Greiner's Powder consists of nitro-cellulose, nitro-benzol, graphite, and lampblack.

B.N. Powder.—This powder is of a light grey or drab colour, perfectly opaque, and rough to the touch. It consists of a mixture, nitro-cellulose and the nitrates of barium and potassium. Its composition is as follows:—

Insoluble nitro-cellulose 29.13 parts Soluble nitro-cellulose 41.31 " Barium nitrate 19.00 " Potassium nitrate 7.97 " Sodium carbonate 2.03 " Volatile matter 1.43 "

This powder is a modification of the Poudre B., or Vieille's powder invented for use in the Lebel rifle, and which consisted of a mixture of the nitro-celluloses with paraffin.

Von Foster's Powder contains nothing but pure gelatinised nitro- cellulose, together with a small quantity of carbonate of lime.

The German Troisdorf Powder is a mixture of gelatinised nitro-cellulose, with or without nitrates.

Maximite is the invention of Mr Hudson Maxim, and is a nitro-compound, the base being gun-cotton. The exact composition and method of manufacture are, however, kept secret. It is made by the Columbia Powder Manufacturing Company, of New York, and in two forms—one for use as a smokeless rifle powder, and the other for blasting purposes.

Wetteren Powder.—This powder was manufactured at the Royal Gunpowder Factory at Wetteren, and used in the Belgian service. Originally it was a mixture of nitro-glycerine and nitro-cellulose, with amyl acetate as solvent. Its composition has, however, been altered from time to time. One variety consists chiefly of nitro-cellulose, with amyl acetate as solvent. It is of a dark brown colour, and of the consistency of indiarubber. It is rolled into sheets and finally granulated.

Henrite is a nitro-cellulose powder.

Normal Powder.—The Swedish powder known as "Normal" Smokeless Powder, and manufactured by the Swedish Powder Manufacturing Company, of Landskrona, Sweden, and used for some years past in the Swiss Army, is made in four forms. For field guns of 8.4 calibre, it is used in the form of cylindrical grains of a yellow colour, of a diameter of .8 to .9 mm. and density of .790—about 840 grains of it go to one gun. For rifles, it is used in the form of grey squares, density .750, and 1 grm. equals about 1,014 grains. One hundred rounds of this powder, fired in eighteen minutes, raised the temperature of the gun barrel 284 deg. F. A nitro- glycerine powder, fired under the same conditions, gave a temperature of 464 deg. F.

This powder is said to keep well—a sample kept 3-1/2 years gave as good results as when first made—is easy to make, very stable, ignites easily, not very sensitive to shock or friction, is very light, &c. Eight hundred rounds fired from a heavy gun produced no injury to the interior of the weapon. Samples kept for eleven months in the moist atmosphere of a cellar, when fired gave a muzzle velocity of 1,450 ft. secs. and pressure of 1,312 atmospheres, and the moisture was found to have risen from 1.2 to 1.6 per cent. After twenty-three months in the damp it contained 2 per cent. moisture, gave a muzzle velocity of 1,478 ft. sees., and pressure of 1,356 atmospheres. In a 7.5 millimetre rifle, 13.8 grm. bullet, and charge of 2 grms., it gives a muzzle velocity of 2,035 ft. secs. and a pressure of 2,200 atmospheres. In the 8.4 cm. field-gun, with charge of 600 grms., and projectile of 6.7 kilogrammes, muzzle velocity was equal to 1,640 ft. secs. and pressure 1,750. A sample of the powder for use in the .303 M. rifle, lately analysed by the author, gave the following result:—

Gun-cotton 96.21 per cent. Soluble cotton 1.80 " Non-nitrated cotton trace. Resin and other matters 1.99 "

100.00

The various forms of powder invented and manufactured by Mr C.F. Hengst are chiefly composed of nitrated straw that has been finely pulped. The straw is treated first with acids and afterwards with alkalies, and the result is a firm fibrous substance which is granulated. It is claimed that this powder is entirely smokeless and flameless, that it does not foul the gun nor heat the barrel, and is at the same time 150 per cent. stronger than black powder.

The German "Troisdorf" powder consists of nitro-cellulose that has been gelatinised together with a nitrate. Kolf's powder is also gelatinised with nitro-cellulose. The powders invented by Mr E.J. Ryves contain nitro- glycerine, nitro-cotton, castor-oil, paper-pulp, and carbonate of magnesia. Maxim powder contains both soluble and insoluble nitro- cellulose, nitro-glycerine, and carbonate of soda. The smokeless powder made by the "Dynamite Actiengesellschaft Nobel" consists of nitro-starch 70 to 99 parts, and of di- or tri-nitro-benzene 1 to 30 parts.

An American wood powder, known as Bracket's Sporting Powder, consists of soluble and insoluble nitro-lignine, mixed with charred lignine, humus, and nitrate of soda. Mr F.H. Snyder, of New York, is the inventor of a shell powder known as the "Snyder Explosive," consisting of 94 per cent. nitro-glycerine, 6 per cent. of soluble nitro-cotton, and camphor, which is said to be safe in use. Experiments were made with it in a 6-inch rifled gun, fired at a target 220 yards away, composed of twelve 1-inch steel plates welded together, and backed with 12-inch and 14-inch oak beams, and weighing 20 tons. The shots entirely destroyed it. The charge of explosive used was 10 lbs. in each shell.

Comparative Tests of Black and Nitro Powders, from "American Field."— The results given in table below were obtained at the German Shooting Association's grounds at Coepenick, Berlin. Penetration was calculated by placing frames, each holding five cards of 1 millimetre in thickness (equals .03937 inch), and 3 inches apart, in a bee-line, at distances of 20 inches. Velocity, pattern, and penetration were taken at 40 yards from the muzzle of a 12-gauge choke-bore double-barrel gun. Gas pressure was taken by a special apparatus. All shells were loaded with 1-1/8 oz. of No. 3 shot, equal to 120 pellets, and the number given below represents the average number in the 30-inch pattern. The number of sheets passed through gives the average penetration. One atmosphere equals pressure equal to 1 kilogramme (2.2 lbs.) on the square centimetre, hence 1,000 atmospheres equal 2,200 lbs. on the square centimetre. The E.C., Schultze, and Walsrode powders were loaded in Elcy's special shells, 2-1/2 inches long. The averages were taken from a large number of shots, and the same series of shots fired under precisely the same conditions.

_____________ Gas Pressure. Velocity. Pattern. Penetration. ___ __ ___ __ ___ Atmospheres. Metres. Sheets. Fine-grained black powder, standard charge 514.2 280 78.6 = 66% 19.O Coarse-grained black powder, standard charge 473.4 281.4 78.2 = 65% 19.4 Schultze powder, 42 grains 921.0 290.0 64.2 = 54% 20.2 Schultze powder, 45 grains 1052.8 305.8 52.2 = 42% 20.6 E.G. smokeless, 42 grains 920.2 298.4 81.4 = 67% 18.8 Walsrode, 29 grains 586.4 280.6 83.0 = 69% 19.0 ___ __ ___ __ ___

Barometer, 760 mm. Thermometer, 30 deg. C. Hydrometer = 65. Wind, S.W.

Picric Powders.—The chief of these is Melinite, the composition of which is not known with certainty. It is believed to be melted picric acid together with gun-cotton dissolved in acetone or ether-alcohol. Walke gives the following proportions—30 parts of tri-nitro-cellulose dissolved in 45 parts of ether-alcohol (2 to 1), and 70 parts of fused and pulverised picric acid. The ether-alcohol mixture is allowed to evaporate spontaneously, and the resulting cake granulated. The French claim, however, that the original invention has been so modified and perfected that the melinite of to-day cannot be recognised in the earlier product. Melinite has a yellow colour, is almost without crystalline appearance, and when ignited by a flame or heated wire, it burns with a reddish-yellow flame, giving off copious volumes of black smoke. Melinite as at present used is said to be a perfectly safe explosive, both as regards manufacture, handling, and storage.

Lyddite,[A] the picric acid explosive used in the British service, is supposed to be identical with the original melinite, but its composition has not been made public.

[Footnote A: Schimose, the Japanese powder, is stated to be identical with Lyddite and Melinite (Chem. Centr., 1906, 1, 1196).]

Picrates are more often used than picric acid itself in powders. One of the best known is Brugere's Powder, which is a mixture of 54 parts of picrate of ammonia and 45 parts of saltpetre. It is stable and safe to manufacture. It has been used in the Chassepot rifle with good results, gives little smoke, and a small residue only of carbonate of potash.

The next in importance is Designolle's Powder, made at Bouchon, consisting of picrate of potash, saltpetre, and charcoal. It was made in three varieties, viz., for rifles, big guns, and torpedoes and shells. These powders are made much in the same way as gunpowder. The advantages claimed for them over gunpowder are, greater strength, comparative absence of smoke, and freedom from injurious action on the bores of guns.

Emmensite is the invention of Dr Stephen Emmens, of the United States. The Emmens "crystals" are produced by treating picric acid with fuming nitric acid of specific gravity of 1.52. The acid dissolves with the evolution of red fumes. The liquid, when cooled, deposits crystals, stated to be different to picric acid, and lustrous flakes. These flakes, when heated in water, separate into two new bodies. One of these enters into solution and forms crystals unlike the first, while the other body remains undissolved. The acid crystals are used mixed with a nitrate.

Emmensite has been subjected to experiment by the direction of the U.S. Secretary for War, and found satisfactory. A sample of Emmensite, in the form of a coarse powder, was first tried in a pistol, and proved superior in propelling power to ordinary gunpowder. When tested against explosive gelatine, it did very good work in shattering iron plates. It is claimed for this explosive that it enjoys the distinction of being the only high explosive which may be used both for firearms and blasting. This view is supported by the trials made by the American War Office authorities, and shows Emmensite to be a useful explosive both for blasting and as a smokeless powder. Its explosive power, as tested, is 283 tons per square inch, and its specific gravity is 1.8.

