MAIN r:r •/ OF TH- v/l "I OF / ' •' M»"l 22, Hi-vised iM'liruary L's, 1'JIO. Hi-print January 24, 1912. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF CHEMISTRY— Bulletin No. 109, Revised. H. W. WILEY, Chief of Bureau SOME TECHNICAL METHODS OF TESTING MISCELLANEOUS SUPPLIES, INCLUDING PAINTS AND PAINT MATERIALS, INKS, LUBRICATING OILS, SOAPS, ETC. BY PERCY H. WALKER, CHIEF, CONTRACTS LABORATORY. WASHINGTON; GOVERNMENT PRINTING OFFICE. 1912. ORGANIZATION OF BUREAU OF CHEMISTRY. H. W. WILEY, Chemist and Chief of Bureau. F. L. DUNLAP, Associate Chemist. W. D. BIGELOW, Assistant Chief of Bureau. F. B. LINTON, Chief Clerk. A. L. PIERCE, Editorial Clerk. M. W. TAYLOR, Librarian. Division of Foods, W. D. BIGELOW, Chief. Food Inspection Laboratory, L. M. TOLMAN, Chief. Food Technology Laboratory, E. M. CHACE, Chief, and Assistant Chi Division. Oil, Fat, and Wax Laboratory, H. S. BAILEY, Chief. Division of Drugs, L. F. KEBLER, Chief. Drug Inspection Laboratory, G. W. HOOVER, Chief. Synthetic Products Laboratory, W. O. EMERY, Chief. Essential Oils Laboratory, under Chief of Division. Pharmacological Laboratory, WM. SALANT, Acting. Chief Food and Drug Inspector, W. G. CAMPBELL. Miscellaneous Division, J. K. HAYWOOD, Chief. Water Laboratory, W. W. SKINNER, Chief. Cattle-Food and Grain Laboratory, G. L. BID WELL, Acting. Insecticide and Fungicide Laboratory, C. C. MCDONNELL, Chief. Trade Wastes Laboratory, under Chief of Division. Contracts Laboratory, P. H. WALKER, Chief. Dairy Laboratory, G. E. PATRICK, Chief. Food Research Laboratory, M. E. PENNINGTON, Chief. Leather and Paper Laboratory, F. P. YEITCH, Chief. Microchemical Laboratory, B. J. HOWARD, Chief. Sugar Laboratory, A. H. BRYAN, Chief. Sections: Animal Physiological Chemistry, F. C. WEBER, in Charge. Bacteriological Chemistry, G. W. STILES, in Charge. Enological Chemistry, W. B. ALWOOD, in Charge. Nitrogen, T. C. TRESCOT, in Charge. Vegetable Physiological Chemistry, J. A. LE CLERC, in Charge. Food and Drug Inspection Laboratories: Boston, B. H. SMITH, Chief. Buffalo, W. L. DUBOIS, Acting. Chicago, A. L. WINTON, Chief. Cincinnati, B. R. HART, Acting. Denver, A. E. LEACH, Chief. Detroit. II. L. Sciiri.x. .\etiug. - Galvcston, T. F. PAIM-I:. .\efiiifi. Honolulu, Hawaiian Islands, H. A. DUNCAN, Acting. Kansas City. Mo.. A. V. II. MORY, Chief. Nashville, It. W. BALCOM, Acting. New Orleans, C. W. HARRISON, C h icf. Now York, It. K. DOOLITTLK. Chief. Omalia, S. II. Ross. Acthitf. Philadelphia. ('. S. P.KINTOX. Chief. Pittshnr^r. M. r. AI.ISKKCH. Chief. Portland, Ore::.. A. L. KMSII.V, Chief. St. I. oil is. I >. I',. I'.ISMKI . .1 '•/ ///.'/. St. Paul, A. S. Mm IIKI.T.. chief. San 1'Yancisro. K. A. Gnri.n. chief. Savannah, W. < '. IU KM T. \<-ting. Seattle, H. M. LOOMIS, Acting. Issued April •_"_'. IMS 'd Fcbnuiry 28, 1910. Reprint January 24, I'M 2. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF CHEMISTRY— Bulletin No. 109, Revised. H. W. WILEY, Chief of Bureau. SOME TECHNICAL METHODS OF TESTING MISCELLANEOUS SUPPLIES, INCLUDING PAINTS AND PAINT MATERIALS, INKS, LUBRICATING OILS, SOAPS, ETC. BY PERCY H. WALKER, CHIEF, CONTRACTS LABORATORY. WASHINGTON: GOVERNMENT PRINTING OF I 1912. LETTER OF TRANSMITTAL U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF CHEMISTRY, Washington, D. C., November 18, 1909. SIR : I have the honor to submit for your approval a revision of the bulletin issued in April, 1908, giving methods of analysis for miscel- laneous supplies such as are examined in the Contracts Laboratory of this Bureau. The work of testing supplies delivered to the Govern- ment on contract grows continually both in extent and in importance, and the wide variation in the nature of the articles received renders the selection of methods for their analysis of peculiar difficulty. The expansion of the work during the last year and the number of new methods which have been tried and found satisfactory make a revi- sion of the bulletin advisable, since the assembling of these methods, tested by experience and modified accordingly, is of service both to those engaged in similar work and to the manufacturers who wish to comply with the specifications. I recommend, therefore, that this manuscript be published as Bulletin 109, Revised, of the Bureau of Chemistry. Respectfully, H. W. WILEY, Chief of Bureau. Hon. JAMES WILSON, Secretary of Agriculture. CONTENTS. Page. action 7 ific-gravity determinations 7 Temperature corrections 7 Westphal balance 8 Vacuum weighing 8 Choice of pyknometers 9 Calibration of pyknometers 10 Hydrometers 10 Arbitrary scales 10 Paint materials 11 Linseed oil 11 Spirits of turpentine 13 Benzin and light petroleum oils 15 Dry shellac 16 Shellac in alcohol . * 17 Damar varnish 17 Oil varnish 18 Japan drier 19 Paints 20 Strength test 20 White pigments 20 White lead 20 Lead zinc whites 22 Calcium pigments 24 Barytes 25 Lithopone 26 General scheme of analysis for a white pigment 27 Black pigments 28 Manufactured blacks 28 Graphite 29 Colored pigments 29 Yellow, orange, and red pigments 29 Chrome yellow 29 Zinc yellow 31 Red lead and orange mineral 31 Vermilion 32 Oxid and earth pigments 33 Blue pigments 34 Prussian blue 35 Ultramarine 36 Green pigments 37 Chrome green 37 Ultramarine and zinc yellow green 38 Mixed paints 39 5 6 CONTENTS. Page. Inks 41 Writing inks 41 Canceling inks having an oil base 44 Rubber-stamp inks 48 Materials used for the manufacture of canceling and other stamping inks. 49 Typewriter ribbons and carbon papers 50 Typewriter ribbons 50 Carbon papers 52 Miscellaneous 1 52 Babbitt metal (rapid analysis) 52 Glue i 54 Glycerin 55 Inking pads ^ 56 Lubricating oils 57 Phenolic disinfectant 59 Pipe covering and cement 61 Rubber 62 Soap 64 ILLUSTRATIONS. Page. Fro. 1. Pyknometer for very fluid liquids 9 2. Pyknometer for viscous liquids 9 3. Machine for testing the life of typewriter ribbons »..<, 51 SOME TECHNICAL METHODS OF TESTING MISCELLANEOUS SUPPLIES, Including Paints and Paint Materials, Inks, Lubricating Oils, Soaps, etc. INTRODUCTION. The Contracts Laboratory of the Bureau of Chemistry is called upon to ex- amine a great variety of articles purchased by the Government. The character of the examinations made varies to such an extent that it is not possible to outline any general methods which are not subject to more or less variation according to the purpose for which the material is intended. In this bulletin are assembled methods which have been found useful in a large number of cases. As a general rule these methods, while not original, have been compiled from a variety of sources and modifications introduced when necessary. No attempt has been made, therefore, except in a few cases, to refer to the original source. Acknowledgments, however, are due to the following authorities: Lunge, Lewkowitsch, Rideal, Fernbach, Hall, Parry, Hurst, Low, and many others. Nor has any attempt been made to give all of the methods for the examination of any of the materials mentioned, but only those which have been found most useful in this laboratory. The methods are presented in this form because of their inaccessibility at present — scattered as they are throughout chemical literature — and because those selected from the mass of material on the subject have been tested in practice and found to be especially satisfactory. SPECIFIC GRAVITY DETERMINATIONS. 1. Temperature Corrections. It is almost the universal custom among oil chemists in this country to express the specific gravity of oils at 15.5° C. (60° F.) as compared with water at the same temperature. Since this temperature is generally far below that of the laboratory, the determination is frequently made at a higher tempera- ture and the specific gravity calculated, using the formula G=G'+K (T— 15.5° C.), in which G=specific gravity at 15.5° C.; #'=speciflc gravity at T, and K=meari correction for 1° C. The correction K varies somewhat for different oils, but the value £"=0.00064 may be taken as sufficiently accurate in most cases for both hydrocarbon and fatty oils.0 These corrections, however, only apply to specific gravities referred to water at 15.5° C. (60° F.). The common custom of determining specific gravity at higher temperatures compared with water at the same temperature 0 For the value of K for different fatty oils see Allen's Commercial Organic Analysis, third edition, 2 (1) : 33, and for lubricating oils see Archbutt and Deeley, Lubrication and Lubricants, page 183. 7 g TESTING MISCELLANEOUS SUPPLIES. as the oil gives values to which this correction does not apply. If the correc- iion is applied to such values, the results obtained for 15.5° C. will contain large errors. This is well illustrated by the following example : The specific gravity of a sample of oil was carefully determined at two tem- peratures, with the following results : (1) Specific gravity at j (2) Specific gravity at r^r =0.9298. If the temperature correction is applied to the second value, the result is as follows: 0.9298+9.5X0.00064=0.93588 (specific gravity at 15.5° C.), an error far too large to be admissible. Now, from the known density of water at vary- ing temperatures," the specific gravity at ^^ ™? be calculated from equa- tion 2, as follows: If to this value tha correction is applied the result is 0.92795+9.5X0.00064= 0.93403, a value differing less than two in the fourth place from that actually determined, which is sufficiently accurate for most commercial work. It is a simple matter to determine the specific gravity at any selected tem- perature as compared with water at 15.5° C., but to do so most easily it is necessary to properly calibrate the pyknometer or plummet. The coefficient of cubical expansion of glass varies somewhat, and, while this correction should be applied when necessary in calibrating, care should always be taken to apply the correction for only a very few degrees. The Bureau of Standards in cali- brating volumetric apparatus assumes the cubical coefficient of expansion of glass per degree centigrade as equal to 0.000025, and if this value is always used, on determinations within a very few degrees of the desired temperature no appreciable error will be made. Since it is not safe to assume that the cubical coefficient of expansion of glass is the same for all temperatures, appa- nitiis should be calibrated at a temperature very near that at which the deter- mination is to be made. 2. Westphal Balance. Tho \\Ystphnl balance, on account of its convenience, is frequently used in makiiiK HjMTiHc-Knivity determinations, but it is difficult to determine the errors for this instrument. The weights can be easily tested, but the divisions <>M the arm are not easily tested, and both may be somewhat incorrect. It is I "-MIT not to use a Westphal balance, but a plummet may be used on an ana- lytical i>:i lance, which, when properly calibrated, is a very useful piece of appa- ratus, though the temperature is rather hard to control. 3. Vacuum Weighing. It is not necessary to reduce weighings to vacuum, since the corrections entering in both numerator and denominator would not make any appreciable difference in the result. It should be understood that both in calibrating the ni'punitiiR and in making the specific-gravity determination all weights are made in air. •Laiulholt. Bornstein, and Meyerhofer, Physikalische Chemische Tabellen, pages 37 to ;$!»: Bui-ran of Standards, Circular No. 19, page 27. SPECIFIC GRAVITY DETERMINATIONS. 9 O.//TJ/TZ snefer 4. Choice of Pyknometers. For the most accurate work with very fluid oils the Sprengel tube is the most accurate form. This apparatus is difficult to clean, and for ordinary work a specific-gravity bottle, with a cap to take up expansion, is the most convenient. A thermometer as part of the apparatus adds to the cost and has little or no advantage; in fact, it has some distinct dis- advantages. Aside from errors in the thermometer, the liquid, if placed in the pyknometer at a low temperature and then allowed to stand at room tem- perature until the thermometer registers 15.5° C., may not be of the same temperature throughout. If a pyknometer without a thermometer is used, it is necessary to bring it to its final temperature by keep- ing it in a bath for some time. By filling the pyknometer at about 14° C., inserting the stop- per, and then placing it in a — O.I mm diameter of/io/e - -2mm diameter of fio/e FIG. 1. — Pyknometer for very fluid liquids (actual size). FK;. i'. — Pyknometer for vis- cous liquids (one-half act- ual size). ith which is kept at 15.5° C. until no more liquid rises out of the capillary tube, one may be sure that the whole of the liquid in the pyknometer is at 15.5° C. The form of pyknometer shown in figure 1 is excellent for very fluid oils, but it is not suitable for those that are very viscous. For the latter use the form shown in figure 2. 21694°— Bull. 100— 12 2 10 TESTING MISCELLANEOUS SUPPLIES. 5. Calibration of Pyknometers. Clean and dry the pyknometer and weigh carefully after it has stood in the balance case for about thirty minutes. Fill with freshly boiled distilled water at about 14° C., insert the stopper, and place in a bath kept at exactly 15.5° C. When the temperature of the water in the pyknometer has risen to 15.5° C., u Inch will be shown by the water ceasing to rise out of the capillary tube of the stopper, wipe off the water on the top of the stopper, place the cap on the stopper, remove from the bath, wipe dry, and let stand until it has come to room tem- perature, then weigh. This will give the water value of the pyknometer at 15.5° C. Calibrate the same pyknometer at 50° C., proceeding in the same way, except that the water used in filling the apparatus is about 47° C. and is brought up to 50° C. in a bath kept at exactly that temperature. Multiply the weight of water at 50° C. by =1.011125. This product is the water value of the pyknometer at 50° C. In determining the specific gravity of any oil pro- ceed as in calibrating, only substitute the oil for the freshly boiled water, and divide the weight of oil by the water value of the pyknometer. Results are 1^5 ''i0 C1 ^50° C1 expressed in specific gravity at ^5 5° Q or -^ 5° Q If it is desired to calibrate the instrument at any other temperature, proceed in a similar manner and multiply the weight of water found by the density of water at 15.5° C. (0.99905) and divide by the density at the temperature used. (Use the tables of Landholt, Bernstein, and Meyerhofer, previously mentioned, page S.) A plummet may be calibrated in a similar manner by calculating the water equivalent at 15.5° C. ef the volume displaced at the different temperatures. 6. Hydrometers. Hydrometers are frequently used by oil chemists, and when properly made are very j: 1 instruments. They should be calibrated at 15.5° C. (60° F.) and used at that temperature when exact results are desired. If a hydrometer is T used at any other temperature (T,), the results should be recorded at1f- K0f^ and also reduce, I t<> i :..:. C. by applying the formula: G=G'+K (T— 15.5° C.). 15.5° C. Unless the determination is made at the temperature at which the hydrometer is calibrated the result should not be relied on beyond the third place. If more exact results are desired, do not use a hydrometer. 7. Arbitrary Scales. Tll(> ' M Tie gravity in terms of arbitrary scales, such as Baume, is very objectionable, since there are a number of Baume tables in print which differ widely. It is sometimes necessary, however, to report specific gravity In degrees Baume. When this is the case determine the specific gravity in the usual mam., i. mince, if necessary, to^'C-, and find the degrees Baum§ by !•>..) t^i. i" an American standard table. Such a table has been recalculated ami published by I lie Bureau ,,f Standards, Circular 19, and is calculated by means of the following formulae: F<>r liquids lighter than water: II I "i liquids heavier (ban water: Specific gravity at 15.5° C. —130. 145 — Specific gravity at 15.5° C. LINSEED OIL. 11 PAINT MATERIALS. LINSEED OIL. 1. Preparation of Sample. All tests are to be made on oil which has been filtered through paper at a temperature of between 15° and 30° C. immediately before weighing, with the exception of tests No. 0, Turbidity; No. 7, Foots; No. 9, Moisture and Volatile Matter, and No. 10, Ash. The sample should be thoroughly agitated before the removal of a portion for filtration or analysis. 2. Specific Gravity. Determine with a pyknometer, plummet, or hydrometer at 15.5° C. 3. Viscosity. Use the Engler-Ubbelohde method, making the determination at 20° C. 4. Flash Point, Open Cup. Set a nickel crucible 60 mm in diameter at the top, 40 mm in diameter at the bottom, and 60 mm in height in a hole in the middle of a sheet of asbestos board 200 mm square. The bottom of the crucible should project about !'."» mm through the asbestos. Support the asbestos on a tripod and suspend a ther- mometer reading to 400° C. in degrees in the center of the crucible, so that the lower end of the thermometer is 10 mm from the bottom of the crucible. Then pour in the oil until its level is 15 mm below the top of the crucible. Place a Bunsen burner below the crucible and regulate the size of flame so that the thermometer rises 9° a minute. As a test flame use an ordinary blowpipe attached to a gas tube. The flame should be about 6 mm long. Begin testing when the temperature of the oil reaches 220° C., and test for every rise of 3°. In applying the test move the flame slowly across the entire width of the crucible immediately in front of the thermometer and 10 mm above the surface of the oil. The flash point is the lowest temperature at which the vapors above the oil flash and then go out. 5. Fire Point. After noting the temperature at which the oil flashes continue the heating until the vapors catch fire and burn over the surface of the oil. The tempera- ture at which this takes place is the fire point. In determining the flash point note the behavior of the oil. It should not foam or crack on heating. Foam- ing and cracking are frequently caused by the presence of water. 6. Turbidity. Note whether the oil is perfectly clear or not. 7. Foots. Let a liter of the oil stand in a clear glass bottle for eight days, and then note the amount of sediment formed. The highest grades of oil show no turbidity or foots by this test. The claim is made that sometimes what would be called foots by the above method is due to the freezing out of fats of rather high melt- 12 TESTING MISCELLANEOUS SUPPLIES. ing point When a sufficient amount of the sample is available, heat one portion to 100° C. and set it aside for the determination of foots, together with a sample Just as it is received. Note also the odor of the warm oil, rubbing it on the hands; a small amount of fish oil may be detected in this way. 8. Break. Heat 50 cc of the oil in a beaker to 300° C. Note whether the oil remains un- changed or " breaks ; " that is, shows clots of a jelly-like consistency. 9. Moisture and Volatile Matter. Heat about 5 grams of oil in an oven at 105° for forty-five minutes ; the loss in weight is considered as moisture. This determination is of course not exact, as there is some oxidation. When a more accurate determination is desired, per- form the whole operation in an atmosphere of hydrogen. 10. Ash. Burn about 20 grams of oil in a porcelain dish and conduct the ashing at as low a temperature as possible. The best oil should contain only a trace of ash. An amount as large as 0.2 per cent would indicate an adulterated or boiled oil. Examine the ash for lead, manganese, and calcium. 11. Drying on Glass. Coat glass plates 3 by 4 inches with the oils to be examined, expose to air and light, and note when the film ceases to be tacky. A good oil should dry to an elastic coherent film in three days. Varying conditions of light, temperature, and moisture have such an influence on drying tests that for comparison of one linseed oil with others all samples must be run at the same time. 12. Drying on Lead Monoxid. Livache's test calls for precipitated lead, but litharge gives equally good results. Spread about 5 grams of litharge over the flat bottom of an aluminum dish 2.5 inches in diameter and five-eighths of an inch high; weigh the dish and the litharge; distribute as evenly as possible over the litharge 0.5 to 0.7 gram of the oil, weigh exactly, expose to the air and light for ninety-six hours, weigh again, and calculate the gain in weight to percentage based on the orig- inal weight of the oil used. 13. Acid Number. Weigh 10 grams of oil in a 200 cc Erlenmeyer flask, add 50 cc of neutral alcohol, connect with a reflux air condenser, and heat on a steam bath for half an hour. Remove from the bath, cool, add phenolphthalein, and titrate the free acid with fifth-normal sodium hydroxid. Calculate as the acid number (milligrams of potassium hydroxid to 1 gram of oil). The acid number varies with the age of the oil, and should be less than 8, though when the oil is re- tiiml with sulphuric acid it may show a higher acid number. Test for sul- phuric acid. SPIRITS OF TURPENTINE. 