LIBRARY OF THE UNIVERSITY OF CALIFORNIA. C/ass Digitized by the Internet Archive in 2007 with funding from Microsoft Corporation http://www.archive.org/details/foodanalysisOOIeffrich SELECT METHODS IN FOOD ANALYSIS HENRY LEFFMANN, A.M., M.D., PH.D. AND WILLIAM BEAM, A.M., M.D., F.LC. SE.COND EDITION, REVISED AND ENLARGED •CClftb One Plate an& 54 ©tber f llustratlons PHILADELPHIA P. BLAKISTON'S SON & CO. I0I2 WALNUT STREET 1905 -v' V Copyright, 1905, by P. Blakiston's Son & Co. PRESS OF . FELL COMPANY HILADELPHIA PREFACE TO SECOND EDITION The rapid sale of the first edition of this work, and the favor- able opinions expressed in reviews and correspondence, have encouraged the authors to prepare a second edition, which it is hoped will be worthy of the position attained by the first. The preparation of the second edition has been considerably de- layed, and in the interval much progress has been made in the field. American work is rapidly becoming the leader in food- analysis. The excellent equipment of the laboratories of the Department of Agriculture at Washington, supplemented by more than two-score of State experiment stations, and by hun- dreds of investigators, connected with Boards of Health and Food Commissioners, enables every problem to be submitted to prompt and searching inquiry. We have endeavored to utilize this material fully. It is to be regretted that the publication of these investigations is still unsatisfactory, important results often appearing in bulletins of local circulation and limited editions. It is to be hoped that some system of international publication, easy of access, will be instituted. In the present edition much alteration has been made. Many paragraphs have been cancelled and much new matter inserted. Among the additions are: Detailed descriptions of special arrangements for polarimetry, distillation and extraction; new processes for detection of natural colors used as substitutes for fruit and egg-colors; improvements in detection of formaldehyde, abrastol and saccharin; rapid methods for examination of vanilla and lemon extracts, and for the determination of fat in condensed milk and cereal foods; determination of boric acid in iii:^5S IV PREFACE TO THE SECOND EDITION fruit-juices; analytic data in regard to fruit-juices, jams and jellies; detection of palm oil in oleomargarin, and many minor modifications of tests and processes intended to simplify or ex- pedite analysis. The purpose of the book has not been modified. It is for the practical worker in the detection of food adulteration. No space has been given to discussion of the effects of adulteration, nor to the principles to be observed in the establishment of food-standards, or in framing or administering food-laws. These are not matters for the analyst. The standards pub- lished by the U. S. Government have been included as official interpretations of analytic data. All temperatures are centigrade. Unless otherwise noted, all readings of scale or arc are positive; sulfuric, nitric and hydrochloric acids and ammonium hydroxid are the standard concentrated pure grades of these reagents; alcohol is 95 per cent. Philadelphia, May, 1905. ADDITIONS AND CORRECTIONS Page 64, after line 3, insert "For special methods for detection and determin- ation of aluminum, see pages 378 and 386." Page 79, after line 4, insert "Aluminum oxyacetate is sometimes used as a meat-preservative; see pages 378 and 386," Page 139, line 3, insert after " Hiibl " the reference-figure ^^. Paj^e 140, line 16 from bottom, for ^^ read ^^. Page 349, line 13, for " lo per cent." read " 16 per cent." NOTE SPECIAL PAGE FOLLOWING INDEX CONTENTS ANALYTIC METHODS Physical Data: pace Specific Gravity — Melting and Solidifying Points — Boiling- point — Polarimetry — Spectroscopy — Fluorescence — Micros- copy, 1-26 Chemical Data: Water and Fixed Solids (Extract) — Nitrogen — Crude Fiber — Ash — Extraction with Miscible Solvents — Extraction with Im- miscible Solvents — Distillation and Sublimation — Apparatus and Chemicals, 27-56 APPLIED ANALYSIS General Methods: Poisonous Metals — Colors — Preservatives, 57-86 Special Methods: 3tarch, Flours, and Meals — Bread — Leavening Materials — Sugars — Honey — Candies and Confections, 86-136 Fats and Oils: lodin Number — Volatile Acids — Saponification Value — Acid Value — Solubility in Acetic Acid — Thermal Reac- tion with Sulfuric Acid — Specific Temperature Reaction — Bromin Thermal Value — Elaidin Test — Refractive Index — Soluble and Insoluble Acids — Cholesterol and Analogs — Acetyl Value — Unsaponifiable Matter — Analytic Data — Special Tests, 137-168 Olive Oil — Cottonseed Oil — Maize Oil — Arachis Oil^-Sesame Oil — Rape Oil — Coconut Oil — Cacao-butter — Lard — Butter- fat, 168-189 Milk and Milk Prodticts: Milk — Condensed Milk — Butter — Cheese — Fermented Milk Products, 190-251 Non-alcoholic Beverages: Tea — Coffee — Cacao, 251-282 Condiments and Spices: Vinegar — Pepper — Long Pepper — Cayenne Pepper — Ginger — Nutmeg — Mace — Allspice — Cinna- mon — Cloves — Mustard — Flavoring Extracts — Fruit-products Table Accessories and Desserts — Egg-substitutes, 282-336 V VI CONTENTS Special Methods {Continued): page Alcoholic Beverages: Cider — Spirits — Whiskey — Brandy — Gin — Rum — Malt Liquors — ^Wine — Alcohol Tables — Malt Ex- tracts, 337-3 72 Flesh Foods: Meats — Meat-extracts, 373-385 Appendix. Tables — References, 386-3S8 Index. FOOD ANALYSIS ANALYTIC METHODS PHYSICAL DATA Specific Gravity. In food analysis, determination of specific gravity of solids is rarely made. Fats are usually tested in the melted condition. The following method for solid fats, due to Hager, is suita- ble for small amounts of material : The sample is melted and allowed to drop slowly from the height of aboilt 3 centimeters into some cold alcohol in a dish. The globules thus obtained are placed in diluted alcohol at 15.5°, the strength of which is so adjusted that the globules float in any part of the liquid. The specific gravity of the liquid is then determined; it is, of course, the same as that of the globules. Many substances when cooled suddenly are liable to have abnormal density, hence it is preferable, as noted by Allen, to use fragments cut from a solid mass cooled under normal conditions and allowed to stand at least twenty-four hours. The specific gravity of a liquid is generally expressed by comparison with water. Confusion and inconvenience have arisen from the fact that results have been referred to water at different temperatures as unity. It is becoming customary to express, as is proper, the temperatures of observation and comparison. ^^ indicates a determination at 100° and com- 3 I 2 FOOD ANALYSIS parison with water at 15.5° as unity. It is best to compare the substance and the standard at the same temperature. Pyknometer or Specific- gravity Bottle. — This is an accurate, generally applicable means of determining specific gravity. It is a bottle with a perforated stopper, adjusted to hold a certain weight of water at a standard temperature, usually 15.5°. Bottles as sold are often inaccurate. The weight of water that a bottle holds should be carefully determined. E. R. Squibb devised a convenient form of pyknometer (fig- ure i) which permits the determination to be made at any _ temperature between o and 25°, and compared with water at the same temperature. The bottle should hold 100 grams of recently-boiled distilled water at 20° at about 58 on a scale of o to 100. In weighing the water into the bottle, the fine adjust- ment to o.ooi gram is made by use of narrow strips of blotting-paper that will pass easily down the bore of the graduated stem. When the 100 grams are in the bottle, and the column stands between 50 and 65 divisions of the scale, the stopper is put in, a leaden ring is put on the neck, and the whole im- ■pj^j J mersed in a bath of broken ice and water until the column of water comes to rest. It should then read at zero of the scale, or not much above it, and the read- ing should be noted. If it reads below zero, the bottle is too large, and the stopper part of the stem must be ground farther into the bottle neck, until the reading, on new trial, brings the column a little above zero. The bottle is then put into a bath at 25° and kept there, with stirring of the bath, until the column comes to rest, when it should read somewhere from 90 to TOO of the scale. Should it read above 100, while the lower limit is as far above the zero, the bottle is too small, and the end of the stopper must be ground off until the reading of the column is within the graduations at both ends of the scale. SPECIFIC GRAVITY Sprengel Tube. — This is a form of pyknometer with which a high degree of accuracy is attainable. It is especially suita- ble for determinations at the boiling-point of water. It con- sists (figure 2) essentially of a thin glass U-tube terminating in two capillary ends bent at right angles and each provided with a ground cap. One of these capillary tubes must have a smaller caliber than the other — not larger than 0.25 mm. The larger tube should bear a mark at m. The tube is filled by immersing h in the liquid under examination, connecting "iW w Fig. Fig. 3. the smaller end with a large glass bulb, and applying suction to the latter by means of a rubber tube, as shown in figure 3. If now the rubber tube be closed, the glass tube will fill auto- matically. It is placed in water, the ends being allowed to project, and the water is brought to the proper temperature. A conical flask may be used to contain the water, the ends of the Sprengel tube being supported by the neck. The mouth of the flask should be loosely covered. As the Hquid 4 FOOD ANALYSIS expands it will drop from the larger orifice. When this ceases, the liquid is adjusted to the mark at m. If beyond the point, a little may be extracted by means of a roll of paper. The tube is then taken out of the bath, the caps adjusted, the whole thoroughly dried, allowed to cool, and weighed. The same operation having been performed with distilled water, the calculation of the specific gravity is made as usual. Westphal Balance. — This affords a convenient means of determining specific gravity. It consists of a delicate steel- FlG. 4. yard provided with a counterpoised plummet. The latter, being immersed in the liquid, the equilibrium is restored by means of weights or riders, the value of which is directly ex- pressed in figures for the specific gravity without calculation. Thus, the rider A' is of such a weight as to express the first decimal place, and will be represented by any of the figures from o to 9 according to its position on the beam.. Similarly the riders A, B and C furnish the figures for the second, third and fourth decimal places respectively. The weight A^ used in the case of liquids heavier than water. is SPECIFIC GRAVITY 5 The ordinary form of Westphal balance is untrustworthy, but good instruments are made by some European manu- facturers. The principle of the hydrostatic balance may be applied by using a plummet (that sold with the Westphal balance will serve) with the ordinary analytic balance. Test-tubes weighted with mercury and sealed in the flame may also be used. The plummet is suspended to the hook of the balance Fig. 5. Fig. 6. by means of a fine platinum wire. The specific gravity of any liquid may be determined by noting the loss of weight of the plummet when immersed in the liquid and dividing this by the loss in pure water. If the determination be made at the boiling-point of water, the arrangements shown in Figs. 5 and 6 may be employed. The temperature of the liquid will not usually rise above 99°. This may be done with a hydrometer or balance, if the cylin- 6 FOOD ANALYSIS der containing the oil be kept for a sufficient time in boiling water. With the Sprengel tube high accuracy may be ob- tained. The weight of the Sprengel tube and that of water contained at 15.5° being known, the tube should be com- pletely filled with the oil, by immersing one of the orifices in the liquid and sucking at the other. The tube is placed in a conical flask containing water which is kept actively boiling, a porcelain crucible-cover being placed over the mouth of the flask. The oil expands and drops from the orifices. When this ceases, the oil adhering to the outside is removed by the cautious use of filter-paper, the tube removed, wiped dry, cooled, and weighed. The weight of the contents divided by the weight of water contained at 15.5° will give the specific gravity at the temperature attained compared with water at 15.5°. When the amount of material is sufficient, the deter- mination may be made by use of the plummet, employing a cylindrical bath with two orifices. One of these is fitted with an upright tube for conveying the steam away from the neigh- borhood of the balance; into the other a test-tube, 15 cm. in length and 2.5 cm. in diameter, fits tightly, the joint being made perfect by cork or india-rubber. The test-tube is filled with the substance to be tested, and the plummet immersed in it. The water in the outer vessel is then kept in constant ebullition, until a thermometer, with which the oil is repeat- edly stirred, indicates a constant temperature, when the plum- met is attached to the lever of the balance, and counterpoised. For temperatures higher than 100° glycerol or paraffin may be used, but considerable care is required in such cases. Hydrometers are much used for the determination of the specific gravity of liquids, but the indications are less reliable than by the methods described above. The instruments as furnished are often not accurately graduated, and the zero point, at least, should be verified by immersing in distilled water at a standard temperature. Sensitive hydrometers with MELTING AND SOLIDIFYING POINTS 7 slender stems, accurately graduated, are now obtainable. These are capable of furnishing good results. Care should be taken to make the reading at the top, center or bottom of the meniscus according to the method used in the grad- uation of the instrument. Instruments intended for use with opaque liquids should be graduated to be read at the top of the meniscus. The actual specific gravity of any substance is the ratio of its density at a given temperature to that of water at the same temperature. Statements made upon any other basis than this may be converted into actual specific gravity by calculation from the table of density of water given in the appendix. Thus, a determination of specific gravity of 0.8000 at 7^ may be converted into actual specific gravity (^) as follows: Density of water at 15° = 0.99916. 100° = 0.95866. too" 100° is" 100° Therefore, 95866 : 99916 : : 0.8000 : 0.8337 (actual specific gravity at 100°). Melting and Solidifying Points. The determination of these is often difficult. Many sub- stances, especially fats, assume conditions exhibiting abnormal melting-points, and also frequently solidify at a temperature very different from that at which they melt. If, in the prep- aration of any substance for determining its melting-point, it is necessary to make a previous fusion, the mass should be allowed to rest not less than twenty-four hours after solidifi- cation before making the experiment. Chemists disagree as to whether the melting-point should be considered to be that at which the substance begins to be liquid or that at which the liquid is perfectly clear. Ordinary thermometers are frequently inaccurate, the error amounting to a degree or more. No ob- servations in which precision is required should be made with unverified instruments. 8 FOOD ANALYSIS The following method for determining melting-points is suitable for many technical purposes. By substituting strong brine or glycerol for the water in the bath observations may be made at temperatures beyond the limits of o° and ioo°: The substance is heated to a temperature slightly above its Fig. 7. Fig. 8. fusing-point, drawn into a very narrow glass tube, and allowed to solidify for not less than twenty-four hours. The tube, open at both ends, is attached by a wire or rubber ring to a. thermom- eter so that the part containing the substance is close to the bulb. The apparatus, immersed in water, is heated at a rate not exceeding 0.5° per minute until fusion takes place, when MELTING AND SOLIDIFYING POINTS 9 the temperature is noted. The temperature is allowed to fall and the point at which the substance becomes solid is also observed. To insure uniform and gradual heating, it is neces- sary to immerse the vessel containing the thermometer and tube in another larger vessel filled with water. Allen sug- gests a flask of which the neck has been cut off, as shown in figure 7. A neater form of apparatus is shown in figure 8, from "Richter's Organic Chemistry." The two following methods are especially adapted to the examination of fats and waxes. The A. O. A. C. method dis- regards the abnormal condition of recently-solidified masses : A. O. A. C. Method. — A mixture of alcohol and water of the same specific gravity as the sample is prepared in the fol- lowing manner: Separate portions of distilled water and 95 per cent, alcohol are boiled for 10 minutes. The water is poured, while still hot, into the test-tube described below until it is nearly half full. The test-tube is nearly filled with the hot alcohol, which is carefully poured down the side of the inclined tube to avoid too much mixing. If the alcohol is added when water is cold, the mixture will contain air- bubbles and be unfit for use. The apparatus (Fig. 9) consists of: A thermometer reading easily and accurately to tenths of a degree; a cathetometer for reading the thermometer (this may be substituted by an eyeglass if held steadily and properly adjusted); a thermom- eter; a tall beaker 35 cm. high and 10 cm. in diameter; a test-tube 30 cm. long and 3.5 cm. in diameter; a stand for supporting the apparatus; some method of stirring the water in the beaker (for example, a rubber blowing-bulb and a glass tube extending to near the bottom of the beaker). The melted and filtered fat is allowed to fall from a drop- ping-tube from a height of from 15 to 20 cm. on a smooth piece of ice floating in recently-boiled distilled water. Disks from I to 1.5 cm. in diameter, and weighing about 200 mg., 10 FOOD ANALYSIS are formed. Pressing the ice under the water the disks float on the surface, and are easily removed with a steel spatula, cooled in the ice-water before using. The test-tube contain- FiG. 9. ing the alcohol and water is placed in a tall beaker containing water and ice, until cold. The disk of fat is then dropped into the tube from the spatula and at once sinks to the part of the tube where the density of the diluted alcohol is exactly MELTING AND SOLIDIFYING POINTS II equivalent to its own. The delicate thermometer is placed in the test-tube and lowered until the bulb is just above the disk. In order to secure an even temperature in all parts of the alcohol mixture in the vicinity of the disk, the thermom- eter is used as a stirrer. The disk having been placed in posi- tion, the water in the beaker is slowly heated and kept con- stantly stirred by means of the blowing apparatus already described. When the temperature of the alcohol-water mix- ture rises to about 6° below the melting-point, the disk of fat begins to shrivel and gradually rolls up into an irregular mass. The thermometer is lowered until the fat particle is even with the center of the bulb. The bulb of the thermom- eter should be small, so as to indicate only the temperature of the mixture near the fat. A gentle rotatory movement should be given to the thermometer bulb. The rise of tem- perature should be so regulated that the last 2° of increment require about ten minutes. The mass of fat gradually ap- proaches the form of a sphere, and when it is sensibly so the reading of the thermometer is taken. As soon as the tem- perature is taken the test-tube is removed from the bath and placed again in the cooler. A second tube, containing alcohol and water, is at once placed in the bath. The test-tube (ice- water having been used as a cooler) is of low enough tem- perature to cool the bath sufficiently. After the first deter- mination, which should be only a trial, the temperature of the bath should be so regulated as to reach a maximum of about 1.5° above the melting-point of the fat under examina- tion. If the edge of the disk touches the sides of the tube a new trial should be made. Second and third results should show a near agreement. TiTER-TEST. — To eliminate error in determining melting- points of intimate mixtures, such as commercial fats and waxes, the titer-test, proposed by Dalican, has been largely adopted. 100 grams of the fat are saponified, the fatty acids separated 12 FOOD ANALYSIS by addition of acid, freed from water, filtered into a porcelain dish, and allowed to solidify overnight under a desiccator. The mass is then carefully melted in an air-bath and sufficient poured into a test-tube i6 cm. long and 3.5 cm. in diameter to fill the tube a little more than halfrfuU. The tube is then placed in a suitable flask, say of 2000 c.c. capacity, and a deli- cate thermometer, indicating one-fifth of a degree, inserted so that the bulb reaches the center of the mass. When a few crystals appear at the bottom of the tube, the mass is stirred by giving the thermometer a rotatory movement, first three times from right to left, then three times from left to right, and then continuously, by a quick circular move- ment of the thermometer, without allowing it to touch the side of the vessel, but taking care that all solidifying portions, as they form, are well stirred in. The liquid will gradually become cloudy throughout, and the thermometer must be observed carefully. At first the temperature will fall, but will soon rise suddenly a few tenths of a degree and reach a maximum at which it remains stationary for a short time before it falls again. This point is called the ''titer" or sohdifying point. Boiling-point. Fig. 10. For the determination of boiling-point the ap- paratus of Berthelot is convenient. Figure 10, from Traube's '' Physico- Chemical Methods," shows the con- struction. The thermometer is inclosed in an outer tube, so that the portion of the scale to which the mercury rises is immersed in the vapor. If this be not done, a correction must be applied for the error produced by the cooling of the thermometer tube. The bulb of the thermometer does not reach into the liquid. A few fragments of pumice-stone or POLARIMETRY 1 3 broken clay pipestems will prevent bumping. The exit-tube at the lower end of the wide tube connects with a condenser. The barometric pressure must always be noted and correction made for the variation from the standard pressure, 760 mm., by the following formula: B = B^ + 0.0375 (760 — ^P); in which B is the boiling-point at normal pressure, B^ the observed boiling-point, P the observed pressure in millimeters. For an apparatus designed for special boiling-point observa- tions see under ''AlcohoHc Beverages." Polarimetry. Polarimeters are instruments used to measure the extent and direction of the rotation of the plane of polarized light. They consist essentially of a Nicol's prism as polarizer, a tube carrying the substance to be tested, and a second Nicol's prism, or analyzer, by which the extent of rotation is meas- ured. In all forms some condition of the field of vision is fixed upon as the zero point, and the rotation of the analyzer or other manipulation necessary to restore this standard field affords the measurement of the rotation caused by the inter- posed substance. Several types of instrument have been de- vised, of which two are most important. In one form, de- vised by Soleil, white light is used and a colored field, known as the transition tint, is taken as the zero point. In the other type white light or monochromatic (yellow) light is used and the zero point determined by equahzing the brightness of the field. Instruments of the first form are unsatisfactory by reason of the difference in susceptibiUty in the eyes of differ- ent person to color-contrasts. The instruments of the second type, commonly designated shadow instruments (more cor- rectly, ''penumbral"), are now more generally employed. In the Laurent apparatus, shown in figure 11, the mono- 14 FOOD ANALYSIS chromatic light passes through the collimating lens A and is polarized by the Nicol's prism B, which is so placed that it may be moved, on its axis, over a small arc by means of the lever C and clamped at any point; by this the brightness of the field may be varied and the sensitiveness of the instrument in- creased or diminished as may be needed. The polarized beam then passes through a quartz plate of even thickness, cut ex- actly parallel to the optic axis, and placed so that it covers a Fig. II. semicircle of the field. At the other end of the apparatus is the analyzing prism E and the eye-piece F fixed to a graduated disk. This combination can be rotated upon its axis in a com- plete circle. Attached arms carry viev^-lenses for reading the angle of rotation, and the instrument is set at zero by an in- dependent adjustment by which the analyzing prism is rotated without disturbing the position of the graduated disk. Ver- niers are provided for close measurement. The monochro- POLARIMETRY 1 5 matic light must be obtained from a sodium flame, since the thickness of the quartz plate is adjusted to these rays. In use, the tube is filled with water, the instrument directed to the source of light, and the adjusting milled head turned until the disk is set at zero. The two portions of the field should now appear equally illuminated. If this is not the case, the position of the analyzer must be altered by means of the independent adjustment, the index remaining undisturbed at the zero point. The tube is filled with the liquid to be tested and again placed in the instrument. If optically active, the plane of the polarized light will be rotated and one-half of the field of observation will appear darker. The extent of rotation, which will depend upon the nature of the substance and its amount, is measured by rotating the analyzer to the right or left, as the case may be, until the halves of the field become equally illuminated. This instrument can be employed to measure the rotatory power of all classes of substances, but other forms give ac- curate indications only with substances which have the same dispersive power as quartz, unless monochromatic light be used. In the Schmidt and Hansch penumbral instrument, the division of the field is obtained by a special construction of the polarizing prism and the restoration is accomplished by the adjustment of compensating quartz-wedges constructed so as to produce in the zero position no rotation. When an optically active substance is interposed in the path of the ray, one of the quartz-wedges must be moved to an extent suffi- cient to overcome this rotation in order to retore the stan- dard field. The effect is dependent upon the fact that by this movement the thickness of the quartz is increased or dimin- ished until it compensates for the rotation produced by the solution. The extent of movement of the quartz is registered upon a linear scale, which is read by means of a lens and ver- i6 FOOD ANALYSIS nier. White light is employed in making the observations. A form of the Laurent instrument, with quartz-wedge com- pensation, and employing white light, is made. An instru- ment has been devised in which the field is divided vertically into three zones, the central one being a broad band. Dupli- FlG. 12. cate Nicol prisms are so arranged that the lateral zones agree in tint, thus making stronger contrast with the central zone. The polarimeter shown in figure 12 is now. the standard instrument. It has been improved lately by the substitution of a heavy iron stand for the rickety tripod, but is still in- complete. It has two serious defects. The illumination of POLARIMETRY 1 7 the scale is awkward, and it is not convenient for examina- tions at temperatures above normal. The illumination of the scale is done by a mirror over the eye-lens which receives light from the main lamp. This in- terferes with the eye reaching its highest sensitiveness. In the laboratory of one of us (L) the following arrangement has been adopted. The polarimeter is in the balance-room, close to a small opening in the board partition, on the other side of which is the source of light. In daylight work the scale can be read without special light, but if greater sensitive- ness of the eye is needed a focussing cloth is thrown over the instrument and operator, and the scale is illuminated by a small incandescent lamp. FJ operated by two dry cells. The lamp is inserted just under the mirror that re- r- flects the scale and is con- I trolled by a make-circuit \_ key as usual. For examinations at tem- 5 peratures above normal, i- Leach employs a double Fig. 13. metal tube, similar to the ordinary condenser, the inner channel being heavily gilded to prevent corrosion by acid liquids. Arrangements must be made for taking temperatures during observation and for expansion and contraction of the liquid in the inner tube when this is closed by the glass fronts. For taking temperature, it is usual to provide a tube at the center, connecting with the annulus, in which a thermometer is inserted. For expansion, Cochran pro- vides a short tube at one end, communicating with the inner tube. Figure 13 is a sketch of a form designed by one of us (L) in which the expansion and temperature tubes are combined. It is made of brass. The inner tube is 197 mm. long. This 3 1 8 FOOD ANALYSIS allows the standard length of 200 mm. to be obtained by washers, against which the glass circles rest. These are held in place by caps, which screw into the solid end-pieces. The inner tube and the surface on which the washers rest should be well gilded. The joints need not be brazed as the tem- perature will never be near that of the melting-point of soft solder. At each end, somewhat above the middle horizontal line and communicating with the annulus, is a short tube about 0.7 cm. in diameter. These are for attachment of rubber tubes carrying water. By placing them above the middle Hne, the tube will He properly in the trough of the instrument. In the middle is a tube 3 cm. high, of the same diameter as the inner tube and communicating with it. It must be in such direction as to be upright when the tube is in position in the instrument. This tube is for expansion and holds the thermometer, which is set down as far as possible without interfering with the observation. The thermometer should be about 20 cm. long, with a scale from 0° to 100°. It is easily fastened by slipping a short piece of rubber tube over it, and over the brass tube. Holes can be cut in a focussing cloth so that the instrument and operator can be in darkness, the scale being read by means of the electric lamp as noted above. A metal vessel holding several liters is provided with heating arrangements, a rubber tube leads from it to one of the water- tubes, and an exit is provided through the other. The water in the vessel is allowed to flow through the observation tube at such a rate as will maintain the proper temperature in it. As many of these examinations are for differential temper- ature readings, it will often be unnecessary to connect up the hot- water apparatus. The observation tube should be closed with corks, the annulus filled with hot water, all its openings similarly closed, and then placed in water at a suitable tempera- ture for at least five minutes. It is removed, wiped dry, the glass fronts fastened in the usual way, and the liquid to be POLARIMETRY I9 examined run in through the thermometer opening. It will be easy to do this without retaining air-bubbles. The thermometer is fastened by the short rubber tube, allowed a few minutes to reach the temperature of the inner liquid, the apparatus placed in the polarimeter and the reading quickly taken. It may be wrapped in some non-conducting material while waiting for the thermometer to reach its highest point. Observation with hot tubes should be made quickly ; if a number are to be made, an interval of a few minutes should be allowed to intervene between each, during which the polarimeter trough should be opened. The dehcate optical train may be injured by much heating. Sources oj Light. — For white light, oil, gas, or electric lamps are employed, of which numerous patterns are furnished. Sat- isfactory results may be obtained by the Welsbach lamp. Wiley recommends the use of the acetylene flame, especially for deeply colored solutions. For monochromatic light, the lamp usually employed is a Bunsen burner with a ledge at the top for holding some solid. sodium compound. A fused mixture of sodium chlorid and phosphate is better than sodium chlorid alone. The fol- lowing is an excellent method for obtaining a steady, strong, yellow light: Strips of common filter-paper 5 cm. wide and about 50 cm. long are soaked in a strong solution of sodium chlorid and thiosulfate, dried, and rolled into a hollow cylin- der of such size as to fit firmly on the top of the Bunsen burner. The cylinder is kept from unrolling by a few turns of fine iron wire. The flame bums at the top of the cylinder, giving for the first few minutes a luminous cone, but soon becoming pure yellow. The cylinder becomes a friable charred mass, but if not disturbed may be used for some time continuously or at intervals. Specific Rotatory Power. — The specific rotatory power of a substance is the amount of rotation, in angular degrees, 20 FOOD ANALYSIS produced by a solution containing one gram of the substance in I c.c. examined in a column one decimeter long. It is usually represented by the symbol [«]. To indicate the light employed in the observation, [«]d or [a]j is used, d stands for light of wave length corresponding to the D line of the solar spectrum (sodium flame) and j (jaune) for the transition tint. It is usual also to indicate in the same symbol the temperature of observation; thus, [aY°. Under ordinary methods of observation the specific rota- tory power is represented by the following formula: W^ = ^; in which [a]^ is the specific rotatory power for the light of the sodium flame, a is the angular rotation observed, c is the concentration expressed in grams per loo c.c. of liquid, / is the length of the tube in decimeters. Comparison oj Scales oj Various Instruments. — Polarimeters are now usually provided with a scale reading to loo when a certain quantity of sucrose, called the normal weight, is dissolved in water and made up to loo c.c. For the German instruments, which are largely used in the United States, this is 26.048 grams. This scale is known as " Ventzke," "Schmidt and Hansch," and "sugar" scale. The instruments made by Schmidt and Hansch are gradu- ated to read correct percentages when the normal weight of sugar is contained in 100 Mohr's cubic centimeters and ob- served in a 2 decimeters tube at 17.5°. With the Laurent apparatus the normal weight of the sugar should be contained in 100 true cubic centimeters. The volume of 100 Mohr's cubic centimeters . is that of 100 grams of water at 17.5° weighed in air with brass weights; it is equal to 100.234 true cubic centimeters. For the nor- mal weight of 26.048 grams in 100 Mohr's cubic centimeters SPECTROSCOPY 21 of solution, may be substituted 25.9872 grams in 100 true cubic centimeters at 17.5°. At the session of the International Commission for Uniform Methods of Sugar Analysis held at Paris, July 24, 1900, it was agreed that the normal weight shall be fixed at 26 grams in 100 true c.c. at 20°, weighed in air with brass weights (see under "Sucrose"). The following factors may be employed for the conversion of data obtained by different instruments: I division Schmidt and Hansch 0,3468° angular rotation D. 1° angular rotation D 2 .8835 divisions Schmidt and Hansch. 1° angular rotation D 0.75 1 1 division Wild. I division Laurent 0.2167° angular rotation D. 1° angular rotation D 4-6154 divisions Laurent. Correction jor Precipitate.— In some cases the volume of precipitate produced by the clarifying agents is considerable, and a correction would be necessary. The error may be eliminated by Scheibler's method: A normal weight of the sample is dissolved in water or proper solvent, treated with the clarifying agent, the liquid made up to 100 c.c, shaken well, filtered, and a reading taken of the filtrate. A second portion of normal weight is treated in the same way except that it is made up to 200 c.c. before filtration. Great care must be taken in the readings. The true reading is obtained by dividing the product of the two readings by their difference. Spectroscopy. In practical analysis the spectroscope is mostly useful in detecting some of the rarer elements in ashes and water-resi- dues. For this purpose the direct vision instrument shown in figure 14 is sufficient. It will often serve for the examina- tion of absorption bands, but for precise research in distinguish- ing colors and specific absorptions a more elaborate instru- ment, as shown in figure 15, will be needed. Zeiss makes a 22 FOOD ANALYSIS direct vision instrument in which the Hght enters by openings placed side by side, but forms spectra that are exactly super- posed. By this means a solution of known composition can be ex- amined in comparison with a material to be tested; or two flame-tests may be compared. This instrument can be mounted as shown in figure 14. For the examination of ashes or water-residues, the material is mixed with a few drops of hydro- chloric acid, a portion of the mass taken up on a loop of clean plati- num wire and held in a non-lumin- ous flame, the spectrum being ex- amined through the instrument. It is important that the first effects should be noted, as some sub- stances volatilize quickly. The Fig. 14. platinum wire should be cleaned by dipping it in a little pure Fig. 15. MICROSCOPY 23 hydrochloric acid and heating it in the gas flame until it im- parts no color thereto. For the observation of absorption-bands of liquids, small flat bottles with ground and polished sides are used. These permit the observation of a thin or thick stratum as desired. Deeply colored solutions should not be used since large por- tions of the spectrum may be cut out by general absorption and the distinctive selective absorption be lost. For some purposes the microspectroscope will be needed, but its use is practically limited to medico-legal work. Fluorescence. This may be detected satisfactorily in the manner described by Allen: A test-tube or cylindrical beaker is nearly filled with a perfectly clear solution of the substance, set upon a dark surface, and observed from above. Another plan is to make a streak of the liquid on a piece of black glass or pol- ished black marble and examine this in a good white light. Tests can also be made by directing a ray of white light from any source through the side of a beaker containing the liquid and looking at it from above. In all the methods the liquid must be perfectly clear or misleading reflection-effects are pro- duced. Microscopy. For preliminary examination of food samples a hand lens is useful, but the practical analysis involves the use of the compound microscope. A good instrument can now be ob- tained at comparatively small cost. It should be supplied with at least two objectives, one of low power, about 16 mm. focus (§ in.), and one of rather high power, 4 mm. focus (J in.). The usefulness of a microscope is much enhanced by the attachment of a sub-stage achromatic condenser and ad- justable diaphragm. Polarizing apparatus, including a selenite plate, is needed, especially for differentiation of starches. 24 FOOD ANALYSIS The instrument shown in figure i6, of American construc- tion, is arranged to receive all accessories. A double nose- piece will be sufficient, as the high-power lens which is shown Fig. I 6. is not needed for chemical work. The outfit, with two lenses and polarizing attachment with selenite, costs about $70. For the better differentiation of objects submitted to ex- amination under the microscope, clearing and staining agents are used. In many cases details of structure are brought out MICROSCOPY 25 sharply by using a dense liquid as a mounting fluid. The following is a list of the important apparatus and reagents: Slides and cover- glasses. Agate mortar^ 2.5 cm. outside diameter, and a somewhat larger glass triturating mortar are useful for preparing mate- rials. The pestles of agate mortars are usually inconveniently short, and are much improved by being mounted in a wooden handle. Dissecting needles are easily made by sawing off the metal portion of an ordinary penholder close to the wood and for- cing the eye-end of a sewing needle under the ferrule which has been thus formed. A neat form of a needle-holder is furnished by the instrument makers. Small forceps and sharp scissors will be needed. Watch-glasses are used for immersing specimens in liquids; still better are the so-called Syracuse glasses, the best form of which has a ground-glass surface for memoranda. Water. Distilled water is best, but any clear, colorless water not containing much mineral or organic matter will answer. Glycerol. A pure article is easily obtained. Alcohol. The commercial 95 per cent, form is used for hardening tissues, but for ordinary microscopic work, a 70 per cent, solution will suffice. Methyl alcohol in the purified form now obtainable may be substituted in many instances for common alcohol. Ether^ chlorojorm, benzene, and carbon disulfid are occasion- ally used for their solvent action, especially to remove oils, waxes, and resins. Carbon tetrachlorid will be also of use. For these extractions it will often be most satisfactory to ope- rate in a small continuous extraction apparatus, with repeated washings, as described under ''Extraction," drying the material at a gentle heat to remove all the solvent, which would inter- fere with the action of watery solutions or glycerol. 4 26 FOOD ANALYSIS Chloral hydrate solution^ — a saturated solution in water. Chloral hydrate and iodin solution, — a portion of the above solution to which a trace of iodin has been added. Potassium iodid and iodin solution, — potas- sium iodid, 0.4 gram; iodin, o.i gram; water, 20 c.c. Zinc chloriodid and iodin solution: Dissolve 5 grams of zinc chlorid and 1.6 grams of potassium iodid in 17 c.c. of water and saturate with iodin. Sodium hydroxid, — 5 per cent, solution. In some instances a strong solution is employed, which is best prepared when required. Acid phloroglucol. This is best prepared when needed by dissolving a few milligrams of phloroglucol in i c.c. of alcohol and adding a drop of hydrochloric acid. Bottles (figure 17) with caps ground on and pipet, are the best for reagents. A little vaselin may be put on the joint to prevent sticking. Fig. 17. WATER AND FIXED SOLIDS 27 CHEMICAL DATA Water and Fixed Solids (Extract). Water is usually determined with sufficient accuracy, pro- vided other volatile bodies are not present, by heating the material (solids should be finely divided) in a flat dish on the water-bath or in the water-oven until it ceases to lose weight. The residue constitutes the fixed solids or extract. Flat platinum dishes from 4 to 8 cm. in diameter and 0.5 cm. high are well adapted to this work. They should rest on porcelain or asbestos rings. Nickel dishes are often applicable, especially the broad shallow crucible covers made in dish form. Dishes of glass — especially the shallow (Petri) dishes used for microbe culture — and porcelain are suitable; aluminum and tin less so. In many cases drying will be facilitated by using an absorbent material such as pure quartz sand, powdered asbestos, or pumice-stone. These materials should be extracted with dilute hydrochloric acid, well washed, and well dried before use. The quantity used should be rapidly weighed, preferably in the dish in which the operation is to be carried out. It is advisable to cover the dish with a nearly flat, thin watch-glass in all the weighings. By a few trials a glass can be selected which fits fairly close to the rim of the dish and restricts evaporation or absorption of water. It is often convenient to weigh a small stirring-rod with the dish and absorbent. In many cases liquid can be measured directly into the dish, the residue being recorded in grams per 100 c.c. or other suitable ratio. Sirupy and gelatinous liquids or those containing much solid matter, especially if this be somewhat difficult to dry, may often be more satisfactorily treated by diluting a weighed portion with several times its weight of water, evaporating a measured or weighed amount of the dilute liquid, and calcu- lating the amount of residue in the original substance. 28 FOOD ANALYSIS The ordinary water-bath and water-oven need no descrip- tion. The temperature of materials heated on the former is usually much less than ioo°; in the latter, slightly below ioo°. By using strong brine a somewhat higher temperature may be obtained. In the case of very hygroscopic or easily Fig. 1 8. decomposable bodies it may be necessary to dry in a current of hydrogen or at reduced pressure. Figure i8 shows a drying oven for use with a current of hydrogen. The apparatus was designed by Caldwell for determining moisture, ether-extract, and crude fiber as pre- scribed by the A. O. A. C, the three data being determined on the same sample. WATER AND FIXED SOLIDS ^9 The bath is made of copper and is 24 cm. long, 15 high, and 8.5 broad. It stands in a piece of sheet-copper bent at right angles along the sides, as shown in the end view; on one side this vertical part need not be over i cm. high, just enough to project a little up the side of the bath, which rests snugly against it; along the other side it projects upward, at a little distance from the side of the bath, about 15 mm., and to about the height of 4 cm.; opposite each of the tubes of the bath a slot is cut in this vertical part, which serves then as a shoulder against which the glass tube rests when in place, to keep it from slipping down and out of position. The tube for containing the substance has at the zone a three small projections on the inner surface, which support a perforated platinum disk of rather heavy platinum foil carry- ing the asbestos filter. This tube is 13 cm. long and 23 mm. inner diameter, and weighs, with its closed stoppers, about 30 grams. The filter is readily made in the same manner as the Gooch filter, the tube being first fitted to a suction flask by an en- largement of one of the holes of the rubber cork, or, better still, by shpping a short piece of rubber tube over it, of such thickness that it will fit tightly in the mouth of a suction flask provided with lateral tube for connection with the suction. A thin welt of asbestos is sufficient; if it is too thick, the gas and ether will not flow through readily. About 2 grams of the substance are put in this tube, pre- viously weighed with the stoppers h and c, and the weight of the substance accurately determined by weighing tube and contents. The stoppers are removed, a band of thin asbestos paper is wound around the end d of the tube, a little behind the slight shoulder at the rim, as many times as may be neces- sary to make a snug fit, when this tube is slid down into the copper tube in the bath, thus preventing circulation of air between the glass and the copper tubes that would retard the 30 FOOD ANALYSIS heating of the former; the stopper e is put in the lower end of the tube for connection with the hydrogen supply, and the stopper / in the upper end; this latter stopper is connected by rubber tube with a glass tube slipping easily through one of the holes of a rubber cork closing a small flask containing a little sulfuric acid, into which this tube just dips; when as many tubes as are to be charged are thus arranged in place and the hydrogen is turned on, the even flow of the current through the whole number is secured by raising or lowering a very little the several tubes through which the outflow passes, so as to get a little more back pressure for one, or a little less for another, as may be found necessary. When the drying is supposed to be completed, the tubes are weighed again with their closed stoppers, and so on. For ether-extraction the unstoppered tube with contents is put directly into the extractor. Carr and Osborne have made an extended series of inves- tigations as to the determination of water, and find that more accurate results may be obtained if the operation be conducted under a diminished pressure at a temperature not exceeding 70° C. Under these conditions it was found possible to dehy- drate levulose completely, without decomposition. The oven is made of a section of metal tubing, from 15 to 20 cm. in diameter and 30 to 40 cm. long. One end is closed air-tight by a brass end-piece, brazed or attached by a screw. The other end is detachable and is made air-tight by ground surfaces and a soft washer. On the top are apertures for the insertion of a vacuum-gauge and for attachment to a vacuum-apparatus, thermostat and thermometer. The aperture for admission of air or hydrogen is best placed at the fixed end. The oven may be heated by a single burner, but a series of small jets is prefer- able. The metal should be protected by sheet asbestos. The temperature of the oven can be kept uniform by a gas regulator, or by attention to the lamp. WATER AND FIXED SOLIDS 3 1 The method of operating is as follows: Clean pumice-stone of two grades of fineness is used, one that just passes through a I mm. mesh and one that passes through a 6 mm. mesh. These are digested with hot 2 per cent, sulfuric acid, washed by decantation until the wash-water is free from acid, placed, wet, in a sand crucible and heated to redness. When the water is expelled, the material may either be placed hot into a desiccator or directly into the drying dishes. In loading the dishes, place a thin layer of dust over the bottom of the dish to prevent the material to be dried from coming in contact with the metal; over this layer place the larger particles, nearly filling the dish. If the stone has been well washed, no harm may result from placing the dish and stone over the flame for a moment before transferring to the desiccator preparatory to weighing. If the material to be dried is dense, it is diluted until the specific gravity is in the neighborhood of 1.08 by dissolving a weighed quantity in a weighed quantity of water. (Alcohol may be substituted in material not precipitable thereby.) Of this,' 2 to 3 grams may be distributed over the stone in a dish the area of which is in the neighborhood of 20 sq. cm., or one gram for each 7 sq. cm. of area. The material is distributed uniformly over the pumice by means of a pipet weighing- bottle (weighing direct upon pumice will not answer), ascer- taining the weight taken by difference. The dishes are placed in the vacuum-oven, which should be maintained at a pressure of not more than 125 mm. of mer- cury. The temperature must not exceed about 70°. All weighings must be taken with the dish covered by a close-fitting plate. The open dish must not be exposed to the air longer than absolutely necessary. Weighings may be made at inter- vals of two or three hours. In the laboratory of the United States Geological Survey a sheet-iron or nickel basin about 10 cm. in diameter and 3 32 FOOD ANALYSIS cm. deep is set upon an iron plate which is heated directly by the burner. A platinum or pipe-clay triangle rests in the basin and supports the dish containing the Hquid to be evap- orated. It is stated that almost any liquid can be evaporated in this way without sputtering. The temperature, however, is liable to be too high for many organic bodies. Parsons has obtained good results in the drying of sensitive organic substances by the following method: A perfectly neutral petroleum oil, free from animal or vegetable oils and mineral substances, sp. gr. 0.920, flash test 224°, fire test 260°, boiHng-point about 288°, is heated to about 1 20° for some time and preserved in a well-stoppered vessel. A quantity of oil about six times that of the weight of the substance to be dried is heated in an evaporating dish in a drying oven to a tempera- ture of 1 1 5°, and then weighed. The weighed portion of the sub- stance is put into the oil; if it be very moist, it is added in small portions. Slight effervescence will usually occur, and the mass should be kept in the drying oven for a short time after effer- vescence has ceased. The evaporating dish containing the oil and substance is weighed; the loss is moisture. The whole operation may be completed in less than half an hour. Nitrogen. Total Nitrogen. — The Kjeldahl- Gunning method is the most satisfactory. The reagents and operation are as follows : Potassium Sulfate. A coarsely powdered form free from nitrates and chlorids should be selected. Suljuric Acid. This should have a sp. gr. 1.84 and be free from nitrates and ammonium. Standard Acid. -^ Sulfuric or hydrochloric acid, the strength of which has been accurately determined. Standard A Ikali. -^ Ammonium hydroxid, sodium hydroxid. NITROGEN 33 or barium hydroxid, the strength of which in relation to the standard acid must be accurately determined. Strong Sodium Hydroxid Solution. 