Abel proposed to use picric acid for filling shells. His _Picric Powder_ consisted of 3 parts of saltpetre, and 2 of picrate of ammonia. _Victorite_ consists of chlorate of potash, picric acid, and olive oil, and with occasionally some charcoal. It has the form of a coarse yellowish grey powder, and leaves an oily stain on paper, and it is very sensitive to friction and percussion. The composition is as follows:—KClO_{3} = 80 parts; picric acid, 110 parts; saltpetre, 10 parts; charcoal, 5 parts. It is not manufactured in England. _Tschiner's Powder_ is very similar to Victorite in composition, but contains resin. A list of the chief picric powders will be found in the late Colonel J.P. Cundill, R.A.'s "Dictionary of Explosives."



CHAPTER VII.

ANALYSIS OF EXPLOSIVES.

Kieselguhr Dynamite—Gelatine Compounds—Tonite—Cordite—Vaseline— Acetone—Scheme for Analysis of Explosives—Nitro-Cotton—Solubility Test— Non-Nitrated Cotton—Alkalinity—Ash and Inorganic Matter—Determination of Nitrogen—Lunge, Champion and Pellet's, Schultze-Tieman, and Kjeldahl's Methods—Celluloid—Picric Acid and Picrates—Resinous and Tarry Matters— Sulphuric Acid and Hydrochloric Acid and Oxalic Acid—Nitric Acid— Inorganic Impurities—General Impurities and Adulterations—Potassium Picrate, &c.—Picrates of the Alkaloids—Analysis of Glycerine—Residue— Silver Test—Nitration—Total Acid Equivalent—Neutrality—Free Fatty Acids—Combined Fatty Acids—Impurities—Oleic Acid—Sodium Chloride— Determination of Glycerine—Waste Acids—Sodium Nitrate—Mercury Fulminate—Cap Composition—Table for Correction of Volumes of Gases, for Temperature and Pressure

Kieselguhr Dynamite.—The material generally consists of 75 per cent. of nitro-glycerine and 25 per cent. of the infusorial earth kieselguhr. The analysis is very simple, and may be conducted as follows:—Weigh out about 10 grms. of the substance, and place over calcium chloride in a desiccator for some six to eight days, and then re-weigh. The loss of weight gives the moisture. This will generally be very small, probably never more than 1 per cent., and usually less.

Mr James O. Handy, in order to save time, proposes to dry dynamite in the following manner. He places 1 grm. of the material in a porcelain crucible 1 inch in diameter. The crucible is then supported at the bottom of an extra wide-mouthed bottle of about 600 c.c. capacity. Air, which has been dried by bubbling through strong sulphuric acid, is now drawn over the surface of the sample for three hours by means of an ordinary aspirator. The air should pass approximately at the rate of 10 c.c. per second. The tube by which the dry air enters the bottle extends to within 1 inch of the crucible containing the dynamite. An empty safety bottle is connected with the inlet, and another with the outlet of the wide-mouthed bottle. The first guards against the mechanical carrying over by the air current of sulphuric acid from the acid bottle into the sample, whilst the second prevents spasmodic outbursts of water from the exhaust from reaching the sample. The method also gave satisfactory results with nitro-glycerine. The dry substance may now be wrapped in filter paper, the whole weighed, and the nitro-glycerine extracted in the Soxhlet apparatus with ether. The ether should be distilled over at least twenty-four times.

I have found, however, that much quicker, and quite as accurate, results may be obtained by leaving the dynamite in contact with ether in a small Erlenmeyer flask for twenty-four hours—leaving it overnight is better— and decanting, and again allowing the substance to remain in contact with a little fresh ether for an hour, and finally filtering through a weighed filter, drying at 100 deg. C., and weighing. This gives the weight of the kieselguhr. The nitro-glycerine must be obtained by difference, as it is quite useless to evaporate down the ethereal solution to obtain it, as it is itself volatile to a very considerable extent at the temperature of evaporation of the ether, and the result, therefore, will always be much too low. The dry guhr can, of course, be examined, either qualitatively or quantitatively, for other mineral salts, such as carbonate of soda, &c. An actual analysis of dynamite No. 1 made by the author at Hayle gave— Moisture, 0.92 per cent.; kieselguhr, 26.15 per cent.; and nitro- glycerine, 72.93 per cent., the last being obtained by difference.

Nitro-Glycerine.—It is sometimes desired to test an explosive substance for nitro-glycerine. If an oily liquid is oozing from the substance, soak a drop of it in filter paper. If it is nitro-glycerine it will make a greasy spot. If the paper is now placed upon an iron anvil, and struck with an iron hammer, it will explode with a sharp report, if lighted it burns with a yellowish to greenish flame, emitting a crackling sound, and placed upon an iron plate and heated from beneath, it explodes sharply.

If a few drops of nitro-glycerine are placed in a test tube, and shaken up with methyl-alcohol (previously tested with distilled water, to see that it produces no turbidity), and filtered, on the addition of distilled water, the solution will become milky, and the nitro-glycerine will separate out, and finally collect at the bottom of the tube.

If to a solution of a trace of nitro-glycerine in methyl-alcohol, a few drops of a solution, composed of 1 volume of aniline, and 40 volumes sulphuric acid (1.84) be added, a deep purple colour will be produced. This colour changes to green upon the addition of water. If it is necessary to determine the nitro-glycerine quantitatively in an explosive, the scheme on page 213 may be followed. Ether is the best solvent to use. Nitrogen should be determined in the nitrometer.

Gelatine Compounds.—The simplest of these compounds is, of course, blasting gelatine, as it consists of nothing but nitro-cotton and nitro- glycerine, the nitro-cellulose being dissolved in the glycerine to form a clear jelly, the usual proportions being about 92 per cent. of nitro- glycerine to 8 per cent. nitro-cotton, but the cotton is found as high as 10 per cent. in some gelatines. Gelatine dynamite and gelignite are blasting gelatines, with varying proportions of wood-pulp and saltpetre (KNO_{3}) mixed with a thin blasting gelatine. The method of analysis is as follows:—Weigh out 10 grms. of the substance, previously cut up into small pieces with a platinum spatula, and place over calcium chloride in a desiccator for some days. Reweigh. The loss equals moisture. This is generally very small. Or Handy's method may be used. The dried sample is then transferred to a small thistle-headed funnel which has been cut off from its stem, and the opening plugged with a little glass wool, and round the top rim of which a piece of fine platinum wire has been fastened, in order that it may afterwards be easily removed from the Soxhlet tube. The weight of this funnel and the glass wool must be accurately known. It is then transferred to the Soxhlet tube and exhausted with ether, which dissolves out the nitro-glycerine. The weighed residue must afterwards be treated in a flask with ether-alcohol to dissolve out the nitro-cotton.

But the more expeditious method, and one quite as accurate, is to transfer the dried gelatine to a conical Erlenmeyer flask of about 500 c.c. capacity, and add 250 c.c. of a mixture of ether-alcohol (2 ether to 1 alcohol), and allow to stand overnight. Sometimes a further addition of ether-alcohol is necessary. It is always better to add another 300 c.c., and leave for twenty minutes or so after the solution has been filtered off. The undissolved portion, which consists of wood-pulp, potassium nitrate, and other salts, is filtered off through a linen or paper filter, dried and weighed.

Solution.—The ether-alcohol solution contains the nitro-cotton and the nitro-glycerine in solution.[A] To this solution add excess of chloroform (about 100 c.c. will be required), when the nitro-cellulose will be precipitated in a gelatinous form. This should be filtered off through a linen filter, and allowed to drain. It is useless to attempt to use a filter pump, as it generally causes it to set solid. The precipitated cotton should then be redissolved in ether-alcohol, and again precipitated with chloroform (20 c.c. of ether-alcohol should be used). This precaution is absolutely necessary, if the substance has been treated with ether- alcohol at first instead of ether only, otherwise the results will be much too high, owing to the gelatinous precipitate retaining very considerable quantities of nitro-glycerine. The precipitate is then allowed to drain as completely as possible, and finally allowed to dry in the air bath at 40 deg. C., until it is easily detached from the linen filter by the aid of a spatula, and is then transferred to a weighed watch-glass, replaced in the oven, and dried at 40 deg. C. until constant in weight. The weight found, calculated upon the 10 grms. taken, gives the percentage of nitro- cellulose.

[Footnote A: If the substance has been treated with ether alone in the Soxhlet, the nitro-glycerine will of course be dissolved out first, and the ether-alcohol solution will only contain the nitro-cellulose.]

The Residue left after treating the gelatine with ether-alcohol is, in the case of blasting gelatine, very small, and will probably consist of nothing but carbonate of soda. It should be dried at 100 deg. C. and weighed, but in the case of either gelignite or gelatine dynamite this residue should be transferred to a beaker and boiled with distilled water, and the water decanted some eight or ten times, and the residue finally transferred to a tarred filter and washed for some time with hot water. The residue left upon the filter is wood-pulp. This is dried at 100 deg. C. until constant, and weighed. The solution and washings from the wood are evaporated down in a platinum dish, and dried at 100 deg. C. It will consist of the potassium nitrate, and any other mineral salts, such as carbonate of soda, which should always be tested for by adding a few drops of nitric acid and a little water to the residue, and again evaporating to dryness and re-weighing. From the difference in weight the soda can be calculated, sodium nitrate having been formed. Thus—

Na{2}CO{3} + 2HNO{3} = 2NaNO{3} + CO{2} + H{2}O.

Mol. wt. = 106 = 170

(170 - 106 = 64) and x = (106 x d)/64

where x equals grms. of sodium carbonate in residue, and d equals the difference in weight of residue, before and after treatment with nitric acid.

The nitro-glycerine is best found by difference, but if desired the solutions from the precipitation of the nitro-cellulose may be evaporated down upon the water bath at 30 deg. to 40 deg. C., and finally dried over CaCl_{2} until no smell of ether or chloroform can be detected, and the nitro- glycerine weighed. It will, however, always be much too low. An actual analysis of a sample of gelatine dynamite gave the following result:—

Nitrocellulose (collodion) 3.819 per cent. Nitro-glycerine 66.691 " Wood-pulp 16.290 " KNO_{3} 12.890 " Na_{2}CO_{3} _Nil._ Water 0.340 "

This sample was probably intended to contain 30 per cent. of absorbing material to 70 per cent. of explosive substances. Many dynamites contain other substances than the above, such as paraffin, resin, sulphur, wood, coal-dust, charcoal, also mineral salts, such as carbonate of magnesia, chlorate of potash, &c. In these cases the above-described methods must of course be considerably modified. Paraffin, resin, and most of the sulphur will be found in the ether solution if present. The solution should be evaporated (and in this case the explosive should in the first case be treated with ether only, and not ether-alcohol), and the residue weighed, and then treated on the water bath with a solution of caustic soda. The resin goes into solution, and is separated by decantation from the residue, and precipitated by hydrochloric acid, and collected on a tarred filter (dried at 100 deg. C.), and dried at 100 deg. C. and weighed. The nitro- glycerine residue is treated with strong alcohol, decanted, and the residue of paraffin and sulphur washed with alcohol, dried, and weighed.

