March, 19501 DETERMINING THE GRADE COLOUR OF FLOUR 133 Experiments in the Photo-Electric Recording of Flour Grade by Measurements of Reflecting Power BY D. W. KENT-JONES, A. J. AMOS AND W. MARTIN (Read at the meeting of the Society on Wednesday, October 5th, 1949) SYNoPsrs-The reliability and usefulness of the instrument described at the same meeting by Kent-Jones and Martin for the determination of the grade colour of flour has been investigated. It is recommended that the grade colour of a flour should be expressed as the dial reading in degrees divided by 10 and corrected to the nearest 0.5 unit. Statistical analysis of the data of reproducibility shows that if the grade colour figure for a given flour is assessed by different operators on different days the results can be expected to agree within 0.5 unit.With flours of ash content below 1 per cent., a range that embraces flours from high-grade patent to 90 per cent. extraction flour, the effect of bleach is within the experimental error of the method. The dial readings of the instrument are linearly related to the logarithms of the corresponding ash contents and the correlation coefficient between dial reading and ash content is 0.959. Some of the aerating ingredients used for self-raising flour influence the dial reading, but it is shown that this influence can be annulled by using a buffer solution in place of water for making the paste. With this modification the instrument gives reliable results with self-raising flour. THE quality of flour is determined by two factors, namely, the suitability of the flour for its intended purpose, eg., bread-making or cake-making, and the colour of the flour.This latter aspect of quality is of such importance that in pre-war days it regulated to a great extent the price of flour. The “colour” of flour is, however, a complex characteristic that is largely governed by two independent factors; these are, first, the brightness or dullness, which is related to the extent to which the endosperm has become contaminated with powdered wheat skins and associated substances during the milling process and, secondly, the degree of creaminess which is dependent upon the amount of carotene in the original wheat and the extent to which this has undergone natural or artificial bleaching. The former aspect of flour colour, i.e., the dullness or brightness, is referred to as the “grade” colour because it is correlated with the degree of extraction, that is, the grade of the flour.Since the two factors which together mainly determine the colour of flour are quite independent, any degree of brightness, or dullness, can be associated with any tint between the full creaminess of an unbleached flour and the extreme whiteness of a highly bleached one. The possible variations in the over-all colour of flour, therefore, cover a very extensive range from a very bright and very white flour on the one hand to a very dull and fully creamy flour on the other. The great commercial importance which attaches to flour colour has led to an evaluation of this characteristic becoming a very frequently applied test in mill and bakery.The earliest method of assessing flour colour, and one which is still the standby of millers and bakers, was a visual test-the Pekar test-in which a comparison is made between the surface of com- pressed slabs of flour before and after they have been dipped in water. The immersion of the slabs in water tends to magnify the colour differences through the action of enzymes,134 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 probably oxidative in nature, which exist in the flour. The visual appearance of such a compressed slab of flour depends, however, not only upon the over-all colour of that flour, but also upon other factors such as the degree of compression to which the slab has been subjected, the moisture content of the flour, the conditions of the immersion of the slab in water and the length of time which has elapsed since the immersion.Although this test does undoubtedly serve a useful purpose in the flour industry, it is only a comparative test, which does not lend itself to numerical expression and which relies upon personal judgment. Attempts have been made to devise methods of evaluating the over-all colour of flour which permit the i-esults to be recorded on a numerical basis. Jagol described a method in which the colour exhibited by the surface of a compressed slab of flour was matched visually against a combination of red and yellow glasses in a Lovibond tintometer. Baker, Parker and Freese2 decided that the over-all colour of flour embraced four colour factors, namely, yellow, red, black and white and that a satisfactory match could be obtained by visual comparison with Maxwell discs rotated at high speed.These methods, however, have never gained popularity in the cereal field and no method of measuring the over-all colour of flour on a numerical basis has become generally accepted. In the milling