13 14. Sapoiiification Number. Weigh from two to three grains of oil ill a 200 cc Erlenmeyer flask, add 30 cc of a half-normal alcoholic solution of potassium hydroxid, connect with a reflux air condenser, heat on a steam bath for an hour, then titrate with half-normal sulphuric acid, using phenolphthalein as indicator. Always run two blanks with the alcoholic potash. From the difference between the number of cubic centimeters of acid required by the blanks and the determinations, calculate the saponification number (milligrams of potassium hydroxid to 1 gram of oil). The saponiflcation number should be about 190. 15. TTnsaponifiable Matter. As the saponification varies somewhat in pure oil, it is sometimes advisable to make a direct determination of unsaponifiable matter. Saponify from 5 to 10 grams of oil with alcoholic potassium hydroxid (200 cc of a half -normal solution) for an hour on a steam bath, using a reflux condenser. Then remove the condenser and evaporate the alcohol as completely as possible ; dissolve the soap in 75 cc of water, transfer to a separatory funnel, cool, shake out with two portions of 50 cc each of gasoline distilled between 35° and 50° C., wash the gasoline twice with water, evaporate the gasoline, and weigh the unsaponifiable matter, which in raw linseed oil should be below 1.5 per cent ; in boiled oil it is somewhat higher, but should be below 2.5 per cent. 16. lodin Number. Weigh in a small glass capsule from 0.2 to 0.25 gram of oil, transfer to a 350 cc bottle having a well-ground stopper; dissolve the cil in 10 cc of chloro- form and add 30 cc of Hanus solution ; let it stand with occasional shaking for one hour, add 20 cc of a 10 per cent solution of potassium iodid and 150 cc of water, and titrate with standard sodium thiosulphate, using starch as indicator. Blanks must be run each time. From the difference between the amounts of sodium thiosulphate required by the blanks and the determination, calculate the iodin number (centigrams of iodin to 1 gram of oil). The iodin number of raw linseed oil varies from 175 to 193, though Gill states that a pure raw oil may give a value as low as 160. Boiled oil may be very much lower. Make the Hanus solution by dissolving 13.2 grams of iodin in 1,000 cc of glacial acetic acid which will not reduce chromic acid, and adding 3 cc of bromiri. 17. Rosin or Rosin Oil (Liebennann-Storch Test). To 20 grams of oil add 50 cc of alcohol, heat on a steam bath for fifteen min- utes, cool, decant the alcohol, evaporate to dryness, add 5 cc of acetic anhydrid, warm, cool, draw off the acetic anhydrid, and add a drop of sulphuric acid, 1.53 specific gravity. Rosin or rosin oil gives a fugitive violet color. SPIRITS OF TURPENTINE.0 1. Color. The best quality of spirits of turpentine should be water white. °If wood turpentine has been carefully refined, it will comply with all the tests given for spirits of turpentine, but it can almost invariably be distin- guished from the latter by its characteristic odor. 14 TESTING MISCELLANEOUS SUPPLIES. 2. Specific Gravity. Determine the specific gravity with a pyknometer, plummet, or hydrometer at 15.5° C. Pure gum turpentine should have a density between 0.862 and 0.875. Wood turpentine may, however, range from 0.860 to 0.910 or even higher. 3. Distillation. Connect a distilling flask of 150 cc capacity with a condenser having a ther- mometer. Introduce 100 cc of turpentine and heat with a Bunsen burner. The initial boiling point should be about 156° C., and 95 per cent should distil over between 153.5° and 165.5° C. 4. Residue on Evaporation. Evaporate 10 grams on the steam bath; the residue should be less than 2 per cent. 5. Refractive Index. Determine with a Zeiss direct reading refractometer at 20° C. The index of refraction for gum turpentine should be from 1.4690 to 1.4740; for wood tur- pentine, 1.4685 to 1.5150. 6. Action of Sulphuric Acid (Polymerization). Measure 6 cc of turpentine in a stoppered, thin-walled tube graduated to 0.1 cc (carbon tubes). Place the tube in cold water and pour in slowly a mixture of four parts of strong sulphuric acid and one part of fuming sulphuric acid. Add the acid slowly, and avoid an excessive rise in temperature. Shake the tube so as to mix the turpentine and the acid, add finally about 20 cc of the acid, stopper the tube, mix thoroughly, cool, allow to stand thirty minutes, and note the volume of unpolymerized oil that collects on top of the acid layer. Then let stand for eighteen hours and again note the volume. A pure turpen- tine should show less than 0.3 cc unpolymerized at the end of thirty minutes, and less than 0.5 cc after eighteen hours. This method will indicate gross adulteration, but will not detect admixtures of very small amounts of mineral oil. Donk has perfected a method which determines the presence of as little as 1 per cent of mineral oil in turpentine. This method is as follows : Sulphuric acid of 38 times the normal strength (101.5 per cent) is prepared by mixing very strong sulphuric acid with fuming sulphuric acid. It must be determined by titration that this reagent is of the exact strength required, for with 37.5 times normal acid (100 per cent) the turpentine is not completely destroyed, and with .-u-id stronger than 101.5 per cent the amount of mineral oil dissolved becomes excessive. Place about 25 cc of the special sulphuric acid in a flask having a narmw graduated neck (n Babcock bottle does very well), cool in ice water, add 5 cc of tlie turpentine to be tested and cool the flask again, shaking it carefully and :n«. hi ing any excessive rise in temperature by frequent cooling. The flask should m>\vr !•<• too hot to hold in the palm of the hand. Then place it in a bath of cold water and heat the bath at such a rate that in about five minutes the tempera MI iv will lu> 65° C. During the heating shake the bottle about every fifty seconds, finally shaking very thoroughly so as to insure the contact of p:irtir|(> of the sample with the acid. Cool to room temperature and add ordinary strong sulphuric acid in sufficient amount to bring the unpolymerized li'inid up in tin- graduated neck. Let stand over night or whirl in a centrifuge and read the volume oil the neck. BENZIN AND LIGHT PETROLEUM OILS. 15 Pure turpentine should leave ;i residue <>f not over 0.04 cc, which is not limpid and which has a refraetive index of not less than 1.r>00. If the unpolyinerized residue is 0.04 cc or less, mineral oil may be assumed to be absent. If the residue is greater, calculate from It the percentage of mineral oil present. This will be, of course, only approximate, for there is some residue from pure turpentine and some mineral oil is dissolved by the acid; but for all practical purposes it may be assumed that the errors balance one another, and hence it is not advisable to apply any correction. 7. Spot Test. Place a drop on filter paper and allow it to dry at room temperature; It should leave no stain. 8. Flash Point. Support a crucible, such as is used in determining the flash point of linseed oil, in a vessel of water at 15° to 20° C. ; the water should cover about two- thirds of the crucible. Fill the crucible to within about 2 cm of the top with turpentine, insert a thermometer, and heat the water bath slowly, 1° per minute. Begin at 37° and test for the flash at each rise of 0.5°. The turpentine should not flash under 40.5° C. BENZIN AND LIGHT PETROLEUM OILS. The term benzin is used for a number of light petroleum oils. In the paint- ing trade it generally refers to a product of about 62° Baum6 (0.7292 sp. gr.). The petroleum benziu of the U. S. Pharmacopoeia is a lighter oil, being a light gasoline. 1. Specific Gravity. Determine with a spindle, pyknometer, or plummet at 15.5° C. The deter- mination can be made at room temperature and corrected to 15.5° C. 2. Sulphur (Sodium Nitroprussid Test). To 100 cc of the sample in a flask add about 1 gram of bright metallic sodium, connect with a reflux condenser, and boil for one hour. Cool, add water drop by drop until the metal is dissolved, separate the aqueous liquid, and test with a drop of sodium nitroprussid solution. A fine violet-blue coloration indi- cates sulphur. 3. Sulphur Compounds and Pyrogenous Products (U. S. P. Test). To 100 cc of the sample add 25 cc of a solution of 10 per cent anhydrous ammonia in 95 per cent alcohol (spirit of ammonia U. S. P.), add 1 cc of silver nitrate solution. Boil gently for five minutes. A brown coloration indicates sulphur compounds or pyrogenous products. 4. Residue on Evaporation. Place 25 cc in a 100 cc platinum dish, heat on steam bath for thirty min- utes, and weigh residue. No residue should be left by this test. 5. Fractional Distillation. Light petroleum oils are usually tested only for specific gravity ; but as light and heavy distillates may be mixed, the specifications would be improved by 16 TESTING MISCELLANEOUS SUPPLIES. requiring that a certain fraction should distil between specified temperatures. To make this determination distil 100 cc in a round-bottom flask 6.5 cm in diameter ; the neck should be 1.6 cm in diameter and 15 cm long, with a side tube set in the middle of the neck at an angle of 75°. The surface of the liquid should be 9 cm below the side tube, and the bulb of the thermometer just below the side tube. 6. Benzol. Mix the sample with 8 volumes of strong sulphuric acid and 2 volumes of nitric acid; heat gently for ten minutes, allow to cool, and note odor. The odor of nitrobenzol indicates benzol. 7. Color and Odor. Note color of sample and odor both in bulk and after rubbing on hands. DRY SHELLAC. 1. Detection of Rosin (Liebermann-Storch Test). To 1 gram of the sample add 15 cc of acetic anhydrid, heat gently until solu- tion is complete, cool, draw off the acetic anhydrid, and add a drop of sul- phuric acid 1.53 specific gravity. A fugitive violet color indicates rosin. 2. Determination of Rosin (Langmuir's Method).0 To 0.200 gram of ground shellac, in a 250 cc dry bottle of clear glass with a well fitting ground glass stopper, add 20 cc glacial acetic acid, and warm gently until the shellac is dissolved (a little shellac wax will remain undissolved, and at least half an hour should be allowed, for the shellac itself is difficultly solu- ble). Then add 10 cc of chloroform and cool the solution down to 21° to 24° C. Add 20 cc of Wijs' iodin monochlorid solution, shake, and keep in the dark at a temperature between 21° and 24° C. for one hour, add 10 cc of potassium iodid rotation, and titrate immediately with thiosulphate — 25 cc may be run in at once and the remainder gradually with vigorous shaking, adding starch solu- tion Just before the end. The end point is sharp; any color returning after half a minute or so may be disregarded. If less than 25 cc thiosulphate is required reject result and repeat determination, using a smaller quantity of The \Vija solution may be made either by dissolving iodin monochlorid in glacial acetic acid or by dissolving iodin in glacial acetic acid and passing in dry chlorin until the halogen content (figured as iodin) has nearly doubled; there may be a very slight excess of iodin, but there must be no excess of chlorin above the theory for Iodin Hilm-id. The halogen content figured as iodin should be :i I H.I n is grams per 1,000 cc. The glacial acetic acid used must be free from n-.lu. in- snhstMiHvs and have a melting point between 14.7° and 15° C. The tltioHulphate solution should he a decinormal sodium thiosulphate solution. The potassium iodid solution is 10 per cent. Blanks must be run each time and the iodin number calculated from the difference between the titration result on the Hunk <>n the sample. Calculate the per cent of rosin, using the formula : ^1 _ ft y=100 y— |/ . where y=per cent rosin, M =Iodin number of shellac, 2Vr=iodin number of rosin, and A=iodin number determined. 0 J. Soc. Chern. Ind., 1905, 24 : 12-17. DAMAR VARNISH. 17 In calculating the percentage of rosin, assume the iodin number for pure rosin to be 228, that for pure unbleached shellac to be 18, and for pure bleached shellac, 10. These figures will tend to give somewhat low results on rosin percentage, but should be used, as it is then safe to say that there is at least as much rosin present as is found by analysis. 3. Alcohol Insoluble. Add to 1 gram of shellac 60 cc of 95 per cent alcohol, heat to boiling, and filter on a Gooch crucible ; wash with about 60 cc of boiling 95 per cent alcohol, dry at 100° C. and weigh. 4. Loss on Drying — Water. Heat 1 gram for two hours in a water oven at 100° C. Loss in weight is cal- culated as water. 5. Ash. 'Determine ash on 1 gram, heating only to dull redness. SHELLAC IN ALCOHOL. 1. Total Solids and Ash. Evaporate 1 gram of the varnish in a porcelain dish on the water bath and dry to constant weight in steam oven (about five hours). Ash residue as in dry shellac. 2. Examination of Solvent. Note the odor of the varnish, as little can be learned from the odor of the distillate. Distil 100 grams of the varnish until incipient decomposition begins ; redistil the distillate, noting the boiling point, and determine the specific grav- ity of the distillate. By this means it is usually an easy matter to determine whether the varnish is made with wood alcohol, grain alcohol, or denatured alcohol. Dilute a portion with three times its volume of water; if benzin is present it will separate. 3. Test for Rosin. Do not use the residue from the total solids determination, but carefully weigh in a capsule a quantity of varnish sufficient to yield 0.2 gram of solid material, evaporate to a paste, and treat the residue as described under "dry shellac," paragraphs 1 and 2, page 16. DAMAR VARNISH. Damar varnish, which should be a solution of damar gum in spirits of tur- pentine, is examined by the methods for oil varnishes (page 18). The percent- age and nature of the volatile oil, specific gravity, and flash point are important tests. It should yield little ash, which should contain no lead or manganese. When flowed on glass it should set to touch in a very short time (less than thirty minutes) and mixed with zinc white should yield a good white enamel surface when applied to wood. In testing for rosin by the Liebermann-Storch test (page 19), the light claret color characteristic of damar should not be mistaken for the violet color char' 21694°— Bull. 109—12 3 18 TESTING MISCELLANEOUS SUPPLIES. acteristic of rosin. The Hirschsohn test for rosin, which is less delicate than the Ldebermann-Storch method, is as follows : Digest about 1 gram of the dried resin with approximately 20 cc of am- monium hydroxid for twenty minutes, filter and acidify with acetic acid. With pure damar the liquid remains clear, but if rosin is present a precipitate forms, and when the proportion of rosin is large the entire acid liquid may become gelatinous. The acid number of damar is lower than that of rosin and higher than that of the hard varnish gums. An oil varnish made of hard gums will generally have an acid number of less than 12. A damar varnish will have an acid number between 15 and 20, a higher number generally indicating the presence of rosin. OIL VARNISH. The methods of analysis for varnish are far from satisfactory. A number of methods for determining the amount of gums in varnish have been proposed, but none of them are reliable. This determination, therefore, is not made, but the combined percentage of oil and gums is reported. 1. Appearance and Odor. Transfer the sample if it is in a metal container to a glass-stoppered cylinder or bottle and note its appearance, color, transparency, body, and whether any sediment is deposited. The presence of light petroleum oil or wood turpentine may often be detected by noting the odor of the sample. After making these observations, the sample should be thoroughly mixed before making the re- maining tests. 2. Specific Gravity. Determine with a pyknometer, plummet, or hydrometer at 15.5° C. 3. Flash Point. Determine as in spirits of turpentine; but begin testing at 25° C. and stir the varnish while heating. A low flash point indicates light petroleum oil. 4. Viscosity. The determination of viscosity is seldom necessary ; but if desired, make the determination by the Engler-Ubbelohde method at 20° C. 5. Volatile Oils. Weigh 100. grams of the varnish into a 500 cc flask, connect with a spray trap ami a vertical condenser, and pass through it a current of steam, first heating the flask in an oil bath at 100° C. ; with the steam still passing through, raise the temperature of the bath to 130° C. Catch the distillate in a small separa- tory funnel; continue distillation until 300 cc of water has been condensed. Portions of this water may be drawn from the cock of the separatory funnel from time to time, but care must be taken not to draw out any of the volatile oil. Let the distillate stand until it separates into two layers, then draw off the water and weigh the light oils in a tared stoppered flask. Filter the light oils through dry p.-ipor and examine as directed under spirits of turpentine (page 13). A slight error is caused by the solubility of turpentine in water; i liis amounts to about 0.3 to 0.4 cc for each 100 cc of water. When sufficient varnish is available, it is well to take another portion and dis- til, without steam or spray trap, placing the weighed flask in an oil bath. Note JAPAN DRIER. 19 the temperature of the bath at which distillation begins, and continue distilla- tion at a temperature of 185° C. in the oil bath, finally raising the temperature to 200° C. This method generally tends to give lower results on volatile oils than the steam distillation method; but the distillate can be tested for water- soluble volatile liquids, which would be lost by the steam distillation. 6. Fixed Oil and Gums. The percentage of fixed oil and gums is obtained by subtracting the percent :ig<- of volatile oils from 100. A check upon this determination is obtained by weigh- ing the residue from the dry distillation. 7. Acid Number. Determine the acid number in the usual way, using 10 grams of varnish. 8. Rosin. After getting the acid number, decant the alcohol, evaporate, and apply the Liebermauu-Storch test for rosin (page 16). 9. Ash. Determine the ash in 10 grams, using a quartz or porcelain dish and carrying out the determinations at a low heat, best in the muffle. Determine the reac- tion of the ash to litmus paper and make a qualitative analysis. It is fre- quently well to make a quantitative determination of lime, a large amount of which indicates rosin. It is sometimes advisable to determine the percentage of lead and manganese. Some lead will, however, be lost in the ashing, and if a correct lead determination is required, follow the method described for mixed paint vehicles (page 41). 10. Miscellaneous. It is not possible from such an examination as has just been described to de- cide on the value of a varnish for any particular purpose, though valuable in- formation as to the materials used in its manufacture may be obtained. An examination of the varnish film should, therefore, always be made. The films are best made by flowing the varnish on carefully cleaned plates of glass, and allowing to dry at room temperature in a vertical position. Observe the time of setting and the nature of the film after twenty-four hours, noting the trans- parency, hardness, elasticity, tendency to dust by scratching, etc. After thor- ough drying, immerse the plate in water over night, dry, and note appearance. The working quality of a varnish must be determined by application on wood, and it is best to make this test on well seasoned and perfectly smooth white pine. Apply a thin coat of the varnish to wood and allow to dry, sandpaper down smooth, and then apply the coat to be tested. Observe how the varnish works under the brush, character of coat, etc. This panel, after drying, may be used for further testing as to whether the varnish will stand rubbing, etc. JAPAN DRIER. The analysis of Japan drier is conducted in the same way as that of varnish. The following specifications for Japan drier for the Philadelphia and Reading 20 TESTING MISCELLANEOUS SUPPLIES. Railroad are given in an article by Job<* on Practical Testing and Valuation of Japan : The material desired consists of a pure turpentine hardener and oil drier, conforming to the following: 1st When equal parts by weight of the Japan and of pure turpentine are thoroughly mixed and poured over a slab of glass, which is then placed nearly vertical at a temperature of 100 degrees Fahrenheit, with a free access of air but not exposed to draft, the coating shall be hard and dry, neither brittle nor sticky, in not exceeding twelve minutes. 2d. When thoroughly mixed with pure raw linseed oil at the ordinary temper- ature in proportions of 5 per cent by weight of Japan to 95 per cent by weight of raw linseed oil, no curdling shall result, nor any marked separation or set- tling on standing. 