To separate the paraffin from the sulphur the residue is heated with a solution of ammonium sulphide. After cooling the paraffin collects as a crust upon the surface of the liquid, and by pricking a small hole through it with a glass rod the liquid underneath can be poured off, and the paraffin then washed with water, dried, and weighed. Sulphur is found by difference. Mr F.W. Smith (Jour. Amer. Chem. Soc., 1901, 23 [8], 585-589) determines the sulphur in dynamite gelatine as follows:—About 2 grms. are warmed in a 100 c.c. silver crucible on the water bath with an alcoholic solution of sodium hydroxide, and where the nitro-glycerine is decomposed, the liquid is evaporated to dryness. The residue is fused with 40 grms. of KOH and 5 grms. of potassium nitrate, the mass dissolved in dilute acetic acid and filtered, and the sulphates precipitated in the usual way. If camphor is present, it can be extracted with bisulphide of carbon after the material has been treated with ether-alcohol. In that case the sulphur, paraffin, and resin will also be dissolved. The camphor being easily volatile, can be separated by evaporation. Let the weight of the extract, freed from ether-alcohol before treatment with bisulphide of carbon, equal A, and the weight of extract after treatment with CS{2} and evaporation of the same equal B; and weight of the residue which is left after evaporation of the CS{2} and the camphor in solution equal C, the percentage of camphor will be A - B - C. The residue C may contain traces of nitro-glycerine, resin, or sulphur.

Camphor may be separated from nitro-glycerine by means of CS{2}. If the solution of camphor in nitro-glycerine be shaken with CS{2}, the camphor and a little of the nitro-glycerine will dissolve. The bisulphide solution is decanted, or poured into a separating funnel and separated from the nitro-glycerine. The two solutions are then heated on the water bath to 20 deg. C. and then to 60 deg. C., and afterwards in a vacuum over CaCl{2} until the CS{2} has evaporated from them. The camphor evaporates, and leaves the small quantity of nitro-glycerine which had been dissolved with it. The other portion is the nitro-glycerine, now free from CS{2}. The two are weighed and their weights added together, and equals the nitro- glycerine present. There is a loss of nitro-glycerine, it being partly evaporated along with the CS{2}. Captain Hess has shown that it is equal to about 1.25 per cent. This quantity should therefore be added to that found by analysis. Morton Liebschutz, in a paper in the Moniteur Scientifique for January 1893, very rightly observes that the variety of dynamites manufactured is very great, all of them having a special composition which, good or bad, is sometimes of so complicated a nature that the determination of their elements is difficult.

The determination of nitro-glycerine in simple dynamite No. 1 is easy; but not so when the dynamite contains substances soluble in ether, such as sulphur, resin, paraffin, and naphthalene. After detailing at length the methods he employs, he concludes with the observation that the knowledge of the use of acetic acid—in which nitro-glycerine dissolves—for the determination of nitro-glycerine may be serviceable. Mr F.W. Smith[A] gives the following indirect method of determining nitro-glycerine in gelatine dynamite, &c. About 15 grms. of the sample are extracted with chloroform in a Soxhlet apparatus, and the loss in weight determined. In a second portion the moisture is determined. A third portion of about 2 grms. is macerated with ether in a small beaker, the ethereal extract filtered, and the process of extraction repeated three or four times. The united filtrates are allowed to evaporate spontaneously, and the residue warmed gently on the water bath with 5 c.c. of ammonium sulphide solution, and 10 c.c. of alcohol until the nitro-glycerine is decomposed, after which about 250 c.c. of water and sufficient hydrochloric acid to render the liquid strongly acid, are added, and the liquid filtered. The precipitate is washed free from acid, and then washed through the filter with strong alcohol and chloroform into a weighed platinum dish, which is dried to constant weight at 50 deg. C. The contents of the dish are now transferred to a silver crucible, and the sulphur determined. This amount of sulphur, deducted from the weight of the contents of the platinum dish, gives the quantity of substances soluble in chloroform with the exception of the nitro-glycerine, moisture, and sulphur. The amount of the former substances plus the moisture and sulphur, deducted from the total loss on extraction with chloroform, gives the quantity of nitro-glycerine. Nitro-benzene may be detected, according to J. Marpurgo, in the following manner:—In a porcelain basin are placed two drops of liquid phenol, three drops of water, and a fragment of potash as large as a pea. The mixture is boiled, and the aqueous solution to be tested then added. On prolonged boiling nitro-benzene produces at the edge of the liquid a crimson ring, which on the addition of a solution of bleaching powder turns emerald- green. And nitro-glycerine in ether solution, by placing a few drops of the suspected solution, together with a drop or two of aniline, upon a watch-glass, evaporating off the ether, and then adding a drop of concentrated sulphuric acid to the residue, when, if nitro-glycerine is present, the H{2}SO{4} will strike a crimson colour, due to the action of the aniline sulphate upon the nitric acid liberated from the nitro- glycerine.

[Footnote A: "Notes on the Analysis of Explosives," Jour. Amer. Chem. Soc., 1901, 23 [8], 585-589.]

Tonite.—The analysis of this explosive is a comparatively easy matter, and can be performed as follows:—Weigh out 10 grms., or a smaller quantity, and boil with water in a beaker, decanting the liquid four or five times, and filter. The aqueous solution will contain the nitrate of barium. Then put the residue on the filter, and wash two or three times with boiling water. Evaporate the filtrate to dryness in a platinum dish. Dry and weigh. This equals the Ba(NO{3}){2}. If the sample is tonite No. 3, and contains di-nitro-benzol, treat first with ether to dissolve out this substance. Filter into a dish, and evaporate off the ether, and weigh the di-nitro-benzol, and afterwards treat residue with water as before. The residue is dried and weighed, and equals the gun-cotton present. It should then be treated with a solution of ether-alcohol in a conical flask, allowed to stand some three hours, then filtered through a weighed filter paper, dried at 40 deg. C., and weighed. This will give the gun-cotton, and the difference between this last weight and the previous one will give the collodion-cotton. A portion of the residue containing both the gun- cotton and the soluble cotton can be tested in the nitrometer, and the nitrogen determined.

Cordite.—This explosive consists of gun-cotton (with a little collodion-cotton in it as impurity), nitro-glycerine, and vaseline—the proportions being given as 30 per cent. nitro-glycerine, 65 per cent. gun- cotton, and 5 per cent. vaseline. Its analysis is performed by a modification of the method given for gelatines. Five grms. may be dissolved in ether-alcohol in a conical flask, allowed to stand all night, and then filtered through a linen filter. The residue is washed with a little ether, pressed, and dried at 40 deg. C., and weighed. It equals the gun-cotton. The solution contains the nitro-glycerine, soluble cotton, and vaseline. The cotton is precipitated with chloroform, filtered off, dried, and weighed. The two ether-alcohol solutions are mixed, and carefully evaporated down in a platinum dish upon the water bath at a low temperature. The residue is afterwards treated with strong 80 per cent. acetic acid, which dissolves out any nitro-glycerine left in it. The nitro-glycerine is then obtained by difference, or the method suggested to me privately by Mr W.J. Williams may be used. The residue obtained by evaporation of the ether-alcohol solution, after weighing, is treated with alcoholic potash to decompose the nitro-glycerine, water is added and the alcohol evaporated off. Some ether is then added, and the mixture shaken, and the ether separated and evaporated, and the residue weighed as vaseline.

The moisture should, however, be determined by the method devised by Mr Arthur Marshall, F.I.C., of the Royal Gunpowder Works, Waltham Abbey, which is carried out as follows:—The cordite or other explosive is prepared in the manner laid down for the Abel heat test, that is t say, it is ground in a small mill, and that portion is selected which passes through a sieve having holes of the size of No. 8 wire gauge, but not through one with holes No. 14 wire gauge.



The form of apparatus used is shown in Fig. 40. It consists of an aluminium dish A, having the dimensions shown, and the glass cone B weighing not more than 30 grms. Five grms. of the cordite are weighed into the aluminium dish A. This is covered with the cone B, and the whole is accurately weighed, and is then placed upon a metal plate heated by steam from a water bath. It is left upon the bath until all the moisture has been driven off, then it is allowed to cool for about half-an-hour in a desiccator and is weighed. The loss in weight gives accurately the moisture of the sample. For cordite of the original composition, one hour's heating is sufficient to entirely drive off the moisture; for modified cordite containing 65 per cent. of gun-cotton, two hours is enough, provided that there be not more than 1.3 per cent. of moisture present.

If the proportion of nitro-glycerine be higher, a longer heating is necessary. The aluminium dish must not be shallower than shown in the figure, for if the distance between the substance and the edge of the glass cone be less than half an inch, some nitro-glycerine will be lost. Again, the sample must not be ground finer than stated, else some of the moisture will be lost in the grinding and sieving operations, and the result will be too low. In order to be able to drive off all the moisture in the times mentioned, it is essential that the glass cone shall not fit too closely on the aluminium dish, consequently the horizontal ledge round the top of the dish should be bent, so as to render it slightly untrue, and leave a clearance of about 0.02 inch in some places. If these few simple precautions be taken, the method will be found to be very accurate. Duplicate determinations do not differ more than 0.01 per cent.[A]

[Footnote A: "Determination of Moisture in Nitro-glycerine Explosives," by A. Marshall, Jour. Soc. Chem. Ind., Feb. 29, 1904, p. 154.]

The Vaseline (C{16}H{34}), or petroleum jelly, used has a flash-point of 400 deg. F. It must not contain more than 0.2 per cent. volatile matter when heated for 12 hours on the water bath, and should have a specific gravity of 0.87 at 100 deg. F., and a melting point of 86 deg. F. It is obtained during the distillation of petroleum, and consists mainly of the portions distilling above 200 deg. C. It boils at about 278 deg. C.