industry, of course, interest is taken not only in the over-all colour of flour, but in each of the two main independent colour factors, since one, the grade colour, reflects the efficiency with which the milling operation has been performed, and the other, the creami- ness, is a measure of the degree of bleach. It is only natural, therefore, that attention should have been given to the independent determination of each of these factors.The determination of the degree of creaminess is a relatively simple matter involving only the extraction of the carotene with a suitable solvent and the measurement of the intensity of colour in the resulting solution. I t is not surprising, therefore, that numerous methods of performing this test have been recorded. Investigations into this determination have been reported by Kent- Jones and Herd3; Markley and Bailey4; Binnington, Hutchinson and Ferrari6; Binnington, Sibbett and Geddese ; Ferrari and Bailey’ p 8 p9 ; Simpsodo ; Visser’t Hooft and De Leeuwll ; Kent- Jones and Herd12; Kent-Jones and Amos.ls The determination of the “grade” colour of flour is a more complex problem, but it also has received attention in the past.Kent-Jones and Herd3 reported a method involving the extraction of the pigments of the admixed bran powder, i.e., the powdered wheat skins and the measurement of the colour of the resulting solution against a standard solution. This method furnished reliable and reproducible results with flours of 75 per cent. extraction or less, but did not prove to be so satisfactory when applied to long extraction flour such as the present 85 per cent. National flour. The grade colour or brightness of flour is, however, related to the extent to which a smooth surface of the flour reflects light-as is evident from the long-used Pekar test-and Kent- Jones and Martin14 have recently u tilised this relation- ship as the basis of a new procedure for measuring grade colour. They have devised and described an instrument which is capable of measuring fine differences in the proportion of incident light reflected by a flour surface and, by employing a flour paste and light of a prescribed wavelength, have made the measurement independent of variations in the granularity of flour and in the degree of bleach.The technique of Kent-Jones and Martin14 is an advance upon previous procedures for assessing grade colour in two directions. First, the method is identical in principle with the procedure employed throughout the flour industry for judging flour colour in that it depends upon the reflecting power of a flour surface and, secondly, relying as it does upon the reactions of photo-electric cells, it is independent of personal judgment. A numerical evaluation of the grade of flour as provided by the instrument of Kent-Jones and Martin,14 supplemented by a numerical measure of the degree of bleach obtained, for example, by the method of Kent- Jones and Amos,ls should provide a complete picture of the over-all colour of the flour and, moreover, reveal the extent to which each of these two main colour factors contributes to the whole.The purpose of this paper is to present the data which we have accumulated during our experience with the method of Kent-Jones and Martin, and to illustrate the value which the method can have for those in the various sections of the cereal industry who are called upon to make frequent assessments of flour colour. Although we have restricted our present review to wheat flours, the instrument can be used to advantage for measuring the colour of many other cereal products and undoubtedly has possibilities in other fields.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 136 METHOD OF EXPRESSING RESULTS- The circular scale carried by the Kent- Jones and Martin instrument and from which the final readings are taken is calibrated in degrees.With the instrument adjusted so as to give a reading of SO" when balance is effected after an 80 per cent. transmission screen has been inserted in front of the standard cell, as advocated by Kent-Jones and Martin, pre-war patent flours give readings of about 15" to 40°, while flours of 85 per cent. extraction give readings of about 100" to 140". There is much to be said for the use in the cereal industry of a less widely spread scale, particularly as many millers in this country are accustomed to the grade colour scale of Kent-Jones and Herd,3 which ranged from 4.5 to 6.0 for patent flour to 8.5 to 10.0 for basic grades. A scale in this region for the grade colour, which is easily understood by the miller, has the further advantage that it falls in line with the widely used bleach figure scale of Kent-Jones and Herd3 which runs from about 2 to 10.With these considerations in mind we decided to narrow the scale and we have done this by accepting as the "grade colour figure" the figure obtained by dividing the dial reading in degrees by 10 and correcting to the nearest 0.5. REPRODUCIBILITY OF DIAL READINGS- In