Analyst, October, 1968, Vol. 93, $9. 669-679 669 The Automatic Determination of Original Gravity of Beer Part I. Introduction and Determination of Reducing Sugar after Hydrolysis BY R. SAWYER AND E. J. DIXON (Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) The determination of the “extract” gravity of sound beers by automatic means is presented and discussed. A method is proposed for correlating “extract” as determined by classical distillation methods with the total reducing sugar after hydrolysis as determined in an automatic analytical system operating at 30 samples per hour, and this is shown to have the form- E - 1000 = 4.33 S + 1-2, where E is the extract weight (expressed relative to water, sp. gr. 1000) and S is the percentage of sugar as anhydrous glucose.The use of the method as part of a screening technique for examina- tion of samples against declaration is discussed. EACH year more than 20,000 samples of sound beer are examined in this laboratory for the purpose of controlling duty. This duty is based on the “original gravity” of the sugar solution, or wort, from which the beer was brewed. The method of determining this, as laid down by Statute,l requires that a measured volume of beer, taken at a fixed temperature, be distilled, and the distillate made up to the volume of the original sample at the same tempera- ture and its specific gravity determined. A table in the Statute gives a value for the “gravity lost” which, when added to the gravity of the distillation residue, again at the original sample volume and temperature (residual sugar or “extract”), furnishes the “original gravity.” As the determination of original gravity by the Statutory method is a relatively lengthy process, a rapid method of screening samples of beer against a declared value of original gravity is desirable.One such method, in which manual measurements of the physical constants, refractive index and present gravity of the beer are usedJ2 is now in use. The potential operating speed and reduced operator attendance of automatic analysis systems offered an alternative to the rather tedious operation that gave laboratory juniors little satisfaction in their work. Preliminary indications of the possible use of automatic analytical techniques in this field appeared in published methods by the Brewing Industry Research F o ~ n d a t i o n .~ ~ ~ s6 These papers, in which methods for the automatic determination of reducing sugars and alcoholic content of beers were described, appeared to form the basis of the measurement of the two parameters necessary to determine original gravity. The major portion of the distillation residue in beer samples is a mixture of poly- saccharides and the fragment sugars from hydrolysed starches together with any residue from added cane and other priming sugars. A survey of the relationship between “extract weight” and reducing value after hydrolysis was carried out in this laboratory by using manual techniques ; the correlation between these two factors was sufficiently encouraging to allow the development of automatic techniques to be pursued.Preliminary reports of the results of this work have been described elsewhere.6 This paper is concerned with the development of the technique of determining the reducing value of the residual sugars in beer in such a way that the results could be used to calculate the extract weight of the beer. 0 SAC; Crown Copyright Reserved.670 SAWYER AND DIXON AUTOMATIC DETERMINATION [Analyst, VOl. 93 DETERMINATION OF REDUCING VALUE- The “extract” specific gravity for the majority of beers of all types lies between the limiting values 1.004 to 1.048, which are about equivalent to a range of reducing values after acid hydrolysis of 1 to 10 per cent., expressed in terms of glucose. As the colour of beer varies from a pale straw colour for lagers to dark red - black for stouts, any colorimetric method applied to the whole range of samples must allow for these extremes of colour.Automatic methods with dialysis techniques were found to be unsatisfactory because of the varying amounts of coloured materials that dialysed from widely different samples. Methods of determining reducing sugars in solution by manual techniques are based on the use of alkaline oxidants such as complexes of copper salts, e.g., Fehling’s solution and the many variants of this formulation; those most frequently encountered involve the use of gravimetric or titrimetric finis he^.