Acetone (CH{3}CO.CH{3}), or dimethyl ketone, is formed when iso-propyl alcohol is oxidised with potassium bichromate and sulphuric. It is also produced in considerable quantities during the dry distillation of wood, and many other organic compounds. Crude wood spirit, which has been freed from acetic acid, consists in the main of a mixture of acetone and methyl- alcohol. The two substances may be roughly separated by the addition of calcium chloride, which combines with the methyl-alcohol. On subsequent distillation crude acetone passes over, and may be purified by conversion into the bisulphite compound.

Acetone is usually prepared, however, by the dry distillation of crude calcium or barium acetate.

(CH_{3}.COO)_{2}Ca = CH_{3}.CO.CH_{3} + CaCO_{3}.

The distillate is fractionated, and the portion, boiling between 50 deg. and 60 deg. C., mixed with strong solution of sodium bisulphite. The crystalline cake of acetone sodium bisulphite, which separates on standing, is well pressed, to free it from impurities, decomposed by distillation with dilute sodium carbonate, and the aqueous distillate of pure acetone dehydrated over calcium chloride. Acetone is a colourless, mobile liquid of sp. gr. .792 at 20 deg. C., it boils at 56.5 deg. C., has a peculiar, pleasant, ethereal odour, and is mixible with water, alcohol, and ether in all proportions.

The acetone used in the manufacture of cordite should conform to the following specification:—

SPECIFICATION FOR ACETONE.

1. The acetone to be not more than 0.802 specific gravity at 60 deg. F. When mixed with distilled water it must show no turbidity, and must leave no residue on evaporation at 212 deg. F. On distillation, four-fifths by volume of the quantity taken must distil over at a temperature not exceeding 138 deg. F. The residual matter left after this distillation must not contain, besides acetone, any ingredient that is not a bye-product incidental to the manufacture of acetone.

2. One c.c. of 0.10 per cent. solution in distilled water of pure permanganate of potash, added to 100 c.c. of the acetone, must retain its distinctive colour for not less than 30 minutes. This test should be made at a temperature of 60 deg. F.

3. The acetone tested by the following method must not show more than 0.005 per cent. of acid, calculated to acetic acid:—

To 50 c.c. of the sample diluted with 50 c.c. of distilled water, with 2 c.c. of phenol-phthalein solution (1 gramme to 1,000 c.c. of 50 per cent. alcohol) added as an indicator, add from a burette N/100 sodium hydrate solution (1 c.c. 0.0006 gramme acetic acid), and calculate to acetic acid in the usual manner.

The water used for the dilution of the acetone must be carefully tested for acidity, and the pipettes used for measuring should not be blown out, as it would be possible thus to neutralise nearly 2 c.c. of the soda solution.

The presence of water in a sample of acetone may be detected by Schweitzer and Lungwitz's method (Chem. Zeit., 1895, xix., p. 1384), which consists in shaking together equal volumes of acetone and petroleum ether (boiling point, 40 deg. to 60 deg. C.), when if present a separation of the liquid in layers will take place.

Estimation of Acetone.—Kebler (Jour. Amer. Chem. Soc., 1897, 19, 316- 320) has improved Squibb's modification of Robineau and Rollins' method. The following solutions are required:—

(1.) A 6 per cent. solution of hydrochloric acid.

(2.) A decinormal solution of sodium thiosulphate.

(3.) Alkaline potassium iodide solution prepared by dissolving 250 grms. of potassium iodide in water, made up to a litre; dissolving 257 grms. of sodium hydroxide (by alcohol) in water, likewise made up to a litre. After allowing the latter to stand, 800 c.c. of the clear solution are added to the litre of KI.

(4.) Sodium hypochlorite solution: 100 grms. of bleaching powder (35 per cent.) are mixed with 400 c.c. of water: to this is added a hot solution of 120 grms. of crystallised sodium carbonate in 400 c.c. of water. After cooling, the clear liquid is decanted, the remainder filtered, and the filtrate made up to a litre; to each litre is added 25 c.c. of sodium hydroxide solution (sp. gr. 1.29).

(5.) An aqueous solution of the acetone, containing 1 or 2 per cent. of acetone.

(6.) Bicarbonated starch solution prepared by treating 0.125 grm. of starch with 5 c.c. of cold water, then adding 20 c.c. of boiling water, boiling a few minutes, cooling, and adding 2 grms. of sodium bicarbonate.

To 20 c.c. of the potassium iodide solution are added 10 c.c. of the diluted aqueous acetone, an excess of the sodium hypochlorite solution is then run in from a burette and well shaken for a minute. The mixture is then acidified with the hydrochloric acid solution, and while agitated, an excess of sodium thiosulphate solution is added, the mixture being afterwards allowed to stand a few minutes. The starch indicator is then added, and the excess of thiosulphate re-titrated. The relation of the sodium hypochlorite solution to the sodium thiosulphate being known, the percentage of acetone can be readily calculated.[A]

[Footnote A: See "The Testing of Acetone," Conroy, Jour. Soc. Chem. Ind., 31st March 1900, vol. xix.]

Dr S.J.M. Auld has recently (Jour. Chem. Soc., Feb. 15, 1906, vol. xxv.) worked out a volumetric method for the estimation of acetone, depending on the formation of bromoform, and its subsequent hydrolysis with alcoholic potash. The hydrolysis is probably expressed thus—

3CHBr{3} + 9KOH + C{2}H{5}OH = 3CO + C{2}H{4} + 9KBr + 7H{2}O

as it has been shown by Hermann and Long that exactly 3 volumes of carbon monoxide to 1 of ethylene are evolved. The residual potassium bromide is estimated by means of standard silver nitrate solution. Bromoform is specially suitable for this purpose for several reasons. It is very readily formed by the action of bromine and potash on acetone, and although very volatile in steam, it is not liable to loss due to its own evaporation. Further, its high molecular weight and large percentage of bromine conduce to accurate results, 58 grms. of acetone being responsible for the formation of 357 grms. of KBr. The method of carrying out the analysis is as follows:—

A known quantity of the solution to be tested, containing acetone to the extent of 0.1 to 0.2 grm., is pipetted into a 500 c.c. round-bottom flask, diluted with a little water, and mixed with 20 to 30 c.c. of a 10 per cent. solution of caustic potash. The flask is connected with a long reflex condenser, and is also fitted with a dropping funnel containing a solution of bromine in potassium bromide (200 grms. of Br and 250 grms. of KBr to 1 litre of water). The bromine solution is allowed to flow into the mixture until it has acquired a faint yellow tinge, the flask and its contents being then heated on the water bath at about 70 deg. C. for half-an- hour. Bromine solution is added drop by drop until the slight coloration is permanent, excess of bromine being got rid of by boiling for a minute or two with a little more caustic potash. The mixture is then distilled until the distillate is free from bromoform, halogen being tested for in the usual manner. Water is added to the contents of the flask if necessary. It may be here observed that no acetone can be detected in the distillate by means of the mercuric oxide test, and free bromine is also absent. The condenser having been washed out with a little alcohol, in order to remove any traces of bromoform which may have collected, the distillate and washings are mixed with 50 c.c. of alcohol and sufficient solid caustic potash to make an approximately 10 per cent. solution. The mixture is then heated on the water bath under a reflux condenser until the bromoform is completely decomposed. This generally occupies about three-quarters of an hour. The liquid is allowed to cool, evaporated to smaller bulk if necessary, and exactly neutralised with dilute nitric acid. It is then diluted with water to 500 c.c., and an aliquot part titrated with N/10 silver nitrate solution, using potassium chromate as indicator; 240 parts of bromine correspond to 58 parts of acetone. The complete analysis can be performed in one and a half to two hours. It is imperative that the bromine used should be pure, as crude bromine frequently contains bromoform. The method is suitable for the estimation of acetone in wood-spirit, the spirit being diluted to 10 times its volume, and 5 c.c. of this solution employed for the determination. For example—

(1.) Three c.c. of a solution containing 9.61 per cent. acetone gave 1.7850 grm. KBr. Acetone found = 9.66 per cent.

(2.) Ten c.c. of a solution containing 0.96 per cent. acetone gave 0.5847 grm. KBr. Acetone found = 0.95 per cent.

Nitro-Cotton.—The first thing upon opening a case of wet cotton, or in receiving a sample from the "poacher," that requires to be determined is the percentage of water that it contains. It is best done by weighing out about 1,000 grms. upon a paper tray, which has been previously dried in the oven at 100 deg. C. for some time, and become constant in weight. The trayful of cotton is then placed in a water oven, kept at 100 deg. C., and dried as long as it loses water. The loss gives the percentage of water. It varies from 20 to 30 per cent. as a rule in "wet" cotton.

OUTLINE SCHEME FOR THE ANALYSIS OF NITRO-EXPLOSIVES Exhaust dried substance with Anhydrous Ether in Soxhlet's Fat Extraction Apparatus. Solution Divide into two parts A. and B. A. Allow ether to evaporate spontaneously. Dry residue in vacuo over H{2}SO{4} and weigh. Equals nitro-glycerine, resin, camphor, and paraffin. The nitro-glycerine in this residue may be decomposed by heating with a solution of alcoholic potash. Water may then be added, and the alcohol evaporated off on the water bath. From this solution the resin may be precipitated by HCl, filtered off, dried, and weighed. Solution containing the paraffin is treated with AmS solution and heated. On cooling the paraffin separates, and may be separated. Residue may be shaken with CS{2} to remove camphor. B. Add phenol-phthalein and titrate with alcoholic potash, 1 c.c. normal KHO = .330 grm. resin, and add considerably more KHO. Evaporate, dissolve residue in water, shake with ether, and separate. Ethereal Solution evaporated leaves paraffin. Aqueous Solution Add bromide, acidify with HCl, separate any resin and precipitate, filtrate with BaCl{2} BaSO{4} x .1373 = Sulphur. Residue Dry, weigh, and exhaust with water preferably in Soxhlet. Solution Residue Contains metallic Dry, weigh, and agitate an aliquot part with nitrates, chlorates, with H{2}SO{4} and Hg in nitrometer. If soluble carbonates, nitro-cellulose is present, treat remainder of the sum of which residue with ether-alcohol. (except AmCO{3}) can be determined by evaporating down at Solution 100 deg. C. to dryness Evaporate and weigh. Residue consists of and weighing. soluble nitro-cellulose. Nitrates can be determined by Residue Dry and weigh and determine hexa-nitro- cellulose in nitrometer, if present. Exhaust remainder with acetic ether. Solution Residue Hexa-nitro-cellulose Dry and weigh, ignite (Gun cotton). and reweigh. Loss = Cellulose. Residue consists of sawdust, charcoal, coal, chalk, guhr, or mineral matter, &c.

