air pump, or more simply by a stream of carbonic anhydride. The procedure is as follows:-A combustion tube (fig. 8), about one metre long, has placed in it next to its closed ends about 30 grains of hydric sodic carbonate (a b), then a layer of cupric oxide (bc), then the mixture of the weighed substance and cupric oxide (cd), a layer of cupric oxide (de) added, and the rest of the tube filled with copper turnings of clean metallic surface; a gas delivery tube is then attached by means of a perforated cork. The tube so filled (a b, fig. 9) is laid in the furnace, united to the bent delivery tube (ef), and the end of the latter dipped under mercury. One-half of the hydric sodic

carbonate is then heated to brisk decomposition; the evolved gases, carbonic anhydride and water vapour, expel the air contained in the apparatus, which escapes in bubbles through the mercury. From time to time samples of the escaping gas are collected in a test tube completely filled with potash solution and mercury; when such a sample is absorbed by the potash without leaving any residue, it is certain that the air has been completely expelled. A measuring tube filled with mercury, first having about 50-60 cc. of concentrated potash

solution run up into it, is placed over the end of the delivery tube and the combustion commenced. The cupric oxide at de and bc, and the metallic copper at e f (fig. 8), are heated to bright redness, and the heating gradually carried to the mixture of cupric oxide and substance. As soon as the whole tube between b and ƒ is red hot the evolution of gas ceases. All the organic substance is then burnt, and the greater part of the products of combustion collected in the measuring tube. The remaining portion of the hydric sodic carbonate is then heated, in order to displace any nitrogen remaining in the tube. After some time, when the complete absorption of the carbonic anhydride by the potash solution is quite certain, a small porcelain capsule is placed

under the open end of the measuring tube, and the tube and contents removed to a deep cylinder filled with distilled water, the capsule being then removed. The mercury and potash sink out of the tube and are replaced by water. The measuring tube is now sunk so far into the cylinder that the surface of the water stands at the same level inside and outside the tube. It is then allowed to stand until both gas and water have acquired the temperature of the room; the volume of the gas in cubic centimetres. (v) is then read off the tube, and at the same time the temperature (t C.) and the barometer (b in mm.) noted. The weight of nitrogen obtained is then, with correction for the tension of aqueous vapour (w in mm.)—

16. Estimation of Nitrogen as Ammonia.-As already explained, most organic bodies, when heated with alkaline hydrates, give off all their nitrogen as ammonia. The analysis is usually conducted as follows: The substance to be analysed is mixed with five to six times its weight of soda lime in a warm porcelain mortar, and then transferred

to a combustion tube which has already been drawn out to an oblique point and partly filled with soda lime; the mortar is then washed out with fresh soda lime and the tube nearly filled therewith, a plug of asbestos put lightly in, and a bulb apparatus containing hydrochloric acid fixed on by means of a well-fitting cork. The tube a (fig. 10) is then heated from the anterior backwards. Ammonia is evolved, which is absorbed by the acid, being converted into ammonic chloride. Free hydrogen and hydrocarbons are generally also evolved, the latter colouring the liquid in the bulb. At the close of the combustion, when the mass, which blackens at first, must have become quite white again, the point of the tube is broken off and air sucked through in order to bring all the ammonia into the bulb. This latter is then removed, the contents emptied into a porcelain basin, and the bulb rinsed repeatedly with pure water. The acid and wash waters are then mixed with excess of platinic chloride and the whole evaporated to dryness on the water bath. The residue consists of ammonic platinic chloride, N,H,Cl,Pt, together with excess of platinic chloride. This latter is dissolved out by a mixture of alcohol and ether, and the insoluble ammonic platinic chloride collected on a dried and weighed filter, completely washed with ether alcohol, and dried at

8 6

100° until constant in weight. Every 100 parts of the double salt contain 6-28 parts of nitrogen. The double salt on ignition leaves a residue of platinum, which can be weighed; every atom of platinum found corresponds to two atoms of nitrogen, or 100 parts of platinum obtained corresponds to 14:172 parts of nitrogen.

Many organic substances, when heated with alkalies, give their nitrogen-at least in part-not as ammonia but in combination with carbon and hydrogen, in the form of organic bases, whose hydrochlorides give with platinic chloride compounds resembling the ammonic double salt, in which there is still one atom of platinum to two of nitrogen. It is, therefore, advisable always not only to weigh the double salt, but also the platinum left on ignition, the nitrogen being finally calculated from the latter.

166. Instead of estimating by precipitating as above, the contents of the bulb may be evaporated to dryness and the chlorine in the residual ammonic chloride determined by addition of a standard solution of argentic nitrate; one atom of nitrogen must have been present for every atom of chlorine found.

Estimation of the Halogens.

17. In all cases where organic halogen compounds are not directly decomposed by water with complete conversion of the halogen into its hydro-acid (as happens with the chlorine, bromine, &c., compounds of the organic acid radicals), the halogens must be rendered capable of precipitation with argentic nitrate by the action of energetic reagents.