500 grams should be added to 500 c.c. of water, the mixture allowed to stand until the undissolved matter settles, the clear liquor decanted and kept in a stoppered bottle. It will be an advantage to de- termine approximately the quantity of this solution required to neutralize 20 c.c. of the strong sulfuric acid. Indicator. Cochineal solution is recommended by the A. O. A. C, but methyl-orange and azolitmin are satisfactory. Phenolphthalein is not well adapted to titration of ammonium compounds. (See under ''Indicators.") Digestion Flasks. Pear-shaped round-bottomed flasks of hard, moderately thick, well-annealed glass, about 22 cm. long, maximum diameter of 6 cm., tapering gradually to a long neck, 2 cm. in diameter at the narrowest part, and shghtly flared at the mouth. Distillation Flasks. Jena-glass flasks of about 550 c.c. ca- pacity. A copper flask, such as sometimes used in the manu- facture of oxygen, may be substituted. Combined Digestion and Distillation Flasks. Jena-glass round-bottomed flasks with a bulb 12.5 cm. long and 9 cm. in diameter, the neck cylindrical, 15 cm. long and 3 cm. in di- ameter, flared slightly at the mouth. Process. 0.7 to 3.5 grams, according to the proportion of nitrogen, are placed in a digestion flask. Then 10 grams of powdered potassium sulfate and 15 to 25 c.c. (ordinarily about 20 c.c.) of the strong sulfuric acid are added and the digestion conducted as follows: The flask is placed in an inclined posi- tion and heated below the boiling-point of the acid for from five to fifteen minutes, or until frothing has ceased. Excessive frothing may be prevented by the addition of a small piece of paraffin. The heat is raised until the acid boils briskly. A small, short-stemmed funnel may be placed in the mouth 34 FOOD ANALYSIS of the flask to restrict the circulation of air. No further atten- tion is required until the liquid has become clear and colorless, or not deeper than a pale straw. When Kjeldahl operations are carried out in limited number, the arrangement used in the laboratory of one of us (L) has been found very satisfactory. A double-Y, terra cotta drain-pipe, about 20 centimeters internal diameter, is connected by an elbow directly with the chimney-stack. The digestion flasks are supported as shown in the rough sketch, figure 20 (not drawn exactly to scale). Two flasks can be operated at once. The Fig. 19. Fig. central opening is convenient for other operations producing fumes. Openings not in use are closed by circles of heavy asbestos. The apparatus shown in figure 19 is used when many de- terminations are made. As corrosive vapors are given off, it must be placed under a hood. The central opening in the ventilating pipe shown in figure 20 will be satisfactory; the mouths of the flasks should be well inside the margin of the pipe. When the liquid has become colorless or very light straw NITROGEN 35 yellow, it is allowed to cool, diluted with icx) c.c. of water if the smaller form of flask has been used, the liquid transferred to the distilling flask, and the digestion flask rinsed with two portions of water, 50 c.c. each, which are also transferred to the distilling flask. With the larger form of flask the dilution is made at once by the cautious addition of 200 c.c. of water. Granulated zinc, pumice stone, or 0.5 gram of zinc dust is added. 50 c.c. of the strong sodium hydroxid solution, or sufficient to make the reaction strongly alkaline, should be slowly poured down the side of the flask so as not to mix at once with the acid solution. It is convenient to add to the acid liquid a few drops of phenol- phthalein or azolitmin solution, to indicate when the hquid is alkaline, but it must be noted that strong alkaline solutions destroy the former indicator. The flask is shaken so as to mix the alkaline and acid liquids and at once attached to the condensing apparatus. The receiving flask should have been previously charged with a carefully measured volume of the -^ acid (100 c.c. is a convenient amount). The distillation is conducted until about 150 c.c. have passed over. The acid is then titrated with standard alkali and methyl orange, cochineal, or azolitmin, and the amount neutrahzed by the distilled am- monium hydroxid determined by subtraction. Each c.c. of -7- acid neutrahzed is equivalent to 0.007 nitrogen. The distillation in this operation requires care, as the amount of ammonium hydroxid is determined by its neutrahzing power, hence solution of the alkali of the glass will introduce error. Common glass is not satisfactory. Block-tin is the best material for the Kjeldahl- Gunning form, but Moerrs has shown that it is not adapted to the methods in which mercury oxid is employed. He found that Jena-glass tubes resist the action of the ammonium hydroxid. The most satisfactory condensing arrangement for general laboratory use is a copper tank of good size, through which 36 FOOD ANALYSIS several condensing tubes pass. Such an arrangement is shown in side-view in figure 26. A more detailed view of the con- struction as applied to Kjeldahl distillations is shown in figure 21, which is a rough sketch, not drawn to scale. The flask is the standard Jena-glass distilling flask, about 12 cm. diameter, the tank should be high enough to allow of a condensing tube 60 cm. long. The connection of this with the receiving flask is made by means of a bulb tube to allow for occasional drawing- back of the liquid. The cork through which this tube passes into the flask must not fit closely, as opportunity must be given for expansion of the air. The safety tube connecting the distilling flask with the condenser should terminate a httle below the water level in the tank. The apparatus may be satisfactorily heated by the low temperature burner, as shown in figure 31. To avoid spurting of the boiling liquid, it is usual to interpose a safety-tube between the distilling flask and the condenser. Many forms have been suggested. Those shown in figures 22 and 23 are most in use. Figure 23 is the more complex, but is satisfactory. The distillation will be hastened if this tube be covered with non-conducting material. In some determinations (as in pepper) the Kjeldahl- Gunning method must be replaced by Arnold's modification: i gram of the sample is mixed with i gram of crystallized copper sulfate and I gram of mercuric oxid. The potassium sulfate- sulfuric acid mixture as given above is added and the mass heated cautiously until frothing ceases, when the temperature is raised Fig. 21. NITROGEN 37 and the digestion completed. The liquid is diluted for dis- tillation, 50 c.c. of a solution of commercial potassium sulfid (40 grams to 1000 c.c.) are added, and sufficient sodium hy- droxid as usual. The liquid is liable to bump. Modification jor Nitrates. If nitrates are present in the material, the weighed sample is well mixed with 35 c.c. of sulfuric acid containing 2 per cent., by weight, of salicyHc acid, and the mass shaken frequently during ten minutes ; 5 grams of sodium thiosulfate are added and 10 grams of potassium sulfate. Fig. 22. Fig. 23. The mixture is heated very gently until frothing ceases and then according to the usual method. The nitrogen in the distillate will include that derived from the nitrogen of the nitrates. Albuminoid Nitrogen. — Stutzer's method for this deter- mination requires a special reagent: Copper Hydroxid Mixture. 100 grams of copper sulfate are dissolved in 5000 c.c. of water, 25 c.c. of glycerol added, and then a dilute solution of sodium hydroxid until the hquid is alkaline. The mass is filtered, the precipitate is mixed well with water containing 5 c.c. of glycerol per 1000 c.c. and 38 FOOD ANALYSIS washed until the washings are no longer alkaline. It is then rubbed up with a mixture of 90 per cent, water and 10 per cent, glycerol in sufficient quantity to obtain a uniform magma that can be measured with a pipet. The quantity of copper hydroxid per c.c. should be determined. It should be kept in a well-closed bottle. Analytic Method. A suitable amount of the material, gen- erally about 0.7 gram, is heated with 100 c.c. of water to 100°, and a quantity of the copper hydroxid mixture containing about 0.5 gram of solid added, stirred well, allowed to cool, filtered, washed well with cold water, and the filter and pre- cipitate treated by the Kjeldahl- Gunning method. Substances rich in starch are best subjected to about ten minutes' warming in the water-bath instead of direct boiling. With substances containing much phosphate a few cubic centimeters of alum solution should be well stirred in before adding the copper hydroxid. Crude Fiber. The A. O. A. C. method is substantially as follows: 2 grams of the substance, well extracted with ether (see under "Ex- traction"), are mixed in a 500 c.c. flask with 200 c.c. of boiling water containing 1.25 per cent, of sulfuric acid; the flask is connected with an inverted condenser, the tube of which passes only a short distance below the rubber stopper of the flask. The liquid is brought to the boiling-point as rapidly as possible and maintained there for 30 minutes. A blast of air conducted into the flask may serve to reduce the frothing of the Hquid. The mass is filtered, washed thoroughly with boiling water until the washings are no longer acid; the undissolved sub- stance rinsed back into the same flask with the aid of 200 c.c. of boiling water containing 1.25 per cent, sodium hydroxid, nearly free from sodium carbonate; again brought to the boiling-point rapidly and maintained there for 30 minutes as ASH 39 • directed above. The liquid is filtered by means of a Gooch crucible; washed with boiling water until the washings are neutral to phenolphthalein ; dried at iio°; weighed and incin- erated completely. The loss of weight is crude fiber. The filters used for the first filtration may be linen, glass, wool, asbestos, or any form that secures clear and reasonably rapid filtration. Hardened-paper filters may serve. The sul- furic acid and sodium hydroxid must be made up of the specified strength, determined by titration. Some analysts use stronger solutions. Hehner used 5 per cent, acid and alkali. It would be convenient if normal sul- furic acid and normal sodium hydroxid were adopted as solvents. It is probable that carbon tetrachlorid could be advantageously substituted for ether in the preliminary extraction. Crude fiber should not be called cellulose. Ash. The ash of food materials may usually be determined by heating several grams in a platinum or porcelain crucible at a low red heat. Higher temperature may cause loss of volatile salts — e. g., chlorids. If a white ash cannot be obtained thus, the material should be heated only to a temperature sufficient to produce charring, the charred mass thoroughly extracted with water, and the insoluble matter collected on a filter, which may then be returned to the crucible and ashed. To this residue the filtrate containing the soluble matter is now added, the liquid evaporated to dryness, heated to low red- ness, cooled, and weighed. A muffle, heated by gas, will often be very useful in the incineration of organic bodies. A light draught of air should be maintained during the operation. Ash Soluble in Water. — The ash obtained as above is treated with boiling water, the solution filtered through an ashless filter, and the filter and contents again ignited and weighed. 40 FOOD ANALYSIS The soluble ash is determined by difference. If desired, the filtrate may be filtered to dryness, heated just below redness, and weighed. The first method is the most convenient. Alkalinity of ash is often an important datum. It will differ with the indicator used and whether tested by direct titration or upon the portions soluble and insoluble in water. The following method will furnish data of value in many cases. The ash is mixed with water, heated nearly to boiling, filtered and washed until the filtrate measures about 50 c.c. An in- dicator (phenolphthalein is usually employed) is added to the filtrate titrated to neutrality with ^ hydrochloric acid. Methyl orange is added and the titration carried to neutrality again. The filter and contents are dried, ignited, and added to the residue in the dish. Excess of standard acid and methyl orange are added and the material titrated to neutrality with sodium hydroxid. It is often sufficient to titrate the ash directly, using a single indicator and not separating the portions soluble and insoluble in water. In this case azolitmin may be satisfactory. Ash Insoluble in Acid. — ^The residue insoluble in water is treated with hydrochloric acid and the portion undissolved is well washed on the filter with water, dried, ignited, and weighed. The ash of jats is conveniently determined by the following method : A weighed quantity is melted in a platinum dish, and a smaller filter, free from ash, is folded in four, placed upright in the melted fat, and lighted. The fat is quickly burnt off. The following is a compilation of methods proposed for the determination of the ash of sugars, molasses, honeys: (i) 5 to 10 grams of the material are heated in a platinum dish of from 50 to 100 c.c. capactiy at 100° until the water is expelled, and then slowly over a flame until intumescence ceases. The dish is placed in a muffle and heated at low red- ness until a white ash is obtained. If the substance contain iron or any other metal capable of uniting with platinum, a dish of some other material must be used. For soluble ash the ash obtained as- above is digested with water, filtered EXTRACTION WITH MISCIBLE SOLVENTS 4 1 through a Gooch crucible, washed with hot water, and the residue dried at 100° and weighed. The difference of weights equals the soluble ash. (2) To 25 grams of molasses or 50 grams of sugar, 50 mg. of zinc oxid are added, and the mass incorporated thoroughly by adding dilute alcohol and mix- ing. It is then dried and ignited as above. The weight of zinc oxid is deducted from the weight of the ash. (3) The mass is carbonized at low heat, the soluble salts dissolved with hot water, the residual mass burned, the solution of soluble salts added, and evapo- rated to dryness at 100°, ignited gently, cooled in a desiccator, and weighed. (4) The sample is saturated with sulfuric acid, dried, ignited gently, then burnt in a muffle at low redness. One-tenth of the weight of the ash is deducted to calculate the percentage. Extraction with Miscible Solvents. For thorough extraction, especially with difficultly' soluble materials and volatile solvents, the continuous extraction ap- paratus devised by Szombathy, but commonly called the Soxhlet tube, is most suitable. The apparatus, as shown in figure 24, is provided with a globular metal condenser, but any form may be employed. The material may be placed in a fat-free paper thimble and covered with a plug of cotton to prevent loss of fine particles. In place of the cotton plug a Gooch crucible may be used, as shown in the cut! The top of the thimble should be a short distance below, and the top of the crucible a short distance above, the bend of the siphon. The thimble should be supported by a section of glass tubing, i to 2 cm. long, with rounded edges; the edge on which the thimble rests should be a little uneven to prevent a close joint, which would hinder the siphoning of some of the liquid. Another method is to use a glass tube open at both ends, the material to be extracted being held in position by loose plugs of cotton placed above and below. Loss of solvent by leakage often occurs. It may be dim- inished somewhat by soaking the corks in rather strong hot gelatin solution, draining them quickly and then exposing them for some hours to formaldehyde vapor. S 42 FOOD^ANALYSIS ^ The solvents most generally employed are ether and petro- leum spirit, but chloroform, carbon tetrachlorid, carbon disulfid, benzene, acetone and absolute alcohol have special applications. Carbon tet- rachlorid is well adapted for extraction purposes as it has high solvent power and is not easily inflammable. When extraction is completed, the carton and materials may be removed from the tube, and, replacing the parts of the apparatus, much of the solvent may be redistilled into the extractor, thus recovering the Hquid. Care must be taken not to distil the contents of the flask closely or heat strongly, lest some of the more volatile of the dissolved matters pass into the distillate. The tedious process of extraction may often by replaced by direct solution as follows: A convenient amount of the material, finely powdered, is placed in a flask, a definite volume of solvent, {e. g. loo c.c.) poured on, the flask tightly corked, the mixture gently shaken at convenient intervals for some hours, and allowed to remain in overnight. Care must be taken that the solvent does not come in contact with the cork. The mix- ture, after standing, is again shaken a few times, allowed to settle somewhat and an ahquot part (e.g. 50 c.c.) rapidly filtered off, evaporated as usual and ' weighed. The process is adapted for use with shghtly volatile solvents such as alcohol, but with care may be used with ether, petroleum Fig. 24. EXTRACTION WITH IMMISCIBLE SOLVENTS 43 spirit, and carbon tetrachlorid. It has value as a sorting method. Extraction with Immiscible Solvents. Solvents not miscible with water are employed for extracting substances by shaking the solvent thoroughly with the aqueous solution, allowing the liquids to separate, and removing one of them. The process is most conveniently performed in a stoppered separator. The principal difficulty is the liabil- ity of some liquids to form emulsions which separate only after long stand- ing. Separation may sometimes be hastened by cooling the mixture or by adding more of the solvent. One of the most satisfactory methods when operating upon small amounts of liquid is to whirl the mixture for a short time in a high-speed centri- fuge. Figure 25 shows a special appa- ratus for use with solvents lighter than water. The cylinder A should hold about 1000 c.c. Two openings are not necessary, since both tubes may pass through the cork, but the arrangement shown is more convenient. 600 c.c. of the solution are placed in the cylinder, 300 c.c. of solvent added and the mixtures well shaken. The rest of the apparatus is then attached. The flask B has a capacity of 200 to 300 c.c. ; the solvent in it is heated by a water-bath. The vapor passes by a into b, the condensed liquid flows to the bottom of A and rises through the solution; the upper layer Fig. 25. 44 FOOD ANALYSIS returns through c into B. The tube c should not extend into the liquid in B. A small quantity of aqueous liquid may collect at intervals in B and should be removed. Distillation and Sublimation. Retorts and alembics are now but little used, but are service- able in some cases. With glass vessels the irregular percussive boiling, commonly called "bumping," is liable to break the Fig. 26. vessel or to spurt portions of the undistilled Hquid into the con- densing apparatus. This may often be prevented by the ad- dition of a few fragments of pumice, clay pipe, or platinum foil. Dry pumice floats on most liquids. It may be made to sink either by soaking it in water for a day or so or by heating the fragment to redness and quenching it in the liquid. With DISTILLATION AND SUBLIMATION 45 inflammable liquids, the latter method must be used cautiously. Bumping may often be prevented by using the burners shown" in figures 31 and 32. Condensing apparatus is made in considerable variety; Glass and block-tin are the materials for tubes. The glazed porcelain tubes made for pyrometers would probably be well adapted for straight condensing tubes. Glass tubes are liable to crack at the point at which the cooling action begins. To avoid leakage and the contact of hot vapors with corks or rubber tubes, the connections should be as few as possible. Fig- ure 26 shows a copper tank through which the condensing tube passes. This apparatus is especially adapted to the so-called ** ammonia" process for water-analysis. The neck of the retort being inclined slightly, as shown, causes any material thrown into it to return to the boiling liquid. Figure 27 shows an improved form of distilling apparatus devised by R. S. Weston. The condenser tube is of copper or japanned galvanized iron. The details of construction and arrangement are sufficiently indicated in the drawing. The apparatus is shown as arranged for water analysis. When Kjeldahl distillations are being made the lower end of the block-tin tube should be extended by means of a bulbed glass tube, as noted elsewhere. Safety bulbs may also be placed between the flask and condensing tube in such a way as to avoid rubber- tube connections. Materials are added by means of long-stemmed funnels. Weston uses a Bunsen burner, but it is probable that the burners figures 31, 32, would be more satisfactory. Figure 28 shows Cribb's condenser, which may be attached to any distiUing apparatus. The distillation tube is attached at A. The walls are double; condensation occurs in the space between them, and the distillate flows out by the tube E. The cooling water flows through F to the bottom of the inner space, overflows at J into the catch-basin below, escaping by G, 46 FOOD ANALYSIS The stopper / serves to steady the tube F, and should have several large notches cut in it to allow the water to escape freely. It is usually necessary to wrap a piece of muslin around the C/a/9t/o\ S*fpfiorf^ Ct^* Fig. 57. outside of the apparatus to cause the overflowing water to run properly. The condenser may be made of glass, block-tin, or tinned copper. Experience shows that the apparatus will be DISTILLATION AND SUBLIMATION 47 more satisfactory if some of the dimensions are changed from those indicated in the figure, which is taken from Cribb's paper. The annular space should be larger, especially at the bottom; the catch-basin must be roomy, and G should have a caliber at least three times that of F. The catch-basin is held in place by rubber tubing. The condenser is supported by a strong clamp. L is for attachment of an air-pump for distillation under diminished pressure. Distillation of small amounts of material may be made with the ordinary extractor, terminating the operation before the distillate reaches the level of the bend of the siphon. For many distillations the simple appa- ratus shown in connection with determin- ation of the volatile acids of butter will serve, but a side-neck flask, as shown in figure 29, is more generally useful. In this figure the condensing tube is represented relatively too short; for the volatilp bodies encountered in food anal- ysis the condenser should be at least 50 cm. long. This form of flask permits of introduction of materials without discon- necting the apparatus and also of distilla- tion in a current of steam or of indifferent gas. Fig. 28. For distillation in a current of steam, a generator is needed. A Jena flask of good size is most con- venient. It should be provided with a stopper with two tubes, one about 0.5 cm. caliber, reaching to near the bottom of the flask, the other about i.o cm. caliber, reaching just below the level of the stopper. The latter is connected with a tube passing nearly to the bottom of the side-neck flask. The smaller tube in the generator is for safety in case of obstruc- 48 FOOD ANALYSIS tion. Its upper opening should be directed so that no damage will be done if the hot liquid is thrown out. With steam distillation, a moderate heat should be maintained under the distillation flask, and the water in the generator kept boiling actively. The junction between the two flasks should be by tubes which touch as closely as possible, held by a rubber sleeve. Fig. 29. Inverted Condenser. — For prolonged boiling in water without concentration, the simplest arrangement is a flask fitted with a cork carrying a tube about 2 meters long. The lower end should be cut off obliquely. If the boihng is moderate, the vapors will condense and run back. For volatile liquids or special cases, regular condensers are used. The ordinary straight form, made of glass, is usually employed, but the ball- APPARATUS AND CHEMICALS 49 form, shown in figure 24, is compact. This can be obtained of glass. Fractional distillation is best carried out with the bulb-tubes devised for attachment to ordinary flasks so that the vapor may be partially condensed and succeeding portions washed with the liquid which runs back continuously into the flask. The most used are the Le Bel-Heninger and Glynsky tubes. The former bears from two to six bulbs. The upper part has an inclined side tube for connection with the receiver and an opening through which the thermometer can be passed. Each of the bulbs is connected with the one just below by a side tube. At the constricted part of each bulb a small thimble of platinum, copper, or nickel gauze rests. The vapor condenses in the cups and washes the vapor subsequently formed. The liquid runs off from each bulb, back to the flask. The flame should be regulated so as to keep all the cups full, and cause the distillate to fall from the end of the tube in separate drops. In the Glynsky bulb, glass balls replace the gauze. The United States revenue-law requires all distilling ap- paratus to be registered, no matter for what purpose it is used. Heavy penalties are imposed for using non-registered stills. No fee is imposed for registry, which is made on blanks furnished by the Collector of Internal Revenue. Sublimation may be performed in a narrow test-tube or watch-glasses with concavities facing, the upper glass being slightly small so that it may fit well. A gentle heat is applied to the lower dish. By substituting a beaker containing water for the upper watch-glass a better cooling effect will be ob- tained. Apparatus and Chemicals. These can now be obtained generally of good quality at almost all times and places, but a few suggestions may be of value. C FOOD ANALYSIS Centrifuge. — Centrifugal apparatus is of much advantage in laboratory work. The slow-speed machines made for milk analysis are of limited application; much better results are obtained by the high-speed apparatus of the type shown in figure 30. In operating such machines, the load on the revolving arms must be balanced or the center of gravity will not coincide with the center of revolution, and an objectionable vibration will be produced. The machine should be attached to a firm table or shelf and kept properly oiled and protected from dust. The tubes usually fur- nished are narrowed at the bottom, and, as solid material is apt to be packed closely by the centrifugal action, it is sometimes difficult to J- ^,c..oK.r3fiill dislodge it, but care should be taken to get all such material out of the tube so as not to contaminate the substance used in a subsequent ex- periment. If it be desired to use vessels not narrowed at the base, small glass tubes closed by cork at one end may be substituted. In this case, however, the lower end of the tube-holder should be packed with cotton to such a height that the cork cannot be driven into a part of the tube narrow enough to hold it tightly. If this precaution be neglected, the rotation will push the glass tube so far into the tube-holder that it may be impossible to draw it out without leaving the cork. Glassware suitable for most laboratory work is now made in the United States, but the Bohemian and Jena glass still Fig. 30. APPARATUS AND CHEMICALS 5 1 shows important merit which will lead to preference for it in many cases. For the cleaning of glass and porcelain, espe- cially when working with fatty matters, the commercial triso- dium phosphate is of much use. Vessels cleaned with it must be well rinsed. A bath of so-called battery fluid (potassium dichromate or sodium dichromate, or, better, the crude chromic acid sold for the purpose, 250 grams; water, 2000 c.c; sul- furic acid, 300 c.c.) will make an efficient cleaning solution for all non-metallic articles. These should be cleansed with soap, sodium phosphate, or sodium carbonate to get rid of the greasy matters, rinsed, and then soaked in the liquid overnight. The solution gives off no fumes and its color guards against imperfect rinsing. It is of little value when it has become brown or green, but may be freshened by adding crude chromic and sulfuric acids. As the liquid is very corrosive, all waste from it should be washed down the drain-pipes with a free flow of water. Strong sulfuric acid is used by some chemists, especially for cleaning greasy apparatus. Organic materials such as corks and rubber tubes should, of course, not be put in these cleaning solutions. For heating beakers and flat-bottomed flasks the hot-plate is much used, but the thin cast-iron plates commonly furnished are unsatisfactory. A better form is a rolled plate at least i cm. thick. Nickel wire-gauze is a good substitute for the common wire-gauze. The Chaddock burner, made of non- corrodable materials, is now obtainable, and is adapted to use in the fume-box. Electric heating apparatus has been brought to considerable efficiency, and will in time supplant all present methods, but the installation and operation are as yet costly. An incandescent lamp may be arranged as a heating apparatus, and is especially satisfactory in extractions and distillations with inflammable materials. The low-temperature burner and evaporating burner shown in figures 31 and 32 are convenient in many operations, especially in heating liquids liable to bump. 52 FOOD ANALYSIS The inlet of the former is too short; it should be lengthened by a piece of metal tube, or the rubber connection will become hot. In default of this lengthening the joint may be kept cool by wrapping around it a piece of muslin, the ends of which dip in a vessel containing water. Filter-papers are furnished in great variety, adapted to all purposes. The so-called hardened filters are serviceable in several operations, such as determination of crude fiber, insolu- ble matter, and extraction with volatile solvents, for with care the wet precipitate can be scraped off without removing an appreciable amount of the filter-paper. Slightly flattened Fig. 31. Fig. 32. glass rods or round rods bent at the middle to an obtuse angle are convenient because they arc not liable to roll off of beakers or funnels. Reagents, especially those used only in small amounts, are most conveniently kept in capped bottles, each with small glass tube or pipet, the tube being long enough to reach above the top of the bottle (figure 17). In this way the solution will not get in contact with the neck of the bottle. Solids should be kept in hood-stoppered bottles, — i. e., those in which the fiat top of the stopper is close to the bottle, — so as to give less chance for de- posit of dust. All chemicals in general use should be kept in closed cases, ammonium hydroxid and ammonium carbonate APPARATUS AND CHEMICALS 53 being separate from the common acids. The stock bottles for acids and standard solutions should be protected from dust by placing over the stopper of each, an inverted tumbler large enough to rest on the top of the body of the bottle. Platinum ware requires care to'prevent staining and crack- ing. Substances containing any of the easily-reducible metals must not be heated in contact with platinum; even iron com- pounds in the presence of reducing agents — e. g., filter-paper — will do harm. Sudden cooling of platinum should be avoided, as it tends to make the metal brittle. After being heated to redness the metal, when cold, should be lightly rubbed with very fine sea-sand (not river-sand nor powdered quartz or pumice), by which the metal will be burnished and its texture preserved. The platinum-pointed forceps should be treated in the same way. Platinum dishes may often be cleaned by rubbing them with sodium amalgam, decomposing this by immersion in water, and driving the mercury off by heating to redness. Some stains may be removed by melted potassium acid sulfate. Nickel dishes may be substituted for platinum in cases in which only gentle heating is required, but nickel is apt to be injured by direct heating with gas. For lubrication of glass stopcocks, the following mixtures, devised by Phillips, are useful : Pure rubber, 70 parts Pure rubber, 70 parts Spermaceti, 25 " Unbleached beeswax, ... 30 " Vaselin, 5 " The rubber must be fresh and pure; rubber scraps will not answer. It should be melted in a covered vessel, the other materials added, and the mixture well stirred while hot, care being taken not to scorch it. It must not be exposed to air longer than is necessary during heating, and should be kept in well-closed bottles. These mixtures may be removed from 54 FOOD ANALYSIS stopcocks by a little strong nitric acid which loosens the lubri- cant so that it may be rinsed off. All the largely used chemicals are obtainable of good quality, as a rule, but in important investigations tests for purity and strength should be applied. The following notes will assist in this. Alcohol. — Ethyl alcohol, commonly called "grain alcohol," contains in its strongest commercial form about 95 per cent, of ethyl hydroxid, notable quantities of esters, aldehydes, fusel oil, and traces of acid. For some purposes — e. g., making standard solutions of alkali — it must be purified by redistilla- tion over sodium hydroxid. The absolute alcohol sold by dealers usually contains some water. The presence of water in alcohol may be detected by the evolution of acetylene when a little calcium carbid is added. This may also be used for removing small amounts of water, the liquid being redistilled, but hydrogen sulfid, hydrogen phosphid, and ammonium compounds may be introduced. Anhydrous copper sulfate is turned blue by alcohol containing water. Methyl alcohol. Crude wood-alcohol is of limited use in laboratory work. It contains much acetone. A purified article is now furnished, under the trade name ''Columbian Spirit," which is about 98 per cent, methyl hydroxid and is free from notable amounts of impurities. It may be used with economy as a substitute for ethyl alcohol in many cases. It is more volatile, but traces of strong-smelhng foreign matters may cause the odor to persist longer than with refined alcohol. Ether. Commercial ether contains notable amounts of al- cohol and water, but much purer samples can be obtained from dealers in laboratory supphes. To obtain good results with ether it is essential that it be as nearly as possible free from alcohol and water. The method of purification recommended by the A. O. A. C. is as follows: Commercial ether is washed with two or three successive portions of distilled water and soHd sodium hydroxid added APPARATUS AND CHEMICALS 55 until most of the water has been extracted. Carefully-cleaned metallic sodium, cut into small pieces, is added until there is no further evolution of hydrogen. The ether thus dehydrated must be kept over metallic sodium, and should be only lightly stoppered in order to allow hydrogen to escape. Chlorojorm, benzene, petroleum spirit and carbon tetrachlorid are usually obtainable of good quality. All are liable to contain water. This may be removed by shaking with anhydrous calcium sulfate or anhydrous copper sulfate and redistillation. Commercial chloroform is liable to decomposition, by which it becomes acrid. All volatile solvents are liable to contain appreciable amounts of non-volatile materials, and should be tested by evaporating a measured amount and weighing the residue. If this is appreciable the solvent should be distilled. Carbon tetrachlorid is well adapted for fat extraction when an open flame is used. Light petroleum, commonly known as benzin and gasolin, and often by other trade-names, should be purified by redistillation, selecting the portions which distil over below 50°. Sodium hydroxid. Several brands sold for household use are suitable for ordinary purposes, such as making standard alkali or in the Kjeldahl- Gunning process. Potassium hydroxid. The specially purified grades should be used. Sand and asbestos intended for moisture and extract deter- mination must be selected with care, and dried thoroughly before weighing. Common sand contains much material other than quartz ; asbestos fiber is often of inferior quality. Indicators. — Numerous indicators have been proposed, but for ordinary laboratory work litmus, phenolphthalein, and methyl-orange are usually preferred. Litmus. Litmus solution is now little used, but azolitmin, a pure blue color obtained from it, is a sensitive indicator. It is freely soluble in water but insoluble in alcohol. The solution 56 FOOD ANALYSIS must be kept in an open bottle. Intermediate litmus-paper, which is convenient for ascertaining the reaction of liquids, is prepared as follows : A clear, fresh solution of litmus is divided into two equal portions; one of these is rendered purple-red (not bright red) by the cautious addition of dilute nitric acid; the other portion is then added and strips of good filter-paper soaked in the liquid and dried quickly. This paper will be affected by ordinary acid or alkaline solutions. It should be kept in the dark, protected from dust. Phenol phthalein. A solution of i gram in 100 c.c. of good (methyl or ethyl) alcohol is sufficient and keeps well. Methyl-orange. A solution of o.i gram in 100 c. c. water will be satisfactory. In titrating with methyl-orange very little of the indicator should be used. Cochineal. Many prefer this indicator for titrating am- monium hydroxid. 3 grams of powdered cochineal are macer- ated for several days, with occasional shaking, in 100 c.c. alcohol of about 20 per cent., and the solution filtered. Starch Indicator. — This is much used in titrations with iodin. As it spoils quickly, it is usually made as needed. Moerk has found that oil of cassia acts as a preservative without interfering with the efficiency of the solution. 5 grams of good starch (preferably arrow- root) are mixed with about 100 c.c. of cold water, and the mixture poured into 500 c.c. of boil- ing water with active stirring. The liquid is allowed to cool, 2 c.c. of oil of cassia added, made up to 1000 c.c, shaken and preserved in a well-stoppered bottle. Standard acid. — The strength of dilute sulfuric acid can be accurately determined by adding to a carefully measured quantity a slight excess of pure ammonium hydroxid, evapora- ting in a platinum basin to dryness and weighing the ammonium sulfate. The solution to be valued must contain nothing but sulfuric acid and water, and the ammonium hydroxid must be entirely volatilized by evaporation on the water-bath. APPLIED ANALYSIS 1 GENERAL METHODS POISONOUS METALS The elements included under this title are mercury, arsenic, lead, tin, copper, zinc and chromium. Some very poisonous elements not likely to be encountered in foods, are not con- sidered in this connection. A. H. Allen has devised a general process for the detection of poisonous metals. A convenient quantity of the substance, say 25 grams, is mixed by degrees with sufficient strong sulfuric acid to moisten the mass thoroughly without making it fluid. About 2 c.c. will generally be required. Liquid material should be evaporated to dryness or nearly so at a low temperature before being treated with the acid. The mass is heated for a short time on the water-bath, after which the temperature is gradually raised to a point just below that required to volatilize the sulfuric acid, and maintained until the action seems to be complete. It is not necessary to carry on this part of the process until all the carbon is burnt off. The mass is allowed to cool, about I c.c. of strong nitric acid added, and the heating con- tinued until red fumes are evolved. Allen recommends the use of a porcelain crucible in these operations, but the Kjeldahl digestion flasks of Jena glass would probably serve. Recently ignited magnesia, in the proportion of 0.5 gram for each cubic centimeter of the acid used, is incorporated with the mass and the mixture burned off at a dull red heat, preferably in a muffle. After cooling, the ash is moistened with nitric acid, again burned off, and the process repeated until all the carbon is con- 57 58 FOOD ANALYSIS sumed. The residue is treated with 0.5 c.c. of sulfuric acid, heated until fumes are evolved, cooled, boiled with water, diluted without filtration to about 100 c.c, saturated with hydrogen sulfid, the solution filtered and examined according to the following scheme : Aqueous Solution may contain zinc and iron. Precipitate and Residue may contain lead Add bromin water to destroy hydrogen sulfid. sulfid, stannic oxid, copper sulfid, or calcium convert iron into the ferric state, boil, then sulfate. Fuse in porcelain crucible for lo add excess of ammonium hydroxid, boil minutes with 2 grams of mixed potassium again, and filter. and sodium carbonates and i gram of sulfur. When cool, boil with water and fiher. Precipi- Filtrate if blue, contains nickel. Residue. Boil with strong hy- Filtrate. tate may Divide into two portions: drochloric acid as long as hy- Acidulate contain 1 drogen sulfid is evolved, add with ace- iron (and a few drops of bromin water tic acid. p h s - to complete the oxidation of A yellow phates). the copper sulfid, and filter if necessary. To the filtrate add excess of ammonium hydroxid, when a blue coloration will be indicative of copper. Acidu- late the Hquid with acetic acid precip- itate of stannic sulfid in- dicates tin. and divide into two portions: I. Heat to boil- n. If zinc I. Add potas- II. Add potas- ing and add found in I, sium chro- sium ferro- potassium ferrocyanid. for its deter- mate. A yel- cyanid.; A mination, low precipi- brownish White pre- acidulate tate indicates precipitate cipitate or the ammoni- lead. or coloration turbidity in- acal solution indicates cop- dicates zinc. strongly with acetic acid, filter, if nec- essary, and precipitate the zinc from the fihrate by hydrogen sul- fid. Any nickel pres- ent will also b e precipi- tated. per. Allen's scheme does not include chromium, which may be present as a constituent of lead chromate and will be found almost entirely in the precipitate and residue insoluble in water. For its detection a portion of this or of the original ash should be fused with sodium carbonate and potassium chlorate; the yellow melt, containing chromate, is dissolved in POISONOUS METALS 59 the smallest possible quantity of water and slightly acidulated with hydrochloric acid. The liquid is then added to a test- tube containing a small amount of hydrogen dioxid overlaid with a little ether. In the presence of a chromate the water will acquire a blue color, which on slight shaking will pass into the ethereal layer. When iin is known to be present, the amount may be found by treating the precipitate of stannic sulfid with strong nitric acid, igniting the metastannic acid formed, and weighing the resultant stannic oxid. For the detection of tin it is recom- mended to treat the stannic sulfid with hydrochloric acid and bromin water and boil the filtered liquid with iron wire to reduce to the stannous condition. The liquid is diluted and decanted from the undissolved iron and any precipitated material, and the tin detected by adding a drop of mercuric chlorid solution, which will produce a white or gray turbidity according to the amount of tin present. Copper may be estimated colorimetrically by means of ammonium hydroxid or potassium ferrocyanid. According to Bodmer & Moor, for very small amounts the ferrocyanid method is more accurate. Paul & Cownley determine copper as follows: The sample is carbonized in a platinum dish and extracted with a little hydrochloric acid; the insoluble residue is ignited with a little nitric acid, hydrochloric acid added, and the resulting mixture added to the original extract. The solution is then concentrated to about 30 c.c, placed in a weighed platinum dish, and the copper deposited with pure zinc. If the deposit is not of true copper color, it is dissolved in a little nitric acid and the copper determined colorimetrically. Zinc. — Evaporated fruits are liable to derive zinc from the trays on which the drying is conducted. Wiley gives the follow- ing process for determination : The sample is placed in a large platinum dish and heated slowly until dry and in incipient com- bustion. The flame is removed and the combustion allowed to or THE UMlVrDGi-r^ 6o FOOD ANALYSIS proceed, the lamp being applied from time to time, in case the burning ceases. The mass, when burned out, consists of ash and char. It is ground to fine powder and extracted with hydrochloric or nitric acid, the residual char is burned to whiteness at a low temperature, the ash extracted with acid, the soluble portion added to the first extract, and the whole filtered. A drop of methyl orange solution is placed in the liquid and ammonium hydroxid added until it is only faintly acid. The iron is precipitated as ferric oxyacetate by adding 50 c.c. of a solution of ammonium acetate, 250 grams to the liter, and raising the temperature to about 80°. The precipitate is separated by filtration, washed in water at 80° until free from chlorid, the filtrate saturated with hydrogen sulfid, allowed to stand until the zinc sulfid settles, and poured on a close filter. It is often necessary to return the filtrate several times before it becomes limpid. The collected precipitate is washed with a saturated solution of hydrogen sulfid containing a little acetic acid. The precipitate and filter are transferred to a crucible, dried, ignited, and the oxid weighed. Arsenic, if present in notable amount, may be detected by ReinscWs test, a liberal amount of hydrochloric acid being used, since arsenates do not otherwise respond to the test. Some water strongly acidulated with hydrochloric acid is placed in a test-tube, about half a square centimeter of bright copper foil added, and the liquid boiled gently for a few minutes. If the copper remains bright, showing that the reagents contain no arsenic, the material to be tested is added and the liquid again boiled for several minutes. If arsenic be present, a steel-gray stain will appear on the copper. The slip is removed, washed with distilled water, dried by pressure between filter-paper, placed at the closed end of a narrow glass which has been previously dried by heating nearly to redness. The tube is gently heated at the point at which the copper rests. The arsenic will be converted into arsenous oxid, which will collect on the cooler portions of the tube in octahedral crystals. POISONOUS METALS 6 1 Reinsch's test cannot be applied in the presence of active oxidizing agents, such as chromates, chlorates, or nitrates. GutzeWs test, which is more delicate, is as follows: Place in a tall test-tube about a gram of pure zinc, 5 c.c. of diluted sulfuric acid (6 per cent.), and i c.c. of the sample. The mouth of the test-tube is covered with a tightly-fitting cap of three thicknesses of filter-paper. A drop of strong solution of silver nitrate is placed on the upper paper and the tube allowed to stand for 10 minutes in the dark. If arsenic be present, a bright yellow stain will appear on the filter-paper^ which, on the addition of water, becomes black or brown. A blank test should aways be made to establish the purity of the reagents. Sulfids (which may be detected by substituting lead acetate for the silver nitrate in the above test) must be oxidized to sulfates before applying the test. The test is delicate. A less rigorous one may be made by substituting a drop of a saturated solution of mercuric chlorid for the silver nitrate. If no yellow coloration appears after 10 minutes, the sample may be considered free from arsenic. The purity of the reagents must be carefully ascertained' before applying any of these methods. For the detection of minute amounts of arsenic. Marsh's test is used. The details as given by Haywood are generally applicable. The apparatus consists of a flask holding about 100 c.c, with a rubber stopper through which passes a long-stemmed separa- tory funnel — the tube of which should reach nearly to the bottom of the flask — and an exit tube bent at a right angle. The flask should stand in a basin containing cold water. The exit connects with a bulb-tube containing a small amount of lead acetate solution, to absorb sulfur, selenium, and tel- lurium. To this is connected a calcium chlorid tube, and, finally, a tube of very resistant glass, about 20 cm. long and not over 0.5 cm. caliber. It must be drawn out to nearly capillary 62 FOOD ANALYSIS narrowness about the middle. A piece of fine wire-gauze is wrapped around the tube for a few centimeters on the wide part nearer the flask. The gauze must not reach to within a centi- meter of the narrow part. Two Bunsen burners must be ar- ranged so as to be used at once to heat the gauze. The general arrangement is as figure 33, except that the protecting gauze, extra burner and stem of the separator are not shown. The burners are placed so that the flames meet and the gauze is at that point. The bulb-tube may be placed in water. The extra-tube, closed by a pinch-cock, is convenient but not neces- sary. If used, care should be taken that the pinch-cock closes it well. Fig. 33. For use, three grams of arsenic-free zinc are placed in the flask and then 30 c.c. of dilute pure sulfuric acid (i to 8). The apparatus is connected and the hydrogen allowed to flow for 15 minutes, after which the gauze is heated strongly for 20 minutes. No deposit should appear in the tube. The prepared material (see below) is placed in the funnel and gradually run into the flask. The action is continued for about an hour, the portion of tube within the gauze being kept POISONOUS METALS 63 very hot all the time. The tube is allowed to cool and the extent and appearance of the deposit compared with tubes of known value. The sample is best prepared by mixing a small weighed portion in a porcelain basin with from i to 5 c.c. of a mixture of nitric and sulfuric acids. The mass is heated with a low flame until it has granulated and fumes of sulfuric acid are not abun- dant. The charred mass is broken up, mixed with a little water, and boiled to get rid of sulfurous acid. It is filtered, the residue washed, and the filtrate and washings made up to a definite volume (about 40 c.c). It is then ready for the de- termination. The comparison tubes are made by using measured volumes of a standard solution of arsenous oxid in such amount as will contain the following fractions of a milligram of elementary arsenic J operating with each solution as directed above: 0.005; o.oi; 0.02; 0.03; 0.04; 0.05; 0.06; 0.07. These deposits (mirrors) should be sealed and kept in the dark. Even then they fade, and for accurate observation should not be over three weeks old. . The standard solution is made by dissolving o.oi 32 gram of dry pure arsenous oxid and o.i gram pure sodium acid carbonate, in 100 c.c. of water. The mixture is kept hot until the arsenous oxid is dissolved, cooled, slightly acidified with sulfuric acid, and made up to 1000 c.c. Each c.c. of this solution contains 0.00001 gram of elementary arsenic. Aliquot portions are used for making the standard mirrors. As this test is extremely delicate, great care must be taken to ensure purity of all reagents. It must be borne in mind that most natural substances will give slight reactions for arsenic by it. All junctions must be as tight as possible. The connected points of the different pieces should be of the same diameter and the junctions made by short, close-fitting, pure rubber tubing. 