NOTE.—Camphor is found by difference. Sulphur is only partially soluble in ether. It is better, therefore, to extract some of the original substance with water, and treat residue with alcoholic KHO. Add bromide, acidify, and precipitate as BaSO.

The Solubility Test.—The object of this test is to ascertain, in the case of gun-cotton, the percentage of soluble (penta and lower nitrates) cotton that it contains, or in the case of soluble cotton, the quantity of gun-cotton. The method of procedure is as follows:—Five grms. of the sample which has been previously dried at 100 deg. C., and afterwards exposed to the air for two hours, is transferred to a conical flask, and 250 c.c. ether-alcohol added (2 ether to 1 alcohol). The flask is then corked and allowed to digest, with repeated shaking, for two or three hours. The whole is then transferred to a linen filter, and when the solution has passed through the filter, is washed with a little ether, and pressed in a hand-screw press between folds of filter paper. The sample is then returned to the flask, and the previous treatment repeated, but it will be sufficient for it to digest for one hour the second time. The filter is then again pressed first gently by hand, then in the press, and afterwards opened up and the ether allowed to evaporate. The gun-cotton is then removed from the filter and transferred to a watch-glass, and dried in the water oven at 100 deg. C. When dry it is exposed to the air for two hours and weighed. It equals the amount of gun-cotton and unconverted cotton in the 5 grms. The unconverted cotton must be determined in a separate 5 grms. and deducted.

The method of determining the soluble cotton now used in the Government laboratories is as follows:—Fifty grains of the nitro-cotton are dissolved in 150 c.c. of ether-alcohol, and allowed to stand, with frequent shakings, in a 200 c.c. stoppered measure for six hours; 75 c.c. of the clear solution are then drawn off by the aid of a pipette and evaporated in a dish on the water bath, and finally in the water oven at 120 deg. F. (49 deg. C.), until constant in weight. The weight found equals the quantity of soluble cotton in the 75 c.c., which, multiplied by 4, equals the percentage, thus: Suppose that 2.30 grains was the weight found, then

(2.3 x 150)/75 = 4.6 in 50 = 9.20 per cent.

A method for the determination of soluble nitro-cellulose in gun-cotton and smokeless powder has been published by K.B. Quinan (Jour. Amer. Chem. Soc., 23 [4], 258). In this method about 1 grm. of the finely divided dry sample to be analysed is placed in an aluminium cup 1.9 inch in diameter and 4-1/8 inch deep. It is then covered and well stirred with 50 c.c. of alcohol, 100 c.c. of ether are then added, and the mixture is stirred for several minutes. After removing the stirrer, the cup is lightly covered with an aluminium lid, and is then placed in the steel cup of a centrifugal machine, which is gradually got up to a speed of 2,000 revolutions per minute, the total centrifugal force at the position occupied by the cups (which become horizontal when in rapid rotation) is about 450 lbs. They are rotated at the full speed for ten to twelve minutes, and the machine is then gradually stopped. By this time the whole of the insoluble matter will be at the bottom of the cup, and the supernatant solution will be clear. It is drawn off to within a quarter of an inch of the bottom (without disturbing the sediment), with the aid of a pipette.

Care must be taken that the solution thus withdrawn is perfectly clear. About 10 to 15 c.c. of colloid solution and a film of insoluble matter remain at the bottom of the cup; these are stirred up well, the stirrer is rinsed with ether-alcohol, about 50 c.c. of fresh ether-alcohol are added; the mixture is again treated in the centrifugal apparatus for about eight minutes; the whole washing process is then repeated until all soluble matter has been removed. This may require about seven or eight (or for samples with much insoluble matter ten or twelve or more) washings, but as the extraction proceeds, the period of rotation may be somewhat reduced. After extraction is completed, the insoluble matter is transferred to a Gooch crucible with the usual asbestos pad, dried at 100 deg. C., and weighed. The residue may, if wished, be dried and weighed in the aluminium cup, but then it cannot be ignited. The whole time for an analysis exclusive of that required for drying, is from one to two hours—average time, 1-1/4 hour. The results are satisfactory both as to accuracy and rapidity. Acetone-soluble nitro-cellulose may be determined by the same method.

The Unconverted or Non-nitrated Cotton.—However well the cotton has been nitrated, it is almost certain to contain a small quantity of non- nitrated or unconverted cotton. This can be determined thus:—Five grms. of the sample are boiled with a saturated solution of sodium sulphide, and then allowed to stand for forty-eight hours, and afterwards filtered or decanted, and again boiled with fresh solutions of sulphide, and again filtered, washed first with dilute HCl and then with water, dried, and weighed. The residue is the cellulose that was not nitrated, plus ash, &c. It should be ignited, and the weight of the ash deducted from the previous weight.

Acetone, and acetic-ether (ethyl-acetate) may also be used as solvents for the nitro-cellulose. Another process is to boil the gun-cotton, &c., in a solution of sodium stannate made by adding caustic soda to a solution of stannous chloride, until the precipitate first formed is just re-dissolved. This solution dissolves the cellulose nitrates, but does not affect the cellulose. Dr Lunge found the following process more satisfactory in the case of the more highly nitrated products:—The reagent is an alcoholic solution of sodium-ethylate prepared by dissolving 2 to 3 grms. of sodium in 100 c.c. of 95 per cent. alcohol, and mixing the filtered solution with 100 c.c. of acetone. It has no effect upon cellulose, but decomposes nitro-cellulose with the formation of a reddish brown compound, which is soluble in water. In the determination, 5 grms. of gun-cotton are heated to 40 deg. or 50 deg. C. on the water bath with 150 c.c. of the reagent, the liquid being shaken at intervals for twenty to thirty minutes; or the mixture may be allowed to stand for a few hours at the ordinary temperature. The brown-red solution is decanted from the undissolved residue, and the latter washed with alcohol and with water, by decantation, and then on the filter with hot water, to which a little hydrochloric acid is added for the final washings. For ordinary work this cellulose is dried immediately and weighed, but in exact determinations it is washed with alcohol, again treated with 50 c.c. of the reagent, and separated and washed as before. The cellulose thus obtained, gives no trace of gas in the nitrometer, and duplicate determinations agree within 0.1 to 0.2 per cent. when the weight of unchanged cellulose amounts to about 0.2 grm. Gun-cotton, which is completely soluble in acetone, contains only traces of cellulose, and when as much as 0.85 per cent. is present it does not dissolve entirely. This method is not applicable to the determination of cellulose in lower nitrated products, and Dr Lunge attributes this to the fact that these being prepared with less concentrated acid invariably contain oxy-cellulose.

Alkalinity.—Five grms. of the air-dried and very finely divided sample are taken from the centre of the slabs or discs, and digested with about 20 c.c. of N/2 hydrochloric acid, and diluted with water to about 250 c.c., and shaken for about fifteen minutes. The liquid is then decanted, and washed with water until the washings no longer give an acid reaction. The solution, together with the washings, are titrated with N/4 sodium carbonate, using litmus as indicator.

Ash and Inorganic Matter.—This is best determined by mixing 2 or 3 grms. of the nitro-cotton in a platinum crucible with shavings of paraffin, heating sufficiently to melt the paraffin, and then allowing the contents of the crucible to catch fire and burn away quietly. The temperature is then raised, and the carbonaceous residue incinerated, cooled, weighed, &c., and the percentage of ash calculated. Schjerning proceeds in the following way:—He takes 5 grms. of the nitro-cotton in a large platinum crucible, he then moistens it with a mixture of alcohol and ether, in which paraffin has been dissolved to saturation, and filtered and mixed with one-fourth of its volume of water. Some fragments of solid paraffin are then added, and the ether set on fire. Whilst this is in progress the crucible is kept in an oblique position, and is rotated so that the gun-cotton may absorb the paraffin uniformly. The partially charred residue is now rubbed down with a rounded glass rod, and the crucible is covered and heated for from fifteen to twenty minutes over the blow-pipe, the lid being occasionally removed. The residue is soon converted into ash, which is weighed, and then washed out into a porcelain basin and treated with hydrochloric acid heated to 90 deg. C. The oxide of iron, alumina, lime, and magnesia are thus dissolved, and the silica remains as insoluble residue. The rest of the analysis is conducted according to the well-known methods of separation. The percentage of ash as a whole is generally all that is required.

Examination of Nitrated Celluloses with Polarised Light.—Dr G. Lunge (Jour. Amer. Chem. Soc., 1901, 23 [8], 527) has formed the following conclusions:—The most highly nitrated products appear blue in polarised light, but those containing between 13.9 and 13.0 per cent. of nitrogen cannot be distinguished from each other by polarisation. As the percentage of nitrogen rises, the blue colour becomes less intense, and here and there grey fibres can be observed, though not in proportion to the increase in the nitrogen. Below 12.4 per cent. of nitrogen, the fibres show a grey lustre, which usually appears yellow when the top light is cut off. Below 10 per cent. of nitrogen, the structure is invariably partially destroyed and no certain observations possible. It is only possible to distinguish with certainty, firstly any unchanged cellulose by its flashing up in variegated (rainbow) colours; and secondly, highly nitrated products (from 12.75 per cent. N upwards), by their flashing up less strongly in blue colours. The purple transition stage in the fibres containing over 11.28 per cent. of N (Chardonnet) was not observed by Dr Lunge.