3d. When the above mixture is flowed over a slab of glass, which is then placed nearly vertical at a temperature of 100 degrees Fahrenheit, with free access to air but not exposed to draft, the coating shall dry throughout, neither brittle nor sticky, in not exceeding two hours. 4th. When five cubic centimeters of the Japan are poured into 95 cubic centi- meters of pure turpentine at the ordinary temperature, and thoroughly shaken, a clear solution shall result, without residue, on standing one hour. 5th. After evaporation of the turpentine, the solid residue must be hard and tough, and must not " dust f' when scratched with a knife. 6th. Benzin or mineral oil of any kind will not be permitted. This, with the following specification of the U. S. Navy Department, will fur- nish a very good basis for an examination of Japan drier : 1. Japan drier must not flash below 105° F. (open tester) ; must be of the best quality and made from pure kauri gum, pure linseed oil, pure spirits of turpen- tine, and lead manganese driers, and be free from adulterants and all other foreign materials ; must set to touch in from one-fourth to one hour, dry elastic in from 18 to 21 hours at a temperature of 70° F., and must not rub up or pow- der under friction by the finger. When mixed with pure raw linseed oil in the pmiiortion of 8 parts of oil to 1 part of drier, must remain clear for two hours and set to touch in from 6 to 7 hours at a temperature of 70° F. PAINTS. STRENGTH TEST. The strength of pigments does not bear a very close relationship to the chem- ical composition, hence a quantitative analysis does not necessarily give the in- formation actually desired. For this reason the "strength test" is of great importance. A standard color is usually selected, but in case no standard is used all of a sorios of similar colors may be compared with one another. Weigh eqn.-i! portions of the samples to be tested, and, if available, a standard sampi*-: add to oach equal portions of pure zinc oxid, the amount depending upon the character of the color tested (this may vary from 5 to 60 times that of tho color). Rub up each mixture on a glass plate or, better, a stone slab, until there is no change in shade on further rubbing. Rate the samples in the order of their depth of tint. White pigments may be rated in a similar way by grinding up each with the same colored pigment, such as Venetian red. With yellows use chrome green or Prussian blue instead of a white diluent. Add the same number of drops of raw linseed oil to all samples in grinding for the strength test. WHITE PIGMENTS. 1. White Lead. Pure white lead Is basic carbonate of lead and should approach the follow- ing con>iK>sttlon : 2PbCOsPb ( OH ) .. It should be completely soluble in acetic aci.l r l.iiriiiin and calcium by tho Uaim- tost. Tsc- an iron wire, as platinum will bo ruinod l>y the lend. o Chemical Engineer, 1906, 4: 198-202. WHITE PIGMENTS. 21 (a) TOTAL LEAD. Weigh 1 grain of the sample, moisten with water, dissolve in acetic acid, filter, and wash, ignite, and weigh the insoluble impurities. To the filtrate from the insoluble matter add 25 cc of sulphuric acid (1:1), evaporate and heat until the acetic acid is driven off; cool, dilute to 200 cc with water, allow to stand for two hours, filter on a Gooch crucible, wash with 1 percent sulphuric acid, ignite, and weigh as lead sulphate. Calculate to total lead (PbSO4X 0.68292=Pb), or calculate to basic carbonate of lead (white lead) by multi- plying the weight of lead sulphate by 0.85258. The filtrate from the lead sulphate may be used to test for other metals, though white lead is only rarely adulterated with soluble substances; test, however, for zinc, which may be present as zinc oxid. Instead of determining the total lead as sulphate it may be determined as lead chromate by precipitating the hot acetic acid solution with potassium dichromate, filtering on a Gooch crucible, igniting at a low temperature, and weighing as lead chromate. (b) COMPLETE ANALYSIS. If it is desired to determine the exact composition of a pure white lead, honi 1 gram of the pigment in a porcelain boat in a current of dry, carbon-diox id- free nitrogen, catching the water in sulphuric acid and calcium chlorid and the carbon dioxid in soda lime or potassium hydroxid (1.27 specific gravity). By weighing the residue of lead monoxid in the boat all the factors for deter- mining the total composition are obtained. Figure the carbon dioxid to lead carbonate (PbCO3), calculate the lead monoxid corresponding to the lead car- bonate (PbCO3) and subtract from the total lead monoxid, calculate the re- maining lead monoxid to lead hydroxid (Pb(OH)2), calculate the water corre- sponding to lead hydroxid and subtract from the total water, the remainder being figured as moisture. This method assumes the absence of acetic acid. Thompson ° states that acetic acid varies from 0.05 per cent in Dutch process white lead to 0.7 per cent In some precipitated white leads. It is then more accurate to determine the carbon dioxid by evolution ; this is especially the case when working with a lead extracted from an oil paste, as the lead soap and unextracted oil will cause a considerable error by the ignition method. In determining carbon dioxid by the evolution method, liberate the carbon dioxid with dilute nitric acid, have a reflux condenser next to the evolution flask and dry the carbon dioxid with cal- cium chlorid before absorbing it in the potassium hydroxid bulbs. (c) ACETIC ACID. It is sometimes necessary to determine acetic acid. The Navy Department specifications demand that white lead shall not contain " acetate in excess of fifteen one-hundredths of 1 per cent of glacial acetic acid." Thompson's method ° is as follows : Eighteen grams of the dry white lead are placed in a 500 cc flask, this flask being arranged for connection with a steam supply, and also with an ordinary Liebig condenser. To this white lead is added 40 cc of syrupy phosphoric acid, 18 grams of zinc dust, and about 50 cc of water. The flask containing the material is heated directly and distilled down to a small bulk. Then the steam °J. Soc. Chem. Ind., 1905, 24:487. 22 TESTING MISCELLANEOUS SUPPLIES. is passed into the flask until it becomes about half full of condensed water, when the steam is shut off and the original flask heated directly and distilled down to the same small bulk — this operation being conducted twice. The distillate is then transferred to a special flask and 1 cc of syrup phosphoric acid added to ensure a slightly acid condition. The flask is then heated and distilled down to a small bulk — say, 20 cc. Steam is then passed through the flask until it contains about 200 cc of condensed water, when the steam is shut off and the flask heated directly. These operations of direct distillation and steam distil- lation are conducted until 10 cc of the distillate require but a drop of tenth- normal alkali to produce a change in the presence of phenolphthalein. Then the bulk of the distillate is titrated with tenth-normal sodium hydroxid, and the acetic acid calculated. It will be found very convenient in this titration, which amounts in some cases to 600-700 cc, to titrate the distillate when it reaches 200 cc, and so continue titrating every 200 cc as it distils over. If the white lead contains appreciable amounts of chlorin it is well to add some silver phosphate to the second distillation flask and not carry the distil- lation from this flask too far at any time. The method used by the chemists of the Navy Department is as follows: Weigh 25 grams of white lead in an Erlenmeyer flask, add 75 cc of 25 per cent phosphoric acid, distil with steam to a 500 cc distillate, add to the distillate some milk of barium carbonate, bring to a boil, filter, keeping the solution at the boiling point (it is not necessary to wash), add an excess of sulphuric acid to the filtrate and determine the barium sulphate in the usual manner ; subtract 53 milligrams from the weight of the barium sulphate and calculate the re- mainder as acetic acid (BaSO4X0.515=CH3COOH). The object of this rather indirect method is to avoid any error that might arise from fatty acids being carried over by the steam distillation. For white lead that has not been ground in oil, Thompson's method is to be preferred. 2. Lead Zinc Whites. French zinc is practically pure zinc oxid, being made from metallic zinc. American zinc is made by roasting zinc ores, and almost always contains some lead. Sublimed lead is made by volatilizing galena, and as sphalerite is usually associated with the galena, it nearly always contains zinc oxid in addition to the basic sulphate of lead. Zinc lead is essentially lead sulphate and zinc oxid. (a) MOISTURE. Dry 2 grams for 2 hours at a temperature of 105° C. (b) SOLUBLE SULPHATE. Boil 1 gram of the sample with 75 cc of water and 25 cc of alcohol, filter, wash with a mixture of alcohol and water (1:3), and determine sulphuric acid In the filtrate by the usual method of precipitation with barium chlorid. The soluble sulphate may be calculated to zinc sulphate. (c) INSOLUBLE AND TOTAL SULPHATE. Treat 1 gram in a GOO cc beaker with 10 cc of water, 10 cc of strong hydro- chloric acid, and 5 grams of ammonium chlorid, heat on a steam bath in a cov- ered beaker for five minutes, add hot water to make about 400 cc, boil for five minutes, and filter to separate any insoluble material. (A pure color should be completely dissolved.) Wash with hot water, ignite, and weigh the insoluble. WHITE PIGMENTS. 23 To the boiling hot filtrate add slowly, with stirring, 20 cc of a 10 per cent barium chlorid solution ; let stand for two hours on the steam bath, filter, wash, ignite, and weigh as barium sulphate. Calculate the total sulphate (SO4), deduct from it the soluble sulphate (SOO, and figure the remainder as lead sulphate. (d) LEAD AND ZINC. (1) IN THE ABSENCE OF OTHER METALS. Dissolve 1 gram by boiling fifteen minutes with 250 cc of water and 20 cc of strong nitric acid, add 5 cc of sulphuric acid, evaporate, and heat until all of the nitric acid is driven off and copious fumes of sulphuric anhydrid begin to come off; cool, dilute, filter on a Gooch crucible, wash with 1 per cent sul- phuric acid, ignite, and weigh as lead sulphate. Deduct from this the lead sulphate obtained from the sulphate determination and figure the remainder to lead monoxid. Evaporate the filtrate from the lead sulphate to 100 cc, cool, add 5 grams of microcosmic salt dissolved in . water, then add ammonium hydroxid until the solution is just neutral (test with litmus paper). Then add 2 drops of ammonium hydroxid and 1 cc of acetic acid, stir vigorously, heat on the steam bath for an hour (the precipitate should assume a crystal- line character and settle well at the end of that time), filter on a Gooch cruci- ble, wash with hot water, ignite at first at a very low heat and finally heat t« redness, and weigh as zinc pyrophosphate (Zn2P2O7), calculate to zinc oxid (Zn2P2O7X 0.53412 =ZnO). Deduct the zinc oxid present as zinc sulphate and report the remainder as zinc oxid. (2) IN THE PRESENCE OF OTHER METALS. I With a sample containing calcium or magnesium salts the lead should be pre- cipitated as sulphid from hydrochloric acid solution, then dissolved in nitric acid and determined as sulphate. The filtrates from the lead sulphid can be used to determine the zinc. If it is desired to make a gravimetric determination of zinc, boil off the hydrogen sulphid, add a little bromin water to oxidize any trace of iron, then add ammonium hydroxid in excess; only in rare cases will any iron appear here, but if it does, filter, redissolve, and reprecipitate. Manga- nese will be present only very rarely ; if present, it can be precipitated by add- ing bromin and warming. Make the zinc solution acid with acetic acid, heat to a boil, and pass hydrogen sulphid into the hot solution until it is completely saturated (for twenty to thirty minutes). Allow the zinc sulphid to settle for five hours, filter, wash with hydrogen sulphid water, dissolve the zinc sulphid in hot dilute hydrochloric acid, boil off the hydrogen sulphid, and determine the zinc as pyrophosphate, as described. The hydrogen sulphid method is very troublesome on account of the difficulty of filtering and washing zinc sulphid, and when calcium or magnesium is present it is generally better to determine the zinc volumetrically. This is best done by Low's method.0 In the absence of iron and manganese take the filtrate from the lead sulphid, add ammonium hydroxid until alkaline, then hydrochloric acid until just acid, then add 3 cc of strpng hydrochloric acid, dilute to about 250 cc, and titrate with standard ferrocyanid just as in standardizing the solution. When iron and manganese are present follow the method for oxidized ores as described by Low. a Technical Methods of Ore Analysis, p. 209. 24 TESTING MISCELLANEOUS SUPPLIES. (e) STANDARDIZING THE FEBBOCYANID SOLUTION. Prepare a solution of potassium ferrocyanid containing 22 grams of the crystallized salt to the liter. Standardize this solution as follows: Weigh carefully about 0.2 gram of pure ziiic and dissolve in 10 cc of strong hydro- chloric acid (sp. gr. 1.20), using a 400 cc covered beaker. Dilute somewhat, place a bit of litmus paper in the solution as an indicator, and make faintly alkaline with ammonium hydroxid. Again acidify faintly with hydrochloric acid and then add 3 cc excess of the strong acid. Dilute now to about 250 cc and heat nearly to boiling. Titrate the hot liquid with the ferrocyanid solution as follows: Pour off about one-third of the zinc solution and set it aside in a beaker. Titrate the remainder by running in a few cubic centimeters at a time until a drop, when tested on a porcelain plate with a drop of a 5 per cent solution of uranium nitrate, shows a brown tinge. Now add the greater part of the reserved por- tion and continue the titration more cautiously until the end-point is again passed. Finally, add the last of the reserved portion, and to save rinsing out the beaker, pour a large part of the solution back into it again and then empty it once more. Finish the titration very carefully, ordinarily by testing after each addition of two drops. Instead of using a single drop of the zinc solution for the test, the reaction is sharper if a quantity equivalent to several drops is taken. If this is done near the end of the titration the amount of zinc lost thereby will be insignificant. A convenient way of making the test is to use a medicine- dropper and place a single drop of the uranium solution in each depression of the test-plate at the outset. By using a glass tube instead of a rod for a stirrer, any desired quantity of the solution can be quickly removed for a test. When the final brown tinge is obtained, note the reading of the burette, and then wait a minute or two and observe if one or more of the preceding tests do not also develop a color. The end-point is always passed by a test or two and the burette reading must be corrected accordingly. A further correction must also be made for the amount of ferrocyanid required to produce a color under the same conditions when no zinc is present. This is ordinarily 2 drops. One cubic centimeter of the standard solution will equal about 0.005 gram of zinc. 3. Calcium Pigments. Whiting is calcium carbonate and is used extensively as an adulterant. Plaster of Paris is dehydrated calcium sulphate and gypsum is the natural hydrated calcium sulphate (CaSO4+2H3O). Besides these a precipitated cal- cium sulphate IB used, especially as a basis upon which to precipitate anilin colors. A microscopic examination is therefore of importance. Free lime is objectionable and is especially to be looked for in whiting. Boil 2 grams of the sample for five minutes with 100 cc of water and add phe- nol phthalein. If red color develops, free lime may be assumed to be present, though the alkaline reaction may possibly be due to hydrates of sodium or potassium. For quantitative analysis, when such is necessary, boil 1 gram of the sam- ple with 26 cc of hydrochloric acid '(1:1), add 150 cc of water, boil, filter, wash, ignite, and weigh the insoluble. Precipitate iron and aluminum hydrates from the filtrate with ammonium hydroxid, dissolve in hydrochloric acid and n precipitate, wash, ignite, and weigh. Unite the filtrates from the iron and WHITE PIGMENTS. 25 alumina precipitation, heat to boiling, add slowly 25 cc of a boiling solution of ammonium oxalate (the solution should be saturated at room temperature) ; let stand on the steam bath for one to two hours, filter, wash free from chlorids (cold water is better for washing than hot water), and determine the calcium by igniting to constant weight in a platinum crucible. A blast lamp is generally recommended for this purpose, but if a platinum triangle is used to support the crucible, from fifteen to thirty minutes' ignition with a good Bun- sen burner will convert the oxalate entirely to oxid with less danger of loss than with a blast. If great accuracy is required, it is well to dissolve the calcium oxalate in hydrochloric acid and reprecipitate. Instead of weighing the calcium as CaO, it may be determined volumet- rically by titrating with potassium permanganate as follows: Precipitate as described, and wash free from ammonium oxalate (when free the filtrate will not react for chlorids). Return the moist precipitate to the beaker by washing with a stream of water, remove the small quantity remaining on the paper with hot dilute sulphuric acid, add 20 cc of sulphuric acid (1:1), dilute to 300 cc, heat to about 80° C., and titrate with tenth-normal potassium perman- ganate (1 cc=0.0028 gram of calcium oxid). Evaporate the filtrate from the calcium oxalate to about 100 cc, make neutral by the addition of hydrochloric acid, if necessary cool in ice water, add an excess of a solution of sodium ammonium phosphate, stir, allow to stand for fifteen minutes, then add 20 cc of strong ammonium hydroxid, stir well, allow to stand for two hours, filter in a gooch, wash with 2 per cent ammonium hydroxid, ignite gently at first and finally to a bright red, and weigh as magnesium pyrophosphate (MgzPzOi). 4. Barytes. Barytes, being one of the cheapest white pigments, is seldom adulterated. It should be white and well ground. A niiscroscopic examination should be made to determine the evenness of grinding, the size and angularity of the par- ticles, and whether they are amorphous or crystalline. (a) MOISTURE. Dry 2 grams of the sample at 105° C. for two hours. The loss in weight rep- resents the moisture. (b) INSOLUBLE. Boil 1 gram in hydrochloric acid (1:3), evaporate to dryness, moisten with hydrochloric acid, dilute with water, boil, filter, and wash. If a pre- vious qualitative examination has determined the absence of lead and other easily reduced metals, ignite in platinum, weigh, and calculate the total insol- uble; this may generally be safely calculated to barium sulphate (BaSOO, but to determine any silica that may be present add hydrofluoric acid and sul- phuric acid, evaporate the acids, ignite, and weigh; calculate the loss as silicon dioxid. (c) ALUMINA AND OXIDS OF IRON. Add ammonium hydroxid to the filtrate from the total insoluble, boil, filter, wash, ignite, and weigh as oxid of iron and alumina. in<;!>-(°— nun. io<)— 12 4 26 TESTING MISCELLANEOUS SUPPLIES. (d) LIME. Precipitate with ammonium oxalate from the filtrate from oxid of iron and alumina, filter, wash, ignite, and weigh as calcium oxid or titrate with potas- sium permanganate. (e) SOLUBLE SULPHATE. Boil 1 giam with 20 cc of strong hydrochloric acid, dilute to 200 cc with hot water, boil, filter, wash, and determine the sulphate in the filtrate by precipi- tation with barium chlorid in the usual manner. Calculate to calcium sul- phate. If carbonates are present calculate the remaining calcium oxid to cal- cium carbonate (whiting) ; if carbonates are absent calculate the remainder as lime. (f) Loss ON IGNITION. Ignite 1 gram in a platinum crucible. The loss in weight represents water (both free and combined with gypsum and kaolin), carbon dioxid from whiting, and organic matter. Barytes should contain only small amounts of foreign matter, and there should be 95 per cent or more of barium sulphate present. 5. Lithopone. Lithopone should contain about 65 per cent of barium sulphate, the remainder being zinc sulphid with small amounts of zinc oxid and carbonate. This sub- stance is also called ponolith and Beckton white. (a) ANALYSIS OF PURE LITHOPONE (DEAWE). (1) MOISTURE. Heat 2 grams of the sample for two hours at 105° C. There should be less than 0.4 i>er cent of moisture. (2) INSOLUBLE AND TOTAL ZINC. Take 1 gram in a 200 cc beaker, add 10 cc of strong hydrochloric acid, mix, and add in small portions about a gram of potassium chlorate; then heat on the water bath until about half of the liquid is evaporated. Dilute with hot water. :i. At least one black on the market contains a large amount of .1 magnetic oxid of iron, which may be readily detected by the action of a nuiKiiet. The ash of ivory and bone blacks should contain a large amount of •alcium phosphate. COLORED PIGMENTS. 