^ g 8 Several variants in colorimetric finish of these exist ; the Somogyig modification of Nelson’slo reagent with molybdoarsenate reagent to measure reduced copper in suspension has been used in both manual and automated methods.11,12 These modifications are used with sugar solutions at a final concentration of 0.001 to 0.008 per cent.of dextrose, which is attained with a dilution ratio of 1 : 1250 for all beers. At this dilution the optical density of the dark stouts is in the region of 0-005 per cm at 500 t o 550 mp and may be regarded as analytically significant if an optimum optical density range of 0 to 0.3 per cm is used in the colorimetric determination. Furthermore, it has been statedll that at the upper levels of sugar concentration the reagent may be prone to deposit copper oxide. Intermittent precipitation of this nature cannot be tolerated in the automatic analytical system, as particulate matter in suspension produces extremely noisy traces and promotes base-line drift because of deposition of solid matter on flow-cell windows.For these reasons the copper reduction methods were not considered further. Alkaline ferricyanide was first proposed as an oxidising agent for sugars by Gentele13 and, subsequently, interest was revived by the introduction of the reagent in the analysis of blood sugars by Hagedorn and Jensen.14 Colorimetric versions of the technique were proposed by Hoffman,15 who determined loss of ferricyanide colour, and by Folin,16 who determined the Prussian blue formed on addition of an acid iron(II1) salt to the reduced ferricyanide. Automatic analytical systems with the Hoffman mode of analysis were proposed by Hill and Kessler,l7 and reports of various modifications of this method have appeared from ’time to time.l89l9 The “colour loss’’ technique is limited, however, as measurement of optical density is made at 420mp with solutions at concentrations in the range of 0.0006 to 0.007 per cent. ; again the colour variation of the diluted samples has analytical significance with these methods, as a 1000-fold dilution of a dark beer yields a solution with an optical density of 0.02 per cm at 420 mp.Automatic methods based on the formation of coloured species from the ferricyanide produced in solution have been described by Fuller,20 who used Prussian blue formation and by Fingerhut, Ferzola and Marsh,21 who used a molybdo- phosphate complex of ferrocyanide in acidic solution.These two approaches were considered to be worthy of further investigation as finishes in the sugar determination. It was necessary to develop a satisfactory finish technique before a study on the conditions of hydrolysis could be carried out and initially the Prussian blue technique, which we found satisfactory, was used, but for reasons given below this was later abandoned in favour of the molybdophosphate finish.21 Methods given that combined acid hydrolysis with a colorimetric determination, such as those based on sulphuric acid a n t h r ~ n e , ~ ~ ~ ~ ~ were rejected because of the sample colour variation and the undesirability of 27 N sulphuric acid for routine use. EXPERIMENTAL In all our operations on the AutoAnalyzer with colorimetric finishes our aim has been to optimise the chemistry of the finish to produce a solution with an optical density within the range 0 to 0.3 referred to blank reagents at the working wavelength and with one of the standard flow cells in the system.This mode of operation has been used so that the 50 to 100 per cent. transmission portion of the recorder chart could be expanded to full scale by the use of a x 2 range expansion. The resulting calibration graph is then spread over the maximum chart width and produces a calibration approaching a straight line with most methods.October, 19681 OF ORIGINAL GRAVITY OF BEER. PART I 67 1 APPARATUS- The components of the AutoAnalyzer assembly used were as follows. Voltage stabiliser; Sampler I1 operating at 30 samples per hour with de-ionised