Determination of Nitrogen by Lunge Nitrometer.—The determination of the percentage of nitrogen in a sample of gun-cotton or collodion is perhaps of more value, and affords a better idea of its purity and composition, than any of the foregoing methods of examination, and taken in conjunction with the solubility test, it will generally give the analyst a very fair idea of the composition of his sample. If we regard gun-cotton as the hexa-nitro-cellulose, the theoretical amount of nitrogen required for the formula is 14.14 per cent., and in the same way for collodion-cotton, which consists of the lower nitrates, chiefly, however, of the penta- nitrate, the theoretical nitrogen is 12.75 per cent., so that if in a sample of nitro-cotton the nitrogen falls much lower than 14 per cent., it probably contains considerable quantities of the lower nitrates, and perhaps some non-nitrated cellulose as well (C{6}H{10}O{5}){x}, which of course would also lower the percentage of nitrogen.

The most expeditious method of determining the nitrogen in these nitro bodies is by the use of Lunge's nitrometer (Fig. 41), and the best way of working the process is as follows:—Weigh out with the greatest care 0.6 grm. of the previously dried substance in a small weighing bottle of about 15 c.c. capacity, and carefully add 10 c.c. of concentrated sulphuric acid from a pipette, and allow to stand until all the cotton is dissolved. The nitrometer should be of a capacity 150 to 200 c.c., and should contain a bulb of 100 c.c. capacity at the top, and should be fitted with a Greiner and Friederich's three-way tap. When the nitro-cotton has entirely dissolved to a clear solution, raise the pressure tube of the nitrometer so as to bring the mercury in the measuring tube close up to the tap. Open the tap in order to allow of the escape of any air bubbles, and clean the surface of the mercury and the inside of the cup with a small piece of filter paper. Now close the tap, and pour the solution of the nitro-cotton into the cup. Rinse out the bottle with 15 c.c. of sulphuric acid, contained in a pipette, pouring a little of the acid over the stopper of the weighing bottle in case some of the solution may be on it. Now lower the pressure tube a little, just enough to cause the solution to flow into the bulb of the measuring tube, when the tap is slightly opened. When the solution has run in almost to the end, turn off the tap, wash down the sides of the bottle, and add to the cup of the nitrometer; allow it to flow in as before, and then wash down the sides of the cup with 10 c.c. of sulphuric acid, adding little by little, and allowing each portion added to flow into the bulb of the nitrometer before adding the next portion. Great care is necessary to prevent air bubbles obtaining admission, and if the pressure tube is lowered too far, the acid will run with a rush and carry air along with it.



The solution being all in the measuring tube, the pressure tube is again slightly raised, and the tube containing the nitro-cotton solution shaken for ten minutes with considerable violence. It is then replaced in the clamp, and the pressure relieved by lowering the pressure tube, and the whole apparatus allowed to stand for twenty minutes, in order to allow the gas evolved to assume the temperature of the room. A thermometer should be hung up close to the bulb of the measuring tube. At the end of the twenty minutes, the levels of the mercury in the pressure and measuring tubes are equalised, and the final adjustment obtained by slightly opening the tap on the measuring tube (very slightly), after first adding a little sulphuric acid to the cup, and observing whether the acid runs in or moves up. This must be done with very great care. When accurately adjusted, it should move neither way. Now read off the volume of the NO gas in cubic centimetres from the measuring tube. Read also the thermometer suspended near the bulb, and take the height of the barometer in millimetres. The calculation is very simple.

EXAMPLE—COLLODION-COTTON.

0.6[A] grm. taken. Reading on measuring tube = 114.6 c.c. NO. Barometer— 758 mm. Temperature—15 deg. C.

[Footnote A: 0.5 grm. is enough in the case of gun-cotton.]

Since 1 c.c. NO = 0.6272 milligramme N, and correcting for temperature and pressure by the formula

760 x (1 + d^{2}) (d = .003665), for temperature 15 deg. = 801.78,[A]

then

(114.6 x 100 x 750 x .6272)/(801.7 x. 6) = 11.22 per cent. nitrogen.

[Footnote A: See Table, page 244.]

The nitrogen in nitro-glycerine may of course be determined by the nitrometer, but in this case it is better to take a much smaller quantity of the substance. From 0.1 to 0.2 grm. is quite sufficient. This will give from 30 to 60 c.c. of gas, and therefore a measuring tube without a 100 c.c. bulb must be used.

EXAMPLE.

0.1048 grm. nitroglycerine taken gave 32.5 c.c. NO. Barometer, 761 mm. Temperature, 15 deg. C.

Therefore,

(3.25 x 100 x 761 x .6272)/(801.78 x.1048) = 18.46 per cent. N. Theory = 18.50 per cent.

Professor Lunge has devised another form of nitrometer (Fig. 42), very useful in the nitrogen determination in explosives. It consists of a measuring tube, which is widened out in the middle to a bulb, and is graduated above and below into 1/10 c.c. The capacity of the whole apparatus is 130 c.c.; that of each portion of the tube being 30 c.c., and of the bulb 70 c.c. The upper portion of the graduated tube serves to measure small volumes of gas, whilst larger volumes are read off on the lower part.



F.M. Horn (Zeitschrift fuer angewandte Chemie, 1892, p. 358) has devised a form of nitrometer (Fig. 43) which he has found especially useful in the examination of smokeless powders. The tap H is provided with a wide bore through which a weighed quantity of the powder is dropped bodily into the bulb K. From 4 to 5 c.c. of sulphuric acid which has been heated to 30 deg. C. are then added through the funnel T, the tap H being immediately closed. When the powder has dissolved—a process which may be hastened by warming the bulb very carefully—the thick solution is drawn into the nitrometer tube N, and the bulb rinsed several times with fresh acid, after which operation the analysis is proceeded with in the usual way.

Dr Lunge's method of using a separate nitrometer in which to measure the NO gas evolved to the one in which the reaction has taken place, the gas being transferred from the one to the other by joining them by means of indiarubber tubing, and then driving the gas over by raising the pressure tube of the one containing the gas, the taps being open, I have found to be a great improvement.

1 c.c. NO gas at 0 deg. and 760 mm. Equals 0.6272 milligrammes (N) nitrogen. " 1.343 " nitric oxide. " 2.820 " (HNO_{3}) nitric acid. " 3.805 " (NaNO_{3}) sodium nitrate. " 4.523 " (KNO_{3}) potassium nitrate.

Champion and Pellet's Method.—This method is now very little used. It is based upon the fact that when nitro-cellulose is boiled with ferrous chloride and hydrochloric acid, all the nitrogen is disengaged as nitric oxide (NO). It is performed as follows:—A vacuum is made in a flask, fitted with a funnel tube, with a glass stopper on the tube; a delivery tube that can also be closed, and which dips under a solution of caustic soda contained in a trough, and the end placed under a graduated tube, also full of caustic soda. From 0.12 to 0.16 grm. cotton dissolved in 5 to 6 c.c. of sulphuric acid is allowed to flow into the flask, which contains the ferrous chloride and hydrochloric acid, and in which a vacuum has been formed by boiling, and then closing the taps. The solution is then heated, the taps on the delivery tube opened, and the end placed under the collecting tube, and the NO evolved collected. The NO gas is not evolved until the solution has become somewhat concentrated. Eder substituted a solution of ferrous sulphate in HCl for ferrous chloride. Care must be taken that the flask used is strong enough to stand the pressure, or it will burst.

The same chemists (Compt. Rendus, lxxxiii. 707) also devised the following method for determining the NO{2} in nitro-glycerine:—A known quantity of a solution of ferrous sulphate of previously ascertained reducing power is placed in a flask, acidified with hydrochloric acid, and its surface covered with a layer of petroleum oil. About .5 grm. of the nitro-glycerine is then introduced, and the flask heated on the water bath. When the sample is completely decomposed, the liquid is heated to boiling to remove nitric oxide, and the excess of ferrous sulphate ascertained by titration with standard permanganate; 56 of iron (Fe) oxidised by the sample correspond to 23 of NO{2} in the sample of nitro-glycerine.

The Schultze-Tieman Method for determining nitrogen in nitro-explosives, especially nitro-cellulose and nitro-glycerine.—The figure (No. 44) shows the general arrangement of the apparatus. I am indebted for the following description of the method of working it to my friend, Mr William Bate, of Hayle. To fill the apparatus with the soda solution, the gas burette is put on the indiarubber stopper of basin W, and firmly clamped down. Then the taps A and C are opened, and B closed. When the burette is filled with soda solution half-way up the funnel Y, A and C are closed, and B opened. The arrows show the inlet and outlet for the cooling water that is kept running through the water jacket round the nitrometer tube. To collect the gas, raise the nitrometer off the rubber stopper, and place the gas tube from the decomposition apparatus in the glass dish W and under the opening of the nitrometer.



For the estimation of nitrogen in nitro-cellulose take .5 to .65 grm., and place in the decomposition flask f (Fig. 45), washing in with about 25 c.c. of water by alternately opening clips D and E. The air in the flask is driven out by boiling, whilst the air is shut off by the tube i dipping into the basin W, which is filled with the soda lye, and tube K is placed in the test tube R, which contains a few c.c. of water. As soon as all the air is completely driven out, clips D and E are closed, and the gas jet is taken away. (This flask must be a strong one, or it will burst.) Into test tube R, 25 c.c. of concentrated solution of protochloride of iron and 10 to 15 c.c. concentrated hydrochloric acid are poured, which are sucked up into the developing flask f by opening clip E, air being carefully kept from entering. The clip E is now closed, and tube i is put underneath the burette, and the development of NO gas is commenced by heating the contents of the flask f. When the pressure of the gas in the flask has become greater than the pressure of the atmosphere, the connecting tube begins to swell at i, whereupon clip D is opened, and the boiling continued with frequent shaking of the bulb, until no more nitrous gas bubbles rise up into the soda lye, the distilling over of the HCl causes a crackling noise, the clip D is closed, and E opened. The burette is again put hermetically on the indiarubber stopper in basin W, and the apparatus is left to cool until the water discharged through P shows the same temperature as the water flowing through (into the cooling jacket) Z. If the level of the soda solution in the tube X is now put on exactly the same level as that in the burette by lowering or elevating the tube X as required, the volume of NO obtained in c.c. can be read off within 1/10 c.c., and the percentage of nitrogen calculated by the usual formula.