29 2. Graphite. Graphite is more of a gray pigment than a black. The finer grades are used as lubricants and in making lead pencils. All commercial graphite con- tains a considerable quantity of ash. For analysis, determine the moisture as usual, boil with hydrochloric acid, filter, dry, and weigh, then ignite ; it may be necessary to use oxygen to insure complete combustion; the loss is taken as carbon. If a previous test has shown the absence of carbonates, the treatment with hydrochloric acid may be omitted. Graphite used as a lubricant is some- times adulterated with soft coal. Good Ceylon graphite is practically unaffected by hot sulphuric acid and potassium sulphate, while soft carbon is completely decomposed. A fairly good quantitative determination of soft coal in graphite is made as follows: Determine the moisture and the ash in one portion, and calculate the total carbon by difference. Treat 1 gram in a Kjeldahl digestion flask with 30 cc of strong sulphuric acid and 2 grams of potassium sulphate. Heat for two hours at a temperature about the fuming point of the sulphuric acid, cool, add water, filter, wash, dry at 105° C., and weigh; then ignite until all of the carbon is burned; the loss here represents the carbon as graphite, and the difference between this determination and that of total carbon represents the soft carbon. COLORED PIGMENTS. 1. Yellow, Orange, and Bed Pigments. The most important yellow pigment is chrome yellow, which varies in shade from a light lemon yellow to a deep orange. The lighter shades generally con- tain lead sulphate as well as lead chromate, while the deep orange contains some basic lead chromate. Pure chrome yellow should contain only lead chromate, lead sulphate,0 and possibly some basic lead in the darker shades. Zinc yellow is zinc potassium chromate of a composition approximating the following for- mula : (a) CHBOME YELLOW. (1) MOISTURE. Dry 2 grams for two hours at 105° C. (2) INSOLUBLE IMPURITIES. Treat 1 gram with 20 cc of strong hydrochloric acid, cover and heat on a steam bath for fifteen minutes, add 100 cc of hot water, boil (the solution should be complete), filter, wash, ignite, and weigh the insoluble impurities. (3) LEAD. Evaporate the filtrate from the insoluble to about 2 cc, but not to dryness, add 100 cc of 95 per cent alcohol, stir, allow to stand for one hour, cool in ice water, filter on a Gooch crucible, wash with 95 per cent alcohol, dry for one hour at °Lead sulphate may be replaced by other insoluble lead compounds. A chrome yellow should be considered adulterated if it contains anything besides insoluble lead compounds. 30 TESTING MISCELLANEOUS SUPPLIES. 110° C., cool, and weigh as lead chlorid. Add 0.0085 gram to the weight of the lead chlorid to correct for its solubility in alcohol. Multiply the result thus obtained by 0.74478 to obtain the amount of lead present. (4) CHROMIUM. Evaporate the filtrate from the lead chlorid determination to dryness on a steam bath to expel the alcohol, add about 5 cc of hydrochloric acid, then about 50 cc of water, and warm till the solution is complete. Cool, transfer to a volumetric flask, and make up to a definite volume. To an aliquot add sodium peroxid in sufficient amount to render the solution alkaline and to oxidize the chromium to chromate, boil until all of the hydrogen peroxid is driven off, cool, make acid with sulphuric acid, add a measured excess of standard ferrous sulphate and titrate the excess of iron with standard potas- sium dichromate. (5) SULPHATE. In another aliquot of the filtrate from the lead chlorid determination, deter- mine the sulphate in the ordinary way by precipitation and weigh as barium sulphate.0 (6) TREATMENT IN THE PRESENCE OF SOLUBLE IMPURITIES. If the sample contains no soluble impurities, the preceding method will give a complete analysis. When the qualitative test shows calcium it may be neces- sary to make the analysis in a somewhat different manner, for gypsum would by this treatment be precipitated with the lead chlorid and render the deter- mination worthless. In such a case dissolve and separate the insoluble as above. Dilute the filtrate from the insoluble to 400 cc, cool, pass in a rapid stream of hydrogen sulphid until all of the lead is precipitated as lead sulphid, filter, wash with water containing some hydrogen sulphid, dissolve the lead sulphid in dilute nitric- acid (hot), add an excess of sulphuric acid, heat to fumes and complete the determination of lead as sulphate in the usual way. In the presence of chromium the solution must be very dilute and the hydrogen HI 1 1 'hid passed in rapidly to get a good separation of lead as sulphid. Trans- f«-r the filtrate from the lead sulphid to a 500 cc flask, add ammonium hydroxid until it is slightly alkaline, pass in hydrogen sulphid until all of the zinc present is precipitated (in case no zinc is present the addition of hydrogen sulphid is nnnet-essary) ; make up to volume, mix, allow to settle, draw off aliquots of the clear liquid, and determine the calcium (and, if necessary, the magnesium) in tin- usual way. All the zinc and chromium will still be in the 500 cc flask; dissolve in hydrochloric acid, make up to the mark, mix, take out an aliquot, and d. -tcrminr the /inc volumetrically by the ferrocyanid method as described under lead /.in«- whites. In another aliquot determine chromium as already dMcribed. Sulphate ran he determined by dissolving another portion in hydrochloric tilteriim from it the insoluble, and precipitating with barium chlorid from tile dilute hot filtrate; if the solution is kept dilute and hot, the barium sul- phate pn-< -ipitate will not he contaminated by lead. <; S(" •hroiurs contain harium phosphate instead of lead sulphate, in which case test for phosphoric arid, or barium sulphate, calcium sulphate, etc., may replace lead sulphate. Lead citrate is also sometimes present instead of lead sulphate. COLORED PIGMENTS. 31 (b) ZINC YELLOW. (1) MOISTURE. Dry 2 grains for two hours at 105° r. (2) INSOLUBLE IMPURITI l s. To 1 gram of the dry pigment add 25 cc of strong ammonium hydroxid, stir well, and dilute to about 100 cc with water, stir frequently for about thirty minutes, carefully breaking up all lumps, but do not heat. Filter and wash with dilute ammonium hydroxid and then with water, dry residue at 105° C. to constant weight, and weigh, then ignite and weigh again. If desired, dissolve the residue in hydrochloric acid and analyze in the usual manner. (3) SOLUBLE ZINC, CHROMIUM, AND POTASSIUM. Heat the filtrate from the determination of insoluble impurities to boiling in a graduated flask, pass in hydrogen sulphid until all of the chromate is reduced and the zinc is precipitated as zinc sulphid and chromium as chromium hydroxid. Fill to the mark, mix, allow to settle, draw out an aliquot of the clear liquid, and determine alkalis by adding sulphuric acid, evaporating to dryness, igniting, and weighing as mixed sulphates. If desired, potassium may be determined as potassium platinic chlorid (K2PtCl0). Add hydrochloric acid to the residue in the flask and determine zinc and chromium as in chrome yellow. (4) IMPURE SAMPLES. The preceding methods will serve only for samples which are pure or adul- terated only with substances insoluble in ammonium hydroxid. Very impure samples should be examined as described under chrome yellow, page 29. (c) RED LEAD AND ORANGE MINERAL. These pigments in the pure state are oxids of lead (approximately PbsOO, being probably mixtures of compounds of varying proportions of lead nionoxid and lead dioxid. Moisture, insoluble impurities, and total lead may be deter- mined by the methods given under chrome yellow; or, in the absence of alkaline earth metals, the lead may be determined as sulphate in nitric acid solution (dissolve by adding a few drops of hydrogen dioxid) by evaporating with an excess of sulphuric acid until fumes of sulphuric anhydrid are evolved. Deter- mine as sulphate in the usual way. The lead dioxid (PbO2) may be determined as follows: Weigh 0.5 gram of the very finely ground pigment into a 150 cc Erlenmeyer flask. Mix in a small beaker 15 grams of crystallized sodium acetate, 1.2 grams of potassium iodid, 5 cc of water, and 5 cc of 50 per cent acetic acid. Stir until all is liquid, pour into the Erlenmeyer flask containing the lead, and rub with a glass rod until all of the lead is dissolved ; add 15 cc of water, and titrate with tenth-normal sodium thiosulphate, using starch as indicator. A small amount of lead may escape solution at first, but when the titration is nearly complete this may be dissolved by stirring. The reagents should be mixed in the order given, and the titration should be carried out as soon as the lead Is in solution, as otherwise there is danger of loss of iodin. One cubic centimeter of tenth-normal sodium thiosulphate corresponds to 0.011945 gram of lead dioxid, or 0.034235 gram of red lead (Pb3O4). Test red lead for dye and water soluble; look for carbonate, nitrate, nitrite, and sulphate. 32 TESTING MISCELLANEOUS SUPPLIES. These colored lead pigments may have their color modified by the addition of organic coloring matters. As a general rule, such adulteration may be detected by adding 20 cc of 95 per cent alcohol to 2 grams of the pigment, heating to a boil, and allowing to settle. Pour off the alcohol, boil with water, and allow to settle, then use very dilute ammonium hydroxid. If either the alcohol, water, or ammonium hydroxid is colored, it indicates organic coloring matter. The quan- titative determination of such adulteration is difficult and must generally be estimated by difference. (d) VERMILION. True vermilion, or, as it is generally called, English vermilion, is snlphid of mercury. On account of its cost it is rarely used in paints, and is liable to gross adulteration. It should show no bleeding on boiling with alcohol and water and no free sulphur by extraction with carbon disulphid. A small quantity mixed with five or six times its weight of dry sodium carbonate and heated in a tube should show globules of mercury on the cooler portion of the tube. The best test for purity is the ash, which should be not more than one-half of 1 per cent. Make the determination in a porcelain dish or crucible, using 2 grams of the sample. Ash in a muffle or in a hood with a very good draft, as the mercury fumes are very poisonous. It is seldom necessary to make a determination of the mercury ; but if this is required, it may be determined by mixing 0.2 gram of the vermilion with 0.1 gram of very fine iron filings, or better " iron by hy- drogen." Mix in a porcelain crucible and cover with a layer 10 mm thick of the iron filings, place the crucible in a hole in an asbestos sheet so that it goes about half way through, cover with a weighed, well fitting, gold lid which is hollow at the top,*fill this cavity with water, heat the crucible for fifteen minutes with a small fiame, keep the cover filled with water, cool, remove the cover, dry for three minutes at 100° C., and thirty minutes in a desiccator, and weigh. The increase in weight is due to mercury. The mercury can be driven off from the gold by heating to about 450° C. A silver lid may be used, but gold is much better. Another method is to place in the closed end of a combustion tube, 45 cm long and 10 to 15 mm in diameter, a layer of 25 to 50 mm of roughly pulver- ized magnesite, then a mixture of 10 to 15 grams of the vermilion with four «.r I'm* times its weight of lime, followed by 5 cm of lime, and plug the tube with asbestos. Draw out the end of the tube and bend it over at an angle of about 00°. Tap the tube so as to make a channel along the top, and place it in n combustion furnace with the bent neck down, resting with its end a little below some water in a small flask or beaker. Heat first the lime layer, and carry the heat back to the mixture of lime and pigment. When all the mer- « -iiry has been driven off, heat the magnesite, and the evolved carbon dioxid will drive out the last of the mercury vapors. Collect the mercury in a globule, wash, dry, and weigh. <;«-tniin,. \.-niiilii>n is at the present time little used in paints. Organic lakes are used for most of the brilliant red, scarlet, and vermilion shades. These organic coloring matters are sometimes precipitated on red lead, orange mint-nil. <>r /in,- nxid: but as a usual thing the base is barytes, whiting, or china clay. I'lininiinmllin red, a compound of diazotized paranitranilin and beta-naphthol is largely employed; but a number of colors may be used. To test for red colors in such a lake the following method from Hall ° may be of °The Chemistry of Paints and Paint Vehicles, p. 29. COLORED PIGMENTS. 33 v;i!nc, though other colors may be employed, which makes the table of only limited use. Determination of red colors in organic lake. Source of color. Reageut. Alizarin. Eosin. Para-nitrani- ' lin. Ortho-anisi- din. Scarlet (2R). Sulphuric acid, cone. Dark brown with reddish under- tone becoming Changes to yel- low; fluores- cent solution Purple; color returning on diluting. Purple: red on diluting. Darkened; lighter on diluting. light yellow on diluting. with excess of sodium hy- droxid. Hydrochloric acid, cone. Color becomes "muddy." Changes to vel- low; reddish Color slightly darkened; Dark purplish red; lighter Darkened; lighter on fluorescent solution with lighter on diluting. on diluting. diluting. • excess of so- dium hydrox- id. Sodium hy- d ro x i d , Dark reddish brown; little Little change; fluorescent Color changed to brownish Little change. Reddish solu- tion on di- cone. sol. change on di- solution on red; light luting. luting. diluting. red on dilut- Alcohol Insoluble.. Reddish fluo- rescent solu- ing. Slight yellow- ish, orange Slight reddish solution. tion. solution. Sodium hy- Purplish; dark Purple; color Little change. Color darker; d ro x i d , brown on dilut- returning on reddish solu- cone, and al- ing. diluting. tion on di- cohol. luting. It is well also to try the action of reducing and oxidizing agents such as stannous chlorid, ferric chlorid, etc. (See Zerr, Bestimmung von Teerfarb- stoffen in Farblacken ; also Schultz and Julius, A Systematic Survey of the Organic Coloring Matters.) Paranitranilin red is soluble in chloroform. It is also well to try fhe solvent action on different reds of sodium carbonate, etc. The amount of organic pig- ment present in such reds is generally very small, and when it can not be determined by ignition owing to the presence of lead, zinc, or carbonate, it is best determined by difference. 2. Oxid and Earth Pigments. Iron oxid is extensively used as a paint. The native oxid naturally varies very much in its composition. In general, however, only the poorer grades of native hematite are used as paints. Artificial iron oxid pigments, made by calcining copperas, may be practically pure ferric oxid. Umbers, ochers, and siennas are earthy substances containing iron and manganese oxids and more or, less organic matter. The methods of analysis are very much the same as those for iron ores. It is generally sufficient to determine moisture, loss on ignition, insoluble in hydro- chloric acid, ferric oxid, and manganese dioxid. If much organic matter is present, roast 2.5 grams in a porcelain dish at a low temperature until all organic matter is destroyed, add 25 cc of hydrochloric acid, cover with a watch glass, and heat on the steam bath for two hours ; then add 10 cc of sulphuric acid (1:1), evaporate, and heat until fumes of sulphur trioxid are evolved and all the hydrochloric acid is driven off. Cool, add 50 cc of water, boil until all of the iron sulphate is dissolved, filter into a 250 cc graduated flask, fill to the mark, mix, and take out a portion of 50 cc for iron determinations. For the 21694°— Bull. 109—12 5 34 TESTING MISCELLANEOUS SUPPLIES. determination of iron reduce with zinc, titrate with standard potassium per- manganate, and calculate to ferric oxid. For the determination of manganese transfer 100 cc of the solution (corresponding to 1 gram) to a 500 cc flask, add sodium hydroxid solution until nearly but not quite neutral. Then add an emulsion of zinc oxid in water in slight excess, shake until all of the iron is thrown down, fill to the mark, mix, let settle, filter through a dry paper, and use portions of 100 cc (corresponding to 0.2 gram) for the manganese deter- mination. Transfer to a 300 cc Erlenmeyer flask, heat to boiling, titrate with standard permanganate. The flask must be shaken during the titration and some experience is necessary to determine the end point, which is best seen by looking through the upper layer of liquid and observing when the pink tinge from the permanganate does not disappear on shaking. A standard potassium permanganate solution, the iron factor of which is known, may be used to de- termine manganese. The iron factor multiplied by 0.2952 gives the manganese factor. In some cases this method of attack will not separate all of the iron from the insoluble matter. In such a case the insoluble must be fused with a mixture of sodium and potassium carbonates, dissolved in water, evaporated with excess of sulphuric acid, filtered from the insoluble, and this solution added to the first one. Another method is as follows: Roast 5 grams of the powder, digest with 25 cc 'of hydrochloric acid, evaporate to dryness, moisten with hydrochloric acid, dissolve in water, filter, and wash the residue ; ignite the residue in a platinum crucible, add sulphuric acid and hydrofluoric acid, drive off the latter, and heat until copious fumes of sulphuric anhydrid come off. Add potassium hydrogen sulphate and fuse, dissolve in water, filter from any remaining insoluble (barium wilphate), unite the two solutions, make up to 500 cc, and use aliquots for the iron and manganese determinations. For the determination of iron place 100 cc in a flask, add about 3 grams of zinc, put a funnel into the neck of the flask, heat when the action slackens; if basic salts separate out add a few drops of hydrochloric acid. When all of the iron is reduced, add 30 cc of sulphuric acid (1:2), and as soon as all of the zinc is dissolved and the solution is cool, titrate with potassium permanganate. For the determination of manganese use 50 cc, evaporate to a very small bulk, add strong nitric acid and evaporate the hydrochloric acid; add 75 cc of strong nitric acid, which should be free from nitrous acid, and 5 grams of potassium chlorate, heat to boiling and boil fifteen minutes; then add 50 cc of nitric acid and more potassium chlorate. Boil until yellow fumes cease to come off, cool in ice water, filter on asbestos and wash with colorless, strong nitric acid; su« k dry and wash out the remaining nitric acid with water, transfer the precipitate and the asbestos to a beaker, add a measured excess of standard solution of ferrous sulphate in dilute sulphuric acid, stir until all of the manganese dloxid Is dissolved, and titrate the remaining ferrous sulphate with potassium permanganate. A ferrous solution of about the proper strength Is made by dissolving 10 grams of crystallized ferrous sulphate in 900 cc of water and 100 cc of sulphuric acid (specific gravity 1.84). This solution is titrated with standard potassium permanganate. The reaction taking place tin- manganese dloxid, acts on the ferrous sulphate is MnO2+2FeSO4-h •jii so«=MnSO«+Fea(SO4)«+H2O. Hence the iron value of permanganate multiplied by 0.491 gives the value in manganese. 3. Blue Pigments. There are two important blue pigments used in painting, Prussian blue and nltramarine. Of minor Importance and very little used in paints for anything but artists' colors are cobalt blue and smalt. COLORED PIGMENTS. 