rosaniline acetate tinted milk or by judging the first onset of pink colour. A number of laboratories recently collaborated in a comparative study of the two methods in connection with the formulation of a standard method to be issued with a British Standard, and Mr. H. B. Hawley undertook a statistical analysis of the results. On the whole there was a decided preference for the comparative method. Barkworth and Evans9 also reached this conclusion, but their statistical analysis referred only to results obtained by one operator working always in the same laboratory. In order that an observer shall be in no doubt as to the “end-point” of a titration, it is essential that the rate of change of A study of the diagrams shows why this should be so.Sept., 19511 THE ACIDITY OF MILK 507 colour shall be reasonably great compared with the rate of change of pH.In titrating strong acids with strong bases the rate of change is very great, but when the rate of change of pH is small and uniform with addition of alkali, as occurs with milk (Fig. 2), the circumstances are quite different. Here the rate of change of colour follows an asymptotic curve and does not reach a point where a change is readily observable until a pH of about 8.3 is reached.This point is approximately that of the rosaniline tinted milk on the colour curve, and this fact explains why replicates between observers and laboratories are more easily obtained by this method. I t is the nearest point to the first perceptible colour change at which a rapid change of colour is observed on the addition of alkali and is probably the best point that could have been chosen for the average observer. The milk recorded in Figs. 1 and 2 is that of a bulk-delivery milk issued in the London area by large dairy companies. Milks from Jersey cows are distinctly yellow and the curve 100 mi 9.0 ml 8.0 ml 7.0 ml PH Fig. 2. Rate of colour change (Fig. 1) indicating colour change with change of pH for these milks will be approximately parallel to that shown, but appreciably to the right.I t is unfortunate that acidity as given by the rosaniline standard is slightly higher than that judged from a direct titration, but it seems certain that greater uniformity among laboratories would be achieved by the rosaniline comparative method. In view of the fact that the “titratable acidity” of milk is an empirical value, having no exact equivalent in terms of a given acid (although usually reported as lactic acid), the use of the comparative rosaniline method. is justified. 1. 2. 3. 4. 5. 6. 7. 8. 9. REFERENCES Davis, J. G., and Sadek, G. M., Milk Ind., 1942, 22, 33. Barkworth, H., Dairy Ind., 1944, 9, 20. Pizer, N. N., Chem. and Ind., 1936, 55, 708.Sommer, H. H., and Minos, J., J . Dairy Sci., 1931, 14, 136. Kruisheer, C. I., V I I C0ng.p.. intern. indus. agr. Paris, 1948, Quest. II-H. Smith, T., and Guild, J., Trans. Opt. SOC., 1931, 33, 73. Schofield, R. K., J . Sci. Instr., 1939, 16, 74. Judd, D. B., J . Res. Nat. Bur. Stand., 1935, 14, 41. Barkworth, H., and Evans, E. M., Daivy Ind., 1944, 9, 640. GOVERNMENT LABORATORY STRAND, LONDON, W.C.2508 JOHNSON AKD KING: THE DETERMINATION OF [Vol. 76 DISCUSSION THE PRESIDENT remarked that this was an important paper. The acidity of milk had often been regarded as an easy determination to make and i t was usually carried out by the most junior staff of a food laboratory. But anyone who had had experience of the determination knew how relatively uncertain was the end-point and how i t was dependent on the amount of indicator used.While comparative results could be obtained by the prescribed technique of any one laboratory, the results from different laboratories might differ considerably; so much so that one wondered whether there was such a thing as “acidity of milk” that could be measured by titration involving the use of indicators. The different proportions of buffering salts in different samples of milk might mean that titratable acidity had little significance. It was a good thing that the process had been submitted to critical review. MR. KING in reply stated that any method relying on a change of pH was purely empirical and was by no means a measure of lactic acid. If, however, universally uniform results were to be attained, a carefully standardised method was essential.He welcomed any further suggestions on this point. DR. J. G. DAVIS complimented the authors on the way in which they had scientifically investigated the interesting and important phenomenon of the colour change of phenolphthalein in milk. His interest in the problem of the determination of the acidity of milk arose on visiting a creamery many years ago when he had been told that brine-cooling of the milk resulted in a fall in acidity. The basis of this statement was that the acidities measured by the creamery receiving the milk as compared with those measured before despatch were consistently lower to the extent of about 0.02 per cent. of “lactic acid.” Investigation of the technique used in the two laboratories showed that the first laboratory used only three drops of phenolphthalein, which may be described as the classical method in the dairy industry, while the receiving creamery used 1 ml.This had prompted him to study the effect of concentration of indicator and it had been easy to demonstrate that the pH of the combined milk and indicator, a t the point where the first faint