view of the possibility of this method being adopted by chemists in both the milling and the baking industries, that is by representatives of buyers and sellers of flour, it is most important that the results it gives should show good reproducibility.It is not sufficient that duplicate readings on a given flour paste should show good agreement; the spread of results obtained upon a flour by different operators on different days must be small. The reliability of the method from this angle was therefore determined by statistical analysis of appropriate experimental data. The design of this experiment involved the determination of the dial readings of five flours by two operators on each of 4 days. The operators were unqualified assistants accustomed to performing routine tests. The experimental results appear in Table I.TABLE I DIAL READINGS OF FLOURS DETERMINED BY TWO OPERATORS ON DIFFERENT DAYS Dial readings Flour No. A I 3 1 2 Operator A 1st day . . .. .. 104 122 2nd $7 . . . . .. 108 120 3rd 39 . . .. .. 101 121 4th >> . . * . . . 96 118 Operator B 1st day . . .. . . 100 123.5 2nd 93 . . .. . . 98 119.5 3rd 99 . . .. .. 111 132 4th 33 . . .. .. 102 126 Statistical analvsis of these- data shows that the ratio 3 4 5 146 159 34 150 161 33 147 158 35 143 157 32 153 160 - 150 158 31.5 148 161 32 146 158 32.5 of "between-operators" variance for the "between-dad' variance to the error varianck is 2.01, and the corresponding figure is 2.13; for these ratios to attain significance at a probability level of 0.05 t h e i would need to exceed 4-84 and 3-59 respectively.The standard deviation is 3.2". These data mean that if the grade colour figure for a given flour is assessed by different operators on different days, the results (expressed as dial readings divided by 10) may be confidently expected to agree within h0.5 unit. INFLUENCE OF DEGREE OF BLEACH ON DIAL READINGS- If the instrument of Kent-Jones and Martin does in fact do what it purports to do, that is measure the grade colour or brightness of a flour, then it is essential that the dial reading given by a flour should be uninfluenced by any bleach, natural or artificial that has been conferred upon the flour. Kent-Jones and Martin state that they have rendered the dial reading substantially independent of degree of bleach by the insertion of a filter in the light path.In their paper they quote three tests in support of their contention but state that further tests may be desirable. Since the maximum difference between the dial readings before and after treatment obtained in the tests quoted by Kent-Jones and Martin was136 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 less than the standard deviation established by our experiments, we extended the study to a wider range of flours. A wider discrepancy than that recorded by Kent-Jones and Martin between the readings given by a flour in the unbleached and in the bleached state would not be perturbing, provided it was still within the experimental error of the method. In these experiments flours of various ash contents, obtained from several mills, were tested on the instrument in the unbleached state and also after the flours had been bleached by the addition of one of the bleaching agents in use in the milling industry.The bleaching agents were applied in the proportions in which they are commonly used in the industry. The data obtained in these experiments are given in Table 11. TABLE I1 EFFECT OF BLEACHING REAGENTS ON THE DIAL READING Colour figures (dial reading divided by 10) 1 7- Dial readings Sample Ash, Bleached Unbleached Bleached Unbleached % Benzoyl peroxide, 1/16 oz. per 280 1b.of flour K1 K2 K3 K4 K5 B1 B2 s 1 s 2 s 3 s1 s 2 s 3 s 1 s 2 s 3 0.49 0.61 0.67 0 75 0.88 0.43 0.99 0.42 0.46 1-15 0.42 0.46 1-15 0.42 0.46 1-15 54 77 86 93 116 42 126.5 34 48 124 53.5 72.5 84 93 114.5 31.5 124.5 28 40 119.5 Nitrogen trichloride, 7 g.per 280 Ib. of flour 34 29.5 48 43 124 119 Chlorine, 1 oz. per 280 Ib. of flour 34 31 48 41.5 124 119.5 5.5 7.5 8 5 9 5 11.5 4 0 12-5 3.5 5-0 12.5 3.5 5.0 12.5 3.5 5-0 12.6 5.5 7.0 8 5 9.5 11.5 3.0 12-5 3.0 4.0 12.0 3.0 4-5 12.0 3.0 4.0 12.0 As we had anticipated, differences appreciably greater than the 2.5” quoted by Kent- Jones and Martin were obtained in this extended series of tests of unbleached and bleached samples. The encouraging feature of the tests is, however, that the effect of bleach upon flours with ash contents below 1 per cent.