35 (a) PRUSSIAN BLUE. Under the name Prussian blue are included all ferrocyanid blues such as Antwerp blue, Chinese blue, TiirnmiU's blue, etc. These blues are all ferric ferrocyanids, ferrous ferricyanids, or double iron potassium salts of hydro- ferrocyanic or hydroferricyanic acids. The analysis of these blues, as is gen- erally the case with pigments, does not necessarily give results which can be used to grade samples, the strength and color tests being most important. Most text books say that Prussian blue is ferric ferrocyanid, Fe4[Fe(CN)e]8, but this substance is not known commercially. Commercial Prussian blue is a mixture of Williamson's blue, KFe[Fe(CN)6], with other iron-alkali cyanids and often with aluminum-iron cyanids, altogether a most complex substance.0 (1) MOISTURE. For the determination of moisture dry 2 grams for two hours at 100° C. blue should contain less than 7 per cent of moisture. (2) INSOLUBLE IMPURITIES. Dry Ignite 1 gram in a porcelain dish at a low temperature. The ignition should be carefully carried out at a temperature just high enough to decompose the last trace of blue, but not high enough to render the iron insoluble, or very difficultly soluble in hydrochloric acid. Cool, add 15 cc of hydrochloric acid, digest for an hour on the steam bath covered with a watch glass, evaporate to a sirup, add water, boil, filter from the insoluble, wash, ignite, weigh, and determine the nature of the insoluble, probably barium sulphate. In pure Prus- sian blue solution should be complete. (3) TOTAL IRON. Decompose as in the foregoing determination, reduce, and determine the iron in the ordinary way. There should not be less than 30 per cent, calculated on the dry pigment. (4) TOTAL NITROGEN. Determine on a 1 gram sample by the official Gunning method, digesting for three hours.6 The percentage of Prussian blue may be obtained with suffi- cient accuracy for commercial purposes by multiplying the percentage of nitro- gen by 4.4 and by multiplying the percentage of iron by 3.03. Eight samples of pure Prussian blue examined by Parry and Coste gave the mean results from which these factors are calculated. The following table shows the accu- racy with which these factors give the percentage of Prussian blue in the eight samples : Parry and Coste's determination of the percentage of Prussian blue in the dry matter of eight samples. Factors. No. 1. No. 2. No. 3. No. 4. No. 5. No. 6. No. 7. No. 8. Nitroeren X 4 4 94 69 100 93 107 58 94 91 98 95 99 83 100 80 102 52 Iron X 3.03 94.63 101.02 109.71 97.11 94.84 98.99 102.47 100 26 a Parry and Coste, The Analyst, 1896, 21 : 225-230. 6 U. S. Dept. Agr., Bureau of Chemistry, Bui. 107, Revised, p. 7. 36 TESTING MISCELLANEOUS SUPPLIES. (5) OTHER DETERMINATIONS. It is seldom worth while to make any further determinations. If desired, however, the iron and aluminum may be precipitated as hydrates by ammonium hydroxid and weighed together as oxids, and the aluminum obtained by differ- ence after determining the iron volumetrically and calculating to ferric oxid. The filtrate from the iron oxid and alumina precipitate may be made up to a definite volume and one aliquot taken for the determination of sulphate and another evaporated with sulphuric acid, ignited, and weighed. Determine whether the alkali is sodium or potassium and subtract the alkali metal corre- sponding to the sulphate (SO4) found. The remainder is double alkali iron ferrocyanid. Well-washed blues should be neutral in reaction. The red shade may be due to organic red. Test the solubility in alcohol, etc. (b) ULTRAMARINE. Ultramarine is a compound of unknown constitution made by heating clay, soda, sulphur, and charcoal together. It appears to be complex silicate and sul- phid of aluminum and sodium. Ultramarine may be readily distinguished from Prussian blue by being dissolved in hydrochloric acid with the evolu- tion of hydrogen sulphid. It can not be mixed with lead pigments on account of the sulphid present. Good ultramarine gives little tarnish on polished cop- per. An analysis is of little value, but if desired may be conducted as follows : (1) MOISTURE. Dry 2 grams for two hours at 105° C. (2) SILICA. Treat 1 gram with 20 cc of hydrochloric acid (1 : 1) in a covered dish, heat until decomposed, remove the cover and evaporate to dryness on a water bath, moisten with hydrochloric acid and again heat to dryness, add 1 cc of hydro- chloric acid, let stand five minutes, add hot water, filter, and wash the insoluble. If great accuracy is desired, evaporate the filtrate to dryness, take up with hydrochloric acid and water, filter, and add the precipitate to the main insolu- ble. Ignite the insoluble, weigh as silicon dioxid, and test the purity with hydrofluoric acid. (3) ALUMINUM. Determine the aluminum in the filtrate from the silicon dioxid by precipitat- ing with ammonium hydroxid, igniting, and weighing as aluminum oxid. (4) SODA. Add sulphuric acid to the filtrate from aluminum hydroxid, evaporate to drynees, Ignite, and weigh as sodium sulphate (Na2SCM. Calculate to sodium monoxid (NaaO). If calcium is present it should be precipitated with am- monium oxalate In the filtrate from aluminum hydroxid, ignited, and weighed as calcium oxid and the soda determined in the filtrate from the calcium oxalate, as described previously. (5) SULPHUR. Mix 1 prnm of ultramarine with 4 grams of sodium carbonate and 4 grams of sodium perozid in a nickel crucible, cover with about a gram of sodium car- COLORED PIGMENTS. 37 bonate, fuse, using an aluminum or asbestos shield to prevent the sulphur from being taken up from the gas. Dissolve the fused mass in water, making acid with hydrochloric acid ; there is no insoluble residue. Precipitate with barium chlorid and determine total sulphur by weighing the barium sulphate. (6) SULPHUR AS SULPHATE. Weigh 1 gram in a beaker, add water, then hydrochloric acid, boil until the hydrogen sulphid is expelled, then add barium chlorid, and determine the sul- phur as usual by weighing the barium sulphate. (7) SULPHUR AS SULPHID. Subtract sulphur as sulphate from the total sulphur. Pure ultramarine will generally have about the following composition : Per cent. Silicon dioxid (SiO2) 39 to 43 Alumina (A12O3) 24 to 30 Sodium monoxid (Na2O) 17 to 21 Sulphur (S) 10 to 16 4. Green Pigments. The most important green pigment is chrome green, which is a mixture of Prussian blue and lead chromate. A green made of ultramarine and zinc yellow is occasionally encountered. If absolute permanence is required chro- mium oxid (Cr2O3) is sometimes used, but this pigment is not common. The very brilliant copper arsenic compounds (Paris green, etc.) are little used in paints.0 (a) CHROME GREEN. (1) GENERAL DISCUSSION. A pure chrome green should contain only Prussian blue, lead chromate, and lead sulphate. Any other substances should be considered adulterants. Greens made by precipitating the ingredients together are superior to those made by mixing the blue and yellow after they are precipitated separately. To dis- tinguish the two, rub up the pigment in oil and allow to stand. A badly made green will show blue, but one properly made by precipitation will not. A micro- scopic examination of the dry pigment should also be made. A badly made green will show yellow and blue particles, as well as green. A well-made green will show green and some blue particles, but no yellow. As different grades of lead chromate vary greatly in shade, and as the same is true to a somewhat less extent with Prussian blue, actual determinations of the percentages of Prussian blue, lead chromate, and lead sulphate are of little real value, and the analysis of chrome green is made with difficulty. It is, however, important to determine impurities, which may be barium sulphate, silica and silicates (clay), or calcium and magnesium compounds. (2) IMPURITIES. To examine for impurities proceed as follows: Place 1 gram in a small cas- serole, add 10 cc of water and sodium peroxid, a little at a time (about 1 gram 0 For methods of analysis see U. S. Dept. Agr., Bureau of Chemistry, Bui. 107, Revised, pp. 25-30; also, Lunge, Chemisch-techuische Untersuchungsmethoden, 2:807-810. 38 TESTING MISCELLANEOUS SUPPLIES. in all), to oxidize the Prussian blue; add about 25 cc more of water and boil for some time, so as to decompose all of the hydrogen peroxid. Acidify with hydrochloric acid, boil, filter, and wash. Transfer the residue again to the casserole and boil with a solution of ammonium chlorid (slightly alkaline) to dissolve the last trace of lead ; filter, wash, ignite, weigh the insoluble impuri- ties, and determine their nature. They are probably barytes or silica. Heat another 1 gram portion very gently in a casserole or small porcelain dish, and remove from the flame as soon as it begins to glow. The heating should be carried out very carefully, the object being to decompose the Prussian blue but not to heat it enough to render the iron difficultly soluble. Allow to cool, add hydrochloric acid, and digest covered on the steam bath for an hour. As a general rule this treatment brings everything except insoluble impurities into solution. Cool, dilute to 300 cc, and pass in a rapid stream of hydrogen sulphid until all of the lead is precipitated; filter, wash with water containing some hydrogen sulphid, boil off the hydrogen sulphid from the filtrate, add bromin water to oxidize the iron, add ammonium hydroxid in slight excess, filter, dissolve the precipitate in hydrochloric acid, reprecipitate with ammonium hydroxid, filter, wash, unite the filtrates, and determine calcium and magnesium in the usual way. (3) QUANTITATIVE DETERMINATION OF YELLOW AND BLUE. The examination for impurities is usually sufficient, but if a quantitative determination of yellow and blue be desired, the sulphate may be determined in the united filtrate from the insoluble impurities by precipitating from hydro- chloric acid with barium chlorid. The ferric ferricyanid present will not cause any difficulty if the barium sulphate is filtered off soon after precipitating; but if it is allowed to stand a long time a reduction is likely to take place with the re-formation of some Prussian blue. Determine the lead in the second portion by dissolving the lead sulphid in nitric acid, boiling the residues with ammonium acetate, uniting the filtrates, adding sulphuric acid, heating to fumes, and completing the determination of !••; id as sulphate in the usual manner. For iron and chromium dissolve the ammonia precipitate in hydrochloric acid, make up to definite volume, deter- mine the iron in an aliquot by reduction and titration with potassium perman- ganate or potassium dichromate. Determine the chromium in another aliquot a< described under chrome yellow. (4) MISCELLANEOUS DETERMINATIONS. If aluminum is present, precipitate another aliquot with ammonium hydroxid, wash. dry. i-nite. and weigh. From the iron and chromium determinations calculate the combined weight. of ferric oxid and chromium oxid and obtain aluminum «»\'n! by difference. Determine the moisture and the nitrogen as i ibed under Prussian blue. If (he presence of ultramarine is suspected, test with hydrochloric acid, and if hydrogen sulphid is evolved, ultramarine is present. (b) ULTBAMARINK AND ZINC YELLOW GREEN. (1) GENERAL DISCUS'SION. No lead compounds should be in this green, since with lead and ultramarine the | .Mint will soon be blackened by the formation of lead sulphid. Owing to the widely varying composition of ultramarine, it is not possible to make a MIXED PAINTS. 39 quantitative determinat ion of the blue, but its presenee ean lie detected by means of hydrochloric acid, and the combined peiventa-e of ultramarine and insoluble impurities thus be determined. (2) DETERMINATION OF YELLOW. Treat 1 gram of the dry powder with 25 cc strong ammonium hydroxid, stir well, and dilute to about 100 cc with water. Stir frequently for al.oni ihiny minutes, carefully breaking up all lumps, but do not heat. Filter, wash with dilute ammonium hydroxid, and then with water; dry the residue at 105° C. to constant weight, and weigh. The loss in weight gives the amount of zinc yellow if no other soluble compounds are present. The filtrate may be tested as in the examination of zinc yellow. (3) ULTBAMABINE AND INSOLUBLE IMPURITIES. The insoluble material remaining from the determination of yellow (2) may contain both ultramarine and other insoluble substances, but its color strength can be determined as compared with a standard pure ultramarine, and the result calculated to percentage based on a standard ultramarine. MIXED PAINTS. Single pigments are generally sold either dry or ground in oil alone; rarely do they contain thimiers or driers. Ready-mixed paints, however, contain, in addition to oil, turpentine or gasoline and usually some drier. More compli- cated yet are enamel paints, which are pigments ground in oil varnish, and con- tain, beside the substances in the vehicle of mixed paints, varnish gums. The can of mixed paint should not be shaken, but should be weighed as a whole, opened, and as much of the clear vehicle removed as possible ; this should be weighed and kept in a stoppered flask (use cork, not rubber, stopper). After removing the clear vehicle mix the contents of the can thoroughly and transfer to another jar. Wipe and weigh the can. These figures will give the neces- sary data for calculating the analytical results back to the original material. 1. Extraction of the Vehicle. Weigh from 15 to 35 grams of the thoroughly mixed pigment in a tall, nar- row Erlenmeyer flask of 300 cc capacity. Add about 10 cc of gasoline, mix thoroughly, then add 150 cc of gasoline, mix well, and stopper with a cork; allow to stand until the pigment has settled, and decant the liquid into a beaker; some pigment will frequently go over with the gasoline, so it is well to allow this to settle and decant into a second beaker. Repeat the treatment with gaso- line; after the first treatment settling is generally much more rapid. Decant the gasoline as completely as possible the second time. Then add 150 cc of benzol (C6Ha), shake, and allow to settle. Decant the benzol and treat in same way with 150 cc of ether. (If a large number of samples are examined at the same time this method is as rapid as any ; but when working with a small num- ber of samples much time is saved by extracting in tubes and throwing down the pigment by whirling in a centrifuge.) This method of extraction is often more satisfactory than any method using a continuous extraction apparatus, for frequently the pigments can not be held by extraction thimbles. This method of treatment will generally give an almost complete separation of the vehicle, but in some enamel paints it is well to follow the gasoline treatment by a treatment with turpentine, and then remove the turpentine with gasoline 40 TESTING MISCELLANEOUS SUPPLIES. before treating with benzol and ether. Some paints are very difficult to settle when treated with a single solvent, and mixtures of solvents must then be tried. Ma in ilia nil (Select Methods of Paint Analysis) recommends 60 volumes of benzol and 40 volumes of 95 per cent wood alcohol. No system of extraction will remove absolutely all the vehicle, the insoluble portion being probably metallic soaps or linoxyn. After removing all of the soluble vehicle, dry the pigment, first at a low temperature in a gentle current of air, and then at 105° C. ; weigh, and from the loss in weight calculate the percentage of vehicle and pigment; then from the total weight of mixed paint and the weight of clear vehicle drawn off calculate the percentage of vehicle and pigment in the original paint. 2. Analysis of Pigment. Follow the methods for dry colors. In general determine the amount in- soluble in hydrochloric acid and the nature of the insoluble, and make an analy- sis of the soluble portion. In the absence of carbon and hydrated silicates (clay, etc.), a fairly close determination of unextracted vehicle may be made by drying the insoluble at 105° C., weighing, and then igniting and weighing. This loss is called combined water and carbonaceous matter, and in many cases may be safely considered as unextracted vehicle. 3. Analysis of Vehicle. Weigh 100 grams of the vehicle into a 500 cc flask, connect with a spray trap and a vertical condenser, and pass through it a current of steam, first heating the flask in an oil bath at 100° C. ; with the steam still passing through, raise the temperature of the bath to 130°. Catch the distillate in a small weighed separatory funnel ; continue distillation until the funnel contains 300 cc of water. Let the distillate stand until separated into two layers, then draw off the water and weigh the light oils. Examine as under turpentine, page 13. A slight error is caused by the solubility of turpentine in water ; this amounts .to about 0.3 to 0.4 cc for each 100 cc of water. When suflicient vehicle is available it is well to take another portion and distil, without steam, placing the flask in an air bath. Note the temperature of the bath at which distillation takes place, and continue the distillation at a temperature of 185° C. in the air bath. This method gives somewhat lower results on volatile oils than the first method, but the distillate can be tested for water-soluble volatile liquids which would be lost by the steam distillation, llnsaponiflable matter should be determined in this residue, or in some of the original vehicle. (See examination of linseed oil, p. 13.) The residue Is frequently too pasty for the determination of the specific gravity, which should be made on the original vehicle. Determine specific gravity with a plummet or hydrometer. The small amount of suspended pig- ment which frequently remains in the vehicle will introduce an error when a pyknometer is used. Determine the acid number of the residue and of the nal vehicle. Determine the percentage and the character of the ash either from the residue or the original vehicle. The iodin number is some- times a useful index ; but the constants of linseed oil which has been mixed with pigments, especially lead compounds, may be so much altered that an iodin number as low as 100 can not be taken as any proof of the presence of other fatty oils. Test for ronin na described under linseed oil. Another test for rosin may be nimlr l.y dissolving in carbon disulphid and adding a solution of stannic WRITING INKS. 41 bromid or chlorid in carbon disulphid. Use a wbite porcelain dish. If no water is present in either solution the presence of rosin is shown by the ap- pearance of a violet color. This test is not as delicate as the Liebermann- Storch test described under linseed oil. When much lead is present it may be lost in ashing, and for a correct de- termination of metals the following method is best: Place 25 grams of the vehicle in a 500 cc separatory funnel, dilute with 25 cc of a mixture of equal parts of gasoline and turpentine, add 50 cc of nitric acid (1 : 1), and let stand one hour, shaking every ten minutes. Then immerse the funnel in hot water, loosen the stopper, and shake gently. This drives off nearly all the gasoline. Remove from the hot water, let it separate, draw off the lower layer, and wash the upper oily layer four or five times with warm water. Add the washings to the main acid portion and determine the metals in the ordinary manner. If the paint is enamel paint, treat the vehicle as a varnish. (See page 18.) INKS. WRITING INKS. Almost without exception all black inks suitable for record purposes are iron-tannin inks. The determinations of an analytical character which are of value are as follows: 1. Specific Gravity. Determine with a pyknometer at 15.5° C. 2. Total Solids. Weigh 10 grams of ink in a flat-bottomed platinum or porcelain dish, evapor- ate to dryness on the water bath, and then heat in an oven at the temperature of boiling water for two hours ; cool in a desiccator and weigh. 3. Ash. Burn the residue from the determination of total solids at a low temperature, preferably in a muffle. In order to avoid loss by foaming the dish used should not be too small, not less than 50 cc capacity. 4. Iron. Transfer the ash to a small beaker and dissolve in 15 cc of hydrochloric acid with the addition of staunous chlorid at the temperature of the steam bath ; reduce by adding stannous chlorid, drop by drop, to the hot solution until the color is destroyed, and then add one or two drops in excess. Wash the reduced iron solution from the small beaker into a 600 cc beaker and dilute to about 250 cc with cold water, add all at once an excess of mercuric chlorid, stir, allow to stand a minute, and titrate with standard potassium dichromate solution. 5. Sulphuric Anhydrid (SO8.) Determine both iron and sulphuric anhydrid in the same sample. Place from 10 to 15 grams of ink in a platinum dish, add 1 to 1.5 grams of sodiuin carbonate previously dissolved in water, evaporate to dryuess, ash, extract with water, add bromin water to the extract, boil, render acid with hydro- chloric acid, boil off the bromin, and determine the sulphuric anhydrid (SO«) 42 TESTING MISCELLANEOUS SUPPLIES. by precipitation with barium chlorid. Dissolve the insoluble residue in hydro- chloric acid and determine iron as described in the preceding section. Or, in case the small amount of platinum (which always goes into solution when iron oxid is dissolved by hydrochloric acid in platinum) causes trouble with the potassium dichromate method, add sulphuric acid, evaporate to fumes, dilute, reduce with zinc, and titrate with potassium permanganate. 