water wash between samples, the ratio of sample time to wash time being 2 : 1 ; two proportionating pumps (two speed) with accommodation for 15 tubes on each pump; one heating bath, set to operate at 95" C, with two delay coils and a colorimeter with 8-mm flow cell equipped with 600 mp or 480 mp filters; and a two-pen recorder with range expander on each amplifier.PRINCIPLES OF MANIFOLD DESIGN- A manifold has been designed to combine dilution, acid hydrolysis, reaction with alkaline ferricyanide, and subsequent reaction with a colour-forming species, to measure the hydrolysed sugars in the full range of beer samples. In order to obtain a suitable optical-density scale with the Prussian blue finish for the beer samples a dilution of 25,000 is necessary through the four stages of the method, which is more than required for removal of beer colour variation.The manif old design has been successfully developed by particular attention to detail in assembly of components to give the minimal flow path lengths for unsegmented liquid streams. Re-cycling and dilution stages are carried out by introduction of polythene tube (0-03 x 0.48 inch) into a close-fitting PVC insert in the lower limb of all C5 de-bubblers. The tube is inserted directly into a pump tube cut off immediately in front of the end block stop. A further length of polythene tubing is used to carry the unsegmented stream after pumping from a shortened end of the pump tube to the metal insert H3 type of connecting piece (Fig. 1). The full manifold layout is indicated in Fig.2, together with details of the H3 I 0.025 inch i.d. pump tube b\d fi 0.030 x0.048 inch h'Yt\ tube PVC insert \- pol y t hene Fig. 1. Connection of debubbler to pump tube and return to manifold for dilution reagents used. Observation of these basic principles of operation is essential for successful application of a multiple dilution method at sampling rates of 30 per hour with maintenance of adequate sample discrimination as shown in Fig. 3. PRUSSIAN BLUE FINISH- A typical calibration obtained over the range 1 to 11 per cent. of dextrose with the Prussian blue finish is indicated in Fig. 4; the method shows a pronounced curvature of response with cut-off at sugar concentrations lower than 1 per cent. Experimentation with the reagent system shows that the cut-off level can be adjusted by altering the sugar dilution or by adjusting the iron(II1) ion concentration of the colour-forming reagent.An increase in iron(II1) ion produces a colour within the "toe" of the curve but the same reagent produces a precipitate at higher sugar levels (z.e., higher resultant ferrocyanide) ; similar effects are obtained by increasing the sugar concentration in the oxidising solution. Comparison of this standard curve with that obtained with the molybdophosphate procedure detailed below, indicates that the threshold limit on the product [Fe3+J, [Fe(CN),4-], is influenced by the complex formed with free [Fe3+] by the phosphoric acid used as the672 SAWYER AND DIXON : AUTOMATIC DETERMINATION [Analyst, Vol. 93 1-2Air ", 3-9 Water A O.'I Sample 30/hr + Debubble t o waste 1.6 Hydrochloric acid 30 per cent.v/v + Debubble t o waste I -6 Potassium ferricyanide 0.25 per cent. wlv Sodium hydroxide 2-5 per cent. w/v X 0.42 A i r --t Debubble to waste Ammonium iron (111) sulphate Sodium lauryl sulphate 0.2 per cent. wlv 0.1 per cent. w/v Debubble t o waste Phosphoric acid 2.5 per cent. w/v I 2.4. Waste 4 Colorimeter 600 mp I 0.8-cm flow cell Range expander Recorder Fig. 2. Manifold for the total hydrolysable sugars in beer by the Prussian blue method acidifying medium, and the solubility product is exceeded only when an amount of ferro- cyanide is produced equivalent to the oxidation of sugar at the 1 per cent. level in the original sample. The reagent system shown is the best compromise solution operating over the range 1 to 11 per cent.of sugar in the original samples taken. The early report of this work6 describes the use of the Prussian blue technique for the sugar determination ; subsequently, this finish was abandoned in favour of the molybdophosphate determination. The principal reason for this change is that the threshold limit, and hence the calibration