This can frequently be accomplished by the action of sodium amalgam upon an aqueous solution of the organic body, é.g. monochloracetic acid is decomposed according to the equation:

C2H ̧CIO2+Na,=C2H ̧NaO2+ NaCl,

the chlorine being then readily precipitated from the sodic chloride by addition of silver salts.

In cases where this method is not available the decomposition can be effected by heating with pure lime in a combustion tube. The apparatus is quite similar to that used in estimating nitrogen as ammonia, only no absorption tube is required. The ignited residue is dissolved in nitric acid, precipitated with silver solution, the silver salt collected on a filter, washed, ignited, and weighed. As in this operation a large excess of lime must be employed a considerable volume of liquid is obtained, which makes the filtration a very wearisome operation.

Latterly this method has been nearly entirely superseded by another, which consists in the complete oxidation of the organic body by nitric acid at high temperatures. In many cases this reagent can be used dilute (1-2-1-3 sp. gr.), and acts completely even at 130°-150°; very few organic bodies can resist the action of fuming nitric acid at 180°-200°. The employment of temperatures so much above the boiling point of nitric acid necessitates the use of strong walled, sealed tubes, in order to resist the high pressures produced.

The substance for analysis is sealed up in thin glass bulbs, usually with bent, and therefore easily broken, capillary tubes. It is then

FIG. 11.

enclosed, together with excess of argentic nitrate and of moderately concentrated, or if necessary of fuming, nitric acid, in a strong walled glass tube, care being taken that in sealing the drawn-out point is well thickened. Fig. 11 shows the whole apparatus. The glass bulb is now broken by vigorous shaking, the enclosed substance coming in contact with the nitric acid, and the tube is then heated for several hours in an air bath, heated to the necessary temperature. The air bath usually employed is shown in fig. 12; it contains several wrought iron tubes closed at one end, and which are placed in an inclined position in the bath. If one of the glass tubes explodes, it cannot destroy the other tubes being heated in the bath at the same time, nor can it damage the operator.

When the reaction is completed the apparatus is allowed to cool, the glass tube carefully removed, wrapped in a cloth, except the capillary tube, and the latter softened in a flame. The internal pressure blows the point out, and the strongly compressed gases (carbonic anhydride and reduction products of nitric acid) are evolved without danger. The glass tube is then cut open, the contents completely washed into a beaker, and after sufficient dilution of the nitric acid, the haloid silver salt, and the fragments of the glass bulb collected on a filter, washed, dried, ignited, and weighed. By subtraction of the weight of the filter, ash, and glass bulb the weight of the chloride, bromide, or iodide of silver is obtained, from which the weight of halogen can be calculated.

Estimation of Sulphur and Phosphorus.

18. Many sulphurised organic bodies can be completely oxidised by nitric acid, with conversion of their sulphur into sulphuric acid, the

method above described being employed, the sulphuric acid being precipitated as baric sulphate from the largely diluted oxidation product. If halogens are estimated in the same operation the filtrate from the silver precipitate must be treated with hydrochloric acid, so as to remove all silver before precipitating the sulphuric acid.

If the substance oxidised by nitric acid contained phosphorus, arsenic, or metals, they will be converted respectively into phosphoric acid, arsenic acid, or metallic nitrates, and can be estimated according to the usual methods of quantitative analysis.

Frequently organic sulphur compounds resist the action of nitric acid, so that their decomposition can only be effected by fusion with nitre and an alkaline carbonate. The mixture must be made in such proportions that there shall be no explosion on heating. Volatile substances are generally partly oxidised in a sealed glass tube by nitric acid, whereby non-volatile bodies are formed; the acid liquid is then more than saturated with dry alkaline carbonate, and the whole evaporated to dryness and fused in a platinum dish. As soon as all carbon is burnt, and the mass has become completely white, it is allowed to cool, dissolved in water, and the sulphuric acid precipitated.

Estimation of Oxygen.

19. Oxygen is usually estimated indirectly; the sum of the percentages of the other elements present is deducted from one hundred, and the difference taken as oxygen. For instance, the combustion of 0.3 gram of cane sugar, which contains carbon, hydrogen, and oxygen only, gave 0-463 gram carbonic anhydride and 0·170 gram water. This, calculated into percentages, gives

DETERMINATION OF THE CHEMICAL FORMULE OF ORGANIC COMPOUNDS.

20. The results of the elementary analysis of an organic body are always expressed in percentages; from these empirical formulæ can be obtained by dividing the percentage number by the atomic weight of the element to which it refers, the quotients then expressing the proportions in which the atoms of carbon, hydrogen, oxygen, &c., are united to each other. These quotients, generally consisting of fractions, are reduced to the nearest round numbers, and then represent the empirical formula of the body.

A

In chemical analysis, as generally in any operation involving measuring or weighing, there are slight deviations from accuracy. formula so obtained is still admissible, although the percentage numbers calculated from it do not quite agree with those found in the investigation, provided that the deviation does not exceed a certain limit. This limit of error is found to be about 0.3 per cent. for carbon and 0.2 per cent. for hydrogen; for instance, in the combustion of pure glacial acetic acid the following numbers were obtained :