FRACTIONAL CRYSTALLIZATION, PRECIPITATION AND SOLUTION. 9

often more valuable for ascertaining the purity of the body than the use of ordinary chemical reagents.

(1040) It frequently happens that two or more substances, of different degrees of solubility in the same menstruum, occur together: in such a case, if crystallizable, they may often be separated by repeated crystallizations, care being taken always to separate those crystals which are first deposited, from those which are formed subsequently: by successive recrystallizations of the first portions from each crop, the least soluble compound may be obtained in a state of purity. In other cases, if the different compounds yield precipitates with the same reagent, but differ in the degree of their attraction for it, a separation by fractional precipitation may be effected by adding a quantity of the precipitant insufficient to throw down, for example, more than onehalf of the amount of the substances that may be present. The salts of silver or of lead may thus be added to mixtures of various organic acids, particularly to such as, like the fatty or the resinous acids, greatly resemble each other in properties. The precipitate is then suspended in water, and decomposed by means of sulphuretted hydrogen, or a dilute mineral acid, in order to separate the metal, and liberate the acid, which is again submitted to a similar incomplete precipitation; and this process is repeated until the fusing point becomes constant, or till some other guarantee of the purity of the body is obtained.

In other cases, where a mixture of two solids of unequal solubility in the same menstruum is operated on, fractional solution may be employed. In this case, by using successive small portions of the same solvent, a separation of the more soluble ingredient from the less soluble one may be gradually effected, till at length the last portions which remain undissolved consist of the least soluble ingredient nearly pure.

In the rectification of mixtures of volatile liquids, analogous principles are acted upon; in these cases the boiling point is the guide. The liquid is placed in a retort, into the tubulure of which a thermometer is fitted, and the temperature is raised till the liquid boils. So long as the temperature of the boiling liquid continues to rise, the substance which distils must consist of a mixture; but as soon as the point of ebullition becomes nearly stationary, the distillate may be supposed to have a composition nearly uniform; and by rectifying again those portions which distil at the same fixed temperatures, the different components of a mixed liquid may be obtained in a state of purity. Where

10

FRACTIONAL DISTILLATION.

large quantities can be operated on, this process of fractional distillation is well adapted to the separation of liquids the boiling points of which differ considerably from each other; but it is much less satisfactory when the quantities are small, or where the boiling points are within 20° or 30° (10° or 15° C.) of each other.

A considerable improvement in the mode of conducting fractional distillation has recently been introduced by Warren, which he has applied very successfully in his examination of the volatile products of American petroleum. (Memoirs of the American Academy, New Series, vol. ix.) This mode of distillation will be understood from an examination of fig. 377, in which a shows

the retort or two-necked flask containing the liquid for distillation and a thermometer f; it is connected with an ascending spiral worm b, which is maintained at a steady temperature by passing through the oil bath c, regulated by the aid of the thermometer, t. d is a vessel ordinarily empty, but which can, if necessary, be filled with ice. The upper end of the worm e is connected with a Liebig's condenser or other suitable apparatus for condensation. By this arrangement it is easy to ensure the condensation of nearly all the less volatile portions, which flow back into the retort,

FRACTIONAL SATURATION-DESICCATION-ULTIMATE ANALYSIS. 11

whilst the more volatile portions pass on and are obtained in the colder part of the apparatus. The whole should be made of brazed copper, the worm about 5 fcet, or 15 metre long, and inch or 10 millimetres in internal diameter.*

Liebig has applied a method, which may be termed fractional saturation, to the separation of certain volatile acids from each other; and it may be employed generally in such cases with advantage. This process consists in saturating a portion only of the acid with potash or with soda, and then proceeding with the distillation; the more volatile acid passes over with the distillate. Suppose that it be desired to effect the separation of butyric from valeric acid :—a certain amount of potash is added, and if this be more than sufficient to saturate the valeric acid, the distillate will consist of butyric acid only, while the residue will be a mixture of potassic valerate and butyrate. If, on the other hand, the valeric acid be more than sufficient to saturate the potash, the residue in the retort will consist of pure potassic valerate, and the distillate will be a mixture of butyric and valeric acids, which may be separated by a repetition of the same process.