64 FOOD ANALYSIS Great care must be taken that the apparatus is thoroughly cleaned between each use. In cases in which the results are to be used in criminal prosecutions the apparatus should be new. COLORS At present, the colors used in food-articles are mostly synthetic products, commonly called "anilins," but largely derived from other coal-tar materials. Natural organic colors — annatto, cochineal, turmeric, indigo, saffron, and chlorophyl — are used to a limited extent, but mineral colors are rarely employed. Ferric oxid is used in some chocolate substitutes. The identification of individual colors in mixture with foods or beverages is difficult, often impossible, with methods at present available. It is possible in many cases to distinguish between artificial and natural colors. The following method is generally applicable for distinction between these classes. Pure white wool (the material known as ''nun's veiling" is satisfactory) is cleaned by boiling for a short time in soap- suds, washed thoroughly with water, well-dried, and cut into slips about 3 X 10 cm. They should be kept in a closed bottle. A convenient quantity of the material, depending on the amount of color, is placed in a beaker. For ordinary liquids, 100 c.c. will suffice; for soHds and semi-solids from 5 to 25 grams. In the latter case, water should be added to make the bulk about 100 c.c. The beaker is placed in a water bath, i c.c. of hydrochloric acid added and a slip of the cleaned wool. The liquid is kept in the boihng water for a moderate time. If not appreciably dyed in fifteen minutes it may be assumed that no coal-tar color is present. In most cases, however,- some color will be imparted, even if only natural colors are present. The slip is washed well with cold water, warmed for a few minutes in very dilute hydrochloric acid, again washed well, and im- COLORS 65 mersed in about 25 c.c. of water to which 2 c.c. of strong ammonium hydroxid have been added. By this means, the color will generally be dissolved promptly from the slip, but it may be necessary to allow much longer action. When the cloth is nearly or quite decolorised, it is taken out of the liquid. The latter is diluted to about 50 c.c, rendered moderately acid by ad- dition of hydrochloric acid, another slip of cleaned wool im- mersed and the liquid heated in the water bath. Coal-tar colors and some lichen colors (archil, cudbear, litmus) will give marked second dyeing. Lichen colors, including a sulfonated orcein, now often used in food articles, are distinguished by Tolman's method,^ depending on the fact that amyl alcohol removes them from the ammonium hydroxid solution. If, therefore, a double dye - ing is obtained, the process should be repeated, but the am- monium hydroxid solution should not be acidified but shaken with pure amyl alcohol. If this acquires a purplish red tint, it is evaporated on the steam bath, the residue dissolved in water and the solution mixed with a little tin and hydrochloric acid. Lichen dyes are bleached by this method and are restored by ferric chlorid. These reactions exclude all azo-dyes and ma- genta. Some tests adapted specially to the recognition of colors in particular foods will be described in connection with such foods. When dyes intended for food-coloring are to be examined in bulk, the following methods are advantageous : A small quantity of the sample (o.i to 0.25 gram) is heated on platinum foil. Nitro-colors show more or less deflagra- tion at first. Sulfonated colors form a fusible residue, in which the carbon bums with difficulty. It will be advanta- geous to add some oxidizing agent (potassium nitrate, potas- sium chlorate, or sodium nitrate). It is not necessary to bum off all the carbon. The mass is allowed to cool, boiled up with water acidulated with hydrochloric acid (this may 7 66 FOOD ANALYSIS cause the evolution of a little hydrogen sulfid), and barium chlorid added. A copious white precipitate will occur if the color is a sulfonated one. For detection of arsenic the Reinsch test may be applied or the color may be examined for all the important poisonous metals by the scheme given on page 58. Identification of colors may sometimes be accomplished by routine methods, several of which are given in the following pages. The first is Green's adaptation of Weingartner's tables. It is reproduced without modification of spelling or nomenclature from Allen's '' Commercial Organic Analysis," edited by Matthews. The reagents required are as follows: Tannin solution. Tannin, i gram ; sodium acetate, i gram ; water, 10 c.c. Zinc dust. Dilute hydrochloric acid: Hydrochloric acid, 5 c.c; water, 15 c.c. Ammonium hydroxid solution. Chromic acid solution: Chromium teroxid, i gram; water, 100 c.c. Chromic- sulfuric acid solution: Chromium teroxid, i gram; strong sulfuric acid, 2.5 c.c..* water, 100 c.c. Strong sodium hydroxid solution: Sodium hydroxid, 33 grams; water, 67 c.c. Dilute sodium hydroxid solution: Sodium hydroxid, 5 grams ; water, 95 c.c. Alcohol. 70 per cent. In applying the scheme a primary division is made into dyes soluble and insoluble in water. The former are divided by means of the tannin solution into the so-called basic and acid groups. The dyes which in aqueous solution are pre- cipitated by tannin solution are termed basic dyes. The reduction with zinc dust is best made by adding a little of the zinc dust to the hot dyestuff solution contained in a COLORS 67 test-tube, agitating, and adding dilute hydrochloric acid drop by drop until decolorized. Excess of acid should be care- fully avoided. When the color acid is quite insoluble, the reduction is made with zinc dust and ammonium hydroxid. The reduced solution is decanted upon a small filter; if the color does not return in a few minutes, the paper is moistened with chromic acid solution. In the case of acid colors the chromic-sulfuric acid solution should be used. As some dyes do not show their color in presence of free acids, the paper should be exposed to the fumes of strong ammonium hydroxid solution before deciding as to whether the color will return. 68 FOOD ANALYSIS 5^ ^■^ " 3 JO 3 5o 8S <1 7 ;l5 «w 1^ p. - 23 nil Oi o r O o^g>.pq;,. ,^^5 2^3 2i o S'IS;§^1 60§K "NTS O C N H* 03 OS o .CO « s .I' E o >. "3.2 >.'5 o o s 11 (i5 G '> o5 1 11 ^ 'E C il If 1 "^ fl'O o ^ COLORS 71 rota's scheme for recognition of colors^ Two special reagents are used. Stannous chlorid 10 per cent, solution in hydrochloric acid. Potassium hydroxid 20 per cent, solution in water. The material may be tested in solution in water or alcohol. It should be diluted with water or alcohol, as required, until the color is not deep. Turbid liquids njust be filtered. A com- parison test of the solution should be made with hydrochloric acid alone, as many effects of the stannous chlorid reagent are due to the acid and not to the tin compound. Some colors require considerable time to effect a change. To a portion of the solution a small amount of stannous chlorid reagent is added, the mixture shaken and heated to boiling. The same test is applied to another portion, using hydrochloric acid alone. 1 . The stannous chlorid decolorizes the liquid (see A) . 2. The color is not affected more than by hydrochloric acid alone (see B). A. The liquid is mixed with either ferric chlorid or hydrogen dioxid or is shaken with air. The color does not return. Nitro-, nitroso-, azo- and hy- drazo-colors. Picric acid, naphthol yellow, Ponceau, Bordeaux, Congo-red. The color is restored. Indogenid, imido-quinones> methylene blue, safranin, indigo-carmine. B. A part of the original solution is mixed with some of the potassium hydroxid and warmed. The liquid is decolorized Amido-derivatives of di- and or rendered turbid. triphenylmethane, auramins, acridins, quinolins and colois from thiobenzinil. The reagent produces no discoloration or turbid- ity. Monamid, diphenylmethane, oxyketone, eosins, aurin, aliz- arin and most natural colors. Many of the powders and pastes sold for imitating natural vegetable colors are mixtures of several coal-tar colors often 72 FOOD ANALYSIS representing several types, so that the above schemes will give confusing results. The identification of the ingredients of such mixtures can generally be done only by expert color-chemists, but some information may be obtained by dyeing successive por- tions of wool in the same bath. The color with the strongest attraction is taken out in greater amount in the first dyeing, and a series of dyed slips will be obtained showing the principal tints of the mixture. Information is also often gained by dye- ing in different baths. The color material to be tested is made up with about loo c.c. of water, a few grams of sodium sulfate and 2 c.c. of strong sulfuric acid. Another bath is made with a few grams of alum in loo c.c. of water. A separate piece of wool is dyed in each bath. If more than one color is pres- ent a notable difference in the dyeing may be obtained. The following process for cochineal is due toGirard & Dupre:^ The material is dissolved in water if not already in solution, moderately acidulated with hydrochloric acid, and shaken out with amyl alcohol. If cochineal is present, the alcohol will be colored. It is separated, washed with water until neutral and divided into two portions. To one, a dilute solution of uranium acetate is added. Cochineal produces a characteristic emerald green. To the other portion is added a little ammonium hy- droxid. Cochineal gives a violet solution, but this reaction is not characteristic, as it is given by many fruit colors. See also pages 65 and 74, and the detection of carmine in meat, under ''Flesh Foods." COLORS 73 For the detection of colors used in egg substitutes, Winton & Bailey give a special scheme:^ The material is treated with 95 per cent, alcohol. A. The color dissolves. 1. Filter-paper is dipped in the solution, dried, moistened with a mixture of hydrochloric and boric acids and again dried. b. The color becomes cherry-red, changed to grayish-blue on addition of ammonium hydroxid. Turmeric. 2. The color is not affected by these reagents. a. The alcoholic solution on evaporation leaves a deposit soluble in water; the solution is partly decolorized by hydrochloric acid. Nitro-colors. b. The deposit from alcohol is soluble in water. True egg color. B . The yellow color is not soluble in the alcohol. I. The material is treated with a mixture of 90 per cent, of alcohol and 10 per cent, hydrochloric acid. It dissolves with an orange color. Filter-paper dipped in this and dried becomes rose-red on drying at room-tempera- ture. Azo-colors. The annexed synopsis of reactions of natural colors with some common color- reagents is from results obtained by La Wall from authentic samples. The ammonium hydroxid, hydrochloric acid and stannous chlorid are the ordinary laboratory solutions, added in small amounts to water-solutions of the color. The other reagents are : 1. Double dyeing as described on page 64. The figures refer to the order of the dyeing. "Amm" following abbreviation of a color-name, means the effect produced by ammonium hydroxid on the first dyeing. As a rule, second dyeing gave no noteworthy effect. 2. Good kaolin was shaken with a portion of the solution and the liquid filtered. 3. A piece of zinc was dropped into the hydrochloric acid solution of the color. 74 FOOD ANALYSIS t3 Q O r- H ac ^ OJ 2; w J r^sS 5§ ^ a «g ^ 55. •SP J < C^ 0^ C! ^v :d ^ o ^ < ^ o% z PS 8? u- i-T Q w w § < w ^ s ^ ^ z h-l d P u < w |2 is 15 i o Oi LJU u 'd o Pi o &J K a .t_i CA) y <-> ^ o<; tJD Q 'C 4> a u 0) a u a o c?5 '^ S ^ bnbJD^L) Jh ^H o o ^ o o hn u CJ CJ a ^ 5:5^75 PPcJoQl^Q CO Q Q (U o bJDbJ0biDbJDbJDtiJ3W)W)b/}O W) bX)^> .> bJD.bf ar.W)_> cT CJ pq T3 PQ bJDO) bCtJDO) bJD (72 Td P 0) X! biD pq -s . CJ Oh Tl d (U <1) u 0) P^ « • tT • ^ ... CJ c75 J bO I ^11 si J I bcg- •S g ^ "5 a c nj ^ . 51 Q ft^ J rt o o Oh (U (L) Q Oh (u o bo u o :z; -^ bo bo C a o o bO C bo C cj u o "S c; ^ a ^ O) ^ 1^ ij^ Cj -S ^cJ" -o C/3 ;3 P U fe I ^1 -s ^ be o 2 o .bO I ■S bC .bp pq C i 2 c5? J5 u 76 FOOD ANALYSIS PRESERVATIVES The decomposition of food is prevented by sterilization or by addition of preservatives. Some preservatives — e. g., com- mon salt, niter, acetic acid, and wood smoke — have been known from early times and are still in vogue. Among the more important of the newer forms are salicylic acid, benzoic acid, sodium benzoate, beta-naphthol, saccharin, abrastol, formal- dehyde, fluorids, silicofluorids, sulfites, boric acid, and borax. Others, mostly synthetic coal-tar derivatives, have been sug- gested and, to a limited extent, used. Most acids are antiseptic. Each of the substances above named has special adap- tabilities; some of them are widely applicable, and hence are largely used. The permissible food-preservatives are not distinctly germicidal and must remain in the food if continued preservation is desired. Salicylic acid is a white crystalline powder, soluble in 500 parts of cold water, more freely in alcohol. Ether, petroleum spirit, chloroform and carbon tetrachlorid dissolve it readily and remove it from an acidified water-solution. It distils in a current of steam. Its most characteristic reaction is the violet produced by ferric salts. Salicylates exist normally in many vegetable substances; in a few in considerable amount, in many, such as common edible berries, in very small amounts, but still recognizable by delicate tests. Care must be taken therefore in interpreting the results of such tests. Sodium benzoate is usually sold as a granular white powder which has a slight aromatic odor and a nauseous taste. It is freely soluble in water. In the United States it is the usual preservative for catsups, jams, jellies, mince-meat," and pre- serves. . . Benzoic acid is not frequently used in food articles, but some of it may be formed from sodium benzoate by the action of acids or acid salts in the food. PRESERVATIVES 77 Saccharin. Commercial saccharin is somewhat variable in composition. It is a white, crystalHne, intensely sweet powder, soluble in looo parts of cold and loo parts of boiling water. It is more soluble in alcohol, glycerol, and ether, and very slightly soluble in chloroform, benzene, and petroleum spirit. Ether removes it from its aqueous solutions. Pure saccharin is slightly volatile at ioo° and leaves no ash, but impurities may be present in the form of sodium salts, and considerable ash, principally sodium sulfate, may be left upon ignition. ,3-naphthol is a white crystalHne powder, shghtly soluble in water, freely in alcohol, ether, chloroform, benzene, fats, and alkaline solutions. It is wholly volatile on ignition. It is liable to contain small amounts of a-naphthol. The so-called hydronaphthol is substantially the same as i9-naphthol. Ahrastol or asaprol (calcium /5-naphthol-a-monosulfonate) is a colorless or light reddish powder freely soluble in water and alcohol. In dilute solution in water it produces with a solution containing mercuric nitrate and nitric acid, a canary-yellow liquid. Stronger solutions produce a yellow precipitate. .Formaldehyde is a gas freely soluble in water, from which solution a polymeric modification is easily obtained as a white solid, volatilized only at a temperature above the boiling- point of water. Formaldehyde is principally sold as a 40 per cent, watery solution designated by the copyrighted name ''formahn." More dilute solutions are sold under a variety of fanciful and misleading names. The 40 per cent, solution is a colorless liquid with a slight, somewhat acrid odor and a faint acid reaction, the last property being probably due to small amounts of formic or acetic acid produced by oxida- tion. When this solution is boiled, most of the formaldehyde distils readily with the steam; but if the fresh distillate be evaporated at a lower temperature, — as, for example, on a shallow dish placed over boiling water, — a large part is con- verted into the solid form. All the modifications of formalde- 78 FOOD ANALYSIS hyde have active reducing qualities and exhibit strong tendency to combine with proteids so as to form insoluble bodies. In the preservation of food, the commercial formalin is almost ex- clusively used. Sulfites. The acid salts are more active than the neutral form and are more used. Calcium sulfite is also frequently employed. Sulfites are white solids freely soluble in water and glycerol, but not appreciably in alcohol, or the solvents immiscible with water. Their antiseptic action being strongly exerted upon yeast, they have been used largely to control or prevent alcoholic fermentation. The detection of sulfites being based upon the recognition of the sulfurous acid derived from them, a specific description of each will not be needed. Boric acid, Borax. A mixture of these is frequently sold under trade names, such as '' Preservaline " and " Rex Magnus." They are also used separately. Both are white powders soluble in water; borax is practically insoluble in alcohol, boric acid freely soluble. Both are non- volatile at a red heat, but a watery solution of boric ac* ' ■ tnnot be evaporated without considerable of the acid passing A with the steam. Borax has an alkaline reaction; boric acid is acid to litmus, but turns turmeric paper brown when its solution is evaporated on it. When boric acid is heated with glycerol, tritenyl borate is produced as a thick sirup miscible in all proportions with cold water and decomposed by hot water. By evaporation it can be obtained in the form of a transparent, glassy, brittle mass which absorbs water readily. A preparation made by dissolv- ing borax in glycerol has also been offered as a preservative, but is little used. These glycerol preparations have been sold under various names, such as ''boroglyceride" and "glyceride of boric acid." Borates are present in appreciable amount in many fruit- juices. FluoridSj borofluorids, and silicofluorids. The sodium, potas- PRESERVATIVES 79 slum and ammonium compounds, have been principally used, being among the few forms soluble in water. They are white powders, not volatile at a red heat except ammonium fluorid. The last has been sold under the riame "antisepticum." Detection of Preservatives. — Owing to the difference in the chemical character of preservatives and of the food articles in which they are used, few general methods can be given; the examination must be conducted with reference to the material likely to be present. The following are suggestions in this direction: In meats, boric acid and sulfites; in milk and milk products, formaldehyde and boric acid, occasionally salicylic acid. In jams, jellies, mince-meat, and table dehcacies, benzoic and salicylic acids or their salts; occasionally boric acid. In cider and some other fruit juices, salicylic acid and sulfites. In fermented beverages and malt extracts, salicylic acid, sulfites, fluorids, silicofluorids, borofluorids ; abrastol may be employed, but the data in regard to it are limited. Saccharin is likely to be present in beer*, wines, and sweetened articles. Several preservatives are easily extractt ' ' ^m food articles by shaking with ether which dissolves them. The solution should be slightly acid. If not, a little sulfuric acid should be added. If the extraction be repeated with several portions of the solvent an approximate quantitative determination may be made. The shaking must be vigorous, so as to bring the solvent in contact with all parts of the sample. In many cases this will produce an emulsion which separates very slowly. The application of the centrifugal method will be useful in this case. The addition of more of the solvent and the cooling of the material is also advised. The following descriptions are adapted especially to the conditions under which the different preservatives are likely to be found. As they are somewhat soluble i^ water, solid or semi-solid materials may be exhausted with vvater and the liquid concentrated at a low temperature. In many cases the 8o FOOD ANALYSIS sample may be strained through muslin and the tests applied to the filtrate. The volatility of some preservatives, especially in a current of steam, is occasionally serviceable. Formaldehyde may be thus obtained from milk. Benzoic acid, saccharin and sulfites may be separated by mixing about 200 grams of the sample with 5 c.c. of a 20 per cent, solution of phosphoric acid, and distiUing nearly to dryness. Benzoic and sulfurous acids distil, and the saccharin remains in the flask. Sulfuric acid may also be used. A current of steam through the distilKng flask is more efficient. Salicylic acid. This is usually detected by extraction with an immiscible solvent. 25 to 50 c.c. of the sample are rendered feebly acid with a few drops of sulfuric acid and shaken vigor- ously with about an equal bulk of a mixture of equal parts of ether and petroleum spirit, the liquids are allowed to separate, as much as possible of the solvent is drawn off, filtered, and evaporated at a gentle heat. When salicylic acid has been added as a preservative, distinct needle-like crystals will be usually seen. A few drops of water should be added and then a drop of very dilute ferric chlorid solution. The reaction of salicylic acid is distinct. When a crystalline deposit cannot be obtained, a larger quantity of the sample may be concentrated at a gentle heat and extracted as above. (See under * 'Alcoholic Beverages.") Some analysts prefer chloroform as the extracting liquid. In this case the shaking should be done in a stoppered sepa- rator, that the solvent may be readily drawn. A solution of ammonio-ferric alum is in some respects preferable to ferric chlorid as a testing agent. If 50 c.c. of the sample properly extracted does not give a visible deposit of the acid, it is not likely that it has been added as a preservative. Saccharin, A suitable amount of the sample (50 or 100 c.c.) is acidified with dilute (25 per cent.) sulfuric acid and extracted with a mixture of equal parts of petroleum spirit PRESERVATIVES 8l boiling below 60° and ether. The solvent is evaporated at a gentle heat. The presence of saccharin in the residue may be detected by the taste. 2 c.c. of a saturated solution of sodium hydroxid are added and the dish heated until the residue dries and then to 210-215°, and maintained thus for half an hour. The saccharin is converted into salicyHc acid, which may be detected in the residue by acidulating it with sulfuric acid and applying the ferric chlorid test. If salicylic acid be present originally in the sample, the residue from the petroleum spirit and ether solution is dissolved in 50 c.c. of dilute hydrochloric acid, bromin water added in excess, the liquid shaken well, and filtered. Sahcylic acid is completely removed as a bromi- nated derivative. The filtrate is made strongly alkaline with sodium hydroxid, evaporated, and fused as described above. A substance capable of giving a reaction by this method often exists in wine. For the elimination of this error, see under ** Alcoholic Beverages." Benzoic acid and benzoaies. Mohler's method: About 100 grams of the sample are made alkaline with sodium hydroxid and' evaporated to a paste, which is then acidified with hydro- chloric acid, mixed with sand, and extracted with ether. The ether is evaporated spontaneously, the residue moistened with 2 c.c. of sulfuric acid, heated until acid vapors escape (at about 240°), and a few decigrams of sodium nitrate added in small portions, until the liquid becomes colorless. The liquid is poured into excess of ammonium hydroxid and a drop of ammonium sulfid solution added. Benzoic acid is indicated by a yellow, changing to reddish-brown. Peter's method: The material is made slightly acid and ex- tracted with chloroform, which is then evaporated sponta- neously. The vessel containing the residue is placed in melting ice, 2 c.c. of sulfuric acid added, and stirred until the residue is dissolved. Barium dioxid is dusted into the mass, with con- stant stirring, until the liquid begins to foam, when 3 c.c. of 82 FOOD ANALYSIS hydrogen dioxid (3 per cent.) are added drop by drop. The dish is then removed from the cold bath, the contents diluted with water to convenient bulk, and filtered. The acid filtrate is extracted with chloroform. The benzoic acid will have been converted into salicylic acid by the process and the latter may be detected by dilute solution of ferric chlorid or ammonio-ferric sulfate. Boric acid and borax. These may be detected in many food-articles, especially milk and milk products, by the follow- ing test: A few drops of the sample or of a solution obtained by shaking some of it in water are mixed with a drop of strong hydrochloric acid and a drop of strong alcoholic solution of turmeric, evaporated to dryness at a gentle heat, and a drop of ammonium hydroxid added to the residue when cold. A dull green stain shows that boric acid is present. Borates being normal constituents of many fruits, quali- tative tests are not sufficient to determine if the preservative has been added. For methods of quantitative determination, see under "Alcoholic Beverages." Fluorids. 100 grams of the sample are made slightly alka- line with ammonium carbonate, heated to boiling, a few centi- meters of calcium chlorid solution added, and heating con- tinued for 5 minutes. The precipitate is collected, washed, dried, transferred to a platinum crucible, and ignited. When the mass is cold, a few drops of strong sulfuric acid are added, and the crucible covered with a piece of glass partly protected on the lower side by paraffin. The bottom of the crucible is then heated for an hour at a temperature between 75° and 80°. The glass is etched if fluorids are present. Borofluorids and silico fluorids. 200 grams of the sample are made alkaline with calcium hydroxid solution, .evaporated to dryness, incinerated, and the ash extracted with sufficient acetic acid to decompose carbonates. The residue is col- lected on a filter, washed, again extracted with acetic acid. PRESERVATIVES 83 and filtered. The filtrate contains any boric acid that may be present and is tested for this substance as directed on page 82. The insoluble residue contains the calcium silicate and calcium fluroid. The filter and residue are ashed, a portion of the mass mixed with a little precipitated silica and 2 c.c. of sulfuric acid, and placed in a short test-tube to which is attached a small U-tube containing a few drops of water. The test-tube is heated cautiously in a water-bath; any sili- con fluorid that may be formed from fluorin present will pro- duce a gelatinous deposit in the U-tube. If boric acid has been found in the filtrate noted above, it may be assumed that any fluorin is in the form of borofluorid; but if boric acid is not present, the other portion of the ash from the filter and residue is treated with sulfuric acid without previous addition of silica. If gelatinous silicic acid be formed, the compound was originally siHcofluorid. Formaldehyde. The tests for formaldehyde have been mostly adapted to its detection in milk. One of the most delicate and positive reactions of formalde- hyde is as follows: To a few c.c. of the suspected liquid, a pinch of phenylhydrazin hydrochlorid is added, the liquid shaken and a drop of a dilute solution of sodium nitroprussid added and then a few drops of sodium hydroxid. A deep blue color is at once produced with formaldehyde. The nitro- prussid solution should be fresh. The test is applicable to milk, but the color is grayish-green. Another test is the addition of a small amount of a solution of I per cent, of phloroglucol and about 25 per cent, of sodium hydroxid in water. This produces a rose-red. The test is best applied by running the test solution by means of a pipet under the suspected liquid. Formaldehyde may be obtained pure by distillation of the sample, especially in a current of steam. An investigation by Leonard, H. M. Smith, & Richmond showed that with or- 84 FOOD ANALYSIS dinary aqueous solutions, about 30 per cent, of the formalde- hyde has passed over when 20 per cent, of the liquid has been distilled, and nearly 50 per cent, when 40 per cent, of the liquid has been distilled. A larger proportion distils if sulfuric acid be added to the liquid. For details of this and for other tests for formaldehyde, see under ''Milk." Determination oj Formaldehyde. B. H. Smith,^ who also in- vestigated the methods for this purpose, finds that the choice will depend on the strength of the solution. For moderately strong solutions the iodin method of Romijn is satisfactory. 10 c.c. of the solution, which should be diluted so as not to contain more than 3 per cent of formaldehyde, are mixed with 25 c.c. ^ iodin solution and sufficient strong sodium hydroxid solution added to make the liquid bright yellow. After stand- ing 10 minutes, hydrochloric acid is slowly added until a marked brown liquid is produced. The iodin is then titrated with thiosulfate in the usual way. The amount of iodin that has been taken up, multiplied by 0.118, will give the amount of formaldehyde. A blank experiment should be made and any necessary correction appHed. For dilute solutions, the potassium cyanid method is best. 30 c.c. of ^ silver nitrate solution are acidulated with 15 drops of nitric acid. 10 c.c. of this solution are mixed with 10 c.c. of normal potassium cyanid solution (6.5 grams in 1000 c.c), then water to make 50 c.c, the liquid shaken, filtered through a dry filter and 25 c.c. set apart for titration as below (Volhard's method). Another 10 c.c of cyanid solution are mixed with a measured amount of the formaldehyde solution (which must not contain more than 0.03 gram of formaldehyde), the mixture added to another 10 c.c of the acid silver nitrate solution, shaken, made up to 50 c.c, filtered and 25 c.c. of the filtrate taken as before. The two solutions contain excess of silver, but the second con- tains more, because the formaldehyde converts the cyanid into a compound that does not precipitate silver. PRESERVATIVES 85 Standard thiocyanate solution is prepared by dissolving 10 grams of potassium thiocyanate (or 8 grams of ammonium thiocyanate) in water to make 1000 c.c. The solution is ap- proximately -7^. Its value in silver must be determined thus : 50 c.c. of -7^ silver nitrate are mixed with i c.c. of nitric acid and i c.c. of saturated solution of ammonium ferric sul- fate, and thiocyanate solution added until a faint permanent brown is produced. The titration of the acid filtrates is conducted in the same manner. To each filtrate is added i c.c. of ferric sulfate and then the thiocyanate until the faint permanent brown is ob- tained. If the thiocyanate is exactly -^, the difference in c.c. required for the two filtrates multiplied by 0.006 will give the amount of formaldehyde in the quantity originally taken. If the thiocyanate is not -^ the result must be reduced to that basis. For detection of sulfites see under "Alcoholic Beverages." 13-naphthol. Several allied antiseptics of this type may be detected by the following method: 200 grams of the sample are acidified with sulfuric acid and distilled with open steam until 150 c.c. of distillate are obtained. This liquid is shaken with 20 c.c. of chloroform, the latter withdrawn, rendered alkaline with potassium hydroxid, and heated almost to boiling for a few minutes. Color changes as follows : Salol, light red. Phenol, light red, to brown, to colorless. /?-naphthol, deep blue, to green, to brown. A portion of the distillate may also be tested as follows : 2 5 c.c. are made faintly alkaline with ammonium hydroxid, then faintly acid with nitric acid and then a drop of strong sodium nitrite solution. /3-naphthol develops a rose red, but the reaction is sometimes uncertain and seems to be affected by light. The so-called hydronaphthol gives the same effect. 86 FOOD ANALYSIS Ahrastol (Asaprol). A characteristic reaction for abrastol is that described by Pintus'^ ; the yellow produced by acid mercuric nitrate solution prepared as directed for the clarification of milk (see under Milk). It can be extracted from jellies, fruit juices, wines and similar articles by acidulation with dilute sulfuric acid and agitation with ether, petroleum spirit, chloroform or carbon tetrachlorid. On adding to the immiscible solvent a small amount of mercuric nitrate solution and shaking the liquids for a few seconds, the watery liquid will become yellow, rapidly changing to bright red.^ The following method, devised by Sinabaldi, especially for wine, is applicable to other food-articles. 50 c.c. of the sample are made alkaline by cautious addition of ammonium hydroxid, shaken gently for two minutes with amyl alcohol, and the liquids allowed to separate. If this does not occur a little common alcohol should be added. The amyl alcohol is decanted, filtered if turbid, and evaporated to dryness. The residue is thoroughly mixed with a mixture of i c.c. of nitric acid and i c.c. of water, heated on the water-bath until half of the liquid is evaporated, transferred to a test-tube by the aid of I c.c. of water, 0.2 gram of ferrous sulfate added and then ammonium hydroxid to excess with constant shaking. If the resulting precipitate is reddish, it is dissolved in a few drops of sulfuric acid and treated with ferrous sulfate and ammonium hydroxid as before. As soon as a dark greenish precipitate has been obtained, it is dissolved in sulfuric acid, the Hquid well shaken and filtered. In the absence of abrastol the filtrate is light yellow, with abrastol in appreciable amount it is red. SPECIAL METHODS STARCH Detection. The reaction with iodin affords a delicate method for detect- ing starch. The color is shown by undissolved granules, but it is more satisfactory to dissolve it by boiling with water, allowing the solution to cool and adding the iodin, preferably as potassium iodid-iodin solution (p. 26). If the proportion of. starch be large, an almost black precipitate will be formed. The depth of color will be some indication of the amount present, but exact determinations cannot be made by this method. In the undissolved condition, starch may be recognized by the microscope and its source often determined. A magnify- ing power of from 150 to 300 diameters will be required. The characteristics of the granules are seen more vividly by mount- ing them in a dense medium such as chloral hydrate solution or glycerol (p. 26) and arranging the reflecting mirror so as to throw an oblique light upon the object. By this means distinct markings, termed hilum and concentric rings, are recognized. If the chloral-hydrate iodin solution (p. 26) be employed for mounting, or if a drop of the potassium iodid-iodin solution be introduced under the cover of a glycerol- or water-mounting, the granules will become blue. With polarized light, many starches show on the dark field — i. e., with crossed nicols — dark bands radiating from the hilum, giving the appearance of a Maltese cross. For this examina- tion the object is mounted uncolored in one of the denser media and the light thrown directly from below. By inserting a selenite 87 88 FOOD ANALYSIS plate between the object and the lower nicol, colors will be produced with many starches. Muter employed a selenite giving a green field, but red and red- violet fields are also suitable. The successful application of these methods requires good apparatus and considerable practice. A careful study of starch- granules of authentic origin should always be made before deciding as to the nature of any specimen. The size, appearance and effect on polarized light may be. much altered by heating starch, and possibly by some other manufacturing operations. A synopsis of the characters of the principal starches is pre- sented in the annexed tables. A micron (o.ooi miUimeter) may be converted into thousandths of an inch by multiplying by 0.03937. The factor 0.04 will be near enough for most cases. The classification is essentially that of Muter, the basis being the predominating form of the granule, the distinctness and position of the hilum and markings, the appearance under polarized Hght, with or without selenite plate. Muter indi- cated five groups, each group designated by the name of an important type of starch, as follows : Potato Group. — Oval or ovate granules, showing hilum and concentric rings clearly, cross and colors usually distinct. Legume Group. — Round or oval granules, hilum marked, rings faint, but rendered visible in cases by chromic acid solu- tion, cross and colors feeble. Wheat Group. — Round or oval granules, hilum and rings generally invisible, feebly-marked cross and colors. Sago Group. — ^Truncated granules, hilum distinct, faint rings, cross and colors fairly marked. Rice Group. — Polygonal granules, hilum distinct, rings faint, cross and colors usually faint. In the description of individual starches, the term "eccen- tric" denotes that the hilum is not in the apparent center of the granule. The granule is often described as oval, circular STARCH 89 or polygonal, terms which are strictly applicable to surfaces. It will be understood, therefore, that such terms refer to the apparent cross-section of the granule as it is usually viewed. The dimensions given must be regarded as the most frequent; granules not included within the limits will often be found. Polarized hght is affected to some extent by almost all starch granules, if very close observation is made. Source. Size in Microns. Potato, 60-100 Canna, 45-135 Maranta, 10-70 Natal arrow- root, 35-40 30-60 Turmeric, Ginger, 40 Mother-cloves, 20-60 Banana, 40-80 General Character or Granules. Smaller granules round, large ones ovate; hi- lum a spot, eccentric; rings numerous and complete. Irregular ovate; hilum annular, eccentric; rings incomplete, narrow and regular. Ovate; hilum eccen- tric, circular or linear, often cracked; rings numerous, not very distinct; sometimes a projection at one end. Ovate to circular, ir- regular projections; hilum eccentric, cracked; rings dis- tinct. Ovate, often much nar- rowed at one end; hilum eccentric, dot- like; rings indistinct. Ovate, many with a projection on one end; hilum and rings scarcely visible. Ovate; hilum a dis- tinct spot, eccentric; rings visible. Ovate but often very narrow in proportion to length; hilum a spot, eccentric; rings distinct. With Polarizer. Without Selenite. Well-marked . cross. Well-marked cross. Well-marked cross. Well-marked cross. Well-marked cross. Faint Well-marked cross. Faint cross. With Selenite. Well-marked colors. Well-marked colors. Well-marked colors. Well-marked colors. Well-marked colors. Faint colors. Well-marked colors. Faint colors. go FOOD ANALYSIS With Polarizer. Size in Microns. General Character OF Granules. Source. Without Selenite. With Selenite. Bean, 3S Reniform or ovate; Cross indis- Colors very hilum stellate or fur- tinct. faint. row-like; rings very faint. Pea, 15-30 Reniform or ovate; hi- lum elongated; rings Cross indis- tinct. Colors very faint. very faint. Lentil, 30 Reniform or ovate ; Cross indis- Colors very hilum elongated, dis- tinct. faint. tinct; rings visible. Nutmeg, 5-50 Rounded, collected in groups of two to four; lilum stellate; rings invisible. Cross faint. Colors very faint. Wheat, 2-50 Mostly roundish, chief- Cross not well Colors very ly the smallest and marked. faint. largest sizes present; hilum indistinct, near- ly central; rings in- distinct. Barley, 15-40 Resembles wheat but Cross not well Colors very some granules slightly marked. faint. angular or elliptical; rings more distinct than wheat. Rve 20-60 Resembles wheat; hi- lum distinct, stellate; Cross not well marked. Colors verv ^^j ^> faint. rings often visible. Distorted forms not infrequently occur. Dhoura, 1-3 Round, hilum faint. Cross. Colors. 12-33 Round; no hilum. Cross faint. Colors faint. Acorn, 20 Round or nearly so; Cross not well Colors not hilum eccentric. marked. well marked. Cacao, 5-10 Round; hilum and Cross not well Colors not rings indistinct. marked. well marked. Saffo, 25-66 Ovate, truncated; hi- lum a circle or spot; Well-marked cross. Well-marked KJCVgW, colors. rings faint. Prepared sago. Characters less distinct than in raw sago. Tapioca, 8-22 Circular; hilum a slit, Well-marked Well-marked nearly central. cross. colors. Prepared tapi- oca, Characters less distinct than in raw form. STARCH 91 With Polarizer. Source. Size in General Character Microns. OF Granules. Without Selenite. With Selenite. Cinnamon, 8-20 Truncated at one end, Well-marked Well-marked two to four granules cross. colors. often joined; hilum distinct, nearly cen- tral; rings invisible Rice, 5-10 Pentagonal, hexagonal, occasionally triangu- Cross distinct. Colors dis- well marked. tinct. lar with sharp angles; hilum distinct under high power. Buckwheat, . . . 5-20 Polygonal, angles Cross d i s - Colors dis- somewhat rounded; tinct. tinct. hilum central, spot or star; granules often compound. Oat, 5-30 Mostly polygonal, a few spherical; hilum Faint cross. Faint colors. and rings visible only with high power; often compound. Maize, 5-20 Round to polygonal, angles usua ly round- ed; hilum central; crack or star; rings nearly invisible. Faint cross. Faint colors. Pepper, 0-5 -5 Polygonal, very small. Cross with Color with sometimes showing high power. high power. Brownian movement, sometimes united into large irregular masses; hilum only seen with high power. According to Lintner* potato-starch becomes pasty suddenly at 62-64°; cereal starches become pasty gradually at from 80-85°. Diastase acts on ungelatinized cereal starches at comparatively low temperatures ; ungelatinized potato-starch is hydrolyzed only at a comparatively high temperature. Barley. Pea. Potato. Oat. Maize. Rice. Bean. Wheat. Rye. Arrowroot. Buckwheat. 92 STARCH 93 Determination. The exact quantitative determination of starch is difficult. The proposed methods have been carefully investigated by Wiley & Krug, who have shown that in the presence of vegetable tissue containing pentosans or similar carbohydrates the diastase method is alone trustworthy. The first method is applicable to assaying commercial starches. Hydrochloric Acid Method. — ^3 grams of the substance are treated with about 50 c.c. of cold water for an hour, with frequent stirring ; the residue is collected on a filter and washed with sufficient water to make a total of 250 c.c. This Hquid contains the soluble carbohydrates. The undissolved residue is heated for 2J hours with 2.5 per cent, hydrochloric acid (200 c.c. water and 20 c.c. hydrochloric acid, sp. gr. 1.125) in a flask provided with an inverted condenser, cooled, neutralized with sodium carbonate, made up to 250 c.c, filtered, and the dextrose determined in an aliquot portion of the filtrate. The weight of dextrose multiphed by 0.9 gives the weight of starch. Diastase Method. — 3 grams of the finely-powdered sub- stance are extracted on a hardened filter with five successive portions of 10 c.c. of ether, washed with 150 c.c. of a 10 per cent, alcohol, and then with a little strong alcohol. The residue is mixed in a beaker with 50 c.c. of water. The beaker is immersed in boihng water, the contents stirred constantly until all the starch is gelatinized, cooled to 55°, and 20 c.c. of malt-extract added. The liquid is maintained at 55° for i hour, heated again, boihng for a few minutes, cooled to 55°, 20 c.c. of malt-extract added and maintained at 55° until a micro- scopic examination of the residue shows no starch with iodin. It is cooled and made up directly to 250 c.c. and filtered. 200 c.c. of the filtrate are placed in a flask with 20 c.c. of a 25 per cent, solution of hydrochloric acid (sp. gr. 1.125), connected with a reflux condenser, and heated in boiling water for 2 J hours. It 94 FOOD ANALYSIS is nearly neutralized, while hot, with sodium carbonate, made up to 500 c.c, mixed, poured through a dry filter, and the dextrose determined in an aliquot part. Calculate the dextrose to starch by multiplying by 0.9. Preparation of Malt Extract. — 10 grams of fresh, finely ground malt are macerated overnight at about 25° with 200 c.c. of water, filtered, the amount of dextrose in a given quantity of the filtrate after boihng with acid determined as in the starch determination, and the proper correction noted. If diastase - be used, a correction will be unnecessary. A good diastase is now easily obtainable. Commercial malt extracts are liable to be destitute of diastatic power. In the application of the diastatic method, the material must be ground very fine and the preHminary extraction with ether must not be omitted. In many cases it will be more convenient to make the extraction in the continuous extractor. If a large tube is used, several samples may be treated at once by tying each in filter-paper. The centrifugal apparatus may also be used. The fine material is shaken up with ether in the proper tubes, whirled for a short time, the ether poured off, fresh ether added and again whirled, and the operation repeated until the necessary amount of solvent has been used. The Hquid may be poured off closely each time. Extraction with carbon tetrachlorid may be better, but the result may not be equivalent to that with ether. FLOURS AND MEALS Meal is coarsely ground, flour is finely ground material. Most of the forms used as foods are derived from plants be- longing to the order Graminece, but buckwheat, banana, and potato are not of this order. The distinction between the different flours and meals is based in part on the microscopic characters of the starches as indicated under that head, but chemical tests are in some cases available. STARCH 95 The commercial value of wheat flour depends upon its color and texture and upon quantity and quality of gluten. The latter differs much in different varieties and in the same variety grown in different localities. In whole-wheat flour containing about lo per cent, of gluten the quantities of the chief proteids are about as follows : Globulin, 0.70 Albumin, 0.40 Proteose, 0.30 Gliadin, 4.25 Glutenin, 4.35 Good wheat flour will yield from 20 to 40 per cent, moist gluten and 10 to 18 per cent, gluten dried at 100°. Rye flour contains gliadin, but no glutenin. COMPOSITION OF CEREAL GRAINS Car- bohy- drates Weight other OF 100 THAN Kernels Moist- 6.25 Ether Crude Crude IN Grams, uke. N. Extract, Fiber. Ash. Fiber. Typical u n h u 1 1 e d barley, 10.85 ii.o 2.25 385 2.5 69.55 Typical American maize, 38.0 10.75 lo.o 4.25 1.75 1.5 71-75 Typical wheat, . . . 3.85 10.6 12.25 1-75 2.4 1.75 71.25 Sweet corn, 19 sam- ples (Richardson), 8.44 11.48 8.57 2.82 1.97 66.72 Typical American buckwheat 3.0 12.0 10.75 2.0 10.75 1.75 62 75 Typical rye 2.5 10.5 12.25 i-5 2.1 1.9 71.75 Typical unhulled oats, 30 100 12.0 4.5 12.0 3.4 58.0 Typical rice, un- hulled 3.0 10.5 7.5 1.6 9.0 4.0 67.4 Typical rice, hulled, but unpolished, . . 2.5 12.0 8.0 2.0 i.o i.o 76.0 Typical rice, pol- ished, 2.2 12.4 7.5 0.4 0.4 0.5 78.8 Typical rye 2.5 10.5 12.25 i-5 2.1 1.9 71.7 Typical wheat, . . . 