-a range which embraces flours from high grade patent flours to 90 per cent. extraction flours-is within the experimental error of the method, i.e., 3~0.5 units. CORRELATION BETWEEN DIAL READING AND ASH CONTENT- We have explained in the introduction to this paper that the “grade” colour of flour, which Kent-Jones and Martin claim to be measurable by their instrument, is related to the extent to which the endosperm of the grain has become contaminated with powdered wheat skins and associated substances during the milling process.Provided a flour contains no extraneous mineral matter, the proportion of powdered wheat skins it contains is correlated with its ash content, because the ash content of pure endosperm is 0.3 per cent. or less, whereas the ash content of pure wheat skin is in the region of 8 to 10 per cent. It follows, therefore, that if this instrument is to provide a reliable measure of the “grade” colour of flours, there must be a correlation between dial reading and natural ash content.Our next step in this investigation, therefore, was to determine the dial readings and the ash contents of various mill stocks from several mills. As the instrument was designed for use with commercial flours the stocks used in this experiment were restricted to those with ash contents not substantially greater than 1 per cent. The data obtained in this series of tests are given in Table 111.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE TABLE I11 DIAL READINGS AND ASH CONTENTS OF A SERIES GF MILL STOCKS Sample A flour . . B 99 .. c 3) .. D 79 .. B, 79 . . A flour . . B 99 .. c '9 .. D 31 .. A flour .. B 99 .. c '7 .. D 39 .. B, 39 . . A flour . . B 9' .. c 79 .. D 99 .. €3, 9) . . .. .. .. .. .. .. .. .. .... .. .. .. .. .. .. .. .. .. .. .. * . .. .. .. Ash, % Mill 1 . . 0.52 . . 0.54 . . 0.46 . . 0.80 . . 1.11 Mill 2 . . 0.43 . . 0.42 . . 0.43 . . 0.87 Mill 3 . . 0.53 . . 0.42 . . 0-47 . . 1.08 . . 1.06 Mill 4 . . 0.44 . . 0.44 . . 0.51 . . 0-68 . . 1.08 Mill 5 . . 0.37 . . 0.43 . . 0.43 . . 0-66 . . 1.10 . . 0-79 Dial reading 61 59.5 49 99 121 40.5 36 47 108 54 38.5 50.5 122 110 55 56 62.5 98 130 39 39.5 39.5 96 149.5 99 Sample A flour . . .. B 79 .. * . c 9' .. .. D 73 .. .. 1st Bk. flour . . A flour .. .. B '9 .. .. c 99 .. . . A flour . . . . c 11 .. * . D 99 .. . . B, 9' . . . . B.M.R. .. .. A flour .. .. B 99 .. .. c 99 .. .. D n .. .. A, B, C, D flours A flour . . .. B 99 .. .. c 9) .. .. D 3) .. . . 1st Bk. flour . . 2nd Bk. flour . . B, flour . . .. Ash, % Mill 6 .. 047 .. 0.44 . . 0.43 . . 0.74 . . 0-60 Mill 7 . . 0.44 . . 0.42 . . 0-47 Mill 8 . . 0.48 . . 0.44 . . 0-66 . . 0.50 . . 0-77 Mill 9 . . 0.46 . . 0.62 . . 0.47 . . 0-58 . . 0.61 Mill 10 . . 0-45 . . 0.43 . . 0.49 . . 0.61 . . 0.59 . . 0.59 . . 0.59 137 Dial reading 48 43.5 40-5 93.5 82 36.5 37.5 51 56 39.5 80.5 57 96 47.5 51.5 48 68 54 39 39 47 66 80 75.6 70.5 St at istical data reveals that for these ex~eriments the correlation analysis of these coefficient between dial reading and ash content is 0.959. This very' high degree of correlation leaves no doubt that the colour figure of a normal uncontaminated flour furnished by this instrument is a reliable index of the grade of that flour. When the dial readings of Table I11 were plotted against the corresponding ash contents, the scatter diagram thus produced showed that the relationship between the two parameters was not linear.Re-examination of the data established that a linear relationship did exist, however, between dial reading and the logarithm of the ash content. The regression of log ash on dial reading was therefore calculated and was found to be- It must be emphasised that this equation connecting dial reading with log ash applies only to the instrument we used and at its present setting. Each setting and each instrument, therefore, will have its own regression equation. In Fig. 1 the regression line corresponding to this equation has been drawn and on the same diagram the dial readings of Table I11 have been plotted against the logarithms of the corresponding ash contents.Although the degree of correlation between dial reading and ash content is so high, as can be seen from the scatter diagram of Fig. 1, it is possible that two flours of significantly different ash contents may furnish the same dial reading and, conversely, two flours of identical ash contents may yield significantly different dial readings. Examples from Table I11 are A flour from Mill 5 and A flour from Mill 10, flours which give the same dial reading of 39 but have ash contents of 0.39 and 0.45 per cent. respectively; D flour from Mill 5 and B.M.R. flour from Mill 8, each of which gives a dial reading of 96 but which have respectively ash contents of 0.66 and 0.77 per cent.; B flour from Mill 4 and A flour from Mill 7, each of which has an ash content of 0.44 per cent.and yet which give Log ash = 0.0044 x dial reading - 0.5454.138 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 dial readings of 56 and 36.5 respectively; and D flour from Mill 5 and D flour from Mill 8 which are identical in ash content of 0.66 per cent., but which give respectively dial readings of 96 and 80.5. Occasional apparent discrepancies of this nature are not surprising and are, in fact, to be expected. The ash contents of different types of wheat show an appreciable spread and the ash content of the pure endosperm of one wheat may be different from that of another. Furthermore, the various skins on a wheat kernel vary among