pink colour was perceptible, varied appreciab1.y according to the concentration of the indicator used.1 They had therefore recommended the standard use of 1 ml of a 0.5 per cent. solution of phenol- phthalein for every 10 ml of milk as a reasonable compromise. It was true that this did not settle the vexed question of the colour end-point. A variety of methods had been tried, including the use of pink stirring rods, pink tiles and a special Tintometer glass, and the best of these appeared to be the use of the Tintometer glass, perhaps because the glass could be adjusted to match the end-point colour most accurately.Un- fortunately this method was very expensive and for this reason alone it had not been recommended. In practice, experienced dairy laboratory analysts worked to a very early end-point. Thus, in titrating milk of average quality, a distinct greyness 01 characteristic change in colour was perceptible when the titration was a t about 0-12 per cent. of “lactic acid” and the first perceptible pink colour appeared a t about 0.14 per cent. The inexperienced analyst nearly always worked to a later end-point (more intense pink) and so returned a somewhat higher figure for acidity.Obviously from the scientific point of view a colour standard was required, and the rosaniline acetate method, while giving a reasonably good match, required a special solution to be available, which was a further complication in the test; for this reason i t had not been generally adopted. Dr. Davis emphasised that the test really was a coinplete anomaly in that the value usually stated for fresh milk (0.14 per cent.) was not really lactic acid at all, although this expression was always used. The test really measured the buffer value of the milk from its initial pH (about 6.6) to the pH of the visually detected end-point with phenolphthalein (about 8.4). If the test really measured lactic acid it might be worthwhile to bring out a more elaborate and accuIate form of the test, but for farmers’ milks he suggested that the test should be abolished.Mastitis, which affected about one-third of the cows in this country, resulted in the lowering of the acidity of milk to a variable extent, while a high proportion of solids or feeding on certain types of land could result in a high acidity. As a measure of souring or hygienic quality for herd milks the test was, therefore, not reliable. What was wanted was a specific test for the amount of lactic acid formed in the milk, and work was proceeding on these lines. From the point of view of dairy control i t would have to be a very quick test. The industry urgently wanted a still quicker test that would give in half or a quarter of a minute a reasonably accurate measure of the degree of souring.These objections did not, of course, apply to bulk milks, e.g., as from a 3000-gallon tanker, for the bulking levelled out the variations of acidity in individual herd milks. The titratable acidity test, therefore, was a reliable measure of souring for bulk milks provided that a standardised technique could be developed. Whatever test was recommended would have to be simple and of a type that could be used in the ordinary dairy control laboratory. MR. KING, in his reply, agreed with Dr. Davis on the completely empirical nature of the acidity determination by any method in which phenolphthalein was used as an indicator and further agreed that a rapid and specific determination of lactic acid would afford far more information as to the condition of milk.Until phenolphthalein should cease to be used in determining milk acidity i t was most desirable to standardise the procedure on a sound basis. DR. EGAN said that i t might be that some mixed indicators increased the number of least perceptible differences that accompanied the indicator colour change. Kolthoff and Stenger2 have described the use The present official rejection test, the rapid resazurin test, took 10 minutes.Sept., 19511 THE ACIDITY OF MILK 509 of phenolphthalein in conjunction with a dye, methyl green, the colour of which was complementary to magenta, for the titration of milk: unfortunately the pH a t which the best colour change occurs is 9.0. Dr. Davis had described an effect that may be observed in favourable conditions without the addition of a dye or a second indicator; namely, that the yellow colour of the milk being titrated gave way to a chalky neutral tint immediately before the phenolphthalein colour appeared. Had Mr. King any experience of the colour changes or any comments to make on them? Assuming that the average milk had a spectral reflection curve that was preferential in the yellow part of the spectrum, the superimposition of a slight magenta colour-which was complementary to yellow and given by the first change of phenolphthalein-would result in a chalky-white neutral colour, immediately before the pink - magenta given by a further addition of alkali to the milk. He agreed that the range of pH covered by Kolthoff’s suggested mixed indicator was far too great to give agreement with the usually accepted figures for the acidity of milk. REFERENCES TO DISCUSSION Davis, J. G., and Zadek, G. M., Milk Ind., 1942, 22, 33. Kolthoff, I. M., and Stenger, V. A., “Volumetric Analysis,” Volume 11, Interscience Publishers MR. KING replied that he thought there was a sound basis for Dr. Davis’s observations. 1. 2. Inc., New York, 1947, p. 57.