3.85 10.6 12.25 1.75 2.4 1.75 7«-25 A detailed description of the proteid and other constituents of cereal grains has been published by the United States De- 96 FOOD ANALYSIS partment of Agriculture. The table on page 95 has been taken from this. The proteids are calculated by multiplying the nitrogen by the factor 6.25, but the investigations by Osborne, Chittenden, and Voorhees indicate that the following factors would be better: Maize, 6.23; barley, rye, and wheat, each 5.68; oats, 6.10. A recalculation of the proteids by corrected factors will change the proportions of the carbohydrates, since these were determined by difference. Wheat Flour. — Good wheat flour is a fine white powder with a very faint yellow tinge. Several tests are recognized for its examination, among which are the following : Color Test. — The sample may be compared with one of known quality by laying out heaps of equal size, say, 3 cm. by 8 cm., and 0.5 cm. deep. If this be done on a colorless glass plate, the examination may be made with both white and colored background, and the plate may subsequently be immersed in water (not over 35°) so that the colors produced on wetting may also be observed. Doughing Test. — This consists in making a dough with 15 grams of the sample and 10 c.c. of water and comparing color, firmness, elasticity, and compactness. Gluten Test. — 10 grams of the sample are mixed with suf- ficient water to make a stiff dough and allowed to stand for one hour. The mass is kneaded in a piece of linen in running water until the washings are clear. The fresh gluten thus obtained should have a faint yellow tinge, be tough and of such consistency that it can be pulled out into threads. Gray and red glutens indicate inferior samples. Good gluten swells at 150° and assumes the appearance of bread. Adulterations. — Flour may be mixed with mineral matters to increase weight, with alum or copper sulfate to improve its appearance, or with cheaper flours or starches. It may also contain seeds of weeds, may be damp or decomposed, or may contain fungi. STARCH 97 In examining for these adulterations, determintions of ash, crude fiber, ether extract, and total nitrogen are of consider- able value. The following table gives some data on these points, but the limits must not be rigidly interpreted. The figures, except the first column, have been calculated on the water- free substance: COMPOSITION OF FLOURS Wheat, . . . Rye Barley, . . . Buckwheat, Rice Oat (meal), Maize (meal), Graham, . . Moisture. Max. Mill, 150 J4.0 150 18.0 15.0 10.0 18.0 150 9.0 12.0 lO.O 12.5 lO.O 6.0 8.0 1 1.0 Ash. Max. Min. 0.3 0.5 1.0 0.8 0.3 2.0 i.o 1.8 6.25 N. Max. Min. 15.0 1 1.0 12.0 9-5 lo.o 18.0 "•5 1.5.0 8.0 6.0 8.5 .SO 7.0 14.0 8.0 lO.O Fiber. Ether Extract. Max. Min.iMax. Min 1.0 0.6 0.6 0.6 0.4 1-4 3.5 2.4 0.1 2.0 0.4 0.3 0.3 0.1 0.7 0.7 2.0 1.0 2.0 2.0 0.6 9-5 6.0 2.2 05 0.9 0.5 0.8 0.3 6.5 2.5 1-9 N-FREE EXIRACT. Max. Min. 82.0 88.0 87.0 84.0 85.0 72.0 63.0 70,0 Alum. Logwood Method. — An alkaline solution of logwood is pre- pared as follows : Half a gram of fine logwood chips, preferably freshly cut from the log, is macerated for 10 hours in 15 c.c. of alcohol; 10 c.c. of the solution are poured off and mixed with 150 c.c. of water and 10 c.c. of a saturated solution of am- monium carbonate. To make the test, 50 grams of the flour are made into a thin paste with water, a few drops of the log- wood solution (freshly prepared) added, and the mixture allowed to stand several hours. Alum produces a lavender- blue lake. Chloroform Method. — 200 grams of flour are shaken in a separatory funnel with a sufficient amount of chloroform, allowed to stand overnight, and the materials which subside carefully removed through the stopcock. This material may be further purified by shaking a second time with a little chlo- roform and then transferred to a watch-glass and the chloro- form evaporated. The residue is treated with water, the solu- 98 FOOD ANALYSIS tion separated from the insoluble portion and allowed to evaporate, when the crystals of alum will be observed. The crystals may be dissolved in water and tested for sul- fates, aluminum, potassium, and ammonium. The residue insoluble in water should be examined under the microscope for mineral matters. The steps in the treatment of the residue insoluble in chloroform will be assisted by the use of a centrifuge. Copper sulfate can be detected by the ferrocyanid method as described under Bread. Ergot in Rye Flour. — A preliminary test may be made to determine if the flour has been damaged by fungi. Vogel advises that the sample be stained with anilin violet and exam- ined with the microscope. Any starch granules that have been injured by fungus will be deeply stained. Gruher^s test: A little of the flour is moistened with water on a microscope-slide, a cover- glass placed on, and the mass heated to the boiling-point on a hot plate or water-bath. After cooling it is examined with a power of 120 diameters. Ergot will be recognized by its high refracting power, furrows, and color-7-deep violet on the edge, greenish-yellow within. A second examination with a power of about 300 diameters will enable any doubtful particles to be recognized. Chemical Tests. — 200 grams of the sample are digested with boiling alcohol as long as any color is extracted. The solu- tion is treated with i c.c. of sulfuric acid (1:3). In the presence of ergot the solution will be red, and if it be diluted with a large volume of water, the color may be extracted from separate portions by means of chloroform, ether, petroleum spirit, or amyl alcohol. 10 grams of the sample are macerated for about 30 minutes with a mixture of 20 c.c. of ether and 10 drops of dilute sulfuric acid (i : 5); the liquid filtered, washed with ether until the filtrate amounts to 15 c.c. This is shaken with 5 drops of a ^ STARCH 99 saturated solution of sodium bicarbonate. The chlorophyl remains in the ether; the sodium bicarbonate solution remains clear if the flour be from sound grain, but takes on a deep violet color if ergot be present. Mixed Flours. — The following data are taken, with but few changes, from the contributions of Bigelow & Sweetser and Kraemer: Gluten obtained from a mixture of wheat and rye flours is dark and viscous, without homogeneity; from a mixture of wheat and barley flours, dark, non-viscous, and dirty reddish- brown; from a mixture of wheat and oats, dark yellow; from a mixture of wheat and maize, yellowish and non-elastic; from a mixture of wheat and leguminous flour it varies from a grayish-red, in the case of vetch or beans, to green, in the case of peas, and has the characteristic odor and taste of leguminous products. The ash of leguminous flour is deli- quescent, high in chlorids, and turns turmeric paper brown; cereal ash is the reverse. The aqueous extract of the legu- minous flour is acid; that of cereal flour is faintly alkaHne. If the filtrate from the gluten determination of flour contain- ing leguminous flour be made alkahne with ammoniurji hy- droxid, allowed to stand overnight, and the clear liquid de- canted, dilute sulfuric acid will precipitate legumin. For the detection of potato flour a portion of the sample is rubbed in a mortar until a stiff paste is obtained, thinned with more water, filtered, and the clear filtrate tested with a drop of a dilute solution of iodin. Potato flour produces a deep blue, while with pure wheat flour the result is yellow or light orange. If a mixture of cereal and potato flours be dried, spread in a thin layer on a glazed black surface, and examined with a lens, the potato is indicated by bright and glassy particles in the other- wise dull white substance. Vogel extracts the flour with 70 per cent, of alcohol, to which 5 per cent, of hydrochloric acid has been added. The lOO FOOD ANALYSIS extract is colorless if the flour consist only of wheat or rye, pale yellow if adulterated with barley or oats, orange yellow with pea flour, purple red if made from mildewed wheat, and blood red if made from ergotized wheat. Rice in Buckwheat Flour. — When pure buckwheat is mixed with water into a thin paste, the addition of calcium hydroxid produces a dark green, which becomes red when acidified with hydrochloric acid. Rice flour gives a yellow color with potas- sium hydroxid and white with hydrochloric acid. A mixture of buckwheat and rice flours made into paste is changed to a light green color by potassium hydroxid and becomes flesh- colored when acidified with hydrochloric acid. Wheat in Rye Flour. — Kleeburg has advised the follow- ing test: A pinch of the sample is mixed on a small glass plate (a microscope-slide will serve) with water at about 45° in suflicient quantity that the particles of flour still float. The mixture is spread over a considerable part of the glass and a similar glass laid upon it so that about one-fourth of each glass protrudes at the ends. The two glasses are pressed together, the exuded liquid wiped off, and the glasses rubbed on e^ch other several times. If wheat flour be present, white spots will be observed, which will form threads on being rolled ; these are short and thin if the proportion of wheat be small, and thicker and longer with larger amounts. An admixture of 5 per cent, of wheat flour with rye is said to be thus recogniz- able. Maize in Wheat Flour. — Kraemer has devised the follow- ing test, which, he states, will detect 5 per cent, of maize in wheat flour: i gram of the sample is mixed with 15 c.c. of good glycerol and heated to boiling for a few minutes. An odor recalling that of popcorn indicates maize. It is alleged that cheap flours have been adulterated with sawdust. G. A. LeRoy applied the following test for detect- ing this addition: A small amount of the sample is gently STARCH lOI warmed with the acid solution of phloroglucol (page 26). Ordinary wood-fiber quickly acquires a bright red tint, while bran particles are but slightly affected. BREAD Bread is made by baking the mass obtained by kneading flour with water. This gives the so-called unleavened bread, but it is usual to add a little common salt to the water and make the dough light by inflating it with carbon dioxid. This may be done by the use of baking powder, or by mixing the flour with water containing carbonic acid under pressure (aer- ated bread), but commonly yeast is added to the dough and the mixture, called the ** sponge," allowed to stand for some hours and then baked. The slight fermentation which occurs liber- ates carbon dioxid. The chemical composition of bread is approximately that of the flour from which it is made. The moisture usually ranges from 30 to 40 per cent., and will depend, among other condi- tions, upon the quantity and quality of the gluten, and the size and shape of the loaf. On the size and shape will also depend the relative proportion of crust to crumb, the latter containing about twice as much moisture as the former. The addition of potato flour or rice flour will enable a bread to be prepared containing a much larger proportion of water than usual. The addition of about i per cent, of mashed potatoes to the dough is said to render the bread white without any notable increase in the amount of moisture retained. The proportion of fat in bread, as determined by the ether extract, is apt to be less than that of the original flour, owing to decomposition of the fat in the crust, by heat, and also to the inclosure of the fat particles in such a way as to render them difficult of extraction. On the other hand, the propor- tion of fatty matter may be increased by the use of milk or by the material used to grease the pans. 102 FOOD ANALYSIS When bread is raised by yeast, some solid matter is lost by the fermentation. According to Lawes and Gilbert, this is prob- ably less than J of i per cent., and appears to be due to the decomposition of the sugar. The unchanged starch is not appreciably altered during the short time that the yeast acts. The ash of bread will be higher than that of the flour if salt or baking powder has been added. Orig- inal Sub- stance Moist- ure. Vienna, average of lo sam- ples, Home-made, average of 2 samples, Graham, average of 9 sam- ples, . Rye, average of 7 samples, Quaker, average of 3 sam- ples, Miscellaneous, average of 9 samples, . ... 38.71 3302 34-8 3342 36.16 3441 Pio- teids, NX 5- 70. 10.8 2.51 ;i.86 10.59 In the Dry Substance. Ether Ex- tract. Crude Fiber. Ash. Salt. 1-73 2.91 1.02 2.21 0.97 0.36 1.74 095 0.41 0.46 1-95 1-55 2.29 2.79 1.68 1-53 0-93 0.84 1.07 1-5 0.92 0.76 Carbo- drates, exclud- ing Fiber. 83.1 84.75 82.06 84.36 85.41 85.66 The table represents the average composition of various breads of commerce according to analyses published by the Department of Agriculture. The loaves weighed approximately one pound each. Trade names are given in most cases. Adulterations. — These may consist in the use of damaged flour, of flours other than that purporting to be present, presence of excess of water, or addition of alum or copper sulfate to improve appearance. Alum.^Th^ bread is moistened with water and then with some of the alkaline logwood solution (see p. 97). If alum be STARCH 103 present, the bread will become lavender-blue in an hour or two. Pure bread assumes a light red-brown tint. The blue color, however, is not proof of the presence of alum unless it is permanent at the temperature of boiling water. Blyth gives the following test: 150 grams of the material are macerated for two days in 2 liters of water. The- solution is strained through muslin and evaporated at a gentle heat to small volume; a strip of gelatin immersed in this liquid, and then in the alkaline logwood solution, will acquire the lavender color if alum is present to the extent of 0.03 per cent. These tests are not applicable to sour bread. Vander- planken recommends the following modification to meet the difficulty: 15 grams of the sample are triturated to a paste with water and some pure sodium chlorid and 10 drops of a freshly-prepared solution of logwood in alcohol, and then 5 grams of pure potassium carbonate are added. The mass is well mixed, washed with 100 c.c. of water into a beaker, and is allowed to settle. In a few minutes the liquid becomes reddish-violet if alum is absent, grayish-blue to deep blue when it is present. The quantitative estimation of the alum is made as follows: The ash from 100 grams or more of the bread is boiled with hydrochloric acid and the solution filtered. The filtrate is boiled and added to a concentrated solution of sodium hy- droxid, the mixture being again boiled and filtered while hot. A little disodium acid phosphate is added to the filtrate, which is then slightly acidulated with hydrochloric acid and finally made feebly alkaline by addition of ammonium hydroxid. The precipitate of aluminum phosphate is filtered, washed, ignited, and weighed. Flour contains a small proportion of aluminum, which, in the ash, is probably in the form of silicate. The amount of silica is approximately equal to that of alum equiva- lent to the aluminum normally present. It is the practice, therefore, to determine the silica and subtract it from the I04 FOOD ANALYSIS amount of alum calculated from the aluminum phosphate found. The remainder, multiplied by 3.8 or 3.7, will give approximately the potassium alum or ammonium alum respectively. Copper sulfate may be detected by the brown produced when a thin slice of bread is immersed in a dilute solution of potassium ferrocyanid. Foreign flours may be sought for by the microscope, but the starch granules are usually so altered by heat as to render identification impossible. For detection of maize in wheat bread and pastry, Ottolenghi proposes the following test based on the reaction of proteids pecuHar to maize as elucidated by Donard & Labbe.^ 100 grams of crumb are dried at 40°, powdered finely, treated with 500 c.c. of a 0.3 per cent, solution of potassium hydroxid for 12 hours, with frequent shaking. The liquid is strained through muslin, the residue again treated with the alkaline solution for 3 hours, after which the mass is poured on the muslin strainer and well pressed. The filtrate is evaporated below 70° to dryness, the residue broken up as finely as possible, transferred to a flask, mixed with 40 c.c. absolute iso-amyl alcohol, an inverted condenser is attached to the flask and the liquid boiled in an oil-bath for 6 hours. The solvent is filtered hot. If no maize is present, the yellowish-brown filtrate re- mains clear, but with maize it becomes turbid. The admixture of the filtrate with 3 volumes of pure benzene increases the tur- bidity if maize is present, but produces no effect if the original substance was pure wheat flour. The following adulterants are said to be employed abroad, but their use does not appear to have been attempted in this country : Soap is said to be used to render the bread light and soft. It is said to be added in solution containing emulsified oil. Terra alba and gypsum have been found; they are readily detected in the ash. LEAVENING MATERIALS 105 Stannous chlorid is a comrr|on constituent of ginger cake, to which it is added, with pota^ium carbonate, in order to give the product the color ordinarily produced by honey or mo- lasses. It is said to render a product made of poor flour and molasses of the same color as that produced by a good flour and honey. Tin may be detected as described on page 59. LEAVENING MATERIALS The yeast cakes sold for leavening purposes are usually mixtures of common yeast with potato starch. The study of yeast is practically limited to those connected with the fermentation industries. Cream of tartar and baking soda are commonly employed as leavening agents. Baking Soda, Sodium Acid Carbonate, is not subject to serious adulteration. Cream oj Tartar, Acid Potassium Tartrate, is frequently adulterated with starch, alum, acid calcium phosphate, calcium sulfate, and potassium acid sulfate. Many samples will be found to contain no tartrate, but merely a mixture of starch, calcium phosphate, and alum. For the detection of tartaric acid see under "Fruit Juices." If starch is present the sample should be treated with cold water for a while, filtered and the residue evaporated on the water bath and tested. Allen devised the following method for the examination of commercial cream of tartar : 1. 881 grams of the dried material are dissolved in hot water and titrated with t,- sodium hydroxid and phenolphthalein. If tartaric acid and acid sulfates are not present, each c.c. will represent i per cent, of acid potassium tartrate. 1. 881 grams of dried material are ignited for 10 minutes, the residue boiled with water, filtered, and washed. The filtrate is titrated with ^ hydrochloric acid and methyl-orange. Io6 FOOD ANALYSIS With pure tartrate the amount of acid consumed will be the same as that of the alkali in the first experiment. Each cubic centimeter of deficiency is equivalent to 0.36 per cent, calcium sulfate, or 0.72 per cent, acid potassium sulfate. If the amount of acid be more than equivalent to that of the alkali used in the former experiment, it suggests the presence of neutral tartrate, each cubic centimeter of excess representing 0.6 per cent, thereof. The amount of sulfate can be determined by precipitating with barium chlorid in the usual way. The residue is ignited, dissolved in 20 c.c. of -^ acid, filtered from any insoluble residue, and the filtrate titrated with -^ alkali. Each c.c. corresponds to 0.5 per cent, of calcium tartrate, or 0.36 per cent, of anhydrous calcium sulfate. The cream of tartar substitutes commonly sold contain starch, alum, and calcium phosphate. Starch can be detected by the iodin test and by the microscope. Quantitative examin- ation of such samples will be conducted as described under "Baking Powders." Baking Powders. — These contain acid sodium carbonate, some acid salt, e. g., acid potassium tartrate, acid calcium phosphate, or alum, with inert material, starch or flour, to prevent caking. Many powders contain both alum and acid calcium phosphate. The following methods for examining baking powders were published by Crampton : The value of baking powder depends on the gas liberated when it is mixed with water. The determination may be by the apparatus of Knorr (figure 34). The flask A holds the weighed portion of sample. The condenser D, attached by a ground joint, serves to condense the steam formed when the liquid in A is boiled. B contains either recently-boiled water or dilute sulfuric acid, according to whether the available carbon dioxid or total carbonates are to be determined. It has a soda-lime tube attached by a ground-joint to pre- LEAVENING MATERIALS 107 vent admission of carbon dioxid from the current of air which is drawn through the apparatus during the operation. The junction of this portion with the fiask should be by ground or fused joint. The evolved gas is dried in E by sulfuric acid and absorbed in F. Fig. 34. Available carbon dioxid, which gives the leavening power, is determined as follows: The flask A is dried thoroughly, a weighing tube is charged with about 2 grams of the powder, accurately weighed, the contents emptied into the flask, and the tube weighed again. The exact amount of powder taken Io8 FOOD ANALYSIS is thus known. Recently-boiled water is put into B, the apparatus connected tightly, and the water allowed to flow in slowly from B, the aspirator attached to G being put in opera- tion. When the effervescence in A has ceased, the liquid in it is boiled for a few seconds, the lamp removed, and aspiration through G continued for 15 minutes. The absorption ap- paratus F is weighed, and the increase represents carbon dioxid. Total carbonates are determined by substituting 10 c.c. dilute sulfuric acid for the water in B. Starch. — 5 grams are mixed in a flask with 200 c.c. of 4 per cent, hydrochloric acid. A condensing tube about i meter long is attached by means of a cork (an inverted condenser may be used) and the liquid boiled for 4 hours. The contents are cooled, rendered slightly alkaline by sodium hydroxid, and the dextrose determined as given, and multiplied by 0.9. For powders not containing appreciable amounts of alum, direct washing with water, and drying the residue, will often give determinations of sufficient accuracy. Since the residual liquid in properly-made baking powders is alkaline, due to slight excess of baking powder, the diastase method for starch may be applicable. The liquid should be filtered and the insoluble residue well washed. The aluminum hydroxid may interfere with this method. If flour be used as filler, which may be ascertained by inspection, the starch found may be roughly calculated to flour by the table on page 97. Aluminum and Phosphates. — McElroy devised the following method : 5 grams are charred in a platinum basin, mixed with strong nitric acid, and filtered into a 500 c.c. flask. The residue is washed slightly, the filter and residue returned to the basin, burned to whiteness, mixed with sodium carbon- ate, fused, and cooled. Nitric acid is added, the liquid evapo- rated to dryness, again acidified with nitric acid, and the whole mass washed into the 500 c.c. flask. The liquid is made up to the mark and filtered through a dry filter, 100 c.c. of the filtrate LEAVENING MATERIALS IO9 are nearly neutralized with ammoniurn hydroxid, ammonium nitrate and ammonium molybdatc solution added, the mass digested at a low heat for a few hours, and filtered. The filtrate contains the aluminum, which may be precipitated as hydroxid by adding ammonium hydroxid. The precipitate is dissolved in ammonium hydroxid and the phosphate deter- mined in the usual way. Calcium. — 5 grams are mixed in a 500 c.c. flask with 50 c.c. of water and 30 c.c. of strong hydrochloric acid, the mixture made up to the mark, shaken well, and allowed to settle. 50 c.c. are collected through a dry filter, nearly neutralized by ammonium hydroxid, acetic acid added in small amount, then ammonium acetate, and the liquid boiled. If any precipitate forms it should be removed. The clear liquid is precipitated by ammonium oxalate. Suljates. — 0.5 gram of the sample are digested with strong hydrochloric acid until everything has dissolved, the liquid is diluted considerably, brought to boiling, and precipitated with barium chlorid, taking care not to use a large excess. The precipitate is weighed in the usual manner. Ammonium Compounds. — These may be determined by adding to the water filtered from a known weight of the powder sufficient sodium carbonate to make it distinctly alkaline distilling until half the liquid has passed over and titrating the distillate with standard acid. The best commercial baking powders yield about 12 per cent, by weight of gas. 10 grams would, therefore, yield 1.2 grams, occupying at ordinary temperature about 600 c.c, which will be much increased in baking. Many powders yield much less gas. no FOOD ANALYSIS SUGARS Detection. Most of the tests for sugars except the phenylhydrazin, fer- mentation, and optic tests depend on their reducing effect. Su- crose possesses less reducing action than other common sugars, does not give any precipitate with phenylhydrazin, and is not directly fermentable. By the action of dilute acids or inver- tase (yeast-enzym) it is converted (hydrolyzed) to equal parts of dextrose and levulose, a change commonly termed "in- version," the mixture being known as "invert-sugar." This responds to all the above tests. Cobalt Nitrate Test. — Wiley has experimented with this method and has obtained satisfactory results. He describes it as follows: 5 c.c. of a 5 per cent, solution of cobaltous nitrate are well mixed with 15 c.c. of sugar solution, and 2 c.c. of a 50 per cent, solution of sodium hydroxid added. Sucrose gives an amethyst-violet solution, which is made somewhat more blue by boihng, but regains its color on cooling. Dextrose gives a turquoise-blue, which in the course of two hours passes into a pale green. A slight fiocculent precipitate is noticed in the. tube containing dextrose. Maltose and lactose act very much as dextrose, but in the end do not give so fine a green color. If the solution containing dextrose, lactose, or maltose be boiled, the original color is destroyed and a yellow-green color takes its place. In mixtures of dextrose and sucrose the sucrose coloration predominates — one part of sucrose in nine parts of dextrose can be distinguished. Impurities such as gum arable or dextrin should be removed by alcohol or lead subacetate before the application of the test. Dextrin may also be thrown out by treatment of the solution with barium hydroxid and ammoniacal lead acetate. The reaction may be applied to the detection of cane-sugar in wines after SUGARS III they are thoroughly decolorized by means of lead acetate and bone-black. Sucrose may be detected in fresh or condensed milk after the disturbing matters have been thrown out by lead acetate. Sucrose may be detected in honey. Phenylhydrazin Test. — Phenylhydrazin hydrochlorid is usu- ally employed. The commercial article is often contaminated with anihn hydrochlorid; It may be purified by solution in hot water, precipitation by strong hydrochloric acid, and recrys- tallization from hot water. For the test, o.i gram of the sample, about 0.2 gram phenyl- hydrazin hydrochlorid, and 0.3 gram of sodium acetate are dissolved in 5 c.c. of water and heated on the water-bath for some time. Sucrose forms no precipitate, but with many sugars crystalline compounds called osazones separate. Dextrose and levulose yield the same compound, which may be termed ''glucosazone." It crystallizes in needles melting at 204-205°, and reduces Fehling's solution. Maltosazone crystalhzes in plates that melt with decomposi- tion at 206°. Lactosazone crystallizes in prisms melting at 200°. Sucrose forms no osazone. After hydrolysis it yields glu- cosazone. Lactose^ after boiling with dilute sulfuric acid, yields a mixture of glucosazone and galactosazone. The latter is distinguished by its melting-point, 193°. Starch and dextrin^ after hydrolysis, yield maltosazone and glucosazone. Maltose and lactose produce with ammonium hydroxid a characteristic red, a reaction that distinguishes them from other common carbohydrates. Wohlk,^^ to whom this test is due, describes the following manipulation: About 0.6 gram of the sample are dissolved in a test tube in 10 c.c. of 10 per cent, ammonium hydroxid and the tube im- mersed in water that has just ceased boiling. This causes the 112 FOOD ANALYSIS ammonium hydroxid to pass off without the liquid reaching the boihng-point or being ejected. In about 20 minutes the color appears. Determination. The preparation of sucrose for use as a standard in polar- imetry and reduction-tests was the subject of formal action at the third session of the International Commission for Uni- jorm Methods oj Sugar Analysis, Paris, July 24, 1900. Purest commercial sugar is selected and dissolved by satu- ration in hot water, and ethyl alcohol added sufficient to pre- cipitate the sugar. The precipitate is whirled in a centrifuge and washed with alcohol. The material obtained is put through the whole process a second time, and the washed material is dried on pure bibulous paper and kept in stoppered glass ves- sels. It still contains moisture, which must be determined and allowed for in making standard solutions. The temperature of the water is not given. Blotting-paper is mentioned in the original test, but filter-paper is better, as commercial blotting-paper is of uncertain composition. For the standardization of solutions for the determination of sucrose and invert-sugar, 2.5 grams of pure sucrose should be dissolved in a mixture of 75 c.c. of water and 5 c.c. of hy- drochloric acid (sp. gr. 1.188 at 15°), hydrolyzed according to the method on page 119, the acid neutralized with sodium carbonate, and the solution diluted to one liter. 2.5 grams of sucrose yield 2.6316 grams of invert-sugar. The number of cubic centimeters of sugar solution used, multiplied by 0.00263, will give the weight of invert-sugar required to reduce com- pletely 10 c.c. of the test solution under the conditions of the experiment. CHEMICAL METHODS. These methods, when applied to the determination of sucrose, must be preceded by hydrolysis, for which see page 119. The following are standard reagents : SUGARS 113 soxhlet's modified copper solution (a. o. a. c). Copper suljate solution. 34.639 grams of pure crystallized copper sulfate are dissolved in sufficient water to make 500 ex. Alkaline tartrate solution. 173 grams of pure potassium sodium tartrate and 50 grams of sodium hydroxid are dis- solved in sufficient water to make 100 c.c. A convenient method is to use 100 c.c. of a solution containing 500 grams of sodium hydroxid in one liter. Potassium acid tartrate, now obtainable of very good quality, may be used instead of potassium sodium tartrate, in which case the proportion required will be 133 grams of potassium acid tartrate and 80 grams of sodium hydroxid made up to 500 c.c. The copper and alkaline tartrate solutions must be kept separate in well-stoppered bottles and mixed only when needed. approximate volumetric method for rapid work. 5 c.c. of each of the solutions are placed in a large test-tube, 10 c.c. of distilled water added, the liquid heated to boiling, and small portions of the solution to be tested gradually added until the copper has been completely precipitated, boiling to complete the reaction after each addition. When the end reaction is nearly reached and the amount of sugar solution can no longer be judged by the color of the solution, a small portion of the liquid is removed by means of a filtering-tube, placed in a porcelain crucible or on a testing plate, acidified with dilute acetie acid, and tested for copper by solution of potassium ferrocyanid. The sugar solution should be of such strength as will require from 15 to 20 c.c. to complete the reduction, and the number of additions of solution should be as few as possible. It is best to verify the first experiment by a second, based on the approximation which the first gives. Boiling for 2 minutes should be required for complete precipi- tation when the full amount of sugar solution has been added in one portion. The factor for calculation varies with the minute details of manipulation; every operator must determine the 114 FOOD ANALYSIS individual factor by using a known amount of the form of sugar that is to be determined and maintaining conditions as uniform as possible. Figure 35 shows filter-tubes suitable for obtaining a small quantity of the liquid. Wiley's tube (A) is a thick- walled glass tube about 4 cm. long on one of which a flange has been made, over which a piece of fine linen is tied. Knorr's tube (B) is much narrower, and has a perforated platinum disk sealed into the lower end. The tube is dipped into water containing suspended asbes- tos, and by aspiration a thin felt is formed over the lower surface of the platinum disk. The tube, thus prepared, is dipped into the boiling copper solution and by aspiration a drop is drawn into the tube. The Wiley filter requires that the liquid be poured from the tube when it is to be tested, but with the Knorr tube the asbestos is wiped off, the liquid expelled through the platinum, and the drop is tested for copper as noted. Another method is to remove a drop of the boiling solution by means of a rod and place it on a piece of pure filter-paper. The pre- cipitate remains in the center of the moistened spot. A drop of potassium ferrocyanid solu- FiG. 35. tion, acidulated with acetic acid, is then placed near it ; as the spot spreads, a brown stain will appear where the liquids meet, if copper still be in solution. soxhlet's exact method. An approximate determination of the reducing sugars in the sample is made by one of the titration methods and a solution is prepared which contains nearly, but not more than, I per cent, of these sugars. 50 c.c. of copper sulfate solution and 50 c.c. of alkaline tartrate solution are mixed, added to a SUGARS II volume of the solution of the sample estimated to be suffi- cient for the complete precipitation of the copper, boiled for 2 minutes, some of the solution filtered rapidly, and the filtrate tested for copper. The process is repeated until two proportions of the solution of the sample are determined which differ by O.I c.c, one giving complete reduction and the other leaving a small amount of copper in solution. The means of these volumes is the amount of solution required for the volume of Fehling solution taken. Under these conditions, which must be rigidly observed, the volume of solution used will contain 0.475 gram of dextrose or 0.494 gram of invert-sugar. As the weight of the sample which is in this amount of solution is known, the percentage of either sugar may be calculated by simple proportion. allihn's method for dextrose. Copper suljate solution. See page 113, Alkaline tartrate solution. 1 73 grams of pure potassium sodium tartrate and 125 grams of potassium hydroxid are dissolved in water and made up to 500 c.c. The substance to be tested is dissolved in water in such proportion that the solution shall not contain more than i per cent, of dextrose. 30 c.c. of each of the reagent solutions and 60 c.c. of water are mixed and heated to boiling, 25 c.c. of the solution to be exam- ined are added, the boiling continued for 2 minutes, and the liquid immediately filtered without dilution, as directed in connection with the reduction or electrolytic methods of determination of copper. The precipitated cuprous oxid is usually converted into free copper and weighed as such. Two methods may be employed for reduction : by hydrogen or by electrolysis. Reduction by Hydrogen. — The cuprous oxid is collected on an F'IG. 36. Il6 FOOD ANALYSIS asbestos filter. This is arranged most conveniently in a special filtering tube, which is shown in figure 36. The wider part is about 8 cm. long and 1.5 cm. in diameter, the narrower portion about 5 cm. long and 0.5 cm. in caliber. A perforated platinum disk is sealed in just above the point of narrowing. The asbes- tos is placed on this disk, washed free from loose fibers, dried well and the tube weighed. The filtering tube is attached to an exhaustion apparatus by passing narrower portion through the cork, and a small funnel is fitted tightly in the top of the tube. The object of this funnel is to prevent the precipitate collect- ing on the upper part of the tube. The lower end of the funnel should project several centimeters below the bottom of the cork through which it passes. The filtering apparatus must be arranged prior to the pre- cipitation, so that the cuprous oxid may be filtered without delay. The precipitate is transferred as rapidly as possible to the filter, well washed with hot water, alcohol, and ether successively, dried, and the cuprous oxid reduced by gentle heating in a current of dry hydrogen. When the reduction is complete, the heat is withdrawn, but the flow of hydrogen is continued until the tube is cold. It is then detached and weighed. The amount of sugar is determined by reference to the table on page 117. Quantities of copper intermediate be- tween those given in the table may be converted into the equiva- lent in sugar by allowing for each o.ooi of copper, 0.0005 of dextrose for figures in the first column, 0.00055 for figures in the second column, and 0.0006 in the third column. Reduction of Copper by Electrolysis. — The filtration is per- formed in a Gooch crucible with an asbestos-felt film and the beaker in which the precipitation was made is well washed with hot water, the washings being passed through the filter, but it is not necessary to transfer all the precipitate. When the asbestos film is completely washed, it is transferred with the adhering oxid to the beaker; any oxid remaining in the (university j SUGARS 117 crucible is washed into the beaker by use of 2 c.c. nitric acid (sp. gr. 1.42), added with a pipet. The crucible is rinsed with a spray of water, the rinsings being collected in the beaker. The liquid is heated until all the copper is in solution, filtered, the filter washed until the filtrate amounts to at least 100 c.c, and electrolyzed. EQUIVALENTS FOR ALLIHN'S METHOD Copper. Dextrose. Copper. Dextrose. Copper. Dextrose. O.OIO 0.0061 0.170 0.0869 0.330 0.1 731 0.020 O.OIIO 0.180 0.0921 0.340 0.1787 0,030 0.0160 0.190 0.0973 0.350 0.1843 0.040 0.0209 0.200 0. 1026 0.360 0.1900 0.050 » 0.0259 0.210 0. 1079 0.370 0.1957 0.060 0.0308 0.220 0.II32 0.380 0.2014 0.070 0.0358 0.230 0.1 185 0.390 0.2071 0.080 0.0408 0.240 0.1239 0.400 0.2129 0.090 0.0459 0.250 0.1292 0.410 0.2187 0. 100 0.0509 0.260 0.1346 0.420 0.2245 0.1 10 0.0560 0.270 0. 1400 j 0.430 0.2304 0.120 0.061 1 0.280 0.1455 j 0.440 0.2363 0.130 0.0662 0.290 0.1510 , 0.450 0.2422 0.140 0.0713 0.300 0.1565 0.460 0.2481 0.150 0.0765 0.310 0.1620 0.463 0.2499 0.160 0.0817 0.320 0.1675 0.465 0.2511 Electrolytic apparatus has been constructed in a great variety of forms. When the operation is carried out frequently, it is best to have an electrolytic table. A platinum basin holding not less than 100 c.c. is used. A cylindrical form with flat bottom is convenient. It should rest on a bright copper plate, which is connected with the negative pole of the electrical supply. The positive pole should be also platinum, either a spiral wire, cylinder, or flat foil. Many operators use a funnel- shaped perforated terminal for the negative pole ; in which case a glass beaker or casserole will be a suitable container, the posi- tive terminal being placed within the negative. Il8 FOOD ANALYSIS Four cells of a gravity battery will suffice for a single de- composition, and will operate two, but more slowly. It is usual to arrange the apparatus so that the operation may be continued during the night. When the electricity is taken from the general supply of the laboratory, it is usually neces- sary to interpose resistance and to have some means of meas- uring the current-flow. This is sometimes done with a gas evolution cell and incandescent lamp, but an ammeter and adjustable rheostat is better. OPTIC METHODS. The general principles of polarimetry have been explained elsewhere. For the decolorization and clarification of solu- tions, the following standard reagents are employed : Lead suhaceiate. Solution of lead acetate is boiled with excess of lead monoxid for 30 minutes, filtered, and brought to a specific gravity of 1.250. Sohd lead subacetate may be. used in preparing the solution. The clarification of sugar solutions may often be more con- veniently effected by the addition of solid lead subacetate, ac- cording to the suggestion of Horne.^^ The weighed material is dissolved in water and made up to 100 c.c. Finely-powdered lead subacetate is added in small quantities, with shaking until the precipitation is complete, allowing each portion to dissolve before adding more. When the last portion has dissolved, the solution is shaken, filtered and the reading taken. No allowance for precipitate is required. Excess of lead may be removed from these solutions by Sawyer's method.^^ A solution of double normal potassium oxalate (184.4 grams in 1000 c.c.) is used. 10 c.c. of this are added to 80 c.c. of the clarified solution, allowed to stand at room temperature for 15 minutes, and filtered. The oxalate is in large excess; this does not interfere with the polarization but renders the precipitate granular and easily filtered. Alumina-cream. A cold saturated solution of alum is SUGARS 119 divided into two unequal portions; a slight excess of ammo- nium hydroxid is added to the larger portion and the remainder is added until a faintly acid reaction is obtained. For sugars and molasses the normal weight for the instrument is weighed out, washed into a 100 c.c. flask, and water added to make about 80 c.c. When the material has dissolved as far as possible, lead subacetate is added until all precipitable matter has separated. (With molasses sufficient acetic acid should be added to convert the lead subacetate into acetate.) The flask is filled to the mark, — using, if necessary, a little ether spray to break bubbles, — filtered with a dry filter, the first 15 c.c. re- jected, and the reading taken on the remainder of the filtrate. If the liquid is very dark, some dry finely-powdered pure bone- black should be used instead of paper and the first 40 c.c. of filtrate rejected. All observations should be made as nearly as possible at the temperature for which the instrument is adjusted. A change of 5° in the interval between filling the flask and making the reading will cause, by change of volume, an error of about 0.1 per cent, in samples containing 90 per cent, of sucrose and an error of about 0.5 per cent, in samples containing 50 per cent, of sucrose. With juices or other dilute materials, weighing may be omitted, and 100 c.c. of the sample measured off, powdered lead subacetate added (page 118), filtered and a reading taken. A. O. A. C. INVERSION METHOD (hYDROLYSIS). A clear solution is made according to one of the methods given above. 50 c.c. of the filtrate are placed in a flask marked at 50 and 55 c.c, 5 c.c. of pure fuming hydrochloric acid added, and the liquids well mixed. The flask is heated in water until the thermometer, with the bulb as near the center of the solution as possible, marks 68°. About 15 minutes should be required for this heating. The flask is then removed, cooled quickly to room temperature, and polarized, noting the temperature. If the sample originally contained invert-sugar, the secoud I20 FOOD ANALYSIS polarization should be made at approximately the same tem- perature as the first. The calculation of the amount of sucrose is made by the following formula : S = - 143 -i 2 a being the first and h the second reading, which are added when of opposite signs and subtracted when of like signs ; that is, the algebraic difference is taken, in either case. With dark-colored materials it will often be* advantageous to add an excess of alumina cream. Alumina cream alone will often suffice for clarification. When lead subacetate is used with liquids containing levu- lose, it is usual to render the filtrate acid in order to break up a compound which the levulose forms with lead, but it is likely that potassium oxalate method (page 1 1 8) would be satis- factory. Hydrochloric acid affects slightly the rotatory power of these solutions. In observations at high temperatures, the expansion of the liquid also introduces an error. These interferences are usually disregarded in food analysis. GERMAN OFFICIAL METHOD. 26.048 grams of the sample are dissolved in a sugar flask and the solution made up to 100 c.c; 50 c.c. of this solution are transferred by means of a pipet to a flask graduated at 50 and 55 c.c, enough lead subacetate solution added for clarification, the volume made up to the 55 c.c. mark, and the liquid thor- oughly shaken and filtered. The filtrate is then polarized, the reading being corrected for the extra 5 c.c. The liquid ad- hering to the pipet is washed into the 100 c.c. flask containing the remaining 50 c.c. (13.024 grams), 5 c.c. of concentrated hydrochloric acid (38 per cent., specific gravity 1.188 at 15°) added, and the flask placed in a water-bath the temperature of SUGARS 121 which is 70°. The contents of the flask should reach a tem- perature of 67^-70° in two or three minutes, when the tem- perature should be maintained within this limit for exactly five minutes, keeping the temperature as nearly 69° as possible. (See international agreement, page 21, as to standard weight of sugar.) SUCROSE Under the term sucrose all forms of table sugar are included. The principal sources are: the sugar-cane, Saccharum offici- narum L. ; beet. Beta vulgaris L. ; sorghum, Sorghum sacchar- atum Persoon; sugar maple, Acer saccharinum L. In the crude state there is a noticeable difference, but so far as is known, the sucrose is identical in all cases. Adulterations are few. The addition of glucose, especially to the lower grades, formerly extensively practised, now rarely occurs. The difference in the grades depends largely upon the extent to which the molasses and mineral matter have been removed. Maple sugar is sold in the crude condition and is often adulterated. The usual examination of commercial sugar is determina- tion of the amount of water, ash, sucrose, and reducing sugar. Water and ash are determined as on pages 27 and 39. In the best grades of sugar these wjll often not amount to more than 0.1 per cent. In the lower grades ash may be 3 per cent., and water between 10 and 15 per cent. The higher proportions of ash are found in beet-sugar. The estimation of sucrose is most conveniently made by the polarimeter. The direct read- ing is usually sufficient, but the result may be checked by hydrolysis, and reading at ordinary temperature and at 86°. The best grades will give a direct reading closely approximat- ing 100 per cent. In some cases the direct reading will slightly exceed icx), due to a small proportion of rafhnose. The lower grades of sugar contain some invert-sugar, and the proportion 122 FOOD ANALYSIS of sucrose may be even below 80 per cent. Maple sugar usually contains about 85 per cent, of sucrose. Coloring-matters. — Granulated and loaf sugars often con- tain ultramarine blue, added to improve color. It may be separated by dissolving a considerable quantity of the sample in water, allowing the coloring-matter to subside, and washing it with water several times by decantation. Ultramarine blue is decomposed by hydrochloric acid, the color discharged, and hydrogen sulfid Uberated. Tin chlorid is sometimes employed in order to give sugar a bright, lasting, yellow tint. The color appears to be the result of action on the sucrose. As a rule, the finished product contains but traces of tin, the greater portion being removed with the molasses. The so-called Demerara sugar is prepared in this way. Demerara sugar is frequently imitated by the addition of artificial coloring, usually to beet-sugar. To sepa- rate such added coloring-matter Cassel recommends the following method : About 100 grams of the sample are shaken with alcohol of 90 per cent. This will often remove the color in a single washing. In some cases it is advisable to use alcohol of 75 or 80 per cent. The solution is filtered from the sugar, evap- orated to dryness, the color again taken up with alcohol, and a skein of silk or wool (preferably slightly mordanted with aluminum acetate) treated with the solution, warmed for some time, and subsequently well washed with water. The skein will be dyed of a more or less yellow color in the presence of artificial dye. A sample containing only such coloring- matter as is natural to sugar, even by repeated washing with alcohol of 90 per cent., does not leave absolutely colorless crystals, and does not give a solution capable of permanently dyeing silk or wool. It is probable that the wool test described on page 64 might be successfully apphed to a solution in water. See also Crampton & Simon's test for caramel ^ page 125. SUGARS 123 The occasional occurrence of artificial sweetening substances (e. g. saccharin, glucin) as substitutes for sugar in confections, fruit juices, jams, and similar articles must not be overlooked. The possibility of commercial glucose and invert-sugar con- taining arsenic and lead derived from the sulfuric acid must also be borne in mind. MOLASSES AND SIRUP Molasses is the uncrystallizable sirup produced in the manufacture of sugar. It properly differs from treacle in that it comes from sugar in the process of making, while treacle is obtained in the process of refining, but the two terms are often employed interchangeably. Treacle, often called refiner's molasses, may contain 35 per cent, or more of sucrose, which is prevented from crystallizing by the associated substances. Ordinary table-molasses is made from cane, sorghum, or maple. Molasses from raw cane-sugar contains considerable invert-sugar, from which beet-root molasses is comparatively free. The latter, however, contains raffinose and a great variety of other bodies; the proportion of salts being some- times 15 per cent. These impurities render it unfit for table use. Beet-sugar partially or wholly refined is free from these ingredients and may be used in the preparation of table sirups. Maple sirup is molasses from the maple. Some so-called maple sirup or "mapleine" is made by addition of extract of hickory-bark to sucrose or glucose sirup. Molasses and maple sirup are often adulterated by the addi- tion of glucose sirup. The product is usually sold as molasses, but is sometimes designated "mixed goods" or "table sirup." Glucose sirup produces a pale liquid, of good body, and many samples consist almost entirely of this material, flavored by the addition of a small proportion of the lowest grades of refuse molasses. The addition of glucose to molasses is readily detected by 124 FOOD ANALYSIS means of the polariscope. The normal or half normal quantity for the instrument is prepared as described on page 119 and the reading taken. A portion of this solution is hydrolyzed, as described on page 118, and two readings taken, one at or near the same temperature as the direct reading, and a second at 86° (see page 17). Pure molasses generally gives on direct reading at a temperature of 20° a deviation corresponding to 40 or 50 on the cane-sugar scale. After hydrolysis, the reading at the same temperature will be — 10 or — 20, and at a tempera- ture of 86° will be zero or near it. Sirups made by the solution of sucrose in water will usually give a rather higher direct reading, but after hydrolysis the results will be the same as with molasses. In the presence of any considerable quantity of glucose the direct reading is nearly always above 60 and may rise to 120 or more. After hydrolysis, the sample remains strongly dextrorotatory even at 86°. For determination of glucose see page 126. Dark molasses is often bleached. Bone-black is sometimes used, but ozone, hydrogen dioxid, sulfurous acid, sulfites, and sulfuric acid have been employed. One method consists in the addition of zinc dust and sodium sulfite, the zinc being subsequently removed by the addition of oxalic acid. The bleached molasses is liable, therefore, to contain either zinc or oxalic acid. As noted above, some samples of sugar are prepared by the use of stannous chlorid; the latter may pass into the mo- lasses in such proportion as to be dangerous. Copper is occa- sionally present, derived from the apparatus of the refinery. For the detection of metallic impurities in molasses, not less than 50 grams should be ashed and examined as described on page 40. The U. S. standard for molasses is not more than 25 per cent, of water nor more than 5 per cent, of ash. Caramel is a dark brown mass, soluble in water and weak SUGARS 125 alcoholic liquids, obtained by heating sucrose to about 200°. It is largely used as a coloring-matter in foods and beverages. It is now occasionally adulterated or imitated by artificial coal-tar colors. The wool test will serve in many cases to detect these. Caramel as a coloring agent is most easily recognized by a method due to Crampton & Simons: The liquid is well shaken with a small quantity of fuller's earth and filtered. Coloring matters from charred or uncharred wood are not removed, but if caramel be present the filtrate will be noticeably paler than the original liquid. See also under "Alcoholic Beverages." GLUCOSE Commercial glucose consists principally of dextrose with considerable maltose and gallisin and some dextrin. In trade the term "glucose" is restricted to the sirup; the solid is called "grape-sugar." Inferior quahties of glucose may con- tain sulfurous or sulfuric acid, calcium sulfate, arsenic, and lead. Glucose is often termed "corn sirup." The following are analyses of commercial glucoses; Nos. i and 2 are by Moritz & Morris, 3 and 4 by Stern. In Stern's analyses some figure has been determined by difference, prob- ably that given as " unfermentable bodies," in which the gallisin and nitrogenous matters are included. Dextrose, Maltose, Dextrin, Gallisin, Nitrogenous matters, . . Unfermentable bodies, Ash, Water, No. I 50.58 No. 2. 