themselves in both ash content and colour, and hence the relationship between the dial reading and the ash content of an individual mill stock can be influenced by the origin of the stock in question.0.100 0,050 0 - 1.950 T.900 T-850 - 1.800 - 1.750 ‘ - 1.700 ’i. 650 . ’T.550 T500 30 40 5 0 6 0 70 80 90 100 110 120 130 140 150 Fig. 1 Although these facts have not been unrealised in the past, there has been a tendency to ignore them and to accept ash content as a definite measure of flour colour. Our experience, however, has proved that ash content is not always a fully reliable index of the brightness of the flour as deterrnined by its ability to reflect light, which is what is meant by “grade” colour. If an instrument is to be a reliable means of measuring the “grade” colour of flour, the readings it gives will bear a high degree of correlation with the corresponding ash contents, but since the ash content is not always a sure index of the colour, occasional deviations from the relationship should occur.This instrument conforms to these expectations and it is our opinion that providing, as it does, a direct measure of reflecting power, it is a more reliable means of evaluating “grade” colour The position is, in our opinion, as follows.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 139 than is the indirect estimate from the ash content, and hence, where a discrepancy between the two methods occurs, it is the instrument that provides the truer index of the colour. As has been mentioned, the miller in pre-war days paid considerable attention to the ash content of his flour because this figure was accepted as an index of the grade colour.To-day, however, the miller is without this guidance because by Government decree all bread flour contains a proportion of added calcium carbonate (Creta praeparata) and the reported ash content of a flour is not, therefore, a measure of the natural ash content of the flwr. Furthermore, it is not possible to apply a standard correction for the effect of the addition of the statutory proportion of calcium carbonate because it is impossible to effect completely uniform distribution of this addendum. This new method for the measurement of grade colour should, therefore, prove exceptionally valuable at the present time in that it will enable the miller to obtain a numerical assessment of the success of his milling operations despite the influence of added mineral matter upon the ash determination.In Table IV are given the grade colour figures, Le., dial readings divided by 10, of a series of commercial samples of 85 per cent. National flour received recently at our laboratories for analysis and it will be seen that between the worst and 'the best of these, for samples covering the extreme range likely to be encountered under normal conditions, there is a spread of 4 units of grade colour. Included also in this table are the ranges of colour figures to be expected from pre-war white flours and from 90 per cent. extraction flours of the type made for a short period during the war. TABLE IV TYPICAL GRADE FIGURES FOR COMMERCIAL FLOURS Sample Pre-war patent flours (ash contents 0.32 to 0.40 per cent.) Pre-war straight-run flours (72 per cent.extraction) (ash contents 0-44 to 0-50 per cent.) . . . . .. .. National flour (85 per cent. extraction) . . .. .. n * n >¶ .. .. .. n n n n .. .. .. 91 * n n .. n I> n n .. .. .. .. .. I3 n n n n n n n .. .. .. .. .. * . * 99 9s n n n n n .. .. .. .. .. .. n n n n .. .. * . n n n n n n 19 n n n n n n 99 n n .. .. .. .. .. .. .. .. .. .. .. .. 90 per cent. extraction flour . . .. .. .. Grade colour figure 1-4.0 4.5-6.5 9.5 10.5 11.0 9.0 10.5 13.0 10.0 9.5 11.0 13.5 10.0 10.0 12.0 9.5 13*0-16*0 DETERMINATION OF GRADE COLOUR OF SELF-RAISING FLOURS- The evaluation of the grade colour of self-raising flour has always been a problem. The inclusion of over 7 lb. of mineral aerating ingredients in 280 lb.of flour and the possibility of minor variations in the distribution of the ingredients renders the ash content of a self- raising flour valueless as an index of colour. Separation of the added mineral ingredients by high-speed centrifuging prior to the determination of ash content has been suggested,ls but the method has not been widely adopted. The present colour method, therefore, seemed to offer a simple solution to the problem, provided any interference caused by the added chemicals could be overcome. A preIiminary experiment (Table V) revealed that the addition of acid calcium phosphate and sodium bicarbonate to a flour in the proportions used for self-raising purposes significantly diminished the dial reading given by a flour by the normal technique.The addition of normal proportions of acid sodium pyrophosphate and sodium bicarbonate did not, however, seriously alter the dial reading. In view of the interference caused by the presence of acid calcium phosphate, which is the commonly used acid ingredient of self-raising flours, it was decided t o investigate the effect of using a buffer solution in place of water in the preparation of the flour paste. Experi- ments showed that if the paste were made with a citrate buffer solution, the reading of a plain140 KENT- JONES, DIAL READINGS Natural ash content, yo . . AMOS AND MARTIN: EXPERIMENTS I N THE [Vol. 76 TABLE V OF FLOURS CONTAINING MINERAL ADDITIONS PASTE MADE WITH WATER Flour A B C D A f \ .. .. .. . . 