curve shape has been found to be susceptible to the quality of the phosphoric acid used in preparing the ammonium iron( 111) sulphate reagent. The use of different batches of analytical-reagent grade phosphoric acid (especially from one manufacturer) produced precipitation of Prussian blue to varying degrees in the upper part of the calibration scale. Furthermore, in studying the variation of acid concentration necessary to give optimum hydrolysis conditions, it hasOctober, 19681 OF ORIGINAL GRAVITY OF BEER.PART I 673 been found that precipitation also occurs at the colour finish stage when the hydrolysis is carried out with hydrochloric acid at concentrations over 25 per cent. v/v. Although the curved response line could be overcome by addition of an appropriate amount of ferrocyanide to the ferricyanide reagent, the susceptibility to what appears to be extraneous iron contamination could not be readily eliminated. The molybdophosphate reagent system appears to tolerate those variations in reagent quality that were troublesome in the Prussian blue method, I oc U C V L aJ a C .- U 0) E U 6.0 I I I Fig. 3. Dextrose standards, Prus- sian blue method u C 8 L u a C U u E .- n I 2 3 4 5 6 7 8 9 10 Glucose, per cent.Fig. 4. Calibration curve, Prussian blue finish, of sugar as glucose, per cent. w/v, against full scale deflection, per cent. MOLYBDOPHOSPHATE FINISH- The method of Fingerhut, Ferzola and Marsh,21 which is based on the reaction between molybdate ion and ferrocyanide in acidic s o l ~ t i o n , ~ ~ ~ ~ ~ p26 has been adapted to satisfy the conditions required for analysis of beers containing 1 to 10 per cent. of reducing sugar after hydrolysis. A manifold is described by Fingerhut in which molybdophosphoric acid is added directly to the alkaline ferricyanide oxidation stream directly on emergence from the 95" C heating bath and the resulting colour measured at 450 mp. We have found that if this course of action is followed the traces on the recorder chart are not suitable for application to our method of analysis. The curves owe their irregularity to three major faults.Firstly, 4ml per minute of reagent are added to an existing reaction mixture that is pumped at 1.2 ml per minute, causing momentary stoppages A typical set of traces is illustrated in Fig. 5. - 0 8 U I 0 a C 0 Y) .- .- 5 E I- I00 Fig. 5. Response, per cent. of dextrose by the original method674 SAWYER AND DIXON : AUTOMATIC DETERMINATION [Autdyst, VOl. 93 and surges in the main reaction stream. Secondly, direct addition of the acidic reagent to the alkaline reaction stream causes initial local concentration variations in the solution before mixing takes place, and at acid pH values and high ferrocyanide concentrations precipitation from solution of the reaction product occurs ; this precipitate forms around the liquid junction and is detached in fragments by the flowing stream.Thirdly, at the wavelength chosen for measurement (450 mp) the reagent system at zero sugar concentration has an appreciable optical density, as is shown in Fig. 6. Irregularities in the addition of reagent are amplified by measurement at such a point on the spectrum. A modified system designed to avoid these features has been assembled and the layout is indicated in Fig. 7; this manifold is the one finally used for the sugar determination in the beer samples. I .o - 0.5 - 1 I -4. * . . . . ._. ...- ....... ......- 400 450 500 550 Wavelength, my Fig. 6. Spectrum plots of molybdophos- phoric acid - ferrocyanide system for: A, reagents against water; B, reaction mixture for 10 per cent.sugar against water; C, reaction mixture against reagents DETERMINATION OF OPTIMUM ACIDITY FOR HYDROLYSIS- The choice of acid concentration and time of hydrolysis of samples to obtain maximum reducing sugar yield for a given set of samples are dictated by the geometry of the apparatus and by the selection of suitable solution pumping rates. Choice of pumping rate is pre-deter- mined by that ratio of liquid to segmenting air which will provide a satisfactory bubble pattern with good sample discrimination, after allowing for expansion and contraction of the air bubble on entering and leaving the heating bath. The slowest rate of