47-71 No. 3. 70.0 No. 4. 67.4 14.19 1.76 12.29 2.98 5-1 II.O 15-59 1. 18 15.90 0.81 14.08 4-3 1.44 1-39 0.2 1.6 16.49 20.77 9.9 15-7 101.23 101.85 lOO.O 126 FOOD ANALYSIS Leach ^^ found that the glucose commonly used in adulterating molasses, maple sirup and honey gives a direct reading of 87.5 with a half-normal weight and 200 mm. tube, equivalent to a full reading of 175. He has, therefore, proposed to calculate the glucose on this basis, by the following formula : „ 100 (a — 5) This may be simpUfied to : = 0.561 (a — S) in which G is glucose, S, sucrose and a the polarimetric read- ing before hydrolysis. The amount of sucrose must be calcu- lated by the formula on page 120 from the reading before and after inversion. Some samples used for jellies and jams may show a reading as low as 150. If glucose of this quality is suspected, the constant in the above formula should be 0.666. The method therefore is approximative and suggestive. Freshly made solutions of dextrose show bi-rotation (as described under lactose). This disappears on standing at room temperature for 24 hours. It does not occur with sirups or the glucose used in adulterating sugar- or fruit-products, but must be borne in mind in dealing with solid articles. The examination of glucose samples may be conducted as follows : Arsenic may be detected by Reinsch's test ; lead by the routine method given on page 58. The amount of free acid is deter- mined by titration of a known weight with standard alkali, using phenolphthalein as indicator. Sulfurous acid may be detected by adding some of the samples to dilute hydrochloric acid, with a few fragments of zinc in a test-tube, and covering the mouth of the tube with a piece of filter-paper containing some lead acetate. A spot of lead sulfid indicates reducible sulfur compounds. Calcium sulphate or other mineral matter may be determined by the weight and composition of the ash. SUGARS 127 LACTOSE Commercial lactose is usually obtained from the whey of cows' milk. Inferior qualities contain notable amounts of nitrogenous matter, mineral substances, bacteria, and spores of fungi. Pure lactose is a white crystalline powder, not very soluble in water and feebly sweet. When crystalHzed by evaporation at low temperature, it retains one molecule of water, but this is easily removed. The freshly made solution in water has a dextrorotatory power much greater than normal ; upon standing for 24 hours, or immediately upon boiling, it acquires its normal rotatory action. This phenomenon, known as "birotation," must not be overlooked in examining samples of lactose or concentrated milk-products. Lactose has high reducing power, especially upon alkaHne copper solutions. Under the influence of some common organisms it is rapidly converted into lactic acid; by special methods it may be converted into ethyl alcohol. For qualitative tests for lactose see page in. Quantitative determinations are made either with a polarimeter or an alkaline copper solution. The details of these methods are given in connection with the analyses of milk. The examina- tion of commercial samples should be directed to the deter- mination of the amount of nitrogen, ash, lead, copper, and zinc. The sample should not be acid, nor contain any appre- ciable amount of matter insoluble in water. MAPLE SIRUP AND MAPLE SUGAR These are substantially sucrose with minute amounts of special flavors. Sucrose from other sources is often added- Adulteration with maple sirup glucose is also common. Much attention has been given to the standard composition of pure maple sugar, in order to determine adulteration with sucrose from other sources. Analytic methods. Glucose is detected by examination with 128 FOOD ANALYSIS the polarimeter before and after hydrolysis. Pure maple sugar is inverted, glucose is but slightly affected. The follow- ing results obtained by Ogden illustrate this : Percentage Polarimeter Reading. Sucrose. Direct. After Hydrolysis. Maple sirups free J 53.1 — 22.2 56.0 from glucose : ( 59.6 — 2 1 .9 60.6 Maple sugars: { 84.1 —28.8 85.9 88.0 —28.3 87.6 Maple sirups con- f 80.0 18.9 taining glucose: ( loo.o 45.6 The methyl alcohol method for detecting glucose in honey will be of some value. With pure maple sirup, the precipitate is abundant and flocculent but not adherent to the glass. On standing, crystals of sucrose appear. When considerable glucose is present, a more granular precipitate appears which adheres to the glass. For determination of glucose, see page 126. The water in maple sirup is determined in the usual way, but it will be advantageous to use a dilute liquid and spread it over a large surface. Maple sirup should be diluted with its weight of water, and maple sugar dissolved in twice its weight of water. The drying should be completed in the water-oven. The most important data for judging of the addition of sucrose are the amount and alkalinity of the ash, the amount of lead subacetate precipitate and the malic acid value. Frear,^'' who examined sirup and sugar made under his own observation, suggests a minimum relation of ash to sucrose of i to 160, that is, the ash should not be less than 0.625 ^^ ^^^ total sucrose. The ash must be determined with care, as some of the consti- tuents are volatile. Burning in a muffle at as low a tempera- ture as possible is preferable. The weighing must be done promptly, as the ash is deliquescent. In some cases, the data of alkalinity of the water-soluble portion to phenolphthalein and methyl orange, and the alkalinity of the insoluble portion, will be needed as described on page 39. SUGARS 129 @ A The lead subacetate precipitate is measured by volume after concentration by a centrifuge, according to the method of Hortvet.^^ A special tube and holder, shown in figure 37 on a scale of one-half, is used. Each tube must have a holder. Tubes and holders must be closely balanced in pairs, so that the centrifuge will be evenly loaded. The holder may be made of soft wood. Instrument-makers can, however, make aluminum holders that will be satisfactory. The narrow part of the tube should be graduated in c.c. and fractions. 5 c.c. of sirup or 5 grams of sugar are placed in the tube, 10 c.c. of water added, and the contents well-mixed, sugar being allowed to dissolve completely before the final mixing. 0.5 c.c. alumina cream and 1.5 c.c. of lead subacetate (see page 118) are added, the mixture again shaken and allowed to stand for an hour, the tubes being occasion- ally rotated to facilitate settling. Tubes must, of course, be made up in pairs. They are placed in the centrifuge, run for about 10,000 turns within six minutes, and exam- ined ; any material that may be adhering to the wider portion is loosened with wire at the end, the tubes again rotated for six minutes, and the volume of the precipitate noted, read- ing to o.oi c.c. if possible. Each operator must by trial with samples of definite origin establish standards applicable to the centrifuge used. Using an instrument with a radius of 18.5 cm., Hortvet obtained with pure maple sirups 1.2 to 2.5 c.c. and with pure maple sugars 1.8 to 4.0 c.c. Adulterated articles give much less. Experiments with pure sucrose and precipitants must be made and the volume of precipitate noted as a correction. The so-called ''malic acid value," of use in judging the quality of maple products, is obtained by Hortvet's modifica- tion of the method of Leach & Lythgoe.*^ Fig. 37. 130 FOOD ANALYSIS 6.7 grams of the sample are weighed into a 200 c.c. beaker, water added to make the volume 20 c.c, the solution made slightly alkaline with ammonium hydroxid, i c.c. of a 10 per cent, solution of calcium chlorid and then 60 c.c. of 95 per cent, alcohol added. The beaker is covered and heated for one hour on the water-bath, the heat withdrawn and the liquid allowed to stand overnight. The precipitate is collected by filtering through good filter paper (probably the hardened paper will be satisfactory), washed with hot 75 per cent, alcohol, until all calcium chlorid is removed, dried and ignited. 20 c.c. -^ hydrochloric acid are added, the solution warmed until the lime is dissolved and the excess of acid determined by titration. One-tenth the number of c.c. of acid neutralized is the provi- sional malic acid value. With pure maple products the figure will not be below 0.80. HONEY Honey consists principally of dextrose and levulose with small proportions of mineral and flavoring matters and often formic acid. In some cases small amounts of sucrose and mannitose and a considerable proportion of carbohydrates of the dextrin class are present. Microscopic examination will usually show pollen, portions of insects' wings, and spores of fungi. Crystallized dextrose is occasionally present. The color of honey varies from light amber- yellow to brown- ish-black, according to the source, and time and manner of storage. White clover honey is nearly colorless. Strained honey is that freed from comb by straining. Extracted honey is freed from comb by centrifugation or settling. The proportion of water ranges within the limits of 12 and 22 per cent. The reducing bodies calculated as dextrose usu- ally amount to from 60 to 75 per cent. If sucrose be present in but small amount in the nectar of the flowers, it may be en- HONEY 131 tirely hydrolyzed in the bee or after deposition in the hive, the honey being quite free. Honey contains no true dextrin, but many samples yield, with strong alcohol, precipitates of carbohydrate intermediate between starch and sugar, the proportion being as high as 40 per cent, or more in the case of honey of coniferous origin. Dextrorotatory samples, apparently pure, have been reported. They were probably of coniferous origin. They have been disregarded in the official standard. U. S. Standard. Honey is the nectar and saccharine exudation of plants, gathered, modified and stored in the comb of the honey-bee {Apis mellifica). It is levorotatory. Water should not be over 25.0 Ash should not be over 0.25 Sucrose should not be over 8.0 Adulterations. — Bees are often fed with cane-sugar, which they hydrolyze partially. Ogden gives the following results of polarimetric examination of honey obtained in this way : Direct, i8°.5. Temperature, 25.2°. After hydrolysis, — 9.0. Temperature, 24°. The common adulterants of strained honey are invert-sugar and glucose sirup. It is usually impossible to detect with certainty the addition of invert-sugar. An ash higher than 0.3 per cent., containing a notable quantity of calcium sul- fate, may point to invert-sugar or to glucose sirup. Samples are frequently encountered which give a direct polarimetric reading of — 14 to — 20 on the cane-sugar scale, and, after inversion, slightly higher figures ; these in many cases probably contain added invert-sugar. The direct addition of sucrose to honey is not usual, but has been practised in some cases. Its presence in considerable quantity will be indicated by the high right-handed rotation, 132 FOOD ANALYSIS decidedly reduced on hydrolysis. A sample of so-called "hoar- hound honey" examined in the chemical laboratory of the U. S. Department of Agriculture was found to consist mainly of a solution of sucrose with some alcohol. A common method of adulteration consists in pouring glucose sirup over honeycomb from which the honey has been drained, and allowing the mixture to stand until it has acquired a honey flavor. Such samples give a high positive polarimetric reading, but little affected by hydrolysis. Dextrin is a constant constituent of commercial glucose sirup, and the attempt has been made to detect the latter by the formation of a precipitate when the sample is diluted with alcohol. It has been shown, however, that many samples of honey contain a considerable material precipitable by ethyl alcohol, amounting in some instances to 50 per cent. Accord- ing to Beckmann, better results may be obtained by the use of methyl alcohol. Pure honey, both the ordinary form and the dextrorotatory variety, that might be regarded as adul- terated with glucose, was found to yield, when largely diluted with methyl alcohol, only a slight flocculent precipitate, which did not adhere to the walls of the vessel. Glucose sirup yielded a precipitate of dextrin amounting to about 31 per cent., which produced with a solution of iodin in potassium iodid the red characteristic of erythrodextrin. The reaction is also obtained by direct addition of the iodin solution to honey containing glucose sirup. The quantitative determina- tion is made by diluting 8 grams of the sample with 8 c.c. of water and diluting the mixture to 100 c.c. with methyl alcohol. The precipitate is filtered off, washed with methyl alcohol, dissolved in water, and the solution evaporated on the water- bath with repeated addition of methyl alcohol until quite dry. Adulteration with solid glucose (so-called grape-sugar) cannot be detected by this method, since in the preparation of this the hydrolysis is carried further. Methyl alcohol produces only a slight turbidity. HONEY 133 Beckmann has also proposed the following test for solid glucose and glucose sirup: 5 c.c. of the honey solution (20 grams in 100 c.c. of water) are mixed with 3 c.c. of a 2 per cent, solution of barium hydroxid, 17 c.c. of methyl alcohol added, and the mixture shaken. Pure honey remains clear, but in the presence of dextrin, glucose, or glucose sirup a considerable precipitate is formed. The test was applied quantitatively by increasing the amount taken to 50 grams, the methyl alcohol added rapidly to avoid deposition on the glass, the liquid well shaken once^ the precipitate collected on a tared asbestos filter, washed with methyl alcohol and ether, and dried at 55° to 60°. Excessive shaking was avoided in order to prevent the action of air on precipitate. It was found that the quicker the working, the more accurate the results. In some cases it was found necessary to determine the sulfates and phosphates and to correct the results accordingly. The mean results in test analyses, calculated to i gram of the material taken, were: Dextrin, 0.916 gram; glucose sirup, 0.455 gram; solid glucose, 0.158 gram. Admixture of dex- trorotatory conifer honey to the extent of 90 per cent, was not found to increase the amount of precipitate, but, on the con- trary, to diminish it slightly. The following are results obtained on samples of natural honey rich in dextrinous bodies. Sp. is the specific rotatory power for yellow light : Apple honey, Sp. = — 12.2. Precipitate by ethyl alcohol 23.7 per cent. Barium precipitate 5 c.c. 10 per cent, solution gave 0.0044 gram. " " 5 c.c. 20 per cent. " " 0.0072 " Umbellifer honey, Sp. = — 4.6. Precipitate by ethyl alcohol 29.1 per cent. Barium precipitate 5 c.c. 10 per cent, solution gave 0.0148 gram. " " 5 c.c. 20 per cent. " " 0.023 " Conifer honey, Sp. = 16.9. Precipitate by ethyl alcohol 41 .9 per cent. Barium precipitate 5 c.c. 10 per cent, solution gave 0.0132 gram. " " 5 c.c. 20 per cent. " " 0.0248 " It appears from these data that even under unfavorable cir- 134 FOOD ANALYSIS cumstances it is possible to recognize the addition of from 5 to 10 per cent, of ordinary dextrin, 10 to 20 per cent, of glucose sirup, and 30 to 40 per cent, of solid glucose to conifer contain- ing as much as 40 per cent, of natural dextrinous matter. With ordinary samples, such as the apple honey just noted, adultera- tion would be much more easily detected. For the determination of glucose Leach ^^ recommends hy- drolyzing in the usual manner, taking the reading at 87° (see page 17) and dividing by 175. The quotient is the approximate percentage of glucose. (See page 126.) Konig and Karsch have proposed the following method for detection of glucose : 40 grams of the sample are made up to 40 c.c. with water, well mixed, 20 c.c. placed in a 250 c.c. flask, and absolute alcohol added, by very small portions at a time, with constant shaking, until the flask is filled to the mark. The mixture is allowed to stand for several days with occasional shaking. The solution is again shaken well and quickly filtered. 100 c.c. of the filtrate are evaporated to remove alcohol, but not to dryness, the residue made up to 20 c.c. by addition of lead subacetate and water, the solution filtered and examined in the polarimeter. The precipitate produced by alcohol is washed several times with 90 per cent, alcohol and then dissolved off the filter with water, evaporated on the water-bath, dried in the water-oven, and weighed. The following are some of the results obtained : Percentage of Reduc- POLARIMETRIC READING. ing Carbohydrates Before Treatment After Treatment Precipitated by with Alcohol. with Alcohol. Alcohol. Pure honey, .... -6.4 -8.5 3-2 —12.4 134 1-7 -16.7 17.0 —117 ir.7 3-3 —9.2 13.2 —7.7 9.9 9-7 —9.9 12.5 — —7.5 6.2 34.0 Honey containing 75 per cent, glucose, .... 25.5 2.4 20.6 CANDIES AND CONFECTIONS 1 35 Molasses is said to have been added to honey, but its use is infrequent. The ash of molasses is high and contains considerable chlorids. Beckmann suggests its detection by the production of a precipitate on addition of a solution of lead subacetate in methyl alcohol, the formation of which is at- tributed to the presence of raffinose. 5 grams of the solution are mixed with 22.5 c.c. of methyl alcohol and 5 c.c. of a solu- tion of the honey (which should not contain more than 25 per cent.) are added. If the honey be pure, the solution will re- main clear, but in the presence of molasses a precipitate will be formed. The amount of precipitate varies according to the particular sample of molasses present, but Beckmann claims that it will usually be possible to detect as low as 10 per cent. CANDIES AND CONFECTIONS These terms include many articles, some complex mixtures, the composition of which is secret. The main ingredient is usually sucrose, but invert-sugar, dextrose, starch, mucilaginous substances, gelatin, colors, and flavors are largely employed. Among the objectionable ingredients are paraffin, clay, calcium sulfate, mineral colors, fusel oil, and metal foil. Preservatives are usually unnecessary. The use of mineral colors has declined much of late years, owing to the cheapness and superior brilli- ancy of artificial organic dyes, but some of the chocolate con- fections contain considerable amounts of brown ferric hydroxid. The plain candies, such as rock candy, molasses candy, and candy toys are usually only crystallized or melted sucrose with flavors and colors. Actual experiment by manufacturing confectioners has furnished the following data for proportion of color: One part of auramin will color 30,000 parts of melted sucrose to the deepest yellow required. One part of eosin or fluores- cein will give the average tint to 28,000 parts of "cream goods" (such as used in high-class "mixtures") or 21,000 parts of 136 FOOD ANALYSIS clear and hard candies, or 12,000 to 24,000 parts of some other types. These figures are for "soUd" coloring — that is, the whole mass is dyed; when merely surface-coloring is done, the quantity needed is about i part to 50,000. The ash of candies and confections is generally below one per cent. The flavors are often artificial. A brand called *'Rock and Rye Drops" is often flavored with fusel oil. The colors employed are numerous and constantly chang- ing. At present various eosins {e. g., rhodamin B, rose bengale, erythrosin) are much used for red, fluorescein and auramin for yellow, malachite green and sulfonated allies for green. Cochi- neal and vegetable colors, such as chlorophyl, cudbear and fustic, have come largely into use of late. Bismarck brown is apt to be employed in chocolate colors. Analytic Methods. — The examination of candies will be usually limited to identification of the coloring-matters and detection of starch, clay, calcium sulfate, paraffin, and poison- ous metals. Determinations of sucrose, invert-sugar, dextrose, and gum are difficult and of no practical interest. Glucose may be detected and approximately determined as in honey and maple sugar. A weighed portion of the sample is stirred in cold water until all soluble matter is taken up, the liquid is filtered in a Gooch crucible, the residue washed with cold water, trans- ferred to the crucible, dried at a low heat, weighed, burnt off, and again weighed. The figures for insoluble residue and ash will be obtained. The aqueous solution will usually contain the coloring and some of the flavoring material; the former may often be identified by the tests given on pages 64 to 75. Many flavoring agents may be recognized by odor. If a moderately large sample is dissolved, fractional distillation as described in connection with fruit juices may give information. Starch may be detected by iodin. Any notable amount of gelatin or albumin will be indicated by the Kjeldahl method. Clay, calcium sulfate and iron oxid will be found in the ash. FATS AND OILS 137 FATS AND OILS The methods for determining melting and solidifying points and specific gravity of fats and oils have been fully described in the introductory part. Some comparative data are given in this section, together with methods applied almost exclu- sively to this class of food-products. Specific gravity determined at temperatures other than 15.5° may be reduced to this by a correction of 0.00064 for each degree. This figure is derived from results obtained by Allen. The specific gravity of fats and oils changes by time. The follov^ing table, due to Thomson & Ballentyne, shows this fact ; the figures are for -J^^ : One Month. Three Months. Six Months. 0.9187 0.9208 0.9246 0.9237 0.9261 0.9320 0.9213 0.9233 0.9267 0.9183 0.9188 0.9207 Olive, .0.9168 Cottonseed, 0.9225 Arachis, 0.9209 Rape, 0.9168 Color-tests. — Many color-tests for oils and fats have been proposed. The reactions are in some cases dependent on natural impurities and may fail when the sample has been pro- duced under unusual conditions or subjected to special treat- ment. Thus, cottonseed oil by heating loses susceptibility to several color-tests, while lard derived from animals fed liberally on cottonseed products will give distinctly the cotton- seed oil reactions. Special color-tests applicable to particular oils or fats will be described in connection with these. The following general reactions are much used : Sulfuric Acid Test. — A drop or two of strong sulfuric acid is placed in the center of about 20 drops of oil, allowed to rest a few moments, the color change noted, the mixture stirred, and the effect again noted. The charring action which often obscures the reaction may be avoided by dissolving a 13 138 FOOD ANALYSIS drop of the oil in 20 drops of carbon disulfid and agitating this with the sulfuric acid. Nitric Acid Test. — Bach's method is to agitate 5 c.c. of the sample with 5 c.c. of nitric acid, sp. gr. 1.30. The color reaction is noted, the mixture immersed in boiHng water for 5 minutes, and the condition again noted. The reaction may be violent, and care must be taken to protect persons and apparatus against injury. Massie's method is to agitate 10 grams with 5 c.c. of nitric acid, sp. gr. 1.40, and note the color at the end of one hour. Lewkowitsch states that an acid of specific gravity 1.375 is preferable. In some cases the mixture should stand 24 hours before the final observation is made. Mixtures of strong sulfuric acid and strong nitric acid have been used, but the results are not of material use with food oils. The following data, compiled by Allen, will illustrate the value of these color- tests : Olive, Cotton- seed. Sesame. Arachis. Rape. Sulfuric Acid.— Before stirring, . Yellow- Red-brown. Yellow to Yellow green orange. with or brown. red rings. After stirring, . Brown or Dark red- Green or Brown. green. brown. brown. Nitric Acid.— Bach's test : After agitation Pale- green. Yellow- brown. White. Pale rose. Pale rose. After heating, Orange- Red-brown. Brown- Brown- Orange- yellow. yellow. yellow. yellow. After 1 2 hours' standing, . Solid. Buttery. Liquid. Solid. Solid. Massie's test, . Yellow- green. Orange-red Yellow- orange. Pale red. Orange. Time for solidifica- tion (minutes), . 60 105 105 200 FATS AND OILS 1 39 lodin Number. — This, also called iodin value, is the per- centage of iodin absorbed under specified conditions. Baron Hiibl discovered that a solution of iodin and mercuric chlorid is more uniform in action than iodin alone, and this solution, commonly known as Hiibl's reagent, is usually employed. The following reagents are used in the process : Iodin solution. 25 grams of iodin are dissolved in 500 c.c. of 95 per cent, alcohol. Mercuric chlorid solution. 25 grams of mercuric chlorid solution are dissolved in 500 c.c. of 95 per cent, alcohol and the solution filtered, if necessary. Starch solution. See page 56. • Potassium iodid solution. 15 grams in 100 c.c. of water. Potassium die hr ornate solution. 3.874 grams of pure potas- sium dichromate in 1000 c.c. of water. For use, equal parts of the iodin and mercuric chlorid solu- tions are mixed and allowed to stand at least 12 hours. The strength of the thiosulfate solution is determined as follows: 20 c.c. of potassium dichromate solution, 10 c.c. of potassium iodid solution, and 5 c.c. of strong hydrochloric acid are mixed in a glass-stoppered flask, and the solution of sodium thiosulfate is allowed to flow in from a buret until the yellow color of the mixture has almost disappeared. A few drops of starch solution are then put in and the addition of the thio- sulfate continued until the blue color just appears. The num- ber of cubic centimeters of thiosulfate solution used, multiplied by 5, is equivalent to i gram of iodin. Not more than i gram of fat is weighed in a glass-stoppered flask holding about 300 c.c, and 10 c.c. of chloroform or carbon tetrachlorid are added. After complete solution 30 c.c. of the iodin solution are added and the flask is placed in the dark for three hours, with occasional shaking. 20 c.c. of potassium iodid solution and 100 c.c. of water are added to the contents of the flask. Any iodin which may be noticed upon the stopper of I40 FOOD ANALYSIS the flask should be washed back into the flask with the potas- sium iodid solution. The excess of iodin is now titrated with the sodium thiosulfate solution, which is run in gradually, with constant shaking, until the yellow color of the solution has almost disappeared. A few drops of starch-paste are added, and the titration continued until the blue color has entirely disappeared. Toward the end of the reaction the flask should be closed and violently shaken, so that iodin remaining in the chloroform may be taken up by the potassium iodid solution. A sufficient quantity of sodium thiosulfate solution should be added to prevent a reappearance of any blue color in the flask for five minutes. At the time of adding the iodin solution to the fats,Hwo flasks of the same size as those used for the determination should be employed for conducting the operation without fat. In every other respect the performance of the blank experi- ments should be just as described. These blank experiments must be made each time the iodin solution is used. Iodin monobromid, used as suggested by Hanus,^^ is a satis- factory substitute for Hiibl's solution. It is prepared by dis- solving 13 grams of iodin in a liter of glacial acetic acid and adding 3 c.c. bromin, by which the halogen content is doubled. The acetic acid must be free from substances that reduce a mixture of chromic and sulfuric acids. The iodin mono- bromid keeps for several months and the maximum absorption occurs in 30 minutes, but oils of high iodin number should be given an hour. The solution is used similarly to that of Hiibl, except that an excess of at least 70 per cent, of unabsorbed iodin is necessary, and only 10 c.c. of the potassium iodid solu- tion are added, the solutions being well mixed before the dilut- ing water is added. Especial care is needed in measuring the solution, as the co- efficient of expansion of acetic acid is high and slight changes in temperature will cause appreciable errors. FATS AND OILS 141 loDiN Number of Liquid Acids. — This determination is sometimes of value for detection of admixture of vegetable oils vv^ith animal oils. The separation of the oleic and other liquid fatty acids is best made by the method of Muter & De Koningh, as follows : 3 grams of the fat are mixed with 50 c.c. of alcohol and a fragment of potassium hydroxid in a flask furnished with a long tube. The mixture is boiled until saponi- fication is complete, when a drop of phenol- phthalein solution is added and acetic acid until the solution is slightly acid. AlcohoHc solution of potassium hydroxid is added drop by drop until *a faint permanent pink tint is obtained, when the liquid is poured slowly, with constant stirring, into a beaker containing a boiling solu- tion of 3 grams of neutral lead acetate in 200 c.c. of water. The solution is rapidly cooled and stirred at the same time, and, when cold, the clear liquid is poured off. The precipitate is well washed with boiling water by decantation, transferred to a stoppered bottle, mixed with 120 c.c. of ether, and allowed to remain 12 hours. Wallenstein & Finck use a Drechsel gas- washing flask having the tube shortened about two-thirds, to contain the ethereal solution, and pass a current of hydrogen through it for about a minute. In the case of white fats the liquid is said to remain colorless at the end of 12 hours, but if free access of air is permitted, a dark-yellow solution is produced by oxida- tion. Lead oleate, hypogeate, linolate, or ricinolate will be dis- solved by the ether, leaving lead laurate, myristate, palmitate, stearate, and arachidate undissolved. Lead erucate is sparingly soluble in cold ether, but readily in hot. The contents of the bottle are filtered through a covered filter into a Muter separating- FlG. 38. 142 FOOD ANALYSIS tube (Fig. 38), 40 c.c. of dilute hydrochloric acid (1:4) added, and the tube shaken until the clearing of the ethereal solution shows that the decomposition of the lead soaps is complete. The aqueous Hquid, containing lead chlorid and excess of hydro- chloric acid, is run off through the bottom tap, water added, and agitated with the ether and the process of washing by agitation repeated until the removal of the acid is complete. Water is then added to the zero mark and sufficient ether to bring the ether to a definite volume (e. g., 200 c.c). An aliquot portion of this (e. g., 50 c.c.) is then removed through the side tap and the residue weighed after evaporation of the ether in a current of carbon dioxid. Another aliquot portion of the ethereal solution should be distilled to a small bulk (avoiding complete evaporation), alcohol added, and the solution titrated with decinormal sodium hydroxid and phenolphthalein or methyl-orange, from which the fatty acids may be calculated from the result, or their mean combining weight deduced there- from. A third aliquot part of the ethereal solution should be evaporated at about 60° in a flask traversed by a rapid stream of dry carbon dioxid. When eyery trace of ether is removed, 50 c.c. of the iodin- mercuric chlorid solution (p. 139) should be added, the stopper inserted, and the liquid kept in absolute darkness for 12 hours, after which an excess of potassium iodid solution is added and 250 c.c. of water, and the excess of iodin ascertained with thiosulfate solution in the usual way. From the result the iodin number is calculated. The Hanus method may be used instead of the Hiibl method. Volatile Acids. — This method was first suggested by Hehner & AngelV^but was systematized by Reichert,^^and hence is gen- erally called the Reichert process. In this form it is carried out by saponifying 2.5 grams of the fat, adding excess of sulfuric acid, distilling a definite portion of the liquid, and titrating the distillate with ^ alkali. The number of cubic centimeters of this solution required to overcome the acidity of the distillate FATS AND OILS 1 43 is called the Reichert number. E. MeissP^ suggested the use of 5 grams, and the number so obtained is called the Reichert- Meissl number. Alcoholic solution of potassium hydroxid was originally used for saponification, but the solution devised by Leffmann & Beam," namely, sodium hydroxid in glycerol, is more satisfactory. The reagents and operation are as follows : Glycerol-soda. — 100 grams of pure sodium hydroxid are dissolved in 100 c.c. of distilled water and allowed to stand until clear. 20 c.c. of this solution are mixed with 180 c.c. of pure concentrated glycerol. The mixture can be conveni- ently kept in a capped bottle holding a 10 c.c. pipet, with a wide outlet. Suljuric Acid. — 20 c.c. of pure concentrated sulfuric acid, made up with distilled water to 100 c.c. Sodium Hydroxid. — An approximately decinormal, accu- rately standardized, solution of sodium hydroxid. Indicator. — Solution of phenolphthalein or methyl-orange. A 300 c.c. flask is washed thoroughly, rinsed with alcohol and then with ether, and thoroughly dried by heating in the water- oven. After cooling, it is allowed to stand for about 15 minutes and weighed. (In ordinary operation this preparation of the flask may be omitted.) A pipet, graduated to 5.75 c.c, is heated to about 60° and filled to the mark with the well- mixed fat, which is then run into the flask. After standing for about 15 minutes the flask and contents are weighed. 20 c.c. of the glycerol-soda are added and the flask heated over the Bunsen burner. The mixture may foam somewhat; this may be con- trolled, and the operation hastened by shaking the flask. When all the water has been driven off, the liquid will cease to boil, and if the heat and agitation be continued for a few moments, complete saponification will be effected, the mass becoming clear. The whole operation, exclusive of weighing the fat, requires about five minutes. The flask is withdrawn from the heat and the soap dissolved in 135 c.c. of water. The first 144 FOOD ANALYSIS portions of water should be added drop by drop, and the flask shaken between each addition in order to avoid foaming. When the soap is dissolved, 5 c.c. of the dilute sulfuric acid are added, a piece of pumice dropped in, and the liquid distilled until no c.c. have been collected. The condensing tube should be of glass, and the distillation conducted at such a rate that the above amount of distillate is collected in 30 minutes. The distillate is usually clear; if not, it should be thor- oughly mixed, filtered through a dry filter, and 100 c.c. of .^^ Fig. 39. the filtrate taken. A little of the indicator is added to the distillate, and the standard alkali run in from a buret until neutralization is attained. If only 100 c.c. of the distillate have been used for the titration, the number of cubic centi- meters of alkali should be increased by one-tenth. The distilling apparatus shown in figure 39 is that recom- mended by the A. O. A. C. (and since adopted in Great Britain), and the directions for preparing the flask are also from the same source, but when it is intended merely to distinguish FATS AND OILS I45 butter from oleomargarin, it will be sufficient to measure into a flask 3 or 6 c.c. of the clear fat, and operate upon this directly in an ordinary distilling apparatus. A blank experiment should be made to determine the amount of standard alkali required by the materials employed. With a good quality of glycerol, this will not exceed 0.5 c.c. Most fats give distillates containing but little acid. Saponification Value. — Koettstorfer Number. — This is the number of milligrams of potassium hydroxid required for the saponification of i gram of fat. Its use was suggested by Berthelot, and it was applied to the examination of butter by Koettstorfer.^^ If the saponification value be divided by lo, the result will be the percentage of alkali required for saponi- fication. The reagents and process are as follows: Alcoholic potassium hydroxid. 40 grams of good potassium hydroxid are dissolved in sufficient alcohol to make 1000 c.c. The solution should be clear and light yellow. Alcohol that becomes brown is unfit for use. Purified methyl alcohol and sodium hydroxid may be sub- stituted. The saponification value of sodium hydroxid may be converted into the standard number by multiplying by 1.4. Halj-normal hydrochloric acid accurately standardized. Phenol phthalein solution. The process is as follows: About 1.5 grams of the sample are accurately weighed into a small flask, 25 c.c. of the alcoholic alkali added, and the mass saponified. The same amount of the alkaline solution must be used in all comparative experi- ments, and it must be accurately measured. The flask is provided with an inverted condenser or, more simply, with a tube about 50 cm. long and 0.5 cm. caliber passing through the cork. It is heated on the water-bath for 30 minutes, being occasionally given a rotatory motion. The alcohol should not boil actively. A drop of the indicator solution is added, the liquid allowed to cool somewhat, the flask being 146 FOOD ANALYSIS closed, and then titrated with the standard acid. A blank test should be made, which will eliminate some of the errors of experiment. The number of cubic centimeters used for titra- tion of the saponified mass, subtracted from the number used in the blank experiment, will give the acid corresponding to the alkali which has been neutralized by the fat. From this, the amount of alkali can be determined and calculated by simple proportion to i gram of fat. Flasks of the same kind of glass should be used in com- parative experiments, as some of the cheaper forms of glass are notably affected by alkali. A special form of saponification flask and method of heating used by the A. O. A. C. are shown in figure 40. The flask is arranged so that the cork can be tied down. Allen suggested the use of the figure representing the grams of fat saponified by 1000 c.c. of normal alkali. This would render the method independent of the alkali employed, but the suggestion has not been generally followed. The datum was called by Allen saponi- fication equivalent. It may be ob- tained in any case by dividing 56100 by the saponification number. Similarly, the saponification number may be obtained by dividing 56100 by the saponification equivalent. Acid Value. — This is the amount of free fatty acid. The reagents required are ^ sodium hydroxid and neutral alcohol. The latter is prepared by adding to a good quality of alcohol a drop or two of phenolphthalein solution and sodium hydroxid drop by drop with stirring until the color change occurs. 10 grams of the sample are placed in a bottle provided with a glass Fig. 40. FATS AND OILS 1 47 stopper, about 50 c.c. of the neutral alcohol and i c.c. of phenol- phthalein solution added, and the mass heated to boiling by immersing the bottle in hot water. The bottle is then stoppered and well agitated and the liquid titrated with standard alkali, the bottle being vigorously shaken after each addition until a faint pink coloration persists for a minute or two. On long standing the alkali acts upon the fat itself, i c.c. of -^ alkali is equivalent to 0.0256 gram of palmitic acid, 0.0284 gram of stearic acid, or 0.0282 gram of oleic acid. As the acid present may not be known, it is usual to express the result as the milligrams of potassium hydroxid required to neutralize i gram of fat. This is called the acid number. When sodium hydroxid is used for -titration, the acid number may be calculated by multiplying the quantity of sodium hydroxid required for i gram of sample by 1.4. Solubility in Acetic Acid. — Valenta's Test. — Fats and oils are arranged by Valenta into three classes, according to their solubility in acetic acid. Equal volumes of the oil and acid are placed in a test-tube, thoroughly mixed, and, if no solution takes place, warmed. Class I. — Completely soluble at ordinary temperature: Olive kernel oil ; castor oil. Class 2. — Completely soluble or nearly so at temperatures ranging from 23° up to the boiling-point of glacial acetic acid: Palm oil; coconut oil; olive oil; cacao-butter; sesame oil; cottonseed oil; arachis oil; beef tallow; butter, etc. Class 3. — Not completely dissolved even at the boiling- point of glacial acetic acid: Oils obtained from the seeds of the Crucijerce; rape-seed oil; mustard-seed oil; hedge-mustard oil. For the practical appHcation of the test the method of Chatta- way, Pearmain, & Moor is satisfactory: 2.75 grams of the sample are weighed in a short, rather thick tube with a well-fitting stopper, 3 c.c. of acetic acid 148 FOOD ANALYSIS (99.5 per cent.) are added, the tube closed, placed in a beaker of warm water, and the heat increased until, after well shaking the tube, the contents become quite clear. The source of heat is then removed, and the test-tube so placed that it is in the center of the beaker of heated water, and, by means of a ther- mometer attached to the tube by a rubber band, the whole is allowed to rest until the change from brilliancy to turbidity takes place. The change is very definite, and can be repeated as often as is wished, with a maximum error of about 0.25°. Thermal Reaction with Sulfuric Acid.— Maumene's Test. — Maumene^* found that on mixing sulfuric acid with drying oils a higher temperature is produced than with non- drying oils. With the same sample the temperature will depend upon the acid. The strength of acid employed should be determined by titration, since the specific gravity of the acid of 96 per cent, and of 99 per cent, is practically identical. L. Archbutt recommends the following method of operating: 50 grams of the sample, weighed closely, are placed in a beaker of 200 c.c. capacity, and, together with the bottle of acid, placed in water until both have acquired its temperature, the thermom- eter having been placed in the oil. The beaker is removed, wiped, and placed in a nest of cardboard having hollow sides stuffed with cotton. (A beaker, lined with cotton, or, better, a vacuum jacketed test-tube, may also be used.) The temperature having been noted, 10 c.c. of acid are rapidly with- drawn from the bottle, which is immediately closed, the acid is allowed to flow into the oil while it is being stirred with the thermometer, and the stirring is continued until no further rise of temperature is observed. The stirring must be so managed as to effect as perfect admixture of the oil and acid as possible, thereby insuring an even development of heat throiighout the mixture. The best results are obtained with an acid about 97 per cent. It is desirable to keep on hand a ^stock of oil of known FATS AND OILS - 1 49 purity, and to test some of this with each set of samples examined. Specific Temperature Reaction. — The discrepancies ob- served in Maumene's method may be largely eliminated by that devised by Thomson & Ballentyne," which is to compare the rise of temperature with oil and with an equal volume of water under similar conditions. The number obtained by dividing the oil figure by the water figure is multiplied by loo to eliminate decimals, and the datum so obtained is called the specific temperature reaction. Bromin Thermal Value. — Hehner & Mitchell^® ascertained that the heat evolved in the reaction of bromin with unsaturated fatty bodies furnishes more definite data than does sulfuric acid. As the action of bromin may be violent, it is moderated by a diluent such as chloroform, carbon tetrachlorid, or glacial acetic acid. The latter has the advantage, owing to its high boiling-point, of allowing a wider range of temperature. The procedure is as follows: The bromin, oil, and diluent are all brought to the same temperature, i gram of the oil is dissolved in 10 c.c. of chloroform in a vacuum- jacketed, test-tube. Ex- actly I c.c. of bromin (measured by means of a pipet, connected at the upper end with a narrow tube filled with caustic lime, and having an asbestos plug at each end) is added and the rise of temperature determined by a thermometer graduated into fifths. Acids are dissolved in glacial acetic acid instead of chloroform. A definite relation exists between the iodin number and the heat produced by bromin. In Hehner & Mitchell's experi- ments it was found that if the rise of temperature in degrees was multiplied by 5.5, a close approximation to the iodin number was always obtained, except with rape and linseed oils, but each observer must ascertain the factor applying to particular cases. Wiley" has made this method more accurate and more easy 150 FOOD ANALYSIS of application. A solution of bromin in four parts by volume of chloroform or carbon tetrachlorid is employed. This is to be made up in quantity sufficient for one day's use, and kept in the dark. Dissolving the sample in similar solvents is an additional convenience. 10 grams of the sample, in sufficient chloroform or carbon tetrachlorid to make 50 c.c. of solution, will suffice for nine determinations. At least four determina- tions should be made. The apparatus is shown in figure 41. The tube for holding the reagent and thermometer is about 40 cm. in length, and 1.5 cm. internal diameter. It is conveni- ently held in a drying jar, being fitted air-tight by a rubber stopper. Air is withdrawn from the jacketing jar through the side tubulure. The bromin solution is contained in a stout- walled conical flask with a side tubulure provided with a rubber bulb. Through the stopper passes a pipet, and the flask may be rendered air-tight by gentle pressure on the stopper. The thermometer should be graduated to 0.2° and be read to a tenth by a lens. The operation should be conducted in a room at uniform temperature. The solutions and apparatus are allowed to stand until all reach a uniform temperature. 5 c.c. of the solution of the sample are placed in the inner tube by means of the pipet, without allowing any of the solution to run down the walls of the tube, the thermometer is inserted, and the bromin so- lution is forced up into the pipet by compressing the rubber bulb until the Hquid has passed the mark on the stem. The top of the pipet is closed by the finger, the stopper of the flask loosened, and the liquid allowed to run out until it reaches the mark, when it is transferred to the mixing tube and allowed to flow directly into the solution of fat, but it is now not neces- sary to prevent the liquid running down the side of "the tube. The empty pipet is returned to the flask and the thermometer is observed at once by means of a lens, since the bromination is practically instantaneous, the mercury reaching its maximum FATS AND OILS 151 height in about a minute after the pipet is withdrawn. When the mercury begins to fall, air is admitted to the jacketing space, the mixing tube is withdrawn, its contents emptied, and Fig. 41. the tube held inverted until the residual bromin vapor escapes. The tube may be cleaned by wiping it with a long test-tube cleaner or may be used again without cleaning, after standing 152 FOOD ANALYSIS inverted for half an hour. Traces of brominated oil which may remain upon the side of the tube do not interfere unless they obscure the thermometer. By the above manipulation the thermometer soon returns to the room temperature, and a second determination may be made in half an hour. As noted by Hehner & Mitchell, each analytic system must be separately standardized and the factor for calculating the iodin absorption determined. It is important not to stir or churn the mixture of oil and bromin further than is produced by the running in of the solution itself. Carbon tetrachlorid is the preferable solvent, but the rise of temperature is slightly higher with chloroform. Gill & Hatch^^ have proposed to facilitate the comparison of tests made with different apparatus by employing a standard- izing material, and recommend sublimed camphor for this purpose. 7.5 grams of the camphor are dissolved in carbon tetrachlorid, the solution made up to 25 c.c, and portions of 5 c.c. each brominated. The temperature increase obtained with various oils is divided by the rise observed with camphor, giving a specific temperature increase, analogous to that sug- gested by Thomson & Ballantyne (see p. 149). By dividing the iodin value of an oil by the specific temperature increase, a figure will be obtained by which the iodin value may be ap- proximately calculated. Elaidin Test. — i c.c. of mercury is dissolved in 12 c.c. of cold nitric acid of 1.42 specific gravity. 2 c.c. of the freshly- made deep green solution are shaken in a wide-mouthed stop- pered bottle with 50 c.c. of the sample to be tested and the agitation repeated every ten minutes during two hours. When treated in this manner, oils consisting of nearly pure olein or of mixtures of olein with solid esters, such as palmitin and stearin, give more or less solid product. OKve oil is remarkable for the firmness of the canary or lemon-yellow mass formed. After 24 hours the product is impervious to a glass rod, and some- FATS AND OILS 153 times rings when struck; but this character is also possessed by the elaidins yielded by thfe arachis and lard oils. In making the test, it is important to note the time required to obtain a "solid" product, which will not move on shaking the bottle, as well as the final consistence. The temperature should be kept nearly constant, or erratic effects will occur. The behavior of the more important oils, when tested in the foregoing manner, is described by Allen as follows : A hard mass is yielded, among others, by olive, almond, lard, and sometimes arachis oils. A product oj the consistency oj butter is given by mustard, and sometimes by arachis and rape oils. A pasty or buttery mass which separates jrom a fluid portion is yielded by rape, sesame, cottonseed, sunflower, and some- times mustard oils. Liquid products are yielded by linseed, hempseed, walnut and other drying oils. The results of the elaiden test must be accepted with caution, since it is affected by many conditions, such as temperature, shape of the containing vessel, and the mode of preparation of the acid liquid. The extent to which the sample has been exposed to light and air is a still more important factor; it has been shown that olive oil after exposure to sunlight for two weeks may fail to respond to the test. Index of Refraction. — This datum differs notably in dif- ferent oils, but it is not of much value in detecting adulteration unless considerable of the adulterant be present. Several in- struments have been devised for making refraction determina- tion; the familiar ones are the refractometer of Abbe (figure 42) and the butyro-rcfractometcr of Zeiss (figure 43). The butyro-refractometer has been strongly recommended for the examination of butter. It is equally adapted for the general examination of fats and oils, and may be used for the determination of the index of refraction as well. As these instruments are made by only one firm and are furnished with directions for use, further description will not be required. 154 FOOD ANALYSIS Drying Property. — Livache's Test.