0.41 0.56 0.78 1.20 Dial reading with water (normal procedure) .. .. 39 68 105 169 Dial reading witth water after addition of 341b. NaCO, and 4+ lb. A.C.P. per 280 lb. (normal self-raising flour) 25 48 87 151 Dial reading with water after addition of 34lb. NaCO, and 64 lb. proprietary pyrophosphate mixture per 280 lb. (normal self-raising flour) . . .. .. 39 62 105 168 flour was not altered and, moreover, the conversion of the flour to a self-raising flour by the addition thereto of sodium bicarbonate and acid calcium phosphate or acid sodium pyro- phosphate did not markedly affect the dial reading. Furthermore, when the buffer solution was used to make the paste, a flour containing a marked excess of acid calcium phosphate, acid sodium pyrophosphate or sodium bicarbonate, gave a dial reading not vastly different from that of the untreated flour.A series of results from these experiments is given in Table VI. TABLE VI PASTE MADE WITH BUFFER SOLUTION DIAL READINGS OF FLOURS CONTAINING MINERAL ADDITIONS Flour Natural ash content, % . . .. .. . . .. Dial reading with water (normal procedure) . . . . Dial reading with buffer . . .. .. * . .. Dial reading with buffer after addition of 34 lb. NaHCO, and 44 lb. A.C.P. per 280 lb. (normal self-raising flour) Dial reading with buffer after addition of 3) lb. NaHCO, and 64 lb. proprietary pyrophosphate mixture per Dial reading with buffer after addition of 4+1b. A.C.P. Dial reading with buffer after addition of 64 lb. pro- prietary pyrophosphate mixture per 280 lb. (excess of Did reading with buffer after addition of 34 lb.NaHCO, 280 lb. (normal self-raising flour) . . .. .. per 280 lb. (excess of acid) . . .. .. . . acid) .. .. . . .. . . .. per 280 lb. (excess of alkali) . . .. .. .. r A 0.4 I 39 41 40 39 41 39 35 B C D‘ 0.56 0.78 1.20 68 105 169 65 103 172 67 106 164 62 105 168 70 I 173 174 - - A stock buffer solution is prepared by dissolving 110 g. of disodium hydrogen phosphate and 77 g. of citric acid in water and making up to 1 litre. When required for testing self-raising flour this solution is diluted fivefold. These tests show, therefore, that the method can be applied to self-raising flours equally as well as to plain flours provided the flour is made into a paste not with water but with a citrate buffer solution of specified strength. By this technique it is possible to obtain a reliable numerical assessment of the colour of a self-raising flour.This instrument, therefore, provides a long-felt want in that it enables the grade colour of flour to be expressed on a numerical basis, even when that flour contains mineral additions which invalidate the use of the ash test as a measure of colour. SUMMARY The instrument of Kent-Jones and Martin enables the grade coldur to be expressed numerically and the results it gives have good reproducibility in the hands of different operators (irrespective of the day upon which the test is made). The “grade colour figure” obtained by dividing the dial reading by 10 can be determined within f0-5 unit irrespective of the degree of bleach on the flour. The dial readings obtained upon this instrument are linearly related to the logarithms of the corresponding ash contents and there is a very high degree of correlation between dial reading and ash content.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 141 The grade colour of self-raising flours can be determined by this instrument by employing a buffer solution in the preparation of the paste.in the 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. It is with pleasure that the authors place on record their appreciation of the ready willingness of Dr. E. C. Wood to guide them in the theoretical and practical issues involved statistical aspects of this paper. REFERENCES Jago, W., and Jago, W. C., The Technology of Breadmaking, Simpkin, Marshall, Hamilton, Kent Baker, J . C., Parker, H. K., and Preese, F.B., Cereal Chem., 1933, 10, 437. Kent-Jones, D. W., and Herd, C. W., Analyst, 1927, 52, 443. Markley, M. C., and Bailey, C., Cereal Chem., 1935, 12, 33. Binnington, D. S., Hutchinson, W. S., and Ferrari, C. G., Ibid., 1941, 18, 10. Binnington, D. S., Sibbitt, L. D., and Geddes, W. F., Ibid., 1938, 15, 119. Ferrari, C. G., and Bailey, C. H., Ibid., 1929, 6, 218. t , Ibid., 1929, 6, 347. -- , Ibid., 1929, 6, 457. Sirnison, A. G., Ibid., 1935, 12, 569. Visser’t Hooft, F., and De Leeuw, F. J. G., Ibid., 1928, 5, 351. Kent-Jones, D. W., and Herd, C. W., Ibid., 1929, 6, 33. Kent-Jones, D. W., and Amos, A. J., Modern Cereal Chemistry, 4th Edition, Northern