6. Keeping in Ink Wells. Allow the bottle of ink to stand perfectly still at room temperature for three days. Carefully remove the stopper without shaking the bottle and draw out about 50 cc with a pipette from the middle of the bottle. Filter this through a ose to sunlight. Expose other strips to the weather uncovered mid keep some strips in ditfnsed daylight in an atmosphere free from laboratory fumes for eight days. I'sc some of these last strips for the tests WRITING INKS. 43 with reagents and preserve some for final comparisons. If the ink is a copying Ink, make press copies of some of the strips as soon as they are thoroughly dry and treat the press copies and the copied original as in the other cases. In testing with reagents, immerse the strips in the reagents and observe the effect at the end of fifteen minutes, of one hour, and of twenty-four hours. Test, using the following reagents, though others may also be employed : Water ; 95 per cent alcohol ; 90 volumes of water and 10 volumes of ammonium hydroxid (0.90 sp. gr.) ; 90 volumes of 95 per cent alcohol and 10 volumes of ammonium hydroxid (0.90 sp. gr.) ; 2 per cent hydrochloric acid; 2 per cent sodium hydroxid; bleaching powder solution, N/200 available chlorin. The tests used are relative, and it is well to have a standard ink for com- parison. Such a standard may be prepared by the following formula: 23.4 grams of pure dry tannic acid ; 7.7 grams of pure crystallized gallic acid ; 30.0 grams of pure crystallized ferrous sulphate ; 10.0 grams of gum arable ; dilute hydrochloric acid in sufficient amount to contain 2.5 grams of hydrochloric acid (25 grams dilute hydrochloric acid U. S. P.), and 1.0 gram of crystallized carbolic acid. All of these chemicals should be of U. S. P. quality, and in addition the purity of the tannin should be determined by the hide powder method.0 Dissolve the tannin and gallic acid together in about 50 cc of warm water, dissolve the ferrous sulphate in about 150 cc of cold water, dissolve the gum arabic in about 150 cc of warm water. Allow the warm solutions to cool, add the hydrochloric acid to the ferrous sulphate and immediately mix all the solu- tions and make up with distilled water to 1,000 cc. Mix thoroughly and allow to stand for at least four days at room temperature. Without shaking the bottle draw out the ink to be used in making comparisons. This standard ink will not be of a good color but will make a dirty gray green mark which will finally turn black. It may be colored to match the ink under examination by addition of a small amount of soluble dyes. 11. Bating. The rating of a number of samples of ink is entirely relative, and any sys- tem may be adopted. The exposure to sunlight is, of course, the most impor- tant test, and the following scheme answers very well for record inks. Rate the standard ink as follows : Exposure to sunlight 70 Exposure to reagents 10 Keeping quality, penetration, stickiness, fluidity, and action on steel pens - 15 Composition 5 Total 100 Other record inks are given values above or below the figures for the standard ink, as the judgment of the analyst may indicate, so the total for a very good ink may be over 100. Rate copying inks as record inks and make a separate rating on the copy, comparing one with another, and giving a weight of 50 per cent to clearness and depth of copy, and 50 per cent to resistance to reagents and sunlight. For °U. S. Dept Agr., Bureau of Chemistry, Bui. 107, Revised, p. .'iu. 44 TESTING MISCELLANEOUS SUPPLIES. further information consult Lunge,0 Schluttig and Neumann,6 Mitchell and Hepworth,0 and Munson.d The chromate logwood inks and the vanadium inks are not permanent. Among the inks examined by Munson no sample with a specific gravity below 1.035 was permanent, and with one exception no sample containing less than 0.60 per cent of iron. The exception, however, was a very permanent ink, in spite of the fact that it contained only 0.32 per cent of iron. CANCELING INKS HAVING AN OIL BASE.* The following methods have been devised for the purpose of ascertaining the suitability of canceling inks for the use of the Post-Office Department. Many of these methods will be found of assistance in passing upon the quality of stamping inks for miscellaneous uses. It is important that the ink used by the Post-Office Department for post- marking i>ossess in the highest possible degree certain properties. The ink, first of all, must produce an indelible cancellation ; that is, it must be relatively indelible as compared with the ink used for printing the postage stamps. The I>ostmark made with the ink must dry quickly in order that the mail matter may be handled immediately without any blurring or smearing of the post- mark. Both this property and the property of indelibility involve the question of the rate at which the ink penetrates or is absorbed by the fiber of the paper. A satisfactory ink does not harden or form a crust on the ink pad on exposure to the air. There must be no deposition of solid matter on the bottom of the vessel in which the ink is stored, and the pigments, on which the indelibility of the ink depends, if insoluble, must not settle out in such a way as to make it possible to pour off from the top of the container a portion of the ink which contains little or none of the insoluble pigment or pigments. The following methods have been found of value for the purpose of ascertaining the quality of a given sample of ink as well as the appropriateness of certain materials used for the manufacture of canceling inks. 1. Preparation and Care of the Sample. Since canceling inks contain more or less insoluble and volatile matter, spe- cial attention must be given to the preparation and care of the sample. It (i i ust be carefully mixed by shaking before each portion is removed for analysis. and the container must be left open no more than is absolutely necessary for the removal of ix>rtions of the ink. 2. Determination of Matter Volatile at Ordinary Temperatures. Place a carefully weighed quantity (between 5 and 5.2 grams) of the ink in a flat-bottomed aluminum dish 102 mm (4 inches) in diameter. Distribute the ink rompl.-tHv over the surface of the bottom of the dish by gently tilting 1h<« same. This quantity of ink should be sufficient to completely cover the bottom of the dish. Place the dish on a horizontal shelf or table where air will have free access to it and where it will be screened in such a way that no dust can fall into it. 0 ChemiHch-Technische Untersuchungs Methoden, 8: 762-793. 6 Die EisengalhiHtinten. c Inks, Composition and Manufacture. *Thi» Examination of Writing Inks, J. Auier. Chem. Soc., 190G, 28: 512. «Th<> work on • •.-nicelm- inks is a revision of the methods previously issued as Circular 12 of this Bureau. CANCELING INKS HAVING AN OIL BASE. 4f» Reweigh the dish at the end of 18 hours, iM hours, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 14 days, 21 days, and 2s days. Calculate the total per- centage loss of weight at the end of each period of drying. The loss of weight should be gradual and should not exceed 15 per cent during the first 7 days, nor 25 per cent during 28 days. This test shows the absence of highly volatile ingredients and the absence of an excess of matter volatile at the ordiiuiry room temperature. The constituents of a canceling ink should be such that the vola- tile matter will not exceed the above limits when the ink is exposed under the conditions named to a summer temperature of 80° F. and upward. 3. Determination of Relative Penetrating Power, (a) APPARATUS. (a) Homeopathic shell vials about 8 cm long and 2 to 2.5 cm in diameter. (6) Strips of white blotting paper, which for a given series of determinations should be cut from the same sheet and of exactly the same dimensions. A con- venient size is 12 mm wide and 25 cm long. (c) A pair of dividers with arms 15 cm long or longer. (d) A millimeter rule. (b) DETERMINATION. Place approximately 5 cc of the ink or other material to be tested in one of the " shell vials " described, and if several samples are to be tested arrange the vials in a row. Place 5 cc portions of distilled water in each of two of the vials, and put one of the vials containing water at each end of the row of vials con- taining samples to be tested. Proceeding from left to right, insert a strip of blotting paper in each of the vials, recording the exact time the paper was placed in each vial. The blotting paper should maintain a nearly upright position. The liquids gradually ascend the strips by capillarity; the strips, however, should be in such a position that the liquid does not ascend by capillarity be- tween the edges of the strips and the sides of the vials. At the expiration of exactly fifteen minutes from the time each strip is inserted in the vial, measure the height to which the liquid has ascended the strip of paper by means of a pair of dividers and record the distance in millimeters. Make a second set of readings at the end of forty-five minutes. After all of the measurements have been recorded reduce the results to the terms of the penetrating power of distilled water, taking the penetrating power of distilled water as 100. This is accomplished by dividing each result by the average of the results obtained for the distilled water contained in the vials and multiplying the quotient by 100. Ten samples may conveniently be tested at one time by working as described. The method gives good comparative results, and has been applied not only to canceling and other stamping inks, but to the liquids used for the manufacture of these inks. In observing the penetrating power of a given sample of ink it is important to remember that the value of a canceling or stamping ink depends upon its power to penetrate the paper during the first minute or fraction of a minute following its application to the paper. It is well, however, to keep the tests under observation for several hours, as information can thus be obtained in regard to the extent to which the coloring matter contained in the ink follows the liquid base of the ink as it passes through the paper. In some cases the coloring matters keep pace with the liquid portions of the ink; in others an uncolored band at the top of the portion of the paper which is wet with the ink shows that the coloring matter does not proceed through the paper as rapidly 46 TESTING MISCELLANEOUS SUPPLIES. as the base of the ink. This may or may not be an undesirable result. If the colorless band is due to a difference in the rate of penetration, it is undesirable. If it shows that the dye contained in the ink has an affinity for the fibers of the paper, it is evidence of a valuable quality. Additional information can be obtained from the penetration test by removing the strips of paper from the vials, cutting off the part of the paper which has actually been immersed in the ink, and treating the upper part successively with petroleum, ether, alcohol, and other solvents for removal of the constitu- ents of the ink soluble in these liquids. The extent to which the dye or dyes contained in the ink resist the action of these solvents and the extent to which the lampblack has passed up the strip of blotting paper are indices of the quality of the ink. An examination of the strips with the microscope will give valuable informa- tion regarding the rise of carbon in the paper and the affinity of the dye for the fiber of the paper. With many inks the carbon will not rise above the surface of the liquid, while with others it penetrates the paper to the same height as the dye. With this class of inks it is important that the base of the ink have the power to carry the carbon well into the fibers of the paper. To make the determination allow the strips to remain in position until the next day, remove, dry between blotters, and examine for a rise of dye or carbon. If either dye or carbon rises as far, or nearly as far, as the base of th» ink the rise is pronounced " satisfactory." Less than this is not acceptable except in the case of glycerin inks, which rarely give any rise of carbon. A small amount of rise in the latter inks must be accepted as satisfactory. To deter- mine whether the rise is carbon or simply dye, pick off particles of the paper and adhering ink at intervals on the strip and mount on slides with water or alcohol. Examine with the microscope, using low power; note size of the car- bon grains. The grains adhere to the outside of the paper fibers in clots, as a rule, but it is often difficult to distinguish them. 4. Sedimentation Test, (a) APPARATUS. (1) Glass-stoppered cylinders, graduated for 200 cc and fractions thereof, the distance between the bottom and the 200 cc mark being 25 cm (10 inches). If unobtainable, other cylinders may be. substituted, marks being placed at distances 25 cm (10 inches) and 16 mm (}§ inch) from the bottom. (2) A pair of dividers with arms 15 cm long or longer. (3) A millimeter rule. (4) Pipettes made from straight tubing (7 mm), at least 30 cm in length and having a capacity of from 10 to 15 cc. (b) DETERMINATION. By means of the special pipette introduce carefully, drop by drop, into one of the 200 cc cylinders, exactly 16 mm depth of the ink to be tested. The ink should be previously tested to determine a proper solvent for both base and dye. Alcohol is generally the solvent to use for rosin inks. It may be neces- sary to use other solvents, such as gasoline (boiling point 50° to 60° C.), ether, brnxol, etc. Dissolve the ink in the cylinder in the appropriate solvent and dilute up to the 200 cc mark, stopper, and shake thoroughly. Allow the cylin- der to stand, and record from time to time, by using the dividers and milli- meter rule, the height of the top of the layer of sediment which collects in the CANCELING INKS HAVING AN OIL BASE. 47 bottom of the cylinder, expressing results in millimeters. During the first hour observations should be made at intervals of fifteen minutes ; later, each hour for several hours successively, and then twice daily for a week to ten days. After the settling of the top of the layer of sediment has entirely ceased, the height of the sediment should equal or exceed 16 mm, the amount of ink taken for the test. The rate of sedimentation is an index of the state of division of the carbon, some inks showing no appreciable layer at the expiration of a ten-day test. In the case of some inks the supernatant liquid above the sediment is of such a dark color that there is difficulty in locating the top of the sediment, even when the cylinder is inspected by light reflected at various angles. In this event the use of a dark room with a light placed so as to give a strong ray through a small aperture will locate the top of the layer of sediment in all cases except when the ink contains a very large percentage of a dense dye. Frequently the layer may be located by holding an incandescent electric light back of the cylinder and noting where the carbon filament can not be seen. The test is somewhat crude and only approximate, but it serves to give an idea of the fineness and amount of carbon, and, as a rule, agrees fairly well with the carbon determinations. 5. Determination of Lampblack. Load a porcelain gooch with asbestos, using a felt about one-fourth of an inch thick. After washing the felt thoroughly with water to remove fine particles, finally wash with alcohol and ether, dry, and weigh. Weigh out about 5 grams of ink in a small beaker, dilute with a suitable solvent (alcohol is used in case of rosin-oil inks), transfer to the gooch and wash until all oil and soluble color is removed. Finally, wash with alcohol and ether, dry, and weigh. 6. Determination of Ash. The lampblacks prepared for the manufacture of canceling ink yield less than 0.5 per cent of ash when burned, and the coal-tar dyes employed should contain no mineral matter other than that which is an essential part of the molecules of the substances to which the tinctorial power of these dyes is due. For the determination of the ash, place 2 to 3 grams of the ink in a porcelain dish, which must be of such size as to avoid loss of ink due to the foaming which is likely -to attend the beginning of the incineration. Heat the dishes thus charged in a muffle at a low red heat, until all organic matter and uncom- bined carbon have been burned. Cool and weigh. If an excessive percentage of ash is found, the percentage of mineral matter contained in the alcohol extract should be determined by incineration of the residue obtained after evaporation of this extract. If either the total ash or the ash of the alcoholic extract is high, a qualitative examination should be made. 7. Resistance of Pigments and Dyes to Light and Reagents. It is necessary in the case of canceling inks, and important, if not necessary, in the case of many stamping inks, that the pigments and dyes employed in their manufacture be as resistant as possible to means which may be employed for the erasure of marks made with them on paper. Under this heading may be men- tioned also the importance of the use of dyes which possess considerable affinity for vegetable fibers. It is not practicable to enumerate the agents which should 48 TESTING MISCELLANEOUS SUPPLIES. be employed in experiments to ascertain the resistance of a given dye to erasure, as light, heat, and all of the solvents and reagents known to the chemist are available for the use of persons who might desire them for use in assisting them in making fraudulent erasures. For the purpose of canceling postage stamps, it is necessary that the cancel- ing marks be substantially indelible, because the inks used in printing many of the stamps are very resistant. Stamping inks used for other purposes, however, do not require absolute indelibility. In making the tests use several layers of blotting paper as a pad, pour on this a small quantity of the ink and distribute carefully ; see that all excess has been absorbed by the pad before using the stamp. Make a sufficient number of impressions at one time to suffice for all tests and leave some in reserve. Hav- ing made the impressions, arrange them in groups according to the color and kind of ink, and rank them according to the following scheme : 1. Of highest rank. 2. Very good, but not of the best. 3. Good. 4. Fair. 5. Poor. Having exposed the impressions to the various reagents, as described herein- after, each sample is again rated according to the effect of the reagents, as follows : 1. Unaffected. 2. Slightly affected. 3. Much affected. 4. Almost effaced. 5. Effaced. The wet reagents used are pure water, water with the addition of 10 per cent of strong ammonia (sp. gr. 0.90), pure alcohol (95 per cent), alcohol with the addition of 10 per cent of strong ammonium hydroxid, 2 per cent hydrochloric acid, and 200th-normal bleaching powder. Expose each impression in a small Erlenmeyer flask to about 50 cc of the reagent for 24 hours, noting its appear- ance at the end of 15 minutes, 1 hour, and 24 hours. Then rinse, dry, and rate. For the sunlight test expose impressions under glass for ten days to direct sun, rating at the end of the third, seventh, and tenth days. The tests with reagents are considered of less use than the other tests, and are not always applied RUBBER-STAMP INKS. 1. Preparation and Care of Sample. The precautions given in regard to the care of samples of inks made with an oil base should be observed. 2. Change of Weight on Exposure to Air. This determination should be conducted in the manner described for the determination of volatile matter in inks made with an oil base. Rubber- stamp inks, however, gain or lose in weight according to the constituents used in their manufacture and according to atmospheric conditions. A rubber- stamp ink should not, however, undergo very much greater changes in weight when exposed to the air under given conditions than diluted glycerin containing 75 per cent of glycerin and 25 per cent of water by volume. MATERIALS FOR STAMPING INKS, ETC. 49 3. Penetrating Power. This test should be conducted in the manner described for inks made with an oil base. 4. Sedimentation Test. This test should be conducted as described for inks made with an oil base, with the exception that the portions of ink should be diluted with water instead of with organic solvents. 