^® — The so-called drying of oils (a process of oxidation) is hastened by admixture with finely divided lead. This is prepared by precipitating lead acetate by zinc, washing the precipitate rapidly with water, alcohol, and ether in succession, and drying at very low pres- sure. (Probably drying in nitrogen gas would be preferable.) I gram of the dried lead is mixed on a watch-glass with not Fig. 42. Liu. 43. more than 0.7 gram of the sample by dropping the latter so that it is distributed over the mass, of the lead. The glass is allowed to stand at room temperature exposed to light, but reasonably protected from dust. Drying oils absorb the maximum quantity of oxygen after from 18 hours to 3 days, but non-drying oils do not begin to gain weight until after 4 or 5 days. Fat-acids, except those FATS AND OILS 155 from cottonseed oil, behave the same as the fats. Livache's results are given in the following table. The figures show the percentage of increase in weight after the time specified. A drying oil (linseed) is added for comparison with the food oils. The figure for maize oil is given by Vult6 & Gibson. Oil. 2 Days. 7 Days. 10 Days. Olive, o 1.7 Cottonseed, 5.9 Maize, 5.0 Arachis, o 1.8 Sesame, o 2.4 Rape, o 2.9 Linseed, 14.3 Soluble and Insoluble Acids. — This method, due to Hehner & Angell,^" has been much modified by other investigators. The proportion of acids insoluble in water is often called the Hehner value. The following method, described by Allen, is some- what different from that recommended by the A. O. A. C, but will serve for practical purposes, it being understood that blank tests and tests with standard oils should be made for comparison : About 5 grams of the sample, accurately weighed, are placed in a saponification flask, 50 c.c. of a solution of 40 grams of sodium hydroxid to looo c.c. of alcohol added, the flask closed, and the mixture heated in a steam-bath until complete saponi- fication has occurred. The flask is cooled, the soap solution acidulated with sulfuric acid, the aqueous Hquid separated from the layer of fatty acids, and the latter several times boiled with a considerable quantity of water in a flask furnished with a reflux condenser. The liquids resulting from these operations are separated from the insoluble fatty acids, which it is desirable to boil again with a moderate quantity of water, while driving a current of steam through the flask in which they are con- tained, collecting the distillate, and treating it like the wash- ings. The acidulated aqueous liquid first separated from the 156 FOOD ANALYSIS layer of fatty acids is then distilled to a small bulk, and the dis- tillate exactly neutralized with standard sodium hydroxid, using phenolphthalein as an indicator. The first washings from the insoluble fatty acids are then added to the contents of the dis- tilling flask, and the liquid again distilled to a small bulk, the process being repeated with the succeeding washings. The different distillates should be titrated separately with decinor- mal alkali and phenolphthalein, so that the progress and com- pletion of the washing may be followed, and some information obtained as to the nature and relative proportions of the lower jatty acids present. The neutralized distillates should be united and evaporated gently to dryness, and the residue dried at 100° until the weight is constant. It consists of the sodium salts of the acids that passed over in the distillation. If the number of cubic cen- timeters of -^ sodium hydroxid employed for neutralization be multiplied by 0.22, and the product be subtracted from the weight of the dry residue, the difference will be weight of the volatile acids. When coconut oil and palmnut oil are treated in this man- ner, the distillate will be found to contain lauric acid, which, though almost insoluble in water, is volatile in a current of steam. It may be separated from the more soluble volatile fatty acids by filtering the distillate. Acetyl Value. — This determination, originally suggested by Benedikt, is most conveniently carried out by the method of Lewkowitsch^^: 10 grams of the sample are boiled for two hours with an equal volume of acetic anhydrid in a flask pro- vided with an inverted condenser; the mass is then transferred to a larger beaker, diluted with several hundred cubic centi- meters of water, and boiled for 30 minutes, with a slow current of carbon dioxid passed through by means of a tube drawn out to a fine opening at the lower end. This prevents bumping. On cooling, two layers are formed. The water-layer is drawn FATS AND OILS 1 57 off by a siphon and the other portion washed three times by boiling with convenient measures of water. Prolonged wash- ing should be avoided. The acetylated product is freed from water by filtration through a dry filter in a water-oven at ioo°. 5 grams of the substance are saponified as noted on page 145, the alcohol is evaporated, and the soap dissolved in water. The subsequent operations may now be completed by two methods, "distillation" or "fikration." The latter is the shorter and more convenient. Distillation Method. — The liquid is made up to a volume of several hundred cubic centimeters in a flask fitted with an arrangement for passing in steam or for adding water from time to time. Sufficient dilute sulfuric acid (i part of acid to 10 of water) is added to make the liquid slightly acid, and dis- tillation is carried on until about 700 c.c. are collected. The distillate is filtered and titrated with decinormal alkali. Phe- nolphthalein is recommended as an indicator, but probably methyl- orange will serve as well. The number of cubic cen- timeters of solution required to neutralize the distillate, mul- tiplied by 5.61 and the product divided by the weight of the acetylated material, gives the acetyl number. Filtration Method. — The solution of the saponified acetyl- ated substance is mixed with sufficient standard sulfuric acid to be equivalent to the alkali added for saponification, and the mixture warmed gently. The acids will separate as an oily layer. The layer is removed, washed with boiling water until the washings are not acid, titrated with decinormal alkali, and the acetyl number calculated as-above. The acetyl number is the number of milligrams of potas- sium hydroxid required for neutrahzin^ the acetic acid ob- tained from I gram of the acetylated substance. In this process cholesterol and phytosterol are included in the acetyhzation. Substances yielding volatile acids give an acetyl number 158 FOOD ANALYSIS too high ; this condition will affect the distillation method more than the filtration method. To ehminate most of this error, the percentage of volatile acid should be determined and the figures obtained deducted from the acetyl number. The water used in both methods should be freed as far as possible from carbon dioxid. Even the water used in pro- ducing the open steam should be brought to active boihng be- fore the steam is let into the flask. Waters rich in carbonate are especially objectionable. A slight excess of sulfuric acid causes the insoluble acids to separate better, but this must, of course, be known accurately and allowance made for it. It is possible that the data elucidated by Richmond with regard to the rate of distillation of acids of the acetic series could be applied to the distilla- tion method with advantage, but a special investigation will be needed to determine the point. Viscosity . — Practical determina- tions of viscosity are comparative only and are of little value unless uniform methods are employed. Many forms of viscosimeter have been devised. They are of two types, resistance and flow instruments. In the former, the viscosity is measured by the resistance to the movement of an immersed solid; in the latter, the time required for the flow of a given volume of liquid is measured. Doolittle's torsion viscosimeter is the best of its class; Reilly's (figure 44) is the best of the second class. Descriptions of these instruments and of methods of operation are unnecessary, as they are made according to standard patterns and full working directions are furnished with them. Fig. 44. FATS AND OILS 1 59 Blasdale^^ investigated the relative viscosities of solutions of soap from different grades of olive oils and found the figures of much value. He used the torsion viscosimeter. The prepa- ration of the solution is as follows: 15 grams of the sample are saponified with a mixture of 10 c.c. of alcohol and 30 c.c. of water containing 7.5 grams of potassium hydroxid. The mass is washed into a large dish, heated until the alcohol is re- moved, diluted to 500 c.c. at 20°, and the viscosity determined. The result is expressed by Blasdale in the number of grams of sugar that it would be necessary to add to a liter of water to get the same readings. With some oils it would be necessary to dilute the solution to 1000 c.c. Blasdale's results were as follows: Oils. Viscosity. Olive (California), 573-655 Cottonseed, 280 Arachis, 220 Sesame, 415 Rape, 670 Sweet almond, 645 Mustard-seed oils give high viscosity figures, and a mixture of these with cottonseed oil in some proportions would escape detection by this test. Unsaponifiable Matter. — Most fats and oils contain un- saponifiable matters, the extraction and examination of which are useful data. The operation is most conveniently performed by saponifying with a solution of sodium (or potassium) hy- droxid in alcohol, evaporating the alcohol, dissolving in water, and extracting this solution with ether. The extraction of the dry soap with ether is not so satisfactory. The use of the watery solution is due to Allen. The operation is most con- veniently carried out in a stoppered separator. Separation does not always occur readily, but may often be induced by cooling the contents by adding a little sodium hy- l6o FOOD ANALYSIS droxid solution, more ether, or a few cubic centimeters of alcohol and rotating the mass gently. The aqueous liquid is run out, a few drops of sodium hydroxid solution and lo ex. of water are added, gently agitated, and run off. This treatment is repeated, after which the ether is run off in a tared flask, the aqueous liquid is agitated with a fresh portion of ether, which is washed and poured into the tared vessel as before. This process is again performed, when it will be complete. The ethereal solu- tion may be fluorescent if petroleum products are present. The greater portion of the ether should be distilled off in a recovering apparatus and the rest evaporated in the water bath. If the mass retains globules of water, the flask should be held horizontally and rotated rapidly so as to spread the residue in a thin layer. When no more water is visible and the odor of ether is very slight, the flask is placed on its side in the water- oven for 15 minutes, cooled, and weighed. Long heating should be avoided, as some hydrocarbons are sensibly volatile at 100°. Spermaceti and waxes yield in this process a large percentage of unsaponifiable matter, hence it is not available for the detection of paraffin in such substances. In ordinary cases the distribution of the bodies will be as fol- lows, but some resins will pass into the water in the form of sodium salts : In the Ether: In the Water: Hydrocarbons. Sodium salts. Mineral oils. Glycerol. ParaflSn. Sodium hydroxid. Neutral resins. Coloring-matters from palm oil. Cholesterol and analogs. Cholesterol and Analogs. — In the examination of commer- cial edible oils, the cholesterols are the most important of the above ingredients. Cholesterol is a member of a series of alco- hols, having physical characters somewhat like those of fats. There are a number of homologs, but the individual members FATS AND OILS l6l of the group with a few exceptions have been but little studied. Cholesterol occurs abundantly in some animal fats, such as wool-grease, and has been supposed to be present in olive oil as an exception among vegetable oils, but the investigations of Gill & Tufts ^^ have made this doubtful. Vegetable oils contain anal- ogous bodies. Among the most common of these is phytosterol. Some cereals contain a homolog termed sitosterol^ and oils from these seeds will be liable to contain it. A general method for the extraction of these substances is that of Foster & Riechelmann: 50 grams of the fat are twice boiled, for about 30 minutes at a time, with 75 c.c. of alcohol in a flask fitted with an inverted condenser, the flask being mean- while well shaken. The alcoholic solution is mixed with 15 c.c. of 30 per cent, sodium hydroxid solution, and boiled on the water-bath in a flask fitted with a condensation tube until about one-fourth of the alcohol is evaporated. The fluid is then evaporated nearly to dryness in a porcelain basin and the residue shaken with ether. The ethereal solution is evaporated to dryness, the residue dissolved in about 40 c.c. of water, shaken out with a mixture of 75 c.c. of ether and 3 c.c. of alcohol, the solvent removed, washed three times with water, evaporated, and the residue crystallized from alcohol. Von Raumer determines the amount of crude cholesterol as follows: 50 grams of the oil are saponified with alcoholic po- tassium hydroxid. The resulting soap is evaporated to dryness, reduced to powder, and extracted with 50 to 75 c.c. of ether in a Soxhlet apparatus, plugs of fat-free cotton being placed above and below the layer of soap. The residue is saponified again with 10 c.c. of half normal alkali evaporated to dryness with sand, and re-extracted as before during two hours. When the work is carefully 5one, the second Saponification and ex- traction is unnecessary. The following amounts of residue calculated to 100 grams of sample were obtained by this method: Cottonseed oil, 0.719 15 l62 FOOD ANALYSIS gram; sesame oil, 1.3 14 grams to 1.325 grams; lard, 0.217 gram. These substances are insoluble in water, sparingly soluble in cold alcohol, freely in boiling alcohol, and in the other com- mon solvents immiscible with water such as ether, chloroform, petroleum spirit. They are distinguished from each other by melting-point, crystalline form and some color reactions as fol- lows : Cholesterol. Phytosterol. Sitosterol. Melting-point. 145° 132-4° 137-8° Crystals from hot Rhombic plates, Needles, Narrow plates. alcoholic solu- often with re- grouped i n with pointed tion. entering angles. tufts. terminals. Solution- in Bluish green Clear green dilute acetic becoming changing to anhydrid with reddish yellow. pure yellow. sulfuric acid. Solution in Blood red. Blood red Blood red chloroform becoming becoming with sulphuric cherry red. purple. acid. The color reactions are obtained by dissolving a little of the sample in a few c.c. of the solvent, adding strong sulfuric acid, shaking, and allowing the liquid to stand for some time. The results are somewhat vague and it is not impossible that a por- tion of the action is due to unknown impurities. According to Salkowski, cholesterol gives with chloroform and sulfuric acid the following effects: The solution immediately becomes blood red, afterward cherry red and purple; the last tint re- mains for several days. The sulfuric acid layer under the chloroform shows a strong green fluorescence. On pouring a few drops of the purple chloroform layer into a porcelain basin, the red color changes rapidly to blue, green, and finally to yellow. On diluting the purple chloroform solution with more chloro- form, it becomes nearly colorless, or acquires an intense blue; FATS AND OILS 1 63 if it now be shaken again with the sulfuric acid layer, the former coloration appears. These latter changes of color are due to traces of water in the chloroform. The solution of phytosterol gives the same reaction with sulfuric acid, but there is the slight difference that the coloration obtained with the former passes after a few days into a bluish- red, whereas the cholesterol solution remains more of a cherry red. In the crystallization from alcohol, if a mixture of chole- sterol and phytosterol is present, the crystals show one form either approximating to that of phytosterol or, if cholesterol is present in the greater quantity, differing from the pure crystals of either body. Analytic Data. — The data, commonly termed "con- stants," obtained by the processes described in the preceding pages, are subject to uncertainty, owing to the want of abso-' lute standards. Fats and oils, being mixtures of several in- gredients, will vary with conditions of growth of the animals or plants yielding them, methods of extracting and refining, exposure to light, heat, and air, and, doubtless, from unrecog- nized causes. Samples prepared in the laboratory do not necessarily serve as standards for commercial products. Er- rors of observation from defective apparatus, especially in- accurate thermometers, are by no means uncommon. The data for specific gravity and for melting and solidifying points given in the following tables have been compiled from the best accessible sources, and will give a general idea of the range of figures in commercial samples : 164 FOOD ANALYSIS SPECIFIC GRAVITIES OF FATS, OILS AND FATTY ACIDS. Oils. Acids. (iS.S°.) (100°.) (ioo°.) Olive, 0.914-0.918 0.875 Cottonseed, 0.922-0.925 0.8725 0.882 Maize, 0.922-925 0.871 1 Coconut, 0.912 0.868-0.874 0.844 Arachis, 0.916-0.922 0.847 Sesame, 0.922-0.924 Rape, 0.913-0.917 0.875-0.879 Cacao-butter, S 0.948-0.976 0-857 Lard, 0.932-0.938 0.859-0.864 0.837-0.840 Tallow, 0.893-0.898 0.870 Butter-fat, 0.926-0.940 0.909-0.914 Coconut olein, 0.926 0.907 lODIN NUMBERS OF FATTY ACIDS. Oil or Fat. Mixed Acids. Liquid Acids. Olive, 86-90 Cottonseed, 111-116 147 Maize, 1 13-125 140 Arachis, 95-103 1 28 Sesame, 109-1 1 2 Rape, 99-105 Coconut, 8.5-9 54 Cacao-butter, 32.5-39 Butter-fat, 28-31 Lard, 64-81 104 MELTING AND SOLIDIFYING POINTS AND TITER- The titer-tests were determined by Lewkowitsch. Oil or Fat. Acids Melting. Solidifying. Melting-point. Olive, 4 to — 2 24 to 27 Cottonseed, i to 10 35 to 40 Maize, not above — 10 18 to 20 Coconut, 20 to 28 14 to 23 24 to 2 7 Arachis, — 5 28 to ^^ Sesame, — 4 to — 6 23 to 3 1 Rape, — 6 to — 10 18 to 22 Cacao-butter, 30 to 34 20 to 27 481052 Lard, 281045 271044 351047 Butler-fat, 29 to 35 20 to 30 36 to 46 (insol.) Beef tallow, 36 to 49 33 to 48 43 to 47 Mutton tallow, 36 to 49 33 to 48 46 to 54 TESTS. Titer-test. 16.9 to 26.4 32.2 1037.6 21.2 to 25.2 28.1 to 29.2 21.2 to 23.8 II. 7 to 13.6 48.0 to 48.2 41.4 to 42.0 37.9 to 46.2 40.1 to 48.3 1 66 FOOD ANALYSIS Special Tests. — Several tests are of value for recognizing particular oils or fats. The indications for their use will be given in connection with these. Carbon disulfid-suljur test. — Halphen's test. — This is in- tended for the recognition of cottonseed oil. It is applicable both to oils and mixed acids. Carbon disulfid containing about i per cent, of sulfur in solution is mixed with an equal volume of fusel oil. Equal volumes of this reagent and the sample (about 3 c.c. of each) are mixed and heated in a bath of boiling brine for 15 min- utes. If no red or orange tint is produced, i c.c. of the re- agent is added, and if after 5 or 10 minutes more heating no color is shown, a third addition of i c.c. may be made. It is possible to detect very small quantities of cottonseed oil by this test. Lard and lard oil derived from animals fed on cot- tonseed meal will often give a faint reaction. Silver nitrate test. — Bechi^s test. — This is a test for cotton- seed oil. Several modifications are in use. According to Del Torre, the following reagents are required : A Silver nitrate, i .0 gram. Alcohol, 200.0 c.c. Ether, 40.0 c.c. Nitric acid, " o.i gram. B Fusel oil, loo.o c.c. Rapeseed oil, 15.0 c.c. 10 C.C. of the oil to be examined are mixed in a test-tube with I c.c. of reagent A, and then shaken with 10 c.c. of reagent B. The mixture is next divided into two equal por- tions, one of which is immersed in boiling water for 15 minutes. The heated sample is then removed from the water- bath, and its color compared with the unheated half. Cotton- seed oil is indicated by the reddish-brown of the heated portion. FATS AND OILS 167 Only the purest alcohol should be used, and the rapeseed oil used should be "cold drawn," and only slightly colored; it should be filtered in a hot-water oven before preparing the re- agent. To guard against errors from impurity of the reagents, a blank test should be made. It is stated that old and rancid samples will not react unless the rape oil be present. Most chemists, however, do not use it, especially in testing lard. Hehner uses reagent A, add- ing I volume to 2 volumes of oil and heating for 15 minutes. Milliau uses A with the mixed fatty acids; but experience has shown that in some cases, in which cottonseed oil was present and responded to the test, the fatty acids failed to give a similar reaction. After heating to 240° or on long keeping, both oil and fatty acids may fail to respond to the test. Furjural test. — Badouin^s test. — This is a test for sesame oil. In its original form, the sample was shaken with a mix- ture of sucrose and strong hydrochloric acid, when a crimson is produced if sesame oil be present. As furfural is a prod- uct of the action of hydrochloric acid on sucrose, and is the active agent in the test, Villavecchia & Fabris have substi- tuted an alcoholic solution of the latter for the sugar. The solution is made dilute (2 per cent.), as furfural itself gives a violet tint with hydrochloric acid. The modified test is ap- plied in one of the following forms : (a) 0.1 c.c. of the 2 per cent, furfural solution is placed in a test-tube^ lo c.c. of the sample and 10 c.c. of hydrochloric acid (sp. gr. 1.19) added, the mixture shaken for half a min- ute, and allowed to settle. In the presence of even less than I per cent, of sesame oil, the aqueous layer will become crimson. In the absence of sesame oil the lower layer is either colorless or, at most, becomes, as in the case of very rancid though pure olive oil, dirty yellow. {h) 0.1 c.c. of the furfural solution is mixed with 10 c.c. of the sample and i c.c. only of hydrochloric acid added; the 1 68 FOOD ANALYSIS mass shaken thoroughly and separation brought about by addition of lo c.c. of chloroform, or by a centrifuge, when the aqueous layer will be crimson with even less than i per cent, of sesame oil. Pyrogallol test (Tocher^ s test). — i gram of pyrogallol is dis- solved in 15 c.c. of hydrochloric acid and shaken with an equal volume of the sample. After separation, the watery liquid is boiled. Sesame oil produces a solution that is red by transmitted, and blue by reflected, light. Brulle's test, o.i gram of finely powdered egg albumin and 2 c.c. of dilute nitric acid (3 c.c. of nitric acid and i c.c. of water) are mixed with 10 c.c. of the sample, the mixture heated in a test-tube, without stirring, to boiling, and then shaken cautiously until the albumin dissolves. Care must be taken in this as the action may be violent. Cottonseed, arachis, rape and sunflower oils give red solutions; olive oil and lard yield an elaidin but no color. OLIVE OIL Olive oil is obtained from the fruit of the Olea europcea L. Its color usually ranges from light yellow to golden yellow, but some forms are deep green from presence of chlorophyl. The quality of the oil depends on many conditions; that intended for food is always expressed cold. Olive oil contains about 28 per cent, of solid fat, consisting of palmitin and a little arachidin. The remainder is mostly olein, with a little linolin. Hehner & Mitchell found no stearin. Appreciable amounts of cholesterol are present, differing from most vegetable oils, which contain phytosterol. The un- saponifiable matter ranges from i to 1.5 per cent. Free fatty acid is always present, amounting in the best grades to about 1.5 per cent., and in the lowest grades to 25 per cent. Adulteration. — Olive oil is very liable to adulteration. In this country, cottonseed oil and arachis oil are the additions OLIVE OIL 169 most commonly employed. In many cases the article con- tains no olive oil, cottonseed oil or a mixture of cottonseed and arachis oil being substituted. Other adulterants are sesame, rape, poppyseed, and lard oil. Still more rarely, curcas oil, and even castor oil have been employed. It is stated that 15 or 20 per cent, of the latter may be present with- out affecting the taste. In the lower grades of oil, not intended for table use, any ordinary oil, including refined petroleum, may be present. Specific Gravity. — The specific gravity of olive oil usually ranges from 0.914 to 0.917, or even 0.918 in the case of Cali- fornia oils. Commercial, usually brown, oils, expressed at a high temperature, and containing a higher proportion of •palmitin, may range as high as 0.925. A specific gravity of 0.918 or over, in a sample of light color, w^ould give rise to suspicion of adulteration with cottonseed, poppyseed, or se- same oil. Solidifying- point. — Olive oil has usually a higher solidifying- point than any other of the vegetable oils. Mixtures of olive with other oils have, as a rule, a lower melting-point than either constituent alone. The melting and solidifying points of the mixed acids are also of some value, but, according to Dieterich, less than 25 per cent, of adulteration cannot be detected with certainty. Saponification Value. — This determination is of use only in the case of adulteration with a considerable proportion of rape oil. lodin Number. — This determination furnishes the most valuable indications of the purity of olive oil. The figure for pure oil usually ranges between 81.5 and 85 per cent. Values as high as 88.6 have been reported from some California oils, but such samples are exceptional, and a figure above 85 should give rise to suspicion of adulteration. Heat of Bromination. — Specific Temperature Reaction. — The 16 lyo FOOD ANALYSIS thermal values of olive oil are lower than those of other vege- table oils and the determination is frequently of use. Elaidin Test. — Olive oil yields the hardest elaidin of all the oils, and in the shortest time, but, as noted on page 153, too much reliance must not be placed upon the indications of this test. The following figures, obtained by Blasdale from fresh Cahfornia oils, of known purity, serve to show that the times required to form a solid product may differ much : Time Required for Brand of Oil. Elaidin Test. Uvaria, 6 hours. Pendulina, 4 " Redding Pecholine, 3 " Nevadillo bianco, 2 " Manzanillo, 30 minutes. Refractive Power. — The refractive power of olive oil is less than that of any other of the vegetable oils. The determination of the refractive index gives reliable indications only in the presence of a considerable proportion of the adulterant. The most satisfactory results are obtained by the butyrorefracto- meter. (See table on page 165.) Nitric acid test. — This will detect small amounts of cotton- seed oil in olive oil. Some operators employ acid of 1.41 spe- cific gravity, but, according to Lewkowitsch,^* one of 1.375 gives better results. He recommends that the mixture be allowed to stand about 24 hours, when olive oil containing cottonseed oil becomes pure brown; but if rape oil be present, the mixture becomes more yellowish. Attention has been called to the fact that some highly purified cottonseed oils react so faintly with nitric acid that samples containing as much as 10 per cent, showed no reaction. The following is a summary of tests adapted to detection of the particular adulterations noted : Cottonseed Oil. Halphen's test; nitric acid color test; COTTONSEED OIL 171 Bechi's test; iodin number; Livache's test; temperature re- actions; viscosity of soap solution. Brulle's test. Arachis Oil. Viscosity of soap solution; determination of arachidic acid; iodin number. Brulle's test. Rape Oil. Iodin number; Palas' test; melting and solid- ifying points of acids ; acetic acid test ; refractive index. Sesame Oil. Furfural tests; pyrogallol test; iodin absorp- tion; temperature reactions; saponification value. Some true olive oils give a reaction simulating sesame oil with the furfural test, but this confusion may be avoided by using the mixed fatty acids; the olive oil acids do not give the reaction. Lard Oil. Melting-point of fatty acids; odor of lard on warming. Seed Oils collectively. Separation of cholesterol analogs. Castor Oil. Solubility in acetic acid in the cold; solubil- ity in absolute alcohol ; specific gravity. CuRCAS Oil. Iodin value; saponification value. Treated with nitric acid and copper, an intense reddish-brown is pro- duced in presence of as little as 10 per cent, of curcas oil. Hydrocarbon Oils. Determination of unsaponifiable mat- ter. Green olive oil has been imitated by coloring other oils with copper acetate. All green oils should be tested for copper by boiling with hydrochloric acid and testing the acid solution, as described on p. 58. COTTONSEED OIL Cottonseed oil is obtained from seeds of several species of Gossypium. The crude product is dark red. It is refined by treatment with alkali. The refined oil is pale yellow, of pleas- ant flavor, and neutral, but becomes rancid gradually, when free acid is also formed and a so-called ''stearin" deposited. The better grades of oil are sold after being freed from stearin by chilling or long standing. The refined oil is used for cooking 172 FOOD ANALYSIS purposes and as a salad oil, as an adulterant for olive oil, butter, lard, and lard oil, and in the manufacture of butter substitutes. It is so cheap that it is but little liable to adulteration, except possibly with mineral oils. Cottonseed oil contains stearin, palmitin, olein, and linolin. A small proportion of a hydroxy-ester is said to be present. Cottonseed Stearin. — This is a commercial name of the solid fat deposited on standing or by cooling the oil and pressing. The product differs according to the completeness with which the oil has been separated. The proportion of true stearin appears to be very low. A sample examined by Hehner & Mitchell yielded only 3 per cent, of stearic acid. As ordinarily obtained the fat is light yellow and of the consistency of butter. It is largely used in the preparation of substitutes for butter and lard. The following are some of the constants of this fat: o -a -J. 15-5° 100° o/r i o/: 100° Specific gravity, ^^ = °-9^2> ^^o ^ 0.864 to 0.869 ^^. Solidifying-point, 26° to 40°; titer test, 16°. . Saponification value, . 194-195. lodin value, 89-104. Mixed Fatty Acids. Solidifying-point, 35°. Melting-point, 27° to 30°. lodin number, 94. Cottonseed stearin responds to the color tests for cottonseed oil. Another variety of so-called cottonseed stearin is the solid portion of the fatty acids separated from the oil in the pro- cess of purification by alkali. It consists chiefly of stearic acid and is employed in soap-making. MAIZE OIL CORN OIL Maize oil is obtained by expression from the seeds of the Zea Mays L., either directly or after they have been used for MAIZE OIL CORN OIL 1 73 the preparation of alcohol. The latter product contains much free acid. The most recent and extended investigation of this oil is that made by Vult^ & Gibson.^^ Data furnished by them, together with some from other sources, have been incorporated in the tables on pages 164 and 165. The following additional figures are from their paper. Acid value, 2.25. Free add (percentage), 1.12. Insoluble acid, 92.2. Elaidin test, Orange-yellow deposit. Bechi's test, Dark brown. Many esters are present, as the following acids have been obtained from the saponified material: Formic, acetic, stearic, palmitic, arachidic, hypogeic, oleic, HnoHc, ricinolic (probably), and, according to some investigators, caproic, caprylic, and capric. The results of different investigators do not agree in some points. Hehner & Mitchell were unable to find stearin in a sample examined by them. J. C. Smith found volatile a^ids equivalent to a Reichert number between 2 and 3. Hop- kins found no volatile acids in the sample examined by him. The oil is practically without drying power at the ordinary temperature. According to Smith, no decided siccative prop- erties are communicated to it by simply ''boiling" or by the addition of litharge. On passing a current of air through it for an hour at a temperature of 150°, it becomes slightly darker and rather more viscous, but not to the same extent as cotton- seed oil. If to the oil so treated a small quantity of manganese borate be added, slight siccative properties are acquired, and a thin film on lead dries in from 10 to 20 hours, but not com- pletely. Hopkins found that on heating the untreated oil in the water-oven, a small amount of oxygen was absorbed, the increase in weight amounting to about i per cent, at the end of 24 hours. The unsaponifiable matter was high in the samples exam- 174 FOOD ANALYSIS ined by Vulte & Gibson, the cholesterol analog (probably si- tosterol) being 1.4 per cent, and lecithin about i.i per cent. Gill & Tufts propose to detect maize oil in cottonseed oil by applying the method described on page 161. From known mixtures of the two oils, they obtain the following weights of material the melting-point of which in each case coincided with that of sitosterol : Pure cottonseed, 50 grams yielded 0.095 Cottonseed 45, maize 5, " " 0.120 40, " 10, " " 0.16 A characteristic reaction of the oil is to dissolve it in carbon disulfid, add a drop of sulfuric acid and allow the mixture to stand for 24 hours, when it will become violet. ARACHIS OIL Arachis oil — also called peanut, ground-nut, and earth-nut oil — is obtained from the seed of the Arachis hypogcBa L. The cold expressed oil from the first runnings is nearly colorless, and that of the second expression usually of a pale greenish- yellow. It has an agreeable odor and flavor, but may be ob- tained nearly odorless and tasteless. It contains olein, pal- mitin, stearin, arachin, lignocerin, and probably hypogein. It is used as a salad oil. So-called "peanut butter" consists simply of the ground roasted nuts. The principal use of the oil is as an adulterant for olive oil. The specific gravity and chemi- cal constants of the two oils are so nearly alike that the detec- tion of the admixture by these data is hardly possible. The determination of the iodin value is occasionally of use, but the only reliable method is that of Renard, depending upon the estimation of the amount of arachidic acid, or, more properly speaking, of the arachidic and lignoceric acids, since later in- vestigation has shown that the body separated and weighed as arachidic acid consists of both, lignoceric acid being in larger proportion. The method is laborious, and requires considerable ARACHIS OIL 175 care in its performance ; many shorter methods have been pro- posed, none of which are as satisfactory as the original method, which in its most improved form is described by Archbutt, as follows : 10 grams of the oil are saponified in a deep porcelain basin, using 8 c.c. of aqueous sodium hydroxid solution (50 grams in 100 c.c.) and 70 c.c. of alcohol. The basin is covered, the mass gently evaporated to about 20 c.c, rinsed with hot water into a separating funnel, mixed with slight excess of hydrochloric acid, and shaken with ether to dissolve fatty acids. Two extractions are sufficient. After washing the ether with water, it is distilled in a 250 c.c. flask, the fatty acids dried by heating the flask on a steam-bath and sucking out the vapor, and then dissolved in the hot flask in 50 c.c. of 90 per cent, alcohol. The solution must not be allowed to cool below about 38°, lest crystals of lignoceric or arachidic acid should separate. 5 c.c. of a 20 per cent, aqueous solution of lead acetate are added and the mixture cooled to about 15°, shaken, allowed to stand for half an hour, washed only once with ether, the mass rinsed back into the flask with a spray of ether, and digested with ether for a short time; then again filtered and again rinsed back. After doing this about four times, the lead oleate will be dis- solved. The filter is opened in a large plain funnel placed iii the neck of a separating funnel, and the soaps at once rinsed into the separator with a jet of ether. The material that adheres to the paper and flask is decomposed and transferred by rinsing with hot dilute hydrochloric acid, followed by ether. About 20 c.c. of hydrochloric acid (i.io sp. gr.) are poured into the separator, shaken well to decompose the lead soaps, the aqueous liquid drawn off, the ether repeatedly washed with small quan- tities of cold water until the lead chlorid is removed, distilled in a 250 c.c. flask, and the residual fatty acids thoroughly dried by heating on a steam-bath. 50 c.c. of ethyl alcohol of exactly 176 FOOD ANALYSIS 90 per cent, strength (sp. gr. 0.834) are poured into the flask, which is then closed with a cork carrying a thermometer, heated cautiously until the fatty acids have completely dissolved, and cooled to 15°, when lignoceric and arachidic acids, if present, will crystallize out, either at once or shortly. To estimate the amount, the flask should be kept for one hour, with occasional agitation, in a water-bath at either 15° or 20°, or at some intermediate fixed temperature which is nearest to that of the laboratory, the crystals collected on a small filter, using only the filtrate to rinse the flask, and washed with three portions of 10 c.c. each of 90 per cent, alcohol, at the same fixed temperature. A paper filter may be used, but a Gooch filter, used with gentle suction, is better, as the mother liquid is more perfectly removed and the washing more thorough. The filtrate and washings with 90 per cent, alcohol are poured into a measuring cylinder, and the acids thoroughly washed with 70 per cent, alcohol, in which arachidic and lignoceric acids are quite insoluble, until some of the washings give no precipitate when diluted with water. These washings are thrown away. It is not absolutely necessary, but it is advisable to redissolve the fatty acids thus obtained in 50 c.c. of 90 per cent, alcohol, and recrystallize them, filtering and washing as before, adding the filtrate and washings with 90 per cent, alcohol to the first quantity in the measuring cylinder. Pure arachidic and lignoceric acids are thus obtained, and are dissolved off the filter with boiling ether, distilled down, and weighed in a tared flask after drying at 100° for an hour. To the weight obtained is to be added the amount dissolved by the 90 per cent, alcohol, which is calculated from the following table, based on deter- minations made by Tortelli & Ruggeri, and confirmed by Arch- butt. It will be noticed that the amount dissolved varies ac- cording to the weight of mixed acids obtained : SESAME OIL 177 Weight of Arachidic and Correction per 100 cc of 90 Per Cent. LiGNOCERic Acids Alcohol Used for Crystallization (Gram). and Washing (Gram). (iS° C.) (17.5° C.) (20° C.) 0.1 or less, 0.033 0039 0.046 0.2 0-3 0.4 0.5 0.6 07 0.8 .0.048 0.056 0.064 .0.055 0.064 0.074 .0.061 0.070 0.080 .0.064 0.074 0.085 .0.067 0.077 0.088 .0.069 0.079 0.090 .0.070 0.080 0.091 0.9 and upward, 0.071 0.081 0.091 The proportion of arachidic and lignoceric acids which has been obtained by different observers from arachis oil is very fairly constant, averaging about 5 per cent., so that the amount of these acids found in any given mixture of oils, multiplied by 20, will give a close approximation to the amount of arachis oil present. SESAME OIL Sesame oil (also called Gingli and Teel oil) is obtained from the seeds of the Sesamum orientale L. and S. indicum L. The col^ expressed oil is yellow and of pleasant taste. It consists of stearin, palmitin, olein, and linolin, with other bodies not clearly understood. Sesame oil has been used as a compulsory addition to butter- substitutes, in order to facilitate the detection of these. It is readily recognized by the furfural and pyrogallol tests. Adulteration. — Sesame oil is liable to adulteration, more especially with cottonseed, arachis, poppyseed, and rape oils. These may be detected as follows: Cottonseed oil. Halphen's, nitric acid, and Bechi's tests; Livache's test; melting-point of the fatty acids. Rape oil. Saponification value; specific gravity; solidifying and melting points of the fatty acids. Poppyseed oil. lodin value; temperature reactions. Arachis oil. Specific gravity; determination of arachidic acid. 178 FOOD ANALYSIS RAPE OIL Rape oil is obtained from several varieties of the Brassica campestris L. The oils derived from all of these are, as a rule, described indiscriminately rape oil or colza oil; but on the continent of Europe- ''colza oil" is sometimes taken to mean only that from a particular variety (napus). The physical and chemical characters of all the varieties appear to be practically identical. Rape oil is pale yellow, has a peculiar smell, and rather an unpleasant taste. It consists chiefly of stearin, olein, and erucin. It also contains a small proportion of arachidin. About 0.4 per cent, of arachidic acid is said to have been sepa- rated from it. It is very Hable to adulteration, but is of interest here only as an adulterant of olive oil. The physical and chemical characters are given in the tables on pages 164 and 165. Palas^ test. — A dilute solution of fuchsin (about i per cent.) and a strong solution of sodium acid sulfite (about 30 per cent.) are prepared separately. 20 c.c. of each of these are mixed, 200 c.c. of water added and 5 c.c. of strong sulfuric acid. When the solution is decolorized, 10 c.c. of the sample should be shaken with it. A partial restoration of color will occur if rape oil be present. It will be well to shake in a vessel full of the mixture, as contact of air may produce color. It must also be borne in mind that several aldehydes, especially formalde- hyde, will produce color with this test. COCONUT OIL Coconut oil is obtained from kernels of the coconut (species of Cocos), being usually expressed with aid of heat. It is nearly white and about the consistency of butter; has the taste and odor of the coconut. It contains palmitin and stearin, much myristin and laurin, with some caprin, caproin, and caprylin. It gives, therefore, a notable amount of volatile acids and soluble acids. CACAO-BUTTER 1 79 By treatment with alcohol and animal charcoal, a white neutral product of agreeable flavor and good keeping qualities is ob- tained which is sold for food purposes under fanciful names, such as "vegetable butter," ''vegetaline," ''laureol," ''nuco- line." By submitting the oil to pressure products termed "coconut olein" and "coconut stearin" are obtained. From samples of these, Allen has obtained the following data : Sp. Gr. (water at 15.5°= i) Solidifying- Melting- Reichert AT 15.5°; AT 98-99°. . POINT. POINT. NuMBER. Olein, 0.926 0.871 4 rising to 8 5.6 Stearin, solid 0.869 21.5 rising to 26 28.5 3.1 For its recognition, the Reichert-Meissl number is most satisfactory. (See the constants on page 165.) CACAO-BUTTER Cacao-butter is the fat expressed from cacao beans. It is yellowish-white, becoming paler on keeping, possesses the pleasant odor and flavor of chocolate, is solid at ordinary tem- peratures, but easily melts in the mouth. It consists chiefly of ste^arin, palmitin, and laurin, with small proportions of arach- idin, linolin, formin, acetin, and butyrin. It is insoluble in 90 per cent, alcohol, but dissolves in 5 parts of boiling absolute alcohol. Adulteration. — The common adulterants of cacao-butter are tallow, stearic acid, lard, paraffin wax, beeswax, coconut and arachis oils. The constants will usually suffice for their detection. Stearic acid is indicated by the high acid value ; Paraffin or beeswax^ by the low saponification value and high proportion of unsaponifiable matter; Vegetable oils, by the increased iodin value and lower melt- ing-point of the fatty acids; Coconut oil by the low iodin value, high saponification value, and moderately high Reichert number. The following special tests are also useful: l8o FOOD ANALYSIS Bjorkland^s test. — 3 grams of the fat are mixed in a test- tube with 6 grams of ether, the test-tube closed with a cork, and solution effected, if possible by shaking. When wax is present, a cloudy liquid results which is not changed on warming. If the solution is clear, the tube is placed in melting ice and the time observed after which the solution begins to become milky or to deposit white flakes; then the temperature is noted at which the mixture becomes clear on removing from the ice-water. Pure cacao-butter solution becomes cloudy in 10 or 15 minutes, and becomes clear again at 19° to 20°. With cacao-butter containing 5 per cent, of tallow, these figures are 8 minutes and 22° respectively; 10 per cent, of tallow, 7 minutes and 25°. Filsinger suggests a modified ether test: 2 grams of the fat are melted in a graduated tube with 6 c.c. of a mixture of 4 volumes of ether (sp. gr. 0.725) and 2 volumes of alcohol (sp. gr. 0.810), shaken, and set aside. The pure fat gives a solution that remains clear, even on cooling to 0°. Hager recommends the following test: About i gram of the fat is warmed with 2 to 8 grams of anilin until dissolved; the mixture is allowed to stand one hour at 15° or one and a half hours at 17° to 20°. Pure cacao-butter floats as a liquid layer on the anilin. If the sample contain tallow, stearic acid, or a Httle paraffin, particles, which remain hanging on the upper wall on gentle agitation, are formed in the oily layer. If wax or much paraffin be present, the layer solidifies. If much stearic acid be present, layers will not form, but the whole will solidify to a crystalline mass. The oily layer from pure cacao- butter hardens only after many hours. A parallel test should be made with a sample of known purity. LARD Strictly speaking, lard is the fat obtained from the mem- branes about the kidneys and intestines of the common hog. LARD l8l Commercial lard consists of the mixed fat from various parts of the animal. U.S. Standard. Lard is the rendered fresh fat from slaughtered, healthy hogs, free from rancidity, and containing not more than i per cent, of substances not fat (other than fatty acids), necessarily incorporated in the process of rendering. Leaf lard is the lard rendered at moderately high tempera- tures from the internal fat of the abdomen of the hog, excluding that adherent to the intestines, and has an iodin number not greater than 60. Neutral lard is lard rendered at low temperature. The following grades have also been given, but are not included in the official definitions. The requirement of not over 60 for iodin number of standard lard seems somewhat severe. Choice Kettle-rendered Lard. — Choice Lard. — Portions of the leaf, together with the fat cut from the backs, are rendered in steam- jacketed open kettles. The hide is removed from the back-fat before rendering. Prime Steam Lard. — The whole head of the hog, after the removal of the jowl, is used for rendering. The fat from the small intestines and fat attached to the heart are afso used. The back-fat and trimmings and the leaf may also be used. Prime steam lard, therefore, may represent the fat of the whole animal, or only portions. A lower grade is made from intestines. The definition of the term as used by hog-packers is: everything inside of a hog except the lungs and the heart, or, in other words, the ab- dominal viscera. Lard consists of stearin, palmitin, and olein, with a small amount of linolin. Hehner & Mitchell obtained stearic acid in proportions varying from 6 to 16 per cent. The unsaponifi- able matter is small; Allen & Thomson found 0.23 per cent. 1 82 FOOD ANALYSIS American and European lards differ appreciably in some analytic characters, as exhibited in the following table : Sp. Gr. — --• lODiN Number. 15° American Lards: From head, 0.8632 65.9 " back, 0.8616 63.8 " leaf, 0.8626 61 .4 European Lards: From back, 0.8607 60.5 " kidney, 0.8590 52.6 " leaf, 0,8588 53.1 More marked differences in the iodin value of fat from dif- ferent parts of the animal have been noted by other observers. Fresh lard usually contains little free acid, generally from 0.1 to 0.4 per cent., but the proportion may rise above i per cent. On exposure to the air the amount increases consider- ably. Spaeth has made a number of determinations of free acid of samples kept in loosely-corked flasks. The following is a summary of the results obtained : Fresh. i Year Old. 3 Years Old. Free acid calculated as oleic, . . 0.013 to 0.45 0.51 to 6.05 2.81014.2 Iodin number, 63.2 to 51.7 55.4 1036.7 41.11021.5 Adulteration. — Lard is much adulterated, especially with cottonseed oil, cottonseed- stearin, beef- stearin, and excess of water. Articles containing no lard have often been sold under the name ''refined lard." More recently such preparations have been designated "lard compound" or "compound lard." Maize, sesame, and arachis oils may be present in these articles. Much attention has been given to the examination of com- mercial lards, and the following is a summary of the more trustworthy of the methods. A comparison of constants will be found on pages 164 and 165. Specific Gravity. — The specific gravity of lard is usually be- tween 0.860 and 0.861. The usual adulterants, except beef- LARD 183 stearin, tend to raise the specific gravity, but they may be corrected by addition of vegetable oils. Wainwright ob- tained valuable data by compressing the sample in muslin or linen at ordinary temperatures and examining the more fluid portion. Melting-point. — This datum is usually of Httle value. Goske obtained some useful results by applying the titer-test (p. 11). Pure lards gave figures ranging from 23° to 30°; lard adul- terated with tallow and lard oil, from 29.7° to 36°. The solidi- fying-point of the fatty acids may be of value in detecting maize oil. lodin Number. — This differs considerably according to the part of the animal from which the sample is derived. The following table has been compiled from the results of many observers: American Lards. Head, 63. to 85 ; average, 75. Foot, 63, to 77 ; average, 70. Ham, 66. to 69 ; average, 67.8. Back,.. 61. 5 to 66.7; average, 64.1. Leaf, 52.51066.7; average, 59.6. Intestines, 60. English lards may give figures 6 or 8 units lower. American steam-lard derived from different parts of the animal has an iodin value of about 59 to 66, but the effect of age on this must not be forgotten (see page 182). As a rule, the iodin value of mixtures of lard, beef-stearin, and lard oil is well within these limits, so that normal iodin value is not proof of purity. The addition of vegetable oils raises the figure notably, but, according to Lewkowitsch,^^ the iodin value of the liquid fatty acids is the best method of detecting admixture of vegetable fats. With American lard, the figure is between 97 and 106; and with European lards, between 90 and 96. Should a sample give a value within the above limits, it must 184 FOOD ANALYSIS be further examined for beef- stearin and coconut oil, since these may be added with a vegetable oil to bring the figure within the limits of normal lard. Thermal Test. — The rise of temperature with sulfuric acid, and more especially the heat of bromination, is of service in the detection of cottonseed products. The results with Mau- mene's test, as reported, differ greatly. It is advisable to per- form tests with samples of pure lard and cottonseed oil side by side with the suspected sample. The initial temperature may be about 35° or 40°. Care should be taken that the sample contains no water. Rejractometric Examination. — The examination of lard by the refractometer or the butyrorefractometer is of value. Vege- table oils are readily detected, but the indications in the case of beef tallow and stearin are not so satisfactory. According to Jean, better results are obtained by operating on the liquid fatty acids. The following table is compiled from the results of Jean, Dupont, and other observers. The figures were obtained by means of a refractometer different from those figured on page 154, but the table has value for the comparative results. The liquid fatty acids may be prepared as described on page 141. Jean, whose figures are given in the table, prepared them by Sear's process: 50 grams of the lard are saponified, the fatty acids separated by addition of acid, washed with hot water, and mixed in a flask together with 250 c.c. of carbon disulfid and 8 to 10 grams of zinc oxid. The zinc salts of the liquid fatty acids dissolve in the carbon disulfid, and can thus be separated from the solid fatty acids. The carbon disulfid is evaporated, the fatty acids liberated with hydrochloric acid, well washed with hot water, and dried at a temperature of 120°. LARD 185 Degrees in Oleoreiractometer. P . Liquid Fatty Acids. American lard, mixed, _ 7 " leaf, _ii.5 " " foot, back, head, etc., — 4 to— 11 European " —12 to— 13 —30 " " stearin, — 10 to — 1 1 Beef tallow, — 16 to — 1 7 — 40 " stearin, _ 34 Veal " -19 Coconut oil, — 54 Cottonseed oil, + 1 2 to + 23 usually 4- 20 +10 " stearin, +25 +20 Arachis oil, +3.5 to + 7 — 15 Sesame " + 13 to -|- 18 — 18 European lard with 20 per cent, cottonseed oil, — 6 " " " 10 " " stearin, —7 u u cc 30 " " " .. _3 34-2 34-3 35 33-6 33.8 33-9 340 34-2 34-3 34.5 34.6 34-7 34.9 35-0 35-2 35-3 °C. 10 10.5 II. I 11.6 12.2 12.7 ^3-3 13.8 14.4 15.0 15-5 16.1 16.6 MILK AND MILK PRODUCTS 199 and dish is noted. 