pumping to satisfy these conditions is about 2 ml per minute total volume, with a liquid-to-air ratio of 4: 1.In these circumstances the time in a standard coil in the heating bath is about 6$ minutes per sample, which can be modified by varying the air segmentation rates from about 3 to over 8 minutes, but at both extremes the liquid segmentation is erratic. Initial chromatographic examination of samples after acid hydrolysis with varying con- centrations of hydrochloric acid indicated that an almost complete hydrolysis of higher polysaccharides is obtained with a working acid concentration of 2.3 N. The results of a further examination of various representative samples of beer with 15 to 40 per cent. v/v hydrochloric acid and an acid-to-sample ratio of 16 : 3, and with the full colorimetric finish, are indicated in Fig.8. The graphs presented show that an optimum yield is obtained with most beers when the hydrolysis acid has an initial concentration in the region of 30 to 40 per cent., the highest reducing value for primed beers is obtained with 30 per cent. v/v acid while the lagers require an acid concentration approaching 40 per cent. v/v. The con- centration chosen is a compromise value to include the whole range of beer types withinOctober, 19681 OF ORIGINAL GRAVITY OF BEER. PART I 675 A9 A2 A4 I Cold c5 95” c c / n 4 0.32 Sample 3.4 Water 3.4 Water 3.4 Air Debubble to waste 0 *23 Concentrated hydrochloric acid 35 per cent. v/v 1 -6 0.42 Air Debubble to waste 0.32 o.8 Potassium ferricyanide 0.50 0.8 Sodium hydroxide 8-4 per cent.0.8 Air per cent. w/v wlv Debubble to waste 0.32 3.9 Concentrated hydrochloric acid 1.2 Air 4.5 per cent. v/v 0.6 Molybdophosphoric acid 3 -9 Pumps A and C ml/min Fig. 7. Manifold for total hydrolysable sugars in beer one screening method. then a hydrochloric acid concentration suitable for that beer type can be chosen. If, however, the method is to be used with a single type of beer, METHOD REAGENTS- Hydrochloric acid for hydrolysis-Dilute 3.5 litres of analytical-reagent grade concen- trated hydrochloric acid to 10 litres with distilled water. Hydrochloric acid, dilute-Dilute 450 ml of analytical-reagent grade concentrated hydrochloric acid to 10 litres with distilled water. Potassium ferricya.nide-Dissolve 25 g of analytical-reagent grade potassium ferricyanide in 5 litres of distilled water.Sodium hydroxide-Dissolve 420 g of analytical-reagent grade sodium hydroxide in 5 litres of distilled water. Molybdophosphoric acid reagent-Add 350 g of analytical-reagent grade molybdenum trioxide and 5 0 g of analytical-reagent grade sodium tungstate (hydrated) to 1 litre of distilled water in a 5-litre beaker, stir, add 200 g of sodium hydroxide, and boil the mixture for 20 to 30 minutes. Cool the solution, add 1 litre of distilled water and cautiously add676 SAWYER AND DIXON : AUTOMATIC DETERMINATION [A%@&$, VOl. 93 1-25 litres of 88 per cent. analytical-reagent grade orthophosphoric acid, cool and dilute to 5 litres. Standard solutions-Prepare standard solutions, in the range of I to 10 per cent. (w/v) in unit steps, of analytical-reagent grade anhydrous glucose.In practice these standards contain similar unit volumes of alcohol for the alcohol method reported in Part I1 of this series (p. 680). 2 2 3 W v) 2 I * #.* - - - - -+ - - - - - , . . - M- IS 20 25 30 35 40 Hydrochloric acid, per cent. v/v for hydrolysis A = Continental beer G = Brown ale B = Bitter ale H = Stout C = Danish Pilsner J = Milk stout D = Pale ale K = Milk stout E = Export ale L = Milk stout F = Polish lager M = Strong ale Fig. 8. Variation of sugar determined after hydrolysis with different acid concentrations PROCEDURE- The samples for analysis are de-gassed by rousing and filtration through a Postlip filter-paper (18.5 cm, Evans Edlard and Co.) contained in a polythene funnel. The first runnings are rejected and the bulk sample is collected in a squat 250-ml beaker.A series of standard solutions is placed in 10-ml sample cups and placed in position in the Sampler 11, and a series of samples is then placed in rotation following the standards. Individual check standard solutions are then placed at the rate of one in every twenty samples. The sampler is