(1041) B. ULTIMATE ANALYSIS.-Before proceeding to the ultimate analysis of a body, it is necessary first to dry it carefully, either in the water oven (1039), or by the means of an apparatus similar to that shown in fig. 378, in which a represents a bent glass tube, for the reception of the substance to be dried; B, is a vessel of glass or of metal, containing water, or some saline solution; c, an aspirator, by which a current of atmospheric air can be maintained through the apparatus; and D, a

Wanklyn has shown (Proceed. Roy. Soc. xii. 534) that the proportions in which the mixed liquids distil over must vary not only with the relative amount of each in the mixture, and with the tension of the vapour of each at the boiling point of the mixture, but also with the adhesion of the liquids to each other and with the vapour density of each. In most cases the effect of adhesion is trifling, but the variation in vapour density is important. According to his experiments, a mixture of equal weights of two liquids of different boiling points will furnish a distillate in which the proportion of each liquid may be calculated by the product of the tension of its vapour into the specific gravity of the vapour. In consequence of this circumstance, the less volatile liquid may distil over more rapidly than the other; for example, a mixture containing 18 parts of wood spirit (boiling at 151°, vapour density 1107) and 17 parts of ethylic iodide (boiling at 1616, vapour density 5'397) yielded in the first third a mixture consisting of 6 parts of wood spirit and 87 of ethylic iodide. If the vapour density and the tension at the boiling point of the mixed liquid are inversely proportional in the case of the two liquids, the mixture distils over without change. These observations explain the facility with which bodies of high boiling but of considerable vapour density, such as the essential oils, pass over with the vapour of water. Berthelot has made similar observations.

12

DETERMINATION OF MINERAL COMPONENTS.

bent tube filled with fragments of pumice-stone, soaked in oil of vitriol, for drying the air as it enters. The desiccation must be con

FIG. 378.

tinued so long

as the tube A

ceases to lose weight. If the desiccation be

effected in the

covered whilst being weighed, in order to prevent the absorption of moisture.

(1042) Determination of the Mineral Components.-If any mineral substances be present in the compound under examination, the amount of these must be ascertained by the incineration of a weighed quantity of the material. This operation, simple as it may appear, is one, the exact performance of which is attended with considerable difficulty, which arises from the partial volatilization of the alkaline chlorides at a red heat, and the reduction of the sulphates to the form of sulphides. The incineration should therefore always be performed at a low temperature, which need not exceed a barely visible red, but must be prolonged for many hours if the mass for incineration be considerable. Caillat finds it advantageous in a large number of cases, before incinerating to treat the substance with dilute nitric acid, by which means almost the entire saline ingredients may be extracted in the soluble form; on incinerating the insoluble portion, scarcely anything is left except silica and a small quantity of ferric oxide. The quantity of saline matters thus obtained always exceeds that furnished by the former method; since in the digestion with nitric acid the loss occasioned by reduction of the calcic sulphate, and the partial expulsion of sulphur from the calcic sulphide so formed is avoided; the proportion of sulphuric acid obtained is therefore always higher when nitric acid is used as a pre

liminary to incineration than when simple incineration without the use of acid is practised.

(1043) Analysis of Organic Compounds by Combustion.-Since the four elements, carbon, hydrogen, oxygen, and nitrogen, constitute the bulk of most organic substances, the operation of organic analysis resolves itself mainly into the accurate quantitative determination of these elements. We are indebted to Gay Lussac and Thénard for the fundamental principle that regulates our operations in this respect. The process proposed by them has subsequently been modified and improved by many chemists, especially by Berzelius, Prout, and Liebig, and it was by the last named eminent philosopher that the method now in general use was contrived.

The object of this operation being to determine the relative proportion in which each of the ultimate elements exists, it becomes necessary to the success of any analytical process that these elements should be procured either separately in a state of purity, or in the form of definite compounds that can easily be collected. In practice it has been found most convenient to supply the substance to be analysed with a quantity of oxygen sufficient to convert the carbon which it contains into carbonic anhydride, and the hydrogen into water; arrangements having been made for absorbing the carbonic anhydride by means of a known weight of a solution of potash; whilst the water, by passing it over a substance that has a powerful attraction for it, such as chloride of calcium or puraice moistened with sulphuric acid, is collected and weighed; the nitrogen is collected and measured in the form of gas, over mercury.

In certain cases where nitrogen is present, the substance to be analysed may be mixed with caustic soda or potash, and heated; all the nitrogen is thus converted into ammonia, in which form, like carbonic anhydride and water, it admits of being weighed. It is easy to find by calculation the weight of the carbon, the hydrogen, and the nitrogen respectively contained in the carbonic anhydride, the water, and the ammonia collected. When the proportion of saline matter has been determined by incineration of a portion of the mass, the quantity of oxygen which the substance contains may be known by deducting the united weight of the carbon, the hydrogen, the nitrogen, and the salts, from the total weight of the body analysed; the deficiency (supposing sulphur and phosphorus not to have been present) is reckoned as

oxygen.

Scrupulous attention to the purity of the matter submitted to