2 or 3 c.c. of the sample are run into the dish from the pipet, the watch-glass placed on, and the weight taken as rapidly as possible. The glass prevents appreciable loss from evaporation during an ordinary weighing. The cover is removed, the dish heated on the water-bath or in the water- oven, and weighed from time to time (with cover on it) until the weight is sensibly constant. The percentage of residue can be easily calculated. About three hours may be required to secure constant weight. Find the temperature of the milk in one of the horizontal lines and the specific gravity in the first vertical column. In the same line with this and the tempera- ture the corrected specific gravity is given. 63 64 65 66 67 68 69 70 71 72 73 74 75 21.3 21.4 21-5 21.6 21.7 21.8 22.0 22.1 22.2 22.3 22.4 22.5 22.6 22.3 22.4 22.5 22.6 22.7 22.8 23.0 23.1 23.2 233 23.4 23.5 23-7 23.3 23.4 235 23.6 237 23.8 24.0 24.1 24.2 24.3 24.4 24.6 24.7 24.3 24.4 24-5 24.6 24.7 24.9 25.0 25- 1 25.2 253 255 25.6 257 25-3 254 25.5 25.6 25-7 25.9 26.0 26.1 26.2. 26.4 26.5 26.6 26.8 26.3 26.5 26.6 26.7 26.8 27.0:27.1 27.2 273 27.4 27.5 27.7 27.8 27.4 27.5 27.6 27.7 27.8 28.0 28.1 28.2 28.3 28.4 28.6 28.7 28.9 28.4 28.5 28.6 28.7 28.8 1 29.0 i 29.1 29.2 29.4 295 29.7 29.8 29.9 29.4 295 29.6 29.8 29.9 30.1 30.2 30-3 30.4 30.5 30.7 30.9 31.0 304 30.5 307 30.8 30.9 311 312 3t.3 315 1 31.6 318 31-9 32.1 314 315 317 31.8 32.0 32.2 32.2 32.4 32.5 32.6 32.8 330 33'^ 32.5 32.6 32.7 32:9 330 33.2 33-3 33-4 33.6 33-7 33-9 340 34.2 33.5 33.6 33.8 33-9 340 34.2 34-3 34.5 34.6 34-7 34-9 35-1 35-2 34.5 34.6 34-8 34.9 350 35.2 35.3 35-5 35.6 35.8 36.0 36.1 36.3 35-5 35-6 35-8 35-9 36.1 36.2 36.4 36.5 36.7 36.8 370 37-2 37.3. tt: ~ 18.3 18.8 19.4 20 20.5 21. 1 21.6 22.2 22.7 23-3 23.8 200 FOOD ANALYSIS The A. O. A. C. method is: Heat at ioo° to constant weight, about 3 grams in a tared platinum, aluminum or tin dish of 5 cm. diameter, with or without the addition of 15 to 30 grams of sand. Cool and weigh. The use of aluminum or tin as substitutes for platinum is inadvisable, much better results will be obtained with nickel, porcelain or glass. Ash. — The residue from the determination of total solids is heated cautiously over the Bunsen burner, until a white ash is left. The result obtained in this manner is apt to be slightly low from loss of sodium chlorid. This may be avoided by heating the residue sufficiently to char it, extracting the sol- uble matter with a few cubic centimeters of water, and filtering (using paper extracted with hydrofluoric acid). The filter is added to the residue, the whole ashed, the filtrate then added, and the liquid evaporated carefully to dryness. The ash of nofmal milk is about 0.7 per cent, and faintly alkaline. A marked degree of alkalinity and effervescence with hydro- chloric acid will suggest the addition of a carbonate. The method of the A. O. A. C. is as follows: In a weighed dish put 20 c.c. of milk from a weighing bottle; add 6 c.c. of nitric acid, evaporate to dryness, and burn at a low red heat till the ash is free from carbon. Fat. — Many methods for fat determination have been de- vised. The following will suffice for all practical work: Bahcock Asbestos Method. — ^This is recommended by the A. O. A. C. : Provide a hollow cyHnder of perforated sheet metal 60 mm. long and 20 mm. in diameter, closed 5 mm. from one end by a disk of the same material. The perforations should be about 0.7 mm. in diameter and 0.7 mm. apart. Fill the cylin- der loosely with from 1.5 to 2.5 grams of freshly ignited woolly asbestos free from fine or brittle material. Cool in a desiccator and weigh. Introduce a weighed quantity of milk (about 4 grams) and dry at 100°. The cylinder is placed in the ex- MILK AND MILK PRODUCTS 20I traction tube and extracted with ether in the usual way. The ether is evaporated and thb fat weighed. The extracted cyl- inder may be dried at ioo° and the fat checked by the loss in weight. A higher degree of accuracy is secured by performing the drying operation in hydrogen. Adams' Method. — This consists essentially in spreading the milk over absorbent paper, drying, and extracting the fat in an extraction apparatus; the milk is distributed in an extremely thin layer, and by a selective action of the paper the larger portion of the fat is left on the surface. A paper, manufac- tured especially for this purpose by Schleicher & Schuell, is obtainable in strips of suitable size. Each of these yields to ether only from o.ooi to 0.002 gram of extract. Coils made of thick filter-paper, cut into strips 6 by 62 cm., are thoroughly extracted with ether and alcohol, or the weight of the extract corrected by a constant obtained for the paper. From a weighing bottle about 5 grams of the . milk are trans- ferred to the coil by means of a pipet, care being taken to keep dry the end of the coil held in the fingers. The coil is placed, dry end down, on a piece of glass and dried for one hour, preferably in an atmosphere of hydrogen; it is then transferred to an extraction apparatus and extracted with absolute ether, petroleum spirit of boiling-point about 45° or, better, carbon tetrachlorid. The extracted fat is dried and weighed. The above procedure is very satisfactory, but the drying in hydrogen may usually be omitted. After the coil has re- ceived at least twenty washings, the flask is detached, the ether removed by distillation, and the fat dried by heating in an air- oven at about 105°, and occasionally blowing air through the flask. After cooling, the flask is wiped with a piece of silk, allowed to stand ten minutes, and weighed. Richmond states that to perform a rigidly accurate deter- mination attention to the following points is necessary: The ether must be anhydrous (drying over calcium chlorid and 202 FOOD ANALYSIS distilling is sufficient). Schleicher & Schuell's fat-free papers should be used, and one should be extracted without any milk on it, as a tare for the others. Four or five hours' extraction is necessary, and the coils should be well dried before extraction is begun. Thimble- shaped cases made of fat-free paper are now ob- tainable and are convenient for holding the absorbent material on which the milk is spread. The fine texture prevents un- dissolved matter escaping. A case may be used repeatedly. Sour milk may be thinned with ammonium hydroxid before taking the portion for analysis. Werner- Schmid Method. — This is suita- ble for sour milk and for sweetened con- densed milk. I oc.c. of the milk are meas- ured into a long test- tube of 50 c.c. capac- ity, and 10 c.c. of strong hydrochloric acid added, or the milk may be weighed in a small beaker and washed into the tube with the acid. After mixing, the liquid is boiled ij minutes, or the tube may be corked and heated in the water-bath from 5 to 10 minutes, until the liquid turns dark brown. It must not be allowed to turn black. The tube and contents are cooled in water, 30 c.c. of well-washed ether added, shaken, and allowed to stand until the line of acid and ether is distinct. The cork is taken out, and a double-tube arrange- ment, like that of the ordinary wash-bottle, inserted. The stopper of this should be of cork and not of rubber, since it is difficult to slide the glass tube in rubber, and there is a pos- sibility, also, of the ether acting on the rubber and dissolving it. The lower end of the exit-tube is adjusted so as to rest im- mediately above the junction of the two liquids. The ethereal Fig. 45- MILK AND MILK PRODUCTS 203 solution of the fat is then blown out and received in a weighed flask. Two more portions of ether, 10 c.c. each, are shaken with the acid liquid, blown out, and added to the first. The ether is then distilled off and the fat dried and weighed as above. Centrijugal Methods. — Among the processes for the rapid determination of fat, those employing centrifugal action have been found most convenient. The following method, devised by Leffmann & Beam in 1889,^^ has proved satisfactory on the score of accuracy, simplicity, and ease of manipulation. This process, which antedates in its successful operation and public exhibition all the rapid centrifugal methods except the De Laval, is sometimes called the "Beimling" method, but Beimling was merely a patentee of a crude form of cen- trifugal machine, and had no part in devising the mixture for freeing the fat. The distinctive feature is the use of fusel oil, the effect of which is to produce a greater difference in surface tension between the fat and the liquid in which it is suspended, and thus promote its readier separation. This effect has been found to be heightened by the presence of a small amount of hydrochloric acid. The test-bottles have a capacity of about 30 c.c. and are provided with a graduated neck, each division of which repre- sents 0.1 per cent, by weight of butter fat. 15 c.c. of the milk are measured into the bottle, 3 c.c. of a mixture of equal parts of amyl alcohol and strong hydro- chloric acid added, mixed, the bottle filled nearly to the neck with concentrated sulfuric acid, and the liquids mixed by holding the bottle by the neck and giving it a gyratory mo- tion. The neck is now filled to about the zero point with a mixture of sulfuric acid and water prepared at the time. It is then placed in the centrifugal machine, which is so arranged that when at rest the bottles are in a vertical position. If only one test is to be made, the equilibrium of the machine is main- tained by means of a test-bottle, or bottles, filled with a mixture 204 FOOD ANALYSIS of equal parts of sulfuric acid and water. After rotation for from one to two minutes, the fat will collect in the neck of the bottle and the percentage may be read off. It is convenient to use a pair of dividers in making the reading. The legs of these are placed at the upper and lower limits respectively of the fat, allowance being made for the meniscus; one leg is then placed at the zero point and the reading made with the other. Ex- perience by analysts in various parts of the world has shown that with properly graduated bottles the results are reliable. As a rule, they do not differ more than o.i per cent, from those obtained by the Adams process, and are generally even closer. For great accuracy, the factor for correcting the reading on each of the bottles should -be determined by comparison with the figures obtained by the Adams or other standard process. Cream is to be diluted to exactly ten times its volume, the specific gravity taken, and the liquid treated as a milk. Since in the graduation of the test-bottles a specific gravity of 1.030 is assumed, the reading must be increased in proportion. A more accurate result may be obtained by weighing in the test-bottle about 2 c.c. of the cream and diluting to about 15 c.c. The reading obtained is to be multiplied by 15.45 and divided by the weight in grams of cream taken. The mixture of fusel oil and hydrochloric acid seems to be- come less satisfactory when long kept. It should be clear and not very dark in color. It is best kept in a bottle provided with a pipet which can be filled to the mark by dipping. Rigid accuracy in the measurement is not needed. See also Cochran's method under "Condensed Milk." Calculation Methods. — Several investigators have proposed formulae by which when any two of the data, specific gravity, fat, and total solids, are known, the third can be calculated. These vary according to the method of analysis employed. That of Hehner and Richmond, as corrected by Richmond, was deduced from results by the Adams method of fat extrac- MILK AND MILK PRODUCTS 205 tion, and has been found to be the most satisfactor>\ It is as follows: T = o.25 G + 1.2 F + 0.14; in which T is the total soHds, G the last two figures of the specific gravity (water being looo), and P the fat. A table based upon this formula is annexed. A formula has been devised by Richmond by which the lac- tose and proteids may be calculated (approximately), the specific gravity, fat, total soUds, and ash being known. Thus: G P = 2.8T + 2.5A — 3.33 F — 0.7 D in which P is the proteids, T the total solids, A the ash, F the fat, D specific gravity (water at 15.5° being taken as i), and G 1000 D — 1000. The difference between the total solids and the fat, proteids, and ash gives the lactose. In this formula it has been assumed that everything that is not fat, proteids, or ash, is milk-sugar, an assumption which is not strictly correct, and which intro- duces a small error. Another slight error is introduced by the fact that the ash in milk is not the same as the salts existing in the milk. Total Proteids. — For practical purposes the total pro- teids are best estimated by calculation from the total nitrogen obtained by the Kjeldahl-Gunning method. Milk contains, however, a sensible proportion of non-proteid nitrogen. Ac- cording to Munk, this may range, in cows' milk, from 0.022 to 0.034 per cent., and from 0.014 to 0.026 per cent, in human milk. By these figures, the average proteid nitrogen in cows' milk would be 94 per cent., and in human milk 91 per cent., of the total nitrogen. The determination of total nitrogen as recommended by the A. O. A. C. is to place in the digestion flask a known weight (about 5 grams) of the sample and proceed, without evaporation, as 206 FOOD ANALYSIS ON a o "^ CNJ ^ NO 8n S VO M CM s- M "^ N N to M CO M CO fO I-I ro CO M CO CO Tf l^- "* ">!*• 'il- oo ^ 00 8s CO o »r> 00 M ^ CO »o VO VO •s. a CO O VO "S % Tf M N N fO ro ro CO CO CO CO h4 -^ -"i- ^ -* t>* to ^ 00 ON ? NO 00 CO »o ^ 00 ON ? VO 00 CI ^ W i-i M N M ro ro fO CO CO CO CO CO "^ M vO ^ ^ ON On 3- NO ON ^ VO ^ ON ON t NO ■^ rj M N H4 M M CO CO CO »-l CO CO CO CO it Tl- to ;^ NO ON ON o t^ On CI VO VO On ? ■^ N N W cs IX N CO CO CO CO CO I-I CO CO ■^ M '^ t^ ^, . d. CO i^ 01 g^§2^ i2 vq 00 q On VO ^ Form OOLO Choi Firj Cro u vd CO d oi ^ 00 CO O d ^ VO '"' • u rs jT B 'o '^ rt Oj ^ _c "o ^ _c • ^a (J 1 3 D 1 c "c a d i £ < i tt 1 h C 1 TEA 255 It is probable that the proportion of caffein in the above analyses is slightly underestimated as the determination was made by treating the watery extract with magnesia, evapo- rating to dryness, and extracting with ether. The tea leaf is ovate-lanceolate with short 5tem not sharply distinguished from the blade. The distal two-thirds of the leaf is marked by serrations with slightly curved spines. At the insertion of these spines the leaf tissue is thickened. This structure is wanting in young leaf buds. The venation is a Fig. 47. — Epidermis of Under Surface of Tea-leaf. 5/>, stoma; h, hair; m, cells containing chlorophyl. (X 160.) midrib running to the extreme end of the leaf with frequent lateral nearly opposite branchings anastomosing near the edge and sending off secondary branches to the extreme edge. The apex of the tea leaf is often distinctly notched, whereas most other leaves are pointed. The stomata and hairs are fairly characteristic. Figure 47 is from Moeller's work.*' Adulteration. — The substitution of inferior grades of tea for those of finer aroma and strength is the common adulter- ation of tea. Other forms are: additions, such as sand, ex- 256 FOOD ANALYSIS hausted leaves, foreign leaves, and materials to increase astrin- gency, especially catechu. Green tea is often colored or "faced" with Prussian blue, indigo, or turmeric, and black tea with graphite. Lie tea is an imitation made of dust and sweepings of tea or other leaves along with mineral matter of various kinds and held together by means^of starch or gum. It is readily detected by the addition of hot water, when the mass breaks down into the fragments of which it is composed. The following analyses of spurious teas, received from the United States consuls at Canton and Nagasaki (Japan), were made by Battershall''^: I. 2. 3. 4. Total ash, 8.62 8.90 7.95 12.58 Ash insoluble in water, 7.98 6.04 4.95 8.74 Ash soluble in water, 0.64 1.86 3.00 3.84 Ash insoluble in acid, 3.92 3.18 1.88 6.60 Extract, 7.73 14.00 1276 22.10 Gum, 10.67 7-3° 11.00 11.40 Insoluble leaf, 70.60 70-55 67.00 60.10 Tannin, 3.13 8.01 14.50 15.64 Caffein, 0.58 none 0.16 0.12 1. Partially exhausted and refired tea leaves, known as '^Ching Suey^' (clear water), which name doubtless has reference to the weakness of a beverage prepared from the article. 2. "Lie tea," made from Wampan leaves. 3. A mixture of 10 per cent, green tea and 90 per cent, "lie tea," sometimes sold as "Imperial" or "Gunpowder" tea. 4. "Scented caper tea," consisting of tea dust made up into little shot-like pellets by means of "Congou paste" (i. e., boiled rice). Analytic Methods. Water. — This is determined as on page 27. A slight amount of caffein may be lost in the drying and counted as water, but the error is negligible. Ash. — Soluble ash and alkalinity of soluble ash. (See page 39)- TEA 257 Extract. — 2 grams of the finely powdered tea are boiled for an hour in a flask provided with a reflux condenser. The liquid is decanted and the residue boiled for a short time with suc- cessive portions of 50 c.c. of water until this is no longer colored. The solutions are mixed, heated, filtered through a tared filter, to which the insoluble leaf is also transferred. After washing with boiling water, the filter and contents are dried to constant weight. The extract is determined by difference, or, if desired, the filtrate is made up to a definite volume, and an aliquot por- tion evaporated and dried at 100° and weighed. Nitrogen. — Total and albuminoid nitrogen is determined by the methods described on pages 33 and 37. Cafjein. — This is best determined by Allen's method: 6 grams of the finely powdered tea and 600 c.c. of water are boiled under a reflux condenser for six or eight hours; 4 grams of lead acetate in powder are then added and the liquid again boiled for ten minutes. If, on removing the source of heat, the pre- cipitate does not curdle and settle readily, leaving the liquid colorless or nearly so, a further addition of lead acetate must be made and the boiling repeated. When clarification is ef- fected, the liquid is passed through a dry filter, 500 c.c. of the filtrate (5 grams of the tea) are evaporated to about 50 c.c, and a little disodium hydrogen phosphate is added to precipitate the remaining lead. The liquid is filtered, the precipitate washed, and the filtrate further concentrated to about 40 c.c, when the caffein is extracted by at least four agitations with chloroform. The separated choroform solutions are mixed, and distilled in a tared flask immersed in boiling water. While the flask is still hot the last traces of chloroform are removed by a current of air, and the residual alkaloid is weighed. Determinations of caffein based upon the treatment of the leaves with boiling lime water or alkali are valueless, as is also the process of Paul & Cownley, in which the leaves are mixed with magnesia, dried and exhausted by alcohol. 23 258 FOOD ANALYSIS The following volumetric method, due to Gomberg, has been reported upon favorably by Ladd: A weighed quantity of the tea is boiled with water as above, the solution made up to a known volume, and filtered. An aliquot portion of the filtrate is treated with lead subacetate so long as a precipitate is formed. After standing, the pre- cipitate is filtered off, the excess of lead carefully removed by hydrogen sulfid, the filtrate from the lead sulfid boiled to re- move hydrogen sulfid, and divided into two equal parts. One portion is acidified with sulfuric or hydrochloric acid and ex- cess of decinormal iodin solution added; after standing 5 to 10 minutes is is filtered and the filtrate titrated with decinormal thiosulfate solution. If in the other portion potassium iodid- iodin solution (page 26) produces a precipitate, a correction is necessary, i c.c. of decinormal thiosulfate corresponds to 0.00458 gram of caffein. Facing. — The coloring-matter used in facing is usually present in minute amount, and is best detected by the microscope, the leaf being examined by reflected light. A good plan is to shake some of the leaves with water, allow the suspended matter to settle, and examine the sediment by the microscope and chemic- ally. Prussian blue may be distinguished from indigo by the fact that the color of the former is discharged by addition of sodium hydroxid. Indigo forms a deep blue solution with sulfuric acid. Turmeric is detected as on page 73. Graphite may be detected by examination under the microscope. Added Mineral Matter. — Any considerable addition of mineral matter will be shown by the increased proportion of ash, which usually ranges from 5 to 6.5 per cent., and only in exceptional cases rises to 7.5 per cent. Magnetic iron oxid and particles of iron have been found in tea, and may be readily separated from it by the magnet. Sand and powdered brick have also been found. The former may be accidental. Exhausted Tea Leaves. — The detection of admixture of TEA 259 moderate proportion of added tea leaves is difficult. Con- siderable addition will be indicated by the decreased propor- tion of extract and caffein, and especially of soluble ash and its alkalinity. The soluble ash of pure tea is from 2.5 to 4 per cent., and is usually over 3 per cent., v^hereas that of exhausted tea is generally not over 0.8 per cent. The alkalinity of the soluble ash expressed as potassium oxid is from 1.25 to 2 per cent, (cal- culated on the dry tea). In exhausted tea the alkalinity is likely to be less than 0.3 per cent. The soluble ash is best calculated to percentage of total ash. The interference of sand may be eliminated by calculating the proportion of ash soluble in water to that soluble in acid. Wigner obtained the following average results from the ex- amination of 67 samples of tea: Siliceous matter, 7.96 per cent. Soluble in acid, 37-54 " " " "water, 54.50 " " Alkalinity of soluble ash, 25.09 " " Excluding the portion insoluble in acid, the figures become: Soluble in water, 59-2 1 per cent. Alkalinity of soluble ash, 27,26 " " If the soluble ash is less than 40 per cent, of the total ash or less than 45 per cent, excluding siliceous matter, adulteration with exhausted leaves may be suspected. The minimum proportion of extract yielded by pure tea is, ac- cording to the standard fixed by the Society of Public Analysts in 1874, not less than 30 per cent. The proportion usually found much exceeds this figure, but congou may contain less. The proportion of caffein found by different observers ranges from 1.8 to 4 per cent., the lower proportions being found in Japan teas. Exhausted leaves have in some instances been found to be partly unrolled or much frayed and broken, and more posi- 26o FOOD ANALYSIS tive indications might be had by the examination of selected leaves of suspicious appearance. Foreign Astringents. — Catechu is sometimes added, espe- cially to ''lie" or ''caper" tea, or to mask the presence of ex- hausted leaves. It may be detected by Hager's test: About a gram of the sample is boiled with water, the extract treated with excess of lead monoxid, and filtered. A solution of silver nitrate is added to clear the filtrate; in the presence of catechu, a yellow flocculent precipitate, which rapidly becomes dark, is formed. Pure tea gives only a slight grayish precipitate of silver. Allen recommends the following process, which should be applied to the suspected tea, side by side with a genuine sam- ple: I gram of the pure tea, and an equal weight of the sus- pected sample, are infused in separate portions of loo c.c. each of boiling water, and the strained liquid precipitated while boiling with a slight excess of neutral lead acetate. 20 c.c. of the filtrate from the pure tea (which should be colorless), when cautiously heated and treated with a few drops of sil- ver nitrate solution, avoiding excess, gives only a very slight grayish cloud or precipitate of reduced silver; but the same tea containing 2 per cent, of added catechu gives a copious brown- ish precipitate, the liquid acquiring a distinctly yellowish tinge. With a somewhat larger proportion of catechu, the filtrate from the lead precipitate gives a bright green color on adding one drop of dilute ferric chlorid, while the solution from pure tea gives only a slight reddish color, due to the presence of acetate. On allowing the liquid to stand, the adulterated tea gives a pre- cipitate of a grayish or olive-green color, the pure tea under- going no change. Foreign Leaves. — A small proportion of foreign leaves, such as those of the rose, jasmine, and orange, are sometimes added to impart bouquet, but these are usually removed before pack- ing. Other foreign leaves, especially the sloe, willow, elder, Chloranthus inconspicuus, Camellia sasanqua, and Eurya TEA 261 chinensis, have been added in considerable quantity, but the practice, so far as concerns the tea shipped to the United States, seems to be less common than formerly. The detection of such additions is best made by the appearance of the leaf and the microscopic examination, but a few chemical tests have been proposed v^hich may be of some assistance. Blyth proposes to utilize the presence of manganese, which is a constant con- stituent of the ash of tea. The suspected leaf is ashed and the ash treated on fused platinum foil with potassium nitrate and carbonate. The distinct green color due to a manganate is readily recognized. Allen has applied the test to various leaves and found manganese to be present in the following: Species of Thea (tea), Camellia sasanqua, C. japonicay coffee, beech, blackberry, and sycamore. Manganese was absent from the leaves of the hawthorn, ash, raspberry, cherry, plum, and rose, and only faint traces were detected in the leaves of the Ilex Paraguay ensis J elm, birch, lime, sloe, elder, willow herb, and willow. Blyth has also proposed the following test, de- pending upon the isolation of caffein and recognition by its crystalline form under the microscope: The leaf or fragment is boiled for a minute in a watch-glass with a very little water, an equal bulk of calcined magnesia is added, and the whole heated to boiling and rapidly evaporated to a large-sized drop. This drop is transferred to a subliming cell, and if, after heating to about 110°, no crystalline sublimate of caffein is obtained, the leaf cannot be a tea leaf. If, however, a sublimate of caf- fein is obtained, it is not conclusive evidence, since other plants contain the alkaloid. More satisfactory results are obtained by the examination of the shape and venation of the leaf. The sample should be softened by soaking in hot water, carefully unrolled, trans- ferred to a microscope slide, and examined with a hand lens. Such examination will usually be sufficient, but in doubtful cases it may be necessary to use higher powers. 262 FOOD ANALYSIS COFFEE Coffee is the seed of species of Cojfea, cultivated in sub- tropical climates. The fruit usually consists of two seeds surrounded by a pulp, which is removed by fermenting and washing. The membranous pericarp removed by machinery is sometimes roasted and used as a substitute for coffee. The following are the more important constituents of raw coffee: An essential oil, fat, caffetannates, caffein, and caf- fearin. The essential oil has been little studied. The fat of coffee is soluble in alcohol, but its composition is not yet clearly ascertained. Caffetannic acid is crystalline, astringent, soluble in water, less soluble in alcohol, and very sparingly in ether. It gives a dark green coloration with ferric chlorid, and does not pre- cipitate gelatin. Coffee contains a fairly constant proportion of caffein (see page 263). According to Paladino, there is also present a nar- cotic alkaloid, which he calls caffearin. Paladino's results seem to be corroborated by those of Forster & Riechelmann, who found an alkaloid distinguished from caffein by the fol- lowing characteristics: failure to respond to the murexid test, precipitability by picric acid, and insolubility in chloroform. Roasted coffee contains a small amount of sugar, which, according to Spencer, consists largely of sucrose. It appears to be absent from raw coffee and is derived from the decom- position of the glucosids (tannins). The aroma of roasted coffee is due to cafjeol, which may be separated by distilling with water, agitating the distillate with ether, and evaporating. It is an oily liquid, slightly soluble in hot water, but easily soluble in alcohol and ether.. By fu- sion with caustic soda it yields sodium salicylate. The phy- siological effects of coffee are attributed to the caffeol, caffein, and caffearin. The roasting of coffee results in a notable reduction of some COFFEE 263 of the constituents, especially the caffein, fat, and sugar. When properly conducted, the total loss in weight amounts to from 12 to 18 per cent., of which about 8 per cent, represents moisture. Konig gives the following figures, calculated as percentage of moisture-free material : Raw. Roasted. Soluble in water, 30-93 28.36 Total nitrogen, 2.21 2.38 Caffein, 1.33 1.42 Fat, 14-91 16.14 Sugar, 3.66 1.35 Fiber, 31.24 25.07 Other nitrogen-free matter, 34.55 39.84 Ash, 3.92 3.87 Coffee is sometimes glazed with sugar before roasting. Ac- cording to Konig, when so treated it retains much more mois- ture. According to Hilger & Juckenack, glazed coffee requires to be heated to a much higher temperature, which results in about double the usual loss of caffein and fat. R,aw coffee is subject to less adulteration than roasted and especially ground coffee. Coffee beans dififer considerably in size and quality according to their origin, and the inferior kinds are sometimes so treated as to give them the appearance of the better quahties. West India coffee is for the most part even-sized, pale and yellowish, firrn and heavy, with fine aroma, losing little weight by the roasting process. Brazil coffee is larger, less solid, greenish or white, usually styled by the brokers "low" or ''low middlings." Java coffee is smaller, shghtly elongated, pale in color, light and deficient in essential oil. Ceylon coffee is of all descriptions, but the ordinary planta- tion products are even-colored, slightly canoe-shaped, strong in aroma and flavor, heavy, and permit of adulteration more than other kinds. 264 FOOD ANALYSIS Mocha coffee is usually considered the best, but very little reaches the United States. Porto Rico coffee is often called Mocha. The grains of Mocha coffee are small and dark yellow. Java coffee, when new, is pale yellow, and is then cheaper than when old and brown. This color is partly the effect of curing as well as the result of age. Java coffee, being of high price, has been imitated by color- ing the cheaper grades with dyes or mineral pigments. According to Waller, Java coffee is imitated by exposing South American coffee to a high moist heat, by which the color is changed from green to brown. Raw coffee is heavier than water. Fade gives the specific gravity of raw coffee berries at from 1.041 to 1.368. Dam- aged coffee that has been washed and partially roasted to im- prove the color may have a specific gravity less than i . Roasted coffee has a specific gravity of from 0.500 to 0.635, t>ut samples that have been made to take up much water by steaming and then coating with glycerol or sugar (see page 263) may possess a specific gravity appreciably higher (0.650 to 0.770). Implicit reliance should not be placed on these figures, since over- roasted coffees may be heavier than water. The specific gravity of raw coffee may be determined by immersing the beans in strong brine and cautiously adding water until they remain suspended in the liquid. The specific gravity of the liquid is then determined as usual. In the case of roasted coffee the brine is replaced by petroleum spirit to which is gradually added ordinary petroleum. Adulteration with exhausted coffee beans is reported by Roos. The samples examined yielded only i per cent, of ether extract. Facing. — The following are reported to have been used as ''facing" for coffee. Scheele's green, chrome yellow, ochre, silesian blue, burnt umber, Venetian red, charcoal, indigo. COFFEE 265 ultramarine blue, clay, gypsum. A blue color is also said to be produced by shaking the beans with finely powdered iron. The beans are sometimes polished by rotating in a cylinder with soapstone. The examination for facing should be made with the micro- scope, and also by shaking with water, and examining the sediment, as described under tea (page 258). Artificial colors may usually be detected by treating the beans with strong alcohol, evaporating to dryness, and testing the residue (see pages 64 and 66). Imitation beans have frequently been sold for use in mixing with coffee. In some cases these are molded in close imita- tion of the true beans. The material used for the purpose is sometimes clay, but more frequently one or more of the fol- lowing: Wheat flour, chicory, bran, rye, peas, and acorns. These are often mixed with molasses. Ferrous sulfate has also been found. Most imitation coffee is heavier than water, but the readiest means of detection is by means of the microscope, the appli- cation of the iodin test for starch, and determination of the ash. Many substances have been used as substitutes for coffee as well as for its adulteration; among these are chicory, Mogdad and Mussaenda coffee, roasted cereals and leguminous seeds, cocoa husks, and figs. Coffee contains no starch, a constituent of many adulter- ants, such as cereals and acorns. It may be detected by Allen's method: The coffee is boiled for a few minutes with about ten parts of water. When the liquid has become perfectly cold, some dilute sulfuric acid is added, a strong solution of potassium permanganate is dropped in cautiously, with agita- tion, until the coloring-matter is nearly destroyed, when the liquid is strained or decanted from the insoluble matter and iodin added. A distinct reaction occurs in the presence of even i per cent, of starch. In identifying the starch granules 24 266 FOOD ANALYSIS hp— - qu with the microscope it is advisable to make a preHminary ex- traction of the sample with ether, and subsequently with alcohol. Chicory is the root of the Cichorium intyhus L. Its micro- scopic structure distinguishes it from coffee. The cells of the parenchyma are large, smooth-walled, and regular. The milk ducts are branched and filled with a coarsely granular material. The body of the root contains long, pointed cells presenting a characteristic dotted appear- ance. (See Fig. 48.^') It contains no starch. Dande- lion and other sweet roots \ V^l '^; i/^^^^t^^ '/fi^l present a somewhat similar \ V ^lli- i = :^ ^^^^0 \mlli structure, but the ducts are \ ^^ipi^^^^S if / / scaliform, the cells larger, and milk vessels are absent. Rimmington recommends the following method for the detection of chicory: The sample is boiled for a short time with water containing a little sodium carbonate; the solution is decanted and the residue treated with a solution of bleaching pow- der for several hours, when decolorization will be ef- fected. The coffee will be found as a dark stratum at the bottom of the beaker and the chicory as a light stratum above it. Analytic Methods. — The following preliminary .tests may be of value. A small quantity of the ground material is sprin- kled on cold water. Coffee will usually float, and impart very little color to the water. Chicory and most other additions sink, and the caramel contained in them dissolves quickly, forming Fig. 48. g, Vascular tissue; hp, parenchyma; fibers; m, medullary rays. COFFEE 267 a dark and usually turbid solution. Coffee grains are hard, whereas chicory and some other adulterants, after maceration for some hours in water, are quite soft. At the end of this time if the mixture be transferred to a piece of stretched cloth and rubbed with a pestle, the chicory will pass through. The proportion of the adulterant which has been detected by the microscope or the preliminary tests just mentioned may often be determined with a fair degree of accuracy by chemi- cal examination, especially by the determinations of fat, caf- fein, water extract, and ash. The actual amount of coffee present may be determined by calculation from the caffein present determined by the process given on page 257, using double the quantity of material. In the presence of chicory the extracted alkaloid is liable to be strongly colored, and Allen recommends that it be redis- solved in water, a few drops of sodium hydroxid added, and the liquid again extracted with chloroform. Caffetannic acid may be determined by Krug's method*®: 2 graijis of the material finely powdered is digested for 36 hours with 10 c.c. of water at a moderate temperature, then 25 c.c. of 90 per cent, alcohol added and the digestion continued for 24 hours. The liquid is filtered and the precipitate washed with 90 per cent, alcohol. The filtrate is heated to boiling, and a boiling concentrated solution of lead acetate added. When the precipitate (lead caffetannate) has become floccu- lent, it is separated, washed on the filter with alcohol (90%), until the washings are free from lead (ammonium sulfid being used as a test), and then with ether, until free from fat. It is dried at 100° and weighed. The weight multiplied by 0.516 gives the caffetannic acid. The proportion of caffein in roasted coffees, ranges from 0.8 to 1.3 per cent. In the better grades it probably does not go below I.I per cent., 1.2 might be taken as a basis of calcula- tion. 268 FOOD ANALYSIS Fat. — The fat of coffee may be determined by extracting with petroleum spirit or carbon tetrachlorid the material dried at 1 00°. According to Macfarlane, the petroleum spirit ex- tract from previously dried coffee usually ranges from 10 to 12 per cent. Only one sample out of nearly fifty showed less than 10, and 12.5 per cent, was reached only in a few cases. Water-extract. — Valuable indications are often furnished by the determination of the amount of water-extract, which is fairly uniform and little affected by the usual variations in ex- tent of roasting. The determination is simplified by the ob- servation of the specific gravity of the solution in water as recommended by Graham, Stenhouse, and Campbell. One part of the sample is treated with ten parts of water, the liquid heated to boiling, cooled to 15.5°, and the specific gravity taken. The following figures were obtained in this manner: Mocha cofifee, 1008.0 Turnips, 102 1 .4 Neilgherry coffee, 1008.4 Dandelion, 102 1 .9 Plantation Ceylon coffee, 1008.7 Red beet, 1022. i Native Ceylon coffee, 1009.0 Marigold wurzel, 1023 .5 Java coffee, 1008.7 Lupins, 1005 .7 Jamaica cofifee, 1008.8 Peas, 1007.3 Costa Rica cofifee, 1009.0 Beans, 1008.4 " average, 1008.7 Brown malt, 1010.9 {1019.1 Black " 1021.2 to Rye meal, 102 1 .6 1023.6 Maize, 1025.3 " average, 1021.0 Bread raspings, 1026.3 Parsnips, 1014.3 Acorns, 1007.3 Carrots, 1017.1 Spent tan, 1002. i According to McGill, the specific gravity of the infusions of coffee and chicory are materially affected by the fineness of powder and the time occupied in heating the solution to boiling, and the duration of the boiling. He recommends the following process: 10 grams of the dried, finely powdered sample are heated with 100 c.c. of distilled water in a flask provided with a reflux condenser. The heat is adjusted so that ebulHtion COFFEE 269 commences in 10 to 15 minutes, and the boiling continued for exactly one hour; the liquid is allowed to stand for 15 minutes, and then passed through a dry filter. The average specific gravity of the decoction from pure coffee was found to be 1009.86 at 17°, and that of chicory, 1028.21. The amount of coffee present in a mixture of coffee and chicory may be approximately calculated by deducting the observed gravity from 1028.21 and multiplying the remainder by 5.45. Macfarlane has determined the water extract by boiling with water the dried residue from the determination of fat (page 268) and redrying and weighing the residue. The water-extract is determined by difference. The following results were obtained : Cofifee (Santos, Mocha, and Java), 20.4-22.4 per cent. Chicory, 77.7 " " Hehner has found highly roasted chicory to give a water-ex- tract as low as 54.1 per cent, and a specific gravity of the 10 per cent, solution of 1019. Cassal has found genuine coffee to give a water-extract as high as 29 per cent. More recently several observers have called attention to the fact that the proportion of water-soluble matter in commercial chicory may be markedly greater than that found in the above samples, examined years ago. This appears to be due, as pointed out by Dyer, to the less roasting to which it is subjected. The following results, due to Dyer, were obtained by boiling the sample with water, washing, dry- ing, and weighing the insoluble residue, and determining the soluble matters by difference. The moisture varied in extreme cases from i to 4 per cent., but the results were calculated as percentage of the dried material: 270 FOOD ANALYSIS Insoluble Ash IN Ether- Nitro- Total Soluble in Water. extract, gen. Ash. Water. Sand. Chicory "nibs" described as "medium roast," 22.40 2.57 1.53 4.63 2.50 0.70 Chicory "nibs" described as "dark roast," 5o-30 2.43 1.67 4.70 2.99 0.30 f 21.50 1.90 1.23 5.33 1.60 0.77 Ground chicory, 9 samples, . j to to to, to to to i 37.80 3.87 1.52 8.23 3.30 3.97 In eight out of the eleven samples the matter insoluble in water ranged from 21.50 to 23.50 per cent. One sample con- tained 35.50, one 37.80, and one 50.30 per cent. Graham, Stenhouse, and Campbell have suggested the tinc- torial power of the infusion as a means of determining adul- terants in coffee. As a rule, the coloring power of chicory is about three times as great as that of coffee. The method may be useful in the detection of added caramel or of added sugar which has been caramelized in roasting. The infusion should be compared with that from pure coffee. The ash of coffee is usually 3.5 to 4.5 per cent., and rarely, if ever, 5 per cent. Of this, about 80 per cent, is soluble in water. It contains mere traces of silica, and is almost in- variably white. A red ash usually indicates adulteration. A notable amount of potassium is present, but sodium may be present in small amount. Analyses by Ludwig indicate that the composition of coffee ash is subject to marked variation according to soil. Chicory contains about 6 per cent, of ash, of which only from 30 to 40 per cent, is soluble in water. It may contain several per cent, of silica and usually carries con- siderable admixed sand. Sodium is always present, often to a considerable extent. The ash of cereals and leguminous seeds is usually less than that of coffee (see page 95). The following table, due to Konig, gives some results ob- tained from the examination of various coffee adulterants : COFFEE 271 Water- extract Calcu- lated NiTRO- Ether- on the GENOUS EX- N-FREE DrY MA- Water. Matter, tract. Sugar. Matter. Fiber. Ash. terial. Chicory, roasted,.. 1 3. 1 6 6.53 2.74 17.89 41.42 12.07 6.19 70.50 Figs, roasted, 1250 4.57 2.96 32.50 31.92 12.34 5.21 82.50 St. John's bread ' — . — ' (carob bean), ... 5.35 8.93 3.65 69.83 10.15 2.09 63.71 Cereals (rye, etc.),.. 1 2. 50 12.15 3-57 4-i2 55-66 8.45 3.55 48.53 Malt, 7.08 13.05 2.25 15.67 51.74 7.38 2.83 65.00 Mogdad cofiFee(Ca5- sia occidentalis),. II. og 15.13 2.55 46.69 21.21 4.33 30.00 "Congo" cofifee, raw, ....13.72 39.82 1.26 37.09 4.41 3.70 "Congo" coffee, roasted, 4.22 27.06 1.19 3,25 39.74 19.28 4.63 22.50 Acorns, shelled and roasted, 12.50 6.78 4.35 69.27 5.02 2.07 28.88 Date stones, 9.27 5.46 8.50 52.86 23.97 1.44 12.87 Fruit of wax palm, raw, 9.37 6.54 10.57 1.67 25.48 44.31 2.06 13.41 Fruit of wax palm, roasted, 3.76 6.99 14.06 1.25 33.25 38.45 2.24 14.03 A number of methods have been proposed for the deter- mination of the caramel in coffee roasted with sugar. A method due to Hilger is as follows: 10 grams of the whole coffee are shaken for half an hour each time with three suc- cessive portions of 100 c.c. of a mixture of equal parts of water and 85 per cent, alcohol. The united solutions are made up to 500 c.c, filtered, the residue dried at 100°, weighed, and the ash determined and deducted. It is necessary to decant the liquid from the berries before filtering, since the extra time considerably increases the relative amount of ash in the extract, due to the more complete extraction of the constituents of the berry itself. Fresenius & Griinhut consider that the best results are had by deducting from the result a mean constant for the materials extracted from the cofifee itself. The following results were obtained. The roasting of the 272 FOOD ANALYSIS coffee without sugar was performed in the normal manner; i. e., the loss on roasting was about 18 per cent. : Soluble Residue (Less Ash). Yellow Java, 0.71 Green " 0.62 Blue " 1.39 Maracaibo, 0.60 Average, 0.83 Percentage of Ash- free Soluble Matter Less 0.83. Yellow Java roasted with 7I per cent, of sugar, 2.21 2.83 2.06 3-46 2.55 4.00 2.78 3-39 9 Green " " 7h 11 (( " 9 Blue " 7i t( .&• .s On 0) (/I o CI. 1/3 >; '5b o a; *u a 0. P T^ t-^ uTi On rj- vO PO t~^ t^ vooo ro 00 Ox o o o o SO X 6 6 6 6 6 d d Q- . t^iO •^ PO On X Tt M O O -^ M VO w rj ^J CN N 1^ m rn < d d d d d d d Q g Q y 1-1 1-1 VO lO IN 00 -I U P ^ lO lOvO On t^oo < 5 ^ <•'■. „ M « - ro M M K U 1-1 N c^ < 6 6 6 S H d d d d d d d W « y X Cfl " J fc < ! ■===' z s OCl r^ ro M r^ N 1-1 DO ■^ t^ « On Tf 00 n' r^. ro rn m' i-J fO •< O 30 w^ r^ N l^ N CS 00 ONOO vO D M d d -; M •-<" ci \n J> Q h <. t^ uD vO ^ 6 6 6 6 6 d d U < OO O rj- « Tt- ui C^ ""d- ro N 00 CO Tt o: ro t-« ro ^^ <-o i^ r^ HH •^ ON -^ (N \d u^ i/> r-^ U-) irjvd 'd vrj uS VO u^r^ U J o I O rOvO Os ro ro CO t^ 'd- VT) U^ Ti- C\ rOsC r^ t-^ r^ I^ ro Tf M >0 :^ Tf ro ro 'tf n" 4 '^^^ \r\ u-i vd liSo < . . . uT (U >%- V, - ., .. rt " - " " 2 tl OS OnXJ u^ ^vO 00 O tn o cxD M f^T^ •IMC) Vm o ^ 1) c a; < i; i :: :: 2 - : < < oT oc^ 2 ^ ]« U <^X "y I- 2^ uT ~ 1 5 ^xx s « 2 aj s ^ f •J2 .si rti k- i- . ^ c2 ":« "^ 1 1^^^^ rt 3 3 MOO ts ft! WINE 345 The addition of preservatives, especially salicylic acid, sodium fluorid, sodium silicofluorid, and of sulfites is very common. Sodium bicarbonate is also added in order to cor- rect acidity. The quantity of chlorids may, at times, be con- siderable, due either to the addition of salt, or to the presence of chlorids in the water used in making the mash. The direct addition of salt is probably infrequent. The following recommendations as to standards of composi- tion of beer were made in 1897 to the Association of Official Agricultural Chemists by the referee on food adulteration: "The glycerol content of beer should not be less than 0.4 gram per 100 c.c. The ash should not be less than 0.12 nor greater than 0.30 gram per 100 c.c. The presence of less than o.io gram indicates that some malt substitute low in ash, such as starch, has been used in the preparation of the beer, while if the ash content be greater than 0.30 gram per 100 c.c, and the volatile acids, calculated to acetic acid, less than 0.075 gram per 100 c.c, it is probable that an excess of acid has been neu- tralized by sodium carbonate, and the ash of the beer should be examined for both sodium and carbonic acid. The phos- phoric oxid should not be less than 0.05 gram nor greater than 0.10 gram per 100 c.c. If less than 0.05 gram, it is probable that a portion of the malt has been replaced by starch or similar substance." WINE Wine has been defined to be the fermented juice of the grape with such additions as are essential to the stability or keeping of the liquid. The method of preparation is, briefly, as follows: The grapes are crushed, the stem being removed in the case of the better grades of wine, and the juice expressed. The juice or ''must" is sometimes allowed to stand in contact with the skins for several days in order to extract additional ''bou- quet." In the case of red wines, the expression of the juice 346 FOOD ANALYSIS and removal of the skins do not take place until the fermentation is practically completed. The juice of most varieties of grapes is colorless, but in the presence of alcohol formed by the fer- mentation the red coloring-matter of the skin is extracted; red wine contains a greater proportion of tannin than white wine. The chief fermentation of the wine usually takes place in from four days to several weeks, according to the temperature at which it is conducted. After this, the liquid is drawn off into casks, where a secondary quiet fermentation takes place. The wine is then allowed to age or ripen, a process which in- volves chiefly direct oxidation, and during which potassium acid tartrate is deposited, along with a considerable proportion of the coloring-matter, and, by the interaction of the alcohols with the acids and other constituents present, various esters are formed which give flavor and bouquet. The yeast that ferments the must is found on grape skins. There are many varieties, some of which produce special flavors, and by the application of these in special cases the flavor of the wine may be modified. Wines prepared as above usually contain very little sugar, and are termed dry wines, as distinguished from "full-bodied" or sweet wines. Some wines are prepared by adding to the must a certain proportion of alcohol, which causes the fermenta- tion to cease before the complete conversion of the sugar is effected. Port and sherry are manufactured in this way. Champagne is usually prepared as follows: The pressed grapes are fermented as rapidly as possible until but little sugar is left. The clarified wine is blended with other wine to bring it to the quality desired, and pure sugar (about 2 per cent.) is added and the liquid placed in strong bottles," which are tightly stoppered and placed horizontally until fermenta- tion is completed, and then with the necks downward, and, as the wine clarifies, the yeast- sediment collects on the stop- per. This is promoted by frequent turning and manipula- WINE 347 tion of the bottle. The bottle is then skilfully uncorked and a small portion of the wine, carrying with it the sediment, removed. The space so emptied is filled by the addition of wine and a certain proportion of so-called liqueur, and the bottle recorked and wired. The operations are performed so quickly that there is but little loss of carbon dioxid. The liqueur consists of a mixture of sugar, wine, and cognac. Cham- pagne is sometimes prepared by adding the liqueur to the fer- mented wine and charging the liquid with carbon dioxid under pressure. The normal constituents of wine are water, alcohol and its homologues, acetic acid, succinic acid, various compound ethers, sugar, gum, pectin, glycerol, tannin, coloring-matters (in red wine), tartaric acid, calcium or potassium tartrates, phosphates, and other mineral matter. The sugar in wine is apt to be chiefly levulose, dextrose being more readily fermentable. The table on page 348 gives the composition of must and wines from various sources expressed in grams per 100 c.c. The data are derived in most cases from the examination of a great many samples. Adulteration. — The fact that the composition of wine varies within notable limits renders it impossible to assign ab- solute standards and allow a margin for the addition of water and other substances without so far changing the composition as to enable the chemist to determine whether a given sample is or is not genuine. Usually it can only be stated that the sample conforms in composition to that of genuine wine. In some cases additions to the wine or must are regarded as legitimate. Thus, it has been found that a certain propor- tion of acid to sugar in the must is best adapted to the pro- duction of good wine; and in cases in which this proportion does not obtain, it is the practice, in some localities, to make such additions as are necessary to bring these constituents within the proper limits. 