Publ. Co., Ltd., Liverpool, 1947, p. 473. Kent-Jones, D. W., and Martin, W., Analyst, 1950, 75, 127. Gustafson, C.B., Cereal Chem., 1931, 8, 475. & Co., Ltd., London, 1911, p. 709. -- 88, MADELEY ROAD LONDON, W.5 DISCUSSION MR. A. L. BACHARACH called attention to the difference between “dial readings” for bleached and unbleached flour. Although the maximum difference between pairs of readings was 10, and although this represented only twice the expected difference from the mean (5 units), the effect of bleaching might well be significant because all the differences produced by i t were in the same direction. DR. J. H. HAMENCE asked whether there was any means of checking the instrument against a standard. MR. D. M. FREELAND pointed out that the correlation of colour units with ash content demonstrated by the authors could apply only these days to “pure” mill streams.Most users of flour found ash figures were increased because of the presence of 14 oz. of prepared chalk per 280 lb. of flour and the ash - colour relation was thereby disjointed. Would not some approximate ranges of colour units to the usual observa- tions of the Pekar test be more useful to the flour buyer? MR. E. SEAL said that, as the grade colour was really that of the flour-water paste and not the flour itself, it was not exactly a measure of flour colour. A soft flour would appear much brighter in the dry state than a strong granular flour of equal grade. Had Dr. Amos observed these actual differences between weak and strong flours of equal grade-figure as measured on this instrument? He also asked whether the use of the buffer with plain flours would extend the period of time during which a measurement could be accurately made on the paste.DR. C. R. JONES endorsed the claim that there was need for a method of measuring brightness of flours, and considered the results given most encouraging, subject to the important question already raised as to reproducibility of results with different instruments. He suggested the term “flour-water suspensions” be used in place of “flour pastes” to avoid confusion with gelatinised preparations. DR. A. GREEN said that he considered that the reference by the authors to the presence of carotene in wheat flour was not justified. Both Zechmeister and von Euler had failed to find it (see also Bacharach, Analyst, 1941, 66, 36). The degree of bleaching of the test flours was not in general that which is used in commercial practice.Would the application to the more branny stocks of heavier treatments than those described by the authors still yield practical identity in light reflection for the same mill product when unbleached and when bleached ? In view of the fact that nitrogen trichloride and chlorine dioxide have different effects on the pigments of bran, had the authors tested the performance of their instrument on mill products containing finely divided bran, before and after these products had been treated with chlorine dioxide ? It is possible that light of a different wavelength might be needed in this case. DR. K. A. WILLIAMS mentioned that he used a very similar method in determining the colour of oils and fats. In that case he had preferred to determine the ratio of the transmissions a t two wavelengths rather than a figure a t only one wavelength.Attention had recently been paid to this method in America. DR. J. STRAUB called attention to the fact that greyness of flour is measured by the apparatus much more accurately than it could be assessed visually. He felt that such precision would prove its usefulness in research in preventing greyness of flour. When trying out new techniques of milling for the removal of142 KENT- JONES, AMOS AND MARTIN [Vol. 75 greyness, the slightest invisible change in greyness would show whether progress was being made. The laboratory was t o be congratulated on its results. DR. N. L. KENT asked whether a high correlation between dial reading and ash content was obtained if a series of flours taken right down the reduction of a mill, from A to K, L or M, Le., including both high grade and low grade flours, was used.He felt that the method which had been described for measuring the colour of flour in a positive way would be welcomed by millers, and more particularly by the miller abroad. For the first time i t would be possible for him to compare the colour he obtains in a simple and reliable way with that achieved by millers in this country. Therefore the speaker asked for an assurance that no local factors would upset the figures he might obtain, thus preventing a true comparison with the figures obtained in this country. He asked whether the character of the water would affect the result, and whether the temperature of the water would alter the readings.DR. AMOS said in reply that the fact that the slight effect of bleach upon the dial readings had always been in one direction might have some significance, but the purpose of the quoted figures was t o