operated at the rate of 30 samples per hour; in normal operation this feeds a double manifold for sugar and alcohol but consideration of the alcohol method has been omitted at this stage. An aliquot for sugar determination at 0.32ml per minute is fed into air-segmented water at 6.8 ml per minute; the stream is mixed and re-cycled to the pump. This procedure conveniently dilutes out the residual dissolved gases and reduces the alcohol content of the samples to less than 0.5 per cent.v/v. An aliquot of this dilution is then pumped at 0.23 ml per minute into an air-segmented stream of hydrochloric acid (35 per cent. v/v). The streams are mixed and passed to a standard length delay coil in the heating bath at 95" C , the emergent stream of hydrolysed sugars is cooled, de-bubbled and an aliquot is pumped at 0.32 ml per minute into an air-segmented mixture formed by 0-8 ml per minuteOctober, 19681 OF ORIGINAL GRAVITY OF BEER. PART I 677 sodium hydroxide (2.1 N), 0.8 ml per minute potassium ferricyanide (0.5 per cent.) and air at 0.8 ml per minute. The resultant alkaline mixture is then passed to the second standard delay coil in the heating bath where the reducing sugars are oxidised; the stream is then cooled and re-cycled for the final colorimetric stage.A 0.23 ml per minute aliquot of the 10 Fig. 9. Response peaks of dextrose, molybdophosphate method ferricyanide - ferrocyanide couple is first diluted and acidified by passing it into an air- segmented stream of hydrochloric acid (4.5 per cent.) at 3*9ml per minute. This step is essential to avoid the localised precipitation of the reaction product with molybdophosphate referred to under Molybdophosphate finish. The mixed and diluted ferricyanide - ferro- cyanide then receives an addition of molybdophosphate reagent that is pumped at 04ml per minute. After mixing, the resulting stream is de-bubbled, and a fraction of the liquid stream is pumped through an 8-mm flow cell in the colorimeter, which is equipped with I *5 I -0 0.5 0 0.5 I .o + Difference, [€,=4.335+ 1001 * 21- E Difference between calculated extract and observed extract samples used in preparing regression equations 480-mp narrow band filters.Fig. 9 indicates the normal response peaks obtained with this method, and Fig. 10 indicates the plot of sugar concentration against peak height readings with a x2 range expansion on the recorder. Fig. 10.678 SAWYER AND DIXON AUTOMATIC DETERMINATION [A?ZfZ&St, VOl. 93 RESULTS AND DISCUSSION A study has been made of the correlation between the “extract weight” as determined by the distillation procedure with the 1000 notation for specific gravity, and the reducing sugar value after hydrolysis, expressed as anhydrous glucose.A total of 79 samples of different beers of all types with extract weights varying from 1008-0 to 1039.0 has been used to establish such a correlation having the form- * E - 1000 = 4.33 s + 1.2 where E = extract weight, expressed relative to water (sp.gr. lOOO), and The statistical constants obtained for this line are : regression coefficient = 0.996, standard error of the estimate of the extract weight = 0.62. A histogram plot of calculated extract weight against observed extract weight for the samples used in deriving this equation is shown in Fig. 11. S = percentage of sugar as anhydrous glucose. Glucose, per cent. Fig. 11. Calibration curve, glucose, per cent., A further series of samples was chosen on the basis of the difference between calculated and observed extract values to examine whether the deviations for a particular type of beer were constant and, if so, what significance could be placed on a “label correction.” Values for these observations are indicated in Table I and probability levels of significance TABLE I COMPARISON OF EXTRACT WEIGHT FROM SUGAR DETERMINATIONS WITH THAT OBTAINED BY against full scale deflection, per cent.DISTILLATION OF VARIOUS BEERS Mean extract Mean difference Standard (-=--, between pairs deviation Type of beer samples from sugar distillation observed differences Number of Calculated Observed on calculated - of Milk stout . . .. 6 28.2 28.1 + O * l 0.24 stout . . .. .. 6 26.3 24.8 + 1-at 0.22 stout . . .. .. 6 16.7 16.4 + 0.3 $ 0.10 Continental lager . . 