5. Determination of Lampblack and Other Constituents. A scheme of analysis similar to that described for inks made with an oil base should be employed. Some experiments will be necessary in most cases to ascertain the proper solvent to be used in the case of each sample of ink to be examined. Alcohol, however, will generally be found to be satisfactory for rubber-stamp inks. This test is difficult and can not be carried out without a slight loss. To make the loss as small as possible use a very thick felt of asbes- tos and make the filtration continuous — never allow the gooch to run empty ; if this occurs it is generally better to repeat the test. 6. Besistance to Light and Reagents. The remarks made in regard to the investigation of the resistance of cancel- lations made with oil inks apply in general to canceling and other inks for use with rubber stamps. MATERIALS USED FOR THE MANUFACTURE OF CANCELING AND OTHER STAMPING INKS. 1. Volatility and Penetrating Power. The methods which have been described above will be found useful in de- termining the suitability of liquids for use as bases or constituents of bases of canceling and other stamping inks. 2. Sedimentation Test. A modification of the sedimentation test described may be employed with good results for the purpose of ascertaining the suitability of lampblack and other pigments for use in the manufacture of canceling and other stamping inks. The results, of course, are mainly of value for purposes of comparison. The conditions of the test may be modified to suit the purposes of the investi- gation and the character of the materials to be compared. The writer has ob- tained good results in the comparison of lampblacks and other blacks rich in uncombined carbon by the following method, which was so planned that the re- sults might be applied to stamping inks made with either a water-solution base or an oil base : Mix 0.5 gram of the black to be tested in a mortar with dilute glycerin (87.5 cc of glycerin diluted with water to 1 liter). Rinse the mixture into a 100 cc Nessler cylinder and dilute to the 100 cc mark, using the same dilute glycerin. After having prepared a series of tubes, each containing a portion of one of the blacks to be tested, close each tube with a cork and shake thoroughly each tube 50 TESTING MISCELLANEOUS SUPPLIES. successively, performing the operation as quickly as possible in order that the time of settling may be approximately the same in the case of each sample. Allow the cylinders to stand at rest in a place free from jar, and record from time to time the height of the sediment formed by the deposition of the blacks. When submitted to this test, a black which is suitable for the manufacture of a canceling or stamping ink should occupy a volume of not less than 25 cc when the sediment has stopped settling. 3. Ash, etc. Blacks, dyes, and other substances used for the manufacture of canceling and other stamping inks should be carefully examined to insure the absence of con- siderable percentages of substances which are not essential to the production of :HI ink of good quality. It can generally be assumed that the presence of con- siderable quantities of any substance which does not actually contribute to the desirable qualities of the ink will detract therefrom. Black pigments rich in carbon of high specific gravity due to the presence of a large percentage of ash are highly unsuitable for the manufacture of stamping inks. Only the concen- trated brands of coal-tar dyes should be used, unless the substances with which the less concentrated are diluted have been found to actually contribute to the working qualities of the ink to be produced. TYPEWRITER RIBBONS AND CARBON PAPERS. TYPEWBITER RIBBONS. The important things to be determined in a typewriter ribbon are length, width, typefilling, character of record (and copy, if it is a copy ribbon), life of ribbon, and resistance of record and copy to reagents and sunlight. 1. Length and Width. Length and width are generally specified and should be stated on the report, but the rating should be based entirely on quality. 2. Typefilling. Tako a sufficiently long section of the ribbon on a suitable typewriter (for wide ribbons use a Remington No. r» and for narrow ribbons use an Under- wood No. 5).° Carefully clean tin- small letter r, and, using the standard paper employed for ink testing, make records of this letter, striking the key with as uniform a stroke as possible. Continue writing the letter c until the loop fills so as to necessitate cleaning the type, or until 8,000 impressions have been made. Ribbons which give 8,000 clear impressions are given a maximum ratinir on typefilling, though the test may be carried as much further as may be desired. • 3. Character of the Becord. The general eha met eristics of the record may be determined by examining the sheets from the tyi^fllling test. Note color, clearness, whether the ribbon 0 These are the machines thai were used in the test, but any others may be employed for the purpose equally well, providing that the same machines are used throughout the tests, otherwise the results would have no comparative value. TYPEWRITER RIBBONS. 51 appears overinked or uuderinked, and tendency to smear. If the ribbon is a copy ribbon, take a press copy of the record and make the same observations on the copy and copied original as on the original record. 4. Life of Ribbon. For this test the special machine shown in figure 3 is used. This machine was designed and constructed by F. M. Allen, tool maker for the Bureau of Chemistry.0 The machine has ten types, including the vowels, both capitals and small letters, each type being mounted on the end of a plunger weighing 60 grams. The plunger is raised by a cam, and a uniform blow is obtained by allowing it to fall freely for 1.8 cm. The plungers fall, one after another, in a manner similar to that on a typewriting machine, except that each key strikes a different part of the ribbon. The cams are controlled by a geared mechanism which is operated by an electric motor connected with the testing machine proper by a belt. When a complete line has been struck the platen roll is moved forward one space automatically. The ribbon is fastened firmly at one end and the other end passes over a bar and has clamped to it a weight of 45 grams, thus insuring a uniform tension on all ribbons tested. The paper is fed from a long roll 14 cm wide, or it may be in strips of this width and MS. FIG. 3. — Machine for testing the life of typewriter ribbons. 41 cm long (5£ by 16 inches). It works under the ribbons as in a typewriter. The life of a ribbon is measured by the number of lines of writing which can be made before one or more holes appear in it or before one or more types fail to make a clear impression. 5. Resistance to Sunlight and Reagents. Cut up the sheets used in making the typefilling test, and subject the original, the copy, and the press-copied original to the same tests as to resistance to sunlight and reagents as were recommended for writing inks. Expose to sun- light for at least fourteen days, and continue the exposure as long as possible. 6. Rating1. Length and width are recorded, but are not included in the rating, which is based on quality only, and the ribbons are marked on a scale of 100. Type- filling is given a maximum rating of 5 and is marked as follows : 1,000 or fewer clear impressions of the small letter e, rate as 1 ; 1,000 to 3,000 such impres- sions, as 2; 3,000 to 5,000, as 3; 5,000 to 8,000, as 4; and 8,000 or more, as 5. Application for patent requested. 62 TESTING MISCELLANEOUS SUPPLIES. The character of the record is given a maximum rating of 20 ; life or durability of the ribbon, 10; resistance to reagents, 20; and resistance to sunlight, 45. Ribbons rated below 5 on typefilling or below 18 on character of record are rejected without further test. CARBON PAPERS. The practical testing of carbon papers consists in making a carbon copy on the special machine used for testing the life of ribbons, using a good quality of medium weight writing paper, such as is used in testing typewriter ribbons. The record is examined in the same manner as the ribbon record and is rated as follows : Character of record 35 Resistance to reagents 20 Resistance to sunlight 45 Total maximum 100 No test need generally be made of the life of a carbon paper, since when struck in the same place all papers are very soon exhausted, no practical dif- ference in this respect having been observed. Of course in actual use a number of impressions can be made, since the type seldom strikes in the same place. MISCELLANEOUS. BABBITT METAL (RAPID ANALYSIS)." 1. Copper. Weigh 1 gram of the alloy into a 250 cc beaker, add 20 cc of hydrochloric acid and 5 cc of water, heat, and complete the solution by adding nitric acid in small amounts; with most alloys solution can be effected in a very few minutes and without adding more than 1 or 2 cc of nitric acid. Evaporate off the acid on a st. •;! in bath. It is not necessary to carry to complete dryness, but practically all the acid should be driven off and the residue should be pasty. Add 25 cc of a solution made of 200 grams of tartaric acid and 260 grams of potassium hydroxid, the whole being made up to 500 cc with water. Heat on the steam bath until solution is completed, add 25 cc of water, boil, add 25 cc of a 0.2 per cent Invert sugar solution, boil for two minutes, filter through asbestos, wash the precipitate of cuprous ox id with water, dissolve in nitric acid, catching the copper solution in a 200 cc flask, and determine copper by any good volumetric method. Equally good results can be obtained by following Low's iodid method,* or Jamieson, Levy, and Wells's thiocyanate and iodate method.0 The results nre uniformly a little low. This error is not due to the volumetric methods employed, both of which give exceedingly accurate results, but to the fact that y »; i>er cent of the copper present is not precipitated as cuprous oxid. This loss is uniform, for if we add 6 per cent of the copper determined, the result will be the per cent of copper in the alloy. Tin- statement is frequently made that if a babbitt metal is decomposed by nitric acid, evaporated to dryness, taken up with nitric acid and filtered, copper 1 •• determined in the filtrate with an error of not more than one or two tenths of 1 p<-r cent. This is not the case, ns the error with an alloy containing r cent of cop^r will frequently he from o.r, to 0.7 per cent, while by th« difficulty may be experienced in decanting the water; this is some- times ;i voided by using a cage of fine wire gauze or a piece of bolting cloth to hold the ground glue. The temperature at which the water absorption is determined may be varied, but the figures obtained at one temperature MIV not- to |M» ennipiircMl with those at another. The determination can not be riirricat 4 or 5 cc of the oil in a test tube with a piece of caustic soda (solid) for fifteen minutes at a temperature of 230° to 250° C. Allow to cool. The presence of fatty oil is indicated by the gelatinizatiou, complete or partial, of the con- tents of the tube. 11. Quantitative Determination. In tho prosonoo of largo quantities of minernl oil, sjiponifiention with alco- holic potash takes M long time, since the mineral oil prevents the potash coming in '•"mart with tin- fjitty oil. Schreiber's method0 gives good results in a short time. Weigh 5 grams of the oil in a 200 cc Erlenmeyer flask, add 25 to 50 cc of half-normal alcoholic potash and sufficient benzol (CeH«) to dissolve the oil \vln-n warmed (generally *2r> cc is enough, but with heavy cylinder oils as much as 50 cc may be necessary ; in this case it is well to add 25 cc of neutral alcohol). Connect the flask with a 3-foot condenser and set it on the iron plate that forms the top of the steam bath, so that the steam will not strike it directly, and regu- 0 J. Amer. Chem. Soc., 1907, 2» : 74. PHENOLIC DISINFECTANT. 59 late the heat so that the condensing liquid will not be forced to the top of the condenser. In this way the contents of the flask can be boiled without ap- parently losing any of the solvent. Boil for thirty minutes. Cool, add phenolph- thalein, and titrate the excess of potassium hydroxid with half-normal sulphuric acid. On adding the acid the liquid separates into two layers and the change in color can be seen in the lower layer ; the titrating, however, must be conducted slowly. Subtract the cubic centimeters of sulphuric acid from the amount used on blanks and calculate the saponiflcation number. For all practical pur- poses 195 may be considered as the saponiflcation number of the fatty oils used in lubricating oils, and hence if S equals the determined saponification number, 100 S divided by 195 equals the percentage of fatty oil present. When the oil is very dark, or when soaps are present, it may be necessary to determine the hydrocarbon oils directly. For this purpose saponify as de- scribed, cool, add 25 cc of benzol, transfer to a separatory funnel, add 100 cc of water, shake, allow to settle, draw off the water layer, and wash the benzol three times with 100 cc of water; wash the soap solution with three portions of 50 cc of benzol, evaporate off the benzol solution, and weigh the nonvolatile hydrocarbon oils. 12. Acid Number. See linseed oil, page 12. 13. Bosin Oil. See linseed oil, page 13. PHENOLIC DISINFECTANT. The phenolic soap mixtures can generally be identified as they form an emulsion with water. They consist of phenolic soaps and hydrocarbon oils and are quite variable in composition. The following scheme for analysis is only a slight modification of that given by Allen.0 The only important modi- fication consists in the recovery of light hydrocarbons, which by Allen's process would be lost. 1. Volatile Hydrocarbons. Mix the sample thoroughly, weigh a portion of somewhat more than 100 grams in an Erlenmeyer flask with a funnel. Pour about 100 grams into a distilling flask of 150 to 200 cc capacity, using the funnel to prevent the disinfectant from getting on the neck of the flask, weigh back the Erlenmeyer flask and the funnel, and get the exact weight by difference. The delivery tube of the distilling flask should be bent so that it can connect with a small vertical con- denser; all connections are of cork. Heat the flask in a paraffin bath, slowly and carefully, not exceeding 130° C., for one and one-half hours, or until the substance becomes solid or ceases to give off volatile matter. The receiver is a small separatory funnel with a short tube, the tube of the condenser being con- nected with it by a cork having a groove cut into it for the free passage of air. The funnel is weighed empty, and again with the total volatile liquid; then the water is drawn off and the remaining liquid weighed as volatile hydrocarbon. 2. Pyridin. Dissolve the residue in the distilling flask in water, it is usually necessary to add 60 cc of water ; heat slowly to avoid foaming and leave the flask in the steam bath for about an hour. Wash the soapy solution into a 500 cc separa- a Commercial Organic Analysis, 3d ed., 2(2): 262. 60 TESTING MISCELLANEOUS SUPPLIES. tory funnel, add dilute sulphuric acid (1:3) until the liquid is distinctly acid to litmus, cool, shake with ether, drain off the acid layer and shake it twice with additional portions of ether, wash the ether twice with water, unite the acid water solutions, transfer to a large round-bottomed flask, and render dis- tinctly alkaline with sodium hydroxid, but avoid a great excess of alkali, as it will cause foaming. Use a spray trap and a large condenser. In all dis- tillations tie the stopper in the flasks and condensers, as they are liable to be blown out, and distil in a current of steam until all of the pyridin has passed over. This is usually a tedious operation, but in most cases the pyridin can be driven over in the first 600 or 700 cc. To ascertain whether all of the pyridin has come over, mix 10 cc of the distillate with 10 cc of a solution of 1 drop of half-normal sulphuric acid in 50 cc of water colored with methyl orange. The 10 cc of distillate should not change the color of the solution. Transfer the distillate to a 1,000 cc graduated flask, add an excess of half- normal sulphuric acid, fill to the mark, mix, and titrate aliquots back with fifth- normal sodium hydroxid using methyl orange. From the amount of half- normal sulphuric acid neutralized, calculate the percentage of pyridin (1 cc of half -normal sulphuric acid corresponds to 0.0395 grain of pyridin). 3. Heavy Hydrocarbons. The ethereal layer above the acid solution in which the pyridin is, contains hydrocarbons, phenols, and fatty and resin acids. Add a 25 per cent solution of sodium hydroxid in amount sufficient to combine with all the fatty and resin acids and phenols, shake, add water to dissolve the soaps, shake again and allow to settle. It will frequently be observed that there are not two layers, but three, and care must be taken to distinguish between the dark ethereal layer and the intermediate dark alkaline layer below it. Draw off the liquid underneath the ether, shake out twice again with ether, wash all ether por- tions three times with water, unite the ether solutions in a weighed Erlenmeyer flask, distil off the ether, dry for three-quarters of an hour on a steam bath, wipe out the moisture from the neck of the flask, and weigh as heavy hydro- carbons. Concordant results can not be obtained by drying to constant weight, hence it is always much better to leave the Erlenmeyer flask on the steam bath exactly three-quarters of an hour after distilling off the ether. 4. Phenols. Heat the alkaline liquid separated from the heavy hydrocarbons on the steam bath to drive off the ether, transfer it to a round-bottomed flask, make distinctly acid with sulphuric acid (1:3) and distil in a current of steam until no more oily drops pass over (do not use a spray trap) ; to the distillate add an excess of sodium hydroxid, evaporate to about 20 cc, transfer to a narrow graduated MopiK-n-d rviimi.T. acidify with sulphuric acid (1:3) keeping it cool, mix, allow to settle over night, and measure the layer of phenol. The volume in cubic centimeters multiplied by 1.05 gives the number of grams of phenols. 5. Fatty and Resin Acids. Cool thr n-siihn- in a distilling flask, extract with ether, distil off the ether. dry on u steam bath tlirc<- -quarters of an hour, cool, and weigh as fatty and resin acids. PIPE COVERING AND CEMENT. 61 6. Other Tests. Ash a 10 grain portion of the disinfectant and determine the alkalinity of the ash, and whether the alkali is soda or potash. In determining the relative values of disinfectants of this class, the sample should be submitted to a bacteriological test. In the absense of such a test the substance may be graded by the sum of the percentage of phenols and pyridiu. This is, of course, only a general approximation, as the hydrocarbons and soaps doubtless have some germicidal action. PIPE COVERING AND CEMENT. 1. Sampling. The insulating covering for steam pipes is generally made of a hydrated basic carbonate of magnesium and serpentine asbestos. Such material is difficult to sample, but by cutting out pieces of about 1 gram weight from 12 or more different places on the section of pipe covering, grinding in a mortar and thor- oughly mixing, a satisfactory sample can generally be obtained. Preserve this sample in a tightly stoppered bottle. The cement for pipe covering is of the same composition as the cover itself and is examined in the same way, but it is much easier to sample. 2. Moisture (Loss at 105° C.). Dry 1 gram of the prepared sample for two hours at 105° C., weigh, and calculate loss in weight as moisture. 3. Loss on Ignition. Ignite sample used for the moisture determination to constant weight. It is necessary to stir the material several times to insure heating all parts of the sample. 4. Asbestos. Treat 5 grams with 100 cc of 6 per cent acetic acid, heating on the steam bath to hasten solution. Filter on a Gooch crucible, wash, dry at 120° to 130° C., and weigh. This result is considered to represent asbestos. Then ignite to determine the water in the asbestos (about 14 per cent generally), which is deducted from the total, as it has already been included in the " Loss on Ignition." By this method all matter insoluble in acetic acid is assumed to be asbestos. If the presence of other insoluble material is suspected, filter on paper and examine the residue with the microscope or make an analysis of it. This, however, is seldom necessary. Serpentine asbestos is slightly soluble in acetic acid, but this error may be neglected. 5. Alumina and Iron Oxid. Add ammonium chlorid and ammonium hydroxid to the filtrate from the asbestos, filter, dissolve precipitate in hydrochloric acid, reprecipitate with am- monium hydroxid, filter, wash, ignite, and weigh as alumina and iron oxid. 6. Magnesia. Unite the filtrates from the alumina and iron oxid and make up to 500 cc. On 25 cc of this solution determine magnesia in the ordinary manner. If 62 TESTING MISCELLANEOUS SUPPLIES. the absence of all other metals has been proved the magnesium can be rapidly determined by adding sulphuric acid, evaporating to dryness, carefully heating to drive off ammonia salts and excess of sulphuric acid, and weighing as magnesium sulphate. Report the magnesium as magnesia (MgO) and calculate the equivalent of magnesium carbonate (MgCO3), and also of hydrated basic magnesium car- bonate [(MgCO3)4Mg(OH)25H2O]. The last is not a definite compound, the formula being only an approximation, but it approaches in composition the magnesium compound used in manufacturing these articles. BUBBEB. 1. Preparation of Sample. No general rule can be laid down for sampling all kinds of rubber goods, but pieces should be taken from as many parts of the article to be. analyzed as possible, and shredded by passing first through a meat chopper and then through a coffee mill. Hard rubber is best subdivided by rasping. 