348 FOOD ANALYSIS 1 •H 00 00 5:» rovO ^00 % 04 M CO fO • " < H d d d d d o -v-' D ° d vO -^ On M M U^ VO VO N VTi 00 t^ 5 rrl 1 N vO t~^ ro M M M T^ N ■^ ON ••= ro d d d d d d d d ^ d d -e g LO N vn "I vn •■'^ vo 00 U-) N l^ 00 M 00 0\ 00 TT 00 m <^ M Tl-vO *-< VO (/I 3 2 r^ NH i-i fo ►-: HH •H Tf M l-l rO d d d d d d d d d d d d d d d u^ N cs t^ 00 Tt ir> ro « N N 00 w v£> "-i -^ M f^ N PO ►- ir^ tnoo ►-» ■+ 0000 q 0000 ^£0 6 6 6 6 d d d d d d 6 6 6 6 . fO vO 0\ >-" On PJ t^ 00 m 00 NO 00 fO N t^ ro 3 d d d 11 fo Tf N ro n ■^ M rj « CO d d M On d d 6 6 6 6 d d d « 6 6 6 6 d d t^ rl- ro -^ On ON li^ tJ- ro - tJ- On vO Tt- w ro 10 ^ M rj- u-i fOOO fOOO N t-^ >-. On rt * rooo fot^ rooo Tf r^ M Tf 04 r^ m m Tj- 1^ d « d d d d d 6 6 6 6 d d d d d d 6 6 6 6 II- td H < 00 vO 00 00 fO Lr> On ro vO w^ ON 00 ro v£) t^ M 10 06 ON On fO 'O M «- rOOO 1-. Tj- \0 ON 00 M 11* fO N ir> N -^ M m' w pi vd ds M 00 4^0 00 d u J X On N t^ ^ ^S.^? 00 "^ M r^ 00 00 U-> Ti- -^ li^ 00 M 2^00 f fo vo 9 On «^ I^30 On 00 8 vO On ON J;^ 06 M oQ M 00 ^ uS r-^ M On 4no 00 d >j f* i-i < id • .s i A i .s d .s d id i d •sd .S d .£ d s s h" B B B B B B B B S B s s s s s s e a •— Y— ' g ^~y-^ ^V— ' ~— Y— ' ^-Y— ' ^,— ' W-Y— ' ^-v— ' v-v— ' ^— V— ' oT ._ s rt u s if c 1^ 1 .2 c i '. a 3 ^ -s 1 3 ■g >. 2 t > v2 3 1 1 Co fa ?) c-^- ^^ the filtrate are mixed with 1.5 c.c. of a saturated solution of 37© FOOD ANALYSIS sodium carbonate and 1.5 c.c. of water, again filtered, and ex- amined in the polarimeter. The reading must be multipHed by 1.2 to compensate for the dilution. For red wines the same amount of sample is taken, and 6 c.c. of lead subacetate solution are used without addition of water. 33 c.c. of the filtrate are treated with 3 c.c. saturated sodium carbonate solution, filtered, and the reading multiplied by 1.2. With sweet wines 100 c.c. are mixed with 2 c.c. of lead subacetate solution and 8 c.c. of water and filtered. 55 c.c. of the filtrate are mixed with 0.5 c.c. of saturated sodium carbonate solution and 4.5 c.c. of water, filtered, and the reading multiplied by 1.2; 33 c.c. of the filtrate, prior to the addition of the sodium carbonate, are mixed with 3 c.c. of hydrochloric acid and the liquid inverted according to the method on page 119. The liquid is cooled quickly, filtered, the reading taken at known temperature, and multiplied by 1.2. 50 c.c. of the sample are freed from alcohol by concen- tration, made up to the original volume with water, mixed with some well-washed beer yeast, and the mass kept at 30° until fermentation is complete, which will usually require from 48 to 72 hours. The liquid is then transferred to a 100 c.c. flask, a few drops of acid mercuric nitrate added (p. 213), then some lead subacetate solution, followed by the saturated sodium carbonate solution. The flask is filled to the mark, the liquid mixed, filtered, and the reading multiplied by 2. The polarimetric data obtained in the above examinations are interpreted according to the following schedule: If the direct examination shows no rotation, the sample may nevertheless contain invert-sugar associated with the dextro- rotatory unfermentable impurities of glucose or with sucrose. If inversion results in a levorotation, sucrose was present. If fermentation results in dextrorotation, it shows that invert - sugar (or some other levorotatory fermentable carbohydrate) and the unfermentable constituents of glucose were present. If the inversion or fermentation produces no change, sucrose, MALT-EXTRACTS 371 unfermentable constituents of glucose, and levorotatory sugars are absent. If the direct examination shows dextrorotation, sucrose and the unfermentable constituents of glucose may be present. If after inversion it is levorotatory, sucrose was present; if dextrorotatory to more than 2.3 divisions of the sugar scale, the unfermentable impurities of glucose were present; if the dextrorotation is less than 2.3 divisions and more than 0.9, a portion of the original specimen must be submitted to the following treatment: 21b c.c. are mixed with i.i gram of potassium acetate and evaporated to a thin sirup, which is mixed with 200 c.c. of 90 per cent, alcohol, with constant stirring, the solution is filtered, the alcohol removed by distil- lation until about 5 c.c. remain, the residue is mixed with washed bone-black, filtered into a graduated cylinder, and washed until the filtrate amounts -to 30 c.c. If this filtrate shows a dextrorotation of more than 1.5 divisions on the sugar scale, the impurities of glucose were present. If the direct examination shows levorotation, and this is increased by inversion, sucrose and levorotatory sugar were present. If the sample after fermentation shows levorotation of 3 divisions, it contains only levorotatory sugars. If after fermentation it rotates to the right, levorotatory sugars and the unfermentable impurities of glucose were present. MALT-EXTRACTS Some commercial malt-extracts are semi-solid mixtures of diastase with products of hydrolysis of starch, such as, maltose, dextrose, and dextrin. No alcohol is present; preservatives and coloring-matters are not likely to be used. Other extracts are dark- colored liquids, containing from 3 to 7 per cent, of alcohol, 5 to 15 per cent, of solids, mostly organic, but little, if any, active diastase. Preservatives are liable to be used in this class, salicylic acid being the most common. 372 FOOD ANALYSIS The usual examination of malt-extracts will involve detec- tion of preservatives, determination of alcohol, solid matter, and diastatic power. Qualitative tests for diastase may be made as follows: 50 c.c. of a solution of 5 grams arrow- root starch in 1000 c.c. of water, made as directed below, are mixed with about i gram of the extract to be tested, and the mixture heated in a water-bath within the limits of 35° and 45°. Every few minutes a drop of the liquid is tested on a porcelain plate with a drop of iodin solution (page 26), until the blue color ceases to appear. It is not worth while to continue the ex- periment beyond a half hour, as a malt-extract that will not transform the starch in that time is of no diastatic value. The solution should not be acid. For quantitative measurement, it is necessary to determine the reducing sugar formed in presence of a large amount of starch. 10 grams of arrowroot starch are stirred into about 100 c.c. of cold water, the mixture added, with constant stirring, to 250 c.c. of boihng water, and the boihng continued until the starch is well diffused through the mass. The solution is diluted to 500 c.c. when cold. 50 c.c. of this solution are mixed with 0.5 gram of the sample and the mix- ture kept at a temperature between 35° and 45° for half an hour. The reducing sugar is measured by the volumetric method described on page 113, care being taken that the liquid is sufficiently diluted. An experiment without addition of starch must be made to determine the amount of reducing substance in the extract. In some cases rough comparative approximations may be made by comparing the color produced by iodin at the end of the heating, but the liquid must be largely diluted, and the indications are merely suggestive. Alcohol and solids are determined as in alcoholic beverages. FLESH-FOODS 373 FLESH-FOODS Descriptions of anatomic and histologic characters of flesh- foods need not be given here. The following table, from data compiled by Allen, will show the principal constituents of some meats. The figures are percentages; they must be regarded as approximations, as the analytic processes are imperfect. The proteid was obtained probably by multiplying the total nitrogen — found by the Kjeldahl method — by 6.25 or approxi- mate factor. Meat from: Water. Ox (lean), 76.7 Ox (fat), 55.4 Mutton, 76.0 Mutton (fat), 48.0 Pig, 72.6 Horse, 74.3 Hare, 74.1 Deer, 75.7 Chicken, 76.2 Pigeon, 75.1 Lobster, 76.6 Oyster, 80.3 Herring, 74.6 Mackerel, 71.2 Salmon, 64.3 Cod, 82.2 Grindley^^ has investigated the action of pure water at a tem- perature not over 10° on raw and cooked beef. Some of his results are given in the annexed table. The nitrogenous com- pounds are in all cases obtained by multiplying the Kjeldahl nitrogen by 6.25. Cold Water Extract of Raw. Boiled. Raw Beef. Boiled Beef. *OTEID. Fat. Ash. 20.7 1-5 1.2 17.1 26.3 I.I I7.I S-7 1-3 14.8 36.4 0.8 19.9 6.2 I.I 21.6 2-5 1.0 23-3 I.I I.I 19.7 1.9 I.I 19.7 1.4 1-3 22.1 I.O 1.0 I9.I I.I I.I I4.I 1-5 2.7 14-5 9.0 1-7 19.4 8.0 1-3 21.6 12.7 1-3 16.2 0.3 1-3 Total proteids, 19.96 37 Coagulable proteids, Albumoses, Peptones, Meat-bases, Acid, (calculated as lactic), . . . Ash, 70 2.18 0.05 0.08 0.12 0.03 O.IO 1.05 0.87 1.09 1. 14 I.I4 0.85 374 FOOD ANALYSIS The higher proteid content of boiled beef was due largely to the lower proportion of water. Grindley found that after extraction of raw meat with pure, cold water, a lo per cent, solution of sodium chlorid will ex- tract much additional matter, largely coagulable proteids. Very little proteid matter is extracted from boiled beef by pure water or sodium chlorid solution. Adulteration. — Meats are not adulterated in the sense in which that word is commonly used, but cheap meats are sub- stituted for dear {e. g., horse meat in sausages and mince- meat), the meat of diseased and immature animals is often sold, preservatives are employed, and appHcations made to improve color or texture. The detection of entozoa is a matter of importance. Tests for incipient and actual decomposition may be required. Analytic Methods. Water. — 5 grams of the finely divided material are dried according to the methods described on pages 27-32, Parson's method being especially worth trial in this connection. Ash. — The dry residue obtained in the water determina- tion is incinerated according to the methods given on pages 39 to 41. Total Nitrogen. — The Kjeldahl- Gunning process is em- ployed. The nitrogen, mulitplied by 6.25, will give an ap- proximation to the proteids present. If nitrates are present, as will be the case with some preserved meats, the modified process, page 37, must be used. Fat. — Much of the fat in meat samples can be removed by mechanical methods, but some adheres obstinately to the muscle-tissue, and it is probable that errors have beeii made in this respect, as with condensed milk. It has been suggested that the muscle-tissue be digested with pepsin and hydro- chloric acid and the fat extracted from the mass. Good re- sults have been claimed for the following process: 2 grams FLESH- FOODS 375 of the material are shaken frequently for six hours with 200 c.c. of ether and 2 c.c. of mercury and the fat determined in an aliquot part of the mixture. Most investigators use too much material. It is probable that results near enough for practical purposes could be ob- tained by continuous extraction for some hours of a few grams of the material, but care should be taken that the sample rep- resents a fair average of the specimen and that it is very finely divided without loss of fat. If the fat is to be examined, a large amount of it should be extracted by mechanical means, and not with solvents, unless there are special reasons to the contrary. Horseflesh. — The detection of horseflesh is difficult. Many processes have been proposed, but they are all open to objec- tion. The principal reliance is upon the detection of glycogen, which is present in horseflesh in much greater proportion than in most other flesh. A brief qualitative method may be used for glycogen (Cour- ley & Coremons^*) : 50 grams of the material are boiled for 30 minutes with water, strained, and a portion of the filtrate mixed with a few drops of potassium iodid-iodin solution (page 26). With a large percentage of horse-meat, the glycogen will produce a dark brown liquid, destroyed by heating and reappearing on cooling. If starch is present, it must be removed, by adding to a portion of the filtrate 2 volumes of glacial acetic acid, again filtering and testing this filtrate as above. The following quantitative method (Pflueger & Nerking^") is provisionally recommended by A. O. A. C. 50 grams of the finely-macerated meat are digested on the water-bath with 200 c.c. of 2 per cent, solution of potassium hydroxid, until solution is practically complete. The liquid is cooled, diluted to 20 c.c. with water, shaken, filtered through a dry filter, and 100 c.c. of the filtrate mixed with 10 grams of 376 FOOD ANALYSIS potassium iodid and i gram of potassium hydroxid, which are stirred in until dissolved. 50 c.c. of alcohol are added and the mixture allowed to stand overnight. The glycogen will separate. It is collected by filtration, washed with a solution containing i c.c. of a 73 per cent, solution of potassium hy- droxid, 10 grams of potassium iodid, 100 c.c. of water and 50 c.c. of alcohol. The material is then washed with a mixture of 2 volumes of alcohol and i of water, containing sodium chlorid in the proportion of 0.007 gram per liter, the residue dissolved in water the remaining proteids removed by solution of potas- sium mercuric iodid. Filter if necessary, add sodium chlorid in the proportion 0.002 gram per 100 c.c, precipitate the glycogen again by the alcohol-sodium chlorid solution noted above, wash with alcohol containing 0.007 gram sodium chlorid per liter, then with absolute alcohol, finally with ether, dry to constant weight and weigh. As control, the glycogen may be hydrolyzed by boiling for 3 hours with hydrochloric acid diluted with 10 parts of water, and the reducing sugar determined as on page 113, multiplying the result by 0.9 for glycogen. Bremer states that the most definite test for horseflesh is the character of the intramuscular fat. For this test, all visible fat is removed from a sample, the mass finely minced, and heated in water for an hour at 100°. The fat that floats is poured off with the water, the flesh washed several times with boiling water, dried for twelve hours at 100°, and the material then extracted for several hours with petroleum spirit of low boiling- point. Part of the fat thus obtained may be set aside for the determination of iodin number, but most of it should be sa- ponified, the excess of alkali carefully neutralized with acetic acid, and any alcohol that may have been used in the saponifi- cation removed by evaporation on the water-bath. The glyc- erol- soda method would seem to be applicable here. The soap is dissolved in water, a hot solution of zinc acetate added FLESH-FOODS 377 in the proportion of i part of the salt to 2 of fat, the precipitate washed with hot water and alcohol, pressed between folds of filter-paper, and heated with ten times its volume of anhydrous ether for thirty minutes under a reflux condenser. The solu- tion is cooled, filtered into a separating funnel, mixed with dilute hydrochloric acid, the ethereal layer, which contains the acids, washed with water, and parts of it filtered into weighed flasks, the ether evaporated, and the iodin number determined. It is stated that horseflesh always gives a reddish-brown tint to the petroleum spirit solution, but bull's flesh also gives such a tint. If, however, glycogen has been detected by the tests already mentioned, the petroleum spirit solution is reddish- brown, the iodin number of the fat exceeds 65 and that of the liquid acids, obtained as above, is considerably over 95, the presence of horseflesh may be inferred. Starch is often added in large amount to sausage, deviled meats and similar articles. It may be detected as noted on page 87, but it must be remembered that it may be used in small amount to facilitate mixture, and may occur in spices, and in some brands of table-salt. A slight reaction should be dis- regarded. The determination of starch cannot be carried out by the standard reduction methods on account of interference of some of the meat-constituents. For approximation the method of Ambiihl, with slight modification, is suggested by A. O. A. C.*" 2 grams of the sample are thoroughly macerated with 100 c.c. of water, then boiled for 30 minutes and the liquid made up to 200 c.c, mixed, filtered, an aliquot portion taken, tested with the potassium iodid-iodin solution (page 26) and the color compared with a solution containing a known amount of the same kind of starch as that in sample. The last point may be determined by microscopic examination. Coloring-matter. — Meats are not infrequently colored to give them a fresh look or to improve naturally pale samples. 378 FOOD ANALYSIS Sausage meats are often colored. Carmine and coal-tar colors, especially the latter, are often employed. Fuchsin and eosin are among these, but Allen states that benzopurpurin is the most common. The detection of artificial colors will generally be acomplished satisfactorily by the test on page 64. E. Spath has found that heating the material for a short time on the water-bath with a 5 per cent, solution of sodium salicylate will often dissolve out colors not otherwise soluble. Ordinarily, water or alcohol will take out sufficient for the wool-test. For the detection of carmine, the method of Klinger and Bujard, modified by Bremer, may be used : 20 grams of the minced mate- rial are heated for several hours with a mixture of equal parts of glycerol and water slightly acidulated with tartacic acid. The yellow liquid is freed from fat, filtered, and the coloring-matter precipitated as a lake by the addition of alum and ammonium hydroxid. This is washed, dissolved in a little tartaric acid, and examined in the spectroscope. Absorption bands lying between the position of b and D of the solar spectrum are characteristic of carmine. Improvers and Preservatives. — Mixtures of potassium ni- trate, sodium chlorid, and other mineral preservatives with a little coloring-matter — the latter almost always a coal-tar color — are sold for improving the appearance of meat. Sulfites are also used as improvers — acid sodium sulfite being a common form — in quantity equivalent to 0.5 to i per cent, calculated as sulfur dioxid. Salicylic acid and borates are also used. As these are all soluble in cold water, they may be extracted by simple maceration, the watery solution being concentrated at a low temperature and treated as directed on pages 78 to 85. Formaldehyde is not likely to be used in meat on account of its hardening action on proteids. Chace ^^ found aluminum oxyacetate (basic acetate) as a preservative in canned sausage in amounts yielding from 1 1.2 to 31.3 of aluminum oxid to 100 grams of material. The FLESH- FOODS 379 qualitative test is made as follows: 25 grams of the material are partially ashed in a platinum vessel, exhausted with hy- drochloric acid, sodium hydroxid added in excess, the liquid boiled, filtered, the filtrate acidified with hydrochloric acid, and ammonium hydroxid added. Aluminum hydroxid and aluminum phosphate are thrown down. Aluminum is not a constituent of normal flesh in appreciable amount. For quantitative methods, the process of Fresenius & Wacken- roder is used. (See page 386.) Putrejaction. — To detect incipient putrefaction, Ebers pro- posed the following test: A rod moistened with a mixture of hydrochloric acid i c.c, alcohol 3 c.c, and ether i c.c. is held over the suspected material. The formation of fumes of am- monium chlorid shows that putrefaction has begun. Care must be taken not to mistake the fumes of the hydrochloric acid for those of ammonium chlorid. Nitrates. — These are generally in the form of added potas- sium nitrate and may be determined by the following method, which so far as the preparation of the sample is concerned is due to Given.^^ The operation must be preceded by a deter- mination of the chlorids present, as these interfere with the process. This can easily be done by titrating in the usual manner a cold water solution of the finely divided meat, i gram in 200 c.c. will be convenient For nitrates, i gram of the sample is placed in a 100 c.c. flask, 50 c.c. of water added, and the mixture kept in hot water for 20 minutes, with occasional shaking. For each i per cent, of sodium chlorid present, 3 c.c. of a saturated solution of silver sulfate are added, then 10 c.c. of lead subacetate and 5 c.c. of alumina-cream, shaking after each addition. The liquid is made up to 100 c.c, shaken, filtered through a plaited, dry, filter, the filtrate being returned until it is clear. 20 c.c. of the fil- trate are evaporated on the water-bath in a shallow porcelain dish to dryness and mixed with i c.c. of the phenoldisulfonic 380 FOOD ANALYSIS acid described below, the acid being stirred over the whole dish with a glass rod so as to touch all parts of the residue. Heat is not needed. The liquid is diluted with water, rinsed into a nesslerizing glass, the dish rinsed several times, these rinsings being added to the first, and then ammonium hydroxid or sodium hydroxid is added to distinct alkaline reaction. The nitrates form picric acid, the alkali forms a picrate; the depth of color of this is proportional to the amount present. The determination is made by comparing the color with that produced by a solution of potassium nitrate of known strength treated in the same manner, that is, evaporation on water-bath, admixture with i c.c. of the phenoldisulfonic acid, and addition of alkali. The phenoldisulphonic acid is prepared as follows: 37 grams of pure sulfuric acid and 3 grams of pure phenol are heated for six hours in a flask immersed in boiling water. The reagent may crystalhze on cooling, but can be easily liquefied by gentle warming. The nitrate solution for comparison may be made by dis- solving o.ioo gram of pure dry potassium nitrate in water to make 100 c.c. i c.c. of this is evaporated in a porcelain dish on the water bath, the residue mixed with i c.c. of phenol- disulfonic acid, stirred, diluted with water and rendered alkaline as noted above. The solution is diluted to the same volume as that of the solution from the meat and the colors compared. The nitrate indicated in the solution of the sample is one- fifth of that present, since 20 c.c. out of the 100 c.c. are taken. The standard nitrate solution is such that i c.c. contains o.ooi gram of potassium nitrate. If the two solutions are of equal tint, 0.005 gram of potassium nitrate was in the sample, i. ^., o. 5 per cent. If the two solutions are very different in depth of color, evapo- ration of a second portion of standard nitrate solution must be made, taking, as far as can be judged, enough, more or less, FLESH- FOODS 38 1 to approximate closely to the other solution. When the depth of color is not widely different in the two solutions, they can be compared by pouring out the deeper solution until, when placing the glasses side by side upon a pure white surface and looking down through the hquids, the tints are sensibly equal. The relative volumes of the liquid will then be a basis for calcula- tion. For example: I gram of sample treated as directed is made up to 50 c.c, which volume contains the picrate equivalent of the nitrate in 0.2 gram of the sample; if, now, i c.c. of standard nitrate also treated and made up to 50 c.c. gives a liquid which is the same depth of color as 25 c.c. of the liquid from the sample, then: 50 c.c. from standard = o.ooi potassium nitrate. 25 c.c. from sample = o.ooi potassium nitrate. 50 c.c. from sample = 0.002 potassium nitrate. 0.002 X 5 = o.oio potassium nitrate = i per cent. Injected Meats. — The lower animals are subject to para- sitic diseases communicable to human beings. The most important are two species of so-called tapeworm and the Trichina spiralis. One species of tapeworm, Tcenia saginata, is found in one stage of development in beef; another species, T. solium^ is found in pork. This condition is often termed "measles." Trichina spiralis is principally found in pork. Many other animal parasites are known, but recognition of them belongs to pathology and biology. Tcenia saginata Goeze, also called T. mediocannellata, occurs in beef as little white cysts among the muscular fibers, Hke knots in wood. The mature animal is developed from the cysts when the meat is eaten. It is the common tapeworm of the United States. Tcenia solium L. occurs in the flesh of the hog. Trichina spiralis Owen is a worm that occurs in hog-flesh as light-colored cysts, smaller than a pin's head, and usually 382 FOOD ANALYSIS lying with the long diameter in the direction of the muscular fiber. The cysts contain immature worms, which are released when the cyst is digested; the worm quickly reaches matur- ity, multipHes rapidly, and distributes itself through various tissues of the host. The detection of the various parasites of meat can often be attained by examining with a good hand-glass. With higher powers, the organism can be seen in more detail. Canned Meats. — These are now usually prepared on a very large scale at estabhshments under inspection and hence are but little liable to adulteration. Preservatives, except common salt and niter, are not likely to be employed. If any other preservative should be used it will probably be boric acid or possibly salicylic acid, either of which can be easily detected in the extract with cold water by methods given elsewhere. Tin and sometimes lead are absorbed in small amounts from the can or solder. These may be tested for by the methods given on page 58. Examination under moderate magnifying power will detect parasitic infection. (See pages 378 and 386.) Meat-extracts. — These are now offered in great variety. Some contain partly digested proteids (proteoses and peptones), but in many samples the most abundant nitrogenous ingredients are the so-called meat-has es^ a class of amido-derivatives of which kreatin, kreatinin, and xanthin are examples. Many pro- prietary articles, intended especially for invalid feeding, con- tain much alcohol and carbohydrates (maltose, lactose, dex- trine). Some contain notable amounts of iron and manganese. Many investigations of these preparations have been made, but the processes of analysis are still in dispute and the results obtained by different observers do not agree. The following methods are compiled from the work of Allen, Mitchell and Grindley. Water ^ Ash, and Total Nitrogen are determined as indicated under those titles in the introductory part. FLESH-FOODS 383 Fat is usually present in but small amount, and is extracted more accurately by petroleum spirit or carbon tetrachlorid than by ether, applying the methods described on pages 41 to 43. Insoluble matter, which may include some meat-fiber, is de- termined by treating from 5 to 25 grams (depending on whether the preparation is sohd or liquid) with cold water, filtering, and drying the residue at 100°. A microscopic examination of this should be made. Proteids, Peptones, and Meat-bases. The following method has been suggested by Allen, *^ partly from his own experiments and partly from those of Bomer: 50 c.c. of a solution of a known weight of the sample, of such strength as to contain about 1.5 grams of nitrogenous bodies, are freed from insoluble material, mixed with i c.c. of diluted sulfuric acid (i to 4), and saturated with zinc sulfate by stirring in the powdered salt until no more dissolves. Zinc sulfate containing the full amount of water of crystallization disssolves in about half its weight of water at room tempera- ture, but the commercial salt is usually partly effloresced, and will often cake when added to the solution. When the liquid is saturated with zinc sulfate, the precipitate is assumed to contain all the albumin and gelatin and immediate derivatives (proteoses), but no peptone. It is separated by filtration, washed with a saturated solution of zinc sulfate, and the filter and precipitate treated by the Kjeldahl- Gunning method. The nitrogen obtained, multiplied by 6.25, will give approximately the amount of nitrogenous bodies precipitated. The filtrate and washings are made up to 200 c.c, mixed, and 100 c.c. transferred to a flask of the larger form described on page 33, enough hydrochloric acid added to make the liquid strongly acid to litmus, and then bromin water by moderate portions, with active shaking or stirring, until there is an excess of bromin present. The precipitate may be flocculent at first, but most of it soon becomes viscous and adherent. It 384 FOOD ANALYSIS is allowed to stand until the free portions have settled, when it is decanted through an asbestos filter either in a Gooch cru- cible or in an apparatus similar to that described on page 115. The precipitate is washed several times with cold water con- taining some hydrochloric acid and bromin, but it is advisable to keep the washings at first separate from the main filtrate. The contents of the filter-tube are returned to the vessel in which the precipitation was made, 10 c.c. of sulfuric acid added, and the mass cautiously treated until it chars and vapors of bromin are evolved, after which 10 grams of potassium sul- fate are added and the operation conducted as described on pages 33 to 37. The nitrogen, multiplied by 6.33, will give ap- proximately the peptone. The process of A. O. A. C. suggests liquid bromin (2 c.c.) instead of bromin water. By deducting from the total nitrogen the sum of the nitro- gen figures obtained from the zinc sulfate and bromin precipi- tates, and multiplying the remainder by 3.12, an approxima- tion to the meat-bases will be obtained. These meat-bases are in the filtrate from the bromin precipitate, but the bromin, hydrochloric acid, and zinc sulfate will be likely to interfere with the determination of the nitrogen. The zinc sulfate, can be removed by cautious addition of either potassium carbon- ate or barium hydroxid, but the bromin will be apt to form hypobromites, which will decompose some of the meat bases with evolution of nitrogen. A more satisfactory plan seems to be that outlined by Bau- mann and Bomer: The remaining portion, 100 c.c, from the zinc sulfate precipitate is mixed with excess of sodium phos- phomolybdate (see page 274), by which the meat^bases, pep- tones, and ammonium compounds are precipitated. This precipitate is removed by filtration under pressure, so as to draw out as much as possible of the mother liquor, and the nitrogen determined as usual. The nitrogen due to peptone being FLESH- FOODS 385 known, that due to meat-bases and ammonium compounds can be calculated. To determine the ammonium compounds, a known weight of the original sample should be distilled with barium carbonate, the distillate being collected in a known quantity of standard acid, which is afterward titrated. Meat-extracts may contain coagulable proteids. These may be estimated by rendering the filtrate solution distinctly acid with acetic acid and boiling for five minutes. The coagulum may be weighed directly or the nitrogen in it estimated by the Kjeldahl-Gunning method and multiplied by 6.25 for proteid. As solutions of proteids, proteoses, and peptones are strongly levorotatory, while most of the meat-bases that occur in these extracts are inactive, some information might be gained by con- centrating the liquid from the zinc sulfate precipitate and ex- amining it in the polarimeter, filtering if necessary. A solution that has no appreciable optic activity will not be likely to contain much peptone. Another special test that may be ap- plied to this liquid is the so-called biuret reaction. Bomer ap- plies this as follows: The filtrate from the zinc sulfate pre- cipitation is decolorized by shaking with animal charcoal and the zinc sulfate decomposed by excess of sodium carbonate or cautious addition of barium hydroxid. The filtered solution is rendered alkaline with sodium hydroxid and a drop or two of very dilute solution of copper sulfate added. Peptones give a rose-red tint. Preservatives may be added to meat-extracts, although this is not usual. Boric acid will be most likely to be used, and the methods on page 367 will suffice for its detection. Poi- sonous metals are not likely to be present, but may be sought for, if deemed necessary, by the methods given on pages 57 to 64. Some preparations may require examinations for iron and manganese. These will be obtained in solution by heating the ash in strong hydrochloric acid, and may be separated and determined by the standard methods of mineral analysis. 34 386 FOOD ANALYSIS Addendum to page 378. — Fresenius & Wackenroder's process for the determination of aluminum, as described by Chace'^: A weighed amount of the finely comminuted sausage is heated over a low flame until danger of spurting is past. (The low-temperature burner, page 52, figure 31, will be sat- isfactory.) The mass is then heated until thoroughly charred, cooled and digested for some time on the water-bath with hydrochloric acid, filtered, slightly washed, and the filter and residue ashed. This ash should be gray and small in amount; it is dissolved in hydrochloric acid, the solution filtered and the filtrate added to the other solution. Any appreciable residue on the filter should be tested for aluminum. The combined filtrates are made slightly alkaline by ammonium hydroxid, and barium chlorid added until no further precipi- tate is formed. This consists of barium phosphate, aluminum hydroxid and aluminum phosphate. It is washed, and dis- solved in the least possible amount of hydrochloric acid. This solution is saturated with barium carbonate. Potassium hydroxid is added in excess and the mass digested for some time; then sodium carbonate is added, the barium carbonate and phosphate separated by filtration and thoroughly washed. The filtrate is acidulated with hydrochloric acid, and the aluminum determined in the usual way. SPECIFIC GRAVITY OF WATER. 387 SPECIFIC GRAVITY OF Water at 0° = 0.99987 WATER FROM Water at 4° - 0° TO icxj° = 1. 00000 I 0.99992 26 0.99686 51 0.9S772 76 0.97438 2 96 27 60 52 25 77 0.97377 3 99 28 33 53 0.98677 78 16 4 1 .00000 29 05 54 29 79 0.97255 5 0.99999 30 0.99576 55 0.98581 80 0.97194 6 97 31 77 56 34 81 32 7 93 32 47 57 0.98486 82 0.97070 8 88 33 0.99485 58 37 83 07 9 82 34 ■ 52 59 0.98388 84 0.96943 ID 74 35 18 60 38 85 0.96879 II 65 36 0.99383 61 0.98286 86 15 12 54 37 47 62 34 87 0.96751 13 43 38 10 63 0.98182 88 0.96687 14 29 39 0.99273 64 28 89 22 15 16 40 35 65 0.98074 90 0.96556 16 00 41 0.99197 66 19 91 0.96490 17 0.99884 42 58 67 0.97964 92 23 18 65 43 18 68 08 93 0.96356 19 46 44 0.99078 69 0.97851 94 0.96288 20 25 45 37 70 0.97794 95 19 21 04 46 0.98996 71 36 96 0.96149 22 0.99782 47 54 72 0.97677 97 0.96079 23 60 48 10 73 18 98 08 24 36 49 0.98865 74 0.97558 90 0.95937 25 12 50 19 75 0.97498 100 0.95866 388 FOOD ANALYSIS Correspondence of Centigrade and Fahrenheit Degrees 1 2 3 4 5 6 7 8 9 20 392.0 393-8 395-6 397.4 399-2 401.0 402.8 404.6 406.4 408.2 19 374-0 375-8 377-6 379-4 381.2 3830 384.8 386.6 388.4 390.2 18 356.0 357.8 359.6 361.4 363-2 3650 366.8 368.6 370.4 372.2 17 338.0 339-8 341.6 343-4 345-2 347.0 348.8 350.6 352.4 354-2 16 320.0 321.8 323.6 325-4 327-2 329.0 330-8 332.6 334.4 336.2 15 302.0 303.8 305.6 307-4 309.2 311. 312.8 314.6 316.4 318.2 14 284.0 285.8 287.6 289.4 291.2 293-0 294.8 296.6 298.4 303.2 13 266.0 267.8 269.6 271.4 273-2 275.0 276.8 278.6 280.4 282.2 12 248.0 249.8 257.6 253-4 255-2 257.0 258.8 260.6 262.4 264.2 II 230.0 231.8 233-6 235.4 237-2 239.0 240.8 242.6 244-4 246.2 10 212.0 213.8 215.6 217-4 219.2 221.0 222.8 224.6 226.4 228.2 9 194.0 195-8 197.6 199.4 201.2 203.0 204.8 206.6 208.4 210.2 8 176.0 177.8 179.6 181.4 183.2 185.0 186.8 188.6 190.4 192.2 7 158.0 159.8 161. 6 163.4 165.2 167.0 168.8 170.6 172.4 174.2 6 140.0 141.8 143-6 H5-4 147.2 149.0 150.8 152.6 154-4 156.2 5 122.0 123.8 125.6 127.4 129.2 131.0 132.8 134.6 136.4 138.2 4 104.0 105.8 107.6 109.4 III. 2 113.0 114.8 116.6 118.4 120.2 3 86.0 87.8 89.6 91.4 93-2 95-0 96.8 98.6 100.4 102.2 2 68.0 69.8 716 73.4 75-2 77-0 78.8 80.6 82.4 84.2 I 50.0 51.8 53-6 55-4 57.2 59.0 60.8 62.6 64.4 66.2 32.0 33.8 35.6 37-4 39-2 41.0 42.8 44.6 46.4 48.2 15.55° C. = 60° F. -I -2 -3 -4 -5 -6 -7 -8 -9 32.0 30.2 28.4 26.6 24.8 23.0 21.2 19.4 17,6 15.8 -I 14.0 12.2 10.4 8.6 6.8 50 3-2 1-4 -0.4 -2.2 -2 -4.0 -5-8 -7.6 -9.4 -ir.2 -13.0 -14.8 -16.6 -18.4 -20.2 -3 -22.0 -23.8 -25.6 -27.4 -29.2 -31.0 -32.8 -34.6 -36.4 -38.2 -40° C. = -40° F. REFERENCES [" Bulletin " refers to the publications of the Div. of Chem., U. S. Dept. of Agric] > J. A. C. S., 1905, 25. 2 Bulletin 65. 3 J.A. C.S., 1905, 141. * Advance sheets, Amer. Jour. Pharm. ^ J. A. C. S., 1903, 1028. ' Abst. Analyst, 1900, 292. ' Unpublished; to appear in Chem. Zeit. * Tollens, Handb. d. Kohlenh., 2, 207. ' Abst. Analyst, 1904, 306. •" Z. Anal. C, 1905. " J. A. C. S., 1904, 186. '2 J. A. C.S., 1904, 1631. '3 Bulletin 65, also 32d Ann. Rep. Mass. St. B. of H. (1900), 658. '* Private communication to authors. IS J.A. C.S., 1904, 1523. »" Bulletin 65. *^ Ding. Polyt. Jour., 253 (1884), 281. ^« Bulletin 77. »» Z. Anal. C, 1877, 145. 20 Z. Anal. C, 1879, 69. 2' Ding. Polyt. Jour., 233 (1879), 229. " Analyst, 1891, 153. " Zeit. Anal. C, 1879, 199. 2* Compt. rend., 35 (1851), 573. " J. S. C. I., 1891, 233. ^' Analyst, 1895, 147. " J. A. C. S., 1896, 378. 28 T ^^ Q^ g^ " J. C "S. I., 1886, 494- 3« Zeit. Anal. C, 1877, 145- ^' Chem. Anal. Oils, Fats and Waxes, 165. 32 J. A. C. S., 1895, 935. 33 J. A. C. S., 1903, 251,498- 3* Chem. Anal. Oils, Fats and Waxes. 35 J. A. C.S., 1900,453; 1901, I. 3' Chem. Anal. Oils, Fats and Waxes, 574. 3' Much misrepresentation has been made of this matter. Several American chemists have ignored our claims to the devising of the process. The Gerber method is merely a modification of it. This fact is known to chemists of the Department of Agriculture at Washington, yet in the "Provisional Methods of Food Analysis," the Gerber method is men- tioned as an alternative, as if it were entirely original with Gerber. 389 390 REFERENCES '' J. A. C. S., 1899, 503. 39 J.A. C. S., 1904, 1195. *° J. A. C. S., 1904, 1195. *^ Russky Vratch. Abst. Jour. Am. Med. Ass'n., 44 (1905), 1235. " J. A. C. S., 1900, 207. *^ Stokes & Bodmer suggested 10 minutes' boiling, but Watts & Tempany (Analyst, 1905, 119) show that at least 30 minutes' boiling is necessary. ** Bulletin 65. « J.A. C. S., 1905,270. *^ Mikroscopie der Nahrungs- und Genussmittel. *'' Food Adulteration and its Detection. *^ Bulletin 13. *^ Rep. State Board of Health of Mass., 1902, 485. ^° Winton, Bulletin 65. ^' Rep. State Board of Health of Mass., 1903. ^2 J. A. C. S., 1899, 721. ^^ Brooks, Rep. Lab. Hyg. N. J., 1903. '* J. A. C. S. 1899, 257. ^^ J. A. C. S., 1902, 1129. ^^ Food Inspection and Analysis, in place. " Bulletin 65. ^^ The reference {Chemist &* Druggist, 57, 732) directs the use of "25 per cent, sulfuric acid." It is assumed that proportion by weight is meant. 59 Rep. State Board of Health of Mass. «° J- A. C.S., 1901,349. «i J.A. C.S., 1905,613. *^ Leach, Food Inspection and Analysis, 261. «3 Bulletin 65. «* Bulletin 64. «5 J.A.C.S., 1905, 137. «« J. A. C. S., 1905, 138. " J.A. C.S., 1903,16. °^ Analyst, 1904, 301. 89 Edition of 1890. '» J. A. C. S., 1904, 1627. '^ Analyst, 1905, 124. ^2 J. A. C.S., 1900,810. ^3 Allen's Com'l Org. Anal., i, in place. '* Allen's Com'l Org. Anal., i, in place. '^ Amer. Chem, J., 1899, 266. '8 Analyst, 1904, 301. " J. A. C. S., 1904, 1086. " Bulletin 65. 79 Bulletin 65. 8° Bulletin 65. ^^ Bulletin 65, Some errors in the Bulletin description have been corrected here. ^^ Com'l Org. Analysis, vol. 4. *3 J. A. C. S., 1904, 662; also Fresenius' Quantitative Anal., Amer.-Ed., 1904. INDEX Abrastol, 77, 86, 220 Acetyl number, 156 value, 156 Acidity, total, 359 Acid mercuric iodid, 229 nitrate, 2 1 value, 146 Acorn starch, 90 Acrinyl isothiocyanate, 317 Adams' method, 201 Agar, 334 Albumin, 190, 208 Albuminoid nitrogen, 37 Alcohol, detection, 352 determination, 353 ethyl, 54 methyl, 54 detection, 365 — tables, 354-5-6 Alcoholic beverages, 337 Ale, 343, 344 AUihn's method, 117 Allspice, 311 Allyl isothiocyanate, 317 Almen's reagent, 208 Alum in bread, 102 in flour, 97 Alumina-cream, 118 Aluminum acetate, 378 detection, 378 determination, 386 Ammonium in baking powders. Amphoteric milk, 192 Annatto, detection, 215, 217 Antisepticum, 79 Apparatus, 49 Arachidic acid, 175 Arachidin, 174 Arachis oil, 174 Arnold's method, 36 Arrow-root starch, 89, 92 Arsenic, detection, 60-62, 220 Asaprol, 86, 87, 220 109 Ash, 39 Azolitmin, 55 Babcock's method, 200 Baking powders, 107 soda, 105 Banana starch, 89 Barley, 97, 100 starch, 90, 92 Bases, meat-, 382 Baudguin's test, 167 Bean starch, 90, 92 Bechi's test, 166 Beef fat, 188 stearin, 186 Beer, 343 root, 343 Benzoates, 76, 81 Benzoic acid, 76, 81 Birotation, 213 Bitters in beer, 366 Biuret reaction, 385 Bjorkland's test, 180 Boiled milk, detection, 220 Boiling-point, 12 Borax, 78, 82, 239 Boric acid, 78, 82 Borofluorids, 78, 82 Brandy, 341 Bread, loi commercial, 102 Bromas, 277 Bromin, thermal value, 149 Brulle's test, 168 Buckwheat, 97, 100 starch, 91, 92 Bumping, prevention, 44, 45 Burners, 51, 52 Butter, 230 cacao-, 179 colors, 237 composition, 230 391 392 INDEX Butter fat, 189 milk, 193 peanut, 174 vegetable, 179 Butyrorefractometer, 154 Cacao, 273 butter, 179 essence, 277 husks, 277 masse, 277 red, 275 starch, 90 Caffearin, 262 Caffein, 253, 257 determination, 257 Caffeol, 262 Caffetannic acid, 267, 292 Calculation methods for milk, 204 Candies, 135 Cane-sugar, 121 Canna starch, 89 Caper tea, 256 Caramel, 124, 362 Caryophyllin, 315 Casein, 190, 208 Cassia, 313 oil, 314 Catsup, S33 Centrifuge, 50, 203 Cereals, 95 starches, 91 Champagne, 346 Cheese, 240 Chemicals, 49 Chicory, 266 Ching suey, 256 Chocolate, 273 nuts, 276 Cholesterol, 160 Chromium, detection, 58 Cider, 337 vinegar, 284 Cinnamon, 312 oil, 314 starch, 91 Citric acid, determination, in milk, 191 Clove oil, 315 Cloves, 315 Cobalt nitrate test, no Cochineal, 56, 72 Cochran's method, 224 Cocoa, 277 Cocoas, soluble, 277 Coconut oil, 178 olein, 179 stearin, 1 79 Coffee, 262 essence, 272 extracts, 272 Colors, 64-75 in butter, 237 in candies, 136 in meat, 377 in milk, 215 in wine, 361 test for oils, 137 Colostrum, 195 Colza oil, 178 Condensed milk, 222 Condensers, 45-8 Condiments, 282 Confections, 135 Congou paste, 256 tea, 256 Constants for oils, 164, 165 Copper, detection, 59 hydro xid mixture, 37 in bread, 104 in flour, 98, 104 Coriander seed, 301 Corn, Dhoura, 301 meal, 97, 100 oil, 172 starch, 91, 92 Cottonseed oil, 171 stearin, 172 Cream, 193 evaporated, 222 of tartar, 105 Cribb's condenser, 45 Crude fiber, 38 Cryoscopy of milk, 214 Cumarin, 324 Dalican's titer test, 11 Desserts, 335 Dextrin in honey, 131 in wine, 359 Dextrose, determination, 113 Dhoura corn, 301 starch, 90 Distillation, 44 Doughing test, 96 Drying of oils, 154 ovens, 28, 30 INDEX 393 Drying property, 154 Dry wine, 346 Egg colors, 72 detection, 335 Elaidin test, 152 Electrolytic methods, 1 1 6 Ergot, 98 Erucin, 178 Essence of cacao, 277 of coffee, 292 Ether purification, 54 Eugenic acid, 315 Eugenol, 315 Evaporated cream, 222 Extract, 27 Extraction apparatus, 41 Facing coffee, 264 tea, 258 Fat of milk, 190, 200 Fats, 137 Fehling's solution, 113 Fermented milk, 248 Fiber, crude, 38 Inlter-tubes, 114, 115 Fixed solids, 27 Flesh-foods, 373 Flour, 93, 97 Fluorescence, 23 Fluorids, 78, 82 Foreign leaves in tea, 260 Formaldehyde, 77, 83, 218 Formalin, 77 Fractional distillation, 49 Furfural test, 167 Fusel oil, determination, 363 GaLACTOSAZONE, III Gallisin, 125 Gelatin, detection, 217 Gin, 341 Ginger, 306 starch, 89 Gingli oil, 177 Gliadin, 95 Globulin, 95 Glucosazone, in Glucose, 125 Glutenin, 95 Gluten test, 96 Glycerol in wine, 350 Glycerol soda, 143 Glycogen, 375 Graham flour, 97 Grape-juice vinegar, 283 Grape-sugar, 125 Gum in wine, 359 Gutzeit's test, 61 Gypsum in bread, 104 Hager's test, 352 Halphen's test, 166 Hanus' reagent, 140 Hehner value, 155 Honey, 130 Horseflesh, detection, 375 Hiibl's reagent, 139 Hydrometers, 6 Hydronaphthol, 77 Ice-cream, 335 Immiscible solvents, 43, 55 Improvers, meat, 378 Index of refraction, 153 Indicators, 55 Indigo, detection, 258 Insoluble acids, 155 Inversion methods, 119, 227 Inverted condenser, 48 Invert-sugar, 119, 227 lodin number, 139 value, 139 Iron, detection, 58 Jams, 327 Jellies, 327 Kefyr, 249 Kjeldahl-Arnold method, 36 Gunning method, 32 Kottstorfer number, 145 Kumiss, 248 LACTOSAZONE, III Lactose, no, in, 191, 210 Lager beer, 343, 344 Lard, 180 Laurent polarimeter, 13 Laureol, 179 Laurin, 178 Lead, detection, 58 subacetate, 118 394 INDEX Leavening materials, 105 Leffmann-Beam method, 203 Leguminous flours, 99 Lemon extract, 324 juice, 330 — sirup, 330 Lentil starch, 90 Lieben's test, 352 Lie tea, 256 Lignoceric acid, 175 Litmus, 55 Livache's test, 154 Long pepper, 301 Low wine, 283 Mace, 308 false, 309 Maize, 97, 100, 104 oil, 172 starch, 91, 92 Malic acid, 290 value, 129 Malt extract, 93, 371 liquors, 342 vinegar, 285 Maltosazone, in Maple sugar, 127 sirup, 127 Maranta starch, 89 Marsh's test, 62 Maumene's test, 148 Mead, 343 Meal, 93 Meat bases, 382 extracts, 382 Meats, 373 Melting-point, 7 Mercuric iodid, acid, 229 nitrate, acid, 211 Metals, poisonous, 57 Methyl alcohol, 54 detection, 365 orange, 56 Microscope, 23 Milk, 190 ash, 191, 200 • — boiled, 193, 220 condensed, 222 enzyms, 191 — ■ serum, 193 • sugar, 191 Miscible solvents, 41 Mixed flours, 99 Mohr's centimeters, 20 Molasses, 123 Mother cloves, 315 starch, 89 Must, 348 Mustard, 317 oil, 317 Myristic acid, 308 Myristicol, 308 Myronic acid, 317 My rosin, 317 Naphthol, 77, 85 Nickel, detection, 58 Nitrogen, albuminoid, 37 total, 32 Normal weight, 20 Nucoline, 179 Nutmeg, 307 starch, 90 oil, 307 Nutshells, 300 Oats, 97, 99 Oat starch, 91, 92 Oil, arachis, 174 cassia, 314 cinnamon, 314 cloves, 315 coconut, 178 colza, 178 corn, 172 cottonseed, 171 gingli, 177 maize, 172 mustard, 317 nutmeg, 307 olive, 168 peanut, 174 pepper, 293 rape, 178 sesame, 177 teel, 177 Oleomargarin, 232 Oleorefractometer, 153 Olive oil, 168 stones, 298 Original solids, 287 Ovens, 28, 30 Paraffin in oleomargarin, 240 Peanut butter, 174 oil, 174 INDEX 395 Pea starch, 90 Penumbral polarimeters, 13 Pepper, 293 cayenne, 304 long, 302 starch, 91 Pepperette, 298 Peptones, determination, 383 Perry, 337 Petroleum spirit, 55 Phenol, 85 Phenolphthalein, 56 Phenylhydrazin test, 1 1 1 Phytosterol, 161 Pintus, A. S., 86 Piperidin, 293 Piperin, 293 Plastering of wine, 350 Platinum, care of, 53 Poisonous metals, 57 Poivrette, 298 Polarimetry, 13 Porter, 344 Potato flour, 99 starch, 89, 91, 92 Preservaline, 78 Preservatives, 76, 378 Process butter, 236 Proteids, determination, 205 Proteoses, determination, 383 Prune juice, detection, 362 Prussian blue, detection, 258 Putrefaction, detection, 378 Pyknometer, 2 QuERCiTANNic acid, 312 Rape oil, 178 Reagents, 25, 52 Recknagel's phenomenon, 192 Refraction index, 153 Refractometer, 153 Reichert-Meissl number, 143 Reichert number, 143 Reinsch's test, 60 Renovated butter, 236 Rex magnus, 78 Rice starch, 91, 92 Ritthausen method, 207 Root beer, 343 Rum, 341 Rye flour, 96, 99 starch, 90, 92 Saccharin, 77, 80, 361 Sago starch, 90 Salicylic acid, 76, 80, 362 Salol, 85 Saponification equivalent, 146 value, 145 Sausage, adulteration of, 378 Sawdust in flour, 100 Scales for polarimeter, 20 Scheibler's method, 21 Schmidt and Hiinsch polarimeter, 15 scale, 20 Separated milk, 193 Sesame oil, 177 Silicofluorids, 78, 82 Sirup, 123 Sitosterol, 161 Sodium benzoate, 76 phosphomolybdatc, 274 Solidifying-points, 7 Solids, original, 287 Soluble acids, 155 cocoas, 277 Solvents, immiscible, 43, 55 miscible, 41 Soxhlet's method, 210 Specific gravity, i bottle, 2 rotatory power, 19 temperature reaction, 148 Spectroscope, 21 Spices, 291 Spirits, 338 Sprengel tube, 3 Standard acid, 56 Stannous chlorid in bread, 105 in sugar, 122 Starch, 87 detection, 87 determination, 93 indicator, 56 Starches, characters of, 89, 90, 91 I Stearin, beef, 186 coconut, 179 cottonseed, 172 Stout, 343 Stutzer's method, 37, 246 Sublimation, 44, 49 Sucrose, 121 Sugar, cane-, no Sugars, no Sugar scale, 20 Sulfites, 78, 359 Sulfur chlorid test, 186 Sulfuric acid in vinegar, 289 396 INDEX Sulfurous acid determination, 360 Szombathy's tube, 45 Table accessories, ^;^^ Taenia, forms of, 381 Tallow, 187 Tannin, determination, 292 Tapeworm, 381 Tapioca starch, 90 Tartaric acid, 332 Tea, 252 Teel oil, 177 Terra alba in bread, 104 Thein, 253 Theobromin, 273 Thermal reactions, 148-150 Tin, detection, 58, 59 in bread, 105 Titer-test in sugar, 122 Tocher's test, 168 Treacle, 123 Trichina, 381 Turmeric, 310 starch, 89 Ultramarine blue, 122 Unsaponifiable matter, 159 Valenta's test, 147 Vanilla extract, 320 Vanillin, 323 Vegetable butter, 179 Vegetaline, 1 79 Vinegar, 282 cider, 284 malt, 285 spirit, 285 wine, 283, 284 Viscosity, 158 Volatile acids, 142 Water determination, 27 specific gravity of, 386 Weissbier, 343 Werner-Schmid method, 202 Weston distillation apparatus, 45 Westphal balance, 4 Wheat, 96, 99 starch, 90 Whey, 193 Whiskey, 339 Wild's scale, 21 Wine, 345 low, 283 vinegar, 283 Wool test, 64 Zinc, detection, 58, 59 If this leaf is sent to the publishers six months or more after the publication of this book, under sealed cover and enclosing two cents in U. S. stamps (or the equivalent in un- cancelled stamps of any other nation) a slip will be returned containing memoranda of errors that have been noted or re- ported. The preparation of this list will be begun about October, 1905. This lea} must in all cases accompany the request, with name and address oj sender. The authors will be pleased to receive memoranda of errors noted by users of the book. Address (postage full-paid) P. BLAKISTON'S SON & CO., 1012 WALNUT ST., PHILADELPHIA, PA., U. S. A. Name, Address, TTNIV TtSTTY OF CAttr^oRNJ4 LTP-»- 7X^ v.;j/ iC 93639 j^^-ffni «)»r