11 18.2 17.7 + o.5t 0.30 Continental lager .. 12 10.8 10.4 + 0.3-r 0.13 Continental lager . . 12 14-3 13.8 + O W 0.2 1 U.K. lager . . . . 12 16-4 14.6 + o w 0.26 Pale ale . . .. .. 12 16.6 16.3 +0.3$ 0.36 Strong ale . . .. 6 27.6 29.7 - 2.1 t 0.37 Brown ale . . .. 6 17.0 16.7 + 0.3 0.46 t Significant at 0.1 per cent. probability level. $ Significant at 1 per cent. probability level. * The equivalent correlation determined by manual hydrolysis and titration with Fehling’s solution was E - 1000 = 4.32 S + 1-9 for 39 samples.October, 19681 OF ORIGINAL GRAVITY OF BEER. PART I 679 obtained by Student’s t-test are indicated. There is evidence of a deviation characteristic of a particular label within the proposed method. In the practical situation when the method is used as a screening test, this value would be weighed against any tolerance allowed within the laboratory for such a screening test.A further examination of the calculated extract from duplicate determinations of sugar for various beer samples showed that for 131 pairs the mean extract difference is 0.27 and the standard deviation for the population is 0.29; this compares with a similar average figure for the standard deviation of differences between calculated extract and true extract. The larger value of standard error of the estimate of extract from sugar values obtained in the regression analysis confirms that the differences between observed and calculated values for specific samples shown in Table I are significant. The development of a method for the determination of “gravity lost” from the alcohol content of the various types of beer forms the basis of a further paper (Part 11), which will also include a comparison of the deter- mination of original gravity by distillation and by a combination of the two methods.The authors wish to thank Miss L. M. Grisley for the manual determinations used in establishing the correlations shown, and Miss B. G. Cox for experimental work on the molybdophosphate method. The Government Chemist is also thanked for his permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. REFERENCES Customs and Excise Act 1952, The Beer Regulations 1952, S.I. 2232, H.M. Stationery Office, Simmonds, C . , “Alcohol : Its Production, Properties, Chemistry and Industrial Applications, ” Cooper, A.H., Hudson, J . R., and MacWilliam, I. C., J . Inst. Brew., 1961, 67, 432. Ashurst, P. R., Ibid., 1963, 69, 457. Hudson, J . R., “Investigational and Routine Automatic Techniques Applied to Brewing,” Technicon Symposium Lecture, London, 1964. Sawyer, R., and Dixon. E. J., i n “Automation in Analytical Chemistry, 1966,” Technicon Symposia, Mediad Inc., New York, 1967, Volume 11, p. 111. Soxhlet, F., J . prakt. Chem., 1880, 21, 227. Lane, J. H., and Eynon, L., J . Soc. Chem. Ind., 1923, 42, 32T. Somogyi, M., J . Biol. Chem., 1952, 195, 19. Nelson, N., Ibid., 1944, 153, 375. Fuller, K. W., in “Automation in Analytical Chemistry, 1965,” Technicon Symposia, Mediad Inc., -, in “Automation in Analytical Chemistry, 1966,” Technicon Symposia, Mediad Inc., New Gentele, A., Chem. ZentBZ., 1861, 1, 91. Hagedorn, H. C., and Jensen, B. N., Biochem. Z., 1923, 135, 46. Hoffman, W. S., J . Biol. Chem., 1937, 120, 51. Folin, O., Ibid., 1928, 77, 421. Hill, J, B., and Kessler, G., J . Lab. Clin. Med., 1961, 57, 970. Ough, L. D., and Lloyd, N. E., Cereal Chem., 1965, 42, 1. Baum, E. H., Ann. N . Y . Acad. Sci., 1960, 87, 894. Fuller, K. W., i n “Automatic Determination of Reducing Sugars, ” Technicon Symposium Fingerhut, B., Ferzola, R., and Marsh, W. H., Clinica Chim. Acta, 1963, 8, 953. Fairbairn, N. J., J . SOC. Chem. Ind., 1953, 86. Pinnegar, M. A., in “Automation in Analytical Chemistry, 1965,” Technicon Symposia, Mediad Barbieri, G. A., Ber. dt. chem. Ges., 1927, 60, 2415. Witekowa, S., Zesz. Nauk. Politech. L d d i , Chem., 1957, 6, 23; Chem. Abstr., 1958, 52, 16115b. Frum, F. S., and Medvedeva, L. P., Uchen. Zap. Gorkov. Gos. Univ., 1958, 32, 139; Chem. Abstr., First received August 14th, 1967 Amended October 16th. 1967 London. Macmillan and Co., London, 1919, p. 473. New York, 1966, p. 78. York, 1967, Volume 11, p. 57. Lecture, London, 1964. Inc., New York, 1966, p. 80. 1960, 54, 5317h.