2. Ash. Determine ash on a 1 or 2 gram sample in a porcelain crucible, distilling off as much of the material as possible at the lowest possible temperature and completing the ashing at a very low red heat. To completely burn out all the carbon it will be necessary to break up the ash. In case the only mineral fillers are silica or barium sulphate, the percentage of ash will give a fairly good estimate of the amount of mineral filler used ; but with lead and zinc compounds there will always be more or less loss of metal and with carbonates there will also be loss. A qualitative analysis should be made of the ash. 3. Acetone Extract. Extract from 1 to 2 grams of the sample with acetone, using an extractor having no rubber or cork connections and one in which the hot vapors pass around the extraction tube so that the solvent in contact with the rubber is near its boiling point. With such an apparatus a five-hour extraction is usually Bufflcient. Evaporate the solvent and weigh the extract. As a check the ex- tracted rubber may be dried in a current of hydrogen or coal gas at 100° O., and weighed ; but it is better to determine the extract by direct weighing. The acetone extract contains the free sulphur in the rubber, the resins, and the fatty and mineral oils. 4. Free Sulphur. iMermine the free sulphur by adding about 25 cc of fuming nitric acid to the acetone extract, heating on the stoara bath ovor night, boiling to expel nitrons fumes, diluting t<> about 500 cc and determining sulphur in the ordi- nary way as barium sulphate. Deduct the free sulphur from the total acetone extract nnd report " sulphur-free acetone extract." 5. Alcoholic Potash Extract (Fatty Substitutes). Transfer the residue from the acetone extract from the extraction tube or eartridKe to a small beaker, add 20 cc of approximately normal alcoholic potash, and heat on the steam bath for thirty minutes. Dilute with 20 cc of water, RUBBER. 63 filter through Hie extraction tube (first decanting off flic <-Ic;ir supernatant liquid) ;ui(l wash with water. Add MII excess of sulphuric acid, lie;il for twenty minutes on the steam bath, transfer the liquid to a separately funnel, shake out with el her, evaporate the ether, and weigh the residue. This weight divided by 0.8 gives the weight of the alcoholic potash extract (fatly substitutes). 6. Chloroform Extract (Bitumens). Wash the residue from the alcoholic potash extract with alcohol and ether, dry in a current of hydrogen or coal gas, extract for three hours with chloro- form, evaporate the chloroform, and weigh the extract. A very small amount of chloroform extract is always obtained even when bitumens are not present, but this is generally not over 1 per cent. 7. Total Sulphur (Other than as Barium Sulphate).0 To 1 gram of the sample in a 100 cc nickel crucible, add about 2 cc of water and about 5 grams of anhydrous sodium carbonate. Mix thoroughly, using a nickel or platinum rod. Add sodium peroxid in about 0.5 gram portions, mixing thoroughly after each addition. Continue adding the peroxid until the mixture becomes granular and nearly dry, using about 5 or 6 grams in all. Heat the crucible with a very low alcohol or gasoline gas flame until the contents are fused, stirring all the time and carefully avoiding any overheating. (If the material ignites, the determination is ruined.) Allow to cool somewhat and cover the hardened mass with a layer of peroxid to a depth of about 0.5 cm. Heat carefully to complete the fusion and keep the material fused for ten minutes. Cool somewhat, place the crucible in a casserole, and carefully add water. After violent action has ceased, remove the crucible, washing it out with water, make slightly acid with hydrochloric acid, and boil to make sure that any barium sulphate, which may have been used as a filler and was partially decomposed by the fusion, is precipitated. Filter and determine sulphur in the filtrate by precipitation with barium chlorid in the usual manner. This method gives the total sulphur except that present as barium sulphate, and in case no mineral matter other than barium sulphate is present will con- tain no sulphur from the mineral matter. If the rubber contains sulphur other than as barium sulphate in the mineral matter, this will be included in the total sulphur. 8. Fillers. As a rule the ash may be considered as fillers, but if carbon or hydrated or carbonated mineral matter has been used the ash will be less than the fillers and consequently the gum reported by difference will be too high ; on the other hand, the free and combined sulphur is more or less fixed by the amount of mineral matter. An accurate method for determining fillers is much needed : but all methods proposed fail on certain kinds of rubber goods, and frequently the ash determination must be used. For most samples, except hard rubber, the following method is very satisfactory : Extract a fresh portion of rubber with acetone; transfer 1 gram of the extracted residue to a weighed beaker of 200 cc capacity ; add 50 cc of camphor oil, and heat on the steam bath until the rubber is dissolved (this usually takes a U. S. Dept. Agr., Bureau of Chemistry, Bui. 107, Revised, p. 23. 64 TESTING MISCELLANEOUS SUPPLIES. about two days). When the rubber is all dissolved (as indicated by the appear- ance of the insoluble matter), filter through a counterpoised paper. (If the papor proves unsatisfactory use a Gooch crucible with a thick layer of asbestos.) Dilute the filtrate with benzol and whirl in a centrifuge. In this way very finely divided mineral matter, which may pass through the filter, can frequently be determined. Wash the paper and beaker containing the mineral matter with hot benzol, gasoline, ether, and alcohol; then wash again with ether, dry at 105° C., and weigh. 9. Gum. Add together the sulphur-free acetone extract, the alcoholic potash extract, the chloroform extract, the total sulphur, and the fillers ; subtract the sum from 100 to obtain the percentage of gum present.0 SOAP. 1. Sampling. It is somewhat difficult to secure a fair sample from a cake of soap, as the moisture content of the outer layer may be very different from that of the in- terior of the cake. Weigh the whole cake, cut it in two in a diagonal direction, and cut from one of these fresh surfaces in thin layers a sufficient quantity for all determinations, taking care to cut entirely across so as to get a fair propor- tion of the outer and inner parts. When the cake is not too soft it is sometimes a good plan to run the whole cake through a meat chopper and reduce it all to fine particles. By running through the chopper several times a homogeneous sample is obtained. Preserve the sample in a tightly stoppered bottle and weigh the different portions required as soon as possible. 2. Moisture. The determination of moisture is at best unsatisfactory and it is better to include it In the " Undetermined." If a determination is required, heat 2 grams of the finely shaved soap for two hours in an oven at 105° C. The loss is con- sidered as water, though, of course, this is not accurate, as volatile oils may constitute part of the loss, and on the other hand, all of the water may not be driven off, and if the soap contains free alkali, carbon dioxid will be ab- sorbed from the air. This last source of error may be avoided by drying in vacuo. F.ihrion's method6 is much quicker and probably as exact. The author states that it gives results within 0.5 per cent of the truth. For this method have a supply of oleic acid which has been heated to 120° with stirring until oil of the water is driven off and preserved in a well-stoppered bottle. Weigh from 2 to 4 grams of the soap in a platinum dish, add three times its weight of oleic acid, and weigh agsiin. Heat the dish cautiously with a small flame until all of the water is driven off and nil of the soap is dissolved, being careful not to heat enough to decompose the acid. At the moment when all of the water is 'hi \cn off the soap forms a clear solution with the acid if there are no fillers present. Cool in a desiccator and weigh. This determination does not require more than fifteen minutes. 0 For further information on this subject consult the standard works on the subject, such as Weber on the Chemistry of India Rubber, Allen's Commercial Organic Analysis, and Lunge, Chemisch-technische Untersuchungsmethoden. 6ZtB. angew. Chem., 1906, 19:385. SOAP. 65 3. Insoluble Matter. Dissolve 5 grams of soap in hot water (use about 75 cc). Filter on a gooch or weighed paper and wash with hot water; dry at 105° C., weigh, and calculate the percentage of total insoluble. Ignite the residue and calculate the percentage of insoluble mineral matter. 4. Total Fatty Matter. To the filtrate from the insoluble matter add 40 cc of half-normal sulphuric acid, heat on the water bath until the fatty acids have collected in a clear layer on top, cool in ice water, remove the layer of fatty acids, and wash with water ; then heat again with water, cool, remove fatty layer, wash with ice water, dry with filter paper, unite the acid liquids, transfer to a separatory funnel, and shake out with two portions of 50 cc each of gasoline ; wash the gasoline twice with 20 cc of water. Evaporate off the gasoline, add the cake of fatty acids, etc., dry at 100°, and weigh as total fatty matter. This may represent fatty acids, uncombined fat, rosin and hydrocarbons. 5. Total Alkali. Heat the acid liquid obtained from the determination of total fatty matter to drive off traces of gasoline, cool, add methyl orange, and titrate the excess of acid with fifth-normal alkali. From this titration calculate the amount of acid neutralized by alkali in soap and figure to percentage of sodium monoxid. A more rapid method is to ash the soap, dissolve the ash in water, and titrate. This method gives good results in the absence of silica or other mineral fillers, but can not be used in their presence. 6. Free Alkali, (a) ALCOHOLIC METHOD. Treat the freshly cut surface of the soap with a few drops of an alcoholic solution of phenolphthalein ; if it does not turn red it may be assumed that free caustic alkali is absent. If free alkali is present, dissolve 2 grams of the soap in 100 cc of neutral alcohol (heat is necessary to complete the solution), filter frpm the undissolved sodium carbonate, etc.; wash with alcohol, add phenol- phthalein, titrate with standard acid, and calculate to per cent of free alkali as sodium hydroxid. Should the alcoholic solution be acid instead of alkaline, titrate with standard alkali and calculate the percentage of free fatty acid as oleic acid. Wash the portion insoluble in alcohol with water, add methyl orange to the washings, and titrate with half-normal sulphuric acid. Calculate to percentage of sodium monoxid present as carbonate, or possibly as borate or silicate. If borax is present boil off the carbon dioxid after neutralizing exactly to methyl orange ; cool, add mannite and phenolphthalein, and titrate the boric acid with standard alkali. (b) DEVINE'S METHOD. The alcohol method of determining free alkali is open to the objection that if the soap contains free alkali and free fat also, by the heating with alcohol the fat will be partly or completely saponified and the free alkali may not be 66 TESTING MISCELLANEOUS SUPPLIES. obtainable. Devine's method a avoids this danger. For this method three stand- ard solutions are used : 1. Tenth-normal hydrochloric acid. 2. Tenth-normal caustic soda or potash in alcohol. 3. Tenth-normal stearic acid in alcohol. Solution No. 1 is used to standardize No. 2; Nos. 2 and 3 should be titrated against one another warm, using phenolphthalein as indicator. Weigh 2 grams of soap (which needs no drying) into a round-bottomed flask of about 300 cc capacity, and pour 50 cc of alcohol upon it. Run in sufficient tenth-normal stearic acid from a burette to neutralize the free alkali in 2 grams of the soap, add some phenolphthalein, and stopper the flask with a cork stopper through which passes a glass tube about 30 inches long and of about £ inch internal diameter, having the lower end ground to a point on a grindstone. The puri>ose of the tube is to serve as a reflux condenser. Place the flask and con- tents on a steam bath and heat for thirty minutes. At the expiration of this time the solution should be quite clear and show no alkali with the phenol- phthalein. If the solution turns red during the boiling, showing that an insuffi- cient quantity of stearic acid has been added at first, add more of that solution until the color disappears, then several cubic centimeters in excess, and heat twenty minutes longer. Remove the flask from the bath and, after cooling it a few minutes, titrate with tenth-normal caustic soda. The difference between the number of cubic centimeters of stearic acid solution added and the number of cubic centimeters of caustic soda used to neutralize it is equivalent to the total free alkali present. While the first flask is heating, weigh in a similar flask 2 grams of soap, ;wld 50 cc of alcohol, and place on the steam bath. When the first test is finished, calculate roughly the total alkali, assuming the total quantity to be carbonate. Now add to the second flask sufficient 10 per cent barium chlorid solution to precipitate the alkali found,6 heat a few minutes, add phe- uolphthalein, and titrate with tenth-normal stearic acid. The titratiou must take place slowly and with thorough agitation of the liquid because the sodium or potassium hydroxid reacts with the barium chlorid added and forms sodium rliinrid and barium hydroxid. The latter is not very soluble in the alcoholic liquid, and sufficient time and pains must be taken to insure its complete neutralization by the stearic acid. Make a blank test on 50 cc of the alcohol, since this frequently contains carbon dioxid, and add the amount of tenth- iMirmul caustic soda necessary to neutralize the free acid in this quantity of alcohol to the reading of the stearic acid burette in the second test. This corrected reading gives the number of cubic centimeters of tenth-normal stearic arid used to neutralize the caustic alkali in 2 grams of soap. The difference between the total alkali found and the caustic soda will, of course, give tin- carbonate. For example: To 2 grams of soap and 15 cc of teuth- nonnal st-earic acid add :j.U <•«• «»f truth-normal caustic soda to neutralize. Then ]7t cc — 3.2 cc = 11.8 cc of tenth-normal stearic acid equivalent to total free alkali. To ncntrali/.r the caustic soda in the sample treated with barium chlorid 4.1 cc of tenth-normal stearic acid was required. Fifty cubic centimeters of the alcohol used required 0.2 cc of tenth-normal caustic soda, hence the total Is 4.o cc. Since it requires 1..", cc of tenth-normal stearic acid to neutralize free caustic alkali, 11.8 cc — 4.3 cc = 7.5 cc of tenth-normal stearic acid required t<> neutralize the carbonated alkali. ".I. Anier. ('hem. So<-., I'.MK). 2'J :»;'.):;. 6 One cubic centimeter of tenth normal stearic acid corresponds to 0.0122 gram BaCh 2HaO or 0.122 cc of a 10 per cent barium chlorid solution. SOAP. 67 One cubic centimeter of tenth-normal stearic acid = 0.004 gram of caustic soda or 0.0053 gram of sodium carbonate. These liguros calculated to percent- age would be: 0.86 per cent of caustic soda and 1.99 per cent of sodium car- bonate. It is to be noted that a rubber stopper can not be used in the flasks used for dissolving the soap on account of the sulphur in the rubber which decolor- izes an alcoholic solution of phenolphthalein. The method is applicable to all soaps that do not contain fillers which react with the standard solutions employed. 7. Unsaponified Matter. Dissolve 5 grams of soap in 50 cc of 50 per cent alcohol ; if any free fatty acids are present add just enough standard alkali to neutralize them and wash into a separatory funnel with 50 per cent alcohol. Extract with 100 cc of gaso- line (B. P. 50°- 60°). Wash the gasoline with water, evaporate, and weigh as unsaponified. This may consists of fat that has not been converted into soap or of hydrocarbon oils. 8. Rosin. If much unsaponified matter is present, prepare dry fatty and rosin acids from the portion from which the unsaponified matter has been removed. Dis- solve 3 grams of the dry mixed acids in 25 cc of absolute alcohol in a 100 cc stoppered flask, place the flask in cold water and shake; add 25 cc of absolute alcohol saturated with dry hydrochloric acid, shake occasionally, and let the action go on for twenty minutes ; then add 10 grams of dry granular zinc chlorid, shake, and allow to stand for twenty minutes. Pour the contents of the flask into 200 cc of water in a 500 cc beaker, and rinse out the flask with dilute alcohol. Put a small stick of metallic zinc in the beaker and boil off the alcohol. Cool, transfer to a separatory funnel, extract with gasoline, wash the gasoline until free from hydrochloric acid, distil off the gasoline, dissolve in neutral alcohol, and titrate with standard alkali using phenolphthalein as indicator. (One cubic centimeter of normal alkali corresponds to 0.346 gram of rosin.) The rosin may be determined gravimetrically by washing the gasoline extract with water. It is not necessary to wash absolutely free from acid. Then treat the extract in the funnel with a solution of 5 grams of potassium hydroxid and 5 cc of alcohol in 50 cc of water. Upon shaking, the rosin is rapidly saponified and the two layers separate well. Draw off the rosin soap, wash the gasoline once with water, add the washings to the soap solution, add an excess of acid, extract the rosin with gasoline, evaporate the gasoline, and weigh the rosin. 9. Silica and Silicates. Insoluble silicates, sand, etc., are present in the insoluble; but sodium silicate fillers will only show by forming pasty liquids. When sodium silicate is sus- pected, ash the soap (10 grams), add an excess of hydrochloric acid to the ash, evaporate to dryness; add hydrochloric acid and again evaporate to dryness; cool, moisten with hydrochloric acid, dissolve in water, filter, wash, evaporate the filtrate to dryness, and again take up with hydrochloric acid and water, filter, and wash. Unite the precipitates, ignite, and weigh as silicon dioxid; calculate to sodium silicate (Xa-Si^o). If metals other than the alkalis are suspected, the filtrate from the silica determination may be examined. 68 TESTING MISCELLANEOUS SUPPLIES. 10. Glycerin. Dissolve 20 to 25 grams of soap in hot water, add a slight excess of sulphuric acid, and heat on a water bath until the fatty acids separate in a clear layer. Remove the fatty acids and filter the acid solution into a graduated flask. Remove the chlorids and the soluble fatty acids by adding crystals of silver sulphate, cool, make up to the mark, mix, allow to settle, filter through dry paper, take an aliquot corresponding to 5 grams of soap, and determine by Hehner's bichromate method (see under Glycerin). This method can not be used when sugar is present as it also would reduce the bichromate. When sugar is present, remove the fatty acids as just described, neutralize an aliquot with milk of lime, evaporate to about 10 cc, add 2 grams of sand and milk of lime containing about 2 grams of calcium hydroxid, and evaporate almost to dryness. Treat the moist residue with 5 cc of 96 per cent alcohol, rub the whole mass into a paste, heat the mixture on a water bath stirring con- stantly, and decant the liquid into a 250 cc flask, wash the residue five or six times with small portions of alcohol, cool the contents of the flask to 15°, fill to the mark with 96 per cent alcohol, mix, and filter through a dry paper; evap- orate 200 cc of the filtrate to a sirupy consistency on a water bath, transfer to a stoppered cylinder with 20 cc of absolute alcohol, add three portions of 10 cc each of absolute ether, mixing after each addition; let stand until clear, pour off through a filter, and wash the cylinder on the filter with a mixture of two parts of absolute alcohol and three parts of absolute ether. Evaporate to a sirup, dry for one hour at the temperature of boiling water, weigh, ignite, and weigh again. The loss multiplied by 5/4 is the weight of the glycerin in the aliquot taken. Instead of weighing the glycerin it can be titrated after driving off the alcohol and ether. 11. Sugar. Dissolve 5 grams of soap in water, add an excess of hydrochloric acid, heat on a steam bath for thirty minutes, cool, separate the fatty acids, and make up to a definite volume. Add an aliquot, depending upon the amount of sugar supposed to be present, to 50 cc of a mixed Soxhlet solution, make the total volume 100 cc, heat to boiling, boil two minutes, filter on a weighed gooch, wash with water, then with alcohol and ether, dry at 100° for thirty minutes, cool, and weigh the cuprous oxid. Calculate the percentage of sugar as invert sugar according to Munson and Walker's tables.0 0 J. Amer. Chem. Soc., 1906, 28 : 663-686. O 1 RETURN CIRCULATION DEPARTMENT 202 Main Library LOAN PERIOD 1 HOME USE 2 3 4 5 6 ALL BOOKS AAAY BE RECALLED AFTER 7 DAYS 1 -month loans may be renewed by calling 642-3405 5-month loans may be recharged by bringing books to Circulation Desk Renewals and recharges may be made 4 days prior to due date DUE AS STAMPED BELOW Vf'O DEC 2 1981 BEG. cm. DEC 07-32- FORM NO DD6 40m 10 '77 UNIVERSITY OF CALIFORNIA, BERKELEY BERKELEY, CA 94720 U. C. BERKELEY LIBRARIES