show that the magnitude of this effect upon the readings was within the experimental error of the test. One of these, which was mentioned in the paper, was to insert an 80 per cent. transmission screen in one light path and to adjust the instrument so that a given reading was obtained upon the dial. The other method was t o adjust the instrument so that the readings of two or three flours of known ash contents were those required by the regression equation for the instrument in question. The ideal method would probably be calibration against a standard white surface, but this would necessitate a standard which was absolutely permanent, and it had not been possible to meet this essential criterion.Mr. Freeland had perhaps misunderstood the position) The authors had demonstrated that their dial readings were closely correlated with the true ash contents of flours as proof that their readings were a measure of grade colour. The fact that this correlation would not be apparent in National flours because of the presence of chalk was beside the point; the instrument was designed to measure grade colour, and this it did whether chalk was present or absent. Although the reading was made on a flour paste, it was nevertheless a true measure of the grade colour of the flour, as had been shown by the high degree of correlation between the readings and ash contents. Certainly a soft flour appeared brighter than a strong flour of equal grade because of the influence of granu- larity, and i t was for this very reason that the authors had eliminated the influence of granularity by the use of a paste.The use of the buffer with plain flours had not been tried, but there seemed no point in investigating the possibility of extending, by this means, the time that a paste could be kept before the reading was taken, since with a water paste no change in the reading occurred in 30 minutes. The rates of bleaching treatment that were employed were average values for the commercial bleach- treatment of straight-run flours, and they had not significantly affected the readings when applied t o flours ranging from patent grade to a grade equivalent to 90 per cent.extraction. It was true that in commercial practice certain very low-grade stocks would receive a much higher dosage, but this point had not arisen because in this paper the authors had restricted their investigations to stocks with ash contents not significantly greater than 1 per cent. The use of chlorine dioxide was not included in the bleaching experi- ments because it was not in commercial use in this country, but in view of the fact that it was likely to be so used at some time in the future, tests with this product were in hand. The correlation coefficient between dial reading and ash content quoted in the paper applied to a series of flours ranging from high patent grade to those with ash contents slightly higher than 1 per cent. Tests with low-grade stocks having ash contents of 3 per cent. and over had shown a very good degree of correlation. The nature of the water used to make the paste would not be likely to influence the results seriously; the authors had tried distilled water and tap water and had found no difference in their results. Tests made during the summer had not been affected by the temperature, but the high temperature experienced in some countries overseas might lead to a slight difference in the readings; this could be tested and due allowance made. MR. F. T. HOLDEN said that he was a miller from British East Africa. The instrument could be standardised by two methods. THE CALCULATION OF THE BOTANICAL COMPOSITION OF WHEAT FLOURS AND OFFALS FROM THE CHEMICAL ANALYSIS AT the meeting at which the above two papers were read, a lecture with this title was delivered by J. Straub, Chem. Ing., of the Central Institute for Nutrition Research, Utrecht. The following is a summary of Mr. Straub’s lecture, which has been published in full in Rec. Trav. Chim. Pays-Bas, 1950, 69, 141. A method was described for calculating the botanical composition, in terms of the percentage of pericarp, aleurone and endosperm in wheat flours and offals, from the amounts of crude fibre, phosphorus and starch in the milled product; these three constituentsMarch, 19501 EL-SOKKARY AND HASSAN : EGYPTIAN MILK 143 being characteristic respectively of the pericarp, the aleurone and the endosperm. Hence, from a knowledge of the amounts of fibre, phosphorus and starch that characterise the different botanical tissues it is possible to calculate the relative proportions of each that might be present in a sample, without having to separate them in a pure state. As a further consequence of the work it was possible to calculate the protein and fat content of the botanical fractions. It was also shown that the endosperm consists of two distinct parts, an inner and an outer, of widely differing starch and protein contents, and of which the proportions could be determined in milled products.