The solution of protochloride of iron is obtained by dissolving iron nails, &c., in concentrated HCl, the iron being in excess. When the development of hydrogen ceases, it is necessary to filter warm through a paper filter, and acidify filtrate with a few drops of HCl. The soda solution used has a sp. gr. of 1.210 to 1.260; equals 25 deg. to 30 deg. B. The nitro-cellulose is dried in quantities of 2 grms. at 70 deg. C. during eight to ten hours, and then three hours in an exiccator over H{2}SO{4}. The results obtained with this apparatus are very accurate. The reaction is founded upon that of MM. Champion and Pellet's method.

The Kjeldahl Method of Determining Nitrogen.—This method, which has been so largely used by analysts for the determination of nitrogen in organic bodies, more especially perhaps in manures, was proposed by J. Kjeldahl,[A] of the Carlsberg Laboratory of Copenhagen. It was afterwards modified by Jodlbauer, of Munich,[B] and applied to the analysis of nitro- explosives by M. Chenel, of the Laboratoire Centrale des Poudres, whose method of procedure is as follows:—0.5 grm. of the finely powdered substance is digested in the cold with a solution of 1.2 grm. of phenol and 0.4 grm. phosphoric anhydride in 30 c.c. of sulphuric acid. The mixture is kept well shaken until the solution is complete. From 3 to 4 grms. of zinc-dust is then cautiously and gradually added, the temperature of the mass being kept down until complete reduction has been effected. Finally, 0.7 grm. of mercury is added, and the process continued in the usual way, according to Kjeldahl; that is, the liquid is distilled until all the ammonia has passed over, and is absorbed in the standard acid. The distillate is then titrated with standard ammonia.

[Footnote A: J. Kjeldahl, Zeitschrift Anal. Chem., 1883, xxii., p. 366.]

[Footnote B: Jodlbauer, Chemisches Centralblatt, 1886, pp. 434-484. See also Arms and Explosives, 1893, p. 87.]

The NO_{2} group is at the moment of solution fixed upon the phenol with the production of mono-nitro-phenol, which is afterwards reduced by the action of the zinc-dust into the amido derivative. During the subsequent combustion, the nitrogen of the amido-phenol becomes fixed in the state of ammonia. M. Chenel is perfectly satisfied with the results obtained, but he points out that the success of the operation depends upon the complete conversion of the phenol into the mono-nitro derivatives. This takes place whenever the organic compound forms a _clear solution_ in the cold sulphuric acid mixture. Substances like collodion or gun-cotton must be very finely divided for successful treatment. The following table shows some of the results obtained by M. Chenel:—

Total Nitrogen. Substances Analysed. Calculated. Found. Saltpetre (KNO{3}) 13.86 13.91 13.82 13.73 13.96 Ammonium nitrate 35.00 35.31 34.90 34.96 Barium nitrate 10.72 10.67 10.62 Nitro-glycerol 18.50 18.45 Di-nitro-benzol[A] 16.67 16.78 16.57 Para-nitro-phenol 10.07 10.03 Picric acid[A] 18.34 18.42 18.43 Ammonium picrate 22.76 22.63 22.67 Di-nitro-ortho-cresol 14.14 14.10 13.98 Tri-nitro-meta-cresol 17.28 17.57 17.27

[Footnote A: Dr. Bernard Dyer obtained 18.39 per cent. for picric acid and 16.54 per cent. for di-nitro-benzol.—Jour. Chem. Soc., Aug. 1895.]

When Chenel endeavoured to apply Jodlbauer's modification of Kjeldahl's process to the examination of the tri- and tetra-nitrated naphthalenes, he found that good results were not obtainable, because these compounds do not dissolve completely in the cold sulphuric acid. It may, however, be used if they are previously converted into the naphthylamines, according to the plan proposed by D'Aguiar and Lautemann (Bull. Soc. Chim., vol. iii., new series, p. 256). This is rapidly effected as follows:—Twelve grms. of iodine are gradually added to a solution of 2 grms. of phosphorus in about 15 or 20 c.c. of bisulphide of carbon, this solution being contained in a flask of 250 c.c. capacity. The flask and its contents are heated on the water bath at 100 deg. C. with constant attention, until the last traces of the carbon bisulphide have distilled away. It is then cooled, and the iodide of phosphorus is detached from the sides of the flask by shaking, but not expelled. The next step is to add about 0.5 to 0.6 grm. of the substance that is to be analysed, after which 8 grms. of water are introduced, and the flask is agitated gently two or three times. As soon as the reaction becomes lively, the contents of the flask are well shaken. It is usually finished about one minute after the addition of the water. The flask is now cooled, and 25 c.c. of sulphuric acid, together with 0.7 grm. of mercury, are gradually added; hydriodic acid (HI) forms, and the temperature of the flask must be raised sufficiently to expel it. The remaining part of the operation is as in the ordinary Kjeldahl process.

M. Chenel has found this process the best for the analysis of the nitro- naphthalenes, and for impervious substances like collodion or gun-cotton. Personally, I have never been able to obtain satisfactory results with this process in the analysis of nitro-cellulose, and I am of opinion that the process does not possess any advantage over the nitrometer method, at any rate for the analysis of gun-cotton.

Table giving the Percentages of Nitrogen and Oxide of Nitrogen in Various Substances used in or as Explosives:

Name FORMULAE NITROGEN NO_{2} per cent. per cent.

Nitroglycerine C{3}H{5}(ONO{2}){3} 18.50 = 60.70 Hexa-nitro-cellulose C{12}H{14}O{4}(ONO{2}){6} 14.14 = 46.42 Penta-nitro-cellulose C{6}H{8}O{5}(ONO{2}){5} 11.11 = 36.50 Nitro-benzene C{6}H{5}NO{2} 11.38 = 37.39 Di-nitro-benzene C{6}H{4}(NO{2}){2} 16.67 = 54.77 Tri-nitro-benzene C{6}H{3}(NO{2}){3} 19.24 = 63.22 Nitro-toluene C{7}H{7}NO{2} 10.21 = 33.49 Nitro-naphthalene C{10}H{7}NO{2} 8.09 = 26.53 Di-nitro-naphthalene C{10}H{6}(NO{2}){2} 12.84 = 42.12 Nitro-mannite C{6}H{7}(NO{3}){6} 23.59 = 77.37 Nitro-starch C{6}H{8}O{4}(HNO{3}) 6.76 = 22.18 Picric acid (Tri-nitro-phenol) C{6}H{2}OH(NO{2}){3} 18.34 = 60.15 Chloro-nitro-benzene C{6}H{3}Cl(NO{2}){2} 13.82 = 45.43 Ammonium nitrate NH{4}NO{3} 35.00 = Sodium nitrate NaNO{3} 16.47 = Potassium nitrate KNO{3} 13.86 = Nitric acid HNO{3} 22.22 = Barium nitrate Ba(NO{3}){2} 10.72 =

Analysis of Celluloid.—The finely divided celluloid is well stirred, by means of a platinum wire, with concentrated sulphuric acid in the cup of a Lunge nitrometer, and when dissolved the nitrogen determined in the solution in the usual way. To prevent interference from camphor, the following treatment is suggested by H. Zaunschirm (Chem. Zeit., xiv., 905). Dissolve a weighed quantity of the celluloid in a mixture of ether- alcohol, mixed with a weighed quantity of washed and ignited asbestos, or pumice-stone, dry, and disintegrate the mass, and afterwards extract the camphor with chloroform, dry, and weigh: then extract with absolute methyl-alcohol, evaporate, weigh, and examine the nitro-cellulose in the nitrometer.

Picric Acid and Picrates.—Picric acid is soluble in hot water, and to the extent of 1 part in 100 in cold water, also in ether, chloroform, glycerine, 10 per cent. soda solution, alcohol, amylic alcohol, carbon bisulphide, benzene, and petroleum. If a solution of picric acid be boiled with a strong solution of potassium cyanide, a deep red liquid is produced, owing to the formation of potassium iso-purpurate, which crystallises in small reddish-brown plates with a beetle-green lustre. This, by reaction with ammonium chloride, gives ammonium iso-purpurate (NH_{4}C_{8}H_{4}N_{5}O_{6}), or artificial murexide, which dies silk and wool a beautiful red colour. On adding barium chloride to either of the above salts, a vermilion-red precipitate was formed, consisting of barium iso-purpurate. With ammonio-sulphate of copper, solutions of picric acid give a bright green precipitate. Mr A.H. Allen gives the following methods for the assay of commercial picric acid, in his "Commercial Organic Analysis":—

Resinous and Tarry matters are not unfrequently present. They are left insoluble on dissolving the sample in boiling water. The separation is more perfect if the hot solution be exactly neutralised by caustic soda.

Sulphuric Acid, Hydrochloric Acid, and Oxalic Acid, and their salts are detected by adding to the filtered aqueous solution of the sample solutions of the picrates of barium, silver, and calcium. These salts are readily made by boiling picric acid with the carbonates of the respective metals and filtering: other soluble salts of these methods may be substituted for the picrates, but they are less satisfactory.

Nitric Acid may be detected by the red fumes evolved on warming the sample with copper turnings.

Inorganic Impurities and Picrates of Potash and Sodium, &c., leave residues on cautious ignition.

General Impurities and Adulterations may be detected and determined by shaking 1 grm. of the sample of acid in a graduated tube with 25 c.c. of ether, the pure acid dissolves, while any oxalic acid, nitrates, picrates, boric acid, alum, sugar, &c., will be left insoluble, and after removal of the ethereal liquid, may be readily identified and determined. For the detection and determination of water and of oxalic acid, 50 c.c. of warm benzene may be advantageously substituted for ether. Sugar may be separated from the other impurities by treating the residue insoluble in ether or benzene with rectified spirit, in which sugar and boric acid alone will dissolve. If boric acid be present, the alcoholic solution will burn with a green flame. Mono- and di-nitrophenic acids lower the melting point (122 deg. C). Their calcium salts are less soluble than the picrate, and may be approximately separated from it by fractional crystallisation, or by precipitating the hot saturated solution of the sample with excess of lime water. Picric acid may be determined by extracting the acidulated aqueous solution by agitation with ether or benzene, and subsequently removing and evaporating off the solvent. It may also be precipitated as the potassium salt.

Potassium Picrate [KC{6}H{2}(NO{2}){3}O]. When a strong solution of picric acid is neutralised by carbonate of potash, this salt is thrown down in yellow crystalline needles, which require 260 parts of cold or 14 parts of hot water for their solution. In alcohol it is much less soluble.

Ammonium Picrate is more soluble in water than the above, and sodium picrate is readily soluble in water, but nearly insoluble in solution of sodium carbonate.

Picrates of the Alkaloids.—Picric acid forms insoluble salts with many of the alkaloids, and picric acid may be determined in the following manner:—To the solution of picric acid, or a picrate, add a solution of sulphate of cinchonine acidulated with H{2}SO{4}. The precipitated picrate of cinchonine [C{20}H{24}N{2}O(C{6}H{2}N{3}O{7}){2}] is washed with cold water, rinsed off the filter into a porcelain crucible or dish, the water evaporated on the water bath, and the residual salt weighed. Its weight, multiplied by .6123, gives the quantity of picric acid in the sample taken.

Analysis of Glycerine.[A] Glycerine that is to be used for the manufacture of nitro-glycerine should have a minimum specific gravity of 1.261 at 15 deg. C. This can be determined, either by the aid of a Sartorius specific gravity balance, or by using an ordinary specific gravity bottle. One of 10 or 25 c.c. capacity is very convenient.

[Footnote A: See also Sulman and Berry, Analyst, xi., 12-34, and Allen's "Commercial Organic Analysis," vol. ii., part i.]

Residue[A] left upon evaporation should not be more than 0.25 per cent. To determine this, take 25 grms. of the glycerine, and evaporate it at a temperature of about 160 deg. C. in a platinum basin, and finish in an air bath. Weigh until constant weight is obtained. Afterwards incinerate over a bunsen burner, and weigh the ash.

[Footnote A: Organic matter up to .6 per cent. is not always prejudicial to the nitrating quantities of a glycerine.]

Silver Test. A portion of the sample of glycerine to be tested should be put in a small weighing bottle, and a quarter of its bulk of N/10 silver nitrate solution added to it, then shake it, and place in a dark cupboard for fifteen minutes. It must be pronounced bad if it becomes black or dark brown within that time (acrolein, formic, and butyric acids).

The German official test for glycerine for pharmaceutical purposes is much more stringent, 1 c.c. of glycerine heated to boiling with 1 c.c. of ammonia solution and three drops of silver nitrate solution must give neither colour or precipitate within five minutes.

Nitration. Fifty grms. of the glycerine are poured from a beaker into a mixture of concentrated nitric acid (specific gravity 1.53) and sulphuric acid (1.84), mixed in the proportions of 3 HNO_{3} to 5 H_{2}SO_{4} (about 400 c.c. of mixed acids). The mixed acids should be put into a rather large beaker, and held in the right hand in a basin of water, and the glycerine slowly poured into them from a smaller one held in the left. A constant rotatory motion should be given to the beaker in which the nitration is performed. When all the glycerine has been added, and the mixture has been shaken for a few minutes longer, it is poured into a separator, and allowed to stand for some time. It should, if the glycerine is a good one, have separated from the mixed acids in ten minutes, and the line of demarcation between the nitro-glycerine and the acid should be clear and sharp, neither should there be any white flocculent matter suspended in the liquid. The excess of acids is now drawn off, and the nitro-glycerine shaken once or twice with a warm solution of carbonate of soda, and afterwards with water alone. The nitro-glycerine is then drawn off into a weighed beaker, the surface dried with a piece of filter paper, and weighed; 100 parts of a good glycerine should yield about 230 of nitro-glycerine. A quicker method is to take only 10 c.c. of the glycerine, of which the specific gravity is already known, nitrate as before, and pour into a burette, read off the volume of nitro-glycerine in c.c. and multiply them by 1.6 (the specific gravity of nitro-glycerine), thus: 10 grms. gave 14.5 c.c. nitro-glycerine, and 14.5 x 1.6 = 23.2 grms., therefore 100 would give 232 grms. nitro-glycerine. The points to be noted in the nitration of a sample of glycerine are: the separation should be sharp, and within half an hour or less, and there should be no white flocculent matter formed, especially when the carbonate of soda solution is added.

Total Acid Equivalent. Mr G.E. Barton (Jour. Amer. Chem. Soc., 1895) proposes to determine thus: 100 c.c. of glycerine are diluted to 300 c.c. in a beaker, a few drops of a 1 per cent. solution of phenolphthalein and 10 c.c. of normal caustic soda solution are added; after boiling, the liquid is titrated with normal hydrochloric acid (fatty acids are thus indicated and roughly determined).

Neutrality. The same chemist determines the neutrality of glycerine thus: 50 c.c. of glycerine mixed with 100 c.c. of water and a few drops of alcoholic phenolphthalein[A] are titrated with hydrochloric acid or sodium hydroxide; not more than 0.3 c.c. normal hydrochloric acid or normal soda solution should be required to render the sample neutral; raw glycerines contain from .5 to 1.0 per cent. of sodium carbonate.

[Footnote A: Sulman and Berry prefer litmus as indicator.]

Determination of Free Fatty Acids. A weighed quantity of the glycerine is shaken up with some neutral ether in a separating funnel, the glycerine allowed to settle, drawn off, and the ether washed with three separate lots of water. The water must have been recently boiled, and be quite free from CO_{2}. All the free fatty acid is now in the ether, and no other soluble acid. A drop of phenolphthalein is now added, a little water, and the acidity determined by titration with deci-normal baryta solution, and the baryta solution taken calculated as oleic acid.

Combined Fatty Acid. About 30 grms. of the glycerine are placed in a flask, and to it is added about half a grm. of caustic soda in solution. The mixture is heated for ten minutes at 150 deg. C. After cooling some pure ether is added to it, and enough dilute H{2}SO{4} to render it distinctly acid. It is well shaken. All the fatty acids go into the ether. The aqueous solution is then removed, and the ether well washed to remove all H{2}SO{4}. After the addition of phenolphthalein the acid is titrated, and the amount used calculated into oleic acid. From this total amount of fatty acids the free fatty acid is deducted, and the quantity of combined fatty acids thus obtained.

Impurities. The following impurities may be found in bad samples of glycerine:—Lead, arsenic, lime, chlorine, sulphuric acid, thio-sulphates, sulphides, cyanogen compounds, organic acids (especially oleic acid and fatty acids[A]), rosin products, and other organic bodies. It is also said to be adulterated with sugar and glucose dextrine. Traces of sulphuric acid and arsenic may be allowed, also very small traces indeed of lime and chlorine.

[Footnote A: These substances often cause trouble in nitrating, white flocculent matter being formed during the process of washing.]

The organic acids, formic and butyric acids may be detected by heating a sample of the glycerine in a test tube with alcohol and sulphuric acid, when, if present, compound ethers, such as ethylic formate and butyrate, the former smelling like peaches and the latter of pine-apple, will be formed.

Oleic Acid, if present in large quantity, will come down upon diluting the sample with water, but smaller quantities may be detected by passing a current of nitrogen peroxide, N{2}O{4} (obtained by heating lead nitrate), through the diluted sample, when a white flocculent precipitate of elaidic acid, which is less soluble than oleic acid, will be thrown down. By agitating glycerol with chloroform, fatty acids, rosin oil, and some other impurities are dissolved, while certain others form a turbid layer between the chloroform and the supernatant liquid. On separating the chloroform and evaporating it to dryness, a residue is obtained which may be further examined.

Sodium Chloride can be determined in 100 c.c. of the glycerine by adding a little water, neutralised with sodium carbonate, and then titrated with a deci-normal solution of silver nitrate, using potassium chromate as indicator.

Organic Impurities of various kinds occur in crude glycerine, and are mostly objectionable. Their sum may be determined with fair accuracy by Sulman and Berry's method: 50 grms. of the sample are diluted with twice its measure of water, carefully neutralised with acetic acid, and warmed to expel carbonic acid; when cold, a solution of basic lead acetate is added in slight but distinct excess, and the mixture well agitated. The formation of an abundant precipitate, which rapidly subsides, is an indication of considerable impurity in the sample. To ascertain its amount, the precipitate is first washed by decantation, and then collected on a tared, or preferably a double counter-poised filter, where it is further washed, dried at 100 deg. to 105 deg. C., and weighed. The precipitate and filter paper are then ignited separately in porcelain, at a low red heat, the residues moistened with a few drops of nitric acid and reignited; the weight of the lead oxide deducted from that of the original precipitate gives the weight of the organic matter precipitated by the lead. Raw glycerines contain from 0.5 to 1.0 per cent.

Albuminous Matters. An approximate determination of the albuminous matters may be made by precipitating with basic lead acetate as already described, and determining the nitrogen by the Kjeldahl method; the nitrogen multiplied by 6.25 gives the amount of albuminous matter in the precipitate.

The Determination of Glycerine. The acetin method of Benedikt and Canton depends upon the conversion of glycerine into triacetin, and the saponification of the latter, and reduces the estimation of glycerine to an acidmetric method. About 1.5 grm. of crude glycerine is heated to boiling with 7 grms. of acetic anhydride, and 3 to 4 grms. of anhydrous sodium acetate, under an upright condenser for one and a half hours. After cooling, 50 c.c. of water are added, and the mixture heated until all the triacetin has dissolved. The liquid is then filtered into a large flask, the residue on the filter is well washed with water, the filtrate quite cooled, phenolphthalein is added and the fluid exactly neutralised with a dilute (2 to 3 per cent.) solution of alkali. Twenty-five c.c. of a 10 per cent. caustic soda solution, which must be accurately standardised upon normal acid, are then pipetted into the liquid, which is heated to boiling for ten minutes to saponify the triacetin, and the excess of alkali is then titrated back with normal acid. One c.c. of normal acid corresponds to .03067 grm. of glycerine.

Precautions.—The heating must be done with a reflux condenser, the triacetin being somewhat volatile. The sodium acetate used must be quite anhydrous, or the conversion of the glycerine to triacetyl is imperfect. Triacetin in contact with water gradually decomposes. After acetylation is complete, therefore, the operations must be conducted as rapidly as possible. It is necessary to neutralise the free acetic acid very cautiously, and with rapid agitation, so that the alkali may not be locally in excess.